ISSN:0003-2654    

DOI:10.1039/AN97196FX041

出版商:RSC    

年代:1971

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97196BX043

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

iv SUMMARIES OF P-iPEKS I N THIS ISSUE [November, 1971Summaries of Papers in this IssueA Flexible Program for Calculations inEmission- spectrographic AnalysisA computer program has been written in FORTRAN 1V to evaluate forspectrochemical analysis spectral information stored on photographic plates.With this program, ratios of line intensities are calculated from microphoto-meter readings ; it includes the computation of the emulsion Calibrationfunction using either the two-step or seven-step method. A formatlessdata input technique is employed and use is made of descriptive headingsand labels to facilitate the introduction of data. These headings and thecodes used for labelling the readings are chosen so as to be readily understoodby the spectrographer. The program is characterised by considerable flexi-bility with regard to the number of spectra and number of analytical andinternal standard lines that can be used.The maximum number of spectrathat can be evaluated is sixty and up to nineteen analytical lines can bepresent in each spectrum. The program contains many data checks, togetherwith appropriate diagnostic messages, which are designed to avoid thepresence of undetected errors in the results.B. L. TAYLOR and F. T. BIRKSAnalytical Sciences Division, Atomic Energy Research Establishment, Harwell,Djdcot, Berks.Analyst, 1971, 96, 753-763.The Determination of Alumina in Iron Ores, Slags and RefractoryMaterials by Atomic-absorption SpectroscopyAs chemical methods of determining alumina are time-consuming, thesimpler, more rapid atomic-absorption technique was adapted to iron ores,slags and refractory materials.The basic technique comprises fusion of the sample with a sodiumtetraborate - sodium carbonate mixture, extraction with dilute nitric acid,appropriate dilution and spraying under recommended conditions foraluminium.As the flux materials enhance the aluminium absorption, sampleand corresponding calibration solutions should contain equivalent amounts.Interference tests on synthetic solutions of the other constituents showedthat the inhibiting effect of silica was suppressed by lanthanum chloride. Astandard addition of lanthanum chloride was therefore incorporated andcalibration conditions were developed to cover the ranges 0 to 1 per cent.,0 to 5 per cent., 0 to 10 per cent.and 10 to 35 per cent. of aluminium.Applications to miscellaneous B.C.S. materials showed good agreementwith accepted values and reproducibilities of 3 0-035 per cent., &0*12 percent., fO-18 per cent. and f0.38 per cent. of aluminium, respectively, a tthe above levels.W. D. COBB and T. S. HARRISONGroup Chemical Laboratories, British Steel Corporation, Scunthorpe Group, P.O.Box No. 1, Scunthorpe, Lincolnshire.Analyst, 1971, 96, 764-770.The Atomic-fluorescence Determination of Mercury by theCold Vapour TechniqueA method for the determination of low levels of mercury by using atomicfluorescence in conjunction with the cold vapour technique is described, andshows significant advantages over the corresponding absorption technique.K.C. THOMPSON and G. D. REYNOLDSShandon Southern Instruments Ltd., Frimley Road, Camberley, Surrey.Analyst, 1971, 96, 771-775vi SUMMARIES OF PAPERS I N THIS ISSUEX-ray Spectrometric Determination of Rare Earth Elementsby Using a Fusion TechniqueA general method for the determination of rare earth elements byX-ray fluorescence spectrometry is described. The sample is fused withsodium tetraborate, chromium is added to act as internal control-standard,and the resulting bead is analysed directly. This method overcomes variationsin sample form and particle size. Individual rare earths can be determined a tlevels from 0.1 to 100 per cent., and by using the L spectra and a lithiumfluoride 220 crystal, line overlap is kept to a minimum.C.PLOWMANCentral Electricity Generating Board, Scientific Services Department, Kirkstall,Leeds, LS4 2HB.Analyst, 1971, 96, 776-778.[November, 1971The Use of a Shield for the Reduction of Fogging on PhotographicPlates in Spark- source Mass SpectrographyIn spark-source mass spectrography, secondary effects cause pronouncedfogging of the photographic plate in the vicinity of the isotope lines of themajor elements, and the limit of detection of elements whose lines fall withinthis area is considerably raised.A shielding device has been designed to reduce the amount of ionsproduced by secondary effects and to prevent them from reaching the photo-graphic plate. This reduces the background in the region of the major isotopelines and enables elements in this region to be determined with increasedsensitivity and precision.C.W. FULLER and J. WHITEHEADTioxide International Limited, Billingham, Teesside.Analyst, 1071, 96, 779-784.Automatic Radiofrequency Titration of Acids inTertiary Butyl Alcohol - Acetone MediumThe mixed solvent t-butyl alcohol - acetone can be used as a differen-tiating medium for the titration of acids with a standard solution of tetra-n-butylammonium hydroxide dissolved in a mixture of toluene and methanol.Titration curves of various shapes can be obtained and the simultaneousdetermination of certain acids achieved. The differences in curve shapesobtained by radiofrequency titration can be used as the basis of separation.The low toxicity of t-butyl alcohol (compared with other solvents a t present inwidespread use for the titration of weak acids) combined with its low solvatingpower and good differentiating properties make it a suitable medium forthe routine determination of such acids in admixture.W.J. SCOTT and G. SVEHLADepartment of Analytical Chemistry, The Queen’s University, Belfast.Ajtalyst, 1971, 96, 785-797.Improvements to the Nitrite - Diazo Dye (Blom’s) Method ofDetermining Hydroxylamine as Used in the Determinationof Residues of AldicarbThe precision of determination of hydroxylamine, and hence of residuesof aldicarb, by Blom’s method has been improved by controlling the pHof the solutions at the diazotisation and coupling stages and by removingthe excess of iodine, after oxidation of the hydroxylamine, by extraction intobromobenzene. The latter previously required chemical reduction to iodide.D. F. LEE and J. A. ROUGHANMinistry of Agriculture, Fisheries and Food, Plant Pathology Laboratory, HatchingGreen, Harpenden, Herts.Analyst, 1971, 96, 798-801

ISSN:0003-2654    

DOI:10.1039/AN97196FP165

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

xiv SUMMARIES OF PAPERS IN THIS ISSUE [November, 1971Determination of Orthophosphate in the Presence of Nitrate IonsA method is described for the spectrophotometric determination oforthophosphate in the presence of nitrate. Reduction of nitrate ions toammonia is effected by aluminium metal and sodium hydroxide solution,and after the removal of all the ammonia produced (by heating) the phos-phate is determined by the formation of the phosphomolybdate complexand reduction to molybdenum blue. The method is suitable for the deter-mination of orthophosphate a t a concentration of M in the presence ofmore than 10-1 M of nitrate. I t also overcomes interference by fluoride andother halide ions in the molybdenum-blue method.E. J. DUFF and J. L. STUARTDepartment of Preventive Dentistry, Turner Dental School, The University,Manchester, M15 6FH.Analyst, 1971, 96, 802-806.The Determination of Epichlorohydrin in Aqueous Solutions in thePresence of Glycerin, Monochlorohydrin and GlycidolEpichlorohydrin can be extracted from aqueous solutions containingglycerin, monochlorohydrin and glycidol into carbon tetrachloride and deter-mined by infrared spectrophotometry.The high-frequency component ofthe complex band a t a wavenumber 1274 cm-l is used as the analytical band.A three-fold excess of carbon tetrachloride over the aqueous phase is necessaryfor the quantitative extraction. A cell thickness of up to 1 mm can be used; theconcentration limit of the solute in the aqueous phase is then 0.3 per cent.v/v.PETR ADAMEKInstitute of Chemical Technology, Prague 6, Czechoslovakia.and VLASTIMIL PETERKAResearch Institute of the Fat Industry, Rakovnik, Czechoslovakia.Analyst, 1971, 96, 807-809.The Use of Back- flushing with Electron-capture GasChromatographyAn apparatus is described that enables materials with a longer retentiontime than those of interest to be back-flushed out of electron-capture gaschromatographs. Shorter analysis times, a longer column life and lessfrequent cleaning of the detectors have been achieved with this device.B. T. CROLLThe Water Research Association, Medmenham, Marlow, Buckinghamshire.Analyst, 1971, 96, 810-813.An Enzymic Method for the Determination of Dried SkimmedMilk in Mashed Potato PowdersAn enzymic method is described for the determination of dried skimmedmilk in mashed potato powders. The method is based on the determinationof free lactose by its hydrolysis with /iI-galactosidase, the glucose formedbeing determined by the hexokinase method. The determination is freeof interference from varying concentrations of reducing sugars found indried potatoes. Accuracy and reliability are greater with this method thanwith other methods currently in use.R. K. BAHLJ. Sainsbury Ltd., Stamford House, Stamford Street, London, S.E.1.Analyst, 1971, 96, 814-816.The Determination of Carbarsone in Animal FeedsReport prepared by the Prophylactics in Animal Feeds Sub-committee.ANALYTICAL METHODS COMMITTEE9/10 Savile Row, London, W1X 1AF.Analyst, 1971, 96, 817-823

ISSN:0003-2654    

DOI:10.1039/AN97196BP173

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

NOVEMBER, I97 I THE ANALYST Vol. 96, No. I148 A Flexible Computer Program for Calculations in Emission-spectrographic Analysis BY B. L. TAYLOR AND F. T. BIRKS (Analytical Sciences Division, Atomic Energy Research Establishment, Harwell, Didcot, Berks) A computer program has been written in FORTRAN IV to evaluate for spectrochemical analysis spectral information stored on photographic plates. With this program, ratios of line intensities are calculated from microphoto- meter readings; it includes the computation of the emulsion calibration function using either the two-step or seven-step method. A formatless data input technique is employed and use is made of descriptive headings and labels to facilitate the introduction of data. These headings and the codes used for labelling the readings are chosen so as to be readily understood by the spectrographer. The program is characterised by considerable flexi- bility with regard to the number of spectra and number of analytical and internal standard lines that can be used.The maximum number of spectra that can be evaluated is sixty and up to nineteen analytical lines can be present in each spectrum. The program contains many data checks, together with appropriate diagnostic messages, which are designed to avoid the presence of undetected errors in the results. THE ratio of the light intensity of an analytical line to that of an internal reference line is a quantity of prime importance in analytical emission spectrography. In general, the calcu- lation of this intensity ratio has to be carried out for several analytical lines in each of several spectra recorded on a photographic plate.Writing a computer program to calculate these ratios is not difficult when the type of analysis is fixed and the analytical and reference lines do not alter. However, if a versatile program is required that is to be applicable to any emission-spectrographic- determination, the task becomes more difficult. This paper des- cribes one method of writing such a program. It is written in FORTRAN IV for use with the Harwell IBM 360/75 computer. The mode of data input is punched card, punched paper tape or keyboard, but for simplicity the following description is given in terms of punched card input. GENERAL CONSIDERATIONS The following standard procedure is used in this laboratory for photographically recording spectra and measuring them for the purposes of spectrochemical analysis.On each plate a two or seven-step iron-arc spectrum is recorded by using a step filter or a rotating step sector; this provides the intensity pattern used for the calibration of the photographic emulsion. The spectra for standards and analysis samples, with or without a two-step filter, are recorded on the remainder of the plate. The transmittances ‘of lines and backgrounds are measured with a modified Hilger and Watts microphotometer which gives a reading proportional to the amount of light reaching the detector. In order to determine the relative transmittance* the ratio of the reading for the exposed portion of the emulsion to the reading for a clear portion is taken. (The latter is maintained at an arbitrary value of 50 during the measurements,) The main problem in writing a versatile program to compute ratios of intensities is concerned with the input of data.Each microphotometer reading has to be introduced in a way that indicates the line and spectrum to which it refers. In addition, the data for the emulsion calibration and a list of the required intensity ratios have to be read in. The * The nomenclature used throughout follows the recommendations given in (a), “Nomenclature, Symbols, Units and their Usage in Spectrochemical Analysis-I (Tentative),’’ I.U.P.A.C., 1969, and (b), “Methods for Emission Spectrochemical Analysis,” 5th Edition, A.S.T.M., Philadelphia, 1968. 0 SAC; Crown Copyright Reserved.753754 TAYLOR AND BIRKS: FLEXIBLE COMPUTER PROGRAM FOR CALCULATIONS [Autdyd, VOl. 96 usual method used in the FORTRAN language for identifying data depends entirely on the sequence and column positioning of the data in the card deck. This method suffers from a number of drawbacks, especially when there are several types of input data and the amount of data varies from one computation to another. Under these circumstances failure of the user to satisfy the rigid requirements of a formatted data input technique is a frequent source of job failure. One method of removing some of the restrictions was suggested by Franke and Poste.1 They used a method in which each numerical value submitted was identified by a two-digit numerical code which preceded the number.This removed the necessity of submitting the data in a fixed sequence but exact column positioning was still necessary. In addition, numerical coding presents difficulties both in punching and checking the data which might well discourage potential users of the program. In our program an alternative approach has been used that is based on the use of easily understood key words and non-numerical codes for labelling the different types of data. These codes, which are chosen by the user, are used to associate each microphotometer reading with a particular line. They are also used when specifying the ratios required. The need for exact column positioning has been removed by using a formatless-data input routine that already exists in the Harwell Subroutine Library.This subroutine enables the data to be read, one word at a time, irrespective of the position of the word on the data card. A word is defined as a string of characters separated from any other string by at least one blank. CARD TYPE 29/8/69 A1 542 1 1 1 TITLE RFB89 2 I EMULSION CALIBRATION 44.13 33.25 23-56 12.22 6.08 2-88 1.45 I 3 4 { 5 I RATIO INTERNAL STD CO LINES MN CR NI AL MO CU W I I DATA SPECTRUM 1 LINE SI 8-78 48.96 19-99 49-04 12.51 49.46 I I DATA SPECTRUM 2 LINE MN * * * * * * 2.59 44.35 I I xxxxxxxxx I Fig. 1. Examples of types of data card DESCRIPTION OF THE PROGRAM There are five types of card used for data input, each defined by a key word, which is the first word on the card. Some typical examples are shown in Fig. 1. The key words specifying the type of data, and the other words on the cards, are designed to aid both in punching and checking the data.There is no restriction concerning the column positioning of any of the information other than that both words and numbers must be separated by at least one blank. The five types of card must be submitted in a fixed order but the order of the cards within any one class is irrelevant. As an additional aid in presenting the input data, sets of cards can be made available with the labels pre-punched on them, each class on a different coloured card. The five types of card are as follows. Title--This contains the title of the job. With the IBM 360/75 system, in addition to the characters 0 to 9 and A to 2, there are twenty-four special characters, such as *, /, %, ) and -, that can be used in the title.Emzllsion calibration-The program is designed to use either the two-step or seven-step emulsion calibration procedure.2 If more than seven microphotometer readings are punched, the program assumes that the two-step method is being used and carries out the appropriate function-fitting procedure. For the two-step method the microphotometer readings are punched in pairs of strong and weak steps, e.g., 4.1 8.9 2.1 4.7 11.3 23.2 etc. For the seven- step calibration, the readings are punched in descending order, that is, weak step first. The transmission ratio of the calibration device is treated as a local constant and is not read in as data. If the word COEFFICIENTS is punched after EMULSION CALIBRATION, the two previously determined coefficients of the emulsion calibration function can be read in from this card (see APPENDIX).November, 19711 IN EMISSION SPECTROGRAPHIC ANALYSIS 755 If the wavelength range of the analytical lines is such that a single emulsion calibration does not adequately cover it, the data are divided into wavelength sets and each set, plus the corresponding emulsion calibration data, is submitted for a separate cycle of the program (see section on End card).Intensity ratios-The cards of this type serve both to introduce the codes used to label the microphotometer readings and to specify the intensity ratios required. The way in which each microphotometer reading for a line is processed is determined solely by the label given it and the ratios listed on the RATIO cards.Each internal reference line code is punched on to a separate card and the analytical lines to be compared with it follow on the same card. The line codes are chosen by the user, subject to the limitations that: no more than eight characters are used; there are no embedded blanks; the first character is a letter from A to z; and only letters and numbers are used (i.e., no decimal points are allowed). For the microphotometer readings for the copper line at 324.75 nm, valid codes would be cu32475, LINE2, cu, cu2, cuw (where w could be used to indicate a weak step reading of a spectral line), etc. This is a very versatile method of data input. For example, it is possible, by introducing the same data more than once with a different label each time, to obtain the ratios of a particular line against a number of different internal reference lines.This procedure may be useful when developing an analytical method. Again, if it is considered necessary to measure a particular reference line more than once, in order to compensate for drift in the microphotometer, each reading can be given a different code and the analytical lines to be compared with that reading are listed on the appropriate RATIO card. In the most general instance every reading could be given a different code, thus allowing it to be processed independently of all the other readings. More usually, however, the spectra to be evaluated are related and the same analytical and reference lines are used for each spectrum. In these circumstances, the same line codes can be used in each spectrum.Microphotometer readings-The microphotometer readings for the analytical and reference lines are punched on to the cards together with the line codes and the spectrum number. The spectrum number is of the form IS.IR where IS and IR are integers and the maximum value of IR is 4. The numbering sequence is therefore 1.1, 1.2, 1.3, 1.4, 2.1, 2.2, etc. This method of numbering the spectra is designed to cope with the situation when replicate spectra are taken for each sample. The number IR can then be used as the replicate number. If necessary, the average ratio of intensities for up to four replicates can be calculated and printed with the results. Only the IS portion of the spectrum number is punched on to the card and the pairs of readings, up to four in number, in the order line followed by back- ground, are automatically numbered IS.1 to IS.4.It should be emphasised that the way in which the spectrum number is used is left to the user and spectra numbered 2.1, 2.2, 2.3, 2.4 can be completely independent spectra or four replicates from a single sample. The maximum value of IS is fifteen so that data for sixty spectra can be evaluated. When less than four pairs of readings are available for a card the gaps in the data are indicated by punching a single asterisk (e.g., card 4B, Fig. l), except when the gap is at the end of the card in which case the space is left blank (e.g., card 4A, Fig. 1). A single asterisk can be used in place of a background reading when no correction for background is required. If, for a particular line, there are no data for any of the four spectra, no card is submitted.With this method of presentation of the microphotometer readings, no distinction is made between data for the analytical and reference lines. This contrasts with a possible alternative technique in which the reference line reading is identified by being punched immediately after that for the analytical line. This last method suffers from several dis- advantages, among which is that it generally involves considerably more data punching. End card-A card with at least eight x symbols is used to indicate the end of the data. If instead a second title card is placed at this point, the calculation will be repeated with a second set of data. Originally the mode of data input was punched cards, but more recently the micro- photometer has been modified so that the readings are transferred directly to paper tape.The key words, line codes, etc., are added from a keyboard. However, these paper tape data are still in card image form.756 TAYLOR AND BIRKS: FLEXIBLE COMPUTER PROGRAM FOR CALCULATIONS [AnaZyst, Vol. 96 Ratio card found indicating end calibration data I s word in printing of error message __--- of emulsion and job termination of values store Read key word on 2nd card t p EMULSION Read 3rd word I , 1 CO E F F I CI ENTS calibration coefficients t t 7-step data into arrays of strong and weiik step t 2-step data into arrays of strong and L, weak step Call emulsion calibration su broutir,e EMCAF reference line Store word No as line Word is item data ..store Fig. Z(a). Flow diagram of main programNovember, 19711 I N EMISSION SPECTROGRAPHIC ANALYSIS Read 3rd word on data card + word a number f Yes I Word i s a spectrum number (N) t I Note position of I ine symbol in dictionary (M 1 Read micro- photometer values for spectrum, N line, M End card ___ found Read key word of next card + + DATA ___--- subroutine GRlNT t 0- FLAG = 1 t No Finish E 757 Fig. 2(a)--continued758 TAYLOR AND BIRKS: FLEXIBLE COMPUTER PROGRAM FOR CALCULATIONS [AfidySt, VOl. 96 Subroutine GRlNT are microphotometer G and BG. and background, Emulsion respectively. calibration GCP is clear plate function reading I I I b L Calculate intensities of linesand backgrounds message t Subtract background from line intensities 1 reference line is determined from the position of line Ratio corrected cOc*c intensities of line and internal reference line Print ratios of intensities Subroutine EMCAF Calculate transmittances from emulsion calibration subroutine to fit data to equation (4) (see appendix) I f Print error message 4 faulty data coefficients of emu I sio n calibration Finish r l J Finish 1 Fig.2 (c). Flow diagram of subroutine EMCAF Fig. 2 (b). Flow diagram of subroutine GRINT METHOD OF CALCULATING INTENSITY RATIOS With the input data correctly read in the program calculates the intensity ratios in a conventional mannere2 The steps in the calculation are as follows: firstly, the emulsion calibration data (two or seven-step) are used to derive the emulsion calibration function (see Appendix) ; secondly, the microphotometer readings for the analytical lines, reference lines and their backgrounds are converted to transmittances and the intensities calculated using the emulsion calibration function ; and thirdly, background corrections are made for the analytical and reference lines and the ratios of the corrected intensities are calculated.Simplified flow diagrams of the main parts of the program are shown in Fig. 2 (a), (b) and (c). With standard samples, having determined the intensity ratios, the constructionNovember, 19711 IN EMISSION SPECTROGRAPHIC ANALYSIS 759 cpl SI FE MG GA B PB SN $Y ek MN CR NI cuw cpl SI FE MG GA B PB SN :!r MN CR NI AL cuw Mpl cpl SI FE MG GA B PB SN $7 MN CR NI AL cuw MP 1.1 1.82 48.34 8.78 48.96 1.62 48.12 0.89 40.99 4.86 46.48 33.87 46.49 7.55 40.30 31.53 37.26 24.42 42.43 1.75 47.53 5.37 40.22 6.02 39.42 7.21 27.27 7.23 24.01 11.63 20.42 9.32 40.88 2.1 2.99 48.33 5.88 47.97 3.64 47.94 1.82 44.76 4.55 44.82 33.81 46.73 5.57 44.49 34.74 40.70 77.53 44.52 2.88 47.19 23.74 43.30 11.68 35.40 14.04 31.35 8.03 25.40 7.22 42.61 - - 3.1 2.39 48.15 4.64 48.33 18.67 47.97 0.49 39.10 4.12 44.32 0.67 43.05 2.62 34.24 1.38 35.06 1.01 40.47 2.28 47.01 1.15 39.17 11.98 39.07 3.13 25.74 3.49 21.90 2.01 17.82 0.71 38.39 RFB 89 29/8/69 A1 542 INPUT DATA SPECTRUM NUMBER 1.2 4.45 48.86 19.99 49.04 6.43 47.88 3.62 44.99 5.13 46.79 39.98 47.13 4.56 43.44 33.77 40.78 13.75 43.92 4.30 48.03 16.17 43.91 22.70 42.90 17.96 33.92 23.86 32.14 13.20 23.92 8.16 41.77 2.2 2.56 49.85 3.78 49.98 2.08 48.52 1.00 44.07 3.98 45.27 34.11 46.77 5.59 44.89 34.18 39.45 18.70 45.23 2.44 49.12 16.64 44.09 9.30 35.74 8.67 31.17 6.71 26.54 6.90 43.84 - - 3.2 4.17 48.43 3.72 48.19 22.13 48.71 0.37 40.29 4.96 44.66 0.41 43.80 1.08 32.39 0.82 38.14 0.66 42.91 4.10 47.43 0.91 40.17 12.80 42.25 5.71 34.29 4.63 29.09 1.58 22.69 0.56 41.54 1.3 2.50 48.87 12.51 49.46 2.26 48.29 1.28 44.74 4.02 46.77 34.97 46.07 4.70 42.80 32.92 40.23 12.46 45.06 2.46 48.16 6.69 44.50 9.74 43.74 8.98 34.87 13.00 32.18 10.47 26.93 4.84 43.13 2.3 ,2.95 49.38 6.15 49.10 4.74 49.48 2.11 46.26 4.96 46.33 34.99 47.99 6.03 46.90 38.68 44.38 20.05 47.85 2.82 48.65 27.81 45.42 16.83 41.99 17.63 39.26 10.60 33.25 10.68 45.44 3.3 3.13 46.61 1.28 46.73 6.24 46.44 0.12 33.36 4.57 42.87 0.20 39.19 0.35 19.56 0.33 33.66 0.31 38.51 3.16 46.01 0.46 34.33 3.92 38.64 1.69 28.43 2.19 24.88 0.50 18.07 0.30 38.61 - - 1.4 - - - - - - - - - - - - - - - - - - - - - I - - - - - - - - - - 2.4 2.51 49.56 7.56 49.65 29.75 49.55 1.24 45.01 4.40 46.12 1.72 45.58 7.08 42.58 3.55 41.92 1.55 46.31 2.50 48.51 2.59 44.35 21.48 44.86 11.23 39.19 6.23 34.25 3.95 29.09 1.83 43.57 3.4 1.53 48.09 26.64 49.04 78.01 48.33 9.95 42.57 3.79 44.65 28.73 46.09 33.97 42.54 24.68 43.90 1.48 47.21 - - - - - - 21.80 31.52.13.61 25.35 2.97 41.43 - - Fig. 3 ( 4 . Microphotometer readings760 TAYLOR AND BIRKS: FLEXIBLE COMPUTER PROGRAM FOR CALCULATIONS [Analyst, vol. 96 of the analytical curve is carried out manually. The computer is not used at this stage as it is considered desirable to use the ability of an experienced spectrographer to recognise and reject faulty data.Once the ratios have been obtained the effort required to produce the analytical curve is relatively small. The total execution time on the IBM 360/75, with the data shown in Fig. 1, was 9.3 s, consisting of 4.1 s of central processor time and 5.2 s wait time. RFB89 29/8/69 A1 542 INPUT DATA FOR EMULSION CALIBRATION 44.13 35.25 23.56 12.22 6.08 2.88 1.45 **OUTPUT FROM MA14A** NO. OF EQUALITY CONSTRAINTS = 0 NO. OF EQUATIONS OF CONDITION = 6 NO. OF VARIABLE PARAMETERS = 2 SOLUTI ON ** **** * * ******** 1 0.174056E 01 2 0,295192E 01 RESl DUALS ** * *** ** *++***** 1 -0.154729E-01 2 -0.422842E-01 4 0.875895E-02 5 0.110919E-01 6 -0.976565E-01 3 0.11139'lE 00 END OF MA14A OUTPUT***** *I*** SUM OF SQUARES OF RESIDUALS 0.24172E-01 NO. OF REJECTED POINTS 0 TWO-STEP KAl SE R COE FF ICI ENTS D(1) = 0.52396E 00 D(2) = 0.88862E 00 RATIOS SPECIFIED GROUP 1 INTERNAL STD CO LINES S I FE MG GROUP 2 INTERNAL STD GA LINES B PB SN AGW GROUP 3 INTERNAL STD CO1 LINES MN CR N I AL MO cuw Fig.3(b). Emulsion calibration information and list of requested ratios DETECTION OF ERRORS IN THE DATA INPUT Although the method of data input has been designed to minimise, as far as possible, the occurrence of user errors, they can still occur. It is important that when they do they should be detected by the user's program and a message given stating clearly the nature of the error. If this precaution is not taken one of two things will happen; either the error will go undetected and the final intensity ratios will then contain hidden errors, or, less seriously, the computer supervisor program may detect the error and terminate the program.The latter situation is undesirable because the message given in these circumstances is not always easy to correlate with the user's data input error. The following tests are made on the data as they are read in: the order of the types of card is checked by means of the key words on each card; each microphotometer reading is tested to ensure that it is less than the clear plate reading and that the reading for a line is less than that for the corresponding background; the number of internal reference lines and the total number of line codes used are checked to make sure that they do not exceed 10 and 20, respectively; and, for the two- step method of calibration, the data are checked to ensure that they are in pairs of strongNovember, 19711 I N EMISSION SPECTROGRAPHIC ANALYSIS 761 and weak-step readings. Also, any numerical value that is written in a form not conforming with FORTRAN IV rules generates a message clearly indicating the error and the point a t which it has occurred.RESULTS The printed output from the program for which the data in Fig. 1 were the input is shown in Fig. 3 (a), (b) and (c). The input data are printed for recording and checking purposes. Also printed are the coefficients of the emulsion calibration function, the sum of squares of the residuals and the standard error (o), all obtained from the data-fitting routine. If, during the data fitting, any points are found to lie outside the 30 range they are rejected and a second fitting is carried out.The intensity ratios are printed in an easily comprehended tabular form. RFB 89 29/8/69 A1 542 INTENSITY RATIOS SI FE MG B PB SN AGW MN CR NI AL cuw MId SI FE MG B PB SN AGW MN CR N1 AL cuw Mer SI FE MG B PB SN AGW MN CR NI MA; cuw 1.1 0.222 1.1 12 1.899 0.07 1 0.608 0.036 0.1 24 0.331 0.293 0.210 0.1 98 0.078 0.184 2.1 0.522 0.829 1.569 0.069 0.816 0.032 0.219 0.082 0.204 0.141 0.28 1 0.393 - 3.1 0.536 0.1 17 4.187 5.390 1.494 2.753 3.702 1.853 0.169 0.677 0.584 1.007 2.875 SPECTRUM NUMBER 1.2 0.199 0.692 1.193 0.042 1.102 0.042 0.332 0.224 0.128 0.1 44 0.062 0.171 0.497 2.2 0.699 1.201 2.326 0.059 0.714 0.025 0.1 78 0.1 26 0.234 0.239 0.307 0.354 - 3.2 1.1 13 0.161 9.089 10.054 4.1 23 5.351 6.545 3.988 0.283 0.66 1 0.796 2.291 6.201 1.3 0.207 1.095 1.83 1 0.05 1 0.84 1 0.035 0.301 0.372 0.246 0.244 0.141 0.171 0.512 2.3 0.496 0.641 1.348 0.076 0.827 0.031 0.214 0.064 0.136 0.1 18 0.218 0.257 3.3 2.289 0.51 1 19.107 17.773 10.602 11.316 1 1.999 5.788 0.787 1.722 1.318 5.258 8.525 - 2.4 0.351 0.065 1.885 2.420 0.604 1.209 2.667 0.9 54 0.090 0.1 99 0.375 0.589 1.317 3.4 0.048 0.083 0.1 53 0.083 0.038 0.102 - - - 0.029 0.064 0.514 - Fig.3( c). Intensity ratios762 TAYLOR AND BIRKS: FLEXIBLE COMPUTER PROGRAM FOR CALCULATIONS [Analp$, vol. 96 CONCLUSIONS The computer method of calculation has been used in the emission spectrography labora- tory of the Actinide Analysis Group, A .E . R . E . , for about 3 years and has been found versatile enough to deal with the wide variety of analyses encountered in this work. Also available in this laboratory is an alternative method of computation in the form of a RESPEKTRA calculating board.3 Comparison of these two methods has led to the following observations. Conveniefice-The computer technique has been proved to be simple to use by personnel not particularly familiar with computer programming, and failure of the computation caused by a user’s error is very rare. The ease with which the required ratios are specified encourages the use of more internal reference lines than might otherwise be practical. By contrast the RESPEKTRA board is not found easy to use by untrained staff. Reliability-For a period of several months after the introduction of the computer method, all data were evaluated by both methods.In general the agreement was good, and where significant differences did occur, the cause was traced in all cases to user errors on the RESPEKTRA board. Provided the input data have been correctly introduced the computer method has so far proved to be completely reliable. Speed-Using a keyboard method for input and output of data, the turn-round period between the time the input data enter the computer and the time the results arrive back at the keyboard is about 10 to 20 minutes, depending on the length of the job queue. For all normal applications this time is independent of the amount of calculation to be done and therefore the gain over the RESPEKTRA board increases with the complexity of the calculation.If, for example, the total number of ratios required was fifteen the RESPEKTRA method might be quicker, but it should be appreciated that during the period results are being awaited from the computer the spectrographer is free to carry out other tasks. One disadvantage of the computer compared with the RESPEKTRA board method is that, owing to machine “down-time,” the former cannot be relied on to be available at all times. Mostly, however, convenience and reliability of the calculations are more important than speed. These comparisons pertain to the computer and RESPEKTRA board methods of calculation. If a RESPEKTRA board is not available and manual computation is the only alternative, the computer method becomes even more attractive.Copies of the source listings of the program can be obtained on application to the authors. Appendix THE EMULSION CALIBRATION FUNCTION The emulsion calibration function relates the transmittance (T) of the photographic image of a line to the spectral line intensity ( I ) . As the exact form of this function is not known, an approximation has to be made which predicts as closely as possible the observed relationship between I and T. The approximation used in our calculation is log1 =- 1 P + log10 Y or log I,=! P Y .. .. where y and I,, are constants and IR is a relative intensity defined such that IR = 1 when P = 0. P is the Baker - Sampson - Seidel transformation* P = K log (1 - T ) - log T . . .. .. (2) where K is the transformation constant. Combining (1) and (2) .. .. .. (3) K 1 Y Y In the two-step method of calibration the data are in the form of pairs of transmittances for the strong (s) and weak (w) steps, and the transmission ratio of the calibrating device (7) is known. lo& =- log(1- T ) - -log T I , I , 7 =--November, 19711 I N EMISSION SPECTROGRAPHIC ALAYSIS 763 Taking logarithms and substituting from (3) log7 = 2 log (isw) + ;log (2) .. - ’ (4) Y Substitution of the strong and weak-step transmittances into equation (4) results in an K 1 over-determined set of linear equations in the constants - and - The values of these constants, which minimise the sum of the squares of the residuals (S) are calculated by using a subroutine from the Harwell Subroutine Library. Y Y’ where M is the number of pairs of readings. etc., and the same method of determining the constants can then be used. to relative intensities. For the seven-step method the readings are divided into pairs, 1 and 2 , 2 and 3, 3 and 4, In calculating the ratio of intensities, equation (3) is used to convert the transmittances REFERENCES 1. 2. 3. 4. Franke, H., Poste, K., and Schmotze, W., 2. analyt. Chem., 1965, 212, 269. “Methods for Emission Spectrochemical Analysis,” Fifth Edition, American Society for Testing Kaiser, H., Spectrochim. Acta, 1951, 4, 351. Margoshes, M., and Rasberry, S. D., Ibid., 1969, 24B, 497. Materials, Philadelphia, 1968. Received November 16th, 1970 Accepted Jusze 14th, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600753

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

764 Analyst, November, 1971, Vol. 96, pp. 764-770 The Determination of Alumina in Iron Ores, Slags and Refractory Materials by Atomic-absorption Spectroscopy BY W. D. COBB AND T. S. HARRISON (Group Chemical Laboratories, British Steel Covporation, Scunthorpe Group, P.O. Box No. 1, Scunthorpe, Lincolizshire) As chemical methods of determining alumina are time-consuming, the simpler, more rapid atomic-absorption technique was adapted to iron ores, slags and refractory materials. The basic technique comprises fusion of the sample with a sodium tetraborate - sodium carbonate mixture, extraction with dilute nitric acid, appropriate dilution and spraying under recommended conditions for aluminium. As the flux materials enhance the aluminium absorption, sample and corresponding calibration solutions should contain equivalent amounts.Interference tests on synthetic solutions of the other constituents showed that the inhibiting effect of silica was suppressed by lanthanum chloride. A standard addition of lanthanum chloride was therefore incorporated and calibration conditions were developed to cover the ranges 0 to 1 per cent., 0 to 5 per cent., 0 to 10 per cent. and 10 to 35 per cent. of aluminium. Applications to miscellaneous B.C.S. materials showed good agreement with accepted values and reproducibilities of *0.035 per cent., fO.12 per cent., f0.18 per cent. and f0.38 per cent. of aluminium, respectively, a t the above levels. SENSITIVITIES reported for aluminium are largely dependent upon the type of flame and apparatus used and range from nil with an air - acetylene mixture1 to the 1 pg ml-1 obtained by Amos and Willis2 when using the nitrous oxide - acetylene flame.Experiments with oxygen-enriched air - a~etylene,~ fuel-rich oxygen - and other flames697 have met with reasonable success, but the nitrous oxide - acetylene flame is now generally preferred for its superior sensitivity and comparative safety in operation. Most workers prefer the 309.3 nm line as it appears to provide the best signal-to-noise ratio. Other factors governing sensitivity are the spectral intensity of the source, efficiency of the nebuliser coupled with burner design, and performance of the monochromator and amplification systems. Aluminium has been determined at concentrations of up to about 5 per cent. in a range of metallurgical materials9 and more recent work has covered the range 3-3 to 6.4 per cent. in cements.1° The particular application described in the present paper arose from a considerable increase in the demand for analyses of refractory materials; prior to the adoption of the atomic-absorption technique, chemical determinations of alumina in our works’ materials were made by the current B.S.procedure,ll which is tedious and time-consuming. Recently, more rapid new techniques published by the British Ceramic Research Asso- ciation have also been used with success, but it was soon realised that the input of samples could be matched only by the use of a method based on an even faster principle such as atomic absorption. The object of the present investigation, therefore, was to devise suitable procedures for determining alumina in home and foreign iron ores, iron and steel-making slags and various types of refractory materials over wide ranges of content.Performance in the higher ranges (about 20 to 60 per cent.) was of particular interest, and the results obtained compared favour- ably with those of the classical method. A considerable saving in time was also achieved. 0 SAC and the authors.COBB AND HARRISON 765 EXPERIMENTAL The instrument used was the Techtron Model AA-4 with the following conditions and settings- Wavelength 309.3 nm Setting Gauge reading 18 p.s.i. Lamp High spectral output Fuel Acetylene, cylinder pressure Current 8 mA 11 p.s.i. Slitwidth 100 pm Flow reading The observed reading for the Spectral band width 0.33 nm optimum flame conditions Flame Nitrous oxide - acetylene head* Scale expansion Ranges 1, 2 and 3: x 5; range Observation height 1 mm 4: none Support gas Nitrous oxide, cylinder pres- sure 36 p.s.i.was within the range 6 to 9 Burner A.B. 50 with grooved titanium Damping D * This burner is of special design, which considerably reduces carbon build-up. The stainless-steel nebuliser has a platinum - iridium capillary and a Teflon throat to reduce acid attack. MATRIX EFFECT- The borax - carbonate fusion followed by an acid extraction constitutes a rapid and effective means of achieving complete dissolution of the types of sample under consideration. but it was readily appreciated that the flux, consisting of alkali salts, would affect the absorp- tion reading.At the high temperature (about 3000 "C) of the nitrous oxide - acetylene flame some ionisation of the aluminium will occur, thus reducing the number of atoms in the ground state. The addition of a readily ionisable alkali metal is known to counter this effect by raising the proportion of non-ionised atoms, thus producing an enhancement of the absorbance, and so an experiment was designed to study the matrix effect. Solutions were prepared containing the various matrix constituents and 20 p.p.m. of aluminium. They were sprayed under the conditions given in the procedure. The recoveries are compared in the upper part of Table I. TABLE I THE EFFECT OF FLUX CONSTITUENTS IN VARIOUS DILUTIONS In 50 ml of solution containing 20 p.p.m. of aluminium- Aluminium found,* p.p.m.Constituents of test solution 5 ml of nitric acid (sp.gr. 1-42) . . .. .. .. a . 20.09 plus 1-0 g of sodium carbonate . . . . .. .. 22.57 plus 0-5 g of sodium tetraborate . . .. .. .. 23-54 plus 1.0 g of sodium carbonate plus 0.5 g of sodium tetra- borate . . .. .. . . .. .. .. 24.69 In 100 ml of solution containing 20 p.p.m. of aluminium- Aluminium Blank, p.p.m. of Constituents of test solution found,* p.p.m. aluminium A. 10 ml of nitric acid (sp.gr. 1.42) plus 2.0 g of sodium car- bonate plus 1.0 g of sodium tetraborate . . .. .. 25.22 1.58 B. 20 ml of A diluted to 100 ml . . .. .. .. .. 22.66 0.79 C. 10 ml of A diluted to 100 ml . . .. .. .. .. 21.67 0.59 ranges of aluminium content. aqueous solution. The dilutions A, B and C are those recommended in the Procedure to give the three lowest * The found aluminium contents are derived from a calibration graph prepared by using pure Further solutions containing different concentrations of sodium carbonate and borax, together with 20 p.p.m.of aluminium, were then prepared and sprayed. The amounts used corresponded with the dilutions used when calibrating for the three lowest ranges of aluminium content. The recoveries are compared in the lower part of Table I.766 COBB AND HARRISON: DETERMINATION OF ALUMINA I N IRON ORES, [AflUjySt, Vol. 96 These tests showed that sodium carbonate and sodium tetraborate enhance the alu- minium absorbance, the effect being greater when they are combined. Progressive dilution diminishes the effect. A slight increase due to nitric acid is also ob~erved.~ Hence the calibration solutions should contain the same concentrations of these reagents as the corresponding sample solution.EFFECTS OF OTHER MAJOR ELEMENTS- I Silicon dioxide . . .. Iron(II1) oxide.. .. Calcium oxide . . .. Manganese(I1) oxide . . Magnesium oxide . . Phosphorus pentoxide. . Chromium(II1) oxide . . Titanium(1V) oxide . . It was appreciated that interference tests should preferably be made over the lowest range of aluminium content, viz., 0 to 1.0 per cent., at which the concentration of salts would be at its highest. Additions of the other elements were therefore made to solutions having final concentrations of 10 per cent. v/v of nitric acid, 2 per cent. w/v of sodium carbonate, 1 per cent. w/v of sodium tetraborate and 20 p.p.m.of aluminium, and the solutions were sprayed. The recoveries are compared in Table 11. The results indicate an interference by silica, when it is present alone, which increases with concentration. This could be ascribed to the formation of aluminium silicate in the flame, thus reducing the number of aluminium atoms in the ground state and hence the absorption. This effect is counteracted by some elements when they are present, probably as a result of the Dreferential formation of other silicates, e.g., of calcium. TABLE I1 INTER-ELEMENT INTERFERENCE Range: 0 to 1.0 per cent. of aluminium Additions to the base solution,* per cent. w/w 12.0 8.0 30.0 30.0 - - - - 100.0 - - - 15.0 20.0 10.0 - - - - - - 90.0 - - 50.0 - - - - - - - 50.0 40.0 40.0 - 10.0 - - - -- - - 10.0 - - -- - - 20.0 - -- - - 50.0 - 6.0 2.0 10.0 - - 5.0 10.0 5.0 - - - 12.0 20.0 5.0 - - - - _ - - - - - - - - - Aluminium found,? p.p.m.19-41 19.32 19-41 16.44 19-32 19.49 19-49 19-67 13.73 19.49 20.03 19.55 * The base solution contained 10 per cent. v/v of nitric acid (sp.gr. 1-42), 1 per cent. w/v of t The found aluminium contents are derived from a calibration graph prepared by using the sodium tetraborate, 2 per cent. w/v of sodium carbonate and 20 p.p.m. of aluminium. base solution. When signal suppression occurs, a spectroscopic buffer is usually employed as a “releasing agent’’ to increase the ground-state atom population of the element concerned. Further tests were therefore devised, as in Table 111, to test the effect of adding lanthanum chloride solution to synthetic solutions of high silica content, which showed various degrees of suppression, depending on the nature and amount of boric oxide present.The recoveries of aluminium recorded on spraying with and without 2 ml of lanthanum chloride solution (= 0.2 g of lanthanum) indicated, in every instance apart from 100 per cent. of silica, the elimination of interference and a considerable improvement in the consistency of absorption. Regarding the matrix effect, the use of nitric acid in conjunction with the lanthanum solution resulted in an aluminium recovery of 20.61 p.p.m. ; the inclusion of sodium carbonate increased this value to 25.21 p.p.m., i.e., a little above the previous representative value for the flux contents reported in Table I.The lanthanum chloride addition was therefore incorporated in the method as described below, the calibration solutions being made up to contain the appropriate concentrations of lanthanum chloride and the flux reagents.November, 19711 SLAGS AND REFRACTORY MATERIALS BY ATOMIC ABSORPTION 767 TABLE I11 SUPPRESSION OF INTER-ELEMENT INTERFERENCES BY THE ADDITION OF Range: 0 to 1.0 per cent. of aluminium LANTHANUM CHLORIDE Aluminium found,? p.p.m. Additions to the base solution* with and without ‘ Without With 0.2 g of ’ lanthanum chloride lanthanum lanthanum 100 per cent. of silicon dioxide . . .. ,. .. .. .. 40 per cent. of silicon dioxide . . .. .. .. * . .. 40 per cent. of silicon dioxide plus 20 per cent. of iron(II1) oxide . . 40 per cent. of silicon dioxide plus 4 per cent.of calcium oxide . . 40 per cent. of silicon dioxide plus 40 per cent. of calcium oxide . . 40 per cent. of silicon dioxide plus 10 per cent. of manganese(I1) oxide 40 per cent. of silicon dioxide plus 10 per cent. of magnesium oxide . . 40 per cent. of silicon dioxide plus 10 per cent. of phosphorus pentoxide 13-53 18-04 17.25 17.61 10.06 17.90 19.04 16-00 19.64 20.41 20-41 20.41 20.21 20.79 20.41 20.41 * The base solution contained 10 per cent. v/v of nitric acid, 1 per cent. of sodium tetraborate, The found aluminium contents are derived from a calibration graph prepared from the 2 per cent. w/v of sodium carbonate and 20 p.p.m. of aluminium. base solution with and without added lanthanum. METHOD APPLICATION- slags, magnesite, silica brick, chrome ore, fire-brick and sillimanite.RANGE AND REPRODUCIBILITIES- The method is suitable for home and foreign ore, blast-furnace and basic open-hearth Aluminium per cent. r 1 Range Reproducibili ty 1. 0 to 1.0 f 0.035 2. 0 to 5.0 f 0.12 3. 0 to 10.0 &O-lS 4. 10.0 to 35.0 f 0.38 INSTRUMENT CONDITIONS- As under “Experimental. ” NOTE- In general the optimum burner height is found by raising the burner until it just begins to obstruct the light path. This position is indicated by a left deflection of the meter needle. The burner is then lowered by 1 mm. The optimum flame setting is found by starting with a fuel-rich flame and slowly reducing the acetylene flow until the white luminosity just disappears, leaving a red “feather” approximately 15 mm in height.REAGENTS- Sodium carbonate, anhydrous-AnalaR grade. Sodium tetraborate, fused-AnalaR grade. Nitric acid, 20 per cent. v/v, pure. Hydrogen peroxide-100-volume concentration, AnalaR grade. Laatkanum solution-Dissolve 26.7 g of analytical-reagent grade lanthanum chloride in Standard aluminium solution- water and dilute to 500ml. water and dilute to 100ml. 1 ml of solution = 0.1 g of lanthanum. (a) For the lower ranges-Dissolve 0.8792 g of potassium aluminium sulphate in 1 ml of solution = 100 pg of aluminium. (b) For the highest range-Dissolve 1 g of pure metal in 20 ml of hydrochloric acid (sp.gr. 1-18) and cautiously evaporate to dryness. Add a further 10ml of nitric acid (sp.gr. 1.42) and again cautiously evaporate to dryness. Add 10 ml of nitric acid (sp.gr.1-42), digest until dissolution is complete, cool and dilute to 1 litre. Store solutions in polythene containers when appropriate. 1 ml of solution = 1000 pg of aluminium.768 COBB AND HARRISON: DETERMINATION OF ALUMINA IN IRON ORES, [ATUdyst, VOl. 96 PREPARATION OF CALIBRATION SOLUTIONS- 1. Range 0 to 1 per cent. of aluminium-To each of six 150-ml conical beakers transfer 2 g of sodium carbonate and 1 g of sodium tetraborate and add 50 ml of 20 per cent. v/v nitric acid. Heat to dissolve solids and expel carbon dioxide. Cool and add 2 ml of lanthanum solution. Add 0, 4, 8, 12, 16 and 20 ml of standard aluminium solution (a) (= 0, 0.2, 0.4, 0.6, 0.8 and 1.0 per cent. of aluminium) and dilute to 100ml in calibrated flasks. Store in polythene containers.2, 3 and 4. Ranges 0 to 5, 0 to 10 and 10 to 35 per cent. of aluminium-Dissolve 5 g of sodium carbonate and 2-5 g of sodium tetraborate in 125 ml of 20 per cent. v/v nitric acid and heat to expel the carbon dioxide. Cool, add 5 0 m l of lanthanum solution and dilute to 250ml for the base solution. (2) Range 0 to 0.5 per cent. of aluminium-To 20-ml fractions of the base solution add 0, 4, 8, 12, 16 and 20 ml of aluminium solution (a) (3 0, 1, 2, 3, 4 and 5 per cent. of alu- minium) and dilute to 100ml in calibrated flasks. (3) Range 0 to 10 per cent. of aluminium-To 10-ml fractions of the base solution add 0, 4, 8, 12, 16 and 20 ml of aluminium solution (a) (= 0, 2, 4, 6, 8 and 10 per cent. of alu- minium) and dilute to 100 ml in calibrated flasks. Store in polythene containers.(4) Range 10 to 35 per cent. of aluminium-To 20-ml fractions of the base solution add 4, 6, 8, 10, 12 and 14 ml of aluminium solution (b) (= 10, 15, 20, 25, 30 and 35 per cent. of aluminium) and dilute to 100 ml in calibrated flasks. Store in polythene containers. Store in polythene containers. PREPARATION OF SAMPLE SOLUTION- Mix 0.2 g of sample with 2 g of sodium carbonate and 1 g of sodium tetraborate in a small platinum capsule. Melt over a Meker burner and, when dissolution is apparent, raise to full heat (900 to 1000 "C) and maintain this temperature for about one minute (Note 1). Cool, place the capsule in a 150-ml tall-form beaker containing 50 ml of 20 per cent. v/v nitric acid warmed to a temperature not exceeding 80 "C, add two or three drops of hydrogen peroxide (100-volume) and swirl to effect dissolution.Cool, transfer to a 100-ml calibrated flask, add 2ml of lanthanum solution and dilute to the calibration mark with water. This solution serves for the range 0 to 1 per cent. of aluminium (Note 2). For the higher ranges, dilute as follows- Range Dilution 0 - 5 per cent. of aluminium 0 - 10 per cent. of aluminium 10 - 35 per cent. of aluminium NOTES- 20 ml to 100 ml 10 ml to 100 ml 20 ml to 100 ml 1. Some refractory materials may require an additional period of 5 to 10 minutes in a muffle 2. For samples containing up to 40 per cent. of silica complete dissolutjon should be achieved. With silica bricks some silica hydrolyses out of solution so, after dilution to 100 ml, the solution furnace maintained a t 1100 "C.is filtered through a Whatman No. 41 paper preparatory to spraying. DETERMINATION OF ALUMINIUM- Set the instrument according to the table of instrument conditions. Lower ranges-Spray the appropriate calibration solutions, followed by the sample Spray water between each test and set to 100 per cent. Plot absorption (per cent .) against element concentration, draw a calibration graph and solutions. transmission. read off the percentage of aluminium. Repeat this procedure. Percentage of aluminium x 1.8895 = alumina, per cent. NOTE- For greater accuracy the system used under "High range" can be used. High range-Spray the sample solution and select two calibration solutions containing concentrations of aluminium immediately above and below those present in the sample.November, 19711 SLAGS AND REFRACTORY MATERIALS BY ATOMIC ABSORPTION 769 4 8 12 16 20 Aluminium concentration,p.p.m. Fig.1. Aluminium calibration curve Then spray the sample and calibration solutions several times, spraying water between tests and setting to 100 per cent. transmission. From the mean absorbance readings calculate the percentage of aluminium in the sample in relation to each calibration solution. Convert the results into percentage of alumina. ACCURACY AND REPRODUCIBILITY TRIALS- The above procedure was then applied to a series of British Chemical Standards including home and foreign ore, blast-furnace and basic open-hearth slag, magnesite, silica brick, chrome ore, firebrick and sillimanite. TABLE IV ALUMINA IN IRON ORES, SLAGS AND REFRACTORY MATERIALS: ACCURACIES AND REPRODUCIBILITIES (LOWER RANGES PROCEDURE) B.C.S.No. 302 Northants Type of material iron ore Alumina, per cent., 7-24 certificate value Alumina, per cent., 7-07 by atomic-absorption 7-26 spectroscopy 7-22 7.26 7-07 7-30 7-36 7.34 7.37 7.17 Mean 7.24 Reproducibility 0.22 (95 per cent.) f 2 s 301 Lincs iron ore 4-26 4.19 4.07 4.22 4-15 4.19 4.19 4.15 4-24 4-34 4.02 4.18 0.18 367 Blast- furnace slag 20.0 20.18 20.00 20.27 20.00 20.18 20.37 20.18 19-80 20-09 20.00 20.1 1 0.33 17511 Liberian iron ore 1.10 1.12 1.08 1.08 1.09 1.1 1 1.06 1.11 1.08 1-09 1-06 1.09 0.04 17412 319 Basic Mag- slag nesite 0.77 1.00 0.78 1.01 0.72 0-96 0.7 1 0.93 0.68 0.91 0-72 0.97 0.71 0.92 0.69 0.98 0-73 1.00 0.71 0.97 0.70 0.93 0.72 0.96 0.05 0.07 314 Silica brick 0.77 0.79 0.74 0.72 0-75 0-79 0.75 0.79 0-77 0.75 0.7 7 0.76 0.05770 COBB AND HARRISON For the blast-furnace slag a further calibration point was prepared containing the equivalent of 12 per cent.of aluminium, although the range 0 to 10 per cent. of aluminium will cover modern blast-furnace practice. A typical calibration graph is shown in Fig. 1. Determinations on each sample were carried out on ten separate occasions. The results shown in Tables IV and V were obtained. Comparison of the certificate values with the corresponding mean values found by atomic- absorption spectroscopy shows an acceptable degree of accuracy for the method, whilst the reproducibilities compare favourably with those of existing chemical procedures. Six determinations can be completed in less than three hours, compared with about three days by the standard hydroxyquinoline chemical procedure.TABLE V ALUMINA IN IRON ORES, SLAGS AND REFRACTORY MATERIALS: (HIGH RANGE PROCEDURE) ACCURACIES AND REPRODUCIBILITIES B.C.S. No. Type of material Alumina, per cent., certificate value Alumina, per cent., by atomic-absorption spectroscopy Mean Reproducibility (95 per cent.) f 2 s 269 Firebrick 33.9 33.35 32.85 33.7 33.95 33.75 33.65 34.0 33.55 33.3 33.95 33.61 0.72 309 Silliinanite 61.1 61.1 61-3 61.05 61.45 61-75 61-15 61-0 61-6 61.6 61.6 61.36 0-55 308 Chrome ore 19.4 19.7 19.3 19-55 19.8 19.65 19.5 19-4 19-6 19.7 19.6 19.58 0.30 The authors thank Messrs. W. W. Foster and T. E. Clayton of this Department for useful discussion and the management of the Scunthorpe Group , British Steel Corporation, for permission to publish this paper. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. REFERENCES Gilbert, P. T., Analyt. Chew., 1962, 34, 210R. Willis, J. B., Nature, 1965, 207, 715. Amos, M. D., and Thomas, P. E., Analytica Chim. Acta, 1965, 32, 139. Slavin, W., and Manning, D. C., Analyt. Chem., 1963, 35, 253. Elwell, W. T., and Gidley, J. A. F., “Atomic Absorption Spectrophotometry,” Second Edition, Pergamon Press, Oxford, London, Edinburgh, New York, Toronto, Sydney, Paris and Braun- schweig, 1966, p. 76. Chakrabarti, C. L., Lyles, G. R., and Dowling, F. B., Analytica Clzim. Ada, 1963, 29, 489. Slavin, W., “Atomic Absorption Spectroscopy, ’’ Interscience Publishers, New York, 1968. Nikolaev, G. I., and Aleskovskh, V. B., Zh. Analit. Khim., 1963, 18, 816. Capacho Delgado, L., and Manning, D. C., Analyst, 1967, 92, 553. B.S. 1121 : Part 43, British Standards Institution, London, 1963. , , , Ibid., 1963, 28, 392. --- Received April Ist, 1970 Amended March 19fh, 1971 Accepted July 13th, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600764

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

Analyst, November, 1971, Vol. 96, &5. 771-775 771 The Atomic-fluorescence Determination of Mercury by the Cold Vapour Technique BY K. C. THOMPSON AND G. D. REYNOLDS (Shandon Southern Instruments Ltd., Frimley Road, Camberley, Surrey) A method for the determination of low levels of mercury by using atomic fluorescence in conjunction with the cold vapour technique is described, and shows significant advantages over the corresponding absorption technique. THE determination of mercury in solution by using atomic absorption in conjunction with the cold vapour technique has been reported by many Manning' has published a review on the subject. The methods utilise the fact that mercury is the only element (other than the inert gases) that has an appreciable vapour pressure at room temperature, and whose vapour is practically wholly monatomic.Mercury also has a very low affinity for oxygen, and therefore a relatively high concentration of mercury atomic vapour can be maintained in air at room temperature. Because of its relatively high vapour pressure, no thermal energy from a flame (or other source) is required for the vaporisation and atomisation of elemental mercury. At 20 "C, 1 litre of air saturated with mercury contains 14 pg of mercury. Most methods generate mercury vapour by first reducing it to the elemental state, e.g., with tin(I1) chloride or sulphate, and then blowing air through the solution to expel the mercury. The stream of air is passed through a cell and the absorption of the 253.7-nm mercury resonance line emitted by a low pressure mercury lamp is monitored.To improve the sensitivity, various methods are used to concentrate the vapour. In one method, air is continuously circulated in a closed system until all of the mercury is liberated before an absorption reading is taken.4s6 This procedure can be time consuming, and also requires the use of a drying column to remove traces of water vapour, which can cause slight spurious broad band absorption at 253.7 nm. Traces of organic vapours can also cause broad band absorption. The use of automatic background correction by measuring the broad band absorption with a hydrogen lamp at 253.7 nm can overcome these problems6 as long as the broad band absorption at 253.7 nm is not too great. Other methods liberate the mercury by using the same principle, but the mercury is then absorbed on a suitable substance such as iodised charcoals or ~ i l v e r .~ The absorbing substance is heated and the mercury evolved is carried into a suitable absorption cell and the absorption at 253.7 nm is monitored. This method is time consuming and errors can be caused by organic contaminants being absorbed by the absorbent and subsequently driven off, causing broad band absorption at 253-7 nm. The fluorescence technique shows distinct advantages over the absorption technique in that it is simple, there is no enclosed cell (hence no fogging of the cell windows), no spurious signal from broad band absorption by organic contaminants, no re-circulation system or drying column required, the sensitivity is high, and the calibration graph is linear over a wide concentration range.EXPERIMENTAL A Southern Analytical A3000 atomic-absorption - emission spectrophotometer with the burner and lenses removed is used. A low-pressure OZ4W Philips mercury lamp is mounted in a suitable housing in the fluorescence position, so as to minimise specular reflection from components mounted near the monochromator. It is essential that the specular reff ection is minimised. The light from the mercury lamp is directed over the top of a 10-mm i.d. Pyrex tube, which is connected to a mercury generation system. The apparatus is shown in Fig. 1. Mercury is liberated in cell A, and air or argon is bubbled through this cell and through expansion chamber B. This tendsto smooth out the flow and also prevents any accidental carry-over of corrosive solutions caused by sudden excessive gas flows into the 0 SAC and the authors.772 THOMPSON AND REYNOLDS : THE ATOMIC-FLUORESCENCE DETERMINATION [Artdyst, Vol.96 instrument compartment. The optimum size of the expansion chamber was found to be approximately 200 ml. The mercury passes up through tube C and fluoresces in region D. It is not necessary to use a drying column. The fluorescence signal is displayed on a Bryans Series 27 000 recorder. The measurement procedure was as follows. The reducing agent [tin(II) chloride, hydrochloric acid and sulphuric acid] was added to cell A and argon was bubbled through this solution until a stable base-line on the recorder was obtained. The mercury solution, containing between 0.01 and 2 p g of mercury, was then added and the fluorescence signal monitored on the recorder.in Fig. 1. Schematic diagram of mercury fluorescence detector OPTIMISATION OF OPERATING PARAMETERS- Damping-The A3000 was operated at maximum damping (the time constant was 5 seconds). This usually gave well defined peaks, but occasionally sharp spikes were observed above the smooth trace. The recorder was further damped by using an RC network attached to the recorder terminals (R = 12 kR, C = 400 pF, time constant = 4.8 seconds). With the extra damping, reproducible peaks as shown in Fig. 2 were obtained. The recorder output of the A3000 was set to 0 to 25 mV full-scale deflection and the recorder to 0 to 10mV full-scale deflection. Choice ofgas-For similar amounts of mercury, the peak height with argon was approxi- mately four times greater than that with nitrogen and thirty-five times greater than that 100 50 Time--+ Typical peaks obtained for mercury by using spectral band pass, 6 nm; 0-06 pg (1 ml containing 0.06 p.p.m.) of mercury added to cell A; and argon a t a flow-rate of 1.4 1 min-l Fig.2.November, 19711 OF MERCURY BY THE COLD VAPOUR TECHNIQUE 773 with air. This is obviously caused by the quenching of excited mercury atoms in the 3P, state by nitrogen and oxygen molecules. Oxygen is a far more efficient quenching agent than nitrogen because the excited mercury atoms and the ground-state oxygen molecules are both in the triplet state, thus facilitating energy transfer. The optimum flow-rate was 1.4 1 min-l for air, nitrogen and argon.The fluorescence signal was not very dependent on the gas flow-rate. At low flow-rates a spiky trace was obtained and at high flow-rates a decrease in peak height was observed. Argon was used in all further measurements unless otherwise stated. Choice ofgas bubbling arrangement-To increase the rate of uptake of the mercury from the solution, a glass sinter was sealed across the end of the gas inlet to cell A (Fig. 1). This should have increased the solution - gas interface area appreciably. However, an increase in the blank value, caused by fine droplets of solution being carried through the system and causing specular reflection, was observed. On adding a known amount of mercury, the glass sinter increased the signal by 5 per cent.This showed that an efficient uptake of mercury occurred without the sinter, and that all the sinter did was to saturate the gas with water vapour and also cause a fine mist of droplets to be carried up to region D. The sinter was not used in further studies. Throughout this work, no condensation was observed in ex- pansion chamber B, which indicated that when the sinter was removed the gas was not saturated with water vapour. The diameter of tube C was fairly critical, the optimum internal diameter being 1 cm. When this tube was sheathed with argon, only a very slight increase in signal was observed, which indicated that little mixing with air occurred in region D. The sheath was not used in further measurements. Lamp current-The optimum lamp current was found to be 0.32 A.Reagents for mercary generation-Various systems have been reported in the literature. In this study the following reagents were used. Solution 1 was a 10 per cent. w/v solution of tin(I1) chloride in 5 N hydrochloric acid, and solution 2 was a mixture of 2 N nitric acid and 10 N sulphuric acid. The mercury solutions were freshly made before use from a stock solution containing 1000 p.p.m. of mercury that was prepared by dissolving 1 g of mercury in 50 ml of nitric acid and diluting it to 1 litre with distilled water. All reagents were of analytical grade. RESULTS For mercury concentrations between 0.01 and 2 p.p.m., 10ml of solution 1, 20ml of solution 2 and 20 ml of distilled water were added to cell A and argon was bubbled through until a stable base-line was obtained.This procedure ensured that any traces of mercury, either in solutions 1 and 2 or accidentally transferred to cell A, would be removed from the system. This effective elimination of the blank from the reagents was important when working at maximum sensitivity. Between 0.1 and 1 ml of the mercury solution was added to cell A, the peak recorded (Fig. 2), and the peak height measured. The recorder returned to the base-line within 5 minutes of adding the sample and the next sample could then be added directly to the contents of cell A. This procedure could be repeated until the total volume reached 60 ml. [The use of argon prevents the oxidation of tin(I1) salts in cell A by atmospheric oxygen.] For mercury concentrations below 0.01 p.p.m. the same procedure was followed except that between 5 and 10 ml of the sample were added to cell A and the solution in the cell was discarded after only one or two measurements, i e ., when the total volume reached 60ml. For the determination of mercury in urine samples, the use of potassium permanganate to break down any protein mercury present would be nece~sary.~,~ In the limited time avail- able these procedures could not be tested. However, standard additions of mercury(I1) were made to acidified urine samples (5 ml of urine and 1 ml of nitric acid), which were then added to cell A. To prevent foaming, a trace of silicone anti-foaming agent was added to the cell. The response was linear with respect to the amount of mercury added, and as little as 0.003 pg of mercury could be detected in a 5-ml sample of urine.The calibration graphs for mercury in air, nitrogen and argon are shown in Fig. 3. EFFECT OF ORGANIC SOLVENTS- The presence of 2 per cent. v/v of acetone in cell A (which, in an original 1-ml sample added to the 50 ml in cell A, would represent 100 per cent. of acetone in the original sample)774 THOMPSON AND REYNOLDS : THE ATOMIC-FLUORESCENCE DETERMINATION [Artahst, vol. 96 did not affect the base-line, and with a 0.3-pg sample of mercury (1 ml = 0-3 p.p.m. of mercury) gave a 33 per cent. reduction in peak height when argon was used. With 5 per cent. of acetone in cell A, the base-line was not affected, but the peak height for a 0.3-pg sample of mercury was reduced by 60 per cent. This decrease was thought to be caused by the quenching of excited mercury atoms by acetone molecules.When air was used instead of argon, no change in the base-line was observed until 5 per cent. of acetone was present in the cell. A 2-pg sample of mercury (1 ml = 2 p.p.m. of mercury) and 2 per cent. of acetone in cell A did not affect the peak height, while the presence of 5 per cent. of acetone in cell A gave only an 8 per cent. decrease in peak height. It would appear that the quenching effect of nitrogen and oxygen molecules had minimised any additional quenching from the low concentration of acetone vapour in the gas. 70 .s 60 2 50 2 f 40 - 30 = 20 10 =I + .- m --. L 0) al A! (TI .- 0.2 0.4 0.6 0 8 1.0 1-2 1.4 1 Amount of mercury added to cell A/pg Fig. 3. Mercury calibration graphs at a spectral band-pass of 6 nm and a gas flow-rate of 1.4 1 min-1 for: A, argon; B, nitrogen; and C, air The presence of 2 per cent.of ethanol, chloroform or benzene in cell A did not affect the base-line when argon or air was used. However, when argon was used, the signal for 1 pg of mercury was reduced by 25,22 and 93 per cent., respectively, and when air was used, the corresponding figures were <5, <5 and 60 per cent. The large decrease with benzene and air is thought to be caused by relatively strong absorption of the 253.7-nm radiation over the small path lengths of region D (Fig. 1). Although the presence of organic solvents gave a negligible change in the base-line, the absolute signal magnitudes were decreased appreciably when argon was used.It appears that air would be the better choice of carrier gas when organic solvents are known to be present, although this would reduce the sensitivity considerably (Fig. 3). Alternatively, the method of standard additions could be used with argon. An absorption system similar to that of Hatch and Ott4 was constructed. A cell with a 12-cm path length was used in conjunction with a re-circulating pump. The mercury generation system was similar to that shown in Fig. 1. A l-pg sample of mercury gave an absorbance of 0.27. The presence of 2 per cent. of acetone in cell A (no mercury present) corresponded to 0 4 p g of mercury, while 2 per cent. of benzene (again with no mercury present) gave complete absorption of the radiation at 253.7 nm. PRECISION- A mixture of 10 ml of solution 1, 20 ml of solution 2 and 20 ml of distilled water was placed in cell A and argon bubbled through it until a stable base-line was obtained.Then 1 ml of a 0.06-p.p.m. mercury solution (containing 0.06 pg of mercury) was added to the cell, the peak recorded and, when the recorder had returned to the base-line, another lml of the solution containing 0.06 p.p.m. of mercury was added and the peak recorded. This was repeated until ten readings (spread over a period of 90 minutes) had been obtained. The relative standard deviation calculated from the peak height was 5 per cent.November, 19711 OF MERCURY BY THE COLD VAPOUR TECHNIQUE 775 SENSITIVITY- The noise on the base-line gave a 3a detection limit of 0.002 pg of mercury. The steady blank signal (caused mainly by specular reflection) corresponded to a peak fluorescence signal of 0.004 pg of mercury.It was essential to position the source so that this specular reflection signal was minimised. The peak height on adding 1 ml of distilled water to cell A corresponded to less than 0-0005pg of mercury. The sensitivity with the A3300 atomic- absorption - fluorescence - emission spectrophotometer, which has improved amplification characteristics, was a factor of 2.5 times better. It is difficult to compare the sensitivity of the fluorescence technique with that of the absorption technique because the sensitivity of the former is directly dependent on the source intensity and on the detector sensitivity. Hatch and Ott4 used a re-circulation system to obtain a steady absorbance reading of 0.29 for 1.1 pg of mercury.Lindstedt,5 who used a straight-through system with small volumes, obtained a limit of detection of 0-002 pg of mercury in a l-ml sample. The present system yielded a 30 detection limit of 0.002pg of mercury. CONCLUSIONS The determination of mercury in solution by using atomic fluorescence in conjunction with the cold vapour technique is a simple, sensitive and specific method for determining low levels of mercury. The main advantages over the corresponding absorption technique are : its simplicity, as no re-circulation system or drying column is required; the fluorescence is not viewed from an enclosed cell so that the cell window cannot fog and the memory effect is minimised; and background correction facilities to compensate for broad band absorption are not required. Broad band (molecular) absorption will be recorded as mercury in the absorption technique unless background correction facilities are used. In the fluorescence technique , any broad band absorption (which should be small because of the low path lengths used) will not be recorded as mercury. Even more important, molecular fluorescence at the 253.7-nm exciting wavelength is unlikely to occur with most common volatile organic compounds. We thank the Directors of Shandon Southern Instruments Ltd. for permission to publish this work and Mr. R. C. Rooney for many helpful suggestions. REFERENCES 1. 2. 3. 4. 5. 6. 7. - , Ibid., 1970, 9, 97. 8. 9. Ballard, A. E., and Thornton, C. W. D., Ind. Engng Chem., Analyt. Edn, 1941, 13, 893. Poluektov, N. S., and Vitkun, R. A., Zh. Analit. Khim., 1963, 18, 33. Poluektov, N. S., Vjtkun, R. A., and Zelyukova, Y . V., Ibid., 1964, 19, 873. Hatch, W. R., and Ott, W. L., Analyt. Chem., 1968, 40, 2085. Lindstedt, G., Analyst, 1970, 95, 264. Manning, D. C., Atontic Absorption Newsletter, 1970, 9, 109. Moffitt, A. E., and Kupel, R. E., Ibid., 1970, 9, 113. Kalb, G. W., Ibid., 1970, 9, 84. Received March 2nd, 1971 Accepted June 23rd, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600771

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

776 Analyst, November, 1971, Vol. 96, pp. 776-778 X-ray Spectrometric Determination of Rare Earth Elements by Using a Fusion Technique BY C. PLOWMAN (Central Electricity Generating Board, Scienti3c Services Department, Kirkstall, Leeds, LS4 2HB) A general method for the determination of rare earth elements by X-ray fluorescence Spectrometry is described. The sample is fused with sodium tetraborate, chromium is added to act as internal control-standard, and the resulting bead is analysed directly. This method overcomes variations in sample form and particle size. Individual rare earths can be determined at levels from 0-1 to 100 per cent., and by using the L spectra and a lithium fluoride 220 crystal, line overlap is kept to a minimum. X-RAY fluorescence spectrometry has been used to determine rare earth elements in so1ution,lp2p3 catalysts4 and ores,5 and traces of rare earths in high-purity oxides.6 The purpose of this work was to develop a general method for the analysis of lanthanum, cerium, praseo- dymium, neodymium, samarium, europium and gadolinium in a wide range of materials including rare earth ores, alloys containing a high percentage of a rare earth, and mixed oxides.Analysis of alloy steels and nickel-base alloys containing rare earths was not attempted. The use of a fusion technique reduces sample preparation to a minimum as solid samples can be fused directly. Various fluxes were tried, including sodium tetraborate, lithium tetra- borate, lithium tetraborate - lithium carbonate in various proportions, and potassium pyro- sulphate.Sodium tetraborate was found to be at least as satisfactory as other fluxes, and had the advantage of being readily available in a pre-fused form. EXPERIMENTAL APPARATUS- A Philips 1220 semi-automatic spectrometer was used with the following conditions : the X-ray tube target was gold and the tube required 50 kV at a current of 40 mA; the crystal was a lithium fluoride 220 crystal; the collimator opening was 160pm, the X-ray path was through a vacuum; the counter was flow-proportional with a counting time of 40 s; and automatic pulse height analysis was used. It is possible to use either the K or L spectra of the rare earths for quantitative analysis. Even though the lithium fluoride 220 crystal gives reasonably good resolution of the K spectra, the resolution of the L spectra is much better.No advantage is gained by use of the K spectra from the point of view of intensity because it has been shown that, for those rare earths heavier than praseodymium, greater intensity is obtained from the L lines (under vacuum conditions) .' Preliminary work showed that the use of an internal control-standard8 was desirable; a suitable one for the rare earth L lines is the chromium Ka line.112 This compensates satisfactorily for variations in the sample preparation technique by different operators, and serves for all the rare earth lines used (Table I). TABLE I ANALYTICAL LINES Spectral line Lanthanum LcY.. .. .. Cerium La . . .. . . Praseodymium L/3 . . .. Neodymium L/3 . . .. Samarium L/3 . . .. .. Europium L/3 .. .. .. Gadolinium L /3 .. .. Chromium Kcc . . .. .. Line "28 138-70 128.11 104.97 99-02 89-10 84-78 80.86 107.11 Background "28 137.60 127.00 106.00 98.00 90.00 83.90 53-90 - 0 SAC and the author.PLOWMAN 777 Some of the La lines were subject to line overlap but in most instances the intensity of the L/3 line was equal to, or greater than, the intensity of the La line. This is because of the greater absorption in the X-ray path and flow-counter window of the La compared with the L/3 line. It is therefore more marked if an air path is used. The only serious line overlap occurs when holmium is present (the holmium La line is very near to the gadolinium Lp line) so that a correction must be made for holmium. PROCEDURE- Mix thoroughly 1-Og of sample, 0.10g of chromium(II1) oxide (Cr,O,) and 15g of sodium tetraborate and place them in a 95 per cent.platinum-5 per cent. gold crucible. Particular care must be taken to weigh the chromium(II1) oxide accurately. The crucible is heated a t 1100 "C in a muffle furnace for 30 minutes, with swirling at 5-minute intervals. The mixture is then poured into a graphite mould at about 400 "C on a hot-plate and covered with a graphite lid. The mould is allowed to cool slowly to room temperature (for about 1 hour), and its contents are then analysed directly on the spectrometer by using the settings mentioned above. Standards are made up by mixing individual rare earth oxides in various proportions. The mixes are dissolved in hydrochloric acid, precipitated as oxalates and ignited at 1000 "C for 1 hour.They are immediately weighed (because they pick up moisture and carbon dioxide very rapidly) and fused as for a sample. RESULTS AND DISCUSSION Table I1 shows the comparison between the total rare earth analysis obtained by sum- ming the results for the individual rare earths and that obtained by chemical analysis of some typical samples. TABLE I1 COMPARISON OF X-RAY FLUORESCENCE AND CHEMICAL RESULTS Total rare earth oxides, per cent. I 1 Sample By X-ray fluorescence By chemical method Rare earth - iron silicide 1 . . .. .. 30.8 31.0 Rare earth - iron silicide 2 . . .. .. 39.1 39-8 Rare earth - calcium silicate slag 1 . . .. 25.6 24-8 Rare earth - calcium silicate slag 2 . . .. 31.7 31-7 Ore concentrate .. .. .. .. 70.6 70.9 Mixed oxides . , .... .. .. 98.9 99.8 The standard deviation was calculated from 126 separate determinations, in which the concentrations ranged from 0.1 to 100 per cent., and was found to be 0.123. Thus the 99-7 per cent. probability (30) is h0.369. The main factor causing errors in the repeatability of the results is the sample preparation. No significant difference was found between the determinations of the separate elements. On analysing a number of routine samples, it was invariably found that any lack of agreement between duplicates resulted from an inhomogeneous melt. LIMIT OF DETECTION- This was taken as the concentration yielding a line intensity that exceeds the back- ground by three standard deviation^.^ Calculated on this basis, the detection limit was found to be about 0.02 per cent., although for lanthanum and cerium it was about 0.01 per cent. The author thanks the London and Scandinavian Metallurgical Co. Ltd., Rotherham, in whose laboratories this work was carried out.778 PLOWMAN REFERENCES Rekholainen, G. I., and Ivanova, L. I., Zav. Lab., 1968, 34, 172. Rekholainen, G. I., Ibid., 1965, 31, 442. Tertian, R., “Advances in X-ray Analysis,” Volume 12, Plenum Press, New York, 1969. Stone, I. C.. and Rayburn, K . A., Analyt. Chem., 1967, 39, 356. Lytle, F. W., Botsford, J. I., and Heller, H. A . , Re$. Invest. U.S. BUY. Mines, 1957, No. 5378. Lytle, F. W., and Heady, H. H., Analyt. Chem.. 1959, 31, 809. Payne, K. W., in “Proceedings of the 5th Conference on X-ray Analytical Methods,” Philips, Bertin, E. P., “Principles and Practice of X-ray Spectrometric Analysis,” Plenum Press, New Jenkins, R., Hurley, P. W., and Shorrocks, V. M., Analyst, 1966, 91, 395. Eindhoven, 1966. York, 1970. Received March 12th, 1969 Amended January 14th, 19 7 1 Accepted July lst, 1971 1. 2. 3. 4. 6. 6. 7. 8. 9.

ISSN:0003-2654    

DOI:10.1039/AN9719600776

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

Analyst, November, 1971, Vol. 96, @. 779-784 779 The Use of a Shield for the Reduction of Fogging on Photographic Plates in Spark-source Mass Spectrography BY C. W. FULLER AND J. WHITEHEAD ( Tioxide International Limited, Billingham, Teesside) In spark-source mass spectrography, secondary effects cause pronounced fogging of the photographic plate in the vicinity of the isotope lines of the major elements, and the limit of detection of elements whose lines fall within this area is considerably raised. A shielding device has been designed to reduce the amount of ions produced by secondary effects and to prevent them from reaching the photo- graphic plate. This reduces the background in the region of the major isotope lines and enables elements in this region t o be determined with increased sensitivity and precision.ALTHOUGH the photographic plate is widely used in spark-source mass spectrography for recording spectra, it has several disadvantages. The method involving its use is slow, the intensity range over which the plate has a linear reponse is limited and the high background in the vicinity of the isotope lines of the major elements restricts the measurement of adjacent trace-element concentrations. The evaluation of mass spectra has been speeded up by the introduction of automatic microdensitometry with coupled punch tape and computer facili- ties,l but the problem of high photoplate background remains. The introduction of electrical detection2J has eliminated the photographic plate and made it possible to produce directly a semi-quantitative record of mass spectra.Rapid quantitative determination of elements within a mass range of M to 2M for one setting of the magnet current is also made possible by using electrical detection and peak switching. This method is useful when a small number of elements has to be determined quantitatively, but it becomes less advantageous as the number of elements to be determined increases2 and the mass range is extended. A very high degree of resolution is also required to discriminate instrumentally between interfering spectra, and the eye can more easily resolve such interferences on a microdensitometric trace from a photographic plate. In these circumstances the photographic plate is still a useful technique. The method would be considerably improved by eliminating the background interference in the vicinity of the isotope lines of the major elements and a simple device to achieve this end is described in this paper.The general background of the photographic plate under recommended conditions of operation is satisfactory. It is only in the region of the isotopes of the major elements, where secondary ionic species are produced in large numbers, that it becomes unacceptable. This caused by the geometry of in that environment, which latter effect is species behave the instrument and the way in which charged is illustrated in Fig. 1 as follows. Fig. 1. Causes of secondary emission fogging on photographic plates: a, b and c are described in the text 0 SAC and the authors.780 FULLER AND WHITEHEAD: SHIELD FOR REDUCTION OF FOGGING [A%dyst, vol.96 (a) As an ion beam strikes the photoplate a small fraction of the ions is reflected. The energy of these ions is considerably reduced and the ions are rapidly deflected by the magnetic field so that they reach the photoplate in a slightly higher mass position. (b) A situation arises, when a large ion current strikes the photoplate and raises the potential of the emulsion surface, where electrostatic repulsion of some of the ions occurs. These ions follow a path of small radius and return to the photoplate. (c) Some of the ions leaving the photoplate may not return to the photoplate by the routes described in (a) and (b) because they collide with metal parts of the instrument in the neighbourhood of the photoplate. In doing so, they release secondary electrons, which are attracted to the photoplate by localised charge accumulations. X-rays and luminescence are also produced when the positive ion beam strikes the photographic plate.All of these factors combine to give the characteristic halo effect on photographic plates around the lines of the major components. Several attempts have been made to reduce the problem of fogging, the mode of attack falling into two categories: removal of the factors contributing to fogging, and special treatment of the photographic plate at the manufacturing and emulsion development stages. METHOD 1- (a) The segment of the photographic plate and plate holder where the major component normally strikes is removed.*y5y6 In this way the ion beam due to the major component passes through the aperture and none of the interferences described above occur.The disadvantages are numerous, however, and are as follows: (i) each photoplate must be individually cut to size; (ii) the photoplate holder is modified in such a way that it can be used only for the analysis of one particular type of sample; (iii) the two pieces of the photoplate must be accurately aligned in the holder, which will be particularly important when determination of an element of mass number close to that of the major element is required; and (iv) the photoplate holders are no longer light-tight and precautions have to be taken when loading and unloading the plate holders. (b) An earthed metallic strip is placed over the segment of the emulsion struck by the ions of the major component to avoid build-up of positive charge on the photographic plate.59' This technique is not very successful and results are not as good as those obtained by method 1 (a).(c) A suppressor plate is fitted at a high positive potential to reduce the effects of secondary electron^.^ This removes only the least important cause of fogging. METHOD 2- Kennicott899 investigated the rBle of photographic developers in spark-source mass spectrography and their relationship to photoplate background. By using a surface developer to obtain more reproducible results he found that the photoplate exhibited less fogging than usual. The surface developer resulted in an image of less intensity but this was compensated for by a large increase in signal-to-noise ratio.He also used an internal image developer and obtained a substantial reduction in background and an improvement in the reproducibility of the developing process. There was again a reduction in the absolute intensity of the image, but the increased signal-to-noise ratio compensated for this effect. CavardlO has also reported similar results. METHOD 3- Some work11 has been carried out on photoplates containing a layer of a conducting oxide between the glass plate and the emulsions. The oxide layer is intended to conduct away excess charge so that fogging of the plate does not occur. Although a reduction in back- ground is obtained, the plates have not been used other than for experimental purposes, because of manufacturing difficulties. The methods described above have their inherent advantages and disadvantages, no one method being completely satisfactory.Thus the methods 1 (a), (b) and (c) show no decrease in intensity of signal but are limited in use and are not very practical. The techniques of method 2 are more widely applicable and, in general, show a greater reduction in background.November, 19711 ON PHOTOGRAPHIC PLATES I N SPARK-SOURCE MASS SPECTOGRAPHY 781 They also show a reduction in the absolute signal obtained and require a stricter control of the developing technique. The work described here falls into the list of modifications given under method 1. Fig. 1 shows in diagrammatic form the way in which fogging can occur on the high mass side of a high- intensity ion beam. If a shield is placed on the photographic plate as shown in Fig.2, then the primary ion beams of the intense line and the weak lines can still reach the photographic plate, but the secondary ion beams now strike the shield and do not reach those parts of the photographic plate which are of interest. As no secondary ions reach the photographic plate, fogging effects on the high mass side should be removed. Fig. 2. Effectiveness of inserting a small shield a t an angle of about 45' to the photographic plate to remove the causes of secondary emission fogging EXPERIMENTAL An MS7 double-focusing instrument of the Mattauch-Herzog type* was used. Ilford QZ photographic plates were used for recording the mass spectra, and a microdensitometert was used for measuring line intensities in the spectra.Electrode preparation and instrument parameters have been described previously.12 DISCUSSION The main problem is to design a shield that can be fitted into the limited space available. It is important that the shield and its fixtures do not protrude beyond the dimensions of the plate holder because otherwise the plate holder itself will not fit into its outer cover, p c (a) Fig. 3. Modified MS7 photoplate holder and single shield (a) and double shield (b) * A.E.I. Scientific Apparatus Limited. t Joyce-Loebl and Company Limited, Model RlliITIC.782 FULLER AND WHITEHEAD: SHIELD FOR REDUCTION OF FOGGING [Afidyst, Vol. 96 0.2 4 20 60 4 30 3 30 + c m aJ r Y m p. .- 0 1 Fe Mn , Cr ~ W - U L L u 0.2 4 10 60 1.5 30 3 30 ' 1 u w u " 1 ' ' 0.15 4 6 60 3 30 3 30 Ex posu re/nC Fig.4. Microdensitometer scans of 56Fe, 55Mn, 52Cr and 51V in a titanium dioxide matrix: (a), with no shield; (b), with single shield; and (c), with double shield. The exposures recorded on the photoplate were 60, 40, 30, 20, 10, 6, 4, 3, 2, 1.5, 1, 0.6, 0-4, 0.3, 0.2 and 0.15 nC. The longest and shortest exposures measured on the microdensitometer are shown; - - - represents the background level due to foggingNovember, 19711 ON PHOTOGRAPHIC PLATES IN SPARK-SOURCE MASS SPECTROGRAPHY 783 and that the position of the shield should be adjustable to allow greater versatility and to enable minor adjustments to be made to compensate for any variation in magnet current that would cause a shift in the position of the ion beams. Fig. 3 shows the modified MS7 plate holder with the single shield in place.The design of the double shield is also shown. The essential features of the modification are as follows. The sides of the photoplate holder have been lowered over a length of about 50mm to a depth of 1 mm to accommodate a piece of 0.79-mm brass sheet. A small hole has been drilled at a position coinciding approximately with the major isotope lines of titanium on both sides of the holder, and each hole has been tapped to take a small screw. The shields have been made from 0.79-mm brass sheet to the designs illustrated. The choice of 0.79-mm (&inch) brass sheet gives sufficient rigidity to the shield. The thickness also represents about one half of the distance between two adjacent mass numbers on the photographic plate, thus allowing some leeway in the position and angle of the shield. The shield is made so that it subtends an angle of approximately 45" to the photographic plate.The exact shield dimensions and angle are found by experiment. The shield is held in place by the two screws. The metal strip at the inside front end of the outer case for the photoplate holder is Thin photoplates are used without the aluminium backing plate. The design enables the position of the shield to be varied over a range of about 20 mm, which corresponds to 16 mass units about a mean mass number of 50 (magnet current 248 mA). Although it has not been necessary in the present work it should be possible to cover a much larger mass range, which could be achieved by increasing the length of the modified part of the photoplate holder and by drilling a series of holes at regular intervals along the edges of the photoplate holder.To adjust the shield to the correct position, insert an exposed photoplate of the type of sample to be analysed into the holder, and slide the shield along until the base edge of the shield just comes to the high mass side of the major isotope line to be shielded. Tighten the two retaining screws. Remove the exposed plate and insert an unexposed photoplate. Spark a sample in the normal way and then develop the photoplate. Any slight adjustments to the position or angle of the shield will be apparent from the exposed plate. After making the adjustments, repeat the above procedure to check that the shield is functioning satis- factorily. One or two plates exposed in this manner are normally sufficient to fix the position of the shield.No further adjustments will be necessary unless there is any drift in the magnet current. Removal of the two screws and the shield leaves the photoplate holder available for any other analyses that may be required. The shield has been applied to the determination of trace amounts of vanadium, chro- mium, manganese and iron (mass numbers 51,52,55 and 56, respectively) in titanium dioxide (titanium mass numbers 46 to 50 inclusive). Without the shield there is a very high level of background in the required mass range 51 to 56 [Fig. 4 (a)]. With the use of the single shield some improvement in the observed background is obtained [Fig. 4 ( b ) ] , but it has not been completely eliminated.As there are five titanium isotope lines involved, the single shield is obviously inadequate in shielding all of them and some secondary ions arising from the titanium-46, titanium-47 and, possibly, titanium-48 lines are still able to pass the shield and cause an increase in background. To overcome this a modification was made by fixing a second shield to the first shield [Fig. 3 ( b ) ] in such a way that direct shielding was applied at the titanium-48 and titanium-50 lines. In this way, the background on the high mass side of the titanium lines was almost completely eliminated [Fig. 4 ( c ) ] . Normally, a single shield should give adequate results but for shielding several lines a double-shield arrangement is clearly superior.The result of the work reported above is, therefore, to increase the signal-to-noise ratio with no corresponding loss of absolute intensity. The modifications outlined are simple and, if required, the photoplate holders can be used normally after removal of the shield. Finally, and most important, once the shield has been positioned no change is required in the normal mode of operation used for photographic detection. removed.784 FULLER AND WHITEHEAD The authors acknowledge the technical assistance of Mr. B. R. Ellis and Mr. A. R. Collins in the manufacture and design of the photoplate shield. This work is published by permission of the Directors of Tioxide International Limited. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. REFERENCES Degreve, F., and Champetier de Ribes, D., Int. J . Mass Spectrom. Ion Phys., 1970, 4, 125. Hull, C. W., Ibid., 1969, 3, 293. Bingham, R. A., Powers, P., and Wolstenholme, W. A., ASTM, 17th E14, Dallas, 1969, p. 261. Ahearn, A. J., and Malm, D. L., AfiPZ. Spectrosc., 1966, 20, 411. Mai, H., “Advances in Mass Spectrometry,” Volume 3, Elsevier Publishing Co., Amsterdam, -, J . Scient. Instrum., 1966, 42, 339. Hannay, N. B., Rev. Scient. Instrum., 1954, 25, 644. Kennicott, P. R., Analyt. Chem., 1965, 37, 313. Cavard, A., “Advances in Mass Spectrometry.” Volume 4, Elsevier Publishing Co., Amsterdam, “Proceedings of the Sixth Annual MS7 Mass Spectrometer Users’ Conference,’ Associated Elec- Jackson, P. F. S., and Whitehead, J., Analyst, 1966, 91, 418. 1966, p. 163. -, Ibid., 1966, 38, 633. 1968, p. 419. trical Industries, Scientific Apparatus Department, Manchester, March, 1966, p. 23. Received February 1 Sth, 197 1 Accepted July lst, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600779

出版商:RSC    

年代:1971

数据来源: RSC

摘要:

Analyst, November, 1971, Vol. 96, $$. 785-797 785 Automatic Radiof requency Titration of Acids in Tertiary Butyl Alcohol- Acetone Medium" BY W. J. SCOTT AND G. SVEHLA (Defiartment of Analytical Chemistry, The Queen's University, Belfast) The mixed solvent t-butyl alcohol - acetone can be used as a differen- tiating medium for the titration of acids with a standard solution of tetra- n-butylammonium hydroxide dissolved in a mixture of toluene and methanol. Titration curves of various shapes can be obtained and the simultaneous determination of certain acids achieved. The differences in curve shapes obtained by radiofrequency titration can be used as the basis of separation. The low toxicity of t-butyl alcohol (compared with other solvents at present in widespread use for the titration of weak acids) combined with its low solvating power and good differentiating properties make it a suitable medium for the routine determination of such acids in admixture.THE use of t-butyl alcohol as a non-aqueous solvent for the titration of acids was first investigated by Fritz and Marple,l who noted the low acidity of the compound compared with methanol or ethanol and its good solvent properties, and pointed out its potential uses as an amphiprotic solvent. Other workers reported on its use in potentiometri~,~s3s~~V condu~timetric~~~ and vis~a1~9~ titrations. It has been observed that this alcohol possesses differentiating properties for weak acids, which are attributed to the high degree of dissociation of ion-pairs, and to the fact that no appreciable association of the acid anion with free acid occurs in the solvent.1,2s3s495s6 No thorough investigation has so far been reported on the differentiating titrations of carboxylic acids in the solvent, although an investigation on phenols has shown some encouraging resulks Radiofrequency end-point detection has not so far been reported.The differentiating effect of the alcohol, coupled with the advantages of radiofrequency end-point detection, offers considerable selectivity and precision for the titration of mono- and dibasic acids and mixtures of these acids. The low dielectric constant (10-6) of the pure solvent imposes some limitations on instrument response, because the operational frequency of the oscillator (of which the titration cell forms an integral part) must be rather low if reasonably high admittance is to be achieved.The radiofrequency titrimeter des- cribed by Scott, Quigg and SvehlalO is equipped with a continuously variable operational frequency covering the range 0.5 to 30 MHz and could be applied to monitor such titrations at the lower end of this frequency range. Experiments showed that the addition of about 20 per cent. v/v of acetone to the solvent increases the dielectric constant slightly (to 13.5). The optimum operational frequency for such media, which is about 2 MHz, is safely within the working range of the instrument. The presence of acetone does not interfere in the resolution of mixtures of acids as it is not itself a levelling solvent.ll To ensure that the solvent does not freeze, 10 per cent.v/v of toluene has to be added also. If the operational frequency is correctly selected, instrument response is satisfactory and there is no super- imposition of background noise on the titration curves. With tetra-n-butylammonium h y d r 0 ~ i d e ~ ~ ~ ~ ~ , ~ ~ ~ ~ ~ s ~ ~ ~ ~ 7 as a titrant dissolved in a mixture of toluene and methanol, sharp, well defined end-points can be obtained. EXPERIMENTAL REAGENTS- (9 + 1 v/v).t 24th to 28th, 1969. Titrant-A 0.1 M solution of tetra-n-butylammonium hydroxide in toluene - methanol This solution is standardised against analytical-reagent grade benzoic acid * Presented at the National Conference of Analytical Chemistry, Mamaia, Romania, September t This solution is available from B.D.H. Chemicals Ltd. 0 SAC and the authors.786 SCOTT AND SVEHLA: AUTOMATIC RADIO FREQUENCY TITRATION OF [ArtdySt, VOl.96 with radiofrequency end-point detection. The titrant must be protected from contamination with carbon dioxide and moisture by the use of a drying tube containing soda-lime and calcium chloride. Solvent-t-Butyl alcohol was dried by the method of Lund and Bjerrumls by using ethyl bromide as a catalyst. As the pure alcohol just melts at room temperature (m.p. 25.5 “C), a volume of analytical-reagent grade toluene sufficient to give a 1 + 9 mixture was added immediately after distillation so as to depress the melting-point; as toluene is used as a solvent for the titrant, it does not interfere in the titration. To raise the dielectric constant of the solvent 20 per cent.by volume of anhydrous, analytical-reagent grade acetone is added to the solvent immediately prior to titration. The acid blank of the solvent was often checked and was found to be less than 0.001 milli-equivalents per 50ml. During the titration an atmosphere of pure dry nitrogen was maintained over the solution in the cell (described below) so as to exclude carbon dioxide. If titrations are completed within a few minutes, this precaution can be omitted except when the compounds are sensitive to atmospheric oxygen (as are, for example, dihydric phenols). Acids and fiheutoZs-The acids and phenols were, whenever possible, of analytical-reagent grade ; if these were not available, general-purpose grades were used. APPARATUS- The radiofrequency titrimeter described earlier by the authorslo was used. An automatic burette and a T - Y chart recorder were added to the basic instrument, thus making the whole set-up completely automatic.A Radiometer ABU l c piston burette of 25-ml capacity and a Sunvic pen recorder were used. The volume of dispensed titrant is read on a digital counter of the burette. To slow down the rate of delivery and to increase the precision of volume reading, we used the slow gear system (normally used with the 0-25-ml burette tube) and the faster counting rate; therefore, the full delivery of 25.00ml corresponded to a reading of 100-00 on the counter. The rate of delivery was about 2 ml min-1 and the chart speed was usually 5 cm min-l. A special switch was built into the burette, in parallel with the push- button by which the burette could be operated manually, which activated both the burette and the chart-delivery motor (T-axis) of the recorder.The optimal operational frequency for these titrations was 1-9 MHz. A low, but not zero, beat frequency was adjusted when commencing the titration and the variation of beat frequency was plotted on the Y-axis of the chart recorder. The titration cell is a thin-walled glass cylinder, 5 cm in diameter, with two sheets of copper foil (each 2 x 7 cm) acting as capacitor plates, which were held firmly to the outside by Sellotape. The Quickfit stopper is equipped with three outlets, to accommodate an “Electrothermal” stirrer, a gas inlet tube and a combined burette inlet - gas outlet tube. The polythene outlet tube of the burette, 0.2 cm in diameter, enters the cell through the burette inlet - gas outlet orifice and terminates in a fine 5-cm long glass capillary.The tip of the capillary dips below the surface of the solution to be titrated. Photographs of the cell and of the instrument panel are shown in Figs. 1 and 2. PRECISION OF VOLUME MEASUREMENT- With the automatic titration system used in these investigations, the accuracy of the results depends mainly on the precision with which the volume is measured. This measure- ment is made in the following way: the starting position of the pen is marked on the chart and the titration carried out beyond the end-point. The final position of the pen is marked again on the chart and the corresponding figure, a, on the digital counter is read.(Because of the gear setting used, the volume of titrant delivered is Qaml.) The titration chart is then examined, the end-point determined graphically and the full I , distance between starting and final position as well as the I , distance from the start to the end-point are nieasured with a ruler. The V , volume corresponding to the end-point is then calculated from the equationFig. 1. View of the capacitive cell complete with nitrogen supply Fig. 2. View of the instrument panel of the radiofrequency titrator [To face page 786November, 19711 ACIDS IN TERTIARY BUTYL ALCOHOL - ACETONE MEDIUM 787 By applying the law of propagation of errors to these measurements, the s,/V coefficient of variation of volume measurements can be expressed as where s,/a is the coefficient of variation of the burette delivery.We determined this quantity by operating the burette to the same digital read-out several times and determining the volume of the liquid delivered by weighing it. This s,/a coefficient of variation was found to be &0-0017. The coefficient of variation of chart travel, SJZ, was determined by operating the burette again to the same digital read-out several times and by measuring the distance travelled by the chart. This s,/Z coefficient of variation was found to be *O.OOS. By using these values in equation (1) the coefficient of variation of volume measurement was found to be SV - = ,t0.009 or k0.9 per cent. V which is the degree of reproducibility that one can expect in the final results. With a different burette and a different chart recorder a different level of reproducibility can be expected.PROCEDURE- Accurately weigh a sample containing about 0.3 milli-equivalents of active ingredient and transfer it into the titration cell with 50ml of solvent. (This amount will consume 2 to 3 ml of titrant up to the first equivalence-point.) Add 10 ml of acetone and adjust the stirring rate so that the vortex does not extend below the top of the capacitor plates. With the operational frequency set at 1.9 MHz, tune the reference oscillator with the capacitor control to give a low beat frequency. After filling the burette with the titrant and adjusting the chart recorder, start the titration by operating the start switch. At a suitable time after the equivalence-point, stop the titration by using the same switch.Determine the volume of the titrant corresponding to the end-point from the titration graph and calculate the result. STANDARDISATION OF THE TITRANT- (equivalent weight, 122.13). By following the procedure given above, titrate analytical-reagent grade benzoic acid RESULTS AND DISCUSSION TITRATION OF SINGLE ACIDS- The groups examined were (i) inorganic acids, (ii) monocarboxylic aliphatic and aromatic acids, (iii) dicarboxylic aliphatic and aromatic acids and (in) phenols. (i) Inorganic acids-Hydrochloric, sulphuric, nitric and phosphoric acids were investi- gated, mainly for the sake of comparison. The titration curves were easy to evaluate. Hydrochloric and nitric acids give titration curves of the “elbow” type, the end-point being marked by the change in slope of the curve.Sulphuric acid gives a “chair”-shaped curve Group Aliphatic Aromatic TABLE I TITRATION OF MONOCARBOXYLIC ACIDS Acid (phenolic OH) Formic 3.76 - Acetic 4.76 - Hexanoic 4.88 - G1 ycollic 3.83 - Lactic 3-08 - Benzoic 4.19 - Mandelic 3.85 - Anthranilic 6-97 - Salicylic 2.97 13-40 3-Hydroxybenzoic 4-00 9.92 4-H ydrox ybenzoic 4.48 9.32 Number of end-points 1 1 1 1 1 1 1 1 1 1 1 Shape of curve Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow788 SCOTT AND SVEHLA: AUTOMATIC RADIO FREQUENCY TITRATION OF [ArtdySt, VOl. 96 with two changes in the slope corresponding to the two equivalence-points. Titration curves of phosphoric acid show one sharp and a second less distinct end-point, which obviously correspond to the neutralisation of the first and second acidic hydrogen atoms, respectively. These results were, however, not quantitative and reproducible, because the strong acids react readily with t-butyl a1cohol.l The titration of strong acids is, of course, possible in aqueous solutions.1° (ii) Aliphatic and aromatic monocarboxylic acids-The titration curves of these acids consist of two lines (sometimes slightly curved) of different slopes, the end-point corresponding to their intersection.Organic hydroxy-acids fall into two categories. Some of these acids, mainly aliphatic acids but also salicylic acid, show only one end-point, corresponding to the carboxyl group only, while others, mainly phenolic acids but excluding salicylic acid, show two end-points with a chair- shaped titration curve corresponding to the neutralisation of the carboxyl and phenolic hydroxyl groups.The monobasic behaviour of salicylic acid has been reported by other workers.lg Results are summarised in Table I. The pK values (corresponding to aqueous solutions) are shown for the sake of comparison.20 The high value given for the phenolic hydroxyl group of salicylic acid is in accordance with the fact that it shows only one end-point. These elbow-like titration curves can easily be evaluated. (b 1 t 1 c a, 3 u 2 n t + m 0 t-Volume of titrant/mi Fig. 3. (a) N-curve and (b) chair-curve for the homologous series of aliphatic dicarboxylic acids (iii) Aliphatic and aromatic dicarboxylic acids-The lower aliphatic dicarboxylic acids, both saturated and unsaturated, show two very sharp end-points, which correspond to the neutralisation of the first and second acidic hydrogen atoms.These titration curves resemble the letter N [see Fig. 3 (a)]. As one proceeds along the homologous series, the strength of these acids decreases and the difference in the pK values of the first and second dissociations decreases also (cf. Table 11), and therefore the sharpness of the two end-points decreases as well. The N-shaped titration curve changes into a chair-like trace [see adipic acid, Fig. 3 ( b ) ] , then the sharpness decreases further until an elbow-type titration curve is obtained in which only the second end-point is clearly marked [see pimelic and sebacic acids; cf. Fig. 4 (a) and ( b ) ] . The sharpness of end-points and the pK values of the first and second dissociations (referred to aqueous media) are in a similar correlation also with aromatic dicarboxylic acids.November, 19711 ACIDS IN TERTIARY BUTYL ALCOHOL - ACETONE MEDIUM TABLE I1 TITRATION OF DICARBOXYLIC ACIDS Number of Group Acid pK, (H,O) pK, (H,O) end-points Aliphat j c Oxalic 1-23 4.19 2 Malonjc 2.83 5.69 2 Succinic 4-16 5.6 1 2 Glu taric 4.34 5-41 2 Pimelic 4-71 5.43 2 Suberic 4-52 5.41 2 Azelaic 5.30 6-40 2 1 Sebacic 5-59 - Maleic 1.83 6.07 2 Fumaric 3.08 4.44 2 I tacon j c 3-85 5.45 2 Malic 3-40 5-11 2 Aromatic Phthalic 2.89 5.51 2 Adipjc 4.43 5-41 2 Isophthalic 3.62 4.60 2 Terephthalic 3.54 4.46 Insoluble 789 Shape of curve N N N N Chair Chair Chair Chair Elbow N N N N N Chair - These results obtained for the titration of dicarboxylic acids illustrate well the advantages of the solvent system.While the trend in the sharpness and separation of end-points, as discussed above, has been noted also for other sharpness and separation deteriorate more quickly in most of these solvents than in t-butyl alcohol as the chain length of the compound investigated increases. The explanation of this phenomenon lies in the fact that the solvent has a very low solvating power,21 which is a distinct advantage in differen- tiating titrations. It is difficult to make more quantitative comparisons, or predictions of the behaviour of acids in other solvents, because the dissociation constants of these acids and the ionic conductivities of their dissociation products in non-aqueous solvents are generally not known. 0 - Volume of titrant/rnl 0 Fig.4 (a) Intermediate-curve and (b) elbow- curve for the homologous series of aliphatic dicarboxylic acids790 SCOTT AND SVEHLA: AUTOMATIC RADIO FREQUENCY TITRATION OF [AlzdySt, VOl. 96 ( i v ) Phenols-The phenols investigated, as distinct from phenolic acids, which were dis- cussed under (ii), yield elbow-shaped titration curves in all instances apart from the trihydric phenol phloroglucinol, which gave a titration curve with three distinct end-points correspond- ing to the three hydroxyl groups. The shape of the titration curve resembles the letter N but has a horizontal ending (N-). Results are shown in Table I11 and are in agreement with low-frequency conductimetric results.2s TABLE 111 TITRATION OF PHENOLS Group Compound Monohydric Phenol 2-Nitrophenol 3-Nitrophenol 4-Nitrophenol 2-Bromophenol 3-Bromop hen01 4-Bromophenol 0-Cresol m-Cresol #-Cresol 2-Phenylphenol 2,PDinitro -1-naphthol Picric acid 2,3-Xylenol 2,PXylenol 2,5-Xylenol 2,B-Xylenol 3,4-Xylenol 3,5-Xylenol Resorcinol Quinol Dihydric Catechol Trihydric Phloroglucinol PK, (H2O) 9.99 7-23 8.38 7-15 8.39 8.87 9-34 10.26 10.04 10.20 10.01 0-29 10.50 10.45 10.21 10.59 10.32 10.15 9.12 9.15 9.91 8.45 ( Number of pK, (H20) end-points 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 12.08 1 11-32 pK, (H,O) 1 12-03 1 - - - - - - - - - - - - - - - - - 8-88 (?I 3 Shape of curve Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow Elbow - TITRATION OF MIXTURES OF ACIDS- The titration of binary mixtures of the acids mentioned was also attempted.Instead of classifying them according to chemical structure we classified these acids according to the shape of titration curves (N-acid, chair-acid and elbow-acid), as preliminary experiments showed that resolution of acids by such titrations is dependent mainly on the form of indi- vidual titration curves. Mixtures that gave good resolution can be classified into four groups : T T 0 f--- Volume of titrandm1 Fig. 5. Differentiating titration of a mixture comprising 3-nitrophenol and 2,6-xylenolNovember, 19711 ACIDS IN TERTIARY BUTYL ALCOHOL - ACETONE MEDIUM 791 (i), elbow plus elbow; (ii), elbow Plas N; (iii), elbow plus chair; and (iv), N plus chair.Other combinations, such as N plus N, etc., were tested but did not yield good resolution and are not mentioned here. (i) Elbow plus elbow mixtures-Mixtures of compounds characterised by elbow curves were resolved without difficulty. Fig. 5 shows the results obtained with the titration of a mix- ture of 3-nitrophenol and 2,6-xylenol. (The break in the middle of the curve originates from the fact that as the titration proceeded and the beat frequency became very low, it was felt necessary to increase the sensitivity of the beat-frequency meter, that is, to “magnify” the frequency axis of the titration graph. Information on the resolution of such titrations is shown in Table IV. None of the ternary mixtures investigated gave good resolution, which is in agreement with the results of potentiometric titrations,l although other workers* were able to resolve the isomers of cresol.The volume axis remains the same.) TABLE IV TITRATION OF MIXTURES OF ELBOW-ACIDS (PHENOL MIXTURES) Mixture 4-Nitrophenol 2,G-Xylenol . . 3-Ni trophenol 2,G-Xylenol . . 4-Bromophenol 2,G-Xylenol . . 2,4-Xylenol . . 2,G-Xylenol , . 2,3-Xylenol . . 2,6-Xylenol . . $-Nitrophenol %Nitrophenol 3-Nitrophenol 2-Bromophenol 3-Bromophenol 4-Bromophenol 3,5-Xylenol . . 3,4-Xylenol . . 2,6-Xylenol . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. PI( (H2O) 7.15 10.59 8.39 10.59 9.34 10.59 10.45 10.59 10.50 10.59 7.15 7.23 8.38 8.39 8-87 9.34 10.35 10.32 10.59 A PK 3.44 2.21 1-25 0.14 0.09 0-08 1.15 0.45 0-47 0.17 0.37 Results Two end-points Two end-points Two end-points Two end-points Two end-points One end-point One end-point One end-point Results in Table IV indicate that resolution might be expected if the difference between the aqueous pK values of the acids involved is larger than 0.1. Such rules are usually drawn with potentiometric titrations, for which the shapes of the titration curves are closely and simply related to the equilibrium constants of the reactions involved.The correlation between curve shape and equilibrium constants in radiofrequency titrations is much more complicated because factors such as ionic mobilities, relaxation times and the variation of dielectric constant during the titration influence the actual shape of the titration curve.An attempt to express the resonance frequency of the oscillator, i.e., a quantity that differs only by a constant from the variable monitored in these titrations, as a function of hydrogen- ion concentration was made by K. I. Millar of this Department. The very complex equation (covering several pages) is not given here, as its practical application requires a special computer programme. Details of such a calculation will be published shortly in a Ph.D. thesis.** (ii) Elbow plus N mixtures-When titrating elbow and N-acids together, two types of titration curves are obtained. If there is complete resolution, as with malonic and acetic acids (Fig. S), three end-points are shown on the titration curve. Experiments with various amounts of the two acids showed that the first end-point corresponds to the neutralisation of the first acidic hydrogen atom of malonic acid, the second to that of acetic acid and the third to the neutralisation of the second acidic hydrogen atom of malonic acid. This order corresponds to the order of aqueous pK values of these acids.Although the order established by experiments generally followed this rule, there were some notable exceptions, e.g., salicylic792 SCOTT AND SVEHLA: AUTOMATIC RADIO FREQUENCY TITRATION OF [Autalyst, vol. 96 Volume of titrant/ml ' Fig. 6. Complete resolution of a mixture of N- and elbow-type acids. Malonic + acetic acids and malonic acids showed end-points in an unexpected order (first the salicylic acid and then the two malonic acid end-points).Other pairs of elbow fibs N-acids showed only two end-points. The general rule is that initially the first acidic hydrogen atom of the dibasic acid and the monobasic acid are titrated together, while the difference between the second and first end-points corresponds to the neutralisation of the second acidic hydrogen atom of the dibasic acid. This was found to be true even if the aqueous pK values suggested a different pattern, e.g., titration of oxalic and sebacic acids together. Predictions from aqueous pK values cannot reliably be made; titration of salicylic and glutaric acids, for example, gave only two end-points, although from the pK values one would expect three end-points. In any event, if mixtures of acids are titrated, the order of end-points must be established experimentally. Table V shows the pairs of acids that can be titrated, together with aqueous pK values and the numbers of end-points on the titration curves.TABLE V TITRATION OF AN ELBOW-TYPE AND AN N-TYPE ACID TOGETHER Elbow-acid N-acid Compound Salicylic Lactic Formic Glycollic Mandelic Benzoic Acetic Hexanoic Anthranilic Sebacic o x j ~ i c 2 2 3 2 2 2 3 2 3 2 1-23 4-19 Maleic 3 2 2 2 3 3 3 3 3 3 1.83 6.07 Malonic 3 3 3 2 3 3 3 3 3 3 2.83 6.69 Phthalic Fumaric Malic 2 2 2 3 2 3 2 2 .- 2 2 1 9 2 2 2 3 2 3 2 2 3 2 2 3 3 2 3 2 2 3 2.89 3.03 3.40 6-51 4.44 5.11 9 Itaconic 2 2 2 2 2 3 2 3 3 3 3.85 5.45 Succinic 2 2 2 2 2 2 3 3 3 2 4-16 5-61 Glutiric 2 3 2 2 2 2 2 2 2 3 4.34 5.41 2 G two end-points on the titration curve. 3 E three end-points on the titration curve.November, 19711 ACIDS IN TERTIARY BUTYL ALCOHOL - ACETONE MEDIUM 793 I - 2 m l i -Volume of titrant/rnl Fig.7. Complete resolution of a mixture of chair- and elbow-type acids. 3-Hydroxybenzoic + glycol- lic acids (iii) Elbow plus chair mixtures-Mixtures of elbow plus chair-acids show complete reso- lution in some instances, that is, three end-points are found on the titration curves, as with glycollic and 3-hydroxybenzoic acids (see Fig. 7). The order of end-points follows the order of the aqueous pK values. Other mixtures show only two end-points and their order follows the rule mentioned under (ii), that is, generally the first end-point corresponds to the neutralisation of the first acidic hydrogen atom of the dibasic acid together with the mono- basic acid, while the second end-point corresponds to the neutralisation of the second acidic hydrogen atom of the dibasic acid.Mixtures with adipic acid yield only two end-points, although according to the aqueous pK values one would expect three. Mixtures of anthranilic and isophthalic acids cannot be resolved as they yield only one end-point corresponding to the total acid content of the system. Results of titrations are summarised in Table VI. TABLE VI TITRATION OF AN ELBOW-TYPE AND A CHAIR-TYPE ACID TOGETHER Elbow-acid Compound Salicylic . . Lactic . . Formic . . Glycollic . . Mandelic . . Benzoic . . Acetic . . Hexanoic .. Anthranilic. . Sebacic . . PKS (H*O) . . 2.97 . . 3-03 , . 3.75 . . 3-83 . . 3.85 . . 4-19 . . 4.76 . . 4.88 . . 5-00 .. 5.59 PKa1 ( H a PKaz (H@) 3-H ydroxybenzoic 3 3 3 3 3 2 2 2 2 2 4.06 9.92 Chaj r-ad d A 3 Adipic 4-Hydroxybenzoic Isophthalic 2 3 2 2 3 3 2 3 2 2 3 2 2 3 2 2 2 2 2 2 2 2 2 2 2 2 1 2 2 2 4-43 4.48 3.62 5.41 9.32 4.60 2 E two end-points on the titration curve. 3 = three end-points on the titration curve. (iv) N plus chair mixtzlres-Most of the pairs formed from N- and chair-acids show four end-points giving total resolution. Fig. 8 shows the titration curve for a mixture of malic acid and 3-hydroxybenzoic acid. The four end-points follow generally the order794 SCOTT AND SVEHLA: AUTOMATIC RADIO FREQUENCY TITRATION OF [AflabSt, VOl. 96 3-Hydroxybenzoic acid ( 1 1 V' c c- Volume of titrant/ml 0 Fig. 8. Complete resolution of a mixture of N- and chair-type acids.3-Hydroxybenzoic + malic acids expected from the aqueous pK values. In a few instances, when two pK values were close together, only three end-points were obtained when the two corresponding acidic hydrogen atoms were titrated together. Mixtures with fumaric acid resulted in two end-points only, one when the first acidic hydrogen atoms of both acids were titrated together, followed by a second end-point when the second acidic hydrogen atoms were titrated. Unfortunately, therefore, the two acids cannot be resolved at all, the titration curve giving information only on the total acid content. In some instances the order of end-points does not follow the order predicted from the aqueous pK values. Thus, a mixture of maleic and 3-hydroxybenzoic acids was titrated in the order of maleic (l), 3-hydroxybenzoic (l), 3-hydroxybenzoic (2) and maleic (2), while one would expect the last two end-points to appear in the reverse order.The good resolving power of the solvent can again be emphasised here. While the mixture of oxalic and adipic acids, for example, yields four end-points (that is, giving complete resolu- tion), in pyridine only one end-point occurs.z5 Results of titrations are shown in Table VII. TABLE VII TITRATION OF A CHAIR-TYPE AND AN N-TYPE ACID TOGETHER N-acid Compound Oxalic Maleic Malonic Phthalic Fumaric Malic Itaconic Succinic Glutaric PKai (H2O) ~ K a 2 W2O) 1-23 4-19 1.83 6.07 2.83 5.69 2.89 5.51 3.03 4.44 3.40 6-11 3.85 5-45 4-16 5.61 4-34 5-41 PKai (HP) ~ K a 2 (H2O) Chair-acid 3-Hydroxybenzoic Adipic 4-Hydroxybenzoic Isophthalic r A 7 4 4 4 3 4 4 4 4 4 4 4 4 4 3 4 4 2 2 2 2 4 3 4 4 4 4 4 4 4 4 4 3 4 3 4 4 4-06 4.43 4.48 3-62 9-92 5-41 9-32 4.60 4 = four end-points on the titration curve.3 = three end-points on the titration curve. 2 = two end-points on the titration curve.November, 19711 ACIDS I N TERTIARY BUTYL ALCOHOL - ACETONE MEDIUM 795 TITRATION OF TERNARY MIXTURES- In the course of this study attempts were made to titrate ternary mixtures consisting of N-, chair- and elbow-acids. The titration curves resembled with each attempt those obtained with a mixture of an N- and a chair-acid, the elbow-acid being titrated together with one of the acidic hydrogen atoms of the other acid. In some instances, therefore, the amounts of all of the component acids could be determined from the four end-points, while in other instances, when only three end-points were obtained, only one of the acids could be deter- mined separately, with the sum of the other two.Even in the most favourable case, i.e., when four end-points were obtained, one of the acids is determined indirectly from a difference of two volumes so that we did not pursue this line of approach, the error in the indirect determination being always considerably higher than that in a direct determination. A comprehensive study of ternary mixtures would have required the examination of a large number of combinations, but because of the limitations in accuracy such a study has not been undertaken. ACCUFUCY AND PRECISION- The accuracy of the determination of each acid was checked by titrating various amounts of the acid.In one range ten different samples were taken, the titrations were carried out, and the proportionality between sample weight and volume was checked by calculating the correlation coefficient and the intercept of the volume ‘ueysus weight curve with the volume axis, which is supposed to be zero. The standard deviation of the intercept was calculated and the presumption that the intercept is different from zero was tested with 95 per cent. statistical significance. The correlation coefficient was in all instances better than (+)0.99, and the intercept was found to be zero in each instance. The slope of these curves is propor- tional to the equivalent weight of the acids, and these slopes were calculated for all samples.An “absolute” test of the accuracy, based on stoicheiometry, was not possible, as the titrant was obtained already made up, and the standardisation with benzoic acid presupposes the stoicheiometry in these reactions. Evidence from our work and other experiments, however, leaves no doubt about the stoicheiometric nature of such reactions. The precision of these titrations was checked by six parallel measurements by titrating approximately equal amounts of the acids taken from the middle of the useful range of sample sizes. As all of the samples had to be weighed with a swinging balance by the method of swings, these samples could not be completely identical, and therefore the ratios of the weights calculated from the end-points and the weights of the samples were determined (this should be exactly unity) and the standard deviation of these ratios was also determined.As benzoic acid was used as a standard, precision of its titration was checked in four sets of different sample sizes. TABLE VIII RESULTS OF TITRATION OF SUCCINIC ACID VCYSUS 0.0929 N TETRA-n-BUTYL- AMMONIUM HYDROXIDE Volume to Volume to Acid end-point/ml Acid Amount fOund end-pointlml Acid hnount found taken/mg (1) found/mg Amount taken (2) found/mg Amount taken 32-38 2.89 31.70 0.979 31-80 2.84 31.15 0.980 31.60 2-82 30.94 0.979 32.38 2.87 31.48 0,972 32.47 2-93 32.14 0.990 32.67 2.93 32.14 0.984 Mean .. .. .. . . 0.981 Standard deviation . . . . 0.006 Coefficient of variation, per cent.. . 0.580 Error, per cent. . . .. . . -1.9 6.03 33-07 1.021 5-80 31.88 1.003 5.75 31.57 0.999 5.89 32.31 0.998 6.01 33-00 1.015 6.04 33.16 1.015 Mean .... .. . . 1.008 Standard deviation . . . . 0.010 Coefficient of variation, per cent.. . 0.990 Error, per cent. . . .. .. -1.0 A typical set of results obtained for succinic acid is shown in Table VIII. The values given in this table indicate that the error and the coefficient of variation of these determina- tions is in the range of 1 to 2 per cent., a value similar to that predicted from the precision of volume measurements. With a more precise burette and recorder system the precision796 SCOTT AND SVEHLA: AUTOMATIC RADIO FREQUENCY TITRATION OF [ABUt?ySt, VOl. 96 could be increased, though we found that the variation in temperature, which results in changes in the specific volume of the titrant, is a considerable source of error.We did not attempt, however, to overcome this problem, as such variations in room temperature would occur normally in laboratories. In laboratories with air conditioning a better degree of precision might be expected. Such accuracy and precision can be achieved when titrating acids in the range 10-4 to 2 x gram-equivalents. With a slightly reduced accuracy and precision, samples containing not less than 6 x gram-equivalents can still be analysed. In the higher concentration range more acetone might be needed to provide good titration curves that are free from instrument noise. Binary mixtures of acids can be resolved if the molar ratios are between 10: 1 and 1 : 10. Accuracy and precision of titrations of acids in mixtures can be judged from Table IX, in which results for the titration of mixtures of malic acid and 3-hydroxybenzoic acid are shown.The coefficient of variation is in the same range (1 to 2 per cent.) as for the titration of single acids. The errors are somewhat larger and increase as higher end-points are used for evaluation because these errors accumulate. In extreme instances, e.g., the determination of the lesser component in a 1: 10 mixture, the reproducibility of the results is lower; the coefficient of variation may then be as high as 3 to 4 per cent. M takenlmg 27.28 27.25 28.19 28-53 28.64 28-47 HB taken/mg 29.13 28.52 28.86 29.13 28.67 28.91 TITRATION OF MIXTURES OF MALIC AND 3-HYDROXYBENZOIC ACIDS Volume to end-pointlml M Amount found (1) found/mg Amount taken 2.09 26-68 0.978 2-11 27.00 0.991 2.17 27.71 0.983 2.18 27.93 0.979 2.20 28.10 0.981 2.2 1 28.38 0.997 Volume to end-pointlml M Amount found ( 3) found/mg Amount taken 0.987 26-92 2.10 2-12 27.06 0.993 2.19 28.05 0-996 2.22 28.42 0.99G 2-24 28-61 0.999 0.991 28.21 2.20 Mean * ... .. . . 0.985 Mean .. .. .. . . 0.994 Standard deviation . . . . 0-007 Standard deviation . . . . 0.006 Coefficient of variation, per cent.. . 0.7 Coefficient of variation, per cent.. . 0.6 Volume to Volume to end-pointlml HB Amount found end-pointlml HB Amount found (2) found/mg Amount taken (4) found/mg Amount taken 2-32 30.64 1.052 2.29 30-27 1.039 2-25 29.75 1.043 2-34 30.83 1-081 2.32 30.62 1.061 2-33 30.68 1.063 2.29 30.18 1.036 2-36 31-08 1-067 2.31 30.42 1.061 2.29 30.22 1.054 2-34 30.79 1.065 2.33 30.70 1.062 Mean .... .. .. 1.049 Mean .. .. .. . . 1.064 Standard deviation . . . . 0.012 Standard deviation . . . . 0.014 Coefficient of variation, per cent.. . 1-2 Coefficient of variation, per cent.. . 1.4 M = malic acid; HB = 3-hydroxybenzoic acid. A more comprehensive account of the whole of the work carried out is available in the form of a thesis.2s The authors thank Professor C. L. Wilson for his interest in this work. One of us (W.J.S.) thanks the Government of Northern Ireland for the provision of a Postgraduate Studentship. We are indebted to Mr. R. Donelly who took the photographs of the instrument. REFERENCES 1. 2. 3 . 4. 5. 6. 7. Fritz, J. S., and Marple, L. W., Analyt. Chem., 1962, 34, 921. Crabb, N. T., and Critchfield, F. E., Talanta, 1963, 10, 271. Kreshkov, A. P., Drozdov, V. A., and Kolchina, N. A., Zh. Analit. Khim., 1967, 22, 123. Wetzel, F. L., and Meloan, C. E., Analyt. Chem., 1964, 36, 2474. Ljn, S., and Blake, M. I., J . Pharm. Sci., 1966, 55, 781. Marple, L. W., and Fritz, J. S., Analyt. Chem., 1963, 35, 1223. -,- , Ibid., 1963, 35, 1305.November, 19711 ACIDS I N TERTIARY BUTYL ALCOHOL - ACETONE MEDIUM 197 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Riolo, C. B., and Soldi. T. F., Annuli Chirn., 1964, 54, 923. Lickhova, N. V., and Dulova, V. I., Im. Vjjssh. Ucheb. Zaved., Khirn. Khim. Tekhnol., 1964, Scott, W. J., Quigg, R. K., and Svehla, G., Analyst, 1970, 95, 113. Fritz, J. S., and Yamamura, S. S., Analyt. Chern., 1957, 29, 1079. Cundiff, R. M., and Markunas, P. C.. Ibid., 1956, 38, 792. Bruss, D. B., and Harlow, G. A., Ibid., 1958, 30, 1836. Marple, L. W., and Fritz, J. S., Ibid., 1962, 34, 796. Harlow, G. A., Ibid., 1962, 34, 1482 and 1487. Buell, B. E., Ibid., 1967, 39, 762. hlorales, R., Ibid., 1968, 40, 1148. Lund, H., and Bjerrum, J., Ber. dt. chern. Ges., 1931, 64, 210. \‘an Meurs, N., and Dahmen, E. A. M. F., Analytica Chim. Acta, 1959, 21, 10. Kortiim, G. F. A., Vogel, V., and Andrussow, K., “Dissociation Constants of Organic Acids in Aqueous Solutions,” IUPAC Commission on Electrochemical Data; Pure Appl. Clzem., 1960, 1, 187. Van Meurs, N., and Dahmen, E. A. M. F., J. Electroanalyt. Chem., 1959160, 1, 458. Masui, M., J. Pharm. Soc. Jafian, 1955, 75, 1519. Marple, L. W., and Scheppers, G. T., Analyt. Chzem., 1966, 38, 553. Millar, K. I., Thesis to be submitted for the Ph.D. degree, Queen’s University, Belfast, 1972. Van Meurs, N., and Dahmen, E. A. M. F., Analytica Chirn. Acta, 1959, 21, 443. Scott, W. J., Ph.D. Thesis, Queen’s University, Belfast, 1970. 7, 10. Received March 31st, 1970 Amended AfiriE 28th, 1971 Accepted June 7th, 1971

ISSN:0003-2654    

DOI:10.1039/AN9719600785

出版商:RSC    

年代:1971

数据来源: RSC