US 20080128759 A1
An array of continuous diffusion regions and continuous gate electrode structures is formed over a semiconductor substrate. Interconnecting diffusion region portions and interconnecting gate electrode portions are removed to electrically isolate transistor circuitry. The removal of interconnecting diffusion region portions and gate electrode portions can be performed sequentially, at substantially the same time, and before or after forming source/drain contacts.
1. A semiconductor device comprising:
a first diffusion region and a second diffusion region;
a first gate electrode and a second gate electrode; and
a first opening between the first gate electrode and the second gate electrode, wherein the first opening electrically separates the first gate electrode from the second gate electrode.
2. The semiconductor device of
3. The semiconductor device of
4. The semiconductor device of
5. The semiconductor device of
6. The semiconductor device of
7. A semiconductor device comprising:
a first structure and a second structure, wherein the first structure is adjacent the second structure;
an opening between the first structure and the second structure, wherein the opening electrically separates the first structure from the second structure.
8. The semiconductor device of
This application is a divisional application of prior application Ser. No. 10/956,320 which was filed on Sep. 30, 2004 and for which a Notice of Allowance was mailed on Oct. 4, 2007.
Embodiments of the present invention relate generally to semiconductor processing and more specifically to methods for isolating semiconductor structures.
In conventional semiconductor processing, discrete diffusion regions are typically formed in a semiconductor substrate and then discrete gate electrodes structures are aligned and patterned relative to them. For example, referring to
Proper x-axis and y-axis alignment of gate electrodes relative to the diffusion regions is an important consideration with respect to transistor operation and yield. For example, an inability to align the gate electrode 208B (in
One possible solution for reducing the occurrence of gate electrode/diffusion region misalignment includes increasing the length of the gate electrode 208B to insure that more of poly end cap 224 is available to overlap the diffusion region 112. However, this solution may be undesirable because it imposes a limit on the ultimate scalability of the device. In other words, to the extent that gate electrodes lengths are increased, spaces 220 between adjacent gate electrodes must also be increased to reduce the likelihood of encountering problems wherein the gate electrode 208A overlaps onto adjacent transistor regions. As semiconductor scaling continues, conventional patterning limitations such as these have the potential to limit the overall reliability, yield, and scalability semiconductor devices.
It will be appreciated that for simplicity and clarity of illustration, elements illustrated in the drawings have not necessarily been drawn to scale. For example, the dimensions of some of the elements are exaggerated relative to other elements for clarity. Further, where considered appropriate, reference numerals have been repeated among the drawings to indicate corresponding or analogous elements.
In the following detailed description, a method for electrically isolating adjacent transistors is disclosed. Reference is made to the accompanying drawings within which are shown, by way of illustration, specific embodiments by which the present invention may be practiced. It is to be understood that other embodiments may exist and that other structural changes may be made without departing from the scope and spirit of the present invention.
In one embodiment, semiconductor devices are fabricated by electrically isolating transistors from each other after the gate electrode patterning operation. Unlike conventional processing, which electrically isolates diffusion regions and gate electrodes at each respective patterning operation, one embodiment of the present invention forms a grid of continuous parallel gate electrode structures overlying a grid of continuous parallel diffusion regions. The grid of continuous diffusion regions and the grid of continuous gate electrode structures are perpendicular to each other. The combination of the two grids forms an overlapping array. With the exception of the masks that define the continuous diffusion regions and gate electrode structures, the etch, implant, and photo processes used to form the continuous diffusion regions and continuous gate electrode structures are conventional to one of ordinary skill.
After the continuous gate electrode structures have been formed, source/drain regions and gate electrodes have been partially formed. But, respective transistors associated with the source/drain regions and gate electrodes could not be functional since the continuous diffusion regions and continuous gate electrode structures are still interconnected. Specific circuit interconnectivity (i.e., transistor isolation) is accomplished by selectively removing specific interconnecting diffusion regions and specific interconnecting gate electrode regions. Interconnecting regions can be removed using conventional patterning and etch processing, laser ablation, focused ion beam processes, or the like. The removed material can be replaced with a dielectric material. Source/drain and gate electrode contacts can be formed either before or after interconnecting regions are removed.
Selective removal of the interconnecting diffusion regions can be done independent of the selective removal of the interconnecting gate electrode regions. Or, both the selective removal of the interconnecting diffusion region portions and the selective removal of the interconnecting gate electrode portions can be done at substantially the same time.
In alternative embodiments, instead of using both continuous gate electrode structures and continuous diffusion regions, one layer can be a continuous region or structure and the other can be a discrete region of structure. These embodiments and variations thereof may be better understood with respect to
Patterned diffusion regions that include regions 306, 308, 310, and 312 have been formed on or within the semiconductor substrate 302. Here, the patterned diffusion regions are separated laterally from each other via isolation regions 314. The isolation regions 314 can be formed using any number of isolation schemes, such as shallow trench isolation, local oxidation of silicon (LOCOS), or the like. The presence of isolation region 314 is not a requirement of embodiments of the present invention. However, the isolation regions 314 may improve planarity of the substrate surface and may be advantageous. In embodiments where the substrate is a silicon-on-insulator (SOI) substrate (as shown here), the isolation areas 314 can be formed overlying a buried oxide layer 304 and between the diffusion regions 306, 308, 310, and 312. In alternative SOI embodiments, where isolation 314 is not present, the substrate surface would not be planar. In this case, subsequently patterned gate electrodes would wrap around the top and sidewalls of the diffusion regions, thereby forming tri-gates or similar structures.
After the patterned openings 504, 506, 508, 510, and 512 have been formed in resist layer 502, the exposed portions of the diffusion regions 531-533 and the exposed portions of the gate electrodes 521-526 can be removed. Removal can be accomplished using conventional etch processes. Removal of exposed diffusion region portions 531-533 and exposed gate electrode portions 521-526 forms breaks in the continuous diffusion regions and continuous gate electrode structures, thereby converting them from continuous structures to discontinuous structures. The breaks electrically isolate adjacent diffusion regions and gate electrodes from each other and thereby define the interconnectivity between transistors.
More specifically, the removal of the exposed gate electrode portion 521 electrically isolates adjacent gate electrodes 5211 and 5212. The removal of the exposed gate electrode portion 522 electrically isolates adjacent gate electrodes 5221 and 5222. The removal of the exposed gate electrode portion 523 electrically isolates adjacent gate electrodes 5231 and 5232. The removal of the exposed gate electrode portion 524 electrically isolates adjacent gate electrodes 5241 and 5242. The removal of the exposed gate electrode portion 525 electrically isolates adjacent gate electrodes 5251 and 5252. The removal of the exposed gate electrode portion 526 electrically isolates adjacent gate electrodes 5261 and 5262.
Similarly, the removal of the exposed diffusion region 531 electrically isolates adjacent diffusion regions 5311 and 5312. The removal of the exposed diffusion region 532 electrically isolates adjacent diffusion regions 5321 and 5322. And, the removal of the exposed diffusion region 533 electrically isolates adjacent diffusion regions 5331 from a diffusion region which is beyond the confines of
In embodiments where the patterned openings 604, 606, 608, 610, and 612 are formed prior to forming source/drain and gate electrode contacts (not shown), the fill material can have a composition that is dissimilar to, or that has an etch selectivity that is different from, the dielectric 602, which will be etched again when the source/drain and gate electrode contact openings are formed. This may improve device reliability and yield by preventing or minimizing the loss of fill material 702 in patterned openings 604, 606, 608, 610, and 612 during the subsequent contact etch. Exposure of fill material during contact etch could occur if the contact openings were to be misaligned or otherwise overlap with one of the patterned openings 604, 606, 608, 610, and 612. However, in this embodiment, to the extent that there is overlap of the contact opening and the patterned opening, the contact etch process will not substantially remove fill material. Moreover, to the extent that the fill material and the ILD layer can be etched selectively with respect to each other, the fill material may additionally function as a self-aligning structure for the etch that forms the contact opening. For example, if the ILD layer is a layer of silicon dioxide, then silicon nitride, which can be etched selectively with respect to silicon dioxide, can be used to form the fill material.
From this point on, processing to form the semiconductor device is considered conventional to one of ordinary skill. Contacts, ILD layers, interconnects, vias, or the like, and any other structure necessary for fabrication of a semiconductor device can then all be fabricated over the semiconductor substrate to produce semiconductor circuitry.
Embodiments of the present invention are particularly useful in forming arrays of memory cells where the transistor layout is repeated over a relatively large area, such as for example, in arrays that include SRAM cells. In such arrays, scaling is critical to fabricating advanced integrated circuits (ICs). Embodiments of the present invention have advantages over conventional methods for forming transistors because they facilitate reducing layout dimensions for memory array circuitry.
For example, using embodiments of the present invention, alignment tolerances can be widened with respect to aligning the gate electrode to the diffusion regions. And, greater alignment tolerance facilitates scaling of device geometries, layout dimension, and reduced processing costs. Instead of aligning discrete gate electrode regions to discrete diffusion regions, which requires critical alignments in both the x-axis and y-axis directions, the primary consideration now becomes merely the orthogonal position of the continuous gate electrode to the continuous diffusion regions. The alignment and the critical dimensions of the patterned openings to electrically isolate the continuous sections of gate electrodes and diffusion regions is not as critical as the alignment of the discrete gate electrode portions to the discrete diffusion regions.
In addition, the array of regularly spaced gate electrode and diffusion regions provides a consistent pattern of repeating structures over the substrate. The repeating structures can reduce variations in the topography of the substrate surface. Also, chemical mechanical polishing (CMP) CMP process latitude is improved because now there is less surface topography variation and also because processing variations due to CMP macroloading and microloading effects are reduced.
In addition, embodiments of the present invention can facilitate using tighter design rules for smaller layout areas. Forming the isolation cuts (patterned openings) after forming the contact openings allows designers to use tighter design rule for a semiconductor device layout. This can be better understood with respect to comparing
As described previously with respect to
One of ordinary skill appreciates that many alternative embodiments to the present invention exist, For example, instead of using both continuous gate electrode structures and continuous diffusion regions, one layer can be a continuous region or structure and the other can be a discrete region of structure. For example, where a grid of continuous gate electrode structures is formed over discrete patterned diffusion regions, specific interconnectivity can be obtained by removing only interconnecting gate electrode structure portions. Or, where discrete gate electrode structures are formed over continuous diffusion regions, specific interconnectivity can be obtained by removing specific interconnecting diffusion region portions. This could be advantageous where lithography or etch process limitations, such as tool costs or etch rate, exist. In addition, while much of the discussion herein focuses on embodiments surrounding the isolation of transistors via the removal of interconnecting diffusion regions, gate electrode portions, or both, one of ordinary skill will appreciate that the post patterning isolation methods disclosed herein can be incorporated into other front end and back end semiconductor processing steps as well. For example, instead of forming discrete metal segments using a traditional lithography process, a continuous metal line (for example a metal line used to form interconnects) can be deposited over a semiconductor device. And then, isolation cuts can be applied to specific regions of the metal line after it is formed to obtain the functional interconnectivity required for a circuit. In this way, the long metal line can be arranged to provide a density of features that can be used to improve CMP process uniformity by reducing overall topography variations, CMP microloading, and CMP macroloading effects.
While the present invention has been described with reference to various illustrative embodiments thereof, the present invention is not intended that the invention be limited to these specific embodiments. Those skilled in the art will recognize that variations and modifications can be made without departing from the spirit and scope of the invention as set forth in the appended claims. The various implementations described above have been presented by way of example only, and not limitation. Having thus described in detail embodiments of the present invention, it is understood that the invention defined by the appended claims is not to be limited by particular details set forth in the above description, as many apparent variations thereof are possible without departing from the spirit or scope thereof.