ISSN:0003-2654    

DOI:10.1039/AN98106FX013

出版商:RSC    

年代:1981

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98106BX015

出版商:RSC    

年代:1981

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98106FP041

出版商:RSC    

年代:1981

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98106BP047

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

APRIL 1981 The Analyst Vol. 106 No. 1261 Spectrophotometric Determination of Dissolved Titanium in Sea Water after Sodium Diethyldithiocarbamate Pre-concentration C. Y. Yang, J. S. Shih and Y. C. Yeh Institute of Nuclear Energy Research, P.O. Box No. 3, Lung-Tan, Taiwan, Republic of China A simple and rapid method for the determination of dissolved titanium in sea water is described. Using sodium diethyldithiocarbamate as precipitant, titanium, iron(III), zinc and part of the magnesium and calcium ions that are naturally present in sea water are coprecipitated. After collecting the precipitate by filtering through a 0.45pm Metricel membrane filter, the qualitative recovery of titanium and the coprecipitation of titanium with other ions were studied by X-ray fluorescence spectroscopy.The precipitate was easily dissolved in dilute hydrochloric acid and then a sensitive 4,4’-diantipyrylmethane - thiocyanate extraction method was applied, followed by spectrophotometric determination of titanium. The concentration of dissolved titanium in Tong Shiau sea water was found to be 0.11 pg 1-l with a relative standard deviation of 5%. Keywords : Titanium determination ; sea water ; spectroph otometry ; sodium diethyldithiocarbamate The concentration of titanium in sea water was first reported by Griel and Robinson1 who used thymol as a chromophoric reagent to determine titanium spectrophotometrically, but they admitted that a more sensitive method was required. Subsequently a few other papers2-5 appeared, but in most of them attention was concentrated on the determination of the titanium content of sediments and particulate matter in sea water using emission spectrography.However, accurate data6 for the dissolved titanium content of sea water were still not available. As titanium, like other heavy metal ions, is often concentrated in the particulate phase5 and the total concentration of titanium in sea water is only a few parts per 1Og,1s5s7 and also it is easily hydrolysed and has a short residence time, the determination of very low concentrations of dissolved titanium is a challenging problem. Hydrous titanium dioxide is well known as a good adsorbent8 for extracting uranium from sea water. In order to obtain information on the mass balance of titanium in the process of extraction of uranium from sea water, the determination of titanium in sea water is necessary.Sodium diethyldithiocarbamate is commonly used in spectrophotometric analysis9 and recently it and other dithiocarbamates have been used to form precipitates with transition metal ions from matrices.lO-l3 In this work, titanium was first coprecipitated with zinc and iron(II1) ions by using sodium diethyldithiocarbamate as the precipitant and also coprecipitated with magnesium hydroxide, then a membrane filter made of cellulose triacetate (pore size 0.45 pm) was used to retain the precipitates for analysis. Owing to the high void volume of membrane filters, large numbers of sea water samples could be handled in a short time by using high suction flow-rates. The chloroform extraction of ion pairs of titanium thiocyanates and 4,4’-diantipyrylmethane14 was one of the most sensitive methods for the spectrophotometric determination of titanium.The ion pair [(HDAM+),J [Ti(SCN)62--]15 was formed in aqueous solution, and the distribution coefficient of the complex was so high that the pre-concentration of titanium by extraction with a small volume of an organic extractant was easily achieved and only a single extraction was required. 385386 Samples Sea water was collected 10 m below the surface, 1.5 km off the shore of Tong-Shiau, Taiwan, and first filtered through a plastic coarse-pore filter to remove muds and grains of sand, then through a 0.45-pm cellulose triacetate membrane filter. Subsequently the water was acidified to pH 2 with hydrochloric acid and stored in acid-leached polyethylene containers until taken for analysis.YANG et al.: TITANIUM IN SEA WATER AFTER Experimental Analyst, Vol. 106 Apparatus A Gelman magnetic filter funnel of polysulphone construction with an alkylbenzenesulphonate (ABS) plastic support screen, magnetic coupling and an effective filtration area of 9.62 cm2, and a 0.45-pm GA-6 Metricel cellulose acetate membrane filter were used to retain particulates for analysis. X-ray fluorescence spectroscopy was carried out using a Tracor North, Model NS-880, instrument equipped with a lithium-drifted silicon energy-dispersive detector. A Varian Techtron 634 ultraviolet spectrophotometer was used to determine titanium. Chemicals All chemicals were of analytical-reagent grade. Water. Water was distilled in quartz after de-ionisation with a Millipore reverse-osmosis Titanium could not be detected spectrographically in this water following Obtained from Merck, purified by isoosmotic distillation Obtained from Merck and purified by isoosmotic distillation. This was purified by extraction with 4- methylpent an-2-one.Prepared by dissolving ammonium titanium oxalate in 1% hydrochloric acid at 50 "C. This is commercially available, but can be synthesised by a combination of ZAtka and Hoffmann's method17 and Van de Velde et al.'s method.l* Diantipyrylmethane. Tokyo Kasei Co. Tin(l1) chloride. Merck. Potassium thiocyanate. membrane system. Hughes et a1 .'s procedure16 for evaluating ultra-pure water quality. and diluted as necessary. Hydrochloric acid, concentrated.Ammonia solution, sp. gr. 0.88. Sodium diethyldithiocarbamate solution, 4% m/V. Standard titanium solution, 1 pg ml-l. Ammonium titanyl oxalate [ammonium dioxalatooxotitanate(1V) 3. Merck. Dissolve in water to give a 30% m/V solution. Procedure To 1-6-1 volumes of acidified sea water in polyethylene containers, add 0, 0.5 and 1.0 ml of standard titanium solution (in recovery tests only), then to each add 1 ml of 4% sodium diethyldithiocarbamate per litre of sea water, ammonia solution (sp. gr. 0.88) to adjust the pH to 9.0-9.5, mix for 5 min and filter through a 0.45-pm Metricel membrane filter using a water aspirator. After completing the filtration separation, dissolve the precipitate in 10 ml of 2 N hydrochloric acid and heat gently. To the clear solution, add 5 ml of 30% m/V potassium thiocyanate solution, 3 ml of 5% m/V diantipyrylmethane solution, 2 ml of con- centrated hydrochloric acid and 2 ml of 10% m/V tin(I1) chloride solution.Extract the titanium complex with 5 ml of chloroform for 2 min and measure the absorbance at 420 nm using a reagent blank as reference. Perform the reagent blank test using 4 1 of de-ionised, distilled water and the same procedure as for sea water samples. Results and Discussion The presence of trace amounts of titanium in sea water was confirmed by the absorption spectrum shown in Fig. 1. The molar absorptivity of [(HDAM+),] [Ti(SCN),2-] at 420 nm in chloroform is 6.0 x lo4 1 mol-l cm-l. A qualitative recovery test was performed by adding 2 pg of titanium to 1 1 of sea water, precipitating with sodium diethyldithiocarbamate and collecting the filtrate on a membrane filter, followed by counting the membrane directly using an energy-dispersive X-ray fluores- cence spectrometer.The titanium was coprecipitated with zinc and iron(II1) ions in seaApril, 1981 SODIUM DIETHYLDITHIOCARBAMATE PRE-CONCENTRATION 387 0.2 a 0 C OJ + 0.1 2 n a 380 420 460 Wavelengt h/n rn Fig. 1. Absorption spectrum of [(HDAM+)z][Ti(SCN)6z-] in chloroform using a 6-1 sea-water sample. ZnKa f I 0 ’ u Channel number d Fig. 2. Coprecipitation X-ray fluorescence spectrum 2 pg of titanium spiked in 1 1 of sea water. of water, as shown by the X-ray fluorescence spectrum (Fig. 2). Unfortunately, owing to its low sensitivity, coprecipitation followed by X-ray fluorescence spectroscopy cannot be applied to the determination of titanium in sea water; further studies on improving the sensitivity by varying the experimental conditions are currently being carried out in our laboratory. Standard titanium solution (1 ml) was added to 5 1 of sea water in order to carry out a quantitative recovery test by using precipitation pre-concentration followed by spectrophoto- metric measurement.In four assays carried out during the precipitation and separation steps, the recovery of titanium was 95 & 5%. The titanium concentration of the Tong Shiau sea water is reported in Table I. Results for the seven aliquots that were spiked with titanium are also included. The effect of pH on the recovery of titanium from sea water spiked at the level of 1 pg 1-’ is shown in Fig.3. As pH values higher than 9.8 cause too much precipitation of magnesium hydroxide and interference in the subsequent filtration process, the pH value in the procedure was fixed between 9.0 and 9.5. A 1-ml volume of 4% m/V sodium diethyldithiocarbamate precipitant per litre of sea water was found to be sufficient for a satisfactory recovery of titanium. The only four ions in sea water that, according to Marczenko,lS could be considered to be potential interferents in diantipyrylmethane - thiocyanate extraction and direct spectro- photometric determination of titanium were Mo(VI), Cu(II), Co(I1) and Fe(II1). Therefore, 20 pg of Cu(II), 80 pg of Mo(VI), 5 pg of Co(I1) and 30 pg of Fe(II1) were added to 4 1 of sea water (these levels are about double those found in natural sea water) and the proposed procedure was carried out.Zinc ions form a colourless complex with diantipyrylmethane, so they do not interfere in the deter- mination of titanium. Iron(II1) ions were reduced with tin(I1) chloride, so that they would no longer interfere. In order to prevent adsorption losses, the collected sea water should be acidified to pH 220 and all experimental procedures be completed within 24 h. Contamination control was always checked by carrying out a blank test. The blank value was about 10-20 ng of titanium. Extra care was taken to prevent contamination from labora- tory vessels, reagents and filters. The cleaning procedures for plastic containers suggested by ’ No interference from these four ions was observed.4 6 8 1 0 2 60 40 > a 20 PH Fig. 3. Effect of pH on the recovery of titanium from sea x x r a t e r cnilrorl Qt fho 1 . . r r l - l388 YANG, SHIH AND YEH TABLE I CONCENTRATION OF TITANIUM FOUND IN SEA WATER SAMPLES AND TOTAL TITANIUM FOUND IN THE SAME SAMPLES SPIKED WITH ADDITIONAL TITANIUM with 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Volume of sea water/l 1 1 1 2 2 2 4 4 4 4 5 5 5 5 6 6 6 6 Amount of titanium added/pg 0 0.5 1.0 0 0.5 1 .o 0 0 0 1 .o 0 0 0 1 .o 0 0 0 1 .o * -, Below detection limit. Total titanium Titanium concentration found/pg in sea water/pg 1-1 - * - 0.55 - 1.00 - 0.65 1.20 0.10 0.44 0.11 0.44 0.11 0.40 0.10 1.40 0.10 0.55 0.11 0.50 0.10 0.55 0.11 1.55 0.11 0.65 0.11 0.64 0.11 0.66 0.11 1.68 0.11 - - Moody21 were followed.bench to prevent contamination from dust. had generally been decreasing over the years. Samples were handled inside the laminar flow hood of a clean air Christensen22 mentioned that the reported concentrations of some trace metals in sea water The results of this work are in accordance such a trend. References Griel, J . V., and Robinson, R. J., J. Mar. Res., 1952, 11, 172. Silvey, W. D., U.S. Geol. Surv. Water Supply Pap., 1967, No. 1535-L. Corraine, A., Bull. Sew. Caste Geol., 1967, 20, 257. Emelyanov, E. M., Geokhimiya, 1974, 4, 610. Vanderstappen, M., and Grieken, R. V., Fresenzus 2. Anal. Chem., 1976, 282, 25. Riley, J. P., and Skirrow, G., Editors, “Chemical Oceanography,” Second Edition, Volume 1, Academic Press, London, 1975, p. 462. Riley, J. P., “Introduction to Marine Chemistry,” Academic Press, London, 1971, p. 65. Keen, N . J., J . BY. Nucl. Energy SOL, 1968, 178, 7. Marczenko, Z., and Mojski, M., Anal. Chim. Ada, 1971, 54, 469. Watanabe, H., Berman, S., and Russell, D. C., Talanta, 1972, 1363. Holynska, B., and Bisinilk, K., J. Radioanal. Chem., 1976, 31, 159. Kessler, J. E., and Hitchell, J. W., Anal. Chem., 1978, 50, 1645. Fishman, M. J., and Erdmann, D. E., Anal. Chem., 1979, 51, 317R. Tananaiko, M. M., and Nebylitskaya, S. L., Zavod. Lab., 1962, 28, 263. Babko, A. K., Tananaiko, M. M., and Lozovik, A. S., Zh. Neorg. Khim., 1969, 14, 1618. Hughes, R. C., Murau, P. C., and Gundersen, G., Anal. Chem., 1971, 43, 696. ZAtka, V., and Hoffmann, O., Analyst, 1970, 95, 200. Van de Velde, G. M. H., Harkema, S., and Gellings, P. J., Inorg. Nucl. Chem. Lett., 1973, 9, 1169. Marczenko, Z., “Spectrophotometric Determination of Elements,” Ellis Honvood, Chichester, 1976, Rattenetti, A., Nut. Bur. Stand. (U.S.) Spec. Publ., No. 422, Volume 1, p. 633. Moody, J. R., Anal. Chem., 1977, 49, 2264. Wong, C. S., Cretney, W. J., Piuze, J., and Christensen, P., Nut. Bur. Stand. (U.S.) Spec. Publ., Received September 8th, 1980 Accepted November 1 lth, 1980 p. 560. No. 464, p. 249.

ISSN:0003-2654    

DOI:10.1039/AN9810600385

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, April, 1981, Vol. 106, $9. 389-393 Determination of Residues of Furalaxyl and 389 Metalaxyl in Nutrient Solution, Peat Compost and Soil Samples by Gas Chromatography David J. Caverly and John Unwin Ministry of Agriculture, Fisheries and Food, Agricultural Development and Advisory Service, Olantigh Road, Ashford, Kent, TN25 5EL Shefield City Polytechnic, Pond Street, Shefield, S1 1 WB Rapid and sensitive techniques for the determination of residues of the fungicides furalaxyl and metalaxyl in agricultural samples are described. Air-dried soils and peat composts are extracted with acetone in a Soxhlet apparatus; the extract is applied directly to the gas chromatograph and detected using a nitrogen-selective detector. Plants are macerated with acetone and, after filtration and dilution with water, partitioned with chloro- form ; nutrient solutions are similarly extracted with chloroform.The extracts are subjected to gas chroniatography after removal of chloroform and dissolution in acetone. Recoveries are generally better than SOY0 with detection limits of 0.5 mg kg-1 for peat, 0.1 mg kg-l for soils and plants and 0.02 mg kg-1 for nutrient solutions. Keywords : Furalaxyl determination ; metalaxyl determination ; fungicides ; agricultural samples ; gas chromatography Furalax yl* [methyl N - (2,6-dimet hylphenyl) -N-( 2-furo yl) alaninate] (I) and metalaxyl* [methyl N-(2,6-dimethylphenyl)-N-(2-methoxyacetyl)alaninate] (11) are new fungicides with residual and systemic activity against fungi of the order Peronosporales, which attack a wide range of commercial crops.Maier- Bode and Riedman2 made a study of the nitrogen-selective detector and its application to the detection of nitrogen-cont aining pesticides. Over-application of chemical can lead to crop damage.l I I\ The procedures described here are sensitive and rapid and can be applied to a wide range of agricultural materials. Soils and peat composts are dried and extracted with acetone in a Soxhlet apparatus, and the extract is analysed, without clean-up, by gas chromatography using a nitrogen-selective detector. Plant samples are chopped and stored deep-frozen; they are macerated with acetone and, after filtration and dilution of the extract with water, the fungicides are extracted into chloroform. The solvent is removed and the residue dissolved in acetone for gas chromatography.Nutrient solutions are extracted with chloroform, which is then treated as previously described for plant samples. Experimental Reagents Blank determina- tions should be carried out frequently and interferents in solvents removed by redistillation. Analytical-reagent grade materials should be used whenever possible. * Proposed common name. Crown Copyright.390 CAVERLY AND UNWIN : RESIDUES OF FURALAXYL AND Autalyst, VoZ. 106 The pesticides used were the 25% formulated products, pure specimens being prepared by Acetone. Chloroform. Sodium sulphate solution, 5% mlV. Fungicide stock solutions. Fungicide working standard solution. Soxhlet extraction with acetone and recrystallisation from ethanol. Dissolve 20 mg of fungicide in acetone and dilute to 200 ml with Dilute 10 ml of stock solution to 100 ml with acetone.acetone. This solution contains 10 pg ml-l of fungicide. Apparatus A Pye, Series 104, gas chromatograph equipped with a rubidium chloride thermionic detector and a glass column (0.9 x 4 mm i.d.) packed with a 5% high-vacuum silicone grease on 80- 100-mesh Gas-Chrom Q was used [stationary phase as supplied by ICI (grade M494) or BDH Chemicals]. The carrier gas was nitrogen at a flow-rate of 40 ml min-l. The flow-rate of hydrogen was 40 ml min-1 and that of air was approximately 250 ml min-l. The column temperature was 190 "C for metalaxyl and 210 "C for furalaxyl. An almost linear response of peak height and amount of fungicide applied was obtained with 30 and 20 ng of furalaxyl and metalaxyl, respectively, giving full-scale deflect ion.This apparatus consisted of an electrically heated series of mantles, Soxhlet extractors with thimbles (28 x 120 mm) and 150-ml flat-bottomed flasks. Extraction apparatus. Separating funnels. Homogeniser. Capacity 100 and 250 ml. "ATO-MIX MSE" emulsifier with a 1-1 stainless-steel container, obtainable from Scientific Supplies Ltd., London. Procedure for Soils and Compost To 25 g of soil or 10 g of peat compost, dried at room temperature and ground to pass a 2-mm mesh sieve, add 2 and 4 ml of water, respectively, mix and allow to stand for 1 h. Trans- fer the mixture into an extraction thimble, measure approximately 100ml of acetone into the flask, assemble the Soxhlet extraction apparatus and extract the sample for 8 h at a siphon rate of approximately six changes an hour.Adjust the volume of solvent and take aliquots (2-8 pl) for gas chromatography. Procedure for Plant Samples Transfer 50g of a finely chopped sample into the homogeniser container and blend with 100 ml of acetone for 3 min using a loosely fitting cover. Filter using suction and wash the container and filter with 50ml of acetone. Dilute the extract to 250ml with water and transfer 50 ml into a 250-ml separating funnel containing 50 ml of water, 10 ml of 5% m/V sodium sulphate solution and 20 ml of chloroform. Shake vigorously for 30 s and, after separation of the phases, transfer the lower layer into a small conical flask containing approxi- mately 5 g of anhydrous sodium sulphate and allow to stand for 10min with occasional swirling.Repeat this extrac- tion of the aqueous phase with a further 20 ml of chloroform using the same sodium sulphate to dry the extract. After decanting into the first extract, wash the sodium sulphate twice with 5 ml of chloroform. Evaporate the combined chloroform extracts to dryness using a water-bath at 80 O C , exercising much care in the final stages with the application of slight heat. In order to remove the last traces of chloroform, dissolve the residue in a few millilitres of acetone and again evaporate to dryness; repeat the addition and removal of acetone twice more. Dissolve the residue in a convenient volume of acetone for gas chromato- Decant the dried extract into a 150-ml flat-bottomed flask.graphy. Procedure for Nutrient Solution Add 10 ml of 5% m/V sodium sulphate solution, shake vigorously with 20 ml of chloroform for 30 s and follow the above procedure for plant samples from ". . . and, after separation of the phases. . . ." The sample to solvent ratio may be 25: 1 for nutrient solution, 5 : 1 for plants and 1 : 1 for peats and soils. Transfer 50 ml of sample into a 100-ml separating funnel.April, 1981 METALAXYL BY GAS CHROMATOGRAPHY 391 ( a ) A Time + Fig. 1. Gas chromatograms of (a) stan- dard containing 8 ng of furalaxyl (A); ( b ) extract from peat compost not treated with furalaxyl; and (c) extract of peat in (b) after addition of furalaxyl a t 20 mg kg-' (2 pl of 1 g per 5 ml). Column temperature 210 "C. Detector temperature 250 "C.Attenuation 50 x 1. Time + 1 II 1 Time __+ Gas chromatograms of (a) standard containing 6 ng of metalaxyl (B) ; ( b ) extract from soil A not treated with metalaxyl; and (c) extract of soil A after addition of metalaxyl a t 20 mg kg-l(6 pl of 1 g per 20 ml) . Column temperature 190 "C. Detector tempera- ture 250 "C. Attenuation 50 x 1. Fig. 2. Fig. 3. Gas chromatograms of (a) standard containing 8 ng of furalaxyl (A) ; ( b ) extract from soil B not treated with furalaxyl; and (c) extract of soil B after addition of furalaxyl at 10 mg kg-l (4 p1 of 1 g per 5 ml). Column temperature 205 "C. Detector temperature 250 "C. Attenuation 50 x 1.392 CAVERLY AND UNWIN: RESIDUES OF FURALAXYL AND Analyst, VoZ. 106 Time ---+ Fig. 4. Gas chromatograms of (a) standard containing 8 ng of furalaxyl (A); ( b ) standard containing 8 ng of metalaxyl (B); (c) extract of nutrient solution containing neither fungicide ; and ( d ) extract of nutrient solution in (c) after addition of furalaxyl and metalaxyl at 5 mg kg-l(2 p1 of 4 ml of nutrient solution in 5 ml of acetone).Column temperature 205 "C. Detector temperature 250 "C. Attenuation 50 x 1. (a) It I Time --b Fig. 5. Gas chromatogram of (a) standard containing 4 ng of furalaxyl (A) and metalaxyl (B) ; (b) extract (2 pl of 1 g ml-l) of field-treated lettuce containing 0.3 mg kg-' of metalaxyl; and (c) extract as in (b) after addition of furalaxyl and metalaxyl a t 4 mg kg-l (2 p1 of 1 g per 2 ml). Column temperature 210 "C. Detector temperature 250 "C. Attenuation 50 x 1.April, 1981 METALAXYL BY GAS CHROMATOGRAPHY 393 Results and Discussion The recovery of the two fungicides from peat composts and soils was measured by the addi- tion of stock solutions to the air-dried materials and, after thorough mixing and standing for 24 h, extraction with acetone as detailed above; additions to plant samples were made prior to maceration with acetone and, with nutrient solutions, as a direct addition to the aqueous phase.The mean recoveries of duplicate determinations are given in Table I and are illus- trated in Figs. 1-5. TABLE I RECOVERY OF FUNGICIDES ADDED TO SAMPLES Furalaxyl Me talax yl / A \ I A \ Sample Peat compost . . .. Soil A (calcareous silty loam; pH 7.6; organic matter 3.57%) . . . . . . Soil B (sandy loam; pH 5.7; organic matter 3.82%) .. Nutrient solution . . . . Lettuce . . . . . . Hops . . . . .. .. Added/ mg kg-' 200.00 100.00 50.00 20.00 10.00 5.00 200.00 100.00 50.00 20.00 10.00 5.00 200.00 100.00 50.00 20.00 5.00 20.00 10.00 5.00 1 .oo 0.50 0.20 0.10 20.00 4.00 2.00 0.50 - Found/ mg kg-' 190.00 105.60 44.00 17.00 7.50 4.00 200.00 111.00 49.30 23.00 7.50 4.30 147.00 80.00 39.00 16.00 4.50 19.20 9.40 4.80 0.89 0.42 0.19 0.11 19.40 4.20 2.00 0.49 - Recovery, % 95 106 88 85 75 80 100 111 99 115 75 86 74 80 78 80 90 96 94 96 89 84 95 110 97 105 100 98 Added/ rng kg-I 400.00 200.00 100.00 40.00 20.00 - 40.00 20.00 10.00 5.00 1.00 - 20.00 10.00 5.00 1.00 20.00 10.00 5.00 1 .oo 0.50 0.20 0.10 20.00 4.00 2.00 0.50 5.00 - Found/ mg kg-1 313.00 200.00 88.00 40.00 21.00 - 30.00 16.00 8.30 4.00 1.10 - 17.00 7.70 4.50 0.70 - 19.30 9.60 4.90 0.96 0.44 0.16 0.08 19.60 4.10 1.80 0.48 4.90 Recovery, % 78 100 88 100 105 - 75 80 83 80 110 - 85 77 90 70 97 96 98 96 88 80 80 98 103 90 96 98 - Conclusions The procedures described are rapid and easy to operate and, although they contain no time- consuming clean-up steps, are free from interferences. Recovery of added fungicides is generally better than SOY0 with detection limits of 0.5 mg kg-l for peat, 0.01 mg kg-l for soils and plants and 0.02 mg k g l for nutrient solutions. The staff of the first author's laboratory, especially Mrs. A. Dray, are thanked for assistance in the preparation of the analytical data and of samples for analysis. References 1. 2. Norman, R., Grower, 1978, October, 658. Maier-Bode, H., and Riedmann, M., Residue Rev., 1975, 54, 113 Received February Ist, 1979 Accepted October 21st, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600389

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

394 Analyst, April, 1981, Vol. 106, pp. 394-402 High-performance Liquid Chromatographic Determination of Four Biogenic Amines in Chocolate W. Jeffrey Hurst and Paul B. Toomey Hershey Foods Corporation, Technical Center, Hershey, Pa. 17033, USA Some biogenic amines occur in a wide variety of foods including cheese, fish, bakery products, milk products and chocolate. This study was undertaken to analyse and quantify four of the biogenic amines thought to occur in chocolate. Tyramine, tryptamine, 2-phenylethylamine and serotonin (5- hydroxytryptamine) were chosen as the amines of interest. Two high- performance liquid chromatographic (HPLC) systems were used for the final analysis of amine extracts. Both systems employed dual detection, with the first using ultraviolet absorbance a t 254nm and the formation of a post- column o-phthaldehyde derivative. The second method used ultraviolet absorbance at 254 nm and the natural fluorescence of the four amines.Thin- layer chromatography (TLC) was performed on all of the extracts to provide further confirmation. All four amines were detected and quantified a t varying levels in extracts of several kinds of chocolate and chocolate liquors. Keywords : High-performance liquid chromatography ; biogenic amines ; chocolate ; food Tyramine, tryptamine, 2-phenylethylamine and serotonin are members of the pressor amine group and tend to cause a rise in blood pressure.lS2 Many foods contain amines that are members of this pressor amine group including tomatoes3 and other fruits, as well as fish, meat and poultry produ~ts.4~~~6~7 As early as 1919 Pognic et al.suggested that allergies could cause migraines and especially implicated chocolate and the amines found in ch~colate.~ Recent literature contains conflicting information about the presence or absence of these various amines in choc~late.~-~~ This study was undertaken to determine the levels of four of these amines in various chocolate products and cocoa liquors. Samples were de-fatted and extracted using recog- nised procedures for biogenic The final determination was carried out on one of two HPLC systems. Both used reversed- phase HPLC with dual detection and ultraviolet absorbance at 254 nm as one of their detec- tion modes. The systems differed in that one used the formation of a post-column o- phthaldehyde derivativela of the biogenic amines while the other used natural fluorescence. These two analytical systems gave comparable results ; recovery and precision studies of samples and standards showed the methods to be satisfactory.Results using both HPLC systems are presented for several chocolate and cocoa liquor samples. Experiment a1 HPLC System 1 This was a modular system consisting of an M6000A solvent delivery system, Model 440 absorbance detector at 254 nm and U6K universal injector (Waters Associates). The column was Bondapak Reversed Phase (Waters Associates). The flow-rate was 1.0 ml min-l. The fluorescence detector was a Gilson, Spectra Glo Filter Fluorimeter equipped with excitation filters (340 nm) and emission filters (418 nm) (Gilson Medical Electronics).The post-column reaction apparatus consisted of a Milton Roy Mini-Pump (Laboratory Data Control) used to pump the o-phthaldehyde solution into a mixing chamber. The pump was equipped with a home-made pulse damper in the form of a standing air column. The pump effluent was passed into one port of a three-port mixing chamber at a flow-rate of 1.0 ml min-l. After mixing, the o-phthaldehyde - amine complex was carried into a reaction coil consisting of 10 m of 0.009-in i.d. stainless-steel tubing kept at a constant temperature of 40 "C by immersion in a circulating water-bath.HURST AND TOOMEY 395 HPLC System 2 This was a modular system using the same solvent delivery system, ultraviolet detector, injector and HPLC column as the HPLC System 1.The fluorescence detector used was a Varian SF-330 spectrofluorimeter equipped with an HPLC flow cell [excitation wavelength = 285 nm, emission wavelength = 320 nm (cut-off filter)]. The flow-rate was 1.3 ml min-l. Reagents Acetic acid (0.2 M) at pH 2.8 in HPLC water, the latter was prepared by passing distilled water through ion-exchange and organic absorber cartridges, and finally through a 0.1-pm filter. An 80 + 20 (V/V) mixture of 0.2 M potassium dihydrogen phosphate solution adjusted to pH 3.7 with orthophosphoric acid and methanol. This mixture was prepared, de-gassed and passed through a 0.1-pm final system filter. T L C developing solvent. The solvent system was chloroform - methanol - concentrated ammonia solution (28% m/V), 12 + 7 + 1. Boric acid bufer.Boric acid solution (0.4 M) adjusted to pH 10.3 & 0.2 with solid potassium hydroxide. o-Phthaldehyde reagent. This reagent was prepared by dissolving 0.32 g of o-phthaldehyde in 100 ml of ethanol and diluting to 1 1 with boric acid buffer. Ninhydrin solution. A mixture of 0.300 g of ninhydrin, 100 ml of butan-1-01 and 3 ml of acetic acid. Standard solutions. Tyramine and trypt amine (Calbiochem-Behring) ; 2-phenylet hylamine and serotonin (Sigma Chemical Co.). The 2-phenylethylamine was re-distilled prior to use. All standards were prepared in HPLC mobile phase to a final concentration of 0.1 pg p1-l. Chocolate samples were of nationally distributed types while chocolate liquor samples were obtained from the Hershey Chocolate Company. HPLC mobile phase 1. HPLC mobile phase 2.SampZes. Apparatus Centrifuge. Oven. Thelco Blue M. Sorvall Omni-Mixer. DuPont Instruments. Dual pen recorder. T L C developing tank. Silica gel plates. Capable of 2000 rev min-l. 5-60 (0.25 mm) from EM Labs. Extraction All samples were de-fatted with petroleum spirit (boiling range 36-60 "C) prior to extraction. One gram of de-fatted chocolate or cocoa liquor is mixed with 20 ml of 0.1 N perchloric acid using a Sorvall Omni-Mixer at setting 7 to 10 min. Alternatively this extraction could be carried out by using a wrist action shaker for 45 min. The homogenate is transferred into a centrifuge tube and centrifuged at 2000 rev min-l for 10 min. The supernatant liquid is adjusted to pH 10.3 & 0.1 with concentrated ammonia solution and then stored overnight in a refrigerator at -4 "C.The resulting solution is filtered through a Whatman No. 41 filter- paper or its equivalent. The filtrate is saturated with solid sodium chloride and then extracted four times with 5 ml of an ethyl acetate - acetone (2 + 1) mixture. After each extraction the mixture is briefly centrifuged to help separate the two layers. The first three extractions are transferred into a clean test-tube using a Pasteur pipette. The fourth extraction is filtered through Whatman IPS phase-separating paper. The organic extracts are combined and dried with anhydrous sodium sulphate. After decanting, the sodium sulphate is washed with an additional 2 ml of the ethyl acetate - acetone (2 + 1) mixture. The water-free extracts are evaporated to dryness under nitrogen a t 20 "C and then dissolved in I ml of the HPLC mobile phase.The procedure of Kissinger16 was modified for the extraction of the four compounds.396 HURST AND TOOMEY: HPLC DETERMINATION OF Analyst, VoZ. 106 Analysis The use of two HPLC systems was a result of the arrival of new equipment that would allow direct measurement of the amines by natural ff~orescence~~ ; earlier studies used the post-column o-phthaldehyde derivative formation HPLC method. Inject 20 p1 of the extract and compare with injections of standards. Calculate the concentration in the extract by comparison of the peak heights of the sample and standards. Figs. 1-3 show chromato- grams of standards using ultraviolet, post-column derivatisation and natural fluorescence detection, respectively.Figs. 4-6 depict chromatograms of extract using the same methods of detection. D B I 1 I 0 5 10 1 Timehin E 20 Fig. 1. Chromatogram of standards using ultra- violet detection. A is the injection point. Sample contained 20 pg each of : B, tyramine ; C, serotonin ; D, 2-phenylethylamine ; E, tryptamine. Column, pBondapak C18. Mobile phase, 0.2 M acetic acid. Detector, Waters Associates, Model 440, at 254 nm, 0.02 a.u.f.s. Thin-layer chromatography. Spot 10 pl of each of the extracts and standards on to Si-60 TLC plates. Develop to 12-14 cm with chloroform - methanol - concentrated ammonia solution (28%) (12 + 7 + 1). Air dry the plates, then spray with ninhydrin solution and heat at 110 "C for 10 min to reveal the spots. Compare the R, values obtained for the standards with those of the extracts.Results Recovery studies were conducted on the matrices of cocoa liquor and milk using the o-phthaldehyde post-column reaction and subsequent determination. Spiking for all recoveries was performed by the addition of standard to the perchloric acid extract. Additional recovery studies were carried out on the milk chocolate matrix using native fluorescence detection. The four amines were added at four different levels to the cocoa liquor, three different levels to the whole milk and three different levels to the milk chocolate. Tables 1-111 show the averages of duplicate determinations. Tables 1-111 show good method accuracy using either the o-phthaldehyde derivative or natural fluorescence detection methods. Table IV shows the results for five 1-g samples of de-fatted milk chocolate assayed in duplicate using natural fluorescence detection.April, 1981 BIOGENIC AMINES IN CHOCOLATE 397 D 0 5 10 15 20 Ti me/min Fig.2. Chromatogram of standards after post- column formation of o-phthaldehyde derivatives using fluorimetric detection. A is the injection point. Sample contained 2 pg each of: B, tyramine; C, serotonin ; D, 2-phenylethylamine ; E, tryptamine. Column, pBondapak CIS. Mobile phase 0.2 M acetic acid. Detector, Gilson Spectra-glo fluori- meter. 0 5 Time/min Fig. 3. Chromatogram of standards with detection by means of natural fluores- cence. A is the injection point. Sample contained : 0.5 pg tyramine (B); 0.5 pg serotonin (C); 14 pg 2- 0.54 pg tryptamine (E). Column, pBondapak CIS.Mobile phase, 0.2 M potas- sium dihydrogen phosphate solution, pH 3.7 containing 20% of methanol by volume. Detector, Varian SF-300 with HPLC flow cell. phen ylethylamine (D) ; Tables V and VI give information pertaining to standard and sample precision studies. The ultraviolet data was included only for completeness as the ultraviolet mode was used as an additional confirmatory method and not for quantitative purposes. The lower limits of detection of four of the amines are shown in Table VII. The average biogenic amine contents of eight selected chocolate liquors are given in Table VIII and the average amine contents of five chocolate products are given in Table IX. These tables show results obtained using both the o-phthaldehyde derivative and natural fluorescence methods.Discussion The results described indicate the presence of the biogenic amines of interest in chocolate liquor and chocolate products. As shown in Tables VIII and IX, these amines occur at varying levels and in varying ratios. It was not possible to gain information about fermenta- tion patterns from the amine concentrations in the various liquor types, these levels vary as would be expected in natural products. One of the factors that complicates this problem is the tryptamine level. Tryptamine is a likely precursor of the plant growth hormone indole acetic acid.19 In an attempt to gain further information about the fermentation patterns, similar samples should be analysed during all stages of growth and fermentation. From this information it might be possible to arrive at some meaningful conclusions.HURST AND TOOMEY: HPLC DETERMINATION OF Analyst, VoZ.106 B i 0 5 10 Time/mi n I I I 5 10 15 20 Ti melmi n Fig. 4. Chromatogram ot cocoa extract Fig. 5. Chromatogram of cocoa extract after using ultraviolet detection. A is the injec- post-column formation of o-phthaldehyde tion point ; B, tyramine ; and C, serotonin. A is Column, pBondapak C18. Mobile phase, the injection point; B, tyramine; and C, 0.2 M acetic acid. Detector, Waters tryptamine. Column, pBondapak C18. Mobile Associates. Model 440. at 264 nm, 0.02 phase, 0.2 M acetic ,acid. Detector, Gilson derivatives using fluorimetric detection. a.u.f .s. Spectra-glo fluorimeter. 0 5 10 Time/m i n Fig. 6. Chromatogram of cocoa extract with detection by means of natural fluorescence.A is the injection point; B, tyramine; C, serotonin ; D, 2-phenylethylamine ; and E, tryptamine. Column, pBondapak C18. Mobile phase, 0.2 M potassium dihydrogen phos- phate solution, pH 3.7, containing 20% of methanol by volume. Detector, Varian SF 330 with HPLC flow cell.April, 1981 BIOGENIC AMINES IN CHOCOLATE TABLE I AVERAGES OF DUPLICATE RECOVERIES FROM COCOA LIQUOR USING O-PHTHALDEHYDE DERIVATISATION DETECTION Sample size = 1 g of de-fatted material in each instance, n = 2. Amount added/ Amine r g g-l Tyramine .. .. .. 10 20 50 100 Tryptamine .. .. Serotonin . . .. .. 2-Phenylethylamine . . 10 20 50 100 10 20 50 100 10 20 50 100 Amount recoveredl rg g-l Recovery, yo 9.56 96.6 18.90 94.5 46.80 93.6 93.60 93.6 Average = 94.3 9.49 94.9 18.40 92.0 45.50 91.0 88.60 88.6 Average = 91.6 8.96 89.6 19.20 96.0 48.70 97.4 95.80 95.8 Average = 94.7 8.45 84.5 17.80 89.0 44.50 88.9 92.30 92.3 Average = 88.7 399 The analytical data presented in this work are in agreement with those obtained by Kenyhercz and Kissinger who found tyramine levels of 8-11 p g g l in c0c0a.l~ The data also seem satisfactory when compared with that of Ingles et aZ.14 where tyramine was deter- mined in one sample at the 5 pg g-l level by HPLC.However, other reports have suggested that tyramine is absent from cocoa1-3 so that the literature is confusing concerning even the presence or absence of these compounds in cocoa. Examination of whole milk extracts indicates the presence of compounds with a primary amino group.This work did not indicate the presence of the four biogenic amines in whole milk extracts. TABLE I1 RECOVERY FROM WHOLE MILK USING O-PHTHALDEHYDE DERIVATISATION DETECTION Amount added/ Amount recovered/ Amine r g g-l r g g-' Recovery, yo Tyramine . . .. .. 20 18.40 92.0 50 44.90 89.8 100 96.80 96.8 Average = 92.9 Tryptamine .. .. Serotonin . . . . .. 2-Phenylethylamine . . 20 50 100 20 50 100 20 50 1 no 17.60 88.0 46.80 93.6 91.80 91.8 Average = 91.1 16.30 81.5 44.70 89.4 94.20 94.2 Average = 87.8 17.80 42.40 89.40 89.0 84.8 89.4400 HURST AND TOOMEY: HPLC DETERMINATION OF Analyst, T/d. 106 TABLE I11 AVERAGE OF DUPLICATE RECOVERIES FROM MILK CHOCOLATE USING NATURAL FLUORESCENCE DETECTION Sample size = 1 g. Amine Tyramine . . .. .. Tryptamine .. .. Serotonin . . .. . . 2-Phenylethylamine . . Amount added1 10 25 50 Pg g-l 10 25 50 10 25 50 10 25 50 Amount recovered/ Pg g-' Recovery, % 9.78 97.8 23.40 93.6 47.10 95.6 Average = 95.7 9.61 96.1 24.30 97.2 47.80 95.6 Average = 96.3 9.48 94.8 23.60 94.4 46.55 93.1 Average = 94.1 9.34 93.4 24.75 99.0 49.65 99.3 Average = 97.2 TABLE IV MULTIPLE MILK CHOCOLATE ANALYSES Tryptamine could not be detected in any of the samples. Sample number 1 2 3 4 5 Mean . . .. . . Standard deviation . . Coefficient of variation, % T yramine 12.02,12.02 11.97,12.07 12.03J1.99 12.06,11.98 11.95,12.09 12.03 0.045 0.38 Amine contentlpg g-1 2-Phenylethylamine 0.42,0.46 0.44,0.44 0.41,0.47 0.43,0.45 0.40,0.48 0.44 0.025 5.68 A v Serotonin 27.10,27.30 27.30,27.10 26.80,27.60 26.92,%7.48 27.03,27.37 27.2 0.25 0.92 TABLE V HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY PRECISION STUDIES ON STANDARDS Amount injected/ Amine Pg Tyramine .. . . .. 20 5 Tryptamine . . .. , .. 20 5 5 2-Phenylethylamine .. 20 15 5 Serotonin . . . . . . 20 5 5 n 9 5 5 9 5 5 9 5 5 9 5 5 Detection mode Ultraviolet Post-column derivatisation Fluorescence Ultraviolet Post-column derivatisation Fluorescence Ultraviolet Post-column derivatisation Fluorescence Ultraviolet Post-column derivatisation Fluorescence Coefficient of variation, yo 0.78 1.44 1.52 2.27 4.63 3.91 7.44 8.74 6.02 2.01 2.97 2.23April, 1981 BIOGENIC AMINES I N CHOCOLATE TABLE VI HIGH-PERFORMANCE LIQUID CHROMATOGRAPHY PRECISION Amount injected/ Amine Pg n Detection Tyramine . . . . .. 20 9 Ultraviolet 401 STUDIES ON SAMPLES Coefficient of mode variation, yo 2.19 5 5 Post-column derivatisation 1.53 5 5 Fluorescence 0.93 Tryptamine . ... .. 20 9 Ultraviolet 3.17 5 5 Post-column derivatisation 5.03 5 5 Fluorescence 4.46 2-Phenylethylamine 20 9 Ultraviolet 6.79 15 5 Post-column derivatisation 6.08 7 5 Fluorescence 5.58 Serotonin . . .. . . 20 9 Ultraviolet 1.98 5 5 Post-column derivatisation 2.36 5 5 Fluorescence 2.92 This study is not yet complete as other amines have been found to occur in chocolate. Kenyhercz and Kissinger,16 using HPLC with an electrochemical detector, found octopamine, metanephrine and synephrine in chocolate. These amines need to be investigated further and quantified as they are conversion products of tyramine. The other conversion products of tyramine6120,21 are methyltyramine, dopamine, hordiene, methyloctopamine and noradrenaline.These and other amines can be formed by several biochemical pathways, including amino acid decarboxylation, aldehyde amination, phospholipid decomposition and thermal amino acid composition.6 Other amines found in chocolate are6 methylamine, butylamine, di- methylamine, isobutylamine, ethylamine, isoamylamine, trimethylamine and triethylamine. TABLE VII DETECTION LIMITS FOR FOUR OF THE AMINES Amine Detection mode Lower 1imitlp.g g-I Tyramine .. .. . . Post-column derivatisation 1 .o Natural fluorescence 0.5 Tryptamine .. . . Post-column derivatisation 0.5 Natural fluorescence 0.5 Serotonin . . .. . . Post-column derivatisation 2.0 Natural fluorescence 0.25 2-Phenylethylamine .. Post-column derivatisation 9.0 Natural fluorescence 0.25 TABLE VIII AVERAGE BIOGENIC AMINE CONTENTS OF CHOCOLATE LIQUORS In each instance the sample size was 1 g and the mean of 2 determinations is shown. Amine contentlwg g-l Tyramine Serotonin Tryp tamine 2-Phenylethylamine' A -l P P I--- Post-column Natural Post-column Natural Post-column Natural Post-column Natural r Type derivatisation fluorescence derivatisation fluorescence derivatisation fluorescence derivatisation fluorescence Costa Rica .. .. 7.96 7.96 15.80 16.10 3.54 3.52 N.D.' 4.34 New Guinea . . 14.70 14.70 9.19 9.23 2.88 2.93 N.D. 6.56 Lagos . . .. 2.69 2.73 3.96 3.85 0.99 1.04 N.D. 4.38 Equador . . .. 8.37 8.42 12.40 12.21 2.41 2.38 N.D. 5.13 Light Lagos . . 2.06 0.92 0.52 N.D. 2.19 Malaysian .. . . 2.74 0.18 1.07 N.D. 3.28 Sanchez . . .. 0.73 0.79 0.71 N.D. 2.55 Caracas . . .. 1.09 3.82 2.04 N.D. 8.02 * N.D. = not detected.402 HURST AND TOOMEY TABLE IX AVERAGE BIOGENIC AMINE CONTENT OF SOME CHOCOLATE PRODUCTS In each instance sample size = 1 g and the result given is the mean of 2 determinations. Amine contentlwg g-1 I L 1 Tyramine Serotonin Tryptamine 2-Pheylethylamine P z z G r x G z P N a t u r a i -1 Post-colurnnNaturai Product type derivatisation fluorescence derivatisation fluorescence derivatisation fluorescence derivatisation fluorescence Milk chocolate A . . 11.90 12.02 26.60 27.20 N.D.* N.D. N.D. 0.44 Milk chocolate B . . 5.96 6.04 8.51 8.33 N.D. N.D. N.D. 2.13 Milk chocolate C . . 3.83 3.76 5.32 5.25 N.D. N.D. N.D. 6.60 Milk chocolate D .. 4.27 4.41 1.04 1.02 N.D. N.D. N.D. 4.40 Dark chocolate . . 11.90 12.02 8.46 8.64 N.D. N.D. N.D. 3.84 * N.D. = Not detected. The levels in chocolate are low when compared with levels found in other foods. Cheese has been reported as having tryptamine and 2-phenylethylamine levels ranging from below the limits of detection to over 4OOpgg-l. Sausage has been reported to contain levels of tyramine ranging from below the limit of detection to over 350 pg g-l, 2-phenylethylamine levels from below the limit of detection to almost 7OOpgg-1 and tryptamine levels from below the limit of detection to over 50 pg g-l. Meat and cheese are not the only foods that contain amines. This work is not a comprehensive study of the amines that occur in chocolate and more investigations need to be made; however, it provides a method for the traction, detection and quantitation of the four biogenic amines at trace levels in chocolate products. References 1.2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Sandler, M., Youdim, M. B. H., and Hanington, E., Nature (London), 1974, 250, 335. Cochraine, A. L., Editor, “Background to Migraine, 3rd Migraine Symposium,” Heinemann, London, Chaytor, J . P., Crathorne, B., and Saxby, M. J., J . Sci. Food Agric., 1975, 29, 593. Waalkes, T. P., Sjoerdsma, A., Crevelin, C. R., Weissbach, H., and Undenfriend, S., Science, 1958, Sen, N. P., J . Food Sci., 1969, 34, 22. Maga, J. A., CRC Crit. Rev. Food Sci. Nutr., 1978, 10, 373. Lovenberg, W., Editor, “Toxicants Occurring Naturally in Foods,” Second Edition, National Academy of Sciences, Washington, D.C., 1973, p. 170. Rowe, A. H., and Rowe, A. H., Jr., Editors, “Food Allergy. Its Manifestation and Control, and the Elimination Diets. A Compendium : With Important Consideration of Inhalant, Drug and Infectant Allergy,” C. C. Thomas, Springfield, Ill., 1972, pp. 324 and 584. 1970, p. 113. 127, 648. Marley, E., and Blackwell, B., Adv. Pharmacol. Chemother., 1970, 8, 185. Riggin, R. M., and Kissinger, P. T., J . Agric. Food Chem., 1976, 24, 900. Dietrich, P., Lederer, E., Winter, M., and Stoll, M., Helv. Chim. Acta, 1964, 47, 1581. Marion, J . P., Muggler-Chavan, F., Viani, R., Bricout, J., Reymond, D., and Andeglic, R. H., Helv. Weurman, C., and DeRooij, C., J . Food Sci., 1961, 26, 239. Kenyhercz, T. M., and Kissinger, P. T., Phytochemistry, 1977, 16, 1602. Ingles, D. L., Tindale, C . R., and Gallimore, D., Chem. Ind. N . Y . , 1978, 12, 432. Kenyhercz, T. M., and Kissinger, P. T., Lloydia, 1978, 41, 130. Kohler, P. C., and Eitenmiller, R. R., J . Food Sci., 1978, 43, 1245. Roth, M., Anal. Chem., 1971, 43, 880. Smith, T. A., Phytochemistry, 1977, 16, 171. Wheaton, T. A., and Stewart, I., Phytochemistry, 1969, 8, 85. Schultz, H. R., Editor, “Origin of Biologically Active Amines Found in Man,” Pergainon Press, Received May 30th, 1979 Accepted September 29th, 1980 Chim. Acta, 1967, 50, 1509. Oxford, 1969, p. 148.

ISSN:0003-2654    

DOI:10.1039/AN9810600394

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, April, 1981, Vol. 106, pp. 403-411 403 Standard Atmosphere Generator: a Dynamic System for the Controlled Dilution of Organic Vapours in Air B. I . Brookes Strathclyde Regional Council, Department of the Regional Chemist, 8 Elliot Place, Clydeway, Glasgow, G3 8E J An apparatus is described for the production of standard atmospheres of compounds a t the concentrations normally encountered in air pollution studies. I t is a glass-blown, flow system with continuous syringe injection of the compounds into a vaporiser, prior to their dilution in the main flow line. Statistical analyses of performance tests on a range of compounds show that stable atmospheres are produced within 5min of start-up. The instrument produces the predicted concentrations of non-polar compounds with a high degree of accuracy, and tests on atmospheres of a relatively involatile compound, l-methylnaphthalene, reveal no diminution in per- formance even with concentrations approaching that of its saturated vapour.Keywords Standard atmosphere ; syringe pump ; organic vapour ; air sampling For the most accurate assessment of the efficiency of air sampling equipment in collecting many varied pollutants, it is necessary to have an apparatus that can produce standard atmospheres containing the compounds at concentrations similar to those encountered in practice. The apparatus described here, a dynamic system with liquid injection, has been designed to provide standard atmospheres with a minimum of procedure, but with concentra- tions that can be accurately pre-determined, and without a requirement for lengthy periods of st abilisation.A dynamic system was chosen in preference to a static system with which inaccuracies arise because of losses to the container walls, and liquid injection was chosen as the means of intro- duction of the pollutant into the gas flow. Design features specific to this apparatus were vaporisation of the liquid before it mixed with the main air flow, and the incorporation of a smoothing system. The alternative methods of producing the vapour and controlling its dilution were dismissed because they did not meet the requirements set out above. Some systems are based on the production and dilution of a saturated vapour, but truly saturated vapours can be difficult to obtain. Other procedures depend on diffusion-controlled dilution of a vapour either through a defined ~rificel-~ or across a semi-permeable membrane,* but the operator has limited control over the precise concentration that will be produced.In these alternative systems the pro- duction of atmospheres containing several components may require separate saturators or diffusors for each vapour. Standard Atmosphere Generator The generator is illustrated in Fig. 1. I t is essentially a glass-blown construction using 7 mm i.d. tubing for the main flow line. Air is drawn through the apparatus using a Charles Austen, 4N, diaphragm pump, and the necessary flow control is achieved with the combination of a fixed leak and a TF/6/18 Rotaflo valve. The flow-rate is measured in the range 0.005- 0.025 m3 min-l on the Fisher Controls, Series 1100, Rotameter and the incoming air is purified by passage through a 500-ml sintered-glass gas wash-bottle packed with 250 ml of Linde 13X molecular sieve and 250 ml of 30-60-mesh activated charcoal.Pressure can be measured at both ends of the apparatus with a mercury manometer. The Rotameter was calibrated under its operating condition using a gas meter. The vaporiser section and injection port are constructed from a Pye, Series 104, injection head attached to a 6 mm 0.d. x 1 mm i.d. x 50 mm long glass tube, and wound with a 120-W heating cord connected to a variable transformer power supply. Helium carrier gas at 40 ml min-l transports vapours into the main flow line. The organic compounds are injected404 BROOKES : STANDARD ATMOSPHERE GENERATOR CONTROLLED Analyst, VOl.106 using a SAGE, Model 341, syringe pump fitted with SGE or Hamilton gas-tight syringes. The latter has 0.1 mm bore needles and the former 0.2 mm bore needles. The syringe pump and the diaphragm pump are operated from the same power supply so that they can be switched on or off at the same time. t P Fig. 1. Standard atmosphere gen- erator: A, air intake filter; B, Rotameter gauges; C, syringe and syringe pump; D, vaporisor; E, 5-1 mixing bulb; F, sampling port; G, fixed leak; H, main air flow pump; and I, mercury mano- meter. There are two ports for the removal of standard atmospheres from the main flow. The first consists of a 6 mm 0.d. x 1 mm i.d. x 50 mm long glass tube fitted with a PTFE coupling (Chemcon Ltd.).The standard atmosphere can be withdrawn continuously from this port using an electric pump. The 1-mm constriction is small enough to permit the port to be opened during operation of the apparatus with little effect (less than 2.5% diminution) on the main flow. The second port is intended for the withdrawal of syringe samples and consists of a 40-mm length of 6 mm i.d. glass tubing terminated with a Thermogreen, half-hole type, cylindrical septum (Supelco Ltd.) . With the arrangement shown in Fig. 1, the apparatus produces atmospheres at pressures slightly less than ambient. It can be operated equally well at pressures above ambient by connecting the air pump to the inlet filter, instead of the generator outlet, so that it blows air through the apparatus.When not in use, the PTFE union is plugged with a 6 mm 0.d. glass rod. Principles of Operation dilutionlmixing and smoothing. The operation of the apparatus can be divided into four functions : injection, vaporisation, Injection The syringe pump moves the syringe plunger by a series of pulses, each corresponding to a movement through 6.9 x mm. The flow-rate can be varied either by changing the pulse rate (from a minimum of 1.3 up to 3000 min-l) or by varying the size of the syringe (from 25 pl to 50 ml). Vaporisat ion When the temperature of the vaporiser is just below the boiling-point of the injected liquid, vaporisation will take place rapidly at the tip of needle, or on the walls of the chamber if the syringe flow-rate is very high. At temperatures above the boiling-point of the liquid, vaporisa-April, 1981 DILUTION OF ORGANIC VAPOURS I N AIR 405 tion may occur within the bore of the needle.This second mode of operation is necessary to ensure the vaporisation of very high boiling components when a mixture of compounds is injected. Fractionation during vaporisation, with the less volatile species tending to remain in the liquid phase and the gas phase concentration correspondingly less than predicted, is not a major problem as it is self-compensatory with time, there being rapid enrichment of such com- pounds in the immediate vicinity of the liquid front, and their rates of.vaporisation increasing to equal their rate of injection. Dilutionlmixing The efflux velocity from the narrow bore of the vaporiser is 0.1 m s-l (for a carrier gas flow-rate of 40 ml min-l) and the air velocity in the main flow line falls in the range 0.2-1.1 m s-l. The atmosphere subsequently passes through Drechsel heads fitted into round-bottomed flasks and the velocity of the incoming air induces rapid mixing.A simple T-junction suffices as the preliminary mixing stage. Smoothing The mixing process could occur as adequately in small flasks as in the two 5-1 flasks chosen for this apparatus. However, these large flasks have the advantage of greatly increasing the volume of the mixing system and enable it to smooth any irregularities in concentration intro- duced during injection and vaporisation. At its slowest rate the syringe pump produces pulses with a time interval of 0.78 min and this is comparable to the retention times for the 10-1 mixing volume, which are in the range 0.4-2.0 min.Procedures for Using the Generator Purging Switch on the vaporiser heater to give a temperature of up to 250 OC, switch on the diaphragm pump to give a main flow-rate of 0.025 m3 min-l, and switch on the carrier gas flow to give a flow-rate of 40 ml min-l. Leave the apparatus in this condition for about 10 min to purge it of previously adsorbed material. Conditioning Fill the syringe with the liquid to be injected, ensuring there are no entrained air bubbles. Switch on the pumps (pre-set to produce the required atmosphere) and insert the needle 30 mm tlirough the injection head. Leave the system running for at least 10 min before using the standard atmosphere. To change the concentration when the apparatus is already running, make the appropriate adjustment to the syringe flow-rate or the main flow-rate and permit at least 10 min for the system to stabilise.The predicted concentration for the standard atmosphere, C,, is given by the expression C, = fc mg m-3 where f p l min-1 is the flow-rate from the syringe, c mg $1 (or g ml-l) is the concentration in the liquid and Fm3min-1 is the main flow-rate. For pure liquids, c is equal to the density in grams per millilitre. Note: In the tables of results that follow, the measured concentrations are referred to by the symbol C,. Test Procedure and Apparatus Chemicals Halothane (I,], 1 -trifluoro-2-chloro-2-bromoethane) was supplied by ICI. cals were of analytical-reagent grade. Nitrobenzene-d, (perdeuteronitrobenzene) was obtained from Koch-Light Laboratories Ltd.All other chemi-406 BROOKES : STANDARD ATMOSPHERE GENERATOR: CONTROLLED Analyst, VoZ. 106 Sample Tubes The sample tubes (see Fig. 2) had a 12.5 mm 0.d. x 57 mm long stahless-steel tube welded at each end to short tubes, 6.3 mm 0.d. x 12.5 mm long (G. N. Instrumentation Consultancy Ltd.). The body of the tube was packed with 1 g of 60-80-mesh Tenax GC and plugged with silanised glass-wool. The sorption characteristics of these tubes have been described previou~ly.~,~ C 12.5 m 12.5 m m 4 1-57 rnm-4 Fig, 2. Adsorption tube: A, 60-80-mesh Tenax GC; B, stainless-steel body; and C, silanised glass-wool. Sampling and Analysis Three methods of sampling were used. Atmospheres containing halothane were drawn through a Carlo Erba gas sampling valve via a 1.6 mm 0.d. PTFE tube.The sample loop volume was 0.5 ml. Suction was provided by a Charles Austen Dymax 2A diaphragm pump, regulated by a needle valve, and the pressure in the sampling line was monitored on a mercury manometer. Direct injection was made on to a 2 m x 3 mm i.d. column packed with 5% OV-101 on 80-100-mesh Supelcoport, and fitted to a Carlo Erba flame-ionisation detector. The carrier gas flow-rate was 15 ml min-l(101 kN m-2, 20 "C) and the column temperature was constant at 60 "C. For comparison, 1-p1 portions of standard solutions of halothane in carbon tetrachloride were injected through the septum injection port of the gas chromatograph. A t least five injections of each solution were made.Some atmospheres were sampled on adsortion tubes by connecting the tube directly to the PTFE union and drawing off the standard atmosphere at 100-200 ml min-l using an electric pump for large volumes, or a gas-tight syringe for volumes of less than 500 ml. Before sampling, the tubes were loaded with the appropriate internal standard (2 pg of anisole for hydrocarbon analysis, or 0.1 pg of nitrobenzene-d, for nitrobenzene-do analysis) according to the method of Brookes6 The samples were heat desorbed into a liquid nitrogen cooled trap using a GN concentrator (G. N. Instrumentation Consultancy Ltd.) and flash desorbed on to a 30 m x 0.25 mm i.d. SE-30 WCOT gas-chromatographic column. For the hydrocarbon analysis this was programmed from 20 to 110 "C at 4 "C min-l and the peaks were monitored using a VG Organic MM16F mass spectrometer in its integrated ion mode of detection.The same column was programmed from 60 to 100 "C at 5 "C min-l for nitrobenzene analysis using the mass spectrometer in its selective ion mode to monitor the nitrobenzene-do and -d, parent ions. Calibration graphs for these analyses were obtained by making injections of known amounts of the compounds of interest on to adsorption tubes, together with the appropriate internal standard, and analysing the tubes in the above manner. All of the analyses were carried out within 36 h of the preparation of the standard solutions and standard atmospheres. For continuous monitoring a 3.1 mm 0.d. PTFE tube was connected between the PTFE union and a Carlo Erba flame-ionisation detector.With the air pump blowing air through the apparatus at a rate of 0.025 m3 min-l, an exit pressure of 0.9 kN mb2 above ambient was obtained, and this caused a small portion (30 ml min-l) of the standard atmosphere to flow directly into the detector. The output was recorded on a chart recorder and data for statistical analysis were obtained by measuring the response at regular 0.5- or l-min intervals. This was a variation of the method of MacDonald and MacKenzie.7April, 1981 DILUTION OF ORGANIC VAPOURS IN AIR 407 Test Results All of the work reported here was taken from data obtained when producing standards with concentrations in the range from ambient to the Threshold Limit Value8 or sometimes in the region of the Odour Threshold Value.g The time required for the apparatus to produce a stable atmosphere of a volatile compound was determined whilst monitoring a halothane atmosphere continuously for 60 min. Details of the statistical analysis of the detector response measured at regular l-min intervals are given in Table I.The two values for the mean responses measured for the periods 5-20 min and 45-60 min after start-up showed no difference when compared in an unpaired t-test at the 0.05 level of significance, indicating that stabilisation had effectively occurred within 5 min of start-up. TABLE I STATISTICAL ANALYSIS OF HALOTHANE, 1-METHYLNAPHTHALENE AND NITROBENZENE ATMOSPHERES, SHOWING STABILISATION Compound * Halothane . . 1 -Methylnaphthalene Nitrobenzene . . .. Predicted concentration/ mgm-5t 1 77 77 15.2 15.2 61.1 61.1 18.0 18.0 36.5 36.5 Air flow- rate/ m3 min-lt 0.025 9 0.025 9 0.025 9 0.025 9 0.0259 0.025 9 0.0259 0.025 9 0.025 9 0.0259 Syringe flow-rate/ p1 min-' 1.07 1.07 0.386 0.386 1.55 1.55 0.386 0.386 0.785 0.785 Period of data analysis/ min (after start-up) 5-20 45-60 5-20 45-60 5-20 45-60 5-20 29-44 5-20 20-35 Number of data 16 16 16 16 16 16 16 16 16 16 Mean response f S.E./mm 91.7f0.3 92.5f 1.5 100.2&0.8 101.0f0.5 438 f 2 37.3f0.8 37.5f0.7 211f2 209 f 2 435f 3 * The pure compounds were injected with the following vaporiser temperatures : halothane, 100 "C ; t At 102.2 kN m-2 and 20 "C.l-methylnaphthalene, 220 "C ; and nitrobenzene, 190 "C. The accuracy with which the apparatus produced atmospheres at pre-determined concentra- tions was measured for a succession of standard atmospheres of halothane in the range 20- 280 mg m-3 (20 "C and 1 0 1 .3 kN m-2) by varying the syringe flow-rate and holding the main flow-rate constant. Halothane peak heights resulting from the gas-chromatographic analysis of the standards are plotted with respect to time in Fig. 3, which also contains the procedural information. Each sample represented the mean concentration produced by the generator over a 0.5-s period. On the two occasions when a freshly charged injector syringe was inserted into the vaporiser, the halothane con- centration exceeded the steady-state level. This phenomenon occurs as a result of the rapid vaporisation of halothane in the dead space of the needle when it is inserted.The mean values of the peak heights for each steady state are plotted in Fig. 4 with respect to the amount of halothane injected on to the gas-chromatographic column, the amount injected being calculated from the predicted concentration, which is also shown on the hori- zontal co-ordinate. In addition, Fig. 4 shows the results for the liquid standards, which were injected directly on to the gas-chromatographic column and analysed in the same way as the gas standards. For each set of data the line of regression, the standard errors (S.E.) of the constants and the correlation coefficients (R) were calculated. Describing the line of regression by an equation of the form The syringe was filled with pure halothane. y = a x + b where y represents peak height and x the amount of halothane injected, the data given in Table I1 were calculated.408 BROOKES : STANDARD ATMOSPHERE GENERATOR : CONTROLLED Analyst, VoZ.106 TABLE I1 STATISTICAL ANALYSIS OF HALOTHANE DATA FOR THE LINE OF REGRESSION^ = ax + b a f S.E./ b -+ S.E.1 Number of Correlation Standards injected mm ng-1 mm x lo2 measurements coefficient, R Gas standards . . .. 4.46 f 0.05 +0.13 f 0.12 41 0.998 Liquid standards . . .. 4.66 f 0.11 -0.27 -J-- 0.32 32 0.992 An unpaired t-test on the values of a and b for the two sets of data showed no significant difference a t the 0.05 level of significance. Tests were also carried out on a relatively involatile compound. Standard atmospheres of 1-methylnaphthalene were produced by varying the rate of injection of the pure liquid and holding the main flow-rate constant.Sampling via a gas sampling valve or a glass syringe 25 20 E E N. 15 X E .g 10 1 0 50 100 Time/m i n Fig. 3. Stabilisation of halothane atmos- pheres with respect to time for a series of different flow settings. S , Syringe flow-rate; P, syringe pulse-rate; C,, predicted concentration (101.3 kN m-2 and 20 "C). At points x on the graph a freshly charged syringe was inserted into the vaporiser. Vaporiser temperature, 100 "C (boiling-point of halothane, 50 "C) ; main flow-rate, 0.025 m3 min-l; sampling period, 0.5 s. E 25 E . 20 15 E '6 10 F Dl x Y a 3 5 0 50 100 150 Amount injectedlng 0 50 150 250 Concentration of gas standard/mg m-3 Fig. 4. Mean gas-chromatographic GC responses (& 2 x standard error) for the injection of halothane gas stan- dards ( 0, A) and liquid standards ( A, B) and plotted against the amount of halothane injected.For gas standards this was calculated from the predicted concentration, which is also shown on the graph. In all but two instances the error bars were smaller than the size of the point. proved to be unsuitable. The former method gave gas-chromatographic peaks with extensive tailing and the latter evidenced losses of the compound to the walls of the syringe. Continuous monitoring was the only satisfactory procedure available. The results and procedural informa- tion for the stabilisation studies are given in Table I. In each of the two atmospheres tested, the t-test showed no difference at the 0.05 level of significance between the mean responses measured near the beginning and at the end of each run.The amospheres were generated on different occasions and no correlation tests were carried out on the data because of variation in gas-chromatographic sensitivity. To test specifically for correlation a series of 1-methylnaphthalene atmospheres were gener- ated in the range 3.18-124 mg m-3 (102.2 kN m-2, 20 "C), again by varying the syringe flow- rate and holding the air flow-rate constant. Each atmosphere was allowed 5 min to stabilise and the detector response was measured at regular 0.5-min intervals over the subsequent 10-min period. The statistical analysis of the data is given in Table 111.April, 1981 DILUTION OF ORGANIC VAPOURS I N AIR 409 TABLE I11 STATISTICAL ANALYSIS OF 1 -METHYLNAPHTHALENE DATA FOR THE LINE OF REGRESSION y = ax + b Predicted concentration/mg m-3* 3.18 7.03 15.2 29.5 61.1 124 Air flow-rate/mS min-'* .. . . 0.0259 0.0259 0.0259 0.0259 0.0259 0.0259 Syringe flow-rate/pl min-1t . . 0.0806 0.179 0.386 0.749 1.55 3.14 Mean response S.E./mm . . 20.510.5 51.0k0.5 112*1 214&1 438&2 908&4 Total number of measurements. . Correlation coefficient . . .. 0.999 6 \ V I 126 * At 102.2 kN m-2 and 20 "C. t Pure 1-methylnaphthalene injected at 220 "C. Further confirmation of the performance of the instrument was obtained from quantitative measurements of standard atmospheres containing mixed hydrocarbons at predicted individual concentrations of 0.0535 mg m-3 (101.3 kN m-2, 20 "C), which were produced on two separate occasions by injection of a methanolic solution of the hydrocarbons.The results of the gas chromatography - mass spectrometry of adsorption tube samples of these atmospheres are given in Table IV. In order to test the apparatus for the production of atmospheres containing high-boiling polar compounds, a number of measurements were made using nitrobenzene. Table I gives the results of statistical analyses of two nitrobenzene atmospheres, which were monitored continuously using the flame-ionisation detector. In both instances the t-test demonstrated stabilisation within 5 min of start-up, but the detector response showed a greater degree of variation compared with the atmospheres of other compounds. Quantitative measurements were also carried out. The difficulties in the production of atmospheres of compounds such as nitrobenzene apply equally to their sampling and gas-chromatographic analysis.These analytical problems were minimised by sampling on adsorption tubes with nitrobenzene-do for the standard atmosphere and nitrobenzene-d, as the internal standard. The results presented in Table V were obtained with standard atmospheres produced on four separate occasions. Pure nitrobenzene-do was injected for the first three atmospheres and a methanolic solution was injected to produce the fourth. The results show a much greater degree of variation than those obtained for other compounds. Discussion All of the continuous monitoring tests reported in Table I demonstrated the stabilisation of each standard atmosphere within 5 min of start-up.With 1-methylnaphthalene the test was acutely sensitive because the degree of variation about each mean value was very small, and great care was needed to eliminate interferences from outside contamination of the air supply, as the normal variations would have considerably reduced the sensitivity of these tests. It may be that the minimal variation exhibited by 1-methylnaphthalene was a function of its comparative involatility in that it permitted a smoother vaporisation process. TABLE IV ANALYTICAL RESULTS AND OPERATING CONDITIONS FOR HYDROCARBON ATMOSPHERES Operating conditions: main flow-rate = 0.010 m3 min-l a t 99.3 kN m-2; liquid flow-rate from the 100-p1 SGE syringe = 1.05 pl min-l; syringe solution = 0.0005 mg p1-1 of each hydrocarbon in methanol; syringe pulse rate = 77 min-l; injection temperature = 100 "C; conditioning time = 15 min; sampling time = 60 min; sample volume = 0.00952 m3 at 101.3 kN m-2 and 20 "C.Trimethyl- Parameter Test o-Xylene p-Xylene benzene Naphthalene Boiling point/"C (101.3 kN m-3) . . - 144 138 165 218 Predicted concentration, (C,)/mg n ~ - ~ * Both tests 0.0535 0.0535 0.0535 0.053 5 Measured concentration, (C,) /mg m-3* First test 0.054 3 0.054 5 0.0539 0.051 3 Second test 0.0560 0.057 2 0.053 8 0.051 3 Deviation, (C, - C,)/C,, % . . First test f1.4 +1.9 f0.7 - 4.0 Second test f4.7 $6.9 f0.6 -4.0410 BROOKES : STANDARD ATMOSPHERE GENERATOR : CONTROLLED Analyst, VoZ. 106 For the purposes of the quantitative and correlative measurements on halothane, the onset of stabilisation was estimated by visual inspection of the data in Fig.3. These estimates agreed with that expected from the detailed analysis reported in Table I. The degree of variation about each mean value was very small and in Fig. 4 the 95% confidence limits for the gas standards data were too narrow to be represented. These observations, together with the high correlation and the comparability with liquid standards further confirmed that stabilisa- tion effectively took place within 5 min of start-up. TABLE V ANALYTICAL RESULTS AND OPERATING CONDITIONS FOR NITROBENZENE ATMOSPHERES Atmosphere number f A I Parameter 1 2 3 4 Main flow-rate/m3 min-'* . . .. . . . . . . 0.024 3 0.024 3 0.024 3 0.023 8 Inlet pressure/kN m-2 . . . . . . . . . . 98.6 98.6 98.6 98.6 Syringe liquid flow-ratelpl min-' .. . . . . . . 0.007 65 0.007 65 0.007 65 1.05 Nitrobenzene concentration in the syringelmg p1k1 . . 1.20: 1.20: 1.201 0.0005§ Syringe type . . . . . . . . . . . . H t Ht Ht st Syringe pulse rate/min-l . . . . . . . . . . 1.3 1.3 1.3 77 Vaporisor temperature/'C . . .. . . . . - . 190 190 190 190 Sample volume/m3* . . . . . . . . .. 0.000 1 0.000 1 0.000 1 0.002 0 Predicted concentration, C,/mg m-3* . . . . . . 0.377 0.377 0.377 0.022 Conditioning timelmin . . . . . . . . . . 60 65 90 15 Sampling time/min . . . . .. .. . . . . 0.5 0.5 0.5 15 Measured concentration, Cm/mg m-3* . . . . .. 0.372 0.45 0.46 0.018 Deviation, (Cm -Cp)/Cp, yo . . . . . . . . -1.3 f19 + 22 - 18 * At 101.3 kN m-2 and 20 "C. 7 H = 5O-pl Hamilton syringe; S = 100-p1 SGE syringe.: The density of pure nitrobenzene. 3 Solution in methanol. The correlation coefficients for halothane and methylnaphthalene indicated a highly reproducible and predictable performance in the production of atmospheres from compounds of very different volatility. Calibrations of the gas-chromatographic analyses were made with liquid standards administered to the gas chromatography and gas chromatography - mass spectrometry systems either by direct syringe injection, which allowed the rapid production of a statistically significant number of data, or via adsorption tubes with an internal standard, which permitted very accurate indi- individual measurements. The halothane measurements in Fig. 4 and Table I1 were calibrated by direct syringe injec- tion. The statistical treatment revealed high individual correlation coefficients and small standard errors, a situation in which the application of the t-test to the two sets of data was especially critical; even so, the test showed no significant difference between the gas-chroma- tographic responses to the gas and liquid standards.Adsorption tubes were used for the performance tests with mixed hydrocarbon atmospheres (Table IV) and of nitrobenzene atmospheres (Table V) . For hydrocarbons the differences between the measured concentra- tions were small and within the experimental error of the analytical procedure. In particular, naphthalene, which is relatively involatile, showed deviations of only 4% in both tests. Nitrobenzene, a compound of similar volatility to naphthalene but more polar, had much greater deviations of up to 22% in the four tests.The poor performance of the apparatus for nitro- benzene was not unexpected in view of the physical properties of this substance and the likelihood of it having irregular sorption characteristics on the glass surfaces of the instrument. An apparatus constructed from PTFE might overcome this problem. The performance of the equipment and the stability of the atmospheres might have been expected to decrease for any compound as concentrations approached that of its saturated vapour. Extrapolation of the vapour pressure data for 1-methylnaphthalene at 244 and 107 O C l 0 leads to a value of 60 mg m-3 at 20 "C, compared with the concentrations of 61 and 124 mg m-3 in Tables I and 111. The errors in this extrapolation are very large, but the The quantitative tests supported this conclusion.April, 1981 DILUTION OF ORGANIC VAPOURS I N AIR 41 1 concentrations of l-methylnaphthalene in this study were certainly close to, and may have exceeded, that of the saturated vapour without showing any loss of correlation or stability.If the concentration of the saturated vapour was exceeded, the atmosphere must have existed either in a state of supersaturation or with partial formation of an aerosol. Conclusion This apparatus was designed for the production of standard atmospheres of both pure com- pounds and mixtures, at pre-determined concentrations, with a high degree of accuracy and a minimum of procedure. For all of the compounds tested the atmospheres have been observed to stabilise within 5 min of start-up and the apparatus met all of the above requirements for non-polar compounds. For a compound that was both polar and high boiling the accuracy of the apparatus in achieving pre-determined concentrations was diminished. There was no evidence of any diminution in performance with decreasing volatility of the non-polar com- pounds or with the production of concentrations close to that of the saturated vapour pressure. The author thanks Mr. R. S. Nicolson, the Regional Chemist to Strathclyde Regional Council, for his encouragement of this project, Mr. I. MacDonald of the Department of Anaesthesia, Glasgow University, and his colleagues Mr. W. Swanson and Mrs. L. Naismith for their help and advice. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Barratt, R. S., Jones, R. L., and Thompson, J . M., Br. J . Anaesth., 1975, 47, 1177. Raymond, A., and Guiochon, G., J . Chromatogr. Sci., 1975, 13, 173. Altshuller, A. P., and Cohen, I. R., Anal. Chem., 1960, 32, 802. O’Keeffe, A. E., and Ortman, G. C., And. Chcm., 1966, 38, 760. Brookes, B. I,, Jickells, S. M., and Nicolson, R. S., J . Assoc. Publ. Anal., 1978, 16, 101. Brookes, B. I., Analyst, 1979, 104, 698. MacDonald, I . , and MacKenzie, J. E., B r . J . Anaesth., 1976, 48, 519. Health and Safety Executive, “Threshold Limit Values for 1978,” Guidance Note EH15/78, H.M. Fazzalari, F. A., Editor, “Odour and Taste Threshold Values Data,” American Society for Testing Weast, R. C., Editor, “Handbook of Chemistry and Physics,” Fifty-third Edition, Chemical Rubber Stationery Office, London, 1979. and Materials, Philadelphia, Pa., 1978. Company, Cleveland, Ohio, 1972-1973, p. PC-382. Received M a y 2nd, 1980 Accepted October 22nd, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600403

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

412 Analyst, April, 1981, Vol. 106, p p . 412-418 Comparison of Some Porous Polymers as Adsorbents for Collection of Odour Samples and the Application of the Technique to an Environmental Malodour Roger D. Barnes,* L. Maria Law and Alexander J. MacLeodt Department of Chemistry, Queen Elizabeth College, University of London, Campden Hill Road, London, W8 7AH A range of Chromosorbs and Tenax-GC have been compared with regard to their efficiency as adsorbents for volatile odorous compounds using a very simple model system. When desorption was accomplished by solvent elution using acetone, Chromosorb 103 provided the best recoveries of those investi- gated (between 90 and 95%). In situations where thermal lability of trapped compounds is not a problem, use of Tenax-GC and thermal desorption a t about 250 "C would be recommended on the basis of this survey, during which this approach consistently provided recoveries greater than 96% from small sample volumes.The latter procedure was applied to the analysis of an industrial malodour from an animal rendering factory, and using gas chromatography - mass spectrometry over 35 compounds (92% of the total odour sample) were positively identified. In particular, a range of alkylthiophenes may be characteristic of this particular odour. Keywords : Odour analysis ; porous polymer adsorbents ; Chromosorbs ; Tenax-GC ; animal rendering odour Many industrial and related processes give rise to objectionable malodours, which frequently are allowed to escape into the environment to the discomfort of the local community.I t is likely that legislation will be enforced to control such pollution, but in order to enact this adequately it is necessary to identify as far as possible the particular compounds responsible for the objectionable odours. Further, with this information it should then be more feasible to devise and introduce rational abatement procedures for the problem compounds. From an analytical point of view it is necessary to develop procedures whereby representative samples of the offensive odours can be collected in a form suitable for examination by the tech- nique ideally suited to the analysis of complex mixtures of volatile components, namely gas chromatography - mass spectrometry. However, the compounds of interest are present in the atmosphere in relatively small amounts so some form of concentration of the odorous air sample is usually necessary before analysis.Frequently sampling and concentration can be achieved in one step by some trapping procedure. There are three main methods: cryogenic tra~ping,l-~ adsorption on charcoa14p5 or adsorption on a porous polymer.6-11 Of these, the first two meth- ods suffer disadvantages not exhibited by the third. For example, with cryogenic trapping the nature of the coolant can be problematical. Although liquid oxygen has been used successfully* it is a hazardous material for routine use. If liquid nitrogen is used then liquid air will also be condensed and cause difficulties. Further, whatever the coolant, considerable amounts of water will be condensed from the air sample as ice, with the possibility of the formation of solid blockages in the trap.The main problem with charcoal as an adsorbant is that it is often difficult, if not impossible, to effect complete desorption for analysis, particularly of polar compounds. Porous polymers, on the other hand, can usually be readily and completely desorbed either simply by heat or by elution with an appropriate solvent. Equally, as they are hydrophobic they do not retain large amounts of water at ambient temperatures, although in some instances sufficient can be retained to cause some difficulties. For these and other reasons, porous polymers are a good choice as potential trapping materials for odorous air samples. However, there are many such commodities on the market and clearly it is necessary * Present address : Beecham Pharmaceuticals Ltd., Brockham Park, Surrey.t To whom correspondence should be addressed.BARNES, LAW AND MACLEOD 41 3 to evaluate these with regard to their efficiency and suitability for the analysis in question. This paper therefore partly describes a survey undertaken to assess the performance of some porous polymers as adsorbents, using a model system consisting of a simple mixture of volatile, odorous compounds. Having determined a suitable adsorbent in this manner, and at the same time optimum experimental conditions for its use, this system was then applied to a real analytical problem, namely the offensive odour associated with the process of hot rendering of animal by-products. Little previous work has been reported in this particular area, although Doty et aLl2 conducted a preliminary survey of the problem as long ago as 1972.Experimental Adsorbent Traps The porous polymer adsorbents examined were a selection of Chromosorbs (Phase Separa- tions Ltd., Queensferry, Clwyd, UK) and Tenax-GC (Field Instruments, Twickenham, Middle- sex, UK). Traps were prepared by packing a glass collection tube (200 x 4 mm i.d.) to a length of 75 mm with porous polymer, and were conditioned before use at the maximum recommended temperature for that particular adsorbent for 1 h in a flow of helium. All were 60-80 BSS mesh. Evaluation of Adsorbents Using a Model System The model system consisted of equal volumes of pentanal, pyridine, butyl acetate and diallyl disulphide.A glass tube (400 x 8 mm i.d.) was connected to the adsorbent trap using a PTFE sleeve, and an exact aliquot of the standard mixture (2 pl) was introduced into the horizontal sample tube via a vertical inlet protected by means of a silicone-rubber septum. A flow of dry nitrogen (variable within the range 50-750 ml min-l) was passed over the liquid sample and through the trap at ambient temperature for various lengths of time (initially mainly 15 or 30 min). In some instances a second trap of Chromosorb 103 was connected in series beyond the test trap, again by means of a PTFE sleeve. Initially the traps were desorbed by elution with acetone (exactly 500 pl) and 1 p1 of the eluate was then analysed. Thermal desorption is described later. Collection of Odorous Air Samples At first, samples were collected both by on-site trapping and by on-site collection of the atmosphere in poly(viny1 fluoride) (Tedlar ; Du Pont de Nemours International S.A., Geneva, Switzerland) bags of about 30-1 capacity.Subsequently it was found that samples stored in bags were unchanged after a period of up to 1 week, so thereafter only collection in Tedlar bags was used as this was much more convenient for replicate analyses. The bags were fitted with PTFE valves and sampling lines (Production Techniques Ltd., Fleet, Hampshire, UK) and were mounted in plastic carboys for portability. The carboy was evacuated on site by means of a battery-operated pump so that samples could be drawn into the bag. Samples were taken from the effluent pipes of the animal rendering cookers, where possible after condensers which removed the bulk of unwanted water.Odorous air (2 1) from the bag was passed through a trap of Tenax-GC at ambient temperature at a rate of 500 ml min-l. A second trap in series beyond the first cooled to -78 "C in a solid carbon dioxide - acetone bath could be used to collect highly volatile, gaseous organic components that passed through the main trap. Adsorbed odour samples collected in this manner could be stored unchanged for a few days at room temperature provided that both ends of the trap were securely stoppered. Analysis by Gas Chromatography A Pye-Unicam 104 instrument with a heated flame-ionisation detector was used. The columns employed were 1.5 or 5.5 m x 4 mm i.d. glass tubes packed with PEG 20M (10%) or PEG 6000 (5%) on Diatomite C, acid-washed and HMDS-treated (100-120 BSS mesh).The column temperatures employed were 90 "C isothermal for the examination of the model system and programmed from 70 to 175 "C at 2.5 "C min-l for the odorous air samples. The flow-rate of the carrier gas (helium or nitrogen) was always 30 ml min-l. Solvent desorbed samples (1 pl) were injected directly into the column in the normal manner. Thermal desorp- tion of Tenax-GC traps was achieved as follows. The trap was connected at the beginning of414 BARNES et al. : POROUS POLYMERS AS ADSORBENTS Analyst, VoZ. 106 the gas-chromatographic column using a suitable valving system and a carrier gas by-pass line (effectively a sample loop). The trap was then flushed with helium carrier gas at ambient temperature for 5-10 min, as it was found that if Tenax-GC was heated above 200 "C in the presence of oxygen some degradation occurred, and this caused the appearance of spurious peaks on the chromatograms.The gas flow was then stopped (i.e., passed directly through the gas-chromatographic column) and the trap heated as rapidly as possible to 260 "C using a heat- ing coil. The gas flow (30 ml min-l) was then re-diverted through the trap, and the sample thus rapidly introduced on to the column as a discrete band. The chromatogram was then developed. Total odour samples thermally desorbed in this manner from traps back into Tedlar bags were found to possess the same odour as the original odorous air samples. Analysis by Gas Chromatography - Mass Spectrometry An AEI MS 30 double-beam, double-focusing instrument was used, fitted with an integrated Pye-Unicam 104 gas chromatograph connected via a heated membrane separator interface.The mass spectrometer was linked on-line to an AEI DS 50 data processing system equipped with the facility for double-beam accurate mass measurement. Broadly the same gas- chromatographic conditions as already summarised (including sample introduction) were employed. Relevant mass spectrometric operating parameters were as follows : ionisation potential, 70 eV; ionisation current, 500 PA; source temperature, 200 "C; resolving power, 1500; and scan speed, 3 s per decade (repetitive throughout run). Results and Discussion Probably the most widely used porous polymer adsorbents for similar work (e.g., collection of food aroma volatiles) are the Chromosorbs, Tenax-GC and, to a lesser extent, the Porapaks.In this project a range of Chromosorbs and Tenax-GC were evaluated using a model system consisting of a mixture of arbitrarily selected volatile, odorous compounds (pentanal, b.p. 102-103 "C; pyridine, b.p. 115-116 "C; butyl acetate, b.p. 126-127 "C; diallyl disulphide, b.p. 174 "C). The components covered a range of compound type and polarity, and are typical of those which might be expected in environmental odour samples. An empty, horizontal glass tube was connected to a trap of the adsorbent under examination and 2 p1 of the sample mixture were injected into the tube. Trapping was then accomplished by allowing a stream of dry nitrogen (variable, but typically about 250 ml min-l) to sweep over the sample mixture at ambient temperature and through the adsorbent trap for various lengths of time.Preliminary experiments on the Chromosorbs indicated that Chromosorb 103 was probably the most efficient and it provided excellent recoveries of all components of the mixture. Detailed results for this adsorbent are given in Table I, and it can be seen that there was good reproduci- bility on replicate determinations. Improved collection was obtained using a sampling (nitrogen sweep) time of 30 min compared with 15 min. Longer periods did not provide any further advantage. The improved recovery between 15 and 30 min was not due to less com- plete evaporation of the test mixture from the glass tube during the shorter time.Other experiments showed that at least 96% of all components were evaporated within 5 min at a flow-rate of 500 ml min-l (see results for thermal desorption from Tenax-GC tubes, discussed later). Solvent elution from the trap was then employed, using acetone. TABLE I PERCENTAGE RECOVERIES OF COMPONENTS OF A SAMPLE MIXTURE FROM A CHROMOSORB 103 TRAP USING ACETONE AS ELUENT AND A SAMPLING FLOW-RATE OF 250 ml min-l Sampling timelmin Component 15 30 h f -l Pentanal . . . . . . .. 89, 85, 87 94, 94, 91 Butyl acetate . . .. . . 86, 84, 84 94, 94, 94 Diallyl disulphide . . . . . . 80, 85, 83 90, 93, 93 Pyridine . . .. .. . . 87, 80, 80 92, 90, 92April, 1981 FOR COLLECTION OF ODOUR SAMPLES 415 Similar preliminary assays using Tenax-GC and the same sampling rates (250 ml min-1) gave poorer results overall (about 50% at 30 min and 80% at 15 min).However, with this adsorbent, clearly efficiency was inversely proportional to the length of these sampling times, in contrast to the results for the Chromosorbs. Hence the optimum time was likely to be even less than 15 min, and with Tenax-GC increasing sampling periods beyond a certain point was counter-productive. Presumably the initially adsorbed sample was being desorbed and eluted in the nitrogen stream with time. Obviously, sampling time is a most important operational parameter when using Tenax-GC. A more detailed comparative survey of the adsorbents in question was then carried out as follows. The same system as already described was used, except that a second trap containing the known high-efficiency Chromosorb 103 was placed beyond the first sample trap to ascertain any loss of sample volatiles by incomplete adsorption.This also provided a valuable means of developing optimum sampling times, nitrogen flow-rates, etc., for various samples, as any circumstances that permitted sample components to “break through’’ the main trap into the second Chromosorb 103 trap were clearly unacceptable. Table I1 gives recoveries of the com- ponents from the same sample mixture from all the porous polymer adsorbents examined under conditions of flow-rate, sampling time, etc., optimised for that particular adsorbent (i.e., in all of these instances quoted, no components were detected in the second, reference trap).The optimum sampling time for the Chromosorbs was about 30min and did not vary with the particular type, whereas that for Tenax-GC was about 5 min. The nitrogen flow-rate was not so critical, but about 200mlmin-l was appropriate for the Chromosorbs and about 500 ml min-l for Tenax-GC. It can be seen from Table I1 that, for any particular adsorbent, recoveries of all components of the test mixture were virtually the same, within experimental error, except for Chromosorb 102 with which the recovery of diallyl disulphide was distinctly low and that of pyridine was extremely poor. As neither of these components was collected in the second trap, the problem here is incomplete elution, and it is possible that in these instances there is some interaction between solute and adsorbent.Whatever the reason, these results illustrate the necessity to assess adsorbents carefully before embarking on a real analytical problem. Clearly Chromo- sorb 102 is suspect for this type of project. With this exception, the recoveries for particular components were similar for the other three Chromosorbs and Tenax-GC, and at an acceptably high level of about 90% or just below. Chromosorb 103 would seem to be consistently slightly the best. As no components were detected in the second trap in these experiments, the weakness of the procedure, although slight, is the solvent elution. Clearly the alternative, thermal desorption, could have disadvantages in dealing with thermally susceptible components, but as this could not be a problem with regard to the animal rendering sample for analysis in this project, this method of elution was also assessed on samples trapped as already described.The particular procedure adopted was similar to that of Cropper and Kaminsky.13 Thermal desorption possesses some advantages in being simpler and less liable to quantitative inaccuracies. How- ever, using this method of elution the Chromosorbs were much less satisfactory owing to their TABLE I1 PERCENTAGE RECOVERIES OF COMPONENTS OF A SAMPLE MIXTURE FROM VARIOUS POROUS POLYMER ADSORBENTS UNDER INDIVIDUAL OPTIMISED TRAPPING CONDITIONS USING ACETONE AS ELUENT Recovery, yo * r 1 r Chromosorb A 1 Component 101 102 103 105 Tenax-GC Pentanal . . . . . . 88 84 93 87 87 Pyridine . . . . . . 84 54 91 83 87 Diallyl disulphide . . . .85 75 92 83 84 Butyl acetate . . . . 88 82 94 83 84 * Recoveries are averages of four determinations. Typical standard deviations for the figures quoted range from 0.866 to 3.55%.416 BARNES et aZ. : POROUS POLYMERS AS ADSORBENTS Analyst, VoZ. 106 relatively low thermal stability. At temperatures at which they did not degrade to produce an unacceptably high interfering background, the efficiency of desorption was very poor (in particular, chromatographic performance was bad). Solvent elution is hence the better pro- cedure for these adsorbents. Tenax-GC, on the other hand, has a high thermal stability and reputedly it does not degrade below 350 "C. However, desorption temperatures of about 260 "C were adequate to provide almost quantitative elution. Thus, using Tenax-GC traps with relatively short sampling times (5 min), a flow-rate of 500 ml min-l and thermal desorp- tion, it was possible to obtain recoveries for all components in the sample mixture consistently and reproducibly better than 96%.Individual results obtained were as follows (the figures given are the averages of four determinations rounded to whole numbers, and standard devia- tions are given in parentheses) : pentanal, 96% (1.32) ; pyridine, 98% (0.78) ; butyl acetate, 98% (1.41) ; and diallyl disulphide, 97% (0.50). On the basis of these results, this system was then adopted as the best for the analysis of the odorous air samples obtained from factories concerned with the hot rendering of animal by-products. It should be emphasised that al- though this system is entirely appropriate for this particular project, Chromosorb 103 and solvent elution would be the procedure of choice for the examination of odorous samples and volatiles produced by other than thermal means.Although Tenax-GC has a very high affinity for organic compounds, it was found that when the odorous air sample was pumped through the single trap using the optimised conditions just described, a few gaseous, organic components were not retained. However, these compounds were readily collected by incorporating in series beyond the main trap a second, identical Tenax-GC trap but which was cooled to -78 "C in a carbon dioxide - acetone bath. This arrangement then retained all volatile components in that none was eluted from a third trap (sub-ambient), and the combined eluates from the two traps possessed the same aroma as the original sample.The second trap merely contained the expected collection of gases that emanate from the majority of such industrial processes, namely ammonia, hydrogen sulphide, methanethiol and trimethylamine. These can usually be prevented from escaping into the atmosphere from the factory by "good housekeeping". Examination of the contents of the main trap by routine gas chromatography showed them to be very much more complex than those of the second trap, consisting of at least 60 com- ponents, and clearly these were more specific to the particular sample in question. The majority of these components were identified by gas chromatography - mass spectrometry using an AEI MS 30 instrument equipped with a DS 50 data processing system.Details of identities are given in Table I11 together with the KovAts retention indices of the gas-chroma- tographic peaks and the approximate concentration in parts per million of each component in the original total odour sample. The latter were assessed using the TIC monitor of the gas chromatograph - mass spectrometer and suitable reference standards, and were corrected to allow for the over-all recovery of the analytical procedure. Initially identities of unknown components were suggested from their mass spectra on the basis of knowledge of fundamental fragmentations. These assignments were then confirmed by comparison with literature spectra. In all instances where positive identifications are quoted in Table 111, agreement with literature spectra was nearly perfect and within experimental ( i t ? ., instrumental) variability. None of the compounds identified is particularly unusual chemically, and mass spectra are not quoted here as all appear in the existing literature. The background subtraction facility and the retrospective single-ion monitoring facility of the mass spectrometer data system were extensively employed in evaluating the results, and were particularly valuable in assigning some alkanals and aliphatic hydrocarbons that were not resolved by gas chromatography - mass spectrometry (see Table 111). Thus a range of straight- chain aliphatic hydrocarbons from C, to C,, inclusive was identified in the odour sample. A similar collection of alk-1-enes, but with a few unspecified, was also detected.In general the former were produced in greater amounts than the latter. A range of straight-chain alkanals from C, to C, inclusive, and of alkylbenzenes up to hexyl, were also identified. Some of the latter were produced in large amounts, but a significant contributor to the specific odour would probably be 3-methylbutanal at 8.1 p.p.m. Much more important to the characteristic odour, however, would be the interesting range of alkylthiophenes from C, to C, inclusive. Although in general they were formed in only trace amounts, such compounds do have very low odour threshold values. For example, the threshold in water for a C,-substituted thiophene is A number of series of compounds can be recognised in Table 111.April, 1981 FOR COLLECTION OF ODOUR SAMPLES 417 1.3 x 10-3 p.p.m.,14 compared with 5.0 x p.p.m.for methanethi01.l~ Alkylthiophenes have been previously identified in certain heat- treated foods, e.g., propyl- and butylthiophenes in coffee16 and ethyl-, butyl-, pentyl- and octylthiophenes in boiled beef .17918 Particularly with respect to the latter occurrence, their liberation on the hot rendering of animal by-products would not, therefore, be unexpected. However, numerous other compounds have been determined in cooked meat aroma that were not detected in the factory odorous air sample, so clearly the alkylthiophenes have special significance here. The origin of such a series is intriguing but no explanation can yet be offered. p.p.m. for hydrogen sulphide and 2.0 x TABLE I11 ANALYSIS OF THE VOLATILE COMPOSITION OF ANIMAL RENDERING ODOUR Component Unknown .. .. Heptane . . .. Hept-1-ene . . .. Octane . . . . An octene . . . . An octene .. .. Butanal . . .. An octene . . . . Nonane . . . . 3-Methylbutanal . . Non-1-ene . . .. Benzene . . . . Pentanal . . .. Decane . . . . Dec-1-ene . . .. Toluene . . . . Dimethyl disulphide Hexanal . . . . Undecane . . . . A xylene . . . . An undecene.. .. A xylene . . . . A xylene . . .. Heptanal . . . . Dodecane . . .. Ethylbenzene . . Unknown .. .. Unknown .. .. C,, branched-chain hydrocarbon . . C,, branched-chain hydrocarbon . . 2-Propylthiophene . . Styrene . . .. * tr = trace. Kovkts retention index 650 700 735 800 820 830 840 875 900 910 925 930 970 1 000 1045 1060 1070 1070 1100 1120 1140 1150 1170 1185 1 200 1225 1225 1240 1250 1260 1270 1270 Approximate abundance in odour sample, p.p.m.* 0.6 5.5 0.4 5.5 0.3 0.3 0.4 0.4 0.3 8.1 t r 0.6 0.6 0.6 t r 2.8 1.3 1.3 0.6 0.6 tr 2.0 t r 2.0 1.3 11.3 1.3 0.5 0.5 0.4 0.6 t r Component Octanal .. .. Tridecane . . . . 1,3,5-Trirnethylbenzene Butylbenzene . . Unknown . . .. A tridecene . . .. 2-Butylthiophene . . Unknown . . . . Tetradecane . . . . A tridecene . . . . Dimethyl trisulphide Unknown . . .. Pentylbenzene . . Unknown . . . . 2-Pentylthiophene . . Unknown . . . . Decan-2-one . . . . Hexylbenzene . . C, unsaturated aldehyde (C,H,,CHO) . . 2-Pentylpyridine . . Unknown . . . . 2-Hexylthiophene . . C , unsaturated aldehyde (C,H,,CHO) . . Unknown . . . . C, unsaturated aldehyde (C,H,,CHO) .. C, unsaturated aldehyde Unknown .. . . (CcSiXHO) . - Kovkts retention index 1290 1300 1320 1345 1355 1360 1370 1380 1385 1400 1410 1415 1415 1440 1445 1465 1475 1490 1510 1520 1540 1550 1560 1570 1600 1610 1630 C,~H~,O~isomer . . . . 1675 Unknown . . . . 1730 Approximate abundance in odour sample, p.p.m.* 0.6 0.6 t r 2.8 0.5 0.3 0.1 t r 0.4 1.3 tr 0.3 0.1 2.1 0.6 tr 0.5 0.6 2.8 0.6 0.5 0.6 tr 0.4 0.5 tr 0.4 tr 0.6 Of the 61 components detected in the trapped sample of odorous air (Table 111), 36 were positively identified with a further 13 being partially characterised. Overall, about 92% of the total odour sample was fully identified and the remaining 8% was distributed over about 25 minor components, many of which were characterised to some extent. As the odour sample was representative of animal rendering odour and as about 92% of it was positively identified, it is reasonable to claim that the majority of compounds that contribute to animal rendering odour have now been determined.This work was partly sponsored by a grant from the Department of the Environment. We are particularly grateful to the staff at Warren Spring Laboratory, Stevenage, Hertfordshire418 BARNES, LAW AND MACLEOD (especially J. C. Bailey, N. Hurford, A. A. North and N. J. Viney) for productive discussion and consultation, and also for the collection and provision of some odorous air samples from animal rendering factories. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Burnett, W. E., Environ. Sci. Technol.. 1969, 3, 744. Kato, T., US Nat. Tech. Inform. Serv., PB Report N4-11933/TR-289-73. Williams, I. M., Anal. Chem., 1965, 37, 1723. Grob, K., and Grob, G., J . Chromatogr., 1971, 62, 1. Mueller, F. X., and Miller, J. A., Int. Lab., 1974, 34. Spinner, E., J . Chromatogr. Sci., 1975, 13, 181. Bertsch, W., Chang, R. C., and Zlatkis, A., J . Chromatogr. Sci., 1974, 12, 175. Perry, R., and Twibell, J. D., Atmos. Environ., 1973, 7, 929. Mieure, J. P., and Dietrich, M. W., J . Chromatogr. Sci., 1973, 11, 559. Zlatkis, A., Lichenstein, H. A., and Fishbee, A., Chromatographia, 1973, 6, 67. Pellizari, E. D., Bunch, J. E., Berkley, R. E., and McRae, J., Anal. Chem., 1976, 48, 803. Doty, D. M., Snow, R. H., and Reilich, H. G., “Investigation of Odour Control in the Rendering Industry,” Environmental Protection Agency, Washington, D.C., Report EPA-R2-72-088, 1972. Cropper, F. R., and Kaminsky, S., Anal. Chem., 1963, 35, 735. Boelens, M., de Valois, P. J., Wobben, H. J., and van der Gen, A., J . Agric. Food Chem., 1971, 19, Stahl, W. H., “Compilation of Odour and Taste Threshold Values Data,” American Society for Vitzthum, 0. G., and Werkhoff, P., 2. Lebensm. Unters. Forsch., 1976, 160, 277. Hirai, C., Herz, K. O., Pokorny, J., and Chang, S. S., J . Food Sci., 1973, 38, 393. Garbusov, V., Rehfeld, G., Wolm, G., Golovnja, R. V., and Rothe, M., Nahrung, 1976, 20, 235. 984. Testing and Materials, Philadelphia, Pa., 1973. Received September 2nd, 1980 Accepted October 31st, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600412

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, April, 1981, Vol. 106, $9. 419-428 419 Simultaneous Determination of Trace Metals in Sea Water Using Dithiocarbamate Pre-concentration and Inductively Coupled Plasma Emission Spectrometry C. W. McLeod National Institute for Environmental Studies, Yatabe, Ibaraki 305, Japan, and University of Tokyo, Bunkyo-ku, Tokyo, Japan A. Otsuki and K. Okamoto H. Haraguchi and K. Fuwa National Institute for Environmental Studies, Yatabe, Ibaraki 305, Japan Department of Chemistry, University of Tokyo, Bunkyo-ku, Tokyo, J a p a n National Institute for Environmental Studies, Yatabe, Ibaraki 305, Japan, and Department of Chemistry, University o j Tokyo, Bunkyo-ku, Tokyo, Japan A method based on dithiocarbamate pre-concentration and inductively coupled plasma emission spectrometry is described for the simultaneous determination of cadmium, copper, iron, molybdenum, nickel, vanadium and zinc in sea water.The metals are extracted from 500g of sea water with ammonium tetra- methylenedithiocarbamate - diethylammonium diethyldithiocarbamate in chloroform and back-extracted into nitric acid ; the sea water concentration factor is 250 or 500. Advantages of the method include high precision, simplicity of calibration and a detection capability in the nanograms per litre range. The method has been applied to Japan Sea, Pacific Ocean and Atlantic Ocean samples. Keywords Trace metals ; sea water ; dithiocarbamate pre-concentration ; inductively coupled p l a s m a emission spectrometry The National Institute for Environmental Studies, Japan, is involved in studies to determine background levels of selected pollutants for' monitoring purposes and one objective is the determination of trace metals in sea water. Graphite furnace atomic-absorption spectrometry (GFAA) , an extremely sensitive analytical technique, has found widespread use in sea water analysis and an extensive review of the sub- ject is available.1 Invariably, as indicated in recent reports,2-6 when utilising the graphite furnace technique, some form of pre-concentration is required for quantification of the extremely low levels of trace metals in open-ocean water.In a study on cadmium, copper, nickel and zinc by Bruland et a1.,2 for example, the minimum metal concentrations determined for a vertical sea water profile (from the surface to a depth of about 3000 m; 37"05'N, 123"22'W) were 15 ng 1-1 of cadmium (upper value 118 ng 1-l), 92 ng 1-1 of copper (upper value 240 ng 1-l), 228 ng 1-1 of nickel (upper value 693 ng 1-l) and 7 ng 1-1 of zinc (upper value 628 ng 1-l).The analytical results, among the lowest reported in the literature, were con- sidered to reflect the total metal content of sea water, i.e., the dissolved, colloidal and particulate fractions. A report has recently been published' describing the application of inductively coupled plasma (ICP) emission spectrometry with ultrasonic nebulisation, and ion-exchange pre- concentration to sea water analysis. The ICP technique lacks the sensitivity of GFAA but provided relatively high pre-concentration factors are achieved, the sensitivity limitation can be circumvented and the analytical advantages of the ICP technique such as (i) a simultaneous multi-element capability, (ii) a relative absence of matrix interference and (iii) a wide dynamic range, can be realised for sea-water analysis. In the afore-mentioned ICP study,' a 100-fold pre-concentration of sea water allowed quantification of manganese, iron, zinc, copper and nickel in a relatively unpolluted coastal water sample.A multi-element pre-concentration420 MCLEOD et al. : TRACE METALS IN SEA WATER BY Analyst, Vol. 106 procedure based on co-precipitation and flotation with indium hydroxide, and offering a 240- fold concentration factor, has also been used in conjunction with ICP emission spectrometry for the determination of heavy metals in water and artificial sea water.* Of the pre-concentration procedures developed for GFAA, the solvent extraction method of Bruland et al.,2 based on metal dithiocarbamate formation, is considered suitable for ICP analysis for several reasons.The dithiocarbamate system is applicable to many trace elements, the extraction process eliminates the bulk of the sea-water salts and problems related to the nebulisation of extracts having high dissolved solids contents are therefore avoided. Further- more, a relatively high concentration factor, a basic requirement for the ICP analysis of sea water, was demonstrated in the study by Bruland et aL2 Briefly, the reported procedure involved a double extraction of acidified sea water (250g) using chloroform and combined ammonium tetramethylenedithiocarbamate (APDC) - diethylammonium diethyldithiocarba- mate (DDDC) followed by back-extraction of the metal carbamates into nitric acid.Subse- quent sample evaporation - oxidation and semi-quantitative volumetric transfer gave a 200-f old sea-water concentrate. In this paper a modified dithiocarbamate extraction procedure in combination with ICP emission spectrometry is evaluated for the simultaneous determination of cadmium, copper, iron, molybdenum, nickel, vanadium and zinc in sea water. Method development considera- tions are discussed with attention focused on maximising the sea-water concentration factor, minimising the potential matrix interference effects in the sea-water concentrate and minimis- ing the number of steps for sample processing.The analytical results for open-ocean samples indicate that the ICP method offers a high detection capability, together with good precision and reliability. Experimental Purification procedures and sample processing steps were performed in a class 1000 clean- room and as much as possible at a laminar-flow clean-air bench. Pre-cleaned PTFE laboratory equipment (Nalgene FEP, Nalge Co., New York) was used unless otherwise stated, for the preparation and storage of purified reagents and for sample processing steps. Miniature PTFE beakers with tight fitting caps (4-ml capacity) (Sanai Co., Nagoya, Japan) were used in the final stage of sample processing. The pre-cleaning procedure for laboratory equipment was similar to that reported by Patterson and Settle.g Much information regarding reagent purification, amounts of reagents, shaking and standing times for extraction, etc., was obtained from the references already cited.Reagents and Purification Procedures Water. A quartz sub-boiling still utilising Millipore water as the input was used. The product was stored in a 2-1 pre-cleaned PTFE bottle . Nitric acid. Sub-boiling distilled nitric acid was prepared in an all-PTFE still of similar design to that reported by Mattinson.lO The nitric acid feed was atomic-absorption grade (Kanto Chemical Co., Tokyo, Japan). Spectroscopic grade chloroform (400 ml ; Wako Pure Chemical Industries, Osaka, Japan) was extracted 3-times with nitric acid (100 ml, 2 M; Kanto, atomic-absorption grade) in a 500-ml separating funnel with fresh nitric acid being used for each extraction.The chloroform was next washed 3 times with high-purity water and then transferred into a PTFE bottle. When not in use, the PTFE bottle was placed in a black polyethylene bag and stored in a refrigerator. Usually 2 1 of purified chloroform were prepared at one time, this solvent being stable for about 3 weeks. APDC - DDDC solution. APDC (0.5 g , Wako) and DDDC (0.5 g, Wako) were weighed into a 50-ml measuring cylinder and dissolved in 50 ml of water . The solution was then transferred into a 250-ml separating funnel and extracted using 10 ml of purified chloroform. The pro- cedure was repeated a further 5 times using fresh chloroform. The purified extracting reagent was stored in a pre-cleaned 100-ml PTFE bottle.A sea-water extraction experiment indi- cated that the solution was stable for at least 24 h. Equal volumes (500 ml) of glacial acetic acid (Kanto EL, high- purity grade) and 25% ammonia solution (Kanto, atomic-absorption grade) were mixed slowly over a 2-h period, in a polypropylene container placed in a water-bath. The acetate . Chloroform. Ammonium acetate bNffer.April, 1981 ICP SPECTROMETRY AFTER PRE-CONCENTRATION 421 buffer solution was then purified as follows. Purified APDC- DDDC solution (5 ml) was added to 400 ml of the buffer solution and the mixture extracted 6 times with separate 40-ml portions of purified chloroform. The purified buffer solution was stored in a pre-cleaned 2-1 PTFE bottle. Multi-element standard solutions were prepared from single element stock solutions (1 000 pg ml-1) by appropriate dilution. High-purity metals or oxides were used to prepare the stock solutions.Synthetic sea water prepared from analytical-reagent grade chemicals3 and Japan Sea water were used for method development. Open-ocean samples, obtained from the Atlantic Ocean and the Pacific Ocean, were acidified on collection to approximately pH 2 by the addition of 1 ml of nitric acid (sub-boiling distilled) to 1 1 of sea water. The Pacific sample was obtained during the Hako Maru cruise (KH 80.2, April 26th-June 18th, 1980; University of Tokyo Oceanographic Institute) at station 10 (32"01'N, 144"60'E), from a depth of 4500 m. An all-stainless-steel sampler of Niskin design was used for collection and the sample was stored in a pre-cleaned 5-1 polyethylene container.The Atlantic water, made available to participants attending the Intergovernmental Oceanographic Commission's inter- calibration workshop, Bermuda Biological Station, January 10-26th, 1980, was obtained from a high-volume holding tank (polyethylene lined, 500-1 capacity) containing sub-surface water that had been sampled about 2 miles off Bermuda (32"22'N, 64'44'W) utilising an all- PTFE pump with polyethylene tubing. Approximately 20 1 of sea water were withdrawn from the tank into a pre-cleaned polypropylene container. Pre-concentration Procedure Weigh accurately about 500 g of acidified sea water (pH about 2) into a 1-1 beaker on a pan balance. Transfer into a 500-ml separating funnel and add 1.4 ml of acetate buffer, 2 ml of APDC - DDDC solution and 10ml of chloroform.Shake the mixture on a mechanical shaker for 3 min and then allow 5 min for the phases to separate and drain off the chloroform layer into a 125-ml separating funnel. Add a further 10 ml of chloroform, shake for 3 min, allow to stand for 5min and then combine the two chloroform extracts. Add0.20mlof concentrated nitric acid to the chloroform extracts, shake for 1 min and then allow to stand for 5 min. Add high-purity water (about 2ml), shake for 1 min and then allow to stand for 5 min, discard the chloroform layer and drain the aqueous layer into a 4-ml PTFE beaker, rinse the separating funnel with about 2 ml of high-purity water, adding the washings to the PTFE beaker. Then evaporate the extract to dryness using an infrared lamp on a clean-air bench and dissolve the residue in 0.3 ml of concentrated nitric acid.Continue to heat the nitric acid solution under the lamp to a small volume (about 50 pl) and then add water (exactly 1.0 or 2.0 ml). Instrumentation and Calibration Inductively coupled plasma. A Jarrell-Ash Atomcomp MKII direct-reading emission spectrometer with pneumatic nebulisation (cross-flow type nebuliser) was used. The wave- lengths for the elements programmed into the polychromator and instrument detection per- formance under normal plasma operating conditions are listed in Table I. The values for the principal plasma operating parameters (i.e., R F power 1.1 kW, observation height 18 mm and sample uptake rate 1.0 ml min-1) were determined from a previous optimisation study.ll Standardisation of the instrument was based on a two-point calibration procedure, using a multi-element solution containing cadmium, copper, iron, molybdenum, sodium, nickel, vanadium and zinc (1 pg ml-1 of each in 0.4 M nitric acid) as the high standard and distilled water (zero metal concentration) as the low standard.The sample volume of both the 1 and 2 ml concentrates permitted the acquisition of 3 independent intensity measurements and the print-out in concentration units therefore consisted of three values. For the 1-ml extract, however, none of the sample remained after these measurements. A sample pre- rinse time of about 20 s was used. Standardisation of the instrument was usuaily repeated after the assay of each set of sea-water extracts to minimise the effects of instrument drift.Distilled water and a check standard were also included in the analysis sequence. An IL atomic-absorption instrument (IL 151) with a graph- ite furnace attachment, Model 555, was used. The atomic-absorption wavelengths and operat- ing conditions for the graphite furnace atomiser were those recommended by the manufacturer. Standard solutions. Sea water. The solution is then assayed directly from the PTFE beaker. Atomic-absorption spectrometer.422 MCLEOD et al. : TRACE METALS IN SEA WATER BY Analyst, VoZ. 106 TABLE I ANALYTICAL WAVELENGTHS AND DETECTION LIMITS FOR ICP INSTRUMENTATION See text for principal plasma operating conditions. Element Cadmium .. .. .. Copper . . .. .. Iron: . . .. .. Molybdenum .. .. Nickel: . . .. .. Vanadium1 .. .. Zinc .. .. .. Wavelength /nm 228.8t 324.8 259.9 202.0 231.6t 292.4 213.9 Detection limit*/pg 1-' 1 1 3 4 8 3 1 * Detection limit is defined as the concentration that corresponds to an emission t Refers to 2nd order. : Refers to ion line. signal equivalent to twice the standard deviation of the background noise. * Results and Discussion Comments on the Pre-concentration Procedure The solvent extraction procedure of Bruland et aL2 was considered suitable for ICP analysis but was modified for a number of reasons. To compensate for the poorer sensitivity of the ICP technique relative to GFAA, a procedure offering, at maximum, a 500-fold concentration was devised. The sample volume of 1 ml represents the minimum acceptable volume for ICP analysis using commercially available nebulisers.A 2-ml sea-water extract however, may be preferable (see Pre-concentra- tion Procedure) as the sample volume is then not exhausted by ICP analysis and further data may be obtained by other techniques, e g . , GFAA. Unless otherwise stated the results in this section are based on a 250-fold conceutration factor. An additional step, taken to ensure the maximum ICP detection capability, was to minimise the nitric acid concentration in the final sea-water extract. The ICP detection sensitivity is known to deteriorate in solutions of high nitric acid concentration and this difficulty was noted by McLaren et aZ.7 in a sea-water analysis study utilising an ion-exchange pre-concentration procedure. In the present study only a slight deterioration in detection performance was noted for sea-water extracts 0.8 M in nitric acid (the approximate acid strength for the 1-ml concentrate).No significant deterioration was found for the solution 0.4 M in nitric acid (the approximate acid strength for the 2-ml concentrate). As indicated later for the assay of an open-ocean sample the agreement between data for the 1- and 2-ml concentrates confirmed the absence of acid interference. An important feature of this method is that ICP analysis is performed directly from the PTFE vessels used to receive the sea-water extracts after back-extraction. The elimination of an additional sample transfer step reduces the possibility for analyte loss, volumetric errors and contamination.It should be noted that evaporation of sea-water extracts to small volumes is time consuming (at least 6 h) relative to solvent extraction and ICP analysis, but the evaporation can be conveniently performed overnight and unattended. The formation of the 5O-pl droplet, following oxidation of the residue, is rapid (about 30 min) and reproducible and this, together with the final volumetric step (the addition of 1 or 2 ml of high-purity water), ensures good reproducibility for the procedure. In a normal working day up to 24 sea-water samples may be processed to the evaporation stage. This was achieved by concentrating 500 g of sea water to 1 ml. Effect of pH on Extraction Efficiency The combined APDC - DDDC extracting reagent was first proposed by Kinrade and Van Loon12 for the pre-concentration of trace elements from natural waters, primarily because of its ability to complex many metals and also because there is a broad pH range over which the extraction process is equally efficient.In recent applications to sea water pre-concentration the method of Danielsson et aL3 utilised pH 5 for the extraction of cadmium, cobalt, copper, iron, nickel, lead and zinc whereas Bruland et aZ.2 used pH 4 for cadmium, copper, nickel and zinc. For this study the effect of pH on extraction efficiency was investigated as informationApril, 1981 ICP SPECTROMETRY AFTER PRE-CONCENTRATION 423 on molybdenum and vanadium, utilising the combined extracting reagent, is not documented. As can be seen in Fig. 1 the extraction efficiency was constant over the pH range 2-6 for many metals, with the curves for cadmium, copper, iron, nickel and zinc being similar to those reported by Danielsson et aL3 For molybdenum and vanadium, particularly the former, extraction efficiency was sensitive to pH with the curves exhibiting maxima at about pH 4. Based on these findings precise control of the pH of the medium is required to obtain quantita- tive data for molybdenum and vanadium and for this work pH 4.l&O.l was utilised.Matrix Interference It has been demonstrated that the ICP technique does not suffer appreciably from matrix interference. The deterioration in detection performance for solutions of high acid content is well known and has already been discussed for nitric acid. The possibility of stray light and nebuliser-related effects associated with extracts containing high concentrations of alkali and alkaline-earth metals was considered, but was not found to be significant.Fortunately the extraction process removes the bulk of the sea-water salts and the concentrations of sodium, calcium and magnesium for the 250-fold sea-water extracts were typically in the ranges 0.1- 60, 0.05-3 and 0.2-12 pg ml-l, respectively. In routine analysis it is recommended that the multi-standard solution contains an appropriate concentration of sodium (or calcium or magnesium), e.g., 1-10 pg ml-l to check the level of salt carry-over. Reports in the literaturel3J4 have indicated that plasma background shifts occur when solutions containing organic matter are nebulised. The occurrence of such phenomena due to retention of residual chloroform in the sea-water extracts was considered to be unlikely as oxidation of the residues was carried out.In an attempt to detect such interference the pre-concentration procedure was performed with omission of the oxidation step, i.e., sea- water extracts were evaporated directly to 50 pl before the final volumetric addition. The analytical data for sample and blank solutions by the above technique and by the recommended procedure were indistinguishable and background shifts were not observed. The oxidation step in sample processing is recommended, however, as the addition of nitric acid was found to improve the reproducibility of droplet formation. Method Sensitivity and Blank The method sensitivity for the elements based on the respective ICPlower limits for quanti- fication (LLQ) and a 500-fold concentration factor are presented in Table 11.A compilation of the trace element concentrations in sea water is also given for comparison purposes. Intended as a rough guide, it can be seen from Table I1 that the sensitivity of the proposed ICP method may be sufficient for quantifying at the lower end of the concentration scales. As indicated in recent sea-water analysis however, trace element concentrations in unpolluted water may be significantly lower than those indicated in Table 11. The present discussion assumes that the analytical blank levels do not represent significant fractions of the total analytical signals. Minimisation and control of the blank levels are essential for the achievement of reliable determinations of trace metals in sea water on a routine basis.TABLE I1 METHOD SENSITIVITY FOR ELEMENTS AND APPROXIMATE SEA-WATER CONCENTRATIONS Element Cadmium . . .. Copper . . . . Iron . . .. Molybdenum . . Nickel . . . . Vanadium . . Zinc .. .. Method sensitivity* I ng 1-l .. 10 . . 10 .. 30 .. 60 . . 80 . . 30 .. 10 Sea-water concentrationt I ng 1-l 70 600-2 000 100-62 000 2 100-18800 1 100-4 000 200-4 000 100-50 000 * Based on 500-fold concentration factor and ICP lower limit The LLQ is taken as 5 times the for quantification (LLQ). detection limit,.e.g., LLQ for cadmium is 5 pg 1-'. t See reference 15.424 1 v 1 I I I I L MCLEOD et al.: TRACE METALS IN SEA WATER BY Analyst, Vol.106 0.06 - cu 0.04 - 0.01 0.02 - 0 0 0.6 0.4 - I - E 0, 5 C -..s 0.2 h 4- C a, C 0 0 0 I 0.04 O . O 8 I 1 I I 6 I I Ll 0 1 2 3 4 5 6 7 8 PH I 3.0 I 0.4 0.2 0 0 1 2 3 4 5 6 7 8 PH Fig. 1. Effect of pH on extraction efficiency. A, Japan sea water; a, synthetic sea water: the solution was enriched with multi-element solution before extraction ; actual concentrations for Cd, Cu, Fe, Ni and Zn were 10 times greater than indicated; buffer, 1 M sodium phosphate solution (3 ml). Solution pH was adjusted by addition of HNO, or NH, solution.April, 1981 ICP SPECTROMETRY AFTER PRE-CONCENTRATION 425 Blank extracts were obtained by performing the entire procedure (equivalent to a 250-fold concentration) in the absence of a sea-water sample and ICP analysis indicated that the con- tamination levels were below the respective LLQ except with iron.Therefore, the more sensitive graphite furnace technique was employed and the analytical results for the metal blank levels are given in Table 111. The molybdenum and vanadium blanks were not detected by GFAA and contamination levels for the remaining elements, except iron, were generally insignificant in relation to sea-water concentrations (see Table VII). The high iron blank represented a significant fraction of the analytical signals and in an attempt to obtain reliable data, a subtraction of the blank was performed. A lower limit for iron determinations in sea water is considered to be about 500 ng l-l, where the contamination level accounts for about 20% of the ICP signal. Experiments indicated that iron contamination originated primarily from the chloroform and nitric acid used in the sample processing.With respect to the purifi- cation of nitric acid a report16 has revealed that the PTFE sub-boiling still can be a source of iron contamination and all-quartz apparatus would be preferable. TABLE I11 METHOD BLANK AS DETERMINED BY GRAPHITE FURNACE ATOMIC-ABSORPTION SPECTROMETRY Values refer to blank sample concentrates. concentration divide data by 250. To obtain method blank in terms of an equivalent sea-water Number of Uncertainty limits expressed as standard deviation. replicate analyses, 12. Element Cadmium . . .. .. Iron* . . .. .. .. Nickel . . .. .. . . Vanadium . . .. . . Zinc . . .. .. . . Copper . . . . .. .. Molybdenum . . .. . . Concentration/ pg 1-1 0.3 f 0.1 0.4 f 0.1 23 f 8 < 1.0 3 f 1 < 1.0 0.3 f 0.1 * By ICP emission spectrometry.Precision and Accuracy Method precision was determined by performing replicate analyses of a Japan Sea sample (n = 8) and a synthetic sea water (n = 6). The results are given in Table IV together with those for instrumental measuring precision. The values for the latter were derived from analy- sis of a multi-element solution with a composition similar to the Japan Sea water concentrate. The data in Table IV indicate high precision for the procedure, the relative standard deviations TABLE IV r Element Cadmium .. .. .. Copper .. .. .. Iron . . . . . . . . Molybdenum . . .. .. Nickel . . .. . . Vanadium .. .. .. Zinc . . . . . . . . METHOD AND INSTRUMENTAL PRECISION Synthetic Multi-element Japan solution Sea water sea water Concentration/pg I-' RSD*, % Concentration/pg 1-1 RSD*, % RSD, % pg ml-' L > r & , Instrumental concentration/ 0.016 31 0.076 8.4 1 2 0.01 0.308 5.8 0.948 3.8 2.0 0.1 1.26 7.6 14.0 3.3 1.5 0.2 7.76 4.6 (4.0 1.2 0.2 0.236 10 2.35 3.2 5.8 0.05 1.46 1.1 <3.0 1.3 0.5 1.42 2.1 7.84 3.1 0.9 0.2 RSD = relative standard deviation.(RSD) being as good as or better than those obtained using methods based on GFAA. For certain metals (e.g., cadmium and nickel) whose concentrations in the Japan Sea concentrate were extremely low, determinations were performed at, or close to, the respective LLQ and therefore ICP measuring precision contributed significantly to the total method precision (cf., corresponding data for the synthetic sea water).For a critical assessment of accuracy there is no substitute for a sea-water standard reference material (SRM) certified for elemental composition but, unfortunately, such a standard has yet to be prepared. The accuracy of the procedure was therefore tested by performing a spike426 MCLEOD et d.: TRACE METALS IN SEA WATER BY Analyst, VOl. 106 recovery experiment and by analysing SRM 1643a, Trace Elements in Water (simulated fresh-water standard, US National Bureau of Standards). For the former, two 500-g aliquots of the Japan Sea sample were spiked with multi-element solution (0.6 ml; see Table V for composition) just before extraction. Element recoveries, as indicated in Table V, were essentially 100 yo except for molybdenum. An additional experiment indicated that recoveries for a single extraction using only 10 ml of chloroform were approximately 10% lower than the values listed in Table V.TABLE V RECOVERY RESULTS FOR SPIKE ADDITION TO JAPAN SEA SAMPLE Element Cadmium . . .. Iron . . .. .. Nickel . . .. Vanadium .. .. Zinc . . . . .. Copper . . .. Molybdenum .. Original* concentration/ pg ml-1 0.004 0.077 0.315 1.94 0.059 0.365 0.355 Amount added?/ - pg ml-l 0.0 125 0.125 0.25 2.5 0.125 0.5 0.5 Amount found*/pg ml-1 A \ 1 0.017 0.205 0.532 4.16 0.185 0.830 0.849 2 0.017 0.202 0.542 4.02 0.196 0.846 0.855 Recovery, % 100 101 95 92 104 97 100 * Values for original and found refer to sea-water concentrates. To obtain concentrations for Japan Sea sample divide original value by 250. t 0.5 ml of the multi-element solution added to two 500-g aliquots of sea water; multi-element solution composition Cd 0.05 pg ml-l, Cu 0.5 pg ml-l, Fe 1.0 pg ml-l, Mo 10 pg ml-l, Ni 0.5 pg ml-l, V 2 pg ml-1 and Zn 2 p g ml-l.In a further attempt to assess accuracy, SRM 1643a was analysed after appropriate dilution. The SRM water, which contains trace elements at the nanogram per gram level and may be analysed directly by ICP emission spectrometry, was diluted 100-fold (10 ml to 1 1 in a polypropylene calibrated flask) with sub-boiling distilled water and acidified to pH 2. The samples were then processed in the normal manner and the results are presented in Table VI. The results for direct analyses were in good agreement with the certificate values except TABLE VI ANALYTICAL RESULTS FOR TRACE ELEMENTS IN WATER SRM 1 6 4 3 ~ Means f 95% confidence levels. Amount found/ng ml-' Element Cadmium .. .. Iron . . .. Molybdenum . . Nickel . . .. Vanadium . . Zinc . . .. Copper . . .. Certificate* value/ ng g-l 10f 1 1 8 5 2 88f4 95f6 55f3 53f3 72f4 f Direct? 1 0 5 2 19&2 90&2 99f6 51 f 14 48&2 61 f 4 and pre-concentration: lO&l 19f2 106f 15 82f8 54&8 44&2 60f4 * To convert to nanograms per millilitre, multiply by specific gravity of SRM; the specific gravity at t Total number of measurements, 18. $ Results derived from analyses performed on two separate occasions: number of replicates (z.e., of 500-g 23 "C is 1.017 g ml-l. aliquots), 6; total number of measurements, 18. for zinc and therefore the accuracy of the prepared multi-standard solution (see Experimental) was confirmed.The disparity in zinc concentrations was significant at the 95% confidence level, the relative error being 15%. An independent ICP analysis also failed to reach agreement with the certified zinc value. The pre-concentration data were generally satisfactory but particularly so for cadmium, copper and nickel. Based on the previous recovery study the relatively low results for molybdenum (relative error 14%) and zinc were not unexpected.April, 1981 ICP SPECTROMETRY AFTER PRE-CONCENTRATION 427 A blank correction was not performed for iron in order to assess the magnitude of the error arising from contamination. The relatively poor precision for the iron determination is also revealed by the magnitude of the uncertainty limits.High precision was obtained for the remaining elements; it should be noted that the ICP measuring precision was the limiting factor rather than the precision associated with pre-concentration (cf., uncertainty limits for direct and pre-concentration data). Analysis of Open-ocean Samples The analytical results for a deep (4500 m) Pacific Ocean sample and a sub-surface Atlantic Ocean sample are given in Table VII. In view of the low metal concentrations expected, concentration factors of 250 and 500 were employed for the latter sample. The nickel concentration for the Atlantic Ocean sample was such that determinations for the 250-fold concentrate were close to the LLQ (42 pgml-l compared to 4Opg1n-l) and improved measurement precision was realised for the 500-fold concentrate.Even with the adoption of the maximum concentration factor, the cadmium concentration for the Atlantic sample was below the ICP detection capability and GFAA was required for quantification, the cadmium value being 4 ng 1-l. Agreement in the results for both sea-water concentrates confirmed the reliability of the final volumetric step in the procedure. Further, as mentioned earlier, although the nitric acid concentration for the 500-fold concentrate was approximately double that for the 250-fold concentrate the good agreement in data confirmed the absence of acid interference. TABLE VII TRACE METAL CONCENTRATIONS FOR OPEN-OCEAN SAMPLES See Experimental for sample description. Number of replicate analyses, 3 ; total number of measurements, 9.Results are in micrograms per litre. The values in parentheses are standard deviations. Atlantic r Element Pacific Cadmium .. .. 0.096 (0.004) Copper . . . . 0.348 (0.012) Iron .. .. 1.05 (0.10) Molybdenum . . 8.36 (0.58) Nickel . . . . 0.612 (0.032) Vanadium .. 1.41 (0.04) Zinc . . .. 2.30 (0.04) 1 Concentration Concentration factor factor of 250 of 500 < 0.02 0.392 (0.007) 0.544 (0.011) 7.80 (0.36) 0.168 (0.018) 1.33 (0.01) 0.556 (0.007) < 0.01 0.390 (0.012) 0.546 (0.026) 7.48 (0.25) 0.164 (0.006) 1.30 (0.02) 0.544 (0.01%) The data for the Pacific and Atlantic samples reveal a similarity in concentration levels for certain elements, and for cadmium, copper and nickel the relatively low concentrations are of similar magnitude to those reported in the study of Bruland et aL2 The agreement in the results for molybdenum and vanadium is an interesting finding considering the differences in geographical location and sample depth. Molybdenum and vanadium concentration levels for Japan Sea samples (sub-surface, 800 m and 3000 m) were also similar to the data in Table VII.In addition, it is of interest to note that a recent study17 on molybdenum and based on /3-naphthoin oxime pre-concentration and neutron activation analysis reported the molyb- denum concentration for Atlantic Ocean samples to be within a narrow range of 7.31-7.95 pg 1-l. An earlier study1* for the North East Atlantic Ocean reported dissolved molybdenum to be within the concentration range 9.1-13.0 pg l-l, whereas that of vanadium was 0.83- 1.57 pg 1-l.Unfortunately in this study, sampling bottles and storage containers for the Pacific and Atlantic samples were dissimilar (see Experimental) and, therefore, caution is required in evaluating data, particularly for zinc and iron, elements prone to c0ntamination.l Finally, it should be mentioned that the present method has been used in a recent trace metal inter- calibration study* designed to determine the optimum sampling procedure for deep-ocean water and it is hoped that much needed information on sampling and storage will be provided when the official document is published. * Trace Metal Intercalibration Exercise of Intergovernmental Oceaongraphic Commission, Bermuda, January 10-26th, 1980.428 MCLEOD, OTSUKI, OKAMOTO, HARAGUCHI AND FUWA Conclusion The present study has attempted to determine background levels of selected trace metals in open-ocean water, utilising dithiocarbamate pre-concentration and ICP emission spectrometry.Similarities for trace metal concentrations were demonstrated for Pacific and Atlantic Ocean samples, and cadmium, copper, nickel and zinc found at the nanogram per litre levels are con- sistent with recent independent analyses based on atomic-absorption spectrometry. Analyti- cal merits for the ICP method included simultaneous multi-element analysis, high precision, simplicity of calibration and high detection capability. The method blank was significant for iron and further work is needed to reduce the iron contamination. Although accuracy of the method has been tested, the possibility of contamination and/or adsorption losses associated with the sampling and storage of sea water should be kept in mind.Future work is to be directed at accumulating data for coastal and open-ocean areas and it is hoped that such studies can indicate the extent and trends of heavy-metal pollution. We are grateful to S. Fujiwara and Y. Nojiri (Chemistry Department, University of Tokyo) for the collection of Japan Sea and Pacific Ocean samples. M. Nishikawa (Analysis Section, NIES) deserves mention for his constant support during the analysis programme. Discussion at the trace metal analysis workshop of the Intergovernmental Oceanographic Commission was invaluable and the assistance and co-operation given by participants and staff of the Bermuda Biological Station is acknowledged. C.W.M. wishes to thank the Japanese Ministry of Education and Science (Monbusho) for financial support. 1. 2. 3. 4. 6. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Chakrabarti, C. L., Subramanian, K. S., and Nakahara, T., National Research Council of Canada, Bruland, K. W., Franks, R. P., Knauer, G. A., and Martin, J. H., Anal. Chim. Acta, 1979, 105, 233. Danielsson, L. G., Magnusson, B., and Westerlund, S., A&al. Chim. Acta, 1978, 98, 47. Kingston, H. M., Barnes, I. L., Brady, T. J., Rains, T. C., and Champ, M. A., Anal. Chem., 1978, 50, Smith, R. G., and Windom, H. L., Anal. Chim. Acta, 1980, 113, 39. Sturgeon, E. E., Berman, S. S., Desaulniers, A., and Russell, D. S., Talanta, 1980, 27, 85. Berman, S. S., McLaren, J. W., and Willie, S. N., Anal. Chem., 1980, 52, 488. Hiraide, M., Ito, T., Baba, M., Kawaguchi, H., and Mizuike, A., Anal. Chem., 1980, 52, 804. Patterson, C. C., and Settle, D. M., Nut. Bur. Stand. (U.S.) Spec. Publ., 422, 1976, 1, 321. Mattinson, J. M., Anal. Chem., 1972, 44, 1715. McLeod, C. W., Furuta, N., and Nishikawa, M., “Elemental Analysis of Pepperbush (Standard Reference Material) by ICP Emission Spectrometry,” in “The Preparation, Analysis and Certifica- tion of Pepperbush Standard Reference Material,” National Institute of Environmental Studies, Ibaraki, Japan, 1980. Publication No. 17530, 1979. 2064. Kinrade, J . D., and Van Loon, J. C., Anal. Chem., 1974, 46, 1894. Ward, A. F., ICP Inform. Newsl., 1977, 3, 90. Windsor, D. L., and Bonner Denton, M., Appl, Spectrosc., 1978, 32, 366. Brewer, P. G., in Riley, J. P., and Skirrow, G., Editors, “Chemical Oceanography,” Volume 1, Dabeka, R. W., Mykytiuk, A., Berman, S. S., and Russell, D. S., Anal. Chem., 1976, 48, 1203. Kulathilake, A. I., and Chatt, A., Anal. Chem., 1980, 52, 829. Morris, A. W., Deep-sea Res., 1975, 22, 49. Academic Press, New York, 1975, p. 415. Received July 271h, 1980 Accepted November 61h, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600419

出版商:RSC    

年代:1981

数据来源: RSC