RESEARCH AND DEVELOPMENT TOPICS 23 Microprocessor-controlled Laser Remote Sensing System A. R. Morrisson and B. L. Sharp Allocaiday Iszstitute f o r Soil Research, Crazgiebucltler, A berdeua, d B9 2QJ There has been growing concern during the last decade about the protection and conserva- tion of the environment. Modern industrial processes and urbanisation produce vast amounts of waste products, many of which are toxic, adversely affecting animal life either directly or through the food cycle.The most widespread pollutants are those which are emitted directly into the atmosphere. Sulphur dioxide, produced mainly by the combustion of fossil fuels, is a prime example and its effect on soil acidity and plant growth rate are subjects of current research and discussion.There has been much controversy in recent years concerning the emission of sulphur dioxide by the industrialised countries of Western Europe and the effect of this pollution on tree growth in the forests of Scandinavia. Yearly all of the analytical techniques currently employed use point sampling methods. A sampling unit is located at some point and the sample collected over a pzriod by filling a plastic bag with the gas or pumping the gas through a suitable absorber. A variety of analytical methods can then be employed in the laboratory to measure the concentration.The disadvantages of these methods are that they yield average values of the concentration with poor spatial and temporal resolution and have little or 110 ability to track the path of the pollutant.In addition, no information can be obtained about pollutant concentrations at inaccessible points. Laser Remote Sensing Developments in laser technology have yielded a variety of coherent optical sources having outputs whose wavelengths range from the ultraviolet to microwave regions of the spectrum. Continuously tunable lasers are now available covering the ranges 280-340 nm, 400-690 nm and various selected ranges in the infrared region.The high power density and collimation of laser beams enables them to be transmitted over considerable distances and still yield measurable power levels at a receiver point. These factors have led to their use in rernote sensing of the atmosphere, whereby a combination of laser transmitter and adjacent optical receiver are used to detect and determine the concentration of molecules at a location remote from the observation point.The analogy with RADAR techniques has resulted in the term LIDAR (light detection and ranging) being applied to such methods. -1 \.ariety of spectroscopic phenomena can be employed in the detection scheme, for example Raman and resonance Raman scattering, electronic absorption and fluorescence, and rotational and vibrational absorption.Because of the high quenching rate of excited states in air, absorption is the most sensitive technique for the in sit% measurement of sulphur dioxide. Two types of measurement are possible, of which the simpler is long path absorption. The laser beam is tuned to an absorbing wavelength and is attenuated over an absorbing path of several kilometres, according to Beer’s law.Although this technique has the advantages of simplicity and low24 RESEARCH AND DEVELOPMENT TOPICS Proc. Analyt. Din. Chenz. SOC. source power requirement, it has the serious limitation that it is necessary to place a receiver a t the far end of the absorption path or a retro-reflector to return radiation to a local receiver.This is not practical when measurements a t finite elevation or over difficult terrain are required; also, no range information is obtained as the laser light has to pass through the entire absorbing path before reaching the detector. A further difficulty exists in stabilising the relative positions of the transmitter and receiver as small angular shifts in either (e.g., 1 mrad) produce linear displacements of several metres when subtended by a range of several kilometres.An alternative method is that of differential LIDAR, in which the atmosphere is used as a distributed reflector (Fig. 1). Two sequential pulses are transmitted, one at the peak of the absorption band and one at an adjacent non-absorbing wavelength (for sulphur dioxide, 300.1 and 299.5 nm, respectively).The backscattered intensities are then ratioed to give a measure of the sample absorbance and, from a knowledge of the absorption cross-section and the path length, the concentration can be calculated. The range from which back- scattered signals can be detected is limited by the cross-section for Rayleigh and Mie scattering from atmospheric gas and particulates, and the transfer function of the receiver - detector system.The selected sampling range is defined by the overlap of the laser beam with the telescope field of view (Fig. 1) and the timing of the transmitting and receix-ing cycles. The detection limit is determined by the ability of the detector to discriminate small differences in the intensities of the two beams (about 1% absorption is normally detectable).The laser pulse duration is 1 ps and therefore the range resolution (given by CT/2 where T is the pulse length1) is 150 m. Fig. 1. The differential absorption LID_IK technique. Instrumental System The source is a modified Electro-Photonics, Model 23, flash-lamp pumped tunable dye laser using Rhodamine 6G dye. The light from two high-pressure xenon flash lamps is focused on to the dye cell and the laser cavity is formed by two 100yo reflecting mirrors.Tuning to 600 nrn is accomplished by an intra-cavity voltage-controlled interferometric filter. The light is polarised, intra-cavity, by a Glan - I’homson polariser and focused on to a frequency-doubling crystal, which halves the wavelength of the output to the required ultraviolet value.The phase-matching angle for the crystal is selected by rotation of a stepper motor. The output mirror has a low reflectivity at 300nm, which allows the frequency-doubled light to pass out of the cavity and through a beam expander to yield aJanizaiy, 1.979 RESEARCH A%XD DEVELOPhlENT TOPICS 25 transmitted beam with a diameter of about 25mm. Beam steering is accomplished with two adjustable mirrors.The laser is operated a t about 1 Hz and produces a 0.1-mJ pulse a t 300 nm with a beam divergence of 2 mrad. The laser power is monitored by focusing the small fraction of light that passes through one of the beam-steering mirrors on to a photodiode. There is a shot to shot variation in the output power of about 10% and this measurement allows the collected data to be normalised.The laser is mounted beside a 0.5m diameter Cassegrain telescope on a mechanically rigid table. The transmitted ultraviolet laser light is fired in the desired direction and the backscattered light collected by the telescope. The collected light is first filtered to reduce the background radiation by the use of a 5 nm band pass, 12y4 transmission interference filter with centre wavelength at 299.8 nm.It then passes on to an EM1 9789QB photo- multiplier tube whose signal is pre-amplified and digitised using a Tektronix R7912 transient digitiser, which records the temporal variation in backscattered intensity. The operating functions and data handling in the system are controlled by an Intel SBC 80-10 microcomputer.The on-board computer memory is augmented with an 8 kbyte random-access memory expansion built in the laboratory. A 250 kbyte floppy disk drive is linked to the computer for bulk data storage. The programmed experimental sequence consists of tuning the laser, firing the laser pulse, acquiring the digitised backscattered waveform from the transient digitiser and measuring the transmitted pulse energy.This sequence is then repeated at the two wavelengths alternately and the normalised data from several hundred shots are averaged. Signal averaging is necessary in order to enhance the signal to noise ratio and hence to improve the measurement precision. The LIDAR system should not pose an eye hazard to the general public, both because of the transmission path used and as the beam expanded pulse yields an energy flux below the maximum permissible exposure for direct ocular viewing.2 In the event of an emergency the operator can activate the computer interrupt system, which suspends the experimental sequence.Thereafter, the experiment can be continued or terminated by entry of a character on the keyboard. The prototype differential LIDAR system has been built and preliminary testing is under- way.Theoretical studies3 indicate that a wide range of concentrations of sulphur dioxide down to 10 p.p.b. (parts per lo9) can be measured at ranges of several kilometres. This compares with a sensitivity of about 3 p.p.b. for the best point sampling instruments. Future work will include extension of the system for the measurement of nitrogen dioxide and ozone.References 1. 2. 3. Hinkley, E. D., Edztor, “Topics in -$pplied Physics,” Yolume 14, Springer-Verlag, Berlin, 1976, I3ritish Standards Institution, “Draft Guide to the Protection of Personnel against Hazards from Xdrain, R. S., Brassington, D. J., Sutton, S., Tozer, B. A,, and \‘are)-, R. H., Central Electricity pp. 76-78, Laser Radiation,” BS 4803, Document 76/31221 DC.Kesearch Laboratories Laboratory Note Xo. RU/L/N 120/77‘, 197’7. Problems in the Analysis of Chromite Ores: Precision Spectrophotometric Determination of Total Iron M. E. M. Abdel Aziz and D. Thorburn Burns Depni,tiizriit of Chenzistvj~, The Queen’s Uizivrvsity of Belfast, Belfast, BTS L 4 G The long established method for the determination of total iron in chromite ores is by dichromate titration of reduced solutions of the ores1 Recently, the British Standards Institution has adopted a spectrophotometric method for iron2 as a major constituent in chrome-bearing materials using 1,lO-phenanthroline, a reagent that is normally used for trace-level determinations.Following the success of earlier work on the determination of chromium in chromite ores,3 precision spectrophotometry has now been applied to the determination of total iron based on the absorbance of the iron(II1) chloro complex at 342 nm.Problems were encountered in the development of the method a t the ore decomposition and filtration stages and in selecting the chromogenic species for final spectrophotometric26 RESEARCH AND DEVELOPMENT TOPICS Proc.Analyt. Div. Chem. SOC. measurement. It was not possible to use an iron(I1) species as was the original intent. The optimised conditions are as follows. A 0.45-g amount of finely powdered chromite ore, intimately mixed with 5 g of sodium peroxide, is sintered for 3 h at 510 10 “C in a zirconium crucible. The sintered mass is leached with 50 ml of distilled water and filtered by gentle suction through a Whatman GF/C glass fibre filter-paper.The residue is washed with hot distilled water and then dissolved in 250 ml of concentrated hydrochloric acid and diluted to 500 ml with distilled water in a calibrated flask. The final concentration of hydrochloric acid should be about 6.0 M; previous workers4 have shown that this is the least critical concentration for variation of absorbance with acidity.The solution is mixed, thermostated at 25.0 “C, made up to volume and its absorbance measured at 342 nm using a thermostated cell holder in a Pye Unicam SP 3000 spectrophotometer. The iron contents are calculated from the pre-determined specific absorptivity of the iron( 111) chloro com- plex; Beer’s law holds up to 1.5 absorbance units.Sample masses are chosen so as to give absorbaaces in the range 0.700-1.200 using 0.2-cm cells. The results except for sample 49gG show acceptable accuracy and precision. Sample 49gG was analysed by the British Standards Institution method2; the figure obtained, 14.43 & 0.04% of FeO, is in agreement with the chloro complex determination. The cause of the apparent discrepancy is being investigated. Results are shown in Table I for the analysis of standard samples.TABLE I DETERMINATION OF TOTAL IRON -4s FeO IN STANDARD REFERENCE CHROME ORES Sample Certificate value, ()& Result, BCS 308 . . . . .. .. 15.3 15.38 + 0.07 Student’s sample 49f . . .. 15.3 15.40 & 0.05 Student’s sample 49gG . . .. 15.2 14.63 & 0.02 We thank the Pye Unicam Company for the gift of the SP 3000 spectrophotometer and the Industrial Research and Consultancy Institute, Khartoum, Sudan, for leave of absence and financial assistance for one of us (M.E.M.A.) References I .2. 3. 4. Brearlcy, H., and Ibbotson, F., “The XnaIysis of Steei-Works Materials,” Longmans, London, 1902. HS 1907 : P a r t 2C : 1974. 4bdel Aziz, M.E. M., and Thorburn Burns, D., unpublished results. Desesa, M. A,, and Rogers, L. R., Awzlytica Chim. Acta, 1952, 6, 534. Development of Fluorescence lmmunoassay Methods of Drug Analysis G. Handley and J. N. Miller Department of Clzernistry, University of Technology, Loughborough, Leicestershire, LE11 3T U and J. W. Bridges Uepavtmeizt of Biochemistry, Univevsity of S u ~ r e ~ y , Guildford, Surrey Fluorescence immunoassay methods for drug analysis are being developed in an attempt to overcome some of the disadvantages of radioimmunoassay (RIA).The intention is to produce assays that are rapid, homogeneous (i.e., do not require a separation step) and do not involve the use of hazardous materials or expensive equipment. One such method for the analysis of thyroxine (T4) involves the phenomenon of fluorescence enhancement.It has been found that when T, is labelled with a fluorescent moiety, such as fluorescamine (4-phenylspiro [f uran-2 (3) - 1 ’-pht halan] -3,3-dione) or MD PF [2-me thoxy-2,4-diphenyl-3 (2H) - furanone], the fluorescence of the complex is greatly increased on binding with anti-T, antibodies. This report describes how this effect of fluorescence enhancement can be utilised in a fluorescence immunoassay.January, 19 79 RESEARCH AND DEVELOPMENT TOPICS Thyroxine (T4) Fluorescent derivative o f fluorescarnine R \ Fluorescent derivative of MDPF 27 Experimental The first involved the rapid addition of 0.5 ml of a 0.03:/, solution of the label in acetonel to 2 ml of a 10 pg ml-1 solution of T, in phosphate buffer (0.1 M, pH 8.0) that was being mixed on a vortex mixer.The second technique involved the incorporation of the label into a cycloheptaamylose complex, after the method of Nakaya et aZ.2 The resultant stable solid was added in excess to a solution of T, in phosphate buffer and incubated for 30 min at 37 "C. Residual cycloheptaamylose was removed by centrifugation.No further purification was deemed necessary as neither the labels nor their hydrolysis products are fluorescent. The latter technique was the preferred method of labelling and fluorescamine the preferred label as it is commercially available and MDPk. is not. Both fluorescarnine and MDPF were obtained from Roche, anti-T, serum from Calbiochem and cycloheptaamylose from Sigma.A11 dilutions were carried out in barbital buffer (0.075 ni, pH 8.6) and a Raird Atomic Fluoricord spectrophotometer with excitation and emission wavelengths set at 390 and 490 nm, respectively, was used for all fluorescence measurements. An antibody dilution graph was constructed by the addition of a series of 1-ml antiserum dilutions to 1 ml of fluorescarnine-labelled T, solution (10 ng ml-l).After incubation at room temperature for 15 min the fluorescence of each mixture was measured. In order to correct for the background contribution of the antiserum, the fluorescence of the anti- serum dilutions was also measured separately in the absence of T,. A standard graph was constructed using 0.5-ml aliquots of known T, dilution. These were added to 0.5 ml of labelled T, (20 ng ml-l), followed by 0.5 ml of anti-T, serum.The antiserum was used in an initial dilution of 1 : 200 (1 : 600 final dilution) on the basis of information gained from the antibody dilution graph. After incubation at room temperature for 15 min the fluorescence of the mixtures was measured and corrected for the fluorescence of the antiserum. Therefore, a series of T, solutions of known concentration were prepared in pooled T,/T,-depleted serum.Aliquots (0.25 ml) of these solutions were diluted to 0.5 ml with barbital buffer and 0.5 ml of labelled T, solution and 0.5 ml of anti-T, serum added as before, in order to construct a standard graph. The high background signal produced by serum could be caused by either scattering of light or the fluorescence of species present in the serum.In order to investigate the possi- bility that scattered light was responsible a horizontal polarising film was inserted into the excitation beam and another standard graph constructed. A standard graph of T, in serum was also constructed using MDPF as the fluorescent label, using the same procedure as that used for the fluorescamine label.Thyroxine was labelled by one of two methods. I t is, however, normal to wish to measure T, concentrations in serum.28 RESEARCH ASD DEVELOPMEST TOPICS P V O C . A%L~J$. D i V . Chem. SOC. Results From the antibody dilution graph (Fig. 1) it can be seen that the fluorescence of labelled T, is enhanced on binding with antibod!.. The data obtained from this graph allow the antiserum dilution for the standard graph to be determined.The standard graph of T, in barbital buffer (Fig. a), which summarises the pooled results of several experiments, shows that unlabelled T, competes for antibody binding sites, causing a decrease in observed fluorescence with increasing concentration of added T,. co +J ._ 2. P -fi 40 a- m 35 z P +-’ 3 - ; 1 5 # 30 “I 10 1 : l O O 1 1 000 1:10000 I nit i a I antiserum d i I II tion \ +\ Fig. 1.Antibody dilution graph. Fig. 2. Standard graph of fluorescence enhance- The results of the standard graph in serum (Fig. 3 line A) show that the effect is still present, but do not indicate the high background signal obtained, which is several times greater than the fluorescence due to fluorescamine. ment of T, in buffer.I I 100 1000 Initial thyroxine concentrationhg ml-’ Fig. 3. Standard graphs of fluorescence enhancement assay of T, in serum, (A) using fluorescamine-labelled T,, (B) using fluo- rescamine-labelled T, and a horizontally polari- sed excitation beam and (C) using MDPF- labelled T,. The presence of a horizontal polariser in the excitation beam (Fig. 3 line B) produced no When MDPF was used as the fluorescent label in the construction of a standard graph Therefore, improvement in the standard graph and reduced the background only slightly.(Fig. 3 line C) the results were comparable to those obtained using fluorescamine.Jizir iinvy, 1979 RESEA4RCH ASD DEVELOPMEKT TOPICS 29 the only advantage of using MDPF would be that its derivatives are slightly more stable than the derivatives of fluorescamine.Discussion Iodine is well known for its quenching of fluorescence by the “heavy atom e f f e ~ t . ” ~ Therefore, it is possible that the fluorescence of labelled T, is normally quenched by the iodine atoms of the iodothyronine moiety of T,. If so, the enhancement of fluorescence on binding with antibody might be explained in terms of inhibition of this quenching.This possibility is still being investigated. Horizontally polarised light has been shown to reduce the effect of scattered light,, improving the sensitivity of standard fluorescence graphs. However, as no improvement of the standard graph was obtained and the background was reduced only slightly it appears that the high background signal was due mainly to the fluorescence of species present in serum, The enhancement of fluorescence on binding of labelled T, by antibody and the reduction of the effect when unlabelled T, is presznt, results which compare favourably with those obtained by Smith,j using fluorescein isothiocyanate-labelled T,, indicate that it would be possible to assay serum T, by a fluorimetric immunoassay technique.This assay would be rapid as there is only a 15-min incubation period, homogeneous as no separation step is required and cheap as the assay is followed by conventional fluorimetry. Unfortunately, serum has a high intrinsic fluorescence that will vary from sample to sample. It would therefore be advantageous to use a label that fluoresces at a wavelength outside the fluorescent spectrum of serum.Such a compound is rhodamine, with an emission maximum at 595 nm; work is currently being carried out to determine if the quenching effect of the iodothyronine moiety of T, will apply to rhodamine. References 1. 2. 3. 4. 5. Bohlcn, P., Stein, S., Dariman, \Y., and Udenfriend, S., Arch. Biochem. Biophys., 1973, 155, 213. Xakaya, I<., Yabuta, M., Iinuma, F., Kinoshita, T., and Nakamura, Y., Biochem.Biophys. Res. Wehry, E. L., in Guilbault, G. G., Editor, “Practical Fluorcscence,” Marcel Dekker, New York, r i m , C. S., Miller, J . N., and Bridges, J . \V., -4nnlytica Chiwz. .-lcta, in the press. Smith, D. S., FERS Lett., 1977, 77, 2 5 . Commun., 1975, 67, 760. 1973, pp. 87-91. Determination of Atmospheric Pollutants by Gas Phase Auger Electron Spectrometry G.N. Killoran and J. F. Tyson Chemistry Department, University of Technology, Loughbovough, Leicestevshire, LE11 3T U The potential of Auger electron spectrometry (AES) as an analytical technique for gases and vapours has been under study a t Loughborough University of Technology for the past few years. These studies have concentrated on the molecular spectra and the qualitative aspects of AES.ly2 Applications of AES in general and with reference to gases have recently been reviewed.3 In this paper some of the more recent studies conducted at Loughborough are described.The good resolution and high counting rates available, together with unique molecular and sensitive elemental identification, make AES a potentially powerful analytical tool., The analytical system chosen to continue the study of the analytical possibilities of AES for gases was the determination of the common pollutant gases found in air.The gases studied in- cluded carbon monoxide, carbon dioxide, hydrogen sulphide, sulphur dioxide, nitrogen oxide, nitrogen dioxide and ammonia, as well as oxygen, nitrogen and argon. These gases were studied as pure gases or in simple mixtures.The Auger spectra of the individual gases were obtained under similar conditions, so that they could be more easily compared. *4s AES is an elemental identification method the30 RESEARCH AND DEVELOPMENT TOPICS Proc. AnaZyt. Diu. Chenz. SOC. compounds were studied in groups having a common element: for example, sulphur dioxide, hydrogen sulphide ; carbon monoxide, carbon dioxide ; nitrogen oxide, nitrogen dioxide, ammonia.In order to assess the quantitative analytical potential the sensitivity, limit of detection and analytical calibration graph were determined for each gas in a binary mixture. Analytical Basis of Auger Electron Spectrometry AES is related to both X-ray fluorescence spectrometry (XFS) and X-ray photoelectron spectrometry (XPS), as shown in Fig.1 , in that they all depend on the ionisation of an inner shell, X, of an element. The energy of the ejected inner-shell electron is the basis of SPS. The inner shell vacancy is filled by an electron falling from level Y; this results in either the creation of an X-ray photon or the release of an Auger electron from level Z.The measure- ment of the energy and intensity of these electrons forms the basis of Auger electron spectro- metry. A e- Auger electron i L, 1 2 p ” * I \ \ Excitation e I ec tron s, X-ray Fig. 1. The Auger proccss. The Auger electron ejected is designated an XYZ electron. For the elements studied the Auger spectra are the KLL spectra for the first row elements and LMM for the second row elements. The energy of the XYZ electron can be calculated from the following equation: E X Y ~ = Ex - (Er + Ez) where Ex,, is the calculated Auger energy.Ex is the energy of the level where the initial vacancy occurs and Ey and EZ are the binding energies of the levels from where the “down” and Auger electrons originate, respectively. Also, E , is the binding energy of an electron in the shell of an ion having a single vacancy in the Y level.The Auger energy depends chiefly on the energy of the initially ionised level, which is characteristic for a given element, so that elemental identification is possible by AES as well as by XPS and XFS. The Auger spectra are complex and difficult to interpret. However, they are also characteristic of a given free molecule, hence molecular identification is possible.Under a given set of spectrometer operating conditions the intensity of a selected Auger peak will be proportional to the partial pressure of that component in the gas mixture, and thus AES is capable of quantitative analysis. The spectrometer used was a Vacuum Generators AFM2 and it is described in reference 1 .The excitation source was an electron gun operated at a beam energy of 5 kV. The other parameters were set for maximum signal to background ratio for qualitative analysis and for maximum figure of merit ( 2/Ip - l / l b ) for quantitative analysis, where I, and I b are the peak and background intensity, respectively.~~n1lllnYy, 1979 RESEARCH AND DEVELOPMEKT TOPICS 31 Results and Discussion Qualitative Analysis The Auger spectrum of each pure gas obtained could be easily distinguished from other spectra. In an equal concentration mixture of sulphur dioxide, carbon tetrachloride, carbon dioxide, argon and nitrogen, each element was easilv identifiable in the spectruni (Fig.2 ) . 2 I I 1 I 1 uc) 200 300 1 Electron kinetic energy/eV 0 Fig. 2 . Auger spectrum of a gas mixture.Problems of inolecular identification arise when gases in the mixture contain the same ele- ments, such as carbon and oxygen. Changes in the chemical environment of an atom can cause Auger peak shifts, intensity changes and changes in shape. The spectra of the elements common to each set of compounds were therefore compared to determine the possibilities of determining each compound in a mixture.The carbon and oxygen spectra for carbon mono- oxide, carbon dioxide and a mixture of the two gases are shown in Figs. 3 and 4. It can be seen that the carbon Auger peaks of carbon monoxide at 251 and 154 eV enable it to be identi- fied, whereas the oxygen peaks of carbon dioxide a t 494 and 503 el’ makes its presence distinguishable. J!q co/co, 1 I 1 ! 1 2 50 260 Electron kinetic energylell Fig.3 . Carbon Auger spectra of CO, CO, and a CO - CO, mixture. I I I I I ! 490 500 Electron kinetic energyfev Fig. 4. Oxygen huger spectra of CO, CO, and a CO - CO, mixture.32 RESEARCH AND DEVELOPMENT TOPICS Proc. Analyt. Div. Chew. Soc. Thus, a compound without elemental spectral interference can be determined qualitativelj. in a mixture, whereas compounds with common elements are more difficult to determine simultaneously, unless of course, they contained other elements, for example a sulphur dioxide - hydrogen sulpliide mixture in which only sulphur dioxide contains oxygea.For mixtures such as carbon monoxide - carbon doxide, if the spectral profile was known for each element in each component it might be possible to determine small amounts of one in the other by spectral deconvolution. Quantitative Analysis The sensitivity, limit of detection and analytical calibration graph were obtained for each compound studied.A series of binary mixtures of each compound in either argon or nitrogen were prepared, with concentrations ranging from 0 to 20 mole percent of the analyte.The sensitivity (counts per second per mole yo) ranged from 65 (for carbon dioxide oxygen) to 3 800 (for argon). The limits of detection were calculated from the formula: Limit of detection = 3 Ib S T 21- where S is the sensitivity, Ib the background intensity and T the counting time (100 s). These limits ranged from 0.24 (for carbon dioxide carbon) to 0.012 (for argon) mole :;, or 2 400 to 120 v.p.m. (see Table I).Most analytical calibration graphs were linear over the range from 0 to 20 mole :$. TABLE I Eleineii t C C 0 0 S S -1r SESSITIVITIES AND LIMITS OF DETECTION Sensitivity1 count s-l Limit of detection, Compound mol ? L - l v.p.m. co 340 950 co 66 1700 H,S 1 7 0 0 360 SOL 920 700 Xr 3 800 120 co, 140 2 400 co, 65 1 900 In addition to the well known problem of the high secondary-electron background ( I b ) , these studies revealed an instability in the electron gun’s output (owing to interaction between certain gases and theheated filament) and therefore also in the Auger electronintensity.Detec- tion limits could be improved by increasing the sensitivity and/or reducing the background. The stability could be improved by keeping the electron gun more isolated from the sample.References 1 . 2 , 3. 4. Thompson, M., Hewitt, P. X., and Wooliscroft, D. S., Analyt. Chrm., 1976, 48, 1336. Thompson, M,, Hewitt, P. X., and Wooliscroft, D. S., .4na?yt. Chewa., 1978, 50, 690. Thompson, M., Talanta, 1977, 24, 399. Carlson, T. A,, “Photoelectron and Auger Spectroscopy,” Plenum Press, London, 1975. Selectivity Rating of Calcium Ion-selective Electrodes G.J. Moody, N. S. Nassory and J. D. R. Thomas Chemistyy Department, Universzty of Wales Instatute of Science and Technology, Cavda ff, CF1 3 S C’ Interference Equations proposal^^-^ for rating selectivities of calcium ion-selective electrodes include the widely used selectivity coefficient, k::;, a selectivity parameter, K,, which is derived from a consideration of interference potentials, and a more arbitrary parameter, Ki.The origin of hF:i5, recommended in IUPAC nomenclature recommendation^,^ lies in the following form of the Nicolsky e q u a t i ~ n ~ , ~ : E = constant -t- . . . .Janiiavy, 1979 RESEARCH AND DEVELOPMENT TOPICS 33 where E is the emf of the calcium ion-selective electrode coupled to a reference electrode, a,, and a, are the activities of calcium ions and interferent B ions, respectively, and z, is the valence of the B ions.By measuring the emf, El, of a solution containing only calcium ions and the emf, E,, o: a solution containing calcium ions at the same activity and interfering ions, it is possible to deduce kF;k from R, T and F have their normal significance.. by plotting eAf:?”/RT against . ~ This corresponds to the Srinivasan and Rechnitzs (a,) 2’2n (%a) method for high k::; values (Method IIC of references 9 and 10). Another such potential, E,, has been defined in terms of the difference between an Eidea, and an E,, representing the theoretical and experimental difference between the emf of a pure calcium chloride solution and that of a mixed calcium chloride - interferent chloride solution, respec- tively.2 The two solutions are of equal ionic strength.For the theoretical case, Eidea, assumes that there is no interference and also that the single-ion activity coefficient of Ca2+ depends only on the total ionic strength, but not on composition. The results for E , ob- tained in this way for calcium chloride solutions plus sodium chloride or lithium chloride at a total ionic strength of less than 0.6 have been shown2 to fit The AE term of equation (2) can be regarded as an interference p~teni-ial.~ E , = - ~ In 1 + K , - ] .. . . . . RT F [ (a,,)‘ ,411 arbitrary relationship for calculating selectivity arises from misgivings3 concerning the power term in equation (1) : E , - E - - In 1 + ~ i 51... . . . . . - RT 2F [ a,, (4) Here, the selectivity parameter, Ki, is based3 on a principle of equal affinity whereby an electrode “has an equal affinity for a primary ion and an interfering ion if the same measured potential is obtained for solutions of the primary ion and the interfering ion which have the same activities, regardless of the charges on the ions.” Critique of Interference Equations The definitive basis of equation (4) demands an intersection between the calibration line for primary calcium ions and that for interfering B ions, when Ki = 1 .M’hen the calcium ion calibration line is at a more positive emf than that for the interfering ions Ki is less than 1; a t more negative emfs Ki is greater than 1. For calibration graphs of calcium and interfering ions, respectively, that are of equal slope, the condition of equal affinity holds at all points for superposable calibrations and Ki = 1 over the complete calibration.This situation occurs for calcium - magnesium or “water hardness” electrodes based on calcium dialkylphosphate as sensor with decan-1-01 solvent as mediator.lY1l However, when calibrations of equal slope are not superposable the definitive requirement of equation (4) cannot be fulfilled; nevertheless, Ki can still be calculated and will be less than 1 for calcium ion calibrations that lie at more positive emfs than those for interferents and more than 1 for calibrations with more negative emfs.Misgivings concerning the power, 2/zu of aB in equation (1) have arisen3 because of k,t& values of greater than 1 for electrodes that demonstrate selectivity of calcium ions over sodium ions.’These occurrences need cause no concern as the relationship . . . . - (5) a,, = kE& (aNa)2 . . . . used for calculating k:::, by the mixed solution Method IIA of references 9 and 10 shows that when a,, is less than 1 electrodes can usefully be used for calcium ion determinations,34 RESEARCH AND DEVELOPMENT TOPICS Proc.Aaalyt. Div. Chcm. SOC. even for apparently high kgka values.12 For example, when RzJa = 46, determined in the presence of a 10-2 M sodium ion background, the calcium ion selective electrode still responds down to about a 4.6 x M concentration of calcium ions with little interference from the background sodium ions.I2 In fact, the useful calcium ion range will be even better because the true activity of sodium will be less than 1 0 - 2 ~ .Any misgivings concerning the significance of the power term in equation (1) need to be of a more fundamental origin, particularly because the calibration graphs obtained by use of an electrode for interfering ions, B, frequently do not match that expected from the charge on B and are frequently sub-Nernstian or hyper-Nernstian.Furthermore, the graphs may be curved, as with magnesium ions for the first PVC-matrix membrane calcium ion selective electrode based on calcium bisdidecylphosphate sensor with dioctyl phenylphosphonate solvent,l3 where the slope gradually increased from 14.2 to 23.6 mV decade-l. Because no allowance is made for such features, variations occur in the selectivity coefficient, k:::, and while their significance is relatively unimportant in the recommendati~n~~~~ for quoting the background level of interferent in relation to k::: values, there are problems when applying graphical-type relationships like equations (1)-(4) over extended ranges.15 Any fundamental physicochemical deductions made from such equations need to be cognisant of slope changes that must be related to rather more complicated features of membrane - solution interface ion-exchange and membrane diffusion processes than would be the case for constant calibra- tion slopes.Some Practical Considerations Because of the likelihood of non-Nemstian slopes or curvatures in calibration graphs of ion- selective electrod3s for interEering ions, linear graphs emanating from equations (2) and (3) are frequently just a coincidence rather than a regular feature.15 This phenomenon presents difficulties in conveying helpful data.In any case, the nature of kp,t is such that its values for B ions of different valencies are not comparable and, of course, it is of paramount import- ance for kyi data to be accompanied by information on how they were obtained.14 Despite the difficulties of theoretical interpretation, such background information on kE data en- hances their utility in assessing the practical scope of ion-selective electrodes.Improved Calcium Ion Sensors Selectivity coefficient data derived by the mixed solution method (Method IIA of references 9 and 10) and accompanied by the actual background level of interferent59l4 constitute a conven- ient way of conveying information on the selectivity of ion-selective electrodes.Thus, the lower k:$ values for a calcium ion selective electrode based on calcium bis[di(p-1,1,3,3- tetramethylbutylphenyl)phosphate] as sensor and dioctyl phenylphosphonate solvent as mediator confirm greater selectivity towards calcium ions than is obtained by use of an electrode based on Orion 92-20-02 calcium liquid ion exchanger (Table I).Tripentyl phos- phate is a good alternative to dioctyl phenylph~sphonatel~ (Table I). Table I establishes calcium bis[di(#-l,l,3,3-tetram~thylbutylphenyl)phosphate] as a superior calcium ion sensor, like its octyl isomer.16 It can be even better than is indicated in Table I, TABLE I SELECTIVITY DATA FOR PVC MATRIX MEMBRANE CALCIUM ION-SELECTIVE ELECTRODE+ k;:; for various interferents, B I 1 Membrane components Na* K* Mgt Srt Bat Mnt Cut Nit Znt Orion 92-20-02 liquid ion-exchanger 0.045 0.062 0.13 0.14 0.058 0.23 0.16 0.96 $ Calcium bis[di(p-l,1,3,3-tetramethyl- butylphenyl) phosphate] plus dioctyl phenyl- phosphonate 0.017 0.018 0.021 0.041 0.000 0.040 0.014 0.013 0.30 Calcium bis [di(p-l,1,3,3-tetramethylbutyl- phenyl)phosphate] plus tripentyl phosphate 0.021 0.022 0.062 0.091 0.043 0.22 0.086 0.082 0.48 * B ion level = 5 x M.t B ion level = 5 x M. 1 Calibration graph for Ca2+ in the presence of zinc ions did not, a t any stage, coincide with the normal Ca*+ calibration.January, 1979 RESEARCH AND DEVELOPMENT TOPICS 35 and Table I1 presents k;c data, obtained in various studie~~~J5@J9 and compared with those for a neutral carrier sensor,17 that were obtained for a lower level of background sodium ion interferent.These factors all point to a similar low interference from sodium ions for (~-1,1,3,3-tetramethylbutylphenyl)phosphate sensor for the various electrodes to that for the neutral carrier system, although the phosphate is synthetically more accessible.20 TABLE I1 k,Pit,, DATA FOR CALCIUM ION-SELECTIVE ELECTRODES BASED ON CALCIUM BIS[DI($-1,1,3,3- TETRAMETHYLBUTYLPHENYL)PHOSPHATE] SENSOR Sensor DTMBPP + DOPP* Limit of useful Ca2+ Pot Na+ level/M range from eqn./M Reference kcam 0.017 4.3 x 10-5 11 2.5 x 10-5 15 0’01 0.001 } 0.05 2.5 x 18 t 4.6 x 19 o-nitrophenyl octyl ether 0.01 0.01 1.0 x 10-6 17 Neutral carrier sensors + * DTMBPP, di@- 1,1,3,3-tetramethylbutylphenyl)phosphate ; DOPP, dioctyl.phenylphosphonate. t Sodium interference not measurable. ’+ Calibration limit. § Neutral carrier ssnsor = The relative freedom from sodium ion interference when using (@-1,1,3,3-tetrarnethylbut~+ pheny1)phosphate sensor can be seen from the “at a glance” graphs of Fig.1 (curves B and C compared with E and F). Such at a glance graphs emphasise the utility of k,Pi\ because for interferents at any single level and of the same charge, the lowest k::: values yield the longest interference-free calibration ranges with respect to calcium ions (compare curve B with E, and C with F, in Fig.1). For any one interferent, it frequently occurs that k,P;L is greater for low levels of B than for higher levels. This stresses the importance of quoting the interferent level alongside k , E and also the usefulness of equation ( 5 ) , as the higher k;;; values give the more extensive range with respect to calcium ions (compare curve B with C, and E with F, in Fig. 1). Studies on Di[p-( 1,1,3,3-tetramethylbutyl)-o-nitrophenyl]phosphate Sensor Effect of Zinc Interference Continuation of the studylg of the behaviour of a calcium ion-selective electrode with a more electrophilic octylphenyl group in the phosphate sensor, as with the calcium salt of di[$- (1,1,3,3-tetramethylbutyl)-o-nitrophenyl]phosphate, has shown an interesting aspect concern- ing zinc interference.15 A long recovery time is normally required by calcium ion-selective electrodes based on dialkylphosphate sensors after exposure to zinc.13 This is not the case with the nitrated sensor but unfortunately its calibration range is shorter.Also, calcium electrodes made from the nitrated sensor with dioctyl phenylphosphonate solvent mediator exhibit less interference than the non-nitrated sensor when only low levels of zinc are present (Fig.2).36 > E > u! c! 3 h c m - 5 RESEARCH AND DEVELOPMENT TOPICS R o c . Analyt. Div. Chem. SOC. F E D B A I I I I I l l I I I 6 5 4 3/16 5 4 3 PCa Fig. 1. Calibration of calcium ion-selective electrodes containing (I) calcium bis [di(p-1,1,3,3-tetramethylbutylphenyl)phosphate] sensor with dioctyl phenyl- phosphonate solvent mediator and (11) Orion 92-20-02 calcium liquid ion exchanger trapped in a PVC matrix membrane.15 A and D, Calibration with calcium chloride standards; B, C, E and F, calibration with calcium chloride standards in background sodium chloride solutions containing 0.05 M (B and E) (hgka = 0.01 and 0.11, respectively) and 0.15 M (C and F) = 0.002 8 and 0.050, respectively) sodium chloride. pH Interferences In an alternative mixed solution method for expressing interferences (Method IIB of refer- ences 9 and lo), the primary ion level is kept constant while the interferent level, say pH, is varied.Normally with calcium dialkylphosphate sensors and dioctyl phenylphosphonate solvent mediator a dip appears in the pH interference curve which, with the more electro- philic di(p-octylpheny1)phosphate sensors, occurs at lower pH values.This dip was not observed for PVC matrix membrane electrodes based on calcium di(@-nitropheny1)phosphate sensor with dioctyl (m-nitropheny1)phosphonate solvent mediatorlg or with the same mediator in conjunction with di [@- (1,1,3,3-tetramethylbut yl) +nit rophenyl] phosphat e ~ens0r.l~ I I I / 11 I I 5 4 3 '5 4 3 P Ca Fig. 2.Calibration of calcium ion-selective electrodes containing (I) calcium bis [di( p - 1, 1 ,3,3-tetramethylbutylphenyl) phosphate] with dioctyl- phenylphosphonate solvent mediator and (11) calcium bis {di[P-l,1,3,3- tetramethylbutyl(o-nitro)phenyl]phosphate} with dioctyl phenylphosphonate solvent mediator trapped in a PVC matrix membrane.15 A, Calibration with calcium chloride standards ; B-D, calibration with calcium chloride standards in background zinc chloride solutions containing 0.000 5 M (B), 0.005 M ( C ) and 0.05 M (D) zinc chloride. The slight dip in the pH curve with the lower permittivity mediators like dioctyl phenyl- phosphate or tripentyl phosphate in conjunction with the nitrated sensor becomes moreJuiiiiary, 1979 PESTICIDE RESIDUE AN4LTiSIS 37 prominent with the non-nitrated di[p-( 1,1,3,3-tetramethylbutyl)phenyl]phosphate sensor.Such trends seem to suggest that the existence of a dip in the pH interference curves for the phosphate ester type sensors of calcium ion-selective electrodes are related to the acid strength of the diester. Thus, the sensors of high pK,, with their strong affinity for protons, exhibit dips in the pH interference curves, while the sensors of low pK, are less likely to do so, particu- larly with solvent mediators of sufficiently high permittivity such as dioctyl (m-nitropheny1)- phosphonatr .Conclusion The selectivity coefficient, k::;, associated with the level of interferent at wh‘ch it is determined is a satisfactory parameter for expressing the practical scope of calcium ion selective electrodes.The best electrode is based on calcium dire-( 1,1,3,3-tetramethylbutyl)- phenyl] phosphate as sensor plus dioctyl phenylphosphonate as solvent mediator for reasons of convenience of fabrication and freedom from interference. The nitrated form of the sensor, although appearing to give greater freedom from low level zinc interference and also from pH interference when used in conjunction with dioctyl (9%-nitrophenyl)phosphate, does not offer a sufficient additional advantage over its non-nitrated counterpart to compensate for the loss of calcium calibration range. The authors are grateful to the University of Technology, Baghdad, Iraq, for financial support (to XSN). 1 . ? -. 3. 4. 8 . 6. -3 8 . 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. References Koss, J . U’., in Durst, K. A, Editov, “Ion-Selective Electrodes,” Special Publication 314, National Bagg, J . , Nicholson, O., and Vinen, R., J . Phys. Chem., 1971, 75, 2138. Cattrall, K. W., and Drew, D. M., Analytica Chim. Acta, 1975, 77, 9. Moody, G. J., Nassory, N. S., and Thomas, J. D. R., Hung. Sci. Instvum., 1977, 41, 23. IIJPAC, “Recommendations for Nomenclature of Ion-Selective Electrodes,” Puvp A p p l . C h e w , 1976, Nicolsky, B. P., and Schultz, M. M., Zh. Fix. Khim., 1962, 36, 704. Nicolskq-, B. P., Schultz, M. M., Belijustin, A. A., and Lev, A. A,, in Eisenman, G., Editor, “Glass Srinivasan, K., and Rechnitz, G. A., ,4nalyt. Chem., 1969, 41, 1203. Moody, G. J . , and Thomas, J. D. R., Lab. Pvact., 1971, 20, 307. Moody, G. J . , and Thomas, J. D. R., “Selective Ion Sensitive Electrodes,” Merrow, \Vatford, 1971. Moody, G. J., Nassory, N. S., and Thomas, J . D. R., Analyst, 1978, 103, 68. Craggs, X., Keil, L., Moody, G. J., and Thomas, J . D. R., Talanta, 1975, 22, 907. Moody, G. J , Oke, R. B., and Thomas, J. D. R., Analyst, 1970, 95, 910. Moody, G. J . , and Thomas, J . D. R., Talanta, 1971, 18, 1251. Moody, G. J., -?;assory, S. S., and Thomas, J . D. R., t o be published. KbEiCka, J., Hansen, E. H., and Tjell, J . Chr., ilnaZ&a Chzm. Acta, 1973, 67, 155. Amman, D., Bissig, R., Guggi, M., Pretsch, E., Simon, W., Borowitz, I. J., and Weiss, L., Helv. Chim. Acta, 1975, 58, 1535. Birch, B., Craggs, A,, Moody, G. J . , and Thomas, J. D. R., i n Pungor, E., and Buzas, I., Editors, “Ion-Selective Electrodes,” Akademiai Kiadb, Budapest, 1978, p. 335. Iieil, L., Moody, G. J . , and Thomas, J . D. K., Analytica Chirn. Acta, 1978, 96, 171. Craggs, Bureau of Standards, Washington, D.C., 1969, p. 57. 48, 127. Electrodes for Hydrogen and Other Cations,” Marcel Dekker, New York, 1967. Delduca, P. G., Keil, L., Key, B. J . , Moody, G. J . , and Thomas, J . D. R., J . Inovg. Xucl. Chem., 1978, 40, 1483.