The present invention relates generally to solar cells and more particularly to methods and apparatus for fabricating solar cells.
Silicon is the most commonly used component for forming a photovoltaically active material, since silicon is abundant, inexpensive, and environmentally responsible. Of the various forms of silicon, hydrogenated amorphous silicon (a-Si:H) film deposited by plasma enhanced chemical vapor deposition (PECVD) is the least expensive used in fabricating solar cells. However, the current state of the art solar cell manufacturing technology employs large inexpensive PECVD machines.
Despite the large capital equipment requirements, an a-Si:H layer that is formed using conventional processing by material properties which limit photovoltaic efficiency to approximately 10%, as compared to the 30% or 40% level which would be achieved with ideal material properties. “Photovoltaic efficiency” is defined as a ratio of the electric power produced by a photovoltaic device to the power of the light incident on the device. Despite the low photovoltaic efficiency, solar cell production technology has reached a price-point threshold that triggers large market response. The goal is to decrease the cost per Watt of power that is generated, as compared to fossil fuel, hydroelectric and nuclear alternatives.
Two factors which are significant in determining the photovoltaic efficiency upper limit are the absorptivity and the carrier lifetime properties of the layer. Solar absorptivity is the fraction of the incoming solar energy that is absorbed by the layer. Since absorbed photons generate the charge carriers (free electrons or holes), increasing the absorptivity of a particular material is likely to increase the generation of charge carriers. The carrier lifetime is the average time a charge carrier exists before recombination.
In a conventional a-Si:H structure, a disordered silicon atom arrangement enables a higher absorptivity than would be possible with crystalline silicon. It is possible that approximately one hundred times more light is absorbed per unit thickness by the a-Si:H structure. However, while absorptivity is an important requirement for low cost solar cells, conventional atomic disorder also result in a high rate of recombination of photo-generated carriers. That is, amorphous silicon exhibits a lower carrier lifetime than does crystalline silicon. In disordered a-Si:H material, a high fraction of the photo-generated electrons and holes recombine before drifting to electrodes, thereby preventing their contribution to photo voltage or photo current. The hydrogenation of the a-Si:H structure plays a role of dramatically reducing the density of recombination traps, but high recombination rates remain as a major leveling factor for achieving a high photovoltaic efficiency. Crystalline silicon does not have the same problem, since it exhibits the higher carrier lifetime, but thicker photovoltaically active layers are required in order to compensation for the lower absorptivity of crystalline. This is significant, since the increase in layer thickness increases the overall expense of a solar cell.
In accordance with the invention, at least one photovoltaically active layer of a solar cell is formed as a “hybrid” of regions in which first regions exhibit high absorptivity and second regions have a longer range order in a direction generally perpendicular to the major surfaces of the layer, thereby exhibiting a longer carrier lifetime than the first regions. The photovoltaically active layer is a film which is homogenous in a lateral direction (i.e., parallel to the major surfaces) with respect to film constituents, but is non-homogenous with respect to photovoltaic properties. The high absorptivity exhibited by the first regions ensures generation of sufficient charge carriers for a given layer thickness, while a medium or long-range order (in either atomic positions or stoichiometry) enables high mobility channels to be formed within and around the first regions.
Conventional approaches to fabricating a silicon-based photovoltaic structure encounter a tradeoff between absorptivity and carrier lifetime, since amorphous silicon more readily absorbs incident photons to generate charge carriers, but crystalline silicon is superior with respect to the carrier lifetime. Using the “hybrid approach” circumvents this tradeoff. Nano-layered transitions from amorphous to nanocrystalline define the high absorptivity first regions adjacent to the high carrier lifetime second regions. While both regions are compatible with generating the charge carriers and both regions enable carrier mobility, the second regions provide medium or long range order in the vertical direction (normal to the major surfaces), so as to achieve sufficient carrier lifetime to allow a greater percentage of the photo generated electrons and holes to reach electrodes on the major surfaces.
While the “hybrid approach” is described primarily with respect to silicon-based layers, the approach may be used in other applications. A way of example, the photovoltaically active layer may be based upon Ge, GaAs, SiGe or other semiconductor materials. For the silicon-based structure, the first regions are hydrogenated amorphous silicon (a-Si:H), while the second regions are hydrogenated crystalline silicon (c-Si:H). However, the benefits of the invention apply to more complex structures, such as multi-junction (e.g., triple junction) structures in which material constituents are charged through a sequence of layers/films in order to provide multiple band gases. Then, different portions of the solar spectrum are converted at different junctions, thereby increasing overall efficiency.
In accordance with a method of forming the photovoltaically active layer, the “vertically ordered” silicon in the second regions may be formed by providing sufficient energy to a growing chemical vapor deposition (CVD) surface so as to give the silicon atoms mobility to move energetically favored areas, but with less energy than would result in the silicon atoms being removed from their positions within the second regions. The short range order in the first regions is energetically favored when there is an incident high energy pulse of silicon reactive species during fabrication, instantaneously forming amorphous material. After reaching the growing surface, the silicon atoms of the amorphous layer tend to rearrange into energetically preferred locations, with preference being given to locations which have already developed a “template” of amorphous and nanocrystalline locations. The formation of the original “template” may be controllable, but a randomization in the locations of the second regions at the onset of the layer deposition is within the scope of the invention. As will be readily understood by a person skilled in the art, “islands” of nanocrystallization will occur during the CVD processing. The process parameters are then controlled to continue the location of the nanocrystallization as the layer is grown. A high mobility channel through the layer may be continuous throughout the thickness or may be aligned but separated regions exhibiting medium to long range order.
The process that is well suited for forming the photovoltaically layer in accordance with the invention employs the hollow cathode effect and high power DC pulsing. The pulse duration is a particularly important factor. Longer pulses increase the amount of time for the silicon atoms to establish their energetically preferred locations. Nevertheless, the time between two high powered DC pulses is not less than the duration of each pulse. An example of a PIN solar cell junction formed using this method is as follows: a hollow cathode chamber containing the solar cell substrate (e.g., a stainless steel foil) is heated in hydrogen at a 50% duty cycle. A p-type junction may be formed using silane and 2% diaborane (or other p-type dopant), followed by an intrinsic silicon layer (typically the layer of focus with respect to the present invention). Two frequencies can be used for power application to the plasma, with the higher frequency pulsing preferably being at 25 kHz and a 50% duty cycle. The lower frequency “burst” at 25 Hz can be run (but may be deleted in some embodiments) at a 10% duty cycle for a total duty cycle of 5%. High current per pulse is used for the higher frequency pulsing, such as a current in the range of 25 mA/2.54 cm2 to 100 mA/2.54 cm2, (e.g., a 25 kHz pulse at 60 mA/2.54 cm2). Following this, the n-type layer may be formed using silane and 2% phosphine (or other n-type dopant) in hydrogen. Preferably, the pulse duration is within the range of 5% to 50% of the cycle. As one possibility, 1,000 volt pulses (20 W/cm2) have a 10 microsecond “on” time and 90 microsecond “off” time, yielding a morphology which has the correct regime, according to TEM evaluation (transmission electron microscopy).
The “hollow cathode effect” as used herein. In case a large increase in current as compared to convention plasma glow. The increase is due to the “oscillation motion” of fast (hot, accelerated) electrons between opposite space charge sheaths, which enhances the excitation and oscillation rates in the plasma several times higher than the conventional glow discharge. Because this electron pendulum motion is related to the mean free path of the fast electrons, there is a relationship of the hollow cathode effect to pressure within the hollow cathode and the spacing between the two or more electrodes. That is, a hollow cathode with a small spacing will operate at a higher pressure than a hollow cathode with a larger spacing.
While the fabrication of one or more photovoltaically active layer in accordance with the invention may take place within a tube, the end product need not be tubular. For applications in which the layer is formed within a cylindrical workpiece, the workpiece may be cut into sections which then are used to generate solar energy. Arcs of 120 degrees to 180 degrees substantially increase the collected solar power when exposed diffused light, such as in cloudy or hazy conditions.
In another embodiment, the photovoltaically active layer is formed on a substrate that is progressed through an area in which the proper deposition conditions are established. For example, a flexible substrate may be progressed through one or more tubes in which a hollow cathode effect is established. The substrate may temporarily or permanently cover the wall of the tube. Alternatively, the substrate may cover only a portion of the tube wall, such as a spiraling substrate that is progressed through the tube.
In
In a conventional a-Si:H structure, a disordered silicon atom arrangement enables a higher absorptivity than would be possible with crystalline silicon. However, as previously noted, conventional atomic disorder also results in a high rate of recombination of photo generated carriers. The hydrogen is used for dramatically reducing the density of recombination traps, but high recombination rates remain a major limiting factor for achieving high output energy efficiency in silicon-based solar cells. Crystalline silicon does not have the same problem, but thicker photovoltaically active layers are required because of the lower absorptivity of crystalline silicon. Increases in thickness significantly increase the overall expense.
In the acquisition of the TEM of
Variable high power pulsing enables deposition of nanolayered transitions from amorphous to nanosilicon, with tunable short range order. This is illustrated in
The ability to fabricate nanolayer transitions changing the degree of morphological order (either atomic positions or stoichiometry) is an important process capability in optimizing the a-Si:H photovoltaic properties. As background, researchers at the Lawrence Berkeley National Laboratory have identified the use of small-size material domains (nanometer range) for absorbing more photon energy than would be possible for homogenous semiconductor material. Semiconductor nanocrystal-based cellular imaging is described in U.S. Patent Publication No. 2003/0113709 to Alivisatos et al. Within the patent document, a “quantum dot” is defined as a semiconductor nanocrystal, which is a protein-sized crystal of organic semiconductor nanocrystals, initially developed for optically electronic applications. The concept is to create a special type of structure that enables incoming photons to release more than one set of electron-hole pairs. This special structure is a small bounded domain with nanometer dimensions. The primary interaction is between a photon and a small domain. When the photon has sufficient energy for more than one electron-hole pair, a second electron-hole pair is more likely to be formed if the photon energy is confined to the quantum dot rather than being dissipated in a surrounding material. The process provides primary homogenous a-Si:H produced with high power transient DC pulse verse, a 5-50 W/cm2 DC with a peak voltage of 1,000 volts used to trigger the discharge.
In comparison to the processing for providing the “quantum dots”, the present invention relates to photovoltaically active layers arising from unique energy distribution of plasma species present in a hollow cathode discharge with high power pulsed DC operation. A stacked nanolayer 3-dimensional quantum well array may be defined, with grain boundaries (precipitates and defects) as lateral quantum well boundaries and with nanolayer edges as “top” and “bottom” quantum well boundaries. High efficiency is enabled by multiple electrons being associated with a high energy photons.
While
In
In
In the embodiment of
From the foregoing it can be seen that the present invention departs from conventional silicon-based photovoltaically active layers, since the hybrid structure is defined. The first areas exhibit high absorptivity, while the second areas exhibit high carrier mobility in the vertical dimension. In the method of forming the photovoltaically active layer, the vertically ordered silicon in the second regions may be formed by providing sufficient energy to a growing chemical vapor deposition surface so as to give the silicon atoms sufficient mobility to move to energetically favored areas. However, the applied energy should be less than that that would result in the silicon atoms being removed from their positions within the second regions. The short range order in the first regions is energetically favored when there is an incident high energy pulse of silicon reactive species during fabrication, instantaneously forming amorphous material. During initial growth, “islands” of nanocrystallization will occur. These “islands” create a “template” of amorphous and nanocrystalline locations. The process parameters are then such that this template is maintained throughout the growth process.
In forming a solar cell of which the photovoltaically active layer is a component, a hollow cathode chamber is used to establish the hollow cathode effect. In some applications of the invention, the solar cell layers are formed directly on the chamber walls. For example, a cylindrical stainless steel substrate may be used as the substrate, whereafter the substrate may be cut into 90 degree or 180 degree sections for end use. Alternatively, a stainless steel foil may be passed through one or more chambers, as will be described with reference to
The first process chamber 53 may be dedicated to cleaning the substrate, rather than forming a layer. A stainless steel foil may be cleaned in hydrogen at a duty cycle of approximately 50%. The following process chambers may be used to deposit the layers which form the solar cell. Thus, process chambers 55, 57, and 59 may sequentially deposit an p-type hydrogenated silicon layer, the process chamber 57 may deposit an i-type hydrogenated silicon layer, while the following process chamber 59 may deposit a n-type hydrogenated silicon layer. The three-layer sequencing may be repeated.
The thicknesses of the various layers deposited within the tube processors 53-63 of
By way of example, a p-type layer may be formed using silane and 2% diaborlane. Following formation of the p-type layer, an intrinsic silicon layer is formed.
In order to establish a hollow cathode effect within the different tube processors 53-63, the conductive processors are biased. Two frequencies can be used for power application to the plasma within a processor. The higher frequency is a pulsing that preferably occurs at 25 kHz, with a 50% duty cycle. The lower frequency “burst” occurs at 25 Hz and may have a duty cycle of 10%. Thus, the combined pulsing has a duty cycle of 5%. Each pulse of the higher frequency should have a high current, such a current in the range of 25 mA/2.54 cm2 to 100 mA/2.54 cm2 (for example, a 25 kHz pulse at 60 mA/2.5 cm2).
In the illustrated embodiment, the substrate 81 may be a stainless steel foil. A zinc oxide layer 83 is formed on the substrate. Then, an amorphous silicon 85 is provided.
A significant difference between the structure of claim 13 and the conventional approach of forming a solar cell is that nanocrystalline silicon columns are formed using the techniques which are described above. Thus, high power DC pulsing is applied and the hollow cathode effect is established in order to provide layer texturing as shown in the n silicon layer 91. Additional layers are then provided. For example, the sequence of the three layers 87, 89, and 90 may be repeated as described with reference to
In
This application claims priority from U.S. provisional application Ser. No. 60/993,567, filed Sep. 12, 2007.
Number | Date | Country | |
---|---|---|---|
60993567 | Sep 2007 | US |