Europäisches Patentamt
`
`European Patent Office
`
`Office européen des brevets
`
`(cid:6)(cid:27)&(cid:11)(cid:11)(cid:12)(cid:18)(cid:14) (cid:12)(cid:15) (cid:23)(cid:12)(cid:6)
`EP 1 732 142 A1
`
`(11)
`
`(19)
`
`(12)
`
`(43) Date of publication:
`13.12.2006 Bulletin 2006/50
`
`EUROPEAN PATENT APPLICATION
`(51) Int Cl.:(cid:3)
`H01L31/068(2006.01)
`
`(21) Application number: 05105081.3
`
`(22) Date of filing: 09.06.2005
`
`(84) Designated Contracting States:
`AT BE BG CH CY CZ DE DK EE ES FI FR GB GR
`HU IE IS IT LI LT LU MC NL PL PT RO SE SI SK TR
`Designated Extension States:
`AL BA HR LV MK YU
`
`(71) Applicant: Shell Solar GmbH
`81739 München (DE)(cid:3)
`
`(72) Inventors:
`• Froitzheim, Armin
`81739 München (DE)(cid:3)
`
`• Münzer, Adolf
`85716 Unterschleissheim (DE)(cid:3)
`
`(74) Representative: Rau, Manfred
`Rau, Schneck & Hübner
`Patentanwälte
`Königstrasse 2
`90402 Nürnberg (DE)(cid:3)
`
`(54)
`
`Si solar cell and its manufacturing method
`
`(57)
`A solar cell comprising a silicon body layer of n-
`type bulk doping having a front side and a rear side; a
`non-(cid:3)alloyed p-(cid:3)type doped region at the rear side of the
`body layer and forming a p/n-(cid:3)junction therewith; a rear
`
`side contact in electrical connection with the p-(cid:3)type
`doped region; and a front side contact in electrical con-
`nection with the n-(cid:3)type silicon body layer, and a method
`of manufacturing such a solar cell.
`
`Printed by Jouve, 75001 PARIS (FR)
`
`EP1 732 142A1
`
`Petitioner Canadian Solar Inc. - Ex. 1006, p. 1
`
`

`

`Description
`
`EP 1 732 142 A1
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`Field of the Invention
`(cid:3)[0001] The present invention relates to a solar cell and to a method of manufacturing a solar cell.
`
`Background of the Invention
`(cid:3)[0002] Conventional silicon (Si) based solar cells comprise a base layer of crystalline Si doped with boron. At the side
`facing the sun during normal operation (the front side), an n-(cid:3)doped layer is provided e.g. by diffusion of phosphorous
`into the base layer. This so-(cid:3)called emitter layer forms a p/n-(cid:3)junction with the p-(cid:3)doped base layer. Front and rear side
`contacts are arranged that are in electrical connection with the n-(cid:3)doped layer and the p-(cid:3)doped base layer, respectively,
`to withdraw a photocurrent from both sides of the p/n-(cid:3)junction.
`(cid:3)[0003] A known problem in such solar cells is degradation of the boron doped base layer, which is also referred to as
`light induced degradation (LID). The degradation adversely affects the minority carrier lifetime, i.e. electron lifetime in
`the boron (p) doped base layer, and is observed in both monocrystalline as well as in multi- or polycrystalline materials.
`(cid:3)[0004] Research efforts have been directed in the past towards understanding the mechanisms underlying the deg-
`radation. The degradation is generally attributed to the formation of boron- and oxygen related defects under illumination,
`or minority carrier injection in the dark, although the precise mechanism is not yet fully understood.
`(cid:3)[0005] Degradation of minority carrier lifetime impairs the efficiency of the solar cell, so one of the prime parameters
`of such solar cells is subject to degradation.
`(cid:3)[0006] Several ways have been explored to circumvent this problem. For example, it has been proposed to use gallium
`(Ga) instead of boron as dopant. However, although Ga doped Si shows a much better stability indeed, this is not seen
`as a practical alternative for commercial cell production as Ga-(cid:3)doped crystalline Silicon is much more difficult to grow
`in a Czochralski process due to a difference in segregation coefficients.
`(cid:3)[0007] Another alternative that has been considered is to reverse the doping of the base layer and the front side emitter
`layer, so that an n-(cid:3)type base layer of Phosphorous doping is used, at the front side of which a p-(cid:3)doped emitter layer is
`provided that forms a p/n-(cid:3)junction with the base layer. In such a cell, the base layer shows much better stability than
`boron-(cid:3)doped base layers in conventional cells, however degradation mechanisms of the front surface passivation of the
`p-(cid:3)type front emitter are observed here.
`(cid:3)[0008]
`It has further been proposed to arrange the p/n junction at the rear side of an n-(cid:3)type Silicon base layer, wherein
`the p/n-(cid:3)junction is formed by alloying Alumina with Silicon. Such cells have been disclosed in Cuevas et al., Proc. 3rd
`World Conference on Photovoltaic Energy Conversion, 2003, vol. 1, p. 963-966 and Schmiga et al., 19th European
`Photovoltaic Energy Conference, 7 June 2004. The efficiencies of such solar cells with Czochralski (Cz) silicon was
`however not higher than 15.8%, which is well below efficiencies of above 17% that are achieved with conventional cells
`made on the basis of a p-(cid:3)type Si body layer.
`(cid:3)[0009]
`It is an object of the present invention to provide a new type of Si solar cell that shows improved stability and
`good efficiency.
`
`Summary of the Invention
`(cid:3)[0010] To this end the invention provides a solar cell comprising
`a silicon body layer of n-(cid:3)type bulk doping having a front side and a rear side;
`-
`a non-(cid:3)alloyed p-(cid:3)type doped region at the rear side of the body layer and forming a p/n-(cid:3)junction therewith;
`-
`a rear side contact in electrical connection with the p-(cid:3)type doped region; and
`-
`a front side contact in electrical connection with the n-(cid:3)type silicon body layer.
`-
`(cid:3)[0011] Applicant has found that by arranging a non-(cid:3)alloyed p-(cid:3)type doped region at the rear side of an n-(cid:3)type Silicon
`body layer, a rear junction solar cell is obtained that can achieve significantly higher efficiencies than known from cells
`based on an Al-(cid:3)Si alloy. Preferably, the non-(cid:3)alloyed p-(cid:3)type doped region is a doping region obtained by diffusion of a
`dopant into the base layer, in particular by diffusion from a p-(cid:3)type dopant from a liquid that is applied to the rear side.
`Preferably the p-(cid:3)type dopant is Boron, but other dopants are also possible.
`(cid:3)[0012] Not using an Al alloy has the further advantage that an Al alloy is opaque, whereas a doping region as obtained
`by diffusion only without alloying is transparent. This allows in a special embodiment to allow a maximum of light to be
`received as well through the rear side, such as diffuse or reflected light. In principle only the rear contacts provide shading.
`(cid:3)[0013] Furthermore, the use of n-(cid:3)type Si as body layer overcomes the degradation problem of conventional solar cells
`with a boron doped body layer, and at the same time the invention circumvents the use of a p-(cid:3)doped front layer that can
`
`2
`
`Petitioner Canadian Solar Inc. - Ex. 1006, p. 2
`
`

`

`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`EP 1 732 142 A1
`degrade as well. The placement of the p/n-(cid:3)junction at the rear side of the cell has the further advantage that the photo-
`current maximised since light cannot be unproductively absorbed in a front-(cid:3)side emitter, in particular in the so-(cid:3)called
`dead zone of such an emitter. This maximises in particular the contribution of short wavelength light to the photocurrent,
`because such light is predominantly absorbed close to the front side, whereas longer wavelength light penetrates deeper
`into the cell.
`(cid:3)[0014]
`In fact, in principle no doping is required at the front side for a functioning cell, and in particular not in regions
`that are not covered by front contacts anyway.
`(cid:3)[0015] Preferably, the cell further comprises a n+ doped region between the front contact and the n-(cid:3)type silicon body
`layer. Such n+ doped regions form a n+/n junction with the base layer underneath the front contacts, and thereby create
`a driving force for photogenerated minority carriers away from the front contact. This can also be referred to as creating
`a ’front surface field’, using similar terminology as the well-(cid:3)known ’back surface field’ created by a p+/p junction at the
`rear side of a conventional cell.
`(cid:3)[0016] A further advantage of arranging an n+ doped region underneath the front contact is that a better electrical
`connection is obtained, in particular when screen-(cid:3)printed/(cid:3)fired front contacts are used. In this way a possible metal/
`semiconductor Schottky contact can be overcome by tunnelling.
`(cid:3)[0017]
`In a particular embodiment the n+ doped region is a shallow n+ doped region. The term shallow can refer to a
`relatively light n+ doping and/or to a doping that does not extend very deep, such as less than 1 micrometer, into the
`body layer. A shallow doping is preferred for minimizing recombination at the front surface. The doping is preferably just
`sufficient for a good electrical contact with the front contact.
`(cid:3)[0018] Typically, the front contact has a finger structure striking a compromise between minimum shading of the front
`surface and optimum withdrawal of photocurrent. The n+ doped region between the front contact and the body layer
`can extend over the entire or substantially all of the front surface, or it can only be arranged surrounding the front contact.
`In the latter case, the n+ doping on the front side suitably covers at the highest 200% of the front surface area that is
`covered by the front side contact, preferably at the highest 150%, more preferably 130% thereof. In this way, a selective
`front surface field can be obtained, wherein the n+ doping underneath the contacts can be chosen such that a sufficient
`front surface field is achieved and for good contacting is provided, but recombination at the front surface area not covered
`by the front contact can be minimized. It can however be desired to arrange a region at the front side, not covered by
`the front contact, with additional n-(cid:3)type doping, in particular at a level between that of the n+ region and that of the n-
`type Si body layer. This can be advantageous for example for optimum passivation of the front surface.
`(cid:3)[0019] The solar cell further suitably comprises a passivation layer at the front side, in order to minimize recombination
`losses at the front surface. The layer can preferably for example be an oxide such as SiO2, a nitride such as SiN, undoped
`amorphous silicon, or n+ doped amorphous silicon layer.
`(cid:3)[0020]
`In a particular embodiment of the invention the rear contact of the solar cell covers only part of the rear side.
`In this way, less material such as Ag or Al is required for the contacts, which could have the shape of a finger structure.
`Moreover it can be arranged that light can also be received through the rear side.
`(cid:3)[0021]
`In particular, the p-(cid:3)doped regions at the rear side can only be arranged between the rear-(cid:3)side contact and the
`n-(cid:3)doped body layer, i.e. immediately surrounding the rear side contacts but not in substantially all of the remaining area
`on the rear side that is not covered by the rear side contact. In this way recombination at the rear side can be minimized.
`The p-(cid:3)dopes region can cover for example 200% of the rear surface area that is covered by the rear contact, preferably
`at the highest 150%, more preferably 130% or less thereof. In a rear surface region between the p-(cid:3)doped region a
`weaker p-(cid:3)doping can be applied as well.
`(cid:3)[0022] The invention further provides a method of manufacturing a solar cell, the method comprising
`providing a silicon body layer of n-(cid:3)type bulk doping and having a front side and a rear side;
`providing a non-(cid:3)alloyed p-(cid:3)type doped region at the rear side of the body layer so as to form a p/n-(cid:3)junction therewith,
`by applying a p-(cid:3)dopant containing liquid to the rear surface and allowing the p-(cid:3)dopant to diffuse into the body layer;
`providing a rear side contact in electrical connection with the p-(cid:3)doped region; and
`providing a front side contact in electrical connection with the n-(cid:3)type silicon body layer.
`
`-
`-
`
`-
`-
`
`Brief description of the Drawings
`(cid:3)[0023] The invention will now be described in more detail and with reference to the accompanying drawings, wherein
`Figure 1 shows schematically a first embodiment of a solar cell of the invention in cross-(cid:3)section;
`Figure 2 shows schematically a top view of a finger contacting structure of a solar cell;
`Figure 3 shows schematically a second embodiment of a solar cell of the invention in cross-(cid:3)section;
`Figure 4 shows schematically a third embodiment of a solar cell of the invention in cross-(cid:3)section;
`Figure 5 shows schematically a fourth embodiment of a solar cell of the invention in cross-(cid:3)section.
`
`3
`
`Petitioner Canadian Solar Inc. - Ex. 1006, p. 3
`
`

`

`EP 1 732 142 A1
`Figure 6 shows Internal quantum efficiency IQE as a function of wavelength l in nm of an n-(cid:3)type solar cell with front
`p/n junction (curve a) and a conventional p-(cid:3)type (curve b) solar cell, respectively, and for the p-(cid:3)type cell also the for
`rear side illumination (curve c);
`Figure 7 shows IQE of a front (curve a) and rear (curve b) junction n-(cid:3)type silicon solar cell;
`Figure 8 shows IQE for cells according to the invention at three thicknesses, a) 150 Pm, b) 200 Pm, and c) 250 Pm;
`Figure 9 shows IQE of solar cells with a standard front surface field diffusion (curve a) and a shallow front surface
`field diffusion (curve b);
`Figure 10 shows effective lifetime teff at an injection level of 1015.cm-3 excess charge carriers, measured on textured,
`both side phosphorous diffused, and SiN coated n-(cid:3)type silicon wafers before (wide bars) and after firing (narrow bars);
`Figure 11 shows Voc (closed squares) and JSC (open circles) of solar cells processed in parallel to the lifetime
`samples from Fig. 10;
`Figure 12 shows the Current-(cid:3)Voltage (I-(cid:3)V) characteristic of a large area (148.9 cm2) n-(cid:3)type solar cell according to
`the invention, fully screen-(cid:3)printed, provided with a shallow Front surface filed and SiN passivated; and
`Figure 13 shows degradation of efficiency on a relative scale with respect to starting efficiency as a function of
`illumination time T for the cell of Figure 12 (curve a) and a standard p-(cid:3)type cell (curve b).
`
`Where the same reference numerals are used in different Figures, they refer to the same or similar objects.
`
`Detailed Description of the Invention
`(cid:3)[0024] Reference is made to Figure 1 showing schematically a cross-(cid:3)section through a first embodiment of a solar
`cell 1 according to the invention. The solar cell 1 comprises
`a silicon body layer 3 of n-(cid:3)type bulk doping having a front side 6 and a rear side 7. The silicon body layer 3 is made from
`crystalline silicon, and can be of the mono-, multi- or polycrystalline type. Preferably, Czochralski grown Silicon is used.
`The thickness of the body layer can be typically between 10 and 400 Pm, suitably between 100 and 300 Pm, preferably
`between 150-250 Pm. A suitable dopant is for example phosphorous. The doping level is generally in the range of 1013-
`1018 atoms per cm3, typically in the order of 1015-(cid:3)1016. Suitably, the doping is such that a resistivity higher than 0.5 and
`lower than 150 Ohm.cm is obtained, preferably higher than 1 and lower than 20, more preferably lower than 10 Ohm.cm.
`(cid:3)[0025] At the rear side 7, a p-(cid:3)doped layer 10 is arranged suitably such as discussed further below, and a p/n-(cid:3)junction
`12 with the body layer is formed.
`(cid:3)[0026] The doping level of the p-(cid:3)region can be higher that what one would use at a front contact, since at the rear side
`that is away from the light-(cid:3)receiving front side there most of the light has already been absorbed and there is less risk
`for recombination. A typical surface resistivity for the p-(cid:3)doped region in the present invention is in the range of 3 Ohm/sq
`or higher and 60 Ohm/sq or lower, preferably 5 Ohm/sq and higher, preferably 50 Ohm/sq and lower, such as 6 Ohm/sq.
`Sometimes a p-(cid:3)doped region in the doping levels required for such surface resistivities would is also referred to as a p+
`region. Ohm/sq (equivalent to Ohm/(cid:3)square) is a commonly used parameter to characterized surface resistivity, which
`is also referred to as sheet resistivity. The depth of the p-(cid:3)emitter is typically in the range of 0.1-6 Pm, when Boron diffusion
`from a liquid source is used typically 1-5 Pm, often 1-3 Pm.
`(cid:3)[0027] A rear side contact 15 is arranged in electrical connection with the p-(cid:3)doped layer 10. The rear side contact is
`shown covering a large part or substantially all of the rear side of the cell.
`(cid:3)[0028] At the front side 6, screen-(cid:3)printed front side contacts 16 are provided in electrical connection with the n-(cid:3)type
`silicon body layer. The front side contacts form part of a finger structure 19 such as sketched in top view in Figure 2
`covering typically around 10% of the front service when screen printed contacts are used. The electrical connection is
`provided via a shallow n+ doped layer 17 on the front side 6, i.e. between the front contact and the n-(cid:3)type silicon body layer.
`(cid:3)[0029] The n+ doped layer can have a surface resistivity of typically between 40 and 250 Ohm/sq, preferably 50
`Ohm/sq and higher, more preferably 80 Ohm/sq and higher, in particular 120 Ohm/sq and higher, such as between 130
`and 200 Ohm/sq. The n+ doping level is typically less than 4 Pm deep, in particular 3 Pm or less. The layer typically
`exhibits a concentration profile of n-(cid:3)dopant, which can be determined by SIMS, e.g. having a surface concentration of
`6.1021 atoms/cm3 down to 1018 at 2-3 Pm into the material.
`(cid:3)[0030] The front surface is further provided with a passivation layer 18. The passivation layer can for example be an
`oxide such as SiO2 a nitride such as SiN, undoped amorphous silicon, or mildly n+ doped amorphous silicon layer, which
`can be arranged using methods known in the art. It is a further advantage of the present invention that n or n+ doped
`layers can generally easier and better be passivated than p or p+ doped layers. An advantage of using SiN as passivation
`layer is that it has intrinsic charges.
`(cid:3)[0031] Suitably further, an anti-(cid:3)reflection coating is provided as known in the art.
`(cid:3)[0032] During normal operation of the solar cell 1 the front side 6 faces the sun and light is received through the area
`of the front surface that is not covered by the front contact 16. Light is absorbed predominantly in the n-(cid:3)doped body layer
`3, and excess hole-(cid:3)electron pairs are created. The holes form minority charge carriers in the n-(cid:3)type material and travel
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`4
`
`Petitioner Canadian Solar Inc. - Ex. 1006, p. 4
`
`

`

`EP 1 732 142 A1
`towards the rear side over the p/n-(cid:3)junction 12 where they become majority carriers in the p-(cid:3)layer that are withdrawn at
`the rear side contact thereby providing photocurrent. The photogenerated excess electrons are prevented from diffusing
`to the front contacts by the field exerted by the n+/n junction leading between the body layer 3 and the p+ layer 18.
`(cid:3)[0033] Reference is made to Figure 3 showing a second embodiment of a solar cell 31 according to the invention.
`Cell 31 differs from cell 1 of Figure 1 in that the rear contacts 35 are arranged in a finger structure covering only part of
`the rear side, e.g. similar to that of Figure 2. A typical coverage is in the range of 10-40% of the rear surface. Between
`the fingers a passivation layer 34 is arranged, for which the same type of layers can be used as for the front side. If the
`cell is suitably arranged for operation such that reflected or diffracted light is received at the rear side not facing the sun,
`additional photocurrent can be created by photons absorbed predominantly in the body layer 3 from the rear side.
`(cid:3)[0034] A further difference with the cell of Figure 1 is the structured front surface field. Regions 38 of high n+ doping
`are arranged surrounding the front contacts 16, and regions 39 of weaker n+ doping in between front contact fingers,
`wherein the weaker n+ doping can be achieved by a doping concentration intermediate between that of the n-(cid:3)type body
`layer and that of the n+ doped regions 18, and/or in that regions 39 are shallower than regions 38.
`(cid:3)[0035] Reference is made to Figure 4 showing a third embodiment of a solar cell 41 according to the invention. Cell
`41 is similar to cell 31 of Figure 3, but in this case the front surface field is only provided by n+ doped regions 38, and
`the remainder of the front side of the body layer 3 is not provided with further doping in addition to the basic doping of
`the body layer.
`(cid:3)[0036] Reference is made to Figure 5 showing a fourth embodiment of a solar cell 51 according to the invention. Cell
`51 is similar to cell 41 of Figure 4, however the p doping at the rear side is only provided in regions 55 surrounding the
`rear contacts 35. In this way a maximum amount of, in particular short-(cid:3)wave, light that is received through the rear side
`of the cell can contribute to the generation of photocurrent. This embodiment allows an optimizing of the fractional rear
`surface area used for p-(cid:3)doped regions, independent of the area covered by rear side contacts. This is a further advantage
`over cells with Al alloyed rear contacts, in which the area of p-(cid:3)doped regions cannot be chosen larger than the area
`covered by opaque Aluminium, there Al serves both as dopant and provides the contacting structure.
`(cid:3)[0037]
`Light that is received through the rear side can be reflected light or diffuse light.
`(cid:3)[0038] The method of manufacturing a solar cell according to the invention will now be discussed.
`(cid:3)[0039] A silicon body layer of n-(cid:3)type bulk doping, typically a wafer in <100> orientation that was sawed from a Cz
`grown and Phosphorous doped crystal, is provided. If desired a short, alkaline crystal-(cid:3)oriented etching can be used to
`texture the surface in order to improve the geometry of incidence of light in order to prevent reflection.
`(cid:3)[0040] Providing a non-(cid:3)alloyed p-(cid:3)type doped region is done by a process which introduces dopant, preferably boron,
`only by diffusion into the rear surface of the body layer. Several processes for boron diffusion are known in the art,
`including diffusion from the gas phase or from liquid and solid phases applied to the surface. A preferred process is that
`known for the making of a p+ back surface field on the rear side of conventional p-(cid:3)type cells. This process as well as
`prior art is described in USA patents No. 5899704 and 6096968, which are incorporated by reference.
`(cid:3)[0041]
`In general terms, a diffusion source layer that contains boron as a dopant is applied, e.g. spun, only onto the
`back of the n- doped silicon wafer. The diffusion source layer is one that contains boron and out of which the boron is
`thermally driven. The diffusion source layer is preferably applied by means of a boron doping resist. In addition to boron
`or its compounds, this resist contains powdery SiO2 in a suspension. This doping resist is normally used for generating
`high dopings in power semiconductors. It may be applied in liquid form and may for example be spun on.
`(cid:3)[0042] A suitable doping liquid is Siodop (TM) of Merck. Then the wafer is treated in an atmosphere that contains
`oxygen at a temperature of 900 to 1200 degree centigrade, preferably between 1000 and 1100 degree C, to generate
`an oxide layer and to drive in the dopant.
`(cid:3)[0043] The further steps for providing the front side n+ layer and the front and rear side contacts for manufacturing a
`cell of the invention can also be conducted according to the steps disclosed in US 5899704 and 6096968, incorporated
`by reference.
`(cid:3)[0044] So, after the p-(cid:3)doped region has been formed the oxide layer can be removed from the front surface, e.g. using
`HF, and optionally also the diffusion source layer and the oxide layer from the rear surface, followed by a diffusion of
`Phosphorous from the gas phase to form a front side n+ layer.
`(cid:3)[0045]
`If a selective front surface field is to be arranged, a mask can be arranged on the front surface before removing
`the oxide layer, so that the oxide layer can be selectively etched away, followed by deep n-(cid:3)dopant (typically Phosphorous)
`diffusion into the body layer through the openings in the oxide layer. Then, the oxide layer can be fully removed, and if
`desired a shallow n-(cid:3)doping can be applied, this can be applied to the entire surface as the n+ regions will remain.
`(cid:3)[0046] Similarly, if a structured p-(cid:3)emitter is to be provided as in Figure 5, a mask can be applied to the oxide on the
`rear side and those p-(cid:3)doped regions can be etched away that are not desired, leaving the desired emitter structure in
`place. The desired emitter structure can be precisely adjusted in this way.
`(cid:3)[0047]
`If required after the formation of the n+ doped layer, a separation at the edge of the wafer may be performed,
`e.g. in a plasma.
`(cid:3)[0048] Front and rear side contacts are preferably screen-(cid:3)printed in an industrial production process, and contacting
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`5
`
`Petitioner Canadian Solar Inc. - Ex. 1006, p. 5
`
`

`

`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`EP 1 732 142 A1
`
`is done at elevated temperatures. These can be 700-800 degree centigrade, but lower temperatures are also possible.
`(cid:3)[0049] The squeegee paste for rear side contacts can contain silver particles, filler, oxides, and also an amount of
`Aluminium. It will be understood that at firing the paste in order to make electrical contact with the p region, a thin alloy
`might be formed at the interface between the p-(cid:3)doped region and the contact, but that the p-(cid:3)doped region itself towards
`the p/n junction is still a non-(cid:3)alloyed p-(cid:3)doped region. The front side contact is also suitably produced by screen printing
`and firing.
`(cid:3)[0050] Passivation and antireflection coating can be applied using known methods, before or after contact formation.
`(cid:3)[0051] Overall, a considerable advantage of the solar cell of the invention is that it can be produced with only minor
`adjustments of a known and industrially applied process for conventional p-(cid:3)type cells, and also using the advantages
`of that process as described in the cited US patents. Adjustments concern the adjustment of the doping to achieve
`desired sheet resistances.
`(cid:3)[0052] For further illustration of the invention as a whole and of particular special advantageous features we will now
`discuss experimental data obtained from specific examples of solar cells according to the present invention.
`(cid:3)[0053] Bare wafers have been measured by the El(cid:3)(y)(cid:3)mat-(cid:3)method in order to determine the diffusion length LD of
`charge carriers in the wafer. This method uses a HF-(cid:3)solution in order to passivate the surface and to form a Schottky
`barrier on the rear side oft the wafer. This contact is used to extract minority charge carriers, generated by a laser
`operating at 680 nm that illuminates the wafer from the front side with a spot size of about 1 mm2. The extracted current
`can be used to determine LD.
`(cid:3)[0054]
`In order to clarify whether phosphorous doped silicon (n-(cid:3)type) wafers do not show a degradation, and whether
`LD of the bulk silicon is superior to p-(cid:3)type boron doped silicon, El(cid:3)(y)(cid:3)mat measurements have been taken from a boron
`and phosphorous doped wafers before and after 50h illumination. Area averaged diffusion length LD on 125 x 125 mm2
`for various wafers was determined. It was found that the area average diffusion length decreased from 400 to 220 Pm
`due to illumination of the boron wafer, but that the Phosphorous doped wafer had an stable LD of 440 Pm before and
`after illumination. Taking into account that the wafer thickness is usually below 300 Pm, the n-(cid:3)type wafers are suitable
`for fabricating a stable solar cells.
`(cid:3)[0055] The final solar cells have been investigated by current-(cid:3)voltage (I-(cid:3)V) and internal quantum efficiency (IQE)
`measurements. The quantum efficiency analysis has been performed by illumination of the front and rear side, respec-
`tively.
`(cid:3)[0056] Solar cells were manufactured from solar grade monocrystalline 125 mm x 125 mm Cz (Czochralski)- grown
`p and n-(cid:3)type silicon wafers as body layers, of about 250 Pm thickness as sawn. After etch removal, texture etching was
`applied to the wafers. Three types of cells were manufactured and compared, conventional cells with p-(cid:3)doped body
`layer and front side p/n junction, n-(cid:3)type body layer cells with front p/n junction, and n-(cid:3)type body layer cells with rear p/n
`junction according to the invention.
`(cid:3)[0057] For p-(cid:3)type doping of p- or n-(cid:3)type silicon wafers, to produce a back surface field (BSF) or a p-(cid:3)doped region,
`respectively, boron from a dopant source was diffused into the silicon.
`(cid:3)[0058] Boron diffusion for BSF formation in p-(cid:3)type silicon suitably gives a sheet resistance of e.g. below 20 Ohm/sq.
`A front side emitter on an n-(cid:3)type Si suitably has a sheet resistance of 50 Ohm/sq or above, otherwise recombination
`losses are too high.
`(cid:3)[0059] A phosphorous diffusion as used to arrange front side emitter for p-(cid:3)type silicon, and rear BSF for front contact
`n-(cid:3)type or front surface for n-(cid:3)type silicon. Again, the Phosphorous diffusion process is conducted differently for p- and n-
`type silicon, in order to adopt the diffusion doping to the application, e.g. above 50 Ohm/sq for emitter formation for p-
`type Si wafers and below 20 Ohm/sq for BSF formation for n-(cid:3)type Si wafers. For the FSF in a structure as depicted in
`Figure 1 a weaker/(cid:3)shallower doping of the layer 17 is used as compared to a BSF, typically 50 Ohm/sq and higher.
`(cid:3)[0060] SiN was used in all cases as antireflection as well as surface passivation coating for n- and p-(cid:3)type silicon.
`Screen printing on the front and rear side is used for contact formation. The process sequences are in all cases similar
`except for the doping levels (sheet resistances) as discussed, and can be used on industrial scale.
`(cid:3)[0061] First, the results for conventional p-(cid:3)type and front contact n-(cid:3)type cells will be discussed.(cid:3)
`
`p-(cid:3)type
`17.2
`
`35.5
`
`619
`
`78.3
`
` (%)
`JSC (mA/cm2)
`VOC (mV)
`FF (%)
`
`difference (%)
`
`- 18.7
`
`- 16.6
`
`- 1.6
`
`- 1.1
`
`Table I
`n-(cid:3)type front p/n junction
`14.0
`
`29.6
`
`609
`
`77.4
`
`6
`
`Petitioner Canadian Solar Inc. - Ex. 1006, p. 6
`
`h
`

`

`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`EP 1 732 142 A1
`(cid:3)[0062] The standard process on p-(cid:3)type Si results in 17.2% efficiency, while the process on n-(cid:3)type Si results in 14.0%
`efficiency. The main difference is due to low short circuit current JSC, e.g. only 29.6 mA/cm2 for n-(cid:3)type compared to 35.5
`mA/cm2 for p-(cid:3)type silicon. The other parameters, open circuit voltage Voc and fill factor FF are reasonable.
`(cid:3)[0063]
`In order to investigate the origin of the low JSC, Internal Quantum Efficiency (IQE) measurements have been
`performed on n- and p-(cid:3)type cells, respectively. Figure 6 show IQE measurements on conventional p-(cid:3)type silicon solar
`cells (curve b) and on n-(cid:3)type front junction solar cells (curve a) as a function of wavelength l in nm. Additionally, a rear
`illuminated IQE measurement of the p-(cid:3)type solar cell is shown in curve c. The n-(cid:3)type silicon solar cell has a drastically
`reduced response in the short wavelength regime.
`(cid:3)[0064] SIMS measurements on wafers that have been diffusion doped in parallel with the tested solar cells show a
`deep emitter profile of about 2 Pm. As the passivation seems not to be very effective, most of the charge carriers
`generated in the emitter recombine rather than separate within the p/n-(cid:3)junction. Anyhow, the blue response is significantly
`enhanced compared to a rear illuminated p-(cid:3)type cell, as the recombination in the BSF is even more enhanced than in
`the boron emitter of n-(cid:3)type cells due to lower boron doping. The red (long wavelength) response of the IQE measurements
`indicates a higher diffusion length for n-(cid:3)type compared to p-(cid:3)type silicon, as already measured