Efficient fabrication process for dual well type structures

[0002] 1. Field of the Invention

[0003] The present invention relates generally to the field of integrated circuit fabrication, and more specifically to a fabrication process for use when two well types are required on a single integrated circuit.

[0004] 2. Description of the Related Art

[0005] It is often necessary or desirable to fabricate both n-type and p-type devices on a single substrate. In order to fabricate both types of devices in a single substrate, typically a large area of a substrate of one type is doped to the other type. This large area is referred to as a “well.” For example, FIG. 1 illustrates a portion 100 of an integrated circuit on a p-type substrate 110. The substrate 110 has an n-well (a well formed of n-type material) 120 formed therein. Inside the n-cell 120 are a p-type source 141 and drain 142 on either side of a gate stack 150, which is formed of a gate dielectric 151, a gate electrode 158 and an insulating cap 159. Together, the gate stack 150 and the source and drain 141, 142 form a p-type field effect transistor 160. Field oxide regions 130 isolate the transistor 160 from other devices on the substrate 110.

[0006] Conventional wells are formed by implanting dopants at the well locations and then diffusing them (usually through a thermal process) to the desired depth. Such wells are also referred to as diffusion wells for this reason. One drawback associated with diffusion wells is that the diffusion occurs laterally as well as vertically, e.g. the diffusion well gets wider as it gets deeper. A second drawback associated with diffusion wells is that relatively large spaces between the edges of the wells and device active areas is required.

[0007] A second type of well, referred to as a retrograde well, attempts to overcome the lateral spreading problem by implanting high energy dopants to the desired depth so that thermal diffusion is not necessary. Retrograde wells require less space between the edges of the well and device active areas than required by diffusion wells. Retrograde wells are therefore desirable for high density applications.

[0008] In certain applications, it makes sense to use both well types on a single integrated circuit. For example, in flash memory applications, which contain a low voltage peripheral circuit portion and a high-voltage peripheral circuit portion, it is desirable to have the low voltage peripheral circuit portion implemented in retrograde wells and the high voltage peripheral circuit portion implemented in conventional diffusion wells. This type of structure is discussed in Watanabe et al., “Novel 0.44 μm²Ti-Salicide STI Technology for High Density NOR Flash Memories and High Performance Embedded Application,” IEDM 98-975 (IEEE 1998). Another example of a device using both diffusion and retrograde wells on a single integrated circuit is described in U.S. Pat. No. 5,428,239, to Okamura et al.

[0009] Although these references disclose use of dual well type structures, they do not disclose an efficient method for fabrication such structures. Thus, integrated circuit manufacturers are faced with choosing between the benefits of using a dual well type structure one hand and the less complicated fabrication processes required for single well type structures on the other hand.

[0010] What is needed is an efficient method for fabricating a dual well type structure.

[0011] The invention overcomes to a great extent the aforementioned problems by providing an efficient method for fabricating structures with two types of wells which uses the same number of masks as used in single well type structure fabrication. In a preferred embodiment, the current invention allows low voltage and high voltage n-channel transistors and low voltage and high voltage p-channel transistors to be formed in a single substrate. One mask is used for forming a diffusion well, a second mask for both forming a retrograde well and doping the retrograde well to achieve an intermediate threshold voltage in that well, and a third mask for both differentiating the gate oxides for the low voltage devices and doping the areas of the substrate corresponding to the low voltage devices to achieve the final threshold voltages.

[0012] The foregoing and other advantages and features of the invention will become more apparent from the detailed description of the preferred embodiments of the invention given below with reference to the accompanying drawing in which:

[0013] FIG. 1 is a cross-sectional diagram of a portion of an integrated circuit having a field effect transistor formed in a conventional diffusion well;

[0014] FIG. 2 is a cross-sectional diagram of a portion of a prior art integrated circuit having high and low voltage diffusion wells at an early stage of fabrication;

[0015] FIG. 3 is a cross-sectional diagram of the integrated circuit portion of FIG. at a later stage of fabrication;

[0016] FIG. 4 is a cross-sectional diagram of the integrated circuit portion of FIG. 3 at a later stage of fabrication;

[0017] FIG. 5 a cross-sectional diagram of the integrated circuit portion of FIG. 4 at a later stage of fabrication;

[0018] FIG. 6 is a cross-sectional diagram of the integrated circuit portion of FIG. 5 at a later stage of fabrication;

[0019] FIG. 7 is a cross-sectional diagram of the integrated circuit portion of FIG. 6 at a later stage of fabrication;

[0020] FIG. 8 is a cross-sectional diagram of the integrated circuit portion of FIG. 7 at a later stage of fabrication;

[0021] FIG. 9 is a cross sectional diagram of a portion of an integrated circuit having a high voltage diffusion well and a low voltage retrograde well at an early stage of fabrication in a fabrication process according to one embodiment of the present invention;

[0022] FIG. 10 is a cross-sectional diagram ofthe integrated circuit portion of FIG. 9 at a later stage of fabrication;

[0023] FIG. 11 is a cross-sectional diagram of the integrated circuit portion of FIG. 10 at a later stage of fabrication;

[0024] FIG. 12 a cross-sectional diagram of the integrated circuit portion of FIG. 11 at a later stage of fabrication;

[0025] FIG. 13 is a cross-sectional diagram of the integrated circuit portion of FIG. 12 at a later stage of fabrication;

[0026] FIG. 14 is a cross-sectional diagram of the integrated circuit portion of FIG. 13 at a later stage of fabrication;

[0027] FIG. 15 is a cross-sectional diagram of the integrated circuit portion of FIG. 14 at a later stage of fabrication;

[0028] FIG. 16 is a cross-sectional diagram of the integrated circuit portion of FIG. 15 at a later stage of fabrication;

[0029] FIG. 17 is a cross-sectional diagram of the integrated circuit portion of FIG. 16 at a later stage of fabrication;

[0030] FIG. 18 is a cross sectional diagram of a portion of an integrated circuit at an early stage of fabrication according to another embodiment of the present invention;

[0031] FIG. 19 is a cross sectional diagram of the integrated circuit portion of FIG. 18 at a later stage of fabrication;

[0032] FIG. 20 is a cross sectional diagram of the integrated circuit portion of FIG. 19 at a later stage of fabrication;

[0033] FIG. 21 is a cross sectional diagram of the integrated circuit portion of FIG. 20 at a later stage of fabrication;

[0034] FIG. 22 is a cross sectional diagram of the integrated circuit portion of FIG. 21 at a later stage of fabrication;

[0035] FIG. 23 is a cross sectional diagram of the integrated circuit portion of FIG. 22 at a later stage of fabrication;

[0036] FIG. 24 is a cross sectional diagram of the integrated circuit portion of FIG. 23 at a later stage of fabrication;

[0037] FIG. 25 is a cross sectional diagram of the integrated circuit portion of FIG. 24 at a later stage of fabrication;

[0038] FIG. 26 is a cross sectional diagram of the integrated circuit portion of FIG. 25 at a later stage of fabrication;

[0039] FIG. 27 is a cross sectional diagram of the integrated circuit portion of FIG. 26 at a later stage of fabrication; and

[0040] FIG. 28 is a circuit diagram of a six transistor SRAM cell.

[0041] The present invention will be illustrated through a description of the fabrication of a portion of an integrated circuit containing two types of wells, diffusion and retrograde, for high voltage and low voltage devices, respectively. Numerous specific details, such as materials, thicknesses, etc., are set forth in order to provide a thorough understanding of the present invention. It will be obvious, however, to one skilled in the art that the present invention is capable of many different embodiments and that the present invention may be practiced without the specific details set forth herein. Accordingly, the drawing and description herein are to be regarded as illustrative in nature and not as restrictive.

[0042] A prior art process for fabricating both low and high voltage field effect transistors in diffusion wells will be illustrated first. It should be noted that the drawings herein are not to scale. This is especially true of the thickness of the gate oxide regions. It should also be appreciated that the following drawings illustrate wells with only a single transistor therein solely for the purpose of illustrating fabrication processes. Those of ordinary skill in the art will appreciate that, in practice, two or more devices, and often many devices, are placed in a single well. For example, in SRAM memory circuit applications, a common configuration for an SRAM memory cell 500 consists of 4 n-channel transistors and 2 p-channel transistors, as shown in FIG. 28, formed on a p-substrate. Typically, a single n-well is formed for the two p-channel devices. However, in the peripheral portion of the circuit, it is common to have many p-channel devices in a single well.

[0043] FIG. 2 illustrates a portion 200 of an integrated circuit on which four devices will be fabricated at the indicated places: a high voltage p-channel transistor, a low voltage p-channel transistor, a high voltage n-channel transistor and a low voltage n-channel transistor. A photoresist mask 270 with two diffusion well openings 271 is formed on substrate 210. The well openings 271 are of equal size. Low energy (less than approximately 200 keV) ion implantation occurs in the well openings 271. The entire integrated circuit is then heated to cause the implants to diffuse both vertically and horizontally. The resultant diffusion n-wells 220, 221 are shown in FIG. 3 which also shows removal of the mask 270. It should be noted that the n-wells 220, 221 are wider than were the well openings 271 in the mask 270 of FIG. 2. This is because, as discussed above, diffusion occurs horizontally as well as vertically.

[0044] Next, as shown in FIG. 4, field oxide areas 230 are formed by a process such as STI (shallow trench isolation). A thin oxide layer 251 is then grown over the entire wafer 210 as shown in FIG. 5. The oxide layer 251 is typically thermally grown. The oxide layer is formed such that its thickness is less than the required gate oxide thickness for the high voltage devices by approximately the thickness of the gate oxide required for low voltage devices.

[0045] Next, a second photoresist mask 272 is formed with openings 273 over only the low voltage n-wells 221 as shown in FIG. 6. Then the low voltage n-wells are doped to adjust the threshold voltage of a transistor to be formed later in the n-well 221. It should be noted that the low voltage n-well 221 will be further doped to its final value in a subsequent step as discussed below. Because the subsequent processing involves doping both low voltage devices with p-type material such that the n-channel low voltage device threshold is appropriately set, the doping that occurs through openings 273 at this point involves doping the n-well 221 with a concentration of n-type material higher than that which would be required to properly set the threshold if no subsequent doping were to occur. The excess positive doping is offset by the subsequent negative doping in the steps below. The mask 272 is then removed.

[0046] Then a photoresist mask 274 is formed with openings 275 over the fixture locations of all low voltage devices, including both the n-type and p-type low voltage devices as shown in FIG. 7. The oxide in the openings 275, which correspond to the low voltage devices, is then removed. Then, both the n-channel and p-channel low voltage devices are doped to their final values through the mask openings 275. Then the mask 274 is removed. Additional gate oxide is then grown over the entire wafer such that the total thickness of the gate oxide over the high voltages areas is approximately equal to the required gate oxide for high voltage devices and the gate oxide thickness over the low voltage areas is approximately equal to the required gate oxide for low voltage devices. The resultant structure, with thin gate oxide 251 over the low voltage areas and thick gate oxide 251, 252 over the high voltage areas, is shown in FIG. 8. The process of forming oxides of different thickness over the high and low voltage devices is called gate oxide differentiation. Processing steps after gate oxide differentiation, which involves forming gate stacks, passivation layers, interconnection layers, and other subsequent processing steps, depending upon the specific application, are conventional and will not be discussed in further detail herein.

[0047] The processing illustrated above in FIGS. 1-8 is simple in that it requires three masks to complete processing up to gate oxide differentiation. However, because this processing uses a diffusion well for both high and low voltage devices, the low voltage devices are not as tightly packed as would be possible if retrograde wells were used for the low voltage devices. Next, one possible process conceived by the inventor for fabricating an integrated circuit with both retrograde and diffusion wells and both n-type and p-type high and low voltage devices will be illustrated. FIG. 9 is a cross-sectional view of a portion 300 of a substrate 310 on which four devices will be fabricated at the indicated places: a high voltage p-channel transistor, a low voltage p-channel transistor, a high voltage n-channel transistor and a low voltage n-channel transistor. A diffusion well will be formed for the high voltage p-channel transistor and a retrograde well will be formed for the low voltage p-channel transistor. Processing begins by forming a mask 370 with an opening 371 corresponding to the desired location of the diffusion well for the high voltage device on the substrate 310 as shown in FIG. 9.

[0048] After the mask 370 is formed, low energy ion implantation through the mask opening 371 is performed. The wafer 300 is then heated such that the implanted ions diffuse to form the diffusion n-well 320 shown in FIG. 10. Shallow trench isolation regions 330 are then formed using a conventional process as shown in FIG. 11. The shallow trench isolation (STI) process requires some thermal treatment; therefore, it is performed before retrograde wells are formed so that the thermal treatment used in the STI process does not adversely affect retrograde wells formed later.

[0049] After the isolation regions 330 are formed, a second mask 372 is formed over the substrate 310 as shown in FIG. 12. The mask 372 has an opening 373 to allow high energy (approximately 200 keV to approximately 1 MeV) ion implantation into the substrate 310 to form the retrograde well 321 shown in FIG. 13. The retrograde well 321 is typically more shallow and narrow than the diffusion well 320. A gate oxide layer 351 is then grown over the entire substrate 310 as shown in FIG. 14.

[0050] The retrograde well 321 must be “excessively” doped with n-type material as discussed above in connection with FIG. 6. To accomplish this, a third mask 374 with an opening 375 corresponding to the retrograde well 321 is formed over the substrate 310 as shown in FIG. 15. Excessive doping occurs through the opening 375. The mask 374 is then removed.

[0051] Next, a fourth mask 376 is formed over the substrate 310 with mask openings 377 over both the n-channel and p-channel low voltage transistor locations as shown in FIG. 16. Both the n-channel and p-channel low voltage devices arc doped to their final values through the mask openings 377. The gate oxide 351 is then stripped in the areas corresponding to the mask openings 377. Then the mask 376 is removed. Next, an additional oxide layer 352 is grown over the entire substrate 310 as shown in FIG. 17. The total thickness of the oxide layers 351, 352 over the high voltage device wells is sufficient for high voltage devices, while the thickness of the oxide layer 352 over the low voltage device wells (corresponding to the openings 375) is sufficient for low voltage devices. At this point, processing up to gate oxide differentiation is complete. As discussed above in connection with the single well type fabrication process, subsequent processing is conventional and will not be discussed further herein.

[0052] Four masks were used in the foregoing dual well process to complete the gate oxide differentiation step of the fabrication process. In contrast, only three masks were necessary to reach the same point in the single diffusion well process. Next, an improved method in accordance with the present invention for fabricating a dual well structure requiring only three masks to complete the gate oxide differentiation step will be illustrated.

[0053] Processing for a dual well structure in accordance with the invention is shown starting with FIG. 18. A first mask 470 with an opening 471 for a diffusion well is formed on the substrate 410. Low energy ion implantation occurs through the opening 471 after which the mask is removed and the entire structure 410 is heated to diffuse the implanted ions. The resulting diffusion well 420 is shown in FIG. 19. Isolations regions 430 as shown in FIG. 20 arc then formed by an STI process. Next, as shown in FIG. 21, an oxide layer 451 is grown over the substrate 410. Once again, the thickness of the oxide layer 451 is less than the final desired thickness of the gate oxide for high voltage devices.

[0054] Next, a second mask 472 with an opening 473 is provided over the substrate 410 as shown in FIG. 22. High energy ion implantation occurs through the opening 473 to form a retrograde well 421 as shown in FIG. 23. At this point, the processing deviates from the four-mask processing discussed above in connection with FIGS. 9-17.

[0055] After the high energy ion implantation occurs through opening 473 and the retrograde well 421 is formed, the well is doped through the same opening 473. This results in the doping of non-active areas (i.e., source, drain and area under the gate stack) in the retrograde well, but this is acceptable. The doping that occurs at this point results is “excessive,” meaning that the well 421 is more positively doped at this point than it will be at the end of processing. Further doping, as discussed below, will set the gate threshold to the proper level. The mask 472 is removed after the retrograde well has been initially doped as shown in FIG. 24.

[0056] After the mask 472 is removed, a third mask 476 with openings 477 over both the n-channel and p-channel low voltage device locations is formed over the Substrate 410 as shown in FIG. 25. The gate oxide layer 451 is stripped in areas exposed by the openings 477.

[0057] Next, the exposed areas corresponding to the openings 477 are doped such that the gate threshold of both the low voltage and high voltage n-channel devices is at the proper value. Then the third mask 476 is removed, and a second oxide layer 452 is grown as shown in FIG. 26. Gate differentiation is now complete. Again, conventional back end processing is performed to form gate stacks 450, source areas 441 and drain areas 442 to complete the transistors 470 as shown in FIG. 27. Further processing, such as interconnection/metallization and passivation, which are all well-known in the art, is also conventional and hence will not be discussed in further detail herein.

[0058] An efficient fabrication method for producing a structure with two different types of wells has been shown. This method requires only three masks to complete processing to the gate oxide differentiation step. In a preferred embodiment, the different well types arc diffusion n-wells and retrograde n-wells. However, the method is easily modified to form diffusion p-wells and retrograde p-wells in an n-type substrate. Additionally, the method is applicable in situations where not all device types are required. For example, in the preferred embodiment, the method does not depend upon whether high voltage n-channel transistors are required.

[0059] While the invention has been described in detail in connection with the preferred embodiments known at the time, it should be readily understood that the invention is not limited to such disclosed embodiments. Rather, the invention can be modified to incorporate any number of variations, alterations, substitutions or equivalent arrangements not heretofore described, but which are commensurate with the spirit and scope of the invention. Accordingly, the invention is not to be seen as limited by the foregoing description, but is only limited by the scope of the appended claims.