Before discussing multi-junction solar cells, let’s review the basic reasons for the energy conversion limitations of single-junction solar cells. Consider an ideal solar cell made of a material with a semiconductor bandgap. When the photon energy is incident inside the cell, the photon will be absorbed and converted into electrical energy, but the electrical energy converted at this time has only excess energy, which will become heat dissipation. And when h v < E, the photons will not be absorbed into electricity. And only in the case of hν>E, the efficiency of converting light energy into electrical energy can reach the highest. It should be noted that even in the case of hv = E, the maximum energy conversion efficiency is still below 100%.
The range of the solar spectrum is very wide, ranging from 0 to 4 eV, so the conversion efficiency of single-junction solar cells to sunlight is naturally much lower than that of single-frequency light. A simple idea to solve this problem is to divide the solar spectrum into several bands according to its energy, and each band uses a solar cell with a suitable energy bandgap to convert light energy into electrical energy, so as to improve the efficiency of converting light energy into electrical energy. E.g. Divide the solar spectrum into blue bands hv1 , hv2 , hv3 , hv4 , hv4 , hv5 , where hv1 <hv2 <hv3 and the photons in these spectral ranges will use three different band gaps of the solar energy Eg1 = hv1 , Eg2 =hv2, Eg3 =hv3 to convert light energy. If the number of spectral bands can be divided into more, higher energy conversion efficiency can theoretically be achieved.
The scholar Henry has calculated under the conditions of A Ml. 5 and 1 sun. Energy conversion efficiency limit of multijunction solar cells. When the number of junctions of the solar cell is 1, 2, 3 and 36, the highest energy conversion efficiencies are 37%, 50%, 56% and 72%, respectively. When the energy conversion efficiency is changed from a single-junction solar cell to a 2-junction solar cell, the increase is most obvious, and the improvement of the energy conversion efficiency of more junction pairs becomes limited. This calculation is good news in some ways, because it is very difficult to actually make a solar cell with 4 or 5 junctions, and if a multi-junction solar cell is to really improve the energy conversion efficiency, each It is very important whether the energy band gap of a sub cell material is selected correctly.
Using high-efficiency single-crystal semiconductor solar cells on the surface to generate electricity, whether it is a single-crystal Si solar cell or a GaAs solar cell, the cost of the battery is too expensive. An alternative is to use a concentrating solar cell power generation system, using cheap concentrating modules to reduce the usage of single crystal semiconductor solar cells to one percent, or even one thousandth.
Table 1[Z] shows the structural parameters of concentrating GaAs solar cells, including P on-N type and N-on-P type. Compared with non-concentrating GaAs solar cells, the structure of concentrating GaAs solar cells has an extra layer under the base layer to reflect secondary carriers back to improve cell efficiency. The concentrating design can not only reduce the amount of solar cells, but also improve the conversion efficiency of solar cells, which is mainly due to the increase of the open circuit voltage Voe and the increase of the fill factor. At present, the highest conversion efficiency of GaAs solar cells is 27.6% (AM 1.5, 255 suns) under concentrated light conditions.
In concentrated solar cell power generation system. Since a single solar cell produces currents of several amperes or more, series resistance has a very large effect on efficiency. Too high series resistance will reduce the fill factor and severely reduce the cell efficiency. The main sources of series resistance include the resistance Re of the grid electrode line, the contact resistance Rs between the metal and the semiconductor (Fig. 1), and the lateral sheet resistance of the current passing through the emitter layer. There are several ways to reduce the series resistance of the battery itself. click here to open to learn more.
(1) Increase the density of grid electrodes. Increasing the density of the grid electrode can simultaneously reduce the grid electrode line resistance, contact resistance and lateral sheet resistance of the entire solar cell, but increasing the grid electrode density often increases the shading area, thus reducing the efficiency of the cell.
(2) Increase the conductivity of the emitter layer. Increasing the conductivity of the emitter layer can reduce the lateral sheet resistance R day. Since N-type GaAs has better conductivity than P-type GaAs, the N on P structure is a better choice.
(3) Reduce contact resistance For concentrating GaAs solar cells, the contact resistance Re must usually be small. A thin layer of alloy, such as Au/Zn/Au or Au/Ge/Ni/Au, is added between the gate electrode and the semiconductor, and with suitable thermal fusion conditions, the contact resistance can be effectively reduced to the required level scope.
Figure 2 shows the electrode design of a concentrator solar cell. The radial grid electrodes guide the current from the central area of the solar cell to the periphery, although the grid electrode density is actually much higher than shown in the figure. In order to reduce the series resistance and reduce the shading area, the shading ratio of the grid electrode will be controlled at 4%~8% as much as possible.
Figure 3 shows the efficiency comparison of Pon-N type and N-on-P type concentrating GaAs solar cells under different concentrating conditions. The dense grid electrode design is used to reduce the shading area and achieve higher efficiency. However, when the concentrated light intensity exceeds 400 times, the efficiency of N-on-P type GaAs solar cells designed with low m-density grid electrodes begins to saturate, while the efficiency of N-on-P type GaAs solar cells designed with high-density grid electrodes begins to saturate. The efficiency of on-N type GaAs solar cells can continue to increase as the concentration of light increases by a factor of 1000.
This result shows that the design of the grid-like concentrating GaAs solar electric electrode is the key point of the concentrating solar cell, according to the design of the grid-like electrode of the cell. The actual grid concentration magnification optimizes the design of the grid electrode, effectively reducing the density of the grid electrode, which is much higher than the series resistance shown in the figure, and at the same time reducing the shading area as much as possible to increase the photocurrent, in order to improve the efficiency of the battery.
GaAs is a material with a direct energy band gap, and the absorption effect of light is quite good. When the photon energy is larger than the bandgap of GaAs, the absorption coefficient of the material for light is about 104cm-1. Figure 1 shows the light absorption coefficients of semiconductor materials such as Si, GaAs and InP for different wavelengths, so the thickness of the GaAs solar cell structure is only about 4µm. The high absorption rate of light by the direct energy gap material is an advantage. But other problems arise when making solar cells.
There are many dangling bonds on the surface of semiconductor materials, forming a large number of surface defects. These surface defects can cause electron-hole pairs generated by illumination. Recombination on the surface of the solar cell makes the light energy unable to be converted into electrical energy, reducing the energy conversion efficiency of silver solar energy and earth. This is beneficial, and the effect has a greater impact on direct-energy semiconductor solar cells. Generally, we will use the surface recombination speed (surface recombination), the unit is cm / hour, this parameter to evaluate the influence of surface defects. The surface recombination speed of GaAs material is about 107 cm/s, which is very high. Figure 2 shows the effect of different surface recombination speeds on the optical spectral response of GaAs solar cells. The higher the surface recombination speed, the worse the quantum conversion efficiency of short-lived photons. The surface recombination speed must be reduced to below 104 cm/s, and the GaAs solar cell will have good conversion efficiency.
In order to improve the energy conversion efficiency of GaAs solar cells, the influence of surface defects must be reduced. Generally speaking, there are four methods to reduce the influence of surface defects.
(1) The first is to try to passivation the crystal surface and reduce the defect density on the surface. For example, the generation of SiO2 on the Si surface can greatly reduce the surface recombination rate. However, there is no suitable and stable oxide in GaAs material to achieve this effect.
(2) The second is to make the PN interface as close to the surface of the solar cell as possible. For materials with high light absorption efficiency such as GaAs, the depth of the PN junction must be controlled at 50 nm to reduce the influence of surface defects.
(3) The third is to use the front surface electric field (front surface field). This is the same as the principle of the back surface field of the Si solar cell, that is, a high/low doping structure is used to form a built-in electric field to prevent the secondary carriers generated by illumination from diffusing to the surface of the solar cell, so as to avoid the diffusion of the secondary carriers generated by the illumination to the surface of the solar cell. Reduce the impact of surface defects. For example, for a solar cell that was originally a P-type emitter IN-type base (P-type emitter IN-type base) structure, the P-type emitter can be changed to a high and low doped structure, and the F layer must be very thin, Reduce light absorption.
(4) The fourth is to grow another layer of window layer on the surface (window layer). The window layer is a layer of material with a large energy band gap, which can allow most of the incident light to pass through, and can prevent electrons and holes from diffusing to the surface of the solar cell and recombining under the influence of surface defects. In addition to selecting a material that allows most of the sunlight to pass through the window layer, the interface defect density between the window layer and the solar cell emitter must be very low, which means that the lattice constant of the window layer material must be very different from the emitter layer material. In order to reduce the defects caused by lattice mismatch.
GaAs solar cells can be said to be the most researched and best developed type of single-junction field V semiconductor solar cells. In the production of single-junction solar cells, the advantages of GaAs in terms of material properties are as follows
(1) The energy band gap of GaAs material is closest to the theoretical optimum value of single-junction solar cells.
(2) The radiation resistance of GaAs solar cells is better than that of Si solar cells.
(3) GaAs solar cells have a good window layer material: the lattice constant of GaAs is almost completely matched with that of GaAs, which makes it difficult to generate stacking fault defects (dislocation) when growing on GaAs, regardless of the composition ratio of Al. The high aluminum content is an indirect energy gap material, which can almost penetrate the window layer as long as the photon energy is lower than 3 eV.
(4) The conversion efficiency of GaAs solar cells is less sensitive to temperature. The energy band gap of Si semiconductor is relatively low, so when the temperature increases, the carrier density in the material changes greatly. In addition, and it is an indirect energy gap semiconductor, its carrier life time is greatly affected by phonon (phonon), that is, lattice vibration (lat t ice vibration). The carrier lifetime decreases rapidly. Therefore, the energy conversion efficiency of Si solar cells decreases rapidly as the temperature increases.
(5) The GaAs substrate is cheaper than InP, and at the same time, there are cheaper alternative substrates. The lattice constant and thermal expansion coefficient of GeoGe are almost the same as those of GaAs, which makes the Ge substrate a very good alternative substrate for the manufacture of GaAs solar cells. Moreover, the mechanical strength of Ge material is twice that of GaAs, so the thickness of GaAs solar cells made of Ge substrate can be reduced to about U = 90µm, which greatly reduces the weight of GaAs solar cells. The efficiency of solar cells made of Ge substrates is almost the same as that of solar cells made of GaAs substrates. The pool is just as good.
GaAs solar cells have been developed since the 1960s. In 1972, IBM used the LPE method to produce a heteroface GaAs solar cell. The structure of this solar cell is a P-type GaAs emitter (emitter), an N-type GaAs base (base), and a P-type AlGaAs window layer on top of the P-type GaAs. Since the AlGaAs window layer blocks the diffusion of electron-hole pairs to the surface of the GaAs solar cell and reduces the influence of surface defects of the material, the efficiency of the GaAs solar cell is greatly improved. The GaAs solar cell structure of this heteroface was also later developed into the basis of GaAs solar cells. While using LPE to develop GaAs solar cells, many people also use metal-organic vapor phase epitaxy to develop GaAs solar cells. Since metal-organic vapor phase epitaxy has excellent control over the thickness and doping concentration of the device structure, it is possible to produce high-efficiency Quite high GaAs solar cells (22%, AMO).
Figure 3 [Z] is a structural diagram of a typical GaAs solar cell, which includes an N-type epitaxial substrate, an N-type GaAs buffer layer, an N-type GaAs base, a P-type GaAs emitter, a P-type window layer, and an anti-reflection film. The epitaxial substrate can be GaAs or Ge. If it is a Ge substrate, two issues must be paid attention to in the epitaxial time.
(l) GaAs is a polar semiconductor material, while Ge is a non-polar semiconductor material. If the epitaxial conditions of the GaAs/Ge interface are not properly controlled during the epitaxial delay, an ant knock base domain will be formed on GaAs, resulting in stacking fault defects, reducing GaAs. Efficiency of solar cells.
(2) In addition, due to the high temperature of the epitaxial time, Ga and As atoms will diffuse into the Ge substrate, which will activate the Ge substrate to generate PN junction. In theory, the activated Ge substrate can form another solar cell and improve the efficiency of the GaAs solar cell, but in fact, due to the mismatch of the photocurrent between the GaAs cell and the Ge cell, the efficiency of the GaAs solar cell is reduced, so GaAs is fabricated on Ge. In the case of solar cells, it is generally tried to keep the Ge from being activated.
Solar cell structure The key to solving the above two problems lies in the N-type GaAs buffer layer. Appropriate buffer epitaxy conditions can avoid the generation of anti phase domains and Ge activation.
N-type GaAs base thickness is generally about 3. 5µm, and the doping degree is 1. 0 × 10[7 2. 0 × 1017 cm 3 o Some studies have shown that the doping material will affect the life of holes, such as using As an N-type doped GaAs material, Se has a higher hole life than Si-doped GaAs, so the conversion efficiency of the solar cell is also higher. The thickness of the P-type GaAs emitter is about 0.5 m. Generally, Zn is used as the P-type dopant, and the concentration is about 2.0 × 1018 cm. The window layer is a very important part. The higher the Al content, the higher the light transmittance of the window layer. The higher it is, it is also easy to be oxidized and deteriorated. Generally speaking, the Al ratio of the window layer is 80% ~ 85%, and then an anti-reflection film is coated as soon as possible in the solar cell manufacturing process to protect the window layer from oxidation. The thickness of the window layer is about 50 nm, and the doping concentration should be as high as possible to reduce the potential array and series resistance of the hetero interface.
The effect of antireflection coatings on solar cell performance is also very important. The optical index of refraction of GaAs is about 3.6, if there is no anti-reflection coating, more than 30% of sunlight will be reflected and cannot enter the solar cell. If a suitable thickness of SiN, single-layer anti-reflection film is used, the light reflectivity can be reduced to about 10%, while the use of MgF2/ZnS double-layer anti-reflection film can further reduce the reflected light to about 3%.
The range of III-V compound semiconductor materials is quite wide, and the properties of various materials, epitaxy technology and device use can become a big topic. This time, I will give a brief introduction to III-V semiconductors, especially the materials and epitaxy technology related to solar cells.
Introduction to III-V semiconductor materials
In simple terms, group III-V compound materials are compounds formed by group III elements and group V elements on the periodic table. The III-V compounds formed by different elements have very different properties, ranging from semi-metals to semiconductors and insulators. In the common field V group semiconductor materials, the H group elements include aluminum CAD, gallium (Ga), indium Cln), etc., while the V group elements include nitrogen (N), phosphorus (P), arsenic (As), bromine (Sb) ) Wait. Binary materials formed by these III-V group elements (referring to a semiconductor formed by a group H element and a group V element, such as GaAs, InP, etc.), with the different elements, the crystal lattice size is related to the semiconductor. The energy band gap is also different. → Generally speaking, the smaller the lattice constant, the larger the energy band gap of the semiconductor. The ternary or quaternary materials formed by mixing various III-V elements can achieve the required lattice size or energy band gap size by adjusting the element composition ratio in the material. III-V compound semiconductors are widely used in the manufacture of various optoelectronic materials, including light-emitting diodes, laser diodes, photodetectors, solar cells, and high-frequency electronic devices.
Figure 1 shows the lattice constant (symbol a, unit A) and energy band gap (symbol Eg, unit eV) of different III-V compound semiconductors. In the figure, binary materials, such as GaAs, lnP, GaSb, etc., are represented by open circle symbols, while lattice constants and GaAs are represented by solid circles. The line between two binary materials shows the change of lattice constant and energy band gap when the composition is changed; such as the line between GaAs and InAs, it indicates that the ternary materials of Ga, ln1 and As are in different Ga and In Lattice constant versus bandgap at scale.
The lattice constant of ternary materials has a linear relationship with the material composition, which is called Vega’s law. This simple law allows us to analyze the composition of ternary materials using high-resolution X-ray diffraction methods. It is generally believed that quaternary materials also obey Vegard’s law.
III-V semiconductor materials related to solar cells
When referring to the semiconductor materials related to solar cells, the most important ones should be GaAs series materials, and the so-called GaAs series actually refers to materials whose lattice constants are consistent with GaAs, including Al, Ga1, As materials (X= O~1) and Ge etc. In addition, in the development of III-V semiconductor solar cells, solar cells such as InP solar cells and GaSb have also appeared.
GaAs series (1) GaAs:GaAs is a material with a direct energy band gap. At room temperature, the lattice constant of GaAs is 5.6532, and the energy band gap is 1.424 eV. The energy band gap is theoretically suitable for making single-junction solar energy. Battery. The application of GaAs wafers is also quite common and is often used to manufacture various optoelectronic devices.
(2) AIAs is an indirect energy band gap material with an energy band gap of 2.14 eV and a lattice constant of 5.660, which is almost identical to GaAs. Since the lattice constant of AlAs is very consistent with that of GaAs, the density of defects such as stacking faults and dislocations grown on GaAs is very low. Table 1 lists the compositional change versus energy bandgap of many ternary III-V semiconductor materials at 300 K. The transition from direct bandgap to indirect bandgap occurs when the Al ratio is about 45%. In single-junction GaAs solar cells, it is mainly used as a window layer to reduce the influence of GaAs surface defects, and the Al content is 80%~90%. In addition, in the application of double junction solar cells, it can also be used to make upper stratum solar cells.
Introduction of epitaxy methods for III-v semiconductor materials
The device structure of a solar cell is mainly a PN junction. At the beginning, the PN junction fabrication method of III-V semiconductor solar cells with industrial materials such as GaAs, InP, etc., mostly adopted the impurity diffusion method. The main mode of structure of compound semiconductor solar cells. The epitaxy methods of III-v semiconductors can be mainly divided into blue types: liquid phase epitaxy (LPE), molecular beam epitaxy (MEE), and organic metal vapor phase epitaxy (OMVPE). At present, high-efficiency III-V solar cells, whether single-junction GaAs solar cells or stacked multi-junction solar cells, all use metal-organic vapor phase epitaxy as the epitaxy method for fabricating solar cell device structures.
3.1 Liquid phase epitaxy
Liquid-phase epitaxy (LPE) was often used in early research on III-V or II-VI semiconductors. The liquid phase epitaxy method melts the elements as components of the epitaxial layer into a liquid state at high temperature, controls the temperature to keep the liquid in a saturated or supersaturated state, and then immerses the epitaxial substrate in the solution or tries to make the surface of the substrate contact the liquid to make the supersaturated liquid Ficus liquid grows semiconductor crystals on the substrate. Figure 2 is a schematic diagram of three methods commonly used in liquid phase epitaxy, where (a) is the dipping method, (b) is the tipping method, and (c) is the sliding boat method.
The advantages of phase epitaxy are its simplicity and the ability to grow high-quality materials with low impurities and low point defect concentrations. Another advantage is that liquid phase epitaxy has the function of removing oxygen when growing materials containing aluminum components (such as AlGaAs), because aluminum will combine with oxygen to form stable, floating on the surface of the liquid. This makes the quality of AlGaAs grown by liquid phase epitaxy far superior to other epitaxy techniques in the early days.
In 1972, IBM’s Hovel and other researchers used liquid phase epitaxy to grow PN GaAs solar cells. The surface of this cell used a P-type doped AlGaAs window layer to reduce the influence of GaAs surface defects. In 1979, researchers such as Yoshida further achieved a conversion efficiency of 19% under AMO conditions.
However, the liquid phase epitaxy method also has many disadvantages: the liquid phase epitaxy method cannot grow too many complex epitaxial structures, the interface between the epitaxial layers cannot be very steep, and the thickness of the epitaxial layer cannot be controlled very thinly and accurately, so many compound semiconductors The epitaxial growth of the device structure is currently carried out by molecular beam epitaxy and organic metal vapor phase epitaxy.
3.2 Molecular beam epitaxy
Basically, molecular beam epitaxy (MBE) is a method of vacuum evaporation of solid-state thin films. The molecular beam epitaxy method heats the epitaxial substrate in an ultra-high vacuum environment, and then shoots the molecular beam or atomic beam of the element onto the surface of the epitaxial substrate to grow the epitaxial layer. Since the epitaxy process is a physical reaction without complex chemical reactions, the growth kinetics of the epitaxy becomes relatively simple and easy to study, and the composition control of the epitaxial film is relatively simple. Molecular beam epitaxy uses elements as raw materials for epitaxy. The elemental material is placed in a special “collapse” called effusion cells, and the emission amount of the molecular beam is adjusted by controlling the temperature of the effusion cells, and the baffle can control whether the molecular beam is emitted to the surface of the epitaxial substrate n Figure 3 is old A diagram of the vacuum chamber setup of a molecular beam epitaxy system.
Since the reaction chamber of the molecular beam epitaxy system is an ultra-high vacuum system, it is convenient to install various analysis equipment on the reaction chamber for various real-time analysis, such as Auger Electron Spectroscopy (AES) and reflective high-energy electric hand. Diffraction Equipment (RHEED) etc. AES is mainly used to monitor the surface state of the epitaxial substrate and confirm the cleanliness of the substrate surface. RHEED provides the reconstructed state, microstructure and surface flatness of the epitaxy process surface, which is helpful for the analysis of the growth mechanism, and can also provide information such as epitaxy rate and thickness.
Molecular beam epitaxy can provide atomic-level thickness control and a very steep epitaxial interface, which is very suitable for making special epitaxial structures such as quantum wells or superlattice. However, molecular beam epitaxy also has many disadvantages, including not being suitable for growing mixed-side and phosphorus (As/concave) compound semiconductors, not easy to grow antimonide semiconductors, and not easy to grow nitride (Nitride) semiconductors, as well as low productivity, Equipment is expensive etc.
Gas source molecular beam epitaxy (gas source MBE) and metal-organic molecular beam epitaxy (metal-organic MBE) are improved versions of molecular beam epitaxy. They manage to use gaseous materials such as P-phosphate, NH, or N2, etc., as well as organometallic materials in molecular beam epitaxy systems. These improvements expand the variety of compound semiconductors that can be grown by molecular beam epitaxy.
Based on the consideration of mass production, in the past, except for a few academic research papers, almost no field V solar cells were epitaxy using molecular beam epitaxy. However, in the future, some new research findings may change this situation. For example, the quantum dot structure solar cells that are currently being studied by the academic community, or the solar cells of ln N materials, etc. At present, molecular beam epitaxy is more stable than organometallic vapor phase epitaxy for the control of lnAs and InGaAs quantum dots.
3.3. Organometallic vapor phase epitaxy
Organometallic vapor phase epitaxy (MVPE) is also known as organometallic chemical vapor deposition method. The English abbreviation of this epitaxy method may appear in various writing methods such as OMVPE, MOVPE, OMCVD, MOCVD, etc., but they all refer to the same epitaxy method. Organometallic vapor phase epitaxy is currently the most commonly used epitaxy method for making III-V compound semiconductor devices, including various light-emitting diodes, light-emitting diodes, photodetectors, solar cells, and microwave devices such as HBT and HEMT.
Organometallic vapor phase epitaxy uses organometallic materials and hydrides as raw materials for epitaxy. Almost most of the III-V group elements and the II-V-VI group elements used for doping have corresponding organometallic raw materials or hydride raw materials, so the range of materials that can be grown by organometallic vapor phase epitaxy is very wide. Whether it is nitride, compound, phosphide, Jin compound, or various ternary or multi-component compound semiconductor materials. All can be epitaxially grown by metal-organic vapor phase epitaxy.
Figure 4 is a system diagram of an organic metal vapor phase epitaxy system. Generally, an organic metal vapor phase epitaxy system can be divided into eight subsystems, and the eight parts will be described below. (1) carrier gas module (carrier gas module). Organometallic vapor phase epitaxy requires high-purity gases to push the reactants into the reaction chamber, and also requires these gases to establish a stable gas flow field in the reaction chamber. The most commonly used carrier gas is hydrogen (H2), which is generally purified using a palladium cell. Nitrogen (N2), ammonia (He) or oxygen (Ar) are sometimes used as the carrier gas, and adsorption purifiers are mainly used to purify the carrier gas.
(2) hydride gas module (hydride gasmodule). Hydride is mainly used as group V raw material and part of N-type doping raw material. The main hydrides include As, PH, and N. These hydrides, especially AsH3 and PH3, are highly toxic and dangerous gases. It needs to be used and handled with caution. Commonly used N-type doping raw materials include SiH, , Si2H6 (providing Si impurity of group N), H2Se (providing Se impurity of group VI), and DETe belonging to organometal (providing Te impurity of group VI), etc. These N-type dopant materials are diluted with hydrogen or nitrogen to reduce the density to the range of 100-1000 ppm for ease of use.
(3) Organometallic raw material module (metal-organic module). Organometallic raw materials mainly provide D group raw materials, as well as some P-type doping raw materials. Commonly used Hui family organometallic raw materials include Zn, etc., while P-type doping raw materials include DMZn, DEZn (providing Zn of H group), Cp2Mg (providing Mg of H group), etc. Most of the organometallic raw materials are liquid at room temperature, and a few are homomorphic. Organometallic raw materials are generally packed in metal cylinders with high cleanliness. When using, the cylinder is immersed in a constant temperature water tank to keep the temperature of the organic metal raw material stable. Thereby fixing the vapor pressure of the organometallic raw material. Introduce high-purity hydrogen or nitrogen as the carrier gas, and take the vapor of the organic metal raw material out of the cylinder and send it into the reaction chamber. When the carrier gas is passed in, bubbles will be generated in the liquid organic metal, so the cylinder of the organic metal raw material is also called bubbler. To stabilize the supply of organometallic raw materials, the temperature of the cylinder must be strictly maintained and the pressure of the cylinder must be controlled. The flow of the organometallic feedstock is then controlled by controlling the flow of the carrier gas.
(4) Dilute module (dilute module). Sometimes very low amounts of reactants are required for epidemiology, and the concentration of the reactants can be reduced further using the dilution module. The dilution module adds the carrier gas to the reactant raw material, and then takes out the required amount and then flows into the reaction chamber. For example, when 0.1sccm of SiH4 is required in the experiment, 10sccm of SiH4 can be mixed with 990sccm of hydrogen. Then the mixed gas is taken 10 sccm into the reaction chamber, and the equivalent flow rate of 0.1 sccm can be achieved.
(5) switching module (switching manifold). In order for the epitaxy system to grow multiple quantum wells or superlattice structures, the reactants must be rapidly switched during the epitaxy process, and during the switching process, the flow rate of the reactants, the flow rate of the carrier gas in the reaction chamber and the A stable balance of pressure in the reaction chamber is required to maintain the steepness of the heteroepitaxial interface. The key point of the design of the switching module is to minimize the dead space on the valve pipeline, so as to avoid the continuous release of a small amount of residual reactant into the reaction chamber after the reactant is closed, and pay attention to make appropriate compensation for changes in flow and pressure during switching.
(6) the design of the reaction chamber (reactor design). The design of the reaction chamber is the core part of the metal-organic vapor phase epitaxy, in order to achieve the uniformity of the epitaxy and improve the utilization efficiency of the reactants. The design of the reaction chamber must take into account the gas flow field, the temperature gradient distribution in the reaction chamber, and the concentration distribution of reactants. Now, due to the improved understanding of the CVD reaction mechanism and the help of numerical simulations using computers, the new organometallic epitaxy reaction chamber has been achieved. Quite a large production volume.
Figure 5 is a typical example of several reaction chamber designs. ① Horizontal reaction chamber. This is one of the commonly used reaction chamber forms in early metal-organic vapor phase epitaxy systems. It is still commonly used in small experiments and research and development. ② Vertical reaction chamber. This is also one of the early commonly used reaction chamber forms, and is still commonly used in small experiments and research and development. ③ Barrel-shaped reaction chamber. The design of this reaction chamber can be said to be a multi-chip reaction chamber design in which the horizontal reaction chamber shown in Figure 5 is rolled up. Compared with the horizontal reaction chamber, the uniformity becomes better because the boundary on both sides disappears, and the susceptor can be managed to rotate. In the past, it was mainly used to grow GaAs, AIGaAs materials.
④ There is a multi-wafer planetary reaction chamber (multi-wafer planetary reaction or) at the current inlet. This epitaxial reaction chamber design is one of the reaction chambers currently used for mass production of III-V semiconductor devices. Through the revolution and rotation of low speed. It can effectively control the epitaxy uniformity of multiple chips.
⑤ High speed rotating-disk reactor. This is also one of the reaction chambers currently mainly used for mass production of III-V semiconductor devices. Using the plug flow mode (plug flow mode) generated by adjusting the rotation speed of the epitaxial base, and adjusting the distribution of reactants at the air inlet, good uniformity control of multi-chip epitaxy is achieved. (7) the exhaust system. The exhaust system consists of a set of air extraction pumps and throttling. to control the pressure in the reaction chamber. At the same time, this part also includes some filtering devices to filter out some products of the epitaxy process as much as possible, so as not to block the pump and the exhaust pipeline at the rear end.
(8) Exhaust gas treatment system (toxic gas scrubbing system). Since the organometallic vapor phase epitaxy system uses toxic gases and produces many toxic compounds at the same time, it is necessary to have a good waste gas treatment system to treat the exhaust gas. At present, waste gas treatment methods are roughly divided into three categories: ① Adsorption type, which uses activated carbon or resin products to adsorb toxic substances; ② Combustion type. Add an appropriate amount of oxygen to burn toxic substances or decompose them at high temperature; ③ wet treatment type. Use chemicals to react toxic substances into other compounds. Many exhaust gas treatment systems may combine two or three of these for best results.
At present, in the epitaxy technology of III-V semiconductors, including III-V semiconductor solar cells, metal organic vapor phase epitaxy has the most technical and mass production advantages. Of course, the disadvantages of expensive equipment and raw materials, and the use of toxic gases in the epitaxy process must also be considered in a detailed evaluation.
To use solar cells in space. Basically, there are three important points to consider. (1) High energy conversion efficiency The solar spectrum (AMO) on the satellite or in space is different from the surface < AM I. 5), and the difference is that the AMO is mainly short-wavelength light. The efficiency of solar cells will be poor for AMO, and can only reach 0.85~0.9 of AM 1.5. In the past, silicon solar cells and GaAs solar cells were mainly used on satellites. Generally, the efficiency of silicon solar cells for satellites is 12.7%~14.8%, and the high-efficiency silicon solar cells can reach 16.6%. The efficiency of single-junction GaAs solar cells is 19%, the efficiency of double-junction III-V solar cells is 22%, and the efficiency of triple-junction → V group solar cells is up to 26.8%.
( 2 ) Good radiation resistance There are radiations of various energies in space, ranging from 50keV to 50 MeV. Space radiation can cause defects inside solar cells and reduce conversion efficiency. Therefore, if the factor of energy conversion efficiency attenuation is not considered into the system design. After the solar cell runs for a period of time. There will be a problem of insufficient power supply. Among single-junction solar cells, InP solar cells have the best radiation resistance, GaAs solar cells are the next, and silicon solar cells are the worst.
(3) Lightweight The cost of launching a satellite or spacecraft is about $10,000 per stand, and in order to reduce the launch cost, the weight of the solar cell must be considered, or the power/weight ratio CW/kg). Silicon solar cells have a good advantage in this regard, and their weight per unit area is 0.13~0.50 kg/m2. The GaAs substrate is too heavy and has weak mechanical strength, so Ge is used as the substrate. 8 ~ 1. Okg/m2 0 For GaAs solar cells with Ge substrate or double junction and blue junction solar cells, the weight per unit area is still 0.8 ~ 1.0kg/m2 0 Due to its high energy conversion efficiency and good radiation resistance, III-V solar cells have gradually replaced silicon solar cells and are used in satellites and spacecraft.
In recent years, GaAs solar cells have replaced silicon solar cells on newly launched satellites. GalnP/GaAs/Ge solar cells are very similar to GaAs solar cells in the assembly and integration of solar cell modules, with higher energy conversion efficiency and as good radiation resistance as GaAs solar cells, while adding the advantages of high voltage and low current, Therefore, it is expected to become the next generation of space solar cells.
Surface power generation
At present, solar power generation systems are widely used on the surface, from the power supply of small consumer products to large power plants. As mentioned in the preface, because the cost of Tian V solar cells is too expensive, Tian V solar cells are rarely used to make flat-panel solar cells except for special purposes such as satellites. According to the calculation, if we want to achieve an acceptable power generation cost, the concentrating conditions must be at least 400 times (400suns) above.
Using GalnP/GaAs/Ge high-efficiency solar cells, under high concentration conditions (such as 1000 suns), the power generation cost has the opportunity to drop to 0.07 US dollars / (kW h). At present, the annual production capacity of solar cells for space is about 0.5 MW. If these production capacities are converted into 1000 × concentrated solar cells, the annual power generation can reach about 0.5 GW. The current highest efficiency record for GalnP/GaAs/Ge solar cells is 34 % (AM 1.5 G, 210 suns), while the conversion efficiency of outdoor modules is 25 % 29 % at low concentration magnification. These data show that III-V solar The potential for batteries to be used for surface power generation is very high. Of course, before the actual large-scale investment in power generation systems, stable and reliable products are still required.
In the future, thin-film solar cell products will be similar to today’s computer CPUs, and the demand for products can also be stimulated by the ever-improving energy conversion efficiency. Judging from a certain area. The roof of the house, the exterior of the building, the surface of the clothes bag, etc. Some possible application and installation places of thin-film solar cells. Needless to say, its power output increases proportionally with efficiency. How to further improve the energy conversion efficiency of the product is a major challenge, and it is also the key to the decisive victory of the product. Although reducing the manufacturing cost of the product is another factor, compared with the two, the former also has the effect of reducing costs. should be given more attention.
The more traditional method to improve the efficiency of CIGS solar cells is to use a tandem structure, that is, PN junctions made of different materials are stacked in order from top to bottom according to the size of the energy gap. According to theoretical calculations, the series connection of three cells can achieve the most economical Affordable results, see Figure 1. The light-transmitting conductive layers can be connected in series between the cells, each responsible for the absorption of sunlight in different bands, which can improve the efficiency to more than 30%. Taking the Hybrid V compound semiconductor as an example, the current maximum has reached 40%. The IHB VI family also has different material combinations with adjustable energy gaps for device design. The related material data is shown in the series structure of the compound, as shown in Figure 2.
Another research direction to improve the efficiency of CIGS solar cells is to deposit an ultra-thin absorber layer (ETA cell for short) in nanostructures, as shown in Figure 3. Its operation principle is similar to that of dye-sensitized TiO2 solar cells. similar. If quantum dots can be deposited and formed, using their light absorption characteristics different from those of bulk materials, when the size of the material is reduced to the point where the particle size is lower than the mean free path of carriers. It can reduce the carrier recombination generated by light, and can control the energy gap value of nanocrystals by adjusting the size, improve the utilization rate of sunlight energy, and also allow photons with more energy than the energy gap to generate more than one pair of electrons. and holes to improve battery efficiency. This new type of design is called a third-generation solar cell. Since it is an emerging technology, there are still problems to be overcome. If the efficiency can be effectively improved to more than 10% in the future, it may become one of the market mainstreams of thin-film solar cells. To read more about batteries click here to open.
Figure 3 ETA cell compares with crystalline silicon solar cells up to 200µm thick, and thin film solar cells can build their product features on flexible substrates. Figure 4 is a comparison of the measured efficiency values of flexible CIGS solar cells fabricated on different substrates. High-efficiency CIGS solar cells are usually prepared at temperatures above 500 °C,
It is feasible to use metals such as stainless steel as the substrate. However, if the mass is to be greatly reduced in order to facilitate the use of personal power sources or to facilitate the application of power sources in outer space, lightweight polymer substrates are often used. From Figure 4(b), it can be seen that the efficiency value of CIGS solar cells fabricated at different substrate temperatures starts to drop when the temperature is above 450 °C , because this type of substrate has cracking phenomenon at this temperature, so a low-temperature process is required. Preparation of CIGS films. The German company Solarion has developed a roll-to-roll process to manufacture CIGS solar cells on flexible polymer substrates. As shown in Fig. 5, the device structure is fabricated by Polyimide/Mo/CIGS/CdS/ZnO stack, in which the CIGS film is fabricated by ion beam assisted evaporation, but the cell efficiency is only about 8%. Defects will affect the film properties. resulting in low power generation efficiency. In our laboratory, high-quality CIS epitaxial films can be grown at 300°C by UV-assisted evaporation. This method should be a better choice for the research and development of low-temperature processes.
There is a worrying question in the mass production of CIGS solar cells, that is, whether In will run out of material when there is a lot of demand. Although some arguments suggest that In mineral stocks are sufficient, they are not necessarily credible. When planning ahead, there is indeed a research on the idea of replacing In with Zn and Sn, but the method is successful because of the concomitant generation of three phases. It can be tried to use CuGaTe2 with similar energy gap. In addition, reducing the amount of use is also one of the solutions. If you can seek a breakthrough in the design of the device structure, try to make use of the extremely high light absorption coefficient of CIGS to reduce the thickness of CIGS by half or even to a quarter, which is another solution. Way. These may be the focus of future research and development.
Due to the significant increase in demand for crystalline silicon solar cells in recent years, raw material manufacturers of silicon materials strategically do not cooperate with the simultaneous expansion of production, resulting in the dilemma of material shortages. Those who are eager to catch a ride and participate in the investment and establishment of factories are discouraged. Turning to the production of thin-film solar cells, because the mass production technology of amorphous silicon solar cells is mature and easy to obtain, several amorphous silicon solar cell manufacturing plants were established around the beginning of 2007. However, the efficiency of amorphous silicon solar power modules Low, it can meet the demand of the towel field in the short term, but it is not optimistic in the long term. Another option on the tabletop of static film solar cells is CIGS, whose power and module efficiency is close to that of products made of polysilicon materials. More turn-key mass production equipment manufacturers with a full set of technologies came to Taiwan to sell.
The above information and development have made this industry attract attention. We can see that in the field of solar cells, there can be different materials, different processes, and different products. Higher efficiency and lower cost are the keys to winning, and there is no clear answer at present. Research in this area is still worthy of further exploration, and we have noticed that new ideas are advancing with the times, and new combinations of technologies and materials are constantly being introduced. Today, the research and development of solar cells has escaped the old framework and is heading for a new future, which has the opportunity to become a cheap and clean energy source in human daily life.
The device structure commonly used in high-efficiency CIS solar cells is shown in Figure 1(a). The energy gap values of each layer material are Zn(): 3.30eV, CdS: 2.42eV, CulnSe2: 1.02eV, decreasing from top to bottom , which can cover a broad absorption range of the solar spectrum. Usually, the P-type Cu-rich CIS film is firstly plated on the indium-coated glass substrate to form a good ohmic contact, and then the In evaporation temperature is increased to continue plating the Intrin sic In-rich CIS film. At this time, its grain structure inherits the underlying CuCIS-rich film, which is also a large grain and a rough surface. The N-type CdS buffer layer on the CIS is designed for the best matching of the heterojunction electrical properties, and a chemical bath deposition method is used to cover the slightly rough surface intact. Finally, ZnO and Al/Ni beer films were grown successively by sputtering to complete the entire device structure. The energy conversion efficiency of solar cells with CIS as the main absorber layer is currently up to 15.4%; while the efficiency of the recent CIGS solar cells is close to 20%, and its device structure is shown in Figure 1(b).
The combination of materials in the device structure of the above-mentioned high-efficiency CIS solar cells has been in the same vein since the publication of the first CIS solar cell in 1976, but it is the same as the 10% breakthrough of the CIS solar cell structure by Piyin in the 1980s. compared to the material. However, there have been some important changes, including the use of uranium glass for the Jin glass substrate, the replacement of CIS with CIGS, and the slight changes in the anti-reflection layer and light-transmitting layer. The selection and characteristics of materials for each layer are briefly described below.
In the selection of glass substrates, borosilicate glass was initially used, and later, lower-cost uranium glass was used. Nano-atoms have strong diffusion ability in solid-state materials, and can pass through the aluminum layer into the CIS coating during the CIS evaporation process. The temperature at that time was about 500 °C and the time was about 60 min. Uranium has some unexpected but very positive effects in the growth of CTS films, and some preliminary understandings have been obtained in related studies: ① it inhibits the formation of crystalline defects; ② it contributes to the P-type conductivity. Although clear experimental evidence has not yet been obtained, indirect experimental results indicate that these effects and cell efficiency have been significantly improved. In this experiment, a thin NaF-containing film was deliberately plated on the indium layer to regenerate the CIS and complete the above device structure. , Table 2 shows that the cell efficiency value is improved by about 3%.
In terms of the ohmic contact of the back electrode, it has been experimentally confirmed that there is only a few atomic layers of M0Se2 between CIS and Mo when the CIS film is grown, which improves the CIS/Mo interface with Schottky contact characteristics. quality, while showing good ohmic contact characteristics.
For a long time, CdS is considered to be the best match to form a PN junction with CIS, and the energy band structures formed by the two have been discussed in many papers. Some believe that Cd will diffuse into CIS and replace Cu to become N-type dopant, causing the position of the PN junction to move inward to the CIS, so that the internal electric field formed by the PN junction is separated from the original CdS/CIS heterojunction due to the lattice. Defects due to lattice mismatch create regions of carrier recombination.
Originally in the 1980s. The role of N-CdS is mainly a light-transmitting layer (window layer). But because Cd is a heavy metal element, the thickness of the layer is greatly reduced, and now it is often called a buffer layer (buffer layer). Because the surface of the CIGS thin film is slightly uneven, the preparation of the CdS thin layer is often carried out by chemical bath deposition (CBD) to ensure that the thin film can cover the CIGS surface continuously and completely. Recent studies have found that discontinuous CdS The thin film will make the efficiency of CIGS solar cells not higher than 18% [1:. The relevant data is shown in Figure 3. This buffer layer also has the effect of blocking the atomic impact of ZnO and reducing the generation of defects. The current research also tends to use Cd-free buffer layer materials. Based on the current research results, a direct comparison under the same conditions is shown in Figure 4. Some materials such as ZnS and other buffer layers are matched with CIGS. For solar cells, their efficiency is already on par.
In the previous CIS solar cell structure, ZnO is the anti-reflection layer (anti-reflection layer), but in the high-efficiency CIGS solar cell structure, it has multiple roles, so the i-ZnO, which is an electrical near-intrinsic semiconductor, is Light-transmitting layer, doped with Al, ZnO:AlC with good conductivity, also known as AZO) is the outermost layer with multiple functions such as light-transmitting conductivity and anti-reflection.
Since the 1950s, I-III-VI 2 compounds have been studied to understand their crystalline phases and material properties, and Shay and Wernick published the results of these studies in 1975. The device characteristics of PN junction solar cells fabricated with CulnSe2 and CdS were also published in 1976. After that, the electrical efficiency of CIS solar energy has been gradually improved with the continuous improvement of the process and matching materials. The current highest efficiency I-III-VI 2 thin film solar cell, its main absorber layer, its Ga/In ratio must be maintained within 0.3. . If it is higher than 0.3, the solar cell efficiency will be greatly reduced. The efficiency of solar cells made with the ternary compound CulnSe2 is slightly lower because although the material has a rather high light absorption coefficient, the energy gap of CulnSe2 is only about 1.0 eV. The addition of Ga makes the material a quaternary compound, which can increase the energy gap of CulnSe2, and also increase the open circuit voltage of the battery and reduce the current to reduce the resistance loss, thereby achieving higher battery efficiency. Before I accidentally browsed an article, the author wrote very detailed and comprehensive on the various factors affecting battery efficiency. If you are interested, please visit tycorun.com.
CulnSe2 is the most studied of all I-III-VI 2 compounds, and of course good results have been obtained. However, even if Ga is added to increase the energy gap to improve the cell efficiency, as mentioned above, when the value of x exceeds 0.3, the expected effect cannot be obtained, and when r = 0.3, the energy gap can only reach 1.15 eV. There have been some attempts to change the energy gap by means of the composition of quaternary compounds, or even pentads, in an attempt to improve cell efficiency, but this has not been the case, but has reduced the efficiency. It seems that for a long time, the successful experience of material matching and improvement obtained in CulnSe2 and Se2 cannot be directly applied to other I-III-VI 2 series compounds, especially CulnSe2, which is in addition to CulnSe2, additionally by composition. The I-III-VI 2 compounds of N-type and P-type can be obtained by formulating, as shown in Table 1, its energy gap is about 1.5 eV, as mentioned above, it is a kind of solar cell with appropriate energy gap value Material. Often these changes in material properties include interfacial and intrinsic defects.
The combined effect is enough to lead to considerable negative results, and it is clear that enough research efforts are still needed for these I-III-VI 2 series compounds to show their proper device performance. Therefore, the following is an introduction to the device structure and process of high-efficiency solar cells.
Thin film process of CuInSe2
High-efficiency CIS solar cells use co-evaporation (co-evaporation) or solarization (selenization) reaction method to coat CIS thin films. Other methods, such as co-sputtering, have defects such as defects caused by high-energy film surface impact, and insufficient control of film composition due to In repulsion. At present, high-efficiency solar cells cannot be produced by this process; For example, the thin films plated by low-cost electrochemical deposition (electro deposition) are not suitable for the fabrication of high-efficiency solar cells due to their poor quality.
The three-source (CIS) or four-source (CIGS) co-evaporation process uses an elemental evaporation source to evaporate at a substrate temperature of 450 to 600 °C. Because the vapor pressure of the Se element is high, the ratio of Cu to Group III elements 1), the grain size is about several hundreds or even thousands of nanometers, and the film has a rough surface; while the grain size of the In-rich film is less than tens of nanometers, so the surface of the film is smooth bright. In the process of CIS thin film evaporation, two binary phases (Cu2Se and In2Se3) are formed first, and then the CuInSe2 ternary phase is formed by further reaction.
In the 1980s, the CIS solar cell produced by the Boeing laboratory in the United States broke through 10% of the energy conversion efficiency, reaching 12%, and its CIS thin film was successively prepared under Cu-rich and In-rich evaporation conditions. CJS films with large grains can be obtained under Cu-rich conditions but contain Cu2Se secondary phases. During the growth of polycrystalline films, Cu2Se is liquid and appears on the surface and grain boundaries, which helps Grain growth; followed by In-rich growth conditions, the excess In2Se3 binary phase will completely react with the Cu2Se secondary phase in the Cu-rich CIS film and be eliminated, resulting in a large-grain and single-phase CIS The film, of course, the component formed first in the film contains a little more copper and exhibits P-type conductivity, while the part formed later is a component with more In, so it has a high resistivity N-type close to intrinsic (in trinsic) ) Conductivity properties of semiconductors.
At the same time, ARCO solar company also developed the CIS solarization process. Although the efficiency of CIS solar cells produced by this process is slightly lower than that of solar cells produced by evaporation process, it is close behind. The tanning process is to first coat Cu and In metal films with specific thicknesses to achieve a specific atomic number ratio, and then place them in H2Se gas or Se vapor, and react them into Cu2Se at a temperature above 400 °C. The composition uniformity of the thin film prepared by the solarization method is slightly inferior to that of the vaporizer, but it can still meet the requirements of high-efficiency solar cells. Process, suitable for mass production planning. In addition, the new tanning method adopts rapid thermal annealing (rapid thermal ling, the heating rate is at least 10°C/s or more), and the Cu/In/Se three-element pre-plating layer (precursor film) on the extraction plate is heated at a temperature of 400~500°C. The drying reaction is completed in a very short time (1~5 min), obviously this process has the advantages of large output and low cost.
If the solarization method is gradually heated, the film formation process is a series of reactions that are carried out successively, that is, the formation of polythematic binary mesophases such as copper resistance and induration, and finally the synthesis of ternary. CIS single phase.
The formation of intermediate phases can be skipped if rapid heating is used. Directly synthesizing CIS compounds, the materials prepared in this way will not cause loss of cell efficiency. As for the mechanism of film formation, it is difficult to know because the synthesis speed is too fast, but indirect methods can still be used to find better materials.
At present, the highest conversion efficiency of solar cells with CIGS as the main absorber layer has reached 19.2%, which was proposed by NREL in the United States in 2003. NREL uses an improved evaporation method called a three-stage process. As the name suggests, this method divides the entire process into three stages to modulate the substrate temperature and control different element sources and their evaporation temperatures.
The first stage is to heat the uranium glass plated with aluminum (Mo) metal to 260°C, and to provide the elements In, Ga and Se to grow On, Ga) precursors at the same time; in the second stage, the elements In and Ga are turned off and replaced to provide Elemental Cu and vat, well heat the substrate to 560°C, at this time the quaternary compound Cu On, Ga) begins to form, accompanied by the formation of a secondary phase on the film surface, which is liquid and helps to generate large grains And dense columnar crystals; the third stage is to turn off the element Cu and keep the substrate temperature at 560 ° C, and then continue to provide In, Ga and Se for a short period of time and also turn off In and Ga, so that enough elements In and Ga are combined with the secondary phase. Reaction again to form Cu2(In,Ga)4Se7 or Cu1(In,Ga)3Se5 film on the surface, so Cu(In,Ga)4Se2 film with slightly insufficient copper (Cu-poor) will eventually be formed. Its composition The ratio is 0.93<Cu/(In+Ga)≤0.97. Finally, in the elemental Se atmosphere, do high temperature annealing for 5~10 min for recrystallization.
Cross-sectional grain structure of CIGS thin films grown by three-stage evaporation process The depth distribution of the constituent elements of the CIGS thin film prepared by this method is shown in Figure 1. By adding Ga to form a quaternary compound, the energy gap value of the main absorber layer can be increased, so that the open circuit voltage can be increased by 20-30 mV. As for the concentration distribution gradient of Ga, the energy band design concept of a-SiGe solar cells is used to make the conduction band form a V-shaped double slope. The opportunity for recombination at the back metal contact interface enhances the collection of charges, and the light incident direction benefits from the design of the slope of the energy gap, which expands the coverage of the light absorption band. Overall, the short-circuit current can be increased.
The successful application of the above-mentioned material and device design concepts is the main reason for pushing the efficiency of CIGS solar cells above 15%. In fact, replacing part of the tanned with sulfur into a quaternary compound can also increase the energy gap to obtain the same effect as CIGS, but the vapor pressure of sulfur is much higher than that of tanning. If the evaporation method is used, a special evaporation source design is required for control. The output of sulfur vapour is regulated by a special valve. At present, in the mass production process of CIG tanning, it is also possible to simultaneously insert tanning and sulfur to synthesize CIGSS five-membered compounds.
In order to reduce the cost of CdTe films, polycrystalline structures are often the mainstay, which can be prepared by various methods, including evaporation, close-spaced sublimation (CSS), and vapor transport deposition (vapor transport deposition). , VTD ) , chemical vapor deposition ( chemical vapor deposition , CVD ) , electrochemical deposition ( electro deposition . ED ) , screen printing ( screen printing ) , thermal plate spraying ( spray pyrolysis ) , sputtering ( sputtering ), etc. has been used . Figure 7.4 organizes various CdTe coating processes and related process conditions for reference and comparison . Because CdTe has ionic bonds. It is easy to form Cd and Te bonds during the film formation process, and different coating methods are used. It is not difficult to control the stoichiometric composition. Even elemental films of Cd and Te can be plated separately at room temperature and then heated up. to form compounds.
It is worth mentioning that CdTe solar cells often require an additional heat treatment procedure. Also CdCl2 is usually mixed into the air with a partial pressure of 0.04 and heat treatment at a temperature of 425 ℃ for 20 min. This promotes the growth of CdTe grains and the passivation of the crystal interface, so that the high resistance at the crystal interface disappears, and at the same time, it promotes the mutual diffusion of materials between PN heterojunctions, forming ternary compounds (such as CdS/CdTe near the interface. Produces CdTe1, S,) with a graded composition, thereby reducing the interface defect density caused by the lattice mismatch between CdTe and CdS. This procedure helps to greatly improve the efficiency of solar cells; while 0 in the air has The effect of increasing the P-type conductivity of CdTe and promoting CdS/CdTe interdiffusion. There is also a saying: Since this is a necessary follow-up process. It does not matter which method is used for the CdTe coating, as much as possible using the low-cost method.
Electrical properties and device technology of CdTe
The combination of N-CdS/P-CdTe in II-VI solar cells can reach 16.5% energy conversion efficiency. The large-area CdTe solar cell module has entered the mass production stage in recent years, and its energy conversion efficiency has exceeded 10%. Different companies use different preparation methods for the P-CdTe film in the device structure, which are coating technologies such as close-range volatilization, electrochemistry, and vapor transport.
The P-type doping of CdTe is dominated by P or As, and it is usually doped under Cd overpressure, that is, Cd-rich conditions. The hole concentration of 2 × 101; cm – 3 can be obtained for the CdTe epitaxial film prepared by CVD or MOCVD, while the polycrystalline film is only 6 × 101 cm . As doped CdTe prepared by photo-assisted molecular beam epitaxy (PAMBE) has the highest hole concentration of 6. 2 × 1018 cm – 3, which is conducive to the activity of atoms on the surface under illumination , and also makes Te more easily desorbed, and the effect on the doping amount is not obvious. In the MBE process, some people have also used low-energy ion beam doping, and its hole concentration is only 2 × 1017 cm3, but the material defects caused by the ion beam impact make its electrical properties worse, and it often needs to be regrown on it. A layer of undoped CdTe to avoid reducing the short-circuit current of the solar cell.
In terms of ohmic contact, because the work function of P-CdTe is as high as 5.7 eV, no metal can form an ideal ohmic contact with it, so the metal alloy containing Cuo.12Au0.88 is rich in Te after treatment. The surface of CdTe is diffused from Cu into the lattice position where CdTe replaces Cd, which promotes the formation of high content of P-type doping on the surface of CdTe, and holes are conducted through strong electric field tunneling. In the module mass production process, the 1-contact resistance contact of the P-CdTe polycrystalline film is often fabricated by electron beam evaporation on the surface of CdTe after CdCl2 annealing treatment → Cu thin layer, and then heated to promote the diffusion of Cu into CdTe, achieve surface doping.
In summary, the fabrication of high-efficiency N-CdS/P-CdTe solar cells should be based on the following designs and steps: ① Coating the CdTe film on the transparent CdS/TCO/Glass substrate; ② The thinner the CdS light-transmitting layer, the more It can improve the spectral response of short wavelength; ③ After the CdTe coating, a heat treatment procedure is required in the atmosphere of CdCl2 and O2 to improve the grain structure and P-type conductivity of CdTe, and also promote the interaction between CdTe and CdS. The interdiffusion of CdTeS is formed to reduce the number of interface defects; ④ The CdTe layer is chemically treated to make its surface rich in Te, then plated with Cu and heated to diffuse into CdTe to increase the amount of P-type doping, and finally plated on Conductive materials such as Ni or graphite.
Future development of CdTe thin film solar cells
CdTe solar cells are mass-produced by one or two companies in the United States and Germany. The energy conversion efficiency of large-area CdTe modules can currently reach 10.5% and the service life is as long as 25 years. Its unit cost has the opportunity to be $1 per watt. The following are the thin-film solar cells with the lowest cost at present. Furthermore, Cd and Te elements can be obtained from the smelting process of basic metals such as copper and iron ore. However, a negative factor in the application of this product is that the Cd contained in the material is a heavy metal element, and the environmental protection problems caused by its process and disposal have attracted much attention. Although the industry emphasizes that the pollution of Cd compounds is very different from that of elemental Cd, and that the process and waste recycling treatment procedures can meet strict requirements, a company in the United States even takes long-term samples of its employees in the production line. Examination of the Cd content in the body has not detected higher than usual results, but it is still worrying. Under the competition of other materials, although the EU’s restrictions on the Cd content of most electronic products have not yet included solar cells, the manufacturers of CdTe solar cells mass-produced in the short term have fewer competitors to produce the same products, and the products are more expensive. This is lower than solar cells made of other materials, thus making profits. Manufacturers even give a plan to pay the cost of recycling solar panels in the future, but in the long run, under the stricter environmental protection requirements in the future, it is not conducive to CdTe solar cells. development of.
As shown in the periodic table of elements in Fig. 1, group E elements (including Zn, CD, Hg) and group H elements (i.e. s, Se, Te, etc.) constitute group II and VI compounds with semiconductor properties, Its basic properties are shown in Fig. 1[5l. Except that CdTe and ZnSe have n-type and p-type conductive properties, other II-VI compounds can only obtain a single conductive form, such as p-type ZnTe and n-type CDs, CdSe, ZnS, etc. Therefore, there are some restrictions on the matching of materials. According to theoretical calculation, PN homojunction (Homo J function) The energy conversion efficiency of solar cell is the highest, and its energy gap is 1 About 5 EV, so CDT has naturally become the most suitable material for solar cells in h-group compounds.
II-VI compounds have stronger ionic bonds than V group. In addition, the vapor pressure of group II and VI elements is similar and their value is not low. Therefore, in the preparation of materials, it is easy to form compounds by self interactive coordination. For the growth of thin films, if the substrate temperature is set so that the same atoms (ii-ii or vi-vi) cannot form valence bonds, the elements of group H and group m only have interactive valence bonds to form group II-VI compounds; Under normal coating conditions, the temperature is not high, about 300 ℃.
Some of the crystalline structures of II-VI compounds are cubic sphalerite structures, such as ZnSe, ZnTe, CdTe, etc., while others are hexagonal structures, such as CdSe; CDs and ZnS have one or both of the above two structures according to the material preparation conditions. Different structures have different properties.
In semiconductor materials, the strength of valence bond is directly related to the energy gap. From the periodic table in Fig. 1, those with small atomic numbers of constituent elements in II-VI compounds have strong ionic bonds. For example, Zn Se is stronger than Zn te, and CD-s is also stronger than CD te; Their energy gap also reflects the strength of bidding bond. Compared with CD te, Zn te has stronger valence bond and larger energy gap.
I-ill-vi group 2 compounds
Group I VI 2 compounds can be regarded as three element compounds derived from group II VI, that is, the elements of group H are replaced by group I (Cu, Ag) and group I (al, GA, in) and combined with group VI (CS, Se, TE). Because there are two kinds of atoms arranged alternately and orderly in the lattice position occupied by cations, the unit cell seems to be formed by stacking the unit cells of two amphibolites. However, the unit length ratio of c-axis and a-axis is different due to the valence bond strength of two different cations and anions, which is not equal to 2. Such a crystalline structure is called chalcopyrite structure, as shown in Figure 2.
As can be seen from the periodic table of Fig. 1, this series of compounds include 18 compounds such as group I elements Cu and Ag, group IV elements Al, in and GA and group M elements s, Se and te. Like other compound semiconductors, their energy gaps can form quaternary or even quinary compounds through the mutual substitution of homologous elements in a certain range. The biggest difference between it and tianv or HVI compounds is that there is no composition very close to the chemical ratio. The composition stability range of single phase can reach 3% ~ 5%, and the degree of deviation from the fixed ratio composition of 1:1:2 is considerable. The above situation can be seen from the pseudo binary phase diagram of Fig. 2. The situation of deviating from the fixed specific composition will produce intrinsic point defects in the material. There are 12 kinds of ternary compounds of III-VI group, and the distribution of the number of various point defects is closely related to the chemical composition and the formation energy of defects.