CA1189603A - Method of making a photovoltaic panel and apparatus therefor - Google Patents

Abstract

ABSTRACT
The specification describes a method of making a photovoltaic panel which comprises the steps of forming a roll of a web of a flexible substrate, unrolling the substrate roll substantially continuously into a partially evacuated space which includes at least one silicon depositing region therein where there is deposited over at least a portion of said substrate at least two thin, flexible silicon alloys which are of opposite conductivity (p and n) type. One or more of the alloys form one or more photovoltaic depletion regions. Subsequently, on the silicon alloys a thin flexible electrode-forming layer is applied on at least some of the photovoltaic depletion regions. An apparatus adapted to make a photovoltaic panel is also described.

Description

This is a DiYisional Application o-F Canadian Pa-tent Applica-tion Serial rlo. 377,664, filed May 15, 1981.
This invention relates to a method of makin~ more efficiently p-doped silicon films ~/ith higher acceptor concentrations and clevices made therefrom so that improved p-n and p-i n devices can be now produced in a batch or continuous process involving the successive deposition and formation of all or partially amorphous p and n type silicon films. While the invention has uti'lity in making di~des, switches and amplifier devices like transistors, it has its most important application in the making of photoconductive devices li~e solar cel's or other energy cor,version devices l~hen crystalline semiconductor technolo~y reached a commercial state, it became the foundation o~ the present huge semiconductor devices manu-Facturing indus-try. This was due to the ability of scientist to grow substantially defect free germanium and particularly silicon crystals, and then turn them into extrinsic materials with p-type and n--type conductivity regions therein. This was accomplished by diffusing into such crystalline materia'l parts per million of donor (n~ or acceptor (p) dopant materials introduced as substitutional impurities into the substantially pure crystalline materials, to increase their elec-tri-~ 3 \

cal conductivity and to control their being eitherof a p or n conduct.ion type~ The fabrication processes for making p-n junction and photocon-ductive crystals involve exi:remely complex, ~ime consuming, and expensive procedur~s. Thus, these cr~7stalline materials useful i.n solar cells and current control devices are produced under ~ery carefully controlled conditions by gro~ing in-dividual slng]e silicon or ~ermanium crystals, and when p-n jun~_tions are required, by dopin~ such sinyle crystals with e~t~emely small an~l cri~i~a~
amounts of dopants.
These crystal growing processes produce such relati.vely small crysta:Ls tnat solar cel's re~luire lS the assembly o~ many single cryst3.1s to eEIccmpass the desired area of only a single solar cell ~a~-el. The amount o~ energy necessary tc n~ake 2 solar cell in this process, the limitation caused by the size limitations of the silicon crystal, and the necessity to cut up and assemb).e such a crystall.ine mat:erial has a.Ll resulted in an i~-possible economic barrler to the lar~e scale use oE crystalline semi.conductor sola~ cells ~or en-er~y con~7ersion. Fllrtiler, crystalline silicon has 5 an indi~ect optical edge which results in poo--light absorption in the material. Because o~ the poor light. absorption, crystalline solar cells have to be at least 50 microns thick to absorb the incident sunlight. E~en if the crystalline ma-terial is replaced by polycrys~alli.lle silicon withcheaper production processes, the indirect optical edc~e .is still maintained; hence th~ material t.h:ick-ness iS tlO;- reduced. The polycrystalline materi.al also involves the addition of grain bvundaries and ot.heL problem defects. On the other han~1, amor-phous sillcon has a direcc optical ed.ge and only one--micron-tll.lck material. ls necessary to ah~orb the same amount of sunligl-lt as crystalline sili--con~
Ac~ordingl~, a consi.derable effort has been macle to ~evelop processes ~or readily depositing amorphous semiconduclor ~.ilms, each oL ~hich car encompass relatively large areas, i desir.ed, limit~cl only by the siæe of the deposition equlp-mel~t, and which could be readily doped to lorm p-type and n-t~pe materials ~here p-rl junction devi.ces are to be made therefrom equivalent to those produced by their crystalline counterpaLts.
For many years such work was substantially un-productive. Amorphous silicon or germanium (Group IV) films were ~'ound to have microvoids and d~n-gling bonds and other defects which produce a high density o~ localized states in the energy gap thereof. The presence of a high density oE J~cal-5 ized states i.n the energy gap of amorphous siliconsemiconductor films results in a low degree o~
photoconductivity and ShOft diffusi~n lengths, ma~ing such fi.lms unsuitable for solar cell ap-plications~ Additionally, such films canno~ be successfull.y doped or other~l~ise modifi.ed to shi-.t the Ferml level close to t:he conduction o ~alence bands" makin~ t.h~rn unsuitable fOL m.cl'~ing Scho~t.~y barrier or p-n jullctions for solar cell and cur-rent control device applications.
In an attempt to minimize the aforement:icned problems lnvolved wi.th amorphous silicon and ~er-manium, W. ~. Spear and P. G. Le Cor,lbe-- o:~- Carlne-gie Laboratory of Physics, University o l~undee, in Dundee, Scotland~ did scme work on "Substitu-~ tional Dopin~ o~ Amorphous Silicon~-l as rep~rt~
in a paper published in Solid Sta'~e Communica--tions, Vo.l. 17, pp. 1193-1196, 1975, towar~ the end o~ reducing the localized states in the ener~y ~ap in tlle amorphous silico!l or germanium to ma~e the-same approxirnate more close~Y intrinsic crys--tall;.ne silicon or germanium and of substitu-tionally doping said amorphous materia~s with suitable classic dopants, as in dopin~ crystalline mat.erials, to make them eY.trinsic and of p or n conducti.on types. The reduction of the localized states was accomplished by glow dlscharge deposi-tion of amorphous silicon films ~7herein a gas o silane (SiH4) ~as passed through a reaction tu~e where the gas was decomposed by an r.f. glo~7 dis~
charge and deposi.ted on a substrate at a subs~rate temperature of about 500-600K (227~C-327C~ The material so deposited on tl.le substrate was a~l intrins:;c amorpshous materia]. conslstirL~ of silicon and hydrogen. To produce a doped amorphous ma terial a gas of phosphine (PH3) for n-type con-duction or a gas of diborane (B2E~6) fo~ p-type conducti.on ~7ere premi~ed witll the silarle ~as a~d passecl through the glow discharge re2ct.icl. tube under the same operating conditiorls. The ~y~sec)u.s concentra~ion of the dopants used W-lS between about 5 x 10 6 and 10 2 parts per volu!ne. The ma~erial so deposited included supposedly sub-stitutional phosphorus or boron ~opant and ~ras shown ~o be extrinsi.c ar.d of n OL p conductio.n type. ~owever, the dopirl~ eficiency for the same ~ -5-3~8~

amount of aclded dopant material was much poorer than that of crystalline silicon. The electrical conductlvity fox highly doped n or p materif~1 was low, being about lo-? or 10-3 (S~ cm)~l. In acldi.-tion~ the band gap was narrowed due to th~ addi-tion of the dopant materials especially in the case of p-doping using c1iborane. These resu~ts indicate that diborane did not efficiently dope - the amorphous silicon bul: created localize~l states in the band gap.
As expressed above, amorphous silLcon, and also germanillm, is normall.y four-fold ccordinared and normally has rnicrovoidci and dangling bonds c.r other cteI-ectlve confiyurationsr p~oduci.na l.ocal~
ized states in ~he eneryy gap. Whlle it ~!~as not kno~n by these researchers, :;t is now known by the work of others that the hyd.-.oclen in the silane :combirles at an optimum tempeLature with many of the dangling bond.s of the silicon during the ~lo~
discharge deposition, to decrease substan~ic~ y the densl~y of ti-le localized states in the ene-r~y gap toward the end of making the amorphous ma-terial approximate more nearly the correspon~ing cr~stalline material.
llowe~er, the incorporation oE hydrogen no'L

only has lîmitations based upon the fixec3 ratio v~
hydrogen to silicon in silane, bu~, most irmpor-tantly, various Si:M bonding configurations in-troduce new anti.bonding states which ean ha~e deleterious eonsequences in these materialsO There-fore, there are basic limitations in redueinc~ the density of loeali~ed states in these mater;als whieh are particularly harmful in terms o~ ef-fective p as well as n doping. The resll~.tinc~ -unaecepta~le density of states of the silQne-deposited material.s leads to a narrow depleti width, which in turn limits the effit_iencies Gf solar eells and other de~i.ces whose operation depends on ~he drift of free carrier.s. The Ine~h~c'~
Of ma~ing these materials by the use o, orlly si].i-eon and hydrocJen also results in a high density o~
surfaee states ~hich arlects all the d~OVe pararn-eters.
After the developrnent of the 910~ diseharg~
deposition of silicon from silane gas ~7as carried out, WOL ~ as done on the sputter deposi~incJ o amorp'nous silicon films in the atmospheLe or a mixture of ar~on (recluired by the sputtering a~p-os;.t:ion process) and molecular hy~lro~;en, ~o d~-terlnine the results Gf such molecular hydro~en on ~ . ~

~ 6~3 the characteristics of the deposi~ed amorphoussilieon film. This research indiea-ted that the hydrogen acted as an alterinq agent which bonded in sueh a way as to reduce the locali~ed states in the energy gap. Ho~ever, the degree to which the locali~ed states of the en~rgy gap were reduc~d in the sputter deposition process was much less than that achieved by the silane deposition plocess ~escribed above. The abo~e described p and n dc~pant gases also were introduced i.n tlle spu~ter-ing process to pxoduc~ p and n doped mateLials.
These materials had a lower doplng efficienc~ th~.n the ma~eri.als produced in the glow discha3-ge pro-cess. Nei'cher process produeed ef~icient p-dope~
materials witll sufficiently h:igh accep~or ~on-ce3ltrations for produeing commercial p-n or p-i-n junction deviees. The n-doping effiel.enc~F was below clesirable acceptable commercial le~Fel.s and . the p-doping was partieularly undesirable sin~e it reduced the width of the band gap and increased the numbers o, :locali~ed states in the ba~d ~ap~
The prior deposition of amorphous si~icon, which has been altered by hydrogen from ~he silane ~as in an attempt to ma~.e it more C1OSe1Y reSemb1e crystalline silicon and wllich has been doped in a .

6~3 manner like that of dopiny crystalline sili.con, has characteristics which in all important re--spects are inerior to those of doped crystalline silicon. Thus, inadequate doping efficiencies and conductivlty were achieved especially in the p-typ~ materia]., and the photoconductive ~nd photo--voltaic qualities of these silicon films left m~ch to be desired.
substantive breakthrouyll in forming amor-0 phou5 silicon films with a very lo~l derlsît~ ofstates was achieved by the inven'~ions disclosed in our U.S. Pa.tents Nos. 4,217,374 and 4,226,~38 which produced amorphous fi.lms, particularly sili--~on amorphous fll.ms, ha~ing the rel~ti~e favorable l~ attributes of crystalline semiconductor matetsial~.
~The former patent discloses the deposition o improved amorphous silicon fil~s using vapor ~ep-~osition therof an~ the latt~r patent disclose~ t:he deposition of improved amor~hous silicon films by glow discharge of silicon-containing gases.~ Th~
improved amorphous intrinsic silicon fillll.s pro~-duced by th~ prvcesses disclosed therein h~ve reduced nu~ber of states in the band gap in the intrinsic ma.~erial and provi(3e for ~r atly in-creased n-doping e'ficiencies, high photocon _~ _ 6~:3 d~lctivity and increased mobility, long diffusion length of the carriers, and low darl~ intrinsic electrical conductivity as desired in photovoltaic cells. Thus, such amorphous semicondJctor -films can be useful in maki.ng more efficient devices, such as so:lar cells and current cont.roiling de-vices including p-n junction devices, diodes, transistoLs an~ the like.
The invention ~h.ich achi.eved these results incorporates into the amorphous .ilms, E~referably as they are being deposited, alterant o~ compen-sating materlalc. ~7hich a.re beli.e~ed to fornl alloys with the amorphous semiconductor materia~s and modify tlle same so as to greatly red~lce the lo-calized states in the energy gap thereof to malcethe same equivalent in many respests to intrinsic crystalline silicon. In the prosess of forming silicon films disclosed in said Paten, No. 4,226,~
a compoulld includins silicon as an elem~nt theJ:eo is decomposed by glo~l discharge decomposition t:o deposit amorphous silicon on a subs,rate a~ong with the incorporation of a plurality of alterant elements, preferably activated fluorine and hy-drogen, dl~ring the ~Jlow discharge deposi_ion.
In these spec.ific embodiments of ,he in---10 ~

vention disclosecl in the latter application, sili-con is deposited in a batch mode at a substrate temperature of about 380C by the glow dischar~e oE silicon tetrafluoride (SiF~) which supplies the silicon in the deposited amorphous films and flu-orine as one alterant or compensating element.
While silicon tetrafluoride can orm a plasma in a glow discharge, it is not ~y itself most effec~ive as a starting material for glow dischclrge dep-osition OL silicon. The atmospheL^e for the glowdischarge i~ made Leactive by adding a gas li~e molecular hydrogen (E12), which is ma~e reactive ~r the glow discharge by changing it to atomic hy-drogen or h~drogen ions or the like. This re-active hydrogen reacts in the glow discharge ~iththe silicon tetrafluoride so as to more readily cause decomposit1on thereof and to deposit amoL-phous silicon therefrom on the substrate. At tne same time, ~luorine and various silicon suhflu-orides are relcased and made reactive by the gJowdischarge. The reacti~e h~drogen and the reacti~e fluorine species are incorporated in the a~orphou~
silicon host matrix as it is being deposited an~
crea',e a new intr;nsic material ~hich has a low number of defect 5 ta~es. A simple way to consider 6~3 the new all.oy is that there is a satiatiorl of capping oE dangling bonds and the elimina~ion of other defects. Hence, these alterant eler,lents reduce substantially the densi.ty of the localized states in the energy gap, with the foregoing bene-ficial results accruing.
When it is desired to provide n-type and p-type conduction in the arnorphous si.licon semi conductor matri~ the latteL application recom-mends incorproation of modifier elemerlts in gas-^
eous rorm during the ~lo~ deposition of the fil~l.
The re^ornmended modifier elements or dopan~s ~or n-~type conductj.on are phosphorus and arsenic in the form of ~he ~ases phosphine (PH3) and arsi~e lS (AsH3). The recornmended mc~i~ier elements or dopants ~or p-type conduction are boron, alum.inum~
gallium and indiurn, in the form of the gases dl--borane (B2H6), Al~C~Hs)3, ~a(C~i3)3 a~d In~C~313.
The modifier elements were added under the same

2.0 deposition conditiolls as described for the in-tri.nsic material ~ith a substrate temper~ture of about 380C.
While the pLOCeSS for making deposited .5ili-con devices in the aforesaid applications reprc~-sen~.s a significan. improvement, making possi~]e 6~3 the production of improved solar cells and oth~r devices, the p-doped deposited silicon material did not have a p-type conductivity as efficlenJ~ as desired. ~s reported in the Journal of Non ~rY-stalline Solids, Volumes 35 and 36~ Part I, Jan-uary/February, 1980, pp. 171-181, with the add~-tion of 500 ppm PEI3 in the cleposition gasest cor-respondin~ to an nt layer and the intrinsic m~-teria3. With the addition of diboLane (B2I161 in ~he cleposition gases, sign~ficant changes in opt7-cal absorption ta~es place. The implication is that a new alloy invo]Yin~ boron has been syn-thesized which possesses a more narrow band t~ap and exhibits p-type characteristicsO It i~ os-sible that three-center bonds unique to boron ~re responsible in part for this behavior. Thi, is in contrast to the results obtained when phosplloru~s or arsenic are aclded where a conventional n-~y~e material is produced.
While devices like a Schottky barrier or MI5 ~evice can be ma~e with or without p-doped ~ilr.lsF
they are difficult to manuEacture since the prop-erties of the thin barrier layer commonly used therein i5 dif~icult to con~rol ~nd fLequently th~
thin layer cannot be efficiently encapsulatecd ~o ~9~;~3 preven~ diEfusion of environrnental elements there-through with the result that the device is Ere-quently unstable. In addition, such structures lead to a high sheet resistance in the upper level S of the device. It appears that a photovolcaic c~ll having desirecd efficiency and stability re-quires utilizing a p~n or p-i-n junction. ~or this ~urpose, an i.mproved p~c]oped material is desirable i.o increase the efficiency of the cell~
In making the fluorine and hyclrogen compen-sated glow dischar~e depositecl si licon filln.s dis-closed in ~he latter aforesaid patent, the silicoi^
is preferabl~- deposited at a s~bstrate tempe~ature oF about 3~0C. Above this substrate tempera~ure~
1~ the ef~iciency of the h~d.rogen compensation grad-uall~ decreases and at temperatures abov~ about 450.C redu.ces signifi~antly, because the ny~co~en does not readily combine with the cleposi.tincJ si~
con at such temperatures.
As noted above, it has heen disco~Jerecl th~t the introductioll of the gaseous p-dopan-t l~ater-ials, while produc~ng a p-type material, do not produce a materia' i7ith a p-type conduction eE-ficiency as ~:ould be theori2ed if only the d2sire~
four-sidecl or t~trc,hedral bondin~ were ta~ing ~1~--:

place. It appears that at the glow discharge substrate temperatures of 400C or belo~, which are necessary for the mosc ef~icient hydrogen compensation of the silicon material some of the would-be p~dopant materials are threefold rather than tetrahedrally coordinated, because of th~
absence of crystalline constrainLs, thus leac~ing to additlonal states in the gap and no dopi~g.
Other processes involving cliborane lead to the fv~mation of three-centered bonds or other less efficient combinations be~ause the metallic or boron parts thereof do not readily disassociat-e completely from their hydrocarbon or hydrogen companion substituents and so do not in such ~orm provide an efficient p-doping element in the sili-con host matrix. ~`urthermore; states are ad~ed in the band gap of such materials which are belie~7ed to reduce the p-doping efficiency achie~e~.
Therefore, appreciable effort has been made to improve the p~doping efficiency of said p-doping elements in glow discharge deposited sili-con material. Glow discharge deposition of ~
con for photovoltaic and other applications re-quiring intrinsic layers or p-n juncticn formed 5 deple ion regions presently appears to be the ~re~erred ~e~osit,ion metho~ thexe~or, ~inc,~e the~egreQ of hydrogen a~d fluorine compensation and reduced density of states in the reSulting ma-$eria1 ~re Superior to that obtained by vapor ~epositiorl or sputtering of si]icon.
~ l'he present invention has to do with a method ~ more efficiently p-doping material in a glow d,ischarge silicon deposition batch or con~inuous - `
-~rocess ~o procluce more eff:icien~ly dop~ p-t~pe lP materials and p-n and p-i-n j~lnction devices in-corpora~ing tlYe more efficic-ntl~ p-doped silicon materials. The methods~of maklng p-dope~ material ir! the prior art have been limited to use c~ ~on-ventional dopant gases, such as diborane, unde~
the de~osition conditions optimized for the in--trinsic materials. No one heretofore considered p-dopant ~aseous boride compounds ~such as B2~.~6) and p-dopant metal g~seous compounds useful'in ~low discharge deposition of amorphous (or poly-crystalllne) silicon deposited at substrate te~n-peratures above about 450C which has been con-sidered to be outside the temperature range re-quired for the preparation of the useful amorpho~s silicon.
?5 ri'he pLesent inven~ion also encompasses the 36~3 method Oc making a more eficiently p-doped glow discharge deposited silicon by depositing the material above about 5QQ C. The loss oE the ad-vantages of hydrogen compensation in the silicon materials deposi~ed at these high tempeLatures i5 more than overcome by the increased efficiency of the p-doping achieved, especially where the p-doped deposited layer is to form an ohmic p~ in-terface with the associated electrode. ~s pre-viously stated, it zppears that at these hightemperatures the boron or metal p-dopan. elements are so substantially disassociated frorn the hy-drogen and hydrocarbon elements of the gaseous conpound used that the three center or other un-desirable bonding configurations are eliminated.The desired four-sided (te rahedral) bondillg which is efficient for p-doping is thus obtained. Al-though p-dopant metal (i.e. Al, Ga, In, Zn and Tl) compound gases were also not effective as p-type dopants in the glow discharge ~leposition o sili-con using su~strate temperatures a~ or below about 400C, these elements are good p-dopants in gas-eous compound form using the much higher silicon glow discharge substrate temper~tures descri~ed herein ~that is temperatures at least about 500C).

, It should be noted that although the high sub-strate temperatures above about 500 C can result in inefficient hydrogen compensation of the sili~
con material, the material is still effectively fluorine compensated since fluorine efficiently combines with the deposited silicon at substrate temperatures up to the range of 700C to ~00C.
For amo~phous silicon deposited without h~--drogen or fluorine compensation, the.cry~talliza-tion pr~cess ~ecomes important at substra'e tem-peratures of about 550C. For depositin~ a~or-phous silicon with hydrogen compensa_ion and/or alloying the amorphous state substantiall~ is maintained up to substrate temperatures of about 550C. For amorphous silicon compensated with hydrogen and doped with boron, the amorphous state remains to substrate tempel~atures about 700C.
The addition of fluorine such as in the materials of this invention, extend the amorphous state oE
the deposited materia]. t.o still hi~h~r substrate temperatures. From this it is clear tnat the present process produces fluorine co~pensated amo~phous silicon dopecl with boron at su~strate t-.e~,peratures above 700C. "oping le~els ~c,hieved with deposition substrate temperatures such that .

the hydrogen and fluorine compensated silicon film remains substantially amorphous, wi31 be suF-ficient for certain doping applications. For still higher doping levels, hiyher deposi~ion substrate temperatures may be used such that the amorphous material will become mixed with cr~7-stallltes of silicon, or become subst~tially polycrystalline.
The inclusion of crystalllte macerial into the amorphous deposited silicon or th~ use of substantially polycrystalline p-doped material does no-~ impair the efficiency of a p-n or p~r-i-n+
photovoltaic device. The efficielcy is not iin-paired because the eE-E;ciency of p doping in poly-crys~alline silicon is well known, and becaus~ tneoptical absorption of the crystallites will be lower than that of the amorphous material, so ~he photon absorption in the photoactive layer will not be affected. Eor amorphous materials ~lith high absorption coefficients, the p~~ layer in a n~~ structure is kept as thin as possible, less than 1000 angstroms, to minimi~.e absorp~ion of photons since it is a non--photoacti~e layer.
The layer thickness still provides enouqh ~sitive carriers to bend the conduction and valence bands between ~ne p~~ and the intrinsic layer in the device for efficient photovoltaic action. The admixture of silicorl crystallites into the amor-phous silicon, not only does not impair the ef-ficien:cy of a p+-i-n+ device, but also may assist the efficiency of a p-n photovoltaic device be-cause of the increased hole mobility and inc,eased photoconductivity of the crystallin2 p material compared with amorphous p material.
The present invention also discloses the method of eliminating the difficult~ of p-doping by utilizing an unconventional non-gasec~us ma~
terial as a dopant. The method includes heatinq a solid metal to a h.igh temperatllre to evaporate t~e 15 metal and then feed the metal vapo~r directly into the glow discharge chamber with the siliccn dep-ositi.on gases continuously or intermittentl.y~ The p-dopant metals in a vaporized me~allic ~o~m are efective in the glow discharge deposition of 20 silicon at lower substrate temperat-lres, T.~here fluorine and hydrogen compensation is desired.
These evaporated p-dopant metals can also ~e util-ized ~Tith glow discharge silicon deposited rilrn at higher substrate ternperatur.es where h,drogen coln-25 pensation is not needed.

~`~, ~ tilizing the present invention, p-dopant boron and metal materials may be deposited in a continuous process combined ~7ith n and intrinsic type glow discharye deposlted amorphous materials to manufacture improved p-n and p-i-n junction photovoltaic and the like devices~ In the con-tillUOUS process, the materials are glow discharcie deposited upon a web substrate as it is contin-uously or step~lse moved through separate dep-osition stations eacll having the substrate ~em-perature and other environmQntal conditions nec essary to efficiently deposit the particular de-sired p and n and/or intrinsic type sillcon films on the con~inuous web. In the continuous man-u~acturing process of the invention, each dep osition station is dedicated to depositing one layer (p, i, or n), because the deposition ma-terials contaminate the station bacl~ground en-vironment and are not easil~ removed.
While the principles of this invention apply to the aforementioned amorphous and polycrystal-line type silicon semiconductor materials, for purposes of illustration herein and as settin~
fortll preferred embodiments of this inven~ion~
specific reference is made -to gaseous boron and gaseous and evaporated metal p-dopant material glow discharge deposited with the silicon material at substrate temperatures of between 500c to 700C. The deposited film will be fluoride com-pensated throughout the substrate temperaturerange, but the hydrogen compensation will decrease with increasing substrate temperature. Also, the evaporated metal p-dopant materials may be glow discharge deposited with the silicon material at substrate temperatures below 400C to form a hy~
drogen and fluoride compensated p-doped material.
In summary, to bring the significance of the present invention into focus, it is believed that the present invention enables the fabrication of more efficient p-type amorphous semiconductor films for use in the manufacture of solar cells and current devices including p-n and p-i-n de-vices. Additionally, the present invention pro-vides for viable mass production of the various devices in a glow discharge environment ~ith boron or at least one of the metals Al, Gal In, Zn or Tl providing the p-dopant material at prescribed substrate temperatures.
We have found that the above disadvantages may be overcome by depositing a silicon containing ~8~316~3 .

film or alloy with an evaporated metal p-dopant, or example, aluminum, gallium, indium, zinc and/or thalliu~, to form a p-type alloy. We have also found that the improved p-type allo~ may be formed by depositing the silicon containing alloy wi-th a gaseous compound cont~ining at least the p-dopant element, ~or example, aluminum, gallium, indium, zinc, thallium and/or boron, which is disassociated at substrate temperature of 450C or above. Various devices~
including for ex~mple p~n and p-i-n solar cells, can be made incorporating these p-type alloys and preferably are formed at least in part by a continuous web deposition process.
Embodiments disclosed in this applica~ion are also disclosed and claimed in Applicant?s copendi~g Application Serial Number 377~664D
Therefore, in accordance with one aspect o~ the present invention there is provided a method Qf making a photovoltaic panel comprising the steps of forming a roll of a web of a flexible substrate, unrolling the substrate roll substantially continuously into a partially evactlatea space including at least one silicon depositing region therein w~ere there is deposited over at least a portio~ of the substrate at least two thin, flexible silicon alloys which are of opposite conductivity (p and n) type, one or more of the alloys forming one or more photovoltaic depletion re~ions.
The substrate is formed in a substantially continuous web and each of the silicon alloy films i5 deposited at a se~ara~e ~low discharge region past which the web is moved to form a substantially continuous deposition process.

"' , cr/l'i 6~3 In accordance with a second aspect there is provided an apparatus for making a photovoltaic panel from a roll of a web of a flexible substrate material. The apparatus comprises means for forming the substrate in a substantially continuous web; means for unrolling the substrate roll substantially continuously into a partially evacuated space, the space including at least one silicon depositing region therein, the region including means for depositing over at least a portion of the substrate at least two thin flexible silicon films which are of opposite conductivity (p and n) type, one or more of the films forming one or more photovoltaic depletion regions. The means for depositing the silicon film include means for depositing each of the silicon films at a separate glow discharge region past which the web is moved to form a substantially continuous deposition process.
The preferred embodiment of this invention will now be described by way of example with reference to the drawings accompanying this specification in which:
Fig. 1 is a partial schematic and partial diagrammatic illustration of the process steps for making semiconductor devices including the p-doped material of the invention.
Fig. 2 is a diagrammatic illustration of an apparatus for continuously depositing the semiconductor film of the invention.

- 23a -cr/\

Fig. 3 is a block diagram o~ one illustra~ive apparatus for performing the process steps of Fig.
1 to continuously foxm the improved p-doped semi-conductor devices of the invention.
Referring to Fi~. 1, the first step ~A~ in the manufacture of the devices incoxporating the improved p-type material o~ the invention includes forming a substrate 10, The substrate may be formed o~ a non-flexible material such as glass where a batch process is involved o~ of ~ ible web such as aluminum or stainless steel, especial-ly where a continuous mass production process is involved. Thus, the flexible substrate web 10 may be utilized in a continuous process to deposit ~he Yarious layers of metal electrode-forming and silicon layers as the web is drawn through various deposition stations to be described hereinafter with respect to Figs. 2 and 30 The aluminum or stainless steel substrate 10 preferably has a thickness oE at least about 3 mils, and preEerably about 15 mils and is of a width as desired. In the case where the web 10 is a thin~ flexible web it is desirably purchased in rolls.

r~

The second step ~B) includes depositing aninsulating laye~ 12 on top of the aluminum ~r stainless steel substrate 10 so that spaced in-sulated electrode-forming layers are formed, if desired, thereon. The layer 12, for instance, about 5 microns thick can be made of a metal ox-ide. For an aluminum substrate, it preferably is aluminum oxide (A12O3) and for a stainless steel substrate it may be silicon dioxide (SiO2) or other suitable glass. The substrate can be pur-chased with the insulating layer 12 preformed thereon or the insulating layer 12 can be laid upon the top of the substrate surface 10 in a conventional manufacturing process such as ~y chemical deposition, vapor deposition or anodizing in the case of the aluminum substrate. The two layers, substrate 10 and oxide layer 12, form an insulated substrate 14.
The third step (C) includes depositing one or more electrode-forming layers 16 on the insulated substrate 14 to form a base electrode substrate 18 for the junction device to be formed thereon. The metal electrode layer or layers 16 prefera~ly is deposited by vapor deposi-tion, which is a rela-tiYely fast deposition process. The electrode layers preferably are reflective metal electrodes of molyb-denum~ aluminum, chrome or stainless steel for a photovoltaic device. The reflectiYe electrode is pre-ferable since, in a solar cell, non-absorbed light which passes through the semiconductor material is re~lected from the electrode layers 16 where it again passes through the semiconductor material which then absorbs more o~ the light energy to increase the device 10 ' efflciency.
The base electrode substrate 18 is then placed in a ylow discharge deposition environment, such as the chamber described in said Patent rlo~ 4,226,898~
or a continuous process apparatus as discussed herein-after with respect to Figs. 2 and 3. The specific examples shown in Dl-D5 are merely illustrative of the various p-i-n or p-n junction devices which can be manufactured utilizing the improved p-doping methods and materials o~ the invention. Each of the devices js formed using the base electrode substrate 18. Each of the devices illustrated in Dl-D5 have silicon films having an overall thickness of between about 5000 and 30,000 angstroms. This thickness jb/``~

~896~P3 ensures that there are no pin holes or other phys-ical defects in the structure and that there is maximum light absorption efficiency. A thicker material may absorb more light, but at some thick-S ness will not generate more current since ~hegreater thickness allows more recombination of the light generated electron-hole pairs. (It should be understood thAt the thicknesses of the various layers shown in Dl-D5 are not drawn to scale.) Referring first to Dl, an n-i-~ device is formed by first depositing a heavily doped n~
silicon layer 20 on the substrate 18. Once -the n~~
layer 20 is deposited an intrinsic (i) silicon layer ?2 is deposited thereon. The intrinsic layer 22 is followed by a highly doped conduc-ti~e p~ silicon layer 24 deposited as the final semi-conductor layer. The silicon layers 20, 2~ and 24 form the active layers of an n-i-p device 26~

~0 While each of the devices illustrated in Dl-D5 may have other utilities, they will be now described as photovoltaic devices. Utilized as a photovoltaic device, the selected outer, p+ layer 24 is a low light absorption, hi~h conductivity -27-~

6~3 layer. The intrinsic layer 22 is a high abscr~-tion, low conductivity and high photocondu~tive layer over a low light absorption, high conduc-tivity n+ layer 20. The overall device thickness S between the inner surface of the electrode layer 16 and the top surface of the p~~ layer 24 is, as stated previously, on the order of at least about 5000 angstroms. The thickness of the n~ doped layer 20 is preferably in the range of about 50 to 500 angstroms. The thickness of the amorphous intrinsic layer 22 is preferably between a~out 5000 angstroms to 30,000 angstroms. The thickness of the top p+ contact layer 24 also is preferably between about 50 to 500 angstroms. Due to ~he shorter difEusion length of the holes, the p~
layer generally will be as thin as possible on the order of 50 to 150 angstroms. Further, the ou,er layer (here p+ layer 24) whether n+ or p+ will be kept as thin as possible to avoid absorption of light in that contact layer.
Fach of the layers can be deposited upon the base electrode substrate 18 by a conventional glow discharge chamber described in the aforesaid U.S.
Patent No. 4,226,898, or preferably in a con--28~

6~3 tinuous process described hereinafter with respec~
to Figs. 2 and 3. In either case, the glow dis-charge system initially is evacuated to approxi mately 20 mTorr to purge or eliminate impurities in the atmosphere from the deposition system~ The silicon material preferably is then fed into the deposition chamber in a compound ~aseous form, m~ .t advantageously as silicon tetrafluoride {SLF4)~
The glow discharge plasma preferably is obtained from a silicon tetrafluoride and hydrogen ~H2~ gas mixture, with a preferred ratio range o~ from about 4:1 to 10:1. Preferably, the deposition system is operated at a pressure in the range of about 0.3 to 1~5 Torrl preferably between 0.6 to 1.0 Torr such as about 0.6 Torr.
The semiconductor material is deposited from a self-sustained plasma onto the substrate which is heated, prefeL-ably by infrared means to the desired deposition temperature for each layer.
The p-doped layers of the devices are deposited at specific temperatures, dependlng upon the Eorm oE
the p-doping material used. The evaporated p-dopant metal vapors can be deposited at the lower temperatures, at or below about 400C, where a ~8~6~3 well compensated silicon material is desired, butit can be deposited at higher temperatures up to about 1000C. The upper limitation on the sub-stratP temperature in part is due to the type of metal substrate 10 utilized. For aluminum the upper temperature should not be above about 600C
and for stainless steel it could be above about 1000C. If a well compensated amorphous silicon layer is to be produced,.which is necessary to form the intrinsic layer in a n-i-p or p-i-ll de-vice~ the substrate temperature should be less than about 400C and preferably àbout 300~
To deposit an amorphous p-doped hydrogen compensated silicon material utilizing the evapor-ated metal vapors of the invention, the substrate temperature is in the range of about 200C to 400C, preferably in the range of about 250C to 350C, and desirably about 300C~
To deposit the sillcon semiconductor material utilizing the p-dopant gases of the invention, the substrate temperature is in the range of about 450C to 800C, preferably in the range of about 500C to 700C~

~39~3 ~ he doping concentrations are varied to prc-duce the desired p, p~~, n or n+ type conductivity as the layers are deposited for each device. Fo~
n or p doped layers, the material is doped wit~ S
to 100 ppm of dopant material as it is deposited~
For n~ or p+ doped layers the material is doped with 100 ppm to over 1 per cent of dopant material as it is deposited. The n dopant material can ~e phosphine or arsine in the above amounts. The p dopant material can be those of the invention deposited at the respective substrate temperatures preferably in the range of 100 ppm to over 5000 ppm for the p~ material.
The glow discharge deposition process in-cludes an a.c. signal generated plasma into which the materials are introduced. The plasma pref-erably is sustained between a cathode and sub-strate anode with an a.c. signal of about lkHz to 13.6 MHz.
Although the p-doping method and materials of the invention can be utilized in devices wit:h various silicon amorphous semiconductor material layers it is preferable that they are utilized with the fluorine and hydrogen compensated glow ~ 39~q~3 discharge deposited materials disclosed in said U.S. Patent No. 4,226,898. In this case, a mix-ture of silicon tetra~luoride and hydrogen is deposited as an amorphous silicon compensated m~terial at or below about 400C, for ~he in-trinsic and n-type layers. In the examples shown in D2, D3 and DS, the p+ layer which is placed upon the electrode layer 16 can be deposited a~ a higher substrate temperature above about 45OOC
which will provide a material which is ~luorine compensated. The material will then not be ef-ficiently hydrogen compensated since the hydrogen does not efficiently deposit with the silicon a~
the higher substrate temperature ranges, and will be swept away with the exhaust gases.
The devices illustrated in Dl and D4 where the p-~ layers are on the outer side of the in-trinsic "i" layer may not have high temperature deposited p+ layers, since substrat~ deposition temperatures above about 450C would destroy the hydrogen compensation underlying character of the layers, the intrinsic "i" layer being one which may be a well hydrogen and fluorine compensate~
amorphous layer in a photovoltaic de~ice. The n 896~3 and n~ type layers in each of the devices are al~oprefera~ly deposited in amorphous fluori.rle and hydrogen compensated form. The conventional n doparlt materials are readily deposited with th~
silicon material at the lower temperatures below about 400~C and result in high doping ef~iciency.
Thus~ in Dl and D4, in these structures each of the layers is amorphou.s silicon and, the p~ layer is best formed with one of the evaporated p-dopant 1~ metal vapors at a substrate tenlperature of at ~r less than about 400C. Using gaseous metal or boron compound p-dopant materials requiring high substrate temperatures is also useful, provided the temperature does not reach a value ~hich des-troys the characteristics of the underlying amor-phous layers.
The second device 26' illustrated in D2 ha~
the opposite configuration from the Dl p-i-n de vice. In the device 26' a p~ layer 28 is firs~
deposited on the base electrode substrate 18, followed by an intrinsic layer 30 and an outer n~
layer 32. In this device, the p~ layer can be deposited at an~ substrate temperature in the range of the invention.

g6~3 The devices 26" and 26'l' illustrated in ~3 and ~4 also are of opposite configuration, bein~
respectively p-n and n-p junction devices. In the device 26", a p~ amorphous silicon layer 34 is deposited on the base electrode substrate 18, followed by an amorphous silicon p layer 36 t then an amorphous silicon n layer 38 ancl finally an amorphous silicon n~ outer layer 40. In the de-vice 26''' the inverse order is followecl with an n+ amorphous silicon layer 42 depositecl first followed by an n layer 44, a p amorphous silicon layer 46 and finally an outer p~ amorphous silicon layer 48.
A second type of p-i-n junction device 26l'"
is illustrated in D5. In this device a first p~
amorphous layer 50 is deposited, followecl by ~n intrinsic amorphous silicon layer 52, an amorphous silicon layer 54 and an outer n~~ amorphous s;licon layer 56. (The inverse of this structure, not illustratecl, also can be utilized.) Following the glow discharge of the vario~s semiconductor layers in the desired order, a Eitth step (E) is performecl preferably in a separate deposition environment. Desirably, a vapor dep-~3~-96~3 osition environmen~ is utilized since it is afaster deposition process than the glow discharge process. In this step~ a TCO layer 58 (trans-parent conductive oxide) is added, for exam~le ~o device 26, which may be indium tin oxide (ITO)~
cadmium stannate (Cd2SnO~, or doped tin oxide (SnO2) .
Following the TCO layer 58, an optional s-t~p six (F) can be performed to provide an electrode grid 60. The grld 60 can be placed upon the top of the TCO layer 5S depending upon the final size of the devices utilized. In a device 26 ha~-ing an area of less than 2 square inches or so, the TCO
is sufficiently conductive such that an e7ectrode lS grid is not necessary Eor good efficiency. If ~he device has a greater area or if the conductivity of the TCO layer is such that it is desired~ the electrode grid 60 can be placed upon the TCO la~er to shorten the carrier path and increase the con-duction efficiency of the devices.
As discussed above, the devices 2G to 26"'1can be formed as described in a conventional gl~w discharge chamber, but preferably are folmed in a continuous process as generally illustrated in 2S Fig. 20 -6~3 In ~ig. 2 a diagrammatic illustratIoll oE ~he continuous processing wherein one deposition area is illustrated. The base electrode substrate 18 is unwound from a payout reel 62 around a pair of rollers 64 and 66 forming a planar deposition area 68 therebe.ween. The substrate 18 is in electri-cal contact with the roller 66 which is coupled to ground by a lead 70. The substrate in the planar area 68 forms an anode adjustably spaced from a cathode plate 72. The cathode is coupled to the output terminal of an r.f. source 74. The area between the anode area 68 and the cathode 72 forms a plasma glow discharge deposition region 76~
Although not illustrated, each of the ele-1~ ments in Fig. 2 is enclosed within an e~acuatea space to isolate the ylow discharge regi~n 76 from the surrounding environment. The deposition gases are introduced into the plasma region 76 as i1_ lustrated by an arrow 78. The dopant material can be introduced in a second flow stream as illus-trated by an arrow 80 or the dopant input can be combined with the deposition gases. The exhausted gases are removed from the plasma region 76 and the system as indicated by an arrow 82.

?3 The deposition area of Fig. 2 can ~Q utilizedin a batch mode by introducing the pr~per ~ix o~
gases to orm each desired layer in succession~
In a continuous process, only one type of material can be deposited in a si~gle pass of the substrate 18 thxough the plasma area ~rom the payout reel 62 to a takeup reel 84; however, the operation of the reels can be reversed at the end of ~he web 18 and a second and succeeding layers can be deposited in successive passes through the plasma region 76 with the introduction of the desired dopant ma-terial in each pass. The temperature of the sub-strate 18 can be controlled by one or more in-frared heat lamps or other sources 86. The glow discharge deposition may occur at a airly slow rate of 2 to 5 angstroms of material thickness deposited per second. ~ssumin~ the deposition of the semiconductor material 5000 angs~roms thick on the substrate 18, the 5000 angstrom layer at 5 angstroms per second would take about lOOQ seconds to complete. This is, of course, feasible but it is prefered to deposit the layers on the substrate 18 in a number of deposition stations to increase the deposition rate, as illustrated in Fig. ~.
Referring to Fig. 3, an overal~ system block ~, 96q~3 diayram is .illustrated to perfor~ the processes of the steps C, D and E of Fig. 1. Step C can be performed in a vapor deposition chamber 88. The oxidized substrate 14 is fed off a payout reel. 90 into and through the chamber 88 where the elec-trode layer is deposited thereon to form the base electrode substrate 18 and ~hen to a takeup re~l 92. The deposition process may be observed through a viewing po.rt 94 visually or by monitoring and control instrumentation.
The electrode layer can be formed with a ~rid : pattern by a mask 96 in the form of a similar web to the substrate 14. The mask 96 is fed off the payout reel 98 into registry with the substrate 14 as it passes through the chamber 88 and then to a takeup reel 100.
Following the deposition of the electrode layer, the base electrode substrate 18 is fed successively into and through a plural.ity of glow discharge chambers 102, 102' and 102", each in-cluding a plasma area like 76 and the other glow discharge elements illustrated in FigO 2. The~
same numerals have been utilized in each Fig. to

-3~-iden~ify identical or s~bstantially identical elements. It also is feasible that all of the chamber deposition areas 76 be enclosed in a single chamber isolated one from another.
The n~i-p device 26 of Dl will be utilized to describe the following specific continuous dep-osition example. In this case, the base electrode substrate 18 is fed off the payout reel 52 into the chamber 102. The deposition gas, such as premixed silicon cetLafiuoride and hydrogen, is fed into the deposition region 76 as indicated by the arxow 78. The dopant material, such as phos-phine, is fed into the deposition region 76 as indicated by the arrow 80. The exhausted yases are removed from the chamber as indicated b~ arrow 82.
Depending on the deposition speed desired and the thickness of the n+ layer 20 to be deposited~
there can be one or ~ore chambers 102 each dep-ositing the n~ doped layer 20~ Each of the cham-bers 102 is connected by an isolation passageway 104. The exha~st 82 from each chamber 102 shou~d be sufficient to isolate each of the chambers;
however, an inert carrier gas can be ble~ int~
each passageway 104 as indicated by an arrow 106 ~96q~3 to sweep the passageway 104 clear of any gases from the chamber on either side of the passa~eway.
'~he doping concentrations can be varied in each of the successive chambers to grade the layers if desired.
The chamber 102' is only fed the premixed ~ deposition ~ases silicon tetrafluoride and h~-drogen shown by the arrow 78' in this example, since it deposits the intrinsic layer 22 without Ony dopant material being introduced. Again, there can be a plurality oE chambers lG2' to in-crease the deposition speed of the layer 22. F~r-: ther, since each of the chambers 102, 102' etc. is depositing on the same continuous web the number of deposition areas 7r, for each layer and thesizes thereof are matched to deposit the desired layer thicknesses for each type of layer for ~he device to be formed, here n-i-p device 26.
The substrate 18 is then fed into the chamber 102" which is fed the deposition gases as in-di~ated by the arrow 78". The p-dopant material is fed into the deposition area as indicated by the arrow 80". In this example, the p-dopant is the evaporated metal vapor since the p+ layer 24 is being applied over the amorphous n~~ and i lay-ers. Again, there can be one or more chambers 102~' and the film 26 from the final chamber 102"
is taken up on the takeup reel 84~
~ A mask 108 compatible with the electrode mask 96 can be fed off a payout reel 110 and passed through the successive chambers 102 in regis~ry with the subs-trate 18. The mask 108 is taken up on a takeup reel 112 following the las~ cham~er 102".
The device film 26 then is ed into a vapor deposition chamber 114 to deposit the TCO layer 58 of step E. The film 26 is fed off a payout reel 116 through the chamber 114 to a takeup reel 118.
A suitable mask 120 can be utilize~ ~ed from a payout reel 122 to a takeup reel 124. If the electrode grid 60 is desired, it can be applied in a similar vapor deposition chamber with a suitable mask (not i~lustrated).
For manufacturing a particular devi~e such ~s the ~-i-n device 26', each of the chambers 102, 102' and 102" are dedicated to depositing a par~
ticular film layer. As stated above each of the chambers is dedicated to depositing one iayer ~p, i, or n) since the deposition materials for other 5 layers contaminate the chamber background en--36~3 vironment. To optimize each layer of the p-n or p-i~n device, it is critical that dopants ~rom the other types of layers are not present since they will interfere with the p~eferable electrical characteristics of the layer. For exampler de-positing a p or n layer first, contamination of the following intrinsic layer by the residual p or n dopant creates localiæed states in the intrinsic layerO The efficiency of the device thus will be reduced by the contamination. The problem of contamination which causes the lower efficiency or the devices has been encountered when a speciric deposition chamber was used ~or making successive layers of p-n or p-i-n devic~s. The contamination of the chamber environment is not easily removed so that it presently is not ~easible to utilize a single chamber for more than one layer in a con-tinuous process, since other layers are contam-inated by the residual materials remaining in the background environment.

Claims (26)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A method of making a photovoltaic panel comprising forming a roll of a web of a flexible substrate, unrolling said substrate roll substantially continuously into a partially evacuated space including at least one silicon depositing region therein where where is deposited over at least a portion of said substrate at least two thin, flexible silicon alloy films which are of opposite conductivity (p and n) type, one or more of said alloy films forming a photovoltaic depletion region, wherein said substrate is formed in a substantially continuous web and each of said silicon alloy films is deposited at a separate glow discharge region past which said web is moved to form a substantially continuous deposition process.

2. The method of claim 1 wherein depositing said p-type alloy film includes depositing a material including at least silicon by glow discharge of a compound containing at least silicon in said partial vacuum atmosphere and during the glow discharge deposition of the alloy introducing an evaporated metal p-dopant element into the silicon depositing glow discharge region which metal element is deposited with the glow discharge deposited silicon alloy to form the p-type alloy.

3. The method of claim 2 wherein depositing said p-type alloy film includes depositing on said substrate heater to at least above a temperature of 500°C and below 800°C a material including at least silicon by glow discharge of a compound containing at least silicon in said partial vacuum atmosphere and during the glow discharge deposition of the alloy introducing a p-dopant gaseous compound into the silicon depositing glow discharge region, said p-dopant gaseous compound including at least a p-dopant element and a non-p-dopant substituent and which gaseous compound disassociates into said p-dopant element and said non-p-dopant substituent at said substrate temperature of at least above 500°C, the p-dopant element then combining with the depositing silicon material to form the p-type alloy film.

4. The method of claims 2 or 3 wherein said p-dopant element is at least one of the group consisting of aluminum, gallium, indium, zinc or thallium.

5. The method of claim 3 wherein said p-dopant element is boron.

6. The method of claim 5 wherein said gaseous compound is diborane.

7. The method of claim 1 wherein there is deposited between said p and n doped silicon alloy films an intrinsic amorphous sillcon-containing alloy film by the glow discharge thereof without a p or n dopant element present therein.

8. The method of claim 1 wherein at least one of said silicon-containing alloy films is a substantially amorphous alloy film, and there is included in said silicon compound forming each such alloy film one density of states reducing element and at least one separate density of states reducing element not derived from the compound is introduced into said glow discharge region so that these elements are incorporated in each said substantially amorphous silicon-containing alloy film deposited on said substrate to alter the electronic configurations and produce a reduced density of localized states in the energy gap thereof.

9. The method of claim 1 wherein said p-doped silicon-containing alloy film is deposited to a thickness of less than 1000 angstroms.

10. The method of claim 1 wherein said compound includes fluorine.

11. The method of claim 1 wherein said compound includes at least a mixture of SiF4 and H2.

12. The method of claim 1 wherein said compound includes at least a mixture of SiF4 and H2 in the ratio of 4 to 1 to 10 to 1.

13. The method s defined in claim 1 comprising the further steps of forming one or more electrode-forming regions on said substrate, depositing said films over at least some of said electrode-forming regions and forming said flexible electrode-forming layer separately as to each of said photovoltaic depletion regions.

14. An apparatus for making a photovoltaic panel from a roll of a web of a flexible substrate material comprising:
means for forming said substrate in a substantially continuous web;
means for unrolling said substrate roll substantially continously into a partially evacuated space, said space including at least one silicon depositing region therein, said region including means for depositing over at least a portion of said substrate at least two thin, flexible silicon films which are of opposite conductivity (p and n) type one or more of said films forming one or more photovoltaic depletion regions, said means for depositing said silicon films including means for depositing each of said silicon films at a separate glow discharge region past which said web is moved to form a substantially continuous deposition process.

15. The apparatus is defined in claim 14 wherein said means for depositing said p-type film includes means for depositing a semiconductor host matrix film including at least silicon by glow discharge of a compound containing at least silicon in said partial vacuum atmosphere and means for introducing an evaporated metal p-dopant modifier element into the silicon depositing glow discharge station during the glow discharge deposition of said film, said means for depositing said metal modifier element depositing said element with the glow discharge deposited silicon film to modify same to form the p-type film.

16. The apparatus as defined in claim 14 wherein said means for depositing said p-type film includes means for depositing on said substrate heated to at least above a temperature of 500°C and below 800°C a semiconductor host matrix film including at least silicon by glow discharge of a compound containing at least silicon in said partial vacuum atmosphere and means for introducing an evaporated metal p-dopant modifier element into the silicon depositing glow discharge station during the glow discharge deposition of said film, said means for depositing said metal modifier element depositing said element with the glow discharge deposited silicon film to modify same to form the p-type film.

17. The apparatus as defined in claim 14 wherein said means for depositing said p-type film includes means for depositing on said substrate heated to at least above a temperature of 500°C and below 800°C a semiconductor host matrix film including at least silicon by glow discharge of a compound containing at least silicon in said partial vacuum atmosphere, means for introducing a p-dopant gaseous compound into the silicon depositing glow discharge region during the glow discharge deposition of said film, and means for combining a p-dopant modifier element with the depositing silicon to modify the silicon semiconductor material to form the p-type film.

18. The apparatus as defined in claim 14 wherein said means for depositing said silicon films includes means for depositing at least one of said silicon-containing films from a silicon-containing compound as a substantially amorphous film, for introducing into said silicon films at least one density of states reducing element not derived from said silicon-containing compound so that said element is incorporated in each said substantially amorphous silicon-containing film deposited on said substrate to alter the electronic configurations and produce a reduced density of localized states in the energy gap thereof.

19. The apparatus as defined in claim 14 further comprising means for depositing between said p and n doped silicon films an intrinsic amorphous silicon-containing film by the glow discharge thereof without a p or n dopant modifying element present therein.

20. The apparatus as defined in claim 19 wherein each said silicon depositing region is dedicated to depositing one type film, each of said regions being isolated from other regions.

21. The apparatus as defined in claim 20 wherein each of said dedicated silicon depositing regions is located in one of a plurality of dedicated glow discharge chambers isolated from one another.

22. The apparatus as defined in claim 21 wherein said dedicated glow discharge chambers are isolated from one another by gas flow means.

23. The apparatus as defined in claim 22 wherein said dedicated glow discharge chambers are coupled to one another by an isolation passageway through which said substrate is passed.

24. The apparatus as defined in claim 21 wherein each of said dedicated chambers includes means for heating said substrate.

25. The apparatus as defined in claim 20 wherein each of said depositing regions is of a length along which said substrate is passed proportional to the thickness of the total film to deposit the proper thickness of film thereon in a single pass therethrough.

26. The apparatus as defined in claim 14 including means for forming one or more electrode-forming regions on said substrate, means for depositing said films over at least some of said electrode-forming regions and means for forming said thin flexible electrode-forming layer separately as to each of said photovoltaic depletion regions.