Tribute to Alan Walsh from the analytical chemistry group at Imperial College, London, 1963–1975

Tribute to Alan Walsh from the analytical chemistry group at Imperial College, London, 1963–1975

Spectrochimica Acta Part B 54 Ž1999. 2017]2022 Tribute to Alan Walsh from the analytical chemistry group at Imperial College, London, 1963]1975 q T.S...

76KB Sizes 0 Downloads 54 Views

Spectrochimica Acta Part B 54 Ž1999. 2017]2022

Tribute to Alan Walsh from the analytical chemistry group at Imperial College, London, 1963]1975 q T.S. WestU Chemistry Department, Uni¨ ersity of Aberdeen, Aberdeen, Scotland, UK Received 17 April 1999; accepted 28 July 1999

Keywords: Atomic absorption spectrometry; Atomic fluorescence spectrometry; Furnace; UV spectrophotometry; History

1. Introduction I am indebted to the editors of this special edition of Spectrochimica Acta B in honour of the late Sir Alan Walsh, FRS, for the invitation to make a personal contribution in relation to his influence on my own work and that of the staff and postgraduate students of the research school of Analytical Chemistry at the Imperial College of Science and Technology of the University of London. Alan’s enormously significant contribution to the development of a laboratory technique for measuring atomic absorption during the earlyto-mid-1950s had a profound effect on my thinking about how the new research team I was about to set up there in late 1963 could best make a worthwhile contribution to the determination of traces of metal ions and to the training of postgraduate students in relevant modern techniques. q

Dedicated to the memory of Sir Alan Walsh and published in this Memorial Issue. U Correspondence address: Professor Emeritus in Chemistry, 31 Baillieswells Drive, Bieldside, Aberdeen AB15 9AT, Scotland, UK.

2. Trace metal analysis before the Walsh era

To set this out logically it is necessary, first of all, to review the current technology of the time and to follow this with a brief account of what we were doing just before Alan’s seminal contribution of atomic absorption spectroscopy ŽAAS. to the scene. At that time there was generally no simple way of determining traces of metal cations in solution. In most cases the analytical procedures that could be used demanded many laboratory operations before the desired analyte could be presented to the measuring instrument. The whole process was fraught with the possibility of contamination and loss of sample and was, in modern parlance } though we did not think so then } generally time-consuming and labour intensive. The principal technique used in most laboratories of those days was molecular absorption spectrometry. It relied on the use of chelate complexes formed by the metal cations with reagents such as dithizone and oxine, and the selective extraction and preconcentration of these com-

0584-8547r99r$ - see front matter Q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 5 8 4 - 8 5 4 7 Ž 9 9 . 0 0 1 1 6 - 0

2018

T.S. West r Spectrochimica Acta Part B: Atomic Spectroscopy 54 (1999) 2017]2022

plexes into organic solvents such as chloroform. Spectral measurements were made in mainly crude ‘absorptiometers’ which relied on optical filters to isolate the necessary wavebands from a continuum light source. The gradual introduction of more sophisticated spectrometers using prisms and gratings in the late 1940s to early 1950s scarcely improved matters. The problem was that the metallochromic absorption spectra were invariably overlaid with those of the reagents and were rather broad, covering at least 100]200 nm, so that the spectral effect of chelate formation was not capable of being used efficiently. These reagents were extremely unselective in their reactions with metal cations. Sometimes masking agents such as cyanide ion could bind some interfering ions, e.g. Zn and Cd in the determination of Cu, sufficiently to prevent them from reacting with the metallochromic agent. Demasking of Cd was possible subsequently by adding an excess of hydroxylamine, which by cyanhydrin formation removed the cyanide ion from some of the more weakly bound cyano-complexes, e.g. Cd, so that the liberated metal cation could now react with the metallochromic agent. Various other clean-uprpreconcentration techniques such as adsorption chromatography, ion exchange, adsorption]flotation and even, on the odd occasion, distillation were used in addition to solvent extraction as precursors to the final analytical operation. Another spectral technique much used by laboratories which carried out large numbers of repetitive analyses on mainly fixed major matrices, such as steels and other alloys, was emission spectrography using electric arcs and sparks between carbon electrodes, first with photographic plates and subsequently with photomultiplier arrays. It was an extremely empirical technique, but worked well in certain restricted areas. Flame emission spectrometry ŽFES., using the nebulisation of solutions Žadopted subsequently in AAS., was useful for the alkali metals and the alkaline earths and, though also prone to matrix effects, it too was quite widely used in some areas. Flames generally were insufficiently energetic to excite most elements other than those.

Variations using ‘total consumption’ burners and air]acetylene flames extended the range of metals that could be determined by flame emission. Flame emission was largely replaced in due course by inductively coupled plasma emission spectrometry ŽICPES. which was developed principally by Velmer Fassel in the USA and Stan Greenfield in the UK. ICPES, like AAS, is much used at the present time. Other spectroscopic techniques very much in vogue at present such as spark source mass spectrometry were also undergoing development during the early 1960s. It is not particularly relevant to this discussion, but I feel that I should mention the electrochemical technique of polarography devised by Heyrovsky and which subsequently won him the Nobel Prize. It was quite widely used at the point when AAS was undergoing development and, in its various DC-, AC-, square wave-, Tast- and other forms, is still used today though not nearly so widely on the inorganic scene as AAS. It strikes me as sadly ironic that those of us who drew Alan Walsh’s development of the AAS technique to the attention of more than one Nobel Prize Committee were unable to win a similar accolade for him.

3. Pre-AAS at Birmingham University As a newly appointed lecturer at the University of Birmingham in 1955 I had become intensely interested in Schwarzenbach’s technique of complexometry and was more particularly interested in the metallochromic indicators, largely o,o9dihydroxyazo phenyl and naphthyl dyestuffs which he used to detect the disappearance of the last traces of free metal cations in solution as he approached the end-point of his EDTA titrations. It appeared to me that this was a possible source of metallochromes for the absorptiometric determination of traces of metals. The problem with these, which we set out to resolve, was that they were very unselective. Just before AAS hove on the scene, we had attained a considerable degree of success by developing ‘Calcichrome’, a

T.S. West r Spectrochimica Acta Part B: Atomic Spectroscopy 54 (1999) 2017]2022

tris-hydroxy-tris-cycloazo chelate cage molecule with three o-hydroxyl groups in the centre alternated by three azo bonds. At pH 10, when two of the hydroxyls had ionised, the molecule formed a complex with calcium ions, but not with strontium or barium, both of which are bigger, we had achieved a conditionally-specific calcium reagent w1x. Another series of experiments on the introduction of a metal-complexing iminodiacetic acid group adjacent to the o-hydroxy group in dihydroxyalizarin sulphonic acid produced another pleasing success. The cerium ŽIII. complex of this reagent was, and I believe still is, the first ever specific Žand positive. colour reaction for the fluoride ion w2x. We called the reagent Alizarin Complexan: manufacturers called it Alizarin Fluorine Blue. Despite these successes we were very much aware that Alan Walsh’s new AAS technique was virtually completely specific for nearly every metallic element in the Periodic Table and that, as analytical chemists, we could not ignore this remarkable achievement. At first, because of the lack of commercial equipment on the market we carried on with our search for highly selective chromogenic chelate systems for metal ions and experimented further with molecular fluorescence systems in the realisation that fluorescence signals were inherently more sensitive than absorption ones. Electronic amplification of fluorescence signals was beneficial within noise limitations and optical gathering of the fluorescence emission from as wide a spatial area as possible also increased sensitivity. Then, suddenly, an AAS instrument appeared on the UK market which was based on the Hilger UvispeK spectrometer and we successfully petitioned the Department of Scientific and Industrial Research for one. It was a beautiful ruggedlooking instrument with all the AAS gear set out on a length of an optical bar along the axis of the spectrometer just as described in Alan’s publications. Wildly excited by the prospects of experimentation in front of us, one of my young colleagues suggested to one of the local breweries that we would be willing to check if any of their

2019

beers contained traces of lead by this new ‘wonder technique’. A supply of cans arrived, were opened and approximately 5 ml were extracted from each and aspirated into the flame. No lead was found. The rest of the samples were put to traditional use! However, we quickly found that the instrument had been developed to operate only on an air]propane, or similar, slow-burning flame. This was a serious limitation because there was insufficient energy in the flame to atomise most elements efficiently though it worked beautifully for some such as cadmium, lead and zinc. In those pre-‘Health and Safety Act’ days we did not hesitate to set out to try and convert it to air] acetylene with little knowledge of the critical design limitations. We were a little shaken and a bit nervous by the time we succeeded, as were many others in the laboratory, but eventually we were able to maintain a reasonably steady premixed flame on a 10-cm long burner head of curious design and could aspirate solutions into it. Then we ran into the second limitation. The designers had used an ingenious classical optical arrangement to defocus light emitted in the flame from the entrance slit of the spectrometer whilst allowing the radiation from the hollow cathode lamp ŽHCL. to be focused on it. This worked well with the low emitting air]propane flame, provided there was not sufficient extraneous material in the analyte solution to cause appreciable flame emission. However, it was not very successful with air}acetylene. The amplifier of the spectrometer and the power supply to the HCL apparently both operated DC so we could not chop the radiation from the HCL and modulate the detector to the same frequency. In later times we would have taken the power supplies to pieces and installed a suitably tuned amplifier ourselves, but in those days we were just simple chemists and knew little about electronics. So we did not get very far beyond zinc, lead and cadmium. When the Australian Techtron AA4 spectrometer came on the market we petitioned the DSIR again for a new instrument, but unfortunately they took the view that we had only recently been supplied with the Hilger and ‘would we kindly get on with it’!

2020

T.S. West r Spectrochimica Acta Part B: Atomic Spectroscopy 54 (1999) 2017]2022

4. Analytical atomic spectroscopy at the Imperial College Shortly afterwards, in 1963, I was appointed to the vacant Readership in Analytical Chemistry at the Imperial College of Science and Technology, University of London, London, and was able to recruit two former Birmingham research students, Roy Dagnall and Gordon Kirkbright, to lectureships. The DSIR responded positively to a request for an AA4 and we set out to develop and expand into AAS. The AA4 was an ideal instrument for those who wanted to investigate and experiment with AAS technology as well as for those who wanted to apply it routinely to analysis. There was so much space around the ‘business end’ of the instrument that virtually any configuration or arrangement of physical parameters could be undertaken. I was not surprised to learn from Max Amos later on that this flexibility probably arose from the influence of Alan Walsh on its design. Based on our Birmingham experience, Roy Dagnall and a small group of our postgraduate students quickly set the optical bar at right angles to the spectrometer’s optical axis and we started to research atomic fluorescence spectroscopy ŽAFS., beginning with the inorganic chemical spectroscopist’s Žand inorganic polarographer’s. favourite element, cadmium, and using a cadmium street lamp as irradiation source. This was a logical extension of the work we had done on molecular absorption and then molecular fluorescence spectroscopy. We assumed that if molecules could become excited by incident radiation and then re-emit it later as fluorescence or phosphorescence, atoms should also be able to do so. At Florida State University, based I am sure on similar experience with molecular species, Jim Winefordner appeared to have been asking the same question. Jim published ahead of us, but it was a great thrill for me to give a public demonstration of AFS at my inaugural lecture as the first full Professor of Analytical Chemistry in Imperial College in January 1966 w3x. Intense atomic line emitting sources were the key to specificity and sensitivity in AFS, so we followed this up with extensive work on the construction of

microwave-excited electrodeless discharge lamps ŽEDLs. containing iodine and a fragment of metal or halide salt as intense atomic line sources. The metal halide was formed when the sealed silica tube was evacuated and ‘cooked’ for the first time in the quarter- or three-quarter-wave cavity of the microwave generator. Subsequently the EDL gave strong line emission for the metal concerned. Air }propane flames were used when practical and later on specially adapted air]acetylene flames to measure the fluorescence signals generated by the corresponding elements aspirated into the flame in the usual way. The construction of EDLs was the main source of success in this area at the outset and much of this depended on the skill, in noting cause and effect, of the first group of postgraduates to engage in this work, particularly Clive Thompson, and on their ability to draw logical conclusions from their experimental observations w4x. Meanwhile, partly based on our early problems in Birmingham with ‘noisily emitting’ flames, Gordon Kirkbright and another small group set out to explore how to get windows into the interconal areas of flames where free atoms were formed immediately above the turbulent and strongly emitting primary combustion zone w5x. To a limited extent this could be done by putting a silica tube around the flame so that the secondary combustion zone, supported by the diffusion of atmospheric oxygen and which emits a continuum in the process, now burned, out of the way, on top of the tube. These ‘singing flames’ as they were sometimes called, were much beloved by Victorian demonstrators of scientific curiosities. However, for scientific observation of absorbing or emitting atomic species in a flame they were very advantageous in removing flame background radiation from the hot atom zone. We used these mechanically separated flames in the upright mode for AFS measurements and in a horizontal mode under slightly reduced pressure from a water pump for AAS. Signal-to-noise ratios were considerably improved and some refractory oxide forming elements could be more easily determined as a result. These silica flame separators were, however, eventually subject to sooting up, devitrification, and contamination by analyte

T.S. West r Spectrochimica Acta Part B: Atomic Spectroscopy 54 (1999) 2017]2022

species and had to be replaced periodically. From these we progressed to nitrogen- and argonsheathed air]acetylene and nitrous oxide] acetylene flames in which a concentric flow of the inert gas around the periphery of the burner head lifted off the diffusion mantle well above the primary combustion zone and allowed spectroscopic measurements } particularly of AFS } to be made with perfect transparency into the centre of the flame against virtually no background emission. Long pathlength rectangular burner heads could also be used very satisfactorily for AAS measurements of refractory oxide forming elements such as aluminium. The shielding or separation gave transparency in the UV area and minimised oxide formation by markedly reducing oxygen diffusion into the interconal zone. Like most flame spectroscopists we found premixed oxy-acetylene flames rather too hazardous to use. However, one hardy student, Roger Stephens w6x, did some very interesting and spectacular experiments with great sang froid on oxyacetylene flames burning under reduced pressure in a large commercial borosilicate glass four-way crosspiece fitted with suitable optical windows. A vacuum pump prevented the flame from striking back and from forming an outer mantle. There were few results of worthwhile analytical significance to be won from these particular experiments, but there were some spectacular emission spectra observed in the greatly extended primary reaction zone of the flame, probably due to free radical and chemiluminescence reactions induced by the aspiration of solutions. We did not pursue this further but deduced that the poor absorption and fluorescence signals obtained with this rather cumbersome arrangement were due to greatly reduced residence time and hence atomisation of the aspirated metal cations in the hot zones of the flame. The resistively heated graphite tube furnace first devised by Massmann at the Institute of Spectroscopy in Dortmund as a means of producing free atoms for AAS measurements was of considerable interest to us. However, we substituted a carbon rod or filament so that we could reap the benefits of using very small samples,

2021

which produced relatively dense transient plumes of free atomic vapour above the surface of the heated rod in an inert atmosphere of nitrogen or argon. Spectroscopic measurements of absorption or fluorescence were easily made despite their transience. At first we used an inert gas flushed dome over the filament to exclude atmospheric oxygen, but later switched to simple shielding with a concentric wall of inert gas w7x. Generally there was considerably enhanced sensitivity over flame-based methods, particularly for fluorescence. In a subsequent development we used inverse Cassegrainian mirror microscope objectives to collect and focus fluorescence signals on the entrance slit of the photomultiplier in non-dispersive AFS measurements. Just as Alan Walsh’s lock-and-key mechanism of the signal from the hollow cathode lamp fitted the slightly broader flame absorption profile of the same free atomic species as the cathode metal and virtually no other, similarly the emission lines from the appropriate EDL stimulated fluorescence of the same species in the atomic plume rising off the carbon rod and virtually no other. In this case there was no flame background and no flameinduced luminescence, so that with an appropriate optical slit arrangement there was no need to use a dispersing monochromator with its invariable loss of signal strength w8x. We were also able to demonstrate non-dispersive time-resolved AFS signals from more than one element in the sample. The temperature profile of the resistively heated rod was a function of time between switching on and the attainment of maximum temperature. Elements such as cadmium atomised well at low temperatures and produced their atomic plume well before others such as bismuth or zinc, which required higher temperatures. When the irradiating EDL contained the requisite elements, in cases such as the above, cleanly separated fluorescence spikes were obtained as each plume responded to the exciting line radiation from the source w9,10x. In conclusion, following these few examples of the work that Alan inspired at Imperial College, I should like to tell a small story about him that, for me at least, typified his humorous and good-

2022

T.S. West r Spectrochimica Acta Part B: Atomic Spectroscopy 54 (1999) 2017]2022

natured outlook on life, his approachability and his cheerful goodwill to all who were privileged to meet him. I have already mentioned Velmer Fassel of Iowa State University in connection with his pioneering work on inductively coupled plasma emission spectroscopy. There has always been a friendly rivalry between the aficionados of absorption and emission spectroscopy. In the days when AAS was running rife Žas it were. Velmer was frequently known to pounce at symposia on any loose statement made by speakers on absorption. One day it was my turn and as I was sitting, feeling rather disconsolate and exposed by the faux pas which Velmer had just featured, a huge paw landed on my shoulder from the row behind and turned me round. And there was Alan with an enormous grin on his face. ‘Congratulations Tom’, he said, ‘You’ve just made it. Welcome to the club. By the way, what took you so long?’

Finally, I acknowledge with gratitude the tremendous debt I, and I am sure all my colleagues at Imperial College and subsequently The Macaulay Research Institute in Aberdeen, owe to Alan. He was universally liked and admired as a man as well as a scientist by all of us. It is difficult to express how profoundly he and his work affected our careers and consequently our lives, but there is no doubt whatsoever that we all benefited enormously by the excitement and the prospects we saw through the window of opportunity he opened for us. John Keats put my feeling of this ‘fascinating open window’ that Alan showed us more aptly than I ever could, in his Ode to a Nightingale:

wHex Charm’d magic casements, opening on the foam of perilous seas in faery lands forlorn.

That really was what Alan Walsh did for us all. It was magic! The world is a greyer place without him. References w1x R.A. Close, T.S. West, A new selective metallochromic reagent for the detection and chelatometric determination of calcium, Talanta 5 Ž1960. 221]230. w2x M.A. Leonard, T.S. West, Chelating reactions of 1,2dihydroxyanthraquinone-3-yl-methylamine-N, N-diacetic acid with methyl cations in aqueous media, J. Chem. Soc. Ž1960. 4477]4486. w3x T.S. West, Inaugural Lectures, Imperial College of Science and Technology Vol. 1964]1965 Ž1967. 221; Some sensitive and selective reactions in inorganic spectroscopic analysis, Analyst 91 Ž1966. 69]77. w4x R.M. Dagnall, T.S. West, Some applications of microwave-excited electrodeless discharge tubes in atomic spectroscopy, Appl. Optics 7 Ž1968. 1287]1294. w5x G.F. Kirkbright, T.S. West, The application of separated flames in analytical flame spectroscopy, Appl. Optics 7 Ž1968. 1305]1311. w6x R. Stephens, T.S. West, Some analytical properties of low pressure flames, Spectrochim. Acta Part B 27 Ž1972. 515]526. w7x T.S. West, Developments in atom reservoirs and line sources for atomic absorption and atomic fluorescence spectroscopy, Pure Appl. Chem. 23 Ž1970. 99]126. w8x A.F. King, T.S. West, Time-resolved non-dispersive atomic fluorescence by the carbon filament atomization technique, Proc. Soc. Anal. Chem. 10 Ž1973. 279]280. w9x T.S. West, Some recent developments in atomic fluorescence spectroscopy, Pure Appl. Chem. 50 Ž1978. 837]843. w10x M. Hargreaves, A.F. King, J.D. Norris, A. Sanz-Medel, T.S. West, Non-dispersive AFS with a carbon filament atom reservoir, Anal. Chim. Acta 104 Ž1979. 85]92.