摘要:

Analyst, July, 1979, Vol. 104, pp. 645-659 645 Determination of Noble Metals by Carbon Furnace Atomic-a bsorption Spectrometry Part 1. Atom Formation Processes W. B. Rowston and J. M. Ottaway Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow, G1 1 X L Evidence is presented to support the theory that the formation of atoms of noble metals during carbon furnace atomisation proceeds via direct evapora- tion of the metal. The evidence includes (i) thermogravimetric investigation of noble metal salts and their aqueous solutions in an argon atmosphere, (ii) X-ray diffraction studies of the residues formed on heating aqueous solutions of some noble metal salts in a carbon furnace atomiser and in the microfurnace of a thermobalance, (iii) measurements of the appearance temperatures of noble metals in a carbon furnace atomiser using aqueous solutions and vacuum-deposited metal films and (iv) activation energies, Eg, and vapour pressure data relating to noble metals at elevated temperatures.Experimental conditions giving the best sensitivity for the determination of seven noble metals (osmium was not detected) in the Perkin-Elmer HGA-74 carbon furnace atomiser are presented. On the basis of a 20-p1 sample volume these gave sensitivities (1% absorption) of 0.019, 0.0058, 0.0045, 0.00021, 0.038, 0.023 and 0.00089 pg ml-l for ruthenium, rhodium, palladium, silver, indium, platinum and gold, respectively. Keywords : Noble metal determination ; atomic-absorption spectrometry ; carbon furnace atomisation Numerous investigations concerning the use of flames as atom sources for the determination of noble metals" by atomic-absorption spectrometry have been rep0rted.l Most workers have described widespread interference effects on the atomic-absorption signals of noble metals.The extent of such interferences depends on the analyte element, the concentration of interferent, the nature of the counter ions in solution, the type of flame and the stoicheio- metry of the flame.2 Although releasing agents or high-temperature flames are often used successfully to minimise these effects, their causes have not been explained satisfactorily. During the past decade, various designs of electrothermal atomisers have been developed for use as atom cells in atomic-absorption spectrometry.3~4 Nickel, platinum, tungsten and tantalum are among the metals that have been used as atomiser materials but most commercial atomisers are fabricated from graphite either as a tube furnace or filament.The literature relating to carbon furnace atomisation of noble metals has dealt mainly with analytical applications and few studies of a mechanistic nature have been reported. Adriaessens and Knoop5 indicated the feasibility of determining iridium, platinum and rhodium in a glass matrix using a graphite furnace atomiser. The effects of increasing amounts of nitric acid and hydrochloric acid on the absorption signals for the three metals and some inter-element interferences were reported but the causes of the effects were not explained. Pearton and Mallett6 compared the use of a mini-Massman furnace, a NIM-rod7 and a carbon furnace atomiser for the determination of noble metals.Of the three atomisers, the carbon furnace device was found to be the most sensitive and precise, to have the lowest limits of detection and to be accompanied by the least interference. The authors reported depressive interference of the absorption signal for palladium in the presence of the other noble metals. To account for this signal depression, it was suggested that palladium volatilisation is suppressed by elements having high boiling-points and as such are involatile under the analytical conditions. Optimum conditions for the determination of some noble metals by carbon rod atomic- absorption spectrometry were described by Everett.8 The author also examined the inter- * For the purposes of this paper, silver is included and considered to be one of the noble metals.646 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, Vol.104 ference effects of the precious metals on each other at ratios from 1- to 100-fold mass excess of interferent over analyte. The effects were mainly depressive with the exception of palladium as analyte, with which, in contrast to the results above, the signal was found to be constant from 5- to 100-fold excess of other :noble metals. Aggett and West9 and GuerinlO noted a depressive interference on the atomic-absorption signal for gold in the presence of an excess of palladium. The former workers9 explained this effect as being due to a loss of gold atorris by vapour-phase reactions while GuerinlO postulated the formation of a gold - palladium intermetallic compound that retarded the volatilisation of gold.The mechanism of atom formation of noble metals using a graphite furnace atomiser has not been reported and explanations for the many observed interference effects have been speculative. The aim of the present study was to establish the mechanism of atomisation of the noble metals in a carbon furnace. In :Part I1 of this series some aspects of inter- element interferences for noble metals'will be considered in detail and the mechanisms of some of the interferences explained. Experimental Reagents Noble metals, noble metal salts and JM2204 graphite rod were of Specpure quality supplied by Johnson Matthey Chemicals Limited.All 'other reagents were of the highest available purity and glass-distilled water was used throughout for dilution purposes. This was prepared by dissolving 0.786 g of silver nitrate in 50 ml of 0.1 M nitric acid followed by dilution with water to 500 ml in a calibrated flask. Stock osmium solution, 1000 pg ml-l of osmium. This solution was prepared by extracting a 5-ml ampoule of 1% osmium solution with 6 ml of 0.1 M nitric acid followed by dilution with water to 50 ml in a glass-stoppered calibrated flask. Stock ruthenium, rlzodizlm, palladium, iridium, platinum and gold solutions, each 1000 pg ml-I of metal. Individual stock solutions were prepared by dissolving the appropriate amounts of the hydrated salts RuCl,, RhCl,, PdCl, (anh:ydrous), IrC13, H,PtCl, and HAuC1, in 50 ml of 0.1 M hydrochloric acid followed by dilution with water to 500 ml in a calibrated flask. Solutions of the same noble metal concentrations were also prepared in 0.01 M nitric acid using the same salts.Stock lead, cadmium, aluminium and zinc solutions, 1000 pg ml-l of individual metal. These were prepared by dissolving the appropriate amounts of each metal nitrate salt in 50 ml of 0.1 M nitric acid followed by dilution with water to 500 ml in a calibrated flask. Dilute solutions for calibration purposes were prepared as required by dilution of the stock solutions with 0.01 M acid of the corresponding counter ion. For the thennogravimetric study, stock solutions of each noble metal (except osmium) were prepared on. a semi-quantitative basis.Each solution contained about 0.05 g of noble metal salt in 1 ml of 0.01 M acid of the same anion. Stock silver solution, 1000 pg ml-l of silver. Apparatus Carbon furnace atomisation A standard Perkin-Elmer 360 atomic-absorption spectrometer equipped with an HGA-74 graphite furnace and a deuterium arc background corrector was used in conjunction with a Perkin-Elmer Model 56 strip-chart recorder. The HGA-74 atomiser has variable time and temperature (digits) selectors for sequentially drying, ashing and atomising samples. The device can also be heated more slowly using one of nine controlled heating rates up to the maximum furnace temperature of 2920 K. An Ircon automatic optical pyrometer was used to measure the temperature of the furnace wall at the sample entrance port at various digital settings.The results obtained were in good agreement with the temperature - digit correlation table supplied in the manufacturer's handbook, provided argon was used as the purge gas. The temperature - time profiles for heating rates 9 and 8 were also calibrated using the optical pyrometer.July, 1979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 647 Eppendorf micropipettes with disposable plastic tips were used for sample injections. Hollow-cathode lamps supplied by Pye Unicam (rhodium, palladium, platinum, gold, iridium and silver) and Perkin-Elmer (osmium and ruthenium) were used as spectral line sources. Thermogravimetric analysis The apparatus used to study the thermal breakdown pattern of noble metal salts and their aqueous solutions was a Stanton Redcroft TG750 thermobalance. This consists of a microfurnace and electronic balance, a programmer to control furnace heating at heating rates ranging from 1 to 100 Kmin-1 and a two-pen Speedomax strip-chart recorder.A switched range of sensitivities permits the use of 1-250-mg samples for full-scale deflection (f.s.d.) on the 10-mV recorder. For most of the test runs the sample in solid or solution form was contained in the 40-pl platinum crucible supplied by the manufacturer but a graphite crucible weighing about 150 mg, and machined from JM2204 graphite rod, was also used. Vacuum deposition apparatus deposit metal films on the inside wall of the HGA-74 graphite atomiser tubes. A General Engineering Co. Ltd. Genevac Vacuum evaporation apparatus was used to X-ray powder difraction analysis A Philips X-ray generator equipped with a Debye - Scherrer camera was used to record the X-ray diffraction patterns of residues formed on heating noble metal salts or their solutions in the HGA-74 carbon furnace atomiser or in the Stanton Redcroft TG750 micro- furnace.Procedures (a) Thermogvavimetric analysis To simulate reactions taking place in the carbon furnace, the thermal decomposition of noble metal salts and their aqueous solutions was investigated in the presence and absence of graphite in an argon atmosphere. The procedure involved taring the platinum crucible on the thermobalance through which argon flowed at 20 ml min-l. The noble metal salt was added to the crucible and the thermogravimetric curve recorded as the sample was heated at 50 I< min-1 from ambient temperature to the maximum available temperature of 1270 K.As the mass and composition of the starting compounds were known, mass losses were converted into formula mass losses (Fig. 2). Procedure TG2 (noble metal salts in the presence of graphite). This procedure was the same as procedure TG1 except that the platinum crucible was replaced with the specially machined graphite crucible referred to under Apparatus. Procedure TG3 (aqueous solutions of noble metal salts in the absence of graphite). The 4O-pl platinum crucible was tared on the microbalance at an argon flow-rate of 20 ml min-l and 30 pl of test solution, containing about 1.5 mg of noble metal salt in 0.01 M acid solution of the same counter ion, was transferred into the crucible using an Eppendorf micropipette.The furnace temperature was then increased to 350 K and maintained at this temperature until a nearly constant mass was observed on the 2.5 mg f.s.d. sensitivity scale. The thermogravimetric curve for the residue was then recorded as for procedure TG1. Lower maximum temperatures were used in the study of silver and gold residues. Procedure TG4 (aqueous solutions of noble metal salts in the preseme of graphite). About 1.5 mg of powdered graphite (ground from an HGA-74 atomiser tube) were weighed into the previously tared platinum crucible. The crucible plus graphite was then re-tared before adding 30pl of test solution. The subsequent heating programme was the same as that described in procedure TG3.Procedure TG4 was considered to be the best simulation of the conditions encountered in carbon furnace atomisation. However, as will be shown under Results and Discussion, it was the most difficult to quantify. Four sampling procedures were adopted, as follows. Procedure TG1 (noble metal salts in the absence of graphite).648 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, Vol. 104 (b) The geometric arrangement of the graphite tube, substrate holder and the tungsten coil heater of the vacuum evaporation apparatus are shown in Fig. 1. With the graphite furnace atomiser in the position indicated, 100 mg of a selected pure metal were placed in the tungsten heater before placing the glass dome in position on top of the vacuum sealing ring.The chamber was then evacuated using the rotary and diffusion pumps situated below the evapora- tion chamber. On passage of current through the tungsten heater the temperature was increased to about 2700 K and metal evaporation occurred. The resulting metallic vapour passed through the 1 mm diameter hole in the substrate holder, through the sample entrance port of the HGA-74 atomiser tube and was deposited over a small area on the opposite wall of the graphite tube. Vacuum deposition of metal jlms HGA-74 Carbon furnace tube Substrate ho I de r Tungsten conical - heater and metal evaporant Glass dome Vacuum - sealing ring -Steel base p date Power ’/ \., supply ‘To rotary and diffusion pumps Apparatus for deposition of metal films in vacuum.Fig. 1. (c) Optimum conditions with respect to analyte wavelength, hollow-cathode lamp current and spectral band pass were established in the usual manner to give the best sensitivity and these are given in Table I for seven noble metals. Three other variables associated with electrothermal atomisation were optimised using the following procedures. A tomic-absorption determination of noble metals ming carbon furnace atomisation TABLE I HGA-74 CARBON Parameter Waveleng th/nm Spectral band widthjim : Lamp current/mA .. . . Sample volume/wl* . . . . Temperature of sequential FURNACE ATOMIC-ABSORPTION DETERMINATION OF NOBLE METALS : OPTIMUM CONDITIONS .AND SENSITIVITIES Ru Rh Pd Ag 0 s Ir Pt Au 349.9 343.5 247.6 328.1 290.9 264.0 265.9 242.8 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 25 15 15 5 25 15 15 10 20 20 20 20 20 20 20 20 heating programme: 30-~dry/K .... 373 373 373 373 373 3 73 373 373 30-s ash/K . . . . 1700 1340 1050 680 (1000) 1650 1450 910 10-s atomise/K . . . . 2920 2 920 2 920 2 550 2 920 2 920 2 920 2 890 Sensitivity/wg ml-l per 0.0044A: Gas flow .. .. .. 0.19 0.0058 0.0055 0.00077 Not detected 0.38 0.041 0.003 1 Gasstop .. _. .. 0.28 0.0067 0.0045 0.000 21 Not detected 0.50 0.023 Flame? .. .. .. 0.5(AA) 0.3(AA) 0.25(AA) 0.06(&4) l.O(NA) 8.0(AA) 2.0(AA) 0.25(AA) Mini-gas stop . . . . 0.24 0.0063 0.0051 0.000 32 Not detected 0.46 0.030 0.001 1 0.000 89 * Analytical solutions: silver and osmium standards prepared in 0.01 M nitric acid, other metals in 0.01 M hydrochloric acid. -f Sensitivities of noble metals included for comparison.AA 1 air - acetylene, NA = nitrous oxide - acetylene.JUZY, I979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 649 (i) Effect of ashing temperature. A three-stage thermal programme (dry, ash and atomise) was used. For each measurement 2 0 4 of the noble metal solution were dried for 30s at 373 K, ashed for a further 30 s at a pre-selected temperature then finally atomised for 10 s at the maximum furnace temperature of 2920 K. Peak absorption signals were recorded during the final stage only. The ashing temperatures given in Table I were fixed at a nominal 70% of the appearance temperature for each noble metal. (ii) Effect of atomisation temperature. The effects of increasing atomisation temperature on the atomic-absorption signals for seven noble metals were examined under conditions of constant drying (30 s at 373 K) and ashing (30 s at 70% of the appearance temperature).Within the temperature range of the HGA-74 carbon furnace atomiser maximum absorption plateaus were observed for silver (2550 K) and gold (2780 K) and these were selected as the optimum atomisation temperatures. For the other metals maximum absorption signals were observed at the maximum available furnace temperature of 2920 K although it was clear that at higher atomisation temperatures even higher sensitivity could be achieved, especially for ruthenium and iridium. Lower atomisation temperatures may well be acceptable in instruments with more rapid heating rates than the HGA-74 used here. In this respect the choice of inert gas was found to be important, as nitrogen and particularly helium were found to decrease the maximum attainable furnace temperature for reasons of diffusivity and thermal conductivity.11 Argon was therefore selected as the most suitable inert gas.(iiz) Effect of inert gas $ow-rate. The effects of argon gas flow (1.5 ml min-l), mini gas flow and gas stop conditions on the atomic-absorption signals for the noble metals were investigated. As can be seen from Table I, marked improvements in the sensitivities of the more volatile noble metals were noted under gas stop conditions, owing to the increased residence time of the atomic vapour in the measuring cell. The less volatile noble metals gave poorer sensitivities under the same conditions. The sensitivities of the seven noble metals (osmium was not detected) were determined using the optimised conditions, and the results are shown in Table I.All results are based on a 2 0 4 sample volume. The sensitivities for each noble metal using a flame atom ceW2 are included in Table I for comparison. ( d ) Atomic-absorption determination of the appearance temperatztres of metals with carbon furnace atomisation The appearance temperatures for the noble metals were recorded using aqueous solutions and vacuum-deposited metal films. The wavelengths, spectral band passes and hollow- cathode lamp currents used in the investigation were as stated in Table I. A 20-4 volume of the analyte solution, con- taining 1 mg ml-1 of metal, was transferred into the HGA-74 carbon furnace tube using an Eppendorf micropipette.The sample was dried at 373 K (122 digits) for 30 s, then cooled to ambient temperature (000 digits) for a further 30 s. The furnace temperature was increased to 2920 K (999 digits) using a heating rate of 6.24 digits s-1 (programme 8) and during this period the absorption-time (temperature) profile of the element was recorded at a chart speed of 120 mm min-l. (ii) From vacuum-deposited metaZ$lms. To determine the minimum atomisation tempera- tures of pure metals that had been vacuum deposited on the inside wall of the HGA-74 atomiser tube, the heating programme described in (i) above was used after insertion of the graphite tube, prepared as in procedure (b), in the HGA-74 atomiser unit. (i) From aqueous solutions of metal saZts.(e) X-ray powder diffraction analysis X-ray diffraction analysis was used to confirm the existence of elemental noble metals on the surface of the HGA-74 graphite atomiser and the final products of thermogravimetric investigations. For the graphite furnace the heating programme involved drying a 200-4 aliquot of solution at 373 K for 30 s. Thereafter, the furnace was heated to a selected temperature and maintained at this temperature for 60s. About 5mg of material were scraped from the inside surface of the cooled tube for analysis. After thermogravimetric analysis, a scalpel was used to remove a similar amount of sample from the TGA crucible. A fine glass650 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, voi. 104 needle was then wet with Canada balsam cement and rolled in the sample so as to produce a cylinder of powder approximately 0.5mm in diameter.The needle was placed in the sample holder of the camera, which was loaded with X-ray sensitive film and placed on the optical bar of the X-ray generator. The wavelength of monochromatic X-radiation, exposure time, voltage and current used to produce the diffraction patterns are stated under Results and Discussion. The nature of the deposits were determined from the X-ray patterns through calculations of lattice parameters or d spacings and compared with the standard values of known materials13 and/or compared with measurements on known materials. Results and Discussion Thermogravimetric Investigation of Noble Metal Compounds and Their Aqueous Solutions Published thermochemical data have been used previously in attempts to explain the atom formation processes of metals in flames14 and electrothermal atomi~ers.15~16 Such data, however, usually relate to the thermal behaviour of metallic compounds in air and should not be used without correction to forecast reactions occurring in other atmospheres.In these studies, the thermal characteristics of several noble metal compounds, and their aqueous solutions, were investigated under an argon atmosphere. The decomposition patterns of the solid compounds were also compared using platinum and graphite crucibles (procedures TG1 and TG2), while the residues formed on heating solutions of noble metal salts were examined in the presence and absence of powdered graphite (procedures TG4 and TG3).The tests that involved the absorption of salt solutions into powdered graphite, rather than solid compounds on the surface of a graphite crucible, should provide a closer comparison to the reactions occurring during carbon furnace atomisation. As less than 5mg of noble metal compound or residue from the aqueous solutions was used for each test, intermediate compounds formed during the heating programme were not chemically analysed. Intermediate compounds were often identified from the loss in mass of starting material (for solids) or by working back from the mass of the final product that was identified from its X-ray diffraction pattern (for solids and residues from solutions). The mass losses for seven solid noble metal compounds were recorded as a function of temperature using platinum and graphite crucibles and procedures TG1 and TG2.The 0- I 300 500 700 900 1 100 1300 Temperature/K Fig. 2. Thermogravimetric graphs for seven noble metal salts: graph (a), HAuC1,.4H20; (b), IrC1,.-3H20; (c), PtC4; (d), RhC1,.-3€120; (e), RuCl,.- -3H20; ( f ) , PdCl,; (g), AgNO,. Condi- tions: heating rate 50 K min-1, platinum crucible and argon atmosphere (procedure TG1).J d y , 1979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 651 mass losses were converted into formula masses on the basis of the known starting materials and plotted against temperature. The results for platinum and graphite crucibles were similar and the thermogravimetric curves on platinum only are shown in Fig. 2. All of the compounds investigated are reduced to the elemental state at temperatures below the maximum temperature of the microfurnace (1 273 K) and the temperatures at which the conversion to pure metal is complete agree very closely for the platinum and graphite crucibles (Table 11).The existence of metallic noble metal at the end of each thermo- TABLE I1 TEMPERATURES AT WHICH NOBLE METAL SALTS ARE CONVERTED INTO METAL Temperature a t which conversion into metal is complete/K Starting compound RuC13.-3H,0 . . .. Rh.C1,.-3HZO . . .. PdCl, . . .. .. AgNO, .. . . .. IrC1,.-3H20 . . .. PtC1, .. .. .. HAuC14.4H,0 . . .. Procedure TG1 (1 100 1180 1070 850 1230 1030 620 Procedure TG2 < 1100 1165 1060 830 1230 1020 620 Procedure TG3 < 1000 1103 1030 7 70 1190 933 613 Procedure TG4 < 950 1030 980 740 1090 863 590 Final temperature of all TG programmes/K 1273 1273 1273 973 1273 1273 973 gravimetric experiment was confirmed by X-ray powder diffraction analysis, the results of which are indicated in Table I11 (C) for procedure TG2.With ruthenium, the final mass of metal was lower than that expected on the basis of the mass of starting material. Experi- ments with silver and gold salts were stopped at 973 K as substantial volatilisation of the metal was noted at higher temperatures. TABLE I11 X-RAY POWDER DIFFRACTION ANALYSIS OF THE DEPOSITS OBTAINED ON HEATING NOBLE METAL SALTS OR THEIR AQUEOUS SOLUTIONS Samples heated in the TG750 thermobalance microfurnace or in the HGA-74 carbon furnace atomiser. X-ray conditions: source, Cu Ka (0.154 nm) radiation; filter, nickel; exposure, 90 min; current, 25 mA; voltage, 40 kV.Sample volume for A and B: 30 p1 (equivalent to 1.5 mg of noble metal salt). Final TG750 or HGA-74 Lattice parameters/nm L temperature/ > Metal species Furnace Form of sample K Calculated Literature identified A: TG750 (Dissolved solids) (procedure TG4) RuC1,.3H20 1273 0.2690-0.4268 0.2704-0.4282 Ruthenium RhC13.3H,0 1273 0.3811 0.3803 Rhodium PdC1, 1273 0.389 6 0.3890 Palladium &NO3 IrC1,.3H20 PtCI, 1273 0.3925 0.3924 Platinum HAuCl,. 4H,O 973 0.407 4 0.407 8 Gold RuC13.3H,0 1250 0.2690-0.4268 0.2704-0.4282 Ruthenium RhC13.3H,0 1250 0.3790 0.3803 Rhodium PdC1, 1250 0.387 8 0.3890 Palladium AgNO, 900 0.408 0 0.408 6 Silver oso, 1250 0.2730-0.4309 0.2733-0.431 9 Osmium IrC13.3H,0 1250 0.382 8 0.3839 Iridium PtC1, 1250 0.391 8 0.392 4 Platinum HAuC14.4H,0 900 0.407 0 0.407 8 Gold 973 0.408 0 0.408 6 Silver 1273 0.384 1 0.383 9 Iridium B: HGA-74 (Dissolved solids) C: TG750 (Solid noble metal salts) (procedure TG2) The X-ray powder diffraction patterns obtained on examination of the residues formed on heating the seven solid noble metal salts listed under A above to the stated temperatures, in a graphite crucible, were recorded.The X-ray patterns derived from A above were used as fingerprint patterns to confirm the presence of Ru, Rh, Pd, Ag, Ir, Pt and Au metals after heat treatment of solid metal salts.652 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, vd. 104 Tests were also carried out using aqueous solutions of the noble metal salts in the presence and absence of powdered graphite (procedures TG3 and TG4).The thermogravimetric curves were very similar and the final products of the experiments, in which solution was absorbed into powdered graphite, were identified as elemental noble metal by X-ray diffrac- tion analysis, as shown in Table I11 (A). When powdered graphite was present in the platinum crucible (TG4), the thennogravimetric curves appeared to be stretched out at elevated temperatures compared to the curves in the absence of graphite. An example of this is shown for H2PtCl,.6H20 in Fig. 3. This made quantification of the results on a strict mass basis difficult. The effect may be due to absorption of the analyte solution into the graphite or to the reaction of graphite with nascent chlorine at elevated temperatures. This could give rise to a volatile carbon product, which would cause a decrease in the mass of the tared crucible plus graphite.The temperatures at which conversion to metal is complete were in all instances lower with solutions (TG3 and TG4) than with metal salts (TG1 and TG2), as shown in Table 11. This is probably a function of the more rapid heat transfer to a solid evaporated on to the surface of the crucible or powdered graphite. Results by all four procedures (TG1, TG2, TG3 and TG4) gave similar mass losses and indicated that under these conditions all of the noble metal salts investigated, whether as solids or in aqueous solution, produced metallic noble metal at temperatures below those at which atoms of these elements appear in the graphite furnace.- 1.8 E m - 1.6 Q) - 1.4 - 2 -. 0 L + 0.8 r“ 1 % Furnace temperature/K Fig. 3. Thermogravimetric graphs for solutions of H,PtCl,. 6H,O as recorded by using procedures TG4 (graph A) in powdered graphite and TG3 (graph B) in the absence of graphite. Trace C shows the loss in mass of 1.6 mg of powdered graphite. Measurement of Appearance Temperatures The lowest temperature at which atoms of a particular element are released from the surface of the atomiser into the vapour phase, the “appearance temperature,” will be a function of a number of parameters, such as the nature of the atomiser surface, the chemical form of the analyte present on the atomiser surface and the chemical reaction that takes place in the conversion of surface analyte into atoms.It will also depend to a smaller extent on the amount of analyte present and on the volume of the atomiser cell. Measurements of the appearance temperature, coupled with a knowledge of the chemical form of the analyte and atomiser surface, allow the nature of the atomisation reaction to be understood. Several workers have attempted to elucidate the reaction mechanism from such information but differences in the methods of measurement of the appearance temperatures have produced confusion and conflicting data. Campbell and Ottawayl’ first proposed the term appearance temperature but correlated it with the atomisation temperature set on the HGA-70 atomiser used, rather than the actual temperature of the atomiser at the time of atomisation.Aggett and Sprottl8 also measuredJ”y, 1979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 653 the “lowest temperature at which atomic absorption signals were observed,” but gave no indication of how the temperature was measured or how this was correlated with the atomic- absorption signals. The most satisfactory approach is that of Sturgeon and co-workers,19-21 who first adopted the term “delay temperature”19-20 as it coincided with the appearance of the atomic-absorption peak which started at the delay time. However, these workers subse- quently referred to their measurements as appearance temperatures21 Their method of measurement, which is essentially the same as that adopted in the present work, is illustrated in Fig.4. A typical absorbance peak is obtained under a given set of atomisation para- meters (specifically, final temperature and heating rate) and the delay time, rdelay is measured. The temperature at Tdelay can then be determined from a calibration graph of temperature against time measured under the same set of atomisation conditions. In the present work it was identified as the minimum temperature at which the absorbance on the leading edge of the atomic-absorption peak first exceeded the upper noise level (peak noise) of the background signal observed during the delay time. W 0 111 -2 2 a Peak absorbance, Appearance temperature (delay) \ \ I I I Peak display ! I I . . . :Delay time: time I I I I I / + / Start of End of a torn i sat ion atomisation Time/temperature increase -+ Fig.4. Characteristics of the atomic-absorption time pulse in carbon furnace atomisation. A similar approach was adopted recently by Czobik and Matousek,22 who used the atomic- absorption signal equal to twice the standard deviation of the base line as the point of measure- ment and referred to the corresponding temperature as the 20 atomisation temperature. Their work was concerned with the effect of anions (Le., interferences) on the appearance temperature and not on the use of appearance temperatures to predict atomisation reaction processes, but highlighted a factor that had not been considered by previous workers. The TABLE IV APPEARANCE TEMPERATURE OF PALLADIUM AS A FUNCTION OF SOLUTION CONCENTRATION 20 pl of PdCl, solution. Dry at 373 K for 30 s, delay 30 s, atomise from ambient to 2920 K in 140 s.Concentration of palladium/pg ml-1 10 20 30 50 125 250 500 1000 Appearance temperature/K 1573 1533 1533 1518 1493 1483 1473 1478654 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, VoZ. 104 first appearance of atoms of the analyte in the vapour phase has been observed in all instances by the measurement of atomic-absorption signals. The measurement of the start of the atomisation peak will therefore depend on the sensitivity of the instrumental system and on the amount of analyte present on the atomiser surface. The effect of concentration is illustrated for palladium in Table IV using a fixed 20-4 aliquot of each solution and identical atomisation conditions. The peak absorbance increases but the delay time is also shorter and the appearance temperature lower as the concentration of analyte is increased. Both the delay time and appearance temperature were found to reach minimum values at 500 pg ml-l for all of the elements studied and concentrations of 1000 pg ml-1 were used for all measurements.Although peak temperatures are more independent of concentra- tion,22 only the appearance temperature can be correlated satisfactorily with the atomisation reaction m e c h a n i ~ m . ~ ~ , ~ ~ All temperatures were obtained from a temperature - time cali- bration graph measured using the optical pyrometer. The absorption time profiles for the eight noble metals were recorded at a heating rate of 160 s from ambient to maximum furnace temperature, and the traces are shown in Fig.5. Both nitric acid and hydrochloric acid media were used when possible. A further four metals (zinc, cadmium, lead and aluminium) were measured for comparison using the same technique but in nitric acid media only. From the curves, appearance temperatures were determined and the results are shown in Table V together with the melting-points and boiling-points of each metal. I I L 1000 900 800 700 600 500 400 300 200 100 2923 2758 2538 2473 1973 1573 1113 733 488 348 L I I I I I I 1 I I 0 s R u Ir P t Rh Pd Au Ag Blank Digits Ternperature/K Fig. 5. Absorption - time (temperature) characteristics of the eight noble metals Heating rate from ambient to 2923 K a t A 20-pl volume of a 1000 pg ml-l solution of noble metal was in a HGA-74 carbon furnace atomiser.6.24 digits s-l. used in each instance. For the noble metals, and with the narrow exception of platinum, the appearance tempera- tures for aqueous solution are lower than the melting-points of the metals. In contrast, the appearance temperatures of zinc, cadmium, lead and aluminium in nitric acid solutions are well above the corresponding melting-points. For the noble metals, close agreement was found between the appearance temperatures in hydrochloric and nitric acid media, suggesting a common atomisation process. Thin films (weighing approximately 50pg) of pure palladium, silver, gold, zinc, lead, cadmium and aluminium metals were individually deposited on the inside walls of seven HGA-74 carbon furnace atomiser tubes using the vacuum deposition apparatus [procedure (b)].Each prepared tube was then used to obtain the absorption versus time traces for the seven metals as described in procedure [d ( 4 1 . From the results, the appearance tempera- tures of metal atoms were determined as shown in Table V.JUZY, 1979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 655 TABLE V APPEARANCE TEMPERATURES OF METALS IN CARBON FURNACE ATOMISATION Saturated vapour pressure Appearance temperature of of metal at the appearance Melting Boiling Aqueous solution Vacuum- Vacuum- temperature temperature ,-*-, deposited Aqueous deposited metal/K temperature/Pa* I A >1 Element of metal/K of metal/K HCl HNO, metal solution metal 3 x 10-3 I 4 x 10-3 - 2 x 10-2 1 x 10-2 Ruthenium . . 2773 5 200 2430 2410 t Rhodium ... . 2240 4 800 1910 1900 t Palladium . . . . 1825 4250 1480 1500 1470 Silver . . . . 1234 2 480 - 970 980 5 x 6 x 2 900 t Osmium . . . . 3300 5 300 - Iridium . . .. 2710 5 600 2350 2370 t Platinum . . . . 2042 4 800 2070 2070 t Gold . . . . 1336 3 240 1290 1310 1290 Zinc .. .. 693 1180 - Cadmium . . .. 594 1040 - - - 9 x 10-3 - 3 x 10-2 - 9 x 10-4 5 x 10-4 970 530 2 x 104 2 x 10-2 770 450 5 x 103 9 x 10-3 Lead . . .. 600 1893 - 938 770 5 x 10-1 3 x 10-3 Aluminium .. 933 2 330 - 2 090 1180 1 x 103 2 x 10-3 * Vapour-pressure data were calculated on the basis of Honig and Kramer data (1969).2j t Not evaporated at highest temperatures available. For palladium, silver and gold very close agreement was obtained between the appearance temperatures using aqueous solutions of the metals and films of pure metal.This evidence supports very strongly the conclusion that these three metals are present in the elemental state on the surface of the atomiser prior to the detection of atoms in the vapour phase. In contrast, the appearance temperatures of metal films of zinc, lead, cadmium and aluminium are much lower than the corresponding temperatures found for aqueous solutions. This evidence suggests that these four metals do not exist in the elemental state prior to atomisa- tion when aqueous solutions are injected, as otherwise much lower appearance temperatures would be expected. The results support the conclusion that the atomisation mechanism for these metals involves a thermal or chemical decomposition of a metal salt or oxide at the appearance temperature and not before it.17s21 Vacuum deposition of the other noble metals on the atomiser surface was attempted.Unfortunately, ruthenium and iridium could not be deposited because their high evaporation temperatures were unattainable in the vacuum evaporation chamber. Platinum and rhodium were also difficult to evaporate owing to an alloy action of the metals with the tungsten heater coil, which fractured and destroyed the electrical continuity of the heating circuit. Therefore, evidence to support the conclusion of the thermogravimetric and X-ray diffraction studies, which indicated an atomisation process involving evaporation of solid metal, had to be obtained indirectly. The saturated vapour pressures at which the seven vacuum-deposited metals commence atomisation (ie., at the appearance temperature) were calculated and were found to be very similar (see Table V).Similar values were obtained for the seven noble metals in aqueous solution but completely different values were found for zinc, cadmium, aluminium and lead in atomisation from aqueous solution. The mean value from the metal deposits was found to be 6.4 x Pa, in close agreement with the vapour pressure at which rapid evaporation of metal takes place according to Holland23 and L’vov.~~ The temperatures at which all twelve metals considered in Table V attain a saturated vapour pressure of 6.4 x Pa were calculated25 and plotted against their appearance temperatures from aqueous solution. All of the points for the seven noble metals lie on a straight line (Fig.6) whereas those for zinc, lead, cadmium and aluminium exhibit no correlation. Much better correlation for these metals is achieved, as expected, if the appearance temperatures of the vacuum- deposited metals are plotted instead of those from aqueous solutions. Allowing for experi- mental error and the unknown validity of the vapour-pressure data, the excellent correlation indicates a common atomisation mechanism for all of the noble metals. That is, the final step in the atomisation process for noble metals involves the evaporation of free metal when the temperature of the atomiser surface has been increased to that at which the saturated656 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, VoZ. 104 2 2500- Lc gY 2000- e ? 2 s 1000- 2 cl: I 2 00) E I500 a - i! a 500 0) 3 000 I- - - - 1 I 0 500 1000 1 500 2000 2500 Temperature at which the analyte metal has 3 saturated vapour pressure of 6.4 x Pa/K Fig.6. Relationship of appearance tempera- ture and temperature a t which the analyte metal has a saturated vapour pressure of 6.4 x Pa. 0, Aqueous solution; and + , evaporated metal. vapour pressure of the metal exceeds about of the atomiser surface and the vapour phase have been shown to be negligible.26 Pa. Differences between the temperature Further evidence to support a mechanism of the general form noble metal(s) -+ noble metal,,, . . .. .. * (1) at the appearance temperature was obtained by calculation of the activation energy of the atomisation process for each metal using the procedure proposed by Sturgeon et aZ.21 In their method, a plot of log (absorbance) versus 1/T is constructed for results in the region of the appearance temperature and the activation energy is obtained from the slopes.For this purpose, the absorption time profiles shown in Fig. 5 were used to construct the log Pd K\ Rh -2.0 9.0 9.2 9.4 , 9.6 9.8 I I I I I I I I I 6.2 6,4 6.6 6.8 7.0 7.2 7.4 7.6 7.8 Au.Pd 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 Pt,Rh,lr,Rti I I 1 1 I I I I I I I T ' ~ 104/ti-1 Fig. 7. Activation energy graphs (of log A verssus 1/T) for 20 pg of seven noble metals derived from the absorption - time pulses shown in Fig. 4.July, 1979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 657 A versus 1/T plots shown in Fig. 7. The values of the activation energy obtained from these linear plots are given in Table VI and compared with the values of the enthalpy of evaporation of the corresponding metals.27 In all instances agreement is satisfactory and within the limitations of the experimental method and data used.TABLE VI COMPARISON OF CALCULATED E , VALUES WITH THE ENTHALPIES OF ATOMISATION FOR NOBLE METALS Element Ruthenium .. .. Rhodium .. .. Palladium . . .. Silver . . . . .. Osmium . . .. Iridium .. .. Platinum . . .. Gold . . .. .. Activation energies of the limiting step in the atomisation process, E,/k J mol-1 625 55 1 37 1 201 625 564 322 - Enthalpy of atomisation of noble metals,27 z.e., M(s) + M(g)/kJ mol-l 640 556 390 289 782 665 565 369 The relationship between the activation energy (E,) and the appearance temperature is linear, as shown in Fig.8, and allows the prediction of the appearance temperature of osmium atoms, which could not be measured in this work. An enthalpy of atomisation of 782 k J mol-1 for osmium gives an appearance temperature for osmium of about 2800 K. It should there- fore be possible to measure osmium in a graphite furnace atomiser, particularly on systems that exhibit faster heating rates near the maximum temperature than are achieved by the HGA-74. 600 - I 0 -7 Y \ - 500- k? 400- 300 - ' R u I I I I I I I I I I I i [OSI I I I 200 I I 1 I 500 1000 1500 2000 2500 3000 Appearance temperature/K Fig. 8. Relationship of activation energy of atomisation of noble metals to the appearance temperature in carbon furnace atomisation.It is now clear that for the seven noble metals examined, the mechanism of atomisation involves evaporation of free metal at the appearance temperature and above. The thermo- chemical processes that occur below the minimum atomisation temperature will depend on the specific metal and on the nature of the acid media and other chemical conditions of the sample solution. Thermogravimetric and X-ray diffraction analysis of metal salts and residues recovered from the graphite atomiser can be used in conjunction with the above data to describe the complete atomisation process that follows the injection of the sample658 ROWSTON AND OTTAWAY: DETERMINATION OF NOBLE METALS BY Analyst, ‘VOl. 104 solution. Typical mechanisms for some noble metal salts dissolved in hydrochloric acid solution are shown below.The final stage in each instance represents sublimation of the metal at its appearance temperature. 950 K* 2430 Kt RuC13.3H20 } 400 ~ Ill-defined ____, Ru(s, Rue, 0.01 M HC1 species 400 K 1030 K* 1910 K t tfFk$y } - RhCl,(,, - Rh,,, - Rh,,, 980 K* 1480 K t } -f!!-% PdCl,(,, ---+ P ~ ( s ) - Pd(g) PdC1, 0.01 M HCl * Temperature a t which conversion into the elemental noble metal is complete. t Sublimation temperature of the noble metal. Minimum temperature that is required to produce a detectable atomic vapour, i.e., appearance temperature. With different salts or solutions, the first two stages in these processes will be different, but in all of the examples of solutions containing a single noble metal (k, in the absence of interfering elements), the third stage has been identified as the atomisation process.In this paper, the atomisation processes of noble metals have been described as a result of experimental studies aimed at identifying species formed or likely to be formed on the surface of the atomiser and theoretical evaluation of the absorption - time pulse. It is our view that these two approaches are complementary. The thermogravimetric studies can be criticised on the grounds that the experiment does not correspond perfectly to that which occurs under carbon furnace atomisation owing to differences in the amount of analyte required and heating rate. On the other hand, experimental errors and the use of inaccurate fundamental data can lead to erroneous conclusions if the fundamental approach is adopted alone.The combination of information derived from several techniques here has allowed an unequivocal elucidation of the mechanism of atomisation of noble metals and could be used to advantage in the study of atomisation and interference mechanisms of other metals. In Part I1 of this series, some interferences observed with noble metals will be described and their mechanisms interpreted. The authors thank the British Steel Corporation, Ravenscraig Works, for providing the Perkin-Elmer 360 spectrometer and HGA-74 carbon furnace atomiser, and Glasgow College of Technology for the use of the Philips X-ray spectrometer, therrnogravimetric apparatus and other facilities. References 1. Sen Gupta, J. G., Miner. Sci. Engng, 1973, 5, 207. 2. Mallett, R. C., Pearton, D. C. G., Ring, E. J., and Steel, T. W., Taluntu, 1972, 19, 181. 3. Fuller, C. W., “Electrothermal Atomization for Atomic Absorption Spectrometry,” Analytical 4. Kirkbright, G. F., and Sargent, M., “Atomic Absorption and Atomic Fluorescence Spectroscopy,” 5. Adriaessens, E., and Knoop, P., Analytica Chim. Acta, 1974, 68, 37. Sciences Monograph No. 4, Chemical Society, London, 1977. Academic Press, London and New York, 1974.July, 1979 CARBON FURNACE ATOMIC-ABSORPTION SPECTROMETRY. PART I 659 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. Pearton, D. C. G., and Mallett, R. C., Report No. 1598, National Institute of Metallurgy, Johannes- Steele, T. W., and Guerin, B. D., Analyst, 1972, 97, 77. Everett, G. L., Analyst, 1976, 101, 348. Aggett, J., and West, T. S., Analytica Chim. A d a , 1971, 55, 349. Guerin, B. D., Jl S. Afr. Chem. Inst., 1972, 25, 230. Littlejohn, D., and Ottaway, J. M., Analyst, 1978, 103, 595. Slavin, W., “Atomic Absorption Spectrophotometry,” Interscience, New York, 1969. Cullity, B. D., “Elements of X-ray Diffraction,” Addison-Wesley, London, 1959. Halls, D., and Townshend, A., Analytica Chim. Acta, 1966, 36, 278. Ottaway, J. M., Proc. Analyt. Div. Chem. SOL, 1976, 13, 185. Maessen, F. J., and Posma, F. D., Analyt. Chem., 1974, 46, 1439. Campbell, W. C., and Ottaway, J. M., Talanta, 1974, 21, 837. Aggett, J., and Sprott, A. J., Analytica Chim. Acta, 1974, 72, 49. Sturgeon, R. E., Chakrabarti, C. L., Maines, I. S., and Bertels, P. C., Analyt. Chem., 1975, 47, 1240. Sturgeon, R. E., Chakrabarti, C. L., and Bertels, P. C., Analyt. Chern., 1975, 47, 1250. Sturgeon, R. C., Chakrabarti, C. L., and Langford, C. H., Analyt. Chem., 1976, 48, 1792. Czobik, E. J., and Matousek, J. P., Talanta, 1977, 24, 573. Holland, L., “Vacuum Deposition of Thin Films,” Chapman and Hall, London, 1970. L’vov, B. V., “Atomic Absorption Spectrochemical Analysis,” Adam Hilger, London, 1970. Honig, R. E., and Kramer, D. A., RCA Rev., 1969, 21, 360. Littlejohn, D., and Ottaway, J. M., Analyst, 1979, 104, 208. Stark, J. G., and Wallace, H. G., Editors, “Chemistry Data Book,” (SI Edition), John Murray, burg, 1974. London, 1975. Received November 15th, 1978 Accepted February 2nd. 1979

ISSN:0003-2654    

DOI:10.1039/AN9790400645

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

660 Analyst, July, 1979, Vol. 104, pp. 660-667 Separation of Organotin Compounds by Using the Difference in Partition Behaviour Between Hexane and Methanolic Buffer Solution Part 1. Determination of Butyltin Compounds in Textiles by Graphite Furnace Atomic-absorption Spectrophotometry Shigeo Kojima National Institute of Hygienic Sciences, 1-18, Karniyoga, Setagaya-ku, Tokyo, Japan A method for determining tributyltin compounds (TBTs) and dibutyltin compounds (DRTs) in textiles has been studied by graphite furnace atomic- absorption spectrophotometry. They are successfully extracted from textiles with methanol containing 0.05% of hydrogen chloride. Their partition between hexane and a methanolic buffer solution (pH 8.5) was studied by varying the methanol concentration of the aqueous buffer phase.Only TBTs can be extracted with hexane from a methanolic buffer solution con- sisting of 40 ml of methanol and 30 ml of buffer solution. TBTs are adsorbed on an alumina column and eluted with a small volume of dichloromethane. DBTs remaining in the aqueous phase are extracted with dichloromethane as a complex with 4-(2-pyridylazo)resorcinol. After evaporation of the dichloromethane, the residue is heated with concentrated nitric acid and its tin concentration is determined by atomic-absorption spectrophotometry. Keywords : Organotin determination ; liquid - liquid partition ; butyltin corn- pounds ; textiles ; graphite furnace atomic-absorption spectrophotometry Tributyltin compounds (TBTs), which have strong biocidal properties, are used for the mildew-resistant finishing of textiles, wood preservation, slime control in the paper industry and anti-fouling paints for ships’ bottoms, whereas dibutyltin compounds (DBTs) are used as stabilisers for poly(viny1 chloride) and other plastics.Because TBTs for factory use contain DBTs as impurities and vice v m a , and they have different toxicities, it is necessary to determine them separately. Various methods have been proposed for the separation of organotin compounds, vix. , thin-layer,lS2 c o l ~ r n n ~ ~ ~ and gas - liquid6-10 chromatography and liquid - liquid extraction,11-13 but they do not always give satisfactory results because thermally unstable organotin compounds often decompose or disproportionate during gas chromatographygJo and volatile TBTs are partly lost during the evaporation of the solvent .2 Freitag and Bock13 reported a separation method for triphenyltin and diphenyltin com- pounds involving their partitioning between chloroform and phosphate - citrate - EDTA buffer solution (pH 8.5).However, attempts to apply this method to mixtures of TBTs and DBTs were not successful in this laboratory. Subsequently the partition behaviour of some organotin compounds between hexane and methanolic buffer solution has been studied by varying the methanol concentration of the aqueous buffer phase. It was found that the effect of the methanol concentration on the partition behaviour is characteristic of the type and number of substituents, and that only TBTs can be extracted with hexane from a methanolic buffer solution consisting of 40ml of methanol and 30ml of buffer solution, while the DBTs remain in the aqueous phase.This paper describes a method for the determination of TBTs and DBTs by use of liquid - liquid extraction, pre-concentration of TBTs on an alumina column and graphite furnace atomic-absorption spectrophotometry, which permits their individual determination without loss. The method has been applied to the determination of butyltin compounds in textiles. Experimental Apparatus The instrument was a Perkin-Elmer, Model 403, atomic-absorption spectrophotometer equipped with a Perkin-Elmer, Model HGA-2100, graphite furnace atomiser and atomisation control unit, and coupled to a Hitachi, Model QPD-54, recorder.KO JIMA 661 All of the glassware was cleaned thoroughly with a potassium dichromate - concentrated sulphuric acid mixture, washed with water, rinsed with 3% nitric acid and distilled water and dried before use.Reagents and Samples All reagents were of analytical-reagent grade unless otherwise stated. Methanol containing 0.05% of hydrogen chloride. Dilute 1.5 ml of concentrated hydro- chloric acid to 1 1 with methanol. Phosphate - citrate - EDTA bzc$er solution. Dissolve 42.3 g of disodium hydrogen ortho- phosphate dodecahydrate, 7.7 g of citric acid and 2.0 g of disodium ethylenediaminetetra- acetate (EDTA) in 800 ml of distilled water, adjust the pH of the solution to 8.5 and dilute to 1 1 with distilled water.13 Dilute three-fold with distilled water before use. 4-(2-Pyridylazo)resorcinoZ solutiort, 2.5 x M.Dissolve 107 mg of 4-(2-pyridylazo)- resorcinol (PAR) in 200 ml of methanol. Capriquat solution. Dissolve 20 mg of trioctylmethylammonium chloride (Capriquat) in 100 ml of hexane. Hydrated alumina. Prepare hydrated alumina containing 10% m/m of water by adding distilled water to ICN Pharmaceuticals’ aluminium oxide W 200 neutral, activity grade Super I (anhydrous, about 200 mesh) and shaking it vigorously for 30min. Keep it in a tightly capped container. Organotin compounds. Bis(tributy1tin) oxide (TBTO, 97% purity) and tributyltin chloride (TBTC, 97%) , acetate (TBTA, 99%) and fumarate (TBTF, 99%) were supplied by Yoshitomi Pharmaceutical Industries, Ltd. , and used without purification. Dibutyltin dichloride (DBTDC) (Tokyo Kasei Kogyo Co. Ltd.) was purified by recrystallisation from benzene.Dibutyltin dilaurate (DBTDL, Nitto Kasei Kogyo Co. Ltd.), diacetate (DBTDA) and maleate (DBTM) (Tokyo Kasei Kogyo Co. Ltd.) were used without purification. Standard tin stock solution, 0.287 mg ml-l. Dissolve 0.287 g of tin shot (99.999% purity; Wako Pure Chemical Industries, Ltd.) in concentrated hydrochloric acid by heatmg and dilute to 11 with 1 M hydrochloric acid. Standard tin working solution, 0.287 pg ml-l. Dilute 1 ml of the stock solution to 1 1 with 3% nitric acid. Prepare this solution just before use. Mildew-resistant Jinishing agent, Sana TS. This agent , purchased from Yamato Chemical Industries Co. Ltd., is a xylene solution of TBTO (about 20%) and a surfactant. Standard samples. Standard samples were prepared by soaking wool fabrics in a dilute aqueous emulsion of Sana TS (liquid to fibre ratio 30: 1) at 60 “C for 1 h in a flask fitted with a reflux condenser.After cooling, the fabrics were washed with distilled water, dried at room temperature for 5 days and weighed (mass m). The theoretical content of TBTO in the fabrics (“TBTO calculated”) was determined as follows. The flask and the reflux condenser were rinsed with small volumes of methanol and hexane. The washings, the residual processing solution and the washings from the fabrics were combined, and the remaining TBTO (mass t2) was extracted with hexane and determined by the recommended procedure. In addition, a volume of the dilute aqueous emulsion of Sana TS identical with that used for processing was independently taken and extracted with hexane, and the total amount of TBTO added (tl) was determined in the same manner as for t,.“TBTO calcu- lated” was given by (tl - tz)/m. Commercial samples were purchased at a department store in Tokyo in 1974. Flame-resistant fabrics were provided by Dr. Teruo Kan of the Tokyo Metropolitan Research Laboratory of Public Health. Commercial samples. Recommended Procedure Extraction from textiles and separation of TBTs and DBTs Subdivide a textile sample into pieces approximately 10 x 10 mm in size. Weigh accurately 1.0 g of the sample and transfer it into a 200-ml round-bottomed flask. Add 100 ml of methanol containing 0.05% of hydrogen chloride, reflux for 30 min and filter the methanolic extract through a glass frit (porosity 2).Wash the flask, the reflux condenser and the glass frit with 20 ml of methanol and add the washings to. the filtrate. To the methanolic solution, add 1 ml of 10% sodium hydroxide solution and 90 ml of the buffer662 KO JIMA SEPARATION OF Analyst, Vol. 104 solution. Extract the methanolic buffer solution with four 30-ml portions of hexane and wash the combined hexane extract with 10ml of the buffer solution. Transfer the hexane phase into a flask, add 5 g of anhydrous sodium sulphate and stand for more than 4 h. Then add 1 ml of the Capriquat solution to the hexane solution and pass the solution through a 10 mm i.d. column packed with 1.5 g of hydrated alumina. Elute the adsorbed TBTs from the column with 10ml of dichloromethane and use this eluate as the test solution for TBTs.To the residual methanolic buffer solution, add 5 ml of PAR solution and concentrate the solution to about 70 ml on a rotary evaporator. Extract the resulting DBT - PAR complex with three 10-ml portions of dichloromethane. Centrifuge the emulsified mixture in the extraction procedure, whenever necessary. Concentrate the combined extract to 2-3 ml at 40 "C on a rotary evaporator and pass the concentrate through a small glass frit (porosity 2) packed with a small amount of anhydrous sodium sulphate. Wash the glass frit with dichloromethane and add the washings to the filtrate. Use this solution as the test solution for DBTs. Determination of tin concentration by atomic-absorption spectrophotometry Remove the dichloromethane from the test solution at 40 "C on a rotary evaporator, but do not evaporate completely to dryness, then dry with a stream of air or nitrogen.Add 2 ml of concentrated nitric acid to the residue, fit a reflux condenser to the flask and heat for 5min. After cooling, wash the reflux condenser with 3% nitric acid. Transfer the oxidised solution into a 10-ml calibrated flask through a glass frit (porosity 3) and dilute to the mark with distilled water. Withdraw a 20-4 aliquot of the solution with a high-precision Eppendorf micropipette and determine its tin concentration by injection into a graphite furnace atomiser, operated under the following conditions : Wavelength . . . . . . 286.3nm Lamp current . . . . .. 20mA Spectral band width . . . . 0.2nm Readout mode . . .. . .Absorbance Ashing conditions . . .. 5OO0C, 50s Atomisation conditions . . . . 2500 "C, 10 s Shielding and purge gas Drying conditions . . . . 110 "C, 20s . . Nitrogen, flow-rate 30 ml min-l, interrupted mode Sample volume . . . . .. 2 0 4 Results and Discussion Separation of TBTs and DBTs by Partitioning Between Hexane and Methanolic Buffer Solution Fig. 1 shows the partition behaviour of TBTs and DBTs between 1 O m l of hexane and methanolic buffer solutions consisting of various volumes of methanol and 30 ml of the buffer solution. All of the tributyltin compounds examined (TBTO, TBTC, TBTA and TBTF) show the same partitioning behaviour. They are found exclusively in the hexane phase in the absence of methanol, and the concentrations in the organic phase gradually decrease with increasing the methanol concentration of the aqueous phase.In contrast, most of the dibutyltin compounds, such as DBTDC, DBTDA and DBTM, remain in the aqueous phase in the absence of methanol, while more than 70% of DBTDL partitions into the hexane phase. However, the concentration of DBTDL in the hexane phase decreases rapidly with increasing methanol concentration in the aqueous phase. The addition of 40 ml of methanol ensures that all of the DBTs, including DBTDL, remain in the aqueous phase, while TBTs can be extracted successfully with four 10-ml portions of hexane. The extraction pattern for TBTs is shown in Fig. 2. The influence of the pH of the buffer solution on the partition of TBTs was studied. The concentration of TBTs in the hexane phase was almost constant in the pH range 7.0-12.0.Sawyer14 extracted DBTs, which were eluted from poly(viny1 chloride) and other plastics into water, as the complex with PAR with chloroform. Because PAR forms complexes selectively with dialkyltin compounds in the presence of EDTA, we attempted to extract DBTs remaining in the methanolic buffer solution as the PAR complex with dichloromethane.Jub, 1979 ORGANOTIN COMPOUNDS. PART I 663 0 20 40 60 80 100 Volume of methanol/ml Fig. 1. Partition of TBTs and DBTs between 10 ml of hexane and methanolic buffer solution consisting of various volumes of methanol and 30 ml of phosphate - citrate - EDTA buffer solution (pH 8.5). 0, TBTs; A, DBTDL; and A, other DBTs. 100 a9 0 L $ I- 6 50 a a 0 Number of extractions Fig.2. Extraction pat- tern of TBTO from meth- anolic buffer solution [con- sisting of 40 ml of meth- anol and 30 ml of phos- phate - citrate - EDTA buffer solution (pH 8.5)], with five 10-ml portions of hexane. However, this failed in the presence of a large volume of methanol. When the methanol was removed from the aqueous phase, the DBT - PAR complex was successfully extracted with dichloromethane. The results for tests of the recovery of TBTO and DBTDC from the methanolic solution by the recommended procedure are given in Table I. TABLE I RECOVERIES OF TRIBUTYLTIN AND DIBUTYLTIN COMPOUNDS FROM METHANOLIC SOLUTIONS BY THE RECOMMENDED PROCEDURE Amount found in the test solution/pg* (recovery, %, in parentheses) Amount added/pg* - I 1 h NO. TBTO DBTDC TBTs DBTs 1 5.5 - 5.3 (96.4) N.D.1.2 109.0 - 103.2 (94.7) 0.6 3 - 4.7 N.D.1. 4.6 (97.9) 4 - 93.2 N.D. t 93.2 (100.0) 5 5.5 4.7 5.3 5.0 5.5 4.8 5.3 4.6 5.2 4.9 6.6 4.8 Mean: 5.4 (98.2) 4.8 (102.1) 4.7 107.1 5.2 105.0 5.3 109.6 6.1 Mean: 107.2 (98.3) 5.2 (110.6) 5.6 93.2 6.6 94.7 5.2 92.0 6.5 89.0 Mean: 5.4 (98.2) 91.9 (98.6) * Calculated mass of tin taken or found. t Not detected. 6 109.0 7 Pre-concentration of TBTs with an Alumina Mini- column Evaporation of the solvent and wet ashing of the residue are essential procedures in the determination of organometallic compounds in organic solvents by atomic-absorption664 KO JIMA SEPARATION OF Analyst, Vol. 104 spectropho tome try with electrothermal atomisation. When the solvent was completely evaporated from the hexane solution of TBTO on a rotary evaporator, more than 20% of the TBTO was lost during the procedure because of its high volatility. Therefore, the pre- concentration of TBTs on an alumina mini-column was examined.The recovery of TBTs from the alumina column was tested as follows. A 150-ml volume of hexane solution containing 20.7 pg of TBTO was passed through a column packed with 1.5 g of hydrated alumina and the total amount of TBTO was adsorbed on the alumina. TBTO was eluted from the column with 10ml of dichloromethane and determined by the recommended procedure. As shown in Fig. 3, the recovery of TBTO from the column increased with increasing water content of the hydrated alumina and became constant at water contents greater than 12y0. However, the recovery was only 75%, and 25% of the TBTO was irreversibly adsorbed on the alumina.It was found that when 100 pg of dioctyltin dichloride (DOTDC) was present in the hexane solution, TBTO was recovered quantitatively, independent of the water content of the hydrated alumina at levels above 5% (Fig. 3). It was thought that DOTDC inhibited the irreversible adsorption of TBTO on the alumina. On the assumption that analogous cationic compounds should have the same effect as DOTDC, three cationic surfactants, viz., tetrabutylammonium chloride, trioctylmethylammonium chloride (Capri- quat) and distearyldimethylammonium chloride, were examined. In the presence of 200 pg of these surfactants in the hexane solution, the recoveries of TBTO from the column packed with hydrated alumina containing 10% of water were 100.7, 101.4 and 99.8y0, respectively.0 5 10 15 Water content of hydrated alumina, % Fig. 3. Influence of water content of hydrated alumina and addition of DOTDC or Capriquat on the recovery of TBTO from alumina mini-column. 0, TBTO, 20.7 p g ; A, TBTO, 20.7 p g , and DOTDC, 1OOpg; and 0, TBTO, 20.7 pg, and Capriquat, 200 pg. Consequently, Capriquat was employed for this purpose because of its solubility in hexane. The amounts of the hydrated alumina (1.0-2.0 g) and Capriquat (100-300 pg), the volume of hexane (50-200 ml) and the rate of elution did not affect the recovery in the given range. Washing and drying of the hexane extract are essential pre-treatments. By this method, a small amount of TBTs in a large volume of hexane can be concentrated without loss.Decomposition of Butyltin Compounds and Atomisation When a dichloromethane solution of TBTO or DBTDC was injected into the graphite furnace atomiser without wet ashing, the calibration graph was not linear, as shown in Fig. 4. On the other hand, if the dichloromethane solution was evaporated to dryness, the residue heated with concentrated nitric acid and the oxidised solution diluted with distilled water, the aqueous solution obtained gave a linear calibration graph at tin concentrations in the range 0-0.50 pg ml-1. This graph was identical with that for the standard tin solution.July, 1979 ORGANOTIN COMPOUNDS. PART I 665 0 0.25 0.50 Concentration of tin/pg mI-' Fig. 4. Calibration graphs for butyltin compounds. m, Standard tin solution; 0, TBTO solution in dichloromethane; a, TBTO solu- tion, heated with concentrated nitric acid; A, DBTDC solution in dichloromethane; and A, DBTDC solution, heated with concentrated nitric acid.Application to the Determination of TBTs and DBTs in Textiles Kaniwa et a1.l5 reported that dieldrin (1,2,3,4,10,10-hexachloro-6,7-epoxy-l,4,4a,5,6,7,8,8a- octahydro-exo-l,4-ertdo-5,8-dimethanonaphthalene), a typical f at-soluble chemical like TBTs, was successfully extracted with methanol from wool fabrics, but only partly with hexane. Freeland and HoskinsonlS used chloroform for the extraction of TBTs from wool fabrics, but obtained only poor recoveries. Williams4 used a 0.05% solution of hydrochloric acid in ethanol for extracting TBTs and DBTs quantitatively from preserved softwoods.On this basis, three solvents, vix., hexane, methanol and methanol containing 0.05% of hydrogen chloride, were chosen and their efficiencies in the extraction of TBTO from standard samples were examined. The results in Table I1 show that, although only 40-50% of TBTO was extracted with methanol alone, on the addition of 0.05% of hydrogen chloride the recovery became quantitative. DBTs were also extracted quantitatively with acidified methanol from wool fabrics that had been processed with DBTDC in the same manner as with TBTO (Fig. 5). In addition, TABLE I1 RECOVERIES OF TBTO FROM STANDARD SAMPLES WITH THREE EXTRACTION SOLVENTS TBTO found/pg g-1 (recovery, %, in parentheses) f A \ Standard TBTO calculated/ Methanol - 0.05% sample Pf? g-l Hexane Methanol HCl A 63.1 4.5 (7.1) 30.4 (48.2) 64.3 (101.9) B 177.3 7.7 (4.3) 84.1 (47.4) 178.3 183.1 185.0 175.4 180.9 Mean: 180.5 (101.8) 36.4 (43.9) 83.8 82.5 84.3 79.6 83.3 Mean: 82.7 (99.6) C 83.0 4.3 (5.2)666 KO JIMA : SEPARATION OF Aaalyst, Vol.104 Fig. 6 shows that TBTs and DBTs were successfully extracted with acidified methanol from a commercial fabric (flame-resistant fabric A in Table IV). On the basis of these results, methanol containing 0.05% of hydrogen chloride was selected for extraction. The state of subdivision of the fabrics did not affect the efficiencies of extraction of TBTs and DBTs from these fabrics with acidified methanol. Y- u) -0 5 l o 2 - P 5 C > 3 .- c, +d - a, I I 1 0.1 0.2 x 0 Concentration of hydrogen chloride in methanol, % Fig.6. Influence of the concentration of hydrogen chloride in methanol on the extraction of TBTs and DBTs from wool fabrics. 0, TBTs; and e, DBTs. ,, c A Because many kinds of surfactant are used as softeners for textile products, the potential interference of anionic, non-ionic and cationic surfactants in the determination of TBTs and DBTs was examined. Amounts of sodium dodecylbenzenesulphonate, polyoxyethylated nonyl phenyl ether and distearyldimethylammonium chloride in up to 500-fold excess did not interfere in the determination of 1-100 pg of TBTs or DBTs; 50-fold excesses of metallic ions, such as tin(II), tin(IV), chromium(VI), nickel, mercury, lead, zinc, iron(III), cadmium, copper and silver, did not interfere. As shown in Table 111, triphenyltin chloride (TPTC) and dioctyltin dichloride (DOTDC) caused positive relative errors in the TBTs and DBTs determinations.However, triphenyltin compounds are rarely used for the mildew-resistant finishing of textiles. TABLE I11 INTERFERENCES IN THE DETERMINATION OF TBTs AND DBTs CAUSED BY TRIPHENYLTIN AND DIOCTYLTIN COMPOUNDS Amount found in the 7 - 7 & Amount added/pg* test solution/ pg* No. TPTC DOTDC TBTO TBTs DBTs - 3.8 4.7 1 9.1 - 2 9.1 - 10.9 14.5 4.9 3 - 112 - 0.3 2.4 * Calculated mass of tin taken or found. Butyltin compounds in some commercial textile products were determined by the recom- mended procedure, and the results are given in Table IV. Large amounts of DBTs were found in diaper cover A and flame-resistant fabrics A-C. It is thought that DBTs have been used as stabilisers for nylon 6 or polychlal in these fabrics.A small amount of TBTs was found in overcoat A, which had caused dermatitis in a garment-manufacturing plant in Japan. This fabric had probably been treated with an agent containing TBTs in some process during manufacture.J d y , 1979 ORGANOTIN COMPOUNDS. PART I 667 TABLE IV ANALYTICAL RESULTS FOR TBTs AND DBTs IN SOME COMMERCIAL TEXTILE PRODUCTS Amount foundlpg 8-l r A \ Sample Diaper . . Diaper covers : A .. B .. . . A .. .. B .. . . A .. .. B .. .. A .. . . Socks : Pantyhose : Overcoats : B .. .. .. .. .. .. .. .. .. .. .. . . . . . . . . . . .. .. .. . . Flame-resistant fabrics : A .. .. .. .. B .. .. .. . . c .. .. .. . . Material* Cotton Nylon 6 Wool Polyester,§ nylon 6, wool Cotton Nylon 6 Nylon 6-6 Wool Wool PolychlalT TBTOt N.D.: 1.9 1.5 2.0 Mean: 1.8 N.D.1 N.D.1 N.D.1 N.D.: N.D.: 7.5 6.9 7.9 Mean: 7.4 N.D. 2 12.1 11.8 12.4 Mean: 12.1 21.7 Polychlal 80%, nylon 6 20% Polychlal 85%, cotton 15% 5.9 * Confirmed by the American Association of Textile Chemists and t TBTs and DBTs are calculated as TBTO and DBTDC, respectivelv. Method 20-19771' and infrared spectroscopy. DBTDCt * N.D.f 210 201 198 203 N.D.: N.D. $ N.D.: N.D.: N.D. 1 0.3 0.2 0.5 0.3 N.D.1 555 578 569 567 245 261 Colorists' Test . * : Not detected. 9 Poly(ethy1ene terephthalate) . T[ A blended polymer of poly(viny1 chloride) and poly(viny1 alcohol), manufactured by Kohjin Co. Ltd., Japan, under the trade-name Cordelan. Triphenyltin and tricyclohexyltin compounds are used as agricultural chemicals, and can also be separated by using the difference in their partition behaviour between hexane and methanolic buffer solution. The method will be reported elsewhere. The author thanks Dr. Teruo Kan, Tokyo Metropolitan Research Laboratory of Public Health, for the flame-resistant fabrics. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. References Herold, B., and Droege, K. H., 2. Analyt. Chem., 1969, 245, 295. Woggon, H., and Jehle, D., Nahrung, 1973, 17, 739. Figge, K., and Bieber, W. D., J . Chromat., 1975, 109, 418. Williams, A. I., Analyst, 1973, 98, 233. Tonge, B. L., J . Chromat., 1965, 19, 182. Jitsu, Y., Kudo, N., Sato, K., and Teshima, T., Bunseki Kagaku, 1969, 18, 169. Gauer, W. O., Seiber, J. N., and Crosby, D. G., J . Agric. Fd Chem., 1974, 22, 252. Neubert, G., and Wirth, H. O., 2. Analyt. Chem., 1975, 273, 19. Matsuda, H., and Matsuda, S., Kogyo Kagaku Zasshi, 1960, 63, 108. Neumann, W. P., Angew. Chem., Int. Edn Engl., 1963, 2, 165. Aldridge, W. N., and Cremer, J. E., Analyst, 1957, 82, 37. Getzendaner, M. E., and Corbin, H. B., J . Agric. Fd Chem., 1972, 20, 88. Freitag, K.-D., and Bock, R., 2. Analyt. Chem., 1974, 270, 337. Sawyer, R., Analyst, 1967, 92, 569. Kaniwa, M., Mohri, J., Kojima, S., Nakamura, A., and Oba, T., Eisei Kagaku, 1977, 23, 7. Freeland, G. N., and Hoskinson, R. M., Analyst, 1970, 95, 579. American Association of Textile Chemists and Colorists, Tech. Man. Am. Ass. Text. Chem. Color., Received October 20th, 1978 Accepted February 13th, 1979 1977, 53, 57.

ISSN:0003-2654    

DOI:10.1039/AN9790400660

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

668 Analyst, Jub, 1979, Vol. 104, $9, 668-679 Method for the Rapid Detection of Organic Pollutants in Water by Vapour-phase Ultraviolet Absorption Spectrometry K. C. Thompson and K. Wagstaff Malvern Regional Laboratory, Severn-Trent Water Authority, 141 Church Street, Malvern, WR14 2A N A novel method for the rapid detection of organic pollutants in water utilising vapour-phase ultraviolet absorption spectrometry is described. The water sample is extracted with hexane or chloroform and a small amount (0.1-10 pl) of the extract is placed in a small graphite tube, which is slowly heated while the absorption of the sample is monitored. For optimum sensitivity for most substances a wavelength of 190 nm was used, and other wavelengths can additionally be used to characterise further and fingerprint the sample.Each trace can be completed within 2 min and the technique responds to many substances that are difficult to characterise or detect by gas - liquid chromatography. Keywords : Organic fiollutant detection ; vapour-phase ultraviolet absorption sfiectrometry ; oil GharaGterisation A difficult problem in most water laboratories is the identification of gross pollution in watercourses by oil and related substances where a rapid indication of the type of substance(s) present is required. Often it is necessary to compare the sample from a pollution incident with other samples from nearby watercourses or with samples of various wastes in order to identify the source of the pollution. A common screening technique is to extract the sample with a suitable volume of hexane and inject a small volume of this extract into a gas - liquid chromatograph fitted with a flame-ionisation detector.Common pollutants have a wide range of volatilities and temperature programming is normally used for this type of analysis. This can be time consuming if large numbers of samples are involved but, more important, many substances are difficult to detect using typical routine operating conditions. For example, vegetable oils and fats, many heavy oils, tar residues and resinous materials give poorly defined, non-specific gas chromatograms. Thus, if the substance is a mixture of a volatile (e.g., 28 s fuel oil) and a relatively involatile substance (e.g., heavy lubricating oil), the latter substance can easily be missed.This paper describes a novel technique that is simple, rapid and capable of comparative evaluations of many substances by monitoring the gas-phase ultraviolet absorption, at one or more wavelengths, of a very small amount of a sample as it undergoes slow evaporation and/or thermal pyrolysis in an atmosphere of nitrogen. The peak shape(s) and the appearance time(s) of each peak, and also subsequent measurements at other wavelengths, allow many types of substance to be characterised. Possible applications of the technique are indicated and Table I lists the wide range of materials examined during this study. Experimental Apparatus A Varian AA6 atomic-absorption spectrophotometer, a Cathodeon deuterium hollow- cathode lamp, a Varian CRASO electrothermal atomiser and a Servoscribe pen recorder were used.The CRASO unit was modified (see text) so that the minimum atomise ramp rate could be reduced from 25 "C s-1 to 11 "C s-l. The recorder chart speed was set to 120 mm min-l. Glass syringes with stainless-steel needles and capacities of 1 and 10 p1 were used to introduce the samples into the graphite tube. Solvent evaporations were carried out in specially modified 60-ml (25 mm i.d., 150 mm length) test-tubes1 with a 1.2-ml capacity (7 mm i.d., 30 mm length) bottom section that was calibrated at the 1-ml mark. The tubes were fitted with ground-glass stoppers.THOMPSON AND WAGSTAFF 669 TABLE I MATERIALS EXAMINED IN THIS STUDY Type of material Material Type of material Material substances. . . . Light lubricating oils substances.. .. Cooking oils Mineral oil related Vegetable oil related 20/50 lubricating oils Margarines Heavy lubricating oils Butter Gear oils Beef dripping Cutting oils Linseed oil Vaporising oil Transformer oils Soaps and detergent 28 s, 35 and 1 500 fuel oils related substances Various proprietary Creosotes anionic detergents Tars Various proprietary non- Vacuum oil ionic detergents Greases Various proprietary soaps Various fatty acids Other substances . . Anti-freeze Brake fluid Pentachlorophenol Polychlorinated biphenyls 2,4-Dichlorophenoxyacetic acid Reagents Analytical-reagent grade materials were used whenever possible. The sodium sulphate used in drying the extracts was heated at 500 "C in a muffle furnace and stored in a desiccator. Initial Studies It was found that if a small amount (0.1-1 pl) of a hexane solution of various types of fuel oils, lubricating oils, vegetable oils, etc., was introduced into the graphite tube of the Varian CRASO electrothermal atomiser and a slow rate of heating was applied during the atomisation stage, specific reproducible traces of absorbance versus time could be observed as the tube was heated from ambient temperature to about 900 "C. By changing the wave- length and running further traces, additional information about a given sample could rapidly be acquired.Optimisation of the Operating Conditions Rate of heating the graphite tube (ramp rate) Early in this study it was found that if the rate of heating (ramp rate) of the graphite tube was reduced from its original nominal minimum value of 25 "C s-l to 11 "C s-l, the resolution of the technique was considerably enhanced.This was achieved by adding an extra 4.7-pF capacitor across C110.2 This capacitor could be switched out of the circuit for normal use of the device. In order to prevent damage to the integrated circuit [No. MA103(308)] the 4.7-pF capacitor should only be switched in or'out of the circuit with the unit switched off. A ramp rate of 11 "C s-1 gave satisfactory sensitivity and resolution and thus was used in all further characterisation studies (see Table 11). It should be noted that this ramp rate was calculated from the time taken for the temperature (as indicated on the temperature readout meter) to increase from 100 to 500 "C. The ramp rate tended to increase with increasing temperature.The important criterion for this technique is that any variation in the actual ramp rate should be reproducible from firing to firing and this was found to be so. It was not feasible to reduce the ramp rate of the CRASO much below 11 "C s-l because of the risk of overheating the workhead, but a further reduction in the ramp rate may possibly improve the resolution of the technique. This would involve, however, greater volume loadings into the tube. Some work was carried out with a ramp rate of 50 "C s-l (see Temperature settings) when determining trace levels of certain substances after extraction into chloroform. Although this reduced the resolution of the technique, it considerably improved the sensitivity.670 THOMPSON AND WAGSTAFF: RAPID DETECTION OF ORGANIC POLLUTANTS Analyst, VOl. 104 Type of graphite tube The use of a grooved tube (Varian P/No.56-100223-00) tended to minimise the spreading of hexane solutions compared with the plain tube (Varian P/No. 56-100157-00). Each new tube was initially fired at 2300 "C ten times in order to "age" the graphite surface. The tube lifetime was in excess of 300 firings and towards the end of its useful lifetime a shoulder was often observed on the tail end of a normal peak. A Varian graphite cup (VarianP/No. 56-100159-00) was also evaluated but this gave less resolved traces than the grooved tube. If only one mask was fitted, as for trace metal work, the reproducibility was much poorer. Gas flow-rates The standard Varian light masks were fitted to both sides of the CRASO workhead in order to reduce air entrainment. A nitrogen flow-rate of 6 1 min-l was found to be optimum; at lower nitrogen flow-rates poor reproducibility was observed and at higher flow-rates no significant improvement was observed.An argon - methane (9 + 1 V / V ) flow of 0.20 1 min-l was used to prolong the life of the graphite elements3 It was also essential to have an efficient fume extraction system mounted directly above the device in order to prevent poor reproducibility caused by diffusion of any volatilised components away from the tube along the optical axis. Temperature settings Hexane sol.utions. The evaporation stage temperature of the CRASO was set to 35°C with a time of 20 s so that the device would always start from a fixed temperature slightly above ambient.The atomisation temperature was set to 900 "C with a hold time of 2.5 s and a ramp rate of 11 "C s-l. This resulted in a typical trace time of about 80 s and did not result in overheating of the metal support pillars of the CRASO. After approximately 25 firings the ramp rate was temporarily increased to 200 "C s-l and the maximum temperature to 1750 "C in order to volatilise any refractory residues that might have built up on the tube surface. No significant peaks or residues were ever observed on injecting 5 p1 of a hexane blank solution. For most typical oil pollution samples, emulsions were not observed in the hexane extract and thus anhydrous sodium sulphate was not required in order to dry the hexane extract (see below). Chloroform solutions.The evaporation stage temperature was set to 65 "C with a time of 30 s. It was essential to evaporate completely the chloroform as it exhibits very strong absorption at 190nm. The dry ashing stage was not used. For characterisation studies the atomisation temperature hold time and ramp rate settings were similar to those used for hexane (900 "C, 2.5 s and 11 "C s-l, respectively), but after each firing an additional firing was made with an increased atomisation temperature of 1750 "C and a ramp rate of 200 "C s-l. This higher atomisation temperature ensured that any sodium sulphate, used to dry emulsified chloroform extracts, that was transferred into the tube was completely volati- lised. For trace level studies of chloroform-extractable substances the atomisation tempera- ture was set to 1750 "C, with a ramp rate of 50 "C s-l and a hold time of 2.5 s.This increased ramp rate significantly increased the sensitivity, but decreased the selectivity of the technique. If an atomisation temperature of 1750 "C was used in conjunction with a ramp rate of 11 "C s-1, overheating of the CRA9O workhead occurred. The dry ashing stage was not used. Choice of wavelengths All substances were initially examined at 190 nm as many substances or their pyrolysis products exhibit significant absorption at this wavelength. This was considered to be the lowest useable wavelength of the spectrometer. At lower wavelengths complete absorption of radiation by oxygen molecules in the air path of the spectrometer will O C C U ~ .~ Two other wavelengths (210 and 253.7 nm) were also used. At 210 nm a number of common pollutants exhibit maximum absorption5 (e.g., pentachlorophenol) ; also, 253.7 nm is frequently used to monitor absorption of organic compounds.6 Almost all common solvents that are immiscible with water absorb appreciably at 190 nm and consequently will not allow conventional spectrophotometric measurements to be carried out a t this wavelength.July, 1979 IN WATER BY VAPOUR-PHASE ULTRAVIOLET ABSORPTION SPECTROMETRYZ 671 Reference standard preparation Solutions containing 10% m/V or V/V of each reference substance in hexane were prepared. For soaps and detergents 1 O O m l of a 2% m/V aqueous slurry were heated to boiling, cooled, acidified with 10ml of 25% V/V hydrochloric acid and extracted with 1OOrnl of chloroform, as the latter was found to be a better extracting solvent than hexane for these materials.Anhydrous sodium sulphate was used to reduce any emulsions that were formed and also to dry the chloroform extracts. Some of these final anionic detergent extracts were still appreciably emulsified. For brake fluid and anti-freeze 10% V/V solutions in water were used. In all instances it was found that if the neat substance was directly added to the tube, poor reproducibility was sometimes observed and often complete absorption of the light beam occurred even with sample volumes as low as 0.1 pl. Sample preparation Water samples that were obviously polluted with oil-type substances were extracted with a suitable volume of hexane or, if the pollutant was present as a separate phase, an approxi- mately 10% V/V solution of the pollutant in hexane was prepared.Ideally the concentra- tion(s) should finally be adjusted to produce peak heights similar ( &25y0) to the corresponding standard(s) for a given injection volume. For trace levels of relatively non-volatile substances 500-1000 ml of the sample (depending on available sample volume) were extracted with a suitable volume of chloroform and any emulsion was broken using sodium sulphate, which was also used to dry the extract. The extract was then evaporated to a final volume of 1 ml using a stream of dry nitrogen with the sample tube (see Apparatus) placed in a water-bath at 40 "C. NOTE- Chloroform vapour is very toxic and this operation must be carried out in an efficient fume cupboard.Operating Conditions Table I1 lists the operating conditions used in this study. 2 min and required 0.2-5 p1 of the sample extract solution. different wavelengths can take as little as 6 min. Each evaluation took less than Consequently, scans at three TABLE I1 Parameter Nitrogen flow-rate/l min-l . . .. Evaporation temperature/"C .. Evaporation time/s . . .. . . Atomisation temperature/"C . . .. Ramp rate/"(= s-1 . . .. . . Hold time/s . . .. . . . . Argon - methane flow-rate/l min-l . . Recorder chart speed/mm min-l . . OPERATING CONDITIONS Hexane solutions, Chloroform solutions, Chloroform solutions, characterisation characterisation* trace level work 6.0 0.2 35 20 900 11 120 2.5 6.0 0.2 65 30 900 11 120 2.5 6.0 0.2 65 30 1750 50 120 2.5 * After each trace was run an additional firing was made with an increased atomisation temperature of 1750 "C and a ramp rate of 200 "C s-l to remove any involatile residuals (see text).Results Oil and Petroleum Products Table I lists the substances examined, and all were found to give relatively characteristic reproducible traces, as shown in Figs. 1-5. Fig. 1 shows the traces obtained for a range of substances of widely differing volatilities and it can be seen that the shape of the traces and the ratio of the absorbances at 190 and 253.7nm give some form of fingerprint for each substance. For example, the ratio of the peak maxima absolute absorbances at 190 and 253.7 nm for the 28 s fuel oil (curve A) was 17, while for the light lubricating oil (curve C) it was 3.5 and for the earliest peak of the 1500 s fuel oil (curve E) it was 1.8.The appearance times of the peak maliima were reproducible to about -&2%, which re-emphasises the usefulness of the technique for characterising unknown samples. The peaks tended to672 THOMPSON AND WAGSTAFF : RAPID DETECTION OF ORGANIC POLLUTANTS Analyst, VoZ. 104 30 60 Time/s 0 30 60 Time/s Fig. 1. Typical oil traces. All solutions 10% V / V of oil in hexane; injection volume 0.5 pl; wavelength A, 28 s fuel oil; B, 35 s fuel oil; C, light lubricating oil; (a) 190 nm and (b) 253.7 nm; ramp rate 11 "C s-l. D, SAE 20/50 lubricating oil; and E, 1 500 s fuel oil. broaden very slightly as the graphite tube aged, probably owing to increased porosity of the graphite tube, and therefore trace comparisons were always carried out by sequentially running the necessary traces. Initial studies indicated that the peak heights were directly proportional to the concentration of the solute(s).Fig. 2 shows traces for a number of proprietary lubricating oils at 190 and 253.7 nm and again it can be seen that some form of fingerprint for each oil can rapidly be obtained. Fig. 3 shows traces for two mixtures of fuel oils and lubricating oils at 190 and 253.7 nm and in these particular instances resolved peaks were observed. Fig. 4 shows traces for a range of vegetable oils. At 190nm a very characteristic trace was obtained, which was virtually identical for all tested edible vegetable oils (three brands of margarine and five brands of cooking oil were tested).However, at 253.7 nm one of the edible vegetable oils gave an additional peak and the presence of this peak could be used to characterise that particular oil. Also, the appearance times of the main peak maxima at 253.7 nm showed a significant 0 30 60 Time/s 0 30 60 Ti me/s Fig. 2. Typical lubricating oil traces. All solutions 10% V / V of oil in hexane; injection volume 0.5 A, SAE 20/50 lubricating oil; B, pl; wavelength (a) 190 nm and (b) 253.7 nm; ramp rate 11 "C s-l. SAE 20/50 lubricating oil (different brand) ; C , SAE 30 lubricating oil; and D, SAE 85 lubricating oil.JUZY, 1979 IN WATER BY VAPOUR-PHASE ULTRAVIOLET ABSORPTION SPECTROMETRY 673 0, 8 + g 2 0 30 60 0 30 60 Time/s Fig. 3. Mixtures of fuel oil and lubricating oil.All solutions contained 10% V/V of each constituent in hexane; ramp rate 11 "C s-l. (a) A, 35 s fuel oil and SAE 85 lubricating oil, wavelength 190 nm, injection volume 0.5 p1; and B, as A, wavelength 253.7 nm, injection volume 2 p1. (b) A, 28 s fuel oil an'd SAE 20/50 lubricating oil, wavelength 190 nm, injection volume 0.5 p1; and B, as A, wavelength 253.7 nm, injection volume 2 pl. increase compared with those observed at 190 nm for all of the vegetable oils tested. This indicates that only the components in the tail of the peak observed at 190 nm significantly absorb radiation at 253.7 nm. Fig. 5 shows traces for an aqueous anti-freeze solution and an aqueous hydraulic brake fluid solution. Unlike many other samples, these exhibited no significant absorption at 253.7 nm.Fig. 6 shows traces for three different brands of chassis lubrication greases and it can be seen that the traces are substantially different. Fig. 7 shows traces for a new and a used motor lubricating oil, The used oil gave a similar trace to the new oil except for an initial hump, probably caused by petrol condensation products. The technique has also proved very useful when attempting to trace the origin of a gross pollution incident where a number of possible sources are known. In many instances a large proportion of the sources can be eliminated with a few minutes' work and in some instances the source can be positively identified. 0.5 0) C fl g 2 I 60 Time/s 0.2E 0) C e 2 2 Timels Fig. 4. Vegetable oil traces. All solutions 10% V/V of oil in hexane; wavelength (a) 190 nm and (b) A, B and C, three different brands of cooking oil, injection volumes (a) 0.2 pl 263.7 nm; ramp rate 11 "C s-l.and (b) 2 p l ; and D, linseed oil, injection volumes (a) 0.1 pl and (b) 0.6 pl. Soaps and Anionic and Non-ionic Detergents greater response than stearic or palmitic acid. Fig. 8 shows traces for some fatty acids and it can be seen that oleic acid gives a much Fig. 9 shows traces for three brands of soap674 THOMPSON AND WAGSTAFF: RAPID DETECTION OF ORGANIC POLLUTANTS Analyst, Vd. 104 1 A Time/s Fig. 6. Anti-freeze and hydraulic brake fluid trace. Both solutions 10% V/V in water; wavelength 190 nm; ramp rate 11 "C s-l; and evaporation temperature 96 "C. A, Anti-freeze, injection volume 2 pl; and B, brake fluid, injection volume 1 p1.and oleic acid, and it can be seen that the three soaps and the oleic acid exhibit similar trace shapes. Fig. 10 shows some traces for three brands of washing powders and sodium dodecyl- benzenesulphonate. The two brands that gave similar traces originated from the same manufacturer. It should be stressed that even though the extracts were dried with sodium sulphate, the washing powder and the sodium dodecylbenzenesulphonate extracts were badly emulsified. It was also found that for soaps and detergents it was good practice to remove any involatile residual substances by making an additional firing after each trace with an increased atomisation temperature of 1750 "C and a ramp rate of 200 "C s-l. Any build-up could lead to a variation in trace shape for some substances (e.g., oleic acid).A pollutant present at a low concentration in a small stream was tentatively identified as a soap by using this technique. This was achieved after extraction of an aliquot of the Time/s Fig. 6. Chassis lubrication grease traces. Grease (1 g ) shaken with warm hexane (10 ml) ; injection volume 1 pl; wavelength 190 nm; and ramp rate 11 "C s-l. A, B and C, three different brands of chassis lubrication grease.J d y , 1979 IN WATER BY VAPOUR-PHASE ULTRAVIOLET ABSORPTION SPECTROMETRY Ti me/s 0.5 0, C 5 2 675 G 5 :: :: : : : * : * . . . . . . . . . '. . . . . . . . . - . . * . : ; : ; 0 - 9 f A ..* 0 30 60 Fig. 7. New and used lubricating oil traces. Both solutions 10% V / V in Timeh hexane; injection volume 1 pl; wave- Fig.8. Some fatty acid traces. All solutions 2% length 190 nm; and ramp rate 11 m/ V in chloroform; wavelength 190 nm; and ramp rate "C s-l. A, Used SAE 20/50 lubricating 11 "C s-1. A, Stearic acid, injection volume 5 pl; B, oil; and B, new SAE 20/50 lubricating palmitic acid, injection volume 5 p1; and C, oleic acid, oil (same brand as A). injection volume 0.5 pl. acidified sample with chloroform and concentration by evaporation of the extract ten times. A similar extraction of another sample aliquot that was made alkaline did not give any significant response. Fig. 11 shows traces of four proprietary non-ionic detergents at the 5000 pg ml-l level in chloroform. These traces show that for a given amount of Triton X-100 (isooctylphenoxy- polyethoxyethanol) and Lissapol NX (alkylphenol 9-ethylene oxide) very similar traces are A B C Time/s Fig.9. Some typical soap and oleic acid traces. All solutions nominally 2% m/V in chloroform (see Reference Standard Preparation) ; wavelength 190 nm ; ramp rate 11 "C s-I. A, B and C, three different brands of soap, injection volume 2 pl; and D, oleic acid, injection volume 0.6 pl.676 THOMPSON AND WAGSTAFF: RAPID DETECTION OF ORGANIC POLLUTANTS Analyst, vd. 104 0.5 D ; 00 00 0 30 60 lime/s Fig. 10. Some washing powders and dodecylbenzenesulphonic acid traces. All solutions nominally 2'7; m/V in chloroform (see Reference Standard Preparation) ; wavelength 190 nm; ramp rate 11 "C s-l. A, B and C, three different brands of washing powder, injection volume 5 pl; and D, dodecylbenzenesulphonic acid, injection volume 1 pl.obtained, while for Tween 80 [polyoxyethylene (20) sorbitan monooleate] and Brij 35 (poly- oxyethylene lauryl ether) significantly different traces were observed. It can be seen from Fig. 11 that a sensitive response is obtained from polyethoxyethanol non-ionic detergents. These substances are particularly difficult and time consuming to determine by conventional techniques' and it appears that this method might be useful for 0.5 m 2 +? 2 a n ( Time/s Fig. 11. Some non-ionic detergent traces. All solutions contained 5 000 pg ml-1 of non-ionic detergent in chloroform; injection volume 4 p1; wavelength 190 nm; and ramp rate 11 O C s-l. A, Lissapol NX; B, Triton X-100; C , Tween 80; and D, Brij 35. rapid screening, and ultimately even as a measurement procedure for these types of sub- stances.A very brief investigation was conducted by taking 500-ml aliquots of a blank, a river and the same river spiked with 2 pug ml-l of Triton X-100 and extracting with three 20-ml aliquots of chloroform; the resulting extract was dried using anhydrous sodium sulphate and then 50 ml of the extract were evaporated to a volume of 1 ml. A 5-p1 volume of the extract was then added to the graphite tube and Fig. 12 shows the very significantJ%@, 1979 IN WATER BY VAPOUR-PHASE ULTRAVIOLET ABSORPTION SPECTROMETRY 677 signal for 2 pg ml-1 of a typical non-ionic detergent. To improve the sensitivity a ramp rate of 50 “C s-1 was used. This relatively rapid technique has been used for screening rivers for specific non-ionic detergents using the method of standard additions; most routine samples give a “less than’’ figure; any samples that give a positive result are then investi- gated using conventional procedures.’ 0.25 aJ t -9 a a Q 0 30 0 3c Time/s Fig.12. Screening for Triton X-100 in a river sample. Concentration factor 500 times (see text) ; injection volume 5 pl; wavelength (a) 190 nm and (b) 210 nm; and ramp rate 50 “C s-l. A, River sample plus 2 pg ml-l Triton X-100; B, same river sample as used in A; and C, blank. Other Applications Fig. 13 shows a trace for the extract from a river that was thought to have been grossly polluted with pentachlorophenol and a sample from the same river spiked with 1 pgml-l of sodium pentachlorophenate.A 500-ml volume of the sample was taken, 10 ml of 1 M sodium hydroxide solution were added and the sample was extracted with 50 ml of chloroform and the chloroform layer rejected; the sample was then acidified with 20 ml of 1 M hydro- chloric acid and extracted with 25ml of chloroform. The chloroform extracts were evaporated to a final volume of 1 ml and 5 pl were added to the graphite tube. It can be seen from Fig. 13 (b) that the river contained less than 0.05 pg ml-l of pentachlorophenol. It can also be seen that the absorption for pentachlorophenol is much greater at 210nm than at 190 nm, whereas for Triton X-100 (see Fig. 12) the absorption is much greater at 190 nm. General Pollution Monitor for Chloroform-soluble Substances Although it is possible to monitor the ultraviolet absorption spectrum of aqueous samples directly,6 this technique has several disadvantages, vix., absorption by inorganic ions [e.g., iron(III), nitrate] limits the lower useful wavelength to about 230 nm and many substances do not exhibit significant absorption above 230 nm; no indication of the volatility of the pollutant is given; and there is less likelihood of resolving mixtures of compounds and the sensitivity of the technique for many substances is not very good.The proposed vapour- phase ultraviolet absorption spectrometry effectively overcomes these disadvantages, viz., measurements can be made down to 190nm (or even lower if nitrogen purged optics are used4) and thus almost all extractable substances will exhibit some absorption; the appearance times of the various peak(s) in conjunction with their wavelength dependence can act as a general fingerprint for a sample and give an indication of the range of volatility of the678 THOMPSON AND WAGSTAFF: RAPID DETECTION OF ORGANIC POLLUTANTS Analyst, VoZ.104 extracted material; and the sensitivity can be increased by increasing the sample volume to final extract volume ratio. A I Time/s Fig. 13. Screening for pentachlorophenol in a river sample. Concentration factor 600 times (see text) ; injection volume 5 pl; wavelength (a) 190 nm and (b) 210 nm; and ramp rate 50 "C s-l. A, River sample plus 1 pg ml-l of sodium penta- chlorophenate; B, same river sample as used in A; and C, blank. Fig. 14 shows some traces for some rivers and a sewage works final effluent after extrac- tion of 1 1 of the sample with 50 ml of chloroform, placing 30 ml of the extract into a modified test-tube (see Apparatus) maintained in a water-bath at 40 "C and then evaporating it to a final volume of 1 ml using a stream of dry nitrogen.The traces were run at ramp rates of 11 and 50 "C s-l and it can be seen that the former ramp rate gives better resolution while the latter gives better sensitivity. Thus, by regularly monitoring a given point on a river any significant change in the extractable material can be observed and possibly give early warning of pollution (see Figs. 12 and 13). Ultimately it should be possible to develop a continuous monitor for this aspect of the work. For continuous monitoring applications the evaporation step could be performed by slowly introducing a large volume (500~1) of the initial extract on to the actual heating element whilst maintaining the element at 65 "C prior to the ramp atomisation stage.A volatile solvent such as pentane that does not absorb radiation at 190 nm could be used so that more volatile substances could be detected, but this could result in a poorer extraction efficiency for some substances. Conclusions This preliminary study has shown that monitoring the vapour-phase ultraviolet absorption of hexane or chloroform extracts of natural water samples at two or more wavelengths appears to be a rapid and sensitive method for both the detection and the comparative evaluation of gross organic pollution. Characterisation of some types of pollutants is possible using appearance time@) and peak shape(s).It should be possible to develop methods to allow rapid screening for trace levels of a number of organic substances (e.g., non-ionic detergents). The technique has been used at the Malvern Regional Laboratory for some time and has already proved very useful for tracing the cause of various organic pollutions. It is also a useful indication when setting up the gas-chromatographic conditions for identification of a pollutant and the technique should be regarded as a rapid complementary technique to gas chromatography. It should be stressed that although these results were obtained by using an electrothermal atomiser and an atomic-absorption spectrophotometer, a much simpler system could beJuuly, 19’19 IN WATER BY VAPOUR-PHASE ULTRAVIOLET ABSORPTION SPECTROMETRY 679 envisaged.A low-pressure mercury lamp, interference filters and a solar-blind photo- multiplier should allow for isolation of the wavelengths 194.2, 253.7 and 313.2nm. A simple electrically heated graphite rod or tube, a metal strip or an alumina tube could feasibly be used in place of a sophisticated electrothermal atomiser. A maximum tempera- ture of 1800 “C with a variable heating rate from about 10 to 50 “C s-l would be required. 0.E aJ C -e 2 a n 0 30 Ti me/s 0.2 Q) C -E % 2 a 30 60 Ti me/s Fig. 14. Some river and a sewage works final effluent traces. All solutions represent extraction of 1 1 of sample into 1 ml of chloroform (see Sample Preparation); wavelength 190 nm; (a) ramp rate 60 “C s-1, injection volume 5 p1; (b) ramp rate 11 OCs-l, injection volume 10 pl. A, Sewage works final eBuent; B, C and D, river samples (sample D was from a particularly clean river) ; and E, blank. The technique could also possibly be adapted as a liquid-chromatographic detection system. It might also prove useful for comparison and fingerprinting studies of various types of samples in forensic analysis (e.g., paints, oils and dusts). It is felt that this technique has considerable potential and numerous other applications and instrumental modifications can be envisaged. The authors thank Mr. W. F. Lester, Director of Scientific Services, Severn-Trent Water Authority, for permission to publish this work and also Mr. D. Evans of Varian Associates Limited for carrying out the electronic modifications to the Varian CRASO electrothermal atomiser. References 1. Thompson, K. C., and Wagstaff, K., “Detection of Organophosphorus and Organosulphur Com- pounds in Rivers and Some Effluents,” Malvern Regional Laboratory Report, No. ML17, Malvern, 1977, Appendix 1, p. 8. “Varian CRASO Carbon Rod Atomizer Handbook,” Varian Associates, Springvale, Australia, 1976, Fig. 8.6. Thompson, K. C., Godden, R. G., and Thomerson, D. R., AnaZytica Chcim. Acta, 1976, 74, 289. Thompson, K. C., and Reynolds, R. J., “Atomic Absorption, Fluorescence and Flame Emission Spectroscopy, A Practical Approach,” Charles Griffin, London and High Wycombe, 1978, p. 213. Gore, R. C., Hannah, R. W., Pattacini, S . C., and Porro, T. J., J . Ass. Ofl. AnaZyt. Chem., 1971, 54, 1040. Briggs, R., and Melbourne, K. V., J . SOC. Wat. Treat. Exam., 1968, 17, 107. Department of the Environment, “Analysis of Raw, Potable and Waste Waters,” H.M. Stationery Received December 14th, 1978 Accepted February 2nd, 1979 2. 3. 4. 6. 6. 7. Office, London, 1972.

ISSN:0003-2654    

DOI:10.1039/AN9790400668

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

680 SHORT PAPERS Analyst, July, 1979 Cathode-ray Polarographic Determination of Molybdenum in Serum, Plasma and Urine Gary D. Christian* and Gaston J. Patriarche Institut de Pharmacie, Universite' Libre de Bruxelles, Campus Plaine 2051 1, Brussels, Belgium Keywords Molybdenum determination ; blood ; urine ; cathode-ray polarography Molybdenum was identified in 1953 as a factor in xanthine oxidase1y2 and is now known to be an integral part of aldehyde oxidase and other flavoen~ymes.~ This element is of bio- logical interest, it occurs as an ultra-trace element, but it is difficult to analyse. Molybdenum(V1) exhibits a catalytic polarographic wave in the presence of perchloric4 or nitric acid5 that is 100 times more sensitive than conventional diffusion-controlled waves ; the limit of detection in 2.4 M nitric acid is 5 x 10-8 M of molybdenum6 Bikbulatova and Sinyakova7 have used this catalytic wave with 1 M sulphuric acid or 1 M potassium nitrate solution for the polarographic determination of trace amounts of molybdenum in soils and plant ash.We have recently investigated the cathode-ray polarography of this sensitive catalytic wave, and found that in this mode a concentration of 3 x M (0.0003 p.p.m.) of molybdenum(V1) could be determined in 2 M potassium nitrate solution at pH 1.6-2.2.lO A 400-fold excess of copper(I1) and a 200-fold excess of cadmium(I1) did not interfere in the determination. This cathode-ray polarographic method has been applied to the determination of molyb- denum in human plasma and urine. The molybdenum is separated by a convenient solvent- extraction procedure for rapid determination. Toropova et aZ.ll have determined molyb- denum in blood using a catalytic bromate ion wave on a graphite-disc electrode.A 1500-fold excess of iron(II1) did not interfere, but chromium did interfere. Cathode-ray polarography8 99 offers added sensitivity over conventional polarography. Experimental Reagents Reagent-grade chemicals were used throughout. A stock 1000 p.p.m. molybdenum(V1) solution was prepared by dissolving 2.52 g of sodium molybdate (Na2Mo04.2H20) in 1 1 of water.12 Working standard solutions of 0.05, 0.1 and 0.2 p.p.m. of molybdenum for the blood analyses and 1, 2 and 4 p.p.m. of molybdenum for the urine analyses were prepared by dilution of the stock solutions. The supporting electrolyte was 2.0 M potassium nitrate solution with the pH adjusted to 2.2 with sulphuric acid.Apparatus Polarographic measurements were made with a Chemtrix Single Sweep Polarographic Analyzer System, Model SSP-3. The working electrode was a dropping-mercury electrode with a drop time of 8 s (open circuit) and the potential was scanned during the lifetime of a single drop. The anode was a mercury pool electrode. The polarographic cell was as previously described.13 This was designed to hold as little as 1 ml of solution. The starting potential was set at -0.3 V against the mercury pool anode. A 0.5-V scan in the negative direction, started 1.0s after the start of the drop, was made in 0.1 s. The solution was de-aerated with nitrogen for 10 min before running the polarogram. A nitrogen atmosphere was maintained above the solution during the running of the polaro- gram.* Permanent address : Department of Chemistry, University of Washington, Seattle, Wash. 98196, USA.SHORT PAPERS 681 Procedure Dry ash 10 ml of serum, plasma or urine at 600 "C as follows: evaporate to dryness over a Bunsen burner or steam-bath in a porcelain crucible; heat gently until smoking ceases and increase the heat until most of the charring is complete and then place in a furnace at 600 "C for at least 3 h. Cool and transfer with two 2-ml portions of 6 N hydrochloric acid into a 25-ml separating funnel, taking care to break up the residue with a glass rod. Add 2 ml of pentyl acetate and extract the molybdenum into this by shaking for 1 min.Allow the layers to separate and discard the aqueous layer. Back-extract the molybdenum from the organic layer into 1.00 ml (pipetted) of the supporting electrolyte by shaking for several seconds. Add the aqueous electrolyte layer to the polarographic cell, de-aerate and run the polarogram. Measure the height of the polaro- graphic wave. Prepare a blank, starting with the solvent-extraction step (4 rnl of 6 N hydrochloric acid). Prepare a calibration graph in a similar manner by taking 1 ml of the 0.05, 0.1 and 0.2 p.p.m. molybdenum standards used for the serum and plasma analyses, or 1 ml of the 1, 2 and 4 p.p.m. molybdenum standards used for the urine analyses. Correct the samples and standards for any blank reading and determine the amount of molybdenum in the sample from the calibration graph.Results and Discussion A typical recording of the molybdenum(V1) catalytic wave is shown in Fig. 1. No appreciable molybdenum was observed in the blank. The half-wave potential of about -0.68 V is 100 mV more positive than that obtained at pH 2.2,10 indicating that some of the acid is back-extracted to decrease the pH slightly (at pH 1.6 the would be -0.65 V). -0.4 -0.5 -0.6 -0.7 -0.8 -0.9 E versus Hg pool Fig. 1. Reproduction of cathode- ray polarogram of molybdenum cata- lytic wave. Molybdenum(V1) (0.2 p.p.m); 2 M KNO,, pH 2.2. Initial potential -0.3 V vevsus Hg pool, 0.05 V per division; scan range 0.5 V; current scale 1 pA per division. Equimolar concentrations of electroactive substances, which are reduced or oxidised at the same potential at which the molybdenum wave occurs, do not interfere, and at least 500-fold excesses can be tolerated if the waves are separated by 0.1-0.2 V.lO The molyb- denum(V1) is quantitatively separated from the bulk sample matrix by solvent extraction from 6 N hydrochloric acid into pentyl acetate.14 Of the elements present in those which would be expected to extract15 in amounts nearly equal to or greater than the amount of molybdenum include only arsenic, iron, vanadium and, to a lesser extent, cobalt.The last two elements would be present in amounts too small to interfere, even if they were polarographically reducible at the same potential as molybdenum. The maximum concentration of arsenic in normal blood would be expected to be about 0.2 p.p.m., approxi- mately 300-fold greater than the molybdenum concentration.Iron in serum or plasma is682 SHORT PAPERS Analyst, Vol. 104 usually about 1 p.p.m., but in whole blood it is 500 times this concentration. Only arsenic and iron in urine would be expected to be extracted in amounts equal to molybdenum and, to a lesser extent, antimony. These would not be expected to interfere. In addition, there has been one report of appreciable amounts of germanium16 and tellurium17 occurring in the urine; the former would be extracted efficiently while the latter would be extracted only partially. It is doubtful if appreciable amounts of tellurium would remain after dry ashing, and considerable volatilisation of arsenic would occur. Of the elements mentioned above, only large amounts of iron have been found to interfere in the molybdenum determination.Ten parts per million of arsenic added to urine and run through the procedure did not interfere. Molybdenum measurements at the concentra- tions encountered in this work could be made in the presence of up to 1 p.p.m. of iron(III), but measurements could not be made with 10 p.p.m. of iron present. As iron(II1) is extracted efficiently from 6 N hydrochloric acid, this precludes measurements of molybdenum in whole blood. Measurements can be made satisfactorily with serum or plasma; for highest accuracy, the polarogram should be run within 15 min, as the iron tends to cause interference after standing for a period of time. Molybdenum could be determined in whole blood by extracting it with benzoin a-oxime from acidic solution into chloroform,ls and then evaporating the solvent and digesting the complex to obtain inorganic molybdenum(VI), as has been done with soil and plant samples'; the molybdenum is extracted from 1-2% sulphuric acid with 3 ml of a 2% alcoholic solution of benzoin a-oxime into 10 ml of chloroform7 and of all the common metal ions only molyb- denum and tungsten extract.ls Direct solvent extraction of added molybdenum(V1) from plasma or urine samples acidified with 6 N hydrochloric acid resulted in a recovery of less than 5%, suggesting that the molyb- denum is tightly bound to proteins or other constituents.Hence, the ashing step is necessary . Table I summarises the results obtained by the recommended procedure in the analysis of human urine and plasma samples.A sample of each was also analysed by a catalytic procedure12 and the results are given in the table. Excellent agreement was obtained for each sample, indicating that the suggested method is accurate. In addition, the over-all results of the polarographic method agree well with those found previously by the catalytic method.12 They also agree very well with what other workers have found for molybdenum in whole blood or ~ e r u m l ~ - ~ ~ ; molybdenum is evenly distributed between plasma and the red ~ e 1 l s . l ~ ~ ~ ~ (See reference 12 for a summary of the results of other workers.) TABLE I CATHODE-RAY POLAROGRAPHIC DETERMINATION OF MOLYBDENUM IN HUMAN PLASMA AND URINE Molybdenum found, p.p.m. f A > Sample* Polarographic method Catalytic methodl2 Plasma .... 0.006, 0.006 0.006, 0.005 Plasma . . .. 0.005 Plasma . . . . . . 0.008 Urine .. .. 0.22, 0.22 0.21, 0.20 Urine . . .. 0.22 Urine .. .. 0.40 * Volume of sample taken was 10 ml. Samples of 1 O m l were taken for analysis in order to obtain maximum sensitivity of the method, however, is sufficient that 1-ml samples could be accuracy. The analysed. The polarogriphic method is preferred over the catalytic method, where applicable, bkcause it is convenient, it is more selective and it is operable over a wider range of concentrations. However, it is not applicable when large amounts of iron are present unless the molybdenum is separated from the iron, whereas large amounts of iron do not affect the catalytic method. Thanks are expressed to the Commission for Educational Exchange between the USA, Belgium and Luxembourg (G.D.C.) and the Fonds de la Recherche Scientifique (F.N.R.S.Belgium) (G. J. P.) for support of this research.July, 1979 SHORT PAPERS 683 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. De Renzo, E. C., Kaleita, E., Heytler, P., Oleson, J . J., Hutchings, B. L., and Williams, J. H., Archs Biochem. Biophys., 1953, 45, 247. Richart, D. A., and Westerfield, W. W., J . Biol. Chem., 1953, 203, 915. Christian, G. D., and Feldman, F. J., “Atomic Absorption Spectroscopy. culture, Biology, and Medicine,” John Wiley, New York, 1970, p. 293. Haight, G. P., Analyt. Chem., 1951, 23, 1505. Johnson, M. G., Analyt. Chem., 1952, 24, 366.Violanda, A. T., and Cooke, W. D., Analyt. Chem., 1964, 36, 2287. Bikbulatova, R. U., and Sinyakova, S. I., Agrokhimiya, 1967, 123. Randles, J. E. B., Trans. Faraday SOC., 1948, 44, 344. Davis, H. M., and Kooney, R. C., J . Polarogr. Soc., 1962, 8, 25. Christian, G. D., Patriarche, G. J., and Vandenbalck, J. L., Analytica Chim. Acta, in the press. Toropova, V. F., Vekslina, V. A., Vashchenko, S. A., and Kitaeva, L. N., Nov. Polyarogr., Tezisy Dokl. Vses. Soveshch. Polarogr., 6th, 1975, 183. Christian, G. D., and Patriarche, G. J., Analyt. Lett., 1979, 12, 11. Newberry, C. L., and Christian, G. D., J . Electroanalyt. Chem., 1975, 9, 468. Taylor, R. P., Thesis, Rutgers University, 1960. Morrison, G. H., and Freiser, H., “Solvent Extraction in Analytical Chemistry,” John Wiley, New Schroeder, H. A., and Balassa, J. J., J . Chron. Dis., 1967, 20, 211. Schroeder, H. A., Buckman, J., and Balassa, J. J., J . Chron. Dis., 1967, 20, 147. Jeffrey, D. G., Analyst, 1956, 81, 104. Mertz, D. P., Koschnick, R., Wills, G., and Pfeilsticker, K., 2. Klin. Chem. Klin. Biochem., 1968, Allaway, W. H., Kubota, J., Losee, F., and Roth, M., Archs Envir. Hlth, 1968, 16, 342. Butt, E. M., Nusbaum, R. E., Gilmore, T. C., Didio, S. L., and Sister Mariano, Arch Envir. Hlth, Brune, D., Samsahl, K., and Westev, P. O., Cliozica Chim. Acta, 1966, 13, 285. Bala, Y. U., and Lifshits, V. M., Fedn Proc. Fedn Am. Socs Exp. Biol. (Trans. Sufipl.), 1966, 25, Received January 19th, 1979 Accepted February 27th, 1979 Applications in Agri- York, 1957. 6, 171. 1964, 8, 52. T370.

ISSN:0003-2654    

DOI:10.1039/AN9790400680

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

July, 1979 SHORT PAPERS 683 Improved Method for the Determination of Mercury in Fish Tissue Using 50% Hydrogen Peroxide and a Hot Block J. W. Davidson Department of Fisheries and Environment, West Vancouver, British Columbia, Canada, V7 V 1 N8 Keywords Mercury determination ; fish ; 50% hydrogen peroxide digestion Trace mercury determination in fish tissue by the cold vapour atomic-absorption method is well documented. Preliminary tissue destruction with acid - permanganatel in this labora- tory, failed to oxidise completely fatty tissue such as the liver of dog fish (Squalzcs sp.) and rat fish (Hydrolagus colliei). The non-destruction of fats and oils resulted in globules of floating fat in the final digest. Complete digestion of fish tissue is essential in order to destroy any fat capable of adversely affecting mercury volatilisation after reduction,2 to destroy volatile compounds that absorb in the mercury 253.7-nm region (aromatic hydrocarbons, sulphur dioxide, nitrogen d i ~ x i d e ) ~ and to separate mercury from the biological matrix.4 Other combinations of acids and oxidants4-6 were evaluated to assess their properties in maximising destruction of organic matter and recovery of mercury and in minimising analysis time, oxidation temperature, sample handling, amounts of reagents and utilisation of complicated glassware.These criteria were achieved by combining the excellent oxidation characteristics of sulphuric acid and 50% hydrogen peroxide’-9 with the simplicity of a “hot block.”lO684 SHORT PAPERS Analyst, Vol.104 Experimental Apparatus A 100-ml, Pyrex brand, “Student” graduated cylinder (CGW 3075-100 without plastic base) was used as the reaction tube. The cylinder was truncated at the 100-ml mark to ensure a constant total volume from tube to tube. Sample weighing, digestion and reduction were performed in the reaction tube to eliminate sample transfer and minimise surface contact. The long tubular shape minimised mechanical and volatile losses during digestion and enhanced mercury diffusion during volatilisation because of the extended bubble contact time. The hot block was fabricated from three 0.75in thick aluminium sheets. Two sheets were welded together and a series of 1.25 in diameter holes were drilled through the sheets. The third sheet was welded to the drilled sheets to form a base plate.An open system (Fig. 1) was adopted, as a closed system involved a time-consuming summation of absorbance peaks’ and incurred reproducibility prob1ems.l The drying tube was omitted because no condensation occurred on cell windows when the apparatus, reagents and samples were at room temperature and the cell was constantly purged between samples. Small-diameter connecting tubing and elimination of the drying tube lowered the total volume of the system, thereby minimising memory problems. The reduced volume also perrnitted a smaller ratio of pre-cell to cell volume, thus increasing ~ensitivity.~J Argon was used as a sweep gas to eliminate potential mercury contamination (Rains and Menisll found that in all their tests argon was superior to air as argon exhibited a 25% increase in absorbance for mercury).The 15 cm long absorption cell was mounted in the beam path of a Jarrell-Ash, Model 810, atomic-absorption spectrophotometer. Mercury was determined at 253.7 nm with simultaneous background correction at 254.7 nm using a tin hollow-cathode lamp. I G Fig. 1. Open cold-vapour system for atomic-absorption spectrometric determination of mercury. A, Argon inlet; B, flow meter; C, three-way valve; D, purge; E, reductant pipettor; F, gas dispersion tube; G, reaction tube; H, absorption cell; and I, mercury trap. Reagents All reagents were of analytical-reagent grade. De-ionised water was used in all reagent preparations. Sdfihuric acid, 4 + 1. Four parts of concentrated sulphuric acid were added to one part of water and the mixture was allowed to cool before use.Hydrogen peroxide, 50% V/V. This reagent was tested for the presence of mercury in the following manner. To 4ml of 50% hydrogen peroxide were added 45ml of water. The solution was covered and left overnight, then 10 ml of 4 + 1 sulphuric acid and 1 mlJuly, 1979 SHORT PAPERS 685 of 5% m/V potassium permanganate solution were added. A similar blank was prepared with water instead of 50% hydrogen peroxide. Both blanks were then analysed in the usual manner. At twice the peroxide concentration used in the digestion, no detectable difference between the blanks was found. Potassium permanganate solution, 0.1 yo m/V. One gram of potassium permanganate was dissolved in 1 1 of water and refrigerated before use.As this reagent is the main potential source of mercury contamination, it may contain mercury itself or may extract it from container walls and/or air.12 Reducing solution ; 2% m/V hydroxylammonium sulpphate, 1 yo m/V lzydrazirtium sulphate, 3% m/V tin(l1) chloride. Ten grams of hydroxylammonium sulphate [(NH,OH),.H,SO,] were dissolved with stirring in 300 ml of water, 5 g of hydrazinium sulphate (H,NNH,.H,SO,) were added and stirred until dissolved, then 15 g of tin(I1) chloride (SnC1,.2H20) were added and dissolved. The entire solution was diluted to 500ml with water. The hydro- xylammonium sulphate destroyed excess of permanganate and maganese( IV) oxide. Hydrazinium sulphate accelerated the reaction of the tin(I1) ch10ride.l~ Calibration standards were prepared by dilution of a 1000 pg ml-I mercury reference solution (Fisher Scientific Co., Certified Atomic Absorption Standard).Each standard contained 10.0 ml of 4 + 1 sulphuric acid, 46.0 ml of 0.1% m/V potassium permanganate solution and was diluted to 60.0ml with de-ionised water. A range from 0.01 to 0.5 pg of mercury covered most mercury concentrations found in marine biota. Mercury standard. Procedure An approximately 0.100-0.200-g portion (less if high mercury levels are known to exist) of homogenised, freeze-dried and ground tissue (or 0.500-1.00g wet mass) was weighed into each reaction tube. A 10-ml volume of 4 + 1 sulphuric acid was added to each tube, and the tubes were covered and left to stand overnight. A 4-ml volume of cold (4 "C) 50% m/V hydrogen peroxide was mixed in and the tubes were placed on the hot block.This setting resulted in the hot-plate obtaining, without the block, a surface temperature of approxi- mately 250 "C within 1 h. The block itself was cooler and its temperature varied with loading, but the temperature of the sample solutions did not exceed 80 "C. As the reaction gradually became more vigorous, the solutions began to froth. After frothing had stopped (approximately 1-1.25 h), reaction was considered to be complete, at which time the solutions were clear and colourless and the tubes were removed from the hot block. (When the tubes were left too long on the hot block, the solutions became slightly coloured; however, the effect on the accuracy of the results was negligible.) The tubes were cooled in a cold water-bath and 46.0 ml of cold (4 "C) 0.1% m/V potassium permanganate solution were added in a steady stream to ensure complete mixing.The required final volume was 60 ml. With argon flowing through the gas dispersion tube at 40 ml min-l (determined empirically for maximum absorption), a reaction tube that contained a standard or sample was placed on the cold vapour apparatus and 6.0 ml of reductant were injected forcefully to facilitate mixing. When the maximum peak height had been obtained, the tube was removed, the cell was purged and the argon flow was switched to the gas dispersion tube. This process was repeated for the remaining tubes. Tests showed that waiting for a return to the base line between samples was not necessary and rinsing the gas dispersion tube between samples was also not necessary as carry-over was not detectable in either instance.A calibration graph of amount of mercury versus peak height was prepared, and was linear up to 0.5 pg of mercury. The base-line noise level was approximately 0.25 unit, the blank was 2 units, the 0.01-pg mercury standard was 4 units and the 0.5-pg mercury standard was 100 units. The concentration of mercury in the sample was calculated on a wet and a dry basis. Heating was started by turning on the hot-plate to a high-medium setting. Results and Discussion Mercury Determinations Mercury content was determined for Albacore tuna [National Bureau of Standards (NBS), Research Material 501 and for eight samples received as part of an on-going mercury check sample programme instituted by Environment Canada, Fisheries and Marine Service,686 SHORT PAPERS Analyst, Vol.104 Winnipeg, Manitoba. Each of the eight samples was composed of a homogeneous mixture of varying proportions of tissue from three fish species, namely pickerel (Stizostediolz viZveum) , pike (Esox Zucizls) and lake trout (SalveZinus namaycush). No samples were spiked by our laboratory as the validity of spiking is questionable.2 Nine replicate samples of NBS tuna were analysed by the 50% hydrogen peroxide method to determine the repeatability of the method. The mean and standard deviations were 1.00 pg g-1 and 0.02 pg g-l dry mass, respectively. The NBS reported a value of 0.95 & 0.1 pg g-1 and indicated that 80-90% of the mercury content is present as methylmercury.Each of the eight check samples was analysed in triplicate and the mean value and range reported on a calculated wet mass basis for comparison with other reported values (Table I). The corrected mean values and standard deviations of the results reported by the partici- pating laboratories were calculated by rejecting those values which were greater than two standard deviations from the mean, then repeating the process until all values were less than two standard deviations from the mean. TABLE I RESULTS FOR MERCURY CONTENT IN EIGHT SAMPLES OF FISH TISSUE MIXTURE Mercury content/pg g-1 Sample A No. This method Other methods 1 0.24 f 0.01 0.24 & 0.03 2 0.96 f 0.01 0.91 f 0.07 3 0.50 f 0.01 0.52 f 0.04 4 0.39 f 0.01 0.42 f 0.03 5 1.11 & 0.00 1.04 f 0.06 6 0.57 f 0.02 0.56 f 0.03 7 0.46 f 0.01 0.47 f 0.03 8 0.33 f 0.01 0.33 f 0.04 -\ No.of reporting Range laboratories 0.18-0.30 22 0.77- 1.04 20 0.46-0.59 21 0.37-0.48 21 0.92-1.15 22 0.53-0.6 1 21 0.44-0.54 24 0.27-0.4 1 26 Advantages of the method No loss of mercury has been associated with this procedure.14Js Differences in moisture content due to natural causes, variations in sample preparation and variations in freeze and storage time were eliminated by freeze- drying. As only a small portion of the sample was taken for mercury determination, the remainder was stored for subsequent analysis of metals other than mercury. The use of 4 + 1 sulphuric acid eliminated sample charring during dissolution and reduced the heat evolved during dilution with hydrogen peroxide.The maximum initial critical oxidation temperature reached 45 "C and the maximum final oxidation temperature on the hot block reached approximately 75-80 "C. These low temperatures contributed to mini- misation of mercury losses. The dilute permanganate solution destroyed any excess of peroxide and maintained the sample in an oxidising media. Refrigeration of the permanganate solution ensured that when it was added to the sample solution, the solution reached room temperature more quickly. In a %day period, one person can analyse 100 freeze-dried or wet tissue samples in dupli- cate (200 determinations) with a detection limit (twice the blank level) of 0.01 pg of mercury, or approximately 0.07 pg g1 dry mass or 0.013 pg g-l wet mass, based on a sample size of 0.15 g dry or 0.75 g wet mass, respectively.The limits can be adjusted by varying sample size or by using an absorption cell that has a longer path length. In summary, the 50% hydrogen peroxide method using a hot block quickly and com- pletely digested a variety of fish tissues and other marine biota using a minimum number and amounts of reagents and glassware. Results of mercury determinations in the test fish were precise and in excellent agreement with those obtained by other methods. All samples were freeze-dried before oxidation. The author thanks the staff of the Pacific Region Laboratory Services in Vancouver for their technical support and advice during the development of this method.July, 19 79 SHORT PAPERS References 687 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Uthe, J. F., Armstrong, A. J., and Stainton, M. P., J . Fish. Res. Bd Can., 1970, 27, 805. Uthe, J. F., Armstrong, A. J., and Tam, K. C., J . Ass. Off. Analyt. Chem., 1971, 54, 866. Kopp, J . F., Longbottom, M. C., and Lobring, L. B., J . Am. Wat. Wks Ass., 1972, Jan., 20. Stuart, D. C., Analytzca Chim. Acta, 1978, 96, 83. Munns, R. K., and Holland, D. C., J . Ass. 08. Analyt. Chem., 1977, 60, 833. Analytical Methods Committee, Analyst, 1977, 102, 769. Down, J. L., and.Gorsuch, T. T., Analyst, 1967, 92, 398. Analytical Methods Committee, Analyst, 1967, 92, 403. Analytical Methods Committee, Analyst, 1976, 101, 62. Hendzel, M. R., and Jamieson, D. M., Analyt. Chem., 1976, 48, 926. Rains, T. C., and Menis, O., J . Ass. 08. Analyt. Chem., 1972, 55, 1339. Christmann, D. R., and Ingle, J. D., Jr., Analytica Chirn. Acta, 1976, 86, 53. Hoover, W. L., Melton, J . R., and Howard, P. A., J . Ass. 08. Analyt. Chem., 1971, 54, 660. LeFleur, P. O., Analyt. Chem., 1973, 45, 1534. Turkey, A., and Hornung, H., Atom. Absorption Newsl., 1978, 17, 59. Received Sefitember 19th, 1978 Accepted February 12th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9790400683

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

July, 1979 SHORT PAPERS 687 Preparation and Operation of Selenium Electrodeless Discharge Lamps for Use in Atomic- f I uorescence FI a me Spectrometry R. G. Michel, J. M. Ottaway and J. Sneddon and G. S. Fell Department of Pure and Applied Chemistry, University of Strathclyde, Cathedral Street, Glasgow, G1 1 X L Department of Biochemistry, Royal Infirmary, Glasgow, G4 OSF Keywords : Atomic fEuovescence ; flame atomisation ; electrodeless discharge lamps ; selenium A previously reported methodl for the preparation and operation of microwave excited electrodeless discharge lamps (EDLs) included a definition of the factors that had an effect on the radiant output and stability of EDLs. A cadmium EDL was taken as an example and the ten factors optimised by using the Simplex algorithm.With this method it was demonstrated that EDLs could be prepared with no failures when using the optimised conditions. In an argon-separated air - acetylene flame cadmium atomic fluorescence detection limits were as good as the best reported in the past. Further, the detection limits obtainable for successively prepared EDLs were reproducible with a relative standard deviation of 25%. Later work2 has shown that the method can be applied while employing either an A antenna or a $-wave Broida cavity during preparation and operation. This paper describes the use of the method for the preparation and operation of selenium EDLs. The performance of the selenium EDL, defined by flame atomic fluorescence detection limits, ~7as optimised with respect to the same ten factors that were applied to the cadmium EDL.The Simplex algorithm was again used and the &wave cavity employed for both preparation and operation. The results have indicated that the method can equally well be used for selenium EDLs. Detection limits for the determination of selenium by flame atomic fluorescence were comparable with or better than literature values, depending upon the choice of flame. The reproducibility was similar to that found for cadmium EDLs. Experimental Past researchers have prepared selenium EDLs by adding the element or the iodide' to the EDL blank. However, the method described in reference 1 involves the addition of an aqueous solution of the element to the EDL during the preparative stage.688 SHORT PAPERS Analyst, Vol.104 Selenium iodide is not appreciably soluble ; therefore its use could not be considered without jeopardising the reproducibility of the method because the alternative of adding small amounts of solid is difficult to carry out reproducibly. Initial experiments with selenium(1V) oxide, which is soluble, gave measurable limits of detection. It was therefore decided to carry out the full, ten-factor optimisation on solutions of this compound. Any oxygen produced by the breakdown of the selenium(1V) oxide during the discharge stage of the preparation was expected to be removed by the vacuum system. This was similar to the manner in which iodine was removed in order to optimise the ratio of cadmium to cadmium iodide as discussed in reference 1. The above approach appears to have been successful in the instances of both cadmium and selenium, as demonstrated by the excellent detection limits obtained.The radiant outputs of the selenium EDLs prepared during the Simplex optimisation were monitored by determining the magnitude of selenium atomic fluorescence signals at 204.0 nm in a nitrogen-separated air - acetylene flame. Details of the instrumentation and operating conditions, which were identical with those used for cadmium, are given in references 2, 8 and 9. Solutions of selenium(1V) oxide in de-ionised water acidified with hydrochloric acid to 0 . 0 4 ~ were aspirated into the flame for all fluorescence measurements. The Simplex search was designed and carried out as described in references 1 and 2. Results and Discussion The ten-factor Simplex search generated 84 vertices and involved the construction of 140 EDLs.Vertex 70 gave the best detection limit (0.11 pg ml-l for a count time of 1 s on a photon counter and a signal to noise ratio of 2). The Simplex was terminated when there appeared to be no further improvement in the detection limit. The reproducibility of the preparation and operation of the selenium EDL was determined by preparing a set of eighteen EDLs under conditions identical with the co-ordinates of vertex 70. These 18 EDLs were produced in three batches of six, with a between-batch interval of several months. For these eighteen EDLs (Table I) a mean limit of detection of 0.17 pg ml-1 was obtained with a relative standard deviation of 26%. A confidence interval of 0.1-0.2 pg ml-1 resulted from using the t distribution and a 99% confidence level.TABLE I REPRODUCIBILITY OF PREPARATION OF SELENIUM EDLS Fluorescence measured in a nitrogen-separated air - acetylene flame. EDLs prepared using factor levels resulting from vertex 70. as in references 1 and 2 (l-s count time, signal to noise ratio of 2). Eighteen identical Detection limits measured A t distribution and a 99% confidence level gave a confidence interval of 0.1-0.2 pg ml-1. Detection limit/ EDL pg ml-l 1 0.19 2 0.17 3 0.23 4 0.08 6 0.13 6 0.11 Detection limit/ EDL* pg ml-l 7 0.18 8 0.22 9 0.10 10 0.13 11 0.21 12 0.15 Mean detection limit = 0.17 pg ml-l Standard deviation = 0.043 Relative standard deviation = 26% * EDLs 7-12 were prepared 1 month after EDLs 1-6.t EDLs 13-18 were prepared 3 months after EDLs 1-6. Detection limit/ EDLt pg ml-1 13 0.22 14 0.17 15 0.18 16 0.13 17 0.18 18 0.20 More information about the response surface in the region of the optimum EDL was obtained by carrying out a univariate search, in which each of the factors was varied in turn while the remaining ones were kept constant at their Simplex optimum level (the co-ordinates of vertex 70). From the univariate search it was possible to infer the range over which each factor level could be varied while maintaining the detection limit within the confidenceJuly, 1979 SHORT PAPERS 689 interval. The results indicated the degree of control, over each of the ten variables, which is necessary in order to maintain the detection limit at the optimum value.The selenium control ranges are shown in Table I1 and were obtained in an exactly similar manner to the same experiments carried out for the cadmium EDLs discussed in reference 2. The cadmium control ranges are reproduced in Table I1 for comparison purposes. TABLE I1 CONTROLRANGEFOREACHFACTOR Factor* W, Mass of selenium or cadmium . . .. .. .. t, Time under vacuum after water removal . . .. A , Argon pressure under microwave excitation during preparation . . . . .. . . . . .. P, Microwave power for discharge during preparation . . t2 Time interval for microwave discharge during prepara- tion . . .. . . .. . . . . .. t , Time for EDL to cool before evacuation . . . . t, Time under vacuum after cooling period . . . . A , Final argon pressure . . .. .. .... P2 Operating microwave power . . . . . . . . T Operating temperature . . . . . . .. . . * These factors are fully defined in reference 1. t From references 1 and 2. Units PLg S Torr w S S S Torr w "C Selenium control Cadmium control range ranget 220-260 680-750 500-1 000 0-1 000 4.0-6.0 1.0-3.0 110-140 80-90 6.2-9.5 9.5-12.5 10-270 240 10-300 500 5.5-8 .O 6-9 31-45 45-55 1 40- 1 60 210-230 I t can be seen from Table I1 that the same factors for selenium and cadmium have mostly different control ranges and that in some instances, notably W,, A,, P,, t, and T , there are large differences. The difference in volatilities of selenium(1V) oxide and cadmium iodide (Table 111) probably explains the different control ranges required for these two elements. For example, it is plausible that a highly volatile material will require less time to volatilise during the discharge stage (tz) and a lower operating temperature (T) in order to maintain the correct vapour pressure of material in the lamp.TABLE I11 MELTING- AND BOILING-POINTS~~ OF SELENIUM, SELENIUM IODIDE, SELENIUM(IV) OXIDE, CADMIUM AND CADMIUM IODIDE Element or compound Melting-point/"C Boiling-point/"C Se .. .. .. 217.4-220 SeI, (%,I,) . . . . 68-70 100 (decomposes) Cd .. . . . . 320.9 767 f 2 CdI, . . . . .. 387 796 684.8 SeO, .. . . . . 340-350 (sublimes) - In Table IV detection limits in four flames are given. These results were obtained under optimised conditions for our instrument ation. Previously reported atomic fluorescence detection limits for selenium have been obtained by using the argon - hydrogen - entrained air flame,l1*l2 although selenium fluorescence in an air - propane flame has been reported.13 Detection limits in the literature are 0.2 pg ml-1,11 0.4 pg ml-1 1 2 and 0.15 pg mlk1.13 Hell and Ricchiol* reported a result of 0.04 pg ml-1, although details of their experimental con- ditions are not readily available.The best argon - hydrogen detection limit reported here, 0.02 pg ml-l, is marginally better than that of Hell and Ricchio but significantly better (by one order of magnitude) than all other literature values. This improvement can in part be attributed to a reduction in noise from stray light by a factor of 5.6, which resulted from the use of a double monochromator. Despite the favourable detection limits in hydrogen flames the present authors prefer the use of the nitrogen-separated air - acetylene flame because it has a greater freedom from interferences. For example, chemical interferences are always a problem in hydrogen flames.Also, the (This reduction is discussed in detail el~ewhere.~)690 SHORT PAPERS Analyst, Vol. 104 scatter of source radiation by matrix particles in the flame is relatively more troublesome in hydrogen flamess and will lead to degraded detection limits in most instances in which real samples are analysed for selenium. TABLE IV DETECTION LIMITS FOR THE DETERMINATION OF SELENIUM IN FOUR FLAMES Detection limits are given for a signal to noise ratio of 2. The noise was taken as the square root of the total background (flame background plus scatter background) measured on a photon counter.O Flame conditions were adjusted to give maximum signal to noise ratio.Spectral band pass was 1 nm a t the analytical wavelength of 204.0 nm.* EDL number 14 (Table I) was used to obtain these results. Flame Detection limitlpg ml-1 I A 3 Count time of 1 s Count time of 10 s Nitrogen-separated air - acetylene . I 0.09lt 0.027 Air - hydrogen . . . . .. . " 0.29 0.093 Nitrogen-separated air - hydrogen . I 0.08 0.025 Argon - hydrogen - entrained air . . .. 0.066 0.022 * The 204.0-nm line gave signals four times greater than signals a t the 196.1-nm Nitrogen flushing of the monochromator had no effect on the signals a t either t This detection limit is slightly better than that given in Table I for EDL number All detection limits in this table line.line. 14 as a result of fine tuning of the optical system. were obtained under the same conditions and are therefore comparable. Other analytical characteristics of the optimised EDLs were similar to those reported by previous workers. The sensitivity dropped by a factor of 2 4 after approximately 100 h in an analogous manner to the cadmium EDLs.lp2 Dagnall et aL4 reported no failure of their selenium EDL after 50 h of operation. The linear dynamic range extended from the detec- tion limit to a selenium concentration of 70pgml-l, which is comparable with the upper limit of 100 pg ml-l reported previ~usly.~ Conclusion The ten factors involved in the preparation and operation of EDLs were defined carefully in reference 1 and as a result it has been possible to optimise the radiant output of cadmium and selenium EDLs rigorously for use in atomic-fluorescence spectrometry.It is clear, from the control ranges for these two elements, that the levels of each of the ten factors are both critical and quite different. This is probably a result of the differing volatilities of the materials added to the EDL during preparation. It may be that, as the method is extended to other elements, control ranges will only prove to be similar or identical in the unlikely instances where two EDLs of different elements contain materials of similar volatility. The detection limits reported here for the atomic fluorescence of selenium are superior to all flame atomic fluorescence literature values. However, after allowing for instrumental factors (the use of the double monochromator and a photon c o ~ n t e r ~ ~ ~ ) , these selenium EDLs give comparable or slightly better detection limits than others in the atomic-fluorescence literature.This result was expected because the method used is well controlled, but in other respects similar to other methods. The reproducibility of preparation of the selenium EDLs is the same as that obtained by us for cadmium E D L S , ~ ~ ~ comparable with the prepara- tive method using hydride generation described by Bently and Parsonsll and superior to all other methods as described in the literature. The authors acknowledge the support of the Scottish Home and Health Department for the purchase of the equipment used in this project and for the award of a Postdoctoral Fellowship (to support R.G. M.). They also thank the Eastern District, Greater Glasgow Health Board for maintenance grant in support of J. S. References 1. 2. Michel, R. G., Coleman, J., and Winefordner, J. D., Spectrockim. Actu, 1978, 33B, 195. Michel, R. G., Ottaway, J. M., Sneddon, J., and Fell, G. S., Analyst, 1978, 103, 1204.July, 1979 SHORT PAPERS 691 3. 4. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. Mansfield, J. M., Bratzel, M. P., Norgordon, H. O., Knapp, D. O., Zacha, K. E., and Winefordner, Dagnall, R. M., Thompson, K. C., and West, T. S., Talanta, 1967, 14, 551. Dagnall, R. M., Thompson, K. C., and West, T. S., Atom. Absorption Newsl., 1967, 6 , 117. Cooke, D. O., Dagnall, R. M., and West, T. S., Analytica Chim. Acta, 1971, 56, 17. Marshall, G. B., and West, T. S., Analytica Chim. Ada, 1970, 51, 179. Michel, R. G., Hall, M. L., Ottaway, J. M., and Fell, G. S., Analyst, 1979, 104, 491. Michel, R. G., Hall, M. L., Rowland, S. A. K., Sneddon, J., Ottaway, J. M., and Fell, G. S., Analyst, “Handbook of Chemistry and Physics,” 46th Edition, Chemical Rubber Co., Cleveland, Ohio, 1965. Bently, G. E., and Parsons, M. L., Analyt. Chem., 1977, 49, 551. Zacha, K. E., Bratzel, M. P., Winefordner, J. D., and Mansfield, J. M., Analyt. Chem., 1968, 40, Dagnall, R. M., Thompson, K. C., and West, T. S., Talanta, 1967, 14, 557. Hell, A., and Ricchio, S., Presented a t 21st Pittsburgh Conference on Analytical Chemistry and Received December 7tk, 1978 Accepted February 2nd, 1979 J. D., Spectrochim. Ada, 1968, 23B, 389. 1979, 104, 505. 1733. Applied Spectroscopy, Cleveland, Ohio, March 1970.

ISSN:0003-2654    

DOI:10.1039/AN9790400687

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

July, 1979 SHORT PAPERS 691 lodimetric Determination of Milligram Amounts of Certain Aliphatic Acids S. N. Nema and R. M. Verma Department of Post-Graduate Studies and Research in Chemistry, University of Jabalpur, Jabalpur, M . P., India Keywords : Aliphatic acids detewnination ; solid iodate and iodide reagent; iodinzetric titration Acids can be determined iodimetrically by utilising the reaction 10, + 51- + 6H+ --f 31, + 3H20 The iodide present is oxidised, in an amount equivalent to the amount of acid present, to iodine, which is titrated with thiosulphate; however, the equilibrium is pH dependent. The iodimetric method gives excellent results with strong acids, but weak acids cannot be titrated directly by this method because the pH at the end-point is too high to permit complete reaction.Moderately strong acids, such as oxalic, tartaric and acetic acid, also react incompletely, even after 24 h.1 In order to use the reaction for the iodimetric deter- mination of weak acids, Kolthoff2 suggested a modification that consisted in adding a known excess of thiosulphate to a 0.1 N acid solution containing iodide and iodate solution followed by back-titration of the thiosulphate after about 15-30min. The reaction period may be longer with a more dilute solution of the acid. This modified procedure has been adopted by several workers for determining the acid content in various products, but the drawback of these procedureW is the long reaction period required (2-24 h); further, during this period there is a possibility of reaction between the thiosulphate and the acid.In order to evolve a rapid method for the determination of milligram amounts of certain aliphatic acids, the influence of various factors on their reaction with iodide and iodate was studied. As the reaction mixture could not be heated on account of the volatility of iodine, the only means left to speed up the reaction was to increase the concentration of iodide and iodate. If iodide and iodate are added in the form of their aqueous solutions, as was done by previous workers, the reaction does not proceed to completion even when large volumes of iodide and iodate solutions are added and sufficient time is allowed. This is caused by the increase in the over-all volume of the reaction mixture, as there is a sharp fall in the reaction rate with increase in the volume of the reaction mixture.692 SHORT PAPERS Analyst, VoZ.104 In order to keep the volume of the reaction mixture at a minimum, suitable amounts of iodide and iodate were added in the solid form, when the reaction was found to be quantitative in about 3 min at room temperature (30 "C). Increasing the amount of iodate was found to be more effective than increasing the amount of iodide in enhancing the reaction rate. This finding is in agreement with the expression for the rate equation5 3 9 9 Rate = k[12]z[H+]E[I0,-]a Experimental Reagents All solutions were prepared using analytical-reagent grade chemicals. Sodium thiosuZfJhate. Standardised with potassium iodate solution. This was diluted to give 0.05, 0.01 and 0.00625 N solutions.Starch solution, 1% m/V. Solutions of organic acids (0.05 N) were prepared and standardised with standard alkali solution; 0.025 and 0.00625 N solutions were then prepared by dilution of the 0.05 N solution with conductivity water. Stoicheiometry of the Reaction Before using the reaction for the iodimetric determination of a number of different organic acids, the stoicheiometry of their reactions with potassium iodide and iodate was examined. The effect of variation in the amount of solid potassium iodate and iodide added and in the reaction period, on the reaction rate, was investigated for each of the organic acids. The minimum amounts of solid potassium iodate and iodide required for the completion of the reaction in the given period of time (3 min) at room temperature (30 "C) were found to be different for the different acids.However, as their addition, even in large excess, did not affect the results, a general procedure applicable to all of the acids tested has been proposed. It was also found that although the reaction was complete in about 3 min, the reaction mixture could be kept for 15-30 min with no adverse effect on the results. Procedure Pipette 5-10 ml of the acid solution containing 0.03-0.12 mequiv of carboxyl group into a 100-ml iodine flask. Introduce about 2 g of solid potassium iodate and stir the solution for about 3 min using a magnetic stirrer. Add 3 g of solid potassium iodide and wait for 2-3 min, swirling the flask frequently. Titrate the liberated iodine with 0.01 N sodium thiosulphate solution from a microburette using starch as indicator.1 ml of 0.01 N thiosulphate solution = 0.45 mg of carboxyl group Results and Discussion The proposed iodimetric procedure was applied to eight different organic acids. The analysis was carried out a t three concentration levels, the amount of the acid taken being between 0.03 and 0.12 mequiv. The results presented (see Table I) are the averages of six determinations. The recovery studies showed that the maximum error is 0.3% and the standard deviation is 0.1-0.2%. In acid - base titrations, the inflection of the neutralisation graph progressively shortens as the concentration of the acidic and alkali solutions is reduced. If a visual end-point is detected by means of an indicator, the inflection range should be greater than 2 pH units.6 Hence, there is a limit to the dilution where a colour change of the indicator can be observed.When 0.01 N alkali solution is used for titrating a dilute solution of a weak acid, it is not easy to observe the colour change of phenolphthalein as the pink colour fades rapidly.' There is not much difficulty in detecting an iodimetric end-point when a 0.01 N or an even more dilute solution of thiosulphate is used as the titrant. Some natural products, in which the acid content is to be determined, are coloured; this colour interferes with the perception of the colour change at the end-point during a neutralisa- tion titration. Iodimetric determination of acidity in such instances provides a betterJuly, 1979 SHORT PAPERS TABLE I DETERMINATION OF CERTAIN ALIPHATIC ACIDS Amount of samplelmg Acid Isobutyric ..Hexanoic .. Adipic . . Acrylic . . Fumaric . . Lactic .. Laevulinic . . Cyanoacetic . . .. .. .. .. .. .. .. .. * . .. .. .. .. .. .. .. r Taken 8.880 5.505 2.752 9.524 8.478 3.623 8.220 4.559 2.338 9.007 5.512 2.485 5.803 2.408 1.915 7.206 4.684 2.702 9.293 7.260 3.717 8.506 5.528 2.764 1 Recovered* 8.859 5.492 2.745 9.519 8.474 3.616 8.218 4.556 2.333 8.999 5.501 2.478 5.797 2.403 1.909 7.200 4.675 2.694 9.281 7.255 3.708 8.497 5.524 2.756 Deviation, % - 0.23 - 0.23 - 0.25 - 0.05 - 0.04 -0.19 - 0.02 - 0.06 -0.21 - 0.08 -0.19 -0.28 -0.10 - 0.20 -0.31 - 0.08 -0.19 - 0.29 - 0.13 - 0.07 -0.24 -0.10 - 0.07 - 0.29 693 Standard deviation, % 0.10 0.19 0.16 0.10 0.11 0.16 0.14 0.08 0.19 0.14 0.16 0.12 0.10 0.11 0.18 0.10 0.11 0.18 0.10 0.15 0.16 0.10 0.10 0.12 * Average of six determinations.alternative. Further, the iodimetric method is to be preferred for determining the acid content of substances that contain a component capable of reacting with alkali. For example, in determining carboxylic acid in lac, the alkalimetric method is unsatisfactory as the lac resin is readily saponified by alkali8 The proposed method is simple, rapid and accurate and being iodimetric gives sharp end- points; hence, it is more suitable for the quantitative analysis of dilute solutions of the organic acids examined. The authors thank the University of Jabalpur for the award of a Fellowship to one of them (S. N. N.). References 1. 2. 3. 4. 5. 6. 7. 8. Kolthoff, I. M., Belcher, R., Stenger, V. A., and Matsuyama, G., “Volumetric Analysis,” Volume 3, Kolthoff, I. M., Chem. Weekbl., 1926, 23. 260. Nabar, G. M., and Padmanabhan, C. V., Proc. Indian Acad. Sci., 1950, 31A, 371. Dassler, H. G., Pharm. Zentralhalle Dtl., 1962, 101, 409. Pickering, W. F., “Modern Analytical Chemistry,” Marcel Dekker, New York, 1971, p. 413. Smith, T. B., “Analytical Processes,” Second Edition, Edward Arnold, London, 1952, p. 228. Cheronis, N. D., and Ma, T. S., “Organic Functional Group Analysis,” Wiley, New York, 1964, Kamath, N. R., and Mainker, V. B., Analyt. Chem., 1950, 22, 724. Interscience, New York, 1957, p. 276. p. 485. Received June 27th. 1978 Accepted Decembey 14th, 1978

ISSN:0003-2654    

DOI:10.1039/AN9790400691

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

694 SHORT PAPERS Analyst, Vol. 104 A Variation of the Monier-Williams Distillation Technique for the Determination of Sulphur Dioxide in Ginger Ale Bronislaw L. Wedzicha and Michael K. Johnson Procter Department of Food Science, University of Leeds, Leeds, LS2 9 JT Keywords Sulphur dioxide determination ; ginger ale analysis The Monier-Williams distillation technique1 for the determination of sulphur dioxide in a wide range of foods has so far been unsurpassed in its simplicity for the reliable determination of the additive. Disadvantages include lengthy distillation times, the need for large samples to provide titratable distillates at low sulphur dioxide levels and the risk of interference from volatile components. Here we describe a variation of the technique in which the distillate is analysed for sulphur dioxide by a spectrophotometric method. The procedure is an adaptation of the method of Humphrey et al.2 for the determination of sulphite ion in aqueous solution based on the reaction between 5,5’-dithiobis(%nitro- benzoic acid) (Ellman’s reagent, DTNB) and sulphite ion to form, as one of the products, the highly coloured 5-mercapto-2-nitrobenzoic acid (E = 15500 mol-l dm3 cm-l at 412 nm) when present as the thiol anion. The reaction is reported to be quantitative in the pH range 6-9.All concentrations of sulphite species will be expressed as concentrations of sulphur dioxide, as the analysis involves conversion into the gaseous product. Experimental Distillation Procedure A semi-micro Kjeldahl assembly (Quickfit 21/100MC) with a capacity of 100 cm3 was used as the distillation apparatus with oxygen-free nitrogen as the carrier gas.The outlet of the condenser was placed below the surface of a solution of DTNB reagent (50 cm3, 2.5 x mol dm-3) in phosphate buffer (1.6 x 10-2 mol dmW3 Na,HPO, + 1.8 x 10-3 mol dm-3 KH,P04, pH 8.0) containing 10% V/V of ethanol to aid dissolution of the reagent. For analysis, sulphuric or hydrochloric acid (50 cm3, 2 mol dm-3) was placed in the Kjeldahl flask, de-aerated by heating to boiling in an atmosphere of nitrogen and the test solution (total sulphur dioxide content 1-8 x 10-6mol) added to the boiling acid by way of the air lock. Analysis of Distillate When the distillation was complete the solution in the receiver was diluted to 100 cm3 with buffer (pH 8.0) and the absorbance measured in l-cm cells using a Corning-EEL long- cell absorptiometer with a 601 filter in the light path.Solutions containing hydrogen sulphite ions were prepared from sodium metabisulphite (BDH, laboratory-reagent grade) and contained EDTA (1.0 x 10-4mol dmw3) to reduce oxidation. Solutions were standardised iodimetrically and standards for the spectrophotometric determination of sulphite were prepared by direct addition to buffered DTNB reagent. A reagent blank was used in the reference compartment in all instances. When necessary, assays were checked using a recognised distillation pr~cedure.~ Stability of DTNB Reagent To determine the stability of DTNB reagent (1.0 x low2 mol dm-7 as a function of pH in the pH range 7.0-9.2, the reagent was dissolved in appropriate buffer solutions to which 10% V/V of ethanol had been added to aid dissolution.The absorbance of the reagent was measured as a function of time, as before.July, I9 79 SHORT PAPERS 695 Preparation of Sulphited Ginger Ale Solutions Ginger ale solutions containing known amounts of sulphur dioxide were prepared by dilution of an unsulphited concentrate and included appropriate amounts of standard hydrogen sulphite ion solution. mol dm-3) was added to reduce oxida- tion and the mixtures were not carbonated. EDTA (1.0 x Results and Discussion Preliminary Investigations Stability tests on the DTNB reagent showed that its absorbance at 412nm remained negligible over several days at pH 7-8, but at higher pH there was evidence of attack by OH- at the S-S linkage, leading to the formation of highly absorbing thiol.It was con- sidered that a high pH was consistent with efficient trapping of sulphur dioxide from the gas phase and therefore the pH value chosen for the solution in the receiver was 8.0. Under these conditions and with [DTNB] = 10-3 mol dm-3 and [SO,2-] = 10-5 mol dm-3 the reaction between the DTNB and sulphite is quantitative and instantaneous. For the analysis of standard solutions, when the total amount of sulphur dioxide analysed was in the range 2-8 x mol with a distillation time of 240 s and carrier gas flow-rate of 0.2 cm3 s-l, a 92 & 3% analysis for sulphur dioxide was obtained with hydrochloric acid in the distillation flask and 97 &- 3% when sulphuric acid was used.Under the same conditions and with sulphuric acid in the distillation flask, 96% of the recovered sulphur dioxide appeared in the first 90 s, increasing to 98% after 240 s and reaching a steady value ( l O O ~ o ) after 600 s. A distillation time of 240 s with sulphuric acid was chosen for subse- quent experiments. Under the conditions used for these analyses, the production of the chromophore through an interchange reaction with 2-mercaptoethanol (as an example of a volatile thiol) was quantitative when total amounts of up to lop5 rnol were analysed. Organic thiols and any compounds capable of producing -SH species in aqueous solution at pH 8.0 are therefore likely to interfere with the same sensitivity as the analysis for sulphur dioxide. Experiments on Sulphited Ginger Ale The results for analyses on ginger-ale samples prepared with controlled addition of HS03- are summarised in Table I.The last two entries relate to analyses of sulphited concentrate by the spectrophotometric and also standard Monier-Williams techniques. The unsulphited beverage was found to give a negligible absorbance when the distillate was analysed. The regression line drawn for a graph of recovered sulphur dioxide values against actual amounts TABLE I DETERMINATION OF SULPHUR DIOXIDE IN GINGER ALE Volume of sample used*/cm3 10.0 10.0 10.0 10.0 5.0 5.0 10.0 5.0 5.0 5.0 5.0 5.0 5.0 1.0 1.0 [SOJ added/ mol dm-3 x 1.23 1.23 2.47 2.47 4.94 4.94 4.94 7.41 7.41 7.41 9.88 9.88 9.88 52.07 52.0t [SO,] found*/ mol dm-3 x 1.16 1.16 2.34 2.34 4.70 4.59 4.47 7.20 6.80 7.00 9.59 9.59 9.09 52.5 51.5 Recovery, 93 93 95 95 95 93 91 97 92 95 97 97 92 101 99 % * Proposed variation of the Monier-Williams technique.t Amounts found to be present by the standard Monier-Williams technique using 50-cms samples.696 SHORT PAPERS present has a slope of 0.96 with an intercept of -0.38. For an intercept at the origin the line computed by the method of least squares has a slope of 0.95. An efficiency of 95% recovery was therefore obtained, which could be increased to nearly 98% if the distillation time was increased to 600 s. Unlike the conventional Monier-Williams procedure, this rapid technique gives little time for hydroxysulphonate-type adducts, which are stable under acidic condition^,^ to decompose. Free HSO, in solution is converted rapidly into gaseous sulphur dioxide when the sample is introduced but alkaline pre-treatment of the sample, under which conditions hydroxysulphonates are quantitatively decomposed, did not increase significantly the yield for ginger-ale samples.Re-use of the acid for distillations was found not to affect the results adversely after four analyses, although accumulation of products from non-enzymic browning was evident. Conclusion The method described offers improvements in sensitivity and speed over the conventional Monier-Williams technique. For a 240-s distillation a recovery of 95% is expected when a total of 2-8 x 10-6 mol of sulphur dioxide are analysed. The experimental error showed a spread of 5 4 % in this result at the 95% confidence limit assuming a normal distribution. This is to be compared with about lo-* mol required for reliable titrimetric Monier-Williams analysis and distillation times of the order of 3000s. No interfering compounds were present in the beverage analysed , although organic t hiols and inorganic sulphide are expected to interfere. We thank Schweppes Ltd. , London, for suggesting this project, providing experimental samples and helpful discussions. References 1. 2. 3. 4. Monier-Williams, G. W., Analyst, 1927, 52, 415. Humphrey, K. E., Ward, M. H., and Hinze, W. L., Analyt. Chem., 1970, 42, 698. “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edition, Association of Official Analytical Chemists, Washington, D.C., 1970, p. 466. Vas, J., J . SOC. Chem. Ind., 1949, 68, 340. Received November ZOth, 1978 Accepted February 12th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9790400694

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

696 SHORT PAPERS Communications Material for publication as a Communication must be on an urgent matter and be of obvious scientijk importance. Rapidity of publication i s enhanced if diagrams are omitted, but tables and formulae can be included. Conamunications should not be simple claims for priority: this facility for rapid publication i s intended for brief descriptions of work that has progressed to a stage at which it i s likely to be valuable to workers faced with similar problems. A fuller paper ynny be offered subsequently, i f justilfied by later work. Manuscvipts are not subjected to the usual examination by refevees and inclusion of a Communication i s at the Editor’s discyetion. Sources of Error in the Flame Photometric Determination of Sodium and Potassium in Portland Cements Keywords : Sources of ervor ; pame photometry ; sodium determination ; potassium determination ; cement analysis The determination of sodium and potassium in cement has been carried out almost exclusively by flame emission spectrometry for more than two decades.The high calcium content of cements, about 45%, emits a substantial amount of radiation from molecular band systems in the relatively cool flames of most of the instruments commonly used, and a significant proportion of thisJuly, 1979 COMMUNICATIONS 697 radiation passes through optical filters having their maximum transmission at, or close to, the characteristic emission wavelength of the sodium doublet at 589nm. One of the methods of overcoming this problem is to add an amount of calcium to the calibration solutions to match the calcium content of the cement being ana1ysed.l An alternative is to add aluminium to the cement solutions to suppress the calcium molecular band interference, an equal amount being added to the calibration solutions.This second method is specified in a British Standard2 and is widely used within the UK and elsewhere. Some recent work by one of us (P. G. D.) in which the two methods were compared indicated that previously unsuspected erroneous results were being obtained for both sodium and potassium when using the standardised method of aluminium addition. For sodium it was observed that not all of the calcium radiation from the sample solutions was being suppressed by the aluminium, leading to positive errors equivalent to about 0.02% of sodium oxide in cement.For potassium it was found that the added aluminium depressed the emission when calcium was absent as in the calibration solutions, but not when calcium was present as in acidic solutions of cement. These phenomena gave falsely high results for potassium, the magnitude of the error being approxi- mately 6% of the content. It was found experimentally that by adding sufficient calcium to the calibration solutions containing aluminium both of the aforementioned errors could be avoided. A Cement Manufacturers Federation working party, which included representatives from the leading UK cement manufacturers, was set up for the purpose of verifying the above findings, particularly with respect to the effects of using different types of flame photometers and different combustion gases.The working party’s results confirmed the existence of defects in the British Standard method, and that the addition of calcium to the calibration solutions would usually overcome them. However, a constant bias towards slightly lower results for sodium was observed when one particular instrument, the Evans Electroselenium Model 100, was used. Further experimental work by the group showed that this was due to suppression of calcium radiation by the silicic acid in the cement solutions, indicating that insufficient aluminium was present to complex the calcium fully. Although the proportion of calcium complexed by the silica, and the consequent reduction in calcium emission, was likely to be the same whichever instrument had been used, the effect was more pronounced with the EEL 100 instrument because its gelatin filters were shown to be much less efficient at blocking calcium molecular band emission than the interference filters used in more modern instruments. The effect was overcome by increasing the concentration of added aluminium from 850 to 2000 mg 1-l.The addition of both calcium and aluminium to the calibration solutions has three advantages over the ASTM methodl in which only calcium is added. Firstly, this modified method can be used with simple flame photometers that do not have sufficient “backing off” capability to cope with the unblocked calcium radiation from the calibration solutions of the ASTM method. Secondly, by adding aluminium and calcium to solutions of materials that have relatively low but unknown levels of calcium, e.g., clays, rocks and glasses, the alkali elements can be determined using the same calibration solutions as for cements.Thirdly, the ASTM alternative procedure in which silica is removed before determining the alkalis is not necessary even when simple flame photometers are used. I t is concluded that the method specified in BS 4550 for determining sodium oxide and potassium oxide should be modified by increasing the aluminium addition to both calibration and cement solutions from 850 to 2000 mg 1-1 and by adding 630 mg 1-1 of calcium oxide to the calibration solutions only. References 1 . 2 . “Chemical Analysis of Hydraulic Cement,” ASTM Standards 1977, Part 13, C114, American Society “Methods of Testing Cement,” BS 4550: Part 2: 1970. for Testing and Materials, Philadelphia, Pa. Received April 25th, 1979 Blue Circle Technical, Research Division, Greenhithe, Kent, DA 9 J O P. G. Deane Cement and Concrete Association, Research and Development Division, Wexham Springs, Slough, Buckinghamshire, SL3 6PL T. P. Lees

ISSN:0003-2654    

DOI:10.1039/AN9790400696

出版商:RSC    

年代:1979

数据来源: RSC

摘要:

698 COMMUNICATIONS Analyst, Vol. 104 Gas Analysis Using an Internal Standard in Adsorption Tubes Keywords : Gas analysis ; adsorption tube ; internal standard ; anisole ; gas c hrornatograp hy Air sampling for organic contaminants is commonly carried out by drawing the air through a sample tube containing a solid adsorbent. When carbon or silica is used, contaminants can be extracted by mixing the adsorbent with a small volume of solvent containing a suitable standard. The contaminant and standard equilibrate between the surface phase and the solvent, and can be analysed on a gas chromatograph by injecting a portion of the solvent. This procedure is adopted in many of the standard NIOSH meth0ds.l However, as only a portion of the solvent can be injected, typically 5 ~1 of a 2-ml volume, the sensitivity of the analysis is correspondingly reduced ; also, solvent extraction is an unsuitable technique when the contaminant might be obscured by the solvent peak.Hence, when analysing for low concentrations, especially of volatiles or unknowns, i t is preferable to adsorb the contami- nants on a medium that permits their thermal desorption in total directly on to the gas-chromato- graphic column. A full description of the method has been given by Brookes et aZ.2 A drawback to this approach has been the absence of an internal standard in the adsorbed sample. This communication describes how the problem has been overcome by injecting 1 p1 of a solution of a standard on to the Tenax GC sample tube prior to sampling and purging off the solvent with a flow of clean air.The solvent is chosen so that it is readily purged and free of contamination. The standard must be free of contamination, strongly retained by Tenax GC, totally released during thermal desorption, unlikely to be present in the sampled air and, for analysis by gas chromatography - mass spectrometry, should have a readily identifiable mass spectrum. Tenax GC is commonly used as the adsorbent in this technique. Anisole in methanol meets these requirements. Experimental Standard Solution Ltd.) a t a concentration of 2 g 1-l. The standard solution consists of anisole (Griffin and George Ltd.) in Distol methanol (Fisons Apparatus and Materials Adsorption tubes Tenax GC and plugged at each end with glass-wool. The tubes (3.25 x 0.25 in), made of stainless steel, were packed with 0.3 g of 60-80-mesh Desorption For this work the GN Concentrator (GN Instrumentation Consultancy Ltd., 75 Craven Gardens, London, S.W.19) was used, but desorption can be effected by simple home-made systerns.8~~ The GN Concentrator consisted of a desorption oven and liquid nitrogen cold trap connected to the inlet of the gas chromatograph, with a carrier gas flow-rate of 10 nil min-l and desorption for 13 min a t 250 "C into the cold trap. The sample was flashed out of the cold trap by rapid heating to 220 OC, and passed through a heated capillary tube to the gas chromatograph.Gas chromatography A Pye Unicam Series 104 gas-chromatographic oven was fitted with an SGE Ltd. inlet splitter set to a splitting ratio of 20: 1.The column was a 25 m x 0.25 mm i.d. Carbowax 20M capillary (Column Technology, Newcastle upon Tyne) temperature programmed from 0 to 200°C at 5 "C min-1. The initial sub-ambient temperatures were achieved with a liquid nitrogen cold finger inserted in the gas-chromatographic oven. The detector was a VG Micromass Ltd. MM16B mass spectrometer. Syringe use with the above standard solution. A Hamilton 75 5-p1 syringe with the plunger in the barrel was used. It was reserved forJuly, 1979 COMMUNICATIONS 699 Purging gas Air, purified by passage through a charcoal adsorption trap, was used as the purging gas. Procedure A l-pl volume of standard solution was injected into the sample tubes and 5 1 of purging gas were drawn through at 200 ml min-l to remove the methanol.The same direction of flow was maintained when sampling so that the anisole would be drawn further into the tube. The tubes were analysed according to the conditions set out above, anisole appearing between 95 and 105 “C in the temperature programme. Results and Discussion The anisole peak heights resulting from analyses on various tubes used over a range of sample volumes were as follows: Volume of air sampled/l . . .. .. .. 1 1 3 10 30 Peak height of anisole/mm . . .. .. 184 178 162 184 184 Deviation from the mean peak height, yo . . + 3 +0.2 -9 +3 f 3 The reproducibility of the anisole peak height is consistent with the errors resulting from day-to- day variations in the mass spectrometer response, the syringe error and the errors inherent into the operation of an inlet splitter coupled to a flash injection device.By relating pollutant peak height to an internal standard, all except the syringe error would be eliminated, and these results show that in this procedure anisole can function as such an internal standard for air samples of up to at least 30 1. Application to the Extraction of Volatiles from Aqueous Samples Volatile organic compounds are readily extracted and concentrated from aqueous samples using adsorption tubes in the gas stripping technique. 7.2 1 of purified nitrogen are bubbled at 2 ml s-1 through the sample, immersed in a boiling water-bath. The moisture is removed by passing the vapour through a water-cooled condenser and the volatile compounds are trapped on a Tenax GC adsorption tube. Preliminary work has shown that if anisole is introduced into the aqueous sample as an internal standard, for example, by injection of 1 p1 of the above standard solution into 250 ml of sample, at least 80% is transferred into the tube, and no significant amounts of the methanol solvent are retained.Both of these procedures are being increasingly used in the Regional Chemist’s Department of Strathclyde Region, and over the wide range of work that has been undertaken no interference of the standard with the sample has been observed. In the method of Brookes et The author is indebted to Mr. R. S. Nicolson, Regional Chemist for Strathclyde Region, for his support in the development of these techniques. References 1. 2. 3. 4. “NIOSH Manual of Analytical Methods,” Second Edition, Volumes 1-3, National Institute for Brookes, B. I., Jickells, S. M., and Nicolson, R. S., J. Ass. Publ. Analysts, 1978, 16, 101. Halliday, M. M., and Carter, K. B.. BY. J. Anaesth., 1978, 50, 1013. Brookes, B. I., Jickells, S. M., and Nicolson, R. S., J. Assoc. Publ. Analysts, 1978, 16, 117. Occupational Safety and Health, US Government Printing Office, Washington, D.C., 1977. Received May lst, 1979 Department of Regional Chemist, Strathclyde Regional Council, 8 Elliot Place, Clydeway, Glasgow, G3 8EJ B. I. Brookes

ISSN:0003-2654    

DOI:10.1039/AN9790400698

出版商:RSC    

年代:1979

数据来源: RSC