methods of analysis by the u.s. geological survey national water

Methods of Analysis by the U.S. Geological Survey National WaterQuality Laboratory— Determination of Ammonium Plus Organic Nitrogenby a Kjeldahl Digestion Method and an Automated Photometric Finishthat Includes Digest Cleanup by Gas Diffusion

Open-File Report 00–170

U.S. Department of the InteriorU.S. Geological Survey

Methods of Analysis by the U.S. Geological Survey National WaterQuality Laboratory— Determination of Ammonium Plus Organic Nitrogenby a Kjeldahl Digestion Method and an Automated Photometric Finishthat Includes Digest Cleanup by Gas Diffusion

By Charles J. Patton and Earl P. Truitt

U.S. Geological SurveyOpen-File Report 00–170

Denver, Colorado2000

U.S. DEPARTMENT OF THE INTERIORBRUCE BABBITT, Secretary

U.S. GEOLOGICAL SURVEYCharles G. Groat, Director

The use of brand, firm, and trade names in this report is for identificationpurposes only and does not constitute endorsement by the U.S. Government.

For additional information write to: Copies of this report can be purchased from:

U.S. Geological Survey U.S. Geological SurveyChief, National Water Quality Laboratory Branch of Information ServicesBox 25046, Mail Stop 407 Box 25286Federal Center Federal CenterDenver, CO 80225-0286 Denver, CO 80225-0286

Contents III

CONTENTS

Abstract 1Introduction 2Analytical method 3

1. Application 32. Summary of method 43. Interferences 44. Instrumentation 65. Apparatus 76. Reagents 77. Calibrants 98. Sample preparation 109. Instrument performance 11

10. Calibration 1111. Procedure and data evaluation 1112. Calculations 1313. Reporting results 1414. Detection levels, precision, and accuracy 15

Discussion of results 17Analytical methods 17Statistical analysis 24

Conclusions 29References cited 30

FIGURES

1. Schematic showing analytical cartridge for colorimetric determinationof Kjeldahl nitrogen in filtered samples (I-2515-91) and whole-watersamples (I-4515-91) by U.S. Geological Survey methods 5

2–9. Graphs showing:2. Typical calibration plot for determination of ammonium ions in

Kjeldahl digests prepared by methods I-2515-91/4515-91 123. Duplicate data for 564 randomly selected samples determined at the

National Water Quality Laboratory by methods I-2515-91/4515-91between 2/11/92 and 2/13/93 18

4. Spike recovery in relation to Kjeldahl nitrogen concentration determinedby methods I-2515-91/4515-91 during the July –August, 1991 experiment 19

5. Concentration ranges for Kjeldahl nitrogen blind blank samplesanalyzed at the U.S. Geological Survey National Water QualityLaboratory from January 1989 through September 1999 21

6. Isopleths of nicotinic acid concentration (a), in milligrams nitrogen perliter, found in each tube in relation to the position of each tube in theblock digester (b) 22

ContentsIV

7. Kjeldahl nitrogen concentration recovered in digests of a high-suspended-solids sample, an aqueous solution of adenosine5’ triphosphate, and an aqueous solution of nicotinic acidby methods I-2515-91/4515-91 as a function of high-temperaturedigestion time 23

8. Relation between Kjeldahl nitrogen concentrations determinedby methods I-2552-85/4552-85 and I-2515-91/4515-91 in the April andJuly–August experiments for each of the four water types 26

9. Concentration differences between Kjeldahl nitrogen determined bymethods I-2515-91/4515-91 and I-2552-85/4552-85 in the April (A) andJuly–August (B) experiments for each of the four water types 27

TABLES

1. Kjeldahl nitrogen concentrations determined by methods I-2515-91/4515-91 for resolvated Heidelberg sample digests before and afterpassage through a 0.45-micrometer nylon syringe filter 6

2. Calibrant preparation protocol 103. Suggested block protocol for determination of Kjeldahl nitrogen by

methods I-2515-91/4515-91 114. Suggested tray protocol for automated determination of ammonium

ions in resolvated digests by methods I-2515-91/4515-91 135. Data used to estimate the method detection level (MDL) for Kjeldahl

nitrogen determination by methods I-2515-91/4515-91 166. Repeatability (within-run precision, May 3, 1991) data for 55 replicate

determinations of Kjeldahl nitrogen concentration in the Heidelbergsample by using methods I-2515-91/4515-91 16

7. Between-day (April 8–26, 1991) precision of digested calibrants andthe Heidelberg sample for Kjeldahl nitrogen determination by usingmethods I-2515-91/4515-91 17

8. Between-day (April 8–26, 1991) accuracy of digested U.S. EnvironmentalProtection Agency and U.S. Geological Survey reference samples forKjeldahl nitrogen determination by using methods I-2515-91/4515-91 19

9. Statistical data and regression analysis results for Kjeldahl nitrogenmethods I-2515-91/4515-91 (new methods) and I-2552-85/4552-85 (former methods) 20

10. Median difference between paired Kjeldahl nitrogen concentrationsdetermined by methods I-2515-91/4515-91 and I-2552-85/4552-85 25

11. Effect of nitrate plus nitrite concentrations on median differences betweenKjeldahl nitrogen concentrations determined by U.S. Geological Surveymethods I-2515-91/4515-91 and I-2552-85/4552-85 (April data set) 28

12. Effect of nitrate plus nitrite concentrations on median differences betweenKjeldahl nitrogen concentrations determined by U.S. Geological Surveymethods I-2515-91/4515-91 and I-2552-85/4552-85 (July–August data set) 28

13. Average concentrations of nitrate plus nitrite and ammonium for samplesubsets with nitrate plus nitrite concentrations greater than 1 milligramnitrogen per liter 29

Conversion Factors and Abbreviations V

CONVERSION FACTORS, ABBREVIATED WATER-QUALITY UNITS, ANDOTHER ABBREVIATIONS

Multiply By To obtaincentimeter (cm) 3.94 x 10-1 inchgram (g) 3.53 x 10-2 ounce, avoirdupoisliter (L) 2.64 x 10-1 gallonmicroliter (µL) 3.38 x 10-5 ounce, fluidmicrometer (µm) 3.94 x 10-5 inchmilligram (mg) 3.53 x 10-5 ounce, avoirdupoismilliliter (mL) 3.38 x 10-2 ounce, fluidmillimeter (mm) 3.94 x 10-2 inchnanometer (nm) 3.94 x 10-8 inch

Degree Celsius (°C) may be converted to degree Fahrenheit (°F) by using the followingequation:

°F = 9/5 (°C) + 32.

Abbreviated water-quality units used in this report are as follows:

µg/L microgram per litermg-N/L milligram nitrogen per liter

Other abbreviations also used in this report:

A/D analog-to-digitalASTM American Society for Testing

and MaterialsATP adenosine triphosphateFW formula weighth hourHz hertzID identificationLCL lower control limitM molarity (moles/liter)MDL method detection levelN normality (equivalents/liter)NASQAN National Stream Quality

Accounting NetworkNWQL National Water Quality

LaboratoryPC personal computerQC quality controlRSD relative standard deviations secondsp gr specific gravity

TKN total Kjeldahl nitrogenTP total phosphorusUCL upper control limitUSEPA U.S. Environmental Protection

AgencyUSGS U.S. Geological Surveyv/v volume/volumew/w weight/weightw/v weight/volume˜ nearly equal to< less than= greater than or equal to= less than or equal to± plus or minus

Abstract 1

Methods of Analysis by the U.S. Geological Survey NationalWater Quality Laboratory— Determination of Ammonium PlusOrganic Nitrogen by a Kjeldahl Digestion Method and anAutomated Photometric Finish that Includes Digest Cleanupby Gas Diffusion

By Charles J. Patton and Earl P. Truitt

ABSTRACT

The National Water Quality Laboratory(NWQL) determined ammonium plus organicnitrogen (Kjeldahl nitrogen) by using semi-automated, block digester methods for filteredand whole-water samples from 1986 untilOctober 1, 1991. During that time,phosphorus was determined by a persulfatedigestion method. In 1991, projectedincreases in demand for both tests by the U.S.Geological Survey’s National Water-QualityAssessment Program led the NWQL todevelop and validate methods for determiningboth analytes in a common digest.

This report describes a rapid andaccurate method to determine Kjeldahlnitrogen. The batch, high-temperature (blockdigester), Hg (II)-catalyzed digestion stepused in the new methods I-2515-91/4515-91 issimilar to U.S. Geological Survey methodsI-2552-85/4552-85 and U.S. EnvironmentalProtection Agency method 351.2 except thatsample and reagent volumes are halved.Prepared digests are desolvated at 220 degreesCelsius (oC) and digested at 370oC in separateblock digesters set at these temperatures,rather than in a single, temperature-programmed block digester. This approachpermits 40 calibrants, reference materials, andsamples to be digested and resolvated in aboutan hour. Ammonium ions originally presentin samples, along with those released duringthe digestion step, are determined photo-metrically by an automated, salicylate-hypochlorite Berthelot reaction procedure at a

rate of 90 tests per hour. About 100microliters of digest are required perdetermination. The upper concentration levelis 10 milligrams per liter (mg/L) with amethod detection level of 0.05 mg/L.Repeatability for a sample containing about4.1 mg/L of Kjeldahl nitrogen in a highsuspended-solids matrix is 3.1 percent.Between-day precision for the same sample is4.8 percent.

A gas diffusion cell in the air-segmentedcontinuous flow analyzer eliminatesparticulates and ions that otherwise wouldinterfere in the photometric finish. A single-channel analyzer can process the resolvateddigests from two pairs of block digesters eachhour. Statistical analysis of paired data forabout 1,500 samples determined by U.S.Geological Survey methods I-2552-85/4552-85 and I-2515-91/4515-91 during methodvalidation revealed a median concentrationdifference between the former and the lattermethods of about 0.1 mg-N/L. This result wasexpected because digestion blank concentra-tions (nearly equal to 0.1 mg/L) were notsubtracted from concentrations reported bymethods I-2552-85/4552-85. A 10-yearrecord of National Water Quality LaboratoryKjeldahl nitrogen blind blank concentrationdata also supports a step-change decrease inKjeldahl nitrogen concentrations of about 0.1mg/L after methods I-2552-85/4552-85 werereplaced by methods I-2515-91/4515-91 onOctober 1, 1991. Somewhat largerconcentration differences between the two

Determination of Ammonium Plus Organic Nitrogen2

methods were observed for a subset of about350 samples with nitrate plus nitriteconcentrations greater than 1 mg-N/L.

INTRODUCTION

From 1986 until October 1, 1991, theNational Water Quality Laboratory (NWQL)determined ammonium plus organicnitrogen (Kjeldahl nitrogen) by using semi-automated, block digester methodsI-2252-85 (filtered samples) and I-4551-85(whole-water samples). These methods aresimilar to U.S. Environmental ProtectionAgency (USEPA) method 351.2 (U.S.Environmental Protection Agency, 1993).During that time, phosphorus wasdetermined by a persulfate digestion methodsimilar to USEPA method 365.1 (U.S.Environmental Protection Agency, 1983a).In 1991 projected increases in demand forboth tests, brought on by the NationalWater-Quality Assessment program, led theNWQL to develop and validate methods fordetermining both analytes in a commondigest. Details of the phosphorus method,which was developed and validatedconcurrently with the Kjeldahl nitrogenmethod described here, can be found inPatton and Truitt (1992).

Previous work suggested thatstreamlining the Kjeldahl digestion stepimproved data quality for these tests andalso increased production capacity,decreased production costs, and loweredanalysts' exposure to corrosive and toxicchemicals. Jirka and others (1976), forexample, reported that digestion times forKjeldahl nitrogen could be reduced fromabout 4 hours to 90 minutes by halvingsample and reagent volumes prescribed inUSEPA method 351.2. Later, Bowman andDelfino (1982) demonstrated that furtherreduction in digestion time could beachieved by replacing a single, temperature-

programmed block digester with a pair ofblock digesters set at the desolvation(≈220°C) and digestion (370°C)temperatures. Methods I-2515-91/4515-91,which are described here, have been inroutine use at the NWQL since October 1,1991. They combine the Jirka and others(1976) and Bowman and Delfino (1982)digestion procedure improvements with anovel, highly robust automated colorimetricfinish (salicylate-hypochlorite reaction) forseparating ammonia in resolvated Kjeldahldigests from ionic and particulateinterferents by in-line, gas diffusion across amicroporous, polypropylene membrane. Abatch of 40 calibrants, reference materials,and samples can be digested and made readyfor colorimetric analysis in about 1 hour byusing these methods.

Data in this report result from twodistinct studies in 1991. The aim of the firststudy, conducted in April, was to assess thefeasibility of using a common (Kjeldahl)digestion procedure for Kjeldahl nitrogenand phosphorus determinations and todocument the performance of updatedphotometric finishes associated with thesedeterminations. Analytical performance ofthe newly developed methods wasdocumented, and quality-control (QC)guidelines for routine operation of the newmethods were established. In many cases,data that appear in this report relating toanalytical figures of merit— precision andaccuracy of results, method detection levels,and blank concentrations— were collectedduring the April study. At that time, allsamples received at the NWQL with testrequests for Kjeldahl nitrogen andphosphorus also were analyzed for Kjeldahlnitrogen by USGS method numbers I-2515-91 (filtered samples), I-4515-91 (whole-water samples) and phosphorus by USGSmethod numbers I-2610-91 (filteredsamples), I-4610-91 (whole-water samples).

Analytical Method 3

At that time, these methods were in theinitial stages of validation (Patton and Truitt,1992). The resulting data set consists ofabout 400 pairs of Kjeldahl nitrogenconcentrations that had been determined byUSGS method numbers I-2552-85/4552-85and I-2515-91/4515-91, which were used toassess relative performance of the twomethods.

The second study, conducted in July,August, and September, was performed toaddress U.S. Geological Survey (USGS)concerns that the full range of water types,which are commonly sent to the NWQL fornutrient determinations, was not adequatelyrepresented in the April study. During thesemonths, about 1,100 additional samples,selected by the USGS, were determined byboth methods. Results of this comparisonare included here, but typically QC andother analytical performance-related datathat closely matched those from the Aprilstudy are not.

This report describes USGS methodsI-2515-91/4515-91 for determining Kjeldahlnitrogen developed for use at the NWQL.The method was implemented in the NWQLon October 1, 1991. All aspects of methodsI-2515-91/4515-91 are described fromsample preparation through calculation andreporting of results. Precision and accuracydata are included. Method I-2515-91/4515-91 supplements other methods of the USGSfor determination of inorganic substances inwater that are described by Fishman andFriedman (1989) and Fishman (1993).

Several individuals merit specialthanks for assistance with various aspects ofwork reported here. Andrea M. Jirka (U.S.Environmental Protection Agency, KansasCity, Kan.), George T. Bowman (StateLaboratory of Hygiene, Madison, Wis.), andJack W. Kramer (Water Quality Laboratory,

Heidelberg College, Tiffin, Ohio) aregratefully acknowledged for helpfuldiscussions during the planning phases.Kramer also supplied high-particulate,agricultural run-off samples (Heidelbergsample) used as a control sample throughoutthis work. Dr. Ivan Sekerka (NationalWater Research Institute, Canada), BertinFrancoeur (National Laboratory forEnvironmental Testing, Canada), andAndrea Jirka served as colleague reviewersfor a previous draft of this report.

The authors especially thank JeffreyW. Pritt (U.S. Geological Survey), whoperformed statistical analysis on surface-water data included in this report.Assistance from several other individuals atthe National Water Quality Laboratory, theBranch of Quality Systems, the Office ofWater Quality, and the Branch of SystemsAnalysis also is gratefully acknowledged.

Former USGS chemist Earl P. Truitt(deceased) made substantial contributions tothe development of this analytical method.The principal author gratefully acknowledgesEarl's able assistance and analytical workupon which this report is based.

ANALYTICAL METHOD

Parameter and Codes:

Ammonium plus organic nitrogen,dissolved, I-2515-91 (mg/L as N):00623 Ammonium plus organicnitrogen, total, I-4515-91 (mg/L as N):00625

1. Application

This method is used to determineKjeldahl nitrogen in water, drinking water,wastewater, brines, and water-suspendedsediment. The suitability of this method fordetermination of Kjeldahl nitrogen in


bottom materials has not been investigated.The analytical range of this method is 0.1 to10.0 mg/L of nitrogen.

2. Summary of Method

2.1 Organic nitrogen is converted toammonium ions at a temperature of 370ºC in areaction medium of sulfuric acid, potassiumsulfate, and mercury (II). In principle, nitrateand nitrite are not reduced to ammonium ionsunder these conditions; USGS nomenclaturerefers to Kjeldahl nitrogen as ammonium plusorganic nitrogen to emphasize this distinction.In practice, however, nitrate and nitrite mightinterfere positively or negatively (seesection 3, Interferences).

2.2 The digestion procedure wasadapted from the method of Jirka and others(1976) and Bowman and Delfino (1982),which is identical to USEPA method 351.2(U.S. Environmental Protection Agency,1983b), except that sample and reagentvolumes are halved, as is the time required fordigestion.

2.3 An air-segmented continuous flowanalyzer is used to automate the photometricdetermination of ammonium ions in resolvatedKjeldahl digests by the salicylate analog(Reardon and others, 1966) of the Berthelotreaction (Patton and Crouch, 1977; Harfmannand Crouch, 1989). Resolvated Kjeldahldigests often contain suspended particulates(clays) and ions that interfere with thephotometric finish. Both classes ofinterferents are eliminated by means of an on-line gas diffusion cell, which consists of acontinuous-flow, parallel-plate dialyzerassembly that replaces the dialysis membranewith a hydrophobic, microporous,polypropylene membrane. Gases pass throughthe microporous polypropylene membrane;particles and ions do not. Before passage

through the gas diffusion cell, ammonium ionsin acidic resolvated Kjeldahl digests mix withthe alkaline donor stream and are converted toammonia—

NH4+

(aq) plus -OH → NH3 (g)↑ + H2O.

Inside the diffusion cell, gas-phase ammoniain the donor stream passes through thepolypropylene membrane and is trapped in theinterferent-free recipient stream (see fig. 1).

The authors determined ammonium infiltered (nominal pore size, 0.45µm) andunfiltered portions of highly turbid Heidelbergsample digests during the April experiments todemonstrate that the gas diffusion celleffectively removed suspended particles(clays) from resolvated Kjeldahl digests.(Samples were provided for this study by theWater Quality Laboratory at HeidelbergCollege in Tiffin, Ohio.) Equivalentammonium concentrations in unfiltered andfiltered sample digests are listed in table 1.Furthermore, the authors observed no surfacefouling or mechanical failure of the porouspolypropylene gas-diffusion membrane duringthe course of this work, and photometer flow-cell clogging was never a problem.

3. Interferences

3.1 As described in section 2.3, in-linedigest cleanup by gas diffusion (Seifter andothers, 1971) eliminates all particulate andpotential ionic interference in the photometricfinish. Thus, resolvated digests containingsuspended particulates (clays) do not requirefiltration prior to analytical determinations.Likewise, colorimetric reagents do not containcomplexing agents, such as citrate, tartrate, orEDTA, that are required to preventprecipitation of calcium (II) and magnesium(II) in the alkaline analytical stream ofmethods that lack a gas-diffusion cleanup step.

Analytical Method 5


Table 1. Kjeldahl nitrogen concentrations determined by methods I-2515-91/4515-91 for resolvatedHeidelberg sample digests before and after passage through a 0.45-micrometer nylon syringe filter[µm, micrometer; mg-N/L, milligram nitrogen per liter]

Concentration (mg-N/L) Concentration (mg-N/L)Juliandate Unfiltered Filtered

Juliandate Unfiltered Filtered

98 3.96 4.03 107 3.78 3.8099 4.28 4.25 107 4.15 4.0899 4.11 4.10 108 4.38 4.46

100 3.54 3.62 108 4.44 4.34100 4.26 4.28 109 4.32 4.32101 4.17 4.11 109 4.15 3.99101 4.14 4.07 113 4.12 4.25101 4.09 4.07 113 4.29 4.29101 3.95 4.02 114 3.90 3.71102 3.95 4.05 114 4.11 4.09106 4.33 4.29 115 4.05 4.27106 4.12 4.11 116 4.39 4.30106 4.16 4.20 116 4.38 4.44

Average concentration (mg-N/L):Standard deviation:Number of samples:

Relative standard deviation (percent):

4.140.20

264.88

4.140.20

264.83

3.2 Once samples have been acidified,they are subject to contamination by ammoniain the laboratory atmosphere. The digestionprocess, therefore, must be performed in ahood that is located in an ammonia-free areaof the laboratory. Other analytical orhousekeeping procedures with potential tocontribute ammonia vapor to the laboratoryatmosphere may not be performed in or nearthis hood. Avoid delays between samplepreparation, sample digestion, and digestanalyses to minimize the risk of ammoniacontamination.

3.3 Nitrate can exert both a positiveand negative interference in Kjeldahl nitrogendeterminations. As stated by the AmericanPublic Health Association (1992, p. 4–94):

During [Kjeldahl] digestion, nitrate inexcess of 10 mg/L can oxidize a portion ofthe ammonia released from the digested

organic nitrogen, producing N2O andresulting in a negative interference. Whensufficient organic matter in a low state ofoxidation is present, nitrate can be reducedto ammonia, resulting in a positiveinterference. The conditions under whichsignificant interferences occur are not welldefined and there is no proven way toeliminate the interference… .

4. Instrumentation

4.1 A third-generation, air-segmentedcontinuous flow analyzer (Alpkem RFA-300™ ) was used to automate photometricdetermination of ammonium ions in resolvateddigests. Modules in this system include a 301sampler, 302 peristaltic pump, 313 analyticalcartridge base, 314 power module, 305Aphotometer, 311 recorder, and a dataacquisition and processing system for use with

Analytical Method 7

a personal computer (PC). Alternativeprocedures to automate the photometricfinish by using flow injection analyzers orother second or third generation continuousflow analyzers also could be implemented(Patton and Wade, 1997).

4.2 Photometric data from thecontinuous flow analyzer were acquired andprocessed with a software package (AlpkemSoftPack™ ), operated on a PC equipped witha 12-bit, analog-to-digital (A/D) converterplug-in card. This A/D converter provided aresolution of 1 part in 4,095 (≈ 0.0012 voltswhen 5 volts is full scale), which isconservatively 20 times better than thatafforded by a 10-inch (26-centimeter) strip-chart recorder. The A/D converter must beable to acquire data at frequencies in the rangeof 0.5 to 2 Hz— that is, 30 points/min to 120points/min. As a general rule, data acquisitionfrequencies for air-segmented continuous flowanalyzers should match the roller lift-offfrequency of the peristaltic pump (Patton andWade, 1997)— that is, 0.5 Hz for TechniconAutoAnalyzer II ™ and 1.5 Hz for AlpkemRFA-300 equipment. Most PC-based dataacquisition and processing systems sold byvendors of continuous flow analyzers meet orexceed these specifications.

4.3 Operating characteristics for thisequipment follow:

Analytical wavelength 660 nmFlow cell path length 10 mmStandard calibration control setting ≈1.4Dialyzer1 4.3 in. (10.9 cm)Segmentation frequency 1.5 HzReaction coil volume 2 mLReaction coil temperature 37oCSample time 20 sWash time 20 sAnalysis rate 90/h, 1:1_______________1Microporous polypropylene membrane, 0.1 µm pore size(Gelman Metricel®)

5. Apparatus

5.1 Tecator Digestion System 40,Model 1016 block digesters or equivalent,which accommodate 40, 75-mL tubes, areused to desolvate and digest samples.

5.2 In this procedure, block digestersare operated in pairs. Prepared samples aredesolvated in one block set at 220°C andimmediately digested in another one set at370°C. Time required for desolvation is 30minutes and for digestion is 15 minutes.

6. Reagents

6.1 Digestion Reagents

CAUTION: Heat is produced whenconcentrated sulfuric acid is mixed withwater. Wear protective eyeglasses, gloves,and clothing. Hot sulfuric acid solutions arehazardous.

NOTE: All references to deionized watershall be understood to mean ASTM Type Ideionized water (American Society forTesting and Materials, 1995).

6.1.1 Sulfuric acid, 3.6 M: Cautiouslyadd 200 mL of concentrated sulfuric acid[H2SO4, sp gr 1.84] to ≈700 mL of deionizedwater contained in a 1-L volumetric flask withconstant mixing. Allow this solution to cool,dilute it to the mark with deionized water, andmix it well. Transfer this reagent to a plasticbottle where it is stable indefinitely at roomtemperature.

6.1.2 Mercury (II) sulfate reagent: Add25 mL of 3.6 M sulfuric acid to 4.0 g of redmercury (II) oxide [HgO, FW = 216.59]contained in a 100-mL Griffin beaker. Placethe beaker in an ultrasonic bath to speeddissolution. Use the resulting solution


immediately to prepare the digestion reagentas described in the next paragraph.

6.1.3 Digestion reagent: Add 268 g ofpotassium sulfate [K2SO4, FW = 174.27] to≈1,300 mL of deionized water contained in a2-L volumetric flask. Cautiously add 400 mLof concentrated sulfuric acid [H2SO4, sp gr =1.84] with constant mixing, and then add themercury (II) sulfate reagent. Stir the mixturemagnetically or place the flask in an ultrasonicbath to speed dissolution. Allow this solutionto cool, dilute it to the mark with deionizedwater, and mix it well. Transfer this reagentto a glass bottle or dispensing apparatus andstore it at or above 20°C to preventprecipitation of potassium sulfate.

6.2 Colorimetric Reagents

6.2.1 Sampler wash reservoir solution(≈1.3 N sulfuric acid + ≈0.3 M potassiumsulfate): Cautiously add 78 mL ofconcentrated sulfuric acid [H2SO4, sp gr 1.84]to ≈1,500 mL of deionized water contained ina 2-L volumetric flask with constant mixing.Add 54 g of potassium sulfate [K2SO4, FW =174.27 g], and after it has dissolved, allow thesolution to cool. Then dilute it to the markwith deionized water, and mix it well.Transfer this solution to plastic bottles whereit is stable indefinitely at room temperature.

NOTE: This solution has hydronium andsulfate ion concentrations similar to those ofresolvated digests. It can be used as thematrix for undigested ammonium ioncalibrants.

6.2.2 Sodium hydroxide, 5.0 M (alkalinerecipient stream reagent): Cautiously add 200g of sodium hydroxide pellets [NaOH, FW =40.00] to ≈700 mL of deionized watercontained in a 1-L volumetric flask. Swirl theflask repeatedly to dissolve its contents.Allow the resulting solution to cool, dilute itto the mark with deionized water, and mix it

well. Transfer this reagent to a plastic bottlewhere it is stable indefinitely at roomtemperature.

CAUTION: Heat is produced when sodiumhydroxide pellets are mixed with water.Wear protective eyeglasses, gloves andclothing. Hot sodium hydroxide solutionsare very hazardous.

6.2.3 Sodium hydroxide, 2.5 M(alkaline donor stream reagent): Add 50mL of 5 M sodium hydroxide solution to 50mL of deionized water contained in a 125-mL plastic bottle. Swirl the bottle to mix itscontents. Prepare this reagent daily.

6.2.4 Brij-35 surfactant (30 percentw/w): Most commercially available solutionsare satisfactory.

6.2.5 Sodium bicarbonate buffer (0.1 M,pH ≈ 10.5): Add 16.8 g of sodiumbicarbonate (NaHCO3, FW = 84.00), and 15mL of 5 M sodium hydroxide solution to ≈1,500 mL deionized water contained in a 2-Lvolumetric flask. Shake the flask vigorouslyto dissolve its contents. Dilute the resultingsolution to the mark with deionized water andmix it well. Transfer this reagent to plasticbottles where it is stable indefinitely at roomtemperature.

6.2.6 Sodium hypochlorite reagent(STOCK, 5.25 percent w/v): Mostcommercially available solutions of householdchlorine bleach containing 5.25 percentavailable chlorine, Clorox®, for example, aresatisfactory.

6.2.7 Sodium hypochlorite reagent(WORKING): Add 375 µL of stock sodiumhypochlorite reagent to 250 mL of bicarbonatebuffer reagent contained in a plastic bottle.Swirl the bottle to mix its contents. Preparethis reagent daily.

Analytical Method 9

6.2.8 Sodium salicylate/sodiumnitroferricyanide reagent (STOCK):Dissolve 150 g sodium salicylate [C7H5O3

-

Na+, FW = 160.10] and 0.30 g sodium nitro-ferricyanide [Na2Fe(CN)5NO • 2H2O, FW =297.95] in about 800 mL deionized water.Dilute the resulting solution to the mark withdeionized water and mix it well. Store thisreagent in an amber-colored glass bottle,where it is stable for several months at roomtemperature.

6.2.9 Sodium salicylate/sodiumnitroferricyanide reagent (WORKING): Add100 µL of Brij-35 surfactant per 100 mL ofstock solution and mix it well.

7. Calibrants

7.1 Primary calibrant, inorganic(1.000 mL = 2.50 mg-N): Dissolve 4.7736 gof ammonium chloride [NH4Cl, FW=53.49]previously dried at 110°C for ≈2 h and storedin a desiccator, in ≈400 mL of deionized watercontained in a 500-mL volumetric flask.Dilute this solution to the mark with deionizedwater and mix it well. Transfer this calibrantto a plastic bottle and store it in a refrigeratorwhere it is stable for several months.

7.2 Primary calibrant, organic (1.000mL = 2.50 mg-N): Dissolve 9.9500 g ofglycine hydrochloride [C2H5NO2 • HCl, FW =111.5] in ≈400 mL of deionized watercontained in a 500-mL volumetric flask.Dilute this solution to the mark with deionizedwater and mix it well. Transfer this calibrantto a plastic bottle and store it in a refrigeratorwhere it is stable for several months.

7.3 Spike solution, organic (1.000 mL= 0.1 mg-N): Use an adjustable pipet (100 to1,000 µL) to dispense 1,000 µL of organicnitrogen primary calibrant into ≈20 mL ofdeionized water contained in a 25-mLvolumetric flask. Dilute the resulting solution

to the mark with deionized water and mix itwell. Transfer this solution to a 40-mL amberglass vial and store it in a refrigerator where itis stable for about a month.

NOTE: When Kjeldahl nitrogen andphosphorus are to be determined in the samedigest, prepare mixed primary calibrant andspike solutions that contain both nitrogenand phosphorus.

7.4 Working calibrants: Use twoadjustable pipets (ranges 10 to 100 µL and100 to 1,000 µL) to dispense the volumes ofprimary calibrant, listed in table 2, and1.000 mL of field preservative solution1 (asolution containing 1.3 g HgCl2 + 10.0 g NaClin 100 mL of deionized water) into a series of250-mL volumetric flasks that each contain≈240 mL of deionized water (U.S. GeologicalSurvey Office of Water Quality TechnicalMemorandum No. 94.16, 1996; U.S.Geological Survey Office of Water QualityTechnical Memorandum No. 99.04, 1999).Rinse these flasks with a dilute solution ofhydrochloric acid (≈5 percent v/v) anddeionized water just prior to calibrantpreparation. Dilute the contents of the flasksto the mark with deionized water, and shakethem with repeated inversion to ensurethorough mixing. Transfer calibrants toplastic bottles and store them in a refrigeratorwhen they are not in use. Prepare workingcalibrants as needed or biweekly, whichevercomes first.2

1The USGS no longer amends nutrient samples with mercuric

chloride as it did at the time this method was validated. Thepreparation of calibrants, blanks, and check standards in a matrixthat closely matches that of samples being analyzed is necessary toensure the accuracy of analytical results. Beginning in 1996, somenutrient samples received at the NWQL for analysis have beenamended with sulfuric acid. Since January 1999, all whole-waternutrient samples sent to the NWQL require sulfuric acidamendment.

2Undigested calibrants, which are useful for assessinginstrument function, can be prepared from the primary ammoniumcalibrant by using the sampler wash reservoir solution (see section6.2.1) as the matrix.


Table 2. Calibrant preparation protocol

[µL, microliter; mg-N/L, milligrams of nitrogen per liter; mL, milliliter]

Calibrantidentification

Primarycalibrant

volume (µL)

Fieldpreservativevolume (µL)

Nominalconcentration1

(mg-N/L)C1 1,000 1,000 10.0C2 500 1,000 5.0C3 250 1,000 2.50C4 125 1,000 1.25C5 50 1,000 0.50C6 10 1,000 0.10

(Blank) 0 1,000 0.001Based on a final volume of 250 mL.

8. Sample Preparation

8.1 Dispense calibrants, blanks(generally, ASTM type I deionized water),reference materials, and samples into digestiontubes as follows. Be sure the matrix ofcalibrants, blanks, and reference materialsmatch that of samples amended withpreservatives— sulfuric acid, for example— atcollection sites. Note that three positions ineach block are reserved for duplicate andspiked samples. To spike samples at aconcentration of 1.0 mg-N/L, dispense 100 µLof the organic spike solution (see section 7.3)directly into appropriate tubes prior todigestion.

8.2 Rinse all glassware first with adilute solution (≈5 percent v/v) ofhydrochloric acid and then with deionizedwater before each use. This step is critical toachieve consistently low-concentrationdigestion blanks for both nitrogen andphosphorus.

8.3 Use an adjustable (5.0 to10.0 mL) pipet to dispense 10.0-mL aliquotsof calibrant, blank, reference, and samplesolutions into digestion tubes. The suggested

block protocol can be found in table 3.Vigorously shake sample containers andimmediately aspirate aliquots to avoidsampling errors.

NOTE: When samples contain largequantities of suspended solids, continuousstirring during sample aspiration mayprovide the only means of obtainingrepresentative aliquots.

8.4 Use an adjustable (2.0 to 100.0 µL)pipet to dispense 100 µL of the organic spikesolution, equivalent to 1 mg-N/L when addedto 10 mL of sample (see section 7.3), directlyinto samples designated as spikes.

8.5 Use an adjustable volume, syringe-based repetitive dispenser to add 2.00 mL ofdigestion reagent into each tube. Then addseveral acid-rinsed (≈5 percent HCl), Teflon™boiling chips to each tube.

8.6 Desolvate prepared digests under ahood for 30 minutes in a block digester set at220°C. At the end of this time, digestvolumes should be less than 3 mL.

Analytical Method 11

Table 3. Suggested block protocol for determination of Kjeldahl nitrogen by methods I-2515-91/4515-91

[ID, identification; mg-N/L, milligrams nitrogen per liter; SRWS, U.S. Geological Survey Standard ReferenceWater Sample]

Blockposition Sample ID Block

position Sample ID Blockposition Sample ID

1 CALIBRANT 1 15 SAMPLE 29 SAMPLE2 CALIBRANT 2 16 ORG_CHK1 30 SAMPLE3 CALIBRANT 3 17 SAMPLE 31 SAMPLE4 CALIBRANT 4 18 SAMPLE 32 BLANK5 CALIBRANT 5 19 SAMPLE 33 SAMPLE6 CALIBRANT 6 20 SAMPLE 34 SAMPLE7 BLANK 21 SAMPLE 35 SAMPLE8 BLANK 22 SAMPLE 36 SAMPLE9 SAMPLE 23 SAMPLE 37 SAMPLE

10 SAMPLE 24 SRWS 38 SPIKE2

11 SAMPLE 25 SAMPLE 39 DUPLICATE2

12 SAMPLE 26 SAMPLE 40 DUPLICATE2

13 SAMPLE 27 SAMPLE14 SAMPLE 28 SAMPLE

1A solution of glycine (7.5 mg-N/L) is recommended.2Samples for spiking and duplicate determinations should be chosen randomly.

8.7 Immediately transfer (CAUTION)desolvated digests to the block digester set at370°C, and leave them there for 15 minutes.At the end of this time, digest volumes shouldbe less than 0.5 mL and appear as a clear,syrupy bead at the bottom of each digestiontube. Crystallized digests indicate a problemin the acid to sulfate ratio of the digestionreagent.

8.8 Cautiously remove digestion tubesfrom the block digester and allow them to coolfor about 10 minutes in the hood. Withextreme caution, immediately dispense10.0 mL of deionized water into each tubewith vigorous agitation, by using an adjustablevolume, syringe-based repetitive dispenserand a vortex mixer.

9. Instrument Performance

When a pair of block digesters is usedas described in the Introduction, 40calibrants, check standards, and samples canbe digested and made ready for colorimetric

analysis in about 1 hour. The air-segmentedcontinuous flow analyzer used in thismethod can perform 90 ammoniumdeterminations per hour with less than 1percent interaction. Thus, a single channelanalyzer can process the resolvated digestsfrom two pairs of block digesters each hour.

10. Calibration

With a second-order polynomial least-squares curve-fitting algorithm (y =a+bx+cx2, where y is the corrected peakheight and x is the concentration), thecorrelation coefficient (r2) of the calibrationplot should be greater than 0.999. A typicalcalibration plot for Kjeldahl nitrogencalibrants in the concentration range of 0.1to 10.0 mg-N/L is shown in figure 2.

11. Procedure and Data Evaluation

Set up the analytical cartridge of thecontinuous flow analyzer as shown infigure 1. Turn on electrical power to all


0 1 2 3 4 5 6 7 8 9 100.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5 x y10.0 4.406 7.5 3.335 5.0 2.295 1.25 0.588 0.5 0.224 0.1 0.049

y = a + bx + cx2

Parameter Value Errora -0.0011 0.0130b 0.4718 0.0081c -0.0032 0.0008

Corr. Coef.r2 0.99994 0.0183

CO

RR

EC

TED

AB

SO

RB

AN

CE

, IN

VO

LTS

NOMINAL AMMONIUM + ORGANIC NITROGENCONCENTRATION, IN MILLIGRAMS PER LITER

Figure 2. Typical calibration plot for determination of ammonium ions in Kjeldahl digests preparedby methods I-2515-91/4515-91. Corr. Coef = correlation coefficient.

system modules and put fresh sampler washreservoir solution and reagents on-line.After about 10 minutes, verify that theoutput of sample and reference detectors is≈5 volts. A suggested sampler tray protocolfor automated determination of ammoniumions in resolvated digests is provided intable 4.

NOTE: To minimize errors that result fromcontaminated analyzer cups, rinse themseveral times with the solution they are tocontain before placing them on the analyzersampler tray.

NOTE: The full-scale absorbance rangecontrol (STD CAL) of the photometershould not require daily adjustment.Between-run/between-day variations inbaseline-absorbance level and calibrationcurve slope of about ±5 percent areacceptable. Adjustment of the STD CALcontrol to compensate for larger variationsin sensitivity or baseline (reagent blank)levels will only mask underlying problems,such as incipient light-source failure,partially clogged flow cells, or contaminatedor improperly prepared reagents, any ofwhich could compromise analytical results.Dwell time for the analyzer may increase by


Table 4. Suggested tray protocol for automated determination of ammonium ions in resolvated digestsby methods I-2515-91/4515-91

[ID, identification]

Cupnumber

SampleID

Cupnumber

SampleID

Cupnumber

SampleID

1 SYNC 15 SAMPLE 29 U_BLANK1

2 CALIBRANT 1 16 SAMPLE 30 SAMPLE3 CALIBRANT 2 17 SRWS3 31 SAMPLE4 CALIBRANT 3 18 SAMPLE 32 SAMPLE5 CALIBRANT 4 19 SAMPLE 33 SAMPLE6 CALIBRANT 5 20 SAMPLE 34 SAMPLE7 CALIBRANT 6 21 SAMPLE 35 SAMPLE8 U_BLANK1 22 SAMPLE 36 SAMPLE9 BLANK 23 SAMPLE 37 SAMPLE

10 BLANK 24 SRWS3 38 SAMPLE11 ORG_CHK2 25 SAMPLE 39 SAMPLE12 SAMPLE 26 SAMPLE 40 SPIKE13 SAMPLE 27 SAMPLE 41 DUPLICATE14 SAMPLE 28 SAMPLE 42 DUPLICATE

1Undigested blank (sampler wash reservoir solution, see 6.2.1).2Organic nitrogen check sample; see note 2 on table 3.3USGS standard reference water sample.

a few seconds per day as pump tubes wear,but it need not be checked or updated morethan once per week.

12. Calculations

12.1 Instrument calibration requires thepreparation of a set of solutions (calibrants) inwhich the analyte concentration is known.These calibrants are digested along withsamples and used to establish a calibrationfunction that is estimated from a least-squaresfit of nominal calibrant concentrations (x) inrelation to peak absorbance (y). A second-order polynomial function (y = a+bx+cx2)usually provides improved concentrationestimates at the upper end of the calibrationrange than the more conventional linear model(y = a+bx). Accuracy is not lost when asecond-order fit is used, even if the calibrationfunction is strictly linear, because, in this case,the value estimated for the quadraticparameter c will approach zero.

12.2 Before the calibration function canbe estimated, the baseline absorbancecomponent of measured peak heights,including drift (continuous increase ordecrease in the baseline absorbance during thecourse of an analysis), if present, needs to beremoved. Baseline absorbance in continuousflow analysis is analogous to the reagent blankabsorbance in batch analysis. Correction forbaseline absorbance is an automatic functionof most data acquisition and processingsoftware sold by vendors of continuous flowanalyzers.

NOTE: These correction algorithms arebased on linear interpolation between initialand intermediate or final baselinemeasurements, and so they do not accuratelycorrect for abrupt, step-changes in baselineabsorbance that usually indicate partialblockage of the flow cell. It is prudent,therefore, to reestablish baseline absorbanceat intervals of 20 samples or so.


12.3 After peaks are baseline corrected,they need to be digestion blank corrected.

This correction can be applied inseveral ways:

1. Subtract the baseline-correctedabsorbance of the digestion blank—compute an average concentration ifmultiple digested blanks areincluded in each block— from thebaseline-corrected absorbance of allcalibrants, check standard, andsamples in the block. Then estimateregression parameters (a, b, and cterms) for the calibration functionby using a second-order polynomialleast-squares algorithm. Forsecond- and higher order calibrationfunctions, use the Newton-Raphsonsuccessive approximationsalgorithm (Draper and Smith, 1966;Swartz, 1976, 1977, 1979) toconvert corrected peak heights intoconcentrations.

2. Designate digestion blanks as acalibrant with a nominalconcentration of zero. Thecalibration function estimated asdescribed in section 12.3 then willhave a positive y-intercept, with amagnitude that will approximate thebaseline-corrected absorbance of thedigestion blank. If this method isused, be sure that the curve-fittingalgorithm does not force a zero y-intercept by including one or more"dummy" (0,0) points in the data setused for calibration.

3. Designate digested blanks as baselinecorrection samples. In this caseinitial, intermediate (if included),and final baselines are interpolatedbetween digested blank peakmaxima. Thus, baseline and

digestion blank corrections areperformed in a single operation.

NOTE: Digestion blank corrections fordata in this report were performed bymethod 1. Note, however, that analyticalresults calculated by the other two methodsshould be equivalent, although in theauthors' opinion the second correctionmethod is most sound statistically. In anycase, choose the digestion blank correctionalgorithm that is most easily implementedwith the software package available.Regardless of the algorithm chosen, makesure that it is documented in the standardoperating procedure (SOP) and that it isunderstood by analysts. This process alsoapplies to algorithms used to convertcorrected peak heights to concentrationunits. The SOP for this method must beupdated whenever any changes in dataacquisition and processing software or incalculation algorithms are implemented.

12.4 Most software packages provide adata base for entering appropriate dilutionfactors. Usually, these factors can be enteredbefore or after analyses are performed. Ifdilution factors are entered, then reportedconcentrations will be compensatedautomatically for the extent of dilution. Thedilution factor is the number by which ameasured concentration must be multiplied toobtain the analyte concentration in the samplebefore dilution. For example, dilution factorsof 2, 5, and 10 indicate that sample and diluentwere combined in proportions of 1+1, 1+4,and 1+9, respectively.

13. Reporting Results

Report Kjeldahl nitrogen, dissolved (testID = 00623), and total (test ID = 00625)concentrations as follows: 0.1 to 0.9 mg-N/L,one decimal; 1.0 mg-N/L and greater, twosignificant figures.


NOTE: The laboratory reporting level(LRL)— about twice the method detectionlevel (MDL) as defined in section 14.1— forthis method was reduced from 0.2 mg-N/Lto 0.1 mg-N/L on November 10, 1997 (U.S.Geological Survey National Water QualityLaboratory Technical Memorandum 98.07,1998).

14. Detection Levels, Precision, andAccuracy

14.1 A method detection level (MDL)of ≈0.05 mg-N/L was estimated for methodsI-2515-91/4515-91 by using the protocol setforth in the U.S Environmental ProtectionAgency (1990). This MDL estimate wasobtained with homogenous solutions;therefore, the MDL for particulate-ladensamples might be somewhat higher because ofthe increased difficulty of obtaining arepresentative sample for analysis (seetable 5).

14.2 Within-run precision (repeatability)for methods I-2515-91/4515-91 on the basis of55 replicate determinations of the same highsuspended-solids sample was 3.1 percent. Theaverage concentration and standard deviationof these replicates was 4.1 ± 0.1 mg-N/L (seetable 6).

Between-day precision based on 26replicate determinations of the same samplebetween April 8 and April 26, 1991, was 4.8percent. The average concentration andstandard deviation of between-day replicateswas 4.1 ± 0.2 mg-N/L (see table 7). Figure 3shows duplicate data for 564 randomlyselected samples determined for Kjeldahlnitrogen at the NWQL by method I-2515-91/4515-91 between the dates of 2/11/92 and2/13/93. The top panel of this figure presentstrial 1 concentra-tion in relation to trial 2concentration as a scatter plot about the line ofequal relation. The bottom panel shows a

scatter plot of the percent difference betweenduplicates in relation to the trial 1concentration. About 90 percent of the 300samples with Kjeldahl nitrogen concentrations≥ 0.5 mg-N/L (10 MDLs) agreed to within ±25 percent.

14.3 Between-day accuracy for methodsI-2515-91/4515-91, which is based onconcentrations determined for referencematerials provided by the USEPA and theUSGS, is listed in table 8. In all cases,Kjeldahl nitrogen concentrations determinedby this method were within published controllimits for these reference materials.

14.4 The average recovery for 1.00 mgof glycine nitrogen added to 94 randomlyselected samples analyzed during the July–August experiment was 1.04 ± 0.10 mg-N/L.Figure 4 shows that concentrations of Kjeldahlnitrogen in samples did not affect spikerecovery.

14.5 When the 1,487 data pairs fromApril and July–August experiments werecombined, the median difference betweenKjeldahl nitrogen concentrations determinedby the present (I-2515-91/4515-91) and old(I-2552-85/4552-85) methods was about 0.1mg-N/L, which is highly significant(p < 0.0001) on the basis of a Wilcoxonsigned rank test. Linear regression analysis ofthese data pairs shows a statisticallysignificant (p < 0.0001), nonzero intercept,which also suggests a possible additivedifference between the former and newmethods. Table 9 lists mean and medianKjeldahl nitrogen concentration differencesdetermined for filtered- and whole-watersamples by the former and the new methods aswell as the results of the regression of theformer method in relation to the new methods.A concentration difference of this magnitudewas expected because digestion blankconcentrations (≈ 0.1 mg-N/L) were


Table 5. Data used to estimate the method detection level (MDL) for Kjeldahl nitrogendetermination by methods I-2515-91/4515-91

[mg-N/L, milligrams nitrogen per liter. Measured concentrations pertain to eight replicate digestions ofa 0.25 mg-N/L glycine calibrant that was prepared in deionized water that contained an appropriate volumeof field preservative solution.]

Replicatenumber

Nominal concentration(mg-N/L)

Measured concentration(mg-N/L)

1 0.25 0.1942 .25 .1963 .25 .2024 .25 .2295 .25 .1926 .25 .2047 .25 .1668 .25 .196

Average concentration (mg-N/L):Standard deviation (mg-N/L):

Number of samples:Degrees of freedom:

One-sided t value (99 percent confidence):Method detection level (MDL):

.197

.017872.9980.052

Table 6. Repeatability (within-run precision, May 3, 1991) data for 55 replicate determinationsof Kjeldahl nitrogen concentration in the Heidelberg sample by using methods I-2515-91/4515-91

[mg-N/L, milligrams nitrogen per liter]

Trialnumber

Concentration(mg-N/L)

Trialnumber


Trialnumber


1 3.97 20 4.12 38 4.132 4.36 21 4.29 39 4.063 4.16 22 4.29 40 4.254 4.19 23 4.09 41 3.945 4.36 24 4.20 42 4.066 4.07 25 4.25 43 4.227 4.21 26 3.99 44 3.938 4.37 27 4.13 45 4.119 4.04 28 4.20 46 4.23

10 4.21 29 3.88 47 3.9911 4.26 30 4.03 48 4.1812 4.09 31 4.15 49 4.2113 4.33 32 3.94 50 3.9714 4.22 33 3.98 51 4.0415 4.15 34 4.11 52 4.1916 4.34 35 3.89 53 3.9017 4.11 36 4.06 54 4.1118 4.14 37 4.22 55 4.1419 4.24

Average concentration (mg/L):Standard deviation (mg/L):

Relative standard deviation (percent):

4.130.133.1

Discussion of Results 17

Table 7. Between-day (April 8–26, 1991) precision of digested calibrants and the Heidelberg sample forKjeldahl nitrogen determination by using methods I-2515-91/4515-91

[ID, identification; mg-N/L, milligrams nitrogen per liter; std. dev., standard deviation; RSD, relative standarddeviation; ≈, nearly equal to; µm, micrometer]

SampleID

Nominalconcentration

(mg-N/L)

Numberof

samples

Averageconcentration

(mg-N/L)

Std. dev.(mg-N/L)

RSD(percent)

10.0 24 9.99 0.04 0.387.5 25 7.52 .08 1.105.0 23 4.98 .08 1.652.5 25 2.50 .06 2.301.25 25 1.27 .06 4.39.5 25 .48 .06 3.51

3 ≈ 4.1 26 4.14 .20 4.88

CALIBRANT 1CALIBRANT 2CALIBRANT 3CALIBRANT 4CALIBRANT 5CALIBRANT 6Heidelberg1

Heidelberg2 3 ≈ 4.1 26 4.14 .20 4.831Diluted digest was shaken and poured into analyzer cup.2Diluted digest was dispensed into analyzer cup through a 0.45-µm nylon syringe filter.3National Water Quality Laboratory consensus concentration.

not subtracted from concentrations reportedby the former methods. Indeed, a stepchange of about 0.1 mg-N/L in NWQLKjeldahl nitrogen blind blank concentrationsis evident after October 1, 1991, when thenew methods became operational, as shownin figure 5. Complete details of the methodcomparison studies can be found in theDiscussion of Results section that follows.

DISCUSSION OF RESULTS

Analytical Methods

Before beginning validation studies, theauthors performed several experiments tocharacterize block digester and continuousflow analyzer performance. For example, toestimate the extent of temperature variationthroughout the 40 positions within the blockdigester, a batch of 40 test solutions asdescribed in section 8 was prepared anddigested. The first eight tubes in row 1 of theblock digester contained calibrants and blanks(see table 3 and fig. 6b). The remaining 32tubes in rows 2 through 5 contained anaqueous solution of nicotinic acid with anominal concentration of 7.5 mg-N/L.

Nicotinic acid was chosen for this studybecause it is one of the most difficultcompounds that contain organic nitrogen todigest by the Kjeldahl method (Bradstreet,1965; Jirka and others, 1976; Bowman andDelfino, 1982). Digestion time was limited to90 minutes at 370°C— less than the timerequired for 100-percent recovery of nicotinicacid. The authors hypothesized that recoveryof nicotinic acid in each tube would indicatethe relative temperature at each block position.That is, tubes with lower nicotinic acidrecoveries would be correlated with coolerpositions in the block. Results of thisexperiment are presented in figure 6a, whichshows isopleths for nicotinic acidconcentration found in each tube relative to itsposition in the block digester. This figureshows that recovery of nicotinic acidapproached 100 percent only in tubes near thecenter of the block digester and decreasedradially from the digester's center to itsperimeter.

Next, the authors performed a set ofexperiments to assess how the temperaturegradient detected within the block digesterwould affect method performance for


0.1 1 10-100

-75

-50

-25

0

25

50

75

100

(b)

10 method detection levels (MDLs)

DIF

FER

EN

CE

BE

TWE

EN

DU

PLI

CA

TES

, IN

PE

RC

EN

T

LOG 10 KJELDAHL NITROGEN CONCENTRATION,

IN MILLIGRAMS NITROGEN PER LITER, TRIAL 1

0.1

1

10

(a)

Line of equal relation

LOG

10 K

JELD

AH

L N

ITR

OG

EN

CO

NC

EN

TRA

TIO

N, I

N M

ILLI

GR

AM

SN

ITR

OG

EN

PE

R L

ITE

R, T

RIA

L 2

Figure 3. Duplicate data for 564 randomly selected samples determined at the National Water QualityLaboratory by methods I-2515-91/4515-91 between 2/11/92 and 2/13/93. About 90 percent of the 300samples with Kjeldahl nitrogen concentrations ≥ 0.5 milligram nitrogen per liter (10 MDLs) agreed towithin ± 25 percent. Note that several extreme outliers shown in graph (a) do not appear in graph (b),because the y-axis scale is limited to differences between trial 1 and trial 2 of plus or minus 100 percent.


Table 8. Between-day (April 8–26, 1991) accuracy of digested U.S. Environmental Protection Agencyand U.S. Geological Survey reference samples for Kjeldahl nitrogen determination by using methodsI-2515-91/4515-91

[ID, identification; mg-N/L, milligrams nitrogen per liter; ±, plus or minus; std. dev., standard deviation; RSD,relative standard deviation; LCL, lower control limit; UCL, upper control limit; µL, microliters; mL, milliliters;SRWS, standard reference water sample]

SampleID

Nominalconcentration

(mg-N/L)

Numberof

samples

Averageconcentration

(mg-N/L)

Std. dev.(mg-N/L)

RSD(percent)

USEPA LOW1

(LCL - UCL)22.50 ± 0.34

(1.82 – 3.18)13 2.40 0.19 8

USEPA3

(LCL - UCL)24.95 ± 0.44

(4.07 – 5.83)10 5.00 .20 4

SRWS N-28(LCL - UCL)

0.26 ± 0.17(-0.08 – 0.60)

11 .154 (0.03 – 0.35)

.10 67

SRWS N-29(LCL - UCL)

1.21 ± 0.21(0.79 – 1.63)

13 1.024 (0.88 – 1.17)

.09 8

1A solution containing 500 µL USEPA "Nutrient 2" and 400 µL National Water Quality Laboratory HgCl2/NaClpreservative solution diluted to 100 mL with deionized water.

2Confidence interval of 95 percent.3A 1,000 µL USEPA "Nutrient 2" solution and 400 µL National Water Quality Laboratory HgCl2/NaCl

preservative solution diluted to 100 mL with deionized water.4Concentration range.

0.1 1

8 0

1 0 0

1 2 0

SP

IKE

RE

CO

VE

RY

, IN

PE

RC

EN

T

K J E L D A H L N I T R O G E N C O N C E N T R A T I O N ,I N M I L L I G R A M S P E R L I T E R

Figure 4. Spike recovery in relation to Kjeldahl nitrogen concentration determined by methodsI-2515-91/4515-91 during the July–August, 1991 experiment. Number of samples = 94, nitrogen added(as glycine) = 1.0 milligram nitrogen per liter (mg-N/L), average nitrogen found = 1.04 ± 0.10 mg-N/L, andaverage percent recovery = 104 ± 10.


Table 9. Statistical data and regression analysis results for Kjeldahl nitrogen methodsI-2515-91/4515-91 (new methods) and I-2552-85/4552-85 (former methods)

[Mean and median concentration differences, in milligrams nitrogen per liter, are the result of subtracting former-method concentrations from new-method concentrations. Slopes and y-intercepts are the result of using new-methodconcentrations as the independent (x) variable. Data for surface-water and ground-water samples are combined.]

Sampletype

Number ofsamples

Mediandifference

Meandifference Slope y-intercept

Filtered 508 -0.074 -0.091 0.922 0.159Whole-water 979 -0.115 -0.049 0.962 0.096

several real and synthetic samples thatcontain various forms of organic nitrogen.Percent recovery and precision of analyticalresults were used to evaluate methodperformance. In this second set ofexperiments, full blocks for each of threedifferent test solutions were prepared anddigested. As before (see section 8), the firsteight tubes in row 1 of the block digestercontained calibrants and blanks. Theremaining 32 tubes in rows 2 through 5contained an aqueous solution of nicotinicacid with a nominal concentration of 7.5mg-N/L in batch one, an aqueous solution ofadenosine triphosphate (ATP) with anominal concentration of 1.0 mg-N/L inbatch 2, and a high suspended-solids samplewith a consensus concentration of 6.3 ± 0.5mg-N/L in batch 3. Each batch wasdesolvated in a block digester set at 220°Cfor 30 minutes and then transferred to thehigh-temperature block digester set at370°C. After 15 minutes, all eight tubeswere removed from row 1, which containedcalibrants and blanks, and tubes 1 through 4from row 2. At 15-minute intervalsthereafter, tubes 5 through 8 were removedfrom row 2, tubes 1 through 4 from row 3,tubes 5 through 8 from row 3, tubes 1through 4 from row 4, tubes 5 through 8from row 4, tubes 1 through 4 from row 5,and tubes 5 through 8 from row 5 (see fig.6b). This procedure yielded 4-tube subsets

of each test solution that had undergonedigestion at 370°C for 15, 30, 45, 60, 75, 90,105, and 120 minutes.

Results of these experiments aresummarized in figure 7, where symbolsplotted as a function of digestion timerepresent the average Kjeldahl nitrogenconcentration measured in each 4-tube subsetfor each of the three test solutions. Error barsassociated with the symbols represent ±1standard deviation. As shown in figure 7,recovery of nitrogen from digested ATPsolutions and high suspended-solids sampleswere ≥ 95 percent after 15 minutes andremained constant at longer digestion times.Standard deviations for these two testsolutions fluctuated randomly as digestiontime increased. Typically, however, shorterdigestion times are desirable because digestionblank concentrations and variability tend toincrease as digestion time increases. Incontrast, but in keeping with the block digestertemperature gradient experiment, digestiontimes greater than 90 minutes were required toapproach nitrogen recoveries ≥ 95 percent innicotinic acid solution digests. The authorsinterpret the increased precision of nicotinicacid concentrations observed at the longestdigestion time as an indication that the effectof the block digester's temperature gradient onnicotinic acid recovery decreased substantiallyas digestion time approached 120 minutes.


1 2 3 4 5 6 7 81

2

3

4

5 (b)

RO

W N

UM

BE

R, F

RO

M F

RO

NT

TO B

AC

K

TUBE POSITION IN BLOCK DIGESTER,FROM THE FRONT, LEFT TO RIGHT

4.3

4.7

4.75.1

5.15.5

5.55.9

5.9

6.3 6.3

6.7

7.1

2

3

4

5(a)

Figure 6. Isopleths of nicotinic acid concentration (a), in milligrams nitrogen perliter, found in each tube in relation to the position of each tube in the block digester (b).


15 30 45 60 75 90 105 1200

1

2

3

4

5

6

7

8C

ON

CE

NTR

ATI

ON

, IN

MIL

LIG

RA

MS

NIT

RO

GE

N P

ER

LIT

ER

Adenosine triphosphate (ATP)

Nicotinic acid

High suspended-solids sample

DIGESTION TIME AT 370 DEGREES CELSIUS, IN MINUTES

Figure 7. Kjeldahl nitrogen concentration recovered in digests of a high-suspended-solids sample, anaqueous solution of adenosine 5' triphosphate, and an aqueous solution of nicotinic acid by methodsI-2515-91/4515-91 as a function of high-temperature digestion time. Error bars represent +1 standarddeviation for the average of four determinations of each sample type at specified digestion times.

On the basis of these results and statisticalanalysis of Kjeldahl nitrogen concentrationsdetermined by USGS methods I-2552-85/4552-85 (370°C digestion time ≈90minutes) and I-2515-91/4515-91 (370°Cdigestion time =15 minutes), the authorsconclude as follows:

1. Digestion times of 15 minutes at370°C are adequate for fullrecovery of organic nitrogencompounds in natural-water

samples typically received foranalyses at the NWQL, and

2. Longer digestion times required forquantitative recovery of nicotinicacid might contribute to over- orunderestimation of Kjeldahlnitrogen concentrations in samplescontaining nitrate concentration ≥ 1mg-N/L (see discussion thatfollows).


Statistical Analysis

The lack of precision in the formerUSGS methods (I-2552-85/4552-85) forseveral years preceding validation of the newmethods (I-2515-91/4515-91) is indicated byNWQL blind blank data for the years 1989through 1999 (see fig. 5). In an effort tomitigate errors arising from statisticalcomparison of methods with differingprecision, the authors used medians andmedian-based, nonparametric statistics(Wilcoxon signed rank test) in the discussionsthat follow in preference to means and means-based, parametric paired t-tests. As listed intable 9, when filtered- and whole-water-sample data from the April and July–Augustexperiments are combined, median Kjeldahlnitrogen concentrations determined by thenew methods are about 0.1 mg-N/L less thanthose determined by the former methods. Thiswas expected because Kjeldahl nitrogenconcentrations reported for the formermethods were not corrected for digestionblank concentrations, which are approximatelyequal to this difference. Median concentrationdifferences between the two methods werecomparable during the April and July–Augustexperiments, as listed in table 10.

When data pairs from the April andJuly–August experiment were grouped intothe categories filtered and whole surface-watersamples and filtered and whole ground-watersamples, similar trends were observed. Datafor each of these four categories are shown asscatter plots in figure 8 and as box plots infigure 9. As shown in these figures, thedeviations of plotted data from the line ofequal relation, which represent theconcentration difference between the new andformer methods, was least for unfilteredsurface-water samples from both studies.Additional trends in the data emerge when

method median differences are calculated forthe subset of samples in these four categorieswith nitrate plus nitrite concentrations eitherless than or equal to 1.0 mg-N/L or greaterthan 1.0 mg-N/L. The results of such ananalysis are presented for data from the Aprilexperiment (table 11) and July–Augustexperiments (table 12).

Inspection of tables 11 and 12 revealsthat median concentration differences betweenthe two methods in all four categories moreclosely matches the median digestion blankconcentration in the subset of samples withnitrate plus nitrite concentrations less than orequal to 1 mg-N/L than for the entirepopulation. Likewise, larger concentrationdifferences than can be accounted for by thelack of blank correction in the former methodsare observed for the subset of samples withnitrate plus nitrite concentrations greater than1 mg-N/L. For a substantial number ofsamples in these subsets, the differencesbetween Kjeldahl nitrogen concentrationsdetermined by the former and new methodsincreases as nitrate plus nitrite concentrationsincrease. The authors hypothesize that thiseffect is caused by complex and poorlycharacterized interactions among nitrate,ammonium, and organic matter duringKjeldahl digestion as discussed in section 3.3.The authors further speculate that the muchlonger 370°C digestion time of the formermethods in relation to the new methods— 90minutes rather than 15 minutes— intensifiesthese undesirable side reactions.

The concentration difference betweenthe new and former methods was positive in asingle case in the July–August experiment.This difference was observed in the subset ofwhole surface-water samples with nitrate plusnitrite concentrations greater than 1 mg-N/L(see table 12).


Table 10. Median difference between paired Kjeldahl nitrogen concentrations determined bymethods I-2515-91/4515-91 and I-2552-85/4552-85

[Median differences are the result of subtracting methods I-2552-85/4552-85 concentrations from methodsI-2515-91/4515-91 concentrations. Data for surface-water and ground-water samples are combined.mg–N/L, milligrams nitrogen per liter]

Median Kjeldahl nitrogen concentration (mg-N/L)Data source Number

of samples I-2515-91I-4515-91

I-2552-85I-4552-85

Mediandifference

April 411 0.495 0.566 -0.071July–August 1,076 .422 .519 -.097Both 1,487 .437 .531 -.094

Careful inspection of original dataeliminated inadvertent transposal of datacolumns for the two methods during initialstages of data analysis as a possibleexplanation for this apparent anomaly. In afurther effort to account for the anomaly, theammonium concentrations were examined inthe subset of samples with nitrate plus nitriteconcentrations greater than 1 mg/L (see tables11 and 12). As indicated in table 13, theaverage nitrate plus nitrite and ammoniumconcentrations in the July–August subset havehigher than average ammonium concentrationsand, in general, lower than average nitrate plusnitrite concentrations than the April subset.Furthermore, the authors found that 14samples in the July–August subset (126samples)— seven each from two USGSstations near minimally treated sewagedischarge outfalls— were largely responsiblefor the observed positive concentrationdifference between the two methods (table12). Average ammonium and nitrate plusnitrite concentrations were 11.1 and 2.2 mg-N/L for the seven samples from station9419700 and 9.4 and 2.6 mg-N/L for the sevensamples from station 9419753. Samples withKjeldahl nitrogen and ammoniumconcentrations in this range would have

required dilution before analysis. The authors,therefore, can only speculate whether dilutionerrors, the atypically high ammonium tonitrate ratio (table 13), other unknown factors,or some combination of all three areresponsible for the positive concentrationdifference between the new and formermethods in this data set.

This study provides evidence that nitrateplus nitrite concentrations greater than 1 mg-N/L might cause positive or negativeinterference in Kjeldahl nitrogen determina-tions. This information should serve as acaution to data users who calculate totalnitrogen concentrations by summing Kjeldahlnitrogen and nitrate plus nitrite concentrations.

Data analysis in this section alsoprovides insight into the difficulties ofcomparing Kjeldahl nitrogen (ammonium plusorganic nitrogen) concentrations with totalnitrogen concentrations determined by high-temperature combustion methods (U.S.Environmental Protection Agency, 1997) oralkaline persulfate digestion (D’Elia andothers, 1977) methods as discussed inprevious studies (Ameel and others, 1993;Kroon, 1993).


0.050.05

0.25

0.50.5

2.5

55

25

5050

Line of equal relationAPRIL DATA

SET F9

n = 43


SET R9

n = 317

0.050.05

0.250.50.5

2.5

55

25

5050


SET F6

n = 32


SET R6

n = 19

0.050.05

0.25

0.50.5

2.5

55

25

5050

Line of equal relationJULY DATA

SET F9

n = 416


SET R9

n = 601

0.050.05 0.25 0.50.5 2.5 55 25 50500.050.05

0.25

0.50.5

2.5

55

255050


SET F6

n = 17

0.050.05 0.25 0.50.5 2.5 55 25 5050

LOG

10 K

JELD

AH

L N

ITR

OG

EN

CO

NC

EN

TRA

TIO

N, I

N M

ILLI

GR

AM

S P

ER

LIT

ER

(ME

THO

DS

I-25

15-9

1/45

15-9

1)

LOG 10 KJELDAHL NITROGEN CONCENTRATION,

IN MILLIGRAMS PER LITER (METHODS I-2252-85/4552-85)


SET R6

n = 42

Figure 8. Relation between Kjeldahl nitrogen concentrations determined by methods I-2552-85/4552-85 and I-2515-91/4515-91 in the April and July–August experiments for each of the four water types.Sets F9, R9, F6, and R6 were composed of filtered and whole surface-water samples and filtered andwhole ground-water samples, respectively. The letter n represents the number of points in each set.


Table 11. Effect of nitrate plus nitrite concentrations on median differences between Kjeldahl nitrogenconcentrations determined by U.S. Geological Survey methods I-2515-91/4515-91 andI-2552-85/4552-85 (April data set)

[mg-N/L, milligrams nitrogen per liter; ≤, less than or equal to; >, greater than]

All samplesMedian Kjeldahl nitrogen concentration (mg-N/L)Water type

Number ofsamples

I-2515-91I-4515-91

I-2552-85I-4552-85 Difference

Filtered surface 43 0.475 0.636 -0.161Whole surface 317 .561 .592 -.031Filtered ground 32 .143 .325 -.182Whole ground 19 .144 .281 -.137

Samples with NO3- + NO2

- concentrations (≤ 1 mg-N/L)Filtered surface 25 .375 .442 -.067Whole surface 249 .459 .496 -.037Filtered ground 20 .231 .33 -.099Whole ground 10 .149 .207 -.058


- concentrations (> 1 mg-N/L)Filtered surface 18 .843 1.281 -.438Whole surface 68 1.217 1.48 -.263Filtered ground 12 .075 .319 -.244Whole ground 9 .144 .451 -.307

Table 12. Effect of nitrate plus nitrite concentrations on median differences between Kjeldahl nitrogenconcentrations determined by U.S. Geological Survey methods I-2515-91/4515-91 andI-2552-85/4552-85 (July–August data set)

[mg-N/L, milligrams nitrogen per liter; ≤, less than or equal to; >, greater than]

All samplesMedian Kjeldahl nitrogen concentration (mg-N/L)Water type Number of

samples I-2515-91I-4515-91

I-2552-85I-4552-85 Difference

Filtered surface 416 0.352 0.452 -0.100Whole surface 601 .566 .616 -.050Filtered ground 17 .138 .579 -.441Whole ground 42 .109 .382 -.273


- concentrations (≤ 1 mg-N/L)Filtered surface 341 .311 .413 -.102Whole surface 475 .455 .53 -.075Filtered ground 7 .272 .446 -.174Whole ground 15 .064 .172 -.108


- concentrations (> 1 mg-N/L)Filtered surface 75 .582 .713 -.131Whole surface 126 1.123 1.017 +.106Filtered ground 10 .073 .573 -.5Whole ground 27 .122 .493 -.371

Conclusions 29

Table 13. Average concentrations of nitrate plus nitrite and ammonium for sample subsetswith nitrate plus nitrite concentrations greater than 1 milligram nitrogen per liter

[mg-N/L, milligrams nitrogen per liter]

Average concentration (mg-N/L)

April July – AugustWater type

NO2- + NO3

- NH4+ NO2

- + NO3- NH4

+

Filtered surface (F9) 11.1 0.12 2.5 2.1

Whole surface (R9) 6.0 .62 2.4 1.6

Filtered ground (F6) 5.0 .00 19.6 .04Whole ground (R6) 8.3 .00 6.1 .06

CONCLUSIONS

U.S. Geological Survey (USGS)methods I-2515-91/4515-91 for determiningKjeldahl nitrogen in filtered and whole-watersamples were implemented at the NationalWater Quality Laboratory (NWQL) onOctober 1, 1991. The methods, in which high-temperature block digestion is followed byautomated, colorimetric ammoniumdetermination that includes digest cleanup bycontinuous flow gas diffusion, have methoddetection levels of about 0.05 mg-N/L asestimated by using U.S EnvironmentalProtection Agency protocols. Within-runprecision (repeatability) and between-dayprecision for filtered and whole-water samplesanalyzed by methods I-2515-91/4515-91 areabout 3 percent and 5 percent, respectively.Duplicate determinations for these methodstypically agree to within 25 percent, and spikerecoveries are typically 100 ± 20 percent.Data provided in this report demonstrate thatdigestion times of 15 minutes at 370°C areadequate for full recovery of organic nitrogencompounds in natural-water samples typicallyreceived for analyses at the NWQL.Furthermore, these data suggest that digestiontimes longer than 15 minutes might contributeto over- or underestimation of Kjeldahlnitrogen concentrations in samples with nitrateplus nitrite concentrations ≥ 1 mg-N/L.

Statistical analysis of paired datacollected during method validationexperiments and NWQL blind blank data areconsistent with a step-change decrease ofabout 0.1 mg-N/L in Kjeldahl nitrogenconcentrations reported by the NWQLbeginning on October 1, 1991. On that date,USGS methods I-5225-85/4525-85 werereplaced by USGS methods I-2515-91/4515-91. Generally, this concentration differenceresulted from the lack of digestion blankcorrection in methods I-5225-85/4525-85.

This work documents the importance ofconsidering factors, such as digestion time andwhether digestion blank corrections have beenapplied, when Kjeldahl nitrogen dataproduced at different laboratories arecompared. It also provides insight into thedifficulties of comparing Kjeldahl nitrogenconcentration with true total nitrogenconcentrations. This work also indicates thatnitrate plus nitrite concentrations greater than1 mg-N/L might cause positive or negativeinterference in Kjeldahl nitrogendeterminations. This information should serveas a caution to data users who either combineKjeldahl nitrogen concentrations fromdifferent laboratories, or who calculate totalnitrogen concentrations by summing Kjeldahlnitrogen and nitrate plus nitrite concentrations,or both.


REFERENCES CITED

Ameel, J.J., Axler, R.P., and Owen, C.J.,1993, Persulfate digestion fordetermination of total nitrogen andphosphorus in low-nutrient waters:American Environmental Laboratory,no. 10/93, p. 1, 8, 10, and 11.

American Public Health Association, 1992,Standard methods for the examinationof water and wastewater (18th ed.):Washington, D.C., American PublicHealth Association Inc., p. 4–94.

American Society for Testing and Materials,1995, Annual book of ASTMstandards, Section 11, Water (D1193,Standard specification for reagentwater): Philadelphia, v. 11.01,p. 122–124.

Bowman, G.T., and Delfino, J.J., 1982,Determination of total Kjeldahlnitrogen and total phosphorus insurface waters and wastewaters:Journal of the Water Pollution ControlFederation, v. 54, no. 9, p. 1324–1330.

Bradstreet, R.B., 1965, The Kjeldahl methodfor organic nitrogen: New York,Academic Press, p. 239.

D’Elia, C.F., Steudler, P.F., and Corwin, N.,1977, Determination of total nitrogenin aqueous samples using persulfatedigestion: Limnology andOceanography, v. 22, p. 760–764.

Draper, N.R., and Smith, Harry, 1966,Applied regression analysis: NewYork, Wiley, 407 p.

Fishman, M.J., ed., 1993, Methods ofanalysis by the U.S. Geological SurveyNational Water Quality Laboratory—Determination of inorganic and organicconstituents in water and fluvialsediments: U.S. Geological SurveyOpen-File Report 93–125, 217 p.

Fishman, M.J., and Friedman, L.C., eds.,1989, Methods for determination ofinorganic substances in water andfluvial sediments: U.S. Geological

Survey Techniques of Water-Resources Investigations, book 5,chap. A1, 545 p.

Harfmann, R.G., and Crouch, S.R., 1989,Kinetic study of Berthelot reactionsteps in the absence and presence ofcoupling reagents: Talanta, v. 36,p. 261–269.

Jirka, A.M., Carter, M.J., May, Dorothy, andFuller, F.D., 1976, Ultramicrosemiautomated method forsimultaneous determination of totalphosphorus and total Kjeldahl nitrogenin wastewaters: EnvironmentalScience and Technology, v. 10, no. 10,p. 1038–1044.

Kroon, Henie, 1993, Determination ofnitrogen in water: Comparison of acontinuous flow method with on-lineUV digestion with the originalKjeldahl method: Analytica ChimicaActa, v. 276, p. 278–293.

Patton, C.J., and Crouch, S.R., 1977,Spectrophotometric and kineticsinvestigation of the Berthelot reactionfor the determination of ammonia:Analytical chemistry, v. 49,p. 464–469.

Patton, C.J., and Truitt, E.P., 1992, Methodsof analysis by the U.S. GeologicalSurvey National Water QualityLaboratory— Determination of totalphosphorus by a Kjeldahl digestionmethod and an automated colorimetricfinish that includes dialysis: U.S.Geological Survey Open-File Report92-146, 39 p.

Patton, C.J., and Wade, A.P., 1997,Continuous flow analyzers, in Ewing,Galen, ed., Analytical instrumentationhandbook (2d ed.): New York, MarcelDekker, p. 125–220.

References Cited 31

Reardon, J., Foreman, J.A., and Searcy,R.L., 1966, New reactants fordetermination of ammonia: ClinicaChimica Acta, v. 14, p. 403–405.

Seifter, Eli, Kambosos, Demetrios, andChanas, Alexandra, 1971,Determination of dissloved CO2 inbiological fluids, in Advances inautomated analysis, TechniconInternational Congress, 1970, VolumeI (Clinical Analysis): Miami, Florida,Thurman Associates, p. 509–511.

Swartz, L.M., 1976, Nonlinear calibrationcurves: Analytical Chemistry, v. 48,no. 14, p. 2287–2290.

________1977, Nonlinear calibration:Analytical Chemistry, v. 49, no. 13,p. 2062–2068.

________1979, Calibration curves withnonuniform variance: AnalyticalChemistry, v. 51, no. 6, p. 723–727.

U.S. Environmental Protection Agency,1983a, Methods for chemical analysisof water and wastes: Cincinnati, Ohio,Environmental Monitoring andSupport Laboratory, USEPA-600/4-79-020, p. 365.1-1 through 365.4-1.

________1983b, Methods for chemicalanalysis of water and wastes:Cincinnati, Ohio, EnvironmentalMonitoring and Support Laboratory,USEPA-600/4-79-020, p. 351.2-1through 351.2-5.

________1990, Guidelines establishing testprocedures for the analysis ofpollutants (App. B, Part 136,Definition of procedures for thedetermination of the method detectionlimit— Revision 1.11): revised as ofJuly 1, 1999, p. 537–539.

________1993, Methods for thedetermination of inorganic substancesin environmental samples: Cincinnati,Ohio, Environmental Monitoring andSupport Laboratory, EPA/600/R-93/100, August 1993, 79 p.

________1997, Determination of carbonand nitrogen in sediments andparticulates of estuarine/coastal watersusing elemental analysis: Cincinnati,Ohio, National Exposure ResearchLaboratory, Office of Research andDevelopment, Method 440.0, 10 p.,accessed May 18, 2000, at URLhttp://www.epa.gov/nerlcwww/m440_0.pdf

U.S. Geological Survey, 1994, Newpreservation techniques for nutrientsamples: Office of Water QualityTechnical Memorandum 94.16,accessed May 18, 2000, at URLhttp://water.usgs.gov/public/admin/memo/QW/qw94.16.txt

________1998, Reporting level changes forvolatile organic compounds (schedules2020/2021), inductively coupledplasma–atomic emission spectrometry(ICP–AES), ammonia plus organicnitrogen and phosphorus (micro-Kjeldahl) in water methods at theNational Water Quality Laboratory:National Water Quality LaboratoryTechnical Memorandum 98.07,accessed February 23, 2000, at URLhttp://wwwnwql.cr.usgs.gov/Public/tech_memos/nwql.98-07.html

________1999, Changes in field treatmentprotocols and bottle types for whole-water samples collected for totalammonium plus organic nitrogen andtotal phosphorus determinations:Office of Water Quality TechnicalMemorandum 99.04, accessedMay 18, 2000, at URLhttp://water.usgs.gov/public/admin/memo/QW/qw99.04.txt

methods of analysis by the u.s. geological survey national water

Documents