REFRACTORIES FOR HAZARDOUS

WASTE INCINERATION AN OVERVIEW

JAMES A. CAPRIO and H. EDWARD WOLFE Harbison-Wal ker Refractories

Division of Dresser Industries, Inc.

Pittsburgh, Pennsylvania

ABSTRACT

With the ever increasing generation of hazardous wastes and recent RCRA regulations changing the

country's hazardous waste management philosophy,

incineration has become a very viable and efficient alternative to burying and recycling. Since incinera­tion systems such as the rotary kiln and liquid in­jection units seem to be the most popular and

promising designs, the reaction between the refrac­tory lining and the waste system has taken on new importance. Alkalies introduced as salts and halides

from halogenated hydrocarbons work in conjunc­tion with higher efficiency temperatures to put new strains on linings. The basic effects of these detrimental compounds on various refractories have been investigated through actual incinerator slag tests and the results of these tests are discussed in order to determine a general rule of thumb on

selecting refractory materials.

BACKGROUND -

When the Environmental Protection Agency's

Resource Conservation and Recovery Act Regula­tion (RCRA) became effective in November of 1980, it had an impact on refractories, as well as the environment.

RCRA established "cradle to grave" standards

for industrial wastes considered by regulation to

be hazardous, requiring more thorough methods of disposal and disposal monitoring. All areas of hazardous waste disposal, from chemical treatment

to landfill dumping, are covered in RCRA.

139

Hazardous waste management is a complex issue

where no single process can be applied to all waste streams. The following are several options for man­

aging waste: 1. Wastes can be recycled. 2. They can be treated and recycled. 3. Wastes can be treated and disposed. There are certain hazardous materials that are

hazardous simply because of the structure of their molecules rather than the properties of the ele­

ments contained. For these types of hazardous materials, high temperature incineration is the best

means of disposal. Carbon dioxide, water vapor and inert ash are usually all that remain after in­cineration. Because of these factors, incineration has become the fastest growing method of disposal

in the 1980's. According to the EPA, there will be

a 55 percent increase in the amount of wastes

handled through incineration in 1982 and in 1983

a 44 percent increase is projected.

In contrast, the amount of hazardous wastes disposed of in landfUls is expected to decrease by

9 percent in 1982 and chemical treatment of in­dustrial wastes should rise at a rate of only half that projected for incineration.

HIGHER STANDARDS, HIGHER

TEMPERATURES

As RCRA tightens the regulations on traditional

dumping and burying practices, waste generators and handlers look to incineration as a complete and cost-effective disposal practice, a trend already experienced in Europe. However, RCRA has in-

creased the efficiency requirements for hazardous waste incineration. The RCRA performance criteria for industrial incineration require that all incinera­tors burning hazardous wastes achieve a destruction and removal efficiency of 99 . 99 percent. For in­cinerator operators this usually means higher burn­

ing temperatures and longer burn times.

Although not required, the EPA suggests a mini­mum retention time of 2 sec at 1,830 F (1,000 C) combustion temperature, or 2,200 F (1,205 C) if the wastes contain halogens. Retention time is the amount of time that the hazardous waste molecules must be maintained at these temperatures to break down and thereby achieve the proper combustion efficiency. Until now, temperatures of 1,500 F -1,700 F (816 C - 927 C) would have been con­sidered sufficient for hazardous waste incineration.

Thus, many incinerator operators will increase efficiency by the cranking-up of furnace tempera­tures beyond limits previously considered sufficient - a move that hits home with refractory users be­cause of the new strains put on linings used in in­cinerators. Incineration of hazardous wastes has always been a complex matter when it comes to refractories, because the waste stream contains compounds which can be very destructive. And with the higher temperature needs required by RCRA, some refractories once thought adequate for an in­cinerator will have to be replaced by others. In cer­tain cases, incinerator linings just won't hold up as long as could previously be expected. But it's a sure bet that the increasing demands for environmentally safe disposal through incineration will bring new and tougher demands for refractory performance.

TOXIC CHEMICALS: A MAJOR ISSUE

Toxic chemical wastes represent one of the most serious problems facing society today. Hazardous wastes, the by-products of modern manufacturing processes, were produced,at a rate of 57 million tons (5 l .7 X 106 t) a year in 1980 and that number is rising at a rate of 3. 5 percent yearly. It is alarming to note that only 10 percent of 1980's hazardous waste was disposed of properly based on current Government regulations.

Wastes are produced chiefly from chemical, petrochemical and metallurgical industries. Paper production, printing and the electronics industries also produce some by-products in the form of hazardous wastes.

Affiong industrial wastes are groups of com­pounds called toxic organics and waste salts. One

of the main groups of toxic organics that falls under EPA regulation are the halogenated hydro­carbons, which include chemicals like polychlori­nated biphenyl (PCB), a suspected carcinogen. The halogenated compounds contain fluorides and chlorides. These can be destructive to refractories at the higher efficiency temperature required to­day.

And when alkalies in the form of salts are added to the incineration process, the burden on the re­fractories is even greater.

TYPES OF INCINERATORS

Generally speaking, there are five major types of incineration processes in use today, the most com­mon being the rotary kiln and liquid injection methods (Sketch 1) [1] . Rotary kilns have the advantage of being able to burn most forms of toxic organics, as well as the barrels which contain them. Rotary kilns generally have secondary cham­bers which burn gases at higher temperatures than the primary chambers. Operating temperatures in rotary kilns have ranged from 1,500 F - 3,000 F (816 F - 1,650 C) and these systems can handle most forms of toxic organics.

Liquid injection involves a vertical or horiz�ntal vessel with wastes atomized and sprayed through nozzles into the combustion chamber. Injection systems are limited to use with pumpable liquids and slurries and their operating temperatures range from 1,200 F - 3,000 F (649 C -1 ,650 C).

Other methods of industrial incineration are: pyrolysis, the thermal decomposition of a substance in the absence of oxygen; multiple hearth, a system of vertically stacked hearths used mainly on sewage sludge; and fluidized bed, where wastes are injected into a hot agitated bed of inert granular particles.

Lastly, a version of incineration that is recently gaining favor in the U.S. after its introduction eleven years ago in Europe is waste incineration at sea. This method of incineration, involving a fleet of vessels equipped with rotary or liquid injection incinerators, could be a vital part of tomorrow's waste manage­ment solution.

A major reason "at sea" incineration could be advantageous is its handling of the vapor form of hydrochloric acid. On land, incinerators require scrubbers to remove the HCI, which readily assumes its liquid form. However, the ocean seems to be capable of breaking down the problem gas without environmental side effects, thus eliminating the need for troublesome scrubbers [2].

140

APPROX. 30 FT.

Rotary

Kiln

Unit

APPROX.

Liquid

Injection

Unit

--- 10 FT. ---�

PRIMARY

Liquid injection systems are the most common industrial Incinerators in use in the United States. Liquid injection sys­tems, however, can only handle pumJ> able liquids and slurries, such as waste oils and chemicals, paints, solvents, dyes, soap, pesticides and chlorinated hydrocarbon wastes.

The rotary kiln Incinerator can handle most organic wastes'and is well suited for burning solids, sludges, liquids and gases. An advantage of the rotary burner is that it will consvme wastes as well as the drums which contain them. Heavy metals will also decompose, with the remaining solids dropping out in the form of clinkers. The fumes are burned at even hotter temperatures In a secondary chamber before being released into a pollution control device, such as a scrubber.

SECONDARY

SKETCH 1 •

All types of incineration processes are covered by the RCRA regulations, which only complicate the already complex refractory-waste stream rela­tionship. As efficiency specifications increase in all types of incinerator systems, so will the demand on refractories.

TESTING AND EVALUATION

ALKALIES AND HALIDES: DESTRUCTIVE TO

REFRACTORIES

At increased operating temperatures, the waste

put into an industrial incinerator can be especially

141

destructive to the lining. Two destructive compo­nents found in industrial wastes are alkalies and halides. .

Alkalies can have a detrimental effect when they react with fire-clay and high alumina refractories. The alkalies flux the alumina-silica (40-55 percent Al203) system, causing glass formation on the sur­face of the refractory. At low temperatures, this can be beneficial since the glazing seals the refrac­

tory surface (Fig. 1). However, if the temperature is raised higher, the glass could become fluid and drain from the hot face.

FIG. l GLAZED BRICK DUE TO ALKALI ATTACK

If temperature conditions are such that a glass densified zone forms, rapid cooling can cause spall­ing of the altered portion of the brick. This phe­nomenon results from structural stresses set up by the differential rates of thermal expansion between the unaltered refractory and densified zone (Fig. 2).

Therefore, if the incinerator operator attempts to comply with RCRA by increasing the tempera­ture, the refractory may need to be upgraded. How­ever, alkali reactions may still occur. Compositions higher in A1203 can show both glass formation and the presence of expansive alkali-alumina-silica phases. If potash is present, the phases typically observed are leu cite (K20.A1203.4Si02) and kalsilite (K20. A1203.2Si02). If soda is present, the expansive phase typically observed is nepheline (Na20.A1203. 2Si02). Bricks with greater than 70 percent alumina normally contain free alumina (corundum) which can also react to form an expansive alkali-alumina phase. The presence of these e�pansive phases re-

FIG. 3 EXPANSIVE REACTIONS DUE TO ALKALI ATTACK

sults in disruption of the brick structure and loss of strength (Fig. 3).

Thus, reaction can occur with all types of re­fractory materials, but certain brick classes have better resistance to reaction at certain temperatures. This has also been observed by laboratory slag test­ing (Fig. 4). In cyclic alkali cube testing at 2,200 F (1,205 C), alumina-silica compositions in the 40-55 percent A1203 range show a glazing of the surface. These samples appear relatively unaffected after six test cycles. AI; we increase alumina content, materials in the 55-65 percent range show both glazing and minor cracking after six test cycles (Fig. 5) . In the 70 percent alumina range, severe de­terioration was observed after three cycles (Fig. 6) .

FIG. 2 SPALLED GLASS DENSIFIED ZONE

142

Discussions concerning higher temperature applica­tions can be found in the conclusion and summary section.

Another group of compounds found in waste streams harmful to refractories are the halides: Chlorides, fluorides, and related compounds. Of

FIG. 4 ALKALI TEST (43 PERCENT AI,Og)

FIG. 5 ALKALI TEST (60 PERCENT A!,Og)

'!'I'lil'I'!'r -Trf ')Ches' 2 3 4 5 FiG. 6 ALKALI TEST (70 PERCENT AI,Og)

143

these, the effects of chlorine and chloride are prob­ably best understood.

Chlorine reaction with silica and alumina is a severe problem. Chlorine reacts with these oxides to form volatile chlorides. The removal of silica and alumina will increase the porosity and reduce the strength of the refractory. An example of this

can be seen by examining two high alumina compo­sitions before and after exposure to a chlorine-bear­ing atmosphere. A 99 percent A1203 composition showed loss of the fme matrix bond phase (Fig. 7). A mullite composition shows distruption of matrix and coarser fractions due to removal of silica via volatile SiCl4 formation' (Fig. 8).

(before)

(after)

FIG. 7 99 PERCENT AI203 BEFORE AND AFTER CHLORINE ATTACK (TWO PHOTOS)

144

( before)

1. mullite calcine 2. mu II ite bonded matrix 3. void

4. epoxy , , '

(after)

3. epoxy 4. altered mullite calcine

• ,

FIG . 8 MULLITE COMPOSITION BEFORE AND AFTER CHLORINE ATTACK

These volatile chlorides may flow out of the

incinerator and condense downstream where they

have been known to cause blockages in cooler areas

such as gas cleaning equipment or waste heat boilers.

When free chlorides are present with alkali, these compounds may condense within the lining forming expansive alkali-chloride phases. The presence of these phases can result in cracking of the refractory (Fig. 9). Chlorine gas will also react with the bond­ing phase in castables. If the material is calcium

aluminate cement bonded, calcium chloride is

formed. If the material is sodium silicate bonded, sodium chloride is formed. The presence of these

phases reduces the strength of the castable, plus

these phases are water soluble. Thus, in the pres­ence of water or high humidity conditions, a com­

plete removal of the bond is possible (Fig. 10-11).

145

N

-, . -

•

,

•

• •

•

• •

•

FIG.9 CRACKED SAMPLE AS A RESU L T OF ALKALI CHLORIDE CONDENSATION

The effect of fluorine is similar to chlorine.

Fluorine has been found to act as a flux in glassy

silicates, but chlorine has not, indicating that fluo­

rine is more active than chlorine. Like alkalies, halides can be extremely corrosive

in contact with most refractories. Most problems usually begin at lower temperatures downstream or

at the backside of the lining where halides can

condense. Depending upon the temperature, a dense

high-fired superduty brick (40-50 percent Ah 03)

or a dense 60 percent Ah 03 brick lining with low

porosity and a minimum of excess Al203 or Si02 will usually keep reactions to a minimum.

FIG. 10 VARIOUS CASTA BLES AFTER EXPOSURE TO CHLORIDE ATMOSPHERE

FIG. 11 CASTA BLES EXPOSED TO CHLORIDE AFTER SOAKING IN WATER FOR TWO DAYS

INCINERATOR SLAG TESTS

One way of determining which incinerator lining materials should be used is to perform slag tests us­ing a slag from an actual incinerator operation. The following is a compilation of four slags recently evaluated and tested.

1. The first example is a case in which a com­pany burned hazardous waste in a rotary kiln. Bar­rels containing hazardous wastes were dumped into the kiln which had been running successfully with a superduty brick at approximately 1,800 F (983 C). With the new RCRA regulations, the oper-

146

ating temperatures were increased to the 2,370 F-2,740 F (1 ,300 C-1 ,506 C) range, and both a super­duty brick and a 70 percent composition proved unsuitable. Analysis showed that the slag, Slag" A, " consisted of high concentrations of iron oxide and silica with minor alumina and lime (Fig. 12). At these higher temperatures, these oxides caused flux­ing of the alumina silica bond in a 70 percent Al203 system (Fig. 13) resulting in corundum and anorthite (CaO.Al2 03 .2Si02) in a low melting Fe ­Si -Ca -Al glass. In an effort to find a more suitable refractory lining material at these high tempera­tures, cup slag tests were performed at 2/) 10 F

Slag A Slag B Slag C Slag 0

Silica (SiO,) 23.90 percent 4.20 percent 23.70 percent 30.30 percent

Alumina (AI,03) 10.50 11.40 28.70 6.92

Titania (TiO,) 3.26 0.38 2.38 0.23

Iron Oxide (Fe,03) 51.90 0.41 t.20 31.50

Chromic Oxide (Cr,O,) 0.64 0.17

Lime (CaO) 6.48 0.35 0.05 4.15

Magnesia (MgO) 1.31 0.13 0.01 4.20

Manganese Oxide (MnO) 0.55

Phosphorous Pentoxide (P,Os) 0.27 31.50 18.70 9.67

Soda (Na,O) 0.90 32.40 25.20 3.14

Potash (K,O) 0.30 0.38 0.01 0.04

Chlorine (CI) trace 1.82 trace trace

Total Analyzed 100.00 percent 98.60 percent 99.90 percent 97.60 percent

FIG.12 SLAG CHEMISTRIES

FIG. 13 ALTERATION OF 70 PERCENT AI,03 BRICK WITH SLAG A

1. magnetite 2. Si -Ca -Fe phase 3. Fe-Si-Ca-AI glass 4. a north ite 5. corundum 6. altered alumina-silica calcines 7. void

(1,600 e) using 100 g of slag per cup. Results showed that an alumina-chromic oxide composition provided the best slag resistance in comparison to two 85 percent Al203 phosphate bonded compositions and a 90 percent Ah 03 composition (Fig. 14) .

147

2. Another series of slag tests were peFformed using a liquid injection incinerator slag, Slag "8," rich in phosphorous pentoxide and soda with minor alumina (Fig. 12). Testing consisted of cup slag evaluations at the operating temperature of 2,400 F (1,317 e) using 100 g of slag per cup. Testing indi­cated that a fireclay composition (42 percent Al2 03) showed glass formation and fluxing (Fig. 15). A 60 percent Al203 composition showed retention of much of the slag within the cup with only limited reaction about the cup exterior (Fig. 16).

A direct bonded magnesite-chrome composition was weak and friable after testing due to the reac­tion of the alkali with the chromite spinel which disrupted the direct bonding (Fig. 17). A 100 per­cent magnesite composition showed no reaction with the slag but the slag had penetrated the sample and filled the porosity (Fig. 18).

3. A third series of tests involved a l.iquiq injection slag, Slag "e," containing significant concentrations of soda, alumina, silica, and phosphorous pentoxide. Note the concentrations of soda and phosphorous pentoxide approximate those of "Slag "B" . How­ever, the quantities of alumina and silica are much greater in Slag "e" (Fig. 12).

Testing consisted of cup slag .tests at both 2,200 F (1,205 e) and 2,600 F (1,428 e) on the high alumina and basic brands already discussed and drip slag tests at 2,650 F (1,456 e) on selected brands.

After the 2,200 F (1,205 e) tests, all samples showed trace to no penetration or alteration. How-

FIG. 14a CUP SLAG TESTS USING SLAG "A" WITH 85 PERCENT AllO, PHOSPHATE BONDED BRICK

FIG. 14 b CUP SLAG TESTS USING SLAG "A" WITH A 90 PERCENT MULLITE

BONDED All 0, COMPOSITION (LEFT) AND AN ALUMINA CHROME COMPOSITION (RIGHT)

FIG. 15 CUP SLAG TESTS USING SLAG " B" WITH 42 PERCENT AllO,

148

FIG. 16 CUP SLAG TEST USING SLAG " B" WITH 60 PERCENT AllO,

I 2 3 4 6

FIG. 17 CUP SLAG TEST USING SLAG "B" WITH

DIRECT BONDED MAGNESITE

CHROME COMPOSITION

I 5 6

FIG. 18 "cUP SLAG TEST USING SLAG "B" WITH 100

PERCENT MgO COMPOSITION

• •

I

ever, the phosphate bonded 85 percent Al203 com­

position exhibited slight penetration (Figs. 19-22).

Mter the 2,600 F (1,428 C) cup slag tests, the

phosphate bonded 85 percent Al203 compositions

again showed the most severe alteration. The fire­clay and high alumina compositions showed a glaz­ing of the slag cup with no significant penetration

(Figs. 23-25). The 100 percent magnesite composition again

showed no reaction with the slag. However, the slag was absorbed into the brick structure (Fig. 26).

Mineralogical examination revealed that all the

frreclay and high alumina samples showed the same type of alteration mechanism, that is, solution of the bonding matrix and coarser fractions resulting

in altered alumina-silica calcines and corundum in a glassy Na - P - Al- Si matrix. The important dif­ference was the rate at which the reactions occurred. The phosphate bonded compositions exhibited the poorest resistance. The 60-90 percent Al203 ma-

• terials exhibited alteration only at the slag-brick

interface (Figs. 27. 28). As expected at 2,600 F (1,428 C) slightly greater quantities of glass were

detected in the fireclay sample (Fig., 29).

Five compositions were selected for the more

dynamic drip slag test: From the cup slag tests, we know that lower melting glassy phases were formed, but these phases just collected within the cups. The

viscosity of this glass is extremely important in re­lation to how quickly the refractory can be eroded

away. The drip slag tests consisted of dripping

1.50 kg of slag on to each test sample and meas­uring the amount of ef0sion in cubic meters.

Results again showed that two 85 percent Ai203 phosphate'bonded brick compositions were the

most altered while the 90 percent alumina compo-

4

FIG. 19 CUP SLAG TEST USING SLAG "C" AT 2, 200 F (1, 205 C) 85 PERCENT

PHOSPHATE BONDED AI203 FIRED (LEFT) AND60PERCENT AI203 (RIGHT)

149

FIG. 20 CUP SLAG TEST USING SLAG "C" AT 2,200 F (1,205 C) 85 PERCENT PHOSPHATE BONDED AI203 DRIED (LEFT) AND60 PERCENT AI,03 (RIGHT)

FIG. 21 CUP SLAG TEST USING SLAG "c" AT 2,200 F (1,205 C) 60 DOMESTIC AI,03 (LEFT) AND 90 PERCENT MULLITE BONDED AI,03 (RIGHT)

FIG.22 CUP SLAG TEST USING SLAG "C" AT 2,200 F (1,205 C) HARD BURNED SUPER DUTY 43 PERCENT AI,O, (LEFT) AND 100 PERCENT MgO (RIGHT)

150

FIG. 23 CUP SLAG TESTS USING SLAG "C" AT 2,600 F ( 1,428 C) 85 PERCENT PHOSPHATE BONDED AI,O) (LEFT) AND 60 PERCENT AI20) (RIGHT)

FIG. 24 CUP SLAG TESTS USING SLAG "C" AT 2,600 F (1,428 C) 85 PERCENT PHOSPHATE BONDED AI20) DRIED (LEFT) AND 60 PERCENT AI,O) (RIGHT)

FIG. 25 CUP SLAG TESTS USING SLAG "C" AT 2,600 F (1,428 C) 60 PERCENT DOMESTIC AI20) (LEFT) AND 90 PERCENT MULLITE BONDED AI,O) (RIGHT)

sition showed only minimal erosion at these ele­vated temperatures (Figs. 30, 31). The 60 percent Al203 compositions were glazed and showed trace to no erosion (Figs. 32, 33,34).

151

4. The last series of slag tests were performed with an iron oxide-silica-phosphorous pentoxide rich slag, Slag "0, " containing only minor soda. As. with Slag" A, " Slag "0 " represents a rotary kiln

FIG. 26 CUP SLAG TESTS USING SLAG " c" AT 2,600 F ( 1,428 C) HARD BURNED SUPER DUTY 43 AI203 (LEFT) AND 100 PERCENT MgO (RIGHT)

(unused) (altered)

2. Na -P -AI- Si - Ca - Fe glass

5. void

FIG. 27 MINERALOGICAL EXAMINATION OF 85 PERCENT PHOS· PHATE BONDED AI203 (UNUSED AND ALTERED) AFTER SLAG

TEST WITH SLAG "c" AT 2,600 F (1,428 C)

152

•

,,,

•

, . , • •

. ' .' J' '0 • �.. . . '

• , • •

• ..

• . .,:... .

• • , .

. . " . � ... • . .� ..... •

,/f' ... . ' • . . • '1 .. /. ,,� . " . "' .... . .. . . '

. " - ,

. , , . . •

,

"

•

•

• ,

(unused)

4. Alumina·silica calcine 5, mullite bonded matrix

6. void

•

,

, •• •

." . • • .. . ...... ' • ..

t .. •

(altered)

1. densified zone 2. corundum 3. Na·P·AI-Si-Fe-Ca glass

4. alumina-silica calcines 5 .. mullite bonded matrix

6. void 7. epoxy

FIG. 28 MINERALOGICAL EXAMINATION OF 60 PERCENT DOMESTIC AI,03 (UNUSED AND ALTERED) AFTER SLAG

TEST WITH SLAG "C" AT 2,600 F (1,4 28 C)

type operation where barrels of waste are incin­erated. However, the quantities of phosphorous

pentoxide and soda are greater in Slag "D". Both cup and drip slag tests were performed on

a fireclay composition (42 percent Alz03) and two 60 percent Alz03 compositions.

The cup tests were run at 2,375 F (1,303 C) with a 24-hr hold using 0.1 kg of slag per cup. Microscopic examination revealed that an alteration occurred in all compositions similar to that pre­

viously described: solution of the bonding matrix

and coarser fractions resulting in altered calcines in

a glassy matrix. Significantly greater slag penetra­

tion was evident with the fireclay composition than

with the 60 percent Al203 brick at these tempera­

tures (Fig. 35).

153

The drip slag tests were run at 2,400 F (l ,317 C)

using 1.200 kg of slag per sample. Greater penetra­tion and erosion was again evident with the hard burned fireclay composition (Fig. 36).

To further identify possible reasons for the dif­ference in slag resistance of the three brands, addi­

tional tests were run on each sample. Physical prop­erties showed that the 60 percent Al203 composi­tions had a finer pore size distribution and a lower permeability than the fireclay material, even though the apparent porosity of the 60 percent Al203 com­

position is higher than the hard burned fireclay.

Microscopic examination of the unused samples also

revealed that the 60 percent compositions had a

more continuous interlock mullite crystallite net­work and lower quantities of glass than the fireclay

(unused)

3. fi reclay calc ine 4. mullite and glass

bonded matrix 5. void 6. epoxy

(altered)

1. altered coarse alumina-silica calcines 2. Na -P -AI- Si glassy matrix 5. void 6. epoxy

FIG . 29 MINERALOGICAL EXAMINATION OF 43 PERCENT AI,O, HARD BURNED BRICK (UNUSED AND ALTERED) AFTER SLAG TEST WITH SLAG "C"

AT 2,600 F ( 1 ,428 C)

1�'II' I 21' r 31 I I 41' I 51

FIG. 30 DRIP SLAG TESTS AT 2,600 F (1,428 C) WITH SLAG "C" 85 PERCENT PHOSPHATE BONDED AI,O,

FIRED

r I I ' I ( i ! I I I j I I I I I i I I I I inches I 2 :3 4 5

FIG. 31 DRIP SLAG TESTS AT 2,600 F (1,428 C) WITH SLAG "C" 85 PERCENT PHOSPHATE BONDED

AI,O, DRIED

154

brick. These properties would minimize slag pene­

tration and lengthen service life at these tempera­tures (Fig. 37).

CONCLUSIONS AND SUMMARY

HOW TO CHOOSE AN INCINERATOR LINING

An example of how RC RA impacts on refrac­tories is the hypothetical case of a chemical waste

incinerator using a superduty brick. Before the

EPA regulations were enacted, it probably would

have heen run between 1,500 F (816 C) and 1,700 F (9 27 C) and superduty would work ex­tremely well in the presence of chlorinated hydro­carbons and alkalies. But as the temperature is raised to meet the RCRA efficiency level, problems could occur. At the higher temperatures the alkalies could flux the brick, causing it to become soft and

to melt. If high alumina brick were substituted for superduty, expansive reactions could develop, but service life would be longer. If the incinerator tem­perature and salt content is extremely high, an inert basic brick would probably be tried if thermal cycling, expansion and moisture problems could be accommodated.

As the above situation illustrates, the following two variables should be taken into consideration while choosing a lining:

Phosphate Bonded High Alumina

(fired)

Drip Slag Test at 2,650 F (1 ,456 C) Using 1 ,500 kg of Synthetic Slag Volume Eroded, m

3 4 . 1 X 10-5

FIG. 32 DRIP SLAG TESTS AT 2,600 F (1,428 C) WITH SLAG "C" 90 PERCENT MULLITE BONDED AI,03

FIG. 33 DRIP SLAG TESTS AT 2,600 F ( 1,428C) WITH SLAG "C" 60 PERCENT AI203

Phosphate Bonded High

60 percent Alumina 60 percent 90 percent AI203 (Dried) AI203 AI203

Trace 7.9X 10-5 Trace 2.1 X 10-5

FIG. 34 SLAG TEST RESULTS

FIG. 35 CUP SLAG TESTS AT 2,375 F ( 1,303 C) WITH SLAG "D" (FROM LEFT TO RIGHT: 60 PERCENT AI,03' 43 PERCENT AI203, 60 PERCENT AI203

155

1" ' 1 i nches I

12\' 13\1 14

(60 percent AI, °3)

(60 percent AI,O,)

FIG.36 DRIP SLAG TESTS AT 2,400 F (1,317 C) WITH SLAG "0" (SEE PHOTOS FOR IDENTIFICATION)

156

60 percent Unused 42 percent Unused 60 percent AI,03 (Published AI203 (Published AI203

Sample Designation A Data) B Data) C

kg/m3 2,547 2,499 2,563 2,323 2,307 2,371 2,563 Bulk Density, pet Apparent Porosity, percent

Air Permeability *

Pore Size Distribution (Method 4P)

Percent Pores Finer Than 60 Jlm 40 Jlm 20 Jlm 10 Jlm

4 Jlm 2 Jlm

Jlm 0.5 Jlm

Drip Slag Tests At 2,400 F (1,317 C) Using 1.200 kg gms of slag

Volume Eroded, m3

159 14.1

0.16

100 99 94 79 61 40

7 2

0.4 X 10- 5

156 to 160 145 13'to 16 11.2

0.88

98 66 26 10

8 7 7 7

1.2 X 10-5

'Permeability is measured as cu in./sec/sq in. of area/in. thickness/psi pressure.

144 to 148 160 9 to 12.5 11.8

0.05

100 99 97 67 18

9 9 8

0.5 X 10-5

FIGURE 37

1. Maximum operating temperature of the fur­

nace (RCRA). 2. Chemical analysis of the feed. ln general, a rule of thumb in selecting hazardous

waste incineration refractories is as follows: Let the

temperature dictate the Alz03 content of the lining. For lower temperature applications, lower AlZ03 materials generally lengthen service life. However,

as the temperature increases, the AlZ03 content of the lining should also increase. The rule of thumb is meant to accommodate the incineration of halide and alkali bearing compounds in the following

specific units.

ROTARY KILN SYSTEMS

As shown in the analysis of Slags" A" and "D," the

' waste stream can contain high quantities of

FeZ03 from the drums, alkalies from the salt wastes, halides, and phosphorous pentoxide. At operating temperatures in the 1,800 F (983 C) range, U.S.

and European research and experience indicates that a high Hred superduty fireclay should minimize reactions and maximize service life [3,4]. At these low temperatures, the excess alumina in a higher

alumina material could combine with the alkalies and halides to form destructive, expansive reactions.

157

However, as the temperature in the kiln increases to

the 2, 100 F-2,500 F (I ,150 C-l ,372 C) range, the

fireclay can be fluxed by the alkalies and damaged

by the increased reactiveness of the Fez 03, As

illustrated by tests with Slag" D," brick linings in

the 60 to 70 percent Alz 03 range have been a cost effective approach under these temperature situations. Yet, temperatures approaching the 2,700 F-2,900 F (I ,483 C -1,595 C) range should substantially shorten the life span of a 70 percent Alz03 kiln lining. As shown under Slag "A," lin-' ings of higher Alz 03 content should increase ser­

vice life. However, a solid solution bonded alumina­chromic oxide composition was shown to minimize the reaction mechanism of the Fez 03 at these temperatures.

The secondary chamber of the rotary kiln sys­tem can be exposed to similar temperatures and contaminants. Therefore, the same rule of thumb concerning temperature and alkali/halide reactions would apply. However, the secondary chamber will not be exposed to high quantities of Fez 03, negat­

ing the need for the alumina-chromic oxide ma­

terial at very high temperatures. The cooler areas of the system will probably re­

quire acid resistant refractory materials because of

acid condensation. Experts on acid resistant mason­ry should be consulted.

LIQUID INJECTION SYSTEMS

As shown in the analysis of Slags "B" and "C," the quantities of alkalies introduced in the liquid units are generally much higher than those intro­duced in rotary systems. Hence, the refractory will be subject to very distructive conditions as the tem­perature increases.

In general, the same rule of thumb should apply to most refractory selections in liquid units. At low temperature in the 1,800 F-2,000 F (983 C-1 ,094 C) range, high fired superduty fireclay materials should maximize lining life. Along with the 2,200 F (1,205 C) tests with Slag "C," both U.S. and Euro­pean researchers have performed alkali tests under 2,000 F (1,094 C) and found that fireclay materials with 40-50 percent Al203 and less than 15 percent porosity show excellent resistance against alkali attack [4]. At these temperature, increased Al203

content does not seem to increase service life. However, operating temperatures of 2,200 F-

2,500 F (1 ,205 C-1 ,372 C) will cause increased reac­tion at the hot face, resulting in the fluxing of the fireclay. Research has shown that a 60 percent Al203 brick should maximize life in the units with slags similar to Slags " B" and "C". Higher Al203

materials and pure mullite composition do not de­crease reactions at these temperatures. In fact, the higher Al203 materials could develop expansive reactions and crack prematurely.

In liquid units that operate in the range of 2,500 F-2,800 F (1,372 C-1,539 C) the 60 Al203

materials are sometimes fluxed. Experience shows that a high purity 90 percent Al203 brick minimizes reactions at these high temperatures. This observa­tion concurs with the 2,600 F (1,428 C) drip slag tests with with Slag "C".

PRECAUTIONS

Basic Brick

In tests with Slags " B" and "C," magnesite chrome and magnesite brick show little or no reac­tion with alkali/halide slags due to the chemical inertness of the brick composition. Yet, the slag will penetrate the basic brick, possibly forming a spalling situation due to differences in thermal expansion. In very high temperature units where excess alkali attack is a problem, basic brick could be an alternative. However, the design must accom-

modate expansion and the brick's tendency to hydrate when exposed to moisture.

Phosphate Bonded Al203 Brick

During tests with Slags " A" and "C, " 85 percent Al203 phosphate bonded brick compositions showed a higher degree of reaction than the 90 percent Al203 high purity mullite bonded brick. Whether the phosphate bonded bricks were fired or merely dried, the phosphate additions appear to act as a flux, increasing the alkali-slag reaction. Thus, the higher purity 90 percent Al203 mullite bonded brick should be a better choice in areas of rotary or injection systems that experience temperatures exceeding 2,500 F. Experience in the U.S. tends to confirm this approach.

Monolithic Materials

In most cases, because the ceramic matrices of pressed and fired brick are generally more dense and less porous, brick resist alkali reactions better than castables and plastics. Research has shown that the ceramic bond in fired refractories is less susceptible to attack by alkalies and chlorides than the phosphate, cement or clay bonds in monoliths. When monoliths must be used, however, there are criteria for choosing the material with the highest resistance to attack. Some fireclay based castables have additions to form glasses which help seal the surface against further attack. These will be effec­tive at the lower temperature ranges (1 ,800 F -2,200 F) (1,983 C-1 ,205 C) where glasses are very viscous. In most cases phosphate bonded plastics should be avoided for much the same reasons as were dis­cussed for brick, I.e., the phosphate could act as an internal fluxing agent speeding the attack of alkalies. In limited tests, clay bonded plastics have performed better in resisting alkalies in the 1 ,800 F-2,400 F (983 C-1 ,317 C) range because, as was discussed earlier, a glassy seal forms at the surface, halting further penetration. In this temperature range, the more resistant clay bonded plastics and calcium aluminate bonded castables have Al203

contents between 40 and 60 percent. Chlorides, as was discussed previously, are ex­

tremely detrimental to all types of plastics and castables. If monoliths must be used, plastics should perform better than castables. Chlorides will attack the calcium-alumina bond in the cement forming calcium chloride salts which are soluble in water. The high alumina (70-99 percent) plastics and castables should be avoided because of chloride reactions with free corundum (present in most high

158

alumina refractories). The plastics in the 40-60 per­

cent alumina range, because of higher densities and

fmer pore size distribution, should give better ser­vice than most castables.

Viscosity

Although temperature and waste chemistry are the most critical furnace variables, viscosity of the waste slag at temperature is very important. An alkali-laden waste with a low viscosity will move

quickly across the face of the refractory, maximiz­

ing wear by washing away the reacted material and the mortar. In contrast, a sticky viscous slag will

move slowly across the refractory face, enabling

some refractory brick to form a protective coating and minimize wear. In most cases, the incinerator

operator should attempt to strike a balance between

a very viscous slag that could clog the unit and a

flowing slag stream that shortens refractory life.

Design

The selection of the refractory based on tem­

perature and waste chemistry is the most important

component of a successful refractory lining. Usually,

alkali attack from salt wastes is the most destruc-

tive reaction that limits service life. A brick lining

selected according to the above rule of thumb should

maximize performance. In most cases, the brick

selected to combat alkali reactions will also minimize damage from halides.

Thermal shock, expansion allowance, mortar usage, installation, thermal gradient and overall brick configuration are also important variables

and should be discussed with refractory experts.

REFERENCES

[1] Shen, T. T., Chen, M., and Lauber, J., "Incinera­

tion of Toxic Chemical Wastes," Pollution Engineering,

October 1978, p. 45. [2] Piellisch, R., "Waste Disposal Problem Spurs New

Interest in Incinerator Ships," Chemical Marketing Re­

porter, September 29, 1980, p. 39. [3] Fabian, H. W., Schon, M., Capek, K., and Lever­

kusen, "Problems in the Incineration of Refuse from the

Chemical Industry."

[4] Subramanyam, A. V., Aetekar, V. A., Narasimhan,

M. D., "Alkali Attack by Molten Salts on Alumina-Silicate

Samples," Refractories Journal.

Key Words

Efficiency

Hazardous

High Temperature

Incineration

Refractory

Rotary Kiln

Slag

159