摘要:

ANALYST JUNE 1986 VOL. 111 645 Determination of Ca Mg Na Cd Cu Fe K Li and Zn in Acid Mine and Reference Water Samples by Inductively Coupled Plasma Atomic Fluorescence Spectrometry Richard F. Sanzolone and Allen L. Meier US Geological’Survey Box 25046 Denver Federal Center MS 955 Denver CO 80225 USA An inductively coupled plasma atomic fluorescence spectrometric (ICP-AFS) method was used for the determination of nine elements in natural water. Reference and acid mine water samples were analysed by this method to demonstrate its usefulness for hydrogeochemical exploration. The elements were determined in two groups based on the compatibility of operating conditions and consideration of element abundance levels in natural water. Ca Mg and Na were determined as a group using one set of instrumental conditions and a 1 + 99 dilution of the sample and Cd Cu Fe K Li and Zn were determined using another set of conditions and the undiluted sample.The detection limits for the elements are as follows Ca 1.4; Mg 1.7; Na, 2.0; Cd 1.8; Cu 6.2; Fe 15.8; K 3.5; Li 0.3; and Zn 1.2 ng ml-1. Each element has a linear range spanning about four orders of magnitude. The method has good precision and accuracy as shown by statistics on replicate analyses and by the agreement between values obtained and those recommended for the reference water samples and also those obtained by atomic absorption spectrometry for the acid mine water samples. Keywords Inductively coupled plasma atomic fluorescence spectrometry; water analysis; major and trace element determinations The realisation that “generally trace elements in waters can be used as pathfinders for mineralisation” has prompted the widespread application of hydrogeochemical prospecting to the search for buried base metal deposits.1 The need for rapid, inexpensive sensitive multi-element techniques to meet the increasing analytical demands imposed by hydrogeochemical surveys has become apparent.Useful methods for water analysis should meet certain criteria they should be sensitive enough not to require pre-concentration of the sample; accurate and precise enough to detect subtle differences in element concentrations; and ideally be multi-element tech-niques to increase the speed of analysis and to decrease its cost while also reducing sample volume requirements and helping in the interpretation of inter-element associations.Elemental analysis of’water is typically carried out using single-element flame and electrothermal atomic absorption (AA) techniques. Recently however inductively coupled plasma atomic emission spectrometry (ICP-AES) techniques have brought simultaneous multi-element capabilities to water analysis. This paper describes a multi-element analytical technique for the determination of nine major and trace elements in water using a commercially available inductively coupled plasma atomic fluorescence spectrometer (ICP-AFS). The results of this study demonstrate that the ICP-AFS procedure is a viable alternative to AA and ICP-AES methods; it is a rapid simultaneous multi-element technique that has a larger dynamic range and more freedom from chemical interferences than AA techniques and it has a more stable base line and much more freedom from spectral interferences than ICP-AES techniques.Experimental Apparatus A commercially available inductively coupled plasma atomic fluorescence spectrophotometer the Plasma/AFS * manufac-tured by the Baird Corporation was used for all fluorescence * Use of trade names is for descriptive purposes and does not imply endorsement by the US Geological Survey. measurements in this study. Demers et aZ.2 described the operational configuration of the instrument. The instrument has a capacity of 12 simultaneous determinations using interchangeable self-contained element modules each dedi-cated to the determination of an element.Nine element modules were used in this study (Table 1). The ICP-AFS was interfaced with an Apple 11+ computer and an Epson MX-80 printer for the collection and reduction of data generated by the instrument.3 Line voltage regulation to the ICP-AFS was provided by a Sola CVS constant voltage transformer. The sample delivery system incorporated a Gilson Minipuls I1 pump set at 500 units (about 2 ml min-I) a cross-flow nebuliser with 0.013 in i.d. capillary tubing and the standard quartz torch supplied with the instrument. Under nominal pre-burn (5 s) and integration times (10 cycle; about 25 s) the sample volume per measurement was about 1 ml. A Varian AA-6 atomic absorption spectrophotometer and an IL-951 dual-channel atomic absorption spectrophotometer were used to cross-check the results for this study.Reagents and Standards Stock solutions each containing 1000 pg ml-1 in 5% VWnitric acid were prepared from Specpure metals or oxides for Cd, Cu Fe K Li and Zn and from reagent-grade salts or oxides for Ca K Li Mg and Na. A combined standard solution containing 1 pg ml-1 of Cd Cu Fe K Li and Zn and individual 1 pg ml-1 standard solutions for Ca Mg and Na were prepared by dilution of the stock solutions and were used for instrument calibration. Water Samples The acid mine’water samples used in this study were collected from water flowing from four gold mines in the Idaho Springs, Colorado area. The samples were filtered through 0.45-pm membrane filters into acid-washed polyethylene bottles and acidified with 1 ml of concentrated nitric acid per 100 ml of sample.Reference water samples were obtained from the US Geological Survey SRWS (Standard Reference Water Sam 646 ANALYST JUNE 1986 VOL. 111 Table 1. Instrument module settings for elements in this study Observation height*/ Element mm Mg . . . . . . 127 Ca . . . . . . 125 Na . . . . . . 132 Cd . . . . . . 140 c u . . . . . . 140 Fe . . . . . . 127 K . . . . . . . . 140 Li . . . . . . . . 105 Zn . . . . . . 133 * From tip of injection nozzle of ICP torch. Waveleng thl nm 285.2 422.7 589.0 228.8 324.8 248.3 766.5 670.8 213.9 Lamp current/ mA 18.2 18.2 7.9 6.3 10.3 18.2 7.9 11.1 11.1 Photomultiplier/ V 570 600 705 550 600 760 725 760 825 ple) programme.4 They are natural matrix water samples that have been analysed by a variety of methods in more than 25 independent laboratories.Most probable mean concentration and a confidence interval are reported in this paper for each element determined. Procedure Optimisation A systematic study was carried out to determine the operating parameters to obtain the best detection limit for each of the elements. Employing the plasma conditions used in this study each individual elemental module was independently opti-mised for lamp current photomultiplier voltage and obser-vation height (Table 1). Of these parameters observation height was found to be the most critical. Table 1 lists the optimum module settings. Plasma conditions are determined by adjustments of radiofrequency (RF) power carrier coolant and propane gas flow-rates.RF power was found to be the most important factor affecting detection limit followed by in order of importance coolant carrier and propane gas flow-rates. For simultaneous multi-element determinations compromise plasma conditions for the elements studied were used. Consideration of both expected element abundance in natural water samples and the compromise plasma conditions necessi-tated the sorting of the elements into two groups. Ca Mg and Na are determined as a group using a 1 + 100 dilution of the sample. The plasma conditions for this group are as follows RF power 600 W; carrier flow-rate 1.55 1 min-1; coolant flow-rate 12 1 min-1; and propane flow-rate, 60 ml min-1.Cd Cu Fe K Li and Zn are determined as a group using the undiluted sample. The plasma conditions for this group are RF power 425 W; carrier flow-rate 1.40 1 min-1; coolant flow-rate 11 1 min-1; and propane flow-rate 30 ml min-1. Detection limits Detection limits are the best indicators of ICP-AFS performance. Increased signal intensity and apparent increased sensitivity are often accompanied by increased background signal and noise. Thus real performance is poorer than it would appear. By calculating the detection limit real performance is evaluated better. In this study data were collected and detection limits calculated by an Apple 11+ computer using a program written for this purpose.3 The detection limit for an element is calculated by the following equation: 2 v c Detection limit = -S - B where V is the standard deviation of the signal intensity of ten replicates of a blank solution; C the concentration of the standard solution used; S the signal intensity of the standard solution; and B the signal intensity of the blank solution.The signal intensity (in counts) is obtained using the test mode program of the instrument. The detection limits were used not only to evaluate the operating parameters but also to assess the stability of the ICP-AFS and the effect of integration time. Detection limits for each element (Table 2) were determined five times each on two consecutive days to determine the stability. The detection limits were also determined with a four-fold increase in integration time (duty cycle of 40).Sample analysis Element determinations in reference water samples and acid mine water samples were carried out using the menu mode program of the ICP-AFS. In this mode the instrument is calibrated using a blank solution and the 1 pg ml-1 single- and mixed-element standard solutions for two-point calibration. The samples were analysed in two sets. The first set, consisting of Ca Mg and Na required a 1 + 99 dilution of the samples. The second set consisting of Cu Cd Fe K Li and Zn was determined on undiluted samples although a few samples required dilution for the determination of Zn. The first set was analysed on days 1 and 3 and the second set was analysed on days 2 and 4. The two sets were each analysed five times and the instrument was calibrated between each of the five runs.Flame atomic absorption analyses of the water samples were made using standard procedures and instrument settings recommended by the manufacturers. Methods for Mg and Ca required the addition of lanthanum as a releasing agent. Samples were analysed ten times consecutively for each element and the instrument was calibrated between each run. Results and Discussion Detection Limits The mean and standard deviation of the detection limits for each of the elements on separate days and a pooled mean and standard deviation are shown in Table 2. The short-term precision of the instrument is demonstrated by the relatively small deviations obtained within each day and longer term stability is illustrated by the reproducibility of detection limits determined on the two different days.Improvement in detection limits can be achieved by increasing the sample measurement integration time. The improvement is possible because the ICP-AFS background noise is “white” at the 2000-Hz modulation frequency; thus a longer measuring time decreases the magnitude of the noise as the square root of the integration time.5 The detection limits obtained with a 40-cycle integration or four times the normal measuring time are shown in Table 2. Detection limit value ANALYST JUNE 1986 VOL. 111 647 Table 2. Mean and standard deviation of detection limits in ng ml-1 determined from five runs on separate days and the pooled detection limits and standard deviations of those determinations Pooled Detection Day 1 Day 2 limit Mean Mean Mean with detection Standard detection Standard detection Standard 40-cycle Element limit deviation limit deviation limit deviation integration M g .. . . . . 1.6 0.4 1.7 0.3 1.7 0.3 0.8 Ca . . . . . . 1.4 0.3 1.4 0.3 1.4 0.3 0.7 Na . . . . . . 2.2 0.5 1.8 0.6 2.0 0.5 1 .o Cd . . . . . . 1.7 0.6 1.9 0.3 1.8 0.4 0.8 c u . . . . . . 6.4 1.3 5.9 0.6 6.2 1 .o 3.7 Fe . . . . . . 1.4.5 5.1 17.1 3.5 15.8 5.0 6.5 K . . . . . . 3.4 1.3 3.5 1.9 3.5 1.3 1.8 Li . . . . . . 0.3 0.05 0.3 0.07 0.3 0.06 0.2 Zn . . . . . . 0.9 0.2 1.5 0.2 1.2 0.3 0.7 100000, + i 4- .- yo! E 1000 m .- “‘““i :. A A 0 10 0.01 0.1 1 10 100 1000 Concentration/pg ml-1 Fig. 1. Diagram showing the relationship of signal intensit (in counts) to concentration of K and Fe.A K (766.5 nm); and d Fe (248.3 nm) are improved by about a factor of two for all elements with 40-cycle integration as expected. Propane is added to the spray chamber as an auxiliary gas to improve atom formation for some elements. Adding this “sliver” of propane6 to the plasma significantly improved the detection limits of Fe K Mg and Na by two- to five-fold under the proposed operating conditions. The addition of propane caused an increase in signal intensity for these elements without a corresponding increase in noise. The presence of propane did not affect the detection limits or signal intensities of Cd Cu or Zn. Propane increased both the signal and the noise of Ca resulting in a similar detection limit.An increased lamp current further improved the detection limits up to two-fold. However this causes the lifetime of the lamp to decrease significantly. The detection limits reported here are for multi-element determinations and can be improved by selecting optimum plasma conditions for a specific element. I I I I I Linear Working Range Diagrams showing calibration graphs were constructed to determine the linear working range for each element. Blank-corrected signal intensities for single-element solutions were plotted against element concentrations. The element concen-trations at which deviations of 5 and 10% from linearity, respectively occur are Mg 4 and 6; Ca 3 and 5; Na 8 and 9; Cd 5 and 10; Cu 15 and 35; Fe 50 and 60; K 8 and 13; Li 10 and 15; and Zn 3 and 5 pg ml-1.In this study the 5% deviation level was considered to be the limit of the linear range. The calibration graph in Fig. 1 begins to flatten at approximately 200 pg ml-1 and starts to bend towards the concentration axis at about 750 pg ml-1 or 1.5 orders of magnitude above the linear range of 50-60 pg ml-1. This curvature is typical for most of the elements studied. The curves begin to reverse at 1-2 orders of magnitude above the linear range. This folding over can result in a signal intensity that is the same for two vastly different element concen-trations. This problem is mostly circumvented by measuring the rise time of the signal by means of a sensing circuit in the detection electronic system of the ICP-AFS. If the concen-tration is out of the linear range of the element the ICP-AFS indicates a message of “signal high” (SH).A range of concentrations from three times the upper limit of the linear range to as high as three orders of magnitude above the linear range was tested and the instrument correctly indicated SH for all of the elements except K. The curve (Fig. 1) bends over sharply at a concentration of approximately 60 pg ml-1 of K. Concentrations from 20 to 75 pg ml-1 of K triggered the SH response; however concen-trations of greater than 75 pg ml-1 of K did not. Attempts to decrease significantly the severe bending of the K curve by adjusting the operating parameters of the instrument were unsuccessful. Therefore the same signal may indicate two possible concentrations of K.Two ways to help avoid this situation are to determine K in the undiluted sample and then make an additional measurement of a 1 + 99 dilution of the sample or to monitor the K signal while determining Ca Mg and Na in a 1 + 99 dilution of the sample. Monitoring the K signal using the operating conditions for Ca Mg and Na would give only an estimate of the K concentration because the K sensitivity is greatly reduced under these conditions but the signal would indicate samples with concentrations that might be above the instrument’s working range. Analytical Data Precision The precision of the method is directly related to instrument stability. The precision of the SRWS and acid mine water samples is illustrated (Tables 3 and 4) by mean and standard deviation values for ten determinations five determinations on each of 2 d.The values determined by ICP-AFS are similar to those reported for the SRWS samples. A comparison of standard deviation values for ICP-AFS calculated for each of the 2 d shows very reasonable agreement and thus good day-to-day stability (data not presented). In general within-day precision is only slightly better than day-to-day precision. The precision for the acid mine water samples analysed by ICP-AFS is poorer than that for AA analysis partially becaus Table 3. Mean and standard deviation values* for Cd Cu Fe K Li and Zn in standard reference and acid mine water. All values ng ml-1 Standard reference Cd c u Fe K water sample 1 2 3 1 2 3 1 2 3 1 2 3 1 52 61 64 65 75 80 81 84 85 86 2.7f 1.2 3 f 2 N.d.2.4f0.6 3 f 1 N.d. N.d. N.d. N.d. 11.8f 1.8 13 f 1 N.d. 7 . 9 f 1.2 9 f 2 N.d. N.d. N.d. Ned. 8.65 1.4 1 2 f 1 N.d. N.d. N.d. N.d. 2.63f 1.22 4 f 1 N.d. N.d. N.d. N.d. 7 9 f 8 8 9 f 4 11025 1 2 5 t 6 N.d. N.d. 482 f 25 503 k 12 1 0 2 f 7 1 1 3 f 7 N.d. N.d. 29f3.9 3 6 f 5 N.d. N.d. 65 f 4 N.d. N.d. 54.1 f 4.2 N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. 270 f 20 93 f 17 N.d. 499 f 20 372 f 26 N.d. 704 f 29 N.d. 188 t 14 N.d. 323 t 31 115 f 18 N.d. 530 f 16 395 f 19 N.d. 759 f 30 N.d. 214 t 22 N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. N.d. 8270 f 1010 N.d. N.d. 4370 f 290 N.d. 4740 k 580 N.d.4620 f 460 1220 f 90 2320 t 110 8190 f 470 2910 f 135 2800 f 155 4630 f 280 3670 f 235 5070 f 280 5010 f 299 4880 f 250 1260 f 75 2280 f 76 N.d. 3150 f 98 2920 f 73 N.d. 3750 f 139 N.d. 5140 t 217 N.d. 8 6 f 7 3 2 f 4 N.d. 105f10 9 3 f 1 2 N.d. N.d. N.d. 514 k 48 29.2 f 6.7 Acid mine water sample 1s N.d. 41 t 1 47 t 3 N.d. 1910t 75 1748 f 25 N.d. 58900 If 4500 56500 t 640 N.d. 8400 f 560 8830 f 45 N.d. vc N.d. 292 f 12 303 f 45 N.d. 797 t 40 747 f 7 N.d. 83900 f 8200 72oOO t 1720 N.d. 10800 f 1800 10240 f 460 N.d. LM N.d. 3 f 2 9 f 3 N.d. 2 3 f 6 3 1 f 6 N.d. 138f28 100f10 N.d. 5430f511 5580f380 N.d. RT N.d. 24 f 2 24 f 3 N.d. 1450 f 65 1340 f 20 N.d. 17110 f 1390 16800 k 570 N.d. 3670 + 307 3620 t 134 N.d.* 1 Reference value; 2 ICP-AFS value; 3 AA value. t N.d. = not determined ANALYST JUNE 1986 VOL. 111 649 ~ ~~ ~ Table 4. Mean and standard deviation values* for Mg Ca and Na in standard reference and acid mine water samples. All values in g ml-1 Standard Mg Ca Na reference water sample 1 2 3 1 2 3 1 2 3 64 64.6 4 4.3 68.3 f 2.1 N.d.t 166 4 6 180.4 k 5.5 N.d. 160 f 4 172.8 4 5.1 N.d. 80 12.3 f 0.7 12.5 k 0.4 N.d. 49.6 4 2.2 53.4 4 2.1 N.d. 32.0 4 1.1 35.0 4 1.6 N.d. 84 51.8 k 2.3 52.2 k 1.7 N.d. 86.7 4 3.2 90.3 4 3.7 N.d. 79.7 k 2.4 82.0 f 4.2 N.d. 86 28.3 f 1.1 28.4 4 1.0 N.d. 71.5 f 3.4 75.2 k- 3.9 N.d. 77.2 k 2.7 77.8 k 4.3 N.d. Acid mine water sample 1s N.d. 167.2 f 6.8 154.3 k 2.4 N.d. 402.2 4 18.4 402.7 k 5.0 N.d. 58.4 k 2.4 59.5 4 0.8 vc N.d.134.2 2 5.4 124.3 4 2.6 N.d. 399.0 4 19.4 393.8 k 2.7 N.d. 382.0 k 2.6 409.9 k 0.6 LM N.d. 63.2 4 2.6 56.4 4 0.9 N.d. 219.9 k 10.3 226.3 4 2.7 N.d. 30.9 2 1.3 34.8 k 0.5 RT N.d. 61.7 4 2.4 59.6 4 0.8 N.d. 124.9 2 7.4 123.5 f 1.7 N.d. 22.4 4 1.0 26.6 k 0.4 * 1 Reference value; 2 ICP-AFS value; 3 AA value. t N.d. = not determined. precision for AA was calculated on ten consecutive determi-nations on the same day. Relative standard deviations (data not presented) for values determined by ICP-AFS are generally less than 5%. This precision agrees well with other published multi-element water analysis techniques.7.8 Accuracy The difficulty in achieving pre-determined levels of accuracy for water analysis has been shown by several inter-laboratory studies.Studies commonly show that most laboratories do not meet the minimum requirements for accuracy for numerous if not all elements invo1ved.P-11 One such recent study of Cd, Cu Pb Ni and Zn in river water doubled the original requirement for accuracy to within 20% of the determined concentration or 10 pg 1-1 whichever was greater and most participating laboratories still did not meet the elevated criteria.8 The accuracy for the SRWS analysed by ICP-AFS in this study meets the requirements of the above study as the values are well within the 10 pg ml-1 or 20% level of accuracy. The ICP-AFS values tend to be slightly higher than the reference values; however the values are within or very near the confidence interval of the reference values (Tables 3 and 4). Values for the acid mine water determined by ICP-AFS also agree well with AA values for samples Tables 3 and 4).Analysis of 156 ground-water samples from Pennsylvania by the proposed method gave values that agree very well with flame and electrothermal AA values determined for seven elements. Correlation coefficients between the two sets of Pennsylvania data are as follows Ca 0.99; Mg 0.99; Na 0.97; Cu 0.99; Fe 0.99; K 0.98; and Zn 0.99. Inter-element Cross-talk Inter-element cross-talk can occur when a signal from one channel spills over into another channel. This can happen if the detector of one module detects light from or produced by the hollow-cathode lamps of other modules. To check the existence of cross-talk with the ICP-AFS individual solutions of 1000 pg ml-1 of Ca Mg and Na were aspirated separately into the ICP-AFS and the base-line signals of the other channels were monitored.No change in the base-line signal intensities was observed. The same procedure was followed for the second group of elements in this study using individual 1000 pg ml-1 solutions of Cd Cu Fe K Li and Zn. Again, the aspiration of these solutions had no effect on the base-line signals of the other elements. Thus ICP-AFS analyses are free from cross-talk interferences at concentration levels above those commonly observed in natural water analysis. Conclusion Surveys to determine the range of element concentrations in various natural water sample types reveal concentration levels that are within the detection limit capabilities of the ICP-AFS technique.12-14 Thus the ICP-AFS method is an efficient procedure for the determination of trace and major elements in a wide variety of natural water samples.The procedure has good precision and accuracy in a multi-element system that compares favourably in sensitivity and precision to !CP-AES and flame AA techniques. The analysis of 100 samples per day per analyst can be easily accomplished using this procedure. The authors thank T. T. Chao for suggestions for the study and for his assistance in the preparation of the manuscript, Walter Ficklin for his help in collecting the acid mine water samples and for supplying the samples and AA analysis for the Pennsylvania water and Elwin Mosier and Arthur Hubert for their help in the preparation of the manuscript. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. References Miller W. R. Geol. Surv. Can. Econ. Geol. Rep. 1979 31, 479. Demers D. R. Busch D. A. and Allemand C. D. Am. Lab. 1982 14 167. Meier A. L. and Bigelow R. C. U.S. Geol. Surv. Open-file Rep. 1984 84-698. Skougstad M. W. and Fishman M. J. AWWA Technical Conference Proceedings 1974 XIX-1. Demers D. R. and Allemand C. D. Anal. Chem. 1981,53, 1915. Demers D. R. Spectrochim. Acta Part B 1985 40 93. Garbarino J. R. and Taylor H. E. Appl. Spectrosc. 1979, 17 84. Floyd M. A. Halouma A. A. Morrow R. W. and Farrar, R. B . Am. Lab. 1985 3 84. Analytical Quality Control (Harmonised Monitoring) Commit-tee Analyst 1985 110 1. Dybcznski R. Tugsavul A. and Suschny O. Analyst 1978, 103 734. Ellis A. J. Geochim. Cosmochim. Acta 1976 40 1359. Kopp J. F. and Kroner R. C. “Trace Metals in Waters in the United States (Oct. 1 1962 to Sept. 30 1967),”Federal Water Pollution Control Administration Water Quality Surveillance Program Federal Water Pollution Control Administration, Cincinnati 1970 p. 32. Meranger J. C. Kunnath S. S . and Chalifoux C. Anal. Chem. 1979 13 707. Shvartsev S. L. Udodov P. A. and Rasskazov N. M. J . Paper A51356 Geochem. Exp. 1975 4 433. Received October 7th 1985 Accepted December 3Ist 198

ISSN:0003-2654    

DOI:10.1039/AN9861100645

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST JUNE 1986 VOL. 111 651 Determination of Lead in Blood by Atomic Absorption Spectrometry with Electrothermal Atom isation Ian L. Shuttler and H. Trevor Delves Trace Element Unit University Chemical Pathology and Human Metabolism University of Southampton, Southampton General Hospital Southampton SO9 4XY UK A method is described for the routine determination of lead in whole blood by electrothermal atomisation from a L'vov platform and atomic absorption spectrometry. The sample preparation is a simple 1 + 19 dilution with a diluent containing a mixture of NH3 solution NH4H2P04 and (NH4)2H2EDTA. Further chemical modification is achieved using in situ oxygen ashing during electrothermal sample decomposition at 550 "C. Desorption of oxygen at 950 "C without losses of lead gives tube lifetimes of 200-250 firings.The calibration graph established with matrix-matched standards and using integrated absorbance measurements is linear for concentrations up to 5 pmol 1-1. The detection limit is 8.3 x 10-12 g or 0.08 pmol 1-1 of lead in blood. The within- and between-run precisions are 6.9% and 7.3%' respectively at 0.5 pmol 1-1 of lead and the accuracy of the method is demonstrated by the excellent agreement with the results of micro-sampling flame atomic absorption spectrometry and with group mean values in external quality assessment programmes. Keywords Lead determination; blood; electrothermal atomisation atomic absorption spectrometry; L'vov platform; oxygen ashing Blood lead concentrations provide probably the best index of recent exposure to this element and have been used for the biological monitoring of populations exposed to environmen-tal lead in many countries.l.2 One source of environmental lead tetraalkyllead was to be reduced in the UK from 0.40 to 0.15 g 1-1 in petrol by the end of 1985.3 The effect that this will have on the exposure to lead of the general population is to be determined by monitoring in the authors' laboratory lead concentrations in blood specimens taken from selected groups from different locations in the UK over the period 1984-87.The magnitude of the analytical challenge posed by this work is apparent when one considers that lead added to petrol is only one of many sources of lead and possibly contributes only 2&30°/0 of the total blood lead.4.5 Clearly anaIyticaI accuracy must be maintained within limits of 3-5% at the median lead level for the UK population of 0.5 pmol l-1 (100 pg 1-1) over the whole of the 4-year study period in order to detect any change in the population's blood lead levels.The method currently used in the authors' laboratory for such work microsampling flame atomic absorption spec-trometry6 (MSFAAS) achieves this accuracy routinely but in view of the importance of the study it was considered necessary to be able to demonstrate minimum temporal changes in analytical bias. The analytical protocol devised for this purpose7 consists of a novel internal quality control (IQC) procedure re-analysis within our laboratory using a different method of at least 20% of the ca. 3000 samples collected each year re-analysis of at least 10% of the survey samples by a different laboratory using a different method and participa-tion in external quality assessment (EQA) programmes.The method described here was developed specifically for the within-laboratory re-analysis aspect of this protocol. It is based on atomic absorption spectrometry with electrothermal atomistion (ETA-AAS) using a L'vov platform and Zeeman-effect background correction. The physico-chemical prin-ciples of this method are sufficiently different from those of MSFAAS that the inherent errors of both methods are also likely to be different. The sample preparation is simple using a single diluent containing NH3 solution (NH4)2H2EDTA and NH4H2P04, which gives a diluted homogeneous blood solution that is stable for up to 24 h.Further chemical treatment of in situ oxygen ashing during thermal decomposition of the sample in the electrothermal atomiser eliminates matrix effects that could occur with the processing of large numbers of samples. The method has a characteristic mass of 14-16 pg per 0.004 A s which is in agreement with other published data,Q and is partly automated giving an analytical throughput of 40 samples per day. The between-batch precision over the concentration range 0.2-1.5 pmol 1-1 (40-310 pg 1-*) ranges from 3 to 10% and is 7% at the median blood lead level of 0.5 pmol l-1 (100 pg 1-1). The method gives excellent agreement with the MSFAAS procedure; the equation of the line is [ETA-AAS] = 1.029 [MSFAAS] -0.029 r = 0.997 over the range 0.1-4.0 pmol 1-1 (20-830 pg l - l ) and the accuracy of both methods is established by good performance in EQA schemes.Experimental Instrumentation A Perkin-Elmer Zeeman/5000 atomic absorption spec-trometer equipped with an HGA-500 graphite furnace an AS-40 autosampler and a Model 056 chart recorder was used. The instrumental conditions are given in Table 1. Atomic signals were measured with a Perkin-Elmer 3600 Data Station using HGA graphics software (Part No. VE16-2407/D) which allowed the simultaneous display of analyte and background signals and calculations of both integrated and peak absorbance data. Pyrolytically coated furnace tubes and L'vov platforms were employed. Trace metal-free polycarbonate tubes (Tak Lab) were used for sample preparation.Reagents All of the water used to wash the laboratory ware and to prepare solutions and standards was glass distilled and then further purified by de-ionisation with a Milli-Q system (Millipore). All reagents were of Aristar Spectrosol (BDH Chemicals) or Specpure (Johnson-Matthey) grade unless stated other-wise 652 ANALYST JUNE 1986 VOL. 111 ~ Table 1. Instrument and furnace conditions for the determination of lead in whole blood Zeeman/5000 spectrophotometer : Hollow-cathode lamp current . . . . 10 mA Wavelength. . . . . . Slitwidth . . . . . . Background correction . . Measurement mode . . Integration time . . . . Chartrecorder . . . . HGA-500 graphite furnace Samplevolume . . . . . . . . . . 283.3 nm .. . . . . 0.7 nm low . . . . . . Zeeman . . . . . . 2 s . . . . . . Integrated absorbance calculated by HGA software 40 mm min-1; 0.5 V full-scale Pyrolytically coated graphite tubes fitted with a . . . . . . . . . . . . L'VOV platform . . . . . . lop1 Step Parameter TemperaturePC . . . . Ramp/s . . . . . . . . Hold/s . . . . . . . . Flow-rate of internal gas, Ar/mlmin-l . . . . . . Flow-rate of alternative gas, Recorder magnet on . . Read . . . . . . . . Base-line offset correction . . 02/mlmin-1 . . . . . . 1 2 3 4 5 6 7 . . 150 550 950 950 1900 2200 20 . . 1 1 10 0 0 1 5 . . 30 30 30 10 2 2 5 . . 300 300 0 0 300 300 . . 50 . . -4 X 5 Ammonia solution 14 M. Ammonium dihydrogen phosphate. Chelex-100 chelating ion exchange resin.Bio-Rad Labora-Ethylenediaminetetraacetic acid diammonium salt. Hydrochloric acid 11 M. Lead nitrate standard solution 1 mmoll-l. Nitric acid 16 M. Triton X-100. tories. Stock standard solutions To a series of ten 25-ml calibrated flasks add 5 ml of de-ionised water and 0.5 ml of 16 M nitric acid followed by 0 0.25 0.5, 0.75 1.0 1.5 2.0 2.5 3.0 and 4.0 ml of 1 mmol 1-1 lead nitrate standard solution. Dilute to volume with water and mix well. These solutions contain 0,10,20,30,40,60,80,100,120 and 160 pmol l-1 of lead respectively. Working standard solutions To a series of ten 20-ml calibrated flasks add 0.50 ml of each of the above stock standard solutions dilute to volume with water and mix well. These solutions contain 0,0.25,0.5,0.75, 1.0 1.5 2.0 2.5 3.0 and 4.0 pmol 1-1 of lead respectively, and 0.7 M nitric acid.Prepare freshly each week. For routine use transfer approximately 2 ml of these working solutions into plain 2-rnl stoppered polycarbonate tubes. These are replaced daily. Autosampler wash solution Dissolve 1 ml of Triton X-100 in water add 2 ml of 11 M hydrochloric acid and dilute to 1 1 with water. Chemical modifier - diluent solution Dissolve 1.667 g of ammonium &hydrogen phosphate in water add 5 ml of 14 M ammonia solution and dilute to 500 ml with water. Pass this solution through a column of Chelex-100 chelating ion-exchange resin in the ammonium form at a flow-rate of 0.5 ml min-I then add 0.556 g of ethylenedi-aminetetraacetic acid diammonium salt. Store in an acid-washed polypropylene bottle.This solution contains 0.14 M NH3,0.003 M (NH4)2H2EDTA and 0.029 M NH4H2P04. Collection of Blood Samples Blood samples were collected by venepuncture and trans-ferred into 2-ml trace metal-free polycarbonate tubes contain-ing 3.5 mg of K2H2EDTA as anticoagulant. All samples were stored at -20 "C until required for analysis. Sample Preparation Transfer triplicate 900-pl aliquots of chemical modifier -diluent solution into plain trace metal-free polycarbonate tubes and add 50 p1 of the zero lead working standard solution, followed by 50 pl of well mixed whole blood. Stopper the tubes and mix initially by gentle inversion followed by 5-10 min on a roller mixer. These diluted samples are stable for up to 24 h I when stored at 4 "C.Use the conditions in Table 1 to determine the concentra-tion of lead in the samples. Establish a calibration graph by adding 50-p1 aliquots of the lead working standard solutions to plain polycarbonate tubes containing 900 pl of the chemical modifier - diluent solution and 50 pl of bovine blood with a known low lead concentration and mix as for the samples. Prepare similarly internal quality control samples. Analyse singly each of the triplicate sample solutions and run a minimum of three internal quality control samples after every batch of five samples. Calculate the blood lead concentrations from the integrated absorbance measurements and the calibration graph. The chart recorder trace is used as a visual check that the analysis is proceeding satisfactorily.If the results for the internal quality control samples do not fall within previously established ranges repeat the analysis. The rigorous internal quality control protocol established within this laboratory is to be reported in detail el~ewhere.~ Results and Discussion Sample Preparation The sensitivity of ETA-AAS for lead allows blood specimens to be diluted between 1 + 5 and 1 + 19 to give analyte signals that are within the linear range of calibration graphs.+" The role of the diluent is particularly important with large numbers of specimens that may be prepared hours in advance of the analysis. Diluents such as water or dilute acids give rise to ANALYST JUNE 1986 VOL. 111 0.3 0.2 0.1 0 C m n o :::I * \ I 171 600 700 800 900 100011001200 600 700 800 900 1000 11001200 0.2 I- \ \ C E I \\ 1-i L b L J 8tO 9iO Id001 lb012b0 ' 6;O 7b0 8;)O Secondary ashing temperature/"C Fig.1. Effect of secondary ashing temperature on lead atomic signals from various matrices after initial ashing at 550 "C in oxygen. Open symbols 0.31 ng of aqueous lead; solid symbols 0.32 ng of blood lead. A Water diluent; B NH4H2P04 (30 pg) + Mg(N03)2 1 pg). Atomisation off the platform at 1900 C from 500 "C maximum power heating zero gas flow t30 pg); c NH4H2PO4 (30 !@); D7 Mg(NO$2 (30 pg); and E7 Pt (Iv) Table 2. Effect of carbon monoxide on the lead atomic signal. Whole blood containing 3.09 pmoll-* of lead diluted 1 + 19 with NH4H2P04 solution was ashed in oxygen at 550 "C followed by desorption of oxygen at 950 "C and atomisation at 1900 "C from a L'vov platform under zero gas flow Integrated Peak Internal gas absorbance absorbance Argon .. . . . . . . 0.077 0.223 10% Coinargon . . . . 0.027 0.057 Argon . . . . . . . . 0.080 0.227 ____ "'y Aqueous 653 slow precipitation of red cell membranes and produce a turbid heterogeneous suspension after about 15 min. The addition of detergents such as Triton X-100 eliminates this by lysing the red cells but can cause problems because of variable diffusion of the diluted solution into the graphite. Dilute aqueous ammonia solutions give complete lysis of red cells and within 2 min produce a clear homogeneous solution that does not diffuse significantly into the graphite. During development work it was found that a 1 + 19 dilution of whole blood with 1% V/Vammonia solution would give an acceptable working calibration range for the method, and complete lysis.At the resultant pH of 8.5-9.0 it is necessary to add (NH4)2H2EDTA to prevent losses of lead by precipitation and chemisorption. The concentration of (NH&H2EDTA chosen of 0.003 M was ten times greater than that commonly used as an anticoagulant. The addition of NH4H2P04 as a chemical modifier enables a high ashing temperature to be used in the graphite atomiser without loss of lead. The concentration selected (0.029 M) was the minimum amount of NH4H2P04 necessary to ensure no loss of 0.32 ng of lead in a blood matrix heated to 950 "C. Combining these two reagents in 1 % V/V ammonia solution gives a single chemical modifier - diluent which produces a clear homogeneous solution within 2 min when 50 pl of blood are added to 900 pl of diluent plus 50 p1 of zero lead working standard in a polycarbonate tube and mixed on a roller mixer.The solutions of diluted blood samples are stable for up to 24 h when stored at 4 "C and can easily be left for 3-4 h in the autosampler cups. Other chemical modifiers were examined during this study (Fig. 1). During the development of this method measure-ments were made using both peak and integrated absorbance modes. Fig. 1 shows peak absorbance data but identical conclusions could be drawn from the integrated absorbance data. Magnesium nitrate was used in conjunction with ammonium dihydrogen phosphate by Pruszkowska et aZ.9 and, although this mixture enabled an ashing temperature of 950 "C (el Aqueous " I I', (9) Aqueous I I I I 1 I 0 1 .o 2.0 3.0 0 1 .o 2.0 3.0 Desorption time at 950 "C/s 0.15 Fig.2. Effect of chemical modifiers and oxy en ashing at 550 "C on lead atomic signals; 0.31 ng of a ueous lead 0.32 ng of blood lead. -Oxygen ash at 550 "C followed by desorption of oxy en at 950 "C in argon; (c) ( d ) char at 550 "C in argon followed by further char at 950 in argon; (e) m oxygen ash at 550 "C followed%y flushing with argon at 550 "C; and (g) ( h ) char at 550 "C in argon followed by further char at 550 "C in argon. Atomisation at 1900 "C off the platform from 550 "C maximum power heating zero gas flow Aqueous dilution; - - - dilution with N$H2P04 (30 pg on platform); and dilution w i 8 Pt(1V) (1 pg on platform).(a) 654 (a) Argon ANALYST JUNE 1986 VOL. 111 ( b ) 10% V/VCO in argon to be used it offered no improvement over the sole use of ammonium dihydrogen phosphate. Platinum(1V) showed great promise as a chemical modifier with aqueous solutions of lead allowing an ashing temperature of 1050 "C before loss of lead. Unfortunately this effect could not be produced in the presence of blood presumably because the blood proteins complexed with the platinum thus preventing the formation of a lead - platinum compound during the initial thermal pre-treatment. Further work in this area is in progress. Use of Oxygen Ashing Pyrolysis of a blood matrix within an electrothermal atomiser in the presence of argon at 500-600 "C produces after a number of firings a carbonaceous residue.12 This causes several problems (1) irreproducible sample deposition in the atomiser (2) variable rates of atomisation from the residue and (3) partial occlusion of the light beam.The combination of these effects leads to a rapid reduction in sensitivity and precision. This residue is resistant to removal by use of a high-temperature "clean" stage in excess of 2700 "C. The use of oxygen ashing at 500-600 "C eliminates completely the above effects and in addition reduces the molecular signal on subsequent atomisation.12 The optimum temperature for chemisorption of oxygen on the active carbon sites as the surface oxide is 500-600 "C and the optimum desorption temperature is 950 "C. 13 Subsequent atomisation without desorption of oxygen leads to severe surface attack and a short graphite tube lifetime owing to the rapid formation and removal of carbon monoxide.Inspection " I 0.2 1 L 0 0.5 1.0 1.5 0 0.5 1.0 1.5 Timeis Fig. 3. Effect of carbon monoxide on lead atomic signals. (a) ( b ) 0.32 ng of blood lead; and (c) ( d ) 0.31 ng of aqueous lead. Diluted with NH4H2P04 solution (30 pg on platform). Atomisation at 1900 "C off the platform from 950 "C maximum power heating zero gas flow 0.3 1 I I I I I I 0 10 20 30 40 50 Desorption time at 950 "Cis Fig. 4. Effect of time of desorption of oxygen from graphite at 950 "C. A peak absorbance; and B integrated absorbance. 0.43 ng of blood lead 30 pg of NH4H2P04 atomisation at 1900 "C off platform from 950 "C maximum power heating zero gas flow of a tube after 30 firings with oxygen ashing followed by a short (10 s) argon flushing step at 550 "C showed complete removal of the pyrolytic coating and almost total destruction of the platform.Desorption for 30 s at 950 "C produces negligible attack by oxygen on the graphite during subsequent atomisa-tion and gives tube lifetimes in excess of 200-250 firings. The use of an oxygen desorption step at 950 "C does however, require the use of chemical modification to overcome the loss of more volatile analytes at this elevated temperature. As already discussed NH4H2P04 was selected as the chemical modifier for this method. Eaton and Holcombel4 questioned the need for the addition of phosphate salts when using oxygen ashing for the determination of lead in blood but as Figs.1 and 2 show losses of lead occurred at ashing temperatures in excess of 650 "C in the absence of NH4H2P04. The use of oxygen and the presence of a blood matrix has a significant effect on the magnitude of the subsequent char - desorption temperature that may be used before loss of lead occurs. Fig. 1 shows a steady fall in the lead atomic signal with increasing desorption temperature for aqueous lead solutions and in the presence of various chemical modifiers. Large differences were observed between the lead atomic signals from either an aqueous or a blood matrix depending on the chemical modification and thermal treatment (Fig. 2). The atomic signals for both aqueous and blood lead show similar thermal shifts towards later appearance times following incomplete 0.09 0.08 3 0.07 6 0.06 $ 0.05 2 0.04 I 0.03 fn 0 e D Q, (5, 4-z 0.02 0.01 0 -Concentration of leadipmol 1-1 Fig.5. Effect of species on blood lead calibration graphs. A Bovine blood; B human blood; and C avian blood 30 I I A 0 1 .o 2.0 3.0 4.0 Concentration of leadipmol I-' Fig. 6. Precision of blood lead analysis. A Within-batch precision (12 > n > 9); and B between-batch precision (36 > n > 20 ANALYST JUNE 1986 VOL. 111 655 4 s r I 0 - -5 8 3 a a 3.0 w cn v) 3 8 0 .-.- 4- 2.0 c 8 a 8 8 -0 al -1.0 0 - rn 0 1 .o 2.0 3.0 4.0 Blood lead concentration using MSFAASipmol I-’ Fig. 7. Comparison of blood lead results using ETA-AAS and MSFAAS.y = 1.029~ - 0.029; Y = 0.997; ~1 = 200. removal of chemisorbed oxygen prior to atomisation as has been found by others.13 These atomic signals and ashing curves indicate that within a graphite electrothermal atomiser oxygen plays a significant role in atom formation subsequent to its use as an ashing aid. As part of this study we examined the effect of 10% V/V carbon monoxide in argon on the lead atomic signals. Fig. 3 and Table 2 show a drastic reduction in the lead atomic signal in the presence of carbon monoxide. Further evidence of the adverse effect of carbon monoxide on the lead atomic signals was found on investigating the length of time necessary to desorb oxygen at 950 “C. A steady increase in the lead atomic signal was found with desorption times up to 30-40 s after which no further increase was seen (Fig.4) suggesting that during atomisation the production of carbon monoxide from chemisorbed oxygen on the graphite depletes the free atom concentration possibly by reversal or suppression of the reaction This emphasises the crucial importance of a desorption stage in any furnace programme that uses oxygen ashing. Optimisation of Furnace Parameters The advantages of using L’vov platforms have been well documented elsewhere.15J6 The parameters for the so-called “stabilised temperature platform furnace”15J7 were critically examined during the development of the method using the microcomputer and associated software to monitor accurately the shapes of the atomisation signals.It is important to optimise all furnace parameters in the presence of the sample matrix because the appearance time and shapes of the lead signals are clearly dependent on the matrix (Fig. 2). Simul-taneous monitoring of the background and analyte signals showed that the use of oxygen ashing reduced the background absorbance to less than 0.05 A. This is easily within the capabilities of both the Zeeman and conventional deuterium source background correction systems. Halls18 showed that critical examination of the functions and steps in a furnace programme can lead to major savings in r 4.0 - 0 5 3j B 2 3.0 w m v) 3 8 0 .-.-c 2.0 I c 8 al 8 8 g 1.0 -0 -Blood lead concentration mean EQA valuesipmot I Fig.8. Performance in external quality assessment (EQA) pro-grammes. y = 1.036~ - 0.039; Y = 0.998; n = 120. the time of analysis. All of the steps in the developed programme have been carefully examined and optimised. It should be noted that the within- and between-batch variation in the thermal performance and lifetime of the furnace tubes and platforms may require some minor modifications to be made to the drying step. Analytical Performance A number of workers9.16 have reported that the use of a L’vov platform enables a calibration to be made using simple aqueous calibrants for the determination of lead in blood. In the development of the method described here it was not possible to achieve an acceptable level of analytical accuracy using aqueous calibrants and matrix-matched standards were essential.The calibration graph was constructed by additions of aqueous standards to a bovine blood with a known low lead content and was linear for concentrations up to 5 pmol 1-1 using integrated absorbance measurements. The validity of using a calibration graph established with a bovine blood matrix for the determination of human blood lead concentra-tions is apparent from Fig. 5 which shows no significant difference between the calibration graphs for human bovine and avian blood matrices. Precision data are given in Fig. 6. The within-run precision was calculated from the IQC results over a period of 3 months. Frequently a within-batch precision of less than 1% was found. Measurements made at or near the blank value gave a limit of detection (30) of 0.08 pmol 1-1 and a limit of determination (100) of 0.25 pmol 1-1.The between-batch precisions at these two limits were 16% and lo% respectively. The established method for blood lead determinations within this laboratory is MSFAAS and a number of comparisons were made between MSFAAS and the method described. The results are shown in Fig. 7; the equation of the line is [ETA-AAS] = 1.029 [MSFAAS] -0.029 r = 0.997. This indicates a relative bias of 3% and a fixed bias of -0.03 pmol l-1 between the two methods over the range 0 . 1 4 pmol l-l. This was considered acceptable for the purposes of our studies. Some external appraisal of the accuracy of the method described here was obtained by participation in a number of national and international EQA programmes.The results are shown in Fig. 8. The equation of the line is [ETA-AAS] = 1.036 [EQA] -0.039 r = 0.998 656 ANALYST JUNE 1986 VOL. 111 Conclusion The method developed demonstrates that acceptable accuracy and precision can be achieved in the routine use of ETA-AAS if careful attention is paid to the chemistry of the system under consideration. We thank Professor B. E. Clayton for her interest in and encouragement of this work Mrs. S. Diaper and Mrs. J. North for the blood lead analyses by MSFAAS and Miss D. Pain of the Edward Grey Institute of Field Ornithology for providing samples of avian blood. This work which forms part of a larger project concerned with environmental lead monitoring, is funded by the Department of the Environment.References 1. CEC Directive Ref. 77/312/EEC Off. J . Eur. Comrnun. 28 April 1977 L105/10-17. 2. Department of the Environment “European Community Screening Programme for Lead United Kingdom Results for 1979-80,” Pollution Report No. 10 HM Stationery Office, London 1981. “The Neuropsychological Effects of Lead in Children A Review of Recent Research 1979-1983,” Medical Research Council London 1984. “Lead and Health,” Report of a DHSS Working Party on Lead in the Environment HM Stationery Office London 1980. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. “Isotopic Lead Experiment Status Report July 1982,” EUR 8352 EN CEC Joint Research Centre Ispra Establishment, Ispra Varese Italy 1982. Delves H. T. Analyst 1970 95 431. Delves H. T. et al. in preparation. Slavin W. “Graphite Furnace AAS-A Source Book,” No. 0993-8139 Perkin-Elmer Norwalk CT 1984. Pruszkowska E. Carnick G. R. and Slavin W. Ar. Spectrosc. 1983 4 59. Fernandez F. J. Clin. Chem. 1075 21 558. Hinderberger E. J. Kaiser M. L. and Koirtyohann S. R., At. Spectrosc. 1981 2 1. Delves H. T. and Woodward J. At. Spectrosc. 1983 55, 946. Salmon S. G. Davis R. H. Jr. and Holcombe J. A. Anal. Chem. 1981 53 324. Eaton D. K. and Holcombe J. A. Anal. Chern. 1983,55,65. Slavin W. Manning D. C . and Carnick G. R. At. Spectrosc. 1981 2 137. Ottaway J. M. At. Spectrosc. 1982 3 89. “Analytical Methods for Furnace AAS,” Publication B332, Perkin-Elmer Norwalk CT 1984. Halls D. J. Analyst 1984 109 1081. Paper A51402 Received November 4th 1985 Accepted January 15th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100651

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST JUNE 1986 VOL. 111 657 Determination of Aluminium in Human Tissues and Body Fluids by Zeeman-corrected Atomic Absorption Spectrometry* Jan Rud Andersen** and Susanne Reimert Royal Danish School of Pharmacy Department of Chemistry AD 2 Universitetsparken DK-2 100 Copenhagen Denmark A procedure is described for the determination of aluminium in samples of human origin. It is based on Zeeman-corrected atomic absorption spectrometry and the method of standard additions is used for the quantitation. Plasma and urine samples were analysed directly after dilution and tissue samples after wet digestion with nitric acid. For plasma samples the precision was approximately 5% (relative standard deviation) at the 50 yg I-' level and the accuracy of the method was evaluated by analysing the NBS reference material bovine serum.For tissue samples the accuracy and precision were determined from recovery experiments. Examples of aluminium concentrations found in human plasma urine bone liver and brain are given. Keywords Aluminium determination; Zeeman graphite furnace; atomic absorption spectrometry; human tissue and body fluids; renal failure patients The biological effects of aluminium have recently increased considerably in interest. In 1976 aluminium was implicated as the cause of the clinical conditions dialysis ecephalopathy and dialysis osteomalacia observed in renal failure patients under-going long-term haemodialysis. 1 In these patients aluminium accumulates in the body causing the above-mentioned disorders in the brain and in the skeleton.At first the aluminium was believed to originate from the water used in preparing the dialysis baths ,2 but after using decontamination procedures such as reverse osmosis and ion exchange rou-tinely for a number of years in purifying the water in dialysis centres it has become obvious that water is not the only agent supplying the aluminium. In the recent past it has been shown that the aluminium-containing phosphate binders given orally to dialysis patients in order to keep their plasma phosphate low are absorbed to a small extent from the intestine,3 and it is now believed that this contribution of aluminium is just as serious as the contribution from the dialysis baths if not more so. There are no adequate alternatives to aluminium-containing phosphate binders so far; the water can be effectively purified.With this background the search for reliable procedures for the determination of aluminium in human matrices is easily justified and several proposals have been made4-13. However as the "normal" values for tissue and body fluid aluminium concentrations have continued to decrease with time and as yet are known with little certainty,l4 it has become increasingly clear that the sensitivity of the graphite furnace is called for in the determination of aluminium concentrations in the range normally encountered in samples of human origin. Most of the graphite furnace methods described both emission and absorption methods have dealt with the determination of aluminium in blood plasma or serum and only a few procedures concerning aluminium in human soft and hard tissue have appeared.15J6 This is unfortunate as doubts have arisen about the value of plasma or serum aluminium concentration as a monitor for the body burden of the element." In this paper we present a graphite furnace atomic absorption spectrometric procedure employ-ing Zeeman-effect background correction and we give examples of its applicability both to serum and plasma samples and to human brain liver bone and urine samples.* Presented at the Colloquium Spectroscopicurn Internationale ** To whom correspondence should be addressed. XXIV Garmisch-Partenkirchen September 1985. Experimental Instrumentation A Perkin-Elmer Zeeman 5000 atomic absorption spec-trometer equipped with a Perkin-Elmer AS-40 autosampler was used.The atomisation signals were displayed on a Perkin-Elmer R100-A recorder and their areas printed out on a Perkin-Elmer PRS-10 printer. Pyrolytically coated graphite tubes with platforms of solid pyrolytic graphite inserted were used throughout. The instrumental conditions are given in Table 1. Contamination Control Contamination is a serious problem when dealing with samples containing low concentrations of aluminium because of the ubiquitous nature of the element. Therefore all utensils sample cups pipette tips sample containers etc., were carefully decontaminated before use and the use of glassware was generally avoided. The cleaning procedure consists in soaking the items in 4 M nitric acid for at least 1 week followed by washing with copious amounts of Milli-Q water (see below).Another 1-week soaking cycle in Milli-Q water affords contamination-free surfaces as determined by leaching experiments with 4 M nitric acid 0.1 M Na2EDTA and low-aluminium human plasma in combination with the atomic absorption spectrometric procedure described here. Certain items have to be dried before use e.g. blood collection tubes, and this was done using a class 100 laminar air flow clean bench (TL 2448; Holten LaminAir Allergd Denmark). Sample Collection The samples originated from hospitals in the Copenhagen area and were collected by the medical staff. Decontaminated containers were used and the knives and cannulae were made of aluminium-free stainless steel. The medical staff were informed about the importance of contamination control prior to sampling.The samples were kept at -22 "C until analysed. Reagents Nitric acid was of Suprapur quality Triton X-100 was of scintillation grade and methanol and chloroform were analy-tical-reagent grade all purchased from E. Merck (Darmstadt, FRG). A certified 1 g 1-1 aluminium reference solution (Titrisol; E. Merck) was used and aliquots of this were dilute 658 ANALYST JUNE 1986 VOL. 111 Table 1. Instrumental conditions for the determination of aluminium in human matrices Table 2. Tube performance quantitation of plasma aluminium as function of tube lifetime Wavelength . . . . . . 396.2 nm Spectral bandpass . . . . 0.7nm Lampcurrent . . . . . . 20mA Samplevolume. . . . . . 2 0 ~ 1 Graphite furnace programme* : Step Temperature/"C Ramp/s Dry1 .. . . . . 120 30 Dry11 . . . . . . 200 20 Aluminium concentration f s.d./ Correlation Determination No. coefficient * 1 49.5 f 1.5 0.9994 2 46.2 f 1.4 0.9998 0.9998 3 44.0 f 0.7 Hold/s 4 36.6 f 1.2 0.9995 20 5 36.8 f 1.1 0.9997 10 Aged tube 33.8 f 0.9 0.9997 . . . . . . Char 1400 40 30 * From regression analysis. Atomise . . . . . . 2500 O t 7 Clean . . . . . . 2700 1 2 Cool . . . . . . 20 3 5 * The internal argon gas flow was stopped and the Zeeman background correction was on during atomisation. As temperatures may vary slightly between instruments for a given setting they should be regarded as approximate values only. t Maximum power heating. with a solution containing nitric acid and Triton X-100 (see below) to yield working standards.Milli-Q water which is a type I ultrapure water prepared using a Milli-Q deionisation unit (Millipore Bedford MA USA) was used throughout. so however for the other matrices dealt with here; for tissue and urine samples background correction is mandatory. Mg(N03)2 is the recommended matrix modifier for aluminium determinations but we add none as it is endogenous to the samples. Biological materials contain magnesium (in plasma, for example it is in the millimolar range) and nitrate is added with the diluent or via the digestion. For tissue samples high in aluminium which are diluted perhaps 100-fold the amount of Mg(N03)2 present is probably not sufficient to act as a matrix modifier. On the other hand no losses of aluminium have been observed at 1400 "C.6J9 Sample Pre- treatmen t Tube Conditioning Plasma and urine samples are analysed directly after dilution (normally 1 + 1 or 1 + 3) with a solution 10-3 M in nitric acid and 0.1% with respect to Triton X-100.A solution containing the same ingredients but 10 times more acidic was proposed by Osters as a diluent for plasma samples. We found that protein precipitation occurs when the latter solution is used, giving rise to analytical signals that change with time. This is avoided by decreasing the content of nitric acid without any loss in analytical performance. Soft tissue e.g. post mortem brain and liver samples is freeze-dried and subsequently decomposed under pressure as follows 100 mg of dried tissue and 2 ml of concentrated nitric acid are placed in a 25-ml Uniseal decomposition vessel (Uniseal Haifa Israel) which is then closed and heated at 130 "C for 3 h.When cold the clear digest is transferred into an acid-rinsed polystyrene tube and diluted to 10 ml with Milli-Q water. Small liver biopsies from living individuals (dry mass 10-20 mg) are conveniently digested directly in the poly-styrene tube with 1 ml of concentrated nitric acid at 70 "C overnight. Bone samples are washed with methanol and dried before defatting with chloroform. The dry defatted samples are digested in polystyrene tubes with concentrated nitric acid (1 ml per 100 mg) at 70 "C overnight. Digested tissue samples normally have to be diluted further in order to be within the linear range defined by the instrumental conditions and this is done with the same solution as is used for dilution of plasma and urine samples.Results and Discussion Instrument Settings The furnace programme given Table 1 is the result of a compromise between analysis time and graphite tube lifetime. It is our experience that with this rather slow programme up to 300 firings are achievable before the tube has to be discarded because of lack of sensitivity or physical breakdown. The programme adheres to the "stabilised temperature platform furnace" (STPF) concept as proposed by Slavin et aZ.18 The charring step at 1400 "C for 30 s is sufficient to reduce the organic content of the samples to such an extent that virtually no background signal is present in plasma samples.This is not The pyrolysed tube equipped with a L'vov platform is superior to unpyrolysed tubes in terms of both sensitivity and lifetime; this has been noted before.8-'0712J3 However the tube has to be thoroughly conditioned before reliable results are obtained. When starting with a fresh tube we initially go through a conditioning sequence of heating four times to 2650 "C with a ramp of 60 s (approximately 44 "C s-1) and twice to the same temperature with a 2 s ramp (1315 "C s-1) and 5 s hold as proposed by Slanina et ~ 1 . 2 0 This is however not sufficient to yield a low tube blank. Next a series of firings with pure diluent (see Experimental) is carried out until the blank signal is below 0.002 A s. Then it is advisable to run a real sample several times as in our experience the graphite tube still has not reached a stable level where accurate results are obtained.This is illustrated in Table 2 which shows a sequence of determinations on the same plasma sample with a new tube treated as mentioned above. The quantitation is effected by means of two standard additions and although fair correlations are obtained for all determinations the value found decreases before it finally reaches a stable level. Once this level has been reached the values obtained change very little although the tube loses some sensitivity with time. The last result given in Table 2 is for the same plasma sample run just before the tube was discarded. Analytical Line For aluminium determinations by means of Zeeman-effect atomic absorption spectrometry the 396.2-nm line is to be preferred over the 309.3-nm line normally used.The decrease in sensitivity (of the order of 10%) is compensated for by the far better linear range obtained. With the experimental conditions described here the 309.3-nm line gives a linear response up to approximately 0.4 A s whereas the 396.2-nm line gives linearity up to close to 1 A s. This is of special importance when the quantitation is effected by the method of standard additions (see below). Also without Zeeman correction the 396.2-nm line may be the better choice for certain matrices." The decrease in sensitivity is larger but the linear range is still far better. However it should be borne in mind that the 309.3-nm line is within the compensation range of the often used deuterium arc background corrector ANALYST JUNE 1986 VOL.111 659 Table 3. Recovery experiment aluminium in freeze-dried homo-genised porcine liver (n = 3) Aluminium found, Sample mean 2 s.d./mg kg-1 Liver . . . . . . . . 7.6 2 0.7 Liver + 5mg kg-l A1 . . 11.9 k 1.0 Liver + 10mg kg-1 A1 . . 17.0 k 0.9 Table 4. Plasma aluminium in a uraemic patient starting on oral phosphate binders Aluminium concentration, Day mean k s.d./pg 1-1 1 2 3* 9 23 51 87 129 * First dose administered here. 3.2 f 0.6 3.2 2 1.7 7.9 k 1.8 38.2 k 1.7 41.3 k 3.2 34.8 k 1.3 44.5 k 1.9 26.3 k 1.3 whereas the 396.2-nm line is not; if deuterium arc background correction is used this limits the use of the latter line to samples with virtually no background such as plasma samples and highly diluted tissue samples.Quantitation There appears to be some controversy as to the preference of quantitation by means of a calibration graph or by the method of standard additions.lOJ3 It has often been stated that quantitation by standard additions is time consuming and works only when a calibration graph approach also works. We however prefer and recommend quantitation by means of the method of standard additions for three reasons. Firstly, unless a strict class 100 working environment is available for the quantitation all samples must be run at least in duplicate in order to remove erroneous results stemming from airborne particulate contamination. When duplicate or triplicate analy-ses are mandatory the addition of standards to samples two and three during dilution do not increase the analysis time.We routinely analyse three separate aliquots of each sample, containing no one and two standard additions respectively. Secondly we find that the sensitivity of the graphite tubes changes during their lifetime. Other workers have made the same observation.10,13,19 There appears to be a period with very little change after proper conditioning has been carried out but prudence is nevertheless called for. Thirdly it is not possible to construct a good calibration graph. Tissue samples show sensitivities that differ according to their dilutions and severe chemical interferences are often observed on our instrumentation even with STPF conditions.This is most pronounced with bone samples. Plasma samples at least from uraemic patients also differ in sensitivity. Chemical interfer-ences however are not a problem with plasma samples and characteristic masses21 of 12-14 pg are normally observed, corresponding to a “quantitation limit” of approximately 2 pg 1-1 for a 20-pl injection of a sample diluted 1 + 1. We consider that these reasons make quantitation by standard additions mandatory if accurate results are required. Precision and Accuracy During a 10-month period we have evaluated the precision of the method as far as plasma samples are concerned by analysing aliquots of the same spiked plasma sample repeat-edly. The 5-ml aliquots were kept at -22 “C and were analysed at regular intervals during the 10-month period.We found an aluminium content of 50.9 k 2.4 pg 1-1 (mean k s.d.) based on 20 determinations. The range of results was 46.4-54.9 pg 1-l. In addition to giving a measure of the precision these results also show that the samples are stable under the storage conditions. The accuracy of the method was more difficult to assess as no appropriate certified reference material existed. For-tunately a new reference material has become available from the US National Bureau of Standards namely RM 8419, bovine serum.22 This material is stored and distributed frozen (-20 “C). The recommended aluminium content is 13 k 5 pg 1-1 and we obtained 14.1 -t 1.9 pg 1-1 (n = 5 ) . No standard materials relating to tissue samples are available to our knowledge.In view of this,we prepared our own “house standard” in order to conduct a recovery experiment to validate the recovery and precision after a digestion procedure was carried out. The “house standard” is porcine liver bought fresh from a butcher and freeze-dried and homogenised. Experience has shown it to be homogenous at the 100-mg level but for smaller sample sizes inconsistent results are found. We therefore carried out the experiment on 100-mg samples utilising the bomb digestion technique described previously. Known amounts were added as aqueous standards directly to the freeze-dried liver tissue in the digestion vessel. The results are given in Table 3; it should be noted that no special considerations were given to contamination during the preparation of the “house stan-dard,” i.e.during freeze-drying and homogenisation and the results obtained probably do not reflect the true aluminium content in porcine liver. Application The method described has been used for more than 1 year in our laboratory for the determination of aluminium in samples of human origin. A few examples are given here to illustrate its applicability . Table 4 gives plasma aluminium levels as a function of time for a uraemic patient starting on oral aluminium-containing phosphate binders.23 The first dose was administered at day 3. An increase in the concentration of aluminium in the plasma is clearly seen. This is believed to be caused by the oral phosphate binder alone as the aluminium concentration in the dialysis bath was below our “quantitation limit” of approximately 2 pg 1-1.The decrease in the level observed in this patient between days 87 and 129 was also noticed in other patients perhaps indicating the onset of aluminium accumula-tion in the tissue. For long-term dialysis patients we have occasionally found values exceeding 600 pg 1-1. In comparison for unexposed persons we normally find values close to or below our “quantitation limit,” and we agree with Frech et aZ.9 that the true “normal value” cannot be quantitated with contemporary analytical equipment. A second example is the determination of aluminium in bone samples i.e. normally in iliac crest biopsies. We have found the content of aluminium in this kind of sample from healthy individuals to be 1.0 If 0.5 mg kg-1 ( n = 8) with a range of 0.61.7 mg kg-1.24 Uraemic patients have signifi-cantly elevated levels and concentrations up to 400 mg kg-1 have been found in long-term dialysis patients.As far as soft tissue concentrations are concerned we only have a limited number of results and only from long-term dialysis patients. The values in human liver show a very wide spread from 30 to 300 mg kg-1 (dry mass). Two sets of values from the (necessarily dead) human brain have been obtained: 24.8 k 3.3 mg kg-1 (n = 2) from a long-term dialysis patient who died of causes unrelated to the kidney disease and 45.6 k 2.1 mg kg-1 (n = 2) from a patient who died of dialysis encephalopathy. Both sets of results are based on dry masses and the samples were taken from the grey matter part of the brain .2 660 ANALYST JUNE 1986 VOL.111 Finally the concentration of aluminium in urine from healthy individuals was found to be 8.3 k 1.4 pg 1-1 (n = 3). Conclusion Aluminium in samples of human origin can be determined accurately and with good precision by utilising the Zeeman-corrected atomic absorption spectrometric procedure des-cribed here. For plasma and urine samples which are analysed after simple dilution quantitation of 2 pg 1-1 of aluminium is achievable. The determination of aluminium in tissue has a poorer sensitivity compared with e.g. plasma but as the values found in tissue are much higher this does not constitute a problem. The help of Hanne Blaehr in providing the clinical samples is gratefully acknowledged.1. 2. 3. 4. 5. 6. 7. 8. 9. References Alfrey A. C. LeGendre G. R. and Kaehny W. D. N. Engl. J. Med. 1976 14 184. Kaehny W. D. Alfrey A. C. Holman R. E. and Shorr, W. J. Kidney Int. 1977 12 361. Kaehny W. D. Hegg A. P. and Alfrey A. C. N. Engl. J. Med. 1977,296 1389. Smeyers-Verbeke J. Verbeelen D. and Massart D. L. Clin. Chim. Acta 1980 108 67. Oster O. Clin. Chim. Acta 1981 114 53. Gardiner P. E. Ottaway J. M. Fell G. S. and Halls D. J., Anal. Chim. Acta 1981 128 57. Parkinson I. S. Ward M. K. and Kerr D. N. S . Clin. Chim. Acta 1982 125 125. Leung F. Y. and Henderson A. R. Clin. Chem. 1982 28, 2139. Frech W. Cedergren A. Cederberg C. and Vessman J., Clin. Chem. 1982,28 2259. 10. 11. 12. 13. 14. 15. 16.17. 18. 19. 20 * 21. 22. 23. 24. 25. Brown S . Bertholf R. L. Wills M. R. and Savory J. Clin. Chem. 1984 30 1216. Guillard O. Tiphaneau K. Reiss D. and Piriou A. Anal. Lett. 1984 17 1593. Bettinelli M. Baroni U. Fontana F. and Poisetti P., Analyst 1985 110 19. Gardiner P. E. Stoeppler M. and Niirnberg H. W. Analyst, 1985 110 611. Cornelis R. and Schutyser P. Contrib. Nephrol. 1984, 38 1. D’Haese P. C. Van de Vyer F. L. deWolff F. A. and DeBroe M. E. Clin. Chem. 1985 31 24. Baxter D. C. Frech W. and Lundberg E. Analyst 1985, 110 475. Gilli P. Malacarne F. and Fagioli F. Lancet 1983 1 656. Slavin W. Manning D. C. and Carnrick G. R. At. Spectrosc. 1981 2 137. Manning D. C. Slavin W. and Carnrick G. R. Spectrochim. Acta Part B 1982 37 331. Slanina P. Falkeborn Y. Frech W. and Cedergren A., Food Chem. Toxicol. 1984 22 391. Slavin W. and Carnrick G. R. Spectrochim. Acta Part B, 1984 39 271. Veillon C. Lewis S. A. Patterson K. Y. Wolf W. R., Harnley J. R. Versieck J. Vanballenberghe L. Cornelis, R. and O’Haver T. C. Anal. Chem. 1985 57 2106. Blaehr H. Madsen S . and Andersen J. R. in Taylor A., Editor “Aluminium and Other Trace Elements in Renal Disease,” Baillikre Tindall London 1986 p. 71. Heaf J. G. Podenphant J. Joffe P. Andersen J. R., Fugleberg S . and Braendstrup O. Nephron submitted for publication. Blaehr H. Madsen S . and Andersen J. R. Ugeskr. Laeg., submitted for publication. Paper A51395 Received October 31st 1985 Accepted January 15th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100657

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST JUNE 1986 VOL. 111 661 Elimination of the Effects of lnterferents on the Quantitative Determination of Deuterium Oxide in Biological Samples by Infrared Photometry Ralph N. Arnold Eric J. Hentges and Allen Trenkle Department of Animal Science 301 Kildee Hall Iowa State University Ames IA 5001 I USA Modifications to a procedure utilising sample purification by lyophilisation and automated infrared analysis were necessary to measure accurately the deuterium oxide content of various tissues and gastrointestinal tract contents of ruminants. Volatile fatty acids produced in the digestive tract of ruminants interfered with the measurement of deuterium oxide by infrared analysis. The volatile fatty acids were removed from samples by making the pH of the samples basic prior to lyophilisation.Contaminants in water extracted from tissues by lyophilisation of samples allowed the removal of the contaminants. The precision (S;) attained with this procedure was 3-6 mg 1-1 for the range of deuterium oxide concentrations between 120 and 800 mg I-'. Keywords Deuterium oxide determination; infrared photometry; lyophilisation A recent procedure ,1 modified with an automated sampling technique,2 permits the rapid measurement of tracer levels of deuterium oxide in biological fluids. Problems have been encountered however in measuring the deuterium oxide concentration of the gastrointestinal tract contents of rumi-nants. It was hypothesised that the volatile fatty acids produced in the gastrointestinal tract of ruminants interfered with the measurement of deuterium oxide by infrared photometry.Volatile fatty acids have been removed from tritiated plasma samples during lyophilisation by using sodium carbonate3 to make the plasma alkaline in order to prevent volatilisation of the acids during lyophilisation. It was also observed that the water extracted by lyophilisation from the tissues was cloudy indicating that contaminants were present. The purposes of this paper are to present the modifications that permit the accurate measurement of deuterium oxide in these samples and to present the precision obtainable with the assay for the concentrations of deuterium oxide normally used in experiments in which deuterium oxide is used as a tracer. Experimental Apparatus The procedures and equipment used were essentially those presented elsewhere112 with exceptions as indicated.Lyophilisation A rectangular bath (76 cm x 15 cm x 18 cm) insulated with Styrofoam (5.1-cm thickness) was constructed from galvan-ised sheet metal. A lid (60 cm x 90 cm) was constructed from 0.64-cm thickness hardboard and insulated with Styrofoam (5.1-cm thickness) to allow two rows of 11 vacuum traps (Curtin Matheson No. 233-114) to be introduced into the bath. One end of the bath was fitted with a removable lid to allow shell freezing of the flasks containing the sample. The bath was filled with 14 1 of methanol which was kept refrigerated at -84 "C by a flexible cooling probe (FTS Systems Model P2) attached to a Liquid Cooler (FTS Systems Model LC-100). The vacuum traps were attached by vacuum tubing to a vacuum manifold constructed from a 76-cm length of 5.1-cm diameter PVC pipe into which 13 mm valves (Labconco No.75900) were tapped. Liquid samples were lyophilised in 50-ml calibrated flasks and non-liquid samples were lyophilised in 500-ml round-bottomed freeze-drying flasks (Labconco No. 75608) connected to vacuum traps by vacuum tubing. After the flasks were connected to vacuum traps the traps were evacuated to purge the trap of atmospheric water before being placed in the methanol bath. Infrared Photometry A Miran I 4-pm fixed-filter infrared analyser (Foxboro Analytical Model No. 063-5651) with a calcium fluoride flow cell (Foxboro Analytical No. 0067024) with 0.2-mm spacers was used for deuterium oxide analysis.A water-jacketed cell holder was constructed from a block of aluminium metal (7.5 cm X 7.2 cm x 1.7 cm) by boring out the centre to accommodate the cell and boring passageways around the periphery of the block. The block was attached to the back plate of a standard cell mount (Foxboro Analytical No. 018-5321). Equilibration of temperature between the cell and cell holder was facilitated by filling empty spaces with a heat-sink compound. A thermostatically controlled heater and circulating pump (Braun Model Thermomix 1480 BKU) mounted on an 18-1 stainless-steel bath (Braun No. 850208/0) was used to pump water regulated at a constant temperature, through the cell holder. The thermostat was adjusted so that the temperature of the water in the bath was at room temperature.An automated sampling system2 consisting of an Auto Analyzer sampler (Technicon Model 11) and proportioning pump (Technicon Model I) was used to deliver the sample to the infrared analyser. The pump tubing used for sampling had a flow-rate of 2.0 ml min-1 and the pump tubing for drawing the sample through the flow cell had a flow-rate of 1.6 ml min-1. Triton X-100 was used as a surfactant in the rinse water at a concentration of 1 ml l-1. A sampling cam of 50 samples per hour with a 1 2 sample to wash ratio was used. To ensure that the temperature of the sample stream was the same as that of the flow cell a 20-cm length of transmission tubing was passed through a side-arm test-tube through which water returning from the flow cell holder to the circulator bath was passed.A small plug of glass-wool placed in a glass-T de-bubbler was used to filter the water stream before entry into the flow cell. The cell was cleaned by alternately forcing and drawing air bubbles and Triton X-100 rinse water through the cell with a syringe. When installing or removing the cell from the cell holder the heat-sink would often adhere to parts of the cell, blocking the light path. Detergent water and a soft bristle brush were found to be adequate for removing any part of the heat sink without harming the cell 662 ANALYST JUNE 1986 VOL. 111 A Hewlett-Packard Reporting Integrator (Model 3390A) was used to record the absorbance. The infrared analyser was set at a high gain an absorbance range of 0-0.1 and a 40-s integration response time.A voltage divider was constructed with a variable resistor to reduce the output of the infrared analyser to the integrator by 1000-fold. By adjusting the variable resistor greater sensitivities were obtainable. Dupli-cate analyses were achieved by allowing the sampler tray to rotate twice. A set of standards was run in each tray. Peak heights of standards were regressed against the known concentration of standards and the concentration of samples was determined from the regression equation. Procedures Effect of volatile fatty acids on deuterium oxide measurement and their removal The effect of individual volatile fatty acids on the infrared analysis and their removal by the addition of sodium carbo-nate to samples before lyophilisation was evaluated.Three solutions of acetic acid (45.2 50.9 and 56.9 mM) three solutions of propionic acid (10.4 15.9 and 21.8 mM) and three solutions of butyric acid (5.4,10.5 and 15.2 mM) were made up to the approximate concentrations of these acids that were measured in the rumen of sheep consuming diets containing 100 40 or 10% of hay.4 In addition three solutions that were approximately of the same concentration (26.6 mM acetic acid, 25.2 mM propionic acid and 25.5 mM butyric acid) and a solution (30.6 mM) that was 1.5 times the highest concen-tration of propionic acid that was measured in the sheep were also made up. Sufficient amounts of deuterium oxide were added to each of these solutions to make a final concentration of deuterium oxide of 200 mg 1-1.Duplicate 10-ml volumes of all the solutions except the three equimolar solutions were lyophilised both with and without 0.3 g of sodium carbonate. Measurements of the deuterium oxide were then made on all the lyophilised solutions and the three equimolar solutions. After the measurements of deuterium oxide were completed, the duplicate samples were pooled and the pH values of the solutions were measured. Precision and Sensitivity of the Assay To evaluate the precision of the assay standards having a deuterium oxide concentration of 120 200 400 and 800 mg 1-1 were each analysed in 13 replicates and the coefficient of variation was calculated at each concentration. These analyses were repeated four times and the average coefficient of variation was calculated.A linear calibration calculation5 was made on 85 calibration graphs measured over a 4-month ,period. The linear calibration calculation was used to provide a measure of the precision of the predicted deuterium oxide concentration (&) obtained at the mean concentration for the standards used to produce the regression. The average coefficient of determination and the average linear calibration calculation were calculated for calibration graphs with similar ranges of standards. Application to Biological Samples Samples of blood ruminal contents duodenal contents ileal contents faeces and urine were obtained from cattle that were fed either a high-roughage or high-grain diet and were dosed with deuterium oxide. Water was extracted from the samples by lyophilisation and analysed for deuterium oxide by infrared photometry.Sodium carbonate was added to the extracted water in sufficient amounts (approximately 0.03 g ml-1) to make the water alkaline. The samples were lyophilised a second time and analysed for deuterium oxide. Comparisons for statistical differences between the measured concen-trations of deuterium oxide before the addition of sodium carbonate and the measured concentration after the addition of sodium carbonate were made by a Student’s t-test.5 Tissue samples were removed from cattle and sheep that had been dosed with deuterium oxide. Water was extracted from the tissues by lyophilisation. After extraction activated charcoal was added to the water in an attempt to remove the contaminants and lyophilisation was performed a second time.Evaluation of Some Potential Errors on the Measurement of Deuterium Oxide The degree of isotopic separation was evaluated to determine if samples need to be lyophilised to completion. A solution was made up to a concentration of 500 mg 1-1 of deuterium oxide and aliquots were lyophilised to different degrees of completion. One 10-ml aliquot was lyophilised for 30 min, three 10-ml aliquots were lyophilised for 90 min one 10-ml aliquot was lyophilised for 110 min and seven 10-ml aliquots were lyophilised to completion. The concentrations of deuterium oxide were measured in both the water that was collected in the vacuum traps and the water remaining in the calibrated flasks. The degree of isotopic separation owing to adsorption on to activated charcoal was also evaluated.One gram of activated coconut charcoal (10 mesh) was added to 10-ml aliquots of solutions made up to contain 192 or 800 mg 1-1 of deuterium oxide and allowed to stand at room temperature for 2 d. The aliquots were then lyophilised with or without charcoal. The procedure was performed on six replicates for both deuterium oxide solutions and the concen-tration of deuterium oxide was measured in the lyophilised water. Student’s t-test5 was used to determine if statistical differences in deuterium oxide concentration occurred owing to isotopic separation. Results and Discussion Effect of Volatile Fatty Acids on Deuterium Oxide Measure-ment and Their Removal Acetic propionic and butyric acids absorbed infrared radia-tion in direct proportion to their concentrations with no evident differences between the acids (Fig.1). Spectral analyses6 of most carboxylic acids show that these acids exhibit a similar degree of infrared absorption at 4 pm. The absorption due to the 0-H stretching of the carboxylic group occurs in a broad peak at 3.6 pm which ranges from 3 to 4 ym; thus at 4 ym the shoulder of the peak would be measured. Making the solution alkaline by the addition of sodium carbonate before lyophilisation evidently prevented the volat-ilisation of these acids and thus prevented the collection of the acids in the extracted water (Fig. 1). The pH values of the volatile fatty acid solutions were less than 4.0 but when sodium carbonate was added the pH was over 8.0.After lyophilisation of the solutions with the added sodium carbo-nate the pH of the extracted water was around 5.5 which was similar to the pH of the water used to make up the solutions. The pH of the water extracted by lyophilisation of the volatile fatty acid solutions without the addition of sodium carbonate was less than 4.0. The pK of acetic propionic and butyric acids is 4.8.7 The total amounts of volatile fatty acids in the solutions were collected in the vacuum traps when lyophilised in the acidic state; the equimolar solutions (approximately 26 mM) that were not lyophilised absorbed infrared radiation proportionally to the concentration of volatile fatty acid to a similar extent as solutions of volatile fatty acids that were lyophilised without sodium carbonate (Fig.1). Precision and Sensitivity of the Assay The precision suggested1 for the determination of deuterium oxide by manually injecting the sample into the flow cell was ANALYST JUNE 1986 VOL. 111 663 600 -5 400 -ti 0 10 20 30 40 50 60 [ V F A ] / ~ M Fig. 1. Effect of volatile fatty acids on the determination of 200 mg 1-1 deuterium oxide standards by infrared analysis after lyophilisa-tion with or without the addition of sodium carbonate. 0 Acetate; . acetate + Na2C03; A propionate; A propionate + Na2C03; 0, butyrate; 0 butyrate + Na2C0, Table 1. Concentration of deuterium oxide measured in samples from cattle before and after the addition of sodium carbonate during lyophilisa tion Average concentration of samples removed from cattle High-roughage High-grain diet/ diet/ mg 1-I mgl-l Standard error/ Sample Before After Before After mgl-' Blood .. 391 382 434 423 12 Rumen . . 259 246 359 271* 7 Duodenum . . 394 327* 395 358* 12 Ileum . . . . 331 326 360 324t 14 Faeces . . 286 279 428 326* 9 Urine . . . . 313 302 343 329 8 of sodium carbonate. of sodium carbonate. * P < 0.05 differs from the mean concentration before the addition t P < 0.10 differs from the mean concentration before the addition mg 1-l. The coefficient of variation with the use of an automated sampling system2 has been determined to be less than 1% at all standard concentrations. The coefficient of variation in the procedure described in this paper was 1.9% at deuterium oxide concentrations of 120 mg 1-1 1.7% at concentrations of 200 mg 1-1 1.5% at concentrations of 400 mg I-' and 1.2% at concentrations of 800 mg 1-1.The concentration of deuterium oxide in most samples usually fell in the range between 120 and 800 mg 1-1 when the animals were injected with 0.2 g of deuterium oxide per kg live mass. The precision of the estimated deuterium oxide concen-tration (S;) predicted using the linear calibration calculation, was 3 mg 1-1 for 39 calibration graphs (r2 = 0.9990) with standards between 120 and 400 mg 1-1 and 6 mg 1-1 for 34 calibration graphs (r2 = 0.9994) with standards ranging between 160 and 800 mg 1-1. In order to measure deuterium oxide concentrations near background levels the voltage output from the voltage divider was increased so that a 320 mg 1-1 deuterium oxide standard caused full-scale usage of the integrator chart.This increased the sensitivity so that standards as low as 10 mg 1-1 were measurable. A 5 mg 1-1 standard was detectable but was more variable. The precision (Sa) as determined by the linear calibration calculation of 11 calibration graphs for the concentration of deuterium oxide from 10 to 320 mg 1-1 was 2 mg 1-l with an average coefficient of determination (r2) of 0.9995. The precision (S;) obtainable (3-6 mg 1-1) using the described technique for the range of deuterium oxide concen-trations usually encountered in biological systems would be adequate for determining the body water content of rumi-nants. Application to Biological Samples Impurities were present in water extracted from gastrointesti-nal tract contents as evidenced by an odour and a greater measured concentration of deuterium oxide than present in the blood.After addition of sodium carbonate to the water and lyophilising a second time the odour was removed and lower measured concentrations of deuterium oxide were observed for some samples removed from the gastrointestinal tract (Table 1). With the exception of duodenal contents, over-estimation of deuterium oxide occurred only for water extracted from samples removed from the gastrointestinal tract of cattle fed a predominantly maize grain diet. The measured concentrations of deuterium oxide in water extrac-ted from blood and urine from cattle fed either the high-roughage or the high-grain diets and in water extracted from ruminal contents ileal contents and faeces of cattle fed the high-roughage diet decreased by 14 mg 1-1 or less.This small change may have been due to impurities but was not statistically significant and was close to the precision of the assay. Attempts were not made to identify directly the contaminant(s) in water extracted by lyophilisation from gastrointestinal tract contents but volatile fatty acids would be a likely possibility. It is not clear why the high-grain diet produced a greater over-estimation of deuterium oxide concentration in samples removed from the animals than the high-roughage diet. It has been commonly observed that, when ruminants are fed high-grain diets the pH of the rumen is lower than when fed high-roughage diets.8 When ruminants are fed high grain diets the total concentration of volatile fatty acids are higher than when high roughage diets are fed.4 Tissue samples containing high levels of fat produced water extracts that were more cloudy than water extracts from tissues containing low levels of fat.These impurities did not interfere with the infrared absorbance of deuterium oxide but the degree of variation between replicates was greater and the flow cell required more frequent cleaning. Water was extrac-ted from tissues removed from an animal not dosed with deuterium oxide and no detectable absorbance was measured. The addition of approximately 1 g of activated coconut charcoal (10 mesh) was usually sufficient to remove the impurities from 10 ml of extracted water within 2 d at room temperature.The water was then separated from the charcoal by lyophilisation. Although contaminants that interfered with the determi-nation of deuterium oxide were not present in all of the samples it is advisable to treat all samples in a similar manner because of the uncertainty of the source of the contaminants. Thus the routine procedure adopted for lyophilisation of water from tissues and digesta samples that contain solids was first to extract water by lyophilisation add sodium carbonate and charcoal to the extracted water and then perform a second lyophilisation. The routine procedure adopted .for liquid samples such as blood urine or rumen fluid was to add sodium carbonate to the sample before lyophilisation.Evaluation of Some Potential Errors in the Measurement of Deuterium Oxide The average concentration of deuterium oxide in water that had been collected after partial lyophilisation was 486 mg 1-1 with a standard error (SE) of 3 mg 1-1 whereas the concentration of deuterium oxide in the water that remained in the flasks was 502 (SE 3) mg 1-1. The average concentration of deuterium oxide for the seven samples lyophilised t 664 complete dryness was 490 (SE 2) mg 1-1. The lower concentration in the collected water after partial lyophilisation (4 mg 1-1 lower than the average of the samples lyophilised to completeness) was not a significant decrease. The 12 mg 1-1 greater concentration in the water that remained in the flask was significant at the P<O.O1 level.Differences in the concentration of deuterium oxide were not evident between the lengths of times for which the samples were lyophilised. Whereas some separation of isotopes was evident in the water that remained in the flasks the amount of over-estimation was close to the precision of the assay. No separation of the isotopes was evident in the collected water which would be the fraction used for the measure of concentration of deuterium oxide and therefore lyophilisation to complete dryness is not essential. Isotopic separation has been ob-served9 to occur during evaporation or freezing owing to the difference in properties of the hydrogen and deuterium atom. The separation of isotopes owing to the adsorption of charcoal did not occur (P>O.10). The average concentrations of deuterium oxide for the 192 and 800 mg 1-1 solutions were 193 (SE 1) and 802 (SE 6) mg 1-1 when charcoal was added but not lyophilised 194 (SE 1) and 800 (SE 3) mg 1-1 when the added charcoal was lyophilised with the solution and 192 (SE 1) and 798 (SE 3) mg 1-1 when charcoal was not added. Thus, the charcoal does not need to be lyophilised with the sample. Water adsorbed on to charcoal or silica gel was foundlo to have an increase in density. Contamination of lyophilised water with dried blood or sodium carbonate which were drawn from the calibrated flasks during lyophilisation also occurred. The presence of red blood cells did not seem to interfere with infrared absorption. Sodium carbonate solutions were found to absorb infrared radiation proportionally to concentration at 4 pm.The degree of absorption for a sodium carbonate solution of 1 mg ml-1 corresponded to a concentration of deuterium oxide of approximately 12 mg 1-1. Absorption of infrared radiation by the hydrogen carbonate ion has been reported11 to occur between 3.7 and 4.1 pm. Thus when particulate matter was drawn over contaminating the lyophilised water the samples had to be lyophilised a second time. After prolonged use of the infrared analyser the peaks plotted by the integrator were frequently rounded instead of being sharp. Cleaning of the cell corrected the problem. The presence of impurities in water extracts required the cell to be cleaned more frequently and increased the variation between replicate analyses.Caution must be exercised as to which ANALYST JUNE 1986 VOL. 111 solutions are used to clean the cell. Although calcium fluoride is seemingly resistant to most chemicals except ammonium salts,12 damage to the precision-sealed flow cell was sustained when either acetone or dilute hydrochloric acid were used as cleaning solutions. The precision-sealed calcium fluoride cell, as noted previously,l tended to leak after prolonged use unless the edges of the cell were glued. Contact cement and quick-drying epoxy glues were found to be unsuitable but slow-drying epoxy glues were found to be adequate. Journal Paper No. 5-11979 of the Iowa Agricultural and Home Economics Experimental Station Ames project No. 2454. This study was supported in part by grants from the Iowa Corn Promotion Board and the Iowa Beef Industry Council.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Byers F. M. Anal. Biochem. 1979,98 208. Ferrell C. L. and Jenkins T. G.,J. Anim. Sci. 1980 Suppl. 1 , 358. Jackson H. D. Black A. L. and Moller F. J. Dairy Sci., 1968 51 1625. Sutton J. D. in Ruckebush Y. and Thivend P. Editors, “Digestive Physiology and Metabolism in Ruminants,” AVI Publishing Company Westport CT 1980 p. 271. Snedecor G. W. and Cochran W. G. “Statistical Methods,” Iowa State University Press Ames IA 1980 pp. 55 and 169. Pouchert C. J. “The Aldrich Library of Infrared Spectra,” Aldrich Chemical Co. Milwaukee 1970 p. 223. Weast R. C. Editor “Handbook of Chemistry and Physics,” The Chemical Rubber Co. Cleveland OH 1972 p. D120. Church D. C. “Digestive Physiology and Nutrition of Ruminants,” 0 & B Books Corvallis OR 1976 p. 291. Emeleus H. J. James F. W. King A. Pearson T. G., Purcell R. H. and Briscoe H. V. A. J . Chem. SOC. 1934, 1207. King A. James F. W. Lawson C. G. and Briscoe H. V. A., J. Chem. SOC. 1935 1545. Robinson J. W. Editor “Handbook of Spectroscopy,” CRC Press Cleveland OH 1974 11 p. 78. Smith A. L. “Applied Infrared Spectroscopy Fundamentals, Techniques and Analytical Problem Solving,” Wiley New York 1979 p. 111. Paper A51394 Received November 31st 1985 Accepted December 30th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100661

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST JUNE 1986 VOL. 111 665 Background Correction Method for the Determination of Ascorbic Acid in Soft Drinks Fruit Juices and Cordials Using Direct Ultraviolet Spectrop hotometry Oi-Wah Lau Shiu-Fai Luk and Kit-Sum Wong Department of Chemistry Chinese University of Hong Kong Shatin N. T. Hong Kong A background correction method has been developed for the determination of ascorbic acid in soft drinks, fruit juices and cordials using UV spectrophotometry. The sample blank is produced by the catalytic oxidation of the acid by copper(l1). To correct for the absorption due to copper(ll) EDTA was added after the oxidation, and standard solutions were prepared to contain the same concentration of the copper(l1) - EDTA complex, which does not catalyse the oxidation of ascorbic acid at room temperature.The decomposition of ascorbic acid in real samples should be carried out at 50 "C because citric acid which is usually present in the samples under study retards the copper(l1)-catalysed oxidation of ascorbic acid. Absorbance measurements were made at 267 nm and pH 6. The calibration graph was linear for 0-20 pg ml-1 of ascorbic acid. The precision was 0.1-0.5% for ascorbic acid concentrations in the range 5 1 3 1.19 ml-1. The proposed method is selective and many ingredients commonly found in soft drinks fruit juices and cordials do not interfere. The method was utilised to determine ascorbic acid in a wide range of samples. Keywords Ascorbic acid determination; ultraviolet spectrophotometry; background correction; copper(//)-catalysed oxidation; soft drinks fruit juices and cordials analysis Ascorbic acid is present in large amounts in citrus fruits and tomatoes.A prolonged lack of ascorbic acid results in scurvy, and less severe deficiency of the acid produces alterations in connective-tissue structure and may also cause decreased resistance to some infections.1 A number of methods have been used for the determina-tion of ascorbic acid. Among them is a titrimetric method using 2,6-dichloroindophenol2 as the titrant. Although the method is rapid the reagent itself is unstable and must be standardised before use. Moreover the method is unsuitable for coloured samples which interfere with the detection of the end-point. Ascorbic acid can also be determined by a microfluorimetric method 3 which is very time consuming.Differential-pulse polarography4.5 has been used to determine ascorbic acid at the pg ml-1 level but the method suffers from interferences from electroactive impurities present in the sample. Ascorbic acid absorbs strongly from 243 to 267 nm depending on the acidity of the medium. Before background correction techniques were introduced the ultraviolet spec-trum of ascorbic acid was seldom used for the direct determination of the acid in real samples with complex matrices because of the very high background absorption in the UV region. There are various methods available in the literature to correct for background absorption in the ultraviolet region for the determination of ascorbic acid such as thermal decompo-sition,6 an enzymatic method,7ts direct ultraviolet irradiation9 and alkaline treatment.10 The concentration of ascorbic acid can be calculated from the difference in absorbance at a wavelength selected before and after the decomposition of ascorbic acid.The enzymatic method requires a long incuba-tion time of 30 min at 30 "C and is also too specialised for general use. The destruction rates of ascorbic acid using thermal decomposition and UV irradiation were observed to be very slow.10 Tannin and caramel interfere seriously with the method of Fung and Luk,lO and the method is also unsuitable for the analysis of grapefruit and apple juices. l1 Thus there is a need for a new background correction method for samples with complex matrices. It has long been known that copper(I1) catalyses the oxidation of ascorbic acid.12 Onishi and Haral3 measured the oxidation rates of ascorbic acid in the presence of various kinds of metal ions and the catalytic effect of copper(I1) was found to be much better than that exhibited by other ions.Moussa'S used copper(I1) sulphate to oxidise selectively ascorbic acid in order to prevent its interference in the determination of other compounds dispensed together with ascorbic acid. The purpose of this work was to develop a background correction technique based on the metal-catalysed oxidation of ascorbic acid which will be subject to less interference and applicable to a wider range of samples. The method has been applied to the determination of ascorbic acid in soft drinks, fresh and canned fruit juices and cordials.Experimental Apparatus The spectra were recorded with a Hitachi 323 recording spectrophotometer using matched 10 mm silica cells. The pH measurements were made with a Chemtrix Type 60A pH meter. Reagents All reagents used were of analytical-reagent grade. Ascorbic acid stock solution 1000 pg ml-1. Prepared by dissolving 100 mg of the acid in 100 ml of distilled water. From this stock solution working standard solutions were freshly prepared by appropriate dilution with distilled water. Copper(1l) solution 5 pg ml-1 (7.78 X 10-5 M) p H 6. Prepared by pipetting 5.0 ml of a 1000 pg ml-1 copper(I1) solution followed by the addition of 200 ml of 1 M sodium acetate solution and 7 ml of 1 M acetic acid and diluting the solution to 1 1 with distilled water.The copper(I1) - EDTA complex solution (6.2 x 10-5 M) was freshly prepared before use by mixing the copper(I1) solution with 5 X M EDTA solution (4 + 1 V/V). Procedure The wavelength was set at 267 nm. The calibration graph was prepared by measuring the absorbances (at 267 nm) of the ascorbic acid standard solutions against the reagent blank. The standard solutions (0-20 pg ml-1) were prepared b 666 ANALYST JUNE 1986 VOL. 111 mixing 1.0 ml of the appropriate ascorbic acid solution containing 0-120 yg of the acid with 5.0 ml of the copper(I1) -EDTA solution in dry test-tubes. The slope (S) of the calibration graph was determined. In real sample analyses the canned fruit juice soft drink or cordial sample was mixed thoroughly by shaking vigorously to ensure uniformity.Juice from a fresh citrus fruit sample was obtained using a squeezer. The sample was then centrifuged at 750 rev min-1 for 15 min and the supernatant liquid was taken out and diluted X-fold with distilled water; X i s usually 5 or 10 depending on the ascorbic acid content and the colour of the sample. To 1.0 ml of the dilute sample solution were added 5.0 ml of the copper(I1) - EDTA solution and the absorbance of the solution at 267 nm designated as A(I) was measured. To another 1.0-ml aliquot of the dilute sample were added 4.0 ml of the 5 pg ml-1 copper(I1) solution at pH 6. This mixture was heated in a water-bath at 50 "C for 15 min and then 1.0 ml of 5 x 10-4 M EDTA solution was added. The absorbance of the resulting solution at 267 nm was measured and was designated A(I1).The ascorbic acid content was calculated using the equation ascorbic acid content = 6(XAA)/S where X is the dilution factor AA the difference between A(1) and A(II) S the slope of the calibration graph and the factor of 6 included to account for the dilution by the buffer solution. Results and Discussion Selection of the Wavelength and pH for the Measurement of Absorbance Natural molecules of ascorbic acid readily ionise in solution with a pK1 of 4.12 and a pK2 of 11.51.16 As the ascorbate-containing species have different absorption spectra the wavelength of the absorption maximum of ascorbic acid is closely related to pH. The wavelengths of the absorption maximum for ascorbic acid at different values of pH were determined.It was found that at pH 2 where ascorbic acid is expected to exist entirely in the non-ionised form the absorption maximum occurred at 245 nm. At pH 12 where the acid is expected to dissociate fully the absorption maximum occurred at 300 nm. The wavelength of the absorption maximum remained essentially constant at 267 nm at pH 5-10 where the predominant species is the monoanion of ascorbic acid. Further the absorbance was found to be highest at pH 6 and lowest at pH 12. Thus it appeared that the optimum pH should be 6 and the absorbance would then be measured at 267 nm. The stability of ascorbic acid at different values of pH was also studied by measuring the initial absorbance of 30 yg of ascorbic acid at different pH and measuring the absorbance again 1 and 2 h later.It was found that the percentage loss at pH 2 after 2 h was only 1.4%. The percentage loss increases as the pH increases and at pH 7 or above ascorbic acid was completely oxidised within 1 h and within minutes at pH 12. The percentage loss at pH 6 after 1 h was 35 which is considerable so that some measure is necessary to stabilise ascorbic acid at this pH if absorbance measurements are made at this wavelength. Choice of Catalyst An ideal catalyst for the oxidation of ascorbic acid for the purpose of background correction should be one which can speed up the oxidation of ascorbic acid whilst having negligible absorbance in the UV region. Onishi and Haral3 have found that copper(I1) is the most effective catalyst for the oxidation of ascorbic acid followed by iron(II) cadmium(II) zinc(II), aluminium(II1) and iron(II1).As copper(I1) and iron(I1) and their complexes absorb in the UV region whereas cad-mium( 11) zinc( 11) and aluminium( 111) and their complexes do not the latter three metal ions will cause less interference in the determination of ascorbic acid when they are used as catalysts for the oxidation of the acid. Thus experiments were performed to investigate whether their catalytic power could be improved by increasing the temperature and the results are shown in Table 1. It can be seen that at 50 "C the rate constants of the reaction in the presence of these three non-transition metal ions are still small compared with that found in the presence of copper(II) and are not large enough to make them useful as catalysts for the complete destruction of ascorbic acid.Thus copper(I1) was chosen as the catalyst for the oxidation of ascorbic acid despite the fact that copper(I1) and its complexes absorb in the UV region. The reaction time for the complete oxidation of ascorbic acid is expected to decrease when the concentration of copper(I1) increases. The effect of the concentration of copper(I1) on the time for the complete oxidation of ascorbic acid at 25 "C was studied and the results are shown in Table 2. It was observed that the time for the complete oxidation of ascorbic acid decreased rapidly with increasing concentrations of copper(II) but levelled off at Cu(I1) concentrations of 5.2 X 10-5 M.As copper(I1) complexes absorb in the UV region, the lowest possible copper(I1) concentration (Le. 5.2 X 10-5 M) was chosen where the time for complete oxidation of ascorbic acid at 25 "C was 12 min. Choice of Masking Reagent The next step was to select a chelating agent that could mask the catalytic effect of Cu(I1) on the oxidation of ascorbic acid for the correction of the absorption due to Cu(I1). The resulting copper(I1) complex must not catalyse the oxidation of ascorbic acid in the sample solution. The effect of chelating agents on the rate of the copper-catalysed oxidation of ascorbic acids was studied by Onishi and Hara.14 Their results indicated that four chelating agents, namely nitrilotriacetic acid (NTA) ethylenediaminetetra-acetic acid (EDTA) quinolin-8-01 and a,a'-dipyridyl are very effective in suppressing the catalytic effect of Cu(I1).However quinolin-8-01 absorbs strongly at 243 nm whereas a,a'-dipyridyl absorbs strongly at 234 and 282 nm. On the other hand NTA and EDTA absorb only slightly above 250 nm and thus these two chelating agents are more suitable masking agents for this work. The absorbance at 267 nm of the Cu(I1) - NTA and Cu(I1) - EDTA complexes in acetate buffer of pH 5.6 were measured against distilled water giving the values of 0.12 and 0.19 for the Cu(I1) - NTA and Cu(I1) -Table 1. Rate constants at 50 "C for the oxidation of ascorbic acid (10 pg ml-l) in the presence of metal ions at pH 5 Metal ion concentration Rate constant at x 1 0 " ~ 50°C x 103/rnin-' Cd(I1) Zn(I1) AI(II1) .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0.67 1 1 26.0 17.8 8.2 Cu(I1) . . . . . . . . . . . . . . . . 0.05 14.3 x 103 Table 2. Effect of copper(I1) concentration on the oxidation of a 10 pg ml-l ascorbic acid solution at pH 6 and 25 "C Concentration of Cu(I1) x lO5/ Time for complete oxidation M of ascorbic acid/min 1 .o 2.1 3.1 4.2 5.2 6.3 45 25 16 15 12 1 ANALYST JUNE 1986 VOL. 111 EDTA complexes respectively where both complexes had a concentration of 1.6 X M. Further NTA has a greater suppressive effect on the Cu( 11)-catalysed oxidation of ascorbic acid than EDTA e.g. the rate of oxidation at pH 6 was 0.0134 min-' in the presence of NTA (8 X 10-5 M) compared with 0.0402 min-1 in the presence of EDTA (8 x 10-5 M).Nevertheless NTA is a suspected carcinogen and a less common reagent than EDTA so EDTA was chosen as the masking agent for copper(II).The absorption due to the Cu(I1) - EDTA complex constitutes part of the reagent blank and can be readily corrected. 667 Effect of EDTA on the Stability of Ascorbic Acid Fogg and Summan17 observed that ascorbic acid could be stabilised by the addition of EDTA. The effect of EDTA on the stability of ascorbic acid in the presence of copper(I1) was studied and the results are shown in Table 3. It can be seen that when the concentration of EDTA was the same as that of copper(II) ascorbic acid was completely decomposed within 3 h. However as soon as a slight excess of EDTA was present, the acid was significantly stabilised.Further increases in EDTA concentration produced much smaller differences in the stabilising effect. It seems possible that the oxidation of ascorbic acid may also be catalysed by trace amounts of metals in the reagents and in distilled water. Thus a slight excess of EDTA seems to be sufficient to complex these impurities and suppress the oxidation of ascorbic acid. Hence the copper(I1) - EDTA solution used in the proposed method contained a slight excess of EDTA. An experiment to assess the effect of pH on the stability of ascorbic acid in the presence of 5.2 x M of Cu(I1) - EDTA complex and 3.1 x 10-5 M of EDTA was performed as described previously in the absence of any Cu(I1) complex.The stability of ascorbic acid in the presence and absence of Table 3. Effect of the concentration of EDTA on the stability of a 10 pg ml-1 ascorbic acid solution containing 5.2 x 10-5 M Cu(I1) and at pH 6 (initial absorbance 0.863) Concentration of EDTA x 1 0 5 / ~ Absorbance at 267 nrn after 3 h 5.2 0 5.3 0.826 5.4 0.829 8.3 0.846 1 .ooo 0.800 8 0.600 c ru e $ La Q 0.400 0.200 0.000 210 250 300 Wavelengthhm 350 Fig. 1. Effect of pH on the UV spectrum of ascorbic acid (10 pg ml-l) in the presence of the Cu(I1) - EDTA complex (5.2 X M) against reagent blank. The number on each curve indicates the pH used the Cu(1I) - EDTA complex and EDTA is compared in Table 4 by considering the percentage loss of the acid after 1 h at various pH values.It can be seen that ascorbic acid decom-poses less rapidly in the presence of the complex - EDTA mixture and the loss of ascorbic acid at pH 6 was reduced to 1%. Effect of pH on the UV Spectrum of Ascorbic Acid in the Presence of the Cu(I1) - EDTA complex The UV spectra of ascorbic acid solution at different pH but at a fixed concentration of Cu(I1) - EDTA (5.2 X 10-5 M) and EDTA (3.1 x 10-5 M) are shown in Fig. 1. The series of spectra in the presence of the mixture of complex and EDTA are very similar to those in their absence and there were no shifts in the wavelengths for the absorption maxima. The absorbance was observed to increase slightly in the presence of the mixture at pH 7-12 possibly owing to the greater stability of ascorbic acid in the presence of EDTA.The maximum absorption again occurred at pH 6. As noted previously the loss of ascorbic acid in the presence of the mixture was reduced to 1% within 1 h at this pH so that pH 6 could now be chosen as the optimum pH for the proposed method. To sum up copper(I1) has been shown to be an effective catalyst for the complete oxidation of ascorbic acid and it was converted into the EDTA complex after the oxidation. Background absorption could then be obtained. The absor-bance due to ascorbic acid was measured in the presence of the Cu(I1) - EDTA complex which had practically no catalytic effect on the oxidation of ascorbic acid to compensate for the absorption due to copper. A pH of 6 in the presence of a small amount of EDTA and a measurement wavelength of 267 nm, were chosen.Interference Studies A number of organic acids such as citric acid tartaric acid, malic acid lactic acid tannic acid and fumaric acid and a number of sugars such as glucose sucrose and fructose are known to be present in fresh fruit juices. Preservatives such as benzoic acid and sorbic acid and other artificial colouring Table 4. Comparison of the stability of 30 vg of ascorbic acid in 3 ml of buffer at different pH with and without the [Cu(II) - EDTA complex] -EDTA mixture Loss after 1 h YO PH 2 4 5 6 7 8 10 12 Buffer alone 0.6 4 18 35 100 100 100 100 With [Cu(II) - EDTA complex] - EDTA mixture 0.2 0.2 0.8 1 12 25 68 100 Table 5. Effect of temperature on the oxidation of 10 pg ml-1 of ascorbic acid at pH 6 containing 0.17% of citric acid and 5.25 X M of Cu(1I) (initial absorbance 0.863) Temperature/"C ~~~~ ~ 25 30 40 50 Absorbance at 267 nm after 15min .. . . . . . . 0.476 0.300 0.172 0.001 Decomposition % . . . . . . 46 66 80 99. 668 ANALYST JUNE 1986 VOL. 111 -Table 6. Effect of the common ingredients of soft drinks fresh and canned fruit juices on the determination of a 10 pg ml-l ascorbic acid solution at pH 6 containing the copper(I1) - EDTA complex (5.2 x 10-5 M) Foreign substance added Citricacid . . . . . . Fumaricacid . . . . . . Lacticacid . . . . . . Malicacid . . . . . . Tartaricacid . . . . . . Fructose . . . . . . . . GIucose . . . . . . . . Sucrose . . . . . . . . Aspartame . . .. . . Saccharin . . . . . . Caramel . . . . . . . . FD & C Yellow No. 5 . . FD & CYellow No. 6 . . Carboxymethylcellulose . . Pectin . . . . . . . . Starch . . . . . . . . Xanthan . . . . . . . . Caffeine . . . . . . . . Sodiumsulphite . . . . Benzoicacid . . . . . . Sorbicacid . . . . . . Tannicacid . . . . . . EDTA . . . . . . . . Concentration of foreign substance . . 0.2% . . 8.3p.p.m. . . 0.13% . . 0.17% . . 0.25% . . 0.58% . . 0.8% . . 2.8% . . 220p.p.m. . . 17p.p.m. . . 20p.p.m. . . 66.7p.p.m. . . 83p.p.m. . . 0.1% . 0.12% * . 0.1% . . 92p.p.m. . . 20p.p.m. . . 33p.p.m. . . 17p.p.m. . . 1.5p.p.m. . . 4.2p.p.m. . . 19.5p.p.m. Ascorbic acid found/pg ml-1 9.56 10.48 9.55 9.58 9.58 10.45 9.56 9.55 9.50 9.52 9.59 9.53 9.52 9.57 9.52 10.48 10.45 10.50 9.50 10.47 9.57 9.54 9.60 Error YO -4.4 4.8 -4.5 -4.2 -4.2 4.5 -4.4 -4.5 -5.0 -4.8 -4.1 -4.7 -4.8 -4.3 -4.8 4.8 4.5 5.0 -5.0 4.7 -4.3 -4.6 -4.0 Table 7.Results of recovery test Ascorbic acid Ascorbic acid Sample added/pg ml-1 found/pg ml-l Recovery YO - Limecordial . . . . 0 0 Blackcurrant drink . . 0 408.5 k 0.8 -497.5 499.6 k 3 100.4 500 901.1 ? 7.0 98.5 500 824.0 k 10.5 100.2 500 697 ? 24.0 99.8 - Orangedrink . . . . 0 323.0 k 0.4 Lemontea . . . 0 198 k 6.4 -matter and sweeteners are very common in canned fruit juices. It is therefore necessary to assess the effect of these substances on the determination of ascorbic acid using the proposed method.EDTA citric acid tartaric acid and acetic acid are known to form complexes with copper(I1) with logarithms of the stability constants at pH 6 and 25 "C being 12.5 6.7 3.2 and 1.1 respectively. 18 Although EDTA forms a stable complex with copper(II), its concentrations in soft drinks and fruit juices are unlikely to be high. Citric acid was found to retard the copper(1I)-catalysed oxidation of ascorbic acid. Citric acid in citrus fruit accumulates to about 1 YO at commercial maturity.19 Commer-cial fruit beverages are expected to contain around 1% or less of citric acid. The fruit juice samples are diluted at least six times before analysis so that the concentration of citric acid in the sample solutions would be below 0.16%.The time for complete oxidation of ascorbic acid in the presence of 0.17% of citric acid was found to be 1 h compared with 12 min in its absence. Next the effect of temperature on the copper(I1)-catalysed oxidation of ascorbic acid in the presence of 0.17% of citric acid was studied. The results in Table 5 show that a temperature of 50 "C is high enough to reduce the interference from 0.17% of citric acid and the ascorbic acid could be decomposed within 15 min. However it is not advisable to reduce the time needed for the complete decomposition of ascorbic acid by increasing the temperature further as other constitituents commonly found in fruit juices may also be decomposed. Because citric acid is present in nearly all the fresh and canned fruit juices the complete copper(I1)-catalysed oxida-tion of ascorbic acid for the determination of its content in real samples should be carried out at 50 "C.It is expected that tartrate and acetate will not cause any interference at 50 "C as both of them form less stable complexes with copper(I1) than citrate. The effects of a number of chemicals commonly present in soft drinks or fruit juices were examined by applying the method to a 10 yg ml-1 ascorbic acid solution at pH 6 and containing the copper(I1) - EDTA complex (5.2 X 10-5 M) and varying the concentration of each substance being studied. Results of the interference studies are shown in Table 6. The criterion for an interference was an absorbance varying by +5% from the expected value. It is evident that none of the chemicals cause any interference at the levels studied.It was noted that components that are UV active such as fumaric acid tannic acid caffeine saccharin colouring matter etc. cause high background absorption. However if they have no side reaction with the copper(I1) ion they will not affect the net absorbance of ascorbic acid. Further their concentrations in fruit juices are low and it is unlikely that they will cause serious interference and their effect can usually be minimised by diluting the sample where necessary. It is also worth noting that tannic acid and caramel which interfere in the alkaline treatment method of Fung and Luk,l0 did not interfere at the concentration given in Table 6. Calibration Graph and Precision As the standard solutions of ascorbic acid were prepared to contain a known concentration of the copper(I1) - EDTA complex the absorbance due to ascorbic acid could be readily obtained by making the absorbance measurement against the reagent blank.As noted previously only 1% of ascorbic acid decomposed within 1 h at 25 "C in the presence of the Cu(I1) -EDTA complex solution (5.2 x 10-5 M) and the error resulting from the oxidation catalysed by the copper(I1) complex would be very small. The calibration graph passed through the origin and was linear up to 20 pg ml-1 of ascorbic acid. Its slope was 0.086 (p.p.m.)-l and gave a correlation coefficient of 0.9995. The molar absorptivity at 267 nm for ascorbic acid was calculated using the equation A = ECL to be 1.52 x 104 1 mol-1 cm-1 ANALYST JUNE 1986 VOL.111 669 Table 8. Determination of the content of ascorbic acid in fresh fruit juices canned fruit juices soft drinks and cordials ( i ) Fresh fruit juice: Ascorbic acid content/pg ml-* Sample Grapefruit . . Lemon . . . . Orange . . . . Proposed Polarographic method method DCIP" . 347.8 335.0 366.9 . . 541.4 517.5 543.4 . . 600.4 580.0 606.0 (ii) Canned fruit juices: Ascorbic acid contentlpg ml-Sample Apple juices: BrandA . . . . . . B . . . . . . c . . . . . . C + 60p.p.m. of ascorbicacid . . D . . . . . . D + 60p.p.m. of ascorbicacid . . E . . . . . . E + 60 p.p.m. of ascorbicacid . . BrandA . . . . . . B . . . . . . B + 300p.p.m. of ascorbicacid . . c . . . . . . C + 300 p.p.m.of ascorbicacid . . BrandA . . . . . . B . . . . . . Prunejuice . . . . . . (iii) Soft drinks: Alkaline electrolyte drink . . Blackcurrant . . . . . . Guavajuicedrink . . . . Grapefruit juices: Orange juices: Honeylemon . . . . . . Kalamansijuice . . . . Lemondrink . . . . . . Lemon tea: BrandA . . . . . . B . . . . . . BrandA . . . . . . B . . . . . . BrandA . . . . . . B . . . . . . c . . . . . . D . . . . . . Mango juice drinks: Orange juice drinks: E . . . . . . Sugarcane . . . . . . (iv) Cordials: Grape . . . . . . . . Honeylemon . . . . . . Lemon . . . . . . . . Lime . . . . . . . . Orange . . . . . . . . * DCIP titration with 2,6-dichloroindophenol. t NBS titration with N-bromosuccinimide. . . .. . . . . * . . . . . . . . . . . * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proposed method 0 20.6 0 57.9 0 60.5 6.6 70.2 291.8 359.5 659.9 253.0 541 .O 502.1 492.0 0 484.1 432.0 211.7 11.6 224.7 18.6 141.8 264.3 290.5 319.3 82.4 362.4 332.8 592.7 480.9 85.2 0 105.6 54.4 0 27.0 DCIP* -Not detected -307.3 356.8 650.2 252.9 544.7 526.2 482.7 -492.8 222.8 12.0 229.9 18.6 -146.4 272.2 306.3 322.9 84.9 377.0 341.2 607.5 489.8 82.0 -107.2 56.3 0 -The precision of the procedure was checked by calculating the relative standard deviation of ten replicate determinations of several standard solutions and a real sample.The results for the 5.0 8.3 and 13.3 pg ml-1 standard ascorbic acid solutions were found to be 0.5 0.1 and 0.2% respectively. The precision for the determination of ascorbic acid in a canned fruit juice was O.6% where the average ascorbic acid content in the sample was 555 pg ml-1 670 ANALYST JUNE 1986 VOL. 111 Recovery Tests Recovery tests using the proposed method were performed using four different samples and the test for each sample was carried out in triplicate. As shown in Table 7 the recoveries of ascorbic acid added to lime cordial orange drink and lemon tea were all close to 10070 whereas the result for blackcurrant drink was 98.5%. The results of the recovery tests are very good.Determination of Ascorbic Acid in Soft Drinks Fresh and Canned Fruit Juices and Cordials The ascorbic acid contents of some soft drinks cordials and fresh and canned fruit juices were determined by the proposed method. The results are summarised in Table 8. The results for most samples were checked by a titrimetric method using 2,6-dichloroindophenol2 as the titrant. However titration with N-bromosuccinimide20 was employed for highly coloured samples where titration using 2,6-dichloroindophenol failed. The results for fresh fruit juices were also checked by the differential-pulse polarographic method.5 Each determina-tion was carried out in duplicate. There is a close agreement between the results obtained by the proposed method and the established methods indicating the accuracy of the proposed method.It was noted that accurate results were obtained for grapefruit and apple juices. Thus the proposed method is more widely applicable than the alkaline pre-treatment background correction method proposed by Fung and Luk,lo as their method was unsuitable for the determination of ascorbic acid in these two types of sample.’’ The proposed method is also applicable to coloured samples where the simple titrimetric method using 2,6-dichloroindophenol is unsuitable because end-point detection is difficult if not impossible. Further the proposed method is very simple as it requires no special instrument or reagents. However the method did not work well for samples such as prune juice, which absorb very strongly in the UV region.Conclusion A simple rapid precise and accurate method has been developed for the determination of ascorbic acid in soft drinks fruit juices and cordials using direct UV spectropho-tometry. The background correction method recommended is one in which the sample blank for the determination of ascorbic acid is produced by the catalytic decomposition of the acid by copper(I1). The decomposition of ascorbic acid in real samples should be carried out at 50 “C because citric acid, which is usually present in these samples retards the copper(I1)-catalysed oxidation of ascorbic acid. The reagents used are cheap safe stable and readily available in most laboratories. The proposed method is selective and many ingredients commonly found in soft drinks fruit juices and cordials do not interfere; it is also applicable to a wide range of samples.The working range of the method is wide and can cover the normal concentration range of ascorbic acid present in various types of samples. The method is particularly attractive to those less well-equipped laboratories possessing UV - visible spectro-photometers but not other more sophisticated instruments such as polarographic analysers or spectrofluorimeters. The proposed method should also be applicable to other types of beverages in which the matrices do not absorb or scatter strongly in the UV region. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. References Lehninger A.L. “Biochemistry,” Second Edition Worth Publishers New York 1975 p. 259. Horwitz W. Editor “Official Methods of Analysis of the Association of Official Analytical Chemists,” Thirteenth Edition Association of Official Analytical Chemists Washing-ton DC 1980 p. 43.056. Horwitz W. Editor “Official Methods of Analysis of the Association of Official Analytical Chemists,” Thirteenth Edition Association of Official Analytical Chemists Washing-ton DC 1980 p. 43.061. Lindquist J. and Farroha S. M. Analyst 1975 100 377. Sontag G. and Kainz G. Mikrochim. Acta 1978 I 175. Baczyk S . and Swidzinska K. Mikrochim. Acta 1975 I 259. Tono T. and Fujita S . Agric. Biol. Chem. 1981 45 2947. Tono T. and Fujita S . Agric. Biol. Chem. 1982 46 2953. Delaporte N. and Macheix J . J. Chirn. Anal. 1968,50 187; Anal. Abstr. 1969 17 1162. Fung Y.-S. and Luk S.-F. Analyst 1985 110 201. Luk S. F. M . Phil. Thesis 1985. Weissberger A. and LuValle J. E. J. A m . Chem. SOC. 1944, 66 700. Onishi I. andHara T. Bull. Chem. SOC. Jpn. 1964,37,1314. Onishi I . andHara T. Bull. Chem. SOC. Jpn. 1964,37,1317. Moussa A. F. A. Pharmazie 1977 32 724; Anal. Abstr., 1978 34 5E6. Kumler W. D. and Daniels T. C. J. Am. Chem. SOC. 1935, 57 1929. Fogg A. G. and Summan A. M . Analyst 1983 108 691. Ringbom A. “Complexation in Analytical Chemistry,” Inter-science New York 1963 p. 355. Tressler D. K. and Joslyn M. A. Editors “Fruit and Vegetable Juice Processing Technology,” Second Edition AVI Publishing Co. Westport CT 1974 p. 43. Barakat M. Z . El-Wahab M. F. A and El-Sadr M. M., Anal. Chem. 1953 27 536. Paper A5133 I Received September 16tk 1985 Accepted December 31st 198

ISSN:0003-2654    

DOI:10.1039/AN9861100665

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST JUNE 1986 VOL. 111 67 1 Trace Determination of Hydroperoxides by Spectrophotometry in Organic Media Jaroslav Petrdj Chemopetrol Research Institute of Macromolecular Chemistry Tkalcovska 2 656 49 Brno, Czechoslo va kia Susanne Zehnacker Laboratoire d’Etude de la Degradation et de la Stabilisation des Polymeres lnstitut C. Sadron CNRS, 6 Rue Boussingault 67083 Strasbourg Cedex France Jiii Sedlai Chemopetrol Research Institute of Macromolecular Chemistry Tkalcovska 2 656 49 Brno, Czechoslovakia and Jean Marchal Laboratoire d‘Etude de la Degradation et de la Stabilisation des Polymeres lnstitut C. Sadron CNRS, 6 Rue Boussingault 67083 Strasbourg Cedex France Two spectrophotometric methods for the trace determination of hydroperoxides in benzene - methanol medium are compared.The first is based on complex formation between Fell and 1,lO-phenanthroline (tentatively denoted by FellP) and the other makes use of the complex formed between Fell1 and the thiocyanate anion (tentatively denoted by FelllT). The thiocyanate method when modified by adding more benzene to the system and employing active silver to suppress the blank values proved to be less laborious, more sensitive and faster than the determination with 1,IO-phenanthroline. The yield of FelIlT is almost stoicheiometric which is not so for the Fe”P complex. Owing to its simplicity great sensitivity and reproducibility the modified thiocyanate method is suitable for series determinations. Keywords Hydroperoxide determination; spectrophotometry Spectrophotometric methods making use of coloured iron complexes have been used extensively for the determination of peroxides that are readily reduced by iron(I1) ions (organic hydroperoxides hydrogen peroxide peroxy acids diacylper-oxides).Employing this approach either the concentration of iron(II1) ions formed is determined after their complexation with NH4SCNI-12 or the consumption of iron(I1) ions is deduced from the absorbance of the FeII - 1,lO-phen-an throline complex. 1 13,14 Although the reaction of hydroperoxides (ROOH) is faster with iron(I1) ions than with other transition metal i0ns~15-19 it is difficult to achieve the ideal stoicheiometry of the funda-mental reactions: (1) . . (2) In practice the alkoxy radicals RO’ undergo competitive reactions producing free radicals which may induce ROOH decomposition and which if oxygen is present are capable of producing an additional amount of hydroperoxides.20 Hence the over-all mechanism is complex the individual reactions being sensitive to the reaction medium hydroperox-ide structure and oxygen concentration.Kolthoff and Medalias7 presented a detailed analysis of error sources and showed the idiosyncrasies of this method. In methods for ROOH determination the polarity of the reaction medium should be reduced as much as possible. This is of particular importance for the determination of hydroper-oxides in organic and even polymeric materials where a non-polar medium is necessary. The thiocyanate method was originally developed for work in acetone5 or methanoliclO media.Bolland et a1.,3 Smiths and Lea9 used benzene -methanol (7 + 3) and Lebelll reduced the proportion of methanol to 12%. Zeppenfeldl2 reduced the concentration of methanol even more (8%) for the determination of hydroper-oxides in PVC. BoCekl4 adapted the 1 10-phenanthroline ROOH + Fez+ + H+ + RO’ + Fe3+ + H20 . . RO’ + Fez+ + H+ + ROH + Fe3+ method for use in benzene - methanol (8 + 2) and applied it to the determination of hydroperoxides in atactic polypropylene. We have modified the method of Zeppenfeldl2 by reducing the proportion of methanol to 4% and by using active silver.14 The low polarity of the solvent system in our method makes it particularly suitable for the determination of organic hydroperoxides in degraded polyolefins which will be the subject of our next paper.21 The use of benzene although a suspected carcinogen in previous work and in this work is deliberate because no structure less inert towards oxidation can be found among the commonly available non-polar solvents.The requirement of negligible oxidisability of the reaction medium for trace determinations of hydroperoxides is obvious. In this paper we present the results of determinations of three model hydroperoxides and compare them with those obtained when using the phenanthroline method due to BoCe k. 14 Experimental Reagents 75% solution in di-tert-butyl peroxide. 80% solution in cumene. tert-Butyl hydroperoxide. Fluka “for synthesis” grade. Cumyl hydroperoxide. Merck analytical-reagent grade, Hydrogen peroxide. Merck analytical-reagent grade 30%.Standard solutions of hydroperoxides ca. 0.1 M. Prepared by appropriate dilution with methanol from which oxygen was removed by stripping with an inert gas (nitrogen). The actual concentrations of hydroperoxides in these standards were determined iodimetrically (see below). Solvents. The spectrophotometric microanalyses were car-ried out either after a 1000-fold dilution of the standard with benzene or by using a Hamilton 701 microsyringe (10 pl) fo 672 ANALYST JUNE 1986 VOL. 111 direct sampling of the undiluted standard. In the latter instance the tip of the needle had to be immersed in benzene in order to obtain satisfactory results. The solvents i.e. benzene (Prolabo Normapur) and methanol (Merck analytical-reagent grade) were stripped with nitrogen for 3 h prior to use.Caution-Benzene is a suspected carcinogen and appropriate precautions should be taken. Nitrogen. Air Liquide. U-grade O2 < 5 p.p.m. Used without further purification. Acids. Acetic acid (glacial) sulphuric acid (sp. gr. 1.83), nitric acid (sp. gr. 1.33) and orthophosphoric acid (sp. gr. 1.70) were Prolabo products of analytical-reagent grade. Sodium thiosulphate solution 0.1 M. Prepared by dissolving a Normanal standard (Merck) in water. Iron(III) chloride anhydrous. Merck analytical-reagent grade. Used without further purification. Cumene. Roth purity > 99%. Purified prior to use by passage through a column packed with Woelm W 200 basic alumina. Other reagents. Potassium iodide ammonium iron(I1) sulphate [Fe(NH4)2(S04)2.6H20] ammonium thiocyanate, silver nitrate and 1,lO-phenanthroline were Prolabo products of analytical-reagent grade. Active silver. Prepared according to the procedure des-cribed by BoCek.14 A 30-g amount of silver nitrate is dissolved in 400 ml of water acidified with HN03. The solution is stirred with a sheet of copper (electrolytic grade). The silver powder that precipitates is then washed several times with distilled water until no positive response to copper(I1) ions in the filtrate is observed. The powder is further washed with methanol and dried. The active silver thus prepared is used to remove undesir-able Fe3+ ions from the reagent solutions of Fez+. For this purpose the solution is either filtered through a 5-cm layer of active silver or it is shaken directly with the silver powder using a special vessel (Fig.1) that permits rapid pipetting of the solution without contamination of the liquid with sedimented silver. When this method is applied which was preferred in most instances it is advisable to use only the minimum amount of silver necessary to produce discolouration of the reagent solution. I L P i p e t t e SR 5 cm -Fig. 1. Design of vessel for preparation of the reagent solution with active silver Apparatus The spectrophotometric measurements were carried out on a Zeiss PMQ I1 spectrophotometer with an automatic slit setting. The differential measurements were made at wavelengths corresponding to the absorption maxima of the complexes [512.5 nm for Fe"' - thiocyanate (FeIIIT) and 500 nm for Fe" - 1,lO-phenanthroline (FeIIP)].The reference cell, path length 1 or 0.5 cm was filled with pure benzene. Procedures Iodimetry This method was assumed to be absolute and was employed for the standardisation of hydroperoxide solutions. The oxidation of KI was carried out in acetic acid medium and the liberated iodine was determined titrimetrically with 0.1 M Na2S203 solution. The experimental details can be found elsewhere .22-24 Phenanthroline method according to Boc'ek14 The following solutions are first prepared. Fe2+ standard 5 x 10-4 M. Weigh accurately 490 mg of Fe(NH4)2(S04)2-6H20 into a 25-ml calibrated flask add a few millilitres of methanol and 0.35 ml of concentrated H$04 and shake well. After dissolution of the salt dilute to the mark with methanol.Before use dilute this solution 100-fold with methanol. 1,lO-Phenanthroline solution. Prepare a 0.5% mlV solution of 1 ,lo-phenanthroline in benzene. H3P04 solution 4 x 10-3 M. Weigh accurately 46 mg of orthophosphoric acid into a 100-ml calibrated flask and dilute to the mark with methanol. Procedure. Place a known amount of the solution to be analysed into a 25-ml calibrated flask that contains a few millilitres of benzene. The total amount of hydroperoxides should be of the order of 0.05-1.0 pmol per flask. Adjust the volume of liquid in each flask to 10-12 ml with benzene. Add 0.5 ml of H3P04 solution and 5.00 ml of Fe2+ standard solution; the latter is carried out by the means of a burette filled with active silver.Wash the walls of the flask with benzene and stir the contents well. After 15 min (or 3 h see Results and Discussion) add 1 ml of 1,lO-phenanthroline solution and dilute to the mark with benzene. Wait a further 30 min and measure the absorbance of the red complex formed at 500 nm. Ammonium thiocyanate method (this work) The following solutions are first prepared. Solution A Fe2+ solution (10-2 M Fez+) and H2S04 (4.5 x M) in methanol. Weigh 0.4 g of (NH4)2Fe(S04)-6H20 into a 100-ml calibrated flask and add a few millilitres of methanol and 0.25 ml of concentrated H2SO4. Dilute to the mark with methanol. Solution B NH4SCN solution 2 x 10-1 M. Weigh 1.5 g of ammonium thiocyanate into a 100-ml calibrated flask and dilute to the mark with methanol.Reagent solution. Always prepare this freshly before each series of analyses by mixing equal volumes of Fe2+ and NH4SCN solutions preferably in the vessel shown in Fig. 1. Add a small amount of active silver (see Reagents) to remove the red colouration that develops after mixing. After the sedimentation of silver powder the reagent is ready for use. Procedure. Introduce a known volume of the hydroperox-ide solution into a 25-ml calibrated flask containing 10-15 ml of benzene. The total amount of hydroperoxides taken for the determination should be within the range 0.02-1.0 pmol. Add 1 ml of the decoloured reagent solution. Avoid contamination of the ground-glass joint of the flask with undiluted reagent as its subsequent oxidation by air may be a source of serious errors.It is therefore advisable to wipe off the external part o ANALYST JUNE 1986 VOL. 111 673 the pipette tip prior to dosage. Pipetting of the reagent should be carried out just above the benzene level in the flask. Stir the contents and dilute to the mark with benzene. Allow the solution to stand in darkness for 15-30 min then measure its absorbance at 512.5 nm. The red complex (FeIIIT) is sensitive to light so it is advisable to reduce the intensity of the incident beam in the spectrometer by using as narrow a slit width as possible (in our case 0.02 mm) and to protect the solutions from undesirable exposure to light. Calculation The absolute amount y (pmol) of complex formed in the flask is given by the relationship where AEl represents the difference between the molar absorptivities of the solution being analysed and that of blank (reagent diluted with benzene) as measured in 1-cm cells E being the molar absorptivity of the complex.The E values measured at the wavelength of maximum absorption were determined by using standard Fe" (for FeIIP) and Fe"' (for FeIIIT) solutions. The absorbance of the two complexes complies with the Beer - Lambert law over the whole of the investigated concentration range. For the complex FeIIP the value of E calculated by the method of least squares was found to be E~~ nm = (1.12 k 0.03) x l o 4 1 mol-1 cm-1. This is in good agreement with the result obtained by BoCek14 of (1.10 _+ 0.02) x 104 1 mol-1 cm-1. For the FeII'T complex the value found was E512.5 nm = (1.68 k 0.06) X lo4 1 mol-1 cm-1 which differs from that reported by Zeppenfeldl2 of (1.34 k 0.04) x l o 4 1 mol-1 cm-1.If the amount x (pmol) of the hydroperoxide introduced into the flask is known (i.e. determined iodimetrically) then it is possible to calculate the experimental value of the stoicheiometric factor e = y/x for a selected spectrophoto-metric method. This value is then used for the calculation of results bearing in mind that it may be different from the theoretical value of 2 for the reasons already mentioned. The concentration of hydroperoxide in the spectro-photometrically analysed sample is then calculated from the following relationship: 2.5 x 1 0 7 ~ 4 e E h l ax. cROOH (mmol-1) = . . . . (4) where v p1 is the volume of the sample in the flask.Results and Discussion 1,lO-Phenanthroline Method When this method was applied in its original form as described by BoCek14 to solutions of model hydroperoxides or to peroxidised atactic polypropylene samples we observed that the recommended reaction time of 15 min was not sufficient for constant stoicheiometry to be achieved. Experiments with solutions of model hydroperoxides25 have shown that the reaction of peroxide with Fe*+ requires at least 2 h before it may be considered quantitative. For this reason, we allowed a minimum of 3 h for the reaction before adding the complexing agent. Only under these conditions could we observe a linear relationship between the concentration of the I' / / 2 - / / 0 0.5 1 .o 106xlmol Fig.2. Formation of FeI* - 1,lO-phenanthroline complex as a function of tert-butyl hydroperoxide level. Reaction time 1 15 min; and 2 3 h. 0 Pipette; 0 microsyringe. Broken line theoretical stoicheiometry 2 1 Table 1. Determination of stoicheiometric coefficient e for tert-butyl hydroperoxide analysed by the 1 ,lo-phenanthroline method. Symbols cROOH = concentration of standard solution; S = microsyringe (pl); P = pipette (ml); t = time of reaction; x = amount of hydroperoxide in the flask; y = amount of complex formed Volumelpl (S) Series No. CROOH/M or ml (P) xlpmol AEI ylpmol e 1 1.58 x 10-1 (S; t = 15 min) 1 0.158 0.176 0.393 2.49 2 0.316 0.350 0.781 2.47 3 0.474 0.494 1.103 2.33 4 0.632 0.626 1.397 2.21 5 0.790 0.726 1.620 2.05 Av.-* 2 1.58 X lo-* (S; t = 3 h) 1 0.158 0.185 0.413 2.61 2 0.316 0.366 0.817 2.59 3 0.474 0.534 1.192 2.51 4 0.632 0.733 1.636 2.59 5 0.790 0.898 2.004 2.54 Av. 2.57 k 0.04 3 1.75 X (P; t = 3 h) 1 0.175 0.200 0.446 2.55 2 0.350 0.428 0.955 2.73 3 0.524 0.622 1.388 2.65 4 0.699 0.860 1.919 2.75 5 0.874 1.012 2.259 2.58 Av. 2.65 k 0.09 * No average value given e decreases systematically (see Fig. 2 graph 1) 674 ANALYST JUNE 1986 VOL. 111 Table 2. Determination of stoicheiometric coefficient e for cumyl hydroperoxide analysed by the 1 ,lo-phenanthroline method. Symbols as in Table 1 Series No. CROOHIM Volumelml xlpmol AEl ylymol e 4 1.613 X (P f = 3 h) 1 0.161 0.244 0.545 3.38 2 0.324 0.464 1.036 3.21 3 0.484 0.698 1.558 3.22 4 0.645 0.922 2.058 3.19 5 0.806 1.060 2.366 2.94 Av.3.19 k 0.16 Fe"P complex and that of the hydroperoxides actually present. The need to prolong the reaction time to 2 h was also noted by Amin et ~1.~26 who applied this method to the determination of hydroperoxides in low-density polyethylene (LDPE). They attributed this phenomenon to an excessively slow - . penetration of the reagent into the films. OG results obtained with model hydroperoxides in benzene solutions are given in Table 1 and Fig. 2 for tert-butyl hydroperoxide and in Table 2 for cumyl hydroperoxide. It follows from Fig. 2 that at higher hydroperoxide concentrations (exceeding 5 x 10-7 mol per 25 ml) a considerable deviation from a straight-line dependence occurs when a reaction time of only 15 min is allowed.On the other hand no differences in the results is noted at lower hydroper-oxide levels whether the reaction was carried out for 15 min or 3 h. This clearly indicates that under the given conditions the effect of oxygen on the results is negligible and that the curvature of line 1 is due to an insufficient time period being allowed for the reaction. The stoicheiometric factor e calcu-lated for t = 3 h is on the other hand fairly constant (see Table l) but is greater than the theoretical value. Moreover, it seems to depend on the nature of the hydroperoxide being analysed (compare Tables 1 and 2). At this stage it is difficult to decide whether the stoicheiometric factor depends on the nature of the hydroperoxide or also on the other components present in commercial products.It is our opinion that the concentration of Fez+ recom-mended by BoCek14 is too low with regard to that of the hydroperoxide to be determined. For example in a 25-ml flask the total amount of Fez+ available for the ROOH reduction is 2.5 x mol which in the most unfavourable instance Le. for 1.0 X 10-6 mol of ROOH to be determined, represents an excess by a factor of only 1.25. This insufficient excess of Fez+ is believed to be responsible for the above-described phenomena (curvature long reaction time require-ments). It can be concluded however that in spite of the above-mentioned disadvantages the 1 ,lo-phenanthroline method is sufficiently reproducible and precise. This follows from Table 1 and Fig. 2 which show the results obtained using different methods of sampling (direct or with a syringe).Unfortunately the working procedure is relatively compli-cated and time consuming. As the method is based on the measurement of FeII consumption it is necessary to know precisely the amount of the reagent added initially. The major disadvantage however is the fact that the strongest coloura-tion is always observed at zero concentration of ROOH. Therefore when very low concentrations of hydroperoxides are determined one has to compare two relatively high absorbances which obviously reduces the precision of the determination. Ammonium Thiocyanate Method Methods involving the use of the reaction of formation of an FeIII - thiocyanate complex give a colouration that increases with increasing concentration of the hydroperoxide being determined.They are therefore better adapted for analyses at relatively low hydroperoxide levels. Another advantage of 0 1 2 3 4 1 03Csc~-/M Fig. 3. Molar absorptivity (E) for the FeIII- thiocyanate complex as a function of NH4SCN concentration. [Fe3+] = 8 X lop5 M; [HZS04] = 9 x 10-4 M. 1 Fez+ absent; and 2 [Fez+] = 1.2 X M these methods is that the amount of Fez+ reagent does not have to be known exactly and that it may be applied in an excess with reagent to ROOH. The excess of Fez+ accelerates reaction (2) thus suppressing undesirable reactions of alkoxy radicals. Under these conditions the experimental value of the stoicheiometric factor should approach the theoretical value of 2. When applying the method of Zeppenfeldl2 we observed in accordance with his results that a deposit was slowly formed during the experiments this being particularly undesirable for the spectrophotometric measurements.Our analysis of this precipitate has shown that it is ammonium sulphate. As the solutions contain an excess of NH4+ it is obvious that its formation will be controlled by the S042- content in the system. We therefore investigated the relative proportions of iron(I1) ions thiocyanate ions and H2SO4 that would ensure optimal conditions for the analysis. An increased concentration of Fe2+ up to the solubility limit of Fe(NH4)2(S04)2.6H20 in methanol favours reactions (1) and (2) which should bring the experimental stoicheiometric factor close to the theoretical vahe.The more concentrated Fe2+ solution is however oxidised more readily which in turn leads to increased blank values. Moreover more H2S04 is required in addition to an increase in the NH4SCN concentration; under these conditions the precipitation of (NH4)2S04 is considerably enhanced. After a series of experiments we concluded that the optimum reagent composi-tion after mixing stock solutions A and B (1 + 1) (see Experimental) is 5 x 10-3 M Fe2+ 2.25 x 10- M H2S04 and 10-1 M NH4SCN. It is worth noting that on adding 1 ml of this reagent to a 25-ml flask containing the recommended amount of hydroperoxide the available concentration of Fez+ is at least twice as high as that in the 1,lO-phenanthroline method. The excess (4.5-fold) of H2S04 with respect to Fe2+ ensures the solubility and stability of the reagent and also a sufficient concentration of H+ in reactions (1) and (2).A 20-fold excess of NH4SCN is recommended on the basis of a study in which the absorption intensity of the FeIII ANALYST JUNE 1986 VOL. 111 675 complex was measured as a function of NH4SCN concentra-tion (see Fig. 3 line 1). The introduction of 1 ml of this reagent into a 25-ml flask represents 5 X 10-6 mol of Fez+. In the proposed solvent mixture it is possible to increase this amount to 1.5 ml without the risk of precipitation of (NH4)2S04 which begins to be significant after the addition of 2 ml or more. The maximum amount of ROOH taken for the determina-tion should not exceed 10-6 mol. This corresponds to the 1.4 ' I I 2 0 20 40 60 [ C H 3 0 H ] ,% Fig.4. Molar absorptivity (E) for the Fe"' -function of methanol concentration in benzene. [Fez+] = 1.2 X l o - 4 ~ ; [H2S04] = 9 x l o - 4 ~ ; M 0 0.5 1 .o 1 O6xim o I Fig. 5. Formation of FelI1 - thiocyanate complex as a function of tert-butyl hydroperoxide level. 0 Pipette; and 0 microsyringe oxidation of 2 x 10-6 mol of Fe2+ i. e. 40% conversion (if 2 1 stoicheiometry is assumed). As the intensity of colouration can be influenced by the excess of Fez+ ions we measured the concentration of NH4SCN that is necessary to obtain the maximum intensity of colouration when 2 X 10-6 mol of Fe3+ are to be complexed in the presence of 3 X 10-6 mol of Fez+ in a 25-ml flask. It can be seen from Fig. 3 line 2 that the recommended concentration of 4 X 10-3 M NH4SCN is sufficient.Similarly the effect of methanol concentration on the intensity of colouration of the FeIIIT complex was followed. As shown in Fig. 4 the methanol to benzene ratio may be increased to 1 2 without any significant influence. Finally we recommend carrying out the spectrophoto-metric measurements no sooner than 15-30 min after the addition of reagent. After this period no further colour changes are observed if the samples are sufficiently protected from artificial light or daylight. In this way they may be stored, if necessary for up to 2 h without introducing any additional errors. The oxidation of Fez+ by hydroperoxides (including H202) and peroxy acids is almost instantaneous whereas diacyl-peroxides react less readily.Peroxides other than those mentioned do not react at a11.12 The results of the analysis of three model hydroperoxides by employing the modified ammonium thiocyanate method are presented in Table 3 and Fig. 5 for tert-butyl hydroperoxide in Table 4 for H202 and in Table 5 for cumyl hydroperoxide. These results confirm that the reaction time of 15 min is sufficient. The concentration of the FeIIIT complex is directly proportional to the level of hydroperoxide in all three instances. The values of the stoicheiometric factor e are very close to the theoretical value of 2 for tert-butyl hydroperoxide and H202. The higher value of e ~ 2 . 2 5 found for cumyl hydroperoxide was originally believed to be caused by additional autoxida-tion of a sensitive substrate (cumene) present in the commer-cial product.Experiments in which pure cumene was added to the analysed hydroperoxide showed however that this is not so. No significant influence on the value of e was observed even when cumene was present at levels exceeding that of the hydroperoxide by several orders of magnitude (see Table 6). Hence the value of the stoicheiometric factor e seems to be an inherent property of the hydroperoxide being determined and it should be determined for each experiment using an absolute met hod. Table 3. Determination of stoicheiometric coefficient e for tert-butyl hydroperoxide analysed by the ammonium thiocyanate method. Symbols as in Table 1 Volume/yl (S) Series No. CROOH/M or ml (P) xlymol AEl ylymol e 5 1.028 X 10-l (S) 6 1.750 X 10-4 (P) 4 0.411 0.543 0.808 1.97 6 0.617 0.858 1.277 2.07 8 0.822 1.153 1.716 2.09 10 1.028 1.440 2.143 2.06 1 0.175 0.211 0.314 1.79 2 0.350 0.461 0.686 1.96 3 0.525 0.709 1.055 2.01 4 0.700 0.931 1.385 1.98 5 0.875 1.157 1.722 1.97 Av.2.05 f 0.06 Av. 1.94 k 0.09 Table 4. Determination of stoicheiometric coefficient e for hydrogen peroxide analysed by the ammonium thiocyanate method. Symbols as in Table 1. Series No. CROOH/M Volume/pl xlymol AEl ylymol e 7 9.10 X (S) 4 0.364 0.482 0.717 1.97 6 0.546 0.741 1.103 2.02 8 0.728 0.990 1.473 2.02 10 0.910 1.230 1.830 2.01 Av. 2.00 k 0.0 676 ANALYST JUNE 1986 VOL. 111 Table 5. Determination of stoicheiometric coefficient e for cumyl hydroperoxide analysed by the ammonium thiocyanate method.Symbols as in Table 1. Volumelpl (S) Series No. CROOH/M or ml (P) xlymol AEl ylvmol e 8 1.125 x 10-1 (S) 4 0.450 0.661 0.984 2.19 6 0.675 1.035 1.540 2.28 8 0.900 1.324 1.970 2.19 10 1.125 1.635 2.433 2.16 Av. 2.20 k 0.05 9 1.613 x 10-4 (P) 1 0.161 0.232 0.345 2.14 2 0.323 0.495 0.737 2.28 3 0.484 0.758 1.128 2.33 4 0.645 0.990 1.473 2.28 5 0.806 1.225 1.823 2.26 Av. 2.26 k 0.07 Table 6. Influence of cumene on the stoicheiometric coefficient e for cumyl hydroperoxide analysed by the thiocyanate method. Symbols as in Table 1; c = amount of cumene in the flask Series No. xlpmol clymol AEl Y clx e 10 0.313 0.313 0.313 0.313 0.313 0.1* 0.470 0.699 0.32 2.23 35.3 0.455 0.677 11.0 2.16 70.5 0.450 0.669 22.1 2.14 176.1 0.440 0.655 55.1 2.09 352.1 0.460 0.684 110.2 2.19 Av.2.16 k 0.04 * The amount of cumene introduced when sampling commercial cumene hydroperoxide. Conclusion The modified ammonium thiocyanate method for the trace determination of hydroperoxides in organic media is less laborious more sensitive and faster than the method due to BoEekl4 making use of the Fe”’ - phenanthroline complex. The yield of the coloured complex is almost stoicheiometric (e = 2.0) for H202 and tert-butyl hydroperoxide. A higher value (e = 2.25) is obtained for cumyl hydroperoxide the mechanism of the reduction of which by Fez+ is probably more complicated than that considered in equations (1) and (2). It is our opinion that the ideal stoicheiometry (e = 2.0) is probably unattainable for the general case when Fe’J oxidation is employed.Therefore a calibration by an absolute method is always necessary if more precise absolute values are required. We believe however together with Lea,9 that for most purposes the sensitivity reproducibility rapidity and sim-plicity are the most important features. References 1. Martin A. J. “Organic Analysis,” Volume IV Wiley-Interscience New York 1960 pp. 1-64. 2. Swern D. “Organic Peroxides,” Volume 11 Wiley-Interscience New York 1971 pp. 579-635. 3. Bolland J. L. Sundralingam A. Sutton D. A. and Tristram, G. R. Trans. Inst. Rubber Ind. 1941 14 29. 4. Chapman R. A. and McFarlane W. C. Can. J. Res. Sect. B, 1943 21 133. 5. Kolthoff I. M. and Medalia A. I. Anal. Chem. 1951 23, 595. 6. Kolthoff I. M. and Medalia A.I. J. Am. Chem. SOC. 1949, 71 3777,3784 and 3789. 7. Medalia A. I. and Kolthoff I. M. J. Polym. Sci. 1949 4, 377. 8. Smith G. H. J. Sci. Food Agric. 1952 3 26. 9. Lea C. H. J. Sci. Food Agric. 1952 3 586. 10. Young C. A. Vogt R. R. and Nieuwland J. A. Znd. Eng. Chem. Anal. Ed. 1936 8 198. 11. Lebel P. Thbe No. d’Ordre 543 University of Paris 1957. 12. Zeppenfeld G. Makromol. Chem. 1966 90 169. 13. Laitinen H. A. and Nelson J. S. Ind. Eng. Chem. Anal. Ed. 1946 18 422. 14. BoCek P. Chem. Prtlm. 1967 17 439. 15. Swern D. “Organic Peroxides,” Volume 11 Wiley-Interscience New York 1971 pp. 102-152. 16. Denisov E. T. and Emanuel N. M. Russ. Chem. Rev. 1960, 29 645. 17. Hiatt R. Mill T. and Mayo F. R. J . Org. Chem. 1968,33, 1416. 18. Hiatt R. Irwin K. C. and Gould C. W. J. Org. Chem., 1968 33 1430. 19. Kochi J. J. Am. Chem. SOC. 1962 84 1193. 20. Swern D. “Organic Peroxides,” Volume 11 Wiley-Interscience New York 1971 pp. 153-268. 21. Petrfij J. Zehnacker S . SedlBi J. and Marchal J. Polym. Deg. Stab. in the press. 22. Bartlett P. D. and Altschul R. J. Am. Chem. SOC. 1945,67, 812. 23. Cass W. E. J. Am. Chem. SOC. 1946 68 1976. 24. Swern D. “Organic Peroxides,” Volume 11 Wiley-Interscience New York 1971 p. 585. 25. Petrfkj J . unpublished results. 26. Amin M. U. Scott G. and Tillekeratne L. M. K. Eur. Polym. J . 1975 11 85. Paper A51130 Received April 1 Oth 1985 Accepted January 17th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100671

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST JUNE 1986 VOL. 111 677 Ternary Complexes in Solution Complex Formation Between the 1 1 Thorium(lV) - Alizarin Maroon Complex and Thiosalicylic Acid Mohamed M. Seleim Kamal A. Idriss Magda S. Saleh and Elham Y. Hashem Department of Chemistry Faculty of Science Assiut University Assiut Egypt The stability constant of the ternary Th(lV) complex containing alizarin maroon (AZM) and as a second ligand thiosalicylic acid (TSA) in a 1 1 1 molar ratio was determined potentiometrically in 40% V/Vethanol -water medium ( I = 0.1 M NaC104; 25 _+ 0.1 "C). Under identical conditions the binary complexes of thorium with both reagents have also been examined. For the equilibrium Th(AZM);! + Th(TSA)2 2 Th(A2M) (TSA) the corresponding constant was log X = 3.6(0.75). The constant given in parentheses is due to AlOgKTh = The mixed-ligand complex is considerably more stable than the corresponding binary complexes.The reaction of Th(lV) with AZM and TSA was investigated spectrophotometrically and a sensitive method was developed for the microdetermination of thorium based on the formation of the Th(lV) - AZM - TSA ternary complex. The solution spectrum of the complex is characterised by an absorption band with I,,,. at 585 nm within the pH range 4.4-5.5. The ternary system obey's Beer's law between 1.16 and 18.57 1.19 ml-I of thorium with a molar absorptivity of 9.2 x lo4 I mol-1 cm-1. Keywords Thorium(lV) mixed-ligand complexes; stability constants; alizarin maroon; thiosalicylic acid; spectroph o tom etr y logK%(i%] (TSA) - IOgKT:(AZM). In continuation of our studies on the complexation equilibria of mono- and biligand systems in solution,1-4 we have studied the complex formation between Th(IV) - alizarin maroon (AZM) 1 1 complex and thiosalicylic acid (TSA).This study was undertaken to explore the potential of thiosalicylic acid in mixed-ligand complexes also containing an acid anthraqui-none dye. There is no example of this behaviour in the literature. The working conditions were established in order to obtain fundamental information about the stability of the complexed biligand system and the possible equilibria that exist in solution. Recent work in this laboratory"6 showed that the ligand AZM (3-amino-l,2-dihydroxyanthraquinone) possesses a unique combination of properties that enable it to be utilised in various analytical applications.The basic characteristics of the mixed-ligand complex of thorium(1V) with AZM and thiosalicylic acid in a 1 1 1 molar ratio were investigated potentiometrically . The Irving and Rossotti pH-titration technique7 and its related modifica-tions8.9 were employed. Fundamental conditions for the spectrophotometric determination of thorium(1V) using the Th(IV) - AZM - TSA ternary complex have also been established. Experimental Chemicals All chemicals were of analytical-reagent grade unless stated otherwise and doubly distilled water was used for the preparation of solutions. The solutions were diluted as necessary to prepare standard working solutions. Alizarin maroon (AZM) stock solution 5 X 10-3 M in ethanol.Obtained from Merck. Thoriurn(lV) standard solution 5 x 10-3 M. Prepared using AnalaR-grade thorium(1V) nitrate. Thiosalicylic acid stock solution 10-3 M. Prepared directly from the pure reagent. Sodium hydroxide standard solution 0.1 M . EDTA stock solution 10-1 M. Obtained from BDH Chemicals. Other solutions. Buffer solutions sodium perchlorate and perchloric acid standard solutions and solutions of the diverse ions used for interference studies were prepared as described previously.6 Apparatus The pH titrations were carried out using an Orion pH meter (M601A) with combined glass - calomel electrode. The measured pH values of the solutions were corrected as described by Douheret.10 The absorption spectra of solutions were recorded on a Pye Unicam SP 8000 spectrophotometer in the range 350-650 nm using 1-cm matched silica cells.All the potentiometric and spectrophotometric measure-ments were peformed in 40% V/V ethanol - water medium at 25 f 0.1 "C. The ionic strength of all solutions measured was adjusted to I = 0.1 M (sodium perchlorate). Dissociation Constants of the Ligands and Stability Constants of the Binary and Ternary Complexes The acid - base properties of AZM in aqueous media and in water - organic solvent mixtures have been discussed by Idriss and co-workers.11312 The predominant form of this reagent in acidic medium is the triprotonated form (H3A)+ which undergoes stepwise ionisation on increasing the pH of solutions. The proton-association constant P K ~ ~ and the dissociation constant P K & ~ of AZM in 40% V/V ethanol were determined by potentiometric titration of 50 ml of 2.5 X 10-3 M HC104 and NaC104 ( I = 0.1 M) in the presence and absence of the ligand (0.8 x 10-4 M) with standard carbhate-free NaOH solution (1.05 x lo-* M).The dissociation constant of TSA corresponding to the ionisation of the carboxyl hydrogen was also determined potentiometrically in 40% V/V ethanol; pK was found to be 4.30. The dissociation constant for the sulphydryl hydrogen of TSA as reported by Kumar et al. I 3 was 9.52. The dissociation constants of AZM and TSA as obtained from the titration graphs are given in Table 1. The stability constants for the binary Th(1V) - AZM complexes were calculated from titration curves in which the metal to ligand ratio was 1 2.The corresponding data for the secondary ligand (TSA) was determined under identical conditions. The equilibrium constants for the ternary systems were calculated from titration curves obtained for a 1 1 1 molar ratio of Th - AZM - TSA systems. The concentration of Th(IV) was 8 x 10-5 M . Multi-titrations were carried out with the Th(IV) - AZM - TSA ternary system. Details regardin 678 ANALYST JUNE 1986 VOL. 111 Table 1. Negative logarithms of acidity constants of the ligands and logarithms of the stability constants of their binary Th(1V) complexes (I = 0.1 M; 25 "C 40% V/V ethanol) Ligand PKE,L PKZL LWKfPlL LogKfk LogmL, AZM* . . . 9.10 11.2 (ref. 11) 6.30 4.75 11.05 TSA . . . . 4.30 9.15 4.45 3.90 8.35 * pKEqL+ = 6.75 for AZM.the potentiometric titration method have been reported earlier6-6.12. For the following equilibria in binary systems containing AZM: ECThAZM) Th + AZM Th(AZM) . . . . (1) GI(AZM), Th + 2AZM Th(AZM)2 * . (2) the formation constants were calculated taking into account the species H H3A H2A HA Th Th(HA) and Th(HA)2, where H2A is the neutral form of AZM. For ternary systems the formation constant for the equilib-rium B:f:k(AZM) (TSA) Th + AZM + TSA - Th(AZM) (TSA) . . (3) was calculated by considering the species H H3A H2A HA, H2L HL L Th ThL ThL2 Th(HA) Th(HA)2 and Th(HA)(L) where H2L represents the netural form of the secondary ligand TSA. The stepwise formation constants for the equilibria GktEEj, Th(AZM)+ AZM Th(AZM)2 . . (4) K;!{t$% (TSA) Th(AZM) + TSA .Th(AZM)(TSA) . . (5) were calculated considering the relevant data or the acid dissociation constants and the cumulative binary and ternary constants. Results and Discussion The potentiometric titration curve for AZM in the cationic form (H3A)+ (Fig. 1 I) shows a moderate inflection at a = 1 followed by a steep inflection at a = 2 (a = moles of base added per mole of ligand). The constants PKfl3A.t and pK&A corresponding to the stepwise dissociation of the ligand, AZM are given in Table 1. The titration curve for a system containing a 1 2 molar ratio of Th(IV) to monocationic AZM exhibits two inflections at rn = 2 and rn = 3 (rn = moles of alkali added per mole of metal ion) indicating the formation of mono- and bis-binary complexes.The corresponding equilibria may be represented by Th4+ + H3A+ [ThHA]3+ + 2H+ . . (6) [ThHAI3+ + H2A [Th(HA)2]2+ + H+ . . (7) For the secondary ligand TSA (H2L) and its binary complex with Th(IV) the constants corresponding to the following equilibria were also determined under our experimental conditions: G 2 L H;?L G HL-+H+ . . . . . . (8) G L Th4+ + HZL [ThL]'+ + 2H+ . . (9) KEL* [ThL]2+ + H2L s Th(L)2 + 2H+ . . (10) 0 1 2 3 4 m Fig. 1. Potentiometric titration curves of binary and ternary complex systems of thorium(1V) [I = 0.1 M (NaCIO,); 40% V/Vethanoi]. For I and 11 the abscissa represents the moles of alkali added per mole of ligand ( a ) ; I AZM; 11 TSA. 1 1 1 Th(1V) - AZM; 2 1 1 Th(1V) -TSA; 3 1 2 Th(1V) - TSA; and 4 1 1 1 Th(IV) - AZM - TSA Ternary Systems The potentiometric titration curves for ternary systems containing Th(IV) AZM and TSA in a 1 1 1 molar ratio exhibit a single steep inflection at rn = 4 (cf.Fig. 1). The composite curve drawn by adding the horizontal distance of the Th - AZM titration curve to the TSA curve is not super-imposable with the mixed ligand titration curve thereby confirming the formation of the Th - AZM - TSA complex. The stability constants of a ternary system are defined by equations (11)-(13). The over-all stability constant, BfazM) (L) which was determined experimentally is connec-ted with K#(AzM) (L) and KE[tgB] (L) by equations (14) and (15) respectively. M + AZM + TSA M(AZM) (TSA) BD(AzM) (TSA) = [M(AZM) (TSA)]/([M] [AzMI [TSA]) (11) M(TSA) + AZM M(TSA) (AZM) KE(-E?# (AZM) = [M(TSA) (AZM)I/[M(TSA)I [AZM] (12) M(AZM) + TSA S M(AZM) (TSA) KE[E%](TsA) = [M(AZM) (TSA)I/[M(AZM)I [TsA] log KM(TsA] (AZM) = log B&TSA) (AZM) - 1% M(TSA) log KM[AZM](TSA) = log BD(AZM) (TSA) - log KE(AZM) (13) (14) (15) One way to quantify the stability of ternary complexes is according to equation (16),'4 i.e.by comparing the differ-ences in stability e.g. for the reaction between M(TSA) and (AZM): M TSA M AZM Alog KM = log KE@%] (AZM) - log KB(AZM) M AZM = log &t[AZMj (TSA) - log K E v s A ) . . (16 ANALYST JUNE 1986 VOL. 111 679 The value of Alog KM is the logarithm of the equilibrium constant due to equation (17): M(TSA) + M(AZM) e M(AZM) (TSA) + M . . (17) In general a negative value for Alog KM [equation (16)] is expected as usually KEL > KfiL2.This is in accordance with the statistical values obtained for the coordination of two different bidentate ligands to a regular and to a distorted octahedral coordination sphere Alog Koh = -0.4 and Alog Kdo == -0.9 respectively.15 The other approach commonly used to quantify the stability of a ternary complex is based on the equilibrium constant X , as defined by equation (18)16317; log X may be calculated according to equation (19). M(AZM)2 + M(TSA)2 2M(AZM)/(TSA) X = [M(AZM) (TSA)]2/[M(AZM),] [M(TSA),] . . (18) log x = 2 log B&TSA) (AZM) - [log B&AZM)2 -k log BE(TSA)2] = log [KE[&#] (TSA) - log KE(%i]2] = log [K$@%] (AZM) - log K3[kg&1 * (19) The statistical value for log X i s the same for all geometries of the coordination sphere of a metal ion and is 0.6.'7.l8 Comparing the graph resulting from the titration of TSA and Th(1V) in a ratio 1 1 with that where in addition AZM was present (molar ratio 1 1 l) it is obvious that the deprotonation of TSA in the ternary system in the presence of AZM occurs at a lower pH.This means that the Th(IV) -AZM - TSA ternary complex is more stable than the corresponding binary complex and a positive AlogK value is obtained. The acidity constants of the ligands and the stability constants of the binary Th(IV) complexes that were used for the calculation of the stability constant of the ternary complex are given in Table 1. The results obtained for the formation of a ternary complex between the 1 1 Th(IV) - AZM complex and TSA are shown in Table 2.The constant calculated from the experimental data was log @&HA) ( L ) which is due to the overall equilibrium Th4+ + H3Af + HZL [Th(HA) (L)]+ + 4 H+ By using the results in Tables 1 and 2 the values for AlogK [equation (16)] and log X [equation (19)] were calculated. The experimental data show that the formation of the ternary complex of Th4+ with AZM and TSA shifts the buffer region of the ligand to lower pH values which indicates that the ternary complex is more stable than the binary complex. Further the deprotonation of AZM in the ternary system occurs at a lower pH than that of the corresponding binary complex. This means that the ternary Th - AZM - TSA complex is more stable than the Th - AZM binary complex and hence a positive AlogK value was obtained.Spectrophotometric Studies In 40% V/V ethanol - water media in the pH range 3.5-11.4, the ligand AZM exists in four different forms having Table 2. Logarithms of the equilibrium constants of Th(1V) - AZM -TSA and some related data Parameter Value LogBFk(&7M)(TsA)* . . . . . . 11.50 LOgGk TSA (AZMlI. . . . . . . 7.05 Logh?&$& (TSA)$ . . . . . . 5.20 LogX§ . . . . . . . . . . 3.6 ALogKy . . . . . . . . . . 0.75 * Equation (11). t Equations (12) and (14). $ Equations (13) and (15). § Equations (18) and (19). 7 Equation (6). absorption maxima with A at 410 480 510 and 550 nm. The absorbance versus pH graphs at various wavelengths exhibit three inflections indicating the possible existence of three equilibria in the pH range under study.These are as follows: H3A+ HZA + H+ (PH d 7.0) H2A HA- + H+ (PH 3 9.0) HA- A2- + H+ (pH 2 11.3) Absorption Spectra and Optimum pH The visible spectra of AZM show an absorption band at 410-420 nm within the pH range 3.5-6.0. The spectrum of TSA (1 x 10-4 M) exhibits no measurable absorbance in the visible region and there are no significant changes in colour or absorbance in the presence of Th(IV) ions. However the solution containing AZM and TSA undergoes an observable change in colour from yellow to pink - violet when mixed with 2.5 ml of 10-4 M Th(1V). The spectrum of the reaction mixture against a blank solution containing the same concentration of the two ligands exhibits a new band at 585 nm.The latter band is unambiguously due to the formation of the mixed-ligand complex of Th(IV) with AZM and thiosalicylic acid with maximum colour development at pH 4.4-5.5 (cf. Fig. 2). Spectrophotometric Determination of Th(1V) Procedure A standard solution of thorium(1V) containing 60-560 yg of thorium was introduced into a 25-ml calibrated flask then 2.5 ml of 10-1 M EDTA solution and 2.5 ml of 10-3 M AZM solution were added. Buffer solution was added to adjust the pH to 4.6-5.5 2.5 ml of 10-3 M TSA were added and the solution was diluted to volume with doubly distilled water and the requisite amount of ethanol (40% VlV). After thoroughly mixing the reaction mixture the absorbance was measured at 580 nm against a blank similarly prepared but containing no thorium.To examine the effect of interfering ions solutions of such ions were pipetted first followed by Th(IV) EDTA, AZM buffer and TSA solutions. The above method was followed. Effect of masking agents and foreign ions The addition of EDTA solution as a masking agent at concentrations up to a 100-fold molar excess over Th(IV) had no effect on the sensitivity of the method. The effect of foreign ions at levels of 5-14 mg per 25 ml on the determination of 360 440 520 600 680 Fig. 2. Absorption spectra of Th(IV) - AZM - TSA ternary complex in 20% V/Vethanol [pH = 4.4-5.5; Th(1V) = 1 x M . A AZM, AZM - TSA (both give the same s ectrum); B 1 1 Th)IV) - AZM (versus reagent blank); C 1 1 1 TK(IV) - AZM - TSA versus buffer containing 20% ethanol; and D 1 1 1 Th(1V) - AZM - TSA (versus reagent blank) Wavelengthh 680 0.5 0.4 8 2 3 0 0.3 0.2 0.1 ANALYST JUNE 1986 VOL.111 -----0.6 1 I I I I 1 1 I I J 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 (A) 1 I I I 0 0.2 0.4 0.6 0.8 (B) (C) Molar fraction of metal ion (A B) or TSA (C) Fig. 3. Job plots A Th IV) - TSA (in the presence of excess of AZM); B Th(1V) - AZM [in the presence of excess of TSA); and C, TSA - AZM [in the presence of excess of Th(IV)] Th(1V) by the given procedure was examined. The effect of each ion was tested individually in the presence of 5 X 10-3 M EDTA. There was no interference from 14 mg (ca. 200-fold excess) of Li+ Na+ Ba2+ Pb2+ C1- I- NO3- and S042- 8mg (ca. 100-fold excess) of Mg2+ Ca2+ Mn2+ Fez+ Fe3+ Co2+, Zn2+ A13+ Br- C032- NO2- and S032- or 4 mg (ca.50-fold excess) of Cr3+ Ni2+ and S2- The presence of 1 mg of U(V1) caused a negative error corresponding to 7 pg of Th(1V). Of the anions investigated fluoride and dihydrogen phosphate caused serious negative errors. The presence of F-seems to prevent any reaction between Th(1V) and AZM. Experiments to determine the stoicheiometry of the reaction of fluoride with solutions containing AZM and TSA were inconclusive. Calibration graphs and reproducibility The system Th(1V) - AZM - TSA followed Beer’s law over the range 5 X 10-6-0.8 x 10-4 M of Th(1V). The molar absorptivity was 9.2 x l o 4 1 mol-1 cm-1. Ten identical samples each of final Th(IV) concentration 4.0 X M, were tested according to the recommended procedure and their absorbances were measured.The mean absorbance was 0.36 with a standard deviation of 0.005 absorbance unit. Stoicheiometry of the complex Job’s method of continuous variation19.20 was applied to establish the composition of the ternary Th(1V) - AZM - TSA complex. The molar fractions of two of the components were varied continuously keeping their combined concentration constant and keeping the third component in a large excess for all solutions in the series. Under these conditions the ternary system was modified to a quasi-binary system. The results shown graphically in Fig. 3 indicate that the over-all Th(1V) - AZM - TSA composition is 1 1 1 at the pH used. The stoicheiometry of the ternary system was also determined by applying the molar ratio method.21 1.2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. References Idriss K. A. Hassan M. K. Abu-Bakr M. S. and Sedaira, H. Analyst 1985 109 1389. Idriss K. A. Harfoush A. El-Shahawy A. S. and Sedaira, H. Analyst 1985 110 709. Idriss K. A. Seleim M. M. Abu-Bakr M. S. and Sedaira, H. Polyhedron 1985,4 1521. Idriss K. A. Seleim M. M. Hassan M. K. Abu-Bakr M. S., and Sedaira H. Analyst 1985 110 705. Idriss K. A. Seleim M. M. Abu-Bakr M. S. and Saleh, M. S. Ann. Chim. (Rome) 1984 74 845. Idriss K. A. Seleim M. M. Abu-Bakr M. S. and Saleh, M. S. Analyst 1982 107 12. Irving H. M. and Rossotti H. S. J. Chem. SOC. 1953,3397; 1954 2904. Chidambaram M. V. and Bhattacharya P. K. J. Inorg. Nucl. Chem. 1970 32 3271. Agrawal B. H. Dwivedi K. Chandra M. Agrawala B. and Dey A. K. J. Indian Chem. SOC. 1977 54 931. Douheret G. Bull. SOC. Chim. Fr. 1967 1412; 1968 3122. Idriss K. A. Seleim M. M. and Abu Bakr M. S. Bull. SOC. Chim. Fr. 1981 5-6 180. Idriss K. A. Seleim M. M. Abu-Bakr M. S. and Saleh, M. M. Indian J. Chem. 1982 21A 395. Kumar A. N. Nigam H. L. and Seth T. D. J. Polarogr. SOC. 1966,3 83. Fischer B. E. and Sigel H. Inorg. Chern. 1979 18 425. Sigel H. Angew. Chem. 1975 87 391; Angew. Chem. Znt. Ed. Engl. 1975 14 394. Martin R. B. and Prados R. J. Inorg. Nucl. Chem. 1974,36, 1665. Dewitt R. and Watters J. I. J . Am. Chem. SOC. 1954 76, 3810. Kida S. Bull. Chem. SOC. Jpn. 1958 29 805. Job P. Ann. Chim. (Rome) 1928 10 113. Shirif F. G. and Awad A. M. J. Inorg. Nu& Chem. 1962, 24,79. Yoe G. H. and Jones A. L. Ind. Eng. Chem. Anal. Ed., 1944 16 111. Paper A51368 Received October 14th 1985 Accepted December 19th 198

ISSN:0003-2654    

DOI:10.1039/AN9861100677

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, JUNE 1986, VOL. 111 681 Study of Ternary Thorium Complexes with Some Triphenylmethane Reagents and Cationic Surfactants Maciej Jarosz Department of Analytical Chemistry, Technical University, ul. Noakowskiego 3, 00-664 Warsaw, Poland Nine ternary thorium systems with Chrome Azurol S (CAS), Eriochrome Cyanine R (ECR) and Pyrocatechol Violet (PV) and the cationic surfactants cetyltrimethylammonium (CTA), cetylpyridinium (CP) and zephir- amine (zeph) ions were examined as basis for spectrophotometric method forthe determination of Th. Chrome Azurol S (with CP or CTA as the third component of the system) is recommended for the determination of thorium (the molar absorptivity for the method with CAS and CP is 1.40 x lo5 I mol-1 cm-l) and other transition metal ions hydrolysed in neutral or weakly acidic media.Ternary systems with ECR (and mainly with zeph) can form the basis of sensitive spectrophotometric methods for the determination of some light main group metal ions. Keywords : Thorium determination; ternary complexes; triphen ylmethane reagents; cationic surfactants; spectroph otometry Spectrophotometric methods based on metal ion (M) - triphenylmethane reagent (R) - cationic surfactant (CS) systems can exhibit very high sensitivities.132 Many papers concerning these systems with metal ions that are hydrolysed in neutral or weakly acidic media have been published recently. The variety of chelating triphenylmethane reagents and cationic surfactants has led to the development of many spectrophotometric methods. However, no systematic studies have been carried out on the sensitivity and specificity of ternary systems for the determination of particular metal ions, although some work has been carried out in an attempt to explain the mechanisms of such complex formation.This work has been particularly concerned with the influence of the surfactant on the properties of the complex.s7 The most sensitive spectrophotometric methods that have been reported1 for determining metal ions are based on ternary systems with Chrome Azurol S (CAS), Eriochrome Cyanine R (ECR) and Pyrocatechol Violet (PV) and the cationic surfactants cetyltrimethylammonium (CTA), cetyl- pyridinium (CP) and tetradecyldimethylbenzylammonium (zephiramine, zeph). 1 The ternary complexes of thorium have not yet been examined in detail as a basis for spectrophoto- metric methods.Three papers concerning Th - CAS - CTA879 and Th - PV - CTAlO have been published to date, but there is no information for complexes with ECR or other frequently used cationic surfactants such as CP or zeph. This paper describes the study of nine Th systems, each containing a triphenylmethane reagent (CAS, ECR or PV) and a cationic surfactant (CTA, CP or zeph). After consider- ing all the different parameters, the most sensitive of the nine systems for the spectrophotometric determination of Th has been evaluated. With reference to other work,11-14 some features of the reactivity of triphenylmethane reagents with respect to ternary complex formation are elucidated. Experimental Apparatus Absorbances were measured in 1-cm cells using a Unicam SP-800 spectrophotometer or a Specord UV - visible spectro- photometer.The pH measurements were carried out using an Acidimeter 325 pH meter. Reagents To prepare the thorium standard solution (1 mg ml-I), about 2.5 g of thorium (IV) nitrate tetrahydrate (analytical-reagent grade) were dissolved in water with 5 ml of concentrated nitric acid and the solution was diluted to 1 1. This solution was standardised gravimetrically as thorium(1V) oxide by evapor- ation of a portion in a platinum vessel. Working solutions were prepared as required by suitable dilution with 0.1 M nitric acid. Chrome Azurol S, Eriochrome Cyanine R and Pyro- catechol Violet (supplied by BDH Chemicals, Poole, UK, Loba-Chemie, Vienna, Austria and Polskie Odczynniki Chemiczne, Gliwice, Poland, respectively) were purified by re-precipitation from ca.1.5 M hydrochloric acid. Solutions of these reagents were then made up to concentrations of 5 X Cetyltrimethylammonium bromide, cetylpyridinium chloride and zephiramine (supplied by International En- zymes, Windsor, UK, and Plainview, NY, USA, Loba- Chemie and ICN Pharmaceuticals, Inc. , Plainview, NY, USA, respectively) were used as aqueous solutions (5 x 10-3 or M for CAS and ECR and 1 x 10-3 M for PV. 1 x M). General Procedure A solution containing an appropriate amount of thorium (to give Th concentrations in the final volume of 5 x 10-6 M, 1 X 10-5 M and 1.5 x 10-5 M for experiments with CAS, ECR and PV, respectively) was placed in a 25-ml beaker. The solutions of triphenylmethane reagent [in a 10-fold, or optimum (after its evaluation), molar excess over Th] and cationic surfactant (in a lOO-fold, or optimum, molar excess over Th) were added and the pH was adjusted to the required value with ammonia solution.The solution was transferred into a 25-ml calibrated flask and diluted to the mark with water. After the dppro- priate period of time, the absorption spectrum of the complex solution was recorded against a reagent blank. Results Ternary Systems of Thorium with Chrome Azurd S and a Cationic Surfactant The optimum pH ranges for ternary complex formation are as follows: 5.3 k 0.3 (with CP), 5.8 k 0.5 (CTA) and 6.6 2 0.1 (zeph) (Fig. 1). Below pH 4 the complexes absorb at shorter wavelengths than under the optimum conditions.Relatively high molar excesses of CAS with respect to Th [&12 (zeph); 10-15 (CP); 10-20 (CTA)] are required in order to obtain maximum absorbances. The absorption maxima do not shift with the change in excess of the reagent. When CS is added to each system, a point is reached where turbidity appears. Further addition of CS causes the turbidity to disappear and682 ANALYST, JUNE 1986, VOL. 111 I 1 I I I 7.0 0.5 ' 4.5 5.0 5.5 6.0 6.5 PH Fig. 1. Effect of pH on the following ternary systems: 1, Th - CAS - CP (630 nm); 2, Th - CAS - CTA (630 nm); and 3, Th - CAS - zeph (635 nm). CTh = 5 x 10-6 M; cCAS = 5 x 10-5 M; and ccs = 5 x 10-3 M the complex of analytical interest to be f0rmed.l' Stable absorbances are reached with the following molar excesses of CS with respect to Th: 50-80 (CTA), 75-150 (CP) and 100-150 (zeph).The formation of complexes is negligible before the turbidity range is reached. Increasing the CS concentration only causes bathochromic shifts of A,,,, in comparison with the binary systems. Reactions forming ternary complexes require about 30 min to go to completion. After this time the maximum absorbances are obtained. The molar ratios of CAS to Th for all three complexes were established by Job's method to be about 3.5. An evaluation of CS to Th molar ratios in the ternary complexes is impossible, because the analytically interesting range in which CS is present in excess falls after the turbidity range and micelles, rather than CS ions, participate in this reaction. Ternary Systems of Thorium with Eriochrome Cyanine R and a Cationic Surfactant Eriochrome Cyanine R reacts with Th and CS over a lower pH range than that required for CAS systems.The optimum pH ranges are 4.2 k 0.2 (with CP), 4.3 k 0.3 (CTA) and 4.4 k 0.2 (zeph). In contrast to CAS, ECR systems exhibit batho- chromic shifts of A,,,. above pH 5. The appropriate molar excesses of ECR with respect to Th for maximum ternary complex formation are 10-15 (with CTA or CP) and 10-20 (with zeph), the excesses of CS being 75-100 (CTA or CP) and 90-120 (zeph). Only in the Th - ECR - zeph system do increasing molar excesses of ECR produce a small hypsochromic shift in A,,,,. None of these complexes are formed below the turbidity range. Maximum and stable absorbances are exhibited immedi- ately when zeph is the third component; after 15 min with CP or 30 min with CTA.A precise evaluation of the molar ratio of ECR to Th in ECR complexes is difficult, but data obtained by Job's method indicate a molar ratio of approximately 4 for all three of the complexes studied. Ternary Systems of Thorium with Pyrocatechol Violet and a Cationic Surfactant The optimum pH ranges for the formation of the ternary complexes of thorium with PV and cationic surfactants are 7.9 f 0.1 (CP), 8.0 _t 0.1 (CTA) and 8.6 k 0.2 (zeph). In the Th - PV - zeph system the formation of two distinct pH-dependent complexes is observed (Fig. 2). One of them (Amax, = 580 nm) exists at a pH of about 4, and the other, of analytical interest, at pH 8.6 as described above. When CTA or CP is used, small hypsochromic shifts in A, occur above the optimum pH.The necessary molar excesses of PV, with respect to Th, for obtaining the maximum absorbances of the ternary complexes are 8-10 when zeph is used and 10-12 for CTA or CP. The absorption maxima of the complexes are shifted about 15 nm towards shorter wavelengths with greater molar excesses of PV. Cationic surfactants must be used in the following molar 450 550 650 750 850 Wavelengthhm Fig. 2. Effect of pH on absorption spectra in the Th - PV - zeph System. pH: 1,4.0; 2,5.0; 3,6.0; 4,8.0; 5,8.5; 6,9.0. CTh = 1.5 X lop5 M; Cpv = 1.5 x lop4 M; and CZeph = 1.5 x lop3 M. excesses with respect to Th: 40-60 (CTA), 75-100 (CP) and 75-175 (zeph). The ternary complexes are formed only at CS concentrations above the turbidity range, as with CAS or ECR systems.A small hypsochromic shift in ymax. appears in the absorption spectra when molar excesses of CS are greater than the optimum ones. The ternary complexes of thorium with PV and CS are not stable. For the complexes with CTA or CP the absorbances begin to decrease after 15 rnin and with zeph after 10 min. The molar ratio of PV to Th in the ternary complexes, established by Job's method in the optimum pH range, is about 3. In the Th - PV - zeph complex with a A,,,. of 580 nm at pH 3.8 the molar ratio is 2. Spectrophotometric Determination of Thorium Based on Ternary Systems with CAS, ECR and PV in the Presence of CTA, CP and Zeph The nine ternary systems described above can each form the basis of a spectrophotometric method for the determination of Th.Some characteristics of these systems are given in Table 1. The most sensitive systems are the ones containing CAS and, amongst these, the CAS - CP system appears to be the best for the determination of Th. The procedure for a method based on this system is as follows. To a solution containing 7-40 pg of thorium, placed in a 25-ml beaker, add 5 ml of 5 X 10-3 M CP solution and 5 ml of 5 x 10-4 M CAS solution. Adjust the pH to 5.3 k 0.3 with ammonia solution. Transfer the solution into a 25-ml calibrated flask, dilute to the mark with water and, after 30 min, measure the absorbance at 630 nm against a reagent blank. This method is not selective as numerous metals that readily hydrolyse in slightly acidic media [e.g., aluminium, gallium, indium, scandium, yttrium, and uranium(VI)] interfere when present in any amounts. Rare earth elements give positive errors only when present in about a 50-fold ratio to thorium.Hence, prior to the determination, thorium should be separated from other metals, e.g., by extraction or co- precipitation with carriers. The proposed method has been applied to the analysis of lanthanum chloride (pure, supplied by Reakhim, USSR). Thorium was separated from the lanthanum matrix by extraction with thenoyltrifluoroacetone in chloroform and re-extraction with dilute nitric acid, according to the procedure proposed by Kondrateva and Merisov.15 The content of thorium in the sample examined (sample mass 0.25 g) was (3.8 f 0.3) x lO-3% (for five determinations). A recovery of 93% was obtained on adding 10 pg of thorium.ANALYST, JUNE 1986, VOL.111 683 Table 1. Some data on spectrophotometric methods for Th based on ternary systems with CAS, ECR and PV and cationic surfactants* System CAS - CP CAS - CTA CAS - zeph ECR - zeph ECR - CP ECR - CTA PV - zeph PV - CP PV - CTA Lax./nm 630 630 635 585 590 590 675 660 660 Determination range/yg ml-l 0.3-1.6 0.3-1.6 0.3-1.6 0.2-3 .O 0.2-3.0 0.2-3.0 0.6-3.5 0.6-4.0 0.6-4.0 ~ / 1 mol-1 cm-1 14.0 13.8 13.0 7.8 6.6 6.4 7.0 6.2 6.0 x 104 * Binary systems of R - CS (blank solutions) absorb at 430,470 and 600 nm for CAS, ECR and PV, respectively. Discussion and Conclusions For the spectrophotometric determination of Th the most sensitive systems are those containing CAS and CS. CAS is a versatile reagent, and methods based on CAS systems are sensitive for many metal ions, especially transition metals.For the following metal ions higher sensitivities are attained with CAS compared with ECR: Th ( E = 1.40 X 105 for CAS - CP and 7.8 x 104 for ECR - zeph); Sc13 (E = 1.50 x 105 for CAS - zeph and 9.8 x l o 4 for ECR - zeph); and Fe(III)14 ( E = 1.32 x lo5 for CAS - CTA and 1.10 X 105 for ECR - zeph). The affinity of CAS for transition metal ions is due to the coordination of the M - CAS bonding. The most suitable third components among the CSs examined are CTA and CP, which interact with R mainly electrostatically. l2 Zephiramine, for which hydrophobic interactions prevail in the ion association compounds with reagents, gives poorer results (systems with lower molar absorptivities at Amax.).For small metal ions, especially those from the main groups of the periodic system, the sensitivities of methods based on ternary systems with ECR and CS are comparable or even higher than those of methods with CAS, e.g. , A13+ 11 ( E = 1.24 X 105 for ECR - zeph and 1.10 x 105 for CAS - CS); Ga3+16J7 (E = 1.20 x 105 for ECR - CTA and 1.02 x lo5 for CAS - CP). This indicates that M-ECR bonding is mainly of an ionic character. The best results (the highest E) are obtained when zeph is the third component of the system. The difference in reactivity of CAS and ECR is caused by the different structure of their molecules. In the ECR molecule the sulphonic group occurs in the isolated benzene ring in the ortho-position with respect to the central carbon atom of the molecule, close to the carboxylic groups in the neighbouring benzene ring.This makes possible hydrophobic interactions between CS molecules themselves (bonded to SO3- and COO- groups) and between CS molecules and ECR molecules. For CAS, steric hindrances caused by two chlorine atoms occurring in the ortho-position with respect to the central carbon atom (the sulphonic acid group exists in the metu-position in this benzene ring) do not allow these interactions. The main difference between the reagents discussed above (CAS, ECR) and PV is that there are no carboxylic groups in this molecule. Phenolic groups have less affinity to quaternary ammonium cations and interactions between these cations and PV molecules are weaker than similar interactions involving CAS or ECR.At low pH, PV reacts in the R- form and the phenolic protons are not activated by CS. Therefore, the reaction with thorium ions is not completed and a ternary complex with a molar ratio of Th to PV of 1 : 2 is formed. Only at a pH of about 8, when PV exists in the R2- form, is it possible to obtain a complex which has a Th to PV ratio of 1 : 3. In this complex the interactions between CS and PV2- make the protons in the functional groups more mobile. However, even in the optimum pH range the spectrophotometric methods based on ternary systems with PV and CS are not as sensitive as the ones based on the systems with CAS or ECR. Two factors determine the sensitivity of these systems-the composition of the ternary complex formed and its stability.The molar ratio of R to M increases when the reaction takes place under pH conditions in which the metal occurs as the hydroxide and the protons of the functional groups of R are activated by interactions with CS. Larger complexes are formed with heavy, multi-valent metals such as Th. Stable complexes can be obtained when an appropriate chromogenic reagent and cationic surfactant are used (e.g., CAS and CTA or CP for transition metals, ECR and zeph for light main group metals). References 1. Tikhonov, V. N., Zh. Anal. Khim., 1977, 32, 1435. 2. Marczenko, Z . , Crit. Rev. Anal. Chem., 1981, 11, 195. 3. Chernova, R. K., Zh. Anal. Khim., 1977, 32, 1477. 4. Martynov, A. P., Novak, V. P., and Reznik, B. E., Ukr. Khim. Zh., 1978, 44, 203. 5. Savvin, S. B., Chernova, R. K., Belousova, V. V., Sukhova, L. K., and Shtykov, S. N., Zh. Anal. Khim., 1978, 33, 1473. 6. Savvin, S. B., Marov, I. N., Chernova, R. K., Shtykov, S. N., and Sokolov, A. B . , Zh. Anal. Khim., 1981,36, 850. 7. Chernova, R. K., Shtykov, S. N., Belolipceva, G. M., Sukhova, L. K., Amelin, V. G., and Kulanina, E. G . , Zh. Anal. Khim., 1984, 39, 1019. Shijo, Y., and Takeuchi, T., Bunseki Kagaku, 1969, 18, 469. Evtimova, B., Anal. Chim. Acta, 1974,68, 222. Tikhonov, V. N., and Pavlova, 0. K., Zh. Anal. Khim., 1982, 37, 1815. Marczenko, Z . , and Jarosz, M., Analyst, 1982, 107, 1431. Jarosz, M., and Marczenko, Z., Analyst, 1984, 109, 35. Jarosz, M., and Marczenko, Z . , Anal. Chim. Acta, 1984, 159, 309. Jarosz, M., and Malht, M., Microchem. J . , in the press. Kondrateva, T. M., and Merisov, U. I., Zavod. Lab., 1977,43, 913. Marczenko, Z., and Kafowska, H., Mikrochim. Acta, 1979,II, 507. Ganago, L. I., and Istchenko, N. N., Zh. Anal. Khim., 1980, 35, 1718. Paper A51332 Received September 16th, 1985 Accepted December 16th, 1985 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

ISSN:0003-2654    

DOI:10.1039/AN9861100681

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, JUNE 1986, VOL. 111 685 Determination of Trace Amounts of Phosphate in Water Samples by Ion-exchange Resin Thin-layer Spectrophotometry Kiichi Matsuhisa" Asahikawa National College of Technology, Shunkodai 2-2, Asahikawa, 070, Japan and Kunio Ohzeki Department of Chemistry, Faculty of Science, Hokkaido University, Sapporo, 060, Japan A simple but highly sensitive method for determining trace amounts of phosphate in water has been developed. The method involves concentrating molybdenum blue in a thin layer prepared from a finely divided anion-exchange resin and measurement of the absorbance of the coloured layer. The dependence of the absorbance on the concentration of sulphuric acid and ammonium molybdate resulting from the procedure for producing molybdenum blue was reduced by adding gelatin to the sample solution.The calibration graph deviated slightly from linearity and the relative standard deviation was 1.95% ( n = 4 ) for 0.5 pg of phosphate. This method was applied to the determination of trace amounts of phosphate in several water samples. Keywords: Phosphate determination; water analysis; molybdenum blue; spectrophotometry; anion-exchange resin thin layer To determine phosphate at trace levels of 5-6 pg 1-1 in environmental samples, several sensitive methods have been reported, such as Fe(I1) - ferrozine complex indirect spectro- photometry,l long capillary cell absorption spectro- photometry2 and laser-induced thermal lens effect spectro- photometry.3 However, all these methods have the disadvan- tage of being complicated, particularly as regards the appa- ratus.We have been studying a spectrophotometric method in which finely divided ion-exchange resin suspensions of par- ticle diameter less than 30 pm were employed as reagents for the concentration and separation of coloured sample species and the results of studies on determining phosphate ion,475 sulphide ion6 and nitrite ion7 have been reported. This method does not require any special apparatus and its great advantage is that by using a common spectrophotometer determinations about 100 times more sensitive than the solution method employing 1-cm cells are possible. This paper reports the results of the use of an anion- exchange resin suspension alone for determining phosphate ion. The variation in the colour of molybdenum blue due to variations in the concentrations of sulphuric acid and ammo- nium molybdate could be reduced by adding gelatin to the sample solution.Also, in determining total phosphorus, an improvement in the decomposition of samples according to the EPA methods by omitting the neutralisation step was tried. Experimental Reagents De-ionised, distilled water at 25 "C was used in all the experiments. Standard phosphorus solution. Potassium dihydrogen phosphate was dried at 110 "C for 3 h then dissolved in water to give a 50 mg 1-1 solution of phosphorus, which was sub- sequently stored. In actual usage, the stored solution was diluted to a phosphorus concentration of 1 pg ml-1. Mixed reagent solution. Ammonium molybdate (0.75 g) and antimony potassium tartrate (0.0137 g) were dissolved in 20 ml of water and to the solution were added dilute sulphuric acid ~ ~~~~~~~~~~ ~ * To whom correspondence should be addressed.prepared from 7 ml of concentrated sulphuric acid plus water, to give a total volume of 50 ml. In determining total phosphorus, dilute sulphuric acid prepared from 3.5 ml of concentrated sulphuric acid was used, as sulphuric acid was employed in decomposition. This reagent and 0.6 M ascorbic acid solution were not used after they were more than 2 d old. Gelatin solution, 1.1 mg ml-1. A 0.055-g amount of board-formed gelatin (Merck) was boiled and subsequently added to cold water. The mixture was heated to dissolve the gelatin and diluted to 50 ml. Anion-exchange resin suspension ( A RS) . An anion- exchange resin suspension was prepared from Amberlyst A-27 (Rohm & Haas) as reported previously.9 The exchange capacity of the suspension, determined by conductimetric titration, was 8.9 pequiv.ml-1. The ion-exchange resin precipitated when the suspension was allowed to stand, so it was necessary to shake it just before use. Apparatus A Shimadzu UV-240 spectrophotometer, Toyo KG-25 and KG-47 filter holders, a Pyrex CTF screw-mouthed bottle (decomposition bottle) and a Hirayama HP-15 autoclave were used. Procedure A 50-ml sample containing 0.05-0.5 pg of phosphorus was placed in a 100-ml beaker and kept at 25 "C. Then 0.2 ml of gelatin solution, 4 ml of mixed reagent solution and 1 ml of ascorbic acid solution were added successively to the sample, which was subsequently stirred with a magnetic stirrer for 10 min, then 4 ml of ARS were added to the sample, which was stirred for a further 10 min.The mixture was filtered through a 0.8-pm Millipore filter (AA type) set on the KG-25 filter holder, leaving on the filter a circular thin layer of coloured resin 17 mm in diameter and about 0.3 mm thick. This layer was soaked for 10 min in 0.18 M sulphuric acid, which had the same concentration as it did when used in the colour- producing process. The layer was subsequently attached to the side of the cell holder of the spectrophotometer to measure its absorbance at 695 nm against a sheet of white paper that acted as a dimming plate. Another sample containing no phos-686 ANALYST, JUNE 1986, VOL. 111 phorus was also taken and a thin layer was prepared as a blank using the same procedure.Filtration and Decomposition of the Samples A 47 mm diameter HA-type Millipore filter of pore size 0.45 ym were used for filtration to determine orthophosphate. Before the filtration, the filter was washed by passing 100 mi of de-ionised, distilled water through, and the first 100 ml of the sample leaving the filter were discarded so as to prevent any phosphorus contained by the filter from mixing with new filtrate. In determining total phosphorus, the same amount of sulphuric acid that was contained by 10 ml of the decomposed sample was omitted from the mixed reagent, and neutralis- ation after the decomposition was also omitted, which was a partial improvement on the EPA method.8 For decomposition, a sample containing 0.1-0.9 pg of phosphorus was placed in a 50-ml decomposition bottle and water was added to give 15 ml of solution, to which 1.4 ml of 6 M sulphuric acid and 2 ml of a 4% mlVpotassium persulphate solution were added.The mixed solution was autoclaved for 30 rnin at a gauge pressure of 2 kg cm-2 (132 "C). After removal from the autoclave and cooling, 10 ml of the decomposed sample were transferred into a 100-ml beaker and water was added to it to give a volume of 50 ml. The subsequent procedure was identical with that for the deter- mination of orthophosphate. Results and Discussion Experiments were performed with the volume of the solution after the addition of the colour former being maintained at 55.2 mi. Effect of Shaking Time after Addition of the Coiour Former and the Anion-exchange Resin Suspension Studies were made with the shaking time after the addition of the colour former varying from 5 to 30 min but remaining at 10 rnin after the addition of the suspension.From the fact that the absorbance remained constant throughout this procedure, it was assumed that the generation of molybdenum blue had been completed in the first 5 rnin before the examination. Next, the shaking time was kept at 10 min after the addition of the colour former but was varied from 5 to 30 rnin after the addition of the suspension. In this instance, the absorbance increased at 0.005 A min-1 as the shaking time increased. This indicated that shaking after the addition of the suspension helped the molybdenum blue to concentrate on the resin.The shaking also caused the resin itself to coagulate as a bulky material that was easy to filter and separate. Effects of Concentrations of Sulphuric Acid and Ammonium Molybdate and of Amounts of Gelatin and Triton X-100 Added Fig. 1 shows the effects of the concentration of sulphuric acid and the amount of gelatin added. Fig. 2 shows the influence of the concentration of ammonium molybdate and the amounts of gelatin and Triton X-100 added. The effect of the addition of gelatin was substantial, and the dependence of the absorbance on the concentrations of sulphuric acid and ammonium molybdate was considerably reduced. In the generation of molybdenum blue, the effects of the concentrations of sulphuric acid and ammonium molyb- date were complementary.This resulted from abnormal colouringlo due to the reduction of molybdate at low concentrations of sulphuric acid and at high concentrations of ammonium molybdate , and from hindrancelo to molybdenum blue generation under the opposite conditions. However, both the abnormal colouring and the hindrance to molyb- denum blue generation were controlled by the formation of a 1.1 I I t 0.1 OS3 * 0.125 0.175 0.225 0.275 Sulphuric acid concentrationh Fig. 1. Effect of varying the concentration of sulphuric acid on the colouration of 0.25 g of phosphorus (A) in the presence of 0.22 mg of elatin and (B) in its absence. Ammonium molybdate concentration: ! A) 0.109 and (B) 0.072% mlV. Ascorbic acid concentration: (A) 10.9 and (B) 5.45 mM. Amount of antimony: (A) and (B) 0.4 mg I I I I , 0.03 0.06 0.09 0.12 0.15 0.18 Ammonium molybdate concentration,% m/V Fig.2. Effect of varying the concentration of ammonium molybdate on the colouration of 0.25 g of phosphorus in the presence of (A) 0.05 mg of gelatin, (B) 0.22 mg of gelatin, (C) 0.44 mg of Triton X-100 and (D) in the absence of these surfactants. Sulphuric acid concentration: 0.18 M in each instance. Ascorbic acid concentration: (A-C) 10.9 and (D) 5.45 mM. Amount of antimony: 0.4 mg in each instance complex of gelatin with an excess of molybdate ion and by gelatin acting as a protective colloid on molybdenum blue micelles.11 When the amount of gelatin was small and the concentration of ammonium molybdate was over 0.109°/~ mlV, the absorbance tended to increase, as shown in Fig.2 (A). When 0.44 mg of gelatin was added to the sample solution, however, the absorbance showed the same ten- dency, as shown in Fig. 2(B). However, adding the same amount of Triton X-100 was not as effective, and when more Triton X-100 was used the resin thin layer became in- homogeneous. Effect of Ascorbic Acid Concentration, Amount of Antimony and the Colour Stabilisation of the Resin Thin Layer The concentration of ascorbic acid had nearly the same effect as when no gelatin was added. As the ascorbic acid used became more concentrated, the absorbance showed a gradual increase. As regards the effect of the amount of antimony, the range in which the absorbance became constant was 0.4-0.8 mg, whereas it was 0.05-0.8 mg when no antimony was added. The colour of the resin thin layer remained stable for at least 30 rnin after the filtration on soaking it in 0.18 M sulphuric acid, which is the same concentration as used in the colour- producing process.This effect was the same as that when no gelatin was added.ANALYST, JUNE 1986, VOL. 111 687 Effect of Coexisting Substances Studies were made of the influence of arsenate(V) and silicate, which are believed to cause a considerable hindrance to colouration, and also of surfactants in river water into which city sewage flows (Table 1). Arsenate(V) and silicate gave a larger error when gelatin was present in the sample solution than when no gelatin was present. On the other hand, the influence of the surfactants tended to be reduced when gelatin was added. Effect of Decomposition Time When the autoclave was kept at a gauge pressure of 2 kg cm-2, all the adenosine 5'-monophosphate, disodium salt (5'-AMP, 2Na) was completely decomposed in 30 min.As a practical sample, water was taken from the Ishikari River at the Kamikawa-ohashi Bridge, Kamikawa. With this river water, 20 min were required before a constant concentration of 67.5 pg 1-1 was reached. However, triphenylphosphine was not decomposed even after 50 min. Tests were also made on a standard phosphorus sample by using hydrogen peroxide in place of potassium persulphate. When 2.5 yl of 30% mlV hydrogen peroxide was added to the sample, no colour due to 1 .o 0.8 0, 0.6 + a 2 0.4 0.2 0 I I 0.2 0.4 c P hos p ho ruslyg Fig. 3. Calibration graph. The abscissa shows the amount of phosphorus as phosphate in 50 ml of sample solution.Molybdenum blue, produced from the final 59.2 ml of the solution, was concen- trated on a resin thin layer 17 mm in diameter and about 0.3 mm thick Table 1. Effects of coexisting substances on the determination of 0.25 pg of phosphorus as phosphate Added amount/ Relative error, 70 * Substance Pg with 0.22 mg gelatin Arsenate . . . . . . 0.125 14.2 Silicate . . . . . . . . 500 12.3 1250 25.0 Anionic surfactantt . . 100 0.0 Non-ionic surfactant$ . . 10 0.0 ml containing 0.22 mg of gelatin. * Mean of three determinations; absorbance at 695 nm. Sample, 50 t Sodium lauryl sulphate. $ Heptaoxyethylene dodecyl ether. molybdenum blue was observed even after the sample had been autoclaved for 50 min; in addition, there was no effect when the amount of ascorbic acid used was doubled.The reason for this is that coexisting phosphate functioned as a negative catalyst for the decomposition of hydrogen peroxide, and the remaining hydrogen peroxide affects the colouration of molybdenum blue. Calibration Graph The calibration graph had a slightly convex shape (Fig. 3). The mean values and standard deviations (n = 4, 95% confidence limits) were 0.511 k 0.010 for the blank, 0.126 k 0.007 for 0.05 pg of phosphorus and 0.985 k 0.019 for 0.5 yg of phosphorus. Application to Water Samples Change in concentration of orthophosphate in stored samples River water from Kamikawa-ohashi Bridge, Kamikawa, was filtered through a 0.45-ym, 47-mm diameter Millipore filter (HA type and then transferred into two 250-ml poly(viny1 fluoride) bottles, which were subsequently stoppered.One of the bottles was left in the laboratory and the other was kept in a Table 3. Determination of phosphorus in 0.2-40-ml samples of tap water. Samples were taken in the laboratory of Asahikawa National College of Technology Date Phosphorus found/pg* 1984) 0.2ml 0.4ml 0.6ml 20mlt 40mlt 9 0.060 0.116 0.178 6.0 12.0 12 0.065 0.130 0.200 6.5 13.0 16 0.078 0.148 0.210 7.0 14.0 19 0.065 0.130 0.195 6.8 13.6 23 0.190 26 0.190 (March 0.160$ 0.165$ 0.1803: * Mean of two determinations. Phosphorus concentration (95% confidence limits): 329.4 k 25.3 pg 1 - 1 ( n = 34, this method); 328.7 k 11.5 yg 1-1 (n = 16, solution method). 1- Using the solution method (JIS K 0102-1981).12 3: Recovery of 0.1 pg of phosphorus as phosphate: 100.8%.Table 4. Determination of phosphorus in lO-40-ml samples of seawater. Sample taken at Rumoi facing the Sea of Japan Date Phosphorus found/pg* (April 1984) 10 ml 20 ml 40 ml o*lll 2 0.029 0.060 5 0.042 0.082 0.157 * Mean of two determinations. Phosphorus concentration (95% t Recovery of 0.2 pg of phosphorus as phosphate: 98.3%. 0.260t 0.32 1- 0.2781- 0.352-t confidence limits, n = 20): 3.47 k 0.29 pg 1 - I . Table 2. Determination of phosphorus in 50-ml portions of eluate from membrane filter Phosphorus found/pgt Millipore filter* Pore size/pm First 50 ml Second 50 ml Third 50 ml Fourth 50 ml HA . . . . . . . . 0.45 0.143 0.040 0.021 0.020 GS . . . . . . . . 0.22 0.471 0.099 0.054 0.046 * 47 mm diameter; effective filtering area when used with KG-47 filter holder, 9.6 cm2.T Mean of four determinations.688 ANALYST, JUNE 1986, VOL. 111 Table 5. Determination of phosphorus in river water. Samples taken from Ishikari River . . Sampling point (place name, distance from source) Takane Bridge, 4 km Phosphorus concentration/ pg dm-3* Date Residue (September 1984) Orthophosphate Total phosphorus on filter? 15 14.4 14.5 - 22 14.8 15.0 - 29 13.2 14.0 2.63 Av. 14.1 f 0.63: Av. 14.5 k 0.43: Sounkyo, Kamikawa Town,30km . . Kanetomi, Aibetsu Town, 70 km 15 6.56 31.6 - 22 9.20 28.0 29 5.44 27.6 4.07 - Av. 7.06 k 1.50$ Av. 29.0 k 1.6$ - 15 7.20 16.4 22 7.36 16.0 - 29 6.08 16.4 3.50 Av. 6.88 f O X $ Av. 16.2 5 O . l $ Kyokusei Bridge, Asahikawa City, lOOkm . . . . . * Mean of two determinations.15 6.00 36.4 - 22 6.88 37.0 - 29 7.04 37.6 6.50 Av. 6.64 2 0.413: Av. 37.0 k 0.2$ t Dry mass of the solid matter in 350 ml of sample solution. 3: 95% confidence limits. refrigerator at 5 "C. The concentrations of orthophosphate in the two bottles showed the same tendency to change: the concentration continued to increase slightly for 7 d after the storage began, and then remained unchanged for 19 d. The increase in orthophosphate concentration during the first 7 d was about 6% in the bottle left in the laboratory and about 4% in the bottle kept in the refrigerator. Determination of phosphorus eluted into de-ionised, distilled water from the membrane filter When a 47 mm diameter Millipore filter was set on a Toyo KG-47 filter holder, the filter had an effective filtering area of 9.6 cm2.Through this filter was passed 200 ml of de-ionised, distilled water in 50-ml portions, and the phosphorus eluted into the water from the filter was determined (Table 2). It can be seen that 0.224 and 0.670 pg of phosphorus were eluted into the 200 ml of eluate from the HA-type filter of pore diameter 0.45 pm and from the GS-type filter ofpore diameter 0.22 pm, respectively. This indicated that those filters should be washed with de-ionised, distilled water before use. Determination of phosphorus in tap water The phosphorus present in the tap water in the laboratory of the Asahikawa National College of Technology was deter- mined (Table 3). Tap water is underground and, as the amount of phosphorus in the tap water was closely related to its volume, and as the average recovery of spiked phosphorus was 100.8% (n = 6), it was possible to determine the amount of phosphorus in the sample by using a calibration graph.The amount of phosphorus determined was in good agreement with that obtained by the solution method.12 Determination of phosphorus in seawater The phosphorus in surface seawater from Rumoi, facing the Sea of Japan, was determined (Table 4). The determination was performed by the calibration graph method, because when coexisting substances were examined it was found that there was no change in the calibration graph even when 3.05% mlm of sodium chloride were present. The validity df employing the calibration graph method was supported by the facts that the recovery of phosphorus from the seawater sample was 98.6% (n = 8) and that the amount of seawater tested was closely related to the amount of phosphorus detected in it.Determination of phosphorus in river water Water was taken at four points from the upper reaches of the Ishikari River: under the Takane Bridge near the source; at Sounkyo in Kamikawa, about 30 km below the source; at Kanetomi in Aibetsu, about 70 km below the source; and under the Kyokusei Bridge in Asahikawa, about 100 km below the source. The orthophosphate and the total phosphorus after the autoclave decomposition of the sample with potas- sium persulphate were determined (Table 5). The calibration graph method was applied to the determination, because the amount of the phosphorus detected was the same even though the amounts of river water tested were different, and also because the recovery was nearly 100%.The results showed that there were no large variations in concentration in either orthophosphate or total phosphorus throughout the period when the measurements were performed. It was also found that there was no particular correlation between the concen- tration of orthophosphate and that of total phosphorus. In the sample taken near the river source, most of the phosphorus existed in the form of orthophosphate, the concentration of which was nearly double that in the other three samples. Correlation was recognised, however, between the dry mass of the solid matter, or the residue on the filter, in the 350-ml sample taken on September 29th and the concentration of total phosphorus. 1. 2. 3. References Bet-Pera, F., Srivastava, A. K., and Jaselkis, B., Anal. Chem., 1981, 53, 861. Fujiwara, K., Wei, L., Uchiki, H., Shimokoshi, F., Fuwa, K., and Kobayashi, T., Anal. Chem., 1982, 54, 2026. Wei Lei, Fujiwara, K., and Fuwa, K., Anal. Chem., 1983, 55, 951.ANALYST, JUNE 1986, VOL. 111 689 Abe, M., Ohzeki, K., and Kambara, T., Bull. Chern.Soc. Jpn., 1978,51, 1090. Going, J. E., and Eisenreich, S. J . , Anal. Chirn. Acta, 1974,70, 95. Harvey, H. W., J. Mar. Biol. Assoc. U. K., 1948, 27, 337. “Testing Methods for Industrial Waste Water,” JIS K 0102- 1981, Japanese Standards Association, Tokyo, 1981. 9. 10. 11. 12. 4. Matsuhisa, K., Ohzeki, K., and Kambara, T., Bull. Chern. SOC. Jpn., 1981, 54, 2675. 5. Matsuhisa, K., Ohzeki, K., and Kambara, T., Bull. Chem. SOC. Jpn., 1982, 55, 3335. 6. Matsuhisa, K., Ohzeki, K., and Kambara, T., Bull. Chem. SOC. Jpn., 1983, 56, 3847. 7 . , Matsuhisa, K., and Ohzeki, K., Nippon Kagaku Kaishi, 1983, 1593. 8. “Method for Chemical Analysis of Water and Wastes. Tech- nical Transfer,” US EPA-625-/6-74-003a, Environmental Protection Agency, Washington, DC, 1976. Paper A51377 Received October 22nd, 1985 Accepted January 13th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861100685

出版商:RSC    

年代:1986

数据来源: RSC

摘要:

ANALYST, JUNE 1986, VOL. 111 69 1 Use of a Silver - Gelatin Complex for the Microdetermination of Hydrogen Sulphide in the Atmosphere Tarasankar Pal, Ashes Ganguly and Durga S. Maity Department of Chemistry, Indian Institute of Technology, Kharagpur 72 1302, India Gelatin forms a weak complex with silver(1) in alkaline medium, which was found to be suitable for the determination of 0.6-35 p.p.b. of hydrogen sulphide by spectrophotometry and 0.3-30 p.p.b. of hydrogen sulphide using a column packed with silica gel impregnated with the complex. The method was applied successfully to the determination of hydrogen sulphide under different environmental conditions. The molar absorptivity of the solution is 2.06 x lo4 I mol-1 cm-1 for the spectrophotometric procedure, which has a relative standard deviation of 1.2%, a confidence limit (for 20 determinations) of 11.976 k 0.023% and a Sandell sensitivity of 1.65 x 10-3 pg cm-2.The relative standard deviation and confidence limit for the column method (20 determinations) were 0.12% and 4.99 k 0.003, respectively, even in the presence of sulphur dioxide and carbon monoxide. Keywords: Hydrogen sulphide determination; spectrophotometry; silica gel column; silver - gelatin complex Most recent methods for the determination of trace amounts of hydrogen sulphide in air depend on gas chromatographic,l flame photometric,2 adsorption - desorption,3 electroana- lytical,4>5 solid sensor6 and laser techniques.7 They are sensitive but require expensive instrumentation. On the other hand, the old methylene blue method8 is widely used for the determination of soluble sulphides, hydrogen sulphide being trapped either in a cadmium salt or a 1% zinc acetate solution. Lodge and Pate9 reported that the results obtained by the methylene blue method were not precise.Daniel and Robert10 showed that zinc acetate is a better absorbent than cadmium salts for the determination of hydrogen sulphide. For the spectrophotometric determination of hydrogen sulphide (0.1-1.2 pg), Balasubramanian and Ramakrishnall extended the use of the zinc acetate solution method with the production of a benzene-extractable ion pair with Rhodamine 6G. These methods range from fairly sophisticated instrumental to less elaborate spectrophoto- metric techniques. Recently we reported a method for the determination of ascorbic acid12 using a silver - gelatin complex in alkaline medium, and observed that even trace amounts of hydrogen sulphide interfered.The method is based on the reduction of the silver - gelatin complex with ascorbic acid in alkaline solution to produce a silver sol, and in situ stabilisation of the silver sol helps in the determination of ascorbic acid. This observation has led to the development of spectrophotometric and column procedures for the determination of 0.6-35 and 0.3-30 p.p. b. of hydrogen sulphide, respectively, by direct reduction techniques. Experimental Apparatus A Varian Cary-17D spectrophotometer was used for measur- ing absorbance in 10-mm quartz cells. Impinger A simple impinger was used for gas absorption. Reagents All reagents used were of analytical-reagent grade without any further purification.Absorbing solution The reagent solution was prepared by dissolving 0.5 g of gelatin powder (E. Merck) in 100 ml of boiled, distilled water, with continuous stirring with a magnetic stirrer, and adding 10 ml of 10-2 M silver nitrate solution. Turbidity, if any, was removed by adding a few drops of 2 M sodium hydroxide solution. The solution was warmed to ca. 50 "C, the pH of the solution was adjusted to about 8 for practical purposes13 and the volume of the solution was made up to 200 ml. The solution was stored in an amber-coloured bottle in the dark. The solution remained stable for more than 10 weeks. The silver - gelatin binding is now a known interaction.13 Calibration Graph (Spectrophotometry) To calibrate, 10 ml of reagent solution were placed in each of three impingers connected in series by rubber tubing, and were exposed to a known concentration of hydrogen sulphide (0.6-35 p.p.b.) by bubbling a hydrogen sulphide - air mixture through the impinger with the absorbing solution at a flow-rate of 600 ml min-1.The hydrogen sulphide - air mixture was prepared by successive dilution of a pure hydrogen sulphide stream with streams of purified air, inert or nitrogen gas, and rotameters were used to measure the flow-rates (see Results and Discussion). It was observed from the absorbance measurements that 80% of the gas was absorbed in the first, 16% in the second and the 4% in the third impinger. On completion of the exposure the solutions were mixed and the absorbance was measured against a reagent blank at 415 nm after 30 min.The molar absorptivity was 2.06 X l o 4 1 mol-1 cm-1. The absorbance was plotted against concentration of hydrogen sulphide and a linear graph that passed through the origin was obtained; Beer's law was obeyed over the range 0-10 p.p.b. of hydrogen sulphide. The relative standard deviation, confidence limit (20 determina- tions) and Sandell sensitivity were 1.2'30, 11.976 k 0.023 and 1.65 X 10-3 pg cm-2, respectively. Silica Gel Tube Glass tubes (i.d. 8.5 mm) packed with 2 g of desiccator-dried silica gel (6S120 mesh) impregnated with silver - gelatin complex were used for the determination of hydrogen sulphide in the ambient atmosphere. A column 87 mm long was packed with 2 g of impregnated silica gel.The length of the tubes may be altered depending on the hydrogen sulphide concentration in the atmosphere.692 ANALYST, JUNE 1986, VOL. 111 Preparation of Silica Gel Column Silver - gelatin complex (35.0 ml) (silver concentration 0.80 X 10-2-0.80 X 10-3 M) was added to 26.0 g of silica gel in a beaker. The wet impregnated silica gel was dried in a vacuum desiccator and stored therein for several months. The concen- tration of silver nitrate used in the column method (0.00182 g per gram of silica gel) is optimal. At lower concentrations the exposed silica gel becomes indistinguishable from the unex- posed part, and at higher concentrations the column is too sensitive to ambient environmental conditions. Calibration Graph (Column Method) A glass tube (100 mm x 8.5 mm i.d.) was packed uniformly with 2 g of impregnated silica gel (60-120 mesh).In order to measure the length accurately, a sharp edge is needed, which was achieved by placing a sintered-glass disc on both the sides of the packing. The column was finally plugged with cotton on both sides. The effective length of the silica gel column was maintained at 87 mm throughout the course of the evaluation in order to obtain reproducible results. One side of the tube was connected with a chamber containing a known mixture of air and pure hydrogen sulphide and the other end was connected to a water-suction pump aspirating at a rate of 1000 ml min-1 through the column. The flow-rate was measured with a rotameter having a minimum count of 100 ml min-1.Reaction of hydrogen sulphide with the impregnated silica gel produced a steel-grey colour instantaneously. The change in colour gave a very sharp boundary between the coloured, exposed portion and the white, unexposed portion, which facilitated the accurate measurement of length. The measure- ment of length was the key point in the whole experiment, because it was finally utilised to evaluate the hydrogen sulphide concentration in air with the help of a calibration graph of exposed column length versus concentration of hydrogen sulphide, which was linear over the range 0-10 p.p.b. of hydrogen sulphide. The confidence limit for 20 determinations was 4.99 k 0.003 with a relative standard deviation of 0.12%. Results and Discussion Effect of Light Indoor lighting and direct sunlight did not affect the silver - gelatin complex in aqueous medium.For a silica gel column with a higher concentration of silver nitrate, however, both types of light reduced the column packed with impregnated silica gel. Therefore, it is desirable to cover the tube with Table 1. Effect of diverse gases on the determination of 12 p.p.b. of hydrogen sulphide by the spectrophotometric method Gas H2S.. . . NH3 . . ( 2 3 2 f CI, . . . . HCHO . . N2 . . . . H2 . . . . C6H5CH3. . C6HsOH . . so,. . . . co . . . . NO2 . . C6H6 . . C6H5NH2 CO, . . NO . . . . O3 . . . . CH3COCH3 Concentration, p.p.b. - . . . . 20 . . 30 . . 10 . . 30 . . 50 . . 50 . . 30 . . 30 . . 20 . . 10 . . 20 . . 20 . . 20 . . 30 . . 40 . . 40 . . 40 Absorbance 0.350 0.365 0.350 0.368 0.350 0.350 0.350 0.350 0.350 0.342 0.356 0.344 0.350 0.350 0.350 0.350 0.350 0.350 Difference, YO +4.3 0 +5.1 0 0 0 0 0 -2.3 +1.1 -1.7 0 0 0 0 0 0 - carbon paper to prevent the photoreduction of the silver - gelatin complex.Effect of Temperature13 Variations of up to 10 "C (below 50 "C) did not affect the determination. However, with silica gel columns the tempera- ture should be below 35 "C; temperatures above 40 "C cause deterioration of the column. Stability of the Colour13 The yellow colour of a silver sol appears in the aqueous phase. Its intensity remains stable for 1-2 d at room temperature (30 "C) . Effect of Diverse Gases Spectrophotometry Carbon monoxide and sulphur dioxide interfered in the determination of hydrogen sulphide, but hydrogen, nitrogen, chlorine and its oxide, ozone, acetyene and benzene did not.The technique used in the interference studies was to produce the appropriate concentration of gas under study in an exposure chamber. Two techniques were used to prepare standard atmospheres for either the spectrophotometric method or the column method. Using permeation tubes,14 hydrogen sulphide - air, sulphur dioxide - air and nitrogen dioxide - air were prepared and were suitably placed in the air stream in the dilution system. The carbon monoxide - air, hydrogen - nitrogen, nitrogen - argon, acetylene - air, chlorine - air, benzene - air and other air mixtures were prepared by using a motor-driven syringe to inject the mixtures into the hydrogen sulphide - air stream between two exposure cham- bers or just before an exposure chamber.The plunger speed was regulated to give the required concentration of the gas under study in the exposure chamber. The same regulated syringe technique was checked using formaldehyde solutions (30%) following the method of Lyles et al. 15 The results of the interferences studies are given in Table 1. The synthetic gas mixtures were analysed by a standard method. 16 Table 2. Effect of diverse gases on the determination of 5 p.p.b. of hydrogen sulphide by the column method Gas H2S . . NH3 . . C2H2 . . HCHO . . c12 . . . . N2 . . . . H2 . . . . C6HSCH3 C~HSOH so, . . co.. . . NO2 . . C6H6 . . C6H5NH2 co2 . . NO . . CH3COCH3 O3 . . . . Concentration, p.p.b. - . . . . 200 . . 500 . . 100 . . 500 . . 500 . . 500 .. 500 . . 500 . . 500 . . 200 . . 500 . . 500 . . 400 . . 500 . . 500 . . 500 . . 500 Length of coloured portion/ mm 18.0 18.0 18.0 18.5 18.0 18.0 18.0 18.0 18.0 18.0 19.0 18.0 18.0 18.0 18.0 18.0 18.0 18.0 Difference, YO 0 0 +2.6 0 0 0 0 0 0 +5.1 0 0 0 0 0 0 0 -ANALYST, JUNE 1986, VOL. 111 693 Table 3. Results of field evaluation at 30 “C Present methods* Spectro- photometric method, Date P.P.b. R.S.D.,% 10.7.85 . . . . . . 15 0.80 11.7.85 . . . . . . 13 0.78 12.7.85 . . . . . . 11 0.87 13.7.85 . . . . . . 9 0.96 14.7.85 . . . . . . 7 1.14 15.7.85 . . . . . . 12 0.87 16.7.85 . . . . . . 10 0.94 * Average values from six determinations. Column method, p.p.b. 15 14 11 8 8 13 10 R.S.D., Yo 0.10 0.08 0.11 0.12 0.10 0.18 0.12 Reference method,l6 p.p. b. 18 15 12 9 8 14 10 Column method Carbon monoxide, formaldehyde , acetylene , benzene, phe- nol, toluene, nitrogen, hydrogen, dinitrogen oxide, ozone and sulphur dioxide did not interfere in the determination.Various gas - air mixtures were prepared as in the previous procedure and the same injection technique was followed. The flow-rate was maintained at 1000 ml min-1 for the column method and the results are given in Table 2. Field Evaluation The evaluation was carried out in a shelter in a parking lot, surrounded by buildings on four sides. Air was aspirated at the flow-rates described under Calibration Graphs from a point adjacent to the shelter through a filter to remove suspended particles from the air, then passed through exposure chambers as described under Experimental and the concentrations of hydrogen sulphide were determined from the absorbance values and exposed column length by the spectrophotometric and column methods respectively. The results for both methods under the same environmental conditions were compared (Table 3).The results of the field studies show that the present methods gave very good reproducibility. Conclusion The methods reported here are simple, rapid, inexpensive and reliable for the determination of average concentrations of hydrogen sulphide in the ambient atmosphere. The tech- niques require no electrical or mechanical connections or complex equipment and no tedious sampling technique is necessary. The methods suffer from few interferences. The column technique may be applied to environmental analysis even in the presence of carbon monoxide.Packed tubes of various diameters can be used, depending on the concen- tration of hydrogen sulphide in the atmosphere. The reagent is stable at ordinary temperatures for up to 10 weeks. The absorbance value of the blank at 30 “C is 0.033 and this value varies by less than kO.001. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Ichinose, N., Nakamura, K., and Shimizu, C. J., J. Chromat- ogr., 1984, 292, 393. Mangani, F., Bruner, F., and Penna, N., Anal. Chem., 1983, 55, 2193. Berezhina, L. G., Eleftrova, N. A., and Sukhodolova, V. I., Zh. Anal. Khim., 1984, 39, 242. Opekar, F., and Bruckenstein, S . , Anal. Chem., 1984, 56, 1206. Zhao, Z., Wu, Z . , and Chen, Y., Fenxi Huaxue, 1984,12,179. Lalauze, E . , and Pijolat, C., Sens. Actuators, 1984, 5, 55. Pokrowsky, P., Appl. Opt., 1983, 22,2221. Matheson, N. A., Analyst, 1974, 99, 577. Lodge, J. P., and Pate, J. B., in Kolthoff, I. M., Elving, P. J., and Stross, F. H., Editors, “Treatise on Analytical Chemistry, Part 111, Volume 2, Analytical Chemistry in Industry,” Wiley-Interscience, New York, 1970, pp. 27-109. Daniel, F. L., and Robert, J. E., Environ. Sci. Technol., 1976, 2, 159. Balasubramanian, N., and Ramakrishna, T. V., Indian J. Chem., Sect. A , 1983,22, 550. Pal, T., and Maity, D. S . , Anal. Lett., 1985, 18B, 1131. Pal, T., and Maity, D. S., Analyst, 1986, 111, 49. Scaringelli, F. P., O’Keefe, E. A., Rosenberg, E., and Bell, J. P., Anal. Chem., 1970,42, 871. Lyles, G. R., Dowling, F. B., and Blanchard, V. J . , J. Air Pollut. Control Assoc., 1965, 15, 106. Ruch, E., “Quantitative Analysis of Gaseous Pollutants,” Ann Arbor Science Publishers, Ann Arbor, MI, 1970, p. 132. Paper A51358 Received October 8th, 1985 Accepted January 6th, 1986

ISSN:0003-2654    

DOI:10.1039/AN9861100691

出版商:RSC    

年代:1986

数据来源: RSC