1. 2.5D/3D Chip Stacking Supply
Chain Integration
Kurt Huang, Ph.D.
Director, Corporate Marketing, UMC
ChipEx, 2013

2. Outline
 2.5D/3D Applications
 Ecosystem Work Flow
 Summary
P. 2

3. 2.5D/3D Applications

4. 2.5D Si Interposer Stacking
 Logic/logic: FPGA, networking infrastructure
 Logic/memory: Gaming, HPC

5. 3D Logic/Memory Stacking
- Via-Middle TSV 28nm Logic + Memory Cube
 Mobile WideIO, Computing WideIO, HMC

6. Never-Enough Memory Bandwidth
- Mobile AP Driving Wide IO Memory
Dimensioning use case:
 3D video streaming playback to external display via wireless
+ on-line 3D gaming local

7. Keeping Power Under Hood
•LPDDR3 @800Mhz
dual-channel
•LPDDR3 @800Mhz

8. Real Value Drives Adoption
Motivations:
 Higher BW, lower W/BW, smaller form-factor
Opportunity of return on 3D IC investment:
 Chip process node optimization
 Homogeneous partition
 Cross-node combinations
 BOM cost optimization
 Less demanding substrate/PCB,
lighter cooling assembly, ...
 Ultimately: better product, better margin
Xilinx Virtex 7
Micron HMC

9. Application Examples

10. Market Segment Overview
 High performance driving
2.5DIC
 Mobile AP driving 3DIC
 SoC + Mobile Memory
 Next wave:
heterogeneous 3DIC
 Logic + Memory +
Analog/RF
 Ecosystem more
complex
 Expect driving force
emerges after 3DIC
matures

11. Ecosystem Work Flow

12. Example 2.5D Stacking Flow

13. Various Work Models
 Service scopes distinguished by MEOL inclusion
 Consult your foundry/OSAT
 Work flow optimization may depend on BOM cost,
stack recipe and test strategy

14. Foundry TSV Design Collaterals
 Consider TSV a passive device with rule decks/models
 Typical foundry engagement applies under ecosystem work flow
(UMC 2.5D Si interposer documents)

15. Innovations by Open Eco-System
 Wafer thinning and handling
 Thermal/stress
consideration
 Testability
 Reliability
 3D EDA tool
 Seamless business model
 Cost
 …

16. Evolution Of the Supply Chain
 Technology exploration
 Feasibility study
 Interface definition
 Handover criteria
 Model convergence
 Flow standarization
 Cost down
 Service differentiation
 Further innovation

17. UMC Ecosystem Effort
1Q1
2
2Q1
2
1Q1
3
2Q1
3

18. Summary

19. Summary
 Foundry TSV process demonstrated
 Applicable to both 2.5D/3D
 Leverage existing CMOS process technology
 Key process issues identified & conquered
 Ecosystem work flow
 Typical foundry/OSAT engagement flow applies for both
2.5D/3D, among other models
 Foundry TSV next step: ecosystem focus
 Product level reliability assessment
 3D package level reliability demonstrated for open
ecosystem model
 Potential EDA collaboration for emerging 3D tools

20. BTW, Dreams Do Come True!

21. Thank you for your attention!
Contact: kurt_huang@umc.com