1. Wieman: 1 LBNL Micro-vertex STAR Collaboration Meeting Aug 2003 LBNL Howard Wieman, Fred Bieser, Robin Gareus (Heidelberg), Howard Matis, Marcus Oldenburg, Gulshan Rai, Fabrice Retiere, Kai Schweda, Hans- Georg Ritter, Eugene Yamamoto UCI Yandong Chen, Stuart Kleinfelder BNL Instrumentation Div Consulting OSU Ivan Kotov Purdue Dennis Reichhold

2. Wieman: 2 LBNL Today's topics Monolithic APS CMOS, new development of photo-gate –Purpose –Modeling –First silicon results Mechanical work –Vibration measurements –Light weight silicon support, another approach –Thin beam pipes

3. Wieman: 3 LBNL CDS removal of fixed pattern noise and KTC reset noise on the chip (standard diode requires CDS off chip) Increase signal by reducing signal spreading to adjacent pixels. The photo gate permits large geometry without adding capacitance to the sense node. Photo gate purpose - addresses standard diode limitations P- P P+ Standard diode geometry Standard APS diode structure

4. Wieman: 4 LBNL photo gate source follower gate reset gate transfer gate row select gate sense node drain Large photo-gate to collect large fraction of the charge on a single pixel, directly on the p- epi layer Small transfer gate also directly on p- epi layer Small drain (minimum capacitance) connected to source follower gate (sense node) Photo-gate geometry 20 m x -2 m- 1 m 5 nm 8 m 0.1 m 0.4 m P epi 1.4x10 15 1/cm 3 N+ 1x10 20 1/cm 3 photo gate transfer gate drain x = 0.4 and 0.8 m (simulation quantities)

5. Wieman: 5 LBNL Photo-gate issues in standard CMOS No double poly process – possible poor transfer between gates because of low transverse field Floating n well between gates, a bad solution to the transfer problem with single poly Sub-micron process may solve problem single poly, limitation of CMOS double poly, standard for CCDs photo gate transfer gate drain floating n well

6. Wieman: 6 LBNL Photo gate/transfer gate operation photo gate trans gate 1.8 V V drain 2.4 V photo gate trans gate drain V phg = 0.8 V transfer mode V phg = 2.4 V collection mode

7. Wieman: 7 LBNL Photo gate/transfer gate with 800 nm separation photo gate trans gate 1.8 V V drain 2.4 V photo gate trans gate drain V phg = 0.8 V transfer mode V phg = 2.4 V collection mode

8. Wieman: 8 LBNL Drain current after light injection 400 nm between photo gate and transfer gate, no floating n well Nano amp drain current Rapid electron transfer - complete in 60 ns photo gate transfer gate drain Light injection 60 ns

9. Wieman: 9 LBNL First silicon tests, comparing photo-gate with standard diode structure Photo-gate directly to sense node drain DC bias: V photo-gate 0.6 V V drain 2.4 V Output signal for Fe 55 X-ray test Issue: ADC Linear ADC Log diode Photo-gate 1 Photo-gate 2 diode - full charge collection Why is the signal spread out – is it surface traps under the gate?

10. Wieman: 10 LBNL Mechanical Concept Single end support for rapid installation and removal –For beampipe bake out –Insurance for beam excursion damage Readout electronics in end support module Low Mass Carbon fiber tube Thin silicon ladders under tension Aluminum Kapton cables under tension

11. Wieman: 11 LBNL Tension concept

12. Wieman: 12 LBNL Tension concept

13. Wieman: 13 LBNL 50 m Silicon Thinned and polished wafers, a standard industrial process Thinned Silicon wire bonded to cable, both supported under tension Used in wind tunnel test

14. Wieman: 14 LBNL Air cooling and vibration 1-2 m/s air cools 100 mW/cm 2 TV holography shows 2 m vibration for tensioned silicon structure TV holography vibration map Lucite wind tunnel Thin silicon ladder, tension support Laser light source Camera

15. Wieman: 15 LBNL Thin stiff ladder concept Under construction Will be tested for vibration in our wind tunnel carbon composite (75 m) aluminum kapton cable (100 m) silicon chips (50 m) 21.6 mm 254 mm

16. Wieman: 16 LBNL Self supporting ladder concept Outer shell carries extended beam pipe gravitational loads Free end access for detector insertion

17. Wieman: 17 LBNL A thinner beam pipe Elastic Bucking Instability Euler 1757 Eigen solution Critical buckling force The challenge: increase moment of inertia along y without increasing material 16 cm 4 cm Aluminum 150 m thick critical pressure 5.5 atmospheres Increase I with corrugations 0.03 atmos. with no corrugation

18. Wieman: 18 LBNL Conclusion Present monolithic CMOS APS detector technology suitable for slow 10-20 ms readout (examples at LBNL/UCI and LEPSI/IRES) Fast readout not ready yet, but progress is being made Radiation hardness good enough for RHIC Mechanical concepts progressing R&D is fun, but… Proposal due this December

19. Wieman: 19 LBNL Test of diode variation No Fe55 signal Will test with more statistics p- epi n p p Puzzle:

20. Wieman: 20 LBNL Rejection of primary tracks For large rejection ratios must set cut at several times multiple scattering angle At these large angles single coulomb scattering dominates and materials contribute linearly Beam pipe x/X0 =.17 % Si x/X0 =.05 % C Comp x/X0 =.03 % Al Kapton x/X0 =.03 %

21. Wieman: 21 LBNL Photo-gate Fe 55 X-Ray test Charge collection histogram for CMOS APS – comparing two photo-gate structures to standard diode structure Photo-gate coupled directly to drain – no transfer gate Photo-gate shows more complete charge collection, but… ADC LogLinear diode Photo-gate 1 Photo-gate 2 full diode charge collection

22. Wieman: 22 LBNL Drain current after light injection, floating n well delay V photo gate transfer gate drain floating n well Light injection Potential before and after injection V=Q/C (very small) drain current log drain current linear 200 s Smaller the signal the higher the effective resistance and the slower the transfer

23. Wieman: 23 LBNL Radiation resistance Radiation tests of the LBNL APS at the 88 cyclotron 55 MeV protons 30 year RHIC operation at 4 x design luminosity

24. Wieman: 24 LBNL Properties, LBNL APS MIP (most probable)440 e Node C, measured with Fe 55 Xray6.1 fF Gain ~26 V/e Noise ( 1 pixel, CDS, I leak subtr )17 e rms Signal/Noise (9 pixel sum, CDS)9 Signal/Noise (potential, single pix)26 kTC reset noise (measured)50 e rms kTC reset noise (expected)30 e rms Leakage Current0.9 fA

25. Wieman: 25 LBNL Silicon Cost Ladders per Wafer5 Ladders per Detector24 Yield60% Number of Detector Copies4 Number of Wafers32 Wafer Cost Each2-5 k$ Wafer Costs64-160 k$ Mask Cost150-200 k$ Total214-360 k$ 8 inch wafers 20 mm x 170 mm ladders