New Scientist 28Mar2020

"The founders of

quantum theory

dreamed of

performing

experiments

like this one"

problems, which explains why physJclsts

have beenarguingvehementlyandhugely

fndtlesslyibrdecades about which is best.

withoutanyhardevidencetohelpthem

decide. This is enctlywhat Minev and his

colleagues sought to rectify last year with an

ambitious experiment designed to probe

quantummeasurementmore sensitively

than ever before.

To understand their results, theytwned

toalesser known formulation of quantum

mechankslcnownasquantumtrajectoiy

theory, orQlT. Itwas developed in the 1990s

totmckthepaththataquantumobjecttakes

through the space of all its possible states as

it is measured. "CDT is a wonderfuJ. and widely

used, tool for descdbing quantum evolutions;

says Max Schlosshauer at the University of

Portland in Oregon.

There is nothing in QlTthatdeviates from

regular quantum mecbanics. But it does have

a unique selling point Unlike the Sdu6dinger

equation. which only describes a quantum

system in isolation. it can describe the way

quantumobjectsintelactwitbtheir

enviromnentdispeningtheirquantum-ness

in a process known as decoherence, and

gettingJ06tled by the environment in retum

in another process known as back-action.

Minevc:onsidenQ'.IT"arefinedanclevolwd

venionof quantum mechanics~

The trouble is that QlTis maddeningly

difficult to use for analysing experiments

because you need toknowprettymuch

everything that happens. Imagine, fbrinstmce,

attemptingtomonitoranatom that might

make a quantum "jump" betweeneneigystates,

emitting a photon ofllghtas it does so. To apply

QlT, JOU need to keep chec1cingat incredibly

short time intervals to see ff the photon bas

been emitted. You am't afford to miss a single

photon.Andm!qtimeyoucheck, youmust

considertheefl'edtheresultingbaclc-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 nowchangedthanb to a team led

by Minev's PhD supervisor MichelDevoret.

anclindudingtheoriltHoward<:armic:hael

at the University of Auckland, New Zealand,

whohelpeddevelopQIT. They used

Quantum corrections

llleMPla ..... lllOCllll brwhlchthefuaJ,

wm.:ldaclquntumwurlcllllQl-..Uthe

flmllllrC81'1111ntielof ._.. .. .,.,,...

....., ................ o1 •• ......,

stNet. Butma,benatfarnu:hlonglr.

'lhephvlldltlblhhlareant

mcperfmentthatgneusatlrstglmpse

fnlldltthls111aa111(Saemafnlblry)nuw

thil*thav might be able tu nM1118 lttDo

-ar..tthatmuldbDClllthequllttDdellwr

mefuilquantwnanp......., ........ t

thestNngeproperU.ofq&mUftphvsics

........ artaincalallatlons.

Lmt,..,GaogleclaimedtD hne

rachedthemo.tm.oflllllmtga

~mnputsthatcansalwa

prabllln In mlnua thatwauld 1111111 the

beltconwntlanalanputwthauands

Df,_...Butaneoftheblggatlmdi.

......... COl'l'QIUllrlrautfnely .....

........... dproblemlisthatitllhardtD

Today's quantum computers

are delicate and error-prone

correct far errors that lnevitablf oa:urwllln

• ..,..,, bit,orq...,lt,whldl lltllean
  mmpanmtof•quantum..........,,
  randamlJtlpsllsstal8-frama 1 to•O.
  say. n..11-.aclllnaloflUChemn
  ...11ng1nt11edllkal8..,...mmat
  theqlllllslllldfarmmputatlan.andlwJ
  a911bt.,Wthatmnder.laalailaftarl
  lfthav.,..,. ,iidad.
  a, ~Ing a11mllll'tflnlllio11 known•
  aqunum...,,,intheld,1ndlftticiPlilfnl
  whlnltWBllbouttDhlppln,ZlatlmMIMv

atYlle Uniuerl1ty1ndhllcalalg&m._.

sugg....i•pcilllllesalutton. 111e ldla Is

that, with sufftd8ntlr-detllctlan

...,, Vallcauld manltarq..nssacandullr

that,.. mukl antldp8'8and1mmedlatlllr

imdoemn,ar..,.pravamthem •

Armn:llngtD ... themalhod ~--

the PGtlntlal tD 1m. ... in the lllOCllll br

whldlnoileocanln1QU1mtumcempu1S,

ellowlng UltDpotantidv ..... it...,. it

canmudc ....... mmputmtl~

Wlllllm DllUll', 11'1 aperlmental phylldlt

attheMasaduettsntftubtrd'lldmalaW

allo-tllepalantlal • .,,......there Isa

slgnatunl that indlcatas a Jump Is about tD

oma/'lwsaps-anethatdaem'tlldullllr

nMlllthe .......... inthequblt,whkli

wouldcll*oJtheq~

CIUbitstateonwhichtllecomput9tlon

dlpendl-..,..aneanlmaglne

ca11eclmgfarlUChfwnps•tMJhappen.•

otherenar-mnctlngtllchnlqms ..

alrudrblmgdmllaped,butthlsane

cauld tlldclethe prablam at Its samat.