Behind The Paper: Efficient readout of a silicon-based quantum computer

Behind The Paper: Efficient readout of a silicon-based quantum computer


I think what I enjoy most
about quantum computers is actually trying to build them. I really like being hands-on to really understand the device physics and learn how to make the device
work for you, how to make it do what it should do. The story of this paper
really starts in 2017, when a few groups
started to work on blueprints of a quantum classical processor. After a series of demonstrations, which have shown that silicon-based
cubits are promising candidates for large-scale quantum computing. In this work recently published
on nature electronics, we combined radio frequency
measurement techniques with concepts of random access, which is found in modern memory devices. Using this, we addressed the challenges of read-out of the large-scale devices, by reading out one device
after another using the same line. To implement such a cryogenic
control circuit, we decided, with out colleagues
at Hitachi Cambridge laboratory, and CEA-LETI, to use CMOS technology. CMOS technology is the basis
for conventional processors. It has driven the digital revolution
we’ve seen over the last decades, due to reliable fabrication
of complex circuits consisting of millions
and billions of transistors. Now a proof of concept experiment,
we combine CMOS transistors with CMOS quantum devices all on the same chip. First we had to check
that these transistors actually still work
at cryogenic temperatures. Because, for a selected cubit, they should allow
the read-out signal to be delivered without disturbance. Additionally, for the selected cubit,
signals should be blocked, and the cubit
should not be disturbed. The transistors we chose
actually work quite well and we showed sequential read-out
of two quantum devices. However, the approach
can be easily extended to a large number of devices, where each individual cell consists of
a CMOS transistor and a cubit each. Using this, we can make a two-dimensional array
of quantum devices which can be randomly addressed
similar to conventional memory chips and then read out all of these devices
using a single line. We start off with a big wafer, which we cut into small pieces, each still containing many transistor
and quantum devices. To actually measure the device, and bring it to life, we need to make connections
from the nano-scale device to the outside world. For this, we take a chip and glue it
onto a printed circuit board. We use wires thinner than a hair to actually make the connection. Now that we have the chip
wired to our PCB, we take this PCB and connect it
to a cylindrical sample holder which we can then attach to the coldest
part of our dilution refrigerator which also makes
the electrical connection from the temperature
to our device. We’re very proud of this work. We managed to demonstrate
the CMOS technology can help to solve the challenges of a large-scale quantum computer. Maybe quantum computer chips will not look so different
from conventional processors. From the circuit models
we have developed, we could next go ahead and design fully-optimised and integrated
multi-cubit circuits. There’s lots of exciting work ahead.

Leave a Reply

Your email address will not be published. Required fields are marked *