_2001_99
TRANSCRIPT
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 1/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 2/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 3/250
October 27, 2000
Dear Colleagues:
The Food and Drug Administration (FDA) has been pleased to cooperate with the
International Society for Pharmaceutical Engineering in the development of the Baseline®
Pharmaceutical Engineering Guide for Water and Steam Systems. We appreciate the
coopertive efforts and the dedicated intensive work of the engineers who voluntarily
initiated the development of this Guide. This is an excellent example of how through public and private cooperative efforts, both industry and consumers can benefit.
This document covers engineering aspects of design, construction and operation of new
water and steam systems. It expands on existing FDA guidance on water systems.
This Guide is solely created and owned by ISPE. It is not an FDA regulation, standard or
guidance document and water and steam systems built in conformance with this Guide
may or may not meet FDA requirements. FDA has provided comments for ISPE's
consideration in preparing this Guide. It should be helpful to the engineering profession
and the industry for the design, construction and operation of new water and steam
systems
FDA is pleased with the development of this document and we look forward to a contin-
ued working relationship as future Baseline® Pharmaceutical Engineering Guides are
developed.
Sincerely,
Janet Woodcock, M.D.
Director, Center for Drug Evaluation and Research
Dennis Baker
Associate Commissioner for Regulatory Affairs
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 4/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 5/250
3
WATER AND STEAM SYSTEMS
ISPE PHARMACEUTICAL ENGINEERING GUIDE
FOREWORD
For many years, the pharmaceutical industry has experienced increases in the cost of new facilities. These
increases in cost have been driven in part by uncertainty about the requirements for regulatory compliance.Some significant areas of concern are validation, particularly related to automation systems, and the trend tovalidate back to source utilities. The absence of a consistent and widely accepted interpretation of some
regulatory requirements has led to one-upmanship. This practice of building increasingly technically ad-vanced facilities has led to increased cost, longer lead times and, in some cases, delays in bringing new
products to market.
In May 1994, engineering representatives from the pharmaceutical industry engaged in a discussion with theInternational Society for Pharmaceutical Engineering (ISPE) and the Food and Drug Administration (FDA).As a result of that discussion in November 1994, ISPE began work on nine facility engineering Guides, now
known as the Baseline ® Pharmaceutical Engineering Guides. The first, “Bulk Pharmaceutical Chemicals,”was published in June 1996. The second, “Oral Solid Dosage Forms,’” was published in February 1998. The
third, “Sterile Manufacturing Facilities,” was published in February 1999. This is the fourth such Guide, cover-
ing Pharmaceutical Water and Steam Systems. Each Engineering Guide was created by, and is owned solelyby ISPE. FDA provided comments on this and previous Guides, and many of their suggestions have beenincorporated.
As with the BPC Guide, OSD, and Sterile Guide, the Water and Steam Systems Guide has been sponsoredby ISPE’s Pharmaceutical Advisory Council, made up of senior pharmaceutical engineering executives from
owner companies, and ISPE senior management. Overall planning, direction, and technical guidance in thepreparation of the Water and Steam Systems Guide was provided by a Steering Committee most of whom
were involved in the BPC Guide. The Water and Steam Systems Guide itself was produced by a Task Team ofindividuals who expended a great deal of their own time in its preparation and development.
The Water and Steam Systems Appendix contains material considered “informational” which, although nec-essary, would have been detrimental to the clarity of the dedicated chapter. The Appendix has not been
reviewed by and therefore is not endorsed by the FDA.
Editors’ Disclaimer: This Guide is meant to assist pharmaceutical manufacturers in the design and construction of new and renovated facilities that are required to comply with the requirements of the Food and Drug
Administration (FDA). The International Society for Pharmaceutical Engineering (ISPE) cannot en- sure, and does not warrant, that a facility built in accordance with this Guide will be acceptable to
FDA.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 6/250
4
WATER AND STEAM SYSTEMS
ACKNOWLEDGEMENTS
CHAPTER WRITERS AND REVIEWERS
The following individuals took lead roles in the preparation of this document:
Gerald L. Geisler, Bristol-Myers Squibb Co. was the Task Team Chairperson for the Water and SteamSystems Guide. Moe Elmorsi, Boehringer Ingelheim, acted as the Guide mentor.
Technical Documents Steering Committee Chairperson
Paul Lorenzo, (Retired)/ Paul D’Eramo, Johnson & Johnson
The Core Team on the Water and Steam Systems Guide comprised:
Gerald L. Geisler, Bristol-Myers Squibb Co. Jeff Biskup, Clark, Richardson & Biskup
Robert Myers, Kvaerner Bob Bader, Kinetics
The Chapter Credits are as follows:
Gerald Geisler, Bristol-Myers Squibb Co. Chapter 1: Introduction
Brian Owens, H2O Pure, Inc.
Gerald Geisler, Bristol-Myers Squibb Co. Chapter 2: Key Design PhilosophiesBrian Owens, Water Pure, Inc.
Jeffrey Biskup, Clark, Richardson & Biskup Chapter 3: Water Options and ProgrammingMaria Capote, Source Tech
James C. Cox, Merck & Co.Gerald L. Geisler, Bristol-Myers Squibb Co.
Ryan Schroeder, Clark, Richardson & Biskup
Sidney Brookes, DuPont Merck Pharmaceuticals Chapter 4: Source Feed Water Quality andPretreatment
Michael Partow, Pfizer Inc. Chapter 5: Final Treatment Non-CompendialAndrew Zaske, Osmonics and Compendial Purified Water
Gary Zoccolante, U.S. Filter
Sharif Disi, Meco Chapter 6: Final Treatment Compendial WFIBrian Owens, H2O Pure, Inc.
Brian Owens, H2O Pure, Inc. Chapter 7: Pharmaceutical SteamBob Bader, Kinetics
Robert Myers, Kvaerner Chapter 8: Storage and Distribution Systems
Gary Gray, East GroupBob Bader, KineticsRandolph Brozek, Pfizer, Inc.
James Cox, Merck & CoPaul Skinner, Clark, Richardson & Biskup
Gerald Geisler, Bristol-Myers Squibb Co.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 7/250
5
WATER AND STEAM SYSTEMS
John Linder, CE & IC Chapter 9: Instrumentation and ControlDebra Nahas, Eli Lilly & Co.
John Fadool, Glaxo Wellcome Chapter 10: Commissioning & Validation
Robert Myers, KvaernerManfred Septinus, Roche Carolina, Inc.
Phil DeSantis, Fluor Daniel
Dominick Smith, Regeneron Chapters 11 and 12: Appendices
Phil Desantis, Fluor DanielSidney Brooks, DuPont Merk Pharmaceuticals
James C. Cox, Merck & Co.Sharif Disi, Meco
Brian Owens, Water Pure, Inc.Michael Partow, PfizerPaul Skinner, Clark, Richardson & Biskup
Gary P Zoccolante, U. S. FilterPat H. Banes, Oakley Services Co.
The above Guide Task Team worked on one or more chapters and volunteered countless hours to attendmeetings, and review the many drafts, which were prepared over an 18-month period.
The following members of the Water and Steam Systems Task Team also worked on one or more of thechapters and reviewed drafts:
Georgia Keresty, Ph.D., Bristol-Myers Squibb Paula Soteropoulis, Genzyme Corp.
Alex Konopka, Eli Lilly & Co. John Trentacosti, Johnson & Johnson
Carl Roe, Abbott Labs
FDA ReviewersWe would like to thank the following FDA review team for their input to this Guide:
Sharon Smith Holston (Deputy Commissioner for External Affairs)
Joseph Phillips (Deputy Regional Food and Drug Director, Mid-Atlantic Region)
Tracy Roberts (CDER, Office of Compliance)
Robert Coleman (Atlanta National Expert)
Richard Friedman (CDER, Office of Compliance)
Nancy Rolli (Investigator, Drug Specialist, New Brunswick, NJ Inspection Post)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 8/250
6
WATER AND STEAM SYSTEMS
Also, ISPE acknowledges with gratitude the following companies who supplied the start-up funding
for this project:
Alcon Laboratories Bayer Corp. Boehringer Ingelheim
Bristol-Myers Squibb Co. Eli Lilly & Co. Glaxo Wellcome Inc.
Hoffmann-La Roche Inc. Merck & Co., Inc. Pfizer Inc.
Pharmacia & Upjohn Inc. Wyeth-Ayerst Laboratories Zeneca Pharmaceuticals
Zenith Goldline Pharmaceuticals
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 9/250
7
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ........................................................................................................................... 4
1. INTRODUCTION
1.1 BACKGROUND..................................................................................................................... 111.2 SCOPE OF THIS GUIDE ...................................................................................................... 11
1.3 SOME APPLICABLE FDA CURRENT REGULATIONS AND GUIDES FORPHARMACEUTICAL WATER SYSTEMS .............................................................................. 111.4 KEY CONCEPTS .................................................................................................................. 12
1.5 GUIDE STRUCTURE ............................................................................................................ 13
2. KEY DESIGN PHILOSOPHIES
2.1 INTRODUCTION................................................................................................................... 152.2 UNITED STATES PHARMACOPOEIA (USP) ........................................................................ 152.3 SPECIFICATION OF PHARMACEUTICAL WATER QUALITY .............................................. 20
2.4 CRITICAL PROCESS PARAMETERS .................................................................................. 212.5 CGMP COMPLIANCE ISSUES............................................................................................. 21
2.6 DESIGN RANGE VERSUS OPERATING RANGE ................................................................ 22
3. WATER OPTIONS AND SYSTEM PLANNING
3.1 INTRODUCTION................................................................................................................... 253.2 WATER QUALITY OPTIONS................................................................................................. 253.3 SYSTEM PLANNING ............................................................................................................ 29
3.4 SYSTEM DESIGN ................................................................................................................. 33
4. PRETREATMENT OPTIONS
4.1 INTRODUCTION................................................................................................................... 354.2 PROCESS DESIGN OF PRETREATMENT .......................................................................... 354.3 FEEDWATER TO PRETREATMENT QUALITY: TESTING AND DOCUMENTATION ............ 37
4.4 OUTPUT WATER FROM PRETREATMENT: QUALITY OF FEEDWATER TOFINAL TREATMENT .............................................................................................................. 38
4.5 CONTROL OF FOULING: REMOVAL OF TURBIDITY AND PARTICULATES ...................... 394.6 CONTROL OF SCALING: REMOVAL OF HARDNESS AND METALS ................................. 39
4.7 REMOVAL OF ORGANICS ................................................................................................... 404.8 SYSTEM DESIGN FOR CONTROL OF MICROBIAL GROWTH .......................................... 414.9 REMOVAL OF MICROBIAL CONTROL AGENTS................................................................. 42
4.10 CHANGES IN ANION COMPOSITION / CONCENTRATION................................................ 424.11 THE IMPORTANCE OF PH IN PRETREATMENT................................................................. 43
4.12 MATERIALS OF CONSTRUCTION AND CONSTRUCTION PRACTICES............................ 434.13 PRETREATMENT SUMMARY .............................................................................................. 44
5. FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.1 INTRODUCTION................................................................................................................... 475.2 ION EXCHANGE................................................................................................................... 48
5.3 CONTINUOUS ELECTRODEIONIZATION (CEDI) ............................................................... 515.4 REVERSE OSMOSIS ........................................................................................................... 53
5.5 POLISHING COMPONENTS - NON-IONIC CONTAMINANTS REDUCTION....................... 56
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 10/250
8
TABLE OF CONTENTS
6. FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.1 INTRODUCTION................................................................................................................... 636.2 US PHARMACOPOEIA ISSUES ........................................................................................... 63
6.3 DISTILLATION ...................................................................................................................... 646.4 DISTILLATION APPLICATIONS AND CAPACITIES .............................................................. 64
6.5 PROCESS AND SYSTEM DESCRIPTION ........................................................................... 656.6 REVERSE OSMOSIS (RO) ................................................................................................... 706.7 USP - WATER FOR INJECTION SYSTEMS COMPARISON ................................................ 73
7. PHARMACEUTICAL STEAM
7.1 INTRODUCTION................................................................................................................... 75
7.2 CGMP ISSUES ..................................................................................................................... 757.3 TYPES OF STEAM ............................................................................................................... 767.4 BACKGROUND AND INDUSTRY PRACTICES .................................................................... 77
7.5 SYSTEM PLANNING ............................................................................................................ 807.6 PHARMACEUTICAL STEAM PURITY DECISION TREE...................................................... 82
7.7 PROCESS AND SYSTEM DESCRIPTION ........................................................................... 83
7.8 SIZING THE CLEAN STEAM SYSTEM................................................................................. 867.9 COST IMPLICATIONS .......................................................................................................... 907.10 STEAM “QUALITY” ............................................................................................................... 90
7.11 DISTRIBUTION ..................................................................................................................... 907.12 FOUR EXAMPLES OF CORRECT PIPING PRACTICE ....................................................... 93
8. STORAGE AND DISTRIBUTION SYSTEMS
8.1 INTRODUCTION................................................................................................................... 958.2 SYSTEM DESIGN ................................................................................................................. 95
8.3 SYSTEM DISTRIBUTION DESIGN....................................................................................... 968.4 MATERIALS OF CONSTRUCTION..................................................................................... 1138.5 SYSTEM COMPONENTS ................................................................................................... 115
8.6 COMPARISON OF WFI SYSTEMS WITH STORAGE TANK AND WITHOUTSTORAGE TANK ................................................................................................................. 117
8.7 MICROBIAL CONTROL DESIGN CONSIDERATIONS ....................................................... 1198.8 CONTINUOUS MICROBIAL CONTROL ............................................................................. 122
8.9 PERIODIC STERILIZATION/SANITIZATION....................................................................... 1248.10 SYSTEM DESIGN FOR STERILIZATION/SANITIZATION .................................................. 126
9. INSTRUMENTATION AND CONTROL
9.1 INTRODUCTION................................................................................................................. 1299.2 PRINCIPLES ....................................................................................................................... 129
9.3 GENERAL INSTRUMENTATION REQUIREMENTS ........................................................... 1309.4 DESIGN CONDITIONS VERSUS OPERATING RANGE .................................................... 1349.5 INSTRUMENTATION SPIKES ............................................................................................. 135
9.6 CONTROL SYSTEMS......................................................................................................... 135
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 11/250
9
10. COMMISSIONING AND QUALIFICATION
10.1 INTRODUCTION................................................................................................................. 13710.2 SYSTEM QUALIFICATION DOCUMENTATION .................................................................. 137
10.3 SYSTEM QUALIFICATION SAMPLING PROGRAM ........................................................... 13810.4 ACCEPTANCE CRITERIA................................................................................................... 140
10.5 QUALIFICATION REPORTS ............................................................................................... 14110.6 CHANGE CONTROL AND REQUALIFICATION ................................................................. 141
11. APPENDIX
11.1 USP REGULATED WATER QUALITY ................................................................................. 14311.2 EUROPEAN PERSPECTIVE .............................................................................................. 151
11.3 PASSIVATION ..................................................................................................................... 15311.4 PRETREATMENT PROCESSES ........................................................................................ 16211.5 FINAL TREATMENT FOR NON-COMPENDIAL AND COMPENDIAL PURIFIED
WATER SYSTEMS .............................................................................................................. 17511.6 DISTILLATION FOR HIGH PURITY WATER SYSTEMS ..................................................... 182
11.7 CLEAN STEAM - CLEAN STEAM GENERATORS ............................................................. 187
11.8 MICROBIAL CONTROL BASICS, TESTING, AND STERILIZATION/SANITIZATIONEQUIPMENT DESIGN AND INSTALLATION ISSUES ........................................................ 190
11.9 FABRICATION/PROCEDURES FOR DISTRIBUTION SYSTEMS ...................................... 198
11.10 DESIGN OF WFI/PURIFIED WATER DISTRIBUTION SYSTEM ......................................... 20511.11 FABRICATION OF A WFI/PURIFIED WATER DISTRIBUTION SYSTEM............................ 20611.12 ABBREVIATIONS AND DEFINITIONS ................................................................................ 213
TABLE OF CONTENTS
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 12/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 13/250
INTRODUCTION
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 14/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 15/250
11
INTRODUCTION
1. INTRODUCTION
1.1 BACKGROUND
The design, construction, and validation (commissioning and qualification) of water and steam systems for
the pharmaceutical industry represent key opportunities for manufacturers, engineering professionals, andequipment suppliers. These systems are required to meet current Good Manufacturing Practice cGMP regu-lations while remaining in compliance with all other governing codes, laws, and regulations.
The cost of bringing these systems on line is highly variable, owing to interpretation of regulatory require-
ments and overly conservative design approaches. This Guide is intended to offer a practical, consistentinterpretation, while still allowing flexibility and innovation.
This Guide was prepared by ISPE, with feedback from industry representatives from all areas and disci-plines, and comments provided by FDA. It reflects ISPE’s current thinking related to engineering of new water
and steam systems.
It is recognized that industry standards evolve, and this document reflects the understanding of them as of
the publication date.
1.2 SCOPE OF THIS GUIDE
This Guide is intended for the design, construction, and operation of new water and steam systems. It is
neither a standard nor a detailed design guide. The validation of water and steam systems, which comprisescommissioning and qualification activities, will not be discussed in-depth in this Guide, but is covered in the
Commissioning and Qualification Baseline ® Guide.
The purpose of this Guide is to focus on engineering issues, and provide cost effective water and steamsystems. Where non-engineering issues (e.g., microbiological topics) are covered, the information is includedto stress the importance of such topics and the impact they have on water and steam system design. Such
non-engineering topics, therefore, are not covered comprehensively, and specific advice from QA depart-ments and technical experts must be sought where technical input is required.
This Guide is intended primarily for regulatory compliance for the domestic United States (US) market, and
follows US standards and references. European and other non-US standards and references may be incor-porated in future revisions.
1.3 SOME APPLICABLE FDA CURRENT REGULATIONS AND GUIDES FOR PHARMACEUTICAL
WATER SYSTEMS
• Food and Drug and Cosmetic Act
• The United States Pharmacopoeia XXIV
• Title 21 CFR, Part 211
• FDA Guide To Inspections of High Purity Water Systems
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 16/250
12
INTRODUCTION
1.4 KEY CONCEPTS
The following key concepts covered in this Guide are:
a) Methodology for defining the required water quality and configuring a water delivery system.
b) Critical process parameters.
c) Good Engineering Practices.
d) Design Options
a) Methodology for defining the required quality and configuring a water delivery system:
Perhaps the most critical step in a new pharmaceutical water or steam system, from a regulatory as well
as technical and financial standpoint, is the specification of water or steam quality required. The specifi-cation established is likely to have a larger impact on lifecycle costs of the system than any of the subse-
quent design decisions. In addition, regulated industries must consider the costs of noncompliance and
water system failures. Therefore, it is essential for the designer to seek advice from the Quality unit andtechnical experts early in the process.
Once process water and/or steam requirements are determined, system design options need to beaddressed. This Guide presents alternative baseline water and steam system building blocks and asso-ciated advantages and disadvantages of each. These baseline building blocks are qualified relative to
such things as capital costs; feed water chemistry; product water quality; chemical handling; water con-sumption; energy consumption; outside service costs maintenance requirements; and chemical/micro-
bial/endotoxin removal performance.
Guide emphasis is on how the system design should be determined based on the quality of feed water;the design of the pretreatment and final treatment system; the storage and distribution system design;and operation/maintenance procedures.
The Guide aims to improve consistency of pharmaceutical water and steam quality throughout the indus-
try, as a result of system performance and reliability improvements. It also provides the user with alterna-tive basic system building blocks to permit reliable and consistent generation of the required water or
steam quality.
b) Critical process parameters
Critical parameters are defined as those parameters that directly affect the product quality. For example,
since microbial quality cannot be directly monitored in real time, the parameters relied upon to controlmicrobial growth are normally considered critical. These may include temperature; UV intensity; ozone
concentration; circulating systems under positive pressure; etc. In regard to chemical purity, the qualityattributes themselves (properties of water produced), may be monitored at or after each process step,and the proper performance of that operation confirmed directly. For a system producing compendial
water, properties mandated in the official monograph obviously constitute critical parameters.
Critical instruments are those instruments that measure critical quality attributes. This concept is dis-cussed in Chapter 2 and used as a basis for subsequent chapter discussions where appropriate.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 17/250
13
INTRODUCTION
c) Good Engineering Practice (GEP)
GEP recognizes that all systems in a facility, whether they are water systems, steam systems, elevators,process reactors, safety valves, or rest rooms, require some form of commissioning and/or qualification.
Nearly all systems require documentation, inspection, and field testing. Good Engineering Practice capi-talizes upon this practice suggesting that manufacturers engage all stakeholders (engineers, operators,
Quality Assurance, and others) very early in the planning, design, construction, commissioning/qualifica-tion phases to ensure that systems are documented only once.
d) Design Options
The Guide emphasizes that a water system can be designed in many different ways, yet meet the overallrequirements of the system. It encourages a well-thought-out, planned approach to the design with input
from many areas of the organization including Quality Assurance.
1.5 GUIDE STRUCTURE
The structure of the Guide is shown in Figure 1-1 below. The chapters have been organized to assist in a
logical decision process to determine the type of water required and the system design needed to provide it.
Figure 1-1 Pharmaceutical Water and Steam Baseline ® Guide Structure
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 18/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 19/250
KEY DESIGN PHILOSOPHIES
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 20/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 21/250
15
KEY DESIGN PHILOSOPHIES
2. KEY DESIGN PHILOSOPHIES
2.1 INTRODUCTION
Pharmaceutical water is the most widely used ingredient in drug manufacturing and the main component in
equipment/system cleaning. Therefore, systems for the production of pharmaceutical water constitute a keycomponent in every manufacturing facility. The nature of producing pharmaceutical waters is to minimize oreliminate potential sources of contamination. This Guide considers this and the means by which engineers
can design out, or ensure control of the risk.
The quality of Pharmaceutical Water and Steam is not only critical from a regulatory point of view, but alsofrom a financial point of view. The Pharmaceutical Water and Steam specification has the largest impact on
lifecycle costs of the system.
It must be demonstrated that all pharmaceutical waters (non-compendial and United States Pharmacopoeia
(USP) monograph compendial waters) can be produced consistently to specification. Establishing the level ofmicrobial control needed in a pharmaceutical water and steam system used in the manufacture of a non-
sterile product requires an understanding of both the use of the product and the manufacturing process.
Manufacturers need to define the appropriate water purity based upon sound process understanding andsystem equipment capability. They must determine the specific purification capability for each processing
step, the limitations of the unit operation, and the critical parameters, which affect the specified water/steamquality - chemically, physically, or biologically. Expert QA advice should be sought to provide further detailsabout this important area.
USP covers two compendial water qualities (USP Purified Water and USP Water for Injection). This Guide
supports both these water qualities plus additional non-compendial waters including “Drinking Water”. It iscommon practice to name non-compendial waters (exclusive of “Drinking Water”) used in pharmaceutical
manufacturing by the final treatment step (i.e., Reverse Osmosis/RO water, deionized water/DI water, etc.).
Guidance on establishing specifications for monographed USP water is provided in the United States Phar-
macopoeia (USP). Additionally, the FDA Guide to Inspections of High Purity Water Systems (which wasdeveloped for FDA personnel) also provides useful information to the user.
2.2 UNITED STATES PHARMACOPOEIA (USP)
USP is a Guide to producing medicinal products for consumption within the US. USP specify standards of
quality, purity, packaging, and labeling for a number of waters including two bulk waters, “Water for Injection”and “Purified Water” used in the preparation of compendial (USP) dosage forms. This Guide is concerned
with the production of these two compendial (USP) waters and does not address the other “packaged waters”monographed by the USP. USP 24 (and supplements) is the current version, at the time this Guide was
prepared.
2.2.1 USP Purified Water
Official monograph requirements for “Purified Water” require that “Purified Water”:
• Is obtained from water complying with the “U.S. Environmental Protection Agency National Primary Drinking
Water Regulations, or comparable regulations of the European Union or Japan, and will be referred tosubsequently as “Drinking Water”.
• Contains no added substance
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 22/250
16
KEY DESIGN PHILOSOPHIES
• Is obtained by a suitable process
• Meets the requirements for Water Conductivity
• Meets the requirements for Total Organic Carbon (TOC)
2.2.2 USP Water for Injection (WFI)
Official monograph requirements for “Water for Injection” require that “Water for Injection”:
• Meets all of the requirements for “Purified Water”
• Is obtained by a suitable process and purified by distillation or Reverse Osmosis
• Meets the requirements of the Bacterial Endotoxin test and contains not more than 0.25 USP EndotoxinUnit per ml
• Is prepared using suitable means to minimize microbial growth
2.2.3 Non-Monographed but accepted requirements
The USP “General Information” provides background information, which clarifies regulatory intent. The fol-
lowing information is included in Chapter 11:
• Purified water systems require frequent sanitization and microbiological monitoring to ensure water of
appropriate microbiological quality at the points of use.
• Water for Injection is “finally subjected to distillation or Reverse Osmosis”, implying that the Still or ROunit is the last unit operation. “The system used to produce, store and distribute water for injection must
be designed to prevent microbial contamination and the formation of microbial endotoxins, and it must bevalidated.”
• An action limit of 100 colony forming units per ml (10,000 CFU/100 ml) for “Purified Water” is suggested.
• An action limit of 10 colony forming units per 100 ml (10 CFU/100 ml) for “Water for Injection” is sug-gested.
• Minimum sample sizes are 1 ml for USP Purified Water and 100 ml for WFI. (FDA recommends 100 ml forPurified Water and 250 ml for WFI).
Note: “It should be emphasized that the above action guidelines are not intended to be totally inclusive for
every situation where ingredient waters are to be employed. It is therefore, incumbent upon the manufacturerto supplement the general action guidelines to fit each particular manufacturing situation” [USP24, page
2163]. When designing a pharmaceutical or medical device water system, it is critical for the designer toconsult with the manufacturers technical experts to ascertain what purity levels must be achieved.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 23/250
17
KEY DESIGN PHILOSOPHIES
2.2.4 USP testing and instrument requirements
THREE STAGE CONDUCTIVITY TESTING
CONDUCTIVITY INSTRUMENT REQUIREMENTS FOR ACCEPTABLE REGULATORY
MEASUREMENT
Temperature Measurement ± 0.25°C Accuracy Resolution < ± 0.1 µS/cm
Conductivity cell constant ± 2% Accurate Reading accuracy < ± 0.1 µS/cm
Location of In-Line meters Must reflect the quality of the water used. Typically, the optimum location in adistributed water loop is following the last “point of use” valve, and prior to thestorage tank return connection.
Instrument type The above procedure is based on the use of a “dip” or “flow through” conduc-
tivity cell. Conductivity readings used to control USP compendial waters mustbe non-temperature-compensated measurements.
2.2.5 Total Organic Carbon (TOC) and Requirements for TOC Control
TOC is an indirect measure, as carbon, of organic molecules present in high purity water. USP replaced theUSP 22 “Oxidizable Substance” wet chemistry test with an In-Line capable, TOC test. A limit was determined
by USP to be 0.5ppm or 500ppb, based on the results of studies and an industry wide survey of pharmaceu-tical water systems.
SYSTEMS AVAILABLE FOR MEASURING TOC
Instruments are available for measuring TOC In-Line from slipstreams and from grab samples manuallyremoved from the water system. Automatic Off-Line sample introduction systems are available for processing
large numbers of grab samples. USP have not prevented acceptable technologies from being used, but limitthe methods to the following instruments that are capable of completely oxidizing the organic molecules to
Carbon dioxide (CO2), measuring the CO2 levels as carbon, discriminating between Inorganic carbon (IC)and the CO2 levels generated from the oxidization of the organic molecules, maintaining an equipment limit ofdetection of 0.05 mg per liter, or lower, and periodically demonstrating an equipment “suitability”.
Acceptance Criteria
Use the stage 1 table from the latest revisionto USP to determine the conductivity limit.
When change does not exceed a net of 0.1µS/cm over 5 minutes, measure the conduc-
tivity. If less than 2.1 µS/cm the water meetsthe requirements.
Use the stage 3 table from the latest revisionto USP to determine the conductivity limit. If
either the measured conductivity is greater thanthe limit value or the pH value is outside the
range of 5 to 7, the water does not meet the
requirements.
Method of Measurement
Use In-Line or grab sample and measure theconductivity and operating water temperature
Retest at least 100 ml of the stage 1 grabsample for conductivity after vigorous mixing
and temperature normalization to 25°C ±1°C
If the stage 2 test does not meet the require-ments, retest the sample within 5 minutes while
maintaining temperature. Add 0.3 ml per 100ml of saturated potassium chloride solution and
determine the pH to the nearest 0.1 pH unit.
Stage
One
Two
Three
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 24/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 25/250
19
KEY DESIGN PHILOSOPHIES
2.2.6 USP 23 Microbial and Endotoxin Testing
Microbial contaminants and Endotoxins are traditionally sampled at the points of use in a water system.
USP 24 has made no changes in this area.
2.2.7 USP 23 pH testing
Testing water for compliance with the USP 24 pH limits is required as part of the stage 3 Off-Line conductivity
testing. (pH must be confirmed as being in the range of 5 to 7.) Testing may use calibrated Off-Line meters.Calibration should be performed using solutions of a known pH, covering the range of 5 to 7. The frequency
of calibration should ensure that the levels of accuracy are maintained. Refer to manufacturer for specificrecommendations on both method and frequency.
Table 2-1 In-Line compared with Off-Line TOC Monitoring
Features
Installed cost
Operating cost
Recommended
test frequency
Frequency ofSuitability and
Limit Responsetesting
In-Line
Monitor should include built in alarms andbe programmable in respect to the “out
of spec.” excursions. Should have con-venient method of conducting Limit Re-
sponse and Suitability Tests.
Medium, based on above features and a
single installed unit. High, if multiple unitsare installed.
Low to high, depending on instrument
capability for suitability and limit responsetesting and the number of instrumentsinstalled.
4 to 48/day
The recommended frequency is based
on the specific system requirement fortrending or concern for “out of spec.” ex-cursions and their subsequent investiga-
tions. See paragraph on “Special Re-quirements”.
Based on documented history
Off-Line
Laboratory instrument should be capableof achieving robust oxidation levels and
should include automatic Off-Line sampleintroduction systems, for processing
large numbers of grab samples. A gen-erous supply of scrupulously clean poly-
mer based sample containers is required.Laboratory instruments will require re-agents and carrier gases.
High, based on above features
High
1/shift
1/shift
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 26/250
20
KEY DESIGN PHILOSOPHIES
2.2.8 Validated Backup Instrumentation
Failure of a monitoring instrument should not be precluded when making decisions concerning type, locationand the extent of validation. Since each excursion from the acceptable limits must be investigated, In-Line
installations should be supplemented with a calibrated laboratory instrument as backup. Validation shouldinclude the operation in Off-Line mode as a supplement or alternate to In-Line instrumentation. Off-Line
laboratory testing should also include a backup instrument to be maintained calibrated in case of failure of theprimary unit.
2.3 SPECIFICATION OF PHARMACEUTICAL WATER QUALITY
2.3.1 Specifying Water Quality
The quality of water supplied in any pharmaceutical process must be consistent with the quality required forthe final product. It may not be sufficient to specify a water quality that meets the specification of the two
compendial grades of water outlined in the USP. These grades, USP Purified water and WFI, are minimumstandards. A more stringent specification could be required depending on the intended use of the product
and on the process used to manufacture that product. It is the responsibility of each drug manufacturer to
establish the logic for their water quality specification based on the required quality of the end product.
Pharmaceutical water uses can be categorized as:
• An ingredient in a dosage form manufacturing process
• An ingredient in an Active Pharmaceutical Ingredient (API) process (the term API is used interchange-ably with BPC, meaning Bulk Pharmaceutical Chemical)
• Equipment cleaning or rinsing
Water intended for use as a dosage form ingredient must be USP monograph water and must be producedconsistently to specification. Evidence of control is required for all critical process parameters that may affect
the final drug characterization. USP WFI water would be expected to be used for parenteral manufacture,some ophthalmic and some inhalation products.
The monographs for USP Purified and WFI compendial pharmaceutical waters stipulate the baseline require-
ments for water used in production, processing, or formulation of pharmaceutical activities.
For some applications where there are no requirements for compendial waters, the manufacturer may estab-
lish quality specifications equivalent to USP-WFI or Purified Waters, depending on the specific application.
Specifications for water used as an ingredient (exclusive of sterile bulks) in the manufacture of API’s or as thewash solvent in the wash or rinse cycles must be determined by the manufacturer. In some cases “Drinking
Water” may be acceptable, or certain chemical or microbial or endotoxin quality specifications may be estab-lished, or one of the compendial waters may be used. The specification should be based on the potential forcontamination of the final drug product. Any decision about water usage must be made with the approval of
Quality Assurance.
With the appropriate justification, non-compendial pharmaceutical waters (including “drinking waters”) maybe utilized throughout pharmaceutical operations including production equipment washing / cleaning as well
as rinsing, laboratory usage and as an ingredient in the manufacture or formulation of bulk active pharmaceu-tical ingredients. Compendial water must, however, be used with preparation of (as an ingredient) compendialdosage forms. In both compendial and non-compendial waters, the manufacturer must establish an appropri-
ate microbial quality specification per the FDA “Guide To Inspections of High Purity Water Systems”. The
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 27/250
21
KEY DESIGN PHILOSOPHIES
significance of microorganisms in non-sterile pharmaceutical products should be evaluated in terms of theuse of the product and the nature of the product and the potential harm to the user. Manufacturers are
expected to establish appropriate microbial alert and action levels for microbial counts associated with thetypes of pharmaceutical waters utilized. These levels must be based on process requirements and the his-
torical record of the system in question. The US Pharmacopoeia states action levels that are generally con-sidered acceptable are 500 CFU/ml for drinking water, 100 CFU/ml for Purified Water, and 10 CFU/100 ml for
WFI, and may be more stringent depending on its use. Microbial system design considerations are discussedlater (see Chapter 8).
The user should consider whether microorganisms in pharmaceutical water could threaten product preser-vation or product stability, or whether water may contaminate product with pathogenic bacteria or endotoxins.
Specific microbiological objectives and standards suitable to the needs of the products manufactured mustbe defined. A water system must meet these objectives and a monitoring program must be established /
implemented to document that the standards are consistently being met.
Engineers involved in water system design must understand the chemical and microbial quality attributes in
the water delivered to use points.
The final quality of pharmaceutical water and steam is determined by the manufacturing process and end
product, quality of feed water, pretreatment and final treatment sub-systems, storage and distribution systemdesign and operator/maintenance procedures. Expert QA advice should be sought out to give further detailsabout this important area.
2.4 CRITICAL PROCESS PARAMETERS
Critical parameters are defined as those parameters, which directly affect the water quality at, or after, a
treatment step. For example, water temperature during a heat sanitization cycle has a direct effect on waterquality.
Regarding chemical purity, the quality attributes may be monitored at or after each critical process step, andthe proper performance of that operation confirmed directly. Since microbial quality cannot be directly moni-
tored in real time, the parameters relied upon to control microbial growth are usually (depending on thesystem) considered critical, such as temperature, UV intensity, ozone concentration, circulation rate, saniti-
zation procedures, positive pressure, etc.
For a system producing compendial water, properties mandated in the official monograph (including bioburdenand endotoxins) constitute critical attributes. Critical instruments are those instruments, which measure criti-cal attributes or parameters.
2.5 cGMP COMPLIANCE ISSUES
Satisfying regulatory concerns is primarily a matter of establishing proper specifications, and using effectiveand appropriate methods to verify and record that those specifications are satisfied. Issues such as quality ofinstallation, sampling and testing procedures, operating and maintenance procedures, record keeping, etc.
often have greater significance than the particular technologies selected to purify and distribute the water.
Fundamental conditions expected to aggravate a microbial problem typically include system design condi-tions such as stagnant conditions, areas of low flow rate, poor quality feed water etc.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 28/250
22
KEY DESIGN PHILOSOPHIES
Measures to alleviate such problems include:
• Continuous, turbulent flow
• Elevated or reduced temperatures
• Smooth, clean surfaces that minimize nutrient accumulation
• Frequent draining, flushing or sanitizing
• Flooded distribution loop (maintenance of positive distribution loop pressures)
• Properly designed, installed and maintained system
While the control of chemical quality is important, the primary challenge in a pharmaceutical water system ismaintaining the microbial quality. The industry and the regulatory community have recognized the effective-
ness of maintaining a continuously recirculated system at high temperatures (65°C-8O°C) in preventingmicrobial growth. Distillation has a long and well-documented history of success, but need not be the only
technology considered for producing water with endotoxin limits. Reverse osmosis is the only other technol-
ogy accepted by the USP for WFI. Ultrafiltration has been successfully used to produce waters with strictendotoxin limits that meet WFI attributes, but it cannot, by regulation, be used to produce compendial gradeWFI.
Each pharmaceutical steam and water treatment system must be viewed in its entirety, as design and opera-tional factors affecting any unit operation within the system can affect the whole system. It is useful to identify
both the quality parameters of water entering the system and the quality parameters of the water or steam tobe produced. Water quality should be enhanced with each successive step. It does not necessarily follow that
measures enhancing one quality attribute (such as conductivity, particulate level or color) will always en-hance another (such as microbial population).
2.6 DESIGN RANGE VERSUS OPERATING RANGE
This Guide recognizes the distinction between “Design Range” and “Operating Range” and the impact this
distinction has upon validation and facility system operation. These criteria are defined as:
See Figure 2-1.
Design Range: the specified range or accuracy of a controlled variable used by the designer as a basis todetermine the performance requirements for an engineered water system.
Allowable Operating Range: the range of validated critical parameters within which acceptable product
water can be manufactured.
Normal Operating Range: a range which may be selected by the manufacturer as the desired acceptable
values for a parameter (i.e., conductivity) during normal operations. This range must be within the Allow-able Operating Range.
a) While a water or steam system should meet all stated Design Conditions, the acceptability of the systemfor operation from a cGMP standpoint depends on operating within the Allowable Operating Range.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 29/250
23
KEY DESIGN PHILOSOPHIES
b) Performance criteria for a Pharmaceutical Water Generation System may require a final product waterquality conductivity of 0.5 µS/cm. (2 Mohm-cm) or better as a Design Condition. The Allowable Oper-
ating Range for this pharmaceutical water may, however, allow for generation of water quality with aconductivity of 1.3 µS/cm. (0.77 Mohm-cm) or better. The Normal Operating Range for generating water
may, in the end, be set by the manufacturer at conductivity value approaching 1.0 µS/cm. (1.0 Mohm-cm)or better to provide a comfortable environment for the operation.
c) Normal Operating Range cannot exceed the Allowable Operating Range for the product water. TheDesign Condition selection should reflect Good Engineering Practice.
d) It is also good practice for manufacturers to apply the concept of Alert and Action limits along with
Normal Operating Range. Alert and Action limits should be based on the actual capability of the system.Alert Limits are based on normal operating experience and are used to initiate corrective measures
before reaching an Action Limit, which is defined as the process condition established by product accep-tance criteria. The Action Limit deviations must be kept as a par t of the batch record as they representdeviations from validated parameters.
Figure 2-1 Values of Critical Parameters for Product Water
Note: These are general guidelines.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 30/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 31/250
WATER OPTIONS
and
SYSTEM PLANNING
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 32/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 33/250
25
WATER OPTIONS AND SYSTEM PLANNING
3. WATER OPTIONS AND SYSTEM PLANNING
3.1 INTRODUCTION
This chapter outlines basic water system design criteria and, along with subsequent chapters, aims to pro-
vide a better understanding of pharmaceutical water, how it is used, and how it can be provided. The primarygoal of this chapter is to provide the user with a methodology for:
a) Evaluating water quality options for product manufacturing
b) Evaluating basic system configurations available to provide the water
Detailed information on unit operation design, maintenance and cost factors is addressed in later chapters.
The chapter also outlines the system planning effort for pharmaceutical water systems. This planning startswith the selection of water quality based upon product requirements, processing operations, and end use. Adecision tree concept is included to assist in selection of compendial and non-compendial waters for produc-tion, cleaning, and support. The program then provides steps to guide the user through a use-point and
system analysis, to set-up the water system distribution strategy. Finally, evaluation points are provided forthe selection of the primary system configurations.
3.2 WATER QUALITY OPTIONS
Quality requirements for water used in pharmaceutical manufacturing and product development are driven bythe product characteristics, manufacturing processes, and the intended use of the product. To aid in the waterselection process, the USP Monographs define minimum requirements for general types of pharmaceuticalwater used in almost every pharmaceutical application. However, there is also the opportunity for a manufac-turer to determine water quality requirements, different from those in the USP, based on specific productcharacteristics and processing operations. If this option is taken, the product manufacturer is responsible forassuring that water used to manufacture the product is appropriate, to reliably produce safe product.
Though water quality requirements are product specific, it is impractical to reliably produce special water thatis specific to each situation. Manufacturing operations typically generate and distribute only a few, or perhaps
just one, quality of water. Therefore, products and operations requiring similar water qualities are commonlygrouped. The most common segmentation is that defined in the USP.
Manufacturers agree that in many if not most cases, the requirements defined in the USP are adequate forproduction of safe product. More stringent water quality specifications may be appropriate for some productsand processes. Others may be appropriately less stringent. Typically, more stringent requirements may applyto some processing operations involving significant concentration steps or products comprised of high watercontent, which may be applied in large volume doses. Likewise, processes involving reliable sterilization andpurification steps which remove impurities may, in some cases, not require water qualities as strict as thosedefined in the USP. Other process characteristics can affect water quality requirements as well.
In manufacturing operations with only one quality of water, the water system must be designed to meet themost stringent requirements of the most demanding product or process. With more than one quality of water,products and processes are often categorized and fed by the most appropriate system. The number of typesof water generated is most often a function of volume of water consumed and variation of quality. Largeconsumers may find it economical to generate and distribute multiple grades of water, while small users oftenwill generate only one quality of water.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 34/250
26
WATER OPTIONS AND SYSTEM PLANNING
The three main categories of water used in pharmaceutical manufacturing are:
Drinking water: meeting EPA national primary drinking water regulations. In Figure 3-1 drinking water isincluded in the category Suitable Non-Compendial.
• Compendial water: meeting the compendial requirements for specific types of water in USP Monographs
(i.e., Purified Water USP, Water for Injection USP).
• “Suitable” non-compendial process waters: meeting the requirements of drinking water, but with addi-tional treatment to meet process requirements. It may, or may not, contain added substances for micro-bial control and does not have to meet full compendial requirements for USP Water. In this Guide, wename the non-compendial process waters used in manufacturing by the final/major process step (i.e.,Reverse Osmosis - RO water, Deionization - DI water, etc.).
Non-compendial water is not necessarily less critical, or less costly to produce or to qualify, than compendialwater. It can enable the manufacturer to set product specific quality and/or test criteria that are appropriate forthe specific product and processes.
Generally, more highly purified water is more expensive than less purified water. However, the specifics of
each operation are different. For example, a plant with existing excess capacity of WFI might elect to use WFIover other grades even when unnecessary. In the example case, documentation defining water quality shouldidentify the quality required for the product and why the WFI was used instead.
Figure 3-1 provides the framework of a diagram that can be developed by a manufacturer to show the re-quirements for water used in the pharmaceutical manufacturing processes. This diagram should be accom-panied by documentation supporting the options chosen, with review and approval of Quality Assurance. Theoptions chosen should be based on product and process specific requirements. Ultimately, water supplied toany process must meet or exceed the requirements, as defined by the manufacturer, for the safe and reliablemanufacture of that product.
Figure 3-1 provides an overall summary of water requirements for a manufacturer supported by the neces-sary justification for specific products, processes, and areas. It is almost impossible to provide one generic
decision tree due to the diversity it would have to cover.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 35/250
27
WATER OPTIONS AND SYSTEM PLANNING
Notes:
1) By test procedure definition, some analytical methods require USP Compendial waters. Quality shouldmeet the needs of the analytical methods.
2) Labs performing both cGMP and Non-cGMP operations should follow the cGMP path.
3) Non-compendial water may be more highly purified than compendial water. Endotoxin and microbialquality is based on the process and quality standards of the product. Non-compendial water must at aminimum meet EPA (or comparable EU or Japanese standard) drinking water requirements for microbio-logical quality.
4) Quality of final rinse water is determined by the type of product and subsequent processing steps. Whereproduct contact surface is subsequently sanitized, final rinse with Suitable Non-Compendial water maybe acceptable. Such practice may necessitate more stringent qualification criteria for the subsequentsanitization steps.
5) Where product is purified downstream
Figure 3-1 Pharmaceutical Water Quality Decision Tree
Note: Commitments made in drug applications override suggestions of this decision tree.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 36/250
28
WATER OPTIONS AND SYSTEM PLANNING
6) More stringent endotoxin requirements (e.g., WFI quality) should be employed for water used in the finalpurification step for NON-Sterile parenteral grade APIs.
Figure 1-1 provides “Baseline” requirements for most product contact water applications. Water quality cri-
teria for pharmaceutical manufacturing and product development are driven by the product charac-teristics, manufacturing process, and the intended use of the product. Specific product and process
characteristics may dictate that more or less stringent criteria than shown are appropriate. Figure abovegives engineers some general guidance on selection of pharmaceutical water quality. Expert QA adviceshould be sought to give further advise on this critical of pharmaceutical water selection.
Once water needs are determined based on usage, Table identifies common design options for various typesof pharmaceutical water in the industry. The order of components and actual installed equipment varieswidely throughout the industry. Primary criteria in evaluating the options are:
• To have suitable specification for water criteria (i.e., it must be adequate for the process and product)
• To produce water consistent in composition and quality
• To monitor key performance indicators for assurance that specifications are met.
DESCRIPTION
Product Staged RO
System
Either Conventional
Regenerable or Off
Site Regenerated Ion
Exchange/Mixed Bed
System
Variations of Single
and Double Pass RO
Followed by Mixed Bed
DI System
Single Pass RO &
Electrodeionization
System
Regenerable Mixed
Bed/Ultrafiltration
System
Single Pass RO/
Non-Regenerable (or
off site regenerated)
Mixed Bed/
Ultrafiltration System
Ultrafiltration often with
some pretreatment
Various configurations
of stills often with some
pretreatment
Ultra
filtration
X
X
X
PHARMACEUTICAL WATER TYPE
TYPICAL
PROCESS
WATER
TYPES
Double Pass
RO Water
DI Water
RO/DI Water
RO/EDI Water
DI/UF Water
RO/DI/UF
Water
UF Water
Distilled Water
Primary
Filtration
X
X
X
X
X
X
X
Softening
X
X
X
X
X
X
Activated
Carbon
Filtration
X
X
X
X
X
X
X
Ion
Exchange
(Cation/
Anion -
1st Stage)
X
X
RO
(1st
Pass)
X
X
X
X
RO
(2nd
Pass)
X
Ion Exchange
(Mixed Bed -
2nd Stage)
X
X
X
EDI
X
Still
XVarious configurations of pretreatment,
primarily to prolong the still life.
PROCESS UNIT OPERATION
Table 3-1 Typical Pharmaceutical Process Water Types
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 37/250
29
WATER OPTIONS AND SYSTEM PLANNING
3.2.1 Cost Implications
Determining the economics of pharmaceutical/medical device water production is complex. Costs are quitepredictable, but vary greatly depending on scale of operation, system design, actual usage, etc. The totaloperating cost to produce pharmaceutical waters is obtained by adding the cost of feed water to the costs ofpretreatment (e.g., media filtration, carbon filtration, softening, and chemical addition) and final treatment
(e.g., primary ion removal and polishing).
Other significant costs should be anticipated for validation, ongoing QA/QC, as well as waste treatment andsewerage. In addition, regulated industries must consider the risks (cost) of noncompliance and water sys-tem failures. Municipal feed water ranges from $1-3 per thousand gallons with even wider variations outsideof the U.S. Feed (surface or ground) water quality, generation technology and its associated capital cost, andproduct water specifications are then utilized to determine the total pharmaceutical water system net presentvalue (NPV). The type of pharmaceutical water system design option selected is typically based on feedwater total dissolved solids and hardness levels, organic and colloidal content, as well as anticipated watersystem utility costs (acid, caustic, salt, power, and source water). Consideration should also be given tomaintenance requirements and available resources.
Although water treatment systems for generating either compendial (USP purified) or non-compendial phar-
maceutical process waters significantly vary in system operational costs, NPV for each of these varioustypes of process waters are quite similar. The only exception is DI process water generated through the useof a non-regenerable mixed bed bottle system, typically regenerated off site. However, membrane basedsystems do marginally produce the lowest net present values for pharmaceutical water generation. The NPVanalysis is usually based on the water system capital cost and a five-year system operating cost. The periodchosen has to be long enough to allow operating cost to be a significant factor, but short enough for reason-able analysis of operating cost returns versus increased capital expenditures.
Cost savings opportunities can be found in other places than just the quality of water and method of genera-tion. Wastewater from the pretreatment or treatment systems can often be used for miscellaneous loads suchas lawn irrigation, humidification, boiler feed, etc. Each chapter of this Guide also addresses cost savingsissues based upon design criteria and approach for independent unit operations and systems.
3.3 SYSTEM PLANNING
High purity water and steam are the most widely used, and often the most expensive raw material or utility ina pharmaceutical facility. Improper sizing or selection of a steam or water system could limit or even shutdown production if under sized; or compromise the reproducible quality and increase the capital cost if over-sized. However, system sizing is not the starting point in design. Proper definition of water quality require-ments and usage can save construction as well as operational costs.
Figure 3-2 shows a graphic representation of the system boundaries, limitations, and restrictions the de-signer faces when planning a pharmaceutical water system. Initial system planning reveals primary bound-aries that establish the cornerstone for design criteria. These primary system boundaries are Water Quality,Use-Point Criteria, and System Criteria.
During initial planning, the limits of each boundary need to be established. The arrows encircling each bound-ary represent limitations that establish more specific operating strategies and ranges. When documentingthese limitations, the designer should always indicate ranges of acceptability, rather than a specific value orposition. This allows more flexibility in final planning and detailed design decisions.
The reality of certain restrictions will sometimes force a specific strategy. As long as the decision leads to ananswer that is within the limits of the system boundaries, this is perfectly acceptable. An example is a facilitywhere the use-point criteria require non-compendial water with microbial control. However, there happens to
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 38/250
30
WATER OPTIONS AND SYSTEM PLANNING
be an oversized WFI system in an adjacent area, so the designer decides to provide WFI to the use-point. Inthe example case, documentation defining water quality should identify the quality required for the productand why the WFI was used instead.
The primary emphasis of this section is to outline a systematic approach to planning a pharmaceutical watersystem. Figure 3-2 outlines a planning methodology that begins with the selection of water quality, given its
own system constraints and limitations. Then the use point criteria are established, followed by an initialsystem planning exercise. Often, these sequential steps are repeated as information in the design processiterates, and further criteria about the overall system boundaries are identified.
Figure 3-2 Pharmaceutical Water System Planning
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 39/250
31
WATER OPTIONS AND SYSTEM PLANNING
3.3.1 Establish Water Quality
The first step in the evaluation of water systems is the selection of water quality required for the specificproduct and process operation. Selection is based primarily on the dosage and form, and the microbiologicaland chemical purity criteria set for the product for which the water is used. The selection must considerunderlying factors that have impacts on quality control; installed and operating cost; maintenance and practi-
cality.
See Section 3.2 in establishing possible water quality via development of the decision tree. Making notes asthe water quality is designated for each use-point, indicating the basis for each decision. Simple annotationsfrom the supporting documentation will be useful in later stages of the planning process. System designconstraints may provide the motivation to challenge water quality or other criteria, particularly when it can bedemonstrated that the change does not affect product quality or manufacturing controls.
3.3.2 Characterize Use Point
Once the initial selection of water quality has been established, the operational criteria should be character-ized for each use point. A matrix should be developed to outline the primary criteria required for systemdesign.
Each use point should be annotated with the proper values for pressure, flow, and the temperature range ofwater entering unit operation, or process point from the water supply system. Establishing a range, ratherthan a fixed value, increases opportunities for system optimization by allowing a more flexible approach tofinal design.
This data can be organized in many ways, but a well-planned spreadsheet can simplify the planning processand provide clear decision pathways for future detailed design activities. Table 3-1 shows an example of aspreadsheet used to characterize use-point flow and system demand. Flowrate is primarily used to size lines,whereas Daily Use leads to storage and generation decisions. The Diversity Factor is one way to level-outanticipated usage, assuming that not all loads happen every day or at the same time. This table indicates aClean in Place (CIP) system and stopper-washer that are both likely to be used on the same day, but never atthe same time. Therefore, only the higher flowrate is relevant to loop sizing as shown in the Design Flowrate
column. Demand flowrates are eventually used for branch line sizing.
FLOWRATE DAILY USE
EQUIPMENT
NAME
CIP Wash cycle
Stopper Washer
DEMAND
(LPM)
40.0
20.0
DIVERSITY
FACTOR
1
0
DESIGN
(LPM)
40.0
0.0
DEMAND
(LPD)
1200
460
DIVERSITY
FACTOR
1
1
DESIGN
(LPD)
1200
460
COMMENTS
Assume a recirculating
cycle in 4 steps for a
total of 23 minutes.
Assume one cleaning
cycle per day, 100liters/rinse, 3 rinse/
cycle, 1 overflow rinse/
cycle @ 2 LPM for 80
mins.
Table 3-2 Use Point Criteria
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 40/250
32
WATER OPTIONS AND SYSTEM PLANNING
Once the location and qualities are finalized, the various properties can be charted on a requirements analy-sis histogram. This can be done with the aid of a computer and either simulation or spreadsheet software forlarger systems, or done manually for small systems. At this point, basic Process Flow Diagrams also providea good pictorial view of the water qualities, locations and the point-of-use properties.
Figure 3-3 Water Usage Chart
3.3.3 Establish System Criteria
Histogram analysis is beneficial for determining overall system peak demand(s), average demand, and therelationships between peak demand time periods and their flow rates. Figure 3-4 below shows a hypotheticalstorage tank profile using the 24-hour demand profile from Figure 3-3.
There is no “Rule-of-Thumb” for minimum water level, or the optimum water level to turn on a still. However,these charts provide the tools for creating various scenarios to simulate recovery times from a failure, futureexpansion or reduction capabilities and analyze other factors that allow design of a properly sized watergeneration, storage and distribution system.
System planning and analysis also reveals other restrictions that influence design, and often lead the de-signer to re-evaluate the primary boundaries as discussed earlier. These restrictions might include items
such as:
• Must the system be available at all times?
• What are the constraints on a shutdown?
• Is the plant/personnel able to handle chemicals properly? Are permits in place?
• Is production batched or continuous?
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 41/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 42/250
34
WATER OPTIONS AND SYSTEM PLANNING
using an ambient or cold distribution versus a hot system. Plant shift operating hours must also be consideredsince there may be an inability to perform regular heat sanitizations of cold systems, for example.
The boundaries, limitations, and restrictions that were identified in the initial planning stages should now beintegrated into the design approach. Further considerations might include the physical area a system needsfor support, one production area, one building, or multiple buildings on a site. This could determine the size of
the system and whether it is made up of multiple tanks, or multiple loop storage and distributed systems. Forexample, central systems are higher in initial capital, but lower in operation and maintenance and possiblylower overall cost on per unit basis. The capital is higher primarily due to the larger generation, storage, anddistribution equipment or system. Alternatively, multiple generation systems may require less initially for eachsmaller system but more cost in terms of capital and operating and maintenance for the same total capacity.
All systems have a fixed capacity and will eventually have a failure. Therefore, if a piece of equipment fails aplan should be in place to deal with the down time. Having backup generation equipment for the criticalcomponents such as a still or deionization equipment, should be considered. The backup equipment can beused in a lead-lag type operation and/or to meet a specific duration of a peak demand.
More detailed descriptions of the alternatives for the various unit operations required for production of phar-maceutical grade water are discussed in the following chapters. Rationale is provided for decisions that will
surface regarding quality, cost, performance, maintenance, and reliability as the system is developed indetail.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 43/250
PRETREATMENT OPTIONS
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 44/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 45/250
35
PRETREATMENT OPTIONS
4. PRETREATMENT OPTIONS
4.1 INTRODUCTION
Pretreatment is all process steps or unit operations prior to the last (final) water treatment step. Pretreatment
is a series of unit operations to modify the feed water quality so that it will be of adequate quality to be fed toa final treatment step. This final step may be Reverse Osmosis, Ultrafiltration, Multi or Mixed Bed Deionizationor Distillation. These final steps are discussed in Chapters 5 and 6.
Reverse Osmosis is unique since it can be a pretreatment step, in addition to being a final treatment step.
Reverse Osmosis applications in pretreatment are discussed in this chapter and Chapter 11, but ReverseOsmosis as a technology is discussed in Chapters 5 and 6.
The initial sections of this chapter discuss the process design (programming issues) for pretreatment designincluding feed water quality and output water quality from pretreatment. The chapter then discusses the
selection of treatment options (i.e. unit operations) for four groups of impurities:
Control of fouling--removal of turbidity and particulates
• Control of scaling--removal of hardness and metals
• Removal of organics and microbiological impurities
• Removal of microbial control agents
Pre-treatment options are summarized in Figure 4-1 at the end of the chapter.
The final sections of the chapter discuss the importance of anion composition/concentration, pH, materials of
construction, and pretreatment system control.
This discussion is based on the description of these technologies presented in Chapter 11.
4.2 PROCESS DESIGN OF PRETREATMENT
Process design of the pretreatment system is the specification of the unit operations or process steps to treatthe feed water. Typical information includes flow rates, temperatures, pressure, and composition of all streams.Detailed mechanical design of the equipment for a given unit operation or process step is beyond the scope
of this Guide.
The process design (programming issues) for a pretreatment system may include:
a) Required quantity and quality of the water from the final treatment process.
b) Temperature constraints on the water used in a pharmaceutical process and the approach to microbial
control.
c) The final treatment option that has been chosen, as this defines the required water quality leaving pre-treatment.
d) Quality of the feed water that is the input to the pretreatment system (water quality to be validated over aone year period).
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 46/250
36
PRETREATMENT OPTIONS
e) Difference between input water quality and desired output water quality. The difference determines impu-rities that must be removed by the pretreatment system. The difference is determined by performing a
material balance. Attention should be paid to impurities and minor components.
f) Pretreatment options to provide the desired removal of impurities considering other factors such ascapabilities of the labor force, economics, waste disposal, environmental considerations, validation, and
the available space and utilities.
In addition to defining the options for removal of impurities, the approach taken to microbial control is an
integral part of the process design of the pretreatment system. Considerations include:
a) If the drinking quality water to the pretreatment system comes from a municipality in the United States, itwill typically contain chlorine, or chloramines, as a microbial control agent. In Europe, ozone is the more
common microbial control agent. The concentration of the agent should be sufficient to protect the initialsteps of the pretreatment.
b) If the quantity of microbial control agent is insufficient, additional microbial control agent may be added orprovision made to periodically sanitize the initial equipment in the pretreatment system. This is likely if
water comes from a source other than a municipality. Increased monitoring of feed water and the initial
steps may be warranted.
c) At some point in the pretreatment process, the microbial control agent must be removed before going to
the final treatment. At this point, a means of either continuous or periodic sanitization needs to be se-lected for the treatment steps following this removal.
The USP requirement that compendial waters should contain “no added substances” eliminates addition ofchemicals to “Purified Water” or Water For Injection. However, addition of chemical agents is not prohibited in
pretreatment. Substances are frequently added in pretreatment and subsequently removed in the pretreat-ment or final treatment. Some examples are:
• Chlorine (to control microbial growth, removed in later stages of pretreatment)
• Sodium ions (in softener with exchange for multivalent ions, removed in ion removal process)
• Acid (for degasification to remove carbon dioxide, counter ions, removed in a subsequent ion removalprocess)
• Sulfite (to reduce chlorine to chloride, or chloramines, to ammonium and chloride while forming sulfate,removal by softening or ion removal process)
• Sequestrants (to prevent scaling in final treatment, removed by RO in final treatment)
• pH control agents (removed in ion removal process)
Added substances are an issue if they result in an increase in microbial growth or endotoxins.
A final consideration is the relationship between investment and operating dollars in pretreatment, and theperformance and cost of the final treatment process. The following are generally true:
• A final treatment system will not operate reliably over the long term, without reliable operation of the
pretreatment system.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 47/250
37
PRETREATMENT OPTIONS
• Inadequate operation in pretreatment (breakthrough of particulates, hardness, or chlorine) may not im-mediately affect water quality from final treatment, but it will be reflected in long term maintenance and
operating reliability, and possibly in water quality.
• Investment in pretreatment capability and reliability can return many times the investment in final treat-ment maintenance costs.
• Pharmaceutical water systems are expected to generate water meeting final pharmaceutical productwater standards. The system should be designed to control impurity spikes in the incoming water quality,
or seasonal impurity profile changes. A robust pretreatment system design handles impurity spikes det-rimental to final treatment.
There is no single “right” answer to the process design of the pretreatment system. Pretreatment system
process design is a series of choices and options, each with advantages and disadvantages.
4.3 FEEDWATER TO PRETREATMENT QUALITY: TESTING AND DOCUMENTATION
Compendial pharmaceutical water systems are required to use feed water complying with “Drinking Water”
standards.
Most pharmaceutical manufacturers utilize municipal water supplies. This water generally meets “Drinking
Water” quality standards and is treated with a microbial control agent. Historically in the US, the microbialcontrol agent is chlorine, but chloramine is now used with increasing frequency. Either feed water composi-tion or microbial control agent concentration may be subject to occasional and seasonal variations. These
variations may negatively impact water quality, and can be detected only by extensive sampling. In addition,water quality at the plant site may not be equivalent to that from a municipal treatment facility, due to potential
for contamination or loss of microbial control agent in the distribution system. Documentation of feed waterquality is recommended either by use of municipality testing (if applicable) supplemented by some testing at
the plant side or by extensive testing of feed water quality.
Typical contaminants in feed water include:
• Particulates Silt, dust, pollen, pipe scale, iron and silica, undissolved minerals and organics
• Inorganics Calcium and magnesium salts, heavy metals (iron, aluminum, and silica) with their corre-
sponding anions
• Organics NATURALLY OCCURRING BYPRODUCTS OF VEGETATIVE DECAY, I.E., HUMIC AND
FULVIC ACIDS AND “MAN-MADE ORGANICS” SUCH AS PESTICIDES AND AUTO-MOTIVE POLLUTION (OILS)
• Bacteria BACTERIAL CONTAMINATION AND ITS BYPRODUCTS, ENDOTOXINS, AND
PYROGENS
Testing recommendations include:
• Documentation that feed water meets drinking water quality. This may be based on results of testing
by the municipality, possibly supplemented by local or in-process testing. Frequency of in-process testingwill be affected by reliability of the municipality, importance of monitored variables, and company philoso-
phy.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 48/250
38
PRETREATMENT OPTIONS
• Monitoring for microbial control agent levels at the start of the pretreatment system. Chlorine levelis affected by pH. A chlorine level of 0.2 - 1.0 ppm is generally considered adequate to control microbial
growth and generally has negligible effects on pretreatment equipment or performance.
• Specific testing for contaminants known or suspected of being present in the feed water. This is todetermine if data from the municipality is adequate; e.g., feed water from a surface source for pesticides
in an agricultural area where run off from farms may be seasonal.
4.4 OUTPUT WATER FROM PRETREATMENT: QUALITY OF FEEDWATER TO FINAL TREATMENT
The goals for pretreatment are to provide water quality that minimizes the operating and maintenance prob-lems in the final treatment equipment and to permit the final treatment step to produce water meeting the
desired specifications for final treatment.
The impurities that must be removed in the pretreatment process to permit reliable operation of the final
treatment step depend on the final treatment step selected and the tolerance of a final treatment step for theimpurities. If pretreatment is inadequate, resulting problems can become very large in magnitude, as seen in
Table 4-1 below:
Pretreatment requirements for feed water to the final treatment process usually include:
FOR MEMBRANES
The concerns are fouling by suspended solids (particulates) and scaling (precipitating solids) as water isremoved. A typical goal for control in pretreatment might be a silt density index (SDI) of 3-5 and hardness of
<1 grain/gallon for on-site analysis. Membranes tolerate chlorides but only some membranes tolerate chlo-rine.
Table 4-1
DEGRADATION:caused by
chlorine
Large*
Large*
Large
Large
Large
MAGNITUDE OFPROBLEMS IN FINALTREATMENT CAUSED
BY TYPE OF IMPURITY
Reverse Osmosis
Other Membrane Processes
Single Effect Distillation
Multi-effect Distillation
Vapor Compression
Distillation
IMPURITY
FOULING:caused by
particulates
Large
Large-moderate
Moderate
Large-moderate
Moderate
SCALING:caused by
hardness andminerals
Large
Large - moderate
Moderate
Large - moderate
Moderate
CORROSION:caused by
chlorides
None
None
Moderate - large
Moderate - large
Small
*Membrane dependent
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 49/250
39
PRETREATMENT OPTIONS
FOR DISTILLATION
The concerns are scale formation due to hardness and corrosion due to chlorides. Typical water quality mightbe as high as 1.0 Mohm/cm, which often requires additional treatment beyond pretreatment, i.e. RO or ultra-
filtration. Distillation has no tolerance for chlorine due to corrosion and carryover to the product. Distillationhas some tolerance for particulates.
Pretreatment generally has little effect on the other parameters indicative of water quality such as anions,microbial levels, conductivity, total organic carbons (TOC), and volatiles.
Selecting pretreatment to reliably provide the required feed water quality to final treatment, in spite of spikes
in feed drinking water quality, will reduce operating and maintenance costs in final treatment.
4.5 CONTROL OF FOULING: REMOVAL OF TURBIDITY AND PARTICULATES
The principal methods for removing particulates and reducing turbidity are:
• Clarification and the accompanying operations of flocculation, coagulation, and sedimentation
• Depth or Media filtration including single and multimedia filtration (particles retained by the media)
The definitions, filtration mechanisms and typical removal processes for these are outlined in Chapter 11.
Clarification is not applicable, as feed water sources are potable quality or better.
Depth or media filtration is used in pharmaceutical water systems and is often the first step in a pretreat-
ment system. Multi-sized sand is the most common media, but other media may provide better performancewith some feed waters. Removal of particulates down to 10 microns is possible and depends on selection of
media. Microbial growth is a key concern in a media filter, unless the feed water contains a microbial controlagent. Otherwise, microbial control in the depth filter is required (e.g., periodic sanitization using either heator a chemical sanitizing agent).
4.6 CONTROL OF SCALING: REMOVAL OF HARDNESS AND METALS
When water is separated from its impurities in the final treatment process, those compounds with low solubil-ity are concentrated to the point where they precipitate. This precipitation, or scaling, is the result of exceed-ing the solubility of the divalent and trivalent cations, usually as a sparingly soluble salt such as carbonate or
sulfate. The methods of control are:
• Removal by ion exchange. These are principally calcium and magnesium and may include divalent andtrivalent ions such as iron, aluminum and silica. Pretreatment is usually water softening, (exchanging the
ions causing hardness and scaling for sodium ions).
• Removal of carbonate by acidification. Acidification converts the carbonate to carbon dioxide, which is
removed by subsequent degasification.
• Removal of the offending compound by a barrier filtration process such as nanofiltration. Waterpasses through the membrane and compounds are retained by the membrane and removed as a purge
stream.
These removal processes are detailed in Chapter 11.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 50/250
40
PRETREATMENT OPTIONS
Water softening ion exchange, which removes divalent and trivalent ions and replaces them with sodium, isa very common process used in pretreatment of pharmaceutical water. It is applicable for all flow rates and all
hardness levels, and is well understood and easy to operate. It involves the handling of salt only, and pro-duces a non-hazardous waste stream. However, the high total dissolved solids (TDS) in the waste stream
may limit disposal options. Water softening is also easily controlled manually or with a PLC.
For large flow rates (>50 gpm or 0.18 m3
/min) and high hardness (>50 ppm) degasification (after acidification)may be the process of choice. This degasification process is often employed between the two stages of anRO and involves the handling of acid and base for two pH adjustments:
• Lowering of pH before first stage of RO
• Increasing of pH before second stage of RO
The principal advantage is that the carbon dioxide is released to the atmosphere rather than being a liquidwaste stream requiring disposal.
Nanofiltration is a membrane process that may be applicable with certain feed waters and specialized situa-
tions. The filtration is usually cross-flow and involves a significant purge stream. It is much like RO, the
differences being pore size in the membrane and the corresponding effect on ion removal. Removal of diva-lent ions can be greater than 98%.
Chemical injection is an alternate method to control the ions or compounds that contribute to scaling. Thisprocess injects a compound (usually a proprietary organic compound) to the final treatment feed water.These compounds are called sequestrants and act “to tie up and complex” the offending ions or compounds
to form a complex, or compound, that is more soluble and will not precipitate in the final treatment process.The “complexed ion and sequestrant” have a large molecular weight and are removed as a purge stream in
the final treatment process. Sequestrants are almost always proprietary compounds, which require testing toverify applicability and dosage level for the particular feed water, and analysis to verify removal in the final
treatment process.
A key choice in the process design of the pretreatment system is location of the softener. The two options are
either before or after removal of the microbial control agent (often chlorine) that is in the feed water, or whichmay have been added for control of microbial growth.
Softener located prior to removal of microbial control agent: The principal advantage is protection of the
softener from microbial growth by the microbial control agent present in the feed water. If the microbial controlagent is chlorine, it will have only a minor effect on resin life and efficiency at the chlorine levels typicallyencountered in chlorinated municipal feed waters (<1 ppm).
Softener located after removal of microbial control agent: The advantage is better resin life and capacity
(due to absence of chlorine, if it is the microbial control agent). However, this must be balanced by the needto protect the softener from microbial growth and endotoxin load (i.e., by periodic sanitization with the asso-
ciated cost of heat or chemicals, labor, down time, and waste stream disposal).
4.7 REMOVAL OF ORGANICS
The types of organics and microbiological impurities typically present in water systems and the methods forremoval of them are discussed in Chapter 11. The methods for removal of organics are:
• Ozone
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 51/250
41
PRETREATMENT OPTIONS
• Strong Base Ion Exchange
• Barrier filtration (microfiltration, ultrafiltration or Reverse Osmosis)
• Polymer Flocculant
• Carbon
Ozone is a powerful oxidant that controls microbial growth and reduces the concentration of organics due to
oxidation, but requires compatible materials of construction.
Strong base ion exchange removes organics but results in a purge stream containing high concentrations ofbrine and organics, due to regeneration of the resin.
Barrier filtration, when appropriately sized, captures organics and microbial growth on the barrier and can beaided by addition of a polymer flocculant. A potential problem with barrier filtration is microbial growth “grow-
ing through the barrier” which results in microbial contamination on the downstream side of the barrier.
Carbon is probably the most common method of reducing organics. It is used because it provides multiple
functions, including removal of organics as well as removal or reduction in the amount of chlorine and chloram-ines (if these are present and the carbon filter is appropriately designed). The advantages of using carbon arethat it is a frequently practiced technology, it performs multiple functions, and effectively “cleans up the feed
water”, and microbial growth can be controlled by periodic sanitization. The disadvantage is that it is a sourceof microbial growth, as well as a source of nutrients.
4.8 SYSTEM DESIGN FOR CONTROL OF MICROBIAL GROWTH
The methods for control of microbial growth are summarized in Chapter 11. The methods used in pretreat-
ment to control microbial growth are:
• Microbial control agent such as chlorine or chloramine
• Periodic sanitization (heat or chemical)
• Ultraviolet l ight
• Avoiding dead legs and avoiding water stagnation
A common strategy in the design of the pretreatment system is to leave the microbial control agent providedby the municipality in the water through as many pretreatment steps as possible, in order to protect these
steps from microbial growth.
However, at some point the microbial control agent (chlorine or chloramine) must be removed since it is notcompatible with the final treatment processes. At this point, the only option is periodic sanitization, either withheat or a chemical disinfectant. This must be included in the design of the pretreatment system, along with
the provisions for validation and monitoring its effectiveness via sampling and testing. If a chemical disinfec-tant is used, provisions to remove it and monitor its removal are also required.
Ultraviolet light (UV) is effective in inhibiting microbial growth but is only effective when the light is present. UV
light is often used before a unit operation to minimize the microbial growth in the unit operation by controllingthe microbial counts in the feed water. The most common places for use of UV light are before ReverseOsmosis units and some filters.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 52/250
42
PRETREATMENT OPTIONS
4.9 REMOVAL OF MICROBIAL CONTROL AGENTS
At some point in pretreatment, microbial control agents must be removed because of their detrimental effecton final treatment equipment and performance. Chlorine causes deterioration of most Reverse Osmosis
membranes and is corrosive in distillation. Chloramines can pass through pretreatment and decompose inthe distillation process with an adverse effect on water quality.
The methods for removal of chlorine and chloramines are similar and are detailed in Chapter 11.
For chlorine removal, activated carbon is a straightforward process for the absorption of chlorine. The carbonwill reduce some of the chlorine to chloride ion, which is then removed in the final treatment ion removal
process. Sulfite reduction is also straightforward, with sulfite being oxidized to sulfate and chlorine beingreduced to chloride ion.
Chloramine removal is more complex. Chloramine adsorption on carbon occurs at a much slower rate thanchlorine, necessitating longer contact times and lower hydraulic flow rates. The potential for dissociation of
the absorbed chloramines into ammonium ion and ammonia is a problem. Ammonium is removed by ReverseOsmosis but decomposes to ammonia in a distillation process. Ammonia passes through both Reverse
Osmosis and distillation processes in final treatment.
Sulfite reduction for chloramines results in ammonium and chloride ions. These can be removed by ReverseOsmosis. The ammonium ion partially decomposes to ammonia in the higher temperature distillation pro-
cess, resulting in carryover and affect on the water quality.
Removal of ammonia (from chloramine) and carbon dioxide requires proper pH control to maintain these
species as ions for removal in an RO. The equilibrium of carbonate, bicarbonate, and carbon dioxide is pHdependent, with alkaline conditions required to maintain the ionic species. The equilibrium between ammo-
nium and ammonia is pH and temperature dependent, with acidic conditions required to maintain the ionicspecies. At no single pH point are these species all carbonate and ammonium ions. Thus two pH adjustment
steps followed by the appropriate removal technologies are required to remove both chloramines and carbondioxide.
4.10 CHANGES IN ANION COMPOSITION / CONCENTRATION
Pretreatment systems typically remove non-ionic impurities and cations. Thus, any change in anionic compo-
sition or concentration is usually secondary. However, some distillation processes in final treatment are af-fected by chlorides, which can be removed by an RO prior to the final treatment step.
The pretreatment processes that affect anionic composition are:
• Deionization
• Degasification
• Carbon bed filtration for removal of chlorine and chloramine
• Reduction to remove chlorine and chloramine
• Barrier filtration (nanofiltration, ultrafiltration and Reverse Osmosis)
Ion exchange resins are designed to remove either cations or anions. An ion exchange resin that is designedto remove anions (anionic resin) will typically exchange the anions (chloride, sulfate, nitrate, and carbonate;
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 53/250
43
PRETREATMENT OPTIONS
and bicarbonate if the pH is appropriate) for the hydroxyl ion. The ion exchange may be in a single bed, mixedbeds, or twin beds and will affect anionic composition if an anionic resin is present. Ion exchange as a
deionization process to specifically remove anions is discussed in Chapter 5.
Degasification and the accompanying process of acidification, for removal of hardness, changes anioniccomposition. The water is acidified with a non-volatile acid (usually sulfuric, based on cost and ease of
removal of the resulting anion i.e., sulfate) to convert carbonate and bicarbonate to dissolved CO2, which isremoved by degasification. The net effect is replacement of bicarbonate and carbonate with sulfate, (seeChapter 11).
As discussed above, carbon bed filtration adsorbs chlorine and chloramines from feed water. However, some
of the chlorine is reduced to chloride and is removed in a subsequent ion removal process, usually in finaltreatment.
The removal of chlorine and chloramines by reduction, often with bisulfite, changes ionic composition, andconcentration, as the bisulfite is oxidized to sulfate and the chlorine, or chloramines, are reduced to chloride
and ammonium.
Some barrier filtrations (particularly nanofiltration) remove some of the larger anions. Reverse Osmosis may
be used to remove chloride ion prior to some distillation processes.
4.11 THE IMPORTANCE OF PH IN PRETREATMENT
The effect of pH on the equilibrium between carbonate, bicarbonate, and carbon dioxide is discussed in
Chapter 11.
EPA drinking water standards require a pH range of 6.5-8.5. In reality, the pH range of most drinking feedwater is narrower, due to the corrosive nature of acidic water and the scaling potential of alkaline waters.
The pH of the feed water and its seasonal variations need to be known because of its impact on pretreatmentand final treatment process design. The pH determines the form of the carbon dioxide, its scaling potential
and where carbon dioxide (carbonate) is removed (see Chapter 11).
A complicating factor in pretreatment design is the potential presence of ammonia as a result of chloraminepresence in the feed water. Ammonia is a dissolved gas at the pH values where carbon dioxide is an ion
(carbonate), and exists an ion (ammonium) at pH values where carbon dioxide exists as a dissolved gas.Thus it is not possible remove both carbon dioxide and ammonia at one pH. If both are present, two pHadjustment steps are required:
• pH adjustment followed by removal of either carbon dioxide or ammonia
• A change in pH to remove the other compound
These operations may be part of pretreatment or final treatment.
4.12 MATERIALS OF CONSTRUCTION AND CONSTRUCTION PRACTICES
Piping to the pretreatment system may be copper, galvanized steel, or a suitable thermoplastic. Piping in the
pretreatment system, where high temperatures are not encountered, is usually plastic (PVC, CPVC, polypro-pylene, or other material) based upon cost and corrosion resistance. Leaching from some plastics such asPVC and CPVC may make these materials undesirable to the user. Vessels may be fiberglass, lined carbon
steel, or stainless steel.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 54/250
44
PRETREATMENT OPTIONS
The piping and equipment in a portion of the pretreatment system may encounter high temperature (periodicheat sanitization) or high pressure (RO plus degasification). In these portions, piping is typically stainless
steel or a plastic that can be heat sanitized, such as PVDF. Equipment designed for high pressure may becarbon steel, lined carbon steel, or stainless steel. Mill finish is satisfactory for these materials; electropolishing
is unnecessary.
The cost of sanitary construction practices such as orbital welding and sanitary fittings may not be warrantedin the pretreatment system. Use of plastic pipe that is solvent cemented or heat fused, stainless steel pipethat is welded or flanged with mill finish, or tubing with compression fittings is common. Ball or diaphragm
valves predominate for flow diversion, with globe and needle valves for flow control. Selecting the minimumcost piping components that will not degrade water quality is an area for major cost savings.
Sample points should be provided upstream and downstream of each piece of equipment for monitoring and
for troubleshooting. Points for field measurement of pressure and temperature are also useful for trouble-shooting.
4.13 PRETREATMENT SUMMARY
The philosophy of control selected for pretreatment can have a major impact on both investment and continu-ing operating cost. Reliable operation and control of pretreatment can significantly reduce operating andmaintenance costs in final treatment. The important process steps in pretreatment are:
• Removal of turbidity and particulates to minimize membrane and equipment fouling
• Removal of hardness and metals to prevent scale formation in final treatment
• Removal of organics and microbiological impurities
• Control of microbial growth and removal of microbial control agents to prevent degradation of final treat-ment
These process steps are important because of their immediate effect on water quality from final treatment ortheir long-term effect on final treatment equipment performance and hence, their indirect effect on water
quality from final treatment.
Pretreatment, like other parts of the water treatment system, should be subject to Good Engineering Prac-tices. Validation of pretreatment, as a component of the water treatment system, is required as part of theentire water treatment system validation and should include microbiological monitoring.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 55/250
45
PRETREATMENT OPTIONS
Figure 4-1
Note: The order of unit operations may be different than shown.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 56/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 57/250
FINAL TREATMENT OPTIONS:
NON-COMPENDIAL and COMPENDIAL
PURIFIED WATER
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 58/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 59/250
47
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5. FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.1 INTRODUCTION
This chapter discusses the final treatment technologies and basic system configurations related to the manu-
facturing process of USP Purified Water and non-compendial water.
Various system configurations are presented, and reflect a significant shift from ion exchange based systems
to membrane based systems. Equipment and system materials, surface finish and other design factors arediscussed to promote the use of Good Engineering Practice for proper selection of components, piping,
instrumentation, and controls.
USP Purified Water and non-compendial water can be produced by an almost unlimited combination of unitprocesses in various configurations. The most common pretreatment and final treatment technologies usedin purified water production are shown in Figure 5-1, Figure 5-2, and Table 5-4 at the end of the chapter. This
chapter discusses the final treatment unit processes currently utilized, including ion exchange, reverse os-mosis, electrodeionization, ultrafiltration, microfiltration, and ultraviolet light. These technologies as well as
distillation (see Chapter 6) are utilized in thousands of systems for the successful production of purified and
non-compendial water.
Ion exchange based systems were the dominant systems for decades in purified water production and are
still successfully utilized in facilities today. The last decade has seen the growth of reverse osmosis mem-brane based systems increase to the point where over 90% of new systems employ primary reverse osmo-sis, with final polishing by continuous electrodeionization, ion exchange, or a second reverse osmosis stage.
Membrane based systems usage has increased due to chemical consumption reduction, contaminant rejec-tion (ionized solids, organics, colloids, microbes, endotoxins, and suspended solids), reduced maintenance,
consistent operation, and effective lifecycle cost.
The various membrane based system configurations are compared with ion exchange and distillation in Tableat the end of this chapter.
Equipment construction is discussed for each unit process section to promote proper selection of materials,surface finishes, and other design factors. The total system capital cost is influenced more by equipment
design details than by process selection. Many aspects of equipment can be “overdesigned” and hence,become unnecessarily costly. Proper thought must be given to the individual component’s function, location,
required microbial performance, sanitization, and other factors, to optimize design. It is not necessary toconstruct every makeup system component with the same level of surface finish and detail as the distributionsystem for successful operation in most cases.
Many material selections are made erroneously to conform to cGMP requirements that do not actually dictate
the details of construction for most final treatment components. Good Engineering Practice should be em-ployed to optimize the system for consistent operation to specifications and lifecycle cost optimization. Part of
the consideration is the need to replace system components (e.g., filters, RO membranes) at a frequencythat meets GMP.
This chapter does not differentiate between compendial and non-compendial water system equipment. Non-compendial water is often manufactured and validated in a manner consistent with compendial water.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 60/250
48
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.2 ION EXCHANGE
5.2.1 Description
Cation and anion exchange resins are regenerated with acid and caustic solutions, respectively. As waterpasses through the ion exchange bed, the exchange of ions in the water stream for the hydrogen and hydrox-
ide ions, held by the resin, occurs readily and is driven by concentration. Thus, the regeneration process isdriven by excess chemical concentrations. The important parameters of this system include resin quality,regeneration systems, vessel linings, and waste neutralization systems. The operation of the system can be
monitored by conductivity (resistivity) of the product water.
A two-bed ion exchange system includes both cation and anion resin tanks. Two-bed ion exchange systemsoften times function as the workhorse of a strictly deionization (DI) water system in terms of salt removal.
Mixed-bed ion exchange systems are typically used as a secondary or “polishing” system. Mixed-bed DI unitsconsist of a single tank with a mixture of anion and cation removal resin. A cation bed can also be used as a
“polishing” DI step, rather than a mixed-bed DI.
Ion exchange resins are available in on-site and off-site regenerable systems. On-site regeneration requires
chemical handling and disposal, but allows for internal process control and microbial control. Off-site regen-eration can be accomplished through new resin to be used one time, or through repeated regeneration of theexisting resin. New resin provides greater capacity and some possible quality control advantages, but at a
higher cost. Regenerated resin produces a lower operating cost, but may raise quality control issues, such asresin segregation, regeneration quality, and consistency.
Additional details on ion exchange can be found in Chapter 11.
5.2.2 Application
The major purpose of ion exchange equipment in USP purified water systems is to satisfy the conductivityrequirements of the USP. Deionization (DI) systems are often times used alone or in conjunction with reverseosmosis to produce USP Purified Water. Typical ion exchange systems do not effectively remove other con-
taminants noted in the USP purified water specification. In the ion exchange process, salt ions, which arecommon to potable water, are removed from the water stream and replaced with hydrogen and hydroxide
ions. Ion exchange systems are available in various configurations that include two-bed DI and mixed-bed DI.Both configurations are available in on-site and off-site regeneration systems.
5.2.3 Pretreatment Requirements
Ion exchange systems require pretreatment to remove undissolved solids from the water stream and to avoidresin fouling or degradation. Although dechlorination is also recommended to avoid resin degradation by
oxidation, the low levels of chlorine commonly found in most potable water supplies normally demonstrateonly long-term effects on most ion exchange resins.
5.2.4 Cost Savings Factors
Most of the cost savings opportunities for these systems revolve around the correct choices in materials ofconstruction, pretreatment options, instrumentation, and sizing of the DI system. Acceptable piping materials
of construction can vary from PVC to 316L SS. A correctly designed system will minimize the equipment sizeand maximize the amount of time between regenerations, considering microbial control and maintenance.
Choosing to monitor only the critical parameters such as conductivity (resistivity), flow, pressure, etc., canminimize instrumentation.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 61/250
49
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
There are also cost savings choices that will need to be made with respect to capital purchase and on-goingoperating costs. These choices will steer you towards DI off-site regenerable bottles, on-site regenerable DI
vessels (with automatic or manual controls) or another water treatment unit operation.
5.2.5 Advantages and Disadvantages
Advantages:
• Simple design and maintenance
• Flexible in water flow production
• Good upset recovery
• Low capital cost for single train DI systems
• Removes ionizable substances (ammonia, carbon dioxide, and some organics)
Disadvantages:
• High cost of operations on high total dissolved solids (TDS) in-feed-water
• Requires chemical handling for on-site regenerable DI (safety and environmental issues)
• Full on-site DI system can take significantly more floor space due to primary vessels, chemical storage,
and neutralization system
• Off-site DI systems will require outside service and significant costs for regeneration services
• Off-site regeneration involves consequent loss of control over the use, handling, and care of DI vessels
• DI vessels are excellent places for microbial growth to occur between regenerations
5.2.6 Sanitization
All ion exchange resins can be sanitized chemically with various agents. The degree of resin attrition is a
function of resin type and the chemical agent. Chemical cleaners include peracetic acid, sodium hypochlorite,and others. Some resins are capable of hot water sanitizations at temperatures between 65°C to 85°C. Ionexchange resins suitable for limited thermal sanitizations include: strong acid cation resin, and standard
polystyrene cross-linked with divinylbenzene Type 1 strong base resin.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 62/250
50
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
Table 5-1 Comparison for ion exchange unit operations
Off-Site Regenerated On-Site Regenerated
Chemical Use: N/A Extensive
Sanitization Method: Change Out or Hot Water Regenerate
Capital Cost: Minimal Extensive
Water Consumption: None Medium
Energy Consumption: Minimal Minimal
Maintenance Requirements: Minimal Medium
Outside Service Used: Extensive Low
Reliability: Good* Good
Upset Recovery Operations: Good, Replace Good
*Note: Having the DI bottles regenerated by an outside service does not relieve the manufacturer of theresponsibility to have quality control of their ion exchange system.
Table 5-2 Limits of Operation and Expected Performance
Feed Quality:
• Total Suspended Solids (turbidity):
• Chlorine Tolerance:
• Total Dissolved Solids (TDS):
• Temperature:
• Conductivity:
• Regeneration and Chemical Efficiency:
• Feed TOC:
• Product TOC:
• Filtration of 10 micron is recommended
• Varies with type of resin, generally at 0.5 ppm,some resins are rated up to 1 ppm
• < 200 ppm, operation at higher TDS levels ispossible but operating costs can be high
• Most cation resin up to 121°C; most anion resin
40-70°C; some anion resin up to 100°C
• Can achieve conductivity below 1.0
microsiemen/cm depending on the system
pretreatment and regeneration schedule
• Linear variation is inverse to the feed water total
dissolved solids - best below 200 ppm
• Ability to avoid fouling varies with type of resin
• May increase or decrease incoming TOC levels
depending on resin type and feed water - difficultto predict.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 63/250
51
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.3 CONTINUOUS ELECTRODEIONIZATION (CEDI)
5.3.1 Description
Electrodeionization removes ionized or ionizable species from water using electrically active media and anelectrical potential to effect ion transport. Electrodeionization is distinguished from electrodialysis or oxida-
tion/reduction processes by the use of electrically active media, and is distinguished from other ion exchangeprocesses by the use of an electrical potential.
The electrically active media in electrodeionization devices functions to alternately collect and dischargeionizable species and to facilitate the transport of ions continuously by ionic or electronic substitution mecha-
nisms. Electrodeionization devices may comprise media of permanent or temporary charge and may beoperated batchwise, intermittently, or continuously. The devices can be operated so as to cause electro-
chemical reactions specifically designed to achieve or enhance performance and may comprise electricallyactive membranes such as, semi-permeable ion exchange or bipolar membranes.
The continuous electrodeionization (CEDI) processes are distinguished from the collection/discharge pro-cesses (such as electrochemical ion exchange or capacitive deionization) in that the process is continuous
rather than batch or intermittent, and that the ionic transport properties of the active media are a primary
sizing parameter, as opposed to ionic capacity. Continuous electrodeionization devices typically comprisesemi-permeable ion exchange membranes, permanently charged media, and a power supply that can createa DC electrical field.
A continuous electrodeionization cell is formed by two adjacent ion exchange membranes or by a membraneand an adjacent electrode. CEDI units typically have alternating ion depleting (purifying) and ion concentrat-
ing cells that can be fed from the same water source, or different water sources. Water is purified in CEDIdevices through ion transfer. Ionized or ionizable species are drawn from the water passing through the ion
depleting (purifying) cells into the concentrate water stream passing through the ion concentration cells.
The water that is purified in CEDI units passes only through the electrically charged ion exchange media, andnot through the ion exchange membranes. The ion exchange membranes are permeable to ionized or ioniz-able species, but not permeable to water.
The purifying cells typically have permanently charged ion exchange media between a pair of ion exchange
membranes. Some units incorporate mixed (cationic and anionic) ion exchange media between a cationicmembrane and an anionic membrane to form the purifying cell. Some units incorporate layers of cation and
anion ion exchange media between ion exchange membranes to form the purifying cell. Other devices createsingle purifying cells (cationic or anionic) by incorporating a single ion exchange medium between ion ex-change membranes. CEDI units can be configured with the cells in a plate and frame, or spiral wound con-
figuration.
The power supply creates a DC electric field between the cathode and anode of the CEDI device. Cations inthe feed water stream passing through the purifying cell are drawn to the cathode. Cations are transported
through the cation exchange media and either pass through the cation permeable membrane or are rejectedby the anion permeable membrane. Anions are drawn to the anode and are transferred through anion ex-change media and either pass through the anion permeable membrane or are rejected by the cation perme-
able membrane. The ion exchange membranes are oriented in a manner which contains the cations andanions removed from the purifying cells in the concentrating cells so that the ionic contaminants are removed
from the CEDI unit. Some CEDI units utilize ion exchange media in the concentrating cells, while others donot.
As the ionic strength of the purified water stream decreases the high voltage gradient at the water-ion ex-change media interfaces can cause water decomposition to its ionic constituents (H+ and OH-). The H+ and
OH- ions are created continuously and regenerate the cation and anion exchange media, respectively, at the
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 64/250
52
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
outlet end of the purifying cells. The constant high level of ion exchange media regeneration level allows theproduction of high purity water (1 to 18 Mohm-cm) in the CEDI process.
5.3.2 Application
In some cases, where drug microbiological quality is of lesser concern, CEDI units may be utilized down
stream of reverse osmosis (RO) units in production of USP Purified Water or non-compendial water to in-crease the life of the CEDI units. For USP WFI water, the CEDI units are utilized up stream of reverse osmosis(RO) units.
5.3.3 Limitations
CEDI units cannot remove all contaminants from water. The principal removal mechanism is for ionized or
ionizable species. CEDI units cannot purify 100% of the feed water stream, as a concentrate stream is alwaysrequired to remove the contaminants from the system. CEDI has temperature limitations for practical opera-tion. Most CEDI units are operated between 10 - 40°C (50 - 104°F).
5.3.4 Pre-treatment Requirements
CEDI units must be protected from scale formation, fouling and thermal or oxidative degradation. The RO/ pretreatment equipment typically reduces hardness, organics, suspended solids, and oxidants to acceptablelevels.
5.3.5 Performance
CEDI unit performance is a function of feed water quality and unit design. Ionized solids reduction is generallygreater than 99% allowing production of 1 - 18 Mohm-cm quality water from reverse osmosis feed water.
Organic rejection typically varies from 50% to 95% depending upon the type of organic material present inthe feed stream. Ultraviolet light (185 nm) upstream of CEDI units can substantially increase organic rejec-
tion. Dissolved carbon dioxide is converted to bicarbonate ion and removed as dissolved ion. Dissolved silicaremoval is in the range of 80 - 95%, dependent upon operating conditions.
5.3.6 Cost Savings Factors
Most of the cost savings opportunities revolve around the correct choices in materials of construction, instru-mentation, and post-treatment equipment selection. Acceptable materials of construction for piping can vary
from PVC to 316L SS. Choosing to monitor only the critical parameters, such as resistivity, flow, and pressurecan minimize instrumentation. Many applications for Purified Water require no post-treatment afterelectrodeionization. Some systems incorporate ultraviolet light and/or sub-micron filtration to either reduce
sanitization requirements or to provide microbial levels well below those allowed for Purified Water productionas outlined in the USP.
5.3.7 Advantages and Disadvantages
Advantages:
• Attainment of stage 1 conductivity
• Elimination of chemical handling
• Elimination of outside service (off-site regenerated resin)
• Electric field in membrane/resin module provides some bacterial control
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 65/250
53
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
• Removal of ionizable substances (e.g., carbon dioxide, ammonia, and some organics)
Disadvantages:
• Does not remove non-ionic contaminants
• Unique designs for each manufacturer (modules are not interchangeable)
• May require UV, sub-micron filtration, or reverse osmosis (RO) for further bacterial reduction
• May require reverse osmosis pretreatment
• Rinse up after chemical sanitization may take hours to reach peak resistivity and TOC
5.3.8 Sanitization
CEDI units are typically chemically sanitized with a number of agents including: peracetic acid, sodiumpercarbonate, sodium hydroxide, hydrogen peroxide, and others.
5.4 REVERSE OSMOSIS
5.4.1 Description
Reverse osmosis (RO) is a pressure driven process utilizing a semi-permeable membrane capable of remov-
ing dissolved organic and inorganic contaminants from water. A semi-permeable membrane is permeable tosome substance such as water, while being impermeable to other substances such as many salts, acids,
bases, colloids, bacteria, and endotoxins.
RO membranes are produced commercially in a spiral wound configuration for pharmaceutical water produc-tion. Membranes are available in two basic materials; cellulose acetate and thin film composite (polyamide).All of the membrane types have advantages and disadvantages. Membrane operating parameters are shown
in Table 5-3, below.
RO membranes without leading edge brine seals, allow controlled flow between the membranes and pres-sure vessels to minimize bacterial growth.
Table 5-3 RO Membrane Operating Parameters
Cellulose Acetate Polyamide/TFC
pH 4-7 2-11
Chlorine Limit, mg/l 1.0 0.05*
Resistance to Bacteria Poor Good
Operating Temperature Range °C 15-28 5-50Rejection - % 90-98 97-99
Sanitization Temp. Limit, °C 30 50-80
Typical TDS Feed Range, mg/l 30-1000 30-1000
Silt Density Index, Max 5 5
*BEST OPERATION AT 0.0
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 66/250
54
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.4.2 Application
Reverse osmosis can be successfully implemented in pharmaceutical systems in several ways. RO units canbe utilized upstream of regenerable deionizers, or off-site regenerated deionizers, to reduce regenerant acid
and caustic consumption, or to minimize resin replacement costs. Two-pass RO units (product staged) withproper pH control are generally capable of producing water that meets the requirements of the USP for TOC
and conductivity.
5.4.3 Limitations
Reverse osmosis cannot remove 100% of contaminants from water and has very low to no removal capacity
for some extremely low molecular weight dissolved organics. RO, however, quantitatively reduces bacteria,endotoxins, colloids and high molecular weight organics from water.
RO cannot purify 100% of a feed water stream. A concentrate flow is always necessary to remove thecontaminants that are rejected by the membrane. Many users of RO utilize the waste stream from the RO unit
for cooling tower make-up water or compressor cooling water, etc.
Carbon dioxide passes directly through the RO membrane and CO2 will be in RO product stream at the same
level that present in the feed water stream. Excess carbon dioxide in the RO product stream may increase theproduct conductivity beyond the USP Stage 1 limit. Carbon dioxide contributes to the loading of anion resin,which may be downstream of the RO units.
Reverse osmosis has temperature limitations for practical operation. Most RO systems operate on feed waterbetween 5°C and 28°C.
5.4.4 Pretreatment Requirements
Reverse osmosis membranes must be protected from scale formation, membrane fouling, and membrane
degradation. Scaling is possible since the contaminants present in the feed water stream are being concen-trated into the waste stream, which is an average of 25% of the feed stream. Scale control is normallyprevented by the use of water softening upstream of the membranes, the injection of acids to lower the pH of
the feed water stream, or an anti-scalant compound to prevent precipitation.
Reverse osmosis membrane fouling is reduced through the use of back-washable multi-media filters or car-tridge filters for suspended solids, greens and filtration or softening for colloidal iron removal, and various
microbial control pretreatment methods to reduce biological fouling.
The principal causes of membrane degradation are oxidation of certain membrane materials and heat degra-
dation. Membranes, which cannot tolerate chlorine normally, incorporate activated carbon or injection ofvarious sodium sulfite compounds for dechlorination. Protection against high temperature is normally incor-
porated where the feed water is preheated and the membrane material cannot tolerate high temperature.
The reverse osmosis pretreatment unit operations are reviewed in Chapter 4.
5.4.5 Performance
A single stage of reverse osmosis elements typically reduces the level of raw water salts, colloids, organics,
bacteria, and endotoxin by 90 to 99%. Single stage reverse osmosis product water does not normally meetthe requirements of the USP without further purification steps. Some two-pass units (two sets of RO mem-
branes in series) produce water that can pass the USP 24 Stage 1 conductivity requirements, allowing On-Line testing. Those units that do not meet the Stage 1 requirement normally meet Stage 2 or 3. Membraneselection should be based upon pretreatment requirements, operating performance characteristics, sanitiza-
tion options, warranties, capital and operating costs, and the feed water source.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 67/250
55
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.4.6 Advantages and Disadvantages
Advantages:
• Reverse osmosis units eliminate or significantly reduce chemical handling and disposal, relative to re-generable ion exchange systems
• Generally, RO has more effective microbial control than ion exchange systems
• Integrity testing can be accomplished by salt challenge and measurement of differential conductivity
• RO removes a wide variety of contaminants including ionized solids and non-ionic materials (e.g., col-loids, bacteria, endotoxin, and some dissolved organics)
Disadvantages:
• Water consumption can be significantly higher than ion exchange systems unless the wastewater isreused
• Energy consumption is generally higher than ion exchange and less than distillation
• No removal of dissolved gases (e.g., carbon dioxide and ammonia)
5.4.7 Cost Saving Factors
Capital costs can be minimized by reducing membrane area to the minimum suitable for the feed waterquality and membrane selected. Piping material and finish significantly impact capital cost. Some systems
incorporate PVC low-pressure piping and welded mill finish stainless steel high-pressure piping. Instrumentcosts can be minimized by appropriate selection of critical and non-critical parameters of operation. These
parameters include:
• Flow
• Pressure
• Temperature
• Conductivity
5.4.8 Waste Water Reuse
RO wastewater is frequently used as cooling tower make-up, or for non-contact cooling for compressors, orother heat loads. Wastewater is sometimes re-purified in a wastewater reverse osmosis unit for reintroduction
as system feed water. RO wastewater is sometimes used for filter backwash. The wastewater from the sec-ond pass of a two pass RO is normally returned to the feed water stream of the first pass RO.
5.4.9 Sanitization
All RO membranes can be sanitized with some chemical agents that vary as a function of membrane selec-tion. Specially constructed membranes are available for hot water sanitization at 60° to 80°C.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 68/250
56
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.5 POLISHING COMPONENTS - NON-IONIC CONTAMINANTS REDUCTION
5.5.1 Ultrafiltration
5.5.1.1 Description
Ultrafiltration (UF) is a cross-flow process similar to reverse osmosis (RO). A pressurized feed stream flowsparallel to a porous membrane filtration surface. A pressure differential forces water through the membrane.The membrane rejects particulates, organics, microbes, pyrogens, and other contaminants that are too large
to pass through the membrane. UF does not reject low molecular weight ionic contaminants, as does reverseosmosis.
Membranes are available in both polymeric and ceramic materials. Polymeric membrane elements are avail-
able in spiral wound and hollow fiber configurations. Ceramic modules are available in single channel andmultiple channel configurations.
5.5.1.2 Application
Ultrafiltration is utilized in several ways in Pur ified Water systems. UF is frequently used down stream of ion
exchange processes for organic, colloidal, microbial, and endotoxin reduction. Purified Water with low endot-oxin levels (<0.25 Eu/ml) is utilized by some manufacturers in ophthalmic solutions, topicals, and bulk phar-maceutical chemicals that will be utilized in parenteral manufacturing and other applications.
Ultrafiltration is frequently used in still feed water systems, in combination with ion exchange, to limit theendotoxin and colloidal silica feed levels to the still.
5.5.1.3 Limitations
Ultrafiltration cannot remove 100% of contaminants from water. No ionic rejection occurs and organic rejec-
tion varies with the various membrane materials, configuration, and porosity. Many different nominal organicmolecular weight rejection ratings are available. Dissolved gasses are not rejected by UF.
Most ultrafilters require a waste stream to remove the contaminants on a continuous basis. The waste streamvaries, but is usually two to ten percent. Some UF systems run dead-ended.
5.5.1.4 Pretreatment Requirements
Pretreatment can include multimedia filters, activated carbon filters, ion exchange, membranes, or others.The UF flux rate and cleaning frequency vary widely as a function of feed water and pretreatment. Most UF
membranes are chlorine tolerant and do not require dechlorination of the feed water.
5.5.1.5 Performance
UF is utilized to remove a variety of contaminants. The proper UF membrane must be selected to meet theperformance requirements. Organic molecules can be rejected well, but the rating of UF membranes variesin molecular weight cutoffs from 1,000 to 100,000. Reduction of typical raw water organics is not as effective
as reverse osmosis. Pressure drops vary with membrane selection and operating temperature. Some UFmembranes are capable of continuous operation at temperatures up to 90OC, to provide excellent microbial
control.
UF reduction of endotoxin (pyrogens) varies from 2 log10 to 4 log10 as a function of membrane selection. UFhas been shown to be capable of consistent production of water meeting the USP WFI endotoxin limit of 0.25Eu/ml in typical system applications. UF produces excellent microbial reduction with typical ratings of 3 log10
to 4 log10 reduction.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 69/250
57
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
UF produces excellent particle reduction and is frequently used in other applications, such as semiconductorproduction when particle control is far more critical than pharmaceutical water.
5.5.1.6 Advantages and Disadvantages
Advantages:
• UF can remove some contaminants, such as endotoxin and organics, better than microfiltration
• UF can have more effective operating costs than microfiltration, in high particle loading applications.
• Some UF elements can tolerate more rigorous sanitization procedures using steam or ozone, than someother membrane filters (MF or RO).
• The waste stream is generally much less than waste from reverse osmosis units
• Ultrafiltration is generally less energy intensive than reverse osmosis
Disadvantages:
• UF cannot remove ionic contaminants, where reverse osmosis can
• UF generally requires a waste stream, which can be a significant cost factor
• UF membranes are sometimes more difficult to integrity test than microfiltration cartridges
5.5.1.7 Cost Savings Factors
Capital costs can be influenced by the optimum sizing of membrane area and membrane selection. Piping
material and finish significantly impact capital cost. Some systems incorporate various plastic piping materi-als while others utilize sanitary 316L SS. The sanitization method selected is a major factor in material selec-tion. Instrument costs can be minimized by appropriate selection of critical and non-critical parameters of
operation.
5.5.1.8 Sanitization
UF membranes are sanitized in many different ways. Most polymeric membranes are tolerant of a widevariety of chemical sanitizing agents such as sodium hypochlorite, hydrogen peroxide, peracetic acid, so-dium hydroxide, and many others. Some polymeric membranes can be hot water sanitized and some can
even be steam sanitized.
Ceramic UF elements can tolerate all common chemical sanitizing agents, hot water, steam, and ozone insanitization or sterilization procedures.
5.5.1.9 Waste Water Recovery
Most pharmaceutical UF units are fed deionized water for USP Purified Water production or special non-compendial water applications. The wastewater is therefore still low conductivity water that can be recycled
upstream to reverse osmosis units or fed directly to boilers, cooling towers, or other uses.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 70/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 71/250
59
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
5.5.3 Ultraviolet Light Treatment
5.5.3.1 Description
Ultraviolet light rays strike microorganisms (bacteria, virus, yeast, mold, or algae) and break through theirouter membrane to modify the DNA. The modified DNA code brings about the destruction of the organism.
The ultraviolet radiation is a point of use application with no residual radiation characteristics. Proper prefiltrationshould be implemented to keep particulate from shielding organisms from UV light. (See Chapter 8.)
5.5.3.2 Advantages and Disadvantages
Advantages:
• Simple design and maintenance
• 254 nm design for microbial reduction
• 185 nm design for TOC reduction
• No waste stream
• Heat, ozone, and chemical sanitization are possible
Disadvantages:
• Can be used only as a safety net for microbial production
• No ion or endotoxin removal
• No disinfection residual
• Particulate can shield organisms from UV light
5.5.3.3 Performance
The UV light is used as a final treatment step to address microbial control and TOC reduction (where neces-
sary), after deionization processes.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 72/250
60
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
Figure 5-1 Purified Water
Figure 5-2
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 73/250
61
FINAL TREATMENT OPTIONS: NON-COMPENDIAL AND COMPENDIAL PURIFIED WATER
Table 5-4 Purified Water Systems Comparison Chart
Ratings: N = None L = Low M = Medium H = High
Notes for Table 5-4:
1) High-water consumption unless wastewater is reused - cooling tower makeup, etc.
2) Total chemical requirement dependent upon pretreatment selection
3) Total water consumption dependent upon pretreatment selection.
4) USP TOC requirement is met in most cases but may not be if feed water is high TOC (>2 ppm)
5) High microbial performance refers to low microbial count in relative terms
CAPITAL COST
CHEMICAL
HANDLING
ENERGY
CONSUMPTION
WATER
CONSUMPTION
OUTSIDE
SERVICE COSTS
OPERATIONAL
MAINTENANCE
PRODUCT
CONDUCTIVITYMICROSIEMEN /
CM @ 25°C
PRODUCT TOC.PPB
MICROBIAL
PERFORMANCE
Off-Site
Regenerated
Ion Exchange
L
N
L
L
H
L
1.0 - 0.06
(4)
L
Reverse Osmosis/
Off-Site Regenerated
Ion Exchange
M
L
M
H (1)
M
M
1.0 - 0.06
<500
M
On-Site
Regenerated
Ion Exchange
M
H
L
M
L
M
1.0 - 0.06
(4)
L
Reverse Osmosis/
On-Site Regenerated
Ion Exchange
M
M
M
H (1)
L
M
1.0 - 0.06
<500
M
Reverse Osmosis/
Continuous
Electrodeionization
M
L
M
H (1)
L
M
1.0 - 0.07
<500
M
Two-Pass
Reverse
Osmosis
M
L
M
H (1)
L
M
2.5 - 0.5
<500
H
Distillation
H
L (2)
H
M (3)
L
L
1.0 - 0.1
<500
H
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 74/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 75/250
FINAL TREATMENT OPTIONS:
WATER FOR INJECTION (WFI)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 76/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 77/250
63
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6. FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.1 INTRODUCTION
This chapter addresses the USP approved final treatment methods for the production of compendial WFI.
WFI is the purest grade of bulk water monographed by the USP and would be expected to be used for themanufacture of parenteral, some ophthalmic and inhalation products, and for finishing steps of parenteralgrade active pharmaceutical ingredients (API’s).
Recommended systems include either distillation or RO as the final processing step, but may also include
ultrafiltration (UF), deionization (DI) and/or ion exchange (IX), to compliment the RO or distillation unit opera-tion.
The technology, operation, maintenance, and relative cost issues for the approved process methods arediscussed. This chapter includes USP monograph information, regulatory issues, and subsections that cover
the unit operations:
• Single effect (SE) distillation
• Multi-effect (ME) distillation
• Vapor compression (VC) distillation
• Reverse osmosis (RO)
Feed water pretreatment is covered along with economic factors such as construction materials, surface
finishes, and instrumentation and controls. A comparison table on USP-WFI final treatment options andrelative attributes is provided.
6.2 US PHARMACOPOEIA ISSUES
The United States Pharmacopoeia (USP) allows WFI to be “purified by distillation or by RO”. This statement
does not imply that the regulated process step is the only process step, but the USP advisory section doesimply that it is the final step in the process.
• Only distillation may be used to produce WFI under current European regulations
• Distillation, RO and UF are allowable methods to produce WFI under the Japanese regulations
There are few regulations, which govern the design and construction of pharmaceutical water purificationsystems. There are no existing regulations governing materials of construction, type, or level of instrumenta-
tion, surface finish, or operating temperatures. Most practices commonly followed with respect to these andother issues, have been adopted based on many factors.
Among U.S. government publications, including the Code of Federal Regulations (CFR) and the FDA Guideto Inspection of High Purity Water Systems, there are few stipulations related to design and construction of
WFI processing equipment. Two notable stipulations are:
• “Heat exchangers, other than the double concentric tube type or double tube sheet type, must employ apressure differential and a means for monitoring the differential.”
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 78/250
64
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
• “All stills and tanks holding liquid requiring microbial control shall have air vents with non-fiber releasingsterilizable filters capable of preventing microbial contamination of the contents.”
6.3 DISTILLATION
The pharmaceutical still chemically and microbiologically purifies water by phase changes and entrainmentseparation. In this process water is evaporated, producing steam. The steam disengages from the waterleaving behind dissolved solids, non-volatiles, and high molecular weight impurities. However, low molecular
weight impurities are carried with the water mist/droplets, which are entrained in the steam. A separatorremoves fine mist and entrained impurities, including endotoxins. The purified steam is condensed into WFI.
Distillation systems are available to provide a minimum of 3 log10 (99.99%) reduction in endotoxin concentra-tion. Specific endotoxin loading limits should be reviewed with the manufacturer.
A variety of different designs are available including single effect (SE), multi-effect (ME), and vapor compres-sion (VC). The distilled water quality expected from an SE still is equivalent to an ME design, by virtue of the
fact that water is distilled only once in both systems. The benefit to the user of ME versus SE distillation arethe significantly lower operating costs associated with utilities.
In an ME system, purified steam produced by each effect is utilized to heat water and generate more steamin the subsequent effect. Due to this staged evaporation and condensation process, only the first effectrequires heat from an external source, and only the purified steam produced by the final effect is condensed,
using an external cooling medium.
VC stills can produce similar quality water using a different technique. Energy imparted to the generated
steam, by a mechanical compressor, results in compressed steam with increased pressure and temperature.The higher energy steam is then discharged back into the evaporator/condenser vessel to generate more
steam in a continuous cycle.
Areas of concern are carry over of impurities, evaporator flooding, stagnant water, and pump and compres-sor seal design. These concerns may be addressed using mist eliminators, high water level indicators, use ofsanitary pumps and compressors, proper drainage, adequate blow down control, and conductivity sensing to
divert unacceptable water to drain.
6.4 DISTILLATION APPLICATIONS AND CAPACITIES
The majority of USP WFI currently produced in the United States is produced by distillation. WFI productionis shared by both ME and VC stills. SE stills are found in areas where only small quantities of WFI are
required. However, where large amounts of WFI are required, economics of operation dictate the use of eitherME or VC.
Table 6-1 shows typical capacities and temperatures of WFI produced by each process.
Table 6-1 Capacities and WFI Temperature Options
Vapor compression
100 - 6,000
• Ambient• 80 - 100
• Combination Ambient/Hot
CAPACITY RANGE IN GPH
WFI Temperaturerange/Options, °C
Single-Effect
1 - 100
80 - 100
Multi-Effect
25 - 3,000
37 - 100
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 79/250
65
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.5 PROCESS AND SYSTEM DESCRIPTION
6.5.1 Single Effect Distillation (SE)
SE systems incorporate a single evaporator heat exchanger, separator mechanism, and a condenser.
SE systems are available in electrically or steam powered versions, although electrical units are limited tovery small production rates (<30 gallons per hour).
Steam powered units typically require 30-60 psig plant steam. Cooling fluid is required for both steam andelectric powered versions. When water is the coolant, the rate is approximately 8-10 gallons per gallon of WFI
produced, based on a supply temperature of 4 - 16°C, and temperature rise of 67°C.
SE systems typically operate at atmospheric pressure and 100°C, and incorporate non-ASME code vessels.
WFI is delivered at atmospheric pressure and 80 - 100°C; thus a distillate transfer pump is required, unless
the WFI tank is at a lower elevation than the still.
6.5.2 Multi-Effect Distillation (ME)
ME systems incorporate two or more evaporator heat exchangers, separator mechanisms, and a condenserinto a staged evaporation and condensation process. Typical systems have 3-8 effects. Each effect includes
an evaporator and a separator (see Chapter 11).
ME systems typically require plant steam at 80 - 120 psig, and cooling fluid at a supply temperature of 4°C -
16°C, based on a temperature rise of 65°C - 70°C. The quantity of steam and cooling fluid required variessignificantly based upon the WFI production rate and the number of effects. Capital costs increase while
steam and cooling fluid consumption decrease, as the number of effects to produce a given quantity of WFIincreases. ME systems operate under pressure, and typically deliver WFI at 80°C - 100°C.
Normally, water used for cooling is not the same as the feed water, and does not require special pretreatmentfor the purpose of scale prevention. However, corrosion prevention measures, such as chlorine and chloram-
ine removal, are necessary.
Some designs deliver the water at atmospheric pressure and require a transfer pump unless the WFI storagetank is at lower elevation than the still. Other designs which may operate at 5-10 psig condenser pressure, do
feature a distillate transfer pump for higher pressure deliveries.
6.5.3 Vapor Compression Distillation (VC)
VC is a distillation method where water is evaporated inside, or outside, a bank of tubes arranged in a
horizontal or vertical configuration. The horizontal design is normally of the forced circulation type with re-circulation pump and spray nozzles, while the vertical design is of the natural circulation type.
Major system components are the evaporator, compressor, heat exchangers, deaerator, pumps, motors,valves, instruments, and controls.
The VC process operates on the same principle as the mechanical refrigeration cycle.
In a VC still, feed water is evaporated on one side of the tubes. The generated steam passes through the
disengagement space, through the separator, and into the compressor.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 80/250
66
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
The energy imparted by the compressor results in compressed steam with increased pressure and tempera-ture. The higher energy steam is then discharged back into the evaporator/condenser vessel. There, the
steam condenses and gives up its latent heat, which is transferred through the tube wall to the water. Morewater is boiled off, generating more vapor, and the process is repeated. The outgoing distillate and blow down
streams preheat the incoming feed water, thus saving energy. Since the latent heat is recycled, there is noneed for a stand-alone condenser as in the SE or ME systems.
6.5.4 Distillation pretreatment requirements - General
All distillation units are susceptible to scaling and corrosion, if the appropriate feed water pretreatment is notprovided. VC and some SE stills operate slightly above atmospheric pressure, and the removal of calcium
and magnesium, by way of water softening, is normally required as a minimum. ME stills operate at a muchhigher pressure and temperature, and require higher quality feed water in order to prevent scaling and corro-
sion. Normally, ion exchange beds are employed as feed water pretreatment to a multiple effect still. RO isalso used as feed water pretreatment for either the VC or multiple effect stills. All distillation units will invariablyexperience some form of scale build-up and must therefore include routine visual inspections plus cleaning of
the still during shutdown periods when appropriate. Both types of stills are susceptible to attack by chlorine.Chlorine removal is essential if damage is to be avoided. Activated carbon filters and sodium bisulfate injec-
tion are effective and common methods for chlorine removal.
From a microbiological perspective, the bacterial and endotoxin load should be consistently controlled to alevel that does not overload the still.
6.5.5 Pretreatment requirements – Specific
6.5.5.1 Pretreatment for Single Effect Still (SE)
See the “Pretreatment for ME still” paragraph and above for general information on distillation pretreatment.
6.5.5.2 Pretreatment for Multi-Effect Still (ME)
The baseline pretreatment for ME must provide very low TDS feed water, preferably less than 10 mg/l, and
less than 1 mg/l silica. Some manufacturers offer ME stills to operate on softened water. Others allow higherlevels of silica, up to 5 mg/l. The pretreatment must also remove chlorine and objectionable volatiles, such as
ammonia, if present.
Figure 6-1 A baseline system to achieve very low TDS may be DI or RO.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 81/250
67
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
See Chapters 4 and 5 for more information on pretreatment.
6.5.5.3 Pretreatment for Vapor Compression Still (VC)
See Section 6.5.4 for general information on distillation pretreatment requirements. The baseline pretreat-ment for VC stills is softening, the removal of chlorine, and other objectionable volatiles such as ammonia, if
present.
6.5.5.4 Economics
a) Economics of the Single Effect Still: Commercially available SE systems are inherently simple indesign, configured similarly, and offered with significantly fewer options, compared to ME and VC sys-
tems. As a result, fewer factors affecting costs are applicable by comparison. Operating costs of SEsystems are associated mainly with plant steam and cooling fluid. Utilities consumption rates are fairly
consistent among SE manufacturers.
b) Economics of the Multi-Effect Still: Although all commercially available ME systems are configuredsimilarly and supplied with the same basic components, opportunities for cost savings exist in the areasof construction materials, surface finishes, and instrumentation. Operating costs of ME systems are
associated mainly with plant steam and cooling fluid. Utilities consumption rates vary among ME manu-facturers.
c) Economics of the Vapor Compression Still: Significant opportunities exist to reduce capital cost asso-
ciated with selection of construction materials, surface finishes, and instrumentation used in the con-struction of VC stills. Operating costs of VC systems are associated mainly with electrical power.
6.5.6 Recommended Construction Materials
Materials shown Table 6-2 are based on available designs by leading manufacturers. However, other materi-
als may be utilized based on the technical application.
Figure 6-2 Baseline Treatment for VC Stills
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 82/250
68
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.5.6.1 Surface Finish
Mechanical polishing (MP), electropolishing (EP), and passivation processes are implemented in stainlesssteel distillation systems in order to improve corrosion resistance. These processes are neither necessary,
nor applicable, to other alloys such as tin-coated copper, titanium, and Inconel, based on differences in metalchemistry.
MP and EP/passivation processes affect the microscopic amplitude and chemical composition, respectively,of the stainless surface. These processes are not considered necessary to control microbial growth due to
the relatively high operating temperatures. MP is advocated for final finishing of mechanical welds and EP/ passivation for all stainless steel surfaces to optimize the formation of the corrosion resistant chromium oxide
barrier.
The impact of progressive mechanical polishing on the capital cost of stills and other equipment is consider-able, and often can account for 25% to 30% of a ME or VC still cost. MP processes, except when used tosmooth out a mechanical weld or misalignment etc., may be removed from the applicable specification with-
out fear of compromising the water quality.
*Some manufacturers may use sanitary clamps on distillate piping only.
Vapor Compression
304/L SS
304/L SS
304 SS
316 SS
316/L SS
304L SS
316/L SS, Sanitary
clamps*
316 / 316L SS with
Sanitary clamps*
316 SS / Inconel
316 SS
Carbon Steel
All 316 / 316L SS
EvaporatorShell
Tubesheets
Tubes
Separator
HeatExchangers
Deaerator
Piping
Pumps
Compressor
Valves
Skid/Frame
Optional
Single-Effect
SS / Tin coated copper
SS / Tin coated copper
SS
SS / Tin coated copper
SS or Tin coated copperwhen used
Not used
SS / Tin coated copper
SS when used
Not used
SS
Carbon Steel
Multi-Effect
316L SS
316L SS
316/L SS; Titanium
316 SS
316/L SS
Not Used
316/L SS, Sanitary
clamps*
316 / 316L SS with
Sanitary clamps*
Not Used
316 SS
Carbon Steel
Other tube material
options are available
Table 6-2 Materials of Construction
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 83/250
69
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.5.6.2 Instrumentation and Controls
For WFI applications, the level of instrumentation should be sufficient to monitor parameters consideredcritical because they relate to ensuring proper hydraulic/thermodynamic functionality and the production of
the appropriate quality of WFI. Instrumentation for critical operating parameters should be calibratable usingNational Institute of Standards and Technology (NIST) traceable equipment
6.5.6.3 Advantages and Disadvantages
Table 6-3 Process and Steam Comparison
Configuration
• Evaporator
• Condenser
• Feed/Blowdown heatexchangers
• Compressor
• ASME Coded
• Distillate pump
• Blowdown pump
• Feed booster pump
Makeup heat
• Steam pressure
Cooling water
Plant Steam Pressure, psig
Feed Water Pressure, psig
Condensor Cooling Water
Pressure, psig
MAY BE USED TO
GENERATE CLEAN STEAM
Single-Effect
YES
YES
NO/optional
NO
Normally NO
Normally NO
NO
NO
Steam or electric
LOW
YES
30 - 60
30 - 50
30 - 50
YES (not common
practice)
Multi-Effect
YES
YES
YES
NO
YES
Optional
NO
YES
Steam
HIGH
YES
100 - 120
75 - 90
30 - 50
YES
Vapor Compression
YES
NO
YES
YES
NO
YES
YES
NO
Ambient WFI: -Steam
or electric
HOT WFI: -Steam
LOW
NO
30 - 40
30 - 50
Not required
NO
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 84/250
70
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.6 REVERSE OSMOSIS (RO)
RO employs a semi-permeable membrane and a relatively high pressure differential to force water throughthe membrane to achieve chemical, microbial, and endotoxin reduction, critical in USP WFI applications. The
feed water is converted into two streams, permeate and reject. The permeate water flows through the mem-brane and is produced cold and as such does not have the temperature protection for microbial growth
afforded by the alternate distillation processes. The reject stream discharges comparatively smaller volumethan the permeate, and contains vir tually all of the feed water contaminants.
6.6.1 Application
RO systems are used as USP WFI pretreatment for distillation processes, or as final treatment for USPPurified water systems. RO is also an accepted means of producing WFI, and may provide a low capital and
operational cost alternative to distillation.
Membranes that are hot water sanitizable at 80°C are now available for pretreatment and final treatment, thus
eliminating the need for chemical sanitization and simplifying the validation process. These membranes stillrequire periodic chemical cleaning.
Membranes, which may allow for continuous operation at 80°C, are under development. This may have asignificant impact on the use of RO as a means of producing USP WFI, since operation of the system at 80°Cmay nearly eliminate biological concerns. Failure of a membrane or seal will result in permeate contamina-
tion. These problems may be controlled by:
• Pretreatment of the feed water
• Appropriate membrane material selection
• Latest technology membrane design
• Integrity challenges
• Periodic sanitization
• Monitoring of microbial levels, conductivity, total organic carbon and differential pressures
6.6.2 Description
Semi-permeable RO membranes are produced commercially for water purification in spiral wound and hol-
low fiber configurations. RO membranes are permeable to some substances such as water and dissolvedgases, while impermeable to other substances such as salts, high molecular weight organics, acids, bases,
colloids, bacteria, and endotoxins. Membranes are available in four basic materials; cellulose acetate, polya-mide, thin film composite, and polysulfone. (Polyamide membranes are virtually identical in performance to
thin film composite membranes.) All three membrane types have advantages and disadvantages. (See Chapter5 for more details.)
Bacteria and endotoxin removal, required for WFI applications, can be performed at ambient temperatures.This significantly reduces utility costs compared to alternative elevated temperature processes (distillation).
By operating at ambient temperatures, distribution piping may not require insulation and may not need to beconstructed of stainless steel.
For WFI applications, opportunities exist for enhanced control of the single pass unit, by utilizing multi-pass-product-staged or other combination designs. These configurations improve reliability and efficiency, while
improving water quality and quality assurance over the single pass design.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 85/250
71
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.6.3 Pretreatment Requirements
RO, as the final processing step, may require pretreatment using ion exchange, deionization, RO, and/orultrafiltration to improve operability and quality attributes.
Pretreatment requirements normally include gross particle filtration, scale prevention, and chlorine removal.
Carbon dioxide and ammonia gas, are not removed by the RO process, and may be removed by degasification,caustic addition, ion exchange, or electrodeionization, prior to the final RO process step. (See Chapter 5 formore details.)
Due to the stringent microbial and endotoxin control required for parenteral and other critical applications, the
pretreatment prior to the RO should incorporate additional provisions for control and monitoring of microor-ganisms.
Disinfectants, such as chlorine or chloramine, should be maintained, when tolerable, at appropriate levelsthroughout the pretreatment chain. Stagnant water resulting from surge tanks or dead legs should be avoided
by design, or by the inclusion of recirculation systems, which should include In-Line microbial control devices,such as UV sterilizers.
Regular and appropriate sanitization and cleaning of all unit operations subsequent to and associated with,the disinfectant (chlorine or chloramine etc.) removal should be scheduled, to maintain and complete themicro-organism control of the pretreatment system. (See Chapter 4 for further details.)
6.6.4 Economics
Opportunities are available to reduce capital costs associated with the selection of construction materials,surface finishes, and instrumentation used in the construction of RO units without compromising the water
quality. Operating costs of RO systems are associated mainly with replacement membranes, water concen-trate discharge, electrical power, cleaning and sanitizing chemicals, replacement filters, and pretreatment
cost.
6.6.5 Construction Materials
Construction material selection for RO systems are driven by:
• Structural integrity, based on high operating pressure
• Structural integrity, based on low pressure sections ahead and after the membranes
• Chemical compatibility with the contact fluid and its constituents
• Need to control micro-organism growth
The low operating temperature of the RO system allows the use of non-metallic construction materials.Sanitary piping and valves are generally optional features for RO systems, based on the specific manufac-turer and location of the RO in the treatment chain.
For the final purification step, it may be very cost effective to utilize mill finish 304 stainless steel for the feed
and concentrate waste piping for the system, maintaining 316L stainless steel or PVDF and sanitary designfor the product piping only.
6.6.6 Surface Finish
MP and EP processes are not applicable to non-metallic systems.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 86/250
72
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
6.6.7 Instrumentation and Controls
RO control systems usually use local control and indication, and do not typically require programmable logiccontrollers (PLC) as a standard feature. The type and level of instrumentation is similar among manufactur-
ers. The level of instrumentation should be sufficient to monitor parameters considered critical because theyrelate to ensuring proper hydraulic functionality and the consistent production of quality WFI. Instrumentation
for critical operating parameters should be calibratable using NIST traceable equipment. (See Chapter 9 formore details.)
The typically monitored operating parameters for an RO system are feed pH, feed conductivity, and productquality (TOC and conductivity). These three parameters should be measured using calibratable, NIST trace-
able instruments. Recording data may be accomplished manually or electronically using analog instrumentsand paper/paperless recording systems.
6.6.8 Advantages and Disadvantages
Multi-pass can, in most cases, produce water quality consistent with the minimum requirements of USP WFI.In cases where the feed water quality is such that this is not possible, the use of some type of deionization
(e.g., additional RO, UF, ion exchange, or electrodeionization) as a pretreatment may be required. This is to
allow the final point of purification to remain RO and the system to generate consistent and reliable waterwithin the USP WFI specifications.
Advantages associated with the design and operation of RO units used as the final treatment step for theproduction of WFI are:
1) Depending on cost and complexity of pretreatment, RO systems designed for production of USP WFImay provide for significantly reduced capital costs when compared with distillation processes, while
maintaining the appropriate USP WFI quality.
2) The utility requirements are significantly lower for RO systems (electricity for pump horsepower) than fordistillation, resulting in lower operating costs, which may be a very significant factor over the lifetime ofthe system.
Disadvantages associated with the design and operation of RO units used as the final treatment step for the
production of WFI are:
1) Membrane fouling and integrity
• Bacteria grow-through
• Seal leakage or by-passing
• Seal failure or damage caused by chemical attack etc.
• Membrane damage during installation etc.
• Membrane damage due to chemical or high temperature attack
2) Membrane material sensitivity to bacteria and sanitizing agents.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 87/250
73
FINAL TREATMENT OPTIONS: WATER FOR INJECTION (WFI)
3) Inherent sanitization limitations
• Periodic chemical or hot water sanitization may be required
• Periodic chemical cleaning may be required
4) Pretreatment cost may be high
RO systems provide a method for consistently producing ambient water in accordance with the USP WFI
requirements. This not only reduces utility requirements, but also may reduce installation costs, since thermalinsulation may not be required for storage and distribution.
6.7 USP - WATER FOR INJECTION SYSTEMS COMPARISON
Table 6-4 WFI Systems Comparison
Ratings: L = Low M = Medium H = High
Notes:
1) All Indicators are relative to each other within the specific category2) Optimum design and operating conditions are assumed
3) Total water consumption is dependent on pretreatment selected4) RO may not meet USP TOC levels if feed water TOC is high (>3 ppm)
RO DISTILLATION
UNIT OPERATION (1,2) 2 pass RO SE ME VC
Capital Cost M M H H
Chemical Consumption L N/A N/A N/A
Energy Consumption M H H H
Water Consumption M (3) H M M
Outside Service Costs L L L L
Operational Maintenance L L L L
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 88/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 89/250
PHARMACEUTICAL STEAM
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 90/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 91/250
75
PHARMACEUTICAL STEAM
7. PHARMACEUTICAL STEAM
7.1 INTRODUCTION
This chapter aims to simplify and standardize the process of selection, programming, and design of pharma-
ceutical steam systems. Guidelines, information, and options are provided, along with advantages and disad-vantages, based on the best and most cost effective of current and proven practices and technologies.
The absence of regulations governing the use of steam in pharmaceutical processes has resulted in theproliferation of differing practices and interpretations. Most interpretations are made on the side of conserva-
tism. Unfortunately, in addition to increasing cost without an associated increase in benefits, excessive con-servatism can result in system complexity, and possibly reduced reliability. One example is the use of clean
steam (non-utility boiler produced steam) where a form of utility steam (utility boiler produced steam) wouldbe adequate to maintain product quality. The installation of a clean steam generator when a simple steamreducing station would suffice results in added equipment and the associated impact on cost, complexity, and
reliability.
In some instances, interpretations are based on inaccurate assumptions of what is important or critical. An
example is the over specifying of pretreatment or using WFI as feed to solve the perceived problem.
The chapter establishes standard definitions for terms commonly associated with pharmaceutical steam and
provides information that facilitates making correct and cost effective decisions.
7.2 cGMP ISSUES
The user has the ultimate responsibility for system design and performance, and for ensuring that the propertype of steam is used for a given process.
There is no FDA or USP minimum standard for clean steam. However, cGMPs for large volume parenterals(LVPs) issued in 1976 indicated that feed water for boilers supplying steam that contact components, drug
products, and drug product contact surfaces shall not contain volatile additives such as amines or hydra-zines.
Few regulations govern the design and construction of clean steam generators. There are also no regulations
governing materials of construction, type or level of instrumentation, surface finishes, or operating tempera-tures.
Among US Government publications, the FDA’s Code of Federal Regulations (CFR) provides culinary steamrecommendations and stipulations related to heat exchanger and tank air vents design and construction. The
Culinary steam recommendations apply to food applications only.
US Public Health Service/Dairy Industry Committee, 3A Sanitary Standards, Number 609-02, adds addi-tional limitations to Culinary steam feed water additives for food applications. It should be noted that boilerfeed water additives permitted in food for human consumption may not be acceptable in drinking water or
orally ingested drug products.
7.2.1 Steam Attributes
7.2.1.1 Quality
The term “Quality” when referring to steam indicates the level of steam saturation. There are no FDA or USP
regulations relating to minimum “steam quality” or the level of non-condensable gasses present in pharma-ceutical steam. (See Section 7.4.)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 92/250
76
PHARMACEUTICAL STEAM
European regulators have defined specific criteria for pharmaceutical steam used for equipment sterilization.(European Standard EN 285 - Steam Sterilizers - reference section 13.3) These cover acceptable levels of
saturation or dryness, the level of superheat, and the volume of non-condensable gases present.
7.2.1.2 Purity
Purity requirements for steam used in pharmaceutical manufacturing and product development are driven bythe product characteristics, manufacturing process, and the intended use of the product. The product manu-facturer is responsible for ensuring that steam used to process the product is appropriate.
Though steam purity requirements are product specific, it may be impractical to reliably produce special
steam for each situation. Manufacturing operations typically generate and distribute only one or two steampurity grades, commonly grouped.
7.3 TYPES OF STEAM
Pharmaceutical steam is classified into two (2) types based on their respective sources. These are:
1) Utility-Boiler produced steam, hereafter called Utility Steam.
2) Non-Utility Boiler produced steam, hereafter called Clean Steam.
7.3.1 Utility Steam
Utility steam is characterized with usually having:
• Chemical additives to control scale and corrosion
• Relatively high pressure with the potential of generating superheat during expansion
• Relatively high pH
Chemical additives: Utility steam is produced, in most cases, using conventional fire-tube steam boilers,
normally of steel construction. Such boilers are almost always provided with systems that inject additives inthe feed water to protect the boiler and steam distribution piping from scale and corrosion. Some of these
scale and corrosion inhibitors may, and often do, include amines and other substances that may not beacceptable in steam being used in pharmaceutical processes. The user must determine what additives areused, and verify if they are acceptable in the particular application, i.e., do not add any impurities or create a
reaction in the drug product.
Utility steam can be filtered to remove particulate matter, but filtration does not remove dissolved substancesand volatiles such as amines.
Superheat: Superheated steam is produced in water tube boilers by reheating the steam or by generatingthe steam at a higher pressure in a fire tube boiler and then reducing the pressure through a regulating valve.
When the pressure is reduced, the energy in the higher temperature steam is dissipated to generate steam atthe lower pressure and produce superheated steam above the corresponding saturation temperature. Super-
heat is dissipated downstream of the regulating valve due to heat loss in the line.
pH control: In order to protect carbon steel from corrosion by the steam, it is necessary to use additives toraise the pH to between 9.5 -10.5.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 93/250
77
PHARMACEUTICAL STEAM
7.3.2 Clean Steam (CS)
Pharmaceutical clean steam is generated from treated water free of volatile additives, such as amines orhydrazines, and is used for thermal disinfection or sterilization processes. It is considered especially impor-
tant to preclude such contamination from injectable drug products:
Clean steam is characterized as having:
• No additives
• No generated superheat except when the generated pressure is significantly higher than the use pres-
sure of the steam. (See Section 7.3.1 - Superheat.)
• Relatively low pH
There are many terms used in the pharmaceutical industry to describe Clean Steam. These include Clean
Steam, Pure Steam, Pyrogen Free Steam, WFI Steam, and USP Purified Water steam. There is no standardor accepted definition for any of these terms. However, the most commonly used terms are “Pure Steam” and
“Clean Steam”. In this Guide, the term “Clean Steam” is used in lieu of all others.
The condensate of Clean Steam has no buffer, and may have a relatively low pH compared to that of utilitysteam.
7.4 BACKGROUND AND INDUSTRY PRACTICES
7.4.1 Purity of sterilizing steam
When steam or the resulting condensed water comes in direct or indirect contact with the drug product, the
purity should be equivalent to the water purity acceptable for final rinsing of the drug contact surfaces.
Note: A continuous supply of Dry Saturated Steam at the point of use is considered necessary for efficient
steam sterilization. Water carried by the steam in suspension may cause damp loads and superheated steamis considerably less effective than saturated steam when used for sterilization. Non-condensable gases if
contained in the steam may prevent the attainment of sterilization conditions in parts of the sterilizer load.
7.4.2 Steam used for humidification
When steam is used for indirect humidification, such as injection into HVAC air streams prior to final air
filtration, the steam does not need to be purer than the air that it is being mixed with. However, when humidi-fying process areas, the potential level of impurities, including amines and hydrazines should be evaluated in
order to ascertain the impact on the final drug product. This is particularly important in areas where openprocessing takes place, such as aseptic filling suites and formulation areas. If the diluted water vapor is found
to contribute significantly to the contamination of the drug, a purer grade of steam should be selected.
7.4.3 Common practices
It is common practice to generate pharmaceutical steam from compendial waters and test the steam conden-
sate for equivalency to the compendial standard. This practice ignores the ability of the pharmaceutical steamgenerator to remove impurities. This overprocessing is wasteful and unnecessary. An exception is when the
steam quantity is small and the cost and maintenance of a dedicated feed water pretreatment system ex-ceeds the cost of using compendial water. Pharmaceutical Clean Steam is commonly used in applicationswere utility steam would suffice; such as non-critical room humidification and high purity water heat exchang-
ers.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 94/250
78
PHARMACEUTICAL STEAM
Table lists the commonly accepted industry standards and highlights the trend in the pharmaceutical indus-try to provide “purer than necessary” steam and over-specified feed water.
7.4.4 Industry and Baseline Practices in the Production of Steam
Table 7-1 Practices in the Production of Steam
Method of generation of steam
The use of a sanitary clean steam generator with
entrainment for the control of endotoxins & liquidcarry-over (SCSG) is both baseline and commonindustry practice.
The use of an SCSG is both baseline and com-
mon industry practice.
SCSGs are commonly used; however, utility steam
is the acceptable baseline application.
WHILE THE USE OF A SCSG IS COMMONPRACTICE, AN ALTERNATIVE APPROACH IS
TO USE UTILITY STEAM PLUS HOT USPWATER, FLUSHING & WASTE TESTING.
SCSGs are commonly used and are the Baselineapplication.
SCSGs are commonly used but utility steam maybe totally acceptable.
Where open processing takes place and wherethe potential level of amines, hydrazine’s etc. in the
condensate has been determined to have adetrimental effect on the drug product the baselineand common practice is the use of a SCSG.
However, if it has been determined that the
impurities have an insignificant effect on the drugproduct, a utility steam source would qualify as thebaseline approach.
It is common practice to use a SCSG as theenergy source. The baseline approach would be to
use a utility steam source coupled with a cGMPheat exchanger design.
It is common practice as well as the baseline
approach to use utility steam.
Intended Use of Steam
Parenteral and Non-Parenteral Dosage form
applications where steam is in direct contact withthe drug.
Critical step in the manufacture of API where
steam is in direct contact with the Active Pharma-ceutical Ingredient (API).
Non-Critical step in the manufacture of an API
where added impurities may be removed in asubsequent step.
Sterilization of USP water systems.
Process humidification for dosage form applicationwhere steam is in direct contact with the drug,
where open processing takes place and where thepotential level of amines, hydrazine’s etc. in the
condensate has been determined to have adetrimental effect on the drug product.
Humidification of non-critical HVAC systems suchas rooms and areas where the drug is not directly
exposed to the room atmosphere.
HUMIDIFICATION OF PROCESS & CRITICALCLEANROOMS.
Energy source for non-critical & cGMP heatexchangers.
Sterilization of fermentation vessels.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 95/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 96/250
80
PHARMACEUTICAL STEAM
7.5 SYSTEM PLANNING
Pharmaceutical Steam System Planning, shown in the Figure 7-1 is a graphic representation of the systemboundaries, limitations, and restrictions. Initial system planning reveals primary boundaries that establish the
cornerstone for design criteria. These system boundaries are Steam Requirements, System Design, UsePoint Criteria, and Distribution System requirements.
The arrows encircling each boundary represent limitations that establish more specific operating strategiesand ranges. To allow more flexibility in final planning and detailed design the designer should always indicate
ranges of acceptability, rather than a specific value or position.
7.5.1 Steam Requirements
The planning process starts with the listing of all steam requirements and applications that include:
• Company standards including QA/QC requirements and published Sop’s
• The categorization of use-point by:
• Type of application (Humidification, critical or non-critical, API, and Dosage for applications)
• Purity selection (this is based primarily on the application and the endotoxin and chemical purity
criteria set for the product for which the steam, or its condensate, will be in contact with. The selectionmust consider underlying factors which have impacts on purity control, installed and operating cost,maintenance, and practicality)
• Steam quality (dryness, non-condensable limits, and maximum superheat)
7.5.2 System Design
Pharmaceutical steam is generated using different methods. The most appropriate method for each applica-tion must be selected. (See the Pharmaceutical Steam Purity Decision Tree, Section 7.6.)
The process continues with an evaluation of the steam system requirements (generation) that includes: the
selection of the type of generation system that would satisfy each category, which would include:
• The types of generation systems available. (If both pyrogen free clean steam and clean steam withoutendotoxin limits is required, the practicality and economy of producing only the higher grade should beraised.)
• The source of utility steam or electrical power (The plant steam requirement for clean steam as well as
utility steam and the option of electric powered steam generators should be considered.)
• The type and number of systems required based on feedback from the “Distribution System” evaluation
• The condensate sampling needs
• Safety considerations
7.5.3 Use Point Criteria
The third step defines the specific delivery requirement ranges for clean steam at the point of use including:
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 97/250
81
PHARMACEUTICAL STEAM
• Utilization, which is determined from each overall system peak demand(s), average demand, and therelationships between peak demand time periods and their flow rates.
• Pressures and flow levels
• Use periods and histogram analysis, if available
• Quality
• Purity
7.5.4 Distribution System
The fourth step includes the distribution system evaluation, which includes:
• Condensate, non-condensable and moisture removal
• Pipe size and Insulation requirements including:
• Materials of construction, sanitary design requirements and surface finish
• Physical location of each use point
• Heat and temperature losses
• Natural drainage
Note: Since the steam quality will decline, due to heat losses, with time, the efficiency of the insulation andthe length of the distribution system, the quality at the use point will not be expected to reflect the generation
quality level.
7.5.5 Re-evaluation of system boundaries and constraints
These sequential steps are repeated and re-evaluated as information in the design process iterates, and
further criteria about the overall system boundaries are identified. (See Figure 7-3.)
In operations with a requirement for only one grade of steam, the steam system is designed to meet the moststringent requirements of the most demanding product or process. With more than one purity grade of steam,products and processes are often categorized and fed by the most appropriate system. The number of types
of steam generated is most often a function of the volume of steam consumed and variation of purity re-quired.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 98/250
82
PHARMACEUTICAL STEAM
7.6 PHARMACEUTICAL STEAM PURITY DECISION TREE
Figure 7-2 Pharmaceutical Steam Purity Decision Tree
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 99/250
83
PHARMACEUTICAL STEAM
7.7 PROCESS AND SYSTEM DESCRIPTION
7.7.1 Utility Steam
Utility Steam is produced in conventional plant utility boilers whose typical design and construction are wellknown and will not be covered in this chapter.
7.7.2 Clean Steam (CS)
Clean Steam is produced in specially designed non-fired generators or from the first effect of multi-effect WFIstills, which do not use scale or corrosion inhibitor additives. The generator is fed with water pretreated for the
purpose of removing elements that contribute to scaling or corrosion, and the materials of construction areresistant to corrosion by steam that has no corrosion inhibitors.
The dedicated CS generator is very similar in design and construction to the first effect of a multi-effect still.For information on multi-effect (ME) stills, see Chapter 6.
7.7.2.1 CS obtained from a ME still
When Clean Steam is obtained from the ME still, the first effect is usually fitted with two valves; one to isolatethe remaining effects and the other to isolate the Clean Steam use points. Depending on the manufacturer,the still may or may not produce steam when the still is producing WFI.
Advantages:
• Does not require a separate generator with the associated cost, space, installation, operation, and main-tenance
Disadvantages:
• Output is limited to the capability of the first effect of the ME still
• May not produce steam when the still is producing WFI. In an ME, the steam generated in the first effectbecomes the motive (power) steam for the second effect, which in turn produces motive steam for the
third effect, etc. Therefore, the impact of the diverted steam is multiplied by the number of effects, andWFI production is significantly reduced.
The still manufacturer should be consulted in advance, if simultaneous production of WFI and Clean Steam isdesired.
7.7.2.2 CS Produced from a Sanitary Clean Steam Generator
CONFIGURATION OF A TYPICAL SANITARY CS GENERATOR
There are various designs of CS generators. All are evaporators.
They can be of the vertical or the horizontal type, depending on the manufacturer and the overhead spaceavailable.
The disengagement space and the separator may be housed in the same vessel as the evaporator or in a
separate vessel.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 100/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 101/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 102/250
86
PHARMACEUTICAL STEAM
7.8 SIZING THE CLEAN STEAM SYSTEM
Figure 7-3 Sizing the Clean Steam System
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 103/250
87
PHARMACEUTICAL STEAM
7.8.1.1 CS Produced from a Simple Clean Steam Generator
There are applications where pyrogen free steam and sanitary construction features are not required, and atthe same time, Utility Steam cannot be used. In such cases, it may be most economical to utilize a simple
Clean Steam generator of the most economical design. Savings may be worthwhile when the elimination ofthe steam separator is combined with non-sanitary features such as:
• Non-Sanitary pipe and fittings
• Non-sanitary instruments and valves
• No polishing
• Minimum controls
The elimination of the separator alone does not provide significant cost savings. It is important to remember
that the separator’s function is more than removal of endotoxins. It removes entrainment, which includes alltypes of contaminants present in the feed water, except volatiles. Without an entrainment separator, impuri-
ties from the feed water may well be entrained in the steam and the moisture content of the steam as it leaves
the generator, can be much higher than in the standard entrained generator. Thus the feed water puritybecomes a critical factor in controlling the steam purity if entrainment is not incorporated in the design.
Independent sanitary entrainment devices are available for installation at, or close to the point of use, andmay be used with typical “Simple CS generators” as well as to control additional moisture build up due to heatlosses in the distribution system of Sanitary CS Generators.
7.8.2 Steam Condensate Sampling
7.8.2.1 Purity Sampling
When required by the process, the steam purity shall be monitored through acceptable sampling techniques.A slipstream of the steam may be passed through a sample condenser/cooler, fitted with a sampling valve.
(See Section 7.2.1 for information on Steam Attributes.)
To ensure that the steam does not contribute to drug product contamination, sampling should be includedduring commissioning, as a good engineering practice, and/or prior to each time the steam is used.
If the sampling requirement is for endotoxin or pyrogen testing, the sample cooler, tubing and valve should beof sanitary construction.
Sample coolers can be fitted to the CS generator, or located in the distribution line, or at the use point
(recommended location), or a combination thereof. It is common practice to fit sample coolers with conductiv-ity monitors and alarms.
Endotoxin removal: The condensate sample from a Clean Steam generator with separator is expected toshow 3-4 log10 level reduction in pyrogens compared to the level in the feed water.
7.8.2.2 Steam “Quality” Sampling
Steam “Quality” sampling may be employed to determine the level of saturation and non-condensable gas-
ses. This can be determined by applying a steam calorimeter and measuring the dryness or saturation level.A steam calorimeter measures the percentage by weight of steam in a mixture of steam and entrained water.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 104/250
88
PHARMACEUTICAL STEAM
7.8.3 Materials of Construction
7.8.3.1 Materials of Construction for Sanitary and Simple CS Generators
Structural integrity and chemical compatibility with the contact fluid and its constituents are two of the morepractical issues that drive construction material selection for CS systems.
The inherent corrosion potential forces CS manufacturers to consider relatively inert metal including stainlesssteel or titanium etc. Sanitary piping and valves, considered unnecessary for utility and simple CS genera-
tors, are often standard features for CS systems based on the specific manufacturer and model. The materi-als chosen should not contribute to contamination of the drug product
Typical materials of construction for Sanitary and Simple CS Generators are:
Evaporator and separator:
Shell, tubesheets, and internals: 300 series S.S
Evaporator tubes: 300 series or titanium, or other suitable alloy
Heat exchangers 300 SERIES(FEED HEATER, BLOW DOWN &SAMPLE COOLER):
Piping: 300 series for water and Clean Steam, and carbon steel for utility
steam contact
Valves: 300 series and elastomers/diaphragms for water
Skid and structural: Carbon steel
7.8.3.2 Materials of Construction for Utility Steam Generation
Chemical compatibility with the Utility boiler generated steam and the carried over feed water chemicals are
required for all materials used to condition the contaminated steam.
Based on the particulate levels in the steam and the required steam purity, more than one filtration stage maybe utilized.
Distribution of Utility Steam following filtration follows similar practices as CS to control condensate build up,non-condensable gases and saturation levels as required for the application.
Acceptable materials must be relatively inert and may include SS or tin-coated copper.
7.8.4 Surface Finish
Mechanical polishing (MP), electropolishing (EP), and passivation processes are implemented in some stainlessCS systems. Chlorine and/or chlorides will damage the generator regardless of the finish.
The operating temperatures of these systems are more than sufficient for inhibiting microbiological growth.
Therefore, MP is advocated for final finishing of mechanical welds, with mill finishes and final passivation tooptimize the formation of the corrosion resistant chromium oxide barrier. Electropolishing will also optimizethe chromium oxide barrier, and should be considered if passivation is not an option.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 105/250
89
PHARMACEUTICAL STEAM
7.8.5 Pretreatment for CS (Sanitary and Simple) Generators
The feed water pretreatment for a CS generator is born from three separate and distinct considerations:
1) Scale formation.
2) Corrosion.
3) Volatiles which carryover with the steam and may affect steam purity.
7.8.5.1 Scale
Scale formation is a function of generator feed water chemistry, concentration (depends on blow down rate),
and temperature. It is independent of design and make, and is outside the control of the generator manufac-turer or the operator.
Because scale inhibitors are not used, and because of the relatively high operating temperatures of the CSgenerator, the total dissolved solids (TDS) of the feed water should be very low. Silica is of particular concern.
Most manufacturers stipulate a level of less than 1 ppm (parts per million); some go as high as 5 ppm. In
addition to having low TDS (Total Dissolved Solids), the feed water should have no measurable hardness. Itis therefore common to use DI or reverse osmosis as pretreatment to the CS generator. All CS generators willinvariably experience some form of scale build-up and therefore must include routine visual inspections, plus
cleaning of the generator during shutdown periods when appropriate.
Using compendial water as feed is wasteful, unless the steam quantity is small and the cost and maintenance
of a dedicated feed water pretreatment system exceeds the cost of using compendial water.
Some manufacturers offer generators to operate on softened water. Usually, the rate of blow down is in-creased in order to maintain low concentration.
Note: If the TDS of the soft water is relatively high, soft scale (such as sodium scale and sludge) can form.
7.8.5.2 Corrosion
The most common cause of corrosion is free chlorine, not chlorides.
Chlorine and chlorides, at any detectable level, are very detrimental to stainless steel. The higher the tem-perature and chlorine level, the more severe is the attack. Chlorine is known to migrate and concentrate inlocalized cells where the level can reach tens, or hundreds of ppm, while the concentration in the main stream
is a fraction of a ppm.
Chlorine can be removed from the feed water by chemical injection of a reducing agent such as sodiumbisulfate, or by passing the chlorinated water through carbon filters.
7.8.5.3 Volatiles
Dissolved gasses and substances that are volatile at the operating temperature of the CS generator willcarryover with the steam. If such substances are objectionable or may potentially compromise product qual-
ity, they must be removed at the pretreatment stage. Ammonia and CO2 (carbon dioxide) are examples ofvolatile gases that will have an effect on the conductivity, such that a condensate sample may not meet USP
requirements for Purified or WFI Water.
For more details on pretreatment and the advantages and disadvantages of the different processes, refer to
Chapters 4 and 5 of this Guide.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 106/250
90
PHARMACEUTICAL STEAM
7.8.6 Treatment of Utility steam
When utility steam is considered, it may be necessary to filter/condition the steam. In certain applications, itmay also be necessary to change the steam boiler treatment and substitute additives that do not contain
amines or hydrazine.
Since the type and degree of conditioning are dependent on the application, as well as on the quality of theutility steam and additives present, this Guide cannot address all possible scenarios.
Prior to the elimination of amines and hydrazines, by the substitution for standard boiler pretreatment addi-tives, the Utility Steam boiler manufacturer should be consulted regarding the impact on equipment warranty,
performance and expected life. Some of the substitute additives are not as effective as the standard.
7.9 COST IMPLICATIONS
Determining the economics of pharmaceutical steam production is complex. Costs are quite predictable, butvary greatly depending on scale of operation, system design, actual usage, etc. The total operating cost to
produce pharmaceutical steams is obtained by adding the cost of feed water to the costs of pretreatment and
final treatment (primary ion removal and polishing). The type of pharmaceutical steam system design optionselected is typically based on feed water TDS, silica and hardness levels, organic and colloidal content, aswell as anticipated steam system utility costs (acid, caustic, salt, power, and source water). Consideration
should also be given to maintenance requirements and available resources.
7.10 STEAM “QUALITY”
Steam quality is defined as the saturation percentage of steam to water or more explicitly, the ratio of thevapor mass to the mass of the steam mixture.
Dry Saturated Steam with minimum superheat is necessary for efficient steam sterilization.
Water can be generated and carried by steam within distribution systems in two ways:
1) In suspension as moisture when the steam is not 100% saturated
2) As condensate separated from the steam
Water vapor carried in suspension may be reduced by: adding more heat or raising the temperature, reducing
the pressure, or adding a steam entrainment separator. Water moisture and condensate may be reduced bysteam traps.
7.11 DISTRIBUTION
Distribution systems for clean steam follow the same good engineering practices commonly used for utility
steam, with the exception that contact materials must be inert to the aggressive nature of clean steam.Corrosion-resistant 304, 316, or 316L grade stainless steel “tubing” or solid-drawn “pipe” are commonly used.
Surface finish is not critical due to the self-sanitizing nature of the clean steam. Mill finish or 180-grit mechani-cally polished pipe or tubing is sufficient. TIG Orbital welding and post-installation passivation is considered
appropriate for this application. Piping must be designed to allow for thermal expansion and to drain conden-sate.
Note: Drains should have air breaks.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 107/250
91
PHARMACEUTICAL STEAM
Sanitary clamps or pipe flanges are most commonly used where the piping must be broken, but weldedconnections are used as much as possible to eliminate maintenance costs and potential for leaks. Threaded
connections may be suitable for instrumentation if positioned to drain condensate and remain hot. Ball valvesare commonly used for isolation because elastomeric diaphragms do not hold up well in this service. Where
diaphragm valves are used, Teflon-encapsulated EPDM diaphragms give the best long-term performance.
Steam quality sampling may be determined during “commissioning” and consistency ensured based on theproper location and subsequent maintenance of traps, entrainment separators, and vents. (The subject ofmaintenance cannot be over emphasized when these devises are involved due to the small orifices required
in the separation of gas and liquid.)
7.11.1 Line Sizing
The steam distribution header should be sized for a maximum velocity of 7,200 feet per minute (120 ft/sec or37 m/sec) to limit erosion and extend the life expectancy of the piping. Condensate line sizing should followgood engineering practices for utility condensate.
7.11.2 Water Moisture Removal
Water vapor forms in steam systems due to heat loss, causing a change in the liquid/vapor ratio or steam“quality”.
Steam may be dried of moisture by reducing the generated pressure just prior to the point of use to coincidewith the steam temperature of saturation at the reduced pressure.
Moisture entrained in the steam can also be removed by installing an In-Line separator at the point of use, justprior to, or just after, the regulator. If the separator is located upstream of the regulator, the regulator should
be protected from water damage (wire drawing) and impingement damage on the regulator diaphragms.
In-line separators are available in sizes from 1/2" to 6" (approx. 1 cm to 15 cm) and remove moisture with aseries of baffles on which the suspended water droplets impinge and fall out by gravity to the drain, whichmust be piped to a trap. Separators have a separation efficiency of better than 99% in the removal of all liquid
and solid entrainment exceeding 10 microns.
7.11.3 Condensate Removal
Condensate is the water that has separated from the liquid vapor mixture and forms in steam systems due toheat losses and natural separation effects. Lines should be designed to prevent the buildup of condensate toavoid dangerous water hammer and to eliminate potential cold spots where bacteria can grow. Any untrapped
vertical length of pipe will quickly fill with condensate. If this condensate is allowed to stand for sufficient time,it can cool and become a breeding ground for bacteria. This bacteria could possibly be entrained back into the
main distribution header and contaminate use points downstream. Worst case condensate removal locationsshould be sampled monthly for presence of bacteria. The following practices are commonly employed to
minimize these concerns:
• Each line is adequately supported to avoid sagging and subsequent condensate accumulation.
• Steam traps are installed at all points where condensate can collect (e.g., at least every 100 feet (30
meters) of line, upstream of control valves, at the bottom of vertical risers, etc.). Steam traps used forclean steam service should be sanitary design and self-draining.
• If the main distribution header is above the use points, the branches to the users should be routed fromthe top of the header to avoid excessive condensate loads at the branch. Each branch should be trapped
to avoid condensate buildup.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 108/250
92
PHARMACEUTICAL STEAM
• An alternative is to run the main distribution header below the use points. Then the branches can drainback to the main distribution header, avoiding the need for additional traps.
• The requirement to trap each branch can be waived for short drops from main headers to vessels or
other equipment that are in frequent use where the sterilization and water hammer is not impacted by thecollected condensate. An example is a drop from a main distribution header to a media storage tank,
which is sterilized daily. The condensate built up in the vertical drop line has only a limited time to cool andis quickly eliminated by the trap at the bottom of the vessel when the block valve is opened. The verticaldrop is sterilized daily with the vessel, so there is little chance for bacteria to grow.
7.11.4 Non-Condensable Gas Removal
Air and other non-condensable gases should be minimized from pharmaceutical steam systems. Since air
acts as an insulator, incomplete sterilization can occur in the process. Air in a system offers a very effectivebarrier to the heat transfer which will lead to a reduced temperature at the surface of a tube, system compo-nent or process equipment.
Air can be discharged using steam traps, however excessive levels may slow down the discharge of conden-
sate. The subcooled condensate can then lead to insufficient sterilization temperatures due to the excess
water.
The removal of air can be achieved by placing thermostatic pharmaceutical steam traps with the inlet in the
upward position. These should be placed in positions where air is prone to collect such as the terminal pointsof the main and large branches of the steam header, high points in the tanks, reactors and sterilizers, etc. Inthe case of air and condensate discharge at the bottom of large vessels, the air and condensate should be
separated by correct piping practices.
7.11.5 Superheat
While higher-pressure steam can be used to compensate for superheat, the latent heat, or killing power of thepotentially superheated steam is reduced at higher pressures; leading to increased sterilization cycles.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 109/250
93
PHARMACEUTICAL STEAM
Figure 7-4 Vessel Sterilization
7.12 FOUR EXAMPLES OF CORRECT PIPING PRACTICE
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 110/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 111/250
STORAGE and
DISTRIBUTION SYSTEMS
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 112/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 113/250
95
STORAGE AND DISTRIBUTION SYSTEMS
8. STORAGE AND DISTRIBUTION SYSTEMS
8.1 INTRODUCTION
This chapter provides an overview of eight common distribution configurations and a decision tree to help
decide which system best suits the operating requirements. A comparison of tank versus tankless systems isaddressed, as well as alternative materials of construction available, and ancillary equipment related tooverall distribution systems. Common industry practices are listed as examples, to help clarify regulatory
requirements.
8.2 SYSTEM DESIGN
8.2.1 General Considerations
A storage system is used to accommodate peak flow requirements against usage rates. The storage systemmust maintain the feed water quality to ensure the appropriate quality of the end use of product. Storage
allows a smaller, less costly pretreatment system to meet peak demand. A smaller treatment system oper-
ates closer to the ideal of continuous, dynamic flow. Large manufacturing sites, or systems serving differentbuildings, may use storage tanks to separate one section of the loop, and others to minimize cross contami-nation.
The main disadvantage of a storage tank is its capital cost, and the cost of associated pumps, vent filters, andinstrumentation. However, this is usually less than the increased cost of pretreatment equipment sized to
handle the peak use rate in the facility.
Another disadvantage of storage is that it introduces a region of slow moving water, which can promotebacterial growth.
8.2.2 Capacity
Criteria affecting storage capacity include the user’s demand profile or the amount of use, duration, timing,and diversity, (in the case of more than one user), balance between the supply of pre- and final- treated
waters, and whether the system is recirculating or non-recirculating. Careful consideration of these criteriawill affect cost and water quality.
The storage tank must provide reserve to minimize cycling of the treatment equipment and to reduce pumpcavitation. It should provide sufficient reserve to enable routine maintenance and orderly system shutdown in
the event of an emergency, which can vary from few to many hours, depending on the size and configurationof the system and maintenance procedures.
8.2.3 Storage Tank Location
It may not be cost-effective to locate storage tanks as close as possible to the point of use, within high-cost,GMP-finished areas. It may be more advantageous to locate them close to the generation equipment, for
ease of maintenance. Utility areas are acceptable for this purpose, if access is provided (and the area is keptclean).
8.2.4 Types of Storage Tanks
Vertical storage tanks are common but horizontal tanks may be necessary if overhead space is limited. Forrecirculating systems, tank design should include an internal spray ball to ensure that all interior surfaces are
wetted for microbial control. Jacketing is usually provided in hot systems, to maintain water temperature over
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 114/250
96
STORAGE AND DISTRIBUTION SYSTEMS
long periods without makeup, or to temper high influent temperatures, to preclude excessive rouging andpump cavitation. To avoid the absorption of carbon dioxide and its effect on conductivity, inert blanketing of
the tank headspace should be considered. Storage tanks must be fitted with a sub-micron hydrophobic ventfilter to reduce bio-burden and particles.
The maximum size of a single storage vessel is often limited by the space available in the facility. It may be
necessary to resort to multiple tanks to obtain the desired capacity. In this case, interconnecting piping mustbe carefully designed to assure adequate flow through all supply and return branches.
8.3 SYSTEM DISTRIBUTION DESIGN
8.3.1 General Considerations
Proper design of both the water storage and distribution systems is critical to the success of a pharmaceuti-cal water system.
The optimal design of any water storage and distribution system must accomplish three things:
1. Maintain the quality of the water within acceptable limits.
2. Deliver the water to the use points at the required flow rate and temperature.
3. Minimize capital and operating expenses.
Although items 2 and 3 are well understood, item 1 is often misinterpreted. It is not necessary to protect thewater from every form of degradation, only to maintain the quality within acceptable limits. For instance, water
stored in the presence of air absorbs CO2, increasing the conductivity. This degradation can be avoided byblanketing the storage vessel with nitrogen. However, for many systems this would be a wasteful expenditure
if the increased conductivity were still within the required specification.
As technology has improved over the years, many design features such as storage at elevated temperature,
constant circulation, use of sanitary connections, polished tubing, orbital welding, frequent sanitization, andthe use of diaphragm valves have become common place. To incorporate all of these features into each new
design typically leads to ever escalating costs with little if any reduction in risk of contamination. Althougheach of these items provides a level of security, it is a mistake to assume that all of them need to be in every
system. Many systems operate successfully with one or more of these design features omitted. In suchcases, the cumulative effect of the other design features is adequate to prevent degradation of the water.
A more reasonable approach is to utilize design features that provide the greatest reduction in contaminationrisk at the most reasonable cost, and add the more expensive features in the design phase, only if they are
required to maintain quality within acceptable limits. The systems should be designed to be robust, so fea-tures do not have to be added later, affecting cost and schedule. The idea of selecting design features based
on “return” on investment where “return” is defined as reduction in contamination risk, can be very helpful incontrolling system cost and in evaluating different alternatives. Ultimately, the effectiveness of each systemdesign is determined by the quality of the water delivered to the users. The challenge for the design engineer
is to know what features to include, to achieve the required degree of protection with the lowest lifecycle cost.
EXAMPLE
A USP compendial water system is designed with a 316L SS storage and distribution system and operatesnormally at 80°C. The tubing is all sanitary, orbital welded, with minimal clamps and zero dead leg diaphragmvalves at the use points. Water is kept circulating through the tubing at a minimum return velocity of 3ft/sec. In
this case, use of high mechanical polish tubing (<20 Ra) with electropolishing may not be required. The risk
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 115/250
97
STORAGE AND DISTRIBUTION SYSTEMS
of contamination for such a system is already low, and the impact of this upgraded surface finish is question-able. The benefits that will be achieved by further improving the quality of finish may not be justified.
However, if the same system were open to the atmosphere, consideration would be given to installing a
0.2micron vent filter on the storage vessel, as the reduction in contamination risk is quite large for a relativelysmall investment. Similarly, if the zero dead leg valves were replaced with less expensive valves with larger
dead legs, you might consider increasing the minimal circulation velocity to help compensate.
The purpose of the following chapter sections is to provide information to help the user evaluate the advan-
tages, disadvantages and cost effectiveness of many of the design features commonly used to protect waterfrom degradation. A method of selecting/optimizing system storage and distribution design is also presented.
As a general rule, a water system is optimized as a result of the following:
1) Minimizing the time the water is held at conditions which favor growth
2) Minimizing changes to water temperature
3) Contacting all areas during sanitization
One system design can be said to be better than another, if it accomplishes these goals to the same degree,but at a reduced lifecycle cost. Examples of storage and distribution concepts commonly used today arepresented in subsequent sections of this Guide, to help demonstrate the idea of optimal system design.
8.3.2 Distribution Design Concepts
The two basic concepts developed for distribution of pharmaceutical waters are referred to as the “batch” and“dynamic/continuous” distribution concepts.
The batch concept utilizes at least two storage tanks. While one is being filled, the other is in service providing
pharmaceutical waters to the various process users. After one tank has been filled from the water finaltreatment system, it is isolated and the water inside is tested. Only after testing is that tank put into service.The water is often drained after 24 hours, but can be validated for longer periods of time. At the completion of
the draining operation, the vessel and distribution system is usually sanitized before refilling.
The dynamic/continuous concept off-sets the peak instantaneous water demand, put on the overall watersystem through utilization of a single water storage vessel which simultaneously receives final pretreatment
system make-up, stores the water in the vessel, and ultimately supplies it to the various process users whilemaintaining water quality.
The advantage of the “batch” distribution concept, over the “dynamic/continuous” distribution concept, is thatthe water is tested before use with tank QA/QC lot release (water used in each product batch lot is traced and
is identifiable). The advantages of the “dynamic/continuous” distribution concept include lower lifecycle costs,as well as less complex piping around the storage vessel, and a much more efficient operation.
Once a system distribution concept has been selected, the following additional storage and distribution de-sign considerations should be carefully evaluated:
• System configuration including whether series or parallel loops are required, distribution loop points of
use, cooling requirements (steam-able, sub-loop or multiple branched heat exchanger assemblies), re-heat requirements, if any, secondary loop tanks versus tankless system considerations, etc.
• Hot (65-80°C), cold (4-10°C), or ambient temperature process use point requirements
• Sanitization method (steam, hot water, ozone, or chemical)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 116/250
98
STORAGE AND DISTRIBUTION SYSTEMS
8.3.3 Distribution Decision Tree
The decision tree in Figure 8-1 is presented to aid in the analysis of distribution design. Most of the systemsin use today are represented by one of these eight configurations, but other designs may also be acceptable.
In evaluating which configuration is optimal for a given situation, the designer needs to consider many factors,including the requirements for Quality Assurance release, the desired specification of water (DI, USP WFI,
etc.), hydraulic limitations, the required temperature at each drop, the number of use points, and the cost ofenergy.
Decision tree guide
1) Batched System
2) Branched/One Way
3) Parallel Loops, Single Tank
4) Hot Storage, Hot Distribution
5) Ambient Storage, Ambient Distribution
6) Hot Storage, Cool and Reheat
7) Hot Tank, Self-contained Distribution
8) Point of Use Heat Exchanger
Each configuration varies in the degree of microbial control provided and in the amount of energy required.Better microbial control is usually achieved by minimizing the amount of time water is exposed to conditions
favoring microbial growth. Configurations that store water at sanitizing conditions such as hot, under ozone,or circulation at turbulent velocities, are expected to provide better microbial control than those that do not.Naturally, hot circulating systems are more forgiving than cold systems from a microbiological perspective.
However, adequate microbial control may be achieved in other configurations provided they are frequentlyflushed or sanitized. In any case, system design should prevent stagnation, which promotes formation of
biofilm.
Energy usage is minimized by limiting the amount of water changing temperature. Configurations storingwater hot but supplying it to the use points at lower temperature must cool the water before use. Energyrequirements are minimized by cooling only that water drawn from the system. Configurations that constantly
cool and reheat water utilize more energy than systems that do not.
The configurations delivering lower temperature water are shown with a single cooling exchanger for clarity.Usually the cooling medium is tower water since this is the least expensive to generate. In most parts of the
world, tower water is not cold enough to allow use temperatures much below 25°C. A second cooling ex-changer using chilled water or glycol must be added if the required use temperature is below 25°C. It isusually cost prohibitive to cool water from 80°C to less than 25°C using chilled water or glycol alone as the
chiller size becomes quite large.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 117/250
99
STORAGE AND DISTRIBUTION SYSTEMS
8.3.4 Example System Descriptions
The following describes the systems, contained in the accompanying decision tree, that can be used suc-
cessfully to store and distribute high purity water. Figure 8-2 through Figure 8-12 present simplified sche-matic diagrams (not meant to be P&IDs) of each configuration.
Figure 8-1 Distribution Decision Tree
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 118/250
100
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-2 Batched Tank Recirculating System
This system is used where QA release is required on the water before it goes into the process. One batchtank supplies water to the process, while the other is filled and tested for QA release (traditionally due to
unreliable means of water production). This is a cumbersome system to operate and is usually limited tosmaller systems. The disadvantages are the high capital and operating costs. In-line conductivity and TOC
measurements can provide nearly the same degree of assurance for less money.
Figure 8-2 Batched Tank Recirculating System
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 119/250
101
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-3 Branched/One Way with Limited Points of Use
This configuration is sometimes used where capital is tight, the system is small, and microbiological quality isof lesser concern. It is also useful where frequent flushing or sanitization of the piping is possible. It is a good
application where water use is continuous. It is less advantageous where water use is sporadic, as the linestays stagnant when not in use. Microbial control is more difficult to maintain. A program must be set up to
flush (e.g., daily) and sanitize the loop to maintain microbial contamination within acceptable limits. Morefrequent sanitization may be required, increasing operating costs. It is also more difficult to use On-Linemonitoring, as indicative of the quality of the water throughout the system, in a non-recirculating system.
Figure 8-3 Branched/One Way with Limited Points of Use
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 120/250
102
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-4 Parallel Loops, Single Tank
This configuration is a combination of any of the loop distribution schemes off one storage tank. Figure 8-4shows a hot storage tank with two separate loops; a hot distribution and a cool and reheat loop. Parallel loops
are very common and are most advantageous where multiple temperatures are required, or where the areaserved is so large that a single loop becomes cost prohibitive or hydraulically impractical. The major concern
is to balance the various loops to maintain proper pressure and flow. This is accomplished either by usingpressure control valves, or by providing a separate pump for each loop. (Note: A different design is intention-ally presented for each loop).
Figure 8-4 Parallel Loops, Single Tank
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 121/250
103
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-5 Hot Storage, Hot Distribution
This is the configuration of choice when all use points require hot (greater than 65°C) water. Temperature ismaintained in the storage tank by steam supplied to the tank jacket or alternatively by a heat exchanger on
the circulating loop. Water is generally returned to the top of the tank through a spray ball to ensure that theentire top surface is wetted. This system provides excellent microbial control and is simple to operate. In
addition, tank and loop sanitization is required less frequently, or not at all, if a temperature of 80°C is main-tained. This type of system is universally accepted by regulatory agencies.
Areas of concern include protecting workers from scalding, cavitation in the circulation pump, moisture con-densation on the vent filter, and the formation of rouge. Scalding is minimized by operating at lower tempera-
ture (65°C) or by proper training and personal protective gear. Cavitation is avoided by accounting for the highvapor pressure of hot water in the net positive suction head (NPSH) calculations. Condensation is prevented
by positioning the hydrophobic vent filter for good drainage and by heating the filter with either a low pressuresteam jacket or electric tracing. Avoid overheating as this can melt the filter cartridge. Rouge formation iscontrolled by passivation and by operating at a lower temperature. It can be eliminated by using non-metallic
or lined components.
Figure 8-5 Hot Storage, Hot Distribution
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 122/250
104
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-6 and Figure 8-7 Ambient Storage, Ambient Distribution
This system is most advantageous when the water is generated at ambient temperature, will be used only atambient temperature, and there is adequate time for sanitization.
Since the water is stored at ambient temperature with no disinfectant, microbial control is not as good as in
hot storage system configurations. However, good microbial control is possible provided sanitization is con-ducted on a frequent basis. Frequent sanitization is usually accomplished by allowing the water level in thestorage tank to drop through use, then heating the remaining contents, and circulating through the loop for a
set amount of time. Reducing the water level limits the energy and time required to sanitize. Heat is providedby steam supplied to the tank jacket, or alternatively, by a heat exchanger on the circulating loop. Cooling may
be required to prevent temperature increases due to heat buildup from the pump, and for cool down aftersanitization. Water consumption is low if the level in the storage tank is allowed to drop through use prior to
sanitization and moderate if it is drained.
The capital and operating costs of this system are minimal. Another advantage is that it can provide high flow
rates of ambient pharmaceutical water, without need for complex points of use heat exchangers. Its majordisadvantage is the time required to sanitize, which is longer than the previously described systems, due to
the need to heat and cool the contents of the storage tank.
Figure 8-6 Ambient Storage and Ambient Distribution
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 123/250
105
STORAGE AND DISTRIBUTION SYSTEMS
Many pharmaceutical water users have found that storing and distributing water at ambient temperatureswith periodic sanitization, (utilizing either clean steam or heating to 80°C for microbial control) to be safe and
cost effective. The ambient system can also be effectively operated with an ozonated storage and a periodi-cally ozonated loop, in lieu of hot water sanitization (see Figure 8-7). Levels of 0.02ppm to 0.2ppm of ozone
protect the water from microbial recontamination. Ozone needs to be completely removed from processwater prior to usage, using UV radiation. Consideration therefore must be given to verifying/assuring that
ozone has been eliminated, such as the use of In-Line monitors.
One advantage of ozonation or chemical sanitization, is that these methods allow the use of plastics as a
material of construction (popular in Europe for purified water systems).
Figure 8-7 Ozonated Storage and Distribution
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 124/250
106
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-8 Hot Storage, Cool and Reheat
This system is most advantageous when the water is generated hot, tight microbial control is required, andthere is little time for sanitization. It provides excellent microbial control and is easily sanitized. It requires less
capital than point of use exchangers, if there are multiple low temperature use points. Hot water from thestorage tank is cooled through the first heat exchanger, circulated to the use points, and then reheated in a
second exchanger before returning to the storage tank. Sanitation of the loop is accomplished by turning offthe cooling medium on a periodic basis. Water consumption is minimized since no flushing is required. Themajor disadvantage of this configuration is it’s high energy consumption, since it cools and reheats the circu-
lating water regardless of whether it is drawn out of the loop.
Figure 8-8 Hot Storage, Cool and Reheat
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 125/250
107
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-9 Hot Storage, Self-Contained Distribution
This configuration is most advantageous when water is generated hot, there are many low temperature waterusers, and energy consumption is critical. It provides the benefits of the cool and reheat loop without the large
energy requirement. Hot water from the storage tank is cooled through the heat exchanger, circulated to theuse points, and then returned to the pump suction bypassing the storage tank. The loop is sanitized on a
periodic basis by turning off the cooling medium and opening up the return to the storage tank, allowing hotwater to flow through the loop. An alternative is to flush the lower temperature water to drain until the loopbecomes hot and then return the flow to the storage tank. The water in the storage tank is kept hot through a
steam jacket or heat exchanger on an external pump around loop.
When water is drawn out of a point of use valve, hot water from the storage tank flows into the loop and iscooled by the heat exchanger. The hot water flushes the short section of line between the storage tank and
the circulation pump preventing a deadleg. In most pharmaceutical installations, this happens many times perday so the line stays relatively hot. If the usage rate is low, a small amount of water can be returned to thestorage tank on a continuous or timed basis, keeping this line flushed. A third alternative is to return the
circulating water to just downstream of the storage tank outlet valve, so the deadleg is negligible.
Figure 8-9 Hot Storage, Self-Contained Distribution
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 126/250
108
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-10, Figure 8-11, and Figure 8-12 Hot Storage, Hot Distribution, Point of Use Heat Exchanger
This configuration is identical to Figure 8-5 except that use points requiring water at lower temperature areequipped with point of use heat exchangers. Figure 8-10, Figure 8-11, and Figure 8-12 show three different
designs for these exchangers. All three allow flushing water to drain to lower microbial counts and adjustingtemperature before opening up the point of use valve. All three also allow for sanitizing the exchanger and
downstream piping when water is not called for at the drop. The schemes differ in capital cost, sanitationmethod, and in the amount of water used for flushing. Sanitization is accomplished using low pyrogen steamin Figure 8-10. In Figure 8-11 sanitization is accomplished by circulating hot water from the loop, through the
point of use exchanger, back to the main loop. The operation in Figure 8-11 can be facilitated by installing ablock valve at the return of the main loop. The valve would be closed immediately prior to starting the sub
loop, to prevent back flow from the main loop. The initial draw of point of use water would be diverted to drain.Figure 8-12 is sanitized by flushing hot water from the main loop once through to drain. Tube-in-tube or
serpentine type coolers could be used, as well as double tube sheet exchangers, which are depicted.
Point of use exchangers are most advantageous when there are both hot and lower temperature water use
points off the same loop, and the number of low temperature users is small. Since they maintain the water hotuntil it is drawn from the loop, they provide excellent microbial control, provided they are frequently flushed or
sanitized when not in use. As the number of low temperature users increases, the capital costs and space
requirements become prohibitive, and one of the other configurations should be considered. Water consump-tion is high due to flushing, although this is minimized by the scheme shown in Figure 8-11. Energy consump-tion is moderate because only water drawn out of the loop is cooled but additional energy must be spent to
make up water flushed to drain. Maintenance requirements are high due to the added exchangers and valves.Complexity is high as each exchanger must be properly flushed and sanitized. Each drop is limited in capac-ity by the sizing of the exchanger. The scheme shown in Figure 8-11 results in added pressure drop in the
main loop, which leads to a larger circulation pump.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 127/250
109
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-10 Single Point of Use, Steamed
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 128/250
110
STORAGE AND DISTRIBUTION SYSTEMS
Figure 8-11 Point of Use Installed in Sub-loop
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 129/250
111
STORAGE AND DISTRIBUTION SYSTEMS
8.3.5 Storage and Distribution Comparison Table
Table 8-1 compares several storage and distribution options currently used in the pharmaceutical industry.Comparisons are made based upon capital, energy, operating costs, maintenance, validatability, and other
factors. Each category is rated low (L), medium (M), or high (H) for each system relative to the other systemspresented. The particular storage and distribution choice for a given scenario will depend upon the specific
situation being addressed, and the priority the end user gives to each of the categories, with quality being theforemost priority.
Figure 8-12 Point of Use Heat Exchanger with Multiple Branch Users
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 130/250
112
STORAGE AND DISTRIBUTION SYSTEMS
Table 8-1 Comparison of Storage and Distribution Options
Legend: L = Low M = Medium H = High Amb = Cold or Ambient Conditions
Notes:
1) Lower with hot water sanitization once every 24 hours.
2) Storage tank is always hot, loop is cold or ambient, and hot water sanitized once every 24 hours. Loop return is heated prior to reentering storage
tank.
3) Frequent hot water flushing or steaming can effectively control bioburden. High turnover of the water in each branch due to use (at least once daily)
can significantly reduce bioburden.
Category
Capital Cost
Water
Consumption
Energy
Consumption
Validatability
Operability
Maintenance
Requirements
Tank Turnover
Line Flushing
Requirements
Ability to
Respond to
Large Peak
Demands
Loop Balancing
and Control
Requirements
Microbial/
EndotoxinGrowth Potential
Most
Advantageous
When:
Least
AdvantageousWhen:
1. Batched
System
H
H
L
Simple
Complex
M
Non-Critical
Critical
Limited by
QA hold
Average
L-M
Method of
Generation is
Not Reliable,
QA release
required
before water
use. Small
system is
required.
Capital and
operatingcost is a
concern
2. Branched /
One Way
L
H
L
Complex
Complex
L
Limited
Critical
Excellent
Simple
H (3)
Tight
Capital,
Continuous
Use,
Frequent
Flushing or
Sanitization.
Sporadic
Demand UseProfile, or
operating cost
a concern
3. Parallel
Loops,
Single Tank
M
M
Depends on
loops
Complex
Depends on
loops
Depends on
loops
Average for
ambient tank,non-critical
for hot tank
Depends on
loops
Average to
Excellent
Critical
Hot = L
Amb = M
Multiple
Temperatures
Required or
Hydraulic
Limitation
Hydraulic
balancing isdifficult
4. Hot
Storage, Hot
Distribution
L
L
L
Simple
Simple
M
Non-Critical
Non-Critical
Excellent
Simple
L
Hot Water is
Required,
Water is
Generated
Hot, or
Microbial
Control is
Critical
Initial Capital
or EnergyAvailability is
Tight
5. Ambient
Storage,
Ambient
Distribution
L - M
L - M
L
Average
Average
L
Average
Average
Excellent
(Cold Surge
Volume)
Average
M (1)
High Peak
Demands for
Ambient or
Cold Water,
Water is
Generated at
Ambient
Temp
Sanitization
will not fit intooperating
schedule
6. Hot
Storage,
Cool and
Reheat
M
L
H
Simple
Average
M
Average
Non-Critical
Average
Average
L - M (2)
Tight
Microbial
Control,
Limited Time
for
Sanitization,
Energy cost
not a
concern,
many low
temperature
users
Per Unit
Energy Costis High
7. Hot Tank,
Self-
contained
Distribution
M
M
M
Average
Average
M
Limited
Average
Average-
Excellent
Average
L - M (2)
High Peak
Demands for
Ambient to
Cold Water
and Unit
Energy
Costs a
concern,
many low
temperature
users
Per Unit
Energy Costis High, or
tank turnover
is a concern
8. Point of
Use Heat
Exchanger
H
H
M
Average
Average-
Complex
H
Non-Critical
Critical
Limited by
exchanger
sizing
Critical
M
Both hot
and warm
temperature
water
required and
number
of low
temperature
users is low
Space, Initial
Capital orEnergy
Availability is
Tight
9. Tankless,
Ambient
Loop
L
M
L
Average
Average
M
Not
Applicable
Average
Average
Average
Hot = L
Amb = M
Space
Constraints
or Tank
Turnover a
concern,
Limited
Capital
High
demands forAmbient or
Cold Water
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 131/250
113
STORAGE AND DISTRIBUTION SYSTEMS
8.4 MATERIALS OF CONSTRUCTION
Pharmaceutical equipment and piping systems rely extensively on stainless steels to provide the non-reac-tive, corrosion-resistant construction needed in manufacturing and heat sterilization. However, thermoplas-
tics are available that may offer improved qualities, or lower cost. Less expensive plastics such as polypropy-lene (PP) and polyvinyl chloride (PVC) may be acceptable for non-compendial systems. Others, such as
polyvinylidene fluoride (PVDF) offering greater heat resistance, may be suitable for compendial waters, al-though they require continuous support in hot applications. The cost of a PVDF system may be approximately10-15 percent lower than the cost of a stainless steel system once factors such as passivation, boroscope,
radiographic inspection, etc., are included. New methods of joining PVDF tubing leave a weld much smootherthan possible with stainless steel. At higher temperatures, however, thermal expansion of the plastic be-
comes a major concern.
While certain changes to higher grade materials (higher alloys such as AL6N and Hastelloy) and methods offabrication to assure compliance can yield minor improvements, others may only provide minor gain despiteconsiderable additional expense.
Material selection should be consistent (all 316L or all 304L etc.) throughout the distribution, storage, and
processing systems, if regular passivation is planned.
For compendial water, the use of 316L stainless steel is preferred.
Insulation for stainless piping should be free of chlorides, and hangers provided with isolators to precludegalvanic corrosion.
304L and 316L stainless steel has been the industry preference in tanks for the storage of compendialwaters. Jacket material in contact with the shell should be compatible, to avoid chromium depletion in the
weld-affected zones. Non-compendial water storage may not require the same level of corrosion resistanceor the use of low carbon nickel chromium alloys and special finishes, depending on the owner’s water speci-
fications.
High purity water distribution systems, using the material and finishes specified by the design, should be
joined using acceptable welding or other sanitary techniques. The distribution and storage systems should beinstalled in accordance with cGMPs and fabricated, manufactured, procured, and installed in strict accor-
dance with explicit operating procedures.
Orbital welding has become the preferred method for joining high purity metallic water piping systems, due tothe greater control over critical weld parameters and the smooth weld bead characteristics of the process.However, hand welding is still employed and may be required in certain situations.
304 and 316 stainless steel have been preferred grades for use in metallic piping systems due to their high
chromium and nickel content and ease of automatic welding. Low carbon and low sulfur grades of stainlesssteel are preferred for compendial systems, and control and inspection of the welding process is necessary
to limit corrosion and crevices in the system. A maximum sulfur content of 0.04% would be ideal for weldingbut any mismatch in the sulfur content of the mating par ts will easily cause the weld to weaken, outweighingthe advantages lower sulfur levels.
Where possible, all fittings, valves, tubing, and weldable pieces of the same nominal size (diameter) should
be purchased and manufactured from steel with the same specification and Heat Number in order to stan-dardize the weld quality for each tubing size.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 132/250
114
STORAGE AND DISTRIBUTION SYSTEMS
8.4.1 Comparison of Materials of Construction for Tanks and Distribution Systems
8.4.2 Workmanship
Fabrication should be performed by certified welders in a controlled environment to preclude contaminationof equipment and material surfaces. Facilities dedicated to the fabrication of stainless steel (or higher grade
alloys) are preferred, to avoid contamination by carbon steel. Fabrication must follow an approved qualityassurance plan. There needs to be adequate documentation in the design and construction of the system,
including up to date P&IDs, system isometrics, weld test reports, etc.
Tubing and piping welds, whether orbital or manual, must have a smooth internal diameter contour without
excessive concavity or convexity, bead wandering, misalignment, porosity, or discoloration. One hundredpercent photographic or radiographic analysis, while utilized to an increasing extent, is neither cost effective
nor infallible. Appropriate sampling is strongly recommended.
Table 8-2 Comparison of the relative values of key factors in the design and installationof water systems
Legend: Y = Yes N = No H = High M = Medium L = LowNotes:
1) Based on skilled labor requirements, ease of welding, ease of visual inspection, shop fabrication requirements, etc.
2) The steam pressure and steam temperature control is critical to keep both below the manufacturer’s ratings.
PVDF ABS POLYPRO PVC 316LSS 304LSS 316LSS 304LSS
TUBING TUBING PIPING PIPING
Installed Cost M M L L M M M M
Ease of Installation (1) H M M H M M H H
Steam Sanitizable Y N N N Y Y Y Y
Hot Water Sanitizable Y N N N Y Y Y Y
Ozone Sanitizable Y N N N Y Y Y Y
Chemical Sanitizable Y Y Y Y Y Y Y Y
Rouging Susceptibility N N N N Y Y Y Y
Corrosion Resistance H H H H H M H M
Availability M L M H H M H H
Extractables L M L H L L L L
Degree of Thermal H H H N/A L L L L
EXPANSION
Joining Method
-TRICLAMP Y N Y N Y Y Y Y
-Solvent N Y N Y N N N N
-THERMAL FUSION Y N Y N N N N N
-WELD N N N N Y Y Y Y
External Support H H H M L L L L
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 133/250
115
STORAGE AND DISTRIBUTION SYSTEMS
8.5 SYSTEM COMPONENTS
8.5.1 Heat Exchangers
Shell and tube, tube-in-tube, and plate and frame heat exchangers are employed. Although plate and frameunits may offer a cost advantage, they are used less often in the distribution portion of the system in compendial
service because of the perceived greater risk of contamination. However, they are common in the pretreat-ment side prior to final purification. In a shell and tube exchanger treated water flows through the tube bundle;the risk of contamination from cooling or heating media can be significantly reduced by means of a double
tube sheet. Complete drainability of the u-tube bundle is achieved by weep holes located at the low point ofeach chamber in the exchanger channel. Ensuring a positive pressure differential on the “clean” side can
further reduce contamination risk. Similarly, a plate and frame unit should be operated with the cleaner waterside at higher pressure than the heating or cooling medium. Conductivity meters may be used for monitoring
leakage. Unit design should permit full drainage and ready access for inspection and cleaning.
8.5.2 Vent Filters
Used on storage tanks in compendial water service to reduce contamination during drawdown. Units are
constructed of hydrophobic PTFE or PVDF to prevent wetting and generally rated at 0.1 to 0.2 microns.
Filters should be capable of withstanding sterilization temperatures and sized for maximum fill or drawdownrates to effectively relieve the negative pressure created by high temperature sanitization cycles. Filters in hotsystems are usually jacketed to minimize condensate formation that could result in blinding vessel exhaust
hydrophobic filters. Storage tanks should be rated for full vacuum, (or have vacuum protection), if steam isused for sterilization. Installation should also allow for drainage of condensate caused by high operating orsanitizing temperatures, and ease of replacement. The filter cartridges need to be appropriate for the filter
housing. Vent filters should be integrity tested for compendial water storage tanks, but may not need to bevalidated as sterile filters.
8.5.3 Pumps and Mechanical Seals
Centrifugal pumps are commonly employed in distribution systems. Performance curves and suction headrequirements should be reviewed to preclude cavitation, which might lead to particulate contamination. The
generation of pump heat over extended periods of low or no draw off should also be considered, sincesignificant temperature rise in cold systems, or cavitation due to vapor pressure in hot systems could occur.
Casing drains allow for full system drainage, where the pumps are at the low point of the distribution. Althoughwith double mechanical seals, with WFI or other compatible seal, water flushing may minimize the possibility
of contamination; single mechanical seals flushed to the outside have also been used. In extremely criticalapplications, polished rotating elements may be warranted. The installation of dual pumps, for standby pur-poses, should ensure flow throughout the system.
8.5.4 Piping System Components
a) Piping and Tubing: Extruded seamless and/or longitudinally welded tubing is commonly used in sys-
tems two inches in diameter and smaller. Recently, welded steel tubing (ASTM A-270), similar to seam-less in appearance, has become available at significantly lower cost. PVDF has also been shown to be aviable alternative.
b) Fittings: Single fittings may be manufactured from as few as one, to as many as five pieces. This can
materially affect the suitability of the end product, in terms of weld content, documentation, and cost.
c) Valves: The trend in the industry has been to use diaphragm valves in high purity systems, particularly inisolation applications. For steam service, sanitary ball valves may be acceptable and require less main-tenance.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 134/250
116
STORAGE AND DISTRIBUTION SYSTEMS
The following is a summary of water system components, listing the common industry practice, and listingadvantages and disadvantages:
8.5.5 System Components Comparison Table
Table 8-3 System Components Comparison
Item
Valves
Gaskets
Vent Filters
Heat Exchangers
Pumps
Mechanical Seals
Industry Practice
Diaphragm
Plug/Ball
Butterfly
Elastomers, including Viton
Silicone
EPDM
Teflon
Teflon encapsulated
0.2 Micron Hydrophobic
Membrane
Steam jacketed or electric traced
housing
Double tube sheet
(Shell and tube)
Single tube sheet
(Shell and tube)
Concentric pipe
Plate and Frame
Centrifugal
Positive Displacement
Double
Single
Advantages
Drainable*
Sanitizable
Cleanable
No steam seal
No body pockets
Low cost
Tight shutoff
Low maintenance
Low cost
Tight shutoff
Low maintenance
Temperature resistant
Less expensive
Temperature resistant
Less expensive
Temperature resistant
Less expensive
Best temperature resistance
Inert
Good temperature resistance
Good chemical resistance
Bioburden and
particulate reduction
Sanitary design
Protection against plant to clean
side leaks
Less expensive than double tube
sheet
Low leak potential
Least Expensive
Commonly available
Commonly available
More efficient when higher
discharge pressure is required
Constantly being flushed
Higher on-stream reliability
Less expensive
Disadvantages
Higher initial cost and
maintenance
Wears out quicker
Not absolute shutoff for high
pressure systems
Need stem seal
Have body pockets where
bacteria may linger, making
sanitization difficult
Need stem seal
Have body pockets
Chemical resistant
Chemical resistant
Steaming not
recommend
Cold flow in service
More expensive
Expensive
Sensitive to pinching
Possible plugging due to wetting
Cost
More expensive
Need to maintain a higher Delta P
on clean side is operationally
difficult
Low heat transfer coefficient,
requiring a large surface area
Greatest leak potential
Needs double gasketing
Less expensive
Lower maintenance
More expensive
Higher Maintenance
More expensive, both installation
and operation
For non-shrouded impeller type,
cleanability an issue
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 135/250
117
STORAGE AND DISTRIBUTION SYSTEMS
8.6 COMPARISON OF WFI SYSTEMS WITH STORAGE TANK AND WITHOUT STORAGE TANK
Figure 8-13 Tankless Ambient Distribution
*If canted at the correct angle, and installed in pitched lines
Item
Connection Types
Tanks
Rupture Discs
Industry Practice
Sanitary clamped
Flanged
Jacketed (1/2 pipe)
Jacketed (full jacket)
Non-Jacketed
Advantages
Minimal crevicing
Ease of inspection
Ease of Disassembly
Easier in piped systems
Good in high pressure
applications
Recommended for >4” OD
Good thermal efficiency
Less welding leading to lower
probability of weld failure
Allows for complete external
inspection of the tank
Safety relief device prevents tank
rupture should vent filter become
blocked
Tank to be designed as anatmospheric tank rather than a
pressure vessel
Disadvantages
Pressure limitations
Size limitations
High cost
Gasket protrusion
Greater chance for crevicing
Significant welding required
Less thermal efficient
Requires an external heat
exchanger
Table 8-3 System Components Comparison
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 136/250
118
STORAGE AND DISTRIBUTION SYSTEMS
It is possible to feed sub-loops off a single main loop without the use of an intermediate storage tank (Figure8-13). This configuration is most advantageous when space or capital constraints are tight. The sub-loop is
generally a circulating loop. Water drawn out of the main loop when a point of use valve is opened cannotreturn to the main since the sub-loop is at a lower pressure. This provides a degree of isolation between the
sub-loop and the main, or other, sub-loops. The major disadvantage is that there is no storage capacity.Usually this capacity is provided by a storage tank on the main loop.
Table 8-4 Comparison of WFI Systems: with Storage Tank and without Storage Tank
STORAGE TANK SYSTEM
Advantages
Provides air break to minimize back contaminationof hot WFI supply.
Minimize WFI cooler capacity by averaging hot
WFI feed flow into tank.
Provide cooled WFI “surge” volume ready ondemand to facilitate production schedules.
Positive point to relieve system pressure due tonormal venting or hot water sanitization.
Once operational, conditions easier to maintain
than tankless system. Potential problems easier toisolate.
Eliminate back pressure control valve cavitation bydividing pressure drop between valve and spray
balls.
TANKLESS SYSTEM
Advantages
Decreased capital expense (no tank, filter, etc.)
Increased perception of improved system sterilitydue to “totally welded tubing”
If system is drained daily, potential WFI lost maybe less than storage tank system.
Disadvantages
Added capital expense of tank, filter, etc.
Sanitization/steamout may be more involved than
for a tankless system.
If the system is drained daily, potential loss of WFImay be greater than tankless loop.
Disadvantages
WFI tank may be required to satisfy peak
demands of ambient system.
Thermal expansion of WFI during hot watersanitization has nowhere to relieve, except flush todrain.
More difficult to isolate than storage tank system.
System hydraulics more difficult to manage/controlthan storage tank system.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 137/250
119
STORAGE AND DISTRIBUTION SYSTEMS
8.7 MICROBIAL CONTROL DESIGN CONSIDERATIONS
In any given water storage and distribution system, there are certain fundamental conditions that can alwaysbe expected to aggravate a microbial problem. Likewise, several basic measures will always tend to counter-
act such problems. Fundamental conditions that typically could aggravate the problem include:
• Stagnant conditions and areas of low flow rates
• Temperatures that promote microbial growth (15-55°C)
• Poor quality supply water
Some basic measures that have been shown to alleviate such problems are:
• Maintaining ozone levels of 0.02ppm to 0.2ppm
• Continuous, turbulent flow
• Elevated temperatures
• Proper slope
• Smooth, clean surfaces that minimize nutrient accumulation
• Frequent draining, flushing, or sanitizing
• Air breaks in drain piping
• Ensuring no leaks in the system
• Maintaining positive system pressure
All pharmaceutical water must meet the EPA standard for microbiological quality of potable water; whichmeans it must basically be free of specific indicator organisms. Beyond that, microbiological quality for non-
compendial water should be based upon its intended use and the types of products that will be formulatedwith it.
It is important to note though that although the required microbial population acceptance level for USPcompendial purified water is 100 CFU/ml, reliance on such a single parameter can be misleading. The 100
CFU/ml limit may generally be applied to the manufacture of solid oral dosage forms. Many times, however,aqueous or topical formulations require tighter controls to maintain product quality. The USP points out that
these types of products have been the subject of recalls when found to be contaminated with gram negativeorganisms, and the typical microbiological flora of water are gram negative organisms.
A common appropriate approach to dealing with this key issue involves the use of trend analysis. Using suchan approach, alert and action levels are related to the system norm. In this context, strategies for responding
to the alert and action levels can, and should, be developed. Even with the most conscientious design, theremay be places in which biofilm can form. Good Engineering Practices, such as eliminating deadlegs, ensur-
ing adequate flow velocities through out the system, and periodic sanitization help control microbial activity.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 138/250
120
STORAGE AND DISTRIBUTION SYSTEMS
It is common practice, therefore, to store and distribute water in a circulating system under any of the follow-ing scenarios:
• At conditions which are self-sanitizing such as above 65°C or under ozone.
• At conditions that limit microbial growth such as below 10°C with periodic sanitization.
• At ambient temperatures where sanitization is determined by the validated methods that control micro-bial growth
8.7.1 Regulatory Clarification to Common Industry Practices
The following are industry practices that are all Good Engineering Practices (GEPs), and have been per-
ceived in the past to reduce the chance of microbial growth.
If you collectively ignore all of these items, you increase the likelihood of having a bioburden problem. These
items include finishes, storage tank orientation, storage tank isolation, storage tank turnover, piping slopeand drainability, deadlegs and velocity.
8.7.1.1 Finishes
Common industry practices typically range from milled pipe to 320 grit (0.38 microns Ra) mechanical polish
with electropolish. Electropolishing is a reverse plating process, which improves the surface finish of me-chanically polished stainless steel piping and equipment. It reduces surface area and removes surface intru-sions caused by mechanical polishing, which may cause subsequent rouging, and/or discoloration. After
mechanically polishing or electropolishing the system, the polishing compounds should be confirmed to havebeen completely removed from the pipe, so as not to accelerate corrosion.
The benefits for electropolish or finishes smoother than 0.76 microns Ra (approx. 180 grit or 30 micro inch)
are questionable.
Systems operating at ambient temperature or with infrequent sanitization may require a smoother surface
finish. The interior surfaces of stainless piping systems, in compendial water service, are typically groundand/or electropolished, at considerable cost, to achieve a smooth surface of minimal porosity (0.4 to 1.0
microns Ra), in order to reduce bacterial adhesion and enhance cleanability. A viable alternative is extrudedPVDF piping, which can produce a smoother surface than most metallic systems, without recourse to polish-
ing, although PVDF has other disadvantages. (See Section 8.4.)
8.7.1.2 Storage Tank Orientation
Vertical orientation is the most common because of the following advantages:
• Lower fabrication cost
• Less dead volume
• Simpler spray ball design
• Less floor space required
• Horizontal vessels are used where height is a constraint
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 139/250
121
STORAGE AND DISTRIBUTION SYSTEMS
8.7.1.3 Storage Tank Isolation
Common practice for compendial and non-compendial waters where microbial contamination is a concern isto use a 0.2 micron hydrophobic vent filter.
For hot storage vessels, the vent filter must be heated to minimize moisture condensation. An alternate
practice is to blanket the tank with 0.2 micron filtered air or nitrogen. Nitrogen blanketing can be used if CO2absorption is a concern, or if final product oxidation is a problem.
8.7.1.4 Storage Tank Turnover
Common practice is 1-5 tank turnovers per hour.
The turnover rate may be important for systems using external sanitization or polishing equipment.
The turnover rate is less important when storage is under sanitizing conditions, including hot storage or
ozone. It may be less important under conditions that limit microbial growth, such as cold storage (4-10°C),but this must be demonstrated by documentation.
Some storage tank turnover is required to avoid dead areas.
8.7.1.5 System Drainability
Systems that will be steam sterilized must be fully drainable to assure complete condensate removal.
Systems that will never be steam sterilized do not need to be fully drainable, as long as water is not allowedto stagnate in the system.
It is good engineering practice to allow for the draining of equipment and associated piping.
8.7.1.6 Deadlegs
Good engineering practice is to minimize or eliminate deadlegs where possible. Common practice is to limitdeadlegs to less than 6 branch pipe diameters or less. This stems from the “6D” rule contained in the pro-
posed CFR 212 regulations of 1976. Recently, industry experts have suggested using a guideline of 3D orless. However, this new guideline causes confusion since the proponents of this standard generally are
discussing the length of dead leg from the outer wall of the pipe, while the original 6D rule describes thedistance from the center line of the pipe to the end of the deadleg. Obviously, if a 1/2” branch is placed on a3” main, the distance from the center line of the pipe to the outer wall of the pipe is already 3D. Thus, even a
zero deadleg valve might not meet the 3D requirement.
To avoid confusion in the future this Guide suggests that the length of the deadleg be considered from theouter wall of the pipe. We propose avoiding a hard rule of thumb for maximum allowable deadlegs.
Ultimately, the water must meet the required quality regardless of the length of the deadleg. Good Engineer-ing Practice requires minimizing the length of deadlegs and there are many good instrument and valve
designs available to do so.
It should be recognize that any one-way system can constitute a deadleg if it is not frequently flushed orsanitized.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 140/250
122
STORAGE AND DISTRIBUTION SYSTEMS
8.7.1.7 Positive Pressure
It is important to maintain the system under positive pressure at all times. One common concern is systemsdesigned with insufficient return flow, which could draw a vacuum at points of use under periods of high water
usage. This causes an unexpected microbial challenge to the system.
8.7.1.8 Loop Velocity
Common practice is to design circulating loops for a minimum return velocity of 3 feet/second or higher, and
for Reynolds numbers in the turbulent region of greater than 2,100.
Return velocities less than 3 feet/second are acceptable for short periods of time, or in systems that do notfavor microbial growth, such as hot, chilled, or ozonated loops.
A minimum return flow is required to maintain the loop in a flooded condition under positive pressure.
8.8 CONTINUOUS MICROBIAL CONTROL
Process water systems generally employ both continuous methods of microbial control, and periodic saniti-zations. This section discusses continuous methods for controlling microbial growth.
8.8.1 “Hot” Systems
The most effective and reliable means of preventing the growth of bacteria is to operate the system at tem-
peratures above which bacteria can survive. If the distribution system is maintained in hot conditions, saniti-zation on a routine basis can be eliminated.
Systems operating at 80°C have a long history of data showing the prevention of microbial growth. More
recently, companies have been validating water systems at 65°C. The advantages of operating at lowertemperature include energy savings, lower risk of injury, and reducing the amount of rouging. Systems oper-ating at the higher end of this range have a greater safety margin with regard to microbial contamination. The
effectiveness of temperatures below 80°C must be verified with test data, on a case by case basis.
Note that these temperature ranges will not destroy endotoxin. As noted in Chapter 6, where endotoxin is aconcern, the treatment system must be designed to remove it.
8.8.2 “Cold” Systems
The use of the term “cold” in this case implies that a system is maintained at a low enough temperature toinhibit microbial growth. While this has been shown to be effective, the energy costs associated with it gener-
ally make this type system costly to operation. Generally, “Cold” systems are operated from 4°C to 10°C.Microbial growth rates drop off significantly below 15°C, so the sanitization frequency of cold systems may be
reduced compared to ambient systems. The effectiveness of a specific temperature, and the associatedsanitization frequency in any particular system, must be determined by statistical analysis, on a case by casebasis.
8.8.3 “Ambient” Systems
Circulation temperatures of any pharmaceutical water system are dictated by either the required microbial
specification or the required temperature for usage. “Ambient” temperature purified water systems usingeither ozone and/or hot water sanitizations are common throughout the industry, and normally result in lowerlifecycle costs, as well as reduced energy consumption compared with either the “hot” or “cold” systems.
However, without increased levels of system sanitizations, the lack of temperature control at the water stor-
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 141/250
123
STORAGE AND DISTRIBUTION SYSTEMS
age vessel and distribution loop could result in the formation of a biofilm within the system, which couldsporadically and unpredictably produce water failing microbiological specifications and necessitate non sched-
uled water system shutdowns.
8.8.4 Ozone
Ozone has been shown effective for microbial control. It is a strong oxidant, which chemically reacts withorganisms and destroys them. The destruction of these organisms results in organic compounds, which maybe further degraded by ozone, ultimately to carbon dioxide. Ozone is twice as powerful an oxidant as chlorine
and needs to be dosed continually to maintain concentration.
In any compendial water system and most other applications, water at the use points is expected to be totallyfree of ozone. Ozone removal is commonly effected through ultra violet radiation. 254 nanometer UV light
converts ozone to oxygen. A common design is to maintain an ozone concentration between .02 ppm and0.1ppm in the storage tank, and use a UV light at the beginning of the distribution loop for removal. To sanitizethe loop itself, the UV light can be turned off during periods of no use, and the ozone will circulate through the
loop. The UV dosage required for ozone destruction is generally 2 to 3 times that required for microbialcontrol. Testing should be done to verify absence of ozone at the use points.
8.8.5 UV Light
Ultraviolet lights have been shown to reduce microbial populations in storage and distribution systems. UV
energy is germicidal in the wavelengths of 200 to 300 nanometers, which falls below the visible spectrum. UVlight de-activates DNA leading to bacteria reduction. A UV light is not a sterilization device, as it is oftenreferred to. The effectiveness of the light will depend on the quality of the water in which it is acting, the
intensity of the light, flow rate of water, contact time, and the type of bacteria present.
8.8.6 Filtration
Along with other particulate matter, bacteria and endotoxins may be removed via filtration. This filtrationmedia can be either of the microfiltration (2-0.07 microns) or ultrafiltration (0.1-0.005 micron) scale. Theintegrity of these filters must be maintained.
8.8.6.1 Microfiltration
Microfiltration includes the use of depth cartridge filters, pleated filters, and cross flow filtration membrane
elements. These filters can remove particles ranging in size from 100 microns down to 0.1 micron. Depth andpleated filters allow water to flow through a wall of fibers perpendicular to the water direction (dead endedfilters). The particles are then trapped on the outside wall of these filters, or within the filter walls (for depth
filters), due to the pore size of the filter. The filter will fill up with these particles and then needs to be replacedwith a new filter.
8.8.6.2 Ultrafiltration
Ultrafiltration can be used to remove organics and bacteria, as well as most viruses and pyrogens from awater source. Filtration is typically from 0.1 micron down to 0.01 micron. Cross flow ultrafiltration forces the
water to flow parallel to the filter media, and the particles which are too large to pass through the membraneelements are then expelled from the system in a concentrate stream to drain (typically 5-10% of the feed
flow). This allows the filters to be self-cleaning and eliminates the need to replace these membrane elementsfrequently. This type of filtration can be used as a “maintenance “ step downstream of the storage tank in
certain situations.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 142/250
124
STORAGE AND DISTRIBUTION SYSTEMS
In general, for any purified water system, filtration downstream of the storage tank is not recommended. Thiscomes from concerns of “grow through” where bacteria will colonize on the upstream side of the filter, and
ultimately be found on the downstream side even though the pore size of the media may theoretically besmaller than the size of the bacteria. An additional concern is the potential for accumulation of nutrients on
the filter media, which may in fact increase the opportunity for microbial growth. However, filters downstreamof a circulation pump are sometimes used in water systems. System designs should be predicated on obtain-
ing the desired water quality upstream of the storage tank, using the treatment train. Filters downstream ofthe storage tank should not be relied on to purify the water.
8.8.7 Circulation
Most new water systems use a circulating loop for distribution. The primary purpose of circulation is to reducethe chance of microbial growth, or microbial attachment to the surfaces of the system. Although the mecha-
nisms are not universally agreed upon, it is thought that the shear forces associated with turbulent water flowinhibit nutrient concentration and attachment of bacteria to surfaces. The velocity required to obtain thesebenefits is generally agreed to be greater than 3 feet per second, or Reynolds numbers greater than 2,100.
Velocity may drop off for short periods of time during high use times without adversely affecting the system,so long as positive pressure is maintained in the system. Circulation also serves to maintain proper tempera-
ture throughout the system in hot and cold systems.
Studies have shown that the velocities required to remove biofilm are higher than practical for a water system(above 15 ft/sec). However, a combination of high velocity (5 ft/sec or greater) with an antimicrobial agent,
such as ozone or chlorine, may, over a long enough period of time, effectively remove biofilm.
A turbulent condition may be maintained in short dead ended pipe stubs if the length of the stubs is limited.
This limiting length varies with the pipe stub diameter and to a lesser degree with the main pipe diameter. Arule of thumb for the maximum dead leg is 6 branch pipe diameters. This “rule of thumb” may be difficult to
achieve in large mains with small branches, and may result in unacceptably long dead legs in large branches.Rather than universally applying “rules of thumb”, it is important to recognize dead legs as an area of concern
and take appropriate steps to prevent them in the original design or implement special provisions to addressthem if unavoidable. Some of the factors to consider include operating temperature, velocity in the main, andfrequency of use (if the dead leg is a use point).
8.9 PERIODIC STERILIZATION/SANITIZATION
Periodic sanitization of storage and distribution systems is generally required. Based on monitoring the mi-crobial quality of the system, a required frequency of sanitization should be formally established. Sanitizationmay also be done in response to reaching an “action limit” during routine testing. Various methods of periodic
sanitization are discussed below.
8.9.1 Chemical
Various chemicals or combinations of chemicals can be used to periodically sanitize storage and distributionsystems. Chlorine solutions on the order of magnitude of 100 ppm are very effective at killing organisms, butare not generally used in distribution systems because of corrosion problems associated with stainless steels.
Hydrogen peroxide in concentrations on the order of 5% is a more practical alternative. Peracetic acid canalso be used, generally in concentrations of 1% or less. Many different mixtures of these and other chemicals
are commercially available for the purpose of sanitization.
Verification of the removal of the sanitizing agent is critical. Commercially available indicators (test strips orsticks) are commonly used to indicate when the amount of rinse water is sufficient. A rinse water analysis isthen required to verify the absence of objectionable chemicals before the system is placed into service.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 143/250
125
STORAGE AND DISTRIBUTION SYSTEMS
8.9.2 Ozone
Sanitation can be done either periodically or continuously with ozone. Storage tanks are typically continu-ously ozonated, and then the ozone is removed prior to the distribution loop or individual use points through
the use of ultraviolet radiation. The distribution system can be periodically sanitized by turning off the UV lightand, if necessary, increasing the ozone concentration while allowing it to circulate through the distribution
loop. Concentrations as high as 1 ppm may be needed for periodic sanitization, particularly if biofilms must beremoved.
8.9.3 Heat
It has been found that periodic sanitization by heating of the process water system is highly reliable andeffective. The frequency at which sanitization must occur will vary depending on many factors.
• System design
• Size of distribution system
• Components of system
• Volume of process water in the system
• Frequency of use of the process water (turnover volume)
• Temperature of the circulating process water
Each distribution system must develop its own microbial profile, and the sanitization cycle and frequency will
have to be developed to fit that system.
The most straightforward method of sanitization is to heat the circulating process water in the distributionsystem to 80°C ± 3°C and hold it at that temperature for a validated period of time. The use of this heatsanitization has been proven to be very effective and if designed properly can also be economical. Controls
needed to perform this cycle of sanitization can be either manual or automatic.
Because of the types of bacteria found in purified water systems, the use of steam is not required for effectivemicrobial kill. Steam sterilization of distribution piping may require additional valving for vents and drains, and
may require a higher pressure rating than otherwise needed. Storage tanks are by their nature more easilysteam sterilized and this practice is common, although not required.
Hot systems inherently are continuously sanitized. Thus, the need for sanitization should be based on micro-bial testing results, or when the system is off line for an extended period of time and the temperature of the
loop drops to below the validated temperature range.
Depending on the process water specification, a conservative initial sanitization frequency should be as-signed for “cold” systems. After the operating characteristics of the system are determined through microbialtesting, the routine sanitization frequency can be determined.
8.9.4 Initial Sanitization (Ambient System)
Steam sanitization has a successful history, and is probably the most reliable method for sanitization. How-
ever, there is no requirement for steam sanitization in Purified Water or WFI systems. The following procedureis suggested as one option for hot water sanitization of an ambient system.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 144/250
126
STORAGE AND DISTRIBUTION SYSTEMS
Immediately after passivation (for a SS system), the system should be flushed with process water at a hightemperature (80°C ± 5°C) and all valves opened and points of use flushed. Normally two (2) times the volume
of the system (after conductivity readings), or rinse water tests indicate that no passivation chemicals aredetected, is required. This is the initial sanitization of the system.
Once it has been determined that the chemical characteristics for the quality of the process water have been
achieved by USP chemical testing, then microbial samples should be taken after each component, the pointsof use, and the storage tank. This initial sampling should show that the distribution system at any samplingpoint has no viable bacterial contamination. Once this is achieved, the system should be brought down to its
operating temperature and allowed to stabilize.
8.10 SYSTEM DESIGN FOR STERILIZATION/SANITIZATION
The following sections highlight particular aspects of storage and distribution system design, which are rel-evant to sanitization.
8.10.1 Materials of Construction
The sanitization methods used must be compatible with the materials in the system. By far the most widelyused material for storage tanks and piping is 300 series stainless steel (generally 316L). This choice providesthe most flexibility with regard to sanitization options. Sanitization with heat, UV, or ozone can be used in
stainless systems practically without restriction. Chemical sanitization must be carefully managed with re-gard to concentration, pH, and temperature to avoid corrosive effects on stainless distribution systems.
Other material used for distribution piping is PVDF. PVDF is susceptible to degradation by UV light. It iscommon to use stainless piping immediately adjacent to the UV light in a PVDF system to compensate for
this problem. The temperature limitation of PVDF is approximately 140°C, which is high enough to allow heatsanitization or sterilization.
In stainless systems, the gaskets used must be reviewed for compatibility with the sanitization method. Awidely used gasket material is PTFE or EPDM, both of which have good thermal memory and excellent
resistance to temperature, ozone, and chemical sanitizers. Other gasket materials must be carefully re-viewed for compatibility with the sanitization methods, and to ensure that they will not leach substances into
the water.
The key is to recognize that the materials of construction “shall not be reactive, additive or absorptive so as toalter the safety, identity, strength, quality, or purity of the drug product beyond the official or other establishedrequirements” (21 CFR 211.65). The sanitization procedures must be considered when selecting materials to
comply with this requirement.
8.10.2 Storage Tank Design
Storage tanks are an area in the system that could be considered at high risk for microbial contaminationbecause of the large surface area, low velocities, the need for venting, and potential for “cold spots” in thehead space.
Tank size is generally based on economic considerations in combination with the pretreatment train sizing.
From a bacterial standpoint, smaller tanks are preferred because they have higher turnover rates, whichreduce the likelihood of bacterial growth. They also reduce surface areas and make it easier for ozone to
permeate the water, if the tank is ozonated.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 145/250
127
STORAGE AND DISTRIBUTION SYSTEMS
Spray balls may be used on the return loop to wet the head space of storage tanks. The use of a spray ballserves to keep the top of the tank at the same temperature as the water, in heated systems, and avoids
alternately wet and dry surfaces, which could promote corrosive action with stainless steel and allow micro-bial growth. Connections on the top head (relief devices, instrument connections, etc.) should be kept as
close to the head as possible to simplify the spray ball design and get the benefit of the spray action. Anexception is the vent filter, which should be removed far enough from the storage tank to avoid direct contact
from the water spray, which could blind the filter. If dip tubes or instruments project down from the head,multiple spray balls may be needed to avoid a “shadow” being created in the spray pattern.
The tanks must be vented to allow filling, and a filter should be used at the vent to avoid airborne particulateand microbial contamination. To avoid the problem of condensation in the filter and the resultant potential for
colonization and grow through, hydrophobic vent filters are used and/or the filters are maintained at a tem-perature above the tank temperature with steam jacketing or electric tracing.
To help avoid microbial growth, and avoid the change in conductivity resulting from absorption of atmosphericgasses into the water, nitrogen blanketing on the head space may be used. This eliminates outside air pass-
ing into the tank through the vent filter. Note that gasses added to storage tanks should be appropriatelyfiltered to avoid objectionable contamination.
Table 8-5 Comparison of Alternate System Designs for Microbial Control in Storageand Distribution
Note 1: All systems are circulating
Note 2: Operating costs and effectiveness will increase with frequency of sanitization
Microbial ControlMethodology
Ambient system, with ozonated
tank, periodic ozone indistribution piping
Ambient system with periodichot water sanitization (note 2)
Continuous “Cold” system
(4-10°C) with periodic hot watersanitization
Continuous “Hot” system
(65-80°C) with multiple Point ofUse Coolers
RelativeEffectiveness/Reliability
Good
Good
Better
Best
Operating Cost
Low
Low
High, unless
cold water isrequired for
process
Medium
Installed Cost
Low
Low
Medium
High
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 146/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 147/250
INSTRUMENTATION and
CONTROL
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 148/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 149/250
129
INSTRUMENTATION AND CONTROL
9. INSTRUMENTATION AND CONTROL
9.1 INTRODUCTION
Instrumentation and controls are often used within pharmaceutical water systems to:
control the operation of equipment and components
• monitor and document the performance of critical equipment
• monitor and document pharmaceutical water quality
The concepts and regulatory philosophy of defining critical versus non-critical parameters is discussed as itrelates to instrumentation and controls. This definition could be summarized as:
“All instruments and control systems should be commissioned following Good Engineering Practices. Critical instruments and control systems should be commissioned and qualified.”
There is no regulatory requirement that requires the use of On-Line instrumentation. A monitoring programmay include a combination of On-Line instrumentation, manual documentation, and laboratory analysis.
If On-Line instrumentation is used to measure or record a critical parameter, action and alert limits may beestablished. The methods of addressing “spikes” are also discussed.
Automation can have a significant impact on the cost and performance of a pharmaceutical water system.There is no single optimum level of instrumentation and control for all systems. The optimum level for a given
system balances the benefits of improved process control, improved documentation, and lower labor costsagainst the cost of procuring, installing, validating, and maintaining the instruments and control systems. In
many cases, the level of automation for a pharmaceutical water system should be consistent with that utilizedfor the manufacturing process it supports.
9.2 PRINCIPLES
a) To achieve GMP compliance, the manufacturer must demonstrate, through documented evidence, that
the pharmaceutical water system is in control and consistently produces and delivers water of accept-able quality.
b) Although many quality attributes can be continuously monitored using On-Line instrumentation, there isno compendial or regulatory requirement for On-Line monitoring of pharmaceutical water quality. A moni-
toring program typically includes a combination of On-Line instrumentation, manual documentation ofoperational parameters, and laboratory analysis of water samples.
c) Instruments and control systems are critical and must be qualified when they are used to measure,monitor, control, or record a critical process parameter. A critical parameter is a processing parameter
that affects the final water quality.
For example, the temperature of the final water product may be considered critical for microbial control. Inthis case, the temperature controls (e.g., sensors and alarms) would be considered critical. However, it is
not necessary to consider the temperature of the heating media (e.g., steam) as a critical parameter.
Documentation should clearly indicate which instruments are critical and which are not. It is also advis-
able to identify non-critical instruments as such on the field device.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 150/250
130
INSTRUMENTATION AND CONTROL
d) All instruments and control systems should be designed, installed, calibrated, and maintained appropri-ately, according to Good Engineering Practice. All critical instruments and controls require qualification.
e) Items that should be recorded in the system documentation include maintenance procedures and main-
tenance work performed, procedures for sampling and analysis, reporting the results, and trend analysisof the laboratory data. The monitoring program during start-up typically defines maintenance frequency
and alert and action levels for the process variables.
9.3 GENERAL INSTRUMENTATION REQUIREMENTS
9.3.1 Instrument Selection and Installation
a) Instruments should be selected for accuracy and reliability over the entire process range.
b) Instruments should be selected and installed in a way that reduces the potential for contamination.
• Water contact surfaces should be constructed of materials that are compatible with the water they
contact. Materials of construction and surface finishes (see Chapter 8) are commonly specified for
instruments installed in distribution systems.
• Sensors in direct contact with waters with strict microbial limits should be of sanitary design. Non-
sanitary instrumentation is commonly used in feed water and pretreatment systems.
• Instruments may be installed directly in the water system or in a side stream that may, or may not, be
returned to the main system.
• Deadlegs should be avoided.
c) When possible, instruments should be installed such that exposure to harsh process conditions, such aspH and temperature extremes, is avoided. For example, In-Line sensors used to monitor effluent from adeionizer should be positioned such that exposure to regeneration chemicals is avoided.
Instruments that are not compatible with passivation agents, sanitization agents, or sanitization tempera-
tures should be installed so that they may be easily removed or bypassed. Such devices may need to besanitized off line.
d) Instruments should be installed in accordance with manufacturers’ requirements to ensure proper opera-tion. For example, flow meters should be installed in the proper orientation and with the correct upstream
and downstream straight run of pipe. The impact of process and ambient conditions on an instrument’saccuracy and reliability should be addressed.
e) Conductivity cells are especially sensitive to the presence of air or steam bubbles, which can be present
where there is turbulence, cavitation, or flashing. Such locations should be avoided.
f) Accessibility for maintenance should be considered, but improving control response is usually more
critical. Poor response time may be a consequence of the poor placement of a device and, in most cases,can be improved by installing the device closer to the point of measurement.
9.3.2 Instrument Calibration
a) The calibration of critical instruments should follow a regular program, which provides evidence of con-sistently acceptable performance. Non-critical instruments may be calibrated on a frequency deemed
appropriate for the service.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 151/250
131
INSTRUMENTATION AND CONTROL
b) Calibration should follow approved procedures and the results should be documented. Each componentin a control loop should be calibrated individually or the loop may be calibrated in its entirety. All calibra-
tions should be traceable to certified standards (e.g., NIST).
c) Vendor-supplied calibration certificates should reference the applicable instrument serial numbers. Theimpact of shipment and installation on the vendor’s calibration should be addressed.
9.3.3 Types of Instrumentation
9.3.3.1 Conductivity
a) Although non-ion specific, conductivity is a valuable tool for measuring the total ionic quality of water andis a critical parameter for many high purity water systems. Conductivity limits for Purified Water and WFI
are specified in the USP.
b) On-Line conductivity instrumentation is frequently used to monitor and control the performance of many
types of purification equipment and to continuously monitor the quality of pharmaceutical waters. On-lineconductivity instrumentation may also be used for final quality assurance testing, thus eliminating the
need for periodic laboratory analysis of water samples.
c) Temperature has a profound impact upon conductivity measurement. To eliminate this temperature de-pendence, most instruments include a temperature sensor in the conductivity probe and one or more
algorithms to correct the actual measurement to a standard temperature. However, due to he inaccuracyinherent in temperature compensation algorithms, compensated conductivity measurements are notsuitable for critical quality assurance testing of USP purified water and WFI. When In-Line conductivity
measurements are used for final quality assurance testing of USP purified water and WFI, a non-com-pensated conductivity value and the water temperature must be measured as required by the USP.
Compensated conductivity values used strictly for process control and monitoring are not subject to USPrequirements.
d) To operate properly, conductivity sensors must be installed such that there is continuous water flowthrough the sensor and air bubbles or solids cannot become trapped inside the electrodes. Air bubbles
will result in lower-than-expected conductivity readings while solids can impact the conductivity in eitherdirection. Clean steam must be condensed prior to conductivity measurement.
e) Conductivity meters may be used throughout a pharmaceutical water system to monitor and control
purification processes or to monitor pharmaceutical water quality. Some examples are:
• Feed water monitoring can detect seasonal or unanticipated quality changes that could impact pre-
treatment equipment operation.
• RO influent and effluent monitoring allows calculation and trending of percentage rejection. Changesin percentage rejection may be a sign of membrane failure, scaling or fouling, seal failure, improper
pH, inadequate feed pressure, or too high a recovery rate.
• Deionizer effluent or in-bed monitoring detects, or predicts, resin exhaustion and allows automatic
initiation of regeneration cycles.
• The conductivity of pharmaceutical water may be monitored after the final treatment step to verifyacceptable quality prior to delivery to a storage tank. In addition, conductivity meters are often in-
stalled in the return piping of distribution loops downstream of the final point of use. Many systemsinclude provisions for automatic diversion to drain or recirculation back through purification equip-ment when water quality entering the tank is outside the acceptable range.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 152/250
132
INSTRUMENTATION AND CONTROL
9.3.3.2 Total Organic Carbon (TOC)
a) Total Organic Carbon (TOC) is a measure of the carbon dissolved in water in the form of organic com-pounds. It is a valuable tool for measuring the aggregate level of organic impurities in pharmaceutical
water systems. A TOC test with a nominal limit of 500 ppb for USP Purified Water and WFI is a requiredtest in the USP.
b) TOC meters are relatively sophisticated analytical instruments. The USP provides guidance on how toqualify an instrument and how to interpret the instrument results.
c) In addition to “continuous” monitoring of equipment performance and pharmaceutical water quality, On-
Line TOC meters may be used for final quality assurance testing, thus eliminating the need for periodiclaboratory analysis. When used for critical assurance testing of USP purified water and WFI, instrument
precision, system suitability, test methodology and calibration procedures must meet USP requirements.Instruments used strictly for process control and monitoring are not subject to USP requirements.
d) TOC is often monitored at several locations in a pharmaceutical water system. Some examples are:
• Feed water monitoring can detect seasonal or unanticipated quality changes that could impact pre-
treatment equipment operation or the potential for resin or membrane fouling.
• Monitoring TOC downstream of carbon filters, organic scavengers, RO units, and UV lights can verify
proper equipment operation and provide advance warning of bed exhaustion, compromised mem-branes, or the need for lamp replacement.
• TOC levels of pharmaceutical water may be monitored after the final treatment step to verify accept-able quality prior to delivery to a storage tank. In addition, TOC meters are often installed in the return
piping of distr ibution loops downstream of the final point of use. Many systems include provisions forautomatic diversion to drain or recirculation back through purification equipment when water quality
is outside the acceptable range.
e) There has been much interest in the possible use of TOC analyzers to indicate endotoxin contamination.
While this type of contamination will lead to higher TOC levels, there is no quantitative correlation to TOClevels. TOC results cannot substitute for microbial or endotoxin testing.
9.3.3.3 PH
a) pH measurement is relatively straightforward for high conductivity water. Reliable results can generallybe obtained using pH indicators or laboratory, field, or On-Line pH meters.
b) Accurate pH measurement is difficult in many pharmaceutical waters due to the low conductivity. Low
conductivity water is susceptible to pH swings due to contaminants introduced from the air, samplecontainers, and test equipment, as well as instrument difficulties associated with measuring low ionic
strength solutions.
c) Common locations for On-Line pH measurement and control include:
• Upstream of cellulose acetate RO membranes, where acid is injected to minimize membrane hy-
drolysis
• Upstream of a degasifier, where acid is injected to increase C02 removal
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 153/250
133
INSTRUMENTATION AND CONTROL
d) While pH limits for purified water and WFI are no longer specified in the USP, On-Line pH meters arerarely used for process control or for final quality assurance testing of pharmaceutical waters for several
reasons:
• Conductivity is a more sensitive measurement of overall ionic quality since changes in pH reflectlogarithmic changes in water quality
• A pH sensor’s reference electrode contains a buffer solution that may leak through the referenceelectrode into the water being measured. To prevent contamination of the pharmaceutical water
system, a pH sensor is installed in a side stream that flows to drain. The water flow rate through themeter must be controlled and held constant to achieve repeatable results.
• pH meters require frequent (daily in some cases) calibration with standard buffer solutions
9.3.3.4 Ozone
a) Dissolved ozone levels should be monitored in storage and distribution systems that utilize ozone formicrobial control. Ozone levels can be determined periodically in the laboratory using several wet chem-
istry methods, or continuously using an On-Line analyzer. On-line analyzers are relatively inexpensive
and easy to maintain, but they should periodically be calibrated against laboratory methods.
b) For effective and safe system operation, ozone levels should be monitored at the following locations:
• At the storage tank discharge to control operation of the ozone generator
• Downstream of the UV light to ensure ozone destruction prior to water use
• In loop return piping to ensure proper ozone levels are maintained during sanitization
c) Since Oxidation Reduction Potential (ORP) analyzers are nonspecific and unable to differentiate ozonefrom other oxidants, ORP analyzers should not be used for controlling ozone levels in pharmaceuticalwater systems.
9.3.3.5 Flow
A wide variety of flow meters may be used in the feed water and pretreatment portion of a pharmaceutical
water system including magnetic flow meters, mass flow meters, vortex shedding meters, and ultrasonicmeters. All meters should be installed according to the manufacturer’s instructions, to ensure proper opera-tion.
Water flow rate (or velocity) may help to reduce microbial growth and maintain temperature within hot or cold
systems. It is commonly verified upon startup, but not continuously monitored. Flow rate may vary. It may bemonitored for information only.
9.3.3.6 Temperature
Temperature is often monitored and/or controlled at various locations to ensure optimum equipment opera-tion and/or for microbial control. Temperature interlocks may be used to prevent damage to membranes,
resins, or equipment if water temperatures drift outside allowable ranges.
In distribution systems where temperature is controlled or where heat sanitization is used, temperature isconsidered critical to ensure proper system operation or effective sanitization.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 154/250
134
INSTRUMENTATION AND CONTROL
9.3.3.7 Pressure
Pressure may be monitored and controlled throughout the purification process to ensure optimum equipmentoperation. Monitoring differential pressure across filters indicates when backwashing or element replace-
ment is needed. Differential pressure measurement across resin beds is useful in detecting resin fouling andpoor flow distribution. Monitoring RO feed, interstage, permeate, and concentrate pressures provides early
warning of membrane fouling and scaling. Back pressure control in distribution systems may be critical, ifminimum pressures are required at points of use.
Pressure is not normally considered a critical parameter, however, the system should maintain positive pres-sure at all times. It may typically be monitored for information only.
9.3.3.8 Level
Various types of level measurement are used in the feed water and pretreatment portion of a pharmaceuticalwater system, including simple float switches, ultrasonic sensors, capacitance sensors, and differential pres-
sure transmitters. The stub from the tank must be kept as short as possible to minimize deadlegs. Calibrationof this type of transmitter is time consuming, since it requires filling the tank to verify proper operation. Tank
nozzles with integral valves minimize deadlegs and allow calibration while the tank is in service.
Tank level may be monitored to control the supply of water into a tank and for control and protection ofdownstream pumps.
In some instances, level may not normally be considered a critical parameter and may be monitored forinformation only. In these cases, it is usually not validated.
9.4 DESIGN CONDITIONS VERSUS OPERATING RANGE
The control system may recognize the distinction between design conditions and operating ranges, and theimpact this distinction has upon validation and facility operation. These criteria are defined as:
• Design Condition: the specified range, or accuracy, of a controlled variable used by the designer as abasis to determine the performance requirements for an engineered system.
• Allowable Operating Range: the range of validated critical parameters within which acceptable water
product can be produced.
• Normal Operating Range: a range that may be selected as the desired acceptable values for a param-
eter during normal operations. This range must be within the Allowable Operating Range.
a) While it is desirable that a facility should meet all stated design conditions, the acceptability of the watersystem for operation from a cGMP standpoint depends on operating within the Allowable Operating
Ranges.
b) Normal Operating Ranges cannot exceed the Allowable Operating Range for the product water. Design
condition selection should reflect Good Engineering Practice.
c) It may be desirable to apply the concept of Alert and Action points along with Normal Operating Range.Alert levels are based on normal operating experience and are used to initiate investigations or corrective
measures, before reaching an Action level. Action levels are defined as the level at which some correctiveaction must be taken to avoid jeopardizing water quality.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 155/250
135
INSTRUMENTATION AND CONTROL
9.5 INSTRUMENTATION SPIKES
“Spikes” may be experienced in the measurement of some parameters. These excursions may be the resultof the measurement technique or sensor and may not be representative of the actual parameter value. If a
spike occurs in a system with a significant physical lag or mass, the rapid changes in a parameter as evi-denced by spikes may be physically impossible and consequently can be treated as instrumentation spikes.
In other cases, it may be decided to treat these spikes as Alert Level deviations based upon their frequencyand duration even though their magnitude may exceed the Action level.
A procedure for defining and handling spikes should be developed in conjunction with Quality Assurancebased on the specific water system.
9.6 CONTROL SYSTEMS
9.6.1 Level of Automation
Selection of a control strategy for a pharmaceutical water system should consider feed water quality and
reliability, the complexity of the purification and/or distribution system, labor costs, personnel skill levels and
capabilities, and documentation and reporting requirements. Options for control include:
a) Local instrumentation with manual control: In this option, a combination of instrumentation, periodic
samples, and visual examination is used to monitor critical process parameters. Data is collected andrecorded manually, and analysis and trending capabilities are limited. Excursions of critical parametersoutside acceptable ranges typically trigger local alarms to reduce the risk of unacceptable water quality.
Satisfactory manual operation requires significant human intervention. This requires detailed operatingprocedures and conscientious documentation of critical quality parameters. This option has the lowest
installed cost, but is very labor intensive and may be subject to human error.
b) Semi-automatic control: These systems use local operator control panels, relay logic control, localchart recorders and printers, and some manual data collection to monitor and control the water system.These systems are less labor intensive over the manual systems, but are still labor intensive due to the
manual data collection and monitoring required to control the process.
c) Automatic control: Automated systems use a computer (PLC or DCS), or computers, to control thepharmaceutical water system. The computer system utilizes appropriate process monitoring instrumen-
tation (conductivity probes, flow meters, etc.) to gather data and make appropriate adjustments to thesystem automatically. As water generation systems become more sophisticated, relying on human inter-vention to control and monitor the water system becomes more difficult and labor intensive. An auto-
mated system requires less operator involvement, but requires a more highly trained maintenance andengineering support staff.
d) Fully integrated systems: These systems include a fully automated system and a wide area network
connected to other computer systems in the building or site. These systems allow for central site monitor-ing, automatic electronic data collection, centralized alarm monitoring with automatic recording, response,and report generation.
Additional information on control system design is available in the Good Automated Manufacturing Practice
(GAMP) Guide and in various guidelines by the Instrument Society of America (ISA).Whichever level of automation is selected, the validation effort should verify operation of the complete sys-
tem, including vendor-supplied sub-systems.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 156/250
136
INSTRUMENTATION AND CONTROL
9.6.2 Control System Software
The software/control system may be used to measure, monitor, control, or record critical process param-eters. Programming and design standards, especially concerning operator interface, control techniques, alarm
handling, and interlock processing should be applied during the development, validation and maintenancephases of the project. The control system software consists of:
a) Firmware, Operating System and Application Software: This is software permanently loaded intomemory that may or may not be accessible to the user. While the functions performed by the control
system may be divided between critical and non-critical functions, it is impossible to divide or isolate thefirmware, operating system, application software, and associated hardware functions. Therefore, if some
of the functions of a control system are considered critical, all of the above software is considered critical,and should be validated.
b) User Configurable Software: The functions of the user configurable software may be defined as critical or non-critical. The critical functions or modules require enhanced documentation, including validation. In
some cases, it may be impossible to divide or isolate software adequately. In such cases, if some of thefunctions are critical, it may be necessary to validate all the software.
The type of process control required is often the determining factor in the type of software needed, andsoftware requirements often define the type of system selected. Major considerations are:
• Number of I/0 points
• Mathematical or statistical functions required
• Reporting features required (particularly if the control system is to be further integrated into higher sys-
tems)
• Whether or not advanced control techniques are required (e.g., neural nets; fuzzy logic controllers; adap-tive gain; dead-time compensation)
9.6.3 Control Hardware and Operation Interface
a) Critical software requires enhanced documentation and should be designed and tested in accordancewith the Lifecycle Methodology.
b) The water system, field instruments and control requirements all affect control hardware selection. Plantstandards, or a large installed base of a particular system may drive the selection.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 157/250
COMMISSIONING and
QUALIFICATION
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 158/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 159/250
137
COMMISSIONING AND QUALIFICATION
10. COMMISSIONING AND QUALIFICATION
10.1 INTRODUCTION
Commissioning and qualification comprise the validation process by which a system is put into service and
demonstrated to consistently produce water of a specified quality, under various conditions, while operatedunder set procedures. Although commissioning and qualification are typically separated within a project sched-ule, they are in essence, one continuous process.
The specific activities and processes during commissioning and qualification will not be discussed in this
Guide. These are considered by a separate ISPE Baseline ® Guide on Commissioning and Qualification, andpharmaceutical water systems are used as examples throughout. A summary of key concepts are listed
below:
a) Due to the interdependence between activities and those involved, excellent communication, planning
and coordination between operations, engineering, commissioning, and validation personnel will enabletimely and cost-effective project completion.
b) Each component of the system should be built in accordance with plans and specifications and should beinspected, tested, and documented by qualified individuals. These activities, and the production of sup-porting documentation, are defined as Good Engineering Practice (GEP).
c) GEP recommends a minimum level of documentation for all systems and equipment. This encompassesdesign, fabrication, vendor testing, construction, field inspection, and commissioning. If these documents
are appropriately planned, organized, and authorized, they may become an integral part of qualificationsupport documentation, thus avoiding redundancy and saving time and money.
d) Design criteria and documentation requirements should be clearly defined early in the design phase, to
ensure clear expectations and appropriate planning, and facilitate timely commissioning and validation.Engineering firms, vendors, and contractors should be required, per the system specifications, to providethe necessary documentation, to avoid unnecessary costs and delays associated with obtaining or cre-
ating these documents.
e) During construction, timely review of documentation and periodic “walk-throughs” can ensure that Instal-lation Qualification requirements are met.
f) Commissioning takes the system from a state of substantial completion to a state of operation. It is thephase of a project that includes mechanical completion, start-up, and turnover. Commissioning incorpo-
rates a systematic method of testing and documenting the system at the conclusion of construction, andprior to the completion of validation activities.
g) Commissioning documents should not be created and executed for the purpose of regulatory compli-
ance. However, commissioning tests and documentation will typically satisfy many installation and op-erational qualification requirements.
10.2 SYSTEM QUALIFICATION DOCUMENTATION
Good Engineering Practice dictates that documentation be developed to provide evidence of the design, and
that the water system operates in accordance with the design. This documentation encompasses engineer-ing, installation, inspection, and testing. Such documentation is common to all system commissioning activi-ties and is partially summarized below:
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 160/250
138
COMMISSIONING AND QUALIFICATION
• A system description stating design intent
• A schematic drawing of the system (P&ID)
• Written system specifications
• Detailed design drawings
• Vendor manuals and drawings
• Field inspection and test reports
• System qualification test results
Because of their critical impact on pharmaceutical production, water systems require additional emphasis oncertain sections of this documentation. Specific design requirements for water and steam systems are con-
tained within the body of this Guide. When compiling documentation related to water systems, particularattention should be paid to the following:
a) Schematic documentation may be enhanced by the inclusion of a system isometric diagram (or dia-grams) indicating location and numbering of welds, relative elevations, slope of lines, and points of drain-age.
b) The system specification should indicate performance criteria, as well as design parameters.
c) Field inspection and test reports should include cleaning and passivation procedure and record, weldparameter documentation and inspection reports, slope verification, and verification of the absence of
“dead-legs”.
d) System qualification tests may or may not be subject to a pre-approved protocol addressing qualificationtest requirements. In either case, test results should be reported in direct comparison to acceptancecriteria derived from system design and operating specifications.
e) System qualification tests should include verification of all automated functions, specified temperature
control, distribution system velocity, and initial water quality determination.
Additional details regarding water system qualification may be found in the associated “ISPE Baseline ® Guideon Commissioning and Qualification”.
10.3 SYSTEM QUALIFICATION SAMPLING PROGRAM
The qualification of water systems is unique in that performance must be proven over an extended period of
time, and is subject to variations in use rate and initial feed water quality. Therefore, the sampling programassociated with pharmaceutical water systems validation is unique and specialized.
Extensive sampling is required to establish and confirm that the entire system will operate within specifiedoperating ranges, to develop and evaluate the system operation and maintenance procedures, and to verify
that the water produced and delivered to the points of use consistently meets the required quality specifica-tions and acceptance criteria. This portion of the program is sometimes termed performance qualification.
Because of the critical impact that water has upon pharmaceutical quality, the sampling program and evalu-ation of results is usually subject to a pre-approved plan or protocol, with clearly defined acceptance criteria.
Also included should be procedures to deal with deviations from specified parameters and analytical results.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 161/250
139
COMMISSIONING AND QUALIFICATION
The sampling program consists of three successive phases, each with a specific purpose and samplingscheme, as outlined below. The initial phase of the sampling program typically begins after the water system
is shown to be fully operational, as demonstrated through integrated system testing in Operational Qualifica-tion.
The water generated during the various phases may be used for manufacturing as long as analytical results
are acceptable. The intended applications and impact of water quality should be considered in determininghow much data is required before use.
Table 10-1 Sampling Program
Primary Objectives
• Develop appropriate operating ranges.
• Develop and finalize operating, cleaning, and maintenance proce-
dures.
• Demonstrate production and delivery of water of the requiredquality.
• Demonstrate consistent operation within established ranges.
• Demonstrate consistent production and delivery of water of therequired quality.
• Demonstrate extended performance.
• Ensure that potential seasonal variations are evaluated andtreated.
Phase
1
2
3
TypicalDuration
2-4 weeks
2-4 weeks
One year
10.3.1 Phase 1
The purpose of this phase is to establish appropriate operating ranges and provide data for the developmentof cleaning and sanitization procedures and frequencies. Sampling should be performed after each step inthe treatment process and from each point of use. In addition, the incoming feed water to the water system
should be tested and verified to comply with the relevant “Drinking Water” regulations. The FDA Guide toInspections of High Purity Water Systems suggests daily sampling for two to four weeks, but recognizes that
other sampling programs may be acceptable.
In devising the sampling scheme, consideration should be given to the system configuration, maintenancecycles, how the water is drawn for use, and the expected or potential variation in chemistry and microbiologi-cal attributes, at each potential sample point. In treatment, chemistry testing is specific for each processing
step and microbiological testing between each component is important to determine the microbial load andthe component’s ability to manage the load.
At the end of this phase, the SOPs for system operation and maintenance should be developed and approved
for continued interim use during the next phase. System logs, documentation for critical parameters (e.g.,conductivity and TOC data, sanitization data, etc.), and responses to critical alarms or action levels should bereviewed, to verify the appropriate procedures are in place. In addition, the process that will be followed to
investigate a confirmed test failure should be developed at this time. The intent of this process is to assess
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 162/250
140
COMMISSIONING AND QUALIFICATION
whether a failure is localized (i.e., isolated to a specific port) or systematic, and to define how different typesof failures will be handled.
10.3.2 Phase 2
The second phase is intended to demonstrate that the system consistently operates within pre-determined
operating ranges and delivers water of the required quality when operated in accordance with the SOPs. TheFDA Guide suggests that the sampling scheme and duration should be the same as for phase one. Duringphases 1 and 2, multiple samples should be taken from each point of use. Sampling methodology should be
representative of the way water will be used. For example, sampling should not involve a lengthy preliminarypurge if water usage will be direct and immediate. If water is used through an attached hose, then the sample
should be taken from the hose.
It is recommended that each point be sampled at least once per week as a minimum. In this manner, localizedcontamination may be discovered. (Note that too frequent sampling of little used points may mask incipientlocalized microbial growth by artificial purging.) Phase 2 allows the gathering of sufficient data to establish
microbial alert and action limits (see Section 10.4).
10.3.3 Phase 3
The third phase is intended to demonstrate that, when operated for an extended time period (typically oneyear), the system produces and delivers water of the required quality, despite possible seasonal variations of
the feed water. Sample locations, frequencies, and test requirements are based on the established proce-dures. For WFI systems, the FDA Guide recommends sampling daily from a minimum of one point of use,with all points of use tested weekly. At the end of this phase (i.e., after a full year of testing), the validation is
considered completed. In most cases, ongoing monitoring will establish a continuing record of water quality.
10.4 ACCEPTANCE CRITERIA
Acceptance criteria for water are dependent upon its use. For both Purified Water and WFI, the chemicalacceptance criteria are clearly described in the US Pharmacopoeia (USP). It is expected that a well designed
water system, operating within specified design parameters will consistently be able to meet these criteria.Therefore, failures in chemical analysis during phases 1 and 2 must be investigated, the reason for failure
corrected, and (except where errors in sampling or laboratory error are clearly indicated) the sampling phaseextended to re-establish consistency of performance.
Microbial quality is not specified by the USP, but is established by the user based upon water use. The USPdoes recommend action limits for the different waters in its General Information chapter. These are 10 CFU/
100ml for WFI and 100 CFU/ml for Purified Water. These may be employed as initial acceptance criteria forsystem qualification, although some flexibility is allowable, depending upon system design and use. It is
permissible that a single excursion, followed by acceptable re-sampling would not constitute a failure. Inaddition, because of the inherently bacteriostatic nature of WFI production and distribution systems, it should
be expected that the large majority of samples should be well below the initial acceptance criterion. There-fore, for WFI it is prudent to establish a sample average acceptance criterion, which will be below the limit fora single sample. Failure investigation would be handled similarly to chemical analysis failure.
During phases 1 and 2, normal system microbial limits may be established. Acceptance criteria may then be
converted to alert and action limits for use during phase 3 and beyond. These would take into accountrepeated excursions from the norm as well as step increases in micro count.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 163/250
141
COMMISSIONING AND QUALIFICATION
10.5 QUALIFICATION REPORTS
Qualification data should be compiled and conclusions written into a summary report. This is to be reviewedand approved by those responsible for operation and quality assurance of the water system. An interim report
should be written and approved at such time during the qualification sampling program, as it is desired to usewater in production activities. A summary report should be prepared at the conclusion of phase 2, periodic
updates provided throughout phase 3 and a major update issued at the conclusion of phase 3.
10.6 CHANGE CONTROL AND REQUALIFICATION
Changes to the system must be assessed with regard to potential impact of the change on the entire system.Required action would be determined based on that assessment. It may involve extensive re-qualification,
localized increase in sampling frequency, or inclusion in the routine monitoring program.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 164/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 165/250
APPENDIX
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 166/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 167/250
143
APPENDIX
11. APPENDIX
11.1 USP REGULATED WATER QUALITY
See Chapter 2, Section 2.3 on USP Regulated Water Quality for more details.
In order to set the maximum allowable conductivity, USP determined the limit concentration of the leastconductive water attribute species used in USP 22 to be chloride at 0.47 ppm and 1.01 µS/cm at the standardtemperature. When actual measurements of conductivity are evaluated at temperatures above the standard25°C, the least conductive water attributes between pH 5 and 7, changes to favor ammonia at the maximum
acceptable concentration of 0.3 ppm, using chloride as the electro-neutrality-balancing counter ion. Thus bymaintaining a water conductivity less than the value corresponding to the least conductive water attributesbetween pH 5 and 7 at the specific water temperature, failure of any single Ionic chemical test in USP 22 isprecluded.
11.1.1 USP Three Stage Conductivity Levels
Stage 1 Primarily intended as an in-line test
Measure water conductivity and temperature using a non-compensated conductivity sensor using a suitablecontainer or an in-line measurement. See USP Stage 1 table.
Stage 2 An off-line test using a “grab” sample
If the sample fails the Stage 1 test, adjust the temperature to 25°C and stir. If the sample stabilizes to theminimum conductivity value listed in the Stage 3 table, the water meets the requirements.
Stage 3 Additional test to account for the variation in conductivity with respect to alkalinity
Measure pH within 5 minutes of the Stage 2 conductivity reading after increasing its ionic strength to allow apH reading, using saturated potassium chloride solution at 3%. See “Stage 3 Conductivity Levels at 25°C”chart.
11.1.2 Derivation of USP Stage 1 Conductivity Levels
The following chart shows conductivity levels for temperatures between 0°C and 100°C for the chloride-ammonia model at the least conductive water attributes between pH 5 and 7.
Due to the difficulty experienced in accurately measuring pH of high purity water, USP have selected theminimum conductivity levels, occurring at pH 5 and 7, to limit the water quality at each temperature increment.These levels are listed in the table following the chart.
DISCLAIMER:The Water and Steam Systems Appendix contains material considered “informational” which, al-though necessary, would have been detrimental to the clarity of the dedicated chapter. The Appendixhas not been reviewed by and therefore is not endorsed by the FDA.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 168/250
144
APPENDIX
11.1.3 In-Line (Stage 1) Conductivity Limits for Temperatures 0° to 100°C
Temperature Range 0 TO 4.9°C 5 to 9.9°C 10 to 14.9°C 15 to 19.9°C 20 to 24.9°C 25 to 29.9°C
Maximum In-Line (Stage 1) 0.6 0.8 0.9 1.0 1.1 1.3
Conductivity (µS/cm)
Temperature Range 30 to 34.9°C 35 to 39.9°C 40 to 44.9°C 45 to 49.9°C 50 to 54.9°C 55 to 59.9°C
Maximum In-Line (Stage 1) 1.4 1.5 1.7 1.8 1.9 2.1
Conductivity (µS/cm)
Temperature Range 60 to 64.9°C 65 to 69.9°C 70 to 74.9°C 75 to 94.9°C 95 to 99.9°C 100°C
Maximum In-Line (Stage 1) 2.2 2.4 2.5 2.7 2.9 3.1
Conductivity (µS/cm)
Figure 11-1 Conductivity Levels for Chloride-Ammonia Model
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 169/250
145
APPENDIX
11.1.4 Derivation of Stage 2 and 3 Conductivity Levels
When the least conductive water attributes are plotted for water exposed to atmospheric CO2 at 25°C, thefollowing chart is produced:
Figure 11-2 Derivation of Stage 2 and 3
11.1.5 Total Organic Carbon (TOC) and Requirements for TOC Control
TOC is an indirect measure, as carbon, of organic molecules present in high purity water. A TOC limit wasdetermined by USP to be 0.5 ppm or 500 ppb, based on the results of studies and an industry wide survey ofpharmaceutical water systems.
Special Requirements
Organic contamination may be detected from different sources, and once a commitment is made to monitorusing a sensitive TOC meter, care must be exercised in controlling TOC in numerous materials used tosupport and maintain the water purification systems. These materials may include:
• Replaced Filter Cartridges
• Particles from Valves, Seats, Gaskets, etc.
• Soap and Detergent
• Chemicals used for General Cleaning, etc.
• Regeneration and Sanitization Materials
• Alcohol and other solvents used to clean and sanitize seals and gaskets
• Plastics used for sampling apparatus (beakers, bottles etc.) or other applications that may leach outorganic chemicals into the water samples to be tested.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 170/250
146
APPENDIX
• DI resin, detected after regeneration or when beds are switched. Sources of TOC from ion exchangeresin beds include: unconverted monomers or low molecular weight polymers; decomposition productsand compounds resulting from oxidation or hydrolysis of the organic resins; organics in water used toregenerate and rinse ion exchange beds.
New or replaced DI resin may be recycled five or six times to remove leachables, or resin may be specified as
preconditioned. A TOC rinse down cer tification curve is available from most resin suppliers.
Sample containers for off-line sampling must be scrupulously cleaned of organic residues and care must beexercised in handling the containers to avoid transferring natural skin oils onto the container surfaces.
11.1.5.1 Systems Available for Measuring TOC
Instruments are available for measuring TOC in-line from slip streams and from grab samples manuallyremoved from the water system. Automatic off-line sample introduction systems are available for processinglarge numbers of grab samples. USP have not prevented acceptable technologies from being used, but limitthe methods to the following instruments that are capable of completely oxidizing the organic molecules tocarbon dioxide (CO2), measuring the CO2 levels as carbon, discriminating between Inorganic Carbon (IC)and the CO2 levels generated from the oxidization of the organic molecules, maintaining an equipment limit of
detection of 0.05 mg/l or lower, and periodically demonstrating an equipment “suitability.”
A number of acceptable methods exist for measuring TOC in high purity water and all share the same basicmethodology, the complete oxidation of the organics to CO2 and the measurement of this CO2
Three general approaches, based on the above concept, are used in a variety of commercially availableinstruments which measure Organic Carbon in a water sample by completely oxidizing the organic mol-ecules to carbon dioxide (CO2) and measuring the CO2 levels as carbon. Four common oxidation methodsand four common CO2 measurement methods are used in different combinations in these TOC analyzers.The Total Carbon (TC) result may be expected to include Inorganic carbon resulting from dissolved CO2 andbicarbonate which must be subtracted from the TC to produce the TOC level in the sample. Some TOCanalyzers remove the IC by acidifying the samples and either gas stripping or vacuum degassing the CO2. Inpharmaceutical waters, the IC levels are generally very low and IC removal processes are not usually re-
quired.
11.1.6 TOC Measurements
USP have applied laboratory quality control procedures, common in a laboratory for setting wide rangeequipment for measurement over a specific range. These include: Standardization (Limit Response Test) andSuitability (USP Suitability Test).
These tests are in addition to calibration requirements and in no way replace or compensate for an accept-able calibration program.
Types of TOC Analyzers
Method of Oxidation Method of CO2 Detection Require the Addition ofChemicals or Gases
High temperature combustion NDIR Yes
Heat activated persulfate NDIR Yes
UV activated persulfate NDIR or CO2 selective conductivity Yes
Laboratory Instruments
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 171/250
147
APPENDIX
These instruments require the injection of pure carrier gasses and/or persulfate to achieve more robustoxidation levels allowing the successful analysis of high levels of TOC in unknown water samples.
Method of Oxidation Method of CO2 Detection Require the Addition of
Chemicals or Gases
UV light Direct conductivity or CO2 Noselective conductivity
In-Line Instruments (Simplified TOC Analyzers)
Laboratory Instruments Capable of Operating In-Line and Unattended
Method of Oxidation Method of CO2 Detection Require the Addition ofChemicals or Gases
Heat activated persulfate NDIR Yes
UV activated persulfate NDIR or CO2 selective conductivity Yes
These instruments are designed for TOC measurement of deionized water, with or without CO2. These con-ditions allow the accurate measurement of the IC in the sample. If they are used in water that has significantlevels of other ions, the IC results and therefore the TOC results will be in error. Waters that do not meet theserequirements should use the laboratory type TOCs (in-line or off-line).
These instruments require the injection of pure carrier gasses and/or persulfate to achieve more robustoxidation levels allowing the successful analysis of high levels of TOC in unknown water samples.
11.1.6.1 Typical Organic Oxidation Methods used in Commercial TOC Analyzers
The four most common oxidation methods for TOC are: high temperature combustion, thermally activatedpersulfate, UV activated persulfate, and UV light only.
The “high temperature combustion” (>500°C) oxidation method is rapid and can easily oxidize large par-ticles, but requires a source of compressed oxygen or air as a carrier gas for the sample.
The “thermally activated persulfate” method uses heat to activate the persulfate to form highly oxidativesulfate and hydroxide radicals which then react with the carbon in the organics to produce CO2.
The “UV activated persulfate” method uses ultraviolet light (<280 nm) to activate the persulfate to form thehighly reactive sulfate and hydroxide chemical oxidizing radicals. In both of these persulfate methods, thepersulfate is a source of oxygen and can be added at higher levels to allow the complete oxidation of higherconcentrations of organics. TOC levels (without sample dilution) of 50 to 100 ppm can be measured. Eachpersulfate method can easily oxidize macromolecules and biomolecules to CO2, but can have difficulty oxidiz-ing large particles (>30 µm) in reasonable periods of time.
The “UV light only” method uses short wave ultraviolet light (<195 nm) to activate dissolved oxygen or waterto produce powerful oxidizing agents such as the hydroxide chemical radical which react with the carbon inthe organic molecules to form CO2. The “UV light only” method does not require the addition of a chemicalreagent, but the upper levels of TOC are limited. With this method, the short wave UV light will activate
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 172/250
148
APPENDIX
dissolved oxygen free water molecules to completely oxidize organics at TOC levels of up to 250 ppb, and inthe presence of oxygen, saturated water (at 25°C) will completely oxidize organics at TOC levels of up to1500 to 2000 ppb (1.5 to 2.0 ppm).
11.1.6.2 Typical CO 2 Detection Methods used in Commercial TOC Analyzers
There are four CO2 detection methods currently used in commercial TOC analyzers: Non-Dispersive Infrared(NDIR), CO2 selective conductivity, direct conductivity, and differential conductivity.
The “NDIR” method measures the infrared absorption of CO2 in the gas phase. This method requires theCO2 in the water sample to be removed and transferred to the IR absorption cell by a carrier gas stream. The“NDIR” absorption detectors also measure water vapor; therefore, it must be removed from the CO2 gas priorto measurement. This detector responds quickly, must be calibrated regularly, has a dynamic linear range ofabout 1.5 to 2 orders of magnitude, and a limit of detection of 2 to 10 ppb. It is the most common TOC detectorused in laboratory type analyzers.
The “CO2 Selective Conductivity” method uses a special membrane to selectively diffuse CO2 from thewater sample into a deionized water collector. The CO2 ionizes in the collector water, the temperature andconductivity are measured and the concentration of CO2 calculated. This detector is slower responding than
an NDIR, but has excellent long term calibration stability (typically six to 12 months), a linear dynamic rangeof five to six orders of magnitude and a limit of detection of 0.05 ppb.
The “Direct Conductivity” CO2 detection method measures the conductivity of the sample water directly. Toaccurately measure CO2 with this method, the water sample must be composed of only CO 2, deionizedwater, OH-, HCO3
- and H+(deionized water in equilibrium with CO2). When this is true, the CO2 concentrationcan be calculated from the conductivity and the temperature measurements of the sample water. If other ionsare present, the accuracy decreases with increasing concentration of these other ions. This method can beapplied to the sample water before oxidation to measure the level of IC. The same method can be appliedafter the oxidation of the sample to measure TC. Although not as specific or selective to CO2 measurementsas the prior detectors, it is fast, has excellent calibration stability (typically six to twelve months), and has alow limit of detection (about 0.05 ppb).
The “Differential Conductivity” CO2 detection method produces a differential conductivity signal (pre andpost oxidation). The analyzers using this method are designed to only partially oxidize the organics to thevery conductive organic acids stage (not completely to CO2) and measure organic acids not CO2. The ana-lyzer measures the difference between the initial sample water conductivity and the post oxidation sampleconductivity, and relates this to specific organic compound calibration tables.
11.1.7 Limit Response and System Suitability Testing
Apparatus Requirements for Limit Response and System Suitability Tests
a) Reagent Water (r W) complying with the USP definition in section 661 of USP 23, having a TOC level of notmore than 0.25 mg/l.
b) Standard Solution (r S ) containing 1.19 mg of sucrose (produced from r W and USP Sucrose RS dried at105° C for 4 hours)
c) System Suitability Solution (r SS) produced from r W and USP 1,4-Benzoquinone RS (containing 0. 5 mg/l ofcarbon)
Instrument Standardization (Limit Response Test)
USP require that TOC instrument must be standardized by performing a test on each water sample or “testsolution.”
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 173/250
149
APPENDIX
The test solution meets the requirement if the response for the standard solution less the response for thereagent water used to produce the standard solution is greater than the test solution response, or:
r t < ( r S - r W ).
For in-line testing, standardization (determination of the limit response ) may not be practically performed with
each TOC reading; therefore, standardization (determination of the limit response ) once per operating day (orless frequent ) may be appropriate. Support data generated from standardizations performed over a suitableperiod could be used to justify modifications to the testing frequency, based on system reliability, repeatabilityand stability, i.e., TOC levels may be maintained, nominally, far below the limit response over a suitableperiod, thus decreasing the risk in supporting less frequent standardization testing.
Standardization (Limit Response) Testing for In-Line Instruments
In-line instruments must be disconnected from the water system to conduct Standardization (Limit Response)tests. At once per day, these tests represent a considerable investment.
TOC meters with built in provisions for automating these tests and minimizing materials have the potential forconsiderable savings in both time and expense.
Instrument Suitability or Response Efficiency
An instrument is “suitable” if its response efficiency is within ±15% of the theoretical response.
Suitability tests may only be performed off-line, or with the in-line instrument disconnected.
TOC measuring equipment must be tested for suitability “periodically” which may range from, each watersample test for off-line lab testing, to once per month for in-line testing, depending upon the experience withthe system and the data produced from suitability tests.
The Instrument Suitability Response Efficiency is equal to 100 times the ratio of the Suitability solution re-sponse and the Standard solution response, or:
100[(r SS - r W)/(r S - r W)] = 100% ± 15%.
Suitability Testing for In-Line Instruments
In-Line instruments also must be disconnected from the water system to conduct Suitability tests. Suitabilitytests may be conducted together with the standardization tests.
TOC meters with built in provisions for automating both the standardization and suitability tests have thepotential for further savings.
Prepackaged traceable and certified standards are available from some manufacturers for both standardiza-tion and suitability testing.
Special Requirements
Organic contamination may be detected from different sources. Once a commitment is made to monitor usinga sensitive TOC meter, care must be exercised in controlling TOC in numerous materials used to support andmaintain the water purification systems. These materials may include:
• Replaced Filter Cartridges
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 174/250
150
APPENDIX
• Particles from Valves, Seats, Gaskets, etc.
• Soap and Detergent
• Chemicals used for General Cleaning, etc.
• Regeneration and Sanitization Materials
• Alcohol and other solvents used to clean and sanitize seals and gaskets
• Plastics used for sampling apparatus (beakers, bottles, etc.) or other applications that may leach outorganic chemicals into the water samples to be tested
• DI resin detected after regeneration or when beds are switched. Sources of TOC from ion exchange resinbeds include: unconverted monomers or low molecular weight polymers; decomposition products andcompounds resulting from oxidation or hydrolysis of the organic resins; organics in water used to regen-erate and rinse ion exchange beds.
New or replaced DI resin may be recycled five or six times to remove leachables, or resin may be specified as
preconditioned. A TOC rinse down cer tification curve is available from most resin suppliers.
Sample containers for off-line sampling must be scrupulously cleaned of organic residues and care must beexercised in handling the containers to avoid transferring natural skin oils onto the container surfaces.
Calibration of TOC Meters
Calibration should be performed using solutions of a known carbon content, covering the normal range of theinstrument. The frequency of calibration should ensure that the levels of accuracy are maintained. Refer tomanufacturer for specific recommendations on both method and frequency.
Calibration should be performed independent, and in addition to, Standardization and Suitability testing.
Dual Purpose TOC/Conductivity Meters
A number of TOC meters designed for in-line applications include conductivity measurements that are claimedto be in accordance with USP. These instruments use the conductivity method of CO2 detection and utilize theconductivity reading taken prior to water processing. To use this dual purpose instrument to measure compendialwater conductivity, it is important to be able to calibrate the conductivity cell separately and ensure that itcontains a cell constant adjustable and maintainable to the limits defined; an accurate temperature measure-ment being included since most modern conductivity instruments monitor temperature; a conductivity andresolution accuracy as defined and a conductivity calibration which must be accomplished by replacing theconductivity cell with a NIST-traceable precision resistor, accurate to ±0.1% or by an equivalently accurateadjustable resistance device.
11.1.8 USP Microbial and Endotoxin Testing
Microbial and Endotoxins are traditionally sampled at the points of use in a water system. See Chapter 2,Section 2.3 USP Regulated Water Quality.
11.1.9 USP pH Testing
See Chapter 2, Section 2.3 USP Regulated Water Quality (pH testing to support in-line conductivity testingwas eliminated in May 1998.)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 175/250
151
APPENDIX
11.1.9.1 Calibration of pH Instruments
See Chapter 2, Section 2.3 USP Regulated Water Quality.
11.2 EUROPEAN PERSPECTIVE
11.2.1 Regulatory Structure
The pharmaceutical industry within the nations that make up the European Union (EU) is regulated by theEuropean Agency for the Evaluation of Medicinal Products (EMEA), based in London. This Agency is fundedby the EU under the authority of the European Commission (EC), the EU’s policy making body. EMEA isestablished as an autonomous body to promote public health and free circulation of pharmaceuticals. It ischarged with coordinating the resources of the national drug regulatory authorities and has three majorfunctions:
• Centralized drug approval, granting marketing authorization across Europe. This procedure is mandatoryfor biotech products and optional for other drugs. The EMEA also coordinates the mutual recognition ofmarketing authorization among the Member States when a decentralized method (application to one or
more selected Member States) is used.
• Supervision of continued use of medicinal products, using the resources of the national agencies, includ-ing site inspections and monitoring of adverse effects (“pharmacovigilance”).
• Coordination of European and internal harmonization of guidance and regulations, particularly within theframework of the International Conference of Harmonization.
Within the Member States, each national has established a national agency. Examples are the MedicinesControl Agency (MCA) in the United Kingdom and the Bundeinstitut fur Arzneimittel and Medizinprodukte(BfArM) in Germany. These are responsible for national registrations, inspections, and enforcement of regu-lations within their national boundaries.
Another important body with membership from several nations both within and without the EU, is the Pharma-ceutical Inspection Convention (PIC), which has been superseded by the Pharmaceutical Inspection Coop-eration Scheme (PICS). These bodies provide entirely non-binding guidance to member nations, most ofwhich are harmonized with the EU.
11.2.2 Regulations Governing Water and Steam
The EC recognizes the European Pharmacopoeia (EP) as the source of compendial water standards forPurified Water and Water for Injection. The use of these waters parallels that in the United States. Nationsoutside the EU may maintain national compendia, which differ in some parts from the EP.
The EP does not regulate Clean Steam. In Europe, the requirements for the chemical quality of steam to beused in the sterilization of medicinal products in autoclaves are governed by European Standard EN 285. TheEU cGMP Annex for Sterile Products requires that steam be “of suitable quality and does not contain addi-tives which could cause contamination of product or equipment.” Although similar in concept to US require-ments, interpretation of this standard by European investigators may be quite different. Insight into interpre-tation by inspectors may be found in important guidance on the quality of steam found in Health TechnicalMemorandum HTM-2010, published by the United Kingdom National Health Service (NHS Estates). [Editor’sNote: criteria described in HTM-2010 are of major concern in the sterilization of porous autoclave loads andpresent less of an issue in the sterilization of product in vials, hard goods, or equipment.] In addition, UKHTM-2031 deals directly with the quality of steam to be used in steril izers.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 176/250
152
APPENDIX
11.2.3 Water Issues
• While USP requires that the starting point for compendial waters must meet the national drinking waterstandard, the EP requires only that the history of the starting water be documented.
• The EP does allow the production of Water for Injection by distillation only. There are some pharmacopoeias,
notably among the former Soviet bloc, which allow WFI produced by reverse osmosis.
• The EP requires sample sizes of 100 ml for purified water and 500 ml for WFI. The USP is not definitiveon these sample sizes.
• EP has not accepted the substitution of conductivity/resistivity for the traditional wet chemistry analysesof ionic contaminants. There is some early discussion about adopting a conductivity test, but it is likelythat this may be quite different from the USP test.
• The EP has not added the test for TOC.
• EP continues to require the test for pH as an independent test. This is likely to remain, even in light ofpotential changes.
• European inspectors seem to be sensitive to the drainability of systems through points of use. Additionallow point drains are acceptable only if necessary. Guidance in this area remains unclear.
11.2.4 Steam Issues
• HTM-2010 emphasizes the dryness requirements for steam. If steam is too wet, it may cause dampnessin porous loads (e.g., hospital gowns). If it is “too dry,” it may become superheated upon expansion intothe chamber, reducing the sterilizing effectiveness. [Editor’s Note: considering the use of modern, well-controlled steam generators which produce saturated steam, and the large mass made up by the steril-izer and the load, excess dryness is seldom a problem. Excess wetness is controlled by proper insulationof steam lines and judicious use of steam traps to remove condensate.]
Standard EN-285 requires steam for use in sterilizers to have a dryness value of not less than 0.9 (0.95for metal loads). In practice, dryness between 0.9 and 1.0 will not present a problem if the final pressurereduction into the chamber is around 2:1. A method of testing dryness may be found in HTM-2010, part3.
• Superheat is also an issue of concern. Superheated steam is not as effective a sterilizing medium as issaturated steam. It may result from adiabatic expansion (as across a control valve) an exothermic reac-tion resultant from the re-hydration of a hygroscopic load, or the application of jacket heat above thesaturation temperature of the internal steam. A specification and test method for superheat may be foundin HTM-2010. [Editor’s Note: again, considering the production of saturated steam by modern genera-tors, the control of jacket heat, and the condensate resulting from loss of heat to the steril izer and load,superheat is an uncommon problem.]
• The third area of concern which European inspectors have focused on is Non-Condensable Gas (NCGs).Major NCGs in steam are air and carbon dioxide. These may result from the failure to degas boiler/ generator feed water. They present a problem in those areas where localized “pockets” of gas are allowedto accumulate, preventing the penetration of steam. This may be effected by preheating the feed water ina vented tank at above 80°C. A test of method may be found in HTM-2010. [Editor’s Note: the presenceof air in a well-mixed sterilizer is common practice and does not reduce the effectiveness of sterilizingsteam.]
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 177/250
153
APPENDIX
11.3 PASSIVATION
11.3.1 Introduction
Pharmaceutical equipment and high purity water systems are designed so that product contact surfaces arenot reactive, additive, or absorptive so the drug product is not adversely altered. High purity water systems
are primarily composed of austenitic stainless steel (SS) materials due to their corrosion resistant and con-taminant free properties. Passivation is performed to maximize the metal’s corrosion resistance. The stain-less steel is sulfuric/nitric/hydroflouric acid pickled at the mill to remove manganese sulfide inclusions, scale,and other impurities or imperfections from the surface of the steel. As the steel is removed from the picklingbath, a thin oxide layer forms immediately over the surface. This oxide layer is what renders the stainless steelpassive and non-reactive to corrosion. Any 300 series chromium steels containing 17% or more chromiumthat has been handled, welded, or worked should be passivated prior to service and suitably cleaned prior topassivation.
Passivation is the method used to fortify the steel surface by strong oxidizing chemicals such as nitric acid.The acid depletes the steel surface of acid soluble species, leaving the highly reactive chromium on thesurface in a compounded oxide form.
11.3.2 Advantages of Passivation
When SS systems are fabricated, the welding process destroys the existing passive film and compromisesthe metal’s ability to ward off the corrosive process. This is particularly applicable in those zones that areeither heat affected or have had residues remain in contact with the metal surface for prolonged periods.Passivating then provides a method to restore the integrity of the metals corrosion resistant surfaces thatwere disturbed. Passivation must be proceeded by a cleaning process.
11.3.3 The Chemical Process
Excessive electron depletion of the upper film and an inadequate supply of oxygen (molecular O2) will ensurethe formation of surface corrosion products. When this occurs, the chromium (Crn+) separates from the sur-face and opens the way for oxidation of the iron (Fe) and nickel (Ni), lower in the metal lattice.
Establishing a passive surface or film on austenitic SS is essential to maximize the corrosion resistance thatthe metals offer. Passive surfaces on these metals occur naturally when exposed to an oxidizing environ-ment. Sources of oxygen include air, aerated water, and other oxidizing atmospheres. Formation of a sub-stantial uniform oxidized corrosion resistant surface or film is the result of passivation.
Besides natural occurring passivation, chemical and electro-chemical processes can be used to obtain ananodic oxide film. Nitric acid solution (HNO3), is an oxidizing acid (depletes electron from the metal surface)which erodes the metal. This initial reaction or oxidation resists further chemical reaction on the metal sur-face. Metals that have such a state are called “passive” and the phenomenon itself is called “passivity.”
The chromium oxide film thickness typically ranges from 0.5-5.0 nm, averaging 2.0-3.0 nm. The chrome toiron ratio measured in atomic percent within the chromium oxide should be at least one with ratios of two ormore being optimal.
11.3.4 Passivation Procedures
Numerous procedures are available for passivating; they share the commonality of consisting of four mainsteps which are:
1) Wash (Solvent Degreasing)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 178/250
154
APPENDIX
2) Water Rinse
3) Acid Wash (Passivation Step)
4) Final Water Rinse
Proper preparation of the metal surface to obtain a uniform non-defective passive film mandates the metalsurface be completely clean and void of any organic or inorganic soils, free iron, metallic contaminants, orcorrosive products.
The First Step (Degreasing) of the procedure is designed to remove dirt, dust, oil, and grease. A water-soluble detergent is used to accomplish this, or a solvent.
The Second Step (Water Rinse) is required to remove dissolved and freed soils and the detergent itself fromthe metal being cleaned.
The Third Step (Acid Wash) is to remove free iron, metallic residues, oxides, and other corrosion productsfrom the surface of the metal. By removing these soils from the metal surface and providing an oxidizingatmosphere, the passive film is allowed to form and the passivation is accomplished. Inorganic acids are
typically used in this step of the procedure.
The Fourth Step (Final Water Rinse) - The acidic solution is flushed and the system is rinsed until the qualityof the effluent is equal to that of the influent.
The American Society for Testing and Materials, ASTM A 380-96, “Standard Recommended Practice forCleaning and Descaling Stainless Steel Parts, Equipment and Systems,” is an excellent source of informationabout passivation. It includes cleaning and passivation procedures, chemical applications, methodology, andtesting procedures. The standard is valuable in establishing specific passivation and other specialized clean-ing procedures.
Establishing an effective passivation procedure can be obtained by using the following guidelines:
• Start with an accepted or specified procedure. (See the chart on the next page.)
• Obtain weld coupons from the system or have weld coupons made for testing purposes.
• Perform specified procedure along with alternate procedures to offer a choice, meeting specific situa-tions, or requests.
• Confirm the effectiveness of the procedures tested with specified field and/or laboratory testing.
• This process for confirming the effectiveness of a specified procedure or qualifying alternative proce-dures should be included in the passivation documentation being submitted as part of the final validationpackage.
11.3.5 Passivation Chemical Alternatives
Nitric acid, a strong oxidizing acid, is the most common acid specified for passivation. Besides its ability toproduce a free iron surface, the acid supplies the oxidizing atmosphere needed for passivation to occur.Because nitric acid is a corrosive chemical, extreme care must be used with handling, storage, and use.Federal Specification QQ-P-35C (1988) is an excellent reference for obtaining guidelines when using nitricacid on a variety of stainless steel alloys.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 179/250
155
APPENDIX
Although nitric acid has traditionally been the preferred passivating acid, the trend in use of passivatingsolutions is to reduce chemical aggressiveness and to make safety, cost, and the environmental impact of thewaste solution effluents a consideration.
Citric acid and ammonium citrate (ammoniated citric acid) are gaining popularity as alternatives to using nitricacid. The safety these chemicals offer the personnel and the work environment are desirable qualities. The
ASTM Standard A 380 (1996) refers to these acids as cleaning acids, not passivating acids. This distinctionhas probably been made because the acids are not oxidizers as is nitric acid. The standard states that thecitric acid-sodium nitrate treatment is the least hazardous for removal of free iron and other metallic contami-nation and light surface contamination. To achieve a true oxidation chelating agents in conjunction with citricacid and ammonium citrate has recently been introduced to the pharmaceutical/biotech industry.
Phosphoric acid is a weak oxidizing acid sometimes specified in passivation procedures; however, there is noformal documentation referencing the use of phosphoric acid as a passivating acid.
Chelants, otherwise known as sequestering agents or co-ordination compounds, which include all the stan-dard water softening compounds such as Sodium tri-polyphosphate (STPP), Nitrilotriacetic acid (NTA), andEthylene Diamine Tetra Acetic acid (EDTA) may be compounded into acid passivation solutions to enhancemetal ion extraction.
Orbital welding in conjunction with the increased use of electropolished tubing decreases the aggressivenessrequired of the passivating acids during the initial passivation. Decreasing acid contact time, temperature,and/or concentration accommodates the quality of the welds and already passive surface of the electropolishedstainless steel.
11.3.6 Chemical Application Methods
Passivation can be accomplished using a variety of applications. Among these are:
When detergent washing, agitation or impingement provides the best results. During the acid wash step,chemical contact is usually sufficient. Recirculation is the preferred application method for performing passi-vation procedures. Recirculating allows flow rate criteria, usually specified at 5 feet per second (1.5 m/sec), tobe achieved. Meeting flow rate requirements of a procedure should not be confused with particle removal.Many people assume when high flow rates are used that particle removal will be achieved. This is not true.Particle removal is achieved by including the total linear feet of the system into the appropriate mathematicalequation. A recirculating water system of 1000 feet (300 meters) with a consistent tube diameter wouldrequire as much as 25 hours of filtered recirculation time for total particle removal.
Circulation Recirculating through distributionsystems
One Way Intermittent Flow Large non-recirculating Long one way pipe runs systemsdistribution
Spraying Tank interiors
Tank Immersion Numerous small parts Prefabricated tubing
Swabbing/Wiping Isolated Areas/Tank/Equipment Equipment that does not allowExteriors spraying or other applications
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 180/250
156
APPENDIX
11.3.7 Tests for Cleanliness and Passivity
There are several tests available to determine an acceptable level of cleanliness. Should confirmation ofcleanliness be required prior to continuing with the passivation procedure, the water break free surface test,wipe test, or ultraviolet light testing are just a few of the tests that could be performed. These tests are forgross cleanliness inspections as stated in the ASTM Standard A 380 (1996).
Once the passivation procedure is completed, a test method should be used to confirm or establish confi-dence that the passivation procedure has been successful. One inexpensive method is the Ferroxyl Test forfree iron as set forth in the A380 (1996). The test is used to detect surface iron contamination, i.e., iron saltresidue from pickling, iron tool scratches on the stainless steel surface, iron deposits at weld areas, and ironoxides. The testing solution is applied to the surface being tested. A blue stain appearing within 15 seconds ofapplication indicates presence of free iron.
Testing for a passive surface is usually accomplished by looking for traces of free iron on the metal surface.The assumption is made that if there is no detectable free iron, the metal surface is clean enough for auniform oxide film to develop. Another excellent source for specific testing methods is the Military Standard753B (1985). Both Standards discuss specific tests for detecting free iron. They include Water Immersion/ Water Wetting and Drying Test, High Humidity Test, Copper Sulfate Test, and Ferroxyl/Potassium Ferricya-
nide-Nitric Acid Test.
Direct testing for a passive surface can be accomplished by X-ray Photoelectron Spectroscopy (XPS) testingwhich is used to measure the oxidation states of elements found on the metal surface. Another direct, de-structive testing method is Auger Electron Spectroscopy (AES) which measures the elemental chrome/ironratio on the metal surface and sub-surface with depth profiling. The direct testing methods for passivity supplydetailed information about the oxide film itself rather than indirect observations. XPS or AES testing offersdirect evidence as to whether the passivation procedure being used is effective or not. These methods oftesting are more costly than the other above mentioned tests and are ideal for use with weld coupons todetermine the effectiveness of the passivation procedure for the system.
11.3.8 Modified Passivation Procedures
A passivation procedure can be modified to deal with a variety of soils, surface finishes and weld areaconditions. Adjusting contact times and solution’s temperature and concentration would be the simplest wayto modify a specific procedure. Sometimes detergent wash or acid wash chemicals are changed or modifiedwith additives to remove certain soils. For example, when removing rouge, solutions containing sodium hy-drosulfite can be substituted for the detergent wash step of the procedure. Citric and Phosphoric Acid alsocould be used as they do have some ability to remove light rouging. Another example would be the use ofHydrofluoric Acid, or more specifically, Ammonium Bifluoride to remove silica scale. The descaling step andassociated rinse would necessitate additional steps being added to the standard procedure.
It is important when developing a passivation procedure, that laboratory testing is performed to determine theeffectiveness of your procedure. Without preliminary laboratory testing, an educated guess would have to bemade and the results may not prove satisfactory.
Below is a guide that can be used for passivating and de-rouging stainless steel components, piping, andequipment. The chart has some possible options for determination of the contamination and a course ofaction.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 181/250
157
APPENDIX
11.3.9 Contamination Analysis
Method 1 Filtration of 1 liter sample through a 0.2-0.5 µm filter and inspect.
Method 2 Quantitative analysis of the specified metals and organic compounds with wet chemistry tech-niques or as available.
Method 3 SEM or Auger Electron microprobe/spectroscopy for analysis of surface chemistry and contami-
nation.
11.3.10 Cleaning and Passivation Method
Method 1 Clean surface with aqueous cleaning solution, apply passivation paste to surface, rinse surfacewith DI water until traces of chemicals are removed.
Method 2 Circulate cleaning solutions through piping or vessels by circulation method. Circulate cleaningsolutions as required by procedure. Circulate passivation solution as per recommended condi-tions. Rinse surfaces once through with DI water until conductivity of inlet and outlet fluids arewithin tolerances.
Method 3 Spray cleaning and passivation solutions onto surfaces of vessels, containers, and equipmentas per recommended conditions. Rinse surfaces for minimum of 30 minutes per each rinsestage, and perform triple rinse.
Method 4 Soak components or equipment items in treating solutions or tanks as per recommended condi-tions. The minimum soak time per each solution is two hours. Process requires cleaning, passi-vation, and rinsing as a minimum. The cleaning system should include circulation, filtration, andheating.
Condition/Status Contamination Cleaning & System ProcedureAnalytical Method Passivation Chemistry
Method
New Component N/A 2,4 3 2Electropolished
Component Newly Welded N/A 1,3,4 1,2,3 1,2
New System - Tubing N/A 2 2,3,4 2
Component/System Discolored 1 1,2,3,4 1,2,3,4,6 2(Gold Color)
Component/System Discolored 1,2 2,3 4,5,6 3(Brown, Red/Brown Color)
Component/System Discolored 2,3 2,3 4,5,6 3(Black, Blue/Black Color)
Cleaning and Passivation
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 182/250
158
APPENDIX
11.3.11 System Chemistry
Chemistry 1 Nitric acid passivation is performed at Ambient Temperature for 30 to 60 minutes and at 50-60°C for 20 to 40 minutes.
Chemistry 2 Alkaline degreasing is performed with detergents (phosphates, sodium hydroxide, and po-
tassium hydroxide), pH buffers, and surfactant. The process will remove organic films andparticulate debris from the surface of the stainless steel. Utilize approximately 1.0-2.0%detergent, 0.2-0.5% buffer and 0.01-0.2% surfactant.
Chemistry 3 Citric acid/chelant passivation is performed with chelants, reducing agents, surfactants, andpH buffers. These systems are proprietary processes and the exact chemistry and percent-ages are not available. The chelant systems are able to remove most metal contaminationfrom the surface including iron, manganese, aluminum, and copper. The systems include3.0-5.0% Citric acid and a variety of chelants, reducing agents, pH buffers, and surfactants.
Chemistry 4 Mineral acid cleaning and passivation can be performed for iron oxide removal or passiva-tion. Typical mineral acids include phosphoric, sulfuric or sulfamic acid. These acids can beutilized at 3.0-10.0% concentrations and at a variety of temperatures. Sulfuric acid is not
typically used due to its highly hazardous nature.
Chemistry 5 Intensified acid/chelant systems are utilized for removal of high temperature iron oxide films,silica scales, and organic/carbon films. These systems are a citric based solution with addi-tional organic acids, strong reducing agents, and acid chelants. These systems can usefluorides for silica removal. After strong acid cleaning in a reducing environment, it is recom-mended that an oxidizing flush be used to ensure oxidation at the surface, removal of or-ganic films, and sanitization of the system.
Chemistry 6 Sodium Hydrosulfite, a strong reducing agent, typically used at 5% by weight at 120 to 160°Ffor two to four hours.
11.3.12 Procedures
11.3.12.1 Procedure 1
Clean surface of organic film and other debris.
a) Rinse surface with DI water.
b) Apply gelled acid onto surface at ambient temperature.
c) Brush passivating agent on surface very 15 minutes, maintain a wet surface.
d) After one hour minimum, brush surface with sodium bicarbonate solution until all reaction ceases.
e) Rinse surface with DI water until all traces of chemicals are removed.
11.3.12.2 Procedure 2
a) Fill system with DI water and perform leak test with circulation pump.
b) Circulate for a minimum of one to two hours with alkaline degrease stages and heat to 60-80°C withfiltration.
c) Drain and rinse with DI water.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 183/250
159
APPENDIX
d) Circulate for a minimum of one to two hours with passivating acid solution and heat to 60 - 80°C.
e) Drain and rinse with DI water.
11.3.12.3 Procedure 3
a) Fill system with DI water and perform leak test with circulation pump.
b) Circulate for a minimum of two hours with alkaline degrease stages and heat to 60 - 80°C with filtration.
c) Drain and rinse with DI water.
d) Circulate for a minimum of eight hours with intensified passivating acid solution and heat to 60 - 80°C.
e) Drain and rinse with DI water.
f) Flush with oxidizing/sanitization solution.
g) Drain and rinse with DI water.
11.3.13 Rouging
Rouging is seen in many water systems, usually high temperature (80°C) distilled water and clean steamsystems. Rouge is not limited to storage and distribution systems; it also can be found in distillation and cleansteam generating equipment. The main constituent of the rouge film is ferric oxide, but it can contain iron,chromium, and nickel of different forms. From Auger Electron Spectroscopy, it has been found that the outerlayer of a rouge film is carbon rich, and the underlying region is iron and oxygen rich, probably iron oxide.Over time, the film uniformly distributes itself throughout the system. The exact mechanism of the rougeformation and proliferation is unknown. Because the phenomenon occurs in systems that offer the mostcorrosive environment, it is thought that low molecular weight ions of the stainless steel, such as iron, aredrawn to the metal surface or are dissolved and uniformly re-deposited throughout the system. Others feelthe rouge is an external contaminant probably colloidal in nature that once in the system, uniformly deposits
itself.
Rouging would seem to be very site (facility) specific because of the variety in appearance and texture.Rouge can be observed in a variety of colors including; orange, light-red, red, reddish-brown, purple, blue,gray, and black. It can be a very loose film, dust like in appearance and texture that can be readily wiped offto a tight pertinacious film that requires scraping with a sharp instrument to be removed. In addition to thediversification already discussed, rouge can be multi-layered exhibiting different colors and textures. Tradi-tionally the red rouges are most common in high purity high temperature water systems, while the blue/blackrouges are typically found in clean steam systems.
Evidence of the migration of rouging in distribution systems can be demonstrated by monitoring a systemover a period of time. Key places to look for rouging are still and clean steam generator discharge lines, tankwater/vapor interface, pump heads, Teflon ® diaphragms on diaphragm valves, interior surface of tank sprayball, and heat effected area of welds. Rouge deposition seems to have an affinity for Teflon ® and would be oneof the first places to look for signs of system rouging.
In some cases, the rouging appears as quickly as a month or two after system start up. In other cases, it isseveral years before rouging is observed. In either case rouging is an industry wide phenomena. In a specificcase, a facility cold WFI system would re-rouge within a week of being derouged and passivated. The systemwas derouged and passivated a total of three times. Each time, within a week, the system was totally rougedagain. The specific cause was never determined.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 184/250
160
APPENDIX
The presence of rouge in high purity water systems has not been proven to effect water quality. The FDA hasno written position addressing rouging, its existence, or presence in high temperature high purity water andclean steam systems. Their criterion has and remains to be in meeting established USP standards for waterquality. There is some fear that as the unwanted film develops, it might eventually slough off and be dispersedthroughout the system. This, in fact, does occur and is manifested in systems with filtered use points. Filtersbecome discolored with the typical reddish-brown rouge color.
Phosphoric, citric, oxalic acids, and ammonium citrates are used depending on the severity of the problem.Oxalic acid solutions are used for the worst cases of rouge. Passivation with nitric acid is required after anoxalic acid rinse.
11.3.14 Preparing Systems for Passivation
Hydrostatic pressure testing is the first test in preparing a system for passivation. All newly constructed ormodified systems require pressure testing prior to implementing any chemical procedure. The second checkprior to passivation is to confirm the compatibility of the system, its components, and the passivating solu-tions. This would include in-line instrumentation, flow meters, regulating valves ultraviolet lights, pumps, pumpseals, filter membranes, gasket and seal materials, and other specialized in-line devices. The manufactureror supplier should be consulted to determine whether their equipment is compatible with passivating solu-
tions. Items that are not compatible should be removed from the system and replaced with a blank, valve,spool piece, or temporary jumper hose. In some cases with in-line instrumentation, chemical incompatibilitymay lie in the effect it has on instrument calibration. Incompatible components should be processed indepen-dent of the main system.
Once the system/chemical compatibility has been established, the system to be passivated should be iso-lated from existing systems, process equipment, utility tie-ins, etc. In most cases, in-line heat exchangers(excluding plate and frame design) and small filter housings (filter elements removed) are left in place andflowed through. This is acceptable as long as the ability to vent and drain is available.
Isolated equipment that requires passivation should be handled independently from the main system unless,by agreement, it is left in-line and flowed through. All isolation points must be valved to avoid forming deadlegs in the system being passivated.
Elimination of all dead legs is critical to ensure chemical contact and complete rinsing.
High point vents and low point drains are desirable for complete filling and draining of systems. In distributionsystems where high point vents are not installed, high velocity flow and flow restriction techniques can beused to ensure complete filling of the system.
After the system has been pressure tested, compatibility has been confirmed, the system isolated and deadlegs valved, consideration must be given to automated controls that govern the system.
Are all the automatic valves operational?
Will valve alignments atypical of normal system operation be permitted?
Will in-line temperature sensors open diverter valves if unusual temperatures are detected?
Can the desired flow path safely and effectively be achieved.
Passivation contractors generally supply temporary equipment such as circulating vessels, pumps, heatexchangers, flow meters, filters, hoses, spray heads, fittings, specialized adapters or transition fittings, andneutralization vessels. All this equipment should be inspected to assure it meets the requirements for itsintended use.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 185/250
161
APPENDIX
11.3.15 Disposal of Passivation Chemicals
Disposal of waste solutions is an important issue. The chemicals discussed for cleaning and passivating areall water-soluble and can easily be neutralized. Except for heavy metals dissolved in the acid wash solution,the only criterion that makes the waste solutions hazardous is having the pH outside the range of 2 to 12.5. Itis the heavy metals contained in the waste effluent that can cause an environmental or disposal problem. Of
the 13 priority pollutant metals tested for, two are found in elevated levels in passivation waste effluent waters.The two heavy metals are chromium and nickel.
Fluids discharged must meet the site’s discharge temperature requirements.
There are three options for dealing with waste solutions generated when passivating:
• They can be put into chemical drains. This can only be done where compatible drain and treatmentsystems are available.
• Neutralize waste solutions in contractor-supplied equipment and discharge through chemical drain tosite treatment system.
• Off-site disposal is the final option. It is the most costly form of disposal.
Should a site waste treatment system not be available, permission could be obtained from the municipal orprivate sewer authority to put neutralized waste solutions to sanitary drains. Under no conditions should anywaste solutions in any form be permitted to enter storm sewer systems
You will, however, receive documentation confirming proper disposal of waste solutions. Documentation wouldinclude a bill of lading or hazardous waste manifest and receipts from the state certified treatment facilitywhere the waste solutions are being transported and treated. When off-site disposal is being used, it isimportant to verify the credentials of the hauler and final destination site before utilizing their services.
Ultimately, disposing of waste solutions in a proper and legal manner is the responsibility of all involvedparties. The owner of the property where the waste solutions are generated, contractors, subcontractors
involved with the use of the chemicals, haulers, and the final waste treatment facility would all have someliability for proper disposal of waste solutions.
11.3.16 Documentation
Complete and detailed documentation should be kept as the procedure is being performed. Specifics onchemical concentrations, temperatures, contact time, quality of rinse water supply, and effluent sample read-ings should all be recorded.
Some contractors use job log sheets to record chronological job data including specifics from the time thecontractor arrives on-site until the time he leaves. In addition to job log sheets, passivation log sheets shouldbe completed. Detailed information, as discussed above, can be plugged into a “fill in the blank” form suppliedby passivation contractor, validation firm, or owner. No matter how the information is recorded, the importantthing is that detailed and accurate documentation is kept. The following information can be submitted to theowner and become incorporated into the final validation documentation:
• Passivation Procedure
• Miscellaneous Pertinent Information
• Procedure Development Data
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 186/250
162
APPENDIX
• Testing Procedure and Equipment
• Calibration Documentation
• Passivation Log Sheets
• Chemical Batch Record Information
• Marked up system drawings, completed using point check list or line identification list.
11.4 PRETREATMENT PROCESSES
11.4.1 Turbidity and Particulates
11.4.1.1 Definitions
Particulates are insoluble suspended materials present in the water. Concentrations are measured in mg/l.Sources of particulates are dust, pollen, silica, insoluble minerals, and corrosion products.
Turbidity is a cloudy appearance in water caused by the presence of suspended and colloidal materials.Rather than a physical property, it is an optical property based on the amount of light reflected by the sus-pended particles and is measured in Nephelometric Turbidity Unit (NTU). The EPA limit for turbidity in drinkingwater is 1 NTU. Turbidity cannot be related to particulates since it is affected more by particle size, shape, andcolor rather than concentration. Light colored particles reflect more light than dark colored particles and manysmall particles reflect more light than a few larger particles of equivalent concentration. Removal of particu-lates and turbidity is required to prevent fouling/plugging of final treatment processes using a membrane(RO).
11.4.1.2 Filtration Mechanisms
The principal methods for removal of turbidity and particulates are:
• clarification and the accompanying operations of flocculation, coagulation, and sedimentation
• media filtration including single and multimedia filtration
• barrier filtration included pre-coat filter, surface, and depth media such as cartridges and finer barrierssuch as nanofiltration or ultrafiltration
Factors affecting the removal of turbidity and particulates are:
• particle size and shape relative to the filtration media
• tendency of the particles to adhere to each other or the media which may be enhanced by addition of aflocculating agent or alum (agglomeration)
• surface effects including surface tension, hydrogen bonding, and electrostatics
11.4.1.3 Clarification
Clarification is one method used by municipalities and large water treatment suppliers for removal of particu-lates and turbidity. Addition of alum, lime/FeCl3, or other flocculating agents, and pH adjustment aids thesedimentation and clarification to remove particles larger than 25µm. Flow rates are generally large and cost/
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 187/250
163
APPENDIX
gallon is low. This process is typically not found in the production of purified water because it would beredundant to the treatment typically done by a municipality. The scale of these systems is much larger thanmost purified water systems and clarification is not 100% efficient so that some additional filtration methodwould be required to prevent particulates from being retained by filtration and cause blockage in subsequentpretreatment operations such as ion exchange, carbon beds, and fine barrier filtration.
11.4.1.4 Media Filtration
Media filtration using a depth filter is the most common method of removing particulates from the water inpharmaceutical water systems. It also may have some minor effect on turbidity. Design can be with either asingle size media or multi-sized media in a tank that has the means to support the media. With multi-sizedmedia, the larger media is typically at the top with the main flow direction downward through progressivelyfiner layers of media. The porosity of the media bed selected permits removal of particles down to a size of10-40 µm. Accumulated particulates are removed by a back flush operation based on increase in pressuredrop or time. This back flush in the upward direction also decompresses the filter bed and is followed by arinse to resettle the media and remove fines. This back flush is generally considered a sanitary rather thanchemical waste and is typically 3-10x the design flow rate for a period of about 30 minutes. Following the backflush, a short flush to drain in the direction of the process flow is required to resettle the bed and remove fines.
The filtration media in a depth filter may be sand, anthracite, carbon, or manganese. Sand is the most com-mon because of cost and availability in a wide range of sizes and purities. Anthracite might be used whereleaching of the silica from a sand filter is a problem due to high temperatures or alkalinity. Depth filters usinganthracite often have higher filtration rates over extended runs and require less back washing (and regenera-tion) because of the sharply angular particles rather than the rounded silica particles. A depth filter usingcarbon might be selected if the water has a high loading of organics, or if there is a particular reason tocombine removal or particulates, organics, and chlorine. The carbon is usually a course retention layer underan extended layer of an activated granular carbon such as coconut, lignite, or anthracite. A depth filter usingparticles coated with potassium permanganate or manganese zeolite as the depth media might be selectedfor water having high concentrations of iron or manganese. It also may oxidize sulfur or hydrogen sulfide.Generally, an oxidant, potassium permanganate or chlorine and permanganate is added prior to this filter toconvert the metals to the higher oxidation states that are insoluble. Removal down to levels of 0.03 mg/l ofiron and 0.05 mg/l of manganese are possible.
Microbial growth is a key consideration in any filter, but particularly a depth filter. This occurs because of thelarge surface areas and relatively low velocities. In the case of a carbon filter, the media also is a source ofnutrient. Design of the system should include the presence of a disinfectant such as chlorine or chloramine inthe feed water, an added disinfectant, or the ability to periodically sanitize with a disinfectant or heat. The filterbed also may be designed with constant recycle to ensure continuous flow through the bed to minimizestagnation and growth.
Advantages: filtering material; works well in a chlorinated environment; large capacity at low cost
Disadvantages: filter out only large particles; can be a source of microbiological growth
11.4.1.5 Barrier Filtration
Barrier filtration includes cartridge filtration, pre-coat filtration, ultrafiltration, and nanofiltration. This type offiltration has a “barrier” through which the water must flow. The barrier retains particulates that are removedby changing the barrier (cartridge and pre-coat) or by a purge stream (ultra and nanofiltration).
Barrier filtration is typically not used as the primary method for removal of particulates because of cost of thebarrier, labor, and the frequent need for replacement of the barrier due to the relatively high particulateloading in the water entering pretreatment. It is frequently used as a “final clean-up” after the other pretreat-ment process such as ion exchange or carbon bed filtration, and before going to a final treatment step suchas reverse osmosis.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 188/250
164
APPENDIX
As a “final clean-up,” the barrier is often cartridges that might typically have a nominal particle size retentionof 1-10 µm for just the removal of particulate carry-over from the previous operations. If, however, the objec-tive is to remove insoluble forms of silica and iron and achieve a SDI (silt density index) of less than 5 for feedto an RO system, the absolute particle size retention might be <1 µm and as such are at the lower end ofmicrofiltration or the upper end of ultrafiltration. An absolute rated cartridge filter or a pre-coat filter canachieve this. The later is usually not used unless the high particulate loading offsets the associated labor
cost.
Ultrafiltration (0.1-0.001 µm) and nanofiltration (0.005-0.0005 µm) are methods of barrier filtration. They arenot used as for primary removal of particulates for the reasons noted above, particularly cost and rapidblinding of the membrane. Ultrafiltration is discussed in Section 11.4.4.3E.
Advantages: low initial cost; can vary particle size retention by cartridge selection
Disadvantages: Cartridge filtration is an expensive way to remove solids due to the cost of cartridges andlabor if solids concentration is high; potential for microbial growth in a non-chlorinated envi-ronment.
11.4.2 Hardness and Metals-Ion Exchange
An ion exchange system consists of a tank containing small beads of synthetic resin. The beads are treatedto selectively absorb either cations or anions and exchange these ions based upon their relative activitycompared to the resin. This process of ion exchange will continue until all of the available exchange sites arefilled, at which point the resin is exhausted and must be regenerated by the appropriate chemicals. Forremoval of hardness and metals, a cation exchange system will remove positively charged ions (metals) andexchange them for sodium ions. Ion exchange resins that remove cations or anions and replace them withhydrogen and hydroxyl ions are discussed in Chapters 5 and 6.
The presence of calcium (Ca) and magnesium (Mg) in a water supply is commonly known as “hardness.” It isusually expressed in grains per gallon (gpg). Ion exchange is the principal method of removing hardness fromwater in a pretreatment system. The process of removing hardness is often called “softening.” This is requiredto prevent scale formation in final treatment operations such as RO and distillation.
11.4.2.1 Water Softening
Hardness in a water supply can result in “scale formation,” which is a deposit of minerals left over after thewater has been removed or evaporated. This can be found in boilers, cooling towers, reverse osmosis ma-chines, clean steam generators, and distillation systems.
The function of an ion exchange water softener is removal of scale forming calcium and magnesium ionsfrom hard water. In many cases, other multivalent ions such as soluble iron (ferrous) and ionized silica alsoare removed with softeners.
A standard water softener has four major components: a resin tank, resin, a brine tank, and a valve orcontroller. The softener resin tank contains the treated ion exchange resin - small beads of polystyrene.Capacity depends on volume of the resin bed. The resin beads initially absorb sodium ions during brineregeneration. The resin has a greater affinity for the multi-valence ions such as calcium and magnesium thanit does for sodium. As a result, when hard water is passed through the resin, calcium, magnesium, and othermultivalent ions such as iron and silica adhere to the resin, releasing the sodium ions until equilibrium isreached. The water softener has exchanged its sodium ions for the calcium, magnesium, and iron ions in thewater.
Regeneration is achieved by passing a sodium chloride (NaCl) solution through the resin, exchanging thehardness ions for sodium ions. The resin’s affinity for the hardness ions is overcome by using a highly con-centrated solution of NaCl (brine). The spent brine solution plus the associated water back-flushes and rinses
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 189/250
165
APPENDIX
are waste streams and might typically approximate the nominal throughput for one hour for each regenera-tion cycle.
Water softening is a simple, well-documented ion exchange process. It solves a very common form of watercontamination: hardness. Regeneration with brine is simple and inexpensive and can be automatic with nostrong chemicals required.
Advantages: Industry standard, low cost and effective. Works well in a chlorinated environment for micro-bial control with chlorine having only minor effect on resin life and efficiency. Inherent in theregeneration is flushing of some microbial growth, but this should not be relied on as the solemeans of microbial control.
Disadvantages: salt handling for brine regeneration and disposal of spent brine solution
11.4.2.2 Demineralization/De-ionization
Three multivalent ions (iron, silica, and aluminum) present unusual removal problems.
a) Iron
Iron is a common water contaminant. It is one of the more difficult contaminants to remove because itmay change valence states--that is, change from the water-soluble ferrous state to the insoluble ferricstate.
In solution, ferrous iron behaves like calcium and magnesium; however, when oxygen or an oxidizingagent is introduced, ferrous iron becomes ferric and precipitates, leading to a rusty (red brown) appear-ance in water.
Certain bacteria can further complicate iron problems. Organisms such as Crenothrix, Sphaerotilus, andGallionella use iron as an energy source, eventually forming a rusty, gelatinous sludge that can plug uppiping and equipment, particularly barrier processes such as nanofiltration and reverse osmosis. Oneremoval method for iron in the oxidized state is a replaceable barrier filtration such as a cartridge filter
with an absolute rating of <1 µm. It also can be removed by nanofiltration or reverse osmosis with somemembrane fouling, or with sequestering agent dosing that is removed further downstream in RO.
b) Silica
Like iron, silica may be present in more than one form and is a major problem in some parts of thewestern United States. It may be a soluble ionized species or an insoluble material, sometimes as acolloidal mixture with organics and other metals. The concentration of ionized silica will be reduced by awater softener and insoluble silica forms can be removed by a replaceable barrier filtration with an abso-lute rating of <0.5 µm (ultrafilter). The insoluble silica also can be removed by nanofiltration or reverseosmosis with some membrane fouling or by strong base ion exchange.
c) Aluminum
Like iron and silica, aluminum can exist in multiple valences and its chemistry is complex. It also can bea component of colloid complexes. Its solubility, particularly as hydrated oxide compounds, is a functionof pH. Aluminum may be present in the water either naturally or as a result of the alum treatment used bya municipality as part of coagulation. Aluminum that is present as a colloidal component can be removedby fine barrier filtration. Softening or de-ionization removes aluminum in an ionized form. Aluminum alsocan be removed by reverse osmosis if the pH is <6 or >8. However, between these pH values that areoften common in an RO unit, the hydrated aluminum oxides are only partially (about 80%) rejected by ROand often lead to fouling in an RO. If aluminum is a major problem, softening or de-ionization followed bypH adjustment and then RO may be required for removal.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 190/250
166
APPENDIX
11.4.2.3 Ion Removal
Ion exchange (softening) is often part of the pretreatment process where it is used for removal of hardness.Ion exchange as a total ion removal process (cations are exchanged for hydrogen ions and anions areexchanged for hydroxyl ions) is discussed in Chapters 5 and 6.
11.4.3 Hardness and Metals-Other Processes
This section covers two alternate means (acidification/degasification and nanofiltration) for removing scaleforming components in water destined to be further purified in RO systems.
11.4.3.1 Acidification and Degasification
a) Carbon Dioxide
Use of a pH adjustment step is common in the process design to favor or inhibit the formation of CO2. Asdiscussed in the section on pH (Section 4.11), injecting acid to a pH of approximately 5.5, will maximizethe concentration of CO2 whereas, the addition of a base to a pH of 8.3 will minimize the formation of CO2,converting it to carbonate ion. See Sections 4.6 and 4.11.
If high levels of CO2 are present in the water, it can be removed down to a concentration of about 5-10 ppmwith an atmospheric degasifier. An atmospheric degasifier has the potential of increasing bacterial burdenand should be located where bacterial control measures are available. One example is to locate the degasifierbetween the stages of a two pass RO system.
b) The Acidification/Degasification Process
The process is well known and accepted in water purification systems. It is usually used where there is ahigh flow rate (>50 gpm or 0.18 m3 /min) or high hardness (>50 ppm). The incoming water is acidifiedbefore the RO unit and a degasifier is used to remove residual CO2 prior to moving on to a second passRO or a mixed bed de-ionization (DI) unit.
In this pretreatment process, the incoming water is adjusted to a pH in the range of 3.8-4.2 with sulfuric acid.The acidified water is sent to a packed column degasifier for removal of free CO2 by air. Removal efficiency ofCO2 is better than 98% (typical commercial degasifiers are designed to reduce outlet CO 2 to less than 5ppm). This residual CO2 should not pose a problem for downstream single and mixed bed de-ionization unitsor electro-deionization. The residual CO2 also can be removed by addition of a base to increase the pH to ³8.5 that converts it to CO3
= which is removed in the second stage of the RO.
Commercial degasifiers are typically from 18 in. in diameter for 50 gpm to 72 in. diameter for 680 gpm (0.46m for 0.18 m3 /min to 1.83 m for 2.57 m3 /min). Fan power requirements will range from 1/2 HP to 10 HP for thepreceding sizes. Smaller and larger units are possible to meet exact needs. Standard packed tower designmethods are used.
The pH after removal of CO2 will be in the range of 6.5-7.0. Just prior to feeding the RO unit, the pH should beadjusted to approximately 8.0-8.5 in order to minimize the amount of free CO2 still remaining in the water andenhance removal of remaining carbonate in the second pass of the RO.
The acidification/degasification process has some associated problems. Air borne bacteria, if a problem orconcern, can be removed by a HEPA filter in the inlet-air line. The air also may oxidize any iron present to formsolids.
Water from the degasification column is usually collected in a holding tank. Further treatment in this tank ispossible for TOC removal and microbial control. A multimedia filter usually follows the degasifier for removal
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 191/250
167
APPENDIX
of initial incoming solids and any solids generated in the degasification step. When using gasification as a“softening” process, addition of an anti-scaling agent is recommended just before the RO (if used as the firststage in the final treatment step). The anti-scaling agent will be removed along with minerals, high molecularweight organics, and endotoxins in the RO. Monitoring for removal of the anti-scaling agent by the RO isrequired.
Advantages: Replaces softener and the handling of large amounts of salt for softener regeneration; theCO2 is released to the atmosphere rather than being purged as an ion in a waste steam; theadded sulfate ion from acidification is easier to remove in RO than added sodium ion fromsoftening.
Disadvantages: Handling acid for acidification; instrumentation and chemical handling for two pH adjust-ments.
11.4.3.2 Nanofiltration
Nanofiltration is a pressure driven membrane process with performance characteristics between RO andultrafiltration. The theoretical pore size of the membrane is one nanometer (10-9 meter). These membranesare sometimes referred to as “softening membranes” and will remove anions and cations. The removal of the
larger anions (sulfate for example) is easier than the removal of a smaller anion (chloride) as discussedearlier.
The nanofiltration membrane offers high rejection of salts of divalent anions as well as organics with molecu-lar weights above 200. This includes color bodies, trihalomethane precursors, and sulfates. The rejection islower, but effective for salts with monovalent anions or non-ionized organics with a molecular weight above150. Typical rejections (based on pure salt solution--mixtures may differ) are shown in the following table:
Solute Descriptive Solute formula MW Rejection-%
Sodium chloride NaCl 58 60
Calcium bicarbonate Ca(HCO3)2 162 80
Magnesium Sulfate MgSO4 120 98
Glucose C6H12O6 180 98
Sucrose C12H22O11 342 99
Final product conductivity will range from 40-200 µS/cm depending on the inlet water total solids and mineralspecies make-up. A single pass RO unit will produce conductivity of 5-20 µS/cm.
The investment cost and size of a nanofiltration system is about the same as for a RO system. Energy use islower because they operate at 70-150 psig (4.76-10.2 bar) as opposed to 200-350 psig (13.6-23.8 bar) forreverse osmosis membranes. Operating pressures are always a function of temperature, feed water salinity,and recovery.
Nanofiltration membranes, like other membranes, are to a large extent application dependent. Key factors arethe quality of the feed water and the quality of the product water required. The feed water should be pro-cessed through a multi-media filtration system prior to going to the membranes. Potential applications are:
• Removal of color
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 192/250
168
APPENDIX
• Removal of trihalomethane precursors and organic carbon compounds from surface waters
• Removal of hardness, radium, and TDS from well water
• Feed waters with high silica
Industrial uses are where moderate water quality is required.
11.4.4 Organics and Microbiological Impurities
11.4.4.1 Introduction
Organic and microbiological contaminants need to be addressed in water treatment systems. The concernsare twofold: contaminants entering the system and contaminants created/growing in the system. Organicsusually enter with the feed water, but also may leach from some non-metallic materials of construction.Microbiological contaminants may enter with the feed water or grow in the system and are classified as viableand non-viable. Viables are those organisms that can proliferate, given specific conditions. Non-viables arederived from a breakdown of or a product of a viable organism.
The first issue to consider is water source since it affects organic loading. If the water is drawn from a well,organic loading is usually not very great. Surface water (lake, river, or reservoir) will probably contain rela-tively high levels of organics and the composition and quantity may show seasonal variation.
Water from a municipal system is usually chlorinated, sometimes with ammonia added to form chloramines.Microbiological content of the feed water will be low and will generally be inhibited until the chlorine/chloram-ine is removed.
The second issue to address is biological growth occurring within the water pretreatment system. Mostpretreatment systems are designed to keep an oxidant in the water for as long as possible to minimize thepotential for growth. Special design and maintenance requirements need to be addressed in all equipmentthat operates without a microbial control agent, chlorine, or chloramine present. These include materials ofconstruction and piping layout (set up and fittings for sampling and periodic sanitization and instrumentation
for monitoring) compatible with the sanitization method selected.
11.4.4.2 Organic Contaminants
The organic contaminants found in many water sources are:
a) Bacterial Contamination
Bacterial contamination is usually expressed as “total viable microbial counts per ml” or as “ColonyForming Units (CFU) per unit volume.” CFUs are determined by counting the growth resulting from incu-bating samples. Each colony is assumed to form from one bacterium.
b) Pyrogenic Contamination
Pyrogens are substances that can produce a fever in mammals. The pyrogens are often endotoxins,organic compounds (lipopolysaccharides) that are shed by bacterial cells during growth, or are the resi-due of dead cells. They are chemically and physically stable and are not necessarily destroyed by condi-tions that kill bacteria. Their molecular weight may vary, generally 12,000 to 320,000. Pyrogen levels arequantified in Endotoxin Units (EU) per milliliter. Pyrogens are of great concern to the pharmaceuticalindustry, since high concentrations may cause responses in humans ranging from fever to shock ordeath.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 193/250
169
APPENDIX
c) Total Organic Carbon (TOC)
TOC is a measure of organic materials contaminating the water and is specified in mg/l or µg/l. TOC is adirect measure of the organic material that is oxidizable. TOC is a very fine measurement used in sophis-ticated water treatment systems where any organic contamination can adversely affect product quality.TOC is not a good measure of microbial contamination.
d) Dissolved Organic Compounds
Organics occur both as the product of the decomposition of natural materials and as synthetic com-pounds such as oils or pesticides. Naturally occurring organics include Tannin, Humic acid, and Fulvicacids. They detract from the aesthetics of water (i.e. color), but unless they come in contact with certainhalogens, they have no known health consequences in normal concentrations. Under conditions of freehalogen compounds (principally chlorine and bromine), they form chlorinated hydrocarbons andtrihalomethanes (THMs), which are suspected carcinogens.
11.4.4.3 Removal of Organics
A number of technologies are available to remove organic materials, and these have differing benefits and
drawbacks. The use of chlorine and chloramines to remove bacterial contamination are the most commonand are discussed in the next section. Treatment devices used to remove one or more of the other types oforganic material are:
a) Ozone
Ozone is twice as powerful an oxidant as chlorine. It will prevent microbial growth as well as reduce theconcentration of organics. Ozone is not used frequently in pretreatment systems due to the preferencefor chlorine and materials of construction that are readily degraded by ozone. Ozone is discussed morefully in Chapter 8, Storage and Distribution Systems.
b) Strong Base Ion Exchange
Organic scavengers or traps are ion exchange resins that contain strong-base anion resins and areregenerated with sodium chloride brine. Most naturally occurring organics have a slightly negative chargeand are absorbed by the anion resin. After the resin is loaded, the organics can be displaced by highconcentrations of chlorides during regeneration.
Advantages: removes most natural organics; can be regenerated
Disadvantages: disposal of brine and organic solution; requires chemicals for regeneration; brine carryovermay result after regeneration
c) Carbon
A carbon bed containing activated carbon will remove organics by adsorption of the organics on thecarbon. Periodically the carbon must be replaced when its capacity to adsorb diminishes.
Carbon is also one of the methods used to remove chlorine. Use of carbon for this and its advantagesand disadvantages are discussed in Sections 4.7 and 11.4.4.4.
d) Microfiltration
Microfiltration includes the use of depth cartridge filters, pleated filters, and cross-flow filtration mem-brane elements. These filters can remove particles ranging in size from 100 µm down to 0.1 µm, thus
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 194/250
170
APPENDIX
capturing bacteria, giardia cyst, and large molecular weight organics. Depth and pleated filters allowwater to flow through a wall of fibers perpendicular to the water direction. The particles are trapped on theoutside wall of these filters, or within the filter walls (for depth filters), due to the pore size of the filter. Thefilter will fill up with these particles and then needs to be replaced with a new filter. Cross flow microfiltrationforces the water to flow parallel to the filtering media, and the particles which are too large to passthrough the filter are then expelled from the system in a concentrate stream to drain (typically 5-10% of
the feed flow). This allows the filters to be self-cleaning and eliminates the need to replace these filtersfrequently.
See Section 5.5 for additional discussion.
e) Ultrafiltration
Ultrafiltration can be used to remove organics and bacteria, as well as viruses and pyrogens from a watersource. Filtration is typically from 0.1 down to 0.001 µm. Cross flow ultrafiltration forces the water to flowparallel to the filter media, and the particles which are too large to pass through the membrane elementsare then expelled from the system in a concentrate stream to drain (typically 5-10% of the feed flow). Thisallows the filters to be self-cleaning and eliminates the need to replace these membrane elements fre-quently. The UF membrane elements will need to have any suspended solids removed from the feed
stream prior to the UF system.
Advantages: effective filtering barrier; no by-products; works with chlorine
Disadvantages: medium to high capital cost; 10% constant concentrate stream; can be source of micro-bial growth
f) Reverse Osmosis (RO)
RO, if included in a pretreatment system to remove anions and/or cations, also will remove organics andmicrobiological impurities. Like ultrafiltration, a purge stream removes impurities that are too large topass through the RO membrane. Advantages and disadvantages are similar to ultrafiltration but toler-ance to chlorine depends on membrane selection. RO as a unit operation is discussed in Chapter 5.
11.4.4.4 Control of Microbiological Growth
The methods of controlling microbiological growth in pretreatment systems are periodic sanitization, ultravio-let (UV) light and chlorine/chloramine.
a) Periodic Sanitization
Periodic sanitization methods, employed on a scheduled or as needed basis, include heat, chemicalsanitization, regeneration or replacement of media, flushing, or drainage. With heat, USP indicator organ-isms are killed above 60°C and the majority of pathogenic organisms will not proliferate. Temperaturesabove 80°C result in complete kill. Sanitization times might be one to two hours at temperature. Totalcycle time including heat-up and cool down might be four to eight hours. Heat is commonly used incarbon beds, filters, and distribution systems.
Chemical sanitization agents (when chlorine cannot be used) include hydrogen peroxide, iodine, ammo-nium compounds, and organic or inorganic per-oxygen compounds. Sanitization times might be 0.5-4hours with additional time for set up to feed the sanitization agent and to flush it from the system. Totalcycle time might be eight hours.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 195/250
171
APPENDIX
Controlling temperature to minimize microbial growth permits increasing the period between sanitiza-tions. Temperatures below 15°C slow microbial growth. Avoiding stagnation and dead legs also mini-mizes microbial growth. Recycle loops around various unit operations can be utilized during shutdownperiods, i.e. recycle around depth filter and softener while sanitizing carbon bed or while cleaning andsanitizing RO.
Times for specialized periodic sanitization methods such as regeneration, replacement of media, anddrainage will depend on the equipment piece and specific design.
All sanitization methods (frequency and length of sanitization) will be system and sanitizing agent depen-dent, and must be validated.
b) Pressurized Carbon Dioxide
Periodic microbial growth inhibition coupled with shutdown is sometimes practiced where the high cost ofelectricity and water favors shutdown of pretreatment and RO systems during normal non-work periodssuch as third shift or weekends. During the shutdown, rather that maintaining water flow to drain to inhibitmicrobial growth, the cartridge filters and RO are deadheaded and pressurized with carbon dioxide to 2-4 bar. This lowers the pH to <5.5 which inhibits microbial growth in the stagnant water as well as dissolv-
ing scaling chemicals such as carbonate from the RO. Before start-up, the acidic water and carbonatesalts are flushed to drain. The advantages include savings on energy and water consumption coupledwith microbial inhibition and de-scaling during a normal shutdown period. In addition, this technology hasthe potential advantage of eliminating the need for a softener before the RO if there is a match betweenshutdown frequency and fouling frequency of the RO without a softener present in pretreatment.
c) Ultraviolet Light (UV)
Treatment with UV light is a popular form of microbial control and disinfecting due to ease of use. Wateris exposed, at a controlled rate, between ultraviolet light waves. The UV light de-activates DNA in themicrobes preventing duplication and hence leading to bacteria reduction. See section on UV light sys-tems in Chapter 8, Storage and Distribution Systems for additional information. In pretreatment systems,UV is used when chlorine/chloramine and heat are not available or possible. The feed water to a UV
needs to be free of suspended solids, which can “shadow” bacteria, preventing adequate UV contact. UVis typically used in controlling feed water to an RO unit that cannot accept chlorine or heat, and incontrolling non-chlorinated water re-circulation during system idle time. The UV system does not leave aresidual in the treated water, and therefore is only effective if there is direct UV light contact with mi-crobes.
d) Chlorine
Municipalities frequently use chlorine to disinfect the water before and during, distribution. Chlorine is fedinto the system to kill bacteria at typical dosage levels of 0.2 to 2.0 ppm. In order to maintain the “killpotential,” an excess of chlorine is fed into the supply to maintain a chlorine residual. The chlorine level atoutlying distribution points is targeted at about 0.2 to 0.5 ppm; however, if the water supply is heavilycontaminated with organics, the chlorine may react and form certain chlorinated hydrocarbons(trihalomethanes or THM’s). In other cases, chlorine can dissipate and no residual level is maintained atoutlying points in a municipal distribution system. Chlorine concentration should be monitored in the feedwater and in parts of the pretreatment system prior to its removal.
Molecular chlorine can have adverse effects on the components in a water purification system. It willcause oxidative deterioration of the membranes, particularly polyamides, used in ultrafiltration and RO. Italso will cause degradation, embrittlement, and loss of capacity in de-ionization resins (oxidation ratevaries with resin type) although the amount is low to moderate at chlorine concentrations usually found indrinking water. It also will cause corrosion of stainless steel, particularly at elevated temperatures and
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 196/250
172
APPENDIX
may carryover into the product in a distillation system. Therefore, in most systems making purified water,the chlorine is removed at some point.
Advantages: low capital cost; common treatment; compliments municipal water treatment; maintainsa residual; easy to test and maintain levels
Disadvantages: can create THMs; does not affect all organics; residual chlorine is not acceptable inmany final treatment systems
The two principal methods of chlorine removal are activated carbon and reduction, often with sulfite.
Activated Carbon
Activated carbon removes the chlorine by adsorbing it onto the carbon particles in a carbon bed. There isalso some reduction of chlorine to chloride. Removal efficiency depends on bed depth, face velocity andadsorptive capacity of the carbon. Design is based on rate of adsorption with adsorption rates typicallybeing more rapid for chlorine than organics if this is done in the same operation. Design based on chlo-rine removal will occur with bed depths of as little as 2-3 feet (0.61-0.91 m) and hydraulic rates of 2-4gpm/ft3 (270-540 l/min/m3) of empty bed volume. Carbon bed volume is a balance between total adsorp-
tive capacity and the frequency of replacement of the carbon bed.
Use of carbon to remove chlorine provides the perfect conditions for microbiological growth: slow flowrates in a warm media with lots of nutrient present. Hence a program to periodically sanitize the carbonbed is required. Heat (either steam [plant steam can be used, but often is not] or hot water at 190°F) iseffective with sanitization frequency varying from daily to a couple of times a week or less. With a propersanitization program, microbial growth in carbon beds can be controlled. Following the sanitization thecarbon bed is usually rinsed to remove fines before being put back in service.
Advantages: removes low molecular weight organics; removes color; removes chlorine effectively;technically not complex; relatively low cost
Disadvantages: high potential for increase in bioburden; medium to high capital cost; shedding of fines
requires downstream filtration; periodic replacement of the spent carbon
Reduction
The addition of a reducing agent will reduce the chlorine to chloride. Sulfite, usually as sodium bisulfite, isgenerally the reducing agent of choice. The chemistry is:
SO3- + Cl2 + H2O ----> 2Cl- + 2H+ + SO4
=
The addition of sulfite, also may require an accompanying pH adjustment step. The chloride and sulfatethat are formed may be removed by a subsequent de-ionization step or RO.
Advantages: effectively removes chlorine; lower capital cost than carbon filters that can be heat sani-tized; no regeneration or replacement required; low operating cost
Disadvantages: technically more complex, chemical handling including sodium bisulfite and acid/basefor pH adjustment; potential for microbial growth in sulfite feed tank requires frequent (<5days) preparation of sulfite solution; higher capital cost for feed systems and monitors;higher cost than disposable carbon.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 197/250
173
APPENDIX
e) Chloramines
Chloramines are formed by the reaction of chlorine and ammonia. Municipalities add ammonia (nearly25% in 1990, principally in the southeast and Midwest US) to form a longer acting disinfectant thanchlorine and to reduce the formation of trihalomethanes during the chlorination of municipal water. Chloram-ines are three compounds: monochloramine (NH2Cl), dichloramine (NHCl2), and trichloramine (NCl3).
Dichloramine is a particularly strong biocide.
Chloramines present problems since the removal is typically not a single step. The methods for chloram-ine removal are:
• activated carbon
• reduction (sulfite injection)
Activated Carbon
Chloramines, like chlorine, can be removed by carbon; however, the absorption is much slower than forchlorine or organics. Chloramine adsorption will require hydraulic rates as much as 3-6X less than chlo-
rine and empty bed contact times of 3-6X those required for chlorine.
The removal of chloramines by activated carbon results in dissociation of some of the chloramines toammonium ion and ammonia. The ratio is dependent on pH and temperature. The ammonium ion can beremoved by cation exchange (water softening). Thus, if chloramines are present in the feed water, it maybe desirable to locate the carbon bed for removal of microbial control agent prior to the water softeningoperation in the pretreatment system design.
The advantages and disadvantages of carbon are similar to those for chlorine. The potential dissociationof chloramines to form ammonia is a disadvantage and can cause problems in final treatment. SeeSection 4.9.
Reduction
Reduction with sulfite will convert chloramines to ammonium ion and chloride ion. These are removed byan ion exchange operation or the ion removal process in final treatment. Again, if chloramine is present,it may be desirable to locate the microbial agent removal prior to the water softening operation in thepretreatment design. The advantages and disadvantages of sulfite reduction are similar to those forchlorine.
11.4.5 pH and Carbon Dioxide
pH, the negative log of the hydrogen ion concentration, is a measure of the concentration of hydrogen ions(H+) in a water-based solution. The more hydrogen ions that are present, the lower the pH and the more acidicthe solution.
The concentration of H+ ions (pH) is very important because it affects the chemistry of the water. For in-stance, the pH of the water, along with other parameters, can tell us if the water will corrode piping or if certaincontaminants (carbonates) are likely to precipitate and cause scaling.
In water or aqueous solutions, a certain ratio of water molecules, H2O, separates (or “dissociates”) into ions,H+ and OH-.
H2O <----> H+ + OH-
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 198/250
174
APPENDIX
Because of the properties of water, when the concentrations of hydrogen (H+) and hydroxyl (OH-) present inany water-based solution are multiplied together, the value is always the same. This number is the “equilib-rium ion product”, Kw, which has been determined to have the value shown below:
Kw = [H+] x [OH-] = 1.01 x 10-14 at 25°CWhere [H+] = Concentration of H+ (moles/liter)
[OH-
] = Concentration of OH-
(moles/liter)
Free carbon dioxide in water is produced by the decay of organic matter, dissolution of carbon dioxide fromunderground sources, and solution from the atmosphere. Since the carbon dioxide content of the atmo-sphere is quite low (less than 0.04%), this is not a major source of carbon dioxide in the water and surfacewaters normally are relatively low in free carbon dioxide; however, well waters usually contain an appreciablequantity of free carbon dioxide.
Free carbon dioxide is the term used to designate carbon dioxide gas dissolved in water. The designation“free” carbon dioxide differentiates a solution of carbon dioxide gas from combined carbon dioxide present inthe form of bicarbonate and carbonate ions. In the case of high purity water, low levels of carbon dioxide fromthe atmosphere can cause the pH to drop from 7.0 to 5.5 and the conductivity to increase from 0.1 µS to 1 µS.Low levels of CO2 also can prevent a water purification system such as two-pass RO from producing water
with a conductivity of <1 µS.
The pH of the water causes the equilibrium between free carbon dioxide (gas) and bicarbonate alkalinity(dissolved ion) to shift to more or less carbon dioxide. The determination of the level of CO2 present in thewater as it proceeds through the treatment process is important to understand because it can affect the finalwater quality or it can cause premature exhaustion of ion exchange systems.
The approach to pH in the pretreatment system affects the equilibrium between carbonate, bicarbonate, andcarbon dioxide. As the pH is lowered the equilibrium is shifted toward carbon dioxide which is a neutralspecies dissolved in the water with the ionic charge being maintained with anions from the added acid andthe net formation of water. As the pH is increased, the equilibrium is shifted toward bicarbonate and thencarbonate with the ionic charge being maintained by the addition with cations from the added base and thenet formation of water.
11.4.6 Importance of Feed Water pH
EPA drinking water standards require that the pH of the water be within a range of 6.5-8.5; however, the rangeof most water is much narrower due to the corrosive nature of water with an acidic pH and a scaling potentialat a high pH. The feed water pH is very important when designing the pretreatment for a purified watersystem. Also, the pH is an important parameter is designing a RO system or an ion exchange system.
If the pH of the feed water is less than 8.3, the feed water will have to be analyzed for the amount of CO 2
present in the water. The lower the pH from a pH of 8.3, the higher the potential capacity for dissolved CO2.The CO2 will directly pass through pretreatment and an RO membrane and depress the conductivity and pH,making it difficult to meet the USP conductivity requirements. If the system has an ion exchange systemfollowing the RO, high levels of CO2 will produce a high ionic loading on the system. High CO2 may require theuse of a degasifier to remove the CO2. See Section 11.4.3.1.
If the feed water pH is between 6-10, the RO system has the potential to incur hardness scaling. Adding acidto the feed water controls the deposition of scale, but this converts carbonates to CO2 that will pass throughboth the RO and distillation final treatment processes. On the other hand, the addition of base converts thebicarbonate to carbonate, and carbon dioxide (CO2) to bicarbonate. These ions will be removed by an ROunit, but also will cause scaling. Most pharmaceutical companies incorporate the use of ion exchange soften-ing in order to prevent scaling from occurring in the RO membrane.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 199/250
175
APPENDIX
Chloramine in the feed water can result in ammonia or ammonium. The pH as well as temperature affects theequilibrium between ammonia and ammonium. Acidic conditions are required to maintain the ionic speciesfor removal in an RO unit. In final treatment ammonia will pass through an RO unit. In distillation, the highertemperatures shift the equilibrium from, ammonium towards ammonia. Ammonia will affect conductivity andpH, making it difficult to meet the USP conductivity requirements.
Based on these situations, some water systems will include the capability to add either acid or a base to thewater in order to optimize the performance of the system. The most common acid used for pH adjustment issulfuric acid because it is readily available and less corrosive than hydrochloric acid. The most common baseused for pH adjustment is sodium hydroxide (caustic soda) because it is readily available, and the finaltreatment process will remove sodium ions.
11.5 FINAL TREATMENT FOR NON-COMPENDIAL AND COMPENDIAL PURIFIED WATERSYSTEMS
11.5.1 Ion Exchange for Purified Water Applications
11.5.1.1 Ion Exchange Use in USP Systems
The primary purpose of ion exchange equipment in USP high purity water systems is to satisfy the ionicquality portion of the specification. Ion exchange systems can effectively reduce organics in many applica-tions with proper ion exchange resin selection and maintenance. Ion exchange systems may not meet USPTOC requirements without additional membrane processes in certain applications where high feed waterTOC levels exist. For most water supplies, both two-bed and mixed-bed units in series can be utilized to meetthe USP water specifications.
Ion exchange systems require pretreatment to remove undissolved solids from the water stream and to avoidresin fouling or degradation. Although dechlorination also is recommended to avoid resin degradation byoxidation, the low levels of residual chlorine commonly found in potable water supplies, in worst cases,demonstrate only long-term effects on most ion exchange resins. Typical pretreatment for an ion exchangesystem includes a filter and/or carbon filter for removal of undissolved solids and chlorine.
11.5.1.2 Functionality
With the exception of one-time use (virgin) resins (which are not regenerated), cation and anion exchangeresins are regenerated with acid and caustic solutions, respectively. As water passes through the ion ex-change bed, the exchange of ions in the water stream for the hydrogen and hydroxide ions held by the resinoccurs readily and is driven by concentration gradient. Similarly, the regeneration process is driven by excesschemical.
Cation Exchange: Cation exchange is the exchange of cations (Ca, Mg, Na, etc.) in water for hydrogen ion(H+). Hydrogen cycle operation of cation exchangers is the term used when regeneration is accomplishedwith dilute acid (generally sulfuric (H2SO4) or hydrochloric (HCl). All salts are converted to correspondingacids following the cation exchange process.
Anion Exchange: Anion Exchange is the exchange of anions (SO4=, HCO3
-, Cl-, etc.) in water for hydroxideions (OH-). This exchange following cation exchange completely demineralizes water when carried to comple-tion. The anion exchange resin is typically regenerated with sodium hydroxide (NaOH).
Ion exchange resins are available in “strongly” and “weakly” ionized versions. Strongly ionized resins have agreater affinity for all ionized constituents in water and are capable of removing even weakly ionized constitu-ents such as acetates and silica with practical exchange capacities of 18 and 14 kgs as CaCO3 per ft3 of resin(0.5 and 0.4 kgs/m3) for cation and anion resins respectively. Weakly ionized resins are ineffective at remov-
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 200/250
176
APPENDIX
ing weakly ionized constituents; however, their exchange capacities are two to three times that of stronglyionized resins and can be regenerated more efficiently.
Ion exchange resins have a higher affinity for polyvalent ions. A result, common divalent ions such as cal-cium, magnesium, sulfate, etc., are removed first as water passes through the resin bed. Monovalent ionssuch as sodium, potassium, and chloride can be displaced by divalent ions in the exhaustion cycle and will
leak into the product stream first. RO product streams are generally comprised of monovalent ions andpresent a more difficult challenge to ion exchange units than a typical raw water. Mixed bed ion exchangeunits were developed to provide superior removal of monovalent ions.
Two bed (also called separate bed) deionizers use two separate columns for the cation and anion resin.Regeneration of a two-bed unit is relatively simple since chemicals are easily introduced and rinsed. No resinseparation or mixing is required as with the alternative mixed bed deionizer.
Mixed bed (also called mono-bed) deionizers utilize one column with both cation and anion resins intimatelymixed for the service mode. Regeneration is more complex than two-bed regeneration as resin separationand mixing steps are required.
Mixed bed ion exchangers function as an infinite series of two-bed ion exchangers since the resin bed is
comprised of both cation and anion resins thoroughly mixed. Therefore, residual ions that may leak through atwo-bed exchange system are eventually removed by the mixed bed exchanger to achieve optimum ionicpurity.
Among two bed systems, there are two types: co-current and counter-current regeneration units. In co-current regeneration systems, the regeneration fluid flows in the same direction of the process water stream.In a counter-current system, these fluids flow in opposite directions. The practical results of counter-currentregeneration are higher quality product water and approximately 50% reduction in chemical usage.
11.5.1.3 Operating Parameters
From a process standpoint, ion exchange systems require consideration of three basic parameters: flow rate,ionic loading, and product water quality.
There are many parameters to consider on a practical level including cost of operation, capital cost, spacerequirements, chemical handling issues, etc. However, for the purpose of sizing equipment and defining thebasic needs of an ion exchange system, these parameters are most important.
Deionizer Type Two Bed Two Bed Mixed-Bed
Co-Current Counter Current
Regeneration Regeneration
Product Water Quality (µS/cm) 2.0-10.0 0.2-2.0 0.055-1.0
Lbs of Acid/Caustic per ft3 resin (100% Basis) 6-8 4 6-8
Kgs of Acid/Caustic per liter resin (100% Basis) 0.1-0.133 0.067 0.1-0.133
Process Flow Rate (gpm/ft2 of bed area) 5-10 5-15 5-25
Process Flow Rate (m3 /hr/m2) of bed area) 12.21-24.42 12.21-36.63 12.21-61.05
Typical Operating Temperature Range in °F (°C) 40-140 (4-60) 40-140 (4-60) 40-140 (4-60)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 201/250
177
APPENDIX
The table above lists basic operation information for ion exchange systems. Process flows should be main-tained within these ranges to avoid problems such as channeling and high pressure drops. Process flow ratesthat are too low may result in channeling through the ion exchange bed. Channeling is typically recognized bysignificantly shorter run times between regeneration. Excessive flow rates can significantly increase systempressure loss and potentially affect product water quality.
11.5.1.4 Component Description
An ion exchange system is comprised of a tank(s), ion exchange resin, a piping and valve system, water/ chemical distributors internal to the tank(s), a regeneration system, and a conductivity or resistivity meter andcell. Typically, other instrumentation may include a flow meter and pressure gauges. Ion exchange systemsare available in both on-site regenerable and off-site regenerable (rechargeable) versions. In both versions,tanks may be constructed from fiberglass, stainless steel or carbon steel with an inert interior lining such asvulcanized rubber or PVC. Off-site regenerated or rechargeable systems are typically transported off site toa facility that is equipped to either regenerate or replace the resin. For this reason, these units are typicallysupplied with fiberglass or light gauge stainless steel tanks in sizes ranging from less than 1 ft3 (0.03 m3)-50ft3 (1.4 m3) per tank. Larger, off-site regenerated systems are recharged with new resins on-site, and theexhausted resin is returned to an off-site regeneration facility.
On-site regenerated units are designed with a much more complicated valve and piping system to accommo-date on-site chemical regeneration and rinsing. These systems are selected when larger volumes of waterare required on a continuous basis, thus justifying the higher capital investment.
11.5.1.5 Tanks
For pharmaceutical applications, given the typical quantity of water utilized, ion exchange tanks rarely needto be more than 3-4 feet (.91 m-1.22 m) in diameter. Typically, tank straight shells are 6-8 feet (1.82 m-2.44 m).Steel tanks are welded and typically manufactured and designed in accordance with the ASME Code foroperating pressures between 100 and 150 psig (7 and 10.5 kgs/cm 2 gauge). ASME Code stamping is notnecessarily required for this type of equipment; however, local regulations and end user safety concernsshould govern this decision.
11.5.1.6 Distributors
Each ion exchange tank includes distributors at all pipe to tank interfaces. Distributors are required to ensurethat resin does not escape from the tank while water is flowing through the system and to provide adequatedistribution of flow through the vessel. Distributors are typically supplied in stainless steel, PVC, CPVC,polypropylene or PVDF. Structural integrity of a distributor system is a key element in any design since aruptured distributor can cause a significant loss of resin and may require significant time for repair.
11.5.1.7 Piping and Valves
The selection of a piping and valve system depends upon several factors including budget, product waterquality (in terms of chemical analysis), and preferred methods of sterilization. Most ion exchange systemsare provided with schedule 80 PVC or CPVC piping and valves. The advantages of these materials includelow cost, ease of assembly, and high corrosion resistance. Specialty plastics such as polypropylene andPVDF also have been utilized in DI systems to a great extent. These materials are more expensive than eitherPVC or CPVC; however, these materials are superior in terms of the lower level of organic leachables into theprocess water. Furthermore, these materials are available in a piping design that more closely resembles theorbital welding in sanitary stainless piping systems.
Stainless steel piping systems offer greater structural integrity than plastic piping systems and the ability tosanitize using hot water or steam. On the other hand, stainless steel is much more vulnerable to corrosionand more expensive than PVC and CPVC.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 202/250
178
APPENDIX
11.5.1.8 Regeneration System
On-site regenerated ion exchange systems require a regeneration system that includes chemical pumpsand/or eductors, chemical tanks, piping and valves, and related instrumentation. The cost critical componentin any regeneration system is the chemical pump. Pumps better suited for this application require inert mate-rials of construction and the capability to closely meter or regulate the chemical dosage. Positive displace-
ment pumps driven by either electric motors or compressed air are ideally suited for this application.
Chemical eductors offer another option to deliver chemicals to an ion exchange system; however, chemicaldosing may be inconsistent based on variable dilution water pressure and flow.
11.5.1.9 Microbiological Concerns, Cleaning, and Sanitization
Although ion exchange resins beds, due to the hydrogen ion and hydroxide ion exchange sites, have pHvalues at the extreme ends of the range, microbiological activity remains a concern. The regeneration of boththe cation and anion exchange resin beds effectively sanitizes the system; however, as the system processeswater, the resin becomes exhausted and the pH approaches neutral. Organic matter, which may be depos-ited on or absorbed by the resins, particularly the anion resin, and the laminar flow of water through the bedfoster bacteria growth in ion exchange beds. For this reason, regeneration frequency is more important to ion
exchange systems that are not designed with auxiliary microbiological control components such as UV lights.Polishing ion exchange systems are typically positioned in a system with bacterial control elements such assub-micron filters and ultraviolet sterilizers, and may operate for several weeks without requiring regenera-tion.
11.5.1.10 Advantages and Disadvantages
Ion exchange based WFI and purified water systems have an extensive installed based and lengthy history.Major advantages are the considerable flexibility in flow rate of ion exchange systems, lack of sophisticatedmaintenance requirements, consistent production of Stage 1 conductivity, and the ability to use the chemicalregeneration of ion exchange resins as a means of microbial control. The major disadvantages include thenecessity to store and handle acid and caustic, the requirements to neutralize waste chemicals (for on-siteregenerated systems only), and the reduced ability of ion exchange resins to reduce dissolved organics
relative to membrane based systems.
11.5.2 Reverse Osmosis (RO) for High Purity Water Applications
11.5.2.1 Description
RO is a pressure driven process utilizing a semi-permeable membrane capable of removing dissolved or-ganic and inorganic contaminants from water. A semi-permeable membrane is permeable to some sub-stances such as water, while being impermeable to other substances such as many salts, acids, bases,colloids, bacteria, and endotoxin.
RO membranes are produced commercially for water purification in spiral wound and hollow fiber configura-tions. Spiral wound elements are much more forgiving in pretreatment protection against fouling. Membranesare available in two basic materials: cellulose acetate and thin film composite (polyamide). All of the mem-brane types have advantages and disadvantages. Cellulose acetate membranes are the oldest commercially.Cellulose acetate has the advantages of being the lowest initial cost membrane and is chlorine tolerant. Theprimary disadvantages of cellulose acetate membranes are: the fastest loss of rejection among membranetypes; the necessity to operate in a pH range of 5 to 6 to minimize hydrolysis; and the necessity to keep freechlorine in the feed stream to control bacterial consumption of the base membrane material. Cellulose ac-etate membranes also are more resistant to some types of fouling than alternative membranes. Celluloseacetate membranes are relatively intensive in energy considerations since the membranes normally operateat a high pressure (300-500 psig or 21-35 kgs/cm2 gauge) and commonly operate at elevated feed watertemperatures (60-80°F) (15-27°C).
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 203/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 204/250
180
APPENDIX
monitoring of RO ionic rejection. Permeate flow, waste flow, and feed temperature are typically monitored aswell as pressure for feed, concentrate, and permeate. These parameters should be measured using cali-brated, NIST traceable instruments. Some RO units also monitor additional parameters such as membraneinter-stage pressure and conductivity. Scale will tend to form on downstream membranes and pressure dropcan be an indicator. Chlorine or Oxidation Reduction Potential (ORP) monitoring is sometimes utilized toprotect chlorine intolerant membranes against oxidation.
Most RO units incorporate some level of automation. Protective devices are typically included to protect:pumps against low suction pressure and membranes against high pressure and high temperature. Mostvalves on RO units are manually adjusted. Automatic valves are utilized in many units to accomplish productside flushing and system sanitization. Membrane cleaning is performed manually in most systems, but canbe automated.
11.5.2.4 Limitations
RO cannot remove 100% of contaminants from water and has very low to no removal capability for someextremely low molecular weight dissolved organics. RO also quantitatively reduces bacteria, endotoxin, col-loids, and other macro molecules from water. RO cannot purify 100% of a feed water system. A concentrateflow is always necessary to remove the contaminants that are rejected by the membrane.
Recovery is defined as the percentage of feed water that becomes purified product water. Most RO unitsoperate in the range of 75% recovery. Some small units operate at lower recoveries, while large systems mayhave recoveries as high as 85% if water consumption is critical. Many users of RO utilize the waste streamfrom the RO unit for cooling tower make-up water, compressor cooling water, etc. The determination ofrecovery must be a balance of life cycle costs, water, waste, and maintenance factors. A high recovery unitmay have less waste to achieve the desired output rate, but it tends to have high maintenance costs due toeffects of the concentrate.
Carbon dioxide passes directly through the RO membrane and, for design purposes, the CO2 will be in theRO product stream at the same level that is present in the feed water stream. Excess carbon dioxide in theRO product stream may cause product water quality problems directly or may increase the load on the anionresin in deionizers which follow the RO unit to raise the RO permeate to a higher quality level.
Water quality produced by an RO system is dependent upon a number of factors, including, but not limited to:membrane type, operating pressure, and feed water quality. Since RO membranes remove a percentage ofthe contaminants in the feed water, as the feed water quality degrades so will the product quality. As a result,it is conceivable that the feed water quality could change enough so that the product quality from the ROsystem may no longer meet USP quality.
Present RO technology requires ambient RO operation with occasional chemical or hot water sanitization.Operating at ambient temperatures can result in the possibility of biological growth. The ability for RO sys-tems to continuously operate at high temperature (80°C) should alleviate this; however, to date, this may bethe most significant factor in why RO use in a WFI application is rather limited.
Since RO membranes remove a percentage of the constituents in the feed water as the feed water qualitydeteriorates or the membranes degrade, it may become more challenging for the RO system to meet USPWFI product quality. As a result, the RO-based USP WFI system may require more regularly scheduledmonitoring than a distillation system.
The performance of an RO system also is dependent upon potentially numerous “o” rings between the mem-brane elements and between the membrane elements and the membrane element housings. O-ring slippagemay result in poor water quality. This would normally result in high conductivity (low resistivity), providing ameans of monitoring for unacceptable water quality.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 205/250
181
APPENDIX
11.5.2.5 Scaling
RO scale formation in waste streams is predicted by a calculation that utilizes the concentrate stream waterlevels of calcium, alkalinity, total dissolved solids, and temperature. The proper pH to operate a RO unit at(without the use of an antiscalant agent) is a minimum of 0.5 pH units below the pH calculated to producescale.
Scale control through acidification of feed stream to lower the pH is very effective. The principal negativeaspect of this method of scale control is in the formation of free carbon dioxide from bicarbonate that ispresent in the feed water as the pH is lowered.
Antiscalant chemicals also are available for injection into the RO feed water stream. The antiscalant agentsare very effective in minimizing scale formation through a sequestering action that increases the time neces-sary for crystal formation of the precipitate.
The various membrane types have different maximum operating temperatures. Many RO systems utilize feedwater preheating to optimize the membrane area and to minimize the pumping energy required to operate theRO system. It is prudent to have high temperature protection for the membranes when utilizing feed waterpreheating.
11.5.2.6 Cleaning and Sanitization
Virtually all RO units need periodic cleaning. Acid based cleaners are used to remove accumulated metalsand salts from the membranes. Alkaline detergent cleaners are used to remove silt and organic foulants fromthe membranes. Sequential acid and alkaline cleanings are frequently done to assure a thorough cleaning.Cleaning frequency should not be more than four times per year if the pretreatment system is designed andworking properly. Cleaning need is based upon a loss of rejection, an increase in the feed to waste membranepressure drop, or a loss of product flow.
Chlorine tolerant membranes can be sanitized with chlorine solutions. Non-chlorine tolerant membranes canbe effectively sanitized with a peracetic acid/hydrogen peroxide solution. Some membranes are sufficientlyheat resistant to allow thermal sanitization at 80°C.
11.5.2.7 Purification Capability and Efficiency
A single stage of RO elements commonly reduces the level of raw water salts by 90 to 99%. Other raw watercontaminants such as colloids, bacteria, and endotoxin also are reduced by 1-3 logs. Passage of waterthrough a single pass of RO membrane may frequently not purify the water to a level that meets the currentrequirements for USP Purified Water.
Repurification of the water through a second set of RO membranes will raise the quality of water to levels thatcan exceed the requirements of USP Purified Water or Water For Injection in most applications. RO systemsthat utilize two sets of RO membranes in series are commonly called two-pass or product staged RO units.Product staged RO units can produce water from the municipal drinking water supplies to produce productwater which will normally have a resistivity of 0.5 Mohm/cm or greater.
Almost all product staged RO units can produce product water that passes the USP tests for conductivity andTOC; however, product staged RO product water may not necessarily pass the conductivity test in rareapplications.
11.5.2.8 Performance
RO unit must incorporate sufficient membrane area for reliable operation. All membrane manufacturers offerrecommendations for membrane area required as function of the feed water quality. One of the most impor-
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 206/250
182
APPENDIX
tant factors for optimization of membrane area is the understanding of fouling and scaling potential. The SiltDensity Index (SDI) reading offers an indication of the tendency of the feed water to foul membranes as aresult of salt, colloids, or bio-burden. Most membrane manufacturers recommend an SDI reading of three orless with an upper limit for proper membrane operation. In general, the higher the membrane area for a fixedproduct water flow rate, the lower the rate of membrane fouling will be. This is not entirely true since thepercent recovery may differ, and therefore might be a factor. A high recovery for the same flow and membrane
area will tend to foul more. It is obvious that an increase in membrane area will cause an increase in thecapital cost of the equipment due to the increased requirement for membranes and pressure vessels. Anintelligent RO design will optimize the membrane area versus the maintenance and cleaning requirements ofthe RO system.
Advantages: RO is utilized to reduce or eliminate chemical handling. This can be a significant advantagefor RO when compared with regenerable deionization systems. RO generally has betterTOC reduction that ion exchange alone.
Disadvantages: Water consumption due to the relatively high waste flow if wastewater reuse is not employed.RO is utilized in many new purified water systems due to the low chemical handling require-ment and the membrane barrier to bacteria, endotoxin, organics, and salts. Additionally, ifthe RO pretreatment is not designed to properly handle raw water change, RO maintenance
can become expensive.
11.6 DISTILLATION FOR HIGH PURITY WATER SYSTEMS
11.6.1 Thermal Efficiency of Distillation Systems
Heat losses from a distillation system are due to:
• Radiation
• Venting
• Heat exiting with the blowdown and distillate streams which are at higher temperature than the feed
• Heat exiting with cooling water
Because of these heat losses, it takes more than one pound of plant steam to produce one pound of WFI ina single effect distiller. For this reason, Multi-Effect distillers are used in order to improve the performance ofthe system.
The Performance Ratio (R) of a distiller is defined as the amount of distillate produced in relation to theamount of steam consumed, and is given by: R = Md/Ms, where
R = Performance Ratio (dimensionless)
Md = distillate produced (lbs)
Ms = steam consumed (lbs)
Another way to measure the performance of a distiller is often stated in terms of Thermal Economy (E). Thisis defined as the amount of distillate produced in relation to the amount of energy input, and can be given by:E = Md/1000 BTU heat input.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 207/250
183
APPENDIX
A typical single effect distiller, without feedwater pre-heating, has an R of about 0.82 and an E of about 0.89.Pre-heating of the feedwater in the condenser raises the R to 0.93 and the E to 0.96 respectively.
Feedwater pre-heaters are sometimes used to elevate the temperature of the feedwater prior to enteringeach effect in a ME system, and thereby reduce the steam requirements of the distiller.
Typical performance values for a 600-gph VC and 5-effect ME are shown in the table below.
11.6.2 Water Recovery
A portion of the feedwater is continuously discharged to waste in order to limit the total dissolved solidscontent of the feedwater within the evaporator. That portion of feedwater not evaporated is termed blowdown ,and is discharged continuously to waste. The total quantity of feedwater (Mf) required is given by the sum ofthe distillate produced (Md) and the blowdown (Mb) discharged. The Recovery Ratio (Rc) of a distilling unit isdetermined by dividing the product rate (Md) by the feedwater rate (Mf).
Hence the Equations: Mf = Md + Mb and Rc = Md/Mf.
In order to minimize energy consumption and pre-treatment cost, it is desirable to have a high Rc. The Rc islimited by the scaling potential of the water. At low flow rates, it may be limited by practical considerations.
The higher the Total Dissolved Solids (TDS) of the feed, the lower is the recovery. With deionized feedwater,the recovery is 90%-95%. With softened feed, the recovery is typically 80%-85%.
11.6.3 Number of Effects or Columns in an ME Design
Increasing the number of effects in an ME system does not result in increased output, but it reduces theamount of steam and cooling water required to produce the same amount of distillate, compared to a systemwith fewer effects.
In order to maintain a temperature difference for heat transfer between the vapor from one effect and theboiling liquid of the next effect, the pressure within each succeeding effect must be lower than its predeces-sor. The temperature and pressure are highest in the first effect, and are lowest in the condenser and the lasteffect, which are nearly at the same pressure and temperature.
In order to minimize the heat rejection requirements of the condenser, it is desirable to operate the systemover the lowest temperature range possible. Additionally, a lower temperature operation reduces the potentialfor scaling and corrosion. If we call the temperature at which the first effect operates the top temperature andthe temperature of the condenser downstream of the last effect the bottom temperature , that constitutes thetemperature range for the multi-effect distiller. In a non-pharmaceutical application, the bottom temperature
VC 5 ME
Cold Hot Cold Hot
Economy in lbs/1000 BTUs 19.52* 6.8 3.84 3.84
Performance Ratio (lbs of distillate/lb of steam 27.67 7.11 3.39 3.39
* Economy takes into consideration the electrical energy consumption for the VC.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 208/250
184
APPENDIX
can be below the boiling point at the prevailing atmosphere, i.e., operation is under vacuum. This is not anacceptable practice for the production of WFI quality water because of the potential for air leak into thesystem. In addition, a low temperature, below that considered sanitary, is unacceptable.
The top temperature is dependent on the pressure of the heating steam, normally 100-125 psig, while thebottom temperature is dependent on the condenser pressure, which is normally 7-10 psig.
The temperature range and equipment cost limit the number of effects that are justifiable. The fixed chargesfor additional effects ultimately dissipate the savings in energy. It is important to note that there is a certainamount of variety in the designs of multiple effect distillers, and that although their geometry may differ, theyare thermodynamically all the same. The larger the output of the system and the higher the cost of steam, themore economically justifiable it is to increase the number of effects.
Typical ME systems have 3-5 effects, and some have as many as 8 effects.
11.6.4 Non-Condensable, Venting, and Deaeration
Non-condensable gases such as carbon dioxide and oxygen are dissolved within the feedwater and becomeliberated as the temperature of the water increases. These gases, if they are not removed, have two detri-
mental effects on distillation units. Since the gases are non-condensable, they can blanket the heat transfersurface and inhibit heat transfer in the condenser and evaporator. This translates into reduced output. Inaddition, these gases are corrosive and will contribute to the pitting, embrittlement, and cracking of stainlesssteel. For these reasons, it is imperative for designers and operators to ensure proper and adequate ventingof the distilling unit.
The steam carries non-condensables that are liberated from the raw water during evaporation. The gases donot travel alone and will not be vented without the associated steam.
There are two characteristics of non-condensable liberated by ordinary waters which allow designers tolocate and vent these gases from an evaporator or condenser. Oxygen and carbon dioxide are significantlyheavier than the associated steam, and therefore collect mostly at the bottom. Additionally, non-condensablegases tend to migrate to the coldest surface they can find.
Considering the points above, the gases liberated in each effect are processed through the bottom of eacheffect with the distillate and make their way to the final condenser which is vented along its length. Operatorsshould ensure that a steady stream of steam is exiting the distiller vent at all times. Improper venting and/ora malfunctioning distillate level control can result in CO2 being pulled into the distillate pump suction andprocessed across the conductivity cell, resulting in high conductivity reading.
11.6.4.1 Deaerator (also called Decarbonator)
All feed waters have dissolved gasses the amount of which depends on water temperature, composition, andpH. The latter is dependent on the type of pretreatment used. Other gasses occur due to the breakdown ofsome of the constituents during heating, as is the case of alkalinity.
It is very desirable, whenever practical, to remove the gasses from the feed waters prior to entry into theevaporator in order to minimize corrosion and to improve heat transfer. This is often done in VC distillers,where a deaerator is installed between the feed heater and the evaporator. Because of the configuration andrelatively high operating pressure of the ME, it is difficult to incorporate a deaerator within the system. Astand-alone deaerator will be costly, and therefore is not included.
A typical VC deaerator consists of a cylindrical tower in which stripping steam and feedwater flow countercurrent to one another. Prior to entering the deaerator, the feedwater has been pre-heated in the blowdowncooler, distillate cooler, and feedwater pre-heater. As the temperature of the feedwater increases, the non-
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 209/250
185
APPENDIX
condensable contained within the feed have a greater tendency to become liberated. Entering the deaeratorat a temperature greater than 93°C, the feedwater is sprayed through a nozzle, which disperses the water inthe shape of a cone to maximize the contact area with the counter current stripping steam. The majority ofgases within the feedwater are liberated at this point, and are vented to atmosphere. The evaporator vent isthe source of the stripping steam which provides additional heating to the feedwater, prior to entering theevaporator. Deaerators are usually constructed of stainless steel.
11.6.5 Typical Water Analysis
The table below provides typical water analysis of a feed source for a high-purity water distillation system,and is used to compare pretreatment and operating cost for a 5-effect ME with a VC.
Feedwater Analysis City of Ocala, FL. Plant Effluent
Constituent mg/l Constituent mg/l Constituent mg/l Constituent mg/l
Calcium 38 Fluoride 0.7 Nitrate 1.1 Zinc 0.026
Chlorine 13 Iron 0.026 Sodium 8.1 Carbon Dioxide 0.2
Bicarbonate, HCO3 1.2 Magnesium 14.0 Sulfate 111 pH 8.2
TDS 157
High Purity Water Production Costs
600 GPH of high purity water at 82°C 600 GPH of high purity water at 30°C
Consumables 5MEF VC 5MEF VC
Steam in lbs/hr 1,471 700 1,471 180
Electricity (kw) 1.5 26.5 1.5 26.5
Regenerant in lb/yr -- 17,520 -- 17,520
Acid in lb/yr 32,143 -- 32,143 --
Caustic in lb/yr 30,917 -- 30,917 --
Cooling Water in gpm 18 -- 32 --
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 210/250
186
APPENDIX
The above figures are based on 7000 hour per year operation and the following utility and consumables’ costsin US dollars:
Associated Costs 5MEF VC 5MEF VC
Steam ($/year) $79,802 $37,975 $79,802 $9,765
Electricity ($/year) $736 $13,000 $736 $13,000
Regenerant ($/year) $876 $876
Acid ($/year) $1,928 $1,928
Caustic ($/year) $4,019 $4,019
Cooling Water ($/year) $1,513 $2,688
Running Cost ($/year) $87,998 $51,851 $89,173 $23,641
Running Costs in US$/1000 gallons $20.95 $12.34 $21.23 $5.63
For costs of consumables, which are different from those used above, calculations can be made by substitut-
ing other values.
11.6.5.1 Comparisons
ME can be used as a standby clean steam generator; VC cannot be used because of the lower steampressure.
Pretreatment: VC does not require as much pretreatment as ME. See Baseline pretreatment in the mainchapter.
Size Selection: The cost of the VC does not scale down as well as the ME. VC is not offered in less than 100gph, and equipment cost is most competitive with ME in the 300 gph and above.
For a hot (80°C) system, an 8-effect ME will approximately match the thermal economy of a VC. For anambient system, it takes 24 effects to match the economy of VC. Obviously, such number of effects is notpractical.
Chlorine Attack: Although the VC still is much more forgiving because it operates at lower temperature, italso is subject to stress corrosion cracking caused by the chlorine. However, at very low chlorine levels, theproblems may not appear until after years of operation.
At the relatively high temperature of the ME, the attack is very rapid, and failure may occur in weeks ormonths.
Steam $7.75 per 1000 lbs Electricity $0.07 per kWh
Sodium Hydroxide $0.13 per lb Renerant-brine $0.05 per lb
Hydrochloric Acid $0.06 per lb Cooling water $0.20 per 1000 gal
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 211/250
187
APPENDIX
11.7 CLEAN STEAM - CLEAN STEAM GENERATORS
Pharmaceutical steam systems including the generator are relatively simple in design, construction, andoperation. Yet, many users can benefit from a better knowledge of factors which contribute to an optimumbaseline system.
11.7.1 System Selection and Design Considerations
11.7.1.1 Sizing the System
A generator designed to deliver the maximum output at a relatively high clean steam pressure will deliversignificantly more steam at the lowest pressure and vise versa.
This appears to be an advantage and often is; however, there can be problems if the performance is not wellunderstood and is taken into consideration at the design stage and later during operation. The followingexample best illustrates the performance of the generator and the advantages and disadvantages.
Example
A generator is sized to deliver 2000 lbs/hr (909 kgs/hr) clean steam at 60 psig (4.2 kgs/cm2 gauge), utilizingplant steam at 120 psig (8.43 kgs/cm2 gauge). Operating at a reduced level of 30 psig (2.1 kgs/cm2 gauge)clean steam pressure and assuming 5 psi (0.35 kgs/cm2) pressure drop across the steam control valve.
Net steam pressure to evaporator =120 - 5 =115 psig (8.08 kgs/cm2), (347.3°F (175.17°C) saturated tem-perature)
CS temperature @ 60 psig = 307.63°F (153.13°C) CS temperature @ 30 psig = 274.46°F (134.7°C)
Temperature difference available for clean steam generation,
∆ T @ 60 psig = 347.3 - 307.63 = 39.67°F (22°C) ∆ T@ 30 psig = 347.3 - 274.6 = 72.7°F (40.5°C)
The amount of steam produced, lbs/hr, W = U x A x ∆ T
Where: U = heat transfer coefficient in Btu/hr/ft2 /°F and A = CS evaporator surface area in ft2
Since the heat transfer surface is fixed, and the change in the heat transfer coefficient is negligible, theamount of clean steam produced is directly propor tional to the temperature difference.
The amount of clean steam produced @ 30 psig compared to that produced at 60 psig at the same plantsteam pressure is: lb/hr of CS @30 psig = (72.7/ 39.67) x 2000 = 3,665 lbs/hr (1666 kgs/hr), or 183% of thatat 60 psig (4.2 kgs/cm).
Specific volume of steam @ 60 psig = 5.818 ft3 / lbs.
Total volume of 60 psig steam = 5.818 x 2000 =1,636 ft3 /hr and specific volume @ 30 psig = 9.403 ft3 /lbs.
Total volume of steam @ 30 psig = 9.403 * 3665 = 34,462 lbs/hr, or 296% of that @ 60 psig.
Since the steam passage areas including the separator have not changed, the velocity of steam increasesapproximately 300%.
If this increase is not anticipated and planned for at the design stage, there can be a potential for carryover ofcontaminants through the separator due to the higher steam velocity.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 212/250
188
APPENDIX
See the following table to determine the output change at different plant and CS steam pressures.
To use the table follow the above example.
• Find the temperature difference for the current and desired operating pressures.
• Calculate ratio of New/Current and multiply the current output by the ratio. If the current output is notknown, use 100% to obtain the percentage change.
In addition to the high velocity and carryover considerations, operating at a temperature difference above70°F (21°C) is not recommended without consulting the CS manufacturer.
At high temperature difference, a heat transfer phenomenon called “dry wall condition” can occur. Simply put,the steam, in this case, is that much hotter than the water on the other side of the heat transfer surface, thatthe surface cannot be maintained wet. It is not unlike throwing drops of salty water on a very hot stove. Thewater evaporates instantly leaving behind a residue (scale). The problem is more of a concern when the CSfeed water has relatively high TDS, such as with softened water. The example above with 72.7°F (22.6°C)temperature difference would not be recommended.
The opposite of operating with high pressure/temperature differential is operating with very low differential.For a CS generator producing 2000 lbs/hr (908 kgs/hr) @ 30 psig (2.1 kgs/cm2 gauge) clean steam and 120psig (8.4 kgs/cm2 gauge). Plant steam, the output drops to 54.6%, or 1091 lbs/hr (495.3 kgs/hr).
Temperature and Pressure Difference between Plant Steam and CS Steam
Clean Steam PLANT STEAM PRESSURES AND TEMPERATURES
P Temp. 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130
Psig °F. 298 303 307.6 312.1 316.3 320.3 324.2 327.9 331.4 334.8 338.1 341.3 344.4 347.4 350.3 353.1 355.8
25 267.3 30.74 35.70 40.37 44.81 49.03 53.05 56.90 60.60 64.15 67.56 70.86 NR NR NR NR NR NR
30 274.5 23.54 28.50 33.17 37.61 41.83 45.85 49.70 53.40 56.95 60.36 63.66 66.84 69.93 NR NR NR NR
35 281.0 16.97 21.93 26.60 31.04 35.26 39.28 43.13 46.83 50.38 53.79 57.09 60.27 63.36 66.34 69.24 NR NR
40 287.1 10.90 15.86 20.53 24.97 29.19 33.21 37.06 40.76 44.31 47.72 51.02 54.20 57.29 60.27 63.17 65.98 68.72
45 292.7 05.27 10.23 14.90 19.34 23.56 27.58 31.43 35.13 38.68 42.09 45.39 48.57 51.66 54.64 57.54 60.35 63.09
50 298.0 04.96 09.63 14.07 18.29 22.31 26.16 29.86 33.41 36.82 40.12 43.30 46.39 49.37 52.27 55.08 57.82
55 303.0 04.67 9.11 13.33 17.35 21.20 24.90 28.45 31.86 35.16 38.34 41.43 44.41 47.31 50.12 52.86
60 307.6 4.44 8.66 12.68 16.53 20.23 23.78 27.19 30.49 33.67 36.76 39.74 42.64 45.45 48.19
65 312.1 4.22 8.24 12.09 15.79 19.34 22.75 26.05 29.23 32.32 35.30 38.20 41.01 43.75
70 316.3 4.02 7.87 11.57 15.12 18.53 21.83 25.01 28.10 31.08 33.98 36.79 39.53
75 320.3 3.85 7.55 11.10 14.51 17.81 20.99 24.08 27.06 29.96 32.77 35.51
80 324.2
NOTE: Plant Steam Pressure Is Downstream of Steam Control Valve.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 213/250
189
APPENDIX
11.7.1.2 Feed Water Pressure and Pump
The feed pressure is typically 8-10 psi (0.5-0.7 kgs/cm2) above the clean steam pressure, which is typically30-60 psig (2.1-4.2 kgs/cm2 gauge, a feed pressure range of approximately 40-70 psig (2.8-4.9 kgs/cm 2
gauge). Most CS generators have outputs of less than 4,000 lbs/hr (1,816 kgs/hr). Many have outputs of lessthan 1,000 lbs/hr (450 kgs/hr). That is a feed rate of slightly more than 2 gpm (7.5 l/min).
The sanitary centrifugal feed pumps favored by the pharmaceutical industry will be operating near shutoffhead at flow rates of less than 10 gpm (38 l/min), particularly with very small generators. Installing pumpcirculation loops complicates piping, and under certain scenarios may call for installing a heat exchanger toaddress temperature buildup. This adds undue complexity. Consider the possibility of feeding the generatorfrom an existing source with sufficient pressure. The dedicated pump, even at very low flow rate, is most likelythe next best option.
11.7.1.3 Feed Water Contact Surfaces
It is common that users specify mechanical and electropolish and sanitary clamp connections for piping, heatexchangers, and vessels in contact with the feed water although such features are not necessary for achiev-ing the desired clean steam quality.
Often, the reason given for such requirements is attributed to having the feed to the CS piped from the WFI orUSP Purified Water loop. Such practices rule out the use of the less costly Simple CS Generator whichotherwise may be acceptable.
Consider the economics of the proposed feed source at the selection stage. If WFI or USP Purified is themost logical source, consider the possibility of using an atmospheric break or sanitary backflow preventersso that the CS does not become an extension of the source loop.
11.7.2 Utilities
Plant Steam Flow Rate, lbs/hr: Make provisions for 110-120% of the maximum expected clean steamproduction. Generators with feed heaters and low TDS feed water will use closer to the 110% rate, and those
without feed heaters and with high TDS (high blowdown rate) will use closer to the 120% rate.
Plant Steam Pressure should be nominally 50-60 psig (3.5-4.2 kgs/cm2 gauge) above the clean steampressure. See (1) above.
Feed Water Rate: With very low TDS, DI, or equal water, the typical feed rate is 105%-110% of the CSproduction rate depending on the size of the generator.
With higher TDS feed water, the blowdown rate is increased to lower the concentration of salts. The feedvaries and often is 115%-120% of CS rate.
Electric Power: Other than controls, the only power requirement is for the feed/booster pump when used.Since almost all CS generators require relatively small flow rates, the power requirement is mainly dependenton the boost pressure. Motors will vary from 2-5 horsepower. (1.5-3.75 kW).
Instrument Air: 80-100 psig (5.6-7.0 kgs/cm2 gauge) instrument air is required for systems with air operatedcontrols.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 214/250
190
APPENDIX
11.7.3 Periodic Cleaning and Derouging
It is common for CS generators and other stainless steel equipment used in high purity steam and waterapplications, to develop rouge or scale. Some of the cleaning chemicals can be very harsh and may etch thestainless surface. The extra cost on electropolishing becomes wasted after one harsh cleaning. When con-tracting an outside service to perform the cleaning and any subsequent passivation, ensure chemicals used
do not cause etching. It is common to avoid passivation with nitric acid if the surface is electropolished, usingcitric acid and ammonium citrate.
11.7.4 Conductivity Monitoring
The conductivity of the CS condensate sample on a system with relatively high feed conductivity is an indica-tion of carryover through the separator. Therefore, it serves as a good on-line warning that pyrogens may bepresent. When the CS generator is operating on high purity DI feed water, the monitoring of CS sampleconductivity or resistivity is meaningless; especially if the conductivity is lower than the limits set in the USP.
In order to verify that the separator system works as intended, it would have to be challenged. Sodium sulfiteor other non-chloride-based salt can be injected in the feed to determine the corresponding product conduc-tivity. Recommended salt concentration in the feed is 150-200 mg/l. Based on approximately 10% blowdown,
the concentration inside the CS will be 1,500-2,000 mg/l. The test is recommended after fieldwork or cleaningwhich may involve the disturbance of the separator system.
11.8 MICROBIAL CONTROL BASICS, TESTING, AND STERILIZATION SANITIZATIONEQUIPMENT DESIGN AND INSTALLATION ISSUES
11.8.1 Bacteria in Pharmaceutical Water Systems
11.8.1.1 Background
Various types of bacteria can be found in the feedwater to purified water systems. Among the ways that thesebacteria can be classified is Gram positive or Gram negative, which is based on their retention of a dye after
a staining and washing procedure. The results of the Gram test are dependent on cell wall structure, andtherefore indicative of many other characteristics.
Most aquatic bacteria are Gram negative, and bacteria commonly found in purified water systems are gener-ally Gram negative because Gram positives cannot thrive on extremely low nutrient levels. Gram negativebacteria are more heat sensitive than Gram positives, and cannot proliferate at temperatures above 60°C.However, the possibility of Gram positive organisms in the system cannot be completely ruled out, and thesanitization and testing procedures should take this into account.
Bacteria generally range from approximately 0.5 µm to 5 µm. The rod-shaped Pseudomonas diminuta, whichis used for challenging sterilizing filters, has a minimum dimension of approximately 0.3 µm.
When Gram negative organisms are killed they release endotoxins; lipopolysaccharide molecules whichcause a fever when injected into the bloodstream. Because they cause fever, these substances are termed“pyrogens,” meaning generating heat. These substances also have been linked to much more serious reac- tions when introduced into the blood stream, including lethal septic ‘shock’. They are a major concern inproducts intended for injection and may be a concern in other non-oral dosage forms such as transdermals.
Bacteria have a tendency to attach to surfaces and form biofilms. Biofilms form an external polysaccharidelayer. This external film protects the bacteria from antimicrobial agents and makes them much more resistantto sanitization procedures. In addition, disinfection of the biofilm does not necessarily remove it from the
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 215/250
191
APPENDIX
surfaces. Since the dead biofilm provides nutrients and attachment points for new biofilm, its removal must beconsidered when developing a sanitization strategy. Biofilms are difficult to detect since water samples areused to measure free floating (planktonic) bacteria, and these bacterial counts do not directly correlate tobiofilm concentration. Prevention or removal of biofilms can be affected by the various continuous and peri-odic microbial control methods described in this chapter; however, once a biofilm has developed, the severityand frequency of the microbial control techniques may need to be significantly increased.
11.8.1.2 Microbial Levels in USP Purified Water and Non-Compendial Water
Most techniques for measuring microbial populations in a water sample require a time delay of two to threedays, and they do not yield an exact value. For this reason, the USP monograph for purified water does notcontain a microbial limit. Instead, alert and action level guidelines must be established by the end user. TheUSP informational chapter discusses the need to match the microbial quality of the purified water to the enduse. An action level of 100 CFU/ml is suggested in the USP, and the FDA Guide to Inspection of High PurityWater Systems states that any action level over 100 CFU/ml is unacceptable.
Since actual microbial counts are dependent on the method and technique of counting, the specifications foralert and action levels should be based on an understanding of the method, including time, temperature, andmedia. What may be more important than the total count is a continuous monitoring program that indicates
trends before they become a problem. It also is important to identify the specific contaminants in addition tothe total count. The water system should not add objectionable organisms to the final drug product. Objec-tionable organisms are defined in the FDA Inspection Guide as “any organisms that can cause infectionswhen the drug product is used as directed, or any organisms capable of growth in the drug product.”
Alert levels are set according to a statistical review of data to reveal those results that are above the normaloperating range of the system, and are generally used to initiate an investigation with the objective of detect-ing and resolving problems before the action limit is reached. It should be noted that action levels are notpass/fail tests. These are levels which indicate that a problem exists and the problem must be investigatedand an appropriate corrective action taken.
11.8.1.3 Microbial Limits in WFI
According to the USP, Water For Injection is “intended for use in the preparation of parenteral solutions.” Themonograph makes no specific reference to a bacterial limit. WFI is generally expected to be free of microor-ganisms. Since some microbial contamination may be encountered during sampling, an action limit of 10CFU/100 ml is commonly specified. The presence of any amount of bacteria in a WFI system is cause forconcern and should be investigated.
The bacterial requirement which is explicitly stated on the monograph is for bacterial endotoxin. The USPendotoxin limit for WFI is 0.25 EU per ml.
Even bacteria which are destroyed quickly within the storage and distribution system will release endotoxin.Therefore, it is imperative that the treatment system removes bacteria before they get to the distributionsystem, rather than relying on protection measures within the distribution system (such as heat) to removebacteria.
Although bacterial contamination in a WFI system will lead to the proliferation of endotoxin, it may be possibleto pass an endotoxin limit, but not a bacterial limit so both tests should be performed. An endotoxin levelcannot be directly correlated to a bacteria count.
11.8.2 Testing and Documentation
Microbial control in high purity water systems relies heavily upon the use of instruments that directly monitorthe control parameters (e.g., temperature indicators, ozone monitors, UV intensity meters). However, micro-biological monitoring is necessary to assure that the intended water quality is met at the system use points.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 216/250
192
APPENDIX
Microbiological monitoring is performed by removing a sample of the water at designated sample ports in thesystem and performing a heterotrophic plate count. For systems in which endotoxin must be controlled (i.e.,WFI), a Limulus Amebocyte Lysate (LAL) test also is performed. The heterotrophic plate count requireslaboratory manipulation (e.g., filtration/ plating), incubation, and reading, which require several days beforeresults are available. Identification of bacteria found will take considerably longer. Therefore, a documentationsystem needs to be in place so that a correlation can be made between results and the system operating
information at the time of sampling (such as the sanitization logs, flush logs, data from system operationrecords, the identification of the sampling technician, etc.)
11.8.2.1 General Testing Considerations
An initial, more intensive sampling program will be undertaken as part of the validation effort, and is dis-cussed in Chapter 10 and in the Commissioning and Qualification Guide. This section focuses on ongoingmicrobial sampling and monitoring after the system has been placed in production.
All use points in the distribution loop should be represented in a monitoring regimen. Testing of the waterbeing supplied to the loop, prior to dilution in the distribution tank, provides important information on theefficacy of the purification system. For WFI or other water with endotoxin limitations, LAL testing is typicallyperformed concurrently with the heterotrophic plate count. The water system is typically tested daily with
individual sample ports being rotated through to include each point over a specified period. The specifiedperiod between samples for any given use point must be based on historical data. Sampling each point everyweek has generally been found acceptable by the FDA. With smaller systems, this may result in only onesample from the system on any day. Although it would be more costly, sampling multiple points each day canprovide useful information about the system. For example, if one sample is found to be unacceptable, andthere were no other samples taken on that day, it is difficult to know if the problem is from the use point, orthroughout the system. However, if many samples were taken and only one use point had a problem, thequality of water from the remaining use points can be defended. The sampling performed should simulate theproduction use of that water. For example, water used through a hose should be sampled through the samehose. The flush performed prior to sampling (if any) should reflect the flush used prior to production use.Procedures for sampling and testing must adhere to aseptic technique to ensure that contamination is notintroduced.
The quantity of water used for analysis is a minimum of 1 ml for purified and non-compendial, and 100 ml forWFI, according to the USP. Larger quantities such as 100 ml for purified and 250 ml for WFI are recom-mended.
11.8.2.2 Testing Documentation
Sampling procedures should be delineated in an approved protocol (SOP). Testing methods also should bespecified in an approved SOP. This would include the equipment and materials to be used, the procedures,growth promotion testing of the media, and negative controls. Procedures for performing the heterotrophicplate count are delineated in the latest edition of Standard Methods for the Examination of Water and Waste- water. Note that organisms in pharmaceutical water will have adapted to a low nutrient environment, and astandard growth media may shock the organisms and result in erroneously low counts. Therefore, a lownutrient media may be more appropriate for the testing. All testing should follow a validated procedure. Allmethods, compendial or otherwise, should be shown to be effective on the particular system being tested.
For new or modified systems, sampling and testing should be completed as part of a Performance Qualifica-tion (PQ). The PQ should provide an overview of the system or the modifications and delineate the monitoringprogram (the sample ports, frequency of testing, type of testing performed, relevant SOPs, etc.). When com-pleted, the PQ should include testing results, investigations of Out Of Specification (OOS) results, specialstudy results (such as a sanitization frequency study), a section for any deviations, and a summary report.The sampling regimen should include samples taken immediately prior to sanitization in order to exhibit worstcase conditions.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 217/250
193
APPENDIX
Following qualification, testing must be performed routinely in order to ensure that microbiological control ismaintained. The documentation system must provide ready access to test results.
11.8.2.3 Analysis of Results
The primary purpose of reviewing test results is to reveal OOS findings to ensure product is not jeopardized
and so that correction can be made. The identification of the organisms found is an important part of theinvestigation since this will provide strong evidence of the source of contamination. For example, Staphylo- coccus is typical of sample handling problems; Bacillus is typical of environmental contamination of thesample port, and Gram negative rods are typical of water borne contamination. Follow-up sampling per-formed in reaction to the OOS result would be included in the investigation.
11.8.3 Sterilization/Sanitization Designs
Various designs may be incorporated to facilitate heat sanitization of process water distribution systems.In this section, two designs for sanitization are considered:
• Hot Pharmaceutical Water (80°C) System
• Ambient Pharmaceutical Water System
a) See Section 11.4.4.4 on Pretreatment Processes - Control of Microbiological Growth earlier in this chap-ter.
The boundaries for this design will be defined for each of the two designs set forth.
The systems described are based on designs that have been found acceptable in a number of installations.It is recognized that a wide variety of system designs may meet the same objectives and also prove accept-able from a regulatory and operations standpoint.
11.8.4 Hot System Distribution Loop
In this system, the boundaries begin with the storage vessel and include pumps, distribution piping (all drops)and the return loop that terminates at the storage vessel.
The storage vessel should be a fully jacketed vessel with the capability of heating and cooling. Heating of thestorage vessel is usually achieved with steam. Cooling, when required, is achieved with chilled or coolingtower water.
It has been found that a hot distillation system has a tendency to build heat when the effluent from the still isdischarging into the tank. The effluent may be as high as 96°C. If the water is too hot, the distribution pumpmay cavitate or boiling will occur in the pump. In addition, the system may be more susceptible to “Rouge”when water is kept at temperatures exceeding 87°C.
A hot system is generally considered to be self sanitizing as long as the temperature of the circulating wateris maintained between 75°C and 85°C.
When the system is down for maintenance or repairs or modification, it will be necessary to sanitize thesystem to bring down the microbiological load and to reduce the pyrogen burden.
The method used with hot systems is to flush the system, all sampling points, and all points of use with hot(80°C) water. The quality of the flush water must be at least as good as the water to be used in the system.Flushing with two to three volumes of the water in the distribution system should achieve sanitization. Ifsanitization is not achieved, then flushing must be continued and the source of the contaminant removed.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 218/250
194
APPENDIX
The heating medium may be hot liquid or steam. It should be designed to heat the ambient process waterfrom 15 to 80°C within a specified period (generally two to four hours).
If the heating and cooling heat exchangers are not directly installed in the return loop, process water shouldpass through when not in use for heating, eliminating the possibility of a dead leg.
Most systems can be fully sanitized with 80°C circulating process water within two to three hours once 80°Cprocess water temperature is achieved.
As a safety note, all exposed piping and storage tank surfaces should be insulated to prevent accidentalburning of personnel in the heat sanitization cycle. In addition, the points of use must be monitored during theentire sanitization cycle.
11.8.5 Cold System
The boundaries are the same as a hot system. A by-pass line may be provided to recirculate process waterthrough the pretreatment train to avoid stagnant flow conditions.
Sanitization of a cold system may be accomplished using hot water using two stainless steel double tube
sheet heat exchangers, one for cooling and one for heating.
The heat exchanger used for cooling is typically installed after the distribution pump(s). The heat exchangerused for heat sanitization is usually installed on the return loop just prior to the return to the storage tank.
One heat exchanger to heat and cool the process water is a viable alternative, but should be designed for themore extreme temperature range.
The heat exchanger for cooling may be cooled with cooling tower water, chilled, or domestic cold water. Thecool down cycle should be designed to cool 80°C process water to approximately 15°C over a specifiedperiod (based on the operational needs). The initial cooling should be very gradual to prevent thermal shock.
11.8.6 Ozone System
Ozone (O3) is a naturally occurring triatomic form of oxygen. O3 is unstable at atmospheric temperatures andpressures, and decomposes readily into molecular oxygen (O2).
There are no objectionable by-products or residues when water is disinfected with ozone. The presence ofoxidizable substances will generate trace carbon dioxide and in the absence of oxidizable substances, onlyoxygen will be formed.
Very low concentrations of ozone of the order of 0.1 to 0.2 mg/l have been shown to control microbiologicalgrowth to below 1 CFU/100ml.
Due to the limited half life, ozone must be produced on-site where it is required.
11.8.6.1 Materials of Construction
Few organic materials are unaffected by ozone contact. Gaskets, piping, vessels, filters, and ion exchangeresins may be subject to attack, and therefore all materials that come in contact should be specifically se-lected.
PVC will be attacked by ozone, but PVDF and PTFE are not so vulnerable.
Stainless steels that may be pitted by Chlorine resist ozone. Thus, ozone will not damage stainless steel stillsand will be removed in the distillation process.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 219/250
195
APPENDIX
11.8.6.2 Comparisons with Chlorine
Chlorine is relatively stable in water and is not described by any half-life characteristic.
Ozone has a relatively short half-life depending on temperature and pH. At neutral pH and 20°C, this can betwo hours in high purity water. Multiple feed points can reinforce the concentration of ozone in a very large
distribution system.
Removal of chlorine and many of its bi-products are comparatively difficult to remove. Absorption by activatedcarbon or reaction with bisulfite are the common methods used, each having its own specific problems.Chlorine may convert many organic substances into derivatives that have been identified with carcinogenic-ity.
Ozone reverts to oxygen and/or carbon dioxide.
Chlorine requires diffusion through the cell walls of the organism in order to degrade enzymes. Tests indicatethat chlorine required more than 400 hours at a concentration of 0.1 mg/l to destroy 99% of 60,000 CFU of E.Coli.
Ozone, when added to water using an efficient mixer, kills bacteria in seconds. At an equivalent CFU level tothe chlorine test, ozone will destroy more than 99% in less than one minute.
11.8.6.3 Ozone Generation
There are two commercial methods to manufacture ozone: corona discharge and electrolytic generation.
The corona discharge method uses an air/oxygen feed. It produces acceptable quality ozone for pharmaceu-tical purposes if the feed stream to the generator is sufficient low in nitrogen and moisture to avoid productionof harmful level of nitric acid. Excessive levels of nitric acid in the ozone effluent can drop the pH and conduc-tivity of the water below acceptable levels and promote corrosion of stainless steel surfaces. Pharmaceuticaldesign typically employs a dryer and a molecular sieve to reduce nitrogen and moisture from the instrumentair.
The other commercial method of ozone production is electrolytic generation. Low conductivity water is usedas the feed stream to a catalytic generation cell. Electrolytic ozone generators can be operated directlyimmersed in a side stream of the water circulating system or can be used to produce a gas first and then betransferred to the water using a conventional contractor device.
Electrolytic generators typically produce lower contamination levels than corona discharge units, but aregenerally more capital cost intensive. Both can be operated successfully when designed and maintainedproperly.
11.8.6.4 Ozone System Installation
The optimum location for ozone injection is in the loop return, just before it re-enters the storage tank.To ensure ozone does not contaminate the pharmaceutical product, it may be removed shortly before the first“point of use.”
Ozone destruction can be accomplished by various technologies including:
• Catalytic
• Thermal
• UV
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 220/250
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 221/250
197
APPENDIX
Pressure limitations also need to be recognized. Most UV units have a operating pressure limit of 100-120psi. High pressures require special arrangements, such heavier flanges or an ASME code stamp.
Most UV units are installed in a horizontal position to eliminate air pockets; however, consulting the manufac-turer for their recommendations is appropriate.
A log reduction target should be specified when purchasing a unit.
All wetted parts should be similar to the storage and distribution system specifications.
The use of UV units does not negate the need for periodic sanitization of the system.
The specified dosage should be obtained even at the end of lamp life, as all UV lamps degrade and losegermicidal efficiency over time. The units should be sized so that a safety margin is built in so that even at theend of the lamp life (EOL), the minimum dose requirement is always exceeded.
11.8.7.3 UV Light Dosage and Sizing Requirements
Proper UV dosage measured in milli-watt seconds per square centimeter at the chamber wall determines the
performance of a UV unit in storage and distribution systems. Each microorganism has been assigned a killdose, known as D10 that produces a 1 log, or 90% kill rate. E. Coli, for example, has a D10 of 3 milli-wattseconds per square centimeter. This gives a 1 log reduction. To achieve a 3 log reduction, a dose of 6 milli-watt seconds is required. Most pharmaceutical water systems require at least a minimum of 30 milli-wattseconds per square centimeter dosage to maintain germicidal effectiveness.
The following factors are involved in the selection of a UV unit to obtain satisfactory performance in storageand distribution systems:
• Flow-rate - The units must be sized to handle the maximum flow possible from the system.
• Water quality is critical to the success of a UV unit’s performance. Whenever possible, a water sampleshould be obtained, and the water checked for UV transmission through a UV spectrophotometer. UV
absorbing compounds such as iron, manganese, dissolved organics, turbidity, or suspended solids allaffect germicidal efficiency. Although in the storage and distribution systems, only the dissolved organicsshould be seen.
• Identification of microorganisms and concentrations also affect sizing. If there is a high concentration ormultiple organisms, this could dictate a larger UV unit.
• Temperature of the water also affects the choice of UV units. Low pressure units have specific tempera-ture limitations, while medium pressure units do not. In very cold (5-10°C) or hot water systems (above45°C) the arc tube efficiency is greatly reduced.
• Whenever possible water in storage tanks or distribution loops should be recirculated through the UVunit. The ideal rate would be a minimum of four times per hour for the entire volume of the tank or theentire volume of the loop.
11.8.7.4 Characteristics
Conventional low-pressure units usually have multiple lamps in a chamber to achieve the proper amount ofUV energy for germicidal energy. Medium pressure units usually employ a single lamp approach due to themuch higher output of the individual lamp. Each lamp type has advantages depending upon the application.The lamps are enclosed within quartz sleeves which contact the water. All other wetted parts should bemanufactured from an acceptable iner t material, generally 316L SS. UV intensity and frequency should bemonitored continuously and recorded.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 222/250
198
APPENDIX
11.9 FABRICATION/PROCEDURES FOR DISTRIBUTION SYSTEMS
11.9.1 Introduction
Fabrication of the distribution system must be made with extreme care and precision to ensure a smoothinternal finish that will not allow any crevices that will support or promote bacterial growth, metal corrosion, or
particulate generation.
The chapter provides an understanding of the requirements of fabrication, materials of construction, andspecialized equipment required to fabricate a piping distribution system.
Decisions concerning the material of construction, the orbital welding, the special handling, and the specialenvironment required must be made and understood by the pharmaceutical manager responsible for the highpurity water distribution piping fabrication and installation.
Unlike the equipment selection, the piping fabrication will require selections of material with specific compo-sitions, welding using inert gas envelopes, and a cleanroom environment.
11.9.2 Materials of Construction
Material selection should be consistent throughout the distribution, storage, and processing systems. Thematerial should be rigid, capable of withstanding steam sterilizing temperatures (as required), cleaning solu-tions, passivation solutions (as required), and capable of maintaining a durable and corrosion resistant sur-face finish.
The material in common use in pharmaceutical water systems is a stainless metal. Based on cost and easeof fabrication, this material is nickel chromium steel.
Inert non-metallic materials also are available that can withstand steam sterilization (as required). Thesematerials are used extensively in the electronics industry for ultra-pure water systems where they haveproved themselves capable of containing and preserving the high level of purity used in the semi-conductorindustry.
11.9.3 Types of Stainless Steel
Type 316L is the preferred steel for a high purity water generation and distribution system. The “L” designationindicates a low level of carbon compared to the non “L” grade.
Alternates to 316L grades are 317L with its higher chromium and molybdenum contents and 304L with itshigher chromium content but lower nickel and molybdenum levels.
11.9.3.1 Corrosive Resistance
Chromium content is the most important alloying element in stainless steel followed by nickel and molybde-num.
Steels with chromium in excess of 11.5% will form a protective film of chrome oxide on the metal surface. Thepresence of nickel in amounts more than 7% enhances the corrosion resistance over the straight chromiumgrade and improves its ductility.
The performance of stainless steels is governed by the oxidizing characteristics of the environment, similarlyto the oxidization of ordinary steels. Strong oxidizing conditions generate a superior protective coating for thestainless steel and a powdery rust on ordinary steels which eventually, if left unchecked, will consume all ofthe metal.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 223/250
199
APPENDIX
11.9.4 Weldability and Polishability
Austenitic stainless steels such as AISI type 316L contain impurities in addition to the major elements de-scribed. These include:
• Sulfur (S)
• Aluminum (Al)
• Oxygen (O)
• Silicon (Si)
• Manganese (Mn)
• Phosphorus (P)
• Titanium (Ti)
• Calcium (Ca)
The effects of some of these elements are cumulative inasmuch as the oxygen can offset some of thewelding characteristics of low sulfur steels and the weld penetrations experienced with sulfur may be offsetby high levels of aluminum. In addition, the ratio of the trace elements of aluminum to silicon has been shownto effect slag formation during welding which in turn produces inclusions.
Manganese has been shown to combine with the sulfur and form manganese sulfide (MnS) inclusions on orin the surface of the metal.
Even the major element chromium, may contribute toward oxide inclusions in the metal surface.
11.9.4.1 The Impact of Sulfur on Welding and Internal Finishing
Small amounts of sulfur improves machinability as well as weldability; however, with the advent of moderntechniques in the steel refining process, the mills are able to produce steels at a cost with very specificchemical compositions.
Low sulfur AISI 316L steels when based on the specification limit of 0.03% have some advantages to obtain-ing an unpitted polish, but other impurities such as manganese, silicon, oxygen, aluminum, calcium, titanium,and chromium can contribute to the oxide inclusions that we are trying to minimize by reducing sulfur content.
A mid-range sulfur content would be ideal for welding but any mismatch in the sulfur content of the matingparts will easily outweigh the advantages of low or lower sulfur levels.
An ideal compromise would be a level between 0.005% and 0.02% or a modified maximum limit of 0.02%since the lower limit will probably not be attained. However, if not all welded parts can be obtained at similarlevels, the exercise produces very little overall advantage.
11.9.4.2 Heat Number and its Impact on Weldabilty
Weld parameters will remain consistent and orbital welding processors will reproduce similar welds if thematerial remains the same. Since the melting point of steel varies over a range depending upon the concen-tration of each element present in the steel, the amount of heat required to melt and thus attain the liquefac-tion of the steel will vary from concentration to concentration.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 224/250
200
APPENDIX
Each “melt” of steel is assigned a “Heat Number” to identify the mix or exact composition of the steel.316L stainless steel has a wide tolerance of “active” elements such as chromium, molybdenum, nickel, andcarbon which vary as much as 100% and as little as 12%. These variations will change the melting point andthe electrical current input of the welding machine considerably.
While the “Heat Number” can be specified for all tubing purchased for a specific distribution system, the heat
number for connected fittings and valves are more difficult to control.
Where possible, all fittings, valves, tubing, and weldable pieces of the same nominal size (diameter) shouldbe purchased and manufactured from AISI 316L steel with the same Heat Number in order to standardize theweld quality for each tubing size.
11.9.5 Corrosion Hazards of Stainless Steels
While being protected by the chrome oxide film, which will guard against most corrosive attacks, there re-mains five corrosive hazards associated with the successful use of stainless steels.
a) Intergranular Corrosion
Austenitic 300 series stainless steels which do not contain any of the stabilizing elements, titanium, orcolumbium are susceptible to intergranular corrosion which can cause early failure or reduced life.
b) Galvanic Corrosion
Occurs when an assembly of dissimilar metals is immersed an any solution which acts as an electrolyte.Therefore, any transition of drain or condensate piping to a dissimilar metal should use a dielectric unionto prevent electrolysis. It is recommended to maintain drains and condensate from pharmaceutical waterand steam in stainless steel due to corrosivity of fluid.
c) Contact Corrosion
Occurs when small pieces of carbon steel, scale, copper or other foreign material is lodged on the
surface of stainless steel.
d) Pitting or Pinhole Corrosion
Solutions containing chlorides may attack stainless steel in a pitting action. This is usually due to highconcentrations of the chloride ion due to evaporation.
e) Stress Corrosion Cracking
Chloride solutions are the worst offenders in promoting stress cracking. Cracking is most likely to occurin hot rather than cold solutions. High and low stresses in the same member produce a condition likely toresult in stress corrosion if chlorides are present.
Corrective action is to keep the internal strain in stainless steels as low as possible by fully annealing orutilizing stress relieved versions of stainless steels at 1200°F.
Select tubing with good concentricity and close wall thickness tolerance to avoid high and uneven stresseswhen tubes are rolled into tube bundle headers etc.
Avoid joining dissimilar metals and avoid cyclic bending.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 225/250
201
APPENDIX
11.9.6 Corrosion Protection of Stainless Steel
Surface protection of chromium nickel stainless steel is inherent when chromium levels are in excess of11.5%.
A passive or inert surface is established naturally with these steels due to exposure to air, aerated water, and
other oxidizing atmospheres.
Airborne impurities, heating, and other direct contact materials can damage this protective film causing themetal to be compromised in its ability to ward off the corrosive process.
11.9.6.1 Passivation of Stainless Steels to Restore Protective Film
Passivation is the process that establishes a surface or film on nickel-chrome steels and maximizes thecorrosion resistance.
The passivation process cleans the exposed surfaces of the contaminates, soils, and surface impurities thatcause film damage. The process forms a new strong protective film in areas that have become “active” orsensitized due to welding.
11.9.6.2 Rouging or Ferric Oxide Discoloration of SS Systems
Rouging is seen in many water systems, but is more prominent in hot, distilled, and clean steam systems.Rouge can be in the form of dust or a light film that can be wiped off. Rouge also can be in the form of abonded multi-layer that requires scraping with a sharp tool to remove.
Rouge is found in many forms:
• Orange - found in high purity/high temperature systems
• Light Red - found in high purity/high temperature systems
• Red - found in high purity/high temperature systems
• Reddish Brown - found in high purity/high temperature systems
• Purple - found in clean steam and high temperature water systems
• Blue - found in clean steam systems
• Gray - found in clean steam systems
• Black - found in clean steam systems
See appendix section on passivation for more details.
11.9.7 Castings and Forgings for Vessel Components
Most castings and forgings are used for components attached to vessels, such as agitator impeller hubs andinstrument housings. Castings typically contain higher levels of ferrite that can cause potential problems withcomponents in contact with high purity water (rouging). In addition, castings are usually more porous orgrainier, and will not typically take a polish higher than a 4 at best. Forgings are less susceptible to rougingand can be polished up to a 8 mirror finish.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 226/250
202
APPENDIX
11.9.7.1 General: Sanitary Tubing and Piping
• The piping or tubing and valve ends to be joined by welding must be within the set tolerances as far asinside diameter and out of roundness is concerned.
• All the welds should be done with automatic welding machines (orbital welding) with insert gas (Ar)
blanket where possible.
• If unavoidable, the last weld of an assembly may be done manually by a qualified, certified welder.
• All welds should be documented and inspected.
• If orbital welding is done, a video tape (borescope) can be prepared for random welds; however, this isoptional. Also, an isometric or piping drawing should be marked identifying each weld by a unique num-ber, date, and welder ID number.
• The piping (tubing) installation should be done according to approved piping drawings (orthographic and/ or isometrics).
• The piping support method and the support spacing should be in accordance with the piping specifica-tion and applicable drawings. It is very important that the installing contractor follows the design docu-ments and does not use shortcuts as the piping support spacing is based on calculations to assure thepermitted deflection (sagging) of the pipe between the adjacent supports.
• As the piping system larger sections are installed, the piping installation contractor shall make up thelatest issue of the piping drawings verifying that the system is installed in accordance with the applicabledrawings, ensuring all valves, fittings, etc., are installed. If there are deviations from the drawing, thedrawing shall be marked up for the preparation of an “as-built” final revision. The “as-built” revision of thedrawing should be approved by the user. The user should have access to the “as-built” drawings. Thequalification protocol completion report should make reference to the “as-built” drawings.
• The installed piping system should be pressure tested according to the requirements of the piping fabri-
cation specification. During the performance of the pressure test, the pipe slope should be checked anddocumented. During the pressure testing, the piping is not insulated, but is full with water - a conditionclosest to the operating condition. If there are deviations from the required slopes due to faulty installationdue to a larger than specified pipe support span, the installation needs to be corrected. The main purposeof the pipe slope is the assurance of self drainage. The results of the pressure test and slope measure-ments should be documented.
After the piping (tubing) system installation, pressure testing, and pipe slope parameters are satisfied andaccepted, the system is ready for passivation.
11.9.8 Surface Finish of Stainless Steels
11.9.8.1 Cold Rolled Stainless Steel Finishes
300 series stainless steel “sheet” may be produced with a very fine grain, slightly milky appearance throughto a bright highly reflective finish produced on mirror polished rolls.
These terms are used by the steel mills without a good method of quantifying the surface quality, roughness,and texture.
Mills also relate to the ASTM or AIAI finishes which range from 1 to 8, where 8 is a “mirror finish.” This methodalso is subjective and relates to the method used to obtain the finish.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 227/250
203
APPENDIX
11.9.8.2 Stainless Steel Polishes and Improved Finishes
A third identification method commonly used to define surface finishes is the “Grit Finish” which is the numberof scratch lines per inch of surface produced by an abrasive wheel or belt.
While this system qualifies the finish and allows acceptance based on measurable criteria, the criteria is not
a true measure of the surface quality as it does not define the depth of the “scratch lines.”
Grit finishes come in grades related to the standard abrasive tools or surfaces:
20 grit = 100+ RMS and 90+ RA
180 - same as a 2B finish - = 80 RMS and 70 RA approximately
240 - between a 2B and a 3A finish - = 15 to 63 RMS and 14 to 57 RA
320 - same as an 8 finish - is 10 to 32 RMS and 9 to 29 RA
500 - same as a 9 finish - is 4 to 16 RMS and 4 to 14 RA
11.9.8.3 Electro-Polishing of Stainless Steel
Mechanical finishing has certain inherent deficiencies, one of which is the tendency to enlarge the exposedsurface area.
Electro-polishing is able to improve the mechanical finish by rounding off the sharp peaks of the “scratchlines.”
The advantages of electro-polishing are:
• reduces the surface areas
• provides a sanitary acceptable surface finish
• cleans the surface
• passivates the surface of stainless steel with a chromium layer
• removes impurities trapped below folded layers of mechanically formed ridges
• reveals defects that have been hidden by mechanical polishing through smearing effect
Electro-polishing requires a mechanical polish preparation developed with a uniform progressive grit polish-ing application.
11.9.9 Stainless Steel Distribution Piping
Stainless steel pipe is available in heavy gage “schedule” piping, thin wall solid drawn tubing, and thin wallseamed tubing.
11.9.9.1 Solid Drawn Thin Wall Tubing
The selected materials are available in solid drawn tube suitable for the type of service expected from a highpurity water distribution system.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 228/250
204
APPENDIX
This is an acceptable material for a high purity water loop and should conform to ASTM A632.
Tubing should be 16 gauge (.065 in./0.165 cm) for tubing diameters of between 1 and 3 inches (2.54 and 7.62cm) inclusive and 14 gauge for 4 in. (10.16 cm) diameter.
11.9.9.2 Welded Seam Thin Wall Tubing
Acceptable thin wall tubing is available at a reduced cost with a welded seam produced in an oxygen freeenvironment. (TIG)
The tubing is available with a uniform smooth surface in accordance with ASTM A269 and may be polishedmechanically and electrically to a suitable surface finish.
0.065 in. (0.165 cm) thick sheet metal is used for the manufacture of seamed tubing between 2 and 3 inches(5.08 and 7.62 cm) diameter with .083 in. thick for 4 in. (10.16 cm) diameter and progressively larger for 5 in.(12.7 cm) and above to suit the application and location. The metal is cut and cold rolled to form a round tube.The tube is clamped so that the two edges make contact. The inside and outside of the tube is blanketed withArgon gas to expel all oxygen and the seam is welded without filler, producing a consistent smooth andoxidation free surface inside and out.
11.9.9.3 Material and Installation Certifications
• mill test certificate of piping/tubing material
• weld test, spot x-ray of carbon steel piping welds, borescope of stainless steel piping/tubing welds
• pressure test, slope measurement certificate
• weld identification on piping drawings
11.9.10 Elbows/Bends In Tubing
While butt joints can be accomplished leaving a smooth surface free of pits, crevices and oxidation, welds areexpensive and should be minimized.
Current practice is to use tight welded elbows for each directional change in the distribution piping. Consider-ing the extent of high purity water distribution piping involving points of use drops from 25 in a small systemto more than 100 in a large one, the total number of elbows is considerable. These can range from 130 to 500and from 60 to 1000 welds to install them.
The purchase of longer lengths than the standard 20 ft (6 m) lengths (say 40 ft (12 m) lengths) and theutilization of long sweeping bends, instead of the traditional tight welded elbows can reduce the total numberof welds by up to 60% and the total pipe used by 10%.
High purity water systems, in order to maintain flow at all times without stagnation, water must be pumpedcontinuously with the associated energy requirements.
Any reduction in these requirements due to frictional improvements of large sweeping bends compared totight traditional bends are added to the advantages of the change from elbows to bends.
Sweeping bends may be accommodated in many areas and with careful design, the advantageous could beexploited.
Large sweeping bends are preferred in order to reduce the potential of wrinkling and/or damage to theinternal pipe finish.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 229/250
205
APPENDIX
11.9.11 Types of Non-Metallic Materials
Few compatible non-metallic piping materials are available that will withstand the rigors of a pharmaceuticalhigh purity water system, such as:
• 80 to 90°C operation or periodic sanitization
• 121°C steam sterilizing temperatures
• ozone and or chlorine contact
One such material that will suppor t the above requirements and limit extractable to a minimum level compat-ible with stainless steel is polyvinylidene fluoride (PVDF).
This material is available in a compatible range of pipe diameters, surface finishes, and automated orbitalwelding capability.
11.9.11.1 Corrosive Resistance of PVDF
PVDF is inert and will not exhibit any surface corrosion when in contact with 90°C high purity water orcommonly used oxidizers.
11.9.11.2 Weldability and Polishability of PVDF
The surface finish of PVDF is equivalent to electro-polished stainless steel and the fusion welding equipmentand capabilities are similar to stainless steel orbital welders.
11.9.11.3 PVDF Distribution Piping and Fittings
Weldable fittings, elbows, tees, reducers, adapters, diaphragm valves, zero static valves, flow meters, regu-lators, etc. are available for PVDF pipe.
11.9.11.4 Pressure Rating of PVDF Piping Systems
Rated pressures for PVDF piping ranges from 230 psi at 68°C to 50 psi at 149°C.
Due to softening at elevated temperatures, continuous support is recommended for systems that are oper-ated at 65°C or above.
11.10 DESIGN OF A WFI/PURIFIED WATER DISTRIBUTION SYSTEM
The layout and general design of a high purity water system (WFI or purified water) should be consistent andfollow good manufacturing guidelines in respect to installation, support, natural drainage, flow rates, dead orstagnation areas (dead legs), and minimization of areas that may promote micro-organism growth.
11.10.1 Fittings and Equipment
All equipment that, when installed, comes in direct contact with the high purity water should use a suitablestainless steel or non-metallic material for all contact areas, except for valve diaphragms and tri-clamp gas-kets.
This includes all valves and monitoring sensors.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 230/250
206
APPENDIX
Valve design should be sanitary diaphragm type, installed at approximately 60°C to the vertical to optimizedrainability. Hash marks or drain dots are provided on many designs to allow precise orientation of the valvesin horizontal installations.
PTFE Teflon faced diaphragms are preferred for hot water applications, covering a temperature range from -30 to 350°F (0 to 177°C).
Connecting flanges for all valves and fittings should be tri-clamp design.
11.10.2 Natural Drainage
Well designed distribution loops should be installed overhead and point of use valves should be located at aconvenient location within the process or user area, below the distribution loop.
Slopes should be verified externally after hanging and before insulation is installed.
Hangers should use steel clamps/fittings with Viton or similar inserts designed for use with the particulartubing, installed for SS at least on 10 ft (3 m) centers. Support for PVDF piping should be appropriate for theoperating temperatures. Continuous support of plastic pipe should be considered using angle iron or equiva-
lent to prevent slope changes due to expansion and contraction and point stresses from hangers.
Piping distribution from an unavoidable low point in the system, such as the storage tank discharge elbow,should be designed with a clear low point and a sanitary drain valve installed with a maximum of 6 diameterdead leg (no less than 6 diameters from the centerline of the main line to the valve center line based on themain line diameter).
Slope of distribution system process pipe or tubing should be a minimum of 1/16 in./ft (0.52 cm/m).
Location of all drain valves should be totally and easily accessible.
11.11 FABRICATION OF A WFI/PURIFIED WATER DISTRIBUTION SYSTEM
High purity water distribution systems using the material and finishes specified above must be joined usingacceptable welding or other sanitary techniques.
11.11.1 Assurance of Quality Distribution Piping
The distribution piping and storage systems should be installed in accordance with cGMPs and should befabricated, manufactured, procured, and installed in strict accordance with explicit operating procedures.
11.11.2 Operating Procedures Should Include the Following:
Purchase of components for the distribution and storage system from a list of selected/preferred vendors.Procurement of specific components based on preferred part numbers, listed for each vendor where appli-cable. Procurement of other components, such as valves, pumps, filters, instruments, and tanks etc. usingapproved specifications.
Inspection of all components of the system on receipt for compliance and/or damage. Verification of the bill oflading using a detailed list of purchased parts.
The qualification of welding procedures in accordance with a detailed list of set-up parameters for:
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 231/250
207
APPENDIX
a) Receiving, fabrication, cutting, end facing, and welding to be performed in a clean area devoid of anyequipment previously used on “carbon” steels.
b) Tubing specification, heat lot, diameter to be used for the specific application
• Check tubing for crushing and confirm the nominal diameter.
• Examine interior and exterior of tubing for physical damage, scratches (interior) dirt or grease.
• Clean tubing of grease and dirt using an approved cleaning solvent isopropyl alcohol (IPA) and lintfree cloth. Repeat cleaning until lint free cloth is visually free of grease or dirt.
c) Length of tubing and facing specifications.
Cut tubing to length and finish with a square deburred butt end or as recommended by the fusion weldingequipment supplier.
11.11.2.1 Cleaning
• Place a lint free cloth in tubing prior to facing and/or deburring.
• Use only Isopropyl alcohol as a cleaner solvent for SS and other appropriate solvent for PVDF. Usesparingly and ensure that all traces have evaporated prior to bagging.
• Use only lint free cloths such as “Texwipe.”
• Clean both inside and outside of tube ends and do not touch the end area with bare hands or soiledmaterial.
11.11.2.2 Protection
a) Bag the prepared tubing or the ends of the tubing with clear plastic bags and tape the bag to the tubing if
the tubing is not to be welded within two hours or if the tube is to be welded outside the clean fabricationarea.
b) Weld fittings, elbows, tees, etc. should be examined similar to the tubing and cleaned of all dirt or greaseprior to welding. All weld fittings should be received, sealed in a plastic or other lint free, non-metallicsealed bag.
c) The fitting should remain in the bag until just prior to welding or for inspection following receipt.
Valves and instruments, etc. should be procured with tri-clamp or suitable sanitary type fittings and matingfittings should be welded onto the tubing in accordance with the system design. Receipt should be in a totallyenclosed sealed lint free bag, similar to the weld fittings.
Welding of tubing, weld fittings, tri-clamp, or suitable sanitary type fittings should be performed inside thefabrication cleanroom, using TIG orbital welding for SS throughout and PVDF fusion machine for PVDF.
Purging of each assembly with Argon Purge gas should be completed for all SS prior to welding. The gasshould be 99.999% pure and contain less than 1ppm of moisture or 2ppm of oxygen.
The complete removal of all air should be accomplished prior to commencing welding.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 232/250
208
APPENDIX
Welder qualifications should be performed for all welders and re-qualification at the commencement of eachwork-day and/or the change to a different diameter.
11.11.3 Clean Preparation Area
A cleanroom/trailer should be used to ensure that welding, bending, and fabrication of high purity water piping
are not contaminated with metal or non-metallic particulates.
The area should be filtered to control particulates with limited controlled access. The following proceduresshould be followed:
• All cutting and welding equipment should be cleaned of any impurities or particulates.
• The room should be maintained clean at all times and access should be across sticky floor mats.
• All piping and fittings should be cleaned prior to and following welding.
• Assemblies should be bagged following welding and the bags removed just prior to making field welds.
• Welding and fitting technicians should use overshoes and cleanroom gloves. The gloves are worn toprevent finger print oils from getting onto the pipe.
The above procedures are currently typical for a high purity piping system for the semi-conductor industryand is considered advisable for pharmaceutical installations with similar requirements.
Inspection of completed welds is sometimes not possible or practical. Remote borescope inspection is limit-ing and often not precise. Common practice and the practice recommended is to provide a validated processfor all welds, a precise instrument to make the welds, an inert environment around the weld and a cleancontact surface. These goals may not be accomplished if the environment for storage and the welding pro-cess is not controlled. Skin oils and metal particulates will contaminate the surfaces of the mating tubing. Themating surfaces must be cleaned and maintained clean through the welding process to ensure a contamina-tion free weld and therefore a validated weld.
The use of a clean area and clean area procedures will protect the welds and reduce the defect incidentsassociated with contaminates in the welding areas. (This approach has been shown to be cost effective bythe cost sensitive semi-conductor industry.)
Spot checks of the outside and inside surfaces on a percentage or time basis will provide assurance of thesystem efficiency.
11.11.4 Joints Using Fusion Butt Welding
Orbital welding is the recommended industry standard for high purity water systems in biotech, pharmaceu-tical and semi-conductor industries. This is due to the smooth inner weld bead that is characteristic of this
joining process.
11.11.4.1 Orbital Welding
Orbital welding is a welding method used for joining stainless steel piping and tubing with electric arc in awelding machine not using any filler (welding rod) under inert atmosphere argon (Ar). Electric power charac-teristics are displayed on the control panel instruments and, if required, a printout of the welding parameterscan be obtained.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 233/250
209
APPENDIX
11.11.4.2 Borescoping
A borescope is a fiber optic instrument with a monitor (TV screen and optional video tape recorder) used forinspecting the weld inside the pipe or tubing. The fiber-glass cable is a maximum 20 ft. long with a light andlens at the end. Borescoping is done for verification of the weld quality.
With the control of the welding process and the processing area, borescope inspection of the inside of theweld is only considered necessary as a means of control and is recommended as part of the qualifyingprocess for the weld size and daily restart checks. (See details later in this Section.)
Video recording each weld is considered un-necessary and difficult to justify.
11.11.5 High Tech. Orbital Tube and Pipe Welding for Stainless Steel
Semi-automatic, programmable, orbital, inert gas purged welding equipment for both tube and pipe is avail-able from many manufacturers. The basic principals and techniques are the same for all commercial ma-chines.
The system is an automatic “Heliarc” or “TIG”(Tungsten Inert Gas) process where Inert Argon gas from a
cryogenic or low temperature source is used to protect the molten steel from oxidation during the metalfusion.
Liquid argon in Dewers is preferred due to its high purity at 99.999%. A gas purifier such as the Nanochemshould be used during purging to bring the contaminants down to the ppb levels which is preferred.
The above equipment will weld stainless steels, nickel based alloys, titanium, and aluminum.
11.11.5.1 Accessories
Welding accessories are available to assist in set-up, alignment of pipe butts, purging, fitting alignment, pre-tacking fixtures, and finishing of the pipe ends in preparation for welding.
11.11.6 Weld Criteria
• Weld quality must meet the strictest standards.
• All welds must be fully penetrated around the entire weld perimeter with no crevices or entrapment sites.These areas are particularly vulnerable to crevice corrosion.
• All welds should be smooth, uniform, complete and flat, not concave, on the outside.
• The weld should have a uniform and complete weld bead width on the inside with little or no convexity.
• The inner weld bead should contain no concavity.
• There should be no visible signs of oxidation/discoloration of the inner weld.
• The joints should be square facing and properly aligned. Tubing surfaces should not be offset in anyplane or direction by more than +/-0.003 in.
• The weld width should be nominal 1/8 in. wide.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 234/250
210
APPENDIX
11.11.6.1 Welder Qualification
There are many types of metals and even more methods of good joining of metal parts by welding.
Each welder performing welding on certain metals must be qualified for this type of metal and welding. TheAmerican Society of Mechanical Engineers (ASME) and the American Welding Society (AWS) have welding
procedures for welder qualifications. Each qualified welder should be formally qualified with a certificatequalifying him or her for a certain type of weld.
As a general rule, welders and welding procedures are qualified to ASME B 31.3 (chemical plant and petro-leum refinery piping), that incorporates ASME Section IX by reference.
11.11.6.2 Common Test Methods for Installed Piping
Test methods on installed piping are: pressure testing, Verification of slope and the last step of passivationrequiring the testing of the final rinse water.
11.11.7 Weld Defects Examples
11.11.7.1 Joint Misalignment
Poor alignment of joint caused by equipment or procedure failure.
Follow GMP welding procedures/use tube alignment gauge.
11.11.7.2 Lack of Penetration
Weld bead does not completely penetrate the ID of tube caused by poor setting of weld program parameters,material thickness and/or composition did not allow for complete penetration or improper welding amperagecaused by power fluctuations.
Develop proper welding parameters/adjust program to compensate for thickness or material/adjust for am-
perage fluctuation.
11.11.7.3 Excessive Penetration
Weld bead over penetrates causing excessive concavity or spikes caused by incorrect weld program (toohot), excessive welding amperage caused by power fluctuations or material thickness and/or compositionwhich, caused excessive penetration.
Develop proper welding parameters/adjust program/use dedicated circuit.
11.11.7.4 Lack of I.D. Purge
Discoloration/oxidization in the heat effected zone caused by a lack of ID purge or the use of an impure purgegas.
Sugaring of the weld bead caused by a lack of ID purge.
Monitor the purge-flow at the weld site and ensure the hose is not pinched.
11.11.7.5 Interrupted Purge
A short segment of discoloration/oxidation and excessive penetration caused by a momentary purge gaspressure drop commonly caused by a pinched gas line.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 235/250
211
APPENDIX
Monitor purge-flow and ensure the gas hose is not pinched.
11.11.7.6 Purge Gas Purity Problems
Coloration/oxidation along the weld area caused by impurities in the weld purge gas or in or on the metalsurface.
Eliminate purge and surface impurities and check for excessive concentrations of oxygen and moisture.
11.11.7.7 Heat Tint/Haze
Light coloration/haze along weld area caused by minor impurities in weld purge gas or metal surface.
Eliminate impurities in weld zone.
11.11.7.8 Oxidation/Discoloration
Discoloration ranging from blue/brown haze to dark blue with black edges caused by oxygen contaminatedpurge gas, incomplete sealing, or insufficient pre-purge.
Ensure the complete sealing of the purged line and monitor the purge flow and purity.
11.11.7.9 Insufficient Tie-In
Weld on OD is good, but the ID is incomplete and does not overlap weld start caused by incorrect speed ortime in the weld program. (Uncalibrated weld head.)
Use high quality welding equipment and check for appropriate procedures.
11.11.7.10 Electrical Current Fluctuations
Significant narrowing of ID weld bead width and/or lack of penetration of tube ID in affected areas caused by
electrical power fluctuations (not using a dedicated circuit).
Ensure that a dedicated circuit is used.
11.11.7.11 Inclusions/Dross(non-metallic substance)
Non-metallic formation on ID weld caused by tungsten inclusion.
Use high quality base material for welding rod.
11.11.7.12 Pinholing
Small holes in the weld bead caused by the weld puddle being cooled to rapidly.
Ramp (feather) the welding amperage properly and provide post purge.
11.11.7.13 Porosity
Metallic or non-metallic impurities trapped in the weld bead (tungsten or slag) caused by the introduction ofimpurities from outside or from the weld material.
Eliminate impurities from outside or from the metal.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 236/250
212
APPENDIX
11.11.7.14 Cracking
Cracks in either the surface of the weld caused by brittle metal or use of a very hot weld program with rapidcooling.
Use only good quality materials and acceptable welding programs. Do not quench or cool the weld.
11.11.8 High Tech. Orbital Tube and Pipe Welding for PVDF Piping
Semi automatic fusion welding machines are available for orbital welding of PVDF piping.
An elastic pressure element backs up the inside of the fusion zone, totally avoiding an inner bead.
A controlled pressure thermoplastic melt optimizes the homogeneous joint between the plastic parts, produc-ing a consistent joint quality.
11.11.9 Remedial Action for Defective Welds - SS or PVDF Piping
Rewelding or weld repairs are not acceptable.
Weld Replacements
The following is a procedure for replacing defective welds:
a) Cut out rejected weld using a band or cut-off saw.
b) Where possible, retain section for further inspection.
c) Prepare faces of tubing and check for signs of discoloration.
d) Ensure that all traces of any heat effected areas are removed.
e) Cut and shape pipe ends using proprietary facing equipment.
f) Debur and reweld.
g) Assign a “R” suffix to weld number.
11.11.10 Joints Using Sanitary Clamps
The sanitary clamp system of quick disconnect fittings are designed to provide a smooth non-contacting ornon-corrosive environment.
The sanitary joints may be cleaned in place, provide leak tight connections and may be adapted to otherforms of piping.
These joints should be used for all distribution pipe connections to valves, sensor housings, and fittings notadaptable to orbital fusion welding.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 237/250
213
APPENDIX
11.12 ABBREVIATIONS AND DEFINITIONS
11.12.1 Definitions
ABS - A plastic material used to make pipe; based on combinations of acrylonitrile, butadiene and styrene;ABS is a relative inert material and contributes little in the way of contamination to pharmaceutical water.
Absorption - Assimilation of molecules or other substances into the physical structure of a liquid or solidwithout chemical reaction.
Aerobic Bacteria - Bacteria capable of growing in the presence of oxygen.
Anaerobic Bacteria - Bacteria capable of growing in the absence of oxygen.
As-Built Drawings (Record Drawings) - Construction drawings and specifications that represent the physi-cal condition of the plant or system at turnover from the designer or installer at satisfactory operation. Thesedocuments supplement and compliment the system manuals and protocols.
Backwash - The process of flowing water in the opposite direction from normal service flow through a filter
bed or ion exchange bed. The purpose of backwashing a sand filter is to clean it by washing away all thematerial it has collected during its service cycle. The purpose of backwashing a carbon filter is also to cleanit, but primarily to eliminate flow channels that might have formed and to expose new absorption sites.
Bacteria - Single-celled microorganisms measured in high purity water by several means: culturing, highpower microscope, or Scanning Electron Microscope (SEM). The value is reported as Colony Forming Units(CFU), or colonies per milliliter or per liter. The bacteria in the water act as particle contamination on thesurface of the product, or as a source of detrimental by-products. See Pyrogen.
Blowdown - The withdrawal of water from an evaporating water system to maintain a solids balance withinspecified limits of concentration of those solids.
BOD - Biological oxygen demand of water. This is the oxygen required by bacteria for oxidation of the soluble
organic matter under controlled test conditions.
Breakthrough - Passage of a substance through a bed, filter, or process designed to eliminate it. For ionexchange processes, the first signs are leakage of ions (in mixed beds, usually silica) and the resultantincrease in conductivity. For organic removal beds, usually small, volatile compounds (THMs are common inactivated carbon).
Calibration - A comparison of a measurement standard or instrument of unknown accuracy to detect, corre-late, report, or eliminate by adjustment of any variation in the accuracy of the unknown standard or instru-ment.
Cation Exchange Resin - An ion exchange resin which removes positively charged ions.
Certified Vendor Drawings - Drawings prepared by vendors for the fabrication of equipment, specialtycomponents, and skid mounted systems. These are certified as fabricated by the vendor and become theofficial document for the equipment involved.
Commissioning - A prescribed number of activities designed to take equipment and systems from a static,substantially complete state to an operable state.
Conductivity - A measure of flow of electrical current through water. This conductance is high with high TotalDissolved Solids (TDS) water and very low with ultrapure deionized water. Conductivity is the reciprocal ofresistivity (C=1/R) and is measured in micromho/cm (µmho/cm) or microsiemens/cm (µS/cm).
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 238/250
214
APPENDIX
Contaminant - Any foreign component present in another substance. For example, anything in water that isnot H2O is a contaminant.
Critical Instrument - These are the instruments used to measure critical parameters.
Critical Parameter - These are the measured values that would determine process compliance and cause a
system to revert to a fail-safe mode.
Drinking Water - EPA primary drinking water or comparable regulations of the European Union or Japan.
Endotoxins - Pyrogens from certain Gram negative bacteria. Generally highly toxic Lipopolysaccharide-protein complexes (fat, linked sugars, and protein) from cell walls. A marker for these bacteria with a reputa-tion for persistent contamination because they tend to adhere to surfaces. See Pyrogen.
Enhanced Documentation - Collection of Engineering, Quality Control and Regulatory Affairs documentswhich will be required for the operation, validation, maintenance, and regulatory compliance of a pharmaceu-tical plant.
EPA - Environmental Protection Agency.
Extractable - Trace material from piping and/or equipment which have been extracted by the processed fluid.
FDA - US Food and Drug Administration.
General Arrangement - A more specific version of a general layout which includes the system interfacepoints, space requirements, ergonomics, construction issues, manufacturing flow of materials and operators,maintenance requirements, and future expansion or alterations.
General Equipment Layout - A diagram that relates the unit operations of the system to one another. Itsdevelopment should depend on production requirements, product matrix, and possibilities for future expan-sion.
Good Engineering Practices (GEP) - Standards, specifications, codes, regulatory and industrial guidelinesand accepted engineering and design methods to design, erect, operate, and maintain a pharmaceuticalfacilities taking into account not only regulatory compliance, but also safety, economics, environment protec-tion, and operability. Standards and specifications are provided by recognized sources such as establishedengineering contractors and pharmaceutical companies. Codes are provided by local, state or federal juris-dictions, or insurance companies. Guidelines are issued by professional societies, industrial organizations, orregulatory agencies. Engineering design methods have been established in the engineering educationalsystem.
Grains Per Gallon - A unit of concentration. 1 grain/gal = 17.1 mg/l.
Gram Negative Bacteria - A basic classification of bacterial type, along with “Gram positive.” These organ-isms resist straining by the Gram technique. Sometimes considered “bad” bacteria when discussing pollutionor contamination; however, this is an artificial and quite broad classification.
Halogens - Atoms of the chlorine family which also include fluorine, bromine, and iodine.
Hardness - The concentration of calcium and magnesium salts in water. Hardness is a term originally refer-ring to the soap-consuming power of water; as such it is sometimes also taken to include iron and manga-nese. “Permanent hardness” is the excess of hardness over alkalinity. “Temporary hardness” is hardnessequal to or less than the alkalinity. These also are referred to as “non-carbonated” or “carbonate” hardness,respectively.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 239/250
215
APPENDIX
High Purity Water - Water conforming to USP Monographs or equivalent.
Heavy Metals - High molecular weight metal ions, such as lead. Known for their interference with manyprocesses, and “poisoning” of catalysts, membranes, and resins.
Humic Acid - The classical method for fractionating the humic colloids that disperse in the sodium hydroxide
extract is to acidify the suspension with sulphuric or hydrochloric acid, which causes a part of the dispersedorganic matter to precipitate. The part that stays in solution is known as fulvic acid, that which precipitates outas humic acid, and that part of the organic matter which does not disperse in the alkali but remains in the soilas humin.
Hydrocarbons - Organic compounds containing only carbon and hydrogen. Sometimes broadened to in-clude compounds or mixtures of compounds with small amounts of oxygen also.
Hydrophilic - Having an affinity for water. Its opposite, non-water-wettable, is hydrophobic.
Inorganics - Chemical compounds which are not organic in nature. Inorganics that are soluble in watergenerally split into negative and positive ions, allowing their removal by deionization.
Instrument List - A list of instrumentation which includes the instrument tag number, instrument name,manufacturer name, model and serial number, P&ID reference, critical or non-critical, and local or panelmounted.
Ion - An atom or radical in solution carrying an integral electric charge, either positive (cation) or negative(anion).
Ion Exchange (IX) - One of the processes used to further reduce the concentration of ions in water suppliesreferred to as total dissolved solids removal. The process uses anion and cation exchange resin to chemicallyreact with and remove the remaining ions or TDS in the water. This process results in water with virtually noTDS.
Ion Exchange Regeneration - The process by which ion exchange resin that can no longer effectively
remove ions from the water is recharged. This recharging or regeneration is performed by adding an excessof caustic (NaOH) to the anion resin and an excess of either hydrochloric acid (HCl) or sulfuric acid (H2SO4)to the anion resin. These regenerant solutions are allowed to flow through the resin beds at specific flow ratesfor specific periods of time depending on the type of resin, the ionic load, and the final purity desired. Theregenerant solutions react with the ion exchange resin releasing the removed cations and anions which arethen carried away to drain by the flow of the regenerant chemicals. The excess chemical is rinsed from the ionexchange resin with purified water when the bed is ready for another service cycle.
Ion Exchange Resin - A styrene-divinylbenzene or acrylic copolymer formed into small, spherical, andhighly porous beads about the size of a pinhead. These inert beads are chemically treated so that theyperform as if they were chemical compounds.
Langelier Index - A means of expressing the degree of saturation of a water as related to calcium carbonatesolubility.
Membrane - A barrier, usually thin, that permits the passage only of particles up to a certain size or of specialnature.
Micron - The same as a micrometer or 1000th of a millimeter. The typical particle size of importance indeionized water is less than 0.2 µm.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 240/250
216
APPENDIX
Microorganism - Organisms (microbes) observable only through a microscope. Larger, visible types arecalled organisms.
Milligrams Per Liter (mg/l) - A term used to report chemical analyses. Milligrams per liter refers to themilligrams of the compound or element present in 1 liter (1000 milliliters) of water. Another term often used isparts per million (ppm) which is the same for substances in water. 1 mg/l = 1000 ug/l = 1 ppm.
Mixed Bed Ion Exchange - The use of both cation and anion exchange resin mixed together in one tank.
MSDS (Material Safety Data Sheet) - Document produced by the manufacturer that contains the chemicaland physical properties of a substance that are pertinent to safe handling and storage.
NPDES Permit - The National Pollution Discharge Elimination System permit required by and issued by EPA.
Organics - Short for organic chemicals; those compounds that contain carbon to hydrogen bonds and arenot carbonate related.
Orifice - An opening through which a fluid can pass; a restriction placed in a pipe to provide a means ofmeasuring flow.
Osmosis - The passage of water through a permeable membrane separating two solutions of different con-centrations; the water passes into the more concentrated solution.
Oxidizer - A chemical which readily oxidizes more reduced substances. Examples of strong oxidizers areozone, hydrogen peroxide, chloride, persulfates, and oxygen itself.
Ozone - Ozone is a very strong gaseous oxidizing agent. It is used in deionized water systems to kill bacteriaand to reduce, by oxidation, the amount of TOC in the water. Ozone is O3 and due to reaction with other thingsrapidly becomes oxygen (O2). Therefore, it has a short but effective oxidizing potential. It can be destructive toion exchange using membrane filters and other plastic materials in the system.
Particles - A physically measurable contaminant in deionized water. Particles can be bacteria, colloidal ma-
terial or any other insoluble material. Particle counts are reported as number of particles per liter of a particu-lar size measured in micrometers (microns).
Passivation - The means of obtaining the loss of chemical reactivity exhibited by certain metals under spe-cial environmental conditions. More specifically, the state in which a stainless steel exhibits a very low corro-sion rate. Passivation generates an oxide film that covers and protects the surface of the metal.
Pasteurization - A process for killing pathogenic organisms by heat applied for a critical period of time.
Pathogens - Disease-producing microbes.
Permeability - The ability of a body to pass a fluid under pressure.
Piping A cylindrical device used for the conveyance of fluid that is sized by nominal outer diameter dimen-sion.
pH - pH, the negative log of the hydrogen ion concentration, is a measure of the concentration of hydrogenions (H+) in a water-based solution. The more hydrogen ions that are present, the lower the pH and the moreacidic the solution.
Photo Oxidation - The mechanism by which ultraviolet light reduces Total Organic Carbon (TOC) to CarbonDioxide. If halogenated organics are present, both CO2 and mineral acids can be formed.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 241/250
217
APPENDIX
Polypropylene (Polypro) - A plastic material used to make pipe; thermoplastic member of polyolefin familyof plastics; lightest plastic known; polypro is a relatively inert material and contributes little in the way ofcontamination to pharmaceutical water.
Polyvinyl Chloride (PVC) - A plastic material used to make pipe that is used extensively with water. Prone toproduce extractables during start-up in high purity water.
Polyvinylidene Fluoride (PVDF) - A plastic material which is used to make pipe for the distribution of phar-maceutical waters. PVDF is a relatively inert material and contributes little in the way of contamination topharmaceutical water.
Precipitate - An insoluble reaction product; in an aqueous chemical reaction, usually a crystalline compoundthat grows in size to become settleable.
P&ID (Process and Instrument Diagram) - This diagram illustrates schematically, the detailed piping, elec-trical, and control requirements of the system.
Process Flow Diagram (PFD) - A schematic of the system which utilizes graphic symbols and text to illus-trate the steps of an operation in proper sequence. A PFD should present a detailed, accurate, and ordered
flow of raw material or ingredient through each manufacturing phase.
Purified Water - USP Purified Water prepared from water complying with the quality attributes of “DrinkingWater” with conductivity in accordance with stage 1, 2 and 3 tests and the following tables. Total OrganicCarbon at 0.5 mg/l. Less than 100 CFU/ml (10,000 CFU/100 ml) for FDA microbiological acceptability.
Pyrogen - Trace organics which are used as markers of bacterial growth or contamination. Produced byvarious bacteria and fungi. Critical pharmaceutical and biotechnological processes have restrictions on con-tamination by these substances, usually at levels near the limit of detection. Primarily polysaccharide (madeof linked sugars) in nature. Fever producing substances when administered parenterally to man and certainanimals.
Resistivity - The measure of the resistance to the flow of electrical current through high purity water. This is
measured in millions of ohms-cm or Megohm-cm (Mohm-cm). Resistivity is the reciprocal of Conductivity(R=1/C, 1 Mohm-cm = 1 µS/cm). This provides an easy means of continuously measuring the purity of verylow TDS water or ionic concentration.
Reverse Osmosis - A process that reverses (by the application of pressure) the flow of water in the naturalprocess of osmosis so that it passes from the more concentrated to the more dilute solution. This is one of theprocesses used to reduce the ionic TDS, TOC, and suspended materials of feed water through a semiperme-able membrane leaving dissolved and suspended materials behind. These are swept away in a waste streamto drain.
Rouge - Rouging is a form of surface corrosion that occurs in some stainless steel water systems.
Salinity - The presence of soluble minerals in water.
Salt - Neutral compound formed of two or more ions. The salt disassociates into cations and anions whendissolved in water.
Sanitary Design - A system of design that meets standard, specification, codes, regulatory and industrialguidelines, and acceptable engineering design methods to reach a degree of sanitation required by food,pharmaceutical, and cosmetics processing.
Saturation Index - The relating of calcium carbonate to the pH, alkalinity, and hardness of a water to deter-mine its scale-forming tendency.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 242/250
218
APPENDIX
Scale - The precipitate that forms on surfaces in contact with water as the result of a physical of chemicalchange.
Sedimentation - Gravitational settling of solid particles in a liquid system.
Silica - Silicon dioxide (SiO2) and it hydrated forms, classed as reactive and non-reactive. Generally, reactive
silica is removed by the anion exchange resin. Reactive silica is only slightly ionized and is held lightly by theanion resin. It is for this reason that silica is the first thing to break through when the resin nears exhaustion.Non-reactive silica is generally considered to be particulate (colloidal) in nature.
Soda Ash - A common water-treatment chemical, sodium carbonate.
Softening - The removal of hardness (calcium and magnesium) from water. This is a PRETREATMENTprocess which used cation exchange resin to remove the hardness elements from the water. The hardnesselements are calcium and magnesium. The cation resin is regenerated with sodium chloride and during theexchange process, the calcium and magnesium are removed from the water and replaces with sodium ions(Na+). The resulting sodium salts are much more soluble than the salts of calcium and magnesium and do notprecipitate which provides better feed water to the RO system.
Soluble Silica - The silica present in the water that has actually dissolved in the water.
Stability Index - An empirical modification of the saturation index used to predict scaling or corrosive tenden-cies in water systems.
Stainless Steel - Steel to which a significant amount of chromium and nickel has been added to inhibitcorrosion.
Start-Up - The initial operation of equipment to prove that it is installed properly and operates as intended.Start-up is considered complete when the selected equipment will adequately process water as specified.
Sterilization - Refers to the killing of microorganisms in the distribution system. This is normally done peri-odically by flushing a sterilizing solution, such as hydrogen peroxide or ozone, through the distribution piping
system. In some systems, ozone is continuously injected at low levels for continuous sterilization.
Surface Water - Surface water is any water where the sources is above ground. This can be rivers, lakes, orreservoirs. Surface waters are usually higher in suspended matter and organic material and lower in dis-solved minerals than well water.
Thermal Fusion - The joining of two materials (usually metal or plastic) by use of heat only, without anyadditional material. Usually done by the use of automatic TIG welding in alloy steel tubing welding or withspecially designed melting equipment for plastics.
Total Dissolved Solids (TDS) - The term used to describe inorganic ions in the water. Usually measured bymeasuring the electrical conductance of the water corrected to 25°C.
Total Organic Carbon (TOC) - Measure of organics in water by their Carbon content. This is one of theparameters used to determine the purity of Semiconductor Grade water. Feed water will have TOC measuredin parts per million. UPW will have TOC measured in parts per billion (ppb).
Trihalomethanes (THM) - Compounds present in the feed water that are formed by the reaction of chlorineand the organic material in the water. The most common THM found in water is chloroform which is quitedifficult to remove. Activated carbon and degasification can serve to reduce THMs.
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 243/250
219
APPENDIX
Tubing - A cylindrical device used for the conveyance of fluid that is sized by its inside diameter dimension.
Turbidity - A suspension of fine particles that obscures light rays, but requires many days for sedimentationbecause of the small particle size.
Ultrafiltration - Filter technology similar to reverse osmosis that is capable of filtering colloids and large
molecular weight organics out of the water. The filter capability of ultrafiltration filters to 0.005 µm particle size.Ultrafiltration also will remove organic material down to about 1,000 - 10, 000 molecular weight.
Ultraviolet Sterilizer (UV) - Ultraviolet lamps used to kill microorganisms in water. These can be placedanywhere in the water system. The wavelength used for control is 254 nanometers (nm).
Ultraviolet TOC Reduction - A UV source which partially oxidized organic compounds to ionic specieswhich can be removed. Relies on 185 nm radiation from “ozone producing” mercury lamps (along with 254nm germicidal radiation). Generally has a longer contact time than for sterilization alone.
USP Purified Water - See Purified Water.
Vacuum Degasification - The process of removing dissolved and entrained gases from the reverse osmosis
product water by creating a vacuum in a tower through which the RO product water flows. The degasifier maybe located before the reverse osmosis system, but the majority of the time it will be located after. The mostprevalent gas present is carbon dioxide which may be have been generated during pH adjustment of thereverse osmosis feed water. Carbon dioxide can be removed by the anion exchange resin, but that load canbe reduced by using the vacuum degasifier. The other gas of concern is the water is oxygen which also isremoved by a vacuum degasifier.
WFI - USP Water for Injection. Prepared from water complying with the quality attributes of “Drinking Water.”Prepared using Distillation or Double pass reverse osmosis. Conductivity in accordance with Stage 1, 2, and3 tests and conductivity tables. Total Organic Carbon at 0.5 mg/l. Less than 0.1 CFU/ml (10 CFU/100ml) forFDA acceptability. Less than 0.25 USP EU/ml.
11.11.2 Acronyms and Abbreviations
AC - Alternating Current
ACS - American Chemical Society
ANSI - American National Standards Institute
API - Active Pharmaceutical Ingredient (also known as Bulk Pharmaceuticals)
ASME - American Society of Mechanical Engineers
ASTM - American Society for Testing and Materials
Ar - Argon
BOD - Biological Oxygen Demand
BPC - Bulk Pharmaceutical Chemicals
Btu - British Thermal Units
CDI - Electrodeionization Deionization (US Filter)
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 244/250
220
APPENDIX
CFU - Colony Forming Units, i.e., viable bacteria
cGMPs - Current Good Manufacturing Practices
CIP - Clean-In-Place (system)
CO2 - Carbon dioxide
CS - Clean Steam
DCS - Distributed Control System
DNA - Deoxyribose Nucleic Acid
DI - Deionized, Deionizing, Deionization
EDR - Electrodialysis Reversal (Osmonics)
EDI - Electrodeionization (Osmonics and Generic)
EPA - Environmental Protection Agency
EPDM - Ethylene Propylene Diemer
EU/ml - Endotoxin Units per milliliter
FDA - US Food and Drug Administration
gpd - Gallons per day
gph - Gallons per hour
H+ - Hydrogen
HCl - Hydrochloric acid
H2CO3 - Carbonic acid
H2O2 - Hydrogen Peroxide
H3O+ - Hydroxonium (Hydronium) Ion
HCO3- - Bicarbonate Ion
H2SO4 - Sulfuric acid
KHz - Kilohertz
kW - Kilowatt
KWh - Kilowatt-hour
l - Liter
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 245/250
221
APPENDIX
l/min - Liters per minute
LAL - Limulus Amebocyte Lysate
lb - Pound
LSI - Langelier Saturation Index (or Langelier Stability Index)
LVPs - Large Volume Parenterals
m - Meter
m/sec (m/s) - Meters per second
Mb - Megabyte or Distillate blowdown discharged
Md - Mass of Distillate produced
Ms - Mass of steam consumed
MF - Microfiltration or Micro-filter
Mf - Distillate feedwater required
ME - Multi-effect (still)
µ - Micro (one millionth)
µm - Micrometer (micron)
MF - Microfiltration
mg/l - Milligrams per liter
ml - Milliliter
mm - Millimeter
MM - Multimedia Filter
MSDS - Material Safety Data Sheet
MTR - Mill Test Reports
NIST - National Institute of Standards & Technology
NDR - Nondispersive Infrared Analysis
NF - National Formulary (or nanofiltration)
nm - Nanometer
NTU - Nephelometric Turbidity Units
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 246/250
222
APPENDIX
O3 - Ozone
OSHA - Occupational Safety & Health Administration
p/ml - Particles per milliliter
P&ID - Process and Instrument Diagram
PFD - Process Flow Diagram
PLC - Programmable Logic Controller
PMA - Pharmaceutical Manufacturers Association
PP - Polypropylene
ppb - Parts per billion
ppm - Parts per million
psig - Pounds per square inch gauge
PTFE - Polytetrafluoroethylene
PVDF - Polyvinylidene
R - Performance ratio of a Distiller
Rc - Recovery ration for a still
RA - Average RMS of surface (Roughness Averager)
RO - Reverse Osmosis
rpm - Revolutions per minute
SOP - Standard Operating Procedure
SPC - Statistical Process Control
SS - Stainless Steel
SVP - Small Volume Parenterals
TC - Total Carbon
TDS - Total Dissolved Solids
THM - Trihalomethanes
TOC - Total Organic Carbon
TOX - Total Organic Halogens
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 247/250
223
APPENDIX
US EPA - United States Environmental Protection Agency
UF - Ultrafiltration or Ultra-filter
µmho - Micromho
µmho/cm - Micromho per centimeter
µS - Microsiemens
µS/cm - Microsiemens per centimeter
USP - United States Pharmacopoeia
UV - Ultraviolet Light
VC - Vapor Compression (still)
WFI - Water for Injection
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 248/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 249/250
For individual use only. © Copyright ISPE 2001. All rights reserved.
8/17/2019 _2001_99
http://slidepdf.com/reader/full/200199 250/250