USSR Seeks to Build Advanced Semiconductor Industry With Embargoed Western Machinery
6
USSR SEEKS TO BUILD ADVANCED SEMICONDUCTOR INDUSTRY WITH EMBARGOED WESTERN MACHINERY
CONCLUSIONS
The USSR is engagedajor effort toarge, modern semiconductor industry usingMoscow will want to use the newly acquired equipment to accelerate development of more advanced military electronics systems where higher reliability, lower weight, smaller volume, and lower power requirements are crucial. Such systems could be used In strategic missiles, space communications and reconnaissance systems, antisubmarine warfare detection systems, and cryptographic equipment.
Semiconductors are sobd state electronic devices, the most advanced of which are called integrated circuits. These are of major importance in improving advanced weapons systems and sophisticated industrial processes. Current Soviet production is less thanf the US level in volume and even further behind in state-of-the-art. Most Soviet military electronic systems are still based on obsolescent transistor or tube technologies, and output of modern third-generation data processing computers is lagging badly behind planned goals.
oscow has purchasedillion worth of machinery and equipment for the manufacture of semiconductors, much of it the most advanced availablej
equipment that Moscow has ordered covers nearly the'cntire"
range of processes in an integrated, modem semiconductor industry. The equipment already delivered gives Moscow the capability to produce, at leastmall scale, more complex circuits than currently arc in production in the USSR.
lthough most of the equipment ordered has probably been delivered, US export control authorities have been advised of the Soviet acquisitions and
Note: This publication wasby the Office of Economic ResearchresearchOffice of Scienter folaltbrfice.
queries rryarding tins publication are welcomed
*
utc holding up certain Items (hut arc criticalhe operation of automated handling and environmental control systems In the manufacture of Integrated circuits.
he USSR has nol ycl acquired the manufacturing technology to make effective use of Ihc purchased equipmeni. In an ambitious scheme Io obtain such technology, Ihe Soviets ore searchingest European partneroint venture for semiconductor research, developmenl, and production. The partner would hire US experts to train West Europeans who. In turn, would train Soviei engineers and managers. If the USSR acquires all the equipment Ihat has been ordered, and comprehensive manufacturing know-how and training us planned, It may be oblc to produce hundreds of millionsiversified mix of ICs annually.
DISCUSSION
Buck ground
For many years, the Soviets have purchased, or have attempted to purchase, embargoed semiconductor produclionriorowever, such purchases were piecemeal ond few. Some attempts were made Io purchase comnlete integrated circuit (IC) lines or IC plants outside legal channels, but we do nol believe Ihat they were successful. Generally, the Soviets did not acquire or even seek associiledhe special processing techniques needed Io effectively use Western-madend relied Instead on Iheir own installation, maintenance, repair, and operator training.1
During ihese formative years of IC manufacturing, the Soviets seemed determined to make it on theiro buy some items of crucially needed equipment such oi bonders and diffusion furnaces In the West and to copy others (Japanese probe testers) but Io use Soviel-made equipment and Soviei technology wherever possible. Thus, most Soviet plinlsew items of Western equipment with mostly indigenous
The sporadic acquisition of Western equipment, by filling crucial gaps in the production process, probably permitted Ihe USSR lo get ICs inlo produclion somewhat earlier than would have been possible otherwise. However, we believe
t. Semlconduciofl ire electronic eoeiporvcnb that generally htrt replaced vacuum tubei ii Ihe bade boOtOn* btotki of electronic eqalpencnL Major types of at.-nlconddclon ire trininlon. diodei. and. tn thelt moilfoira. Inter*rcultv
1 The ffty tneww majnchn raw ol ihe SWariptobc teller. corkd from i
that Ihe overall Impact on Soviet production capabilities was not large. Western equipment alone could not compensate for tho deficiencies In Soviet-made equipment, backward production technology, and the almost total lack of quality and environmental control In Soviet plants.fter about four year* of production experience, the Soviets were able to produce only relatively simple small-scale Integration (SSI) bipolar devices of poor quality, and yields were low. The gap In semiconductor Icchnology between the USSR and the United States was large and growing, with the United States moving rapidly Into the production of high-density devices.
e believe that up3 the Soviets were producing mostly simple types of semiconductors (transistors and diodes) based on germanium material; the transition to silicon technology, and to the production of advanced types of devices including Integrated circuits based on silicon, had been moving slowly. Thus, the USSR produced only aboutillion ICsess thanf US output' (estimated at moreillione believe that the Soviets were able to achieve even this low level of output only through the application of large manpower resources, through the use of Inefficient trial and error methods, and with stolen or clandestinely acquired Western designs for devices.
Tlie Soviets were unhappy with the lack of progress in IC development and production and3 apparently decided to seek large-scale Western. Kolesnikov/ First Deputy Minister of the Electronics Industrynd head of itsain Administration responsible for all semiconductor development and production, toured the United States and Western Europe, met with leading semiconductor developers and producers, and assessed Western state-of-the-art. Kolesnikov's trip apparently signaled Ihe startajor effort to acquire Western equipment and technologyystematic scale.
34 the Soviets began to seek through illegal channels sizable quantities of machinery, Including the most advanced available. Moreover, they began to seek complete systems, rather than discrete items of equipment.
3. C oulpulI kn thinial of the United Slates.eiU/niteUSSR pnxhtetd let. lhanUIfenompirtd with IS billion In lhc Unite! Sun. fmSorts' IC production, mmt ot Ute ScMol aemlconauctof Irrfu.U* j
x vss tSSSSm lot to psssso-oud ffwi DUcciot el ibe Vomik RxBo Puno outUnoIG In UMo Fint Dtp*n> MtahWf In t'u thanjtmecent report. iuh uneonnmed. nam lhai be Iiio replace AX SbofclnHlnhiei of Iha MET eattr
and launched on ambitious scheme to obtain "know-how" and training (sec. In addition, they initiated efforts to acquire huge plani capacity for the production of silicon.4 the Soviets haveeadiness to spend very large sums of money, at0 million, for semiconductor production machinery, technological assistance,emiconductor materials plant.
Recent Equipment Purchases
that the USSR3 hassemiconductor production machinery ond equipment volucd atmillion.f the purchases hove been delivered toprocess of delivery, and therc being held up by UShave been made aware of the illicit nature of these orders.
scope of recent purchases is shown in Figurehich breaksmanufacturing into five majoraterialsdevice fabrication, assembly, andhe dollar value of those purchases is shown In thet is evident thats trying to build an entire modemimported equipment. Details of individual transactions are described in
USSR: Value of Soviet Purchases of Embargoed Semiconductor Production
of
of
rthsses (Percent)
S)
Proem
growth1
J
ik-nuking
Ine,1
for cklatiiof Soviei purchaiei tnd attempted. 7. By USauiaoritki.
). Additional aak< are expectedhey are not Ircluded in Ihh table (are
*
ecent purchases cover almost entirely two^ major semiconductor manufacturing processes: mask-making arid device fabrication. Together, these processes constitute "front-end"hat is, they Include the operations required for (he actual fabrication of semiconductor elements. The single front-end operation not covered by recent purchases is metallization.1
5. Pioccn of laying' down (he Inon paflti.
The crltkil riclnei In ptodticlojt new type* of rttculli are- maUnc Ihe mnit ind inmfeiflne. Ihe ne" dcilan chemically (pbolotllhopiphy) onto Iho aptia.nU! layer.
he front-end processing capability that has been acquired by_theof extraordinary interest since il could enable the Soviets to move rapidlyleast small-scale production of more advanced, higher density types ofcurrently are
nown orders Include equipment for only about half ol all the operations Involved In semiconductor manufacturing. In particular, recent purchases do not include any equipment forwith the exception of two specialized bonders (see tho Appendix,r for intermediate (probe) testing. These operations normally require largo quantities of equipment. For large-scale manufacturing, the Soviets require several hundred conventional bonders of the chip and wire typearge number jf piobe testers for use on-line In various stages of the manufacturing process.
Suspected Purchases
oscow may have already filled some of these gaps through earlier acquisition of equipment for metallization, bonding, and probe testing although the evidence Is not conclusive.
Metallization equipment (evaporators) may have been acquired fromirm in Switzerland, which has sold evaporators to Poland and has been pushing sales in other Communist countries as well, has had dealings with the USSR.
Bonding equipment may have been acquired from the Ur.Ued States through Yugoslavia. The United States authorized the export ofonders to Yugoslavia, which have never been satisfactorily accounted for. Suspicion that some (or all) of these bonders might have been reexported to the USSR arises because: Yugoslavia Is known to haverade protocol that calls for tho export of semiconductormanufacturing equipment to the USSR; Yugoslavia docs not Itself manufacture semiconductor production equipment; and Yugoslavia is known to have diverted some US electronics equipment to tho USSR.
the USSR mav now be serially producing Its own probeHowever, that technology is
now obsolete and, unless up-dated by the Soviets or aided bymay still be inadequate for modem
high-productimy* semiconductor production.
Attempted frurchases
urrently, the USSR Ii attempting to acquire equipment and technoli for the entire materials preparation process. Negotiations arc under way]
turn'*cey P'nntrt>ducc polycrystalline
silicon wilh minimum capacityetric tons per year. The plant also Is to produco monocrystalUne silicon In unspecified quantities. Tho projected output of polycryitalllne silicon would be equivalent to atons of silicon starling material.1
t present, silicon production in the USSR Is smoll. Hard evidence suggests that the total amount of monocrystalUne silicon presently available from domestic source, for the production of semiconductors is on the order ofons annually. By comparion, US consumption of monocrystalline silicon4 wasons.
n addition, the Soviets are attempting to buy slicers for the silicon production process from the United States through normal export channels. Two US firms recently requested authorization from the Deportment of Commerce tootal oflicing machines valued7 million. That number of slicers would be more than adequate toons of monocrystalUne silicon per year.
It thus appears that every major semiconductor operation except one (scribing) is covered by an actualuspected purchase, or an attempted purchase (sc.
Beyond acquiring equipment for particular processes, the USSR has not yet acquired the "know-how" to tie the various operations and processes together in anner.
Manufacturing Technology (Know-How)
Soviets are going to unusuol lengths to obtain ICThey would like to setoint production venture in ato circumvent embargo controls on technology; the Soviets haveAustria, and Switzerland. Mc.cow would underwrite the costnd yQo'. production facilityillion. Thethat within five yean tbe plant wouldapability to develop andthe whole spectrum of semiconductor and IC devices. Including some
7. ctvi cf poly crystalline illkoo rWd om ton ofJkrxm, ( SpedfkaBy, iV pu.nl would ba abta lo ptodaca larp-acaW Intepatton mrlal-oaldo aemlcondUcloti <LJt/MOS)SI/CMOS. bfiioxliUnipolat IO. Mpolai dxirici, MOS (hWhannel) ducretei and LSI. aUkon-tm-aapphlfa. and powat dartcst
24. The West European partner In the venture would hire US expertsthe "acHity and to train West European engineers and technicians;turn, would train Soviet engineers nnd managers, presumably in theflfc now going onwo or more firms.f
USSR also Is seeking to require manufacturing technologymember countriesuch more limited and less systematic wayexchanges" and visits to industrial facilities. Although studentnot new, the USSR has been steadily Increasing its emphasis on researchUnited States on semiconductors. During the put two years, visits ofhave expanded greatly under SAT agreements with private firms inttose visits have been used to learn about Western manufacturingalso to promote the sale of technology from the West to the USSR.the question ofomplete plant for the prxAuction of LSI devices.
Origin
The largest share of Soviot equipmentbout two-thirds by valuemainly diffusion furnaces, crystal growth equipment, and testers for advanced types of ICs. This equipment apparently was shipped directly ^to tho USSR.
About one-third of the equipment is of US origin and has been ordered by third parties in Western Europe. Often the third partyegitimate West European firm acting on behalfummy West European corporation set up specifically to carry out illicit purchases of semiconductor production mac'.inery in the West for the USSR. Dummy corporations exist in several West European countries, most of them directed by the same officials. In iome cases, US goods arc shipped directly to the legitimate West European firm, routed on to the dummy corporation, and transshipped from there to the USSR. In other cases, US goods ire shipped successively to several locations and then rccrnted and remarked to create the appearanceegitimate shipment to the USSR. It is not known whethersuppliers ara witting of the Soviet connection.
The United States is supplying ill of the equipment for the mask-making process and for two major manufacturing operations - epitaxy and
attha United Sialee *r* aapeneneed ptofeittotiah, rather thin thr Uriel meaning t4 3te lerm.
^ieCVfT
photolithography. Some of tlie equipmeni for each of these areas has already been shipped. That which Is being held up Includes more than two-thirds (in units) of the mask-making machines, nearly all of the epitaxyut of 13
arge amount of photolithography equipment. The potentialof equipment being held up Is discussed below (see.
Potential Impact of Equipment Purchases on Production
Sued on (he quantity of phefollthopiphk equipment thai haa been purchased.
Aiumlrjtay*t the minimum rate of throughput capacity (pedfled by (he Sovietsield of anbte product.
With equipment already delivered, the Soviets could achieve at least moderate increases In output and product mix early in the Tenth Five-Year Plan. However, we expect the Soviets to have difficulty getting any large amount of new capacity activated because of stringent environmental requirements, lack of production experience In the use of advanced Western equipment, and shortages of qualified pcrsonn"1
If the USSR acquires all the equipment that has been ordered, plus manufacturing know-how and training, it may be able toiversified mix of IC devices on the order of hundreds of millions of units annually, including many advanced types of ICs that arc now produced only in small lots or not at all. This capability would elevate the USSR to the statusajor world producer of ICs, trailing only Japan and the United States.
The Soviets apparently are planning to establish nine complete manufacturing lines forhey may be limited to only tvo lines, however, unless all of the equipment now being held up in the United States Isritical bottleneck will arise in the chemical processing phase of the photolithographic process for which theas only two of the ten systems ordered. Indeed, If the chemical systems are not shipped, the Soviets probably will be unable to use most of the automated handling systems purchased for the photolithographic process, since the systems are designed to be complementary.
If Moscow can acquire the needed complementary equipment, the nine lines would have the theoretical capacity to produce up to one billion simple -low-densityCs of Soviet design annually,'1 more thanimes current Soviet output. Although the USSR will want to produce large quantities of simple types of SSI ICs, because such devices can satisfy most of their current design
requirements, we believe they also will want to expand the product mix to include ooth medium-density (MSIs) and high-density (LSIs) devices. The purchased equipment is capable of producing many of the most advanced types of ICs now in production in the West, but Soviet capabilities will be limited by the state of their manufacturing technology. Without Western technology, the Soviets probably will be Hmited to the production of medium-density devices of the bipolar type, linear circuits, and the least complex types of high-density MOS circuits.
Ultimately, the output of ICs that can be achieved using the purchased equipment will depend on the product mix and on efficient equipment utilization. Both productivity and yields can be expected to fall as the mix shifts toward higher density devices. Efficient utilization will require proper installation, operation, maintenance, and integration of the equipment throughout the entire production process.
Byhe Soviets apparently axe hoping toapability in IC production capacity and technology approaching the current US level. Achievement of thisepend on Soviet success in acquiring Western technical assistance of the type envisioned in the proposed joint ventureon-COCOM country. The long-range goal may also depend on Soviet success inon capacity for polycrystaliine silicon production. Such large capacity wouldears to install. The long lead times for the acquisition of know-how and the installation of silicon capacity make it unlikely that the USSR can reach current US levels of semiconductor output before thet the earliest.
Some Implications
Moscow can be expected to attach the highest priority to the use of ICs in the production of military electronics equipment. According to knowledgeablef the ICs produced in the USSR in past years, including all of the best circuits, have been delivered to producers of military equipment. Even so, most Soviet military electronics systems stfll use transistor or tube technology and are In need of modernization.
The Soviets could choose to use newly developed ICs in future strategic missile systems. The use of ICs makesore complex guidanceore powerful computer, and improved reliability while reducing the size and weight of the entire electronics package. In the United States, the development
ll
omplex guidance system based on ICs has made an important contribution to increasing the accuracy of Minutemar II and to the introduction of MIRVs on Minuteman III.
However, ICs inapplications have special requirements such as hardening against radiation. Although purchased equipment can be used to produce even these ICs, unconventional processing techniques and specialized know-how also would be needed.
Purchased equipment also is capable of producing ICs that willide range of other military requirements. These include: most ground support equipment, tactical weapons such as antiaircraft and antitank missiles, artillery fire control systems, and most communications systems for aircraft.
The Soviets can be expected to place strong emphasis on the development of complex MOS/LSI circuits for microprocessors. In the United States, microprocessors arc now being designedide variety of US military hardware, including airborne Inertia! guidance, antisubmarine warfare systems, and airborne radarsook-down capability.
In the dvD sector, the USSR urgently requires modern ICs to improve its ailing computer industry. Although some third-generation RYAD family computers are now in serial production, output is far behind plan, and advanced models in the family, which require advanced types of ICs, stOI are apparently in the prototypee believe that the USSR would want to use some of the capability of imported Western equipment and technology to satisfy the requirements of the computer industry. This could be done without jeopardizing support of military requirements only if the Soviets succeed in activating several new IC production lines. Probably two lines would be needed to meet all current and projected IC requirements for civilian computer and computer peripheral programs during theears.
In addition, the USSR hopes to use ICs to modernize other types of equipment for industry and commerce, including especially, communications apparatus, electronic instruments, and numerical controllers for machine tools. Output of IC-based electronic instruments and numerical controllers, in particular, is to be pushed vigorously. According to recently published plan
Jl DpecuJlr (hendhich .re dfitgrnd lolui. rtUiNe emilttr-eoupltd lc*fe drtulu.
utomatic equipment with small-dimensionCs) electronic digital program vontrol is to develop at an acceleratednd "new forms of Instruments based on the wide utilization of microelectronicsCs) are to be developed and produced."
he USSR also may wish to use ICs In watches nnd cameras now exported to the West to earn foreign exchange. The growing use of ICs in Western electronic watches and cameras will tend to make Soviet products noncompetitive unless they incorporate modem IC design. Other lesser priority areas of civil electronics production that could use ICs to produce better, more reliable products at lower cost Include consumer products produced In high volume for the domestic market (radios and television receivers) and hand-held calculators. Production of hand-held calculators based on complex IC technology could appear to the Soviets an effective way of improving the efficiency of their vast army of planners and managers and perhaps as another way of earning foreign exchange.
BLANK PAGE
APPENDIX
DESCRIPTION OF EQUIPMENT TRANSACTIONS
oyst.nl growth
i. 3he soviets purchased
two models of machines for growing semiconductor
puller, described as an "optoelectronic single crystal pulling
a float zone crystal growthhere are indications
that additional salesand^joat-zoneay
have taken place
entatively, we surmise that float-zone models will be used to produce silicon and that the optoelectronic pullers will be used to produce other types of semiconductor raw materials for specialized applications, such as solid state displays, microwave diodes, infrared detectors, lasers, etc.
photolithography
for photolithographic equipment valued9f the total value of knownlectrochcmical processing lines, and(^tutomatedquaiity (environmental) control systems. most of this equipment isby two us firms. an'estimated! "vuisk-alignmcnt systems, of usbeen shipped f expected to sell an additional
ask-alignrncnt systems
Theie equipment)d to form monocryiUIDne tnaoti of very hlr.lt purity. The infot l> diced Intowhich, after being lapped and pollihed. are proceiied Into actualtentUlon, dlodei, and
The Iwo methodi commonly uacd lo produce tingle crystal aomlconduclor mate Halt ait: float-tone and .
automated handling and environmental control systemsfor delivery in the falle do not know if that schedulemet. these complex systems could be particularly advantageous for thethey are designed to minimize human error and eliminate contaminationwork environment, factors which, in the ussr, nrc major contributorylow yields and poorest european firm apparently hasassist the soviets in putting the equipment into operation.
Dif fusion
Soviets have1 unitshe sameto sell the Soviets anurnacesuvnace is not available. We assume, based on theof other equipment that has been ordered, lhat it is acapable of processing wafers of up to three-inch diameter. Three-inchfurnaces. In use In the United States sincerc largernow produced In the USSR. The Soviets havo only recently describedcapable of handling wafers larger than two inches.
Miik-Msking
for the mask-making process include^generatorsto create the art work or batic circuittothe pattern of thed rePcatreproduce the master patterns in miniature. These purchases, whichabout SS million,arge and div:nified mask-makingingle pattern generator of lhc typeufficient tooutput of integrated circuitsedium-size plant* ;L3ask-making capacity equivalent to any of the largestmanufacturers. The photo repeaters that have been ordered -models representing similar levels of highill permitto hindle circuit designs of the highest complexity (LSIs).
Testing
Testers for final test of
high-density (LSI)have been delivered to
iirnndnrf^t. nl.tnt in
esters were delivered lo unidentified facilities
J. Dinudoa fumicri uecrudil put of the inmleondoctot fibdatton proceahen ire pieced IndlfhiUoo fumcAC* 'or Klecife doplnf of Imporltte.orm thr buk dertee riannUi (Biiuhwi.tc.).
edloewiieby US tianduda.employO minion foncul-purpoM IC* annually.
3. The higher Ihe preelHon (he gte.ler Ihe comole.tlv of dreull detUman be rte^-oed.
C
^unit was delivered ^fc*pccts
IS, In i
^
johe USSRS. In lhc case of the
(rained Soviet engineers In the operation of
3
the icstcrs at Voronezh and provided training for associated computer software and equipment maintenance^
Epitaxy
Soviets have ordered pyutaxialystems' in at leastmodels, valued at2 cllllon. One model, wc know. Isuse- with galliumaterial usod, among'other things, forof advanced microwave devices for prcdomlnanlly military use. Ilthat some of (he other models also are intended for growing otherfor special application devices. However, technical data that wouldlo define further Ihe capabilities arJ functions of these models arc nol available.
Other
addition, the Soviets have purchased, and received, tWo advancedbonders. Tlic bonders arc of an unconventional beam lead type normallylimited processing of special application devices and military devices.
pliulal-.iWiwi in Died lo po*attend Uyei of xmlcondMrlOf nui (ridambniiia, Onlre-lodat. Kditon. indrt tabrkalrd (deTWd) bttn ifcl, Urw.
USSR: PUttWi of Emlmjoed SrakoadVlwftodt*lion EoatpftWM, Iff MS
PruOB
l-l rm
Milepilation Cryml (loth Mkn
FUllctl
Nfctl
fly no itpralen PNxo
neo
Pituin
PllKinoimpntii PiuemltWffl Dijtlal phittrn IJj.ui pV>tien Fit-Miili*
i* ilif oncnt
Uiil. alignment Ma* ittpinvni Mi* alignment EiectiO-oViMlrit linn EleciifKhnokit torn AuinmCrd handling ind quality corn ml linn EWIumh
Fuintcn
FufBlCfl
rUvirMy Boodle*
Boflthil (brim In 0
icn(LSt)
Cam. menu
DeltVriTd
I47J
Fxpeeud uki.
4
lleU up by Department of Co me* ice.
op by Drnirtnvnt ufI9TJ
Bp by Dtpiitment of Common.
ApeNA
_ Ikld up by Department of Comtatia.
cm
field tap by DepaiinvAt of Commie*.
5
HtM up by Department of ComoKict. NA In prrxcu of delivery. W)
I97J eeled
redJg "Utomrtit
a
)
4
EipKtednlei.
19
e.erntrr'"
Original document.
Comment about this article, ask questions, or add new information about this topic: