Method of fabricating a split-gate flash memory

[0002] This application claims the priority benefit of Taiwan application ser. no. 88119924, filed Nov. 16, 1999.

[0003] 1. Field of the Invention

[0004] The invention relates to a method for fabricating a flash memory, and more particularly, to a method for fabricating a split-gate flash memory.

[0005] 2. Description of the Related Art

[0006] At present, nonvolatile memory is widely used in the whole range of electrical devices. In particular, programmable nonvolatile memory having a flash memory structure such as the erasable programmable read-only memory and electrically erased programmable read-only memory has attracted immense interest. In general, a flash memory comprises two gates, a floating gate for charge storage and a control gate for data accessing. The floating gate is in a floating state without being connected to any electrical circuit and is located between the control gate and a substrate while the control gate is connected to a word line.

[0007] FIG. 1 is a schematic, cross-sectional view showing a structure of a split-gate flash memory according to the prior art. In the FIG. 1, a floating gate 102 and a control gate 104 are formed on a substrate 100. A split-gate oxide layer 106 and a dielectric layer 108 separate the floating gate 102 and the control gate 104. Source/drain regions 110a, 110bare respectively formed in the substrate 100 adjacent to the collective structure of the control gate 104 and the floating gate 102. Sometimes, the control gate 104 is referred to as a selective gate. Referring to FIG. 1, the semiconductor process is first to form the floating gate 102 on the substrate 100 and then to form the split-gate oxide layer 106. Subsequently, a conductive layer is formed on the split-gate oxide layer 106. Then, the conductive layer is defined into the control gate 104 as shown in FIG. 1 by photolithography and etching. Afterwards, an ion implantation process is performed to form the source/drain regions 110a, 110b. A distance L₁covered by the control gate 104 between the source/drain region 110aand the structure comprising the dielectric layer 108 and the floating gate 102 is referred to as a channel length of the selective gate.

[0008] According to the prior art, the process is first to form the control gate 104 and then to form the source/drain region 110a; thus, the channel length L₁of the selective gate depends on the accuracy of photolithography for defining the control gate 104. Thus, when the photomask used for defining the control gate 104 is misaligned and the control gate 104 formed is shifted from a desired position, the channel length L₁is increased or decreased, and the reading current and the programming current are varied with the length L₁. When the length L₁is increased, the reading current is reduced; thus, a sensitive sense amplifier is required for detecting the reading current. In addition, the programming current is also reduced while the length L₁is increased; thus, the time for programming is increased, the speed becomes slower, and the operation time is increased.

[0009] According to above, the invention provides a method for fabricating a split-gate flash memory. According to the invention, a drain region and a floating gate are formed before a selective gate is formed in order to fix the distance between the drain region and the floating gate and to decide a channel length thereby. Thus, the invention can stably provide a reading current and a programming current and enhance the reliability of a device.

[0010] The invention provides a method for fabricating a split-gate flash memory. The method comprises the following steps.

[0011] A tunnelling oxide layer, a first conductive layer, and a hard mask layer are formed on a substrate in sequence. A drain opening and a floating gate opening are formed on the hard mask layer by defining the hard mask layer in order to expose the first conductive layer. A first polyoxide layer and a second polyoxide layer are formed on the first conductive layer exposed by the drain opening and the floating gate opening, respectively. The first polyoxide layer and the first conductive layer beneath the first polyoxide layer are removed to expose the substrate in the drain opening. A drain is formed in the substrate in the drain opening. The hard mask layer is removed, and the first conductive layer is etched into a floating gate using the second polyoxide layer as a mask. A split-gate oxide layer and a second conductive layer are formed on the resulting structure in sequence. A control gate is formed by defining the second conductive layer, and a source region beside the floating gate is formed in the substrate.

[0012] The invention provides another option for the method of fabricating a split-gate flash memory. The option comprises the following steps.

[0013] A tunnelling oxide layer, a first conductive layer, and a hard mask layer are formed on a substrate in sequence. A drain opening and a floating gate opening are formed on the hard mask layer by defining the hard mask layer in order to expose the first conductive layer. The first conductive layer exposed in the drain opening is removed to expose the substrate. A drain is formed in the substrate exposed by the drain opening, and an oxide layer is formed on the drain region. A polyoxide layer is formed on the first conductive layer exposed by the floating gate opening. The hard mask layer is removed, and the first conductive layer is etched into a floating gate using the second polyoxide layer as a mask. A split-gate oxide layer and a second conductive layer are formed on the resulting structure in sequence. A control gate is formed by defining the second conductive layer, and a source region beside the floating gate is formed in the substrate.

[0014] It is to be understood that both the foregoing general description and the following detailed description are exemplary, and are intended to provide further explanation of the invention as claimed.

[0015] FIG. 1 is a schematic, cross-sectional view showing a structure of a split-gate flash memory according to the prior art;

[0016] FIG. 2A through FIG. 2I are schematic, cross-sectional views showing a process of fabricating a split-gate flash memory according to a first preferred embodiment of the invention; and

[0017] FIG. 3A through FIG. 3C are schematic, cross-sectional views showing a process of fabricating a split-gate flash memory according to a second preferred embodiment of the invention.

[0018] The preferred embodiments of the invention disclose methods for fixing the distance between a drain region and a floating gate in order to ensure the channel length of a selective gate. Thus, the invention can ensure that the position of the selective gate does not vary with the misalignment of the selective gate.

[0019] First Preferred Embodiment of the Invention

[0020] FIG. 2A through FIG. 2I are schematic, cross-sectional views showing a process of fabricating a split-gate flash memory according to a first preferred embodiment of the invention. Referring to FIG. 2A, a isolation structure (not shown) of a field oxide layer formed by local oxidation or a shallow trench isolation formed by a shallow trench process is formed to define an active area, and a tunneling oxide layer 202 formed by, for example, thermal oxidation is formed to cover a substrate 200.

[0021] In FIG. 2A, a first conductive layer 204 having a thickness of about 1700-2000 Å and comprising polysilicon formed by, for example, chemical vapor deposition is formed to cover the tunneling oxide layer 202. A hard mask 206 having a thickness of about 1500 Å and comprising silicon nitride formed by, for example, chemical vapor deposition is formed to cover the first conductive layer 204.

[0022] The hard mask layer 206 is defined by photolithography. For example, a photoresist layer is formed on the hard mask layer 206. Then, using the first conductive layer 204 as an etching stop layer, the hard mask layer 206 is defined by the photoresist layer to form a hard mask layer 206ahaving a drain opening 208aand a floating gate opening 208bon the hard mask layer 206aand to expose the first conductive layer 204 in the openings as shown in FIG. 2B. According to the first preferred embodiment of the invention, the drain opening 208aand the floating gate opening 208bare in their respective desired positions.

[0023] Referring to FIG. 2C, using the hard mask layer 206aas a mask, the first conductive layer 204 exposed by the drain opening 208aand the floating gate opening 208bis oxidized to form a first polyoxide layer 210aand a second polyoxide layer 210b, respectively. For example, the first conductive layer 204 comprising polysilicon is oxidized in an O₂environment to form a polyoxide layer having a thickness of about 1800 Å.

[0024] Referring to FIG. 2D, a photoresist layer 212 is formed on the second polyoxide layer 210band covers portions of the hard mask layer 206b. Using the hard mask layer 206band the photoresist layer 212 as a mask, first, an anisotropic etching process with an etching recipe for removing oxide is performed to remove the first polyoxide layer 210ato expose the first conductive layer 204 in the drain opening 208a, and an anisotropic etching with an etching recipe for removing the first conductive layer 204 is then performed to remove the first conductive layer 204 in the drain opening 208ato expose the substrate 200 in the drain opening 208aand to form a first conductive layer 204a. The anisotropic etching process is, for example, a plasma etching process. In addition, due to the hard mask 206 and the anisotropic etching process, there may be a remainder 210a′ of the first polyoxide layer 210ain the drain opening 208a.

[0025] Ions are implanted in the exposed substrate 200 in the drain opening 208a, the photoresist layer 212 is removed, and a drive-in process such as a rapid thermal process for the implanted ions is performed to form a drain 216 as shown in FIG. 2E. Since the drive-in process is performed by a rapid thermal process, an oxide layer 218 is formed on the surface of the substrate 200 in the drain region 216. The oxide layer 218 can prevent the drain region 216 from being damaged by a subsequent etching process.

[0026] Referring to FIG. 2F, the hard mask layer 206ais removed by, for example, wet etching to expose the first conductive layer 204acovered by the hard mask layer 206a. Using the second polyoxide layer 210bas a mask, an anisotropic etching process such as a plasma etching process is performed to remove the first conductive layer 204auncovered by the second polyoxide layer 210bin order to form a floating gate 220. Due to the remainder 210a′ of the first polyoxide layer 210a, there is a remainder 204a′ of the first conductive layer 204abeneath the remainder 210a′ during removal of the first conductive layer 204aas shown in FIG. 2F. The remainders 210a′ and 204a′ are then removed.

[0027] Referring to FIG. 2G, a split-gate oxide layer 222 is formed on the resulting structure. A dielectric layer 224 such as a silicon nitride layer is formed on the split-gate oxide layer 222 and then etched back by, for example, an anisotropic etching process to form dielectric spacers 224aon sidewalls of the floating gate 220 as shown in FIG. 2H.

[0028] Referring to FIG. 2H, a second conductive layer 226 such as a polysilicon layer is formed on the resulting structure by, for example, chemical vapor deposition. Photolithography is performed to define the second conductive layer 226 in order to form a selective gate 228 as shown in FIG. 2I. An ion implantation is performed to form a source 230 beside the floating gate 220. Since the drain 216 and the floating gate 220 are completed before the selective gate 228 being formed, a channel length L₂of the selective gate 228 is fixed. That is, the channel length L₂of the selective gate 228 is not affected by the misalignment of the selective gate 228. Thus, the flash memory can obtain a stable reading current and a stable programming current.

[0029] Second Preferred Embodiment of the Invention

[0030] FIG. 3A through FIG. 3C are schematic, cross-sectional views showing a process of fabricating a split-gate flash memory according to a second preferred embodiment of the invention. Description of steps in FIG. 3A that are the same as the steps in FIG. 2A and FIG. 2B of the first preferred embodiment is not repeated here.

[0031] Referring to FIG. 3B, a photoresist layer 300 is formed to cover the floating gate opening 208band portions of the hard mask layer 206a. The first conductive layer 204 exposed by the drain opening 208ais removed by, for example, an anisotropic etching process to expose the substrate 200. Ions 302 are implanted in the exposed substrate 200 in the drain opening 208a.

[0032] Referring to FIG. 3C, the photoresist layer 300 is removed to expose the first conductive layer 204ain the floating gate opening 208b, and a drive-in process for the ions 302 is performed to form a drain 304. Since the drive-in process is performed in an O₂environment, the exposed substrate 200 in the drain opening 208ais oxidized into an oxide layer 306, and the exposed first conductive layer 204ain the floating gate opening 208bis oxidized into a polyoxide layer 308. The subsequent steps are the same as those in FIGS. 2F-2I; thus, the details are not repeated.

[0033] According to the invention, since the drain and the floating gate are completed before the selective gate is formed, the channel length of the selective gate is fixed. Thus, the channel length of the selective gate is not affected by the misalignment of the selective gate during the selective gate process. As a result, the flash memory can obtain a stable reading current and a stable programming current, and the reliability of the flash memory is enhanced.

[0034] Other embodiments of the invention will appear to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.