ISSN:0003-2654    

DOI:10.1039/AN97500FX021

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BX023

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

SUMMARIES OF PAPERS I N THIS ISSUESummaries of Papers in this IssueProblems in the Determination of the Specific Radioactivity ofUniformly 14C - labelled Amino -acids after their Reaction withModifications to the standard Technicon amino-acid analyser are described,which permit a reliable determination of the specific radioactivity of [U-14C]-amino-acids to be made after their reaction with ninhydrin. The loss of thevolatile aldehyde products of the ninhydrin reaction can be prevented bytrapping them with semicarbazide. Glutamate was shown to lose twocarbon atoms when oxidised at pH 5-0 but only one at pH 2-5.R. BISHOP and A. P. SIMSSchool of Biological Sciences, University of East Anglia, University Plain, Nonvich,NR4 7TJ.Analyst, 1975, 100, 369-376.NinhydrinApplication of Differential Pulse Polarography to the Assay ofVitaminsThe determination of low concentrations of the vitamins B,, nicotinic acid,nicotinamide, ascorbic acid, K, and K, in pure solutions by use of differentialpulse polarography has been studied.The approximate limits of detectionachieved for the different vitamins varied from 0.01 to 1 p.p.m., which is10 to 100 times better than for d.c. polarography. A comparative deter-mination of vitamin B, and nicotinamide in three complex multivitaminpreparations showed that differential pulse polarography is more selectivethan d.c. polarography. With the former method the nicotinamide can bedetermined directly, i e . , without prior separation of interfering substances.J.LINDQUIST and S. M. FARROHADepartment of Analytical Chemistry, University of Uppsala, Box 531, S-751 21,Uppsala 1, Sweden.Analyst, 1975, 100, 377-386.Investigations into the Use of Gas-sensing Membrane Electrodesfor the Determination of Carbon Dioxide inPower Station WatersGas-sensing membrane electrodes have been assessed for use in the deter-mination of carbon dioxide in power station waters. The Radiometer,Type E5036, electrode has the most satisfactory performance, and the systemis convenient and simple for the investigation of abnormal levels of carbondioxide that arise a t start-up or during the commissioning stages; the limitof detection (about 0-1 pg ml-l) is not low enough for the routine monitoringof normal operation levels.The electrode has been used for both discreteanalysis and on-line continuous monitoring. Standard deviations of 2.0,0.3 and 0.13 pgml-l were obtained a t concentration levels of 23.2, 5-00 and2.32 pg ml-1, respectively. The response times a t concentrations greaterthan 1 pg ml-l were found to be within the range 3-7 min for a 10-fold increasein concentration. Of the species expected in power stations only sulphite ion,at high concentration, caused interference.D. MIDGLEYCentral Electricity Rescarch Laboratories, Kelvin Avenue, Leatherhead, Surrey,KT22 7SE.Analyst, 1975, 100, 386-399v1 SUMMARIES OF PAPERS I N THIS ISSUEModification of the Iodimetric Titration Method for theDetermination of Bromide and its Application to MixedDomestic - Industrial Waste EffluentsJune, 1975The iodimetric titration method for the determination of bromide involves theobservation of various colour changes, making the method unsuitable for usewith samples that are highly coloured.A modification is described, whichextends the usefulness of the method to highly coloured samples, such assewage and industrial waste effluents. The obvious substitution of a pHmeter for indicators was made and the use of standardised concentrations andamounts of reagents added was incorporated into the method. The pro-cedural steps were studied in order to optimise the sensitivity while mini-mising interfering side reactions. The modified method was applied to amixed domestic - industrial waste effluent spiked with bromide and iodide.The experimental details and statistical parameters for samples spiked atdifferent levels are presented.DANIEL F.BENDERMethods Development and Quality Assurance Research Laboratory, EnvironmentalProtection Agency, National Environmental Research Center, Cincinnati, Ohio46268, U.S.A.Analyst, 1975, 100, 400-404.A Semi-automated Procedure for the Determination of Iodine inPlant Tissue and Soil ExtractsA semi-automated method is described for the determination of total iodinein plant tissue and soil extracts. The automated, colorimetric determinationis based on the catalytic action of iodine on the oxidation of arsenic(II1) bycerium(1V). The reproducibility and accuracy of the proposed proceduresare reported for the materials studied.H.van VLIET, W. D. BASSON and R. G. BOHMERDepartment of Inorganic and Analytical Chemistry, University of Pretoria, Pretoria0002, South Africa.Analyst, 1975, 100, 405-407.Measurement of Low Phosphorus Concentrations in NutrientSolutions Containing SiliconA method has been developed, that permits the determination of phosphorusconcentrations down to 0-04 ~ L M in nutrient solutions that contain silicon.The problem of silicon interference is eliminated by use of ethyl acetate as aselective extractant for the yellow molybdophosphoric acid prior to itsreduction to heteropoly blue (molybdenum blue).S. JINTAKANON, G. L. KERVEN, D. G. EDWARDS and C. J. ASHERDepartment of Agriculture, University of Queensland, St. Lucia, Queensland 4067,Australia.Analyst, 1975, 100, 408-414.Determination of Nanogram Amounts of Bismuth by Means ofCandoluminescence Emission Following Application to theSurface of a Calcium Oxide Based MatrixThe blue emission stimulated by a hydrogen - nitrogen - air flame in a matrixthat is mainly calcium oxide and which contains trace amounts of bismuthions is used to determine 0-04-2-0 ng of bismuth in 04-1.0-pl samples, Upto a 50-fold excess of most ions can be tolerated; only Mn(II), Fe(II), Cr(III),V(1V) and Co(I1) interfere seriously.Applications to the determination ofbismuth in a copper alloy are reported.R. BELCHER, K. P. RANJITKAR and ALAN TOWNSHENDDepartment of Chemistry, Birmingham University, P.O. Box 363, BirminghamB15 2TT.Analyst, 1975, 100, 415421June, 1975 SUMMARIES OF PAPERS IN THIS ISSUEA Specific Method for the Determination of Trace Concentrationsof Tetramethyl- and Tetraethyllead Vapours in AirA method is described for the specific determination of tetraalkyllead vapourin air.The atmosphere under test is passed through a solution of iodinemonochloride in hydrochloric acid, which converts the tetraalkyllead com-pounds into their dialkyllead ionic forms.A procedure is given in which dialkyllead species are preferentially extractedas dithizonates from the buffered sample solution in the presence of ethylene-diaminetetraacetic acid, which complexes any inorganic lead compoundsthat may be present. The dialkyllead dithizonates are then decomposed byshaking with a nitric acid - hydrogen peroxide solution and the concentrationof lead in the resulting solution is determined by atomic-absorption spectro-photometry, using a carbon furnace for atomisation.The method has a limit of detection of 7 ng of lead in the sample.Whenexpressed as a concentration of tetraalkyllead in air the detection limitis 0.2 pg m-3 for a 10-min sampling period and 0.04 pg m-3 for a 1-h samplingperiod.S. HANCOCK and A. SLATERThe Associated Octel Company Limited, Ellesmere Port, Cheshire.Analyst, 1975, 100, 422-429.viiThe Determination of Lead and Cadmium in Paint byAtomic-absorption Spectrophotometry Utilising the DelvesMicro - sampling TechniqueThe Delves micro-sampling technique has been applied to the determinationof lead and cadmium in paints by atomic-absorption spectrophotometryusing both liquid and solid samples.The method involving liquid samplesconsists in suspending the paint sample in an appropriate solvent and trans-fering lop1 of the suspension into a nickel cup. After evaporating thesuspension to dryness and pre-igniting the residue in a mume furnace, the cupis inserted directly into an air - acetylene flame. The concentration of themetal is determined by the method of standard addition. The standarddeviation was found to be j 2 . 6 per cent. a t the 99.4p.p.m. of lead level at283.3 nm and &l-8 per cent. a t the 31.7 p.p.m. of cadmium level at 326.1 nm.The method involving solid samples consists in directly inserting a weighedamount of paint into an air - acetylene flame. Calibration was accomplishedby the method of standard additions. The standard deviation was found tobe k5.5 per cent. at the 230p.p.m. of lead level a t 261.4nm and &20 percent. at the 1-7 p.p.m. of cadmium level a t 326.1 nm.Results for lead and cadmium in several paint samples by the two methodsagreed well with those obtained by the conventional atomic-absorptionmethod. Both methods are applicable to oil- and water-based paints andthe time-consuming steps of ashing and dissolution of the ash required inconventional paint analysis are eliminated. The method with solid samplesis more rapid than that with liquid samples, but is less widely applicablebecause the concentration range of the metal that can be determined islimited.0. W. LAU and K. L. LIDepartment of Chemistry, The Chinese University of Hong Kong, Shatin, N.T.,Hong Kong.Analyst, 1975, 100, 430-437

ISSN:0003-2654    

DOI:10.1039/AN97500FP065

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BP071

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

JUNE, 1975 The Analyst Vol. 100, No. 11 91 Problems in the Determination of the Specific Radio- activity of Uniformly 14C-labelled Amino-acids after their Reaction with Ninhydrin R. Bishop and A. P. Sims School of Biological Sciences, University of East Anglia, University Plain, Norwich, NR4 7TJ Modifications to the standard Technicon amino-acid analyser are described, which permit a reliable determination of the specific radioactivity of [U-14C]- amino-acids to be made after their reaction with ninhydrin. The loss of the volatile aldehyde products of the ninhydrin reaction can be prevented by trapping them with semicarbazide. Glutamate was shown to lose two carbon atoms when oxidised a t pH 5.0 but only one a t pH 2.5. An investigation into the regulation of amino-acid biosynthesis during the cell cycle of the food yeast Candida utilis is in progress in these laboratories.This work requires a knowledge of the concentration and specific radioactivity of amino-acids so that absolute rates of bio- synthesis can be determined at different times in the cycle. The application of high-resolution amino-acid analysers permits their ready separation and quantitation ; moreover, when using [16N]amino-acids, reaction with ninhydrin affords a simple but specific method of converting the a-amino nitrogen to ammonia prior to the determination of isotopic abundance.1 We thought it might be possible to determine the 14C specific radioactivity of amino-acids after reaction with ninhydrin so that the abundance of both isotopes in a single sample could be determined.The reaction of amino-acids with ninhydrin has been well studied: oxidation of the amino- acid releases carbon dioxide and an aldehyde, and the a-amino nitrogen is incorporated in the purple complex diketohydrindylidenediketohydrindamine(DYDA), which is the basis of the colorimetric assay of amino-a~ids.~~~ Most amino-acids are thought to lose one carbon atom during the oxidation, with the exception of aspartate, which loses The value deter- mined for the 14C specific radioactivity in a uniformly labelled sample therefore requires a correction for the loss of carbon dioxide. For this method to be generally applicable, 14C losses other than those by decarboxylation, e.g., those due to the volatility of the aldehyde by-products of the reaction with ninhydrin, must also be accounted This paper reports that conditions that are adequate for reproducible colour production do not necessarily give complete oxidative decarboxylation of amino-acids.The number of carbon atoms lost from glutamate was shown to be dependent on the pH of the ninhydrin reaction mixture. Conditions are described for the reliable determination of the specific radioactivity of [U-14C]amin~-a~id~. Reagents Radiochemicals. These were obtained from the Radiochemical Centre, Amersham, Buckinghamshire. Uniformly labelled L-leucine, L-glutamate, L-arginine and D-glucose were obtained as solutions with specific radioactivities of 311, 265, 336 and 283mCimmol-1, respectively, and at least 83 per cent. of the carbon atoms at each position of the molecule were 14C atoms; the amino-acids were diluted for use to approximately 1 pCi ml-1 in distilled water.Sodium [l*C]carbonate standard solution was obtained at a specific radioactivity of 1.013 pCi per gram of solution and n-[l-14C]hexadecane standard solution at a specific radio- activity of 1.10 pCi per gram of solution; standard solutions were dispensed by mass. Amino-acids. A standard amino-acid mixture in 0.1 M hydrochloric acid was obtained from the Sigma Chemical Company, St. Louis, Mo. ; individual amino-acids were obtained from Cambrian Chemicals Ltd., Croydon. Experimental370 BISHOP AND SIMS : DETERMINATION OF THE SPECIFIC Analyst, V d . 100 Production of Uniformly 14C-labelled Amino-acids of Known Specific Radioactivity C a d d a uutdis NCYC 737 was grown on a glucose (4 g 1-l) - salts - minerals medium1 supplemented with ~-[U-~~C]glucose, at 27 "C for about ten generations in order to ensure that all the carbon atoms in the protein amino-acids of the yeast attained virtually the same specific radioactivities as the carbon atoms in the ~-[U-~~C]glucose from which they were derived.The glucose concentration of the medium was determined with the GOD-Perid blood-sugar test kit supplied by Boehringer Mannheim GmbH. Cells were harvested by filtration on to Whatman GF/C paper, extracted with cold 60 per cent. V/V ethanol and cold 5 per cent. m/V trichloroacetic acid, and washed with water. Proteins were hydrolysed by refluxing in 6 N hydrochloric acid for 15 h ; excess of acid was removed by means of rotary evaporation, washing with 50 per cent.V/V ethanol, and re- drying. The extract was dissolved in a small volume of water and passed through a column of Norit OL charcoal (9 mm i.d. and 17 mm long, pre-washed with absolute ethanol and water) in order to remove purine and pyrimidine bases. The non-aromatic amino-acids were eluted with water, followed by 20 per cent. V/V ethanol; the charcoal retained 99.6 per cent. of the material that absorbed at 260 nm, 19.6 per cent. of the ninhydrin-positive material, and 35.5 per cent. of the radioactivity. The eluted amino-acids were dried by means of rotary evaporation and stored at -20 "C. Ion-exchange Chromatography Amino-acids were fractionated on a Technicon TSM automatic amino-acid analyser accord- ing to the manufacturer's instructions, except that the basic amino-acids were resolved on Chromobead C4 with a buffer solution that was 1.025 M in sodium and 0.1 M in citrate, pH 4-06.The resin in the loading cartridges was replaced with a pile of Whatman GF/C glass- fibre filter circles in order to prevent the hydrolysis of glutamine (A. Wiemken, unpublished).@ Amounts of 2-5 pmol of total amino-acid (dissolved in water) were applied to each column, with 0.125 pmol of norleucine as internal standard and 50 11.1 of 1 M citric acid also applied to the acidic - neutral column. The amino-acids were detected with ninhydrin (by heating for 9 min at 95 "C); the absorb- ances at 570 and 4-40 nm were continuously monitored and recorded. The chart recorder was connected to a Kent Chromalog 3 Wide Dynamic Range Integrator (Kent Instruments Ltd., Luton), linked with a fraction collector so that the whole of the amino-acid peak could be collected as a single fraction to avoid the possibility of differential elution of 14C-labelled and r2CIamino-acids.' The amount of each amino-acid was usually calculated automatically by the integrator although it was occasionally calculated manually from the area of the peak.Determination of a-Amino-acids Several different ninhydrin methods were examined in order to establish which one pro- duces the highest yield of DYDA and gives the most rapid and complete decarboxylation of amino-acid. Standard Technicon reagents were used in the method involving the TSM automatic analyser. The ninhydrin solution consists of 40g of ninhydrin, 2OoOml of 2- methoxyethanol and lo00 ml of 4 M sodium acetate buffer solution (pH 5-51 & 0.02), made up to a total volume of 4000ml with water.The hydrazine sulphate solution consists of 1.049 g of hydrazine sulphate, 3960 ml of distilled water and 40 ml of 30 per cent. m/V Brij 35. The reagents were used in the ratio 0.80 ml min-l of ninhydrin solution: 0.32 ml min-1 of hydrazine sulphate solution: 060 ml min-l of column effluent. The mixture was heated at 95 "C for 9 min. Individual detenninations were carried out manually with a reagent consisting of 1.6 ml of the hydrazine sulphate solution, 1.0 ml of amino-acid solution (up to 0-4 pmol) and 50 pl of [14C]amino-acid solution; in control experiments the amino-acid solution was replaced by 0.8 M sodium acetate buffer solution, pH 5.0.The solutions were heated for 10 min at 100 "C and cooled in cold water for 10 min before measuring the absorb- ance at 570 nm. The method of assay reported by Cocking and Yemm involved the use of potassium cyanide as reducing agents; in the examination of this method the assay mixture used consisted of 1.6 ml of TSM ninhydrin solution, 1.0 ml of a 5 per cent. V/V solution of 0.01 M potassium cyanide in 2-methoxyethanol, 0.5 ml of amino-acid solution and 50 p1 of [14C]amino-acid solution. Ascorbic acid was employed as reducing agent in the dimethyl sulphoxide - ascorbateJune, 1975 RADIOACTIVITY OF UNIFORMLY 14C-LABELLED AMINO-ACIDS 371 (DMA) assay modified from the system of Ferguson and Sims.O To 1 O O m l of dimethyl sulphoxide were added 20 ml of 0.8 M potassium acetate buffer solution, pH 4.9, 1 g of ninhydrin, and 30 mg of L-ascorbic acid dissolved in 1 ml of water; this solution is stable for over a week if kept under vacuum.The assay mixture consisted of 14ml of the above solution, 0.5 ml of amino-acid solution and 50 p1 of [14C]amino-acid solution. The solutions were heated for 20 min at 100 "C. The ninhydrin reagent of Greenstein and Winitz4 was slightly modified as follows: to 1.5 ml of 1 M sodium citrate buffer solution, pH 2.5, were added 0.5 ml of 20 mM L-glutamate and 50 pl of ~-[I1~C]glutamate; dissolved air was removed from the mixture, then 120 mg of ninhydrin were added. The mixture was heated for 10 min at 100 "C. In some experiments the distribution of the radioactivity between water and toluene was examined; 1 ml of the reaction mixture was shaken vigorously with 1 ml of toluene, centri- fuged, and then a 100-p1 portion of each layer was counted.Determination of the amino-acids was performed in a closed water-bath in test-tubes sealed with a glass marble; corrections were made for a slight gain in volume (about 2.3 per cent. in 20min). All absorbance values were corrected for the absorbance of control solutions (without amino-acid) heated for the same length of time. The reaction mixtures of the Cocking and Yemm and the DMA assay systems were diluted with a suitable volume of water before measurement. Determination of Radioactivity The scintillant consists of 6 g of 2,5-diphenyloxazole (PPO), 1000 ml of toluene and 500 ml of Triton X-100 (octylphenoxypolyethoxyethanol) .Radioactivity was determined in the Intertechnique SL 30 liquid-scintillation system, using the pre-set 3H- and 1%-channels. After reaction of the amino-acid with ninhydrin the DYDA formed was destroyed by means of acid hydrolysis, in order to reduce colour quenching. Sufficient 1 M hydrochloric acid (about 0.1 ml) was added to 100 pl of sample from the manual assay systems (mixed with 1 ml of scintillant in a scintillation vial) so as to adjust the solution to below pH 2.5. The mixture was shaken and left for at least 10 min to decolorise. The pale yellow solution was then adjusted with 0.3 ml of 1 N sodium carbonate solution to about pH 5, 9 ml of scintillant were added, and the whole clarified by the addition of 0.6 ml of water prior to counting.DYDA is not produced in the Greenstein and Winitz assay system; hence the sample was adjusted to pH 5, 10 ml of scintillant added and the solution clarified with water. Portions (2 ml) of the total amino-acid peaks eluted from the analyser after reaction with ninhydrin were decolorised and counted in a similar way to those samples obtained from the manual assay system, but adding a total of 15 ml of scintillant. Quenching was corrected by the internal channels ratio method.1° A quench curve for the majority of the protein amino-acids that were eluted from the analyser was obtained by counting a constant amount of [U-14C]glutamate with various amounts of acid-decolorised product from the oxidation of glutamate in the DMA assay system.Reaction of some amino- and imino-acids (lysine, asparagine, ornithine, proline and hydroxyproline) with ninhydrin produces coloured products that are resistant to acid hydrolysis. In these instances separate quench curves should be prepared. The quench correction also varies with the reducing agent employed: hydrazine and ascorbic acid normally reduce the counts by about 25 per cent. and 6 per cent., respectively; this variation can be related to the amount of ninhydrin-based pigments formed.11 Water-soluble compounds were counted at an efficiency of 90 per cent., using sodium [14C]carbonate as an absolute standard ; toluene-soluble compounds were counted at an efficiency of 95 per cent. Stability of Ascorbic Acid Solutions The stability of 2 m~ ascorbic acid dissolved in distilled de-ionised water or in 50 per cent.2-methoxyethanol was examined over a period of 1 week by measuring its absorbance at 256 nm. Results and Discussion The Specific Radioactivity of Uniformly 14C-labelled Amino-acids The feasibility of determining the specific radioactivity of amino-acids after their oxidation with ninhydrin was examined by using a sample of hydrolysed yeast protein of known specific372 BISHOP AND SIMS: DETERMINATION OF THE SPECIFIC Analyst, 'Vd. 100 radioactivity. The amino-acids were fractionated and quantified, and the specific radio- activity calculated on the basis of disintegrations per minute per microatom of carbon (d.p.m. patom-1); a loss of one carbon atom was assumed from each molecule of amino-acid, except for aspartate.The results deviated from the expected specific radioactivity in two ways: lysine, arginine, aspartate, threonine, serine and glycine appeared to have a much higher specific radioactivity than the glucose from which they were derived, whereas in alanine, valine, isoleucine and leucine they appeared to be much lower. From the values calculated it seemed that only about 0.5 carbon atom was lost from lysine and glycine during ninhydrin oxidation and that only about 1.5 carbon atom was lost from aspartate. The simplest explanation for these results is that , under the conditions employed, oxidative decarboxylation of the amino-acids is incomplete. The very low recovery of radio activity in some instances indicated that appreciable losses of carbon other than as the a-carboxyl carbon atoms must have occurred (possibly as volatile compounds). In order to develop conditions where carbon losses could be reliably reproduced we examined the kinetics of carbon loss during the ninhydrin reaction. Carbon Loss from Glutamate and Arginine During Reaction with Ninhydrin Various amounts (0.04-4-0 pmol) of [U-14C]glutamate, in a constant volume, were heated with the DMA reagent (ninhydrin, dimethyl sulphoxide, potassium acetate buffer solution, pH 4.9, and ascorbic acid) for 20 min at 100 "C and then cooled by placing the tubes in water ; TABLE I LOSS OF l4C FROM [U-14C]GLUTAMATE BY OXIDATION WITH NINHYDRIN Radioactivity of [14C]glutamate added : 21 38 1 d.p.m. pmol-1.Glu tamatelpmol 0-04 0.10 0.20 0.40 1.0 2.0 3.0 4.0 Radioactivity remaining per micromole of glutamate added/d.p.m.11 960 12 830 12 930 13 610 12 880 12 790 12 790 13 480 Recovery* of 14C in solution, per cent. 56.0 60.1 60-4 63-8 60.3 59-8 59.8 63.0 * The mean recovery in solution is 12 910 d.p.m. pmol-l (60-4 per cent.). samples of the solution in each tube were then decolorised and counted as described under Experimental. Table I indicates that there is a loss of two carbon atoms per molecule of amino-acid on reaction with ninhydrin and that this loss is independent of concentration. The identity of the carbon compound lost in the reaction was ascertained as follows. [U-14C]glutamate (0.2pmol) was heated with the DMA reagent in Thunberg tubes for TABLE I1 RECOVERY O F VOLATILE COMPONENTS AFTER OXIDATION WITH NINHYDRIN OF GLUTAMATE AND OF LEUCINE Radioactivity collected in side-arm/d.p.m.Initial Radioactivity Amino-acid radioactivity/ remaining in ninhydrin d.p.m. mixture/d . p . m . Glutamate 110 000 66 700 63 160 67 720 61 980 Leucine 175 260 91 890 81 340 44 660 96 270 Sodium hydroxide solution 35 240 - - 34 880 26 930 - - 27 440 Saturated sodium hydrogen Hydrochloric sulphite acid solution - - 40 - 20 - - -June, 1975 RADIOACTIVITY OF UNIFORMLY 14C-LABELLED AMINO-ACIDS 373 20 min at 100 "C; the side-arm of each tube contained 0.2 ml of either 2 M sodium hydroxide solution, 2 M hydrochloric acid or saturated sodium hydrogen sulphite solution. An aliquot of the ninhydrin mixture was decolorised and counted. The entire content of the side-arm was counted directly.Table I1 indicates that a minimum of 32 per cent. of the released carbon was in the form of carbon dioxide and 57-60 per cent. of the carbon remained in the reaction mixture; thevolatile component was confirmed as carbon dioxide, as 99 per cent. of the counts were lost when acid was added to another alkali-containing side-arm (results not shown). The loss of two carboxyl groups from the glutamate molecule during the reaction with ninhydrin has not been reported before. We have also confirmed that using the method of Greenstein and Winitz only one carboxyl group is lost (83.8 per cent. of the radioactivity was retained in solution).4 Manual TSM ninhydrin assay of glutamate produced, after 10 min of heating at 100 "C, quantitative loss of the two carboxyl carbons but only 49 per cent.of the theoretical yield of DYDA (assuming a molar absorptivity at 570 nm of 22 000 1 mol-l crn-l).l2 That amino- acid decarboxylation was not closely associated with DYDA formation under these conditions . A w 1 1 1 - .r I I 0 f 2 . 1 I 5 10 15 20 25 30 Ti me/minu tes Time/minutes Fig. 1. (a) Loss of 14C and (b) colour formation from ~-[U-~~C]glutamate A, and ~-[U-l~C]arginine by oxidation with ninhydrin in the TSM assay. L-arginine ; B, L-glutamate. was confirmed by following DYDA formation and carbon loss, with a dicarboxylic amino-acid (glutamate) and a monocarboxylic amino-acid (arginine) (Figs. 1 and 2). For the TSM assay 0.4 pmol of glutamate and 0.1 pmol of arginine were heated for various times at 100 "C and then cooled; an aliquot of the solution was decolorised and counted as described under Experimental.The remaining solution was used for measurement of the absorptivity at 570 nm; a correction was made for the absorptivity of the samples without amino-acid. 100 80 60 0 E x In w W S -E P a P Timehinutes Time/minutes Fig. 2. (a) Loss of 14C and ( b ) colour formation from ~-[U-l~C]glutamate and A, L- ~-[U-1*C]arginine by oxidation with ninhydrin in the DMA assay. arginine; B, L-glutamate.374 Analyst, VoZ. 100 For the DMA assay 0.2 pmol of glutamate and 0.05 pmol of arginine were taken and the same procedure carried out. The graphs indicate that the times taken to complete the re- actions of the two amino-acids differ. The loss of two carbon atoms was confirmed for glutamate with TSM and DMA reagents; only one carbon atom was lost from arginine.The constancy of 14C recovery up to 30 min indicates that the oxidation products of glutamate and arginine remain in solution. The results also indicate that 14C loss is completed in under half the time required for full colour development. There was, however, a large difference in the time required for loss of carbon dioxide from the two systems; in the TSM system, complete loss of carbon from glutamate required about 15 min, whereas in the DMA system it occurred within 4 min (if ascorbate is used as reducing agent in the TSM system, about 8 min is required). BISHOP AND SIMS : DETERMINATION OF THE SPECIFIC Carbon Loss from Leucine During Reaction with Ninhydrin PCJ-14CIleucine (0.2 pmol) was taken for an assay with the DMA reagent, which was carried out under the same conditions as for glutamate and arginine (see above).Of the 14C added as leucine, 15.5 per cent. was trapped by the sodium hydroxide solution, which indicates that one carbon atom is lost by ninhydrin oxidation of the C, amino-acid (Table 11). However, only 50 per cent. of the counts remained in the reaction mixture; a significant amount of radioactivity was associated with the saturated sodium hydrogen sulphite solution. I I I I I I 0 5 10 15 20 25 30 '-5 Time/mi nutes Fig. 3. Fate of ~-[U-l~C]leucine oxidised by means of the TSM system in tubes sealed with a glass marble. A, total radioactivity in solution ; B, toluene-soluble radioactivity ; C, water-soluble radioactivity.Fig. 3 shows a time-course of changes in the reaction mixture of leucine (TSM assay system). Leucine (0.4 pmol) was heated for various times at 100 "C. After cooling, an aliquot of the solution was decolorised and counted; a further aliquot was divided into water-soluble and toluene-soluble fractions, which were also counted. The rapid fall of total counts and the rise of toluene-soluble counts is probably due in part to the loss of p4C]carbon dioxide from the carboxyl group of [U-14C]leucine; there is then a slow conversion of toluene-soluble counts into water-soluble counts. Another experiment was carried out by using a sealed 5-ml screw-capped bottle in order to prevent loss of products through volatilisation during heating. The total counts fell initially and subsequently rose until after 60 rnin they approached the theoretical yield (83.3 per cent.).This experiment suggests that the volatile substance formed initially is subsequently converted to a water-miscible compound. Attempts were made to retain the volatile product in solution. The ninhydrin reagent of Cocking and Yemms containing potassium cyanide was shown to prevent completely the loss of aldehyde during oxidation of leucine with ninhydrin but the reagent blank was considered too high for satisfactory colorimetric analysis. Semicarbazide was shown to be equallyJane, 1975 RADIOACTIVITY OF UNIFORMLY 14C-LABELLED AMINO-ACIDS 375 effective; however, as it is ninhydrin positive it must be added to the cooled reaction mixture just prior to the measurement of the absorbance of the solution at 570 nm.Modifications to the Automatic Amino-acid Analyser The oxidation time-courses of the amino-acids studied indicated that at least 15 min at 100 "C were required for completion of the reaction. A heating coil which gave 17 min of heating at 95 "C was shown to be satisfactory. Semicarbazide (50 mM adjusted to pH 7 in 50 per cent. V/V 2-methoxyethanol) was in- jected, at a rate of 0.015 ml min-1, into the analyser stream after it had passed the cooling coil. Comparison of a standard mixture of amino-acids in the presence or absence of the trapping agent established that no alteration in sensitivity or peak width of the amino-acid occurred at this injection rate. Hydrazine was replaced by 2 mM ascorbic acid as reducing agent.The stock TSM-ninhydrin solution was retained except that 4 M sodium acetate buffer solution was replaced by 0.8 M potassium acetate buffer solution for increased colour yield and ease of decolorisation. The results of analyses of yeast protein amino-acids using these modifications are shown in Table 111. It can be seen that the specific radioactivities (d.p.m. patom-l) of most amino- acids correspond to within 10 per cent. of the specific radioactivity of the [U-14C]glucose from which they are derived, even with samples as small as 0-02 pmol of amino-acid. TABLE I11 SPECIFIC RADIOACTIVITY OF AMINO-ACIDS ESTIMATED BY THE MODIFIED PROCEDURE Specific radioactivity of ~-[U-1'C]glucose : 6691 d.p.m. patom-' Specific radioactivity* of amino-acid A r \ Amino-acid LYS ASP Thr Ser Glu GlY Ala Val Ile Amino-acid in sample/pmol 0.1 104 0-1032 0.0199 0-0165 0.0109 0.0840 0.1032 0.0064 0.1249 0.1331 0.2034 0-2125 0.0649 0-0774 0.0295 0.0229 0-0298 0.0500 0-0393 0.0488 d.p.m.pmol-l 35 140 31 955 29 120 30 300 14 664 17 799 13 872 20 106 5 328 8 993 12 616 14 318 26 320 26 556 31 425 36 685 30 680 38 540 30 950 30 310 * Up to 3 determinations of radioactivity were presented. d.p.m. patom-lt 7028 0391 5824 6072 7332 5933 6936 6702 5328 8993 6308 7159 6580 6389 6285 7337 6116 7708 6190 6062 per cent. of expected value in d.p.m. patom-' (average) 100.3 89.0 109.6 88-7 103.6 100.2 107-0 100.6 96.9 98.3 99.4 made for each amino-acid: the individual values are t The specific radioactivity in d.p.m. patom-l was calculated assuming the loss of one carbon atom for each amino-acid except for aspartate and glutamate for which the loss of two carbon atoms was assumed.Factors that affect the formation of DYDA were examined; the molar absorptivity at 570 nm was proportional to the ascorbic acid concentration at levels below 2 mM, but varied little above that concentration (the background colour increased linearly at concentrations above 2 mM). Changes in the absorbance at 256 nm of stock 2 mM ascorbic acid prepared in water376 BISHOP AND SIMS indicated that about 50 per cent. was lost within 30 h, hence solutions of this concentration must be prepared fresh daily. However, if 2 mM ascorbic acid is prepared in 50 per cent. 2-methoxyethanol its half-life is increased to about 100 h ; a stock solution of 13 mM ascorbic acid in 50 per cent. 2-methoxyethanol, and suitably diluted by the analyser, is stable for a full working week.Conclusion In this paper we describe how we attempted to modify an automatic amino-acid analyser in order to measure the 14C and 15N content in a single sample. Complete loss of the a- carboxyl group has been achieved with the amino-acids studied. Ascorbic acid was chosen as the reducing agent in the ninhydrin reaction for four reasons: it is free of nitrogen, gives a very low base-line at 570 nm and an almost theoretical yield of DYDA (93 per cent.), and there is less quenching of 14C. The largest source of error in the determination of specific radioactivity is the counting of 14C. The major source of quenching is ninhydrin and ninhydrin-based pigments that are produced in side-reactionsll ; these pigments absorb strongly near the fluorescence wavelength of PPO, and samples that contain large amounts of pigments may show a complete loss of counts from the l*C-channel of the liquid-scintillation spectrometer. 1% may be counted at high efficiency in the Intertechnique SL 30 system with PPO as sole scintillant .Addition of a secondary scintillant, 1,4-bis-(4-methyl-5-phenyloxazol-2-yl)- benzene (dimethyl-POPOP ; 0.6 mg ml-l) to highly quenched samples of decolorised amino- acids restored counts to the 14C-channel and reduced the quench correction factor in all instances. We recommend the addition of dimethyl-POPOP to all amino-acid samples of low specific radioactivity. In order to use the analytical system for the determination of 14C and 15N it is desirable that the nitrogen determined should derive from the amino-acid and not from components in the buffer solutions. Thus, although in principle we have shown that volatile aldehyde products produced by the oxidation of valine, leucine and isoleucine can be trapped by the addition of semicarbazide to the buffer solution, we have as yet been unable to achieve this by using a compound that contains no nitrogen. We are grateful to Mr. S. Howitt for help and advice on the operation of the amino-acid analyser. We are also grateful for financial support from the Science Research Council (Grant No. B/RG/2412). 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Ferguson, A. R., and Sims, A. P., J . Gen. Microbiol., 1971, 69, 423. Moore, S., and Stein, W. H., J . Biol. Chem., 1948, 176, 367. Stein, W. H., and Moore, S., J . Biol. Chem., 1954, 211, 893. Greenstein, J . P., and Winitz, M., Editors, “Chemistry of the Amino Acids,” Volume 2, John Wiley & Virtanen, A. I., Laine, T., and Toivonen, T., Hoppe-Seyler’s 2. Physiol. Chem., 1940, 266, 193. Wiemken, A., unpublished work. Gaitonde, M. K., and Nixey, R. W. K., Analyt. Biochem., 1972, 50, 416. Cocking, E. C., and Yemm, E. W., Biochem. J., 1954, 58, xii. Ferguson, A. R., and Sims, A. P., J . Gen. Microbiol., 1974, 80, 159. Patterson, M. S., and Greene, R. C., Analyt. Chem., 1965, 37, 854. McCaldin, D. J., Chem. Rev., 1960, 60, 39. Friedman, M., and Sigel, C. W., Biochemistry, 1965, 5, 478. Sons Inc., New York, 1961, p. 1322. Received Octobev 22nd, 1974 Accepted January 13th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000369

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, June, 1975, Vol. 100, pp. 377-385 377 Application of Differential Pulse Polarography to the Assay of Vitamins J. Lindquist" and S. M. Farrohat. Department of Analytical Chemistry, University of Uppsala, Box 631, S-761 21, Uppsala 1, Sweden The determination of low concentrations of the vitamins B,, nicotinic acid, nicotinamide, ascorbic acid, K, and K, in pure solutions by use of differential pulse polarography has been studied. The approximate limits of detection achieved for the different vitamins varied from 0.01 to 1 p.p.m., which is 10 to 100 times better than for d.c. polarography. A comparative deter- mination of vitamin B, and nicotinamide in three complex multivitamin preparations showed that differential pulse polarography is more selective than d.c. polarography.With the former method the nicotinamide can be determined directly, i.e., without prior separation of interfering substances. Most of the vitamins can be determined po1arographically.l Even the vitamins of the A and D groups give polarographic waves in solvents such as benzene - acetonitrile,2 dioxan - water3 or dime t hylf~rmamide.~ In textbooks that deal with methods of vitamin assay, several polarographic procedures are given, especially for determining water-soluble vitamin^.^^^ However, most of the applications are for pharmaceutical preparations in which the vitamins are present in relatively high concentrations. Only for ascorbic acid are polarographic procedures given even for deter- minations in biological materials such as fruits and vegetables. In those instances when the method is applicable it is often claimed to be simple, specific and accurate.One limitation of polarography has been the relatively low sensitivity attainable. The method is best suited to determinations at concentrations between about 1 x and 2 x 10-3M.' Analysis of 10-6 M solutions can be performed but only under favourable conditions and in the absence of interfering substances. In order to obtain accurate results it is advantageous if the waves are in the form of easily measurable steps, which condition occurs in the concentration range between 10-4 and The lower limit can, however, be considerably extended by the use of other voltammetric methods. The application of a pulse technique to polaro- graphy results, for instance, in the possibility of detecting reversibly reduced species at concentrations as low as 10-8 M, and irreversibly reduced species at 5 x 10-8 ~ .8 The polarographic method is rather selective, as waves are observed in finite ranges of potentials. A 100-mV interval between two half-wave potentials can be sufficient to permit measurement of the corresponding wave heights if the concentration ratio is not less than 10: 1. The differential pulse technique is even more efficient because it yields peak shapes that approximate closely to the theoretically predicted derivative of an ordinary d.c. polaro- gram. It thus enables one to obtain the maximum possible resolution between closely spaced waves. A 40-mV interval between two peak potentials can be enough for the corresponding peak heights to be measured, even if the concentration ratio is as high as lo4: l.9 The applications of modern polarographic methods to vitamin assay have hitherto been very few; they include the determination of riboflavine by ax.polarography,1° and ribo- flavine, thiamine and nicotinamide in pharmaceutical preparations by cathode-ray polaro- graphy.ll The reason for this deficiency is probably the complex instrumentation that is required for use with modern techniques. Recent advances in commercial instrumentation have, however, provided equipment at a price equivalent to that of the older d.c. polaro- graphs, with which both modern pulse polarography and classical d.c. polarography can be carried out. M. * Present address : Research and Development Laboratories, Astra Lakemedel AB, S-151 85, Sodertalje, t Present address : Chemistry Department, College of Science, Bagdad University, Bagdad, Iraq.Sweden.378 Instrumentation D.c. polarograms were recorded with a Radiometer Polariter PO4 instrument and differen- tial pulse polarograms with a PAR, Model 174, polarographic analyser. A saturated silver - silver chloride electrode (d.c. polarography) and a saturated calomel electrode served as reference electrodes, a platinum wire being employed as auxiliary electrode. All potentials given are referred to the saturated calomel electrode. The solutions were de-aerated with oxygen-free nitrogen and controlled thermostatically at 25-0 & 0.1 "C. A controlled drop time of 2 s was used for the differential pulse polarography.Although longer drop times increase the ratio of faradaic to capacitative current to a maximum, the drop time of 2 s was chosen because it gave better reproducibility (greater drop stability). LINDQUIST AND FARROHA : APPLICATION OF DIFFERENTIAL Analyst, Vol. I00 Experimental Reagents biochemical use were obtained from E. Merck AG, Darmstadt. K,) was obtained from Sigma Chemical Company. The vitamins menadione (KJ, nicotinic acid, nicotinamide and riboflavine prepared for Phytomenadione (vitamin Lithium hydroxide. Pure. All other chemicals used were of analytical-reagent grade. Results and Discussion Riboflavine (Vitamin B,) This vitamin forms a reversible redox system the half-wave potential of which is identical with the actual redox p~tential.~ As hydrogen ions are consumed in the reduction, the potential is shifted to more negative values with increasing pH12 The electrochemical behaviour of riboflavine has been extensively studied.RT [H+-I3 + K,[H+I2 nF [H+l + K , Ei = E" + ~ In where E" = -0.056 V, K, = 5 x lo-' and K , = 0-6 x 10-10. In acidic solution, a characteristic pre-wave appears on the polarogram, which is formed by the adsorption of the reduced form of rib~flavine.~ The polarogram is well developed over a wide pH range but a pH of about 7 is most often recommended for analytical purposes. Calibration graph by d c . polarography The linear concentration range was measured with pure riboflavine solutions in 0.1 ~3 phosphate buffer of pH 7.2. A 100 p.p.m. stock solution was prepared by dissolving 50 mg of riboflavine (previously dried in a vacuum desiccator over concentrated sulphuric acid) in 500 ml of water in a calibrated flask.Before making the volume up to the mark, 5 g of sodium salicylate were added in order to facilitate dissolution. From this solution a series of solutions was prepared by serial dilution, in the concentration range of riboflavine froin 1 to 50 p.p.m. Each solution in turn was transferred into the polarographic cell, de-aerated and polarographed and two or three polarograms were recorded for each solution. The diffusion current (id) was linearly dependent on the concentration down to 1 p.p.m., which is very near the limit of detection (Fig. 1). The relative standard deviation of id/C was &2-3 per cent. for the range 10-50 p.p.m., A0.6 per cent.for 4-12 p.p.m. and 43.3 per cent. for 1-0-2-0 p.p.m. Calibration graph by diflerential pulse polarography Fig. 1 shows a d.c. polarogram and a differential pulse polarogram of a 1 p.p.m. solution of riboflavine in 0.1 M phosphate buffer of pH 7.2. At this low concentration the d.c. polaro- gram is too poorly defined to be useful for the determination of the vitamin. The peak obtained by the pulse technique, on the other hand, is excellent for that purpose. However, when the concentration of the vitamin was diminished, the slope of the base-line changed, which made the evaluation of the peak height difficult even at a concentration of 0.2 p.p.m. (Fig. 2); this difficulty was not expected because the peak was well developed at the 1 p.p.m.level, but is probably caused by the adsorptive properties of riboflavine. Although the double-layer charging current associated with the potential step will decay to zero, there is another effect that limits the sensitivity of differential pulse polarography.June, 1975 PULSE POLAROGRAPHY TO THE ASSAY OF VITAMINS 379 I10 nA 0 -0.15 -0.30 -0.45 -0.60 -0.75 -0.90 Potential versus S.C.E./V -0.1 5 -0-30 -0.45 -0.60 -0.75 Potential versus S.C.E./V Fig. 2. A differential pulse polarogram of 0.3 p.p.m. of riboflavine in phosphate buffer of pH 7.2. Pulse amplitude 25 mV. Fig. 1. A d.c. polarogram (1) and a differential pulse polarogram (2) of 1 p.p.m. of riboflavine in phosphate buffer of pH 7-2. Pulse amplitude 50 mV. This effect is due to the capacitative current that flows at a growing drop and is dependent on the different times, and potentials, at which measurements are made.13 Hence, the changes in the double-layer capacity caused, for instance, by adsorption will be of decisive importance for the shape of the base-line.Strong adsorption not only of leuco-riboflavine but also of riboflavine itself has been demonstrated by using the a.c. polarographic method14 (an a,c. wave of riboflavine at concentrations as low as 2 x ~O-'M can, as a consequence of this fact, be obtainedlO). For the best conditions for differential pulse polarography of riboflavine at low concentrations, a modulation amplitude of 25 mV and a scan rate of 2 mV s-1 were required. A calibration graph from 1 to 0.1 p.p.m. of riboflavine is shown in Fig.3. Nicotinamide The reduction of nicotinamide corresponds to a 2-electron change and the wave is well defined in buffers of pH 8 or above. The half-wave potential is -1.75 V in 0.1 N sodium hydroxide solution.15 For the quantitative determination, sodium hydroxide solution is usually recommended for use as supporting electrolyte.15J6 The wave lies, however, near the discharge of the sodium ion, which makes the evaluation rather difficult. As the lithium ion has a lower reducing potential, lithium hydroxide was chosen instead of sodium hydroxide in this work, and Fig. 4 shows that 0.1 M lithium hydroxide solution is a much more satis- factory supporting electrolyte. Calibration graph by d.c. polarogvaphy A 1-100 p.p.m. concentration range of pure nicotinamide solutions was measured in 0.1 M lithium hydroxide solution.Two stock solutions of 100 and 1000 p.p.m. were prepared by dissolving 50 mg of nicotinamide in 500 and 50 ml of water, respectively. These solutions were diluted with lithium hydroxide solution to give appropriate concentrations of the vitamin. The wave heights (id) determined for the various calibration solutions were found to be linearly dependent on the concentration and the relative standard deviation of i d / C was kl.5 per cent. for the whole of the measured range. The limit of detection is well below380 LINDQUIST AND FARROHA: APPLICATION OF DIFFERENTIAL Analyst, Vol. 100 1 p.p.m., probably 0.2-0.3 p.p.m., i.e., polarograms were similar to that for 1 p.p.m. of riboflavine (Fig. 1). .o 1 0.1 1 .o 10 loo Concentration/pg ml-' Potential -+ 3.Differential pulse-polarographic calibration graphs of the 1 to 2 lowest concentration decades for nicotinamide, nicotinic acid and vitamins B,, C, K, and K,. Each graph is normalised to constant pulse amplitude. Fig. 4. D.c. polarograms of nicotin- amide (8 p.p.m.): 1, in 0.1 M NaOH; and 2, in 0.1 M LiOH solution. Calibration graph by diferential pulse polarography The peaks were well developed in the range 0-1-1 p.p.m. and the detection limit was as low as about 0.01 p.p.m. At concentrations below 0.1 p.p.m. problems in the evaluation of the peak heights arose because of the changing background. Fig. 5 shows some peaks of the interval 0.1-0-01 p.p.m., which are highly reproducible, but the peak heights are not a linear function of the concen- tration if they are measured from the peak to the lowest part of the curve or to the mean of the lowest two parts of the curve.Concentrations up to 1 p.p.m. were studied with the pulse technique. Potential + Fig. 5. Differential pulse polarograms of 1, 0.1 ; 2, 0.08; 3, 0.04; 4, 0.02; and 6, 0.01 p.p.m. of nicotinamide in 0.1 M LiOH solution. The linearity was, however, improved by changing the pulse amplitude from 100 to 50 mV Fig. 3 shows a calibration graph for the The scan rate was 2 mV s-l, with pulse amplitude The whole of the but the limit of detection was thereby impaired. concentration interval 14-02 p.p.m. 25 mV in the range 1-0.1 p.p.m. and 50 mV in the range 0.1-0.02 p.p.m. curve has been normalised to a 25-mV amplitude.June, 1975 PULSE POLAROGRAPHY TO THE ASSAY OF VITAMINS 381 Nicotinic acid Buffers between pH 8 and 9 are normally used for the determination of nicotinic acid because the waves are relatively well developed in that range.5 Both nicotinamide and nicotinic acid can be determined in the same solution; their sum is determined at pH 8.0 and nicotinamide at pH 13, at which pH the acid does not give a wave.5 The linear concentration range for nicotinic acid (by d.c.polarography) is from about 40 to 250 ~ . p . m . ~ Calibration graph by diferential pulse polarography A 0.1 M borate buffer of pH 8.7 was used as supporting electrolyte and concentrations up to 100 p.p.m. were studied. The sensitivity was not as good as that for nicotinamide and the limit of detection was about 1 p.p.m.when a pulse amplitude of 10 mV was used. The best linearity was obtained with an amplitude of 5 mV and a scan rate of 1 mV s-1. Fig. 6 shows some peaks in the range 1-20 p.p.m. and Fig. 3 the entire calibration graph. Nicotinic acid gives a catalytic hydrogen wave in the pH range from 1 to 10. Ascorbic acid The polarographic assay of vitamin C is stated to be more specific and to require fewer steps in sample preparation than colorimetric and titration methods1 The relatively low sensitivity of the method and the limited anodic voltage range of the mercury electrode have been the main limitations. The normal concentration range is between 25 and 250 p.p.m. of ascorbic acid.6 Buffers with pH between 3.4 and 6 are normally used as supporting electro- lytes.In more strongly acidic solutions the anodic wave has a very flat slope and lies too near the background current, and at higher pH values ascorbic acid is too readily oxidised.6 Calibration graph by difeerential pulse polarography The supporting electrolyte was an acetate buffer saturated with sodium oxalate and its pH was 5-5.6 Concentrations up to 25 p.p.m. were studied. Excellent peaks were obtained with concentrations down to 0.2 p.p.m. The peak half-width was 46 mV at 25-mV pulse amplitude, which is close to the theoretical value for a reversible 2-electron process.17 A changing base-line made the evaluation troublesome at concentrations below 0.2 p.p.m., as shown in Fig. 7. If the peak heights are measured from the highest to the lowest parts of the curves, the value at 0.1 p.p.m.is half that at 0.2 p.p.m., but in order to ensure proportion- ality in the 0-1-0.025 p.p.m. range, the peak heights must be measured from the break points indicated by the arrows. The best modulation amplitude was 5 mV with respect to both the detection limit and the linearity. Fig. 3 shows a calibration graph for the range 0.025- 2 p.p.m. L ' I I -1.5 -1.65 -1 *80 Potential versus S.C.E./V Fig. 6. Differential pulse polaro- grams of 1, 1; 2, 5; 3, 10; and 4, 20p.p.m. of nicotinic acid in borate buffer of pH 8-7. 5- 4 Potential + Fig. 7. Differential pulse polarograms of 1, 0.2; 2, 0.1; 3. 0.05: and 4, 0-025 p.p.m. of ascorbic acid a t pH' 515. Pulse amplitude- 50 mV, peak potential +0-217 V.382 LINDQUIST AND FARROHA : APPLICATION OF DIFFERENTIAL Analyst, vol.100 Vitamin K, (Phytomenadione) and K, (Menadione) Supporting electrolytes consisting of a 0.06 M solution of ammonium chloride in 75 per cent. propan-2-o15 or a 0.5 M solution of tetrabutylammonium iodide in acetonitrile18 have been used for the polarography of vitamin K,. Similar electrolytes have been used for vitamin K,,6 but, according to Patriarche and Lingane,lg lower pH values are to be preferred because of the instability of vitamin K, in alkaline media. Calibration graph by diferential pulse fiolarography Concentrations below 5 p.p.m. of vitamin K, in a 0-06 M solution of ammonium chloride in 75 per cent. propan-2-01 and concentrations below 2 p.p.m. of vitamin K, in 0.1 M acetate buffer of pH 5.0 containing 25 per cent.of methanol were studied. Very well developed peaks of vitamin K, were obtained in the concentration range 2-5 p.p.m. and Fig. 8 shows a peak at 2.7 p.p.m. Below about 0-5 p.p.m. (1 PM) the slope of the base-line makes the evalu- ation increasingly uncertain. This effect is also shown in Fig. 8, where some polarograms can be seen for concentrations near the limit of detection (about 0.1 p.p.m.). The pulse amplitude used was 50 mV. Fig. 9 shows two peaks at 0.069 and 0.17 p.p.m. (pulse amplitude 25 mV). Calibration graphs for both vitamins are shown in Fig. 3. The limit of detection for vitamin K, is as low as about 0.02 p.p.m. t CI E 3 -I ,d, I -0.30 -0.45 - I I 15 -0.30 -0 Potential versus S.C.E./V Potential versus S.C.E./V Differential pulse polarograms of (left) : 1, 0.09; 2, 0.18; and 3, 0-45 and (right) : 2-7 p.p.m.of vitamin K, in 0-06 M solution of NH,C1 in 75 per cent. propan-2-01. Fig. 8. Pulse amplitude 50 mV. 5 Fig. 9. Differential Dulse polarvograms of 1, 0.069; aAd 2, 0-17 p.p.m. of vitamin K, in 0.1 M acetate buffer of pH 5.0 containing 25 per cent. of methanol. Pulse amplitude 25 mV. Comparative Determination of Vitamin B, and Nicotinamide in Three Complex Vitamin Preparations In some instances, vitamins (especially vitamin B,) can be determined directly, even in complex vitamin preparations, simply by adding supporting electrolyte to the sample, de-aerating the solution and recording a d.c. polarogram.6 Because of the better selectivity of the differential pulse technique, it should be possible to increase the number of instances when such direct determinations can be made, and in order to verify this conclusion a corn- parison was made between the two techniques.The declared contents of 100 ml of the vitamin mixtures analysed were as follows. Vitalvin : thiamine hydrochloride, 11 mg ; riboflavine, 7 mg ; pyridoxine hydrochloride, 7 mg; nicotinamide, 0.11 g ; ethanol, including the content of added wine, 10 g ; sugar, 28 g; sodium phosphate ; citric acid ; ascorbic acid ; apple syrup ; blackcurrant syrup ; raspberry syrup ; and aromatic agents.June, 1975 PULSE POLAROGRAPHY TO THE ASSAY OF VITAMINS 383 Roburan: caffeine, 0.27 g; vitamin A, 17 000 i.u. ; vitamin D,, 1700 i.u. ; thiamine hydro- chloride, 10 mg ; sodium riboflavine phosphate, 15 mg ; nicotinamide, 80 mg ; pyridoxine hydrochloride, 7 mg ; pantothenol, 27 mg ; ascorbic acid, 0.5 g ; sorbitol; and constituents.Pharmaton : Thiamine hydrochloride, 11 mg ; riboflavine, 7 mg ; pyridoxine hydrochloride, 7 mg; nicotinamide, 0.11 g; disodium methylarsonate, 0.11 g ; caffeine, 0.5 g; ethanol, 20 g; sugar, 33 g; sodium phosphate; citric acid; cherry syrup; aromatic agents; and colouring matter . Determination of Vitamin B, Five 5-00-ml samples each of Vitalvin, Roburan and Pharmaton were pipetted into 50-ml calibrated flasks; 5 ml of phosphate buffer solution of pH 7-2 were added and to the last two flasks in each series of five samples, 1.50 and 2.00 ml of standard vitamin B, solution were added. The solutions were diluted to the marks with water and then about 15ml were de-aerated in the polarographic cell and polarographed.Vitalvin and Roburan gave well developed polarograms of vitamin B, and the concen- trations of the vitamin were calculated (Table I). The polarograms obtained with the Pharmaton samples were, however, affected by the presence of an interfering substance, which made the evaluation inaccurate. Because of the high reducing potential of this sub- stance it was assumed that it was a colour additive (possibly an azo compound) and as most edible colour additives are anions that it would be easy to remove by means of an anion exchanger. A chloride-saturated anion exchanger (Dowex 1 x 8, 50-100 mesh) was effective in removing it; 5.00-ml samples of Pharmaton were pipetted on to the column (1 x 4 cm) and the ion exchanger was washed five times with 5-ml portions of water.The eluates were collected in the calibrated flasks and polarographed as described above. The polarograms were then easily evaluated and the concentration of the vitamin was calculated (Table I). TABLE I RIBOFLAVINE CONTENT OF THE SAMPLES Concentration of riboflavine, p.p.m. L P \ Found by Found by d.c. differential pulse Sample Declared polarography polarography Vitalvin . . .. .. .. 7 6.61 -& 0.16 - Roburan .. .. .. 16 14.52 f 0.30 - Pharmaton . . .. .. 7 6-90 & O*OS* (7.28) * After separation of interfering substance. As the Pharmaton sample could not be analysed by d.c. polarography without applying this separation procedure, it was subjected to differential pulse-polarographic analysis.Fig. 10 shows the polarograms before and after the addition of standard. The riboflavine peak is resolved but the background current changes after addition of the standard, which makes the evaluation very difficult. The peak heights measured to either of the broken lines in Fig. 10 gave results in accordance with those obtained by d.c. polarography after carrying out the separation procedure, but these lines are not satisfactory because similar lines cannot be drawn after the addition of standard. Hence, in this instance the method requires to be examined further. Analysis of the Samples for Nicotinamide Direct d.c. polarography of nicotinamide in multivitamin preparations is usually not possible because there is a greater risk of interference with the nicotinamide wave owing to the low reducing potential of the vitamin (-1.75 V) ; 0.50-ml samples were pipetted into 50-ml calibration flasks, 5 ml of supporting electrolyte (1 M lithium hydroxide solution) were added and the solutions were then diluted to the mark with water.After de-aeration the solutions were polarographed. Nicotinamide could not be determined directly by d.c. polarography except in Roburan samples. A simple anion exchanger6 (hydroxyl-saturated Dowex 1 x 8, 50-100 mesh) could, however, be used even in this instance to remove interfering substances. The results384 LINDQUIST AND FARROHA : APPLICATION OF DIFFERENTIAL Avzalyst, VoZ. 100 TABLE I1 NICOTINAMIDE CONTENT OF THE SAMPLES Concentration of nicotinamide, p.p.m. - \ Found by d.c.differential pulse Found by Sample Declared polarography polarography Roburan .. .. .. 80 83 f 3 - 88 f 1* - Vitalvin . . .. .. .. 110 109 f 3* 113 Pharmaton . . . . . . 110 113 f 1* 111 * After separation on ion exchanger. obtained are given in Table 11. Fig. 11 shows differential pulse polarograms of the samples Vitalvin and Pharmaton that could not be analysed by d.c. polarography without previous separation of interfering substances. The base-lines (broken lines in Fig. 11) were drawn by using polarograms of pure supporting electrolyte. The measured results agreed well with the values obtained by d.c. polarography (Table 11). Potential -+ Fig. 10. Differential pulse polaro- grams of a complex vitamin preparation (Pharmaton) in phosphate buffer of pH 7.2: 1, before and 2, after addition of standard riboflavine solution.Broken lines denote arbitrarily drawn background currents. Potential + Fig. 11. Differential pulse pol- arograms of two complex vitamin preparations [Vitalvin (1) and Pharmaton (2)] in 0.1 M LiOH solution before and after addition of standard nicotinamide solution. Broken lines denote residual current. Conclusion At the limit of detection for d.c. polarography the pulse technique gives excellent peaks and if the limit of determination is taken as about five times the limit of detection, it is possible to determine nicotinamide at 0-05, ascorbic acid and vitamin K, at 0.1, vitamins B, and K, at 0.5 and nicotinic acid at 5 p.p.m. The selectivity is also excellent as nicotinamide could be determined directly in the complex multivitamin preparations.The differential pulse-polarographic method shows promise. References 1. 2. Brezina, M., and Zuman, P., “Polarography in Medicine, Biochemistry and Pharmacy,’, Interscience Takahashi, R., and Tachi, I., Agric. Biol. Chem., Tokyo, 1962, 26, 771. Publishers Inc., New York, 1958.June, 1975 PULSE POLAROGRAPHY TO THE ASSAY OF VITAMINS 385 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Kuta, J., Science, N . Y., 1964, 144, 1130. Mairanovskii, V. G., and Samokhvalov, G. I., Zh. Analit. Khim., 1966, 21, 210. Knobloch, E., “Physikalisch-chemische Vitaminbestimmungsmethoden, ” Akademie Verlag, Berlin , Strohecker, R. , and Henning, H. M., “Vitamin Assay-Tested Methods,” Verlag Chemie, GmbH, Meites, L., “Polarographic Techniques,” Wiley-Interscience, New York, 1965. Barker, G. D., and Gardner, A. W., 2. Alzalyt. Chem., 1960, 173, 79. Schmidt, H., and von Stackelberg, M., “Modern Polarographic Methods,” Academic Press, New York, Breyer, B., and Biegler, T., J . Electroanalyt. Chem., 1959-60, 1, 453. Schertel, M. E., and Sheppard, A. J., J . Pharm. Sci., 1971, 60, 1070. Clark, W. M., “Oxidation-Reduction Potentials of Organic Systems,” The Williams & Wilkins Christie, J . H., and Osteryoung, R. A., J . Electroanalyt. Chem., 1974, 49, 301. Breyer, B., and Biegler, T., Colln Czech. Chem. Commun., 1960, 25, 3348. Kolthoff, I. M., and Lingane, J. J., “Polarography,” Interscience Publishers Inc., New York, 1952. Moore, J. M., J . Pharm. Sci., 1969, 58, 1117. Parry, E. P., and Osteryoung, R. A., Analyt. Chcm., 1965, 37, 1634. Tachi, I., and Takahashi, R., Agric. Biol. Chem., Tokyo, 1962, 26, 238. Patriarche, G. J., and Lingane, J. J., Analytica Chim. Acta, 1970, 49, 241. 1963. Weinheim/Bergstr., 1966. 1963. Company, Baltimore, 1960. Received June 17th, 1974 Accepted December loth, 1974

ISSN:0003-2654    

DOI:10.1039/AN9750000377

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

386 Analyst, June, 1975, Vol. 100, pp. 386-399 Investigations into the Use of Gas-sensing Membrane Electrodes for the Determination of Carbon Dioxide in Power Station Waters D. Midgley Central Electricity Research Laboratories, Kelvin Avenue, Leatherhead, Survey, KT22 7SE Gas-sensing membrane electrodes have been assessed for use in the deter- mination of carbon dioxide in power station waters. The Radiometer, Type E5036, electrode has the most satisfactory performance, and the system is convenient and simple for the investigation of abnormal levels of carbon dioxide that arise at start-up or during the commissioning stages; the limit of detection (about 0.1 p g ml-1) is not low enough for the routine monitoring of normal operation levels. The electrode has been used for both discrete analysis and on-line continuous monitoring.Standard deviations of 2-0, 0.3 and 0.13 pg ml-l were obtained a t concentration levels of 23-2, 5-00 and 2.32 p g rnl-l, respectively. The response times at concentrations greater than 1 pg ml-1 were found to be within the range 3-7 min for a 10-fold increase in concentration. Of the species expected in power stations only sulphite ion, at high concentration, caused interference. Carbon dioxide dissolved in water contributes to feed-line corrosion,lP2 although the concen- tration is usually low (less than 0.1 pg ml-l) and its corrosive effect is minimised by the addition of basic substances. A variety of methods of analysis has been tried: the variation of pH in a solution saturatedwith calcium carbonate3; gas chromatography4; and, in the present British Standards method15 pH titration using a glass electrode.All of these methods involve stripping the carbon dioxide from the aqueous sample and (except for gas chromatography) re-absorption to give a more concentrated solution, which processes are slow and involve the risk of contamination because of the amount of handling needed. Only Wall's method3 is suitable for continuous analysis. The British Standards method requires a very large sample (10 1) and the apparatus is large and complex. Potentiometric sensors, being relatively simple and robust , are well suited to analysis in industrial situations, especially when continuous analysis is required. The successful appli- cation of the ammonia gas-sensing electrode to the analysis of feed-water6p7 has led to the investigation of the analogous carbon dioxide system as a method of analysis that requires a minimum of sample preparation. The gas-sensing electrode should be virtually free of interferences and should have a low limit of detection.Basis of the Method The method is based on the measurement of the change of pH in a film of sodium hydrogen carbonate solution trapped between the surface of a flat-ended glass electrode and a hydro- phobic polymer membrane that is permeable to carbon dioxide, but not to ions or to water. Electrodes working on this principle were first made by Stow, Baer, and Randall8 and Gertz and LoeschckeJS and the theory was considered more fully by Severinghaus and Bradleg.lo The use of such electrodes for the determination of carbon dioxide has so far been confined to determination at relatively high levels in blood, and only a narrow concentration range has been investigated. The reliability has not been tested in an industrial situation, nor have possible interferences been studied.Reviews of the use of carbon dioxide electrodes in clinical analysis have been written by Severinghausll and Smith and Hahn.12 Theory When the electrode is assembled, the flat sensing surface of the glass electrode is pressed against the hydrophobic polymer membrane, excluding all but a thin film of a solution of sodium hydrogen carbonate and sodium chloride. The membrane separates the interior ofMID GLE Y 387 the electrode from the sample and is permeable to carbon dioxide, so that the following equilibria will be established : CO, (sample) + CO, (membrane) + CO, (film) The film is virtually sealed from the bulk of the solution in the reservoir (containing the silver - silver chloride reference electrode) by the pressure of the membrane round the circum- ference of the glass electrode.Although a conduction path between the film and the bulk of the electrolyte is maintained, the flux of carbon dioxide from one to the other is insignificant compared with that across the membrane. The response of the electrode changes according to variations in the pH of the film, this pH being governed by the following equilibria: .. * . (1) CO, (g) + H,O -+ H,CO, . . . . K , = aHzco3/Pcoz .. .. - ‘ (2) H,CO, + H+ + HCO, . . I < a l = aH+ x aaC03-/aH2cOs HCO,- + H+ + CO,* .. .. .. - ‘ (3) Kaz = an+ x aco3a-/a~cos- Allowing for the two different, but effectively constant, ionic media in the sample and the film, we can write where KD is a constant at a given temperature. At equilibrium, the partial pressures of carbon dioxide are equal on each side of the membrane. Therefore, changes in the carbon dioxide concentration outside the membrane are reflected by equal concentration changes in the film. The relationship between the e.m.f. of the electrode and the carbon dioxide concentration in the film can be established by considering the cell below: Ag I AgCl I m M NaHCO,, c M NaCl I I m M NaHCO,, c M NaC1, v M CO, I glass electrode where m M is the concentration of sodium hydrogen carbonate, c M the concentration of sodium chloride and v is the number of moles of carbon dioxide that have diffused through the membrane per litre of solution in the film.The potential generated by this cell can be described by equation (5). where k = 2-303 RT/F. The chloride concentration is constant and, as the system is a t constant ionic strength, the chloride activity is also constant. Collecting the constant terms, we obtain Substituting successively for aH+ [equation (2)], for aH2m3 [equation (l)] and for Pcoa [equa- tion (4)] we obtain E = E’ + k log K,, KpKD - k log aHco3- + k log [CO,] (sample) (7) The dissociation constant, KaI, is so small that changes in the hydrogen carbonate ion concentration in the film that are due to changes in the carbon dioxide equilibrium across the membrane are negligible, provided that the initial sodium hydrogen carbonate concen- tration is moderately high (10-3-10-2 mol l-l), and at constant ionic strength the hydrogen carbonate ion activity can be considered constant.Collecting terms, we can re-write equation (7) as The validity of the assumption about constant hydrogen carbonate ion concentration depends on the concentration of sodium hydrogen carbonate in the bulk electrolyte and the carbon dioxide content of the sample. The theoretical relationship between the pH in the film and the carbon dioxide concentration of the sample is shown in Fig. 1 for three different concen- trations of sodium hydrogen carbonate in the internal electrolyte solution of the electrode. It can be seen that at each sodium hydrogen carbonate concentration the calibration graph is [CO,] (sample) = KD x Pco, (film) .... ‘ - (4) E = E;,, + k log aH+ - E&? + k log acl- . . .. * . (5) .. .. * * (6) E = E’ + k log a=+ .. . . E = E” + klog [CO,] (sample) .. .. - (8)388 MIDGLEY: USE OF GAS-SENSING MEMBRANE ELECTRODES FOR THE Analyst, VOZ. 100 curved at the lower concentrations of carbon dioxide in the sample, but would still be of analytical use. I I I 8 g 10 11 pH of electrolyte film Fig. 1. Predicted response for dif- ferent filling solutions. Concentrations of sodium hydrogen carbonate in filling solutions, expressed as CO, contents: A, 1000 pg ml-l of CO,; B, 100 pgml-l of CO,; and C, 10 pg ml-1 of CO,. Experimental and Results Apparatus The electrodes tested were the Radiometer, Type E5036, P,, electrode used in the Radio- meter, Type D616, glass flow cell (supplied by V.A. Howe and Co. Ltd.) and the Electronic Instruments Ltd. (E.I.L.), Model 19210, Pcoz electrode system. Some experiments were also carried out with a modified E.I.L., Model 8002-2, ammonia probe. It should be noted that although termed “electrodes,” devices of this type are complete electrochemical cells con- taining their own reference electrodes. Electrode potentials were measured with an E.I.L. Vibron, Model 33B-2, electrometer used in conjunction with a Smith’s Servoscribe chart recorder. Delivery of sample, standard and reagent solutions was effected by means of a Technicon, Model I, proportionating pump. A Grant, Model LC10, circulator was used to pump water at a constant temperature through the water-jacket of the D616 flow cell.For continuous on-line analysis of power station waters the Radiometer, Type E5036, electrode and D616 flow cell were used and a Technicon, Model I, proportionating pump delivered the sample, standard and reagent solutions. Potentials were measured with an E.I.L. 9841 electronic amplifier with a display on a Leeds and Northrup Speedomax H chart recorder. Switching between the sample and the two standard solutions was effected by means of two Hook and Tucker valves actuated by a Londex Rotaset adjustable cam timer. A two-level cabinet was used to house the apparatus in the power station. The lower compartment contained the electrode, flow cell, temperature control equipment and peristaltic pump assembly.The temperature of this compartment was maintained at 30 & 1 “C by means of a mercury contact thermometer and electronic relay in conjunction with a 300-W strip heater and an air displacement fan. Reagents All chemicals were of AnalaR grade, except the following: morpholine, hydrazine, octa- decylamine and cyclohexylamine were of laboratory-reagent grade ; and iron, copper and nickel solutions for the interference tests were prepared from BDH Chemicals Ltd. reagents for atomic-absorption spectroscopy (1000 p.p.m. of metal in approximately 1 N hydrochloric acid).June, 1975 389 Sulphuric acid, 1 N. The solution was either bought as 1 N acid from Hopkin and Williams Ltd. or prepared by dilution of concentrated acid. Carbon dioxide standard solutions.A stock solution was prepared by dissolving 1.910 (&0.001) g of sodium hydrogen carbonate in water and making up to the mark in a 1-1 cali- brated flask. 1 ml of solution = 1000 pg of carbon dioxide. DETERMINATION OF CARBON DIOXIDE IN POWER STATION WATERS Standards (100 pg ml-1 and 10 pg ml-l) were prepared by dilution of the stock solution with water taken directly from a mixed-bed de-ioniser. “Carbon dioxide h f e r solzction.” A stock solution of 0.05 mol 1-1 sodium hydrogen car- bonate solution was prepared. Portions (50ml) of the stock solution were transferred by pipette to a series of 100-ml calibrated flasks, 0.1 N sodium hydroxide solution was added from a burette and the solution made up to the mark with distilled water. Volumes of 5.0, 10.7 and 17-8ml of sodium hydroxide solution give solutions with free carbon dioxide concentrations of 0.40,O.ll and 0.006 pg ml-l, respectively, as calculated from the equilibrium constants.13J4 Method of Calibration The electrode itself responds to the partial pressure of carbon dioxide in the sample solution, i.e., to the “free,” non-ionic forms of carbon dioxide in solution.At a fixed pH in the sample, a constant proportion of the total carbon dioxide is converted to the free form. At sufficiently low pH values (less than 3.4), the proportion of free carbon dioxide is greater than 99.9 per cent. The greatest problem in calibration is that of preparing water and standard solutions that are uncontaminated by atmospheric carbon dioxide. The removal of carbon dioxide from water with sufficient efficiency has not proved to be easy.Three different techniques of standardisation have been used in the work described. (1) Sodium hydrogen carbonate solutions corresponding to 10, 100 and 1000 pg ml-l of carbon dioxide were delivered to the flow cell by means of a Technicon pump, which also pumped a stream of 1 N sulphuric acid that joined the standard solutions before they reached the flow cell. The sample and the acid were pumped in the ratio of approximately 10: 1. (2) Sodium hydrogen carbonate solutions (equivalent to either 10 or 100 pg ml-l of carbon dioxide) and 1 N sulphuric acid were pumped in ratios between approximately 10: 1 and approximately 1 : 240, thus giving a range equivalent to carbon dioxide concentrations of between 100 and 0.04 pg ml-l.It was desirable, particularly when the standard solution was being pumped much more slowly than the acid reagent, to place a glass mixing coil between the pump and the flow cell. Failure to do so could result in readings showing sudden deviations of approximately 5 mV and which were sustained for 10 min. (3) The pH of 0.025 M sodium hydrogen carbonate solutions was adjusted by addition of sodium hydroxide solution. The concentration of free carbon dioxide was calculated from the pH and the equilibrium constants, i.e., a series of carbon dioxide standards analogous to pH buffer solutions was prepared. Analytical Procedure Samples should be taken in a glass vessel (approximately 100 ml) with a tight-fitting stopper incorporating a three-way tap and a soda-lime guard tube.A suitable style of apparatus has been de~cribed.~ Connect the vessel to the sampling point by the shortest possible length of plastic or rubber tubing, flush out the sample line and then fill the vessel, which should have been previously flushed with nitrogen. The sample should be analysed as soon as possible. Pump the sample from the vessel to the flow cell, avoiding contact with the atmosphere; air should be allowed into the vessel only through the soda-lime guard tube. Simultaneously pump the 0.1 N sulphuric acid solution, allowing it to mix with the sample in a small glass mixing coil before entering the flow cell. Pump-rates of 2.0 and 0.23 ml min-l for the sample and the acid, respectively, are convenient, but the values are not critical.The Radiometer, Type E5036, electrode in the D616 flow cell is the most versatile system for general laboratory measurements. When a steady reading has been attained note the millivolt reading, V,. Replace the sample stream with a 10 pg ml-1 standard solution and note the millivolt reading, V,. Calculate the millivolt difference, A = V , - Vx.390 MIDGLEY: USE OF GAS-SENSING MEMBRANE ELECTRODES FOR THE Analyst, Vd. 100 Read the concentration of the sample from the calibration graph, or, if A lies within the linear range of the calibration graph, calculate the concentration from the equation where k is the slope of the calibration graph. Preparation of the calibration graph Prepare, by dilution of the stock sodium hydrogen carbonate solution, standard solutions containing 100, 50, 20, 10 and 5 pg ml-l of carbon dioxide.Note the steady potential and calculate A for each solution, as described under Analytical Procedure. These determinations should be repeated at least once on another day and then again, as required, until the calibration graph is defined with the desired precision. Calculate the average value of A for each solution and plot these values on the ordinate against the logarithm of the carbon dioxide concentration on the abscissa. The 10pgml-1 solution should be included as the point (log 10, 0). Range 0.1-10 pg m1-l. Using the 10 pg ml-l standard solution, dilute in situ by varying the relative flow-rates of this solution and the solution of sulphuric acid. First, using the same combination of flow-rates as for the analysis of samples, measure the potential, V,.Call the pump-rate of the standard solution fs and the pump-rate of the acid fA. Calculate the factor R, =- fs . Choose four different combinations of pump tubes for the standard solution and the acid, giving increasing dilutions to cover the range of concentration required. For each combination calculate a value of A as in Analytical Procedure and a factor, R, analogous to R,. These determinations should be repeated until the calibration graph is defined with the desired precision. Calculate the average value of A for each solution and plot these values on the ordinate against the logarithm of the concentration, 10R/Rs on the abscissa. The combination of tubes used for the analysis of samples gives the point (log 10,O).The calibration graph should be linear down to a level of 0-4-0.5 pg ml-l, but will be curved below that region. At a concentration lower than about 0.1 pgml-l the graph will have a slope that is too small to be usable. Procedure for continuous on-line analysis of power station water The sample was delivered to the cell at a rate of 0.6 ml min-l and the 1 N sulphuric acid solution at a rate of 0.23 ml min-l. The two streams joined immediately after the pump and were delivered through a length of PTFE tubing to the flow cell. Every sixth hour the sampling was interrupted automatically and two standard carbon dioxide solutions, first of 10 pg ml-l, then of 100 pg ml-l concentration, were run for 30 min each. C = 10 x antilog (A/K) pg ml-I Range 5-100 pg mZ-l.f s +f+ Preliminary Experiments Assessment of diferent electrodes A number of electrodes are produced for use in clinical analysis; those tested were chosen because they could easily be used with standard laboratory equipment, rather than being part of a specific piece of apparatus. The performance of the electrodes is summarised in Table I. It can be seen that only the Radiometer, Type E5036, was worthy of further consider at ion. TABLE I COMPARISON OF THE PERFORMANCE OF DIFFERENT ELECTRODES Response slope*/mV Time of Limit of linear Electrode per decade responsetlmin responselpg ml-1 of CO, Radiometer E5036 . . .. .. 55-59 7 0-4 E.I.L. 19210 .. .. 36-46 15 t E.T.L. 8002-2 (Klingerilon membrane) 30-33 30 t E.I.L. 8002-2 (modified) .. .. 42-46 30 160 * Difference between readings for solutions containing 10 and 100 pg ml-l of CO,. 7 Time t o go from 10 pg ml-l to 100 pg ml-1 of CO,, including 3 min wash-out time. $ The response was not linear in the range tested (0-1-1000 pg ml-l). A range of results from several assemblies is given.June, 1975 DETERMINATION OF CARBON DIOXIDE IN POWER STATION WATERS 39 1 Radiometer, Type E5036. This electrode has a 0.001 in thick PTFE membrane (Radiometer D602), a liner made of Joseph paper (absorbent tissue paper), a silver - silver chloride reference electrode and an internal filling solution containing 0.005 mol 1-1 of sodium hydrogen car- bonate and 0.02 moll-1 of sodium chloride. A stainless-steel and glass flow cell with a glass water-jacket (Radiometer, Type D616) is made for the electrode.The sensitivity of the electrode was sub-Nernstian in the range 100-0.4 pg d-l, and non- Nernstian at lower concentrations. The response time was short and the readings reproducible. The electrode was easy to assemble and could be used immediately for analysis, without requiring a stabilisation period. Radiometer also supply a silicone rubber membrane (D606), but this did not withstand exposure to the acid added to the samples. E.I.L., Model 19210, Pcoz electrode. This electrode has a latex membrane with a cuprophane liner, a calomel reference electrode and an internal filling solution containing 0.01 mol 1-1 of sodium hydrogen carbonate and 0.1 mol 1-1 of potassium chloride. A stainless-steel and glass flow cell with a glass water-jacket is available.The electrode readings drifted considerably from day to day. On first assembly the drift was very large, probably owing to the need to equilibrate the calomel reference electrode with the internal electrolyte solution. In view of the poor performance of the electrode after testing a number of assemblies, the glass electrode itself was checked in two buffer solutions (pH 4 and pH 7) against a saturated calomel reference half-cell. As a further check, the electrode was assembled with Radiometer membranes and tried again, with no better result. Conversion of E.I.L., Model 8002-2, ammonia probe. The ammonia probe has a poly- (vinyl fluoride)* membrane with no liner and a silver - silver chloride reference electrode. The ammonium chloride filling solution used for ammonia measurements was replaced by the Radiometer filling solution (0.005 moll-1 of sodium hydrogen carbonate, 0.02 mol 1-1 of sodium chloride).The normal end-cap was replaced by a flow-through version. The probe, which could be used immediately on assembly, had a sub-Nernstian response at concentrations greater than 100 pg ml-l, but this response became non-Nernstian at concentrations below that level. Response times were long. Klingerflon Threadseal PTFE tape, which has been shown to be suitable for use in a different ammonia probe,l5 was also tried in the E.I.L. 8002-2. The response was again non-Nernstian and the time of response was long. Eflect of variatiom in exfievimental conditions Choice of acidic reagent.The acidic reagent must be non-volatile, so that it does not itself pass across the membrane of the electrode, non-oxidising, to the extent that at normal temperatures (less than 40 "C) carbon dioxide is not produced from organic matter in the sample, and present in such a concentration that, after dilution and reaction with any alkaline material in the sample, it forms an ionic medium of constant pH and ionic strength. At a dilution of 5.2 volumes of sodium hydrogen carbonate solution to 1 volume of acidic reagent, 1, 0.1 and 0.01 N sulphuric acid and 0.05 M potassium hydrogen phthalate all gave the same response. Use of Aristar and AnalaR grade sulphuric acid gave the same results. AnalaR sulphuric acid (1 N) was used for all subsequent tests. When working at concentrations below 10 pg ml-l, it proved necessary to de-gas the acid solution before use.The effect of de-gassing the acid can be seen in Fig. 2, for which the method of calibration by varying the ratios of standard and reagent solutions was used. The effect is exaggerated compared with that expected for the analysis of samples, as in that instance the ratio of acid to sample is constant at 1: 10, whereas in Fig. 2, at the lowest concentration of carbon dioxide, the ratio is as high as 100: 1. I t was decided to acidify with 1 N sulphuric acid, as at this strength there is no handling hazard and only slight dilution of the sample. The efect offlowrate. A variable-speed Quickfit peristaltic pump (Type 10 PP60) was used to check the effect of flow-rate on the Radiometer electrode.Standard carbon dioxide solution (10 pg ml-l) or de-ionised water was pumped along a 5 mm i.d. tube and 1 N sul- phuric acid solution along a 4 mm i.d. tube, The millivolt readings were recorded at two different The electrode had a sub-Nernstian response and slow response times. No malfunction occurred. * This has since been replaced by a microporous PTFE membrane.392 MIDGLEY: USE OF GAS-SENSING MEMBRANE ELECTRODES FOR THE AnaZyst, VOZ. IOO power settings of the peristaltic pump motor. The sampling chamber of the D616 flow cell has a capacity of only 70 p1, and as expected the flow-rate had no effect on the response time. As the rate of flow of solution had no effect on the potential (see Table 11) the method of calibration by diluting a concentrated sample by means of a metering pump could be used without having to keep the total flow constant.TABLE I1 EFFECT OF FLOW-RATE Flow-ratelm1 min-1 E.m.f./ mV Standard CO, solution, 10 p g ml-I 20 3 2.6 - 122 100 20 16 - 122 De-ionised water . . . . .. 20 3 2.6 - 175 100 20 16 - 173 Power setting, - Sample per cent. 5 mm 1.d. tube 4 mm 1.d. tube Performance of the Radiometer, Type E5036, Electrode Laboratory tests Sensitivity. The calibration graph for the electrode is shown in Fig. 3. The calibration was sub-Nernstian over the range 100-0-4 pg ml-l but the sensitivity was only 50 mV for a 10-fold change in concentration. It was found that although the sensitivity could vary in the range 50-58 mV per decade between assemblies of the same electrode, it was constant for any one assembly.0.1 y..!.... O -200 -1 50 -1 00 E.m.f ./mV Fig. 2. Calibration of Radio- meter electrode: A, with, and B, without, de-gassing of the acid. E.m.f./mV Fig. 3. Calibration of the Radio- meter electrode a t 25 "C. The points denoted by crosses represent calibration by means of mixtures of sodium hydrogen carbonate and sodium hydroxide solutions [method (3)]. The other points were ob- tained by mixing de-gassed sulphuric acid and sodium hydrogen carb- onate solutions in different pro- portions. Precision. Four sets of measurements were made on different days. The standards were prepared by mixing sodium hydrogen carbonate and sulphuric acid solutions in different proportions. Each combination was used once per set. The results, given in Table 111, were normalised between sets by reference to the e.m.f.observed for the standard solution with the highest concentration (50pgml-l). It can be seen that the relative standard deviations at concentrations of 23.2, 5.00 and 2.32 pg ml-l are less than 10 per cent., but that those at concentrations of 6.13 and 1.98pgml-l are up to four times larger. This difference shows the effect of a large dilution factor in preparing the standards. In the first group the stock (100 and 10 pg ml-l) solutions were diluted only by factors of two andJune, 1975 DETERMINATION OF CARBON DIOXIDE IN POWER STATION WATERS 393 approximately four, whereas in the second the 100 Jug ml-l stock solution was diluted approxi- mately 16 and 50 times. TABLE I11 PRECISION OF MEASUREMENTS WITH THE RADIOMETER P,, ELECTRODE AT 25 "C CO, concentrationlpg ml-l Standard deviationlpg ml-I 23.2 2-0 6.13 1-2 5-00 0.3 2.32 0.13 1.98 0.3 Response time.At carbon dioxide concentrations above 1 pg ml-l the electrode required 3 - 7 min to reach equilibrium for a 10-fold increase in concentration and about twice as long for the change in the reverse direction, including about 2 min wash-out time for the flow cell. Below 1 pg ml-l, the response time could be extended to 30 min for a 10-fold decrease in concentration. Some typical response curves are shown in Fig. 4. 15 min I - Time - Fig. 4. Response curves for the Radiometer CO, electrode. Concentrations of carbon dioxide in sample solution: A, 7.7; B, 7 7 ; C, 7-7; D, 5.25; E, 2-47; F, 24.7; and G, 2-47 p g ml-l.Efect of other substances. Interferences can arise from two types of substance : (a) Substances that remove, or produce, carbon dioxide. The proportion of the total carbon dioxide that is present in a form that can diffuse across the membrane of the electrode depends on the pH of the sample, and therefore alkalis, which retain carbon dioxide, can be said to interfere. However, enough sulphuric acid is added to the sample to overcome any variation in its basicity. The presence of metal ions that can form insoluble carbonates could affect the analysis, but the addition of sulphuric acid will overcome this problem. Carbon dioxide could be produced by oxidation of organic matter in the sample, but in the prevailing conditions (0.1 N sulphuric acid, 25-30 "C) it is virtually impossible.( b ) Substances that can diffuse across the membrane and change the internal pH. Volatile basic compounds, such as ammonia and amines, do not interfere because they are retained in an ionic form by acidification of the sample. Ions such as sulphite and acetate may inter- fere, if their acidic forms can diffuse across the membrane in sufficient amounts. The results of the interference tests are given in Table IV. Only sulphite shows a significant interference, and only at the relatively high concentration of 100 pg ml-l, Temperature effects. Over the range tested (5-35 "C) the temperature of the sample had no effect on the readings given by the electrode in the water-jacket thermostatically maintained at 20 "C. Changing the temperature of the water-jacket shifted the e.m.f.corresponding to a given concentration of carbon dioxide and altered the sensitivity and response time of the electrode. Table V summarises the effects of temperature and the characteristic traces that occur a t constant carbon dioxide concentration when the temperature of the water jacket is increased or decreased are shown in Fig. 5. The steady e.m.f. values are shifted by approximately 1 mV "C-l and the sensitivity increases with temperature by the factor (RTIF) In 10 in the Nernst equation. The shape of the e.m.f. trace when the temperature is changed is remarkable. From the upper trace in Fig. 5 , it can be seen that the first movement is in the negative direction, but that this is394 MIDGLEY: USE OF GAS-SENSING MEMBRANE ELECTRODES FOR THE Analyst, VoZ.100 TABLE IV INTERFERENCE EFFECTS Substance* Ca2+ . . . . . . Cu2+ . , . . - . Fe3+ . . . . . . Ni%+ . . . . .. c1- . . .. . . NO3- . . . . .. PO,^- . . . . .. Ammonia . . .. Morpholine . . . . Cyclohexylamine . . Octadecylamine . . Sulphite.. . . .. Mg2+ . . . . .. Acetate . . .. .. . . . . . . . . .. .. .. .. .. .. . . .. .. Concentration of impurity/ p g ml-' 1 1 1 1 1 35 62 97 1 10 4 Saturated 100 10 100 10 Apparent concentration of carbon dioxide ( p g ml-l) due to interference a t actual concentrations of- '1 0 0.1 1 pg ml-1 10 pg ml-l 0.04 0.4 0 - 0.2 - 0.02 - 0.2 0.08 0.2 0.75 4 0 - 0 0 - - * With 95 per cent. confidence, the readings without these substances present should be within 0.1 and 1-6 p g ml-l for the 1 and 10 pg ml-l standards, respectively.reversed after a short time (approximately 3 min) and the potential becomes more positive than it was initially. After 10-15 min, the direction changes again and the potential slowly reaches a steady value. TABLE V EFFECT OF TEMPERATURE Millivolt readings a t carbon dioxide concentrations of r SensitivitylmV Temperature/"C 10 p g ml-l 100 p g ml-1 per decade 16 73 21 62 26 86 30 66 35 99 42 67 A postulated explanation is as follows, based on the fact that the "electrode" is a cell comprising three compartments: (a), the PTFE membrane - electrolyte film - glass membrane sandwich; (b), the bulk of the internal filling solution, including the silver - silver chloride reference electrode; and (c), the interior of the glass electrode, with its own filling solution 25 "C 15 "C lom"[ Change of thermostat temperature I - 10 min 15 "C Time --c Fig.5. The effect of temperature.Jzcne, 1975 DETERMINATION OF CARBON DIOXIDE IN POWER STATION WATERS 395 and silver-silver chloride electrode. Compartment (a) is in fairly good thermal contact with the sample solution and the stainless-steel parts of the flow cell and has only a small volume. Compartment (b) has a larger volume and over most of its length is separated from the stainless-steel parts of the flow cell by an air gap and the plastic wall of the electrode. Compartment (c) is enclosed by the others and separated from them by the glass of the glass electrode. In addition, the top end of the electrode is exposed to the atmosphere, although it is insulated.We can write equation (7) in an alternative form by resolving E' into its components, E'b and E'c, from compartments (b) and (c). The last three terms of equation (9) refer to compartment (a). The term k log K,, x KpKD decreases with temperature1' and as 1 > aHcos- >> [CO,], the sum of the other two terms becomes more negative as the slope, k , increases. The effects on compartments (c) and (b) will be dominated by the shift in the standard potential of the silver - silver chloride electrode, and each will oppose the other. The standard reduction potential of the silver - silver chloride electrode becomes more negative as the temperature increases16 and so the term Elc will decrease, while -E'b will increase. Thus, the first decrease in potential can be attributed to changes in compartment (a), the subsequent increase to the changes in compartment (b) and the final decrease to the changes in compartment (c).A decrease in temperature produces a trace of a different shape; the expected initial increase does not occur. This is probably because the Grant LClO circulator used to control the temperature cools only slowly at temperatures of about 25 "C and the temperature does not change quickly enough for the difference between compartments ( a ) and (b) to appear. Limit of response. The limit of response is governed theoretically by the highest pH that can be reached in the internal filling solution by removal of carbon dioxide (interference by basic gases could cause a higher pH). The highest attainable pH is, theoretically, about 11.7 for a 0.005 moll-1 sodium hydrogen carbonate filling solution, corresponding to a milli- volt reading displaced by 270 mV from that of a 75.5 pg ml-l standard solution; however, the greatest observed difference in potential was about 160 mV, obtained when 1 moll-1 sodium hydroxide solution was used (without acidification) as the sample solution.The origin of this discrepancy is difficult to establish. Trace amounts of impurities that do not interfere at higher carbon dioxide concentrations could be the cause, but in general the concentration of free impurity will be depressed by the high pH of the sodium hydroxide solution as much as the concentration of the free carbon dioxide itself. The assumption that the flux of electrolyte from the reservoir to the film is negligible compared with the flux of carbon dioxide across the membrane may not be valid at extremely low levels of carbon dioxide, in which instance a membrane material that permitted a higher flux of carbon dioxide between the sample and the film would lower the limit of detection by reducing the significance of the flux of electrolyte from the reservoir.Carbon dioxide has a greater permeabi- lity coefficient in silicone rubber than in PTFE,ll but because reproducible measurements were not obtained when silicone rubber membranes were used with the 1 moll-1 sodium hydroxide solution, the predicted effect could not be confirmed. On-line operation at power stations Precision. The reproducibility of the millivolt difference, A, between the carbon dioxide standards (10 pg ml-l and 100 pg ml-l) that were periodically introduced, was determined when the electrode was running continuously between standardisation periods on suitable power station waters.The results are given in Table VI, together with the standard deviation calculated for the 1Opg ml-l standard, taking the 100p.g ml-l standard as the reference solution. The precision obtained shows that operating in industrial conditions has no deleterious effect on the electrode. The results for a brand-new electrode at power station B have greater precision than those obtained with the old electrode, but subsequent use of the new electrode at power station A shows no difference in precision from results with the old electrode at the other locations.Comparison with the results in Table I11 shows that the standard deviations are similar whether the electrode is operating continuously or not. For comparison, results of similar tests at C.E.R.L. are shown.396 MIDGLEY: USE OF GAS-SENSING MEMBRANE ELECTRODES FOR THE Analyst, ‘Vd. 100 TABLE VI PRECISION OF ANALYTICAL RESULTS FOR 10 pg ml-l CO, SOLUTIONS Location u (A)/mV a/pg ml-l Number of results C.E.R.L. . . .. .. 1.4 0.5 23 Power station A . . .. 1.6 0.5 60 Power station B . . .. 1.4 0.5 14 Power station B . . .. 0 0 10 (new electrode) Stability of Yesponse and frequency of standardisation. The millivolt reading observed with a given standard solution drifted from day to day. The drift was assessed as follows. From each reading of the 100 pg ml-l standard that was introduced every sixth hour, the reading observed 6,12 or 24 h or 7 d previously was subtracted and the means and standard deviations of the differences corresponding to the various periods were calculated.The results are given in Table VII for electrodes operating in a variety of situations. The readings drifted in one direction for no more than four consecutive days. It can be seen that the drift over a period of 7 d is highly significant, but that, except for the new electrode at power station B, the drift within 1 d is small and not usually signifi- cantly different from zero. The gain in precision on shortening the period between the introduction of standards is small, and as one result over the 24-h period (at C.E.R.L.) is significantly different from zero while another (at power station B) represents an error of greater than 10 per cent., the best compromise between precision and instrument availability is to standardise every 12 h.The drift with the new electrode is large, although the precision of the normal standardisa- tion procedure (Table VI) is excellent. When this electrode was returned to C.E.R.L. its performance was very similar to that of the old electrode. A drift of -2.4mV represents an apparent increase in concentration of 10 per cent. TABLE VII STABILITY OF RESPONSE - 8 is the mean difference over the stated period, CT the standard deviation of a single difference, n the number of difference readings and 6 the largest drift observed in a single period. Time interval r A \ 6 h 12 h A I A \ f \ - - Location 6/mV u/mV n 6 6/mV a/mV n 6 - - - C.E.R.L. ... . -0.5 1.2 11 &-2 - Power station A . . . . -0.1 1.0 75 -3 -0.2 1.3 72 &3 Power Station €3 . . . . -2.1* 2.2 9 -5 -50* 2.3 8 +s Power station B . . . . -0.8 3.2 12 -8 -1.2 4.1 10 -8 (new electrode) Time interval 24 h 7 d r A \ h --------- -r 7 C.E.R.L. .. . . -1*3* 1.9 17 -6 -9*Of 4.6 13 - 16 Power station A . . . . -0.3 1.5 66 - 4 -2.8t 3.9 46 -9 Power station B . . . . -2.5 6.1 8 -12 - - Power station B . . . . -117 2-6 6 -12 - - Location 6/mV o/mV n 6/mV 6/mV o/mV n 6/mV - - - - (new electrode) * Significantly different from zero at the 5 per cent. level. f Significantly different from zero a t the 0.1 per cent. level. Useful Zifetime of a single assembly. An electrode operated continuously at power station A for 10 weeks without needing attention.There was no loss of sensitivity or increase in response time during this period. Reliability of ap$aratus. No defects were found in the electrodes themselves, although the end caps were not always easy to fit. The plastic union nut that secures the electrodeJune, 1975 DETERMINATION OF CARBON DIOXIDE IN POWER STATION WATERS 397 in the flow cell broke easily. This problem was overcome by using a nut that had been machined out of aluminium. Applications While the electrode has too high a limit of detection for monitoring carbon dioxide con- centrations in power station waters under normal operating conditions (i.e., for the range 0.02-0*06 pg. ml-l), it is nevertheless potentially useful for the continuous examination of carbon dioxide in abnormal conditions, where levels may rise above 0.1 pg ml-l.On start-up, the carbon dioxide content of feed-water may be higher than normal, owing to hydrolysis of carbonates and hydrogen carbonates in the evaporators combined with some loss of efficiency in de-aeration. At power station B a carbon dioxide content of 1.5 pg ml-1 was recorded for a sample of boiler water of unusually low pH (approximately 4). A similar occurrence was observed at power station A (Fig. 6), where the rise and fall in the carbon dioxide concentration of feed-water can be clearly seen as the boiler comes on and off load. When a Magnox boiler is running at reduced load, the pressure difference between the coolant gas (carbon dioxide) and the steam in the low-pressure parts of the boiler may be such that carbon dioxide will enter the steam circuit if there is a leak.Duplicate samples taken from the de-aerator outlets of two turbine sets at a nuclear power station were analysed for carbon dioxide at C.E.R.L. by using both the electrode and the low-temperature section of the Beckman, Model 915, total carbon analyser with a Beckman, Model 215A, infrared analyser, as a detector. The results are presented in Table VIII. The agreement between the results from the two methods is good, considering that the Beckman instrument was operating near its limit of detection. b b a A Off Feed load pump On load I I 1 06.00 12.00 Time - Trace obtained on a feed-water sample at power station A.Fig. 6. CO, contents of standard solutions: a, 10 pg &-I; b, 100 pg rnl-l. At a cross-flow cooling tower test rig the ability of the electrode to measure free and total carbon dioxide was utilised. At the inlet to the tower, the total carbon dioxide content of the cooling water was found to be 2Opgmk1, whereas the free carbon dioxide content (determined by making measurements without adding acid to the sample) was only 2.3 pg ml-1. TABLE VIII COMPARATIVE ANALYSIS OF SAMPLES FROM A NUCLEAR POWER STATION Carbon dioxide content (pg ml-l) obtained by use of the- 8 8 9 9.6 4 3 5.5 6.6 A I 5 Sample Radiometer electrode Beckman carbon analyser A, A2 Bl B2398 MIDGLEY: USE OF GAS-SENSING MEMBRANE ELECTRODES FOR THE Analyst, Vol.100 Further Investigation of the Radiometer, Type E5036, Electrode As the observed limit of response of the Radiometer electrode was too high for the monitor- ing of feed-water and condensed steam under normal operating conditions, i.e., where carbon dioxide concentrations in the range 0.02-0*05 pg ml-l are expected, various features of the electrode were investigated to see if the performance could be improved.The glass and reference electrodes cannot be changed and should, in any event, be reliable; the components studied were, therefore, the membrane and the internal filling solution. The membrane Besides the standard arrangement of a PTFE membrane with a tissue-paper liner, three other membranes were tried: PTFE without the liner, a Radiometer D606 silicone rubber membrane, and Klingerflon Threadseal PTFE tape.Eflect of the tissue-pafier liner. Although gas-sensing ammonia electrodes operate without a liner between the membrane and the glass ele~trode,~J5 the Radiometer carbon dioxide electrode has a layer of tissue-paper in order to stabilise the film between them.11 When the liner was removed, the electrode could operate equally well, but the consistency of performance over a number of assemblies decreased. It is concluded that the difference between the performances of the ammonia and carbon dioxide electrodes arises because the faint surface pattern on the membranes used in the ammonia electrode forms small pockets of solution analogous to the solution trapped between the fibres of the tissue-paper. Silicone rubber membranes. These membranes are supplied as an alternative to the PTFE membranes and a shorter response time is claimed for them.The sulphuric acid used to liberate the carbon dioxide damaged the membrane, but the lifetime of a membrane was at least 1 week when 0-05 mol 1-1 potassium hydrogen phthalate solution was used instead. In the latter instance, the response was slower than with the PTFE membrane and there was no advantage in limit of response. The silicone rubber membrane was easily damaged during assembly. KlingerJlon PTFE membranes. These membranes have been used in an ammonia electrode,l5 but broke down on exposure to sulphuric acid. As the response time was relatively long (about 30 min) during the lifetime of the membrane, it was not considered worthwhile to use an alternative acidic reagent such as potassium hydrogen phthalate.The internal $&ng sohtion In Fig. 1 it can be seen that the theoretical calibration graph is less curved at low concen- trations of carbon dioxide in the sample when the internal filling solution is more dilute. A range of concentrations of internal filling solution was tested, and the results are given in Table IX. TABLE IX EFFECT OF SODIUM HYDROGEN CARBONATE CONCENTRATION IN FILLING SOLUTION Concentration of NaHCO,/pg ml-l of CO, A*/mV Response timetlmin 220 54 10 100 64 20 30 60 6 6 37 > 30 0.6 35 > 60 * Millivolt difference between 10 and 100 pg ml-l CO, solutions. t Time taken to reach equilibrium when the CO, content is changed from 100 to 10 pg ml-1. The concentration of sodium hydrogen carbonate in the filling solution has a different effect from that predicted.At low concentrations of sodium hydrogen carbonate the sensi- tivity falls well below the theoretical value and the response time becomes very long. The concentration has relatively little effect for filling solutions with carbon dioxide contents between 100 and 1000 pg ml-l, although the response time lengthens as the concentration in the filling solution decreases. There is no advantage in reducing the hydrogen carbonate ion concentration.Jztne, 1975 DETERMINATION OF CARBON DIOXIDE IN POWER STATION WATERS 399 Conclusions Of those commercially available electrodes that have been assessed, the Radiometer, Type E5036, electrode has the best over-all characteristics. Although the limit of detection of the electrode is too high for the routine analysis of feed-water or condensed steam, it is of potential use for the investigation of the higher carbon dioxide concentrations observed during start-up or commissioning. The work described in this paper was performed at the Central Electricity Research Laboratories and is published by permission of the Central Electricity Generating Board. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. References Uhlig, H. H., Editor, “The Corrosion Handbook,” John Wiley, New York; Chapman and Hall, Speller, F. N., “Corrosion, Causes and Prevention,” McGraw-Hill, New York, 1961, p. 448. Wall, K. H., Analyst, 1966, 91, 795. Gaunt, H., and Shanks, C., Chemy Ind., 1964, 651. British Standard 2690 : Part 5 : 1967. Midgley, D., and Torrance, K., Analyst, 1972, 97, 626. Midgley, D., and Torrance, K., Analyst, 1973, 98, 217. Stow, R. W., Baer, R. F., and Randall, B. F., Arch. Phys. Med. Rehabil., 1957, 38, 646. Gertz, K. H., and Loeschcke, H. H., Naturwissenschaften, 1958, 45, 160. Severinghaus, J . W., and Bradley, A. F., J . Appl. Physiol., 1958, 13, 515. Severinghaus, J. W., Ann. N.Y. Acad. Sci., 1968, 148, 115. Smith, A. C., and Hahn, C . E. W., BY. J . Anaeslh., 1969, 41, 731. Yuan-Hui Li and Tien-Fung Tsui, J . Geophys. Res., 1971, 76, 4203. Martell, A. E., and SillCn, L. G., “Stability Constants,” Special Publication No. 17, Chemical Hawker, B. W., Midgley, D., and Torrance, K., Lab. Pract., 1973, 22, 724. Ives, D. J. G., and Janz, G. J., “Reference Electrodes,” Academic Press, New York, 1961, p. 189. Received Decembey 17th, 1974 Accepted February 3rd, 1976 London, 1948, pp. 62, 127 and 538. Society, London, 1964.

ISSN:0003-2654    

DOI:10.1039/AN9750000386

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

400 Analyst, June, 1975, Vol. 100, p p . 400-404 Modification of the lodimetric Titration Method for the Determination of Bromide and its Application to Mixed Domestic = Industrial Waste Effluents Daniel F. Bender Methods Development and Quality Assurance Research Laboratory, Environmental Protection Agency, National Environmental Research Center, Cincinnati, Ohio 45268, U.S.A . The iodimetric titration method for the determination of bromide involves the observation of various colour changes, making the method unsuitable for use with samples that are highly coloured. A modification is described, which extends the usefulness of the method to highly coloured samples, such as sewage and industrial waste effluents. The obvious substitution of a pH meter for indicators was made and the use of standardised concentrations and amounts of reagents added was incorporated into the method.The pro- cedural steps were studied in order to optimise the sensitivity while mini- mising interfering side reactions. The modified method was applied to a mixed domestic - industrial waste effluent spiked with bromide and iodide. The experimental details and statistical parameters for samples spiked at different levels are presented. Ingestion of bromide ion is known to produce physical effects such as bromoderma (skin rashes) and increased spinal fluid pressure and neurological effects ranging from depression and confusion, through hallucination and psychosis, to mental deterioration and coma.1J The Methods Development and Quality Assurance Research Laboratory of the National Environmental Research Center, Cincinnati, Ohio, is currently developing and evaluating methods of analysis and studying their applicability to the determination of pollutants in industrial waste effluents.The U.S. Environmental Protection Agency guidelines establishing test procedures for the analysis of water pollutants3 specify ASTM Method D1246-C for bromides.4 The method involves two titrations. First, the iodide present in the sample is determined by oxidising the iodide to iodate with bromine water and destroying the excess of bromine with sodium formate solution. The amounts of these reagents are not specified in the ASTM pr~cedure,~ but are determined by visual observation of the subtle yellow bromine colour. The iodate is then made to react with potassium iodide to liberate iodine, which is then titrated with sodium thiosulphate, or an equivalent reductant, using starch as indicator.Next, the iodide plus bromide is determined by oxidising these ions to iodate plus bromate with hypochlorite, destroying the excess of hypochlorite with sodium formate solution and titrating the iodine released when potassium iodide is added as described above. The bromide concentration is then calculated by difference. This paper describes the investigation of the ASTM method for bromide when applied to a mixed domestic - industrial waste effluent. Some modifications were necessary in order to compensate for the colour that is often found in such effluents, which masks the indicator changes as well as observation of the subtle yellow bromine colour called for in the original method.The final, iodimetric titration, colour change must be observable, of course. Adjustments in the reagent concentrations were necessary in order to compensate for the loss of reagents, caused by reaction with interferents that consume some reagent without other- wise affecting the final titration. Experimental Reagents Reagents were prepared as directed by ASTh14 except for the following. Bromine water. Bromine (0.2 ml) was added to 500 ml of distilled water. The mixture was stirred on a magnetic stirrer with a PTFE-coated follower for several hours, or until the bromine dissolved. Methyl red indicator solution. Potassium hypochlorite solution. This solution was replaced by calcium hypochlorite The mixture was then stored in a coloured, glass-stoppered bottle.Use of this solution was avoided by using a pH meter. solution, which is a commercially available material.BENDER 401 Calcium hypochlorite solution. An amount (35 g) of calcium hypochlorite was placed in a 1000-ml calibrated flask. Approximately 800ml of distilled water were added and the suspension was stirred on a magnetic stirrer for approximately 30 min. The contents of the flask were then diluted to the mark, stirred for a further 5 min, and filtered. The resulting solution was stored in a coloured, glass-stoppered flask. Sodium thiosulphate solution. Commercially available phenylarsine oxide solution (0.0375 N), standardised with a biiodate primary ~tandard,~ was used in place of sodium thiosulphate solution ; either titrant may be used.Commercially available amylose indicator (Mallinckrodt) was used. Starch indicator solution. Procedure Pre-treatment A visible excess of calcium oxide was added to a 400-ml sample in order to eliminate the effects of iron, manganese and organic matter, as has been de~cribed.~ The solution was vigorously stirred for approximately 15min and filtered, the first 75-ml portion of filtrate being discarded, In practice, samples up to 3 1 were pre-treated in this manner. Iodide determination A 100-ml volume of pre-treated sample containing not more than 20 mg 1-1 of iodide, or a fraction thereof diluted to 100m1, was placed in a 150-ml beaker. By use of a magnetic stirrer and a pH electrode, sulphuric acid (1 + 4) was added dropwise until a pH of approxi- mately 7 (or slightly less) was obtained.The solution was then transferred to a 250-ml, wide- mouthed conical flask and the beaker washed with distilled water; the washings were added to the conical flask. Sodium acetate solution (15 ml) and 5 ml of acetic acid (1 + 8) were added and the flask was swirled to mix the solutions. Forty millilitres of bromine water were added and the solutions mixed and allowed to stand for 5 min. Sodium formate solution (2 ml) was added and again the flask was swirled to mix the solutions. The sides were washed down and the flask was allowed to stand for an additional 5 min, a gentle stream of nitrogen being blown in to displace any bromine fumes, Because there was no indication of the presence of iron upon cooling (formation of a precipitate),4 the addition of potassium fluoride was not necessary.The solution was then titrated as described below under Titration. Bromide plus iodide determination A 100-ml volume of pre-treated sample containing not more than 20 mg 1-1 of bromide plus iodide, or a fraction thereof diluted to 100 ml, was placed in a 150-ml beaker and 5 g of sodium chloride were added. Dissolution of the sodium chloride was effected by using a magnetic stirrer. When the sodium chloride had dissolved, the pH was adjusted to approxi- mately 7 by adding hydrochloric acid (1 + 4) dropwise; the solution was then transferred, with washing, into a wide-mouthed conical flask. Calcium hypochlorite solution (20 ml), a further 1 ml of hydrochloric acid (1 + 4) and approximately 0.2 g of calcium carbonate were added.The solution was heated to boiling on a hot-plate and boiling maintained for an additional 8 min. The flask was removed from the hot-plate, 4 ml of sodium formate solution were added, then the sides of the flask were washed down with distilled water; it was returned to the hot-plate and the solution was allowed to boil for an additional 8 min. The sides of the flask were occasionally washed down when evaporation produced a residue on the sides. The flask was then allowed to cool. As, upon cooling the solution, there was no indi- cation of the presence of iron (formation of a pre~ipitate),~ the addition of potassium fluoride was not necessary. Several drops of sodium molybdate solution were added and the sample was titrated as described below under Titration.Titration swirling the flask. aside in the dark for 5 min. oxide solution, the amylose indicator being added as the end-point was approached. blue colour usually reappeared on standing, but this was disregarded. Approximately 1 g of potassium iodide was added to the flask and dissolved by gently Ten millilitres of sulphuric acid (1 + 4) were added and the flask was set The solution was then titrated with the 0.0375 N phenylarsine The402 BENDER 1 DETERMINATION OF BROMIDE I N Analyst, Vol. 100 Calculations The calculations were performed as described in the ASTM standard method4 with the exception that the volume of titrant used by a distilled water blank must be subtracted from the volume of titrant used in the bromide plus iodide determination.The equation thus becomes: H B r (mg 1-l) = 13 320 where A is the volume (ml) of phenylarsine oxide needed to titrate the sample for bromide plus iodide; B, the volume (ml) of phenylarsine oxide needed to titrate a distilled water blank that has been treated with the same reagents as the sample for bromide plus iodide; C, the normality of the phenylarsine oxide solution used to titrate the sample and blank for bromide plus iodide; D, the volume of samples taken (100 ml or a fraction thereof) to titrate the bromide plus iodide; E , the volume (ml) of phenylarsine oxide needed to titrate the iodide in the sample; F , the volume (ml) of phenylarsine oxide needed to titrate a distilled water blank that has been treated with the same reagents as the samples in which the iodide was titrated; G, the normality of the phenylarsine oxide used to titrate the iodide in the sample and blank; and H , the volume of samples taken (100 ml or a fraction thereof) for the iodide titration.Results and Discussion Pre-treatment A synthetic sample containing approximately 10 mg 1-1 of iron(III), 10 mg 1-1 of man- ganese(II), 25 mg 1-1 of bromide and 25 mg 1-1 of iodide ions was pre-treated and then analysed to determine bromide and iodide. The formation of a precipitate upon cooling, which would have indicated the continued presence of iron, was not observed. The pre-treatment was therefore shown to be effective, as stated by ASTM,4 and was not studied further. Iodide determination The use of a pH meter was substituted for the acid - base indicator because industrial waste emuents are often highly coloured.In addition, the presence of colour in the sample can make the various yellow bromine colours (which serve as an indication of how much bromine to add and how much sodium formate is required to quench the bromine) impossible to observe. Therefore, specified volumes of bromine water and sodium formate solution were substituted. The amounts of these reagents required for a stoicheiometric reaction with 25 mg 1-1 of bromide or iodide were calculated and an excess of bromine was added. It was found that if the bromine water was too concentrated, an unreproducible blank resulted. It was also found that when a sewage sample was used, an interference consumed some of the bromine so that a further slight excess had to be added; the amount of bromine water chosen for the method was ascertained by trial and error.Bromide @us iodide determination The required amounts of calcium hypochlorite and sodium formate solutions were deter- mined by using considerations similar to those for the iodide determination. There is a definite titratable blank, therefore a distilled water blank determination must be carried out with each set of samples. Titration The starch - iodine titration is commonly used for dissolved oxygen determinations.5 Whichever titrant or indicator is most convenient can be substituted in the method; phenyl- arsine oxide and amylose indicator were used in this study because they were readily available.The procedure for preparing sodium thiosulphate and starch indicator is described by ASTM.4 For a routine analytical application, it is suggested that 0.0075 N phenylarsine oxide should be used. This concentration would provide a larger volume of titrant in the concentration range of bromide and iodide ions suggested.June, 1975 MIXED DOMESTIC - INDUSTRIAL WASTE EFFLUENTS 403 A nalytical parameters Tables I and I1 show the standard deviation and accuracy (percentage recovery) for mixed domestic - industrial waste samples from the Cincinnati, Ohio, Sewage Treatment Plant with a wide range of iodide and bromide concentrations. In each instance seven replicates were run. The results for the iodide determination are included because the results had to be obtained in order to complete the bromide calculation.TABLE I STATISTICAL DATA ON BROMIDE DETERMINATION USING SPIKED AND UNSPIKED SAMPLES OF MIXED DOMESTIC AND INDUSTRIAL WASTE EFFLUENTS Theoretical total Found total Standard Coefficient of concentration/ concentration/ deviation variation,? Recovery,: per cent. per cent. Spike*/mg 1-l mg 1-1 mg 1-1 (4 0 - 0.3 0.13 43 - 2.5 2-8 2.7 0.37 14 96 5.0 5.3 4.4 0.38 9 83 10.0 10.3 10.0 0.44 4 97 20.0 20.3 20.1 0.42 2 99 * The amount of spike that is listed was added to the sample, which already contained some halides. t Coefficient of variation S - -- found x 100. x 100. found theoretical Recovery, per cent, = The sample used for determining the statistical parameters had a natural concentration of 0.3 & 0.13 mg 1-1 of bromide and 1.6 0.23 mg 1-1 of iodide, as is shown in Tables I and XI.The standard deviations and coefficients of variation for the various levels of spiked sample are shown. As expected, the method becomes more precise as the concentration of the spike increases. The percentage recovery, which is an indication of the accuracy, was generally between 92 and 99 per cent.; however, occasionally levels as low as 80 per cent. were obtained. Although running samples spiked at these levels again produced results at a level above 90 per cent., the values shown in the tables were recorded in order to illustrate that lower yields are occasionally obtained, that spiking is necessary to judge the nature of the sample and also because particular results for yield were obtained from the same spiked sample as those for standard deviation and coefficient of variation.TABLE I1 STATISTICAL DATA ON IODIDE DETERMINATION USING SPIKED AND UNSPIKED SAMPLES OF MIXED DOMESTIC AND INDUSTRIAL WASTE EFFLUENTS Theoretical total Found total concentration/ concentration/ Spike*/mg 1-1 mg 1-1 rng 1-1 0 - 1.6 2.5 4-1 3.3 6.0 6.6 6.4 10.0 11-6 11.2 20.0 21.6 19.8 Standard deviation 0.23 0.17 0.10 0.06 0.50 (4 Coefficient of variation, t Recovery, 4 per cent. per cent. 14 - 5 80 2 97 0.5 97 3 92 * The amount of spike that is listed was added to the sample, which already contained some halides. t Coefficient of variation = S x 100. found found theoretical $ Recovery, per cent. = x 100. The method described by ASTM4 is capable only of determining levels of bromide above 50 mg 1-1 because it involves observing subtle changes in the yellow colour of bromine.It permits the use of highly variable amounts of bromine and sodium formate, the bromine quencher, thereby ensuring sufficient reagent concentration without excess. Excess reagent404 BENDER concentration produces a high blank. As these colour changes are not observable in a coloured sample, the modification described in this paper involves the use of standard concentrations of the reagents in order to ensure that there is a sufficient amount of reagent, yet not so great an excess as to produce a high blank. This modification produces a limitation on the upper limit of the range. The range 2-20 mg 1-1 was determined by calculating the amounts of the reagents (bromine water and calcium hypochlorite solution) that were necessary to determine a maximum of 25 mg l-l, then an excess was added to compensate for reagent consumed by interferents.If a value greater than 20 mg 1-1 was observed the sample was diluted and run again to make sure that the value was not actually above the upper limit. Distilled water was deliberately spiked at the 50 mg 1-1 level and analysed using reagent concentrations calculated for a 25 mg 1-1 maximum. A result of 25 mg 1-1 was obtained, as anticipated. The lower range limit was chosen because of the observation that the bromide - iodide deter- mination gave a blank of approximately 0.5 ml of titrant and the conviction that approxi- mately 1.0 ml of titrant is necessary for the result to be significantly different from that of the blank. Conclusion The bromide method as presented is applicable to highly coloured samples in the range 2-20 mg 1-1 of bromide when the iodide concentration is also in that range or lower. Samples that are more concentrated in either component must be diluted. Spiked samples should be run with each sample type because of the possibility of interferences that could consume the reagent and therefore affect the results. The author is grateful for many helpful suggestions from Robert L. Booth, Morris E. Gales, jun., Robert F. Thomas and Larry B. Lobring of the Methods Development and Quality Assurance Research Laboratory. References 1. 2. 3. 4. 6. Sax, N. I., “Dangerous Properties of Industrial Materials,” Third Edition, Reinhold Book Corpora- Gleason, M. N., Gosselin, R. E., Hodge, H. C., and Smith, R. P., “Clinical Toxicology of Commercial Federal Register, Washington, D.C., October 16th, 1973, 38, 28758-28760. “ASTM Standards, Part 23, Water; Atmospheric Analysis,” American Society for Testing and “Methods for Chemical Analysis of Water and Wastes,” Environmental Protection Agency, National Received October Sth, 1974 Accepted January 2nd, 1976 tion, New York, 1968. Products,” Third Edition, The Williams & Wilkins Co., Baltimore, Maryland, 1969. Materials, Philadelphia, 1973, p. 331, Method D1246-C. Environmental Research Center, Cincinnati, Ohio, 1974, pp. 51-55.

ISSN:0003-2654    

DOI:10.1039/AN9750000400

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, June, 1975, Vol. 100, PP. 405-407 405 A Semi-automated Procedure for the Determination of Iodine in Plant Tissue and Soil Extracts H. van VIiet, W. D. Basson and R. G. Bohmer Department of Inorganic and Analytical Chemistry, University of Pretoria, Pretoria 0002, South Africa A semi-automated method is described for the determination of total iodine in plant tissue and soil extracts. The automated, colorimetric determination is based on the catalytic action of iodine on the oxidation of arsenic(II1) by cerium( IV) . The reproducibility and accuracy of the proposed procedures are reported for the materials studied. The investigation and the routine determination of trace amounts of iodine in plants and soils call for a rapid, accurate and sensitive method of analysis. Studies of iodine in soils and plants, however, are seriously hampered by methods of analysis that are both time- consuming and tedious.As a large number of soil and plant samples are analysed as a matter of routine in this laboratory by use of the Technicon AutoAnalyzer in order to determine several elements, a semi-automated procedure was developed for the determination of iodine in these materials. Experimental Reagents De-ionised water and reagents of analytical-reagent grade were used (except when otherwise stated) throughout this investigation. Measurements of volumes for iodine calibration stan- dards were made with an Agla micrometer syringe. Sodium arsenite solution. Dissolve 10 g of arsenic(II1) oxide and 10 g of sodium hydroxide in 150 cm3 of water. Add a further 300 cm3 of water, 10 cm3 of concentrated sulphuric acid and 15 g of sodium chloride.Allow the mixture to cool and dilute it to 500 cm3 with water. Perchloric acid - nitric acid digestion mixture. Mix 400 cm3 of perchloric acid (70-72 per cent.) and 200 cm3 of nitric acid (55 per cent.). Standard iodide solution, 100 9.p.m. Dissolve 130.8 mg of potassium iodide in 1 dm3 of water. Cerium(1V) suZphate solution. Dissolve 6 g of cerium(1V) sulphate (technical grade) in 500 cm3 of 3.6 N sulphuric acid. Allow the solution to stand overnight and then filter it. Sulphuric acid, 20 per cent. VlV. Sodium hydroxide solution, 2 N. Dissolve 80 g of sodium hydroxide in 1 dm3 of water. Sodium hydroxide, dilute solution. Dilute 75 cm3 of 2 N sodium hydroxide solution to 1 dm3 with water.Apparatus A photocell-type colorimeter equipped with a 15-mm tubular flow cell, a heating bath, a 26-m time delay coil and a Mark I1 proportionating pump were used. Technicon AutoAnalyzer instruments were used as shown in Fig. 1. Separation of Iodine The procedure that was finally adopted for the digestion of plant tissue is a modification of a procedure described for the determination of iodine in plasma and tissue.' This method offers the advantage of greater simplicity and more rapid sample preparation. The procedure used for the separation of iodine from soils is essentially the technique recommended by Whitehead. Various methods for the separation of iodine from a matrix have been investigated.406 4.8 1 ; I D M C H DMC] TDC 65 "C VAN VLIET, BASSON AND BOHMER : PROCEDURE FOR THE Analyst, VoZ.100 Sulphuric acid 20% V/V 1 I Sample 2.5 2.5 1.2 1.6 1.2 Sodium arsenite Air Cerium( IV) solution Colorimeter -- I Fig. 1. Flow diagram for the determination of iodine in soil and plant digests. DMC = double mixing coil, TDC = time delay coil. Procedure Digestion of plant tissae Weigh 0.10 g of dried, finely ground, plant tissue into a 125 x 20 mm test-tube (with a calibration at 15 cm3). Add 3 cms of 20 per cent. V/V sulphuric acid, 1 cms of perchloric acid - nitric acid mixture and 0.8 cm3 of water, shaking the tube after each addition to prevent material adhering to the walls of the tube. Allow the sample to stand for 10 min and place the tube in a pre-heated oven at 265 & 5 "C until the digestion mixture is clear (usually 1 to 1.5 h).Remove it from the oven, allow to cool and dilute it to the 15-cm3 mark with water. Treat the calibration standards and blank solutions in the same manner. Extraction of iodine from soil Boil under reflux 1.0 g of air-dried soil with 15 cm3 of 2 N sodium hydroxide solution for 45 min. Allow the mixture to cool and transfer it to a 50-cm3 centrifuge tube, centrifuging for 5min at 2500r.p.m. Transfer the supernatant solution to a ZOO-cm3 calibrated flask. Add 40 cm3 of water to the soil residue, shake it well and again centrifuge. Combine the two supernatant solutions in the 200-cm3 calibrated flask and dilute to volume with water. Pipette 1 cm3 of the soil extract into a 125 x 20 mm test-tube, add 3 cm3 of 20 per cent.sulphuric acid and 1 cm3 of perchloric acid - nitric acid mixture. Shake the tube well and place it in a pre-heated oven at 265 & 5 "C until the mixture is clear (usually 1.5 h). Finally, dilute the solution with water to the 15-cm3 mark. Treat the calibration standards and blank solutions in the same way but substitute dilute sodium hydroxide solution for the 1 cm3 of soil extract. Each group (the number of samples depending on the size of the oven) of soil or plant samples must include a set of calibration standards and a blank solution. The 15-cm3 volume solutions from both sample sources were finally transferred to the sampler unit of the AutoAnalyzer and the iodine contents determined colorimetrically by the catalytic action of iodine on the oxidation of arsenic(II1) ions from the sodium arsenite solution with cerium(1V) ions from the cerium(1V) sulphate solution. Results and Discussion Reproducibility of Results Table I.and plant tissue, as indicated by the values obtained for the standard deviations. Recovery of Iodine Added to Samples To obtain an indication of the accuracy of the proposed procedures, two plant, soil and animal tissue (liver) samples to which known amounts of iodine (as iodide) had been added, were analysed. The average original iodine concentrations were determined by duplicate The average results of ten analyses of each of the different types of sample are given in A satisfactory precision was obtained with the proposed procedures for both soilsJune, 1975 DETERMINATION OF IODINE IN PLANT TISSUE AND SOIL EXTRACTS TABLE I REPRODUCIBILITY OF THE SEMI-AUTOMATED DETERMINATION OF IODINE IN SOIL AND PLANTS Coefficient Standard of variation, Material Concentration deviation (*) per cent. Plant 8.7 pg per 100 g 0.55 pg per 100 g 6.3 2 1.3 0.96 4.5 45.5 1-50 3.3 7.1 0.14 1-9 1.4 0.08 6- 1 soil 8-6 p.p.m.0.18 p.p.m. 2.1 407 analysis of each sample. Known amounts of iodine were then added (depending on the type of sample) and the average of duplicate determinations was obtained. The values given in Table I1 indicate that acceptable recoveries of iodine were obtained with the proposed procedures. TABLE I1 RECO.VERY OF ADDED IODINE FOR PLANT, SOIL AND TISSUE SAMPLES Iodine recovered, Material Iodine added Iodine found per cent. Plant 15.0 pg per 100 g 24.3 pg per 100 g 102 30.0 15.6 98 Soil 3.0 p.p.m. 8.6 p.p.m. 5.0 1.4 98 97 Tissue 15.0 pg per 100 g 9-0 pg per 100 g 96 (liver) 30.0 14.4 97 References 1. 2. Malvano, R., Buzzigoli, G., Scarlattini, M., Cenderelli, G., Gandolfi, C . , and Grosso, P., Analytica Whitehead, D. C., J . Soil Sci., 1973, 24, 260. Received December 2nd, 1974 Accepted December 19th, 1974 Ckim. Act@, 1972, 61, 201.

ISSN:0003-2654    

DOI:10.1039/AN9750000405

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

408 Analyst, June, 1975, Vol. 100, p p . 408-414 Measurement of Low Phosphorus Concentrations in Nutrient Solutions Containing Silicon S. Jintakanon, G. L. Kewen, D. G. Edwards and C. J. Asher Department of Agriculture, University of Queensland, St. Lucia, Queensland 4067, A ustralia A method has been developed that permits the determination of phosphorus concentrations down to 0.04 p~ in nutrient solutions that contain silicon. The problem of silicon interference is eliminated by use of ethyl acetate as a selective extractant for the yellow molybdophosphoric acid prior to its reduction to heteropoly blue (molybdenum blue). Most published methods for the determination of phosphorus are based on the reaction of orthophosphate with an acidic ammonium molybdate solution to form either the yellow heteropolymolybdophosphate complex or the reduced heteropoly blue (molybdenum blue) ~omplex.l-~ However, as corresponding complexes with closely similar absorption spectra are formed with a number of other elements, including silicon, arsenic and germanium,4 the colorimetric determination of phosphate is difficult to accomplish in the presence of these elements .5 Silicon is a ubiquitous constituent of natural waters and is often present as a contaminant in laboratory supplies of distilled or de-ionised water.In addition, silicon is sometimes deliberately added to plant culture solutions in order to suppress manganese toxicity.6 Various methods have been described for measuring phosphate concentrations from 8 p M 7 to 320 p ~ 8 in the presence of silicon concentrations of up to 1800 p ~ .~ These methods have involved partial dehydration and filtering of samples in order to remove silicon prior to determinati~n,~ careful control of solution acidity7 or the addition of tartaric acidlo in order to prevent the formation of the molybdosilicate complex, the use of differential kinetics of complex formation,8 or separation of the complexes by solvent extraction after formation.ll For work at very low concentrations of phosphorus, solvent-extraction methods offer the possibility of increasing the sensitivity by extracting the molybdophosphate complex from a relatively large volume of solution into a much smaller volume of organic solvent. Solvent extraction of the molybdophosphate complex has been used successfully in the analysis of plant nutrient culture solutions that are 0.04-24 p~ in phosphorus but no added ~ i 1 i c o n .l ~ ~ ~ ~ However, experience at this laboratory has shown that the method, based on extraction with butan-1-01, gives unsatisfactory results with solutions that contain appreci- able amounts of silicon. This paper describes an improved method for phosphate determina- tion in which interference from silicon is overcome by the use of ethyl acetate as an extractant. Experimental Equipment the absorbance of the molybdophosphate complex in all solutions. Reagents All solutions were prepared from analytical-reagent grade chemicals and were stored in polythene bottles so as to minimise the risk of contamination by si1ic0n.l~ Standard phos~horz~s solution.Dissolve 0.136 g of potassium dihydrogen orthophosphate in double de-ionised water and dilute to 1 1. This solution contains 1 pmol ml-l of phosphorus. Phosphorus solution, 50 p ~ . Dilute 50 ml of the standard phosphorus solution to 1 1 with double de-ionised water. Phosphorus solution, 8 p ~ . Dilute 8 ml of the standard phosphorus solution to 1 1 with double de-ionised water. Standard silicon solution. Dissolve 0.212 g of sodium metasilicate in double de-ionised water and dilute to 1 1. Molybdate solution. Dissolve 50 g of sodium molybdate in double de-ionised water and dilute to 1 1 in order to obtain a 5 per cent. molybdate solution. A Beckman, Model DB-G, grating spectrophotometer with 4-cm cells was used to determine This solution contains 1 pmol ml-l of silicon.JINTAKANON, KERVEN, EDWARDS AND ASHER 409 Hydrochloric acid, 10 N.Mix 825ml of concentrated hydrochloric acid with 175ml of double de-ionised water. Sulphuric acid, 9 N. Mix 249.3 ml of concentrated sulphuric acid with double de-ionised water and dilute the mixture to 1 1. T i n ( I 1 ) chloride solution. Dissolve 4.0 g of tin(I1) chloride in 10 ml of concentrated hydrochloric acid. Prepare a fresh solution every week. Reducing agent. Dissolve 1.3 g of ascorbic acid in 20 ml of double de-ionised water, add 1.3 ml of tin(I1) chloride solution and 15 ml of 9 N sulphuric acid. Dilute the solution to 100 ml with double de-ionised water (cf., Pakalns and McAllister15). Ethyl acetate, boiling range 76.5-78.5 "C. Procedures A . Determination of phosphate in samples containing more than 6.4 pmol I-1 of phosphorus de-ionised water and then proceed as in B below.B. Determination of phosphate in samples containing 0.5-6.4 pmol I-1 of phosphorus Transfer 50 ml of the sample into a 100-ml separating funnel, add 5 ml of 10 N hydrochloric acid and mix the solutions thoroughly. Add 6 ml of the molybdate solution, again mix and allow the mixture to stand for 3 min. Next add 13 ml of ethyl acetate, shake the mixture vigorously for 1 min, allow it to stand for 5 min and run off the aqueous layer. Add 3 ml of reducing agent to the organic phase, shake the funnel for 30 s and allow it to stand for 5 min. Then run off the aqueous layer completely, transfer the remaining organic phase into a dry 25-ml calibrated flask, wash the funnel with ethyl acetate and drain the washings into the calibrated flask.Dilute the contents of the flask to volume with ethyl acetate and mix well. Read the absorbance at 720 nm, using 4-cm cells, between 30 and 60 min after reduction. C. Determination of phosphate in samples containing less than 0.5 pmol I-I of phosphorus Transfer 200 ml of the sample to a 250-ml separating funnel and add 20 ml of 10 N hydro- chloric acid, mixing the solutions after addition, Add 24 ml of the molybdate solution, mix, and allow the mixture to stand for 3 min. Then add 25 ml of ethyl acetate, and proceed as in B. In both procedures B and C above, the acid concentration in the final solution, before extraction with ethyl acetate, is 0432 N. All samples of nutrient culture solution were stored in polythene bottles at 3 4 "C while awaiting phosphate determination.The composition of the basal culture solution was as follows : nitrate-nitrogen, 1000 p~ ; calcium, 625 p~ ; sulphate-sulphur, 350 p ~ ; potassium, 250 ~ U M ; magnesium, 100 p~ ; sodium, 35 p~ ; iron (as Sequestrene 138), 20 p~ ; chloride, 15 p ~ ; boron, 3 p ~ ; zinc, 0-5 p ~ ; manganese, 0.25 p ~ ; copper, 0.10 p ~ ; cobalt, 0.04 p M ; and molybdenum, 0-04 p ~ . Silicate would normally be added to the nutrient solution a t a level of 10 p~ but was regarded as a treatment variable in the present investigation. For each sample blanks were prepared by using volumes of basal solution equal to the volume of complete nutrient solution being analysed to determine phosphate.Results and Discussion Acid Concentration Acidic conditions are required for the formation of both the heteropolymolybdophosphate and heteropolymolybdosilicate complexes. Several investigators have recommended pH values ranging from 0.7 to 3.8, including Hurford and Boltz,lG who selected a pH of 1.3 as being the optimum for the formation of both heteropoly complexes. Increasing the final acid concentration to 0.76 N (pH 0.75) to prevent formation of the molybdosilicate complex has been reported.' Wadelin and Mellon,l7 in a detailed study of the extraction of heteropoly acids, found that 0.96 N hydrochloric acid was the optimum concentration for extraction of the heteropolymolybdophosphate into butan-1-01, whereas 0.15 N hydrochloric acid was the optimum acid concentration for extraction of the corresponding silicon complex into the same reagent.Hence, the use of acid concentrations in excess of 0.76 N in the extraction solution should assist in the suppression or removal of silicon interference, both at the complex form- ation and extraction stages. Dilute the solution to give a concentration of phosphorus of 6 . 4 ~ ~ or less with double Phosphate was varied over the range 0.0440 p ~ .410 JINTAKANON et al. : MEASUREMENT OF LOW PHOSPHORUS Analyst, VoZ. 100 In the present study, when ethyl acetate was used as the extractant, optimum colour development of the reduced molybdophosphate complex was obtained with the addition of 3-7 ml of 10 N hydrochloric acid (0.51-1.11 N in the extraction solution), there being no detectable silicon interference at any acid concentration (Table I).The excellent agreement between the absorbance readings for the two series clearly demonstrates that the procedure based on extraction into ethyl acetate was successful in removing any silicon interference. This result agrees closely with that of Wadelin and Mellon,17 who found that ethyl acetate effectively extracted the heteropolymolybdophosphate but not the heteropolymolybdo- silicate complex, TABLE I EFFECT OF SOLUTION ACIDITY ON THE DETERMINATION OF PHOSPHORUS WITH AND WITHOUT THE ADDITION OF 50 pmOl 1-l OF SILICON Values are means of three replicates; ethyl acetate used as blank. r Acidity (lo N H C I ) / d 0 1 2 3 4 6 6 7 Absorbance at 720 nm A 'L Phosphorus : silicon, 2.5 : O* 0.015 0.010 0.350 0.350 0.410 0.450 0.430 0.426 0.430 0.420 0.440 0.435 0-420 0.425 0.435 0.425 Phosphorus : silicon, 2.5 : 50* * Phosphorus and silicon concentrations (p~).Molybdate Concentration An excess of molybdate is required to ensure that all of the phosphate present in the sample is transformed into the yellow molybdophosphate complex. Dickman and Bray1* stated that a 0.3 per cent. final concentration of ammonium molybdate was adequate for phosphorus concentrations up to 16 p~ (25 pg of phosphorus in the sample), while Wadelin and Mellonl7 found that 0.6 per cent. of sodium molybdate in the final solution was required for the deter- mination of up to 25 pg of phosphorus. Hurford and Boltzla used 1-0 per cent. of ammonium molybdate for colour development in solutions containing up to 130 p g of phosphorus and 130 pg of silicon.In the current study it was found that 5 ml of 5 per cent. sodium molybdate solution (0.42 per cent. in the final solution) was sufficient for maximum colour development when 2.5 pmol 1-1 of phosphorus (3.9 pg of phosphorus in the sample) was determined in the presence of 50 pmol 1-1 of silicon (70 pg of silicon in the sample) (Table 11). TABLE I1 EFFECT OF SODIUM MOLYBDATE CONCENTRATION ON DETERMINATION OF PHOSPHORUS FOR SOLUTIONS CONTAINING 2.5 pm01 1-l OF PHOSPHORUS AND 50 pmOl 1-l OF SILICON Ethyl acetate used as blank. 5 per cent. sodium molybdate solution added/ml 0 2.5 5.0 7.6 10.0 12.5 Absorbance a t 720 nm 0.075 0.390 0.420 0.425 0.425 0.430 Extractants Kegginlo found that oxygen-containing compounds are good solvents for the extraction of heteropoly acids, while Scroggie20 concluded that esters, ketones, aldehydes and ethers are good extractants of heteropolymolybdophosphates.Since these early reports, many different extractants have been proposed and used. Thus, DeSesa and Rogersll used isoamyl acetate,Jane, 1975 CONCENTRATIONS IN NUTRIENT SOLUTIONS CONTAINING SILICON 41 1 Lueck and Boltz21 used 2-methylpropan-1-01, Boltz22 used butan-1-01, Zaugg and K n ~ x ~ ~ used octan-2-01, Kirkbright, Smith, and West24 used isobutyl acetate and Hurford and BoltzlB used diethyl ether to extract the molybdophosphoric acid. Wadelin and Mellonl' have also demonstrated that a wide range of organic solvents can extract molybdophosphate complexes from aqueous solution.In the present investigation, butan-1-01, 2-methylpropan-1701, pentan-1-01, isoamyl alcohol and ethyl acetate were found to be efficient extractants of the molybdophosphate complex both in the presence and the absence of silicon (Table 111) when 4 or 5 ml of 10 N hydrochloric acid were added to the solution to be extracted. These results are substantially in agreement with those of Martin and D ~ t y , ~ ~ who found that 2-methylpropan-1-01, isoamyl alcohol and ethyl acetate were effective in extracting molybdophosphates from aqueous solution. How- ever, diethyl ether and isobutyl acetate, which have been reported to be satisfactory extrac- t a n t ~ , ~ * removed only small amounts of the molybdophosphate complex. TABLE I11 COMPARISON OF SEVEN ORGANIC SOLVENTS AS EXTRACTANTS FOR THE MOLYBDOPHOSPHATE COMPLEX I N THE PRESENCE AND ABSENCE OF SILICON AT Test solution concentrationlpM Phosphorus : silicon, 0:o Phosphorus : silicon, 2.6; : 0 Phosphorus : silicon, 2.5 : 60 THREE SOLUTION ACIDITY LEVELS Each extractant used as blank as appropriate.Volume of hydro- chloric acid (10 N) I h present/ 2-Methyl- Isoamyl Ethyl Absorbance a t 720 nm in ml Butan-1-01 propan-1-01 Pentan-1-01 alcohol acetate 6 0.080 0.040 0.035 0.030 0.085 4 0.100 0.045 0.035 0.030 0.045 2 0-760 0.190 0.210 0.150 0.055 6 0.510 0.470 0.400 0.430 0.426 4 0.520 0-455 0.390 0.430 0.415 2 0.930 0.570 0.620 0.510 0.415 5 0.495 0.465 0.390 0.425 0.420 4 0.510 0.470 0.390 0-420 0.425 2 > 2 1-700 1.800 1.600 0.415 Diethyl ether 0.013 0.006 0.003 0.1 10 0.075 0.055 0.130 0.140 0.050 - Isobutyl acetate 0.025 0.020 0.010 0.120 0.100 0.126 0.110 0.118 0.118 While butan-1-01, 2-methylpropan-1-01, pentan-1-01, isoamyl alcohol and ethyl acetate were satisfactory for extracting the molybdophosphate complex in the presence of silicon when 4 or 5 ml of 10 N hydrochloric acid were used, only ethyl acetate was satisfactory in preventing interference by silicon when 2 ml of hydrochloric acid were used.With 2 ml of 10 N acid silicon interference was greater in butan-1-01 than any other solvent. These results are consistent with the observation of Wadelin and Melion17 that ethyl acetate had a considerable advantage over 2-methylpropan-1-01 in that it totally discriminated between the yellow molybdophosphoric acid and molybdosilicic acid.A further advantage of ethyl acetate is that readings on the basal nutrient solution of cultures, to which neither phosphorus nor silicon had been added, were much less affected by solution acidity than those with the other solvents that were found to be effective in extracting the molybdophosphate complex (Table For culture solutions containing phosphorus at a concentration of 0.5 PM or more, 13 ml of ethyl acetate were found to be adequate. However, 25 ml of ethyl acetate were required for the larger sample volumes used at phosphorus concentrations below 0.5 PM. At a phosphorus concentration of 2-5 ,UM, no interference was detected for silicon concen- trations ranging from 0 to 2000 p~ when the solution acidity was adjusted to 0.82 N (equivalent to 5 ml of 10 N hydrochloric acid per sample) and ethyl acetate was used as the extractant (Table IV).Washing Boltz22 recommended that the organic phase should receive two separate washes with water prior to measurement of its absorbance. Hurford and Boltzl6 washed their diethyl ether extract of molybdophosphoric acid with hydrochloric acid - water (1 + 9). On the other 111).412 JINTAKANON et al. : MEASUREMENT OF LOW PHOSPHORUS Analyst, Vol. 100 TABLE IV EFFECT OF SILICON CONCENTRATION ON PHOSPHORUS DETERMINATION IN NUTRIENT SOLUTIONS CONTAINING 2-5 pmol 1-1 OF PHOSPHORUS USING THE ETHYL ACETATE EXTRACTION METHOD ConcentrationlpM Absorbance at 720 nm 0 0.420 60 0-410 100 0-400 200 0.410 300 0.400 400 0.400 500 0.405 700 0.390 900 0.405 1000 0.425 1500 0-425 2001) 0.431) Silicon hand, Pakalns and McAllisterl5 did not wash the isobutyl acetate extract of molybdo- phosphoric acid prior to reduction to heteropoly blue.The present results show that washing, particularly with water, resulted in a marked reduction in the absorbance of the reduced molybdophosphate complex (Table V) . Clearly, no advantage accrues from the washing of ethyl acetate extracts of molybdophosphoric acid. TABLE V ABSORBANCE AT 720 nm OF REDUCED MOLYBDOPHOSPHATE WITH AND WITHOUT WASHING Means of two replicates, phosphorus at 2.5 p ~ , silicon a t 50 p ~ , ethyl acetate blank and 10 ml of washing solution used for each wash. Washing Absorbance at 720 nm None .. .. .. .. .. . . 0.410 Once with water. . .. .. .. . . 0.130 Twice with water .. . . .. . . 0.008 Once with hydrochloric acid (1 + 9) . . . . 0.330 Twice with hydrochloric acid (1 + 9) . . 0-320 Wavelength Phosphate can be determined by measuring either the absorbance of the yellow molybdo- phosphoric acid at 330nm, or the absorbance following reduction to heteropoly blue at wavelengths in the visible region beyond 660 nm. Thus, Jackson3 recommended 660 nm, whereas Durge and Paliwa126 used 675 nm, in determinations of the phosphate in plant digests. However, Pakalns and McAllister15 measured the absorbance of extracts in isobutyl acetate at 725 nm, while Boltz22 recommended 720 nm for the measurement of absorbance in but an- 1-01 extracts. Measurement of the absorbance of ethyl acetate extracts in a range of phosphate concen- trations over the wavelength range from 450 to 800 nm established that maximum absorbance occurred at 720 to 725 nm.The absorption spectrum for a test solution containing phosphorus at a concentration of 2.5 p~ but no silicon is shown in Fig. 1. Time to Colour Measurement Several reports have appeared in the literature regarding the instability of the heteropoly blue and most authors recommend reading the absorbance at a standard time after the reduction of the molybdophosphoric acid. However, Martin and DotyZ5 found that the heteropoly blue colour remained stable for up to 24 h. In the present study it was found that from 15 to 30min were required for full colour development after the reduction with tin(I1) chloride. The stability of the colour of the reduced molybdophosphate complex was most satisfactory over the period from 30 to 60 min after reduction (Fig.2). It is recommended that readings should be made in this period.June, 1975 CONCENTRATIONS IN NUTRIENT SOLUTIONS CONTAINING SILICON 413 I I I 1 I I I 450 500 550 600 650 700 750 800 Wavelengthhm Fig. 1. Absorption spectrum for a test solution containing a 2.5 p~ concentration of phosphorus after extraction into ethyl acetate. Calibration Graphs In the range of phosphorus concentrations from 0.5 to 6.4 p ~ , a standard graph for use in determining the phosphate concentration of complete nutrient solutions was prepared by diluting suitable aliquots of the standard 50 ,UM phosphorus solution to a volume of 50 ml with basal nutrient solution. Thereafter, the procedure was as described under section B of Procedures.In the lower concentration range, viz., below 0.5 ,UM, suitable aliquots of the 8 PM phosphorus solution were diluted to a volume of 200 ml with basal nutrient solution, and then prepared as given under section C of Procedures. Calibration graphs over the phosphorus concentration ranges 0-5-6.4 PM and 04U-0-5 PM, respectively, are given by the following linear equations : A,zonm = 0.1636 P + O*OOO 14 (Y > 0.999) A 720 nm = 0.5964 P - 0.0007 (Y = 0.993) where P pmol l-1 is the phosphorus concentration. I I I I I I 10 20 30 40 50 60 0 Time after mixing/min Fig. 2. Changes in absorbance of ethyl acetate A, extracts as a function of time after reduction. P = 0.5 pM, s i = 50 pM; B, P = 2.5 p M , si = 50 pM. Precision and Accuracy The reproducibility of the method was determined at a very low phosphate concentration (0.04 p~ phosphorus) in the presence of 50 pmol 1-1 of silicon by conducting analyses on seven414 JINTAKANON, KERVEN, EDWARDS AND ASHER separate solution samples.The mean phosphorus concentration determined (& standard deviation) was 0.037 & 0.005 PM. In addition, a satisfactory recovery of phosphate from the basal nutrient solution to which known amounts of standard phosphate solution had been added was obtained (Table VI). TABLE VI DETERMINATION OF PHOSPHORUS IN BASAL NUTRIENT SOLUTIONS CONTAINING 0.04 TO 0.45 pmOll-l OF PHOSPHORUS Values are means of 3 replicates. Phosphorus added/pM Phosphorus found/@¶ Recovery, per cent. 0.04 0.037 92.5 0.15 0-140 93.3 0-18 0.190 105.6 0.25 0.240 104.0 0.40 0.420 105.0 0-45 0.470 104.0 The senior author (S. J.) expresses his appreciation to the Thai Government and Kasetsart University for support of his work.1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. References “Standard Methods for the Examination of Water, Sewage, and Industrial Wastes,” Thirteenth Rieman, W., and Beukenkamp, J., in Kolthoff, I. M., and Elving, P. J., Editors, “Treatiseon Analytical Jackson, M. L., “Soil Chemical Analysis,” Prentice Hall Inc., Englewood Cliffs, N. J., 1958, pp. 135- Halasz, A., Pungor, E., and Polyak, K., Talanta, 1971, 18, 577. Boltz, D. F., and Mellon, M. G., Ind. Engng Chem., Analyt. Edn, 1947, 19, 873. Williams, D. E., and Vlamis, J., PI.Physiol., Lancaster, 1957, 32, 404. Rockstein, M., and Herron, P. W., Analyt. Chem., 1951, 23, 1500. Ingle, J. D., jun., and Crouch, S. R., Analyt. Chem., 1971, 43, 7. Kolthoff, I. M., Sandell, E. B., Mechan, E. J., and Bruckstein, S., “Quantitative Chemical Analysis,” Fourth Edition, MacMillan and Co. Ltd., London, 1969, p. 644. Chalmers, R. A., and Sinclair, A. G., Analytica Chim. A d a , 1966, 34, 412. DeSesa, M. A., and Rogers, L. B., Analyt. Chem., 1954, 26, 1381. Asher, C. J., and Loneragan, J. F., Soil Sci., 1967, 103, 225. Robson, A. D., Edwards, D. G., and Loneragan, J. F., Aust. J . Agric. Res., 1970, 21, 601. Ryden, J. C., Syers, J. K., and Harris, R. F., Analyst, 1972, 97, 903. Pakalns, P., and McAllister, B. R., J . Mar. Res., 1972, 30, 305. Hurford, T. R., and Boltz, D. F., Analyt. Chern., 1968, 40, 379. Wadelin, C., and Mellon, M. G., Analyt. Chem., 1953, 25, 1668. Dickman, S. R., and Bray, R. H., Ind. Engng Chem., 1940, 12, 665. Keggin, J. F., Proc. R. SOC., 1934, 144, 75. Scroggie, A. G., J . Amer. Chem. Soc., 1929, 51, 1057. Lueck, C. H., and Boltz, D. F., Analyt. Chem., 1956, 28, 1169. Boltz, D. F., “Colorimetric Determination of Non-metals,” Interscience Publishers Inc., New York, Zaugg, W. S., and Knox, R. J., Analyt. Chem., 1966, 38, 1759. Kirkbright, G. F., Smith, A. M., and West, T. S., Analyst, 1967, 92, 411. Martin, J. B., and Doty, D. M., Analyt. Chem., 1949, 21, 965. Durge, A. S.. and Paliwal, K. V., PI. Soil, 1969, 31, 374. Edition, American Public Health Association, New York, 1971, p. 527. Chemistry,” Part 11, 1961, p. 317. 181. 1958, p. 34. Received October 29th, 1974 Accepted JaNecary 14th, 1976

ISSN:0003-2654    

DOI:10.1039/AN9750000408

出版商:RSC    

年代:1975

数据来源: RSC