September 25, 2009 - Several weeks ago researchers from the National Institute of Standards and Technology (NIST) and the U. of Maryland touted a method to overcome an obstacle in creating molecular switches: sandwiching organic molecules between silicon and metal. The work was published in the Journal of the American Chemical Society.
While the general concept of molecular switches isn’t new, the NIST/UMaryland work managed to achieve a molecular junction of a densely packed monolayer, chemically bonded to silicon and metal, using a nanoprinting method to help overcome the fragility and susceptibility of organic molecules to semiconductor manufacturing process steps — particularly the high temperatures of metal deposition for attaching to electrical contacts.
Previous efforts, NIST materials scientist Mariona Coll Bau told Small Times in an email exchange, used nanotransfer printing (nTP) to build electrodes on top of fully formed monolayers on materials (e.g. metal, dielectric, semiconductors), but it doesn’t work with the silicon generally used in manufacturing due to the reactive nature of the semiconductor surface. What Coll Bau and colleagues did was use a commercially available nTP to create an ultrasmooth gold surface, and (utilizing gold-thiol surface chemistry properties) "create[d] a well-ordered monolayer on the ultrasmooth Au with an exposed functional group capable of bonding to silicon." The same nTP tool was then used to bring the Au-monolayer together with a chemically cleaned silicon surface to bond the reactive groups — a process they dubbed "flip-chip lamination" (FCL). The flexible substrate also enables conformal contact over a large area to make uniform molecular junctions, she added.
The flip-chip lamination method creates an ultra-smooth gold surface (top), which allows the organic molecules to form a thin yet even layer between the gold and silicon. Gold surfaces created by other methods are substantially rougher (bottom), and would result in many of the molecular switches either being smashed or not contacting the silicon. (Credit: Coll Bau, NIST)
Thus, two challenges were addressed and solved, she explained: Making top electrical contact, with a process that produces a smooth, low-temperature, and conformal electrical contact; and making a dense monolayer chemically attached to silicon. "We utilize well studied Au-thiol chemistry to make highly ordered monolayers first, then attach to Si using FCL," Coll Bau said.
The ultimate goal of their work is a metal-molecule-silicon structure, she noted. Previous work started with silicon and formed a monolayer, followed by metal evaporation — but this was tricky because the monolayers were less dense (than those on Au) and harsh evaporation forms "many shorts and degrad[es] the molecular layers." Evaporation (on molecules on Si) could be replaced with the nanolamination process, she said, but that still leaves a less dense monolayer. "The nanolamination requires a sticky group at the end. Forming monolayers with two sticky groups (one functional group to react with the metal and one for the silicon) doesn’t work well on Si because both stick really well," she said — adding that thiols still well on Au, but not on other species.
Yet another conceivable method would be to evaporate the metal on plastic, form the monolayer on the metal, "and then squish it onto Si," she said. However, this generates a very rough Au layer with grains equal to or bigger than the molecule (see figures), "making it easy to short and hard to have a molecular junction where the electrical properties of the metal-molecule-Si are determined by the molecular layer."
Ultimately, they figured out to lift off the Au to get: an extremely smooth Au surface, shown to make the most dense bifunctional monolayers, and flexible enough to be squished onto the Si wafer and create uniform contacts with the molecules bonded to both Si and metal.
And there’s a bonus to the process, she pointed out. "This fabrication technique can be extended to patterned metal, different molecular layers, different metals and bottom substrates," she said. "In addition to bioelectronics, we could do graphene electronics, nanowires, organic crystals, etc."