Potentially Toxic Organic Compounds in Industrial...
0 downloads
62 Views
2MB Size
5 Potentially Toxic Organic Compounds in Industrial Wastewaters and River Systems: Two Case Studies
Downloaded by UNIV OF ARIZONA on September 6, 2013 | http://pubs.acs.org Publication Date: March 13, 1979 | doi: 10.1021/bk-1979-0094.ch005
RONALD A. HITES, G. A. JUNGCLAUS, V. LOPEZ-AVILA, and L. S. SHELDON Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139 The identification and quantitation of potentially toxic substances in the environment requires the application of sophisticated analytical techniques. Ideally, these should exactly identify each of several hundred compounds present in very complex mixtures even though each species may have an environmental concentration of less than a part per billion. The most generally useful and widely employed analytical tool which meets these requirements is gas chromatography mass spectrometry (GCMS). In this paper, we will briefly review sample isolation methods which are used with GCMS and present two case studies on the organic compounds in industrial wastewaters and river systems which demonstrate these and other principles. Concentration Methods. The GCMS analysis of an environmental sample starts with the isolation of the organic compounds from the matrix (air, water, food, etc.) into a form suitable for introduction into the GCMS instrument, typically a solution in a volatile solvent. This concentration step includes essentially three major methods: vapor stripping, solvent extraction, and lipophilic adsorption. We have recently reviewed the detailed operation of these methods(1),(See also Bellar, Budde and Eichelberger, this volume) but their general features will be outlined here. In the vapor stripping technique, one bubbles an inert gas up through a column of water. This sparging effect strips the volatile organics from the water into the gas phase which is then passed through a cold trap or, more commonly, through a lipophilic trap such as Tenax. This trap is, in turn, placed into the gas chromatographic stream and heated. The organics are thus vaporized and analyzed by the GCMS system. The technique is most applicable to very volatile compounds such as chloro form, but by heating the water, compounds with volatilities up to the equivalent of a C20 normal hydrocarbon can be isolated. Vapor stripping is a very sensitive technique; part per trillion analyses of organic compounds in drinking water have been reported (21 Solvent extraction is a very widely used and simple preconcentration technique. After the sample is extracted with a suitable solvent (such as methylene chloride), the extract is concentrated by evaporation and subjected to analysis. One important requirement is extremely clean solvents; fortunately these are now commercially available. Because of the evaporation step, solvent extraction cannot be used for the analysis of very volatile compounds. Depending on sample size, sensitivities of 0.1 ppb can easily be achieved. In the lipophilic adsorption technique, large volumes of water are passed through 0-8412-0480-2/79/47-094-063$07.00/0 © 1979 American Chemical Society In Monitoring Toxic Substances; Schuetzle, D.; ACS Symposium Series; American Chemical Society: Washington, DC, 1979.
64
MONITORING TOXIC SUBSTANCES
Downloaded by UNIV OF ARIZONA on September 6, 2013 | http://pubs.acs.org Publication Date: March 13, 1979 | doi: 10.1021/bk-1979-0094.ch005
a column packed with a material which has an affinity for organic compounds; materi als such as charcoal, X A D resins, and Tenax have been used. A l l of these adsorbents have a non-polar surface; thus, they tend to accumulate the lipophilic (fat soluble) compounds rather than more water soluble species. After sufficient water has passed through the column, the column is drained and extracted with a suitable solvent. The extract is then evaporated and analyzed by G C M S . For this reason, this technique has the same volatility limitations as solvent extraction. Lipophilic adsorption can, how ever, be used for the analysis of very large volumes of water (up to 500,000 liters; see reference 3), thus sensitivities are usually excellent. These three concentration techniques have different applicabilities. They differ in their sensitivity, in the polarity of the analyte, and in their procedural difficulty. Obviously the proper technique must be selected based on the problem at hand. It may be wise to do an exploratory study with each technique to see what range of concentra tions and compound types are present in a typical sample. The remainder of this paper is a report on two case studies which w i l l present the detailed organic analyses of two industrial wastewater and receiving water systems. In the first study, we w i l l discuss the detailed environmental impact of a small specialty chemicals plant on its receiving water. In the second study, we w i l l report on the detailed organic analysis of Delaware River water, a river which receives wastewater from many large-scale chemical producers. In these cases, we are concerned with the identities of compounds entering the receiving waters and sediments, the compounds already present, and others which may be formed through in situ transformations. Some of this information has appeared elsewhere (4). Specialty Chemicals Plant This plant operates in a batch production mode, generally following a weekly schedule. A wide range of compounds including pharmaceuticals, herbicides, antioxi dants, thermal stabilizers, ultraviolet light absorbers, optical brighteners, and surfac tants is produced. Water is used in synthetic processes, in solvent recovery, in steam jets, and in vacuum pump seals. The wastewater is neutralized in either of two onemillion gallon equalization tanks, passed through a trickling filter for biological degra dation, and clarified in a 150,000 gallon tank with a residence time of two hours. The water spills over from the clarifier at a rate averaging 1.3 χ 10 gal/day and enters the river through an underground pipe about 100 yards away. Only about one fourth of the total BOD, which averages 12,000 lb/day, is removed by the waste treatment system and much of this is in the form of low molecular weight solvents. e
Sampling. The water and sediment sampling sites in the vicinity of the plant are shown in Figure 1. Water samples were collected in one gallon amber glass bottles with Teflon-lined caps. Wastewater samples were collected as the water spilled over from the clarifier. River water samples were collected both upstream and downstream from the plant from bridges and from a small boat. About 300 ml of Nanograde (Mallinckrodt) methylene chloride and 15 ml of 12M hydrochloric acid were added to the water samples at the collection site (except those used for volatile organic analysis) in order to minimize biological degradation and to start the extraction. Sediment samples were collected with a dredge-type sampler from a boat and also with the aid of a diver. One quart glass jars with aluminum foil-lined caps were used for the sediment samples; after collection they were placed in a box containing dry ice. The composition of the river bottom sediments varied from coarse sand (>600μ) in the center of the river to coarse and fine silt toward the banks.
In Monitoring Toxic Substances; Schuetzle, D.; ACS Symposium Series; American Chemical Society: Washington, DC, 1979.
In Monitoring Toxic Substances; Schuetzle, D.; ACS Symposium Series; American Chemical Society: Washington, DC, 1979.
Figure 1. Map of the specialty chemicals plant and its environmental setting. Sampling sites are indicated: letters represent water samples and numbers represent sediment samples. Point C is the clarifier at the plant.
Downloaded by UNIV OF ARIZONA on September 6, 2013 | http://pubs.acs.org Publication Date: March 13, 1979 | doi: 10.1021/bk-1979-0094.ch005
i"
H
o
-*
s
|
>
s
pi
66
MONITORING TOXIC SUBSTANCES
Downloaded by UNIV OF ARIZONA on September 6, 2013 | http://pubs.acs.org Publication Date: March 13, 1979 | doi: 10.1021/bk-1979-0094.ch005
Procedures. When returned to the laboratory, the sediment samples were placed in a freezer and the water samples to be used for volatile analysis were placed in a refrigerator. A Teflon-covered magnetic stirring bar was added to each of the water samples containing methylene chloride for overnight extraction on a magnetic stirrer. Then most of the water was poured into another clean bottle, and the methylene chloride extract was separated from the remaining aqueous phase in a separatory funnel. A plug of pre-extracted glass wool was used to aid in phase separation for those samples containing emulsions. The extracts were rotary evaporated to the de sired volume. It might be noted that Kaderna-Danish distillation may be more suitable for quantitative work. The decanted water from the acidic extraction was made alkal ine with a pre-extracted, concentrated K O H solution and extracted with an additional 200 ml of methylene chloride to recover the basic compounds. The exact concentrations of some of the volatile solvents in the wastewater were determined using direct aqueous injection of 2 μ\ aliquots onto a 2 m χ 0.32 cm ID stainless steel column packed with 0.4% Carbowax 1500 on Carbopak C (Supelco, Inc.) and analyzed by G C / M S . Qualitative analyses of the volatile organic compounds were performed using vapor stripping. About two liters of a river water sample or about 200 ml of a waste water sample were put into a 3 liter glass stripping vessel similar to that described by Novotny et al (5). The water temperature was maintained at about 80°C. Purified helium was passed through the sample from a glass frit located at the bottom of the apparatus at a rate of 120 m l / m i n . Helium and the stripped organics were passed through a water cooled condenser into two glass sampling tubes connected in parallel. These tubes were glass injection port liners from the gas chromatographs and were packed with about 40 mg of precleaned 60/80 mesh Tenax-GC porous polymer adsor bent. The liners were conditioned at 250°C for at least an hour in the injection port of the gas chromatograph prior to use. After vapor stripping for the desired length of time, the pre-columns were removed and stored in Teflon-lined screw-cap test tubes until analyzed by G C and G C M S . The sediment samples were allowed to thaw at room temperature and then sieve-washed through a 2 mm stainless steel screen to remove pebbles and extraneous debris. Excess water was decanted and the wet sediment was Soxhlet extracted for several hours with Nanograde isopropyl alcohol. A further extraction with Nanograde benzene was necessary i n order to isolate the polycyclic aromatic hydrocarbons. The isopropanol extract was evaporated to dryness on a rotary evaporator at 30-40°C; the benzene extract was freed of elemental sulfur by passage through a column of colloi dal copper (6). Instrumentation. Preliminary gas chromatographic analyses were carried out on a Perkin-Elmer 900 gas chromatograph equipped with a flame ionization detector and on a Hewlett-Packard 5730A gas chromatograph equipped with flame ionization and electron-capture detectors. The columns used were 180 cm χ 2 mm ID glass columns packed with 3% SL-2100 (a methyl silicone fluid) on 80/100 mesh Supelcoport; we also used 25 m χ 0.25 mm ID glass capillary columns statically coated (7) with SE-52. Liquid chromatographic separations were performed on a Waters M o d e l A L C / G P C 204 liquid chromatograph equipped with two model 6000 pumps, a model 660 solvent programmer, and a model 440 dual U V absorbance detector. Low resolution (~800) mass spectra were obtained with a Hewlett-Packard
In Monitoring Toxic Substances; Schuetzle, D.; ACS Symposium Series; American Chemical Society: Washington, DC, 1979.
5.
HITES ET AL.
Industrial
Wastewater
Toxins
67
5982A G C M S system with a dual E I / C I source and interfaced to an H P 5933A data system. The quadrupole mass spectrometer was coupled to the gas chromatograph via a glass-lined jet separator held at 300°C. The mass spectrometer was usually operated in the electron impact mode, and spectra were obtained by continuous scanning under control of the data system. The instrument was also operated in the continuous scanning mode during analysis of collected liquid chromatographic fractions introduced with the direct insertion probe. Even when the L C fractions contained more than one compound, some fractionation of the compounds due to differential volatility allowed collection of relatively clean mass spectra for individual components.
Downloaded by UNIV OF ARIZONA on September 6, 2013 | http://pubs.acs.org Publication Date: March 13, 1979 | doi: 10.1021/bk-1979-0094.ch005
High resolution (~20,000) mass spectrometric data were obtained on a Dupont 21-110B instrument with photographic plate detection. These plates were read on an automatic densitometer operated on line to an I B M 1800 computer (8). Compound Identification. Our analyses resulted in the accumulation of thousands of spectra. These were initially compared to those published in the Eight Peak Index (9). Nearly all of the solvents, phenols, and hydrocarbons were found in this source, but the spectra of only a few of the remaining compounds were found here. Standard interpretation procedures were then used to identify the remaining compounds. Unfortunately, due to the normally high degree of aromaticity, the mass spectra of many of the unknowns consisted of only a few peaks, making identification difficult. However, a single high resolution mass spectral analysis of an extract provided the exact masses for most of the major molecular and fragment ions observed during a low resolution G C M S analysis. Thus, we were able to assign probable formulae to the various ions. This information often enabled us to identify structures related to other compounds already identified in the samples. When this was not successful, the Formula Index of Chemical Abstracts and the U.S. Trade Commission Report (10J, which lists production data for individual and classes of compounds, were searched for related compounds produced by the company. The company advertising literature, which presented use and toxicity data for some of its products, was also useful in identifying a few compounds. Computer assisted library search routines (11,12) were not generally used. Because the Eight Peak Index contained most of the compounds in these data bases, time could be more economically utilized in performing the manual procedures described above. Although the company officials and plant personnel cooperated fully in acquiring wastewater samples for analysis, they provided little information concerning identities of reactants of company products. Specific ring substitution patterns are not easily derived from mass spectrometric data and are included here only for cases where the literature specifies the isomer or in cases where we have purchased standards and observed identical mass spectra and gas chromatographic retention times. Results and Discussion. A l l of the compounds identified in the wastewater, river water, and sediment samples are listed in Table I along with their concentration range. The individual concentrations have an estimated error of 20%. The structures of several of the compounds in Table I are given in Figure 2. The identification of these 123 compounds (see Table I) was made possible only by the synergistic application of several analytical techniques. For example, the very high concentrations of a few compounds in most of the samples (e.g., no. 6,10, 46, 81), precluded identification of many of the minor components during G C M S analysis. This dynamic range problem was solved, at least qualitatively, by H P L C followed by mass spectrometry.
In Monitoring Toxic Substances; Schuetzle, D.; ACS Symposium Series; American Chemical Society: Washington, DC, 1979.
68
MONITORING TOXIC SUBSTANCES
Table I. Summary of All Compounds Found in Wastewater, River Water, and Sediment Compound no.
Compound name
Wastewater
C o n c n range, ppm River water
Sediment
Present in tar b a l l 3
A/-Containing heterocyclics 1 2 3 4 5 6 7 8
Downloaded by UNIV OF ARIZONA on September 6, 2013 | http://pubs.acs.org Publication Date: March 13, 1979 | doi: 10.1021/bk-1979-0094.ch005
9 10
Acetylpyridine Dibenzo|b,f jazepine* 10,11 -Dihydrodibenzo [ b.f ) azepine * 5-(3-Dimethy laminopropyl)-10,11 -dihydrodibenzo [ b,f | azepine *
0.05"
