University of Utah physicist Christoph Boehme works on equipment used to demonstrate how practical quantum computers could read data.

To test the concept, Boehme prepared a 300- m-thick silicon crystal doped with phosphorus atoms. Deposited atop the crystal are two gold contacts and an extremely thin layer of silicon dioxide in between. Voltage applied across the contacts sends a small but measurable current through the device.

A constant current indicates half the spins point "up" and the other half "down." Chilling the device with liquid helium to 452°F makes most of the spins point down. Subsequent exposure to a magnetic field and microwave radiation makes the electron spins oscillate up and down in step with the microwave frequency, which shows up as a change in measured current. The technique lets Boehme "read" the net spin of about 10,000 electrons and correlate the data to known — and more stable — nuclear spins.

It's a huge improvement over previous efforts that used magnetic resonance to read the net electron spins of roughly 10 billion phosphorus atoms. But the technique is still a far cry from reading electron spins of single atomic nuclei as would be needed for practical quantum computers.

Such quantum computers could one day render conventional digital computers obsolete. Here's why: A simple 3-bit digital computer, for example, can process one of eight possible combinations of zeros and ones at a time. In contrast, each bit in a quantum computer (qubit) may simultaneously exist as both zero and one. This means a 3-qubit quantum computer could calculate 8 faster than a 3-bit digital machine. Typical modern PCs handle data in 64-bit chunks, so a 64-qubit quantum computer could theoretically run 2 ⁶⁴ or 18 billion billion times faster.