Quality Assurance for the ATLAS Pixel Sensor 1st Workshop on Quality Assurance Issues in Silicon Detectors J. M. Klaiber-Lodewigs (Univ. Dortmund) for the ATLAS pixel collaboration Contents: - role of the pixel sensor - overall QA concept - measurement of bulk parameters - sensor breakdown - time stability - sensor depletion - radiation hardness - cross calibration
Introduction large number of detector parts (1718 modules fitted with one sensor tile and 16 front-end chips each) parts not easily accessible after assembly (central position, cooling and radiation) every bad pixel degrades performance Why systematic quality assurance for the ATLAS pixel detector?» 1.1 10 8 pixel channels in total 2
Detector production Sensors Dortmund Udine, Prague New Mex. Electon. LBL Marseille Bonn MCC Genoa Flex Oklahoma Siegen Vecsel Wuppertal tests => t t t t Ohio, Taipei t Bonn Bumps Genoa Bumps passive Packaging t Marseille components t Okla., LBL t t Cable LBL Wuppertal Bonn Genoa Marseille Dicing Flip-chip t LBL Wuppertal Prague Glue flex Wire bonding Cable mounting Components mounting t LBL,NM,Albany,Marseille t Assembly and testing of detector modules 3
Development, production and QM existing data Technical requirements Research & prototyp. Interdependence of main steps TDR Final prototyp. & design FDR/ PRR Tender & preseries Main product. Assembly technical specs acceptance tests QM Identify relevant qualities Define tests & structures fix criteria cross calib. Data base QA protocols 4
Pixel sensor design requirements pixel size 50µm x 400µm 50 µm pitch 12µm diameter bump connection total active area 1.8m 2 (1718 modules) high yield testability 10 years operation fault tolerance harsh radiation environment up to 10 15 MeV n eq./cm 2 fluence and 500 kgy ion. dose 5
Development Strategy Design studies - performed within ATLAS - prototype sensors concerning - isolation technique - design of the pixel cell Studies on silicon - performed within ROSE - various Si impurities concerning - damage parameters - fabrication process Radiation tolerant sensors 6
Sensor Concept n + -in-n pixel oxygenated Si substrate moderated p-spray isolation bias grid for testability 3 sensor tiles per wafer various test and monitor structures 2 nd prototyping for yield optimization Photo of prototype 2 wafer 7
Quality Assurance Quality test plan for sensor production: I-V characteristics on every sensor before bonding, depletion measurement on every wafer process parameters on special test structures (e.g. p-spray dose) both at vendor and at institutes (acceptance tests) for better control 8
Flow of acceptance testing Main measurement steps With tiles wafer & process data tile data tile meas. wafer data wafer meas. Sensor selection - preliminary acceptance bumping & dicing to flipchipping Final acceptance irrad. meas. irrad. data irradiation Data to DB diced meas. add. data Only test structures 9
Tests to perform Grouped in topical order measurements tile quality tests visual inspection, I-V testing, I-t testing wafer quality tests tests after dicing tests after irradiation visual insp., I-V on sensors, diodes and MOS, C-V on diodes and MOS, I-V gate on GCD and MOSFET, planarity meas. thickness meas., I-V gate on GCD, interpixel measurements, implantation meas. I-V on sensors and diodes, C-V on diodes, I-V gate on GCD and MOSFET qualities tested visible damage, breakdown voltage, time stability alignment, depletion voltage, oxide characteristics, p-spray dose, planarity thickness, bump bonding effects, interpixel resistance and capacitance, sheet resistances radiation hardness 10
Testing responsibilities vendor provides process data tests pixel quality on sensor tiles on wafer level ATLAS institutes check process data against measurements test pixel quality on sensors tiles, single chips and mini chips performs diagnostic tests on wafer level for depletion, oxide quality and capacitance, and p- spray dose perform all diagnostic tests on wafer level and on diced test structures measure test structures after irradiating them 11
Breakdown & leakage current I-V tests before bonding using bias grid punch-through effect across a bias grid allows testing of all pixels using only two probes on wafer p-side ground p-side bias guard ring punch-through 12
Testability of pixel quality 1.00E+02 1.00E+01 I-V tests on test pixels using punch-through 1.00E+05 defective pixel matrix with defective pixels 1.00E+04 1.00E+00 1.00E+03 current [na] 1.00E-01 1.00E-02 1.00E-03 good pixel current [na] 1.00E+02 1.00E+01 1.00E+00 good pixel matrix 1.00E-04 0 20 40 60 80 100 120 140 160 180 200 bias voltage [V] current through single pixel current through punch-through array Leakage current indicative for quality of every pixel 1.00E-01 0 20 40 60 80 100 120 140 160 180 200 bias voltage [V] 13
Breakdown & leakage current I-V measurements of leakage current show pixel quality Tile classification by pixel quality breakdown voltage indicates type of defect tile classification possible taking in account current slope and operation voltage 14
Time stability I-t tests on tiles show if leakage current increases significantly over 15h I-t tests at operation voltage similar tests could be done on mini sensor chips after irradiation long time burn-in could be done after assembly on module level Measured at 150V bias 15
Sensor depletion Diagnostic measurement by diode capacitance guard ring test diodes on production wafer for well defined capacitance measurements p + n + defined n-bulk area full depletion visible by levelling out of C vs. V -1/2 curve (suppression of possible constant stray capacitances) 16
Diode depletion vs. Sensor depletion In unirradiated n + -in-n sensors, two depletion cases are reached pinch-off between neighbouring pixels by growing depletion zones from p-side and p-spray full depletion of whole sensor volume (as in diode) Near full sensitivity at lower voltage 3D-simulation of e - -concentration Double depletion in single chip sensor 17
Sensor depletion after irradiation Oxygenated Si shows slower raise in depletion voltage: less damage by charged particles (e.g.p ± ) saturation of reverse annealing at higher fluences ATLAS scenario for 100 d beam, 3 d at 20 C, 14 d at 17 C p.a.(b-layer) 18
Radiation hardness tests Oxygenation testing after irrad. with 3.1 10 14 MeV n eq./cm 2 protons Bulk damage testing after irrad. with 10 15 MeV n eq./cm 2 protons (design fluence) Surface damage testing after irrad. with 500 kgray low energetic electrons (design dose) depletion measurement on diode I-V measurements on mini chip and diode (small structures) interface generation current measurement on GCD p-spray measurement on MOSFET 19
Testing for irradiation environment Irradiation tests show so far: irradiation with p of different energies and p + show comparable results irradiation in CO 2 and N 2 atmosphere cause no additional damage CO 2 N 2 N 2 Measured at -30 C, normalized to 0 C 20
Calibration of measurement sites Repeated tests on same dedicated structure: show stability of measurements at participating labs (influence of time or environment) can be used to test readiness for new testing period I-V and C-V tests are used 16 I-V tests on diode 16 tests 21
Cross calibration of measurement sites Tests on same structures at different labs: show comparability between labs indicate possible problem sources (handling, transport, humidity, set-up) All tests used Currently ongoing 5 I-V tests on sensor tile 22
Summary QM for the ATLAS pixel sensor has been a dynamic process between different aspects of the project requires close collaboration for a long time period QA procedures have required good understanding of the phenomena to be measured decisions on what exactly is crucial or worthwhile to know QC measurements have, until now given us good tools for acceptance decisions ensured compatibility between different measurements / scenario and test 23