One thing that really irritates me is badly drawn ChemDraw structures (maybe I should get out more…).
Anyway, there is no excuse for a benzene ring not being a perfect hexagon – unless of course you’re showing perspective or drawing a fullerene or nanotube. Double and triple bonds should be well separated so that they don’t just look like thick bonds. And as for angles… unless there’s a really good reason for it not to be 120 or 180 degrees (depending on the hybridisation flavour of your [carbon] atoms), then there is likely something wrong; in flatland, even your sp3 centres should be 120 degrees unless you’re showing the stereochemistry and need to draw all four bonds. Atom labels should be in a consistent font (preferably Helvetica) and should be of a size that does not require one to use some kind of electron microscope to read them.
Simple really… so why do a lot of ChemDraw structures that appear in papers or on slides at conferences look like a six-year old has drawn them with a crayon (and I mean a particularly untalented six-year old at that – one that probably wouldn’t even win a Blue Peter drawing competition)? I think it’s just people being lazy…
I’ve been to a couple of conferences in the last month and gasped in horror (that may be a slight exaggeration) at some of the representations of C60 I’ve seen. Somewhat annoying is when the double bonds are drawn inside the pentagonal rings – please, can everyone just agree from now on to put them in the six-membered rings (surely it makes more sense that way?) . When I confronted one speaker about the ChemDraw crimes they had perpetrated, I was told that they’d just used the template… – and yes, it turns out that some of the C60 structures in the template library have dubious patterns of bonding…
Even worse, I see the occasional 5- and 6-bonded carbon atoms in fullerenes (and pyrene seems to fall into that trap as well). Maybe it shouldn’t get under my skin as much as it does, but this is how organic chemists communicate – and wrong is wrong. You wouldn’t use ‘2’ in an equation if it called for the square root of 2, so let’s be a bit more careful about our structures.
Right, soap-box carefully stowed away now for another day…
Stuart
Stuart Cantrill (Associate Editor, Nature Nanotechnology)
Yes, it’s a delicate balance between providing chemically incorrect information (too many bonds to a carbon, etc.) and just making the chemicals look terrible (even though the connectivity, etc., is correct). At Nat. Chem. Biol., we have recently created a chemistry style guide to deal with the badly drawn part of things, but the incorrectness remains an issue… (and since we’ve spent so much time thinking about what we want our molecules to look like, I am now also obsessed and notice all kinds of badly drawn molecules in talks and elsewhere). The bigger question we’re starting to face is that so much of the literature/internet/etc. is completely inconsistent: we have seen several cases where the structures given for a particular molecule in the published literature vs. PubChem and other depositories vs. online catalogs are all different, which of course propagates the mistakes for people who don’t or can’t get the literature references they need to check the true structure. How do we deal with that?
As a carbohydrate chemist I must say that the sugar structures drawn in your style guide look quite bad, specifically the equatorial bonds. For example when drawing glucose in the “chair” conformation the C4-OH bond should be drawn parallel to the C2-C3 and C5-O bonds, and the C5-C6 bond parallel to the O-C1, etc.
Thanks, Philippe! We were actually using the ChemDraw template for sugars, but I see what you mean! I’ll fix it in the Style Guide 2.0… (maybe these templates (and us for being lazy enough to use them), as Stu points out with the C60 situation, are to blame?)
Those interested in this subject should be aware that detailed recommendations will likely be published by IUPAC ‘real soon now’.
For more information and a draft of the poroposed IUPAC recommendations, see https://www.iupac.org/reports/provisional/abstract07/brecher_300607.html.
Although the official review period for that document ended last month, we might be able to consider additional comments as long as we receive them more or less immediately (say, by the end of this week).
Conveniently, the chemistry style guide referenced by Catherine above appears to already agree pretty much completely with the IUPAC proposed draft.
In my mind, one of the issues leading to poorly drawn structures is the inability/difficulty of getting ChemDraw-like programs to correct poorly drawn structures. I might be missing something, but in applying templates/styles, ChemDraw just resizes poorly-drawn structures, and never changes relative bond lengths or angles. If you have a bad bond angle, length, etc. in the structure, you can either a) erase & redraw, or b) don’t bother, and hope no one notices. Sure, one can argue that as professionals, we shouldn’t mind the effort of fixing badly drawn bonds, but imagine how much easier it would be if the programs we use would be able to take a lopsided benzene, and make it a true hexagon.
ISIS does that, if you set it to snap objects to the hexagonal grid. Not so convenient for five-membered rings, though.
For issues with specific programs, I’d strongly recommend contacting the producers of those programs directly (for ChemDraw, all sorts of contact methods are listed at https://www.cambridgesoft.com/contact/support/.) If you think that a program really ought to be able to do something, there’s a decent chance that someone else agreed with you, and that the software already can do you what you want. If the software cannot already do what you want, that’s a good sign that the software producers haven’t thought of your situation, and it’s unlikely that it will change in future versions unless you set them straight. These comments are not restricted to ChemDraw by any means.
The specific example of making a true hexagon is easy — that’s just simple math. The broader issue of making structures that ‘look pretty’ is, um, ‘somewhat more difficult’ in the general case…
I hope I’m not stating the obvious but using Chemdraw in the “Structure” menu there is a “Clean up structure” option which will usually tidy up badly drawn rings etc.
On a Mac the short cut is “shift-apple-K”
At least organic chemists have the opportunity to quickly and easily present well-drawn structures. Most programs also have a ‘clean up’ function that organises a structure.
Yet still, NO functionality exists for structures containing a metal ion to recognise the coordination number of a given metal ion (at a given oxidation state), OR to recognise that coordination bonds or not covalent bonds. It’s a bloody disgrace.