ISSN:0003-2654    

DOI:10.1039/AN98106FX001

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

ANALAO 106 (1 258) 1-1 28 (1 981 ) January 1981ISSN 0003-2654THE ANALYSTTHE ANALYTICAL JOURNAL OF THE ROYAL SOCIETY OF CHEMISTRY123324047546069768594100CONTENTSREVIEW. An Appraisal o f Mutagenicity Test Systems-Diana Anderson andE. LongstaffDetermination o f Lead in Drinking Water by Atomic-absorption Spectrophoto-metry with Electrothermal Atomisation-M. P. Bertenshaw, D. Gelsthorpe andK. C. WheatstoneLaser Ablation for the Introduction o f Solid Samples into an Inductively CoupledPlasma f o r Atomic-emission Spectrometry-Michael Thompson, John E.Gouiter and Friedrich SieperAtomic-absorption Spectrophotometric Determination of Trace Amounts ofArsenic in Acrylic Fibres Containing Antimony Oxide with Solvent Extrac-tion and Arsine Generation-Takashi KorenagaSimultaneous Determination o f Copper, Nickel and Iron in Heavy Alloy UsingAtomic-absorption Spectrophotometry-Mangala A.Pabalkar, S. V. Naik andN. R. SanjanaNebuliser System for Analysis of High Salt Content Solutions with an Induc-tively Coupled Plasma-Bo ThelinSpectrophotometric and Fluorimetric Determination o f Tri- and Di-organotinand -organolead Compounds Using Dithizone and 3-Hydroxyflavone-W. N. Aldridge and B. W. StreetAmorphous Surface and Quantitative X-ray Powder Diffractometry-StephenAltree-Williams, John G. Byrnes and Bernard JordanBromate Oxidation o f Methyl Orange-R. A. Hasty, F. J. Lima and J. M. OttawayDetermination of Atmospheric Isocyanate Concentrations by High-performanceThin-layer Chromatography Using 1 -(2-Pyridyl)piperazine Reagent-P.A.Ellwood, H. L. Hardy and R. F. WalkerApplication o f an Improved Steam Distillation Procedure in Residue Analysis-Akio Tanaka, Norihide Nose, Akiko Hirose and Akinobu WatanabeDetermination o f Fluocinolone Acetonide i n Pharmaceutical Preparations byDifferential-pulse Polarography-C. P. Leung and S. Y. K. TamREPORT BY THE ANALYTICAL METHODS COMMITTEE105 Determination o f Halquinol (Chloroquinol-8-01s) in Pre-mixes and MedicatedFeeding StuffsSHORT PAPERS114 A Radiochemical Modification o f the Berthelot Reaction for the Determinationo f Ammonia-Swee-Eng Aw11 7 Spectrophotometric Determination of L-Ascorbic Acid in Vegetables and Fruits-K. L. Bajaj and Gurdeep Kaur120 A Discrete Plasma Cholinesterase Assay Adapted for Batch Analysis-J.D.Pryce122 Use of Stop-flow Ultraviolet Scanning and Variable-wavelength Detection forEnhanced Peak Identification and Sensitivity in High-performance LiquidChromatography-James W. Readman, Leslie Brown and Michael M. Rhead127 BOOK REVIEWSSummaries of Papers in this Issue-Pages iv, vi, vii, x, xi, xii, xiii, xivPrinted by Heffers Printers Ltd Cambridge EnglandEntered as Second Class at New York, USA, Post OfficDEBEERS I bINDUSTRIAL DIAMOND __ - AK DIVISION r Chemical Analystsfor research and developmentinto new instrumental analyticaltechniques in South AfricaAs you’re xrobably aware, SouthAfrica is ri ing the crest of a long-running boom. Our industries andeconomy are thrivin andexpansion i s the order of t l e day.We‘d like you to come and shareyour expertise with us, at the sametime sharing in everything SouthAfrica has to offer. N o essays onclimate, high standard of living, easytaxation and great lifestyle arenecessary.Those are legend already.Come and experience it with us.De Beers Diamond ResearchLaboratory, situated in Johannes-burg, is engaged in research into themanufacture of synthetic diamondsand other superhard materials. It isthe largest and most advanced of itskind in the world.We now seek Chemical Analystswith an R & D background for thedevelopment of analytical tech-niques on specialitv materials. You4projects within limited time scales.If you havea B.Sc. in Chemistry oran HNUHND coupled with at least 3years’ experience in modernanalytical techniques and have aninterest in R & D, then you could beideal for one of these positions.To attract the right people we willay a well above average salary and E enefits will include a guaranteed10% annual bonus, 21 workin days’leave, free dining facilities anfgoodmedical aid and pension schemes.We will of course pay air fares forr u and your immediate family,otel accommodation for onemonth, as well as giving yougenerous baggage and settling-inallowances.In the first instance applicantsshould telephone Mr.S. A. A. Bryantfor an application form at (01)353-1545 or write to him at AngloCharter International Services Ltd,wiil be deieloping today’s mostmodern analytical techniques tosolve high technology problems andwill be expected to work inde- career details quoting ref. S A ‘I (Appointments Division), 40Holborn Viaduct, London E.C.l .P.IAJ, enclosing full personal andpendantly and ( :omplete research 82/80.-A236 for further iiiformation. See page

ISSN:0003-2654    

DOI:10.1039/AN98106BX003

出版商:RSC    

年代:1981

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98106BP007

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, January, 1981, Vol. 106, pp. 23-31 23 Determination of Lead in Drinking Water by Atomic-absorption Spectrophotometry with Electrothermal Atomisation M. P. Bettenshaw" and D. Gelsthorpe and K. C. Wheatstone Directorate of Scienti$c Services, Severn-Trent Water Authority, Nottingham Regional Laboratory, Meadow Lane, Nottingham, NG2 3HN Directorate of Scientific Services, Severn-Trent Water Authority, Abelson House, 2297 Coventry Road, Sheldon, Birmingham, B26 3PU The use of electrothermal atomisation to determine concentrations of lead in water, although sensitive, generally suffers from suppressive interference effects that can produce large and variable negative bias. This paper describes a procedure that employs a lanthanum pre-treatment technique, which when tested with a range of 14 drinking water samples is shown to overcome such interference. The pre-treatment technique can be either impregnation of the furnace tube with lanthanum or the addition of lanthanum (as lanthanum chloride) to each sample.The procedure has a limit of detection of less than 1.0 pg l-l, a total standard deviation of less than 1.5 pg 1-l, or 5% of the concentration, and a bias of less than 5 pgl-' or 10% of the concentration over a working range of 0-100 pg 1-l for both manual and automated injection of samples. Keywords : Lead determination ; drinking water analysis ; atomic-absorption spectrophotometry ; electrothermal atomisation ; lanthanum treatment The analysis of environmental samples for lead has become increasingly important in recent years due to the concern over the potential health hazards of the intake of small amounts of lead ingested with food and water and inhaled from the air.l I t has been calculated that drinking water containing 5Opg1-1 of lead would contribute between 25 and 43% of the blood lead of an adult.2 The World Health Organization3 recommends a general limit for lead in drinking water of 100 pg l-l, or a maximum of 300 pg 1-1 after 16 h of contact with the pipes.A more recent draft proposal from the EEC4 is expected to give rise to a directive specifying a general maximum of 50 pg l-l, although for households with lead pipes there is a recommendation that the concentration should neither frequently, nor by an excessive amount, exceed 100 pg 1-l. The proposal by the EEC to reduce the limit caused the Depart- ment of the Environment to initiate a national survey of lead in drinking water.5 The increasing interest in the concentration of lead in drinking water has considerably increased the number of samples and the frequency of the determination.However, the conventional method for the determination of lead by direct flame atomic-absorption spectro- photometry, although simple and rapid, is not sufficiently sensitive for this determination, having a limit of detection in the region of 30-50 pg 1-l. Although pre-concentration techniques may be employed to improve this limit, they have the disadvantage of considerably increasing the time and/or complexity of the determination. The determination of lead in drinking water by atomic-absorption spectrophotometry using electrothermal atomisation is potentially advantageous, as it requires minimum sample pre-treatment, needs only small volumes of sample and the analysis time is only a few minutes per sample.The disadvantage of the technique, particularly for the determination of lead in water, is that inorganic matrix constituents of the sample interfere with the atomisation of lead in the sample causing suppression of the result. Themagnitude of the suppression appears to vary from sample to sample and with different designs of carbon furnace, but can be as large as 80% in some in~tances.~.~ Consequently, much of the work in producing a method for the determination of lead in water has concentrated on procedures Laboratory, 141 Church Street, Malvern, Worcestershire, WR14 2AA.* Present address : Directorate of Scientific Services, Severn-Trent Water Authority, Malvern Regional24 BERTENSHAW e8 at?.: LEAD IN DRINKING WATER BY Analyst, vol. 106 for overcoming the suppressive interference effects. The method of standard additions may be employed to minimise the interference effects, but this significantly increases the analysis time and may not be completely free from bias.8 Several methods employing chemical releasing agents to overcome the suppressive interference effects have been reported, which allow the direct determination of lead and thus retain the advantage of speed of analysis; such methods involve the use of EDTA,9 ascorbic acid,6J0 orthophosphoric acid,ll thiourea12 and lanthanum nitrate.' The last method, in which lanthanum salts can be added to samples or, alternatively, the carbon furnace tube can be impregnated with lanthanum, was selected for detailed evaluation using a wide range of potable water samples and using a different electrothermal atomiser system to that reported earlier.' Experimental Reagents Nitric acid, 70% m/ V , atomic-spectroscopy grade.Lanthanum chloride (LaC1,.7H20), analytical-reagent grade. Lanthanum nitrate [La(NO,) ,.6H20], analytical-reagent grade. Apparatus Absorption measurements were made using an Instrumentation Laboratory IL157 single- beam spectrophotometer fitted with deuterium arc background correction and a Juniper lead hollow-cathode lamp. A Linseis LS4 flat-bed recorder was used to monitor output from the spectrophotometer.The electrothermal atomiser used was an Instrumentation Laboratory IL555 controlled- temperature furnace that was fitted with a single-piece, non-pyrolytically coated furnace tube (Cat. No. 42411). For manual operation, 20 pl of sample were introduced into the furnace by means of a Finnpipette microlitre syringe and for automated operation the sample was introduced by means of an Instrumentation Laboratory IL254 Fastac automatic sampler. The automatic sampler was fitted with disposable polystyrene sample cups of 4 ml capacity (e.g., Technicon AAII or equivalent). For low-level lead determination (below 1 pg 1-l) acid-washed poly- thene sample cups are recommended. Nitrogen was used as purge gas. Procedure background correction, and the wavelength used was 217.0 nm.The spectrophotometer settings were as recommended by the manufacturer for simultaneous The furnace programme was as shown in Table I. TABLE I FURNACE PROGRAMME Stage Drying cycle . . .. Manual operation Ambient to 90 "C: 5 s 90 "C to 110 "C: 45 s Automated operation 175 "C: 5 s Ashing cycle . . . . 110 "C to 350 O C : 5 s 175 "C to 350 "C: 5 s 360 "C to 650 "C: 25 s 350 "C to 650 "C: 25 s Atomisation . . . . 650 "C to 2200 "C: 5 s (ramp) Cleaning cycle . . 650 "C to 2300 "C: 5 s (ramp) Hold at 2300 "C: 5 s , . 2200 "C to 2500 "C: 5 s A 5-s sample deposit time on the autosampler was found to give comparable absorbance to a 20-p1 sample volume manually injected and under these conditions, a calibration graph over the range 0-100 pg 1-l was obtained (Fig.1). All samples tested were collected and stored in polythene containers and were acidified by the addition of 10 ml of nitric acid per litre. Blanks and standard solutions were prepared in polythene calibrated flasks and contained 1 ml of nitric acid per 100 ml.January, 1981 AAS WITH ELECTROTHERMAL ATOMISATION 25 400 o 300 2 % c 200 5 2 a 100 X 13 0 20 40 60 80 100 Leadlpg I-' Fig. 1. Lead calibration graph: A, using a 5-s deposit time with IL254 Fastac; and B, using a 20-4 injection with a Finnpipette microlitre syringe. (For instrumental opera- ting conditions see text.) Where addition of lanthanum to samples was used, the samples were adjusted to contain 0.1% m/V of lanthanum (as lanthanum chloride) for manual injection and O . O l ~ o m/V of lanthanum (as lanthanum chloride) for automated injection. New furnace tubes were pre- conditioned by treatment with five replicate injections of solutions containing 0.5% m/V of lanthanum (as lanthanum chloride). Where tubes were impregnated with lanthanum, this was achieved by immersing the tube in the minimum volume of a concentrated solution of lanthanum nitrate and leaving in a vacuum desiccator until all the solution had been absorbed. Furnace tubes prepared in this way were conditioned by operating the furnace programme in Table I a few times, without sample injection.Results and Discussion In order to assess objectively the performance of the electrothermal atomisation technique, the recommendations of the Water Research Centre (WKC) for performance characteristics of a method for the determination of lead in potable waters8 were adopted.These were as follows : Concentration range .. .. . . 0-100 pg 1-1 Total standard deviation . . .. . . Not more than 1.5 pg 1-l or 5% of the Bias of analytical results . . . . . . Not more than 5pg1-1 or 10% of the Limit of detection . . .. .. . . Not greater than 5 pg 1-l. Standard deviations (within batch) obtained from ten single randomised replicate injections of 1.0, 5.0, 10.0, 20.0 and 40.0 pg 1-1 lead standard solutions under the operating conditions described in Table I and with lanthanum added are given in Table 11. These demonstrate that for standard solutions there is no significant difference between the standard deviations obtained by manual or automated injection.Blanks of distilled water were indistinguishable from background noise and consequently the normal method of calculation of limit of detection, based on the within-batch standard deviation of the blank determination,l3 was not possible. Instead, the within-batch standard deviation of the 1.0 pg 1-1 lead standard given in Table I1 was used as the closest approxi- mation. The limit of detection of the electrothermal atomisation technique was thus calculated as 0.65 and 0.60 pg 1-1 for the manual and automated injections, respectively, both of which were well within the limit of 5 pg 1-1 recommended by the WRC. concentration (whichever is the greater) concentration (whichever is the greater)26 BERTENSHAW t?t at?. : LEAD IN DRINKING WATER BY Analyst, vol. 106 TABLE I1 PRECISION TESTS ON STANDARDS Results obtained from ten randomised replicate analyses of each standard.Injection method Manual injection . . Automated . . . . Lead standard solutionslpg 1-l r 7 Statistical test 1 5 10 20 40 Standard deviationlpg 1-l 0.14 0.13 0.28 0.90 0.84 Relative standard deviation, % 14 2.6 2.7 4.5 2.1 Standard deviationlpg 1-1 0.13 0.23 0.46 0.58 1.47 Relative standard deviation, % 13 4.6 4.6 2.9 3.7 J. A bulk sample of local tap water (sample N in Table IV; total hardness 221 mgl-1 as calcium carbonate) was then collected and acidified, and separate aliquots were spiked with 0.0, 2.0, 5.0, 10.0, 20.0, 50.0 and 100.0 pg 1-I of lead. Lanthanum (as lanthanum chloride) was added to the sample aliquots, which were then analysed, using the operating conditions described, in duplicate and in random order each day for 5 d.The results obtained were analysed using statistical methods13 to produce estimates of the within-batch (S,) , between- batch (&) and total (St) standard deviations, and these were compared with the required performance characteristics (Table 111). It can be seen from Table I11 that the require- ments were met at all levels of concentration. Precision tests were then carried out on a range of potable water samples, for which the concentrations of the major ions are given in Table IV. The total standard deviations obtained with lanthanum additions to the samples are given in Table V, which includes both manual and automated injection of samples. It can be seen that all the results meet the requirements of not more than 1.5 pg 1-1 or 5% of the concentration (whichever is the greater).These results confirm that the electrothermal atomisation technique is capable of producing satisfactory precision for the routine determination of lead in potable water. The bias (accuracy) of the electrothermal atomisation technique using lanthanum treat- ment was assessed by comparing the results obtained for the above range of potable waters with those obtained by the standard additions technique. The standard additions tech- nique, although not entirely free from interference effects, is generally accepted8 as giving the best estimate of the true concentration of a determinand. The results obtained for this comparison are given in Table VI, from which the bias of individual results has been calcu- lated.l3 It can be seen that no bias exceeds the requirement of 5pg1-1 or 10% of the concentration (whichever is the greater). TABLE I11 PRECISION TESTS ON SPIKED TAP WATER The sample used was sample N, Table IV.Concentration of spikelpg 1-l 0.0 2.0 6.0 10.0 20.0 50.0 100.0 Standard deviationlpg 1-1 I A I Within-batch Between-batch Total* ( S W ) (sb) (St) 0.24 0.00 0.24 0.24 0.00 0.24 0.43 0.00 0.43 0.59 0.00 0.59 0.87 1.39 1.63t 2.63 1.44 3.00 (6.0%)? 4.01 0.00 4.01 (4.0%) * Requirement is a total standard deviation not greater than 1.5 p g 1-1 7 Not significantly different from the target a t the 95% confidence level. or 5% of the concentration (whichever is the greater).January , 1981 AAS WITH ELECTROTHERMAL ATOMISATION TABLE IV CONCENTRATIONS OF THE MAJOR IONS IN THE SAMPLES TESTED 27 Sample A ... . B .. . . c . . .. D . . . . E . . . . F . . . . G . . . . H . . . . I . . . . JK a . . . . . . . L . . . . M . . . . N . . . . Type of water sample Treated water Borehole Borehole Borehole Borehole Tap water Tap water Tap water Tap water Tap water Tap water Tap water Tap water Tap water conductivity/ pS cm-I Ca2+ 180 31 626 75 430 52 703 106 960 200 364 55 62 6 62 6 62 6 66 7 430 44 66 6 615 78 300 37 Mg2+ 3.5 9.5 31 40 26 15 1 1 1 1 7 2 28 31 Total hardness (CaCO,) Na+ K+ 92 6.4 3.1 227 32 3.1 258 9.1 2.5 430 17 2.8 607 54 3.8 199 6.0 2.7 19 4.2 0.4 19 4.2 0.5 19 4.2 0.4 22 3.9 0.5 140 24 3.6 23 30 7.6 310 33 5.6 22 1 5.1 4.4 Concentration/mg 1-1 Electrical A \ Soda- 47 71 25 163 432 31 3 3 4 5 39 104 124 60 c1- 21 61 24 26 26 19 12 12 12 13 47 13 59 80 In order to demonstrate the efficiency of the lanthanum treatment technique, some of the potable water samples were spiked with 4Opg1-1 of lead and analysed under identical conditions both with and without lanthanum addition to the samples.The results obtained are given in Table VII and represent the mean of duplicate determinations. These clearly show the benefits of the lanthanum treatment technique, the results without lanthanum present being considerably lower (suppression 2449%) than those with lanthanum present (suppression 7-9%). The samples are tabulated in order of increasing total hardness, (199-607 mg 1-1 as calcium carbonate), and it is interesting to note an apparent correlation between increasing hardness and increasing suppression of results for samples without lanthanum treatment.The apparent correlation between total hardness and suppressive interference effects for the determination of lead by electrothermal atomisation has been reported by other workers8 However, it is thought that total hardness is not the only factor causing suppression of results and that the effects are complex and require further elucidation. TABLE V PRECISION TESTS ON A RANGE OF SAMPLES A B C D E F G H I JK L M N Sample . . .. ,. .. .. .. . . . . . . . . .. .. .. .. Mean concentration/ pg I-' .. 6.6 .. 2.6 .. 0.9 .. 2.8 .. 3.6 .. 2.6 .. 79 .. 100 .. 47 .. 65 .. 11.6 .. 4.0 .. 13 .. <0.6 Total standard deviationlpg l-l* 0.31 0.18 0.12 0.37 0.46 0.20 3.16 (4.0%) 2.70 (2.7%) 1.32 4.16 (6.4%)t 1.35 0.66 1.44 - * Requirement is a total standard deviation not greater than f Not significantly different from the target a t the 95% 1.5 pg 1-I or 5% of the concentration (whichever is the greater).confidence level.28 BERTENSHAW et at?. : LEAD IN DRINKING WATER BY AnaZyst, VoZ. 106 TABLE VI COMPARISON OF RESULTS FOR LEAD IN POTABLE WATERS BY THE STANDARD ADDITIONS TECHNIQUE AND LANTHANUM TREATMENT TECHNIQUE True concentration of lead by standard additions Sample techniquelpg 1-' A .. . . 7.3 B . . . . 2.9 c . . . . 1.0 D . . . . 2.7 E . . . . 3.2 F .. . . 2.7 G .. . . 76 H . . . . 99 I . . . . 45 . . 69 12 L . . . . 3.6 M .. . . 12.4 N . . . . <0.6 JK * . . . . . Concentration of lead by direct analysis in presence of lanthanumlpg 1-l 6.6 f 0.37 2.6 f.0.1 0.9 f- 0.1 2.8 f 0.2 3.6 f- 0.3 2.6 f 0.1 79 & 1.8 100 f- 1.6 47 0.7 65 f- 2.4 11.5 f 0.8 4.0 f 0.4 13.0 f 0.8 <0.6 Mean bias/ Maximum possible CLQ 1-' bias*/pg l-l -0.7 - 1.0 -0.3 -0.4 -0.1 -0.2 +0.1 f0.3 + 0.4 + 0.7 -0.1 -0.2 + 3.0 +4.8 + 1.0 +2.6 + 2.0 + 2.7 - 4.0 -6.4 (9.3%) -0.6 - 1.3 + 0.4 +0.8 + 0.6 + 1.4 * Requirement is a bias of not more than 5 pg 1-' or 10% of the concentration (whichever is the greater). t 90% confidence limits. All the above experiments were conducted with lanthanum added to the samples and/or standards (as lanthanum chloride) before injection. An alternative procedure was employed whereby the furnace tube was impregnated with lanthanum (as lanthanum nitrate) and the sample and/or standard injected directly into the pre-treated tube.Similar results were obtained (Table VIII) as reported above for the addition of lanthanum to the samples and it was concluded that either lanthanum addition to samples or pre-treatment of the furnace tube with lanthanum is equally efficient in eliminating suppressive interference effects in the determination of lead in drinking water. However, whereas the tubes used with the addition of lanthanum to samples had a lifetime of about 500 firings, lanthanum-impregnated tubes lasted only 100-120 firings, and for the electrothermal atomiser used for this evaluation, the former is to be preferred. This is not so with an electrothermal atomiser of another manu- facture (Varian) for which furnace-tube lifetimes are similar under both lanthanum treatment conditions. l4 TABLE VII COMPARISON OF RESULTS WITH AND WITHOUT LANTHANUM TREATMENT Actual* Total hardness lead concentra- Sample as CaCOJmg 1-' tion/pg 1-1 F ..199 42.7 B . . 227 42.9 C . . 268 41.0 D .. 430 42.7 E . . 607 43.2 Lead concentration found with lanthanum/ 38.9 41.2 38.1 42.7 46.2 CLg I-' Lead concentration found without Suppression, lanthanum/ Suppression, 9 32.1 25 4 31.5 27 7 31.2 24 0 26.4 38 - 7 22.0 49 (Average 3) (Average 33) % CLg I-' % * Using data from standard additions technique (Table VI) and each sample spiked with 40 pg 1-1 of lead. Determinations Outside the Calibration Range Unlike other automatic sampling systems, which employ a microlitre syringe for sample injection, the IL254 Fastac injects a sample into the furnace tube by nebulisation and aspiration by a process similar to normal flame atomic spectroscopy, There is a linear relationship between deposit time (which is proportional to volume deposited) and concentra- tion, and decreasing deposit time therefore enables direct measurements to be made onJanuary, 1981 AAS WITH ELECTROTHERMAL ATOMISATION 29 TABLE VIII COMPARISON OF RESULTS OBTAINED USING LANTHANUM ADDITION (AS LANTHANUM CHLORIDE) TO EACH SAMPLE WITH LANTHANUM-COATED (AS LANTHANUM NITRATE) FURNACE TUBES (MANUAL INJECTION) (i) Precision, using 40 pg I-' lead standard solution*- Lanthanum addition Lanthanum-coated t o samples? furnace tube Standard deviationlpg 1-1 .. . . . . 0.84 1.52 Relative standard deviation, % .. . . 2.1 3.8 (ii) Suppression of lead result for samples spiked with 40 pg 1-' of Zead- Lanthanum addition to samples5 Lanthanum-coated furnace tube Actuallead A \ I A \ Concentration found/ Suppression, Concentration found/ Suppression, concentration$/ Sample pg I-' Pg I-' % I-' % D . . . . 42.7 42.7 0 44.0 -3 E . . . . 43.2 46.2 - 7 43.2 0 F . . . . 42.7 38.9 9 41.0 4 * Results obtained from ten randomised replicate injections of the standard. t Using data for manual injection from Table 11. $ Using data from standard additions technique (Table VI) and each sample spiked with 40 pg I-' of 5 Using data from Table VII. lead. samples containing high lead concentrations that would normally require dilution. Con- versely, by increasing deposit time, the system may be used to boost sensitivity so that ultra-trace levels of lead may be measured directly without pre-concentration of samples.Table IX shows results obtained by ten consecutive injections of a blank, standards and a sample (sample N, Table IV) using a deposit time of 100 s (which is 20x the deposit time used for the work described above). I t can be seen that the standard deviations are below the 5% target for the 0.1, 0.5 and 1.0 pg 1-1 standard lead solutions and for the sample (lead concentration approximately 0.7 pg 1-l). Using the within-batch standard deviation of the blank, the limit of detection was calculatedl3 as 0.05 pg 1-1 (50 ng 1-l). The deposit time of a sample using the IL2.54 Fastac can be increased even further, to a maximum of 999 s (which is 2OOx the deposit time used in the work described above) which should, theoretically, result in an even lower limit of detection.Interfering matrix effects may, however, prevent such a low theoretical limit of detection being applied in practice. Interference from Chloride Several authors have noted that chloride produces the most severe interference in the determination of lead in water.l2*l5 In the work reported here, no such interference was encountered and indeed it was found to be preferable to use lanthanum in the form of lanthanum chloride rather than lanthanum nitrate,' as the latter was found to contain small but significant amounts of lead. This small amount of lead is not important if the tube is impregnated with lanthanum as the lead is volatilised in the pre-conditioning stage, but it is important when lanthanum is added to each sample, as is preferred for the electrothermal atomiser used in this work.The samples employed for this investigation were typically low in chloride (the maximum being sample N with 80 mg 1-1 of chloride), but for manual injection of samples using the lanthanum addition technique, 0.1% m/V of lanthanum as the chloride was added, which was equivalent to adding a further 766 mg 1-1 of chloride to the sample. This is far in excess of any acceptable level for chloride in drinking water, .and under these conditions no suppression of lead results was obtained. In a separate experiment, again using manual injection, 500 mg 1-1 of chloride were added to a 40 pg 1-1 lead standard solution and no suppression30 BERTENSHAW et al.: LEAD IN DRINKING WATER BY TABLE IX Analyst, Voi. 106 DETERMINATION OF ULTRA-TRACE LEVELS OF LEAD USING AUTOMATED INJECTION ON TEN CONSECUTIVE INJECTIONS OF A BLANK, STANDARDS AND SAMPLE N The deposition time used was 100 s. Lead concentration/pg 1-’ 0.1 pg I-’ Pb 0.5 pg 1-’ Pb 1.0 p g 1-1 Pb Blank standard standard standard Sample N 0.04 0.11 0.49 1.09 0.73 0.04 0.10 0.51 1.02 0.75 0.04 0.10 0.47 1.06 0.74 0.04 0.10 0.52 0.98 0.73 0.02 0.10 0.51 1.02 0.76 0.04 0.11 0.52 1.02 0.67 0.04 0.11 0.50 0.97 0.70 0.03 0.10 0.50 0.97 0.73 0.02 0.10 0.49 0.98 0.75 0.02 0.10 0.51 1.02 0.70 Mean . . . . .. .. . . 0.033 0.10 0.50 1.02 0.73 Standard deviation . . .. . . 0.009 0.004 0.015 0.039 0.028 Relative standard deviation, yo .. - 4.0 3.0 3.8 3.8 was observed either with or without lanthanum treatment. It was concluded that for the electrothermal atomiser used in this work, no suppressive interference effects were observed from chloride at concentrations likely to be encountered in drinking waters. Conclusions The results presented above demonstrate that lead may be successfully determined in drinking water by electrothermal atomisation, using the lanthanum treatment technique to overcome suppressive interference effects of matrix constituents. Either pre-treatment of the furnace tube by impregnation with lanthanum or the addition of lanthanum (as lanthanum chloride) to each sample is equally effective, but for the electrothermal atomiser used for this investigation the latter provides a much longer furnace tube lifetime and is therefore pre- f erred.Tests on 14 drinking water samples from a variety of sources within the Severn-Trent Water Authority show that electrothermal atomisation together with the lanthanum treat- ment technique meets the recommended criteria of performance* for the determination of lead in potable waters in respect of concentration range, standard deviation, bias and limit of detection. In addition, as the procedure is simple and takes only a few minutes per determination, it is a suitable method for the routine determination of lead in drinking water. The authors thank M. D. Frayn, M. Gardner and G. A. Smith for their contribution to the initial work in the development of the method, and Dr. K. C. Thompson for the preparation of lanthanum-impregnated furnace tubes. The authors also thank W.F. Lester, Director of Scientific Services, Severn-Trent Water Authority, for permission to publish this work. References 1. 2. 3. 4. 5. 6. 7. Department of the Environment, “Lead in the Environment and its Significance to Man,” Pollution Department of Health and Social Security. “Lead and Health,” HM Stationery Office, London, 1980. World Health Organization, “European Standards for Drinking Water,” Second Edition, WHO, Geneva, 1970. Commission of the European Communities, “Proposal for a Council Directive Relating to the Quality of Water for Human Consumption,” Off.. J. Eur. Communities, 1975, 18, C214. Department of the Environment, “Lead in Drinking Water. A Survey of Great Britain, 1975- 1976,” Pollution Paper No. 12, HM Stationery Office, 1977. Regan, J . G. T., and Warren, J., Analyst, 1978, 103, 447. Thompson, K. C., Wagstaff, K., and Wheatstone, K. C., Analyst, 1977, 102, 310. Paper No. 2, HM Stationery Office, London, 1974.January, 1981 AAS WITH ELECTROTHERMAL ATOMISATION 31 8. 9. 10. 11. 12. 13. 14. 15. Ranson, L., and Orpwood, B., “An Evaluation of an Electrothermal Device for the Determination of Lead and Cadmium in Potable Water,” Water Research Centre Technical Report No. 49, 1977. DolinSek, F., and Stupar, J., Analyst, 1973, 98, 841. Regan, J . G. T.. and Warren, J., Analyst, 1976, 101, 220. Hodges, D. J.. Analyst, 1977, 102, 66. Ohta, K., and Suzuki, M., 2. Anal. Chem., 1979, 298, 140. Cheeseman, R. V., and Wilson, A. L., “Manual on Analytical Quality Control for the Water Industry,” Thompson, K. C., personal communication. Manning, D. C.. and Slavin. W., Anal. Chem., 1978, 50, 1234. Water Research Centre Technical Report No. 66, 1978. Received June 6th, 1980 Accepted July 2&h, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600023

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

32 Analyst, Janzcary, 1981, Vol. 106, @p. 32-39 Laser Ablation for the Introduction of Solid Samples into an Inductively Coupled Plasma for Atomic-emission Spectrometry Michael Thompson Applied Geochemistry Research Group, Department of Geology, Imperial College of Science and Technology, London, SW7 2BP John E. Goulter Applied Research Laboratories Ltd., Luton. Bedfordshire, L U4 8PU and Friedrich Sieper CZ Scientific Instruments Ltd., Borehamwood, Hertfordshire, WD6 1 NH A laser - inductively coupled plasma microprobe has been assembled from commercially available instruments, viz., a laser microprobe and an inductively coupled plasma source optical-emission spectrometer (ICP) , together with a simply constructed sample chamber and a light-activated switch. The purpose of the combination was the volatilisation of solid samples into the plasma. The system was tested with a series of standard steel samples, which produced linear calibrations for a range of elements.The absolute detection limits obtained were comparable to those obtained with solution nebulisation on the same ICP and the precision was good for elements homogeneously distributed in the sample. Initial work on cali- bration for silicate rocks was undertaken. Keywords : Laser ablation ; inductively coupled plasma ; atomic-emission spectrometry ; steel samples ; silicate rocks Optical emission spectrometry with an inductively coupled plasma excitation source (ICP) is now well established as a powerful tool for simultaneous multi-element analysis. How- ever, the conventional method of sample introduction, viz., nebulisation of the dissolved sample, has several limitations in the analysis of geological materials, the most serious of which is the problem of devising a multi-element dissolution method.Of the normally available methods,l those based on hydrofluoric acid give incomplete dissolution of some important minerals (e.g., cassiterite and wolframite), and loss by volatilisation for certain elements (e.g., silicon, arsenic, selenium and boron). Lithium metaborate fusions are limited by a large dilution factor, contamination and loss of certain elements by volatilisation and other processes. The direct introduction of solid samples would avoid these problems and also the cost of the dissolution process, but has so far proved difficult to achieve.Insufflation of powdered samples into the ICP has proved to be unsatisfactory because of segregation, erratic sample introduction rate and selective volatilisation eff e c k 2 Electrothermal volatilisation of aqueous solutions from tantalum filament^,^ graphite yarn4 and graphite r0ds~9~ as a method of sample introduction has been demonstrated to retain the valuable properties of the ICP, although matrix effects in the electrothermal vaporisation are likely to remain problematical. The use of lasers for the unselective volatilisation of solids seems an attractive alternative. This method, followed by spark emission spectrography of the resulting plume of vapour, has been widely used as a method of analysing microsamples or the virtually non-destructive analysis of larger objects.' The Lasertrace system (Barringer Research Inc., Rexdale, Ontario) for the analysis of airborne particulate matter, employs a laser to vaporise the sample before introduction of the ablated material into the ICP for multi-element analysi~.~?~ The analytical sequence in this system, in which the sample is collected on adhesive tape, is fully automatic, a multi-element analysis being completed every 10 s.The laser - ICP combination has also been used for analysis of powders and metal samples,1° although details have not been published.THOMPSON, GOULTER AND SIEPER 33 This paper describes an exploratory study of the use of a commercially available laser microanalyser that can be easily interfaced with an ICP for multi-element analysis, to form a laser - ICP microprobe.The material volatilised by the laser is swept into the plasma by a stream of argon. The primary application is intended to be in the field of geological materials, but some initial work was carried out on steel samples, as suites of well characterised and homogeneous standard samples are readily available, in contrast to the situation for geological materials. Experimental Laser Microprobe The LJIX10 Laser Microspectral Analyser (Carl Zeiss Jena) consisted of a ruby laser head, capable of delivering pulses of energy up to 1.0 J, mounted in a binocular microscope. The laser pulse is automatically focused by the microscope optics at the centre of the field of view, thereby providing the means of carrying out localised microanalysis, or probing micro- samples.The effect of the pulse can be modified by controlling the total energy, the beam width and the degree of Q-switching. The minimum crater size that could be produced was 10 pm diameter, although in this work larger crater sizes of 300 pm were required in order to achieve satisfactory sensitivities. Inductively Coupled Plasma Spectrometer The ARL 34000 C/ICP (Applied Research Laboratories) is a 1-m vacuum spectrometer fitted with lines for 36 elements. The instrument was provided with a PDP 11/04 computer for instrument control and data processing. The signal output was integrated for a time period determined by the operator, and a pre-integration time for signal stabilisation could also be set. Alternatively, the raw output from any single channel could be monitored continuously by means of a fast-response chart recorder.The ARL ICP source was run at a forward power of 1250 W, with a viewing height above the load-coil of 14 mm. The gas Bow-rates (all argon) used were coolant 12, plasma gas 0.8 and injector flow 0.5 lmin-l. The sample carrier flow-rate was set at 0.5 1 min-l for a maximised integrated signal output. Ablation Chambers and Connections The samples were subjected to the laser pulse in glass chambers similar to the form shown in Fig. 1 . The height of the cell above the surface of the sample was limited to 17 mm, which was the focal distance from the objective of the microscope. Cells of various diameter, from 20 to 50 mm, The laser window was a 35-mm disc of 2.0 mm thick optical-grade silica.output I I I Plasma object ive torch f I \ I I Light- activated switch Sample ablation optic chamber fibre Schematic diagram showing the laser - ICP microprobe system. Fig. 1.34 THOMPSON et al. : LASER ABLATION FOR SOLID Analyst, Vol. 106 were used to accommodate samples of different size. The three-way tap was of the wide diagonal-bore type, and connections between the chamber and the plasma torch were of 6 mm i.d. flexible PVC tubing (total length 92 cm), except for a short length of narrower connecting tubing . The cells were open bottomed, sealing being achieved by mounting the sample and cell on an inverted rubber bung. The whole assembly was firmly mounted on the microscope stage. While samples were being loaded the tap was used to isolate the plasma, which would other- wise have been extinguished by introduction of air.After re-sealing, a 3-s period was allowed to elapse before reconnecting the injector flow, to flush the air out of the cell. For good microscopic image quality the silica window of the cell should be perpendicular to the optical axis of the microscope. However, no special care was needed for the ablation, because the effect of small divergences from the perpendicular had no noticeable effect on the analytical response. For most sample types no contamination of the chamber window was observed. Ablation of brass samples, however, caused sputtering on to the window, which eventually needed replacement or cleaning. Light-activated Switch The light-activated switch was provided to synchronise the firing of the laser and the initiation of the integration cycle of the spectrometer.The switch was connected in parallel with the normal start button of the ARL 34000 and provided a momentary short-circuit in response to an increase in light intensity. The light pulse was transmitted to the switch by an optic fibre, the end of which was fixed about 10 mm from the cell, directed approximately towards the focal point of the laser. Results and Discussion Time-resolved Output Fig. 2 shows signals obtained from the iron channel after a sample of mild steel was subjected to a pulse of the laser. The signal was unaffected for about 2 s after the pulse, then rose rapidly to a maximum in about 1 s and gradually decayed to the original level with a half-life of about 1.2 s.The decay time seemed not to depend on the volume of the cell, although it was slightly affected by the position of the sample in the cell. This suggests that there was little mixing between the plume of vaporised material and the surrounding gas in the chamber. Trace constituents showed the same decay profile as the major constitu- ents, and sulphide and silicate minerals also gave the same results, so there seems to be no differentiation of the vaporised sample as a function of time. The plasma momentarily increased in brightness at the time of the laser pulse, but no effect on the recorder trace greater than the level of the “noise” could be observed. This 0 5 10 15 Time after laser pulse/s Fig. 2. Instantaneous response as a function of time on the iron channel after laser ablation of a sample of steel, in chambers of different volumes: A, about 5 cm3; and B, about 30 cm3.January, 1981 SAMPLES WITH ICP ATOMIC-EMISSION SPECTROMETRY 35 was demonstrated by running recorder traces at high sensitivity for elements known to be absent from the sample and with analytical lines covering a wide range of wavelengths.Hence no effect equivalent to the pressure pulse in electrothermal vaporisation studies5 was observed in this system, presumably because of the small total energy input (about 1 J). As a result, a blank integration of plasma background (i.e., with no laser pulse) gave the same reading as a complete cycle including the laser pulse for a sample with zero analyte concentration. The integration conditions selected on the basis of the single element traces was 1.0 s pre-integration followed by 5.0 s integration, this combination giving a good signal to back- ground ratio, although possibly the best reproducibility might be obtained by increasing the integration time to include a higher proportion of the sample emission peak.Effect of Laser on Samples The physical effect of the laser under various conditions on a wide variety of sample materials has been already described for the d.c. arc - laser system.' Results obtained with the ICP system were very similar, except that oxidation was prevented by the argon atmosphere. With the most energetic laser pulses, metal samples gave roughly hemispherical craters of about 300 pm diameter. Although this implies a mass loss of about 30 pg, depending on the density, the gravimetric mass loss observed, determined with a micro- balance after ten laser pulses on a thin disc of metal, was of the order of 1 pg.Most of the metal removed from the crater, apparently in the molten state, had been forced out to form a shallow rim to the crater. Minerals generally gave irregular craters, small fragments tending to spa11 off at the edge. The use of the Q-switch facility produced shallow craters, but lower sensitivity, and was not found to be of great advantage, even in the study of surface coatings. Reproducibility Obtained on a Steel Standard The laser - ICP microprobe was used on a standard low-alloy steel BCS 402 (Bureau of Analysed Samples Ltd.), with a flat surface prepared on a linishing machine.A representa- tive selection of results obtained by simultaneous analysis are shown in Table I, the statistics being based on ten successive laser pulses, each on a fresh site on the sample. The results show coefficients of variation (on the raw output less the blank value) of less than 2% for iron, and also for nickel, an element known to be homogeneously distributed in solid solution in the iron phase. Copper and chromium gave higher coefficients of variation, and that of silicon, known to segregate in a minor phase, was 24.6%. The last result emphasised the problems of sampling variance with such small samples (about 1 pg) (see below). However, the fact that several elements gave good precision shows that the stability of the whole system, including the power of the laser pulse, the efficiency of transport of the ablated material and the sensitivity of the ICP combined, is potentially very good if a sufficiently homogeneous sample can be obtained.With this material the effect on reproducibility of using iron as an internal standard was small, and noticeable only on elements for which the precision was comparable to that for iron. Hence the coefficient of variation for the nickel to iron ratio was reduced to 1.1%. However, later work with some other standards gave more variable results, and in these instances ratioing to iron gave a marked improvement in the precision for most minor elements. TABLE I REPRODUCIBILITY OF THE LASER - ICP SYSTEM ON STEEL STANDARD BCS 402 The statistics are based on ten replicate single laser pulses, blank corrected.Element L r \ Parameter Fe cu Ni Cr S1 Xi, Fe Concentration in RCS 302, . . 96.3 0.23 0.73 0.55 0.27 Mcan response/mV . . . . . . 110.33 26.56 25.31 74.81 53.5.5 0 . 0 3 i 3 Standard dcx iation,'in\- . . . . 8,05 1 .'6 0.39 i (10 1 3 . 3 6.4 *. IWJ Coefficient of \ ariation, Oi) . . 1.8 1.7 1.5 9.0 24.6 1.136 Analyst, Vol. 106 Calibration of Minor Elements Calibration graphs for a selection of minor elements were prepared by subjecting a suite of standard analysed steels (BCS 01-10) to the ICP laser microprobe. Each standard was used to produce three or four results based on single laser pulses. These results were plotted in two ways: (a) raw response v e w m certified concentration [e.g., Fig.3 ( a ) ] and ( b ) raw results ratioed to the iron result for the same burn VCYSUS the certified concentration ratioed to the iron concentration [e.g., Fig. 3 ( b ) ] . The results illustrated for nickel are typical. The raw results show a strong linear trend, with some samples giving a much wider spread than others and one yielding results apparently centred at a significant distance from the line (no statistical analysis of the data has been made). The ratioed results show a closer adherence to a straight line, with groups centred more closely to the line. On both lines the plasma blank is coincident with the intercept on the response axis at zero concentration. Clearly, the use of iron as an internal standard has a beneficial effect in most instances. Calibration lines produced in this manner by linear regression for the simultaneous determina- tion of manganese, molybdenum, copper, vanadium, chromium, phosphorus, sulphur and silicon are shown in Fig. 4.The points shown are the means of three or four separate laser pulses. Phosphorus and sulphur visually yield a more scattered plot, but this is due to the larger scale used; they are present at much lower concentrations than the metallic analytes and the absolute errors are similar. Silicon [Fig. 4 ( h ) ] did not give a significant calibration at the levels present. This may have been due to transport problems, as silica may tend to be adsorbed on the chamber walls. The response for other elements, when ratioed, does not seem to depend greatly on the surface finish of the sample, provided that no foreign material is present.Small (about 1.0 mm) rough chippings and turnings from the standards gave results very similar to those obtained on the finished surface of the bulk material. THOMPSON et al. : LASER ABLATION FOR SOLID The results generally are comparable to those for nickel. 800 I 700 600 $ 500 400 $ 300 200 100 . a, 0 a CT 0 1 .o 2.0 3.0 4.0 5.0 Ni concentration, YO bv mass 0 1 .o 2.0 3.0 4.0 5.0 NiiFe concentration ratio, O/O Fig. 3. Calibrations produced by the laser - ICP micro- probe on standard steel samples : ( a ) raw instrument response versus nickel concentration; and ( b ) nickel responses and concentrations ratioed to iron. All samples were analysed in triplicate or quadruplicate, but some individual points are unresolved on the graphs.January, 1981 SAMPLES WITH ICP ATOMIC-EMISSION SPECTROMETRY 20 l 5 - 0 l o - 14 12 - ( a ) Copper 10 - 0 0.2 0.4 0.6 0.8 1.0 5 4 - (c) Manganese 0 1 2 - 0 0 0 0 I , I I I 0 0.1 0.2 0.3 0.4 0.5 0.6 1.0 r 0.4 f 0.3 ' 0 0.01 0.02 0.03 0.04 0.05 0.06 80 ( b ) Chromium 0 60 - 0 1 2 3 0 0.5 1 .o I 1.8 1.6 1.4 1.2 0 0.02 0.04 0.06 0.08 25 1 ( h ) Silicon 5 t 0 0 0 0 i I > 0 0.5 0.1 1.5 37 Fig.4. Calibrations produced on the laser - ICP microprobe on standard steel samples. Horizontal axes show the concentration of the element ratioed to iron (7; by mass). Vertical axes show instrumental response for the element ratioed to iron response. Detection Limits Detection limits for trace elements in steel were established by a study of the reproduci- bility of the system on several standards with low levels of analytes.The results are shown in Table 11. The relative detection limits in the steels are not as good as those obtained by macroteclmiques, except for sulphur and phosphorus. The absolute detection limits, based on an assumed ablation of 1 pg of sample, are broadly comparable to the absolute detection limits obtained on the same ICP instrument with conventional nebulisation, assuming a nebuliser efficiency of 2%. However, the absolute limits for sulphur and phosphorus obtained by the laser technique were significantly lower. No attempt was made in this study to measure the transport efficiency of the ablated material, but further reductions in38 THOMPSON et ad. : LASER ABLATION FOR SOLID TABLE I1 Analyst, Vol.106 SOME ESTIMATED DETECTION LIMITS FOR THE LASER - ICP MICROPROBE COMPARED WITH NEBULISATION ON THE SAME ICP INSTRUMENT Element Cr . . s .. P .. Mn . . Mo .. c u .. v .. Ni . . Concentration detection limit for steel/pg g-l 5 G - z z z r .. 15 2 .. 15 45 .. 10 35 . . 80 2 .. 60 4.5 * . 20 1.5 .. 10 1.5 .. 70 1.0 Absolute detection limitlpg 15 25 15 250 10 150 80 5 60 20 20 12 10 20 70 70 * Assuming 1% solution of the steel. t Assuming 1 pg of sample ablated. $ Assuming 2% efficiency at 0.7 ml min-l uptake. detection limit may be possible if this feature can be improved, e.g., by improving the design of the ablation cell and tubing on aerodynamic principles. The absolute detection limits compared favourably with those obtained by various electrothermal vaporisation methods.3-6 Calibration for Silicate Materials Well characterised standard samples of geological materials are available for testing analytical methods, but (for laser ablation) suffer from the disadvantage of being hetero- geneous powders rather than homogeneous solids.The use of bulk analysed large crystals of minerals as standards is not satisfactory. Visually homogeneous material usually gives large variations in response in different parts when subjected to the ICP laser microprobe. Several approaches have been made in this study to see whether rock standards could be used for calibrating the ICP laser microprobe. In the Barringer Lasertrace system standard rock powders have been crushed on to adhesive tape and good results obtained,8 but in the system described here the shock of the laser pulse cleaned a large area of the tape and resulted in a cloud of relatively large particles reaching the plasma.This could result in incomplete volatilisation in the plasma and may also block the plasma torch. In an attempt to provide a homogeneous material, beads of standard rocks fused with lithium metaborate were prepared and subjected to the laser ablation. In these fairly transparent materials, under the conditions used, the laser beam penetrated deeply into the beads and usually caused them to shatter, rather than ablating material from the surface. Pellets were prepared from the powdered standards by means of a hydraulic press, and also by mixing the powders with epoxy resins. Both of these techniques enabled the standards to be ablated and calibration lines to be drawn, but the results were not as satis- factory as for the steel standards.Further work on calibration for silicates and other minerals can reasonably be expected to improve this situation, as sample preparation methods such as carbonate fusion should give reasonable reproducibility on rock standards, as has already been shown on the spectrographic laser microanalyser. Conclusion A laser - ICP microprobe can be readily assembled from a commercially available ICP gpectrometer and laser microscope, together with an easily constructed ablation cell. The results obtained to date indicate that the system has a good potential for the rapid analysis of bulk materials and small fragments, and alw as a supplementary tcchnique for existing ICP s!,stem.;.How- ever, the amlFtes have to be homogeiieousl\- ciistributtd in the iiirttris >.:;!-I under these conditions \'cry reproducible results and linear- c i i i i l ) ~ ; ; ~ i( b l 1 i (-an be obtained. T-lt r;i-trace analysis does not seem possible at present because ctf ?]-it? ~i:iali aiiioiiiit of s:ii?ipic. t1'2 LSfC'rJtd to the plasma. Further work on cell design mrf prcp::ntion c,i calibration standards 1s 36-element scan requJi-ed a?-)iout 1 Ii-iiil, iiiciiiding s;tl-!i!)le loading:.January, 1981 SAMPLES WITH ICP ATOMIC-EMISSION SPECTROMETRY 39 required. The laser - ICP microprobe has distinct advantages over the original spark excitation method, viz. , better reproducibility, easier calibration and a wider range of elements that can be determined. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. References Walsh, J . N., Spectrochim. Acta, 1980, 35B, 107. Dagnall, R. M., Smith, D. J., West, T. S., and Greenfield, S., Anal. Chim. A d a , 1971, 54, 397. Nixon, D. E., Fassel, V. A., and Kniseley, R. N., Anal. Chern., 1974, 46, 210. Dahlquist, R. L., Knoll, J. W., and Hoyt, R. E., paper presented at Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, 1975. Gunn, A. M., Millard, D. L., and Kirkbright, G. F., Analyst, 1978, 103, 1066. Kirkbright, G. F., and Snook, R. D., Anal. Chem., 1979, 51, 1938. Moenke, H., and Moenke-Blankenburg, L., translated by Auerback, R., “Laser Microspectrochemical Analysis,” Adam Hilger, London, 1973. Abercrombie, F. N., Silvester, M. D., Murray, A. D., and Barringer, A. R., in Barns, R. M., Editor, “Applications of Inductively Coupled Plasmas to Emission Spectroscopy,” 1977 Eastern Analytical Symposium, Franklin Institute Press, Philadelphia, 1978. Abercrombie, F. N., Silvester, M. D., and Soute, G. S., Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, 1977, Paper 406. Salin, E. D., Carr, J.. and Horlick, G., Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, Cleveland, Ohio, 1979, Paper 563. Received July l l t h , 1980 Accepted August 27th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600032

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

40 Analyst, January, 1981, Vol. 106, pp. 4046 Atomic-absorption Spectrophotometric Determination of Trace Amounts of Arsenic in Acrylic Fibres Containing Antimony Oxide with Solvent Extraction and Arsine Generation Takashi Korenaga* Saidaiji Plant, Japan Exlan Co., Ltd., Kanaoka-higashi, Okayama-shi, 704, Japan An arsine generation atomic-absorption spectrophotometric method is described for the determination of 0.04-400 p.p.m. of arsenic in acrylic fibre samples containing antimony oxide. The interferences from antimony and other elements are eliminated by solvent extraction with benzene. The sample is digested with a mixture of concentrated nitric, perchloric and sulphuric acids. Arsenic(V) obtained by the digestion is then reduced to arsenic(II1) with titanium(II1) chloride.After the reduction of arsenic and antimony, arsenic(II1) is extracted into benzene from a sulphuric acid - hydrochloric acid mixture, from which no antimony(II1) is extracted. Arsenic(II1) in the benzene phase is further back-extracted into water. &sine is generated with potassium iodide, tin(I1) chloride and zinc powder from 2.4 N hydrochloric acid solution and introduced into a nitrogen - hydrogen flame of an atomic-absorption spectrophotometer. The method has been tested with various acrylic fibre samples containing antimony oxide, and the limit of detection and precision achieved are 0.04 p.p.m. for 1 g of sample and 7 yo (relative standard deviation) , respectively. Keywords : Arsenic determination ; acrylic fibres ; antimony oxide ; atomic- absorption spectrophotometry ; arsine generation Arsenic is an undesirable toxic impurity originating from the antimony in acrylic fibres containing antimony oxide as a fire-retarding agent and its determination is therefore important.Arsenic in organic materials has been determined by various methods, e.g. , spectrophoto- metry with silver diethyldithiocarbamatel-3 and molybdenum bluel19*-6 polarography7 and atomic-absorption spectrophotometry.* Recently, Kiyosug determined micro-amounts of arsenic in acrylic fibres containing antimony(V) oxide by an arsine generation atomic- absorption spectrophotometric method with addition of a constant amount of antimony. The method did not include an extraction procedure, was not precise (relative standard deviation 10-20%) and was not very sensitive (limit of detection 1 p.p.m.).The addition of antimony(V) oxide was needed, however, in order to adjust to a defined amount of antimony(V) oxide in the test solution, and the arsine generation was subject to interference by the co-existing antimony. Therefore, the Kiyosu method is not of great practical value because of its narrowly defined conditions for arsine generation and its inaccuracy in the determination of arsenic, although its sensitivity is adequate for the determination of trace amounts of arsenic in acrylic fibres. It is not possible to apply the above methods to the determination of trace amounts of arsenic in acrylic fibre samples containing antimony oxide because of their poor sensitivities and interference from co-existing antimony as a matrix element.Therefore, a precise and reliable method is needed for the determination of trace amounts of arsenic in organic material samples such as acrylic fibres containing antimony oxide. The authorlo recently reported a sensitive, selective and precise method for determining arsenic in antimony compounds by arsine generation atomic-absorption spectrophotometry with elimination of interferences caused by co-existing elements (e.g., antimony) by solvent extraction of arsenic(II1) with benzene from concentrated hydrochloric acid. Japan. * Present address : School of Engineering, Okayama University, Tsushima-naka, Okayama-shi, 700,KORENAGA 41 In this work, the method used previously in this laboratoryg has been improved by using solvent extraction with benzene, and also the method developed recently by the authorlo has been extended into an application for the determination of trace amounts of arsenic in acrylic fibres containing antimony oxide by atomic-absorption spectrophotometry using benzene extraction and arsine generation, with satisfactory results.Experimental Reagents All reagents were of analytical-reagent grade. Arsenic(II1) standard solution, 1000 pg 1-l. Dissolve 132.0 mg of arsenic(II1) oxide (purity 99.99%) in 2 ml of 4% m/Y sodium hydroxide solution. Add dilute sulphuric acid until the solution becomes slightly acidic and then dilute to 100 ml with distilled water. Arsenic( V ) standard solution, 1000 pg 1-'. Oxidise the above arsenic(II1) standard solution with a sufficient amount of hydrogen peroxide, and evaporate the excess by boiling.Cool the resulting solution and dilute it to the required volume with distilled water. Use these standard arsenic solutions after accurate dilution. Z i n c tablet. Use about 1 g of zinc powder containing a binder (obtained from Nippon Jarrell-Ash Co., Ltd.) to make a tablet. Titaiziuin(II1) solution, 10% m/V. Dissolve 50.0 g of titanium(II1) chloride in concentra- ted hydrochloric acid (see below) and dilute to 500 nil with concentrated hydrochloric acid. Store this solution in a refrigerator. The hydrochloric acid used was of the grade for measurement of arsenic (guaranteed arsenic content less than 5 parts per lo9). Apparatus A Nippon Jarrell-Ash, Model AA-780, atomic-absorption and flame-emission spectro- photometer (10-cm slit burner) equipped with a Nippon Jarrell-Ash, Model ASD-I A, arsenic measurement unit and an arsenic hollow-cathode lamp (Westinghouse Electric Corp.) were used.A schematic diagram of the apparatus used for the arsenic measurement is shown in Fig. 1 and the analytical conditions are given in Table I. Procedure Preparation of sample solution than 2 pg of arsenic) into a Kjeldahl decomposition flask. Ilieigh accurately a required amount of acrylic fibre sample (10 mg-1 g to contain less Add 3 ml of concentrated nitric 1 I I r I] T I L --.- Fig. I . Schematic diagram of apparatus used for the arsenic measurement. A, Slit burner; B, trap; C, change-over cock; D, buffer tank; E, arsine generator (glass reaction bottle) ; F, magnetic stirrer; G, pressure gauge; H, gas flow meter; I, back-sweep cock; J , stop-cock; and K, pressure controller.a, Sweep; and b, by-pass.42 KORENAGA: ARSENIC IN ACRYLIC FIBRES Analyst, Vol. 106 TABLE I ANALYTICAL CONDITIONS FOR ARSENIC DETERMINATION Wavelength/nm .. .. . . . . Lamp current/mA . . .. . . . . Burner length/cm . . . . . . .. Burner heightlmm . . .. .. .. Mechanical slit width/pm . . .. .. Spectral slit width/nm . . . . .. . . Flow-rate of nitrogen/l min-' . . . . .. Flow-rate of hydrogen/l min-l . . . . Flow-rate of auxiliary nitrogen/l min-' . . 193.7 12 10 (slit burner) 19 100 300 7.0 3.5 4.5 acid, 3 ml of concentrated perchlorik acid and 3 ml of concentrated sulphuric acid. Heat the flask on an electric hot-plate (about 1200 W) until the acrylic fibre sample is completely decomposed.If the sample cannot be decomposed completely, add a further 3 ml of concentrated nitric acid and 3 ml of concentrated perchloric acid to the flask, and repeat the heating. When the solution becomes clear and the nitric acid and perchloric acid have been completely evaporated (when white fumes of sulphuric acid are observed), cool the resulting solution to room temperature. After the digestion of the sample, the amount of sulphuric acid remaining is about 3 ml. Transfer the solution into a 50-ml separating funnel and rinse the K jeldahl decomposition flask with 7 ml of concentrated hydrochloric acid. Add by pipette 5 ml of 10% m/V titanium(II1) chloride in concentrated hydrochloric acid solution to the funnel. Mix well and then allow the separating funnel to stand for 30 min at 60 "C in a water-bath (the funnel must occasionally be shaken and frequently de-gassed in order to avoid an explosion).Cool the solution to room temperature, add 10ml of benzene and shake the funnel for a few minutes. Discard the aqueous phase, add 10 ml of distilled water and shake the funnel for a few minutes. In this way, arsenic(V) in the digested sample is reduced to arsenic(III), and only the arsenic(II1) is extracted into benzene and then back-extracted quantitatively into the aqueous phase. Atomic-absorption spectrophotometry Transfer by pipette the aqueous solution containing arsenic(II1) (not more than 2 pg of arsenic) obtained by the above procedure into a glass reaction bottle (E in Fig. 1). Add 5 ml of concentrated hydrochloric acid, 1 ml of 20% m/V potassium iodide solution (prepared freshly each week) and 0.5ml of 20% m/V tin(I1) chloride solution.Dilute the solution with distilled water to give a final volume of 25 ml (final concentration of hydrochloric acid 2.4 N). Drop a zinc tablet containing about 1 g of zinc powder into the reaction bottle (E) and immediately connect the bottle to the collection tank (D). Store the generated arsine gas obtained by mixing the sample solution with a magnetic stirrer (F) to ensure complete arsine generation in the 100-ml collection tank (D) for about 1 min (until the pressure of arsine generation gas reaches 0.5 kg cm-2). As soon as the pressure of arsine generation gas (H) reaches 0.5 kg cm-2, switch the stopcock (C) from the by-pass (b) and introduce the collected arsine into the flame with a stream of nitrogen carrier gas.Use the peak height on the recordings to determine the concentration of arsenic. According to the analytical conditions given ,in Table I, arsenic(II1) obtained in the aqueous solution was determined at the 193.7-nm absorption line by using arsine generation and atomic-absorption spectrophotometry (Fig. 1). A calibration graph was prepared by using 0.25-2.0 pg of standard arsenic(II1) solution throughout the procedure for the prepara- tion of the sample solution, and was then used for subsequent determinations of arsenic concent rations. Mix well and allow to stand for 20 min. Results and Discussion Arsine Generation For the determination of arsenic in the standard arsenic(II1) solution by arsine generation atomic-absorption spectrophotometry (Fig.1) , operating conditions such as acidity, volume of the solutions and amounts of potassium iodide, tin(I1) chloride and zinc powder were notJanuary, 1981 CONTAINING ANTIMONY OXIDE BY AAS 43 critical. For acrylic fibre samples containing antimony oxide, however, the acidity, the amounts of potassium iodide and tin(I1) chloride and the proportions of arsenic and antimony contained in the antimony oxide severely affected the arsine generati~n.~ Hence, the separation of arsenic from the matrix element (antimony) is necessary in the commonly used techniques. Arsine is then generated with potassium iodide, tin(I1) chloride and zinc powder from 2.4 N hydrochloric acid solution and introduced into a nitrogen - hydrogen flame.Separation of Arsenic from Antimony The author1* previously described the separation of arsenic from antimony by solvent extraction of arsenic(II1) with benzene from hydrochloric acid solution. Wunderlich and Hadelerll also reported the determination of trace impurities such as arsenic in high-purity phosphorus by atomic-absorption spectrophotometry after dissolution in hydrobromic acid and extraction with benzene. In this work, several organic solvents were examined for the extraction of arsenic(II1) from 9 N hydrochloric acid. From the results, the solvents were classified into the following three groups : (1) benzene, toluene, 4-methylpentan-2-one and tributyl phosphate, which gave extractions of arsenic(II1) of more than 90% ; (2) xylene, chloroform, 1,2-dichloro- ethane, diisopropyl ether and butyl acetate, which gave extractions of arsenic(II1) of S0-90~0 ; and (3) cyclohexane, hexane and carbon tetrachloride, which gave extractions of arsenic(II1) of 50-80%.The solvents forming group (1) were then examined for the separation of arsenic( 111) from an antimony(II1) matrix with hydrochloric, hydrobromic and hydroiodic acids. Benzene and toluene were found to be the best extraction solvents, and hydrochloric acid was found to be the best acid for the separation of arsenic(II1). Although toluene, which is less hazardous than benzene, could also be used for the present purpose, benzene was selected because of its superior extraction of arsenic(II1) in this work.The Dredominant extracted species of arsenic(II1) in the benzene phase was probably an ionlassociate of HAsC1,.12 Caution-Benzene is highly toxic and appropriate precautions should be taken. The separation of 1 pg of arsenic(II1) from 50 mg of antimony(II1) by solvent extraction with benzene from a sulphuric acid - hydrochloric acid mixture was examined. In Fig. 2, the extractabilities of arsenic(II1) into benzene from the mixed acids with various concentra- tions of hydrochloric acid are shown. The optimum concentration of hydrochloric acid was about 4 N or more in 15 ml of mixed acids containing 3 ml of concentrated sulphuric acid. A t this concentration of hydrochloric acid, no antimony(II1) could be extracted into benzene. The back-extraction of the arsenic(II1) in the benzene phase was also examined; the arsenic( 111) was found to be completely back-extracted into distilled water.The extraction of arsenic(II1) in both the benzene extraction and the back-extraction was more than 95%. When the extraction of arsenic(II1) with benzene was carried out from a mixture of 3 ml of concentrated sulphuric acid and 12 ml of concentrated hydrochloric acid, the extraction was about 98%. Accordingly, the benzene extraction and back-extraction with distilled water were subsequently carried out between 15 ml of the mixed acids and 10 ml of benzene and between 10 ml of the benzene solution and 10 ml of distilled water, respectively. Interferences from iron, lead, chromium, platinum and sulphide-sulphur reported in the arsine generation atomic-absorption spectrophotometric determination of arsenic13 could also be eliminated by this solvent extraction procedure using benzene.Reduction of Arsenic(V) to Arsenic( 111) As both arsenic(V) and antimony(V) were also extracted into benzene at the above- mentioned concentration of hydrochloric acid in the sulphuric acid - hydrochloric acid mixture, the simultaneous reduction of arsenic(V) and antimony(V) was examined. Titanium(II1) chloride dissolved in concentrated hydrochloric acid was found to be the best reducing agent for reducing both arsenic(V) and antimony(V) to arsenic(II1) and antimony(III), respectively. The amount of titanium(II1) chloride necessary for the reduction of both of 1 pg of arsenic(V) [obtained from the arsenic(II1) solution by oxidation with hydrogen peroxide; see Reagents] and 50 mg of antimony(V) (obtained from 99.99% pure antimony metal by44 : 120 .E .- 2 80- r Y a 2 40 KORENAGA: ARSENIC IN ACRYLIC FIBRES - - Analyst, Vol. 106 120 : 100 . c c a L m 80 .- x 60 a 40 20 0, 0 2 4 6 8 1 0 Concentration of hydrochloric acid/N Fig. 2. Effect of hydrochloric acid concentration in sulphuric acid - hydrochloric acid mixture on the extraction of 1 pg of arsenic- (111) from the mixed acid solution containing 50 mg of antimony(II1). I,,,,, 0 2 4 6 8 1 0 Vo I u me of ti tan i u m( I I I) c h lor ide so I u t i onim I Fig. 3. Effect of volume of 10% m/V titanium(II1) chloride solution on the reduction of 1 pg of arsenic(V) and 50 mg of antimony(V) after heating for 30min a t 60 "C.dissolution with concentrated hydrochloric acid and oxidation with hydrogen peroxide) were studied. Blank tests were necessary for subtracting the arsenic contained in the antimony metal used. A constant peak height was obtained by using more than 3 ml of titanium(II1) chloride in hydrochloric acid solution in a water- bath at 60 OC.l0 Therefore, 5 ml of the titanium(II1) chloride solution were subsequently used for the complete reduction. When the reduction was carried out in a water-bath at 60 OC,l0 both arsenic(V) and antimony(V) were completely reduced to arsenic(II1) and antimony(III), respectively, after heating for more than 15min (Fig. 4). Accordingly, the reduction was subsequently performed at 60 "C for 30 min. The results obtained are shown in Fig. 3.Preparation of Sample Solution by Wet Digestion Chapman et aZ.14 reported the volatilisation of arsenic from solutions of perchloric, hydro- chloric and hydrofluoric acids. For this reason, the procedure for the preparation of the sample solution was examined for the wet digestion of acrylic fibres containing antimony oxide. The results showed that the recovery of arsenic throughout the digestion procedure described under Preparation of sample solution was more than 95%, which was superior to that obtained elsewherel by using both nitric and sulphuric acids and spectrophotometric determination of arsenic. Loss of arsenic by volatilisation could thus be prevented by using a Kjeldahl decomposition flask and a mixture of concentrated nitric, perchloric and €1 a aa 40 0 20 40 60 80 100 120 Reduction timeimin Fig.4. Effect of reduction time on the reduction of 1 pg of arsenic(V) and 50 mg of antimony(V) with 5 ml of 10% m/V titanium(II1) chloride solution at 60 "C. E 120 2 80 a 40 c L m .- Y m a 1 0 0.5 1.0 1.5 2.0 Amount of arsenic/pg Fig. 5. Working graph for the arsenic determination a t 193.7 nm.January , 1981 CONTAINING ANTIMONY OXIDE BY AAS 45 sulphuric acids. One of the reasons why a good recovery of arsenic was obtained in this work is probably that organic substances in the acrylic fibre samples are completely wet digested by the mixed acid solution used. Determination of Arsenic in Acrylic Fibres Containing Antimony Oxide Arsenic(II1) separated into the aqueous phase from the matrix antimony after the benzene extraction and back-extraction with distilled water was determined by using arsine genera- tion atomic-absorption spectrophotometry as described under Atomic-absorption spectro- photometry.A calibration graph was prepared by using a standard arsenic(V) solution and an acrylic fibre containing no antimony oxide with the recommended procedure. Although the calibration graph obtained was not a straight line (Fig. 5 ) , amounts of arsenic in the range 0.25-2.0 pg could be determined with good reproducibility under the conditions given in Table I. The graph was therefore used for determining concentrations of arsenic in acrylic fibre samples. The results for arsenic obtained with various acrylic fibre samples containing antimony oxide are given in Table 11. The relative standard deviations in these determinations were less than 7%.TABLE I1 RESULTS FOR THE DETERMINATION OF ARSENIC IN ACRYLIC FIBRES CONTAINING ANTIMONY OXIDE Sample No. 1 2 3 4 5 6 7 8 9 10 11 12 Supplier A A A A A B C C D E Content of Sb,05,* yo 5.1 5.1 4.5 4.6 5.0 3.0 (Sb,O,) - 2.4 (Sb,O,) 1.0 (Sb,O,) Sb,O, 50 mg + acrylic fibre (no Sb) 1 g Sb,O, 10 mg + acrylic fibre (no Sb) 1 g No. of determinations 4 3 2 6 5 3 2 4 2 4 2 1 Arsenic content, p.p.m. 50 f 1 84 & 4 94 f 7 10.3 & 0.5 4.1 j= 0.2 45 j= 0 180 & 11 103 j= 6 8.5 & 0.5 3.2 & 0.1 0.47 0.03 0.08 * Even when antimony(II1) oxide was present in acrylic fibres (samples 6, 9 and lo), antimony(II1) was easily and completely oxidised to antimony(V) by the wet digestion with a mixture of nitric, perchloric and sulphuric acids.Hence, arsenic in acrylic fibres containing antimony( 111) oxide could be determined as well as that in acrylic fibre samples containing antimony(V) oxide by the proposed method. Known amounts of arsenic(V) (0.50 or 1.00 pg) were added to the samples in order to The results obtained are given in Table In order to check the results of determinations of arsenic in acrylic fibre samples, the For all samples, linear graphs were obtained between study the recovery of arsenic from acrylic fibres. I11 and show recoveries of arsenic in the examined range from 96 to 104%. amount of sample taken was varied. TABLE I11 RECOVERY OF ARSENIC ADDED TO SAMPLES Sample No. Arsenic contentlpg 1 0.50 1.01 1.01 4 1.03 5 0.41 0.41 9 0.86 0.86 11 0.49 Arsenic(V) addedlpg Arsenic found/pg 0.50 1.02 0.50 1.49 1 .oo 1.98 0.50 1.54 0.50 0.93 1-00 1.41 0.50 1.35 0.50 [arsenic(III)] 1.34 0.50 0.97 Recovery, % 104 96 97 102 104 100 98 96 9646 KORENAGA the amount of sample taken and the amount of arsenic determined.Accordingly, the proposed method for the determination of arsenic in acrylic fibre samples containing antimony oxide was quantitative. I t is concluded that the method described here may be applicakle to the determination of trace amounts of arsenic in other organic materials containing antimony compounds. The author is greatly indebted to Professor Kyoji T8ei and Dr. Shoji Motomizu of Okayama University and Mr. Hiromitsu Kiyosu of Japan Exlan Co., Ltd., for their encourage- ment and valuable advice, and also to Japan Exlan Co., Ltd., for permission to publish this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Analytical Methods Committee, Analyst, 1975, 100, 54. BS 4404: 1968, British Standards Institution, London. Magnuson, H. J., and Watson, E. B., Ind. Eng. Chem. Anal. Ed., 1944, 16, 339. Hoffman, L., and Rowsome, M., Analyst, 1960, 85, 151. Powers, G. W., Jr., Martin, R. L., and Piehl, F. J.. Anal. Chem., 1959, 31, 1589. Evans, R. J., and Bandemer, S. L., Anal. Chem., 1954, 26, 595. Hagiwara, K., Bunseki Kagaku, 1970, 19, 563. Griffin, H. R., Hocking, M. B.. and Lowery, D. G., Anal. Chem., 1975, 47, 229. Kiyosu, H., Japan Exlan Co., Ltd., unpublished work, 1975. Korenaga, T., Mikvochim. Acta, 1979, I. 435. Wunderlich, E., and Hadeler, W., Z . Anal. Chem., 1977, 284, 19. Tanaka, K., Bunseki Kagaku, 1960, 9, 574. Terashima, S., Anal. Chim. Ada, 1976, 86, 43. Chapman, F. W., Jr., Marvin, G. G., and Tyree, S. Y . , Jr., Anal. Chem., 1949, 21, 700. Received July 3rd. 1980 Accepted August 4th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600040

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, January, 1981, Vol. 106, $9. 47-53 47 Determination of Copper, Nickel and Iron in Heavy Alloy Using Atomic-absorption Spectrophotometry Mangala A. Pabalkar, S. V. Naik and N. R. Sanjana Research and Development Centre, Sandvik Asia Limited, Poona 411 012, India A method is presented for the rapid and fairly accurate atomic-absorption spectrophotometric determination of the major alloying elements and a minor constituent element of the tungsten-based heavy alloy 90 tungsten - 7.5 nickel - 2.5 copper. The sample is dissolved in a mixture of hydro- fluoric and nitric acids. Ammonium citrate is added to keep the tungsten in solution, which is later neutralised with ammonia and made up to a suitable volume. Flame absorbance measurements are made under specified instrumental conditions for the three elements.The effect of variation of tungsten concentration in the matrix and of the variation of the concentration of ammonium citrate, nitric acid and hydrofluoric acid on each of the mentioned constituents is evaluated. Keywords ; Tungsten-based heavy alloy analysis ; atomic-absorption spectro- photometry ; flame absorption ; nickel, copper and iron determination The heavy alloys were first produced and reported by McLennan and Smithellsl in 1935 and they are in effect three-component pseudo-alloys consisting of tungsten in either a nickel - copper or a nickel - iron system. The development of these heavy alloys was reviewed in detail by Larsen and Murphy.2 In order to achieve the desired physical and mechanical properties such as density, tensile strength, compressibility and elongation, the solubility of tungsten in nickel is of importance.However, although this solubility is as high as 40% m/m, the resulting binary alloy is of no value as it is brittle. Elements such as copper and iron, on addition to the above, limit the solubility of tungsten and thus toughen the matrix, rendering it more ductile.3 As one of the major manufacturers of this alloy (90 tungsten - 7.5 nickel - 2.5 copper) in India, we have tried to devise rapid and reliable methods of compositional analysis in this system. Earlier the analysis of nickel, copper and iron was carried out by either gravi- metric, volumetric or spectrophotometric methods, involving time-consuming procedures owing to the need for the separation of each element from the others and their complexation with a costly reagent to convert them into measurable forms.For example, nickel had to be separated from the heavy alloy and precipitated as nickel dimethylglyoxime for gravimetry, copper had to be determined either by electrolysis or by iodimetric titration after separation by precipitation with sodium diethyldithiocarbamate in the presence of EDTA4 and iron had to be determined spectrophotometrically using potassium thiocyanate after separation as the hydroxide in an ammonia - ammonium nitrate medium. Thus, at least three working days were required to complete a series analysis of, say, four samples of heavy metal mixed powder or sintered heavy metal (alloy). In order to shorten the analysis time and to eliminate the need for some of the more expensive reagents, atomic-absorption spectro- photometry was considered as an alternative technique.This method has been tried for the determination of nickel and copper in tungsten metal.5 In this paper we describe an atomic-absorption spectrophotometric method and its comparison with the methods used earlier. The effect of variations of the tungsten concentration in the alloy matrix and of the concentrations of the various reagents was studied. Experimental Apparatus and an air - acetylene (oxidising-lean, blue) flame was used with the following settings : A Perkin-Elmer 306 atomic-absorption spectrophotometer with a 10-cm, three-slot burner48 PABALKAR et al. : COPPER, NICKEL AND IRON IN Analyst, Vol. 106 Copper Nickel Iron Wavelength .. . . 324.8 nni 341.5 nm 248.3 nm . . . . 4 (0.7 nm) 3 (0.2 nm) 3 (0.2 nm) Slit . . Range.. .. . . Ultraviolet Ultraviolet Ultraviolet The working range, under the above standard conditions, was linear up to concentrations of about 5 pg ml-l for copper, 20 pg ml-l for nickel and 6 pg ml-l for iron. A less sensitive line was deliberately chosen for nickel in order to improve the precision by selecting a lower dilution. This also enabled both copper and nickel to be determined in the same final solution. The linearity of the working range for each element was established by adding increasing volumes of the standard solution of the element to a series of synthetic solutions containing the other three elements at a fixed optimum level of concentration.The linearity range for iron, for example, was established by adding a standard solution of iron in increasing volumes to a series of synthetic solutions having identical concentrations of tungsten, copper and nickel. Reagents Analytical-reagent grade materials were used unless indicated otherwise. Nitric acid, 4 N. Hydropuoric acid, 40% mlm. Ammonia solution, 28% mlm. Ammonium citrate solution. Dissolve 500 g of citric acid in 4000 ml of water + 450 ml of 28% m/m ammonia solution. Stock standard solution of copper, nickel and iron. Dissolve 801 mg of NBS standard 162a nickel - copper alloy (nickel 63.95%, copper 30.61y0, iron 2.19%) as described below under Procedure and dilute to 500 ml with water. A 1-ml volume of this solution contains 0.49 mg of copper, 1.02 mg of nickel and 0.035 mg of iron.Analytical standard solution (synthetic) for atomic-absorption spectroehotometric measure- ments. Take 520 mg of tungsten (the tungsten used was about 99.8% pure and contained 0.048~0 of iron; it was free from nickel and copper) and prepare the solution as described below under Procedure. Add to this solution of tungsten 50 ml of NBS 162a stock standard solution and dilute to 500 ml with water. This synthetic standard solution, when further diluted as described in the procedure for each element, corresponds to a standard sample containing 4.08% of copper, 8.52% of nickel and 0.33% of iron. Procedure The amount of sample and the dilution were selected on the basis of the linearity of the working range for each element determined as described earlier.A reagent blank and a standard were run together with the samples. Weigh 0.6 g of the sample in a platinum bowl or a 100-ml polythene (or PTFE) beaker and add 10 ml of 4 N nitric acid. Warm the container for a few minutes on a water-bath, add 3 ml of hydrofluoric acid and mix by swirling. Cover the container with a platinum cover and allow the sample to dissolve completely with gentle heating until brown fumes are no longer emitted. Rinse the cover into the container and allow the solution to cool to room temperature. Add 40 ml of ammonium citrate solution, mix well and add a further 5 ml of ammonia solution. After cooling, dilute the solution to volume in a 500-ml calibrated flask and mix thoroughly. Withdraw aliquots from this solution as follows: (a) Dilute a 10-ml aliquot to 100-ml in a calibrated flask and mix thoroughly.(b) Iron determination. For up to 0.4% of iron use the original solution (500 ml) directly; for 0.4-0.870 of iron, dilute 25 ml from the original solution to 50 ml in a calibrated flask and mix; for 0.8-1.5% of iron, dilute 25 ml from the original solution to 100 ml in a Cali- brated flask and mix. The absorbances for (a) and (b) are read under the instrumental conditions specified above. On mixing finally, the pH is about 5. Copper and nickel determination.January, 1981 Calculation where M = Cu, Ni or Fe, A absorba.nce of standard, B = HEAVY ALLOY BY AAS 49 A - A , Ast - A0 Bst x % M in standard B X - -___ - = absorbance of sample, A , = absorbance of blanks, A,t = amount of sample (mg) and Bst = amount of standard (mg).Results and Discussion Copper Determination A series of samples of heavy alloy were analysed for their copper contents by the volu- metric method and the atomic-absorption spectrophotometric method. A synthetic solution prepared by adding a suitable volume of a standard solution of NBS 162a nickel - copper alloy to a pure tungsten matrix [Analytical standard solution (synthetic) for atomic-absorption spectrophotometric measurements] was used as an analytical standard in the atomic-absorption spectrophotometric method. Some of these samples were also analysed for their copper content by the electrolytic method. The results in Table I show that the agreement between the atomic-absorption spectrophotometric values and the volumetric and electrolytic results is satisfactory. TABLE I COMPARISON OF RESULTS FOR COPPER BY ATOMIC-ABSORPTION SPECTROPHOTOMETRIC, VOLUMETRIC AND ELECTROLYTIC METHODS Copper concentration, yo , 1 Atomic-absorption spectrophotometric Volumetric Electrolytic Sample No.method* method method & I . . 11.. I11 IV v . . VI VII VIII IX x . . XI XI1 . . . . 2.47 . . . . 2.44 .. . , 2.68 .. . . 2.46 . . . . 2.46 . . . . 2.68 . . . . 2.08 . . . . 2.49 . . . . 2.67 . . . . 2.58 .. . . 2.50 - . . . . - 2.46 2.52 - 2.45 - - 2.71 - - 2.46 - - 2.46 - - 2.65 - 2.09 2.01 2.18 2.48 2.42 2.51 2.75 2.71 2.73 2.60 - 2.64 2.58 2.51 2.57 2.48 - 2.57 * Duplicate results represent values obtained on two different The accuracy of the atomic-absorption spectrophotometric method was further tested as follows.A series of synthetic solutions were prepared by adding various volumes of the standard solution of NBS 162a nickel - copper alloy to solutions of pure tungsten prepared as described under Procedure. All of the solutions contained the same amount of tungsten as the recommended amount of sample of heavy alloy would contain, i.e., 0.54g. The solution with a copper concentration at the higher end of the linear working range was used as the analytical standard for calculating the copper content of the other solutions from their absorbance values. The range of copper concentration was chosen so as to cover the entire specified range of copper content of heavy alloy. Table I1 shows the results in terms of the copper content of each solution calculated from its absorbance compared with that actually present.It can be seen that the maximum relative error is about l.OY,, which can be considered accept able. occasions for the same set of samples. Nickel Determination spectrophotometric method for the determination of nickel in heavy alloy. presented in Tables I11 and IV. The entire procedure for copper was repeated for the evaluation of the atomic-absorption The results are50 PABALKAR et al. : COPPER, NICKEL AND IRON IN TABLE I1 Analyst, vol. 106 DETERMINATION OF COPPER I N SYNTHETIC SAMPLES Copper found by Copper atomic-absorption Relative Sample No. presentlmg spectrophotometry/mg error, yo 1 11.77 11.72 0.4 2 13.73 13.64 0.7 3 15.69 15.67 0.1 4 17.65 17.47 1 .o TABLE I11 COMPARISON OF RESULTS FOR NICKEL BY ATOMIC-ABSORPTION SPECTROPHOTOMETRIC AND GRAVIMETRIC METHODS Sample No.I1 IV V VI VI I VIII I X X Nickel concentration, yo A I -l Atomic-absorption spectrophotometric Gravimetric method method 7.60 7.56 7.74 7.78 7.59 7.60 8.04 7.97 6.72 6.83 7.40 7.48 7.73 7.62 7.61 7.49 TABLE IV DETERMINATION OF NICKEL IN SYNTHETIC SAMPLES Nickel found by Nickel atomic-absorption Relative Sample No. presentlmg spectrophotometry/mg error, yo 1 40.98 40.98 0.0 2 43.03 43.50 1.1 3 45.08 45.39 0.7 4 47.13 47.28 0.3 From Table IV it is clear that the maximum relative error for nickel is of the same order as that for copper, i.e., about 1%. Iron Determination How- ever, it can enter the system if it is present in the starting materials as an impurity and through the various production stages.Thus its presence is estimated at not more than about 1.5% in the final alloy. Table V shows the results for the determination of iron in a series of heavy alloy samples carried out both by atomic-absorption spectrophotometry and spectrophotometry. In the present instance, iron is not a major constituent of the heavy alloy system. TABLE V COMPARISON OF RESULTS FOR IRON BY ATOMIC-ABSORPTION SPECTROPHOTOMETRIC AND SPECTROPHOTOMETRIC METHODS Iron concentration, yo A I 1 Atomic-absorption Spectrophotometric Sample No. spectrophotometric method method VI I 0.64 0.65 VIII 0.62 0.62 I X 0.65 0.66 X 0.24 0.27 XI 0.24 0.26 XI1 0.24 0.27January, 1981 HEAVY ALLOY BY AAS 51 As with copper and nickel, the analytical standard used for both the atomic-absorption spectrophotometric and the spectrophotometric determination of iron was a synthetic solution containing pure tungsten and a suitable volume of a standard solution of NBS 162a nickel - copper alloy.Effects of Tungsten Concentration and Reagent Concentration Investigations were carried out to ascertain the matrix effect, i.e., the effect of variation of tungsten concentration on the absorbances of copper, nickel and iron, and to study the effect of variation of the concentration of the reagents (nitric acid, hydrofluoric acid and ammonium citrate) on the absorbances of these elements. The variations were kept within reasonable limits. E$ect of tungsten on copper, nickel and iron determinations Different amounts ranging from 400 to 700 mg of a sample of pure tungsten metal were taken.The tungsten was free from copper and nickel but contained 0.048y0 of iron. The solutions were prepared as described under Procedure but before making up to 500 ml appropriate volumes of standard solutions of copper, nickel and iron were added so as to give, on further dilution, the final concentration of the elements as shown in Fig. 1. Absorbances for the individual elements were read under the instrumental conditions specified in the method. The absorbance of iron was corrected for the iron content of the tungsten metal. I t can be seen from Fig. 1 that the absorbances of copper and nickel are unaffected by variations of tungsten concentration whereas that of iron is affected above a certain level of tungsten.This indicates that for copper and nickel determinations it is not necessary to match the sample with the standard with respect to its tungsten content, but for iron deter- mination such matching is necessary. Two series of standards were prepared by taking 0-, lo-, 20-, 30-, 40- and 50-ml aliquots of the stock standard solution in each series. One series consisted of the reagents added as in the pro- cedure. The other series consisted of solutions with tungsten added as in the case of the analytical standard solution. Thus the two series were identical except for tungsten content in one instance. After correcting for the reagent blank, the signal versus concentration graphs were all linear and passed through the origin (Fig. 2). For copper and nickel the corresponding absorbances in both the presence and absence of tungsten were the same, giving rise to only one straight line graph each and confirming the noninterference of tungsten.Even with iron the error due to the slight deviation was only about 5%. The absence of any background effect was further confirmed as follows. All of the solutions were made up to 500 ml. 0.110 0.100 2 0.090 0.080 a , I) 0.070 1 " " -$ \ a 0.060 V " C - 0.050 /- , , , , , , , 1 0.040 400 450 500 550 600 650 700 Mass of tungsten in original solutionimg Fig. 1. Effect of variation of tungsten concentra- tion on absorbance of: A, copper (3.0 pg ml-l) ; B, iron (2.8 pg ml-l); and C , nickel (9.0 pg ml-l).52 $ 0.070 0.060 PABALKAR et al. : COPPER, NICKEL AND IRON IN Analyst, VoZ.106 - C - * - A v v I 1 1 , I I I 0.220 r 8 0.140 a 0.120 -fl 2 0.100 9 0.080 0.060 0.040 0.020 C 0 1 2 3 4 5 6 7 8 9 10 11 Concentration/pg ml-' Fig. 2. Calibration graph for copper, nickel and iron in heavy alloy showing absence of background effect: A, copper; B, iron with background; C, iron; and D, nickel. Efect of nitric acid, hydropuoric acid and ammonium citrate concentrations on copper, nickel and iron determinations Three separate sets of experiments were carried out to study the effect of nitric acid, hydrofluoric acid and ammonium citrate concentrations. In each set, several 600-mg portions of a sample of heavy alloy containing 2.60% of copper, 7.61% of nickel and 0.24% of iron were weighed out and treated with various volumes of one of the reagents while keeping those of the others constant as described under Procedure.The dilutions were completed and absorbances read as before. Figs. 3, 4 and 5 show the effect of nitric acid, hydrofluoric acid and ammonium citrate concentration, respectively, on the absorbances of copper, nickel and iron. It can be seen that the absorbances of copper and nickel were unaffected by the concentration of all three reagents. The absorbance of iron also remained unaffected by ammonium citrate but increased with the increasing concentrations of nitric acid and hydrofluoric acid. This result shows that in the determination of iron matching of the standard and sample solutions with respect to reagent concentrations is necessary. Conclusions The method is simple and rapid and the accuracy is comparable to those of existing volu- metric, gravimetric and spectrophotometric methods.On a day-to-day basis the savings on chemicals and^ man hours are substantial. 0.120 A I t e : @ ~ 0.110 2 0.100 s 0.090 2 0.080 2 0.070 0.060 - v 0 0.050 ' ' 1 I 5 8 11 14 17 20 Volume of 4 N nitric acidiml Fig. 3. Effect of variation of nitric acid con- centration on absorbance of: A, copper (3.1 pg ml-1) ; B, iron (2.9 p g ml-l); and C, nickel (9.1 pg ml-1). A 0.120 0.1 10 V V 7 V " V Volume of 40% m/m hydrofluoric acid/ml Effect of variation of hydrofluoric acid concentration on absorbance of: A, copper (3.1 p g ml-l) ; B, iron (2.9 p g ml-l); and C, nickel (9.1 pg ml-l). Fig. 4.January, 1981 0.120 0.110 8 0.100 5 0.090 9 0.080 2 0.070 <0.060- 0.050 HEAVY ALLOY BY AAS - A - v 7 7 7 7 . v - M 0 2 v 0 - B , - * “ - C 1 I I 1 I I I 53 Fig. 5. Effect of variation of ammonium citrate concentration on absorbance of : A, copper (3.1 pg n11-l) ; B, iron (2.9 pg ml-l) ; and C, nickel (9.1 pg ml-l). The determination of copper and nickel is not affected by the tungsten matrix or reagent concentrations but the determination of iron requires careful control of these parameters. The authors are grateful to Mr. William Ohlsson, Director R&D, Sandvik AB, Stockholm, and to hlr. Borje Skog, Managing Director, Sandvik Asia Limited, Poona, for their interest in the progress of this work and permission to publish this paper. References 1. 2. 3. 4. 5. McLennan, J. C., and Smithells, C. J., J . Sci. Instrum., 1935, 12, 159. Larsen, E. I., and Murphy, P. C., Can. Min. Metall. Bull., 1965, 58, 413. Yih, S. W. H., and Wang, C. T., “Tungsten-Sources, Metallurgy, Properties, and Applications,” Sanjana, N. R., and Naik, S. V., Sandvik Asia Limited, unpublished work. “Analytical Methods for Atomic Absorption Spectrophotometry,” Perkin-Elmer Corp., Norwalk, Plenum Press, London, 1979, p. 361. Conn., 1973. Received May 6th, 1980 Accepted July 22nd, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600047

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

54 Analyst, January, 1981, Vol. 106, PP. 54-59 Nebuliser System for Analysis of High Salt Content Solutions with an Inductively Coupled Plasma Bo Thelin Swedish Institute for Metals Research, Drottning Kristinas vag 48, 114 28 Stockholm, Sweden A nebuliser system for analysis of high salt content steel samples using an inductively coupled plasma is described. Because of its design this nebuliser system is capable of continuous operation with solutions of up to 10% salt content without clogging. The fabrication of the nebuliser was effected in a new and simple way. The effects on line to background ratios for several element lines which increased in intensity with higher salt concentrations were studied. Comparisons were also made with a conventional nebuliser system; the new system gave higher line to background ratios for the same salt concentrations. Analytical results for various amounts of steels con- taining trace and normal contents of other elements are discussed.For trace contents the relative standard deviations of the determinations were improved when salt concentrations up to 10% were used. No memory effects were noticed and the nebuliser system had a rapid clean-out time. Analysis was carried out using a computerised image dissector kchelle spectrometer. Keywords : Inductively coupled plasma ; nebuliser systems ; steel analysis ; amage dissector kchelle spectrometry ; trace metal analysis Inductively coupled plasmas (ICPs) have aroused increased interest in recent years as components of analytical systems, mainly because they offer good detection limits and linear calibration graphs for several orders of magnitude.One of the major problems with the ICP has been the absence of a totally satisfactory nebulising device, especially for solutions that have a high salt content. Two of the most popular types of nebulisers used are the cross flow1 and the concentric glass nebulisers.2 Both devices rely on capillary tubes in their design to generate aerosols using low flow-rates of argon gas. Although these nebulisers are capable of efficient aerosol generation, capillary tubes sometimes become partially or completely clogged, especially when analysing high salt content samples. They therefore require frequent cleaning to restore the original rate of aerosol generation or else restandardisation of the ICP is required.Application of the Babington nebuliser to the analysis of high salt content samples by flame atomic-absorption spectrometry has been rep~rted.~ The Babington design has great potential for the analysis of solutions with high percentages of dissolved and suspended solids. Basically it consists of a glass surface with an opening cut into it. A liquid flowing over the surface and across the opening is converted into a fine spray by pressurised air flowing through the opening. As the aspirated solutions are introduced on the outside of the nebuliser, there is very little chance of clogging due to suspended particles or high viscosity. In 1978 Suddendorf and BoyeI"1 described a slot-type nebuliser, which could be used for the analysis of high salt content solutions with an ICP.Their design consisted of a Plexiglas base, gold-plated stainless-steel block with a V-groove and a small hole, a sample feed tube as shield and an impactor rod. The flow-rate of the sample solution was 5 ml min-1 and the argon flow-rate was 1 1 min-1. Compared with a cross-flow nebuliser, improved precision was obtained with this slot-type nebuliser. This nebuliser also showed detection limits corn- parable to the cross-flow nebuliser and was also free of memory effects. This nebuliser was made of glass and was based partially on the Babington design, and was capable of aspirating solutions with suspended solids and solutions with high concentrations of dissolved solids. The device consisted of a PTFE base, a nebuliser tube and a capillary tube used to deliver solution to the nebuliser from a peristaltic pump.The heart of the nebuliser was a V-groove, channel and through-hole which were gromd into the sealed end of a glass tube. In the same year Wolcott and Sobe15 described another slot-type nebuliser.THELIN 55 Fabrication was simple, required very little machine time and was performed with a diamond drill. The nebuliser usually operated with an argon flow-rate of 0.5-0.8 1 min-l and a sample flow-rate of about 0.5-1.5 ml min-l. Improved detection limits were obtained compared with a cross-flow nebuliser (Plasma Therm). This slot-type nebuliser was used on a routine basis with an ICP system analysing several types of solution containing about 20% of dissolved solids.In this work we examined the combination of a slot-type nebuliser with a cyclone spray chamber6 when analysing solutions containing high concentrations of dissolved solids (steels). The design of the nebuliser was similar to that of Wolcott and S ~ b e l , ~ but was fabricated in a different way. One objective with this combination was to determine trace elements in steels with improved detectability and reproducibility. Experimental The ICP unit used in this work was a Plasma Therm (2.5 kW) with an automatic tuning The plasma unit was mounted on a mobile table and could be moved both hori- The optimised experimental para- system. zontally and vertically in relation to the optical axis. meters for this ICP are given in Table I. TABLE I OPTIMISED EXPERIMENTAL PARAMETERS Slot-type nebuliser with cyclone spray Experimental parameter chamber PowerlkW .. . . .. .. 1.2 Argon cooling gas flow-rate/l min-1 . . Nebuliser flow-rate of argon/l min-l. . Sample uptake ratelm1 min-I . . 3 Henry frequency/MHz . . .. 27.12 12 7 Nebuliser pressure/Pa .. . . 131 Observation height/mm . . . . 15 Concentric nebuliser with Scott spray chamber 1.2 27.12 12 1 48 0.8 15 Concentric nebuliser with cyclone spray chamber 1.2 27.12 12 1.4 83 1.2 15 In this paper a simple glass nebuliser, based partially on the Wolcott and Sobel design and capable of aspirating high salt concentrations of steel samples, is described. The heart of the nebuliser is a V-groove, channel and through-hole which are cut into the sealed end of a glass tube. The V-groove and the channel are cut with sharp knives while the glass is still hot and the through-hole is made by puncturing the hot glass with a fine tungsten wire (0.15 mm diameter).This method is a simple and cheap means of fabricating such nebulisers (they are also available from Wicklunds Glasinstrument, Idungatan 7, S-113 45 Stockholm). The sample solution is delivered to the nebuliser via a glass capillary (0.8 mm i.d.) which is connected to a Gilson Mini Pulse I1 peristaltic pump. As shown in Fig. 1 the capillary is placed inside a plastic tube, which directs it into the slot. TABLE I1 COMPARISON OF LINE TO BACKGROUND RATIOS FOR DIFFERENT NEBULISER SYSTEMS Sample concentration = 1 g per 100 ml. Element Copper . . Chromium. . Manganese Cobalt . . Nickel . . Aluminium Molybdenum Line wavelength/ nm .. 324.8 . . 425.4 .. 403.5 . . 345.4 . . 341.5 . . 396.2 .. 386.4 Slot-type nebuliser with cyclone spray chamber 100 185 143 22 31 3.4 7.7 Concentric nebuliser with Scott spray chamber 29 35 27 1.9 7.5 3.1 13 Concentric nebuliser with cyclone spray chamber 71 110 80 2.5 10.7 16 5.056 THELIN: NEBULISER SYSTEM FOR HIGH Analyst, Vol, 106 The spray chamber used was a cyclone type according to Greenfield6 (Fig. 1). This type of spray chamber has advantages for use with high salt concentrations because it directs the small drops from the aerosol into the plasma without using obstructing glassware. Because of the centrifugal force only small drops will remain in the middle of the chamber. The big drops will collide with the walls and run down into the drain.Because of the considerable argon flow in this chamber an over-pressure will be built up in the lower part of the chamber, which will cause the small drops in the middle to drift into the plasma. From the top of the spray chamber a 65 cm long and 16 mm wide (i.d.) plastic tube was connected to the torch in a suitable way. This long plastic tube further improves the process for selecting small drops and it will also improve the stability of the aerosol flow into the plasma. The analytical system used was the image dissector dchelle spectrometer (IDES) system,’ which belongs to the new generation of spectrometers with photoelectric recording. This system, which is very versatile, consists of an dchelle spectrograph with high resolution, an image dissector tube and a minicomputer system.Cyclone spray chamber I I I I Solution from I I 750. Conical Slot-type nebuliser \ . Ar Silicone Teflon tube rubber Glass grinding Teflon holder Glass capillary Drain Fig. 1. The slot-type nebuliser in combination with the cyclone spray chamber. Results and Discussion The efficiency of the nebuliser system was calculated by aspirating a known amount of water through the nebuliser with the cyclone spray chamber attached. The drain-off liquid was collected and measured and the efficiency of the nebuliser calculated. For the slot-type nebuliser the efficiency was 8% with a sample uptake rate of 3 ml min-l. The uptake rate of 3 ml min-l is of the same order of magnitude as that for most atomic-absorption nebulisers. In this table, line to background ratios, at the same concentration (1 g per 100 ml), were calcu- lated for different spectral lines, when NBS 363 was aspirated in the ICP system, In this A comparison was also made between three different nebuliser systems (Table 11).January, 1981 SALT CONTENT SOLUTIONS WITH ICP TABLE I11 57 COMPARISON OF LINE TO BACKGROUND RATIOS FOR DIFFERENT AMOUNTS WITH THE SLOT-TYPE NEBULISER AND CYCLONE SPRAY CHAMBER Sample = NBS 363.Element Chromium. . . . . . Cobalt . . . . . . Nickel - . . . . . Aluminium . . . . Copper . . . . . . Manganese . . . . Line wavelength/nm 324.8 425.4 403.5 345.4 341.5 396.2 Sample concentrationlg per 100 ml 1 5 10 100 420 792 185 667 1631 143 540 1269 3.4 13 24 22 103 203 31 111 240 r A \ TABLE IV RESULTS OF THE DETERMINATION OF VANADIUM, NICKEL AND CHROMIUM IN STEELS Results (2) are in % m/V and are the means of 10 determinations in each instance, with relative standard deviation (RSD).Sample concentrationlg per 100 ml r A \ Iron 1 5 10 7 - - z i F y : wavelength Reference Sample Element determined used/nm value RSD, /! NBS 361.. . . Vanadium at 309.3nm 357.0 0.011 0.012 2.8 0.011 1.7 0.011 1.8 362.. . . 0.040 0.038 0.9 0.038 0.9 0.040 0.7 363 .. . . 0.31 0.31 1.2 0.31 1.2 0.30 2.6 364.. _ . 0.105 0.105 1.5 0.106 0.8 0.106 1.9 J K 8 F . . .. 0.022 - - 0.22 1.9 0.22 1.3 NBS 361 . . . . Nickel at 341.5 nm 344.1 2.00 0.7 2.0 1.7 2.0 2.1 362.. . . . 0.59 0.60 2.0 1.2 0.60 1.7 0.60 2.3 363.. . . 0.30 0.31 1.1 0.32 0.7 0.28 1.1 364.. .. 0.144 0.139 2.4 0.141 1.5 0.144 2.0 365.... 0.041 0.040 1.9 0.040 2.2 0.040 2.5 NBS 361 . . . . Chromium at 425.4 nm 386.0 0.69 0.66 1.5 0.67 1.3 0.69 2.3 362.. . . 0.30 0.29 1.4 0.33 2.6 0.25 2.9 363.. . . 1.31 1.32 1.1 1.31 1.4 1.37 2.5 364.. .. 0.063 0.062 2.7 0.063 3.2 0.061 2.4 365.. .. 0.0070 - - - - 0.0069 3.9 TABLE V RESULTS OF THE DETERMINATION OF MANGANESE, MOLYBDENUM AND COPPER IN STEELS Results (5) are in yo m/V and are the means of 10 determinations in each instance, with relative standard deviations (RSD). Sample concentration/g per 100 ml Sample NBS 361 .. .. 362.. .. 363.. .. 364.. .. 365.. .. NBS 361.. . . 362.. .. 363.. .. 364.. .. 365.. .. NBS 361.. .. 362.. .. 363.. .. 364.. _ . 365.. .. JK8F .. .. Iron wavelength Reference Element determined used/nm value Manganese a t 403.4 nm 392.3 0.66 1.04 1.50 0.255 0.006 0 Molybdenum at 386.4 nm 386.0 0.19 0.068 0.028 0.49 0.005 0 Copper a t 324.8 nm 516.7 0.042 0.50 0.10 0.249 0.005 8 0.052 3 7 I 0.65 1.07 1.53 0.23 0.19 0.068 0.030 0.49 0.006 2 0.042 0.51 0.10 0.25 0.006 1 0.052 6 - 1 Y RSD, % 3.5 1.8 2.7 3.0 2.2 2.4 9.4 1.4 37.3 1.2 1.9 1.7 1.4 4.5 1.9 - 7 I 0.67 1.05 1.52 0.248 0.005 9 0.19 0.068 0.030 0.50 0.005 8 0.044 0.51 0.10 0.251 0.0060 0.052 6 5 - RSD, % 2.0 1.1 1.6 2.0 6.0 2.5 0.6 1.2 2.0 3.3 1.2 2.8 1.2 1.1 1.7 2.2 10 - 0.72 1.4 1.03 2.2 1.52 1.7 0.248 2.7 0.0059 5.8 0.19 1.6 0.069 1.0 0.029 1.1 0.50 2.0 0.0050 3.3 0.044 1.3 0.51 1.9 0.10 1.7 0.252 0.8 0.0059 1.8 0.0526 1.458 THELIN: NEBULISER SYSTEM FOR HIGH Analyst, Vol.106 investigation the ICP system was run with the experimental parameters optimised (Table I).For every nebuliser system run, height optimisation was made for each element. The average values for the optimised heights were calculated for these elements. These values were the same for the various nebuliser systems. Table I1 shows that this slot-type nebuliser in combination with the cyclone chamber gave the highest line to background ratios. A comparison was also made between the Scott spray chamber and the cyclone spray chamber. Table I1 shows that the cyclone chamber is superior to the Scott chamber by a factor of 1.5-2.5. This table shows that the line to background ratio increases by about the same ratio as the increase in mass when different sample masses of NBS 363 are taken. In Tables IV-VI results are shown for some steel samples analysed with this nebuliser system when the optimised parameters are used. Some reference samples from the National Bureau of Standards and standard JK 8F from the Swedish Institute for Metals Research were used.In these analyses vanadium, nickel, chromium, manganese, molybdenum, copper, cobalt, aluminium and titanium were determined from trace to normal levels at steel con- centrations of 1,5 and 10 g per 100 ml. The results show that the relative standard deviations for the trace contents seem to decrease and the precision is also improved when higher concentrations are used, and for the normal contents, the relative standard deviations are about the same as for these masses. These facts indicate that this nebuliser system gives better results for trace contents with higher concentrations of dissolved solids.In the measurements and analyses using the IDES spectrometer system, a computer program for peak measuring was used. With this program every spectral line is measured for 0.1 s on the top at each integration. By iterating 15 times each spectral line is measured for 1.5 s, which is sufficient time to provide accurate results with this spectrometer system. In the steel analyses various iron lines were used as internal standards and the argon 435.8-nm line was used for line adjustment of the spectrum. To study the memory effects in this nebuliser system, NBS 363 (10 g per 100 ml) was aspirated for 5min with a constant nebulisation. After that time a blank solution was aspirated and the measurement of the spectral lines began. This is illustrated in Table VII, where the signals from chromium at 425.5 nm, manganese at 403.5nm and nickel at 341.5 nm are completely eliminated after about 50 s.This shows that the cyclone chamber is washed out very quickly because of its clean glass surfaces and that the small drops in the aerosol are selected without colliding with any glassware. For smaller amounts and lower concentrations the washout times will be shorter. The line to background ratio was also studied for different amounts (Table 111). TABLE VI RESULTS FOR THE DETERMINATION OF COBALT, ALUMINIUM AND TITANIUM IN STEELS Results (2) are in yo m/V and are the means of 10 determinations in each instance, with relative standard deviations (RSD). Sample concentration/g per 100 ml L I 7 Iron 1 5 10 RSD,%' wavelength Reference r-A-, ,Tb7 Sample Element determined used/nm value L RSD, % RSD, /o NBS 361 .. . . Cobalt at 345.4 nm 375.0 0.030 0.022 6.0 0.031 1.3 0.038 3.3 362.. _ . 0.30 0.295 2.0 0.30 2.3 0.30 2.8 3 6 3 . . .. 0.049 0.051 2.1 0.049 1.6 0.051 2.0 364.. . . 0.15 0.15 1.9 0.15 1.7 0.15 1.4 365 .. .. 0.0070 0.0080 17.8 0.0071 1.66 0.0071 1.2 JK8F .. . . 0.125 0.126 1.7 0.124 1.2 0.126 2.2 NBS 361.. . . Aluminium at 396.2nm 386.0 0.021 0.021 3.2 0.020 2.3 0.022 1.5 362.. .. 0.095 0.095 1.7 0.095 1.0 0.096 1.0 363.. .. 0.24 0.24 2.9 0.24 2.8 0.24 0.9 364.. .. 0.014 0,015 3.9 0.014 1.5 0.014 3.5 JKSF .. ._ 0.0026 - - 0.0026 5.1 0.0026 2.6 NBS 361.. . . Titanium at 334.9nm 344.1 0.020 0.020 3.5 0.021 1.8 - - - - - - - - 362.... 0.084 363.. .. 0.05 0.052 3.1 0.05 1.6 364.. .. 0.24 0.24 2.1 0.24 1.6 - - - -January, 1981 SALT CONTENT SOLUTIONS WITH ICP 59 TABLE VII MEMORY EFFECTS FOR CHROMIUM, MANGANESE AND NICKEL IN THE NEBULISER SYSTEM Concentration of steel solution = 10 g per 100 ml. The steel used contained 1.31% of chromium, 1.50% of manganese and 0.30% of nickel. Counts per 0.1 s Time/s 0 14 28 42 56 70 84 r Chromium 3 559 2 255 580 27 1 228 197 229 Manganese 1853 1165 418 247 181 183 180 Nickel 1235 739 232 137 112 117 119 Conclusion of advantages for the ICP technique were obtained : By using this slot-type nebuliser in combination with the cyclone spray chamber a number (a) No clogging problems appeared, which means that there are no problems in aspirating solutions with high salt concentrations, up to at least 10%.( b ) Line to background ratios at higher salt concentrations were increased. (c) The relative standard deviations of the analyses of steel samples were improved for the trace contents, with increased salt concentration. (d) No memory effects were observed and the wash-out time was short. (e) By using a peristaltic pump a constant flow of sample was obtained with no viscosity problems. The only disadvantages of this new nebuliser system are higher consumptions of gas (40% more) and sample (a factor of three more) compared with the concentric nebuliser in combina- tion with the Scott spray chamber. Compared with the consumptions of gas and sample required by the high-power ICP technique the consumptions of the new nebuliser system are insignificant. This work was performed at the Swedish Institute for Metals Research as a project from the Research Department at Jernkontoret and financed by Jernkontoret and STU. The author gratefully acknowledges the helpful criticism and encouragement received from Dr. Lars Danielsson during the preparation of this paper. He is also indebted to Mrs. Ingrid Eklof for running the ICP - IDES system and for the sample preparation. He thanks the Chairman of the trace element committee, Mr. Valtter Palvarinne, Uddeholms AB, for valuable criticism and for showing interest in this project. The author also thanks Wicklund Glasinstrument, Idungatan 7, S-113 45 Stockholm, for the fabrication of the glassware of the nebuliser system. References 1. 2. 3. 4. 5. 6. 7. Kniseley, R. N., Amenson, H., Butler, C. C., and Fassel, V. A., AppZ. Spectrosc., 1974, 28, 285. Meinhard. J. E., ICP Inf. NewsZ., 1976, 2, 163. Fry, R. C., and Denton, M. B., Anal. Chem.. 1977, 49, 1413. Suddendorf. R. F., and Boyer, K. W., Anal. Chem., 1978, 50, 1769. Wolcott, J. F., and Sobel, C . B., AppZ. Spectrosc., 1978, 32, 591. Greenfield, S., Invited Lecture, XXI Colloquium Spectroscopicum Internationale (8th International Danielsson, A.. and Lindblom, P., AppZ. Spectrosc., 1976, 30, 151. Conference on Atomic Spectroscopy), Cambridge, 1979. Received June 16th. 1980 Accepted August 5th, 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600054

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

60 Analyst, January, 1981, Vol. 106, j?+. 60-68 Spectrophotometric and Fluorimetric Determination of Tri- and Di-organotin and -organolead Compounds Using Dithizone and 3-Hydroxyflavone W. N. Aldridge and B. W. Street Molecular Toxicology Section, Toxicology Unit. MRC Laboratories, Woodmansterne Road, Carshalton, Surrey, S M 5 4EF For the spectrophotometric determination of organotins and organoleads using dithizone, a suitable buffer system has been devised with adequate buffering capacity a t pH 9, containing EDTA to complex interfering inorganic metals and a high concentration of sodium perchlorate suitable for increasing the distribution of the lower organotins in favour of the solvent. Tri- organotins and triorganometals form complexes with a ratio of dithizone to organometal of 1 : 1 ; with diorganotins and diorganoleads the corresponding ratio is 2 : 1.Whether in practice this theoretical ratio is achieved depends upon the distribution coefficient (chloroform : buffer) and the affinity constant of the organoinetal for dithizone. Under our experimental conditions the affinity constant of triethyltin for dithizone is 3 x lo5 1 mol-l. Of a large series of tri-organotins and -organoleads, only trimethyltin and triphenyltin form fluorescent complexes with 3-hydroxyflavone (excitation 400nm, emission 514nm). All of the other organotins and organoleads form complexes with absorbance maxima at 385 nm. Dimethyltin, although not forming a fluorescent complex, interferes in the determination of tri- methyltin by quenching its fluorescence. By the measurement of fluores- cence emission at 514 nm and absorbance a t 385 nm mixtures of dimethyltin and trimethyltin can be determined.Keywords : Absorbance spectrophotometry ; dithizone ; jluovescence spectrophoto- metry ; 3-hydroxyflavone ; di- and tri-organotins and -organoleads Organotin and organolead compounds are biologically active at low concentrations.l Many are in use commercially as pesticides, fungicides, stabilisers of plastics, etc., and they are also used for the study of several biological systems such as energy conservation by mitochondria and cellular functions in the brain. A recent observation that trimethyltin is toxic to man2 and that both trimethyltin and triethyllead produce in rats selective damage to certain areas in the brain314 has stimulated the examination of methods for their determination in bio- logical materials.There are several reagents available for the direct determination of organotin compounds (for a review, see ref. 5). Dithizone (diphenylthiocarbazone) has been used for the deter- mination of di- and trieth~ltin,~,’ triethylleada and other homologues.9 A fluorimetric method using 3-hydroxyflavone for the determination of triphenyltinlo has also been shown to be capable of development for the determination of other homologues.11 In this paper, the applicability of dithizone and 3-hydroxyflavone for the determination of organotin and organolead compounds is examined and the conditions necessary for quantitative reaction are defined. Experimental Materials Chemicals were obtained from the following sources : diethanolamine and perchloric acid (60% m/m), from BDH Chemicals, Poole, Dorset ; dithizone and 3-hydroxyflavone from Eastman Kodak, Kirkby, Liverpool; trimethyltin chloride from Aldrich Chemical Co., Milwaukee, Wisc., USA; trimethyllead acetate, triethyllead acetate and triphenyllead acetate from Alfa Products, Beverley, Mass., USA ; tripropyltin acetate, tributyltin acetate, octyl- dimethyltin acetate, triphenyltin acetate, phenyldiethyltin bromide, ethyldiphenyltin acetate, t riprop yllead acetate, t ribut yllead acetate, diet h yltin dichloride, diprop ylt in dichloride, di but yl tin dichloride, dimet hyllead diace tate, diet hyllead diacet ate, diprop y llead diace t ate,ALDRIDGE AND STREET 61 dibutyllead diacetate and diphenyllead diacetate from the Institute for Organic Chemistry, TNO, Utrecht, The Netherlands; and dimethyltin dichloride from Metal and Thermit Corp., Rahway, N.J., USA. Dimethyltin dichloride and trimethyltin chloride were recrystallised from light petroleum (boiling range 40-60 "C) .Triethyltin hydroxide, supplied by the Tin Research Institute, Greenford, Middlesex, was purified as follows. The hydroxide (10 g ) was dissolved in chloroform (60 ml) and the insoluble material was removed by filtration. The chloroform layer was washed twice with 0.1 M sodium hydroxide solution (25 ml) and the aqueous layers were discarded. Acetone was added to 200 ml and the hydroxide was converted into the sulphate as previously described. l2 Sodium [36Cl]chloride (3 mCi g1 C1) and [l13Sn] triethyltin chloride (approximately 5 mC mmol-1) was obtained from the Radiochemical Centre, Amersham, Buckinghamshire.Preparations of triethyltin as purchased often contain impurities that are probably tetra- ethyltin and diethyltin dichloride. A suitable amount of an ethanolic solution of [ll3Sn]triethy1tin chloride was added to 0.02 M nitric acid (5 m1) and the tetraethyltin then removed by extracting with benzene (5 ml). This was carried out in a round separating funnel held on its side and using a magnetic stirrer. The bottom, aqueous layer was delivered into another separating funnel containing 0.04 M sodium hydroxide solution (5 ml) and then extracted in a similar manner with chloroform (10 ml). By this procedure triethyltin hydroxide was extracted into the chloroform layer, leaving the diethyltin in the aqueous layer.The bottom, chloroform layer was delivered into another separating funnel containing 0.02 M nitric acid (5 ml) to transfer the extracted triethyltin into the aqueous layer. The funnel was gently rinsed with chloroform (5 ml), the aqueous layer delivered into a glass-stoppered tube and any remaining chloroform removed by gently bubbling air through the solution. The purified [l13Sn]triethyltin can be stored in the dilute nitric acid solution. As the triethyltin chloride and the impurities will contain the same specific radioactivity due to l13Sn and the relative molecular masses of tetraethyltin and diethyltin dichloride are very similar, the molarity of the purified compound was determined as follows.The molarity of the original solution in ethanol was known from the mass of triethyltin chloride supplied. The radioactivity of this solution and of the solution after purification was determined and the molarity of the latter calculated in proportion to the measured counts. Distribution coefficients (chloroform : buffer, pH 9, containing 52.9 g of diethanolamine, 1.0 g of EDTA, disodium salt, and 24 ml of 60% m/m perchloric acid in 500 ml) for triethyltin and diethyltin are 7.75 and 0.008, respectively. They can be purified as follows. After stirring, the chloroform layer was removed. Radioactivity Measurements meter, taking account of the precautions necessitated by the decay process.13 Radioactive counting was carried out with a Packard Autogamma scintillation spectro- Optical Measurements fluorescence using a Perkin-Elmer MPF 3 fluorescence spectrophotometer. Absorbance measurements were made with a Unicam SP 600 spectrophotometer and Reagents Bufer.Diethanolamine (52.5 g), sodium perchlorate monohydrate (310 g) and EDTA disodium salt (1.Og) were dissolved in water and 60% m/m perchloric acid (37.5 ml) was added. The solution was diluted with water to 500 ml. 3-Hydroxy$avone, 0.01 yo (420 PM) solution in toluene. Dithizone. The pH should be 9.0. A 0.03% solution of dithizone in chloroform was stored in the dark at 5 "C and diluted 10-fold with chloroform to 117 p~ immediately before use. The molar absorp- tivity of dithizone in chloroform was measured and found to be 4.02 x lo4 mol 1-1 cm-l at 610nm.62 ALDRIDGE AND STREET: ORGANOTINS AND ORGANOLEADS Analyst, vol.106 Procedures Reaction of organometals with dithizone To 8 ml of buffer were added various amounts (0-100 p1) of organometal in water or in dimethylformamide. Chloroform (5 ml) and dithizone (1 ml) in chloroform were added and the glass-stoppered tubes shaken vigorously for 10 s. After centrifugation the aqueous layer was removed and the absorbance of the chloroform layer at 610 nm in 2-cm cells was compared with that obtained in the absence of organometal. Reaction of orgartometals with 3-hydroxyjavone To 7 ml of buffer were added various amounts (0-100 p1) of organometal in water or in dimethylformamide. 3-Hydroxyflavone in toluene (7 ml) was added and the layers were mixed vigorously with a mushroom-ended glass rod.After brief centrifugation the fluorescence of the toluene layer was measured using procedure A (see Table I). For high sensitivity, procedure B (Table I) was used. TABLE I RELATIONSHIP BETWEEN THE FLUORESCENCE AND AMOUNT OF TRIMETHYLTIN USING 3-HYDROXYFLAVONE For method see Procedures. For procedure A excitation at 400 nm (slit width 6) and emission at 514 nm (slit width 6; sensitivity 1) were used. For procedure B, slit width 7 and slit width 8, sensitivity 10, were used for excitation (400 nm) and emission (514 nm), respectively. Procedure A- Trimethyltinlnmol . . .. . . 30.5 61 91.6 122.2 Fluorescence, units . . .. . . 27.7 53.7 80.2 96.5 Procedure B- Trimethyltin/nmol . . .. . . 0.61 1.22 1.83 2.44 3.05 Fluorescence, units .. .. .. 13 25.5 39.5 52 63 For both of the above procedures identical results were obtained whether the organometal was added in water or dimethylformamide. The maximum amount of dimethylformamide added was 100 pl, so that its maximum possible concentration in the chloroform or toluene would be approximately 1 .ti%. Results and Discussion Composition of Buffer for the Methods Using Dithizone and 3-Hydroxyflavone Preliminary experiments indicated that the method using dithizone for the determination of triethyltin6 was not suitable for trimethyltin. Further work has shown that the reasons for the insensitivity of the method for the determination of trimethyltin are probably a combination of its high partition into the aqueous medium and, owing to the low affinity of dithizone for trialkyltins (see later), the inability of the organic solvent to extract the dithizone - trimethyltin complex.Several conditions for the aqueous buffer must be met. Ideally, when trialkyltins are to be determined in biological material, the buffer must be adequate to neutralise the acids used to remove proteins: The buffering capacity must be effective at about pH 9 because at this pH the organometal will be carried into the organic solvent as the hydr0xide.1~1~~ As many reagents for organotins also react with many in- organic metals, a chelating agent is required that is capable of selectively complexing such metals in the presence of organotins and organoleads. EDTA was used in the original method using dithizone5 to chelate inorganic metal ions while leaving triethyltin and diethyltin free to complex with dithizone.In order to partition trimethyltin hydroxide into the organic phase, a highly soluble salt is required whose anion does not complex with the organotin, a restriction which also applies to the buffer. Experiments with [113Sn]triethyltin and 36Cl- have shown that 1 M diethanolamine buffer at pH 9 is a suitable buffering system and that high concentrations of sodium nitrate or sodium perchlorate increase the partition of triethyltin into the solvent. In practice weJanuary, 1981 BY SPECTROPHOTOMETRY AND FLUORIMETRY 63 have used sodium perchlorate to avoid the possibilities of reduction of nitrate to nitrite and reaction with amines to form carcinogenic nitrosamines in the laboratory effluent. Other experiments have shown that, using this aqueous medium, the calibration graph for tri- methyltin, the most water-soluble of the trialkyltins, approaches that for other triorganotin and triorganolead compounds ; sodium perchlorate therefore increases the partition of tri- methyltin in favour of the chloroform (cf., Fig.1). 1 .o 0.8 0 Q n n Q 0.6 a ' 0.4 0.2 0 0 A 0 A A I 20 40 60 80 100 120 Amount of trialkylmetal compound/nmol Fig. 1. Relationship between absorbance of dithizone at 600 nm in 2-cm cells and concentration of trialkyltins and triethyllead. The solid line indicates a molar ratio (trialkylmetal to dithizone) of 1: 1. For details of method see Procedures. @, Trimethyltin (8 ml of buffer); A, trimethyltin (1 ml of buffer); A, trimethyltin (8 ml of buffer containing no sodium perchlorate) ; 0, triethyltin ; x , tripropyltin; 0, tributyltin; and ., triethyllead.Determination of Organotins and Organoleads Using Dithizone The absorbance spectra of the triethyltin - and diethyltin - dithizone complexes have maxima at 440 and 485nm and no absorbance at 610nm, the absorbance maximum for dithizone.6 Other organotins and organoleads have been shown in these studies to form complexes with the same characteristics. The reduction of absorbance at 610nm is there- fore a measure of reaction of organotin compounds with dithizone and is shown in relation to amount of organotins and organoleads in Figs. 1 and 2. Using the molar absorptivity for dithizone at 610 nm (4.02 x lo4 1 mol-l cm-l) it can be calculated that 1 mol of dithizone has reacted with 1 mol of triethyllead (Fig.1) and 2 mol per mol of dibutyllead diacetate (Fig. 2). All other organotins and organoleads yield less than these theoretical molar proportions. It is clear that the major factor involved in the limited reaction of trimethyltin with dithizone (Fig. 1) is its low partition into the chloroform. The amount of reacting dithizone is markedly increased by the addition of sodium perchlorate and also by reduction of the volume of the aqueous phase (Fig. 1). Nevertheless, this cannot be the only factor and the possibility exists that the affinity of trimethyltin and other organotins and organo- leads is not high enough to ensure complete reaction in the chloroform. Affinity Constant for the Reaction of Dithizone with Triethyltin Hydroxide in Chloroform For the reaction of 1 mol of dithizone and 1 mol of triethyltin hydroxide, if K is the apparent affinity constant of the assumed reaction of triethyltin hydroxide with dithizone64 ALDRIDGE AND STREET: ORGANOTINS AND ORGANOLEADS Analyst, vol.106 (DZ) in the chloroform phase and D is the distribution coefficient of triethyltin hydroxide between chloroform (c) and the buffer system (b), the following equations can be derived: [Et,Sn - DZ], . . - - (1) .. . . [Et,SnOHlc [DZIC K = [Et ,SnOH] [Et,SnOH], D = From equations (1) and (2) [Et,Sn - DZ], D [DZ], [Et ,SnOH], K = . . - - (2) .. . . (3) . . There are several assumptions inherent in the above derivation. Free dithizone and the triethyltin - dithizone complex are assumed to be entirely in the organic solvent and tri- ethyltin in the aqueous phase is assumed not to be complexed to dithizone and is probably entirely present as the hydroxide14 at pH 9.Using [113Sn] triethyltin, the concentration of triethyltin in both aqueous and chloroform phases was measured and the concentrations of free and reacted dithizone were calculated from changes in absorbance at 610 nm. Other conditions were as described under Pro- cedures. The distribution coefficient for triethyltin hydroxide between buffer and chloro- form was found to be 76.5 & 0.07(6) in favour in the chloroform. The derived apparent affinity constant was 3.1 x lo5 1 mol-l (KD = 3.07 x M). As the total concentration 1.2 0 0 1 .o o, 0.8 0 m +? 0.6 2 ? 0.4 0.2 10 20 30 40 50 60 Amount of dialkymetal (except dimethyltin) :nrnol 1000 2000 3000 4000 5000 Amount of dimethyltin chloridehmol Fig.2. Relationship between absorbance of dithizone at 600 nm in 2-cm cells and concentration of dialkyltin and dialkylleads. The solid line indicates a molar ratio (dialkylmetal to dithizone) of 1: 2. For details of method see Procedures. A, Dimethyltin; x , diethyltin; 0, dipropyltin; 0, dibutyltin; 0, diethyllead; and B, dibutyllead. of dithizone in the chloroform is 1.6 x 10-5~ it is clear that not all of the triethyltin will be complexed, and as the dithizone is removed by reaction with triethyltin hydroxide devi- ations from linearity are expected (cf., Figs. l and 2). Therefore, we conclude that tri-January, 1981 BY SPECTKOPHOTOMETRY AND FLUORIMETRY 65 organotins and trioganoleads react with dithizone to produce complexes with a molar ratio of dithizone to organometal of 1 (cf, triethyllead, Fig.1 ) ; the corresponding molar ratio for diorganotins and diorganoleads is 2 (cf., dibutyllead, Fig. 2). Except for the most water- soluble trimethyltin and dimethyltin, minor deviations from theoretical reaction are probably due to inadequate affinity of the organometal for dithizone. Determination of Organotin and Organolead Using 3-Hydroxyflavone The buffer that was used with dithizone was found to be suitable for the determination of trimethyltin using 3-hydroxyflavone ; the relationship between fluorescence and amount of trimethyltin is linear up to 75 and 2 nmol, respectively, for the two modes of operation (Table I).Although the method was originally published for the determination of triphenyltin,lO~ll the formation of a fluorescent complex is remarkably specific and amongst many triorgano- tins and triorganoleads only trimethyltin and triphenyltin are active (Table 11). Replace- ment of one methyl group in trimethyltin by octyl reduces the fluorescence 7-fold and sequential replacement of phenyl in triphenyltin by ethyl groups reduces the fluorescence 10- and 200-fold, respectively. The low fluorescence obtained with most triorgano-tin and -lead compounds is not due to the absence of complex formation with 3-hydroxyflavone; the fairly constant absorbance at 385 nm for all triorganometals indicates that complexes are formed. I t has been reported that dimethyltin and diphenyltin compounds do not form fluorescent complexes with 3-hydroxyflavone.lo~l1 This finding is confirmed and extended to many diorgano-tin and -lead compounds (Table 11).They do, however, form complexes with 3-hydroxyflavone absorbing at 385 nm with molar absorptivities 2 4 times those for tri- organo-tin and -lead compounds (Table 11). TABLE I1 FLUORESCENCE AND ABSORBANCE OF COMPLEXES OF TRIORGANO- AND DIORGANO-TIN AND -LEAD COMPOUNDS WITH 3-HYDROXYFLAVONE Compound Triovganotin- Methyl . . Ethyl . . Propyl . . Butyl . . Octyldimeth yl Phenyl . . Phenyldiethyl Eth yldiphen yl Trio yganolead- Methyl . . Ethyl . . Propyl . . Butyl . . Phenyl . . Diorganotin- Methyl . . Ethyl . . Propyl . . Butyl . . Diorganolead- Methyl . . Ethyl .. Propyl . . Butyl .. Phenyl .. .. .. .. .. .. . . .. .. .. .. .. .. .. .. * . .. .. .. .. .. .. . . . . 1.5 x 104 . . 2.0 x 104 . . 1.6 x lo4 . . 1.2 x 104 . . 2.1 x 104 . . 1.5 x 104 . . 1.7 x 104 . . 1.5 x 104 . . 2.3 x 103 . . 1.1 x 104 . . 1.2 x 104 . . 1.3 x 104 . . 1.7 x 104 . . 4.7 x 104 . . 4.1 x 104 . . 4.2 x 104 . . 4.0 x 104 . . 2.1 x 104 . . 1.6 x lo4 . . 1.9 x 104 . . 2.5 x 104 . . 2.5 x 104 Fluorescence a t 514 nm (yo of trimethyltin) 100 -1.2 No. 1 -0.7 14 380 -2 36 -2 3 4 5 0 -2 8 10* No. 7 4 . 8 0.5 0.3 0 0 * Fluorescence emission at 520 nm.66 ALDRIDGE AND STREET: ORGANOTINS AND ORGANOLEADS AnaZyst, vol. 106 Influence of Dimethyltin on the Fluorescence of the Trimethyltin - 3-Hydroxy- flavone Complex Although dimethyltin and diphenyltin do not form fluorescent complexes with 3-hydroxy- flavone, the absorbance of the complexes formed (Table 11) has been found to reduce the apparent fluorescence of complexes w t h trimethyltin and the measured fluorescence caused by trimethyltin is reduced proportionately to the concentration of dimethyltin (Fig.3). The fluorescence is a linear function of trimethyltin concentration [Fig. 3(a) and Table I] and absorbance is a linear function of dimethyltin and trimethyltin concentrations [Fig. 3(c) and (41. From the following relationships the concentrations of both dimethyltin and trimethyltin can be determined from the measured absorbance at 385 nm and fluorescence at 514 nm: . . * * (4) F T = b,[Me,Sn] . . .. . . .. . . . . * * (5) F D T = b,[Me,Sn](l - b,[Me,Sn]) .. .. AD, = A T + A D = b,[Me,Sn] + b4[Me,Sn] . . .. - - (6) where F D T = fluorescence of trimethyltin - dimethyltin mixtures; b, = slope of linear function relating fluorescence (FT) to [Me,Sn] [cf., Fig. 3(a)]; b, = slope of linear function relating F D T / F T to [Me,Sn] [cf., Fig. 3(b)]; A D , = absorbance at 385 nm due to a mixture of trimethyltin and dimethyltin; b, = slope of linear function relating absorbance ( A , or ADT) to [Me,Sn] & [Me,Sn] b , = slope of linear function relating absorbance (A,) to [Me,Sn] [cf., Fig. 3(c)]. From equations (4), (5) and (6), the following quadratic equation has been derived: [cf-, Fig- 3(d)] ; The values of b,, b,, b, and b, can be determined from standards (Fig. 3). Solution of the quadratic equation yields the concentration of trimethyltin and the concentration of di- methyltin can then be derived using equation (6).Conclusion In this paper a buffer system containing diethanolamine, EDTA and sodium perchlorate has been shown to be suitable for the determination of di- and tri-organotins and -organo- leads with dithizone or 3-hydroxyflavone. Under the conditions described triethyllead and dibutyllead react with dithizone to yield complexes with molar ratios of organometal to dithizone of 1 : 1 and 1 : 2, respectively (Figs. 1 and 2). Other organotins and organoleads do not yield this theoretical stoicheiometry. Based on the measured affinity constant of triethyltin for dithizone (3.1 x lo5 1 mol-l), it is probable its low value is the reason for this discrepancy with most of the organometals examined.For the most water-soluble compounds dimethyltin and dimethyllead, lack of partition into the solvent is a major factor. Amongst a large series of di- and tri-organotins and -organoleads under our conditions only trimethyltin and triphenyltin form fluorescent complexes with 3-hydroxyflavone (Table 11). With maximum sensitivity 0.5 nmol of trimethyltin (Table I) and 0.13 nmol of triphenyltin (Table 11) can be determined. All of the organometals examined form complexes with 3-hydroxyflavone with absorbance maxima at 385 nm. Experiments with mixtures of trimethyltin and dimethyltin have shown that although the 3-hydroxyflavone - dimethyltin complex does not fluoresce, it does reduce the measured fluorescence of the 3-hydroxy- flavone - trimethyltin complex (Fig. 3).It is likely that this will be generally true for mixtures of trimethyltin or triphenyltin with any of the other organometals (Table 11).January, 1981 BY SPECTROPHOTOMETRY AND FLUORIMETRY 25 50 75 100 Trimethyltin/nmol per 7 ml 1.5 I I a p 1.0 ? s 2 0.5 50 100 150 Dimethyltin/nmol per 7 ml 5 LLO 0.6 - I 1 50 100 150 Dimethyltin/nmol per 7 rnl I 1 a t 0 9 s n a 1.5 1 .o 0.5 50 100 150 Trirnethyltin/nmol per.7 ml 67 Fig. 3. Absorbance and fluorescence of mixtures of complexes of 3-hydroxyflavone with dimethyltin and trimethyltin. (a) The lines represent the fluorescence due t o trimethyltin in the presence of dimethyltin: 1, nil; 2, 33.6 nmol; 3, 67.2 nmol; and 4, 134.5 nmol. Procedure A was used (compare with Table I).( b ) The ratio of the fluorescence due t o trimethyltin f dimethyltin (FDT/FT) derived from the results shown in (a). (c) Absorbance at 385 nm in 2-cm cells due to dimethyltin. (d) Absorbance at 385 nm in 2-cm cells due to trimethyltin in the presence of the following amounts of dimethyltin: 1, nil; 2, 33.6 nmol; 3, 67.2 nmol; and 4, 134.5 nmol. The structure - fluorescence relationships for the di- and tri-organotins and -organoleads are difficult to understand. As the operational absorbance coefficients for all organometals examined (Table 11) are similar for all triorgano-tin and -leads and for all diorgano-tins and -leads the affinity of 3-hydroxyflavone must be high enough to ensure that the organometal partitions into the solvent phase (cf., trimethyltin and dimethyltin, Table 11).However, only trimethyltin and triphenyltin yield complexes that show substantial fluorescence (Table 11). None of the symmetrical organometals examined gave more than 10% of the fluorescence of trimethyltin. The replacement of one methyl group in trimethyltin by octyl reduces the fluorescent yield by 86%; the progressive replacement of the phenyl group in triphenyltin reduces the fluorescent yield by 91% and over 99%, respectively. No explanation of these findings is known. References 1. Aldridge, W. N., in Gielen. M., and Harrison, P. G., Editors, “The Organometallic and &-ordination Chemistry of Germanium, Tin and Lead,” Reviews on Silicon, Germanium, Tin and Lead Com- pounds, Freund Publishing House, Tel-Aviv, Israel, 1978, pp. 9-30. Fortemps, E., Amand, G., Bomboin, A., Lauwerys, R., and Laterre, E. C., Int. Arch. Occup. Environ. Health, 1978, 41, 1. Brown, A. W., Aldridge, W. N., Street, B. W., and Verschoyle, R. D., Am. J . Pathol., 1979, 97, 59. Seawright, A. A., Brown, A. W., Aldridge, W. N., Verschoyle, R. D., and Street, B. W.. in prepara- 2. 3. 4. tion.68 ALDRIDGE AND STREET Price, W. J., and Smith, R., “Handbuch der analytischen Chemie. Dritter Teil. Quantitative Bestimmungs- und Trennungsmethoden. Band 4ay. Element der Vierten Hauptgruppe. Zinn,” (in English), Springer-Verlag, Berlin, Heidelberg, New York, 1978. Aldridge, W. N., and Cremer, J. E., Analyst, 1957, 02, 37. Cremer, J. E., Biochem. J., 1957, 67, 87. Cremer, J. E., Br. J . Ind. Med., 1959, 16, 191. Chapman, A. H., Duckworth, M. W., and Price, J. W., BY. Plast., 1959, 32, 78. Vernon, F., Anal. Chim. Acta, 1974, 71, 192. Blunden, S. J., and Chapman, A. H., Analyst, 1978, 103, 1266. Aldridge, W. N., and Cremer, J. E., Biochem. J., 1955, 61, 406. Rose, M. S., and Aldridge, W. N., Biochem. J., 1968, 106, 821. Tobias, R. S., Organomet. Chem. Rev., 1966, 1, 93. Aldridge, W. N., Street, B. W., and Skilleter, D. N., Biochem. J., 1979, 168, 353. Received May lst, 1980 Accepted August 18th. 1980 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

ISSN:0003-2654    

DOI:10.1039/AN9810600060

出版商:RSC    

年代:1981

数据来源: RSC

摘要:

Analyst, January, 1981, Vol. 106, $9. 65-75 69 Amorphous Surface and Quantitative X-ray Powder Diff ractometry Stephen Altree-Williams Division of Occupational Health and Radiation Control, Health Commission of New South Wales, P.O. Box 163, Lidcombe, Australia 2141 John G. Byrnes Geological and Mining Museum, Geological Survey of New South Wales, 36 George Street, Sydney, Australia 2000 and Bernard Jordan Division of Chemical and Physical Sciences, Deakin University, Geelong, A ustralia 3217 Experimental quantitation constants for direct quantitative X-ray powder diffractometry, K i J , are compared with values calculated assuming a model of an ideal crystal. Differences between experimental and calculated ku values demonstrate the existence of significant amounts of amorphous surface in the samples of quartz and corundum investigated.Using respirable dust specimens from “pure” quartz of diverse origins, ku for the 101 reflection varied from 40.5 to 54.5 net counts cm2 pg-’ for the diff ractometer conditions used. The maximum value expected for quartz (calculated using Linde A corundum to calibrate the diffractometer for intensity) was only 47 net counts crn2pg-l. After considering all possible influences, the variations in KiJ were attributed to amorphous surface in the quartz, while the higher values than the theoretical maximum found for quartz were attributed to amorphous surface in the corundum. KiJ for both corundum and quartz increased after a hot orthophosphoric acid treat- ment, while kiJ for quartz decreased with grinding. These results support the amorphous surface concept.The significance of amorphous surface to quantitative X-ray powder diffractometry is discussed. The likely presence of amorphous surface in powders necessitates that care be taken in the interpretation of quantitative results. It is suggested that theoretically calculated X-ray powder diffrac- tion data may prove useful as a source of quantitation standards that are free of amorphous surface effects. Keywords Quantitative X-ray powder diljcractometry ; amorphous surface ; calculated X-ray powder diffraction data ; quartz ; corundum Quantitative X-ray powder diffractometry (XRPD) assesses the mass of a crystalline phase from the integrated intensity of a diffraction reflection. Traditionally, the standard used for diffractometer calibration is a pure sample of the phase of interest.One problem with this approach is that “purity” of a standard has necessarily been equated with “absence of diffraction reflections of other phases.” Analysts have long recognised that such an operational definition of purity is not completely satisfactory because a significant mass of the standard may be amorphous. This amorphous content will not contribute to the Bragg reflections and will show little recognisable influence on the diffractometer scan. Calculated XRPD data has been suggested as an alternative source of standards for quantitative XRPD.1-5 Further, the use of calculated data in conjunction with experi- mentally measured data has been suggested as a way of detecting preferred orientation and extinction effects.4 Considerations given below indicate that such an approach is also useful in the detection of amorphous surface in the analytical standards used in quantitative XRPD.In direct XRPD the amount of analyte is expressed either as mass fraction or as mass per unit area depending on whether the “infinitely thick” method6 or the thin-layer metho~i-~ is used. The respective quantitation equations are70 ALTREE-WILLIAMS et al. : AMORPHOUS SURFACE AND Analyst, Vol. I06 and where IiJ kiJ X J P J 8, M J = integrated intensity of diffraction line i of phase J ; = the quantitation constant, dependent on analysis conditions and diffraction = mass fraction of phase J in the sample; = mass absorption coefficient of the sample; = Bragg angle for diffraction line i of phase J ; = Bragg angle for a chosen diffraction line from the material supporting the = mass of phase J per unit area; line i of phase J ; thin-layer sample ; L = In(?); IoD, ID = integrated intensity of the chosen diffraction line from the material supporting the thin-layer before and after loading, respectively.The quantitation equations provide compensation for the influence of absorption on the relationship between diffraction intensity and amount of analyte. They do not, however, compensate for the other factors that can influence the mass - intensity relationship. Table I summarises these factors, which include effects from both sample presentation and crystallite quality. TABLE I FACTORS (OTHER THAN ABSORPTION) THAT AFFECT THE INTEGRATED INTENSITY OF A REFLECTION Effect Factor Sample presentation .. .. . . Preferred orientation Granularity Surface roughness Microabsorption Extinction Structural heterogeneity Chemical heterogeneity Crystallite quality . . .. . . Amorphous surface The quantitation constant ki J , which is normally experimentally measured, can also be assessed from theoretically calculated XRPD data as follows5 : ku = K(ASF,)I,'~~ x - . . . . . .. .. 1 P J (3) where K = constant for given diffractometry conditions; ASF, = absolute scale factor of phase J for wavelength X calculated from the crystal lure' = calculated relative integrated intensity of diffraction line i of phase J ; f J = density of phase J ; structure data of phase J ; ASF, and Ii,rel are calculated using given crystal structure data and assuming2~3 (i) a conventional parafocusing powder diffractometer (without monochromator) using a specific wavelength;January, 1981 QUANTITATIVE X-RAY POWDER DIFFRACTOMETRY 71 (ii) a phase J that is free of the effects summarised in Table I ; and (iii) the kinematical theory of the diffraction of X-rays. Assuming the validity of points (i)-(iii), then the hypothesis of Hubbard and Smith4 can be put explicitly as follows : differences between experimentally determined k , values and those calculated from equation (3) will be due to one or more of the factors listed in Table I occurring in the standard used to determine k , (or in the standard used originally for the experimental determination of K ) .This paper describes the application of the above hypothesis to interpret the quantitative XRPD data obtained from pure quartz samples of varying geological history.The differ- ences found in the 'data from these samples are consistent with differences in amorphous surface content. Further experimentation with both quartz and Linde A corundum provided results that also support the amorphous surface concept. The significance of amorphous surface to quantitative X-ray powder diff ractometry is considered. Experimental Phases commercial product Linde A, manufactured by Union Carbide, USA. high-purity corundum with a nominal particle size of 0.3 pm. necessary, were powdered in a ring grinder (tungsten carbide barrel, grinding for 60 s). Preparation of Samples for Diffractometry Each sample was placed in a clean dust chamber and four deposits of respirable dust collected on 0.8-pm pore size Nuclepore filters by use of the Higgins and Dewell cyclonelo (Catalogue No.A7650/1, Casella, London) as described previously.ll Under the conditions used, the cyclone would exclude from collection all spherical quartz particles (density = 2.65g ~ m - ~ ) of diameter greater than 5 pm and would exclude from collection 50% of particles of diameter equal to 3.5 pm. The amount of deposit on each filter was determined by weighing. Care was taken to ensure the selected quartz sample was free of contaminating materials. Any sample that gave a non-a-quartz reflection of peak height greater than 50 counts s-1 for a 2-mg deposit was not used. Diffractometry A Bragg - Brentano parafocusing diffractometer was used, manufactured by Philips, The Netherlands.The vertical goniometer (PW 1050/70) had a radius of 173 mm and was aligned to a 6" take-off angle from the X-ray tube anode. The X-ray path consisted of incident-beam Soller slit, divergence slit, sample, receiving slit, diff racted-beam mono- chromator and counter. The diffracted-beam monochromator (E3-202 GVW) was fitted with a curved graphite crystal (dimensions 16 x 16 mm, radius of curvature 223.5mm) and was manufactured by Advanced Metals Research, Mass., USA. No scatter slit or diffracted-beam Soller slit was used in conjunction with the monochromator. Other diff racto- meter components used for this work were sample spinner, xenon-filled proportional counter and copper anode broad-focus X-ray tube (focus dimensions 2 x 12 mm). Samples were analysed with tube conditions of 45 kV, 30 mA with pulse-height selection and sample spinning.The integrated intensity of a diffraction line was measured by counting while scanning the diffraction line at a rate of g020 min-l for 2 min. Background was counted at the start and finish angles of the scan and the average value used. Net peak area counts was used as the integrated intensity of the diffraction line. An aluminium blank was used as an external standard to correct for long-term instrument drift. The quantitation constants were determined experimentally for the most intense reflections on each filter using equation (2). Silver was used as the diffracting underlay for all filters.ll Calculated XRPD Data given by Hubbard et Corundum and a-quartz were investigated.The corundum used was a sample of the This product is a Quartz samples were collected from sources of varying geological history and, where A 1" divergence slit and a 0.3 mm receiving slit was used. The data given by Smith' were used for the I@ data of a-quartz and corundum and that All calculated data were corrected for the for the ASF, data.72 ALTREE-WILLIAMS et d. : AMORPHOUS SURFACE AND Analyst, VOZ, 106 effect of the monochromator, assuming a polarisation term of 1 + Kcos228,, where K is the polarisation ratio12 of the monochromator. K has been determined experimentally as 0.93 & 0.04 for the diffractometer arrangement used in this work.l3 Treatment of Quartz and Corundum with Hot Orthophosphoric Acid Selected quartz powders were treated with hot orthophosphoric acid according to the method of Jephcott and Wall,14 the samples being held at 230-240 "C for 10 min.Linde A corundum powder was similarly treated except that it was held at 200-210 "C for 5 min. Each treated powder was collected by filtration, washed repeatedly with water, dried over- night a t 115 "C, cooled and placed in the dust chamber and four deposits of respirable dust were collected. Results and Discussion kiJ of the Quartz Samples Experimentally measured k, values for the five most intense reflections of each quartz sample (expressed relative to the 101 reflection) are given in Table I1 together with the k, value for the 101 reflection (written klOl).Each sample shows relative k, values similar to the set predicted by calculated XRPD data. The prepared samples can therefore be considered free of significant influence of preferred orientation, extinction and structural and chemical heterogeneity, each of which would affect relative k, (and, of course, exert specific influence on klol). The k,,, values, nevertheless, still show a 30% variation between samples. Interestingly, some klpl values were greater than the maximum value predicted from calculated XRPD data using Linde A corundum as the reference material for determining K . ~ The most likely explanation for this observed variation in K,,, is different amorphous surface content in each sample. Granularity and surface roughness effects should be less than 2% absolute15 (dl 2 pm, p = 94 cm-l, 26 > 20') and considerably smaller than this within a suite of samples of controlled particle size and composition.Microabsorption is not a relevant consideration in a single-component sample. TABLE I1 EXPERIMENTAL k, OF RESPIRABLE-SIZE FRACTION OF QUARTZ SAMPLES Relative kiJ for a-quartz reflections* k10,tl Host formation A , netcounts 112 211 cm'pg-' 100 101 102 Sample V f No. Description Locality Descrlption Age 1 Clear, euhedral Kingsgate, Pegmatite Permian 18.1 f 0.2 100 f 0.7 7.6 f 0.1 15.3 f 0.1 11.6 f 0.2 45.3 f 0.3 2 Clear. euhedral Herkimer Little Falls Cambrian 18.3 f 0.3 100 f 1.4 7.2 f 0.5 14.6 f 0.3 10.8 f 0.2 43.1 f 0.6 quartz NSW. vugh County. dolomite quartz 34 Aeolian quartz dust 4f Min-U-Sil 5 Q22g Novaculite Q45g Novaculite Calculated kiJ t N.Y., USA Seal Rocks, Palaeosol Quaternary NSW within larae 17.7 f 0.3 100 f 0.8 7.8 f 0.2 15.6 0.4 11.9 f 0.3 51.4 f 0.4 saud dune A high-purity crystalline silica of 18.4 f 0.2 100 f 3 6.7 f 0.2 14.1 f 0.5 10.4 f 0.4 40.5 f 1.1 nominal particle size 5 vm maximum.A commercial product of PGS Floridin Corp., Pittsburgh, Pa., USA Hot Springs Arkansas Devonian to 17.4 f 0.1 100 f 2 7.5 f 0.4 15.8 f 0.4 12.3 0.4 54.5 f 1.0 region, novaculite Mississippian Arkansas, USA Hot Springs Arkansas Devonian to 17.9 f 0.2 100 f 0.8 7.6 f 0.1 15.9 f 0.4 12.3 f 0.2 50.9 f 0.4 region, novaculite Mississippian Arkansas, USA 18.1 100 7.7 14.1 10.5 47.0 f 1.3 Average value of four preparations f mean deviation. t Diffractometer constant K = (3.65 & 0.10) x lo6 net counts ASF-l cm-', determined from experimental measurements on corundum'; A S F j data from ref.3; Zuml data from ref. 1 and adjusted before use assuming a polarisation ratio K = 0.93 * 0.04 for the monochro- mator used with Cu Ka radiation. f Sample existed as powder and was used without grinding. f Powdered sample kindly provided by M. B. Norman, 11 (US Geological Survey).*0 Sample used without grinding.January, 1981 Experimental Investigation of Amorphous Surface QUANTITATIVE X-RAY POWDER DIFFRACTOMETRY 73 Efect of grinding Grinding a quartz sample would be expected to reduce its average particle size and increase If amorphous surface is the dominating influence, the experi- Conversely, if granularity is the dominating The effect of grinding was investigated by grinding samples 2 and Q22 (Table 11) in the The resulting powders were Results are All results are free of obvious preferred orientation, extinction and The k , values decrease with grinding time and it is con- In effect, grinding reduces the its amorphous surface content.mental k,, should decrease with grinding. influence the k , values should increase. laboratory ring grinder for various lengths of time up to 60 s. sub-sampled for respirable dust and ki , values experimentally determined. given in Table 111. structural heterogeneity effects. cluded that grinding increases amorphous surface content. a-quartz content by converting some of the ordered (SiO,), lattice into a disordered state. TABLE I11 EXPERIMENTAL k , OF QUARTZ WITH GRINDING Relative kiJ for a-quartz reflections k1oJ Sample net counts No.Grinding time/s 100 101 102 112 211 cm2 pg-1 2 10 18.1 f 0.4 100 f 0.9 7.0 f 0.4 14.9 f 0.4 11.3 f 0.4 47.0 f 0.4 30' 17.2 100 6.5 14.8 10.0 45.9 60 18.3 f 0.3 100 f 1.4 7.2 f 0.5 14.6 f 0.3 10.8 f 0.2 43.1 f 0.6 r A Q22 As received from M.B. Norman, I1 17.4 f 0.1 100 f 2 7.5 f 0.4 15.8 0.4 12.3 f 0.4 54.5 f 1.0 10 17.4 f. 0.1 100 f 1.6 7.8 f. 0.1 16.0 &- 0.1 12.4 f 0.4 51.0 f 0.8 60 17.8 & 0.1 100 f 0.6 7.5 f 0.2 16.0 f 0.2 12.1 f 0.4 48.7 0.3 * Only one filter preparation was made. Eflect of chemical treatment To confirm further the amorphous surface hypothesis, some quartz samples were treated with hot orthophosphoric acid. Because this treatment dissolves amorphous silica,'* much of the amorphous surface content should be eliminated from the powdered quartz samples and the kiJ values should increase. Quartz samples with high to low measured kiJ values (samples Q45, 1 and 4 in Table 11) were so treated.The treated powders were washed, dried and sub-sampled for respirable- size particles and the k , values were determined. There is no significant difference in relative k,, before and after treatment with orthophosphoric acid for any of the samples. The results can therefore be considered free of obvious preferred orientation, extinction and structural heterogeneity effects. After the chemical treatment, kiJ values increased by 12% for sample Q45, by 14% for sample 1 and by 18% for sample 4. I t is concluded that these increases in ki, values are due to loss of amorphous surface from the treated powders.Results are given in Table IV. TABLE IV EXPERIMENTAL k , OF QUARTZ BEFORE AND AFTER ORTHOPHOSPHORIC ACID TREATMENT Sample - No. Treatment 100 Q45 Before 17.9 f 0.2 After 17.4 f 0.1 1 Before 18.1 f 0.2 After 17.8 f 0.2 4 Before 18.4 f 0.2 After 18.0 & 0.2 Relative k , , for a-quartz reflections A \ 101 102 112 21 1 100 f 0.8 7.6 f 0.1 15.9 f 0.4 12.3 f 0.2 100 f 1.0 7.2 f 0.2 15.7 f 0.3 11.5 f 0.3 100 f 0.7 7.6 f 0.1 15.3 f 0.1 11.6 & 0.2 100 f 2 7.0 f 0.4 15.2 f 0.1 11.3 f 0.1 100 f 3 6.7 f 0.2 14.1 f 0.5 10.4 f 0.4 100 1.0 6.1 f 0.4 14.4 f 0.3 10.2 & 0.3 k,o,l net counts cm2 pg-' 50.9 f. 0.4 57.2 0.6 45.3 f 0.3 51.7 f 1.0 40.5 f 1.1 47.9 f 0.574 Analyst, V O ~ . 106 Linde A corundum As mentioned earlier, one of the interesting aspects of the data in Table I1 is that a number of quartz experimental k,,, values were greater than the maximum value predicted from calculated XRPD data using Linde A corundum as the reference material for determining K .This implies that Linde A is a less perfect specimen of corundum than was assumed. If Linde A contained X mass-% amorphous surface, then calculated ki, values for other phases based on this reference material would all be X% less than their maximum possible values. To test the hypothesis that Linde A corundum contained significant amorphous surface, a sample of it was treated with hot orthophosphoric acid, the resulting material washed and dried, the respirable-size particles were sub-sampled and the k,, data determined.The results are given in Table V, together with the calculated relative ki, values. Effects of preferred orientation, extinction and structural and chemical heterogeneity can be dis- counted from the relative k, data and effects from granularity and surface roughness discounted on theoretical grounds given the known particle size and chemical composition of the sample. The 22% increase in experimentally measured k , values is therefore most likely caused by the reduction of amorphous surface content in the particles due to the chemical treatment. ALTREE-WILLIAMS et al. : AMORPHOUS SURFACE AND TABLE V EXPERIMENTAL ki, FOR LINDE A CORUNDUM BEFORE AND AFTER ORTHOPHOSPHORIC ACID TREATMENT Relative kiJ for corundum reflections k.131 c I , net counts 012 104 110 113 116 124 030 cm’ Vg-l Before treatment .. . . 51 f 0.6 85 f 0.9 37 f 0.5 100 0.9 100 f 3 40 f 1.0 59 f 0.6 11.53 f 0.10 After treatment . . .. 50 f 0.5 84 f 1.4 36 f 0.4 100 & 2 98 f 2 38 f 0.3 69 0.6 14.1 f 0.3 Calculated relative kiJ* . . 60 91 41 100 101 40 62 lure’ data from ref. 1 and adjusted before use assuming a polarisation ratio K = 0.93 & 0.04 for the monochromator used with Cu Ko: radiation. The Amorphous Surface Model Anderson16 provided a review of the evidence with respect to quartz, and Tatlockl’ gave evidence for other mineral phases. Suortti and Jennings,18 in their authoritative review of quantitative XRPD measurements, found lower than expected intensity in their magnesium oxide sample. Skidmore and Schwarz discussed amorphous surface in lead dioxide polymorphs.19 Amorphous surface characterises a volume of material bounding a polycrystalline particle, bounding a crystallite and bounding mosaic blocks within a crystallite.Clearly, the amorphous surface content of two samples of a phase with equivalent particle size distri- butions could differ significantly according to the extent of abrasion received and the number of crystallites per particle. This explains why the different samples of “pure” quartz gave different kiJ values (Table 11). Further, the amorphous surface content of a particle is not simply an external skin. The results of chemical treatment on quartz samples (Table IV) indicate that a significant amorphous content can exist within the particle, presumably associated with crystallite boundaries or with dislocations within crystallites. From the point of view of quantitative analysis, a simplistic (but operational) model for a powdered phase would envisage two components : a crystalline portion (of structure defined by the crystal structure data for the phase), and an amorphous portion.The amorphous portion contributes to the mass of the powdered phase but not to the intensity of its Bragg reflections. Amorphous surface in crystalline materials has been suggested for some time. Quantitative Analysis Using X-ray Powder Diffractometry Quantitation of powdered crystalline phases has a complexity not apparent in the quantita- tion of elements or compounds. As indicated above, this complexity arises from the difficulty of defining what constitutes the analyte.The model of a powdered crystalline phase used for calculated XRPD datal-3 assumes that the randomly oriented crystallites of the analyte are made up totally of mosaic blocks within which atomic order is perfect. ForJanuary, 1981 QUANTITATIVE X-RAY POWDER DIFFRACTOMETRY 75 quartz the model seems to be good for the particle size studied (< 5 pm), except for the large amount of amorphous surface that occurs. A similar observation can be made for Linde A corundum. It seems likely that most samples of a crystalline phase will contain some mass- per cent. of amorphous surface after powdering. Given the validity of the amorphous surface model, the problem for the analyst becomes one of assessing the amorphous surface content in his analytical standards.One approach to this problem is to investigate alternate sources of the phase and/or chemical treatment of the phase, then to use the material with highest ki, values for the analytical standard. This approach is illustrated above for quartz and by the work of Skidmore and Schwarzlg for lead dioxide polymorphs. The second approach to the problem is to rely on calculated XRPD data as standards.1-5 When suitable reference phases (free from the effects listed in Table I and with a known amorphous surface content) become a~ailable,~ then K can be determined for a given diffracto- meter set-up, and hence ki, values that are free of amorphous surface effects can be calculated for any known crystal structure by the use of equation (3).The potential value of calculated XRPD data for analytical standards is probably better than published studies3~5 would indicate, because neither of these studies considered amorphous surface effects. The authors thank Camden R. Hubbard (US National Bureau of Standards) for reviewing an earlier draft of this work, Meade B. Norman 11, (US Geological Survey) for supplying the novaculite samples and Irena Sprogis (NSW Health Commission) for assistance in collecting the diffractometric data. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. References Smith, D. K., Norelco Rep., 1968, 15, 57. Borg, I. Y., and Smith, D. K., “Calculated X-ray Powder Patterns for Silicate Minerals,” Memoir 122, Hubbard, C. R., Evans, E. H., and Smith, D. K., J. Appl. Crystallogr., 1976, 9, 169. Hubbard, C. R., and Smith, D. K., Adv. X-ray Anal., i977, 20, 27. Altree-Williams, S., Anal. Chem., 1978, 50, 1272. Klug, H. P., and Alexander, L. E., “X-ray Diffraction Procedures,” Second Edition, John Wiley, Williams, P. P., Anal. Chem., 1959, 31, 1842. Leroux, J., Davey, A. B. C., and Paillard, A., Am. Ind. Hyg. Ass. J., 1973, 34, 409. Altree-Williams, S., Anal. Chem., 1977, 49, 429. Higgins, R. I., and Dewell, P., in Davies, C. N., Editor, “Inhaled Particles and Vapours, 11,” Altree-Williams, S., Lee, J., and Mezin, N. V., Ann. Occup. Hyg., 1977, 20, 109. International Union of Crystallography, A cta Crystallogr., 1978, A34, 159. Altree-Williams, S., and Jordan, B., Anal. Chem., 1980, 52, 1296. Jephcott, C. M., and Wall, H. F. V., Arch. Ind. Health, 1955, 11, 425. Suortti, P., J. Appl. Crystallogr., 1972, 5, 325. Anderson, P. L., Am. Ind. Hyg. Ass. J . , 1975, 36, 767. Tatlock, D. B., Geol. Surv. Bull., 1966, NO. 1209. Suortti, P., and Jennings, L. D., Acta Crystallogr., 1977, A33, 1012. Skidmore, P. R., and Schwarz, K. R., Analyst, 1979, 104, 952. Murata, K. J., and Norman, M. B., Am. J . Sci., 1976, 276, 1120. Geological Society of America, Boulder, Colo., 1969. New York, 1974, pp. 532-534. Pergamon Press, Oxford, 1967, p. 575. Received May 8th, 1980 Accepted July 14th. 1980

ISSN:0003-2654    

DOI:10.1039/AN9810600069

出版商:RSC    

年代:1981

数据来源: RSC