"Entanglement is the secret of the exponential
computing power of a quantum computer.
Competition for the miracle calculators from Google &
Co. A future quantum computer may calculate with quantum bits made of silicon.
That would have many advantages. Three research groups have created an
important prerequisite.
Even if the first commercial quantum computers are already
finding their way into data centers - such as the "Advantage" system
from D-Wave at Research Center Jülich recently - the development of a
universally programmable and fault-tolerant quantum computer is still a long
way off.
The quantum physical information units, the quantum bits, which are
mainly realized by stored ions or superconducting microwave resonators, are
still quite fragile and prone to failure. Thermal radiation, noise and stray
fields can destroy the states of the qbits, which inevitably leads to
calculation errors. In addition, scalability poses difficulties: the
performance of quantum processors cannot be increased simply by increasing the
number of quantum bits. The more Qbits there are, the more difficult it becomes
to shield and control each one from disturbing influences.
Many research groups have long seen an alternative in
quantum computers made of silicon, especially to the superconducting systems
favored by IBM, Google and D-Wave. This also includes the developers of the
chip manufacturer Intel. Extremely compact and stable quantum bits can be
realized in the semiconductor material. By increasing the working temperature,
the cooling effort could also be reduced in comparison to the superconducting
systems. Another advantage: Quantum circuits based on silicon can be
manufactured using standard semiconductor technologies and accommodated on a
microchip. This makes it possible to scale silicon quantum systems in a
relatively simple manner.
Three research groups from Australia, Japan and the
Netherlands have now been able to clear another hurdle that has so far stood in
the way of practicability. They have developed simple silicon quantum
processors that calculate as reliably as the superconducting quantum bits in
IBM or Google systems already do.
Detect calculation errors early and correct them quickly
In the past, it was possible to keep individual qubits
stable for 35 seconds – an eternity for quantum systems. For comparison: The
superconducting Qbits remain coherent for only 100 milliseconds. However, if
you tried to combine two silicon quantum bits into logic gates in order to
actually carry out arithmetic operations, you had to expect a high error rate
in the results. Using different approaches, the three research groups have
succeeded in reducing the error rate to well below one percent. Errors that
occur despite all measures can now be detected in good time and corrected
quickly, write the scientists in the journal "Nature". This clears
the way for the construction of larger, fault-tolerant silicon quantum
processors.
The Dutch scientists led by Lieven Vandersypen from the TU
Delft and the Japanese physicists led by Akito Noiri from the Riken research
center near Tokyo used the intrinsic angular momentum (spins) of individual
electrons as quantum bits, which were isolated in tiny silicon structures,
so-called quantum dots, like in a kind of cage. Spins are ideal quantum bits:
They can “point” down and up at the same time or – more precisely – exist as a
superposition of both states. The scientists had embedded the quantum dots in
thin silicon layers stacked on top of one another for shielding.
With the help of electric and magnetic fields, the distances
between the quantum bits can be controlled and the alignment of the electron
spins can be manipulated in a targeted manner. In this way, it was possible to
bring two qbits so close that an entanglement occurred.
The result was fast
elementary gates that could be used to perform simple arithmetic operations.
The switching times could be reduced to about a hundred nanoseconds, which was
reflected in the stability and the low error rate of the gates. Another factor:
all groups worked with isotopically pure silicon and used sensitive methods to
read the results of the quantum gates without causing any interference. The Japanese
and Dutch researchers state the reliability of their processors at around 99.5
percent. A value of 99 percent is considered the threshold for fault-tolerant
quantum computers.
The gate of the researchers from the University of New South
Wales in Sydney calculated almost as reliably (99.4 percent). Instead of
electron spins, however, they used the nuclear spins of phosphorus atoms that
they had deliberately injected into a thin layer of silicon. Atomic nuclei have
the advantage that they are less susceptible to interference and more robust
than electron spins. Nuclear spins isolated in silicon can be kept stable for
several seconds. Encoded quantum information can be stored with these Qbits for
a correspondingly long time.
The next challenge awaits
However, it is also much more difficult to combine nuclear
spins into a gate. The researchers led by Andrea Morello used a neighboring
electron as a mediator. Using microwave pulses, it was possible to convert the
two atomic nuclei and the electron into an entangled state. The nuclear spins
were thus coupled to one another and the prerequisites for a stable logic
quantum gate were created.
For the quantum physicist Tommaso Calarco from Research Center Jülich, the results of the three research groups are
convincing. "In the still open race for the scalability of quantum
computers, semiconductor technologies have now caught up closing a key gap."
Errors in a two-qbit arithmetic operation are now comparable with the error
rates of other platforms based on stored ions and superconductors, for example.
However, these quantum computers on stored ions and superconductors are already working with a few dozen quantum
bits, which the current silicon processors are still a long way off.
The IBM quantum computer “Eagle” currently holds the record
with 127 superconducting quantum bits. Efficient correction processes already
ensure that the error rate is low. "The next challenge will now be to
demonstrate comparable performance capabilities using multiple silicon
qbits." It is remarkable for Calarco that three laboratories made the
breakthrough at the same time.
However, the scientists at Research Center Jülich are not
working on silicon processors themselves. They are committed to developing a
quantum computer based on superconducting chips. The goal is that the
"OpenSuperQ" system should have a good hundred quantum bits in the
final configuration. The operating system will be open-source software that, in
principle, every user should be able to access the quantum computer."
Komentarų nėra:
Rašyti komentarą