S Hou Academia Sinica Taiwan 2010 03 04

S. Hou, Academia Sinica, Taiwan 2010 -03 -04 2010 March 04 -05 Mini Workshop of the Joint ATLAS CMS Opto Working

High speed True. Light VCSELs TSA-8 B 12 -000 TSD-8 B 12 -008 TSD-8 B 12 -051 TMC-5 A 45 -447 2. 5 Gbps 4. 25 Gbps 10 Gbps 4. 25 Gbps 12 -ch array bare die single channel bare die TO-can True. Light Ga. As 850 nm VCSELs TSA-8 B 12 -000 TSD-8 B 12 -008 TSD-8 B 12 -051 TMC-5 A 45 -447 Ith power fc dia m. A m. W/m. A Gbps μm 1. 5 0. 4 2. 5 100 1. 8 0. 35 4. 25 70 1. 2 0. 35 10 90 1. 5 0. 12 4. 25 TO + PD 2

Rad-hard tests for opto-electronics Beam tests at Cyclotron facilities − IPAS Tendem: protron 3 Me. V 90/60 μm stop in Si/Ga. As − Tohoku CYRIC: proton 30, 70 Me. V narrow beam profile − Indianna IUCF: proton 74, 200 Me. V low flux IPAS Tendem IUCF beam area CYRIC control room beam chamber 3

Portable waveform measurement l Pulser 300 MHz, rise 670 ps l VCSEL via MM 62. 5 m fiber to ST to Teflon jig on PIN l PIN to High-Speed adapter (Hamamatsu circuit inside a BNC housing by D. S. Su) l Waveform by a 500 MHz Tek scope Limited to 2 GHz Teflon ST-TO adaptor PIN leads PIN bias 4

Wave form of 2. 5 Gb VCSELs VCSEL driven by pulser, 3 V, 670 ps rise light output read by Ham. S 5973 (1. 5 GHz) 2. 5 Gb oxide (LAr module) reference rise = 600 ps channels in array 0. 8 E 14 (74 Me. V) p/cm 2 rise = 550 ps 1 E 14 (200 Me. V) p/cm 2 rise = 550 ps 5

Wave form of irradiated high speed VCSELs 4. 25 Gb TMC 0. 5 E 14 (200 Me. V) p/cm 2 rise time = 480 ps 4. 25 Gb chip 10 Gb chip 0. 8 E 14 (74 Me. V) p/cm 2 rise time = 530 ps 0. 5 E 14 (74 Me. V) p/cm 2 risetime= 460 ps 6

Portable L-I scan setup l Via a adaptor board (by D. S Su) reading current from a large area PIN l DAQ by NI 6024 E l In use since 2004 7

VCSEL irradiation online, and annealing VCSEL degradation is linear to fluence online recorded by rad-hard fiber readout independent to Flux rate Fast annealing by charge injection operation current (10 n. A) applied L-I of VCSEL (oxide) vs. online Fluence L-I of VCSEL (oxide) vs. Annealing time 8

VCSEL annealing in time Ø Charge injection at the nominal 10 n. A laser current Ø Fit to f(t)= f∞ - a exp( -t/ τ ) recovery time ~ 5 hours uniform, low systematics over channels 9

VCSEL light degradation, 2. 5 Gb array linear L-I fast annealing (~5 hrs) after irradiation linear shift toward higher threshold current less damaged with higher proton energy L/L 0 versus Fluence IUCF 2009 tests @ I=10 m. A 10

True. Light VCSEL 4. 25 Gb TMC Flu 200 Me. V 5. e 13 74 Me. V 8. e 13 L/L 0 τ 0. 93 1. 4 0. 93 5. 1 T 0 T(F) τ 2. 19 2. 43 3. 2 1. 88 2. 28 2. 4 11

True. Light VCSEL 4. 25 Gb chip Flu 200 Me. V 5. e 13 74 Me. V 8. e 13 L/L 0 τ 0. 96 0. 3 0. 94 1. 4 T 0 T(F) τ 2. 19 2. 53 1. 4 2. 49 2. 88 2. 0 12

True. Light VCSEL 10 Gb chip Flu 200 Me. V 5. e 13 74 Me. V 8. e 13 L/L 0 τ 0. 98 0. 3 0. 97 1. 4 T 0 T(F) 0. 84 0. 78 1. 36 1. 43 τ 0 1. 8 13

Summary on high speed VCSELs Ø First, tested at IUCF, 74 Me. V, 200 Me. V True. Light 4. 25 Gb, 10 Gb Ga. As oxide confined, in TO, bare die Ø Too few samples, too low fluences systematics can be improved a lot Ø Compared to arrays lower light power at the first glance similar characteristics and irradiation tolerance needs future high statistics tests 14

High Speed PINs Irradiation at IPAS, IUCF 2009 Hamamatsu in TO cans True. Light in chips Hamamatsu VR Volt Si 9055 -2 Si 9055 -01 -2 Si 5973 -3 Ga. As 8522 -01 -5 Ga. As 8522 -02 -5 Ga. As 8522 -03 -5 True. Light fc diam. GHz μm 1. 5 200 2. 0 100 1. 5 400 3. 0 40 1. 9 80 1. 5 120 I/L * A/W 0. 32 0. 20 0. 53 0. 11 0. 25 0. 40 VR Volt Epi. Si (ATLAS) -10 Ga. As TPD-015 -2 Ga. As TPD-052 -2 015 fc diam. I/L GHz μm A/W >1 100 0. 55* 4. 25 90 0. 65 10 70 0. 65 052 Responsivity by VCSEL 850 nm via 62. 5 μm fiber to PIN face 15

Wave. Form, 3 Me. V proton to Ga. As PIN Ø Hamamatsu PINs Ø In house beam test, 3 Me. V protons waveform taken with a 2. 5 Gb VCSEL Hamamatsu Ga. As 8055 -2 PIN 2 GHz Rise 390 ps (0 rad) rise 420 ps (2 E 14 p/cm 2) 16

Proton damage to PIN Ø Proton damage to PINs reduced P-N depletion, lower responsivity Ø Wave form to damage rise/fall time < 1 ns (20%-80%) Truelight epi-texial PIN for ATLAS 640 ps (0 rad) 670 ps (4 E 14 200 Me. V p/cm 2) 17

True. Light PIN Waveforms, 200 Me. V protons Epi- Si (ATLAS) rise 553 ps 200 Me. V 3 E 13 TPD-015 4. 25 GHz 200 Me. V 7 E 13 rise 455 ps TPD-052 10 GHz 200 Me. V 7 E 13 rise 389 ps 18

Hamamatsu PIN Waveforms, 200 Me. V proton Hamamatsu Si PIN 5973 200 Me. V 7 E 13 9055 200 Me. V 5 E 13 9055 -01 200 Me. V 7 E 13 Rise 557 ps Rise 455 ps Rise 643 ps Hamamatsu Ga. As PIN (need stronger light for current) 8522 -01 74 Me. V 5 E 13 Rise 1. 07 ns 8522 -02 Rise 803 ps 200 Me. V 3 E 13 8522 -03 200 Me. V 3 E 13 Rise 552 ps 19

True. Light Epi- Si PINs Truelight Epi-Si PIN 1. Responsivity drops to 45%, independent of proton energy 2. Dark Curent depend on Proton energy 3. Forward Bias depend on proton energy CYRIC 30, 70 Me. V IPAS 3 Me. V 20

PIN responsivity online Ø Proton beam on/off with LED light on/off Ø do online PIN VR scan Truelight Epi Si PIN (ATLAS) LED on LEDoff beam on, beam off Each strip is a 0 -20 V scan Beam off: + LED on + LED off Beam on: + LED on + LED off Quick recovery expelling dark current 21

True. Light Epi- Si PINs Truelight Epi-Si PIN Measurables: 1. I-V with light (responsivity) 2. I-V w/o light (dark current) 3. Forward bias prior to irradiation IUCF 74, 200 Me. V CYRIC 30, 70 Me. V 22

Hamamatsu Si 9055 (1) S 9055 -01, ø=100 μm fc=2 GHz IUCF 200 Me. V CYRIC 30, 70 Me. V 23

Hamamatsu Si 9055 (2) IPAS 3 Me. V, protons S 9055 -01, ø=100 μm fc=2 GHz 24

Hamamatsu Ga. As 8522 to proton damage Performance drops for lower energy protons G 8522 -01 ø=40 μm fc=3 GHz CYRIC 30, 70 Me. V IPAS 3 Me. V 25

True. Light high speed Ga. As PINs TPD-015 ø=90 μm fc=4. 25 GHz IUCF 200 Me. V IUCF 74 Me. V 26

True. Light high speed Ga. As PINs TPD-052 ø=70 μm fc=10 GHz IUCF 200 Me. V IUCF 74 Me. V 27

Summary on PIN proton damage Ø Si PIN proton energy dependence Ø is on Forward Bias, Dark Current, less on Responsivity Ga. As PIN proton damage needs more statistics on Hamamatsu and True. Light samples Summary on Cyric 30, 70 Me. V tests Degradation of responsivity (I/L) proton 2 E 14 Truelight S 9055 -01 S 5973 G 8522 -01 G 8522 -02 G 8522 -03 VR Vol -10 -2 -2 -3 -5 -5 -5 fc diam. I/L GHz μm A/W 100 0. 55 1. 5 200 0. 32 2. 0 100 0. 20 1. 5 400 0. 53 3. 0 40 0. 11 1. 9 80 0. 25 1. 5 120 0. 40 30 Me. V I/L Dark ratio n. A 45% 70 100% 40 100% 15 70% 100 45% 0 40% 0 35% 0 70 Me. V I/L Dark ratio n. A 45% 50 100% 20 100% 10 80% 50 80% 0 72% 0 28

NIEL discussion, discrepancy of proton data to NIEL of Ga. As J. R. Srour et al. , IEEE TNS 50, 653 (2003), Fig 8 J. H. Warner et al. , IEEE TNS 51, 2887 (2004), Fig 1 29