Background of the Invention
1. Field of the invention
The present invention relates to an improved fabrication process for CMOS integrated devices having a reduced gate length and lightly doped drain regions of transistors of both polarities.
2. Description of the prior art
At present in CMOS fabrication processes with a gate length of about a micron (.mu.m) or less, the p-channel transistors are fabricated with dimensions greater than the minimum dimensions proper of the fabrication process because the high diffusivity of the p-type dopant commonly used for p.sup.+ junctions, typically boron, produces an effective channel length sensibly reduced in respect to the one obtained in n-channel transistors having the same gate length, when using conventional heat treatments. In this situation the p-channel transistors are much more subject to develop malfunctioning problems, such as the punch-through phenomenon, than n-channel transistors especially in devices destined to operate with a relatively high supply voltage (12 V), and therefore must necessarily be fabricated with adequately increased dimensions, thus reducing the density of integration which may be achieved. In the alternative the problem may be overcome by means of the known technique of forming lightly doped drain regions close to the sides of the transistors' gate. The use of lightly doped drain regions allows to exploit the minimum dimensions of definition of the fabrication process also for p-channel transistors, by spacing apart the p.sup.+ diffusions by means of spacers of dielectric material which are purposely formed on the sides of the gate, according to well known techniques. The formation of lightly doped drain regions commonly implies the repetition of the n.sup.+ and p.sup.+ implantations maskings in order to repeat the implantation of the relevant dopants before and after forming the spacers. This increases substantially fabrication costs because of the two additional masking operations which are required in respect to a standard process (without the formation of lightly doped drain regions, or LDD) or to a process wherein the formation of said lightly doped drain regions takes place on transistors of a single polarity.
Objective of the Invention
Objective of the present invention is an improved fabrication process for integrated CMOS devices through which it is possible to form n-channel and p-channel transistors both provided with lightly doped drain regions (LDD) by means of a single additional masking step.
Brief Description of the Drawings
The FIGS. 1 to 8 show the essential steps of a fabrication process in accordance with the present invention. The FIGS. 9-10 show an alternative order of carrying out relevant steps of the process.
Description of a Preferred Embodiment
In the case of fabricating n-well architecture devices, the process of the invention comprises the following steps:
1. formation of the n-well diffusion (2) in a monocrystalline, p-type silicon substrate (1), according to common practices (FIG. 1);
2. definition on the front of the wafer of the active areas by growing an isolation field oxide layer (3) among adjacent active areas, according to common practices (FIG. 2);
3. growth of a gate oxide layer (5) over the active areas of the device, according to common practices; and
4. deposition, doping and definition of a polycrystalline silicon gate layer (6) inside the active areas, in accordance with common practices (FIG. 3);
5. unmasked implantation on the entire front of the device of an n-type dopant for forming n.sup.- type lightly doped drain regions (LDD) (FIG. 4);
6. first formation of the "p.sup.+ mask" (7) and implantation of a p-type dopant on the active areas of p-channel transistors, in a dose sufficient to compensate and invert J completely the preceding n.sup.- implantation in order to form p.sup.- type lightly doped drain regions (FIG. 5);
7. deposition and etching of a dielectric material (e.g. silicon oxide) for forming lateral spacers (8), according to common practices (FIG. 6);
8. formation of the "n.sup.+ mask" (9) and implantation of an n-type dopant for forming n.sup.+ junctions, according to common practices (FIG. 7);
9. second formation of the p.sup.+ mask (10) and implantation of a p-type dopant for forming p.sup.+ junctions, according to common practices (FIG. 8);
10. formation of the contacts and of the interconnecting lines and execution of the finishing steps of the fabrication process, according to common practices.
Of course, as it will be evident to the skilled technician, the process of the invention is applicable also for fabricating devices with different architectures such as for example p-well and twin-well devices by appropriately inverting polarities. Moreover any technique for defining the active areas and for forming the isolation structure (field oxide, buried oxide, etc.) as well as for forming gates with polycrystalline silicide or other conducting material or with stacked layers of polycrystalline silicon and silicide and various known techniques for forming the junctions, such as the use of particular dopant species, of self-aligned silicides and for forming spacers with materials different from the silicon oxide may be satisfactorily employed in the fabrication process of the invention.
The 5. and 6. steps of the process may obviously also be carried out in a different chronological order and/or by inverting the polarities of the mask used and of the dopants, as long as the compensation of the second implantation is ensured, as depicted in FIGS 9 and 10.
It is essential that the implantations relative to the steps 5 and 6 which characterize the improved fabrication process of the invention be carried out so that the dose of the first implanted dopant on the active areas of n-channel transistors as well as on the active areas of p-channel transistors is completely compensated and inverted by the subsequent masked implantation on the transistors of one specific channel polarity, in order that the total active concentration of the dopant be adequate to form an efficient lightly doped drain region (LDD) having an effective concentration of about 10.sup.18 atoms per cubic centimeter.
Example
Integrated p-channel transistors with a gate length of 0.8 .mu.m, with different combinations of n.sup.- and p.sup.- implantations have been formed on a p-type monocrystalline silicon by means of a polycrystalline silicon gate CMOS fabrication process, modified in accordance with the present invention in respect to a standard sequence of process steps while following substantially standard procedures for carrying out the different process steps.
Over an n.sup.- well region formed by an unmasked implantation of phosphorus in a dose of 3.times.10.sup.13 atoms per cm.sup.2, carried out at 60 KeV (step 5 of the process sequence described above), different boron doses (p.sup.-) utilizing BF.sub.2, respectively of 4, 7 and 10.times.10.sup.13 atoms per cm.sup.2 and implanted at 60 KeV (step 6 of the process sequence described above), have been tested.
In all three cases the p-channel transistors so formed proved to be satisfactorily functioning up to the contemplated maximum supply voltage of 12 V and had threshold voltages of about -1.1 V in a linear zone.
Transistors made with the same process but with a gate length reduced to 0.6 .mu.m were also operating satisfactorily at a lower supply voltage of 5 V and the breakdown voltage of the p.sup.+ junction was considerably higher in respect to that of a p.sup.+ junction without the lightly doped drain structure formed in accordance with the present invention.