When the Greek philosopher Democritus went to work 2,500 years ago, he did not have access even to a low-power microscope to help him gather crude observational clues about what matter, in its smallest bits, was actually like. A thinker by trade, that did not stop him from building on a fine idea by his teacher Leucippus: The material buck of everything stops with indivisible particles that move around in an infinite void in which they engage in myriad temporary liaisons. The word for these diminutive pieces of matter—atoms—comes from the Greek word for “indivisible.”
Thinker that he was, Democritus also conjured physical details of these invisible atoms, and he used these to explain how people experience the world. He surmised, for example, that sweetness was because of atoms that are round, large, and kind to the tongue, whereas sourness was triggered by large, rough, and angular atoms that are tougher on the tongue.
Since John Dalton's resurrection and development of the atomic theory in the first years of the 19th century, chemists have been sketching presumed chemical structures and reactions in various line drawings, ball-and-stick models, and other stark representations. Simple as they were, these visualizations reflected profound leaps of the scientific imagination, says historian of science Alan J. Rocke of Case Western Reserve University.
“It was chemists who were the first to move beyond high-level philosophical speculation regarding the unseen microworld, pursued since antiquity, into the kind of productive, empirically founded and heuristically powerfully investigative programs that have since become routine,” Rocke writes in his forthcoming book on the role of mental imagery in 19th-century chemistry, to be published next spring by the University of Chicago Press.
Today, chemists have sophisticated tools; centuries’ worth of measurements, experiments, and observations; and a quantum mechanical theory of matter to help them visualize and simulate atoms, molecules, and their reactions and transformations. And with computers, they’ve become adept at generating simulations and animations of molecular processes occurring even in indiscernible fractions of nanoseconds.
Yet do the representations routinely deployed by today’s working chemists go all that much beyond Democritus’ visions from 2,500 years ago and those of the 19th-century developers of chemical structure theory? Flip the pages of a typical organic chemistry textbook, by which the newest generation of would-be chemists becomes indoctrinated with the discipline’s way of viewing chemical structure and dynamics, and the answer has to be, not so much. To be sure, ribbon diagrams of proteins and crystallography-based structural plots depict astonishing amounts of molecular information. Yet even these are static and cartoonlike.
“If we look at how chemists conceptualize the processes that they think are going on, I think that chemists—and physicists, too—are incredibly constrained,” says laser spectroscopist Geraldine L. Richmond, a chemistry professor at the University of Oregon. “I am always kind of stunned at the elementary images that we use to understand very complex systems.”
The representations have plenty to do with the intellectual saga known as chemistry that has been unfolding for centuries. That they have changed little despite the constant evolution of chemical knowledge raises the possibility that these simplistic images, models, and diagrams now hamper chemists’ ability to understand molecular phenomena more deeply. Perhaps representations that depict chemical reality more realistically by including, say, morphological fluctuations or the shifting of electronic densities in a reaction might be better catalysts of creativity. Or maybe the status quo is just right. After all, chemists continue to achieve great things.
To gauge whether working chemists find the standard representations of chemical structure wanting, C&EN asked members of the chemistry community to describe their own imaginings of molecules and reactions, the ones they conjure privately to help them access—in the way that Barbara McClintock, the great American geneticist, famously developed “a feel for the organism”—their own feel for the chemical realm.
With his lifetime of chemical theorizing and a book titled “Chemistry Imagined,” which includes lines such as “You have to put yourself into the life of the molecule,” Roald Hoffmann of Cornell University was the natural go-to guy.
“What is in our minds is not necessarily correct,” Hoffmann laments. But he suggests that mere mortals might be relegated to fallacious mental visualization of chemistry for the most fundamental of reasons. “There is a biopsychological evolution that has occurred by which we see things,” even invisible ones, in terms of familiar ones, he contends. This is behind “the billiard-ball model” as our default mental framework for chemistry-think, Hoffmann suggests.
“On top of that, we impose what we learn,” Hoffmann says. This is where space-filling models that mirror atomic volumes or the cloudlike representations of electrons’ locations in atoms and molecules that reflect quantum mechanical principles come in. “We have grafted onto the already three-dimensional graphic imagination of organic and inorganic chemists a roughly commensurately complex imagining of orbitals,” Hoffmann says.
Even so, these visual representations still look like variations of the line drawings from the 19th century, when the very concept of chemical structure was new. Even the textbook and computer representations of atomic and molecular orbitals come out as mostly smooth translucent spheres, teardrops, and other distorted balls.
There’s even more to be desired in chemical representation. Almost absent from the usually static depictions of molecules are anything but the most rudimentary indications of the maniacal motions of molecules—the various rotations and vibrations of their tiny selves and of their bonds, all on fantastically fast timescales. “If everything were frozen in the geometry of the cartoons that we draw, there would be no chemistry,” says David N. Beratan, a theoretical chemist at Duke University, where he studies, among other things, the ways electrons move in biological molecules such as DNA and proteins.
“Most of us understand things that take hours, minutes, or a few seconds, and most of us can even go so far as thinking about the world in milliseconds,” adds California Institute of Technology chemist Harry B. Gray, whose research crosses many categories of chemistry but centers on electron transfer. “When it comes to millionths, billionths, and trillionths of a second, and to picoseconds, femtoseconds, and attoseconds, you have to have a whole new way of thinking,” he continues, referring to the dynamics side of chemistry. In chemistry texts, dynamic phenomena often show up as little arrows, variously indicating molecular rotation, vibrational modes of bonds, or the redistribution of electronic bonding energy.
And then there is the tactile access to molecular structure and the input that that sensory mode adds to mental imagery. Students used to rely on physical, 3-D models to develop a kinesthetic feeling for structure, Hoffmann notes, but molecular simulations on computers have displaced these. “No one builds 3-D structures now,” Hoffmann observes, although he predicts that a wider availability of virtual-reality systems in the future will help bring the tactile mode more routinely back into molecular representation.
There is at least one arena of solid-state chemistry where Hoffmann, graduate student Robert F. Berger, and colleagues in Cornell’s chemistry department and elsewhere have been pushing into new representational territory. They have been looking at the complexity of families of intermetallic crystals with enormous unit crystals, the smallest structural motifs that define a crystalline solid’s overall structure.
“Most chemists’ knowledge of the solid state is based on the really simple crystal forms, like face-centered cubic and body-centered cubic, which have just a few atoms per unit cells,” Berger says. He, Hoffmann, and their coworkers have been looking at solids that have hundreds of atoms in their unit cells.
“We are challenged by being brought up in a world of three dimensions, and with these structures we scale beyond what our brains can handle,” Berger suggests. “When you have a 300-atom unit cell, thinking about each atomic orbital is a bit much.”
So his team has set out to make such structures more imaginable. Their tack has been to find ways of recasting these huge and unwieldy unit crystals as 3-D projections of more fundamental 4-D (or higher dimensional) structures, much in the way that a 2-D cross-section of an object is a slice of the object’s 3-D form.
Berger readily admits that adding spatial dimensions to the normal set of three doesn’t exactly conjure familiar imagery. “An 8-D image is not simpler than a picture I can see,” he says. But, he adds, “it is mathematically simpler to have an 8-D crystal that has just one atom in a unit cell than a 3-D crystal with hundreds.”
For the moment, the chemists view their hyperdimensional method merely as a heuristic for classifying and examining unusual solids, which may reflect little or not at all on the real features of the intermetallic compounds in their studies. On the other hand, notes Berger with a “what if” inflection in his voice, this new way of imaging structures could be pointing to previously unrecognized and hidden ways that electron density distributes and redistributes in space.
Richmond would resonate with the hopeful perspective that new modes of representation lead to chemical insight. She worries that sticking so universally, dogmatically, and relentlessly to the standard ways of representing chemical and material structure “confines our ability to open up new areas and possibilities.”
Not that Richmond has made it her business to invent new types of representations. That’s a risky path that she says would most likely attract the scorn of peers and could place at risk the future employability of chemists-in-training. “You could start looking like a wacko,” she says. But Richmond leverages existing models with metaphoric overlays that she hopes will help her students develop richer mental imagery about chemistry. “I often use liquid imagery when I am talking to students about reactivity,” she says by way of example. “I use the word ‘sloshing,’ so I might talk about bonding processes and polarity as ‘a sloshing of charge.’ ”
Richmond looks favorably on nanoscience as an arena that is bringing new visualization concepts into the picture. “It has expanded people’s thinking, because you no longer expect a molecule to be the center of attention, but you expect a structure to be there, and these act like big molecules,” she says. “This is making people think more broadly and to broaden the way they visualize macromolecular systems.”
Multiple representations can only help, agrees synthetic organic chemist Phil S. Baran of Scripps Research Institute. To shake loose insights, “draw a molecule as many ways as you can,” he says, echoing the time-tested advice of Robert B. Woodward, one of the giants in the history of synthetic chemistry. And more than that, he says, you need to keep in mind all of a molecule’s personality traits—its retention times in chromatography columns, the form it takes as a solid. The insight about how to synthesize a molecule “comes in an epiphany moment, which is an amalgamation of your knowledge about the physical properties of the molecule, its structure, and your empirical knowledge of its behavior,” Baran says.
There are good reasons that the simplistic representations in chemistry have remained so stable in form, Richmond and others contacted by C&EN say. “If I crawl in deeper in thinking about, say, charged species—and we know that it is not really a little ball with a positive or negative charge sign on it—then I have a hard time crawling back out and thinking about chemistry,” Richmond says. “It is wonderful, but it is a different mental space.”
Caltech’s Gray is not worried at all that chemical representations may bear little verisimilitude to what atoms and molecules are like on their own scales. The lack of such models “has not held anything back,” he says, contending even that digging deeper into the physical reality of such chemistry fundamentals as atoms and electric charge could be counterproductive for many chemists, particularly synthetic chemists who are in the business of making new and interesting molecules. “People who go off and think of various representations that may be more elaborate and closer to the truth will make that their occupation,” which is okay, he says. “It’s a question of what your goal is.”
Duke’s Beratan adds to this chorus by pointing out that there are ineluctable limitations, due to physics, that will forever relegate our mind’s eye to unrealistic visual analogies regarding chemical phenomena. “You can only resolve objects whose size is on the order of the wavelength of light, so the idea of seeing an atom is almost silly,” he says. “You can’t expect our conventional concepts about vision and seeing to be relevant on the scales where chemistry happens.” Even scanning-probe microscope images that are suggestive of atomic detail are derived by computers converting raw signals, such as tunneling currents or van der Waals force measurements, into pointillist maps depicting atomic arrangements in a visual format deceptively photographic in quality.
For his part, Beratan has been developing less detailed, “coarser grain” representations of proteins, in which, for example, each amino acid is treated as a single mass, as a way to get his arms around the behavior of these sometimes thousands-of-atoms monsters. He likens his representations to subway maps, which economically portray an urban transportation system’s connectivity. After all, Beratan says, that’s what is most important to the rider, not the literal geometry of the system.
Richard N. Zare, a laser spectroscopist at Stanford University and winner of the 2010 Priestley Medal, contends that chemists today “rely on a bewildering set of visualizations,” even if these don’t always show up on the printed page or the computer screen. For example, many chemists can imagine molecules’ fantastically fast rotations, vibrations, and other motions as the way in which molecules sample, in a vanishing instant or over a longer time, countless conformations, including the few that can actually participate in a particular reaction. This is one way, he says, that he gets, in a chemical context, that “feel for the organism” that McClintock referred to.
The imagery in dreams too can unlock pathways of discovery, says Zare, noting that a prerequisite to that mode of insight is to become obsessed with a problem. “If you become obsessed with something, you will hear things in what people say” and make novel connections based on what you read and see in papers or in your dreams, he says.
One way to loosen up the the imaginations of budding scientists is to make sure they are exposed to those in other disciplines who think with different sets of imagery and representations, Richmond says. “Just as any musician is enriched by being around other artists who dance or paint, it would be good for chemists to think together more often with physicists and biologists,” she says. Mixing it up this way, she says, could encourage more “improvisation” in the minds of young chemists and help them view the field as a creative science, rather than one in which all the language and all the ways of depicting phenomena already have been written in stone.
Even so, the success of chemistry does depend on some bedrock. “Chemists build a certain base of assumptions based on atomic theory, and then they build on that,” Gray says. “But then if you go back and ask a question like ‘What is electric charge?’ it kind of smacks you in the face. This will drop a chemist in his tracks.” For most chemists, the answer to this question doesn’t much matter. “I have no idea what electric charge is, but I know how to use it,” Gray says, recommending that those who want answers to such questions might wander into the physics department, or, perhaps for better results, the philosophy department, where they just might read about Democritus and Leucippus.