Graded LDD Implant Process for Sub-Half-Micron MOS Devices

This invention relates to semiconductor integrated circuit fabrication and, more particularly, to improved processes for fabricating MOS field effect transistors having graded lightly doped drain and source regions.

Semiconductor integrated circuits are comprised of a plurality of devices which invariably include transistors. Transistors are of two general types, namely bipolar and field effect transistors (FETs).

FIG. 1 shows the common type of FET structure generally used to form metal oxide substrate (MOS) type circuits. In this case, an N-channel MOS FET (NMOSFET) structure comprises a substrate 1 of semiconducting material such as silicon having a region which has been doped to form a "P-well" 2. A detailed description of this process may be found in "Silicon Processing For The VLSI Era"--Volume 2, Process Integration, Lattice Press 1990, pp 428-441. An active area 3 is defined between isolating field oxide regions 4 and 5. A gate region 6 of conductive material such as polysilicon (poly) is separated from the surface by layer 7 of dielectric material such as silicon dioxide (SiO.sub.2). Conductive interconnect material 8 such as tungsten silicide is formed above and in contact with the gate region which interconnects the gate to other circuit devices. Implanted into the surface of the P-well 2 astride the gate region 6 are source 9 and drain 10 regions of N type semiconductor material with the FET channel area 11 formed in between. A cap layer 12 and sidewall spacer structures 13 of insulating material such as nitride protect the gate structures during subsequent processing such as the self aligned implants of the source and drain regions and the formation of conductive structures which interconnect these regions.

MOSFETs in combination with other devices commonly form dynamic random access memory circuits (DRAM) used in memory systems such as computers. Because of the continuous demand for the further miniaturization and speed increase of DRAMs, MOSFET devices have been scaled to the point where the channel length from source to drain falls below 0.5 micron (sub-half micron). As the channel shrinks, the maximum electric field (E-field) in the channel region increases, thereby resulting in higher substrate current and short/long term hot electron reliability problems. Electrons traveling through the channel become more energized by the E-field and have a greater tendency to cross into the gate region 6 and become trapped. These problems are discussed in detail in "Silicon Processing For The VLSI Era--Volume 2",Lattice Press, 1990, pp 428-441.

The reference cited above also discusses various methods employed to partially overcome these problems and maximize performance and reliability. One common method involves adding a first lightly doped region between the drain and channel regions and a second lightly doped region between the source and channel regions. FIGS. 2 and 3 show the typical fabrication sequence for this structure. In FIG. 2, after formation of the insulating sidewall spacer structures 13, a low dosage phosphorous implant and drive create N- regions 14, 15. Due to the relatively high diffusivity of phosphorous, the N- regions extend underneath the spacers toward the FET channel area 11. In FIG. 3, a high dosage arsenic implant and drive creates N+ source 16 and drain 17 regions which supersede most of the lightly doped N- regions. What remains are lightly doped regions 18, 19 separating the source and drain from the channel. This structure has come to be known as a lightly doped drain (LDD) structure. The use of LDD structures to relax the E-field is well known.

However, as the devices get smaller, and FET channels become shorter than 0.4 microns, limitations on fabrication precision result in structures that are far from the ideal one shown in FIG. 3. Due to its high diffusivity, the phosphorous in the N- regions further diffuses into the channel during the high heat drive processes required to create the N+ source and drain regions. This causes severe short channel problems resulting in increased sub-threshold leakage which adversely affects refresh time in DRAMs.

An alternative to the phosphorous LDD (phos-LDD) approach is to use arsenic to create the LDD structures as proposed by H. R. Grinolds, et al. in "Reliability and Performance of Submicron LDD NMOSFET's with Buried-As n-Impurity Profiles," IEDM Tech. Dig., 1985, pp. 246-249 and by C. -Y. Wei, et al. in "Buried and Graded/Buried LDD Structures for Improved Hot-Electron Reliability," IEEE Electron Device Lett., vol. EDL-7, np. 6, June 1986.

The fabrication processes required to create an arsenic LDD (As-LDD) proceed similarly to the phosphorous LDD processes. FIGS. 4 and 5 show that an LDD structure can be created by first implanting a low dosage, self-aligned arsenic implant prior to sidewall spacer formation. This implant is then diffused into the substrate through a heating drive process, resulting in the lightly doped N- regions 20 and 21.

In FIG. 5, after the formation of insulating sidewall spacer structures 13, the N+ source 22 and drain 23 regions are created using a high dosage, self-aligned arsenic implant and drive. The N+ regions supplant portions of the N- regions. There remains, however, a first lightly doped N- LDD region 24 existing between the N+ source region 22 and the FET channel area 11, and a second lightly doped N- LDD region 25 existing between the N+ drain region 23 and the FET channel area 11.

Since low diffusivity arsenic was used to create the N- LDD regions, the resulting LDD structures are much more predictable and do not suffer from the short channel problems plaguing phosphorous LDD structures. However, arsenic's low diffusivity also causes the NLDD regions 24, 25 to have an abrupt end 26,27 below the edges of the gate region 6. This abruptness creates an E-field which is still unsuitable in sub half-micron devices due to the resulting hot electron reliability problem.

To alleviate this problem, a combination phos/As LDD structure has been developed where a phos-LDD implant occurs immediately after an As-LDD implant to grade the channel to LDD junction. Again, due to the diffusivity of phosphorous, during subsequent processing, the short channel characteristics are degraded.

Another method to reduce the E-field involves burying the drain/channel and source/channel junctions. FIG. 6 shows an NMOSFET having buried drain/channel and source/channel junctions. In this FET, the N+ source and drain regions 28, 29 each have a projection 30, 31 which terminates at a junction 32, 33 with the FET channel area 11. The projections exist a distance below the channel/gate dielectric material layer 7. This moves the highest concentration of hot-electrons deeper into the channel area and away from the gate dielectric region. Buried structures incorporating LDD regions and graded combination structures have also been created, but at the expense of device speed. Although these structures offer promise, they are quite costly to construct and, therefore, not currently viable solutions, economically.

It would be desirable, therefore, to have a process which produces a sub half-micron MOSFET with a low E-field and improved short channel characteristics and reliability in an efficient and economical manner.

The primary and secondary objects of this invention are to provide a process for creating reliable and inexpensive sub-half-micron NMOSFETs.

These and other objects are achieved by a process wherein a low dosage N- phosphorous implant occurs after the high dosage N+ arsenic implant and drive which creates the source and drain.

FIG. 1 is a prior art cross-sectional view of a MOSFET having non-graded source and drain electrodes;

FIG. 2 is a prior art cross-sectional view of an in-process NMOSFET during a low dosage phosphorous implant of a pair of N- LDD structures;

FIG. 3 is a prior art cross-sectional view of the in-process NMOSFET of FIG. 2 during a high dosage arsenic implant to create N+ source and drain regions;

FIG. 4 is a prior art cross-sectional view of an in-process NMOSFET during a low dosage arsenic implant of a pair of N- LDD structures;

FIG. 5 is a prior art cross-sectional view of the in-process NMOSFET of FIG. 4 during a high dosage arsenic implant to create N+ source and drain structures;

FIG. 6 is a prior art cross-sectional view of a MOSFET having buried source and drain regions;

FIG. 7 is a cross-sectional view of an in-process NMOSFET using the invention after field oxide deposition, active area definition, sacrificial oxide deposition and threshold adjustment implantation;

FIG. 8 is a cross-sectional view of the in-process NMOSFET of FIG. 7 after removal of the sacrificial oxide, growth of a gate oxide layer and deposition of several layers of gate forming materials;

FIG. 9 is a cross-sectional view of the in-process NMOSFET of FIG. 8 after a gate defining mask and etch process;

FIG. 10 is a cross-sectional view of the in-process NMOSFET of FIG. 9 after a low dosage arsenic implant to form N- LDD regions;

FIG. 11 is a cross-sectional view of the in-process NMOSFET of FIG. 10 after gate sidewall oxide growth and deposition of a layer of nitride;

FIG. 12 is a cross-sectional view of the in-process NMOSFET of FIG. 11 after a sidewall spacer creating etch;

FIG. 13 is a cross-sectional view of the in-process NMOSFET of FIG. 12 after a high dosage arsenic implant to create N+ source and drain regions;

FIG. 14 is a cross-sectional view of the in-process NMOSFET of FIG. 13 after a low dosage phosphorous implant to grade both the N+/N- and the N-/channel junctions.

Referring now to the drawings, FIG. 7 shows a cross-section of a silicon substrate 34 doped to form a "P-well" upon which an active area has been defined between implanted field oxide regions 35, 36 and below a sacrificial oxide layer 37. A threshold voltage adjustment implant may also be performed at this stage.

After the sacrificial oxide layer is removed, FIG. 8 shows that a gate oxide layer 38 is grown. Atop the gate oxide layer, a conductive layer of polysilicon ("poly") 39 is deposited. Atop the poly is a more conductive tungsten silicide (WSi.sub.x) interconnect layer 40 and atop this layer is a protective and insulating nitride layer 41.

A gate definition etch is performed resulting in the structure of FIG. 9. A gate dielectric oxide region 42 separates a poly gate region 43 from the silicon substrate 34. A boron halo implant 44 is then performed to optimize the concentration of p-type charge carriers in areas of the substrate outside the channel area 45.

In FIG. 10, a low dosage arsenic implant and drive creates N- LDD regions 46,47 on either side of the channel. The gate structure 48 itself protects the channel from both the boron halo and arsenic implants.

Next, in FIG. 11, a thin (60 to 120 angstroms) layer of oxide 49 is grown on the poly gate sidewalls and the exposed surfaces of the LDD regions to further protect the gate during the subsequent blanket deposition of an insulating layer of nitride 50.

FIG. 12 shows the in-process MOSFET after an anisotropic etch is performed to remove the horizontal portions of the spacer nitride layer thereby leaving nitride sidewall spacers 51,52. The thin oxide layer has also been removed from those portions of the active area left unprotected by the nitride structures. The spacers cover over a portion 53 of each As-LDD region implanted earlier.

FIG. 13 shows the result of a high dosage arsenic implant and drive to create N+ source 54 and N+ drain 55 regions of the MOSFET. Note the adjacent low dosage N- As-LDD regions 56,57 separating the N+ regions from the channel area 45.

Next, as seen in FIG. 14, a low dosage phosphorous implant and drive is performed to grade both the junctions 58,59 between the N+ and N- arsenic LDD regions and the junctions 60,61 between the N- arsenic LDD regions and the channel. The major problem with phosphorous, its high diffusivity, has been side-stepped by implanting after the high heat, high dosage implant and drive which creates the N+ source and drain regions.

Further steps to realize the completed MOSFET involve steps familiar in the art such as the deposition of a layer of boro-phospho-silicate glass (BPSG), reflow and formation of conductive contacts to the source and drains.

The finished MOSFET exhibits a reduced E-field due to the phosphorous grading of the junctions. Since the phosphorous is implanted late in processing, its diffusive nature is more controlled, allowing for a reliable sub-half-micron device. The elimination of numerous steps, along with the self-aligned nature of the existing steps, results in achieving the previously economically unfeasible sub-half-micron MOSFET.

Another advantage is that the phosphorous implant after the source/drain formation allows grades in the junction between the N+ regions and the boron halo implant, resulting in reduced junction leakage and less N+ junction capacitance.

A possible disadvantage to this scheme involves the addition of a mask to protect in-process devices located on other areas of the wafer during this low-dosage phosphorous implant. A typical area sensitive to phosphorous would be the array on an in-process DRAM chip where field oxide regions are narrow. However, in processes where the spacer etch is performed after the cell poly etch, the array will not be exposed because it is still covered with photoresist at that point. In this case, no extra mask would be required.

While the preferred embodiments of the invention have been described, modifications can be made and other embodiments may be devised without departing from the spirit of the invention and the scope of the appended claims.