ISSN:0003-2654    

DOI:10.1039/AN98005FX021

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BX023

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005FP065

出版商:RSC    

年代:1980

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN98005BP071

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

JUNE 1980 The Analyst Vol. 105 No. 1251 Observations on the Determination of Total Copper, Iron, Manganese and Zinc in Foodstuffs by Flame Atomic-absorption Spectrophotometry W. H. Evans, Dorothy Dellar, Brenda E. Lucas, F. J. Jackson and J. I. Read Department of Industry, Laboratory of the Governmefit Chemist, Cornwall House, Stamford Street, London, SEl 9NQ A method is described for the determination of total copper, iron, manganese and zinc in foodstuffs. Organic matter is destroyed by wet oxidation and measurement is made directly upon the sulphuric acid digests by flame atomic-absorption spectrophotometry . In the measurement, direct inter- ferences from the inorganic species found in foodstuffs are shown to be a function of burner design and usage. Elimination of these interferences may be achieved for optimum conditions of use, thereby avoiding systematic bias.Isolation of the sources of variation during measurement reveals indirect interferences that enhance the standard deviation of measurement, despite individual presentation of solutions. The accuracy of the method in applica- tion to foodstuffs is assessed+for the levels of each element normally present. The standard deviations of results are compared with those for measurement alone and the further influence of indirect interferences on the former is inferred. Derived confidence intervals and limits of detection are given for each element. Keywords : Copper, iron, manganese and zinc determinations ; foodstug analysis ; $ame atomic-absorption spectrophotometry Each of the elements copper, iron, manganese and zinc is considered to be essential to animal and human physiology.For this reason and their relative abundance in foodstuffs, some of the earlier applications of flame atomic-absorption spectrophotometry were directed to measurement of these elements in plant material.lS2 Gorsuch3 originally showed that losses of the four elements did not occur in the initial destruction of organic matter whether by wet oxidation, by any combination of acids normally used or by dry combustion; for the last method some care was necessary, however, for copper. Recent examples of the equivalence of the two destruction techniques are available for plant material, using nitric acid - perchloric acid wet oxidation, with final measurement by flame atomic-absorption spectrophotometry being made in hydrochloric acid s~lution.*~~ Final measurement has also been made on nitric acid solutions of digests,6 but on the occasions that sulphuric acid has been used in the destruction stage, concentration of these elements by chelating ion exchange' or by chelation and extraction into solvent followed by solubilisation with nitric acid8 has been preferred. In this laboratory, methods have been evaluated for cadmium, lead and nickelg and antimony, arsenic and tinlo with the primary objective of qualifying levels recorded for the total diet survey of the UK.For effective use of digests prepared for these six elements it was considered desirable to measure also copper, iron, manganese and zinc. The method of destruction of organic matter uses nitric acid and sulphuric acid wet oxidation, giving digests which are in 3-5% V/V sulphuric acid.Evaluation of the direct measurement of these digests from foodstuffs prompted the observations contained in this paper. Crown Copyright. 529530 Reagents water. Ultrar or an equivalent grade of acid be used. EVANS et al. : OBSERVATIONS ON THE DETERMINATION OF TOTAL Analyst, Vol. 105 Method All reagents should be of the grade specified; solutions should be prepared with distilled Nitric acid, sp. gr. 1.42. It is recommended. that BDH Aristar, Hopkin and Williams Sulphuric acid, sp. gr. 1.84, and dilute (1 + 1!3). Standard solutions. Primary standard solutions, specially prepared for atomic-absorption spectrophotometry, can be purchased from commercial sources and contain 1 .O g 1-1 of the element in 1 N acid.Dilute each primary standard solution 1 + 19 to give solutionscon- taining 50 mg 1-1 of the element and store in cleaned polythene bottles. Take 10 ml each of copper and manganese, 50 ml of zinc and 100 ml of iron solutions containing 50 mg 1-1, add 20 ml of sulphuric acid (1 + 19) and dilute to 500 ml. The resulting solution will contain 1, 1, 5 and 10 mg 1-1 of copper, manganese, zinc and iron, respectively, and should be stored in a cleaned polythene bottle. To 5 ml of sulphuric acid (sp. gr. 1.84) diluted to 30 ml in each of a series of calibrated flasks (0-G), add by pipette 0, 2, 5 , 10, 15, 20, 30 and 50 ml of the above composite solution and dilute to 10Qml with water. Store in pre-conditioned polythene bottles and discard after measurement of 12 series of standards for any of these elements. Recommended grade as for nitric acid.Composite standard solutions. Each solution will contain the concentrations shown in Table I. TABLE I COMPOSITIONS OF COMPOSITE STANDARD SOLUTIONS Concentrations in mg 1-l. Solution Element 0 A B C D E F G Copper . . .. .. 0 0.02 0.05 0.1 0.15 0.2 0.3 0.5 Iron . . .. . . 0 0.2 0.5 1.0 1.5 2.0 3.0 5.0 Manganese .. .. 0 0.02 0.05 0.1 0.15 0.2 0.3 0.5 Zinc . . .. . . 0 0.1 0.25 0.5 0.75 1 .o 1.5 2.5 Apparatus All glass apparatus must be kept permanently full of 1 N nitric acid when not in use. Atomic-absorption spectrophotometers. During this investigation two different double-beam spectrophotometers were used; these are referred to as instruments 1 and 2 but the burner design used with each instrument is the important factor. Background correction was not used except for checking sample levels and single-element hollow-cathode lamps were used as emission sources.Instrument 1 was fitted with a 100-mm single-slot burner, the incident beam being 10 mm above the burner. The air and acetylene flow-rates were 4.5 and 1.0 1 min-l, respectively, and the nebuliser flow-rate was standardised at 5.8 ml min-l. An integration time of 4 s was chosen for digital recording and was operator subjective ; automatic zeroing between readings was not used. Instrument 2 was fitted with a 100-mm triple-slot burner, the incident beam being 30 mm above the burner. Air an& acetylene were operaked at pressures of 4.5 and 4.8 lb in-2 and the nebuliser flow-rate was standardised at 4.6 nil min-l.The integration time chosen for digital recording was 10 s and was instrument objective; automatic zeroing between readings was used. Procedure Prefiaration of sample digests These should be prepared according to method (1)C of the Analytical Methods Cornmittee,ll using precautions subsequently des~ribed.~ The resulting 100 ml of digest, which is inJune, 1980 COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS 531 nominally 5% V/V sulphuric acid, should be colourless and should not contain any suspended solids. Prepare at the same time two reagent blanks from the volumes of acid used in sample oxidation. Measurement Tune each emission source to give maximum sensitivity to noise ratio according to the maker’s instructions, at wavelengths of 327.4 nm for copper, 248.3 nm for iron, 279.5 nm for manganese and 213.9 nm for zinc.Adjust the instrument conditions to the fixed con- ditions described. Arrange the samples, standards m d blank solutions in random order, but with one sample reagent blank in each half of the total series. For instrument 1, zero the instrument and allow a sufficient time to attain a constant reading on the display before presentation of the series of solutions for an individual element. Aspirate each solution in turn, taking a constant digital reading for a 4-s integration time. Re-read in reverse order and again in the original order to give three readings for each solution. For instrument 2, record the digital readings for consecutive 10-s integration times, zeroing between each pair of readings if necessary, and then re-read the series in reverse order to give a total of four readings for each solution.The best calibration may be obtained graphically, but for the results contained in this account calibration was based on a third-order polynomial curve-fitting computer program. Experimental Each variable in each stage of a method should be investigated to ensure the absence of systematic bias for levels normally expected. In this method, the strength of the acid in the final prepared digest may vary and there is a choice in the conditions for measurement. Because of the different compositions of foodstuffs, the time taken for destruction of organic matter may vary and the resulting sulphuric acid digest is usually in the range 3 4 % V/V.For optimum conditions of measurement for each instrument and burner design, typical net absorbance values for single integration readings for standards B and F are tabulated for different acid strengths (Table 11). Blanks were in proportion to the concentration of sulphuric acid used. Within the above acid concentration range no effect is apparent, but an increase in absorbance occurs towards the lowest acid concentration. TABLE I1 NET ELEMENTAL RESPONSES FOR DIFFERENT ACID CONCENTRATIONS Element . . Instrument* . . Standard . . Sulphuric acid concentration, 5 . . .. 4 . . . . 3 . . .. 2 . . .. 1 . . . . Absorbance x 1000 r A. -3 .. Copper Iron Manganese Zinc ..1 2 1 2 1 2 1 2 + +7 A A + +7 h . . B F B F B F B F B F B F B F B F %- . . 28 163 39 234 77 458 98 518 43 221 44 267 76 401 91 517 . . 27 162 40 237 79 464 99 522 35 221 44 267 78 404 94 520 . . 32 170 40 238 83 473 99 525 38 224 48 273 78 408 93 517 .. 32 168 41 240 81 473 103 540 37 227 50 278 81 408 96 528 . . 28 170 42 246 81 485 102 546 37 226 47 281 81 416 97 538 * For instrument 2 a scale expansion of 5 was used for copper and manganese, 2.5 for iron and 0.5 for zinc. Interferences It has been reported that the sensitivity of atomic-absorption spectrophotometric measure- ments is influenced by burner design, the height of the radiation beam in the flame and the fuel to air ratios,12 and further, these may also affect the sensitivity of measuring the same element in different acids.13 Cresser14 has pointed out that both the fuel to air ratio and the height of the radiation beam in the flame may also affect the extent of interference from532 EVANS et al.: OBSERVATIONS ON THE DETERMINATION OF TOTAL Analyst, VOZ. 105 other inorganic species on a particular element. Negative interferences have been recorded during the measurement of manganese when calcium and magnesium ratios to sulphate are less than 2.15 Negative errors have also been noted for iron and manganese, but not copper, when hydrochloric acid solutions are measured containing calcium plus magnesium to phosphorus ratios of less than 2, and these interferences are not necessarily removed by addition of a releasing agent such as lanthanum chloride.16-18 Similar interferences noted in the recovery of all four of the present elements added to plant material suggested the most reliable procedure for the determination of these elements in sulphuric acid solution was separation from high levels of interfering inorganic species8 The maj or inorganic species present and the acid medium of the final digest therefore assume greater importance than maximising the sensitivity of response for the elements measured via fuel to air ratios, burner type and conditions of use of the burner.The major inorganic species in digests from retail foods will be sodium, potassium, calcium, magnesium and phosphorus, the last element being present as orthophosphate after dilution of a concentrated sulphuric acid digest. Inspection of the McCance and Widdowson tables19 indicates that 100-ml volumes of digests prepared from l o g of listed foodstuffs would not contain individual amounts exceeding 200,rng of sodium, potassium and phosphorus, 40 mg of calcium and 50 mg of magnesium, with the exception of inorganic chemicals used in food preparation, some shellfish r6taining saline water or calcium in certain cheese products.The number of foodstuffs that would give digests containing half of these amounts of major inorganic species should not exceed 5% of the listed retail foods. In the UK total diet survey it is unlikely that digests prepared from any homogenate group, other than for calcium in liquid milk, would contain amounts exceeding a quarter of those mentioned. The major inorganic species were therefore tested for interference individually for the listed highest amounts and, when netessary, at a half and a quarter of these amounts, upon composite standards 0, B and F '(in 5% V/V sulphuric acid).Each species was added as sulphate for sodium, potassium 'and magnesium, as carbonate for calcium, and as sodium dihydrogen orthophosphate for phosphorus. Further standards 0, B and F were prepared containing mass ratios of calcium"p1us magnesium to phosphorus of 1.8 and 0.9. Each complete series of solutions was measured with standards 0-G within the same series of readings as detailed in the procedure for both burner (instrument) systems, for different burner heights and different flame conditions. The best curves were obtained graphically. Single integration readings only were taken to obtain trends; 95% confidence limits (95 C.L.) are derived from the repeatability, so (within-series standard deviation) obtained for calibration at the optimum burner'conditions.Such limits are restrictive as they do not include the contribution to variatiori-from the blank or that from solution preparation. The direct interferences found for the major single interfering species, calcium and magnesium, are illustrated in Table 111. For the single-slot burner, the response declined with the height of the radiation beam in the flame, and this was reversed for fuel-rich conditions, whereas for the triple-slot burner response was more consistent for the .listed variations. Each blank 0 containing the inter- fering species gave on average small positive responses above the standard blank for each element tested, with one exception for manganese, and the equivalent of these responses did not exceed the levels for the tested element expected in the AnalaR reagents used for the interfering species.For each element, the response for the 0 solution plus inorganic species was subtracted from those of solutions B and F if positive, and if the response was negative, as could occur from random influences, that of standard 0 was subtracted. All readings were made without background correction and the total dissolved solids were, for example, 6200 mg 1-1 when sodium was present and 7 600 mg 1-1 when phosphorus was present. For manganese with concomitant magnesium the response of blank 0 greatly exceeded that expected, and this level of manganese was confirmed by periodate oxidation and measure- ment of the permanganate.The response of blank solutions containing magnesium reflected in turn the interference illustrated in Table I11 for. different conditions of measurement. There were no significant direct interferences from sodium or potassium for either burner system for the conditions outlined in Table 111. Those for calcium and magnesium are self- explanatory and exist to the extent that for low concentrations of iron and manganese, responses could disappear from a combination of systematic and random bias. (The observa- tion may be made that measurement by a method of standard additions should be treatedJune, 1980 COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS 533 TABLE I11 INTERFERENCES IN THE MEASUREMENT OF COPPER, IRON, MANGANESE AND ZINC FOR DIFFERENT BURNER DESIGNS, BURNER HEIGHTS AND FUEL TO AIR RATIOS Element .. .. .. .. Concentration/mg 1-1 . . .. Instrument l-single-slot burner- 95% confidence limits* . . Height of incident beam above burner base- Copper Iron Manganese Zinc b b - 7 - 0.05 0.3 0.5 3.0 0.05 0.30 0.25 1.5 f0.23 &0.06 h0.14 f0.05 f0.36 f0.06 fO.09 f0.03 Relative difference in response with and without added element Added calcium 40 mg- 6mm .. .. .. 10mm .. . . .. lOmm, lean . . .. 10 mm, luminous .. 14mm , . .. .. 18mm . . .. .. -0.14 0.00 -0.35 -0.21 -0.54 -0.40 -0.01 -0.07 -0.16 +0.04 -0.10 -0.02 -0.36 -0.14 -0.12 -0.01 -0.28 -0.02 -0.10 +0.02 -0.01 -0.02 -0.04 +0.02 -0.17 +0.05 -0.36 -0.20 -0.53 -0.31 0.00 -0.03 +0.29 +0.01 -0.02 +0.03 -0.04 1-0.03 -0.05 +0.01 +0.23 -0.03 -0.07 +0.07 -0.06 +0.01 -0.05 +0.02 Added magnesium 50 mg- 6mm .... . . -0.17 -0.06 -0.58 -0.40 -0.54 -0.37 -0.25 -0.06 10mm .. . . . . -0.08 +0.05 -0.19 -0.06 -0.08 +0.04 +0.03 +0.03 lOmm, lean . . . . -0.13 -0.03 -0.23 -0.04 +0.05 +0.05 +0.03 +0.04 10 mm, luminous . . -0.17 -0.02 -0.57 -0.38 -0.30 -0.27 -0.21 -0.03 14mm . . .. . . +0.17 +0.03 -0.13 +0.05 $0.23 +0.09 +0.09 +0.04 18mm .. .. . . +0.06 -0.01 +0.07 +0.06 -0.01 +0.08 -0.03 f0.01 Instrument 2-triple-slot burner- 95% confidence limits* . . Added calcium 40 mg- 10mm .. .. .. 14mm . . .. .. 18mm .. .. .. 22mm .. . . .. 30mm . . .. , . 30mm, lean . . . . 30 mm, luminous . . Added magnesium 50 mg- 10mm .. . . . . 14mm . . .. . . 18mm . . .. . . 22mm . . .. . . 30mm . . .. .. 30mm, lean . . . . 30 mm, luminous .. f O . 1 1 - 0.04 -0.05 0.00 f0.03 -0.01 +0.04 $0.15 f0.05 -0.01 + 0.01 $0.01 0.00 - 0.03 -0.02 0.00 fO.10 -0.56 -0.45 -0.36 -0.23 +0.02 + 0.02 -0.04 f 0.03 -0.41 -0.29 -0.21 -0.12 0.00 -0.05 - 0.01 f0.23 - 0.72 -0.61 -0.48 - 0.29 -0.08 -0.08 - 0.36 * f0.04 -0.63 - 0.50 - 0.34 -0.17 - 0.05 - 0.06 -0.21 fO.09 -0.07 - 0.08 -0.07 - 0.09 +0.01 +0.05 -0.05 f0.05 -0.15 - 0.08 - 0.06 - 0.04 - 0.02 0.00 -0.05 %. -0.34 -0.31 -1.00 -0.74 -0,95 -0.77 -0.56 -0.44 -0.23 -0.17 -0.73 -0.55 -0.63 -0.54 -0.41 -0.21 -0.13 -0.10 -0.59 -0.39 -0.52 -0.33 -0.30 -0.11 -0.30 -0.03 -0.41 -0.24 -0.31 -0.19 -0.16 -0.04 0.00 -0.02 -0.08 -0.01 f0.03 +0.04 +0.04 +0.01 -0.04 0.00 -0.15 0.00 +0.17 0.00 0.00 +0.02 -0.04 -0.01 -0.15 -0.05 -0.11 -0.08 -0.11 -0.02 * Relevant 95% confidence limits are derived from the repeatability of single readings for both burner s ys tems .with caution.) A significant negative interference occurred with phosphorus for the triple- slot burner at the lowest height of the radiation beam in the flame. The interferences from calcium and magnesium at relevant measurement heights were reduced by the presence of phosphorus, acting as a releasing agent, and this reduction was greater for the mass ratio of 0.9. Clearly, a normal fuel mixture and a height for the radiation beam above the burner base of 30 mm were desirable to remove interferences for the triple-slot burner and a height of 10 mm appeared suitable for the single-slot burner.The cause of such interferences has been outlined previously by Alkemade20 and is defined as a condensed-phase effect in which the elements under consideration are occluded by the interfering salt species resulting in incomplete volatilisation in the flame. This effect is smaller when measurement is made in a higher position in the flame but is accentuated by the cooler fuel-rich flames. To confirm the absence of direct interference for the selected conditions, the exercise was repeated on both burner systems for the optimum conditions described in the procedure, recording three or four integration readings. The 95 C.L. values were the relevant ones534 EVANS et a,!. : OBSERVATIONS ON THE DETERMINATION OF TOTAL Analyst, Vo,!. 105 derived from so and were restrictive for the reasons already stated.The results are given in Table IV. For the triple-slot burner the only significant interference was from calcium on manganese and it appears that calcium concentrations above 100 mg 1-1 in sulphuric acid solution will cause low results. (It may be noted that for digests obtained by the procedure from liquid milk, the ratio of calcium plus magnesium to phosphorus is slightly greater than unity and results would not be discredited from interference considerations.) Also tested were possible interferences from the major trace elements likely to be en- countered in foodstuffs. Tin (1 mg) and combinations of iron (1 mg), zinc (1 mg), copper (0.2 mg) and manganese (0.2 mg) on relevant elements yielded no significant interferences. Similar absence of interference for these major trace elements was noted for the single- slot burner and whenever copper was measured in the presence of the major inorganic species under test.The negative interference from the highest levels tested of calcium on man- ganese and magnesium on iron, apparent in Table 111, was confirmed. There was also a tendency for positive interference from magnesium on manganese and zinc and phosphorus on iron and manganese to occur. Interference from phosphorus may be caused by its action as a releasing agent, but it is difficult to explain why this should happen with magnesium. All interferences with one exception (iron), however, were removed from solutions containing calcium plus magnesium and phosphorus, and in practice it would be expected that sulphuric acid digests from most foodstuffs would contain combinations of these elements at mass ratios of less than unity.For samples of the UK total diet survey, measurement using the triple-slot burner should show no significant direct interferences and, for retail foodstuffs generally, special precautions would be required for only a minimum number of foodstuffs if 5% V/V sulphuiic acid digests were prepared according to the procedure. As this burner is associated with instrument 2 , possessing objective integration recording, this combination was used for the results con- sidered subsequently. Cali bration In previous methods involving measurement of elements such as lead and arsenic9,lO by atomic-absorption spectrophotometry, calculated standard deviations have included the experimental uncertainty associated both with measurement and separation (concentration) and it has not been possible to isolate that frlom measurement alone with ease.In the present method this should be possible. For the results presented for retail and reference materials, 15 series of standards 0-G were measured by the 5 analysts for each element, using the defined procedure, during a period of 2 months with instrument 2 in average con- dition and subjected to rapid changes of use. Two series of standards for each analyst were processed according to the following design. Also included were a pair of total reagent blanks (TRB) associated with each calibration line. Each net response for standards A-G may be expressed in terms of standard B and the average net response calculated for each calibration line. The ratio of the sum of net responses for each standard to the sum of the net average response is recorded in Table V and indicates that calibration is linear for copper and manganese, linear for zinc up to 1 mg l-l, but slight curvature may exist for iron throughout the calibration range; for the purpose of this exercise linearity was assumed, however, up to 2 mg 1-l.(The recorded ratios are within 95 C.L. imposed by the reproducibility, s, of measurement for the linear ranges of each element .) The standard deviation for single 10-s integration readings between pairs of consecutive integration readings for each standard in the 10 series of standards may be calculated for 20 degrees of freedom (d.f.) and expressed as the coefficient of variation (C.V.) from the average net response for each standard; the C.V.for standard 0 and TRB will be obtained from responses for standard A. From these, the standard deviation for the mean of four integration readings may be obtained by dividing by 2 and applies for 10 d.f. (TRB 20 d.f.). This represents the instrument noise, Sn, for measurement at each standard level. In practice, it is desirable to measure as many sample solutions as possible with each series of standard solutions and consequently the total time for measurement for an element may approach 1 h. During this time, instrument response may drift or drift because of the difference in the nature of the solutions presented may occur.By aspirating the entireJune, 1980 COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS TABLE IV INTERFERENCE EFFECTS ON MEASUREMENT OF COPPER, IRON, MANGANESE AND ZINC FOR OPTIMUM BURNER CONDITIONS 535 Element . . .. Copper Iron Manganese Zinc Concentration/nig 1-1 . , 0.05 0.3 0.5 3.0 0.05 0.30 0.25 1.6 Instrument l-single- slot burner- 95 confidence b - 7 - b limits* Added ion Na+ K+ Ca2+ Mg2+ Po 43- Ca2+ Mg2+ Ca2+ Mg2+ Ca2+ Mg2+ &0.16 f0.04 f0.08 50.03 h0.21 f0.04 f0.06 f0.02 Relative difference in response with and without added element$ r A \ a . .. Amount/ mgt 200 200 40 20 10 50 25 12.5 6.25 200 100 50 25 40 100 20 25 25 +0.02 + 0.15 - 0.02 - - +0.15 - - - 0.00 - - - + 0.03 +0.01 - 0.02 - - 0.00 - - - + 0.03 - - - -0.01 +0.01 - 0.04 - - ( - 0.12) (- 0.02) (+O.ll)§ (+0.04)§ - 0.03 0.00 $0.05 - 0.03 0.00 - 0.01 -0.01 - - (-0.03)s ( + 0.04) + 0.01 +0.01 ( + 0.05) 3 (+ 0.05)s + 0.03 f0.01 +0.11 -0.01 (-0.15)s ( - 0.14) ( + 0.1 0) (+ 0.11)s ( + 0.10) § ( + 0.06) 9 (+0.07)§ ( + 0.10) § ( + 0.09) 9 + 0.04s +0.16$ 0.00 + 0.04 (- 0.09) (- 0.01) (+ 0.04) ( + 0.05) § (+ 0.08)s ( + 0.09) 3 (+ 0.09)§ (+O.lO)§ (+ 0.11)s + 0.065 + 0.065 f0.20 -0.01 -0.075 -0.065 +0.32 -0.04 -0.02 -0.02 0.00 +0.03 f0.04 +0.01 0.00 f0.05 0.00 0.00 +0.07 -0.02 +0.06 +0.01 +0.06 +0.01 +0.05 -0.01 P048- Over-all average non-significant differences .. $0.07 +O.Ol 0.00 +0.01 +0.04 f0.01 +0.01 0.00 Instrument 2-triple- slot burner- 95 % confidence limits* . . . . Added Amount/ ion mgt Na+ 200 K+ 200 Ca2+ 40 20 10 50.06 &0.03 f0.05 -f 0.02 50.13 h0.02 50.05 f0.03 0.00 0.00 -0.01 + 0.01 -0.01 - - - - 0.05 -0.01 (-0.04) - - - 0.01 (- 0.04) - (+O.Ol) - 0.05 f0.01 -0.02 (-0.02) - - 0.00 (-0.03) -0.01 - (+O.Ol) - 0.06 +0.06 - 0.04 - 0.01 $0.01 -0.01 - 0.09 + O .l l ( - 0.06) 3 ( - 0.10) 5 + 0.03 +0.04 + 0.09 +0.05 - t 0 . 0 1 + 0.03 (- 0.08) (-0.04)s - 0.02 + 0.02 +0.02 0.00 - -0.01 -0.01 -0.01 - -0.06 + O . l l - - - -0.01 - 0.02 - - 0.00 -0.03 - - Mg2+ 25 50 200 100 Caa+ 40 Caaf Mg2+ 50 PO,+ 100 Over-all average non-significant differences . . -0.01 -0.02 - 0.09 0.00 -0.01 0.00 + 0.01 - 0.02 +0.03 + 0.02 + 0.02 + 0.04 - 0.01 0.00 -0.02 -0.01 0.00 -0.01 + 0.02 +0.01 -0.01 0.00 * Relevant 95% confidence limits are derived from the repeatability of the means of 3 -,>d 4 readings for t For 100-ml of digest prepared from 10 g of foodstuff, the concentration of interfering species as mg kg-' # ( the single- and triple-slot burners, respectively.can be obtained by multiplying by 100. ) Implies a duplicated set of readings. < 5% significance.536 EVANS et aZ. : OBSERVATIONS ON THE DETERMINATION OF TOTAL AnaZyst, VOZ. 105 series of solutions in one direction and then in reverse order, bearing in mind the random order of solutions, the standard deviation within the series of measurement may be obtained. This will originate from two sources of variation, that from instrument noise, sn, and that caused by aspirating at different times within the series and may be obtained for single integration readings (30 d.f.) by an analysis of variance.The standard deviation of the mean of four such counts may be obtained by dividing by 2 (10 d.f.) and expressed in a similar way to sn in relation to net response; this will be the repeatability, so, of measurement for the five analysts concerned in the exercise. If linearity is proved or proved for part of the calibration, the standard deviation of the difference between net mean individual responses at each concentration (A, etc.) with the net average response for individual calibration lines, expressed as for standard B, will be a measure (unconnected with sn and so in derivation) of the between-series variation of measure- ment (the reproducibility, s) for the mean of four integration readings. (The inability to derive s for points in these calibration ranges gives credence to the preference for linear Standard solution . ... Copper . . .. .. Manganese . . .. Iron . . .. .. Zinc . . .. .. Copper- For snr instrument noise so, repeatability . . s, reproducibility . . so, single-slot burner, instrument 1 . . Manganese- For s n, instrument noise so, repeatability . . s, reproducibility . . so, single-slot burner, instrument 1 . . Iron- For Sn, instrument noise so, repeatability . . s, reproducibility . . so, single-slot burner, instrument 1 . . Zinc- For sn, instrument noise so, repeatability . . s, reproducibility . . so, single-slot burner, instrument 1 . . TABLE V CALIBRATION DATA .. TRB* 0 A B C D E F G Ratio of net standard to net average responset r A I .. - - 0.97 0.98 1.00 1.00 1.01 1.00 1.00 ..- - 1.07 1.01 1.00 0.99 1.00 1.01 0.99 .. - - 1.05 1.01 1.01 0.99 0.98 (0.95) (0.88) .. - - 1.01 1.00 1.00 1.00 1.00 (0.98) (0.95) Coefficient of variation of measurement, yo, for the mean of 4 integration readings, t $ . . 6.6 . . 7.1 - .. .. 12 . . 2.9 . . 4.0 - .. . . 8.2 . . 2.2 . . 4.4 - .. . . 5.9 * . 2.2 . . 4.0 - .. . . 4.5 3.7 6.6 - 15 2.2 5.0 - 8.9 2.2 5.4 .- 7.2 2.3 4.3 - 4.2 3.3 4.9 12 20 5.0 6.8 13 24 2.0 5.1 11 8.4 2.1 3.6 5.4 5.4 1.5 2.6 2.9 7.4 3.3 6.5 7.4 10 1.0 2.4 2.8 3.9 1.5 2.2 2.2 2.6 0.7 1.5 1.5 4.2 1.2 2.3 3.4 3.4 0.8 1.7 1.8 1.9 0.6 1.8 1.2 1.5 0.6 0.7 1.1 2.8 0.9 2.0 1.6 3.1 0.7 1.6 1.5 1.6 0.6 1.9 1.7 1.4 0.6 0.9 1.9 2.2 0.6 2.0 2.8 2.6 0.6 1.7 2.8 1.3 0.6 1.8 2.2 1.5 0.5 1.2 1.3 1.6 0.4 1.3 3.2 1.6 0.6 0.8 (6.3) 1.4 0.6 1.2 (3.4) 0.9 0.4 1 .o 1.3 1.0 0.4 1.9 2.6 1.2 0.5 1.3 - 0.9 0.4 0.9 - 0.7 * The coefficients of variation expressed for the total reagent blank (TRB) are for 20 degrees of freedom The ratios of the magnitude of TRB to standard 0 were 1.11, 1.01, (d.f.) ; all the remainder are for 10 d.f.1.07 and 1.09 for copper, manganese, iron and zinc, respectively. The base-line noise levels reported for instrument 2, triple-slot burner, aspirating distilled water, are 0.0005 mg 1-I for copper, manganese and zinc, and 0.001 5 mg 1-1 for iron. For comparison with succeeding tables the C.V.s may be related to concentration as mg 1--1 by reference to Table I, e.g., the noise levels for standard A are 0.0007, 0.001, 0.004 and 0.002 mg 1-1 for copper, manganese, iron and zinc, respectively. For 100 ml of digest obtained from 10 g of sample these concentrations may be converted to amount, as pg, by multiplying by 100.t ( ) Refers to standard solutions not used in the calculation of the mean.June, 1980 COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS 537 calibration previously mentioned.21) No values for s can be given for standard 0 and the TRB, as factors other than measurement contribute to variation. For the calibration standards and TRB, Sn, so and s expressed as C.V. are presented in Table V: Also included, for information, are the values for so obtained for the single-slot burner fitted to the instru- ment 1. The instrument noise, sn, expressed as the C.V. decreases with increasing analyte con- centration, with the exception of zinc, for which it is constant for the greater part of the calibration range.When expressed as concentration, s n is relatively constant for the lower measured concentrations, again with the exception of zinc, for which it increases throughout the calibration range. These findings in relation to the C.V. agree with those of a procedure based upon theoretical considerations for evaluating standard deviations of instrument noise in atomic-absorption measurement^.^^^^^ This procedure has revealed that below an absorbance of 0.2, which will apply to copper, iron and manganese in this work (Table 11, allowing for scale expansion), the C.V. is controlled by one of flame transmission noise, source flicker noise or signal shot noise, or by combinations of these effects, with a resulting decrease in the C.V. as the measured concentration increases.For absorbance measure- ments in the range 0.2-1.0, in general, the C.V. of s n is controlled by analyte absorption noise to give relatively constant values for the range, and reference to Tables I1 and V for zinc reflects this. The analyte absorption noise in turn has sources of variation that may depend on flame changes and fluctuations in flame atom population densities caused by nebulisation variations. Consideration of so reveals similar trends to those in sn for the standard deviations expressed as C.V. and concentration. It might be expected that reasonable agreement would be obtained between so and sn at the same concentrations. The 95 C.L. for the ratio of standard deviations for 10 d.f.against 10 d.f. will be 0.52-1.93 (0.64-1.57 for TRB), and inspection of Table V indicates that 23 of 36 such ratios are exceeded and for standard C and above 17 of 20 such ratios are exceeded. This statistical significance must be caused by instru- ment drift. Despite individual presentation of solutions for measurement, indirect inter- ferences from inorganic species in solutions could occur with subsequent responses, type (d),1° and, or, with subsequent variation, type (c) ,lo in extended series of measurements; for the design of measurement used in the procedure type (d) would reveal an increase in the standard deviation of measurement. These in turn could be caused by nebulisation variations producing fluctuations in flame atom population densities.This, however, should not be confused with the analyte absorption noise, which limits the variation of instrument noise, Sn, for high absorbances. Observations on the trends for s in this exercise reveal a similar situation to sn and so for concentration points within the linear calibration range of the four elements. The 95 C.L. for the ratios of s and so, each teskd for 10 d.f., but excluding standard A, indicate that only two ratios are exceeded. This suggests that there is no statistically significant difference for within- and between-series of readings, and in terms of measurement therefore no signi- ficant difference between the five analysts. The standard deviation of the difference between low concentrations of an element similar in magnitude to the standard blank will be affected by the variation inherent in the blank giving a value up to .\/2 of that obtained for so and this is the source of significance for standard A for three elements.The literature for collaborative inter-laboratory exercises involving instrumental measurement indicates ratios of mean squa.res for between and within laboratories that often are significant at a probability level of 0.05 or less. This significance may be caused by mistaking sn for so, the latter should always be obtained under practical routine conditions with a random distribution of solutions for extended series of measurements. Both sn and so were also calculated for 10 series of solutions of €3 and F containing the high levels of inorganic species that interfered for certain burner measurement conditions but not for the conditions finally accepted.In no instance for any of the elements were 95 C.L. for the ratios of standard deviations with those for standard solutions exceeded. This indicates the absence of indirect interference, type (b) ,lo on variation of measurement Le., in the presence of large amounts of foreign ion. Finally, to eliminate any possible subjective opinion during the calibration that could occur by graphical expression, recourse was made to a computerised calibration based on a third-order polynomial curve-fitting program. This is naturally biased for the upper538 EVANS et aZ. : OBSERVATIONS ON THE DETERMINATION OF TOTAL Analyst, VoZ. 105 standards F and G, but for the remainder there will be a deviation for each point in the measurement range from the best fitted curve from which standard deviations may be calculated.The averages for standards O-E (54 d.f.) for each element were 0.0013, 0.0142, 0.0024 and 0.0067 mg 1-1 for copper, iron, manganese and zinc, respectively. The 95% confidence intervals (95 C.I.) can be calculated from these and, where a calibration point for standards O-E exceeded these limits, the source was sought. A total of 26 of the 360 calibra- tion points were omitted and results are based on the corrected calibration lines. Results To assess the accuracy of the total procedure in the presence of foodstuffs, recoveries of all four elements were made from total diet honiogenates containing the lowest levels present of the elements under consideration.The amounts added were 5, 10 and 25 pg of copper and manganese, 50, 100 and 250 pg of iron and 25, 50 and 125 pg of zinc to 20 g of milk, 5 g of fats and 10 g of the remaining homogenates (cereals, meat, fish, fruit, root vegetables and green vegetables). Recovery was determined in duplicate or triplicate, each base level being simultaneously determined in duplicate and all values for a diet homogenate being contained within a series of determinations. These recoveries were repeated periodically over a number of years, giving averages summarised in Table VI for 21-27 results. Inspection suggests the absence of levels of interfering species, although the recovery level for the fats homogenate for manganese and zinc is lower than average. No further use can be made of the individual results because of the wide range of values possible for the base levels.TABLE VI AVERAGE RECOVERIES OF ADDED COPPER, IRON, MANGANESE AND ZINC Each value is the mean of 21-27 results at three added levels. Food homogenate Cereal . . .. Meat . . .. Fish . . .. Fats .. .. Fruit . . .. Root vegetables . . Green vegetables . . Milk .. .. Average . . .. Copper, 10 pg* .. 96 .. 97 .. 101 .. 99 . . 99 .. 101 .. 100 .. 100 .. 99 Recovery, yo A I Iron, 79 pg* Manganese, 13 pg* Zinc, 73 pg* 100 96 100 101 105 105 101 101 98 97 93 92 100 104 98 100 99 102 101 103 101 98 96 100 100 100 100 * Average content in homogenates. The accuracy of the method in application to foodstuffs by the team of analysts was obtained for standard reference materials, further retail foodstuffs being added to the list to encompass the range of measurement. Duplicate total analyses were made by each analyst, each duplicate being included within the same series of results, and these results are sum- marised in Tables VII and VIII.The mass of reference material digested should reflect the equivalent wet mass and on occasion the highest point of the calibration range was exceeded; dilution was made in order to permit measurement within the calibration range. As one object of this exercise was to follow sources of variation and in particular the source from measurement alone by atomic-absorption spectrophotometry, these dilutions assume some importance and attention is drawn to the footnotes in Table VIII. Discussion Accuracy The accuracy of the method in application to reference foodstuff materials can be assessed by consideration of Table VII.In general, the means of the replicated analyses agree with accepted or concensus values for copper, iron and zinc within the 95 C.I. imposed by the variation of results of this exercise. An exception is NBS liver, for which higher means wereJune, 1980 COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS 539 obtained for iron and zinc. It may be noted that two different containers of this material were used during this exercise and the ratio of the mean from one (four results) to the other (six results) for the four elements was between 1.06 and 1.09. There was no apparent difference in the moisture content of the material in the two containers.The means for manganese are higher than the accepted values and outside the 95 C.I. imposed by this exercise. Although the major inorganic species in these reference materials are within the levels tested (Table IV), the calcium content of biological kale (about 40000 mg k g l ) would inevitably understate the manganese measured even by this exercise. Using 1 g of kale the direction for preparation of digests is satisfied, but wet oxidation of 2 g of kale gives a precipitate of calcium sulphate. For all four elements in kale there was no significant difference for the means obtained on digesting the different masses. This conceals, however, certain subtler effects of disobeying the procedure in the strictest terms, as will be shown in the next section.EVALUATION Reference material . . Coppey- Mean . . .. Certified* f 9 5 C.I. of Lean' Ivan- Mean . . .. Certified* . . .. f95 C.I. of mean Manganese- Mean . . . . Certified* . . . . f 9 5 C.I. of mean Zinc- Mean . . . . Certified" . . . . &95 C.I. of mean TABLE VII OF ACCURACY OF METHOD I N APPLICATION TO FOODSTUFFS Concentration/mg kg-1 . . .. .. .. . . .. .. . . . . . . . . .. . . Bowen's kale, 1 g 4.99 4.99 0.41 117 118 4 15.6t 14.7 0.8 33.1 33.2 1.4 Bowen's kale, 2 g 5.29 4.99 0.33 116 118 5 15.9t 14.7 0.5 32.0 33.2 1.5 NBS tuna NBS spinach NBS liver 3.27 12.1 198 12 193 0.08 0.4 7 - 58.0 545 2781. 550 268 0.9 20 8 - 0.51 170t 1l.Ot - 165 10.3 0.12 4 0.4 14.7 50.8 142t - 50 130 0.5 1.9 4 * The values for Bowen's kale are consensus means.24 t These values are outside expected levels.Variation The variation of results obtained by this method in application to foodstuffs is treated in identical fashion to the design previously described.21 Therefore, the repeatability, so, is the standard deviation of analytical uncertainty obtained by known analysts on representative foodstuffs and is unencumbered by sources of variation other than the homogeneity of the samples examined. The reproducibility, s, reflects the standard deviation obtained for any analyst on any foodstuff to which the method is applicable and includes the variation between series of estimates and the variation inherent in the total reagent blank. Each value for so and s listed in Table VIII is an estimate subject to variation; thus, so for 5 d.f.may vary between 0.62 and 2.45 of the calculated value and s for 9 d.f. may vary between 0.69 and 1.83 of the calculated value. Comparison of so and s for the results with those of standard solutions (Table V) give 95 C.L. for the ratios of these standard deviations which are 0.49 to 2.57 for so and 0.51 to 1.99 for s. The range of amounts of manganese determined is from less than 1 to 27 times the derived limit of detection (see Derived Factors). Both so and s, expressed as C.V., decrease as the determined amounts increase, with the exception of retail dried potato. Expressed as540 Element Copper Manganese Iron . . Zinc . . EVANS et aZ. : OBSERVATIONS ON THE: DETERMINATION OF TOTAL Analyst, VoZ. 105 TABLE VIII REPLICATE ANALYSIS FOR COPPER, IRON, MANGANESE AND ZINC ON FOODSTUFFS AND STANDARD MATERIALS Foodstuffs material* Coffee$ Dried milks Stewed apple Tuna Bowen’s kale Dried potato Bowen’s kale NBS spinach Flour NBS tuna NBS liver§[[ Tuna Dried milk NBS tuna Stewed apple Dried potato Bowen’s kale NBS livers Bowen’s kale Flours NBS spinach811 Coffee11 Dried milks Dried potato Stewed apple Tunas Bowen’s kale Flour Bowen’s kale NBS tuna Coffee$ NBS spinachSlj NBS livers Stewed apples Dried potato Coffee Bowen’s kale NBS spinach Flour$ Bowen’s kales NBS tuna$ Dried milks Tuna$ NBS liver Sample masslg 5 2.5 10 10 1 2 2 1 10 5 2 10 2.5 5 10 2 1 2 2 10 1 5 2.5 2 10 10 1 10 2 5 5 1 2 10 2 5 1 1 10 2 5 2.5 10 2 Mean content/ mg kg-l 0.35 0.75 0.29 0.31 4.99 2.70 5.29 1.40 3.27 12.1 198 0.07 0.59 0.51 0.61 4.75 15.6 11.0 15.9 5.40 170 53.3 19.4 3.02 3.97 7.42 117 116 19.2 58.1 88.5 545 278 0.58 9.46 3.85 33.1 50.8 32.0 14.7 47.9 12.1 5.52 142 Range of content/ mg kg-l 0.26-0.48 0.43-1.01 0.23-0.33 0.27-0.36 4.40-5.59 2.49-3.09 4.70-6.01 1.34-1.47 3.14-3.46 11.3-13.0 187-211 0-0.20 0.30-1 .OO 0.32-0.81 0.53-0.66 4.11-5.67 14.1-16.9 14.6-16.9 10.3-11.7 5.19-5.82 163-178 52.2-54.7 1.76-4.25 3.68-4.30 6.42-8.21 17.3-23.7 107-123 112-126 18.4-19.8 5 6.1-59.9 84.8-90.2 514-586 265-289 0.42-0.71 8.72-10.9 3.33-4.44 30.9-35.1 47.8-56.2 29.5-34.5 13.5-15.9 44.4-52.0 11.1-13.0 5.21-5.85 137-151 Amount/ v s 1.74 1.88 2.89 3.12 4.99 5.41 10.6 12.1 14.0 16.4 395 0.73 1.46 2.55 6.05 9.50 15.6 22.0 31.7 54.0 170 267 7.55 38.9 39.7 74.2 117 192 231 291 443 545 556 5.8 18.9 19.3 33.1 50.8 55.2 64.0 73.5 120 121 284 Repeatability, so ,------A 7 % 9.8 9.3 5.4 7.4 5.1 5.1 3.5 2.1 1.9 2.2 11 - 28 20 11 6.8 5.8 2.2 3.7 1.9 1.7 1.4 14 9.8 4.9 3.8 4.4 1.7 4.3 1.8 0.5 2.0 1.5 7.4 6.0 4.1 4.5 5.3 1.5 3.1 1.8 2.5 1.5 2.2 0.17 0.21 0.27 0.17 0.37 0.28 0.54 0.42 0.30 0.31 8.8 0.67 0.41 0.52 0.41 1.09 0.91 0.49 1.17 1.00 2.9 3.7 1.1 3.8 1.9 2.8 5.2 3.2 5.1 2.1 8.5 0.43 1.14 0.80 1.5 2.7 0.8 2.0 1.3 3.0 1.8 6.2 10 11 Reproducibility, s 95% confidence intervals( &I)/ mg kg-lt % 26 26 10 8.8 8.0 6.6 8.6 4.8 3.2 3.2 4.7 I 44 33 12 7.3 5.8 5.3 4.6 3.6 3.5 1.7 28 11 4.9 8.5 4.4 2.4 5.9 2.2 2.3 5.0 3.8 16 7.8 6.0 5.0 5.3 3.6 6.8 5.0 5.6 5.6 3.5 LG 0.45 0.49 0.30 0.28 0.40 0.36 0.91 0.58 0.45 0.53 18.4 0.67 0.64 0.84 0.44 1.14 0.91 1.17 1.47 1.9 5.9 4.4 2.1 4.3 1.9 6.3 5.2 4.6 6.4 14 10 27 21 0.93 1.48 1.15 1.6 2.7 2.0 4.3 3.7 6.7 6.7 9.9 0.044, 0.10 0.053, 0.11 0.069, 0.068 0.044, 0.063 0.095, 0.090 0.071, 0.081 0.14, 0.21 0.11, 0.13 0.077, 0.10 0.080, 0.12 2.3, 4.2 0.11, 0.14 0.13, 0.19 0.11, 0.10 0.28, 0.26 0.21, 0.21 0.13, 0.26 0.30, 0.33 0.26, 0.44 0.75, 1.33 0.95, 0.99 0.28, 0.47 0.98, 0.97 0.44, 0.44 0.72, 1.42 1.2, 1.2 0.82, 1.04 2.6, 3.2 1.3, 1.4 0.53, 2.3 2.8, 6.1 2.2, 4.8 0.11, 0.21 0.29, 0.33 0.21, 0.26 0.39, 0.34 0.61, 0.61 0.21, 0.45 0.51, 0.97 0.33, 0.84 0.77, 1.5 0.48, 1.5 1.6, 2.2 - * Each row represents 10 results obtained by 5 analysts, except for 1 g of Bowen’s kale for which 8 results were obtained by 4 analysts.t These limits are for a single result, the first value derived from repeatability and the sechnd derived from reproducibility.Concentrations are based on a 10-g sample mass. $ 1% analyst significance. 5% analyst significance. 11 Measurement was made on a prepared solution diluted by a factor o€ 10 for copper and a factor of 5 for iron and manganese. amounts, both so and s are relatively constant for determined amounts up to 6 pg. Except for NBS liver, for reasons already noted, and NEE spinach, 3 of 45 analyst means exceed the relevant 95 C.L. Comparison of the standard deviations with those for measurement of standard solutions for equivalent amounts indicate that ratios fall within 95 C.L. for so with the exception of the retail dried potato and kale (1 g), suggesting inhomogeneous distribution of manganese in the examples examined. Similar comparison for s indicates that ratios of 95 C.L.are not exceeded, except for retail dried potato, The range of amounts of iron determined is from less than 1 to 55 times the derived limit of detection. The gradations for so and s with determined amounts are well defined, with the following exceptions. For so, kale and coffee are atypical. The amount determined in retail dried milk is below the limit of detection and analyst significance is exposed from contributions from the TRB that elevate s. For the remaining foodstuffs, analyst significance is caused by the abnormally low so for retail coffee, for the noted reason for NBS liver, leaving retail tuna and NBS spinach to account for 4 analyst duplicate means of the remaining 40 exceeding the relevant 95 C.L. The 95 C.L.for the ratios of so to the values obt5ined for standard solutions for equivalent amounts are not exceeded in any instance, and as far as it is possible to compare for s there is similarly an absence of significance.June, 1980 541 For copper the amounts determined are from less than 1 to 8 times the derived limit of detection, with one exception, reflecting normal levels in foodstuffs. Suitable gradations for so and s are apparent as the amount determined increases, and expressed as amounts are relatively constant up to 5pg. Both retail coffee and dried milk approach the limit of detection in applications where it would be expected that contributions from the TRB would elevate s. For the remainder of the listed foodstuffs, excluding NBS liver, 3 out of 45 analyst means are outside the relevant 95 C.L.Comparison of so with values obtained for standard solutions shows that the 95 C.L. of the ratios are exceeded only for kale, but a similar com- parison for s shows such ratios are always exceeded for equivalent amounts. For zinc the range of amounts determined is 1-70 times the derived limit of detection. Suitable grada- tions are apparent for so and s as the amount determined increases, with the exception of NBS spinach. It is particularly noticeable that analyst significance is apparent at the upper end of the range determined, at least 2 analyst duplicate means in each instance causing the significance obtained. With the exception of spinach, ratios of so to values obtained for standard solutions do not exceed the 95 C.L., but similar ratios for s are exceeded whenever tested.It may be further noted that s expressed as C.V., is constant for the entire range of determined amounts. To summarise, in the foregoing account explanation has been attempted for the instances that distort a general pattern for each element. In general for manganese and iron, non- significant contributions to the reproducibility may occur, which are attributed to analysts, but for these two elements s for the entire procedure does not differ greatly from that for measurement. This implies that the destruction of organic material by wet oxidation, the only other stage in the procedure, does not contribute significantly to the variation of the total procedure for the range of determined amounts noted.For copper, non-significant contributions to s attributed to analysts may occur and s is invariably significant with respect to the reproducibility of measurement, while for zinc significant contributions to s occur, attributed to analysts, which makes s significant with respect to the reproducibility of measurement. Many investigations of the destruction of organic material by wet oxidation have failed to reveal any weakness in this digestion procedure for copper and zinc. It is difficult to accept that either element exists in forms that destruction of organic material makes un- evenly available because of non-equivalence of analysts, and which could be revealed by a low repeatability of results. The magnitudes of the TRB in this exercise averaged 1.6, 4.4, 23 and 6.7 pg for copper, manganese, iron and zinc, respectively, but these values include the background absorption from sulphuric acid and typically the corresponding true elemental levels are 1.0, 1.0, 5.0 and 2.0 pg.While the variation of TRB would contribute to the reproducibility of results for small amounts, this contribution would be small for the larger amounts determined. The conclusion must be that the source of analyst contributions to s emanates from between-series measurement. During this exercise, it was noted for zinc that aspiration during a series of measurements of a quiescent solution of kale (2 g) containing calcium sulphate sediment caused the response to drift for some time, and this clearly affects the responses of solutions subsequently aspirated.Under Calibration it was shown that the repeatability of measurement is dependent on sources of instrument noise to a small extent, but despite individual presenta- tion of solutions it is clearly more influenced by instrument drift. The latter may be caused by flame or nebulisation differences from aspirating different solutions giving indirect inter- ferences with subsequent responses and variation, i.e., indirect interferences type (d) and (c) ,lo respectively. This increase in the standard deviation of within-series measurement on standards may be accentuated in the results derived from different calibration lines (i.e., a double effect), and this may be the cause of the significant between-series variation with measurement for copper; this effect could equally have become apparent for iron and manganese. A similar explanation is offered for zinc, except that in this instance high absorbances are measured (greater than 0.2 for the last eight samples in Table VIII), which theoretically are subject to even greater effects from flame or nebulisation variation causing enhanced indirect interference of type (d) and hence type (c).The result in this exercise for zinc is a consistent level for the reproducibility, expressed as the C.V., but this may be fortuitous even though a similar pattern occurs for zinc for instrument measurement, sn, from analyte absorption noise. COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS542 EVANS et aZ. : OBSERVATIONS ON THE DETERMINATION OF TOTAL Analyst, VoZ.105 Derived Factors The derivation of a practical limit of detection has been described previously and is based on the formula to.oo5s.21 While on occasion derived factors would need to be based on the repeatability of results, so, for normal survey analysis those derived from the reproducibility, s, should be used as comparison would be required of results obtained between series of determinations. For copper, excluding the foodstuffs for which analyst significance is present, the limit of detection calculated from amounts determined up to 5 pg (retail apple, tuna and kale) will be 1.0 pg (24 d.f.); for 10 g of sample digested to produce 100 ml of solution this is 0.1 mg kg-l. Similar considerations for manganese (retail dried milk, tuna and apple) and iron (retail dried potato and apple) give limits of detection of 0.2 and 1.0 mg k g l , respectively.For zinc, the continual increase in s, expressed as amount, reflecting that previously noted for measurement of standard solutions, may cause over- statement of the limit of detection, but calculated for retail dried potato and coffee it appears to be 0.4 mg k g l . It may be noted that the limits of detection derived from s and from so for relevant degrees of freedom are similar for each element. Both sn and so have been calculated for standard 0 and the total reagent blank (Table V). As two total reagent blanks, prepared independently, were measured with each series of standards, an estimate can be calculated for the standard deviation inherent in the blank within a series of measurement.This may be done by pairing the first integration count for each blank in a measurement series, the second pair, and so forth. An analysis of variance will give the standard deviation of a single integration reading from which that for the mean of four such readings can be obtained. The standard deviations so obtained were 0.12, 0.22, 1.0 and 0.32 pg for copper, manganese, iron and zinc, respectively, which differ little from amounts for so obtained for single total reagent blanks. Such standard deviations cannot be used in the calculation of derived factors unless three conditions are satisfied: (a) that the blank receives the same treatment as samples, which in a method involving destruction of organic material is not possible2] ; (b) that a very large number of within-series reagent blank results are available as the equation used in calculation of the limit of detection is a mathematical inequality with a limited number of results21; and (c) that the basing of factors on a standard deviation derived for within-series variation cannot logically be used to qualify results obtained between series of determinations.If the limit of detection had been used on the standard deviation inherent in the reagent blank within series of measure- ments and calculated from to,o,,@2 x standard deviation, the resulting limit for each element would have been 0.4-0.5 of that calculated. These levels are not substantiated and the reasons are clear from the foregoing account. The 95 C.I. for each element are listed in Table VIII for each of the foodstuffs and reference materials.For copper a 95 C.I. of 0.1 mg kg--l should apply for the range 0.1-0.5 mg kg-l and intervals of 20% above 0.5 mg k g l , while for manganese the intervals would be 0.2 mg kg-l for the range 0.2-1.0 mg kg-l and 20% above 1 mg kg-l. For iron a 95 C.I. of 1 mg kg-1 is suggested for the range 1-5 mg kg-l and 20% above this range, and for zinc a 95 C.I. of 20% is suggested for the entire range of determination. These intervals are wider than cursory inspection of atomic-absorption spectrophotometry might reveal initially, but they are derived from a system that must be recognised as routine practice. The results so obtained are not in any way discredited provided that accuracy is assured by elimination of systematic bias.Conclusion A method of determining total copper, iron, manganese and zinc is described, involving destruction of organic matter by wet oxidation and direct measurement on sulphuric acid digests. In this method the measurement stage has been extensively investigated in order to avoid interferences from the major inorganic species present in foodstuffs. In particular, different burners, different heights of the radiation beam in the flame and different flame conditions have been assessed. Serious interferences have been delineated and special attention is drawn to misuse of burner systems. For optimised conditions of measurement and for the normal levels of interfering species encountered in foodstuffs, sulphuric acid digest solutions can be measured with the absence of systematic bias.The sources of variation in the measurement stage of standard solutions for each element, during routine use, haveJune, 1980 COPPER, IRON, MANGANESE AND ZINC IN FOODSTUFFS BY FLAME AAS 543 been isolated for the instrument noise, repeatability and reproducibility. Whereas there is no significance between the latter pair, high statistical significance occurs between the instru- ment noise and repeatability of measurement, which implies instrument drift from indirect interference with subsequent responses and hence enhanced variation for subsequent measure- ments in extended series of measurements. Despite individual presentation of solutions for measurement, the source of this indirect interference is the difference in nature of sample solutions that will cause subsequent nebulisation changes producing fluctuations in flame atom population densities.Application of the total procedure to reference and retail foodstuffs indicates a satisfactory accuracy for the method from agreement with certified or concensus values, with the exception of manganese, for which the means of results for this exercise are significantly higher. Com- parison of the repeatability of results with that of measurement of standard solutions reveals that there is seldom statistical significance, except when there is an inhomogeneous elemental distribution in the foodstuffs examined. Similar comparison of the reproducibility shows that statistically significant differences may or may not occur for results from between-series measurement. The source of this contribution is believed to be the same as the indirect interference for within-series measurement, ie., repeatability, doubly reflected for between- series measurement. This is most marked for zinc for which high absorbances are measured and which would be expected to be subject to greater effects from nebulisation variation than copper, iron or manganese. Because for survey analysis derived factors such as the 95% confidence intervals and the limit of detection should be based on the variation of results, these factors will be larger than might be anticipated. They will reflect, however, the effectiveness of the method for the four elements in foodstuffs for routine practice. This paper is published with the permission of the Government Chemist. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. David, D. J., Analyst, 1958, 83, 655. Allen, J. E., Spectruchim. Acta, 1959, 15, 800. Gorsuch, T. T., Analyst, 1959, 84, 135. Isaac, R. A.. and Johnson, W. C., J . Assoc. Off. Anal. Chem., 1975, 58, 436. Boline, D. R., and Schrenk, W. G., J . Assoc. 08. Anal. Chem., 1977, 60, 1170. Capar, S. G., J . Assoc. Off. Anal. Chem., 1977, 60, 1400. 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. Evans, W. H., Read, J. I., and Lucas, B. E., Analyst, 1978, 103, 580. Evans, W. H., Jackson, F. J.. and Dellar, D., Analyst, 1979, 104, 16. Analytical Methods Committee, Analyst, 1960, 85, 643. Agemain, H., Aspila, K. I., and Chau, A. S. Y., Anal. Chem., 1975, 47, 1038. Fujiwara. K., Haraguchi, H., and Fuma, K., Anal. Chem., 1975, 47, 1670. Cresser, M. S., Lab. Pract., 1977, 26, 171. Bradfield, E. G., Analyst, 1974, 99, 403. Oelschlaeger, W., and Bestenlehner, L., Ladwirtsch. Forsch., 1976, 29, 224. Oelschlaeger, W., Schmidt, S., and Lautenschlaeger, W., Landwirtsch. Forsch., 1976, 29, 21 1. Oelschlaeger, W., Schmidt, S., and Bestenlehner, L., Landwirtsch. Forsch., 1976, 29, 70. Paul, A. A., and Southgate, D. A. T., “McCance and Widdowson’s The Composition of Foods,” HM Alkemade, C. T. J., Anal. Chem., 1966, 38, 1252. Evans, W. H., Analyst, 1978, 103, 452. Bower, N. W., and Ingle, J. D., Anal. Chem., 1976, 48, 686. Bower, N. W., and Ingle, J. D., Anal. Chem., 1977, 49, 574. Bowen, H. J . M., J . Radioanal. Chem., 1974, 19, 215. Stationery Office, London, 1978. Received October l l t h , 1979 Accepted December 21st, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500529

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

544 Analyst, June, 1980, Vol. 105, pp. 544-550 Automated Signal Recording System for the Perkin-Elmer 240 Elemental Analyser D. T. Burns, H. 6. McKnight, R. K. Quigg and W. J. Swindall Department of Chemistry, The Queen's University of €3elfast, Belfast, BT9 5AG, Northern Ireland An automated signal recording system for the Perkin-Elmer 240 elemental analyser is described. The construction of the system involves two main areas, an electronic interface (between the instrument and the digital volt- meter - printer) and a microswitch and cam assembly (to supply signals from the elemental analyser) . Illustrative results are reported for selected standard compounds. Keywords : Elemental analysis ; signal recording ; automation The use of a Perkin-Elmer (PE) 240 elemental analyser for the determination of carbon, hydrogen and nitrogen has been simplified by the incorporation of an automated signal recording system.Construction of the accessory involved two main areas, the ancillary electronic construction for the interface between the instrument, the digital Voltmeter (DVM) and the printer, and the mechanical construction of the system to supply control signals from the instrument to the printer. Over-all Electronic Requirements The method chosen to convert the d.c. analogue voltage from the analyser to digital form was to use a 4;-digit digital voltmeter of 1.9999 V span preceded by an amplifier of x 100 gain. The basic digital voltmeter uses the dual slope integration technique and this, together with an auto-zero correction, ensures that this part of the system is virtually free from zero drift .The x 100 pre-amplifier, in the digital voltmeter, is the main source of measurement error in such a system, but even here zero drift is only about 0.1 pV "C-l and full-scale drift about 40 p.p.m. "C-l after warm-up. The output of the digital voltmeter is presented, for each digit, in four-line binary-coded decimal (BCD) format via optically coupled isolators, thus effectively eliminating earth-loop problems, The active part of each of the six measurement phases of the PE 240 lasts only about 20-30 s and for the rest of the time of each sample run, 12 min, there is no output to the This gives a maximum reading of 19.999 mV with a resolution of 1 pV. Data ident. and command Recorder and DVM B1 I voltag:, output m icro- I 19.999 I m DVM, print command channel ident.-1 9.999 '1 u Printer I switches I P-E 240 Fig. 1. System layout.BURNS, MCKNIGHT, QUIGG AND SWINDALL 545 pen recorder or, in the present case, the DVM-printer. The DVM chosen here, however, normally samples at 1.6 readings per second and the printer prints out at 2 lines s-1; consequently, to avoid being swamped with redundant data it is necessary to arrange that +vcc +vcc +vcc 4.7 K Fig. 2. Print command board. no readings are printed when there is no output from the PE 240. In addition, when there is an output from the PE 240, of duration 20-30 s per element phase, no more readings are taken than are necessary for satisfactory results. It was decided to sample the voltage five times during each of the six active measurement phases.This choice of five voltage samples readily enables the operator to decide if the data were stable or were changing during a given phase. N' m icroswitch ciosed C' microswitch closed Open closed High I C2a low High IC1.3 low High IC2b low High IC2c low U b 0 30 60 90 120 150 180 210 Relative timeh f Fig. 3. Sequence of events for initiating data printout.546 BURNS et al. : AUTOMATED SIGNAL RECORDING SYSTEM Analyst, VoZ. 105 c1 3 4 2 1 +vcc C2 5 6 7 8 9 10 11 12 13 N C H Minus Common Minus Channel 7 x 150s1 0 =1N4148, 18 off Fig. 4. Channel identification board. In addition to arranging for the DVM to read the voltage only at certain times, it is necessary to arrange that these readings are passed on to the print-out system only when they have reached a steady value and when the printer is ready to accept new data, Further, it is necessary to arrange that each voltage printed is accompanied by a letter identifying the element (C, H or N) and another character (minus sign) to distinguish between sample and blank.I_ 1 -I Fig. 5. General layout of cams and microswitches. Once the above requirements had been established it was then considered whether they should be implemented by a microprocessor-based system or by hard-wired logic. Because in the present system the advantages of the more expensive microprocessor control (ie. , speed and flexibility) could not be exploited, a discrete logic system was chosen instead. The basic philosophy of the system having been described it is now appropriate to consider in detail the various parts of the system and how they are inter-related.June, 1980 Mechanical FOR THE PERKIN-ELMER 240 ELEMENTAL ANALYSER System Components 547 In the PE 240 CHN analyser the various steps in an analysis are under the central control of a group of 20 cam actuated microswitches driven by a constant-speed electric motor, which is itself energised intermittently as each step of the analysis is completed.The first stage in the conversion of the PE 240 was to fit an additional bank of four cams on to the end of the existing shaft beside the programme wheel (see Mechanical Construction). Three of these cams, one each for nitrogen, carbon and hydrogen, have two lobes. One lobe actuates its respective microswitch for about 22 s during the blank measuring period and the other lobe actuates the microswitch for about 30s during the sample measuring period.The fourth cam lifts for the total blank measuring duration, about 81 s. These four microswitches are wired to the print command circuitry and hence to the printer to initiate printout of the DVM output together with element and sample or blank identification, letters and characters (see Fig. 1). Fig. 6. Brass fitment. Electrical Printer The printer will not print the appropriate symbols with only a simple microswitch closure, it requires the incoming data to be in transistor - transistor logic (TTL) compatible four-line BCD format. To achieve this, the three element microswitches N, C and H were connected to a diode matrix, the output of which gives the necessary levels for feeding to the letters columns of the printer, a Pye-TMC Series 1008. This device prints out up to eight columns Fig.7. Nitrogen cam. Fig. 8. Carbon cam.548 BURNS et al. : AUTOMATED SIGNAL RECORDING SYSTEM Analyst, Vol. 105 of data, one line at a time, on to a paper strip 58 mm wide at a maximum rate of 2 lines s-1. Column 1 was used for a minus sign (to identify the blank measurements), columns 2 to 6 for the numerical value of the sample and columns 7 and 8 for alphabetic identification of the element. Data input for each of the eight columns is 4-line, TTL-compatible positive logic parallel BCD. The printer has a print inhibit facility which allows printout of data at a rate of one reading every 5 s, but only for the duration of the measurement phase of the PE 240.This is achieved by connecting the output of the print command circuitry (IC1, IC2) to the print- command circuitry in the printer. Connection is made via two Cannon 37-way D connectors. Fig. 9. Hydrogen cam. Fig. 10. Blank/read cam. Digital voltmeter In the digital voltmeter, an EXEL XL 2000, readings are normally repeated at a rate of 1.6 s-1. As already mentioned, it was necessary to modify this reading rate and the printer rate of 2 lines s-1 so that readings are taken only during the blank or the sample measuring periods, of duration 22 and 30 s, respectively, and at the rate of about 0.2 s-l, thus giving about five readings for each measurement phase. TABLE I MANUAL RESULTS ON ACETANILIDE Sample 1 ..2 .. 3 .. 4 .. 5 .. 6 . . 7 .. 8 .. 9 .. 10 .. Meanvalue .. Calculated value . . Standard deviation Nitrogen, % .. .. 10.35 .. .. 10.45 .. .. 10.31 .. .. 10.25 .. .. 10.24 .. .. 10.48 ,. .. 10.38 .. .. 10.36 .. .. 10.32 .. .. 10.40 Carbon, yo 71.16 71.07 71.07 71.15 71.10 71.07 71.09 70.99 71.05 71.15 .. .. 10.35 71.09 .. .. 10.36 71.09 . . .. 0.078 0.053 Hydrogen, yo 6.60 6.74 6.72 6.75 6.67 6.66 6.82 6.78 6.69 6.69 6.71 6.71 0.064 This is accomplished by use of the circuit shown in Fig. 2. IC1 is connected as a relaxation oscillator with two stable states. The output a.t pin 3 is high for about 5 s but then goes low for about 0.05 s. This output is taken to IC2b, a three-input NAND gate, to obtain logic inversion so that the output is normally low but goes high for 0.05 s once every 5 s.A second three-input NAND gate, IC2a, has its three inputs connected to the three micro- switches designated N, C and H in the PE 240. The output of this gate is normally low but goes high when nitrogen, carbon or hydrogen are being measured. The outputs of these twoJune, 2980 FOR THE PERKIN-ELMER 240 ELEMENTAL ANALYSER 549 gates are combined in a third three-input NAND gate IC2c, the output of which gives a 0.05-s duration low signal once every 5 s during the blank and the sample measuring periods. This signal is a print command to the printer (Fig. 3). The analogue millivolt output of the PE 240 goes direct to the x 100 pre-amplifier input of the EXEL DVM, which measures this voltage at a rate of 1.6 measurements s-l with a duration of measurement of 100 ms and displays it on a 44-digit 7-segment light emitting diode (LED) display.The circuits that drive the LED display also provide opto-coupler buffered BCD outputs which are connected to the printer. One problem which now arises is that the BCD data output of the DVM could change during a print cycle; this would, of course, lead to erroneous readings. To prevent this, use was made of two facilities, data hold and converter busy. Data hold. The Exel DVM has a hold facility which freezes displayed and outputted data when held low. The Pye - TMC printer has a data hold facility which goes low during the print cycle. When these two data hold facilities are connected together the desired result is obtained and the data are held steady for the duration of the print cycle.The Exel DVM has a facility which is at logic low only at the end of the transfer of data to stores ready for display and outputting. It returns to high at the commencement of the next DVM measuring cycle. If this facility is taken to the print command logic, IC2, the print command can be enabled only after a DVM measuring cycle has been completed. An additional facility for the convenience of the PE 240 user is function indication. Below the DVM are two LEDs. The left-hand one, a minus sign, is illuminated during measurement of the three blank elements while the right-hand, a 7-segment LED, is pro- grammed, via a &ode matrix, to indicate the component being measured, to light u p l z f o r carbon, / - / for hydrogen and 1-1 for nitrogen (Fig.4). Thus, the actual element being measured can be immediately identified, which is tedious to achieve with the normal pen recorder and impossible with the printer unless the paper strip is fed forward several rows. Converter bztsy. Sample 1 .. 2 .. 3 .. 4 .. 5 . . 6 .. 7 .. 8 .. 9 .. 10 .. TABLE I1 DVM - PRINTER RESULTS ON ACETANILIDE Nitrogen, % . . .. 10.34 . . .. 10.37 . . .. 10.38 . . .. 10.38 . . .. 10.39 . . . . 10.36 . . .. 10.40 .. . . 10.31 .. .. 10.34 . . .. 10.38 Carbon, % 71.04 71.02 71.21 71.21 71.03 71.02 71.04 71.13 71.21 71.03 Meanvalue . . .. .. 10.36 71.07 Calculated value . . . . . . 10.36 71.09 Standard deviation . . . . 0.028 0.086 Hydrogen, % 6.71 6.67 6.68 6.79 6.66 6.69 6.70 6.72 6.79 6.65 6.71 6.71 0.049 Mechanical Construction The mechanical construction of the accessory (for general layout, see Fig. 5) can be carried out by any small workshop.A brass fitment was constructed (Fig. 6) to be fitted on to the existing programmer index wheel and carry four 4 in thick Perspex cams with raised portions to trigger microswitches. Three of the cams were similar: those for nitrogen, for carbon and fqr hydrogen (Figs. 7, 8 and 9). The fourth cam, Fig. 10, was required to indicate whether blank or read signals were being produced. The first lobe on the three similar cams was equivalent to 22 s, i e . , a sector of ll", the second lobe was equivalent to 30 s long (15") and separated from the first lobe by 266 s (133"), 276 s (138") and 286 s (143") for the nitrogen, carbon and hydrogen signals, respectively. The signal lobe on the fourth cam was made 81 s (40.5") long.All the raised lobes on the cams were $in high.550 BURNS, McKNIGHT, QUIGG AND SWINDALL When the four cams were assembled and locked in the appropriate positions, with the microswitches bearing on them, fine adjustment of the length of the lobes was carried out with a small file to ensure that as one microswitch went off another came on. Sample 1 .. 2 .. 3 .. 4 .. 5 .. 6 .. 7 .. 8 .. 9 .. 10 .. TABLE I11 1972 MANUAL RESULTS ON ACETANILIDE Nitrogen, yo .. .. 10.32 .. .. 10.34 .. .. 10.30, .. .. 10.34 .. .. 10.42 .. .. 10.32 .. .. 10.40 .. .. 10.37 .. .. 10.39 .. .. 10.40 Carbon, % 71.01 71.11 71.12 71.12 71.12 71.04 71.07 71.17 71.08 71.05 Hydrogen, yo 6.71 6.71 6.72 6.68 6.71 6.71 6.72 6.71 6.74 6.72 Mean value .... .. 10.37 71.09 6.71 Calculated value . . . . .. 10.36 71.09 6.71 Standard deviation . . .. 0.036 0.048 0.015 Results and Conclusions The device has been in routine daily use for 2 years and has required no servicing attention. Tables 1-111 show results obtained manually, others obtained by using the data handling system and some results obtained manually 7 years ago for 10 samples in sequence using Hopkin and Williams Micro-Analytical Standard acetanilide as the reference compound. Table IV gives results on BCR compound 34, bis(diethy1tin chloride) oxide, which has a low carbon value. The results are seen to be equivalent in practice and, as expected, fall well within the generally accepted spread of &0.3yO.l It is interesting that the results for 1979 are almost as good as those obtained in 1972 when the PE 240 and the Beckmann LM 600 electro- balance were 1 year old. TABLE IV DVM -.PRINTER RESULTS ON BCR 34 [BIS(DIETHYLTIN CHLORIDE) OXIDE] Sample Carbon, yo Hydrogen, yo 1 .. .. .. 21.77 4.61 2 .. .. .. 21.84 4.54 3 .. .. .. 21.69 4.59 4 .. .. .. 21.69 4.54 5 .. .. .. 21.74 4.60 Mean value . . .. .. 21.75 4.58 Certified value . . .. . . 21.81 f 0.08 4.55 f 0.06 Standard deviation . . .. 0.063 0.034 The main advantages of the device are reduced operator attention and strain, and it is not necessary to examine chart recordings in detail or worry in case the chart pen is not functioning. Perhaps the major advantage, however, is that the analyser can be left unattended after insertion of the sample into the furnace because it is not necessary to use the attenuation switches. The wide range of the DVM readily copes with most sample responses. It is hoped in the future to link a microcomputer to the analyser to obtain results expressed as a percentage composition. Reference 1. Cottrell, M. R., and Cottrell, F. H., in Belcher, R., Editor, “Instrumental Organic Elemental Received July 4th, 1979 Accepted January 3rd, 1980 Analysis,’’ Academic Press, London, 1978, p. 47.

ISSN:0003-2654    

DOI:10.1039/AN9800500544

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

Analyst, June, 1980, Vol. 105, pp. 551-557 551 Determination of Cobalt in Blood R. A. Barfoot Department of Chemical Pathology, Wanstead Hospital, Hermon Road, London, E 11 1PA and J. G. Pritchard Department of Chemistry, North East London Polytechnic, Romford Road, London, E l 5 4L.Z Blood and serum have been wet oxidised with a nitric acid - perchloric acid - sulphuric acid mixture, and the cobalt has been extracted from the aqueous ash with 1-nitroso-2-naphthol in chloroform. Cobalt-57 tracer studies showed an over-all recovery of 95% of the cobalt. Carbon rod atomic-absorption spectrophotometry showed the following likely ranges for cobalt in healthy adult human beings: blood, 0.20-0.28 pg per 100 cm3; and serum, 0.12- 0.20 pg per 100cm3. The coefficient of variation of the entire analytical procedure is f 3-4%.The wide discrepancies between literature values for blood cobalt and details of the present method are discussed. Keywords : Blood analysis ; se‘yztm analysis ; cobalt determination ; spectro- photometry; 1-nitroso-2-naphthol Cobalt is a vital trace element that is required in the normal diet of man in the form of the cyanocobalamin complex, vitamin B12. This vitamin is essential to deoxyribonucleic acid synthesis and propionate metabolism, and to the avoidance and control of pernicious anaemia.l In cases of iron-deficiency anaemia, inorganic cobalt may be added to the diet (with iron) to increase the rate of haemogolobin s y n t h e ~ i s . ~ ~ ~ However, cobalt in doses of about 5 mg per day can be toxic to man and may, in the long term, cause heart disease when an adequate diet of normal food is not taken.4 Hence, the correct determination of cobalt in the blood is of obvious importance in human welfare.Over 24 papers have been published on the determination of cobalt in either whole blood, plasma or serum, usually in connection with medical studies in which an observed difference in the cobalt level has been usefully attributed to a clinical condition. However, the reports of the absolute values for normal blood cobalt have varied widely from study to study (Table I). The discrepancies do not appear to be generally associated with any particular method. Neutron-activation analysis has been the most popular method more recently, but the results still vary by a factor of greater than 100.Medical workers have commented that these widely different results cannot arise from variation in the natural blood cobalt levels of the subjects, but must be due mainly to analytical methodologica 1 problems.lO9 26931 ~ ~ ~ 9 3 ~ 9 ~ 5 It appears to us that the unexpectedly large variation observed in the independent determinations listed in Table I arises primarily from the difficulties caused by an actual parts per billion (lo9) level of cobalt in the blood. By carefully checking all possible sources of error, we attempt here to provide a standard method for the determination of cobalt in blood that should be reproducible in other labora- tories and give a correct answer. For the present study, we selected an atomic-absorption spectrophotometer for the measurement of amounts because its calibration and reproducibility are usually very reliable, and most laboratories possess this instrument nowadays.Our Varian AA5 atomic-absorption spectrophotometer permitted the detection of 0.2 pg of cobalt per 100 cm3 at a signal level equivalent to twice the noise level, when a 5-pl aqueous standard cobalt solution was atomised in the hollow carbon rod furnace. This corresponds to a limit of detection of 10-ll g (cf., 7.5 x 10-l2g first quoted for this method by L’vov~~). Preliminary experiments, with multiple additions of 5-p1 aliquots of whole blood and serum in the carbon rod, showed that the natural level of cobalt was only just detectable by this direct method and corresponded most probably to a value near 0.1 pg per 100 cm3.However, only poor precision could obviously be expected with this low-level instrument response, particularly as sample losses tended to occur through frothing of the viscous natural material, and substantial interference552 BARFOOT AND PRITCHARD : DETERMINATION Analyst, Vol. 105 was also observed owing to non-atomic absorption. We have therefore developed a method for the extraction and concentration of cobalt from whole blood or serum to give a concentra- tion of cobalt near 100 pg per 100 cm3, which is suitable for precise determination with the spectrophotometer. Thus, the recovery of radioactive ~ 0 b a l t ~ 0 , ~ ~ s ~ ~ from 57Co-labellecl vitamin B,, in serum was determined after decomposition by a selected wet-oxidation procedure.19937s38 Then the recovery of cobalt was determined after the ash produced by the wet oxidation was taken up in aqueous solution and the cobalt specifically extracted though its l-nitroso-2- naphthoate complex into chloroform (Stary's rr1ethod~~9*~) to give a concentration of cobalt suitable for determination by atomic absorption.As far as we are aware, Stary's method for cobalt has not hitherto been used for the extraction of blood ashes. TABLE I ESTIMATES OF THE AMOUNT OF COBALT IN NORMAL HUMAN WHOLE BLOOD, PLASMA OR SERUM Analytical technique Polarography5 . . Arc spectroscopy6 . . Arc spectroscopy7 . . Spark spectroscopy8 Spectroscopye . . Arc spectroscopylO . . Arc spectroscopylO . . Colorimetryl*JS . . Colorimetryll .. Colorimetry4 . . Colorimetry14 ..Colorimetry16 . . Colorimetry16 . . Atomic absorption6 Atomic absorption17 Atomic absorption18 Atomic absorptionle Atomic absorption20 Atomic a b ~ o r p t i o n ~ l - ~ ~ Neutron activationls Neutron activation25 Neutron activationas Neutron a c t i v a t i ~ n ~ ~ ~ ~ ~ Neutron activationa9 Neutron activation30 Neutron activation31 Neutron activation32 Neutron activation33 .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . . .. .. ,. . . . . .. .. .. . . .. . . . . . . .. I . . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. Date 1952 1974 1964 1968 1956 1955 1955 1953 1957 1968 1963 1958 1960 1974 1970 1979 1975 1961 1977 1979 1956 1972 1973 1978 1964 1976 1978 1976 Sample Whole blood Plasma Whole blood Whole blood Whole blood Whole blood Serum Serum Whole blood Serum Whole blood Whole blood Whole blood Plasma Serum Whole blood Serum Whole blood Whole blood Whole blood Whole blood Serum Whole blood Whole blood Serum Serum Serum Serum Cobalt level or range (pg per 100 cm8) 0.85 6.1 0.1 0.048 0.018 5 0.0072 58 37 6-12 7 4.78 1.0 0.031 1.5 110 0.8-8.6 0.56-0.98 0.4 0.11 1.0-8.0 1.22 0.18 0.06-0.18 0.02-0.13 0.008-0.058 0.003 9-0.027 0.0038-0.026 0.003 Experimental A.Materials Cyanocobalamin BP, 250 pg ~ m - ~ in aqueous solution containing 0.90% of benzyl alcohol as preservative, was obtained from Paines and Byrne Ltd. Cyanocobalamin BP labelled with cobalt-57 was obtained as a 94 n g ~ r n - ~ aqueous solution with a y-ray activity of 10 pCi cm-3 from the Radiochemical Centre, Amersham, and was diluted for use in the experiments.l-Nitroso-2-naphthol and UltraR-grade perchloric acid (sp, gr. 1.54) were obtained from Hopkin & Williams, and AnalaR-grade nitric acid (sp. gr. 1.42), AnalaR- grade sulphuric acid (sp. gr. 1.84) and 1000 p.p.m. cobalt(I1) chloride standard solution in 1 M hydrochloric acid from BDH Chemicals. Blood was withdrawn from human patients by means of disposable polystyrene syringes with rubber plungers and stainless-steel needles. The blood was transferred into poly- styrene tubes (with polyethene stoppers), allowed to clot for 30 min and then centrifuged in the same tubes. Serum was separated and stored in polystyrene tubes at -20 "C.June, 1980 OF COBALT IN BLOOD 553 B. Recovery of Cobalt from Wet Oxidation of Serum Into each of six 50-cm3 Pyrex glass conical flasks were placed human serum (10 cm3), [57Co]vitamin-Bl, solution of 1 pCi ~ m - ~ y-ray activity (0.30 cm3), nitric acid (sp.gr. 1.42) (20 cm3), perchloric acid (sp. gr. 1.54) (4 cm3) and sulphuIic acid (sp. gr. 1.84) diluted to 1 + 4 V/V with water (0.4 cm3). The flasks were allowed to stand at room temperature for 15min to allow the digestion to proceed in a controlled manner. Then the flasks were heated on a hot-plate for about 10min until the contents stopped effervescing and then further for about 20 min at about 200 "C until a white precipitate began to form. Final heating to dryness was carried out over a luminous Bunsen burner flame. The resulting ash for each flask in turn was dissolved in 4 M hydrochloric acid, transferred quantitatively into a Sarstedt polystyrene counting tube and made up to the 5-cm3 mark with water.A standard was prepared from 0.30 cm3 of the 1 pCi CM-~ [57Co]vitamin-Bl, solution made up similarly to 5 cm3 in a counting tube of identical geometry. The counting of y-ray radioactivity was carried out with a Nucleonic well-type scintillation counter employing a thallium-activated sodium iodide crystal, in conjunction with a Panax PX STNI scaler and a PX AAUI pulse-height analyser. The 0.123-MeV emission from cobalt-57 was counted with the operating EHT voltage on a plateau a t 1250 V and the discrimination bias voltage of the pulse-height analyser at a peak value of 56 mV. The counts per minute value minus the background (about 100) for the standard [57Co]vitamin- B,, solution was 38344.The net counts for the six test samples are reported as percentages of the standard (Table 11). TABLE I1 PERCENTAGE RECOVERIES OF COBALT Recovery of radioactive cobalt from acid digestion of vitamin B,, in serum Recovery of cobalt (atomic-absorption signal) after chloroform extraction of aqueous standard solution 99.8 96.8 99.8 93.6 100.0 96.2 99.2 98.1 99.6 100.2 Recovery of radioactive cobalt from acid digestion of vitamin B,, in serum followed by chloroform extraction 93.2 95.2 94.5 95.3 94.0 96.5 C. Recovery of Vitamin BIZ and Cobalt from Polystyrene Containers [57C~]Vitamin-B1, at a concentration of 0.3 ng ~ m - ~ and y-ray activity 0.03 pCi ~ r n - ~ in aqueous solution at pH 7 (the pH of blood) was allowed to stand in a 13-cm3 polystyrene tube for up to 28 days.The radioactivity of the tube and contents was counted, then the contents were drained completely into an unused identical tube and the radioactivity was again counted. This experiment was repeated with a neutralised solution of inorganic cobalt-57 derived by heating [57Co]vitamin-B,, with 0.1 M sodium hydroxide solution. D. Recovery of Cobalt from Extraction of Cobalt( 111) with l-Nitroso-2-naphthol De-ionised water (5 cm3), standard 0.2 p.p.m. cobalt(I1) solution (0.200 cm3) and 100- volume hydrogen peroxide (0.05 cm3) were mixed in a 10-cm3 stoppered glass tube. The pH was adjusted to about 4 with 0.1 M hydrochloric acid, a 1% solution of l-nitroso-2- naphthol in glacial acetic acid (0.2 cm3) was added and the solution was allowed to stand for 30 min at room temperature to allow the formation of the cobalt complex.The solution was extracted with l-cm3 volumes of chloroform and the combined extracts were evaporated to dryness in a 5-cm3 tube. The residue was dissolved in 0.200 cm3 of chloroform and 5-p.1 samples of this solution were examined by atomic-absorption spectrophotometry. The mean recorder displacement for four separate 5-pl samples of the chloroform extract was taken for each of four experiments on this recovery. From each of these mean values was deducted the mean of four atomisations of 5-pl samples of a blank extract derived from the above procedure, except that distilled water was substituted for the standard cobalt solution.654 BARFOOT AND PRITCHARD : DETERMINATION Analyst, Vol.105 The net recorder displacements were compared with the mean displacement obtained from four atomisations of 5-pI samples of the standard aqueous cobalt solution, and the recovery ratios are listed as percentages in Table 11. The 5-pl samples were measured into the carbon rod furnace attachment to the AA5 spectrophotometer by a microlitre auto-pipette fitted with disposable tips, which were specially bent at the ends to prevent inadvertent escape of particularly the more mobile organic solvent during the manipulations of transfer. The following optimum voltage settings on the ashing unit were found to avoid almost completely any non-atomic contri- bution to the absorption (as tested by use of a hydrogen continuum lamp) and to give symmetrical peaks on the chart recorder: drying period at 5 V, 15 s; ashing period at 7.5 V, 20 s; and atomisation at 8 V, 2 s.The scale of the Kipp and Sohnen recorder was adjusted so that measurements for the samples were made at about 60-70% of full-scale deflection, which required a response time of less than 0.5 s. Linearity of the absorbance at 240.6 nm with cobalt concentration was checked for the 0.014.1 and 0.1-0.4 p.p.m. ranges with aqueous standards. E. Recovery of Cobalt from Combined Wet Oxidation and Extraction Six samples of [57Co]vitamin-Bl, solution were treated with serum and acids as described in B above. The ash was dissolved in 4~ hydrochloric acid (5 cm3) and extracted as described in D, the combined chloroform extracts being made up to 5 cm3 with water and compared with a 5-cm3 aqueous solution containing an amount of [57Co]vitamin-Bl, solution identical with that introduced into each of the six recovery test solutions.F. Cobalt Determinations on Blood and Serum Blood samples (25 cm3) were obtained by venepuncture as described in A above from the two groups of human subjects listed in Table 111. A 2-cm3 volume of the blood of each of the five hospital workers was added, before coagulation, to the acid-digestion mixture to constitute a lO-cm3 sample of pooled blood for analysis. The total of 45 samples of blood were otherwise allowed to stand and clot for 30 min at room temperature. The serum was separated by centrifugation, and that from the 40 out-patients was pooled. TABLE TI1 COBALT DETERMINATIONS ON SERUM AND WHOLE BLOOD Number of Standard No.of Pooling of Blood or determina- Cobalt level/ errorlpg patients Type of patients samples serum tions pg per 100 cm* per 100 cma 40 Hospital out-patients Pooled Serum 10 0.149 f 0.002 f 0.009 6 Healthy laboratory Healthy laboratory 6 0.162 workers Not pooled Serum 5 - workers Pooled Blood 1 0.24 Samples of 10 cm3 of blood or serum were treated by the acid-digestion procedure described in €3 above, but omitting the radioactive tracer. The resulting 5-cm3 solutions of cobalt were treated with 0.05 cm3 of 100-volume hydrogen peroxide and extracted with l-nitroso-2- naphthol following procedure D, above, to produce 0.2-cm3 solutions of cobalt concentrated from the original 10-cm3 samples of blood or serum. Then, 5-pl samples of these O.2-cm3 solutions were compared with 5-pl aqueous standard cobalt solutions by the carbon furnace atomic-absorption technique , as described in D above.Separate calibrations were made for whole blood and for serum so as to utilise almost full-scale recordings for both. A hydrogen continuum lamp was used to determine whether any correction was required due to non- atomic absorption (less than 1%). De-ionised water was put through procedures B and D to provide a chloroform extract for use as the reagent blank, the signal from which (about 12%) was deducted from that observed for the blood sample extracts. The thus corrected signals from the chloroform extracts divided by the signal per unit concentra-June, 1980 OF COBALT IN BLOOD 555 tion of cobalt from the original aqueous cobalt standards were further corrected by the factors 0.02 and 100/95 (see, Results and Discussion) to take account of the concentration procedure and the over-all recovery of cobalt, respectively.De-ionised water was put through the blood-taking procedure described above and then formally analysed for cobalt, as above, to check for contamination of the solutions used in the analytical procedure owing to possible loss of cobalt from the stainless-steel needles or from the walls of the syringe and other vessels. The signal from the chloroform extract obtained here was compared with that from the reagent blank (identical). Results and Discussion Recovery Tests The recovery of cobalt-57 from the digestion of vitamin R,, in human serum with a nitric acid - perchloric acid - sulphuric acid mixture in glass vessels was 99.77 If: 0.140/,(standard error) (Table 11).The probability that this result is significantly different from 100.OO~o is at the 92% point (one side) of the $-distribution, and the result is certainly not significantly different from 99.9%. Therefore, it is unlikely that any appreciable amount of cobalt is lost to the glass walls of the reaction vessels or otherwise during the digestion. The recovery of cobalt in the extraction of its aqueous solution with 1-nitroso-2-naphthol in chloroform, determined on 5-p1 samples by atomic-absorption spectrophotometry, is 96.18 0.95y0 (standard error). The experiments here tested a combination of possible losses of cobalt due to (a) incomplete extraction, (b) absorption into the walls of vessels and (c) any differences in behaviour of the chloroform extract on atomisation compared with an aqueous solution of the same concentration in cobalt.The recovery of cobalt-57 from both the digestion procedure carried out on vitamin B,, in serum and the extraction of the cobalt 1-nitroso-2-naphthoate into chloroform is 94.78 & 0.47% (standard error). As a loss of cobalt greater than 0.1:k in the digestion alone is most unlikely, as shown above, we conclude that the extraction procedure causes a loss of, say, 100-94.9% due to causes (a) and (b) above. Now, the difference between the mean values 96.18% and 94.78% is not significant, at the 82% point (two sides) of the t-distribu- tion; and as the former mean is greater than the latter, we conclude that no additional apparent loss is introduced from cause ( c ) above.Moreover, this may introduce an apparent gain of about 1%. In the determination of cobalt by our procedure, therefore, a correction is required due to a deficiency in the recovery of 5%. Determination of Cobalt Determination of the level of cobalt in the serum of pooled samples derived from a larger number of hospital out-patients resulted in 0.149 pg per 100 cm3 as the mean of 10 replicate complete assays on the sample pool (Table 111). The standard deviation was 0.0056 pg per 100 cm3 and the coefficient of variation &3.7y0. This is very satisfactory, considering the complexity of the procedure, and allows the mean value to be quoted with confidence to two decimal places.This value is well within the general scatter of results obtained for our sample of five laboratory workers, the individual results being 0.129, 0.164, 0.173, 0.181 and 0.164 pg per 100 cm3. We therefore assume that there is no significant difference between the serum cobalt of hospital out-patients and healthy laboratory workers, although the scatter in their values amounts to a coefficient of variation of k12-130A. Hence, we can quote our mean values for the cobalt level in human serum and whole blood as 0.16 and 0.24 pg per 100 cm3, respectively, and the corresponding expected ranges at the 99% confidence level are 0.12-0.20 and 0.20-0.28 pg per 100 cm3 for healthy adults. Precautions and Comments The recovery of cobalt in the two main stages of our procedure has been carefully checked, as detailed above.Further, no losses of cobalt were found during storage of either vitamin B,, or inorganic cobalt in the polystyrene vessels used for the storage of blood and serum, and the syringes and other vessels used in the handling of the blood samples were shown to introduce no errors into the analysis. Whilst it is not our purpose to explain in detail the556 BARFOOT AND PRITCHARD : DETERMINATION Amlyst, Vol. 105 wide discrepancies in the reports listed in Table I, it is generally evident that parts per billion amounts of cobalt may, under appropriate circumstances, disappear from solution into the walls of containing vesselsJ41 or cobalt, may appear in solution as a result of impure reagents and contact with rubber, plastics or glass.42 Thus, we have observed a high reagent blank of about 12% of the total signal, which we attribute to cobalt in the l-riitroso-2- naphthol, which, in the course of its preparation, storage and use, acts as a scavenger for cobalt.We suggest that we may have been fortunate in the particular batches of our reagent, and that batches with much higher cobalt levels might be expected on occasion. We have avoided working with blood plasma, the preparation of which requires the use of anti- coagulants that may contain cobalt. In the atomic-absorption technique, the use of the hydrogen continuum lamp to test for signals due to non-atomic absorption is absolutely essential. We found that the ashing procedure required modification until the signals generated at the cobalt wavelength repro- ducibly contained no component (or essentially none) due to non-atomic absorption, otherwise gross errors could arise. The use of the Varian Associates non-por~us~~ high-purity graphite tubes coated internally with pyrolytic graphite for the ashing and pyrolysis of the test samples gave excellent reproducibility, even with 5-p1 samples, and the three-stage heating procedure ensured essentially no difference between signals from chloroform solutions of organic cobalt and from aqueous solutions of inorganic cobalt of the same concentration.The use of nitric acid and sulphuric acid is essential in order to eliminate the otherwise explosive tendency of the perchloric acid used in the digestion r n i ~ t u r e .* ~ * ~ ~ Hydrogen peroxide was used in the extraction to oxidise cobalt(1I) to cobalt(II1) and to eliminate iron(I1) from the l-nitroso-2-naphthol extract, because iron might interfere with the atornic- absorption signal from cobalt. The use of acetylacetone as the extraction reagent, also advocated by star^,^^ was found to be inefficient compared with l-nitroso-2-naphthol. The previous result, which agrees reasonably with our own, 0.18 pg per 100 cm2 of cobalt in serum, was obtained by one of the groups using the neutron-activation technique (Table I) .26 The closest atomic-absorption value, 0.1 1 pg per 100 cm3 of cobalt in whole blood, was obtained by a large group of Japanese workers in Mito, who extracted the dissolved ash from blood with trioctylamine in chloroform and then back-extracted with water prior to electro- thermal atomic-absorption determination of cobalt .22 They reported a sensitivity of 0.01 pg per 100 cm3, a scatter of &lo:(, and a recovery of 97%.An alternative extraction with trioctylmethylammonium chloride has been described by Japanese workers from another laboratory (Kyoto), who recorded the same value for blood cobalt as their colleagues in M i t ~ . ~ * The recent literature records considerable current interest in the use of complexing agents in the determination of cobalt in a variety of material~.~~-~O We hope that the simple method described here may be more widely applied in the future and that attempts will be made to standardise a procedure for cobalt in blood in several laboratories.The met?od described here determines total blood or serum cobalt, which consists of cobalt in ionic form plus organic cobalt in the form of several cobalaniin complexes, one of which is vitamin B12. A reasonable correlation has been claimed (I = 0.8) between the total cobalt and the vitamin B,, content of human serum.51 However, for diagnostic purposes, this vitamin has traditionally been determined in serum by the radioisotopic dilution method.52 The value of this method as a diagnostic tool for pernicious anaemia has recently been questioned, and it appears that the microbiological assay53 of cyanocobalamin activity is probably to be preferred for this a p p l i ~ a t i o n . ~ ~ - ~ ~ Hence, whilst there appears no reason why the determination of total cobalt should not be used to monitor the administra- tion of inorganic cobalt to patients during the treatment of iron-deficiency anaemia, the total cobalt figure may be of only limited use in the diagnosis and treatment of pernicious anaemia.References 1. 2. 3. 4. 5. 6. Smith, E. L., Nature (London), 1948, 161, 638. Harp, M. J., and Scoular, F. I., J . Nutr., 1952, 47, 67. Rohn, R. J., and Bond, W. H., J.-Lancet, 1953, 73, 317. Kesteloot, H., Roelandt, J., Willems, J., Claes, J. H., and Joossens, J. V., Circulation, 1968, 37, 854. Heyrousky, A., Cas. Lek. Cesk., 1952, 91, 680. Maessen, F. J. M. J., Pasrna, F. D., and Balke, J.Y Anal. CKem., 1974, 46, 1445.June, 1980 OF COBALT IN BLOOD 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. Butt, E. M., Nusbaum, R. E., Gilmour, T. C., Didio, S. L., and Sister Mariano, Arch. Environ. Kabiel, A. M., Hanna, 2. G., and Youssif, Y. S., Appl. Spectrosc., 1968, 22, 183. Smoczkiewicz, A., and Mizzalsky, W., Bull. Soc. Amis. Sci. Lett. Poznan, Ser. C., 1956, 8, 51. Thiers, R. E., Williams, J . F., and Yoe, J . H., Anal. Chem., 1955, 27, 1725. Nugmanova, R. N., Musabaev, I. K., Guseva, D. M., and Mukhamedova, I. G., Uzbeksk. Khim. Zh., Pavlova, A. K., Vestsi Akad. Navuk B . SSR, Ser. Biyal. Navuk, 1958, No. 2, 93. Sushko, E. P., Sb. Nauchn. Rub. Inst. Med. Minsk., 1957, 19, 269. Zhukovskaya, E. D., and Idelson, L. I., Lab. Delo, 1963, 9, 16.Ishibaski, M., Igaku To Seibutsuguku, 1958, 47, 50. Haerdi, W., Vogel, J., Monnier, D., and Wagner, P. E., Helv. Chim. Acta, 1960, 43, 869. Welz, B., and Weideking, E., 2. Anal. Chem., 1970, 252, 111. Ward, N. I., Stephens, R., and Ryan, D. E., Anal. Chim. Acta, 1979, 110, 9. Mazzarelli, R. A. A., and Rocchetti, R., Talanta, 1975, 22, 683; 1977, 24, 77. Delves, H. T., Shepherd, G., and Vinter, P., Analyst, 1971, 96, 260. Ishizaki, M., Neno, S., Oyamada, M., Kataoka, F., Murakami, R., Kubota, K., Katsumura, K., and Kataoka, F., Ishizaki, M., Neno, S., Oyamada, N., Murakami, R., Kubota, K., and Katsumura, K., Ishizaki, M., Oyamada, N., Fujiki, M., and Yamaguchi, S., Sangyo Igaku, 1978, 20, 174. Nishuira, S., and Nagada, S., Kyoto-fu Eisei Kogai Kenkyusho Nempo, 1977, 22, 48.Koch, H. J., Smith, E. R., Shimp, N. F., and Connor, J., Cancer (New York), 1956, 9, 499. Kasperek, K., Schicka, H., Siller, V., and Feinendegen, L. E., Strahlentherapie, 1972, 143, 468. Curtis, J . R., Goode, G. C., Herrington, J., and Urdaneta, L. E., Clin. Nephrol., 1976, 5, 61. Coleman. R. F., Herrington, J., and Scales, J. T., BY. Med. J., 1973, 1, 527. Clementi, G. F., Rossi, L. C., and Santoroni, G. P., in “Nuclear Activation Techniques in the Life Pan, R. M., and Taylor, D. M., Biochem. J . , 1964, 91, 424. Versieck, J., Hoste, J., Barbier, F., and Vanballenberghe, L., Lancet, 1976, i, 1403. Versieck, J., Hoste, J., Barbier, F., Steyaert, H., DeRudder, J., and Michels, H., CZin. Chsm., 1978, Lins, L. E., and Pehrsson, K., Lancet, 1976, i, 1191.Underwood, E. J ., “Trace Elements in Human and Animal Nutrition,” Fourth Edition, Academic Schroeder, H. A., Nason, A. P., and Tipton, I. H., J . Chronic Dis., 1967, 20, 869. L’vov, B. V., Sfiectrochim. Acta, 1961, 17, 761; 1969, 24B, 53. Gorsuch, T. T., Analyst, 1959, 84, 135. Doshi, G. R., Sreekumaren. C., Mulay, C. D., and Patel, B., Curr. Sci., 1969, 9, 206. Stary, J., “Solvent Extraction of Metal Chelates,” Pergamon Press, Oxford, 1964, pp. 112 and 183. Jago, J., Wilson, P. E., and Lee, B. M., Analyst, 1971, 96, 349. Robertson, D. E., Anal. Chim. Acta, 1968, 42, 533. Robertson, D. E., Anal. Chem., 1968, 40, 1067. Cresser, M. S., “Solvent Extraction in Flame Spectroscopic Analysis,’’ Buttenvorths, London, 1978, Gorsuch, T. T., “Destruction of Organic Matter,’’ Pergamon Press, Oxford, 1970, pp. 22 and 141. Analytical Methods Committee, Analyst, 1959, 84, 214. Volgin, V. I., and Berezkina, L. M., Byull. Vses. Nauchno-Issled. Inst. Razved. Genet. Skokh. Zhivotn., Singh, R. B., Ray, H. L., Garg, B. C., and Singh, R. P., Talanta, 1979, 26, 898. Kulshreshtha, H., Singh, R. B., and Singh, R. P., Analyst, 1979, 104, 572. Wunsch, G., Talanta, 1979, 26, 177. Armannsson, H., Aszal. Chim. Acta, 1979, 110, 21. Mukhamyedova, I. G., and Mallina, E. R., Med. Zh. Uzb., 1966, 33. Baraket, R. M., and Ekins, R. P., Blood, 1963, 21, 70. Hutner, S. H., Bach, M. K., and Ross, G. I. M., J . Protozool., 1956, 3, 101. Kolhouse, J. F., Kondo, H., Allen, N. C., Podell, E., and Allen, R. H., N . Engl. J . Med., 1978, 299, Cooper, B. A., and Whitehead, V. M., N . Engl. J . Med., 1978, 299, 816. Donaldson, R. M., Jr., N . Engl. J . Med., 1978, 299. 827. Health, 1964, 8, 52. 1953, 7, 20. Noda, M., Ibaraki-Ken Eisei Kenkyusho Nempo, 1977, 15, 71. Ibaraki-Ken Eisei Kenkyusho Nempo, 1977, 15, 65. Sciences, 1978,” International Atomic Energy Agency, Vienna, 1979, p. 527. 24, 303. Press, London, 1977, p. 132. p. 30. 1978, Part 28, 34. 785. Received November 19th, 1979 Accepted November 29th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500551

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

558 Analyst, June, 1980, VoZ. 105, $9. 558-563 Spectrophotometric Determination of Arsenic and Antimony by the Silver Diethyldithiocarbamate Method R. H. Merry and B. A. Zarcinas CSIRO Divisiom of Soils, Glem Osmond, South Australia, 5064 When sodium tetrahydroborate(II1) is used for hydride generation in the spectrophotometric determination of arsenic as a complex with silver di- ethyldithiocarbamate, significant antimony interference may occur. This can be overcome by reading the absorbance at a wavelength other than the absorption maximum or both elements can be determined by reading at two wavelengths. The use of oxidising acids for the digestion of sediment, soil and plant material for arsenic and other metals is not suitable for antimony. This can be overcome by adding a reducing agent in the later stages of digestion and allows both arsenic and antimony to be determined simultaneously on the same digest.Keywords : Arsenic determination ; antimony determination ; spectrophoto- metry ; silver diethyldithiocarbamate The use of sodium tetrahydroborate(lI1) for the generation of arsine prior to the spectro- photometric determination of arsenic has been rep0rted.l We have satisfactorily used an almost identical method over a period of years for the analysis of digests and extracts of sediments, soils and plant materials. Some of these samples were obtained from an area surrounding a lead - zinc smelter complex2 where, amongst other elements, both aTsenic and antimony are polluting elements. A number of reports, including those by Vasgk and Sedi~ec,~ Dubois et al.,4 the Analytical Methods Committee6 and Dal Cortivo et aLJ6 give varying accounts of the extent of interference by stibine when silver diethyldithiocarbamate (AgDDC) is used and whether arsenic and antimony can be measured simultaneously.These workers mostly used a zinc - hydrochloric acid reducing system for hydride generation. This paper reports some findings for a system using sodium tetrahydroborate(II1) for simultaneous arsine and stibine generation, with subsequent spectrophotometric deter- minations using AgDDC in pyridine. Suitable digestion procedures for contaminated sedi- ment, soil and plant materials are also discussed. Experimental Apparatus The hydride generation cell resembled that of Thompson and Thomerson7 with an outlet connected, via a lead acetate trap, to the AgDDC collecting solution in a test-tube. The generation cell was continuously flushed with high-purity nitrogen. A series of six cells was used routinely.Absorbances were measured on Pye Unicam SP6.300 or SP800 spectrophotometers using 10-mm path length standard and 10- or 20-mm path lenffth flow- through cells. Reagents AgDDC solution. A 0.5% m/V solution in pyridine that had been twice distilled was used. Sodium tetra~ydroborate(III) solution. A 5% m/V solution of the reagent grade material, was prepared in NN-dimethylformamide with constant stirring. Arsenic and antimony standards. A range of standards from 0 to 20 pg in 2 N hydro- chloric acid was prepared using arsenic(II1) oxide, potassium antimony(II1) oxide tartrate and antimony(V) chloride.The AgDDC was prepared as described by Powers et aZ.,* but was freeze-dried.MERRY AND ZARCINAS 559 Method The recommended procedure used in this study for the digestion of soils and sediments for spectrophotometric measurements of both arsenic and antimony involved an aqua regia digest (3 + 1 hydrochloric acid - nitric acid) made up to volume with 2 N hydrochloric acid (see Appendix). For plant materials, a sulphuric acid - nitric acid - perchloric acid digestion was used, and the solution was also made up to volume with 2 N hydrochloric acid. For hydride generation, 1 ml of sodium tetrahydroborate( 111) solution was released from an automatic pipette through a side-arm into the standard solution or acidified unknown sample held in the generation cell.Release was gradual over a period of 5 min in order to prevent too rapid evolution of gas [solid sodium tetrahydroborate(II1) was found to react too vigorously for this system]. The cell was then flushed with an increased rate of nitrogen for at least 5 min, by which time full colour development occurred. A 5-ml volume of AgDDC in pyridine was used as the collecting solution in each determination. Sodium tetrahydroborate(II1) in NN-dimethylformamide retains a strong reducing capability for periods in excess of 24 h. This compares with about 1 h for aqueous or alkaline solutions (comparable to the results of Rooneyg and Duncan and ParkerlO), although longer periods have been reported by Smith et aZ.ll and Aggett and Aspell12 after purification procedures. Absorbance measurements were found to be stable over a period of at least 5 h.The wavelengths used are discussed below. The solution of the simultaneous equations used in determining both arsenic and antimony followed one of the sDecial cases described bv Meehanl3 where one of the element complexes does not absorb lengths. atbne of two selected wavl- m 0, 1 6 - X 5 12 e 8 - 2 Q 4 - - .- > .- + \ L \ 0 0 ' - \ I I .I 450 500 550 600 Wavelength/nm Fig. 1. Molar absorptivity of A, arsenic and B, antimony complexes as a function of wavelength. Results and Discussion Molar Absorptivity of Arsenic and Antimony Complexes Molar absorptivities (6) for the arsenic and antimony complexes with AgDDC were calcu- lated for wavelengths from 485 to 600nm, and are shown in Fig.1. Absorption maxima occur at about 530 and 504nm for arsenic and antimony, respectively, for this batch of AgDDC. The maximum E value of about 1.5 x lo4 1 mol-l cm-l for the arsenic complex compares with 1.3 x lo4 1 mol-l cm-l quoted by Jackwerth.14 The maximum molar absorptivity for the antimony complex at 504 nm is approximately 2.5 x lo4 1 mol-l cm-1. T+ L notable that the antimony complex does not absorb at wavelengths greater than 600 nm. -ystallisation and purification of the AgDDC reagent results in a lowering of the colour - isity and a shift of the absorption maximum for arsenic to near 522 nm.43 However, the absorption maximum for the antimony complex does not appear to shift from near ACLF nm560 MERRY AND ZARCINAS : SPECTROPHOTOMETRIC DETERMINATION OF Analyst, Yd.105 Simultaneous Determination of Arsenic and Antimony There have been many conflicting reports describing colour formation with the stibine - AgDDC complex, interference in arsenic determinations and the possibility of measuring the two elements sim~ltaneously.~-~ Only Dal Cortivo et aL6 obtained recoveries of antimony from spiked samples at levels similar to those obtained for arsenic. Dal Cortivo et al. did not use the zinc - hydrochloric acid system of hydride generation common to the others.3-5 The results that we have obtained suggest that, a t least when sodium tetrahydroborate(II1) is used for reduction, simultaneous determinations of both elements are possible. Linear calibrations of absorbance with arsenic and antimony in the range 0-2Opg are shown in Fig.2. Results obtained with antimony(V) are identical with those obtained with antimony(II1). Detailed investigation in the range 0-1 pg showed that linearity was maintained, but when doubly distilled pyridine was not used or the AgDDC had aged, amounts of arsenic lower than 0.2-0.6 pg could not be measured. 2.0 1 1.6 I I $ 1.4 5 1.2 f g 1.0 2 0.8 0.6 0.4 0.2 0 Amount of arsenic or antimony/pg Fig. 2. Arsenic and antimony calibra- tion graphs a t different wavelengths (using a 20-mm path length flow-through cell): A, antimony a t 504 nm; B, arsenic at 540 nm; C, arsenic at 504 nm; and D, arsenic at 600 nm. By measuring absorbances at two wavelengths where differences in molar absorptivities are maximised and solving simultaneous equations (for example, using special cases described by Meehan13), estimates of both elements can be made.Wavelengths of 504 and 600 nm are most useful (Fig. 1). For more sensitive arsenic determinations in the presence of negligible amounts of antimony, a wavelength of 540 nm was used. An AgDDC preparation with an absorption maximum in the region of 540 nm for arsenic is preferable to a purified preparation with an absorption maximum at 522 am for simultaneous determinations because of the wider separation of the absorption peaks. Measurements on mixed standard solutions result in determinations of the individual elements with errors of less than about 5%. Another technique for the separation of the two elements was attempted, using freezing mixtures to separate the hydrides.Difficulties similar to those reported by Skogerboe and Bejmuk15 were encountered and the approach was abandoned because of increasing com- plexity. Digestion Procedures for Sediment, Soil and Plant Material A number of different wet-digestion techniques are commonly used for the analysis of soils and sediments contaminated with heavy metals. In this laboratory, nitric or nitric plus perchloric acids are often used as they have the advantage of beingrelatively rapid and, at least with samples contaminated with lead, zinc, cadmium and copper, were found to give recoveries approaching those obtained using hydrofluoric acid or X-ray fluorescence procedures. The use of nitric or nitric plus perchloric acid digestions for the simultaneous spectrophotometric determination of arsenic and antimony gave reasonable results for arsenic, but the antimony recoveries were very low and variable.June, 1980 561 No loss of antimony occurred with the digestion of a 20-pg standard, but when antimony concentrations comparable to those found in polluted soils and sediments were digested with oxidising acids, a fine, white precipitate appeared that could be dissolved by reducing agents.This precipitate would not be visible in the residue of a soil or sediment digest and conse- quently would be lost on filtration or decantation of the supernatant prior to analysis. Maren16 also reported erratic antimony recoveries, which he attributed to the formation of an insoluble oxide that was found to be readily reduced or further oxidised prior to deter- mination of antimony using the Rhodamine B method.As AND Sb BY THE SILVER DIETHYLDITHIOCARBAMATE METHOD TABLE I RECOVERIES OF ARSENIC AND ANTIMONY (pg g-l) FROM SOIL AND SEDIMENT Calcareous marine sediment 1 Method As Sb X-ray fluorescence . . . . 160* 425* Spectrophotometry following acid digestion- HC1 .. .. - * (a) 48 156 49 334 357 56 HC1-HNO, .. . . (a) 135 (b) 160 323 (b) 125 31 (c) 143 235 (duplicate determinations only) . . .. .. . . 11.9 12.5 HNO, .. .. .. (a)? 101 5* (c) (b) 75 (c) 154 375 HNO, - HClO, . . (a) 122 5 Coefficient of variation, % Calcareous soil r As Sb 120* 80* 87 65 67 82 110 104 110 104 100 103 6* 41 79 79 18 54 70 5 14 39 6.8 5.4 Orchard soil As Sb 120* 4* 85 l$ 92 41$ 96 1 lo$ 103 122: 100 893 90 5$ 100 29$ 103 36: 105 1031 94 5.3 3.6 * Single determination only, others duplicates. t (a) Made up to volume with distilled water.(b) Made up to volume with 2 N HC1. (c) As in (b), but treated with hydroxylammonium chloride. $ 120 pg g-l of Sb added prior to digestion. A summary of an investigation of digestion procedures and treatment of the digested material with a reducing agent (hydroxylammonium chloride) is shown in Table I, and details of the procedures are given in the Appendix. The calcareous samples were taken from near a smelter and also contained considerable concentrations (5000-7000 pg g-l) of both lead and zinc. The orchard soil also contained about 600 pg g-l of lead and 250 pg g-l of copper and, for this study, was spiked with 120 pg g-I of antimony. When the results obtained spectrophotometrically using AgDDC are compared with the X-ray fluorescence results, it appears that a concentrated hydrochloric acid digest gives the best results for antimony recovery, but an aqua regia digest treated with reducing agent is best for the spectrophotometric determination of both elements.At least 2 g of hydroxylammonium chloride was necessary to give a marked improvement to the recovery of antimony in the presence of perchloric acid. Tin(I1) chloride should not be used as a reducing agent because hydrides of tin appear to form a transient coloured complex with AgDDC. Making up the digest to volume with 2 N hydrochloric acid (100 ml in this instance) also considerably improved the recovery of antimony.The large variability of the results for the marine sediment may be partly explained by the suspected presence of ore materials. A sulphuric acid - nitric acid - perchloric acid digest procedure was used prior to deter- minations of arsenic and antimony in plant materials, including the National Bureau of Standards orchard leaves (Standard Reference Material 1571). The mean arsenic concentra- tion of the orchard leaves determined on five occasions was 12.8 pg g-l with a standard deviation of 0.4 pg g-l (compared with the certified value of 14 3 2 pg g-l). A bulk plant sample of Marrubium vulgare L. , which was collected from an area affected by mine tailings, was used to investigate the usefulness of the simultaneous spectrophotometric arsenic and562 MERRY AND ZARCINAS : SPECTROPHOTOMETRIC DETERMINATION OF Auzalyst, Vol.105 antimony procedure. Its arsenic concentration, measured on separate digests on eight occasions over a period of several months, was 12.4 As its antimony con- centratioh was low, a 2Opgg-1 spike was added prior to digestion in order to test the recovery. In each instance 1 g of plant material was digested and made up to 20 ml with 2 N hydro- chloric acid. Where a reducing agent was used, 1 g of hydroxylammonium chloride was added to each and heated at 100 "C for 30 min before making up to the final volume. The 10% recovery of the 2Opgg-l antimony spike increased to S5y0 following treatment with the reducing agent. 0.7 pg gl. The results are shown in Table 11. TABLE I1 DETERMINATION OF ARSENIC AND ANTIMONY I N DIGESTED PLANT MATERIAL Means and standard deviations of three determinations.Sample Aslclg g-l Sb/CLg g-l Original sample . . .. .. .. .. . . 12.3 f 0.21 0.7 f 0.17 Spiked with 20 pg g-' of'Sb . . .. 12.0 f 0.29 3.1 f 0.46 Spiked with 20 pg g-l of Sb and treated with HONHsCl, hydroxylammonium chloride. . .. .. .. . . 11.6 f 0.27 17.8 f 0.76 Conclusion The results for samples analysed spectrophotometrically for arsenic with hydride generation using sodium tetrahydroborate (111) and subsequent colour development in silver diethyl- dithiocarbamate may be in error if significant amounts of antimony are also present. This can be overcome by determining arsenic at 600 nm, where the antimony complex does not absorb. Alternatively, the arsenic and antimony complexes can be used for the simul- taneous measurement of both elements.In this instance, spectrophotometric measurements can be made at 504 and 600 nm and concentrations of both elements calculated using a standard method involving simultaneous equations. An investigation into digestion procedures for arsenic and antimony in polluted sediments, soils and plants showed that rapid wet-chemical digests using oxidising acids may result in the formation of insoluble antimony compounds that are lost with the residue. Treatment of the digest with a reducing agent prior to hydride generation allowed the determination of arsenic and antimony on the same digest. Appendix Digestion Procedures for Soils and Sediments Nitric acid, hydrochloric acid and 3 + 1 hydrochloric acid - nitric acid digests mixture.Heat on a hot-plate for 2 h at 95 "C, cool, then make up to volume with water. Moisten 5 g of soil or sediment, then carefully add 15 ml of concentrated acid or acid Nitric acid - $erchloric acid digest Moisten 5 g of soil or sediment, then add 10 ml of concentrated nitric acid and heat until viscous. Add 5 ml of 3 + 1 nitric acid - perchloric acid, heat again until viscous, then add 2-ml aliquots of perchloric acid until perchloric acid fumes appear; cool and make up to volume with water. To reduce frothing, 5 drops of octan-2-01 can be added to digests of calcareous samples prior to addition of the acid. Redwtiort with hydroxyZammoni.um chloride Prior to making up to the final volume, 4ml of a 50% aqueous solution of hydroxyl- ammonium chloride are added to the digest, which is then heated for a further 30 min in a boiling water bath.June, 1980 As AND Sb BY THE SILVER DIETHYLDITHIOCARBAMATE METHOD References 563 1.2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 16. 16. Aggett, J., and Aspell, A. C., Ariulyst, 1976, 101, 912. Cartwright, B., Merry, R. H., and Tiller, K. G., Aust. J . Soil Res., 1977, 15, 69. VasAk, V., and Sedivek, V., Chem. Listy, 1962, 46, 341. Dubois, L., Teichman, T., Baker, C. J., Zdrojewski, A., and Monkman, J. L., Mikrochim. Actu, Analytical Methods Committee, Analyst, 1975, 100, 54. Dal Cortivo, L. A., Cefola, M., and Umberger, C. J.. Aural. Biochem., 1960, 1, 491. Thompson, K. C., and Thomerson, D. R., Analyst, 1974, 99, 695. Powers, G. W., Martin, R. L., Piehl, F. J., and Griffin, J. M., Anal. Chem., 1969, 31, 1589. Rooney, R. C., Analyst, 1976, 101, 678. Duncan, L., and Parker, C. R., Varian Techtrour, Technical Topics, June 1974. Smith, R. G., Van Loon, J. C., Knechtel, J. R., Fraser, J. L., Pitts, A. E., and Hodges, A. E., Aucal. Aggett, J., and Aspell, A. C., Analyst, 1976, 101, 341. Meehan, E. J., in Kolthoff, I. M., and Elving, P. J., Editors, “Treatise on Analytical Chemistry,’’ Part 1, Volume 6, Wiley, New York, 1964, p. 2753. Jackwerth, E., Arch. Pharm., 1962, 295, 779. Skogerboe, R. K., and Bejmuk, A. P., Anal. Chim. Ada, 1977, 94, 297. Maren, T. H., AfiuZ. Chem., 1947, 19, 487. 1969, 185. Chim. Acta, 1977, 93, 61. Received August 20th, 1979 Accepted November 22nd, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500558

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

564 Analyst, June, 1980, Vol. 105, pp. 564-567 Further Improvement of the Modified Spectrophotometric Method for the Determination of Malathion E. R. Clark and I. A. Qazi Department of Chemistry, University of Aston in Birmingham, Gosta Green, Birmingham, B 4 VET A recently described modified spectrophotometric method for the determina- tion of malathion has distinct advantages over all the existing spectrophoto- metric methods. It suffers from only one drawback in that an expensive spectrophotometer has to be used for the method to be of reasonable sensi- tivity. This is so because the absorption peak of the bismuth complex of dimethyldithiophosphate (the hydrolysis product of malathion) occurs in the ultraviolet region (325 nm). The method described in this paper is an improvement of that modified method but one in which an additional step has to be carried out and the final measurement of the absorbance of an orange - yellow solution is made a t 495 nm.The improvement involves a ligand exchange reaction in which the bismuth - dimethyldithiophosphate complex is transformed quantitatively into the bismuth - dithizone complex. Keywords : Malathion alkaline hydrolysis ; malathion determination ; bismuth complex ; bismuth determination with dithizone The modified spectrophotometric method for the determination of malathion1 has distinct advantages over all the previously described methods in terms of the number of reagents required, the ease of application, avoidance of interferences and the stability of the colour for extended periods of time.The method involves the alkaline hydrolysis of malathion, the final absorbance measurement being made on the bismuth complex of dimethyldithio- phosphate (DMDTP), one of the hydrolysis products. The only disadvantage of the method is that the absorbance of the bismuth- DMDTP complex lies in the ultraviolet region (325nm), and the full potential of the method can be exploited only when an expensive spectrophotometer is used. As this limits the use of the method for routine applications, it was thought desirable either to develop a new method for malathion having all the advantages of the bismuth method, or to improve the method so that the final measurement of the absorbance could be made in the visible region. In this work the second approach was adopted.The fact that the bismuth - DMDTP complex is extracted into carbon tetrachloride only from acidic solutions led to some preliminary studies with the yellow organic extract. It was found that when the carbon tetrachloride solution of the yellow bismuth complex was shaken with pH 10 buffer solution, the yellow colour disappeared completely and the absorbance at 325 nm fell to zero. Also, when the buffer solution from the above experi- ment was acidified with concentrated nitric acid and shaken with another portion of the organic solvent (equal to the original volume), the yellow colour was imparted to the latter. More importantly, the absorbance of the second extract was the same as that of the first. From these observations, it was concluded that when an organic solvent containing the bismuth-DMDTP complex is brought into contact with pH 10 buffer, the complex is completely broken up and the bismuth( 111) and DMDTP ions are quantitatively transferred to the aqueous layer.Hence, an alternative method for the spectrophotometric determina- tion of bismuth could bring about the desired improvement. The only limitation required of such a method was that it should be applicable under the strongly alkaline conditions described above. Fortunately, two well known reagents for the determination of bismuth, sodium diethyl- dithiocarbamate (DEDTC) and dithi~one,~,~ fulfil these requirements. The basic feature in the improvement of the modified spectrophotometric method for the determination of malathion described in this paper is one which involves a ligand exchange reaction, i.e., the bismuth - DMDTP complex is transformed to the bismuth - dithizone (or the bismuth - DEDTC) complex, which is used for the final absorbance measurement.CLARK AND QAZI 565 Experimental Reagents Dimethyldithiophosphate solution (DMDTP), 5 x 10-3 M.Dissolve 0.1752 g of the purified ammonium salt in 200 ml of distilled water. Bismuth solution. Dissolve 0.1 g of bismuth oxide (Bi,O,) (BDH, laboratory reagent) in 3 ml of concentrated nitric acid and dilute to 100 ml [1.0 ml of this solution contains 0.45 mg of bismuth(III)]. Dissolve 0.3880g of bismuth oxide (Bi203) in a few millilitres of concentrated nitric acid and dilute to 100 ml with distilled water. Dilute 10 ml of this solution to 100 ml. Diethyldithiocarbamate solution (DEDTC), 1%.Dissolve 1 g of the sodium salt (BDH, analytical-reagent grade) in 100 ml of distilled water. Dithizone solution. Dissolve 0.5 g of the reagent (Hopkins and Williams, analytical- reagent grade) in 500 ml of purified carbon tetrachloride. Dilute 10 ml of this stock solution to 100 ml with carbon tetrachloride when required. Bismuth solution, 1.66 x 10-3~. Carbon tetrachloride. Bufler solution, pH 10. Re-distil commercial-grade material and store in a glass bottle. Dissolve 26.2 g of ammonium nitrate in the minimum amount of de-ionised water, transfer into a 250-ml calibrated flask and add 142.5 ml of analytical- reagent grade ammonia solution (sp. gr. 0.880). Make up to volume with water. Apparatus Sfiectrophotometer. Unicam SP6-100 spectrophotometer with 1 .O-cm silica cells.Procedure As preliminary Job’s plot studies4 had indicated that bismuth forms a 1 : 3 complex with DMDTP, both DEDTC and dithizone were first studied with 1.66 x 1 0 - 3 ~ bismuth(II1) solution.* Later, the same reagents were employed for the indirect determination of 5 x 10-3 M DMDTP. (i) Extraction of bismuth - DEDTC com$lex To a 50-ml separating funnel were added 2 ml of the buffer solution, about 8 ml of distilled water and 10 ml of carbon tetrachloride. After adding measured amounts (0.25-1.90 ml) of the 1.66 x M bismuth(II1) solution, 1 ml of DEDTC was added and the funnel stoppered and shaken vigorously. The organic layer was allowed to separate and was then transferred to the 1-cm silica cell, via a cotton-wool plug placed in the stem of the funnel. The absorption of the complex was measured at 400nm using carbon tetrachloride as a reference solution.The results are presented as solid circles in Fig. 1. (ii) To the 50-ml separating funnel were added 2 ml of the buffer, 8 ml of de-ionised water and 10 ml of dithizone working solution. The funnel was stoppered and shaken in order to transfer the dithizone to the aqueous alkaline layer. The organic layer was discarded and the buffer solution shaken with another 10-ml volume of carbon tetrachloride. The two layers were allowed to separate and before discarding the second organic layer its absorbance was measured and this value was subtracted from all further readings. A third volume of carbon tetrachloride (10 ml exactly) was added to the funnel followed by 0.01 or 0.02 ml of 1.66 x The contents were shaken, and the absorbance of the organic layer was measured at 495 nm, after transferring it to the 1-cm cell as described above.The results are presented as solid squares in Fig. 1. Extraction of bismuth - dithizone complex M bismuth(II1) solution. (iii) M DMDTP, was extracted into 10-ml portions of carbon tetrachloride as described previously.1 The absorbance of the solution was measured a t 325 nm and a portion of this solution was Indirect determination of DMDTP using dithiocarbamate The bismuth- DMDTP complex, using measured amounts (0.2-1.6 ml) of 5 x * 1:3 = 1.66 x 10-s:5 x566 CLARK AND QMI : FURTHER IMPROVEMENT OF THE MODIFIED Analyst, VoZ. 105 transferred into a 50-ml funnel containing the DEDTC reagent and the buffer (pH 10).It is clear that it is not important to transfer the carbon tetrachloride extract quantitatively to the second funnel and that this simplifies the procedure. The contents of the funnel were then shaken and the absorbance of the organic layer was measured at 400 nm as described in (i) above. The results are presented as open circles in Fig. 1. 2 .o 1.5 a) C $ 1.0 a n < 0.5 0 0.5 1 .o 1.5 2 .o Volume of 5 x M DMDTP or 1.66 x M Bi(lll)/ml Fig. 1. Absorbance versus concentration graphs for the bismuth complexes with different ligands extracted into 10 ml of carbon tetrachloride and measured at the wave- lengths indicated] in 1-cm silica cells with carbon tetra- chloride as reference solution.A, Bi - dithizone complex (A = 495 nm) ; B, Bi - DMDTP complex ( A = 326 nm) ; C, Bi - DMDTP complex (A = 390 nm); and D, Bi - DEDTC complex (A = 400 nm). (iv) M DMDTP) was extracted into 10ml of carbon tetrachloride, as described previously,l followed by the transference of part of the extract into a second funnel containing the dithizone in the buffer solution, which had been purified by repeated shaking with the solvent [see (ii) above]. The contents of the funnel were shaken and the absorbance of the carbon tetrachloride layer was measured at 495 nm. The results are presented as open squares in Fig. 1. Most of the organic solvent used can be recovered by simple distillation and can be re-used. Indirect determination of DMDTP using dithizorte Again, the bismuth - DMDTP complex (using 0.005 or 0.015 ml of 5 x Results and Discussion In addition to the bismuth complexes of DEDTC and dithizone, Fig.1 also presents the absorbances due to the bismuth - DMDTP complex both at 325 nm (solid triangles) and at 390 nm (open triangles). The results shown in Fig. 1 demonstrate that the bismuth - DMDTP can be quantitatively converted into either the bismuth - DEDTC complex or the bismuth - dithizone complex, as the plots of the absorbances versus concentration, i.e., the open and closed squares and the open and closed circles, fall respectively on two independent straight lines. The other important information that can be obtained from Fig. 1 concerns the sensitivity of the two reagents. The use of DEDTC gives a slight improvement in the intensity of the colour but it is not sufficiently sensitive to be useful. On the other hand, when dithizone is used, not only does the almost colourless carbon tetrachloride solution containing the bismuth - DMDTP complex become bright orange, but the sensitivity is also increased about 4fold.Jwze, 1980 SPECTROPHOTOMETRIC METHOD FOR DETERMINATION OF MALATHION 567 It is therefore clear that by the inclusion of one more step in the modified spectrophoto- metric method, in which the final extract is shaken with dithizone solution, malathion could be determined with improved sensitivity using even simple photometers. We have applied this improved method to the analyses of technical products and have found it to be suitable although it is slightly less precise than the direct measurement of the bismuth - DMDTP complex. References 1. 2. 3. 4. Clark, E. R., and Qazi, I. A., Analyst, 1979, 104, 1129. De, A. K., Khopkar, S. M., and Chalmers, R. A., “Solvent Extraction of Metals,” Van Nostrand Marczenko, Z., “Spectrophotometric Determination of Elements,” Ellis Honvood, Chichester, 1976, Clark, E. R., and Qazi, I. A., unpublished work. Reinhold, New York, 1970, p. 238. p. 149.

ISSN:0003-2654    

DOI:10.1039/AN9800500564

出版商:RSC    

年代:1980

数据来源: RSC

摘要:

568 Analyst, June, 1980, Vol. 105, $$, 568-574 Spectrophotometric Assay of Hexamethylenetetramine Aly M. Taha, NawaI A. El-Rabbat" and Fardous Abdel Fattah Faculty of Pharmacy, University of Assiut, Assiut, Egypt The intense charge-transfer band in the ultraviolet spectra of the molecular complex of hexamethylenetetramine with iodine has been utilised in the sensitive spectrophotometric assay of this compound. From the various polar and halogenated solvents that were investigated, 1,2-dichloroethane and chloroform were selected. The hexamethylenetetramine - iodine com- plex has two peaks. Maximum intensity was attained at the shorter wave- length band, 273 nm, E = 2.0 x lo4 1 mol-1 cm-1 in 1,2-dichloroethane and 270nm, E = 1.6 x lo3 lmol-lcm-l in chloroform. The effect of time on the stability of the complex was investigated.A Job's plot of absorbance veww the molar ratio of hexamethylenetetramine to iodine indicated a 1 : 1 hexamethylenetetramine to iodine ratio. Accurate analysis of hexa- methylenetetramine, free and in tablet form, in conjunction with cholic acid or potassium hydrogen phosphate, was possible by this method in the con- centration range 1-12 p g ml-l of hexamethylenetetramine with a relative standard deviation of 0.005. Keywords : Spectrophotometry ; hexamethylenetetramine assay ; iodine complex Amines are excellent n-electron donors and can interact strongly with electron acceptors such as iodine to form charge-transfer complexes. Charge-transfer complexes of amines with halogens and pseudohalogens have been reported.lW6 Some workers have carried out physical methods of study employing vibrational spectra7-11 on the complexes of the tertiary amine hexamethylenetetramine (hexamine ; methenamine) with halogens and inter-halogens.However, no reports could be found of the direct ultraviolet spectrophotometric deter- mination of hexamethylenetetramine, and few approaches are known for its colorimetric assay utilising the nitrogen moiety. For this reason the proposed charge-transfer complexa- tion method has been applied to hexamethylenetetramine assay, free and in pharmaceutical preparations. Experimental Apparatus (Mom, Budapest) using 1-cm cells. Spectra were recorded on a Spectromom 203 ultraviolet and visible range spectrophotometer Samples HexamethylenetetrLm~ne. Pharmaceutical grade.Utilised as a working standard. The following commercial preparations were used in the analytical procedure. Methenamine - cholic acid tablets. Felamine Tablets, Swiss-Pharm, Cairo, containing 225 mg of methenamine per tablet. Methenamine hydrogen Phosfihate tablets. Methenamine - potassium hydrogen phosphate tablets (NFX), prepared according to the cornpendial directions to contain 300 or 500 mg of methenamine per tablet. USP XVIII, prepared according to the compendial directions to contain 300 or 500 mg of methenamine per tablet using Avicel, gum acacia, lactose and starch as diluents. Methenamine tablets. Reagents Solvents. ment.12 * To whom all correspondence should be addressed. Solvents used were of spectroscopic grade or rendered so by appropriate treat-TAHA, EL-RABBAT AND ABDEL FATTAH 569 Re-sublimed iodine (31.75 mg) was dissolved in the selected solvent in a 25-ml calibrated flask.The solution was stable for at least 1 week at 4 "C. Re-sublimed iodine (12.7 mg) was dissolved in 1,2-dichloroethane, or any other selected solvent, in a 50-ml calibrated flask to obtain a 0.255 mg ml-1 solution of iodine. The solution was stable for at least 1 week a t 4 "C. Iodine soZzztion for qualitative analysis (5 x M). Iodine solution for quantitative analysis (1 x 10-3 M). Preparation of Sample Solutions Standard solutions An accurately weighed amount of hexamethylenetetramine was dissolved in 1,2-dichloro- ethane or chloroform and diluted stepwise to obtain a concentration of 50 pg ml-l of hexa- met hylene tetramine.Methenumine tablet solutions An accurately weighed amount, equivalent to 5 mg of methenamine, from a composite of 10 finely powdered tablets was mixed thoroughly with 1,2-dichloroethane or chloroform until extraction was complete and the suspension was then diluted to 100 ml in a calibrated flask. The solution was filtered and then subjected to the assay procedure, the final con- centration being 50 pg ml-l. Assay Procedure A 1-ml aliquot of the prepared solution was transferred into a 10-ml calibrated flask, 1 ml of iodine reagent (1 x 10-3 M) was added and the solution was made up to volume with the selected solvent. The solution was allowed to stand for 30min a t constant temperature (25 & 1 "C). The absorbance of the solution was scanned from 400 nm to the cut-off point of the selected solvent, to determine the absorbance maxima, against a reagent blank treated similarly.For quantitative work, single readings were made for various dilutions at the previously determined maxima. Construction of calibration graph An amount of methenamine (about 25 mg) was weighed accurately, dissolved in the desired solvent and diluted to volume in a 100-ml calibrated flask. The solution was diluted step- wise to give a series of concentrations suitable for construction of the calibration graph in the range 10-120 pg ml-l; 1 ml of each solution was utilised for complex formation with the iodine reagent (1 x 10-3'~) as described under Assay Procedure. Study of Sample Solvents The assay procedure was applied to solutions containing 5.6 pg ml-l of hexamethylene- tetramine in the solvent to be studied.Iodine solution (5 x M) in the same solvent was used. The following solvents were subjected to the procedure: dioxane, 1,2-dichloro- ethane, carbon tetrachloride, chloroform, dichloromethane, methanol and 1,1,2,2-tetra- chloroethane. Study of the Effect of Time on Hexamethylenetetramine - Iodine Complex tetramine. for 12 different periods of time ranging from 5 to 60 min with 5-min intervals. The assay procedure was applied to 12 solutions containing 3 pg ml-1 of hexamethylene- The hexamethylenetetramine - iodine mixtures were left to stand at 25 & 1 "C Effect of Varying the Iodine Concentration on the Intensity of Absorption The assay procedure was applied to dilutions containing 10 pg ml-l of hexamethylene- tetramine in 1,2-dichloroethane, using varying concentrations of iodine ranging from 1 x The absorbance of the complex formed was measured at the specified wavelength against a reagent blank containing 10pgml-l of hexamethylenetetramine.to 5 x 1 0 - 4 ~ with 1 0 - 4 ~ increments.570 TAHA et al. : SPECTROPHOTOMETRIC Analyst, Vol. 105 Stoicheiometric Relationship Standard solutions of iodine (2 x l o - 4 ~ ) and hexamethylenetetramine (2 x 10-4~) in anhydrous chloroform were prepared. A series of the standard solutions of hexamethylenetetramine and iodine in different complementary proportions, totalling 10m1, (from 0 + 10 to 10 + 0 inclusive) were prepared in 10-ml calibrated flasks. After 30min the absorbances were measured at the wavelength of maximum absorption of the complex, which had been previously deter- mined.Job's method of continuous variationlS was employed. 1.2 1 .o 0.8 V m Q $ 0.6 2 0.4 0.2 0 Wavelengt h/nm Fig. 1. Spectra of hexamine (hexamethylenetetramine) - iodine complex (hexamine 4 x 10-5 M ; iodine 5 x 10-4 M) in different solvents: A, dioxan, B, 1,2-dichloroethane ; C, carbon tetrachloride ; D, chloroform ; E, dichloro- methane; F, 1,2-dibromoethane; G, methanol; H, 1,1,2,2-tetrachloroethane; and I, hexamine (100 pg ml-1) in chloroform. Results and Discussion Charge-transfer Spectra in Different Solvents The positions of maximum absorbance and the intensity of absorption of the hexamethylene- tetramine - iodine complex in the different solvents studied are represented in Fig.1 and Table I. It is observed that the uncomplexed hexamethylenetetramine has a non- characteristic and insignificant end-absorption in the accessible ultraviolet region. Con- sidering that the spectra in Fig. 1 were recorded against an iodine blank, then the absorption in this region is due solely to the complex itself. The positions of maximum absorbance and the ratio of the absorption of the intense to weak bands, when compared with the corresponding ratio calculated from previously reported data,5J4 reflect the prevalence of the outer change-transfer complex [R3NI,, equation ~ (l)] in most of the solvents studied, with the exception of the highly polar solvent methanol, where the triiodide anion is probably the predominent species. R3N +'I, ------- R3N.I, .... .. - * (1) - (Outer complex) .. .. * - (2) R3N.I, -2 ,-- (R,N+ - 1)I- (Inner complex) (R,N+ - 1)I- + I, (R3N - 1)+13- .. .. * * (3) (Triiodide ion pair)June, 1980 ASSAY OF HEXAMETHYLENETETRAMINE 571 For methanol, a lower absorption intensity is observed, which may be attributed to the interaction of iodine with methanol leading to high blank readings. It was also noticed that with some grades of chloroform no complex was formed, shown by a lack of absorbance in the wavelength region studied, while with other grades the maximum was red-shifted to 295 nm. These results are presumably due to the presence of trace amounts of ethanol, which is usually added as a preservative to chloroform. These observations agree with some previous reports.15 However, on washing chloroform with water, drying, then distilling it, reproducible absorption readings were obtained at 270 nm, confirming the suggestion that the interference was due to ethanol.TABLE I CHARGE-TRANSFER BANDS OF HEXAMETHYLENETETRAMINE IN SELECTED SOLVENTS Solvent Dioxane.. * . .. .. 1,2-Dichloroethane . . .. Carbon tetrachloride . . . . Chloroform . . .. . . Dichloromethane . . .. Dibromoethane .. . . Methanol . . .. .. 1,1,2,2-Tetrachloroethylene . . Amax,,/ El*/ hmax,2/ nm lmol-lcm-l nm 273 31 000 380 273 29 000 380 2 74 25 000 405 270 21 000 385 270 21 000 383 290 4 000 360 290s 11500 - End-absorption Solvent %*I cut-off 1 mol-l cm-I pointlnm 2 900 245 2 600 235 1500 258 1700 242 2 000 235 - 283 1500 215 250 Solubility t / g in 50 ml N.d. $ 1:50 1: 110 1: 10 N.d.N.d. N.d. N.d. * Apparent Q value based on the relative molecular mass of hexamethylenetetramine. t Experimentally determined average value of a t least three determinations. $ Not determined. $ Small plateau from 290 nm to solvent cut-off point. Selection of sample solvent Although a charge-transfer complex has probably been formed in both 1,2-dibromoethane and 1,1,2,2-tetrachloroethane, as suggested by the change in colour of iodine and the intense end-absorption (Fig. l ) , the high cut-off points of these solvents obscured the scanning of shorter wavelengths and therefore clear-cut spectroscopic evidence for charge-transfer formation could not be ascertained. In principle, it seems that any of the solvents can be used for hexamethylenetetramine assay.However, the low solubility of the drug in carbon tetrachloride (1 : l l O ) , and the poor extraction power of this solvent, restrict its use especially with pharmaceutical prepara- tions. Dioxane is not recommended as an assay solvent owing to the observed deviations from Beer's law. In addition, the Beer's law plot was linear in the concentration range 2-10 pg ml-l. However, the use of dichloromethane as the solvent in a quantitative application was limited by its low boiling-point (39 "C). Consequently, 1,2-dichloroethane and chloroform may be con- sidered to be the most suitable solvents for the determination of this drug, as they donot have such low boiling-points. With dichloromethane, high absorbance readings were obtained. 0 10 20 30 40 50 60 Time/rn in Fig.2. Absorbance versus time graph for hexamine (hexamethylenete- tramine) - iodine complex in chloroform.572 TAHA et al. : SPECTROPHOTOMETRIC Analyst, Vol. 105 Rate of Complex Formation Fig. 2 shows that constant absorption readings were obtained after 30 min when using 1,2-dichloroethane as the solvent. The initial change in absorption with time implies that the reaction is not completely ionic, or it would have been virtually instantaneous. As regard the temperature effects, in princ.iple the intensity of absorption decreases on increasing the temperature16 and reversion of the iodine colour was observed by heating the charge-transfer complex on a water-bath. This indicates the dissociation of the complex back into its components.Thus, the use of a constant low-temperature water-bath (e.g., 15 "C) is recommended for higher and more stable absorption intensities. However, working at a non-fluctuating room tempera- ture (20-25 "C) was found to be adequate for sensitive and precise quantitative work. This is also more convenient practically and has been adopted throughout the quantitative experi- ments. TABLE I1 EFFECT OF VARIATION OF IODINE CONCENTRATION ON THE ABSORBANCE Therefore, lower temperature ranges are preferable. OF HEXAMETHYLENETETRAMINE - IODINE COMPLEX IN CHLOROFORM Hexamethylenetetramine concentration was 10 p g ml-1 in all experiments. Concentration of iodine/ ml-l 10 20 40 60 80 100 Absorbance at 270 nm 0.129 0.164 0.234 0.309 0.360 0.434 Effect of Variation of Iodine Concentration The effect of varying the iodine concentration on the intensity of absorption while keeping the amount of hexamethylenetetramine constant was tested and the results are shown in Table 11.0.6 , 0.5 1 0 0.1 0.3 0.5 0.7 0.9 [Hexarnine] /[hexaminel + [ I 2 1 Continuous variation plot obtained from solutions of hexamine (hexamethyl- enetetramine) and iodine in chloroform (2 x 10-4 M). Fig. 3. It was found that for a fixed hexamethylenetetramine concentration, there is a direct relationship between the intensity of absorption and the amount of iodine. This finding is probably due to the increased change-transfer complex formation owing to the mass effect of the added iodine. However, an iodine concentration of ~ O - * M was found to be moreJune, 1980 ASSAY OF HEXAMETHYLENETETRAMINE 573 suitable and has been selected for the assay procedure.This selection was based on the following considerations : a more concentrated iodine solution may cause oxidation of the amine or it may enhance the iodine interaction with certain type of solvent$; this low concentration of iodine is within the stipulated range that is free from Beer’s law deviations5; and a high concentration of the electron donor is required to allow for the triiodide forma- tionl7; hence both donor and acceptor are kept at the same order of magnitude. TABLE I11 BEER’S LAW PLOT OF HEXAMETHYLENETETRAMINE - IODINE COMPLEX I N VARIOUS SOLVENTS Concentration Solvent Slope Intercept rangelpg ml-l 1,2-Dichloroethane . . . . 0.174 0.019 2-10 Chloroform .. .. .. 0.108 0.018 2-9 Carbon tetrachloride . . . . 0.140 0.016 2-1 1 Hexamethylenetetramine to Iodine Ratio in the Complex The application of Job’s method of continuous variation13 indicated that, in spite of the four nitrogen atoms present in the hexamethylenetetramine molecule, only the 1 : 1 hexa- methylenetetramine - iodine complex is formed when using chloroform as a solvent (Fig. 3). This result is in line with the reported infrared datall on iodine complexes of this compound, and can be explained on the basis that a univalent, partially positively charged hexa- methylenetetramine species may be formed initially during the charge-transfer process, which may not be easily engaged in additional complex formation. This suggestion is supported when examining the behaviour of this compound in non-aqueous media, where it was found to titrate as a monobase.TABLE IV ASSAY OF HEXAMETHYLENETETRAMINE IN BULK DRUG AND DOSAGE FORMS BY CHARGE-TRAN-SFER METHOD Hexamethylenetetramine sample Standard hexamethylene- or tablet tetraminet - Amount takenlmg or mg per tablet Sample or tablet* Bulkdrug . . .. .. .. 28 Bulk drug . . .. .. .. 56 Bulk drug . . .. .. .. 84 Methenamine tablets ‘(USP’ XVIII)’ 300 500 Methenamine hydrogen phosphate tablets . . .. .. . . 300 500 Felamine tablets . . .. .. 225 Bulk drug . . 112 Recovery, $ % 99.8 99.8 99.3 97.7 100.9 100.3 99.2 99.1 99.8 - r Standard deviation, % 1.03 1.14 0.96 2.21 0.91 1.37 0.82 0.90 1.52 A 1 Standard Added/ Recovery, deviation, mg % % 300 100.9 0.91 500 101.1 0.36 300 99.3 0.80 500 100.5 1.35 225 100.2 1.20 * Detailed composition given under Experimental.t Content pre-determined by the compendia1 method.18 $ Average of three determinations. Quantification, Linearity of Beer’s Law Plot, Accuracy and Precision Calibration graphs for the hexamethylenetetramine - iodine complex in 1,2-dichloro- ethane, chloroform and carbon tetrachloride were constructed. The results are presented in Table 111. Conformance to Beer’s law was observed in all solvents over the concentration range 1-12 pg ml-l of hexamethylenetetramine in the final assay solution. The reproducibility of the procedure was determined by running replicate samples, each containing 4.6 pg ml-l of hexamethylenetetramine in the final test solution. At this con- centration level, the standard deviation for five determinations was 6.8 x pg ml-l.574 TAHA, EL-RABBAT AND ABDEL FATTAH Application to Bulk Drug and Dosage Forms The recovery of hexamethylenetetramine from bulk drugs, given in Table IV, indicates the accuracy and precision of the charge-transfer method.The results of the application to dosage forms are also shown in Table IV. The data in Table IV reveal the suitability of the method for determining hexamethylene- tetramine in these preparations without the risk of interference from various additives, thus recommending it for routine work. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. References Mullikens, R. S., J . Am. Chem. SOC., 1950, 72, 600. Papov, A. I., and Rygg, R. H., J . Am. Chem. Soc., 1957, 79, 4622. Yada, H., Tanaka, J., and Nagakura, S., Bull. Chem. Soc. Jpn., 1960, 33, 1660. Kobinata, S., and Nagakura, S., J . Am. Chem. %c., 1966, 88, 3905. Foster, R. , “Organic Charge-transfer Complexes, Academic Press, London, New York, 1969. Rao, C. N. R., Bhat, S. N., and Dwevedi, P. C., in Brama, E. G., Editor, “Applied Spectroscopy Reviews,” Volume 5, Marcel Dekker, New York, 1972, pp. 1-170. Marzocchi, M., and Fenoni, E., Gazz. Chim. Itul., 1961, 91, 1200. Marzocchi, M., Gazz. Chim. Ital., 1961, 91, 1216. Negita, H., Nishi, Y., and Koga, K., Spectrochim. Acta, 1965, 21, 2144. Halpern, A. M., and Weiss, K., J . Am. Chern. SOL, 1968, 90, 6297. Bowmaker, G. A., and Hannan, S. T., Aust. J . Chem., 1972, 25, 1151. Reddick, J. A., and Bunger, W. B., “Organic Solvents,” Third Edition, John Wiley, New York, Rose, J . , “Advanced Physico-Experiments,” Pitman, London, 1964, p. 54. Moriguchi, I., Araki, Y . , and Kaneniwa, N., Chem. Pharm. Bull., 1969, 17, 2088. Tar, H. S. I., Gerlach, E. D., and Dimattio, A. S., J . Pharm. Sci., 1977, 66, 766. Taha, A. M., Ahmed, A. K. S., Gomaa, C. S., and El-Fatatry, H., J . Pharm. Sci., 1974, 63, 1853. Bhat, S. N., and Rao, C. N. R., J . Am. Chem. Soc., 1966, 88, 3216. “The Egyptian Pharmacopoeia,” Fouad I University Press, Cairo, 1953, p. 388. 1970. Received September 20th, 1979 Accepted November 18th, 1979

ISSN:0003-2654    

DOI:10.1039/AN9800500568

出版商:RSC    

年代:1980

数据来源: RSC