Complementary heterogeneous MOSFET using global SiGe substrate and hard-mask memorized germanium dilution for nFET

1. A method, comprising:

providing a substrate that underlies a layer of SiGe;

forming a plurality of fins in the layer of SiGe, each formed fin comprising a hard mask layer on a top surface;

depositing Si on a first subset of the set of fins in what will be an nFET area;

performing a Si—Ge inter-mixing process on the first subset of fins to reduce a concentration of Ge in the first subset of fins while producing a Si—Ge intermix layer;

removing the Si—Ge intermix layer leaving the first subset of fins having the reduced concentration of Ge; and

forming a second subset of fins in what will be a pFET area, the second subset of fins being formed from the layer of SiGe and having a greater percentage of Ge than a percentage of Ge in the first subset of fins.

2. The method of claim 1 , further comprising a step of iterating the steps of depositing Si on the first subset of the set of fins, performing the Si—Ge inter-mixing process and removing the Si—Ge intermix layer.

3. The method as in claim 2 , where the step of iterating is performed until the first subset of fins is comprised of tensile stressed substantially pure Si.

4. The method as in claim 2 , where the hard mask layer preserves a location and shape of the first subset of fins during the iterated steps of depositing Si on the first subset of the set of fins, performing the Si—Ge intermixing process and removing the Si—Ge intermix layer.

5. The method as in claim 1 , where providing the substrate that underlies the layer of SiGe provides a semiconductor on insulator structure having a layer of oxide disposed between a strained SiGe layer and the substrate.

6. The method as in claim 5 , where the first subset of fins and the second subset of fins are formed at the same time, and where the second subset of fins are masked during the steps of depositing Si on the first subset of the set of fins in what will be an nFET area, performing the Si—Ge inter-mixing process, and removing the Si—Ge intermix layer.

7. The method as in claim 1 , where providing the substrate that underlies the layer of SiGe provides a semiconductor on insulator structure having a layer of oxide disposed between a strain-relaxed SiGe layer and the substrate.

8. The method as in claim 7 , where the second subset of fins are formed subsequent to performing the steps of depositing Si on the first subset of the set of fins, performing the Si—Ge inter-mixing process, and removing the Si—Ge intermix layer.

9. The method as in claim 8 , where the strain-relaxed layer of SiGe is a layer of Si 1-x Ge x where x has a non-zero value=n, and where forming the second subset of fins comprises annealing the second subset of fins to form a set of condensed Si 1-x Ge x fins, where x has a value=m, where m>n, where the step of annealing forms a thermal oxide that covers at least vertical surfaces of individual ones of the set of condensed Si 1-x Ge x fins.

10. The method as in claim 9 , where n is equal to about 0.25, and where m is equal to about 0.50.

11. The method as in claim 9 , where individual ones of the second subset of fins have a first width, where the step of annealing consumes some of the Si 1-x Ge x of the fins to form the thermal oxide, and where individual ones of the fins of the second subset of fins have a second width that is less than the first width.

12. The method of claim 9 , where individual ones of the second subset of fins have a first width, where the step of annealing consumes some of the Si 1-x Ge x of the fins to form the thermal oxide, and further comprising removing the thermal oxide such that individual ones of the fins of the second set of fins have a second width that is less than the first width.

13. The method of claim 9 , further comprising masking the first subset of fins prior to forming the second subset of fins and annealing the second subset of fins.

14. The method of claim 9 , where individual ones of the second subset of fins have a first width defined by a width of the hard mask layer and a width of a spacer disposed at least on sidewalls of the hardmask layer, and where the step of annealing consumes some of the Si 1-x Ge x of the fins to form the thermal oxide, and further comprising removing the thermal oxide such that individual ones of the fins of the second subset of fins have a second width that is less than the first width.

15. The method of claim 9 , further comprising masking the first subset of fins prior to performing the step of annealing.