Quantum Technologies: A Review of the Patent Landscape

Dra

ft

Quantum Technologies: A Review of the Patent Landscape*

Mathew Alex, Relecura Technologies Pvt. Ltd.email:[email protected]

October 10, 2021

Abstract

Quantum Technologies is a term that is getting broader with every passing year. Nan-otechnology and electronics operate in this realm. With the invention of industry-disruptingalgorithms like Shor’s algorithm that can break RSA encryption on a quantum computer andQuantum Key Distribution, which offers unconditional security in theory, investment is pour-ing in. Here we taxonomize and analyze 48,577 patents in this area from 2015 to presentcaptured with a comprehensive query in Relecura’s patent database. The author’s subjectexperience, along with the company’s AI-based tools and scholarly literature, were utilizedto make this highly subjective choice of taxonomy. Though most Patent Landscape AnalysisReports consider a single technology, geography, or company, we have tried to give a holisticoverview of these technologies as a whole due to their collaborative and intertwined nature.The physics of each technology and its role in the industry is briefly explained where possible.

Contents

1 Introduction 1

2 Literature Review 3

3 The Taxonomy 43.1 Overview . . . . . . . . . . . . . 63.2 Top Countries . . . . . . . . . . . 63.3 Top Assignees . . . . . . . . . . . 12

4 Top Technologies 124.1 Nanotechnology . . . . . . . . . . 12

4.1.1 Quantum Dots . . . . . . 144.2 Quantum Computing . . . . . . . 16

4.2.1 Quantum Annealing . . . 174.2.2 Error Correction . . . . . 18

4.3 Quantum Cryptography . . . . . 184.3.1 Quantum Key Distribution 19

5 Emerging Technologies 205.1 Superconducting Devices . . . . . 205.2 Quantum Sensing . . . . . . . . . 21

6 Conclusions 21

7 Acknowledgements 22

References 22

1 Introduction

The extinction of dinosaurs possibly gifted ourancestors a jackpot offer to climb the foodchain to dominate the earth. Over the lastfew centuries, we made unprecedented advancesno other species could ever dream of; to-day, we possess the technology to destroy ourplanet several times over and cure the largestpandemic ever. History closely followed ad-vancements in science, often shaping ideologies,rather than the other way around.

The technological advancements were so-cial revolutions of their time. The steam en-gine and the internal combustion engine maderaw mechanical power so cheap that the econ-omy freed humans to work on more creativepursuits. The information revolution acceler-ated by the transistor’s discovery made com-puters fast and affordable that they enabled usto advance research and industries and to col-laborate internationally to create a truly unified

∗Work in progress.

1

arX

iv:2

102.

0455

2v1

[cs

.DL

] 3

Feb

202

1

Dra

ft

global community. Yet, by the end of the lastcentury, we bumped into a few roadblocks thatcould circumscribe this saga of progress.

To keep up with the exponential pace ofinnovation, we need faster computers. Dennardscaling (Dennard, Gaensslen, Yu, Rideout, Bas-sous, and LeBlanc [1974]) states that shrink-ing the features of a chip allow running it at ahigher clock speed for the same power; in otherwords, power demand scales with the area whileperformance scales with the number of transis-tors. Some ingenious ideas always redeemed thegrowth of computational power through pack-ing more transistors per area every time in his-tory when it was about to fail. CMOS tech-nology took over when scaling bipolar transis-tor logic became impractical. When CMOSplateaued, Dennard scaling ended in 2004, andmulticore technology met the performance de-mand (Shalf and Leland [2015]). As state ofthe art moves from 7nm to 5nm in 2020, theperformance boosts are not stellar, and leak-age currents due to quantum tunneling are be-coming an extensive hassle to circumvent as weprobe deeper. One quantum technology maybe the answer to this roadblock that was antic-ipated decades ago, at least for certain types ofcomputation.

The second roadblock is more profound,thrusting deep into the quantum nature of theuniverse. The memory requirement of com-puter simulations of quantum systems expo-nentially grows with the number of particles(or degrees of freedom), making such systemsimpossible to model even with supercomput-ers. Many emergent phenomena like high-temperature superconductivity remain obscuredue to their quantum description’s complex-ity and the impasse with numerical simulations(Tangpanitanon and Angelakis [2019]). Somescientists are even moving to machine learn-ing approaches to go beyond the traditional ap-proximation techniques where they failed to il-luminate important unexplored problems (Car-leo and Troyer [2017]).

One of the most promising solutions tothese problems came from quantum mechan-ics itself: rather than trying to mitigate quan-tum effects, we can use them to our advan-

tage. If quantum many-body systems involvesuch vast amounts of information, why not usethem to store this information? Richard Feyn-man was one of the first scientists to point thisout- “Let the computer itself be built of quan-tum mechanical elements which obey quantummechanical laws. ” (Feynman [1982]). A quan-tum computer meant that we get access to na-ture’s tremendous storage capacities, her knackfor executing physical phenomena in an instantdespite the complexity in describing them, andworking with the fundamental building blocksof the universe.

A closely related field, which too wascoined by Feynman in his lecture ”There’sPlenty of Room at the Bottom: An Invitationto Enter a New Field of Physics,” 1959 (Feyn-man [1960]) is nanotechnology. It involves ma-nipulating materials at the scale of nanometers,where the quantum nature of matter is appar-ent. The efficiency and prowess of biologicalprocesses come from the molecular machinesfacilitating them and the nanostructures theyuse; nanotechnology is a step in this lead’s di-rection (Grzybowski and Huck [2016]).

Innovation is driven by the synergy be-tween academia, industry, and the economy,and Quantum Technologies are no different.Patents are an excellent indicator of the paceof innovation, and they can be subjected to awide variety of analyses based on the biblio-metric data. Such statistical analyses are in-valuable to researchers and policymakers be-cause patents can quantify otherwise difficultto measure phenomena like technological col-laboration, the evolution of the technologicalspace, geographical and company-wise predis-positions, etc. Increased patenting and theonset of computerization of the field resultedin reports often exploring thousands or evenmillions of documents. The patent analysismethodology usually varies between reports de-pending on the context and individual prefer-ence (Grant, Van den Hof, and Gold [2014]).

The difficulty with a patent landscapeanalysis of Quantum Technologies is multifold.Usually, such reports often focus on a singletechnology or geography. But there is no cleanway to cut away quantum communication from

2

Dra

ft

quantum cryptography or nanotechnology fromsemiconductor devices, at least for this set ofpatents. Moreover, the forefront of R&D inthis field cannot possibly be covered by stick-ing to a single geography or company. NorthAmerica is leading research in quantum com-puting, thanks to the US’s tech giants investedin this field and D-Wave in Canada, the firstcompany to sell computers exploiting quantumeffects. China is making news frequently forredefining the limits of quantum communica-tion and cryptography. The considerable over-lap between some technologies is a hindranceas well that we had to drop some nodes (seeSection 3.1).

In this publication, we try to give a holis-tic account of Quantum Technologies, startingfrom analyzing the patent landscape and mak-ing connections to scholarly literature as oftenas needed. We could not find any study in-corporating Quantum Technologies as a whole,including Quantum Information, Nanotechnol-ogy, Electronics, etc., in patents and scientificresearch contexts. Yet, some excellent reviewsweighing on different techniques have come be-fore, and this publication has benefited greatlyfrom them.

2 Literature Review

Bruck, Rethy, Szente, Tobochnik, and Erdi[2016] used a citation-based recursive patentranking based on PageRank algorithm(Page,Brin, Motwani, and Winograd [1999]), the firstalgorithm Google used to rank webpages, torank patents according to the information flowthey contributed to through forward citations.They studied the results for different valuesof damping factor, d, a measure of how mucheach node benefits from indirect citations, andobserved that d = 0.5 yields best results forpatents rather than 0.85 used by Brin and Pagefor webpages. A web surfer may go up to,say, ten webpages following hyperlinks, but re-searchers are not usually interested in patentscited beyond two levels. Hence, this lower valueof d is intuitive. Apart from pulling the mostrelevant patents from the USTPO database us-

ing PageRank values, they interpreted patentsas a time-evolving complex system in whichnew technologies emerge as recombination ofthe old on a mesoscopic scale. Specifically, thepaper established that laser/inkjet printer tech-nology emerged from existing sequential print-ing and static image production technologiesby studying the interaction between US Classesthrough citations over the years.

While we resorted to subject matter ex-pertise to taxonomize the patents, automatedpatent landscaping is the future, freeing the ex-perts’ valuable time for research. The manualprocess is tedious and expensive, requiring thesubject expert to go back and forth betweenthe query and the result multiple times to cap-ture all the valid patents. Choi, Lee, Park, andChoi [2019] proposed an ingenious automatedpatent landscaping model using deep learningthat achieved the state of the art classifica-tion performance. The model used Graph Em-bedding with Self Clustering (GEMSEC) algo-rithm (Rozemberczki, Davies, Sarkar, and Sut-ton [2019]), which preserves clustering whileembedding the graph, to encode the metadata,CPC, IPC, and USPC codes in this case. Theabstract is handled by a transformer encoderlayer (Vaswani, Shazeer, Parmar, Uszkoreit,Jones, Gomez, Kaiser, and Polosukhin [2017])to learn the latent space of a word2vec embed-ding. They concluded that as it is in man-ual classification, although context-dependentin general, the CPC codes, the most detailedamong the three bibliometrics, guarantee bet-ter classification performance.

Bettencourt, Trancik, and Kaur [2013]conducted an interesting study about patentingin energy technologies that answers the fast-paced innovation in renewable energy in the2000s despite sustained low funding. To ac-count for this observation, they introduced anon-linear relationship between the cumulativequantities patents P , public R&D R, and pro-duction levels C as P = P0R

αCβ , where P0, α,and β are constants. This model accounted forthe discrepancy as a combined effect of publicR&D and increasing market.

Wu, Zhang, Lv, Liu, Yu, Zhao, Chen, andMa [2019] used a neural network to do data-

3

Dra

ft

driven technological forecasting for high-techcompanies, which is often a difficult decision forcompanies to make on human instincts alone.Future research and R&D directions are in-fluenced by a company’s present standing andweaknesses in the patent landscape, sensitivityto their competitors, and market forces. TheDeep Technology Forecasting (DTF) frame-work proposed in this work can predict thecomplex interactions between companies andtechnologies by quantitatively analyzing theirstrengths and weaknesses in the patent land-scape, identifying competitor companies andtheir filing trends, and the collaborative rela-tions among technologies.

Kurek [2020] extensively studied patentsin Quantum Technologies, emphasizing the eco-nomics, geographies, key players, and their in-novation areas. The publication quantitativelyargued that China is leading on Quantum Com-munications while the US remains the leaderin technologies related to the manufacture ofquantum computers. The State Grid Corpo-ration of China, the world’s largest electricitydistributor, is taken as an example to illustrateChinese advances in quantum communication.

A more physics-centered review of Quan-tum Computing can be found in Preskill [2018],although quantum communication, cryptogra-phy, and sensing are cut out. The author, alsoa top scientist in the field, defined the cur-rent state of the art as Noisy Intermediate-Scale Quantum (NISQ) computers and arguedthat physics simulations might be their onlyimmediate use. He encouraged ’optimism tem-pered with caution’ about this technology, andwe closely subscribed to his arguments in thiswork.

This report is a middle ground betweenthe works mentioned in the last two para-graphs, attempting to balance the analysis be-tween patenting trends and physics. We willexplore the patents’ overall trends in the nextsection, including an overlap study betweenfirst-level nodes and a short discussion of topcountries and top assignees. We used a cou-ple of techniques to evaluate the quality of USand Chinese patents; this was necessary sinceChinese patents alone account for half of the

patents dominating all the statistics. In Section4, we centered the discussion the top technolo-gies from the point of view of top assignees ineach field. We also provided a short descriptionof emerging technologies identified by Relecurain Section 5.

3 The Taxonomy

Some experts classify quantum technologiesinto two generations(Dowling and Milburn[2002]). The first generation includes tran-sistors, lasers, etc., which work on quantummechanical principles. The second generationcomprises quantum computers, quantum com-munication, etc., where we actively manipu-late the quantum mechanics underlying the uni-verse for our advantage. This entire article isabout the second generation of quantum tech-nologies, which were difficult to separate fromtheir predecessors. In a broader sense, eventhe incandescent bulb work on a quantum phe-nomenon, the black-body radiation (the prob-lem that gave birth to quantum mechanics).

Quantum technologies is an umbrella termhosting an arsenal of technologies penetratingevery field and finding new applications everyday. With the advances in big data analysisand machine learning, we can analyze quan-tum technologies’ patent landscape holisticallywith limited manual labor. Here we examinethe 44, 394 patents from this area filed in thelast five years, captured using a comprehensivequery in Relecura’s patent database. We donot claim these patents to be exhaustive, butthe tolerances were reasonable. Irrelevant doc-uments that may be present are not enough tosabotage the overall statistics and the followinganalysis. We avoided brute force subtractionbased on keywords to keep the taxonomy re-ceiving and unbiased towards patents and tech-nologies yet to come.

4

Dra

ft

Figure 1: Quantum Technologies taxonomy created in Relecura. See this link for the up to date interactive web version.

Dra

ft

3.1 Overview

The taxonomy has a significant overlap betweenfirst-level nodes (see Figure 2). The Nanotech-nology and the Electric Elements nodes sharemost patents, and Quantum Computing, Quan-tum Communication, and Quantum Cryptogra-phy nodes have an almost symmetric overlapbetween each pair. A subject expert will re-joice in the patent sets’ overlap reminiscent ofthe concepts they represent. Modern electron-ics operate in the nano realm, with new ad-vances in nanotechnology replacing traditionalcircuit elements. Security is a built-in fea-ture in quantum computation and communica-tion where the collapse of the quantum stateon measurement and the no-cloning theoremforbids copying data and eavesdropping (seeQuantum Cryptography). Advancements inquantum communication and cryptography willfacilitate the development of the quantum in-ternet, an ideal way to connect quantum pro-cessors.

The organization of this discussion willkeep these overlaps in its heart. We will takethe Nanotechnology node to be representativeof Electric Elements as well. The differentchild nodes are like looking at these patentsfrom a different perspective. The patents in theSuperconducting Devices child node are uniqueto Electric Elements, and we will discuss themin Section 5. Together with the QuantumComputation, Communication, and Cryptogra-phy trio, these are the main patent clusterswhere innovation is proceeding rapidly. Quan-tum Sensing , too, is an important node, whichwe will discuss last.

The concept map (see Figure 3) for thispatent set generated using Relecura’s con-cept extractor indicates that almost all of thepatents utilize quantum mechanics in someform. The largest bubble, Quantum Mechan-ics, is connected with 89% of the patents, vali-dating this taxonomy.

Figure 2: Overlap between first level nodes

3.2 Top Countries

China surpassed the US in the number of in-ternational patents filed annually in 2019. TheUS has been the center of almost all technolog-ical revolutions, from the incandescent bulb tothe Silicon Valley revolution and the top sourceof patents since the international system wasestablished. The patent numbers in quantumtechnologies, when considered in isolation, sug-gests a Chinese dominance. China holds severaltimes the patents the US has in every first-levelnode except for Quantum Computing, wherethe difference is only a few hundred patents.China is at the forefront of innovation in quan-tum communication, and cryptography and theUS still hold its edge in quantum computing.

6

Dra

ft

Figure 3: Concept Map

Figure 4

Some resources suggest that most Chinesepatents are of inferior quality(Kurek [2020]) be-cause researchers in China patent more at theexpense of novelty to pass the strict objectiveevaluation of their performance. We used Rele-cura star rating, a custom rating system thatconsiders key patent parameters like forwardand backward citations, geographical and fam-ily values, etc., to study this observation. Therating ranges from 0 to 5 in increments of 0.5.Chinese patents caps at 3.5, whereas the US haspatents up to 5. A normal distribution fit on

the star ratings (see Figure 6 ) arguably revealsthat US patents are more valuable. Around68% of a normal distribution is concentrated ina standard deviation distance from the meanvalue. For US patents, it is from 1.37 to 2.81,and in the case of China, it is from 1.13 to 2.27.

One of the crucial factors determining theimportance of a patent is the forward citationsoriginating from it. The 8,935 US patents inthe taxonomy account for 71,755 forward ci-tations, whereas 23,360 Chinese patents have44,612 forward citations. A patent citing aprevious patent means that the citing patentis building on the knowledge available throughthe cited patent.

The citation network provides a represen-tation of the innovation process. Although theprimary technique for evaluating this knowl-edge diffusion, there are concerns over using thecitation network, particularly the potential biasor noise added by examiner citations and self-citations that reflects the inventor’s knowledgerather than knowledge flow from other patents(Alcacer and Gittelman [2006]). But even withthis noise, the citation data can be a goldmineof information if used with discretion (Erdi,Makovi, Somogyvari, Strandburg, Tobochnik,Volf, and Zalanyi [2012]).

7

Dra

ft

(a) a)US (b) b)CN

Figure 5: Concept maps of US and Chinese patents.

We will use the PageRank (PR) algorithm(Bruck et al. [2016]), which Google used in itsinitial days to rank websites, to study the cita-tion network. When the internet was a messyplace, Larry Page and Sergey Brin came upwith this recursive algorithm, which rewardsnodes (webpages) that are well-connected toother reliable nodes. It is extensively usedto study patents after the algorithm’s patent(Patent No. US6285999B1) expired in 2019.PR is recursively computed for the network us-ing the formula

P(t+1)i =

(1− d)

N+ d

ni∑j=1

P(t)j

mj

where P(t)i is the PR value of node i at it-

eration t, N is the total number of the nodes inthe network, ni is the count of incoming linksof nodei, mj is the count of outgoing links ofnode i and d is called the “damping factor”.d controls the contribution to a node throughindirect citations and it is set to 0.5 to studypatents rather than 0.85 used for ranking webpages. This is because on average a web surfermight move through tens of web pages through

hyperlinks but the interest in patents does notusually go beyond citations of citations (Brucket al. [2016]). We will be using this well testedvalue that obtained good results in previousstudies, setting the iteration convergence tol-erance to 10−8.

We have plotted the network of US andChinese patents with the nodes color-coded ac-cording to PR (see Figure 8). Evidently, the USgraph form more localized clusters with higherPR, whereas the Chinese graph is more frag-mented with lower PR values for the scatteredclusters. The drastic difference may be due tothe variety of technologies in which patents arefiled in China. The concept maps for the USand Chinese patents (see Figure 5) is open forsuch an interpretation. Chinese patents aremore evenly distributed among the top con-cepts when compared to the US patents.

The same algorithm is used to pull out thetop 20 patents from the taxonomy (see Table1), and 18 of them are from the US, and Chinais absent from the list.

8

Dra

ft9

(a) (b)

(c) (d)

Figure 6: Countrywise contribution of patents to major first-level nodes.

(a) US

Figure 7: Star rating distribution of US and Chinese patents fitted with a normal distribution.

Dra

ft10

(a) US

(b) CN

Figure 8: Patent graph structure illustrating forward citations. In this plot, the node size isproportional to number of connections, and warmer colors mean a higher PageRank score.

No. Publication Number Title Citations PageRank ×108 Country Code

1 US9199842B2 Quantum dot films, lighting devices, and light... 700 124623 US2 US9037247B2 Non-invasive treatment of bronchial constriction 301 106395 US3 US9049010B2 Portable data encryption device with configura... 297 104505 US4 US9000353B2 Light absorption and filtering properties of v... 527 98960 US5 US9430078B2 Printed force sensor within a touch screen 291 95658 US6 US9019595B2 Resonator-enhanced optoelectronic devices and ... 497 89690 US7 US9115348B2 Endoribonuclease compositions and methods of u... 351 88545 US8 US10000788B2 Rapid and sensitive detection of molecules 254 85708 US9 US9899123B2 Nanowires-based transparent conductors 255 84708 US10 US9666702B2 Advanced heterojunction devices and methods of... 342 81985 US11 US10572684B2 Systems and methods for enforcing centralized ... 223 77716 US12 US8927968B2 Accurate control of distance between suspended... 217 75789 US13 US9457139B2 Kits for systems and methods using acoustic ra... 186 62477 US14 US9178123B2 Light emitting device reflective bank structure 337 59275 US15 US9232618B2 Up and down conversion systems for production ... 178 58059 US16 US8932940B2 Vertical group III-V nanowires on si, heterost... 210 56184 US17 US9006704B2 Magnetic element with improved out-of-plane an... 248 53328 US18 WO2009039854A8 MHC multimers in tuberculosis diagnostics, vac... 145 51330 WO19 US9590089B2 Variable gate width for gate all-around transi... 162 50451 US20 KR101635835B1 Coating method with colloidal graphine oxides 140 49776 KR

Table 1: Top 20 patents in the taxonomy according to PageRank with d = 0.5.

Dra

ft

3.3 Top Assignees

The top assignees in the entire patent set arewell-known companies. Most of them are herein this category because of the variety in theirresearch. There are four US companies (Alpha-bet, IBM Intel, and Northrop Grumman), threeChinese companies (BOE Technology, ChineseAcademy of Sciences and TCL Corporation),two Korean companies (Samsung and LG), anda single Japanese company (Toshiba) in the top10 assignees.

Figure 9

The US companies are almost centered onthe Quantum Computing node with an over-flow to the Nanotechnology node from their in-vestment in the semiconductor industry. 435 ofAlphabet’s 505 patents and 700 of IBM’s 876patents go to Quantum Computing, which isnot a surprise as they are the industry leadersof the field. Intel has got 459 patents in Nan-otechnology and 280 in Quantum Computing ;they are working on building quantum proces-sors collaborating with QuTech in the Nether-lands. Northrop Grumman, one of the world’slargest defense technology providers and makerof the renowned B-2 stealth bomber, is ac-tively exploring new disruptive quantum tech-nologies. Their portfolio is concentrated inQuantum Computing (283) and Superconduct-ing Devices (388 patents).

TCL Corporation (the makers of Black-

berry phones from 2016 to 2020), which manu-factures mobiles display panels etc., is patent-ing heavily in QLEDs (704 patents) and Fabri-cation Technologies (528 patents). BOE Tech-nology, one of the world’s largest producerof LCD, OLED, and flexible displays, has asimilar patent profile (434 patents in Fabrica-tion Technologies and 570 patents in QLEDs).Chinese Academy of Science, the only aca-demic institution in the top 10 assignees, hasa well-balanced portfolio in quantum technolo-gies with patents in all first-level nodes.

LG is more swayed towards Semiconduct-ing Devices and Nanotechnology, which is ex-pected since its primary trade is consumer elec-tronics. The other Korean giant in top as-signees, Samsung, shows similar trend exceptthat the numbers must be scaled. Almost all ofits patents come under Nanotechnology.

Though Toshiba is known for its consumerelectronics products, it has diversified its busi-ness model to include IT solutions like quantumcryptography in 2020. They hold hundred pluspatents in Quantum Computing, Communica-tion, and Cryptography.

4 Top Technologies

This report cannot possibly cover every leaflevel node. We have established that Nanotech-nology and Quantum Computing, Communica-tion, and Cryptography are the major first-levelnodes in this taxonomy. Nanotechnology oreven its child node Quantum Dots taken in iso-lation, account for more than half of the entirepatent set. We follow a weighted approach inour analysis, with these nodes taking the mostprecedence and detailed discussion.

4.1 Nanotechnology

The most populated and dynamic node inthis taxonomy is Nanotechnology, and the low-dimensional nanostructures (Quantum Dots,Wells, and Wires) dominate patents in thisnode. In a way, this node captures the spiritof quantum technologies.

The development of semiconductor tech-nologies in the last century was mostly limited

12

Dra

ft

by the difficulty in synthesizing materials withthe desired energy levels. When you choose amaterial for its energy levels, you have to livewith its chemistry. LEDs were a revolution inlighting technologies with their power efficiencyand long life span, but this innovation was bot-tlenecked by the delay in making a blue LED.The first blue-violet LED was developed us-ing magnesium-doped gallium nitride at Stan-ford University in 1972. Isamu Akasaki, HiroshiAmano, and Shuji Nakamura were awarded theNobel Prize in Physics in 2014 for the inventionof the blue LED by growing high-quality Gal-lium Nitride crystals and creating p-type chan-nels in it.

Figure 10: Top Assignees in Nanotechnology

The addition of blue LED completed thecolor spectrum with the already available redand green LEDs, and this trio made its way intoLED screens and the lighting industry (thankblue LEDs you can read this article). The avail-ability of high wavelength (and energy) blueLEDs meant that other colors could be gen-erated by phosphors or combining them withred and green LEDs. This invention lightenedthe power grid load and expedited solar energyadoption because of its power efficiency. But it

took decades to go from blue-violet LED to apure blue LED.

Nanotechnology, the engineering of func-tional systems at a molecular scale, offers a dif-ferent approach to this difficulty of synthesiz-ing crystals for their energy levels. The elec-trons in a crystal see a practically infinite peri-odic lattice, and the solution of the Schrodingerequation (the differential equation that gov-erns quantum behavior) for this system is en-ergy bands typical semiconductors. However,when the charge carriers are confined in one ormore dimensions to the order of their de Brogliewavelength, the scale at which quantum effectsdominate, they will know that the periodicityis gone, and the bandgap changes.

Confinement doesn’t always mean that thematerial is of the length scale of carrier con-finement. Quantum wells (2-d nanomaterials)are grown as alternate layers of semiconductorswith different bandgaps. The progress of thisfield was entirely due to the developments in ad-vanced crystal growth techniques like molecularbeam epitaxy and microscopy techniques likeScanning Tunneling Microscope, which pro-vided unprecedented visualization of atoms andbonds, and as demonstrated in 1990, capableof moving individual atoms around (see A Boyand His Atom to know how proud IBM is ofthis technology).

While research in the past took the quan-tum energy levels as absolute, nanotechnologygoes to the next level to make novel materi-als with tunable electrical, optical, and physi-cal properties. Devices crafted like this are ag-gressively replacing traditional circuit elementswith their futuristic efficiency and speed.

The low dimensional nanostructures arean exciting area of research working at the in-tersection of physics and chemistry. They op-erate at the mesoscopic scale, between micro-scopic and macroscopic laws of physics, largerthan a few electrons or molecules but smallenough that their degrees of freedom must betreated fully quantum mechanically. Scientistsare dealing with hardcore quantum mechanicswith all its quirks here. The research in thisfield has even unveiled many things about themovement of electrons in traditional devices.

13

Dra

ft

4.1.1 Quantum Dots

Low dimensional nanostructures (32,529patents), particularly quantum dots (25,058patents), absolutely crushes everything elsein sheer numbers in our taxonomy. Since itaccounts for more than half of the patents, wewill go down this node, most representative ofnanotechnology and quantum technologies ingeneral, to discuss its applications.

Lower dimensional materials are mate-rials restricted to atomic scales in one ormore dimensions (Geller [2001]): graphene istwo-dimensional, and a quantum dot is zero-dimensional by this definition. Confinement ofcharge carriers in such structures leads to ex-otic electronic and optical properties tunableby their size and composition. For example,quantum dots (QD) can be adjusted to fluo-resce from blue to red by increasing the par-ticle size. They find potential applications inLEDs, displays, lasers, and biological imaging.We will showcase a few important applicationsof QDs here.

1. Biological Tagging and Labelling Inbiological tagging, QDs have significantadvantages over conventional fluorophores(a fluorescent tag that selectively binds toa specific region or functional group on thetarget molecule). The Green FluorescentProtein (GFP), naturally occurring in aparticular jellyfish, and other chemicallysynthetic fluorescent dyes were the mostcommon biological tagging methods. QDsshine here due to their efficiency in cap-turing light to give a brighter image forthe same irradiation compared to tradi-tional fluorophores. Thanks to their in-organic origin, they don’t undergo photo-chemical degradation from prolonged ex-posure to the excitation source. Their size-tunable nature allows a complete gamut ofcolors. QDs find use in drug discovery, sin-gle protein tracking, and disease detection.Maybe someday, we will have the ability tolight up any biological phenomena in anycolor scheme!

Since this is still an experimental technol-ogy, most assignees are academic institu-

tions: University of Jinan (31 patents),Duke University (23 patents), Harvard (23patents), US National Institute of Health(21 patents), and French National Cen-tre for Scientific Research (18 documents)are among the top 10 assignees. NanocoTechnologies (41 patents) and ShenzhenFortense (31 patents) are the top two as-signees.

2. Image Sensors QDs make excellent pho-todetectors, and QD image sensors are onthe way. Traditional CMOS sensors cap-ture the electrons ejected from the semi-conductor pixels through metal contactsand traces, which reflect part of the light,decreasing efficiency and contributing topoor low light performance. The solutionto this is the back-illuminated sensor, withreadout electronics under the detector, butit costs way more than its predecessor.

Smartphone cameras have developed sofast in the past few years that DSLRs onlyhave an edge under challenging shots thatdo not matter to most consumers, but thesoftware has been doing most of the heavylifting. Google thrived on computationalphotography to establish the Pixel lineuplaunched in 2016; their software was sogood that they haven’t upgraded the imagesensor (Sony IMX 363) for their last threegenerations of phones. But all of its com-petitors caught up in 2020 that the differ-ences between brands are subtle and com-putational photography has plateaued.

QD image sensors have the potential totake the industry from here. A QD sen-sor is fabricated in the same way ex-cept for the silicon photodetector part; in-stead, QDs are suspended in a solution andsprayed onto the wafer, reducing the cost.This setup offers the benefit of a back-illuminated sensor without its fabricationcost. The photoelectron can hop fromone QD to another to reach the nearbyelectrode. A thin layer of QD is enoughbecause it can absorb light much betterthan silicon. They function well in lowand high light scenarios offering a bet-

14

Dra

ft

ter dynamic range(Palomaki and Keuleyan[2020]). Due to their size-tunable nature,QDs can be used in infrared cameras aswell. Since silicon’s bandgap is higher thanthe energy of infrared light, present IRcameras have to go through the added has-sle of integrating a different semiconductorinto the silicon wafer. QD sensor is easy tomanufacture for both IR and visible rangeapplications.

Samsung (67 patents), Canon (60 patents),and Fujitsu (26 patents), all of whom man-ufacture cameras in one form or another,are the top three assignees in QD imagesensors. The third top assignee is ty-ing with Fujitsu, InVisage Technologies,founded by Ted Sargent, who pioneeredQuantumFilm technology when he was aProfessor at the University of Toronto.QuantumFilm is the suspended version ofQDs to be coated on the image sensormentioned previously. The industry washeated up when Apple, a lowkey player inthis area but the company which nailed thecamera performance in every iteration oftheir phones, acquired InVisage. AlthoughApple is silent about its R&D, an iPhonewith a QD image sensor is likely to releasein the near future to claim back the cameraindustry for hardware.

3. QLEDs

Although quantum dots are an emergingtechnology, the industry deemed it ripe forcommercial displays. Several companieslike Samsung are aggressively marketingQLED TVs as the next big thing. Maybe,this might be the only piece of second-generation quantum technology that youcan buy right now.

Out of the 3,138 patents listed underQLEDs, Samsung dominates with 354patents, closely followed by BOE Tech-nology (307 patents), Shenzhen ChinaStar Optoelectronics (192 patents), TCLCorporation (121 patents). HiSense (58patents) and TCL producing QLED TVsto compete against Samsung. Even LG(104 patents), who denounces QLEDs in

their website, labeling it as a gimmickyLCD technology, is actively doing R&D inthis area.

Even though both OLED and QLED areimprovements over the LCD technology,it is hard to crown one of them better.OLEDs can produce light of their to de-liver practically infinite contrast and an al-most 180-degree viewing angle. They canbe manufactured as transparent or flexibledisplays (the reason for the recent onset offoldable phones and bezel-less phones thattuck the display under the visible screen toattach the controls). However, OLEDs areexpensive to manufacture as large panels.The cost will increase as we run out of rareearth elements required, and the screens’longevity is a serious issue.

QLED displays are cheaper than theOLED ones, and they are more power-efficient, lasts like LCDs, and offer morecolor purity and a wider color gamut.While QLED is a fantastic comeback of thebacklit technology and offers healthy com-petition against OLEDs in the TV mar-ket, they are unlikely to make their wayinto smartphone screens due to the avail-ability of flexible OLED screens and theadded weight of the backlighting setup.Yet, these corporations’ serious R&D in-vestment suggests that they can evolveinto the next big thing in time.

However, the cost of manufacture is a sig-nificant hindrance to the early adoption of thistechnology, and the toxicity of chemicals in-volved (for example, Cadmium) is another hit.The EU has issued a temporary relaxation ofRoHS (Restriction of Hazardous Materials) forquantum dots, which will expire soon. Biolog-ical applications are limited from mass adop-tion due to the toxicity factor. An activelyresearched solution is to design QDs to evadebreaking down by immune cells to avoid longterm retention of heavy metals in the body. Ei-ther way a non-Cadmium based QD is urgent(Jin, Hu, Liu, and Wu [2011]). QDs can oxidizein the air, which is not a problem in encloseddisplay panels, but must be taken into account

15

Dra

ft

when designing image sensors. In the mean-time, accelerated research in quantum dots isbreaking these barriers and finding them moreniche uses.

4.2 Quantum Computing

This node might be the most magical realmin all quantum technologies with paradigm-shifting implications if ever realized commer-cially. Quantum computing uses the universe’se fundamental building blocks (single atoms,electrons, protons, etc.) as computational el-ements. However, they cannot execute irre-versible gates like AND and NOT, and theyare never meant to replace consumer comput-ers, even though they are Turing complete. Sofar, only a few algorithms are available that canmanipulate quantum mechanics to achieve ex-ponential speedup over classical computers. Aquantum computer employing Shor’s algorithmcan factor prime numbers used in RSA cryp-tography (RSA numbers), while it is virtuallyimpossible for a classical computer in reason-able timescales. Grover’s algorithm can com-plete an unstructured search in a database ofN elements in a time propotional to

√N . It

will be an awful waste of resources to code aquantum computer to play a movie.

Figure 11: Top Assignees in Quantum Com-puting

Quantum computers ingeniously usequantum superposition, the property of

quantum systems to occupy multiple statesweighted with different probabilities, and en-tanglement, the ability of spacially separatedentangled particles to act on each other, toparallel process vast amounts of data. AnN qubit array can store 2N numbers andparallel process them efficiently(Jozsa [1997]).Quantum computers can evaluate functionsparallelly by encoding the inputs as a super-position state and evaluating once. However,we can not assess all of the outputs as onemeasurement will collapse the state. Eventhough the information is there, we need tocome with algorithms respecting the quantumweirdness to take advantage of it.

Quantum computers work best in prob-lems that are easy for a classical computer toverify an answer’s truth, like factoring a com-posite number. The Shor’s algorithm on aquantum computer offers an exponential speedup to the factorization problem. The RSAcryptography system, which relies on the dif-ficulty of finding prime factors of a compos-ite number, is at stake if this algorithm scalesfrom its record of factoring the number 35 (Am-ico, Saleem, and Kumph [2019]). A quantumcomputer employing a few thousand qubits canbreak the RSA and Elliptic-Curve Cryptogra-phy (ECC) systems.

Niche algorithms have been developed forother mathematical problems as well that quan-tum computers changed the paradigms of clas-sical computing; they even introduced a newcomplexity class (see BQP) in the theory ofcomputation. Even though scientists cameup with revolutionary algorithms decades ago,they have not materialized on a real quantumcomputer to have any physical impact as of yet.The accelerated research in recent times is try-ing to mitigate this gap. The most numberof qubits in a quantum computer is less than100 as of now, and this must shoot up to thehundreds and the thousands for quantum algo-rithms to change the world. But still, they donot pose a challenge to personal computers be-cause only a handful of algorithms are available,offering quantum speedup in very specific prob-lems. Quantum computers are likely to reachout to the world through the cloud as most

16

Dra

ft

of the current realizations operate at cryogenictemperatures, requiring bulky cooling systems.

The best in the business is taking up thischallenge. Even though the world’s largestcloud service provider by market share, Ama-zon, does not have a strong presence in thistaxonomy, they are well placed in the busi-ness performing an integrating role between dif-ferent quantum processors developed by othercompanies. Amazon Bracket, released lastyear, allows customers to build and simulatequantum algorithms on simulators as well ason different quantum computers. They offerD-Wave’s quantum annealing processors (seeSection 4.2.1) and gate-based computers fromRigetti and IonQ.

The big names in the computing industryhere competing against each other, often col-laborating with smaller companies, to be theleader in this technology of the future. Outof the 7,867 patents, the patent champ IBMholds 700 patents to lead in this area. On theirshift to research and consulting services, theykept close to quantum computing to create thebest python library Qiskit, which is widely usedin academia, to simulate quantum circuits andexecute them on IBM’s quantum processorsthrough the cloud. Google offers Circ-q, andMicrosoft offers Q# as competition, but noneof them provide access to real quantum com-puters. Plus, IBM collaborates with fortune500 companies and top universities through theIBM Quantum Network program for acceler-ated research and early commercialization ofquantum computing solutions. The next con-tenders are Google (434 patents), NorthropGrumman (282 patents), Intel (280 patents),and D-Wave Systems (231 patents).

Most of these top assignees work toachieve quantum supremacy, i.e., to demon-strate that a quantum computer can solve aclassically intractable problem. Google claimedin October 2019 that it achieved quantumsupremacy using a new 54-qubit processornamed “Sycamore”(Arute, Arya, Babbush, Ba-con, Bardin, Barends, Biswas, Boixo, Brandao,Buell, Burkett, Chen, Chen, Chiaro, Collins,Courtney, Dunsworth, Farhi, Foxen, and Mar-tinis [2019]). However, this is disputed by IBM

and other researchers in the field. Even thougha milestone in computing, quantum supremacyhas a relaxed definition: the algorithm does notneed to perform anything useful, which meansthat it can be achieved before significant ad-vances in error correction. It is likely to beachieved soon, if not already.

Another specialty of this area in the patentlandscape is the presence of universities. Alongwith the big names like MIT, Caltech, etc., Chi-nese and Korean institutions are filing a com-parable number of patents to industries. TheUS is increasing funding for quantum comput-ing research at universities and the academiccontribution is expected to grow.

4.2.1 Quantum Annealing

’Annealing’ means to heat a material past itsrecrystallization temperature and cool it to re-move defects and internal stresses. Nature lovesto settle to the minimum energy configurationif given time, no matter how complicated theoptimization problem is in the number of vari-ables involved; this is how gems, rather thancoal, form under slow geological processes.

Optimization techniques look at a func-tion’s local geometry to move to a better so-lution, but they are prone to get trapped atlocal extreme values. Stochastic optimizationmethods add an element of randomness to freethe algorithm from being trapped at such val-ues. Classical/simulated annealing is a stochas-tic optimization method in which these prob-abilistic jumps mimic the statistical physicsdescription of the annealing process. Quan-tum annealing improves the classical annealingtechnique by replacing the simulated ’thermalfluctuations’ with quantum fluctuations in realquantum systems.

Quantum annealing uses a quantum me-chanical system in which the cost function actsas potential energy, a term nature tries to min-imize. Whereas the classical algorithm had toclimb the height of the function’s ’hills’ to findthe next minimum, the quantum system cantunnel through the barrier as the width (notheight) gets small. Hence this algorithm out-performs classical annealing, at least in theory,

17

Dra

ft

if the function landscape has many high butthin hills surrounding the minima.

There is a company that does this andonly this. Rather than putting out a univer-sal quantum computer that is Turing complete(capable of computing everything computableand able to run all algorithms like the Shor’salgorithm), they capitalize on quantum anneal-ing processors, which only do quantum anneal-ing. It is D-Wave Systems, a Canadian com-pany, that sold the world’s first computer ex-ploiting quantum effects. They followed the in-dustry canon to put out a first-generation prod-uct in the market than to thrive on claims. D-Wave might be the best example for a companythat successfully capitalized on quantum tech-nology. Soon after its inception at the Univer-sity of British Columbia, its customer base grewto include Lockheed Martin, Google, NASA,and Los Alamos National Lab. Their annealingcomputers are finding more applications everyday in optimization, machine learning, and ma-terial science.

Quantum annealing is that important, andthe promises of speed up are groundbreaking.D-Wave’s latest release, Pegasus, has 5,000qubits, but this is not a significant feat in er-ror correction (see next section) since they aredesigned to do annealing alone. We could notfind any conclusive studies on the D-Wave pro-cessor’s performance.

4.2.2 Error Correction

Decoherence is the Achilles heel of quantumcomputing, confining it to prolonged infancy.Completely isolating quantum systems, qubitsin this case, from the rest of the universe is nextto impossible; anything can couple with the sys-tem’s dynamics to send the quantum state intothe unknown, sabotaging the calculation. Ef-ficient quantum algorithms make use of largescale interference between qubits, which is verydelicate. It is difficult to achieve this coordi-nation between qubits while protecting themfrom unwanted influences. This problem wasonce thought to be so forbidding that quantumcomputing would never work. Error correctionin classical computing relies on redundancy, but

that is not an option in quantum computingdue to the No Cloning Theorem, which statesthat no algorithm can copy arbitrary unknownquantum states.

Multi qubit error correction is so daunt-ing that many introductory reviews shy awayfrom them. However, the mathematics of quan-tum mechanics has led to many elegant formu-lations, including the stabilizer codes (Gottes-man [1997]). Single qubit error correction re-lies on coding a logical bit in a superpositionof multi-qubit states. For example, Shor’s codeencodes one logical bit into a nine qubit state,and subsequent algorithms reduced the numberfrom nine.

Google (38 patents), IBM (32 patents),and Microsoft (24 patents), who are also the in-dustry leaders in quantum computers, are thetop assignees in quantum error correction. IBMis expecting to operate a 1000+ qubit quantumcomputer by 2023, according to their roadmap.At this scale, quantum computers can changethe world and move away from the textbookexample scale problems it solved in its infancy.Better error correction codes are all there is be-tween the 65 qubit and the 1,121 qubit comput-ers mentioned in IBM’s roadmap.

4.3 Quantum Cryptography

All present cryptographic systems rely on hardone-way mathematical problems; for RSA cryp-tography, it is the integer factorization, and forElliptic Curve Cryptography (ECC), it is thediscrete logarithm. When quantum comput-ers catch up, the newly found computationalpower will take the ‘hard’ label away from theseproblems, so post-quantum cryptography is aprimary concern for industries as well as gov-ernments. If quantum computing develops asplanned, it will break both RSA and ECC.However, these two cryptography systems arechallenged by Shor’s algorithm, and other un-challenged classical cryptographies that can re-sist attacks from a quantum computer are al-ready available. On the other hand, there is aquantum solution to this quantum problem.

18

Dra

ft

Figure 12: Top Assignees in Quantum Cryp-tography

4.3.1 Quantum Key Distribution

QKD eliminates the vulnerability in transmit-ting encryption keys when shared through clas-sical communication channels. Whereas pre-quantum cryptography relies on its mathemat-ical complexity to ensure security, QKD re-lies on the no-cloning theorem, our adversaryin quantum error correction. QKD dominatesthe quantum cryptography node with 3,718patents.

QKD provides unconditional security, theholy grail of cryptography, in the exchangeof encryption keys. Relying on hard one-wayproblems in mathematics for encryption has theproblem that it gets vulnerable as computa-tional power catches up with time. Would any-one purchase a lock that gets weak over time?It depends. For example, the health records ofa person must remain private in his lifetime.If health data is compromised, insurance com-panies, employers, etc., can make discriminat-ing decisions to increase revenue. Anyone canrecord today’s internet traffic and the publickeys to decrypt them later, possibly before thedata loses its relevance considering the growthof ordinary computers, let alone quantum com-puters.

The public key shared using QKD (the de-

tails are omitted here for brevity, see this excel-lent review for details (Gisin and Thew [2007]))cannot be intercepted by an eavesdropper, nomatter how equipped he is. The measurementact will change the message carrier’s quantumstate to let the receiver know that someonetried to intercept the transmission.

R&D in this area focused on China, withthis country holding the most patents (2,465patents), way ahead of the US (445 patents).China has recently been in the news for push-ing the limits of QKD to establish a secure com-munication link to the low earth orbit satelliteMicius over a range of 1,200 kilometers(Liao,Cai, Liu, Zhang, Li, Ren, Yin, Shen, Cao, Li,Li, Chen, Sun, Jia, Wu, Jiang, Wang, Huang,Wang, Zhou, Deng, Xi, Ma, Hu, Zhang, Chen,Liu, Wang, Zhu, Lu, Shu, Peng, Wang, andPan [2017]). Terrestrial communication linksare limited to a few hundred kilometers dueto transmission losses in optical fiber; the no-cloning theorem forbids noiseless signal ampli-fication. Free space communication has the ad-vantage that the photons have to get throughroughly 10 kilometers of the atmosphere withnegligible absorption. China’s feat takes uscloser to the dream of a global secure quantumnetwork. Though channel loss in optical fiber isa hindrance, quantum repeaters can solve thisproblem, but this technology is still immaturefor practical implementation. Chinese scien-tists have established optical fiber links of upto 404 kilometers as of now (Yin, Chen, Yu,Liu, You, Zhou, Chen, Mao, Huang, Zhang,and et al. [2016]).

Although four of the top 5 assignees inthis area are Chinese, Toshiba, a Japanese com-pany, leads this group. The company is al-ready offering QKD systems to address the databreach problems that cost a fortune to compa-nies and government organizations. IBM esti-mated that the average cost of a data breachis 3.86 million USD, and it takes an aver-age of 280 days to identify and contain thebreach. Toshiba’s solutions are appealing inthis context of the already shaken communi-cation infrastructure. They have been involvedwith QKD from the 90s onwards after startingthe Toshiba Research Laboratory in Cambridge

19

Dra

ft

(1991). The company plans to own a fair shareof the multibillion-dollar market they expect forQKD in the 2030s.

The major US companies and govern-ment agencies, which are actively filing patentsin quantum computing, have a small numberof patents in Quantum Communication andCryptography. This might be correlated tothe government level policy about QKD; theNSA is vocal about the shortcomings of QKDas post-quantum cryptography. Even thoughthe eavesdropper cannot de-code the message,he can disrupt the communication by readingthe information passed. Plus, the NSA ar-gues that the infrastructure cost is too much,and the theoretically secure QKD is knownto be vulnerable in its physical implementa-tions (Huang, Navarrete, Sun, Chaiwongkhot,Curty, and Makarov [2019]Qi, Fung, Lo, andMa [2005]Wei, Liu, Ma, Yang, Zhang, Sun,Xiao, and Ji [2017]). Research in the at-tacking side is coming up with new strategiesto eavesdrop or interupt the communication,and the defending side is trying to eliminatethese vulnerabilities (Yuan, Dynes, and Shields[2010]Honjo, Fujiwara, Shimizu, Tamaki, Miki,Yamashita, Terai, Wang, and Sasaki [2013]).

NIST is actively looking at post-quantumcryptography proposals; early adoption of suchan algorithm would ensure that the data en-crypted will not be breached while it is sensi-tive. Using a symmetric key cryptographic sys-tem like Advanced Encryption Standard (AES)(champion of the last NIST contest in 2000 andUS national standard since then)with a doubledkey size would be a good idea in the mean-time. Symmetric key systems are vulnerable toGrover’s algorithm, which offers a square rootspeed up in a brute force attack, and a 256-bit key can provide the same security as a 128-bit key in the pre-quantum world (Mavroeidis,Vishi, D., and Jøsang [2018]).

QKD requires specialized hardware, andthe industry is analyzing the price to perfor-mance ratio. Existing cryptography systemsare so optimized over time, and some of themeven made into instruction sets of Intel andAMD processors. QKD is likely to replacehigh-stakes communications at first, consider-

ing the implementation costs and bandwidthlimitations in its present form.

5 Emerging Technologies

5.1 Superconducting Devices

Figure 13: Top Assignees in SuperconductingDevices

Superconductivity is the phenomenon of van-ishing electrical resistance in some materialsat sufficiently low temperatures characteris-tic of the material. It is a purely quantumphenomenon finding application in everythingfrom strong electromagnets in MRI machinesto qubits.

Superconducting qubits are the mostsought after type of qubits in the industry; bothIBM and D-Wave use them on their quantumcomputers. Even though the alternatives aremicroscopic systems like the electron, nucleons,etc., which are well isolated from the externalnoise, superconducting qubit, being a macro-scopic system, is easy to couple with otherqubits (Devoret, Wallraff, and Martinis [2004]).All of the top five assignees are here for thatreason.

20

Dra

ft

5.2 Quantum Sensing

Figure 14: Top Assignees in Quantum Sens-ing

Quantum sensing capitalizes on the extremesensitivity of quantum systems to external per-turbations to make high precision sensors (De-gen, Reinhard, and Cappellaro [2017]). Thoughthe atomic clock is an early example, presentresearch goes hand in hand with quantum com-puters, with qubits often used as sensors toprobe their quantum environment. Quantumdots can also be employed for sensing purposesfor reasons discussed in Section 4.1.1.

6 Conclusions

Quantum mechanics was one of the greatestrevolutions in physics in the last century by anymeans. It explained so much of nature, gavehigh precision results, and drove the geniusesof the day nuts. Einstein’s famous (correctlyattributed) quote, ”God does not play dice,”elucidates his loathing of this theory. Physi-cists eventually learned to live with it becauseit is our best theory of nature. P.A.M. Dirachad the opinion that only such a theory thatlimits the scales at which we can meaningfullylook can become a fundamental theory becauseotherwise, it will be an infinite ladder down

(Dirac [1982]). But quantum mechanics sim-plified some fields of physics like statistical me-chanics, and the underlying mathematical rich-ness of the theory has some elegance to it.

A similar benediction happened in thetech world from the onset of quantum tech-nologies. We have turned every weirdness ofquantum mechanics into something useful. Thestorage demand of quantum simulations scaledexponentially with the degrees of freedom, andwe used it to our advantage in quantum com-puting. Quantum systems are highly sensitiveto their environments, a huge challenge to over-come in the scaling of quantum computers, butthat made them excellent candidates for en-hanced sensors. The no-cloning theorem inval-idated redundancy measures for error correc-tion in quantum computing, but it is the cor-nerstone of quantum cryptography. The wholefield encompasses the fruits of hard work andperseverance.

Quantum technologies are being launchedinto the industry at the right time for ourcivilization and earth. Nanotechnology andelectronics are working in conjunction to putout devices that offer better performance andenergy efficiency. The onset of better light-ing solutions, semiconductor devices, etc., willlighten our energy budget, giving some cushion-ing in the transition towards renewable sources.If ever realized quantum computers could domore than cutting down the energy we spendon computation. We can use quantum simu-lators to study molecular interactions and de-velop efficient industrial catalysts. Through fullchemistry solutions, efficient alternatives to theHaber-Bosch process, responsible for 1% globalemissions, can be studied, and better batterychemistries can be invented. Quantum opti-mization algorithms can help us make betteruse of our resources to accommodate the grow-ing population and climate change and some-day even plan space missions.

While nanotechnology and quantum sens-ing materialized their promises in the formof products in the market, we should takequantum computing and communication witha grain of salt. The current state of the art,Noisy Intermediate-Scale Quantum era, is most

21

Dra

ft

likely to benefit quantum physics simulationsat first (Preskill [2018]), where decoherence er-rors are not undesirable, and maybe optimiza-tion with the success of companies like D-Wave.Many physicists think that quantum computingis decades away from making a significant con-tribution. Researchers are also exploring alter-native computing paradigms like neuromorphiccomputing (Schuman, Potok, Patton, Birdwell,Dean, Rose, and Plank [2017]) and molecularcomputing(Conrad [1990]). Quantum comput-ing is a passion project for most assignees inthe field, who often have huge payrolls. Crypto-graphic systems that are unconditionally securein all computation models, if used correctly, arealready available (Bellovin [2011]), albeit at theexpense of increased computational load. Wecould not find any replicable business modelsin quantum information; investments in R&Din these fields are likely to show returns only inthe long run.

There are no sure or impossible things inthe tech world; there are only novel technolo-gies and opportunity costs. Quantum tech-nologies is a giant by its promises, attractinginvestments more than ever. IBM’s Watsonwas introduced 64 years after the ENIAC, andhumans went to the moon 66 years after thefirst flight. Quantum technologies have alreadyput products on the market, and as a whole,the field is iteratively improving. Governmentagencies and industries are globally collaborat-ing on an unprecedented scale to materializethe promises of this field. Many of today’sbillion-dollar businesses like the internet, nu-clear energy, GPS, etc., were publicly funded ordeveloped for military applications before therewas a commercial market to them. If anything,quantum technology is a prodigy, optimisticallyset to change everyone’s life positively in thenext decade at the latest.

7 Acknowledgements

This report is the culmination of my intern-ship at Relecura Pvt. Ltd. (July-November2020), Bengaluru, India, investigating patent-ing trends in quantum technologies and tax-

onomizing the patents. I thank the Relecurateam, particularly Dr. George Koomullil, Dr.Hariharan Ramasangu, and Rohith Singh, forhelping me understand the dynamics of thepatent world using Relecura’s AI-based toolsand the invaluable insights they shared duringthe drafting of this report. Please reach out toRelecura ([email protected]) for more in-formation about the taxonomy and the toolsused.

References

Juan Alcacer and Michelle Gittelman. Patentcitations as a measure of knowledge flows:The influence of examiner citations. The Re-view of Economics and Statistics, 88(4):774–779, 2006.

Mirko Amico, Zain H. Saleem, and MuirKumph. Experimental study of shor’s fac-toring algorithm using the ibm q experi-ence. Physical Review A, 100(1), Jul 2019.ISSN 2469-9934. doi: 10.1103/physreva.100.012305. URL http://dx.doi.org/10.

1103/PhysRevA.100.012305.

Frank Arute, Kunal Arya, Ryan Babbush,Dave Bacon, Joseph Bardin, Rami Barends,Rupak Biswas, Sergio Boixo, Fernando Bran-dao, David Buell, Brian Burkett, Yu Chen,Zijun Chen, Ben Chiaro, Roberto Collins,William Courtney, Andrew Dunsworth, Ed-ward Farhi, Brooks Foxen, and John Mar-tinis. Quantum supremacy using a pro-grammable superconducting processor. Na-ture, 574:505–510, 10 2019. doi: 10.1038/s41586-019-1666-5.

Steven M Bellovin. Frank miller: Inventor ofthe one-time pad. Cryptologia, 35(3):203–222, 2011.

Luıs M. A. Bettencourt, Jessika E. Trancik,and Jasleen Kaur. Determinants of the Paceof Global Innovation in Energy Technologies.PLoS ONE, 8(10):e67864, October 2013. doi:10.1371/journal.pone.0067864.

Peter Bruck, Istvan Rethy, Judit Szente, JanTobochnik, and Peter Erdi. Recognition of

22

Dra

ft

emerging technology trends. class-selectivestudy of citations in the U.S. patent citationnetwork. CoRR, abs/1602.07928, 2016. URLhttp://arxiv.org/abs/1602.07928.

Giuseppe Carleo and Matthias Troyer. Solvingthe quantum many-body problem with arti-ficial neural networks. Science, 355(6325):602–606, Feb 2017. ISSN 1095-9203. doi:10.1126/science.aag2302. URL http://dx.

doi.org/10.1126/science.aag2302.

Seokkyu Choi, Hyeonju Lee, Eunjeong LucyPark, and Sungchul Choi. Deep patent land-scaping model using transformer and graphembedding, 2019.

Michael Conrad. Molecular computing. In Ad-vances in Computers, volume 31, pages 235–324. Elsevier, 1990.

C.L. Degen, F. Reinhard, and P. Cappel-laro. Quantum sensing. Reviews of Mod-ern Physics, 89(3):035002, July 2017. doi:10.1103/RevModPhys.89.035002.

R. H. Dennard, F. H. Gaensslen, H. Yu, V. L.Rideout, E. Bassous, and A. R. LeBlanc.Design of ion-implanted mosfet’s with verysmall physical dimensions. IEEE Journal ofSolid-State Circuits, 9(5):256–268, 1974. doi:10.1109/JSSC.1974.1050511.

M. H. Devoret, A. Wallraff, and J. M. Mar-tinis. Superconducting Qubits: A Short Re-view. arXiv e-prints, art. cond-mat/0411174,November 2004.

P. A. M. Dirac. The Principles of Quantum Me-chanics (International Series of Monographson Physics) 4th Edition. Clarendon Press,1982.

Jonathan P. Dowling and Gerard J. Milburn.Quantum Technology: The Second Quan-tum Revolution. arXiv e-prints, art. quant-ph/0206091, June 2002.

Peter Erdi, Kinga Makovi, Zoltan Somogyvari,Katherine Strandburg, Jan Tobochnik, PeterVolf, and Laszlo Zalanyi. Prediction ofemerging technologies based on analysis of

the u.s. patent citation network. Scien-tometrics, 95, 06 2012. doi: 10.1007/s11192-012-0796-4.

Richard P Feynman. There’s plenty of roomat the bottom. California Institute of Tech-nology, Engineering and Science magazine,1960.

Richard P. Feynman. Simulating physics withcomputers. International Journal of Theoret-ical Physics, 21(6):467–488, Jun 1982. ISSN1572-9575. doi: 10.1007/BF02650179. URLhttps://doi.org/10.1007/BF02650179.

Michael R. Geller. Quantum Phenomena inLow-Dimensional Systems. arXiv e-prints,art. cond-mat/0106256, June 2001.

Nicolas Gisin and Rob Thew. Quantum com-munication. Nature Photonics, 1(3):165–171,March 2007. doi: 10.1038/nphoton.2007.22.

Daniel Gottesman. Stabilizer codes and quan-tum error correction. PhD thesis, CaliforniaInstitute of Technology, January 1997.

Emily Grant, Megan Van den Hof, andE. Richard Gold. Patent landscape anal-ysis: A methodology in need of harmo-nized standards of disclosure. WorldPatent Information, 39:3 – 10, 2014.ISSN 0172-2190. doi: https://doi.org/10.1016/j.wpi.2014.09.005. URL http:

//www.sciencedirect.com/science/

article/pii/S0172219014001367.

Bartosz A. Grzybowski and Wilhelm T. S.Huck. The nanotechnology of life-inspiredsystems. Nature Nanotechnology, 11(7):585–592, Jul 2016. ISSN 1748-3395. doi: 10.1038/nnano.2016.116. URL https://doi.

org/10.1038/nnano.2016.116.

Toshimori Honjo, Mikio Fujiwara, KaoruShimizu, Kiyoshi Tamaki, Shigehito Miki,Taro Yamashita, Hirotaka Terai, Zhen Wang,and Masahide Sasaki. Countermeasureagainst tailored bright illumination attack fordps-qkd. Optics express, 21(3):2667–2673,2013.

23

Dra

ft

Anqi Huang, Alvaro Navarrete, Shi-Hai Sun,Poompong Chaiwongkhot, Marcos Curty,and Vadim Makarov. Laser-seeding at-tack in quantum key distribution. Physi-cal Review Applied, 12(6), Dec 2019. ISSN2331-7019. doi: 10.1103/physrevapplied.12.064043. URL http://dx.doi.org/10.

1103/PhysRevApplied.12.064043.

Shan Jin, Yanxi Hu, Lei Liu, and Hai-ChenWu. Application of quantum dots in biologi-cal imaging. Journal of Nanomaterials, 2011,06 2011. doi: 10.1155/2011/834139.

Richard Jozsa. Entanglement and QuantumComputation. arXiv e-prints, art. quant-ph/9707034, July 1997.

Michel Kurek. Quantum technologies, patents,publications & investments. September 2020.

Sheng-Kai Liao, Wen-Qi Cai, Wei-Yue Liu,Liang Zhang, Yang Li, Ji-Gang Ren, JuanYin, Qi Shen, Yuan Cao, Zheng-Ping Li,Feng-Zhi Li, Xia-Wei Chen, Li-Hua Sun,Jian-Jun Jia, Jin-Cai Wu, Xiao-Jun Jiang,Jian-Feng Wang, Yong-Mei Huang, QiangWang, Yi-Lin Zhou, Lei Deng, Tao Xi,Lu Ma, Tai Hu, Qiang Zhang, Yu-Ao Chen,Nai-Le Liu, Xiang-Bin Wang, Zhen-Cai Zhu,Chao-Yang Lu, Rong Shu, Cheng-Zhi Peng,Jian-Yu Wang, and Jian-Wei Pan. Satellite-to-ground quantum key distribution. Nature,549(7670):43–47, Sep 2017. ISSN 1476-4687.doi: 10.1038/nature23655. URL https://

doi.org/10.1038/nature23655.

Vasileios Mavroeidis, Kamer Vishi, MateuszD., and Audun Jøsang. The impact ofquantum computing on present cryptogra-phy. International Journal of AdvancedComputer Science and Applications, 9(3),2018. ISSN 2158-107X. doi: 10.14569/ijacsa.2018.090354. URL http://dx.doi.org/10.

14569/IJACSA.2018.090354.

Lawrence Page, Sergey Brin, Rajeev Motwani,and Terry Winograd. The pagerank citationranking: Bringing order to the web. Techni-cal report, Stanford InfoLab, 1999.

Peter Palomaki and Sean Keuleyan. Moveover, cmos: Here come snapshots by quan-tum dots, 2020.

John Preskill. Quantum computing in thenisq era and beyond. Quantum, 2:79, Aug2018. ISSN 2521-327X. doi: 10.22331/q-2018-08-06-79. URL http://dx.doi.org/

10.22331/q-2018-08-06-79.

Bing Qi, Chi-Hang Fred Fung, Hoi-Kwong Lo,and Xiongfeng Ma. Time-shift attack in prac-tical quantum cryptosystems. arXiv preprintquant-ph/0512080, 2005.

Benedek Rozemberczki, Ryan Davies, RikSarkar, and Charles Sutton. Gemsec: Graphembedding with self clustering, 2019.

Catherine D. Schuman, Thomas E. Potok,Robert M. Patton, J. Douglas Birdwell,Mark E. Dean, Garrett S. Rose, and James S.Plank. A survey of neuromorphic computingand neural networks in hardware, 2017.

J. M. Shalf and R. Leland. Computing beyondmoore’s law. Computer, 48(12):14–23, 2015.doi: 10.1109/MC.2015.374.

Jirawat Tangpanitanon and Dimitris G. Ange-lakis. Many-body physics and quantum sim-ulations with strongly interacting photons,2019.

Ashish Vaswani, Noam Shazeer, Niki Par-mar, Jakob Uszkoreit, Llion Jones, Aidan N.Gomez, Lukasz Kaiser, and Illia Polosukhin.Attention is all you need, 2017.

Kejin Wei, Hongwei Liu, Haiqiang Ma, Xi-uqing Yang, Yong Zhang, Yongmei Sun,Jinghua Xiao, and Yuefeng Ji. Feasible at-tack on detector-device-independent quan-tum key distribution. Scientific reports, 7(1):1–8, 2017.

Han Wu, Kun Zhang, Guangyi Lv, Qi Liu,Runlong Yu, Weihao Zhao, Enhong Chen,and Jianhui Ma. Deep technology trac-ing for high-tech companies. 2019 IEEEInternational Conference on Data Mining(ICDM), Nov 2019. doi: 10.1109/icdm.2019.

24

Dra

ft

00180. URL http://dx.doi.org/10.1109/

ICDM.2019.00180.

Hua-Lei Yin, Teng-Yun Chen, Zong-Wen Yu,Hui Liu, Li-Xing You, Yi-Heng Zhou, Si-Jing Chen, Yingqiu Mao, Ming-Qi Huang,Wei-Jun Zhang, and et al. Measurement-device-independent quantum key distribu-tion over a 404 km optical fiber. Phys-ical Review Letters, 117(19), Nov 2016.ISSN 1079-7114. doi: 10.1103/physrevlett.117.190501. URL http://dx.doi.org/10.

1103/PhysRevLett.117.190501.

ZL Yuan, JF Dynes, and AJ Shields. Avoidingthe blinding attack in qkd. Nature Photonics,4(12):800, 2010.

25