METHOD FOR HYDROGEN PASSIVATION AND MULTICHAMBER HOLLOW
CATHODE APPARATUS
Introduction
The present invention relates generally to the field of semiconductor device processing and in particular, the invention provides an improved method of hydrogen passivation in semiconductor material. Background
Polycrystalline silicon (pc-Si) has attracted a great deal of interest as an active material in photovoltaic devices and large area applications such as thin film transistors (TFT). However, to obtain technologically useful material, grain boundary and intra-grain defects in pc-Si must be passivated. Hydrogen passivation has been shown to be an effective method for the reduction of the activity of these defects. Hydrogen introduced into pc-Si is reported to diffuse in the positive charge state (H^) and enhance the electrical conductivity ( N.H. Nickel, N.M. Johnson, and J. Walker, Phys. Rev. Lett. 75., 3720 (1995)). The hydrogenation results in a decrease of the defect activity, thereby improving the electrical properties of the materials and devices (T.J. Kamins and P.J. Marcoux, IEEE Electron Device Lett. EDL-1, 159 (1980), A. Mimura, N. Konishi and etc. IEEE Trans. Electron Devices 36., 351 (1989), F. Boulitrop, A. Chenevas-paule, and D.J. Dunston, Solid State Commun. 48., 181 (1983), R. Pandya and B.A.Khan, J.Appl. Phys. 46., 5247 (1987)). The dangling-bond density is reported to decrease as a function of hydrogenation time to a residual saturation value that strongly depends on the passivation technique and passivation temperature. One of the most thoroughly investigated hydrogenation methods, so far, has been to immerse the pc-Si sample in the plasma produced by an rf glow discharge based on parallel plate reactor (I.Wu, T. Huang, and etc. IEEE Electron Device Lett. 12.,
181 (1991), V. Mitra, B. Rossi and B. Khan, J. Electrochem. Soc. 138, 3420- 3424 (1911), T. Takeshita, T. Unnagami and O. Kogure, Jpn. J. Appl. Phys. 27., L2118-L2120 (1988)). Another method is by immersing pc-Si sample in an electron cyclotron resonance (ECR) plasma (T. Takeshita, T. Unnagami and O. Kogure, Jpn. J. Appl. Phys. 27, L2118-L2120 (1988). R.A. Ditizio, G.
Liu and etc., Appl. Phys. Lett. 56, 1140-1142 (1990)). In both rf and ECR plasmas, the introduction of hydrogen into the device is accomplished through plasma-ion penetration and bulk diffusion. The rate of ion penetration is dependent on the plasma-ion density. An rf hydrogen plasma typically has an ion density of ni«109/cm3, while for ECR hydrogen plasma,
density can be as high as ni«1011/cm3. During hydrogenation, the pc-Si is usually kept near 300-450°C to enhance hydrogen diffusion. The relatively low ion densities characteristic of rf plasmas necessitate long processing times. While an ECR plasma has a higher ion density, the sheath potential limits the hydrogen ion current to the pc-Si devices, which also prolongs processing time. Plasma ion implantation (PII) is reported to provide sufficient hydrogenation of pc-Si within a short processing time (J.D. Bernstein, S. Qin and etc., IEEE Trans. Elec. Devices, 43,1876 (1996), N.W. Cheung, Nucl. Instr. Meth. B55, 811 (1991)). The PII process is performed by repetitively applying a large negative voltage pulse to a sample placed in a hydrogen plasma. Hydrogen ions are accelerated by the target potential and implanted into the sample. Ion energies can range from 1-100 KeV with average ion flux densities as high as 1016/cm"2 sec"1. The primary mechanism for the introduction of hydrogen into the device is ion implantation, but additional ion penetration takes place between the pulses. All these tend to enhance dose rates over methods which rely on surface penetration and bulk diffusion.
It is well known that hydrogenation acts to tie up the dangling bonds in the grain boundaries and intra-grain defects with hydrogen atoms. However, it was observed that the characteristics of the pc-Si devices after hydrogen passivation suffer a low hot-carrier endurance and a low thermal stability (S. Banerjee, R. Sundaresan and etc. IEEE Trans. Electron Devices, 35, 152 (1988), M. Hack, A.G. Lewis, and I-W. Wu, IEEE Trans. Electron Devices, 40, 890 (1993)). To improve this situation, NH3 (F. S. Wang, M. J. Tsai and H. C. Cheng, IEEE Trans. Electron Devices, 16, 503 (1995)) and deuterium (L. Lusson, P. Elkaim and etc. Solid State Phenomena, 37-38, 373 (1994)) based plasmas were employed to enhance both the electrical reliability and the thermal stability of pc-Si devices due to stronger Si-N and Si-deuterium bonds. Surface damage created by the H+ ion bombardment can discount the benefits of hydrogenation with respect to the improvement of material and device performance. In both rf and ECR glow discharge techniques, surface damage due to hydrogenation is reported to be quite significant. For PII technique, the surface damage resulting from H+ implantation is most serious as the projected range of H+ implantation at 100 KeV is approximately 1 μm. The primary damage introduced by this process is the
production of vacancies, with a generation rate of vacancies up to 5xl05/cm/ion (J.F. Gibbons, Projected Range Statistics, 1975).
Several approaches have been adopted to reduce the surface damage introduced during hydrogen passivation. These include use of low energy ion implantation sources and remote plasma sources. The Kaufman ion source is a typical example of the former approach and has been used for hydrogen passivation of multicrystalline silicon (mc-Si) solar cells (J.E. Johnson, J.I. Hanoka and J.A. Gregory, IEEE Photovoltaic Specialist Conference, 1112 (1985)). The latter approach separates the plasma source from the pc-Si sample surface and thus hydrogen passivation process relies on diffusion of H^ from the plasma into the sample (M. Spiegel, P. Fath and etc., 13th European Photovoltaic Solar Energy Conference, P.421 (1995)). The H+ ion flux is not particularly high as many H+ ions recombine in the journey from the plasma source to the sample surface.
Summary of the invention
According to a first aspect, the present invention provides a method of hydrogen passivation of a target of semiconductor material including: a) locating the target of semiconductor material to be passivated adjacent to a cathode electrode of a hollow cathode reactor having two opposed, radio frequency powered cathode electrodes enclosed in, and isolated from, a grounded chamber; b) introducing a source of ions of a species suitable for passivation of the semiconductor material into the reactor; c) applying rf energy to the reactor to create a plasma between the cathodes of the reactor characterised in that the reactor is operated to create conditions in which the ion density in the plasma species suitable for passivation is greater than 1011/cm and the ion implantation energy is less than 80eV. In preferred methods according to the invention, while the hollow cathode glow discharge method of the present invention has primarily been developed for hydrogen passivation of silicon material, it can also be used to generate a high density plasma of deuterium or nitrogen for the passivation of pc-Si material. In the case of hydrogen passivation in a plasma of hydrogen ions, a small proportion of Nitrogen (in the range of 0-10% but typically, 4-6%) is beneficial in allowing processing to occur at higher
discharge confinements without extinguishing the discharge as readily. The hollow cathode technique can also be used to passivate defects in other polycrystalline semiconductor materials other than pc-Si such as Ge and GaAs. Two substrates can be simultaneously passivated in a single chamber, the method utilising the discharge maintained between the two cathode electrodes with one substrate placed on the internal side of each of the electrodes.
Preferably, all surfaces of the electrodes, except the two surfaces where the substrates are placed, are enclosed by grounded metal parts, some of which may be chamber walls, with a gap between them sufficiently small to prevent discharge formation at the process pressures.
Preferably also, the ratio of the internal surface area of the electrodes and the internal open surface area of the chamber is more than 1. In one preferred embodiment the method can be extended to accommodate 2n (n= 1,2,3..) targets passivated simultaneously in each of n pairs of electrodes.
In various embodiments of the invention, discharge characteristics and the related hydrogenation process and device characteristics are controlled by confining the discharge with additional metal parts connected either to the electrodes or to the chamber in a way that preserves the symmetry of the system with respect to each of the substrates.
The confining parts connected to the electrodes are preferably shielded from the grounded chamber by grounded parts that are conforming in shape to the surfaces of the conforming parts.
In preferred embodiments of the invention, the ion bombardment energy on the substrate surface is further controlled by spacing the back surfaces of each of the targets from its corresponding electrode.
The gaps described between the targets and respective electrodes are optionally filled with dielectric material.
In one particularly preferred form of the invention, the two electrodes are fixed to a grounded metal frame, but electrically isolated from it, thus forming a portable process vessel which can be loaded into the chamber for the hydrogenation process and then unloaded from it. Preferably, the chamber is operated at a pressure of 5-10 Pa and with a bias voltage of 80-100 V.
According to a second aspect, the present invention consists in a reaction apparatus including a plurality of reaction stages located within a single vacuum enclosure, the vacuum enclosure containing a support structure arranged to receive the plurality of reaction stages, each reaction stage including two opposed radio frequency cathode electrodes defining a hollow cathode reaction cavity in which are located target mounting means adjacent to each electrode, the apparatus including rf power supply connection means for connecting a source of rf power between the cathode electrodes of each stage and an anode electrode defined by the enclosure and an ion source supply means for supplying ion source material to each hollow cathode reaction stage.
Preferably, the reaction stages are removably located in receiving guides provided in the support structure and are connected to the rf supply connection means when they are located in the guide means. Preferably also, the ion source supply means is a gas delivery system arranged to supply a source gas to the cavity between each pair of cathode electrodes.
Preferably also, the surfaces of the cathode electrodes other than those adjacent to the target are enclosed in metal parts connected to the anode of the enclosure, and spaced sufficiently closely to the cathode electrodes to prevent discharge between the cathode and the metal parts.
Preferably, the anode of the enclosure is grounded and includes the enclosure walls.
In one embodiment, heaters are located behind each cathode electrode to control reaction temperature, however, in alternate embodiments reaction temperatures are controlled by controlling the temperature of the supply of ion source gas.
Preferably, the target mounting means is arranged to mount the target substrate with a gap between it and the underlying cathode. Brief description of the drawings
Embodiments of the invention will now be described, by way of example, with reference to the accompanying drawings in which:
Figure 1 is a schematic diagram of a prior art parallel plate reactor.
Figure 2 is a schematic diagram of a hollow cathode reactor. Figure 3 illustrates an optimised Hollow Cathode reactor design, with a cathode electrode diameter of 14 cm and an electrode spacing 3 cm;
Figure 4 illustrates the reactor design of figure 3 with the target materials spaced from its corresponding electrode;
Figure 5 is a simplified perspective view of a hollow cathode Hydrogen passivation system according to an embodiment of the invention providing a plurality of hollow cathode reaction chambers; and
Figure 6 is a cutaway diagram of a module from the hollow cathode reactor of Figure 5 ; Detailed description of the embodiments of the invention
A new hydrogenation process based on hollow cathode glow discharge technique has been developed in which a high density of H*" ions above
10αl/cm3, are generated in the plasma due to the high input power (500W). However, the bias voltage is extremely small, and is about one order of magnitude lower than other techniques previously reported. This technique ensures a short passivation time due to the high ion density and minimum surface damage due to the low bias voltage.
Hydrogenation by Hollow Cathode fhollow cathodej Glow Discharge
The hollow cathode is an enhanced-discharge configuration which permits a high intensity, low voltage discharge to be obtained at low pressure (CM. Horwitz, Appl. Phys. Lett. 43., 977 (1983)). In this, it has benefits similar to other configurations such as magnetron and microwave ECR in providing enhanced operation in comparison with the standard rf diode plasma configuration illustrated in Figure 1. Hollow cathode can also yield a very high power efficiency, and its symmetric construction can minimise high energy electron bombardment of the substrate surface as well as minimising particulate contamination.
The main conceptual difference between conventional glow discharge 132 and hollow cathode glow discharge 32 is in the electrical connection of the two opposing electrodes 20, 21, 120, 121 (compare Figures 1 & 2): in a conventional system, such as that of Figure 1, one of the two electrodes 120 is rf powered, while the other electrode 121 and the rest of the chamber 122 are grounded. In the hollow cathode system as seen in Figures 2, 3 & 4, both the electrodes 20, 21 are rf powered while the chamber 22 is grounded. The increase in the plasma density achieved with the hollow cathode configuration is attributed to so called "electron mirror" effect (CM. Horwitz, Appl. Phys. Lett. 43., 977 (1983)), in which a negative self-developed bias voltage on the two opposing rf powered electrodes helps to increase the
density of the high energy electrons (responsible for ionisation) and to reduce their recombination rate. The hollow cathode effect can be further enhanced by an appropriate discharge confinement which results in an increase in the plasma density, an increase in the ion flux on the electrodes and a decrease in the average ion energy (for the same input power).
In the embodiments of Figures 3 and 4, additional confinement is provided by guard rings 27 and shields 28. 29 surrounding the cathodes 20, 21 and extending over those surfaces of the cathodes 20, 21 which are not adjacent to the target substrates. The gaps between the confinement elements 27, 28 and 29 and the cathodes 20, 21 must be kept sufficiently small that a discharge cannot form in these gaps at the process pressures in use. All these characteristics satisfy the criteria for an effective hydrogen passivation method i.e. high H+ ion flux (short processing time), minimum surface damage due to H+ ion bombardment and better improvement on device performance.
Experimental Results
Hydrogen passivation of pc-Si solar cells was conducted in a confined hollow cathode glow discharge system as shown in Fig. 3. The system consists of two electrodes with dimension of 14 cm in diameter and a spacing of 3 cm. The gap between the two confinement elements is 4 mm.
Hydrogenation is typically undertaken at a pressure of 5-8 Pa and power of 500W (1.66W/cm2). The bias voltage between the plasma and the anode at this condition is only 80V, one order of magnitude lower than that produced in other types of plasma systems. The density of H+ ions produced under this condition is above 10 /cm3. The optimum passivation time during this experiment was around 30 to 45 minutes for films of 2—3 μm thickness but depends on film thickness.
As previously outlined, the important limitation relating to conventional hydrogen passivation is the surface damage resulting from hydrogen bombardment. This damage causes an increase in the surface recombination, thus lowering the ultimate performance of the device. To achieve good grain boundary and defect passivation with minimum surface damage, a high concentration of hydrogen radicals should be created at a minimum acceleration energy. In a conventional rf glow discharge system, such conditions are difficult to obtain since a high input power is necessary to achieve a high concentration of hydrogen ions, which generates a very
high self-developed negative bias voltage on the sample (up to 1000 V). This leads to a high energy ion bombardment and thus a resultant surface damage. The Kaufman ion source also requires a relatively high acceleration voltage to generate the same concentration of atomic hydrogen as used in our experiments, although the situation is improved compared with the standard rf diode plasma configuration.
The surface damage due to hollow cathode hydrogenation can be further reduced by placing the samples 2 to 3 mm away from the cathode electrodes but still within the dark space of plasma (as illustrated in Figure 4). This arrangement reduces the acceleration energy of hydrogen ions impinging on the pc-Si surface and thus reduced the surface damage. A high plasma density hollow cathode glow discharge system operates at an order of magnitude lower bias voltage than the conventional rf glow discharge system and the Kaufman ion source, thus reducing the surface damage to a negligible level. The surface damage can be characterised by electron beam induced current (EBIC) images taken at low electron beam energy in a scanning electron microscopy. The EBIC analysis indicates that the surface damage introduced by the hollow cathode glow discharge is negligible. Commercial versions of the Hollow Cathode reactor
The Hollow Cathode system can be scaled to accommodate large area substrate as required for solar module and liquid crystal display (LCD) applications. In this process the ratio of the electrode to the ground chamber areas should be maintained at approximately the same level to preserve the "electron mirror" effect responsible for the high plasma density achieved in the hollow cathode. Figure 5 shows a schematic of a multichamber hollow cathode system in which a series of hollow cathode stages 34 are connected, with the unit stage frame commonly grounded. Adjacent stages 34 may be attached permanently and in which case they may share a common chamber wall. Figure 6 shows the detail of a single hollow cathode stage 34 including the chamber 40, plasma confinement element 45, gas delivery tubes 46, cathode electrodes 41, 42 and glass substrate targets 43, 44, being coated with a layer of pc-Si. The chamber frame 40 is grounded with two rf electrodes 41, 42 separately placed on either side of the reactor. Two targets 43, 44 are placed facing each other and with each adjacent to an rf cathode electrode.
Two gas delivery tubes 46 are positioned within the gaps 47 in the plasma confinement elements 45.
Alternatively, without the use of gas delivery tubes 46, the gas can be introduced in the gap behind each of the cathode electrodes 41,42 through a set of holes in the grounded wall of the chamber 40 of the hollow cathode stage and will then enter the hollow cathode discharge zone through the gaps between confining elements 45. The hollow cathode system multichamber system has the advantages of achieving higher values of the utilisation factor and the high throughput. Features
The preferred embodiment of the hollow cathode hydrogen passivation system has the following features which give advantages over conventional rf diode system for hydrogen passivation.
Semiconductor films can be passivated simultaneously on two metal or dielectric substrates 25, 26, 43, 44 from a radio frequency glow discharge 32 maintained in a mixture of gases or vapours with at least one of them containing hydrogen in its molecule. The discharge 32 is maintained between two mutually symmetric, radio frequency powered electrodes 20, 21, 41, 42 enclosed in an isolated grounded chamber 22, 40, with one target placed on the internal side of each of the cathode electrodes 20, 21, 41, 42.
All surfaces of the cathode electrodes 20, 21, 41, 42 except the two surfaces where the targets 25, 26, 43, 44 are placed, are enclosed in grounded metal parts 27, 28, 29, 45, some of which may be chamber walls, with a gap between the cathodes 20, 21, 41, 42 and adjacent metal parts 27, 28, 29, 45 being sufficiently small to prevent discharge formation at the process pressures.
The method can be extended so that 2n (n= 1,2,3..) targets are passivated simultaneously in each of n pairs of the electrodes such as by using a multichamber arrangement particularly well suited to the hollow cathode configuration.
The target temperature during the deposition can be controlled by grounded heaters located behind the electrodes and conforming in shape to the back surface of the heaters.
In the case of the multichamber configuration, the heating of each chamber can be realised by flowing a hot gas through the chamber.
10
Discharge characteristics and related deposition process and film characteristics can be controlled by confining the discharge by metal parts connected either to the electrodes or to the chamber in a way that preserves the symmetry of the system with respect to each of the targets. The confining parts connected to the electrodes can be shielded from the grounded chamber by grounded parts that are shaped to conform with the electrode back surfaces.
The ion bombardment energy on the substrate surface and related film characteristics can be controlled by introducing and varying equal gaps between the back surfaces of each of the targets and the cathode electrodes.
These gaps can be defined by dielectric material.
The two electrodes can be fixed on a grounded metal frame, but isolated from it, thus forming a portable process vessel which can be loaded in the chamber for the passivation treatment and then unloaded from it. Hydrogen gas (H2) is believed to be the most suitable source of H+ ions). The hydrogen can be diluted with noble gases and small amounts of nitrogen.
It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.