ISSN:0003-2654    

DOI:10.1039/AN98409FX005

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BX007

出版商:RSC    

年代:1984

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98409BP013

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST FEBRUARY 1984 VOL. 109 113 Determination of Silver Lead and Bismuth in Glasses by Atomic-absorption Spectrometry with Introduction of Solid Samples into Furnaces James B. Headridge* and Ian M. Riddington Department of Chemistry/ The University Sheffield S3 7HF UK Atomic-absorption spectrometry with an induction or resistively heated furnace has been used for the determination of 0.007-22 pg g-I of silver 1 .!%40 pg g-1 of lead and 0.02-1 5 pg g-1 of bismuth in 0.1-10-mg samples of standard forensic archaeological and ancient stained glass added directly to the furnace. Calibration graphs of peak area versus the mass of trace element have been constructed by using NBS standard glasses for silver and lead and standard solutions of bismuth nitrate. Information is presented on the accuracy and precision of the method for 9-29 glasses depending on the element under investigation.The limits of detection are silver 0.004 lead 0.014 and bismuth 0.02 pg g-1. The significance of the results is described with particular reference to forensic work and stained glasses. Keywords Silver lead and bismuth determinations; glass analysis; atomic-absorption spectrometry; furnace atomisation; solid samples The determination of trace elements in glasses is important when these glasses are used for optical fibres and is helpful in establishing the origins or types of archaeological and forensic glass samples. When forensic scientists wish to establish that suspect and control samples of glass are from the same batch, they resort to the measurement of physical properties such as density and refractive index and to chemical analysis for major and minor elements.However modern glasses are made by carefully controlling the composition of the major elements for batches of the same type of glass and if significant differences in the composition of glasses of the same type are to be revealed this is most likely to be among the minor elements and even more so among the trace elements. Catterick and Hickman’ used inductively coupled plasma emission spectrometry to determine minor and trace concentrations of aluminium barium iron magnesium and manganese in 0,2-0.S-mg samples of glass with limits of detection of 500,5,50,500 and 10 pg g - J respectively for 0.5 mg of glass. However glasses contain many other elements at trace or ultra-trace levels and if the list of elements that can be determined is to be extended significantly then a convenient method is required for concentrations down to 0.1 pg g-1.Lead silver and bismuth are trace elements that occur in glasses at concentrations often considerably less than S pg 6-1 and were selected in this study partly for that reason. It is of considerable importance to establish where archaeo-logical glasses and similar materials were made and chemical analysis for elements in such materials is useful in tracing their origin. This can be illustrated with Scottish faience beads, where the results of chemical analysis for magnesium, aluminium titanium barium and silver enabled archaeolo-gists to refute the theory that these Bronze Age artifacts originated from the Eastern Mediterranean.They were actually made in Scotland and showed that its people had a more advanced civilisation at that time than had previously been believed.? In matching glasses to kiln waste materials, the most convincing evidence for a perfect match is produced when minor and trace element concentrations are in good agreement in addition to a good match for concentrations of the major elements. Trace element concentrations in archaeo-logical glasses are frequently higher than those in modern glasses but for many elements they are still below 10 pg 8-1. Useful elements in archaeological and forensic work are those * Present address University of the South Pacific P.O. Box 1168, Suva Fiji. whose concentrations vary widely throughout a selection of glasses.The most widely used methods for the determination of trace elements in glass have been spark-source mass spectro-metry (SSMS) ,3,4 neutron activation analysis (NAA)”6 and furnace atomic-absorption spectrometry (FAAS).7J All of these have their disadvantages. Expensive instrumentation is required for SSMS and for very low concentrations of trace elements long exposure times are needed. Access to a nuclear reactor is usually required for NAA. In conventional FAAS a dissolution step and often a pre-concentration step are required. This results in long analysis times and the possibility of introducing errors either by contamination or loss of analyte is increased. AAS with the introduction of solid samples into furnaces has been used successfully for the determination of trace elements in metals.’ It was decided to investigate this technique for the analysis of solid chips of glass.Most of the work undertaken refers to silver lead and bismuth but preliminary investigations have also been made for thallium, antimony manganese iron zinc and cadmium. The only solid sampling work done previously with glass has involved mixing powdered glass with either graphitel(l.11 or graphite and potassium chloride. 12 Experimental Materials and Solutions Standard glasses. These were standard reference materials SRM 613,614 and 616 from the National Bureau of Standards, Washington DC USA. Glasses fur analysis. These were obtained from the British Glass Industry Research Association Sheffield.the Society of Glass Technology Sheffield the Metropolitan Police Foren-sic Science Laboratory London and the Department of Ceramics Glasses and Polymers University of Sheffield. The types and forms of the glasses are shown in Table 1. The ancient stained glasses from York Minster were small sheets, approximately 3 mm thick with red staining extending about half-way through the otherwise colourless material. These sheets were sawn parallel to their large faces to produce both red and colourless samples for analysis. All glass samples were cleaned by washing with ethanol and immersing in 1.5 M nitric acid for 30 min. They were then rinsed with distilled water and air-dried between filter-papers. The samples were then wrapped in thick polythene sheeting 114 ANALYST FEBRUARY 1984 VOL.109 Table 1. Types and forms of glasses analysed Code EC1.1* . . . . . . . . . . . . . . EC1.2* . . . . . . . . . . . . . . SGT2t . . . . . . . . . . . . . . SGT3I. . . . . . . . . . . . . . . SGT4t . . . . . . . . . . . . . . SGT5t . . . . . . . . . . . . . . SGT6t . . . . . . . . . . . . . . SGT7t . . . . . . . . . . . . . . SGT9-t . . . . . . . . . . . . . . MC 791 $ . . . . . . . . . . . . YMl,YM3,YM4 YM68 . . . . . . Hutton Rosedale Bagots Park and Kimmeridgelj . . . . . . . . . . A - F $ . . . . . . . . . . . . . . * From British Glass Industry Research Association. t From Society of Glass Technology. $ From Metropolitan Police Forensic Science Laboratory. Type of glass Na,O - MgO - CaO - SiO, Na,O - MgO - CaO - SiO, Na,O - Bz03 - A1,O3 - Si02 K,O - PbO - SiO, Fluoride-opal Na,O - MgO - CaO - SiO, NazO - CaO - SiO, Na20 - CaO - SiOz K,0 - PbO - SiO, Na,O - CaO - SiO, NazO - MgO - CaO - SiO, Ancient stained glass Na,O - K,O - CaO - SiO, Form Float glass Sheet glass Broken rod Broken tumblers White opal sheet Tableware Broken bottles Flat-sided bottles Flat disc Sheet glass Window fragments Ancient containers (ca.1600 A.D.) B YMl and YM3 are 12th century glass; YM4 and YMG'are 14th century glass; all from York Minster. 7 From Department of Ceramics. Glasses and Polymers University of Sheffield. Table 2. Instrumental settings for the induction furnace and the Perkin-Elmer 300s Core temperature/ Resonance Element under study "C lineinm Slit width/nm Lead .. . . . . . . 2 100-2 400 Silver . . . . . . . . 2 300-2 400 Iron . . . . . . . . 2 350 Thallium . . . . . . 2250 Antimony . . . . . . 2250 Manganese . . . . . . 2400 Zinc . . . . . . . . 1 870-2 300 Cadmium . . . . . . 2350 283.3 328.1 248.3 276.8 217.6 279.5 213.9 228.8 0.7 0.7 0.2 0.7 0.2 0.2 0.7 0.7 which has been soaked overnight in 8 M nitric acid and dried, and crushed with a plastic mallet so as to produce chips that weighed up to 10 mg. Such chips were added directly to the furnaces. However for wet-chemical analysis of the glasses, crushing was continued until no single piece weighed more than 0.5 mg. Hydrofluoric acid. Aristar grade 40% mlm BDH Chemi-cals.Nitric acid. Aristar grade sp.gr. 1.42 BDH Chemicals. Standard silver solutions. Dissolve 0.394 g of silver nitrate (Engelhard photographic quality) in 25 ml of distilled water. Add 25 ml of 1.6 M nitric acid and make up to 250 ml so as to give a solution 0.16 M in nitric acid and 1 mg ml-1 in silver. Prepare more dilute solutions when required by dilution with 0.16 M nitric acid. Standard bismuth solutions. Dissolve 0.250 g of bismuth powder (Koch-Light Laboratories 99.999 5 % ) in a minimum volume of 4 M nitric acid. Dilute the solution to 250 ml with 2 M nitric acid to give a solution 1 mg ml-1 in bismuth and approximately 2 M in nitric acid. Prepare more dilute solutions when required by dilution with 2 M nitric acid. Apparatus Induction furnace This was identical with that previously described,13 except that the graphite core and side-arms were made from AGTS (British Acheson Electrodes) or UF4S (Ultra Carbon) grade graphite and the window mounts were modified to accommo-date the flow of a window gas as described by Baker and Headridge.14 The graphite core side-arms and powder were baked under vacuum for 10 h at approximately 1 500 "C (35 h for lead).Absorbances were measured on a Perkin-Elmer 300 atomic-absorption spectrometer fitted with a JJ Instruments CR 552 recorder. Resistively heated furnace This was an Instrumentation Laboratory 555 furnace fitted with a square-section graphite cuvette to accommodate graphite microboats. It was fitted within a Varian AA-6 atomic-absorption spectrometer whose amplifier response had been speeded up according to the modification of Lundberg.15 Absorbance - time peaks were recorded on a JJ Instruments CR 552 recorder.The flow-rate of argon within the furnace was always maintained at 20 standard cubic feet per hour (ca. 9.5 1 min-1). Measurement Procedures for a Series of Solid Samples Induction furnace The furnace was operated in the manner already described. 13 The following flow-rates of argon were employed purge gas, 40 rnl min-1 except when determining silver antimony and thallium when the flow-rates were 60 SO and 30 ml min-1, respectively; stir gas 300 ml min-1; and window gas 1 1 min-1. Other instrumental settings are shown in Table 2. With the appropriate hollow-cathode lamp in position recordings of absorbance versus time were obtained for each sample.Peak areas (absorbance x time in seconds) were determined by multiplying the peak height by the width at half-height. Sample masses were within the range 0.1-10 mg and were determined with a five-place balance. Glass chips could be added to the furnace every 2 min. Each sample should consist of a single chip. Resistively heated furnace A chip of glass (0.1-2 mg) was placed on a microboat (uncoated graphite Ringsdorff RWO 322) and inserted into the cuvette at room temperature. The cuvette was then heated to a high temperature in a programmed manner and allowed to cool to near room temperature before insertion of the next sample. Peak areas were determined as for the induction furnace.Samples could be processed every 2-3 min. These masses were determined with a five-place balance. Calibration graphs with the induction furnace For the determination of silver lead and iron in glass samples, calibration graphs of peak area versus mass of element were obtained by adding increasing masses of SRM 614 or SRM 616 to the graphite core at a constant temperature (see Table 2) t ANALYST FEBRUARY 1984 VOL. 109 115 ~~ ___ ~ ~ _ _ _ ~ _ _ _ _ ___ ~ Table 3. Instrumental settings for the IL 555 furnace and Varian AA-6. The microboats were cleaned by holding at 2800 "C for 5 s at the end of each programme Resonance Slit width/ Element under study line/nm nm Temperature programme/"C ramp step Silver (solid samples) . . . . . . 328.1 0.5 Ambient- 1000 hold 5 s- 2 700 hold 5 s 5 s Silver(so1utions) .. . . . . 328.1 ramp ramp ramp step 20s 20s 20 s 0.5 Ambient- 60 - 120 - 500 hold 5 s + 2 800 hold 5 s ramp step Bismuth (solid samples) . . . . 306.8 0.4 Ambient - 1 200 hold 5 s -2 700 hold 5 s 5 s ramp ramp ramp step Bismuth (solutions) . . . . . . 306.8 0.4 Ambient- 60- 120 -400 hold 5 s -2 700 hold 5 s 20s 20s 20 s produce absorbances up to 1.0. Nine samples were used to obtain a calibration graph. Sample masses should not exceed 10 mg. In a further limited study calibration graphs were also obtained for thallium antimony manganese zinc and cad-mium using standard glass samples added directly to the furnace to assess the potential of the method for determining these elements in glasses. The absorbances were obtained using the experimental conditions shown in Table 2.Calibration graphs with the IL 555 For the determination of silver in glass samples calibration graphs of peak area versus mass of element were obtained by adding increasing masses of SRM 614 (0.2-0.6 mg) to uncoated graphite microboats which were heated under the conditions shown in Table 3 so as to produce absorbances up to 0.7. Nine samples were again used to obtain a calibration graph. For bismuth calibration graphs were produced using 1-10-p1 volumes of a 0.2 pg ml-1 solution of bismuth as nitrate dispensed on to microboats coated with pyrolytic graphite. (Solutions tend to sink into uncoated microboats. This results in curved calibration graphs.) Procedure for the Determination of Silver Lead Bismuth and Iron in Glasses Using the Furnaces Induction furnace When a series of glasses is to be analysed for a particular element suitable masses (six of each glass) are introduced into the induction furnace whose core has been constructed from the densest graphite available (see Discussion).During the same run various masses of a standard glass are also added for the purpose of constructing a calibration graph. The maximum number of glass samples that can be added to the core before its replacement is about 500 corresponding to approximately 2.5 g of glass. When the run involving 33 samples has been completed the calibration graph is drawn and the mass of element in each sample is obtained from the graph. The concentrations of element in the samples are then calculated.Resistively heated furnace The procedure is similar to that for the induction furnace except that samples are placed on microboats which are pushed horizontally into the JL 555 furnace. Glass chips weighing more than 2 mg are too large to pass through the entrance slit of the cuvette. Masses should be restricted within the range 0.1-2 mg. The total mass of samples that can be added to any microboat for repeated introduction to the furnace is 10 mg. Accumulating masses in excess of 10 mg produces a hole in the bottom of the microboat as a result of the reaction of silica with hot graphite. Procedures for the Determination of Silver and Bismuth in Glasses After Dissolution Because only one standard glass (SRM 614 0.42 pg 8-1 of silver) was available with a silver content near to those of the glasses being analysed it was considered necessary to analyse a few glasses for silver after dissolution to check on the accuracy of the FAAS method with solid samples.This was done by the well established method of FAAS using solutions and the method of standard additions. Likewise a glass was analysed for bismuth after dissolution as no glasses of known bismuth content were available and it cannot be assumed without additional evidence that aqueous bismuth nitrate solutions can be used to construct a calibration graph for the accurate analysis of solid samples of glasses added directly to a furnace . I 6 Method for silver Dissolve an appropriate mass of glass within the range 0.1-0.4 g in a PTFE beaker with 10 ml of concentrated hydrofluoric acid.Heat until all the glass has dissolved and evaporate to dryness. Add 10 ml of water swirl the solution until all the solids have dissolved then add 5 ml of concentrated nitric acid and evaporate to dryness. Repeat this procedure. Dissolve the residue in 10 ml of water and dilute to 25 50 or 100 ml in a calibrated flask with nitric acid and water such that the final concentration of nitric acid is about 2 M. Prepare a blank solution in an identical manner by adding all the reagents but no glass. Pipette 5 ml of the blank solution and 5-ml aliquots of the glass solution into dry flasks labelled blank 0 1 2 3 and 4. Add lo- 20- 30- and 40-p1 volumes of standard silver solution into flasks 1 2 3 and 4 respectively and mix thoroughly (the volume change is negligible).The mass of glass taken and the concentration of the standard silver solution should be such that the amount of silver added in the first standard addition (10 pl) should be similar to that to be expected in 5 ml of the glass solution. Pipette 10 pl of blank solution using an all-glass micro-pipette into a new pyrolytic graphite microboat and deter-mine the peak absorbance for silver using the IL 555 furnace with the instrumental settings shown in Table 3. Repeat this operation with the same tray processing the solutions in the following order to correct for any decrease in sensitivity with time another blank 0 0 1 2 3 4 4 3 2 1 0 0 blank, blank. Average the results for each concentration subtract the average blank level and determine the concentration of silver in flask 0 from the standard addition plot.Repeat this procedure for the series of flasks twice more and average the three results to obtain the silver content of the glass. Method for bismuth This is identical with that for silver except that the tempera-ture programme for the IL 555 furnace is that shown for bismuth in Table 3. Results Results for the determination of lead in glasses using the induction furnace and SRM 616 (1.85 pg g-1 of lead) to construct the calibration graphs and a core made from UF 4 116 ANALYST FEBRUARY 1984 VOL. 109 Table 4. Results for the determination of lead in glasses using the induction furnace SGT5 . . SCT6 . . SGT7 . . EC1.l .. EC1.2 . . A . . . . B . . . . c . . . . D . . . . E . . . . F . . . . SIC791 . . Hutton . , Bagots Park SRM614* . . SRM613t . . Glass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lead content foundlyg g ~ 10 11 4.4 5.0 4.8 4.8 5.0 2.3 1.8 2.3 1.5 1.7 2.0 1.8 5 .O 5.2 13 23 36 38 35 2.22 * Certificate value 2.32 pg g-1 of lead. t Certificate value 38.6 pg g-1 of lead. R . s . ~ . 9'" 2 10 5 9 5 7 3 8 5 14 13 18 11 6 4 11 5 12 5 6 2 12 Table 5 . Approximate results for the determination of lead in some glasses using the induction furnace.The standard used was SRM 616 (1.85 ~g g - ' of lead) Lead content Lead content Glass foundipg g- 1 Glass foundlyg g-1 SGT 4 . . . . . . 100 YM3 stained . . >500 Rosedale . . . . 60 YM4 clear . . . . 60 Kimmeridge . . 100 YM4,stained . . 150 YM1 clear . . . . 70 YM6 clear . . . . 50 YMl stained . . 200 YM6,stained . . 250 YM3,clear . . . . 100 graphite are shown in Table 4. Some approximate results for lead in glasses containing more than 40 v g g-1 of lead are shown in Table 5. The characteristic mass of lead i.e. the mass of element for 1% absorption varies with core tempera-ture being 90 pg at 2 100 "C and 30 pg at 2 400 "C. Results for the determination of silver in the glasses using the induction furnace and for some glasses the resistively heated furnace employing SRM 614 (0.42 yg g-1 of silver) to construct the calibration graphs are shown in Table 6, together with the silver contents of three glasses obtained by the standard additions method after dissolution.The charac-teristic mass for silver was 55 pg at 2 400 "C using the induction furnace and 3.3 pg using the IL 5.55. Because no glasses standardised for bismuth were available , it was necessary to construct calibration graphs from standard solutions of bismuth. These are much more readily handled with the IL 555 furnace and hence all the results reported for bismuth in Table 7 were obtained with this furnace. Peak-height rather than peak-area measurements were employed for peaks with absorbances less than 0.05.The characteristic mass for bismuth was 16 pg. Glass SGT3 was also analysed for bismuth by AAS after dissolution and a result of 13 pg g-1 was obtained. From calibration graphs obtained for thallium antimony, manganese iron zinc and cadmium using standard glass samples added directly to the induction furnace the following characteristic masses were obtained after correcting for any background absorption thallium 110 pg (2 250 "C) antimony 550 pg (2 250 "C) manganese 550 pg (2 400 "C) iron 1 100 pg (2 350 "C) zinc 13 pg (2 300 "C) and cadmium 14 pg (2 350 "C); the temperatures in parentheses are those of the induction furnace for the element under study. When SRM 614 was analysed for iron using a calibration graph prepared with the induction furnace from SRM 616 (11 pg 8-1 of iron) an iron content of 13 pg g-1 was obtained identical with the certificate value.Discussion The results in Table 4 for the determination of lead are of acceptable accuracy. It will be noticed that the average result for SRM 613 (36.3 pg g-1 of lead) is slightly low. This is because SRM 613 produces a wider peak of absorbance versus time than does SRM 616 and it is likely that a small amount of lead escapes through the walls of the graphite core before the bulk of the lead reaches the light path. This is borne out by the fact that in earlier determinations the core was constructed from the less dense graphite AGTS (sp.gr. 1.60) and the average lead content for SRM 613 was calculated as 30 pg g-1.Diffusion of lead through the walls is greatly reduced by using the denser UF4S graphite (sp. gr. 1.72). The average relative standard deviation for the determination of lead in glass is 8%, which is similar to that reported for the determination of trace levels of lead in metals.9 A limit of detection cannot be obtained from the results reported in Table 4 because all of these concentrations are very much greater than the limit of detection. However it is possible to calculate an approximate limit of detection using the equation previously reported by Headridge and Nicholsonl7: Estimated limit of detection = mass of element for o.2 average r.s.d. of 1 % absorption themethod (70) mass of sample This was calculated as 0.014 pg 8-1 of lead in glass at 2 100 "C using a 10-mg sample.The results shown in Table 5 are only approximate because the method is too sensitive to produce accurate results for lead at concentrations above 40 pg g-1 owing to the very small masses of glass that have to be analysed. They are included mainly to show that the ancient glasses usually have lead contents much higher than modern glasses. It is interesting that the stained part of the York Minster glasses always contains more lead than the clear part. The staining is caused by copper dispersed throughout the glass and this copper probably contains lead which produces the elevated lead levels in the stained portions. The results in Table 6 for the determination of silver are of acceptable accuracy agreement between the three methods being very satisfactory.The results for silver obtained using the induction furnace are probably slightly low for a few glasses such as SRM 613 because these results were obtained early in the investigation when cores were constructed from AGTS graphite rather than the denser UF4S. However, glasses such as SRM 613 because these results were obtained not much wider than that for the standard (SRM 614) and results that were significantly low because of increased loss of silver by diffusion through the core walls were not expected, in contrast to the situation with lead. Even so it is suggested that the denser grades of graphite should be used for core construction if results of the highest accuracy are to be obtained for silver. The average relative standard deviation for the determina-tion of silver in glass using the induction furnace is 11 YO if the results for MC 791 and the stained York Minster glasses are not included.This is similar to the relative standard deviations reported for the determination of silver in nickel-base alloys. 18 The limit of detection for silver in glass calculated from the data obtained for glass MC 791 was 0.004 yg 8-1. It will b ANALYSI I-LBRUARY 1984 VOL. 109 i i 7 .~ Table 6. Results for the determination of silver in glasses. with relative standard deviations in parentheses Silver content foundipg g- I SGT2 . . SGT3 . . SGT4 . . SGTS . . SGT6 . . SGT7 . . SGT9 . . EC1.l . . EC1.2 . . A . . . . B . . . . c . . . . D . . . . E . . . . F . .. . MC791 . . Hutton . . Bagots Park Rosedale . . Kimmeridge YM1 clear YM1 stained YM3 clear YM3 stained YM4 clear YM4 stained Y M6 clear YM6 stained SRM613* . . * Certificate value 22 pg g-l of lead Glass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Induction furnace 0.035 (14) 0.32 (2) 0.088 (18) 0.024 (16) 0.030 (lo) 0.026 (8) 3.1 (1) 0.028 (13),0.025 (14) 0.058(10),0.065 (13) 0.34 (11) 0.67 (10) 0.41 (7) 0.37 (10) 0.31 (9) 0.40 (14) 0.007 (30) 0.093 (18) 0.16 (14) 0.67 (6) 0.36 (12) 4.9 (26) 0.74 (4) 0.37 (12) 0.39 (11) 7.3 (9) 0.54 (6) 3.8 (32) 0.23 (16) 1.7 (32) 19 (161 21 (71, 19 (7),19 (7) Standard additions method after IL 55s dissolution 0.89 ( 5 ) 0.73 0.11 (39) 0.19 (S) 0.23 (6) 3.0 0.33 (14) 0.89 (10) 0.41 (15) 0.26 (19) 0.46 (9) 0.43 (4) 0.87 (8) 0.40 (9) 0.78 Table 7.Results for the determination of bismuth in glasses using the IL 555 Bismuth content Glass foundlpgg-1 R.s.d. 70 SGT2 . . . . . . . . . . 0.02* 47 SGT3 . . . . . . . . . . 1.5 11 SGT4 .. . . . . . . . . 0.15 20 0.20 14 SGTS . . . . . . . . . . 0.07 4 SGT6 . . . . . . . . . . 0.05* 19 SGT7 . . . . . . . . . . 0.03* 35 Kimmeridge . . . . . . . . 0.49 14 Rosedale . . . . . . . . . . 8.3 3 Bagots Park . . . . . . . . 0.12* 21 * Using peak height. noticed from Table 6 that the silver concentrations of the stained parts of the York Minster glasses are very much higher than those in the clear glasses. As for lead the elevated silver levels in the stained parts probably arise from the presence of silver in the copper used to produce the red stain. Because the induction furnace is situated within a Perkin-Elmer 300 which is not fitted with a background corrector it was necessary to check that background absorption was negligible under the conditions employed for the determina-tion of lead and silver.This was done using a hydrogen lamp at the wavelength employed for lead and silver. At no time was the absorbance in excess of 0.01. The results in Table 7 for the determination of bismuth are only approximate as aqueous bismuth solutions were used to construct the calibration graphs but the results are probably accurate to within 15% as results of this accuracy have been obtained for the determination of bismuth in steels nickel-base alloys and coppers when calibration graphs were also constructed from standard bismuth nitrate solutions. 16 This conclusion seems reasonable because the bismuth content of glass SGT3 obtained by analysing the solid and by AAS after dissolution is 15 and 13 pg g-1 respectively.The dissolution method is not sufficiently sensitive to determine levels of bismuth below 0.5 pg 8-1. For the determination of bismuth in solid samples of glasses added to the IL 555 furnace the average relative standard deviation for results based on peak area was 11% which is usual for this technique. The limit of detection as calculated from the peak-height results for SGT2 SGT6 and SGT7 was 0.02 pg g-1 of bismuth. In conclusion it can be stated that a very sensitive method of acceptable accuracy and precision has been developed for the determination of silver lead and bismuth in glasses. With a background corrector fitted to the Perkin-Elmer 300 atomic-absorption spectrometer used with the induction furnace it should also be possible to determine very low levels of thallium zinc and cadmium in glasses and concentrations of antimony manganese and iron in excess of approximately 0.2 pg g-1.Undoubtedly this list of elements could be extended if trace element contents of other volatile elements in the standard reference materials SRM 613 614 and 616 became available. This would be particularly helpful in connection with the determination of trace elements in optical fibres. The concentrations of light-absorbing elements in glasses used for optical fibres must be extremely low if long-distance laser communication systems using fibre optic waveguides are not to suffer unacceptable levels of attenuation (the power loss must be <20 dB km-1).19 The light-absorbing ions are V(III), Mn(III) Cr(III) Fe(II) Co(II) Ni(I1) and Cu(I1).Th 118 ANALYST FEBRUARY 1984 VOL. 109 quality control of these glasses requires an analytical tech-nique that is both extremely sensitive and reasonably rapid. It should be possible to extend the furnace methods with solid samples to the analysis of optical fibres. It has been shown that iron can be determined in glass using AAS and the induction furnace and it seems reasonable to assume that this method could be applied to other transition elements when suitable standards become available. The concentrations of bismuth and silver in glasses have been found to be more variable than those of lead and for forensic purposes bismuth and silver would appear to be better trace elements than lead for distinguishing between samples.We are indebted to the Science and Engineering Research Council for a Studentship (for I.M.R.). We thank the British Glass Industry Research Association the Metropolitan Police Forensic Science Laboratory and the Department of Ceram-ics Glasses and Polymers of the University of Sheffield for samples of glasses. 1. 2. 3. 4. References Catterick T and Hickman D. A, Analysr 1979 104 516. Newton R . G. Glass Technol. 1980,21 173. Tong S . S. C . Su Y.-S. and Williams J . P. Anal. Chim. Acta 1976 84 327. Dobbs M. G. D. German. B. Pearson E. F. and Scaple-horn A. W. J . Forensic Sci. SOC. 1973. 13 281. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Gills T. E. Marlow W. F. and Thompson B. A . Anal. Chem. 1970 42 1831. Coleman R . F . and Wood G. A . AWRE Rep. No. 03/68, 1978. Fuller C. W. Anal. Chim. Acta 1974 68 407. Williams J. P. Su Y.-S. and Wise W. M. Mikrochim. Acta, 1977,II 527. Headridge J . B. Spectrochim. Acta Part B 1980 35 785. Siemer D. D. and Horng-Yih Wei Anal. Chem. 1978 50, 147. Grushko L. F. Krasil’schik V. Z. Lifshits M. G. and Chupakhin M. S . Zh. Anal. Khim. 1977 32 218. Vul’fson E. K. Karyakin A . V. and Shidlovskii A. I. Zh. Anal. Khim. 1973 28 1253. Andrews D. G. and Headridge J . B. Analyst 1977,102,436. Baker A. A. and Headridge J . B. Anal. Chim. Acta 1981, 125 93. Lundberg E. Chem. Instrum. 1978 8 197. Headridge J . B . and Riddington I. M. Mikrochim. Acta, 1982 11 457. Headridge J . B. and Nicholson R . A. Analyst 1982 107, 1200. Baker A. A . Headridge. J . B. and Nicholson R. A Anal. Chim. Acta 1980 113 47. Campbell D. E. Su Y.-S and Williams J . P Phys. Chem. Glasses 1976 17 108. Paper A31211 Received July 11 th 1983 Accepted September 29th 198

ISSN:0003-2654    

DOI:10.1039/AN9840900113

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST FEBRUARY 1984 VOL. 109 119 Quantitative Determination of Lead and Cadmium in Foods by Programmed Dry Ashing and Atomic-absorption spectrophotometry with Electrothermal Atomisation The0 Muys Division for Nutrition and Food Research TNO ClVO Institutes TNO Zeist The Netherlands This paper describes the simultaneous determination of lead and cadmium in foods by programmed dry ashing in the presence of magnesium nitrate complexing with sodium diethyldithiocarbamate extraction with isobutyl methyl ketone and atomic-absorption spectrophotometry using the graphite furnace technique. Recoveries are 94-100% for lead and 94-109O/0 for cadmium. Keywords Lead and cadmium determination; programmed dry ashing; atomic-absorption spectrophotometry; electrothermal atomisation; foods The determination of lead and cadmium in food products for both man and animals has already received considerable attention and plays an important role in trace element analysis.The principal sources of environmental contamina-tion of foods by lead are exhaust gases industrial discharge and its presence in the soil; industrial food processing can also give rise to lead contamination. Cadmium contamination is largely accounted for by the volatilisation of the metal from metallurgical and other industrial plants but the intensive application of fertilisers can lead to soil being a source as well. 1.2 Because lead and cadmium are both highly toxic suffi-ciently sensitive methods are required for their determination. However the volatile character of the two metals in an elemental as well as in an organic-bound form not only contributes to their spreading over large areas but compli-cates the analysis because their volatilisation must be preven-ted in an ashing procedure prior to determination in the food product.In the literature numerous procedures for the ashing are described including dry ashing with or without ashing aids and with or without temperature programming"-7 and wet digestion with mixtures of sulphuric and nitric acids sulphuric acid and hydrogen peroxide and nitric and perchloric acids. 8,') Further expedients include high-pressure digestion in a decomposition vessel and low-pressure low-temperature ashing. 10 All of these techniques have advantages and limitations. In this study programmed temperature ashing with magnesium nitrate as an ashing aid added in ethanolic solution was chosenll912-ashing can be carried out at night which saves time and a relatively large sample can be analysed ( 5 g).After dissolving the ash in acid sodium diethyldithiocarb-amate (NaDDTC) was added as a complexing agent followed by extraction of the complex in isobutyl methyl ketone (IBMK). In the extract the levels of lead and cadmium were measured by atomic-absorption spectrophotometry with the graphite furnace technique. The method is highly sensitive and has good reproducibility. Experimental Apparatus The following apparatus was used 30- and 50-ml quartz dishes warmed with 7 N nitric acid and rinsed thoroughly with de-mineralised or distilled water; an electrical hot-plate or drying oven; a muffle furnace with temperature programming; 50-ml polythene bottles rinsed thoroughly before use with demineralised or distilled water; a pH-meter; and a Perkin-Elmer Model 430 atomic-absorption spectrophotometer with a Model HGA 500 graphite furnace a Model AS-1 auto-sampler and PTFE sample cups.Reagents Analytical-reagent grade. Magnesium nitrate solution 10% mlV in 95% ethanol. Nitric acid 65% (sp. gr. 1.40). Analytical-reagent grade. Ammonia solution 25% (sp. g r . 0.91). Analytical-reagent Isobutyl methyl ketone. Analytical-reagent grade. Phenol red solution 1% mlV. Dissolve 1 g of phenol red in 100 ml of de-mineralised water heat on a water-bath cool and filter the solution with a white ribbon filter. Potassium cyanide solution 10% mlV in de-mineralised water.Sodium diethyldithiocarbamate (NaDD TC) solution 1 YO mlV in de-mineralised water. Prepare fresh daily and purify by extracting twice with IBMK. Ammonium citrate buffer solution. Dissolve 400 g of citric acid (analytical-reagent grade) in 100 ml of de-mineralised water. Add 400 ml of ammonia solution and 40 ml of a 1 YO phenol red solution. Adjust the pH to 8.8 by adding ammonia solution or citric acid solution using a pH-meter. Bring the volume to 11. Transfer the buffer into a 2-1 separating funnel, add 25ml of NaDDTC solution mix add approximately 80 ml of IBMK and shake for 2min. Allow to separate and discard the organic layer. Repeat the procedure and store the buffer solution in a polythene bottle.Lead stock standard solution (1 ml = 1 mg of lead). Weigh 1.598 5 g of Pb(N03)2 and rinse into a 1-1 calibrated flask with de-mineralised water. Add 70 ml of 65% nitric acid and make up to volume. Store the solution in a polythene bottle. Cadmium stock standard solution (1 ml = 1 mg of cad-mium). Weigh 2.2820g of 3CdS04.8H20 and proceed as described for the lead stock standard solution. Combined lead - cadmium standard solution containing 1 p.p.m. of leadand 0.1 p.p.m. of cadmium. Pipette 500 pl of the lead stock standard solution and 50yl of the cadmium stock standard solution into a 500-ml calibrated flask and make up to volume with 3% VlV nitric acid. Mix and transfer into a polythene bottle. Prepare fresh daily. grade. Procedure Weigh into a quartz dish 1.&5.0 g of sample and add 5 ml of the magnesium nitrate solution.Dry on a hot-plate by gently raising the temperature until the sample is completely dry. Place in the muffle furnace and apply the following tempera-ture programme 2 h at 150 "C 2 h at 250 "C 3 h at 350 "C and 6 h at 450 "C. If the ashed residue contains residual carbon, wet it with 1 ml of de-mineralised water add 1 ml of 65% nitric acid dry on the hot-plate and place the samples in the muffle furnace at 350 "C for 30 min 120 ANALYST FEBRUARY 1984 VOL. 109 Take up the fully ashed sample in 1 ml of water and 5 ml of 65% nitric acid warm on a hot-plate for a few minutes and rinse with 25 ml of de-mineralised water into a 50-ml polythene bottle. Add 4ml of ammonium citrate buffer solution and enough ammonia solution to obtain a deep red colour.Check the pH with the pH-meter and adjust to 8.8 k 0.2. Allow to cool add 1 ml of potassium cyanide solution and mix. After 5 min add 5 ml of the NaDDTC solution mix and allow to stand for 10 min. Add 5.0 ml of IBMK and shake for 1min. After separation of the organic phase transfer the IBMK layer the sample extract into a PTFE cup of the autosampler for atomic-absorption measurement. Prepare reference extracts as follows pipette 25ml of de-mineralised water and 5 ml of nitric acid into each of five polythene bottles and add 0 125 250 375 and 500 p1 of the combined lead - cadmium standard solution. Proceed as above i.e. starting from "Add 4ml of ammonium citrate buffer." The lead concentrations in the reference extracts will thus be 0 25 50 75 and 100 parts per billion (p.p.b.) and the cadmium concentrations 0 2.5 5 7.5 and 10 p,p.b.All extracts are measured in succession on the day of their preparation. The concentrations in the sample extracts are found by comparing their peak heights with those of the reference extracts. Measurement Conditions Lead Using an electrodeless discharge lamp and a molybdenum-coated graphite tube,l3 measurements for lead were carried out under the following conditions background correction, on; wavelength 283.3 nm; slit width 0.7 nm; integration time, 8 s; and recorder 1 V. The peak height was measured using the following pro-gramme: Step 1 100 "C 25 s ramp 20 s hold Step 2 110°C 10 s ramp 10 s hold Step 3 120 "C 10 s ramp 10 s hold Step 4 600 "C 25 s ramp 25 s hold Step 5 2 300 "C 3 s ramp 7 s hold An internal gas flow-rate of 10 ml min-J of argon a 2-s read time and a 2-s record time were also used.Cadmium Similarly using an electrodeless discharge lamp and a normal graphite tube measurements for cadmium were carried out under the following conditions background correction on; wavelength 228.8 nm; slit width 0.7 nm; integration time, 8s; and recorder 1 V. The peak height was measured using the following pro-gramme: Step 1 110 "C 20 s ramp 20 s hold 10 s hold Step 2 110°C 10 s ramp Step 3 120 "C 10 s ramp 10 s hoid Step 4 450 "C 25 s ramp 25 s hold Step 5 2 000 "C 3 s ramp 7 s hold An internal gas flow-rate of 50ml min-1 of argon a 2-s read time and a 2-s record time were also used.Results and Discussion For the determination of lead and cadmium in food products, good results have already been attained with the application of NaDDTC as a complexing agent.14 In order to obviate interference from large amounts of copper zinc or iron which are possibly present in the food product a masking process with potassium cyanide has been simultaneously applied. l4 The addition of phenol red to the ammonium citrate buffer solution speeds up the crude neutralisation of the acidic sample solution by visual observation. The pH value with which an optimum result is obtained (9.0 4 0.5) is beyond the neutral point of the indicator and has to be checked with a pH-meter. As a check on the ashing procedure we conducted a series of tests in which identical amounts of lead (in the 0-500ng range) and cadmium (in the 0-50 ng range) were subjected to analysis without and with previous ashing.The resulting absorbance values are given in Table 1; intercomparison shows that the ashing procedure with the specified temperat-ure programme does not measurably affect the results. Statistically using a paired t-test the difference was found to be insignificant at the 95% confidence level. The extraction of the two elements was then examined. To that end identical amounts of complexed lead and cadmium Table 1. Standard amounts of lead and cadmium subjected to analysis without and with ashing. All results are the means of two duplicate analyses; the differences between the blanks without and with ashing are caused by magnesium nitrate and results are given as peak heights to to give a comparison between the sensitivity to both elements Lead absorbance arbitrary units Cadmium absorbance arbitrary units Without ashing With ashing Without ashing With ashing Corrected Corrected Corrected Corrected Samplehg Uncorrected for blank Uncorrected for blank Sampleing Uncorrected for blank Uncorrected for blank 125 173 125 225 120 12.5 162 121 185 120 250 29 1 243 335 230 25 270 229 299 234 375 407 359 447 342 37.5 3 80 339 410 345 500 535 487 555 450 50 500 459 504 439 - 65 - Blank 48 - 105 - Blank 41 Table 2.Extraction of identical amounts of lead (100 ng) and cadmium (10 ng) in different volumes of IBMK. The values of the absorbance were corrected for the total blank consisting of the constant contribution from the IBMK (lead = 0 cadmium = 4 units) plus the proportional contribution from the blank which for 5 ml was 33 units for lead and 24 units for cadmium Lead Cadmium Concentration after Concentration after IBMW complete extraction complete extraction, ml p.p.b.Absorbance p.p.b. Absorbance 20 25 33 50 100 84 130 167 24 1 500 2 2.5 3.3 5 10 129 165 225 335 67 ANALYST FEBRUARY 1984 VOL. 109 121 Table 3. Effect of internal gas flow-rate on peak height for cadmium. Test solution IBMK extract containing 2.5 p.p.b. of cadmium Internal gas flow-rate/ mlmin I 0 10 30 50 80 100 Peak height 493 42 1 302 193 117 101 Table 4. Recovery of lead and cadmium in samples of chicken meat and eggs spiked with 50 p.p.b.of lead and 5 p.p.b. of cadmium. All results are the means of duplicate analyses. Lead. Recovery Cadmium Recovery, Sample p.p.b. Yo p.p.b. O/” Chicken meat A Chicken meat A, spiked . . Chicken meat B Chicken meat B, spiked . . EggsA . . Eggs A spiked E g g s B . . . . Eggs B spiked . . 13 . . 60 . 19 . . 66 . . 9 . . 59 . . 3 . . 50 6 9s 10.7 97 3 96 7.5 94 <o. 1 100 5 100 0.3 94 5.8 109 Table 5. Determination of lead in milk powder. Results are of the duplicate analyses Lead concentration/mg k g Values found with this Values of inter-Sample method laboratory tests A . . . . 0.70,0.65 0.63,0.68 B . . . . 1.2.5,1.28 1.30,1.35 c * .. . 1.87,1.87 1.94 1.99 1 . . . . 1.08,1.04 1.03 1.07 2 . . . . 2.05,2.05 2.05,2.09 were extracted in 5 ml of IBMK shaking for different lengths of time. The same absorbance was obtained with extraction times of 30,60 and 120 s. In the final extraction a shaking time of 1 min was used. The volume of IBMK was also varied. As shown in Table 2, the absorbance was inversely proportional (as expected) to the volume of IBMK over the range 1-5ml. Therefore a 1-ml, rather than a 5-ml volume of IBMK is to be preferred if the highest sensitivity is required or when the sample size is limited as may be so with blood serum or other biological materials. It proved to be essential to purify the NaDDTC solution by extracting twice with IBMK and the ammonium citrate buffer solution by treating with NaDDTC and extracting twice with IBMK in order to obtain low blanks for lead and cadmium.In the furnace of the HGA 500 an adjustable flow of gas can pass through the graphite tube during the atomisation step (step 5 of the temperature programme). By interrupting or accelerat-ing the gas flow-rate one can adjust the retention time of the vaporised atoms in the light beam and thereby vary the peak height. Table 3 gives the results of a typical test series. The sensitivity of the lead measurement can be increased considerably by coating the graphite tube with molybde-num.13 In this investigation a coated tube was used for more than 200 measurements without any change in sensitivity being observed. Performance Tests The method described has been applied to a number of samples of chicken meat and eggs.Recoveries were checked by adding 250 mg of lead and 25 ng of cadmium. The results, showing a recovery of 97 k 3% for lead and 100 +_ 10% €or cadmium are summarised in Table 4. Five samples of milk powder that had been used in inter-laboratory tests for the presence of lead were analysed as described under Procedure. The results in Table 5 obtained using this procedure are in good agreement with those obtained by the participants of the inter-laboratory test. Six participants used atomic-absorption spectrophotometry and two differential-pulse anodic-stripping voltammetry. Statistic-ally there is no significant difference at the 95% confidence level. From participating in inter-laboratory tests it has been shown during the last few years that this method is also applicable to fish products bread vegetables and animal feedstuffs.If the dry-ashing procedure is replaced by wet digestion with 5ml of concentrated sulphuric acid and nitric acid followed by extraction of the nitrous fumes with de-mineralised water the method can be followed as descri-bed under Procedure starting at “. . . and rinse with 25 ml of de-mineralised water . . .” Statistically there is no significant difference between the methods as results for samples of grass meal milk powder and eggs have shown. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Lagerwerff J . V. and Specht A. W. Environ. Sci. Technol., 1970 4 583. Korkish J . and Hazan L. Anal. Chem. 196.5 37,707. Schonhard G. and Schenke A. D. Landwirtsch. Forsch., 1976 29 248. Roschnik R. K. Analyst 1973 98 596. Reith J . F. Engelsma J. and van Ditmarsch M . 2. Lebensm. Unters. Forsch. 1974 156 271. Boppel B. 2. Lebensm. Unters. Forsch. 1975 158 287. Fiorino J. A. Moffitt R. A. Woodson A. L. Gajan R. J . , Huskey G. E. and Scholz R. G. J. Assoc. Off. Anal. Chem., 1973 56 1246. Mack D. Dtsch. Lebensm. Rundsch. 1979 10 309. Woidich H. 2. Lebensm. Unters. Forsch. 1974 155 72. Dewitt A. Bull. SOC. Chim. Belg. 1975 84 91. Horwitz W. Editor “Official Methods of Analysis of the Association of Official Analytical Chemists,” Eleventh Edi-tion Association of Official Analytical Chemists Washington, DC 1970. Friend M. T. Smith C. A. and Wishart D. At. Absorpt. Newsl. 1977 16 46. Manning D. C. Anal. Chem. 1978 50 1234. Spiegelenberg W. De Ware(n) Chemicus 1978 8 183. Paper A21186 Received July 29th 1982 Accepted August lath 198

ISSN:0003-2654    

DOI:10.1039/AN9840900119

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1984, VOL. 109 123 Extraction - Atomic-absorption Spectrophotometric Determination of Antimony by Generation of Its Hydride in Non-aqueous Media Jose Aznarez, Francisco Palacios, Maria S. Ortega and Juan Carlos Vidal Department of Analytical Chemistry, Faculty of Sciences, University of Zaragoza, Zaragoza, Spain A method for covalent hydride generation in a non-aqueous extraction phase is proposed. The hydride generation is carried out in an aliquot of metal - complex extraction solution by sodium tetrahydroborate(ll1) in N,N-dimethylformamide solution and anhydrous acetic acid. The proposed method gives improved sensitivity, eliminates interferences and enables hydride generation of difficult elements to be carried out. Antimony is extracted by ammonium pyrrolidine-I -carbodithioate into chloroform and after hydride generation by the proposed method, it is determined by flame AAS.The method is applied to the determination of antimony in BCS standard steels with good accuracy and precision. Keywords: Antimony determination; atomic-absorption spectrophotometry; hydride generation; non- aqueous media The determination of As, Bi, Ge, Pb, Sn, Sb and Te by atomic-absorption spectrophotometry (AAS) with covalent hydride generation using sodium tetrahydroborate(II1) solu- tion is a well established analytical procedure. In the future development of determinations with hydride generation, there are two possible areas, improved automation and methods of lowering detection limits,' which may improve as manufacturers introduce improved designs for their instru- ments and reduction vessels or using combinations of hydride generation with techniques such as atomic-fluorescence spec- trophotometry,2 d.c.arc discharges3 or inductively coupled plasmas.3 A further area of research in which much useful work may yet be done is in reduction or elimination of interference effects. In this paper, a procedure for covalent hydride generation from a non-aqueous extraction phase is proposed in which the organic phase after extraction is treated with sodium tetrahy- droborate( 111) solution in N, N-dimethylformamide (DMF) and an acid. This covalent hydride generation in a non- aqueous medium with a previous extraction stage gives improved sensitivity and eliminates interferences, Also, ele- ments that form hydrides with difficulty in aqueous solution, such as lead, are reduced more easily in non-aqueous media, owing to the greater range of electroactivity of some organic solvents than water.In the determination of antimony by hydride generation followed by AAS, an electrically heated silica tube has been used as an alternative to a flame. Detection limits reported with this type of atomisation are 0.5 pg 1-1 of Sbs and 0.2 pg g-1 in geological samples." The Analytical Methods Committee7 reported sufficient sensitivity and quantitative recovery from solutions containing SO pg 1-1 of Sb in organic matter or foodstuffs. Chapman and Dales obtained a sensitiv- ity of 2.6 ng of Sb using an argon - hydrogen flame for atomisation. This lower sensitivity was ascribed to the lower temperature of the flame.In this work, antimony was extracted with ammonium pyrrolidine- 1-carbodithioate into chloroform at pH 6.9 with phosphate buffer solution. The hydride was generated in an aliquot of the chloroform extract with sodium tetrahydro- borate(II1) in DMF and anhydrous acetic acid. The generation of hydride was carried out in an inexpensive generatorY.IO with direct introduction into an air - acetylene flame through the nebulisation device of the spectrophotometer. The proposed method has been applied to the determination of antimony in BCS standard steels with excellent agreement. Experimental Apparatus A Pye Unicam SP-9 atomic-absorption spectrophotometer with a three-slot burner, equipped with a Pye Unicam antimony hollow-cathode lamp, was used. An Orion Research Microprocessor Ionanalyzer 901 was used for pH measure- ments. A Haake mechanical shaker was employed.Reagents and Solutions All chemicals were of analytical-reagent grade, obtained from Merck. Standard antimony solution, 1.000 pg ml-1 of Sb. Dissolve 0.598 6 g of antimony trioxide in 10 ml of concentrated hydro- chloric acid (sp. gr. 1.16) and dilute to S00ml in a calibrated flask with 3 M hydrochloric acid. The working solutions were prepared by diluting this solution immediately before use. Sodium tetruhydroborate( I l l ) solution. Dissolve 1 g of sodium tetrahydroborate(II1) in 100 ml of DMF. This solution can be used for 2 weeks. Sodium sulphite solution, 50 g 1-1. Dissolve 5 g of anhydrous sodium sulphite in 100 ml of water.Prepare this solution as required. Ammonium pyrrolidine-I-carbodithioate solution, 0.25% mlV. Prepare just before use. Sodium diethyldithiocurbumate solution, 0.25% miV. Pre- pare just before use. Clark and Lubs buffer solution @H 6.9). 0.1 M phosphate - sodium hydroxide. 1 1 N,N-Dimethylforrnamide. Purify by distillation, collecting the fraction boiling between 148 and 150 "C. Chloroform. Toluene. Isobutyl methyl ketone (IBMK). Procedure Weigh exactly 0.2-0.4 g of steel (containing up to 250 pg of Sb) and dissolve it in 10 ml of Concentrated hydrochloric acid (sp. gr. 1.16). Add 2 ml of concentratcd nitric acid (sp. gr. 1.47), heat until nitrogen dioxide has been eliminated and boil for 5 min. Dilute to S0ml in a calibrated flask with water. Place 10 ml of the steel solution into a separating funnel and add 1 ml of sodium sulphite solution and 15 ml of buffer124 ANALYST, FEBRUARY 1984, VOL.109 solution (pH 6.9), then 2 ml of pyrrolidine-1-carbodithioate solution. Extract with 10 ml of chloroform by mechanical shaking for 5 min. Leave the phases to separate then dry the chloroform phase by filtering it through cotton-wool. Place 1 ml of the chloroform extract into the hydride generator and add 3 ml of anhydrous acetic acid (sp. gr. 1.05). Inject 4 ml of NaBH4 solution through the septum membrane. Measure the peak height of the atomic-absorption signal of antimony at 217.59 nm. Prepare a calibration graph for known amounts of antimony using the standard antimony solution following the same procedure.Results and Discussion Solubility and Stability of NaBH4 Solution in DMF The solubility of NaBH4 in DMF at 25 "C was studied. Using a gravimetric method, a solubility of 60 g 1-1 was found, After optimising the proposed method, solutions of NaBH4 of concentration 1% mlV in DMF were used. This solution, as distinct from aqueous NaBH4 solution,l2,13 was useful for 2 weeks, as shown by the signals obtained for hydride genera- tion. Its stability was comparable to that of an aqueous solution containing a pellet of sodium hydroxide and filtered through a 0.45-ym membrane filter. Acids for Hydride Generation An acid is needed for covalent hydride generation and the simultaneous evolution of hydrogen by NaBH4 in DMF. The acid used must be miscible with DMF and the extraction solvent.The possibilities studied were hydrogen chloride gas dried and dissolved in DMF or dioxane, anhydrous acetic acid and a mixture of glacial acetic acid and concentrated sulphuric acid (sp. gr. 1.84) (3 + 1). The HC1 contents in DMF or dioxane solutions saturated at 1 atm were determined by neutralisation titrimetry. The HCI concentration in DMF solution was found to be 3.9 M and the solution remained stable for 2 weeks, whereas a dioxane solution (1.4 M HC1 ) rapidly lost the HCl. The best results for the antimony peak height were obtained by using either anhydrous acetic acid or a mixture of acetic and concentrated sulphuric acids. When HC1 in DMF solution or a mixture of acetic and sulphuric acids was used, a white crystalline precipitate was formed (owing to the formation of NaCl or Na2S04, respectively) that was hardly soluble in the organic solvents.With anhydrous acetic acid, the solution remained clear. The hydride generation and the simultaneous evolution of hydrogen were immediate and the reaction was complete in less than 30s. Therefore, there was no need for collection devices and the inexpensive generator described elsewhere9.10 can be used. Solvents A necessary condition for the accurate and rapid evolution of the volatile hydride is miscibility of the extraction solvent, the NaBHj solution in DMF and the acid used. Chloroform, toluene and IBMK were studied as suitable solvents for metal-complex extraction. It was found that there was good miscibility of these solvents with DMF and anhydrous acetic acid or the mixture of acetic and sulphuric acids (3 + 1) in the volume proportions used in the proposed method, so that hydride evolution was carried out in the homogeneous phase.Optimisation of Hydride Generation and Atomic-absorption Signal The optimisation of the atomic-absorption signal with this hydride generation device has been discussed in previous papers."lo Recommended instrumental parameters were acetylene flow-rate, 1.0 1 min-1; air flow-rate, 5.0 1 min-*; burner height, 8 mm; spectral slit width, 0.5 nm; wavelength, 217.59 nm: and lamp current, 12 mA. The conditions for hydride generation were studied and the best conditions were as follows: 1 ml of antimony extract, 3 ml of anhydrous acetic acid or mixture of acetic and sulphuric acids (3 + 1) and 4ml of 1% NaBHj solution 1% in DMF injected through the septum membrane of the generator.Antimony Extraction Antimony has been extracted into toluene from hydrochloric acid solutions, but the extraction yield has always been less than 85%. Therefore, the antimony extraction was carried out with sodium diethyldithiocarbamatel2 or ammonium pyrrolidine-1-carbodithioatel3 into chloroform, which also separates Sb(II1) from Sb(V).l4 In this work total antimony was determined, as the Sb(V) formed was reduced to Sb(II1) by the addition of sodium sulphite solution.7 Calibration Graph The peak height of the atomic-absorption signal for antimony was linear between 0.05 and 5 yg of Sb (in 1 ml of extraction solution), with a regression coefficient of 0.999 1. The noise level of the blanks was indistinguishable from the noise of the determination up to a setting of 4 on the expansion scale of the spectrophotometer.The sensitivity (1% absorption) and the detection limit were 0.03 and 0.01 yg ml-1 of Sb in the extraction solution, respectively. Ten replicate determinations of 2 pg of Sb gave a relative standard deviation of 2.0%. Interference Study Interferences from different elements in the determination of antimony were studied. Up to 300 mg of Al, Ca, Mg, Ni, Cr(III), Fe(III), Fe(II), Sn(IV), Sn(II), As(V), As(III), silicate, fluoride, sulphate, phosphate or tartrate did not interfere in the determination of 50 pg of Sb. Up to 1 mg of Se(IV), Te(IV), Bi(III), Pb(II), Cu(I1) or Ge(1V) did not interfere but larger amounts produced a decrease in the atomic-absorption signal for 50 yg of antimony.As a result of the extraction process and the non-aqueous hydride genera- tion, the interference limits were higher than for generation in an aqueous medium. 1" Applications The proposed method has been applied to the determination of antimony in BCS standard steels, as shown in Table 1. The determination of 25 pg of antimony in spiked samples gave a 98.1% recovery with a relative standard deviation of 1.8% (ten determinations). The method has also been applied to the determination of antimony in air particulates, with good precision. Table 1. Determination of antimony in BCS steels Relative standard O/o BCS 329 mild steel 0.018 0.019 2.0 BCS 457 mild steel 0.029 0.030 1.7 BCS 458 mild steel 0.070 0.068 1.6 Sb certified, Sb found, deviation.* o * Sample % /o * Means and relative standard deviations for ten replicate determi- nations.ANALYST.FEBRUARY 1984, VOL. 109 125 Conclusion The extraction and hydride generation in non-aqueous media with NaBH4 in DMF eliminates interfercnces and improves sensitivity. Elements that have difficulty in forming hydrides, such as Pb. Te and Sn, have also been studied. They are reduced more easily than in water and will be the topic of a later paper. References Godden, R. G . , and Thomerson, D. R., Analyst, 1980. 105, 1137. Nakahara, T . , Kobayashi, S . , and Musha, S . , Anal. Chim. Acta, 1978, 101, 375. Braman, R. S., and Tompkins, M., Anal. Chem., 1978. 50, 1088. Hahn, M. H . , Wolnik, K. A , , and Fricke, F. L., Anal. Chem., 1982, 54, 1048. 5 . 6. 7. 8. 9. 10. 11. 12. 13. 14. Hon, P. K., Lau, 0. W., Cheung, W. C., and Wong, M. C., Anal. Chim. Actu, 1980. 115. 355. Crock, J . G., and Lichte, F. E., Geol. Surv. Openfile Rep. (US). IYS1,81,671. Analytical Methods Committee, Analyst, 1980, 105, 66. Chapman, J. F . , and Dale, L. S . , Anal. Chim. Acta, 1979, 11, 137. Aznarez, J . , Castillo, J . R., Bonilla, A . , and Lanaja, J . , At. Spectrosc., 1981, 2, 125. Castillo, J . R . , Lanaja, J., Martinez, Mcl. C . , and Aznarez, J . , Analyst, 1982, 107, 1488. Meites, L., “Handbook of Analytical Chemistry,” McGraw- Hill, New York, 1963. Nagafuchi, Y., Fukamachi, K., and Morimoto, M., Bunseki Kagaku, 1977, 26, 729. Kamada, T., and Yamamoto, Y . , Talanta, 1977. 24, 330. Shabana, R., and Rif. H., J . Radioanal. Chem., 1978,45,317. Paper A31253 Received August 8th, 1983 Accepted September 23rd, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900123

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1984, VOL. 109 127 Analytical Application of Inorganic Salt Standards and Mixed-solvent Systems to Trace-metal Determination in Petroleum Crudes by Atomic-absorption Spectrophotometry Oladele Osibanjo," Samuel E. Kakulu and Sunday 0. Ajayi Department of Chemistry, University of Ibadan, Ibadan, Nigeria The determination of Ni, Cu, Zn, Na, Pb, Cd and Fe in petroleum crudes by means of a toluene - acetic acid mixed-solvent system, inorganic salt standards and atomic-absorption spectrophotometry is presented. The low systematic errors and good burning characteristics of this mixed-solvent system led to its choice. Good recoveries of metals added t o petroleum crudes are demonstrated. Coefficients of variation of 8.4 for Ni, 7.0 for Cu, 7.8 for Zn, 11.2 for Na, 9.8 for Pb, 10.2 for Cd and 13.9% for Fe are obtained.The validity of the method is shown by comparison of results obtained with those of an established method. Keywords: Metal determination; petroleum crudes; mixed solvents; inorganic salt standards; atomic- absorption spectrophotometry The occurrence of trace metals in petroleum crudes or fractions is useful analytically. For instance, the nickel to vanadium ratio is used in the characterisation of petroleum crudes. 1 Additionally, the determination of elements such as lead, copper and zinc in both fuel oils and gasolines is often required because these metals accelerate the oxidative deterioration of the products or otherwise reduce the storage ability.* The analysis of the metal contents of petroleum crudes and products could give an idea of the amount of heavy metal fall-out into the environment as a result of the burning of oil fuels.Atomic-absorption spectrophotometry has been widely applied to the determination of trace metals in petroleum products. As most of the metals in petroleum are in the form of organometallic compounds, organometallic standards are often used in their determination.3 Omang4 and Chuang and Winefordners have described the determination of metals in mineral oils and jet-engine oils, respectively, using elec- trothermal atomisation with a graphite tube. The analytical application of emulsions in the determination of trace metals in gasoline by atomic-absorption spectrophotometry has also been described by Polo-Diez et aL6 Most methods for determining trace metals in petroleum and their products by atomic-absorption spectrophotometry involve dilution of the sample with a suitable organic solvent, usually xylene, followed by the use of organometallic stan- dards.The organometallic standards are very difficult to obtain and often expensive. Further, different results are obtained when a metal is present in different forms, e.g., the amount of lead obtained with the same organometallic standard would be different when it is present as tetramethyl- lead or tetraethyllead. As a result of these problems, the need to develop a new and relatively inexpensive method of sample preparation prior to analysis by atomic-absorption spectro- photometry has become necessary. Some workers have developed mixed-solvent systems for the determination of metals in lubricating oils by atomic- absorption spectrophotometry using inorganic salt standards.Holding and co-workers'.* used the cyclohexanone - butan-1- 01 - industrial methylated spirit - hydrochloric acid and 2-methylpropan-2-01 - toluene mixed-solvent systems and inorganic salt standards in the determination of trace amounts of Ca, Zn and Ba in unused lubricating oils and obtained results that were in good agreement with those obtained by X-ray fluorescence and established Institute of Petroleum chemical procedures. Wittmannq reported the use of a toluene * To whom correspondence should be addressed. - acetic acid (1 + 4) solvent and inorganic salt standards for the determination of Ca, Mg and Zn in lubricating oils by atomic-absorption spectrophotometry with good precision.The analytical application of inorganic salt standards and mixed-solvent systems for the determination of metals in petroleum crudes has been investigated in this work. Experimental Preliminary Investigation of the Choice of Mixed Solvent The choice of solvents as components of a mixed-solvent system depends on the solubility of the various petroleum crudes and that of the inorganic salt standards. The solubilities of various Nigerian petroleum crudes were tested in several solvents, viz., benzene, toluene, dioxane, 4-methylpentan-2- one, xylene, propanone, 2-methylpropan-1-01 and acetic acid. The petroleum crudes were soluble in all except acetic acid but, because of its good flame characteristics,lO its use as a component of the mixed solvent was considered.The dioxane - acetic acid and 4-methylpentan-2-one - acetic acid systems were found to be unsuitable as there was emulsion formation whenever acetic acid was added to a solution of the oil in either dioxane or 4-methylpentan-2-one. The inorganic salt standards were also insoluble in the mixed solvent systems. Both benzene - acetic acid (1 + 4) and toluene - acetic acid (1 + 5 ) solvent systems were found to be suitable for analysis. The choice of the latter system for this study is due to the less smokey characteristics of the flame produced. Apparatus A Varian AA 475 atomic-absorption spectrophotometer, with various metal hollow-cathode lamps, was used with an air - acetylene flame and other conditions stated in the operations manual.Reagents All chemicals were of analytical-reagent grade unless other- wise stated. Solvents that were not of analytical-reagent grade were purified before use as described in the literature." Benzene. Toluene. Xylene. Propanone. (2) -Meth y lpropan - 1 -01. 4-Methylpentan-2-one.128 ANALYST, FEBRUARY 1984, VOL. 109 Copper. Zinc. AnalaR grade. Sodium chloride. AnalaR grade. Potassium chloride. AnalaR grade. Nickel nitrate. AnalaR grade. Lead nitrate. AnalaR grade. Cadmium chloride. AnalaR grade. Iron(ZZl> nitrate. AnalaR grade. Hydrochloric acid. AnalaR grade. Nitric acid. AnalaR grade. Ethanol. AnalaR grade. Nickel cyclohexylbutyrate. Zinc cyclohexylbutrate. Preparation of Stock Solutions A modification of the method of Wittmang is used for stock solution preparation.Appropriate amounts of copper and zinc metals are dissolved in nitric and hydrochloric acids, respec- tively, diluted with a small volume of distilled, de-ionised water and finally made up to the required volume with ethanol to give a 1 000 p.p.m. solution. Stock solutions of Na, Ni, Pb, Cd and Fe are prepared from their chloride or nitrate salts by dissolving in a small volume of distilled, de-ionised water and diluting with ethanol to the required volume. Potassium chloride, of about 2 000 p.p.m. with respect to potassium, is added to the sodium solution as an ionisation suppressant. Standard Working Solutions These are freshly prepared daily by serial dilution of appro- priate volumes of the stock solution of each metal with toluene - acetic acid (1 + 5 j .Sample Preparation Weigh into a weighing bottle about 0.2 g of crude oil. Add 3.0 ml of toluene and stir well before transferring into a 25-ml calibrated flask. Wash the bottle with more toluene followed by acetic acid and transfer quantitatively into the flask. Finally, dilute to the mark with glacial acetic acid. When benzene - acetic acid is used the oil should be dissolved first in benzene before adding acetic acid, as the reverse order results in precipitation of the oil. The standard additions method was used in the analysis of the crude samples to minimise errors due to matrix effects. Results and Discussion The use of a toluene - acetic acid mixed-solvent system and inorganic salt standards was applied to the analysis of trace metals in Nigerian petroleum crudes from seven terminals.The choice of this system resulted from the relatively low systematic errors (between 2 and 25% j compared with those obtained with the benzene - acetic acid system, which ranged from 15 to 48% for most of the metals determined. This also allowed the use of the inorganic salt standards as there was no precipitation as observed in the preparation of cadmium working solutions with the benzene - acetic acid system. The efficacy of the standard additions method employed was ascertained by recovery studies on some of the petroleum crudes quoted in Table 1. This was carried out by spiking about 0.2 g of sample with known amounts of mixed-metal standards. The recoveries obtained for most of the metals were good, as there was no metal for which the recovery was less than 90%.The coefficients of variation, calculated for seven replicate determinations, were 8.4% for nickel at the 18.6 p.p.m.leve1, 7.0% for copper at the 2.94 p.p.m. level, Table 1. Average percentage recovery of trace metals in some petroleum crudes by the proposed method Recovery of trace metals from spiked crude petroleum samples, '% Sample terminal Ni c u Zn Na Pb Cd Fe Shell Forcardos . . . . . . 105.41 99.33 93.74 130.49 123.81 - 104.57 Brass Blended . . . . . . 103.83 92.48 107.93 120.36 97.28 104.76 92.56 Bonny Medium . . . . . . 97.53 102.49 99.73 103.36 108.38 108.38 94.45 Table 2. Trace metal content as determined by the mixed-solvent system and inorganic salt standards Metal content, p.p.m.Shell Element Forcardos Ni . . . . . . 12.75 f. 1.45 Cu . . . . 4.48k0.11 Zn . , . . 5.23+0.63 Na . . . . 9.21 f 1.10 Pb . . . , . . 6.10 f 0.97 Cd . . . . N.D.* Fe . . . . . . 9.05 f 0.73 * Not detectable by this method. Texaco Gulf M.V. Oloibiri Escravos 21.20 k 2.06 2.68 k 0.15 2.88 + 0.38 17.01 k 1.90 1.47 k 0.11 12.09 k 1.65 19.14 k 4.69 2.00 f 0.24 3.74 + 0.14 6.23 k 0.24 3.38 k 0.17 N.D.* 0.11 * 0.02 5.38 k 0.84 Brass Blended 15.96 1.99 1.69 k 0.24 3.38 + 0.35 3.75 k 0.41 3.94 2 0.45 1.39 k 0.27 4.50 k 0.84 Bonny Light 18.59 k 2.42 2.94 2 0.04 3.62 + 0.08 20.22 f 3.61 1.16 k 0.08 0.44 2 0.06 7.37 2 1.73 Qua Iboe T. Blended 25.28 f 1.23 3.64 f 0.37 4.32 + 0.09 20.35 k 2.18 2.10 k 0.36 3.02 f. 0.41 6.73 k 0.64 Bonny Medium 16.72 k 1.29 3.58 k 0.44 3.40 + 0.36 22.22 * 2.92 2.69 k 0.12 2.10 k 0.14 7.81 k 0.64 Table 3.Comparison of results for nickel and zinc in Nigerian petroleum crudes AAS using organometallic standards and xylene as solvent3 Sample/terminal Ni, p.p.m. Zn, p.p.m. Shell Forcardos . . . . . . . . . . . . . . 9.86 4 1.02 5.73 f 0.72 Texaco M.V. Oloibiri . . . . . . . . . . 20.38 f 1.90 2.81 2 0.32 Gulf Escravos . . . . . . , . . . . . . . 21.05 4 2.98 4.48 f 0.51 Brass Blended . , . . . , . . . . . . . . 14.65 4 2.89 3.53 k 0.19 Bonny Light . . . . . . . . . . . . . . 15.61 f 3.14 4.28 k 0.17 Bonny Medium . . , , . . . . . . . . . . 20.26 k 2.71 2.94 k 0.41 AAS using inorganic salt standards and toluene - acetic acid mixed-solvent system Ni, p .p.m. Zn, p.p.m. 12.75 k 1.45 5.23 f 0.63 21.20 k 2.06 2.88 k 0.38 19.14 f 4.69 3.74 f 0.14 15.96 k 1.99 3.39 f 0.35 18.59 k 2.42 3.62 k 0.08 16.72 k 1.29 3.40 k 0.36ANALYST, FEBRUARY 1984, VOL.109 129 7.8% for zinc at the 3.62 p.p.m. level, 11.2% for sodium at the 20.22 p.p.m. level, 9.8% for lead at the 1.16 p.p.m. level, 10.2% for cadmium at the 0.44 p.p.m. level and 13.9% for iron at the 7.37 p.p.m. level. The levels of these metals found in the petroleum crudes are shown in Table 2. Good agreement is obtained when the results obtained with six of the crudes by this method are compared with those of the established method of Means and Raclifte.3 Nickel and zinc were compared by the two methods as shown in Table 3. For both metals, the Student t-test showed that there was no significant difference between results obtained by the two methods.Conclusion The use of inorganic salt standards and a toluene - acetic acid mixed-solvent system in the determination of trace metals in petroleum crudes by atomic-absorption spectrophotometry possesses considerable advantages of time and cost over the use of organometallic standards. A precision of less than 10Y0 was obtained for most metals determined by the method. The validity of the method was shown by comparison of results obtained with those of an established method. 1 2. 3. 4. 5 . 6. 7. 8. 9. 10. 11. References Brunnock, J. V., Duckworth, D. F., and Stephens, G. G., J. Inst. Petrol., 1968, 54, 310. Karchmer, J. H., and Gunn, E. L., Anal. Chem., 1952, 24, 1733. Means, E. A., and Ratclifte, D., A t . Absorpt. Newsl., 1965,4, 174. Omang, S . , Anal. Chim. Acta, 1971, 56, 470. Chuang, F. S . , and Winefordner, J. D., Appl. Spectrosc., 1974, 28, 215. Polo-Diez, L., Hernandez-Mkndez, J., and Pedraz-Penalva, F., Analyst, 1980, 105, 37. Holding, S. T., and Matthews, P. H. D., Analyst, 1972,97,189. Holding, S. T., and Rowson, J . J . , Analyst, 1975, 100, 465. Wittmann, Z., Analyst, 1979, 104, 156. Allan, J. E., Spectrochim. Acta, 1961, 17, 467. Vogel, A. I . , “A Text-Book of Practical Organic Chemistry,” Third Edition, Longmans, London, 1968, p. 163. Paper A3127 Received Januury 26th) 1983 Accepted September 9th, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900127

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1984, VOL. 109 131 Modifications of Standard Methods for Coal Analysis for Sulphur and Nitrogen Determinations in Oil Shales James F. Alvin, David J. McCarthy and Howard J. Poynton CSIRO, Division of Mineral Engineering, P. 0. Box 3 12, Clayton, Victoria 3168, Australia Investigations were made to determine if methods that are satisfactory for determining total sulphur and total nitrogen in coals are also suitable for determinations when using compounds that contain some organic structures to be expected in oil shale kerogens. Sulphur recoveries in an excess of 98%, and typically better than 99%, were obtained using modifications of ASTM method D3177 (75). It was found that the semi-micro Kjeldahl method in the BS 1016 : Part 6 : 1977 was not suitable for the determination of nitrogen in aromatic nitrogen compounds.For example, the recovery of nitrogen from 8-hydroxyquinoline was increased from approximately 30 to 96% by increasing the digestion time to a minimum of 2 h. Variation in the initial amount of sulphuric acid in the digestion mixture between 3 and 5 ml did not alter the recovery. The use of the modified semi-micro Kjeldahl method for an oil shale sample gave a higher nitrogen recovery and better precision of results than the unmodified standard method. Keywords.: Sulphur determination; nitrogen determination; oil shale analysis; coal Oil shale reserves are a significant component of Australia’s over-all energy resources. Analytical methods of demon- strated suitability for oil shales are therefore of great impor- tance.This paper describes a preliminary investigation of the methods used for determining the sulphur and nitrogen contents of compounds of types expected in oil shales. Sulphur Determination Zinc oxide - sodium carbonate Eschka fluxes are known to be suitable for the determination of sulphur in sulphide minerals, and they are particularly suitable if free silica is present. The zinc oxide reacts with free silica to form zinc silicates, which have lower water solubilities than, for example, the mag- nesium silicates formed when magnesium oxide - sodium carbonate fluxes are used. Many oil shales contain high concentrations of free silica, e.g., typically 39% in both Julia Creek shale (Queensland, Austra1ia)z and Anvil Points shale (Colorado, USA).3 The zinc oxide - sodium carbonate fluxes are unsuitable for sulphur determinations in materials that contain either fluorine or tungsten ,I but as these elements usually occur only at very low concentrations in oil shale^,^-^ a zinc oxide - sodium carbonate flux may be as suitable €or sulphur determination in oil shales as the magnesium oxide - sodium carbonate flux recommended for coal analysis.5 It was for this reason that a preliminary investigation was made in order to determine if the zinc-based Eschka flux, which is suitable for determining sulphur in sulphide minerals, is also suitable for determining the sulphur in the organic forms that can be expected in oil shales.b Nitrogen Determination Aliphatic and aromatic nitrogen compounds, including substi- tuted pyridines and quinolines, have been identified in shale o i p and it is reasonable to assume that both types of compound exist in the parent oil shales.Therefore, any method for the determination of nitrogen in oil shale must be suitable for the determination of nitrogen in both aliphatic and aromatic compounds. Nadkarniy has recently shown that the Kjeldahl method is not generally suitable for determining the nitrogen contents of oil shales. His opinion is that this probably indicates the presence of different specific nitrogen compounds in oil shales that result from their different genesis. Bradstreetlo noted that compounds having pyridine struc- tures are “refractory,” i.e., very stable, and are difficult to determine by the Kjeldahl method. This conclusion was borne out by the results obtained from the initial experiments when only 30% of the nitrogen in 8-hydroxyquinoline was recovered using a method recommended for determining the nitrogen content of coals.5 However, as will be discussed later, the same standard method was much more satisfactory for the determination of nitrogen in diphenylamine.Superficially, the Dumas method should be a suitable alternative to the Kjeldahl method. However, it requires scrupulous control of reagent purity and combustion assembly parameters to obtain reliable results,11.’2 and is then less attractive than the Kjeldahl method as a routine method. It was therefore decided to determine whether the semi-micro Kjeldahl method5 could be modified so as to permit the satisfactory determination of nitrogen in both refractory and non-refractory nitrogen compounds.In the initial stage of his work, the semi-micro Kjeldahl method was modified to give satisfactory recoveries of nitrogen from selected organic compounds. The modified and unmodified methods were then used for determining the nitrogen content of an oil shale and the results were compared. Experimental Methods Sulphur An ASTM method for sulphur determination in coals13 was chosen as the basis for this work. The modifications to the method and the reasons for making them are as follows: A mixture (4 + 1) of zinc oxide and sodium carbonate was used as the flux for the reasons discussed earlier. The flux topping was increased from 1 g to a minimum of 2 g. The effect of varying the amount of topping is discussed below.A 100-mg mass of sodium peroxide was added to the fusion frit -hot water slurry to precipitate any manganese as manganese dioxide before filtration. The bromine water oxidation step was omitted because the sodium peroxide treatment served the same purpose. The precipitated barium sulphate was ignited to constant mass at 800 “C instead of 925 “C, as the lower temperature was found to be satisfactory.132 ANALYST, FEBRUARY 1984, VOL. 109 This method is used in the authors' laboratories for routine determinations of sulphur in ores and concentrates. Nitrogen The modifications made to the semi-micro Kjeldahl method5 that were used in all tests are as follows: (i) A 10-ml volume of 750 g 1-1 sodium hydroxide solution was used instead of the 20 ml of 400 g 1-1 sodium hydroxide solution specified in the standard, as splash- over tended to occur when the larger volume was used.The problem did not occur when the smaller volume of higher concentration solution was used. (ii) A distillation time of 25 min.14 Waanders et a1.14 have shown that the evolution of ammonia from diluted digest, after the addition of alkali, may not be complete if a distillation time of 6 min is used,5 but is complete after 20 min . (iii) A sample mass of 20-40 mg for the organonitrogen compounds. A selenium - vanadium(V) oxide catalyst was chosen for this work because of local difficulties in disposing of spent mercury-based catalyst. The variables investigated were the total digestion time and the volume of concentrated sulphuric acid in the digestion mixture.Materials and Determinations Sulphur The reagents used in the experiments were elemental sulphur, benzothiazole-2-thiol and calcium sulphate dihydrate. Sulphur was chosen as a source of volatile sulphur, as it melts at approximately 119 "C and boils at 445 "C. Benzothiazole-2- thiol was selected as a relatively stable source of organic sulphur as it contains both a thiol group and sulphur in a thiazole ring. It melts at approximately 180 "C. Many oil shales contain calcium carbonate in the form of calcite, e . g . , 40% in Julia Creek shale.* To check for interference, a calcium carbonate - benzothiazole-2-thiol mixture was also used in the experiments. Calcium sulphate may be formed in oil shale processes if the retorted shales are calcitic, contain sulphur and are combusted to provide process heat, and it was therefore included in the list of compounds. Nitrogen The three organonitrogen compounds used in these experi- ments were (i) diphenylamine (399.0% purity), which may be considered as a substituted ammonium compound and hence easily determined by Kjeldahl analysis, (ii) benzothiazole-2- thiol (nominal purity 99"/0), which contains nitrogen in a thiazole ring and was expected to be easily determined owing to the relative instability of the five-membered ring, and (iii) Table 1.Recovery of sulphur Compound CaS04.2H20 . . . . Sulphur . . . . . . Sulphur . . . . . . Benzothiazole-2-thiol Benzothiazole-2-thiol Benzothiazole-2-thiol? Benzothiazole-2-thiol? Approx. sample masslmg ., 270 . . 100 . . 100 . . 230 . . 220 . . 120 . . 110 Flux cover/ g 2 3 6 3 6 3 6 Recovery of sulphur, * Yo 99.6 97.7 99.3 96.6 99.0 95.5 98.2 * Results reported are averages of duplicate determinations. In the T 0.9 g of CaC03 was mixed with the benzothiazole-2-thiol. calculations the assumption was made that the reagents were pure. 8-hydroxyquinoline (299.5% purity), which contains nitrogen in a very stable aromatic ring. A spot sample of oil shale, denoted PC in reference 15, was also used. Results and Discussion Sulphur The results for the recovery of sulphur (Table 1) show that, for the levels of sulphur recorded, significant improvements in recovery were obtained by increasing the covering layer of zinc oxide-based Eschka flux for sulphur and benzothiazole-2- thiol.Provided the topping layer of Eschka flux was approx- imately 6 g, the method proved capable of recovering approx- imately 99% of the theoretical sulphur content of benzothiazole-2-thiol and sulphur samples. The addition of calcium carbonate to the organic compound was found to decrease the recovery of sulphur slightly. Calcium sulphate gave the highest recovery of any of the reagents used (99.6%), even with a 2-g flux covering, as this layer was not required to trap volatile (and hence highly mobile) sulphur prior to oxidation. The results also show that an ignition temperature of 800 "C was satisfactory. The sulphide-sulphur method was thus considered satisfac- tory when applied to elemental sulphur, organic sulphur compounds and calcium sulphate, and hence should be suitable for oil shales.For optimum recovery it may be necessary to adjust the thickness of the covering layer of flux mixture according to the proportion of organic sulphur in the sample. As indicated in Table 1, the amount of flux topping may need to be large if the sample contains a high concen- tration of calcium carbonate. Nitrogen The results for the recovery of nitrogen from the organonitro- gen compounds used are reported in Table 2. The effects of variation of digestion time for 8-hydroxyquinoline and diphenylamine are illustrated in Figs. 1 and 2. The results for determinations of nitrogen in the oil shale are shown in Table 3. It is clear from the results in Fig. 1 and Table 2 that the BS 1016 : Part 6 : 1977 method with a 25-min distillation time is not adequate for determining the nitrogen content of Table 2.Recovery of nitrogen in organonitrogen compounds Total Recovery of digestion No. of nitrogen,* Diphenylamine . . . . . . 0.5 7 97.9 k 3.0 Diphenylamine . . . . . . 2-4 12 97.8 k 1.3 Benzothiazole-2-thiol . . 0.5 7 -85 Benzothiazole-2-thiol . , 2-4 9 97.9 k 1.8 Compound timeih assays Yo 8-Hydroxyquinoline . . . . 0.5 3 -30 8-Hydroxyquinoline . . . . 2-4 10 96.2 -t 1.7 *Results are reported assuming the reagents were pure; where limits are given these represent the 95% confidence level. Table 3. Determination of nitrogen in oil shale Results, Oh m/m Digestion time No. of tests Mean Standard deviation 25-30 min 7 0.60 0.14 3 h 7 0.81 0.06ANALYST, FEBRUARY 1984, VOL. 109 -0 > f 03 6 0 - E ? c 50 ? L m ,E 40 133 - - 100 1 I /R / / / / / I I I I I I 0 Lc o ! l i 10 I Total digest ion t im ei h Fig.1. Effect of digestion conditions on recovery of nitrogen from 8-hydroxyquinoline. Volume of acid in digest: 0, 3.0 ml; 0, 4.0 ml; A , 4.5 ml; and 0. 5.0 ml Total digestion timeih Fig. 2. Effect of digestion conditions on recovery of nitrogen from diphenylamine. Volume of acid in digest: O,3.0 ml; 0,4.0 ml; A , 4.5 ml; and 0 , 5 . 0 ml 8-hydroxyquinoline or of benzothiazole-2-thiol. Much more acceptable results can be obtained by simply prolonging the digestion period from 30 min to 2 h. The data in Figs. 1 and 2 for 8-hydroxyquinoline and diphenylamine, respectively, show that the recoveries of nitrogen were not significantly different after digestion times of 2, 3 or 4 h.Therefore, in Table 2 the results for 2-, 3- and 4-h digestion periods for the organonitrogen compounds were grouped together. The effect of variation in the volume of concentrated sulphuric acid in the digestion vessel from 3 to 5 ml was insignificant (see Figs. 1 and 2). The best results for the recovery of nitrogen were found to be low for all materials used. The error varied from a nominal value of 2% for benzothiazole-2-thiol to nearly 4% for 8-hydroxyquinoline. The assumption of 100% purity of the reagents, as used in the calculations, no doubt contributed to a small error, particularly with benzothiazole-2-thiol; however, this could not explain errors as large as those observed here. When using long digestion times evaporation can cause the temperature to rise to levels where nitrogen dissolved in the digest as ammonium salt can be lost by oxidation.For example, Baker16 has stated that the critical temperature for the loss of nitrogen when using a selenium - vanadium(V) oxide catalyst is approximately 387 “C. However, in these experiments, even when 3 ml of acid were used, digestion temperatures did not rise above 345-350 “C, so that loss of nitrogen from the digestion mixture by oxidation was not the cause of the slightly low results listed here. In further support of this contention, as already stated, the results from 4-h digestions were not significantly different from those from the 2-h digestions. The results obtained for the oil shale show that the nitrogen recovery is improved by using long digestion times and that the precision of the results is probably improved also.However, the repeatability specified in the standard method was not achieved even with the long digestion time. Conclusions A method for determining sulphur based on the use of a zinc-based Eschka flux, which is known to be satisfactory for sulphide ores and concentrates, is also suitable for determin- ing the sulphur contents of compounds that contain sulphur in forms that can be expected in oil shales. The semi-micro Kjeldahl method for determining nitrogen was modified so that it was satisfactory for either aromatic or non-aromatic nitrogen compounds. A comparison of the results from the modified and unmodified semi-micro Kjeldahl method when applied to an oil shale showed that the recovery of nitrogen was increased by using the modified method.The precision of the results was probably improved by using the modified method. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Dolezal, J., Povondra, P., and Sulcek, Z., “Decomposition Techniques in Inorganic Analysis,” Iliffe, London, 1968, Mandelson, J., C.S.R. Ltd., personal communication, 1981. Burnham, A. K., Fuel, 1979, 58, 713. Henstridge, D. A., and Missen, D. D., “The Geology of the Narrows Gragen near Gladstone, Queensland, Australia,” Southern Pacific Petroleum NL and Central Pacific Minerals NL, Sydney, 1981. BS 1016 : Part 6 : 1977. British Standards Institution, London. Smith, J. W., Young, N. B., and Lawlor, D. L.. Anal. Chem., 1964, 36, 618. Regtop, R. A., Crisp, P. T., and Ellis, J., Fuel, 1982, 61, 185. Uden, P. C., Carpenter, A. P., Jr., Hackett, H. M., Hender- son, D. E., and Siggia, S., Anal. Chem., 1979, 51, 38. Nadkarni, R. A., Am. Chem. SOC., Div. Fuel Chern., Prepr., 1983, 28 ( 3 ) , 200. Bradstreet, R. B., “The Kjeldahl Method for Organic Nitro- gen,” Academic Press, New York, 1965. Gustin, G. M., in Kolthoff, I. M., and Elving, P. J., Editors, “Treatise on Analytical Chemistry,” Part 11, Volume 11, Interscience, New York, 1965, pp. 40S415. Drew, H. D., in Streuli, C. A., and Averell, P. R., Editors, “Chemical Analysis,” Volume 28, Part I, Wiley-Interscience, New York, 1970, Chapter 1. ASTM D3177 ( 7 9 , “Total Sulphur in the Analysis Sample of Coal and Coke,” American Society for Testing and Materials, Philadelphia, 1975. Waanders, J., Wall, T. F., and Roberts, J., Chem. Aust., 1980, 47, 274. McCarthy, D. J., Fuel, 1983, 62, 1283. Baker, P. R. W., Talanta, 1961, 8, 57. pp. 167-173. Paper A311 45 Received May 20th, 1983 Accepted September 13th, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900131

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1984, VOL. 109 135 Diff erential-pulse Polarographic Determination of Trace Heavy Elements in Coal Samples Guler Somer* and Osman Cakir Department of Chemistry, Hacettepe University, Ankara, Turkey Ali Osman Solak Department of Pharmacy, Hacettepe University, Ankara, Turkey The conditions for the differential-pulse polarographic determination of trace heavy elements in a coal have been investigated. By means of polarograms taken at two different pHs of a solution of 0.5 M EDTA + 0.5 M sodium acetate, it has been shown that Fe, Cu, Pb, Cd, Ti, Sb, Mo, As and Zn could be determined in coal, a complex matrix, within a reasonably short time. Keywords: Trace heavy element determination; differential-pulse polarograph y; coal analysis Pulse polarography has been used in such diverse research fields as biological, metallurgical, sea water, minerals and pharmaceutical analysis.Although it is suitable for the analysis of single elements, it can also be used for the simultaneous determination of a number of elements in a sample.'-6 The range of its applications has increased since 1960 with the introduction of single-sweep, square-wave and a. c. polarography . In this study, polarography has been used for the determina- tion of elements in coal samples. Although for such complex matrices atomic-absorption spectrometry, neutron-activation analysis and X-ray fluorescence are usually considered as standard methods for the determination of trace elements,7?8 it has been found that polarography can also be used with success.However, if the Epcak values of certain elements are very near to each other or identical, interferences occur. Nevertheless, the possibility of determining a set of elements simultaneously certainly shortens the analysis time required, which is an advantage. Overlapping peaks in one electrolyte can be separated by changing either the electrolyte or the pH. By choosing an electrolyte that is convenient to use at different pH values, it is possible to determine many elements by measuring only two polarograms. In this work the determination of elements in coal was investigated. The most suitable electrolyte and pH values at which the elements gave distinct Ep& values were estab- lished. Experimental Apparatus A pulse polarographic system of standard design and a normal polarograph, both built in the Physics Department at Hacet- tepe University, together with a Tektronix 510 N,D-lS oscilloscope, were used.The normal polarograph was similar to a Heath Model EUW-198 with IC operational amplifiers in place of the older vacuum tube types. It had an additional ramp output section that was connected to the pulse polaro- graph. The system of the pulse polarograph was as follows. The potential applied to the electrodes in the system was obtained from a circuit consisting of an external ramp of a normal polarograph from a unit that produced an initial d.c. potential, from a reed relay and a potential adder circuit. The reference potential was divided by a Zener from a regulated power supply of f 1.5 V. It could be accurately adjusted by means of a * To whom correspondence should be addressed.ten-circuit ( k 3 V maximum) helipot. Potential pulses were obtained by opening or closing the reed relay at convenient times. The adder, consisting of a 741 OpA (operational ampli- fier), served to apply to the electrodes the pulses that were superimposed on the external ramp. Two current samples were taken, one just before the pulse had been applied (t') and the other just before the drop had fallen (T). A difference amplifier and a sample and hold circuit were used to plot i(~)-i(t') versus base potential. Observation of the oscillo- scopic trace ensured that the pulses were being applied just before the mercury drops fell. The ramp and pulse were observed on the screen of this memory oscilloscope.A three-electrode system with a saturated calomel electrode as reference was used. The dropping-mercury electrode was part of a Heath polarography system (Model EUW-198) and the drop time was about 2-3 s. To prevent the capillary from plugging, it was stored in distilled water when not in use. The polarograms were recorded with a Heath - Schlumberger SR-2.55B recorder. Reagents The electrolytes and acids used were obtained from Merck (pro analysi grade). The standard solutions of elements were prepared with analytical-reagent grade salts. The water used was distilled approximately eight times in a specially designed packed distillation column. Its conductivity was less than 1 pmho. The mercury used in the dropping-mercury electrode was obtained from Merck (pro analysi).The contaminated mer- cury was cleaned by passing it successively through dilute nitric acid and water columns in the form of fine droplets. The collected mercury was dried between filter-papers. A polaro- gram of this mercury was taken before use to ensure the absence of impurities. Procedure A coal sample (25 mg) was digested in a Parr acid digestion bomb with 1-2 ml of acid mixture (HC104 - HN03 - H2S04) (24 + 24 + 1) by heating for 2 h at 150 "C. After digestion, most of the acids were fumed, 10-15 ml of electrolyte were added, the pH was adjusted to the desired value and nitrogen was passed through the cell for 30 min. The ramp and reference voltages were adjusted by observ- ing them on the oscilloscope screen, so that the pulse would be applied just before the release of the mercury drops.Another advantage of following the ramp and pulse on the oscilloscope was that a very small peak on the recorder could be136 ANALYST, FEBRUARY 1984, VOL. 109 distinguished from the noise. The applied pulse had an amplitude of 50 mV and the pulse duration could be chosen as 40 or 80 ms, The potential sweep was adjusted between the limits of 0 and -0.8 V or 0 and -1.2 V, depending on the pH of the solution. The recorder rate was adjusted to 1 in min-1 and the potential sweep rate was 0.2 V min-1. A 50-500 pl volume of the standard solution (10-3 M) of one of the elements found in coal was added to the electrolyte under test and its peak was recorded. By repeating the same procedure for the appropriate elements and comparing the clarity of polarograms obtained, the most suitable electrolyte for identification of these elements was chosen. This elec- trolyte was added to the solution of the coal sample and its polarogram was taken.By using the standard additions method the elements were determined both qualitatively and quantitatively. Results and Discussion Although combustion is used most frequently for the digestion of coal samples, partial loss of elements during combustion makes the method unfavourable. The approximate losses at Table 1. Approximate peak potentials for a coal sample using 0.05 M EDTA + 0.5 M sodium acetate electrolyte Fe3+ CU2+ Ti4 + Pb2+ Bi3+ Sb3+ Mo6+ As3+ Cr6+ co2+ Ni2+ Cd2+ Zn . . Element E N vs. S.C.E. . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . pH 5.5 -0.10 -0.25 -0.40 -0.55 -0.67 -0.52 -0.85 No peak No peak No peak No peak No peak -0.93 pH 2.5 -0.05 -0.10 -0.22 -0.38 -0.60 -0.65 -0.53 No peak No peak No peak No peak -0.85 -0.95 0 0.2 0.4 0.6 0.8 1.0 -EN VS. S.C.E. Fig. 1. Seyitomer (pH 5.5) Differential-pulse polarogram of a lignite coal sample from 550 "C are 54% for Pb and Sb, 61% for Zn and 47% for Cd.9 Therefore, acid digestion was used in this study. It was found that an acid mixture consisting of HC104 - HN03 - H2S04 (24 + 24 + 1) was the most favourable mixture for digestion. The coal sample could be digested in a Parr acid digestion bomb at 150 "C in 1-2 h.In an attempt to determine the most suitable electrolyte for a coal sample that was later analysed to contain the elements listed in Table 2, KC1, HC1, HC104, NaN03 and EDTA + sodium acetate solutions were individually tested. Of these, the most suitable electrolyte was 0.05 M EDTA + 0.5 M sodium acetate. One advantage of this electrolyte was that it allowed polarograms to be obtained at different pH values. They differed in the degree of separation of over- lapping peaks, thus allowing the best pH to be chosen. It should also be noted that the peaks were shifted by the variation in pH. This provided valuable information, as polarograms of known elements could be obtained at the same pH values and from the extent of the shift the presence of certain elements in the unknown sample could be ascertained.In quantitative determinations, after taking the polarogram of the sample at a certain pH, the pH of the sample was changed and a new polarogram taken. The concentrations of trace elements obtained at each pH level had to be checked closely before the results were considered to be accurate. Table 2. Differential-pulse polarographic (DPP) analysis of two lignite coal samples Amount k standard deviation, p.p.m. Seyitomer cu2+ . . . . . . l l O k 2 0 Ti4+ . . . . . . 980k90 Mo6+ . . . . . . 3 2 0 f 4 5 Sb3+ . . . . . . 125k30 Cd2+ . . . . . . 40k11 As3+ . . . . . . 105 k 2 5 Fe3+ . . . . . . 3300f 120 Element lignite (DPP) ASkale lignite DPP AAS 185 k 25 705 k 55 - 87 * 22 - 220 * 35 - 95+18 100k 15 55 k 13 - 1210 + 70 - 160 f 30 0 ~~~ 0.2 0.4 0.6 0.8 1.0 -EN VS.S.C.E. Fig. 2. Seyitomer (pH 2.5) Differential-pulse polarogram of a lignite coal sample fromANALYST, FEBRUARY 1984, VOL. 109 137 The peak potentials in coal samples using EDTA and sodium acetate as the electrolyte at pH 2 and 5.5 are given in Table 1. After the peak potentials of individual elements had been determined in coal samples, the polarograms could be analysed for the elements. Figs. 1 and 2 show polarograms of a coal solution at two different pH values. To assign a peak to a certain element, a standard solution of this element had to be added; an appropriate increase in the peak height confirmed the assignment. If the presence of one element was doubtful, the pH was changed and the newly formed peak investigated. If the peak appeared at the expected potential, and if the amount calculated was the same as that found at the previous pH, the identification was confirmed. Fig.1 (pH 5.5) shows a peak that is off-scale, followed by peaks at -0.23, -0.36, -0.50, -0.63, -0.80 and -0.92 V. When the sensitivity was decreased 100-fold the off-scale peak became measurable and, by addition of an iron(II1) salt, it was identified as an iron peak. The peak at -0.23 V, according to Table 1, had to be copper, as there was no other peak at that potential in this medium. At pH 2, however, the copper peak moved to more positive potentials (Fig. 2) and thus over- lapped with the off-scale iron peak. At this pH the peaks were at -0.22, -0.52, -0.65 and -0.86 V. The element that gave peaks at -0.36 V (pH 5.5) and -0.22 V (pH 2.5) was titanium, as on the addition of a standard solution of titanium the peak heights increased proportionally and the amounts of titanium calculated at the two pH values were the same.The presence of molybdenum and antimony in coal was confirmed by the same procedure. The polarogram at pH 2.5 had a small peak at about -0.85 V that moved to more negative potentials with increasing pH. Also, as the addition of cadmium salt increased the peak height, this peak was assigned to cadmium. The peak at -0.90 V (pH 5.5) was similarly shown to be that of arsenic. In order to determine the amount of impurities that may have been introduced by the reagents, water and glassware, a blank test was carried out. No peak was observed when a polarogram was taken at the same sensitivity as for the coal samples.Two different lignite samples, one from the Seyitomer and the other from the ASkale mines in Turkey, were analysed using the above procedure and the results are given in Table 2. Standard deviations were calculated in 5-8 samples. In the analysis of lignite coal samples, at least two standard additions were made for each element. To check the validity of the method, one coal sample was analysed by both AAS and the proposed method. As only Cd and Cu lamps were available, these ions could be determined. Results are given in Table 2, and the agreement is considered to be satisfactory. In conclusion, it may be stated that it is possible to analyse a coal sample for heavy metals, both qualitatively and quantita- tively, by taking polarograms at two different pH values. References 1. 2. 3. 4. 5 . 6. 7. 8. 9. Bhatnagar, R. M., Singh, B. N., and Roy, A. K . , Technology (Coimbatore, India), 1971, 8, 29. Cornell, D. G., and Pallansch, M. J., J . Dairy Sci., 1973, 56, 1479. Gilbert, D. D., Anal. Chem., 1965, 37, 1102. Myers, D. J., and Osteryoung, J., Anal. Chem., 1973,45,267. Necker, H. N., Fresenius Z . Anal. Chem., 1972, 261, 29. Pilkington, E. S . , Weeks, C., and Bond, A. M . , Anal. Chem., 1976, 48, 1665. Hansen, L. D., and Fisher, G. L., Environ. Sci. Technol., 1980, 14, 1111. Kingston, H., and Pella, P. A., Anal. Chem., 1981, 53, 223. Hwang, J . Y., Anal. Chem., 1972, 44, 20. Paper A31125 Received May 6th, 1983 Accepted September 26th, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900135

出版商:RSC    

年代:1984

数据来源: RSC

摘要:

ANALYST, FEBRUARY 1984, VOL. 109 139 Anodic Voltammetry of Butyrophenone Neuroleptics E. Bishop and W. Hussein" University of Exeter, Chemistry Department, Stocker Road, Exeter, EX4 4QD, UK Example drugs have been examined at platinum and gold rotating disc electrodes in 0.1 mol 1-1 sulphuric acid, and for spiperone over the pH range 0-7. Aceperone and haloperidol show neither anodic nor cathodic activity. Benperidol and droperidol are anodically active, but the voltammograms do not obey the Levich relationship, and adsorption destroys electrode activity. Spiperone gives a two-electron wave for oxidation of the diazo ring; electrode kinetic parameters have been evaluated. Although relative standard deviations of less than 1% are obtainable in the determination of spiperone, adsorption is severe and caution is required.Keywords: Butyrophenone neuroleptics; rotating disc electrode voltammetry; oxidation mechanism; electrode kin e tic para meters Butyrophenones are widely used major tranquillisers of recent development. These long acting neuroleptics are particularly useful in the prevention of psychotic relapse in schizophrenic patients after discharge on remission of symptoms. Peak plasma levels occur in 2-6 h and may attain a plateau for 72 h and persist detectably for several weeks. Reduction at a dropping-mercury electrode has been reported.l,2 Coulo- metric determination of the number of electrons in alkaline media gave the value as two and the half-wave potential is pH-dependent . Determinations have been reported by GC,3-6 TLC,7-' fluorimetrylo and radioimmunoassay.11 The anodic behaviour of the compounds listed in Table 1 at rotating disc platinum and gold electrodes in 0.1 moll-' sulphuric acid and for spiperone in buffer media is described in this paper. Table 1.Butyrophenones examined. All compounds were supplied by Janssen Pharmaceuticals Ltd. C.A. number Generic name Structure* Batch No. 807-31-8 2062-84-2 5 4 8 - 7 3 - 2 5 2- 86- 8 % CHz N/C-cH3 Ace per one H 3248 Benperidol Droperidol Haloperidol M R- A1911 A7801 DO301 Proprietary name Ace per one Anquil Droleptan Haldol 749-02-0 Spiperone .-N(-J-JH I A3301 Spiperol 0 w * Present address: Department of Pharrnaceutics, Faculty of Pharmacy, University of Karachi. Karachi-32, Pakistan.ANALYST, FEBRUARY 1984, VOL. 109 140 0 0.2 0.4 Q: E L m 0.6 I 0.8 1 .o 1.2 0.8 1 .o 1.2 0 0.4 Q 5 0.8 -!! I 1.2 1.6 0 0.4 0.8 1.4 1.6 1 .o 1.2 0 0.4 0.8 1.2 0.8 1.0 1.2 1.4 1.6 0.8 1.2 1.2 1.4 1.6 0.8 1.0 1.2 1.4 1.6 E N vs.S.C.E. 1.4 1.6 1.8 1 (fl 0.8 1.2 2.0 1.6 L 0.8 1.0 1.2 1.4 1.6 1.8 Fig. 1. Rotating disc electrode voltammograms of 0.002 mol 1-1 butyrophenones in 0.1 mol 1-1 sulphuric acid: (a) and ( b ) spiperone, (c) and ( d ) benperidol and ( e ) and cf) droperidol, at (a, c and e) platinum and (b, d and fl gold electrodes. Electrode area, 0.503 cm3; temperature, 25 "C. Curves 1-5, nominal rotation speeds 10,20,30,40 and 50 Hz, respectively; offset curves (B) represent a re-scan at 50 Hz without intervening activation of the electrode 0.8 1.2 . 1.6 Lu x 5 0.8 v) 1.2 1.6 PH 7 p H 5 p H 4 p H 3 pH2 p H 1 1 MH2SO4 Fig.2. B, re-scan without intervening activation. Concentration, (pH = 0, 1) or citrate - phosphate buffer adjusted to exact unit pH value Influence of pH on the anodic voltammograms of spiperone at (a) platinum and ( b ) gold electrodes. Curves A, activated electrode; mol 1-I; rotation speed, 50 Hz; temperature, 25 "C; media, sulphuric acidANALYST, FEBRUARY 1984, VOL. 109 141 Table 2. Electrode kinetic parameters for spiperone. Rotation speed, 50 Hz; electrode area, 0.503 cm2; medium 0.1 mol 1 - 1 sulphuric acid; reference potential, E, at 25 "C Platinum electrode Gold electrode krnl kl kml kl cRw 0 - 3 Ep'V 10-6 1 10-6 I E,IV 10 -6 1 10-6 1 moll-' vs. S.C.E. P vs. S.C.E. P cm-2 s-I cm-2 s-1 (3111-2 s-1 cm-2 s-1 1.999 87 0.955 0.35 5.37 5.60 0.940 0.41 5.77 6.43 4.999 69 0.960 0.32 5.05 5.55 0.945 0.39 5.45 6.12 7.999 51 0.965 0.34 4.71 4.78 0.950 0.37 4.97 5.89 9.999 39 0.965 0.34 4.32 4.25 0.955 0.40 4.78 5.65 Experimental Apparatus, procedures, electrode activation and solution manipulation have been described.12 The samples, which were of Drug Standard grade, were supplied by Jenssen Pharmaceuticals Ltd. Results and Discussion Spiperone at platinum electrodes in 0.1 moll-' sulphuric acid gave well formed waves as broad peaks followed by troughs [Fig. l(a)], which virtually disappeared when an unactivated electrode was used. The peak potential moved to more positive values as the rotation speed increased. The behaviour was similar but more marked at gold [Fig. l ( b ) ] , and the peak potential became less positive as rotation speed increased; there was also a diminishing after-wave.The Levich plots of peak current against square root of frequency are poor (the RSD of the slope is 4% at platinum electrodes and 8% at gold electrodes), suffering a change of slope rather than curvature and having non-zero intercepts. The peak current versus concentration plots, however, are exceptionally good, espe- cially at gold, with excellent zero intercepts and RSDs <0.4%. The voltammograms and frequency graphs are characteristic of adsorption, and suppression of the waves at unactivated electrodes identifies the adsorbate as the oxidation product. The influence of pH on the voltammetry of more dilute spiperone is illustrated in Fig. 2. At platinum, a single good wave is formed at all pH values, with a maximum height in 0.1 moll-' sulphuric acid, but the wave dwindles and vanishes at an unactivated electrode as the pH increases.At an activated gold electrode three waves, the second ill defined, in sulphuric acid become two at pH 2-4, one at pH 5-6 and two peaks at pH 7. The first wave attains maximum height in 0.1 mol 1-1 sulphuric acid. Again, increasing adsorption gradually des- troys the wave as the pH increases if an unactivated electrode is used. Benperidol gave good waves at platinum at low rotation speeds, but peaking, overlapping and deviation from Levich dependence occurred at higher rotation speeds, as can be seen in Fig. l(c). Oxidation did occur at gold, Fig. l ( d ) , but the waves showed little dependence on rotation speed.The behaviour of droperidol, Fig. l ( e ) and 0, was similar. For both compounds, a repeat scan at 50 Hz without intervening activation of the electrode showed virtual suppression of the waves, as the offset curves in Fig. 1 demonstrate. This indicates that adsorption of oxidation product generates deviation and malfunction and that this is aggravated by high concentrations and rotation speeds. Determination of the number of electrons involved in the oxidation of spiperone by amperostatic coulometryl2 gave a value of 2.0. Aceperone and haloperidol are anodically inactive, which eliminates the butyrophenone and piperidine fragments, leaving the benzamidazol fragment in benperidol and droperidol and the diazo ring in spiperone as the sites for oxidation.In spiperone, the carbonyl group further fixes the reaction site as shown. 0 0 I C6H5 I C6H5 Here R is the 4-fluorophenyl-4-oxobutyl group, which, by elimination, can be suggested as the adsorptive site. Electrode kinetic parameters have been derived for spiperone by pattern theory,13 and are listed in Table 2. The dominance of adsorption of reaction product on the electrode surface prevents voltammetry from being a useful analytical method for the butyrophenones examined other than perhaps spiperone. The square root of frequency graphs raise no great expectations, but the spiperone calibration graphs are unusually good, and sets of four measurements on each of five Drug Standard solutions in 0.1 mol 1-1 sulphuric acid at activated platinum gave relative standard deviations of 0.53, 0.69, 0.89, 0.20 and 0.34%, while at activated gold the results were 0.57, 0.37, 0.32, 0.20 and 0.54%, in the concentration range 10-3-10-2 mol 1-1.Nevertheless, the potential interference presented by adsorption counsels cau- tion in the application. We thank Janssen Pharmaceuticals Ltd., for the gift of materials listed in Table 1 , and the Royal Society for the SEL transfer standard DVM. W. H. thanks the Government of Pakistan for the award of a Scholarship and the University of Karachi for the grant of leave of absence. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. References Vire, J. C., Fischer, M., and Patriarche, G. J., Talanta, 1981, 28, 313. Mikolajek, A., Krzyzanowska, A., and Fidelus, J., 2. Anal. Chem., 1974, 272, 39. Bianchetti, G., and Morselli, P. L., J. Chromatogr., 1978, 153, 203. Forsman, A., Martensson, E . , and Ohman, R., Naunyn- Schmiedebergs Arch. Pharmacol.) 1974, 286, 113. Moulin, M. A., Camsonne, R., Davy, J. P., and Bigot, M. C., J. Chromatogr., 1979, 178, 324. Pierce, W. O., Lamoreaux, T. C., Urvy, F. M., and Kopjak, L., J. Anal. Toxicol., 1978, 2, 26. Pluym, A., J . Pharm. Sci., 1978, 68, 1050. Hulshoff, A., and Perrin, J. H., J . Chromatogr., 1976, 129, 249. Vinson, J. A., and Hooyman, J. E., J . Chromatogr., 1975,105, 416. Baeyens, W . , Moerloose, P. D. E., and Taeye, L. D. E., J . Pharm. Sci., 1977, 66, 1787. Clark, B. R., Tower, B. B., and Rubin, R. T., Life Sci., 1977, 20, 319. Bishop, E., and Hussein, W., Analyst, 1984, 109, in the press. Bishop, E., Analyst, 1972, 97, 761. Paper A31279 Received August 22nd, 1983 Accepted September 13th, 1983

ISSN:0003-2654    

DOI:10.1039/AN9840900139

出版商:RSC    

年代:1984

数据来源: RSC