Evaluation of the state-of-the-art of butyl- and phenyltin compound determinations in freshwater sediment prior to certification of a reference material F. Ariese,a W. Cofino,a J. L. Go�mez-Ariza,b G. Kramerc and Ph. Quevauvillerd aVrije Universiteit, Instituut voor Milieuvraagstukken, de Boelelaan 1115, NL -1081 HV Amsterdam, The Netherlands bUniversidad de Huelva, Departamento de Quý�mica, Campus Universitario de la Rabida, E-21819 Palos de la Frontera, Spain cInstitute for Reference Materials and Measurements, Retieseweg, B-2440 Geel, Belgium dEuropean Commission, Standards, Measurements and Testing Programme, 200 rue de la Loi, B-1049 Brussels, Belgium Received 12th October 1998, Accepted 29th January 1999 In order to control the quality of butyl- and phenyltin compound determinations in sediment, the Standards, Measurements and Testing Programme (formerly BCR) of the European Commission has started a project, the final aim of which is to certify a freshwater sediment for its contents of a range of organotin compounds (mono-, di- and tributyltin, and mono-, di- and triphenyltin). The first part of this project involved an interlaboratory study which aimed to test the feasibility of preparation of candidate freshwater sediment reference material and to detect and remove most of the pitfalls observed in organotin determinations.This paper presents the main results of this interlaboratory study carried out prior to the certification campaign. The agreement obtained among laboratories for the six compounds determined was considered to reflect the state-of-the-art and was encouraging enough to decide upon the organisation of a certification campaign, which will be concluded in June 1999. Introduction Preparation of the reference material The sediment material used in the interlaboratory study was Butyl- and phenyltin compounds, particularly trisubstituted species, are known to be very toxic to aquatic biota at very collected from a side arm of the North Sea Canal (northwest of Amsterdam, The Netherlands).First, a small batch was low concentrations.1,2 Tributyltin (TBT) is released in the aquatic (marine and freshwater) environment from the leaching collected to test the stability of organotin compounds after diVerent drying procedures, and part of this material was used of TBT-based antifouling paints used on boats and ships, whereas triphenyltin (TPhT) is used both as an antifouling for the interlaboratory study.One year later, a larger batch was collected at the same location for the preparation of the agent and a herbicide formulation; both compounds are present in the water phase and can be released from polluted candidate reference material.The wet material was dispatched to the Institute for Reference Materials and Measurements sediments and harbour dredges. The monitoring of tin compounds is required by the European Community (EC) legis- (IRMM) in Geel (Belgium) for further treatment. The wet material was dried at 60 °C in Teflon-protected stainless steel lation, e.g. under the Council Decisions 75/437/EEC (Marine Pollution from Land-based Sources), 77/585/EEC trays, crushed using a stainless steel jaw crusher and subsequently sieved (<1 mm).As the material was extremely wet, (Mediterranean Sea) and 77/586/EEC (Rhine River), and the Council Directive 80/68/EEC (Groundwater). Compliance to it was necessary to repeat the drying and crushing procedure several times. The moisture content of the sieved material this legislation and the need for a worldwide comparability of data require that the analyses are accurate. However, the (determined by Karl–Fischer titration) was less than 3.5%.Fine grinding of the sediment was carried out with the fraction methods currently used for the determination of organotin compounds involve various analytical steps, such as extraction, of less than 1 mmusing a jetmilling/ultrafine classification system based on the following principles.At the bottom of the grinding derivatisation, separation and final detection which multiply the risks of analytical errors.3 In order to improve and ensure chamber, three nozzles were mounted through which jets of air (6 bar) were blown; the three-dimensional nozzle arrangement a good quality control of tin speciation analysis, a series of interlaboratory studies (including certifications) has been enabled the feed material to be ground completely without residue.The air accelerated thematerial particles, which impacted organised in the past few years.4 The project described in this paper is a continuation of this programme; it aims to evaluate on each other and generated a size reduction without any contamination risks. Fine particles were collected through the classi- the state-of-the-art of butyl- and phenyltin compound determinations in freshwater sediment prior to a certification cam- fying wheel on top of the chamber.Coarse particles flowed back along the walls into the grinding chamber. The grinding process paign. Several drying procedures were tested in order to select the method that would lead to a suitably stable reference was performed with a classifier speed of 4000 rpm.Based on previous experience,4 heat sterilisation tests were material. In addition, a complete set of organotin calibrants was prepared, purified in large batches and made available to carried out at 100 and 130 °C under air to verify the final organotin content, i.e.to determine whether phenyltin all participating laboratories. J. Environ. Monit., 1999, 1, 191–196 191compounds were still present in measurable amounts in the final The excess of Grignard reagent was removed with 0.5 mol L-1 sulfuric acid, and the ether extract was reduced to 1 mL and material. Subsamples of the jet-milled sediment were heat sterilised in Teflon protected stainless steel trays (30 mm powder passed through a Florisil column.The organotin compounds were eluted with pentane, concentrated to dryness under a layers) for 2 h and dispatched to the Institute for Environmental Studies (Amsterdam, The Netherlands) for analytical control. nitrogen stream and dissolved in a suitable solution containing the internal standard.The quantification was by GC-FPD. The organotin contents of the sediment treated at 60, 100 and 130 °C are presented in Table 1. The heat sterilisation at 100 °C The variations in amounts of organotin in the sediment under the diVerent storage conditions are depicted in Fig. 1. All the was considered to be the most adequate treatment for the whole batch, i.e.a suYciently high temperature to sterilise the sediment organotin species were found to be stable in samples stored at -20 °C over the entire duration of the experiment (540 d); (which would not be likely at 60 °C), but not modifying drastically the organotin pattern by degradation (which ismuch higher this result confirms previous studies which demonstrated the stability of organotins in frozen sediments.6–8 Other storage at 130 °C for some compounds, e.g.TPhT). Homogenisation was carried out in a multipurpose cone temperatures (+4 and+25 °C) are also possible for maintaining the integrity of the sediment material during the same period of mixer with semi-automatic filling equipment. After an additional treatment of 2 h at 100 °C, the powder was directly time; however, a temperature of +40 °C drastically aVects the stability of the sediment, and significant losses of TBT were introduced into the cone mixer by a central filling nozzle on top of the mixer.All drives and gearing were placed outside noted after 365 days of storage (see Fig. 1). It was decided to store the material at -30 °C to ensure maximum long-term the mixing chamber so that contamination by oil was excluded.All parts of the mixer in contact with the powder were made stability. The material is suYciently stable to allow for shipment at ambient temperature without the risk of degradation. of polished stainless steel. A feed screw, 100 mm from the bottom of the mixer, enabled the filling of the vials without stopping mixing with given masses of the material.Directly Preparation of calibrants after filling of atle, the feeder turned to the opposite direction to push the powder back into the cone mixer. The purpose of this subproject was to prepare highly purified butyltin and phenyltin compounds (in the form of chlorides), Samples were bottled in amber glass bottles with a polyethylene insert and a screw cap, each containing about 40 g of material.as well as their ethylated and pentylated derivatives, for use as recovery standards and calibration standards. The purity Finally, a shrink sleeve was sealed on top of the bottle around the screw cap. The material was stored at -30 °C. of commercially available organotin compounds is often not adequate, while the alkylated derivatives usually need to be prepared in-house.It was felt that, by providing ultrapure Homogeneity and stability checks standards synthesised and purified in large quantities by an expert laboratory, some errors due to impure calibrants could The homogeneity of the sediment reference material was verified by determining the analytical variability (10 determi- be avoided. The calibrants prepared were mono-, di- and tributyltin and mono-, di- and triphenyltin; these six com- nations of the same extract), the within-bottle variability (10 independent determinations from the same bottle) and the pounds were made available as chloride salts, as pentyl derivatives and as ethyl derivatives. Commercial organotin salts were between-bottle variability (45 independent determinations, corresponding to 3% of the total number of bottles).The tests used as starting materials; purification was carried out by a number of recrystallisation and/or distillation steps. The alkyl- did not reveal any significant inhomogeneities, i.e. the withinand between-bottle variabilities were in the same range as the ated derivatives were prepared via a Grignard reaction using the purified organotin chlorides as starting materials.The analytical variability. One of the most critical aspects of the preparation of a resulting products were again purified by repeated recrystallisations and/or distillations. The purity was checked by elemen- reference material is related to the risks of instability, and a stability study was therefore carried out at various tempera- tal analysis, by 1H and 13C NMR and, in the case of the alkylated derivatives, also by GC-MS.For all 18 compounds, tures. Four aliquots of the heat sterilised material were stored in the dark at -20 °C, +4 °C, +25 °C and +40 °C, and the purity was better than 98%. The organotin chlorides were distributed as pure com- analyses were performed in triplicate after 30, 60, 90, 120, 180, 360 and 540 days of storage.The sediment was pre-treated pounds; the ethylated and pentylated standards were dissolved in hexane and sealed in ampoules. The nominal concentrations with 50 mL of a solution of 151 water–hydrogen bromide, and the organotins were extracted with 50 mL of 0.04% (w/v) were calculated by weighing. It could be shown, using propagation error theory, that the uncertainty in the calibrant tropolone solution in dichloromethane for 2 h.The organic phase was dried with anhydrous sodium sulfate and reduced concentrations was dominated by the uncertainty in the material purity (1%). Other possible sources of error, such as to 1 mL. The solid phase was washed with 5 mL of hexane, then added to the dichloromethane extract and reduced again weighing errors, were found to be insignificant.Participants were instructed to transfer the hexane solutions from the to 1 mL to substitute the solvent for final derivatisation.5 The organotins were derivatised with 4 mL of 1 mol L-1 pentylmag- ampoules to screw-cap vials and to store them in the refrigerator after opening. Participants were also informed that the nesium bromide solution in ether for 1 h at room temperature.Table 1 Organotin results in subsamples of the freshwater sediment resulting from pasteurisation experiments. Organotin contents are expressed in mg kg-1 (as Sn) dry mass Air dried Sn species 60 °C 60° C+2 h at 100 °C 60° C+2 h at 130 °C MBT 270 300 330 320 290 310 DBT 170 250 250 250 220 260 TBT 460 430 340 340 100 110 MPhT 120 93 87 86 59 74 DPhT 38 34 28 30 28 37 TPhT 41 100 21 28 8 21 192 J.Environ. Monit., 1999, 1, 191–196Fig. 1 Stability of organotin species in air-dried sediments followed by pasteurisation under diVerent storage conditions (in the dark at various temperatures). The symbols and bars correspond to the mean and standard deviations of organotin compounds obtained from triplicate analyses.pure salts had been stored under nitrogen and should be summarises the diVerent methods used by these laboratories for the diVerent compounds. Calibration was either by cali- handled under nitrogen, especially phenyl SnCl3 which is very sensitive to moisture in the air. bration graph or standard additions. Samples of one of the test batches produced during the feasibility study were dispatched on dry ice to the participating Analytical techniques used in the interlaboratory laboratories.Pure organotin chloride calibrants were also study provided, as well as ethylated or pentylated derivatives for calibration and/or recovery measurements. Each laboratory Sixteen laboratories from eight European states participated in the interlaboratory study (see Acknowledgements).Table 2 which participated in the exercise was requested to make a J. Environ. Monit., 1999, 1, 191–196 193Table 2 Summary of methods used in the interlaboratory study (extraction, derivatisation, separation and detection) Lab. 1: acetic buVer, methanol extraction, sonication, NaBEt4 derivatisation, GC-QFAAS Lab. 2: acetic acid addition, toluene extraction, microwave leaching, NaBEt4 derivatisation, GC-MIP Lab. 3: HCl addition, hexane/ethyl acetate/tropolone extraction, sonication, NaBEt4 derivatisation, GC-MIP Lab. 4: HCl/NaCl addition, ethyl acetate/tropolone extraction, fluorogenic derivatisation, HPLC-FLUO Lab. 5: addition of HBr/H2O/MeOH, dichloromethane/tropolone extraction, NaBEt4 derivatisation, GC-FPD Lab. 6: addition of HBr/H2O/MeOH, tropolone extraction, Grignard pentylation, GC-FPD Lab. 7a: acetic acid extraction, shaking overnight, NaBH4 derivatisation, GC-QFAAS Lab. 7b: acetic acid extraction, shaking overnight, NaBEt4 derivatisation, GC-FPD Lab. 8: addition of acetic acid/MeOH, hexane extraction, NaBEt4 derivatisation, GC-AED Lab. 9a: HCl addition, MeOH/tropolone extraction, sonication, Grignard pentylation, GC-MS Lab. 9b: HCl addition, MeOH/tropolone extraction, sonication, Grignard pentylation, GC-FPD Lab. 10: acetic acid extraction using microwave, NaBEt4 derivatisation, GC-MIP Lab. 11a: acetic acid/toluene extraction, sonication, Grignard pentylation, GC-MS Lab. 11b: addition of HCl, diethyl ether/tropolone extraction, sonication, Grignard pentylation, GC-MS Lab. 12: addition of acetic acid, MeOH/hexane extraction, sonication, NaBEt4 derivatisation, GC-QFAAS Lab. 13: addition of MeOH, hexane extraction, NaBH4 derivatisation, GC-FPD Lab. 14: acetic acid/toluene extraction, sonication, Grignard pentylation, GC-FPD Lab. 15: acetic acid/tropolone extraction, dichloromethane back-extraction, sonication, post-column NaBEt4 derivatisation, HPLC-ICPAES Lab. 16: HBr/H2O/dichloromethane extraction, tropolone back-extraction, Grignard pentylation, GC-FPD AED, atomic emission detection; HPLC, high performance liquid chromatography; GC, gas chromatography; FLUO, fluorimetry; FPD, flame photometric detection; ICPAES, inductively coupled plasma emission spectrometry; QFAAS, quartz furnace atomic absorption spectrometry; MIP, microwave inductively coupled plasma atomic emission spectrometry; MS, mass spectrometry.minimum of five replicate determinations on at least two some improvements are deemed necessary, the participants considered that certification would be feasible. diVerent days. The results submitted in the interlaboratory study were discussed amongst all participants in a technical The results of dibutyltin (DBT) were subject to systematic errors due to, for example, poor separation (Lab. 15), poor meeting; they were presented in the form of bar graphs showing the laboratory codes and the methods used, the individual extraction (Lab. 1), diVerence in behaviour of spiked and non-spiked compounds during standard addition leading to means and standard deviations and the mean of laboratory means with its standard deviation; Fig. 2 gives an example of low results (Lab. 16), and possible underestimation of the hydride generation response leading to overestimated contents such a bar graph. Table 3 summarises the results obtained in the interlaboratory study. (Lab. 7). The same RSD range as observed for MBT was obtained (Table 3), with similar conclusions on possible certification.Intercomparison results The situation for TBT mirrored that observed for DBT and the same rationale was applied to the results from Labs. 1 and In the case of monobutyltin (MBT), two groups of results were apparent. Lab. 10 had used microwave-assisted extrac- 16. The spread of the results (between-laboratory RSDs of 38% for all data and 27% for selected sets) was felt to be tion, which apparently resulted in a higher extraction eYciency, thus explaining the high results. Lab. 12 also produced a somewhat disappointing in comparison with previous interlaboratory studies (e.g. certification of CRM 462). Participants higher result, although a diVerent extraction procedure had been used, and no explanation could be found. Lab. 9 sus- were recommended to carefully look at their methods prior to certification and to thoroughly follow quality control pro- pected its high results to be due to a pre-extraction during the spiking process; the three sets of data were withdrawn on the cedures to make sure that a better agreement will be obtained at the certification stage.basis of the doubts expressed. The RSD between laboratories ranged from 57% (raw data) to 26% (selected values); while For the three phenyltin compounds, Lab. 9 experienced Fig. 2 Example of bar graph used for the evaluation of the results of the interlaboratory study (MPhT in mg kg-1). The laboratory codes are indicated along with the methods used (abbreviations defined in Table 2). The results plotted correspond to five replicate determinations. M is the mean of laboratory means. 194 J.Environ. Monit., 1999, 1, 191–196Table 3 Summary of the results of the interlaboratory study. The results are given in mg kg-1 of the respective cations Number Number of Mean±s RSD (%) RSD (%) Compound of sets sets accepted (accepted sets) raw data accepted data MBT 16 11 154±40 57.0 25.9 DBT 18 14 540±138 51.8 25.6 TBT 19 17 1283±348 38.0 25.2 MPhT 11 9 56.0±10.3 38.0 18.4 DPhT 9 8 51.2±11.7 82.7 22.8 TPhT 12 10 80.6±19.1 43.1 23.7 instrumental problems. Other problems were noted for Labs. 3 ment of spiked samples along with candidate reference materials was found to be an excellent means to evaluate extraction and 7 which withdrew their results. With the exception of these obvious outliers, the agreement was considered to be recoveries on the same basis. reasonable (ranging from 18 to 24% for the three compounds) and promising with respect to future certification.Conclusions One of the most critical aspects to be considered in the Technical discussion preparation of a possible CRM is the long-term organotin stability. Previous experience with sediment (CRM 462)10 and A systematic comparison of the extraction methodologies demonstrated that a high range of recoveries were obtained, mussel tissue (CRM 477)11 demonstrated the diYculty in stabilising reference materials with respect to their organotin depending on the types of reagent and procedure.9 The six organotin compounds considered in this study do not behave contents.In particular, long-term storage at +4 °C was not suYcient to preserve the butyltin compounds in CRM 462, in the same way for diVerent extraction methods and it is clear that there is no ‘universal’ method which can be optimised for which required re-certification;12 in the case of CRM 477, the stability of the butyltin contents was achieved at -20 °C, but all compounds; in other words, no firm recommendation on the use of one single method was given and participants were phenyltins were not found to be stable under any of the conditions tested.11 Participants recommended that some free to use their own method provided that it was properly validated.Clean-up may result in the degradation of com- samples of the candidate CRM be deep frozen (at a temperature of -70 °C or lower) to enable future long-term stability pounds depending on the conditions employed, and this risk should be faced.checks of the CRM (which will be stored at -30 °C). It was stressed that, while heat sterilisation will eliminate bacterial Some comments on the extraction methods were made by the participants: (i) it was noted that HCl concentrations action, the change of temperature when handling the material (from refrigerator to ambient temperature) may be quite above 1 mol L-1 lead to the degradation of trisubstituted tin species; (ii) extraction with glacial acetic acid, followed by critical for its long-term stability, and this should be stressed in the recommendation of the use of the future reference back-extraction into toluene of deep frozen samples, spiking with isotopically labelled tin and then analysis by HPLC- material.The stability of the MBT and MPhT chloride calibrants is another point of concern, and practical ways should ICPMS, indicated some degradation consistent with a molecular rearrangement; (iii) microwave-assisted extraction of sedi- be developed to prevent degradation. With respect to the results of the interlaboratory study, the ments could lead to the rapid degradation of TPhT unless precautions were taken to avoid the high temperature ramp participants agreed that this range of RSDs corresponded to the state-of-the-art and recommended the organisation of the eVect. Derivatisation is another critical step in tin speciation, and certification campaign. A candidate CRM of freshwater sediment (CRM 646) was prepared in June–October 1997; it was a systematic study was carried out to test the influence of reaction temperature, reagent concentration, reaction time and shipped to the participating laboratories in January 1998.The certification project was conducted during the first half of shaking mixture.6 The reagent concentration was found to be a critical factor governing side reactions, whereas the length 1998 and the results were discussed in September 1998.The results of the certification will be discussed by the ‘BCR of the alkyl substituents played a role in the derivatisation yield, particularly for the less substituted tin compounds, due Certification Committee’ who will report on the suitability of the material as a CRM in June 1999; provided that this possibly to volatilisation. Pre-concentration to near dryness gave rise to variations between substituted forms, and should material is accepted, the new CRM could be on the market by the end of 1999.be avoided when short chain alkyl derivatives are used (cf. pre-concentration to 0.3 mL). Ethylation via Grignard reaction gave slightly better results than the use of sodium Acknowledgements tetraethylborate.General observations on the methods were as follows: This project was carried out under EC contract nr. MAT1-CT94–0071. The collection, preparation, homogeneity GC-AED provided an excellent separation, good sensitivity and flat baselines; GC-MS oVered a good separation and peak and stability studies of the material used in the interlaboratory study were carried out by the Institute for Environmental shapes; GC-FPD showed a poor baseline performance that could not be corrected; GC-AAS gave a poor sensitivity but Studies, Free University of Amsterdam (The Netherlands).The following laboratories participated in the interlaboratory was otherwise satisfactory; LC provided variable peak shapes and potential overlap problems leading to unresolved compo- study: CID-CSIC, Departamento de Quý�mica Ambiental, Barcelona (Spain); ENEA, Divisione di Chimica Ambientale, nents/interferences; polarography oVered a limited sensitivitand selectivity. Rome (Italy); GKSS-Forschungszentrum, Geesthacht (Germany); Kernforschungsanlage, Ju� lich (Germany); RIKZ, With respect to extraction recoveries, it was decided that a protocol would be drawn up for possible guidelines.However, Haren (The Netherlands); MAFF, Burnham-on-Crouch (UK); Technical University of Athens (Greece); Tin Research the participants recommended that laboratories be free to adopt their own (documented) spiking procedures; the ship- Institute, Uxbridge (UK); Universidad de Huelva, J. Environ. Monit., 1999, 1, 191–196 195Speciation in Bioinorganic Chemistry, ed.S. Caroli, John Wiley, Departamento de Quý�mica Analý�tica (Spain); Universitaire New York, 1996, p. 195. Instelling Antwerpen, Analytische Scheikunde, Wilrijk 4 Ph. Quevauviller, M. Astruc, L. Ebdon, H. Muntau, W. Cofino, (Belgium); Universidad de Barcelona, Departamento de R. Morabito and B. Griepink, Mikrochim. Acta, 1996, 123, 163. Quý�mica Analý�tica (Spain); Universita` di Genoa (Italy); 5 J.L.Go� mez-Ariza, E.Morales, I. Gira�ldez and R. Beltra�n, Int. J. Universite� de Bordeaux, Laboratoire de Photophysique et Environ. Anal. Chem., 1997, 66, 1. 6 P.H. Dowson, J. M. Bubb and J. N. Lester, Estuar. Coast. Shelf Photochimie Mole�culaire, Talence (France); Universite� de Pau Sci., 1996, 42, 551. et des Pays de l’Adour, Laboratoire de Chimie Analytique 7 Ph. Quevauviller and O. F. X. Donard, Fresenius’ J. Anal. Chem., (France); Universidad de Santiago de Compostela, 1991, 339, 6. Departamento de Quý�mica Analý�tica (Spain); Vrije 8 J.L.Go� mez-Ariza, E. Morales, R. Beltra�n, I. Gira�ldez and M. Universiteit, Instituut Milieuvraagstukken, Amsterdam (The Ruiz-Bený�tez, Quý�m. Anal., 1994, 13, S76. 9 M. B. De la Calle, R. Scerbo, S. Chiavarini, Ph. Quevauviller and Netherlands). R. Morabito, Appl. Organomet. Chem., 1997, 11, 693. 10 Ph. Quevauviller, M. Astruc, L. Ebdon, V. Desauziers, P. M. Sarradin, A. Astruc, G. N. Kramer and B. Griepink, Appl. Organomet. Chem., 1994, 8, 629. References 11 R. Morabito, P. Soldati, M. B. de la Calle and Ph. Quevauviller, Appl. Organomet. Chem., 1994, 12, 621. 1 R. J.Maguire, Appl. Organomet. Chem., 1987, 1, 475. 12 A. Lamberty, Ph. Quevauviller and R. Morabito, EUR Report, 2 Ph. Quevauviller and O. F. X. Donard, in Element Speciation in EN 18406, EC, Brussels, 1998. Bioinorganic Chemistry, ed. S. Caroli, John Wiley, New York, 1996, p. 331. 3 Ph. Quevauviller, E. A. Maier and B. Griepink, in Element Paper 8/07886F 196 J. Env