Skip to main content
Vacuum and cryogenics

Vacuum and cryogenics

Lab architecture

01 Apr 2008 Robert P Crease

Laboratory design can be complex and risky, as Robert P Crease explains

Landmark labs

There are few more dramatic illustrations of the vicissitudes of laboratory architecture than the contrast between Building 20 at the Massachusetts Institute of Technology (MIT) and its replacement, the Ray and Maria Stata Center. Building 20 was built hurriedly in 1943 as temporary housing for MIT’s famous Rad Lab, the site of wartime radar research, and it remained a productive laboratory space for over half a century. A decade ago it was demolished to make way for the Stata Center, an architecturally striking building designed by Frank Gehry to house MIT’s computer science and artificial intelligence labs (above). But in 2004 — just two years after the Stata Center officially opened — the building was criticized for being unsuitable for research and became the subject of still ongoing lawsuits alleging design and construction failures.

Science historians find laboratory architecture fascinating, for it often reflects and reinforces assumptions about scientific life that may or may not pan out. In an article to be published next month in the journal Historical Studies in the Natural Sciences, Stuart W Leslie — a historian at Johns Hopkins University in the US — explores the tensions that can arise between the visions of a lab’s director and its architects. The article forms part of a larger book project, The Architects of Modern Science, which will include chapters on the likes of I M Pei and also William Pereira, who designed General Atomics’ lab in La Jolla, California.

Corporate image-making

Leslie’s early work examined the career of Charles Kettering, the powerful founding director of the research labs at General Motors (GM) and a man who dominated the firm’s R&D from 1919 until his retirement in 1947. Kettering, who was not a PhD scientist and detested academics, thought that industrial labs should be practical environments closely linked to product improvement. The machine-shop atmosphere of GM’s Delco Radio lab in Kokomo, Indiana, for instance, which Kettering oversaw, mingled scientific equipment with lathes, presses, grinders and other machinery.

But Leslie later became fascinated by the drastically different style of the laboratory that GM built after the Second World War — the high-Modernist GM Technical Center outside of Detroit, Michigan. Designed by the Finnish-born architect Eero Saarinen and completed in 1956 at a cost of $125m, the lab was a huge contrast with GM’s prewar research buildings. As Leslie explains in an article he wrote with Scott Knowles in 2001 (Isis 92 1), top managers at GM — as well as those at other corporate giants like AT&T and IBM — abandoned Kettering’s original vision after the war.

Bosses at these firms decided that “the isolated campus was the ideal model for R&D and that basic research required a new spatial and symbolic identity, and an elaborate public stage”. Indeed, it was to forge this new identity that the help of famous architects like Saarinen was sought. But the outcome, Leslie and Knowles found, eventually harmed the companies by interrupting the link between labs and both products and production.

This came about not only through the often strange designs of the labs but also because they were isolated from other company facilities. “More and more,” Leslie and Knowles wrote, “the crucial task of linking research to production fell to branch laboratories located near manufacturing facilities, such as the IBM laboratory at San Jose, where disk storage was invented, [and] Bell Laboratories’ branch at Allentown, Pennsylvania, where Western Electric [the manufacturing wing of AT&T] produced state-of-the-art vacuum tubes and, later, semiconductors and integrated circuits.”

Similar stories unfolded when Saarinen built IBM’s T J Watson Research Center in 1961 to replace the firm’s lab in North Street and when he designed a new facility for Bell Labs in Murray Hill, New Jersey, the following year. Both buildings won architectural awards for their glass facades, modular interiors and other innovative features. However, as Leslie and Knowles discovered, the architecture served corporate image-making better than science.

The designs tended to cater to the companies’ appearance on the public stage rather than to the performance of those working inside, in a way that would prove detrimental to the corporations. Leslie and Knowles (borrowing a contemporary term meant as praise) dubbed these architectural marvels “industrial Versailles”, for the buildings symbolized “an increasingly dangerous isolation from the outside world that would eventually threaten to undermine the very legitimacy of the regimes they so powerfully expressed”.

Architectural tension

In Leslie’s new article, he considers two labs that, in terms of their architecture, exhibit “the essential tension between an imaginative client and an inspired architect at its best”. In both cases, the interaction between science manager and designer resulted in relatively successful structures that did not pin the research that took place inside to one model, but instead allowed for growth and adaptation.

The first is the Mesa Lab of the National Center for Atmospheric Research (NCAR) in Boulder, Colorado, which opened in 1966 (above left). Its design emerged from a collaboration between the centre’s director, the solar astronomer Walter Roberts, and the architect I M Pei. Roberts wanted something to “represent the dignity and importance of the centre as a national scientific laboratory” and Pei largely delivered on this request, creating a building with a distinctive look that was praised by critics.

However, a few features of the lab proved embarrassingly useless in practice. The most notorious of these was a set of top-floor “crows nests” — accessible only by small circular staircases and equipped with tiny balconies — where Pei envisioned individual scientists retreating to meditate in solitude. Few did. As for the beautiful courtyard that Pei envisioned would bustle with activity, it was mainly deserted.

Moreover, the building, designed for small teams of researchers working independently, was soon at odds with the new world of large, formally managed projects linked to huge computers and international networks of other researchers. Additional underground space for computers helped overcome this, and — unexpectedly — other parts of the building provided the social space and arena for networking that the researchers desired. As a result, the Mesa Lab has remained a viable — even outstanding — lab, while retaining its distinctive architecture.

The other building Leslie looks at is the Salk Institute in La Jolla, California, which is a not-for-profit private lab dedicated to the biological sciences. Both the institute’s founder Louis Salk and its architect Louis Kahn were inspired by the monastery at Assisi — an early model of a planned community — and came to think that a lab also should protect researchers from distractions such as teaching and grant writing.

But both Salk and Kahn came to appreciate that a lab has to be flexible; Salk liked to speak of his laboratory as “like a living organism”, containing working space that “is capable of differentiation in response to evolving needs”. This requirement, plus a sudden budget crisis, inspired Kahn to modify his early, monastic design and open up much of the planned space. This did a lot to contribute to the eventual success of the laboratory, allowing it, as Leslie says, to grow into “a conventional academic laboratory in an unconventional building”.

Pedestrian architecture may suit a laboratory for the simple reason that it does not project, and set literally in concrete, a single way of doing science. This may explain why, as Leslie says in his upcoming article, many of the world’s most renowned laboratories, such as the Cavendish Laboratory in Cambridge, MIT’s Rad Lab, Los Alamos and Bell Labs at Murray Hill, “rank among the least architecturally distinguished”.

Consider the Cavendish, which moved into new premises on the outskirts of Cambridge in 1974. From the outside, it is a grey, grim functional building that looks like a series of giant Portakabins designed by a 10 year old on an Etch-a-Sketch. It is nowhere near as attractive as the Victorian edifice that it replaced, which was poky, cramped and unsuited to modern research. The new lab will never win any architectural awards, but it does its job well when measured by the quality of work that is performed there.

Indeed, the high standard of Cavendish research may be more a function of the quality of the people it attracts and the size of the grants it wins. Would a more attractive lab boost morale and prestige, and thus lead to even better research? Or does the down-atheel nature of the lab inspire a dogged team spirit among its researchers, a case of low living and high ideals?

The critical point

Thoughtful design can help a lab’s scientific mission both internally and externally, and in pragmatic and symbolic ways. Internally, a well-designed building can promote good management, morale and research by fostering collaborations and unexpected connections, and by allowing the lab to adapt to changing research patterns. Externally, architecture can promote a lab’s mission by attracting the recognition, support and pride of politicians as well as the scientific and surrounding communities — a classic case being the “high-rise” central building at Fermilab near Chicago (left). A dramatic building can even attract media attention. The Mesa Lab, for instance, appears in the Woody Allen movie Sleeper.

Yet there are also dangers. A lab’s architecture invariably embodies some interpretation of the scientific life, which often evolves in unexpected ways. “Science often changes,” Leslie told me, “but once it’s physically there, a building can only change so much. You have to be careful not to freeze into your architectural forms a particular way of doing science so that you wind up having to work against the building.”

The notion prevalent in the 1950s, for instance, that scientists were a special breed of people needing a special kind of space to work in a special kind of way turned out to be constraining when embedded literally in concrete. Today, laboratories are nodes in numerous networks — of computers, of other laboratories, of other projects — that a building’s architecture has to facilitate. Leslie’s work indicates that the challenge is to continue to allow scientific life to guide the architecture — and not the other way round.

Copyright © 2024 by IOP Publishing Ltd and individual contributors