Synthetic chemistry machine

    This site uses cookies. By continuing to browse this site, you are agreeing to our Cookie Policy.

    • Synthetic chemistry machine

      I've been seeing this a few times on Facebook, a machine developed by Martin Burke et al at UIUC which is supposed to greatly simplify synthesis of a wide variety of small molecules (hype, not so hype, also not so hype). It appears to be based on something called Suzuki-Miyaura cross-coupling, which if I understand it, uses palladium as a catalyst for joining almost anything to almost anything else. More pretty chemistry stick figures here and here. I'm not a chemist by any means but would appreciate any further elucidation anybody cares to supply.

      The last nanotech conference I went to was almost a decade ago so I feel a bit out of touch with developments, but the two really promising fronts at that time seemed to be Chris Schafmeister's "molecular Lego" building blocks, which were small rigid molecules with (iirc) a pretty standardized set of bonding interfaces with one another, and Paul Rothemund's work on DNA origami, which has been recently used in cancer treatment. DNA origami has led to some interesting developments, but after ten years I'm a little surprised that neither of these fronts has yielded more dramatic advances than we see around us today.

      Back in the old days we used to hear complaints that software wasn't advancing quickly enough. These days there seems to be plenty of progress in software, although most of it seems narrowly focused on either web applications or the internet of things.

      So just being curious, was I over-optimistic in my assessment of the work of Schafmeister and Rothemund, or are we lagging in design software? And is Burke's work as revolutionary as some of the hype is making it out to be? When am I finally going to get my respirocytes and utility fog?
      == Resistance is futile. Capacitance is efficacious. ==
      Resistance is futile. Capacitance is efficacious.
    • I saw that news article earlier this week and at first thought it interesting but made the mistake of watching the 3 minute video they put out. It put me off of looking further into it - looked like hype without being specific. I wanted to see some claim of basic validation like, "We made the following molecules" and then show the chemical structures and/or models for them.

      I really appreciate the links you provided - the "End of Synthesis" was an interesting read - I haven't yet managed to plow through all the followups and fully comprehend all that is being said. I'm also out of touch with nanotech developments in general - I can't tell where things are at. I hope to correct that somewhat.

      So there is not a lot I can add, other than some speculation on the slow progress of DNA origami: technologies like that may provide arbitrary shaped structural building blocks, but what appears to be harder to provide are functional building blocks and a detailed plan to get to the final stage of assembly of the complete product.
       
    • Automated synthesis has been going on since the 1980's with the biggest bottlenecks being high yields at every step to minimize purification.  It's possible to build many building blocks with existing technologies but the conformational space will be limited by the available chemistry and/or the software needed to design flexible structures that thermodynamically adopt the desired conformation.  Overall, I vote for better software to direct synthesis.  Right now you almost always have to pair a computational chemist with an organic chemist to get a molecular design that is actually likely capable of being synthesized and will, maybe, have the properties you want.
      Martin
      Martin
    • I recently commented on this topic there:
      plus.google.com/+LifeboatFoundation/posts/V1ouSsTkA8C?hl=de
      I'll repeat that comment here slightly adjusted to fit the context of this discussion.

      I found these two open papers:
      informal overview: http://www.scs.illinois.edu/burke/files/highlights/pub32_Science.pdf
      details: scs.illinois.edu/burke/files/pubs/pub32.pdf
      (I've only read through the first one in detail yet.)

      Here's what this "molecule making machine" is about in essence:

      The the "Suzuki coupling" is used to make carbon-carbon bonds
      en.wikipedia.org/wiki/Suzuki_reaction
      like so: ...≡C-B(OH)2 + Br-C≡... - > ...≡C-C≡... + waste
      To continue after a first reaction one needs pre-delivered building block molecules that are at least double capped
      like so: Br-C≡...≡C-B(OH)2
      But this would self polymerize. That is it would form "infinite" chains. So the solution found was to temporally cap (chelate) the boron end with a molecule called "trivalent N-methyliminodiacetic acid (MIDA fro short)" like so:
      sigmaaldrich.com/chemistry/che…ights/mida-boronates.html
      For purification it was found that the MIDA caps on the growing product molecules can be locked/released to/from silica particles by different solvents.

      MARTINE hit the nail on the head with his comment.

      I'll put it in different words:

      Limitations that are obviously present:

      Since this is conventional chemistry and not mechanosynthesis the yields are low and the error rates are high. Every synthesis step can easily have losses in the single digit percentual range. Thus this is only suitable for small molecules (as the work says itself) and does not scale up very far.

      Limitations of this process that I'm not totally clear about are:

      Can this synthesis be done hierarchically or only serially? (Hierarchically would extend scalability a bit.)

      The minimal size and available shapes of the pre-delivered molecules (some shown in the more technical pdf) limit what can be made. The more informal pdf says that they can create loops – (How?). Given the shapes of the building block molecules sown in the technical pdf I highly doubt the loops can be made maximally tight and maximally close together. In other words I highly doubt polycyclic diamondoid cages can be synthesized (not to speak of strained ones).

      Out of this reason I doubt that this process can be used for the synthesis of some of the small tool-tips for diamondoid mechanosynthesis (like e.g. DC10c). (These tool-tip molecules would be small enough to not suffer too much from the yield decline problem of conventional non mechanosynthetic chemistry.)

      In light of "early APM" I suspect this technology could still could be useful for:
      * The synthesis of light activated "motor molecules"
      * The synthesis of novel side chains for foldamers (for boosting stiffness / symmetry / covalent cross linking / ...)
      * ...

      With early APM I'm referring to "Coarse-block APM systems" page 33:
      energy.gov/sites/prod/files/20…ntation%20-%20Drexler.pdf
      And "modular molecular composite nanosystems" MMCNs
      (are these two deemed identical by Drexler ??)

      I'm NOT referring to what is shown in the INFAPM workshop video!
      That is (as I currently see it) just about the "System-level tech demos" dead end in the diagram (also page 33).

      ----

      Automated synthesis will by now certainly have gone a long way from where it was in the 1980's, but yes it's still on the very beginning (accelerating progress here). I think cheap 3D-printed microfluidics has the potential to give this area (and the area of foldamer nanosystems) a major boost.
    • Saw this remotely related announcement today:
      Scientists create world’s first ‘molecular robot’ capable of building molecules

      The original article in Nature:
      Stereodivergent synthesis with a programmable molecular machine

      I say "remotely related" because it also looks to use a manufacturing system in a solution that can synthesize some kinds of molecules. But it appears to use a chemically positionable "arm" which would put it in a different class of machines/techniques than the previously mentioned approaches.
    • Yeah, I've got info about this via the lifeboat foundation blog's google+ channel.
      Here: (A)
      # plus.google.com/+LifeboatFoundation/posts/7tAm7K4KVLn
      # lifeboat.com/blog/2017/09/scie…e.com&utm_campaign=buffer
      # scienmag.com/scientists-create…le-of-building-molecules/
      # nature.com/nature/journal/v549/n7672/full/nature23677.html
      But your link has some info-graphics.

      A bit earlier another remotely related thing came in there too.
      This article here: (B)
      # plus.google.com/+LifeboatFoundation/posts/RtY1nQDJjwC
      # m.phys.org/news/2017-09-dna-na…s-predefined-regions.html
      # science.sciencemag.org/content/357/6356/eaan6558

      I'm a bit upset about the authors of the article (A).

      Have you read the very first sentence of their abstract here:
      nature.com/nature/journal/v549…html?foxtrotcallback=true
      and checked their leading three references?

      Citation: "It has been convincingly argued 1, 2, 3 that molecular machines that manipulate individual atoms, or highly reactive clusters of atoms, with Ångström precision are unlikely to be realized."

      NNGNNH! ... Time for analysis why this happened ... again.

      The fact that we still cite these minus sign bearing references suggests that we may still lack knowledge about both:
      # the "modern" far term goal of gemstone based nanofactories and ...
      # the incremental-path towards them.

      (Side-note: I use "we" as in "majority" here since the authors of A are not the only ones citing these references).

      Or it may suggest, that we may ruthlessly dismiss all those ideas that are not viewed by a large group as bringing (with high likelihood) economic returns in our own lifetimes. (Side-note: Personally I'm agnostic in regards to this last point.)

      The work done in (A) (one of the R&D fronts of soft nanomachines) is in some respects (solution phase thermally driven assembly) relatively near to the incremental path. In some other respects very far though (utterly different far term goal - hard/stiff nanomachines ve eternal limitation to soft/compliant nanomachines)! Nonetheless (or exactly because of that) the recent amazing progress that was made in climbing the "stiffness ladder" in "hinged foldamer nanomechanics" seems to have flown totally under the radar.

      I guess we are still standing way to close in front of the "bark" of the "soft nanomachine tree".

      The R&D fronts of soft nanomachines are:
      # by basically everyone well known for not leading up to advanced gemstone based APM and ...
      # by many again and again confused (or claimed equal out of fear**) with the incremental path.

      The incremental path is utterly different in its treatment of soft nanomachines as just a means to get away from those soft nanomachines ASAP and thus actually does lead to advanced APM systems.

      **Part of the problem might also be conformist group-think where we try to keep our own reputation maximally safe by aggressively marginalizing our colleagues. Even if we personally would not aggressively disagree.

      (Sorry for ranting.)

      The post was edited 2 times, last by lsuess ().