From potty idea to Nature papers

University of Groningen chemistry professor Kees Hummelen had a potty idea: he believed it would be possible to use single molecules as integrated circuits. The idea resulted in two PhD theses and papers in high-profile journals, the latest of which appeared in Nature Nanotechnology last week.

The Nature paper made the headlines last week. It described how an essential circuit in mobile phones could be recreated in a single molecule . Physicists from Delft University of Technology were responsible for most of the work, but the molecule they used was designed and synthesized in Groningen.

‘Our friends in Delft have done a great job in showing that there is strong negative differential conductance in this molecule. This is very rare, especially on a molecular scale, and so far can only be commercially achieved in electronic circuits, with different components’, says Hummelen. ‘But the exciting thing for me is not that this might pave the way for molecular electronics in the distant future, but rather that it shows I was right, as were my PhD students.’

To explain his enthusiasm, Hummelen has to go back more than ten years in time. ‘I had this potty idea that it might be possible to predict the conductance of single organic molecules from the distribution of single and double bonds between carbon atoms.’ If conductance was predictable, it would then be possible to design molecules that would act as integrated circuits.

Hummelen gets up and walks to his bookcase. ‘Here’s the PhD thesis that describes it all.’ Marleen van der Veen took the ‘potty idea’ and produced a thesis full of molecular designs and predicted conduction pathways. Such pathways through a single complex organic molecule could form integrated circuits. ‘Most of it is just theory. Testing it is technically impossible, because it would require attaching a single molecule to some six electrodes.’

These electrodes would create the input for the molecular circuit. The different currents would influence each other, and the input would be integrated into one output current, just like the silicon integrated circuits that power our computers.

Van der Veen obtained her PhD degree with honours in 2006. ‘It is great work, but in the end, it was too obscure. We couldn’t get it published in scientific journals.’ Hummelen therefore took a step back with his next PhD student, Hennie Valkenier. ‘She set out to test our assumptions at the most basic level.’

Hummelen’s hunch was that the conductance of linearly conjugated organic molecules would be different from that of cross-conjugated molecules. This requires some explanation, so feel free to skip the next two paragraphs if you don’t feel up to it. Hummelen draws some chemical figures on a piece of paper (see illustration).

Organic molecules have a backbone of carbon atoms. These can be linked by single (-) or double bonds (=). If single and double bonds alternate throughout (like this: -C=C-C=C-C=), it is known as linear conjugation, but if a carbon atom in the line has two single bonds (-C=C-C-C=C-), it is known as cross conjugation.

‘The difference looks minor’, says Hummelen, ‘but we believed it was important.’ So Valkenier set out to prove this. ‘She made three molecules that were almost identical, except for this one carbon atom in the chain that had either two single bonds or one single and one double one.’ Furthermore, the carbon atom with the two single bonds had either one double-bond side chain (cross-conjugated) or two single-bond side chains (unconjugated). Thus, the three molecules were as follows: one linearly conjugated (#1 in the figure), one cross-conjugated (#2) and one unconjugated (#3).

‘We designed and synthesized these molecules, and Valkenier took them to different labs that used different techniques to measure single-molecule conductance.’ When the results came in, Hummelen could see that his hunch was correct. ‘The cross-conjugated molecule proved to be the worst conductor and the unconjugated one also exhibited very poor conductance. However, the linearly conjugated molecule was a good conductor. It was exactly what we had hoped for.’

Valkenier wrote her thesis and was awarded her PhD, also with honours, but that is not quite the end of the story, says Hummelen: ‘At first, it was enough just to know about the conductance, but the researchers who did the measurements wanted to know why the molecules behaved as they did in a quantum-mechanical sense.’

A first clue came when Ryan Chiechi’s PhD student Davide Fracasso, also from the Hummelen research group, measured the molecules . Then a group from Leiden University, led by Sense Jan van der Molen, a former postdoc at the University of Groningen, showed that the lack of conductance in the cross-conjugated molecule (#2 in the figure above) was indeed caused by a phenomenon called ‘ quantum interference ’. They published an article about this in Nature Nanotechnology in 2012. ‘It was proven for the first time that this effect occurs in a single molecule and can be observed simply at room temperature.’

Then Herre van der Zant and his PhD student Mickael Perrin at Delft University of Technology discovered something weird in the unconjugated molecule (#3). ‘In almost any system, the current increases with the voltage’, Hummelen explains. ‘But in this molecule, when the voltage was turned up, at some stage they observed first a decrease then an enhanced increase and then a normal increase again in the current.’

This resulted in last week’s Nature Nanotechnology paper . Hummelen: ‘So we had a potty idea, developed some molecules to prove our point, and all these new discoveries ended up being made. Isn’t that fabulous?’

And it gets even better: ‘To get this paper published, we had to do a truckload of calculations to explain what happens. Theoreticians from Delft did these for us. And the exciting bit is that that work showed us we were right about our new proposal, which is based on our results, on how to design a molecular diode. Such a molecule was already suggested in 1974 , but nobody knew how to make one that actually worked. Now we do!’

What about the potty idea that got it all started? Are we any closer to making molecular integrated circuits? ‘We’ve shown that our ideas are essentially correct, but only in a very simple model system. Building and testing an actual integrated molecular system remains science fiction, I’m afraid.’