P ixel Detector Module using MC M-D Technology for the B -layer of the AT LAS P ixel Detector DETECTOR CHIP Pixel 2000 Workshop O. Bäsken K.H.Becks P.Gerlach Ch. Grah O.Ehrmann M.Töpper J.Wolf Christian Grah grah@ whep.uni-wuppertal.de Univers ity of Wuppertal www.atlas.uni-wuppertal.de June 2000, Genova
Overview T he concept of building modules in MC M-D technology MC M-D modules for the ATLAS P ixel Detector Measurements on P rototypes Lab measurements on full scale module and s ingle chip devices Testbeam measurements on single chip devices Conclusion and Outlook 2
ATLAS Pixel Detector 2200 modules 2.2 m 2 active Si 1 x 10 8 channels 3
The basic structure of modules for the AT LAS P ixel Detector S ensor tile (16.4 mm x 60.4 mm active area) Main components which need to be contacted: 16 read out IC s, each providing 18 x 160 pixel unit cells (preamplifier, discriminator, digital readout; pixel cell size: 400 x 50 µm 2 ) 46080 c onnec tions in the pixel c ell array per module with bumpbonding and flip-chipping as interconnection technique one module controller chip Idea of us ing a T hin F ilm technology to perform the s ignal interconnections and power dis tribution on the active s ens or 4
MC M-D, a Thin Film Technology Multi C hip Module Deposited conductor layers Up to 5 copper layers : magnetron sputtered up to 2 µm Ti/Cu/Ti 10 mω/❷ additive electroplating up to 5 µm T i/c u Minimal width and spacing 10 and 20 µm F inal metallisation: electroless 5µm Ni:P/ 200nm Au dielectric layers Spin-on polymer: B C B (Benzocyclobutene / DOW:CYCLOTE NE ) Photosensitive S pecific dielectric constant ε r = 2.7 Process temperatures : 1h 220 C per layer las t layer 1h 250 C T hicknes s / layer 4-10 µm Via >20 µm, P ad 30µm 5
MC M-D Module Pixeldetector Module Module Controller Chip 16 Frontend Readout Chips Temperature Sensor Opto Package (VCSEL & PIN diode) with 3 optical fibres 78,6 mm DORICp VDCp Connections to outer world 24,4 mm Sensor 6
Advantages of modules in MC M-D technology A robust, easy-to-handle module with bump-bonding as the only interconnection technique S ignal lines in µ-s trip configuration, s o with low cros s talk and well defined impedance Allows routing in the pixel cell array to contact s ens or and electronic cells which are not facing each other 7
Schematic Cross-Section of a Bus System 8
S ome pictures of the MC M-D s tructures C h.g rah DETECTOR CHIP University of W uppertal F eed-throughs 50 µm s ignal bus P ixel 2000, G enova power contact 9
F eas ibility S tudies Voltage [V] -260-240 -220-200 -180-160 -140-120 -100-80 -60-40 -20 0 20 1 IZM01 t1-01 IZM09 t1-01 IZM01 t1-01 (after processing) IZM09 t1-01 (after processing) 0-1 -2-3 -4-5 -6-7 -8-9 -10-11 -12-13 Current [µa] Just two exemplary plots The sensor properties are not affected by the MC M-D technology 10
Yield Test - Thin Film Feed-through structures Dais y-c hain interconnection F our copper layers 1.1. 10 6 monitored vias with a diameter of 25µm Measured defect rate 8.13. 10-6 (9 defects of 1 105 920 vias) BCB etched for better visualisation We expect 1.5 unconnected pixel/module 11
DETECTOR CHIP F ull S cale P rototype Module C h.g rah University of W uppertal F rontend C hips MC C Additional tes t pads contacted by wire bonding P ixel 2000, G enova 12
T hres hold and Nois e (Untuned F ull S cale Module) The MC M-D Module shows encouraging performance regarding Threshold dis tribution and Nois e performance Module: MC M-D T1/Frontend B 13
DETECTOR CHIP S ingle C hip Module C h.g rah University of W uppertal A S ingle C hip Module cons is ts of: S ens or cell array + MC M-D interconnections + F rontend chip Inves tigation of different F eed-through layouts, es pecially routing P icture: F rontend C on S ingle C hip P C B P ixel 2000, G enova 14
Feed-throughs in different Class U 400/600 (two columns at the border of the hybrid) layouts Class R 1 (to neighbouring pixel cell) Class R 2 (skipping one cell) Class R 3 (s kipping two cells ) Class U (most common class) 15
T hres hold dis tribution (S ingle C hip) 2000 1600 standard settings: all TDAC s = 4 threshold adjusted to 825 [e - ] 400 adjusted threshold: mean value: 825 [e - ] standard deviation: 45 [e - ] 300 Threshold [e - ] 1200 800 entries/bin 200 not adjusted threshold: mean value: 825 [e - ] 100 standard deviation: 225 [e - ] 400 0 320 640 960 1280 1600 1920 2240 2560 2880 Pixel number 0 400 600 800 1000 1200 1400 1600 Threshold [e - ] Hybrid: MC M-D S T1/Frontend C 16
Noise distribution (S ingle C hip) 350 600 300 500 250 400 200 Noise [e - ] 150 entries/bin 300 mean value: 86 [e - ] standard deviation: 11 [e - ] 100 200 50 100 0 0 320 640 960 1280 1600 1920 2240 2560 2880 Pixel number 0 20 40 60 80 100 120 140 160 180 200 Noise [e - ] Hybrid: MC M-D S T1/Frontend C 17
Summary of Noise measurements hybrid MCMD FeC- S t1 MCMD FeC- S t2 class mean value ± standard deviation [ e - ] U 400 79 ± 10 130 ± 12 U 600 88 ± 12 n/a U 80 ± 10 126 ± 12 R1 93 ± 10 136 ± 12 R2 96 ± 8 142 ± 13 R3 94 ± 8 151 ± 15 T here is no influence on the performance, due to F eed-throughs in MC M-D. As expected, the crossing of copper lines in different layers (classes R i) increas es the Nois e, due to the higher interpixel capacitance. 18
Crosstalk Measurements Crosstalk = fraction of charge that couples into the neighbouring pixel through the interpixel capacitance Q hits Pixel N (masked to read out) Pixel N+1 (with threshold T) Crosstalk = T / Q F or P ixel N+i s imilar 19
Crosstalk distribution (S ingle C hip) 25 Ri U 600 20 Crosstalk [%] 15 10 5 to neighbouring pixel skipping one pixel skipping two pixel ganged Pixel: These electronic cells are connected to two s ens or cells (by design). 0 0 320 640 960 1280 1600 1920 2240 2560 2880 Pixel number 20
S ummary of crosstalk measurements hybrid to pixel N + 1 FeC- St1 N + 2 N + 3 N + 1 FeC-St2 N + 2 N + 3 class mean value ± standard deviation [%] U 400 6,9 ± 0,3 2,6 ± 0,2 <1 2,8 ± 0,8 1,6 ± 0,4 1,1 ± 0,2 U 600 8,6 ± 0,5 3,9 ± 0,2 < 1 n/a n/a n/a U 7,0 ± 0,3 2,5 ± 0,1 < 1 3,0 ± 0,7 1,3 ± 0,3 < 1 R1 8,2 ± 0,5 2,7 ± 0,1 < 1 4,6 ± 0,9 1,6 ± 0,3 <1 R2 8,9 ± 0,5 4,3 ± 0,2 < 1 5,1 ± 1,1 2,4 ± 0,4 <1 R3 8,6 ± 0,4 5,7 ± 0,2 2,8 ± 0,1 5,0 ± 0,8 3,7 ± 0,5 2,0 ± 0,3 Note 1: There is no influence on the crosstalk, due to the Feed-throughs in MCM-D. Note 2: The performance of class R 1 and R 2 layouts is comparable to the 600µm long sensor cells (U 600 ). 21
S ource measurement Upper 3 cells not connected (by design) The MC M-D hybrid shows a uniform functionality. Defects were recognized as bad bump connections. nr of hits Am 241 : Gamma-rays 22
Testbeam data H8 Testbeam at SPS (CE R N) primary: 450 GeV protons Data was mainly taken with: 180 GeV pions T eles cope with 4 layers of strip-detectors (R esolution: 3 µm) H8 Telescope system All presented measurements: (MCM-D) S S G/Frontend B 23
R econs tructed energy depos ition 2000 1750 1500 1250 1000 750 500 250 0 6000 5000 4000 3000 2000 1000 0 0 10 20 30 40 50 60 70 80 90 100 pulseheight (Ke) C onventional hybrid MC M-D hybrid 0 20 40 60 80 100 MCM-D pulseheight (Ke-) S ingle hit events Double hit events (added charges) No charge loss can be seen, due to the MC M-D structures 24
S ingle hit resolution 600 500 C onventional hybrid 58.66 / 53 P1 0.2662E-03 P2 0.5272E-02 P3 0.1938E-01 P4 21.49 MC M-D hybrid 1000 800 69.20 / 51 P1-0.5997E-04 P2 0.5274E-02 P3 0.1880E-01 P4 39.12 400 600 300 400 200 200 100 0-0.1-0.08-0.06-0.04-0.02 0 0.02 0.04 0.06 0.08 0.1 0-0.1-0.05 0 0.05 0.1 MCM-D resolution: single hits resolution: single hits Difference between predicted (Telescope) and measured particle track P2: sigma of gaussian tail P3: width of plateau 25
Double hit res olution 500 C onventional hybrid 80.18 / 35 Constant 516.7 Mean -0.2504E-03 Sigma 0.5269E-02 MC M-D hybrid 1200 1000 171.9 / 57 Constant 1109. Mean 0.1694E-03 Sigma 0.5137E-02 400 800 300 600 200 400 100 200 0-0.1-0.05 0 0.05 0.1 resolution: double hits analog 0-0.1-0.05 0 0.05 0.1 MCM-D resolution: double hits analog Double hit resolution: 5µm (conventional and MC M-D hybrids) 26
Multi C hip Module-Depos ited Conclusion It is poss ible to build easy-to-handle Pixel Detector Modules with the MC M-D technique. The S ensor is not harmed / damaged by the processing. The signal and power distribution structures are able to drive full modules. No problems appeared due to the necessary connections between electronic and sensor cells. Outlook: E xplore the full potential of the MC M-D technique, modules with a homogeneous resolution may be build. 27
F urther pos s ibilities of the MC M-D technology T he pos s ibility of integrating pas s ive components in MC M-D is under inves tigation. R and C: C urrently pos s ible (due to the high proces s temperature this is not (yet) pos s ible for our application!): 720 pf/mm 2 with T a 2 O 5 as dielectric 10-100 Ω/❷ with TaN as resistor material Inductor in MCM-D Technology 28