superconducting quantum bits like those in
quantum computers to build an artificial.atom
and watchitjumpingfromoneenergystate
to another. Theyusedm!crowavestoezdte
thelr"atom",thenwatcheditemitmicrowave
phot.ons as it returned to its ground state.
Jn truth. what they were watching is a bit
mmec:omplicatedtban that. The "atom" keeps
jumping to the excited state and then, under
the influent-e of the back-action caused by
probing it, falling back down again. It doesn't
"stick'' in the excited state until a true quantum
jump occurs. And now, for the first time, this
jiggling back and forth could be tracked.
"The founders of quantum theoiy dreamed
ofperfonning experiments such as the ones
we can begin to perform now;" says Minev.
WhattheJeSeall'hers sawwuaquantum
jumpunfoldiJlgovertime:aphenomenon that
tums outtohappensmootbly,notsuddeolyas
Bohrandhlscollabomtonbadassumed. The
jumps occurred at random moments, but there
was aldnd of precursor signal when one was
imminent: the j1ggllngcaused by quantum
back-action became unusually quiescent.
Thanbtothisadvancewamhlgofan
impenclingjump, the resean:hers were even
able to fire microwaves at the qubits to catch
and.reverse the jump as it was taking plat"e,
something never before achieved.
What does thisbawtodowithcollapse?
Well, though it wasn't much remarked on
at the time, the quantum jumps experiment
was in13ct also monitoring the prooesa
conventionally regarded as wave function
coDapse-because thatJs an inevitable
consequence of continuous observation. Jn
this case, as the artific:ial atom wu continually
drivm towanis one of its excited states,
measurement kept collapsing it to the ground
state. The same applied to the shift into an
excited state. So the result implies that
"collapse", too, Is a teal, physical and smooth
process-an accumulation of small back-
actions from continuous monitoring of the
system-that can be seen as it unfolds.
Tiie collapse of collapse
Thewmkevenralsesthepmspectofavoiding
"collapse"altogetherwhilemaldnga
measurement. This would mean controlling
the interactiDns of a quantum entity with
the environment so auefullythattbere is
negligible back-action, and thus minimal
disturbance. Such a measurement would
supply information as precisely as could be,
subject to the constraints of Heisenberg's
uncertalntyprindple, which says there Is
a limit on how accurately certain pain of
properties can be measured. It would probe the
system at the so-called "Helsenbergllmi~
free from any external back-action noise.
That has long been a goal for extremely
sensitivequantumdetectionmethods,
such as measuring photon travel times in
a gravitational-wave detector, and it could
haw a serious role to play in making good
on the promise of quantum computing
(see "Quantum corrections," left).
"Given sufticiently powerful:read-out
hardwue, wecanmalmmeaningfulquantwn
measurements wtthoutcollapse," says
Devoret. "We are just atewyean away from
being in a position to dothi1 kind of
measurement."
In the meantime,. the results that the team
aheadyhavegiveusplentytochewon-not
least the implication that the notion ofwave
function collapse was neverreallynecessaiy
in the first place. It ls just aaude way of talking
aboutthechangethatoccunwhenaquantum
syatemgets entangled with. and distmbed
by. its environment. "The whole lexicon of
'collapse'isfatallytiawm." saysMin.ev. "It'sa
remnant ofthe discussions in the 1920s, and
givesthewrongmentalimage. Quantum
tnjectoiythecnypeels away the veil that
bas obscured the mec:banics of collapse,
and shows us there is no such thing."Nawexpll'Jmlfttl
seem ta undermine the
....... worlds-take Dll
quantum theoryThat might sound radical, but it is supported
by anotherrecentexperimentconduded by
Markus Henmich and colleagues at the
University of Stockholm in Sweden, in
collaboration with.Adan cabello at the
UniversityofSevilleinSpainandothers.1hey
were able to perfonn a spedal. "ideal" kind of
quantummeaaurementthatdoem'tdestroy
thequantumstate(uitdoeswhenaphoton
is detected by beingabsoJbed, for instance)
but shifts it to another state that can be
measmed agalll. This applies even for
superpositions: quantum states in which mo~
thanonepossibleoutcomeofameasurement
is possible. Superpositions are nonnally
destroyedbymeaturement, buttheycan
smvive an "ideal" measurement like this.
It has ne\'el' been done before, but Henmich
and.his colleagues pulled it offby measuring
electric:allytnpped strontium ions. And.again
they saw a smooth. gm.dual clumge in state
rather than the abrupt. destructive snap of
conventional collapse. When properly used,
says Cabello, quantum mechanics "describes
measurement as a process that requires time
and tells how the quantum state evolves". >