The invention relates generally to the field of semiconductor manufacturing and more particularly to the fabrication of SiN nanodots in non-volatile memory devices.
The size of semiconductor memory devices, and the thin dielectric films that comprise these devices, has been scaled down over the past years in order to fulfill the requirements of diminishing feature size. Further scaling down presents serious problems. One large problem, in floating gate memory devices, is that the charge retention characteristics of the devices are very sensitive to the presence or absence of defects in the tunnel oxide. A reduction in the thickness of the tunnel oxide, to allow for further scaling down, increases the risk of defects occurring in the tunnel oxide.
One partial solution to this problem is to store the charge in a floating gate comprised of a plurality of nanodots, instead of a monolithic floating gate. This plurality strongly reduces the sensitivity of the device to incidental defects in the tunnel oxide. In such a situation, when a defect is present, a nanodot immediately adjacent the defect might lose its charge but the other nanodots are not affected. See, e.g., U.S. Pat. No. 6,165,842 to Shin et al., issued Dec. 26, 2000; U.S. Pat. No. 5,714,766 to Chen et al., issued Feb. 3, 1998.
Nanodots made from silicon have been useful in making such gates. In recent years, a significant amount of research has been done in the area of silicon nanodots. Silicon nanodots are generally produced by low pressure chemical vapor deposition (LPCVD) and the process conditions can be readily optimized to produce silicon nanodots of a desired size and with a desired density. See, e.g., Baron et al., “Low pressure chemical vapor deposition growth of silicon quantum dots on insulators for nanoelectronic devices,” APPLIED SURFACE SCIENCE, Vol. 164, pp. 29–34 (2000).
More recently, one group has investigated the option of blended, or heterogeneous, silicon nitride nanodots. See, e.g., Koga et al., “Silicon Single-Electron Memory and Logic Devices For Room Temperature Operation,” IEDM 01–143, pp. 7.1.1–7.1.4 (2001). Silicon nitride nanodot devices show superior memory characteristics over silicon nanodot devices. Data retention characteristics are better and write characteristics show that a large memory window is attainable in each SiN dot device because multiple traps are formed in each SiN dot. Another advantage is that the silicon nitride dots are not as susceptible to oxidation. This can be important, for instance, in the case of a doubly stacked, floating dot memory device, where an upper dot is formed over a silicon layer. In the creation of such a device, the silicon layer is mostly oxidized, except for a lower silicon dot where the lower area is shielded from oxidation in a self-aligned manner by the upper dot. However, when the upper dot is formed of pure silicon, the upper dot will partially oxidize during the process and thus, size control of the lower dot is difficult. Using a silicon nitride dot as the upper dot largely eliminates this problem, as silicon nitride is very resistant to oxidation.
The current practice for the creation of these silicon nitride dots is to use low pressure chemical vapor deposition. However, the manufacturing of silicon nitride dots by low pressure chemical vapor deposition (LPCVD) has several disadvantages. First, the deposition temperature for the silicon nitride LPCVD process is significantly higher than the deposition temperature for the silicon process, resulting in higher consumption of the thermal budget of the wafer, which is undesired. Second, the LPCVD seed conditions that produce SiN nanodots of a desired size and density are largely unknown and still need to be explored. Finally, the visibility of silicon nitride nanodots on a silicon oxide film, either with scanning electron microscopy (SEM) or with transmission electron microscopy (TEM), is rather poor, as compared to the visibility of silicon nanodots, thus making the silicon nitride dots hard to examine.
The present embodiments have several advantages, these advantages include avoiding the above disadvantages and providing a method of manufacturing silicon nitride nanodots that can easily be implemented in a production environment.
In accordance with one aspect of the present invention, a method is provided for producing silicon nitride nanodots. Silicon nanodots are produced, followed by nitridation of the silicon nanodots by a nitrogen source gas. According to a preferred embodiment, the silicon nanodots are formed by a chemical vapor deposition (CVD) seed technique. According to a further aspect of the invention, nitridation of the silicon nanodots is carried out by exposing them to nitrogen radicals.
In accordance with another aspect of the invention, a non-volatile memory device is provided. The device comprises a semiconductor substrate and a plurality of silicon nitride nanodots across an area of the substrate. The nanodots have an average size between about 1 nm and 100 nm and cover between about 10% and 50% of the area of the substrate.
In accordance with another embodiment, a non-volatile memory device is disclosed. The device comprises a semiconductor substrate, and a number of nanodots on the substrate, wherein at least some of the nanodots comprise a core region having a nitrogen concentration lower than a surface region of the at least some of the nanodots.
A sequence of process steps according to a preferred embodiment is schematically shown in
In
In one embodiment, the silicon dots can be formed by the LPCVD seed technique using SiH4 as the source gas. The nanodots for sample 1, in
The selected temperature for the formation of the nanodots is preferably between 450° C. and 650° C., more preferably between 500° C. and 620° C., and most preferably between 530° C. and 550° C. When relatively high temperatures are selected, polycrystalline nuclei are formed and when relatively low temperatures are selected, amorphous nuclei are formed. The advantage of amorphous nuclei is that their surface topography is smoother. Consequently, the LPCVD seed temperature is preferably selected such that the formed nuclei are amorphous and remain amorphous throughout the entire LPCVD seed stage.
The selected partial pressure of the silicon source gas is preferably between 0.01 mTorr and 1 Torr, more preferably between 1 mTorr and 100 mTorr and even more preferably between 1 mTorr and 10 mTorr. An exemplary partial pressure is between 5 mTorr and 6 mTorr.
The selected silicon source gas flow depends on the size of the system used. For example, for a batch furnace with a load size of about 100 semiconductor substrates of 200 mm diameter, the silicon source gas flow is preferably between about 0.1 sccm and about 1000 sccm, more preferably between about 1 sccm and about 100 sccm, and even more preferably between about 10 and about 30 sccm. It would be a matter of routine operation for one of skill in the art to determine the desired gas flow. Optionally, an inert carrier gas can be used. It will be clear that by the use of an inert carrier gas, such as nitrogen, while keeping the partial pressure of the silicon source gas constant, the process pressure will increase. In one embodiment, the selected inert gas flow is preferably between about 0.1 slm and about 10 slm. The selected silicon source gas exposure time is preferably between about 0.1 min and about 60 min, more preferably between about 0.5 min and about 30 min, and even more preferably between about 1 min and about 15 min.
In one embodiment, the average size or diameter of nanodots formed is preferably in the range of 1 nm to 100 nm, more preferably between 2 nm and 20 nm, and most preferably between 5 nm and 7 nm. Preferably between 10% and 50% of the substrate surface area is covered with the nanodots, more preferably between 20% and 40% and most preferably about 30%. In one embodiment, the nanodots are approximately hemispherical. In one embodiment, monosilane (SiH4) has been found a suitable source gas, although other gases like disilane (Si2H6) and trisilane (Si3H8) can alternatively be used. In another embodiment, chlorosilanes (SiHxCl(4-x), x=0–4) can alternatively be used, although they will give rise to higher process temperatures.
The nitridation is performed by exposure of the silicon nanodots to a nitrogen containing gas such as, but not limited to N2, nitrogen hydrides (e.g., NH3 and N2H4) and nitrogen halides (e.g., NF3, NCl3 and NI3). In a more preferred embodiment, the nitrogen containing gas comprises nitrogen radicals or atomic nitrogen. Preferably, N2 is used as the source gas for creating the nitrogen radicals. In one embodiment, the nitrogen radicals are generated outside the process chamber, by a nitrogen radical generator in communication with the process chamber.
One advantage of the use of nitrogen radicals, particularly neutral atomic nitrogen, is that they are reactive enough to allow nitridation at temperatures below the crystallization temperature of amorphous silicon, preserving the smooth surface topography of the amorphous film. Thus, the resultant silicon nitride nanodots preferably exhibit a similar smooth topography. This smoothness reduces high field effects at rough edges, effectively increasing the breakdown voltage.
Another advantage is that the fact that known nucleation behavior of silicon LPCVD determines the size and distribution of the nanodots. Advantageously, such conditions are already established in the art and need not be optimized separately for direct silicon nitride deposition, while the advantages of silicon nitride nanodots are nevertheless obtained. This is particularly advantageous because it is difficult to evaluate the distribution of silicon nitride nanodots, since the visibility of silicon nitride is poor with either SEM or TEM technology. Thus, the nanodots have a size and distribution characteristic of the silicon LPCVD, set forth above.
Nitridation of blanket amorphous silicon films by nitrogen radicals resulted in a nitride film thickness of 2.2 nm, as measured by spectroscopic ellipsometry, independent of the nitridation temperature. This indicates that the nitridation process is self-limiting under the preferred conditions, similar to what is known for thermally activated nitridation. The nitridation temperature is preferably selected between 200° C. and 800° C., more preferably between 300° C. and 650° C., and most preferably between 400° C. and 500° C.
In one embodiment, the silicon nanodots are exposed to the nitrogen for 5 seconds or more. In a more preferred embodiment, the silicon nanodots are exposed to nitrogen for 5 seconds to 5 minutes. In a more preferred embodiment, the silicon nanodots are exposed to the nitrogen for greater than 30 seconds, most preferably greater than about 1 minute. In one preferred embodiment, the silicon nitride nanodots are exposed to the nitrogen source until all of the silicon nanodots are converted to silicon nitride.
As will be appreciated by one of skill in the art, the reaction chambers wherein the LPCVD seed process and the nitridation are performed can be of different designs. In the experiments that resulted in the examples presented herein, the LPCVD seed process was performed in a vertical, hot-wall batch furnace designed for a batch size of 150 wafers with a diameter of 200 mm. The nitridation was performed in a cold-wall single wafer reactor operating at atmospheric pressure, and the reactor was provided with a remote nitrogen radical generator. However, both processes can be performed in a single wafer reactor or in a batch reactor. When nitridation is performed in a batch reactor, care should be taken that the radicals live long enough to reach all substrates and all parts of each substrate. The nitridation process can also be performed in-situ, inside the LPCVD reactor, avoiding the need for wafer transport between the two process steps.
Silicon substrates were subjected to an ultra diluted wet oxidation process as described in U.S. patent application Ser. No. 09/723,323 of applicant to form a 2 nm thick layer of oxide. The low pressure chemical vapor deposition “LPCVD” seed process was carried out in a LPCVD vertical furnace from ASM, model A400™-XT, (ASM International N.V., Bilthoven, The Netherlands), designed for 200 mm wafers with a usable load size of 150 wafers. The temperature was 540° C., the pressure 550 mTorr, SiH4 flow 20 sccm, nitrogen carrier flow 1 slm and exposure time 10 min.
The surface of this exemplary sample, after silicon nanodot formation and prior to nitridation, was studied by scanning electron microscopy (SEM) and is shown in
The nitridation was performed in an Epsilon® cold wall single wafer reactor, (ASM America, Inc., Phoenix, Ariz.). The temperature was 450° C., the pressure was 1 Torr, the nitrogen flow was between about 10 slm and about 16 slm. A microwave radical generator, commercially available under the trade name TRW-850 (Rapid Reactive Radicals Technology (R3T) GmbH, Munich, Germany), applying a power of 1000 Watts, was used to generate the nitrogen radicals. The nitrogen radical exposure time was 5 minutes.
Having thus described the preferred embodiments, those of skill in the art will readily appreciate that other embodiments may be made and used within the scope of the claims hereto attached. Numerous advantageous of the embodiments covered by this document have been set forth in the foregoing description. It will be understood, however, that this disclosure is, in many respects, only illustrative. Changes may be made in details, particularly in matters of shape, size, and arrangement of parts without exceeding the scope of the invention.
This application claims the priority benefit under 35 U.S.C. § 119(e) of Provisional Application No. 60/435,314, filed Dec. 18, 2002.
Number | Name | Date | Kind |
---|---|---|---|
4262631 | Kubacki | Apr 1981 | A |
4277320 | Beguwala et al. | Jul 1981 | A |
4298629 | Nozaki et al. | Nov 1981 | A |
4363828 | Brodsky et al. | Dec 1982 | A |
4495218 | Azuma et al. | Jan 1985 | A |
4585671 | Kitagawa et al. | Apr 1986 | A |
4684542 | Jasinski et al. | Aug 1987 | A |
4715937 | Moslehi et al. | Dec 1987 | A |
4834020 | Bartholomew | May 1989 | A |
4851095 | Scobey et al. | Jul 1989 | A |
5112773 | Tuttle | May 1992 | A |
5227329 | Kobayashi et al. | Jul 1993 | A |
5389398 | Suzuki et al. | Feb 1995 | A |
5389570 | Shiozawa | Feb 1995 | A |
5418180 | Brown | May 1995 | A |
5453858 | Yamazaki | Sep 1995 | A |
5607724 | Beinglass et al. | Mar 1997 | A |
5612558 | Harshfield | Mar 1997 | A |
5614257 | Beinglass et al. | Mar 1997 | A |
5648293 | Hayama et al. | Jul 1997 | A |
5656531 | Thakur et al. | Aug 1997 | A |
5695819 | Beinglass et al. | Dec 1997 | A |
5698771 | Shields et al. | Dec 1997 | A |
5700520 | Beinglass et al. | Dec 1997 | A |
5714766 | Chen et al. | Feb 1998 | A |
5786027 | Rolfson | Jul 1998 | A |
5789030 | Rolfson | Aug 1998 | A |
5837580 | Thakur et al. | Nov 1998 | A |
5874129 | Beinglass et al. | Feb 1999 | A |
5876797 | Beinglass et al. | Mar 1999 | A |
5885869 | Turner et al. | Mar 1999 | A |
5907792 | Droopad et al. | May 1999 | A |
5916365 | Sherman | Jun 1999 | A |
6015590 | Suntola et al. | Jan 2000 | A |
6027705 | Kitsuno et al. | Feb 2000 | A |
6083810 | Obeng et al. | Jul 2000 | A |
6087229 | Aronowitz et al. | Jul 2000 | A |
6136654 | Kraft et al. | Oct 2000 | A |
6159828 | Ping et al. | Dec 2000 | A |
6165842 | Shin et al. | Dec 2000 | A |
6171662 | Nakao | Jan 2001 | B1 |
6197669 | Twu et al. | Mar 2001 | B1 |
6197694 | Beinglass | Mar 2001 | B1 |
6200893 | Sneh | Mar 2001 | B1 |
6228181 | Yamamoto et al. | May 2001 | B1 |
6252295 | Cote et al. | Jun 2001 | B1 |
6271054 | Ballantine et al. | Aug 2001 | B1 |
6326311 | Ueda et al. | Dec 2001 | B1 |
6391803 | Kim et al. | May 2002 | B1 |
6455892 | Okuno et al. | Sep 2002 | B1 |
6468924 | Lee et al. | Oct 2002 | B1 |
6503846 | Niimi et al. | Jan 2003 | B1 |
6528530 | Zeitlin et al. | Mar 2003 | B1 |
6537910 | Burke et al. | Mar 2003 | B1 |
6613695 | Pomarede et al. | Sep 2003 | B1 |
6638879 | Hsieh et al. | Oct 2003 | B1 |
6724038 | Mikolajick | Apr 2004 | B1 |
6831315 | Raaijmakers et al. | Dec 2004 | B1 |
20020047151 | Kim et al. | Apr 2002 | A1 |
20020073925 | Noble et al. | Jun 2002 | A1 |
20020098627 | Pomerede et al. | Jul 2002 | A1 |
20020168868 | Todd | Nov 2002 | A1 |
20020197831 | Todd et al. | Dec 2002 | A1 |
20030022528 | Todd | Jan 2003 | A1 |
20030059535 | Luo et al. | Mar 2003 | A1 |
20030082300 | Todd et al. | May 2003 | A1 |
Number | Date | Country |
---|---|---|
1 104 012 | May 2001 | EP |
57209810 | Dec 1982 | JP |
59078918 | May 1984 | JP |
59078919 | May 1984 | JP |
60043485 | Mar 1985 | JP |
S60-43485 | Mar 1985 | JP |
61153277 | Jul 1986 | JP |
62076612 | Apr 1987 | JP |
63003414 | Jan 1988 | JP |
63003463 | Jan 1988 | JP |
01217956 | Aug 1989 | JP |
01268064 | Oct 1989 | JP |
02155225 | Jun 1990 | JP |
H 02-155225 | Jun 1990 | JP |
03091239 | Apr 1991 | JP |
H3-91239 | Apr 1991 | JP |
03185817 | Aug 1991 | JP |
03187215 | Aug 1991 | JP |
H3-185817 | Aug 1991 | JP |
H3-187215 | Aug 1991 | JP |
03292741 | Dec 1991 | JP |
04323834 | Nov 1992 | JP |
05021378 | Jan 1993 | JP |
05062911 | Mar 1993 | JP |
07249618 | Sep 1995 | JP |
08242006 | Sep 1996 | JP |
Number | Date | Country | |
---|---|---|---|
20040224534 A1 | Nov 2004 | US |
Number | Date | Country | |
---|---|---|---|
60435314 | Dec 2002 | US |