US condensed-matter community grapples with availability of crystalline samples Crystal growing for physics measurements has fallen between the cracks in the US; without a turnaround, the country can’t help but lag in the discov- ery of new materials and their applications. The US “is a second-class, if not a third- what if Alcatel— The high-temperature superconductor yttrium barium copper class, citizen” in terms of investment which last year oxide is grown in crystal clusters; this one is about 2 cm from in the synthesis of high-temperature took over Bell’s end to end. Typical experiments use single platelets from such superconductors, heavy-fermion mate- parent company clusters. rials, thin films, single crystals, ultra- (see PHYSICS pure semiconductors, and other spe- TODAY, February RUIXING LIANG cialized samples for condensed-matter 2007, page 26)— experiments, says Cornell University’s pulls the plug? Séamus Davis. US scientists “have to go A smattering cap in hand to the people who lead the of crystal grow- development of new materials in these ers work in na- research fields.” Davis gets samples for tional labs and universi- his spectroscopic imaging scanning ties across the US, but in tunneling microscopy (STM) studies recent years, concern in the from colleagues in Japan, Canada, and condensed-matter community has the UK. “From the pure perspective of been rising about the availability of science,” he says, “things are great. It’s samples, a future generation of sample from the parochial perspective of how growers, and competitiveness in the ference between much belongs to the US that you may discovery and exploitation of new ma- Klaus von Klitzing’s think there is a problem.” terials. A National Academy of Sciences experiment and Horst Stormer and Dan With sample synthesis on the de- report exploring these and related is- Tsui’s experiment [which were awarded cline in the US over the past two sues is due out next year. Nobel Prizes in 1985 and 1998 for the decades or so, increasingly the US quantum Hall and fractional quantum condensed-matter community does The rub Hall effects, respectively] was the crys- think there is a problem. Art Ramirez, At the University of British Columbia, tal. Nothing else had changed.” director of device physics research Doug Bonn and colleagues have been For the 21st century, adds Davis, “we at Bell Labs, notes that Bell, IBM, tuning the high-temperature supercon- need electronic materials that are pat- and a few other companies led the ductor yttrium barium copper oxide for terned in different ways in real space. field after World War II. But around 18 years. Growing a sample can take a Many conventional materials—silicon, 1986, when the first observation of couple of months. “It’s never routine,” gold, aluminum—are translationally high-temperature superconductivity, he says. “If a crystal is dirty, the subtle invariant. Their electronic structure in by Georg Bednorz and Alex Müller in effects that test physicists’ theories are real space is completely boring.” What’s Switzerland, set off a rash of activity lost in the noise,” adds Loren Pfeiffer, more, he says, to extend Moore’s law— around the globe, “industrial invest- the crystal grower at Bell Labs whom the halving in size of electronic compo- ment in basic research began its rapid Eisenstein refers to. “The only dif- nents every two years—“we need com- decline,” Ramirez says. plex materials, for which the electronic “And no one has picked structure is incredibly complicated, and up the slack.” you have to be able to see it in real space The situation is reach- on the atomic scale before you can un- ing crisis proportions, derstand why it has the properties it ROBERT CAVA says Jim Eisenstein of does.” Making such materials, he adds, Caltech. The US, he adds, “is among the most challenging prob- has been the leader in un- lems in physics, certainly in materials covering the physics of physics.” two-dimensional elec- Still, it’s the measurements, not the tron systems, and “the sample synthesis, that are generally great majority of that recognized as “physics.” Says Pfeiffer, success involved sam- “To be considered a physicist, you have ples grown by one per- son at Bell Labs. It’s The floating zone crystal growth fur- unstable to have only nace is “arguably the best thing to hap- one individual at one pen to single-crystal growth in the past institution making ultra- 25 years,” according to Princeton Uni- high-purity semiconduc- tor crystals—like every- versity’s Robert Cava. The US, he adds, one else, he will someday doesn’t have enough of these machines retire.” Worse, he says, by a factor of 10. 26 August 2007 Physics Today © 2007 American Institute of Physics, S-0031-9228-0708-340-8 ROBERT CAVA Crystals (clockwise from left): Nickel olivine (the larger crystal is roughly 1 cm long); single crystals of the filled- skutterudite praseo- dymium ruthenium arsenide (the orange square is 1 mm on a side); and BRIAN MAPLE, UCSD a 4-cm-long specimen of the heavy-fermion superconductor uranium ruthenium silicide lightly doped with rhenium. (Image at upper right courtesy of UCSD’s Brian Maple and Zygmunt Henkie, Institute of Low Tempera- ture and Structure Research, Wroclaw, Poland.) to do measurements. Some say that work well for established scientists, crystal growth is not physics, but you says Robert Cava, a solid-state chemist cannot conceive of the right crystal who grows crystals at Princeton Uni- structure to grow without knowing the versity, “but what about experimental- physics. Therein lies the rub.” More- ists who can’t get their hands on sam- over, he adds, “very often, when I grow ples? Especially for young people, this something for someone, it can take is a problem. Getting good samples re- 20 iterations. The phase space is large: lies on nurturing contacts.” The density can be varied, the width of “When you make a crystal, often the quantum wells, . there is literally demand is exceedingly high,” says an infinity of possibilities. The crystal Pfeiffer. “Then you are in the tricky grower is an intimate part of the business of vetting colleagues.” Since collaboration.” growers of specialized samples tend to But getting tenure as a crystal work in collaborations—rather than as grower is pretty much a nonstarter. supply houses—they generally don’t “Crystal growers have hundreds of pa- give the same type of sample to differ- pers, but if you are talking about prizes, ent researchers doing similar measure- the notoriety goes to the people who ments. For example, Graeme Luke and make the measurements,” says Dale his colleagues at McMaster University Van Harlingen of the University of Illi- in Hamilton, Ontario, send crystalline nois at Urbana-Champaign. As a de- samples for magnetic susceptibility, partment chair, he adds, “If I can hire specific heat, thermal conductivity, op- one person, I may not choose to hire a tical conductivity, IR reflectance, and sample grower. He is going to grow other measurements to a dozen groups samples and give them away. There is a around the world. “We don’t send reluctance to hire them. It’s a risk.” many samples to neutron scattering or “Growing crystals takes a lot of in- NMR [nuclear magnetic resonance] frastructure. It’s difficult to support people in the US because we have good both a growing and measurement pro- people here in our group, and we don’t gram,” says physicist Brian Maple of send to their competitors,” says Luke. the University of California, San Diego. In STM, he adds, “Séamus [Davis] is the Maple, who is one of a handful of best in the world, and we give him sam- university-based crystal growers in the ples. But if we hired an STM person, US, makes and characterizes single that person would get some priority.” crystals and polycrystalline materials. Demand for samples is only likely to In chemistry and materials science, he increase as new facilities such as the says, most crystal growers’interests dif- Spallation Neutron Source at Oak fer from physicists’. Ridge National Laboratory ramp up. “Where are users of the SNS going to Sample bottleneck get samples?” asks Cava. “They are With samples hard to come by locally, going to scrounge them.” Adds Luke, US condensed-matter scientists often “You can have all the great advanced get them from abroad. “Many re- characterization facilities in the world. searchers become members in bigger But it’s worthless if you don’t have high- collaborations that are often run from quality samples to go in them.” And outside the US,” says Davis. This may when top-flight facilities come on line www.physicstoday.org August 2007 Physics Today 27 See www.pt.ims.ca/12309-15 in Japan, China, and elsewhere, US sci- LENA FITTING KOURKOUTIS, YASUSHI HOTTA, DAVID A. MULLER, AND HAROLD Y. HWANG entists will feel a squeeze. Low cultural barrier, high payoff In Japan, where the emphasis on mate- rials took off at about the same time that synthesis in the US began to dwindle, many labs are equipped with several furnaces for growing samples. “Japan started a big investment in science dur- ing the bubble economy in the 1990s,” says Hidenori Takagi, a materials physicist at the University of Tokyo and the RIKEN research institute. “And even after the bubble went away, dur- ing the recession the government strengthened our science.” Takagi notes An ultrahigh vacuum pulsed laser deposition chamber (left) at the University of that in Japan, the “mental, or cultural, Tokyo was used to create a superlattice, whose thin layers of lanthanum vana- barrier between chemistry and physics date and strontium titanate are bright and dark, respectively, in the scanning is surprisingly low compared with the transmission electron microscopy image (right) made at Cornell University.