Skip to main content

Chemical News

The One Source of Perfect Crystals

Unless you’re really into graphene (or other two-dimensional advanced materials) you’ve probably never heard of these guys. Takashi Tanaguchi and Kenji Watanabe at Tsukuba’s National Institute of Materials Science are basically the only source in the world for high-quality crystals of hexagonal boron nitride (hBN), and that is apparently the idea substrate for studying all sorts of other two-dimensional sheets that can lie on top of it. hBN itself is not a rare substance – it’s found in all sorts of high-end lubricants, and it shows up in cosmetics (eyeliners especially), printers, electronics, dental cements, pencils and who knows what else. But growing large defect-free single crystals of it is another matter entirely. That takes a gigantic high-pressure setup like the one in Tsukuba, and that device is apparently given over now almost entirely to turning these out. It’s not even what they had ever planned to do:

Neither Taniguchi nor Watanabe is a graphene researcher, and they had no idea that their gems would become so desirable. The researchers now have several patents related to their hBN-making process, but say they don’t expect to be able to commercialize it — at the moment, only research groups need the highest-purity crystals. There is a sizeable perk, however. Because the pair are credited with authorship on studies using their crys- tals, they have become among the world’s most-published researchers. Together, Taniguchi and Watanabe appeared as authors on 180 papers last year — and, since 2011, they have co-authored 52 papers in Science and Nature, making them the most prolific researchers in these journals over the past 8 years. . .

There are other groups and facilities working on increasing the supply, but as that article points out, researchers are going to have to run head-to-head tests before they’ll trust any other suppliers. Single-sheet graphene will conform to the shape of whatever it’s sitting on, and if it’s not absolutely flat many of its unusual properties disappear. (Note that getting really good graphene to start with is not always so easy unless you’re isolating it yourself). The substrate also has to be extremely nonreactive and pure, so as to reduce the possibility of interaction with the electrons in the delocalized graphene orbitals. And hBN checks all these boxes, as became apparent around 2009. The first work using Tanaguchi and Watanbe’s crystals as a key component showed up in 2010 and seems to have caused a frenzy in the graphene field when people realized just how good they were. Like all crystal-growing, it’s full of witchcraft:

Taniguchi is cagey about the exact recipe: this is his secret sauce, and he likes to change the composition of the barium layer from batch to batch. “Using the same recipe every time is not that fun,” he says. For first-time users, he’ll send some baseline crystals, but with long-time users, he wants feedback on each slight change to the process. By measuring electron mobility in graphene, they can detect impurities in the underlying hBN with more sensitivity than Taniguchi and Watanabe can measure. At first, no one had any complaints about their crystals. Only in the past two years, Taniguchi says, have researchers begun reporting impurities that affect their results — a result of them pushing the limits of the material. And that motivates Taniguchi to improve. “I’m a crystal grower,” he says proudly.

The whole field has been advanced tremendously by these crystals, but at the same time, you have to worry a bit when it all depends on these two dudes in Japan, only one of whom knows all the subtle details and is approaching retirement age, to boot. There are those other suppliers tooling up (and other methods as well, from similar high pressure work to chemical vapor deposition), but for now, it’s the Tsukuba lab or bust. I’ve heard of situations like this in molecular and cell biology, where one particular antibody or cell line is available from this one particular lab (or maybe not!), but this is the first time I’ve come across quite this situation in chemistry/materials science. Anyone have any comparable examples? I’ll bet they don’t involve 180 publications a year!

26 comments on “The One Source of Perfect Crystals”

  1. MattF says:

    The most famous example of a guy with a secret sauce was the gentleman who polished the standard kilogram. Obsolete for about a year now, but before the new standard, numerical values of masses could lurch up or down without warning.

  2. Martin says:

    “Taniguchi is cagey about the exact recipe: this is his secret sauce, and he likes to change the composition of the barium layer from batch to batch.”

    For me this is exact antithesis of science. All the research he produced is practically irreproducible for all others.

    1. Isidore says:

      In many ways this is no different from using reagents or other materials available from a single vendor. I use enzymes and other materials for protein characterization, for some of which there is a single vendor or only one whose quality is satisfactory. Of course, in these cases the vendors do not get authorship in publications.

      1. Nameless says:

        In some cases this is related to patents, not black magic.

  3. Haftime says:

    Crystal growers fulfil an unusual but essential role for condensed matter physicists. This is particularly true for neutron scatterers who might need ~10g pure single crystals in order to carry out neutron spectroscopy experiments capable of detecting the exotic quasiparticles that exist in solids (e.g. magnetic monopoles, spinons etc.).
    Some photos of the kind of crystals needed for these experiments: (this is from Prabhakaran at Oxford)

  4. Earl Boebert says:

    According a metallurgist of my acquaintance, this is the way metallurgy worked well into the second half of the 20th century. A pinch of this, a dab of that, everybody knew that Sears wrenches almost never broke and nobody knew why.

    1. aairfccha says:

      To some extent it still does, especially when you want to cast something at the edge of what is possible.

  5. luysii says:

    “Anyone have any comparable examples?”

    Back in the day, I was told that the old German chemists of the 19th and early 20th century were so great because their beards were full of small crystals that no one else could crystallize

  6. anon says:

    It’s not at all unusual that there’s only one facility/instrument in the world that can do some particular measurement (for example CERN, LIGO/VIRGO, certain telescopes, etc.), but in big science it’s obviously not just one or two persons that control it all.

  7. Anonymous says:

    “For first-time users, he’ll send some baseline crystals, but with long-time users, he wants feedback on each slight change to the process.” – In organic synthesis, do we ever run reactions exactly the same way? Scale up changes heat exchange; syringe to graduated cylinder errors; etc.. We keep track in our head of what might have been important to improve — or ruin — a reaction. W&T want feedback from users to know what is an improvement and what makes things worse.

    In org synth, catalyst prep can vary from lab to lab and result in actives, bad actives, and inactives. Lindlar is one example. It works for some but not for others.

    “There is literature and there is lore.” There was a previous (2018? 2019?) Pipeline on stirring rates and methods affecting reaction outcomes. For some reactants, you must use Aldrich and not Acros; for others you must use Acros and not Aldrich. And so on.

    I see 3 possible motivations: (1) patent / $ protection (2) improve the prep with user feedback AND (3) avoid being accused of publishing methods that don’t work or worse. In org synth, you sometimes have to try several times to repeat a published prep. Same thing for this fancy hBN prep. If people set up a vacuum system and don’t know the “lore” of getting everything just right, they might trash W&T for publishing difficult to reproduce methods.

    1. David says:

      Exactly. Given the “voodoo” surrounding this type of synthesis, I find it eminently plausible that even they don’t know which details are essential and which are optional for reproducibility. Can you use plastic measuring cylinders, glass, etched, or inked increments? etc etc etc

  8. Diver Dude says:

    It’s worth saying that the oriniginal article says that Taniguchi is looking for a protégé to train up but hasn’t found one yet. So, if you’d like 200 publications a year, a job for life and living in Japan, form an orderly line. FWIW, I used to work in Tsukuba and it’s very, very nice indeed.

  9. Dave says:

    In some ways, the hBN process reminds me of the development of the hydrothermal process for growing Quartz crystals. The temperatures and pressures involved are scary. And, there are some steps required to avoid twinning of the crystals being produced.

  10. Barry says:

    And famously, getting the “right” result sometimes depends on imperfections. That was the story of the Kishi-Nozaki coupling. Happily, the irreproducibility of the reaction was explored before the “good” bottle of CrCl2 was exhausted, and the critical “contaminant” of Nickel was characterized.

    1. loupgarous says:

      The US Department of Energy failed to fully document the manufacture of the thermonuclear weapon component “FOGBANK” down before decommissioning the part of the Y-12 facility at Oak Ridge where it was made. When it came time to refurbish the thermonuclear weapons that require it for proper conditions in the interstage between primary and secondary components, the new FOGBANK facility’s product wouldn’t do the job properly.

      DoE spent about $90 million on the new plant ($23M) and additional analysis and research (~$69M) trying to make a new batch of FOGBANK that worked, before they discovered that undocumented “impurity” required for several US thermonuclear weapon designs to work as designed.

      According to a Los Alamos newsletter article, the new plant was too clean, and removed something the old plant inadvertently allowed to enter the mix which was required to create FOGBANK that worked well.

      1. Uncle R says:

        Thank you loupgarous. Oh what a gem that Los Alamos newsletter is, although no doubt Maj Gen Groves will be spinning in his catacomb at mass distribution of what gave 60-odd years later at his TOP SECRET facility in the New Mexico desert. Hopefully up in the Great Nuclear Physics Lab in the Sky Oppie is a bit more relaxed about it all…

        1. loupgarous says:

          Oppie was the guy who said there are no defensible secrets to making nuclear weapons, and as the membership of the “nuclear club” enters double digits, his judgment is vindicated.

          Richard Rhodes’ masterful histories The Making of the Atomic Bomb and Dark Sun: The Making of the Hydrogen Bomb are setting Leslie Groves’ corpse spinning in its grave. There are fine points to the manufacture of fission and thermonuclear weapons Rhodes wasn’t cleared for, but the budding nuclear weaponeers of, say, Myanmar, Bangladesh, Venezuela and Qatar have all the broad hints they could wish for in those books.

          Even without them, coy Mother Nature doesn’t hide her secrets as much as tease her admirers. Soon, India and China’s own “bees of Hell” will surpass us in destructive ingenuity, because those nations are now entering the phase we did in the 20th century, several generations with the leisure to pry Nature’s locks open and national governments wealthy enough to finance the effort.

          We could hope there’s nothing in humanity’s future like the thousand-fold leap in destructive power between Los Alamos’s first efforts and the fifteen-megaton “Shrimp” thermonuclear device detonated nine years later – the first truly deliverable hydrogen bomb.

          But it’s a vain hope. CERN and other places are making and storing antimatter, albeit anti-atom by anti-atom. Just as the Curies sifted through tons of pitchblende for radium, polonium, and the other signposts toward nuclear fission.

          Elon Musk is right. We need humans on more than one planet than our home, and he’s very determined to put them on Mars and as many other places as possible, because our ability to govern ourselves hasn’t kept pace with our advancing ability to kill our own species off.

  11. gcc says:

    I remember reading (I think in the Molecular Cloning manual) that there was a time when Doug Hanahan developed a method of making competent cells that resulted in plasmid transformation efficiencies much higher than anyone else could achieve. Apparently he was happy to distribute the key buffer to people who could justify the need for very high efficiencies, but was reluctant to reveal its composition. It used to be referred to by some as “liquid gold” because it was so valued and because it happened to be gold in colour due to the presence of hexamine cobalt (III).

    Hanahan eventually published the protocol, but even then, people often had difficulty getting efficiencies as high as he could because there were some things that were important that weren’t emphasized in the protocol, such as the requirement for extremely clean glassware free from traces of detergent. I vaguely remember making competent cells myself (once the details were all public knowledge) and our lab used to have glassware that was designated specifically for that purpose.

  12. Paul Brookes says:

    Best example of this in the cell culture field was HL-1 cells, a cardiac cell line that was not available from ATCC and could only be obtained from William Claycomb. You had to use his expensive magic pixie-dust “Claycomb Media” to grow them too. You’ll note that the above sentences are written in the past tense…

  13. Peter S. Shenkin says:

    For a little while I did some experimental work in the rubber business. I learned a bit about it in Akron and did some work at the rubber lab at the University of Akron. I was working for a large company at the time.

    I was learning about rubber formulations for tires. About 1/3 carbon black, 1/3 oil, and 1/3 rubber. Then there were the trace ingredients: mainly anti-oxidants and vulcanizing agents with names like “methyl tuads”.

    So, when going down the list of ingredients for a “standard” tire formulation, I’d ask, “What’s this?” about 10 times. I’d usually try to get down to an explanation that made sense. I stopped trying to do that after the following conversation:

    Me: What’s X?
    Him: Oh, that’s an antioxidant.
    Me: Well, what’s Y?
    Him: Oh, that’s another antioxidant.
    Me: Why two antioxidants?
    Him: Well, we always used X, but some years ago we discovered that Y worked better.
    Me: Well, why do you still use X, then?
    Him: Well, we do accelerated testing, of course, but it’s not perfect. Tires sometimes last 10 years or even longer. We’re afraid to take anything out.

    I guess he was right. If a tire deteriorates prematurely, they could be sued, because “Everyone knows you need X in a standard tire formulation, and you left it out!”

    1. Derek Lowe says:

      Have you come across the story Primo Levi told in “The Periodic Table” about adjusting the pH in a batch of substandard paint at a factory? A very similar situation. . .

      1. Anonymous says:

        (The Chapter on “Chromium.” Required reading for my junior colleagues.)

  14. Conrad Vink says:

    It’s unethical not to reveal the method when your work is funded by taxpayers. The justification of government funding is to allow humanity to discover new knowledge, not to discover new knowledge and then keep it to yourself.

    1. Mike D. says:

      “Using hBN crystals obtained from Tsukuba as a substrate, we proceeded to…”

      That’s enough information to reproduce the process, at least over the duration of the contract.

      The government requires a lot of stuff to be submitted in Microsoft Office format files, but that doesn’t mean I have to put a link to the Office source code in my final report.

  15. Picky picky picky, if you pick it, it won't better says:

    “the idea substrate” -> “the ideal substrate”

  16. gippgig says:

    From the Sept. 2 issue of Bloomberg Businessweek(!), which is a special issue on the elements (recommended), under Ga-Ge-As: In the middle of 1949, Shockley’s group developed the junction transistor, which relied on germanium crystals to work best. But his polycrystalline germanium wasn’t doing the trick. Like Walter White in Breaking Bad, Teal was making a product so much purer than anyone else’s that it seemed almost unthinkable.

Comments are closed.