GB1563945A

GB1563945A - Cadmium sulphide-indium phosphide light responsive devices

Info

Publication number: GB1563945A
Application number: GB13755/77A
Authority: GB
Original assignee: Western Electric Co Inc
Current assignee: AT&T Corp
Priority date: 1976-04-02
Filing date: 1977-04-01
Publication date: 1980-04-02
Also published as: BE853029A; IT1076266B; IL51751A0; FR2346859A1; JPS52120786A; ES457439A1; NL7703585A; DE2714221A1

Description

(54) CADMIUM SULPHIDE-INDIUM PHOSPHIDE LIGHT RESPONSIVE DEVICES (71) We, WESTERN ELECTRIC COM- PANY, INCORPORATED, 222 (formerly of 195) Broadway, New York City, New York State, United States of America, a Corporation organised and existing under the laws of the State of New York, United States of America, do hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following state ment - The invention relates to cadmium sulphide-indium phosphide light responsive devices.

In Patent Specification No. 1526 823 a hetero-junction diode is fabricated by the deposition of n-type cadium sulphide on ptype indium phosphide. While such a cell in a single crystal form has quantum efficiencies of 70 percent in the spectral region between 550 and 910 nm and a solar energy conversion efficiency of between 8.5 and 10.3 percent, its single crystal structure places an upper limit on its physical size and hence the uses to which it can be economically put are correspondingly limited.

For a cell with larger area a polycrystalline structure is preferred. Implicit in the use of polycrystalline photovoltaic cells and their resulting advantageous characteristics, is the need for an efficient method of making polycrystalline layers of doped indium phosphide.

Although such manufacture can be accomplished through a variety of techniques, e.g.

conventional evaporation methods such as molecular beam epitaxy or flash evaporation or by CVD, we have developed a chemical vapour deposition method which is particularly suited for production of polycrystalline p-type indium phosphide.

According to the-present invention there is provided a method of making a lightresponsive device having a hetero-junction between a polycrystalline p-type indium phosphide layer and a light-transmissive polycrystalline n-type cadmium sulphide layer and an electrode on the cadmium sulphide layer arranged to allow light to fall on the cadmium sulphide layer, wherein the indium phosphide layer is grown by a chemical vapor deposition process using a vapour which provides phosphorus molecules and hydrogen chloride in a gaseous carrier which vapour is passed over a heated indium source.

Exemplary devices of polycrystalline solar cells with quantum efficiency of 70% and solar energy conversion efficiency of 2.5 have been made by our method.

Some embodiments of the invention will now be described by way of example and with reference to the accompanying drawings, in which: Fig. 1 is arl illustration of a chemical vapour deposition apparatus useful in producing doped indium phosphide films by a method according to the invention; Fig. 2 is an illustration of a modified chemical vapour apparatus useful in producing doped indium phosphide films by a method according to the invention and Fig. 3 is an illustration of a polycrystalline photovoltaic cell.

Films of p-type indium phosphide can be deposited on various substrates by employing conventional techniques, e.g., evaporation methods (flash evaporation or molecular beam epitaxy) or by utilizing a chemical vapour deposition (CVD) process using phosphorus trichloride, or triethyl indium and phosphine as the source combounds. It hns been discovered that a specific CVD process employing PClS saturated hydrogen gas or phosphine and hydrogen chloride saturated hydrogen gas give films quite suitable for photovoltaic purposes.

Specifically, the apparatus shown in Fig. 1 can be used to manufacture p-type InP films for polycrystalline heterojunction devices. A stream of hydrogen gas which has been purified by diffusion through palladium is introduced into tube 19 and passes through a bubbler 20 containing phophorus trichloride held at a constant temperature typically in the range between 0 and 5"C. (A portion of the pure palladium diffused hydrogen also is diverted through 21 where it passes on the outside of tube 44). The hydrogen gas which is in the bubbler becomes saturated with phosphorous trichloride. The bubbler temperature, the flow rate of H2 through the bubbler, and the flow rate of H2 are adjusted so that the molar fraction of PC13 in the H2 is between 1 and 5%, preferably 2%. The PC13 saturated hydrogen gas is then flowed through a conduit 23 into quartz tube 24 which contains boats 25 and 26. The linear velocity of the hydrogen stream can vary over a large range of velocities between 10 and 50 cmlmin. preferably 20 cm/min., but higher flow rates are usable. The boats 25 and 26 are independently heated by furnaces 27 and 28 and contain respectively a dopant composition and elemental indium. The dopant compound can most advantageously be anhydrous cadmium chloride. Elemental cadmium if used in the configuration of Fig.

1 will be converted to CdC12 by the PCl2 saturated hydrogen gas. However, if elemental Cd or Zn is desired as a dopant, the apparatus of Fig 1 may be modified to achieve this result as shown in Fig. 2. This modification can be accomplished as shown in Fig. 2 by placing the dopant in an extra quartz tube 29 which is concentric to the main deposition tube and which ends shortly before the location of the In source boat 26.

Separate inlets 31 and 32 are provided for the two concentric quartz tubes. Only purified hydrogen gas is flowed through inlet 32 to the dopant. The PC13 saturated hydrogen is injected through inlet 31. The two gas flows are adjusted so that the PCl2 molar fraction is set as previously discussed.

The boat 25 containing preferably elemental cadmium is preheated and maintained between 450 and 550CC, preferably between 4700C and 480"C. If anhydrous cadmium chloride is used, the preferred temperature is 500"C and if elemental Zn is used, a range of 5000-6000C preferably 580"C is appropriate. The indium in boat 26 is preheated and maintained between 700 and 800"C, preferably 740"C. Once boats 25 and 26 have been preheated and a stead flow of PCl3 saturated hydrogen is established, the substrate 22 is pushed in the deposition area 33 in tube 44, using rod 34, where it is heated and maintained between 575"C and 7000. C1 preferably 630"C by an independently controlled furnace 35.

As the PCla saturated hydrogen flows over the indium containing boat, it reacts with tfie;:igClium forming a crust of InP on the surface of the indium. The further reaction of the PC12 with the InP crust causes indium monochloride to be carried on the PC13 saturated hydrogen flow downstream and is mixed with the H2 stream containing the dopant, then carried downstream where Cd doped InP is deposited on the substrate. The deposition process can be continued until the desired thickness of p-type InP is obtained. Since an indium phosphide crust is initially formed, it is also possible to use initially in the boat indium phosphide instead of elemental indium. In such a case the same conditions are used. The rate of deposition can be varied between 1 and 20 ,,um/hr. by controlling the temperature difference between the In source and the substrate. Layers between 1 and 25y have been produced.

Layers larger than 25 can be produced using this method by simply extending the deposition time. Since the absorption length of InP is lJIm, layers thinner than this are undesirable if best efficiency is to be obtained.

The carrier system of PC12 saturated hydrogen gas can also be replaced by a hydrogen-phosphine-hydrogen-chloride gas system (HPA system). When this system is used, the PC13 bubbler is replaced by a source of HC1 and phosphine gas such as a cylinder containing a mixture of HC1 gas in hydrogen and a cylinder containing a mixture of PH3 in hydrogen. The compositions of these gas mixtures are not critical but mole ratios of 10% of HC1 to H2 and 5% PH3 to H2 have been found convenient for controlling the flows of HCl and PH3 into tube 24, Fig. 1. The PC13/H2 mixture described previously reacts inside the heated tube 24, Fig. 1 to form HC1, and phosphorous molecules such as P2 and P4. The same chemical composition of the vapour phase is achieved when HC1, PH3 and H2. are injected instead of the PC13/H2 mixture. The process parameters can remain as described for the CVD process using PCI,/H,. The mole ratio of HC1 and PH3 to H2 is controlled to be equivalent to the ratios of PCl3 to H2 outlined previously. Thus, it can be seen that a carrier system containing both HC1 and -phosphorous molecules is the desired result.

The substrate is composed of materials whose coefficient of thermal expansion substantially matches that of InP and which do not react with InP at temperature below 650 cm, for example, "Corning 7052" glass (a tradename for a low alkali, barium silicate composition), "Nonex"glass -(a tradename for a hard borosilicate glass), carbon, molybdenum, and canary glass. Before use, the glass substrates are cleaned by polishing in a hydrogen flame. The molybdenum subtrates are cleaned by etching for 30 sec. in a mixture of 50 ml. concentrated nitric acid and 30 drops of hydrofluoric acid, followed by washing with distilled water and propanol. Both the carbon and molybdenum substrates receive a final step of heating at 1000"C in vacuum. No cracking or peeling was observed for films on the aforementioned substrate materials.

The acceptor concentration obtained by this CVD process is between 1016 and 102 cm~3. This is within a range which produces semi-conductor material which is quite satisfactory for use in photovoltaic cells. Also the average crystal size of the particles forming the layer was approximately 2 m. This is significant. If the grain size was substantially smaller it would closely approach the absorption length of ptype indium phosphide. Therefore, most charge carriers would necessarily have to cross at least one crystal boundary. This, in turn, would lead to excessive recombination and reduced efficiency. Nevertheless, by the practice of the process very thin layers having a 2pm or larger grain size can be produced.

This combination helps maximize efficiency while minimizing the amount of material which must be used.

Once the p-type InP layer 37 as shown in Fig. 3 has been formed on, for example, a carbon substrate 38, by the aforementioned CVD process the surface which is to be used as an interface surface is treated for 15 seconds in a cold bromine-methanol solution (0.1% by volume bromine) and then rinsed in pure methanol. Then a layer of n-type CdS 39 is deposited on the p-type InP layer 37.

The apparatus for CdS growth is a modification of an apparatus previously described by D. Beecham in Rev. Sci. Instrum., 41 1654 (1970). The device, as modified, has been described by S. Wagner in two articles Applied Physics Letters, 22, 351 (1973) and 1. Applied Physics, 45, 246 ,(1974). These publications basically disclose a coaxial isothermal double source for in vacuo growth of CdS from cadmium and sulphur. A single heating element, comprising a concentric nichrome or tantalum wire wound on a quartz tube, controls both the cadmium and sulphur temperature. The respective fluxes of cadmium and sulphur are controlled by separate effusion orifices. The source is positioned in an oil free vacuum system and is separated from the substrate holder by a mechanically operated stainless steel shutter.

After mounting the InP substrate holder in the apparatus the vacuum station used for the growth of CdS films 39 is closed and evacuated to a pressure < 1 X 1tS Torr.

The coaxial isothermal CdS source (not shown), which contains elemental cadmium and elemental sulphur, is then heated to 350"C. Simultaneously the InP substrate holder is heated to 200--250"C. During the heat-up, the stainless steel shutter remains closed and separates the source compartment from the substrate.

When source and substrate have leached the set temperature, the shutter is opened and a 5 to 10Fm thick n-CdS film 39 is grown on the p-InP layer 37. The rate of film growth is monitored with a quartz crystal oscillator. It is about 0.15 Fm/min.

When the desired film thickness has been reached the shutter is closed and both the source and the substrate heaters are turned off.

The thickness of the CdS layer is not critical and can be typically 1OJLm.

Athough it is not essential, the cell can be annealed in forming gass. Electrical contact to the InP side of the heterojunction can be made in many ways, for example, by attaching a copper tab 40 to the carbon substrate, Contact to the CdS side can also be done by various means, for example, by using conventional techniques to add a thin In or Ga-In alloy contact 41 to a portion of the exterior of the CdS material 39. Either a corner of the CdS surface can be covered or a grid over the entire CdS surface can be used. Typically the In or Ga-In alloy covers 0.1% of the CdS surface area. The cell as depicted in Fig. 3 can then be operated by directing the solar energy through the transmissive CdS layer and into the InP layer.

Since InP has a large absorption coefficient, charge carriers are formed near the interface.

This results in a quantum efficiency of 70% and solar conversion efficiency of 2.5% as measured using a solar light simulator as a source.

The above configuration, however, is only exemplary of the possible cell geometries.

Either a glass protective cover or an antireflection coating can be added on top of the In or Ga-In contact. Various substrates such as molybdenum or zincated metal can be employed instead of carbon. '(However, if zincated metal is used the CVI) method disclosed herein may require some additional precautions since the HCl formed in that method attacks the zinc).

A cell can also be built by depositing a polycrystalline layer of n-type CdS on a glass plate which has a conducting grid on its surface. This grid can be formed, for example, by simply brushing a thin layer of In on the surface of the glass. The CdS layer can be formed by techniques such as CVI) or by using the isothermal source described previously. A thin layer of p-type InP is then deposited on the CdS by the CVI) process. Being formed on a polycrystalline CdS layer it will itself be polycrystalline. The InP layer can be approximately 2q,um thick, while the CdS layer may be much thicker, e.g., 10 to 20cm. An opaque metallic electrical contact can then be attached to the exposed surface of the InP by standard techniques such as electroplating or evaporation.

It has been found that the thermal stability in air of the n-type CdS layer of the device is strongly affected by the net donor concentration in the layer. A net donor concentration in the n-type CdS of at least 5 X 10'8cam~3 produces an increase in thermal stability in air when compared to devices with net donor levels of approximately 2 X 10'8cm-" and lower.

For testing, devices with high net donor levels were made as described before using the isothermal source: However, to obtain the higher net donor levels the source tem--perature was raised from 350 to 375"C.

The orifice areas for the Cd source and S source respectively were 0.47cm and 0.4 x 10-'cm2. The InP substrates were kept at 240"C. Under such conditions net donor concentrations between 5 x 1O'8 and 1 x ll9cm-3 were produced.

An exemplary polycrystalline pInP/nCdS device with a n-CdS layer having a net donor concentration of approximately 1 X 1019 cam~3 showed good stability, i.e. stability up to 600"C for 15 min. anneals in air.

WHAT WE CLAIM IS: 1. A method of making a light-responsive device having a heterojunction between a polycrystalline p-type indium phosphide layer and a light-transmissive polycrystalline n-type cadmium sulphide layer and having an electrode on the cadmium sulphide layer arranged to allow light to fall on the cadmium sulphide layer, wherein the indium phosphide layer is grown by a chemical vapour deposition process using a vapour which provides - phosphorus molecules and hydrogen chloride in a gaseous carrier which vapour is passed over a heated indium source.

2. A method as claimed in claim 1 wherein the gaseous carrier is hydrogen.

3. A method as claimed in claim 2 wherein the vapour is initially provided as phosphorus-trichloride-containing hydrogen gas.

4. A method as claimed in claim 2 wherein the vapour is initially provided as a mixture of hydrogen chloride, phosphine and hydrogen.

5. A method as claimed in any of the preceding claims wherein the indium source is indium phosphide.

6. A method as claimed in claim 5 wherein the indium source is initially provided as elemental indium.

7. A method as claimed in claim 5 or claim 6 wherein the indium source is maintained at a temperature in the range 700"C to 8000 C.

8. A method as claimed in any of the preceding claims wherein the vapour is passed over a heated dopant source.

9. A method as claimed in claim 8 wherein the dopant source is anhydrous cadmium chloride.

10. A method as claimed in any of claims 1 to 7 wherein hydrogen gas is passed over a heated dopant source and then mixed with the vapour.

11. A method as claimed in claim 10 wherein the dopant source is elemental cadmium or zinc.

12. A method as claimed in any of the preceding claims wherein the indium phosphide layer is deposited on a supporting substrate and the cadmium sulphide layer is subsequently deposited on the indium phosphide layer.

13. A method as claimed in claim 12 wherein the substrate is of molybdenum or carbon.

14. A method as claimed in claim 12 or claim 13 wherein the cadmium sulphide layer has a donor concentration of not less than 5 X 10'9 per cubic centimetre.

15. A method as claimed in any of the preceding claims- wherein the indium phosphide layer has a grain size of about two micromeires.

16. A method of making a light-responsive device substantially as herein described with reference to the accompanying drawings.

**WARNING** end of DESC field may overlap start of CLMS **.

Claims

**WARNING** start of CLMS field may overlap end of DESC **. surface of the InP by standard techniques such as electroplating or evaporation. It has been found that the thermal stability in air of the n-type CdS layer of the device is strongly affected by the net donor concentration in the layer. A net donor concentration in the n-type CdS of at least 5 X 10'8cam~3 produces an increase in thermal stability in air when compared to devices with net donor levels of approximately 2 X 10'8cm-" and lower. For testing, devices with high net donor levels were made as described before using the isothermal source: However, to obtain the higher net donor levels the source tem--perature was raised from 350 to 375"C. The orifice areas for the Cd source and S source respectively were 0.47cm and 0.4 x 10-'cm2. The InP substrates were kept at 240"C. Under such conditions net donor concentrations between 5 x 1O'8 and 1 x lûl9cm-3 were produced. An exemplary polycrystalline pInP/nCdS device with a n-CdS layer having a net donor concentration of approximately 1 X 1019 cam~3 showed good stability, i.e. stability up to 600"C for 15 min. anneals in air. WHAT WE CLAIM IS:

1. A method of making a light-responsive device having a heterojunction between a polycrystalline p-type indium phosphide layer and a light-transmissive polycrystalline n-type cadmium sulphide layer and having an electrode on the cadmium sulphide layer arranged to allow light to fall on the cadmium sulphide layer, wherein the indium phosphide layer is grown by a chemical vapour deposition process using a vapour which provides - phosphorus molecules and hydrogen chloride in a gaseous carrier which vapour is passed over a heated indium source.

2. A method as claimed in claim 1 wherein the gaseous carrier is hydrogen.