Two scientists at the Hewlett-Packard (HP) labs in Palo Alto have won a US patent for a method to construct molecular electronics.
The work opens the possibility of memory arrays with wires only 2nm in diameter. The patent has been issued to computer architect Phil Kuekes and Stan Williams, director of quantum science research at HP.
"Since the late 1990s, we have been working on a method of chemically constructing ICs," said Kuekes. "The aim is to separate the construction process from the design."
Kuekes and Williams constructed nanoscale arrays of wires with specifically designed chemical molecules in between them. The molecules act as the equivalent of a series of transistors, providing memory and switching functions.
"HP and other laboratories looking at molecular electronics have used electron beams to connect the wires," said Kuekes. "Our patent is about a chemical-based manufacturing technique to make the connections from the molecular electronics to normal ICs."
The technique involves extending the wires in the nanoscale array and then connecting them with wires that are 100 times their diameter by randomly sprinkling gold nanodots over the structure. These act as the address wires.
"Chemists can only build relatively simple things such as a grid or a completely random structure; both low information structures," said Kuekes. "The problem is that, in an existing memory chip, the wires are regularly numbered. If we could make a regular pattern, we would have solved the lithographic problem."
Kuekes and Williams connected the extended molecular wire to the large wires in a random fashion. They then developed an algorithm to ensure, first, that the odds of any two connection patterns being the same is infinitesimal and, second, to establish which addresses have been used. Kuekes says this establishes a method of constructing the connections between the molecular array and the IC.
"We are currently working on a 4X4 array," said Kuekes. "During the next four years, we will attempt to build a 16000bit memory. This is basic research and we still have to solve a large number of problems. We have to demonstrate that molecules still work when they are between nanoscale wires and then scale it up. It has to be a reliable process."