Headspace determination of evolved carbon dioxide in a biodegradability screening test

ECOTOXICOLOGY

AND

ENVIRONMENTAL

SAFETY

19,204-2 11 (1990)

Headspace Determination of Evolved Carbon Dioxide in a Biodegradability Screening Test J. STRUIJS AND J. STOLTENRAMP Laboratory for Ecotoxicology, Environmental Chemistry and Drinking Water. National Institute of Public Health and Environmental Protection. P.O. Box I, 3720 BA Bilthoven, The Netherlands Received July 31, 1989 A headspace method to measure the carbon dioxide evolved in a screening biodegradability test is described. Use was made of conventional serum bottles, sealed with butyl rubber septa, through which headspace samples were taken after acidification and equilibration of the test solution. For each determination a serum bottle was sacrificed. The gas samples were injected directly into the reaction chamber ofa carbon analyzer. Eighteen chemicals varying in solubility and biological properties were submitted to the test. Where possible, mineralization of test compounds was monitored through dissolved organic carbon analysis in addition to CO2 analyses. The reliability ofthe method was verified by comparing the measured concentration ofdissolved inorganic carbon ofa standard sodium carbonate solution with values calculated from measured concentrations ofcarbon dioxide in the headspace after acidification. Results ofbiodegradability testing are discussed in view of the suitability of the method to poorly soluble compounds. Q 1990 Academic Press. Inc.

I. INTRODUCTION Biodegradation is believed to play a key role in the assessment of the environmental fate of synthetic chemicals. Screening tests to determine the potential of a compound to undergo rapid microbial degradation in the aquatic compartment are recommended by the OECD (198 I), the EEC (1984), and the U.S. Environmental Protection Agency (1982). For regulatory purposes, methods for testing the so-called “ready biodegradability” of chemicals in an initial screening process should be cost-effective, stringent, and appropriate to assessthe ultimate biodegradability (mineralization) of a compound. In addition, this set of test methods should cover substances in a wide range of physicochemical properties. To meet these requirements, nonspecific analytical parameters such as loss of dissolved organic carbon (DOC) and uptake of dissolved oxygen or carbon dioxide evolution are used. As a consequence, initial concentrations of the test compound are in the range of 2-100 mg/liter, while the test substance is the sole carbon and energy source in a mineral nutrient medium which is inoculated with a polyvalent microflora. These tests do not simulate any particular aquatic environment but are applicable to a wide variety of compounds. However, the aqueous solubility of many organic chemicals is below this range, thus restricting the mineralization to be monitored either by O2 uptake or COZ production. In general, oxygen uptake methods are appropriate for testing insoluble substances although an extra oxygen uptake by nitrifying bacteria may cause errors and in some cases may even lead to false positive results. It appears that nitrification is hard to avoid and until now no satisfactory 0147-65 l3/90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

204

CARBON

DIOXIDE

DETERMINATION

IN A BIODEGRADABILITY

TEST

205

solution has been found (Blok et al., 1985). Thus nitrite and nitrate analysis are necessary in these tests in order to distinguish between N-oxidation and C-oxidation. Although nitrifying bacteria utilize CO* and in principle also interfere in methods which rely on evolved carbon dioxide monitoring, their CO2 uptake is rather low with respect to their oxygen consumption. The Sturm test (OECD, 1981; EEC, 1984) and the U.S. EPA guideline method (1982), based on the test described by Gledhill(l975) are the only established procedures that employ CO2 monitoring. In addition to the cumbersome experimental setup, the Sturm test has some drawbacks pointed out by Gerike (1984). Moreover, volatile compounds or metabolites may be stripped off through continuous purging of the carbon dioxide from the test vessels. Specially modified glassware is needed for the U.S. EPA guideline method, and the use of internal vessels for trapping the CO2 limits agitation of the test solution because of spillover. In both tests collection and titration of the trap solution are time consuming and a source of error. Until now little attention has been given to alternative methods for carbon evolution monitoring. Recently, Boatman et al. (1986) described an automated headspace carbon dioxide analysis and its application to biodegradability testing. They conducted the test in 20-ml vials designed for an automated headspace gas chromatograph. However, this procedure limits the usefulness of the proposed test for chemicals that are poorly soluble, because higher test volumes are required if the test chemical is to be weighted in directly. In the present study the use of a carbon analyzer (Dohrmann DC-80) is proposed for monitoring the evolution of carbon dioxide in normal 120-ml serum bottles. The method was applied to estimate the biodegradability of 18 chemicals over a wide range of physicochemical and biological properties. Mineralization of soluble test compounds was monitored as DOC removal in comparison with CO2 production. II. MATERIALS

AND

METHODS

Test compounds. The reference compounds sodium acetate and aniline were of p.a. quality. Beeswax was assumed to have the formula C40.6H80.201 .*a The purity of all other test chemicals exceeded 98%. All soluble compounds were introduced to the test bottles through l-ml additions of stock solutions at the appropriate concentrations. Stock solutions were prepared from twice distilled water. A stock solution of 2-chloro-5-methylphenol(2.22 g/liter) was prepared by dissolving the compound in NaOH solution (pH 12.7) under ultrasonic treatment and adjusting the pH to 7 with 1 M HCl. I-Octanol and di(2-ethylhexyl)phthalate (DEHP) were transferred to the test system through impregnation on small pieces (1 cm2) of triply washed (at 1OoOC) and dried membrane filter paper (0.45 pm Sartorius No. 11106). p-Xylene was introduced directly as a liquid with a pipet. Beeswax (0.5 g) was emulsified with Genapol PF 40 (0.1 g) and nonylphenol lOE0.5PO (0.2 g) in 50 ml distilled water under heating (SOOC)before an appropriate amount was added to the test bottle. The emulsifying reagents were supplied by Akzo Research (Holland). Mineral medium. Mineral nutrient medium was prepared from deionized water and contained per liter 10 ml of solution (a) and 1 ml of solution (b)-(f): (a) 8.5 g KH2P04, 2 1.7 g K2HP04, 33.4 g Na2HP04.2H20, and 0.5 g NH&l per liter; (b) 22.5 g MgS04. 7H20 liter-‘, (c) 36.5 g CaC&. 2H20 liter-‘, (d) 0.25 g FeC&. 6H20 liter-‘, (e) 35.4 mg MnCl?. 4H20, 57.3 mg H3B03, 42.8 mg ZnSO,. 7H20, 34.7 mg

206

STRUIJS

AND

STOLTENKAMP

(NH&Mo70Z4, and 100 mg EDTA liter-‘; (f) 30 mg DIFCO yeast extract in 200 ml. Solutions (a)-(f) were prepared from twice distilled water. Inoculum. Activated sludge was from a local sewage treatment plant (de Bilt, The Netherlands). The concentration of suspended solids was 2800 mg/liter dry matter. The activated sludge was sampled the day before the incubations started and was aerated overnight. Procedure. Test solutions were prepared from mineral nutrient medium, an appropriate amount of activated sludge, and the test chemical. The concentration of the test compound was equivalent to about 10 mg/liter organic carbon except for chloroacetic acid which was also tested at 4.5 mg/liter and the addition of the inoculum corresponded to either 4 or 30 mg/liter dry matter. For the blank series a mineral solution with inoculum was prepared. Separate blank series consisted of mineral medium, inoculum, and the same amounts of dispersion agent or adsorption medium as was used in the test series with poorly soluble compounds. Test and blank solutions were transferred to calibrated 120-ml serum bottles in a headspace/solution ratio of l/2. The serum bottles were sealed with three-part caps consisting of butyl rubber septa, pierceable, and closed aluminum screw tops. Incubations were carried out in the dark at 20°C on a rotary shaker (120 r-pm). Weekly, part of the bottles were sacrificed for CO* headspace analyses and, if possible, DOC measurements of the solution. The CO2 concentrations in the headspace of the acidified solution were related to the corresponding blank from which the level of mineralization was calculated as percent ThCO* (theoretical amount of carbon dioxide of the compound). The test was terminated when percent ThC02 exceeded 70%. Analyses of inorganic carbon. Prior to the analyses a stock solution was prepared from 44 1.7 mg anhydrous sodium carbonate in 1 liter CO*-free water, from which standard solutions of 0, 5, and 10 mg/liter DIC (dissolved inorganic carbon) were made. These standard solutions were transferred to duplicate 120-ml serum bottles in a l/2 headspace/solution ratio. The bottles were sealed with the three-part caps. Bottles to be sacrificed for measurements and bottles containing the freshly prepared standard solutions were acidified by injection of 1 ml 7N H3P04 through the septum. To allow equilibration the serum bottles were shaken in the dark at 120 rpm for 1 hr. After equilibration headspace CO* analyses were performed as follows. Gas samples of 1000 ~1 were taken through the septa and injected directly into the reaction chamber of the Dohrmann DC-80 carbon analyzer. Injection of the gas sample was carried out by piercing a Hamilton gas syringe 1 cm into the injection port and drawing the gas smoothly into the reaction vessel. The concentrations of inorganic carbon produced in the test and in the blank were obtained from the calibration line of the standard solutions. Before gas analyses proper functioning of the carbon analyzer was checked with a 10 mg/liter DOC sodium biphthalate standard solution. DOC analyses. DOC measurements of the soluble test compounds were carried out with a Technicon AA” auto-analyzer on Day 0 and after each carbon dioxide analysis. Samples were centrifuged at 5OOOg before DOC was measured. III.

RESULTS

Partitioning of inorganic carbon in a closed gas-water system and the influence of the test conditions. It can be derived (Stumm and Morgan, 198 1) that if water with a volume V, and an initial concentration of inorganic carbon of DIG, mg/liter is al-

CARBON

DIOXIDE

DETERMINATION

IN A BIODEGRADABILITY

TABLE

cs 2 DIG, (meas.) DIG, (talc.)

207

1

PARTITIONINGOFINORGANICCARBONATVARYING v,lVw

TEST

C'~/V,RATIOS

0.5

1

6.62 + 0.1 1

5.05 + 0.14 (5.00 f 0.06)

3.35 + 0.02 (3.34 + 0.02)

0.99 +f 0.04 6.55 0.02

(0.99 z!z f 0.04) 0.01) 0.98 f+ 0.03 (4.97 4.95 10.39~0.10 9.99 (9.95)

3.44 1.02 f+- 0.13 0.04 (3.42 (1.02 +5 0.08) 0.02)

9.86

2

10.12 (10.09)

Note. C’g and C, are the concentrations in the gas and water phase after acidification (pH < 3) and I hr equilibration. The initial concentration of dissolved inorganic carbon is denoted as DIG,. Results in mineral test medium are given in brackets. All concentrations are in mg C/liter.

lowed to equilibrate in a closed system with a volume V, of COz-free gas, the concentration of carbon dioxide in the gas phase (C, mg/liter) is written as c, = DICo(P/ao + V.VJ’,

(1)

where ,6 is the equilibrium constant for CO2 solubility (C,&‘.J for which a value of 0.95 at 20°C has been reported (Stumm and Morgan, 198 1) and a0 is the concentration ratio of undissociated and total inorganic carbon species, which depends on the PN a0 = (1 + K,/[H+]

+ Kr &/[H+12))l,

(2)

where K, and K2 are the primary and secondary dissociation constants of H2COX. At pH below 3, a0 is approximately equal to 1. Thus using Eq. ( l), the amount of inorganic carbon formed in the water phase due to the mineralization process can be obtained from the carbon content measured in the headspace after acidification (pH < 3) and equilibration. To investigate the partition of inorganic carbon in the test system, a 10 mg/liter DIC solution of Na2C03 was transferred to serum bottles (triplicate) in three different V$ VWratios. After acidification and equilibration, concentrations of carbon in both headspace and solution were determined. The results (Table 1) show that the partition coefficient /3 is close to 1, which is slightly higher than given by Stumm and Morgan (1981). Apparently, the influence of the nutrient mineral salts was negligible. Immediately after the preparation of the Na2C03 solution, the concentration of DICo was determined in triplicate through direct injection into the Dohrmann DC 80. The result of these analyses was 10.39 f 0.10 mg/liter. The concentrations of DICo, calculated from headspace analyses, differed less than 5% from this value. Highly biodegradable compounds. Table 2 summarizes the results of the reference compound sodium acetate and five substances which have been demonstrated to be easily degradable in aquatic systems which are close to the real world (Spain et al., 1980; Rogers et al., 1984; Boethling and Alexander, 1979; Paris et al., 1983). Anticipating that chloroacetic acid might be inhibiting at the applied concentration, the test was conducted at concentrations of 9 and 4.5 mg/liter DOC and with inoculum concentrations of 8.3 and 4.2 mg/liter dry matter, respectively. In spite of the higher inoculum concentration chloroacetic acid did not degrade within 1 week at the con-

208

STRUIJS AND STOLTENKAMP

TABLE 2 TESTRESULTSOFHIGHLYBIODEGRADABLECOMPOUNDSAFTER~WEEKINCUBATION

(mgyier) Blank Na-acetate (ref.) p-Nitrophenol Quinoline p-Acetylphenol Cl-aceticacid s-Butanol

12.0 10.3 11.1 10.9 9.0 4.5 8.1

Inoculum (mg/liter)

c-co2 (mg/liter)

%ThC02

4.2 4.2 4.2 4.2 4.2 8.3 4.2 4.2

1.3-1.3 11.4+ 0.3 10.2-10.8 8.9-10.3 11.0-11.4 3.3-4.2 4.6-4.6 6.8-7.6

84-13 86-92 69-8 1 89-93 14-24 73-73 68-78

Note. C,, is the amount of carbon of the test chemical per liter test solution at t = 0. C-CO2 is the concentration of inorganic carbon calculated from the amount of CO2 evolved at t = 7d. % ThCOz is evaluated from C-CO2 of test compound relative to C-CO2 of the blank.

centration of 9 mg/liter, which was apparently

toxic, whereas 4.5 mg/liter resulted in complete mineralization. All results were confirmed by DOC analyses. Intermediate biodegradability. In the series shown in Table 3, all compounds but p-methoxyaniline have been tested in international ring tests (Painter and King, 1985; CEC, 1985). As usual, aniline was degraded within 1 week (in Table 3 only the results of Days 14 and 28 are shown), whereas pentaerythritol and sodium benzenesulfinate did not pass the 60% level at Day 28. Diethylene glycol exhibits a biodegradability which is consistent with the results of the EC ring test program of 198 1- 1982 (Painter and King, 1985). In the present study m-aminophenol was mineralized after 28 days but only after a lag period of at least 2 weeks. Irregular replication which was found for pmethoxyaniline is a phenomenon frequently encountered when a relatively long lag period precedes the onset of microbial degradation. All results were confirmed by DOC data.

TABLE 3 MINERALIZATION(AS%ThCO*) OFINTERMEDIATELYBIODEGRADABLE COMPOUNDSAFTER~~AND 28 DAYS 28 days

14 days

Blank Aniline (ref.) Pentaerythritol m-Aminophenol p-Methoxyaniline Diethyleneglycol Na-benzenesulfinate

CO (mg/liter)

c-co2 (mg/liter)

% ThC02

c-co2 (mg/liter)

% ThCOz

;6 9.6 10.5 10.5 9.1 8.7

1.8 9.5-10.7 1.4-1.6 1.8-2.1 1.6-7.2 2.9-3.2 1.7-1.7

80-93 0 0 o-54 12-15 0

1.9 10.0-10.2 1.9-2.1 8.9-9.1 2.3-9.8 6.0-6.4 4.3-4.9

84-86 0 67-69 4-79 45-50 35-28

Note. Inoculum: 3.9 mg/liter activated sludge.

CARBON

DIOXIDE

DETERMINATION

IN A BIODEGRADABILITY

209

TEST

TABLE 4 BIODEGRADABILITY

Test compound

OF SPARINGLY

C, (mg/liter) 8.4-l 1.3

1 -0ctanol’

Anthraquinone DEHP Beeswax pXylene

42.5-44.9 34.1-38.7 28.3-29.2 43.7-46.5

2-Cl-5-CH,-phenol

16.6-17.2

(1 Inoculum

concentration

was 4 mg/liter

SOLUBLE

COMPOUNDS

7 days

(RESULTS

14 days

71+3

IN %

2 1 days -

28 days -

nd nd

3-6

18-19 23-30

40-46 45-49

o-2 59 55

nd

nd

0

l-l

ThC02)

10-l

1

28-28 4-5 63 61

0

dry matter.

Poorly soluble compound.s. In the range from poorly soluble (2-chloro-5-cresol) to insoluble (beeswax), six compounds were tested. Except where otherwise indicated in Table 4, all test bottles were inoculated with activated sludge at the level of 30 mg/liter dry matter. Before the test compounds were introduced, the inoculum was allowed to acclimate to the mineral nutrient medium in order to reduce the blank CO2 production. For each different procedure of chemical introduction, corresponding blanks were included (not shown in Table 4).

IV. DISCUSSION The described method ranks among the screening tests for the assessment of “ready biodegradability” of chemicals in water, as recommended by the OECD (198 1) and EEC ( 1984). Arguments are provided by the main features of the test: the amount of inoculum (activated sludge: 3-30 mg/liter dry matter), which is not acclimated to the test chemical, the applied concentrations of the test chemical, which is the sole carbon source, and the use of an easily accessible summary parameter for monitoring the mineralization process. In the EC ring test of 198 1-1982, pentaerythritol was found only inherently but not readily biodegradable (Painter and Ring, 1985). Results of the EC ring test of 1983-1984 (CEC, 1985) were more conclusive as most of the participating laboratories reported that pentaerythritol had passed the mineralization level of 60% of the theoretical oxygen consumption after 28 days. The disagreement with our results may be attributed to the difference in inoculum concentration which was in our test system about 10 times lower than applied during the ring tests. Also the degradative behavior of sodium benzenesulfinate in our test (~60% ThC02) did not match the positive results obtained from the EC ring test of 1983-1984 (CEC, 1985). However, in the same ring test, only 2 out of 12 laboratories reported mineralization (>60% of ThOz) of m-aminophenol at Day 28, whereas in our work ultimate biodegradation started after a lag period of at least 2 weeks but was complete within the following 2 weeks. The suitability to test poorly soluble compounds without the drawbacks of the oxygen uptake methods and the ease of performance with respect to the Strum test are the main advantages of the test. We were able to differentiate several levels of biodegradability of poorly soluble compounds (see Table 4). In these series 1-octanol was included as a highly biodegradable reference chemical. Beeswax and p-xylene

210

STRUIJS

AND

STOLTENKAMP

appeared “readily biodegradable” (>60% ThCOJ, but anthraquinone proved only “inherently degradable.” Mineralization of the latter compound started after a lag period of 2 weeks and was not complete at Day 28. DEHP and 2-Cl-5-CH3-phenol were persistent in the test. In a V$ I’, ratio of l/2, the test is applicable to compounds with a volatility corresponding to Henry’s law constant up to 50 Pa m3 mol-‘. Then the fraction of the test chemical in the headspace of the test bottle will not exceed 1%. The volume of the serum bottles used in this study (120 ml) permits the direct addition of insoluble chemicals to the aqueous phase (80 ml); nevertheless VWis small with regard to volumes used in other screening methods. Boatman et al. (1986) employed 20-ml vials containing 10 ml test solution. In general, small volumes are favorable for automation but unpractical when insoluble compounds are to be introduced at the desired amount. It should be emphasized that the low solubility of many “new chemicals” is a major reason to apply CO2 or O2 biodegradability tests. In the present study we employed different methods to transfer poorly soluble substances to the inoculated medium in the test vessel in order to demonstrate the suitability of the proposed biodegradation test. However, it was not the purpose of this work to address to problem of the introduction of insoluble chemicals in test systems. It is now recognized (Boethling, 1984) that there is a need for a set of working methods, to obtain a satisfactory dispersion of these chemicals in the test medium in order to expose them to the utilizing microorganisms. CONCLUSIONS The experimental conditions in the proposed test method, such as the mineral medium and the inoculum concentration, suit the stringency of “ready biodegradability” according to the OECD hierarchy of testing chemicals. It therefore may serve as a simple alternative for the only carbon dioxide evolution test, i.e., the Sturm test, at this level. Pass levels obtained with highly and intermediately biodegradable substances agree fairly well with reported results of ring test exercises. Chloroacetic acid is highly biodegradable; however, at a concentration of 9 mg/ liter this chemical appeared inhibitory to the mixed microflora. Of the poorly soluble compounds tested in this study, I-octanol, beeswax, and p xylene proved readily biodegradable whereas anthraquinone, DEHP, and 2-chloro5-methylphenol did not pass the test. REFERENCES BLOK, J., DE MORSIER, A., GERIKE, P., REYNOLDS, L., AND WELLENS, H. (1985). Harmonisation ofready biodegradability tests. Chemosphere 14, 1805-l 820. BOATMAN, R. J., CUNNINGHAM, S. L., AND ZIEGLER, D. A. (1986). A method for measuring the biodegradation of organic chemicals. Environ. Toxicol. Chem. 5,233-243. BOETHLING, R. S. (1984). Biodegradation testing of insoluble chemicals. Environ. Toxicol. Chem. 3,5-7. BOETHLING, R. S., AND ALEXANDER, M. (1979). Effect of concentration of organic chemicals on their biodegradation by natural microbial communities. Appl. Environ. Microbial. 37, 12 1 l- 12 16. CEC ( 1985). Degradation/Accumulation Sub-Group, Ring Test Programme 1983- 1984, Assessment of biodegradability of chemicals in water by manomettic respirometry, Final Report (H. A. Painter and E. F. King) Jan. EEC (1984). 84/449/EEC Directive, Annex Part C: Methods for the determination of ecotoxicity.

CARBON

DIOXIDE

DETERMINATION

IN A BIODEGRADABILITY

TEST

211

P. (1984). The biodegradability testing of poorly water soluble compounds. Chemosphere 13, 169-190. GLEDHILL, W. E. (1975). Screening test for assessment of ultimate biodegradability: Linear alkylbenzene sulfonates. Appl. Microbial. 30,922-929. OECD ( 198 I). Organization for Economic Cooperation and Development. OECD guidelines for testing chemicals. Paris. PAINTER, H. A., AND KING, E. F. (1985). A respirometric method for the assessment of ready biodegradability: Results of a ring test. Ecotoxicol. Environ. Suf: 9,6- 16. PARIS, D. F., WOLFE, N. L., STEEN, W. C., AND BAUGHMAN, G. L. (1983). Effect of phenol molecular structure on bacterial transformation rate constants in pond and river samples. Appl. Environ. Microbial. 45,1153-l 155. ROGERS,J. E., Lr, S. W., AND LAWRENCE, L. J. (1984). Microbial transformation kinetics of xenobiotics in aquatic environment. EPA-600/3-84-043. PB84-I 62866, US Environmental Protection Agency, Athens, GA. SPAIN, J. C.. PRITCHARD, P. H., AND BOURQUIN, A. W. (1980). Effects of adaptation on biodegradation rates in sediment/water cores from estuarine and fresh water environments. Appl. Environ. Microbial. 40,126-734. STUMM, W., AND MORGAN, J. J. (198 1). Awuat. Chem. Wiley, New York. US Environmental Protection Agency (1982). Chemical Fate Test Guidelines, EPA 560/6-82-003. NTIS PB82-23308. Washington, DC. GERIKE,