Method of forming cross point type DRAM cell

[0001] This invention relates generally to dynamic random access memory (DRAM) cell devices and particularly relates to a realistic formation method for such devices which does not require poor step coverage of the thin films employed.

[0002] Continued growth in the capacity of dynamic random access memory (DRAM) technology can be enhanced by minimizing the size of individual DRAM cells. As technology develops, DRAM devices need to store more bits of information, use less power per bit of stored information and have the individual memory cells containing the stored bits occupy less area on the semiconductor chip. The smallest theoretical limit of a memory cell is four times the square of the feature size on a chip, or 4F² where F represents minimum feature size. The feature size is the minimum width of the electrical leads or wiring and minimum spacing between wiring on the chip. Each cell comprises a MOS pass transistor and a storage capacitor.

[0003] A memory cell which achieves this type of minimum size is known as a cross point cell. The name cross point cell is derived from the location of the cell in an array of cells at the point where one word line and one bit line cross each other. Such cells normally employ an open bit line architecture. The open bit line architecture can result in detrimental substrate noise coupling into the bit lines, and can require an excessive number of masks to form the cross point cell.

[0004] In an effort to address these problems, the invention described in U.S. patent application Ser. No. 08/934,157 was developed. The 08/934,157 application is expressly incorporated herein by reference, and is co-assigned to Texas Instruments Incorporated. Although the DRAM cell and manufacturing method disclosed in the referenced application are quite promising, the formation method in the referenced application depends on deliberately obtaining poor step coverage of certain materials during the fabrication process. In particular, poor step coverage was desired at the bottom of holes during fabrication of the structures. It is difficult to obtain consistent manufacturing quality when the method requires deliberately seeking poor step coverage.

[0005] Further, in the prior art exemplified in the 08/934,157 application, each functional part of the cross point DRAM cell is formed as a portion of a symmetrical pillar. Therefore, each element has a double layer, due to the symmetry of the cell about a center line. Accordingly, to minimize the cell dimensions, the thickness of each layer had to be quite thin, resulting in added manufacturing difficulties.

[0006] Therefore, a need has arisen for a realistic method of forming a cross point type DRAM cell which does not depend on poor step coverage of the thin films, and which can be formed in an asymmetric fashion so that each element of the cell need not be divided into two excessively thin layers. The present invention provides such a realistic formation method of the cross point type DRAM cell wherein the process does not depend on poor step coverage of the thin films. Further, each individual DRAM cell can be formed in an asymmetric fashion with more realistic thicknesses for the various components which are employed.

[0007] In accordance with the present invention, a dynamic random access memory (DRAM) device comprises a substrate of semiconductor material having a main outer surface; a plurality of substantially parallel word lines which are located outwardly of the main outer surface of the substrate; and a plurality of substantially parallel bit lines which are also located outwardly of the main outer surface of the substrate. Each of the word lines has a word line width defined by first and second word line edges and is separated from adjacent word lines by a word line separation distance. Similarly, each of the bit lines has a bit line width defined by first and second bit line edges and is separated from adjacent bit lines by a bit line separation distance. The bit lines are substantially perpendicular to the word lines.

[0008] The device further comprises a plurality of memory cells each associated with an intersection of a given one of the fist and second world line edges of a given one of the word lines and a given one of the first and second bit line edges of a given one of the bit lines. Each of the memory cells in turn comprises a pillar of semiconductor material; a storage node; a storage node plug; a bit line plug; and a word line plug. The pillar of semiconductor material extends outward from the main outer surface of the substrate and is electrically isolated from adjacent memory cells. The pillar has a pillar outer surface and includes an active region of the semiconductor material which is substantially centered inward of the intersection. The pillar is formed with a hole region which is defined by a hole floor and at least one hole wall. The hole floor is located outwardly of the main outer surface of the substrate.

[0009] The storage node is offset from the active region of the pillar in a direction away from the given one of the word lines and the given one of the bit lines. The storage node plug extends from the storage node, through the hole region in the pillar, to contact the hole floor and to form thereat a storage node contact and one of a drain and a source of a MOS transistor. The storage node plug is electrically isolated from the at least one hole wall.

[0010] The bit line plug extends from the given one of the bit lines inwardly to contact the pillar outer surface and to form thereat a bit line contact and the other of the drain and the source of the MOS transistor. The word line plug extends from the given one of the word lines through the hole region in the pillar and terminates outwardly of the hole floor. The word line plug is electrically isolated from the semiconductor material of the pillar, such that a portion of the word line plug adjacent to and electrically isolated from the semiconductor material of the pillar forms a gate of the MOS transistor. A corresponding portion of the semiconductor material of the pillar forms a channel region of the MOS transistor.

[0011] In accordance with the present invention, a method for manufacturing a dynamic random-access memory device comprises the steps of providing a substrate of semiconductor material; forming a plurality of pillars of semiconductor material extending outward from the substrate; forming a plurality of substantially parallel bit lines; interconnecting the bit lines with outer surfaces of the pillars via a plurality of bit line plugs; forming at least one hole region in each of the plurality of pillars; forming a plurality of substantially parallel word lines; locating a word line plug in each of the hole regions; forming a plurality of storage nodes; and interconnecting the storage nodes with the floors of the hole regions in the pillars via a plurality of storage node plugs. The substrate of semiconductor material which is provided can have a main outer surface. The pillars can extend outwardly from the main outer surface of the substrate, and each of the pillars can have a pillar outer surface. The pillars can be formed using a first mask, and can be electrically isolated from each other. The pillars can include an active region of the semiconductor material.

[0012] In the step of forming the plurality of substantially parallel bit lines, the bit lines can be located outwardly of the main outer surface of the substrate, and each of the bit lines can have a bit line width defined by first and second bit line edges. Further, each of the bit lines can be separated from adjacent bit lines by a bit line separation distance. The bit lines can be formed using a second mask.

[0013] In the step of interconnecting the bit lines with the pillar outer surfaces via the bit line plugs, the result can be the formation of a plurality of bit line contacts and one of drains and sources of a plurality of MOS transistors at the intersections of the bit line plugs with the pillar outer surfaces. In the step of forming the at least one hole region, the region can be defined by a hole floor and at least one hole wall. Such a hole region can be formed in each of the plurality of pillars, and the hole floors can be located outwardly of the main outer surface of the substrate.

[0014] In the step of forming the plurality of substantially parallel word lines, the word lines can be located outwardly of the main outer surface of the substrate, and each of the word lines can have a word line width defined by first and second word line edges. Further, each of the word lines can be separated from adjacent word lines by a word line separation distance. The word lines can be substantially perpendicular to the bit lines, and the word lines can be formed using a third mask. A given one of the first and second word line edges of a given one of the word lines and a given one of the first and second bit line edges of a given one of the bit lines can intersect substantially outward of the center of a given one of the active regions.

[0015] In the step of locating the word line plug in each of the hole regions, each of the word line plugs can extend from a corresponding adjacent one of the word lines inwardly into the hole region and can terminate outward of the hole floor. Each of the word line plugs can be electrically isolated from the semiconductor-material of the pillar, such that a portion of each of the word line plugs adjacent to and electrically isolated from the semiconductor material of the pillar forms a gate of the MOS transistor, with the corresponding portion of the semiconductor material of the pillar forming a channel region of the MOS transistor.

[0016] In the step of forming the plurality of storage nodes, one node can be formed for each of the active regions, and each of the storage nodes can be offset from the corresponding one of the active regions in a direction away from the given one of the word lines and the given one of the bit lines which intersect substantially outwardly of the center of the given one of the active regions. The storage nodes can be formed using a fourth mask.

[0017] Finally, in the step of interconnecting the storage nodes with the hole floors of the pillars via the plurality of storage node plugs, the interconnection can be conducted such as to form a plurality of storage node contacts and the other of sources and drains of the plurality of MOS transistors at the intersections of the storage node plugs with the hole floors. The storage node plugs can extend through the hole regions in the pillars and can be electrically isolated from the hole walls and the word line plugs.

[0018] Technical advantages of the present invention include a formation process which can be carried out with existing process technologies, and which does not depend on deliberate poor step coverage of the thin films. Further, the formation process is enhanced since the symmetrical structure of the prior art cell, with concomitant excessively thin layers, is done away with. However, significant technical advantages of the prior-art 08/934,157 application are retained, including minimization of transmission of substrate noise to the bit line and a reduced number of masks (four) required for fabrication.

[0019] Other technical advantages of the present invention will be readily apparent to one skilled in the art from the following figures, descriptions, and claims.

[0020] For a more complete understanding of the present invention and the advantages thereof, reference is now to be made to the following description taken in conjunction with the accompanying drawings in which like reference numbers indicate like features and wherein:

[0021] FIG. 1 is a plan view of a first embodiment of a DRAM structure according to the present invention;

[0022] FIG. 2 is a plan view of a second embodiment of a DRAM structure according to the present invention;

[0023] FIG. 3 is an enlarged plan view of a single cell of a DRAM structure according to the present invention;

[0024] FIG. 4A is a cross-sectional elevation taken along line A-A of FIG. 3;

[0025] FIG. 4B is a cross-sectional elevation taken along line B-B of FIG. 3;

[0026] FIG. 4C is a cross-sectional elevation taken along line C-C of FIG. 3;

[0027] FIG. 5 is an exploded perspective view of a DRAM cell in accordance with the present invention;

[0028] FIGS. 6A, 6B and 6C through 56A, 56B and 56C depict representative manufacturing steps viewed along lines A-A, B-B, and C-C in FIG. 3 respectively, for a representative DRAM cell manufactured in accordance with the present invention; and

[0029] FIG. 57 is a schematic, cross-sectional plan view through a memory cell of the present invention.

[0030] The dynamic random access memory (DRAM) device of the present invention will now be described with reference to FIGS. 1 and 3-5. The DRAM device, designated generally as 100, includes a substrate 102 of semiconductor material, such as silicon, having a main outer surface 104. Also included are a plurality of substantially parallel word lines 106 which are located outwardly of the main outer surface 104 of the substrate 102. Each of the word lines 106 has a word line width, W_WL, defined by first and second word line edges 108, 110. Each word line is separated from adjacent word lines by a word line separation distance D_WS.

[0031] The DRAM device also includes a plurality of substantially parallel bit lines 112 which are located outwardly of the main outer surface 104 of the substrate 102. Each of the bit lines 112 has a bit line width, W_BL, which is defined by first and second bit line edges 114, 116. Each of the bit lines 112 is separated from adjacent bit lines by a bit line separation distance D_BS. The bit lines 112 are substantially perpendicular to the word lines 106.

[0032] The DRAM device further comprises a plurality of memory cells 118, outlined with a dashed line in FIG. 1, each of which is associated with an intersection of a given one of the first and second word line edges 108, 110 of a given one of the word lines 106, and with a given one of the first and second bit line edges 114, 116 of a given one of the bit lines 112. For example, the memory cell 118 depicted FIG. 1 is associated with the intersection of the first bit line edge 114 of the indicated bit line 112 and the first word line edge 108 of the indicated word line 106. It will be appreciated that FIGS. 4A-4C depict cross sections in the region of a given one of the memory cells 118, and that FIG. 5 depicts an exploded view in the general region of a single one of the memory cells 118.

[0033] Each of the memory cells 118 in turn comprises a pillar 120 of semiconductor material which extends outward from the main outer surface 104 of the substrate 102, and which is electrically isolated from adjacent memory cells. The electrical isolation can be achieved, for example, by trench isolation regions 122 which can be made, for example, of SiO₂ or similar materials. The pillar 120 has a pillar outer surface 124 and includes an active region of semiconductor material which is substantially centered inward of the intersection of the given one of the word line edges with the given one of the bit line edges. With reference to FIG. 1, the active regions are designated by outlines 126. The centering with respect to the intersections of the word and bit lines is best seen in FIG. 1. FIGS. 4A through 4C show that the semiconductor pillars 120, which form the active regions, are located inwardly from the word lines 106 and bit lines 112.

[0034] As best seen FIGS. 4A-4C, the pillars 120 can be formed with a hole region defined by a hole floor 128 and at least one hole wall 130. The hole floor 128 is located outwardly of the main outer surface 104 of the substrate 102.

[0035] Each cell of the DRAM device includes a storage node 132 which is offset from the active region 126 of the semiconductor pillar 120 in a direction away from the given one of the word lines 106 and the given one of the bit lines 112. Also included is a storage node plug 134 extending from the storage node 132 through the hole region defined by the hole wall 130 in the pillar 120 so as to contact the hole floor 128. A storage node contact and drain 136 (which could instead be a source, as discussed below) is formed where the storage node plug 132 contacts the hole floor 128. The storage node contact and drain 136 forms the drain of the MOS transistor for each of the given cells. As best seen in FIGS. 4A and 4C, the storage node plug is electrically isolated from the at least one hole wall 130, for example, by a layer of SiN 138 and by a layer of SiO₂ 140.

[0036] Each of the memory cells 118 also includes a bit line plug 142 which extends from the given one of the bit lines 112 inwardly to contact the outer surface 124 of the pillar 120. Where the bit line plug 142 contacts the outer surface 124, a bit line contact and source 144 is formed for the MOS transistor. Source 144 could instead be a drain. Throughout this application, the drain and source can be interchanged for both the device and the manufacturing method.

[0037] Each memory cell 118 also includes a word line plug 146 which extends from the given one of the word lines 106 through the hole region formed by the hole wall 130 in the pillar 120 and which terminates outwardly of the hole floor 128. The word line plug 146 is electrically isolated from the semiconductor material of the pillar 120 such that a portion of the word line plug 146 which is adjacent to and electrically isolated from the semiconductor material of the pillar 120 forms a gate 148 of the MOS transistor. A corresponding portion of the semiconductor material of the pillar 120 forms a channel region 150 of the MOS transistor.

[0038] As noted above, the pillars of semiconductor material 120 can be isolated from adjacent pillars by trench isolation regions 122, which can be formed, for example, of SiO₂. As also noted, the substrate 102 can be made from silicon. Further, the gate 148 can be isolated from the channel region 150 by a region of gate silicon dioxide 152. The word line width W_WL, the word line separation distance D_WS, the bit line width W_BL, and the bit line separation distance D_BS can all be made substantially equal to a characteristic minimum feature size F. In this case, the planform area of each of the memory cells 118 is substantially 4F².

[0039] Referring particularly to FIG. 1, it will be seen that the storage nodes 132 can be offset from the active regions 126 all substantially in an identical direction for all of the memory cells 118, if desired.

[0040] Reference should now be had to FIG. 2, which is of an alternative embodiment of the present invention 100. Elements in FIG. 2 which are the same as those in FIG. 1 have received the same reference numeral. The difference between the FIG. 2 embodiment and the FIG. 1 embodiment is as follows. In FIG. 1, the direction in which the storage node 132 is offset from the active region 126 is substantially identical for all memory cells. In FIG. 2, the direction is substantially identical for all memory cells associated with a given one of the bit lines 112, however, the direction is opposite, in a direction substantially parallel to the bit lines 112, for all the memory cells 118 which are associated with an adjacent one of the bit lines 112.

[0041] As noted above, one desirable feature of the present invention is the asymmetrical nature of the memory cells 118, as compared to the prior art, which permits the various components to be formed as relatively thick, unitary structures rather than thin-walled symmetrical regions. Reference should now be had to FIG. 57. FIG. 57 is a cross-sectional plan view, in semi-schematic form, depicting the storage-node plug 134, bit line plug 142, and word line plug 146 of a given memory cell 118. The elements shown in FIG. 3, including the section lines A-A, B-B and C-C, have been shown in FIG. 57 in outline form, for reference. It will be appreciated that each of the storage node plug 134, bit line plug 142, and word line plug 146 has an outline when viewed in the plan view. At least two of the plugs 134, 142, 146 are not enclosed by the outlines of any of the other plugs 134, 142, 146. In the exemplary form shown, all three of the plugs 134, 142, 146 are not enclosed by any of the outlines of any of the other plugs. As shown in FIG. 57, the plugs 134, 142, 146 can all be substantially solid when viewed in plan. Further, they can form the above-mentioned substantially asymmetric structure. With reference to FIG. 57, as well as FIGS. 4-A through 4-C and FIG. 5, it will be appreciated that there can be overlap between the different plugs at different levels, as best seen in FIG. 4A and FIG. 4C. However, none of the plugs need be enclosed by any of the other plugs, when viewed in plan at any given section, as shown in the drawings. Thus, the thickness of each plug can be selected for ease in manufacturing, and is not dependent on a double-thickness symmetrical structure, as with the prior art discussed above.

[0042] With reference now to FIG. 4A, each memory cell 118 can include, if desired, additional regions of SiN 154, for purposes which will be appreciated in the discussion of the manufacturing method which follows. Further, each of the bit lines 112 can have a SiO₂ portion 156 for purposes which will also be apparent with reference to the manufacturing method discussed below. Additional insulation, which can be, for example, SiO₂ can be applied as shown at 158, in order to enclose the bit lines 112 and to become unitary with the trench isolation regions 122. A layer of TiN 160 can be provided between the storage node 132 and the storage node plug 134, again, for reasons which will be appreciated when the manufacturing method is discussed below.

[0043] A method, according to the present invention, of manufacturing the inventive dynamic random access memory device will now be described. Continued reference should be had to FIGS. 1-5, and exemplary method steps for one possible manufacturing method will also be discussed in detail with respect to FIGS. 6A-6C through 56A-56C. The method comprises the steps of (a) providing a substrate of semiconductor material 102 having a main outer surface 104; (b) forming a plurality of pillars of semiconductor material 120; (c) forming a plurality of substantially parallel bit lines 112; (d) interconnecting the bit lines 112 with outer surfaces 124 of the pillars 120; (e) forming at least one hole region in each of the plurality of pillars 120; (f) forming a plurality of substantially parallel word lines 106; (g) locating a word line plug 146 in each of the hole regions; (h) forming a plurality of storage nodes 132; and (i) interconnecting the storage nodes 132 with hole floors 128 of the hole regions in the pillars 120.

[0044] In step (a), the substrate can have main outer surface 104. In step (b), the plurality of pillars 120 can extend outward from the main outer surface 104 of the substrate 102, and each of the pillars 120 can have a pillar outer surface 124. The pillars can be formed using a first mask, as discussed further below with respect to FIGS. 6A-6C through 56A-56C. The pillars 120 can be electrically isolated from each other and can include an active region 126 of the semiconductor material.

[0045] Step (c) can include forming the bit lines 112 such that they are located outwardly of the main outer surface 104 of the substrate 102. Each of the bit lines 112 can have a bit line width W_BL, as discussed above, which can be defined by the first and second bit line edges 114, 116. As discussed above for the apparatus, each of the bit lines can be separated from adjacent bit lines by a bit line separation distance D_BS. The bit lines 114 can be formed using a second mask, as will be discussed further below.

[0046] In step (d) the bit lines 112 can be interconnected with the pillar outer surfaces 124 via a plurality of bit line plugs 142 so as to form a plurality of bit line contacts and sources 144 of the plurality of MOS transistors. As before, sources 144 can instead be drains. The bit line contacts and sources 144 can be formed at the intersections of the bit line plugs 142 with the pillar outer surfaces 124, as discussed above. The formation of the bit line contact and source 144 can occur by auto diffusion from the plug 142.

[0047] In step (e), the at least one hole region which is formed can be defined by the hole floor 128 and the at least one hole wall 130 in each of the plurality of pillars 120. The hole floors 128 can be located outwardly of the main outer surface 104 of the substrate 102, as discussed above with respect to the apparatus.

[0048] In step (f), the plurality of substantially parallel word lines 106 can be located outwardly of the main outer surface 104 of the substrate 102, and each can have a word line width defined by first and second word line edges and separated from adjacent word lines by a word line separation distance, all as discussed above. The word lines 106 can be substantially perpendicular to the bit lines 112. The word lines 106 can be formed using a third mask, as will discussed below. As discussed above, and with reference again to FIG. 1, a given one of the first and second word line edges 108, 110 and a given one of the first and second bit line edges 114, 116 can intersect substantially outward of a center portion of a given one of the active regions 126. The outward placement of the word lines and bit lines from the active region is best seen in FIGS. 4A-4C.

[0049] In step (g), each of the word line plugs 146 can extend from a corresponding adjacent one of the word lines 106 inwardly into the hole region defined by the hole wall 130, and can terminate outwardly of the hole floor 128. Each of the word line plugs 146 can be electrically isolated from the semiconductor material of the pillar 120, as discussed above, such that a portion of each of the word line plugs 146 adjacent to and electrically isolated from the semiconductor material of the pillar 120 forms a gate 148 of the MOS transistor, and such that a corresponding portion of the semiconductor material of the pillar 120 forms a channel region 150 of the MOS transistor.

[0050] With reference to step (h), one storage node 132 can be formed for each of the active regions 126, and each of the storage nodes 132 can be offset from the corresponding one of the active regions 126 as described above. The storage nodes 132 can be formed using a fourth mask, as will be discussed below.

[0051] In step (i), the storage nodes 132 can be interconnected with the hole floors 128 of the pillars 120 via a plurality of storage node plugs 134. Where the storage node plugs 134 intersect the hole floors 128, a plurality of storage node contacts and drains 136 can be formed, for example, by auto diffusion from the plug 134. Again, the drains can instead be sources. The storage node plugs 134 can extend through the hole regions in the pillars and can be electrically isolated from the hole walls 130 and the word line plugs 146, as discussed above.

[0052] The method can further comprise the additional step of forming a plurality of trench isolators, or trench isolation regions, 122 between the pillars 120 to obtain the required electrical isolation. Further, the method can further comprise the additional step of forming a layer of gate SiO₂ 152 between each of the gates 148 and the corresponding portions of the semiconductor material of the pillars 120, as discussed above.

[0053] In the method, step (c) can comprise-forming the bit lines 112 such that the bit line width and separation distance are both substantially equal to the characteristic minimum feature size F, discussed above, while step (f) can comprise forming the word lines 106 such that the word line width and word line separation distance are also both substantially equal to the characteristic minimum feature size F, as discussed above. Thus, the planform area of each of the plurality of memory cells 118, one each of which is formed at each of the plurality of pillars 120, will be substantially 4F², as discussed above.

[0054] Step (h) can comprise forming the storage nodes 132 such that the direction in which each of the storage nodes 132 is offset from its corresponding one of the active regions 126 is substantially identical for all of the storage nodes 132. The resulting structure has been discussed above with reference to FIG. 1, embodiment 100. Alternatively, step (h) can comprise the sub-step of forming those of the storage nodes 132 which are associated with even-numbered bit lines 112 with the direction in which each of the storage nodes is offset from its corresponding one of the active regions 126 substantially identical. Further, step (h) can also comprise the sub-step of forming those of the storage nodes 132 which are associated with odd numbered bit lines 112 each offset from its corresponding one of the active regions 126 in a substantially identical direction, which is opposite to the offset direction for those of the storage nodes associated with the even numbered bit lines. The offset can be substantially parallel to the bit lines 112. In performing these sub-steps, the embodiment 100 illustrated in FIG. 2 above can be obtained. It will be appreciated that appropriate odd and even numbering can be selected for the bit lines 112.

[0055] Reference should again now be had to FIG. 57 and the above discussion with respect to FIG. 57. It will therefore be appreciated that step (d) can comprise forming the bit line plugs 142 such that they have a bit line plug outline when viewed in plan view, as in FIG. 57. Further, step (g) can comprise forming the word line plugs 146 such that they have a word ling plug outline when viewed in plan view. Yet further, step (i) can comprise forming the storage node plugs 134 such that they have a storage node plug outline when viewed in plan view. The plugs 132, 142, 146 can be shaped such that at least two of the plugs are not enclosed by any of the outlines of any of the other plugs. In the embodiment illustrated, all three of the plugs are not enclosed by any of the outlines of any of the other plugs. As discussed above, there can be some overlap of the plugs at different levels, as best seen in the cross-sectional views 4A through 4C. The shaping step can be performed such that all of the plugs 134, 142, 146 are solid when viewed in plan, as shown in FIG. 57 and discussed above. The plugs can also be shaped such that they form a substantially asymmetric structure when viewed in plan, as shown in FIG. 57 and discussed above.

[0056] Attention should now be given to FIGS. 6A-6C through 56A-56C, for a description of one specific manner in which the manufacturing method of the present invention can be carried out. It should be noted that the specific method shown in the figures is exemplary, and that the method can be implemented in other fashions as well. As shown in FIGS. 6A-6C, a substrate 102 can initially be provided; for example, a silicon substrate can be provided. Referring now to FIGS. 7A-7C, a layer of SiN, or similar insulating material with preferential etching qualities, is then deposited on top of the substrate 102. With reference next to FIGS. 8A-8C, a suitable layer of photoresist 202 can then be applied over the SiN layer 200.

[0057] With reference next to FIGS. 9A-9C, the resist layer 202 can be patterned in the shape of the aforementioned active regions 126, using a first mask. With reference to FIGS. 10A-10C and 11A-11C, the SiN layer 200 can be etched back to the substrate 102 so that it only remains beneath the photoresist 202, and the photoresist can then be removed.

[0058] With reference to FIGS. 12A-12C, the substrate 102 can be etched so as to form the pillars 120 of semiconductor material discussed above. As shown in FIGS. 13A-13C, the silicon in the substrate 102, including that in pillar 120, can then be oxidized to form an SiO₂ layer 204. With reference now to FIGS. 14A-14C through 16A-16C, additional SiO₂ can be deposited, and polished down to the level of the SiN layer 200, which can subsequently be removed. As shown in FIGS. 16A-16C, void regions 208 are formed where the SiN layer 200 is removed. As shown in FIGS. 17A-17C and 18A-18C, the voids 208 can subsequently be filled via deposition of polysilicon 210 (or other desirable material) which can then be polished down to the level of the additional SiO₂ 206.

[0059] With attention now to FIGS. 19A-19C through 21A-21C, a layer of tungsten 212, or other desirable contact material, can be sputtered or otherwise applied over the additional silicon dioxide 206 and polysilicon 210. An additional layer of SiO₂ 214 can then be deposited (or another insulator with suitable preferential etching characteristics). Finally, an additional resist layer 216 can be applied.

[0060] With reference now to FIGS. 22A-22C, the additional resist 216 can be patterned using a second mask in order to form the above-discussed bit lines 112. The additional silicon dioxide 214 which is not protected by the additional resist 216 can then be etched away, as shown in FIGS. 23A-23C. Subsequently, the tungsten 212 which is not protected by the additional resist 216 can also be etched away, as shown in FIGS. 24A-24C. As shown in FIGS. 25A-25C and 26A-26C, those regions of the polysilicon 210 which are not protected by the additional resist layer 216 can be etched down to a thin film 218 and the additional resist 216 can then be removed.

[0061] Attention should now be given to FIGS. 27A-27C and 28A-28C. As shown therein, a thin film of SiN 220 can be deposited over the upper surfaces of the structure remaining after the steps depicted up to and including FIGS. 26A-26C have been performed. It can then be etched away, on the horizontal surfaces only, to leave the vertical thin films of SiN 220 depicted in FIGS. 28A-28C. Again, an alternative insulating material with suitable preferential etching qualities could be used instead of the SiN.

[0062] As shown in FIGS. 29A-29C, a portion of the silicon or other semiconductor material in the pillar 120 can be etched away to form the aforementioned hole region. The hole region is designated as 222 in the drawings depicting the exemplary manufacturing process. It will be appreciated that a corresponding portion of the SiN thin film 220 is also etched through as the hole region 222 is formed. With attention to FIGS. 30A-30C, the portion of the silicon or other semiconductor in the pillar 120 which is exposed as the hole region 222 is formed can be oxidized to form an SiO₂ layer 224. As shown in FIGS. 31A-31C and 32A-32C, additional polysilicon 226 can be deposited and subsequently etched back.

[0063] With reference now to FIGS. 33A-33C through 35A-35C, an additional tungsten layer 228 can be sputtered onto the upper surfaces of the structure resulting at the end of the steps depicted up to and including FIGS. 32A-32C. As an alternative to sputtering of the tungsten layer 228, TiN can be sputtered and the tungsten 228 can be deposited. Further, it is to be understood that other desirable conductors could be used instead of tungsten. A further SiO₂ layer 230 can be deposited over the tungsten layer 228, and a third resist coat 232 can subsequently be applied. With reference to FIGS. 36A-36C, the third resist coat 232 can be patterned using a third mask so as to form the aforementioned word lines 106.

[0064] Those portions of the SiO₂ layer 230 and additional tungsten layer 228 which are not protected by the third resist coat 232 can be etched away, as shown in FIGS. 37A-37C and 38A-38C. It will be appreciated that some incidental etching of the additional polysilicon 226 may also occur during the etching of the additional tungsten 228, as shown. With reference to FIGS. 39A-39C and 40A-40C, the additional polysilicon 226 which is not protected by the third resist coat 232 can be etched away, and the third resist coat 232 can then be removed. As shown in FIGS. 41A-41C and 42A-42C, an additional SiN thin film 234 can be deposited and the horizontal surfaces thereof can subsequently be etched, to result in a number of substantially vertical portions of the SiN thin film 234. It will be appreciated that the additional SiN thin film 234, in some locations, is in intimate contact with the SiN thin film 220 previously discussed. Accordingly, in these locations where the two films are immediately adjacent to each other, the SiN thin film 220 will not be separately labeled after FIGS. 41A-41C, for convenience.

[0065] With reference to FIGS. 43A-43C, the silicon dioxide layer 224 can be etched in those locations where it is not covered by the additional polysilicon 226, or the additional SiN 234. Deposition of a third polysilicon portion 236 can then be carried out as shown in FIGS. 44A-44C. It will be appreciated that the third polysilicon layer can extend down to the hole floor 128 discussed above due to the previous etching of the SiO₂ layer 224. With reference to FIGS. 45A-45C and 46A-46C, a thin film of TiN 238, followed by a layer of platinum 240 can be deposited on the top of the structure present after the execution of the steps depicted through FIGS. 44A-44C. This can be followed by deposition of a layer of SiO₂ 242 and a fourth resist coat 244, as shown in FIGS. 47A-47C and 48A-48C.

[0066] As shown in FIGS. 49A-49C through 51A-51C, the fourth resist coat 244 can be patterned using a fourth mask to form the above-mentioned storage nodes 132. This can be followed by etching of the silicon dioxide layer 242 where it is not protected by the fourth resist coat 244, and then by removal of the resist coat 244.

[0067] As shown in FIGS. 52A-52C, the platinum layer 240 can be etched away in those regions where it is unprotected by the silicon dioxide layer 242, and the silicon dioxide layer 242 can then be removed. Further, the third polysilicon layer 236 can be etched away in those regions where it is unprotected by the remaining portions of the platinum layer 240. It will be appreciated that the remaining portions of the platinum layer 240 form the storage nodes 132 while the remaining portions of the third polysilicon layer 236 form the storage node plugs 134. Elements 132, 134 have been discussed above. Etching of the third polysilicon layer 236 is shown in FIGS. 53A -53C. Finally, attention should be given to FIGS. 54A-54C through 56A-56C which depict deposition of a final layer of silicon dioxide 246 which is subsequently polished and etched down to a level below the tops of the storage nodes formed by the remaining portions of the platinum layer 240. Where final silicon dioxide layer 246 touches silicon dioxide layer 214, layer 214 is no longer separately labeled. The same is true where final silicon dioxide layer 246 touches silicon dioxide layer 230.

[0068] The pillars of semiconductor material can be, for example, about 0.21 m on a side. Other dimensions can be determined with reference to the several views of the drawings, which are proportional.

[0069] It is instructive to compare FIGS. 4A-4C with FIGS. 56A-56C. As discussed above, remaining portions of platinum layer 240 correspond to the storage nodes 132. Further, the remaining portions of the TiN thin film 238 correspond to the TiN layer 160. Remaining portions of the third polysilicon layer 236 correspond to the, storage node plug 134. The remaining portions of the polysilicon layer 210 correspond to the bit line plugs 142. The remaining portions of the polysilicon layer 226 correspond to the word line plug 146. The remaining SiN layers 220, 234 correspond respectively to the layers 138, 154. Remaining portions of the additional tungsten layer 228 correspond to the word lines 106, while the remaining portions of the tungsten layer 212 correspond to the bit lines 112. Silicon layer 206 corresponds to the trench isolation regions 122 and the final silicon layer 246 forms the additional insulation 158.

[0070] With reference again to FIGS. 4A-4C, the writing charging path is indicated by the heavy dashed line 300 in FIG. 4C. Charge from the bit line 112 goes through the bit line contact plug 142 and then to the bit line contact and source region 144. By turning on the transfer gate 148, an inversion layer is formed between the source region 144 and the drain region 136 (seen in FIG. 4C) at the bottom of the storage node plug 134 and the drain region 136. As shown by the dashed line 300, charge goes up through the storage node plug 134 and then into the storage node 132. The cell capacitance can be formed between the storage node 132 and a plate electrode separated by a suitable dielectric. Neither the plate electrode nor the dielectric are shown in the figures, as they can be formed in any desired fashion.

[0071] Although the present invention has been described in detail, it should be understood that various changes, substitutions and alterations can be made hereto without departing from the spirit and scope of the invention as defined by the appended claims.