PCT
`
`Intematlonal Bureau
`WORLD INTELLECTUAL PROPERTY ORGANIZATION
`
`
`
`(51) International Patent Classification 7 :
`
`H01L 31/068, 31/0352, 31/0224
`
`A1
`
`
`INTERNATIONAL APPLICATION PUBLISHED UNDER THE PATENT COOPERATION TREATY (PCT)
`WO 00/55923
`
`(11) InternationalPublication Number:
`(43) International Publication Date:
`
` 21 September 2000 (21.09.00)
`
`
`PCT/USOO/02609
`(81) Designated States: AE, AL, AM, AT, AU, AZ, BA, BB, BG,
`(21) International Application Number:
`
`BR, BY, CA, CH, CN, CU, CZ, DE, DK, EE, ES, FI, GB,
`GD, GE, GH, GM, HR, HU, ID, IL, IN, IS, JP, KE, KG,
`
`1 February 2000 (01.02.00)
` (22) International Filing Date:
`KP, KR, KZ, LC, LK, LR, LS, LT, LU, LV, MD, MG, MK,
`
`MN, MW, MX, NO, NZ, PL, PT, RO, RU, SD, SE, SG, SI,
`SK, SL, TJ, TM, TR, TT, UA, UG, UZ, VN, YU, ZA, ZW,
`
`(30) Priority Data:
`
`
`ARIPO patent (GH, GM, KE, LS, MW, SD, SL, SZ, TZ,
`17 March 1999 (17.03.99)
`US
`60/ 124,797
`
`
`UG, ZW), Eurasian patent (AM, AZ, BY, KG, KZ, MD,
`09/41 4,990
`7 October 1999 (07.10.99)
`US
`
`
`RU, TJ, TM), European patent (AT, BE, CH, CY, DE, DK,
`
`ES, FI, FR, GB, GR, IE, IT, LU, MC, NL, PT, SE), OAPI
`patent (BF, BJ, CF, CG, CI, CM, GA, GN, GW, ML, MR,
`(71) Applicant: EBARA SOLAR, INC.
`[US/US]; 811 Route 51
`
`NE, SN, TD, TG).
`South, Large, PA 15025 (US).
`
`(72) Inventors: MEIER, Daniel, L.; 264 Barclay Avenue, Pitts-
`
`
`burgh, PA 15221 (US). DAVIS, Hubelt, P.; 1141 Sper—
`ling Drive, Pittsburgh, PA 15221 (US). GARCIA, Ruth, A.;
`
`
`1910 Victoria Lane, Irwin, PA 15642 (US). SALAMI, Jalal;
`1121 Huston Drive, West Mifflin, PA 15122 (US).
`
`(74) Agents: SOCKOL, Marc, A. et 211.; Graham & James LLP, 600
`Hansen Way, Palo Alto, CA 94304—1043 (US).
`
`Published
`With international search report,
`Before the expiration of the time limit for amending the
`claims and to be republished in the event of the receipt of
`amendments.
`
`
`
`(54) Title: AN ALUMINUM ALLOY BACK JUNCTION SOLAR CELL AND A PROCESS FOR FABRICATIN THEREOF
`
`110
`
`
`
`(57) Abstract
`
`A process for fabricating a solar cell is described. The process includes: providing a base layer (118), and fabricating an emitter
`layer (120) Of p—type conductivity on a same side as the non—illuminated surface ( 128) of the base layer to provide a strongly doped p—type
`emitter layer and a p—n junction (124) between the n—type base layer (118) and the p—type emitter layer (120). The base layer of the present
`invention has n—type conductivity and is defined by an illuminated surface (126) and a non—illuminated surface (128) which is opposite to
`the illuminated surface.
`
`
`
`

`

`
`
`FOR THE PURPOSES OF INFORMATION ONLY
`
`Codes used to identify States party to the PCT on the front pages of pamphlets publishing international applications under the PCT.
`Slovenia
`SI
`Lesotho
`ES
`LS
`Albania
`SK
`Slovakia
`FI
`LT
`Lithuania
`Armenia
`SN
`LU
`FR
`Austria
`Senegal
`Luxembourg
`SZ
`Swaziland
`LV
`Latvia
`GA
`Australia
`TD
`Monaco
`Chad
`MC
`GB
`Azerbaijan
`TG
`MD
`GE
`Togo
`Republic of Moldova
`Bosnia and Herzegovina
`MG
`GH
`TJ
`Barbados
`Tajikistan
`Madagascar
`TM
`MK
`Turkmenistan
`GN
`The former Yugoslav
`Belgium
`TR
`GR
`Burkina Faso
`Turkey
`Republic of Macedonia
`TT
`Mali
`HU
`Trinidad and Tobago
`Bulgaria
`UA
`Ukraine
`IE
`Benin
`Mongolia
`UG
`IL
`Mauritania
`Brazil
`Uganda.
`United States of America
`US
`Malawi
`IS
`Belarus
`Uzbekistan
`UZ
`Mexico
`IT
`Canada
`Viet Nam
`VN
`JP
`Niger
`Central African Republic
`YU
`Netherlands
`KE
`Yugoslavia
`Congo
`Zimbabwe
`ZW
`Switzerland
`KG
`Norway
`New Zealand
`KP
`Céte d’Ivoire
`Poland
`Cameroon
`China
`Portugal
`Romania
`Cuba
`Russian Federation
`Czech Republic
`Sudan
`Germany
`Sweden
`Denmark
`Estonia
`Singapore
`
`
`ML
`MN
`MR
`MW
`MX
`NE
`NL
`NO
`NZ
`PL
`PT
`RO
`RU
`SD
`SE
`SG
`
`
`
`AL
`AM
`AT
`AU
`AZ
`BA
`BB
`BE
`BF
`BG
`8.]
`BR
`BY
`CA
`CF
`CG
`CH
`CI
`CM
`CN
`CU
`CZ
`DE
`DK
`EE
`
`Spain
`Finland
`France
`Gabon
`United Kingdom
`Georgia
`Ghana
`Guinea
`Greece
`Hungary
`Ireland
`Israel
`Iceland
`Italy
`Japan
`Kenya
`Kyrgyzstan
`Democratic People’s
`Republic of Korea
`Republic of Korea
`Kazakstan
`Saint Lucia
`Liechtenstein
`Sri Lanka
`Liberia
`
`KR
`KZ
`LC
`LI
`LK
`LR
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`AN ALUMINUM ALLOY BACK JUNCTION
`SOLAR CELL AND A PROCESS FOR
`
`FABRICATION THEREOF
`
`BACKGROUND OF THE INVENTION
`
`The present invention relates to an improved solar cell and a process for
`
`fabricating thereof. More particularly, the present invention relates to a solar
`
`cell including a p-n junction located near a non-illuminated surface of the
`
`solar cell and a process for fabricating thereof.
`
`Solar cells are widely used because they convert easily accessible energy
`
`from a light source, such as the sun, to electrical power to operate electrically
`
`driven devices, e.g., calculators, computers, and heaters used in homes. Figure
`
`1 shows a cross-sectional View of a layered stack that makes up a conventional
`
`silicon solar cell 10. Conventional silicon solar cell 10 typically includes a p-n
`
`junction 24 sandwiched between a p-type base layer 18 and an n—type layer 16,
`
`which is located near an illuminated (front) surface 11. The term “illuminated
`
`surface,” as used herein refers to the surface of a conventional solar cell that is
`
`exposed to light energy when the solar cell is active or under operation. Thus,
`
`the term “non- illuminated surface” refers to a surface that is opposite the
`
`illuminated surface. The basic structure of p—n junction 24 includes a
`
`heavily-doped (about 1020 cm'3) n—type emitter layer (11+) 16 at or near the
`
`illuminated surface 11 and disposed above a moderately-doped (about 1015 cm'
`
`

`

`WO 00/55923
`
`PCT/US00/02609
`
`3) p-type base layer (p) 18. Commercial embodiments of conventional solar
`
`cells typically include an optional anti-reflective coating 14 and a p+ layer 20
`
`that is formed between p-type base layer 18 and p—type silicon contact 22.
`
`A typical depth of p—n junction 24 from the top of the n+ emitter layer
`
`16 measures about.0.5um. A shallow front p—n junction 24 is desired in order
`
`to facilitate the collection of minority carriers that are created on both sides of
`
`p-n junction 24. Each photon of light that penetrates into p-type base layer 18
`
`and is absorbed by base layer 18 surrenders its energy to an electron in a
`
`bound state (covalent bond) and thereby frees it. This mobile electron, and
`
`the hole in the covalent bond it left behind (which hole is also mobile),
`
`comprise a potential element of electric current flowing from the solar cell. In
`
`order to contribute to this current, the electron and hole cannot recombine, but
`
`rather are separated by the electric field associated with p-n junction 24. If
`
`this happens, the electron will travel to n-type silicon contact 12 and the hole
`
`will travel to p-type silicon contact 22.
`
`In order to contribute to the solar cell current, photogenerated minority
`
`carriers (holes in the n+ emitter layer and electrons in the p-type base layer)
`
`should exist for a sufficiently long time so that they are able to travel by
`
`diffusion to p-n junction 24 where they are collected. The average distance
`
`over which minority carriers can travel without being lost by recombining
`
`2
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`with a majority carrier is called the minority carrier diffusion length. The
`
`minority carrier diffusion length generally depends on such factors as the
`
`concentration of defects in the silicon crystal (i.e. recombination centers) and
`
`the concentration of dopant atoms in the silicon. As the concentration of
`
`either defects or dopant atoms increases, the minority carrier diffusion length
`
`decreases. Thus, the diffusion length for holes in the heavily-doped n+
`
`emitter layer 16 is much less than the diffusion length for electrons in
`
`moderately-doped p—type base layer 18.
`
`Those skilled in the art will recognize that the n+ emitter layer 16 is
`
`nearly a “dead layer” in that few minority charge carriers created in emitter
`
`layer 16 are able to diffuse to p-n junction 24 without being lost by
`
`recombination. It is desirable to have n+ emitter layer 16 that is shallow or as
`
`close to surface of emitter layer 16 as possible for various reasons. By way of
`
`example, a shallow emitter layer allows relatively few photons to be absorbed
`
`in n+ emitter layer 16. Furthermore, the resulting photogenerated minority
`
`carriers created in n+ emitter layer 16 find themselves close enough to p-n
`
`junction 24 to have a reasonable chance of being collected (diffusion length 2
`
`junction depth).
`
`Unfortunately, in the conventional solar cell design, the depth of n+
`
`emitter layer is limited and cannot be as shallow as desired. Metal from
`
`3
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`emitter contacts 12, especially those formed by screen-printing and firing, can
`
`penetrate into p-n junction 24 and ruin or degrade it. The presence of metal in
`
`the p—n junction 24 “shorts” or “shunts” the junction. Therefore, although a
`
`shallow and lightly-doped n+ emitter layer 16 is desired in order to enhance
`
`the current produced by the cell, in practice, however, n+ emitter layer 16 is
`
`relatively deeper and more heavily-doped than desired to avoid shunting p-n
`
`junction 24. Consequently, in conventional solar cells, the deep location of
`
`n+ emitter layer 16 compromises the amount of current produced by the cell.
`
`What is needed is a structure and process for fabricating a silicon solar
`
`cell, which has a high minority carrier diffusion length, eliminates shunting of
`
`the p-n junction and does not compromise the amount of current produced.
`
`SUMMARY OF THE INVENTION
`
`In one aspect, the present invention provides a solar cell. The solar cell
`
`includes a base layer having dopant atoms of n—type conductivity and being
`
`defined by an illuminated surface and a non-illuminated surface. The
`
`illuminated surface has light energy impinging thereon when the solar cell
`
`is exposed to the light energy and the non-illuminated surface is opposite
`
`the illuminated surface. The solar cell further includes a back surface
`
`emitter layer that is made from an aluminum alloy to serve as a layer of p-
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`type conductivity. The solar cell further still includes a p—n junction layer
`
`disposed between the non—illuminated surface of the base layer and the
`
`back surface layer.
`
`In another aspect, the present invention provides a process for
`
`fabricating a solar cell. The process includes: (1) providing a base layer,
`
`(2) fabricating an emitter layer of p—type conductivity on a same side as
`
`the non-illuminated surface of the base layer to provide a strongly doped p-
`
`type emitter layer and a p-n junction between the n-type base layer and the
`
`p-type emitter layer. The base layer of the present invention has n—type
`
`conductivity and is defined by an illuminated surface and a non-
`
`illuminated surface. The illuminated surface has light energy impinging
`
`thereon when the solar cell is exposed to the light energy and the non-
`
`illuminated surface is opposite the illuminated surface.
`
`These and other features of the present invention will be described in
`
`more detail below in the detailed description of the invention and in
`
`conjunction with the following figures.
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`The present invention is illustrated by way of example, and not by way
`
`of limitation, in the figures of the accompanying drawings in which:
`
`Figure 1 shows a cross-sectional View of a conventional solar cell.
`
`Figure 2 shows a cross—sectional View of a solar cell, in accordance with
`
`an embodiment of the present invention.
`
`Figures 3A-3D show various stages of fabrication, according to one
`
`embodiment of the present invention, for the solar cell of Figure 2.
`
`Figure 4 shows a graph of a measured I—V curve of an aluminum alloy
`
`back junction dendritic web solar cell.
`
`Figure 5 shows a graph of internal quantum efficiency of an aluminum
`
`alloy back junction dendritic web solar cell.
`
`Figure 6 shows a cross-sectional view of a solar cell, in accordance with
`
`another embodiment of the present invention, with partial metal coverage on
`
`the back side of the solar cell.
`
`Figure 7 shows a cross-sectional View of a solar cell, in accordance with
`
`yet another embodiment of the present invention, whose contacts to the base
`
`layer are at the back of the cell in interdigitated fashion.
`
`6
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`DESCRIPTION OF THE PREFERRED EMBODIMENTS
`
`The present invention will now be described in detail with reference to
`
`the presently preferred embodiments as illustrated in accompanying drawings.
`
`In the following description, numerous specific details are set forth in order to
`
`provide a thorough understanding of the present invention. It will be
`
`apparent, however, to one skilled in the art, that the present invention may be
`
`practiced without some or all of these specific details. In other instances, well
`
`known process steps and/or structures have not been described in detail in
`
`order to not unnecessarily obscure the present invention.
`
`It is important to bear in mind that a minority carrier electron diffusion
`
`length (i.e. electrons in p-type silicon) is generally different from a minority
`
`carrier hole diffusion length (i.e. holes in n-type silicon) for a given doping
`
`concentration. This difference may depend on the nature of the crystalline
`
`defect, which may act as a recombination center. In dendritic web silicon, for
`
`example, the primary defects are precipitates of silicon oxide which nucleate
`
`on dislocation cores. The interface between the silicon oxide precipitate and
`
`the surrounding silicon carries a positive charge because of the mismatch of
`
`the atomic structure of the two materials. A negatively charged minority
`
`carrier electron is attracted to such a positively charged interface. Such a
`
`defect can capture and hold an electron until one of the numerous majority
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`carrier holes wanders by the defect whereupon the electron can fall into the
`
`hole in a recombination event.
`
`The present invention therefore recognizes that for a solar cell substrate
`
`layer, n-type doping may be desirable to achieve a higher minority carrier
`
`diffusion length than that obtained with p—type doping. A solar cell substrate
`
`having n-type conductivity requires an n-type base layer. A p—n junction was,
`
`therefore, fabricated employing the n-type base layer and an emitter layer at
`
`the illuminated side that includes boron, which is a practical choice for a
`
`p—type dopant.
`
`Unfortunately, it was found that boron diffusions frequently leave a
`
`stained silicon surface because of the formation of boron-silicon compounds.
`
`Such a stain seriously degrades the appearance of the surface of a solar cell.
`
`Boron stains may be removed at the expense of additional and unwanted
`
`processing steps such as thermal oxidation of the stained surface followed by
`
`chemically removing the thermal oxide. Thus, according to the findings of the
`
`present invention, for silicon materials in which n-type doping gives higher
`
`minority carrier diffusion length than does p-type doping, the desirability of a
`
`front p—n junction is offset by the presence of a boron stain for the p—type
`
`emitter layer.
`
`Referring back to Figure 1, commercial solar cells 10 have an n'i‘pp+
`
`8
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`structure. The n-type emitter layer 16 is commonly made using phosphorus
`
`diffusion (in which staining is not a problem), the silicon substrate is doped
`
`p—type, and back silicon surface 20 is typically doped p+ to promote ohmic
`
`contact with contact metal layer 22. An ohmic contact has a low resistance.
`
`The voltage developed across an ohmic contact has a linear relationship to the
`
`current flowing through the contact. In contrast, the voltage developed across
`
`a rectifying contact has a non-linear relationship to the current flowing
`
`through the contact. An aluminum—silicon alloy may be used to provide the
`
`p+ doping and the back metal contact. More importantly, the commercial
`
`solar cell 10 has p-n junction 24 located at the front of the cell. As explained
`
`above, therefore, n+ emitter layer 16 is forced to be deeper than desired to
`
`avoid shunting the front junction. If a silicon material with superior minority
`
`carrier diffusion length with n-type doping is to be used with a front p-n
`
`junction, p—type emitter layer is boron—doped to be practical. Boron doping, as
`
`explained above, gives a stained surface, which degrades cell appearance and
`
`which can only be removed with additional and unwanted processing.
`
`The present invention, therefore, provides a p—n junction located at the
`
`back end of the solar cell. Figure 2 shows a solar cell layer stack 110,
`
`according to one embodiment of the present invention, that has the p-n
`
`junction located at the back end (near the non-illuminated surface of the base
`
`

`

`WO 00/55923
`
`PCT/US00/02609
`
`layer), so it is impossible for the front metal contacts 1 12 to shunt the p—n
`
`junction. Since aluminum may be the choice of dopant in the silicon p+ layer,
`
`it is also impossible for the aluminum to shunt the junction.
`
`Solar cell layer stack 110 includes a base layer 1 18 of n-type
`
`conductivity, which is defined by an illuminated surface (front or top surface)
`
`126 and a non-illuminated surface (back or bottom surface) 128. A portion of
`
`base layer 118 near illuminated surface 126 is passivated by a shallow n+
`
`layer 1 16, which is typically formed by phosphorus diffilsion into a silicon
`
`layer. This embodiment, therefore, realizes the advantage of the conventional
`
`n+ layer and does not suffer from boron staining. An optional anti—reflective
`
`(AR) coating 1 14 is disposed atop n+ layer 116. If the coating is present as
`
`shown in Figure 2, silver contacts 1 12 fire through AR coating 1 14 to reach
`
`the surface of n+ silicon layer 1 16.
`
`A p-n junction 124 is formed at non-illuminated surface 128 or the
`
`interface between base layer 1 18 and an aluminum doped silicon layer 120,
`
`which functions as a p+ emitter layer in the solar cell layer stack of the present
`
`invention. An aluminum-silicon eutectic metal layer 122, which results when
`
`aluminum is alloyed with silicon, is disposed beneath p+ layer 120 and serves
`
`as a self-aligned contact to p+ emitter layer 120. It is noteworthy that boron
`
`10
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`doping is absent from the solar cell structure of Figure 2, so that there is no
`
`boron stain to compromise the appearance of the front surface.
`
`The placement of the p—n junction at the back side of cell 110 is a
`
`unique feature of solar cell 1 10 shown in Figure 2. It should be borne in mind
`
`that no other current silicon solar cell design employs such an aluminum alloy
`
`back junction. Furthermore, the depth of n+ layer 116 is of no real concern
`
`regarding shunting because p—n junction 124 is at the back of cell 110.
`
`In accordance with one embodiment, n—type base layer 1 18 is a
`
`dendritic web silicon crystal having a resistivity that is generally between
`
`about 5 Q-cm and about 100 Q—cm, and preferably about 20 Q-cm. The
`
`thickness of the dendritic web silicon substrate crystal is generally between
`
`about 30 um and about 200 um, and preferably about 100 um. In order for
`
`the cell 1 10 to operate at a practical energy conversion efficiency level, the
`
`hole diffusion length in n-type base layer 1 18 is, in one implementation of this
`
`embodiment, slightly less than the thickness of base layer 118. In a preferred
`
`implementation, however, the hole diffusion length is substantially equal to
`
`the thickness of base layer 118 and more preferably exceeds the thickness of
`
`base layer 1 18.
`
`Locating the p-n junction near the back of the cell instead of near the
`
`front of the cell is counterintuitive to one skilled in the art. It is normally
`
`1‘]
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`desirable to have the p-n junction near the front of the cell because the number
`
`of photons absorbed per unit depth decreases as one proceeds from the
`
`illuminated surface toward the back surface. For example, about 52% of the
`
`incident photons are absorbed within the first 5 um of the base layer from the
`
`illuminated surface, while about 10% are absorbed within the second 5 um
`
`slice of depth, and'about 6% are absorbed within the third 5 um slice. In fact,
`
`about 75% of the incident photons are absorbed within the top 30 um of the
`
`base layer.
`
`Since the absorption of photons results in the generation of
`
`electron-hole pairs, and since the minority carrier members of the pairs diffuse
`
`to the junction in order to be collected, placing the p—n junction near the
`
`illuminated surface is reasonable. As a rule of thumb, those minority carriers
`
`created beyond one diffusion length from the junction will not be collected
`
`and so will not contribute to the solar cell photocurrent.
`
`Figures 3A-3D show important intermediate solar cell structures formed
`
`at different stages of a fabrication process, according to one embodiment of
`
`the present invention. The illustrated embodiment of the inventive processes
`
`may begin when a single crystal silicon substrate of n~type conductivity is
`
`fabricated using well-known conventional techniques, e.g., Czochralski, Float—
`
`zone, Bridgman, edge defined film-fed growth (EFG), string ribbon growth,
`
`12
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`cast silicon crystallized by directional solidification, and dendritic web
`
`growth. In a preferred embodiment, however, the dendritic web growth
`
`technique may be implemented to produce an n-type single crystal substrate.
`
`Referring to'Figure 3A, the n-type single crystal serves as a base layer
`
`218, below which a p-n junction of the solar cell is subsequently formed. In a
`
`next step, a strongly doped n-type layer 216 (hereinafter referred to as an “n+
`
`layer”) may be formed atop base layer 218 using diffusion processes well
`
`known to those skilled in the semiconductor and solar cell fabrication art. In
`
`accordance with one embodiment of the present invention, a layer containing
`
`phosphorous is applied to the top surface of base layer 218 and the
`
`phosphorous atoms are driven into base layer 218 from any heat source. The
`
`phosphorous layer'is preferably from a screen printed paste or from a liquid
`
`dopant, and the heat source is a rapid thermal processing (RTP) unit or a belt
`
`fumace maintained at a temperature that is between about 750 °C and about
`
`1050 °C and preferably about 950 °C. At these temperatures, the diffusion
`
`time may generally be between about 30 seconds to about 30 minutes and
`
`preferably be about 5 minutes. As a result, there is formed a phosphorus—
`
`doped n+ passivating layer having a thickness that is generally between about
`
`0.1 um and about 1 um, preferably about 0.3 mm, with a sheet resistance of
`
`generally between about 10 9/square and about
`
`13
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`200 Q/square and preferably about 40 Q/square.
`
`Alternatively, in another embodiment, the top surface of base layer 218
`
`is subjected to phosphorous ion implantation and a portion near the top
`
`surface of base layer is transformed into n+ layer 216.
`
`In accordance with yet another embodiment, base layer 218 is
`
`passivated with a transparent dielectric layer which may have an appreciable
`
`positive or negative charge. In one implementation of this embodiment, the
`
`base layer is passivated with a layer of thermally grown silicon dioxide having
`
`a thickness of between about 50 A and about 500 A, and preferably about 150
`
`A. In this embodiment, the passivation of base layer 218 may be carried out
`
`in a rapid thermal processing unit at temperatures that are generally between
`
`about 800 °C and about 1050 °C and preferably about 1000 °C. Such
`
`temperature treatment is carried out for times that are generally between about
`
`30 seconds and about 30 minutes and preferably about 50 seconds.
`
`Regardless of how the n+ layer is passivated, or whether it is passivated those
`
`skilled in the art will recognize that at the conclusion of forming n+ layer 216,
`
`a high—low junction (n+n) is formed at the top of base layer 218.
`
`An anti-reflective coating 214 may then be deposited on top of n+ layer
`
`216, as shown in Figure 3B. By way of example, plasma-enhanced chemical
`
`14
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`vapor deposition (PECVD) of silicon nitride or atmospheric pressure chemical
`
`vapor deposition (APCVD) of titanium dioxide is carried out to form anti—
`
`reflective coating 214.
`
`'
`
`Next, conventional techniques may be employed to deposit an
`
`aluminum layer 222 on a bottom surface or non-illuminated surface of base
`
`layer 218. According to one embodiment of the present invention, aluminum
`
`is deposited by screen printing a pure aluminum paste onto bottom surface of
`
`base layer 218. In one implementation of this embodiment, aluminum layer
`
`222 is deposited by screen—printing over nearly the entire back of the cell to a
`
`thickness that is generally between about 5 um and about 30 um, and
`
`preferably about 15 um. The partially fabricated solar cell is then heated in a
`
`radiantly—heated belt furnace to a temperature comfortably above the
`
`aluminum-silicon eutectic temperature of about 577 °C, e.g., that may
`
`generally be between about 700 °C and about 1000 °C and preferably about
`
`800 °C, for a long enough duration, e.g., generally between about 0.1 minute
`
`and about 10 minutes and preferably about 2 minutes, to facilitate alloying of
`
`at least some aluminum from layer 222 with base layer 218. During the
`
`alloying process, the temperature treatment inside the belt furnace is high
`
`enough so that aluminum may effectively dissolve silicon. At such high
`
`temperatures, a portion of the base layer that measures approximately one—
`
`15
`
`

`

`WO 00/55923
`
`PCT/US00/02609
`
`third the thickness of the aluminum layer is effectively dissolved to create p+
`
`layer 220 shown in Figure 3C. By way of example, at about 800 °C, a 15 um
`
`thick aluminum layer 222 dissolves a 5 mm portion of a silicon base layer.
`
`Consequently, p+ layer 220 is silicon doped with aluminum and contact layer
`
`222 is an alloy that is about 88% aluminum by weight and about 12% silicon
`
`by weight the eutectic composition.
`
`The p+ emitter layer is created as described hereinafter. As the silicon
`
`cools, it continues to solidify or grow onto the underlying silicon, also
`
`incorporating a modest amount of the aluminum dopant, on the order of
`
`between about 1018 and about 1019 aluminum atoms per cubic centimeter. The
`
`concentration of aluminum atoms is significantly less than the concentration
`
`of silicon atoms, which is about 10,000 times larger at about 5x1022 atoms per
`
`cubic centimeter. As shown in Figure 3C, after cooling the layers, a portion of
`
`base layer 218 near the non—illuminated surface is transformed to a strongly
`
`doped p—type layer. 220 (hereinafter referred to as a “p+ layer”), by liquid
`
`phase epitaxy. Some of the silicon that was dissolved becomes part of the
`
`aluminum—silicon eutectic layer 222. The silicon that was not dissolved by the
`
`aluminum acts as a template for restructuring the silicon crystal as it is
`
`epitaxially regrown.
`
`The aluminum-silicon eutectic metal layer 222 is a good conductor (has
`
`16
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`a low electrical resistance), and serves as a self-aligned metal contact to the
`
`solar cell. The thickness of aluminum-silicon eutectic metal layer 222 is
`
`generally between about 1 um and about 30 um and preferably about 15 um.
`
`Figure 3D shows the structure resulting after depositing metal contacts
`
`and firing through the optional anti—reflective coating. By way of example, a
`
`silver layer is screen-printed onto the n+ layer 216, covering about 7% of the
`
`front surface in a grid pattern. Silver may be deposited by screen—printing to a
`
`thickness that is generally between about 5 um and about 20 um and
`
`preferably about 10 um. The resulting partially fabricated solar cell structure
`
`is then run through a belt furnace maintained at a temperature that is generally
`
`between about 700 °C and about 900 °C and preferably about 760 °C. The
`
`time for running the cell through the furnace is short in order to avoid
`
`allowing impurities in the silver paste from diffilsing into the silicon. In one
`
`embodiment of the present invention, the time for running the cell through the
`
`belt furnace is between about 0.1 minutes and about 10 minutes. In a
`
`preferred embodiment, however, the time is about 40 seconds. The thickness
`
`of silver ohmic contacts 212 to n+ layer 216 is generally between about 1 um
`
`and about 20 um and preferably about 10 um. If an optional anti-reflective
`
`coating is deposited, then the silver layer is fired through the coating. Both
`
`operations may (with or without anti-reflective coating) be performed in a
`
`17
`
`

`

`wo 00/55923
`
`PCT/USOO/02609
`
`radiantly-heated belt fumace. The fabrication steps of the present invention
`
`realize high-throughput and low-cost manufacturing.
`
`Figure 4 depicts an I-V curve, which shows a light energy to electrical
`
`energy conversion-efficiency of about 12.5% for an aluminum alloy back
`
`junction dendritic web silicon solar cell with full back metal coverage of the
`
`present invention. The solar cell employed to obtain the data shown in Figure
`
`4 has a nominal thickness of about 100 um and an area of about 25 cm2.
`
`Other parameters include a short—circuit current density of about 28.2
`
`mA/cmz, open-circuit voltage of about 0.595 V, and fill factor of about 0.747.
`
`The knee of the curve shows a maximum power point, where Vmp is about
`
`0.480V and Imp is about 0.653A.
`
`Figure 5 depicts measurements of spectral reflectivity (solid triangles)
`
`510, external quantum efficiency (open squares) 512, and internal quantum
`
`efficiency (solid squares) 514 for the solar cell of Figure 4. Measured
`
`reflectivity indicates that about 47% of the long wavelength (> 950 nm) light
`
`is reflected from the aluminum~silicon eutectic metal layer at the back surface.
`
`This reflected light then gets a second pass through the silicon and adds to the
`
`overall efficiency. Calculation of the internal quantum efficiency using the
`
`one—dimensional finite element model PClD indicates a hole diffusion length
`
`of about 200 pm in the base of this cell. This is double the cell thickness, as
`
`18
`
`

`

`WO 00/55923
`
`PCT/US00/02609
`
`required for good performance. The internal quantum efficiency curve has a
`
`positive slope over the range of about 500 nm to about 850 nm, indicating that
`
`the p-n junction is at the back of the cell, which is also noteworthy. In a
`
`conventional solar cell, the internal quantum efficiency curve would show a
`
`negative slope, indicating that the p-n junction is at the front of the cell.
`
`Figure 6 shows a solar cell layer stack 300, in accordance with another
`
`embodiment of the present invention, that is substantially similar to the solar
`
`cell 1 10 of Figure 2, except that solar cell 300 has only part of the back
`
`surface converted to a p—type layer and covered with metal. Thus, solar cell
`
`300 is referred to as having “partial” metal coverage on the back side as
`
`opposed to have “full” metal coverage, as shown in Figure 2 for solar cell 110,
`
`where the entire back surface is covered with metal. The partial metal
`
`coverage design of Figure 6 eliminates bowing of the thin cells, which results
`
`from a differential thermal contraction between the silicon and the
`
`aluminum-silicon eutectic metal layer upon cooling from the eutectic
`
`temperature of about 57 7 °C. It is important to bear in mind that such bowing
`
`sometimes occurs when full back metal coverage is used. The segmented
`
`metal also reducesthe cost of aluminum. The exposed n—type base layer
`
`surfaces 330 between metal islands at the back may be passivated in order to
`
`obtain respectable energy conversion efficiency. By way of example, the
`
`19
`
`

`

`WO 00/55923
`
`PCT/USOO/02609
`
`passivation step may be performed after the aluminum alloying step by
`
`growing a thermal oxide of about 115 A thickness in a rapid thermal
`
`processing unit at about 1000 °C for about 50 seconds. An anti—reflective
`
`coating may not be applied in this preliminary effort. Efficiencies up to about
`
`6.8% were obtained for 25 cm2 dendritic web silicon cells with screen-printed
`
`metals but without anit—reflective coating. The expected efficiency with an
`
`anti-reflective coating is about 9.9%.
`
`Figure 7 depicts a solar cell layer stack 400, according to yet another
`
`embodiment of the present invention, that is substantially similar to the solar
`
`cell 300 of Figure 6, except that solar cell 400 has n-type contacts 412 to the
`
`n-type base layer moved from the front of the cell to the back of the cell in an
`
`interdigitated fashion.
`
`This structure shown in Figure 7 has the added advantage of reducing to
`
`zero the shadowing of the front surface by contacts. Interconnecting cells in a
`
`module is also expected to be simpler with both contacts on the back side of
`
`the cell in that surface