54 nets and nothing fishy

Prof Batten tells a tale of why it’s always healthy to be skeptical!

What is it?

One of the most important qualities of a scientist is scepticism. The predisposition to doubt conclusions without solid supporting evidence is vital. Nowhere does this apply more than to your own work.

So when one of my colleagues requested my help with a structure that they thought contained 54 interpenetrating three-dimensional networks my initial reaction was that, in polite terms, they’d made a mistake somewhere and I’d need to find it. Yes, interpenetration of networks is quite common in framework structures – see, for example, the two interpenetrating diamond nets in the structures of zinc cyanide and cadmium cyanide, discussed previously on this blog. But the previous ‘record’ was only 18 nets, an exceptional number in itself. This was three times that – there was no way that so many independent 3D networks could pass through each other without bumping into themselves, nevermind the incredible self-assembly process that must happen for such a structure to form. It simply defied belief.

I was wrong.

The structure did, of course, contain 54 interpenetrating, independent networks. Each network was composed of silver atoms bridged by tri(4-imidazolylphenyl)amine ligands. The ligands bridged in two different ways – one was bound to three metals, while the other coordinated to only two. Furthermore, all the silver atoms connected to only two ligands, meaning that the branching points of the networks, the centres of the 3-connecting ligands, were interconnected by parts of two 3-connecting ligands, pairs of silver atoms, and a 2-connecting ligand. As the ligands themselves were quite large, the nodes were therefore an enormous 36.85 Å apart. This very large distance between the nodes meant that an individual net was extremely spacious, and gave the necessary room for 53 other networks to form and entangle with the first. So a few hours and one headache later I found myself confirming that yes, the structure did indeed have that many unconnected networks all tangled up together.

Another interesting feature of this structure was that the networks formed had the “(10,3)-a” topology. This network is of particular interest because it is chiral – i.e. there are two different versions of the net that are mirror images of each other (in the same way that your left and right hands are mirror images and different). Remarkably, nets of both “handedness” were present in this structure – 27 of each – to give an arrangement that was overall nonchiral (or racemic – see the tartaric acid blog post for an explanation of this applied to discrete molecules rather than infinite networks).

What does it look like?


Two of the 54 interpenetrating networks are shown schematically in the figure. Although distorted from the most symmetrical version of the (10,3)-a topology, the chirality of the nets can be seen in the rectangular spirals. Those of the blue net spiral into the page in an anti-clockwise fashion, while those of the red net spiral into the page in a clockwise fashion. The real structure, of course, squeezes another 52 nets into the space you see here.

Where did the structure come from?

“An Exceptional 54-Fold Interpenetrated Coordination Polymer with 103-srs Network Topology”, H. Wu, J. Yang, Z.-M. Su, S.R. Batten and J.-F. Ma, J. Am. Chem. Soc., 2011, 133, 11406-11409. DOI: dx.doi.org/10.1021/ja202303b



Pasteurized Crystals – Tartaric acid.

What is it?

December 27 marks the 192nd birthday of Louis Pasteur, which means that (a) he’d be really old if he hadn’t died in 1895, and (b) today is the perfect day to talk about tartaric acid.

Tartaric acid occurs naturally in many plants, particularly grapes. You’ve already read about ‘wine diamonds’ (potassium bitartrate), but you may not be aware of the contribution tartaric acid has made to scientific language.

Naturally occurring tartaric acid, first isolated in 1769, was found to rotate plane polarized light to the right. When it was prepared synthetically, tartaric acid had identical properties, except that it didn’t rotate plane polarized light. The synthetic material was thought to be a different compound, and was named racemic acid (from racemus, Latin for ‘a bunch of grapes’). It was subsequently determined that tartaric acid can exist in two different forms; and that naturally occurring tartaric acid was L-tartaric acid, while ‘racemic acid’ was actually an equal mixture of D and L-tartaric acid, mirror image isomers (enantiomers). These enantiomers were optically active in opposing directions, appearing optically inactive; this explained the otherwise identical properties of tartaric acid and racemic acid.

For this reason, ‘racemic’ came to mean ‘an equal mixture of enantiomers’, and this term continues to be ubiquitous in organic chemistry today.

So where does Pasteur fit into this story? Early in his career, before he discovered vaccination, microbial fermentation and invented the process which still bears his name (pasteurization), Pasteur studied crystals of tartaric acid and ‘paratartaric acid’ obtained from wine sediments. In particular, he wondered why (as described above) tartaric acid rotated light, while paratartaric acid did not, even though the chemistry and elemental composition of the two were identical. In one of the most beautiful and famous experiments in the history of science, Pasteur noticed, while squinting down a microscope, that there were two subtlety different types of crystals in the samples of paratartaric acid, each the mirror image of the other (see diagram below). He very carefully (and tediously) separated the two types of crystals into separate piles, redissolved each pile, and found that each did indeed rotate light, but in opposite directions. He had, in effect, separated the two enantiomers from the paratartaric acid (a.k.a. racemic acid) and discovered molecular chirality.

The two types of crystals found in paratartaric acid, which are mirror images of each other.

The two types of crystals found in paratartaric acid, which are mirror images of each other.

What does it look like?

The structure of D-tartaric acid (left) and its mirror image, L-tartaric acid (right).

The structure of D-tartaric acid (left) and its mirror image, L-tartaric acid (right).


Where did the structure come from?

D-tartaric acid can be found under CCDC refcode TARTAC, while L-tartaric acid is at CCDC refcode TARTAL.

Dedication in the face of adversity – Cadmium Cyanide

What does it look like?


What is it?

In April 1941 a Russian scientist by the name of GS Zhdanov published the structure of zinc cyanide. He followed up with a second paper in 1945 describing the structure of the cadmium analogue, which is isomorphous (has the same structure).

The structures consist of infinite arrays of metal atoms bridged by cyanide anions, in an arrangement similar to the carbon atoms in diamond. A beautiful feature of these structures, however, was that they contain two completely separate three dimensional networks that interpenetrate. That is, they passed through each other without ever actually being connected. Similar to links in a chain, except instead of being discrete loops passing through each other, they’re infinite networks. This could well have been the first examples of structures with interpenetrating networks, and it must have been puzzling at first for Zhdanov.

Another remarkable thing about this work is that April 1941 was just two months before Hitler launched Operation Barbarossa (the German invasion of the Soviet Union), and by 1945 the German invasion had been repulsed, and the Russians had returned the favour by marching on Berlin. So while most of Europe was ablaze, Zhdanov was making numerous careful measurements and hundreds of hours of hand calculations, working out the crystal structure of these two materials, which would have taken months if not years. In Moscow. In the middle of World War 2. Now that’s dedication.

Where did the structure come from?

“Crystalline structure of zinc cyanide”, G.S. Zhdanov, Doklady Akademii Nauk SSSR, 1941, 31, 352.

“The crystal structure of cyanides. II Structure of cadmium cyanide”, E.A. Shugam and G.S. Zhdanov, Acta Phys. URSS, 1945, 20, 247.

Interdigitation, Interpenetration, Intercalation

There is an old adage that “Nature abhors a vacuum”. This applies particularly to crystals. When crystals form the molecules try to use up as much of the volume in the crystal as possible. Even when the crystal is forced to form “porous” crystals, when they’re first made those pores are usually full of, at the very least, solvent molecules which must be forced out to access and use those pores.

In structures where the molecules form networks, the spaces within the crystal formed by those networks can be minimised through the (un?)holy trinity of crystal packing – interdigitation, interpenetration and intercalation. This is perhaps best illustrated by three structures which, chemically, are very closely related but display very different crystal packing in each case.

inter            All these three structures consist of copper atoms bridged by the small tricyanomethanide anion and another organic ligand, and all three form simple square grid two-dimensional networks. When the organic ligand is hexamethylenetetraamine, the 2D networks are quite corrugated and the anions project above and below the layers like bristles on a brush. So much so, in fact, that they poke into the holes of the adjoining network, and thus the nets interdigitate (literally meaning they have interlocked digits, much like your own digits do when you clasp your hands together).

When the ball-shaped hexamethylenetetraamine is replace by the rod-like 4,4’-bipyridine, the layers interpenetrate rather than interdigitate. This means that (in this case) there are two networks that pass through each other to form discrete layers. The pairs of sheets are not directly connected to each other but are nonetheless interlocked such that they couldn’t be separated without the breaking of bonds. Fortunately this is not the case with interdigitation, or otherwise we’d need surgery every time we clasped our hands together.

If the length of the ligand is extended slightly (by replacing 4,4’-bipyridine with 1,2-trans(4-pyridyl)ethene) the same square grid is formed, however this time the sheets stack to form channels, and the structure fills the extra space by trapping (intercalating) solvent and 1,2-trans(4-pyridyl)ethene molecules in those channels.

So, on the whole, crystals are quite clever things, with a range of tools in their arsenal when it comes to packing molecules together efficiently. So if we want to make crystals with lots of spaces inside, we need to be even clever…

Source: “Interdigitation, Interpenetration and Intercalation in Layered Cuprous Tricyanomethanide Derivatives”, S.R. Batten, B.F. Hoskins, R. Robson, Chem. Eur. J., 2000, 6, 156.

CCDC Deposition codes: 118844-118846


Iron trans-4,4’-azopyridine thiocyanate – Letting things down…

What does it look like? droop1



What is it?

A feature of coordination polymers (or metal-organic frameworks – MOFs) is that many are porous, allowing the passage of gases and solvent molecules in and out of holes within their structures. An interesting feature of the coordination polymer shown here (Fe2(azopyridine)(NCS)2) is that those molecules coming in and out of the structure illicit a change in the magnetic properties and colour of the material.

It happens through a process known as spin-crossover. Magnetism is caused by unpaired electrons in atoms. Electrons have a ‘spin’, which can be either ‘up’ or ‘down’, and usually like to pair up with another electron in various orbitals around an atom. The two electrons in each orbital have opposite spin and thus cancel each other out. In some atoms, however, there are more orbitals available than pairs of electrons, and thus the electrons have a choice of spreading out amongst the orbitals as much as they can, or forming as many pairs as they can. The former case leads to a maximum number of unpaired electrons (and is known as the high spin state), and the later case has the fewest number of unpaired electrons possible, and is known as the low spin state. In spin crossover the atoms can swap between the high spin and low spin states when the temperature is changed, pressure is applied, or the environment around that atom is changed.

In the example shown above, there are iron atoms that are bridged by long linear ligands. Importantly, the iron atoms also have two short thiocyanate ligands on opposite sides. When the coordination polymer is made, it crystallises with ethanol molecules in the lattice that cause the thiocyanate ligands to lay over somewhat. In this state the material swaps from high spin to low spin as it is cooled, and the colour changes. When the ethanol is removed from the lattice, however, the thiocyanates become ‘erect’, standing up and bonding straight-on. This and other resulting movements in the lattice changes the environment around the iron atoms enough to force them to stay high spin at all temperatures. Exposure of the material to ethanol vapour causes it to reabsorb the ethanol molecules, the thiocyanates flop over again, and it returns to the low spin state.

The fact that the presence or absence of ethanol in the structure has such a large effect on the thiocyanate geometry and hence the overall magnetic properties, particularly since they are not directly bonded to the iron atoms, is fascinating and has very interesting implications for tuning the guest responsiveness and signalling in this and other materials. It might also be the first example of “Brewer’s Droop” on a molecular scale…

Where did the structure come from?

Source: “Guest-Dependent Spin Crossover in a Nanoporous Molecular Framework Material”, G.J. Halder, C.J. Kepert, B. Moubaraki, K.S. Murray, J.D. Cashion, Science, 2002, 298, 1762. DOI: 10.1126/science.1075948

CCDC Deposition codes: 189340-189342

Cadmium tricyanomethanide tetramethoxyborate (try saying that three times!)

More tales from the Batten lab.

What does it look like?



What is it?

The best way to grow nice crystals is for the crystals to form slowly and undisturbed. This allows the molecules to assemble in the most regular fashion they can, with time to correct mistakes in the packing as the crystal grows.

Sometimes, however, crystals take much longer to grow than expected, and are found in crystal growing reactions that normally should have been abandoned and thrown away months beforehand. One such example is the rather complicated structure shown above.

The crystals of Cd(tcm)B(OMe)4.xMeOH took six months to grow, and seemed to appear almost overnight as beautiful and very large elongated octahedral. Note that they didn’t grow slowly and evenly over those six months, but rather the solution was completely devoid of crystals for almost all the time, and then suddenly surprise! Huge Crystals!

So what happened? It turns out that the chemistry is probably the key. The structure contains tetramethoxyborate anions, yet no such anions were added to the reaction. Rather, the boron in the anions comes from another anion that was added – tetraphenylborate. It seems that this anion slowly reacted with the methanol solvent to produce the new anion. The concentration new anion presumably built up over time until there was enough to start growing the crystals. And once it did start growing the crystals, well, it didn’t do it in half measures! They’re still some of the most spectacular crystals I’ve ever grown.

Where did the structure come from?

As for the structure itself, it contains a complicated coordination polymer that has small chiral pores running through it. Those pores contain helices of hydrogen bonded methanol molecules. The chiral nature of the coordination polymer, at that time, was very unusual, so the crystals were certainly worth the wait. Sometimes putting off cleaning up your lab bench of ‘failed’ crystallisation attempts can pay off!

Source: “Solvolysis of [B(C6H5)4] in methanol to give the chiral coordination polymer Cd(tcm)[B(OMe)4].xMeOH, x ≈ 1.6”, S.R. Batten, B.F. Hoskins and R. Robson, Angew. Chem. Int. Ed. Engl., 1997, 36, 636. DOI: 10.1002/anie.199706361


What’s in a name? The Octapi Catenane

Prof Batten tells us about one of the molecules that his group cooked up.

octapiChemists love to give names to things. Sometimes to explain a new concept or phenomenon, sometimes to make a discussion easier, and sometimes to just amuse ourselves and others. So when we discovered the structure shown here, we couldn’t resist. But more on that later. I need to give some background first.

The structure is a type of molecule called a catenane. It consists of two ring-shaped molecules that are locked together like rings in a chain. This unusual arrangement of molecules, while fascinating, does not happen by accident. Usually there are weak interactions between the molecules (much weaker than the bonds that define the molecules themselves) that encourage them to associate and form with this particular arrangement. In this case there are flat parts of the molecules called aromatic rings (look for the hexagons in the picture) that like to stack roughly parallel to each other. This is called π-stacking (so-named for the π electron clouds above and below the rings that facilitate this arrangement), and in this catenane there are eight aromatic rings, four from each molecule, that stack through the centre of the structure and stabilise this unusual entanglement of molecules.

So we decided to amuse ourselves by christening this interaction of the eight π (pi) rings stacked through the structure an “Octapi” interaction. We were then delighted when the prestigious journal Science took a shine to our structure, reporting it as an “Editors’s pick” of all the papers published that week. They even highlighted it with a picture. Of an octopus. Perhaps we outsmarted ourselves…


“Octapi Interactions: Self-Assembly of a Pd-Based [2]-Catenane Driven by Eight-Fold p-Interactions”, Jinzhen Lu, David R. Turner, Lindsay P. Harding, Lindsay T. Byrne, Murray V. Baker and Stuart R. Batten, J. Am. Chem. Soc., 2009, 131, 10372. DOI: 10.1021/ja9041912

See also Science, 2009, 325, 518 (DOI:10.1126/science.325_518d).

CCDC Deposition code: 731890