Method of Constructing CMOS Vertically Modulated Wells (VMW) by Clustered MeV BILLI (Buried Implanted Layer for Lateral Isolation) Implantation

1. Field of the Invention

The present invention relates to simplification of a manufacturing process for complementary semiconductor devices, which are becoming increasingly miniaturized and increasingly complex, to a high component-density semiconductor device having enhanced resistance to CMOS latch-up, and to well formation (single, twin, triple).

2. Description of the Related Art

The importance of CMOS technology in the VSLI field has grown, as a result of requirements for a high noise margin and low power consumption. However, as miniaturization has increased, serious problems have arisen with regard to preventing stray thyristor operation which causes the CMOS latch-up phenomenon to occur between mutually adjacent portions of an n-channel MOSFET and p-channel MOSFET, and with regard to maintaining a sufficient level of withstanding voltage between mutually adjacent elements.

Various forms of device configuration and manufacturing process have been proposed for overcoming these problems. These proposals include the use of a configuration containing wells, formation of a buried high concentration layer, and formation of a self-aligned channel stop at the edge of a well region.

Summaries of these various structures and methods have been presented, for example, in U.S. Pat. No. 5,160,996 at column 1 line 34 et seq. Additional disclosure of this technology appears in the article entitled "MeV implantation technology: Next-generation manufacturing with current-generation equipment" by Borland and Koelsch in the December 1993 issue of Solid State Technology.

The present invention solves the following problems:

1) single, twin and triple well CMOS process simplification:

a) total number of steps

b) 2-3 mask levels

c) manufacturing time and cost reduction

d) clustering well, isolation and channel implants, and V.sub.t implants

e) clustering through a masking layer

2) clustered implant to reduce wafer handling, particles, process simplification

3) CMOS latch-up, SER, .alpha.-particle, GOI, and ESD

4) improved bulk wafer by hydrogen denuded zone

5) elimination of Epi for

a) CMOS

b) SOI wafer bonding

c) CCD

6) oxygen related implant defects

The foregoing problems are solved in the following manner

1) clustered implants up to six into one step

2) BILLI structure formation for latch-up and single, twin and triple well

3) Hydrogen denudation for improved Cz wafer and epi elimination

The novel features of the invention include the following:

1) clustered implants

2) reduction in masks and process for twin and triple well

3) epi replacement for CMOS, SOI, CCD etc.

4) latch-up free

5) hydrogen passivation

6) oxygen-denuded-zone free

7) defect free

8) implant-related-defect free

The features of the present invention may be summarized as follows:

1) CMOS semiconductor device manufacturing process simplification and cost reduction for NMOS and PMOS transistor fabrication:

a) single n or p well formation on p- or n-substrates.

b) twin/double n & p well formation on p- or n-substrates.

c) triple well formation, surface n & p well plus buried n or p well on p- or n-substrate.

d) Vt implant through a masking layer.

e) clustered implants; deep retro-well implant/channel stop implant/Vt implant.

deep retro p or n-well implant/channel stop implant.

channel stop implant/Vt implant

deep retro p or n-well implant/channel stop implant/Vt implant.

deep retro p or n-well implant/channel stop implant/deep retro n or p-well implant/channel stop implant.

p or n-well channel stop implant/Vt implant/n or p-well channel stop implant/Vt implant.

deep retro p or n-well implant/channel stop implant/Vt implant/deep retro n or p-well implant/channel stop implant/Vt implant.

f) clustered implants through a masking layer; deep retro-well implant/channel stop implant/Vt implant.

deep retro p or n-well implant/channel stop implant.

channel stop implant/Vt implant

deep retro p or n-well implant/channel stop implant/Vt implant.

deep retro p or n-well implant/channel stop implant/deep retro n or p-well implant/channel stop implant.

p or n-well channel stop implant/Vt implant/n or p-well channel stop implant/Vt implant.

deep retro p or n-well implant/channel stop implant/Vt implant/deep retro n or p-well implant/channel stop implant/Vt implant.

g) elimination of 1 to 4 masking layers.

h) elimination of up to 4 medium current implants and the associated implanters.

i) improved high energy implanter utilization.

j) improved high energy implanter productivity by 25%.

k) cost savings of $25 to $149 per 200 mm wafer.

l) reduction in masking layer thickness for high energy ion implantation.

2) Improved CMOS latch-up resistance and device scaling/shrink.

a) reduction in lateral current gain/lateral beta (B.sub.L).

b) reduction in vertical current gain/vertical beta (B.sub.v).

c) reduction in well resistances (R.sub.W).

d) reduction in substrate resistance (R.sub.S),

e) improved n+ to p+ spacing.

3) Hydrogen denudation for improved Cz silicon wafer surface properties.

a)improved device performance

gate oxide integrity

oxide QBD

junction leakage

device yield

b) improved oxygen out diffusion

c) lower surface oxygen

d) lower surface defect level

e) equivalent to epi wafer surface quality

f) improved resistance to surface native oxide

g) improved wafer surface roughness

h) pre-process intrinsic gettering for high energy implant device processing

4) BILLI structure plus H.sub.2 denudation=epi replacement

a) CMOS technology

b) CCD technology

The invention may best be understood from the following detailed description thereof, having reference to the accompanying drawings, in which:

FIG. 1 is a schematic view of a BILLI structure in accordance with the invention;

Jig. 2 is a schematic view of the steps involved in manufacturing a BILLI twin-well structure with two masks and clustered implantation;

FIG. 3 is a schematic view of the steps involved in manufacturing a BILLI triple-well structure with two masks and clustered implantation;

FIG. 4 is a schematic view showing an alternative process to that of FIG. 3;

FIG. 5 is a schematic view showing another alternative process to that of FIG. 3.

Referring to the drawings, and first to FIG. 1 thereof, FIG. 1 shows a low cost MeV structure in accordance with the invention at the stage in its manufacture at which PMOS devices can be formed in a retrograde n-well and NMOS devices in a retrograde p-well. Although at this stage in its manufacture neither the PMOS devices nor the NMOS devices have been formed, in FIG. 1 PMOS devices are shown, in order to indicate where they will be formed in steps subsequent to said stage. Isolation areas 5 separate adjacent PMOS devices which can be implanted into a retrograde n-well formed in a p-type substrate 10. Adjacent to the series of PMOS devices in an n-well retrograde p-wells have been formed, into which a series of NMOS devices will be implanted in subsequent steps of manufacture. The retrograde n-well includes an upper layer 6 wherein phosphorus ions have been implanted so as to form phosphorus (n) impurity atoms with a density that is produced by 5E12cm-2! and a lower layer 7 wherein phosphorus ions have been implanted so as to form phosphorus (n) impurity atoms with a density that is produced by 3E13cm-2. The retrograde p-well includes an upper layer 8 wherein boron ions have been implanted so as to form boron (p) impurity atoms with a density that is produced by 5E12cm-2! and a lower layer 9 wherein boron ions have been implanted so as to form boron (p) impurity atoms with a density that is produced by 3E13cm-2. The term "upper layer" signifies the layer which is nearer the active surface of the substrate 10 and is 0.5 .mu.m deep. The term "lower layer" signifies the layer which is farther from the active surface of the substrate 10 and is 1.2 .mu.m deep. These are retrograde wells because the impurity-atom density is higher at the lower part of the well than at the upper part of the well.

The BILLI structure is formed by continuance of the boron layers from the depths at which they are formed beneath a thick masking layer, for example: comprising at least 2 .mu.m of photoresist to the depths at which they are formed in the absence of the thick masking layer. The latter depths are under the retrograde n-well, deeper than both the upper layer 6 and the lower layer 7, and the layers at these depths are designated "buried". The boron layers at these depths thus form a "buried implanted layer for lateral isolation" or "BILLI" structure, since they surround the n-well around the sides and the bottom. This BILLI structure provides best latch-up resistance for miminum n+ to p+ spacing even over epi-wafers. When combined with hydrogen denudation, this BILLI structure on bulk Cz wafers can lead to epi replacement.

FIG. 2, in conjunction with FIG. 1, shows a method of forming the BILLI structure in accordance with the invention. This method of the invention not only has the advantage of resulting in the formation of the novel BILLI structure of the invention, but also results in both process simplification and cost reducton for single, twin and triple well formation. A preferred embodiment of the method of the invention comprehends the following steps. First, isolation areas 12 are formed (or, alternatively, active areas are formed) in the substrate 10. Then a mask 11, such as a photoresist mask having a thickness of more than 2.0 .mu.m is placed on the surface as shown to block the phosphorus ions of maximum energy, and a clustered series of four to six implants are carried out.

The clustered series of four to six implants may be carried out as follows: First, boron ions are directed onto the upper surface of the substrate 10 with an energy of 2.00 MeV, thereby forming the lower layer 9 of p-type material (deep retrograde p-well). Second, without removing the substrate 10 from the vacuum chamber in which it was irradiated with boron ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 10 with an energy of 1.25 MeV, thereby forming the upper layer 8 of p-type material (channel stop implant or shallow retrograde p-well). As an optional third step, without removing the substrate 10 from the vacuum chamber in which it was irradiated with boron ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 10 with an energy of 750 keV, thereby forming a shallow, thin threshhold voltage layer V.sub.t at the surface of the retrograde p-well 8,9. Fourth, without removing the substrate 10 from the vacuum chamber in which it was irradiated with boron ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 10 with an energy of 1 MeV, thereby forming the lower layer 7 of n-type material (deep retrograde n-well). Fifth, without removing the substrate 10 from the vacuum chamber in which it was irradiated with boron ions and phosphorus ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 10 with an energy of 450 keV, thereby forming the upper layer 6 of n-type material (channel stop implant or shallow retrograde n-well). As an optional sixth step, without removing the substrate 10 from the vacuum chamber in which it was irradiated with boron ions and phosphorus ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 10 with an energy of 60 keV, thereby forming a shallow, thin threshhold voltage layer V.sub.t at the surface of the retrograde n-well 6,7. Alternatively, the sixth step may comprise directing boron ions onto the upper surface of the substrate 10 with an energy of 30 keV so as to form this threshhold voltage layer.

The method of FIG. 2 may be modified by omitting the two optional steps and, in lieu thereof, carrying out a medium current blanket Vt implant after the M1 step and before the M2 step. Alternatively, the sixth step may be omitted and, in lieu thereof, a medium current blanket Vt implant may be carried out after the M1 step and before the M2 step.

As shown in FIG. 3, if the steps carried out in the method of FIG. 2 are applied to an n-type substrate, a BILLI triple well structure is formed. Alternatively, a BILLI triple well structure may be formed by the clustered implantation indicated in FIG. 4, wherein the implantation steps are carried out on a p-type substrate in the following sequence.

First, phosphorus ions are directed onto the upper surface of the substrate 110 with an energy of 3.0 MeV, thereby forming the lower layer 109 of n-type material. Second, without removing the substrate 110 from the vacuum chamber in which it was irradiated with phosphorus ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 110 with an energy of 2.25 MeV, thereby forming the upper layer 108 of n-type material. As an optional third step, without removing the substrate 10 from the vacuum chamber in which it was irradiated with phosphorus ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 110 with an energy of 2.0 MeV, thereby forming a shallow, thin threshhold voltage layer V.sub.t at the surface of the retrograde n-well 108,109. Fourth, without removing the substrate 110 from the vacuum chamber in which it was irradiated with phosphorus ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 110 with an energy of 500 keV, thereby forming the lower layer 107 of p-type material. Fifth, without removing the substrate 110 from the vacuum chamber in which it was irradiated with phosphorus ions and boron ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 110 with an energy of 250 keV, thereby forming the upper layer 106 of p-type material. As an optional sixth step, without removing the substrate 110 from the vacuum chamber in which it was irradiated with phosphorus ions and boron ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 110 with an energy of 30 keV, thereby forming a shallow, thin threshhold voltage layer V.sub.t at the surface of the retrograde p-well 106,107.

Alternatively, a BILLI triple well structure may be formed by the clustered implantation indicated in FIG. 5, wherein the bottom ("buried") n-layer is formed prior to the clustered implant steps by a separate implantation of 3.0 MeV phosphorus ions at a current of 3 to 5 E13 through a separate mask, designated "M2" in FIG. 5. The remaining, clustered implantation steps are carried out through a separate mask, designated "M3" in FIG. 5, in the following sequence.

First, boron ions are directed onto the upper surface of the substrate 110 with an energy of 2.0 MeV and a dose of 0.5 to 1 E13, thereby forming the lower layer 107 of p-type material in the p-wells. Second, without removing the substrate 110 from the vacuum chamber in which it was irradiated with boron ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 110 with an energy of 1.25 MeV, thereby forming the upper layer 106 of p-type material in the p-wells. As an optional third step, without removing the substrate 110 from the vacuum chamber in which it was irradiated with boron ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 110 with an energy of 750 keV, thereby forming a shallow, thin threshhold voltage layer V.sub.t at the surface of the retrograde p-well 106,107. Fourth, without removing the substrate 110 from the vacuum chamber in which it was irradiated with boron ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 110 with an energy of 1.0 MeV, thereby forming the lower layer 109 of n-type material. Fifth, without removing the substrate 110 from the vacuum chamber in which it was irradiated with boron ions and phosphorus ions, the parameters of the ion accelerator are changed, and phosphorus ions are directed onto the upper surface of the substrate 110 with an energy of 450 keV, thereby forming the upper layer 108 of n-type material. As an optional sixth step, without removing the substrate 110 from the vacuum chamber in which it was irradiated with boron ions and phosphorus ions, the parameters of the ion accelerator are changed, and boron ions are directed onto the upper surface of the substrate 110 with an energy of 30 keV, thereby forming a shallow, thin threshhold voltage layer V.sub.t at the surface of the retrograde n-well 106,107. Alternatively, the sixth step may comprise directing phosphorus ions onto the upper surface of the substrate 110 with an energy of 60 keV so as to form this threshhold voltage layer.

In the foregoing description of clustered implantation, the sequence of clustered implantions are carried out by the same ion accelerator. However, if many ion accelerators are available at the same installation, the sequence of clustered implantations may be carried out by separate ion accelerators without departing from the spirit and scope of the invention; it is only necessary to ensure that the same mask is used throughout the sequence of clustered implantations.

Also, in the foregoing description of clustered implantation, representative thicknesses of the masks are set forth, and representative energies of the ions being implanted are set forth. However, the invention is not limited to such thicknesses or energies, and the invention comprehends, in general, phosphorus-ion energies which are insufficient to pass through the mask involved, thus being blocked by the mask, and boron-ion energies which are sufficient to pass through the mask involved.

Additional advantages of the invention are achieved if the implantation and masking steps hereinbefore set forth are carried out on a substrate the surface whereof has been subjected to hydrogen annealing. The resultant hydrogen denudation provides an improved Cz wafer and can eliminate epi. Using the BILLI structure in combination with hydrogen denudation for epi replacement results in epi equivalent thin gate oxide quality, excellent junction leakage, improved resistance to RCA wet clean related surface micro-defect formation, improved surface smoothness and very low surface oxygen and defect levels.

In the foregoing description the invention has been described in terms of boron ions and phosphorus ions. However, the invention includes the use of p-type dopants other than boron in lieu of boron ions, as well as the use of n-type dopants other than phosphorus in lieu of phosphorus ions.

Having thus described the principles of the invention, together with several illustrative embodiments thereof, it is to be understood that, although specific terms are employed, they are used in a generic and descriptive sense, and not for purposes of limitation, the scope of the invention being set forth in the following claims.