Patent Application
20250232958
The present invention relates to an apparatus and method for the production of films by chemical vapour deposition.
Metal or metalloid containing films, such as gallium nitride (GaN) films, have applications in a range of devices from light emitting diodes (LEDs) to ultraviolet detectors to transistor devices.
These films have commonly been produced by techniques including molecular beam epitaxy (MBE), metal organic chemical vapour deposition (MOCVD) and remote plasma enhanced chemical vapour deposition (RPECVD or RPCVD). RPECVD has been employed to produce films of high quality at considerably lower temperatures than those used in MOCVD which thereby reduces process costs and allows the use of temperature sensitive preferred substrates for film deposition.
One problem which must be addressed during film production using any chemical vapour deposition (CVD) technique is to obtain an even and controlled distribution of reagents across the substrate of the surface onto which the film is to be grown to thereby achieve uniform thin film growth. At least part of the solution to this problem may be addressed by the design of the distribution systems. For example, in RPECVD a shower head or lattice design may be employed to obtain an even distribution of metal organic reagent across the substrate while a baffle may be used to enhance the even distribution of the plasma stream of active nitrogen species. One such baffle design is disclosed in WO/2010/091470, the disclosure of which is hereby incorporated in its entirety, wherein an ‘inverse pagoda’ style of baffle is used to diffuse and filter the plasma stream.
Many of these approaches focus on growth of a single film and so the reagent, e.g. metal organic, distribution lattice and plasma channel with baffle are generally centred over the substrate location to thereby provide a generally homogeneous distribution of both materials over the entire substrate surface. This form of chamber design is not so effective when it is desired to grow a plurality of films within the same growth chamber to improve productivity.
The use of multiple substrates is particularly desirable in those film deposition techniques where the growth rate is extremely slow. For example, atomic layer deposition (ALD) is a useful growth technique based upon the sequential pulsing of chemical precursor vapours to thereby achieve one atomic layer per pulse. Due to the sequential pulsing arrangement of ALD each reagent pulse reacts with the deposition surface until the reaction is completed. A purge gas is used to carry away excess reagent and reaction side products after each pulse in an attempt to minimise impurities being deposited in the film.
ALD is of interest due to the ability to produce thin uniform films with a high degree of control over film thickness and composition. One of the drawbacks of ALD is the amount of time required to grow a useful film since only a monolayer may be deposited in each complete deposition cycle. The time required for each cycle is limited by the switching speed of the reagent release valves as well as the time taken to purge after each half cycle and rotate the substrate into place. This results in each full cycle taking from 0.5 to a few seconds, further contributing to the slow production.
Further, the purging cycle is not entirely effective which often means that an amount of the metal organic reagent, such as trimethylgallium (TMG), remains in the growth chamber during pulsing of the plasma containing the active second reagent, such as nitrogen. This may result in carbon impurities being incorporated into the growing film, thereby reducing its quality.
Minimising the extent of incorporation of both carbon and oxygen, as impurities, in the growing film is a major challenge in CVD film production.
As well as altering the desired chemical composition of the film these impurities disrupt the lattice matching of forming layers thereby causing defects within the film and negatively impacting on the overall quality of the product.
MOCVD approaches have been relatively more successful than certain other CVD techniques at lowering oxygen incorporation into the growing thin films but levels of carbon incorporation are not ideal. More particularly, MOCVD often involves growth temperatures of about 1000° C. to 1200° C., which thereby results in high equipment costs and rules out the use of temperature sensitive preferred substrates for film deposition.
It would thus be desirable to provide for a CVD apparatus and method which could provide for the advantages in control of film growth afforded by ALD while minimising the drawbacks of that technique. Particularly, it would be useful to provide for a CVD apparatus and method allowing for a reduction in the levels of incorporation of carbon and oxygen as impurities in the film product and preferably a CVD apparatus and method which can be run at lower temperatures than those employed in a standard MOCVD approach.
In a first aspect, although it need not be the only or indeed the broadest form, the invention resides in an RPCVD apparatus for forming a film, the apparatus including a growth chamber comprising:
Preferably, the additional reagent inlet is an ammonia inlet.
Preferably, the Group VA plasma inlet, the Group IIIA reagent inlet and the additional reagent inlet open into the growth chamber at a distance between about 1 cm to about 30 cm from a growth surface of the one or more substrates. More preferably, between about 1 to about 20 cm or 1 to about 10 cm.
Preferably, a ceiling of the growth chamber (reactor) is located less than about 30 cm vertically above the location of the substrates, more preferably less than about 25 cm, even more preferably less than about 20 cm, still more preferably less than about 10 cm. Values of 5 cm and 7.5 cm may be useful with 3 cm to 4 cm as the lower end values.
In certain embodiments at least one of the Group VA plasma inlet, the Group IIIA reagent inlet and the additional reagent inlet end flush with the ceiling of the growth chamber which is located between about 1 to about 30 cm, 1 to 20 cm, 1 to 10 cm vertically above a growth surface of the substrates, preferably, between 4 to 15 cm, 4 to 10 cm, 4 to 8 cm.
Suitably, an opening of the additional reagent inlet opens into the growth chamber in close proximity to the one or more substrates.
The additional reagent inlet may extend downwardly from the ceiling of the growth chamber to end in close proximity to the growth surface of the one or more substrates.
In one embodiment, the additional reagent inlet opens into the growth chamber through a side wall thereof at a height suitable to enable a flow of additional reagent entering therethrough to have a flow path passing over and substantially adjacent to the growth surface of the one or more substrates.
In a preferred embodiment there is a direct flow path between the Group VA plasma inlet and the one or more substrates.
Suitably, the direct flow path between the Group VA plasma inlet and the one or more substrates extends to an unimpeded path between a plasma generator for generating the Group VA plasma and the one or more substrates.
In an embodiment, the Group VA plasma inlet and the Group IIIA reagent inlet end flush with a ceiling and/or side wall of the growth chamber through which they extend.
Preferably, the first deposition zone is substantially isolated from the second deposition zone.
Preferably, rotation of the substrate holder (susceptor) causes the one or more substrates to pass sequentially from the first deposition zone to the second deposition zone. The substrates may be one or more wafers, as are known in the art.
Preferably, the substrate holder is of a turntable design whereby it rotates around a central pivot and is provided with a plurality of recesses, each adapted to hold a substrate, around its periphery.
In one preferred embodiment, the substrate holder includes multi-axis rotation whereby the substrate, or a collection of substrates, are located on a satellite, or collection of satellites that rotate about their centre of axis and also about the reactors central pivot. This multi-axis rotation scheme is typically referred to as a planetary rotation system, and can be used to further improve film uniformity compared to a single axis rotation system.
The Group VA plasma inlet and the Group IIIA reagent inlet may be located centrally within the growth chamber.
When the Group VA plasma inlet and the Group IIIA reagent inlet are located centrally within the growth chamber one or both thereof may be provided with a flow control device to direct the corresponding plasma or reagent into the appropriate first or second deposition zone.
The flow control device may be a flow barrier blocking one or more reagent flow paths within the Group VA plasma inlet or the Group IIIA reagent inlet or a directing portion, such as a shroud, continuous with the first or second reagent inlet.
In one embodiment, the apparatus may further comprise a baffle associated with the Group VA plasma inlet such that the plasma substantially passes therethrough.
The baffle may comprise the flow control device which may be a flow barrier blocking one or more outlets of the baffle.
Preferably, the additional reagent inlet opens into the growth chamber substantially adjacent to the opening of the Group IIIA reagent inlet to promote mixing of said reagents prior to their contacting the one or more substrates.
Suitably, the Group VA plasma inlet is in fluid communication with a plasma generator producing a Group VA plasma comprising an active species.
Preferably, the Group VA plasma is a nitrogen plasma comprising active nitrogen species.
Suitably, the Group IIIA reagent is a Group IIIA metal organic reagent.
In one particularly preferred embodiment, the Group VA plasma inlet and the Group IIIA reagent inlet are located peripherally within the growth chamber.
Suitably, the Group VA plasma inlet and the Group IIIA reagent inlet are located substantially at opposite ends of the growth chamber.
The growth chamber may comprise one or more structures associated with the additional reagent inlet and/or the Group IIIA reagent inlet to promote mixing of said reagents immediately prior to their contacting the one or more substrates.
The apparatus may further comprise one or more heating devices to heat the additional reagent inlet and/or the Group IIIA reagent inlet prior to entering the growth chamber.
In a second aspect the invention resides in a method of forming a thin film on a substrate by RPCVD including the steps of:
Preferably, the additional reagent is ammonia.
Suitably, the additional reagent is introduced into the second deposition zone substantially adjacent the opening of the Group IIIA inlet.
In one embodiment, the additional reagent is introduced into the growth chamber through a side wall thereof.
In one embodiment, the additional reagent is introduced into the growth chamber to form a substantially horizontal flow path passing over and substantially adjacent to the growth surface of the substrate.
The additional reagent and the Group IIIA reagent are preferably being introduced into the growth chamber simultaneously.
Suitably, the Group IIIA reagent is a Group IIIA metal organic reagent.
Preferably, the Group IIIA metal organic reagent is a Group IIIA metal alkyl reagent.
Preferably, the Group IIIA metal alkyl reagent is selected from the group consisting of trimethylgallium, triethylgallium, trimethylindium and trimethylaluminium.
The method may further include the step of heating one or more of the reagents prior to their entering the growth chamber.
The method may further include the step of promoting the mixing of the metal organic reagent and the additional reagent adjacent the one or more substrates.
Suitably, the Group VA plasma inlet is in fluid communication with a plasma generator.
Preferably, the Group VA plasma is a nitrogen plasma comprising active nitrogen species.
The isolation of the deposition zones substantially prevents the mixing of the Group VA plasma and Group IIIA reagent.
The method may further include the step of controlling the flow of one or more of the Group VA plasma or Group IIIA reagent upon exiting the associated inlet to direct that flow to a desired deposition zone.
The method may further include the step of controlling the temperature to be between about 400 to about 1000° C., preferably between about 500 to about 900° C. (inclusive of a temperature of about 500° C., 600° C., 700° C., 750° C., 800° C., 850° C. or 900° C.), more preferably between about 500 to about 850° C.
Any reference herein to a specific gas flow rate or plasma power is made in the context of a given reactor or growth chamber size. It is well-understood and accepted within the relevant industry that larger reactors utilise more gas than smaller reactors and so reagent flow rates, for example, are scaled up or down appropriately based upon the size of the reactor. Typically, this scaling factor is proportional to the available growth area, or rather to the total surface area of the substrate holder. The person of ordinary skill in the art will be aware that it is commonly accepted that the size of the reactor is defined by quantifying how many circular wafers of a given diameter can fit on the substrate holder. For example, a 7×2″ reactor can hold up to seven 2-inch round substrates (wafers). Other common reactor sizes include 19×2″, 31×2″, 56×2″ that can each hold nineteen, thirty-one, and fifty-six 2-inch round wafers, respectively. Reactor sizes can also be quantified relative to larger substrate/wafer sizes, such as for example 5×4″, 6×6″, or 3×8″. Reference is made to the following as an example of this commonly accepted definition for reactor sizes: Heuken, M.; Juergensen, H. (1999) MOCVD production tool for high speed electronic devices. In: Gallium Arsenide Applications Symposium. GAAS 1999, 4-5 Oct. 1999 https://amsacta.unibo.it/id/eprint/1421/1/GAAS_99_059.pdf. It will be understood by the person of ordinary skill in the art that any reference to a specific substrate/wafer configuration, such as 19×2″, is simply a means of quantifying the reactor size, and is not to be construed as any restriction on the actual wafer loading configuration used. For example, a reactor that can support 19×2″ wafers is understood as also being able to support 5×4″ wafers. Whether one chooses to load 5×4″ wafers or 19×2″ wafers, or even only 10×2″ wafers, for example, it is the same reactor presenting the same overall substrate holder area and so wafer capacity. The reagent flows and process conditions are therefore set based upon the surface area of the substrate holder (as a useful proxy for growth chamber/reactor size) rather than for the particular number and size of wafers actually placed within that growth chamber/reactor.
In combination with the presence of an additional reagent gas, preferably ammonia, it has been found that the power of the plasma generator has an effect on carbon incorporation into the thin film and so the method may also include the step of controlling the power of the plasma generator to be between about 500 W to about 5000 W from a single source. Larger reactors may require more than one single source to ensure the density of active nitrogen provided by the plasma is uniformly distributed across the entire growth area. This may be combined with a growth pressure of 2-5 torr and a nitrogen plasma flow of 2000-3000 sccm with an ammonia flow of about 15 to about 1500 sccm, preferably about 20 to about 200 sccm, preferably about 20 to about 100 sccm, more preferably about 20 to about 50 sccm. These flow values are made with reference to a 7×2″ reactor size, and would increase when moving to larger reactor sizes, roughly proportionally with the growth area.
The growth pressure may be between 2-5 torr, 2-4 torr or about 3 torr.
Preferably, the power of the plasma generator is between about 100 watts to about 3000 watts with a nitrogen flow rate of 1000-3000 sccm extending to 100-20000 sccm in a commercial unit. A value for the plasma generator power of about 500 to 5000 W, 500 to 4000, 500 to 3000, 500 to 2000, and preferably about 1500 W is preferred and each value or range of which may be independently coupled with an ammonia flow rate of any one of between 15 to 1500 sccm, such as between 50 sccm to 600 sccm, or 100 sccm to 600 sccm or 50 sccm to 550 sccm or 100 sccm to 550 sccm or 50 sccm to 500 sccm or 100 sccm to 500 sccm are particularly useful in lowering carbon incorporation (for a growth chamber/reactor with a 7×2″ substrate capacity), however, in moving to commercial scales, such as 19×2″, 31×2″, 42×2″, 56×2″ reactor sizes, higher powers and multiple plasma sites are envisaged as being useful.
In a third aspect the invention resides in a method of forming a thin film having a carbon impurity content of less than about 5E+17 atom/cm3, on a substrate by RPCVD including the steps of:
Preferably, the carbon impurity content is less than about 3E+17 atom/cm3, even more preferably less than about 2E+17 atom/cm3, yet more preferably less than or about 1E+17 atom/cm3. A lower limit may be considered to be about the SIMS detection limit for carbon impurities in such films.
In one embodiment, the thin film also has an oxygen impurity content of less than about 8E+17 atom/cm3, even more preferably less than about 6E+17 atom/cm3, yet more preferably less than about 4E+17 atom/cm3, still more preferably less than about 2E+17 atom/cm3, or even less than or about 1E+17 atom/cm3. A lower limit may be considered to be about the SIMS detection limit for oxygen impurities in such films.
The statements made above in relation to the second aspect apply equally well to the third aspect.
In a fourth aspect the invention resides in a film made by the method of the second or third aspects.
In a fifth aspect the invention resides in use of a film of the fourth aspect in a semiconductor device.
Further features of the present invention will become apparent from the following detailed description.
Throughout this specification, unless the context requires otherwise, the words “comprise”, “comprises” and “comprising” will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.
In order that the invention may be readily understood and put into practical effect, preferred embodiments will now be described by way of example with reference to the accompanying figures wherein:
The present inventors have identified a particular RPCVD apparatus and process conditions for the production of high quality films which results in improvements to the film growth rate and growth control, by comparison to standard ALD techniques and other CVD processes, and, importantly, which provides for a surprising level of reduction in oxygen and carbon-based film impurities due to reagent side reactions.
The reagents which may be employed with the present apparatus and method, and hence the nature of the films which can be formed, are not particularly limited. Although the embodiments discussed herein generally employ a nitrogen plasma and a metal organic (typically a gallium containing metal organic such as trimethylgallium) as the reagents, the utility of the present invention is not so limited. The Group IIIA (otherwise known as Group 13 under the current IUPAC system) reagent may comprise an element which is selected from the group consisting of boron (B), aluminium (AI), gallium (Ga), indium (In) or thallium (TI). The Group VA (otherwise known as Group 15 under the current IUPAC system) plasma may be generated from any suitable reagent containing a Group VA element selected from the group consisting of nitrogen (N), phosphorus (P), arsenic (As), antimony (Sb) and bismuth (Bi).
The term “deposition zone” as used herein is used to refer to a distinct region, section or segment of the growth chamber into which one or more reagents are introduced. Individual depositions zones, such as first and second deposition zones, are isolated one from the other such that a substrate or growing film will only be substantially exposed to a particular reagent introduced only into one deposition zone when the substrate or growing film actually enters that deposition zone. The separation or isolation of deposition zones may be spatial only or may be effected by partial or complete physical barriers.
In the embodiments described the reagents employed will be trimethylgallium, a nitrogen plasma and ammonia but the person skilled in the art will appreciate the principles disclosed herein could be applied mutatis mutandis to other reagent combinations.
Without wishing to be bound by any particular theory, the present inventors have demonstrated that an experimentally observed dramatic reduction in the levels of oxygen and carbon incorporated into gallium nitride films produced in an RPCVD apparatus are due to a choice of conditions, including, primarily, the supply of an additional gaseous reagent, preferably ammonia gas, to mix with the Group IIIA reagent and preferably in close proximity to the substrate, which favours formation of a trimethylgallium:ammonia Lewis acid:base adduct. This adduct breaks down to form gallium nitride with the release of methane gas which is not incorporated into the growing film to the same extent as methyl radicals would be.
In typical prior art CVD approaches, and particularly MOCVD due to the high temperatures employed, a molecule of trimethylgallium injected into the growth chamber will decompose thermally to finally produce a gallium atom and three methyl radicals. The gallium will react with the nitrogen source, which may be high ammonia flow rates sufficient to provide active nitrogen to form the GaN film. The reactive methyl radicals are often incorporated into the growing film as an impurity thereby increasing strain and lowering the overall quality of the film product.
Without wishing to be bound by theory, when the growth temperature is instead kept below the thermal decomposition point of trimethylgallium and an additional reagent, preferably ammonia, is introduced into the growth chamber, at low flow rates relative to the substrate holder area available to hold wafers, then the two components form a Lewis acid-base adduct of proposed formula (CH3)3Ga:NH3. It is further proposed that this adduct reacts to form an intermediate (CH3)2Ga:NH2+CH4. A further step in this pathway leads to the formation of an adduct of formula [(CH3)2Ga:NH2]3 with eventual formation, from that structure, of three molecules of GaN and six molecules of CH4 gas. The methane gas is less reactive than a methyl radical and is easily removed via the exhaust of the growth chamber.
The inventors have further postulated that such adduct formation and preference of the formation of methane over methyl radicals can be encouraged by minimising the extent of the reactions which occur in the gas phase i.e. in the upper and central regions of the growth chamber and instead maximising the mixing of the reagents only in the immediate vicinity of the substrates. This can be achieved by introducing the ammonia or other additional reagent into the growth chamber with the Group IIA reagent in a manner which causes it to be present or available only adjacent the substrates and hence the growth surface of the growing film.
It can be seen that the plasma enters an area of the growth chamber 105 directly above the metal organic injector 150 and so, in operation, the plasma containing active neutral nitrogen species and the metal organic reagent mix and react to form the particular metal nitride, such as gallium nitride, which is deposited on the substrate to form the film. Excess reagents, carrier gases, contaminants etc are removed via a waste outlet 155.
Carbon and oxygen are inevitably incorporated as impurities into the film but, these aside, this approach is generally satisfactory for the formation of a film on a single substrate. However, it is often desirable to have the capacity to generate a number of such films at the same time. Hence, an apparatus such as that shown in
A plasma generator 215 receives nitrogen through a nitrogen inlet 220 and the active nitrogen plasma formed passes through plasma inlet 225, which once again is remote from the substrates, and into the growth chamber 205 via a baffle 230, which in the embodiment shown takes the form of an inverse pagoda style baffle as described in WO/2010/091470.
The plasma passes through the baffle 230 and is evenly distributed by its concentric ring-like structure. The distributed plasma flow then passes over a metal organic injector 235 where the metal organic reagent is introduced and mixes with the plasma. The metal nitride formed will then deposit on one or more of substrates 240 located on a substrate holder 245. The substrate holder 245 may be of a turn table design and so may be rotating at high speed throughout the deposition process. Waste is removed via outlet 250.
It will be appreciated that the central placement of the plasma inlet 225 will likely result in the bulk of the plasma flow being focused on the centre of the substrate holder 245, even with the use of a distribution system such as the inverse pagoda baffle 230. It is critical to quality film growth that the reagents be distributed evenly across the surface of the appropriate substrate 240 and the deficiency in this apparatus 200 will not be solved by rotation of the substrate holder 245. This type of apparatus 200 also does not typically provide the advantages of control over film growth and thickness that is provided by atomic layer deposition (ALD).
As for
The active nitrogen species then pass by a Group IIIA reagent injector in the form of metal organic reagent (e.g. trimethylgallium) injector 330. In
Located beneath the metal organic reagent injector 330 are a number of substrates 345 which are supported by a substrate holder 350. The substrate holder 350 may hold any desired number of wafers, for example, from 2 to 56 individual substrates. It will be appreciated that the number of substrates that can be held in the substrate holder will increase with increased growth chamber/reactor size and so will be selected accordingly. The substrates may have a crystal structure that is suitable for growth of the particular film desired. In particular examples, the substrates 345 may comprise sapphire, SiC, silica, soda lime glass, borosilicate glass, Pyrex®, silicon, glass, synthetic sapphire, quartz, zinc oxide, nitride coated substrates and other materials as are well known in the art including free standing bulk semiconductor substrates and nitride templates. As is indicated by the arrow in
The combined effect of the directing action of flow barrier 325 on the plasma pathway and the release of metal organic reagent from only those open ports 335, as can be seen from
The sequential exposure of each substrate 345 to the metal organic reagent and the active nitrogen species will result in formation of subsequent layers of a film, much in the manner of ALD. However, the formation of separate deposition zones means that the delays experienced in ALD in both waiting for valves to be switched and the removal of one reagent by a purge gas before introduction of the second reagent, are avoided. Instead, the growing surface of the film is exposed to each reagent with a minimum of downtime in between due to the ability of the substrate holder 350 to rotate at very high speeds. This greatly accelerates the growth of the films while maintaining control over sample growth.
The substrate holder 350 may be adapted to rotate continuously. Preferably, the substrate holder is capable of rotating at speeds of between 10-2000 rpm. A preferred rotation speed may be between 25 to 100 rpm, more preferably about 50 rpm. The skilled addressee will understand that film growth will be controlled by a combination of the speed of rotation of the substrate holder 350 and the reagent flow rates in the deposition chamber. Higher rotation speeds of the substrate holder 350 will require a higher flow rate of reagents to ensure an overall increase in the growth rate of the film is produced.
It will be appreciated that although only one plasma inlet 310, and associated baffle 320, and one region of open ports 335 have been shown in
It will also be understood that baffle 320 is not an essential feature but may be preferred, under certain process conditions, to prevent or reduce etching due to active nitrogen species which may have relatively high kinetic and/or potential energies. If the baffle 320 was not present in
The apparatus 300 may further comprise one or more heaters to heat the growth chamber and/or one of the reagent inlets. This may be useful to promote increased reaction rates, assist with quality of the growing film or to break or otherwise activate one or more of the reagents before exposure to the substrates.
As was mentioned above, the present apparatus and method are not particularly limited in the type of reagents suitable for use therein. Any reagents which are suitable for use in ALD may be appropriate. A wide range of reagent classes including, nitrogen plasma, nitrogen/hydrogen plasma, ammonia plasma and metal organics may be suitable. When a metal organic reagent is used then preferred examples include alkyl Group IIIA reagents such as but not limited to one or more of trimethylgallium, trimethylindium, trimethylaluminium as well as employing various well known Mg, Si and Zn precursors as dopant sources.
The key difference in
When looking at a section of the growth chamber 405, as represented in
The schematic representation shown in
In the embodiment shown in
Although in
It will be appreciated that in the embodiments described in
However, even employing the approach discussed above it has been found that sufficient amounts of oxygen and carbon impurities are still incorporated in the thin film formed to thereby reduce its quality. The use of RPCVD approaches, while more convenient in many ways than a standard MOCVD approach, are generally accepted in the art as inevitably resulting in moderate levels of oxygen and carbon impurities in the films produced as compared with high end MOCVD produced films. Very low impurity limits can be considered as being at least relatively close to the SIMS detection limits as set out in table 1.
However, the present inventors have found that the levels of these impurities can be lowered significantly by the use of an apparatus as shown in
The growth chamber 605 is also provided with a Group IIIA reagent inlet and, more specifically, in the embodiment discussed herein, a metal organic reagent inlet 620 and an additional reagent inlet which may be a hydrazine inlet, a di-methyl hydrazine inlet or a hydrogen plasma inlet but is preferably an ammonia inlet 625. As was described for
Plasma introduced into the growth chamber 605 will directly contact the substrate 630 placed in the first deposition zone as no baffle, shroud or like blocking or distributing device is in place in the embodiment shown. The present inventors have found that, for a 7×2″ reactor size, when using such an apparatus under conditions of relatively low power of the plasma generator (around 1000 W to 300 W) and temperature (about 750° C. to 850° C.) no significant degree of etching was observed. The use of this arrangement with an injection of between about 50 to about 250 sccm ammonia resulted in a substantial reduction of the levels of oxygen and carbon incorporated into the film product.
In a preferred embodiment a reactor pressure of between about 4 to 10 Torr may be appropriate and potentially going as high as 20 Torr, if required. Additionally, controlling the power of the plasma generator to be between about 500 W to about 4000 W from a single source is useful in the present invention. A nitrogen plasma flow of between about 2000 to 4000 sccm may be used alongside these conditions and, finally, an ammonia flow of about 50 to 500 sccm or 100 to 500 sccm is further preferred. These values are all appropriate for a reactor with a 7×2″ wafer capacity. With the exception of pressure, which generally does not scale with reactor size, these parameter ranges will scale approximately, and in a manner which would be well-understood by a person of ordinary skill in the art, with the growth chamber/reactor size. A useful approximation of reactor size is to consider the area of the susceptor (substrate holder), or put another way, the maximum available growth area. This is typically defined in terms of the wafer capacity, or simply how many wafers of a given size can fit in the reactor on the substrate holder. Table 2 below simply confirms how these ranges would be understood by a person of ordinary skill in the art to scale for some typical reactor capacities that are common to the industry.
Further process runs were carried out varying the power output from the plasma generator. At a power output of about 800 W the level of carbon incorporated into the film was reduced to levels approaching the actual detection limit of secondary ion mass spectrometry (SIMS). Oxygen levels have been reduced to those observed using MOCVD wherein oxygen is effectively removed as an impurity of concern. Such a reduction in the levels of oxygen and carbon in RPCVD produced films has hitherto not been shown.
It is envisaged that the method may also include the step of controlling the power of the plasma generator to be between about 500 W to about 5000 W from a single source. This range would be suitable with a growth pressure of 2-3 torr, a nitrogen plasma flow of 2000-3000 sccm and an ammonia flow of between about 15-1500 sccm, for a 7×2″ reactor capacity.
In embodiments, a reactor pressure of 4-10 Torr, a plasma generator power of 500-5000 W (for a single source), a nitrogen plasma flow rate of 2000-4000 sccm and an ammonia flow rate of 50-500 sccm, or 50-400 sccm or 100-500 sccm or 100-400 sccm may be appropriate for a reactor with a 7×2″ wafer capacity. These conditions can be scaled appropriately for larger reactors based upon the growth area which may be viewed as the substrate holder (susceptor) area. For example, for a 19×2″ reactor the susceptor area is about ×4 greater; for a 42×2″ reactor the susceptor area is about ×6 greater; and for a 56×2″ reactor the susceptor area is about ×8 greater (all compared with the 7×2″ reactor size). Therefore, scaling the ammonia flow values set out for the 7×2″ reactor (100-400 sccm) in this paragraph provides 400-1600 sccm for a 19×2″ reactor; 600-2400 sccm for a 41×2″ reactor; and 800-3200 sccm for a 56×2″ reactor. The other growth parameters in this paragraph, except for growth pressure, would also scale accordingly. This scaling for changing reactor size based upon the susceptor area and so number of wafers which can be accommodated, is well-understood and accepted in the field.
Preferably, the power of the plasma generator is between about 100 watts to about 5000 watts, preferably about 500 to about 3000 W with a nitrogen plasma flow rate of 1000-3000 sccm extending to 100-20000 sccm in a commercial unit. A preferred metal organic reagent flow rate is 1200-2000 sccm which may extend to 100-10000 sccm in a commercial unit. A value for the plasma generator power of about 500 to 1000 W, 500 to 900 W, 500 to 800 W, 600 to 1000 W, 600 to 900 W, 600 to 800 W, 700 to 1000 W, 700 to 900 W and preferably about 800 W is preferred.
Such power levels may be independently coupled with an ammonia injection (in sccm) of between 15 to 1500 sccm. Ranges of between about 15 to 200, preferably 15 to 150, 15 to 100, 15 to 75, 15 to 60, 15 to 50, 15 to 40, 15 to 30, 20 to 150, 20 to 100, 20 to 75, 20 to 60, 20 to 50, 20 to 40, 20 to 35, 20 to 30, 25 to 150, 25 to 100, 25 to 75, 25 to 60, 25 to 50, 25 to 40, 25 to 30, including values of about 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100, 120, 140, 160, 180 and 200 sccm have been found to be particularly useful in lowering carbon incorporation for relatively small growth chamber sizes, for example, of dimensions in the order of 7×2″. However, in moving to commercial scale larger reactor sizes, higher powers and multiple plasma sites are envisaged as being useful. For growth chambers of larger dimensions of, for example, 19×2″, 31×2″, 42×2″, 56×2″, then ammonia injection flows of between about 200 to about 1500 sccm, inclusive of 200 to 1300, 200 to 1100, 200 to 1000, 200 to 900, 200 to 800, 200 to 700, 200 to 600, 200 to 500, 200 to 400, 300 to 1500, 300 to 1300, 300 to 1100, 300 to 1000, 300 to 900, 300 to 800, 300 to 700, 300 to 600, 300 to 500, 300 to 400, 400 to 1500, 400 to 1300, 400 to 1100, 400 to 1000, 400 to 900, 400 to 800, 400 to 700, 400 to 600, 400 to 500, 500 to 1500, 500 to 1300, 500 to 1100, 500 to 1000, 500 to 900, 500 to 800, 500 to 700, 500 to 600, 600 to 1500, 600 to 1300, 600 to 1100, 600 to 1000, 600 to 900, 600 to 800, 600 to 700, 700 to 1500, 700 to 1300, 700 to 1100, 700 to 1000, 700 to 900, 700 to 800, 800 to 1500, 800 to 1300, 800 to 1100, 800 to 1000, 800 to 900, 900 to 1500, 900 to 1300, 900 to 1100, 900 to 1000, 1000 to 1500, 1000 to 1300, 1000 to 1100, are appropriate. A person of ordinary skill in the art will understand that the most appropriate flow rate will depend on the reactor size and will generally be proportional to the surface area of the substrate holder, or the available growth area.
It will be appreciated that higher power output from the plasma generator can be tolerated when a baffle or like device is employed whereas lower values will be preferred when there is an unimpeded flow path between the plasma inlet and the substrates to minimise etching.
Although not shown in
As discussed earlier, the present inventors postulate that it is important for the minimisation of carbon and oxygen impurities in the final film to minimise reactions which occur in the gas phase above the substrates. Instead, it is preferable to encourage the key film forming reactions to occur on or as close to the actual substrate surface as possible. Directing the reactions to occur on the substrate surface may improve the scavenging of the oxygen and carbon impurities.
Thus the lowering of the growth chamber 605 ceiling, with reagent inlets formed therein and having their openings flush with said ceiling, results in delivery of the reagents more quickly and effectively to the substrate surface.
Minimising deadspots and, particularly, optimising the flow of the plasma and reagents in relation to the substrates are considered in
A plasma generator 1015 is located externally to the growth chamber 1005 which may be a high frequency generator acting upon nitrogen received from a nitrogen source (not shown). The nitrogen plasma thereby generated enters the growth chamber 1005 at plasma inlet 1020 which ends flush with the ceiling of the growth chamber 1005 i.e. the plasma inlet 1020 does not, to any significant extent, extend into the interior of the growth chamber 1005. The plasma inlet 1020 may, if required, open into a baffle, shroud, impeller or the like to modify the flow path and energy of the plasma. This is not an essential component and the need for such a device will depend upon the power of the radiofrequency generator. A suitable baffle may be as described in the applicant's prior PCT publication WO 2010/091470 which is hereby incorporated by way of reference in its entirety.
A metal organic reagent source 1025 supplies the metal organic reagent which, in a preferred embodiment, is trimethylgallium (TMG) or triethylgallium (TEG). The TMG or TEG enters the growth chamber 1005 via metal organic reagent inlet 1030 which, in the embodiment shown, is located in a side wall of the growth chamber 1005 and ends flush therewith i.e. the metal organic reagent inlet 1030 does not, to any significant extent, extend into the interior of the growth chamber 1005.
A hydrogen-containing, additional reagent source 1035 supplies the additional reagent which, in a preferred embodiment, is ammonia. The ammonia enters the growth chamber 1005 via additional reagent inlet 1040 which, in the embodiment shown, is located in a side wall of the growth chamber 1005, beneath the location of the metal organic reagent inlet 1030, and ends flush therewith i.e. the additional reagent inlet 1040 does not, to any significant extent, extend into the interior of the growth chamber 1005.
Under some conditions it may be preferable that the additional reagent inlet 1040 enters the growth chamber 1005 through a side wall thereof at a height suitable to enable a flow of additional reagent entering therethrough to have a flow path passing over and substantially adjacent to a growth surface of the substrates 1050. It is also beneficial that the positioning of the exhaust 1010 is at an opposite end of the growth chamber 1005 to the additional reagent inlet 1040 which further encourages a flow path of the additional reagent which passes over the surface of the substrates 1050. Thus, the region of injection of the additional reagent and the provision of the exhaust 1010 generally opposite that create an environment whereby the reagent is in constant contact with the growth surface of the substrates 1050 and growing film.
The provision of the three plasma/reagent inlets all with ends flush with either the ceiling (the plasma inlet 1020) or the side walls (the metal organic reagent inlet 1030 and the additional reagent inlet 1040) or a combination thereof avoids the presence of ‘dead spots’ within the growth chamber 1005. It is preferred that the Group IIIA reagent inlet is flush with the chamber ceiling. It is further preferred that the additional reagent inlet injects the ammonia, or other gas, through entry points, for example view ports, which physically surround the additional reagent inlet to thereby have these two reagents introduced into the growth chamber together to encourage mixing.
The presence of inlets which extend into the growth chamber 1005 would result in adjacent regions therein where reagents can collect and be moved around in vortex like movements, due to the spinning of substrate holder 1045. Dead spots are considered to be unwanted volumes within the growth deposition chamber where there could be depletion or recirculation of gas which does not help with the growth of the film. This would encourage reaction pathways other than the desirable adduct formation outlined above and would result in TMG or TEG degradation with methyl radical production.
The substrate holder 1045 may be adapted to support a single substrate 1050 but it is preferred that it is of a design adapted to support multiple substrates 1050. Suitably, the substrate holder 1045 is rotatable.
The design of the apparatus 1000 shown in
As discussed previously, the ceiling height of the growth chamber 1005 plays a critical role in the deposition process and film quality. Suitable heights have been set out previously. Lowering the ceiling height can help to minimise undesirable non-adduct forming reactions by minimising the space in which they can occur due to the placing of the plasma and reagent inlets close to the substrates 1050.
Although not shown in
Although not shown for the sake of simplicity, the growth chamber 1005 may comprise one or more structures associated with the additional reagent inlet 1040 and/or the metal organic reagent inlet 1030 to promote mixing of said reagents immediately prior to their contacting the one or more substrates 1050. Particularly, it may be desirable to generate some turbulence in the flow path of the introduced additional reagent, preferably being ammonia. Since this reagent is introduced to generate a flow path just above the substrates 1050 this ensures rapid and efficient mixing with the TMG or TEG to promote adduct formation prior to contact with the growth surface of the growing film.
The structures themselves may take the form of a baffle-like structure, vanes or any shape which promotes flow turbulence. They may be in direct contact with the corresponding reagent inlets or may be operatively associated with them such that the reagent must flow through the structure before passing close to the substrates 1050.
The key difference between the embodiments of
The provision of the extended portion 3060 does result in the potential generation of one or two “dead spots” as discussed before but this does not prevent the formation of an improved film product compared with many prior art RPCVD approaches. The design of the apparatus 3000 still ensures that the ammonia, or other additional reagent, is only provided in close proximity to the growing film such that adduct formation and the production of methane, as opposed to methyl radicals, is promoted immediately adjacent said film.
The growth chamber 5005 is defined, in part, by the ceiling 5010 which, in relative terms, is not very distant from the substrates 5035 to minimise the chamber mixing space. A plasma inlet 5015 opens into the growth chamber 5005 through the ceiling 5010 to deliver a plasma, such as a nitrogen plasma. A metal organic reagent inlet takes the form of an injector framework 5020 (details of the port openings not shown) while a additional reagent inlet 5025 opens into the growth chamber 5005 through a side wall thereof at a point underneath the metal organic injector framework 5020 and a height such that a flow path of a additional reagent, preferably ammonia, is created just above the growth surface of the substrates 5035 which are rotating with the movement of the substrate holder 5030.
The design in
It will be appreciated that, in one embodiment, the additional reagent may only need to supply hydrogen in a reactive form if the plasma is a nitrogen plasma and so can be used as the nitrogen source for adduct formation. This would result in the use of a hydrogen plasma generator and inlet along with a nitrogen plasma generator and inlet. Due to the possibility of arcing between these components it would be preferable to physically separate them as much as possible and so one may be located in the ceiling at one end of the growth chamber 5005 and the other in a side wall at an opposite end of the growth chamber 5005.
Although the discussion herein has been of rotation of the substrate holder it will be appreciated that it may be possible for the substrate holder and substrates to remain stationary while the plasma inlet and reagent inlets spin within the growth chamber. This will require a design whereby a rotatable connection mates with each of the plasma inlet, the metal organic reagent inlet and the additional reagent inlet which will both be operated in pulses timed to coincide with the rotation speed to ensure that each is only dispersed where required. Such a design would present greater challenges in operation over those disclosed in the figures which have the substrate holder rotate while the plasma inlet and reagent inlets remain stationary and, hence, is a less preferred approach.
The apparatus may also be adapted to allow additional individual rotations of each substrate relative to the substrate holder, typically referred to as a planetary rotation system, for further improvement of thin film growth uniformity.
Thus, from the various embodiments described above, it will be appreciated that the components of the inventive apparatus described herein may be arranged in a number of different ways while still achieving a reduced carbon and/or oxygen level in the growing films compared with that achieved by standard RPCVD approaches. However, all of the embodiments described share at least the feature of the additional reagent being introduced to the growth chamber in close if not immediate proximity to the Group IIIA reagent introduction point and, preferably, the substrates to promote formation of the adduct directly above the growing film surface. This approach has been found to greatly reduce the level of carbon and/or oxygen incorporation into the film. Further common features which assist in further reducing the level of carbon and/or oxygen incorporation include a low chamber ceiling height and therefore a correspondingly reduced chamber volume along with provision of reagent inlets which end flush with the ceiling and/or side walls to minimise dead spots and positioning of the exhaust to encourage a flow path of additional reagent over the substrate surface.
In one highly preferred embodiment of the present invention any one or more and most preferably all of the Group VA plasma inlet, the Group IIIA reagent inlet and the additional reagent inlet all end flush with the ceiling of the growth chamber, as previously discussed. However, if the Group IIIA reagent inlet and the additional reagent inlet are to extend into the chamber then, in one embodiment, it is useful if they extend into the chamber to be between about 1 to about 10 cm vertically above the substrates inclusive of between 2 to 9 cm, 3 to 6, cm and 4 to 5 cm.
The process of film formation as described in relation to any of the aspects herein may also include a doping step, which may be necessary for films to be employed in devices such as LEDs and solar cells. Preferably, the doping step is a p-type doping step. For p-type doping, the dopants could be Mg or Zn or other suitable elements. Suitable reagents that contain these elements, such as diethyl zinc (DEZn), bis(cyclopentadienyl)magnesium (Cp2Mg) can be selected from those known in the art for p-type doping. p-type doping is known in the art to be particularly challenging but it has been found that the present set of process conditions and apparatus features used to reduce carbon impurities in the growing film also allow for better p-type doping. Values obtained for p-type doping (Hall measurements) are: resistivity of 0.9 Ohm-cm, mobility of 2.7 cm{circumflex over ( )}2/Vs for a carrier concentration of 1.4E18 cm{circumflex over ( )}−3. For n-type doping, the dopants could be Si or oxygen or other suitable elements. Suitable reagents that contain these elements, such as silane, disilane, di-tert-butylsilane, oxygen can also be used for n-type doping.
It will be appreciated from the foregoing discussion that a number of other factors can be controlled to further contribute to the extent of the reduction of impurities achieved in the film product.
For example, the method may further include the step of controlling the temperature to be between about 400 to about 1200° C., preferably between about 500 to about 900° C. (inclusive of a temperature of about 500° C., 600° C., 700° C., 800° C., 900° C. or 1000° C.), more preferably between about 500 to about 850° C. or about 700 to about 850° C. This is a relatively low temperature range in comparison to typical MOCVD and even many RPCVD approaches. The lower temperatures favour adduct formation over TMG thermal degradation and so reduce methyl radical reactions at the film surface.
The method may further include the step of promoting the mixing of the metal organic reagent and the additional reagent adjacent the one or more substrates using a flow perturbation device. Once again, the mixing step is to promote immediate formation of the adduct in the vicinity of the film/substrate surface.
It has been found that the power of the plasma generator has an effect on carbon incorporation into the thin film and so the method may also include the step of controlling the power of the plasma generator to be between about 400 W to about 5000 W from a single source. Preferably, the power of the plasma generator is between about 500 to about 3000 W, 500 to 2750 W, 500 to 2500 W, 500 to 1000 W, 500 to 900 W, 500 to 800 W, 600 to 1000 W, 600 to 900 W, 600 to 800 W, 700 to 1000 W, 700 to 900 W and preferably about 800 W is preferred. A value of about 800 W has been found to be particularly useful in lowering carbon incorporation, however, in moving to commercial scale reactor sizes, higher powers and multiple plasma sites are envisaged as being useful such as about 2500 W from a single source.
Although not wishing to be bound by any particular theory, it is postulated that the surprising results achieved may be as a result of one or more of the following processes. Firstly, it is postulated that the injection of ammonia acts as a scavenger gas for oxygen and/or carbon. Secondly, it is theorized that the improvement in lowering carbon incorporation into the film when the power output of the plasma generator is increased could be due to the carbon atoms being actively removed in favour of nitrogen. Thirdly, and as alluded to previously, a proposed mechanism for removal of carbon from the system is that the ammonia assists in the formation of an adduct with the trimethylgallium, i.e. initially (CH3)3Ga:NH3, which subsequently releases a methane molecule. The methane is not incorporated into the film as readily as a CH3 radical may be. Subsequent decomposition of the adduct releases further methane until all of the carbon of the trimethylgallium has been removed as methane and only GaN is left. It is believed formation of this adduct and subsequent decompositions are occurring at the surface of the substrate. Finally, it is possible that the improved lowering of carbon and oxygen incorporation upon removing the baffle from the system is a consequence of the energy from the UV light emitted from the plasma generator/plasma chamber contacting the growing film ejecting carbon and/or oxygen in favour of nitrogen. In reality, it is possible that all of these mechanisms may play at least some role in providing the results achieved.
In a third aspect the invention resides in a method of forming a thin film having a carbon content of less than about 5E+17 atom/cm3, on a substrate by RPCVD including the steps of:
Preferably, the carbon impurity content is less than about 3E+17 atom/cm3, even more preferably less than about 2E+17 atom/cm3, yet more preferably less than or about 1E+17 atom/cm3. A lower limit may be considered to be about the SIMS detection limit for carbon impurities in such films.
In one embodiment, the thin film also has an oxygen impurity content of less than about 8E+17 atom/cm3, even more preferably less than about 6E+17 atom/cm3, yet more preferably less than about 4E+17 atom/cm3, still more preferably less than about 2E+17 atom/cm3, or even less than or about 1E+17 atom/cm3. A lower limit may be considered to be about the SIMS detection limit for oxygen impurities in such films.
The statements made above in relation to the second aspect apply equally well to the third aspect.
In a fourth aspect the invention resides in a film made by the method of the second or third aspects. Such films will have demonstrably lower levels of oxygen and/or carbon incorporated into their structure in comparison to similar films made by standard RPCVD approaches. In one embodiment, films produced by the method of the present invention may have a carbon content of less than about 10E+16 atoms/cm3. Values of 3E+16 atoms/cm3 have been attained and it is believed that values of less than 1E+16 atoms/cm3 are attainable with process optimisation.
In a fifth aspect the invention resides in use of a film of the fourth aspect in a semiconductor device.
The examples set out in further detail the process runs using the apparatus of the invention and the results thereby obtained. In the examples nitrogen was used as the Group VA plasma and trimethylgallium as the Group IIIA reagent.
Process Runs with Baffle
An apparatus essentially as set out in
A second run was then carried out under essentially similar conditions but with an injection of a 15 sccm ammonia flow into the second deposition zone (the organometallic reagent deposition zone). The ammonia was injected at the same time as injection of the TMG so that the two mixed together prior to deposition. This produced a film with a substantial reduction in the levels of both oxygen and carbon. Specifically, compared with the first run without the injection of ammonia, the level of carbon decreased from about 6E+20 atom per cubic centimetre (atom/cc) to about 3E+20 atom/cc while the level of oxygen decreased from about 3E+20 atom/cc to about 1E+17 atom/cc.
The figure of 1E+17 atom/cc for the oxygen level represents an extremely surprising result in that it ceases, for practical purposes, to be a problematic impurity at that level and the result is comparable to that observed using MOCVD. Although the additional hydrogen provided by the ammonia may be expected to provide some benefit in reducing carbon and oxygen impurities it could not have been predicted, based upon accepted wisdom in the semiconductor field, that such a small injection of ammonia could result in such a large reduction in oxygen and carbon impurities.
This experiment was repeated using different flow rates of ammonia. The results of these process runs, in terms of carbon incorporation into the film, are shown in
Process Runs without Baffle
A number of runs were then carried out using the same apparatus and conditions as described above but with the shower head baffle removed. Thus, a direct flow path between the plasma generator, plasma inlet and substrates was established in the first deposition zone.
The distance between the plasma inlet and the substrates was less than 20 cm and no plasma etching was observed. It is postulated this may be due to the relatively low (500-600 W) power output from the plasma generator employed while still providing enough energy to activate the nitrogen.
The results of this run are indicated on
A further process run was carried out under identical conditions (30 sccm ammonia) but with the power of the plasma generator increased to 800 W. Once again, surprisingly, the resulting film was not etched to a significant degree. More surprising, however, was that the level of carbon incorporated into the film was about 1.7E+17 atom/cc. This result is indicated on
It will be appreciated that there will be an upper limit to the plasma generator power which can be employed without etching occurring. If this point is reached and further increases in plasma generator power are desirable then a baffle may once again be placed between the plasma inlet and the substrates.
Further runs designed to optimise the process have resulted in levels of 3E+16 atom/cc of carbon and 3E+16 atom/cc of oxygen in the grown GaN films. It may assist the appreciation of how low these levels are by considering that the SIMS detection limits for carbon are between about 1-2E+16 and between about 1-3E+16 for oxygen. At the above quoted optimal results the present films are approaching the detection limits for carbon and oxygen. SIMS is one of the most sensitive surface analysis techniques available, being able to detect elements present in the parts per billion range, and is the accepted standard for analysis in this field. The low levels of carbon and oxygen impurities achieved by use of the present apparatus and method are comparable with those observed in GaN layers grown by MOCVD, and have never been seen previously using RPCVD.
The first 0-0.5 μm of the depth profile (indicated on the x axis) represents a film produced by the present apparatus and method while the 0.5-2.7 μm component represents the underlying GaN template produced by an MOCVD process. It is clear that the levels of the various impurities, particularly carbon and oxygen, are similar. The spikes observed in the traces are representative of the interfaces between layers or changes in growth conditions and not an increase in impurity levels.
Triethylgallium Experiments A further series of experiments were performed using triethylgallium (TEG) as the Group IIIA reagent and investigating different ammonia injection rates. The process conditions used for these experiments are set out below in table 3. Also noteworthy is that the plasma inlet, TEG inlet and ammonia inlets all ended flush with the growth chamber ceiling to reduce recirculation of gases i.e. deadspot effects. The ceiling was fixed at a height approximately 5.0-7.5 cm above the substrates and the speed of rotation of the substrate holder was 1200 rpm.
The results of the experiments set out in table 3, in terms of the carbon and oxygen impurity levels in the grown films, are shown graphically in the SIMS data in
It can be seen from
It will be appreciated from all of the foregoing that the use of separate deposition zones in an RPCVD arrangement can be useful in reducing impurity incorporation into films, such as GaN films, however these impurities will still be found in the produced films in noticeable quantities. The use of a growth chamber arrangement whereby the plasma inlet and/or organometallic reagent inlet are between about 1 to about 30 cm vertically above a growth surface of the substrates. However, the introduction of relatively small quantities of ammonia into the second deposition zone, simultaneously with the organometallic reagent, has been shown to dramatically reduce the level of oxygen and, particularly, carbon in the film. Under similar conditions but with the removal of any impediment to direct flow between the plasma generator and/or plasma inlet and the substrates results in an extremely surprising reduction in the incorporation of, particularly, carbon into the film. As a further level of control it has been shown that increasing the power of the plasma generator output can further lower the level of carbon impurities found in the film.
Throughout the specification the aim has been to describe the preferred embodiments of the invention without limiting the invention to any one embodiment or specific collection of features, or to a specific reactor size or substrate loading configuration. It will therefore be appreciated by those of skill in the art that, in light of the instant disclosure, various modifications and changes can be made in the particular embodiments exemplified without departing from the scope of the present invention.
1. An RPCVD apparatus for forming a film, the apparatus including a growth chamber comprising:
2. The apparatus of item 1 wherein the additional reagent inlet is an ammonia inlet.
3. The apparatus of item 1 or item 2 wherein the first deposition zone is substantially isolated from the second deposition zone
4. The apparatus of any one of the preceding items wherein the Group VA plasma inlet and/or the Group IIIA reagent inlet open into the growth chamber at a distance between about 1 cm to about 30 cm from a growth surface of the one or more substrates.
5. The apparatus of item 4 wherein the Group IIIA reagent inlet opens into the growth chamber at a distance between about 1 cm to about 10 cm from a growth surface of the one or more substrates.
6. The apparatus of item 1 wherein at least one of the Group VA plasma inlet or the Group IIIA reagent inlet end flush with a ceiling of the growth chamber which is located between about 1 to about 30 cm vertically above a growth surface of the one or more substrates.
7. The apparatus of item 6 wherein the ceiling is located between about 15 to 30 cm vertically above a growth surface of the one or more substrates.
8. The apparatus of any one of the preceding items wherein the additional reagent inlet opens into the growth chamber substantially adjacent to the opening of the Group IIIA reagent inlet to promote mixing of said reagents prior to their contacting the one or more substrates.
9. The apparatus of any one of the preceding items wherein the growth chamber comprises one or more structures associated with the additional reagent inlet and/or the Group IIIA reagent inlet to promote mixing of said reagents immediately prior to their contacting the one or more substrates.
10. The apparatus of any one of the preceding items wherein there is a direct flow path between the Group VA plasma inlet and the one or more substrates.
11. The apparatus of item 10 wherein the direct flow path between the Group VA plasma inlet and the one or more substrates extends to an unimpeded path between a plasma generator for generating the Group VA plasma and the one or more substrates.
12. The apparatus of item 1 wherein an opening of the additional reagent inlet opens into the growth chamber in close proximity to the one or more substrates.
13. The apparatus of item 12 wherein the additional reagent inlet opens into the growth chamber at a distance between about 1 cm to about 10 cm from a growth surface of the one or more substrates.
14. The apparatus of item 12 wherein the additional reagent inlet extends downwardly from the ceiling of the growth chamber to end in close proximity to a growth surface of the one or more substrates.
15. The apparatus of item 12 wherein the additional reagent inlet opens into the growth chamber through a side wall thereof at a height suitable to enable a flow of additional reagent entering therethrough to have a flow path passing over and substantially adjacent to a growth surface of the one or more substrates.
16. The apparatus of any one of the preceding items wherein the Group VA plasma inlet and the Group IIIA reagent inlet are located centrally within the growth chamber.
17. The apparatus of item 16 wherein at least one of the Group VA plasma inlet and the Group IIIA reagent inlet is provided with a flow control device to direct the corresponding plasma or reagent into the appropriate first or second deposition zone.
18. The apparatus of any one of item 1 to item 15 wherein the Group VA plasma inlet and the Group IIIA reagent inlet are located peripherally within the growth chamber.
19. The apparatus of item 18 wherein the Group VA plasma inlet and the Group IIIA reagent inlet are located substantially at opposite ends of the growth chamber.
20. The apparatus of any one of the preceding items wherein rotation of the substrate holder causes the one or more substrates to pass sequentially from the first deposition zone to the second deposition zone.
21. The apparatus of any one of the preceding items further comprising one or more heating devices to heat the additional reagent inlet and/or the Group IIIA reagent inlet prior to the respective reagents entering the growth chamber.
22. A method of forming a thin film on a substrate by RPCVD including the steps of:
23. The method of item 22 wherein the additional reagent is ammonia.
24. The method of item 22 or 23 wherein the additional reagent is introduced into the second deposition zone substantially adjacent the opening of the Group IIIA inlet.
25. The method of any one of item 22 to item 24 additional reagent and the Group IIIA reagent are preferably being introduced into the growth chamber simultaneously
26. The method of any one of item 22 to item 25 wherein the Group IIIA reagent is a Group IIIA metal organic reagent.
27. The method of item 26 wherein the Group IIIA metal organic reagent is a Group IIIA metal alkyl reagent.
28. The method of item 27 wherein the Group IIIA metal alkyl reagent is selected from the group consisting of trimethylgallium, triethylgallium, trimethylindium and trimethylaluminium.
29. The method of any one of item 22 to item 28 wherein the Group VA plasma is a nitrogen plasma comprising active nitrogen species.
30. The method of any one of item 22 to item 29 further including the step of promoting the mixing of the Group IIIA reagent and the additional reagent adjacent the one or more substrates.
31. The method of any one of item 22 to item 30 wherein the additional reagent flow rate is between 15 to 1500 sccm.
32. The method of item 31 wherein the additional reagent flow rate is between 30 to 1000 sccm.
33. The method of any one of item 22 to item 31 further including the step of controlling the power of the plasma generator to be between about 500 to about 4000 W.
34. The method of item 33 wherein the power of the plasma generator is between about 500 to about 3000 W.
35. The method of any one of item 22 to item 34 wherein the growth pressure in the growth chamber is between 2-5 torr.
36. The method of any one of item 22 to item 35 wherein the plasma flow is between 2000-3000 sccm.
37. The method of any one of item 22 to item 36 further including the step of controlling the temperature in the growth chamber to be between about 400 to about 1200° C.
38. The method of item 37 wherein the temperature in the growth chamber is between about 500 to about 1000° C.
39. The method of item 38 wherein the temperature in the growth chamber is between about 500 to about 800° C.
40. The method of any one of item 22 to item 39 further including the step of isolating the deposition zones to prevent the mixing of the Group VA plasma and Group IIIA reagent.
41. The method of any one of item 22 to item 40 further including the step of controlling the flow of one or more of the Group VA plasma or Group IIIA reagent upon exiting the associated inlet to direct that flow to a desired deposition zone.
42. The method of any one of item 22 to item 41 wherein the additional reagent is introduced into the growth chamber through a side wall thereof.
43. The method of any one of item 22 to item 42 wherein the additional reagent is introduced into the growth chamber to form a substantially horizontal flow path passing over and substantially adjacent to a growth surface of the substrate.
44. The method of any one of item 22 to 43 item further including the step of heating one or more of the reagents prior to their entering the growth chamber.
45. The method of any one of item 22 to 44 further include a step of p-type doping of the growing film.
46. A method of forming a thin film having a carbon impurity content of less than about 5E+17 atom/cm3, on a substrate by RPCVD including the steps of:
47. The method of item 46 wherein the carbon impurity content is less than about 3E+17 atom/cm3.
48. The method of item 47 wherein the carbon impurity content is less than about 2E+17 atom/cm3.
49. The method of item 48 wherein the carbon impurity content is less than or about 1E+17 atom/cm3.
50. The method of item 46 wherein the oxygen impurity content of the thin film is less than about 6E+17 atom/cm3.
51. The method of item 50 wherein the oxygen impurity content is less than about 2E+17 atom/cm3.
52. A film formed by the method of any one of item 22 to item 51.
53. Use of the film of item 52 in a semiconductor device.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2012903023 | Jul 2012 | AU | national |
| 2012903455 | Aug 2012 | AU | national |
This application is a Continuation in Part of U.S. patent application Ser. No. 17/171,468, filed Feb. 9, 2021, which is a Divisional Application of U.S. patent application Ser. No. 14/413,587, filed Apr. 28, 2015, which is a 35 U.S.C. 371 National Stage filing from International Application No. PCT/AU2013/000786, filed Jul. 15, 2013, which claims priority to Australian Application No. 2012903455, filed Aug. 10, 2012, and to Australian Application No. 2012903023, filed Jul. 13, 2012, the contents of each of which are incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 14413587 | Apr 2015 | US |
| Child | 17171468 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 17171468 | Feb 2021 | US |
| Child | 19169397 | US |
