LHC CMS Detector Upgrade Project Summary of the

LHC CMS Detector Upgrade Project Summary of the Si. PM and Packaging Preproduction Review - CMS Phase I Upgrade A. Heering, Y. Musienko, M. Wayne Notre Dame & INR (Moscow) July 22, 2014 1

LHC CMS Detector Upgrade Project Outline ● A bit of background and history ● Si. PM vendor status ● Preproduction review and details of the preproduction order Note: detailed talks can be found at https: //indico. cern. ch/event/325267/ July 22, 2014 2

LHC CMS Detector Upgrade Project Silicon Photomultipliers for the Phase I HCAL upgrade • Over the past decade, Si. PMs have become a well-established technology for HEP • Si. PMs have improved performance compared with the HPDs currently in use in the CMS HCAL • CMS has done years of R&D with manufacturers to optimize device performance for the CMS HCAL • CMS has good experience with ~3000 Si. PM channels for HO • There are two vendors that can provide devices which meet our specifications • The project has developed a detailed schedule for the procurement, testing and installation of Si. PM readout for both the HE and HB detectors • The plan has already passed a DOE CD 1 review and two Fermilab Director’s Reviews July 22, 2014 3

LHC CMS Detector Upgrade Project Vendors • Currently two vendors have produced devices which meet the requirements for use in the CMS HCAL • Hamamatsu, Japan, vendor for HO Si. PMs • KETEK, Germany • R&D has been coordinated with both vendors to tune devices for HCAL requirements • Multiple vendors have been kept in the process until final tests are completed • Vendor cost competition is keeping the prices down • Plan to split the preproduction between Hamamatsu and KETEK, final decision on devices by mid-June, 2014 July 22, 2014 4

LHC CMS Detector Upgrade Project A year ago, Hamamatsu was leading the way Hamamatsu July 22, 2014 KETEK 5

LHC CMS Detector Upgrade Project KETEK R&D to optimize PDE 2013 Wafer section new 3. 3 mm to fit 7 fib New KETEK V 11 design with 70% larger active area than the 2012 V 9 design July 22, 2014 6

LHC CMS Detector Upgrade Project Latest results from KETEK New KETEK V 11 design: at V-VB = 3 volts, now PDE ~30%. Note that there is still room to optimize for wavelengths around 500 nm July 22, 2014 7

LHC CMS Detector Upgrade Project Neutron radiation damage results Hamamatsu, Ketek devices irradiated with 3 x 1012 neutrons/cm 2 (about 14, 000 fb-1 in HB) Devices had ~5 mm 2 area, 15 micron pixels Signal changes < 10% for devices from both vendors Noise for both around 20 -25 photoelectrons RMS for a 50 ns integration time and V-VB – 3. 3 volts July 22, 2014 8

LHC CMS Detector Upgrade Project Comparison after irradiation Comparison of two types of KETEK V 11 design with HPK after 1 E 11 p/cm 2 irradiation. All three are still performant and KETEK competitive with HPK July 22, 2014 9

LHC CMS Detector Upgrade Project Si. PM Specifications Spec Value 15 HPK 15 KETEK-II 15 2. 8 24 24 <90 ~65 ~28 ~43 >2 4. 0 <1000 150 300 60 >25 30 27 33 <700 350 600 420 <600 215 525 330 ≤ 10 10 10 5 <1. 3 1. 18 1. 14 1. 16 Optical Cross-Talk [%] After-pulses [%] <20 17 14 15 <5 <2 <2 <2 d. VB/d. T [m. V/°C] Temperature sensitivity [%/C] <60 58. 5 20. 4 35. 4 <6 3 0. 8 1. 5 Voltage sensitivity [%/V] <60 50 38 40 Neutron noise sensitivity low no no Parameter Cell size [mm] Sens. area [Ømm] Operating temperature [°C] VB [V] Vop-VB (V) Dark Current [n. A] PDE(515 nm) [%] Gain, x 103 Capacitance [p. F] Recovery time [ns] Excess Noise Factor July 22, 2014 10

LHC CMS Detector Upgrade Project Si. PM Specifications (cont. ) Parameter Dark current ( 2*1011 n/cm 2 ) [m. A] ENC (50 ns, 2*1011 n/cm 2 ) [pe] Spec Value HPK KETEK-II <500 140 450 215 <12 8 11 10 Fractional gain. XPDE after 2*1011 n/cm 2[%] Laser resp. , 10 Ohm (Int=90%, ns) Laser resp. , 10 Ohm (Int=95%, ns) Idark(20 °C)/Idark(10 °C) >95 >95 - 30 62 39 - 41 78 50 - 1. 86 1. 66 1. 78 July 22, 2014 11

LHC CMS Detector Upgrade Project Packaging for Si. PMs • Protection for Si. PMs important to avoid damage, humidity effects, etc • Epoxy sealing is not acceptable due to large neutron signals induced • Package design complete, will accommodate 2. 8 mm and 3. 3 mm devices • Thoroughly tested for temperature and humidity effects • Quote from Kyocera, ready to purchase for preproduction July 22, 2014 12

LHC CMS Detector Upgrade Project Preproduction Review on 6 -20 -14 • Why Preproduction? • Provides an opportunity for the vendors to demonstrate that they can produce significant quantities of devices with good performance and high yields. • Provides packaged Si. PM arrays to incorporate in readout modules for beam and bench tests. Important for the final electrical and mechanical designs of the HCAL readout chain. • In order to fully test and understand the preproduction devices, we will develop most, if not all, of the quality control procedures that will be used in for final production. • After fully assessing the preproduction devices, the schedule allows time for design changes before the final production. August 6, 2014 13

LHC CMS Detector Upgrade Project Proposal for Preproduction • Packaging • We have a single vendor – Kyocera. We have a firm quote for preproduction and full production • Preproduction order: 450 ceramic packages, 250 windows, soft tooling • Si. PM • Hamamatsu: We order 175 eight-channel arrays, fully packaged. These are our baseline devices – 15 micron pixel, PDE > 30%, meeting all of our specs • KETEK: First 100 eight-channel arrays already paid for, just need packaging (at vendor cost). We order ~80 arrays from KETEK I with best PDE (V 10 and V 11). ~20 arrays from KETEK II for more detailed testing. • Second order of 75 -100 arrays from KETEK – primarily KETEK II design, and possibly KETEK I with 20 micron pixels for higher PDE. Goals are to improve stability and yield, reduce series resistance and recovery time, and test throughput of new foundry. July 22, 2014 14

LHC CMS Detector Upgrade Project Costs • Packaging • 450 ceramic packages at 101 Euros each 45, 450 EUR • Soft tooling for packages 13, 500 EUR • 250 glass windows at 30 Euros each • Tooling for windows 7, 500 EUR 12, 500 EUR Total is 79 k EUR = $103 k. Budget in the RLS is $103 k. • Si. PM • Hamamatsu: Estimate (firm) is $53 k for 175 arrays, windows included and packaged • KETEK: First 100 eight-channel arrays already paid for by RDMS Quote for second 75 arrays, packaged, is 40 k EUR=$52 k Total is $105 k. Budget in the RLS is $128 k. July 22, 2014 15

LHC CMS Detector Upgrade Project Current Schedule • Packaging • Plan to place order by mid-July • 13 week delivery time • 2 weeks for quality assurance • Packages ready for vendors by end of October 2014 • Si. PM • Place orders for batches 1 and 2 at the end of August 2014 • Batch 1 (100 devices from KETEK) packaged by end of November 2014 • Batch 2 (Hamamatsu and KETEK II) packaged by mid-March 2015 July 22, 2014 16

LHC CMS Detector Upgrade Project Latest news since the review • The NSF Cooperative Agreement that will fund the packaging and Si. PM’s has been approved and funds have been released. • Notre Dame’s procurement office has worked with CERN to set up a plan to make these purchases through CERN. All the agreements have been approved and signed. • The first requisition for the packaging has been created, with ones for the two Si. PM vendors to follow shortly. July 22, 2014 17

LHC CMS Detector Upgrade Project Summary • CMS has worked with Si. PM manufacturers to develop devices optimized for the HCAL readout • Two Si. PM vendors can fabricate devices that meet our specifications • An 8 -channel ceramic package with a glass window has been fully designed and tested • A preproduction plan is in place to fabricate ~3 k channels of Si. PM, package them and then test extensively in a beam and on the bench • The cost and schedule are part of, and consistent with, the overall Phase I upgrade “resource loaded schedule” July 22, 2014 18

LHC CMS Detector Upgrade Project July 22, 2014 Backup 19

LHC CMS Detector Upgrade Project Simulation input • However, HO is a low-occupancy environment with mostly quite small signals. • HB/HE have very high occupancy and a broad dynamic range • Simulation studies were performed to understand the impact of various Si. PM parameters on physics performance. • Key: effective pixel count • Low pixel count results in luminosity-dependent nonlinearity and pedestal shifts July 22, 2014 20

LHC CMS Detector Upgrade Project Recovery time for effective pixel count • Because of the long Y-11 pulse shape, a very fast Si. PM (able to reset its pixels within a few ns) can fire multiple times within the pulse: effective increase in dynamic range • Validated in simulation and on the bench with real Si. PMs Simulation July 22, 2014 21

LHC CMS Detector Upgrade Project “Chronic” Radiation Effects • Chronic damage leads to single-pixel dark currents, which increase noise somewhat • However, with TDR Si. PMs (lower PDE than latest devices), resolution is better for Si. PMs than for HPDs even up to 3000 fb-1 July 22, 2014 22

LHC CMS Detector Upgrade Project Si. PM quality control - 1 2014 -2015 Preproduction run Prepare setups for 8 channel arrays using commercial switching systems 1) Quality control/parameter tests 2) Accelerated aging 3) Radiation testing Set up automated software and database transfer 2016 -2017 Production run 2000 devices from different wafers (out of 25000 devices) used in destructive tests Aging and Radiation tests July 22, 2014 23

LHC CMS Detector Upgrade Project Si. PM quality control - 2 1) Si. PM validation measurements: 3) Si. PM neutron sensitivity tests: (S) - PDE vs. voltage (A) - Am-Be source measurements; - spectral response (S) - measurements at IRRAD-6 (1 Me. V neutrons) - gain vs. voltage (A) - dark current and noise vs. voltage (A) - cell recovery time measurement (A) - dynamic range (S) - temperature coefficient vs. voltage (S) 4) Si. PM radiation hardness tests: (S) - Characterization of Si. PM parameters before, after irradiation - Am-Be source measurements; - capacitance vs. voltage (A) - measurements at IRRAD-6; - cell quenching resistance (A) - measurements using gamma-source (Co-60) 2) 5) Si. PM stability tests: (A) - gain stability tests at constant temperature - PDE stability tests at constant temperature - dark current, noise stability tests at constant temperature Si. PM accelerated aging tests: (S) - characterization of Si. PM parameters before, after accelerated aging/burn in tests; - dark current stability tests at elevated temperature (T=70 C) (A) = all devices July 22, 2014 (S) = subsample 24

LHC CMS Detector Upgrade Project Preproduction 350 arrays Full Production ~3000 arrays total for HE and HB QA HE first, then repeat with HB Si. PM testing and quality control Full Characterization (QA) Database record keeping Final Package evaluation Alignment tolerances QA + tolerance check of Batch 1 ~300 arrays 20 days QA + tolerance check of Batch 2 ~300 arrays 116 arrays longterm/Radiation tests / temp and humidity cycling 240 arrays, build in 32 RMs for Test Beam 2015 / lab B 904 / lab B 28 full system tests Production RMs assembly incl. 20% spares RBX system tests (Noise, Uniformity, Stability Pulse, p. e. /p. s. Resolution) Evaluation and Review Burn-in and installation in UXC 55 8% arrays longterm/Radiation tests / temp and humidity cycling QA + tolerance check of Batch 5 ~300 arrays July 22, 2014 25

LHC CMS Detector Upgrade Project Radiation-damage and Radiation Response • In the CMS radiation field, two primary concerns • Chronic: classic lattice dislocations in silicon causing pixel self-firing => dark current • Acute: large amplitude signals created by neutrons/showers in/near the photodetector • Significant issue for CMS APD used in the ECAL Neutron response for CMS APD during beam spill at CERN neutron IRRAD facility • Theory: Si. PM response should be limited to one or two single-pixel firings (much less than 100 Me. V calorimeter scale) • Intensive R&D program over the last three years to identify high pixel-count devices with good radiation characteristics, establish specifications July 22, 2014 26

LHC CMS Detector Upgrade Project July 22, 2014 HPK 2010 devices (B, C) vs 2012 (barrier layer device, A) 27

LHC CMS Detector Upgrade Project Si. PM Production Schedule and Final Device Selection • Si. PM must meet specifications agreed upon by US-CMS and International CMS: • gain, PDE, recovery time, dark current, … • Vendors must have successfully produced multiple generations of devices with good yield and demonstrated capability to qualify devices during production as required by our QA plan • HE Si. PM review followed by full production order in December 2015 • HB Si. PM review followed by full production order in September 2016 • If multiple vendors qualify, decision will be made based on cost and the S/N for MIPs after neutron irradiation to 3000 /fb • Final device choices for barrel and endcap may be different due to factor-of-six difference in neutron fluence and higher level of radiationdamage in endcap calorimeter • Final device choice for the endcap must be agreed with RDMS (cofunding agency) July 22, 2014 28