ISSN:0003-2654    

DOI:10.1039/AN98207FX033

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207BX035

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207FP089

出版商:RSC    

年代:1982

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98207BP095

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

September 1982 The Analyst Vol. 107 No. 1278 Evaluation of an Inductively Coupled Plasma Emission Direct-reading Spectrometer for Multiple Trace Element Analysis of Foodstuffs W. H. Evans and Dorothy Deliar Departmed of I?tdustry Labovatory of the Government Chemist Cornwall House Stainford Street London, SE1 9NQ An inductively coupled plasma optical emission direct-reading polyclironiator (ICP-OES) has been evaluated for multiple trace element analysis of food-stuffs. Rlajor trace elements (copper iron manganese and zinc) and trace elements (cadmium lead nickel chromium arsenic and tin) have been con-sidered. For each element random bias has been exemplified for routine practice in both sulphuric and hydrochloric acid standard solutions by the standard deviations for instrument noise repeatability and reproducibility of measurement.Systematic bias originating from direct interferences has been categorised and relevant corrections have been calculated. To avoid the worst excesses of these interferences and to concentrate the trace elements to levels that permit measurement a method involving isolation by chelation extraction oxidation and solubilisation in fixed acid solution has been used. Six of the elements are amenable to this treatment copper, manganese zinc nickel cadmium and lead. \Then this treatment is applied t o digests of a selection of foodstuffs no element satisfies the criterion of accuracy required for quantitative analysis in dietary surveys of foodstuffs. The results display comparable confidence intervals to those obtained by flanie atornic-absorption spectrometric measurement however for the first four of the elements listed.Keywords Dived-reading inductively coupled plasnza ; optical emission spectvo-iizetev ; trace elements ; foodstufls analysis Spectrographic analysis of trace levels of elements with the d.c. arc had been applied over many decades to the analysis of substrates including plant materials as solids or in solution,l when in 1964 a more stable energy source the inductively coupled plasma (ICP) was suggested by Greenfield et aZ.2 and Wendt and F a ~ s e l ~ independently for the generation of free atoms and excitation of these released atoms. The principle weaknesses of spectrographic analysis were its relative lack of sensitivity susceptibility to interferences by inorganic species on the analyte element and excessive variation in results.I t became apparent with time that the inherent stability of the plasma energy source when applied to measurement of trace elements in solution gave improved sensitivity and reduced variation in measurement and could remove well known interferences inherent in traditional elemental analysis obtained from arc or flame optical emission spectroscopic analysis. For the past two decades however elemental analysis has been dominated by flame atomic-absorption spectrometry with sequential measurement in solution. I t was suggested that with a simultaneous multi-elemental facility for measurement the inductively coupled plasma optical emission spectrometer (ICP-OES) could seriously challenge this domina-tion.4 Subsequent analytical application of the ICP-OES with its wide linear calibration range was initially reported for a variety of samples at high analyte concentration^,^ to major,* major trace and some trace elements in water and seawater,+* and to major and niajor trace elements in biological material~.~-ll I t is significant that for trace elements in * For the purposes of this investigation of foodstuffs sodium potassium calcium magnesium phosphorus aiid chloride are ternied niajor elements copper iron nianganese and zinc are termed major trace elements and any other elements are termed trace elements.Crown Copyright. 97 978 Analyst VoZ. 107 biological materials there is a dearth of proven application e.g. for standard reference materials only levels of the major elements and the major trace elements copper iron, manganese and zinc have been reported.These four elements together with cadmium lead, nickel arsenic tin chromium cobalt and others are routinely determined in this laboratory by atomic-absorption spectrometry for the United Kingdom total diet survey in other diet surveys and in retail foodstuffs as occasion demands. This paper presents the results of an investigation on the attempted determination of levels of these elements normally found in foodstuffs or dietary foodstuff homogenates with measurement by an inductively coupled plasma optical emission direct-reading polychromator. Direct interferences introducing systematic bias into results may originate from a number of sources with this instrumental technique.1. Interferences on an element by other inorganic species due to compound formation so prevalent in flame chemistry appear to be absent.l2,l3 As ICP-OES tends to measure atom - ion emission to gain sensitivity possible similar ionisation interferences are of more importance. By careful control of operating parameters investigations to date suggest relatively small interferences of this kind although the mechanisms involved are not clearly understood.12-14 Kornblum and De Galan,15 however advise caution and suggest that such interferences may depend on the particular combinations of analyte and interfering elements. Another type of interference originates from nebulisation and more so from transport of the sample into the plasma.It has been shown that the signal intensity can be seriously affected by changes in viscosity from nebulising. mineral acids of different concentration for a high-power plasma source.16 The ranges of acid strength investigated in this instance were particularly wide and a much smaller range of difference would be expected from measure-ments on foodstuff digests. Such ranges have been investigated for perchloric acid,g,lO and it would appear that strict control of acidity is desirable for measurement. Alternatively, the nebuliser may be pumped above the natural rate although this will not necessarily remove the interference from transport effects.16 There remain the spectral interferences which were admitted to be serious as application of the instrumental technique progressed.These have been outlined by several workers and those recognised at present may be classified as (a) spectrometer stray light ( b ) overlap from broadened lines and elemental spectral continuum and (c) spectral line coincidence. (a) Normally the spectral background is a large portion of the total measured signal and, if this background is modified large systematic errors are inherent in the measurement of low concentrations of solutions. Such modifications can occur from stray light and the intense radiation from calcium and magnesium is a principle cause.17 Secondary stray light may be caused by reflection and scattering from optical components while primary stray light originates from for example imperfections in the diffraction grating. Of the methods for reducing these band rejection filters fitted to the entrance slit of a polychromator were originally favoured and stray light effects from calcium were considerably reduced a t some expense of detection limits of analyte elements through loss of t r a n s m i s s i ~ n .~ ~ ~ ~ ~ In more recent instruments the reductions in primary stray light are achieved by replacement of mechanically ruled diffraction gratings with holographically recorded gratings20 and improved optical design. ( b ) Similar treatment with band rejection filters of the stray light effects from magnesium delineated true spectral interferences resulting from spectral overlaps and collisional line broadening and recombination continuum from radiative electron - ion recombination processes.21 Further examples of these phenomena have been described for interferences from the alkaline earth elements and aluminium to which may be added background variation from atomic and molecular constituents of the plasma.22 The latter may arise from such species as nitrogen oxide hydroxyl or atomic argon.A choice of emission line may prevent coincidence and for this purpose a graphics software system has been described.23 With the remaining spectral effects for sequential measure-ment with a monochromator correction of interference is possible although some difficulty may be encountered with highly structured backgrounds. For simultaneous measurement ivitli a polychromator mathematical Correction has been suggested,24 and has been applied in practi~e.~vl~ These corrections are usually found to be lion-liixar and tlie extent to which such corrections would be applicable for nulti-spectral interference is unclear.EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR They may be listed under three headings: 2. 3. (c) To these spectral effects must be added those from spectral line coincidence September 1982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS Experimental The methods used in this investigation do not signify endorsement of the procedure. Reagents 979 All reagents should be of the grade specified with low contents of the elements to be Nitric acid sp. gr. 1.42. It is recommended that BDH Chemicals Aristar grade or an Sulphuric acid sp. gr. 1.84 and dilute (1 + 19). Hydrochloric acid sp. gr. 1.18 and dilute (1 + 3 and 1 + 19).Recommended grade as for Ammonia solution sp. gr. 0.88. Analytical-reagent grade. Ammonium tetramethylenedithiocarbamate [ammonium pyrrolidine dithiocarbamate ( A PDC) 3. If determined. equivalent grade of acid be used. Solutions should be prepared with distilled water. Recommended grade as for nitric acid. nitric acid. This reagent should not be discoloured or have an excessively strong odour of ammonia. necessary it can be purified as described previ0usly.~5 Diethylammonium diethyldithiocarbamate (DDDC) . Chelating reagent solution 1.0% mlV solution of APDC - DDDC. of each reagent in 100 & 1 ml of water. Chloroform. Analytical-reagent grade. Standard solutions Primary standard solutions specially prepared for atomic-absorption spectrometry can be purchased from commercial sources and contain 1.0 g 1-1 of element in 1 N hydrochloric or nitric acid.Dilute each primary standard solution 1 + 19 with distilled water to give solutions containing 50 mg 1-1 of the element and store in clean polyethylene bottles. For tin, prepare this standard immediately before use. If used to prepare composite standards containing sulphuric acid dilute serially solutions of each element to suitable levels ensuring each solution contains the equivalent of 40 ml of sulphuric acid (1 + 19) per litre. If used to prepare composite standards in hydrochloric acid dilute serially solutions of each element to suitable levels to contain the equivalent of 10 ml of hydrochloric acid (1 + 1) per litre. Composite standard solutions Prepare these solutions to contain the concentrations described in Table I for sulphuric acid (1 + 19) solutions 0-14 and in Table I1 for hydrochloric acid (1 + 19) solutions 01-N.Apparatus All glass apparatus must be kept permanently full of 1 PJ nitric acid when not in use; glass apparatus used for preparing standard solutions of tin should be kept full of 1 N hydrochloric acid. Dissolve 1.00 & 0.01 g COMPOSITION OF SULPHURIC ACID (1 Standard solution 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 c u 0 ------0.01 0.02 0.05 0.1 0.15 0.2 0.3 0.5 Mn 0 ------0.01 0.02 0.05 0.1 0.15 0.2 0.3 0.5 Fe 0 ------0.1 0.2 0.5 1.0 1.5 2.0 3.0 5.0 Zn 0 ------0.05 0.1 0.25 0.50 0.75 1 .o 1.5 2.5 TABLE I + 19) MULTI-ELEMENTAL STANDARDS (mg 1-l) AS Cd Co Cr Ni Pb Sn 0 0 0 0 0 0 0 - 0.005 0.005 -0.05 0.001 0.005 0.005 0.01 0.01 0.05 0.1 0.002 0.01 0.01 0.015 0.015 0.1 0.15 0.003 0.015 0.015 0.02 0.02 0.15 0.2 0.004 0.02 0.02 0.03 0.03 0.2 0.25 0.005 0.025 0.025 0.05 0.05 0.25 - - -- - - - - - 980 EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR Analyst YOl.I07 TABLE I1 COMPOSITION OF HYDROCHLORIC ACID (1 + 19) MULTI-ELEMENTAL STANDARDS (mg 1-1) Standard solution 01 A B C D E F G H JK L M N Cu Mn Fe Zn As Cd Co Cr Ni Pb 0 0 0 0 0 0 0 0 0 0 - - - - 0.05 0.01 0.01 0.05 0.05 0.05 - - - - 0.1 0.02 0.02 0.1 0.1 0.1 - - - - 0.15 0.03 0.03 0.15 0.15 0.15 - - - - 0.2 0.04 0.04 0.2 0.2 0.2 0.05 0.05 0.5 0.25 0.3 0.05 0.05 0.3 0.3 0.3 0.5 0.5 0.5 0.1 0.1 1.0 0.5 0.5 - -0.2 0.2 2.0 1.0 -0.5 0.5 5.0 2.5 - - - - - -1.0 1.0 10.0 5.0 -1.5 1.5 15.0 7.5 - - - - -2.0 2.0 20.0 10.0 - -3.0 3.0 30.0 15.0 - - - -5.0 5.0 50.0 25.0 - - - - - -- - - - -- - - - --- - - -- -Sn 0 0.25 0.5 0.75 1.0 1.5 2.5 -------Inductively coupled plasma optical emission spectrometer An Applied Research 137 ICPQ instrument was used consisting of an inductively coupled argon plasma operated at 27.12 MHz with direct reading on a 25-channel 1-m Paschen-Runge quantometer.The normal compromise operating conditions involved an incident power of 1.6 kWl viewing height of 15 mm above the load coil a plasma argon flow-rate of 10 1 min-l, an auxiliary argon flow-rate of 1 1 min-l and a sample flow-rate of 1 1 min-l.Aspiration was through an all-glass concentric Meinhard-type pneumatic nebuliser without pumping. A rejection filter was used to ensure minimal residual spectral interference from calcium. For magnesium] only secondary stray light was removed from adjacent polychromator channels with a filter over the secondary magnesium slit.26 Measurements were made a t the following analytical wavelengths Cu I 324.75; Fe 11 259.94; Mn 11 257.61; Zn 11 202.55; As I, 193.77; Cd 11 226.50; Co I 345.35; Cr 11 283.56; Ni 11 231.60; Pb 11 220.35; and Sn I, 189.99 nm. Procedure Preparation of sample digests These should be prepared according to method ( l ) C of the Analytical Methods Committee2’ using precautions described subsequently.28 The resulting 100 ml of digest which is in nominally 5% V/V sulphuric acid should be co ourless and should not contain any suspended solids.Prepare at the same time two reagent B lanks from the volumes of acid used in sample oxidation. Chelation and concentration 1 ml of sample digest into a 150-ml beaker and using a pH meter adjust the pH to 4 & 0.2 with ammonia solution. When necessary a smaller aliquot must be diluted with sulphuric acid (1 + 19) to give a measured concentration within the calibration range. Remove the electrode assembly and rinse. Transfer the solution into a 100-ml separating funnel with minimum rinsing. Add 5 & 1 ml of chelating reagent solution shake for 15 s and allow to stand for 3 min.Extract with 10 & 1 ml of chloroform shaking for 30 s. Allow the solution to stand until separation is complete then drain the chloroform layer into a dry 150-ml Pyrex beaker. Repeat the extraction with chloroform after further addition of chelating reagent and finally extract with 10 ml of chloroform. Evaporate the combined chloroform extracts to dryness on a boiling water-bath add 15 ml of nitric acid (sp. gr. 1.42) and boil to dryness on a hot-plate. Cool the residue and dissolve it in 2 ml of hydrochloric acid (1 + 3) transfer into a 10-ml calibrated flask washing with several 2-ml volumes of water and dilute to 10 ml. Measure 70 Reagent blank solutions should be treated similarly September 1982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS 98 1 Measurement For the trace elements standardise the sensitivity setting of the polychromator channels to the highest permissible value.Stabilise aspiration with the relevant acid of concentration 1 + 19. Arrange the samples standards and blank solutions in random order but with a reagent blank in each half of the total series. Note the sequence of the standard solutions and move forward two places in the sequence for subsequent measurement series. Aspirate each solution in turn and return to acid (1 + 19) for 1-2 min between each solution. Record the digital readings for consecutive 10-s integration times and then re-read the series of solutions in reverse order to give a total of four readings for each solution. Calibration may be obtained by computer techniques ; for this investigation all calculations were obtained arithmetically .Results and Discussion Calibration For multi-elemental measurement of solutions originating from foodstuffs the standard concentrations used for calibration must vary considerably between elements. Because of this care must be exercised in avoiding incompatible combinations of elements in the acid medium of choice which could give losses from solution by precipitation. Cross-contamina-tion of one element at low concentrations by impurities in an element present at high con-centrations must similarly be avoided as it is not unusual to have impurity levels of % in metals. The multi-elemental standards used for sulphuric acid (1 + 19) and hydro-chloric acid (1 + 19) are tabulated in Tables I and 11; those in hydrochloric acid are in general 10 times greater to allow for the 7-fold concentration obtained by chelation and extraction.The overlap of low concentrations of the major trace elements upon the trace elements indicated in Table I1 did not cause cross-contamination or spectral interference; this was proved subsequently. I t must be recognised that quantitative analysis must be shown to be free from systematic bias obtained from direct interferences while the random bias must be defined for the measured concentration range. With the ICP-OES system experience indicates that correction must be made for known direct interferences. To assess these it is initially necessary to obtain good estimates for the variation of measurement on standard solutions at levels relevant to those expected in foodstuffs.The stipulation must be made that the variation is assessed within a series of measurements containing food digests (which need not be recorded) or solutions containing potential interferences i.e. simulating routine practice. The first seven series of standard readings obtained according to the procedure over a period of 2 months were processed by a previously described system,29 for each element for both sulphuric acid (1 + 19) and hydrochloric acid (1 + 19) solutions. Each standard net response was expressed in terms of the lowest standard concentration and the average net response for all the standards for each calibration was calculated in terms of this concentra-tion for each element.The ratio of the sum of the net responses for each standard to the sum of the net average response indicated the expected linearity for each element. The ratios of means falling outside 95% confidence limits (95 C.L.) imposed by the reproducibility, s when practical measurement was possible numbered 3 out of 122. The standard deviation representing instrument noise sn can be obtained from the paired consecutive readings for each elemental standard solution for the seven calibration lines and expressed as the coefficient of variation (C.V.) from the average net response for each standard. The resulting values will be representative of the noise levels in practical use inherent in a single integration reading for 14 degrees of freedom (d.f.) from which Sn (7 d.f.) for the mean of four readings used throughout this investigation may be obtained by dividing by 2.In practice it is desirable to measure as many samples as possible with each series of standard solutions and hence the time for measurement will exceed 1 h. During this time the instrument response may drift or there may be drift because of the differences in the nature of the samples. The standard deviation for within-series measurements the repeatability so, will consist of such contributions together with that from sn and may be calculated by partial analysis of variance for single readings and hence the mean of four readings (7 d.f.). As linearity has been proved for these calibrations the standard deviation of between-series variation of measurement (which includes the blank variation) the reproducibility s 982 EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR Anhdyst vol.107 for the mean of four readings may be obtained from the difference between net mean individual responses at each concentration with the net average response for individual calibration lines. Each standard deviation Sn so and s for each elemental standard when measurement was possible is recorded in Tables I11 and IV. Because of the insensitivity of the fixed cobalt emission line no information at relevant levels of cobalt can be included. The value for Sn expressed as C.V. decreases with increasing elemental concentration. A constant level is obtained for the major trace elements irrespective of which acid standards are measured and this constant level is of the order of 0.4 yo.This applies over a wide range of concentrations e.g. for a range for iron of 0.5-50 mg 1-l and for zinc of 0.5-25 mg 1-1. Un-doubtedly a similar constant effect would be obtained for the trace elements if measurement was made at elevated levels outside the scope of this investigation. When expressed as con-TABLE I11 CALIBRATION DATA-MAJOR TRACE ELEMENTS Each coefficient of variation is for 7 degrees of freedom. Sensitivity settings for the ICP-OES were reduced for iron and zinc standards in hydrochloric acid (1 + 19). Coefficient of variation of measurement for mean of four integration readings yo Standard solution O* 01* 7 8 9 E 10 F 11 12 G 13 14 H k L M N Copper +-Sn So 0.9 3.6 2.3 3.2 10 31 4.4 13 2.1 4.5 3.6 5.6 1.0 2.8 2.1 2.5 0.6 1.2 0.7 1.6 1.3 1.2 0.9 1.4 0.4 0.8 0.6 0.6 0.3 0.3 0.3 0.4 0.3 0.3 0.3 0.4 0.3 0.5 Iron 7 -7 S Sn So S - 0.4 1.7 -2.4 4.5 - -29 0.9 1.2 3.8 19 0.6 1.2 1.2 4.7 0.4 0.6 0.8 5.0 0.9 0.8 1.6 1.4 0.3 0.6 0.7 1.4 0.4 0.5 1.3 1.0 0.2 0.9 0.8 1.1 0.3 0.9 0.7 1.5 0.3 0.3 2.1 1.8 0.5 1.0 0.7 2.0 0.3 0.8 0.6 0.5 0.3 0.5 0.7 0.8 0.4 0.3 0.5 1.7 0.3 0.4 0.5 1.1 0.3 0.5 0.6 1.3 0.3 0.5 0.6 1.3 0.3 0.4 1.1 Manganese -7 Sn So S 0.2 0.4 -0.3 0.5 -2.3 3.5 24 0.9 1.7 7.2 0.5 1.0 2.3 1.0 1.7 1.1 0.4 0.6 1.5 0.7 1.0 1.1 0.3 1.0 1.1 0.4 0.8 1.6 0.7 1.0 1.3 0.5 1.1 0.9 0.3 0.8 1.6 0.4 0.5 0.4 0.7 0.8 0.6 0.5 1.0 0.8 0.3 1.0 0.2 0.5 0.9 0.8 0.5 0.6 1.1 Standard solution 0.01* 1 2 3 A B * Expressed as micrograms per litre. TABLE IV CALIBRATION DATA-TRACE ELEMENTS Zinc - Sn So 0.8 0.8 1.0 1.7 1.6 1.7 0.8 2.3 0.7 1.7 0.8 1.3 0.3 1.2 0.5 1.1 0.2 1.2 0.3 1.2 0.8 0.8 0.7 1.5 0.4 0.6 0.4 0.6 0.5 0.6 0.4 0.6 0.5 0.7 0.5 0.7 0.4 0.7 I Arsenic & 'n '0 33 32 -5 1 48 - - - -36 44 78 23 30 24 18 29 25 14 15 12 11 13 7.4 44 59 57 30 30 25 33 30 26 2 1 20 9.8 1 3 11 18 - - -Each coefficient of variation is for 7 degrees of freedom. Coefficient of variation of measurement for mean of four integration readings, A Cadmium Chromium Xckel Lead sn so s sg so s s so s s so s 0.6 1.5 - 0.7 2.4 - 1.8 3.7 -1.3 1.6 - 1.4 2.3 - 2.8 2.8 - 28 45 -62 86 150 - - -5 1 90 98 13 27 32 23 32 52 - - -40 89 '72 6.5 23 18 11 24 20 - - -28 38 20 6.5 13 9.0 6.2 17 2 1 - - -29 29 49 4.1 7.3 12 6.5 8.2 14 - - -14 28 43 4 1 6.5 9.1 6 1 8 2 24 'I - - -_ _ _ - - - - - __ ~ .- ~-9.0 ii l o 3.5 3.8 3.7 S:i 16.- -5.3 27 52 65 6.8 7.6 6.9 1.7 1.8 2.1 5.5 5.5 4.6 22 48 42 4.3 5.1 2.3 1.1 1.3 1.1 3.9 3.4 6.2 19 19 15 3.4 4.2 2.7 0.7 1.2 1.3 2.6 2.9 3.2 10 11 17 1.6 3.2 1.9 0.6 0.8 0.9 2.0 2.1 3 .i 9 1 4 25 - - - 0.3 0.4 1.4 1.5 2.1 4.7 5 6 15 I S -5.0 7.4 3.0 1.6 1.1 0.8 0.9 1.4 0.7 0.7 0.7 0.5 1.1 1.2 0.8 1.2 12 Tin - 8.3 7.7 -18 17 -18 2 1 16 'n '0 S - - _ 7.6 6.8 8.1 8.5 8.3 4.8 4.1 4.9 7.3 t: E' 1i.5 } 3.9 3.6 3.5 2.0 1.9 1.9 2.6 2.2 2.0 1.7 1.8 1.5 0.6 1.2 1.9 * Expressed as micrograms per litre September 1982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS 983 centration (see Tables I and I1 for conversion) sI1 is constant for the standard blanks and a very narrow concentration range usually below 0.1 mg kg-l for both trace and major trace elements in either sulphuric or hydrochloric acid solution.The trends for so and s are similar to those for sn but the values expressed as C.V. tend to be higher for the same elemental concentration. Expressed as concentration the narrow constant range obtained for s n is wider and tends to extend over the range 0-0.5 mg 1-1 for each element. The 95 C.L. for the ratios of standard deviations for 7 d.f. against 7 d.f. is 0.45 to 2.23. Inspection of the ratios of so to s n indicates that these are exceeded in half of the comparisons for the major trace elements in sulphuric acid.Similar inspection of ratios of s to so indicate a similar number for manganese and zinc in sulphuric acid and copper and iron in hydrochloric acid. For the trace elements (Table IV) these ratios are seldom exceeded except for the highest standard F. It would appear that as sn declines to a constant low value instrument drift becomes significant and this is indicative of indirect interference which is to be discussed later. Interferences I t has been previously recorded that for a system of measurement involving flame or flameless atomic-absorption spectrometry interferences may be of four distinct types30 : A. Directly upon the response of a measured element in the presence of an interfering species.B. Indirectly on the variation of an element response in the presence of interferingspecies. C. Indirectly upon the variation of subsequent element response during an extended D. Indirectly on the response of subsequent measurement in an extended series of For measurement by the direct-reading ICP-OES direct interference A has to be sub-A l . From nebulisation and transport into the plasma of acidic solutions of different A2. From nebulisation and transport into the plasma of solutions containing varying A3. From compound formation or ionisation effects from other inorganic species. A4. From spectral effects from other inorganic species. series of measurements in routine use. measurements.divided into four categories as follows : concentrations. amounts of total dissolved solids. Direct interference A1 The changes in response caused by transport of and by the viscosity of different acid c o n c e n t r a t i o n ~ ~ ~ ~ ~ ~ ~ were considered for a narrow concentration range of both sulphuric and hydrochloric acids for each of the major trace elements using standard solutions 9,14 F and H. Net responses were obtained by subtraction of the blank from 1 + 19 sulphuric and hydrochloric acid to reflect actual practice but as expected blank responses at different acid concentrations differed in magnitude only slightly. The average ratios for the four major trace elements each at two concentrations for the different acid concentrations are shown in Table V and indicate the systematic bias likely from small changes in acid concentration.The experimental work described in this account was made at fixed acid concentrations and the extraction procedure similarly measured solutions of constant acid strength. For TABLE V CHANGES IN ELEMENTAL RESPONSE FOR DIFFERENT ACID CONCENTRATIONS Sulphuric acid Hydrochloric acid concentration Ratio* Range concentration Ratio* Range 1 + 19 1.00 - 6 $- 94 0.98 0.9 5-0.99 1 + 24 1.04 1.02-1.06 1 + 19 1.00 -1 + 32 1.09 1.06-1.14 1 + 24 1.02 1.01-1.03 1 + 49 1.13 1.10-1.21 3 + 97 1.04 1.03-1.05 * This refers t o the average ratio of eight responses for the major trace elements compared with that in acid (1 + 19) 984 Analyst VoZ. 107 acid digestions prepared in sulphuric acid some loss would be expected depending on the time taken for digestion.While measurement by atomic-absorption spectrometry (AAS) has been shown to be stable over the range 3-5y0 V/V sulphuric acid this would not be so for measurement by ICP-OES. EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR Direct interferences A2 and A3 The major elements (species) in foodstuff digests are sodium potassium calcium mag-nesium and phosphorus. These were tested for interference individually for the highest amounts likely and when necessary at a half and a quarter of these amounts on standards 0 9 and 14 for the major trace elements in sulphuric acid (1 + 19) and similarly on standards 01 F H and N in hydrochloric acid (1 + 19). For the trace elements 0 3 6 and 01 B and F (E for cadmium) were similarly tested.Phosphorus was added as sodium dihydrogen orthophosphate. Three or four series of solutions were measured with relevant standards within the same series of readings. The elemental responses of solutions 0 or 01 plus added species were subtracted from solutions 9 F etc. if greater than standards 0 or 01 ; standards 0 or 01 were subtracted if these were greater. The total dissolved solids (TDS) (maximum 1.1% m/V) were calculated in terms of the acidic anionic species and as orthophosphoric acid for phosphorus (for the latter the contribution of sodium in orthophosphate salt was sub-tracted). The relative differences in response with and without added species are illustrated in Table VI for the major trace elements. The 95 C.L.were calculated from so (Table 111) and are restrictive as they do not contain the variation from the blankor contributions from s o h t ion preparation. Similar tests were made for sulphuric acid solutions containing separately added aluminium (0.5 mg) silica (0.5 mg) tin (0.5 and 1.0 mg) and as relevant iron (2 mg) and a combination of copper and manganese (0.2 mg) with zinc (1 mg). Also tested were hydrochloric acid solutions containing separately added aluminium (5 mg) silica (5 mg) tin (5 and 10 mg) and , as relevant iron (20 10 and 5 mg) copper (1.5 and 0.5 mg) manganese (1.5 and 0.5 mg) and zinc (7.5 and 2.5 mg). None of these added species contributed more than 0.07% m/V TDS in solution. The 95 C.L. were exceeded only occasionally for these added species and all figures are included in the over-all non-significant averages at the foot of Table VI.Conversely most of the figures in Table VI exceed the 95 C.L. in a negative direction. The figures display the random distribution expected from such an exercise particularly for copper but closer inspection suggests the following conclusions. (a) There is a tendency for the negative interference for each species to be proportional to the amount tested, expressed as TDS irrespective of the acid solution measured. (b) The interference from calcium and magnesium is greater per unit amount than from sodium potassium and phos-phorus. (c) The degree of interference from an inorganic species is relatively constant irrespective of the concentration of the major trace element measured.(d) Interferences from mixtures appear to be cumulative of the individual species. If these interferences are expressed as milligrams per kilogram of TDS required to give 1 yo depression in response using for each acid solution the highest concentration standard for each major trace element and the highest amount of interfering species (for magnesium 25 mg were used in hydrochloric acid), the results shown in Table VII are obtained. It seems likely that a 1% depression is obtained from about 1000 mg kg-l of TDS for sodium and potassium and therefore another source of depression must exist for calcium which applies consistently to each of the major trace elements. I t is suggested that this is an ionisation interference but this seems unlikely for copper as an atom emission line is measured.A similar comment applies to magnesium except that the extent differs for iron and is probably absent for copper. For phosphorus a reduced and variable depression from TDS is apparent and another anomaly is iron for which the depression is less for sodium and potassium. Table IV indicates that the extent of these interferences would not be identifiable within the large values likely for 95 C.L. for the trace elements with the exception of chromium in hydrochloric acid in which instance the trends appeared similar to those for iron. Although these effects are often mentioned in the literature no attempt is usually made to quantify them. The results recorded here may be singular to the nebuliser and instrument used. For the purposes of this investigation when required the combined direct inter TABLE VI INTERFERENCE EFFECTS A2 AND A3 ON MEASUREMENT OF COPPER, IRON MANGANESE AND ZINC Relative difference in response with and without added elemental .A Manganese r Element .. . . . . . . Copper Iron \ I - A > -r- r-7 I- v- +-7 r - Acid? . . . . . . . . 1 2 1 2 Concentration/mgl-l 0.05 0.5 0.1 0.5 5.0 0.5 5.0 1.0 5.0 50.b 0.05 0.5 0.1 95% confidence limits yo . . f10.6 f l . 9 f6.0 11.4 f1.2 f1.4 11.9 f1.2 1 1 . 2 1 1 . 0 1 2 . 4 51.9 f2.4 Added Amount1 TDSG x ion Na+ K+ . . Ca2+ . . Mgz+ . . P PO,^-) Cat+ . . Mg2+ Ca2+ . . Mgz+ Ca2+ . . Mg2+ Over-all average nbsignificant P PO,^-')' P (PO4+* P ( ~ 0 ~ 3 - j ' differences11 . . . . . . +2.8 -0.6 -2.5 0.0 +0.1 +0.5 -0.1 $0.1 -0.1 -0.5 0.0 -0.1 -0.2 mgt ' lOS/mg 1-1 200 6.2 100 2.5 200 4.5 100 1.9 40 1.4 20 0.6 50 2.5 25 1.2 12.5 0.6 200 6.1 100 3.1 40 50 6.9 5.4 50 20 25 3.1 50 - 12 - 5.9 - 15 - 5.1 ---- 8.6 + 7.4 - 1.1 - 7.5 -- 6.1 -- 4.57 - 3.97 --- 2.97 - 2.77 -1.5 - 7.1 - 3.27 --- 4.57 - 5.57 --- 10 -(- 16) (- 8.0) (-1.0) -- 3.2 - 6.0 -+1.5 --- 11 -- 3.8 -2.4 - 0.8 - 0.4 - 0.8 + 0.5 - 0.6 ------ 1.1 -- 3.97 -- 2.57 - 2.47 + 1.9 -1.4 + 0.2 - 0.8 ----- 1.67 - 1.0 - 4.3 - 3.6 ---- 6.7 - 3.8 - 0.7 - 3.9 --9.6 - 9.0 -- 4.27 - 2.67 - 3.37 - 4.41 -2.97 - 1.7 - 3.67 ------ 3.7 -(-3.5) -(-2.6) - 1.9 + 0.4 - 1.1 (+ 0.4) --- 2.1 -1.1 -1.2 - 1.2 +0.2 + 0.3 - 1.6 ---- 7.6 - 7.8 - 7.1 --2.47 --0.9 --2.07 --1.1 -8.5 -1.6 -12 1 0 .3 -2.0 - 5.0 -1.2 -----8.47 - --9.27 - -- -5.0 -3.1 - 13 -- 8.6 --3.57 --5.37 --4.47 --4.47 --6.07 (-6.8) -1.67 -6.5 -1.4 -2.7 - 0.5 - -4.8 - (- 4.9) - (- 3.9) -4.77 --107 ---117 -- -8.2 -* ( t Acid 1 is sulphuric acid (1 + 19). acid 2 hydrochloric acid (1 + 19). $ For 100 ml of digest prepared frok 10 g of foodstuff the concentration of interfering species as mg kg-l can be obtained by multiplying by 5 TDS equals total dissolved solids calculated in term'of the anion of the measured acid solution with phosphorus calculated as orthophosphoric fi < .5y0 significance.IlThis average includes differences when aluminium silica tin and relevant major trace elements are added at levels described in the text. ) Implies a duplicated set of readings. acid; in hydrochloric acid the TDS is 2.0 1.0 and 0.5 x 10' mg 1-' for 50 25 and 12.5 mg of added magnesium respectively 986 EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR Analyst Vol. I07 TABLE VII INTERFERENCE EFFECTS A2 AND A3 ON COPPER IRON MANGANESE AND ZINC EXPRESSED AS TOTAL DISSOLVED SOLIDS Added ion Na+ . . K+ . . P ( ~ 0 ~ 3 - 3 . . P (Na+)* . . Mg2+ . . Ca2+ + Mg2+ < 'P ( ~ 0 ~ 3 -Ca2+ . . . . Total dissolved solids which gives 1% depression/mg 1-1 Copper Iron Manganese Zinc . . 1010 1260 990 1320 960 1900 930 960 .. 2900 2 140 2 200 - 600 650 520 1380 360 360 340 290 820 600 400 400 -) . . . . 1480 770 650 520 I A 3 * Calculated on the total dissolved solids from sodium included with the phosphorus and the total depression from sodium and phosphorus. ferences A2 and A3 were used to correct for each of the major elemental species at a constant percentage level irrespective of the major trace element concentration present. In practice, for digests from foodstuffs this positive correction should not exceed 10%. Direct interference A4 In the testing of the above interferences measurement of solutions 0 and 01 with con-comitant elemental species permits an assessment of any direct spectral interferences A4. When the difference between these solutions and standards 0 and 01 exceeded 95 C.L.positively for a trace or major trace element the content of the elemental species was checked independently by flame atomic-absorption spectrometry for impurities. These could be of the order of 1 x 10-5-1 x mg kg-l per mg kg-l and after subtraction the remainder if present was considered to be defined by direct interference A4. The occasions satisfying this criterion are listed in Table VIII. In general these factors were relatively constant for the different levels of interfering species but more reliance would be placed on the higher levels tested; there was no significant difference for the factors calculated from 0 in sulphuric and 01 in hydrochloric acid solution. Both interferences for zinc have been previously defined as spectral line coincidence at 202.55nm but that from magnesium for the remainder of the elements may be caused by primary stray light in the instrument used for this investigation.For the low wavelength of measurement of arsenic and tin broad band continua are probably the cause of interference, while the interference from the crowded iron emission is not surprising. Interference from manganese has not been previously recognised. The list in Table VIII does not reflect an exhaustive trial of elemental species or combinations of these species in digests from food-stuffs. It may be noted however that the interference from magnesium is unaltered in the TABLE VIII SPECTRAL INTERFERENCES A4 Results are milligram per kilogram of element per milligram per kilogram of interfering element.Interfering element Element Copper . . Iron . . Zinc . . . . . . Cadmium . . . . Chromium . . . . Nickel . . . . Lead . . . . Arsenic . . Tin . . Magnesium -2.1 x 10-4 2.6 x 10-5 1.6 x 10-4 1.1 x 10-4 4.4 x 10-4 2.8 x 10-3 1.2 x 10-3 Iron Manganese Aluminium Coppe; - - 3.5 x 10-5 -- - 1.0 x 10-3 -- - - 5.4 x 1 0 - 3 1.4 x 10-4 2.0 x 10-4 9.0 x 10-5 -1.6 x 10-3 3.0 x 10-4 - -- - - -- - - -2.0 x 10-3 6.8 x 1.2 x -4.4 x 10-4 3.8 x 10-3 1.3 x 10-3 Se#tember 1982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS 987 presence of calcium and phosphorus. These spectral interferences will have much greater effect for trace elements such as arsenic and cadmium present at very low concentrations in foodstuffs.The corrections indicated in Table VIII are negative concentration corrections and when required should be applied prior to and separately from that of direct interferences A2 and A3. Indirect interferences Indirect interference B may be assessed from the calculation of s n and so for solutions 9, F etc. containing high levels of species which cause direct interference A2 and A3. The 95 C.L. of the ratio of standard deviations for these solutions with those of standard solutions in hydrochloric acid were not exceeded but in sulphuric acid they were exceeded for zinc and tin measurements. This was traced to solutions containing calcium and would be caused by intermittent coating of the plasma torch injector tip. The main source of interference B is however more subtle.The randomness of results in Table VI has been noted for the lowest concentrations of the major trace elements e.g. copper at 0.05 and 0.10 mgl-I. For solutions 0 and 01 with added elemental species occasions when 95 C.L. were exceeded positively have been dealt with by direct interference A4. There were also several occasions when the 95 C.L. were exceeded negatively. In sulphuric acid this occurred 5 out of 39 times for the major trace elements and once out of 47 times for the trace elements; no interfering species was unduly represented. For hydrochloric acid the incidence was 8 out of 106 for the major trace elements and 17 out of 147 for the trace elements; potassium was unduly represented in the latter (12 times). I t has been noted that changes in acid strength do not alter the spectral background.The presence of some major elemental species would seem sometimes to increase the noise in the spectral background. This indirect interference B while having a diminishing effect at higher measured levels will increase sharply variation i.e. random bias, as measured levels decrease towards the detection limit. Indirect interference D and hence C or C alone is always present in a measuring system involving solution aspiration with existing nebulisers into atomic-absorption or in electrothermal atomic-absorption spectrometry and will increase the variation of measurement a t all concentration levels. These interferences are also apparent in solution nebulisation into the plasma as the calibration data have indicated.In the measurement of sample solutions against such calibration solutions these indirect interferences may be doubly represented. Corresponding increases in derived functions such as confidence intervals and detection limits will ensue (such indirect interferences are absent in many analytical techniques involving discontinuous measurement). The cumulative effects of these indirect interferences together with any direct interferences as yet unrecognised will become apparent only during application of the direct-reading ICP-OES to the measurement of foodstuff digests. Application to Analysis of Foodstuffs The reproducibilities s obtained for standard blank and standard solutions up to 0.5 mg kg-l are constant and permit derivation of elemental limits of detection according to a previously described system based on the formula t,~,,,s.31 These will be only an indication of what may be achieved in practice and may be compared with those actually achieved in application to foodstuff digests by atomic-absorption spectrometric measurement (Table IX).The latter are desired practical limits of detection (for 100 ml of digest from 10 g of foodstuff), but for cadmium chromium lead and nickel allow for a concentration stage prior to measure-ment. Clearly for levels present in foodstuffs only major trace elements and for canned goods tin could be measured directly by the direct-reading ICP-OES. Measurements were made on sulphuric acid digests from reference materials for the major trace elements against multi-elemental standards 0-14 (Table I) as a matter of interest.These values were corrected for positive spectral interference A4 and then for negative direct interferences A2 and A3. The resulting values should reflect direct interference A l and from previous experience for the food matrices concerned should not give ratios to the actual contents greater than 1.09 (Table V). The results of this comparison are shown in Table X. Copper cannot be measured but for the other three elements the expected range of ratios i 988 EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR Analyst VoZ. I07 TABLE IX COMPARISON OF LIMITS OF DETECTION Values by AAS are in application to food digests; for cadmium chromium, lead and nickel they include a 7-fold concentration stage and for tin a 5-fold d i l ~ t i o n .~ * - ~ ~ For 100 ml of digest prepared from l o g of foodstuff used generally in this laboratory concentration as milligrams per kilogram in original foodstuff may be obtained by dividing by 10. Amount per 100 ml solution/pg f A 7 ICP-OES ISulphuricHydrochlonC Element AAS acid (1 + 19) acid (1 + 19) Copper . . . . Manganese . I Iron Zinc . . Cadmium . . Chromium. . . . . . Lead . . Nickel . . Arsenic . . Tin 1 .o 2.0 4.0 0.08 0.2 0.6 0.6 0.2 1.2 10 0.8 1.0 0.5 2.7 0.4 0.6 1.8 8 3.4 -0.7 3.0 0.5 2.0 0.3 0.6 2.1 7.3 12 17 acceptable only for flour and NBS liver. The rapid destruction of organic matter in 1 g of dried kale and spinach could hardly result in loss of half of the sulphuric acid in the resulting digest.The conclusion must be that even after correction of known direct interferences the values obtained continue to reflect considerable positive systematic bias (dilution of these digests would yield solutions beyond useful detectable levels for copper and zinc). Although it has been suggested that effects from direct interferences A1 may be minimised by addition of high concentrations of say sodium it is evident that the trace and major trace elements should be removed from interfering species available in faodstuff digests and solubilised at a constant acid concentration thereby removing direct interferences Al-A3. This can be achieved with a chelating ion-exchange column,32 or by chelation and extraction into solvent.33 Each has the added merit of concentrating trace elements to levels permitting measurement by the ICP-OES.I t was established that organic solvents unbalanced the background continuum of the plasma. Back-extraction into a fixed acid concentration was therefore not permissible and it was necessary after chelation and extraction to remove the solvent oxidise with nitric acid to dryness and solubilise in acid to give a final hydrochloric acid (1 + 19) solution. This pre-treatment for a sample series took one man-working day and involved considerable manipulation. Chelation and extraction of standard solutions 0 3 6 9 and 14 (Table I) in sulphuric acid (1 + 19) followed by measurement with standards 01-N indicated that manganese and zinc were not separated.Adjustment of solutions to pH 4 and similar trials were more successful and the following average recoveries were obtained for standard solutions 0-14 (Table I ) Cu 101; Fe 99; Mn 102; Zn 104; Cd 105; Ni 97; Pb 83; Sn 69; Cr 0; and As 0%. Chelation and extraction were selected for trial. TABLE X RATIO OF RESULTS OBTAINED BY DIRECT MEASUREMENT WITH ICP-OES TO THOSE OBTAINED BY FLAME AAS ON FOODSTUFF DIGESTS Foodstuff material Copper Iron Manganese Zinc Dried milk . . . . . . 6.7 1.36 0.91 1.08 Flour . . . . 1.43 1.05 1.04 1.09 NBS liver . . . . . . 1.05 1.01 0.99 1.04 Bowen's kale . . . . . . 6.7 1.18 1.29 1.21 NBS tuna . . . . 1.15 1.11 1.20 1.12 NBS spinach . . . . 1.64 1.07 1.09 1.0 Se@tember 1982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS 989 The fixed insensitive line used in the ICP-OES precluded measurement of cobalt.Chromium(V1) may be chelated with dithiocarbamates but it is not possible to obtain and retain this state in foodstuff digests; reduction to chromium(II1) followed by a separate chelation with pentane-2,4-dione is necessary.34 Digestion with nitric acid yields arsenic(V), and this valence state is incompatible with the described procedure35 ; further the detection limit would preclude the determination of arsenic in most foodstuffs. The efficiency of the procedure for lead and tin was disappointing and when applied in practice to foodstuff digests iron was only partly recovered because of precipitation at pH 4 or reduction to iron(I1) by the dithiocarbamates. This leaves five of the original eleven elements plus lead, which was persevered with.The comment must be made that possession of a multi-elemental measuring facility does not automatically ensure measurement for a wide range of elements; in practical use it may differ little from a sequential measuring facility. The recovery of an analyte from substrates will only prove the accuracy of a method. The accuracy of results can only be proved by obtaining agreement with certified levels of standard reference materials. For the ICP-OES spectral interference in particular would not effect elemental recovery and hence an exercise to establish the recovery of added elements in discrete form from foodstuffs was not carried out. Two reference materials used for inter-analyst assessment (dried milk and white plain flour) Bowen’s kale and three NBS reference materials (bovine liver tuna and spinach) were used as standard reference materials.Each has been used extensively for the development of methods using atomic-absorption spectrometric measurement and subsequently to prove the accuracy of results during the monitoring of foodstuff homogenates. Digests were prepared using masses of dry materials simulating corresponding wet masses and total analyses were made in triplicate by the two analysts concerned. Each result was corrected for relevant known direct interferences A4 according to the factors in Table VIII. Analyst significance was present in 4 out of 48 element material combinations two of which involved lead (the range of cadmium is elevated to that normally found in foodstuffs).Means and 95% confidence intervals (95 C.I.) for the means obtained by this procedure are displayed in Table X I and compared with means and 95 C.I. for 6-10 total analyses obtained by existing procedures with measurement by atomic-absorption ~pectrometry.~8*~9 Table XI1 displays in greater detail the results by the present procedure but inclusion of this table does not endorse the method. Comparison of the two sets of means in Table XI indicates the number of occasions dis-agreement exists 18 out of 36. Enhanced means using ICP-OES measurement are rare, e.g. cadmium and lead in bovine liver. That for cadmium is singular and it would appear that this material may contribute spectral interference at the ion-emission line for cadmium.Those for lead are not surprising at a first impression because of the lengthy pre-treatment and the ubiquitous nature of lead (see later). Most of the disagreements are negative and again the lengthy pre-treatment must be questioned. The remnants of the extracts were therefore measured by flame atomic-absorption spectrometry for three reference materials. For zinc in kale and NBS tuna the values were in agreement with measurement by ICP-OES but that for NBS spinach was increased to 46 mg k g l while the copper values increased to agree with those previously obtained for kale and NBS tuna but only to a value of 11 .O mg kg-l for spinach. While manganese and cadmium in NBS spinach are in agreement, it would appear that copper zinc and nickel are incompletely chelated for this material.For copper it would appear that the measurement is also at fault. The general lack of agreement with established levels in these reference materials given by the present procedure suggests that it is inaccurate and hence unacceptable. The decreasing gradation of repeatability so and reproducibility s with increasing amount may be judged from Table XI1 for each element; copper is particularly uneven. Comparison of so and s of results expressed as C.V. (relevant concentrations may be obtained by multi-plying amount by 0.07) with that for measurement of standard solutions (Tables I11 and IV) at similar concentrations indicate that the 95 C.L. of ratios of standard deviations are in-variably exceeded. While the lengthy pre-treatment could contribute to these increases in so and s indirect interferences D and C during measurement are probably the main contri-butors.Lead is an exception to the above and a high C.V. is involved with measurement; this is a contributory cause to the disagreements with accepted values previously noted. Both procedures using measurement by ICP-OES and AAS give similar variations reflected Nevertheless more information may be gathered from consideration of variation 990 EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR Analyst Vvol. 107' TABLE XI EVALUATION OF ACCURACY OF METHOD I N APPLICATION TO REFERENCE MATERIALS Concentration/mg kg-1 A r 7 Reference material Sample mass/g . . Copper-Mean by ICP f 9 5 C.I. of mean Mean by AASS . . &95 C.I. of mean Mean by ICP .. 1 9 5 C.I. of mean Mean by AAS . . &- 95 C.I. of mean Mean by ICP . . &95 C.I. of mean Mean by AAS . . f 9 5 C.I. of mean Mean by ICP . . 1 9 5 C.I. of mean Mean by AAS . . 1 9 5 C.I. of mean Mean by ICP &95 C.I. of mean Mean by AAS . . 1 9 5 C.I. of mean Mean by ICP . . f 95 C.I. of mean Mean by AAS . . &95 C.I. of mean Manganese-Zinc-Cadmium-Lead-Nickel-. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dried milk . . 2.5 . . 0.37t . . 0.04 . . 0.63 . . 0.07 . . 0.32t . . 0.02 . . 0.76 . . 0.10 . . 43.4t . . 1.8 . . 45.8 . . 1.2 . . 0.005 . . 0.005 . . 0.003 . . 0.004 . . 0.22t . . 0.25 . . 0.02 . .0.03 . . 0.03 . . 0.05 . . 0.03 . . 0.04 Flour 10 1.32t 0.04 1.36 0.02 5.17 0.08 5.25 0.08 5.257 0.17 5.38 0.08 0.034 0.007 0.037 0.003 0.08t 0.09 0.03 0.03 0.037 0.03 0.01 0.01 * Certified and non-certified (in parentheses) levels for NBS bovine liver and spinach are as follows: bovine liver Cu 193 Mn 10.3 Zn 130 Cd 0.27 P b 0.34 mg kg-l; spinach Cu 12 Mn 165 Zn 50 Cd (1.5) Pb 1.2 and Ni (6) mg kg-l. Disagreements with the means by AAS for bovine liver for Mn and Zn have been discussed in reference 29. 1 These values are outside expected levels imposed by 95 C.1.s of procedures using AAS measure-ment. $ All results are calculated on a dry-mass basis. With the exception of dried milk flour and in part spinach for which means were obtained during this investigation using AAS measurement the remainder of the means for AAS measurement are those quoted in references 28 and 29.Cadmium measurement for bovine liver kale and spinach was made on an aliquot of that normally taken and calculation of 95 C.I. based on amount would be reduced accordingly e.g. to 0.03 0.03 and 0.05 pg, respectively compared with 0.06 0.09 and 0.11 pg by ICP measurement. 5 This result was obtained by applying the procedure to a 10-ml aliquot of digest. by the 95 C.I. of means displayed in Table XI for copper manganese and zinc. A 7-fold concentration for measurement by ICP-OES is compared with no concentration for measure-ment by AAS. It may seem strange that the former measurement does not give improved variations but the standard deviation of measurement is constant over an extremely wide range.Concentration of elements classified as major trace elements present well above detection limit levels therefore does not confer any improvement in the variation of measure-ment or the derived confidence intervals. For cadmium lead and nickel comparison is for approximately the same concentration factor (for cadmium the footnotes to Table XI apply for AAS measurement). Detection limits for results for these three elements may be derived from the three reference materials having the lowest s in terms of amount via the variance, for 15 d.f. (materials returning zero results must be excluded). These derived values are cadmium 0.15 lead 2.4 and nickel 0.5pg and may be compared with the detection limits NBS bovine liver* 2 194 6 198 7 l0.2f 0.4 11.0 0.4 139 5 142 4 0.361 0.028 0.25 0.025 0.43t 0.53 0.25 0.04 0.05 0.05 0.05 0.06 NBS tuna 5 3.07t 0.16 3.27 0.08 0.51 0.01 0.51 0.12 14.2 0.2 14.7 0.5 0.042 0.007 0.046 0.010 0.53 0.06 0.50 0.07 1.01 0.10 1.10 0.13 Bowen's kale 1 4.25t 0.18 4.99 0.33 15.4 1 .o 15.6 0.8 32.0 3.0 33.1 1.4 1.02 0.09 1.01 0.12 2.36 0.56 2.23 0.19 0.94t 0.10 0.78 0.10 NBS spinach* 1 10.2t 1.0 12.1 0.4 1675 7 170 4 43.lt 4.0 50.8 1.9 1.39 0.11 1.40 0.14 1.75t 1.33 1.10 0.06 4.8t 0.7 5.8 0.September, Element Copper Manganese .. Zinc . . Cadmium . Lead Nickel . . I982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS TABLE XI1 REPLICATE ANALYSES FOR TRACE AND MAJOR TRACE ELEMENTS IN REFERENCE MATERIALS Each array represents six results obtained by two analysts in triplicate. Reference material NBS spinach* Bowen’s kale NBS spinach Flour NBS tuna NBS liver* NBS liver . . Dried milk NBS tuna NBS liver* Bowen’s kale NBS liver NBS spinach* Flour . . NBS spinach* Bowen’s kale NBS liver* KBS spinach Flourt NBS tuna Dried milk NBS liver . . Dried milk NBS liver* NBS tuna NBS spinach*? Flour NBS liver Bowen’s kale NBS spinach . . Dried milk Flour NBS liver? NBS spinach? Bowen’s kale NBS tuna . . Dried milk NBS liver Flour NBS spinach* Bowen’s kale NBS spinach NBS tuna .Dried milk Sample mass/g 2.5 1 1 1 10 5 2 2 2.5 5 2 1 2 1 10 1 1 2 1 10 5 2.5 2 2.5 2 5 1 10 2 1 1 2.5 10 2 1 1 5 2.5 2 10 1 1 1 5 Mean content/ Range of content/ Amount/ mg kg-l mg kg-l 0.37 4.25 1.32 3.07 11.0 10.2 171 194 0.32 0.51 10.3 15.4 10.2 167 5.17 43.3 32.0 43.1 14.2 43.4 126 5.25 139 0.005 0.39 0.042 1.55 0.034 0.36 1.02 1.39 0.22 0.084 0.43 1.75 2.36 0.53 0.03 0.05 0.03 5.40 0.94 4.83 1.01 0.32-0.43 8.0-1 3.7 3.96-4.42 9.0-11.6 2.8 2-3.1 7 187-202 0.30-0.34 0.49-0.52 9.7-10.8 9.8-10.6 5.08-5.25 36.2-49.7 27.7-35.6 106-140 38.7-48.4 5.10-5.43 40.9-45.7 134-147 1.28-1.38 139-194 14.5-16.5 156-1 7 3 14.1-14.4 0-0.011 0.19-0.67 0.037-0.048 1.16-2.49 0.026-0.042 0.31-0.39 0.90-1.13 1.24-1.5 2 0-0.63 0-0.16 0-1.13 0.44-3.32 1.6-2.9 0.48-0.62 0-0.11 0.0 2-0.12 0-0.06 2.97-7.49 0.85-1.09 0.91-1.16 3.86-5.i1 WLg 0.92 1.58 4.25 10.2 13.2 15.3 48.8 388 0.79 2.53 2.95 15.4 20.3 23.9 51.7 6.2 32.0 35.9 43.1 52.5 71.5 108 277 0.013 0.11 0.21 0.22 0.34 0.72 1.02 1.39 0.55 0.84 0.86 1.75 2.36 2.65 0.07 0.11 0.29 0.77 0.94 4.83 5.05 99 1 Repeatability Reproducibility 7-7 - YO 11 15 3.3 9.1 2.5 3.3 2.9 4.9 2.2 3.5 3.9 3.3 3.8 1.4 7.4 9.3 8.9 1.6 0.7 3.7 3.5 11 12 -4 .i 15 17 20 7.4 8.2 7.3 ---30 17 11 6 1 29 14 --9.8 9.5 wL6 Yo 0.10 11 0.24 18 0.14 4.0 0.93 9.1 0.33 2.8 0.50 5.0 5.4 16 11.2 2.9 0.039 5.6 0.055 2.3 0.10 4.2 0.60 6.0 0.67 3.3 0.91 3.8 0.73 1.4 0.8 12 2.4 9.0 3.3 12 3.8 8.9 0.9 3.0 0.5 1.0 4.0 3.9 9.6 3.5 0.012 -0.050 57 0.031 15 0.037 46 0.067 20 0.053 7.4 0.084 8.2 0.100 7.3 0.60 -0.81 -0.48 -0.53 73 0.39 22 0.28 11 0.11 -0.07 80 0.17 -0.23 40 0.10 9.8 0.69 14 0.48 9.5 0.10 0.28 0.17 0.93 0.37 0.75 7.8 11.2 0.045 0.059 0.12 0.93 0.67 0.91 0.73 0.8 2.9 4.3 3.8 1.6 0.7 4.3 9.6 0.012 0.063 0.031 0.100 0.067 0.053 0.084 0.100 0.60 0.81 1.00 1.27 0.51 0.28 0.11 0.09 0.23 0.31 0.10 0.69 0.48 * These results were obtained by applying the procedure to a 10-ml aliquot of digest.t 5% analyst significance. shown in Table IX. That for nickel is comparable to measurement by AAS but for lead and cadmium they do not permit the determination of the levels normally found in dietary foodstuff homogenates. Conclusion The attempted application of an inductively coupled plasma optical emission direct-reading polychromator has been described for the simultaneous measurement of the major trace elements copper iron manganese and zinc and the trace elements cadmium lead, nickel chromium arsenic and tin. The measurement of multi-elemental standards has been assessed for a situation simulating routine practice for linearity and variation.The latter has been defined by the standard deviations representing instrument noise the repeatability and the reproducibility. Each is as low as or lower than literature values where these exist, for the instrument used. From the reproducibility limits of detection have been derived that permit an assessment of the application of measurement. From both repeatability and reproducibility confidence limits have been calculated that permit the assessment of inter-ferences likely to be encountered in measurement. Direct interference A1 is that from nebulisation and transport of different acid concentrations; A2 similarly is These interferences have been categorised into seven distinct classes 992 EVANS AND DELLAR ICP EMISSION SPECTROMETER FOR A%Ulyjst Vd.107 that for solutions containing different levels of total dissolved solids; A3 is that from ionisa-tion and compound formation effects with other inorganic species; and A4 is that from spectral effects from other inorganic species. Direct interference A1 may be avoided by using a constant acid strength; A4 may be corrected by a direct concentration factor for each interfering species; A2 and A3 combined may be corrected for each interfering species by means of a percentage correction applied separately to and after correction for A4. The three categories of indirect interferences will enhance the variation of results at all levels of analyte element concentration but in particular at low measured concentrations.Chelation and extraction followed by oxidation and solub ilisation at a fixed hydrochloric acid concentration have been used to avoid the worst excesses of these interferences and to concentrate the trace elements to levels that permit measurement. Of the elements studied, only copper manganese zinc cadmium nickel (and lead) are amenable to this treatment. When the total procedure is applied to standard foodstuff reference materials no element satisfies the criterion of accuracy required for quantitative analysis necessary for dietary survey exercises. The variation of measurement by the ICP-OES is extremely low but in application is sharply increased. Comparable confidence intervals to those for results by AAS measurement however may be attained for copper manganese zinc and nickel.For cadmium and lead the enhanced variation of measurement does not permit determinations at the levels normally found in foodstuffs. While both measurement by ICP-OES and the concentration technique (which is widely used) may have been at fault in this exercise simultaneous measurement for the quantitative analyses of low levels of elements in foodstuffs does not appear to confer the advantages sought. It is possible that optimisation of the conditions of measurement with a mono-chromator ICP-OES would remove some of the anomalies described i.e. hitherto un-recognised spectral interferences and interferences due to ionisation or compound formation from particular combinations of inorganic species permitting sequential measurement of some of the elemental levels required.This paper is published with the permission of the Government Chemist. 1 . 2. 3. 4. 5. 6. 7. 8. 9. 10. 1 1 . 12. 13. 14. 15. 16. 17. 18. 19. SO. 21. 22. 23. 24. 25. 26. 27. 28. 29. References Mitchell R. L. “The Spectrochemical Analysis of Soils Plants and Related Materials,” Common-Greenfield S. Jones I . L. and Berry C. T. Analyst 1964 89 713. Wendt R. H. and Fassel V. A. Anal. Chem. 1965 37 920. Fassel V. A. and Kniseley R. N. Anal. Chem. 1974 46 lllOA. Greenfield S. Jones I. L. McGeachin M. M. and Smith P. B. Anal. Chinz. Acta 1975 74 225. Winge R. K. Fassel V. A. Kniseley R. N. DeKalb E. and Hass W. J. Spectrochim. Acta Part Garbarino J . R. and Taylor H. E. Appl. Spectrosc. 1979 33 220. Berman S. S. McLaren J . W. and Willie S. N. AnaE. Chem. 1980 52 488. Ilahlquist R. L. and Knoll J. W. Appl. Spectvosc. 1978 32 1 . &Quaker N. R. Kluckner P. D. and Chang. G. N. Anal. Chem. 1979 51 888. McHard J. A. Foulk S. J Nikdel S. Ullmann A. H. Pollard B. D. and Winefordner J. D., Larson G. F. Fassel V. A. Scott R. H. and Kniseley R. N. Anal. Chem. 1975 47 238. Boumans P. W. J. M. and de Boer F. J. Spectrochim. Acta Part B 1975 30 309. Boumans P. W. J. M. and de Boer F. J. Spectrochim. Acta Part B 1977 32 365. Kornblum G. R. and De Galan L. Spectrochim. Acta Part B 1977 32 455. Greenfield S. McGeachin H. M. and Smith P. A. Anal. Chim. Acta 1976 84 67. Larson G. R. Fassel V. A. Winge R. K. and Kniseley R. N. Appl. Spectrosc. 1976 30 384. Fassel V. A. Katzenberger J . M. and Winge R. K. Appl. Spectvosc. 1979 33 1. Hassell K. D. Rose D. A. and Warren J. I C P I n f . Newsl. 1979 4 343. Taylor C. E. and Floyd T. F. Appl. Spectvosc. 1980 34 472. Larson G. F. and Fassel V. A. Appl. Spectvosc. 1979 33 592. Ediger R. D. and Hoult D. W. A t . Spectrosc. 1980 1 41. Fisher C. G. Barrett P. and Ediger R. D. A t . Spectvosc. 1980 1 153. Boumans P. W. J. M. Spectvochim. Acta Pavt B 1976 31 147. Analytical Methods Committee Analyst 1975 100 899. Hassell K. D. Rose D. A. and Warren J. I C P I n f . Newsl. 1978 4 261. Analytical Methods Committee A n d y s t 1960 85 643. Evans W. H. Read J. I. and Lucas B. E. Analyst 1978 103 580. Evans W. H. Dellar D. Lucas B. E. Jackson F. J. and Read J. I. Analyst 1980 105 529. wealth Agricultural Bureaux Farnham Royal 1964. B 1977 32 327. Anal. Chem. 1979 51 1613 September 1982 MULTIPLE TRACE ELEMENT ANALYSIS OF FOODSTUFFS 30. 31. 32. 33. 34. 35. Evans W. H. Jackson F. J. and Dellar D. Analyst 1979 104 16. Evans W. H. Analyst 1978 103 452. Baetz R. A. and Kenner C. T. J . Agric. Food Chem. 1975 23 41. Baker A. S. and Smith R. L. J . Agric. Food Chem. 1974 22 103. Jackson F. J. Read J. I. and Lucas B. E. Analyst 1980 105 359. Kamada T. Talanta 1976 23 835. 993 Received November 17th 198 1 Accepted March 1st. 198

ISSN:0003-2654    

DOI:10.1039/AN9820700977

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

994 Analyst, September, 1982, Vol. 107, pp. 994-999 Direct Determination of Lead in Used Engine Oils by Atomic-absorption Spectrophotometry J. M. Palmer and M. W. Rush Engine Laboratory, Associated Octel Co. Ltd., Watling Street, Bletchley, Milton Keynes, MK1 1 EZ An improved method has been developed for measuring lead in used lubri- cating oils. The procedure utilises a mixture of acid and a liquid anion exchanger (Aliquat 336) to dissolve the lead particles in 4-methylpentan-2-one and measurement is made by atomic-absorption spectrophotometry. The method is quantitative in the range 0.1-2.5% m/m and is independent of effects caused by variations in particle size, lead species or oil additives. The method is rapid and has good repeatability. atomic-absorption spectrophotometry Keywords Lead determination ; used lubricating oils ; liquid anion exchangers ; Used lubricating oils are frequently analysed for their total lead content to assist in engineer- ing studies conducted on fuels and lubricants. Oil analysis is also used to monitor the levels of wear metals, including lead, to indicate the mechanical condition of moving parts of engines before wear is visually evident .1*2 Similarly, metal contents are measured before waste oils are burnt, with or without heat recovery, as a method of disposal3** and also prior to recycling (re-refining) .5 9 6 The lead found in used engine oils occurs mainly as decomposition products resulting from combustion of the fuel. The lead is in the form of finely divided particles essentially suspended in the oil.It occurs as complex mixtures' of simple inorganic compounds such as lead(I1) bromide, chloride, oxide and sulphate with possibly some lead( 11) orthophosphate. The lead contents of used oils occur in the range 0.1-2.57, m/m, depending on factors such as the length of time in service, the age and condition of the engine and the lead content of the fuel. The method for the determination of lead used at present in our laboratory is based on a classical procedure developed by the Institute of Petroleum.s The oil sample is completely destroyed by wet ashing with mineral acids to produce a precipitate of lead sulphate, which is dissolved in sodium hydroxide solution and the lead is determined by either polarography or atomic-absorption spectrophotometry (AAS) . This method is very time consuming and prone to losses of lead due to spattering during the ashing stage.A more rapid AAS procedure that avoids the wet-chemical treatment is widely used for wear-metal monitoring.1*2 The oil samples are diluted with 4-methylpentan-2-one and the metals are determined directly by AAS. Standards for calibration are prepared from organo- metallic compounds dissolved in the same matrix. This method, however, is known to be dependent on particle size and not quantitative for some elements. Saba and Eisentraut99lo modified the method to overcome this problem for measuring titanium and molybdenum. Initially they reacted the oil samples with a mixture of acids such as hydrofluoric acid plus either hydrochloric or nitric acid to dissolve the metal particles.This they achieved quickly and simply by shaking the mixture for 2 min before dilution with 4-methylpentan-2-one and measurement. Standard solutions were prepared from finely divided powders of the metals dissolved in the acid - oil mixtures. Application of a similar technique for the determination of lead was thought likely to be simple, rapid and quantitative, and such a method is described in this paper. Experimental Scope of Method calcium, zinc, phosphorus, etc. Outline of Method The method is satisfactory for used oils having metallic additives containing barium, The concentration range is 0.1-2.5% m/m. The sample containing a uniform dispersion of finely divided inorganic lead particles isPALMER AND RUSH 995 reacted with hydrochloric acid and Aliquat 336 by shaking, to dissolve the lead, in 4-methyl- pentan-2-one.The lead is measured in the clear supernatant liquid by atomic-absorption spectrophotometry. The analyser is calibrated with a standard solution of lead(I1) chloride in a similar matrix. Reagents The mixture is centrifuged to remove any insoluble precipitate. Hydrochloric acid, 35.4% mlm. 4-Methylpentan-2-one. AAS grade. Aliquat 336 (trioctylmethylammonium chloride), 9 yo ml V solution in 4-methylpentan-2-one. Lubricating oil. Lead-free base oil. Standard lead solution. at 105 "C for 3 h and cool in a desiccator. add 9Og of Aliquat 336. dissolve the lead chloride. with 4-met h ylpen tan-2-one. AnalaR grade. Aliquat 336 is obtainable from Eastman-Kodak Ltd., Kirkby, Lancashire.Dry some laboratory reagent-grade lead(I1) chloride in an oven Weigh accurately 1.3426 g into a beaker and Dilute to about 200ml with 4-methylpentan-2-one and swirl to Transfer quantitatively into a calibrated flask and dilute to 1 1 1.00 ml of solution = 1.00 mg of lead. The solution should be placed in a tightly stoppered borosilicate glass bottle and stored at Immediately any precipitation or change in concentration of the standard These standard solutions are stable under the above storage con- room temperature. is suspected, discard it. ditions for at least 6 months. Apparatus Stirrer (homogeniser) . Centrifuge. Mechanical shaker. Atomic-absorption spectrophotometer. Analytical balance. Calibrated $asks, grade B. Pipettes, grade B. Calibration (Full Range) For the preparation of reference solutions, pipette into successive 100-ml calibrated flasks 0.0, 5.0, 10.0 and 15.0 ml of standard lead solution.Add 15, 10, 5 and 0 ml of Aliquat 336 solution, respectively, to the successive flasks, then add 4.0 g of base oil; mix and dilute to 100 ml with 4-methylpentan-2-one. This gives reference solutions containing 0, 5.0, 10.0 and 15.0 mg of lead. Over-all Calibration Procedure To check the over-all response and to establish a calibration graph (if needed), set up the analyser according to the manufacturer's general instructions. Ignite the burner and allow it to stabilise whilst aspirating 4-methylpentan-2-one for a few minutes. During this time the acetylene flow-rate should be adjusted (reduced) to minimise luminosity in the flame without causing flame lift-off from the burner head.Then aspirate the top standard and make adjustments to the nebuliser, to the burner height, burner alignment and impact bead (if fitted) to give maximum absorbance (about 0.150). Aspirate the reference solutions in order of decreasing lead content and measure the absorbances at a wavelength of 261.4 nm, recording the values to 0.001 absorbance unit, Prepare a calibration graph by plotting net absorbance on the ordinate against mass of lead on the abscissa. Routine Procedure Turbo-agitator L60, Moritz Chemical Engineering. Air - acetylene flame. 100 ml and 1 1. 5, 10 and 15 ml. The line should pass through the origin. Measure the total mass of oil by using pre-weighed sample cans. Warm the oil in the can to 50-60 "C and stir vigorously with the turbo-agitator until all of the sediment is homogeneously suspended in the oil. Using a pipette, transfer sufficient oil in duplicate to contain about 10-15 mg of lead (Table I), into pre-weighed 100-ml calibrated mixing flasks and weigh to the nearest 0.001 g.1. 2. 3.996 Analyst, VoZ. I07 4. Add 0.25 ml of hydrochloric acid and 15 ml of Aliquat 336 to each flask and shake for 5min. Add base oil to adjust the total to 4.0g, add 4-methylpentan-2-one, mix and dilute to the calibration mark with the solvent. Prepare a standard containing 15.0 ml of lead solution, 4.0 g of base oil and 0.25 ml of hydrochloric acid in a 100-ml calibrated flask, mix and dilute to the mark with the solvent. Transfer portions of the samples and standard into centrifuge tubes and spin for 5 min at 3000 rev min-l.Set up the analyser and optimise the response with the standard solution (supernatant liquid) as specified in the calibration procedure. Re-set at zero with 4-methylpentan-2-one. “Spray” the samples (supernatant liquid) using identical settings and record the absorbances. PALMER AND RUSH : DIRECT DETERMINATION OF 5. 6. 7. 8. TABLE I RECOMMENDED SAMPLE SIZE Sample size A I \ Expected lead content, Approx. volume/ Yo mlm ml Mass/g 0.1-0.4 5 4 0.4-0.8 2.5 2 0.8-1.6 1.3 1 1.6-2.5 0.6 0.5 Calculation absorbance of sample x 1.5 absorbance of standard x sample mass (g) Lead content (yo m/m) = Analytical range: 0-15 mg of lead. Absorbance: 0-0.150. Wavelength : 261.4 nm. Results and Discussion Initially we applied the technique of the previous w0rkers99~~ using hydrochloric acid alone, hydrochloric acid - nitric acid or hydrochloric acid - perchloric acid mixtures with oils of known lead content.All of the test solutions gave poor and variable analyser responses and seemed to indicate that the lead was incompletely dissolved. The operating parameters of the analyser are given in Table 11. Other workersllJ2 had shown that it was possible to dissolve inorganic compounds such as lead( 11) chloride in 4-methylpentan-%one by using liquid anion exchangers, such as the quaternary ammonium compound Aliquat 336 (trioctylmethylammonium chloride). Solubility tests (10 min, mechanical shaker) with an Aliquat 336 - 4-methylpentan-2-one mixture showed that lead(I1) chloride and lead(I1) bromide were completely soluble, lead(I1) oxide and lead(I1) orthophosphate were slightly soluble and lead( 11) sulphate was virtually insoluble.The experiments were repeated but with the addition of a small volume (0.5 ml) of con- centrated hydrochloric acid to each test solution. The results showed that all of the lead compounds were soluble in this mixture and the analyser gave an identical response to lead irrespective of the compound present. The optimum amounts of hydrochloric acid and Aliquat 336 were determined in further tests. In separate experiments, the volume of hydrochloric acid (35.4% m/m) was varied from 0 to 0.5 ml and the Aliquat 336 concentration from 3 to 24 g 1-l. It was found that the optimum practical amounts were 0.25 ml of hydrochloric acid (2.5 ml 1-1 of con- centrated hydrochloric acid) and 13.5 g 1-l of Aliquat 336.Both amounts represented about a 20-fold excess of reagent over the maximum amount of lead expected to be present in sample aliquots. Variations in the concentration of hydrochloric acid of *50% and of Aliquat 336 of -25% to +7Oy0 were not deleterious. So far the experiments had been carried out with pure chemicals but without the base oilSeptember, 1982 LEAD IN USED ENGINE OILS BY AAS 997 TABLE I1 OPERATING PARAMETERS FOR ANALYSER Instrument . . .. . . . . Perkin-Elmer 306 Burner control box . . . . . . Perkin-Elmer, Model 303-0240 Wavelength . . .. .. .. 261.4nm Slit width . . .. .. . . 0.7 nm Grating . . .. .. . . Ultraviolet Lead lamp .... . . . . Juniper 7.5 mA, or Perkin-Elmer 10 mA Air- Cylinder regulator . . . . 50 p.s.i.g. Burner control . . .. . . 30 p.s.i.g. Flow-meter setting . . .. 6 Flow-rate . . .. ., . . 15 1 min-1 Cylinder regulator . . . . 9 p.s.i.g. Burner control . . . . . . 8 p.s.i.g. Flow-rate . . .. .. . . 4.2 lmin-l Acetylene- Flow-meter setting . . - . 3 present. Some of these tests were repeated using the optimised reaction mixture but with the addition of amounts of lead-free base oils. A series of tests was made by varying the amount of oil from 0 to 5 g per 100 ml of solution using seven commercially available and one prototype oil. Here it was necessary to centrifuge the solutions after shaking because of precipitate formation and to analyse the clear supernatant liquids.The analyser response for lead was then constant irrespective of the amount or type of oil present. The effect of variations in reaction time (shaking time) was studied in the range 0-14 min using authentic engine oils containing lead (previously assayed by a wet-chemical method). After reaction each test solution was observed to have thrown down a black, gelatinous precipitate, so measurements were again made on centrifuged solutions. The results showed that the lead particles had dissolved almost instantaneously during hand mixing of the reagents and before the prescribed shaking period. Even so, a shaking time of 5 min (mechanical shaker) was subsequently chosen for the test method. A calibration graph was constructed with results obtained from test solutions containing 0-25 mg of lead (PbCl,) per 100 ml of the diluted reaction - oil mixture.The compositions of the test solutions and the results are given in Table 111. A typical calibration graph is shown in Fig, 1. We found that the analyser response to lead was almost linear up to 15 mg of lead, but that significant curvature was evident at the higher concentrations. For practical reasons, therefore, a linear working range of 0-15 mg of lead (0-150 mg 1-1 of lead) was selected for the test method. The stability and shelf-life of a standard solution of lead(I1) chloride (1.0 mg ml-l of lead) containing Aliquat 336 (9% m/V) was assessed when stored (tightly stoppered) at room temperature. The concentration of this solution was determined periodically by AAS against a freshly prepared standard solution on each occasion.The results showed that the TABLE I11 CALIBRATION DATA Flask contents Aliquat 3361 No. Oil/g HCl/ml mg 1 2.5 0.25 1350 2 2.5 0.25 1350 3 2.5 0,25 4 2.5 0.25 6 2.5 0.25 6 2.5 0.25 7 2.5 0.25 8 2.5 0.25 9 2.5 0.25 1350 2 250 1800 2 250 4-Methylpentan-2-one/ Lead/ ml mg 0 5.0 5.0 10.0 Balance to 100 10.0 15.0 15.0 20.0 25.0 Relative absorbance 0.003 0.047 0.047 0.091 0.091 0,130 0.130 0.169 0.201998 PALMER AND RUSH : DIRECT DETERMINATION OF Analyst, VUZ. 107 Lead/mg Fig. 1. Calibration graph for lead in oil determination. Perkin-Elmer 306 instrument with air - acetylene flame. Wavelength, 261.4 nm;Cslit width, 0.7 nm; and sample volume, 100 ml. TABLE IV ANALYSIS OF TYPICAL USED OILS : COMPARISON OF METHODS Lead content, % mlm r \ A Sample No. Wet-chemical method 1 0.61, 0.58 2 0.48, 0.51 3 0.54, 0.54 4 0.50, 0.51 5 0.63, 0.62 6 0.55, 0.55 7 0.48, 0.49 8 0.32, 0.30 9 0.31, 0.31 10 0.48, 0.50 11 0.40, 0.39 12 0.61, 0.59 13 0.73, 0.77 14 0.74, 0.70 15 0.74, 0.70 16 0.77, 0.83 17 0.44, 0.42 18 0.60, 0.62 19 0.54, 0.55 20 0.49, 0.48 21 0.32, 0.38 22 1.50, 1.46 23 0.46, 0.45 24 0.50, 0.49 25 0.49, 0.49 26 1.03, 0.97 27 0.25, 0.26 28 0.95, 0.85 29 0.78,* 1.06* 30 0.24, 0.24 - F calculated .... F tabulated . . .. .. t calculated . . .. .. t tabulated . . .. .. Repeatability, % . . .. 0.07 - - - * Sample 29 not used in statistical analysis. AAS method 0.60, 0.60 0.55, 0.55 0.56, 0.56 0.50, 0.51 0.64, 0.64 0.57, 0.57 0.50, 0.49 0.36, 0.37 0.38, 0.38 0.54, 0.55 0.53, 0.52 0.71, 0.71 0.86, 0.88 0.81, 0.81 0.70, 0.69 0.82, 0.81 0.44, 0.43 0.60, 0.60 0.64, 0.62 0.54, 0.52 0.46, 0.46 1.49, 1.48 0.41, 0.41 0.45, 0.45 0.55, 0.54 1.04, 1.10 0.26, 0.30 0.94, 0.95 0.98,* 0.98* 0.26, 0.26 1.05 1.84 4.00 2.05 0.04September, 1982 LEAD IN USED ENGINE OILS BY AAS 999 solution was stable under these conditions, with no change in concentration, for at least 6 months.Because of the nature of the oil samples it was not possible to provide synthetic test samples with absolute concentration figures. Checks on the accuracy and precision of the AAS method were made, therefore, by comparing the results with those obtained by the original wet-chemical method. The AAS procedure used is described in detail under Experimental, and both methods were applied to the analysis of 30 typical used oils.The results obtained are given in Table IV. The data were evaluated statistically and the F-test variance ratio and Student’s t-test were ~ a l c u l a t e d . ~ ~ , ~ ~ The results showed that on average the AAS method had a positive bias, giving, therefore, slightly higher results compared with those obtained by the original procedure. The lower results with the original method were, how- ever, thought likely to be due to losses of lead during the ashing stage, as mentioned earlier. An over-all estimate of the re~eatabilityl~ of the AAS method gave a result of 0.04% m/m. The particle size range of lead compounds suspended in used oils and the possible effect on the determination of lead by the “wear-metal method”lP2 was not known.Final tests were made on five used oils, which were analysed by the wear-metal and the proposed AAS procedure. The results are given in Table V and showed that the former method gave lower results than those obtained with the proposed procedure, and on standing they were even lower. This was thought to show that the wear-metal method, as applied here to the deter- mination of lead, was dependent on particle size, whereas the AAS procedure was quantitative and, as expected, was independent of the particle size of suspended lead. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. TABLE V PARTICLE SIZE EFFECTS Lead content, % mlm r A -I Wear-metal method f A -I AAS method Sample Immediate measurement Standing time 1 h (immediate measurement) 1 0.96 0.55 1.10 2 0.39 0.18 0.43 3 0.63 0.46 0.86 4 0.76 0.68 0.87 5 0.21 0.14 0.22 References Klug, R., Society of Automotive Engineers, SAE Report 720372, New York, 1972.Jackson, D. R., Salarna, C., and Dunn, R., Can. Spectrosc., 1970, 15, Part 1, 17. Lederman, P. B., Society of Automotive Engineers, SAE Report 750387, Warrendale, PA, 1975. Emmerson, H. R., Hydrocarbon Process., 1975 (September), 145. Weinstein, N. J., Hydrocarbon Process., 1974 (December), 74. Cotton, F. O., Whisman, M. L., Goltzinger, J . W., and Reynolds, J. W., Hydrocarbon Process., 1977 Newby, W. E., and Dumont, L. F., Ind. Eng. Chem., 1953, 45, 1336. “Standards for Petroleum and Its Products,” Part 1, Volume 1, I P 120/48 (obsolete), Institute of Saba, C. S., and Eisentraut, K. J., Anal. Chem., 1977, 49, 454. Saba, C. S., and Eisentraut, K. J., Anal. Chem., 1979, 51, 1927. Russell, T. J., and Campbell, K., Report OP 77/1, Associated Octel Co. Ltd., Milton Keynes, 1977 “Annual Book of ASTM Standards, Part 25, Petroleum Products and Lubricants,” Method D3237-79 “Annual Book of ASTM Standards, Part 24, Petroleum Products and Lubricants,” Method D2891-73, “Operating Manual for the Compucorp 445 Statistician,” Computer Design Corp., Los Angeles, CA, “Annual Book of ASTM Standards, Part 25, Petroleum Products and Lubricants,” American (September), 131. Petroleum, London, 1948. (available from the present authors on request). American Society for Testing and Materials, Philadelphia, 1979. American Society for Testing and Materials, Philadelphia, 1979. 1972, pp. 32-70. Society for Testing and Materials, Philadelphia, 1979, p. 1059. Received February 16th, 1982 Accepted March 291k 1982

ISSN:0003-2654    

DOI:10.1039/AN9820700994

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

1000 Analyst, September, 1982, Vol. 107, pp. 1000-1005 Study of an Automatically Triggered Digital Integrator for Atomic Spectrometry of 15 Elements Using Discrete Nebulisation lsao Kojima and Chuzo lida" Laboratory of Analytical Chemistry, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya 466, Japan A signal integrator, with an automatic trigger and a voltage to frequency counter, was used in the atomic-absorption spectrometry of silver, calcium, cadmium, cobalt, copper, iron, magnesium , manganese, nickel, lead, strontium and zinc and the atomic-emission spectrometry of sodium and potassium with an air - acetylene flame and the atomic-absorption spectrometry of aluminium, calcium and magnesium with a dinitrogen oxide - acetylene flame using a discrete nebulisation technique.The effects of injection volumes, sample flow-rates and metal concentrations on the variables (peak height, integrated value, aspiration time and the ratio of the integrated value to the aspiration time) were investigated. Each of the integrated values for silver, copper, iron, magnesium, manganese and lead is proportional to the absolute amount of each metal irrespective of the metal concentration and injection volume. The calibration graphs for these six elements were straight lines passing through the origin. The relative standard deviations of the measure- ments were less than a%, even with an injection volume of 20 pl, and less than 1% with an injection of 100 pl. Keywords : Automatic integration ; atomic-absorption and -emission spectro- metry ; discrete nebulisation ; absolute amount method Flame atomic-absorption spectrometry has been studied in detail by use of discrete nebulisa- tion of a small volume of sample1 and applied successfully to the analyses of some elements in various samples.lS2 The sensitivity obtained by the measurement of peak height was, however, too low to enable the determination of the concentration of trace metals.A digital integrator, automatically triggered by the electrical conductance of the nebulised sample solution, was recently proved to be applicable to the flame atomic-absorption spectro- metry of copper by a discrete nebulisation te~hnique.~ The copper contents in some NBS standard biological samples were determined a~curately.~ This paper deals with the fundamental study of the application of the automatically triggered digital integrator to flame atomic-absorption and flame atomic-emission spectro- metry by the discrete nebulisation of solutions of 15 elements with an air - acetylene and/or a dinitrogen oxide - acetylene flame.The effects of injection volume and sample flow-rate on the variables (peak height, integrated value, aspiration time and the ratio of the integrated value to the aspiration time) were investigated. Experimental Apparatus All of the apparatus and equipment, except for the home-made burner head with a slit 6 cm long and 0.5 mm wide for the dinitrogen oxide - acetylene flame, were the same as those reported in a previous paper.4 A small PTFE funnel with a platinum electrode is coupled directly to the nebuliser needle (platinum pipe), which is the counter electrode.When the channel between the platinum electrodes is filled with the injected sample solution, the electrical conductivity between both the electrodes acts as an automatic trigger and the signal is integrated for a pre-set time. Details of the injection system of sample solution and the related materials are also given in references 3 and 4. Operating Conditions The absorbances of aluminium, calcium and magnesium were measured under the same operating conditions with a dinitrogen oxide - acetylene flame : the flow-rate of the acetylene * To whom correspondence should be addressed.KOJIMA AND IIDA 1001 was 5.5 1 min-l (0.5 kg cm-2) ; the flow-rate of the dinitrogen oxide was 4.4 1 min-l (1.5 kg cm-2) for the nebuliser and 2.0 1 min-1 (1.5 kg cm-2) auxiliary; burner height, position 2.5; and sample flow-rate, 5.4 ml min-l.The wavelengths were: silver, 328.1 ; aluminium, 309.3; calcium, 422.7 ; cadmium, 228.8; cobalt, 240.7 ; copper, 324.8; iron, 248.3 ; potassium, 766.5; magnesium, 285.2; manganese, 279.5; sodium, 589.0; nickel, 232.0; lead, 283.3; strontium, 460.7; and zinc, 213.9 nm. Contrary to a previous paper,l in this study the response for emission is faster than that for absorption, i.e., the condenser capacitance of the emission was smaller than that of the absorption. The integration time (5 s) and the other operating conditions are the same as those given in the previous paper.4 Reagents The stock solutions of each element (2000 p.p.m.in 0.5 M nitric or hydrochloric acid) were prepared by dissolving analyt ical-reagent grade compounds in nitric or hydrochloric acid and diluting with nitric or hydrochloric acid and doubly distilled water to give a mass of The working standard solutions of each element were prepared by diluting the stock solutions to appropriate concentrations by mass in 10-ml polypropylene bottles with doubly distilled water and nitric acid using micropipettes. All these solutions contain 0.1 M nitric acid. 100 g. Results and Discussion Effect of Injection Volume on Peak Height of Signal As with copper,l the peak height for the atomic-absorption signal increases with an increase in the injection volume, up to a volume of about 100 pl, under a given sample flow-rate and then it remains constant, the height being the same as that obtained by the continuous nebulisation.The results are exemplified in Fig. 1, together with those of the other variables. All the data are summarised in Table I. The minimum injection volume giving a constant peak height and the peak height itself increases with an increase in the sample flow-rate. The minimum volume is about 50 pl for the emission (potassium and sodium in Table I), as the response of the electrical circuitry is faster than that for absorption. Effect of Injection Volume on the Integrated Value the automatic trigger and digital counter. The area under the spike-like signal was integrated satisfactorily by the combined use of The procedure has been described previ~usly.~ 1 o4 20 101 0' 5 I n' 10 100 1 O( Inject ion volume, !'I 1 o4 20 10 5 1 o3 102 I I 10 100 1000 Fig.1. Effect of injection volume on four variables. Sample flow-rate, 4.5 ml min-'. Con- centration: (a) manganese, 2.3 pg ml-1; (b) sodium, 5 pg ml-l; and (c) magnesium, 0.6 pg d-l. (a) and (b), air - acetylene flame; and (c), dinitrogen oxide - acetylene flame.1002 KOJIMA AND IIDA: AUTOMATICALLY TRIGGERED DIGITAL Analyst, VoZ. 107 TABLE I RELATIONSHIP BETWEEN INJECTION VOLUME AND FOUR VARIABLES Injection volume/Wl* 2 Cd co cu Pe Mg Mn Ni Pb Sr Zn K Na A1 Ca Mg Calibration Element graph . . . . . . Linear . . . . . . Curve . . . . . . curve . . . . . . Curve . . . . . . Linear . . . . . . Linear . . . . . . Linear . . . . . . Linear . . . . . . Curve . . . . . . Linear . . . . . . Curve .. . . . . Curve . . . . . . Curve . . . . . . Linear . . . . . . Curve . . . . . . Curve . . . . . . Linear Aspiration time 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 Count 30 YY 60 40 40 40 40 60 50 150 40 N U NLS NLS NLS N U Count1 aspiration 40 *kt 60 40 40 40 50 60 50 150 40 NLS NLS NLS NLS NLS r Peak ' time height Extrapolation 100 A+ 100 NL$ 3 100 100 100 A t . 100 A t 80 A t 100 A t 100 NLS kL* 100 80 80 N L ~ 50 NLS 50 NLt ;2 100 120 120 N L ~ Calibration range, p.p.m. 0-0.5 0-1.2 0-1.0 0-2.0 0-0.5 0-2.0 0-0.03 0-0.4 0-2.0 0-4.0 0-0.5 0-0.8 0-55 0-47 0-1 801 I 0-1.011 0.0.311 At larger than the figures tabulated (in microlitres) the relationship between injection volume and variables is linear. t A, The absolute amount method is possible. $ NL, Relationship between injection volume and variables is not linear.5 Burner parallel. fl B.umer at right angles. 11 Dinitrogen oxide - acetylene flame. With a constant concentration, the integrated values (counts) for silver, cadmium, cobalt, copper, iron, magnesium, manganese, nickel, lead and zinc increase linearly with an increase in the injection volume above 50 pl. Fig. l ( a ) shows the result for manganese as a typical example. This linearity begins only at a volume of larger than 150 p1 for strontium (column 4, Table I). Non-linear increment is observed for calcium, potassium and sodium with an air - acetylene flame and for aluminium, calcium and magnesium with a dinitrogen oxide - acetylene flame. A linear relationship is observed between the integrated value and the concentration of silver, copper, iron, magnesium, manganese and lead with a constant injection volume (columns 2 and 4, Table I). A non-linear relationship is observed for the other elements (Table I and Fig. 2).These phenomena are the same as those observed in the continuous nebulisation method so far. It is clear from Fig. 2 that the slopes of the calibration graphs increase with an increase in the injection volume. Thus the integrated values for silver, copper, iron, magnesium, manganese and lead seem to be proportional to the absolute amount of the analyte. Figs. l ( b ) and l ( c ) show the results for sodium and magnesium. I 1000 I P 0 2 4 6 8 Cobalt concentration yg ml-' Fig. 2. Calibration graphs for cobalt for different injection volumes. Air - acetylene flame.September, 1982 INTEGRATOR FOR ATOMIC SPECTROMETRY OF 15 ELEMENTS 1003 Effect of Injection Volume on Aspiration Time The aspiration time increases linearly with an increase in the injection volume of more than 30 p1 under a given sample flow-rate (Table I), although the precise feature is not very clear in Fig.1. The aspiration time is defined as the time required for the injected sample solution to pass through the channel between both electrodes. The aspiration time depends mainly upon the viscosity of the sample solution. Effect of Injection Volume on the Ratio of the Integrated Value to the Aspiration Time The ratio of the integrated value to the aspiration time is constant irrespective of the injection volume when it is more than about 40 p1 for silver, copper, iron, magnesium and zinc, more than about 50pl for cadmium, manganese and lead, more than about 60pl for cobalt and nickel and more than about 150 pl for strontium (column 5, Table I).However, such a relationship does not hold for calcium, potassium and sodium with an air - acetylene flame and for aluminium, calcium and magnesium with a dinitrogen oxide - acetylene flame because of the non-linearity between the integrated value and the injection volume. Typical examples are given in Fig. 1 ( b ) and (c). This ratio is dependent only upon the concentration of analyte in the sample solution, as in the peak-height method. Thus, the calibration graph of this ratio against the concentration of analyte is either linear or non-linear, depending upon the elements.Table I indicates that a straight line calibration graph passing through the origin of the coordinates is obtained with an appropriate injection volume of more than &bout 50p1 for silver, cadmium, copper, iron, magnesium, manganese, lead and zinc, more than about 60 p1 for cobalt and nickel, and more than about 150 pl for strontium. A typical example is shown in Fig. 3. A simple calibration graph was not obtained, even by use of a large sample volume, for the other elements (aluminium, calcium, potassium and sodium), because of the non-linearlity between the integrated value and the injection volume. Sensi- tive calibration graphs passing through the origin are also obtained with given injection volumes of less than the volumes mentioned above, irrespective of a linear or non-linear relationship between the integrated value and the aspiration time and the injection volume.000 500 1 ? I rn - - L ._ Q rn m V I I I 0 0.2 0.4 Manganese concentrationiltg ml- ' Fig. 3. Calibration graphs for manga- nese for different injection volumes. Air - acetylene flame. Reproducibility The over-all averages of the relative standard deviations of the variables obtained with 20-, 50- and 100-pl sample solutions of an appropriate concentration of the 15 elements are given in Table 11. The relative standard deviation of each variable obtained with 50- and 100-pl sample solutions is very small and superior to that obtained by the continuous nebulisation method. The relative standard deviation of each variable is rather small even with an injection volume of 20 pl.Thus, the reproducibility of this method is satisfactory. The use of an integrated value or the ratio of the integrated value to the aspiration time instead of using the peak height is especially useful for the determination of the given elements. This is discussed further below.1004 KOJIMA AND IIDA : AUTOMATICALLY TRIGGERED DIGITAL Analyst, VoZ. I07 TABLE I1 REPRODUCIBILITY (%) OF FOUR VARIABLES (n = 10) Injection volume/pl r A \ Variable 20 50 100 Peak height . . . . . . 2-3 1-2 0.5-1 Count . . . . .. . * -2 1-2 0.5-1.5 Countlaspiration time . . .. 2-5 1-1.5 0.5-1 Aspiration time . . .. . . 1-3 1-2 -1 Effect of the Sample Flow-rate on Variables The effect of sample flow-rate was investigated with an injection volume of 100 pl, except for strontium (150 pl) and sodium (50 pl), under the constant flame conditions. A typical example is shown in Fig.4. The aspiration time decreases with increase in the sample flow-rate. The integrated values obtained with an air - acetylene flame are the highest and essentially constant at the given range of sample flow-rates, between 4.4 and 5.5 ml min-1 for all the elements except calcium, strontium and cadmium and between 3.6 and 4.4 ml min-1 for cadmium. Such a result is expected because the same absolute amount of analyte is nebulised. Thus, the ratio of the integrated value to the aspiration time increases with the sample flow-rate. However, the integrated value decreases with increase in the sample flow-rate for calcium and strontium although the reason remains unsolved.m v) S 3 0 .I- . 1500 . .- 4- A _ - I I 5 6 a m 0 4 --- Sample flow-rateimt min-’ Fig. 4. Effect of sample flow-rate on four vari- Concentration of manganese, 2.06 pg ml-l. ables. Injection volume, 100 p1. Air - acetylene flame. Application of the Absolute Amount Method As linear relationships are obtained between the integrated value and the concentration of the analyte at a constant injection volume, and also between the integrated value and the injection volume at the constant concentration, the integrated value is proportional to the product of the concentration of the analyte and the injection volume, i.e., the absolute amount of analyte in the sample solution. Thus, the absolute amount method is applicable, as it is with ~ o p p e r .~ The results for the six elements obtained with an air - acetylene flame and for the two elements with a dinitrogen oxide - acetylene flame are given in Table 111. The concentrations of copper, iron and lead, which are too low to be determined on the straight line calibration graphs in the normal concentration ranges, can be determined on these calibration graphs by the injection of a large volume of the sample solution. The results obtained by this absolute amount method are summarised in Table IV for nine elements. The results for copper, iron, magnesium, manganese and lead with differentSeptember, 1982 INTEGRATOR FOR ATOMIC SPECTROMETRY OF 15 ELEMENTS TABLE I11 COUNTS AT CONSTANT ABSOLUTE AMOUNTS OF METAL Injection Cu volume/yl (156 ng)* 2 000 678 1000 672 600 672 300 675 200 678 150 676 100 675 75 669 60 665 50 659 Mean .... .. 671.9 Standard'deviatidn . . . . .. 6.082 Relative standard deviation, % . . 0.90 * Air - acetylene flame. t Dinitrogen oxide - acetylene flame. Fe (200 ng) * 530 523 - - 530 526 - - - 514 524.6 6.618 1.26 Pb (300ng)* I 473 470 468 473 470 466 467 482 456 462 468.7 6.945 1.48 Ni :300 ng)* 885 852 822 774 725 703 671 683 677 671 746.3 81.23 10.9 Ca Na (200 ng)* (200 ng)* 731 582 731 587 - - - - 682 613 632 636 - - - - - - 629 671 681.0 617.8 50.26 36.79 7.38 5.96 Mg Ca (50 n d t (270 ng)t 262 402 284 441 287 473 300 - 312 490 332 516 333 525 349 540 344 538 311.4 490.6 30.17 49.48 9.69 10.1 - - 1005 injection volumes are in good agreement with results from standard solutions containing known amounts of the metal ions. However, Tables I, I11 and IV indicate that calcium, cadmium, cobalt, nickel, strontium, zinc, potassium and sodium with an air - acetylene flame and aluminium, calcium and magnesium with a dinitrogen oxide - acetylene flame cannot be determined by the absolute amount method.Thus, this method can be applied to the determination of very low concentrations of silver, copper, iron, magnesium, man- ganese and lead in various samples by use of a large volume for injection. Application of this method to standard rocks, glasses and serum will be described elsewhere. TABLE IV DETERMINATION OF LOWER LEVELS OF METALS WITH LARGE SAMPLE VOLUMES Element c u . . Fe . . Mg . . Mn . . Pb . . Ca . . Cd . . Ni . . c o . . Concentration, p.p.m. 0.0155 0.053 0.103 0.021 0.053 0.052 0.155 0.105 0.213 0.159 0.225 Injection volumelpl 2 000 1000 2 000 1000 2 000 1000 1000 2 000 1000 2 000 1000 2 000 1000 1000 500 2 000 1000 1000 Measured, p.p.m. 0.0157 0.0149 0.050 0.051 0.102 0.101 0.022 0.056 0.054 0.053 0.151 0.113 0.107 0.30 0.25 0.202 0.188 0.36 Standard solutions, p.p.m. 0-0.8 0-4.0 0-0.25 0-2.0 0-6 0-4 0-2.5 0-6 0-8 References 1. 2. 3. 4. Uchida, T., Kojima, I., and Iida, C., Anal. Chim. A d a , 1980, 116, 205, and papers cited therein. Berndt, H., and Slavin, W., At. Absorpt. Newsl., 1978, 17, 109. Goto, K., and Uchida, T., Rev. Sci. Instrum., 1980, 51, 49. Uchida, T., Kojima, I., and Iida, C., Analyst, 1981, 106, 206. Received February 2nd, 1982 Accepted April 5th, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701000

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

1006 Analyst, September, 1982, Vol. 107, pp. 1006-1013 Low-temperature Oxygen - Fluorine Radiofrequency Ashing of Biological Materials in Poly(tetrafluoroethy1ene) Dishes Prior to the Determination of Tin, Iron, Lead and Chromium by Atom ic-a bsorption Spectroscopy E. V. Williams* Analytical Chemistry, British Steel Corporation, Tinplate Research Laboratories, Carmarthen Road, Swansea, West Glamorgan, SA1 1HF Low-temperature radiofrequency ashing utilising plasma-excited oxygen alone requires long ashing times and for this reason has not been generally accepted. In this paper it is shown how ashing time can be reduced considerably by introducing fluorine into the oxygen plasma and a novel and efficient method of achieving this is described. Food products and National Bureau of Standards Standard Reference Materials, both of known trace metal content, are prepared by the method described.Analysis of the ash by atomic- absorption spectroscopy shows good recovery of added tin and agreement with certified values for iron, lead and chromium. Keywords : Low-temperature plasma ashing ; oxygen - jluorine ; PTFE ; tin, iron, chromium and lead determination ; atomic-absorption spectroscopy When oxygen is passed into a vacuum chamber held at less than 1.0 mmHg by a two-stage vacuum pump and subjected to a radiofrequency field (RF) oscillating at 13.56 MHz, an excited species of oxygen is formed consisting of atoms, ions and electrons with limited lifetimes of the order of 1 s or 1ess.l These vibrational and atomic states of oxygen were investigated by Gleit and Holland2 who found that oxidation of organic matter was possible at temperatures of less than 200 "C within the influence of an oxygen plasma. The tech- nique is often referred to as low-temperature ashing (LTA).These plasma chemistry oxidation reactions have been utilised in the preparation of food samples prior to the analysis of the ash for tin, iron, lead and chromium by atomic-absorption spectroscopy. The primary objective of these studies was to develop a method to replace existing methods of sample preparation using wet or dry high-temperature oxidation^,^ with more refined procedures capable of retaining the very low levels (less than 0.1 pg ml-l) of some metals found in some foods. The procedure is intended to support and become a potential reference method for more rapid systems of analysis previously publi~hed.~ These alternative technologies are, for reasons of economy, safety and energy conserva- tion, worthy of much closer examination.They have not been widely accepted mainly because of difficulties in completely oxidising, within acceptable time limits, some organic materials when oxygen is used alone. Some manufacturers of plasma-ashing equipment recognised these limitations and added agitators or heaters to assist oxidation, but with only limited success. Few manufacturers recognised that the route to successful and efficient ashing for some metals was by means of a plasma chemistry reaction using atomic fluorine. Investigations a t these laboratories were prompted by the fact that ashing was always significantly more efficient and more rapid when fluorine in the form of hydrofluoric acid (2% m/V) was added to the food product before drying.The same effect was found if poly(tetrafluoroethy1ene) was placed in an adjacent aluminium or silica crucible, during the ashing procedure. After 10-15 min in the oxygen plasma, a greenish glow indicating emission due to the presence of fluorine became evident. In the absence of fluorine, the plasma dis- charge was pink for oxygen alone or blue owing to carbon dioxide being formed as a by- product of sample oxidation. * Present address : British Steel Corporation, Welsh Laboratory, BSC Tinplate Research, Port Talbot, West Glamorgan, SA13 2NG.WILLIAMS 1007 These techniques were acceptable for some elements, but for others, unacceptably high blank values arose.When ashing was carried out in silica crucibles, it was found that volatile silicon tetrafluoride (SiF,) was formed. Impurities such as lead and iron were left on the surface and these were dissolved during acid or alkali dissolution of the ash. These contaminated the sample and gave grossly inflated values for lead and iron. The silica crucibles lost weight during exposure to the oxygen - fluorine plasma and the inside walls of the silicon - quartz reactor chamber became slightly etched. Some manufacturers use aluminium reactor chambers and one of the reasons for selecting aluminium is given below. When aluminium dishes were used for food samples containing chromium, it was found that a protective film of aluminium oxyfluoride, as measured by X-ray photoelectron spectroscopy, was formed on the surface, but during subsequent acid dissolution of the ash, chromium(V1) reacted with the aluminium surface to form a passive film containing chromium.These preliminary experiments were instrumental in directing studies towards using poly(tetrafluoroethy1ene) (PTFE) crucibles or dishes. The method described is based on the accelerating effect of atomic fluorine on oxidation by atomic oxygen. The novelty lies in the use of PTFE as both the source of fluorine and the sample container. Oxygen - fluorine plasma chemistry is not new; tetrafluoromethane (CF,, Freon 14) is recommended by the Branson - International Plasma Corporation5 and they justifiably claim that CF, will also accelerate the rate of oxidation. More recently Carter and Yeoman6 have given clear indications of the advantages of CF, plasma chemistry reactions for the determination of cadmium in blood. In the procedure to be described, there are many advantages in using PTFE dishes as the source of fluorine and there are equally good reasons why the technique is to be preferred to conventional classical wet or dry oxidations for some elements.Plasma Chemistry Reactions A plasma is often referred to as the fourth state of matter, but it will only exist so long as energy is sustained. A chemical reaction inside a plasma reactor will immediately stop when power is switched off; there is no thermal lag. Under the conditions described and inside the plasma reactor, two important reactions occur: firstly, atomic oxygen (O*) and other vibrational states of oxygen are formed by capacitative or inductive coupling of radiofrequency energy to oxygen ; and secondly, atomic oxygen attacks the structure of PTFE7 to produce fragmented CF,, which further reacts with atomic oxygen to produce atomic fluorine (F*) as follows: 20" + CF4+ C + 4F --+ :CF, + 2F +CO, + 4F* The accelerated oxidation rate can be attributed to the smaller diameter of the fluorine atom and hence its greater penetrating power.Also, the closer proximity of the source of atomic fluorine, which has a short lifetime, to the material to be ashed enables high concentra- tions of atomic fluorine to accumulate exactly where they are required. The factors which affect the oxidation rate in an oxygen-fluorine plasma chemistry reactor are as follows.(i) Exposed surface area: the larger the surface area and the smaller the particle size, the faster is the ashing rate. (ii) Sample load: ashing time is generally proportional to sample mass. (iii) Temperature : high temperatures increase oxidation rate but may promote volatilisation of certain elements. High radiofrequency power input increases temperature but only an optimum radiofrequency power level or perfectly tvned matching of parameters will produce the maximum useful population of activated atomic oxygen - fluorine species for the task. (iv) Chemical structure and physical nature: silicone rubber is considerably more resistant to attack than PTFE because of its structure and silicone content.It is for this reason that equipment manufacturers use silicone rubber as seals and gasket materials on radiofrequency reactors. ( v ) Concentration of activated species : input power to the discharge and oxygen pressure control the concentration of activated species. The oxygen pressure controls the amount of power that can be accepted in the efficient production of activated states. (vi) Decomposition products : CO, CO, produced during oxidation retard the oxidation rate, but these diminish because they are exhausted from the system by the vacuum pump.1008 Analyst, VoZ. I07 It is possible that other gases which dissociate and ionise in the plasma will also influence the rate of oxidation. (i) Poly(tetrafluor0- ethylene), following exposure to atomic oxygen, develops unique hydrophobic properties enabling the analyst to manipulate the dissolved ash in ways not possible by other means, such as quantitative transfer of the dissolved ash without washing because the PTFE surface is not wetted.(ii) The crucible or dish is not affected by most acids and may be heated to 250 "C. (iii) The material is self cleaning during reactor exposure and the loss in weight is small enough to allow many hundreds of determinations to be performed before the dish/ crucible becomes unserviceable. (iv) Tetrafluoromethane (CF,) is expensive and does not perform a dual function. (v) The impurity levels found in PTFE are very low and for most applications beyond the detection limit of the method of analysis used. It seems therefore, that even if CF, is used, a resistive material in which to carry out the ashing procedure would be required and this material should be of high purity and immune from reaction with the metals present during dissolution.The advantages of plasma ashing over wet or dry oxidation procedures are as follows: (i) the economic aspects are today of major importance in terms of materials, chemicals, energy and applied man hours; (ii) automatically tuned reactors require little attention and fitted with suitable fail-safe devices can be operated overnight; (iii) the technique is cleaner (lower blanks, less contamination) and less manipulation is required; (in) most of the hazards associated with conventional procedures do not arise; (v) there is no reaction between the container and metals in solution; and (vi) it is possible to monitor the progress of a plasma chemistry reaction by measuring emission intensities indicating the presence of a particular element, e.g., a nitrogen emission line a t 674 nm is used in silicon chip manufacture to indicate when silicon nitride has been removed in a CF, plasma, and photodiodes are used to measure the intensity of the emission line.' The disadvantages of using the procedure are relatively few: (i) it is not acceptable over as wide a range of elements (unless low-temperature traps are used) because of the formation of volatile fluorides of, for example, boron, phosphorus, sulphur, silicon, titanium, uranium and tungsten; (ii) the capital cost is high; (iii) there is a need to dry samples before treat- ment ; (iv) static electricity effects necessitate precautionary delays before retrieval of the sample from the reactor chamber; and (v) most oxygen supply lines already installed are metal and there may be a need to replace these with nylon to avoid contamination by metal oxides.WILLIAMS : OXYGEN - FLUORINE RF ASHING OF BIOLOGICAL Most of the factors given above are to a large extent interdependent. The advantages of using PTFE compared with CF, are as follows. Experimental Apparatus Reactors. Instruments were supplied by each of the following manufacturers : Tracerlab Ltd., Division of Electronics, Richmond, CA, USA, Model LTA 600 (UK Agents, Laboratory Impex Ltd., Lion Road, Twickenham) ; Branson - International Plasma Corporation (IPC), Hayward, CA, USA, Model IPC 4000/104B (UK Agents, Teledyne-Tac, Bath Road, Cran- ford, Middlesex) ; and Nonotech (Thin Films) Ltd., Sedgley Park Trading Estate, Prestwich, Manchester, Model P100.When in use each reactor must be equipped with a silicon or aluminium reactor chamber, silicone rubber gaskets and tube connectors and a vacuum system using a halocarbon oil. As a safety precaution, all units should be fitted with a trap filled with disodium tetra- borate crystals to neutralise fluorine emissions and should be vented to the external atmos- phere. The manufacturers will supply data on radiofrequency shielding and the current leakage levels permitted. Homogeniser. A Kenwood Chef using a liquidiser attachment was used to prepare all foods to a slurry. The original liquidiser blades were replaced by titanium blades made in the laboratory.Electronic balance. A Mettler, Model P1200, electronic balance was used for all weighing operations and was also used to measure (by mass) precise dilution volumes using the tare facility on the balance. The mass of ash remaining after oxidation was sufficiently small (0.02-0.06 g) to enable solutions to be prepared by simply adding fixed volumes of acid within the range 5-25 ml, without correction.September, 1982 MATERIALS IN PTFE DISHES FOR SN, FE, PB AND CR BY AAS 1009 Atomic-absorption spectrometer. A Varian-Techtron atomic-absorption spectrometer, Model A.A.5, was used for flame analysis (tin and iron) and a carbon rod attachment (CRA), Model 63, was used for electrothermal atomisation (lead and chromium). Poly(tetra~uoroethyZene) crucibles.These were machined from 15 mm thick PTFE sheet ("Fluon," Imperial Chemical Industries, Plastics Division, Welwyn Garden City). The usual precautions were observed in machining this material. * The external dimensions were as follows: diameter 60 mm; depth 10 mm; and wall 3 mm. After cleaning in 1 + 1 nitric acid, they were placed in an oxygen plasma to etch and clean the sufrace for 15 min at 100 W. Non-volatile impurities, which accumulated as a result of machining on the surface, were dissolved away by immersing the entire surface in a solution of hydrochloric acid containing hydrogen peroxide. Under the test conditions in a Branson - International Plasma Chemistry Reactor, mass losses of 0.06 g h-l at 100 W were typical, but depended on the type of reactor and to some extent on the position of the PTFE crucible inside the reactor.Reagents All reagents were of Aristar grade (BDH Chemicals Ltd., Poole, Dorset). Hydrochloric acid - hydrogen peroxide. A 40-ml aliquot of hydrochloric acid (sp. gr. 1.17) was diluted with water and 2ml of hydrogen peroxide (30% m/V) were added, giving a total volume of 100 ml. Nitric acid. Standard solutions. Multi-element solutions of chromium, iron and tin were prepared from single element stock solutions (1000 pg ml-l) by appropriate dilution in 20y0 m/V hydrochloric acid. Lead solutions were prepared in 20% m/V nitric acid from a separate stock solution (1 000 p g ml-1). High-purity salts or metals were used to prepare all stock solutions.Nitric acid (40 ml) was diluted with water to 100 ml. Procedure Weigh 1.0-5.0 -& 0.02 g of a representative homogenised sample in a PTFE crucible of known mass using an electronic balance. For the spiked samples, add the appropriate amount of standard solution to the PTFE crucibles and dry at 100 "C in an oven to remove any excess of acid. When dry, add the sample and dry at 120 "C for 2 h. Use the reactor conditions recommended by the manufacturer for the reactor model being used, e.g., for an IPC, Model 4000-104B: oxygen pressure 14-34 kN m-2; oxygen flow-rate 300 ml min-l; vacuum 0.5 mmHg; and radiofrequency power 100 W. Treat single 5.0-g samples for 4-8 h and increase the time proportionately for up to 3 x 5-g samples. Test the treated samples with 1-2 drops of distilled water to establish complete ashing, which will be evident by the absence of any black particles.If ashing is incomplete return to the reactor for further treatment. To avoid ash loss by static electricity, allow the charge to dissipate with the radiofrequency power switched off before sample withdrawal from the reactor. Dissolve the sample ash in an acid or alkali suitable for the analytical technique to be used. Tin and iron can be determined by atomic-absorption spectroscopy because they are usually present in sufficient amounts. A dinitrogen oxide - acetylene flame is used for tin and an air-acetylene flame for iron. The solvent used is hydrochloric acid- hydrogen peroxide at the concentration recommended under Reagents.The same solvent is used for chromium determinations, but the instrumental method relies on carbon rod analyses. The determination of lead requires nitric acid and carbon rod analysis but it is possible to use a flame technique where the concentration is sufficiently high. Dry the sample for 2 h at 120 "C. Results and Discussion The significant acceleration of oxidation due to the presence of fluorine is shown in Fig. 1 in which 5.0-g samples of boysenberries were dried at 120 "C to the same initial dry mass before low-temperature ashing treatment, The same effect was shown by a number of other1010 WILLIAMS: OXYGEN - FLUORINE RF ASHING OF BIOLOGICAL Analyst, Yd. 107 Fig. 1. Comparative ashing rates for 5.0 g of boysenberries in syrup using radiofrequency-excited oxygen with aluminium crucibles (B) and oxygen - fluorine with PTFE crucibles (A).Samples dried for 1 h at 120 "C to the same dry mass of 3.5 g . Values plotted have been corrected for crucible weight loss and each plotted point is a separate sample. Radiofrequency power, 100 W; and oxygen flow-rate, 300 ml min-'. products but for the sake of clarity results ha.ve not been included. I t should also be men- tioned that the mechanism by which fluorine increases the oxidation rate is dominated by an induction period (not shown). The period is probably controlled by the time it takes the atomic oxygen to attack and fragment the PTFE to form activated fluorine. Tin in the ash residue is apparently present as SnO (grey ash residue) and is acid soluble. It may also be present as a soluble fluoride, SnF, or SnF,.If SnO, is present, which is intractable, then it will be necessary to resort to lithium metaborate fusion in platinum a t 900 "C and dissolution of the cooled melt in ZOyo m/V hydrochloric acid. In practice this has not arisen and it is considered that if SnO, is present it is highly dispersed and would dissociate in a dinitrogen oxide - acetylene flame. Tables I and I1 show results obtained using the procedure described with PTFE crucibles. The results agree well with the certified values (Standard Reference Materials, National For lead and chromium determinations, sample mass to dilution volume ratios were Bureau of Standards, USA), and good recoveries of added tin are also obtained. TABLE I DETERMINATION OF TIN, LEAD, IRON AND CHROMIUM IN NATIONAL BUREAU OF STANDARDS REFERENCE MATERIALS (SRMs) The crucibles were PTFE, the sample mass was 1.0 g and all results are in micrograms per millilitre.Tin Iron Lead Chromium I Material Found A Certified' -d kound 'Certified ' c- SKRI 1575, pine needles + SRhl 1571, orchardleaves SRM 1577, bovine liver + 2.6 2.6 0.2 20 pg of tin .. 21.5, 20.3 ? (+20) 203, 215 200 & 10 10.8 10.8 f 0.5 ND* 2.48 0.34 + 20pgof tin .. _. 20 0.34-4.1 284 300 f 20 ND* 45 f 3 ( + 20) 200pgof tin . . . . 200 0.18 (+200) 280 268 & 8 0.37 0.34 & 0.08 ND* 0.088 ?C 0.012 * ND = not determined.September, 1982 MATERIALS IN PTFE DISHES FOR S N , FE, PB AND CR BY AAS 101 1 TABLE I1 COMPARATIVE DETERMINATIONS OF TIN, LEAD AND CHROMIUM IN CANNED BOYSENBERRIES (LABORATORY PREPARED SAMPLES) The crucibles were PTFE, the sample mass was 5.0 g and all results are in micrograms per millilitre.Values in parentheses are those obtained by a procedure published previ~usly.~ Reference No. Tin Iron Lead Chromium 2111 19 (19) 121 (128.8) 0.25 (0.19) 0.034 (0.035) 21/1 19 (19) 129 (128.8) - 0.034 (0.035) 2215 2.0 (<5) 31 (27) 1.2 (1.19) 0.038 - 2215 2.0 ((5) 31 (27) 1.2 (1.19) 0.040 - deliberately reduced so that a flame atomic-absorption result could be obtained. A gain in sensitivity thus obtained, however, resulted in a small volume of solution sufficient for the determination of only two elements. There are a number of atomic-absorption techniques that will overcome this disadvantage, e.g., a micro-sampling accessory. In Table 11, some results on boysenberries are given; these show good agreement with values obtained by a previously published procedure, which uses pressure decomposition in polystyrene containers.In Tables I11 and IV, results obtained in both aluminium and vitreosil (silica) dishes are given. Chromium values in Table I11 are excluded for the reasons already explained. TABLE 111 DETERMINATION OF TIN, IRON AND LEAD IN CANNED FOODS Aluminium dishes were used and the sample mass was 5.0 g. Reference Mass of Volume for Tin found,* Iron found,* Lead found,* No. Material ash/g analysislml p.p.m. p.p.m. p.p.m. 29.2 (27.6) 1.18 (1.18) 26.8 (27.6) 1.04 (1.18) 26.0 (27.6) 1.05 (1.18) _ _ 0.15 (0.18) 22 Boysenberries 0.01 25 5 (5) 22 Boysenberries 0.02 25 5 (2) 22 Boysenberries 0.06 25 5 (2) 24 Strawberries 0.04 10 182 (180) 7.2 (5.9) _ _ 9 Pulped tomato 0.02 10 28 (34) * Values in parentheses were obtained by a procedure reported in reference 4.TABLE IV COMPARISON OF PLASMA ASH AND PRESSURE DECOMPOSITION FOR THE DETERMINATION OF TIN USING VITREOSIL (SILICA) DISHES Material Plums . . . . .. Plums + 200 p g of tin Carrots . . . . .. Carrots + 200 pg of tin Grapefruit . . .. Grapefruit . . .. Grapefruit . . .. Full cream evaporated milk Tin found/pg Plasma ash Pressure decomposition I h \ . . .. 5 5 .. .. 205 205 . . .. 9 10 .. . . 200 200 .. .. 65 73 .. .. 191 196 . . .. 43 46 .. .. 7 7 In Table IV plums and carrots are duplicate samples and show the total tin present before Table V gives an indication of the range of food products and other materials ashed by addition.low-temperature ashing. Recoveries are based on a procedure described in reference 4.1012 WILLIAMS : OXYGEN - FLUORINE RF ASHING OF BIOLOGICAL Analyst, VoZ. I07 TABLE V LOW-TEMPERATURE ASHING OF A RANGE OF ORGANIC MATERIALS OF PLANT ORIGIN Sample Initial mass/g Pine needles* . . . . 1.9836 Plant tissue* . . . . 5.0544 Boysenberries . . . . 5.00 Raw sugar* . . . . 2.001 Orange juice* . . . . 17.4447 Tomato . . .. . . 5.0000 Strawberries . . . . 5.00 Boysenberries . . . . 5.00 Boysenberries? . . . . 5.00 Boysenberries . . . . 5.00 Residue/g Time/h 0.071 9 17 0.6996 193 0.01 9 0.002 2 16 0.221 8 30 0.04 7 0.04 7 0.06 4 1.60 4 0.02 4 Radiofrequency power/W 300 150 100 263 300 100 100 100 100 100 Plasma type 0, 0, 0, 0 2 0,- F 0, - F 0,- F 0, - F 0, 0, - F Sample ashed, yo 96.3 86.1 99.8 99.8 99.8 99.6 99.2 98.8 68 99.6 * Data obtained and published by permission oi LFE Corporation, Waltham, MA, USA, using one of their range of reactors.t Incomplete ashing. Justification for the validity of this technique can only be applied to the elements so far examined, i.e., tin, iron, lead and chromium. Some elements may form volatile fluorides under the test conditions, e.g., boron, uranium, titanium, molybdenum, tungsten, sulphur and silicon, but experimental work to establish this needs to be undertaken. There is also some evidence in the literature of the loss of certain elements in an oxygen plasma, e.g., arsenic, selenium, gold and ~ i l v e r . ~ The mechanism for such losses is not clear.There is in fact some advantage to be gained in the losses of some volatile fluorides; silicon appears to be a troublesome element in reacting with or retaining some of the elements of interest in SRMs. It is considered that this is one of the reasons why good recoveries are obtained for lead and iron in pine needles and bovine liver, SRMs 1575 and 1577, respec- tively. Most of the elements found in biological materials are oxidised to their higher oxidation states by atomic oxygen and further react with atomic fluorine to form oxyfluorides. It has been established by chemical tests that chromium is in the hexavalent form. Aluminium metal reacts to form aluminium oxyfluoride as established by X-ray photoelectron spectro- scopy and silica volatilises as silicon tetrafluoride.The colour of the ash from biological materials containing added metals as their nitrates or chlorides at the 200 pg level are as follows: chromium, orange; tin, grey; iron, rust brown; lead, white; and aluminium, white, Conclusions Standard Reference Materials (SRMs) ashed by the procedure described and analysed for tin, lead, iron and chromium show good recoveries. These good recovery values are to be expected because losses due to retention by silica residues would not arise. Silica is volatilised as SiF,. The comparative results show good agreement for tin, lead, iron and chromium in canned foods and added tin is fully recovered. In view of increasing material, chemical and energy costs and in spite of high initial capital cost there is strong evidence in favour of adopting this technique not only for reference work but also to support more rapid routine procedures.The advantages are real: (i) reduced applied man-hours; (zi) reduced chemical costs; (iii) fewer hazards compared with both dry and wet conventional oxidation; and (iv) less con- tamination, lower blanks, higher sensitivity, less interference are all related to the purity of the ash obtained and the absence of certain elements such as silicon. There is a need to explore the influence of other gases in their vibrational states, e.g., helium, argon, hydrogen, nitrogen and neon, on oxidation by oxygen - fluorine plasma chem- istry reactions because they may significantly increase the concentration of excited oxygen or fluorine atoms.September, 1982 1013 The author thanks the Director, British Steel Corporation-Tinplate for permission to publish this paper. Thanks are also due to the equipment manufacturers who loaned the equipment and to colleagues, in particular Mr. T. L. Williams, who assisted with some of the experimentation and Mr. T. M. English of B.S.C. Swinden Laboratories, Sheffield, for his work using X-ray photoelectron spectroscopy. MATERIALS IN PTFE DISHES FOR SN, FE, PB AND CR BY AAS 1. 2. 3. 4. 5. 6. 7. 8. 9. References Foner, S. N., J . Chem. Phys., 1956, 25, 601. Gleit, C. E., and Holland, W. D., Anal. Chem., 1962, 34, 1454. Lundell, G. E. F., Hoffman, J. I., and Bright, H. A., “Applied Inorganic Analysis,” Second Edition, Williams, E. V., J . Food Technol., 1978, 13, 367. Branson - International Plasma Corporation, Hayward, CA, USA, Product Bulletin 1971, No. 4801. Carter, G. F., and Yoeman, W. B., Analyst, 1980, 105, 295. Hirobe, K., and Tsuchihoto, T., J . Electrochem. Soc., 1980, 127, 234. Imperial Chemical Industries, Welwyn Garden City, Technical Service Note F10, Second Edition, Bock, A., “Handbook of Decomposition Methods in Analytical Chemistry,” International Textbook John Wiley, New York, 1953. Plastics Division (1978) and Note F12/13 (1978). Company, London, 1979. Received March llth, 1982 Accepted April 15th, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701006

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

1014 Analyst, September, 1982, Vol. 107, pp. 1014-1018 Determination of Chloride, Sodium and Potassium in Salted Foodstuffs Using Ion-selective Electrodes and the Dry Sample Addition Method 6. R. Chapman* and 1. R. Goldsmith Applications Laboratory, Pye Unicam Ltd., York Street, Cambridge, CB1 2PX The “dry sample addition” method of ion-selective potentiometry is described. This incremental or spiking method requires the addition of a known mass of solid sample directly to a standard solution containing the ion to be measured. Dissolution of the sample produces a change in electrode potential which is related to the sample composition. Dry sample addition has been applied to the analysis of foodstuffs containing up to 2.8% of salt. The chloride, sodium and potassium concentrations are determined by Philips solid-state and plastic ion-selective electrodes. Comparison of this method with titrimetry and atomic-absorption spectrometry gave satisfactory results.The advantages gained in cost, speed of analysis and reliability using ion- selective measurement are discussed. Keywords 1 Ion-selective electrodes ; dry sample uddition potentiometry ; deter- mination of chloride, sodium and potassium ; salted foodstufls The traditional methods for the determination of salt in foodstuffs have been titrimetric and photometric. Chloride has been determined principally by the Mohr argentometric titration method,l and either atomic-absorption spectrometry or flame photometry have been applied to the alkali metals. Both methods are for many foodstuffs either expensive, insensitive or time consuming.The Mohr titration procedure using coloured indicators is, in particular, an insensitive method for foodstuffs containing colouring agents or insoluble fibrous material. In recent years, ion-selective electrodes have been used for the determination of chloride in a number of foodstuffs, for example cheese2 and corn syrup.3 However, in practice these measurements have required the dispersal of the food in water followed by direct potentio- metry or titration with a silver salt. The application of incremental or spiking methods of ion-selective measurement, with the advantages of reliability and time saving, has not been tested on foodstuffs. This paper describes a dry sample addition method of ion-selective potentiometry .This little known incremental method involves the addition of a known mass of solid sample directly to a standard solution containing the ion to be measured. The electrode potential change arising from the dissolution of the soluble salts from the sample is related to its original composition. Dry sample addition has been applied to the determination of chloride, sodium and potassium in salted foodstuffs. ,4 variety of foods were tested ranging from powdered solids to oil-based pastes and containing up to 2.8% m/m of salt. The dry sample addition method effectively combines the sample dispersion and ion-selective measurement procedures into one time-saving step. Comparison of the results from ion-selective measure- ment with those from titrimetric and atomic-absorption spectrometric methods gave satis- factory results.Experimental Apparatus Three Philips ion-selective electrodes, a solid-state chloride (IS550-Cl), a plastic sodium (19561-Na) and a plastic potassium (IS561-K), were used. Ion-selective electrodes were stored in a l O V 3 ~ solution of the appropriate ion between measurements and overnight. A double-junction calomel reference electrode (Philips, Type R44/2-SD/l) contained salt hridrrp ~ l ~ r t r n l ~ r t ~ c nf fi 1 mnl 1-1 ammnnillm nitrQto fnr r n A ; n m qnrl nn+qcc;**m inn- - n ACHAPMAN AND GOLDSMITH 1015 measurements were made with a Philips digital pH - millivolt meter, Model PW9409, con- nected to a single pen chart recorder (Model PM 8251). All solutions were stirred continu- ously using a magnetic stirrer with small PTFE stirring bars.A Pye Unicam SP9-800 atomic-absorption spectrometer and an SP9 computer were used for the confirmatory tests for sodium and potassium. Reagents All solutions were prepared in de-ionised water from analytical-reagent grade materials. Chloride stock solution, 1000 mg 1-l. Prepared by dissolving 2.102 g of potassium chloride in water and diluting the solution to 1 1. Sodium stock solution, 1000 mg 1-1. Prepared by dissolving 2.544 g of sodium chloride in water and diluting the solution to 11. Potassium stock solution, 1000 mg 1-l. Prepared by dissolving 1.908 g of potassium chloride in water and diluting the solution to 1 1. Standard solutions of chloride, sodium and potassium, 10 and 100mg1-l. Prepared by sequential volume dilution of the appropriate stock solution.Reference electrode salt bridge electrolytes, 0.1 mol 1-1 ammonium nitrate solution and 1 mol 1-1 potassium nitrate solution. Prepared by dissolving 8 g of ammonium nitrate or 101.1 g of potassium nitrate in water and diluting the solution to 1 1. Ionisation bufers for atomic-absorption spectrometry, 10% mlrn sodium and potassium sulphate solutions. Prepared by dissolving either 10 g of sodium sulphate decahydrate or potassium sulphate in water and diluting to 100 ml. moll-l. Prepared by dissolving 0.849 g of silver nitrate in water and diluting to 1 1. Procedure Determination of the electrode sensitivity A preliminary calibration of the three ion-selective electrodes was made by measuring the electrode potential of 100-ml portions of the appropriate 10 and 100 mg 1-1 standard solutions of chloride, sodium or potassium. The electrode sensitivity (millivolt per decade change in concentration) was determined from the difference between the two values of electrode potential for each standard solution.This was carried out before each analysis of a fresh portion of sample using the dry sample addition procedure. Dry Sample Addition Procedure A 100.00-ml portion of the 10 mg 1-1 chloride, sodium or potassium standard solution was pipetted into a 150-ml beaker. The appropriate electrode pair was immersed in the solution and the electrode potential allowed to equilibrate for about 2-3 min and recorded. A portion of the untreated food sample was immediately transferred into the standard 10 mg 1-1 solution.The stirring speed was normally increased to encourage dissolution of the solid sample. The plastic boat was re-weighed and the mass added determined by difference. The final electrode potential after addition of the sample was recorded and the change in electrode potential obtained by difference. Determination of chloride, sodium and potassium by dry sample addition The dry sample addition method requires the addition of a known mass of solid sample directly to a standard solution of the ion to be measured. The salt is leached from the sample, a process encouraged by vigorous stirring of the standard solution. A subsequent increase in the salt concentration of the standard solution is sensed by the ion-selective electrode and a change in the electrode potential incurred.The attainment of a constant potential change confirms that all of the soluble salt has been leached from the sample. This change in electrode potential (AE) is proportional to the increase in ion concentration induced by the sample, Ca (gram ion per litre) and described by equation (1) which was derived from the basic Nernst expression for the ion-selective electrode, Silver nitrate for argentometric titration, 5 x A plastic boat containing a portion of the food sample was weighed accurately.1016 Analyst, Vol. 107 where S is the electrode sensitivity or slope (millivolts per decade change in concentration) obtained from the electrode calibration before measurement and CST is the concentration of initial standard solution (moles per litre).The 4= refers to the sign of the electrode slope and is positive for cations and negative for anions. Rearrangement of equation (1) enables the measured ion content of the solid sample to be calculated from equation 2 : CHAPMAN AND GOLDSMITH : CHLORIDE, SODIUM AND POTASSIUM IN CsTvsTAr [lo-/s-l x 100 . . (2) 1 Ion content of sample (% m/m) =: m where V s T (litres) is the volume of initial standard solution containing the ion to be measured, A , is the relative atomic mass of the measured ion and m is the mass of solid sample added (grams). ConJirmatory tests for determination of chloride, sodium and potassium After each dry sample addition measurement, a 20.00-ml portion of the standard solution containing dissolved foodstuff was taken and titrated with standard silver nitrate solution (5 x loF3 moll-1) using a mixture of potassium chromate and potassium dichromate as an end-point indicator.The back- ground level of 10 p.p.m. of chloride already present in the sample solution was subtracted from the measured concentration. After each dry sample addition measurement, a portion (5-15 ml) of the standard solution containing dissolved foodstuff was diluted and an ionisation buffer incorporated in the final solution a t a level of 2%. Similarly a 276 buffer was incorporated into both standard sodium and potassium solutions. Ionisation buffers of 10 06 sodium sulphate decahydrate and potassium sulphate solution were used for potassium and sodium measurements, respectively. Flame atomic absorption was used with a stoicheiometric acetylene - air mixture.Wavelength settings of 589.0 and 766.5 nm were chosen for sodium and potassium, respectively, with a band pass of 0.5 nm. The background level of 10 p.p.m. of sodium and potassium already contained in the sample solutions was subtracted from the measured concentration. Chloride measurement using the Mohrl titration method. Sodium and potassium measurement by atowtic-absorption spectrometry. Results and Discussion The dry sample addition method was applied to the determination of different salted food products commonly found on the British food market. The foods varied from dried crisp- bread, to powders, pastes and sauces containing up to 2.8% of salt. An accurate determination of electrode sensitivity was made before each fresh portion of sample was tested by dry sample addition. This procedure helped to establish the best accuracy of measurement and was used to examine any detrimental effects from contami- nants (i.e., oils) in the foods on the characteristics of the chloride, sodium and potassium ion-selective electrode membranes.Typical values of electrode sensitivity recorded during a series of six successive replicate chloride analyses of salad cream are included in Table I. Only small changes in the chloride electrode sensitivity were noted during the sequence of measurements. No significant changes in sensitivity were found for the chloride, sodium TABLE I CHANGE IN ELECTRODE SENSITIVITY DURING THE DETERMINATION OF CHLORIDE IN SALAD CREAM Sample portion No. Mass added/g 1 0.307 3 2 0.233 8 3 0.2609 4 0.21 18 5 0.2149 6 0.31 15 Chloride in sample/ g per 100 g 1.61 1.66 1.64 1.63 1.54 1.51 Electrode sensitivity before analysis/ mV decade-l 53.3 53.6 53.0 53.0 53.2 53.9September, 1982 SALTED FOODSTUFFS USING ION-SELECTIVE ELECTRODES 101 7 and potassium electrodes with any of the foods tested by dry sample addition.The dispersal of the food sample in a large volume of standard solution effectively dilutes the level of any contaminants impinging upon the ion-selective membrane surface. The accuracy and reproducibility of the dry sample addition technique is affected signifi- cantly by the magnitude and stability of the change in electrode potential after addition of the sample. In general, the monovalent electrodes of chloride, sodium and potassium should be subject to a potential change in the range 2040mV, and this was achieved for the foods tested.Clearly, small changes in electrode potential would incur significant errors of measurement, whereas changes in excess of a decade of concentration, i.e., greater than 50 mV, would affect the over-all ion concentration of the initial standard solution signifi- cantly, thus invoking the need for some form of ionic strength adjustment. Typical potential changes of between 20 and 40 mV were attained by using a 100-ml portion of the appropriate initial 10 p.p.m. standard solution; the sample masses were in the range 0.1-1 g. A constant potential change after sample addition was an indication of the complete dissolution of all unbound soluble salts in the food sample.Waiting times for dissolution varied from 2 to 10min using the fast stirring speed, the time allowed being dependent on the nature and consistency of the foodstuff. All electrode potential readings were taken as constant when a change of 0.1 mV or less was observed over a 30-s period. Results for the determination of chloride in piccalilli sauce, cream of chicken soup and salad cream are given in Table 11. Further results for the determination of sodium and potassium in salad cream, dried milk powder, peanut butter, dried crispbread and tomato ketchup are included in Tables I11 and IV, respectively. The average ion content of six TABLE I1 DETERMINATION OF CHLORIDE IN SALAD CREAM, PICCALILLI SAUCE AND CREAM OF CHICKEN SOUP BY DRY SAMPLE ADDITION Chloride in Range of sample Range of chloride electrode Range of potential change sample*/ Coefficient of Sample mass added/g sensitivity (S)/mV decade-' after addition (AE)/mV g per 100 g variation, % Salad cream .. . . 0.211 8-0.311 5 53.0-53.9 33.7-41.3 1.60 3.7 Piccalilli sauce . . . , 0.1164-0.2519 55.2-59.3 27.8-41.3 1.68 5.5 Cream of chicken soup (ready to serve). . . . 0.1746-0.3242 54.7-58.5 23.8-34.4 0.89 3.8 * Average of six replicate analyses. TABLE I11 DETERMINATION OF SODIUM IN DRIED CRISPBREAD AND PEANUT BUTTER BY DRY SAMPLE ADDITION Sodium in Range of sample Range of sodium electrode Range of potential change sample*/ Coefficient of Sample mass added/g sensitivity (S)/mV decade-' after addition (AE)/mV g per 100 g variation, % Dried crispbread .. . . 1.0206-1.2618 49.4-51.6 Peanut butter . . . . 0.4092-0.7561 50.0-52.3 Average of six replicate analyses. 28.4-31.1 0.25 2.0 18.8-30.2 0.37 8.7 TABLE IV DETERMINATION OF POTASSIUM IN MILK POWDER, SALAD CREAM, DRIED CRISPBREAD AND TOMATO KETCHUP BY DRY SAMPLE ADDITION Potassium in Range of sample Range of potassium electrode Range of potential change sample*/ Coefficient of Sample mass added/g sensitivity (S)/mV decade-' after addition (AE)/mV g per 100 g variation, yo Milk powder . . . . 0.178 5-0.408 7 50.2-53.1 Salad cream . . . . 0.3961-0.9028 48.6-50.3 Dried crispbread . . . . 0.9915-1.6333 47.5-51.1 Tomato ketchup . . . . 0.5633-0.8914 51.2-52.3 * Average of six replicate analyses. 32.8-45.9 1.78 2.8 4.5-8.6 0.06 6.6 34.7-43.0 0.38 4.7 25.2-31.4 0.37 5.51018 CHAPMAN AND GOLDSMITH replicate complete analyses is presented for each sample. The ranges of electrode sensitivity, sample mass and change in electrode potential are included with the coefficient of variation for the six replicate analyses. The reproducibility of the method is good as shown by CO- efficients of variation ranging from 2.0 to 8.7yi (n = 6) for the foodstuffs tested.A comparison of the potentiometric results with those from a Mohr titration for chloride and atomic-absorption spectrometry for sodium and potassium is made in Table V. Satis- factory agreement was found for the chloride, sodium and potassium results using the different methods with the majority of the food samples. The results show that ion-selective potentiometry using a dry sample addition procedure is a satisfactory method for the determination of chloride, sodium and potassium in a range of salted foodstuffs.Dry sample addition is a particularly sensitive technique for the determination of chloride in samples containing colouring agents or fibrous material which often mask or distort the coloured indicator end-point of an argentometric titration. In addition, ion-selective potentiometry secures a reduction in the cost per analysis with the elimination of expensive silver salts. A significant improvement in analysis time using dry sample addition rather than direct potentiometry is achieved in two ways. Firstly, the combination of sample dissolution and ion-selective measurement steps and, secondly, a simplified procedure for calculation of results directly from equation (2).A further improve- ment in analysis time is envisaged with some foodstuffs if the vigorous mechanical mixing methods, traditionally used in food analysis, were applied to the dissolution step of the sample, for example dried crispbread. Further, microprocessor-based ion-selective instru- mentation, e.g., the Philips PW 9416 Ion-Selective Analyser, provides for the direct read-ou t of sample concentration in percent. by mass by the dry sample addition method. This incremental or spiking procedure may now be applied to the determination of other ions contained in a solid as a soluble species. TABLE V COMPARISON OF DRY SAMPLE ADDITION METHOD WITH MOHR TITRIMETRIC AND ATOMIC-ABSORPTION SPECTROMETRIC METHODS FOR SALTED FOODSTUFFS Ion content of sample*/g per 100 g r--- A \ Ion-selective measurement using dry sample addition Mohr titration Atomic-absorption Sample Ion method method spectrometry Salad cream . . .. .. Piccalilli sauce . . .. . . Cream of chicken soup (ready to serve) . . .. Dried milk powder . . . . Salad cream . . .. .. Dried crispbread . . . . Tomato ketchup . . . . Dried crispbread . . . . Peanut butter . . .. .. Chloride Chloride Chloride Potassium Potassium Potassium Potassium Sodium Sodium 1.60 1.68 0.89 1.78 0.06 0.38 0.37 0.25 0.37 1.64 1.63 0.94 1.65 0.02 0.35 0.32 0.16 0.31 * Average of six replicate analyses, per sample and method. The authors thank Miss D.’C. Webb, Pye Unicam Ltd., for her assistance with the con- firmatory tests using atomic-absorption spectrometry. References 1 . 2 . 3. Vogel, A. I., “Quantitative Inorganic Analysis,” Fourth Edition, Longmans, 1978, p. 337. Randell, A. W., and Linklater, P. M., Aust. J . Dairy Technol., 1972, 27, 51. Jacin, H., Die Starke, 1973, 25, 271. Received March 30th, 1982 Accepted April 14th, 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701014

出版商:RSC    

年代:1982

数据来源: RSC

摘要:

Analyst, September, 1982, Vot. 107, pp. 1019-1025 1019 On4 i ne Electrochemical Detection of Oxidisa ble Organic Molecules of Pharmaceutical Importance W. Franklin Smyth Department of Chemistry, University College Cork, Cork, Ireland J. S. Burmicz Both ScientiBc, Farnborough, Hampshire and A. lvaska Department of Analytical Chemistry, Abo Akademi, 20500 Abo 50, Finland On-line electrochemical detection in the oxidative mode using a wall-jet electrode has been evaluated for 1,4-benzodiazepines and other drug mole- cules of widely differing structure. The method has been applied satis- factorily to the automated analysis of some 1,4-benzodiazepines, coefficients of variation of ca. 1-3% being obtained. Keywords : On-line electrochemical analysis ; 1,4-benzodiazepines The need to develop rapid and inexpensive methods of analysis for compounds of clinical and pharmaceutical importance has resulted in some polarographic and voltammetric on-line methods for their analysis.In these methods either the stream containing the electroactive material is continuously monitored by an electrochemical detector1 ,2 or a large number of samples are analysed by passing them repetitively past the detector and recording the current produced by each The great selection of electrochemical detectors that have been developed for high- performance liquid chromatography (HPLC) can naturally be used in instances when chroma- tographic separation is not neededJg-15 as is the case in this paper. These results are collected in Table I. TABLE I SUMMARY OF ON-LINE ANALYSES OF ORGANIC COMPOUNDS Reference Sample Working No.rate/h-l Compounds studied electrode* 3 60 Selected drugs DME 4 15 Benzodiazepines DME 5 20 Phenols Pt, c 6 - Great variety C 7 30 Benzodiazepines DME 8 120 Proteins DME Relative Concentration standard Mode? range studied deviation, % D.c., a.c. 0.4-10 mg ml-l - S.d.c. 5-25 ng ml-l 1.4 N.p. 4 x 10-6- 1.5 5 x 10-3gml-l 1 x 10-3- 1 x 1 0 - 6 ~ - D.c. D.c. 0.1 mg ml-l 0.3 D.p. 5-50 pg ml-l 1 * DME = dropping-mercury electrode. t D.c. = direct current; a.c. = alternating current; s.d.c. = sampled direct current; n.p. = normal pulse; d.p. = differential pulse. These electrochemical detectors can also be used in flow injection analysis where a small volume of the analyte is aspirated into a flowing background electrolyte. This method has been successfully applied to the analysis of organic compounds.12~16 The usable potential range of the indicator electrode used in this study, namely the solid glassy carbon electrode, is approximately from +1.5 to -1.2 V.The large positive potential range allows for the study of the oxidation of a great number of compounds that do not reduce and hence cannot be analysed by the DME or the thin mercury film electrode (TMFE). High flow-rates can also be used with solid electrodes, increasing the sensitivity of the technique owing to increased mass transport of the compounds to the electrode surface. With the DME, the flow-rate is critical because the form of the mercury drop is influenced by the flow-rate so1020 Analyst, Vol.107' that the drop formation is not always uniform. Solid electrodes are also robust and easy to handle. Their disadvantage is that the surface is not renewed and if the products of the electrode reaction are adsorbed on the surface the signal can be decreased. The effect of fouling can be reduced by selection of a suitable pulse mode so that the potential is applied only for a small fraction of the time required for amperometric detection. Choice of supporting electrolyte (e.g., to include methanol), washing periods between individual samples and electrochemical cleaning of the surface can also be employed in this context. The oxidative behaviour of many organic molecules at various stationary and rotating solid electrodes has been studied.17 The oxidative behaviour of 1,4-benzodiazepines, in particular, at rotating platinum and gold disc electrodes has been investigated by Volke et aZ.,ls who found responses at the electrode for oxazepam, diazepam and flurazepam in a supporting electrolyte of 0.1 M tetraethylammonium perchlorate in acetonitrile.They also found responses at the electrode for amitryptylene, diothiepin and dithiaden, which gave well defined waves in the potential region from +1.2 to +1.3 V. This paper is concerned with an investigation of the link-up of a glassy carbon wall-jet electrode contained in a flow cell with a Carlo Erba automatic analyser for the detection of oxidisable organic molecules of pharmaceutical importance and, in particular, an investiga- tion of its application to the automated determination of 1,4-benzodiazepines. FRANKLIN s h w m et al.: ON-LINE ELECTROCHEMICAL Experimental Reagents All chemicals used were of analytical-reagent grade and the pure benzodiazepines and their formulations were obtained from Hoffmann La Roche, Nutley, NJ, USA. A Britton - Robinson (BR) buffer solution of pH 4 was found to be optimum for the on-line analysis of 1,4-benzodiazepines. All solutions were 10% with respect to the methanol concentration intended to keep the organic molecules in solution and to aid dissolution of the reactants and products of electrode reactions from the surface to the indicator electrode. The 1,4- benzodiazepines and other nitrogen-containing drug molecules were prepared in their solutions at a concentration of 1 x 10-4 M. These were then placed in their respective compartments in the sample turntable prior to on-line electrochemical detection.Instrumentation The electrochemical cell was of the wall-jet typelg and is a slight modification of this design.20 The incoming solution impinged on the surface of the working glassy carbon electrode and when leaving the cell it passed the tip of the reference electrode. A platinum auxiliary electrode was placed near the working electrode. A PAR 174A polarograph was used for the voltammetric measurements and all potentials were measured against the saturated calomel electrode (S.C.E.). Solutions were pumped to the cell using a Carlo Erba automatic analyser, as shown in Fig. 1, which involved a turntable distributor, peristaltic pump and mixing coils.The currents were recorded with a strip-chart recorder and the measurements were performed at 22 "C. The connections were made to the PAR 174A via the colour-coded wires illustrated in Fig. 1. The pulse time was set at 0.5 s and the differ- ential pulse mode of operation was investigated. The time constant was set at 0.3 s and the modulation amplitude at 100 mV. Nitrogen was also introduced at the front end of the mixing coil as illustrated in Fig. 1 so as not to allow too great a dilution of the analyte, whilst allowing satisfactory mixing with the supporting electrolyte. Results and Discussion The 1,4-benzodiazepines flurazepam (I), oxazepam (11) , medazepam (111) , lorazepam (IV) , flunitrazepam (V), prazepam (VI) , potassium chlorazepate (VII), nitrazepam (VIII), amino- nitrazepam (IX) and chlordiazepoxide (X) represent a wide range of molecular structures and, in quiescent solution, give rise to the d.c.voltammetric data illustrated in Table I1 when subjected to electrooxidation at the glassy carbon electrode in BR buffer of pH 4.0.21 Results from a recent study22 using BR buffer of pH 12.0 are also included in Table 11.September, 1982 DETECTION OF OXIDISABLE ORGANIC MOLECULES b” (1) R’ = -(CH2)2N(CZH5)2; R2 = H; R3 = F ( 1 1 ) R’ =H; R2 = OH; R3 = H (IV) R’ =H; R2 = OH; R3 = CI (VI) R’ =cyclopropyl; R2 = H; R3 = H CH3 I N-CH CI C-N H 102 11022 Peristaltic Pump FRANKLIN SMYTH et al. : ON-LINE ELECTROCHEMICAL Analyst, Vol. 107 1r Waste r r- t m Buffer Recorder Mixina coils Fig. 1. Link-up of automatic analyser with the wall-jet electrochemical detector.TABLE I1 VOLTAMMETRIC DATA FOR 1 X loM4 M CONCENTRATIONS OF 1,4-BENZODIAZEPINES (I)-(X) AT THE GLASSY CARBON ELECTRODE PH 4 1 ,kBenzodiazepine Flurazepam (I) . . .. .. .. Oxazepam (11) . . .. .. .. Medazepam (111) . . . . .. .. Lorazepam (IV) . . . . .. .. Flunitrazepam (V) . . .. .. .. Prazepam (VI) . . .. .. .. Potassium chlorazepate (VII) . . .. Nitrazepam (VIII) . . .. .. .. Aminonitrazepam (IX) . . .. .. Chlordiazepoxide (X) . . .. . . r E,/V us. S.C.E. + 1.05 + 1.33 $0.88, +l.lO + 1.38 n.e.* n.e.* n.e.* + 1.49 +0.77, .+ 1.23 + 1.31 1 i d 4 5.0 4 4.3, 4.0 4 n.e.* n.e.* 4 n.e.* 3.8, i.d.p.t 6 pH 12 - E,/V us. S.C.E. +0.66 +0.72 +0.68 +0.8 n.e.* n.e.* +1.04, +l.2 + 1.14 + 0.45 +0.9 7 MtLA 3.2 2 10.1 2 n.e. * n.e.* 4, i.d.p.t 7.1 3.3 3 * n.e.= no electrooxidation occurs in the pH range investigated. t 1.d.p. = ill-defined peak. Certain trends in the mechanisms of the electrooxidation of these 1,4-benzodiazepines can be realised for the purposes of this paper. Oxidation of the l-N-substituted molecule flurazepam (I) takes place on the nitrogen atom of the l-N side-chain as its E , value in pH 12 buffer of +0.66V compares exactly with the E, value for the tertiary amine (CH3)2N(C2H5) in the same buffer. Presumably N-oxidation is the mechanistic route in aqueous supporting electrolytes. It is not expected that the 4-N atom is involved in electrooxidation, as the l-N-hydroxyethyl metabolite of flurazepam is not electrooxidisable in the pH range 4-12.22 In any case, the electron- Volke et aZ.18 also observed well defined peaks for I in acetonitrile with 0.1 M tetrabutyl- ammonium perchlorate supporting electrolyte at a rotating platinum electrode (E, = + 1.05 V).Chlordiazepoxide (X) has its azomethine group “blocked” to electrooxidation, as it is already N-oxidised, but has the electrooxidisable /N = C’ entity present. It \ is probably N-oxidised at one or other of these nitrogen atoms. The mechanisms of electrooxidation of the remaining 1,4-benzodiazepines can be grouped under the following functional group subdivisions : withdrawing effect of the fluorine atom would not facilitate oxidation of the ‘C = N / group. / NHCH,September, 1982 DETECTION OF OXIDISABLE ORGANIC MOLECULES 1023 and potassium chlorazepate (VII),* where R = C1, are reasonably difficult to oxidise.This is in agreement with the observations of Volke et aZ.,18 who found I1 to be oxidised at +1.55 V and that IV gave an ill-defined wave at the rotating platinum electrode. The mechanism probably involves initial formation of a radical cation -NH- followed by coupling reactions. N-Acetyl-9-chloroaniline behaves similarly in that an E , value of + 1.3 V (i, = 5.2 PA) is observed in BR buffer of pH 4 and +0.83 V (ip = 3.7 PA) in BR buffer of pH 12. When R = NO,, as in nitrazepam (VIII), resonance delocalisation of the lone pair on the 1-N atom is extensive and precludes electrooxidation in BR buffer of pH 4. Its metabolite, amino- nitrazepam (IX), is, however, relatively easy to oxidise, as illustrated in Table 11. In this instance it is probable that electrooxidation takes place at both the 1-N atom and the N atom in the 7-position.(ii) (y N- containing molecules, such as flunitrazepam (V) and prazepam (VI), are not electrooxidisable, primarily owing to the steric effect of the CH, group and the electron-withdrawing effect of the carbonyl group in position 2. When this latter group is absent, as in medazepam (111), the electron density is sufficient a t the 1-N atom to allow electrooxidation. The results in Table I1 were then applied to the on-line analysis of the 1,4-benzodiazepines I-X, individually at the The settings on the PAR 174A were set as follows: detector potential, +1.2 V; current setting, 0.05 mA (d.p.); pulse time, 0.5 s; modulation amplitude, 100 mV; and low pass filter (T), 0.3 s.M concentration level (Fig. 2). C 0 -0 m m E Fig. 2. Response of certain 1,4-benzodiazepines at the glassy carbon in the on-line mode. Conditions: medium, pH 4 Britton - Robinson buffer + 10% methanol; detector potential, + 1.2 V; current, 0.05 mA (d.p.) ; pulse time, 0.5 s ; and initial concentration, M (in methanol). * This molecule possesses - C (OH)-0-K+ adjacent to the 1-N atom. I1024 AnaZyst, VoZ. I07 At first, problems with adsorption were encountered, with only three compounds giving reproducible results. The adsorption problem was overcome by the addition of methanol to the buffer (10% by volume) and by extending the wash time slightly. Approximately 30 samples per hour are capable of being analysed using this system (Fig. 1) without poisoning of the electrode.However, after the processing of about 30 samples, the electrode was cleaned with chamois leather as a general precaution. Tablet contents were also analysed by this on-line method and the results are shown in Fig. 3. The tablet contents analysed were flurazepam (Dalmane, 30mg in 10ml of methanol), medazepam (Nobrium, 5 mg in 50 ml of methanol and 10 mg in 50 ml of methanol) and chlordiazepoxide (Librium, 5 mg in 50 ml of methanol). The conditions for the experi- ment were the same as for that shown in Fig. 2. As an indication of the reproducibility of results, six successive on-line analyses of the flurazepam formulation yielded a coefficient of variation of 2.57%, values of 0.85% and 1.250/6 being obtained for pure samples of oxazepam and lorazepam.In addition to the 1,4-benzodiazepines, the following nitrogen-containing drug compounds were found to exhibit oxidative behaviour at the glassy carbon electrode using BR buffer of pH 4 as the supporting electrolyte : paracetamol (XI), phenazone (XII), nicotine (XIII), adrenaline (XIV), dibenzepine (XV), nortryptyline (XVI) and diphenoxylate (XVII). FRANKLIN SMYTH et aZ. : ON-LINE ELECTROCHEMICAL HkCOCH3 XI L=ICH, XI1 /\NHCH3 @OH OH XIV Xlll xv XVI XVll It would therefore appear that on-line electrochemical detection could be applied to the determination of many nitrogen-containing drug molecules as either pure substances or as they are found in drug formulations. Routine tablet quality control analysis can be taken over by this type of detector system, in conjunction with the automatic analyser outlined in this section.The analysis of 30 benzodiazepine samples by reductive polarography using a DRiE would take approximately 3-4 h (i.e., 6-8min per sample), whereas one batch of 30 samples could be handled in 1 h by this system (2 min per sample). This detector, in addition, can obviously be used for the determination of many nitrogen-containing drugs at the low concentrations encountered in body fluids following extraction and HPLC separation.September, 1982 DETECTION OF OXIDISABLE ORGANIC MOLECULES 1025 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Fig. 3. Electrochemical detection applied to the “on-line” analysis of 1,4-benzo- diazepine formulations. Tablet sizes: A, flurazepam (Dalmane, 30mg in 10ml of methanol) ; B, medazepam (Nobrium, 5 mg in 50ml of methanol); C, medazepam (Nobrium, 10 mg in 50 ml of methanol); and D, Librium (5 mg in 50 ml of methanol). References Fehbr, Zs., Nagy, G., T6th, K., and Pungor, E., Analyst, 1974, 99, 699.Hackman, M. R., and Brooks, M. A., J . Pharm. Sci., 1978, 67, 847. Silvestri, S., Pharm. Acta Helv., 1972, 47, 209. Cullen, L. F., Brindle, M. P., and Paperiello, G. J., J . Pharm. Sci., 1973, 62, 1708. MacDonald, A., and Duke, P. D., J . Chromatogr., 1973, 83, 331. Pungor, E., Fehbr, Zs., and Nagy, G., Pure Appl. Chem., 1975, 44, 595. Lund, W., and Opheim, L.-N., Anal. Chim. Acta, 1975, 79, 35; 1976, 82, 245; and 1977, 88, 275. Alexander, P. W., and Shah, M. H., Talanta, 1979, 26, 97. Pungor, E., T6th, K., Fehdr, Zs., Nagy, G., and Varadi, M., Anal. Lett., 1975, 8, ix. Kissiner, P., Anal. Chem., 1977, 49, 447A. Heineman, W. R., and Kissinger, P. T., Anal. Chem., 1978, 50, 166R. Gilgen, P., and Rach, P., Chimia, 1970, 32, 345. Application Bulletin, No. 118, Metrohm, Zurich, 1975. Application Note LC4, EDT Research, London, 1978. Bollet, C., Claude, M., and Rosset, P., Analusis, 1978, 6, 54. Betteridge, D., Anal. Chem., 1978, 50, 832A. Adams, R. N., “Electrochemistry at Solid Electrodes,” Marcel Dekker, New York, 1969. Volke, J., El-Laithy, M. M., and Volkova, V., J . Electroanal. Chem., 1975, 60, 239. Fleet, B., and Little, C., J . Chromatogr. Sci., 1974, 12, 747. Burmicz, J., PhD Thesis, Chelsea College, University of London, 1979. Franklin Smyth, W., Ivaska, A., Davidson, I. E., Burmicz, J. S., and Vaneesorn, Y., Bioelectro- Ivaska, A., and Franklin Smyth, W., to be published. chem. Bioenerget., 1981, 8, 459. Received December lst, 1981 Accepted March l l t h , 1982

ISSN:0003-2654    

DOI:10.1039/AN9820701019

出版商:RSC    

年代:1982

数据来源: RSC