POWER SEMICONDUCTOR DEVICE HAVING REFLOWED INTERMETAL DIELECTRIC LAYER

Abstract

To provide a power semiconductor device.SOLUTION: Power semiconductor devices include multi-layer inter-metal dielectric patterns that include at least one reflowed dielectric material pattern and at least one non-reflowable dielectric material pattern. In other embodiments, power semiconductor devices include reflowed inter-metal dielectric patterns that are formed using sacrificial structures such as dams to limit lateral spread of the reflowable dielectric material of the inter-metal dielectric pattern during the re-flow-process. The inter-metal dielectric patterns may have improved shapes and performance.SELECTED DRAWING: Figure 5E

Description

[Technical field]

[0001] This application claims priority to U.S. Patent Application Serial No. 16 / 413,921, filed May 16, 2019, the entire contents of which are incorporated herein by reference as if set forth in their entirety.

[0002] The present invention relates to semiconductor devices, and more particularly to power semiconductor devices. [Background technology]

[0003] Power semiconductor devices are used to pass large currents and support high voltages. A wide variety of power semiconductor devices are known in the art, including, for example, power metal oxide semiconductor field effect transistors ("MOSFETs"), insulated gate bipolar transistors ("IGBTs"), and various other devices. These power semiconductor devices are typically fabricated from wide bandgap semiconductor materials such as silicon carbide or gallium nitride based materials (as used herein, the term "wide bandgap semiconductor" includes any semiconductor having a bandgap of at least 1.4 eV). Power semiconductor devices are designed to block (in a forward or reverse blocking state) or pass (in a forward conducting state) high voltages and / or currents. For example, power semiconductor devices may be designed to withstand potentials of hundreds or even thousands of volts in a blocking state.

[0004] Power semiconductor devices may have a lateral or vertical structure. In a device with a lateral structure, the terminals of the device (e.g., the drain, gate, and source terminals of a power MOSFET) are on the same major surface (i.e., the top surface or the bottom surface) of the semiconductor layer structure. In contrast, a device with a vertical structure has at least one terminal on each major surface of the semiconductor layer structure (e.g., in a vertical MOSFET, the source and gate may be on the top surface of the semiconductor layer structure and the drain may be on the bottom surface of the semiconductor layer structure). Vertical structures are typically used for very high power applications because they allow for thick semiconductor drift layers that can withstand high current densities and block high voltages. In this specification, the term "semiconductor layer structure" refers to a structure that includes one or more semiconductor layers, such as a semiconductor substrate and / or a semiconductor epitaxial layer.

[0005] A conventional silicon carbide vertical power MOSFET includes an epitaxial layer structure formed on a silicon carbide substrate, such as a silicon carbide wafer. The epitaxial layer structure (which may include one or more separate layers) functions as a drift region for power semiconductor devices. The MOSFET may have an active region formed on and / or within the drift region in which one or more semiconductor devices are formed, and a termination region that may surround the active region. The active region functions as a main junction to block voltage in a reverse bias direction and provide current in a forward bias direction. Power MOSFETs typically have a unit cell structure, meaning that the active region includes many individual "unit cell" MOSFETs that are electrically connected in parallel to function as a single power MOSFET. In high power applications, such devices may include thousands or tens of thousands of unit cells. Summary of the Invention [Means for solving the problem]

[0006] According to an embodiment of the present invention, there is provided a semiconductor device including a wide bandgap semiconductor layer structure, a gate electrode structure on an upper surface of the wide bandgap semiconductor layer structure, an inter-metal dielectric pattern on the gate electrode structure including a non-reflowable dielectric material pattern and a reflowed dielectric material pattern, and a source metallization structure on the inter-metal dielectric pattern, wherein the gate electrode structure is between the wide bandgap semiconductor layer structure and the inter-metal dielectric pattern, and the inter-metal dielectric pattern is between the gate electrode structure and the source metallization structure.

[0007] In some embodiments, a pattern of non-reflowable dielectric material is between the gate electrode structure and the pattern of reflowed dielectric material.

[0008] In some embodiments, the minimum thickness of the non-reflowable dielectric material pattern may be selected to be sufficient to avoid destruction of the inter-metal dielectric pattern during normal operation of the semiconductor device.

[0009] In some embodiments, the thickness of the upper corners of the non-reflowable dielectric material pattern may be selected to be sufficient to avoid destruction of the intermetal dielectric pattern during normal operation of the semiconductor device.

[0010] In some embodiments, the gate electrode structure can include a plurality of gate fingers separated from the wide bandgap semiconductor layer structure by respective ones of a plurality of gate insulating fingers, and the non-reflowable dielectric material pattern can include a plurality of non-reflowable dielectric fingers conformally surrounding an upper surface and side surfaces of each of the gate fingers.

[0011] In some embodiments, the reflowed dielectric material pattern may have a rounded cross-section.

[0012] In some embodiments, the ratio of the thickness of the intermetal dielectric pattern above the center of the top surface of the gate finger of the gate electrode structure to the minimum thickness of the intermetal dielectric pattern may be less than 4:1.

[0013] In some embodiments, the portion of the reflowed dielectric material pattern adjacent an upper corner of the non-reflowable dielectric material pattern may be the portion of the reflowed dielectric material pattern having a minimum thickness.

[0014] In some embodiments, the reflowed dielectric material pattern may include a boro-phospho-silicate glass ("BPSG") pattern.

[0015] In some embodiments, the semiconductor device may be a MOSFET, wherein a plurality of source regions are provided in the wide bandgap semiconductor layer structure, the source metallization pattern is electrically connected to the source regions, and the device further includes a drain contact on the wide bandgap semiconductor layer structure opposite the source metallization structures.

[0016] In some embodiments, the source metallization structure can be a diffusion barrier layer and a metallic source contact layer on the diffusion barrier layer.

[0017] In some embodiments, the semiconductor device may be an insulated gate bipolar junction transistor.

[0018] In some embodiments, there may be a reflowed dielectric material pattern between the gate electrode structure and the non-reflowable dielectric material pattern.

[0019] According to a further embodiment of the present invention, there is provided a semiconductor device including a wide bandgap semiconductor layer structure, a gate electrode structure on an upper surface of the wide bandgap semiconductor layer structure, an intermetal dielectric pattern on the gate electrode structure including a reflowed dielectric material pattern, and a source metallization structure on the intermetal dielectric pattern, wherein a lower portion of the reflowed dielectric material pattern has substantially vertical sidewalls.

[0020] In some embodiments, the upper portion of the reflowed dielectric material pattern may have a rounded cross-section.

[0021] In some embodiments, the gate electrode structure may include a plurality of gate fingers separated from the wide bandgap semiconductor layer structure by respective ones of a plurality of gate insulating fingers.

[0022] In some embodiments, the gate insulating finger may have a first thickness in a direction perpendicular to an upper surface of the wide bandgap semiconductor layer structure, and a lower portion of the reflowed dielectric material pattern may have a second thickness that is greater than or equal to the first thickness.

[0023] In some embodiments, the thickness of the lower portion of the reflowed dielectric material pattern may be at least 0.1 microns.

[0024] In some embodiments, a gate electrode structure may be between the wide bandgap semiconductor layer structure and the intermetal dielectric pattern, and the intermetal dielectric pattern may be between the gate electrode structure and the source metallization structure.

[0025] In some embodiments, the reflowed dielectric material pattern may include a borophosphosilicate glass ("BPSG") pattern.

[0026] In some embodiments, the ratio of the thickness of the intermetal dielectric pattern above the center of the top surface of the gate finger of the gate electrode structure to the minimum thickness of the intermetal dielectric pattern may be at least 1:1 and less than 4:1.

[0027] According to a further embodiment of the present invention, there is provided a method of fabricating a semiconductor device, in which a wide bandgap semiconductor layer structure is formed. A conductive pattern is formed on an upper surface of the wide bandgap semiconductor layer structure. A non-reflowable dielectric material pattern is formed on the conductive pattern. A reflowable dielectric material layer comprising a reflowable dielectric material is formed on the conductive pattern. The reflowable dielectric material is reflowed. A source metallization structure is formed on the inter-metal dielectric pattern. The inter-metal dielectric pattern comprises the non-reflowable dielectric material pattern and a reflowed dielectric material pattern comprising at least a portion of the reflowable dielectric material.

[0028] In some embodiments, a non-reflowable dielectric material pattern may be between the conductive pattern and the reflowed dielectric material pattern.

[0029] In some embodiments, the reflowed dielectric material pattern may be between the conductive pattern and the non-reflowable dielectric material pattern.

[0030] In some embodiments, the method may further include reflowing the reflowable dielectric material layer comprising the reflowable dielectric material to form a reflowed dielectric material layer, and then etching the reflowed dielectric material layer to form a reflowed dielectric material pattern.

[0031] In some embodiments, the method may further include etching a reflowable dielectric material layer comprising the reflowable dielectric material to form a reflowable dielectric material pattern, and then reflowing the reflowable dielectric material pattern to form a reflowed dielectric material pattern.

[0032] In some embodiments, the conductive pattern can include multiple gate fingers.

[0033] In some embodiments, forming the non-reflowable dielectric material pattern can include conformally forming non-reflowable dielectric fingers over each of the respective gate fingers.

[0034] In some embodiments, the reflowed dielectric material pattern can be a borophosphosilicate glass pattern.

[0035] In some embodiments, a pattern of non-reflowable dielectric material may be formed on the sidewalls of the gate fingers.

[0036] In some embodiments, the reflowed dielectric material pattern may directly contact the sidewalls of the gate fingers.

[0037] In some embodiments, the conductive pattern may include a semiconductor pattern, and forming the non-reflowable dielectric material pattern over the conductive pattern may include oxidizing exposed surfaces of the semiconductor pattern.

[0038] In some embodiments, the semiconductor pattern may include a plurality of polysilicon gate fingers and the wide bandgap semiconductor layer structure may include a silicon carbide semiconductor layer structure.

[0039] In some embodiments, the method can further include forming respective sacrificial structures in gaps defined between adjacent ones of the gate fingers, the sacrificial structures positioned to limit lateral expansion of the reflowable dielectric material during reflow of the reflowable dielectric material.

[0040] In some embodiments, the minimum thickness of the non-reflowable dielectric material pattern may be selected to be sufficient to avoid destruction of the inter-metal dielectric pattern during normal operation of the semiconductor device.

[0041] In some embodiments, the thickness of the upper corners of the non-reflowable dielectric material pattern may be selected to be sufficient to avoid destruction of the intermetal dielectric pattern during normal operation of the semiconductor device.

[0042] In some embodiments, the reflowed dielectric material pattern may have a rounded cross-section.

[0043] In some embodiments, the ratio of the thickness of the intermetal dielectric pattern above the center of the top surface of the gate fingers of the conductive pattern to the minimum thickness of the intermetal dielectric pattern may be less than 4:1.

[0044] According to yet another embodiment of the present invention, there is provided a method of fabricating a semiconductor device, in which a wide bandgap semiconductor layer structure is formed and then a plurality of spaced apart gate fingers are formed on an upper surface of the wide bandgap semiconductor layer structure. A reflowable dielectric material layer including a reflowable dielectric material is formed on the gate fingers. A sacrificial structure is formed on the wide bandgap semiconductor layer structure in the gaps between the gate fingers. The reflowable dielectric material is reflowed. A source metallization pattern is formed on the intermetal dielectric pattern. The intermetal dielectric pattern includes a reflowed dielectric material pattern including at least a portion of the reflowable dielectric material.

[0045] In some embodiments, forming the layer of reflowable dielectric material over the gate fingers can include conformally forming a layer of reflowable dielectric material over each respective gate finger.

[0046] In some embodiments, the reflowed dielectric material pattern can be a borophosphosilicate glass pattern.

[0047] In some embodiments, the gate fingers may include polysilicon gate fingers and the wide bandgap semiconductor layer structure may include a silicon carbide semiconductor layer structure.

[0048] In some embodiments, the sacrificial structures may be positioned to limit the lateral expansion of the reflowable dielectric material while reflowing the reflowable dielectric material.

[0049] In some embodiments, the method can further include removing the sacrificial structure.

[0050] In some embodiments, a lower portion of the reflowed dielectric material pattern can have substantially vertical sidewalls and an upper portion of the reflowed dielectric material pattern can have a rounded cross-section.

[0051] According to a further embodiment of the present invention, there is provided a semiconductor device including a wide bandgap semiconductor layer structure, a gate electrode structure on an upper surface of the wide bandgap semiconductor layer structure, an intermetal dielectric pattern on the gate electrode structure, the intermetal dielectric pattern comprising at least a first silicon oxide pattern and a second silicon oxide pattern, the second silicon oxide pattern comprising a different material than the first silicon oxide pattern, and a source metallization structure on the intermetal dielectric pattern, the gate electrode structure being between the wide bandgap semiconductor layer structure and the intermetal dielectric pattern, the intermetal dielectric pattern being between the gate electrode structure and the source metallization structure, and a ratio of a maximum thickness of the intermetal dielectric pattern to a minimum thickness of the intermetal dielectric pattern is less than 4:1.

[0052] In some embodiments, the first silicon oxide pattern can include a pattern of non-reflowable material and the second silicon oxide pattern can include a pattern of reflowed material.

[0053] In some embodiments, the minimum thickness of the intermetal dielectric pattern may be adjacent the upper corners of the gate electrode structure.

[0054] In some embodiments, a maximum thickness of the intermetal dielectric pattern may be above the center of the top surface of the gate finger of the gate electrode structure.

[0055] In some embodiments, the ratio of the maximum thickness of the intermetal dielectric pattern to the minimum thickness of the intermetal dielectric pattern may be greater than 1:1.

[0056] In some embodiments, the ratio of the thickness of the intermetal dielectric pattern above the center of the top surface of the gate finger of the gate electrode structure to the minimum thickness of the intermetal dielectric pattern may be greater than 1:1.

[0057] In some embodiments, a minimum thickness of the intermetal dielectric pattern may be adjacent the upper corners of the gate electrode structure. [Brief description of the drawings]

[0058] [Figure 1] FIG. 2 is a schematic cross-sectional view of the upper source metallization structure for several unit cells of a power MOSFET. [Diagram 2] FIG. 2 is an enlarged cross-sectional view of one of the gate fingers and surrounding structure in the MOSFET of FIG. [Diagram 3] FIG. 2 is a schematic cross-sectional view of the upper source metallization structures for several unit cells of a power MOSFET formed using a reflowable intermetal dielectric pattern. [Figure 4] 1 is a schematic plan view of a semiconductor wafer including a plurality of power MOSFETs in accordance with an embodiment of the present invention. [Figure 5A] 5 is a schematic plan view of one of the power MOSFETs included on the semiconductor wafer of FIG. 4. [Figure 5B] FIG. 5B is a schematic plan view of the power MOSFET of FIG. 5A , omitting the top source metallization structure, gate bond pad, and intermetal dielectric pattern. [Figure 5C] FIG. 5C is a schematic cross-sectional view taken along line 5C-5C of FIG. 5A. [Figure 5D] FIG. 5D is a schematic cross-sectional view taken along line 5D-5D in FIG. 5C. [Figure 5E] An expanded view of one of the unit cells shown in Figure 5C with the source metallization structures omitted. [Figure 6A] 5A to 5E are schematic cross-sectional views illustrating a method of forming an inter-metal dielectric pattern of the MOSFET of FIG. [Figure 6B]5A to 5E are schematic cross-sectional views illustrating a method of forming an inter-metal dielectric pattern of the MOSFET of FIG. [Figure 6C] 5A to 5E are schematic cross-sectional views illustrating a method of forming an inter-metal dielectric pattern of the MOSFET of FIG. [Figure 6D] 5A to 5E are schematic cross-sectional views illustrating a method of forming an inter-metal dielectric pattern of the MOSFET of FIG. [Figure 6E] 5A to 5E are schematic cross-sectional views illustrating a method of forming an inter-metal dielectric pattern of the MOSFET of FIG. [Figure 7A] 5A to 5E are cross-sectional views illustrating another method of forming an intermetal dielectric pattern of the MOSFET of FIG. [Figure 7B] 5A to 5E are cross-sectional views illustrating another method of forming an intermetal dielectric pattern of the MOSFET of FIG. [Figure 8A] 5A through 5E are cross-sectional views illustrating a method of forming an intermetal dielectric pattern that can be used in place of the intermetal dielectric patterns of the MOSFETs of FIGS. 5A through 5E. [Figure 8B] 5A through 5E are cross-sectional views illustrating a method of forming an intermetal dielectric pattern that can be used in place of the intermetal dielectric patterns of the MOSFETs of FIGS. 5A through 5E. [Figure 8C] 5A through 5E are cross-sectional views illustrating a method of forming an intermetal dielectric pattern that can be used in place of the intermetal dielectric patterns of the MOSFETs of FIGS. 5A through 5E. [Figure 8D] 5A through 5E are cross-sectional views illustrating a method of forming an intermetal dielectric pattern that can be used in place of the intermetal dielectric patterns of the MOSFETs of FIGS. 5A through 5E. [Figure 8E] 5A through 5E are cross-sectional views illustrating a method of forming an intermetal dielectric pattern that can be used in place of the intermetal dielectric patterns of the MOSFETs of FIGS. 5A through 5E. [Figure 8F] 5A through 5E are cross-sectional views illustrating a method of forming an intermetal dielectric pattern that can be used in place of the intermetal dielectric patterns of the MOSFETs of FIGS. 5A through 5E. [Figure 9A]5A-5C are schematic cross-sectional views illustrating a method of forming a MOSFET according to a further embodiment of the present invention. [Figure 9B] 5A-5C are schematic cross-sectional views illustrating a method of forming a MOSFET according to a further embodiment of the present invention. [Figure 9C] 5A-5C are schematic cross-sectional views illustrating a method of forming a MOSFET according to a further embodiment of the present invention. [Figure 9D] 5A-5C are schematic cross-sectional views illustrating a method of forming a MOSFET according to a further embodiment of the present invention. [Figure 10] FIG. 13 is a cross-sectional view of one unit cell of a power MOSFET according to a further embodiment of the present invention in which the source metallization is omitted. [Figure 11A] FIG. 2 is a simplified circuit diagram of an n-channel IGBT according to an embodiment of the present invention. [Figure 11B] FIG. 11B is a schematic cross-sectional view of a pair of unit cells of the IGBT of FIG. [Figure 12] 4 is a flow chart of a method of forming a power MOSFET in accordance with an embodiment of the present invention. [Figure 13] 5 is a flow chart of a method of forming a power MOSFET in accordance with a further embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0059] A power semiconductor device, such as a power MOSFET or a power IGBT, includes an upper metallization structure that is used to connect the terminals of the device to an external structure. For example, a vertical power MOSFET includes a semiconductor layer structure, an upper source metallization structure that functions as the source terminal of the device, an upper gate metallization structure that functions as the gate terminal of the device, and a "bottom" drain metallization structure formed on the bottom surface of the device that functions as the drain terminal for the MOSFET. Figure 1 is a schematic cross-sectional view of the upper portion of several unit cells of a conventional power MOSFET 10.

[0060] As shown in FIG. 1, a conventional power MOSFET 10 includes a semiconductor layer structure 20 (only an upper portion of the semiconductor layer structure 20 is shown in FIG. 1), a gate electrode structure 30, an intermetal dielectric pattern 50, and an upper source metallization structure 60 formed on an upper surface of the semiconductor layer structure 20 and the gate electrode structure 30. The semiconductor layer structure 20 may include, for example, a silicon carbide semiconductor substrate (not shown). The silicon carbide semiconductor substrate has a silicon carbide drift region 24 epitaxially grown thereon. A well region 26 having an opposite conductivity type to the drift region 24 is formed in an upper surface of the drift region 24. A source region 28 is formed in the well region 26. Although not shown in FIG. 1, a drain contact is formed on a bottom surface of the semiconductor layer structure 20.

[0061] The gate electrode structure 30 includes a plurality of gate fingers 34 formed on the upper surface of the semiconductor layer structure 20. Each gate finger 34 may have a bar-shaped semiconductor or metal pattern, and the gate fingers 34 may extend parallel to each other. To insulate the gate fingers 34 from the semiconductor layer structure 20, the gate electrode structure 30 further includes a plurality of gate insulating fingers 32 formed between each gate finger 34 and the upper surface of the semiconductor layer structure 20. The gate insulating fingers 32 may include stripes of silicon oxide. An intermetal dielectric pattern 50 includes a plurality of dielectric fingers 52 and is formed on the upper surface of the semiconductor layer structure 20 after the formation of the gate insulating fingers 32 and the gate fingers 34. Each dielectric finger 52 may cover the sidewalls and the upper surface of a respective one of the gate insulating fingers 32 and a respective one of the gate fingers 34. A gap 40 is provided between adjacent dielectric fingers 52 to expose the source regions 28 on the upper surface of the semiconductor layer structure 20. The pitch between adjacent dielectric fingers 52 can be small, so that gap 40 has a correspondingly narrow width (eg, 1-5 microns).

[0062] The upper source metallization structure 60 includes a conductive diffusion barrier layer 62 and a metallic source contact layer 64. The diffusion barrier layer 62 may be conformally formed on the upper surface of the semiconductor layer structure 20 over the intermetal dielectric pattern 50 and the exposed well regions 26 and source regions 28. The diffusion barrier layer 62 may be designed to be a reasonably good conductor that may prevent materials (such as wet etchants) from diffusing into the intermetal dielectric pattern 50. The source contact layer 64 may include a highly conductive metal layer. There may be a recess 66, referred to as a "keyhole," in the upper surface of the source contact layer 64 above some or all of the gap 40.

[0063] During the fabrication of the MOSFET 10, chemicals such as wet etchants may be applied to the source contact layer 64. Completely washing off the wet etchant may be difficult, especially if the upper surface of the source contact layer 64 includes deep keyholes 66. Potentially, any remaining wet etchant (or other corrosive material) may diffuse deep into the upper source metallization structure 60 along the grain boundaries of the source metal. Typically, the conductive diffusion barrier layer 62 prevents further diffusion of chemicals that reach the bottom of the source contact layer 64, but small gaps may exist in the diffusion barrier layer 62, especially within the gaps 40, because the metal of the diffusion barrier layer cannot always completely fill the narrow gaps 40. If gaps exist in the diffusion barrier layer 62, chemicals may pass through the gaps and then attack the intermetal dielectric pattern 50. If the intermetal dielectric pattern 50 includes seams or other passages, chemicals may progress along the seams and / or spread, creating a path between the upper source metallization structure 60 and the gate fingers 34. This could potentially create an electrical short between the upper source metallization structure 60 and the gate finger 34. Such an electrical short could result in damage or destruction of the MOSFET 10, even if it occurs in a single unit cell.

[0064] FIG. 2 is an enlarged cross-sectional view, exaggerated to scale, of one of the gate fingers 34 of the MOSFET 10 of FIG. 1. As shown in FIG. 2, a dielectric finger 52 of an intermetal dielectric pattern 50 is formed substantially conformally over the gate finger 34 and the underlying gate insulating finger 32. The sidewalls of the dielectric finger 52 and the upper surface of the semiconductor layer structure 20 define a pair of inwardly facing angles α1 and α2. As shown in FIG. 2, the angles α1 and α2 may each be about 90 degrees. In the exemplary illustrated embodiment, the angles α1 and α2 are each about 80 degrees, and therefore the sidewalls of the gap 40 extend upward at an angle of about 100 degrees. When the sidewalls of the gap 40 have such sharp angles, it may be difficult to completely fill the gap 40, which may result in the formation of gaps in the diffusion barrier layer 62, as described above.

[0065] 3 is a schematic cross-sectional view of a portion of a MOSFET 10A that is a modified version of the MOSFET 10 of FIG. 1, including an intermetal dielectric pattern 50A having a different shape than the intermetal dielectric pattern 50 of the MOSFET 10 of FIG. 1. Referring to FIG. 3, one technique for reducing the likelihood that gaps will form in the diffusion barrier layer 62 is to reduce the magnitude of the angles α1 and α2, thereby increasing the angle that the sidewalls of the gap 40 subtend with the upper surface of the semiconductor layer structure 20. As the angle defined by the sidewalls of the gap 40 increases, it becomes easier for the metal of the diffusion barrier layer to fill the gap 40 without forming gaps. As shown in FIG. 3, one way to reduce the magnitude of the angles α1 and α2 (and therefore increase the angle defined by the sidewalls of the gap 40) is to conformally form a reflowable intermetal dielectric pattern (not shown in FIG. 3) on the top surface and sidewalls of the gate electrode structure 30 using a reflowable dielectric material, such as, for example, boro-phospho-silicate glass ("BPSG"). The BPSG intermetal dielectric pattern, after being conformally formed on the gate electrode structure 30, can be reflowed at an elevated temperature (e.g., between 500° C. and 1000° C.) to provide a reflowed intermetal dielectric pattern 50A. The reflow process can increase the density of the dielectric material and aid in reducing or eliminating material seams. In addition, the surface tension of the reflowed BPSG material can result in separate dielectric fingers 52A forming the reflowed intermetal dielectric pattern 50A having a generally semi-elliptical or semi-circular cross-section, as shown in FIG. 3. This generally "rounded" contour can reduce angles α1 and α2, making it easier to form a diffusion barrier layer 62A in the gap 40 without forming significant gaps.

[0066] Although forming the intermetal dielectric pattern 50A with a reflowable dielectric material can significantly improve the gap-filling properties of the diffusion barrier layer 62A (and therefore reduce void formation therein), the thickness of the reflowed BPSG material covering each gate finger 34 is not uniform. In particular, as shown in FIG. 3, the thickness of the reflowed BPSG material can have a maximum thickness above the center of the top surface of the gate finger 34 and can be significantly thinner adjacent the lateral edges of the top surface of the gate finger 34. In some cases, the thickness of the reflowed intermetal dielectric pattern 50A adjacent the top edges of the gate finger 34 (see region 53 in FIG. 3) can have a thickness that is only 10%-40% of the thickness of the reflowed BPSG material above the center of the top surface of the gate finger 34. In addition, the reflow process can tend to elongate the lateral width of each dielectric finger 52A, thereby reducing the width of the gap 40. The upper thin "corner" regions of the reflowed dielectric fingers 52A can be vulnerable to breakdown during operation of the device. This can result in a short between the gate finger 34 and the source metallization structure 60, which can result in device failure. Additionally, a smaller gap 40 can increase the source contact resistance and / or increase the likelihood of gaps forming during deposition of the diffusion barrier layer 62A. Thus, while the use of a reflowable intermetal dielectric pattern 50A may solve one problem, it can create another.

[0067] According to certain embodiments of the present invention, a power semiconductor device is provided that includes a multi-layer intermetal dielectric pattern including at least one reflowable dielectric material pattern and at least one non-reflowable dielectric material pattern. The non-reflowable dielectric material pattern may be formed on the exposed top surfaces of the gate fingers and on the gate insulating fingers and exposed side surfaces of the gate fingers. The reflowable dielectric material pattern may be formed on the non-reflowable dielectric material pattern. Alternatively, the reflowable dielectric material pattern may be formed on the exposed top surfaces of the gate fingers and on the exposed side surfaces of the gate insulating fingers and gate fingers. The reflowable dielectric material pattern may then be reflowed to form a reflowed dielectric material pattern, and a non-reflowable dielectric material pattern may be formed on the reflowed dielectric material pattern.

[0068] In either case, the non-reflowable dielectric material pattern can have a sufficient thickness to prevent, for example, destruction of the intermetal dielectric pattern. Thus, even if the thickness of the reflowable dielectric material pattern adjacent the upper corners of the gate fingers is significantly reduced during the reflow process, there will still be a sufficient total thickness of dielectric material to avoid destruction. Furthermore, the combination of the non-reflowable dielectric material pattern and the reflowable dielectric material pattern can have a thickness that is approximately the same as the thickness of the reflowable dielectric material pattern of the embodiment of FIG. 3. Thus, the reflowable dielectric material pattern can have a smaller thickness and therefore will not expand into the gaps between adjacent gate fingers during reflow. Thus, an intermetal dielectric pattern according to an embodiment of the present invention can have the advantages of a fully reflowable intermetal dielectric pattern while reducing or minimizing the potential disadvantages associated with a fully reflowable intermetal dielectric pattern.

[0069] According to a further embodiment of the present invention, there is provided a power semiconductor device including a reflowed intermetal dielectric pattern formed using sacrificial structures, such as dams, to limit the lateral expansion of the reflowable dielectric material of the intermetal dielectric pattern during the reflow process. The use of such dams can increase the amount of dielectric material along the lateral edges of the gate fingers and ensure that the intermetal dielectric pattern has a sufficient minimum thickness to avoid destruction of the intermetal dielectric pattern. The sacrificial dams can prevent the lateral expansion of the reflowable dielectric material, thereby maintaining the size of the gap between adjacent dielectric fingers.

[0070] The intermetal dielectric pattern may include, for example, a multi-layer intermetal dielectric pattern having a non-reflowable layer and a reflowable layer. The ratio of the maximum thickness of the intermetal dielectric pattern to the minimum thickness of the intermetal dielectric pattern is less than 8:1 in some embodiments, less than 6:1 in other embodiments, and less than 4:1 in still other embodiments. In each case, the ratio of the maximum thickness of the intermetal dielectric pattern to the minimum thickness of the intermetal dielectric pattern is at least 1:1.

[0071] Although the discussion herein has focused on power MOSFET devices, it will be understood that the techniques disclosed herein are not limited to such devices, for example, the techniques disclosed herein may also be used in IGBT devices.

[0072] In the following, embodiments of the present invention will be explained in more detail with reference to Figs. 4 to 13, in which exemplary embodiments of the present invention are shown.

[0073] 4 is a schematic plan view of a semiconductor wafer 100 including a plurality of power MOSFETs 110 according to an embodiment of the invention. The power MOSFETs 110 may be formed in rows and columns and spaced apart from one another so that the wafer 100 may later be separated (e.g., diced) to separate the individual power MOSFETs 110 for packaging and testing. In some embodiments, for example, the wafer 100 may include a 4H silicon carbide substrate having one or more silicon carbide layers formed thereon (e.g., by epitaxial growth). Other semiconductor layers (e.g., polysilicon layers), insulating layers, and / or metal layers may be formed on the silicon carbide semiconductor layer structure to form the power MOSFETs 110. In some cases, the silicon carbide substrate may be thinned or removed after other semiconductor layers are formed thereon.

[0074] Figure 5A is a schematic plan view of one of the power MOSFETs 110 included in the semiconductor wafer 100 of Figure 4. Figure 5B is a schematic plan view of the power MOSFET of Figure 5A omitting the top source metallization structure, gate bond pad, and intermetal dielectric pattern.

[0075] As shown in Figure 5A, a gate bond pad 112 and one or more source bond pads 114-1, 114-2 may be formed on an upper surface of a semiconductor layer structure 120 (Figure 5C) of the MOSFET 110. A drain bond pad 116 (shown as a dashed box in Figure 5A) may be provided on a bottom surface of the MOSFET 110. Each bond pad 112, 114, 116 may be formed of a metal (such as, for example, aluminum) to which bond wires may be readily connected by conventional techniques such as thermocompression or soldering.

[0076] As will be described in more detail below, the MOSFET 110 includes an upper source metallization structure 160 that electrically connects the source region 128 of the semiconductor layer structure 120 of the MOSFET 110 to an external device. The upper source metallization structure 160 is illustrated by a dashed box in FIG. 5A because a significant portion of the upper metallization structure 160 is covered by a protective layer 118, such as a polyimide layer. In some embodiments, the source bond pads 114-1, 114-2 may be portions of the upper metallization structure 160 that are exposed through openings in the protective layer 118. Bond wires 119 are illustrated in FIG. 5A that may be used to connect the gate bond pad 112 and the source bond pads 114-1, 114-2 to external circuitry, or the like.

[0077] FIG. 5C is a schematic cross-sectional view taken along line 5C-5C of FIG. 5B. FIG. 5D is a schematic cross-sectional view taken along line 5D-5D of FIG. 5C. It will be understood that FIGS. 5C and 5D illustrate one entire unit cell and also illustrate portions of two unit cells on either side of it for context. FIG. 5E is an expanded view of one of the unit cells shown in FIG. 5C, omitting the source metallization structure.

[0078] 5B through 5D, a gate electrode structure 130 may be provided that includes a plurality of gate insulating fingers 132 (FIG. 5C), a plurality of gate fingers 134 (FIGS. 5B through 5D), a gate pad 136 (FIG. 5B), and one or more gate busses 138 (FIG. 5B) that electrically connect the gate fingers 134 to the gate pad 136. The electrical connections between the gate fingers 134, the gate busses 138, and the gate pad 136 may be conventional and therefore are not described or shown herein. The gate insulating fingers 132 may comprise, for example, silicon oxide, and may insulate the gate fingers 134 from the underlying semiconductor layer structure 120. In some embodiments, the gate fingers 134 may comprise, for example, a polysilicon pattern, although metal or other conductive patterns may also be used. The gate fingers 134 may extend horizontally across the entire device, or may comprise a planar layer that extends across the upper surface of the semiconductor layer structure 120 and has openings through which an upper source metallization structure 160 (described below) is connected to the source regions 128 of the semiconductor layer structure 120. In some embodiments, the gate pad 136 may be directly below the gate bond pad 112 and may be electrically connected to the gate bond pad 112. In other embodiments, the gate pad 136 may also be used as the gate bond pad 112. Other configurations are possible. In an exemplary embodiment, the gate pad 136 and the gate bus 138 may comprise polysilicon and / or metal.

[0079] 5C-5D, the intermetal dielectric pattern 150 can include a plurality of individual dielectric fingers 152 that cover the respective gate fingers 134 and gate bus 138. The intermetal dielectric pattern 150 can insulate the gate electrode structure 130 from the source metallization structure 160. As noted above, the gate fingers 134 of the electrode structure 130 can be polysilicon gate fingers. Thus, it is understood that in some embodiments, the "intermetal" dielectric pattern 150 can be a pattern formed to insulate a semiconductor pattern from a metal pattern, as opposed to isolating two metal patterns from each other.

[0080] An upper source metallization structure 160 may be formed on the intermetal dielectric pattern 150. The upper source metallization structure 160 includes a diffusion barrier layer 162 and a source contact layer 164. The MOSFET 110 includes a number of unit cell transistors that are electrically connected in parallel. Below, the individual layers / patterns of the MOSFET 110 are described in more detail.

[0081] 5C to 5E, the unit cell transistor is doped with n-type impurities (e.g., 1×10 18 atoms / cm 3 ~1×10 21 atoms / cm 3The semiconductor device may be formed on an n-type silicon carbide semiconductor substrate 122, such as a highly doped 4H single crystal silicon carbide semiconductor substrate (e.g., n-type or p-type). As used herein, the "doping concentration" of a semiconductor material refers to the number of dopant atoms present within one cubic centimeter of the semiconductor material that cause the semiconductor material to have a particular conductivity type (i.e., n-type or p-type), as measured using standard measurement techniques such as Secondary Ion Mass Spectrometry (SIMS). In some embodiments, the substrate 122 may have any suitable thickness (e.g., 100-500 microns thick) and may be partially or completely removed. It will be understood that the thickness of the substrate 122 is not drawn to scale in FIG. 5C.

[0082] A drain contact may be formed on the lower surface of the semiconductor substrate 122. The drain contact may act both as an ohmic contact to the semiconductor substrate 122 and as a drain bond pad 116 that provides an electrical connection between the drain terminal of the MOSFET 110 and an external device. In other embodiments, the drain contact may be separate from the drain bond pad 116 (e.g., a second layer may be formed on the drain contact that functions as the drain bond pad 116). In the embodiments described herein, a single metal layer 116 that functions as both an ohmic drain contact and a drain bond pad is formed on the lower surface of the semiconductor substrate 122. The drain contact / drain bond pad 116 may include, for example, metals and / or alloys such as nickel, titanium, tungsten, and / or aluminum, and / or thin layer stacks of these and / or similar materials.

[0083] Lightly doped n-type (n -) is provided on the upper surface of the substrate 122. The n-type silicon carbide drift region 124 may be formed, for example, by epitaxial growth on the silicon carbide substrate 122. The n-type silicon carbide drift region 124 may have a density of, for example, 1×10 14 ~5×10 16 Dopant / cm 3 The n-type silicon carbide drift region 124 may have a doping concentration of, for example, 3 to 100 microns above the substrate 122. It will be understood that the thickness of the drift region 124 is not drawn to scale in FIG. 5C. Although not shown in FIG. 5C, in some embodiments, an upper portion of the n-type silicon carbide drift region 124 may have a higher doping concentration (e.g., 1×10) than a lower portion thereof to provide a current spreading layer in the upper portion of the n-type silicon carbide drift region 124. 16 ~1×10 17 Dopant / cm 3 The doping concentration of the doped silicon is preferably 1000 .mu.m.

[0084] A p-type well region 126 is formed in the upper portion of the n-type drift region 124. A highly doped n-type (n + A silicon carbide source region 128 (e.g., a silicon carbide source region 128) may be formed in an upper portion of the well region 126 by, for example, ion implantation. A channel region 127 is defined laterally of the well region 126. The substrate 122, drift region 124, well region 126, and source region 128 together may comprise the semiconductor layer structure 120 of the MOSFET 110.

[0085] After the n-type source regions 128 are formed, a plurality of gate insulating fingers 132 (collectively forming a gate insulating pattern) may be formed on the upper surface of the semiconductor layer structure 120. Each gate insulating finger 132 may comprise an elongated strip of a dielectric material, such as, for example, silicon oxide, silicon nitride, silicon oxynitride, or the like. A gate finger 134, such as a polysilicon gate finger 134, is formed on each gate insulating finger 132. The gate fingers 134 and the gate insulating fingers 132 collectively form a gate electrode structure 130. As noted above, the vertically extending portions of the well region 126 between the source regions 128 and the portions of the drift region 124 directly below each gate finger 134 form a channel region 127. When a sufficient bias voltage is applied to the gate fingers 134, the channel region 127 electrically connects the n-type source regions 128 to the drift region 124. When a bias voltage is applied to gate finger 134 , current can flow from n-type source region 128 through channel region 127 to drift region 124 and to drain contact 116 .

[0086] An intermetal dielectric pattern 150, which may include a plurality of dielectric fingers 152, is formed in the form of spaced apart stripes of dielectric material. The intermetal dielectric pattern 150 may include a multi-layer dielectric pattern including a first non-reflowable dielectric pattern 154 and a second reflowed dielectric pattern 158. As will be described in more detail below, the second reflowed dielectric pattern 158 is formed by forming a second reflowable dielectric pattern 156 and then heating the device to reflow the material of the second reflowable dielectric pattern 156. The second reflowable dielectric pattern 156 is not shown in FIG. 5C but is shown in FIG. 6D (FIGS. 6A-6E illustrate a method of forming the intermetal dielectric pattern 150 of FIG. 5C).

[0087] The first non-reflowable dielectric pattern 154 may be formed on the exposed top surfaces of the gate fingers 134 and on the exposed side surfaces of the gate insulating fingers 132 and the gate fingers 134 in some embodiments. Thus, the first non-reflowable dielectric pattern 154 may include a plurality of non-reflowable dielectric fingers 152A covering the top and side surfaces of each gate finger 134 and the side surfaces of the underlying gate insulating fingers 132. The first non-reflowable dielectric pattern 154 may be, for example, a silicon oxide pattern, a silicon nitride pattern, or a combination of silicon oxide patterns and silicon nitride patterns, each of which is not a reflowable pattern. For example, a silicon oxide pattern that is undoped, doped only with phosphorus, or doped with phosphorus and a small amount of boron may be a non-reflowable pattern. However, it will be appreciated that a wide variety of other non-reflowable dielectric materials (e.g., aluminum oxide, magnesium oxide, or mixtures of these or other oxides and nitrides with silicon dioxide to form silicate or oxynitride mixed dielectrics) may be used to form the first non-reflowable dielectric pattern 154. Suitable materials may exhibit a relatively large bandgap (e.g., greater than 5 eV) and a relatively high breakdown voltage. In some embodiments, the first non-reflowable dielectric pattern 154 may include multiple layers of non-reflowable dielectric materials.

[0088] As mentioned above, intermetal dielectric patterns according to embodiments of the present invention can include both non-reflowable and reflowable dielectric layers / patterns. Reflowable dielectric layers and patterns have low viscosity when their temperature approaches or exceeds the glass transition temperature.

[0089] The first non-reflowable dielectric pattern 154 may be conformally formed over the gate fingers 134 and the underlying gate insulation fingers 132. Thus, the first non-reflowable dielectric material 154 may have a relatively uniform thickness. In some embodiments, the thickness of the first non-reflowable dielectric pattern 154 may be sufficient to prevent destruction of the inter-metal dielectric pattern 150 during operation of the device. In such embodiments, even if the second reflowed dielectric pattern 158 has extremely thin regions, the first non-reflowable dielectric pattern 154 may have a thickness sufficient to separate the gate fingers 134 from the source metallization structure 160 and prevent destruction during normal operation of the MOSFET 110. For example, the minimum thickness of the reflowed dielectric material pattern 158 may be at least 0.1 microns in some embodiments, or at least 0.2 microns in other embodiments. In other embodiments, the thickness of the first non-reflowable dielectric pattern 154 may be less than sufficient to prevent breakdown of the intermetal dielectric pattern 150, but the minimum combined thickness of the first non-reflowable dielectric pattern 154 and the second reflowed dielectric pattern 158 may be sufficient to prevent breakdown. In yet other embodiments, the thickness of the first non-reflowable dielectric pattern 154 may be greater than or equal to the thickness of the gate insulation fingers 132.

[0090] In some embodiments, the second reflowable dielectric pattern 156 may be formed directly on the first non-reflowable dielectric pattern 154. As described above, the first non-reflowable dielectric pattern 154 may include a plurality of non-reflowable dielectric fingers 152A covering the top and side surfaces of each gate finger 134 and the side surfaces of the underlying gate insulation fingers 132. The second reflowable dielectric pattern 156 may include a plurality of reflowable dielectric fingers 152B (see FIG. 6D) covering the top and side surfaces of each non-reflowable dielectric finger 152A. The second reflowable dielectric pattern 156 may be, for example, BPSG. BPSG is a boron and phosphorus doped silicon oxide material that is reflowable at temperatures between, for example, 500° C. and 1000° C. Other mixed elements other than boron and phosphorus may be used in the flowable glass and / or other flowable dielectric materials may be used. The second reflowable dielectric pattern 156 may be heated to reflow the dielectric material therein to form a second reflowed dielectric pattern 158. The second reflowed dielectric pattern 158 may include a plurality of dielectric fingers 152C. In representative embodiments, the ratio of the thickness of the non-reflowable dielectric pattern 154 to the thickness of the second reflowable dielectric pattern 156 may be 2:1, 1:2, 1:4, 1:7, and 1:10. That is, the ratio of the thickness of the non-reflowable dielectric pattern 154 to the thickness of the second reflowable dielectric pattern 156 may be between 2:1 and 1:10. In most applications, the thickness of the non-reflowable dielectric pattern 154 is less than the thickness of the second reflowable dielectric pattern 156.

[0091] As described in more detail with reference to Figures 6A-6E, the second reflowable dielectric pattern 156 may be conformally formed over the first non-reflowable dielectric pattern 154 and then reflowed by heating the second reflowable dielectric pattern 156 to a temperature of, for example, 800°C to 1000°C to convert the second reflowable dielectric pattern 156 into a second reflowed dielectric pattern 158. When the second reflowable dielectric pattern 156 is reflowed, the surface tension of the reflowed dielectric material that constitutes the second reflowed dielectric pattern 158 may cause the second reflowed dielectric pattern 158 to have a generally semi-elliptical or semi-circular cross-section, as shown in Figure 5C. The thickness of the portion of the second reflowed dielectric pattern 158 closest to the top end of the gate finger 134 (one such portion is enclosed in a dotted circle in Figure 5C) may be substantially less than the thickness of other portions of the second reflowed dielectric pattern 158. For example, as shown in FIG. 5E, which is an enlarged view (omitting source metallization structure 160) of a small portion of FIG. 5C, the thickness T1 of the second reflowed dielectric pattern 158 adjacent to the upper surface end of the gate finger 134 may be significantly smaller than the thickness T2 of the upper portion of the second reflowed dielectric pattern 158. In some embodiments, the thickness T2 may be at least two, three, four, five, or six times the thickness T1. Providing the first non-reflowable dielectric pattern 154 below the second reflowed dielectric pattern 158 can ensure that the total thickness of the intermetal dielectric pattern 150 adjacent to the upper end of the gate finger 134 may be sufficient to prevent damage. For example, as shown in FIG. 5E, the distance T3 representing the thickness of the intermetal dielectric pattern 150 at the upper end of the gate finger 134 may be designed to be sufficient to avoid damage.

[0092] In some embodiments, the maximum thickness of the intermetal dielectric pattern 150 may be above the central top surface of the gate finger 134 (i.e., T4). The maximum thickness of the intermetal dielectric pattern 150 may be adjacent to the upper edge of the gate finger 134 (i.e., adjacent to the upper corner of the gate finger 134). In various embodiments of the present invention, the ratio of the thickness of the intermetal dielectric pattern 150 above the central top surface of the gate finger 134 (i.e., T4) to the minimum thickness of the intermetal dielectric pattern 150 (typically T3 in FIG. 5E) may be less than 8:1, less than 6:1, less than 5:1, less than 4:1, less than 3:1, or even less than 2:1, and may be at least 1:1 in each of these embodiments.

[0093] In some embodiments, the combination of the non-reflowable dielectric pattern 154 and the second reflowed dielectric pattern 158 can have a thickness approximately the same as the thickness of the reflowed BPSG intermetal dielectric pattern 50A in the embodiment of Figure 3. Thus, the second reflowed dielectric pattern 158 can have a thickness less than the reflowed BPSG intermetal dielectric pattern 50A of Figure 3, and therefore the second reflowable dielectric pattern 156 does not expand laterally as much during reflow, thereby maintaining an appropriate width for the gap 140. Thus, the intermetal dielectric pattern 150 according to embodiments of the present invention can have the advantages of a fully reflowable intermetal dielectric pattern while reducing or minimizing the potential disadvantages associated with a fully reflowable intermetal dielectric pattern.

[0094] Referring again to FIG. 5C, n-type source regions 128 are exposed in gaps 140 between adjacent reflowed dielectric fingers 152C. As noted above, in silicon carbide power devices, the width of each gap 140 can be very narrow, such as on the order of 1 to 5 microns. Furthermore, as technology continues to develop, gaps 140 are becoming smaller, and in the near future gaps 140 as small as 0.5 microns may be realized in commercial devices. With such small gaps 140, it may be difficult to fill the upper source metallization structure 160 (described below) without forming voids in the upper source metallization structure 160.

[0095] The second reflowed dielectric pattern 158 may have rounded edges, and thus angles α1 and α2 (see FIG. 5E) may be smaller, resulting in gaps 140 that are easier to completely fill with diffusion barrier layer metal. Thus, the MOSFET 110 may be less susceptible to void formation in the portions of the diffusion barrier layer 162 that are within the gaps 140.

[0096] An upper source metallization structure 160 is formed on the intermetal dielectric pattern 150 and on the well region 126 and the n-type source region 128 of the semiconductor layer structure 120 exposed at the gap 140. The upper source metallization structure 160 includes a diffusion barrier layer 162 and a source contact layer 164 formed on the diffusion barrier layer 162. The diffusion barrier layer 162 can be a metal or metal-containing layer that is relatively inert with respect to adjacent layers and that retards or substantially prevents diffusion of other materials therethrough. The diffusion barrier layer 162 can include a continuous conductive layer that includes, for example, titanium, tungsten, tantalum, nickel, hafnium, and / or indium. For example, the diffusion barrier layer 162 can include titanium, tantalum, nickel, hafnium, tungsten, titanium nitride, tungsten nitride, indium oxide, or tantalum nitride, or an alloy of the aforementioned materials. A diffusion barrier layer 162 may be conformally formed over exposed portions of the semiconductor layer structure 120 (e.g., over the well regions 126 and the source regions 128) and over the inter-metal dielectric pattern 150. The diffusion barrier layer 162 is typically formed of a metal that is less conductive than the metal included in the source contact layer 164. As such, the diffusion barrier layer 162 may be a relatively thin layer to reduce its impact on the resistance of the upper source metallization structure 160.

[0097] The source contact layer 164 may be conformally formed on the diffusion barrier layer 162. The source contact layer 164 may include, for example, metals and / or alloys such as nickel, titanium, tungsten, and / or aluminum, and / or thin layer stacks of these and / or similar materials. In some embodiments, the source contact layer 164 may include an aluminum layer because aluminum is relatively inexpensive, highly conductive, easy to deposit, and can act as a good seed layer for other metals. The source contact layer 164 may be significantly thicker than the diffusion barrier layer 162. The thickness of the source contact layer 164 may reflect a tradeoff between the (desirably low) resistance of the upper source metallization structure 160 and the depth of the keyholes 166 that tend to form in the portions of the source contact layer 164 that overlie the source regions 128. Generally speaking, as the thickness of the source contact layer 164 increases, the resistance of the source contact layer 164 increases, but the average depth of the keyholes 166 is reduced. As shown in FIG. 5C, the gate electrode structure 130 is between the wide bandgap semiconductor layer structure 120 and the intermetal dielectric pattern 150, which is between the gate electrode structure 130 and the source metallization structure 160.

[0098] Figures 6A-6E are schematic cross-sectional views illustrating the formation of an inter-metal dielectric pattern 150 for the MOSFET of Figures 5A-5E. Figures 6A-6E correspond to expanded views of Figure 5E and show a single unit cell for simplicity of drawing.

[0099] As shown in FIG. 6A, a non-reflowable dielectric layer 155 is conformally formed on the top surface of the device to cover the sidewalls of the gate fingers 134 and the gate insulating fingers 132. As shown in FIG. 6B, the non-reflowable dielectric layer 155 is then selectively etched to form a first non-reflowable dielectric pattern 154. The first non-reflowable dielectric pattern 154 includes a plurality of dielectric fingers 152A (only one of which is shown in FIG. 6B), which cover the sidewalls of each of the gate fingers 134 and each of the gate insulating fingers 132. As shown in FIG. 6C, a reflowable dielectric layer 157 is conformally formed on the first non-reflowable dielectric pattern 154. As shown in FIG. 6D, the reflowable dielectric layer 157 is then selectively etched to form a second reflowable dielectric pattern 156. The second reflowable dielectric pattern 156 includes a plurality of dielectric fingers 152B (only one is shown in FIG. 6D) covering each dielectric finger 152A. Finally, as shown in FIG. 6E, the device is heated to a temperature between, for example, 500° C. and 1000° C., thereby reflowing the material of the second reflowable dielectric pattern 156 and transforming the second reflowable dielectric pattern 156 into a second reflowed dielectric pattern 158 and transforming the individual reflowable dielectric fingers 152B into reflowed dielectric fingers 152C. Finally, a diffusion barrier layer 162 may be conformally formed on the second reflowed dielectric pattern 158 and on the exposed portions of the well region 126 and the source region 128. As mentioned above, the inclusion of the first non-reflowable dielectric pattern 154 in the intermetal dielectric pattern 150 can ensure that the minimum thickness of the intermetal dielectric pattern 150 is sufficient to prevent damage during device operation.

[0100] In the manufacturing process for the intermetal dielectric pattern 150 described above with reference to FIGS. 6A-6E, the second reflowable dielectric pattern 156 is first etched to expose the source region 128 and then reflowed to form the second reflowed dielectric pattern 158. However, it will be appreciated that in other embodiments, the second reflowable dielectric layer 157 may be reflowed to form a reflowed dielectric layer, and the reflowed dielectric layer may then be etched to form the second reflowed dielectric pattern 158 exposing the source region 128. Reflowing the second reflowable dielectric layer 157 (i.e., performing the reflow before etching) may be advantageous in certain applications because it allows removal of portions of the reflowed dielectric layer that expand laterally during the reflow process while maintaining the proper dimensions of the opening in the intermetal dielectric pattern 150 exposing the source region 128.

[0101] 7A and 7B are schematic cross-sectional views illustrating another method of forming the intermetal dielectric pattern 150 of the MOSFET 110 of FIGS. 5A-5E. As shown in FIG. 7A, a gate electrode structure including a plurality of gate insulating fingers 132 and a plurality of gate fingers 134 is formed on the upper surface of the semiconductor layer structure 120. The gate insulating fingers 132 and the gate fingers 134 can be wider than the corresponding gate insulating fingers 132 and the gate fingers 134 included in the embodiment of FIGS. 6A-6E, and the gate fingers 134 can be thicker than the corresponding gate fingers 134 included in the embodiment of FIGS. 6A-6E. Next, as shown in FIG. 7B, a non-reflowable dielectric pattern 154 is formed by oxidizing the exposed top and side surfaces of the gate fingers 134. The gate fingers 134 can include, for example, polysilicon gate fingers 134 that are not doped with boron. Therefore, the oxidation process of the gate fingers 134 converts the top and side portions of each gate finger 134 into respective non-reflowable dielectric fingers 152A. The oxidation process may be designed to oxidize a predetermined thickness of the outer portion of each gate finger 134 to form a first non-reflowable dielectric pattern 154 having a desired thickness on each gate finger 134. The oxidation process may be performed at a temperature high enough to oxidize polysilicon but not high enough to oxidize silicon carbide. As a result, the exposed silicon carbide source regions 128 are not appreciably oxidized during the oxidation process. After the first non-reflowable dielectric pattern 154 is formed by the oxidation process, a second reflowable dielectric pattern 156 may be formed and then reflowed in the manner described above with reference to FIGS. 6C-6E. Notably, by using an oxidation process to form the first non-reflowable dielectric pattern 154, the etching step described above with reference to FIG. 6B may be omitted, thereby simplifying the manufacturing process.

[0102] 8A-8F are cross-sectional views illustrating a method of forming an inter-metal dielectric pattern 250 that can be used in place of the inter-metal dielectric pattern 150 of the MOSFET 110 of FIGS. 5A-5E. As shown in FIG. 8A, a gate insulating layer 131 and a gate finger layer 133 can be formed on the semiconductor layer structure 120. Then, as shown in FIG. 8B, a non-reflowable dielectric layer 255 (e.g., a silicon dioxide layer) is formed on the gate finger layer 133. The non-reflowable dielectric layer 255 can be formed, for example, by plasma enhanced chemical vapor deposition, low pressure chemical vapor deposition, or oxidation of the top surface of the gate electrode structure 130. As shown in FIG. 8C, an etching step is performed to etch the gate insulating layer 131, the gate finger layer 133, and the non-reflowable dielectric layer 255 to expose the source region 128. Thereby, the gate insulating layer 131, the gate finger layer 133, and the non-reflowable dielectric layer 255 are transformed into a first non-reflowable dielectric pattern 254 including a plurality of gate insulating fingers 132, a plurality of gate fingers 134, and a plurality of dielectric fingers 252A. Note that in this embodiment, the dielectric fingers 252A only cover the top surfaces of each of the gate fingers 134, and do not cover the side surfaces of the gate fingers 134.

[0103] As shown in FIG. 8D, a second reflowable dielectric layer 257 is conformally formed on the device. As shown in FIG. 8E, the second reflowable dielectric layer 257 is then selectively etched to form a second reflowable dielectric pattern 256. The second reflowable dielectric pattern 256 includes a plurality of reflowable dielectric fingers 252B covering the respective dielectric fingers 252A. Finally, as shown in FIG. 8F, the device is heated to a temperature between, for example, 500° C. and 1000° C. to reflow the material of the second reflowable dielectric pattern 256 to transform the second reflowable dielectric pattern 256 into a second reflowed dielectric pattern 258. The second reflowed dielectric pattern 258 includes a plurality of reflowable dielectric fingers 252C. This technique for forming the intermetal dielectric pattern 250 may further require one less etching step than the technique described above with reference to FIGS. 6A-6E. It will be appreciated that, alternatively, in other embodiments, the second reflowable dielectric layer 257 may be reflowed and then etched to form the second reflowed dielectric pattern 258.

[0104] According to a further embodiment of the present invention, a power semiconductor device is provided that includes a reflowed intermetal dielectric pattern that is formed using sacrificial structures during the reflow process. In particular, the sacrificial structures can be deposited on the device before or after the formation of the reflowable dielectric pattern. These sacrificial structures are left in place during the reflow process and can be used to (1) limit lateral flow of the reflowable dielectric material during the reflow process and / or (2) increase the amount of reflowable dielectric material adjacent the lateral edges of the gate fingers 134 after the reflow process is completed. A method of forming such a power semiconductor device is illustrated in Figures 9A-9D.

[0105] As shown in FIG. 9A, semiconductor layer structure 120 may be formed having gate insulating finger 132 and gate finger 134 formed thereon. Referring to FIG. 9B, a reflowable dielectric layer (not shown) may be formed over an upper surface of the device and then etched to form a reflowable dielectric pattern 356 including dielectric fingers covering the top and sides of each gate finger 134 and the sides of each gate insulating finger 132. A sacrificial dam 390 may then be formed in gap 140. Sacrificial dam 390 may be selectively deposited in gap 140. Alternatively, a sacrificial layer may be conformally formed on the upper surface of the device and then selectively etched to form sacrificial dam 390. In an exemplary embodiment, sacrificial dam 390 may comprise polysilicon since polysilicon may be etched with high selectivity to the BPSG reflowed dielectric pattern.

[0106] As shown in FIG. 9C, the device can then be heated to reflow the reflowable dielectric pattern 356, thereby transforming the reflowable dielectric pattern 356 into a reflowed dielectric pattern 358 including a plurality of dielectric fingers 352C. As described above, such a reflow process generally causes the reflowable dielectric material to change its lateral spreading shape to have a semi-elliptical or semi-circular cross section. However, the sacrificial dam 390 prevents the reflowable dielectric material from lateral spreading. Thus, the size of the gap 140 can be maintained. In this manner, the lower portion of the reflowed dielectric material pattern 358 has substantially vertical sidewalls, as shown in FIG. 9C. In addition, because the reflowable dielectric material is prevented from lateral spreading, more reflowable dielectric material is preserved along the sidewalls of the gate fingers 134. This acts to increase the thickness of the dielectric fingers 352C along the lateral edges of the gate fingers 134. Thus, the use of sacrificial dam 390 can reduce or prevent two potential problems described above with reference to Figure 3 that can occur when intermetal dielectric pattern 150 is formed using a reflowable dielectric material. Referring to Figure 9D, after the reflow process is completed, sacrificial dam 390 can be removed by a selective etch. It should be noted that while in the above description sacrificial dam 390 is formed after reflowable dielectric pattern 356 is formed, in other embodiments sacrificial dam 390 can be formed before reflowable dielectric pattern 356 is formed.

[0107] In the above description of embodiments including an intermetal dielectric pattern having both a non-reflowable dielectric pattern and a reflowable dielectric pattern, the non-reflowable dielectric pattern is formed before the reflowable dielectric pattern, such that the non-reflowable dielectric pattern is between the gate fingers and the reflowable dielectric pattern. However, it will be understood that embodiments of the invention are not so limited. In particular, each of the above described embodiments including an intermetal dielectric pattern having both a non-reflowable dielectric pattern and a reflowable dielectric pattern may be modified such that the reflowable dielectric pattern (or layer) is formed first (e.g., directly on the gate fingers) and then reflowed, and the non-reflowable dielectric pattern is formed over the reflowed dielectric pattern. For example, FIG. 10 is a schematic cross-sectional view (with a portion of its source metallization omitted) of a unit cell of a MOSFET 410, which is a modified version of MOSFET 110 of FIGS. 5A-5E. As shown in FIG. 10, the intermetal dielectric pattern 450 of the MOSFET 410 includes a reflowed dielectric pattern 458 formed directly on the gate fingers 134 and a non-reflowable dielectric pattern 454 formed on the reflowed dielectric pattern 458.

[0108] It will be understood that the MOSFETs described above are n-type devices having a source contact layer 164 on their top surface and a drain contact 116 on their bottom surface, whereas in a p-type device these positions would be reversed. Additionally, while the power MOSFET 110 described above and other devices described herein are shown as being silicon carbide based semiconductor devices, it will be understood that embodiments of the invention are not so limited. Instead, the semiconductor devices may include any wide bandgap semiconductor suitable for power semiconductor device applications, including, for example, gallium nitride based semiconductor devices and II-VI compound semiconductor devices.

[0109] It will also be appreciated that the non-reflowable and / or reflowable dielectric layers / patterns forming the intermetal dielectric patterns described herein may each comprise a single layer or a multi-layer structure, in many applications a multi-layer structure may be more effective as a diffusion barrier and / or to provide chemical compatibility with surrounding layers.

[0110] It will also be appreciated that the improved intermetal dielectric patterns disclosed herein may be used in power semiconductor devices other than vertical power MOSFET devices. For example, these intermetal dielectric patterns may be used in power IGBT devices. A power IGBT device, as known to those skilled in the art, is a combination of a BJT and a MOSFET, where the MOSFET feeds the base of the BJT to convert the BJT from a current controlled device to a voltage controlled device.

[0111] 11A is a schematic cross-sectional view of a small portion of an IGBT 500 according to several embodiments of the present invention. As shown in FIG. 11A, the IGBT 500 includes a pnp silicon carbide BJT 502 having a base 504, an emitter 506, and a collector 508. The IGBT 500 further includes a silicon carbide MOSFET 510 having a gate 130, a source 160, and a drain 116. The source 160 of the MOSFET 510 is electrically connected to the base 504 of the BJT 502, and the drain 116 of the silicon carbide MOSFET 510 is electrically connected to the collector 508 of the BJT 502. By convention, the collector 508 of the BJT 502 is the "emitter" of the IGBT 500, the emitter 506 of the BJT 502 is the "collector" of the IGBT 500, and the gate 130 of the MOSFET 510 is the "gate" of the IGBT 500.

[0112] 11B is a schematic cross-sectional view of a small portion of the IGBT 500 of FIG. 11A. As shown in FIG. 11B, the IGBT 500 is, for example, a heavily doped (p +) p-type silicon carbide layer 121. P-type layer 121 may be epitaxially grown on a silicon carbide substrate, for example, which may then be removed. + Layer 121 serves as the collector of IGBT 500 (and also as the emitter 506 of BJT 502). - 5. A silicon carbide drift region 124 is provided on the P-type layer 121 which functions as the base of the BJT 502 and as the source of the MOSFET 510. A lightly doped p-type well region 126 is provided in the upper portion of the n-type drift region 124. The upper portion of each p-type well 126 is more heavily doped with p-type dopants to form a heavily doped p-type well 126 which also functions as the collector of the BJT 502. + The emitter region can be formed by highly doped (n + ) An n-type drain region 128 may also be formed in an upper portion of each p-type well 126 that serves as a shared drain for the IGBT 500. A source metallization structure 160 is formed in the p-type well regions 126 and the n + formed to contact the drain region 128, + An ohmic contact 116 is formed on the underside of the silicon carbide layer 121 .

[0113] The gate insulating finger 132 and the gate finger 134 are provided on the upper surface of the semiconductor layer structure. An intermetal dielectric pattern 550 is formed, including a first non-reflowable dielectric pattern 554 and a second reflowed dielectric pattern 558 (each of which is described in detail above). Gaps 140 are provided between adjacent dielectric fingers, exposing the n-type source region 128 and the p-type well region 126. An upper source metallization structure 160 is formed on the intermetal dielectric pattern 550 and on the exposed n-type source region 128 and the p-type well region 126. The upper metallization structure 160 includes a diffusion barrier layer 162 and a source contact layer 164. It will be understood that the intermetal dielectric pattern 550 may be substituted for any of the intermetal dielectric patterns disclosed herein.

[0114] FIG. 12 is a flow chart of a method of forming a power semiconductor device according to an embodiment of the present invention. As shown in FIG. 12, operations may begin with the formation (block 600) of a wide bandgap semiconductor layer structure ("SLS") including a plurality of semiconductor layers. Next (block 610), a conductive pattern, which may include, for example, a plurality of gate fingers, is formed on an upper surface of the wide bandgap semiconductor layer structure. (block 620) A first non-reflowable dielectric material pattern is formed on the conductive pattern. (block 630) A second reflowable dielectric material pattern is formed on the first non-reflowable dielectric material pattern. (block 640) The reflowable dielectric material pattern is reflowed to form an inter-metal dielectric pattern including the non-reflowable dielectric material pattern and the reflowed dielectric material pattern. (block 650) A source metallization structure is formed on the inter-metal dielectric pattern and on the exposed upper surface of the wide bandgap semiconductor layer structure.

[0115] FIG. 13 is a flow chart of a method of forming a power semiconductor device according to a further embodiment of the present invention. As shown in FIG. 13, operations may begin with the formation (block 700) of a wide bandgap semiconductor layer structure including a plurality of semiconductor layers. Next (block 710), a plurality of spaced apart gate fingers are formed on an upper surface of the wide bandgap semiconductor layer structure. (block 720) A reflowable dielectric material pattern is formed on the gate finger pattern. (block 730) A sacrificial structure is formed in each gap in the reflowable dielectric material pattern. (block 740) The reflowable dielectric material pattern is reflowed to form an intermetal dielectric pattern. (block 750) A source metallization pattern is formed on the intermetal dielectric pattern.

[0116] Thus, embodiments of the present invention may provide power MOSFETs that may be less susceptible to device failure due to gate-to-source electrical shorts by providing intermetal dielectric patterns with improved shape and increased density while still maintaining a minimum desired thickness to avoid damage. These improved intermetal dielectric patterns may substantially eliminate seams that provide a path for moisture or wet etchants to diffuse through the intermetal dielectric patterns, reducing the likelihood of electrical shorts between the gate and source metallizations. Thus, power MOSFETs according to embodiments of the present invention may have lower failure rates.

[0117] The problems addressed by techniques according to embodiments of the present invention tend to be particular to wide bandgap power semiconductor devices, such as silicon carbide and / or gallium nitride based devices. In such devices, the gaps (e.g., gap 140) between adjacent dielectric fingers (e.g., dielectric finger 152) can be much smaller than the corresponding gaps provided in MOSFETs formed in narrow bandgap semiconductor devices, such as silicon based devices. For that matter, in silicon MOSFETs, gaps in the source contact layer tend not to form in the gaps between adjacent dielectric fingers, and any keyholes that are formed tend to be very shallow because the size of the gaps is so large. Thus, in narrow bandgap devices, wet etchants tend not to weep through the source contact metallization, and therefore there is less or no need to densify the intermetal dielectric pattern and / or round the shape of the intermetal dielectric pattern.

[0118] It should be noted that there are other potential ways to address the problem of corrosive materials diffusing through the top source metallization structure 60 of the conventional power MOSFET 10 described above with reference to FIG. 1. As one example, the aspect ratio of the gap 40 can be reduced, for example, by increasing the width of the source region 28. This can improve the gap-filling characteristics of the top source metallization structure 60. As another example, modification of deposition parameters can be used to improve the gap-filling characteristics of the material of the source contact layer 64 (e.g., increasing the deposition temperature can promote improved gap-filling). Furthermore, the gap-filling characteristics can be improved using a high-precision deposition tool. Additionally, other deposition techniques, such as atomic layer deposition, can be used to form portions of the source metallization structure 160 (e.g., portions of the gap 40) to obtain more consistent coverage and reduce or prevent gaps.

[0119] The present invention has been described above with reference to the accompanying drawings, in which examples of the invention are shown. However, the present invention may be embodied in many different forms and should not be construed as limited to the examples set forth herein. Rather, these examples are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art. In the drawings, the sizes and relative dimensions of layers and regions may be exaggerated for clarity. When an element or layer is referred to as being "on", "connected to", or "coupled to" another element or layer, it will be understood that it may be directly on, connected to, or coupled to the other element or layer, or there may be intervening elements or layers. In contrast, when an element is referred to as being "directly" or "directly" on, "connected to", or "coupled to" another element or layer, there are no intervening elements or layers. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items. Like reference numbers refer to like elements throughout this disclosure.

[0120] It will be understood that although the terms "first" and "second" are used herein to describe various regions, layers, and / or elements, these regions, layers, and / or elements should not be limited by these terms. These terms are only used to distinguish one region, layer, or element from another region, layer, or element. Thus, a first region, layer, or element described below may be referred to as a second region, layer, or element, and similarly, a second region, layer, or element may be referred to as a first region, layer, or element, without departing from the scope of the present invention.

[0121] In this specification, relative terms such as "lower" or "bottom" and "upper" or "above" may be used to describe the relationship of one element to another element as shown in the figures. It will be understood that the relative terms are intended to encompass various orientations of the device, not just the orientation depicted in the figures. For example, if a device in a figure is rotated, an element described as being "below" another element would be oriented to be "above" that other element. Thus, the example term "lower" may encompass both an orientation of "lower" and "upper", depending on the particular orientation of the figure. Similarly, if a device in one of the figures is rotated, an element described as being "below" or "below" another element would be oriented "above" that other element. Thus, the example term "below" or "below" may encompass both an orientation of up and down.

[0122] The terms used herein are for the purpose of describing specific embodiments only and are not intended to limit the present invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the words "comprises", "comprising", "includes" and / or "including" as used herein specify the presence of stated features, elements and / or components and do not exclude the presence or addition of one or more other features, elements, components and / or groups thereof.

[0123] In the present specification, embodiments of the present invention are described with reference to cross-sectional views that are schematic illustrations. As a result, variations from the illustrated configurations (e.g., as a function of manufacturing techniques and / or tolerances) may be envisioned. Thus, embodiments of the present invention should not be construed as limited to the particular shapes of the regions illustrated herein, but rather include deviations in shape that may result from manufacturing, for example. For example, an implanted region illustrated as a rectangle will typically have rounded or curved features at its ends and / or a gradient of implant concentration rather than a binary transition from implanted to unimplanted regions. Thus, the regions illustrated in the figures are schematic in nature, and their shapes are not intended to illustrate the actual shapes of the regions of a device and are not intended to limit the scope of the present invention.

[0124] It will be understood that the embodiments disclosed herein may be combined. Thus, features illustrated and / or described with respect to a first embodiment may be included in a second embodiment as well, and vice versa.

[0125] Although the above embodiments are described with reference to certain figures, it should be understood that some embodiments of the invention can include additional and / or intervening layers, structures, or elements, and / or certain layers, structures, or elements can be omitted. Although some exemplary embodiments of the invention have been described, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without substantially departing from the novel teachings and advantages of the invention. Accordingly, all such modifications are intended to be included within the scope of the invention as set forth in the claims. Accordingly, the foregoing is illustrative of the invention and should not be construed as being limited to the specific embodiments disclosed, but rather, modifications to the disclosed embodiments, as well as other embodiments, are intended to be included within the scope of the appended claims. The present invention is defined by the following claims (with equivalents of the claims included therein).

Claims

1. a wide band gap semiconductor layer structure; a gate electrode structure on an upper surface of the wide bandgap semiconductor layer structure, the gate electrode structure including a gate finger separated from the wide bandgap semiconductor layer structure by a gate insulating finger having a substantially uniform thickness in a direction perpendicular to a plane defined by the upper surface of the wide bandgap semiconductor layer structure; an intermetal dielectric pattern on the gate electrode structure, the intermetal dielectric pattern including a non-reflowed dielectric material pattern and a reflowed dielectric material pattern; a source metallization structure on said intermetal dielectric pattern; Including, the gate electrode structure is between the wide bandgap semiconductor layer structure and the intermetal dielectric pattern; the intermetal dielectric pattern is between the gate electrode structure and the source metallization structure; a thickness of the reflowed dielectric material pattern above a center of the gate electrode structure is at least three times a thickness of the reflowed dielectric material pattern adjacent an upper edge of the gate electrode structure.

2. 2. The semiconductor device of claim 1, wherein the gate electrode structure includes a plurality of gate fingers separated from the wide bandgap semiconductor layer structure by each of a plurality of gate insulating fingers, and the non-reflowed dielectric material pattern includes a plurality of non-reflowed dielectric fingers conformally surrounding top and side surfaces of each of the gate fingers.

3. 2. The semiconductor device of claim 1, wherein a top surface of the reflowed dielectric material pattern has a semi-elliptical cross-section.

4. 2. The semiconductor device of claim 1, wherein a thickness of the non-reflowed dielectric material pattern exceeds a minimum thickness of the reflowed dielectric material pattern and does not exceed a maximum thickness of the reflowed dielectric material pattern.

5. 5. The semiconductor device of claim 4, wherein a portion of the reflowed dielectric material pattern adjacent an upper corner of the non-reflowed dielectric material pattern is the portion of the reflowed dielectric material pattern having the minimum thickness.

6. 2. The semiconductor device of claim 1, wherein the reflowed dielectric material pattern comprises a boro-phospho-silicate glass (BPSG) pattern.

7. 2. The semiconductor device of claim 1, wherein a ratio of a thickness of the reflowed dielectric material pattern above a center of a top surface of the gate finger of the gate electrode structure to a minimum thickness of the reflowed dielectric material pattern is greater than 3:

1.

8. 2. The semiconductor device of claim 1, wherein the reflowed dielectric material pattern over the gate electrode structure has a semi-elliptical top surface, and the reflowed dielectric material pattern has vertical outer sidewalls.

9. 10. The semiconductor device of claim 1, wherein a minimum thickness of the non-reflowed dielectric material pattern is selected to be sufficient to avoid damage to the intermetal dielectric pattern during normal operation of the semiconductor device.

10. a wide band gap semiconductor layer structure; a gate electrode structure on an upper surface of the wide bandgap semiconductor layer structure; an intermetal dielectric pattern on the gate electrode structure, the intermetal dielectric pattern including at least a first reflowed dielectric material pattern and a second non-reflowed dielectric material pattern including a different material than the first reflowed dielectric material pattern; a source metallization structure on said intermetal dielectric pattern; Including, the gate electrode structure is between the wide bandgap semiconductor layer structure and the intermetal dielectric pattern; the intermetal dielectric pattern is between the gate electrode structure and the source metallization structure; A semiconductor device, wherein a maximum thickness of the intermetal dielectric pattern is above a top surface at a center of the gate electrode structure.

11. 11. The semiconductor device of claim 10, wherein the first reflowed dielectric material pattern has a semi-elliptical top surface and vertical outer sidewalls.

12. 11. The semiconductor device of claim 10, wherein a top surface of the first reflowed dielectric material pattern and a top surface of the second non-reflowed dielectric material pattern each have a semi-elliptical cross-section.

Images