The Reformation of Analysis

A decade ago, when I left the laboratory for the sake of the history of science, I had some ideas about the instrumental revolution in chemistry, and I believed I knew the basics concerning the changes in chemical practice brought about by the introduction of analytical instrumentation (1, 2). I was surprised, however, to find that these changes ran deeper than I thought, both socially and epistemologically. The analyst of today is no longer the analyst of the 1930s, and the analyst of the “instrument era” is less than aware of this fact.

Instrumentation shifted the focus of analysis silently from “separation” to the “identification” of compounds. As late as 1940, the analytical chemist’s job – namely to “separate” and “quantitatively manufacture” compounds using their reaction properties – was downgraded and given to available research technicians. On the epistemological level, this means that although analytical chemistry remains the discipline that focuses on “signal production and interpretation,” the very nature of the signal read and interpreted has changed; the focus of analytics has shifted from the chemical properties of substances to the properties of “molecular species” (3). On the sociological level, the analytical scientist of the post-war decades gradually became a professional “manager” of personnel, freed from the “dull” work of separation, and able to focus on the elemental properties of compounds – less a laboratory worker, and more a “scientist” of the type that a physicist is.

In this shift, the chromatographic instruments played a pivotal role – both by altering the separation procedures themselves, and by offering a paradigm for all the subsequent laboratory instrumentation. By the end of the 1960s, gas chromatography was the closest to being the dominant analytical method among all the available methods of instrumental analysis. As a method, GC was, from the beginning, characterized by its versatility: capable of analyzing samples over a broad qualitative range, easily adapted for preparative work, and able to operate at different scales of quantity and precision. Most importantly, the machine was complicated, involving too many different types of knowledge to exist. The production of the apparatuses was delegated outside, away from the chemical laboratories per se, and for the first time in the history of chemical analysis, an important part of the analytical process was effectively made invisible. While nowadays almost any chemist has an inbuilt psychological distance from the instruments that he is using, this distancing of the research apparatuses from their users was then a new and unexplored path.

The outsourcing turned the instruments into commercial objects like any other, creating a vibrant market and giving production companies the incentive to create meaningful R&D facilities. The companies would now compete not only through improved technology and products, but also through service structures, advertisements, “lobbying” and “special relations” with the “clients” – in this case, universities, hospitals, public institutions. For the first time and on a massive scale, there was a clear divide between the users of the instruments and the designers and builders.

The dissemination of the new techniques followed the chaotic paths familiar to us nowadays from the dissemination of computer and communication technologies (4): it took us away from academic curricula, to the modular transmission of knowledge packages – discouraging theoretical depth and instead encouraging “user-friendly” applications. In this way, the transformation of analytical science into a full science has come at a price: the loss of knowledge and control over basic laboratory processes. The new analytical scientist knows more on a higher level; yet the knowledge of acquisition processes has been irremediably black-boxed – and is now out of his or her control.