Patent Application
20100059117
The present invention relates to the field of silicon solar cells and in particular, it relates to a method of making such solar cells using a hybrid technology with improved energy conversion efficiency and reduced fabrication cost.
Solar cells based on p-type silicon wafers are usually fabricated with a shallow n-type region (emitter) on the light-receiving side by diffusion of an appropriate dopant, such as phosphorous, to convert the top surface layer of the wafer into n-type, followed by passivation of the light-receiving side, for example by hydrogenated silicon nitride, and passivation of the back side, for example by a back-surface field created by a more heavily doped p-type dopant such as Al, and then followed by metallization of both sides for electrical contacts.
However, n-type Czochralski (CZ) silicon wafers have significant advantages over the commonly available boron-doped p-type CZ wafers. This is apparently due to problems associated with the simultaneous presence of both oxygen and boron impurities in standard p-type CZ material that lead to the generation of defects that significantly lower the minority carrier lifetimes in such p-type material. In comparison, silicon wafers without significant oxygen concentration (which is achieved by avoiding the CZ process such as through the use of float-zone wafers) or silicon wafers without significant boron concentration (such as n-type or high resistivity Czochralski wafers) achieve much higher minority carrier lifetimes than the standard p-type CZ wafers that are predominantly used in the commercial production of solar cells. However, most existing equipment and/or processes used in the fabrication of commercial solar cells have been developed for compatibility with p-type wafers and not n-type wafers. Therefore the solar cell industry has yet to incorporate n-type CZ wafers into fabrication processes. Furthermore, for n-type wafers, the use of boron doping is the predominant method of producing p-type regions (emitters). Consequently, merely using n-type wafers will still result in cell structures with regions that simultaneously have both high B and O concentration.
It has been proposed to use a heterojunction produced at the interface between crystalline silicon and an amorphous silicon (a-Si) material layer that is created on the light-receiving surface as a means of avoiding boron doped p-type CZ—Si regions. With this approach n-type CZ wafers are used without the use of any boron doped regions, to retain high minority carrier lifetimes throughout the device.
However, with this approach, the amorphous silicon in the heterojunction structure has very poor conductivity and when used at the light receiving surface, it is not feasible to conduct the generated current in the direction parallel to the cell surface to where the metal contacts are located on the a-Si material. This necessitates the use of a conducting oxide layer (such as indium tin oxide) deposited onto the amorphous silicon layer as shown in prior art. This conducting oxide layer collects the generated charge from the amorphous silicon material and conducts it to where the metal contacts are located thereby minimising the necessity for current flow in the amorphous silicon material. However, a conducting oxide layer adds significantly to the costs of fabricating the solar cells while simultaneously degrading the cell performance through unwanted light absorption and resistive losses such as at the interface with the metal contact. The conducting oxide layer also introduces potential durability problems that may degrade the performance of the cells as they age. This effect is well documented in the literature.
The slight variations in the amorphous silicon layer thickness on the light receiving surface can also have a significant impact on performance. For example, if the amorphous silicon is slightly thicker than optimal, significant absorption of light will occur within the amorphous silicon material which cannot contribute to the cell's generated current. This particularly degrades the cell's response to shorter wavelengths of light. On the other hand, if the amorphous silicon is slightly less than optimal thickness, this will lead to poorer effective surface passivation with a corresponding degradation in device voltage. Even the optimal thickness of the amorphous silicon material is a trade-off between these two loss mechanisms with some loss in short wavelength response and some loss in voltage.
According to a first aspect, a solar cell is provided comprising:
i) a crystalline silicon layer having a front, light receiving, surface and a back surface;
ii) an amorphous semiconductor layer forming a heterojunction with the crystalline layer on its back surface; iii) a first contact structure contacting the crystalline layer and a second contact structure contacting the amorphous layer.
The device may be formed on a silicon wafer or on a thin crystalline silicon film on a glass or other suitable substrate.
The second contact structure is in contact with, and located over, the amorphous layer on the rear surface and may be a continuous contact layer or may be an intermittent structure such as a grid or a set of fingers. In the case of a rear-surface n-type self-aligned metallisation interdigitated with the heterojunction structure, the amorphous layer may be continuous over the entire rear surface, or alternatively both the amorphous layer and the second contact grid/fingers may be deposited with the same intermittent structure on the rear so that the metal contact is aligned to the amorphous silicon layer.
The first contact structure may be an intermittent structure such as a grid or a set of fingers located over the front, light receiving surface of the crystalline silicon layer, or in the case of a rear-surface n-type self-aligned metallisation interdigitated with the heterojunction structure, the first contact structure (also on the rear) may be eventually isolated from, but initially located over, the amorphous layer if the amorphous layer is continuous across the entire rear surface. In this case, the first contact will be treated so that it extends through the amorphous layer at spaced locations to contact the back surface of the crystalline silicon layer. In the latter case one of the first and second contact structures will be inter-engaged over the back surface to allow distributed contact to both the crystalline and amorphous regions.
According to a second aspect a method of forming a heterojunction on a rear surface of a precursor to a silicon solar cell, opposite to a front, or light-receiving, surface, comprises:
The method may commence with a silicon wafer or on a thin crystalline silicon film on a glass or other suitable substrate. Preferably, in the case of a wafer, the doped silicon wafer is an n-type silicon wafer, on which surface damage removal, texturing and cleaning are first performed. The front surface of the wafer preferably has a silicon nitride layer applied by a PECVD deposition incorporating phosphorus dopants. This silicon nitride layer is arranged to induce an electron accumulation layer beneath the silicon nitride layer.
The amorphous semiconductor layer is preferably hydrogenated amorphous silicon, hydrogenated amorphous silicon carbide, or hydrogenated amorphous silicon germanium alloy. Hereinafter we shall use hydrogenated amorphous silicon as an example.
The second contact is preferably formed by a layer of metal or layers of metals, such as by sputtering aluminium.
The first contact structure is preferably made with plated metals such as Ni, Cu or Ag on heavily doped n++ regions in an n-type crystalline silicon wafer or an n-type crystalline silicon film. The heavily doped n++ regions are preferably produced by laser doping of phosphorous dopants.
The n++ regions are preferably cleaned before electroless/electro plating of metal contacts, such as nickel followed by copper followed by emersion silver to replace surface atoms of copper with silver. Metal sintering is then preferably performed (if this was not already done after Ni plating.)
Alternatively, in the case of wafer devices, front surface first contacts can be formed before the rear heterojunction formation, in which case an oxide layer is temporarily formed over the rear surface of the crystalline silicon, and removed again prior to forming the amorphous silicon layer of the heterojunction and subsequently the rear metal contacts.
In another alternative method, the front surface structure is formed by;
The resulting front structure then has the first contact added as described above.
In a rear-surface n-type self-aligned metallisation interdigitated with heterojunction structure, the first contact to the crystalline silicon wafer or thin crystalline film is formed on the back surface and is laser-doped either through the rear amorphous silicon layer if the amorphous layer is continuous or through the gaps in the rear amorphous silicon layer if the amorphous layer is intermittent. Formation of both the first contact and the second contact on the rear surface comprises the following actions:
Preferably the process of forming the contacts in this form of the rear heterojunction device comprises:
In the case of silicon wafers, following the formation of the rear contacts to the silicon wafer, PECVD depositions of hydrogenated silicon nitride, incorporating phosphorus dopants, are performed to the front surface of the silicon wafer. This silicon nitride layer is arranged to induce an electron accumulation layer beneath the silicon nitride layer.
When the rear-surface n-type self-aligned metallisation interdigitated with heterojunction structure is applied to a thin-film n-type crystalline silicon on glass device with a rear surface n-type self aligned metallisation through the use of laser doping as above, the method comprises;
In this case a front surface silicon nitride layer, incorporating phosphorus dopants, is preferably applied to the glass substrate before the crystalline silicon layer is applied. Otherwise the preferred process is similar to that for a doped wafer.
Embodiments of the invention will now be described, by of example, with reference to the accompanying drawings in which:
Referring to the accompanying drawings, a number of embodiments of solar cells employing rear heterojunction structures are illustrated.
In these embodiments the heterojunction is located at the rear surface removing the requirement for the conducting oxide layer normally required for lateral conductivity in the case when the heterojunction is located on the light receiving (front) surface and also reducing the sensitivity of performance to the thickness of the amorphous silicon layer within the heterojunction structure. In the embodiments described here, the light passes through the crystalline silicon region first, substantially avoiding the situation of having short wavelength light passing through the amorphous silicon layer. This also facilitates the use of metal across the entire rear surface of the amorphous silicon layer therefore avoiding the need for the conducting oxide layer to carry current in the direction parallel to the cell surface.
However the use of the heterojunction at the rear increases the distance that carriers generated near the light-receiving surface have to travel to the collecting junction at the rear. Therefore high resistivity and high quality wafers are preferably used (regardless of whether the structure is developed for use with n or p-type wafers) or the crystalline region is fabricated as a thin film or both. If using n-type wafers, a contacting scheme for the n-type material is required for the top surface (or else interdigitated with the contact to the heterojunction at the rear surface), whereby heavy doping beneath the metal contact is desirable so as to minimise contact resistance and minimise the contribution of the metal/silicon interface to the device dark saturation current. To avoid degradation of the wafer surface or wafer material, no high-temperature thermal processes should be used prior to depositing the amorphous silicon material needed for the heterojunction. Following the deposition of the hydrogenated amorphous silicon, subsequent device processing should also be compatible with the existing structure to avoid degradation of the heterojunction or surface passivation quality.
Conducting the majority carriers from within the bulk to the n-type metal (first) contact (such as the front-surface metal contact) is a challenge in high resistivity wafers without the use of a separate front-surface diffusion of the same polarity, which in this case is not compatible with the use of the heterojunction on the rear. A conventional front-surface diffusion cannot be used after the formation of the rear heterojunction due to the loss of hydrogen from the amorphous silicon or even damage to the amorphous silicon material such as through crystallisation at the temperatures needed. On the other hand, such a diffusion process is also undesirable prior to heterojunction formation due to problems created at the rear surface during the thermal process and associated handling such as through defect generation, surface roughening, contamination of the surface, surface oxidation, or simply unwanted dopants or other impurities diffusing into the surface. Metal contacting schemes used with any of the current commercial cell technologies (such as screen-printed solar cells, buried contact solar cells, point contact solar cells, etc.) are generally unable to achieve all of the above, primarily due to their dependence on high-temperature thermal processes, either in conjunction with necessary diffusion processes or else firing of the metal contacts.
Referring to
In a first alternative, formation of laser doped transparent conductors as described by Wenham et alia in Australian Provisional applications Nos. AU 2005926552 & 2005926662 “Low area screen printed metal contact structure and method” (incorporated herein by reference) can be used to conduct the current to the self-aligned metal contacts, whereby the transparent conductors preferably run perpendicularly to the metal lines. In this configuration, all the laser doping for the transparent conductors and the self-aligned metallisation can be done in a single process by using different laser conditions for the transparent conductors whereby the overlying dielectric layer and/or antireflection coating and/or diffusion source are not significantly damaged and thereby still mask the silicon surface from the subsequent plating process. Alternatively, the transparent conductors can be formed prior to a subsequent dielectric/anti-reflection coating/surface passivation layer deposition so as their surfaces are subsequently protected from the plating process that follows the laser doping used for the self aligned metallisation formation.
In a second alternative, electrostatic effects can be used at the surface such as through deliberately incorporating significant levels of charge (positive charge if using an n-type wafer, negative charge if using a p-type wafer) into the surface dielectric layer so as to produce an accumulation layer at the surface to enhance the conduction of majority carriers to the location of either the metal contact or the transparent conductors. For example, incorporating high levels of atomic hydrogen into a silicon-rich silicon nitride layer can achieve this outcome. Other elements can also be potentially used to add positive charge into such dielectric layers. If done properly, these electrostatic effects in conjunction with the dielectric layer can be used to provide superior effective surface passivation. Alternatively, a semiconductor material with an appropriately high bandgap and appropriate doping can be used to give similar band bending near the surface to create such an accumulation layer for improved lateral conductivity for an n-type wafer. The equivalent can be done for a p-type wafer whereby holes are accumulated to the surface to improve the lateral conductivity of the majority carriers which in this case are the holes. An example of such a wide bandgap semiconductor that is compatible with rear-surface heterojunctions is doped hydrogenated amorphous silicon. In this material, the released atomic hydrogen can bond with silicon dangling bonds at the interface to remove the mid-gap states to provide enhanced surface passivation effect. Furthermore, by diffusion of certain elements such as nitrogen or oxygen, the sub-surface region of a crystalline silicon substrate may be converted into a dielectric layer, thereby moving the silicon dangling bonds away from the original crystalline silicon surface and minimizing any negative impact from surface contaminants from imperfect cleaning processes.
In a third alternative, large-area diffusion across the entire top surface can be effected through the use of either rapid thermal processing (RTP) or laser doping in a way that the thermal effects will not degrade the heterojunction at the rear surface. Such techniques can be used with rear heterojunction structures in conjunction with the self-aligned metallisation scheme whereby the top surface RTP or laser diffusion is carried out prior to the laser doping for heavily doped regions to be contacted by the plated metal. In this approach, the same dopant source could be used for both the top surface diffusion and the laser doping for the self-aligned metallisation and/or transparent conductors. For example, the phosphorus source can be incorporated into the silicon nitride antireflection coating and then used as the phosphorus source for top-surface diffusion, transparent conductors and self-aligned metallisation.
In the case of using medium resistivity n-type wafers in the range 1-5 ohms-cm, the sheet resistivity of the wafer itself is adequate to avoid the need for the above approaches for enhancing the lateral conductivity of majority carriers in the wafer to facilitate collection by the first metal contact. Such wafers have demonstrated minority carrier lifetimes high enough for compatibility with a rear junction device design provided wafers are not much thicker than about 200 microns.
Examples of the Implementation of a rear-surface heterojunction structure.
In summary, what is described above is a crystalline silicon based solar cell having an amorphous silicon heterojunction on the rear for separation of photon-generated electron-hole pairs and laser-doped localized regions within the crystalline silicon material for majority carrier conduction.
Some embodiments incorporate a front (light-receiving side) passivation structure using an impurity diffusion mechanism comprising dopants of the same polarity as the wafer, to create an interface with the more lightly-doped wafer that has moved inward to the silicon bulk before depositions of passivating dielectric films onto the silicon front surface.
Other embodiments incorporate a front (light-receiving side) passivation structure using an impurity diffusion mechanism comprising dopants such as nitrogen or oxygen, to create an interface with the doped wafer that has moved inward to the silicon bulk before depositions of passivating hydrogenated amorphous silicon films followed by passivating low-temperature dielectrics like silicon nitride.
Some embodiments also incorporate a localized front electrode made by laser doping of the silicon front surface in localised regions while simultaneously damaging the overlying passivating dielectric or amorphous silicon layers so as to expose the laser doped silicon surface followed by self-aligned metallization of such regions while the passivating layers mask the remainder of the light receiving surface from forming metal contact.
Embodiments may also use a layer or layers of metal(s) directly deposited on said amorphous silicon film as a back electrode.
In an alternative arrangement, some embodiments may incorporate an interdigitated positive/negative electrode structure on the rear surface made by laser doping over patterned back electrode followed by metallization.
In some embodiments front contacts employ the use of transparent conductors formed by laser doping in conjunction with a front metallisation scheme described above whereby the transparent conductors run perpendicularly or at an angle to the metal contact lines so that the transparent conductors intersect with the heavily doped regions beneath the first metal contact.
It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/CN2007/000445 | 2/8/2007 | WO | 00 | 11/24/2009 |
