All That Matters

This week my colleagues at Nature Chemistry landed an impressive scoop, the publication of a paper by Michael Strano and colleagues from MIT on self-regenerating solar cells.

The performance of any kind of solar cell tends to degrade over time. This is particularly the case for organic solar cells, where sunlight can easily destroy the structure of the molecules used. Natural light-harvesting processes have a similar problem, for example during photosynthesis. The way plants solve this problem is through a self-repair mechanism.

Schematic of the regenerating solar cell consisting of light-absorbing proteins, lipid disks and carbon nanotubes. Reprinted by permission from Macmillan Publishers Ltd. Nature Chemistry, advance online publication (2010)

Taking cues from such self-regeneration strategies, Strano and colleagues use a concept that is surprisingly simple. They prepare a solution containing carbon nanotubes, bacterial light-harvesting proteins and discs made from lipid molecules — the structural components that form the membrane of cells. Once the surfactant that keeps all these molecules separate is removed the molecules assemble themselves: the proteins bind to the lipids, which then attach to the carbon nanotubes. No assembly required.

During solar cell operation sunlight is absorbed by the proteins and creates electronic charges that are transported along the carbon nanotubes to the electrical contacts of the solar cell. To regenerate the proteins damaged by the sunlight, surfactant is added again, along with a small quantity of new proteins to replace damaged molecules. This dissolves the structure. But once the surfactant is removed the molecules reassemble, fully repaired.

In the study, the solar cells ran on a 40 hour cycle: 32 hours of operation, followed by 8 hours regeneration. Despite so many hours of regeneration overall cell performance was up by a remarkable 300% in comparison to cells that are not regenerated. And this could be just the beginning. At the moment, as soon as the cells are turned on they lose about 60% efficiency within the first hour or so of operation. Delaying the onset of degradation or finding a more efficient way of regeneration should lead to further enhancements. But who knows, nature may have a solution to this problem, too.

Reference:
Ham, M., Choi, J., Boghossian, A., Jeng, E., Graff, R., Heller, D., Chang, A., Mattis, A., Bayburt, T., Grinkova, Y., Zeiger, A., Van Vliet, K., Hobbie, E., Sligar, S., Wraight, C., & Strano, M. (2010). Photoelectrochemical complexes for solar energy conversion that chemically and autonomously regenerate Nature Chemistry DOI: 10.1038/NCHEM.822

Last week I blogged about a Nature paper on graphene transistors with a self-aligned nanowire gate.  Well, as I gather from a blog post by Doug Natelson, largely the same UCLA researchers have now published a paper in Nano Letters that uses a rather similar idea, even though in the latest paper the nanowire gate is made from another material, and it seems the latest transistors are even faster.

However, I am worried about the obviously parallel publication of these two papers. The Nature paper was submitted 23 May, published 1 September. The Nano Letters paper was submitted 16 May and published 3 September.

As Doug says: “Looks like they managed to get two papers in good journals for the price of one technique advance.” And I agree, it looks very much like salami slicing of research results to me. In particular, I like to emphasize that it is editorial policy of Nature journals that editors (like myself) are informed about any related submission made to other journals — see Nature‘s policy on duplicate submissions and plagiarism.

I have not checked this fact with my colleagues and if I would have I could not comment here on my private blog, but I do wonder whether such communication has taken place here. If I were the handling editor, this dual publication would not have been knowingly possible, but others might of course have a different opinion.

Either way, in cases where our policy on duplicate submission is not followed, little can be done if clearly different materials were used, even if a study is based on a similar concept. Apart from increased scrutiny of such authors in future submissions of course.

References:

Liao, L., Bai, J., Cheng, R., Lin, Y., Jiang, S., Qu, Y., Huang, Y., & Duan, X. (2010). Sub-100 nm Channel Length Graphene Transistors Nano Letters DOI: 10.1021/nl101724k

Liao, L., Lin, Y., Bao, M., Cheng, R., Bai, J., Liu, Y., Qu, Y., Wang, K., Huang, Y., & Duan, X. (2010). High-speed graphene transistors with a self-aligned nanowire gate Nature DOI: 10.1038/nature09405

This week two interesting papers have been published that I did not get around to highlight here. In terms of topic they could not be more different, one about a possible new data storage material, and the other one about string theory!

The next big thing in computing could be silicon!

A new memory element made from silicon oxide (SiOx) between two silicon contacts. It promises a silicon-compatible non-volatile memory device. Reprinted with permission from Nano Letters DOI: 10.1021/nl102255r. Copyright 2010 American Chemical Society.

It is not often that advances in condensed matter physics are highlighted on the front page of the New York Times, but the possibility of a new silicon memory technology was compelling enough it seems. The New York Times story is a bit short on details, but one of the authors of the paper published in Nano Letters, Douglas Natelson from Rice University in Houston, describes details of the work on his blog.

What Doug and his colleagues have discovered is that by applying a sufficiently high voltage to an insulating silicon oxide the material suddenly becomes conducting. Increase the voltage further and the devices becomes insulating again. The effect is very pronounced and fully reversible for at least 10,000 cycles.  What happens during the initial switching is that electrochemical processes lead to the formation of electrically conducting silicon nanocrystals within the oxide. The nanocrystals are close together so that electrons can easily hop from one to the other so that the device becomes conducting. At even higher voltages the device becomes insulating again as heating effects mean that probably some of the silicon oxidizes again.

The channels between the conducting silicon nanocrystals are only 5 nanometers wide. This suggests that memory devices could be fabricated that are much smaller than present memory devices with sizes of a few tens of nanometers. The device design is really simple. Switching occurs through the application of voltage pulses, which requires only two wires: one input and one output. Moreover, the fabrication of silicon oxide is fully compatible with present silicon technology. And in comparison to the silicon chips in your computer this memory device is non-volatile, it doesn’t lose the information when the computer is turned off. Rather, it behaves more like the flash memory used in solid-state hard drives.

Obviously, a lot of work is required to bring this to the market. At the same time, conceptually this memory effect is somewhat related to that memristor devices. The same New York Times article mentions how Hewlett-Packard is in the process of developing memristor memory chips, so prospects seem realistic.

Reference:
Yao, J., Sun, Z., Zhong, L., Natelson, D., & Tour, J. M. (2010). Resistive Switches and Memories from Silicon Oxide Nano Letters DOI: 10.1021/nl102255r

Could string theory be proven experimentally?

One of the geometrical objects studied in string theory. The image shows a three-dimensional projection of this six-dimensional object. Image by Lunch via Wikimedia Commons.

String theory promises nothing less than to unify the most fundamental concepts in physics, quantum physics and general relativity. In string theory fundamental particles like electrons are not billiard ball-like particles but vibrating strings that can be imagined as little rubber bands.

Unfortunately, string theory is not nearly as straightforward as this simple picture may suggest. The mathematical description of these strings is very challenging and requires more than the four dimensions we know from our universe (space, time). Indeed, most string theories are based on ten-dimensional spaces. To explain the fact that we observe only four of them, one has to imagine the remaining dimensions to be reduced so that they cannot be observed. It is as if you compress a cubic object in one direction until it almost looks like a flat sheet — effectively you see only two dimensions.

One of the biggest problems of string theory is, however, that it cannot be proven experimentally. Or so we thought. In a Physical Review Letters paper Michael Duff and colleagues from Imperial College London claim to have found an experimental proof. They realized that an identical mathematical formulation describes so-called quantum entanglement experiments with single particles of light, photons, as well as black holes. According to the Imperial College press release Duff had this idea at a conference when listening to a talk on entanglement. “I suddenly recognised his formulae as similar to some I had developed a few years earlier while using string theory to describe black holes. When I returned to the UK I checked my notebooks and confirmed that the maths from these very different areas was indeed identical.”

I cannot claim to understand much about Duff’s paper, but it seems that experiments on the entanglement of four photons should be able to replicate certain properties of black holes — if string theory is correct. String theory is controversial, and I am sure this won’t be the last word on the topic. But I am sure somewhere someone is already planning such entanglement experiments. It is too tempting.

Reference:
L. Borsten, D. Dahanayake, M. J. Duff, A. Marrani, & W. Rubens (2010). Four-qubit entanglement from string theory Phys.Rev.Lett.105:100507,2010 arXiv: 1005.4915v2