Graphene in High-Frequency Electronics
This two-dimensional form of carbon has properties not seen in any other substance
Graphite is made of a huge number of sheets of graphene stacked on top of each other. If a single layer of graphite is peeled off, we have graphene. A light pencil mark on paper may actually leave traces of graphene, as pencil “lead” actually contains graphite. But peeling off a single layer of free-standing graphene without having it buckle and fold is not the same as taking a sheet off a stack of paper.
Geim and Novoselov first obtained graphene by exfoliation using transparent tape: By repeatedly sticking the tape on graphite and removing it, they were able to sometimes transfer small flakes of graphene to another carrier material. This technique let them obtain enough graphene to do the seminal experiments that led to their Nobel Prize. This ease of producing graphene probably led to the intense and rapid interest in the material: Any university with a chunk of high-quality graphite, a supply of tape and patient graduate students could produce enough of the material to do some interesting experiments!
However, the flakes of graphene made this way are tiny: tens of micrometers in width. Such small pieces are enough for experiments but useless for the real electronics applications that our group had in mind. Modern electronic circuits are manufactured by wafer processing: Thin discs with a diameter of 200 or 300 millimeters, made of silicon or other materials, are treated with chemical and optical steps to make dozens or hundred of identical circuits simultaneously. For graphene electronics to be practical, graphene has to be made in larger amounts and sizes, and processed into circuits in a way that can be mass-produced.
At the moment there are two methods of producing graphene on this scale. Both involve processes to grow a single layer of atoms stretching across many centimeters—enormous for this material. Graphene is a picoscale (300 x 10–12 meters) material in one dimension, but macroscopic in its other two dimensions.
One growth method is the formation of an epitaxial layer (where the crystalline structure of the layer aligns to that of the substrate) of graphene on silicon carbide (SiC). At sufficiently high temperatures the silicon–carbon bonds will break and the silicon will evaporate from the surface, leaving exposed carbon atoms. If this event occurs with appropriate gases present, additional carbon atoms will adhere to those on the surface, forming the hexagonal pattern that makes graphene. This method has been used for a few years to produce 50-millimeter wafers with a silicon-carbide support structure covered completely by a layer of graphene, which can be patterned and formed into devices and circuits using many of the processing techniques found in conventional electronics fabrication. Wafers 100 millimeters in diameter are now becoming available, which makes this method even more attractive. Silicon carbide is particularly useful for radio frequency electronics: Unlike silicon, it is an insulator, so unwanted signals will not propagate from device to device through the substrate. Additionally, it is optically transparent. One drawback of the epitaxial graphene is that one of its surfaces is fairly tightly bonded to the base, impacting its mobility and making for a transistor that would be slower than one made of free-standing graphene.
Another method of producing wafer-scale graphene is growth through the chemical vapor deposition (CVD) of carbon on a catalyst material. Good quality graphene has been produced using copper as the catalyst. If a wafer covered with copper is exposed to ethylene (C2H4) under appropriate conditions, a single layer of graphene forms on it. Of course, copper is a conducting material, and it connects all the graphene together, making it useless for circuits, so the graphene layer must be transferred to an insulating substrate before it can be put into production. The graphene is first coated with a polymer, then the copper on its underside is chemically etched away. The graphene/polymer sheet is placed on a carrier wafer, and the polymer is then removed, leaving a single sheet of carbon atoms. This technique can be used to cover 200- or 300-millimeter wafers with graphene, and any final substrate can be used. Therefore it’s now possible to transfer the graphene onto already-fabricated circuits to make a hybrid technology. The drawback of the CVD method is that, at the moment, the mobility of the graphene is not as high as that formed from epitaxial growth, due to residues of the polymer adhering to the graphene, physical domains (places where the grid of carbon atoms doesn’t align, creating boundaries) and even wrinkles in the graphene sheet. But with the amount of effort spent on generating graphene with CVD, this situation will probably improve.