Printing transistors instead of etching them is not a completely new idea. Researchers now show that the process also enables new, flexible material combinations and takes integration on chips a step forward
The 80-kilometer strip between san francisco and san jose, known as the birthplace of the modern electronics industry, is called silicon valley – but if the industry had to make do with silicon alone, there would probably be no cell phones or fiber optic networks today. Silicon is the most important semiconductor used in electronics, but it is not the only one.
3d circuit structure of heterogeneous materials bent around a glass cylinder
This is not a problem in itself. However, things get complicated when you try to combine different classes of semiconductors on one chip using conventional production methods. One of the reasons for this is that silicon is very sensitive to impurities. Scientists at the university of illinois in champaign-urbana have now found a way to combine different types of electronic components (from carbon nanotubes and gallium nitride wires to silicon or gallium arsenide) using a printing process. They present their work in the scientific journal science (doi: 10.1126/science.1132394).
The process consists of several steps: first, the individual components are allowed to grow independently on their own substrates. This has the advantage that all material-specific pretreatment steps can be carried out without risk. The structures are then separated from the substrate and attached to a new, common substrate using a special printing process, which may also be organic, for example. The process mimics transfer printers: the printing stamp first picks up the elements to be transferred (the "ink") and deposits them back onto the target substrate. Finally, conductive connections are made between the individual parts of the newly created 3d structure – and that’s it.
Practical tests showed that the process works very efficiently. The researchers combined tfts made of carbon nanotubes, hemts made of gallium arsenide and mosfets made of silicon nanowires to create a multilayer ic.
Microscope image of a heterogeneous three-layer ic structure. The individual structures are around 20 micrometers in size
In contrast, the circuits that a team of u.S. Researchers wants to create using a kind of potato printing process are completely organic in nature. This week, they are presenting their idea in the scientific journal nature (doi: 10.1038/nature05427). It goes beyond what the industry can already do. The process creates circuits not from thin organic layers, but from organic crystals. These are quite a bit more powerful, but until now they have always had to be placed on the target material by hand – a technique that is unsuitable for industrial production.
The team of scientists led by zhenan bao therefore grows the semiconductor crystals on the desired substrate right away. They are animated to do this by a method not unlike potato printing: they transfer a few micrometers by a few micrometers of coarse dots from the source material to the substrate, where the semiconductor crystals are stimulated to grow. The scientists succeeded in applying the method even to flexible substrates. At the moment, however, it is not yet possible to control the orientation of the crystals – this is necessary because of their anisotropic conduction properties.
Flexible field of organic single crystals around a glass fab (photo: alejandro briseno)