Fundamentals of the technology of production of silicon solar cells

Photovoltaic or solar cells are semiconductor devices that convert sunlight into electricity. Today crystalline silicon and thin-film silicon solar cells are liders on the commercial systems market for terrestrial applications. The article describes the basics of traditional technology, developed in Ukraine at 2001-2005 and implemented into production.

Photovoltaic cells or solar cells (solar cells) — this semiconductor products that convert sunlight into electricity. There are different technologies of solar cells, the design of which is distinguished as the physical principles of conversion of solar radiation into electric current and less important details. Most effective in terms of energy, devices for converting solar energy into electricity are semiconductor photovoltaic cells (solar cells), as it is a direct, one-step transfer of energy. Today the market of commercial PV systems for terrestrial applications are most noticeable crystalline silicon (about 80-85% of the world market) and thin-film solar cells (about 10% of the market). Next we’ll talk about the production of crystalline silicon solar cells, which are a key component of solar panels.

Chemical treatment

The most important and expensive part of any of the solar cell is a silicon wafer. It can be both monocrystalline and multicrystalline. As the name suggests, the monocrystalline wafer is single crystal, of which, for instance, by wire cutting silicon wafers obtained the required thickness and size. Typically monosilicon is grown as round ingot, which is then cut out so-called pseudoquare wafer. This form provides the maximum use of a circular silicon ingot and, at the same time, the most dense surface coverage of the future of solar module (solar panel). Monocrystalline wafer is regular squares represent predetermined size and thickness.

Since each wafer has a surface damaged when cutting bars at the nano level, this damaged layer must be removed. If just a few microns to remove by chemical etching, the surface of the wafer would be smooth and will reflect a substantial portion of the incident radiation. As for more efficient solar cells convert into electricity, it is important how much sunlight, then try to make a rough surface on the micro level. Monocrystalline wafers for such process operation is called texturing.

The textured surface is a set of seemingly randomly distributed micro pyramids. Light incident on the surface of the pyramid, is reflected at the same angle and in most cases falls on the brink of a neighboring pyramid. Thus, only by creating a textured surface can reduce the reflectance of silicon from 35% to 11%.

To solve both problems considered (removal of the damaged layer and the formation of the texture on the surface of the wafers) used wet chemical processing wafers. Composition of solutions, temperature and duration of treatment depends on the type of processed wafers, the surface condition before the treatment, further processing operations and many other factors. Typically, wafers are used for monocrystalline etching in alkaline solutions and one or more acid treatments. Thus important — not overzealous as in pursuit of uniform surface texture can significantly reduce the thickness of the wafers. This will lead eventually to enhance their battle for subsequent operations and, as a consequence, to decrease percent product yield.

At the end, after all the necessary operations, the wafers are rinsed in water and dried. It is also very important operations. For example, the quality is very much dependent on the drying parameters of the diffusion layer produced by the following operation.

The heart of the solar cell

A key element of the design of crystalline silicon solar cells is a p-n junction. What is it? The fact that a semiconductor depending on their conductivity type may be either n-type (electron conduction), and p-type (hole conduction). At the same time, if one type of wafer to create a layer of another type, the point of contact of these areas and will be a p-n junction. Generally p-n junction — this is one of the key concepts of solid-state microelectronics. Using physical effects occurring therein or near work such remarkable products as diodes, transistors, and many other more complex circuits.

One of the main characteristics of the p-n junction is its ability to be an energy barrier for the charge carriers, i. e. they pass only in one direction. It is for this effect and the generation of electric current based solar cells. Radiation falling on the surface of the element, in the semiconductor generates charge carriers with different signs — electrons (n) and holes (p). Due to its properties the p-n junction «shared» blocked skipping every type only «a» half and moving randomly in the volume element carriers are on opposite sides of the barrier can then be transferred to an external circuit for applying voltage to the load connected to the solar cell.

Of course, this description is somewhat simplified, but even it shows that without the p-n junction is virtually impossible to convert solar radiation into electric current (remember that we are talking about the classic terrestrial solar cells, and in fact there are more exotic photovoltaics, design does not imply the existence of p-n transition). So how is possible to form a p-n junction?

As we already know, at first original wafer undergo several stages of chemical processing, whereby their surface becomes necessary for us the structure and purity. Typically in the manufacture of solar cells used with a starting wafer of p-type conductivity. For this step a further silicon ingot doped with appropriate cultivation impurities, e. g., boron. Therefore, to create a n-layer it is necessary in one of the surfaces to introduce another impurity element which compensates for effects of boron and saturate carriers semiconductor n-type. This can be done by introducing phosphorus in silicon or other suitable admixture of the relevant part of the periodic table.

One of the most traditional and economically feasible ways to fill silicon is phosphorus diffusion, i. e. a process in which phosphorus at high temperatures penetrates into the semiconductor. Traditionally, diffusion of phosphorus is carried out in tubular furnaces or conveyor at temperatures around 800 oC. In the first case, the quartz wafer is placed in the cassette, and the tube furnace is filled pairs substance containing phosphorous. Handling Time Spent wafers in the reactor temperature and gas flows within it, technologists receive a p-n junction with the necessary properties to them. In the latter case, the phosphorous material is sprayed onto the surface of the wafer lying on a conveyor belt furnace. Thereafter the wafers are moved using the conveyor to the next zone, which is also subjected to high temperature treatment.

As a result of the diffusion of phosphorus on the surface of the silicon wafer and the ends of a thickness of about 200 microns, a layer of n-type, penetrating to a depth of only about 0.5 microns. I. e. p-n junction occurs at the surface of the solar cell. This is done to ensure that the charge carriers of different signs generated by the radiation as possible were in the area of influence of the p-n junction, otherwise they will meet again with each other and thus balanced by, and to prevent any contribution to the generation of electric current.

Plasma etching

The construction of a solar cell requires a p-n junction in the vicinity of one of its surfaces, which is called the front or working side. The other surface is called back. Typically located on the front surface of current collecting grate, and rear-solid contact. Since the n-layer is formed by diffusion, heavily saturated with impurities, it is a good conductor of electric current. During this diffusion layer is formed not only on the front side of the wafer, but at its end and even on the rear surface of the perimeter. Here we have the electrical circuit between the front and back current collecting contacts.

Typically, this problem is solved by physically removing it from the ends of the wafers. This can be done mechanically, laser, chemical or plasma etching. Without going so far in the analysis of the advantages and disadvantages of each of these methods will say that one of the most rational of them still is plasma etching. Operation is a plasma processing stack tightly pressed against each other silicon wafers. The process of removing the silicon depends on many parameters, among them — the duration of treatment, the plasma composition, the direction of flow of ions in the reactor location and tightness of the wafers, the size of the stack, and others.

If you deviate from the optimal regimes maybe two opposite results:

  • n-layer at the ends of the wafers are not removed to the circuit that carries on the edges of the solar cell after forming contacts;
  • n-layer is removed, not only at the ends of the wafer, but also on the perimeter of the front and back surfaces, i. e. circuiting occurs at the contact regions with different conductivity type on its front side.

Monocrystalline wafer with baited on the face after plasma etching as follows:

In the first case, the problem can be solved by controlling the conductivity type at the ends of processed wafers. The second case is more complicated — spoiled electric parameters and the appearance of the front surface requires full bleed n- layer, and only thereafter can be sent again to the initial processing steps. When this re-texturing reduces its thickness, i. e. increases battle wafers on subsequent operations. Additional processing and low yield ratio leads to increased costs and significantly impair the economy.

Antireflection coating

From the reflection of the surface texture of the wafer is reduced from an average of 35% to 11%. This means one tenth of the radiation incident on the surface of the solar cell, will still be reflected back and will not participate in the generation of electric current. In order to further reduce these losses are classified as optical for the next machining operation on the working surface of the solar cell is applied to a so-called antireflection coating (ARC). Based on the laws of optics, engineers selected thickness and refractive index of the coating, so that we can reduce the reflection up to 1-2%. And this is a very good indicator.

Today there are so many different types of anti-reflection coatings, which are applied by several different methods (APCVD, LPCVD, PECVD, etc.). In practice, as ARC for silicon solar cells are most commonly used titanium oxide film or silicon nitride, the latter being increasingly preferred. Silicon nitride is usually applied by PECVD, i. e. by accelerated plasma chemical deposition from a gaseous phase, in special tube furnaces.

PECVD process assumes that the chemical reagent zone entering the reactor decomposes under the influence of plasma and temperature on the individual elements which are then deposited on the wafer surface and react chemically. As a result, the front surface of the wafer «grown» thin film of silicon nitride which has the desired properties. Its thickness is about 70 nm, which is much smaller than the micropyramid texture and allows to achieve antireflection effect regardless of the surface relief structure.

This method provides a very good uniformity of the coating. Estimate the thickness of ARC is quite easy on the eyes even. Optics such that the thickness unevenness, the greater the coloring surface of the wafer is changed. The same effect can be seen looking at a puddle of spilled gasoline — film all colors of the rainbow, making it clear how changing its thickness. Other known methods of applying ARC often require a preliminary application of the solution on the wafer surface, which cannot be done uniformly (substance will accumulate in «gorges» between micropyramids and will not dwell on their tops). It also finally affects the appearance of the solar cell deteriorates and its parameters.

After applying ARC silicon wafer absorbs a large part of the solar radiation incident upon the surface. Moreover, coating thickness is optimized such that run most efficiently in the most efficient range of the spectrum. Later on, I plan to devote a separate publication this issue, and now I can only say that it is the blue part of the spectrum. It is because of this, all solar cells are beautiful and deep navy blue color.

After applying ARC solar cell is almost ready. Under the influence of radiation inside the device is already happening generation of charge carriers, which are then separated p-n junction and almost ready for use. But they need to pass into the load circuit, and for this it is necessary to form a contact surface of the solar cell.

The front contact metallization

The front surface is first and foremost for maximum absorption of the radiation incident on it, and this defines the technical requirements for the contact metallization. It is for this reason contact located on the working side of the solar cell is performed in a grid, usually consisting of 2-3 wide pads and several dozen thin current collector lines perpendicular to the broad.

When choosing the design engineers have personal contact to solve two conflicting aims. First, to reduce the optical losses due to shading metal work surface, they try to perform the grid lines as thin as possible and place them as far as possible from each other. Secondly, since the surface of the element has a specific electrical resistance (defined modes of formation of the p-n junction by diffusion), at a very great distance between the contact elements of the lattice part carriers simply do not have time to reach the contact and recombine inside the semiconductor material. Therefore, in order to reduce electrical losses for a given surface resistance wafer spacing of the lattice of the contact may not exceed a certain value. The same goes for the width of the lines — the thinner the line, the better optics, but the less current will be able to hold such a contact. Plus the very method of forming metallization has its limitations on the minimum line width. For example, the contact width of 125 microns to do is quite simple and inexpensive, and the contact width of 80 microns — that’s the lot of laboratories and enthusiasts.

Typically, in order to reduce the cost of solar cell contact metallization is applied by screen printing or as it is called — silkscreen. The essence of this method is that by using the so-called squeegee (if very simple, the rubber bar) through a fine mesh stencil pressed paste, composed of metal balls, flux, and various binding agents. On the preformed grid pattern defining places where the paste is to be applied to the wafer, and in which — no. After this paste is dried, and the wafer enters the furnace of brazing, where at temperatures above 800 degrees baked into the metal surface in a solar cell.

Since the paste contains enough solid components, to increase the stability in the manufacture of solar cells used stencils formed by metal meshes. The width of the thin lines of contact metallization puts demands on the screen parameters, which typically can be from 165 to 325 openings per inch. This allows you to get in the end contact width 125 mm, located at the desired location with a positioning accuracy better than 10 microns. More detailed method of screen printing stencils and actually, I will look a little later.

For the formation of the front metallization of solar cells used today complex composition containing silver paste. Type of pasta and its properties are very important to obtain good parameters of the finished product. According to my estimates for only a few years due to technological improvements pastes, producers managed to raise the efficiency of solar cells by 1-1.5%. And given that the efficiency of conventional silicon terrestrial element now lies in the range 15-17%, that such an additive is more than essential.

Besides the above-mentioned problems in the formation of the front contact metallization, I cannot mention another important point. As we said above, p-n junction lies at a depth of about 0.5 microns. When brazing with silver paste, the metal must as much as possible enter in the n-silicon layer, but it does not reach the p-layer. Otherwise there will be a electrical contact between the two types of semiconductor and solar cell would be short-circuited. This requirement complicates the task of brazing paste, which must penetrate into the wafer with a truly exquisite accuracy. For this thermal treatment is carried out in special high furnaces, providing the peak effect on the wafer for a time of about 10-15 seconds.

Of course, many manufacturers use the old-fashioned slow conveyor ovens, in which the wafers are in the area burning in about 3 minutes and also get acceptable parameters of products. But the transition to the use of high-speed furnaces can increase the efficiency of the solar cell has at least 0.5%. And this is a significant advantage in the total capacity of products produced per year.

Backside metallization

Unlike the front metallization serves only contact metallization rear solves another problem. Usually on the back surface of the solar cell is put not one, but two types of metal. One of them is a continuous aluminum layer covering substantially the entire area except for a few holes. And in these holes formed familiar to us silver plating, performing the function of the contact.

Why do I need an aluminum layer? The fact that it serves as a mirror for the charge carriers. But not the optical mirror and energy. As is well known, electrons and holes in semiconductors tend to recombine, i. e. move from free to bound. Recombination occurs, for example, if at one point there are two different charge signs. One of the extreme cases of the so-called surface recombination because Any surface is a set of dangling bonds of the crystal lattice — «traps» for the free charge carriers. It is in order to reduce the impact of this type of recombination in solar cells with aluminum metallization formed the so-called BSF (back side field) — «reflect» carriers, which have not yet give a contribution to the current generation.

The aluminum layer on the back side provides a solar cell of a few tens of mV higher voltage than it would have been without it. I. e. 24 in the case of solar cell, this additive is at least 0.5, which corresponds to around 2% increase in capacity. That’s dropped engineers and increase the efficiency of solar panels, making it more efficient.

As the silver-containing contact and a continuous aluminum layer on the back side of the solar cell are formed in the same screen printing method. Changes only drawing stencil and some parameters utilized in the grid. For example, due to the size of aluminum particles in the ink have to be used with larger mesh cells in comparison with grids for applying silver pastes. Also features taut wire and some other process parameters. After application of each layer of the paste was dried in ovens, and only after the application of the three layers (one on the front and 2 on the rear side) is transmitted to the sheet by heating.

In general, the process of screen printing is very moody and requires experience of all participants — engineers, engineers and operators. For example, viscosity and other properties of the pastes are very dependent on the temperature and humidity in the shop, draft, etc. Changes in the indoor temperature even at 2-4 degrees requires cumbersome reconfiguration process. Also affect the outcome terms and conditions of storage of pastes, state squeegee, printer settings, and even the characteristics of the air supplied to the furnace brazing with. Particularly difficult to reconcile all of these options because the formation of metallization — this is the last operation and the wafers had gathered all the deviations from previous operations. I. e. all dimensions and other parameters wafers differ in the party is much stronger than in the first step, and tunes in to the optimum so much harder. But all the effort is worth it, since the output we get finished solar cell, which remains the only measure in the future, sort and pack the parameters.


Obviously, any article of manufacture prior to delivery must be carefully checked and measure its parameters. Do not avoid this procedure, and in the manufacture of solar cells.

The first solar radiation characterized parameters such as the intensity, i. e. power incident on the surface of a certain area. This parameter is different in different regions of the Earth; the maximum intensity of solar radiation on the planet exceeds 1300 W per square meter, but for convenience in measurement, a standard value at a level of 1 kW / m. m

Another important parameter is the distribution of solar radiation by wavelength, i. e. spectral composition of light. To characterize it introduced the concept of «air mass» (AM — Air Mass): so the spectral composition of the radiation from the air mass corresponds AM0 spectrum of sunlight outside the earth’s atmosphere; AM1 corresponds to radiation at the Earth’s surface, provided that the sun is directly above the point of observation, ie rays of light passed through the 1 atmosphere; measurements is standard AM1, 5, corresponding to solar radiation, which took 1.5 atm.

The final condition of the standard measurements of solar cells is temperature. The fact that the characteristics of this product somewhat reduced when the temperature rises, so it is very important that it has not changed during the measurements. Standard assumed that testing is performed at 25 degrees Celsius.

During testing of the solar cell measured the set of parameters, including short-circuit current, open circuit voltage, maximum power and the efficiency. On the question raised at the beginning of this article provides an answer setting the maximum power of the solar cell efficiency and shows what part of the incident power will lead to the emergence of electric power to the load.

For precision measurements of solar cells used testers or testers / sorters. They are both pulsed and continuous irradiation. Pulse testers interesting because during almost instantaneous measurement element could not heat up and the error below. Testers also differ in the type of lamps, which affects the spectral composition of the radiation.

Usually tester contains a reference solar cell and built-in computer, which converts the measured parameters and brings them to the standard conditions of measurement. At first glance it seems that the measurement is a fairly simple task — you just need to change the voltage and measure the corresponding currents. However, in reality it is necessary to consider a number of additional factors, among which the most important are the following:

  • Since the generated current is directly proportional to the element of light, it must be accurately known and constant.
  • Necessary to achieve high uniformity of light on the surface of the test element.
  • The spectral distribution of light should be as close as possible to the spectral distribution of natural light.
  • You must know the temperature being measured.
  • Necessary to eliminate any voltage drop at the contacts and in the chain, which introduces additional errors in the measurements.

To ensure uniformity of radiation, there are two most common ways: use special optics with reflecting and diffusing elements or use a point source of radiation. In the first embodiment, it is often necessary to control and configure the hardware, and to compensate the influence of the reflector and other optical elements on the spectral distribution using an additional filter, which leads to great difficulties in use. In the second embodiment, the desired uniformity is achieved diversity of the source and the test image to a considerable distance. Most common in the industry are pulsed xenon lamp testers.

So, the above were the fundamentals of traditional technology developed in Ukraine and introduced into production. It was probably the most common technology for creating silicon solar cells with contacts deposited by screen printing. As it may seem, the production process of photovoltaic cells is quite simple compared to traditional microelectronics products. But it is only at first glance. In fact, technology solar cells there are so many difficulties and pitfalls.

Below you can find a visual illustrating the latest achievements in the efficiency of photoelectric solar cells manufactured using various technologies.