FIELD OF INVENTION

The present invention relates broadly to a method and system for forming a metal electrode for a solar cell, and more particular to a method and system of forming metal electrodes on the front and rear surfaces of high-efficiency silicon wafer solar cells.

BACKGROUND

A typical silicon wafer solar cell is a simple p-n junction (diode). Like any other diode, it has to have a low-resistance current path to the external circuit by means of metal conductors. In a typical diode this is achieved by coating the entire front and rear surfaces of the diode with metal. However, unlike a normal diode, a solar cell diode must have at least one of the two surfaces (the one facing the sun) semi-transparent, to allow sunlight to enter it. This is typically achieved by using a metal grid instead of a blanket metal layer on the illuminated surface.

To achieve a high solar cell efficiency, the recombination rate (i.e., loss rate) of light- , generated minority charge carriers (electrons or holes) must be minimised, both in the bulk of the wafer and at its two surfaces (front and rear). Bulk recombination is minimised by using a high-quality silicon wafer and clean processing conditions, which minimise contamination of the wafer during solar cell manufacture. Surface recombination is reduced by depositing a surface-passivating dielectric layer onto the surface or by creating a heavily doped region along the surface, or by a combination of both methods. One of the doped layers forms the p-n junction, whereas the other forms a so-called high-low junction or "back surface field" (BSF) layer. The heavily doped layers facilitate the realisation of low-resistance ohmic contacts with the metal electrodes of the solar cell. In the case of a p-type wafer, the resulting structure is dielectric/n ⁺ pp ⁺ /dielectric. Since dielectrics are electrically insulating materials, contact openings need to be realised in the dielectric films on the front and rear surface of the solar cell to extract the photogenerated current, as shown in Figure 1. If these contact openings cover only a small percentage (< 10%) of each surface, the surface is still adequately passivated while at the same time current can be extracted from the solar cell. For laboratory cells, the openings in the dielectric films are usually created using photolithographic methods or laser ablation. For industrial cells, the openings are usually created by screen-printing of a special metal paste onto a dielectric film (usually silicon nitride), followed by annealing of the sample at high temperature (> 700°C). During the high-temperature anneal the metal paste reacts with the dielectric film, thereby dissolving the dielectric film and creating a good ohmic contact with the underlying heavily doped silicon layer.

To further improve the efficiency of the solar cell, localised heavily doped regions can be added underneath the metal contacts on both surfaces. This reduces contact resistance losses as well as recombination losses at the two surfaces. The resulting high-efficiency solar cell structure is shown in Figure 2. This solar cell structure has a one-sun efficiency potential of over 25%, as evidenced by laboratory cells with 24.5% using this architecture (Zhao et alia, 24.5% efficiency silicon PERT cells on MCz substrates and 24.7% efficiency PERL cells on Fz substrates, Progress in Photovoltaics, volume 7, pages 471-474, 1999).

If the rear dielectric film provides excellent surface passivation of the bare silicon wafer, then the solar cell fabrication sequence can be simplified by eliminating the BSF layer along the rear surface. The fabrication sequence can then be further simplified by eliminating the localised heavily doped regions at the front and rear surface, giving the solar cell structure of Figure 3. Despite these process simplifications, this solar cell structure still has a one-sun efficiency potential of well over 20%.

As of today, the solar cell structures shown in Figures 2 and 3 cannot yet be realised cost-effectively in a factory. As a consequence, today's p-type industrial solar cell structures are simpler in design and hence their efficiency is below 19%. A need therefore exists to realise the solar cell structures of Figures 2 and 3 in a cost- effective and industrially feasible way.

Since a silicon wafer solar cell is a low-voltage, high-current device, the electrical resistance has to be minimised to ensure there are negligible ohmic losses. For example, for a 156 mm x 156 mm wafer under one-sun illumination, the cell produces about 8 Amperes of current at a voltage of about 0.5 Volts. Hence the fabrication of the front grid is a compromise between the shading of the solar cell and the electrical series resistance. Due to the high current, the metal fingers of the front electrode should be highly conductive, be high and narrow (i.e., have a good aspect ratio), be closely spaced, and should have a low contact resistance to the underlying silicon. These requirements are not met by screen-printed metal fingers, as these have rather poor aspect ratio (height < 20 micron and width > 80 micron, giving an aspect ratio of less than 0.25). Also, realisation of a good ohmic contact requires a heavily doped surface layer (about 60 Ohm/square), which reduces the blue response of the solar cell. The standard method in the semiconductor industry for this type of task is photolithography combined with wet-chemical etching of the dielectric. However, photolithography is not a high-throughput process and is too slow for silicon wafer solar cell production.

Alternatively, one could deposit the metal lines before the deposition of the dielectric, negating the need to pattern the dielectric film. However, this approach usually suffers from surface contamination of the silicon wafer during metal deposition, resulting in poor solar cell efficiency.

Over the last 20 years a significant amount of research has gone into finding a suitable replacement technique for screen printing of silicon wafer solar cells. Some of the methods that have shown potential are laser ablation and inkjet printing technologies. The general approach is to first form an opening in the dielectric (for example by laser ablation or chemical etching), followed by the deposition of a thin metal layer ("seed layer", see Figure 4). The metal seed layer (typically nickel) is then thickened to the required height (about 10-20 microns) using a metal plating process (usually copper and/or silver). A need therefore exits to provide a method and system for forming a metal electrode for a solar cell, which seek to address at least one of the above mentioned problems.

SUMMARY

In accordance with a first aspect of the present invention there is provided a method of forming a metal electrode for a solar cell, the method comprising the steps of providing a shadow mask adjacent a dielectric layer at a front or rear surface of the solar cell; and directing a metal ion beam through the shadow mask onto the dielectric layer, wherein a first energy of the metal ions in the metal ion beam is chosen to cause both ablation of particles of the dielectric layer and implantation of the metal ions into the dielectric film, whereby at least some of the metal ions are provided on said front or rear surface of the solar cell to form an electrical contact to the solar cell.

The method may further comprise adjusting to a second energy of the metal ions in the metal ion beam lower than the first energy, to thicken the metal electrode.

The method may further comprise performing an annealing process to improve the electrical contact resistance of the formed metal electrode.

The method may further comprise thickening the metal electrode using a plating process.

The shadow mask may comprise a plurality of through-lines and/or through-holes corresponding to electrode lines and/or points respectively of the metal electrode to be formed.

The method may further comprise forming one or more busbars of the metal electrode using screen printing. The shadow mask may further comprises one or more through-bars corresponding to busbars of the metal electrode to be formed.

The method may comprise accelerating the metal ions using an electric field applied between a metal ion source and the shadow mask.

The electric field may comprise a DC or a pulsed electric field.

The solar cell may be silicon based. The dielectric film may comprise silicon nitride, SiN The metal ions may comprise Ni, W or Co. In accordance with a second aspect of the present invention there is provided a solar cell comprising a metal electrode formed using the method defined in the first aspect.

In accordance with a third aspect of the present invention there is provided a solar cell comprising a metal electrode, wherein the metal electrode comprises metal ions implanted or deposited in a dielectric layer of the solar cell.

In accordance with a fourth aspect of the present invention there is provided a system for forming a metal electrode for a solar cell, the system comprising means for providing a shadow mask adjacent a dielectric layer at a front or rear surface of the solar cell; and means for directing a metal ion beam through the shadow mask onto the dielectric layer, wherein a first energy of the metal ions in the metal ion beam is chosen to cause both ablation of particles of the dielectric layer and implantation of the metal ions into the dielectric film, the system being configured such that at least some of the metal ions are provided on said front or rear surface of the solar cell to form an electrical contact to the solar cell. BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the invention will be better understood and readily apparent to one of ordinary skill in the art from the following written description, by way of example only, and in conjunction with the drawings, in which:

Figure 1 is a schematic of an n ⁺ pp ⁺ silicon wafer solar cell with dielectrically passivated front and rear surface. The front surface is contacted with a grid-like electrode featuring parallel metal fingers and two or three busbars ("H-pattern"). Contact of the blanket rear electrode to the rear surface of the silicon wafer is realised via point or line openings through the rear dielectric.

Figure 2 is the same structure as in Figure 1 , but with localised heavily doped regions added underneath the front and rear metal contacts.

Figure 3 is the same structure as in Figure 2, but without the heavily doped layer (BSF layer) along the rear surface and without the localised heavily doped regions at the front and rear surface. Figure 4 shows a cross-sectional schematic of a thin metal seed layer within an opening in the dielectric layer of a silicon wafer solar cell. The metal seed layer has then been thickened by a plating method.

Figure 5 is a schematic representation of the vacuum system embodying the present invention, containing a metal ion source, a shadow mask, and a non-metallised silicon wafer solar cell.

Figure 6 is a schematic representation of the two processes (sputter removal of ionised atoms/molecules from the dielectric layer, implantation of metal ions) that occur when high-energy metal ions (ion acceleration voltage > 2000 Volts) impinge on a dielectric layer on a silicon wafer, according to an embodiment of the present invention. Figure 7 is a schematic representation of the shadow mask based thin metal layer ("seed layer") deposition process using high-energy metal ions (ion acceleration voltage < 1000 V), according to an embodiment of the present invention. Figure 8 is a schematic representation of a silicon wafer solar cell that has a rear metal electrode formed according to an embodiment of the present invention. The metal/silicon contact areas are either point contacts or line-shaped contacts.

Figure 9 shows a flowchart illustrating a method of forming a metal electrode for a solar cell, according to an example embodiment.

Figures 10a) to 10d) show schematic cross-sectional drawings illustrating formation of a shadow mask according to an example embodiment.

DETAILED DESCRIPTION

The embodiments described provide a method for realising a high-efficiency silicon wafer solar cell in a cost-effective and industrially feasible way. In particular, the method will be described in the context of solar cells made on p-type silicon wafers, but it will be appreciated by a person skilled in the art that, with suitable modifications, the method can also be applied to solar cells made on n-type wafers.

Figure 5 shows a system 500 embodying the present invention, more particular the system 500 comprises a metal ion source 502, a shadow mask 504, a non-metallised silicon wafer solar cell 506 with a dielectric film 508 on its surface, and a high-voltage power supply 510.

The metal ion source 502, shadow mask 504 and the unmetallised solar cell 506 are housed in a vacuum chamber. The metal ion source 502 provides a flux of positively charged metal ions 514. The metal ion source 502 could, for example, be a filtered cathodic vacuum arc (FCVA) system or an ionised metal plasma (IMP) system. The shadow mask 504 is held, in this embodiment, in close proximity of or in direct physical contact with the solar cell 506. The shadow mask 504 has line-shaped or circular openings, whereby these openings cover less than 10% of the surface area of the shadow mask 504.

Both the shadow mask 504 and the solar cell 506 are biased to a high negative potential relative to the metal ion source 502, using a high-voltage supply 510. As a result, the positively charged metal ions 514 are accelerated towards the shadow mask 504 (and solar cell 506). The electrical potential for the shadow mask 504 and solar cell 506 can be either steady state or pulsed.

In example embodiment, the initial potential is preferably high enough to sputter off the dielectric film 508. Typical voltage values can be in the range of about 2 kV to 10 kV. The electrical potential can be applied with time periods ranging from full DC operation to pulsed operation with duty cycles of about 10-80% and a frequency of 1- 20 kHz in different embodiments. Since the ion bombardment also heats up the shadow mask 504 depending on the ion energy, e.g. the ionic flux can be up to a few - Amperes of current, this heating parameter is preferably also considered in determining the applied electrical potential in example embodiments. For the second phase of deposition in the example embodiment, the energy needed can be significantly lower, and may be in the order of about 10-500 V.

Metal ions 514a that arrive outside of the openings of the shadow mask 504 are blocked by the shadow mask 504 and do not participate in the formation process of the solar cell's 506 metal electrode.

Metal ions 514b that arrive at an open region of the shadow mask 504 pass the shadow mask 504 unhindered and impinge on the dielectric film 508 on the solar cell's 506 surface. For a high ion energy (e.g. V > 2000 V), these ions 514b ablate (i.e., "sputter off') atoms and molecules from the dielectric film 508, thereby gradually exposing the underlying silicon wafer. The metal ions 514b themselves are implanted into the dielectric film 508 or the underlying silicon wafer 516, as shown in Figure 6. Once the silicon wafer 516 is sufficiently exposed, the ion acceleration voltage is reduced to low levels (e.g. < 1000 Volts). The ion energy is now too low to cause ablation of atoms/molecules at the impinging interface; instead, the arriving metal ions 514b are now deposited onto the impinging interface, leading to the formation of a thin metal contact. Figure 7 shows a cross section of the sample at this stage of the process.

For a typical solar cell, the thickness of the dielectric film, for example SiN, is quite fixed and standard at around 80 nm. Relatively, the thickness of the emitter (compare n ⁺ layer in Figure 1 ) is in the order of 500-1000 nm. Therefore, since there is typically a vast difference in the thicknesses of those two layers, a slight oversputtering can be tolerated. In one example embodiment, the initial sputter process will be controlled on a time basis. It is noted that in the event that the dielectric firm 508 is undersputtered, the residual thickness is typically quite thin and also very heavily implanted with metal ions. This would still result in a conductive film, which should be suitable for formation of the electrodes. As such, the sputter process margin can preferably be very wide.

It is noted that in another example embodiment, a two-step process may not be necessary. For example, in one embodiment, the mass and energy of the metal ions are high enough such that the interaction of the metal ions and the surface will result in the metal ions being implanted into the surface (depending on energy, for example about 10-100 nm deep), and the atoms at the surface of the dielectric film are ejected off resulting in the dielectric film becoming thinner at the same time. This advantageously eventually leads to the implanted metal ions coming into contact with the underlying emitter layer surface as the dielectric film thins, which can enable formation of the metal seed layer in a single process step.

An optional annealing step at elevated temperature, for example for about 30 minutes at about 250°C, may follow to improve the electrical contact resistance of the deposited thin metal contacts 700. Examples of temperature ranges in different embodiments are about 200-800°C for periods of about 10 sec to 30 minutes. For example, for an aluminium back surface field solar cell, if the electrode formation is introduced after the firing process, the annealing may be at low temperature for long anneal times (about 150°C, about 30 mins). If the electrode formation is introduced before the firing process in another example, the annealing can be performed during the firing (about 10 sec at about 800°C). Depending on the metal chosen, especially for the front metal grid, the electrode formation can be quite immune to long thermal processes. Examples of such metals include metals that are used to form silicides in the semiconductor industry (nickel (Ni), tungsten (W), cobalt (Co)). They typically have low solubility in silicon and do not spike through the p-n junction.

The thin metal contacts 700 serve as the "seed layer" for the following metal plating process which thickens the metal contacts 700 to the desired height. Either electrically driven plating or electroless plating can be used. The plated metal is typically copper or silver, or a stack of these metals. Alternatively, a screen printer can be used for thickening of the thin metal contacts. Thus, the metal electrode in the finished solar cell according to an example embodiment comprises a seed layer of metal ions implanted or deposited in a dielectric layer of the solar cell.

In the example embodiment, an electrode line of about 20 pm width may be formed as the seed layer, followed by plating to 20 pm height. The plating typically will result in the width increasing by an additional about 10-20 pm, making the electrode line about 30-40 pm wide in one example. This can advantageously provide an improved aspect ratio for the electrode lines, compared to e.g. screen printed electrode lines having a width of about 100 pm and a height of about 20 μιτι.

The next step in this example embodiment is the deposition of the metal busbars onto the cell's front surface. These busbars are used for the interconnection of individual silicon wafer solar cells into strings of solar cells, as is understood for solar modules. The busbars typically run perpendicularly across the fingers of the front electrode, forming an H-pattern. The busbars are preferably deposited by screen printing. Alternatively, a second high-energy metal ion process as described above can be used, followed by thickening using e.g. screen printing or plating.

It is noted that the busbars may be formed simultaneously with the electrode lines in another example embodiment, using a suitable mask. However, such a mask may be more complex to design, since the busbar and gridlines are perpendicular, leaving some portions of the mask unsupported. With reference to Figure 8, the method described in the above example for a front surface of solar cells can also be applied to the rear electrode 800 of such solar cells, thereby forming an array of contact structures 802 (lines or point contacts) through the rear dielectric film 804. A blanket deposition of a metal film 806 then follows to interconnect the contact structures 802. The metal film 806 may be immediately deposited with the required thickness (for example via screen printing or sputtering) or may be deposited in two steps (deposition of a thin layer, followed by thickening by, for example, metal plating). The final resulting structure at the rear of the solar cell in one example is shown in Figure 8.

If the type of metal ion is appropriately selected, the high-energy metal ion process can lead to the formation of a localised heavily doped layer at the exposed silicon wafer surface. In the case of p-type wafers, a suitable metal is, for example but not limited to, aluminium, forming a p ⁺⁺ doped region underneath the openings of the shadow mask.

Figure 9 shows a flowchart illustrating a method of forming a metal electrode for a solar cell, according to an example embodiment. At step 902, a shadow mask is provided adjacent a dielectric layer at a front or rear surface of the solar cell. At step 904, a metal ion beam is directed through the shadow mask onto the dielectric layer, wherein a first energy of the metal ions in the metal ion beam is chosen to cause both ablation of particles of the dielectric layer and implantation of the metal ions into the dielectric film, whereby at least some of the metal ions are provided on said front or rear surface of the solar cell to form an electrical contact to the solar cell.

Returning to Figure 5, the apparatus, in one embodiment, used for the fabrication of silicon wafer solar cells consists of the following components: Shadow mask 504:

The shadow mask 504 is fabricated from a silicon wafer. Referring to Figure 10a), each side of a silicon wafer 1000 is coated with a silicon nitride (SiN) film 1002, 1004. A pattern of parallel openings 1006 is then made in one SiN film 1002 as shown in Figure 10b), using e.g. laser ablation or photolithography. The wafer 1000 is then immersed in a solution of potassium hydroxide which etches the silicon exposed in the openings 1006 of the silicon nitride film 1002, as shown in Figure 10c). The etching continues until the entire thickness of the Si wafer 1000 is etched, forming a tapered aperture 1008, as shown in Figure 10d). Instead of wet-chemical etching, the silicon wafer 1000 can also be structured using laser ablation in another example embodiment. Instead of a silicon wafer, a sheet of metal can be used to fabricate the shadow mask in different embodiments.

It will be appreciated by a person skilled in the art that the above is only one example of fabricating a shadow mask suitable for use in embodiments of the present invention. Different fabrication processes/techniques may be used in different embodiments, and the skilled reader is referred to existing literature, for example under the topic "silicon micromachining", for details of other suitable fabrication processes/techniques.

Metal ion source 502:

The metal ion source can be selected from a variety of sources. Typically, a remote source may be used, for preferably decoupling the sample from the high voltage normally present in direct sources, distancing the source from the sample, and reducing the chance of arcing etc. In one embodiment, a Filtered Cathodic Vacuum Arc (FCVA) is used as source. Metal ions are formed in a vacuum arc, where they are extracted and then filtered by a magnetic filter to remove macro particles allowing only single ions to pass through. Preferably, the metal chosen is one with a high molecular weight. In example embodiment, metals that have an atomic weight that is much higher than silicon (28 amu) are chosen, for example nickel (58 amu) or tungsten (183 amu). Under sputter collision, the material of the lighter atomic weight would be ejected. Preferably, the chosen metal forms a good contact with silicon, and is readily platable. An example of a suitable metal fitting such a description would be nickel, however, other metals are also possible, as will be appreciated by a person skilled in the art.

High-voltage power supply 510 and sample stage (not shown):

The specimen stage of the system, in one embodiment, is electrically floating relative to the vacuum chamber 512 and is biased to a negative potential by the high-voltage power supply 510. The specimen stage might also be cooled by water or air to remove the heat generated during deposition. The high-voltage power supply 510 can be a steady-state direct-current (DC) supply with a variable output voltage or a pulsed DC supply, or a combination of both. The negative terminal of the power supply 510 is used to bias the sample stage, whereas the positive terminal of the power supply 510 is connected to the vacuum chamber and metal ion source 502, and is normally grounded.

Solar cell 506:

Embodiments of the present invention are applied to partly processed solar cells. In one embodiment, the only missing part of the solar cell is the front metal grid. A possible process flow for the solar cell 506 can be, but is not limited to:

I. Texturing of a p-type silicon wafer (both sides)

II. Phosphorus diffusion (both sides)

III. Coating of one surface of the solar cell with a surface-passivating antireflection coating (usually silicon nitride, SiN). This surface thereby becomes the front surface of the solar cell.

IV. Printing of the metal contacts onto the rear surface of the wafer. This is usually a busbar electrode consisting of silver (to ensure solderability of the solar cell) and a large-area electrode consisting of aluminium.

V. Annealing of the solar cell in a so-called fast firing furnace.

VI. Deposition of the thin metal grid (seed layer) on the front side with the invention disclosed in this document

VII. Thickening of the thin front metal grid by plating.

In the following, the process steps in the formation of the front metal grid on a partly processed solar cell according to one embodiment will be described, by way of example. However, it will be appreciated that the present invention is not limited to that example.

1. The partially completed solar cell is attached to the shadow mask by use of suitable tape or by clamping them together, with the SiN coated surface facing the shadow mask in a horizontal configuration. The rear metal electrode of the solar cell is already complete. The solar cell and shadow mask combination is now attached to the sample stage. Steps are taken to ensure good electrical conductivity between the shadow mask, solar cell and sample stage. Typically, the shadow mask will be made of a conductive material or coated with a conductive material (such as metal) in example embodiments.

The shadow mask can then be connected to the stage by a conductive element such as a metal tape or a wire (noting that typically the stage is much bigger than the shadow mask in example embodiments). The solar cell is mounted on the stage and clamped down by the mask, therefore becomes electrically conductive through contact of the backside of the solar cell to the stage.

It is noted that in a production machine, the shadow mask may face downwards, with the ions travelling upwards. The solar cell can be placed facing down on the shadow mask. Gravity can ensure that the solar cell and the shadow mask are in close proximity.

The vacuum chamber is pumped down and the bias to the stage is turned on. The ion source is activated, and a flux of metallic ions is generated by the ion source. In the case of FCVA, the ion source is generated by the arcing on the metal target. The arcing results in both single ions and macro-particles being ejected from the target. Macro-particles are filtered out by the magnetic filter, allowing only single ions to go through. The ions are typically singularly charged and are still of very low energy when they leave the ion source. Once inside the vacuum chamber, the ions experience the electric field caused by the bias of the sample stage. Since the sample stage is planar, the electric field experienced by the ions is perpendicular to the stage. The ions are accelerated to the sample stage, forming a parallel ion beam. The energy of the ions is controlled via the bias voltage. For instance, if the voltage of sample and stage are set to 1000 V, the ion energy will be approximately 1000 eV. The bias voltage can be in a DC form or in a pulsed form.

Most of the ions will impinge on the shadow mask, except for those which arrive at the open area of the masks (gridline pattern). These ions will pass the mask and land on the solar cell surface. Since the ions were accelerated before reaching the mask, the ion beams will be parallel to each other and perpendicular to the surface of the mask. Since the beams were accelerated by electric fields perpendicular to the mask, there is minimal divergence of the beam after exiting the mask and arriving on the solar cell.

When the ions land on the solar cell, a few phenomena might happen. Since the energy of the ions is high, particles (e.g. atoms and/or molecules) at the surface, here the dielectric layer, will be sputtered off by the incoming ions. Depending on the molecular weight and the energy of the ions the incoming ions might experience a recoil and bounce back if the molecular weight is low, or be implanted into the substrate if the molecular weight is high (i.e. high momentum). As will be appreciated by a person skilled in the art, the momentum of the ions is determined by the product of mass and velocity. An increase in the voltage would increase the acceleration of the ions by the electric field and hence the final velocity of the ions. The latter case is preferred since the implanted ions will form a conducting channel through the thin dielectric film, connecting the substrate to the surface. The sample is then removed from the vacuum chamber.

The cell then undergoes a second deposition step with a different mask, to form the busbars ("H pattern"). Since the busbars are only for providing a conductive path to the grid lines, they do not need to be formed at high ion energies. Alternatively, they could be formed with sputtering or other methods. The solar cell is then plated in a silver plating solution, using either light- induced plating (LIP) or standard electroplating. The typical plating thickness is about 10-20 rqicrons of silver.

The solar cell is then annealed at about 150°C to improve the contact resistance of the solar cell.

8. The solar cell is now complete.

It will be appreciated by a person skilled in the art that numerous variations and/or modifications may be made to the present 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 to be illustrative and not restrictive.