Patent Grant
7514374
The present invention relates to the production of flat substrates with an extent of at least 2500 cm2 and having a silicon layer deposited by means of a PECVD process in a vacuum reactor. Thereby, substrates are manufactured whereupon, generically, the silicon layer becomes part of semiconductor devices as in manufacturing flat substrates for thin film transistor displays (TFT) or for liquid crystal displays (LCD), or for solar cells or for organic light-emitting displays (OLED).
In today's manufacturing of such flat large substrates with a silicon layer, amorphous silicon (a-Si) is the most commonly used material for such layer. These layers are commonly deposited by a PECVD process. Unfortunately, however, a-Si has relatively poor electronic properties and a-Si-based devices on such flat substrates tend to show important degradations under electrical stress.
It is known in the art that crystalline silicon (μc-Si as well as “polycrystalline silicon”) could be a good candidate for replacement of a-Si, as devices made of more crystalline material exhibit better performance such as higher electron mobility (higher field effect mobility), higher ON-current and tend to degenerate less over time (threshold voltage shift).
It is also known in the art that highly crystalline material can be produced by plasma-enhanced chemical vapor deposition processes (PECVD) using standard equipment and plasma-activating a mixture of silicon-containing gas, halogen-containing gas, hydrogen and a noble gas. We refer, as an example to P. Roca i Cabaroca et al. in “Journal of SID” Dec. 1, 2004, where methods are reported for growing on top of a silicon nitride layer (SiN) μc-Si material from a SiF4—H2—Ar gas mixture. The problem inherent to this methods is that of poor uniformity of thickness and of quality for both, the underlaying dielectric layer—SiN—as well as the addressed μc-Si layers. The degree of uniformity of thickness and of quality is governed by the chemical and electrical homogeneity in the near-the-substrate environment. Indeed, both films, dielectric film and μc-Si film, have been found to be thinner and of different chemical composition at a peripheral portion of the substrate compared with these characteristics in the more central portion of the substrate.
It is an object of the present invention to provide a method for manufacturing flat substrates with an extent of at least 2500 cm2, substantially centrally symmetrically and having a Si layer deposited by means of PECVD processing in a vacuum reactor, whereat uniformity of at least one of layer quality and of layer thickness along the substrate up to its periphery is improved. This is achieved by such method which comprises:
For the dielectric layer as deposited in step d) the thickness non-uniformity mainly results from the higher etching rate at the peripheral portion of the substrate also during subsequent Si-layer deposition. Etching radicals originate especially from the plasma used to grow the μc-Si layer in step e). The increase in etching rate towards the periphery of the substrate is believed to be caused by a higher proportion of these etching radicals near the edge of the substrate compared to such proportion in a more central area of the substrate. This can be understood knowing the good chemical neutrality of aluminum alloys with respect to fluorine, which alloys are customarily used for semiconductor fabrication vacuum reactor's wall. Such neutrality—lack of consumption—leads to an increased proportion of etching radicals along the surface of aluminum alloy of the reactor's inner wall and thus to an increased etching rate at the periphery of the substrate. It must be pointed out that the footprint of the vacuum reactor is customary dimensioned as small as possible in view of the dimension of substrates to be manufacture therein. Therefore the spacing between reactor's metallic wall and the edge or periphery of the substrate to be manufactured is not tailored so large that the addressed effect of such wall could be neglected.
Indeed, a most commonly used aluminum alloy for PECVD reactors is an aluminum magnesium alloy, because it develops a protective fluorine-based outer layer and can therefore easily withstand a cleaning step. Thereby, customarily fluorine radicals, as from NF3 or SF6 in plasma are used during such plasma-enhanced cleaning without corrosion of the vacuum reactor's inner surface for standard PECVD silicon processes.
Due to the chemical neutrality of the inner surface of the PECVD reactor the available quantity of etching radicals is higher near the edges of the substrate and these radicals migrate, as by osmotic effect, onto the substrate's peripheral area. The smaller that the footprint of the reactor's inner space is dimensioned compared with the dimension of the substrate, the more pronounced is this effect. The higher quantity of available etching radicals, fluorine radicals, will affect the μc-Si layer growth at the peripheral area of the substrate as by affecting crystallization. Therefore, the local perturbation of the chemical balance between etching- and growth-contributing radicals leads to lowering the deposition rate as well as changing the Si-material composition between defined crystalline fractions, amorphous fractions and voids at the peripheral area of the substrates. Generically, such decrease in layer thickness and change of material composition towards the edge of the substrate leads to undesired deviations of characteristics of semiconductor devices produced along the substrate and towards the periphery of the substrate, relative to characteristics of such devices more centralized on the substrate.
By depositing on at least a part of the inner surface of the reactor a dielectric precoat, before introducing the substrate, and then performing layer deposition upon the substrate, which includes depositing the μc-Si layer as a silicon layer, the inner surface of the vacuum reactor, which is metallic, as was addressed mostly of an aluminum magnesium alloy, is shielded from the precursor gas as used especially during deposition of the μc-Si layer. It is believed that thereby the proportion of etching radicals along the peripheral area of the substrate is kept substantially equal to such proportion in a more centralized substrate area. With respect to spatial relative density distribution of etching radicals, one will encounter substantially no difference in the central area of the substrate and along its periphery.
Thus, the principal which is followed up by the present invention is to present to the plasma activated gas substantially equal conditions, whether seen at the edge or more centrally of the substrate.
Attention is drawn to the following prior art documents: U.S. Pat. No. 5,177,578, U.S. Pat. No. 5,970,383, U.S. Pat. No. 6,071,573, U.S. Pat. No. 5,981,899, U.S. Pat. No. 5,811,195, JP 2002 289 557, P. Roca i Cabaroca et al. in “Journal of SID” Dec. 1, 2004.
The U.S. Pat. No. 5,970,383 teaches to deal with the time development of vacuum reactor wall coverings during manufacturing of series or batches of silicon-coated substrates by resetting the wall characteristics after each batch. Thereby, a coating is applied on the wall of the vacuum reactor. The influence of reactor's wall upon the local uniformity of the single substrate covering is not addressed.
The U.S. Pat. No. 5,981,899 teaches a capacitively coupled RF plasma reactor of the type as may be used to practice the method according to the present invention.
The U.S. Pat. No. 5,811,195 details the use of aluminum magnesium alloys for the walls of a vacuum reactors for semi-conductor processing.
For industrial processes a required uniformity of material properties and of thickness along large substrates of at least 2500 cm2 is important, considered over the whole substrate area. E.g. in layer deposition for TFT-devices, as used for manufacturing substrates for TFT display panels, the final properties of each transistor device of the backplane depend substantially from the thickness of the dielectric layer as deposited in step d) addressed above and on the structural and electronic properties of the intrinsic Si-layer material. Indeed, transistor response is defined by the threshold voltage and by the charge-carrier mobility. Variation in TFT layer thickness and in TFT material property along various areas of the backplane will lead to variations of the transistor properties. This will result in a non-uniformity of pixel luminance over the panel.
The limitations which have been described with respect to the μc-Si layer deposition up to now as from halogen precursor gas are substantially remedied by the manufacturing method according to the present invention and foreseen for industrial-scale production.
The dielectric precoat, which is applied to cover at least parts of the inner surface of the reactor, leads to a substantially ameliorated uniformity of the chemical environment as “seen” from the substrate and along its entire surface to be coated.
When we speak of providing a dielectric precoat on “at least a part of the inner surface of the reactor”, we do not want to exlude that some areas of such inner surface will not be covered by the precoat, which uncovered parts being of no or neglectable relevance for the electrochemical environment which the overall surface of the substrate to be coated sees during processing.
In one embodiment of the method according to the present invention the reactor is provided with a substrate carrier first electrode and, spaced therefrom, with a gas shower second electrode and the substrate is introduced in the addressed step c) on the first substrate carrier electrode.
Here the substrate is deposited onto the dielectric precoat applied to the carrier electrode. Customarily the surface of the precoat-covered first electrode is larger than the surface of the substrate deposited thereon. Considering the dielectric layer deposition in addressed step d), there results that the subsequent μc-Si layer deposition “sees” practically continuously a dielectric surface along the substrate and along the projecting precoated electrode surface as well as along wall areas of the reactor, neighboring the carrier electrode and the substrate edge. Thereby electric field uniformity too, especially along the edge of the substrate, e.g. of glass material, is greatly improved.
Due to the manufacturing method of the invention substrates of large surface result, the silicon layer thereon being provided as a μc-Si layer with improved uniformity of thickness and quality up to the edge of the substrate. This makes sure that electrical properties of devices formed on the basis of such silicon layer are of substantially improved local uniformity over the whole substrate area, and, due to the μc-Si deposition, are also of improved stability over time.
In one embodiment of the method according to the present invention at least one of depositing in step b) and of depositing in step d) a dielectric material is performed by PECVD processing.
Considering the fact that in step e) of the method according to the present invention, a μc-Si layer is deposited by PECVD, the overall processing is significantly simplified by making use of the same processing type at least for depositing the precoat according to step b) or to apply the dielectric layer according to step d) or to perform both addressed steps b) and d). Nevertheless, in some circumstances it might be preferred to deposit the precoat layer and/or the dielectric layer by means of e.g. reactive physical vapor deposition (PVD) using an RF plasma discharge as well.
When we talk throughout the present invention of applying or generating an RF plasma, this shall not exclude that besides of an RF supply there might be applied to the respective electrodes additional a DC bias or such RF plasma might be generated by pulsating RF or even pulsating DC of high enough repetition frequency.
Additionally, it has to be considered that for some reasons a microwave plasma might be used. Here microwave energy is coupled into the reactive space of the vacuum PECVD reactor. In this case the first electrode is rather used as a substrate carrier, the second electrode rather as a gas shower for homogeneously introducing the respective gas or gas mixture.
In a further embodiment depositing on the dielectric layer according to step e) a μc-Si layer comprises plasma activating a gas or gas mixture, which produces etching and layer growth-contributing radicals.
In one embodiment such gas is selected to be SiF4.
Still in a further embodiment of the method according to the invention in step e) a gas mixture comprising a silicon-containing gas, a halogen-containing gas and hydrogen is plasma-activated.
Still in a further embodiment of the invention step e) comprises activating a noble gas.
Still in a further embodiment of the invention at least the surface of the substrate to be coated is of glass, customarily today the entire substrate.
Still in another embodiment of the present invention at least one of step b) and of step d) comprises depositing of at least one of silicon oxide, silicon nitride, silicon oxynitride, fluorinated silicon oxide.
Still in a further embodiment of the method according to the invention step b) comprises depositing silicon nitride. Thereby, in one embodiment step b) comprises depositing the dielectric precoat with a thickness d for which there is valid:
200 nm≦d≦500 nm.
Still in a further embodiment the addressed thickness d is selected to be
200 nm≦d≦400 nm.
Thereby, it must be considered that the higher this thickness d is selected, the higher will be the risk of precoat material peeling off. Increasing the addressed thickness above the addressed 500 nm will further lead to a decrease in total throughput of the manufacturing equipment due, on one hand to long precoat times and, on the other hand to reduced RF power coupling through such precoat. Coating thicknesses below 200 nm on the other hand have proven to be inefficient to improve the targeted thickness and quality uniformity sufficiently, i.e. to be below a standard specification requiring thickness variation e.g. of less than 10% of an average thickness value considered along the overall substrate surface.
In a further embodiment of the method according to the invention step b) comprises depositing the dielectric precoat as a layer of amorphous material. Thereby, the stress within the precoat layer is significantly reduced. This leads to an improved adherence of such precoat layer to the vacuum reactor metallic inner surface.
In a further embodiment the material of the precoat deposited in step b) and of the dielectric layer deposited in step d) are equal at least considering their composition. Thereby, they need not be equal in structure. Thus, as was addressed the precoat may be deposited as amorphous material which is not done necessarily for the dielectric layer deposited in step d).
Although not always mandatory, in most cases a further embodiment of the present invention comprises plasma cleaning at least the addressed parts of the inner surface of the reactor which are afterwards coated with the dielectric precoat, thus performing such cleaning before performing step b).
In an embodiment the addressed cleaning is performed in plasma-activated SF6 and oxygen.
The method according to the present invention with all embodiments addressed is highly suited for manufacturing thin film transistor display substrates or liquid crystal display substrates or solar cell substrates or organic light-emitting display panels. Thereby, at such substrates the addressed μc-Si layer becomes the intrinsic Si-layer of semiconductor devices, which as perfectly clear to the skilled artisan, are realized following up step e) by additional layers deposited, preferably in the same reactor and without vacuum interruption. Therefore, it has to be emphasized that between the addressed step e) and the addressed step f) one or more than one additional treating steps may be performed before manufacturing of the flat substrate is terminated and before step f) is performed for a subsequent single substrate to be manufactured.
The present invention shall now further be exemplified to the skilled artisan with the help of additional explanations and examples as well as with the help of figures. The figures show:
a and 6b the real and imaginary components of a pseudo-dielectric function which is representative of layer thickness and material structure and as measured at a customarily deposited μc-Si layer, in a center and in an edge area of the substrate, and
a and 7b real and imaginary components of the pseudo-dielectric function according to
Additional members which are mandatory for a vacuum reactor as e.g. vacuum pumping port, input/output loadlock etc., are not shown in the schematical
With an eye on the steps of generating an RF plasma discharge in the reaction space between the addressed electrodes and inletting a respective gas or gas mixture, thereby depositing the addressed dielectric precoat, this should not be understood as being done in the time sequence according to the sequence of wording and mentioning such steps here. E.g. it may be possible to first inlet the gas or gas mixture to the reaction space and then to establish the RF discharge or to establish the RF discharge and then to inlet the gas or gas mixture. Important is that deposition of the precoat starts when both conditions are met, namely RF discharge established and gas or gas mixture present in the reaction space R.
The dielectric precoat 13 must be applied along at least the peripheral area of the substrate carrier electrode 11 and along the surface areas of the wall 3, which laterally surround the edge of the substrate carrier electrode 11.
Especially when making use of a vacuum reactor type as shown in
200 nm≦d≦500 nm
and thereby in a today's embodiment
200 nm≦d≦400 nm.
PECVD processing is thereby controlled so as to lead in today's embodiment to an amorphous material structure of the addressed precoat 13.
Today, before every such precoating, there is performed RF plasma reactive cleaning of the metallic inner surface of wall 3 e.g. making use of a gas mixture comprising SF6—O2. Nevertheless, and as addressed later, such cleaning before each precoating step—which latter is mandatory before each single substrate treatment—may be omitted and replaced e.g. by a cleaning step after each third substrate or even just when it becomes necessary.
With an eye on the parts of the wall 3 which are, according to the present invention, to be covered by the dielectric precoat, the parts as shown at 13 in
In
The substrate 15 is in today's manufacturing of glass.
Note that already when depositing the dielectric layer 17 upon the substrate 15 there is substantially no metallic surface of wall 3 which may be seen from the reaction space 3 adjacent to the substrate 15 to be coated. Therefore, no such metallic surface may influence the etching to deposition equilibrium when depositing the dielectric layer 17 upon the substrate 15. Thereby, already the dielectric layer 17 is deposited with a substantially uniform thickness and with a substantial uniform material characteristics all along the surface 15o of the substrate 15.
After having performed dielectric layer deposition on the substrate 15 upon the addressed dielectric layer 17 a silicon layer of μc-Si is PECVD deposited. This μc-Si layer is shown in
For depositing the Si layer 19 as a μc-Si layer etching radicals are exploited. In order to achieve a fully crystallized material structure of layer 19, it is grown from a gas or gas mixture in today's embodiment from SiF4, which produces growth-contributing radicals (silicon-containing radicals) as well as etching radicals (fluorine-containing radicals). Although the growth mechanism is not fully understood it is known that the growth is mainly governed by the balance or equilibrium between etching radicals and deposition radicals (see P. Roca i Cabaroca et al.) Different radicals/surface interactions may occur when a metallic surface, as an aluminum magnesium surface of the reactor wall—including electrode surfaces—is exposed to the activated gas on one hand and, on the other hand, substrate material, as glass, is exposed to the same plasma-activated gas. As in a standard PECVD reactor e.g. of the type as shown in
It has further to be noted that the precoat step, the deposition step for the dielectric layer on the substrate, the deposition step for the μc-Si layer as well as the deposition steps for subsequent layers before repeating all the addressed steps for a further single substrate, are advantageously performed in one and the same vacuum reactor.
In context of
Bottom gate (BG) μc-Si:H TFTs were produced at 200° C. to about 330° C. First, the reactor wall was cleaned as by plasma cleaning with the SF6/O2 gas mixture. Then the reactor wall was coated with a low-stress SiN-layer. Thereby, as already addressed, a thickness optimum for the precoating is between 200 nm and about 500 nm.
In a Unaxis KAI 1XL PECVD system electrically supplied at an RF frequency of 13.56 MHz, 110 sccm of SiH4, 800 sccm of NH3, 800 sccm of N2 and 500 sccm of H2 were injected. The process pressure was 0.75 mbar and RF power 850 W for the electrode surface of 5000 cm2. The substrate was again a 720 mm×650 mm glass substrate (substrate surface about 4700 cm2). The addressed parameters led to deposition of amorphous silicon nitride having a low internal stress as a precoat dielectric layer 13 according to
The first layer was made of the same dielectric material, thus silicon nitride, as the precoating layer. The typical thickness for this dielectric layer deposited on the substrate is between 200 nm and 500 nm too. For a thickness of the dielectric layer applied to the substrate, layer 17 of
After depositing the dielectric layer on the substrate, a μc-Si layer with a thickness of about 150 nm was deposited from a plasma-activated gas-mixture comprising to a greater part a fluorine rich silicon-containing gas, as SiF4, hydrogen and an inert gas, typically Ar with a ratio of fluorine rich silicon-containing gas to hydrogen in the range from 1:1 to about 1:10 and with a ratio of fluorine rich gas to inert gas from 1:10 to about 1:30. The μc-Si layer was deposited at a rate of approx. 0.1 nm per second.
Then the layer stack was completed in order to achieve a functional TFT device structure, i.e. by deposition of a n+ a-Si:H contact layer with a thickness of about 30 nm. If required an additional a-Si:H material layer may be added prior to the ohmic contact layer deposition, in order to avoid leakage of current when a respective transistor device is in its off state (Ioff). Due to a potential higher deposition rate, the completion of the intrinsic layer of μc-Si with a standard amorphous silicon layer may improve the average machine throughput.
After deposition of a precoating layer as was explained in context with example 1 the glass substrate is transferred into the PECVD vacuum reactor and a TFT layer stack is deposited starting with a different dielectric layer. Thereby, 80 sccm SiH4, 350 sccm NH3, 620 N2 and 1200 sccm H2 were injected. The process pressure was 0.6 mbar at an RF power of 750 W. Then a μc-Si layer with a thickness of about 150 nm was deposited from a gas mixture comprising to a greater part a fluorine rich silicon-containing gas, SiF4, a silicon rich gas, SiH4, hydrogen and an inert gas, Ar, with a ratio for the considered example of fluorine rich to silicon rich to hydrogen to noble gas of 10:1:100:400. The layer stack was completed in order to achieve functional TFT, by deposition of a n+ a-Si:H contact layer with a thickness of about 30 nm. All the parameters not specified here were as in example 1.
BG μc-Si:H TFTs were produced at 200° C. to about 330° C. First and after cleaning, the reactor wall was precoated with a dielectric precoat with a thickness of 400 nm. The precoat layer was of fluorinated silicon oxide. In an equipment used for the examples 1 and 2, such a layer was deposited using 100 sccm of SiH4, 100 sccm of SiF4, 400 sccm of N2O and 100 sccm of Ar at a process pressure of 0.5 mbar and an RF power of 200 W. The substrate extent was as addressed in the examples 1 and 2. These parameters led to a deposition of amorphous fluorinated silicon oxide with low internal stress. After having deposited such precoat layer the glass substrate was transferred into the reaction chamber and a TFT stack was deposited without interrupting the vacuum in the PECVD reactor and with techniques and recipes known in the art. E.g the first layer, the dielectric layer, was deposited by PECVD with the following recipe: 80 sccm of SiH4, 350 sccm of NH3, 620 sccm of N2 and 1200 sccm of H2 at a process pressure of 0.6 mbar and an RF power of 750 W. Then a μc-Si layer with a thickness of about 120 nm was deposited from a gas mixture comprising in a predominant part a fluorine rich silicon gas as SiF4, hydrogen and an inert gas, Ar, with a SiF4/H2 ratio in the range from 1:1:1:10 and a SiF4/inert gas ratio from 1:10 to about 1:30. The μc-Si layer was deposited at a deposition rate of approx. 0.1 nm per second at a reactor temperature of 230° C. Then the layer stack was completed in order to achieve functional TFT, i.e. by deposition of a n+ a-Si:H contact layer of about 30 nm.
By applying the dielectric precoat to the inner surface of the vacuum reactor both layers applied to the substrate, namely the dielectric layer as well as the μc-Si layer, showed enhanced thickness uniformity over the large-area substrate of about 4700 cm2 and did not show any of the common location-dependent properties which were observed under exactly the same processing conditions, but without providing the precoat.
Thus, besides of providing the precoat all processing parameters were kept the same as for processing the substrate leading to the
Further, by providing the dielectric precoat according to the present invention uniformity of the layer quality along the substrate is greatly improved. Such quality uniformity of a layer can be checked by using spectroscopic ellipsometry at different locations of the coated substrate. Spectroscopic ellipsometry is a non-destructive optical technique used to determine the optical properties of substrates with thin films. Using an appropriate model to fit experimental measurements, it allows to have access to structural information such as thickness of layers, type of material, crystalline fraction, roughness, etc.
In order to have insight to film quality, spectroscopic ellipsometry was performed on the μc-Si film deposited on a 400 nm SiN layer on the substrate. The μc-Si layer was grown from a SiF4, H2 and Ar gas mixture to an average thickness of about 120 nm. The SiN dielectric layer and the μc-Si layer were deposited on a glass substrate, again with a KAI 1XL PECVD reactor. Before transferring the glass substrate into the vacuum reactor the reactor was cleaned with plasma-activated SF6/O2 and then precoated with a 400 nm SiN layer of the same type as afterwards deposited on the substrate.
For comparison purposes the same substrate coating with the same processing parameters was performed but without precoating the inner wall of the PECVD reactor.
a, b and 7a, b represent imaginary (Ei) and real (Er) parts of a pseudo-dielectric function of the two μc-Si films deposited as was just addressed, i.e. without (
The complex dielectric function E(w)=Er(w)+iEi(w) describes the response of a medium to a static or oscillating electric field. It is a function of frequency w and thus of energy of the electric field as applied. The dielectric function can be considered as a fingerprint of the material investigated with respect to its reaction on electric field exposure. Therefore, the dielectric function is very dependent on the type of material and on material structure. Materials of different composition will also exhibit different dielectric functions.
In the present case the dielectric function is called “pseudo” for two reasons:
First, because it is not directly measured, but calculated from an optical index that is directly derived from ellipsometric spectra.
Secondly, it can be defined for a stack of material while the dielectric function per se is considered as directed on the intrinsic property of one material. Two stacks of materials having different thicknesses will also appear to have different pseudo-dielectric functions. Moreover, in the pseudo-dielectric function there is some interference that can also reveal a thickness variation which, with an eye on
Behavior of the pseudo-dielectric function is displayed in both
It has been shown that the method according to the present invention leads to improved thickness and quality uniformity of a μc-Si layer deposited on large-surface substrates.
It has further to be noted that even if, after processing a substrate according to the present invention and before processing the next substrate, the inner surface of the reactor is not cleaned each time, but there is just applied the precoating, the precoating will bind loose particles to the wall of the reactor which will be integrated into the precoat material, thereby preventing such particles from flaking off. Thus, by precoating the wall of the reactor before each single substrate processing contributes to establishing a “clean” reactor wall without necessarily having to perform a specific cleaning step.
| Number | Name | Date | Kind |
|---|---|---|---|
| 5177578 | Kakinoki et al. | Jan 1993 | A |
| 5811195 | Bercaw | Sep 1998 | A |
| 5863602 | Watanabe et al. | Jan 1999 | A |
| 5970383 | Lee | Oct 1999 | A |
| 5981899 | Perrin et al. | Nov 1999 | A |
| 6071573 | Koemtzopoulos et al. | Jun 2000 | A |
| 6530992 | Yang et al. | Mar 2003 | B1 |
| 6790793 | Nishino et al. | Sep 2004 | B2 |
| 6828254 | Han et al. | Dec 2004 | B2 |
| 6881684 | Aota et al. | Apr 2005 | B2 |
| 6902629 | Zheng et al. | Jun 2005 | B2 |
| 6974781 | Timmermans et al. | Dec 2005 | B2 |
| 7204913 | Singh et al. | Apr 2007 | B1 |
| 20050214454 | Yang et al. | Sep 2005 | A1 |
| 20050257747 | Wakabayashi et al. | Nov 2005 | A1 |
| Number | Date | Country |
|---|---|---|
| 2002-289557 | Apr 2002 | JP |
| Number | Date | Country | |
|---|---|---|---|
| 20070004220 A1 | Jan 2007 | US |
| Number | Date | Country | |
|---|---|---|---|
| 60695111 | Jun 2005 | US |
