The four players in the Analytical Performance

Abstract

We, analytical chemists, have reason to think ourselves masters of the universe. After all, where would the world be without the massive amount of data that pervade virtually every aspect of society. From law enforcement to health care, from environment protection to food safety, from prehistoric fossils to planetary observations, all rely on the intricate measurements that have become available since the instrumental revolution transformed the analytical profession in the two decades between 1955 and 1975. However, the analytical chemist is only one of four actors who together have welded the analytical necklace to its current brilliance. During my working days in academia and industry I have performed two roles myself, and closely observed the other two, so I feel free to reflect on all four in the order of their appearance on the stage.

---

1. The inventor

By initiating the chain with the inventor of an analytical method, I may meet the objection that I ignore the discoverer of the physical phenomenon that underlies any modern instrumental technique. While it is true that these discoveries form the necessary conditions for the ensuing instrumental revolution, it is also true that the physicists who made the discoveries had no idea about its analytical potential. He (for with a rare exception—e.g., Madame Curie—the discoverer was a male physicist) was primarily interested in the physical phenomenon that he was the first to observe. The most revealing example is the discovery of the resonance transition between the two energy levels of the electron spin of the hydrogen atom by Bloch and Purcell shortly after World War II. When with increasing resolution the broad absorption profile split into increasingly finer detail the physicists were disappointed, but the chemists marveled at the appearance of first the chemical shift and thereafter the spin-spin coupling that form the basis of modern NMR.

It is this transformation of the physical discovery into an analytical tool that is the merit of the inventor. Several received a Nobel Prize for their efforts as enumerated in Table 1.

Table 1 Nobel Prizes for discoveries and inventions related to instrumental analysis

Year	Recipients	Subject
1952	Martin and Synge	Partition Chromatography
1952	Bloch and Purcell	Nuclear Magnetic Resonance
1959	Heyrovsky	Polarography
1964	Hodgkin	X-Ray Crystallography
1977	Yalov, Schally and Guillemin	Radio Immuno Assay
1981	Siegbahn	Electron Spectroscopy for Chemical Analysis (ESCA)
1986	Ruska, Rohrer and Binnig	Scanning Tunneling Microscope
1991	Ernst	2-Dimensional NMR
1993	Mullis	Polymerase Chain Reaction
2002	Bennet Fenn	Electrospray Ionization in Mass Spectrometry

---

Unfortunately, the Swedish committee failed to recognize three excellent protagonists in atomic spectroscopy: Alan Walsh, Boris L'vov and Velmer Fassel.

Alan Walsh worked at an Australian Research Centre (CSIRO) when he published his landmark paper on atomic absorption spectrometry in 1955.¹ In that same year Cees Alkemade, who was to become professor of physics at the University of Utrecht, The Netherlands, also published on the same subject,² but Walsh is rightly identified as the inventor of the analytical technique for two superior components in his optical arrangement. Whereas Alkemade used a flame fed with a heavy dose of the analyte as the light source, Walsh advocated the much more intense and stable hollow cathode lamp that has survived to this very day. Walsh also had the insight to place a chopper between the light source and the flame so that an ac-amplifier prevented the flame's radiation from spoiling the absorption measurement.

Fig. 1 Sir Alan Walsh using his AA spectrometer with hollow-cathode lamp and laminar flame.

Boris L'vov, who worked at the Technical University in Leningrad (now St Petersburg) also had a competitor, namely Hans Massmann in Germany. Both described the small, electrically heated graphite tube as the chamber for the absorbing atoms in 1968.^3,4 Massmann's untimely death and L'vov's continuing improvements of the furnace design make him the recognized authority in furnace-AAS.

Fig. 2 The author in discussion with Professor L'vov in Leningrad (now St Petersburg) in 1980.

Velmer Fassel was professor at the Iowa State University when he published his highly influential paper on Inductively Coupled Plasma Atomic Emission Spectrometry.⁵ It is true that Stan Greenfield from England had already introduced the ICP one year earlier,⁶ but Fassel's design proved superior: with better stability and less prone to interferences. The advantage was the result of the lower argon flow of 1 l min⁻¹ that Fassel used for introducing the nebulized sample solution.

It is noteworthy that these three inventors all had an academic or research institute background. That appears the general observation for this period, although there were of course exceptions. The capillary column in gas chromatography was invented by Golay at Perkin Elmer and X-Ray Fluorescence Spectrometry gained its increasing popularity for the analysis of solid samples mainly through the inventions made at the Philips Research Laboratory in Eindhoven, The Netherlands. But the majority of basic innovations in the prominent instrumental techniques, e.g., chromatography, NMR, MS, and also the many consecutive inventions in atomic absorption and emission spectrometry arose from academic research.

2. The instrument maker

The second player entering the stage is the instrument maker who converts the often crude academic prototype of the novel technique into a reliable instrument. This requires close cooperation between the engineer who designs the instrument and the sales man who convinces the customer of its merits for solving his particular problem. The speed by which the instrumental techniques reached the market varied significantly from one technique to another. Gas chromatography was an immediate success because it was embraced by all oil companies for its astonishing capability to produce a detailed composition of oil fractions, and that in just one or two hours! On top of that, a gas chromatograph was relatively easy to build yourself.

For a variety of reasons the atomic spectroscopic methods took rather more time and effort. In 1955 the distance from Australia and the western world was still rather large, in miles and in appreciation. At the time inorganic analysis was dominated by classical methods and for the determination of elements, i.e., ions, in solution there was a refined collection of volumetric methods available. Finally, the major atomic spectrometry companies were remarkably reluctant to perceive that AAS could be a poor man's alternative to their sophisticated but expensive multichannel direct reader emission spectrometers.

Hence, the introduction of first flame AAS and later furnace AAS was pioneered by two American companies with experience not in atomic but in molecular spectrometry, notably UV and IR. They were Varian that had taken over the AA-prototype of the Australian company Techtron, and Perkin-Elmer that took a lot of coaching from Alan Walsh and some ‘young turks’ in the company to introduce around 1963 the legendary double-beam model 203, that captured the market in no time.⁶ It seemed that the traditional atomic spectroscopy companies had learned their lesson, because some years later they were quick to install the ICP emission source in their multichannel instruments. Yet, the market was reluctant, because the instruments were expensive and customer-designed for up to 20 fixed elements simultaneously. It was therefore again for Perkin Elmer to design a modestly priced, computer-controlled, sequential instrument that could be programmed by the customer to measure any arbitrary, and daily variable set of elements.

By this time, the product managers in the instrument companies were eager to scout the academic groups in their field for continuous innovations. Let me give one example from my own experience. By 1977 furnace-AAS was an accepted technique, but a discussion meeting after the CSI-meeting in Prague revealed that an overwhelming majority of the users relied on the standard addition technique to combat its capricious behavior. Two remedies arose from this meeting. The first was L'vov's proposal for a platform inside the furnace that allowed the atoms to be released in an already heated environment. The second was the Zeeman background correction technique with which we were experimenting in Delft. At the time we had acquired quite an extensive knowledge about atomic spectral lines and its effect on the (non)linearity of calibration curves in AAS.^8,9 We were then visited by the product managers from first Varian and soon thereafter Perkin Elmer who both expressed a flattering interest in our research. Varian asked us to calculate the calibration curves in Zeeman AAS for the two alternative configurations of the magnet: around the furnace or around the hollow-cathode lamp; and could we please do that to concentrations well beyond the range normally encountered in furnace AAS. Perkin Elmer was rather taciturn, but invited me for a return visit to Connecticut where they picked my brains about all aspects of AAS.

It turned out that both companies were already experimenting with the Zeeman correction technique and had made a surprising observation that we academics had missed: at very high concentrations the calibration curve in ZAAS passes through a maximum and reclines back towards the concentration axis! Unfortunately for Varian the theoretical curves calculated from the line profiles did not demonstrate such a phenomenon, but once we were alerted to it we came up with stray-light as the culprit.^10,11 Later, we even devised a technique to overcome the problem by measuring the Zeeman signal at three magnetic field strengths rather than the usual two,¹² but this little invention has only recently gained attention.¹³

3. The customer

Enter now, and only now, the analytical chemist who will buy the instrument and adapt it to his needs. He is the primary beneficiary of the marvels that the instrumental revolution have brought. As illustrated in Table 2 the detection limits gradually decreased from the easily detected ppm of flame-AAS to the proverbially difficult ppb of furnace-AAS and ICP-AES and ultimately to the mind-boggling ppt of ICP-MS. Simultaneously, the analysis time was reduced from days to hours and even minutes. Stable instruments controlled by microcomputers and equipped with autosamplers allowed fully automated, unsupervised measurements. Above all, an enormous range of new techniques became available that expanded the profession from the largely inorganic classical analysis into the domains of the organic chemist and ultimately the biochemist. It is now possible to measure almost anything in almost everything.

Table 2 The gradual decrease of detection limits

Decade	Detection limit	Example	Analogy
1960	ppm 1 : 10^6	Lead in gasoline	First gray hair
1970	ppb 1 : 10^9	Polycyclic aromatics	Needle in a haystack
1980	ppt 1 : 10^12	Dioxine in milk	Contact lens on 100 mile beach

---

But this tremendous progress comes at a price. Most science historians argue that the above quantum leaps in performance constitute a true paradigm shift of the analytical profession and hence they speak of the instrumental revolution. In my opinion the shift lies elsewhere. Whereas classical analysis with its burette and balance was absolute because it relied on analyte-specific chemical reactions run to completion, the modern instrumental methods are all relative because they require a calibration with known standards to convert the physical signal to a chemical concentration. It is their common weakness and has given rise to a whole new vocabulary with such concepts as interferences, matrix effects, dynamic range, line (or peak) overlap. It has stimulated a renewed interest in reliability characterized by the traditional concepts of random and systematic errors. The analytical chemist plays an important role here with his Standard Operating Procedures, international validation, proficiency tests, Certified Reference Materials, control charts, and external certification.

But in our discussion of the analytical chain the analytical chemist has another vital role to play. Undoubtedly, the inventions of all analytical instruments were inspired by technology push. Starting with electronic devices such as the triode amplifier and the photomultiplier tube, followed by solid state devices and microcomputers without which Fast Fourier Transform spectrometers would not have been possible, and continuing to this day with diode-array detectors and ultra-high field NMR magnets.

The ultimate success of a novel method is decided, however, by whether or not it fulfills a customer demand of the analytical chemist. He is eager to accept any new technology as long as it promises to solve a hitherto insurmountable problem. Thus flame-AAS enabled the simultaneous determination of calcium and magnesium, and furnace-AAS provided the detection limits needed to satisfy the rising interest in environmental spoils. The same is true for other instrumental techniques as exemplified in Table 3. Conversely, when an innovation did not offer a unique advantage, it was doomed to failure. Such was the case for atomic fluorescence spectrometry and also for many electrochemical methods, e.g., polarography.

Table 3 Immediate successes of novel instrumental methods of analysis

Flame photometry	Determination of Na and K
Infrared Spectrometry	Functional groups in organic molecules
Flame Atomic Absorption Spectrometry	Simultaneous determination of Ca and Mg
Gas Chromatography	Unraveling the composition of mineral oils
Nuclear Magnetic Resonance	Typifying differently bonded hydrogen atoms
Mass Spectrometry	Library searching to identify organic compounds
Furnace AAS and ICP-AES	Elements at the ppb-level

---

4. The client

The grand finale in our analytical symphony is left to the ultimate receiver of the data provided by analytical laboratories around the world. It may be the consumer in the supermarket inspecting the label specifying the ingredients (and calorie intake) of her food product, but more often it is the factory engineer, the public prosecutor, the hospital physician, or the government official. You would expect them to be overjoyed with the new possibilities that our profession offers them, and indeed they are. On the other hand, it is also true that they have not the slightest interest in the numbers that the analytical chemist delivers, be it cholesterol content, calcium concentration or dioxine traces. The client of our analytical services is not interested in concentrations but in properties. To appreciate the distinction we must take a closer look at the analytical sequence that consists essentially of three steps:

a) It starts by asking what it is you want to know. That is not for the analytical chemist but for the client to decide, and it is not expressed in numbers but in one or more properties of a product, a material, a system, a patient etc.

b) Now it is difficult to determine a property directly, and therefore we want to identify a marker indicative of the desired property, and one that is also easy to measure.

c) It is then the task of the analytical chemist to develop a suitable procedure for the measurement.

It may come as a disappointment for us analytical chemists, but the really difficult step is not the last but the second step, finding a suitable marker. Some examples are shown in Table 4.

Table 4 Markers in the analysis of products or objects

Object	Desired property	Marker	Analytical method
Steel	Strength	C, Ni, Cr,…	AC-Spark
Water	Hardness	Ca, Mg	Flame-AAS
Human Body	Risk of heart failure	Cholesterol	Liquid Chromatography
Mummy	Age	C-14	Mass Spectrometry
Margarine	Microbiological Safety	Water droplet size	Low frequency NMR
Olive oil	Shelf-life	UV-absorption, Oxydation resistance	UV-spectrometer, Rancimat

---

I trust that the top four examples are familiar, but the bottom two may require an explanation. Like butter, margarine starts as a (vegetable) oil-in-water emulsion but intense churning converts it to a water-in-oil emulsion. Weight and health awareness have promoted a drive towards lower fat and thus higher water contents with an inherent increased risk of molding, especially when the spread is kept outside the refrigerator. However, bacteria cannot grow in water droplets smaller than their size and hence the water droplet size is an excellent predictor of the margarine's microbiological safety. The Unilever Research Laboratory which I joined in 1987, had developed for their margarine factories a remarkably simple technique to measure both the average and the spread of the water droplet size using a bench-top low-frequency NMR instrument. Only if the result was satisfactory could that days production of margarine be released.

For the olive oil the situation was less clear. All oils eventually become turbid from crystal formation. Although perfectly harmless, no housewife will buy such an oil. Prediction of shelf-life therefore became of paramount importance, but how? I seconded an adventurous young lady to our company in Italy and she measured no less than 10 different parameters for a large number of olive oils over a period of three months, by which time all oils had crystallized, but some much sooner than others. I handed the massive amount of data to my statisticians and they came up with a simple prediction using only two parameters: the UV-absorbance which was of course easy to measure and the oxidation resistance for which there is unfortunately no marker yet. As is often the case with product properties we then resort to an imitation of real life, but highly exaggerated: heated olive oil is subjected to a stream of pure oxygen and the eventual breakdown is signaled by an abrupt rise in electrical conductivity. In mineral as well as in vegetable oils this is known as the Rancimat method. For some very resistant oils it may take as long as 24 h! So, as we used to say in my company at the end of yet another management course: there is certainly room for improvement!

References

1. A. Walsh, Spectrochim. Acta, 1955, 7, 108.
2. C. Th.J. Alkemade and J. W. M. MIlatz, Appl. Sci. Res., 1955, B4, 289; J.Opt.Soc.Amer. 45, 583 ( 1955).
3. B. V. L'vov, Spectrochim. Acta, 1961, 17, 761 , 24B, 53(1969).
4. H. Massmann, Spectrochim. Acta, 1968, 23B, 215.
5. R. A. Wendt and V. A. Fassel, Anal. Chem., 1965, 37, 920 , 38, 337(1966).
6. S. Greenfield, I. L. Jones and C. T. Perry, Analyst, 1964, 89, 713.
7. H. L. Kahn and W. Slavin, Applied Optics, 1963, 2, 931.
8. L. de Galan and H. C. Wagenaar, Meth. Phys. Anal., Sept. 1971, 10.
9. H. C. Wagenaar and L. de Galan, Spectrochim. Acta, 1975, 30B, 361.
10. L. de Galan and G. F. Samaey, Spectrochim. Acta, 1969, 24B, 679.
11. M. T. C. de Loos Vollebregt and L. de Galan, Appl. Spectrosc., 1980, 34, 464.
12. M. T. C. de Loos Vollebregt, L. de Galan and J. W. M. van Uffelen, Spectrochim. Acta, 1986, 41B, 825.
13. H. Gleisner, K. Eichardt and B. Welz, Spectrochim. Acta, 2003, 58B, 1663.

---