Strain-Direct-On-Insulator Substrate Fabrication via Layer Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for forming Strain-Direct-On-Insulator (SDOI) substrates are complex, costly, and time-consuming due to the requirement of two semiconductor wafers for producing a single SDOI wafer, with the formation of a thick composition-graded SiGe layer being particularly cumbersome.
Innovation Solution
A method involving the formation of alternating stack layers of stress and strain materials on a buffer layer, where a portion of these layers is transferred to a second substrate through wafer bonding and removal, allowing for the production of n−1 SDOI wafers using n substrates, thereby reducing the need for multiple wafers and simplifying the process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If two semiconductor wafers are used to form a single SSDOI wafer through bonding and separation, then the strained silicon layer can be formed with proper lattice matching, but the production complexity and cost increase significantly
Solution Approach 1:
The invention segments the production process by forming multiple alternating stacks of relaxed SiGe layers and strained silicon layers on a single donor wafer before bonding. This allows the complex structure to be built in modular units that can be systematically transferred to multiple substrates, reducing overall production complexity while maintaining lattice matching precision.
Solution Approach 2:
The invention performs preliminary actions by pre-forming the complete alternating stack structure with all relaxed SiGe and strained silicon layers on the donor wafer before the bonding step. This preliminary formation of the entire multi-layer structure eliminates the need to form these layers individually on each substrate after bonding, significantly simplifying the production process.
2Manufacturing precision
If a thick composition-graded SiGe buffer layer is formed for each SSDOI wafer, then the lattice constant can be gradually adjusted to achieve relaxed SiGe layer, but the production time and cost increase
Solution Approach 1:
The invention merges multiple production cycles into one by forming alternating stacks of relaxed SiGe and strained silicon layers on a single donor wafer. This combined structure can then be used to produce multiple SSDOI wafers simultaneously through sequential bonding and separation operations, dramatically improving productivity while maintaining the necessary lattice constant gradient control in the buffer layer.
Solution Approach 2:
The invention discards the need to form thick composition-graded buffer layers on each individual substrate by recovering and reusing the pre-formed alternating stacks from the donor wafer. The buffer layer is formed once on the donor wafer, and the entire stack structure is transferred to multiple substrates, eliminating redundant buffer layer formation steps.
3Productivity
If multiple alternating stacks of stress and strain layers are formed on a single donor wafer, then n-1 SSDOI wafers can be produced from n substrates, but the initial fabrication complexity increases
Solution Approach 1:
The invention segments the final SSDOI product into multiple alternating stacks that are pre-formed on the donor wafer. Each stack contains the complete sequence of relaxed SiGe and strained silicon layers needed for one SSDOI wafer. This segmentation allows systematic transfer and separation operations to produce multiple finished wafers from a single donor, improving efficiency despite the complex initial fabrication.
Solution Approach 2:
The invention performs the complex alternating stack formation as a preliminary action on the donor wafer before the bonding and separation steps. By completing the entire multi-layer stack formation in advance, the subsequent operations become simpler and more efficient, allowing n-1 SSDOI wafers to be produced from n substrates through standardized transfer procedures.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces production time and cost by enabling the fabrication of multiple SDOI wafers using a single donor substrate, improving throughput and eliminating the need for a thick composition-graded SiGe buffer layer in each production cycle.
Implementation Method 1
Silicon has a natural lattice constant of approximately 5.43 Angstroms. Silicon germanium has a natural lattice constant of between 5.43 and 5.66 Angstroms, depending upon the concentration of germanium in the SiGe. The higher the concentration of germanium, the larger the natural lattice constant of the SiGe. With the natural lattice constant of SiGe is higher than that of silicon, the entire silicon epilayer will be in a state of biaxial tensile stress.
Implementation Method 2
A biaxially strained epitaxial layer of the overlying semiconductor material may thereby be formed. This layer allows the device channel region to experience increased carrier mobility which increases device performance.
Implementation Method 3
an underlying thick virtual substrate with a gradual increase or increment of Ge concentration is needed before forming the relaxed SiGe layer
Implementation Method 4
The wafers are bonded followed by separation at a predetermined location (the relax SiGe layer) to form one SSDOI wafer
Data Source
AI summary
Methods (and semiconductor substrates produced therefrom) of fabricating (n−1) SDOI substrates using n wafers is described. A donor substrate (e.g., silicon) includes a buffer layer (e.g., SiGe) and a plurality of multi-layer stacks formed thereon having alternating stress (e.g., relaxed SiGe) and strain (e.g., silicon) layers. An insulator is disposed adjacent an outermost strained silicon layer. The outermost strained silicon layer and underlying relaxed SiGe layer is transferred to a handle substrate by conventional or known bonding and separation methods. The handle substrate is processed to remove the relaxed SiGe layer thereby producing an SDOI substrate for further use. The remaining donor substrate is processed to remove one or more layers to expose another strained silicon layer. Various processing steps are repeated to produce another SDOI substrate as well as a remaining donor substrate, and the steps may be repeated to produce n−1 SDOI substrates.