Thin film cds/cdte photovoltaic cells and their manufacture

PROCESS FOR ENHANCING THE CONVERSION EFFICIENCY OF THIN FILM CdS/CdTe PHOTOVOLTAIC'CELLS This invention relates to thin-film p-n 5 cadmium telluride-cadmium sulfide photovoltaic cells. In Bonnet, New Results on the Development of a Thin-Film p-CdTe-n-CdS Heterojunction Solar Cell, 9th IEEE Photovoltaic Specialist Conference, P. 129 (1972), there is reported a thin-film cell of CdS/ 10 CdTe in the polycrystalline condition. The conversion efficiency of this cell was reported by Bonnet to be 5^. However, a review of his calculations reveals that the conversion efficiency of this cell was actually less than the reported value of 5%.

Photovoltaic cells having polycrystalline cadmium sulphide and׳ polycrystalline cadmium telluride in contiguous layers allow the inexpensive layered formation of a truly thin cell, Inasmuch as the combined thickness of two layers of polycrystalline semi20 conductor need not exceed 100 microns, and preferably does not exceed 10 microns. Such thin-film CdS/CdTe photovoltaic cells are presently limited in their conversion efficiency to less than 5$.

The present Invention is based upon the dis25 covery that the conversion efficiency of thin-film, polycrystalline CdS/CdTe cells can be significantly enhanced if either (or both) of the layers of CdS and CdTe is deposited under conditions such that at least a small amount of oxygen is present in the atmosphere 30 through which the CdS or CdTe passes while its respective layer is being deposited. The present process involves a close-space sublimation technique in the steps that involve deposition of the CdS and CdTe layers. Thin-film polycrystalline CdS/CdTe cells having 35 conversion efficiencies of at least 6% can be manufactured using this process.

Fig, 1 is a plot of conversion efficiencies achieved by cells of the invention, versus the substrate temperature used during the formation of the respective CdTe layers.

Fig. 2 Is a partially schematic section view of a cell prepared in accordance with the invention.

Enhancement of the conversion efficiency of a photovoltaic cell having polycrystalline CdS and polycrystalline CdTe in contiguous layers is achieved by forming one or both of these layers in an oxygencontaining atmosphere. Maximum enhancement occurs when the layers are formed within a preferred range of temperatures in the presence of gaseous oxygen.

As used herein, enhanced conversion efficlency means a conversion efficiency that is more than 10% greater than the conversion efficiency of an otherwise identical control cell produced in an oxygenfree atmosphere during the formation of the CdS and/or the CdTe layer(s).

The process of this invention involves forming the CdS and/or the CdTe polycrystalline layers using an atmosphere which contains oxygen. In my process, the layers of CdS and CdTe are deposited in the vapour phase. As used herein, close-space sublimation means sublimation from a source to a substrate positioned from the source a distance no greater than the square root of the smaller of the surface areas of the source or of the substrate. Thus, the.manipulative steps of the process of this invention are known. What was not known heretofore is the use of oxygen in the atmosphere through which CdS and/or CdTe passes during the step which involves the vapour'deposition of polycrystalline layers of such photoconductors.

During the vapour deposition step, the atmosphere through which CdS or CdTe passes from the source to the substrate can be either pure oxygen or oxygen admixed with other gases, such as air.

In the process of this invention, CdS is preferably deposited from the vapour phase on a suitable electrode, described below, with or without oxygen present in the atmosphere. Thereafter, cadmium telluride and, optionally, a p-type dopant such as gold are similarly deposited in the vapour phase on the previously formed layer of cadmium sulphide, with or without oxygen present in the atmosphere, except that at least one of these two layers is deposited in the presence of oxygen gas.

Thereafter, an electrode capable of making a low impedance contact is formed on at least a portion of the layer of cadmium telluride. A preferred process for the formation of this electrode is vacuum deposition.

The vapour-phase deposition of the CdS and CdTe can'be done either as a batch process, e.g., in a vacuum chamber containing a single source and a single substrate, or as a continuous process in which a substrate is moved through appropriate zones for treatment.

Close-space sublimation is a highly desirable form of vapour-phase deposition in the practice of this invention. The oxygen in the atmosphere during the sublimation can be at a partial pressure between 0.01 and 100 torr, and preferably is at a partial pressure between 0.1 and 10 torr. The spacing between source and substrate preferably is between 0.5 and 5 mm. The temperature of the substrate upon which each of the semiconductor layers is to be deposited can be varied, depending on which material is being sublimed. Preferably, the deposition of each of the CdS and CdTe is conducted from 0.1 sec to 10 minutes onto a substrate held at a temperature of between 300°C and 65O°C. The source temperature is maintained in each instance between 10QC and 500°C (preferably 100-300°C) higher than the substrate temperature during the CdS deposition and preferably 20-50 C higher than the substrate during the CdTe deposition.

When close-space sublimation is used in the practice of this invention, higher conversion effi5 ciencies are achieved when the CdTe layer is sublimed in an oxygen-containing atmosphere on' a CdS substrate at a temperature of at least 575°C, the CdS layer also having been sublimed in an atmosphere that contained oxygen.

One effect of deposition temperature is illustrated in Fig. 1. The CdS substrate in this case was also sublimed in an atmosphere that contained 0.4 torr oxygen. Each data point in Fig. 1 represents a separate cell wherein the only difference in the preparation of the cell was the temperature of the CdS substrate during the CdTe deposition. Whereas the use of a substrate temperature of 500°C resulted in a conversion efficiency of 3.9%, increasing the temperature of the substrate to 550°C raised the con20 version efficiency to 6.3%, while increasing the temperature to 570°C raised it still further, to 8.4%. A further increase in temperature of the CdS substrate to 600°C resulted in a further Increase in efficiency to 8.9%.

It is possible that the temperature of the

CdS substrate above which further substantial increases in conversion efficiency are apparently not achievable can be lowered below 575°C if other variables of the sublimation process are further optimized.

Further enhancement of the conversion efficiency of the CdS/CdTe cells can. be achieved by an optional process step of heating the formed CdTe layer in an oxygen-containing atmosphere for a suitable length of time. Oxygen under ordinary or high pressure can be used, such as air at ambient pressure. Partial pressures of oxygen can also be

-לused in this optional step, as can an alternating non-oxidizing and oxidizing atmosphere of the type described in U.S. Patent No. 4,035,197 . The temperature of such post-treatment can be between 100°C and 600°C, for a time of between 1 second and 20 minutes.

Such post-heating of the CdTe layer can increase the conversion efficiency by 1 or 2$ over that achieved by the same cell lacking the post10 heating step.

It is believed that the vapour deposition of one or both of the semiconductor materials In an oxygen-containing atmosphere produces enhanced conversion efficiencies by the Incorporation of oxygen 15 atoms into the semiconductor layer. The exact amounts of such oxygen Incorporation are so minute that they are difficult to detect or assay.

To complete the cell, electrodes are selected to provide low-impedance contact with at 20 least portions of the exterior surface of each of the two contiguous semiconductor layers. As used herein, low impedance means an impedance that is no greater than that which gives a contact resistance of 50 ohm2 cm . Although such low-impedance contacts at least 25 approach a true ohmic contact, some rectification in the contact can also be tolerated.

Each of the electrodes is in operative contact with its respective layer; that is, in the cell it contacts only its layer. At least one of the 30 electrodes is sufficiently transparent to transmit sufficient radiation to photogenerate an electric potential in the cell. Conveniently, such transmission of radiation is accomplished by rendering one of the electrodes transparent to at least some radiation 35 having wavelengths between 350 and 900 nm;

A useful material having this property is Nesatron brand glass, available from PPG Industries, Inc. ”Nesatron is the Trade Mark for a transparent glass electrode comprising a thin layer of conductive 1^0^(Sn) coated on transparent glass. Such an Electrode typically has a sheet ל resistance of 10 to 50 ohm/square and an optical transmittance of 80$ for visible light. The Nesatron electrode provides a low-impedance contact with the n-type cadmium sulphide, and thereby provides a highly useful substrate for the first semlconductlve layer.

Other conventional electrodes also can be selected, providing they have sufficient conductivity, light transmittance, and the capability of forming the desired low-impedance contact with n-type CdS. Examples include transparent layers of CdO, Sn02.abd

Cd2SnOi|, on glass, with or'without do- pants.

Electrodes for low-impedance contact with the p-type CdTe can be, for example, gold, silver, copper, nickel, tin, platinum, palladium, chromium, 20 iridium, rhodium, and non-metalllc materials such as Cu2S and Cu2Te having high electrical conductance and a work function approximately matching that of p-CdTe.

Fig. 2 illustrates a typical photovoltaic cell 10 produced in accordance with the process of 25 the invention. Transparent electrode 12 comprises a support layer 14 of glass bearing a conductive layer 16 of In2O3(Sn), on which is formed layer 18 of ntype CdS. A contiguous layer 20 of p-type CdTe is formed on the layer 18, and an electrode layer 22 con30 tacts at least a portion of the.layer 20. > Leads 24 attached by solder 26 carry away power from the cell. Because the layers 18 and-.20 are polycrystalline, they can be prepared so that they are extremely thin, for example, no greater than 50 microns thick and 35 preferably from 0.05 to 5 microns thick. The total thickness then of the combined semiconductor layers

-ר18 and 20 Is generally less than 100 microns, and preferably no greater than 10 microns. The electrode layers 16 .׳and 22 are extremely thin being from 100 to 1000 nm and from 3 to

100 nm, respectively. The glass support layer 14 is typically 1000 microns thick.

Photovoltaic cells of this invention can be used to photogenerate electric power; e.g., when exposed to activating radiation. In such use, the 10 potential created across the Junction between the CdS and CdTe layers can be used as a voltage source, or the current generated by the cell can be drawn off. Thus, among other things, the cell can be used as a voltage generator or as an exposure meter 15 operating from the cell’s short circuit current.

The following examples further illustrate the invention.־ Except־where noted,־ pure oxygen atmospheres at stated partial pressures were used. In each example simulated sunlight illumination of 75 2 mW/cm was achieved by using a Kodak ’Carousel’ projector, Model 600, with the heat-absorbing glass removed, and containing a 500 W tungsten lamp, and a spectrum-correcting filter. ’Kodak’ and ’Carousel’ are Trade Marks. Conversion efficiencies were calculated from the measured I , V , fill factor and the known input illumination. As used herein, ”fill factor” means the fraction of the product of I and V_n that is available as power output for the cell. In these examples, the space between the source and the substrate during the deposition of the CdS or CdTe layer was approximately 2 mm.

Example 1

A cell was prepared of the type shown in Fig. 2, as follows:

On a 1 mm thick Nesatron glass substrate a 0.2 micron thick layer of CdS was deposited at a temperature of 550°C, in an oxygen-containing at-8mosphere of 0.4 torr oxygen. The temperature of the CdS source was 700-720°C. A 2 microns thick layer of CdTe was then deposited on the CdS layer while it was heated at a temperature of 59O°C, in an at5 mosphere containing 1.5 torr oxygen and 1.5 torr Argon The temperature of the CdTe source was 65O°C. A 500 nm thick gold electrode layer 6 was vacuumdeposited on the CdTe layer. Under simulated sunp light illumination of 75 mW/Cm , the cell had a V״ 0 ך, c of 0.7 volts, an I of 16 mA/cm , and a fill factor SC of 0.6. The conversion efficiency was'8.9%.

Example 2: Enhancement When Oxygen Ts Present Only During Deposition of the CdS Layer

To demonstrate the effect of the use of an 15 oxygen-containing atmosphere only during the deposition of the CdS layer, a control cell and 5 test cells were prepared as follows: Nesatron substrates having a nominal resistivity of 25 ohm/square were cleaned. The CdS layer was then deposited by close20 space sublimation, wherein the source was a hotpressed disc of phosphor-grade CdS powder obtained from the General Electric Company. The process was carried out with a source temperature of 750°C and a substrate temperature of 535°C. In the case of 25 the control cell, a vacuum (less than 0.01 torr oxygen was used. The five test cells, on the other hand, incorporated pure oxygen at a pressure of 2.5 torr while the CdS was being sublimed. The time for CdS deposition was 2 to 4 minutes, leading to a layer of 30 CdS between 1 micron and 2 microns thick. After cooling the substrate to near room temperature, it was placed on another holder for the deposition of CdTe. The source was a polycrystalline disc sliced from an ingot which had been prepared by reacting 35 99,9999% pure Cd and Te in a melt doped with 10 ppm of Au. A source temperature of 550°C, a substrate temperature of 45O°C and a deposition time of 2 minutes were used to deposit a layer of CdTe 0.5 microns thick. The deposition of the CdTe layer was carried out in a vacuum in the essential absence of oxygen. The CdS/CdTe resulting structure was then heated in air at 325°C for 3 minutes, after which the cell was completed by vapor-depositing a 500 A thick layer of Au on the layer of CdTe.

The photovoltaic response of the control cell (wherein both the CdS layer and the CdTe layer had been deposited in the absence of oxygen) was V = 0.37, Ieo = 7.1 mA/cm2, fill factor of 0.4 0 c sc and conversion efficiency = 1.38$. In contrast, the average photovoltaic response of the 5 test cells

כ׳ was V = 0.48 + 0.03 mV, I = 9-1 + 1-6 mA/cnT, oc - sc fill factor of 0., and conversion efficiency = 2.5 + 0r3?> an improvement inconversion efficiency of more than 1$.

Examples 3 and 3A:

To further illustrate the effect of the presence or absence of oxygen during the deposition of the layer of CdS, two cells were made with the layer of CdS deposited in the presence of air at two different pressures. One of the cells, Example 3, was prepared by subliming the CdS layer in a vacuum of less than 0.01 torr. Sublimation was accomplished with the source and the substrate separated by 2 mm. The layer of CdTe was formed by sublimation for 2 min. onto the CdS substrate which was maintained at a temperature of 610°C, in an atmosphere of 1.5 torr oxygen and 1.5 torr Ar. The temperature of the source was 630°iC. Example 3A was prepared identically as Example 3, except that air, at a pressure of 0.15 torr, was used as the atmosphere for the sublimation of the CdS. Table 1 Illustrates the results.

Table 1

Air Pressure, During

Deposition of the CdS

<td> Example</td><td> Layer, Torr</td><td></td><td> V oc ,mV</td><td>¹sc, ₂mA/cnr</td><td> FF</td>

<td> 3</td><td> essentially zero</td><td> 4.0</td><td> 390</td><td> 15.1</td><td> 0.51</td>

<td> 3A</td><td> 0.15</td><td> 8.3</td><td> 650</td><td> 17.0</td><td> 0.56</td>

Very little improvement in conversion efficiency was noted upon increasing the air pressure above 0.15 torr. This example demonstrated that although the presence of oxygen in the atmosphere for the deposition of the layer of CdTe produced a significant enhancement of conversion efficiency, an even higher value is achieved if oxygen is also present for deposition of the layer of CdS.

Example 4: Enhancement When Oxygen is Present Only During Deposition of the CdTe Layer

To demonstrate the enhancement achievable by depositing only the layer of CdTe in the presence of oxygen gas, eight cells were prepared by the procedure of Example 2 on layers of CdS prepared as in Example 2. Then the layer of CdTe was deposited using an atmosphere having a partial pressure of oxygen equal to 5.0, torr. These eight cells showed an average VQC of 0.6182± 0.012 volts, an average Igc of 11.5 ±

1.5 mA/cm^, a fill factor of approximately 0.5 + 0.01, and an average conversion efficiency of 4.89 + 0.54$. Thus, an improvement of more than 2% in conversion efficlency was achieved by depositing the CdTe in the presence of an atmosphere which contained oxygen. Examples 5-11

To demonstrate the effect of the substrate temperature for the deposition of the CdTe, seven cells were made under the following conditions: a thin layer of CdS was deposited by close^space sublimation on a Nesatron glass substrate having a nominal electrical resistance of 10 ohm/square. The temperature of the source was 720°C and the temperature of the substrate was 550°C. The time for CdS deposition was 5 seconds. The CdS deposition was carried out in oxygen at a pressure of 0.4 torr. Then a layer of CdTe was deposited by the close-space sublimation procedure. The temperature of the source was 630°C to 640°C. The temperature of the substrate was varied as shown in Table II. The time for CdTe deposition was 2 minutes. The CdTe deposition was carried out in an atmosphere consisting essentially of

1.5 torr 02 and 1.5 torr Ar. Cell manufacture was completed by deposition on the layer of CdTe of a 5002 thick layer of gold, using .the technique of conventional vacuum evaporation under a pressure of 10 torr. (For the cells with 600°C or higher substrate temperature, a slightly higher source temperature was used to provide a layer of CdTe having a thickness similar to that of the other cells.) The results are summarized in Table II, and represent the data points in Fig. 1.

Table CdTe Deposition

<td> Example</td><td> Source' Temp., °C</td><td> Substrate Temp,, °C</td><td> &lt;£</td><td> V V OCjHlV</td><td> ^sc</td><td> FF</td>

<td> 5</td><td> 630</td><td> 500</td><td> 3.9</td><td> 560</td><td> 12.5</td><td> 0.42</td>

<td> 6</td><td> 630</td><td> 530</td><td> 1.1</td><td> 610</td><td> 13.7</td><td> 0.83</td>

<td> ז</td><td> 63Ο '</td><td> 550</td><td> 6.3</td><td> 630</td><td> 18.7</td><td> 0.50</td>

<td> 8</td><td> 630</td><td> 570</td><td> 8.4</td><td> 680</td><td> 16.7</td><td> 0.55</td>

<td> 9*</td><td> 630</td><td> 590</td><td> 8.9</td><td> 705</td><td> 16.3</td><td> 0.58</td>

<td> 10</td><td> 6Ί0</td><td> 600</td><td> 8.9</td><td> 695</td><td> 16.7</td><td> 0.58</td>

<td colspan="4"> 11 6ήί| 610 8.9 ^,This is the cell described In Example 1.</td><td> 710</td><td> 16.5</td><td> 0.58 1 H N 1</td>

Example 12:

To further illustrate the effect of the pres-erce of oxygen during deposition of a layer of CdTe, a series of cells was prepared by varying the oxygen pressure during the deposition of CdTe while the other process conditions were kept essentially constant. That is, the deposition of the CdS was carried out in an atmosphere having 0.2J, 0.5 or 1.0 torr oxygen, using a T (source temperature) of sourC Θ 710, 720 or 725°C and a T. (substrate tempsuDSuraL.© erature) of 500°C or 550°C. The CdTe deposition used a TC1,&amp;#1524;T,&amp;#1524;O of 630 or 640°C and a T . of 600 or source substrate

610uC. None of these minor variations are believed to have a significant effect on the conversion efficiencies of the resulting CdS/CdTe cells.

It was found that the cell produced in the presence of oxygen at a partial pressure of 0.5 torr had a conversion efficiency of 4.6%, whereas one produced in the presence of oxygen at a partial pressure of about 1.0 torr or higher, the conversion efficiency was generally constant at 8.75% + 1.25.