RU2377695C1 - Semiconductor photoconverter and method of making said converter - Google Patents

Abstract

FIELD: physics; semiconductors.

SUBSTANCE: invention relates to semiconductor engineering and specifically to photoelectric converters (PC) for direct conversion of solar energy to electrical energy using solar cells. The invention can be used in renewable sources of energy. According to the invention, in the semiconductor photoconverter which consists of monocrystalline silicon wafers with vertical diffused p-n junctions connected to each other into a single horizontal structure by metallic liners, with current leads with at least one light-receiving surface with an inversion layer and a dielectric antireflection coating in the p-type conduction region adjacent to the diffused p-n junction, parallel to the light-receiving surface there is at least one nano-cluster region with n-type conduction at a distance from the inversion layer equal to or less than the sum of the thickness of the space-charge regions of the inversion layer and nano-cluster region. A method of making the photoconverter is also disclosed.

EFFECT: increased efficiency of the photoconverter due to longer mean life of non-equilibrium charge carriers with simultaneous reduction of the cost of the technology of making the photoconverter due to its simple design.

2 cl, 4 dwg

Description

The present invention relates to semiconductor technology, namely to photovoltaic converters for direct conversion of solar energy into electrical energy using solar cells.

Scope - renewable energy sources.

Photoelectric converters (FPs) are known for direct conversion of solar energy of a planar structure with pn junctions located along a light receiving (photosensitive) surface, i.e. perpendicular to the flux of light radiation.

This type of phase transition, in particular, designed to convert concentrated solar radiation (high-current, concentrator phase transition), has a special contact grid for collecting photogenerated charge carriers, which also serves to reduce the series resistance. The use of a contact grid requires a complex and expensive technology for its manufacture, including photolithographic processes (US patent No. 4320250, class H01L 31/0224, publ. March 16, 1982) [1].

A technical solution of a semiconductor FP with pn junctions, which are located vertically to the light receiving surface, i.e. along the stream of light radiation (RF patent No. 2127009, class H01L 31/18, publ. 02.27.99, figure 1-3) [2]. Such a phase transition does not have the maximum possible coefficient of efficiency (COP), since it has a relatively small volume of the space charge region (SCR) of pn junctions and a low conversion efficiency of ultraviolet radiation (UV).

The closest is the design of the semiconductor photoconverter (RF patent No. 2147472, class H01L 31/18, publ. 03/10/1999, figure 1, 2) [3]), consisting of single-crystal silicon wafers with vertically located pn diffusion transitions interconnected in a single horizontal design by metal spacers, with current-carrying contacts, while the light-receiving surface of silicon with an inversion layer is coated with a dielectric light-transparent coating.

In other embodiments, the FP can be without an inversion layer and without a dielectric light-coating.

The design drawback is the small SCR of vertical diffusion pn junctions, which leads to a decrease in the lifetime of minority charge carriers in single-crystal silicon wafers and, accordingly, a decrease in the efficiency of the phase transition, as well as a complex system of current-collecting contacts with interdigital geometry.

In a simplified version, when there is no inversion layer, it is possible to provide double-sided illumination, which gives the advantage of eliminating spurious shunting of the unlit portion of the phase transition, but it is impossible to increase the efficiency of the phase transition by collecting the generated charge carriers in the near-surface regions.

A known method of manufacturing a semiconductor photoconverter with vertical pn junctions (Sater BL at all "The multiple Junction Edge Illuminated Solar Cells" in Conf. Rec. Tenth IEEE Photovoltaic Specialists Conf., 1973, p.188-193 [4], US patent No. 4516314, CL H01L 31/18, publ. 05/14/1985 [5]), in which the connection of single-crystal silicon wafers is carried out by sintering them with aluminum foil. Deep diffusion is preliminarily performed to obtain a pn junction and nn ⁺ junction in the plates to create an isotype barrier. After the creation of pnn ⁺ junctions in the plates, they are collected in a column, and they are sintered with aluminum foil. By cutting the column along planes perpendicular to the pn junction, samples of the required thickness are made, from which, after additional processing, FP with vertical pn junctions is obtained.

Deep diffusion leads to a decrease in the lifetime of minority charge carriers, both in the doped regions of the device and in the base ones, which leads to a decrease in efficiency.

The closest is a method of manufacturing a semiconductor photoconverter (RF patent No. 2124472, class H01L 31/18, published March 10, 1999 [3]) by forming diffusion pn junctions on the surface of single-crystal silicon wafers, metallization of the wafer surface, assembly of wafers into a column with metal gaskets, fusing them, cutting the column into structures, forming external current-conducting contacts and applying a dielectric antireflective coating to the light-receiving surface.

The disadvantages of the method include the technology of formation of diffusion transitions, which leads to a decrease in the lifetime of minority charge carriers in single-crystal silicon and, accordingly, to a decrease in the efficiency of the photoconverter. In addition, a complex system of current collector contacts with interdigital geometry requires the use of expensive photolithographic processes.

The efficiency of solar energy conversion by semiconductor phase transitions decisively depends on the lifetime of minority charge carriers (NEC) in each of the areas of semiconductor wafers. The elimination and weakening of the factors limiting the lifetime of the NCD in semiconductor phase transitions leads to an increase in their efficiency.

The technical result of the invention is to increase the efficiency of the FP by increasing the life of the oil refinery, while reducing the cost of manufacturing technology due to its simplification.

This goal is achieved by the fact that in a semiconductor FP, consisting of single-crystal silicon wafers with vertically arranged pn diffusion transitions interconnected into a single horizontal structure by metal spacers, with current-conducting contacts, with at least one light-receiving surface with an inversion layer and a dielectric antireflection coating, in the region of p-type conductivity adjacent to the diffusion pn junction, parallel to the light-receiving surface, is located, m nshey least one nanocluster region of n-type conductivity in the region of the inversion layer is equal to or less than the sum of the thicknesses of space charge regions of the inversion layer and nanocluster area.

In the method of manufacturing a semiconductor photoconverter, including the formation of diffusion pn junctions on the surface of single-crystal silicon wafers, metallization of the wafer surface, assembly of wafers into a column with metal spacers, their fusion, cutting the column into structures, the formation of external current-conducting contacts, the application of a dielectric antireflective coating on light receiving surface, after applying a dielectric antireflection coating perpendicular to the light receiving surface hydrogen ions are implanted with a dose of 0.5–10 [μCoul] with an energy of 20–200 [keV] to create an n-type nanocluster region of conductivity, then thermal annealing is performed at a temperature of 300–500 [° C], after which they are irradiated with gamma rays with energy from 20 [keV] to 500 [keV] dose of 10 ⁵ ÷ 10 ⁷ [rad] to create an inversion layer by forming an integrated charge in the dielectric antireflective coating, and conduct thermal annealing at a temperature of 100 ÷ 250 [° C].

The distinguishing features of this decision include:

1) in the region of p-type conductivity adjacent to the diffusion pn junction, at least one nanocluster region of n-type conductivity is located parallel to the light receiving surface;

2) the nanocluster region is located at a distance from the inversion layer equal to or less than the sum of the thicknesses of the space charge regions of the inversion layer and the nanocluster region;

3) to create a nanocluster region perpendicular to the light receiving surface, hydrogen ions are implanted with a dose of 0.5 ÷ 10 μCul with an energy of 20 ÷ 200 keV;

4) after implantation, thermal annealing is carried out at a temperature of 300 ÷ 500 ° C;

5) to create an inversion layer, they are irradiated with gamma rays with energies from 20 keV to 500 keV with a dose of 10 ⁵ ÷ 10 ⁷ rad a light-emitting light-coating for forming an integrated charge in it;

6) conduct thermal annealing at a temperature of 100 ÷ 250 ° C.

Known technical solutions with such signs were not found.

The proposed photoconverter has a much larger area of all pn junctions (and, accordingly, the SCR volume) that collect light quanta.

The n-type nanocluster region of conductivity provides additional collection of charge carriers generated by solar radiation by increasing the SCR of this region, which, in comparison with the vertical n-layer adjacent to the diffusion pn junction, has the advantage of the absorption efficiency of the long-wavelength spectrum of solar radiation and separation charge carriers in the SCR volume due to the lower rate of volume recombination of light-generated charge carriers in comparison with the rate of volume recombination in di heavily doped fusional regions, as well as due to the location of this layer along the light receiving surface, which increases the number of pn junctions capable of absorbing solar radiation by charge carriers.

Thus, in the region of the inversion layer, the lifetime of the NSC is much longer than in the vertical diffusion n-layer, which makes it possible to reduce the recombination rate of charge carriers generated by light and to receive photons in the ultraviolet spectrum of solar radiation.

The use of this design allows to increase the efficiency of the semiconductor photoconverter up to 25-30%.

The invention is illustrated by drawings.

Figure 1 shows the design of the photoconverter with a nanocluster region on one side of the light receiving surface.

Figure 2 shows an enlarged section of one part of the structure.

Figure 3 shows an example of the construction of a photoconverter with nanocluster regions located on both sides (symmetrically) relative to the light receiving surfaces.

The semiconductor photoconverter (figure 1) consists of single-crystal silicon wafers containing regions of p- and n-type conductivity, forming diffusion pn junctions 1 and isotypic pn ⁺ junctions 2. Silicon wafers are interconnected into a single horizontal design metal gaskets 3, which are aluminum foil. External current-conducting contacts 4 are located on the lateral surfaces of the phase transitions. The p-type conduction region adjacent to the diffusion transition 1 (Figure 2) consists of two parts with different impurity concentrations: the p-region with a lower impurity concentration 5 and the p ⁺ region 6 with a higher concentration of impurities. The p-region 5 borders on the n-type diffusion region 7 and forms a diffusion transition 1 with it. The phase transition has at least one light-receiving surface, which is coated with a dielectric light-clarifying coating 8. In addition, a nanocluster is located in the depth of the p-region parallel to the light-receiving surface. an n-type region of conductivity 9, and an inversion layer 10 is formed under the dielectric antireflective coating.

In sunlight illumination of the light-receiving surface, light is absorbed by semiconductor layers, where electron-hole pairs are generated throughout their thickness. The most efficient separation of the generated electron – hole pairs occurs in SCRs formed by vertical pn junctions 1, an inversion layer 10 near the light receiving surface, and a nanocluster layer 9, which is located at a distance from the inversion layer equal to or less than the sum of the thicknesses of the space charge regions of the inversion layer and nanocluster region.

An SCR is formed on both sides of the nanocluster region. In this case, the SCR, which propagates towards the light-receiving surface, merges with the SCR of the upper horizontal transition of the inversion layer, forming a single space charge region that effectively collects charge carriers generated by both the short-wave and long-wave solar radiation spectra. The collected charge carriers are pulled from a single region (SCR) towards external current-conducting contacts, mainly due to the drift rather than the diffusion mechanism (i.e., much faster), without having time to recombine, which leads to an increase in efficiency.

With a deeper occurrence of the nanocluster region, the SCR does not close the inversion of the inversion and nanocluster layers, and in the p-region 5 (quasineutral) the diffusion mechanism of charge carrier collection works, which leads to a decrease in efficiency.

Compared with the known constructions, this phase transition ensures efficient collection of generated electrons and hole pairs from a larger SCR volume and their output to external current-conducting contacts, thereby providing a high carrier collection coefficient over the entire range of photoactive absorption of radiation from ultraviolet to infrared.

One of the main features of doping silicon with radiation defects by irradiation with light hydrogen ions compared with, for example, electron irradiation is the localization of the radiation exposure in depth. This locality is determined by the choice of the energy of the implanted ions, and the nature of the distribution of the defect concentration in the semiconductor volume is determined by the nature of the energy loss of the ion when it interacts with the crystal lattice (V.A. Kozlov, V.V. Kozlovsky. “Doping of semiconductors with radiation defects when irradiated with protons and α -particles ”, Physics and Technology of Semiconductors, 2001, Volume 35, Issue 7, pp. 769-795; Kozlovsky VV“ Modification of Semiconductors with Proton Beams ”, St. Petersburg, Nauka, 2003, p.268 [6, 7] )

The appropriate irradiation energies are chosen to provide the required mean free paths of hydrogen ions (H ⁺ ) in silicon, which are given by the dependence of the mean free paths of H ⁺ ions and α-particles (He ⁺⁺ ) in silicon (Fig. 4) [6]. The path lengths of hydrogen ions can range from fractions of a micrometer to tens of microns. At the end of the run, a hydrogen ion captures an electron, turns into a hydrogen atom, and its cross section for interaction with the atoms of the crystal lattice sharply increases (from 10 ^-19 to 10 ^-18 cm ^-2 ) (Ladygin EA "Radiation technology of solid-state electronic devices", M .: Central Research Institute "Electronics", 1976 - 316 p. [8]). As a result, a narrow nanocluster region is formed at the depth of the maximum ion path, which plays the role of a concentrator of electron-hole pairs generated by solar radiation (the interval of irradiation energies is shown in Fig. 4 by vertical lines and a horizontal arrow).

FP can be made using the following technological operations.

For this:

- initially form p-type conductivity in single-crystal silicon wafers, for example, KDB-10 or KDB-1, pn junctions 1 by diffusion of phosphorus and boron;

- metallize the surface of the plates with aluminum;

- assemble the plates in a column with metal gaskets 3 made of aluminum foil or silumin;

- after which they are fused, and then the column is cut into separate structures;

- on the lateral surfaces form external current-carrying contacts 4;

- silicon oxide (SiO ₂ ) and an antireflection coating (for example, Si ₃ N ₄ ) are applied to the light receiving surface;

- for the formation of an n-type nanocluster region of conductivity 9, hydrogen ions are implanted perpendicular to the light receiving surface, for example, on a Vesuvius-5 installation with a dose of 0.5–10 μCul, with an energy of 20–200 keV. When hydrogen is implanted with an energy of less than 20 keV and doses of less than 0.5 μCul, it is not possible to penetrate it to the required depth (from units to tens of microns) and to provide the necessary concentration of nanoclusters 10 ¹⁵ ÷ 10 ¹⁷ [nanoclusters / cm] to obtain additional SCR. The resulting concentration of nanoclusters should exceed the concentration of the acceptor impurity in standard types of substrates, where, for example, it is determined by the type of the initial silicon KDB-10 or KDB-1. When using energy above 200 keV and doses above 10 μCool, the processes of nanocluster formation become poorly controlled, and the technology is costly;

- conduct low-temperature annealing at a temperature of T = 300 ÷ 500 ° C in a thermal furnace for 15 minutes to 1 hour to form secondary radiation defects (V.V. Kozlovsky “Modification of semiconductors by proton beams”, St. Petersburg, Nauka, 2003, p. 268 [7]). Outside the specified range, annealing either passes inefficiently (at low temperatures, the activation energy is not high enough), or leads to the appearance of unnecessary, uncontrolled defects (at high temperatures);

- irradiation with a flux of γ-quanta with energies from 20 keV to 500 keV is carried out with an exposure dose (10 ⁵ ÷ 10 ⁷ ) rad, for example, in the MPX-60 installation, to form a positive built-in charge in the dielectric antireflective coating. In this case, the effective density of the built-in charge Nss [cm ^-2 ], which takes into account both surface and space charges, should be Nss> 5 · 10 ^-11 cm ^-2 , because only at high concentrations is it possible to obtain an inversion of conductivity.

When a dielectric light-illuminating coating is irradiated with a flux of γ-quanta with an energy of less than 20 keV and a dose of less than 10 ⁵ rad, the process of forming the built-in charge is inefficient - the charge will be insufficient for the formation of the inversion layer. Irradiation with a flux of γ-quanta with an energy of more than 500 keV and a dose of more than 10 ⁷ cannot be realized on existing sources of γ-quanta; moreover, large energies will create harmful defects;

- conduct low-temperature annealing at T = 100 ÷ 250 ° C in a thermal furnace for 1-3 hours to fix the built-in charge and to eliminate the resulting unstable defects in the dielectric light-coating.

Figure 3 shows an example of the design of the phase transition when the dielectric light-coating coating and the nanocluster layer are formed symmetrically on both sides of the original semiconductor structure. In this case, it becomes possible to efficiently collect charge carriers in two-way lighting, because this ensures the exclusion of parasitic shunting of the non-illuminated part of the FI, in comparison with the case of unilateral illumination.

Thus, the proposed phase transition has a much larger area of pn junctions separating the electron – hole pairs generated by solar radiation, the lifetime of the NEC in the region of the inversion layer is much longer, which makes it possible to reduce the rate of recombination of electrons with holes and more efficiently convert the photons of the solar UV spectrum radiation. All this leads to an increase in the efficiency of the photoconverter.

In such phase transitions, in comparison with conventional phase transitions at planar pn junctions, better cooling due to metal gaskets is provided. The proposed design promotes the free passage of non-photoactive infrared radiation through almost the entire phase transition without absorption, which leads to a lower equilibrium temperature of the phase transition and, consequently, to an increase in efficiency.

The reduction in the cost of manufacturing technology of phase transitions occurs due to the use of semiconductors modified with hydrogen ions. This is due to the simple and relatively low cost of radiation processes in combination with high accuracy, performance, reproducibility and good compatibility with other technological processes for the manufacture of semiconductor devices. Photolithography is excluded in the FP manufacturing technology, since external current-carrying contacts are formed during fusion of the column on its end surfaces and do not require additional operations, as is necessary when forming contacts using traditional technology.

The use of such phase transitions with light concentrators is especially effective, since in this design there are no shading front (front) contacts, and two-sided lighting provides a more complete use of semiconductor material.

The use of such phase transitions is also possible in the conversion of laser radiation, for example, in energy transfer systems using laser beams.

LITERATURE

1. US patent No. 4320250, CL. H01L 31/0224, publ. 03/16/1982

2. RF patent No. 2127009, cl. H01L 31/18, publ. 02/27/99, figure 1-3.

3. RF patent №2127472, cl. H01L 31/18, publ. 03/10/1999, figure 1, 2 (prototype).

4. Sater B.L. at all "The multiple Junction Edge Illuminated Solar Cells" in Conf. Rec. Tenth IEEE Photovoltaic Specialists Conf., 1973, p. 188-193.

5. US patent No. 4516314, CL. H01L 31/18, publ. 05/14/1985

6. Kozlov V.A., Kozlovsky V.V. "Doping of semiconductors with radiation defects upon exposure to protons and α particles." Physics and Technology of Semiconductors, 2001, Volume 35, Issue 7, pp. 769-795.

7. Kozlovsky V.V. “Modification of semiconductors by proton beams”, St. Petersburg, Nauka, 2003, p.268.

8. Ladygin EA "Radiation technology of solid state electronic devices." - M.: Central Research Institute "Electronics", 1976. - 316 p.

Claims (2)

1. A semiconductor photoconverter, consisting of single-crystal silicon wafers with vertically arranged pn diffusion transitions interconnected into a single horizontal structure by metal spacers, with current-conducting contacts, with at least one light-receiving surface with an inversion layer and a dielectric antireflective coating, characterized in that in the region of the p-type conductivity adjacent to the diffusion pn junction, parallel to the light receiving surface is located, along enshey least one nanocluster region of n-type conductivity in the region of the inversion layer is equal to or less than the sum of the thicknesses of space charge regions of the inversion layer and nanocluster area. 2. A method of manufacturing a semiconductor photoconverter, including the formation of diffusion pn junctions on the surface of single-crystal silicon wafers, metallization of the wafer surface, assembly of wafers into a column with metal spacers, their alloying, cutting the column into structures, the formation of external current-conducting contacts, the application of a dielectric antireflective coating on a light receiving surface, characterized in that after applying the antireflection coating perpendicular to the light receiving surface They implant hydrogen ions with a dose of 0.5 ÷ 10 μCool with an energy of 20 ÷ 200 keV to create an n-type nanocluster region of conductivity, then conduct thermal annealing at a temperature of 300 ÷ 500 ° C, after which they are irradiated with gamma rays with energies from 20 to 500 keV dose of 10 ⁵ ÷ 10 ⁷ rad to create an inverse layer by forming a built-in charge in the dielectric antireflective coating and conduct thermal annealing at a temperature of 100 ÷ 250 ° C.

Images