Incineration Treatment of Arsenic-Contaminated Soil

Larry R. Watdand Chady King Marta K. Richards Robert C. Thurnau

hwy R Wderkznd It mmwger of incineration tecbndogy devebpment pmgrams f'Acurer Corporation in Mountain V k , c4, Cbarly King supervises :be Acrrrex Emlssioru Measurement Croup :bat operates EPA's Incineration Researcb Facility in J e f l m o q AR Robert C Tbumau supervises :be T b d Researcb Section witbin EPA's Risk Reduction Engineering Laboratory in Cindnnatr, OH. Marta R Ricbards It a reseatrb cbemist on :be metmal Researcb Section s t a s

An incinetalion testpmgram wascondutedat the US. Envinmmental Protectionqgency 3 Incineration Research Facility to evaluate thepotential of incineration as a treatment option for contaminated SOUS at tbe Baird and McGuitv SuperJund site in Holho4 Massachwem. Thepuqme of these tests was to evaluate the inciwabiliiy of thesesoils in terms of the fate of arsenic and kad and the destructiim of organic contaminants dumg the incineration process. The tatprogram consisted of a series of bench- scale expajments witb a mumfurnace and a serfes of incineration tests. in a pilot-scale tvtary kiln incineratorsystem.

The study repotted in this paper was funded by the Envimmental Pmtection Agency under Contract 68-CY-0038 to Acurex Cotporation. It has been subjected to the Agency's nwkw and has been approyed for publication. Mention of tr& numes or commerclalproducts does not constitute dorSement or recommendation for we.

One of the primary missions of the Incineration Research Facility OW) is to support EPA Regional Offices in evaluating the potential of incinera- tion as a treatment option for contaminated soils at Superfund sites. One priority site in Region I is the Baird and McGuire site in Holbrook, Massachusetts. EPA Region I requested that test bums be conducted at the IRF to support evaluations of incineration as a treatment technology for the contaminated soil.

The soil at the Baird and McGuire site is contaminated with low levels of several pesticide compounds, and varying levels of arsenic and lead, Several areas of the site have arsenic contamination levels on the order of 100 ppm. Two "hot spots" have arsenic levels up to 3,800 ppm. Lead contamination in the soil is on the order of a few tens of ppm. Thus, with respect to incinerability evaluation, the primary concern surrounds the fate of arsenic and lead in the soil when it is subjected to incineration. The effect of incineration on the fate of arsenic and lead in soil is currently unknown. A second concern relates to whether incineration can effectively destroy the organic pesticide contaminants in the soils. Therefore, the test conditions were designed to evaluate the effects of varying incinerator operating conditions on organic contaminant destruction and on the fate of the arsenic and lead in the soil. Specifically, the test program attempted to answer these questions:

RBMBDI~TIoN/SPRING 1991 227

LARRY R WATERLAND CHARLY KING MARTA K. RICHARDS R D B T C. THURNAU

What is the distribution of arsenic and lead in the discharge streams

To what extent can rotary kiln incineration effectively destroy the

What are the effects of excess air and temperature on organic

during incineration of this metal-contaminated soil?

organic constituents in the soil?

constituent destruction and arsenic and lead distribution?

The test program consisted of two components. Initially a series of bench-scale experiments, using a muffle furnace, was performed to evaluate the leachability characteristics of the arsenic and lead in the soil as a function of the arsenidlead concentration in the soil. The second component of this test program consisted of a set of five incineration tests in the rotary kiln incineration system (RKS) at the IRF. These tests were aimed at evaluating the fate of arsenic and lead in the soil as a function of kiln temperature and excess air level.

MUFFLEFURNACEEXPERIMENTS A typical goal for any on-site remediation treatment process is that the

residue from the process (the treated soil) be able to be landfiiled at the site. For incineration treatment, this would not be possible if the kiln ash residue had TCLP leachable arsenic and lead at levels greater than the toxicity characteristic (TC) limit. In an actual site remediation, soils with very high arsenidlead levels can be blended with soils of low arsenidead contamination to give an incineration feed that results in a low concentra- tion of leachable arsenidead in the kiln ash. But, the aptjorlunknown is how low the feed arsenic or lead concentration must be. To address this unknown, a series of muffle furnace tests was performed. The objective of

were designed to

o f varying incinerator operating conditions on

the erects

Ill organic contaminant destruction and on the fate of the arsenic and lead in the soil*

these tests was to develop data to guide the determination of appropriate maximum feed arsenidead concentrations.

For these tests, the primary variable was the arsenidead concentration in the test mixture. Contaminated soil containing 650 ppm arsenic and 45 ppm lead was mixed with various amounts of a background soil containing less than 5 ppm arsenic and 14 ppm lead to produce seven samples of varying arsenidead concentrations. Each sample was heated in a muffle furnace at 982OC (1,800OF) for 1 hour. Analysis of the soil mixtures and the resultant ash residues showed that

Arsenic volatility increased with soil arsenic concentration. Similarly,

Soils containing less than 150 ppm arsenic produced ash residues

Lead content in all ash samples was constant at about 5 ppm. All ash TCLP leachate lead concentrations were below detection,

Organics and moisture in the soil contributed to 25% weight 10s.

lead volatility increased with soil lead concentration.

below the arsenic TC Iimit of 5 mditer.

regardless of initial soil lead content.

To achieve a secondary objective of determining whether the potential additives lime and alum can affect the distribution of metals to the resulting

228 REMBDIATION/SPRXNG 1991

INCKNBRATION ~ T M B N T OF A R S B M C - C D Sou

soil ash, two additional tests were conducted. In both tests, the mixture consisted of the highly contaminated soil and 2% (by weight) of one of the additives. Analysis of the limited data suggests that:

Lime appeared to reduce the volatility of the arsenic; a greater

Alum appeared to increase zrsenic volatility; less soil arsenic

Neither additive affected the volatility of lead. Lime also decreased the leachability (fraction leachable) of resulting ash arsenic to the extent that lime may be added to soils with arsenic concentrations greater than 150 ppm while yielding a thermal treatment ash that would not possess the toxicity characteristic.

fraction of the soil arsenic remained with the resulting ash.

remained in the resulting ash.

PILOT-SCALE INCINERATION The pilot-scale incineration tests were conducted in the RKS to evaluate

the fate of arsenic and lead in the soil as a function of kiln temperature and excess air level. Figure 1 is a schematic of the RKS. The design character- istics of the system are summarized in Table 1. The RKS consists of a rotary kiln primary combustion chamber followed by an afterburner chamber. Downstream of the afterburner, the combustion gas is quenched by direct water injection. The gas then flows through a primary air pollution control system (APCS). For these tests the primary APCS consisted of a single-stage ionizing wet scrubber fabricated by Air Plastics Company.

~~~~~~ ~

Figure 1. Schematic of the Rotary Kiln System

I I 8 1 8

I

-----

I I I

I SYSTEM I I

I MODULAR PRIMARY AIR REDUNDANTAIR I ROTARY KILN PouuTloNcoNTRoL ~ P O U U W N ~ O L

DEVICES INCINERATOR

~~

RE~~RDIATION/~PRING 1991 229

Table 1. Design Characteristics of the IRF Rotary Kiln System for the VenturVPackedColumn Scrubber Trace Metals Tests

Charncte?istics of the KUn Main Chamber

Length, outside Diameter, outside Length, inside Diameter, inside Chamber volume Construction Refractory

Rotation Solids retention time Burner

Primary fuel Feed system

Liquids Sludges Solids

Temperature

2.61 m (a ft, 7 in) 1.22 m (4 ft) 2.44 m (8 ft> 0.95 m 0 ft 1-l/2 inY 1.74m' (61.4 ft) 0.63 an (0.25 in) thick cold rolled steel 12.7 cm (5 in) thick high alumina castable refractory, variable depth to

Clockwise or counterdockwise 0.2 to 1.5 rpm 1 hr (at 0.2 rpm) American Combustion Burner, rated at 880 kW 0.0 MMBtu/hr) with

Propane

produce a frusvoconical effect for moving solids

dynamic 0, enhancement capability

Positive displacement pump via water-cooled lance Moyno pump via front face, water-cooled lance Metered twin-auger screw feeder or fiber pack ram feeder 1,010" C (1,850" F)

Characteristics of the AJerbumer Chamber

Length, outside Diameter, outside Length, inside Diameter, inside Chamber volume Construction Refractory Gas residence time Burner

Primary fuel Temperature

3.05 m (10 ft) 1.22 m (4 ft) 2.74 m 8 ft) 0.91 m 0 fc) l.8Om3 (63.6 fc9 0.63 cm (0.25 in) thick cold rolled steel 15.24 cm (6 in) thick high alumina castable refractory 1.2 to 2.5 sec, depending on temperature and excess air American Combustion Burner, rated at 440 k W 0.0 MMBtuh) with

Propane dynamic 0, enhancement capability

1,200" c (2,200" F)

CharacteristCcs of the Air Pollution Control System

System capacity

Pressure drop Inlet gas flow

Venturi scrubber Packed column

Venmri scrubber Packed column

107 m3/min 0773 acfm) at 1200" F) and 101 kPa (14.7 psia)

7.5 kPa 00 in WC) 1.0 Wa (4 in WC)

77.2 Wmin (20.4 gpm) at 69 kPa (10 psi& 116 Vmin 00.6 gpm) at 69 kPa (1Opsig) Feedback control by NaOH solution addition

Liquid flow

pH control

After the primary APCS, the flue gas passes through a secondary APCS consisting of a demister, an activated-carbon adsorber, and a high- efficiency particulate (HEPA) filter. The treated flue gas is discharged to the atmosphere via an induced draft fan and stack.

TESTPROGRAM Four tests were performed at different combinations of kiln tempera-

ture (nominally 816" and 980°C [1,500° and 1,800OFJ) and kiln exit flue gas 0, (nominally 6% and 1M). A repeat fifth test was completed at the test conditions that produced the kiln ash with the lowest levels of TCLP leachable arsenic and lead.

A bulk sample (nominally 1350 kg, 3000 Ib) of arsenic-contaminated soil was excavated from the Baird and McGuire site to serve as the test waste. The bulk soil sample was packaged into four 55-gallon drums and shipped to the IRF for testing. At the IRF the soil was repacked into polyethylene-lined 1.5-gal fiber pack drums. Each fiber pack drum held about 4.5 kg (10 lb) of the test soil. In the tests, one fiber pack drum was fed into the RKS with a ram feeder every 5 minutes. Thus, test soil feedrate was nominally 55 kg/hr (120 lbhr). A kiln rotation speed of 0.65 rpm produced a solids residence time in the kiln of about 0.5 hour. Figum 2 identifies the sampling locations for the tests and summarizes the sampling protocol employed.

TESTRESULTS Throughout the test program, maximum CO levels at the scrubber exit

and the stack were a few ppm. Total unburned hydrocarbon levels were similarly low at the afterburner and scrubber exits and in the stack. Average NOx concentrations at the scrubber exit ranged from 32 to 51 ppm, levels that are typical for the rotary kiln system.

Flue gas particulate concentrations ranged from 6 to 17 mg/dscm (at 7% 03 at the scrubber exit. In the stack, concentrations ranged from 2 to 29 mg/dscm (at 7% 03. These levels are below the federal hazardous waste performance standard of 180 mg/dscm (at 7% 03.

The only organic contaminants found in the test soil above method detection limits of 2-4 mgkg were p,p' DDE, p,p'DDD, p,p'DDT, and methoxychlor. As shown in Table 2, DDE was present at 39-74 mghg, DDD at 181-310 mg/kg, DDT at 257-501 mg/kg, and methoxychlor at 5 4 81 mghg. None of these was present in the TCLP leachates of the test soils at a quantitation limit of 0.02 mdliter.

Organic analysis of kiln ash, kiln ash TCLP leachate, scrubber blowdown, and scrubber exit flue gas Method 0010 samples for each test showed that all semivolatile organic hazardous constituents analyzed, including the pesticide contaminants in the test soil, were present at less than method detection limits in all cases. The quantitation limits of the Method 0010 sampling trains, when combined with measured flue gas flow rates and soil feed concentrations and feed rates, confirm that incineration destruction and removal efficiencies (DREs) for the pesticide contaminants in the soil feeds were >99.5% to >99.97% for the tests performed.

23 1

Figure 2.

Sam

plin

g Mat

rix

I

AFE

R- - QU

ENC

H -

lON

lZlN

0 -

BURN

ER

SEC

TIO

N

WET

SCR

UBBE

R KI

LN

I L

CA

RB

ON

- H

EPA

BED

FI

LTER

D

EM

WE

R ---C

1 2

3 4

5 6

7

a A

i

i

I b

Ioni

zing

wet

C

antin

uous

00

10

MS/

m&

Is

trai

n M5

Sam

plin

g K

iln

scru

bber

flu

e ga

s (s

emiv

olat

ile

(par

ticul

ate,

(p

artic

ulat

e po

int

Was

tefd

as

h b

low

dm

m

onito

ring

or

gani

cs)

arse

nic)

an

d H

CI)

1 X

2 X

3 X

X

X

X

X

X

X

X

INCINERATTON TaSATMBm OP ARSBNIC-COKTAMMATBD SOIL

. . . kiln temperature ha9 a clear effect on both arsenic and lead distributions in that kiln ash concentrations of arsenic and lead were lower at the higher incineration temperatures.

~~ ~

Table 2. Semivolatile Organic Hazardous Constituents in Test Soils

Soil

Constituent Drum 15 Drum 16 Drum 18 (m@g> vest 1) (Test 2) crest 3) p,p'-DDE 54 74 45 p,p'-DDD 228 310 197 p ,p'-DDT 334 50 1 247 Methoxychlor 81 73 54 All other semivolatiles analyzed c4 <4 <4

Drum 13 (Test 4)

39 181 277 73

<4

Table 3 summarizes the distributions of arsenic and lead among the incinerator discharge streams (kiln ash, scrubber blowdown, and scrubber exit gas), expressed as fractions of the total metals measured in the three discharge streams. The data in Table 3 show that kiln temperature has a clear effect on both arsenic and lead distributions in that kiln ash concentrations of arsenic and lead were lower at the higher incineration temperatures. Higher incineration temperatures would be conducive to greater volatilization of these metals in the kiln, resulting in decreased kiln ash concentrations. Scrubber blowdown and scrubber exit flue gas concentrations of both metals appear to be increased at the higher incineration temperature. Again, this is consistent with increased metal volatilization in the kiln at the higher temperature.

The data in Table 3 show no clear influence of kiln excess air level (as reflected in kiln exit flue gas 03 on arsenic or lead distributions among the discharge streams at the low kiln temperature conditions. At high kiln temperature conditions, increasing excess air had the apparent effect of increasing the volatility of arsenic, as measured by the decreased kiln ash arsenic fraction. However, this apparent increase is most likely the result of increased ash entrainment and carryover by the increased kiln exit flowrate.

Scrubber inlet flue gas concentrations were not measured in these tests. However, if it is assumed that the total amount of metal measured in the sum of the scrubber liquor and the scrubber exit flue gas equals the amount present at the scrubber inlet, an apparent efficiency can be calculated. This apparent scrubber efficiency is ([scrubber liquor fraction)/(scrubber liquor fraction plus scrubber exit flue gas fraction]).

As shown in Table 4, apparent scrubber collection efficiencies for arsenic ranged from 82% to 92% and were comparable to overall particulate collection efficiencies. The apparent collection efficiencies for lead were significantly lower and ranged from 33% to 43%. Neither arsenic nor lead

233

LhRap R WATBRIAND CHARLY KING MARTA K RICHAXDS w RDBBRT C. THURNAU

Table 3. Normalized Metal Discharge Distributions for the Baird and McGuire Incineration Tests

Test 1 5 2 3 4 (9-26-89) (10-5-89) (9-29-89) (9-27-89) (9-2889)

Kiln exit temperature OC 832 839 844 994 994

Kiln exit 0, % 11.3 11.2 6.8 10.4 7.5 (OF) (1529) (1541) (1552) (1822) (1822)

Kiln exit flue gas flowrate 22.8 22.6 11.1 34.8 21.6 (acdmin).

Arsenic Kiln ash Scrubber liquor Scrubber exit flue gas

Total

Lead Kiln ash Scrubber liquor Scrubber exit flue gas

Total

a/ Actual wet m3/min.

Dtmibution (% of metal measured)

72 23 5

100

89 4 7

100

66 29 5

100

91 3 6

100

76 22 2

100

93 3 4

100

36 55 9

100

69 12 19

100

56 38 6

100

69 13 18

100

collection efficiencies showed any significant variation with test variables (kiln temperature or excess air).

Table 5 shows that, with both increased kiln temperature and decreased kiln excess air, the leachability of the kiln ash arsenic was higher, with excess air level appearing to have the more significant effect. At 11% kiln exit 0,, between 8.3% and 13% of the kiln ash arsenic was leachable. At 7% kiln exit 0,, 28% and 67% of the arsenic was leachable, for kiln temperatures of 844" and 944OC, respectively. These tests suggest that, to minimize arsenic leachability, the appropriate incineration conditions are low-temperature/high-excess air.

In contrast, the Ieachability of lead from the kiln ash was consistently low, and lead was not detected in any of the soil and ash TCLP leachates.

CONCLUSIONS Test conclusions include:

Both arsenic and lead remained predominantly in the kiln ash when

INCINERATION TREATMB~ OF ARSBNIC-CONTAMINATBD Son,

Table 4. Apparent Particulate and Metal Scrubber Collection Efficiencies

Test 1 5 2 3 4 (3-26-89) (1 0- 5-89) (3-29-89) (3-27-89) (3-28-89)

Kiln exit temperature OC 832 839 844 994 994 (OF) (1529) (1541) (1552) (1822) (1822)

Kiln exit 02, % 11.3 11.2 6.8 10.4 7.5

Apparent scmbber ColIeCtron emiency f%)

Arsenic

Lead

Overall particulate

82 85 92 86 86

36 33 43 39 42

92 84 95 90 82

incinerated at a kiln temperature of nominally 840°C (154OOF). Between 66% and 76% of the arsenic discharged and 890h to 93% of the lead discharged were accounted for in this stream. Between 2% and 5% of the arsenic and 4% and 6% of the lead were present in the APCS exit flue gas; 22% to 23% of the arsenic and 3% to 4% of the lead were collected by the APCS.

Lead remained predominantly in the kiln ash when incinerated at a kiln temperature of nominally 990°C (182OoF), with 69% of its amount discharged accounted for in the kiln ash. The APCS flue gas contained 18% to 19% of the discharged lead. The APCS collected 12% to 13% of the lead. A significant amount of arsenic left the kiln in the exit flue gas when the soil was incinerated at a kiln temperature of nominally 990°C (182OOF). At this higher temperature, the kiln ash fraction was reduced to between 36% and 56% by arsenic volatilization andor entrainment. Most of this escaping arsenic was collected in the APCS (38% to 55% of the arsenic discharged is accounted for in the APCS). Between 6% and 9?h of the arsenic discharged is found in the APCS exit flue gas. Incineration at both kiln temperatures noted above effectively

destroys the organic contaminants in the soil. Pesticide constituent contaminants were reduced from soil levels ranging from 39 to 501 mag, to less than 0.4 mg/kg (the detection limit) in the kiln ash. The APCS blowdown discharge contained no detectable pesticide con- stituents at a detection limit of 0.02 mgAiter. No detectable pesticide constituents were found in the APCS exit flue gas at detection limits

REMEDIATION/SPIUNG 199 1 235

L ~ R Y R WATERLAND 0 CHNUY KXNG MARTA K. RIC~IABDS ROBERT C. THURNAU

Table 5. Arsenic Fractions-TCLP Leachable

Test 1 5 2 3 4 (9-26-89) (10-5-891 (9-29-89) (9-27-89) (9-2889)

Kiln exit temperature O C 832 839 844 994 994 (OF) (1529) (1541) (1552) (1822) (1822)

Kiln exit 0, % 11.3 11.2 6.8 10.4 7.5

Fraction of A m & Leachable (%)

Soil feed

Kiln ash

~ ~~ ~~

2.2 2.4 2.4 2.5 2.9

9.3 8.3 28 13 67

of nominally 6 pg/dscm. Increased incineration temperatures caused increased volatilization of both arsenic and lead in the kiln, with the result that kifn ash fractions were decreased as noted above. Changes in kiln excess air level did not affect lead distributions among incinerator discharges and did not affect arsenic distributions at the low temperature (nominally 840°C [15400FD kiln condition. Increasing kiln excess air from a kiln exit flue gas of 7.5% to 10.4% apparently decreased the amount of arsenic discharged in the kiln ash from 56% to 36% of the discharged amount, with a corresponding increase in the APCS collected fraction from 38% to 55%. Changing incineration conditions had no effect on APCS apparent collection efficiency for either arsenic or lead. APCS apparent arsenic collection efficiency was in the 82% to 92% range, the Same as for overall particulate. APCS apparent lead collection efficiency was lower, in the 33% to 43% range. Kiln ash lead was not leachable using the TCLP test CCLP leachates contained no detectable lead at a level of 0.05 mgfliter). Between 9?h and 62% of the kiln ash arsenic was mobile, and found in the TCLP leachate. Increasing kiln temperature marginally increased kiln ash arsenic leachability. Decreasing kiln excess air significantly increased kiln ash arsenic leachability.

Other conclusions from the incineration tests include:

The observed relative volatilities of arsenic and lead agree with expectations from physical vapor pressure data; arsenic was signifi- cantly more volatile than lead at both kiln temperatures tested. No incinerator discharge stream (kiln ash or APCS blowdown) had

236 RBMBDIAToN/SPBING 1991

INCINBRATION TREATMENT OF ARSBNIC-GWLWI"TTED Son.

TCLP leachate concentrations exceeding TC limits for either arsenic or lead. Particulate emissions after the APCS were significantly below the federal hazardous waste incinerator performance standard.

Further conclusions from the muffle furnace tests include:

Adding lime to the site soil significantly decreases both the volatility of arsenic in the soil, as well as the fractional leachability of the arsenic remaining in the soil ash. Adding alum to the soil significantly increased arsenic volatility, but did not affect resulting ash arsenic leachability. Neither lime nor alum affected lead volatility nor resulting ash lead leachability.

The results from the test program suggest that incineration is a viable treatment technology for remediating the Baird and McGuire site. The muffle furnace results combined with the incineration results suggest that a soil with arsenic content below 150 mg/kg can be incinerated under any combination of kiln temperaturekin excess air level to give an organically decontaminated ash with TCLP leachable arsenic below the limit that would prevent its landfill disposal.

The incineration test results suggest that soil with arsenic levels as high as 1,200 mdkg could be incinerated to give a kiln ash with TCLP leachate concentration less than 5 mg/liter, provided incineration was under low kiln temperature (nominally 840°C [1,54OoFD/high kiln excess air (kiln exit flue gas 0, nominally 11%) conditions. The muffle furnace test results suggest that even higher soil arsenic levels could be incinerated to give a kiln ash with TCLP leachate arsenic concentration of less than 5 mgfliter with lime addition to the soil.

For reference, contaminated soil incineration costs in a transportable rotary kiln system brought to the remediation site are generally in the $100- $300/ton range. However, costs are highly site-specific and often driven by the permitting process. B

RBMEDIATION/SPRING 199 1 237