ANALYST, FEBRUARY 1992, VOL. 117 121 High-pressure Microwave Digestion for the Determination of Arsenic, Antimony, Selenium and Mercury in Oily Wastes Milford B. Campbell and George A. Kanert Laboratory Services Branch, Ministry of the Environment, 125 Resources Road, Rexdale, Ontario, Canada M9W5Ll The determination of arsenic, antimony and selenium by hydride generation and mercury by the cold vapour technique, following acid digestion of oily waste samples, can be very difficult if there is any residual organic matter. The preparation stage usually requires prolonged digestion to oxidize the organic matter. The situation is exacerbated if these metals are present as the organometallic derivatives. A method has been developed in which the organic matrices of oily waste samples, after solvent extraction, are completely oxidized.This has been achieved through the use of a newly designed microwave system that uses special high-pressure vessels capable of withstanding internal pressures in excess of 82.2 bar (1200 psi) (1 bar = 105 Pa). The relative standard deviation using organometallic standards ranged from 5.4 to 6.6% and the recoveries for four spiked oily waste samples ranged from 89 to 105%. Keywords: Petroleum-based oily waste; microwave digestion; mercury, arsenic, antimony and selenium determination; cold vapour and h ydride-generation atomic absorption spectrometry The US Environmental Protection Agency Method 1330 'Extraction Procedure for Oily Wastes' is used to determine the mobile metal concentration in petroleum-based wastes' and hence provides a method for screening these types of wastes prior to landfilling.Owing to the expected low concentrations in these types of wastes, arsenic, antimony and selenium are determined by the hydride generation method whereas mercury is determined by the cold vapour method. Both of these methods usually require perchloric acid diges- tion of the samples in order to destroy any organic matter.273 Microwave digestion techniques have been used success- fully to digest different types of samples, ranging from geological4 to biological5 matrices. Fisher4 found that by using a low-pressure system, a decrease in the volatility of the more volatile elements led to improved sample to acid contact and hence a more complete digestion of the samples.Very little work, however, has been performed with these types of samples at internal vessel pressures in excess of 13.7 bar (200 psi). In this study, a microwave technique employing specially designed high-pressure digestion vessels was used to digest oily waste samples in a non-perchloric acid medium at pressures in excess of 82.2 bar (12 psi) (1 bar = 105 Pa). Experimental Microwave Digestion System The Milestone microwave system (Mandell Scientific, Guelph, Ontario) consists of three separate units: oven, fume extraction module and a capping station. The MLS-1200 oven is capable of producing a maximum power of 1200 W. An actual value of 1100 W was found by monitoring the change in temperature of 1 kg of water. The power can be varied from 1 to 50% in 1% increments and then by 25% increments to 100%.A microprocessor is used to control the power. For the acid digestions, a maximum of 50% power was used. High-pressure Vessels The Milestone high-pressure vessels (HPV 80) are constructed of a patented polymer derivative that allows the use of high boiling-point acids such as sulfuric acid at temperatures up to 350 "C. The vessels used had special seal rings and burst discs rated at 100 bar (1460 psi). Four vessels at a time were placed in the oven. Instrumentation A Varian Techtron Model 70 atomic absorption spectrometer capable of operating at 193.70, 217.58 and 196.03 nm for the determination of arsenic, antimony and selenium, respec- tively, was used. The atomizing chamber is a 10 cm long X 6 mm i.d. quartz tube wound with Chrome1 C insulated resistance wire (approximately 1 5;2 per 0.3 m).The operating temperature of 850 "C was controlled by a Variac transformer. The sample flow together with various reagents was maintained by use of a Gilson Minipuls 2 proportional pump (Varian Techtron). Tygon tubing of various inner diameters was used to connect the different units. The sample and reagents were mixed in the two mixing coils prior to introduction into the atomizer in a stream of argon via a centrally placed 2 mm i.d. quartz tube. The output from the atomic absorption spectrometer was fed into a strip-chart recorder set at an input voltage of 10 mV. A schematic diagram of the analytical system is shown in Fig. 1. An LCD/Milton Roy mercury monitor set at 254 nm was used to determine mercury.Reagents All inorganic acids used were of Baker analysed ACS grade, except sulfuric acid which was of BDH analytical-reagent grade. Doubly distilled water was used as a diluent. Toluene and tetrahydrofuran used for the extractions were from Caledon Laboratories (distilled in glass). Organometallic Standards No reference materials for metals in oily wastes were available. As the primary area of interest was in the metals that were present as the organometallic derivatives and not simply as particulates in oil, the following organometallic standards obtained from Conostan Division, Conoco (Ponca City, OK, USA) were used: antimony alkyl sulfonate in white mineral oil (5000 pprn); selenium amine sulfonate in white mineral oil (100 pprn); arsenic amine sulfonate in white mineral oil (100 pprn); and mercury alkyl dithiocarbomate in white mineral oil (100 pprn), A working mixed standard solution of 1 pprn was prepared by diluting the standards appropriately with toluene.122 Proportioning f Pump i L L - I ANALYST, FEBRUARY 1992, VOL.117 Argon 300 cm3 min-' r Heated quartz cell Recorder and atomic absorption spectrometer Sampler = 1:2 Fig. 1 Atomic absorption spectrometer manifold for the determination of arsenic and antimony Inorganic Standards The aqueous inorganic standards for arsenic, antimony and selenium were prepared by diluting lo00 mg dm-3 certified atomic absorption stock solutions (Delta Scientific). The mercury stock standard solution (100 mg dm-3) was obtained from J. T. Baker and diluted appropriately.Oily Wastes The following oily wastes were used: OW-1, oil-saturated clay; OW-2, tank bottom from oil refinery (lumpy solid); OW-3, sludge from oil refining (soft, oily solid); and OW-4, aqueous oily sludge (aqueous brown upper portion and solid lower layer). Owing to the high incidence of volatile pet- roleum products such as diesel fuel in some of the oily waste samples, the solids were air dried for 24 h in a fume hood prior to use. Procedure Duplicate oily wastes were Soxhlet extracted according to the US Environmental Protection Agency Method 1330.1 These extracts were composed of a mixture of toluene and tetra- hydrofuran together with the organics solubilized in them. The combined solvent extracts were filtered through a 0.46 pm glass-fibre filter to remove any large particulates that would pass through the extraction thimble.A 2 ml aliquot of the mixed solvent extracts was pipetted on to an 8.5 x 5 cm piece of Whatman QA-M quartz microfibre filter horizontally supported at one end. The filter was left at room temperature in a fume hood for approximately 20 rnin to allow for the bulk of the solvents to evaporate. It is important that blank determinations are performed on these filters prior to use; they were found to produce fairly high blanks for selenium and arsenic. Therefore, the filters were pre-treated by boiling in 6 mol dm-3 hydrochloric acid for approximately 15 min. The excess of acid solution was drained and the filters were then dried in the microwave oven followed by heating in a muffle furnace at 450 "C for at least 5 h.This treatment should produce blank values below 0.1 ppb for these elements. Each filter was cut into small pieces and then placed in a high-pressure digestion vessel. A 5 ml mixture of sulfuric and nitric acid (3 + 1) was carefully added to each of the vessels. The burst disc was placed in position, then the cap was screwed in place and tightened in the capping station to 10 N m. The vessels were inserted in the rotating safety shield and the whole arrangement was placed in the microwave oven. The samples were digested according to program 1: 20% power for 2 min, 30% power for 3 rnin and 50% power for 6 min. [Caution: It is important with these types of samples that a gradual rate of heating is used, otherwise the pressure build-up may be too rapid.It is also important that, if the waste sample is suspected to contain unstable compounds, the bombs are left in their safety shield in the oven long enough to allow adequate cooling and cessation of any chemical reac- tions. If this is not done, it is possible for one or more of the bombs to vent on removal from the oven.] At the end of the digestion period, the assembly was removed from the oven and immersed in cold water until the vessels had cooled sufficiently that they could be removed from the safety shield. The vessels were opened carefully with the venting tube in place (according to the manufacturer's instructions). In order to ensure complete digestion, a further 3 ml of the sulfuric-nitric acid mixture were added cautiously to each vessel and they were resealed as above.The samples were further digested according to program 2: 25% power for 2 min and 50% power for 10 min. At the end of this time, the vessels were cooled and opened as described above. Approximately 5-10 ml of distilled water were added slowly to each of the digests; care was taken at this stage as there was a vigorous reaction with the evolution of brown fumes. The solutions were allowed to stand for 5 rnin and then gently swirled. The filter pieces were then washed into clean 150 ml glass beakers with doubly distilled water. These solutions and the filters were completely colourless once all the nitrogen dioxide had escaped. The solutions were thoroughly stirred with a glass rod and then filtered through Whatman No.541 filter-paper into 50 ml calibrated flasks. After the initial solutions had drained from the filters, they were rinsed with small volumes of doubly distilled water. The filters were pressed with a glass rod to remove as much of the rinse solution as possible and the contents of the flasks were then diluted to the mark with doubly distilled water. It has been found from this work that the nitric acid remaining within the final solution suppresses the arsenic and antimony signals. In order to counteract this, 0.2-0.3 g of urea was added slowly to a 10 ml aliquot of each of the above solutions contained in 15 ml acid-rinsed sample vials. After the effervescence had subsided, the vials were capped and inverted several times to ensure complete mixing. These samples were then analysed for arsenic, antimony and selenium.ANALYST, FEBRUARY 1992, VOL. 117 123 For the determination of arsenic, antimony and selenium by the hydride generation atomic absorption method, it is important that these metals are in the correct oxidation state.The results obtained from the work performed by Sinemus et a1.6 show that much better sensitivities are obtained for these elements when they are in their lower oxidation states. In the present study, the reduction to the lower oxidation state was achieved by using 10% m/v potassium iodide for arsenic and antimony and 6 mol dm-3 hydrochloric acid for selenium. In each instance these solutions were automatically added by using a proportioning pump and mixed in a ten-turn glass coil as indicated in Fig.1. Reduction to arsine, stibine and hydrogen selenide was accomplished by the introduction of 2% sodium tetrahydroborate(i1i) into the flow system prior to the second mixing coil. Certain transition metals, such as nickel, copper and iron, are known to interfere in the determination of arsenic, antimony and selenium . 7 3 According to Kirkbright and Taddia,g the interference effect could be due to the reduction of these transition metals to the free metals; these could then absorb the hydride product, thus reducing the signal. It has been foundg-10 that 5 mol dm-3 hydrochloric acid extends the analytical range in which these analytes can be determined in the presence of these interfering metals. Inductively coupled plasma spectrometric studies have shown that one or more of these transition metals can be present, in excess of the concentrations known to cause interference problems,g in oily waste extract samples, especially if the particulate matter is not removed from the extract.For this reason, 6 mol dm-3 hydrochloric acid was mixed with the sample solution prior to the reduction stage. Mercury was determined by pouring the remaining diges- tate solutions into small glass beakers. Saturated potassium permanganate was added dropwise until a persistent pink colour was obtained (Landi et al.11 noted the merits of having a strongly oxidizing environment for the determination of mercury). The beakers were covered with watch-glasses, placed on a rotating platform and then inserted in the microwave oven. These samples were slowly heated at 15% power for 15 min.Samples that became colourless during this time were separately treated with potassium permanganate as above and re-heated. The manganese dioxide formed in this process was dissolved by careful treatment with 35% v/v hydrogen peroxide to the point where a pink colour was just obtained. If the hydrated manganese dioxide is not removed, it can hinder the oxidation of the mercury ions." The final solution for the determination of mercury was treated with a 20% m/v solution of hydroxylamine sulfate to reduce the potassium permanganate. The ionic mercury was then reduced by tin(ii) chloride to its elemental form in an automatic mixing system similar to that shown in Fig. 1. Digestion recovery studies were carried out in two stages.The first was by evaluating the recovery of the mixed standards only, and the second by evaluating the effects of the different oily waste samples under investigation on the added standards. In each instance, 0.5 ml of the 1 ppm mixed organometallic standard in toluene was pipetted on to the support and left to evaporate. In the second method, the 0.5 ml volume was added to the support after the solvent from the sample had evaporated. These samples were microwave digested in the same manner as the oily waste extracts. Results and Discussion The recoveries obtained from the digestion and analysis of mixed 10 ppb organometallic standards are shown in Table 1. The lower values found in the 8 ppb range were attributed to a slight loss of pressure in these particular vessels.The short-term precisions (relative standard deviations) are fairly consis tent. Results of the oily waste spiking study are presented in Table 2. The recoveries ranged from 90 to loo%, indicating Table 1 Recoveries (%) obtained from the digestion of the mixed 10 ppb organometallic standards Mercury 9.6 10.4 9.1 8.5 9.4 Average 9.4 Standard Relative deviation (YO) 0.622 standard deviation (Yo) 6.6 Arsenic Antimony Selenium 10.0 10.1 10.1 9.6 10.3 8.8 8.9 8.6 9.5 10.8 9.9 10.3 9.6 10.1 9.8 9.8 9.8 9.7 0.621 0.613 0.523 6.3 6.2 5.4 Table 2 Recovery data from the analysis of spiked duplicate oily waste samples Element Mercury Arsenic Antimony Selenium Parameter Concentration in extract + spike (PPb) (PPb) Average extract concentration Recovery at the 10 ppb spike level (Yo) Concentration in extract + spike (PPb) (PPb) Average extract concentration Recovery at the 10 ppb spike level (Yo) Concentration in extract + spike (PPW (PPb) Average extract concentration Recovery at the 10 ppb spike level (%) Concentration in extract + spike (PPb) (PPb) Average extract concentration Recovery at the 10 ppb spike level (%) Sample OW4 OW-2 OW-3 O W 4 9.7 11.3 10.3 9.6 0.6 1.3 0.8 0.6 91 100 95 90 9.6 9.4 11.4 10.3 0.4 0.5 0.9 0.8 92 89 105 95 10.1 10.5 11.8 9.8 0.7 0.8 2.1 0.3 94 97 97 9s 9.6 10.1 10.5 9.8 4 .3 " 0.5 0.7 <0.3* 96 96 98 98 * Estimated detection limit for selenium made from 30 ( n = S), where o is the standard deviation of the results obtained from the support and solvent blanks. that the digestion method was acceptable for the sample types investigated. Cross-contamination between samples was not a problem owing to the material used in the construction of the high-pressure vessels.Rinsing each vessel with 5% v/v nitric acid followed by rinsing with distilled and doubly distilled water was all that was required. Conclusions Petroleum-based oily waste samples are extremely difficult to oxidize completely via acid digestion. It has been found from earlier studies carried out under normal atmospheric condi-124 ANALYST, FEBRUARY 1992, VOL. 117 tions that residual organic matter in solution will result in numerous problems in the determination of mercury, arsenic, antimony and selenium. With the automated method of analysis used, coating of the mixing tube by an oily film occurs, which results in a gradual decrease in sensitivity.This film also results in drift of the recorder baseline and irregularly shaped peaks. With the high-pressure microwave system and using the method described, all of these problems were eradicated. One of the main disadvantages of this high-pressure technique is that there is no way of accurately monitoring the internal pressure and temperature inside the high-pressure vessels. It was established, however, that the internal pres- sures developed in the vessels were between 82.2 and 95.9 bar (1400 psi) by carefully monitoring the conditions under which a burst disc would rupture. This microwave technique, and other high-pressure systems that are becoming commercially available, will be increasingly used to digest samples with difficult matrices. The authors thank all those who assisted in this project, particularly T. McIntosh and H. Konstantinou for carrying out the hydride measurements and D. Russell for carrying out the mercury determinations. Special thanks are extended to J. Pimenta for technical support and providing different samples. References 1 Extraction Procedure for Oily Wastes, Method 1330, Test Methods for Evaluating Solid Waste, SW-846, United States Environmental Protection Agency, Washington, DC, 1986. May, K., and Stoeppler, M., Fresenius Z. Anal. Chem., 1984, 317, 127. Official Methods of Analysis of the Association of Official Analytical Chemists, Association of Official Analytical Chem- ists, Washington, DC, 13th edn., 1980, sect. 25, p. 110. 4 Fisher, L. B., Anal. Chem., 1986, 58, 261. 5 Abu-Samra, A., Morris, J. S., and Koirtyohann, S. R., Anal. Chem., 1975,47, 1475. 6 Sinemus, H. W., Melcher, M., and Welz, B., At. Spectrosc., 1981, 2, No. 3, 81. 7 Welz, B., and Melcher, M., Analyst, 1984, 109, 569. 8 Welz, B., and Melcher, M., Analyst, 1984, 109, 573. 9 Kirkbright, G. F., and Taddia, M., Anal. Chim. Acta, 1978, 100, 145. 10 Welz, B., and Melcher, M., Analyst, 1984, 109, 577. 11 Landi, S., Fagioli, F., Locatelli, C., and Vecchietti, R., Analyst, 1990, 115, 173. 12 Stainton, M. P., Anal. Chem., 1971,43,625. 2 3 Paper 1102546E Received May 30, 1991 Accepted August 6, 1991