Extending Optical Lithography with C E B L (Complementary E-Beam Lithography) July 13, 2011 4008 Burton Drive, Santa Clara, CA 95054
Outline Complementary Lithography E-Beam Complements Optical Multibeam CEBL Approach Infrastructure for CEBL Conclusion 2
Optical Extension Makes Sense 193nm ArF (193/193i) is a growth engine Mature technology, equipment, infrastructure Advanced materials, modeling, mask making, metrology Accumulated knowledge; amortized scanners Resolution limit driving high cost for critical layers For technical and economic reasons, 193i should be extended with another lithography technology 3
Complementary Lithography 193-nm immersion could work hand-in-hand with EUV or maskless lithography to enable advanced chip designs. Borodovsky, Y., Nikon LithoVision (2010) Cut Masks: Optical: 4 EUV: 1 EBL: 0 Borodovsky, Y., SEMATECH Maskless Workshop (2010) CEBL is complementary use of EBL for Critical Layers: Cutting poly and metal lines Cutting contact and via holes 4
1D Gridded Layout: Scaling Devices IBM, GlobalFoundries, and others working on DFM Ideally, the gates would be vertical on a given pitch... and all the metal lines unidirectional. Liebmann, L. EDPS (2009) TSMC announced 65nm Slim Library with 1D Layout unidirectional poly on a fixed pitch TSMC press release (2010) Intel making 1D-layout Logic devices since 2007 Intel 65nm node (2005) Intel 45nm node (2007) Intel 32nm node (2009) Intel 22nm node (2011) 5
193i + Pitch Division: Tighter Densities Yaegashi (TEL), Litho Extensions Symposium (2010) 6
CEBL Line Cuts: Critical Layer 1D Lines Resist CEBL Patterning Etch & Strip From design to silicon 1D Layout in Design File Line Pattern Cut Pattern 193i with Pitch Division CEBL 1D Layout on Wafer Same cuts for vias and contacts 7
Multibeam CEBL Architecture Multi-column scalable architecture Leveraging SEM; eliminating magnetic field Column optimized for cutting lines and holes Columns small, fast, robust, low cost, arrayable Each module: 5 wph any wafer size incl. 450 mm Multiple modules clustered for high throughput Low volume High product mix Multi-column module Single-module system top view 8
Multibeam CEBL Scanning Vector scan shaped beam Intel 32nm node (2009) Cut-pattern density is low: ~5% Beam deflected to each cut One shot per cut One simple shape Reduce overhead, speed cutting 9
Multibeam CEBL Alignment Global alignment: off-axis, as in optical scanners Local alignment: in situ, in-process, with e-beam Each column equipped with built-in SEM Column-SEM to image alignment marks on wafer Repeat registration to improve Overlay and CDU E-Beam E-Detector in Column-SEM Wafer 10
Beam Energy (kev) Wafer Heating Multibeam CEBL Simulations Beam Profile Beam energy from <5keV to >50keV Large window to operate beam Beam Operating Window 10nm features; 50nm resist 12nm FWHM 50 40 30 20 Low Thruput Optimum Window e-e Repulsion G4 column: 10nm - 20nm features Resolution 10nm achievable D. E. Liu, et al. (Multibeam), SPIE 2011 10 0 0 Electron-Resist Interactions 10 20 30 40 50 Beam Current / Column (na) 11
Support Infrastructure 1D Gridded Layout Mask Writers Mask Shops Yield Enhancement E-Beam Inspection CEBL Ecosystem Line Patterning Line Pitch Division E-Beam Resists CEBL Cutting 12
In Conclusion 13
To Optical Extension With Optical our trusted companion, We made lots of gold from silicon. We now aspire to reach the moon, We ll need CEBL very soon. CEBL is not industry savior, Not even NGL. But CEBL loves critical layers, Optical thinks they re from hell. You know Optical fails in small features. That is a fatal flaw. But CEBL can cut for good measure, And you continue on Moore s Law. So, keep all the infrastructure, And keep your 193i. Yes, you can cut capital expenditure, And enjoy the ROI. CEBL will soon be walking tall, Side-by-side with Optical, Then, no layer will ever at all Try to be cute and critical. 14
Acknowledgement Anchor Semiconductor Gartner SEMI TEL Tela Innovations Multibeam 15