Arch. Environm. Contain.Toxicol.7, 301-315

Archives of

Environmental Contamination and Toxicology

Chlorination of Benzidine and Other Aromatic Amines in Aqueous Environments R. L. Jenkins, J. E. Haskins, L. G. Carmona, and R. B. Baird CountySanitationDistricts of Los AngelesCounty,San Jose Creek WaterQualityLaboratory,1965 South WorkmanMillRoad, Whittier,California90601 Abstract. The fate of aniline, N , N - d i m e t h y l a n i l i n e and benzidine in chlorinated waters was investigated. Conditions were controlled to approximate the process chlorination of raw water supplies and wastewater secondary effluents. As the molar ratio, (C12)/(amine), was increased, amine depletions increased and leveled off at about (C12)/(amine) = 1. Depletions in distilled water with 'free' chlorine were somewhat higher than those in activated sludge/secondary effluent with combined chlorine. For each amine the number and type of products appeared to be independent of the water matrix and the ratio, (Cl2)/(amine). For the monophenyl amines ring chlorination was a significant depletion pathway. Extended chlorination of aniline yielded a precipitated product, while the N-substituted amine, N,N-dimethylaniline did not yield a solid product. In contrast to the monophenyl amines, when benzidine (p,p'diaminobiphenyl) was exposed to chlorinated waters, a solid product resulted immediately. Infra-red analysis of this product indicated a polymeric structure with no ring chlorination. GLC analysis of the chlorination supernatant showed no ring substituted isomers of benzidine. The potential carcinogenicity of aromatic amines has been previously demonstrated (Boyland 1958, Bonser et al. 1958, Clayson 1964). Since aromatic amines can be introduced into waters via industrial discharge, assessing the fate of these compounds in water treatment processes is important. The removal and fate of several aromatic amines in activated sludge reactors has been investigated (Malaney et al. 1967, Baird et al. 1977~. Chlorination is the most common disinfection process in water treatment. Compounds resulting from the chlorination of natural and process waters have been characterized (JoUey 1973). An important model for the chlorination of aromatic compounds has been provided by a study of the kinetics of phenol chlorination (Lee 1967). More recently, chlorination kinetics of polyaromatic hydrocarbons have been reported (Perry et aL 1976). The chlorination of aniline has been previously studied, but the products were not conclusively identified (Manufacturing Chemists Association, EPA #12020 EXG). 0090-4341/78/0007-0301 $03.00 O 1978 Springer-Vedag New York Inc.

302

R . L . Jenkins e t al.

In this paper, we have investigated the chlorination of aniline, N,N-dimethylaniline, and benzidine (Figure 1). Conditions were controlled to approximate the process chlorination of raw water supplies and wastewater secondary effluents. Of the three aromatic amines, benzidine (p,p'-diaminobiphenyl) is the most toxic and has been placed on the EPA's list of toxic pollutants (Environmental Protection Agency 1973) and the OSHA list of prohibited materials (Chemical & Engineering News 1974). Aniline and N,N-dimethylaniline were chosen primarily as model compounds for monophenyl and N-substituted monophenyl amine chlorination, respectively.

NH2

Aniline

N(CH3)2

N,N-dimethylaniline (N,N-DMA)

H 2 N ~ N H 2 Benzidine

(p, pLdi(]minobiphenyl)

Fig. 1. Chemical structures of aromatic amines

Chlorination of Aromatic Amines in Water

303

Methods Chromatographic analysis of chlorination mixtures was accomplished by injecting2 to 5-tzlquantities onto a 10% DC-200/Gas Chrom Q column. Temperature settings were as follows for injector, detector, and column, respectively: aniline, 200", 200~ 1200C; N-N-dimethylaniline, 200", 20&, 1300C; and benzidine, 250", 250~ and 220~ All other parameters were as outlined previously (Jenkins and Baird 1975). UV-visible spectra were recorded with a Beckman Acta HI scanning spectrophotometer. A Perkin-Elmer 621 grating-spectrophotometer was used to record infra-red spectra. To record the IR spectrum of the benzidine chlorination product, the solution was filtered through a 1.0 micron glass fiber filter, the filter washed with methanol, allowed to dry and a KBr pellet of the material was formed. Amine standards were prepared in 10-a N HCI, A standard NaOC1 solution was used for chlorination. Gas chromatography/mass spectrometry analyses were accomplished with either of the following two instruments: 1) a DuPont DIMASPEC consisting of a digitally operated GC/low resolution mass spectrometer interfaced to a DECLAB PDP 10 minicomputer; or 2) a Finnigan 3200F with the GCflow resolution quadrupole mass spectrometer interfaced to a Finnigan 6103 integrated data system. GC conditions were similar to those described above for the routine GLC analyses. Mass spectral search techniques similar to those described previously (Hites and Biernarm 1970) were used to assist in the identification of GC product peaks. All reactions were conducted in 20-mlcalibrated ground-glass stoppered tubes. The chlorine dose was 20 mg/L for all experiments and pH values were maintained in the range, 6.5 to 7.5, by adding HC1 or NaOH. To avoid non-representative pH's and reactant concentrations at the beginning of chlorination, pH adjustments and reactant dilutions were carried out prior to mixing the amine and chlorine solutions. When amines were chlorinated in secondary effluent, pH adjustment and addition of the chlorine solution did not amount to more than 5% dilution of the wastewater. Each reaction set included duplicate tubes of 1) C12 + amine in secondary effluent, 2) Clz + amine in distalled water, 3) amine + secondary effluent at pH 7.0 _+ 0.5, and 4) amine + deionized water at pH 7.0 -+ 0.5. All tubes were magnetically stirred for 20 min at the same rate by placing each at the same radial distance on a single stir plate.

Results a n d Discussion I n all cases the designation, (C12)/(amine), indicates a m o l a r ratio. P r o d u c t ratios are b a s e d o n a weight/weight basis, since in m a n y cases the product(s) could n o t be identified c o n c l u s i v e l y . F i g u r e 2 shows p e r c e n t a m i n e d e p l e t i o n as a f u n c t i o n of the ratio, (C12)/ (amine), in distilled w a t e r or s e c o n d a r y effluent with chlorine doses o f 20 mg/L. F o r r e a c t i o n s in s e c o n d a r y effluent the (C12)/(amine) ratio was c o r r e c t e d for the CI~ d e m a n d o f the effluent. T h e o b s e r v a t i o n o f i n c r e a s i n g p e r c e n t d e p l e t i o n with i n c r e a s i n g chlorine to a m i n e ratio is c o n s i s t e n t w i t h the chemical kinetic findings (Lee 1967) for the c h l o r i n a t i o n of p h e n o l i c s . F r o m F i g u r e 2 the p e r c e n t d e p l e t i o n o f a n y g i v e n a m i n e was greater in distilled w a t e r t h a n in s e c o n d a r y effluent. S i n c e a typical s e c o n d a r y effluent c o n t a i n e d 20 m g / L N H a - N , the chlorine residual existed m a i n l y as c h l o r a m i n e s ( > 9 9 % ) c o m p a r e d to distilled w a t e r w h e r e the chlorine residual existed as HOC1 o r OC1- ( < 9 9 % ) . This o b s e r v a t i o n is c o n s i s t e n t with the findings of p r e v i o u s studies w h i c h h a v e s h o w n c h l o r a m i n e s to be w e a k e r c h l o r i n a t i n g agents t h a n HOC1 o r OC1-

304

R.L. Jenkins et I00| |

I

I

I

i

I_

I

Benzidine ~

~

al.

I

N,N-DMA

" N-DMA

801 / / / ==

Benzldine 0

/

~

70 60

so 40

30

-

:~0

-

er

I01 ii

O|

MSecond~ry Effluent I

1.0

I

2.0

I

3.0

I

I

I

I

I

I

4.0 5.0 6.0 7.0 8.0 9.0 (CI2)/(Amine) Fig. 2. % Aminedepletionas a function of the chlorine to aminemolarratio

I

I0.0

For all three amines the depletion with chlorine was rapid and stabilized in approximately 20 min, however, the reaction pathways and products formed were different for each amine, as described below.

Chlorination of Aniline A typical chromatogram following a 1:1 (Clz)/(aniline) reaction is shown in Figure 3. Attempts to assign a compound identification to product peak #1 by GLC retention time or GC/MS were inconclusive. Product peak #2 has been assigned to ortho-chloroaniline by GLC retention and GC/MS analysis. Product peak #3 has been tentatively assigned by GLC retention time as recta- or para-chloroaniline or a mixture of the two isomers. From standard injections the relative FID response of aniline to o-chloroaniline was determined to be 1.6. Assuming this response factor to also be correct for m- and p-chloroaniline, a plot of the percentage of aniline lost which is converted to each respective product as a function of the ratio, (Clz)/(aniline) can be constructed (see Figure 4). For example, if at a 1:1 (Cl~)/(aniline) ratio 19.13 nag of aniline were depleted and 3.83 mg of o-ch/oroaniline were produced, the plotted point would be (1.00, 20.0%). In general, as (Clz)/(aniline) ratios decreased, the percent yields of the chlorinated products decreased suggesting that ring chlorination of aniline was a less favored reaction pathway at higher relative aniline concentrations. For the predominant GLC product, o-chloroani-

Chlorination of Aromatic Amines in Water

305

i

I

I

I

I

I

I

6 Z ,4'--

r

"qo

o L=,

0..

f,0 to Q. r n-

e,EL

rO

d Z "4="

o O Q.

15

13

I

I

II

9

Time

I

!

9 5 (minutes)

I

I

3

I

0

Fig. 3. A typical chromatogram after a 1:1 (Clz)/(aniline) reaction. Conditions as given in Table I and Reference 9. Attentuation--1 x 2

R. L. I r

3116 ,,

i

i

et at.

i

l

I

ODistilled

I

'

l

Water

i..'lSecondary Effluent 5O

"J

.........

l

o-chloroa -o 25 Re

Q~ r

0

o 20 e,-

~

ev

,-. o

15

o

I0

5 Product No.5 0 0.00

1.00

2.00

Product No, 3 3.00

4.00

5.00

(Ci2)/(Aniline) Fig. 4. Percent depleted aniline converted to respective products as a function o f the ctdorine to aniline molar ratio

Chlorination of Aromatic Amines in Water

307

line, the percent yield ranged from near zero at low reaction ratios to a near constant yield of about 28% at > 1:1 reaction ratios. No curve for product # 1 was plotted because of the poor precision and accuracy in quantitating such small amounts. Chlorination of aniline in distilled water or secondary effluent produced a faint violet color, presumably due to the formation of indoaniline, a colored end product which has been previously identified from the reaction of aniline with concentrated NaOC1 solutions, equations 1 and 2 (NoUer 1958a). Since hydroxylation of aniline was indicated as an intermediate process in the formation C~Hs-NH2 + NaOCI--> p-OH-CsHcNH2 + NaC1 Aniline Para-aminophenol

E

p-OH-C6H4-NH2 + 2NaOC1 ~ | HO-C6H4-N =CsH4=NH [_ Indoaniline

(1)

1

(2)

of indoaniline, standard injections of o-, m-, and p-aminophenol were made, but the GLC system used in the study was unsuitable for these compounds. By analogy to the formation of chloramines from the reaction of HOC1 and NH3, organic amines might be expected to form organic chloramines (equation 3). R - NH2 + C12 ~ R - NHC1 + HCI

(3)

Sodium sulfite is known to destroy chloramines (equation 4). When anilinechlorination mixtures were treated with Na~SO3, no change in chromatographic analysis or in solution color was observed, suggesting the absence of organic chloramines. HzO + SO~- + NH2C1 ~ SO~- + NI-I~ + C1-

(4)

If aniline/chlorine mixtures were allowed to stand overnight, a reddish-brown precipitate formed. However, chromatographic analysis of the supernatant agreed with the original GLC analysis of the mixture, and the solution color remained a faint violet. The dye industry has used the oxidation of aniline to produce aniline black, a polymeric product (Noller 1958b). Hence, it is probable that the precipitate observed from aniline chlorination is a polymeric substance of a nature similar to aniline black (Figure 5). It has been reported that prolonged chlorination of phenolics can lead to oxidative rupture of the aromatic nucleus (Lee 1967). In the present study after the initial 20-min contact time, repetitive GLC analysis and continuous absorbance scanning of the UV aromatic bands of representative reaction mixtures for

H N

N

N

H Fig. 5. Proposed structure of aniline black from Reference 12

N,

308

R.L. Jenkins et al.

six hr showed no significant changes. Therefore, under the chlorination conditions employed in this study no destruction of the aromatic ring was indicated as a result of extended chlorination times. In conclusion, it appears that the reaction of aniline with chlorine may proceed via several pathways which involve ring chlorination, ring hydroxylation and/or polymerization. Mass balance considerations around GLC analyses suggest that up to 30% of the aniline depletion can be due to ring chlorination. The formation of indoaniline, a dimeric type aniline species, previously observed in aniline chlorination experiments, was evidenced by the color which developed in the chlorination mixture. The formation of a precipitate from prolonged reaction times suggests the importance of polymerization as a depletion pathway for aniline. Oxidative rupture of the aromatic ring was not indicated as a result of prolonged chlorination and its occurrence in the early stages of the reaction is unlikely because of the high stability of aromatic nuclei. Organic chloramines were not indicated by GLC analysis or solution color changes following addition of excess Na~SO3 to chlorination mixtures.

Chlorination of N,N-Dimethylaniline (N,N-DMA)

Figure 6 shows a typical chromatogram following a 1:1, (Clz)/(N,N-DMA), chlorination. Product peak #1 was identified by GC/MS as a monochloro species of N,N-DMA. Considering the strong ring activation of the -N(CHa)2 group, the analogous aniline results, and the intense light and heat which would be required for alkyl group chlorination (Roberts and Caserio 1964), peak #1 is most probably due to ring substitution of N,N-DMA. Product peaks #2 and #3 may also be due to one or more chlorinated species, but the unavailability of standards prevented conclusive GC/MS identification. Similar to the aniline results (Figure 4) the percentage of depleted N , N - D M A which yielded a chlorinated N,N-DMA decreased as the ratio, (Clz)/N,N-DMA), decreased (Figure 7). As with aniline this observation is suggestive of an additional reaction pathway which does not yield a chlorinated species as a final product. Assuming the relative FID response of chlorinated versus the parent aniline is 1.6, as found for aniline and o-chloroaniline, the yield of the monochlorinated species tended to be greater for N,N-DMA than for aniline, 22 to 56% versus 19 to 29%, respectively. This result is probably due to the greater ring activation of the -N(CH3)2 group. Upon chlorination N,N-DMA solutions become yellow, secondary effluent samples then turned a greenish color within one min while distilled water samples required about 30 min for this greenish color to develop. Eventually the distilled water sample became colorless while the effluent sample remained greenish. Absorbance spectra for typical distilled water and secondary effluent chlorinations at nominal 1:1 (C12)/(N,N-DMA) ratios are shown in Figure 8. In separate experiments it was shown that the absorbance peaks observed were not present in the spectra of standard N,N-DMA solutions or chlorinated secondary. The 720 nm absorbance peak is present in both spectra and is probably responsible for the greenish color imparted to the chlorination solutions. The spectrum from N,N-DMA chlorination in secondary effluent contains at least two other absorb-

Chlorination of Aromatic Amines in Water

309

~

m

L-

o

6 Z

o

'