Method of etching silicon nitride

Abstract

A method of etching an underlying inorganic substrate through a patterned photoresist, including exposing a structure comprising said inorganic substrate and patterned photoresist to a plasma etchant generated from a plasma source gas including at least one fluorine-comprising gas and sulfur dioxide (SO2). The amount of sulfur dioxide present in said plasma source gas may be varied during the etching process. The method is particularly useful when the photoresist is a DUV photoresist. One of the preferred embodiments of the method is the etching of silicon nitride (SiNx) through a DUV photoresist, where the plasma source gas used to provide the etchant includes at least one fluorine-comprising gas, argon, and sulfur dioxide. Other preferred fluorine-comprising gases include nitrogen trifluoride (NF3), carbon tetrafluoride (CF4), and sulfur hexafluoride (SF6).

Description

STATEMENT OF RELATED APPLICATION

[0001] This application is a continuation-in-part of U.S. patent application Ser. No. 09/291,682, filed Apr. 13, 1999, which is currently pending. This application is related to co-pending U.S. patent application Ser. No. 08/907,448, which was filed on Aug. 8, 1997.

BACKGROUND OF THE INVENTION

[0002] 1. Field of the Invention

[0003] The present invention relates generally to a fabrication process for manufacturing semiconductor devices. More particularly, the present invention relates to a method and etchant for etching silicon nitride (SiNx) when the photoresist used to pattern the silicon nitride is a deep UV photoresist.

[0004] 2. Background

[0005] Subtractive dry etching is conventionally used to pattern silicon nitride for use in semiconductor devices. Conventional dry etching (plasma etching) processes include physical bombardment by inert ions, chemically reactive ion etching, and combinations thereof. Dry etching generally involves exposing the SiNx to a gaseous plasma created by RF excitation of a plasma source gas in a vacuum system. The RF excitation generates ions, free radicals and other excited species from the source gas. The SiNx surface is contacted by the excited species, which typically are designed to chemically react with as well as physically bombard the SiNx surface. Plasma etching relies on diffusion to transport the excited species to the surface being etched, and is generally isotropic. In order to provide anisotropic etching, it is necessary to develop a voltage between the plasma and the substrate surface to be etched, to accelerate excited species (particularly ions) toward the substrate surface. Because of this acceleration, it is possible to achieve anisotropic etching, whereby the etching proceeds at an appreciable rate in a direction normal to the surface of a wafer, but does not proceed at an appreciable rate in lateral directions. However, unless process parameters are carefully controlled, etching in the lateral direction, also referred to as sidewall attack, may occur. Depending upon the particular device or feature being fabricated, some deviation from a perfect anisotropic etch may be tolerable, although it is usually desirable to minimize such deviations. For example, an etched feature that has a sidewall included angle (between the sidewall surface and the horizontal surface of the substrate beneath the sidewall) as low as 80° instead of the ideal 90° (where the sidewall is perpendicular to the surface beneath the sidewall), may be acceptable in some instances, but is not desirable. Such an undercut from top toward bottom of the sidewall can cause problems during subsequent processing steps. For a more detailed description of dry etching processes, see Runyan and Bean, Semiconductor Integrated Circuit Processing Technology, Ch. 6.3, pp. 269-280 (1990), which is incorporated herein by reference.

[0006] A layer of SiNx may be patterned during a dry etching process by protecting portions of the layer surface from the etchant, while exposing other portions, using a mask. These masks are frequently made of photoresist which is typically a polymeric material, and most typically an organic-based material. “Hard masks”, typically comprising a metal or a silicon-comprising compound, or other inorganic materials, may also be used to selectively expose SiNx to plasma. However, it is generally more difficult and more expensive to pattern very small features into a hard mask than to pattern such features into a photoresist. As a result, photoresist masks are preferable for many applications.

[0007] However, the plasma etchant etches the organic-based photoresist masking material simultaneously with the substrate exposed through the mask. Photoresist erosion (etching away) may undesirably lead to punch through. For example, during the etching of SiNx, there may be exposure of mask protected SiNx surfaces prior to complete etching through of SiNx surfaces which are intentionally exposed through the mask. This problem is exacerbated with anisotropic etching processes, which generally etch away photoresist at a faster rate than isotropic etching processes. In addition, an anisotropic etching process generally reduces the selectivity for SiNx relative to photoresist, where the selectivity is the ratio of the etch rate of the SiNx, to the etch rate of the photoresist. While the possibility of punch through of the photoresist may be reduced by increasing photoresist thickness, it is preferable to use a photoresist mask that is as thin as possible for the reasons previously given. The minimum photoresist thickness that may be used without punch through depends upon the thickness of SiNx being etched, and the selectivity of the etching process.

[0008] When an I-line photoresist is used as a patterned mask for the etching of the SiNx layer, the photoresist layer can be relatively thick (as thick as 12,000 to 13,000 Å). However, the minimum feature size for an I-line photoresist is in the range of about 2500 Å. To obtain smaller feature sizes, in the range of about 1500 Å, it is presently necessary to use a Deep UV (DUV) photoresist as the patterned masking material for the etching of the SiNx. This requires a reduction in the thickness of the photoresist layer, since DUV photoresist thickness is typically limited to less than about 7000 Å for optical reasons. The reduced thickness in the DUV photoresist layer makes etch selectivity of the SiNx in preference to the photoresist more critical. When an I-line photoresist is used, a reduction in feature size from about 10,000 Å to about 2500 Å results in a decrease in the selectivity for SiNx in preference to the I-line photoresist. For example, the typical selectivity of SiNx:I-line for a feature size of about 10,000 Å ranges from about 2:1 to about 3:1; while the selectivity for a feature size of about 2500 Å ranges from about 1:1 to about 2:1. In addition, when the feature size is reduced from about 2500 Å to about 1500 Å, the etch rate for the SiNx typically decreases because of diffusion limitations pertaining to the removal of etch reaction by-products from the smaller openings to the smaller features.

[0009] An illustrative example of the use of an I-line photoresist in the patterned etching of SiNx is provided in the PCT published application WO 96/16437, published May 30, 1996, and titled: Silicon Nitride Etch Process With Critical Dimension Gain. The etched SiNx feature size is about 5,000 Å, and the plasma source gas used to etch the silicon nitride is a mixture of CHF3 and O2, where the ratio of CHF3:O2 is at least 6:1.

[0010] To be able to use a DUV photoresist, it is necessary to be able to maintain an adequate selectivity to prevent punch through of the thin photoresist and to be able to obtain an acceptable etch rate for the SiNx.

SUMMARY OF THE INVENTION

[0011] We have discovered a method of providing improved selectivity for an underlying layer during plasma etching of that underlying layer through an organic-based photoresist. In particular, when the plasma source gas includes a fluorine containing compound, sulfur dioxide (SO2) is added to the plasma source gas. The method is particularly advantageous when the organic-based photoresist is a DUV photoresist and when the etched feature size is less than about 2500 Å. The method is especially useful for anisotropic plasma etch processes, but is useful for isotropic etching processes as well.

[0012] A particularly preferred embodiment is the anisotropic etching of silicon nitride (SiNx) through a patterned, DUV, organic-based photoresist mask. The plasma source gas includes a fluorine-comprising compound and SO2. Preferably, the plasma source gas includes argon (Ar) or another inert gas which provides a good source for ions useful in the anisotropic etching process. Preferred fluorine-comprising compounds are inorganic and include nitrogen trifluoride (NF3) and sulfur hexafluoride (SF6), by way of example and not by way of limitation. In addition, hydrogen bromide (HBr) may be added to the plasma source gas to assist in profile control during etching of the silicon nitride feature.

BRIEF DESCRIPTION OF THE DRAWINGS

[0013] FIG. 1 shows a trench etched in SiNx.

[0014] FIG. 2 shows a preferred apparatus for practicing the present invention.

[0015] FIG. 3 shows a cross-section of a preferred apparatus for practicing the present invention.

[0016] FIG. 4 shows a preferred apparatus for practicing the present invention that incorporates a computer for purposes of process control.

DETAILED DESCRIPTION OF THE INVENTION

[0017] Conventional anisotropic etching processes typically include a chemically reactive etchant species, frequently generated from at least one halogen species, and preferably generated from a fluorine-containing compound. Inert gases such as helium, argon, krypton, and xenon may also be used to provide ion bombardment in an anisotropic etching process. Inert gases are preferred for ion production in an anisotropic process, because ions generated from inert gases are effective at bombarding the surface being etched, and are unlikely to chemically react with adjacent surfaces (which may cause lateral etching). Further, the use of an inert gas as a diluent allows for better control and uniformity of the plasma. Argon (Ar) is very effective at providing ion bombardment during the etching of a substrate. However, a combination of a chemically reactive etchant species with active ion bombardment of the substrate surface generally results in a low selectivity for SiNx relative to photoresist. The selectivity range is generally from about 1 to about 1.5. As a result, these conventional anisotropic etching processes require the use of a thick layer of photoresist.

[0018] The inventors have discovered that adding sulfur dioxide (SO2) to a plasma source gas including one or more fluorine-containing gases during the patterned, anisotropic etching of an underlying layer protects the photoresist during etching of the underlying layer. As a result, the selectivity for the underlying layer is improved relative to the photoresist. For example, during etching of a SiNx layer through a DUV photoresist, addition of SO2 to the plasma source gas increases the selectivity for SiNx relative to photoresist. By adding SO2 to a fluorine-comprising source gas, the inventors were able to improve selectivity for SiNx in preference to a DUV photoresist from about 1.5:1 (SiNx:DUV) to about 2.1:1. Further, the inventors estimate that this selectivity may be increased to as high as 3:1 by process optimization. As a result, thinner layers of photoresist may be used, resolution may be increased, and the possibility of shading damage to underlying substrate features is reduced.

[0019] Typically, the addition of an oxygen-containing gas to the plasma source gas increases the etch rate of photoresist (which is an organic material and reacts rapidly with active oxygen species). This decreases the selectivity of the process for SiNx relative to photoresist. For example, U.S. Pat. No. Re. 30,505 to Adir Jacob, reissued Feb. 3, 1981, describes a process to facilitate the etching of silicon nitride (among other materials) using a “cold” plasma (under 325° C.), where the plasma source gas comprises a binary mixture of a halocarbon and oxygen. The preferred halocarbon comprises only one carbon atom, such as tetrafluoromethane. Typically oxygen makes up between about 8.5% and 25% of the volume of the plasma source gas mixture. It is disclosed that “(a further) increase in the mole fraction of oxygen (up to 0.5) tends to result in an excessive etch rate of the dielectric layer (SiO2 layer underlying the SiNx) with associated degradation of the photoresist mask and line-line resolution”.

[0020] The improvement in selectivity observed upon addition of SO2 to the plasma source gas was unexpected by applicants. While not intending to be limited as to any explanation as to why the present invention works, it is believed that the SO2 interacts with the photoresist and fluorine reactive species from the plasma to create a passivation layer on the surface of the photoresist which slows the etch rate of the photoresist. Other oxygen-containing gases which do not contain sulfur do not appear to have a similar effect.

[0021] While anisotropic etching processes are the most likely application for the present invention, and the most preferred embodiment of the invention provides for the use of a plasma source gas which includes argon, the inventors believe that the invention may be applied to a plasma etch process that does not involve ion bombardment. For example, in an isotropic process for etching SiNx, the addition of SO2 to a plasma source gas that includes a fluorine-comprising compound is expected to improve the selectivity of the etchant for SiNx relative to a photoresist. Diluent gases such as He, N2, Ar, Kr and Xe may also be added to the plasma source gas, to provide for better plasma control and uniformity in an isotropic etch process, as well.

[0022] One skilled in the art will recognize that a photoresist itself will generate reactive species during a plasma etching process. Since these reactive species affect the overall chemistry of the etch process, the relative amounts of the constituents of the plasma source gas may need to be adjusted depending on the particular photoresist in use. In any case, one skilled in the art can, with minimal experimentation, optimize a particular source gas for the task at hand.

[0023] We have discovered that by adding SO2 to the plasma source gas, due to the improvement in selectivity for SiNx relative to photoresist, it is possible to significantly reduce the thickness requirement for the photoresist masking layer. For example, a number of semiconductor device fabrication processes require the anisotropic etching of SiNx features to a depth of about 1500 Å-6000 Å in a layer of SiNx, with resolutions of 0.25 microns and lower in a lateral direction. Such features have an aspect ratio of about 0.6-2.4, i.e., the ratio of the etch depth to the smallest lateral dimension of the feature. A trench etched in SiNx having a depth of 2000 Å and a width of 2500 Å is an example of such a feature. To achieve a depth of 2000 Å using a conventional DUV photoresist without punch through, it is generally necessary to use a resist thickness of about 5000-7000 Å. However, by adding SO2 to the plasma source gas, it is possible to achieve a depth of 2000 Å using a photoresist thickness as low as about 3000-4000 Å.

[0024] Typically, the etch rate of the SiNx is selected to be about 1000 to 2000 Å/min., the rate which is achievable with conventional etchants. Slower etch rates would lead to higher processing costs, and faster etch rates may not be controllable using current processing apparatus. This is not to limit the invention to a maximum etch rate of 2000 Å, however. The inventors have observed that the addition of SO2 to the plasma source gas does not appear to have a significant effect on the etch rate of SiNx in a direction normal to the substrate. However, the addition of SO2 does lead to an improvement in the anisotropy achievable in an anisotropic etch process. Without SO2, etched regions are usually narrower at the bottom than the top of a trench, i.e., the sidewalls are undercut from the top toward the bottom of an etched trench. The walls may form angles as low as 80-85° with the horizontal substrate surface. However, with the addition of SO2, the amount of taper is reduced, and the sidewalls are more perpendicular with the substrate, with angles of about 87-90°.

[0025] The present invention may be used in to fabricate any anisotropically etched feature in SiNx. For example, the present invention may be used to fabricate trenches, vias, and/or lines in SiNx. FIG. 1 shows a trench 100 fabricated in a substrate having a layer 110 of SiNx. The trench has a depth 102, a top width 104, and a bottom width 106. The trench has side walls 120 and a bottom surface 130. Bottom surface 130 is in the plane of the substrate. A photoresist layer 140, having an initial thickness 141, was used to pattern trench 100. A portion 142 of photoresist layer 140, having an initial overall thickness 143, was etched away during etching of trench 100 in SiNx substrate 110 to a depth 102. A portion 144 of photoresist layer 140, having a thickness 145, remains after etching. Dashed lines are used to show photoresist portion 142 which was removed during the etching process. Portion 144 is generally removed using a separate process after the etching of trench 100.

[0026] Preferred Apparatus

[0027] The present invention may be practiced in any apparatus adapted to expose a substrate to a plasma. Preferably the apparatus is capable of applying a bias voltage to the substrate. Preferably, the apparatus includes devices for the independent control of a plasma source power used to generate the plasma, from a bias power used to generate substrate bias voltage. This enables the selection of a desired plasma density independently of the selection of the amount of energy with which ion bombardment occurs upon the substrate surface. The CENTURA® DPS™ silicon etch system available from Applied Materials, Inc., of Santa Clara, Calif., is an example of a system which provides for such independent control.

[0028] However, the present invention may be practiced in an apparatus having a plasma source power and bias power which are not separately controllable, or in any other type of apparatus adapted to expose a substrate to a plasma.

[0029] FIG. 2 (prior art) shows a preferred etch chamber 200 of the type used in the CENTURA® DPS™ silicon etch system. Etch chamber 200 includes a ceramic dome 202, and a standard monopolar electrostatic chuck (ESC) 204. Gas is introduced into the chamber via four ceramic gas injection nozzles 208 for uniform gas distribution. Chamber pressure is controlled by a closed-loop pressure control system 210 with a throttle valve 212.

[0030] Etch chamber 200 includes an inductive plasma source generator, typically operating at a frequency tuned at approximately 12.56 MHz, which generates and sustains a high density plasma (i.e., typically having an electron density of at least about 1011 e−/cm3). A 13.56 MHz RF power source is frequently applied to the substrate support pedestal (cathode) to create a bias voltage on the substrate surface. In the preferred apparatus, each watt of bias power generally produces about −1 Volt of substrate bias. The conversion from bias power watts to substrate bias voltage is highly dependent upon chamber geometry, and may be different in a different apparatus. The independently controlled plasma source power permits independent control of ion energy and ion density, which provides highly uniform plasma (<5% variation) with a wide process window over changes in source and bias power, pressure, and plasma source gas chemistry.

[0031] FIG. 3 (prior art) shows a nonscaled schematic of a vertical cross-section of an etch chamber 300, of the kind referenced in FIG. 2. Electrostatic chuck 302 is adapted to hold a substrate 304. Electrostatic chuck 302 overlies electrode (substrate pedestal) 306, which is electrically connected to an independently controlled power source 308. The upper portion 313 of the chamber wall is typically comprised of a ceramic material. The lower portion 310 of the chamber wall is typically comprised of an anodized aluminum surface 322 overlying an aluminum body 310 which is electrically grounded through switch 315. A plasma source gas is introduced into and distributed throughout chamber 300 by means of a gas distributor 312 peripherally disposed above substrate 304. Plasma ions are formed from the plasma source gas by applying an RF current to an inductive coil plasma generator 314, which is connected to an independently controlled plasma generator (RF) power source 316. The cathode electrode 306 is electrically biased with respect to the anode electrode (lower portion 310 of the chamber wall) by applying an RF voltage to the cathode electrode 306 via power source 308, so that the plasma ions formed in chamber 300 are attracted toward and energetically impinge upon substrate 304. Spent process gas and etchant by-products are exhausted from chamber 300 through an exhaust system 318. A throttle valve 320 is provided in exhaust system 318 to control the pressure in the chamber 300. A more detailed description of etch chamber 300 is found in U.S. Pat. No. 5,779,926, issued Jul. 14, 1998, to Ma et al., which is incorporated herein by reference.

[0032] Preferred Process Parameters

[0033] In the preferred apparatus, the plasma source power is preferably between about 200 and 2000 Watts during processing, which should be adequate to ionize a sufficient percentage of the gas molecules to generate a high density plasma. Lower plasma source powers may not generate a high density plasma, and an apparatus capable of providing higher plasma source powers are more expensive. The bias power is preferably between about 50 and 500 Watts, which generates an DC voltage difference between the plasma and the substrate of about 50 to 500 Volts. Lower bias voltages may result in an etch rate that is too slow, and higher bias voltages may cause damage to the substrate, and may lead to faster photoresist erosion. The bias frequency is preferably about 13.56 MHz, which generally gives good profile control, but may vary over a wide range from as low as 30 kHz up to multiples of 13.56 MHz.

[0034] Preferably, the total pressure in the chamber is maintained between about 4 and 70 milliTorr (mTorr) during etching. Lower pressures may lead to rounding of comers in the etched features, and higher pressures may lead to decreased etch rates.

[0035] In the preferred apparatus, the amount of fluorine-comprising gas fed into the chamber is preferably between about 20 and 300 sccm, and more preferably between about 20 to 100 sccm. The amount of Ar gas fed into the chamber is preferably between about 20 and 200 sccm, and more preferably between about 50 to 150 sccm. The amount of SO2 is preferably between about 5 and 50 seem, and more preferably between about 10 and 30 sccm. Providing too much gas in aggregate may make it difficult to maintain a process chamber vacuum in the desired range, which in turn will decrease the plasma density and the etch rate. Greater amounts of fluorine-comprising gas may lead to a higher concentration of reactive species, which may lead to an etch having some degree of isotropy. Lesser amounts of fluorine-comprising gas may lead to a lower concentration of reactive species and undesirably slow etch rates. Lesser amounts of Ar gas may lead to less ion bombardment, which may lead to slower etch rates due to passivation layers forming on the surfaces to be etched, and may result in polymeric residues on the etched feature surface. The use of gas volumes outside of these preferred ranges may work, particularly in an apparatus having significant volumetric and geometric differences from the preferred apparatus, or with process parameters significantly different from the preferred process parameters.

[0036] As used herein, the term “element-containing gas” or “element-comprising compound” refers to a gas or compound which includes an atom of the particular element. For example, carbon tetrafluoride (CF4), nitrogen trifluoride (NF3) and sulfur hexafluoride (SF6) are “fluorine-comprising compounds” which are fluorine-containing gases under the process conditions used during plasma etching. The terms “including” and “comprising” are used, inter alia, to describe a plasma source gas which is a mixture of the particular gases included, and which may include additional gases as well. The terms “including” and “comprising” are not intended to limit when the gases are mixed, such that a plasma source gas including CF4 and Ar may be implemented by feeding the two gases into a vacuum chamber via separate routes.

[0037] In terms of percentages, the amount of fluorine-containing gas is preferably between about 20% and 60% by volume of the plasma source gas, and more preferably between about 25% and 35%. The amount of Ar is preferably between about 20% and 60%, and more preferably between about 50% and 60%. The amount of SO2 is preferably between about 2% and 20%, and more preferably between about 5% and 15%.

[0038] The plasma source gas may optionally include hydrogen bromide (HBr). The addition of HBr to the plasma source gas results in the formation of silicon oxybromide (Si2OBr6) and/or bromosilane (SiH3Br), which provides sidewall passivation and improved etch profile control. The amount of HBr is preferably between about 10% and 60%, and more preferably between about 20% and 40%, by volume of the plasma source gas.

[0039] In terms of the atomic and/or molecular ratios of plasma source gases which provide the desired results, for the preferred embodiment fluorine-comprising gas, CF4, the preferred range for the ratio of CF4:Ar is between about 0.1:1 to about 10:1, and the most preferred range is between about 0.5:1 to about 1:1. The preferred range for the ratio of CF4:SO2 is between about 50:1 to about 1:1, and the most preferred range is between about 8:1 and about 3:1. When the fluorine-comprising source gas is other than CF4, one skilled in the art must take into account the amount of energy which is required to generate fluorine-containing etchant species from the fluorine-comprising source gas.

[0040] Preferably, the substrate is cooled during etch processing, generally by means of a coolant passed to a channel in the substrate support electrode (not shown). In addition, a flow of a coolant such as helium may be passed between the substrate and the substrate support platen to enhance cooling and maintain the temperature of the substrate within the desired range, preferably from about 20° C. to 100° C., and more preferably from about 40° C. to 60° C.

[0041] Note that the optimal value for any particular process parameter (including the amount of feed gas components), is highly dependent upon a variety of factors, including the geometry of the particular apparatus used, as well as the values of the other process parameters. It is therefore possible to depart from the preferred ranges set forth herein without departing from the scope of the invention. In any event, one of skill in the art should be able to determine optimal process parameters (including amounts of feed gas components) for use with a particular apparatus and/or subset of process parameters with minimal experimentation.

[0042] Computer Control

[0043] Preferably, the apparatus used to practice the present invention is adapted to be controlled by a computer. FIG. 4 shows a computer 400. Computer 400 comprises a processor 402, memory 404 adapted to store instructions 406, and one or more ports 408. Processor 402 is adapted to communicate with memory 404 and to execute instructions 406. Processor 402 and memory 404 are also adapted to communicate with one or more ports 408. Ports 408 are adapted to communicate with a plasma etch chamber 412. Plasma etch chamber is adapted to carry out process steps in accordance with signals received from processor 402 via ports 408. Preferably, computer 402 can control the composition and feed rate of the plasma source gas, the temperature, the pressure in the chamber, the bias power, the plasma source generation power. Preferably, computer 402 is adapted to receive measurements that describe the condition in the chamber, and adapt the process variables accordingly. Computer 400 is typically used in combination with a medium (not shown) which contains data which enables control of the apparatus to practice the method of the present invention.

EXAMPLES

Example 1 (Comparitive Example)

[0044] With reference to FIG. 1, a layer of DUV photoresist 140 (obtained from Shipley Company) about 7,000 Å thick was patterned (on top of a layer of SiNx 110), to have an opening 104 about 5,630 Å wide. The SiNx layer 110 was then anisotropically etched with a conventional plasma source gas, using the preferred apparatus described above. The conventional plasma source gas was a mixture of CF4 and Ar, and the CF4 was fed into the chamber at a rate of about 80 sccm, while the Ar was fed into the chamber at a rate of about 120 sccm. The pressure in the chamber was maintained at 4 mTorr. The plasma source power was 1,000 W, and the bias power was 200 W. The substrate support platen temperature was maintained at about 50° C. The etch time was 60 seconds.

[0045] After the etch process, a trench similar to that of FIG. 1 was observed. The depth 102 of the trench was about 4700 Å. The top width 104 of the trench was about 5600 Å, and the bottom width 106 of the trench was about 4700 Å. There was an approximately linear taper from top to bottom. The angle &thgr; formed between the planes of sidewalls 120 and bottom surface 130 was about 86.2°. About 3100 Å resist was eroded during the etch, i.e., thickness 143 was about 3100 Å, for a selectivity of about 1.53 (4750 Å/3100 Å) for SiNx relative to photoresist.

Example 2

[0046] A substrate was processed in a manner identical to that of FIG. 1, except that the plasma source gas included CF4, Ar, and SO2, where the relative gas feed rates were about 80 sccm, about 120 sccm, and about 20 sccm, respectively.

[0047] After the etch process, a trench similar to that of FIG. 1 was observed After a similar etch time, the depth 102 of the trench was about 4750 Å, the same as that of Example 1, indicating that the addition of SO2 did not slow the etch rate of SiNx. There was an approximately linear taper from top to bottom. The angle &thgr; between the sidewall and the substrate was about 88°, closer to perpendicular than in Example 1, indicating that the addition of SO2 enables a more ideal anisotropic etch. About 2300 Å of resist was eroded during the etch, i.e., thickness 143 was about 2300 Å. The selectivity was therefore about 2.07 (4750 Å/2300 Å), higher than that of Example 1, indicating that the addition of SO2 slows the etch rate of photoresist.

[0048] While the foregoing is directed to preferred embodiments of the invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof. The scope of the invention is determined by the claims that follow.

Claims

1. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of at least one inorganic fluorine-comprising gas and sulfur dioxide (SO2), wherein the molecular ratio of said inorganic fluorine-comprising gas to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

2. The method of claim 1, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

3. The method of claim 1, wherein one of said at least one inorganic fluorine-comprising gases is sulfur hexafluoride (SF6).

4. The method of claim 1, wherein one of said at least one inorganic fluorine-comprising gases is nitrogen trifluoride (NF3).

5. The method of claim 1, wherein said plasma source gas includes two inorganic fluorine-comprising gases.

6. The method of claim 1, wherein a temperature of a substrate which includes said silicon nitride is between about 20° C. and 100° C. during exposure to said plasma etchant.

7. The method of claim 6, wherein said temperature is between about 40° C. and 60° C.

8. The method of claim 1, wherein said silicon nitride etch rate is at least two times said organic DUV photoresist etch rate.

9. The method of claim 1, wherein said patterned organic DUV photoresist has a thickness of less than about 4000 Å.

10. The method of claim 1, wherein said exposing of the structure comprising said silicon nitride and said patterned organic DUV photoresist to said plasma etchant results in an etched inorganic substrate having a feature size less than 2500 Å.

11. The method of claim 1, wherein said selectivity of said silicon nitride relative to said adjacent patterned organic DUV photoresist is greater than 1.5.

12. The method of claim 11, wherein said selectivity is greater than about 2.0.

13. The method of claim 1, wherein said plasma etchant is generated from a high density plasma.

14. The method of claim 13, wherein said plasma density is at least 1011 e−/cm3.

15. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of at least one inorganic fluorine-comprising gas, sulfur dioxide (SO2), and a diluent gas selected from the group consisting of Ar, Kr, Xe, and He, wherein the molecular ratio of said inorganic fluorine-comprising gas to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

16. The method of claim 15, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

17. The method of claim 15, wherein a molecular ratio of said inorganic fluorine-comprising gas to said diluent gas ranges from about 0.1:1 to about 10:1.

18. The method of claim 15, wherein said diluent gas is argon (Ar).

19. The method of claim 15, wherein said sulfur dioxide amount is about 2%-20% by volume of said plasma source gas.

20. The method of claim 19, wherein said sulfur dioxide amount is about 10%-15% by volume of said plasma source gas.

21. The method of claim 15, wherein said at least one inorganic fluorine-comprising gas amount is about 20%-60% by volume of said plasma source gas.

22. The method of claim 21, wherein said at least one inorganic fluorine-comprising gas amount is about 25%-35% by volume of said plasma source gas.

23. The method of claim 18, wherein said argon amount is about 20%-60% by volume of said plasma source gas.

24. The method of claim 23, wherein said argon amount is about 50%-60% by volume of said plasma source gas.

25. The method of claim 15, wherein one of said at least one inorganic fluorine-comprising gases is sulfur hexafluoride (SF6).

26. The method of claim 15, wherein one of said at least one inorganic fluorine-comprising gases is nitrogen trifluoride (NF3).

27. The method of claim 15, wherein said plasma source gas includes two inorganic fluorine-comprising gases.

28. The method of claim 27, wherein a total amount of said inorganic fluorine-comprising gases is about 20%-60% by volume of the plasma source gas.

29. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of at least one inorganic fluorine-comprising gas, sulfur dioxide (SO2), and hydrogen bromide (HBr), wherein the molecular ratio of said inorganic fluorine-comprising gas to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

30. The method of claim 29, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

31. The method of claim 29, wherein said sulfur dioxide amount is about 2%-20% by volume of said plasma source gas.

32. The method of claim 31, wherein said sulfur dioxide amount is about 10%-15% by volume of said plasma source gas.

33. The method of claim 29, wherein said at least one inorganic fluorine-comprising gas amount is about 20%-60% by volume of said plasma source gas.

34. The method of claim 33, wherein said at least one inorganic fluorine-comprising gas amount is about 25%-35% by volume of said plasma source gas.

35. The method of claim 29, wherein said hydrogen bromide amount is about 10%-60% by volume of said plasma source gas.

36. The method of claim 35, wherein said hydrogen bromide amount is about 20%-40% by volume of said plasma source gas.

37. The method of claim 29, wherein one of said at least one inorganic fluorine-comprising gases is sulfur hexafluoride (SF6).

38. The method of claim 29, wherein one of said at least one inorganic fluorine-comprising gases is nitrogen trifluoride (NF3).

39. The method of claim 29, wherein said plasma source gas includes two inorganic fluorine-comprising gases.

40. The method of claim 39, wherein a total amount of said inorganic fluorine-comprising gases is about 20%-60% by volume of the plasma source gas.

41. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of at least one inorganic fluorine-comprising gas, sulfur dioxide (SO2), hydrogen bromide (HBr), and a diluent gas selected from the group consisting of Ar, Kr, Xe, and He, wherein the molecular ratio of said inorganic fluorine-comprising gas to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

42. The method of claim 41, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

43. The method of claim 41, wherein a molecular ratio of said inorganic fluorine-comprising gas to said diluent gas ranges from about 0.1:1 to about 10:1.

44. The method of claim 41, wherein said diluent gas is argon (Ar).

45. The method of claim 41, wherein said sulfur dioxide amount is about 2%-20% by volume of said plasma source gas.

46. The method of claim 45, wherein said sulfur dioxide amount is about 10%-15% by volume of said plasma source gas.

47. The method of claim 41, wherein said at least one inorganic fluorine-comprising gas amount is about 20%-60% by volume of said plasma source gas.

48. The method of claim 47, wherein said at least one inorganic fluorine-comprising gas amount is about 25%-35% by volume of said plasma source gas.

49. The method of claim 41, wherein said hydrogen bromide amount is about 10%-60% by volume of said plasma source gas.

50. The method of claim 49, wherein said hydrogen bromide amount is about 20%-40% by volume of said plasma source gas.

51. The method of claim 44, wherein said argon amount is about 20%-60% by volume of said plasma source gas.

52. The method of claim 51, wherein said argon amount is about 50%-60% by volume of said plasma source gas.

53. The method of claim 41, wherein one of said at least one inorganic fluorine-comprising gases is sulfur hexafluoride (SF6).

54. The method of claim 41, wherein one of said at least one inorganic fluorine-comprising gases is nitrogen trifluoride (NF3).

55. The method of claim 41, wherein said plasma source gas includes two inorganic fluorine-comprising gases.

56. The method of claim 55, wherein a total amount of said inorganic fluorine-comprising gases is about 20%-60% by volume of the plasma source gas.

57. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of sulfur hexafluoride (SF6) and sulfur dioxide (SO2), wherein the molecular ratio of said sulfur hexafluoride to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

58. The method of claim 57, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

59. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of sulfur hexafluoride (SF6), sulfur dioxide (SO2), and argon (Ar), wherein the molecular ratio of said sulfur hexafluoride to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

60. The method of claim 59, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

61. The method of claim 59, wherein a molecular ratio of said sulfur hexafluoride to said argon ranges from about 0.1:1 to about 10:1.

62. The method of claim 59, wherein said sulfur dioxide amount is about 2%-20% by volume of said plasma source gas.

63. The method of claim 62, wherein said sulfur dioxide amount is about 10%-15% by volume of said plasma source gas.

64. The method of claim 59, wherein said sulfur hexafluoride amount is about 20%-60% by volume of said plasma source gas.

65. The method of claim 64, wherein said sulfur hexafluoride amount is about 25%-35% by volume of said plasma source gas.

66. The method of claim 59, wherein said argon amount is about 20%-60% by volume of said plasma source gas.

67. The method of claim 66, wherein said argon amount is about 50%-60% by volume of said plasma source gas.

68. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of sulfur hexafluoride (SF6), sulfur dioxide (SO2), and hydrogen bromide (HBr), wherein the molecular ratio of said inorganic fluorine-comprising gas to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

69. The method of claim 68, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

70. The method of claim 68, wherein said sulfur dioxide amount is about 2%-20% by volume of said plasma source gas.

71. The method of claim 70, wherein said sulfur dioxide amount is about 10%-15% by volume of said plasma source gas.

72. The method of claim 68, wherein said sulfur hexafluoride amount is about 20%-60% by volume of said plasma source gas.

73. The method of claim 72, wherein said sulfur hexafluoride amount is about 25%-35% by volume of said plasma source gas.

74. The method of claim 68, wherein said hydrogen bromide amount is about 10%-60% by volume of said plasma source gas.

75. The method of claim 74, wherein said hydrogen bromide amount is about 20%-40% by volume of said plasma source gas.

76. A method of improving etch selectivity of silicon nitride relative to an adjacent patterned organic DUV photoresist during plasma etching, comprising: reacting a surface of said patterned organic DUV photoresist with plasma species generated from a plasma source gas consisting essentially of sulfur hexafluoride (SF6), sulfur dioxide (SO2), hydrogen bromide (HBr), and a diluent gas selected from the group consisting of Ar, Kr, Xe, and He, wherein the molecular ratio of said inorganic fluorine-comprising gas to said sulfur dioxide ranges from about 50:1 to about 1:1, to produce a reacted DUV photoresist surface, whereby the etch rate of said organic DUV photoresist is slowed, while said silicon nitride exposed through said patterned organic DUV photoresist is etched.

77. The method of claim 76, wherein the amount of sulfur dioxide (SO2) present in said plasma source gas is varied during said etching of said silicon nitride.

78. The method of claim 76, wherein a molecular ratio of said inorganic fluorine-comprising gas to said diluent gas ranges from about 0.1:1 to about 10:1.

79. The method of claim 76, wherein said diluent gas is argon (Ar).

80. The method of claim 76, wherein said sulfur dioxide amount is about 2%-20% by volume of said plasma source gas.

81. The method of claim 80, wherein said sulfur dioxide amount is about 10%-15% by volume of said plasma source gas.

82. The method of claim 76, wherein said sulfur hexafluoride amount is about 20%-60% by volume of said plasma source gas.

83. The method of claim 82, wherein said sulfur hexafluoride amount is about 25%-35% by volume of said plasma source gas.

84. The method of claim 76, wherein said hydrogen bromide amount is about 10%-60% by volume of said plasma source gas.

85. The method of claim 84, wherein said hydrogen bromide amount is about 20%-40% by volume of said plasma source gas.

86. The method of claim 79, wherein said argon amount is about 20%-60% by volume of said plasma source gas.

87. The method of claim 86, wherein said argon amount is about 50%-60% by volume of said plasma source gas.