Environ. Sci. Technol. l994,28, 2191-2196
?-Ray Destruction of Individual PCB Congeners in Neutral %Propanol+ Rod E. Arbon' and Bruce J. Mincher
Idaho National Engineering Laboratory, P.O. Box
1625,
Idaho Falls, Idaho 8341 5-4107
Walter B. Knlghton
Montana State University, Bozeman, Montana
597 17-0340
The radioiytic degradation of 25 PCB congeners from nine homologs was investigated in neutral aerated 2-propanol using spent nuclear fuel as the y-ray source. Radiolytic degradation is conveniently described in terms of a pseudofirst-order rate constant expressed in terms of dose, rather than time, and referred to as a dose constant. Dose constants obtained were found to be dependent on both the number and the positions of the chlorines. A general increase in the dose constant was observed with increasing chlorine number. Chlorine substitution in the para and meta position increases the dose constant while substitution in the ortho position decreases it. Dose constant trends follow the energy level of the lowest unoccupied molecular orbital. Previously unreported degradation products consisting of 2-propanol-substituted polychlorinated biphenyls are characterized. Mass balance information is incomplete for some congeners. Based upon experiments using a 14C-labeled tetrachlorobiphenyl, degradation products are in nonvolatile or semivolatile constituents.
Introduction The widespread environmental distribution of polychlorinated biphenyls (PCBs) has been well documented (1-3). The persistence of PCBs in the environment, coupled with their apparent toxicity ( I ) , led the United States to pass the Toxic Substances Control Act (TSCA) in 1976. The current TSCA approved methods for destroying PCBs are high-temperature incineration or high-efficiency boilers (1). Destruction of PCBs by incineration, however, is meeting with increasing public opposition. The public's fear of incomplete incineration and the possible formation of highly toxic dioxins and dibenzofurans, if the combustion temperature is not held sufficiently high, has significantly reduced the general acceptance of this technique. Many PCB-containing solvents are valuable products in the absence of PCB contamination. What is needed is a process that would degrade the PCBs and allow the solvent to be recycled rather than destroyed. Towards this end, radiolysis offers many attractive features such as (a) minimization of gaseous and particulate effluents, (b) potential of recovering the bulk solvent for recycling, (c) ability to verify that the hazardous constituents have been reduced to acceptable limits, and (d) possibility of i n situ destruction in selected applications due to the highly penetrating nature of y-rays. As a result, a number of investigations have explored radiolysis as a means for dechlorinating PCSs and other organochlorine
materials (4-15). The mechanism by which the dechlorination occurs was found to be dependent on the nature of the solvent system employed. For instance, in alkaline 2-propanol, solutions of PCBs (7, 8, 12) and other halogenated material (10, 11)undergo stepwise radiolytic dehalogenation via a chain reaction involving the acetone radical anion. The resulting products are a dehalogenated substrate and free halide. In neutral 2-propanol, solutions of CC14(13),DDT (14),and hexachloroethane (15)undergo efficient radiolytic dechlorination via a chain reaction involving the a-hydroxyisopropyl radical. Whereas, for PCBs, the radiolytic dechlorination in neutral 2-propanol has been observed to be less efficient, with conflicting reports as to the identity of the reactive species responsible for initiating the dechlorination. Possibilities include the acetone anion, the a-hydroxyisopropyl radical, hydrogen atoms, and the solvated electron (16). It is generally agreed that solvated electrons are the primary species responsible for the dechlorination of PCBs in neutral 2-propanol (57) with some possible contribution from the a-hydroxyisopropyl radical (5). To our knowledge there has never been a systematic study which relates radiolytic degradation efficiency to the number and position of the chlorine atoms. Any variation in the radiolytic destruction efficiency as a function of chlorine position and chlorine number may also have important implications in terms of a treatment process. Problems associated with identifying variations in degradation efficiency are likely to be compounded using Aroclor mixtures, which consist of a large number of different congeners. Therefore, we initiated a systematic study of the radiolytic degradation of a variety of individual PCB congenersthat are representative of Aroclor mixtures. In this paper, the observed radiolytic degradation efficiency for individual PCB congeners will be related to the number of chlorines and the chlorine substitution pattern. Of equal importance in a treatment process is the identification of degradation products. Mass balance information is provided that demonstrates that the majority of the degradation products have been identified, but there still may remain an unidentified fraction.
Experimental Section
Work performed under contract to the US. Department of Energy Office of Technology Development, Field Office, Idaho, under Contract DE-AC07-76ID01570. * Author to whom correspondence should be addressed.
Irradiations. Samples were y-ray irradiated using spent nuclear fuel at the Advanced Test Reactor (ATR) at the Idaho National Engineering Laboratory. Spent nuclear fuel is an excellent source of y-rays with an average kinetic energy of 700 keV. Depleted fuel elements from the reactor are stored vertically in grids in an adjacent canal under approximately 6.5 m of water. A dry tube extends from the surface of the water into the grid, and the fuel elements are placed around it. Proper positioning of the fuel elements around the dry tube allows for the selection of dose rates up to 25 kGy h-I. PCB samples were contained in 1.5-mL glass septum vials, which are
0 1994 American Chemical Society
Environ. Sci. Technol., VoI. 28, No. 12. 1994 2191
0013-936X194/0928-2191$04.50/0
Table 1. Ballschmiter Numbers, Observed Dose Constants (kGy-*), and Structure of PCBs
Bz no. (kGy-1) 10.006 f 0.001 2 0.003 f 0.002 3 0.008 i 0.002 4 0.009 f 0.001 6 0.007 f 0.002 110.005 f 0.001
structure Monochlorobiphenyl 2 3 4
Dichlorobiphenyl 2,2‘ 2,3’ 3,3’
type”
Bz no. (kGy1)
M
101 0.014 f 0.002 118 0.015 f 0.002 126 0.028 f 0.003
D
2’,3,4 3,3’,4
M
M
183 0.019 f 0.002
2,2’,4,4’ 2,2’,6,6’ 2,3,4,4’ 2,3’,4’,5 3,3’,4,4’ 3,4,4’,5
D M P*
D
194 0.025 f 0.002 200 0.024 i 0.002 202 0.019 i 0.002
2,2’,4,4’,6,6’ 2,3’,4,4’,5,5’ 3,3‘,4,4‘,5,5’
T M P*
Heptachlorobiphenyl 2,2’,3,4,4’,5’,6
Tr
Octachlorobiphenyl
Tetrachlorobiphenyl 47 0.009 f 0.002 54 0.008 f 0.001 60 0.009 i 0.001 70 0.015 f 0.003 77 0.019 i 0.002 81 0.018 i 0.004
2,2’,4,5,5’ 2,3’,4,4’,5 3,3’,4,4’,5
typea
Hexachlorobiphenyl 155 0.014 f 0.002 167 0.025 f 0.002 169 0.036 f 0.004
Trichlorobiphenyl 33 0.010 f 0.003 35 0.017 f 0.001
structure Pentachlorobiphenyl
T M M P* P
2,2’,3,3’,4,4’,5,5’ 2,2‘,3,3‘,4,5‘,6,6‘ 2,2‘,3,3‘,5,5‘,6,6‘
D T
T
Decachlorobiphenyl 209 0.042 f 0.007
a P = planar; M = monoortho;D = diortho; Tr = triortho; T = tetraortho, * = biochemically most active. The uncertainties shown in the table are at the 90% confidence level and were calculated according to Noggle (26).
then placed in stainless steel vessels along with dosimetry and lowered into the dry tube for a predetermined exposure time. Absorbed doses were measured using FWT-60 radiachromic dye dosimeters and the FWT-100 optical density reader supplied by Far West Technology of Goleta, CA. The details involved in dosimetry calibration and measurements have been described elsewhere (17). There are few neutrons present, so there is no neutron activation. Thus the samples were radiologically clean following depleted fuel y-ray exposure. Analytical Methods and Reagents. Polychlorinated biphenyl measurements were performed using a Hewlett Packard 5995 gas chromatography/mass spectrometer operated in the electron impact mode. A db5-625, 30-m column was used with helium carrier gas at a flow of 2 mL/min. Samples were injected onto the column at 80 “C and held 1min, and then the oven temperature was ramped at 3 OC/min to 310 “C. To ensure the integrity of the analytical measurements, strict quality control measures based on EPA Method 680 (18) were followed. Free chloride ion concentrations were measured by ion chromatography using a Dionex 2010i ion chromatograph and a AS4A separator column with conductivity detection as previously described (5). Carbon-14 measurements were made using a Model 2250CA Packard Tricarb liquid scintillation counter and Ecolume scintillation cocktail. Calibration standards were purchased from Amersham International (U.K.), and the counting efficiency was determined to be 43% for the carbon-14 0,which has an endpoint energy of 156 keV. A l4C-labeled tetrachlorobiphenyl was purchased from Sigma Chemicals (St. Louis, MO) labeled in all 12 positions. The radio-PCB was diluted with stable tetrachlorobiphenyl carrier (Ultra Scientific) and 2-propanol and irradiated in the same manner as the other experiments in this study. Following irradiation, the liquid was counted for carbon14 and compared to the original unirradiated solution. Polychlorinated biphenyl standards were purchased from Ultra Scientific and were greater than 99% pure. When discussing specific PCB congeners, we have adopted thePCB numbering system proposed by Ballschmiter (19). 2182
Envlron. Scl. Technol., Vol. 28, No. 12, 1994
The exact chemical composition for each congener, as well as the Ballschmiter, is given in Table 1. Solvents were reagent grade and were used as purchased. Dose Constant. The figure of merit typically used in radiolysis studies to describe destruction efficiency is the G value. It has been chosen to represent the number of molecules changed (bonds broken or formed) for each 100 eV of energy absorbed. G values in our system exhibit a concentration dependence. This concentration dependence makes the comparison of one study against another, where analyte concentrations vary, difficult. For example, in our system, a 38 mg/L octachlorobiphenyl solution in neutral 2-propanol decomposes initially with G = 0.033 molecules/100 eV, while a 150 mg/L octachlorobiphenyl solution gives G = 0.144 molecules/100 eV. Therefore, we chose not to report PCB destruction in terms of a G value and have adopted the use of a new figure of merit, which we call a dose constant. Determination of the dose constant is based on the fact that, in our studies, pseudo-first-order kinetics have been observed as shown in Figure 1 for PCB 200 as open circles. All of the congeners studied exhibited this same kinetic behavior in the concentration range used. Calculation of the dose constant is performed by taking the slope of the line obtained from the plot of the natural logarithm of PCB concentration versus dose. This is shown in Figure 1 as dark circles. The negative slope of this line, is a firstorder rate constant but with the units of dose.-l Just as a first-order rate constant is independent of concentration, the dose constant is also independent of concentration and thus proved to be more versatile than the traditional G value. For the above example where the G values varied with octachlorobiphenyl concentration, both solutions decomposed with a dose constant of 0.024 kGy’. Use of the dose constant has allowed us to evaluate changes and trends in degradation efficiency for individual PCB congeners previously unreported.
Results and Discussion The radiolytic degradation efficiency or dose constants for 25 congeners are reported in Table 1. If dose constant
--
’
200
I
I
1
I
1 1
I
I
5.5 5
‘2
z
gg
W
E150
K
4.5
88
s2 8z
3
02
2 0
2.5
20
0
40
60
80
ABSOABED W E (my)
100
0,F
Flgure 1. Shown as open circles is PCB 200 concentration versus absorbed dose. There is a steady decrease in PCB 200 concentration with increasing dose. Note the half-llfe dependency on PCB 200 in relation to absorbed dose. Dark circles show the linear relationship betweenthe natural logarithm of PCB 200 concentration and absorbed dose. The correlation coefficient for the above line was 0.996. I I
,-0.04
;
‘a
I I
1
1
1
I
I
I
1
I I
bzo9
I
+
0.03
i g&0.02
i!
-
5-0.
Sa
m
0
2
3.5 50
N
$8
4
100
0.01
0 0
2
4
6
8
10
12
CHLORINENUMBER
Flgure 2. Observed dose constant as a function of the number of chlorines on the biphenyl ring. Specific congeners can be identified by their Ballschmiter number. The exact chemical structure for a given number can be found in Table 1.
is plotted against chlorine number, as shown in Figure 2, a general increase in the dose constant is observed as the number of chlorine atoms on the molecule increases. Within a homolog series, as shown in Figure 2, considerable variability exists demonstrating that the positions of the chlorine atoms also have a significant impact on the observed dose constant. For example, if the dose constants obtained for the hexachlorobiphenyl congeners are evaluated, it can be seen that the dose constant for PCB 169 is over 2.5 times larger than that obtained for PCB 155. Careful review of Table 1 shows that those compounds found to have the largest dose constants for their respective homologs are para/meta substituted only. It appears that chlorine substitution in the ortho position decreases the dose constant. On the basis of the results in our latest paper (5),PCB degradation was proposed to occur by a reductive electron capture process. It is well known that, upon electron capture, the captured electron resides in the lowest unoccupied molecular orbital (LUMO) (20). Thus, a correlation between the observed dose constants and the energylevel or availability of the LUMO may exist. LUMO energies for all 209 PCB congeners have been estimated by Greaves (21,221using a semiempirical general molecular orbital program referred to as MOPAC. (The data generated by Greaves was based on in vacuo conditions
0
1.
1 1
tI
I1
1
0.01
0.02
0.03
0.04
0.05
W E CONSTAM (kGy‘ I )
Flgure 3. Calculated LUMO energy versus observed dose constants. Ballschmiter numbers are also given.
and at 0 K.) If the dose constants obtained are plotted against LUMO energy, as shown in Figure 3, a strong correlation exists. Further inspection of Figure 3 reveals that a decrease in energy level of the LUMO corresponds to an increase in the dose constant. A more complete understanding of the trends shown in Figure 2 is now possible by considering the energy of the PCB LUMOs. Calculations performed by Greaves (21) showed a general decrease in the energy of the LUMO with increasing chlorine content. Organic compounds with positive electron affinities possess conjugated bonds (22) whichlead to a relatively low-lying LUMO. The presence of electron-withdrawing groups, like chlorine, lead to an additional lowering of the LUMO. Thus, as the number of chlorines increases, a concomitant increase in the electron affinity (or lowering of the LUMO) results in a larger dose constant. Congeners that are para/meta substituted only were observed to have the largest dose constant for their respective homolog, as shown in Table 1. Calculated LUMO energy was found to be lowest for the planar congeners within a homolog series (211, consistent with the observed dose constants. These compounds are able to adopt a planar configuration, which results in increased susceptibility to destruction. This is thought to occur because of the extended conjugation that exists between phenyl rings in a planar configuration, resulting in an increase in electron affinity and thus the dose constant. Chlorine substitution in the ortho position was found to decrease the dose constant. PCB congeners that contain o-chlorine substituents are restricted from free rotation around the central phenyl-phenyl bond, and the planar configuration cannot be easily adopted. This ortho effect can be explained by observing the twist angles between the biphenyl rings as function of the number of o-chlorines. The twist angle, measured in degrees, between phenyl rings has been calculated (21). With no o-chlorine atom, the calculated twist angle was approximately 41O. With the addition of one o-chlorine, the angle was increased to 56O. Successive additions of o-chlorine atoms increased bond angles to 76O, 86O, and 87O for two, three, and four chlorine atoms, respectively. Using MOPAC calculations (21),the LUMO wag the orbital whose energy was most affected by increasing the number of o-chlorine atoms. As a general rule, shown in Table 1, increasing the number of ortho substituents did decrease the observed dose constants. Environ. Sci. Technol., Vol. 28, No. 12, 1994 2193
Table 2. Relative PCB Response in an Electron Capture Detector and Relative Dose Constants
Bz no.
chlorobiphenyl
1 2
23-
3
4-
4
3,3'-
11
2,2J-
relative response electron capture dose constant 1.0 0.20
1.10 6.1 5.16
1.0 0.5 1.3 0.83 1.5
Given the above considerations, an unexpected result was obtained when dose constants were compared to LUMO energy for dichloro- and monochloro-substituted congeners. A reversal of this ortho effect appears to have occurred when PCB congeners contain two or less chlorine atoms. Ortho-substituted monochlorobiphenyl and dichlcrobiphenyl congenersdid not have the lowest dose constant for their respective homolog. Reference to Table 1shows that the dose constant obtained for the ortho-substituted congenerswas higher than for meta-substituted congeners. Consistent with other ortho-substituted congeners, calculated LUMO energies are higher than for metasubstituted congeners. This would have predicted a reversal in the observed dose constants. It is currently not understood why this occurs, but this trend has also been observed in the gas-phase electron capture by monoand dichlorobiphenyls (23). The relative electron capture response has been determined empirically for a variety of PCB congeners in the gas phase (23) by measuring the response of individual PCB congeners in an electron capture detector. Table 2 gives the variation in the relative molar electron capture response observed for individual PCB congeners and their relative dose constants. The trend is qualitatively the same. Consistent with our dose constants, ortho-substituted congeners responded better in an ECD than meta-substituted congeners. Toxicity and Dose Constant. As reported by De Voogt et al. (2), the adverse effects of PCBs, particularly the planar PCBs, on wildlife have been documented for 20 years. Some of the more common biological effects, typically expressed in terrestrial animals, include hepatic damage, dermal disorders, reproductive toxicity, body weight loss, and teratogenicity (24, 25). Demonstrating that these toxic planar PCBs are susceptible to radiolytic degradation is of obvious importance. PCB congeners 77, 81,126, and 169are the so-called planar or coplanar PCBs. It has been estimated that the planar PCBs, taken in conjunction with the trace levels of polychlorodibenzop-dioxins and polychlorodibenzofurans typically present, account for most of the observed toxicity of commercial Aroclor mixtures (2). As mentioned, planar PCBs each exhibit the largest dose constant for their respective homolog series. In terms of a treatment process, these high dose constants for planar PCBs are fortunate. Less dose is needed to achieve significant levels of destruction. For example, the dose needed to reduce a given congener concentration in half is readily calculated using the relationship: dose = 0.693Jdose constant Planar pentachlorobiphenyl PCB 126has a dose constant of 0.028 kGy-1, while ortho-substituted pentachlorobiphenyl PCB 101 has a dose constant of 0.014 kGy-'. The dose needed to reduce the concentration of these congeners 2194
Environ. Sci. Technoi., Vol. 28, No. 12, 1994
0
20
40
60
80
100
ABSORBED DOSE (kGy)
Flgure 4. Growthand destructionof less chlorinatedhomologs. Curves a-d correspond to hepta-, hexa-, penta-, and tetrachiorobiphenyis respectively.
F
t
t
t
I
I
1
1
1
1
1
1
I
I
0
20
40
60
80
100
1
ABSORBED DOSE (kGy)
Flgure 5. Simultaneous to the production of less Chlorinated PCBs is the rise in free chloride.
in half would be 24.75 kGy for PCB 126, and for PCB 101 it would be exactly twice that dose or 49.5 kGy. Mass Balance. In any treatment process, identifying the degradation products is fundamental. Exhaustive studies were undertaken to determine the degradation products produced and to provide an accurate carbon and chlorine mass balance. The expected products, assuming dissociative electron capture, would be less chlorinated homologs and free chloride ion. These products are in fact observed. Concurrent with the destruction of PCB 200, shown in Figure 1, is the growth and then the destruction of less chlorinated homologs and the rise in free chloride ion concentration (Figures 4 and 5). This pattern of growth followed by destruction is typical for all of the congeners studied. Figure 6 shows an electron impact (EI) total ion chromatogram for a solution of PCB 200, labeled OCB, in neutral 2-propanol after receiving a 120-kGy dose. The envelope of peaks having retention times shorter than PCB 200 are due to less chlorinated congeners ranging from heptachlorobiphenyl to trichlorobiphenyl. The peaks having a retention time larger than that of PCB 200 are due to the presence of a series of solvent-altered PCB congeners. Solvent-altered congeners were also observed for congeners containing six, five, four, three, and two chlorines. These PCB solvent adducts were identified by a combination of mass spectrometric measurements and derivatization techniques.
Table 3. Carbon and Chlorine Mass Balance When 50% of the Beginning Material Is Destroyeds
PCB 200
7l
lo] 80
Bz dose % carbon % chloride Bz dose % carbon % chloride No. (kGy) recovered recovered No. (kGy) recovered recovered PCB SOLVENT ADDUCTS
i I
I
,
,
,
,
,
40
30
Time (min) Flgure 6. Electron impact total ion chromatogramfor a solution initially containing PCB 200 after receiving a dose of 120 kGy. Peaks are labeled to indicate the level of chlorination on the biphenyl ring. Peaks wlth retentiontimes longer than PCB 200 are solvent altered congeners.
100-
.-,x
3
*
43'
p-qkoH
[M-15]+
CIS
GO-
M.W.
=
450 amu
ri
a)
-
'2 40-
2
20-
Mt
218
A
50
Pentachlorobiphenyl 96 100 101 49.5 95 98 118 46.2 88 80 126 24.8
4 6
Dichlorobiphenyl 77.0 111 99.0 111 91
Hexachlorobiphenyl 93 95 155 49.5 95 93 167 27.7 93 92 169 19.3
33 35
Trichlorobiphenyl 94 90 69.3 106 99 40.8
Heptachlorobiphenyl 97 101 183 36.5
,
20
80-
2 3
Monochlorobiphenyl 88 80 115.5 101 231.0 99 95 86.6
1
150
250
350
L
450
rnlz Flgure 7. Electron impact mass spectrum from solvent altered PCB congeners. Insert shows the proposed structure of the 2-propanolsubstituted PCB assigned to this mass spectrum.
The generalized structure of this PCB-solvent adduct is shown as an insert in Figure 7. This conclusion was based in part by the mass spectra obtained (Figure 7). Inspection of Figure 7 reveals the two major fragment ions, m/z 59 and an isotopic cluster starting at rnlz 435443. Low ion intensity is observed for masses for the molecular ion where a weak molecular cluster at m/z 450458 was observed. This spectrum is consistent with the proposed structure. Confirmation that the 2-propanol group is attached to the biphenyl ring via the CY carbon, as shown in Figure 7, was verified by chemical derivatization with bis(trimethylsily1)acetamide (BSA). An aliquot of irradiated sample was blown dry and then reconstituted in pyridine. A small amount of BSA was then added and refluxed for 10 min. BSA converts free alcohol groups to their corresponding trimethylsilyl (TMS) derivatives. The formation of the TMS derivative was confirmed by the presence of the isotopic cluster starting at rnlz 507, which corresponds to the loss of a methyl group from the TMS derivative. TMS derivatives were observed for all of the adducts identified. If a mass balance is performed on Figure 6, not all of the beginning carbon and chlorine can be accounted for as free chloride, less chlorinated homologs, and PCB
solvent adducts. Approximately 60 % of the beginning carbon and chlorine can be accounted for as either less chlorinated homologs or free chloride. Accurate quantitation of the observed PCB adducts is not currently possible because of the lack of standards. However, if typical response factors obtained from the PCBs themselves are used, it has been estimated that the adducts do not account for more than 10% of the missing carbon or chlorine, If these response factors are correct, this leaves a significant fraction of the mass still missing for PCB 200. Mass balance results for 19 PCB congeners when 50% of the original material has been destroyed are given in Table 3. The ability to recover a complete mass balance appears to vary from congener to congener. An excellent example of this variability is demonstrated by PCB 3 and PCB 1,both monochlorinated congeners. Approximately 95% of the carbon and chlorine is recovered as residual parent PCB, biphenyl, or free chloride for PCB 3, while approximately 80% can be recovered for PCB 1. The amount of mass able to be recovered does not appear to be related directly to absorbed dose but appears to be a function of the individual chemistry of each congener. The majority of the percent recoveries for carbon and chlorine are within one to five percentage points of each other. This is an indication that the unidentified material contains both the carbon and the chlorine. 14C-LabeledTetrachlorobiphenyl. For those congeners for which acomplete mass balance was not possible, a variety of mass spectrometric techniques such as GC/ MS and LC/MS, in positive and negative chemical ionization modes, were used to search for the missing mass. Volatile, semivolatile, and nonvolatile fractions were each evaluated in an unsuccessful attempt to identify the missing mass. Finally, a I4C-labeledtetrachlorobiphenyl congener, PCB 47, was irradiated. Following a series of absorbed doses with a maximum of 149 kGy, 10096 of the activity was determined to be in the bulk irradiated solution. In contrast, using mass spectrometric techniques, only 78% of the original material could be identified as original parent material, less chlorinated congeners, PCBsolvent adducts, and biphenyl. The identity of the missing mass is unknown, and its detection may be hindered by the fact that products which are many steps down a chain of reactions are often present in very small concentrations. In an attempt to determine in which fraction the 14Clabeled products reside, an aliquot was distilled at 93 O C . The activity was not removed at that temperature, thus, it appears to be in a semivolatile or nonvolatile fraction yet to be identified. However no evidence of chlorinated Envlron. Sci. Technol., Vol. 28, No. 12, 1994 2195
dioxans or dibenzofurans as well as the EPA-regulated Method 8270 compound was found. While this is negative evidence, it is significant in that it appears that the degradation products formed are not EPA-regulated compounds. Based on the results that we have to this point, we feel that the most likely location of the missing mass is in solvent-based adducts that have not yet been identified.
Mincher, B. J.; Arbon, R. E.; Knighton, W. B.; Meikrantz, D. H. Appl. Radiat. hot. 1994,45 (8), 879-887. Sawai, T.; Shinozaki, Y. Chem. Lett. 1972, 865-869. Sawai, T.; Shimokawa, T.; Shinozaki, Y. Bull. Chem. SOC. Jpn. 1974,47,1889-1893. Singh, A.; Kremers, W.; Smalley, P.; Bennett, G. S. Radiat. Phys. Chem. 1985,25, 11-19. Schweitzer, J. F.; Born, G. S.; Etzel, J. E.; Kessier, W. V. J. Radioanal. Nucl. Chem. Lett. 1987,118,5-28 and 323-
Conclusions
Merrill, E. W.; Mabry, D. R.; Schulz, R. B.; Coleman, W. D.; Trump, J. G.; Wright, K. A. AICHE Symp. Ser. 1977, 74 (NO.178), 245-248. Shimokawa, T.; Sawai, T. J. Nucl. Sei. Technol. 1977, 14,
329.
The radiolytic degradation of 25 PCB congeners in neutral 2-propanol has been investigated using spent nuclear fuel as the y-ray source. Using spent nuclear fuel as the y-ray source could convert what is now considered a liability into an asset. Degradation occurred for all of the congeners studied, demonstrating that radiolysis could possibly be used as an alternative to high-temperature incineration. Destruction efficiency was found to be dependent on both the number and the position of the chlorines. Those compounds that are para/meta substituted only, the highly toxic planar compounds, were found to have the largest dose constants for their respective homologs. This is attributed to conjugation between the phenyl rings in the planar configuration, which results in an increase in their electron affinity. Chlorine substitution in the ortho position decreases the dose constant. The differences in congener susceptibility to destruction has been explained in terms of the energy level of the LUMO that can accept an electron. Given the above findings, destruction efficiencies or dose constants could be estimated reasonably well for any of 209 PCB congeners of interest. This would be useful in the practical application of a treatment process. A new class of radiolytic compounds has been definitively identified to be an 2-propanolsubstituted polychlorinated biphenyl. Mass balance information is incomplete for some PCB congeners. 14Clabeled experiments demonstrated that the missing mass appears to be in solution. However, no EPA-regulated compounds resulting from the degradation of PCBs have been identified.
Literature Cited (1) Erickson, M. D. Analytical Chemistryof PCBs;Butterworth Publishers: London, 1986; Chapter 1. (2) De Voogt, P.; Wells, D. E.; Reutergardh, L.; Brinkman, U. T. Int. J . Environ. Anal. Chem. 1990, 40, 1-46. (3) Exner, J. H. Detoxication of Hazardous Waste;Ann Arbor Science: Ann Arbor, 1992; Chapter 1. (4) Mincher, B. J.; Meikrantz, D. H.; Murphy, R. J.; Gresham, G. L.; Connolly, M. J. Appl. Radiat. hot. 1991, 42 (111, 1061-1066.
2196 Envlron. Scl. Technol.,Vol. 28. No. 12, 1994
731-736.
Sherman, W. V. J. Phys. Chem. 1968, 72 (6), 2287-2288. Radlowski, C.; Sherman, W. V. J. Phys. Chem. 1970, 74 (16), 3043-3047.
Evans, R.; Nesyto, E.; Radlowski, C.; Sherman, W. V. J. Phys. Chem. 1971, 75 (18), 2762-2765. Sawai, T.; Ohara, N.; Shimokawa, T. Bull. Chem. Soc. Jpn. 1978,51, 1300-1306.
Spinks, J. W. T.; Woods,R. J. An Introduction to Radiation Chemistry, 3rd ed.; Wiley-Interscience: New York, 1990; pp 425-429. Mincher, B. J.; Zaidi, M. K. Radiat. Prot. Dosim. 1993,47, 571-573.
Method 680. Determination of Pesticides and PCBs in Water and SoilJSediment by Gas ChromatographyJMass Spectrometry; U.S. Environmental Protection Agency: Cincinnati, OH, 1985. Ballschmiter, K.; Zell, M. Fresenius’ 2.Anal. Chem. 1980, 302, 20-21.
Kebarle, P.; Chowdury, S. Chem. Rev. 1987,87, 513-534. Greaves, J.; Harvey, E.; MacIntyre, W. G. J. Am. Soc. Mass Spectrom. 1994, 5, 44-52. Greaves, J., University of California, Irvine, CA, Personal communication, 1994. Zitko, V.; Hutzinger, 0.;Safe, S. The Chemistry of PCBs; Robert E. Krieger Publishing Co.: Malabar, FL, 1974. Kimbrough, R. D.; Jensen, A. A. Halogenated Biphenyls Terphenyls, Naphthalenes, Dibenzodioxins, and Related Products, 2nd ed.; Elsevier: Amsterdam, The Netherlands, 1989.
McConnell, E. C. Acute and Chronic Toxicity, Carcinogenesis,Reproduction Tetratogenesis and Mutagenesis in Animals;Kimbrough, R. D., Ed.; Elsevier: Amsterdam, The Netherlands, 1980. Noggle, J. H. Physical Chemistry; Harper Collins Publishers: 1989; Appendix 1.
Received for review March 29, 1994. Revised manuscript received June 23, 1994. Accepted July 26, 1994.’ Abstract published in Advance ACS Abstracts, September 1, 1994.
