Organic compounds in an industrial wastewater. Their transport into

Lopez-Avila, and Ronald A. Hites. Environ. Sci. Technol. , 1980, 14 (11), .... Christopher M. Reddy, James G. Quinn, and John W. King. Environmental S...
0 downloads 0 Views 909KB Size
Organic Compounds in an Industrial Wastewater. Their Transport into Sediments Viorica Lopez-Avila Midwest Research Institute, 425 Volker Boulevard, Kansas City, Missouri 641 10

Ronald A. Hites' School of Public and Environmental Affairs and Department of Chemistry, Indiana University, 400 East Seventh Street, Bloomington, Indiana 47405 H The wastewater from a small speciality chemicals manufacturing plant located on the Pawtuxet River (Rhode Island, USA) has contaminated the water and sediment of that river, the Pawtuxet Cove, the Providence River, and (to a lesser extent) the Narragansett Bay. Since the compounds found in this system cover a wide range of functionalities, polarities, and water solubilities, a detailed study of this system has allowed us to assess the environmental behavior of seueral compound types in one aquatic system. We find that the aqueous concentrations of the various compounds follow the rules of simple dilution and that those compounds with the highest octanol-water partition coefficients (log P ) are strongly associated with the particulate matter in the water and are found in the sediment at the greatest distance from the plant. The sediment concentrations (C) of a given compound can be predicted from its log P value and from its concentration in the wastewater (CO)by log (CoIC) = bo b2(dist/log P ) where dist is the distance of the sediment sample from the plant and bo and bz are constants fitted to the data.

\

\

I

Flgure 1. Map of the Pawtuxet River showing the plant location and sediment core sampling sites (I-VIII and stations 1, 5, 6, and 15).

+

The deposition of potentially toxic organic compounds from the wastewaters of chemical processing plants into aquatic sediments is a known phenomenon. Petroleum hydrocarbons ( I ) and Kepone (2) are two examples. Our previous study ( 3 ) on the environmental impact of a small chemical plant is another. In that study, we identified over 120 different organic compounds in the wastewater of a specialty chemicals manufacturing plant and in the nearby river water and sediment. The maximum concentrations of these compounds were 15 ppm in the wastewater, 0.2 ppm in the river water, and 700 ppm in the river sediment. With the exception of some interesting phenol oxidation reactions (which we will report on elsewhere), there were no chemical transformations observed in this system. Unlike most sediment studies, in which only one or a few coi,ipounds have been identified and quantitated, we measured many compounds covering a wide range of functionalities, polarities, and water solubilities ( 3 ) .This was a rare opportunity to study in one aquatic system the behavior of seueral compound types. Therefore, we have carried out a detailed study of this system. This paper summarizes the results of our field measurements and presents some useful generalizations by which the sedimentary behavior of industrial organic compounds can be predicted.

10 km

-

I 71.30'

Experimental Section Sampling Area. The plant setting and the receiving water system are described by the maps shown in Figures 1 and 2. The plant wastewater is discharged after some treatment into the Pawtuxet River, Rhode Island (USA). This river, in turn, enters the brackish Providence River through the Pawtuxet Cove. The Providence River flows into Narragansett Bay, an estuarine area of high recreational and commercial activity. Figures 1and 2 show the locations of the sampling sites in the Pawtuxet River, the Pawtuxet Cove, the Providence River, and Narragansett Bay. Details of the sampling locations, dates, and types are given in Table I. 1382

Environmental Science & Technology

I 71.1'

I 7r 20'

I 71.15'

Figure 2. Map of the Providence River and Narragansett Bay showing the sediment core sampling sites (stations 2-17).

Procedures. Wastewater samples were collected in 1-gal, amber glass bottles directly from the clarifier tank. The river water samples were taken 0.5 m below the water surface by divers in 1-gal, amber glass bottles or from bridges with a bucket. Sediment cores were taken by divers with a 6 cm i.d. X 80 cm stainless steel sampler or 24 X 16 X 8 cm bottomless cans. The sediment cores taken with the 6-cm sampler were sectioned into 3- or 4-cm layers at the sampling site and were transferred to 1-qtjars; they were kept frozen until analyzed.

0013-936X/80/0914-1382$01.00/0

@ 1980 American Chemical Society

The smaller cores taken with the bottomless cans were kept frozen in the can for 3-4 weeks before they were sectioned (6-8-cm layers) and transferred to glass jars. Details about sample preparation and analytical techniques have been given elsewhere (3, 4 ) . Gas chromatography with flame ionization detection, gaschromatographic mass spectrometry in both electron impact and chemical ionization modes, and high-pressure liquid chromatography with UV detection and mass spectrometric identification were used for the analysis of these samples. Quantitation of the organic compounds was based on peak area measurements relative to external standards. Reported concentrations represent minimum values since they have not been corrected for solvent extraction efficiencies. The extraction recoveries of various compounds spiked into distilled water a t -80 pg/L varied between 85 and 104%. Duplicate analyses had errors of less than f 2 0 % for water samples and less than f 4 0 % for sediment samples, without taking sampling errors into consideration. Under the conditions used, the approximate detection limits for most compounds were 1ppb for the wastewater, 0.1 ppb for the river water, 0.1 ppm for the river sediment, and 0.05 ppm for the bay sediment. Aqueous solubilities of the benzotriazoles and the Cle-BHT ester (see Figure 3 for structures) were measured by the dynamic coupled column, liquid-chromatographic technique developed by May e t al. ( 5 ) .Saturated solutions of the compound to be tested were generated by pumping pure water (high-performance LC distilled water, Baker) a t a flow rate of 2 mL/min through a stainless steel column (60 cm X 7.8 mm i.d.) packed with 60/80 mesh silane-treated glass beads coated with the compound to be tested. The coating of the beads with the compound was done by adding 20 g of beads to 200 mL of

a 0.1% methylene chloride solution and removing the solvent under vacuum with rotation. Before any measurements were taken, 500-700 mL of water was pumped through the generator column to remove fine particles and equilibrate the column. A known volume of pure water was then passed through the generator column and into the analytical column (3.9 mm i.d. X 30 cm p-Bondapak-Cis, 10 pm, Water Associates) where the compound to be tested was retained. Elution of the test compound from the analytical column was done under gradient elution conditions using a combination of acetonitrile and water. For the more water soluble compounds (the trichlorodiphenyl ether and the triazine), the method consisted of equilibrating a mixture of the compound to be tested with water until saturation was achieved. The solutions were then filtered through precleaned Whatman GF/F glass fiber filters (0.7 pm), and the filtrates were extracted with methylene chloride a t pH 7. This method was also used to measure the solubility of stearyl alcohol. The solubility of phenylbutazone was obtained from the literature (6).All solubilities are given in Table I1 as log S . The n -octanol/water partition coefficients (expressed as log P ) were calculated from the water solubilities by using a correlation published by Chiou et al. (7): log P = 5.00

- 0.670 log s

r = 0.985

(1)

where S is in pmol/L. The regression used to establish this equation extended over a factor of los for solubility and lo6 for partition coefficient. The log P value of diphenyl ether was taken from the literature (8);log P values for dibenzoazepine and the phenylnaphthylamine were calculated from a correlation that we established between the high-performance LC

HO dibenzoazepine CAS NO. 256-96-2 log P=4.3

C,-benzotriazole CAS NO. 2440-22-4 log P = 4.6

&triazine CAS NO. 1912-24-9 log P=3.9

C,,-benzotriazole CAS NO. 25973-55-1 log P=5.9

phenylbutazone CAS NO. 50-33-9 log P = 2.7 chloro-benzotriazole CAS NO. 3864-99-1 log P ~ 5 . 8

phenylnaphthylamine CAS NO. 25619-54-9 log P=5.8

diphenylether CAS NO. 101-84-8 log P =4.2

C

/ O\

Q

C H ( C H 1 ,,OH

% ~

‘OH

C

I

trichlorodiphenylether CAS NO. 3380-34-5 log P = 4.2

stearyl alcohol CAS NO. 112-92-5 log P=5.8

Cls-BHT ester CAS NO. 6386-38-5 log P=7.2

CI’

Flgure 3. Structures, abbreviated names, Chemical Abstracts Service registry numbers (CAS No.), and octanol-water partition coefficients (log P) for the 11 compounds studied. Volume 14, Number 11, November 1980

1383

Table 1. Sampling Locations in the Pawtuxet River, the Providence River, and Narragausett Bay sample no.

date

I

9/19/77

I1

9/19/77

111

9/19/77

IV

6/26/78

V

6/26/78

VI

6/26/78

VI1

6/26/78

Vlll

6/26/78

1

6/28/77

2

6/28/77

3

6/28/77

4

6/28/77

5

7/29/77

6

7/29/77

7

7/29/77

8

7/29/77

9

7/29/77

10

9/19/77

11

9/19/77

12

9/19/77

13

9/19/77

14

9/19/77

15

9/19/77

16

9/19/77

17

9/19/77

71' 41' 71' 41' 71' 41' 7 1' 41' 71' 41' 71 41" 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41' 71' 41'

24' 30" 46' 10" 23' 50" 46' 07" 23' 22" 45' 55" 24' 25" 46' 10" 23N 45" 46" 07N 23' 42" 46' 05" 23' 20" 45' 53" 23' 17" 45' 55" 23' 13N 45' 36N 22' 51" 45' 40N 22' 25" 45' 41" 22' 24N 41' 48" 23' 13" 45' 36" 23' 09" 45' 26" 22' 44" 45' 15N 22' 28" 44' 20" 21' 57" 43' 23N 19' 20" 40' 52" 20' 12" 42' 47" 21' 11" 43' 10" 21' 57" 44' 19" 22' 42N 45' 55" 23' 07" 45' 51" 23' OOM 38' lorr 23' 40" 35' 00"

'

depth of water, m

type of sample

sample location

1

Pawtuxet River (near plant)

sediment core (30 cm) water

2

Pawtuxet River (1 km downstream from the plant)

sediment core (12 cm) water

2

Pawtuxet River (near dam)

sediment core (42 cm) water

1

Pawtuxet River (near plant)

sediment core (51 cm) water

2

Pawtuxet River (1 km downstream from the plant)

sediment core (54 cm) water

2

Pawtuxet River (1 km downstream from the plant)

sediment core (39 cm) water

2

Pawtuxet River (near dam)

sediment core (68 cm) water

2

Pawtuxet River (near dam)

sediment core (18 cm) water

2.5

Pawtuxet Cove

sediment core (21 cm) water

3.5

head of channel, entering Pawtuxet Cove

sediment core (21 cm) water

3.5

near Sabin Point, east of Pawtuxet Cove

sediment core (21 cm) water

4.5

near Gaspee Point

sediment core (35 cm)

3.5

Pawtuxet Cove exit

sediment core (18 cm)

2.5

inside Pawtuxet Cove

sediment core (16 cm)

3

buoy 31

sediment core (19 cm)

4.5

buoy 27

sediment core (21 cm)

7.5

buoy 23

sediment core (20 cm)

7

Ohio Ledge

sediment core (13 cm)

7

buoy N-16

sediment core (14 cm)

11

Conimicut Point

sediment core (18 cm)

11

buoy N-28

sediment core (20 cm)

6

buoy N-2

sediment core (9 cm)

3

inside Pawtuxet Cove

sediment core (21 cm)

9

Patience Island

sediment core (21 cm)

11

retention times (expressed as log t,) and partition coefficients calculated from solubilities: log P = -0.54

+ 4.6 log t ,

r = 0.980

(2)

Similar correlations have been reported previously (9).All of the resulting log P values are summarized in Table 11.

Results and Discussion Several hundred quantitative measurements were made on this river and bay system. These data were obtained as a function of specific compound, distance from the chemical plant, depth in core, and date. It would be very difficult to present all of these raw data in a format which makes them comprehensible to the reader. We have, therefore, chosen to present summary data abstracted from the raw information. Furthermore, since only a few of the many compounds found in this environmental system either occur throughout the 1384

station Identification

sediment core (36 cm) water

Environmental Science & Technology

north of Jamestown

system or are present at relatively high concentrations, we will limit our attention in this paper to eleven compounds. Figure 3 gives the structures of these compounds and assigns an abbreviated name to them. The reader is referred to our previous paper on this system (3) for a discussion of the industrial significance of these particular compounds. Figure 3 also gives the logarithm of the octanol-water partition coefficients of these compounds and their Chemical Abstracts Service registry numbers. All of the individual measurements for these eleven compounds and for many others are available ( 4 ) .For example, anyone interested in the concentration of a specific compound at a particular station as a function of depth in the core may obtain the individual values elsewhere ( 4 ) . An additional 32 compounds, not reported previously ( 3 ) , were identified in samples taken from the river and estuary system between 1977 and 1978. A list of these compounds; their concentration ranges in the wastewater, river water, and

The dilution factor of a wastewater in a river is the ratio of river to wastewater flows. The wastewater discharge rate was regulated a t 4.8 X IO6 L/day ( 3 ) ,and the annual average Pawtuxet River flow rate is 1300 X lofiL/day ( I O ) . This is a dilution factor of 270. We calculated river water concentrations from this dilution factor and from the wastewater concentrations; the resulting values are given in Table IV in the last column. With the exception of the river water sample taken near the plant, the river water and cove water concentrations are all within experimental error of the calculated value. This implies that the water concentrations up to the outlet of the cove (see map, Figure 1)are directly related to the concentrations discharged by the plant. In all cases, the water concentration nearest the plant is somewhat higher than predicted. This is probably because contaminated, suspended material is carried over from the waste treatment system and analyzed as part of the water. Since the Providence River is a tidal area, no average flows are available. Thus, we cannot predict an expected concentration. In general, the Providence River concentrations are about a factor of 5 lower than the Pawtuxet River and Cove values. Certainly, the dilution effect is a t least this large. In summary, the water data hold no surprises. The aquatic concentrations of the compounds discussed here follow the rules of simple dilution in the Pawtuxet River and Cove and, presumably, in the Providence River. Sediment Analyses (River). Although analyses of grab sediment samples indicated the accumulation of several organic compounds in the river sediment ( 3 ) ,these data could not provide reliable information about this effect. This information could only be obtained from sediment cores. Therefore, eight sediment cores were taken by divers a t three locations in the river (see Figure I). The core sites were selected for an abundance of fine-grained material, since these would be most likely to give information about accumulation of the industrial compounds.

Table II. Summary of Solubilities, High-Performance LC Retention Times, a and Octanol-Water Partition Coefficients of the 11 Compounds P

log S,prnol/L

log lr, rnln

log

0.551 - 1.37 -1.21

1.15 1.47 1.40

1.25

1.12 1.06 0.92 0.68 1.38

4.6 5.9 5.8 4.2 4.2 4.3 = 3.9 2.7 5.8C 5.8 7.2

C1-benzotriazole Clo-benzotriazole chlorobenzotriazole diphenyl ether trichlorodiphenyl ether di benzoazepine C6-triazine phenylbutazone phenylnaphthylamine stearyl alcohol C18-BHT ester

1.57 3.36 -1.26 -3.33

1.59

High-performance LC conditions: 3.9 rnm i.d. X 30 crn, p-Bondapak-C18, 10 prn, CH3CN/H20. Reported in ref 8. Calculated from log tr - log Pcorrelation. Reported in ref 6. a

sediment samples; and the identification method are given in Table 111. Concentration ranges are also reported for several compounds which were identified previously ( 3 ) ,but for which no concentration data were available a t that time. Water Analyses. Table TV gives the average water concentrations for the 11 compounds. These are geometric averages of two to five values measured a t the specified locations a t different times. The geometric average was selected because it is less sensitive to anomalously high or low measurements; but, in almost all cases, the geometric average, the arithmetic average, and the median did not differ from one another by more than 30%.We see that the C1- and Clo-benzotriazoles, the trichlorodiphenyl ether, stearyl alcohol, the CIS-BHT ester, and the phenylnaphthylamine were discharged a t levels above 1000 ppb. Others appear in the wastewater at lower but consistent levels.

N E A R DAM NEAR

loooor&

looor

DAM n

1000

100 IO E

2

i "

-

a a

I I

KM DOWNSTREAM

I

K M DOWNSTREAM

loot

IO

~

1

1

1

1

1

1

1

1

,

1

1

1

1

1

1

,

~

,

1

1

,

-

NEAR PLANT

PLANT

looor

V

looh

IO00

10

NEAR

u c

0

5

1

1 1 1 I I I 1 1 I I 1 1 1 1 1 1 1 1 I I I I

3

9

15

21

27

39 45 D e p t h i n core ( c m ) 33

51

57

63

Figure 4. C,-benzotriazole concentration profiles in Pawtuxet River sediment cores.

1 1 1 1 1 l l l l l l l I I I 3

9

15

21

27

Depth i n

33

39

45

51

57

63

core ( c r n )

Figure 5. C,o-benzotriazole concentration profiles in Pawtuxet River sediment cores. Volume 14, Number 11, November 1980

1385

Table 111. Summary of New Compounds Identified in the Pawtuxet River Study compd

5-(2-aminoethyl)dibenzo [ b,f ] azepine 2-(2‘-hydroxy- tert-amylphenyl)2Kbenzotriazole 2-(hydroxy- tert-butylmethylphenyl)-5-chloro-2K benzotriazole 2-chloro-4-ethylamino-6-isopropylamin0)-s-triazine 2-chloro-4,6-bis(ethylamino)s-triazine methylquinoxaline dimethylquinoxaline

wastewater

concn range, ppmd river estuary water sediment

rlver water

*

ND

GC/MS (El and CH4CI) HRMS

0.1 (1)

ND

GC/MS (El) HRMS

ND

ND

ND

2-25(8)

ND

GC/MS (El) HRMS

ND

ND

ND

3-1400 (20) ND

GUMS (El)

ND

ND

ND

2-1600 (26) ND

ND

GC/MS (El)

ND

GUMS (El)

ND

GC/MS (El)

0.01-1.6 (5)

0.01-0.03 (4)