How to prepare for the acceleration of quantum technologies

This arti­cle is part of our spe­cial issue « Quan­tum: the sec­ond rev­o­lu­tion unfolds ».

Read it here

Lau­rent Sanchez-Palen­cia and his team are inter­est­ed in under­stand­ing the organ­i­sa­tion of mat­ter at the quan­tum scale, where both wave and par­ti­cle effects, such as inter­fer­ence and entan­gle­ment, are inter­wo­ven, as well as strong inter­ac­tions between par­ti­cles. Lau­rent Sanchez-Palen­cia is also involved in devel­op­ing new train­ing pro­grammes in quan­tum tech­nolo­gies at the École Poly­tech­nique and the Insti­tut Poly­tech­nique de Paris.

His team is cur­rent­ly work­ing on quan­tum sim­u­la­tion to mod­el and under­stand the behav­iour of ‘qua­si-peri­od­ic’ mate­ri­als at low tem­per­a­tures. These mate­ri­als are too com­plex to be ful­ly described at the atom­ic scale and could be bet­ter under­stood using quan­tum sim­u­la­tion. Oth­er exam­ples include high-tem­per­a­ture super­con­duc­tors and quan­tum magnetism.

The the­o­ret­i­cal results obtained for min­i­mal mod­els can be test­ed in real exper­i­ments in con­trol­lable sys­tems known as ‘opti­cal lat­tices’. These are made up of atoms at extreme­ly low tem­per­a­tures – on the order of a few tens of bil­lionths of a Kelvin – and held togeth­er by laser beams. When the right num­ber of laser beams point in the same direc­tion on a plane, the result is an exot­ic sys­tem, halfway between order and dis­or­der (ordered over a long-range but non-peri­od­i­cal­ly ordered), some­thing that is known as quasi-periodic.

Researchers are study­ing what hap­pens when inter­ac­tions between atoms lead to the appear­ance of new quan­tum phas­es called Bose glass­es. These glass­es are a spe­cial type of insu­la­tor that, in the­o­ry, should only appear in struc­tures that are either dis­or­dered or qua­si-peri­od­ic. A stan­dard insu­la­tor has an ener­gy gap between its fun­da­men­tal state and its first excit­ed states. This means that only a pow­er­ful elec­tric field can excite the charges and set them in motion. In a Bose glass, on the oth­er hand, there is no such ener­gy gap, but the charges are con­fined to very localised regions that do not allow for a cur­rent of par­ti­cles to flow.

Bose glass­es were first pre­dict­ed at the end of the 1980s, but have nev­er been observed unam­bigu­ous­ly in an exper­i­ment, even in sys­tems of cold atoms. Indeed, cold atoms are nev­er per­fect­ly cold and, even at tem­per­a­tures as low as a few bil­lionths of a Kelvin, ther­mal fluc­tu­a­tions can destroy quan­tum phas­es. How­ev­er, Lau­rent Sanchez-Palen­cia and col­leagues recent­ly pre­dict­ed a sit­u­a­tion in which, despite these ther­mal fluc­tu­a­tions, a Bose glass could be sta­bilised and observed. They are cur­rent­ly dis­cussing with their exper­i­men­tal col­leagues how to design an exper­i­ment in which they could actu­al­ly observe these exot­ic glasses.

These sys­tems are fas­ci­nat­ing in many ways. For exam­ple, they are often non-ergod­ic, as opposed to con­ven­tion­al ergod­ic. Ergod­ic sys­tems explore all the space at their dis­pos­al and can thus reach ther­mo­dy­nam­ic equi­lib­ri­um, a sit­u­a­tion well described by Boltz­man­n’s the­o­ry, devel­oped at the end of the 19^thCen­tu­ry. Indeed, the behav­iour of ergod­ic sys­tems is com­pat­i­ble with most of the obser­va­tions made to date on objects rang­ing from the size of a micron to that of stars and galax­ies. This the­o­ry is based on the idea that the sys­tem fluc­tu­ates between all pos­si­ble states along paths that allow it to move very quick­ly from one state to anoth­er. In non-ergod­ic sys­tems, on the oth­er hand, this ergod­ic­i­ty is pre­vent­ed by inho­mo­geneities in the sys­tem. The sys­tem remains there­fore trapped in a sub­set of its pos­si­ble con­fig­u­ra­tions, far from equilibrium.

Entan­gle­ment is anoth­er pure­ly quan­tum phe­nom­e­non of inter­est to the team. Here, two or more par­ti­cles can have much stronger cor­re­la­tions than clas­si­cal physics allows. For exam­ple, the observ­able prop­er­ties of a quan­tum par­ti­cle are gen­er­al­ly inde­ter­mi­nate, so that mea­sure­ment results are ran­dom. How­ev­er, when par­ti­cles are entan­gled, deter­min­ing the state of one par­ti­cle instan­ta­neous­ly fix­es the state of the other(s), regard­less of the dis­tance between them. This pow­er­ful “spooky action at a dis­tance”, as Ein­stein called it, seems to tran­scend space and time, so that we can deter­mine the state of one par­ti­cle sim­ply by mea­sur­ing that of its entan­gled part­ner. For exam­ple, if you mea­sure the spin of one par­ti­cle, say an elec­tron, you can deter­mine the spin of the oth­er with­out ever observ­ing it.

We are only just begin­ning to exploit the appli­ca­tions of this spec­tac­u­lar cor­re­la­tion-at-a-dis­tance effect, even though it is already used in the cryp­tog­ra­phy of cer­tain telecom­mu­ni­ca­tions. In sim­ple terms, sup­pose that the sender and receiv­er share an entan­gled pair, so that the results of their mea­sure­ments are ran­dom but iden­ti­cal. To inter­cept the com­mu­ni­ca­tion, an eaves­drop­per has to per­form a mea­sure­ment, the result of which is ran­dom but, more impor­tant­ly, changes the state of the pair, which is no longer entan­gled. The mea­sure­ments of the sender and receiv­er are no longer cor­re­lat­ed, and they can ver­i­fy this by com­par­ing the results of their mea­sure­ments. The strength of such an encryp­tion method is that it is not based on the dif­fi­cul­ty of spy­ing with­out being detect­ed, but on an impos­si­bil­i­ty based on the fun­da­men­tal laws of the quan­tum world.

Anoth­er appli­ca­tion of the ran­dom nature of the mea­sure­ment is the pos­si­bil­i­ty of mak­ing per­fect ran­dom num­ber gen­er­a­tors and total­ly ran­dom cryp­to­graph­ic keys.

These ran­dom num­ber gen­er­a­tors are already on the mar­ket and there is even at least one exam­ple of a mobile phone that uses quan­tum tech­nol­o­gy of this type. Quan­tum entan­gle­ment can also be exploit­ed in a process known as dense cod­ing, which is linked to the fact that an entan­gled state con­tains a phe­nom­e­nal amount of infor­ma­tion com­pared with the infor­ma­tion car­ried by each indi­vid­ual particle.

Each par­ti­cle con­tains infor­ma­tion about its own state, but the infor­ma­tion on cor­re­la­tions is dis­trib­uted among all the infi­nite­ly larg­er sub­sets of par­ti­cles. This effect makes it pos­si­ble to encode an immense quan­ti­ty of infor­ma­tion in struc­tures called quan­tum bits, or qubits. These dif­fer from stan­dard com­put­er bits, which can take the val­ue either 0 or 1. Qubits, how­ev­er, can take both val­ues at once, or any com­bi­na­tion of 0 and 1.

Qubits are the basic build­ing blocks of future quan­tum com­put­ers. Exploit­ing their quan­tum prop­er­ties, in par­tic­u­lar entan­gle­ment, can be applied to solve com­plex com­pu­ta­tion­al prob­lems, mak­ing it pos­si­ble to per­form cer­tain com­pu­ta­tion­al oper­a­tions much faster than the most pow­er­ful com­put­ers avail­able today. This would result in an expo­nen­tial increase in com­put­ing pow­er. Qubits can be fab­ri­cat­ed from a vari­ety of cur­rent­ly avail­able mate­r­i­al plat­forms, such as super­con­duct­ing qubits or trapped atoms and ions. Oth­er future meth­ods include pho­ton­ic quan­tum proces­sors that use light.

Spec­tac­u­lar progress has been made in recent years. How­ev­er, a real ‘quan­tum advan­tage’ is not expect­ed until quan­tum com­put­ers are oper­at­ing with – depend­ing on the esti­mate – between a few hun­dred thou­sand and a few mil­lion qubits. We only know how to build machines with around a hun­dred qubits today, so there is still a long way to go.

We only know how to make machines with a hun­dred or so qubits today.

The main obsta­cle to progress is quan­tum deco­her­ence. This results from the inter­ac­tion of qubits with their envi­ron­ment, which destroys their entan­gle­ment. To avoid this – or at least lim­it it – it is gen­er­al­ly nec­es­sary to oper­ate qubits at a tem­per­a­ture close to 0 kelvin and pro­tect them from the envi­ron­ment. From a fun­da­men­tal point of view, there is noth­ing to pre­vent the cre­ation of large-scale quan­tum com­put­ers, but there are still sci­en­tif­ic ques­tions to be resolved, such as whether deco­her­ence can be fun­da­men­tal­ly coun­tered. Impor­tant engi­neer­ing prob­lems also need to be over­come. Projects to solve these are being fund­ed by impor­tant gov­ern­ment pro­grammes and pri­vate invest­ment around the world.

In the short­er term, hopes are being pinned on machines that are less sen­si­tive to deco­her­ence. These are quan­tum sim­u­la­tors, which can be seen as ded­i­cat­ed quan­tum com­put­ers with an archi­tec­ture that opti­mis­es cer­tain spe­cif­ic tasks. Quan­tum sim­u­la­tors are par­tic­u­lar­ly well suit­ed to the search for ‘min­i­ma of mul­ti­vari­able func­tions’. Such machines are there­fore of inter­est to com­pa­nies that use com­plex net­works and seek to opti­mise them. These net­works con­tain a large num­ber of vari­ables and a quan­tum sim­u­la­tor could opti­mise them in a way that a con­ven­tion­al com­put­er cannot.

Most of the nec­es­sary tech­nol­o­gy already exists, but the ques­tion remains as to which com­put­ing prob­lems quan­tum sim­u­la­tion could be of real eco­nom­ic inter­est. These tech­nolo­gies are still very expen­sive and con­sume a lot of ener­gy. The ques­tion is more open than we might think, but cur­rent advances are paving the way for a host of rad­i­cal­ly new tech­nolo­gies. The mes­sage that Lau­rent Sanchez-Palen­cia is try­ing to get across to his stu­dents is that we need to think not only about quan­tum com­put­ers, but also about all the asso­ci­at­ed tech­nolo­gies. These may seem less spec­tac­u­lar at first glance, but they will very like­ly give rise to new technologies.

There is an urgent need for train­ing at all lev­els in the field of quan­tum mechan­ics. For many years, physics teach­ing at École poly­tech­nique has focused on this sub­ject. As a result, all stu­dents, includ­ing those who do not intend to go on to study physics or engi­neer­ing, are trained in this dis­ci­pline dur­ing their first year of study. We are already reap­ing the rewards: many of the start-ups being devel­oped in France today in the field of quan­tum mechan­ics are run by for­mer stu­dents of the École polytechnique.

In order to antic­i­pate the devel­op­ment of quan­tum tech­nolo­gies, the Quan­tum Sci­ence and Tech­nol­o­gy course was cre­at­ed a few years ago: it focus­es on the most mod­ern aspects of quan­tum physics, in par­tic­u­lar entan­gle­ment and how this effect can be exploit­ed. The course empha­sis­es the link between fun­da­men­tal sci­ence and tech­no­log­i­cal devel­op­ment, because at present, despite what we may read in the press, quan­tum physics is still at the devel­op­ment stage, and there is still much to understand.

The Mas­ter’s and PhD pro­grammes are run in close col­lab­o­ra­tion with the oth­er schools on cam­pus, with­in the Insti­tut Poly­tech­nique de Paris. They aim to recruit the best stu­dents from the best insti­tu­tions around the world. Stu­dents who join the insti­tute for a Mas­ter 1 direct­ly join a research team at IP Paris to rein­force links between train­ing and research. In addi­tion, IP Paris recent­ly start­ed offer­ing con­tin­u­ing edu­ca­tion cours­es for pro­fes­sion­al engi­neers who have nev­er stud­ied quan­tum sci­ence. This train­ing is also impor­tant for start-uppers who need very spe­cif­ic tech­ni­cal skills to devel­op par­tic­u­lar devices and for man­agers of com­pa­nies, whether they are SMEs or larger.

An impor­tant aspect of the devel­op­ment of quan­tum research at IP Paris is that it is being car­ried out hand in hand with the Uni­ver­sité Paris Saclay with­in the Insti­tut Quan­tum-Saclay. This allows the two insti­tu­tions to take advan­tage of their com­ple­men­tary strengths. At the nation­al lev­el, a new con­sor­tium was cre­at­ed last year, fund­ed by the French Nation­al Research Agency.

   Isabelle Dumé