Skip to Content

Asphalt, Up Close and Personal

I wrote not long ago about another use of atomic force microscopy images to determine chemical structures, and now here’s another paper on the same general topic. Pretty soon, this is going to become too common to note, and structure determination will have changed forever (once again!)


Shown are some asphaltenes – and if you’ve never had to think about asphalt from a chemical point of view, maybe you’re lucky. It’s a real brew of high-molecular-weight hydrocarbons and heterocycles, and these are just a few of them. There’s been a lot of analytical chemistry done on these systems over the years (there are more compounds in there to separate than you’d want to think about), but even once you’ve separated them, figuring out the structures is what the folks at the bench call “nontrivial”. You can get a molecular weight by blasting away with the mass spec machines, but there are a lot of possible structures per molecular weight up in that range, and these kinds of molecules tend to look very, very similar by NMR and optical spectroscopy. (Here’s a recent review).

These things are perfect candidates, though, for AFM. The authors have taken an asphalt mixture and scattered it onto a surface of sodium chloride crystal layered onto copper, and then just scanned around the surface, picking out one molecule after another. Three representatives are shown; the paper has over a hundred in total. Comparing material from different sources also shows different patterns in the structural classes, as you might expect, opening up a whole new area for fingerprinting such samples. I’d guess that almost all the molecules in this paper are new to science, in the sense of having been described as separate entities (you could draw such stuff on a whiteboard all day long, if you were so minded). And they look even flatter and less soluble than the drug candidates I’ve worked on over the years, which (in a few cases) is really saying something!

23 comments on “Asphalt, Up Close and Personal”

  1. Anonymous BMS Researcher says:

    Good to see your blog’s new home. When I saw the title I remembered a bicycle accident of my youth in which I got up close and personal with asphalt in a rather painful manner.

  2. Slurpy says:

    CA7 looks like some coronene derivative, it seems to be not very two-dimensional at all.

  3. Semichemist says:

    Cool! Are asphalt constituents usually these huge, flat heterocycles? No other organic or 3D compounds?

    Also been wondering – what do 3D molecules look like under AFM? Surely this method is a bit limited by being restricted to 2D molecules

    1. IBM Research says:

      Our approach is definitely most effective for nearly flat molecules. But also 3D molecules can be imaged, as for example a round buckyball [see In this case, however, one can only access the the top most carbon ring. Usually the molecules also get flattened when adsorbed on the surface or one could use atomic manipulation to change the molecule conformation and reveal different parts of the molecule, which helps to make the method suitable for a larger class of molecules. – Bruno Schuler, IBM scientist

    2. IBM Research says:

      Our approach is definitely most effective for nearly flat molecules. But also 3D molecules can be imaged, as for example a round buckyball [see In this case, however, one can only access the the top most carbon ring. Usually the molecules also get flattened when adsorbed on the surface or one could use atomic manipulation to change the molecule conformation and reveal different parts of the molecule, which helps to make the method suitable for a larger class of molecules – Bruno Schuler, IBM scientist

  4. And then there the polymers, gravel, etc that get added…

  5. Hap says:

    I think the asphaltenes are just the less well-characterized portion of asphalt – they tend to be high molecular weight (?) and not very volatile, so they’re difficult to detect by MS and hard to separate, I think. There’s other stuff, but I assume it’s better characterized. has a little more information.

  6. KazooChemist says:

    @Anonymous BMS Researcher

    Your comment brought several painful memories to mind but also a humorous one. When I was a kid a bunch of us were riding our bikes around a long curve when we came upon a crew “chip-sealing” the road. They had covered the entire surface with liquid asphalt, but had not yet spread the stone. The lead biker thought it was water. He yelled “let’s wash our tires” and hit the slippery curve at high speed. Of course, his bike slid out from under him and he rolled over several times in the tar. Quite a sight! It is amazing to me that we survived our childhood.

  7. Radiochemist says:

    I remember a friend asking me “so these compounds you make how do you know how you made it? I guess some sort of microscope? what do they look like?” and I laughed a bit before saying “no” maybe she was just 20 years ahead of her time!

  8. pete says:

    Q: Why did the AFM cross the road ?
    A1: To look for the asphaltenes.
    A2: It was tied to a chicken.

  9. eugene says:

    It’s a little limited in structural determination yet. The molecule has to be mostly flat and in an isolated five membered ring, they can’t tell the difference between CH2, CH, OH, S, N, or NH… which doesn’t bode well for element identification with this technique. However, in concert with a mass spec technique and if you had the material isolated, it could solve some ambiguities in structure.

  10. Jason says:

    This paper has some awesome images, but they’re very weird to me. You look at their STM images and they look sort of like what I’ve been trained to expect in all my courses ( sort of a cumulonimbus cloud lumpy blob of molecular orbitals. But in the AFM you get these nice fine lines for bonds (Figure 1C zoom in real close). The bright parts on the AFM are repulsive areas so I would have thought all the orbitals would be pretty repulsive, not just a fine line.

    Can any sharper minds explain this to me? Their resolution seems quite good so I don’t think it’s just an artifact of low res.

    1. IBM Research says:

      The contrast mechanisms of STM and AFM are fundamentally different. With STM you can access the electronic structure at the Fermi level that is in conjugated molecules inherently distributed about the entire molecule. AFM on the other hand measures the force resulting from all electronic states. The repulsive interaction (bright contrast) that is sensed stems from so-called Pauli repulsion that is the repulsive interaction when the electron clouds of the molecule on the surface and the tip try to keep out of their ways. The contrast is therefore closely related to the total electron density which exponentially decays in space and therefore provides this high resolution.

  11. Shion Arita says:

    I’m not very familiar with this technique, so I’m going to ask a (maybe) trivial question:

    Why do the rings look all warped?

    1. IBM Research says:

      Indeed, the structures in the molecules are distorted. The reason for that is the flexibility of the CO molecule at the tip, which bends differently depending on the position above the molecule. On the one hand this is beneficial because it increases the lateral contrast by sharpening the bonds that could even be used to distinguish bond orders in a molecule []. On the other hand, the tilting of the CO at the tip distorts the structure.

  12. alig says:

    I found the answer to my question about QD to BID. It varies by indication from 0 to 30 % better compliance with QD. It seems the more life threatening a disease the better overall compliance. See:

  13. PorkPieHat says:

    @Shion Arita, your question about the rings looking warped….wonder if it isn’t because the macro-shape of each of those polycyclic molecules is likely to be cupped and very 3D in its morphology.

  14. Derek Lowe says:

    I think another reason for the warped look of some of the molecules is that the single-atom tip of the probe is at an angle, which can presumably change as the scanning process goes on.

  15. Today on ask a physical chemist says:

    The STM/AFM questions brings up a related one I’ve always wondered about X-ray diffraction that someone might know – if the origin of the X-ray scattering is interaction with electrons why do X-ray solutions of eg. arenes show up as the individual atoms rather than the delocalized orbital? NB: This question may be built upon a number of misconceptions

  16. Derek Lowe says:

    “Ask a physical chemist”, eh? Well, I’m merely an organic chemist, but the closest I can come to answering your question is that I believe that the cross section for elastic X-ray scattering, which is what crystallography depends on (Bragg’s law and all that) is very, very much higher for bound electrons as opposed to delocalized ones in molecular orbitals. So those are effectively invisible by this technique. A real crystallographer may show up and vet this explanation, though.

  17. Anon electrochemist says:

    The scattering cross section of any individual electron is approximately the same, because the electron classical velocities in atoms are roughly equivalent. Therefore X-ray scattering basically corresponds to electron density under normal condition. Extreme effort has been expended to specifically distinguish atoms by their scattering, because this allows simple structure determination of proteins! (MAD phasing)

    There are simply more electrons in core shells than there are in valence shells, and those core shells are tightly localized. The small molecule community maps delocalized “thermal ellipsoids” instead of point charges to reflect this small degree of diffuse character and directional uncertainty in atomic coordinates. For macromolecules the structural resolution is simply insufficient for models to capture this, and point nuclei are used for convenience.

  18. Today on ask a physical chemist says:

    I suspected the relative densities of electrons could be involved, thanks

  19. Actually a physicist says:

    As someone who works in the asphaltene field: the consensus is that these polyaromatic rings do form a (relatively) flat core even in solution, although there is still some debate on whether all of them form a big core in the middle or if there are several small cores linked together. The main part that these (super nice) imaging techniques miss out on is the alkane tails scattered around (and some substitutions of N,S,P etc.). The images in the blog post above are from coal asphaltenes which have shorter and fewer tails than oil asphaltenes (which is where the main research interest / funding is found).

    I see Derek says these were adsorbed onto NaCl on Cu, but I believe the IBM guys said at Petrophase that most of them were on plain (111) Cu? (Or my memory isn’t up to scratch.)

Comments are closed.