Chemical amplifiers for detection of peroxy radicals in the atmosphere

Christopher Anastasi, Robert V. Gladstone, and Michael G. Sanderson. Environ. Sci. Technol. , 1993, 27 (3), pp 474–482. DOI: 10.1021/es00040a004...
1 downloads 0 Views 1MB Size
Environ. Sci. Technol. W03, 27, 474-482

Chemical Amplifiers for Detection of Peroxy Radicals in the Atmosphere Christopher Anastasl,* Robert V. Gladstone, and Mlchael 0 . Sanderson Department of Chemistry, University of York, Heslington, York, YO1 5DD, United Kingdom

The chemistry of atmospheric odd-hydrogen radicals (HO,HOz, RO, and R02) has been the focus of much attention in recent years due to their pivotal role in tropospheric air pollution. A chemical amplification system that uses a chain reaction initiated by the reaction of HOz with NO to form NOz and HO radicals has been shown to have potential as a monitor for peroxy radicals in the troposphere. The HO radicals produced react with a chain-carrying agent to re-form peroxy radicals and thus restart the chain. Previous work in this field has used CO as the chain carrier and demonstrated the potential of this approach in the atmosphere. Presented here are the results of a modeling study on a series of alternative chaincarrying agents: ethene, propene, ethanol, and dimethyl ether. The final NOz concentration was determined for each of the chain carriers over the range of HOz concentrations of atmospheric interest [(l-50) X lo7 molecule cme3]. The chain lengths predicted were higher than those obtained for CO, typically by a factor of 2.5. In addition, the sensitivity of the chain length to the initial concentrations of the organic chain carrier and NO was predicted to be small. Interferences from atmospheric pollutants were also assessed.

Introduction Daytime tropospheric chemistry is predominantly driven by the odd-hydrogen radicals (hydroxyl, HO,and perhydroxyl, H02). The HO radical is formed, for example, by the photolysis of ozone and is responsible for initiating the tropospheric oxidation of almost all emitted gases while HOz is the result of the reactions of HO with volatile organic compounds and with CO. The H 0 2 radical is primarily responsible for the oxidation of NO to NO, and so it promotes the production of ozone in the troposphere. It may also have an important role in acid rain chemistry. Numerous attempts have been made to measure atmospheric HO and HOz concentrations, although the problems associated with the sensitivity required to detect the low concentrations of these radicals, for example, lo5 molecule cm-3 for HO and lo8 molecule ~ m for - HOB ~ (I), have been many. The main focus of attention has been HO detection by laser-induced fluorescence (LIF) (2-5) although there is a significant interference from the selfgeneration of HO in these systems (6). A recent development of the LIF method is fluorescence assay with gas expansion (FAGE) (7,8), which uses low pressures to reduce these interferences. The reaction of HOz with NO, which forms HO, has also been used with FAGE to measure HOz concentrations. Other methods that have been attempted as monitors for HO with varying degrees of success, are as follows: chemical labeling by reacting HO with 14C0(9-11),longpath laser absorption (12), high-resolution spectroscopy (13,14), spin trapping/electron spin resonance ( E ) and , resonance fluorescence (16). Methods used to monitor HOz include millimeter spectroscopy (17) and matrix isolation with analysis by electron paramagnetic resonance or gas chromatography/mass spectrometry (28-20). Recently the method of chemical amplification has received attention as a possible alternative for radical monitoring. First developed to measure the sum of atmospheric radicals (HO, RO, H02, and ROz) by Cantrell 474

Envlron. Scl. Technol., Vol. 27, No. 3, 1993

and Stedman (21,22), further refinement has been carried out very recently by Hastie et al. (23).A chain reaction, shown below, involving CO as the chain carrier, and NO, was used to produce NOz (21,22), which was then measured using a luminol detector (24, 25). HO2 + NO HO

4

+ CO

-+

NO2 + HO C02

+H

H + 0 2 (+M) 4HO, (+M) The chain can also be initiated by ROz and RO ROZ + NO RO + NO2 +

(la) (2a) (3)

(lb)

(2b) RO + 02 RCHO + HOz The sensitivity of the monitor, and hence its detection limit, depends primarily on the length of the chain, which, as in all chain reactions, is the result of a balance between the chain propagation reactions (in this case the reaction of HO with the chain-carrying agent) and the chain termination reactions. Cantrell and Stedman originally believed that the reaction of HO with NO HO + NO (+MI HONO (+M) (4)

-

was the dominant radical loss process. The physical loss of radicals to the reactor walls was neglected and chain lengths of -1000 were reported (21, 22). It was later shown by Hastie et al. (23) that wall loss was in fact the most important chain termination process and that chain lengths for the CO system of -120 were more realistic. Techniques that monitor levels of NO2 in the presence of excess NO under ambient conditions will be subject to interferences from the reaction of NO with ozone and also from the thermolysis of peroxyacetyl nitrate, (PAN, CH3C0OzNO2)and pernitric acid (PNA, HOzNO,), all of which yield NO,: 0 3 + NO 02 + NO2 (5)

-

PAN CH3C002 + NO2 PNA + HOz + NO2

(6) (7)

In practice these interferences cause an increase in the background level of NOz and so can be removed by modulating the signal. To achieve this, Cantrell and Stedman (21,22)periodically switched the flow of CO with N,, thus preventing the amplification process and yielding a baseline value. Hastie et al. found that the same effect could be achieved by switching the point at which CO was added (23). However, if the NOz due to the amplification process is only a small part of a large overall concentration of NOz, then the sensitivity of the system is considerably reduced. This aspect of the chemistry was also investigated in this study. The chain length can be improved in four ways: (i) using a carrier that has a faster reaction with HO; (ii) reducing the radical wall loss rate [Ghim (26) found that wall loss was minimized by coating the inside of the reactor vessel with halocarbon wax]; (iii) using chain carriers which undergo further reactions to produce more than one HOz radical per molecule; (iv) minimizing the side reactions of the carrier. In light of these characteristics, several potential organic agents have been investigated in a modeling

0013-938X/93/0827-0474$04.00/0

0 1993 Amerlcan Chemical Soclety

Table I. The CO Model" reaction

rate consta

HO, + NO HO + NO, HO + CO H + CO2

8.1 X 2.4 x 10-13 1.2 x 10-12 5.2 X 7.7 x 10-12 1.9 x 10-15

no. 1. 2.

3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

+

-+

H + 02 HO2 HO t NO HONO CH302+ NO CH30 t NO2 CH3O + 02 4 CHPO + HOp CH3 + 02 CH302 CH3COOz + NO CH3 + COZ + NO2 HO + CHzO H2O + CHO CHO + O2---* H 0 2 + CO HO + O3 H 0 2 + O2 HO + HN03 H20 + NO3 HO + HONO HzO + NO2 HO + HOzNOz HzO + NO2 + 02 HO + NOz HN03 HO + NO3 H 0 2 + NOz HO + HO HzO + 0 HO + HO HzO2 HO + HOz H20 + 0 2 HO + HzO2 HzO + HO2 HO + CH4 Hz0 + CH3 HO + CH3O2 02 + CHBOH HO + CH302N02 H 2 0 + CHzO + NO3 HO + CH30 H20 + CHzO HOP + HOz --* Hz02 + 02 HO2 + NO2 HOzN02 H 0 2 + O3 HO + 2 0 , PAN CH3COO2 + NO2 H02N02 HOp + NO2 CH3C0OZ+ CH3COOz 2CH3 + 2COZ CH3C002+ CH302 CH3 + CH30 + 4

--- -- +

+

-C

4

02

+ coz

32. CH3CO02+ CH302 CH3COOH + CH20 + 02 33. CH302+ NOz CH3O2NO2 34. CH3O2 + CH3O2 2CH30 + O2 35. CH302 + CH3Oz CH,OH + CHzO + 02 36. CH3O2 + CH302 CH3OOCH3 + 0 2 37. CH302N02 CH3O2 + NO2 38. NO + O3 NO2 + O2 39. NO2 O3 NO3 + O2 40. NO3 + NOz N206 41. NzO5 NO3 + NOz 42. HO wall 43. HOz wall 44. CH302 wall 45. CH,COs wall

--- + -- --

4

+

1.0 x 10-12

1.4 X 1.1 x 5.6 X 6.8 X 1.0 x 4.9 x 4.7 x 1.1 x 2.3 X 1.9 x 4.8 X 1.1 x 1.7 X 7.7 x

10-11

1043

10-12 10-12 10-11 lo-" 10-12

10-10 10-15

1.0 x 10-1'

4.0 X

1.0 x 10-11

1.7 X 1.4 X 2.1 x 3.7 x 8.1 X 1.6 X 4.5 x

10-15 10-4*

lo-" 10-13

4.5 x 10-13 4.0 X 8.0 x 2.5 x 3.0 x 1.9* 1.8 x 3.2 x 1.4 X 5.2 X 2.5* 2.5* 2.5* 2.5*

10-14 10-13 10-14 10-14 10-17

"All rate constants are in cm3molecule-' s-l except those marked with an asterisk, which are in 5-l. All values from ref 23.

study and the results are presented here.

Model Simulations Plug-flow modeling was used to investigate the effectiveness of each of the potential chain carriers. This form of modeling assumes a static reaction vessel with no addition or removal of species, apart from loss to the walls. The model is simply started with appropriate concentrations of particular species and then allowed to proceed. Tables I-V show the chemistry used in each of the models. The four models used here build on the CO model used by Hastie et al., which was used as published (23). The additional chemistry for ethene, propene, ethanol, and dimethyl ether (DME) was simply added to the CO model. All schemes include loss of HO and HOzvia wall reactions, and other reactions not directly involved in the amplification process. There are some reactions in Tables I-V that results show make a negligible contribution to the overall mechanism but are included for completeness (e.g.,

the reactions of CHO, because degradation of the parent CH,O via HO is minimal). In the first instance the models GsGmed that the background air stream was "clean"; i.e. it contained only nitrogen and oxygen. In order to directly compare the results of this study with the work of Hastie et al., the same initial conditions were employed, namely, a reaction time of 1.1s and initial concentrations of chain carrier and NO of 1.97 X 10l8 and 1.48 X l O I 4 molecules ~ m - respectively. ~, HOz concentrations were varied between l x lo7 and 50 X lo7 molecules ~ m - ~This . range is believed to cover all concentrations found in the troposphere (1). The CO model is different from the four organic carrier models in several ways. After the initiation (reaction l), HO reacts with CO as follows: HO + CO + H + COZ (24 H

+ 02 (+M)

-

HOz (+M)

(3)

The regenerated HOz is able to start the chain again via reaction la. As the only other product is COz, which is HOz + NO 4NO2 + HO (la) inert, few complications from side reactions occur. However, this system has two major disadvantages, the most important one being the slow rate of reaction between HO and CO. This means that loss of HO to the walls competes with the reaction between HO and CO. Wall losses are significant even at high CO concentrations. Second, only one HOz radical is produced per molecule of CO reacted. The four potential chain carriers used here react more rapidly with HO and produce more than one HOz radical per molecule reacted, giving higher degrees of amplification. However, in some cases the more complicated chemistry causes new problems to arise. Unlike COz,the other products from the reactions are not inert, and so further chemistry can occur. This may involve HO, HOz, NO, and NOz and affect the degree of amplification of the system. For example, following reaction l a one molecule of ethene would ultimately lead to the production of a further six molecules of NOz, assuming that all the secondary chemistry associated with the degradation of the products occurs. In practice however, some of the subsequent reactions do not contribute to the propagation of the chain under the conditions modeled. However, the major disadvantage of ethene as a chain carrier is due to its reaction with ozone (almost certainly present in ambient air), which is thought to produce approximately 12% HOz per molecule of ethene (27). This will initiate new chains which are not derived from the ambient peroxy radicals, and the latter's concentration will be overestimated. Similar arguments can be constructed for propene, and this aspect is discussed in detail later. Ethanol, like ethene and propene, will produce many HOz radicals per molecule. The chemistry of the ethanol system is reasonably well-known and is similar to that for ethene; however, ita reaction with HO, although faster than CO, is significantly slower than that of the alkenes. DME has the advantage of being gaseous at room temperature and not reacting with other ambient species, such as ozone. Although its rate of reaction with HO is slower than that of ethanol (but faster than that of CO), it has the potential to perform well as a chain carrier because the chemistry which follows the attack by HO results in more conversion of NO to NOz. This is because DME and the first stable product, CH,OCHO, do not produce species which react with HO (except for CHzO),whereas the alkenes do, e.g., HOCHzCHO and CH,CHO. Environ. Sci. Technol., Vol. 27, No. 3, 1993 475

Bool 8004

4.0'10'

'

9.5'10'

'

3.0'10'

' '

2 2.5.10'

Ethene Propene

-

,-' ,.,.

I

,.,.',,/;

Ethanol

-DME

,/- 0

-4

.-

,/'P

**

B

0.10'

1.10'

Z*lO'

3'10'

4.10'

,,,.',

-

0

.-

5'10"

[HO,]/cm-'

Figure 3. Final NO2concentrations: [chain carrler] = 1.97 X 10'' molecule cm? [NO] = 1.48 X 1014molecule ~ m - ~ . 2.8

-

2.6-

---~~~----.____________ ----~-.---~-~---__._.~.~...~~..~~~-....--...~~..~.~~~~~~~

------_-_________ 2.4-

2.0-

-

la-

5001

I:

Fl

RCL 2.2-

cL 400

Ethene Pr 0 pen e Ethanol -DYE

I -.,,,

----

I

I

co

-------___-________________________

1.8-

100

I

0'10'

,/'

8004

1.10'

I

2'10'

I

3.10'

I

4.10'

I

5.10'

Ethene

-

Ethanol -DME

oc^T-

0.10~

2.10' 4.10' [Chain Carrier] /molecule cm-'

8

Flgure 4. Chain length response to carrier concentration: [NO] = 1.48 X loi4molecule [HO,] = 1.00 X lo8 molecule ~ m - ~ .

values are presented in Figure 1. The relative chain lengths were between 1.7 for ethanol and 2.7 for propene and DME (Figure 2). This was due to a faster initial reaction with HO and subsequent reactions that produce further HOz radicals. Also, the chain lengths were found to be virtually constant over the range of H 0 2 concentrations investigated, an important feature in an atmospheric monitor. At all the concentrations of H 0 2 investigated the final concentration of NO2produced (Figure 3) is detectable by existing methods such as luminol chemiluminescence (24,W) and laser-induced fluorescence (28)* Simulated Reactor Conditions. With the exception of the CO system, it was found that varying the amount of chain carrier added by a factor of 2 in either direction from the value used by Hastie et al. had little effect on the final concentration of NOz and hence on the chain length. However lowering the amount of chain carrier by more ~, than a factor of 4, to below 4.9 X 1017molecule ~ m -led to a reduction in the chain lengths observed (Figure 4). This relative lack of response of the chain lengths of the organic systems to the amount of chain carrier present is due to the reactor being saturated with the chain carrier, and thus the reaction of the carrier with HO is the only significant removal process for the radical. At lower concentrations of the carrier other processes, including wall loss, compete. In the case of the CO system, the radical loss processes are still playing a role at the higher CO concentrations because the rate of the reaction between CO and HO is relatively slow. The ability to reduce the concentration of the chain carrier has three advantages: (i) the consumption of the chain carrier is reduced; (ii)

Table 11. The Ethene Model no.

reaction

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. lla. 12. 13. 14. 15. 16. 17. 18. Ilb. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

CzH4 + HO HOCHZCHZ HOCHzCHz + 0 2 HOCHzCHzOz HOCH2CHZOz + NO HOCHzCHzO + NO2 HOCHZCHZO CHzO + CHZOH CHzOH + 0 2 CHzO + HOz CHzO + HOz HOCHzOz HOCH20z CHzO + HOz CHZO + HO CHO + HzO CHzO + HO HCOOH + H HOCHzCHzO + 02 HOCHZCHO + HOz HOCHZCHO + HO HOCHZCO + HzO HOCHZCO + 02 HOCHzCOOz HOCHzCOOz + NO2 HOCHzCOOZNOz HOCH2C002NOZ HOCHzCOOz + NO2 HOCHzCOOz + NO CHZOH + COz + NO2 CHZOH + NO HNO + CHZO HNO + 02"O + HOz CH20H + NO2 -+ HONO + CHZO HOCHZCHO + HO HCOCHOH + HZO HCOCHOH + 02 --c HCOCHO + HOz HCOCHO + HO -+ HCOCO + HzO HCOCO CHO + CO CHO + 0 2 CO + HOZ CHO + 0 2 -* COZ + HO CHO + 02 HCOOp CHO + NO ---+ CO + HNO CHO + NOz HONO + CO CHO + NO2 HCOp + NO HCOz + 0 2 HOz + COZ CzH4 + O3 HOz + products CzH4+ O3 products CHO wall -+

-

-

rate constn

-

+

+

- 4

--+

+

-

+

--+

+

9.00 x 3.00 X 9.00 x 1.50 x 9.20 X 7.90 x 1.50 X 9.21 X 1.90 x 8.40 x 8.00 X 5.00 X 8.40 X 5.80 x 1.40 X 2.40 X 3.31 x 8.30 X 2.00 x 1.56 X

ref 30 31 32 33 b 30 30

10-12

lo-''

10-l2 105*

10-14 lo2*

C

34 33,34 30 d

10-13

10-15

e

e e 35 36 35 30

10-4* lo-''

10-14

10-12 lo-" 1.12 x 10-1' 4.80 x 107* 5.60 X 2.00 x 10-14 4.00 x 10-13 1.23 X lo-" 3.53 x 10'" 2.07 X lo-" 5.60 X 2-04 x 10-19 1.50 X 2.5*

f

30,34 g

30, 37, 38 39 39 37,40,41h 42, 43 42, 43 i

27 27 j

"All rate constants are in cm3molecule-' s-l except those marked with an asterisk, which are in 8-l. *The given rate constant is an average 9.3 (30);8.59 (44);9.5, 10.5, 8.6, 8.8 (all ref 45); and 8.61 (46). 'k is an average of the following values of the following values (all X (all X 1O-lZ): 9.0 (34),11.0 (30),and 8.1 (39). dThe rate constant was taken as that for k(CH3CO+Oz)= 5.0 X (30). 'The rate constants and k(CH3COOz+NO)(30). 'The rate constant were assumed to be the same as k(CH3COOz+NOz),k(CH3C00~NOz~CH3C00z+NOz), used in the model was assumed to be the same as k(CH3CHOH+Oz)= 1.56 X lo-" (31, 47). gThe rate constant was estimated from data quoted by Niki et al. (48) and Atkinson et al. (34). hThe rate constant used is an average of 1.26 X lo-" (37), 1.23 X lo-" (41), and 1.20 X lo-" (40). 'The rate constant used is that for the reaction of CHO with 02,as in reaction 24. 'The rate of various wall reactions were taken bv Hastie et al. (23) to be 2.5 8-l; this rate constant has been adorJted here.

secondary reactions of the carrier (e.g., with ozone in the case of the alkenes) are minimized; (iii) for ethanol, the lower concentrations make addition of vapor from a liquid reservoir into the gas stream much easier. As the vapor pressure of ethanol at 298 K is 1.1X 10l8molecule (29) the subsequent modeling concerned with ambient conditions, Le., the interference studies, was performed with initial concentrations equal to half that used for the other amplifiers, i.e., 9.8 X lo1' molecule ~ m - ~ . As the concentration of NO increases, the amount of H 0 2 that is lost to the wall decreases and the amount of HO that is produced rises, through the reaction HO2 + NO NO2 + HO (la) However, one the most important loss processes for HO is the reaction with NO: (4) HO + NO (+M) HONO (+M) This reaction competes with the loss of HO to the reactor walls and the reaction between HO and the chain carrier. Hence there is a balance between increasing the production of HO by reducing the wall loss of HOz and the loss of HO by reaction with NO. Where this balance lies for each of the chain carriers is again dependent on the rate of their reaction with HO. Hence, for CO, the initial concentration of NO used in this work and by Hastie et al. (1.48 X l O I 4 molecule ~ m - is~ ) optimum while for the other chain carriers increasing the

soool

zoooj

,,,'....

...

,*- ,.....' '

_I

+

-

[NO] /molecule cm-'

Figure 5. Chain length response to NO concentration: [chain carrier] = 1.97 X 10'' molecule ~ m - [HO,] ~ ; = 1.00 X 10' molecule ~ m - ~ .

NO concentration resulta in an increase in the chain length. For ethene and propene the increase does not tail off significantly until 4 times the original level, reflecting the high rate of reaction with HO (Figure 5). As indicated earlier, the most important limitation on the chain length is imposed by the loss of the radicals to the walls of the reactor. Figure 6 shows that, in all the systems, reduction of this rate greatly increases the chain length. Environ. Scl. Technol., Vol. 27, No. 3, 1993 477

Table 111. The Propene Model no.

reaction

rate const"

la. 2. 3. 4. 5. 6. 7.

CH3CHCH2 + HO CH3CHCHzOH CH3CHCHZOH + 0 2 --+ CH,CH(O,)CHzOH CH3CH(OZ)CH20H + NO CH,CH(O)CHzOH + NO2 CH3CH(O)CH~OH-CH3CHO+CH2OH CH3CH(O)CH20H + 02 CHBCOCHZOH + HOz CH3COCH2OH + HO CH3COCHOH + H2O CH3COCHOH + 02 CHSCOCHO + HOz CH3CHCHz + HO CH3CH(OH)CHz CH3CH(OH)CHz + 0 2 CH&H(OH)CH202 CHSCH(OH)CH202+ NO CHSCH(0H)CHZO+ NO2 CH3CH(OH)CH20--c CHSCHOH + CHpO CH,CH(OH)CH20 + 0 2 CH&H(OH)CHO + HOz CH&H(OH)CHO + HO CH,C(OH)CHO + H2O CH,C(OH)CHO + 02 CH3COCHO + HOz CH3COCHO + HO CH3CO + CO + H20 CH3C0 + HO CH2C0 + H 2 0 CHzCO + HO CH2O + CHO CHzCO + HO CHzOH + CO CH,CH(OH)CHO + HO -+ CH,CH(OH)CO + HzO CH,CH(OH)CO + O2 CH3CH(OH)C002 CH,CH(OH)COOZ + NO2 CH3CH(OH)COOZN02 CH3CH(OH)C002N02 CH3CH(OH)C002 + NO2 CH,CH(OH)C002 + NO -+ CH3CHOH + COZ + NO2 CH3CHOH + O2 CH3CH0 + HOz CH3CHO + HO CH3CO + H2O CH3CO + 0 2 CHZCOOz CH20H + O2 CH20 + H 0 2 CHzO + HOz HOCHZOZ HOCH20z CHzO + H02 CHZO + HO CHO + H20 CHzO + HO HCOOH + H CHzOH + NO HNO + CHzO HNO + 0 2 NO + HOz CHzOH + NO2 HONO + CHzO CHO + 02 CO + HOz CHO + 02 COZ + HO CHO + 02 HCOO2 CHO + NO -+ CO + HNO CHO + NO2 HONO + CO CHO + NO2 HCOz + NO HCOz + 02 HOz + CO2 CH3CHCH2+ O3 H02 + products CH3CHCHz+ O3 products CHO wall

1.95 x lo-" 1.16 x 10-13 8.16 X 6.00 X lo6* 3.00 x 10-14 3.00 x 10-l2 1.56 X lo-" 1.05 X lo-" 3.82 x 10-l2 8.40 X 6.00 x 105 6.90 x 10-15 2.00 x 10-12 1.56 X lo-" 1.70 X lo-" 2.00 x 10-1' 9.00 x 10-12 9.00 x 10-12 8.00 x 10-12 3.82 X 8.40 X 5.80 x 10-4* 1.40 X lo-" 1.56 X 1.60 X lo-" 5.00 X 9.20 X 7.90 x 1 0 4 4 1.50 X lo2* 9.21 X 1.90 x 10-13 2.40 X lo-" 3.31 x 10-14 8.30 X 5.60 X 2.00 x 1 0 4 4 4.00 x 10-13 1.23 X lo-" 3.53 x 10-11 2.07 X lo-" 5.60 X 2.82 X 8.48 X 2.50*

+

-+

-

1b. 7. 8. 9.

+

10. 11. 12.

-

+

-t

-+

--

13. 14. 15. 16.

+

-

17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

--

-

-

4

-+

+

-+

+

--

-

--

ref

30,34 49 50, 51b 51 51 52 C

30, 34 49 50, 51d 51 33' 36 C

30 53 34 34 308 h 30' 30' 30' 31,47 30 30, 54

1

30

30 k 34 35 36 35 30, 37, 38 39 39 37, 40, 41' 42,43 42,43 m 27 27 h

All rate constants are in cm3 molecule-' s-l except those marked with an asterisk, which are in s-l. * k(CH3CH(02)CH20H+NO)was Allowing for 4% nitrate formation, k = 8.16 X assumed to be equal to k(i-C3H70z+NO)(30) = k(sec-C,H90z+NO) (50) = 8.5 X "The rate constant used is k(CH3CHOH+Oz). k(CH3CH(OH)CH202+NO)was assumed to equal k(n-C3H702+NO)(30) = k(n-C4HS02+NO) (50) = 8.7 X lo-". Allowing for 4% nitrate formation, k = 8.4 X "he rate constant is for k(HOCH2CH20+Oz)(33). fThe rate constant is for k(HOCH2CHO+HO-HOCHCHO+HzO) (30). #The rate constant is for k(HOCH2CHO+HO-HOCH2CO+H20) (30). The rate constant is for k(CH3CH(OH)CH2+02)(49). 'See footnote f , Table 11. 'See footnote c, Table 11. kSee footnote d, Table 11. 'See footnote h. Table 11. See footnote i, Table 11. See footnote j , Table 11.

Interferences. Ambient ozone in the air stream will react with the excess NO to form NO2: O3+ NO O2+ NO2 (5) This increase of the background NO2signal makes it more difficult to discriminate between the NO2produced by the chemical amplification process and the background. Consequently the sensitivity to atmospheric radicals is reduced. A higher chain length, and so a higher final NO2 concentration, improves the discrimination possible. As indicated earlier, the alkenes also suffer from attack by ozone, which can form H02 radicals (27)and further reduce the discrimination power. For example, in the case of ethene the reaction can be represented by CzH4+ O3 0.12H02 + other products Although the reaction is slow and the branching ratio is only 0.12, it has a significant effect as the H02 radicals produced form new chains independent of the concentra-

-

-

478

Environ. Sci. Technol., Voi. 27, No. 3, 1993

tions of atmospheric peroxy radicals. In order to compare the performances of the various systems in discriminating against the NO2produced by ozone, a discriminating power (DP) of each system is defined as where [N0210is the final NO2in the absence of ozone (and therefore produced by chemical amplification) and [N021totalis the final NO2 in the presence of ozone. The DP,,,, of each of the systems over a range of ambient ozone concentrations [(6-25) X lo1' molecule ~ m - ~ ] are shown in Figure 7. The DP,,,, of the ethene system was improved by reducing the ethene concentration and increasing the concentration of NO, as the ozone is forced to react with the NO rather than the ethene, and the extra NO also improves the chain length. The two ethanol lines show the effect of increasing the NO concentration alone. It should be noted that the ethanol b results do not rep-

Table IV. The Ethanol Model

reaction

la. lb.

CH3CHZOH + HO CH3CHOH + HzO CH3CH20H + HO HOCHZCH, + Hz0 CHSCH2OH + HO CH3CHzO + HzO HOCHzCHz + 02 HOCHzCHzOz HOCHzCHzOz + NO -+ HOCHZCHZO + NO2 HOCHzCHzO CHzO + CHzOH CHZOH + 02 CHzO + HOp CHzO + HOz HOCH2Oz HOCHzOz -+ CHzO + Hop CHzO + HO CHO + H20 CHzO + HO HCOOH + H HOCHzCHzO + 02 -+ HOCHZCHO + HOp HOCH&HO + HO HOCHZCO + HzO HOCHZCO + 02 HOCHzCOOz HOCHZCOOZ + NO2 HOCHzCOOzNOz HOCHzCOOzNOz HOCHZCOOZ + NO2 HOCHzCOOz + NO CHZOH + COz + NO2 CHzOH + NO CHzO + HNO HNO + 02 NO + HOz CHpOH + NO2 HONO + CHzO HOCHZCHO + HO HOCHCHO + HzO HOCHCHO + 02 HCOCHO + HOz HCOCHO + HO HCOCO + HzO HCOCO CHO + CO CHO + 0 2 CO + HOz CHO + 02 COZ + HO CHO + 02 HCOOz CHO + NO CO + HNO CHO + NOz HONO + CO CHO + NO2 -.c HC02 + NO CHSCHOH + 02 CH3CHO + HOz CH3CH20 + 02 -+ CHBCHO + HOz CHBCHO + HO CHBCO + HzO CH3CO + 02 CH3C002 CH3CO + HO CHzCO + HzO CHzCO + HO CHzO + CHO CHzCO + HO CHzOH + CO HCOz + 02 HOz + COZ CHO wall

-4

IC.

2. 3. 4. 5. 6. 7. 8a. 8b. 9. loa. 11. 12. 13. 14. 15. 16. 17. lob. 18. 19. 20. 21a. 21b. 21c. 22. 23a. 23b. 24. 25. 26. 27. 28. 29a. 29b. 30. 31.

-

4

+

+

+

-C

--- --

-C

+

-+

+

+

+

-

+

+

-

-C

ref

rate const"

no.

*

2.67 X 3.90 x 3.40 x 3.00 X 9.00 x 1.50 x 9.20 X 7.90 x 1.50 X 9.21 X 1.90 x 8.40 x 8.00 X 5.00 X 8.40 X 5.80 x 1.40 X 2.40 X 3.31 x 8.30 X 2.00 x 1.56 X 1.12 x 4.80 X 5.60 X 2.00 x 4.00 x 1.23 X 3.53 x 2.07 X 1.56 X 8.00 x 1.60 X 5.00 X 2.00 x 9.00 x 9.00 x 5.60 X 2.50*

lo-'' 10-13 1043

lo-"

10-12 105*

30, 55b b b 31 32 33 C

10-14 lo2*

30 30 d

1043

10-15

34 33,34 30 e

10-4* lo-" lo-" 10-14 10-12

lo-"

10-11 lo7* 10-14 1043

10-l' 10-11 lo-" 10-15 lo-"

10-11 10-12 10-12

f f f

35 36 35 30

g

30, 34

h

30, 37, 38 39 39 37, 40, 41' 42, 43 42,43 31,47 30 30 30, 54 53 34j 34

k 1

"All rate constants are in cm3 molecule-I s-l except those marked with an asterisk, which are in s-l. "he overall rate constant kl is 3.4 X (30). The only known value for k,,/k, is 0.75 f 0.15 (55). klb was estimated as '/,k(HO+t-C4H90H) (56), assuming alcoholic H abstraction is negligible. kl, was estimated at 10% of kl. CSeefootnote b, Table 11. dSee footnote c, Table 11. eSee footnote d, Table 11. (See footnote e, Table 11. gSee footnote f , Table 11. hSee footnote g, Table 11. 'See footnote h, Table 11. 'Overall rate constant for HO + CHzCO is 1.8 X lo-"; the branching ratio was taken to be 1:l as the actual value is not known. See footnote i, Table 11. See footnote j , Table 11. 14003

0

I

I

I

I

1

2

3

4

kv.u /aw'

Flgure 6. Variation of wall loss rate: [chain carrier molecule ~ m - [NO] ~ ; = 1.48 X 10'' molecule cmX lo8 molecule ~ r n - ~ .

1 5

a:=

1.97 X 10" [HO,] = 1.00

resent the optimum conditions for ethanol discrimination, it merely demonstrates the potential for improved performance that is readily available. Despite the high chain length of the propene system it suffered badly from a fast

attack by ozone, with a branching ratio for HOzproduction twice that for ethene. In addition to HOz, several other radicals, such as CH30z,CH3CH02,and CHZOz,cue formed that can cause further interference (27). For these reasons no further simulation work was carried out on the propene system. The thermolyses of PAN and PNA contribute to the background NO2 in two ways. In both cases the products are NO2, which adds to the background directly, and a peroxy radical, CH3C002,from PAN and HOz from PNA. The simulations show that only 0.04% of the PAN would break down in the 1.1 s it is in the chamber, as would be predicted from its relatively long lifetime at 298 K of 45 min (23). As a result, the contribution of the NO2 produced directly is almost negligible compared to other background sources, despite PAN concentrations being typically 2000 times greater than HOz concentrations. Where significant interferences do arise, however, is from the reactions of the CH3C0OZradical. This rapidly undergoes a series of reactions that result in the formation of a further two molecules of NOz and a HOz radical. It is this radical that is the critical aspect of the interference as it can start a new amplification chain, making the three Environ. Scl. Technol., Vol. 27, No. 3, 1993 479

Table V. The DME Model no. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 19. 21. 22. 23. 24. 25.

reaction

rate consta

ref

CH,0CH3 + HO CH30CHz+ H,O CH30CHz + 0 2 CH30CHz02 CH30CHzOz + NO CHaOCHZO + NO2 CHzOCH20 + 02 CH3OCHO + HOZ CH,OCHO + HO CHzOCHO + HzO CHZOCHO + 02 -+ HCOOCHzOz HCOOCHzOz+ NO HCOOCHzO+ NOz HCOOCHZO --+ HCO2 + CHzO HCOZ + 02 -+ HOz + COZ CHzO + HOz HOCH20z HOCHzOz -+ CHzO + HOp CH20 + HO CHO + H 2 0 CH2O + HO HCOOH + H CHzOH + NO HNO + CHzO HNO + O2 NO + HOz CH20H + NOz HONO + CHzO CHO + O2 CO + H 0 2 CHO + 02 COP + HO CHO + 02 HCOOz CHO + NO CO + HNO CHO + NOz HONO + CO CHO + NOz HCOz + NO CHO wall

2.39 x 10-12 3.00 X 7.23 x 10-12 8-40 x 10-15 2.27 x 10-13 4.50 x 7.23 x 10-l2 1-50 x 105* 5.60 x 10-l2 7.90 x 10-14 1.50 X lo2* 9.21 x 10-12 1.90 x 10-13 2.40 x lo-" 3.31 x 10-14 8.30 x 5.60 X 2.00 x 10-14 4.00 X 1.23 X lo-" 3.53 x 10-11 2.07 x 2.5*

57, 58, 5gb 31' 32, 60d 27, 34e 61f 30, 318 32, 60d 27, 34h 30' 30 30 j 34 35 36 35 30, 37, 38 39 39 37, 40, 41k 42,43 42,43

---- --+

+

4

-+

1

All rate constants are in cm3 molecule-' s-l except those marked with an asterisk, which are in s-l. * K is an average of the following values (all X 2.32 (58),2.49 (57), and 2.35 (59). "he rate constant used is k(HOCHzCHz+02),because HOCH2CHzand CH30CH2are isomeric. dThe rate constant is an average of k(HOCHz02+NO)= 5.60 X (60) and k(HOCH2CH,0z+NO) = 9.0 X 10-l2 (32) = 7.3 X 10'l2; allowing for 2% nitrate formation gives 7.23 X lo-''. "he rate constant used is k(HOCH2CHz0+O2)(27,34). /There is an alternative product from this reaction, CH30C0. However, a stu& of similar reactions [HO with HCOOH (621,CH&OOH (63),and CH3COOCH3(64)] shows that the attack predominantly occurs at the alkoxy end and so this second channel is unimportant. gThe rate constant used is an (31) and k(CH3CH2CH2+O2)= 6.00 X average of k(HOCHzCHz+02)= 3.00 X (30). h k is equal to k(HOCHzCHzO~CHzOH+CH20)(27, 34). See footnote g , Table 11. See footnote c, Table 11. See footnote h, Table 11. See footnote j , Table 11. J

100,

o / O.O'lO'

Flgure 7. Ozone discrlmlnation (concentrations in molecule ern-?: [CO] = 1.97X lola,[NO] = 1.48X [ethenea] = 1.97X loi8, [NO] = 1.48X loq4,[ethene b] = 0.15 X loi8,[NO] = 2.95X loi4, [propene] = 1.97X loT8,INO] = 1.48 X [ethanol a] = 9.87 X 1017,[NO] = 1.48X 10' , [ethanol b] = 9.87X loi7,[NO] = 2.95 [DME] = 1.97 X 1Ol8, [NO] = 1.48 X loi4. I n all cases X [HO,] = 1.00 X 10' molecule ~ m - ~ .

molecules of NOz produced more directly from the thermolysis effectively insignificant. Even at relatively low PAN concentrations (1X 1O1O molecule ~ r n - the ~ ) overall contributions to the final NOz concentration is significant (Figure 8). Within experimental error, the DPPm of each system is affected equally, as the additional HOPis released at the same rate in each case. The independence of the thermolysis rate to the system is confirmed by following the amount of NOz produced by the thermolysis. In the PNA interference simulations, the NOz produced directly by the thermolysis is again virtually negligible, although this is due to the low levels of PNA present rather than its lifetime. It was found that approximately 8.5% of the PNA decomposed during the 1.1-sreaction time, but 480

Environ. Sci. Technol., Vol. 27, No. 3, 1993

,

, 5.0'10'

,

'

I

l.O'lO1

I

*

,

,

I

1.5'10' e 2.0'10' [PAN] /molecule cm->

,

,

2.5'10'

Figure 8. PAN discrimination: [chain carrier] = 1.97X 10'' molecule ~ m(except - ~ [ethanoq = 9.87 X 1017molecule ~ m - ~[NO] ) ; = 1.48 X lOI4 molecule cm- ; [HO,] = 1.00 X 10' molecule ~ r n - ~ .

as the PNA concentrations investigated were a factor of 100 lower than those for PAN, approximately double the amount of PNA undergoes thermolysis, again producing a HOzradical that can start a new chain. The overall result is that PNA concentrations in the region of (1-25) X lo9 molecule cm-3 produce a DP equivalent to PAN levels 100 times greater (Figure 9). At present, typical ambient concentrations of PNA are not known, though the short lifetime suggests that the concentration is lower than that of PAN. On the other hand, as the PNA precursor, the perhydroxyl radical, is present in higher levels than the peroxyacetyl radical precursor for PAN, it is also possible that the steady-state concentration of PNA is comparable with that of PAN. Although the chemical amplification technique measures the sum of all atmospheric odd-hydrogen radicals, simulations show that the contribution to the total from ambient HO concentrations is so small as to be well within

100,

1 0.0'10 o

V

-co Ethene -- Ethanol -

-DUE

I O [PNA] /molecule cm-'

1

1

m

Flgur 0, PNA dlsalmlnatlon: [chain carrier] = 1.97 X 10'' molecule ) ; = 1.48 ~ m (except - ~ [ethanog = 9.87 X 10'' molecule ~ m - ~[NO\ X loi4 molecule cm- ; [HO,] = 1.00 X 10' molecule cm-

.

experimental error. This is because HO concentrations are between a factor of 100 and 1000 less than HOz concentrations. Ambient HO concentrationstherefore cannot reliably be resolved with this method.

Summary All the systems investigated appear to perform better than CO as chain-carrying agents. However, they all have weaknesses as well as strengths; the alkenes have a fast initial HO reaction and at least one further conversion of NO to NOz but suffer from the attack by ozone. Alcohols have no important interferences but a relatively low chain length; also, there is a practical problem of ethanol vapor delivery into the reactor from a liquid reservoir. Although the rate of reaction of ethers with HO is slower than that of ethanol, they suffer little from ambient interferences; it appears that ethers could combine the best features of alkenes and alcohols without the problems discussed above. Literature Cited (1) Kanakidou, M.; Singh, H. B.; Valentin, K. M.; Crutzen, P. J. J. Geophys. Res. 1991, 96, 15395-15413. (2) Davis, D. D.; Heaps, W.; McGee, T. Geophys. Res. Lett. 1976,3, 331-333. (3) Wang, C. C.; Davis, L. I., Jr.; Selzer, D. M. J.Geophys.Res. 1981,86, 1181-1186. (4) Hoell, J. M., Ed. NASA Conf. Publ. 1984, No. 2332. (5) Bradshaw, J. D.; Rodgers, M. 0.;Davis, D. D. Appl. Opt. 1984,23, 2134-2145. (6) Davis, D. D.; Rodgers, M. 0.; Fischer, S. D.; Asai, K. Geophys. Res. Lett. 1981,8, 69-72. (7) Hard, T. M.; O'Brien, R. J.; Chan, C. Y.; Mehrabzadeh, A. A. Environ. Sci. Technol. 1984, 18, 768-777. (8) Hard, T. M.; Chan, C. Y.; Mehrabzadeh, A. A,; Pan, W. H.; O'Brien, R. J. Nature 1986, 322, 617-620. (9) Campbell, M. J.; Sheppard, J. C.; Au, B. F. Geophys.Res. Lett. 1979, 6, 175-178. (10) Campbell, M. J.; Farmer, J. C.; Fitzner, C. A.; Henry, M. N.; Sheppard, J. C.; Hardy, R. J.; Hooper, J. F.; Muralidhar, V. J. J. Atmos. Chem. 1986,4, 413-427. (11) Felton, C. C.; Sheppard, J. C.; Campbell, M. J. Nature 1988, 335,53-55. (12) Hubler, G.; Perner, D.; Platt, U.; Tonnissen, A.; Ehhalt, D. H. J . Geophys. Res. 1984,89, 1309-1319. (13) Burnett, C. R.; Burnett, E. B. J. Geophys. Res. 1981,86, 5185-5202. (14) Burnett, C. R.; Burnett, E. B. Geophys. Res. Lett. 1982, 9, 708-711. (15) Watanabe, T.; Yoshida, M.; Fujiwara, S.; Abe, K.; Onoe, A.; Horota, M.; Igarashi, S. Anal. Chem. 1982,54,2470-2474. (16) Anderson, J. G. Geophys. Res. Lett. 1976,3, 165-168. (17) De Zafra, R. L.; Parrish, A.; Soloman, P. M.; Barret, J. W. J . Geophys. Res. 1984,89, 1321-1326.

(18) Mihelcic, D.; Ehhalt, D. H.; Kulessa, K. F.; Klomfass, J.; Trainer, M.; Schmidt, U.; Rohrs, H. Pure. Appl. Geophys. 1978,116, 530-542. (19) Mihelcic, D.; Musgen, P.; Ehhalt, D. H. J. Atmos. Chem. 1985,3, 341-361. (20) Mihelcic, D.; Volz-Thomas, A.; Patz, H. W.; Kley, Do; Mihelcic, M. J. Atmos. Chem. 1990, 11, 271-297. (21) Cantrell, C. A.; Stedman, D. H. Geophys. Res. Lett. 1982, 9,846-849. (22) Cantrell, C. A.; Stedman, D. H. Wendel, G. J. Anal. Chem. 1984,56, 1496-1502. (23) Hastie, D. R.; Weissenmayer, M.; Burrows, J. P.; Harris, G. W. Anal. Chem. 1991,63, 2048-2057. (24) Maeda, Y.; Aoki, K.; Munemori, M. Anal. Chem. 1980,52, 307-331. (25) Wendel, G. J.; Stedman, D. H.; Cantrell, C. A.; Demrauer, L. Anal. Chem. 1983,55,937-940. (26) Ghim, B. T. MSc. Thesis, Faculty of Natural Sciences, University of Denver, 1988. (27) Atkinson, R. Atmos. Environ. 1990, 24A, 1-41. (28) Fincher, C. L.; Tucker, A. W.; Birnbaum, M. Proc. SOC. Photo.-Opt. Instrum. Eng. 1978, 158, 137-140. (29) CRC Handbook of Chemistry and Physics, 63rd ed.; Weast, R. C., Ed.; CRC Press, Inc.: Boca Raton, FL, 1982. (30) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr, J. A.; Troe, J. J. Phys. Chem. Ref. Data 1989,18, 881-1097. (31) Miyoshi, A.; Matsui, H.; Washida, N. Chem. Phys. Lett. 1989,160, 291-294. (32) Becker, K. H.; Geiger, H.; Weissen, P. Chem. Phys. Lett. 1991, 184, 256-261. (33) Anastasi, C.; Muir, D. J.; Simpson, V. J.; Pagsberg, P. J. Phys. Chem. 1991,95,5791-5797. (34) Atkinson, R. Chem. Rev. 1986,86, 69-201. (35) Nesbitt, F. L.; Payne, W. A,; Steif,L. J. J.Phys. Chem. 1989, 93,5158-5161. (36) Fujii, N.; Miyama, H.; Koshi, M.; Asoda, T. Symp. Int. Combust. Proc. 1981,18, 873. (37) Langford, A. 0.;Moore, C. B. J. Chem. Phys. 1984, 80, 4211-4221. (38) Osif, T. L.; Heicklen, J. J.Phys. Chem. 1976,80,1526-1531. (39) Temps, F.; Wagner, H. G. Ber. Bunsenges Phys. Chem. 1984,88, 410-414. (40) Nadtochenko, V. A.; Sarkisov, 0. M.; Suividenkov, W. A.; Ceskis, J. Kinet. Catal. 1980,21, 520. (41) Veyret, B.; Lesclaux, R. J.Phys. Chem. 1981,85,1918-1922. (42) Canosa, C.; Penzhorn, R.-D.; von Sonntag, C. Ber. Bunsenges Phys. Chem. 1979,83, 217. (43) Timonen, R. S.; Ratajazak, E.; Gutman, D. J.Phys. Chem. 1988, 92, 651-655. (44) Wayne, W. A. Int. J. Chem. Kinet. 1988,20,63. (45) Pagsberg, P.; Monk, J.; Anastasi, C.; Simpson, V. J. J. Phys. Chem. 1989,93,5162-5165. (46) Nesbitt, F. L.; Payne, W. A.; Steif,L. J. J.Phys. Chem. 1988, 92, 4030-4032. (47) Anastasi, C.; Simpson, V. J.; Monk, J.; Pagsberg, P. Chem. Phys. Lett. 1989, 164, 18-22. (48) Niki, H.; Mater, P. D.; Savage, C. M.; Breitenvach, L. P. Int. J. Chem. Kinet. 1985,17, 547-558. (49) Miyoshi, A.; Matsui, H.; Washida, N. J.Phys. Chem. 1990, 94, 3016-3019. (50) Atkinson, R.; Aschmann, S. M.; Winer, A. M. J. Atmos. Chem. 1987,5,91. (51) Hatakeyama, S.; Akimoto, H.; Washida, N. Enuiron. Sci. Technol. 1991,25,1884, (52) Dagaut, P.; Liu, R.; Wdington, T. J.; Kurylo, M. J. J.Phys. Chem. 1989,93, 7838-7840. (53) Hatakeyama, S.; Honda, S.; Washida, N.; Akimoto, H. Bull. Chem. SOC.Jpn. 1985,58,2157-2162. (54) Miyoshi, A.; Matsui, H.; Washida, N. J.Phys. Chem. 1989, 93, 5813-5818. (55) Meier, U.; Grotheer, H. H.; Riekert, G.; Just, T. Ber. Bunsenges Phys. Chem. 1985,89, 325-327. (56) Wallington, T. J.; Dagaut, P.; Liu, R.; Kurylo, M. J. Enuiron. Sci. Technol. 1988,22, 842-844. Envlron. Sci. Technol., Vol. 27, No. 3, 1993 481

Environ. Sci. Technol. 7093, 27, 482-488

(57) Wallington, T. J.; Liu, R.; Dagaut, P.; Kurylo, M. J. Znt. J. Chem. Kinet. 1988,20, 41. (58) Wallington, T. J.; Adino, J. M.; Skewes, L.M.; Siegel, W. 0.;Japar, S. M. Znt. J. Chem. Kinet. 1989,21,993-1001. (59) Nelson, L.; Rattigan, 0.; Neavyn, R.; Sidebottom, H. Int. J. Chem. Kinet. 1990,22, 1111. (60) Veyret, B.; Rayez, J.-C.; Lesclaux, R. J. Phys. Chem. 1982, 86,3424-3430. (61) Wallington, T. J.; Dagaut, P.; Liu, R.; Kurylo, M. J. Znt. J. Chem. Kinet. 1988,20, 177.

(62) Singleton, D. L.; Parastevopoulos, G.; Irwin, R. S.; Jolly, G. S.; McKenny, D. L. J . Am. Chem. Soc. 1988, 110, 7786-7790. (63) Singleton, D. L.; Parastevopoulos, G.; Irwin, R. S. J . Am. Chem. SOC.1989, 111 , 5248-5251. (64) Campbell, I. M.; Parkinson, P. E. Chem. Phys. Lett. 1978, 53,385-387.

Received for review May 18,1992. Revised manuscript received August 14, 1992. Accepted October 29, 1992.

Reverse-Burn Gasification for Treatment of Hazardous Wastes: Contaminated Soil, Mixed Wastes, and Spent Activated Carbon Regeneration Laura L. Klnner

Entropy Environmentalists, Inc., P.O. Box

129 1,

Research Triangle Park, North Carolina 27709-229 1

Audrey McGowln ABC Laboratories, 7200 ABC Lane, P.O. Box 1097, Columbia, Missouri 65205

Stanley E. Manahan*

Department of Chemistry, Unlversity of Missouri-Columbia,

Columbia, Missouri 652 1 1

Davld W. Larsen

Department of Chemistry, University of Missourl-St.

Louis,

8001

A unique reverse-burn gasification process (the ChemChar process) employing secondary combustion of the product gases is described. The process has been applied to a variety of hazardous and nonhazardous wastes. Thermochemical destruction of these wastes is accomplished in a primarily reducing atmosphere induced by reactions of carbon, oxygen, and water. This paper describes the results from selected studies involving application of reverse-burn gasification to soil contaminated with polychlorinated biphenyls (PCBs),activated carbon regeneration, and mixed waste treatment. The process is especially useful in destroying hazardous wastes because of the unique characteristics of reverse-burn gasification, which is particularly effective for dehydrohalogenating organohalide compounds without producing undesirable byproducts, such as dioxins, in retaining acid gases, such as hydrogen chloride produced in the destruction of organohalides, and in retaining radionuclides during the destruction of the organic constituents of mixed waste. With second-stage combustion of the product gas, destruction/removal efficiency of greater than 99.9999% (“six nines”) is readily achieved.

Introduction Reverse-burn gasification (Figure l),patented for waste treatment as the ChemChar Process (1, 2), provides a means to treat wastes thermochemically in the forms of solids, liquids, sludges, and soils on a devolatilized coal char matrix (3). Organic constituents of the waste are converted to a combustible gas and to a dry, inert, carbonaceous solid. The solid can be readily mixed with cement to prevent leaching of radioactive or heavy metal constituents retained in the char residue after gasification, or the solid can be converted to a much reduced volume of poorly leachable slag by a subsequent forward-burn gasification (4). Therefore, reverse-burn gasification can be a very effective method for treating mixed wastes (organic waste containing radioactive metals) (5). 482

Envlron. Scl. Technol., Vol. 27, No. 3, 1993

Natural Bridge Road, St. Louis, Missourl 63121 Spent activated carbon can be regenerated by reverseburn gasification. A previous study (6)had indicated that, in addition to destroying sorbed hazardous constituents, partial gasification of spent activated carbon by the reverse-burn process restores the sorptive capacity of the carbon, retains heavy metals on the carbon, and (for alkaline carbons) retains HCl on the carbon matrix. Figure 1shows a diagram of the batch-mode version of the reactor used for reverse-burn gasification (a continuous-feed version is under development). For gasification the reactor is charged with the coal char/waste mixture. Water, which aids gasification and provides a source of hydrogen for waste-destroying free radical induced reactions, such as dehydrohalogenation,may be present on the solid or introduced as steam into the oxidant stream. Oxygen was used as the oxidant in all experiments described in this paper. As reverse-burn gasification occurs with movement of the flame front counter to the oxidant flow, a combustible synthesis gas is evolved from the gasifier. This gas consists of combustible components, COz, H20 vapor, and trace volatile organic constituents. The combustible fraction is approximately 45% CO, 45% Hz, and 10% CH4. The carbonaceous solid residue retains heavy metals and, when alkaline, acid gases, such as HC1. If the flame front is not extinguished by stopping the oxidant flow when the front reaches the top of the solids reactor charge, or when the front is initiated at the top of the reactor charge, forward-burn gasification occurs as the flame front travels in the same direction as the oxidant flow, consuming all combustible material and leaving only the inorganic constituents of the reactor charge. The characteristics of the flame front and the hot reducing zone immediately downstream from it determine the characteristics of reverse-burn gasification that are crucial to its success as a waste treatment process. Figure 2 further explains the thermochemical reactions occurring in the region just above the flame front, within the flame front, and in the reducing zone immediately

0013-938X193/0927-0482$04.0010

0 1993 American Chemlcal Soclety