Determination of tin and triorganotin compounds in sea-water by graphite-furnace atomic-absorption spectrophotometry

0039-9140/89 $3.00 + 0.00

Tdanra, Vol. 36, No. 4, pp. 513-517, 1989 Printed in Great Britain. All rights reserved

Copyright

0 1989 Pergamon Press plc

DETERMINATION OF TIN AND TRIORGANOTIN COMPOUNDS IN SEA-WATER BY GRAPHITE-FURNACE ATOMIC-ABSORPTION SPECTROPHOTOMETRY T. FERRI, E. CARDARELLI and B. M. PETRONIO Dipartimento di Chimica, Universita di Roma “La Sapienza”, Piazzale Aldo Moro, 5-00185, Rome, Italy (Received 26 January 1988. Revised 8 June 1988. Accepted 7 November 1988)

Snnunary-An analytical method based on graphite-furnace atomic-absorption spectrophotometry employing a suitable signal-enhancing medium for determination of inorganic tin and two of its trisubstituted organic derivatives in sea-water has been established. This method allows determination of triphenyltin and tributyltin compounds down to 2 x lo-‘* and 2.8 x 10-‘2M respectively by means of enrichment by collection on graphitized carbon black (enrichment factor up to 8 x 104) and a separation on a small silica-gel column. Inorganic tin, which is not adsorbed on the graphitized carbon black, is isolated from the matrix by liquid-liquid extraction of its pyrrolidinedithiocarbamate complex into dichloromethane. The method gives good recovery (295%) and precision (65%) at the rig/l.. level.

The determination of organotin compounds in the environment is becoming increasingly more important as their production increases.‘*Z Their use is strictly related to their degree of substitution.3 In particular, trisubstituted derivatives are widely employed because of their biocidal properties as algicides, fungicides, molluscicides etc.*” The introduction of such compounds into the environment could cause serious problems mainly due to their high toxicity and tendency to bio-accumulation.* In particular, bioassays with algae,5 oysters,6 crabs,‘** mussel larvae’ and fishi”*” have shown sublethal and lethal effects of tributyltin at levels even lower than 1 pg/l. In addition, these compounds can be accumulated by some organisms such as oyster’* and fish” with factors up to 6000 for rather short exposure periods. However, as these organotin compounds in the environment are degraded into less toxic onesI at different rates, it is important to have highly sensitive (ng/l, level) analytical methods available for their determination, in order to control their concentration levels in the environment so as to prevent health risk. Several techniques have already been employed to determine tin and/or its organic derivatives, including gas chromatography,15*‘6 HPLC,“*‘* polarography*%** and especially atomic-absorption spectrophotometry.*‘*r The objective of this work is to define an analytical method for determination of inorganic tin, tributyltin and triphenyltin in sea-water. The method, suitable for analyte determination at rig/l.. concentration levels, should be very simple and require only the usual laboratory equipment and a graphite-furnace atomicabsorption spectrophotometer. EXPERIMENTAL

Apparatus

A Perkin-Elmer model 2380 atomic-absorption spectrophotometer equipped with a Perkin-Elmer model RlOO 513

recorder, a deuterium-arc background corrector and a Perkin-Elmer model HGA 400 heated erauhite atomizer was employed. Perkin-Elmer untreat& -graphite and pyrolytically-coated furnace tubes were used, but for no more than 100 determinations each. Two GFAAS thermal programmes were utilized: the first, for inorganic tin, reported in Table 1, was defined on the basis of a preliminary study to determine the optimal ashing and atomization temperatures and the second, used for organotin compounds, was that defined by Pine1 et ~1.~~ Reagents

A 1 g/l. “Spectrosol” BDH standard solution for AAS was used as the primary standard solution for inorganic tin. All the less concentrated standard solutions obtained from it by dilution were made 1M in hydrochloric acid. Standard solutions of both triorganotin compounds were prepared with Fluka products (> 98% pure), and those of the di- and mono-organic derivatives were prepared with Aldrich products, all without further purification. The solutions (200 mg/l.) taken as standards were prepared by weight [triphenyltin chloride (TPT) and dibutyltin dichloride (DBT)] or by volume (and density) [tributyltin chloride (TBT) and monobutyltin trichloride (MBT)], in ethanol as solvent. The concentrated standard solutions were stored at 4”, and the dilute ones were prepared daily. All the acids employed were “Suprapur” Merck products. All the other reagents were Carlo Erba products of analytical grade except for the ammonium pyrrolidinedithiocarbamate (APDC) which was from BDH. Ultrapure demineralized water (18 Mn .cm) from a Mini-Q Milhpore system was always used. Artificial sea-wate? with a salinity of about 36 g/kg was prepared from demineralized water and salts in the follow-

Table 1. Thermal programme utilized for inorganic tin; the Ar flow (50 ml/min) was stopped during the atomization step Step Drying Drying Ashing Atomization Cleaning

Time, Temperature, set “C 10 5 10 :

90 120 550 2100 2700

Heating rate, deg lsec 4.5 2::: max max

T. FERRIet al.

514

ing molar concentrations: NaCl 0.4106, MgCl, 0.0298, M&SO, 0.0284, KC1 0.0093. Graphitized carbon black (GCB), Carbo-Pack B (Sunelco) 80-100 mesh. with 100 m2/a specific area was used for ‘preconcentration of organotin co£s, and silica gel (70-230 mesh, ASTM, Merck) was used to prepare the column for their chromatographic separation. Procedures Sample treatment and storage. The collected samples were filtered through 0.45~pm acetate filters, acidified with 5 ml of concentrated nitric acid per litre and stored in polyethylene bottles at 4”. Preconcentration. GCB (100 mg, previously washed with methanol) was put in a glass column (bore 7 mm) provided with a porosity-2 frit at one end (Bio-Rad). After passage of 5 ml of demineralized water to remove the methanol, a fixed volume (up to a few litres) of sample was passed through the column at a flow-rate of 20 ml/mm, followed by 5 ml of demineralized water at the same flow-rate, to eliminate the matrix salts. The triorganotin compounds were then eluted with 2 ml of methanol/dichloromethane mixture (4: 1 v/v) at 2 ml/min flow-rate. The GCB column, after washing with another 5 ml of demineralized water was then ready for further use. The eluate containing the triorganotins was evaporated to dryness under a stream of nitrogen and the residue was taken up in 1 ml of n-hexane. Separation. The chromatographic column was prepared with 200 mg of silica gel between two frits in a Supelco column (bore 4 mm). First 20 ml of n-hexane were passed through the column (without draining it) to eliminate air bubbles. The 1 ml of hexane solution from the preconcentration step was then passed through the column at 0.5 ml/min flow-rate. The TBT was eluted with 3 ml of nhexane/ethyl acetate mixture (2: I), and then with TPT 4 ml of ethyl acetate, at 0.5 ml/mm. Liauid-liauid extraction of SnflVJ The samnle (200 ml) was mixed in a 250-ml separatory funnel with‘2 ml of lk acetate buffer (pH 4.75) and 5 ml of 10% APDC solution. After 10 min the tin-APDC complex was extracted by shaking for 2 min with 10 ml of dichloromethane; the mixture was then let stand for 10 min for phase separation. Spectrophotometric measurements. For all atomicabsorption measurements a 0.04% potassium dichromate/ 2% nitric acid mediumz3 was used. The eluate from the chromatographic separation and the organic phase from the extraction step were evaporated to dryness under a stream of nitrogen and the residues were taken up in a suitable volume (usually 1 ml) of the acid dichromate solution. All measurements were made on 20-/r] samples introduced into the furnace-tube by means of a high-precision Gilson micropipette. Each value used was the average of at least five

different readings corrected for the blank absorption. In accordance with the recommendations of the ACS Committee on Environmental Improvement the signals used in the calculations were always at least ten times the standard deviation of the blank signal (i.e., the limit of quantification). Unless otherwise specified, the values reported correspond to the average of four separate determinations, and refer to the tin content.

RESULTS AND

DISCUSSION

Atomic-absorption determination of metals is particularly sensitive when electrothermal atomization is used.2s The direct determinations in real matrices, however, are often subject to interference by matrix components, especially in determinations at ultratrace levels. For tin determination, in particular, problems may arise during both the ashing and atomization steps, from formation of volatiles and interaction of tin with the carbon of the furnace wallsz6 Several approaches have been made into the solution of these problems, including isolation of tin from the matrix,“.2632 addition of a matrix modifier and pretreatment of the furnace to the sample,23~26*3S36 walls.37-39Matrix modifiers should essentially prevent the tin from being lost by volatilization before it can be atomized, thus decreasing interference problems, but may also be used simply to obtain a signal enhancement.24,N However, in the present case, the enrichment step should also avoid interference problems. A short preliminary study was made to optimize some of the experimental conditions (kind of furnace, wavelength, medium). These results, summarized in Table 2, show that non-treated furnace tubes gave higher sensitivity than pyrolytically-coated ones. The sensitivity was also higher for the 224.6 nm line than the 286.3 mn line. The acid dichromate mediumz3 was used whenever possible, as it enhanced the signal for all the analytes considered. TO allow speciation of the tin, at least two preconcentration steps were required, one for inorganic tin and one or more others for organic forms of tin.

Table 2. Evaluation of the influence of some experimental parameters on the recorded instrumental signal Analyte SnCl, SnCl, S&I, SnCl, SnCl, SnCl, TPT TPT TBT TPT

Medium Demineralized water Demineralized water 1M HNO, 1M HNO, 0.04% K,Cr,O, 2% HNO, 0.04% K&O, 2% HNO,

l Demineralized water 0.04% K,Cr,O, 2% HNOJ { Demineralized water 0.04% K&O, 2% HNO, c

Sensitivity,

absorbance x 10’

Correln. coeff.

1.51 1.52 2.85 1.70

0.7 -0.5 -0.6 0.3

0.9981 0.9997 0.9988 0.9987

Gr

3.31

-0.2

0.9996

Gr

3.06

-0.5

0.9993

-0.1

0.9996

1, nm

Tube

224.6 224.6 224.6 224.6

Gr* Py* Gr Py

224.6 286.3

absorbance , ml.

Y-intercept,

pgg-’

224.6

Gr

3.06

224.6

Gr

3.18

0.0

0.9991

224.6

Gr

1.73

1.0

0.9993

224.6

Gr

2.84

*Gr = graphite tube; Py = pyrolytically-coated

tube.

-0.0

0.9981

Tin compounds in *a-water

Table 3. Reconcentration of inorganic tin from 200 ml of artificial sea-water by liquid-liquid extraction of its APDC complex into dichloromethane at pH 4.75 S&l, found, W 3.9 f 0.3 19.7 f 0.5

SnCl, added, n8 &?3

RSD, %

Recovery, %

7.7 2.6

98 98

For preconcentration of inorganic tin, co-precipitation with iron,” magnesiumgl or calcium,42 and liquid-squid extraction with cupferronr tropolone” or APDC4” are most frequently employed. Since sea-water contains a sufficient amount of magnesium, co-precipitation with this metal was tested on artificial sea-water spiked with tin. Unfortunately, the high magnesium content in the sample thus enriched strongly interferes in the tin dete~ination. Therefore, liquid-liquid extraction of the tin-APDC complex at pH 4.75 into dichloromethane45 was adopted. The results in Table 3 show that practically complete recovery of inorganic tin is obtained. Triorganotin compounds have usually been pre~n~ntrat~ from aqueous samples by liquid-liquid extraction with’sJs~4s or without22,24,3G39 complexing agents. Toluene is one of the most commonly employed extractants, and dichloromethane also extracts these compounds. 39Toluene and di-isopropyl ether were tested. The ether does not extract the test compounds su~ci~tly, and even though toluene extracts both t~organotin compounds practically quantitatively,qb their total content cannot be directly determined by extraction because of the different AAS sensitivities for the two species.46 At least one further step is necessary; either mineralization and determination of the resulting inorganic tin to give the sum of TBT and TPT, or a ~hromatograp~c separation. In this connection, the use of GCB seemed very promising for giving preconcentration and separation at the same time, since it had already been widely employed for chromatographic column preparation:” and preconcentration of pesticides from natural water4’ and of metabolites or catabolites from biological fluids.‘s5’ The complete retention of TBT and TPT by GCB was verified by passing 1 litre of 40 pg/l. TBT or TPT solution through it at 20 ml/min flow-rate, and testing the eluate for them. The maximum adsorption capacity (break-through) of GCB for the two organotin compounds was not determined, because of

515

their extremely low concentration levels in sea-water. It was considered much more important to find and verify the lowest concentration levels at which these compounds may be determined by retention on GCB and subsequent elution and AAS dete~nation, It must be stressed that the presence of inorganic tin in the sample does not interfere with the determination of the organotin compounds by this method, since it is not retained by GCB and thus can be separated from the organic forms. The individual TBT and TPT elution curves with different eluents (methanol, acetonitrile and acetone) were determined. Although acetone and acetonitrile eluted TBT faster than TPT, and methanol eluted both simultaneously, the elution was not quantitative (Table 4), and other eluents were tested. A 2: 1 v/v mixture of methanol and dichloromethane was found to be the best. Table 5 gives the data for the preconcentration of TBT and TPT (at two concentration levels) from both demineralized water and artificial sea-water with this eluent. To evaluate the lowest concentrations that can be determined increasing volumes of sample (in 1-litre steps) but always containing 1 ng (as total tin) of TPT or TBT were analysed by the procedure reported above, except that a final volume of 50 ~1 was used, which allowed only two spectrophotometric readings. The results, together with the relative factors, are reported in Table 6. Determination of TPT and TBT down to 0.25 and 0.33 rig/l. respectively is feasible, but it must be underlined that for more si~ifi~nt results to be obtained larger volumes (>, 100 ~1) of acid dichromate should be used to dissolve the final residue so that at least four readings can be made. This also means that either the limit of quantification will be doubled, or larger sample volumes must be treated. Since the aim was dete~ination of the in~~dual analytes, a chromatographic separation of TBT and TPT after the preconcentration step was needed. Preliminary thin-layer chromatographic experiments showed that these two compounds could be separated on silica gel, with 2: 1 v/v n-hexanejethyl acetate mixture; in column work, however, the TPT (more strongly retained than TBT) gave a rather shallow and broad elution peak. It was therefore preferred to eluted the TBT with the solvent mixture, and then TPT with ethyl acetate alone. No interference is caused by any less substituted phenyltin or butyltin compounds present in the sample. Although they may be at least partially

Table 4. Quantitative recovery tests of TPT and TBT from GCB by selective elution TPT Volume, Retained,

Eluent CH$OCH, CH,CN CH,OH

El&d,

ml

a8

W

: 2

25.0 25.0 25.0

21.9 f 1.0

*Average of two determinations.

TBT W

Eluted, W

Recovery, %

21.8(a) 22.9 (a) 41.2& 1.1

50

% 44:o

Recovery, Retained,

% 88

z

516

T. FERRYet al. Table 5. Preconcentration

tests of TBT and TPT by GCB: eluent 500 ml of methanol/dichloromethane (2 : 1)

Medium

Theoretical, ng

Found, ng

RSD, %

Recovery, %

TPT TPT

Demineralized water Demineralized water

25.0 12.5

24.7 f 0.2 12.4 f 0.2

0.8 1.5

99 99

TBT

Demineralized water

22.0 44.0

21.7 + 0.7 43.5 0.2

3.2 0.5

zz

G TBT TBT

ASW(*) ASW(+) ASW(*) ASW(+)

25.0 12.5 44.0 22.0

24.5 12.3 + f 0.1 0.2 43.2 f 0.2 21.8 + 0.5

0.4 1.6 0.5 2.5

98 98 99

Analyte

+ASW = Artificial sea-water.

Table 6. Evaluation of the limit of quantification Analyte

Theoretical, ng

Volume, ml

TBT TBT TBT TBT TPT

1.0 1.0 1.0 1.0 1.0

z

z TPT

1.0 1.0

Found,* ng

Enrichment factor

1000

0.8 nd 1.0

2x 4x 6x 2x

104 104 104

4000 2000 5000

0.9 1.0 nd

4x 8 x 104 l@ -

104

*Average of two determinations. nd = not detected. Table 7. Recovery tests on sea-water samples (200 ml)

Analyte Sn(IV) TPT TBT

Natural content, ng 3.0 -

Added, Found, ng ng 12.5 15.1 kO.4 10.0 9.5 *0.5 10.0 9.5 *0.5

RSD, %

Recovery, %

2.6 5.3 5.3

97 95 95

retained by GCB (only -25% retention of the monosubstituted species) and co-eluted with the trisubstituted compounds by the methanol/ dichloromethane mixture, the TBT and TPT are selectively eluted in the subsequent chromatographic step. The elution of DBT and MBT (which are more

mobile than the corresponding phenyl derivatives) from the silica gel needs larger volumes of ethyl acetate than that needed for elution of TPT (the RF values of DBT and MBT are a half and a fifth, respectively, of that for TPT).

The method was applied to a sea-water sample taken from the Roman coast. The inorganic tin concentration in the sample was only 15.0 ng/ml, and neither TBT nor TPT was detected. The reliability of the method was checked by recovery tests on the same sample suitably spiked. The results summarized in Table 7 confirm the adequate performance of the method, especially considering the very low concentration levels determined. Acknowledgements-The work was done with the financial support of the Italian CNR. The authors are greatly indebted to Professor G. P. Cartoni and F. Coccioli (University of Rome) for very useful discussions and suggestions.

REFERENCES

1. A. G. Davies and P. J. Smith, Tin, in Comprehensive Organometallic Chemistry, International Tin Research Institute, Publication No. 618, London, 1982. 2. H. Vrijhof, Sci. Total Environ., 1985, 43, 221. 3. G. Bressa and L. Cima, Ambiente Rborsa Salute, 1985, 46,45. 4. P. J. Smith, Toxicological Data on Organotin Compour&. International Tin Research Institute, Publication No. 538, London, 1978. 5. G. E. Walsh, L. L. McLaughlan, E. M. Lores, M. K. Lonie and C. H. Deans, Chemosphere, 1985, 14, 383. 6. M. J. Waldock and J. E. Thain, Mar. Pollut. Bull., 1983, 14, 411. 7. R. B. Laughlin, W. French, R. B. Johannsen, H. E. Guard and F. E. Brinckman, Chemosphere, 1984, 13, 575. 8. R. B. Laughlin, Water, Air, Soil Pollut., 1983, 20, 69. 9. A. R. Beaumont and M. D. Budd, Mar. Pollut. Bull., 1984, 15, 402. 10. P. Seinen, T. Helder, H. Vemig, A. Penninks and P. Leenwang, Sci. Total Environ., 1981, 19, 155. 11. Y. P. Chliamovitch and C. Kuhn, J. Fish. Biol., 1977, 10, 575. 12. M. J. Waldock, J. Thain and D. Miller, Proc. Intern. Council Exploration Sea, CM. 1983/E:52. 13. G. S. Ward, G. C. Cramm, P. R. Parrish, H. Trachman and A. Slesinger, in Aquatic Toxicology and Hazard Assessment, D. R. Brason and K. L. Dickson @is.), p, 183. ASTM, Philadelphia, 1981. 14. A. V. Sheldon, in Recent Advances in Inorganotin Chemistry, A. G. Davis and P. J. Smith (eds.). p. 49. Academic Press, New York, 1980. 15. M. D. Milller, 2. Anal. Chem., 1984, 317, 32. 16. A. Woollins and W. R. Cullen, Analyst, 1984,109,1527. 17. K. L. Jewett and F. E. Brinckman, J. Chromatog. Sci., 1981, 19, 583. 18. L. Battini, A. Casoli, A. Mangia and G. Pedrieri, VI Congress0 Nazionale dell0 Divisione di Chimica Analitica, Bari, Italy, 24-27 September 1985. 19. T. M. Florence and Y. J. Farrar, J. Electroanal. Chem., 1974, 51, 191. 20. K. Hasebe, Y. Yamamoto and T. Kambara, 2. Anal. Chem., 1981, 53, 875. 21. G. Weber, Anal. Chim. Acta, 1986, l&i, 49. 22. P. Nangniot and P. H. Martens, ibid., 1961, 24, 276. 23. R. Pinel, M. Z. Benabdallah and A. Astruc, ibid., 1986, 181, 187. 24. E. S. Parks, W. R. Blair and F. E. Brinckman, Talanta, 1985, 32, 633. 25. W. Slavin. Anal. Chem.. 1982. 54. 685A. 26. E. Lundberg, B. Bergmark and W. Frech, Anal. Chim. Acta, 1982, 142, 129. 27. J. Marenger, J. Assoc. 08 Anal. Chem., 1975,58, 1143.

28. M. Tominoga and Y. Umezaki, Ano!. Chim. Acta, 197% 40. K. C. Thompson, R. G. Godden and D. R. Thornarson, Anal. Chim. Acta, 1975, 74, 389. 110, 55. 29. P. Hocquelht and N. Labyeria, At. Abs. Ned., 1977, 41. V. F. Wodp, S. L. Seidel and E. D. Goldberg, ha/. Gem., 1979, 51, 1256. la, 124. 42. A. G. Tybq and P. D. Bran@n, 30, H. L Trdnnin, z&J&*C&w., 1977,49* f@9& 43. N. H. l&m&, W.*B.-Mason and J. J. F&oh, &id., 31. K. Oh& and M. Suzuki, A& C&m, Acta, 1979, 1M, 1949, 2x, 132.5. 245. 32. H. Fritz&c& W. Wegscheider and G. Knapp, Talantcr, 44. H. A. Minama, T. Burger~Wiersmat, G. Vershis-deHaan and E. C. Geevers, &Won. Sci. Technol., 1978, 1979, 26, 219. 12, 288. 33. L. Zhou, T. T. Chao and A. L, Meier, ibid., 1984, 310 45 K. D. R&age and R. Bock, X. Anal. Gem., 1974,27@, 73. 337. 34 T. Mi. Yidrey, H. E. Hw&, G. Y. Z&r&on and C. f, 46. T. Feni,~pubjjsh~ data. Ram&w, Anai, C&em., 1980, 5% t743. 47. A. Di Corcia and A. Liberti, rfdo, Chromatosw.. 1976. 35. Y. Arakawa, 0. Wada and M. Manabe, ib%, 1983, Ss, 1901. 48. A. Bacahni, 0. Goretti, A. Lags& B. M. Petroniio and 36. S. Kojima, Analyst, 1979, 104, 660. M. Rotatori, Anal. Chem., 1980, 52, 2033. 37. D. R. Hanwn, T. J. Gilfoil and H. K-I.Hill, Jr., Anal.

C&m., 1981, 53,857. 38. R. Pine& M, Z. I&mum

arid M. Astruc, The F@h w5.S~ Cotyc,u~r~~j~ C~U~~~~~ C~~~r~ of ~r~~~ T”urand Ee& Padua, &aIy, 8-10 ~iern~ 1986. 39. C. J. Soderquist and D. C. Crosby, Anai. Gem., IW8, so, 1435.

49, F. Audreolini, A. Di Corcia, A. Lqan& R. Sampi and G. %qxx& Ci&. C&m., 1983,29,2075. 50. F. ~~~* F. Borra~A. Di Cwc-ia, A. Lagan&, R Samperi and G. Rap&, &i& 1984.3@, 742, 51. A. La8anit, G. D’Ascenzo, A. Mar&o and A. M. Tarole, bid, 1986, 31, 508.