ISSN:0003-2654    

DOI:10.1039/AN97500FX029

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BX031

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500FP085

出版商:RSC    

年代:1975

数据来源: RSC

ISSN:0003-2654    

DOI:10.1039/AN97500BP089

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

AUGUST, 1975 The Analyst Vol. 100, No. 1 193 Nitrogen Heterocycle and Polynuclear Hydrocarbon Fluorescence and Adsorption Effects in the Presence of Silica Gel. Applications in High-pressure Liquid and Microcolumn Chromatography J. B. F. Lloyd West Midland Forensic Science Laboratory, Gooch Street North, Birmingham, B6 6QQ A flow-through cell packed with silica gel is used for the spectrofluorimetic examination of adsorbed states and as a detector in high-pressure liquid chromatography. The intensified fluorescence of adsorbed benzo homologues of quinoline and acridine is found to be caused by protonation of the elec- tronically excited states by the gel. Sub-nanogram amounts of such com- pounds can therefore be analysed by high-pressure liquid chromatography. In the presence of silica gel the fluorescence of polynuclear hydrocarbons is enhanced proportionally to their radiative lifetimes.The effect is dis- favoured by adsorbate formation, and can therefore be employed in both adsorption and reversed-phase liquid chromatography in order to improve selectivity and sensitivity towards readily quenched fluorescences. Microcolumns (90 x 1 mm) exhibit a ten-fold reduction in theoretical plate height compared with conventional high-pressure columns of the same adsorbent. Separations in such columns mounted on a microscope stage can be monitored by spectrofluorimetric microscopy of a silica gel particle located at the end of the column. The techniques described are applied to heterocycles in creosotes, poly- nuclear hydrocarbons in gasolines and sump oil, and wood splinters, soot particles and trace amounts of pitch. Transfer from solution to an adsorbed state may extensively modify a fluorescence emission.The effect is, for instance, sometimes noted when fluorescent compounds are analysed by thin-layer chromatography.1 Nicholls and Leermakers2 have recently reviewed the subject, mainly with reference to the nature of adsorbate - adsorbent interactions. Adsorbate fluorescence can readily be applied to enhance the sensitivity and selectivity of fluorescence detection, in high-pressure liquid chromatography, and enables microcolumn separations to be monitored by spectrofluorimetric microscopy. Because the existence of previously undetected interactions is indicated, some of the results are significant, forensically as well as analytically: the detection of certain types of contact trace depends on the formation of fluorescent adsorbate^.^ Experimental Materials Solvents used are commercial spectrophotometric grades, exhibiting no detectable ex- traneous fluorescence emission or ultraviolet absorption.Diethyl ether and ethyl acetate for chromatography are one third saturated with water.* Compounds, not below analytical- reagent grade initially, are purified by thin-layer or liquid column chromatography. The various stationary phases are: neutral aluminium oxide containing 4-5 per cent. m/m of water, 18-30-pm particle size (Woelm) ; Corasil C,,, 37-50 pm (Waters Associates Inc.) ; Porasil C , 37-75 pm (Waters Associates Inc.) ; silica gel CT, 11.5 pm (Reeve Angel Scientific Ltd.) ; and thin-layer chromatographic sheets of silica gel Sil G (Machery-Nagel and Co.).For the last named, scanning electron microscopy gives a mean particle diameter of 21 pm. Crown Copyright. 529530 Analyst, VoZ. I00 High-pressure Liquid Chromatography Columns are dry-packed in stainless-steel tubes (2.15 mm i.d., 04-1-15 m in length) con- nected through a union T-coupling and a length of steel tubing (0.76 mm i.d., 3 m in length), which serves as a flow pulsation damper, to either a Micropump, Series 2 (F. A. Hughes Ltd.), or a Metripump, Type HM (Metering Pumps Ltd.). Samples are injected through the T- coupling. The column outlet is connected by PTFE tubing (0.25 mm i.d., made by stretching 0.5 mm diameter tubing) to a flow-through cell (Spectrosil tube of 1 mm i.d.) mounted in the microcell attachment of a Baird Atomic SF lOOE spectrofluorimeter.Similar apparatus has been described previo~sly.~ Solvents are de-aerated by a stream of “white spot” nitrogen (British Oxygen Ltd.). In order to eliminate oxygen from columns it is necessary to pump with de-aerated solvent for at least 2-5 h. This operation is complete when the fluorimetric response to pyrene peaks no longer increases. Inlet joints of packed detectors are subject to pressures of the order of 100 Ib i r 2 ; the joints are made as follows [Fig. l ( a ) ] . The wall of a flow-through cell (length 60 mm, 0.d. 3.5 mm and i.d. 1 mm) is slightly constricted by heating it at a point 7 mm from one end; the constriction in internal diameter is approximately 0.1 mm. Through the other end of the cell is threaded a stretched piece of PTFE tubing, initially of 1 mm o.d., into the unstretched end of which has been pushed a segment (5 mm) of a size 15 hypodermic needle.This rein- forced part of the tubing is pulled down hard into the constriction. No joint made in this way has failed with use over a period of 2 years with packings down to 10 pm in particle size and with solvent flow-rates of up to 2 ml min-l. The outlet joint is made in a similar fashion, except that the cell walls are unconstricted. The cells are packed with dry adsorbent, which is retained between plugs of cotton- or glass-fibre held in place between the inlet and outlet fittings. LLOYD : APPLICATIONS OF THE FLUORESCENCE-ENHANCING EFFECT ( a ) Quartz flow- Fibre through cell Plug \ \ I----- -__-_ I Outlet 2- - ----- I PTF E (inlet) \ Silica / Hypodermic ( b ) Fibre Silica Aluminium oxide / Plug Capiiiary pipette I Microscope point of focus dimensions are given in t h e text.Fig. 1. Packed detector (a) and microcolumn (b) (cross-sections). The Sample solutions are forced through the detector from a hypodermic syringe until the emitted fluorescence is constant. For chromatography, the detector is purged with solvent until a stable baseline is obtained. Detector lifetime is determined by the accumulation of strongly adsorbed fluorescent material, i.e., by the nature of the samples and by photo- decomposition products. Typically the packing is replaced after 1 week’s use.August, 1975 OF SILICA GEL IN HIGH-PRESSURE LIQUID CHROMATOGRAPHY 531 Microcolumn Liquid Chromatography Columns [Fig.l(b)] are made from Drummond Microcap pipettes (100 p1, length 116 mm and i.d. 1.05 mm). At a distance of 10 mm from the outlet the pipette is heated and pulled down to form a constriction that is 10mm in length and of 0.2mm i.d. Into the outlet end of the constriction is introduced a plug of cotton-fibre, and, from the column inlet, sufficient fragments of silica gel to occupy the lower half of the constriction. The remainder of the constriction and the column, to within 2 mm of the top, are filled with dry adsorbent. Typically, when the adsorbent is aluminium oxide, 85 mg are used. The column is mounted on a microscopy stage illuminated through a dark-field condenser by a Leitz 200 W mercury lamp, from which the 366-nm emission is isolated by UG 1 filters.Optical contact between the restricted part of the column and the surface of the condenser is established by use of immersion oil. The microscope is focused on the junction of the silica gel and the column packing at a magnification such that the fluorescence emitted from a single particle of gel can be collected with a Microscope Spectrum Analyser (Farrand Optical Co. Inc.) mounted on the eyepiece and connected to a strip-chart recorder. Monitoring wavelengths are indicated in the text and in the figure captions. Solvent is forced through the column either by an Agla micrometer syringe driven with a synchronous motor, or by compressed nitrogen. In the latter instance, which is the more suitable for prolonged running times, solvent is fed from a spiral of PTFE tubing connected to the column inlet. Flow-rates in the region of 20 pl min-l require a pressure of 30 lb in-2.Liquid samples, e.g., 10 nl, are transferred by capillary micropipette to the inlet plug of the disconnected column. Solid samples are pushed into the top of the capillary of a micro- pipette where they are extracted with 1-2 pl of solvent, usually ethyl acetate. The extract is discharged into the inlet plug where the solvent is evaporated in a stream of air from the pipette. All of these operations are conducted under a dissecting microscope. Fluorescence Spectra The spectra are recorded with a Baird Atomic SF 1OOE spectrofluorimeter at a half-band width usually of 5 nm; they are uncorrected and subject to variation when other instruments are used.Results and Discussion Heterocyclic Compounds The weak fluorescence of acridine in neutral solution is strongly intensified by the addition of acids.69' Thus, a dichloromethane solution [Fig. 2(a)] yields weak excitation and emission spectra (characteristic of the neutral molecule) that are transformed into the intensified ( x 500) spectra of the protonated form on the addition of trifluoroacetic acid. When a neutral dichloromethane solution of acridine is injected into a flow-through cell packed with silica gel (thin-layer chromatographic material so as to permit the subsequent correlation with R, data) the fluorescence is again intensified, but the excitation spectrum remains essentially that of the neutral molecule, whereas the emission is evidently from the protonated form [Fig.2 (a)]. Hence, the intensification is due to protonation by the acidic gel surface of electronically excited acridine molecules. The effect is not observed on basic adsorbents such as aluminium oxide. On the basis of fluorescence emission spectra, Aleksandrova et aL8 reported that on hydrated silica vaporised acridine is adsorbed in the protonated state. However, in the absence of excitation spectra their conclusion is also explicable in the foregoing terms, Excited states of benzo[h]quinoline and benz[c]acridine are protonated to a reduced extent by the gel so that emissions of both protonated and unprotonated forms are present in the spectra. Quinoline is protonated in the ground state.Carbazoles (carbazole, 13H-dibenzo [a,i]carbazole and 7H-dibenzo [b,g]carbazole) , which are non-basic, exhibit unmodified spectra in the adsorbed state. From the R, value (0.037) of acridine in dichloromethane on the silica gel used for the cell packing, and from the ratios of gel to solvent in the cell and in the thin-layer chromato- gram, it follows that adsorption increases the concentration of acridine in the irradiated part of the cell by a factor of 9.7 relative to an unpacked cell when both are equilibrated with the same solution. After making allowance for this effect, the observed increase relative to the BenzoIflquinoline behaves in the same way as acridine [Fig. 2 ( b ) ] .532 Analyst, VoZ. 100 neutral and acidified solutions in unpacked detectors is x 590 and x 1.2, respectively.Hence, the production of excited acridinium ions by the gel surface or in acidic solution occurs to comparable extents. LLOYD : APPLICATIONS OF THE FLUORESCENCE-ENHANCING EFFECT Wavelengthhm Fig. 2. Fluorescence excitation and emission spectra of (a) acridine and (b) benzo[flquinoline, at 0.1 pg ml-l concentration in dichloromethane. Broken lines are spectra from neutral solutions, dotted lines are from solutions acidified with 10 pl ml-l of trifluoroacetic acid and full lines are from neutral solutions injected into silica gel contained in a flow-through cell. Solvent effects, which do not modify the spectra of the adsorbed state qualitatively, show that adsorption occurs mostly prior to excitation. Thus, the fluorescence intensities from a packed cell, equilibrated with constant amounts of acridine in various solvents, increase according to R, valuesg from thin-layer chromatograms [R, = log (Rp-l - l ) ] and hence according to logarithms of the corresponding distribution coefficient of the ground state. Some results are shown in Fig.3. (Two, off-scale, results for n-hexane and carbon tetrachloride are excluded because, as follows from the corresponding R, value of 0, equilibrium could not be established.) A linear least-squares analysis of the results obtained with n-hexane and ethyl acetate mixtures yields a gradient not significantly different from unity (5 per cent. level). However, increasingly less basic solvents lie increasingly above this line (no correction for the absorption of excitation by nitromethane has been made, hence the result underestimates the high fluorescence yield in the presence of this solvent) and proton-donating solvents lie below, presumably because of the varying extent to which the solvents compete with the gel surface for electronically excited acridine molecules.Clearly, proton-donating solvents should not be used for chromatography if sensitivity is to be at a maximum. If broadened peaks are to be avoided: distribution coefficients of eluted compounds on cell packings should not exceed the corresponding values on column packings ; other chromatographic parameters that characterise the cell should not be inferior to those of the column; and the length of the cell viewed by the spectrophotometer should be small relative to the column length occupied by separated compounds. Ideally, therefore, the detector should simply be an extension of the column. These conditions are not difficult to meet in practice, even when cell and column packings differ.Thus, coupled to a 0.5-m column of aluminium oxide eluted with a dichloromethane and acetonitrile mixture (95 + 5 ) at the rate of 1 ml min-l, a silica gel packed detector gives peak widths corresponding to a theoretical plate height of 1-28 mm (standard error, 0-067), which is not significantly different (5 per cent. level) from the value of 1.16 mm (0.037) obtained for anthracene and benz[a]anthracene on the same type of column in the absenceAugust, 1975 OF SILICA GEL IN HIGH-PRESSURE LIQUID CHROMATOGRAPHY 533 of cell packing.The comparison cannot be made with acridine detected in the unadsorbed state, because in the absence of cell packing no response is obtained. Peak heights increase slightly less than linearly with chromatographed amounts in the range 0.6-1280 ng, but a graph of the data in double logarithmic co-ordinates is linear, and yields a linear least-squares analysis of y = 0.0417 + 0-937x, syjx = 0-0436, sg = 0-0258 and Sb = 0.0132, where y and x are logarithms of fluorescence intensities and amounts of acridine, s values are standard deviations, and a and b are the intercept and slope, respectively. An application of the technique to the differentiation of microlitre amounts of two creosotes is shown in Fig. 4. With excitation and emission wavelengths of 358 and 475 nm (optimum for adsorbed acridine), both exhibit an initial peak mainly of unretained polynuclear hydro- carbons followed by heterocyclic components that differ in relative intensities, and in the presence of an additional, partly resolved, component in sample A.Also in Fig. 4 are shown chromatograms of (C) one of the samples monitored in the absence of cell packing, when only the polynuclear hydrocarbon peak can be seen; and (D) of 0.6 ng of acridine, which represents the practicable limit of detection under these particular conditions. 3.0 1 +J v) .- 2.5 +.' .- a) F 2.0 2 0 1.5 - + o o 3.5 1 - - - - 0 0.5 I.Ol: 0- 0 0.5 1.0 1.5 2 RM Fig. 3. Variation with solvent of fluorescence emitted at 475nm (excitation, 358 nm) from a silica gel packed detector equilibrated with solutions of 0.1 pgrnl-l of acridine in various solvents.RM values are from thin-layer chro- matograms in each of the solvents on the same gel. Closed circles and the corresponding least-squares line represent mixtures of ethyl acetate and n-hexane (containing from 100 to 2 per cent. V / V of ethyl acetate). Open circles in order of increasing fluorescence represent 20, 10 and 5 per cent. V/V of methanol in ethyl acetate, acetone, diethyl ether, acetonitrile, nitro- methane, dichloromethane and benzene. Polynuclear Hydrocarbons D --JL 5 10 15 0 5 10 Ti me/m in Fig. 4. Liquid chromatograms on 0-5 m x 2-15 mm columns of 18-30-pm aluminium oxide eluted with dichloromethane - acetonitrile (95 + 5) mixture at the rate of 1 ml min-l.Fluorescence detection is a t 358 nm (excitation) and 475 nm (emission). Samples A and B are of different creosotes (1 pl injected) monitored with a silica gel packed detector; C is sample A monitored with an unpacked detector; and D is 0.6 ng of acridine monitored with a packed detector. The enhanced fluorescence yields of heterocycles adsorbed on silica gel in the presence of generally strongly quenching solvents, e.g., carbon tetrachloride and nitromethane, suggests that the technique might be used to potentiate fluorescence emission from readily quenched polynuclear hydrocarbons.534 Analyst, VoZ. 100 In the presence of silica gel the excitation and emission spectra of these compounds are not usually varied relative to their dissolved states.Pyrene is exceptional. When a cyclo- hexane solution is injected into a cell packed with silica gel, the relative intensities of the vibronic transitions are considerably modified (Fig. 5) to yield an emission spectrum that is closely comparable with spectra in polar solvents such as methanol. Excitation spectra are not affected. Leermakers and co-workers2$l0 reported that addition of silica gel to cyclo- hexane solutions leaves the spectrum of pyrene monomer unaltered. Under their conditions, however, vibrational fine structure is unresolved. (Effects involving excimers reported by Leermakers and co-workers are not observed at the low concentrations used here.) LLOYD : APPLICATIONS OF THE FLUORESCENCE-ENHANCING EFFECT t I 400 450 B J 400 450 Wave lengthh m 400 450 Fig.5. Fluorescence emission spectra of pyrene (1 pg ml-l) in de-aerated cyclohexane (A), methanol (B) and cyclohexane (C) injected into a silica gel packed detector. Relative changes in fluorescence yields caused by silica gel can be determined by liquid chromatography. Anthracene, because of its relatively short fluorescence lifetime, and consequently reduced sensitivity to quenching under usual circumstances, is used as an internal standard. Anthracene and each hydrocarbon in solution are separated [using 1120 x 2-15 mm columns of 24-pm aluminium oxide, in a mixture of n-hexane and diethyl ether (100 + 6), at the rate of 0.6 ml min-l] with fluorescence detection at wavelengths set in regions of spectral overlap of anthracene and the compound in question.In aerated solvents, i.e., under conditions of oxygen quenching, silica gel packed detectors give varyingly increased fluorescence yields relative to unpacked detectors. In the series fluorene, biphenyl, fluoranthene, benz [alanthracene, chrysene, naphthalene, triphenylene, and pyrene, chromatographic peak heights relative to anthracene vary according to radiative lifetimes. Thus, peak height ratios increase from 1.19 to 4.16 as radiative lifetimes (from Birksll) increase from 15 to 690 ns to give a correlation coefficient of 0.951 (significance level, 0.1 per cent.). Ten peaks from five chromatograms of an anthracene and benz [alanthracene mixture monitored with a packed detector give a value of 1.12 mm (standard error, 0.042), which is not significantly different (0.1 per cent.level) from the value previously quoted for an unpacked detector. When the solvent is de-aerated, the response of an unpacked detector is similar to that of a packed detector in the presence of aerated solvent. Relative to these conditions, the response is further increased when a packed detector functions in de-aerated solvent, Examples are given below. The extent to which the above hydrocarbons are adsorbed on silica gel varies only slightly between them [thin-layer chromatographic mobilities relative to anthracene (1.0) vary from 0-79 to 1.131 and shows no correlation with fluorescence enhancement. Fluorescence yields are reduced if the thin-layer chromatographic gel with which the detector is packed is replaced by Porasil C or by the micro-particulate silica gel CT, 11.5 pm.When a packed detector is de-activated by injection of water into the eluate stream, or similarly with bis(trimethylsily1)- Theoretical plate heights are unaffected by the detector packing.August, 1975 OF SILICA GEL IN HIGH-PRESSURE LIQUID CHROMATOGRAPHY 535 acetamide, the fluorescence response is increased (as much as x 1.8). Hence, adsorption reduces fluorescence yields of polynuclear hydrocarbons, presumably by increasing quenching encounters between adsorbates, or by promoting internal conversion or inter-system crossing processes. The increased fluorescence of non-adsorbed molecules is attributed to a reduction in diffusion-controlled rates of quenching encounters in the fluid occupying the pores and interstices of the ge1.12 From this unexpected result it follows that packed detectors can enhance the specificity and sensitivity of reversed-phase (aqueous) as well as adsorption chromatographic systems.Examples of the modified sensitivities towards polynuclear hydrocarbons induced by silica gel are shown in Figs. 6 and 7. Fig. 6 shows a 98-octane gasoline chromatographed in a mix- ture of n-hexane and diethyl ether on aluminium oxide. Monitored at 325 nm (excitation) and 390 nm (emission), intensities of peaks in the naphthalene and pyrene positions (the first and third peaks) are increased relative to that of anthracene (second peak) by factors of 2.3 and 8.0 when conditions are varied from aerated solvent and unpacked detector to de- aerated solvent and packed detector.De-aerated unpacked and aerated packed detectors give intermediate results. At 340 and 410nm, at which wavelengths naphthalene is not detected, the same variation increases the pyrene peak intensity by a factor of 11.6. The variation in the pyrene result with monitoring wavelengths is attributed to the presence of other, unresolved, compounds at the anthracene and pyrene positions, and to the changes caused in the pyrene spectrum by silica gel. (Peaks due to pure pyrene relative to pure anthracene exhibit a similar dependence on wavelength, but the over-all increases in intensities are in the region of 20-fold.) The addition of up to 0.6 pg of pyrene to the gasoline injected yields peak intensities that increase linearly by factors of up to 7.6. Ti me/m in Fig.6. Liquid chromatograms of a %octane gaso- line (5-pl samples) on a 1.12 m x 2-15 mm column of 18-30-pm aluminium oxide eluted with n-hexane - diethyl ether (100 + 6 ) mixture at the rate of 0-6 ml min-'. Chromatograms: A was obtained with aerated solvent and unpacked detector, and with excitation and emission set a t 325 and 390 nm; B was obtained a t the same wavelengths with de-aerated solvent and packed detector (silica gel) ; and C and D are a similar pair but monitored a t 340 and 410 nm. A reversed-phase separation of a used oil on Corasil C,,, essentially according to the condi- tions used by Vaughan, Wheals and Whitehouse,13 is shown in Fig. 7. From the solvent used (aqueous methanol) silica gel does not adsorb ground states and presumably excited states of polynuclear hydrocarbons.However, in accordance with the foregoing, peaks due to pyrene (at 7.5 min) and chrysene (at 13 min) are increasingly intensified as the conditions are varied from unpacked aerated to packed de-aerated detectors. Relative to the anthracene peak (6.5 min), the over-all increases are by factors of 12.1 and 10.4. Packed and unpacked detectors do not yield peak height ratios sigmficantly different in variance. Although increased noise levels are associated with packed detectors, due to536 A~taZyst, VoZ. 100 increased levels of scattered radiation (the increase is approximately three-f old, depending on excitation and emission wavelengths), this effect is more than offset by the increased fluorescence yields obtained. Hence, with readily quenched fluorescences, the described conditions enable detection limits to be reduced by as much as one tenth.LLOYD : APPLICATIONS OF THE FLUORESCENCE-ENHANCING EFFECT J , I I 9 5 10 15 20 25 Timehi n Fig. 7. Liquid chromatograms of a sump oil (l-mg samples) on a 1.12 m x 2.15 mm column of Corasil C,, eluted with methanol - water (80 + 20) mixture a t the rate of 1 ml min-1. Fluorescence detection is a t 350 nm (excitation) and 380 nm (emission) with (A) aerated solvent and unpacked detector, (B) de- aerated solvent and unpacked detector and (C) de-aerated solvent and silica gel packed detector. Microcolumn Liquid Chromatography The application of spectrofluorimetric microscopy to thin-layer chromatography has been mentioned by P ~ k e r . 1 ~ The technique is readily adapted to microcolumn chromatography, which can be conducted on a microscope stage as described under Experimental. Although fluorescence in the fluid effluent from a microcolumn can be monitored microscopically, signals of much improved intensity and stability result when emission from silica gel fragments is monitored.That the technique is practicable depends on the considerable reduction in theoretical plate height exhibited by the microcolumns, which results in separation efficiencies comparable with conventional 1-m long high-pressure columns of the same adsorbent. Thus, anthracene and benz[a]anthracene separated in a microcolumn (88 x 1 mm) under conditions otherwise as previously described yield a value of 0-0941 mm (standard error, 0.0042) from 18 peaks in nine chromatograms, in contrast to the former values of 1.16 and 1.12 mm for aluminium oxide columns.Apart from the effect of the relatively low flow velocity in the microcolumn, the improved efficiency is probably due also to the reduction of diffusional broadening by reduced column dimensions, and to the complete elimination of the dead-volume effects between column and detector that tend to limit the resolution that can be obtained with conventional columns.AzLgust, 1975 OF SILICA GEL IN HIGH-PRESSURE LIQUID CHROMATOGRAPHY 537 Because the column, and therefore, the point of focus, are disturbed during sample injection, quantitative measurements are best made against internal standards. In the presence of 10 ng of anthracene the relative response to between 10 and 120 ng of benz[a]anthracene monitored at 410 nm yields a linear least-squares analysis of y = 0.0210~ - 0.0156, syIx = 0.0583, sa = 0.0367 and s b = 0.00056, where y is the fluorescence intensity due to x ng of the benzanthracene relative to 10 ng of anthracene, and s, a and b have their previously indicated significance.Similarly, 50 pg to 2 ng of acridine relative to 500 ng of perylene chromatographed in a mixture of n-hexane and ethyl acetate (90 + 10) and monitored at 475 nm yields y = 1.140~ - 0.0403, sylx = 0.0622, sa, = 0.118 and sb = 0.0444. The separation of N-heterocyclic compounds from 10-nl volumes of samples of creosotes is shown in Fig. 8. These chromatograms should be compared with those of the same samples in Fig.4. Although the sample size is reduced by one hundredth, the microcolumn reveals the presence of compounds that were previously undetected. The patterns are reproducible to the extent that relative to the most prominent heterocyclic peak the coefficient of variation of the remainder is 7.51 per cent. An initial, variable injection spike can be seen in some of the chromatograms. Also, the broad, rapidly eluted polynuclear hydrocarbon peak varies in resolution and intensity, probably because with these compounds the column is in an overload condition. When the microscope is focused on an aluminium oxide particle, the polynuclear hydrocarbon peak is the only peak detected. k A c, 5 10 1 i k 5 10 0 5 10 Ti me/min Fig. 8. Microcolumn (88 x 1 mm) chromatograms on 18-30-pm aluminium oxide in n-hexane - ethyl acetate (95 + 5 ) mixture, 15 pl min-l, monitored by spectrofluorimetric microscopy of a silica gel fragment at 366 nm (excitation) and 475 nm (emission).Samples A and B (10 nl) are the same creosotes as in Fig. 4. Sample C is a splinter (76 pg) from fencing coated with A. A chromatogram derived from a small splinter (76 pg) of wooden fencing that had been A variety of other splinters of the same Evidently, it should be possible to correlate items of this sort In Fig. 9 are shown chromatograms of a soot fragment (about 1 pg) and a trace amount The patterns obtained reflect the varying polynuclear hydrocarbon Other materials rich in this class of compounds15 can be coated with one of the creosotes is included in Fig. 8.source give the same pattern. with likely points of origin. of pitch (about 3 pg). compositions of the samples. similarly characterised. Future Developments The techniques already described offer a means, applicable on a microscopic scale, of analysing538 LLOYD : APPLICATIONS OF THE FLUORESCENCE-ENHANCING EFFECT Analyst, VoZ. 100 I 5 10 I I 5 10 Time/min Fig. 9. Microcolumn (88 x 1 mm) chromatograms on 18-30-pm aluminium oxide in n-hexane - diethyl ether (100 + 6) mixture, 20 pl min-’, monitored by spectrofluorimetric microscopy of a silica gel fragment excited a t 366 nm, emission followed a t 430 nm. Sample A is of pitch (about 3 pg) and B is a fragment of soot (about 1 pg). minute amounts of potentially fluorescent compounds, and of studying the nature of elec- tronically excited adsorbed states.However, many other types of detector packing are feasible. For instance, chemically bonded stationary phases might be used to exploit specific fluorescence effects. The phosphorescence observed at room temperatures in the adsorbed state by Schulman and Walling16,17 and luminescence quenching, widely used in thin-layer chromatography and, very recently, in gas chromatography by Schulz and Vilceanu,l8 are of obvious significance. Again, micro-particulate adsorbents should enable microcolumns of even greater efficiency to be produced. Indeed, the full potentialities in other applications are unlikely to be fully realised until such packings are employed in conjunction with spectro- microscopy in order to eliminate entirely the problem of connecting high-resolution columns to detectors. References 1. 2. 3. 4. 5 . 6. 7. 8. Sawicki, E., Talanta, 1969, 16, 1231. Nicholls, C. H., and Leermakers, P. A., Adv. Photochem., 1971, 8, 315. Lloyd, J. B. F., Analyst, 1975, 100, 82. Snyder, L. R., and Kirkland, J. J., “Introduction to Modern Liquid Chromatography,” Wiley- Lloyd, J. B. F., J . Forens. Sci. SOC., 1971, 11, 235. Mataga, N., Kaifu, Y., and Koizumi, M., Bull. Chem. Soc. Japan, 1956, 29, 373. Weller, A., Progr. Reaction Kinetics, 1961, 1. 187. Aleksandrova, G. P., Buchneva, A. I., Ignat’eva, L. A., and Levshin, L. V., Z h . Prikl. Spectrosk., 1970, 13, 255; Chem. Abstr., 1971, 74, 59055. Interscience, New York and London, 1974, p. 270.August, 1975 OF SILICA GEL IN HIGH-PRESSURE LIQUID CHROMATOGRAPHY 539 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Bate-Smith, E. C., and Westall, R. G., Biochim. Biophys. Acta, 1950, 4, 427. Weis, L. D., Evans, T. R., and Leermakers, P. A., J . Am. Chem. SOC., 1968, 90, 6109. Birks, J. B., “Photophysics of Aromatic Molecules,” Wiley-Interscience, New York and London, Giddings, J. C., “Dynamics of Chromatography,” Edward Arnold, London; and Marcel Dekker, Vaughan, C. G., Wheals, B. B., and Whitehouse, M. J., J . Chromat., 1973, 78, 203. Parker, C. A., Awalyst, 1969, 94, 161. Lloyd, J. B. F., J . Forens. Sci. Soc., 1971, 11, 153. Schulman, E. M., and Walling, C., Science, N.Y., 1972, 178, 53. Schulman, E. M., and Walling, C., J . Phys. Chem., 1973, 77, 932. Schulz, P., and Vilceanu, R., J . Chromat., 1974, 100, 27. 1970, Chapter 4 . New York, 1965. Received January 29th, 1975 Accepted February 19th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000529

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

540 Artalyst, August, 1975, Vol. 100, p p . 540-543 A Rapid Gas-chromatographic Method for the Determination of Acetaldehyde in the Vapour Phase of Cigarette Smoke D. J. Evans and R. J. Mayfield C.S.I.R.O., Division of TextiEe Industry, P.O. Box 21, Belmoizt, Victoria 3216, AustraEia A simple and rapid procedure is described for the gas-chromatographic determination of acetaldehyde in the vapour phase of cigarette smoke. The acetaldehyde is efficiently extracted into cold water in Drechsel bottles and is determined by gas chromatography of an aliquot of the solution. Results are comparable with those obtained by more complicated techniques in- volving derivative formation or the direct injection of smoke samples into high-resolution columns. A relative standard deviation of 3.4 per cent.was obtained in the absence of an internal standard. The method is parti- cularly convenient for evaluating the effectiveness of cigarette filters in retaining acetaldehyde. The ciliatoxicity of the gas phase of cigarette smoke has been the subject of several reports.1-5 Much of this ciliatoxic effect has been attributed to the presence in this phase of volatile aldehyde^^.^^^ such as formaldehyde, acetaldehyde and acrolein. To facilitate the develop- ment of cigarette filters that selectively remove ciliatoxic components a simple and rapid procedure was required for the determination of acetaldehyde in the mainstream smoke of cigarettes. Gas-chromatographic procedures have been reported previously for the quantitative deter- mination of acetaldehyde and other volatile carbonyl compounds present in complex mixtures such as tobacco ~ r n o k e , ~ - ~ car exhausts,1° food flavours and aromas,ll and atmospheric samples.12 These methods have involved either prior derivative formation10-12 or direct gas chromat~graphy~-~ of gas samples.Derivative formation is time consuming in that it often necessitates extractions or special clean-up procedures prior to the gas-chromatographic analysis. Direct injection of samples of tobacco smoke into a gas chromatograph involves the use of specialised sampling equipment, high-resolution columns and temperature pro- gramming in order to separate the large number of component^.^-^ A gas-sampling valve alone is insufficient as a smoke sample representative of the whole cigarette must first be collected.Neither of the foregoing techniques is convenient for routine applications because of the slow elution of the less volatile components of cigarette smoke. In this paper, a simple and rapid procedure for the accurate determination of acetaldehyde in the mainstream smoke of cigarettes is described. Acetaldehyde in the vapour phase of the smoke is trapped in gas scrubbers containing cold water and is determined by gas- chromatographic analysis of an aliquot of the resulting solution. Special sampling equipment is unnecessary and interference from other smoke components is minimal as only water- soluble components are effectively collected. The precision of this method compares favour- ably with that of other methods,6-8 but its main advantage is the ease with which it can be applied to multiple analyses for acetaldehyde.Method Materials Cigarettes. AcetaEdehyde. Re-distil analytical-reagent grade (99 per cent .) acetaldehyde immediately before use. Standard acetaldehyde solution. Weigh accurately a stoppered 50-ml calibrated flask con- taining about 40 ml of distilled water. Introduce by means of a cooled syringe 230-270 mg of acetaldehyde, stopper the flask tightly and re-weigh. Swirl the contents and quickly dilute to 50 ml with distilled water. Dilute a 10-ml aliquot of the primary standard to 100 ml in a calibrated flask so as to give a solution containing 460-540 pg ml-l of acetaldehyde, and stopper the flask tightly. Prepare working standards from this solution by removing 4-, 5-, 6-, 7- and 8-ml portions and diluting to 100 ml in calibrated flasks with distilled water.Keep the flasks tightly stoppered at all times and store in a refrigerator when not in use. Condition cigarettes at 60 per cent. relative humidity for at least 48 h.EVANS AND MAYFIELD 541 Apparatus A Bendix, 2500 Series, gas chromatograph equipped with flame-ionisation detectors and a 1.9 m x 2 mm i.d. glass column packed with Chromosorb 101 (Johns-Manville) was employed isothermally at 110 "C. The flow-rate of the nitrogen carrier gas was 30 ml min-l and the detector and injector temperatures were maintained at 125 "C. Samples (4.0 p1) were intro- duced into the column by means of a 5-pl syringe (SGE, Type A). Peak areas were determined by a DISC integrator. Cigarettes were smoked with a CSM 100 4-channel smoking machine programmed to take 35-ml puffs of 2-s duration at intervals of 1 min.The smoke was drawn through a Cambridge filter assembly containing a glass-fibre pad to separate the particulate phase from the vapour phase. The trapping system consisted of two 50-ml test-tubes with B29 ground-glass joints fitted with adjustable Drechsel bottle heads. The tubes each contained distilled water (35 ml) cooled to 2 "C with an ice - water mixture and were placed in series between the smoking machine and the Cambridge filter holder. The puff volume of the smoking machine was adjusted with the traps and Cambridge filters in position. Procedure On the machine, smoke four cigarettes in succession to a constant butt length and pass the vapour phase of the smoke through the cooled traps; take a clearing puff at the end of each cigarette.Immediately on completion of smoking, transfer the cold smoke solutions to a 100-ml calibrated flask, rinsing the traps once with distilled water. Adjust the volume to 100 ml with distilled water, stopper the flask tightly and set it aside. To analyse the smoke solution, inject a 4-p1 sample into the gas chromatograph and record the peak area for acetaldehyde. Wait for approximately 4 min for acetone to elute before injecting another sample. Alternatively, inject a second sample immediately following elution of acetaldehyde from the first injection. After elution of acetaldehyde from the second injection allow 5 min to elapse for the elution of acetone from both injections before proceeding with another sample. Analyse each sample in triplicate, take the mean peak area and calculate the acetaldehyde concentration in micrograms per 100 ml direct from the calibration graph.Standardise the procedure by injecting 4.O-pl samples of the acetaldehyde standards into the gas chromatograph. Repeat this procedure in triplicate, calculate the mean peak area for each standard and prepare a calibration graph by plotting peak area against concentration (micrograms per 100 ml). Results The efficiency of collection of acetaldehyde in the cold-water traps was determined by passing the vapour phase from four cigarettes through a series of four traps and measuring the acetaldehyde content of each trap. Table I shows that all of the acetaldehyde was collected in the first two traps and that 97 per cent.of the total was retained by the first trap. Thus, provided that the traps are cooled to 2 "C during the smoking procedure, a system of two traps collects virtually all of the acetaldehyde from four cigarettes smoked in succession. TABLE I EFFICIENCY OF COLLECTION OF ACETALDEHYDE Collection system consisted of 4 traps, each containing 35 ml of water and cooled to 2 "C. Acetaldehyde content/ Total acetaldehyde, Trap pg per 100 ml per cent. 1 2870 97 2 92 3 3 Not detectable 0 4 Not detectable 0 To assess the stability of the acetaldehyde collected in the cold-water traps, the contents were stored at 2 "C for 7 h and samples removed at regular intervals for gas-chromatographic analysis. No appreciable loss of acetaldehyde occurred under these conditions.Hence, analysis can be delayed for several hours following the smoking procedure, provided that the vapour-phase solution is stored at 2 "C in an air-tight vessel. The stability of acetaldehyde542 EVANS AND MAYFIELD : RAPID GAS-CHROMATOGRAPHIC DETERMINATION Analyst, YoZ. 100 over 7 h in cold aqueous smoke solution does not appear to have been appreciated previously and, consequently, it provides the basis for this method. Table I1 shows the acetaldehyde content, as determined by the present method, of smoke from several commercial cigarettes’ and also gives results obtained by others64 who employed similar smoking procedures in conjunction with gas-chromatographic procedures. TABLE I1 ACETALDEHYDE CONTENT OF SMOKE FROM SOME COMMERCIAL CIGARETTES Acetaldehyde content* by- r > present method * other methodstlpg per puff !-%Per /%Per r h 1 Filter type: cigarette puff A6 B7 CB } 82 24 10”; } 96 Cellulose acetate (20 mm) . ... .. 801 Cellulose acetate - charcoal (17 mm) . . 656 77 101 Wool (20 mm) . . * . .. . . 788 90 Cellulose acetate (17 mm) . . . . . . 823 Treated wool (experimental, 20 mm) . . 593 71 Unfiltered . . .. .. .. . . 625 85 92 81 * Owing to differences between cigarettes and filters strict comparisons are not significant. t Methods of analysis by use of gas chromatography. : Butt length of tobacco column in this method: unfiltered 12 mm; filtered 6 mm. The cold-water traps employed here preferentially retain the water-soluble vapour-phase components. This substantially reduces the complexity of the vapour phase in subsequent analysis and facilitates resolution of acetaldehyde without cryothermal conditions, high- resolution columns or temperature programming.Fig. 1 illustrates the simple gas chromato- gram obtained by chromatography of the aqueous smoke solution. Acetaldehyde is well separated from other, minor peaks, allowing accurate integration of the peak area. A second injection can be made immediately following the elution of the acetaldehyde peak from the first injection, as the retention time of acetone is more than twice that of acetaldehyde and no other peaks interfere in this region. By this means the analysis time required for several samples can be considerably shortened. 1 I I I I I 0 1 2 3 4 5 6 7 8 Timehin Fig.1. Gas chromatogram of the water-soluble components of tobacco smoke. 1, Acetaldehyde; 2, acetone. An internal reference standard was not employed and therefore aliquots of standards and samples for analysis by means of chromatography were accurately measured with a high- quality syringe. Analysis of each standard and sample was carried out three times and the mean acetaldehyde peak area was determined. The reproducibility of the method was established from the results of 12 determinations utilising 48 cigarettes, i.e., the acetaldehydeAugust, 1975 OF ACETALDEHYDE IN THE VAPOUR PHASE OF CIGARETTE SMOKE 543 content was found to be 776pg per cigarette & 3.7 per cent. relative standard deviation (90.2 pg per puff & 3-4 per cent.relative standard deviation). The calibration graph for the acetaldehyde standards was linear and the standards remained unchanged over 48 h if kept in a tightly stoppered flask and stored a t 2 “C when not in use. Conclusion The present method is a simple and rapid one for the accurate determination of the acetalde- hyde content of the vapour phase of cigarette smoke. The results are particularly useful in evaluating the effectiveness of cigarette filters in retaining acetaldehyde. The method could possibly be extended to the determination of other biologically active water-soluble components present in cigarette smoke. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. References Battista, S. P., and Kensler, C. J., Archs Envir. I-Zth, 1970, 20, 318. Kensler, C. J., and Battista, S. P., New Engl. J . Med., 1963, 269, 1161. Wynder, E. L., and Hoffmann, D., “Tobacco and Tobacco Smoke,” Academic Press, New York Dalhamn, T., and Rylander, R., Am. Rev. Resp. Dis., 1968, 98, 509. Dalhamn, T., and Rosengren, A., Archs. Otolar,, 1971, 93, 496. Horton, A. D., and Guerin, M. R., Tob. Sci., 1974, 18, 18. Laurene, A. H., Lyerly, L. A., and Young, G. W., Tob. Sci., 1964, 8, 150. Newsome, J. R., Norman, V., and Keith, C. H., Tob. Sci., 1965, 9, 102. Guerin, M. R., Olerich, G., and Horton, A. D., J . Chromat. Scz., 1974, 12, 385. Papa, L. J., and Turner, L. P., J . Chromat. Sci., 1972, 10, 744. Kallio, H., Linko, R. R., and Kaitaranta, J., J . Chromat., 1972, 65, 355. Smith, R. G., Bryan, R. J., Feldstein, M., Levadie, B., Miller, F. A., Stephens, E. R., and White, Received January 13th, 1975 Accepted March 4th, 1975 and London, 1967, p. 250. N. G., Hlth Lab. Sci., 1972, 9, 75.

ISSN:0003-2654    

DOI:10.1039/AN9750000540

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

544 Analyst, August, 1975, Vol. 100, pp. 544-548 Improvements in the Atomic-fluorescence Determination of Mercury by the Cold-vapour Technique K. C. Thompson and R. G. Godden Shandon Southern Instruments Ltd., Frimley Road, Cambevley, Surrey, G 271 6 6ET Relatively simple modifications to a cold-vapour, mercury-fluorescence de- tector have resulted in a large increase in sensitivity and have decreased the time required for each measurement. A 20 detection limit of 0.02 ng was achieved. Recently a number of papers have appeared that describe various techniques for improving the sensitivity of determinations of mercury using the cold-vapour absorption technique. These techniques have included partitioning the mercury between the liquid phase and a fixed volume of air,l colloid flotation,2 amalgamation with gold,3 concentrating the mercury from the sample into a small volume of acidified potassium permanganate solution,* improved purging efficiency of the carrier gas through the reduced sample6 and pre-concentration in a cold trap.6 This last system, using a liquid nitrogen cold trap to isolate the mercury initially, gave the best absolute detection limit (0.2 ng of mercury).West,' by using a theoretical treatment, has shown that the atomic-fluorescence deter- mination of mercury, using a cold-vapour technique, should be more sensitive and produce considerably less interference from non-specific background absorption than the corresponding absorption technique. Previous reports on the cold-vapour, atomic-fluorescence determina- tion of mercury have borne out these observations.s-ll The detection limits reported vary from 0.5 to 3 ng of mercury.With minor modifications to an existing mercury cold-vapour fluorescence apparatus a detection limit of 0.02 ng of mercury could readily be obtained. Experimental The results were obtained by using a Shandon Southern Instruments A3600 atomic- absorption spectrophotometer, a modified Shandon Southern Instruments A3460 fluorescence mercury detector, and a Shandon Southern Instruments Autograph recorder. The system is depicted in Fig. 1. Argon was bubbled through the tin(I1) chloride reagent contained in cell A, and, on addition of the sample, mercury was released into the argon stream. The mercury then passed up through tube B and fluoresced in region C; a drying column was not required.A sheath unit, D, was constructed from acetyl resin and fitted around the top of tube B. Argon was fed into the sheath unit, which produced a laminar argon shield around the top of tube B. This device minimised entrainment of air into the sample - argon stream issuing from tube B, entrainment of air having been shown to result in severe quenching of the fluorescence radiation.*yg The following modifications were made to the mercury lamp unit of the A3460 instrument : the size of the lamp aperture, E, was reduced from 22 x 10 mm to 15 x 6.5 mm by means of a clip-on mask; a 6 mm 0.d. copper tube, F, was fitted on the rear of the lamp housing directly opposite to aperture E and a small flow of argon was passed into the lamp hous- ing through this tube and issued from the lamp aperture E.The mercury lamp, a Philips OZ4W low-pressure lamp,s was run at an optimum current of 0.38 A. Optimisation of Operating Parameters Wavelength and spectral band pass The 253.7 nm mercury line was used with the maximum spectral band pass of 6 nm. Minimisation of the Constant Background Level The constant background level was mainly attributed to direct specular reflection of light (at 253.7 nm) from the mercury lamp into the monochromator entry aperture. This back-THOMPSON AND GODDEN 545 ground was carefully minimised by removing both hollow-cathode lamps from the lamp turret and removing the remaining lens. Tube B was positioned just below the point at which direct specular reflection from the top of the tube commenced.For maximum sensitivity it was essential to minimise the back- ground level. Damping The A3600 atomic-absorption spectrophotometer was operated at maximum damping (the time constant being 10 s). The recorder output of the spectrophotometer (20 mV for full-scale deflection) was set to twice that of the recorder (10 mV for full-scale deflection) in order to prevent overloading of the amplification system. D' Front elevation I C E&Dl Argon - Scale: 1 cm = 2.5 cm (0.75 I min-'1 Fig. 1. Diagram of the improved mercury fluorescence detector: A, cell; B, 7 mm i.d. Pyrex tube; C, region where fluorescence occurs; D, sheath unit; E, lamp aperture; F, gas inlet pipe; and G, optical axis (orifice of tube B was 75 mm from monochromator entry slit).Gas Flows CeZZ A . Argon gas was used8sg and the optimum flow-rate was 0.75 1 min-l. Sheath. The signal amplitude was doubled, and the noise level on the base-line halved, when the argon sheathing gas was flowing. This effect was attributed mainly to minimised air entrainment. The optimum sheath gas flow-rate was 1.6 1 min-l. Lamp housing. Occasionally trace amounts of mercury would condense on the front face of the mercury lamp, especially with the sheath gas flowing. This condensation resulted in an increase in the constant background (specular reflection) level ; it also considerably increased the base-line noise level and caused drifting of the base-line. However, if argon was directed546 THOMPSON AND GODDEN : IMPROVEMENTS IN THE ATOMIC-FLUORESCENCE Analyst, VoZ.100 on to the rear face of the mercury lamp, through tube F, the mercury condensed on this rear face, resulting in a 30 per cent. increase in sensitivity and a stable base-line. The increase in the signal magnitude could be caused by various factors, e.g., removal of any ozone present between the lamp and tube B (ozone absorbs strongly at 253.7 nm), reduced self-absorption or self-reversal of the 253.7 nm mercury line, or even stepwise line fluorescence at 253.7 nm following excitation by the 185.0 nm mercury line. The argon flow-rate through the lamp housing was not very critical; in fact, a flow-rate of 1 1 min-l was used. A small increase in the constant background level (approximately 10 per cent.) was observed under these conditions. Reagents Mercuvy solutions.The mercury solutions were prepared just before use from a 10 pg ml-l stock solution of mercury in 1 per cent. V/V nitric acid. This last solution was prepared daily from a similar 1000 pg ml-1 stock solution. All of the calibrated flasks and pipettes used in this study were soaked in 50 per cent. V/V nitric acid for 1 week prior to use and all of the mercury and blank solutions contained 1 per cent. V/V of nitric acid.12 In order to achieve good, long-term stability (up to 5 months) of dilute (less than 10 ng ml-l) mercury solutions, the addition of 5 per cent. V/V of nitric acid and 0.01 per cent. m/V of dichromate ion has been re~0mrnended.l~ Tin(l1) chEoride solution, 2 per cent. m/V. A 2-g amount of tin(I1) chloride was dissolved in 20 ml of hydrochloric acid (36 per cent.m/V) and 80 ml of 1-5 M sulphuric acid were then added. Argon, at a flow-rate of 0.3 1 min-l, was continuously bubbled through this solution in order to remove any trace amounts of mercury and prevent oxidation of the tin(I1) chloride by air. Procedure A 1-ml volume of the 2 per cent. m/V tin(I1) chloride solution was introduced through the top of cell A. The cell was then connected to the cell head and the sample, contained in a 1-ml MLA pipette (Shandon Southern Instruments Ltd.), was inserted into the side-arm of cell A. After standing for 5 s, in order to allow for the removal of any entrained air, the sample was injected into the tin(I1) chloride solution and the peak recorded. The cell was then emptied, washed and the procedure repeated.Typical peaks are shown in Fig. 2. The peak width was 1.3 min. The maximum volume of sample that could be added to cell A was found to be 1 O m l . The calibration graph was linear over the range 0.02-200ng of mercury. A A A 1 min -.h Time + Fig. 2. Typical mercury trace. A, 1 ml of 0.001 pg ml-l mercury solution; B, 1 ml of blank solution; C , base-line stability. Voltage to photomultiplier = 770 V.August, 1975 DETERMINATION OF MERCURY BY THE COLD-VAPOUR TECHNIQUE 547 Results The new system was compared with the standard A3460 mercury detector. (This unit has been described previously.8) Table I gives a comparison of the two systems and it can be seen that the performance of the new detector system was much better than that of the standard system.Another advantage of the new system was the smaller change in the base-line level on removal of the pipette from the side-arm of cell A, equivalent to 0.03 ng of mercury for the new system and 1 ng of mercury for the standard A3460 system. This change in level was attributed to a change in the partial pressure of water vapour in the argon stream on removal of the pipette. TABLE I COMPARISON OF THE IMPROVED AND STANDARD MERCURY DETECTORS All absolute mercury masses are based on a l-ml sample volume. New Standard A3460 detector system detector system Relative sensitivity (comparison of peak heights for 2 ng of mercury addition) . . .. . . .. . . .. . . .. 9 1 Constant background level (expressed as ng of mercury) . . .. 0.4 4.5 Time for complete measurementlmin .. .. . . .. 1-5 4.5 2a noise level on base-line (expressed as ng of mercury) . . .. 0.015 0.3 The relative standard deviation (17 measurements) for a l-ml addition of a 1 ng ml-l mercury solution was 4.5 per cent. During the course of this study the blank solution (1 per cent. VjV nitric acid) gave signals equivalent to a mercury level between 0.02 and 0.05 ng ml-l. The standard deviation (13 measurements) on 1 ml of a blank solution giving a signal corre- sponding to 0.05 ng ml-l of mercury was 0.01 ng ml-l, which is equivalent to a 20 detection limit of 0.02 ng. In the limited time available breakdown procedures for organically bound mercury were not studied. However, on adding 1 ml of urine to cell A, containing 1 ml of 2 per cent. m/V tin(I1) chloride solution and 100 pl of 1 per cent.m/V silicone anti-foaming agent (BDH Chemicals Ltd.), reproducible signals were obtained for urine samples from non-exposed subjects. Mercury could also be detected in blood from non-exposed subjects by adding 1 ml of blood - water (1 + 9 V / V ) (which had been subjected to ultrasonic agitation for 15 min) to cell A containing 5 ml of 5 per cent. m/V tin(I1) chloride and 100 pl of 1 per cent. m/V silicone anti-foaming agent. The mercury signal from the blood (unlike aqueous standards) decreased markedly if the tin(I1) chloride con- centration was decreased below 4 per cent. m/V. This indicates a decreasing degree of breakdown of organically bound mercury compounds. For the complete breakdown of organically bound mercury compounds various procedures have been recommended.14-17 The system has been tested on urine and blood samples.Conclusions Iiiiprovements have been made to a previously described mercury-fluorescence detector system. An improvement in the detection limit of approximately 20 times has been achieved. This improvement results from various factors : decreasing the cell volume increased the sensitivity and decreased the time of measurement; an argon sheath minimised air entrain- ment and improved the base-line stability; a restricted lamp aperture resulted in a relative decrease in the constant background (specular reflection) level; and finally, cooling the rear face of the mercury lamp with a stream of argon improved both the sensitivity and the base-line stability and also prevented the formation of ozone (which absorbs strongly at 253.7 nm) between the lamp and the mercury vapour stream. References 1. 2. 3. 4. 6. 6. Ure, A. AT., and Shand, C. A., Analytica Chim. Acta, 1974, 72, 63. Voyce, D., and Zeitlin, H., Analytica Chim. Acta, 1974, 69, 27. O~afsson, J., Analytica Chim. Acta, 1974, 68, 207. Topping, G., and Pirie, J . M., Analytica Chinz. 14cta, 1972, 62, 200. Gilbert, T. R., and Hume, D. N., Analytica Chim. Acta, 1973, 65, 461. Fitzgerald, W. F., Lyons, W. B., and Hunt, C. D., Analyt. Chem., 1974, 46, 1882.548 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. THOMPSON AND GODDEN West, C. D., Analyt. Chem., 1974, 46, 797. Thompson, K. C., and Reynolds, G. D., Analyst, 1971, 96, 771. Thompson, I<. C., Lab. Pract., 1972, 21, 645. Muscat, V. I., and Vickers, T. J., Analytica Chim. Acta, 1971, 57, 23. Corcoran, F. L., Am. Lab., 1974, March, 69. Coyne, R. V., and Collins, J. h., Analyt. Chant., 1972, 44, 1093. Feldman, C., Analyt. Chem., 1974, 46, 99. Lindstedt, G., Analyst, 1970, 95, 264. Magos, L., and Clarkson, T. W., J . Ass. Off. Analyt. Chem., 1972, 55, 966. Umezaki, Y., and Iwamoto, K., Japan Analyst, 1971, 20, 173. Lopez-Escobar, L., and Hume, D. N., Analyt. Lett., 1973, 6, 343. Received March 7th, 1975 Accepted March 25th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000544

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, August, 1975, Vol. 100, $9. 549-554 Suppression of Iron(1ll) Interference in the Determination of Iron(l1) in Water by the 1,lO- Phenanthroline Method 549 Hubert Fadrus and Josef Maly Water Management Board, Vodohospoddkkd Sprdva, Brno, Czechoslovakia A method for the determination of iron(I1) in water in the presence of iron(I1I) with 1,lO-phenanthroline is described, in which the interfering effect of iron(II1) ions is suppressed by masking with complexones. In the absence of a chelating agent for iron(II1) the colour intensity of samples being analysed is unstable, owing to redox processes induced by the effect of 1,lO-phenan- throline on the iron(I1) - iron(II1) system, causing reduction of iron(II1) and formation of the corresponding coloured iron(I1) chelate compound in a stoicheiometrically proportional concentration.The advantages of com- plexones, especially nitrilotriacetic acid, over other chelating agents are discussed. The method for the photometric determination of iron, after converting it into iron(I1) ions, with 1 ,lo-phenanthroline has been thoroughly investigated and proved to be satisfact0ry.l If, however, it is required to differentiate between the two valency forms of iron by using this method, the omission from the basic procedure of the addition of reductant used in the determination of iron(I1) in the presence of iron(II1) ions is recommended.2 The results thus obtained are, however, rather unsatisfactory, especially because of the instability of the resulting colour intensity with time.3 This drawback also becomes apparent in the method in which iron(I1) is determined indirectly by calculation from the difference between the absorbances of the iron(I1) - iron(II1) - 1,lO-phenanthroline system at two specific wave- lengths .4 The applicability of the methods for the determination of iron(I1) with 1 ,lo-phenanthroline in the presence of excess of iron(II1) ions is restricted by the following requirements: the iron(I1) concentration must be higher than 1 mg l-l, an amount of reagent equivalent to a t least 30 times the total iron content must be used and the absorbance read within 10 to 15 min; also, the sample to be analysed must be free from all foreign matter and protected against direct sunlight .3 However, even when these conditions are satisfied, the results obtained are not reliable.5 Similarly, the use of 2,9-dimethyl-l,lO-phenanthroline instead of 1 ,lo-phenanthroline does not give satisfactory results.The cause of the poor reproducibility of the results for the determination of iron(I1) with 1,lO-phenanthroline in the presence of iron(II1) ions is attributable to the different stabilities of the iron(I1) and iron(II1) chelate compounds formed with 1 ,lO-phenanthroline,s the presence of which gives rise to a shift in the redox potential of the Fe2+ - Fe3+ system to more negative values in the reaction Fe3+ + 3(phen) + e- = [Fe(phen),12+ with reduction of iron( 111) ions and formation of the corresponding coloured iron(I1) chelate compound in stoicheiometrically proportional concentration. The reaction takes place slowly, but is more rapid in the presence of reductants as electron acceptors.In order t o achieve accurate results in iron(I1) determinations under such conditions it is necessary t o shorten the time of contact of the iron(II1) ions with the reagent to a minimum, so that errors that arise from the increase in concentration of the component being determined would be negligible. Compliance with this requirement, e.g., by preliminary separation of iron(II1) ions in the form of iron(II1) hydroxide, may introduce further errors caused by interference with the balance between acid - base and redox conditions in the Fez+ - Fe3+ system.' The masking of the iron(II1) ions by the addition of chelate-forming compounds and this method of removing the interference from iron(II1) ions in the determination of iron(I1) with 1 , l O - phenanthroline have been further studied.550 FADRUS AND MAL+: SUPPRESSION OF IRON(III) INTERFERENCE IN THE Analyst, VoZ.100 Experimental The colour intensities of the 1,lO-phenanthroline complexes with iron(I1) and iron(II1) ions in the presence of other compounds were measured in 2-cm cells on a Pulfrich photometer with ELPHO-2 supplementary equipment, using an S-51 filter (510 nm). The necessary concentrations of the components in the final solutions were obtained by adding the reagents in the order given in the figure captions. The required concentration of iron(I1) ions was obtained by diluting ammonium iron(I1) sulphate standard solution (see Reagents). Trace amounts of iron(II1) salts in this solution were reduced with a Jones reductor.The required concentration of iron(II1) was obtained by diluting ammonium iron( 111) sulphate stock solution containing 500 mg 1-1 of iron(II1) and 2 ml 1-1 of concentrated sulphuric acid. The exact iron( 111) concentration was determined gravimetrically. In order to remove trace amounts of iron(I1) the solution was always freshly prepared after evaporation to fumes with perchloric acid and hydrogen peroxide and re-dissolution. Solutions were adjusted to the appropriate pH by use of the following buffers: glycocoll (glycine) for pH 2-5, acetate for pH 5.5 and borax for pH 8.0. Results and Discussion Under the conditions of the determination of iron(II), iron(II1) ions react with 1 ,lo-phenanthroline to form a labile yellow-coloured chelate compound, [Fe(phen) ,I3+, which, in the presence of excess of reagent, changes slowly into a stable red chelate compound, [Fe(phen),] 2+.The rate of the reduction is dependent on the pH of the medium as well as on the reagent concentration (Fig. 1); however, contrary to a previous r e p ~ r t , ~ we found that it is not affected by diffused light. A considerable acceleration in the rate of reduction of the [Fe(phen),13+ chelate is obtained in the presence of other substances that have reduc- ing characteristics (Fig. 2). I D/ 1 I I I I I 10 20 30 40 50 60 Time/min Fig. 1. Colour development in the solution containing 1,lO-phenanthroline and Fe(II1) ions. [Fe(III)], 12.5 mg 1-l. Concentration of 1,lO-phen- antliroline: A, B and C, 1.25 x M ; and D, 3.75 x 10-9 M.pH of solution: A, 5.5; B, 2.5; C, 1.95; D, 2-5. By adding to the solution after the addition of 1,lO-phenanthroline, or simultaneously with it) compounds that mask iron(II1) ions, it is possible to suppress the course of the reduction reactions or to decrease their rate to a minimum. Neither tartaric nor citric acid proved satisfactory for this purpose, not only because of insufficient stability of the corresponding iron( 111) chelate compounds, but especially because, in the presence of 1,lO-phenanthroline, there occurs in the corresponding tartrate and citrate chelate compounds an intramolecular redox process in which the central iron(II1) ion is reduced by its own chelate, forming a ligand to iron(I1) ion, which changes to a stable complex [Fe(phen),12+ (Fig.3, curves A and B). Use of fluoride provides reliable results in an acidic medium, but because of the aggressive effects of hydrogen fluoride on the glass cells of the photometer, it has only a limited applic- ability. In a neutral medium the masking efficiency decreases and in an alkaline medium FeF,,- decomposes as a result of hydrolysis. Diammonium hydrogen orthophosphate, recom-August, 1975 DETERMINATION OF IRON(II) IN WATER WITH 1, 10-PHENANTHROLINE 551 0-30 r *= 0.20 2 +-' - 0.10 C 1 I I I I I 10 20 30 40 50 60 Time/min Fig. 2. Effect of reductants on the colour develop- ment in the solution containing 1,IO-phenanthroline and Fe(II1) ions: A, no reductant added; B, 250 mg 1-1 of NH,HS; C, 500 ml 1-1 of sewage with B.O.D., = 82 mg 1-1 of 0,; and D, 250 mg 1-1 of sucrose.[Fe(III)], 12.5 mg ml-l; [l,lO-phenanthroline], 1.25 x 10-3 RI. PH, 2.5. mended in extraction processes in combination with 2,9-dirnethyl-l,lO-phenanthroline, did not prove satisfactory in direct photometry,8 because the pH (about 2.2) cannot easily be controlled and the gradual formation of a turbidity occurs as iron(II1) phosphate is only slightly soluble. By lowering the pH to 1.5 it is possible to remove the turbidity while maintaining the masking effect of the phosphate ions on iron(II1) ions, but an undesirable retardation of the colour development also occurs. 0.30 I I I I I I I 0 10 20 30 40 50 60 Ti me/min Fig. 3. Colour development in the solution contain- ing Fe( 111) ions, 1,lO-phenanthroline and complexones.[Fe(II)], 0.5 mg 1-l; [Fe(III)], 12.5 mg 1-l; [l,lO-phen- anthroline], 1.25 x 1 0 - S M ; and [citric acid], 1 g1-1, added as a reductant, together with 1, lo-phenanthro- line. Complexones are added after the addition of 1,lO-phenanthroline. pH of Curve solution Complexone A 2-5 - B 8.2 - C 2.5 0.01 M NTA D 8.2 0.01 M NTA + 0.01 M DCTA Reduction of the [Fe(phen) complex with 1 ,lo-phenanthroline can be retarded or stopped by the addition of complexones (Fig. 3, curves C and D). The complicated mechanism of the interaction of the components in this system is a function of the stability of the chelate compound formed by the iron with 1,lO-phenanthroline and with the complexone, which is expressed by the corresponding stability constants. An unsuitable choice of the order of addition of the reagents leads therefore to non-reproducible and erroneous results, thus giving rise to criticism of the use of complexones.8 This can be explained as follows: if to a solution containing both iron(I1) and iron(II1) ions, complexones are added together in a mixture with 1,lO-phenanthroline, they have the effect shown in Fig.4. By adding fluoride0-50 I I I 1 I I 0 10 20 30 40 50 s;f Time/m in Fig. 4. Effect of complexones on the colour development in the solution con- taining 1,lO-phenanthroline and Fe(I1) ions: A, without complexone or with NTA; B, with EDTA; and C , with DCTA. [Fe(II)], 0.75 mg 1-l; [l,lO-phenanthro- line], 1.25 x M ; [complexone], 2 x M. pH, 2.5. Complexones are added together with 1.10-phenanthroline. The probable cause of the reduced absorbance values and, in turn, of the erroneous results of the iron(I1) determination, is that with such an effective lowering of the potential of the Fe2+ - Fe3+ system by the masking of iron(II1) ions with the complexone used, the formation of the [Fe(phen),12+ chelate compound evidently competes with the rapid and parallel oxida- tion of iron(I1) ions by the oxygen dissolved in the given mixture: [Fe(phen),12+ + H2Y2- = Fey- + 3(phen) + e- This assumption was verified by adding fluoride or complexone to the reaction mixture prior to the 1,lO-phenanthroline, which, in the presence of a sufficient concentration of dissolved oxygen, caused an almost immediate oxidation of iron(I1) ions to such an extent that the reaction between 1,lO-phenanthroline and the component to be determined was completely negative.On the other hand, the [Fe(phen),12+ already produced, even in the presence of oxygen, is stable towards complexone inasmuch as the intensity of its colour changes only very slowly with time, according to the conditions in the reaction mixture. With the exception of fluoride and NTA, whose chelate compounds with iron(II1) are less stable than the similar chelate compounds with EDTA and DCTA and which can, therefore, be applied in combination in a mixture with 1,lO-phenanthroline, it is necessary to add EDTA and DCTA to the reaction mixture separately only after the development of the colour is completed. In such an instance, when EDTA or DCTA is added 1 min following the addition of 1,lO-phenanthroline to the solution containing iron in both valency forms, the absorbance of the resulting mixture is identical with that of a similar sample containing only an equimolar concentration of iron(I1) salt.Changes in the colour intensity of the investigated solutions brought about by the redox processes of the iron(I1) and iron(II1) compounds in 1 min are negligible. The significance of the order of addition of reagents in the determination of iron(I1) has to be taken into consideration in all instances when the interfering components are masked by complexones, otherwise even procedures recom- mended in the literature may be valueless. I t is possible, however, to use reaction mixtures that are neutral or alkaline (e.g., by use of a borax buffer of pH 8) when reducing substances that cause less interference, e.g., sulphides, are present.The addition of chelating substances that give soluble complexes with iron(II1) ions in this pH range is, under these conditions, essential in order to prevent the occurrence of turbidity due to the hydrolysis of the iron(II1) salts present. Of these chelating agents, 0 " 0 0 A 0 -August, 1975 DETERMINATION OF IRON(II) IN WATER WITH 1 ,lo-PHENANTHROLINE 553 however, only NTA can be used and can be added together with the reagent, because owing to the action of EDTA or DCTA, which is added after the development of the [Fe(phen),J2+, the iron(II1) hydrolysates formed dissolve only very slowly in the alkaline medium. As the stability of this complex ion is more marked in the presence of EDTA and DCTA, it is advantageous to combine all the complexones in such a way that NTA is added in admixture with 1,lO-phenanthroline followed after 1 min by EDTA or even better by DCTA (Fig.3, curves C and D). Results for the determination of iron(I1) in the presence of iron(II1) in underground water from South Moravia obtained by using the recommended method and the procedure according to reference 2 are given in Tables I and 11. TABLE I COMPARISON OF RESULTS FOR IRON(II) DETERMINATION BY USING THE RECOMMENDED CONTAINING IRON(II) AND IRON(III) SALTS METHOD AND PROCEDURE ACCORDING TO REFERENCE 2 ON PREPARED SOLUTIONS 3.18 5.30 All results are expressed in mg 1-l. recommended method* Iron( 11) by Iron(I1) by procedure in reference 2 added added Found -ce Found -ce 1-06 10 1.10 + 0.04 1.24 +O*lS 25 0.98 - 0.08 1-28 + 0.22 50 1.12 + 0.06 1-40 + 0.34 100 1-15 + 0.09 1.51 + 0.45 10 3.26 + 0.08 3.30 +0*12 25 3.23 + 0.05 3.38 + 0.20 50 3.10 - 0.08 3-56 + 0.38 100 3.32 +0*14 3-60 + 0-42 Iron(I1) Iron(II1) 10 5.32 + 0.02 5.41 + O .l l 25 5-39 + 0.09 5.55 + 0.25 50 5.42 +0.12 5.65 + 0.35 100 5.41 + O . l l 5-82 + 0.52 * Colour intensity measured 30 min after addition of reagents. The differences between the values obtained by the two methods for colour intensity during the first 3 min provide evidence for interpreting the influence of iron(II1) on the iron( 11) determination as described above but cannot be explained theoretically because of lack of knowledge of the reaction mechanism involved in the development of the colour intensity of the complex salt of iron(I1) with 1,lO-phenanthroline during the interval 0-3 min.TABLE I1 COMPARISON OF RESULTS OBTAINED ON UNDERGROUND WATER FOR DETERMINATION OF IRON(II) IN THE PRESENCE OF IRON(III) BY RECOMMENDED METHOD AND PROCEDURE IN REFERENCE 2 Iron(I1) contentlmg l-l, r \ A Iron (111) Procedure in reference 2* Recommended method* content/ A I I mg 1-1 a b (b - a) 3 (b - a) 19.1 1-28 1.40 0.12 1-25 1.28 0-03 31.2 2.95 3.22 0.27 2-90 2-96 0.06 22.3 1-42 1.61 0.19 1-36 1-39 0.03 11-8 0.92 1.04 0.12 0.88 0.90 0.02 25.6 1.85 2.08 0.23 1-78 1.84 0-06 16.0 1.08 1.17 0.09 1-05 1-07 0-02 Method * Colour intensity measured a, 3 min and b, 30 min after addition of reagents. Reagents Bz@''r mixture. Mix 5 volumes of 0.025 M 1,lO-phenanthroline hydrochloride solution, 6 volumes of 0.5 M glycocoll solution adjusted to pH 2.9 with 0.5 N hydrochloric acid and 1 volume of a 0.1 M solution of the sodium salt of nitrilotxiacetic acid (NTA) immediately before use.554 FADRUS AND M A L ~ Standard ammonium iron(II) solution.Dissolve 0.702 g of the reagent in distilled water containing 2 ml of concentrated sulphuric acid and make the volume up to 11; 1 ml of the solution contains 0.01 mg of iron(I1). Its exact concentration is determined by titration with standard permanganate solution. Apparatus Pulfrich photometer with ELPHO-2. S-51 filter (510 nm). Measuring cells, 2 em. Procedure Place a neutral or slightly acidic solution of the sample containing up to 125 pg of iron(I1) and up to 2500 pg of iron(II1) in a 50-ml calibrated flask, adjust the volume to about 25 ml and add 10.0 ml of buffer mixture.Agitate the mixture and dilute it to the mark. During the period 3 to 30 min after addition of the buffer mixture, measure the colour intensity of the solution at 510 nm against a sample prepared in the same way without the addition of 1,lO- phenanthroline (in order to compensate for the colour of the iron(II1) - complexone). The calibration graph, which is linear up to an iron(I1) concentration of 5 mg 1-1, is plotted under identical working conditions. Interferences By using the procedure described for the determination of 0 to 5 mg 1-1 of iron(II), the presence of up to 100mg1-1 of iron(II1) in the analysed solution is eliminated as well as the effect of all ions that form stable chelate compounds with NTA.Tungstate, molybdate, iodide, rhodanide, perchlorate and cyanide precipitate the 1,lO-phen- anthroline - iron(I1) chelate compound or form with it soluble ionic associates, which can be extracted into non-polar solvents. Vanadium(V) compounds interfere by forming a brown colour, phosphate retards the development of the colour and nitrite gives a yellow colour. Conclusion It has been proved that contrary to previous reports the determination of iron(I1) ions with 1,lO-phenanthroline in the presence of iron(II1) ions cannot be achieved simply by omitting the addition of the reductant used in normal procedures because, owing to the lack of stability of iron(II1) ions in the given reaction mixture, in which the action of redox processes is stimulated by the strong chelating effect of the reagent on the iron(I1) ions, erroneous and insufficiently reproducible results are obtained as a consequence of the in- stability of the developed colour. The interference of iron(II1) ions can be suppressed by adding chelating substances, the most suitable of which is NTA.Knowledge of the reaction mechanism as well as of the velocities of the mutually competing processes between iron(I1) and iron(II1) ions, reagent , complexone used and redox active components that may be be present in the solution being analysed permits the introduction of the chosen complexone correctly into,the order of addition of reagents, even in those photometric procedures in which complexones are used for masking other interfering components. When highly reducible compounds are present, the deter- mination of iron(I1) ions in the presence of iron(II1) ions cannot be achieved by this method. References 1. 2. 3. 4. 5. 6. 7. 8. Vydra, F., and Kopanica, M., Chemist AnaZyst, 1963, 52, 88. “Standard Methods for the Examination of Water and Waste water,” Thirteenth Edition, American Lee, F. G., and Stumm, W., J . Am. Wat. W k s Ass., 1960, 52, 1567. Harvey, A. E., Smart, J. A., and Amis, E. S., Analyt. Chem., 1955, 27, 26. Ghosh, M., J . Am. Wat. Wks Ass., 1967, 59, 897. Kolthoff, I. M., Lee, T. S.. and Leussing, 0. L., Analyt. Chew., 1948, 20, 385. Verbeek, F., Bull. SOC. Chim. Bdg., 1961, 70, 423. Clark, L. J., Analyt. Chem., 1962, 34, 348. Public Health Association, New York, 1971, p. 189. Received October 21st, 1974 Accepted February 27th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000549

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, August, 1975, Vol. 100, $9. 555-562 555 A Composite Scheme for the Analysis of Steels by Atomic-absorption Spectroscopy Using the Air - Acetylene Flame W. R. Nall, D. Brumhead and R. Whitham Ministry of Defence, Materials Quality A ssurance Directorate, Bragg Laboratory, Janson Street, Shefield, S9 2L J A composite scheme is described for the determination of chromium, molyb- denum, manganese, nickel and copper in all types of steel. The scheme is based on a single sample solution containing 1 g of steel per 100 ml. Recent developments in technique have enabled all these elements to be determined with the air - acetylene flame and the scheme can be extended as required to include other elements such as lead and cobalt. Previous schemes for the determination of several elements on a single sample mass of alloy have involved the use of the nitrous oxide - acetylene flame1 in order to overcome difficulties in the measurement of refractory elements or they have been concerned with relatively low contents of alloying elements2 A limited scheme3 has been described for determining high contents of manganese, chromium and nickel in high-alloy steels by using an air - acetylene flame.In this last paper the use of the secondary resonance line of chromium was recom- mended but no reference was made to the precautions that are necessary* in the determination of this element in this type of flame. The ability to apply the direct atomic-absorption technique to the determination of larger amounts of alloying elements has long been recognised to be desirable but the accuracy and reproducibility of the method have been such that elaborate schemes of averaging a number of readings have been required for this type of analysis.In recent years, however, there has been such a refinement of instrumentation that the required degree of accuracy should now be attainable directly and modern instruments should bring the method into line with other accepted techniques such as titrimetric analysis and molecular spectrophotometry. If there is no requirement to determine elements that form refractory oxides, such as silicon, aluminium or titanium, it would be useful to have a composite scheme that includes the elements chromium and molybdenum that is based on the use of the air - acetylene flame alone. There are some laboratories in which the use of nitrous oxide is undesirable or pro- hibited on safety grounds and for these situations the proposed scheme would be applicable. The problem of determining molybdenum in an air - acetylene flame5 has been solved by the addition of ammonium chloride to the analyte and recently it has been shown4 that chromium can be determined by incorporating quinolin-8-01 as a releasing agent.Choice of Experimental Conditions The usual working concentration of sample solution recommended for steel analysis1 is 1 per cent. m/V, which is dilute enough to prevent deposition of salt in the burner jaws and yet provides a sufficient concentration of the minor elements in the sample for their accurate determination. In order to accommodate alloying elements that are present in larger amounts, dilutions have to be made to this initial sample solution as suggested by Thomerson and Price.1 They recommended that the iron concentration should be restored to the initial 1 per cent.m/V in order to minimise interference and suppression and to equalise these effects between samples and standards. While it was recognised that a constant iron concentration was required, the actual level does not seem to have been determined directly. This aspect has been examined and the results are given below. As part of the development of the scheme the linear working ranges for each element were also determined and the dilutions given below are such that the analytes Crown Copyright.556 NALL et al. : A COMPOSITE SCHEME FOR THE ANALYSIS OF Analyst, VoZ.100 are always within these ranges. It was found that in all instances the best ranges were those which gave absorbances below 0.30. The effect of various iron concentrations on the absorbance of the five elements chromium, molybdenum, manganese, nickel and copper in the proposed scheme was investigated (Fig. 1) ; the results given were obtained in the presence of the releasing agents described below (see Iron in analyte, per cent. mlV Fig. 1. Effect of various iron concentrations on absorbance. A, Chromium 10p.p.m.; B, copper 2.5 p.p.m. ; C, molybdenum 20 p.p.m. ; D, nickel 10 p.p.m. ; and E, manganese 2-5 p.p.m. Instrument conditions as in Table 11. Method). It was found that copper, nickel and manganese were virt ially unaffected by varying the iron concentration but chromium and molybdenum were subject to serious inter- ference, and it was decided to investigate the effect on these two elements more fully, by measuring the absorbance for chromium and molybdenum in the presence of their appropriate releasing agents and of increasing amounts of iron.The results for chromium are shown in Fig. 2, from which it can be seen that at all con- centrations of quinolin-8-01, decreasing the iron content results in increased absorbance for a given amount of chromium. However, a practical limit is set to the minimum amount of sample (hence the iron content) that can be used in order to satisfy the sensitivity require- 0.075 I I O-O45X/ 0.040 4 0 5 10 15 20 Volume of 5 per cent. quinolin-8-01 solution/mi Fig.2. Releasing effect on 6p.p.m. of chromium of various amounts of 5 per cent. quinolin-8-01 solution in different concentrations of iron solution; final volume 100 ml. Iron solution: A, 0.1 per cent.; B, 0.2 per cent.; C, 0.5 per cent; and D, 1.0 per cent.A ugust , 1975 STEELS BY AAS USING THE AIR - ACETYLENE FLAME 557 ments previously determined. The limiting element in the proposed scheme is molybdenum, which, when present in its lowest range (0-1-0 per cent.), allows only a five-fold dilution in order to maintain the optimum concentration of 0-20 p.p.m. of molybdenum in the analyte (Table I). Consequently, with an initial solution of 1 per cent. m/V, the maximum dilution required to satisfy all conditions is to 0.2 per cent. m/V. TABLE I SAMPLE DILUTION Element Nickel Manganese and copper Chromium Molybdenum Range in sample, per cent.0-1.0 14-4.0 2.5-10.0 5.0-20-0 0-0-25 0.10-1.0 0.5-2.0 1.0-5.0 0-1-0 14-5.0 5*0-10*0 10-20 0-1-0 0*5-2*0 1.0-4.0 Concentration range in stock sample solution, p.p.m. 0-100 100-400 250-1000 500-2000 0-25 10-100 50-200 100-500 0-1 00 100-500 500-1000 1000-2000 0-100 50-200 100-400 Dilution of initial to final volume* /ml 20 to 100 5 to 100 (a) 10 to 100 (b) 20 to 100 (a) 10 to 100 (b) 10 to 100 10 to 100 (a) 25 to 100 (b) 10 to 100 (a) 5 to 100 (b) 25 to 100 (a) 5 to 100 (b) 10 to 100 10 to 100 (a) 10 to 100 (b) 20 to 100 (a) 10 to 100 (b) 10 to 100 (a) 10 to 100 (b) 5 to 100 20 to 100 10 to 100 5 to 100 Dilution factor 5 20 50 100 10 40 80 200 10 50 100 200 5 10 20 Concentration range in analyte, p.p.m.0-20 5-20 5-20 5-20 0-2.5 0.25-2'5 0.625-2.5 0.5-2.5 0-10 2-10 6-10 5-10 0-20 5-20 5-20 Stock iron solution added(m1 0 5.0 5.0 5.0 2.5 5.0 6.0 5-0 2.5 5.0 5.0 5.0 0 2.5 5.0 * (a) and (b) denote two stages of a dilution. The effect of varying the concentration of total solids in the analyte was determined and it was found that the analyte flow-rate through the nebuliser decreased in a linear manner from 3.8 to 3.2 ml min-l as their concentration was increased from zero to 2-0 g per 100 ml. The result of this effect can be seen in Fig. 2, in which there is a significant suppression of absorbance as the iron content is increased from 0.1 to 1.0 per cent. m/V in the presence of 20 ml of 5 per cent. quinolin-8-01 solution. The use of a composite scheme designed for as wide a range of alloys as possible requires that two conditions must be satisfied: the iron content of the analyte should be identical in samples and standards and the concentration of the total solids in the analyte should be constant so as to equalise nebulisation rates between samples and also between samples and standards.Having fixed the concentration of the sample solution at 0-2 per cent. m/V in order to satisfy the above criteria, it can be seen from Fig. 2 that maximum absorbance for chromium is obtained with an addition of 5 ml of 5 per cent. m/V quinolin-S-ol solution and this amount is incorporated in the scheme. Similar considerations applied to molybdenum (Fig. 3) indicated that, with a 0.2 per cent. m/V sample solution, satisfactory absorbance is obtained with 10 ml of 10 per cent.m/V ammonium chloride solution. Ottaway and Pradhan* stated that the interference of iron in chromium determinations was strongly dependent on the flame height at which the absorbance measurements were made. In order to determine the correct burner height with the SP1900 instrument, absor- bance measurements were made at various burner heights for two solutions: (a), 5 p.p.m. of chromium plus 5 ml of 5 per cent. m/V quinolin-8-01 and (b), as for (a) but with 0.2 per cent.558 NALL et aZ. : A COMPOSITE SCHEME FOR THE ANALYSIS OF ArtaZyst, VoZ. 100 of iron added. The results of these determinations are given in Fig. 4, which shows that, at an observation height of 1.0 cm above the burner top, the presence of 5 ml of 5 per cent.m/V quinolin-8-01 completely removes the interference of iron with gas flow-rates of 1.2 1 min-1 for acetylene and 4.8 1 min-l for air. This burner height setting is recommended in the method given below, as is the above flame composition, which is a compromise between complete removal of iron interference and maximum sensitivity. 0.05 0 5 10 15 Volume of 10 per cent. ammonium chloride solution/rnl Releasing effect on 20 p.p.m. of molyb- denum of various amounts of 10 per cent. ammonium chloride solution in different concentrations of iron solution; final volume 100 ml. Iron solution: A, 0.1 per cent. ; B, 0.2 per cent. ; C, 0-5 per cent. ; and D, 1.0 per cent. Fig. 3. A similar experiment was made in order to determine the optimum conditions for the molybdenum determination and the results are shown in Fig.5, from which it can be seen that the interference of iron is completely removed at an observation height of 1.0 cm and, with a flame composition corresponding to 1.8 1 min-l for acetylene and 4.8 1 min-l for air, there is adequate sensitivity. This method of determining the optimum burner height setting for a given fuel to air ratio can be used for any type of instrument and is a necessary exercise if the information is not already established. As a result of the above investigations it was decided to finalise the iron concentration at 0.2 per cent. in the method and to base the scheme on the following criteria: to restrict the scheme to the elements most commonly required to be determined in steel analysis; to use only the air - acetylene flame; and to dilute the sample solutions in order to obtain the optimum element concentration so that linear response could be obtained throughout each working range.Apparatus absorption spectrophotometer with digital read-out. values is 0.0-1.999. Solutions (Note 1) Stock iron solution, 4 per cent. m/V. Dissolve 20 g of high-purity iron (BCS 260/4) in 100 ml of hydrochloric acid (sp. gr. 1-18> and cautiously oxidise it with the minimum amount of nitric acid (sp. gr. 1.42). Cool and dilute the solution to 500 ml with de-ionised water. Quinolin-8-01 solution, 5 per cent. m/V. Dissolve 25 g of quinolin-8-01 in 25 ml of hydro- chloric acid (sp. gr. 1.18). Cool and dilute the solution to 500 ml with de-ionised water. Ammonium chzloride solution, 10 per cent.m/V. Dissolve 25 g of ammonium chloride in de-ionised water and dilute the solution to 250 ml. Primary standard metal solutions (nickel, copper, manganese, rnolybdenwn, chromium). Dissolve Method All the determinations were made with the Pye Unicam SP1900 double-beam atomic- The calibrated range of absorbanceAugust, 1975 STEELS BY AAS USING THE AIR - ACETYLENE FLAME 559 0.09 0.08 0.07 0-06 0.05 m -2 54 2 0.04 0.03 0.02 0.01 0.26 0.24 0.22 0.20 0.18 0.1 6 a 0.14 -2 0, Q 0.12 0.10 0.08 0.06 0.04 0.02 a - - - - - - - - - ,,,,,I 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Height of light path above burner top/cm Fig. 4. Effect of reading height and flame composition on absorbance of 5 p.p.m. of chromium: a, 5 p.p.m.of Cr + 5 ml of 5 per cent. m/V quinolin-8-01; b, 5 p.p.m. of Cr + 5 ml of 5 per cent. quinolin-8-01 + 0.2 per cent. of Fe. Air flow-rate: 4-81 min-l. Acetylene flow-rate: x, 1-6; f , 1.4; 0, 1.2; 0, 1.0; and A, 0.8 1 min-'. I I 01 I I 1 I I 1 0.4 0-5 0-6 0.7 0.8 0.9 1.0 Height of light path above burner top/cm Fig. 5 . Effect of reading height and flame composition on absorbance of 10 p.p.m. of molybdenum: a, 10 p.p.m. of Mo + 10 ml of 10 per cent. m/V NH,Cl; b, 10 p.p.m. of Mo + 10ml of 10 per cent. m/V NH,C1 + 0.2 per cent. of Fe. Air flow-rate: 4.8 1 min-1. Acetylene flow-rate: x , 2.0; +, 1.8; 0, 1.6; @, 1.4; and A, 1.2 1 min-l. 0.5 g of the pure metal in 50 ml of hydrochloric acid (sp. gr. 1-18), warming gently. Cool and dilute the solutions to 500 ml (Note 2) (1 ml of solution = 1 mg of metal).Dilute standard nickel, copper and manganese solutions. Dilute 20 ml of the primary standard nickel solution to 100 ml (1 ml of solution 0.2 mg of nickel). Dilute 5 ml of the primary standard manganese and copper solutions to 200 ml (1 ml of solution = 0.025 mg of metal). Solution A . Transfer 25-ml aliquots of each of the dilute nickel, copper and manganese standard solutions into a 500-ml calibrated flask. Add 25 ml of stock iron solution (5 ml of solution = 0.2 g of iron) and dilute to the mark. This solution contains 1.25 p.p.m. of man- ganese, 1.25 p.p.m. of copper and 10 p.p.m. of nickel. Solution B. Transfer 50-ml aliquots of each of the dilute standard solutions into a 500-ml calibrated flask. Add 25 ml of stock iron solution and dilute to the mark. This solution contains 2.5 p.p.m.of manganese, 2.5 p.p.m. of copper and 20 p.p.m. of nickel. For the blank solution, dilute 25 ml of the stock iron solution to 500 ml with de-ionised water. This solution is for use in determining the elements manganese, copper and nickel. DiZute standard molybdenum solution. Dilute 20 ml of the primary standard molybdenum solution to 100 ml (1 ml of solution = 0.2 mg of molybdenum). Solution C. Transfer a 25-ml aliquot of the dilute standard molybdenum solution into a560 NALL et al. : A COMPOSITE SCHEME FOR THE ANALYSIS OF Analyst, Vol. 100 500-ml calibrated flask. Add 25 ml of stock iron solution and 50 ml of 10 per cent. ammonium chloride solution. Dilute to the mark. This solution contains 10 p.p.m.of molybdenum. Solution D. Transfer a 50-ml aliquot of the dilute standard molybdenum solution into a 500-ml calibrated flask. Add 25 ml of stock iron solution and 50 ml of 10 per cent. am- monium chloride solution. Dilute to the mark. This solution contains 20 p.p.m. of molyb- denum. For the blank solution, transfer 25 ml of the stock iron solution and 50 ml of 10 per cent. ammonium chloride solution into a 500-ml calibrated flask and dilute to the mark with de-ionised water. This solution is for use in determining the element molybdenum. Dilute standard chromium soZution. Dilute 25 ml of the primary standard chromium solution to 250 ml (1 ml of solution = 0.1 mg of chromium). Solution E. Transfer a 25-ml aliquot of the dilute standard chromium solution into a 500-ml calibrated flask.Add 25 ml of stock iron solution and 25 ml of 5 per cent. quholin-8-01 solution. Dilute to the mark. This solution contains 5 p.p.m. of chromium. Solution F. Transfer a 50-ml aliquot of the dilute standard chromium solution into a. 500-ml calibrated flask. Add 25 ml of stock iron solution and 25 ml of 5 per cent. quinolin-8-01 solution. Dilute to the mark. This solution contains 10 p.p.m. of chromium. For the blank solution, transfer 25 ml of the stock iron solution and 25 ml of 5 per cent. quinolin-8-01 solution into a 500-ml calibrated flask and dilute to the mark with de-ionised water. Preparation of Stock Sample Solution Weigh 1 g of sample into a 250-ml conical beaker and dissolve it in 10 ml of hydrochloric acid (sp.gr. 1.18), warming gently. When solvent action ceases, add nitric acid (sp. gr. 1-42) dropwise until oxidation is complete and evaporate the solution just to dryness in order to remove the excess of nitric acid. Dissolve the residue in 10 ml of hydrochloric acid (sp. gr, 1-18), warming to obtain complete dissolution. Cool and dilute the solution with approximately 25 ml of water and note if any insoluble material remains. Filter the solution, if necessary, through a Whatman No. 40 filter-paper, washing the residue well with water. Dilute the filtrate to 100 ml with water. Certain types of high-silicon alloys may produce a further precipitate of silica at this stage, which must be filtered off before proceeding. The above procedure may not give a complete solution of all types of alloy steel and in- soluble portions would have to be filtered off, ignited and treated with hydrofluoric acid in a platinum vessel in order to ensure that all of the sample is brought into solution.In this event it is important to remove fluoride by evaporation with hydrochloric acid before combining the solution derived from the insoluble portion with that of the acid-soluble portion in the first filtrate. Determination of Manganese, Nickel and Copper Transfer a suitable aliquot (see Table I) of the stock sample solution into a 100-ml calibrated flask, add a suitable amount of the stock iron solution (see Table I) and dilute to the mark with de-ionised water. Aspirate the appropriate blank solution and zero the spectrophoto- meter, using the conditions specified in Table 11.Continue as described under Absorption Measurement (see below). This solution is for use in determining the element chromium. TABLE I1 INSTRUMENT CONDITIONS Flame, air - acetylene, 10 cm; burner, in line; air flow-rate, 4.5-5.6 1 min-l; and integration period, 4 or 20 s. Manganese Nickel Copper Chromium Molybdenum Wavelength/nm . . . . 279.5 341.5 324.8 367.9 313.3 Slit width/mm . . .. 0.10 0-10 0.20 0.20 0.10 Observation heightlcm . . 0.8 1.0 1.0 1.0 0.8 Acetylene flow-ra.te/l min-l . . 1-4 1.0 1.0 1.2 1.8 Lamp current/mA . . .. 6 10 4 8 5 Determination of Molybdenum Transfer a suitable aliquot of the stock sample solution (see Table I) into a 100-ml calibrated flask, add a suitable amount of the stock iron solution (see Table I) and 10 ml of 10 per cent.August, 1975 561 m/V ammonium chloride solution, then dilute to the mark with de-ionised water.Continue as described under Determination of Manganese, Nickel and Copper. STEELS BY AAS USING THE AIR - ACETYLENE FLAME Determination of Chromium (Note 3) Transfer a suitable aliquot of the stock sample solution (see Table I) into a 100-ml calibrated flask. Add a suitable amount of the stock iron solution (see Table I) and 5 ml of 5 per cent. m/V quinolin-8-01 solution, then dilute to the mark with de-ionised water. Continue as described under Determination of Manganese, Nickel and Copper. Absorption Measurement In order to obtain the highest accuracy the value for the appropriate standard solution (A-F) should be determined before and after that for each sample and the average absorption value for the standard used for calculation of the result.Some degree of scale expansion can be used when the signal is sufficiently stable. Alternatively, the concentration read-out scale of the SP1900 instrument can be calibrated for a given element, using the high and low standard solutions as described in the instrument instructions, and the results read directly from the scale. KOTES- 1. The stock iron and primary standard metal solutions can be stored in plastic bottles for future use. Similarly solutions A - F and the blank solutions prepared for the determination of each element are stable if stored in plastic bottles in a cool place. 2. In the preparation of the molybdenum solution the pure metal is dissolved in 1OOml of hydro- chloric acid (sp.gr. 1-18) in order to prevent hydrolysis on dilution. The copper can be dissolved in a small volume of nitric acid prior to adding the hydrochloric acid. 3. It is preferable to make this determination after the instrument has been in use for at least 20 min so as to allow the burner t o become thermally stable. This practice reduces any tendency for absorbance readings to drift and is simply arranged by determining the other elements before chromium. TABLE I11 RESULTS FOR ANALYSIS OF BRITISH CHEMICAL STANDARDS STEELS Element blanganese Chromium Molybdenum Pu’ickel Copper Steel type Low alloy Low alloy Ferritic stainless Austenitic stainless Austenitic stainless Mild Low alloy Low alloy Ferritic stainless Austenitic stainless Austenitic stainless Mild Low alloy Low alloy Ferritic stainless Austenitic stainless Mild Low alloy Low alloy Ferritic stainless Austenitic stainless Austenitic stainless Mild Low alloy Low alloy Austenitic stainless BCS No.402 409 342 33 1 336 274 402 409 342 33 1 336 274 402 409 342 336 274 402 409 342 331 336 274 402 409 336 Value found, per cent. 0.197, 0.197, 0.197, 0.196 0.494, 0.488, 0.485, 0.494 0.89, 0.89, 0.89, 0.89 0.76, 0.76, 0.76, 0.76 0.81, 0.81, 0.81, 0.82 0.191, 0.192, 0.185, 0.185 0.56, 0.57, 0.56, 0.55 1.26, 1.23, 1-23, 1-23 16.0, 16-2, 16.1, 16.0 15.3, 15.3, 15.3, 15-3 17.6, 17.6, 17-6, 17.6 0.080, 0.066. 0.074, 0.080 0.166, 0.159, 0.166, 0.153 0.78, 0.77, 0.78, 0.79 0.69, 0.68, 0.70, 0.70 2-41, 2.44, 2.47, 2.44 0.132, 0.132, 0-132, 0.132 0.71, 0.71, 0.71, 0.72 3.12, 3.12, 3.12, 3.12 2.15, 2.17, 2-18, 2.18 6.28, 6.31, 6.28, 6.22 9.44, 9.48, 9-44, 9.40 0.038, 0.038, 0.036, 0.036 0.23, 0.23, 0.23, 0.23 0.23, 0.23, 0.23, 0.23 0.118, 0.118, 0-117, 0.118 Certificate value, per cent.0.19 0.48 0.91 0.78 0-81 0-185 0-55 1-22 16.15 15-2 17-6 0.070 0.16 0.77 0.69 2.43 0.125 0.71 3.14 2-16 6-26 9.48 0.040 0.23 0.23 0.11562 NALL, BRUMHEAD AND WHITHAM Results British Chemical Standards steels were used to evaluate the analytical scheme and the results obtained, compared with the certified figures, are given in Table 111. An integration time of 4 s was used except for high nickel, chromium and molybdenum contents, for which the time was increased to 20 s. A further evaluation of the reproducibility of the instrument was made with standard solutions of each element.The results were expressed as a coefficient of variation and are based on ten replicate measurements at the concentrations shown in parentheses below. Manganese Chromium Molybdenum Nickel Copper (2-5 p.p.m.) (10 p.p.m.) (20 p.p.m.) (20 p.p.m.) (2.6 p.p.m.) Coefficient of variation, per cent. 0.32 0.36 0.7 1 0.34 0.62 Discussion Composite schemes of atomic-absorption analysis have been described for nickel base6 and cobalt base alloys' based on the use of both air - acetylene and nitrous oxide - acetylene flames. The present work enables the analysis of complex steels to be made with the use of the air - acetylene flame only (except for the refractory oxide elements). The effect of iron on the absorption of chromium and molybdenum has been studied and the findings of previous workers c ~ n f i r m e d . ~ ~ ~ , ~ It has been shown, however, that in the presence of releasing agents for chromium4 and molybdenum5 the amount of iron can be reduced considerably, thus enabling the present scheme to be devised. Satisfactory results are given for a wide range of British Chemical Standards steels and this range can be extended, if desired, to include any element that can be determined in the air - acetylene flame. References 1 . 2. 3. 4. 6. 6. 7. Thomerson, D. R., and Price, W. J., Analyst, 1971, 96, 825. Harrison, T. S . , Foster, W. W., and Cobb, W. D., Metallurgia Metal Form., 1973, 40, 361. Gregorczyk, S . , Gralewska, K., and Flabon, J., Chemia Analit., 1973, 18, 1065. Ottaway, J. M., and Pradhan, N. K., Talanta, 1973, 20, 927. Mostyn, R. A., and Cunningham, A. E., Analyt. Chem., 1966, 38, 121. Welcher, G. G., and Kriege, H., Atom. Absorption Newsl., 1969, 8, 97. Welcher, G. G., and Kriege, H., Atom. Absorption Newsl., 1970, 9, 61. Received December l l t h , 1974 Accepted April 4th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000555

出版商:RSC    

年代:1975

数据来源: RSC

摘要:

Analyst, August, 1975, Vol. 100, pp. 563-566 563 A Rapid and Sensitive Spectrophotometric Procedure for the Determination of Diphenhydramine and Related Ethers B. Caddy, F. Fish and J. Tranter* Division of Pharmacogvzosy and Forensic Science, School of Pharmaceutical Sciences, University of Strathclyde, Glasgow, G1 1XW An oxidative procedure for the determination of diphenhydramine and related ethers is described. Its high sensitivity is a consequence of the oxidation of the drug to new products that exhibit high absorption in the ultraviolet region of the spectrum. The developed method is discussed with respect t o precision, specificity, sensitivity and applicability to drug deter- mination in biological samples. The eight compounds examined in this work (Table I) all possess antihistaminic properties and are prescribed for the treatment of various allergic conditions.Orphenadrine, benz- tropine and chlorphenoxamine also have parasympatholytic activity and are mainly used as parasympatholytics. While an oxidation procedure for this group of drugs has previously been reported,l this paper reports a simpler and quicker oxidative - spectrophotometric method which, although less sensitive than the gas-chromatographic procedure of Vessman, Hartvig and Stromberg2 has been found to possess adequate sensitivity for the detection of diphenhydramine (and hence presumably the other ethers) in human urine following the ingestion of less than therapeutic doses. Experimental The reagents and apparatus, with the following exceptions, have been adequately described in an earlier paper.1 Apparatus A Linson Instrument tilt-shaker, modified to accommodate six glass-stoppered tubes of approximately 40-ml capacity, and a roller extractor made in the laboratory following the design described by Moss,3 were used.Preparation of Standard Calibration Graphs Dissolve a suitable amount of each drug (10 to 100 pg, as the salt specified in Table 11) in approximately 4 ml of water contained in a 40-ml glass-stoppered test-tube. Add 5 ml of potassium dichromate solution (4 per cent.), 15 ml of sulphuric acid (66 per cent., prepared by mixing 6 volumes of concentrated sulphuric acid with 4 volumes of water) and 5 ml of spectroscopic grade hexane. Tilt-shake the tube for 30 min at room temperature, then remove an aliquot of the hexane layer and record its ultraviolet spectrum over the range 220400 nm in a cell of 1-cm path length with hexane as reference.Drug Administration A subject was administered one 50-mg capsule of diphenhydramine hydrochloride twice daily for 2 d. The total urine voided over a period of 5 d, inclusive of those 2 d on which the capsules were administered, was collected. The urine collected over certain intervals (Table 111) was bulked and aliquots from each bulked sample were assayed. Extraction from Body Fluids Place 10 ml of urine, 5 ml of 1 N sodium hydroxide solution and 100 ml of diethyl ether in a 500-ml reagent bottle. Roll the bottle for 20 min at approximately 40 rev min-l, separate the ether layer, filter it through a Whatman No. 1 filter-paper, measure the volume and transfer it into a second reagent bottle together with 5 ml of 1 N sulphuric acid.Roll this * Present address : Forensic Science Laboratory, 17th Floor, May House, Police Headquarters, Arsenal Street, Hong Kong.564 Analyst, Vol. 100 second bottle for 20 min, separate the aqueous phase and evaporate 4 ml of it under reduced pressure at 50-60 "C for 3 min in order to remove the last traces of organic solvent. Oxidise the resulting acidic solution as described for the solutions of drug salts used in preparing calibration graphs. Extractions from blood can be carried out in a similar manner, except that the blood is made alkaline by the addition of 2 or 3 ml of concentrated ammonia solution. CADDY et al. : RAPID AND SENSITIVE SPECTROPHOTOMETRIC Results and Discussion The oxidation products were identified by gas - liquid chromatography in conjunction with their ultraviolet spectral characteristics1 (Table I). TABLE I GAS-CHROMATOGRAPHIC DATA AND Amax.VALUES IN HEXANE FOR THE OXIDATION PRODUCTS OF DIPHENHYDRAMINE AND RELATED ETHERS Drug Benztropine . . .. . . Bromodiphenhydramine . . .. Chlorphenoxamine .. .. Deptropine .. .. .. Diphenhydramine .. .. Diphenylpyraline . . .. .. Orphenadrine . . .. .. Embramine .. .. .. Oxidation Relative product retention time* XmB,./nm Benzophenone 1.00 (a) 247 4-Chlorobenzophenone 1.98 (a) 254 DiHDBCHt 0.97 (c) 264 Benzophenone 1.00 (a) 247 Benzophenone 1.00 (a) 247 4-Bromobenzophenone 3.08 (a) 257 2-Methylbenzophenone 1-10 (b) 247 4-Bromobenzophenone 3-18 (a) 257 *(a) 1 per cent.OV-25 a t 160 "C with retention times relative to benzophenone: under these conditions benzophenone has a retention time relative to the solvent front of 2-4 min; ( b ) 10 per cent. Apiezon L, temperature programmed from 150 to 200 "C at 1 "C min-1, retention time relative to benzophenone : under these conditions benzophenone has a retention time relative to the solvent front of 23.2 min; (c) 1 per cent. OV-25 at 190 "C with retention time relative to DiHDBCH: under these conditions DiHDBCH has a retention time relative to the solvent front of 3.3 min. t 10,l l-Dihy~ro-5H-dibenzo[a,d]cyclohepten-5-one. Fifteen replicate assays were carried out on a 100-pg sample of diphenhydramine hydro- chloride in order to assess the reproducibility.Under the experimental conditions employed no decomposition was apparent and the precision was found to be of a high order with a coefficient of variation of 1.7 per cent. The specificity of the procedure for the determination of these drugs in biological samples is good within the limits described below. The extraction procedure eliminates interference from highly ultraviolet absorbing acidic and neutral drugs. As the oxidation is carried out under acidic conditions, highly absorbing basic drugs that are not affected by the oxidation do not partition into the hexane layer and, therefore, do not interfere. Interference from compounds that give rise to highly absorbing basic oxidation products (e.g., basic benzo- phenones from drugs containing the benzodiazepine structure) does not occur as they also remain in the acidic phase.Acidic oxidation products from basic drugs (e.g., benzoic acid from ephedrine), which partition into the hexane phase, can be eliminated by washing the latter with dilute sodium hydroxide solution. The only unavoidable interference that may arise is from neutral oxidation products, with high I?:& values, formed from basic drugs. In this respect, all the drugs considered in this work must be considered as mutually interfering. Even though some of them will give rise to different products, the difference between their A,,,, values is not sufficient to permit the determination of any two in admixture at the same time. Other drugs that may cause similar interference include amitriptyline and some of its derivatives.Obviously, it is necessary to establish qualitatively that only one of these ethers, and no amitriptyline, is present before the oxidative assay can be carried out. Other basic drugs that possess the diphenylmethylene group (Ph,C<) and may be oxidised to benzophenones, but which do not interfere under the conditions used, include cyclizine and related amines and methadone and similar compounds. Because of the high E:& values for benzophenones (about 1000) the sensitivity is high and certainly higher than that obtained with direct ultraviolet spectrophotometry. TheAugust, 1975 DETERMINATION OF DIPHENHYDRAMINE AND RELATED ETHERS 565 minimum concentration in urine, which is easily measured, is approximately 1.5 pg ml-1 in a 10-ml sample volume.Absorbance values obtained for this concentration of different drugs range from about 0.07 for deptropine citrate to 0-14 for diphenhydramine hydrochloride, assuming that all the drugs have recoveries similar to that of diphenhydramine hydrochloride. By using the procedure on the scale described, background absorbance is low (0.02 absorbance unit or less), but if the ratio of urine to hexane is raised so as to increase sensitivity, appreciable background absorbance can arise. For example, with a 50-ml volume of urine sample and 2 ml of hexane and taking absorbance readings in cells of 2-cm path length, the background may be as high as 0.09 absorbance unit. However, it is anticipated that such an increase in sensitivity will not be required. When using the procedure on the scale proposed it is better to use a “blank urine” in order to determine the background rather than subtracting the absorbance at 300nm of a test urine as recommended by Walla~e,~ as “blank urine” determinations usually exhibited a greater absorbance at 247 nm than at 300 nm.Standard calibration graphs for all drugs studied in the range 0-20 pg ml-I (equivalent drug concentration in the final hexane solution) were linear with little scatter (Table 11). Assays for diphenhydramine hydrochloride added to blood and urine in the same concen- tration range also produced good linear absorbance - concentration relationships (Table 11). TABLE I1 ABSORBANCE IN HEXANE OF VARIOUS CONCENTRATIONS OF DIPHENHYDRAMINE AND RELATED ETHERS FOLLOWING OXIDATION Final drug* concentration in hexanelpg ml-l 20 15 10 5 2 Absorbance? a t Amax. Drug Amax.lnni I A 3 I 7 d Benztropine .. .. . . .. 247 0.84 0-63 0.42 0.22 0.09 Bromodiphenhydramine . . . . .. 257 1-11 0.83 0.55 0.27 0.11 Chlorphenoxamine . . .. . . .. 254 1-15 0.88 0.58 0.29 0.12 Deptropine . . .. .. .. 264 0.60 0.45 0.31 0.14 0.07 Diphenhydramine . . .. . . . . 247 1-24 0.94 0.63 0-31 0.14 Diphenylpyraline . . . . . . ,. 247 1.20 0.89 0.61 0.30 0.13 Embramine. . .. .. .. .. 257 0.86 0.64 0.43 0.22 0.09 Orphenadrine . . .. .. .. 247 0.91 0.67 0.45 0.22 0.09 Diphenhydramine extracted from blood: 247 1.05 0.79 0.55 0.27 0.11 Diphenhydramine extracted from urine: 247 1-12 0.80 0.56 0.27 0.11 * As the salt : benztropine mesylate, deptropine citrate, all others as hydrochloride. t +ported absorbance values are the mean of a t least two determinations. : Corrected for ether and acid losses.The results obtained for the assay of diphenhydramine in the urine of a subject given oral doses of this compound are given in Table 111. Diphenhydramine was readily determined during the 2 d of its administration and for up to 30 h after the final capsule had been taken. The level determined between 30 and 54 h after final administration was low (06pugml-l) and must be considered to be near to the minimum concentration detectable by the described procedure. No diphenhydramine was detected after 54 h. TABLE I11 CONCENTRATION IN URINE OF DIPHENHYDRAMINE HYDROCHLORIDE FOLLOWING ITS ORAL ADMINISTRATION The first capsule taken a t 0 h and the fourth (the last capsule) taken at 31 h.All determinations were carried out on 10-ml sample volumes. Sample Time during which Volume Drug code urine collected/h collected/ml concentrationlpg ml-l D1 0-14 980 1.9 D2 14-37 1630 2.4 D3 37-61 1000 3.0 D4 61-85 1290 0.5 D5 85-103 1220 0.0566 CADDY, FISH AND TRANTER That diphenhydramine was determined by this procedure for up to 54 h after the final dose indicates that the method possesses adequate sensitivity for most purposes, especially when it is considered that the dosage was half that of the recommended drug regimen. From a normal therapeutic dose of diphenhydramine the peak level in blood is of the order of 1 pg ml-1 and thus from a 10-ml volume of blood sample the oxidation yield would be at the lower end (2 pg ml-l) of the calibration graph.Although the method as applied to blood may not be sufficiently sensitive for routine use following therapeutic dosage, it should be adequate for detecting overdoses. The drug values recorded are a measure of diphenhydramine and its demethylated metabo- lites. These latter may be expected to occur at low levels as they are suspected by Drach and Howell5 to be intermediates in the formation of the major metabolite, diphenylmethoxy- acetic acid, which exists in both free and bound forms. Diphenylmethoxyacetic acid is not determined by the proposed method because it is not extracted from urine under the conditions used. Similarly, another major metabolite, the N-oxide, reported by Drach, Howell, Borandy and Glazko,6 is not determined. Orphenadrine undergoes similar metabolic transformations to diphenhydramine’ and hence it is possible that some of the other related ethers, particularly those in which the amine on the ether function is demethylated, also undergo similar degradation. In conclusion, it is thought that the reported procedure is adequate for the routine deter- mination of diphenhydramine and related ethers in urine following their therapeutic adminis- tration and also in blood in cases of drug overdosage. References 1. 2. 3. 4. 8. 6. 7. Caddy, B., Fish, F., and Tranter, J., Analyst, 1974, 99, 555. Vessman, J., Hartvig, P., and Stromberg, S., Ada Pharm. Suec., 1970, 7 , 373, Moss, M. S., i n “Identification of Drugs and Poisons,” Pharmaceutical Society of Great Britain Wallace, J. E., Analyt. Chew., 1968, 40, 978. Drach, J. C., and Howell, J. P., Biochem. Pharmac., 1968, 17, 212s. Drach, J. C., Howell, J. P., Borandy, P. E., and Glazko, A. J., Roc. SOC. Exp. Biol. Med., 1970, Ellison, T., Snyder, A., Bolger, J., and Okum, R., J . Pharmac. Exp. Ther., 1971, 176, 248. Symposium, 1965, pp. 27-38. 135, 849. Received Jartzcary 6th, 1976 Accepted February 27th, 1975

ISSN:0003-2654    

DOI:10.1039/AN9750000563

出版商:RSC    

年代:1975

数据来源: RSC