36 | New Scientist | 28 March 2020IBM
RE
SE
AR
CH
/SC
IEN
CE^
PH
OT
O^ L
IBR
AR
Y
problems, which explains why physicists
have been arguing vehemently and largely
fruitlessly for decades about which is best,
without any hard evidence to help them
decide. This is exactly what Minev and his
colleagues sought to rectify last year with an
ambitious experiment designed to probe
quantum measurement more sensitively
than ever before.
To understand their results, they turned
to a lesser known formulation of quantum
mechanics known as quantum trajectory
theory, or QTT. It was developed in the 1990s
to track the path that a quantum object takes
through the space of all its possible states as
it is measured. “QTT is a wonderful, and widely
used, tool for describing quantum evolutions,”
says Max Schlosshauer at the University of
Portland in Oregon.
There is nothing in QTT that deviates from
regular quantum mechanics. But it does have
a unique selling point. Unlike the Schrödinger
equation, which only describes a quantum
system in isolation, it can describe the way
quantum objects interact with their
environment, dispersing their quantum-nessThe mysterious process by which the fuzzy,
undecided quantum world snaps into the
familiar certainties of “classical” physical
reality was always thought of as a one-way
street. But maybe not for much longer.
The physicists behind a recent
experiment that gave us a first glimpse
inside this process (see main story) now
think they might be able to reverse it too- a feat that could boost the quest to deliver
useful quantum computers, which exploit
the strange properties of quantum physics
to speed up certain calculations.
Last year, Google claimed to have
reached the milestone of making a
quantum computer that can solve a
problem in minutes that would take the
best conventional computer thousands
of years. But one of the biggest hurdles
to quantum computers routinely solving
real-world problems is that it is hard to
correct for errors that inevitably occur when
a quantum bit, or qubit, which is the core
component of a quantum computer,
randomly flips its state – from a 1 to a 0,
say. There is always a chance of such errors
happening in the delicate quantum state of
the qubits used for computation, and they
create “noise” that can derail a calculation
if they aren’t corrected.
By catching a similar transition known as
a quantum jump in the act, and anticipating
when it was about to happen, Zlatko Minevat Yale University and his colleagues have
suggested a possible solution. The idea is
that, with sufficiently sensitive detection
gear, you could monitor qubits so carefully
that you could anticipate and immediately
undo errors, or even prevent them.
According to Minev, the method “gives us
the potential to intervene in the process by
which noise occurs in a quantum computer,
allowing us to potentially reverse it before it
can muck up an entire computation”.
William Oliver, an experimental physicist
at the Massachusetts Institute of Technology
also sees the potential. “Provided there is a
signature that indicates a jump is about to
occur,” he says – one that doesn’t actually
reveal the information in the qubit, which
would destroy the quantum-entangled
qubit state on which the computation
depends – “then one can imagine
correcting for such jumps as they happen.”
Other error-correcting techniques are
already being developed, but this one
could tackle the problem at its source.Quantum corrections
in a process known as decoherence, and
getting jostled by the environment in return
in another process known as back-action.
Minev considers QTT “a refined and evolved
version of quantum mechanics”.
The trouble is that QTT is maddeningly
difficult to use for analysing experiments
because you need to know pretty much
everything that happens. Imagine, for instance,
attempting to monitor an atom that might
make a quantum “jump” between energy states,
emitting a photon of light as it does so. To apply
QTT, you need to keep checking at incredibly
short time intervals to see if the photon has
been emitted. You can’t afford to miss a single
photon. And every time you check, you must
consider the effect the resulting back-action
has on the atom. It is hard to overstate how
difficult that is. “Until now, the timescale at
which jumps or collapses occur has been too
fast to measure,” says Minev.
That has now changed thanks to a team led
by Minev’s PhD supervisor Michel Devoret,
and including theorist Howard Carmichael
at the University of Auckland, New Zealand,
who helped develop QTT. They used“ The founders of
quantum theory
dreamed of
performing
experiments
like this one”
Today’s quantum computers
are delicate and error-prone