Evaluation of HgCdTe APDs for Single Photon Counting

1
Evaluation of HgCdTe APDs for Single Photon
Counting

• NIST Workshop on Single-Photon Counting
• 1 April 2003

Work supported in part by NIST contract
SB1341-02-C-0015
Jim Gates 760-730-9768 ji
mgates00_at_aol.com
2
Outline of the Presentation

• NIST objectives
• Approaches to photon counting
• APD selection criteria and test results
• Performance model for single photon counting at
  100K (linear mode)
• Resistive transimpedance amplifier (RTIA)
  simulation
• Future plans
• Summary and conclusions

3
NIST Objectives for Voxtel SBIR

• 1.55-micron photon counting module
• Small size (beer can or smaller)
• 200-micron diameter APD
• Fast counting rate (1-10 Mcps goal)
• Prefer Geiger mode but linear mode is an option
• Low dark counts (cooling required)
• Two modules for correlated photon counting
• Open to other approaches (SBIR Research)

4
Geiger vs. Linear Mode

• Geiger Mode (gated) achieves 300Kcps 10 Mcps
  (at RT)
• Features mature and APD noise not important
• Challenges high effective QE (detection
  efficiency), dark counts, afterpulsing, APD
  failure (need better APDs and quenching
  circuits).
• Linear mode just being developed at 1.55 microns
• Features no afterpulsing, no gating, high QE
• Challenges low excess noise, TIA circuits and
  high linear mode APD gain, better APDs (dark
  current)
• This presentation will discuss linear mode with
  HgCdTe APDs

5
Photon Counting Detective Assembly

• Large Diameter APD 200-micron (or smaller with
  micro lens)
• RTIA amplifier (cooled)
• Packaged in vacuum sealed dewar with TE cooler
  (or Stirling minicooler)

TEC
Coldstage
TIA
Circuit
Temperature Sensor
J1
Board
High Voltage Bias
RTIA
Power
Output Signal
AD590 Temp
High Voltage
Sensor
Bypass
Detector
RTIA
Cooler Power
J2
TEC Power
TE Cooler with
Cooler Controller
Vacuum
Filter Board
Thermistor
Dewar
6
Miniature Stirling Cooler (required if APD cooled
below 200K)

• Cooler Weight 270 gr.
• Input Voltage 12 VDC
• Input Power (170mW _at_80K _at_23C) 4.5W
• Cool down Time (190J _at_23C) 5 Minutes
• MTTF gt 5,000 Hours (Expected)
• Cost (3-10K)

7
HgCdTe Advantages for APDs

• High Q.E Over Wide Spectral Range 1-2 mm or
  Larger
• Can be Optimized for Specific Wavelengths
• QE gt90 (with AR coat)
• Lattice matched growth
• QE maintained when cooled
• Back illuminated Vertical Design Enables
• GHz Bandwidths and double pass
• Low Excess Noise
• k can be engineered to low values lt 0.1
• Low k tracks with detector temperature
• Good Uniformity due to material control
• Low Idark due to material purity, GRings and
  passivation
• Synergy with High Volume Manufacturing of IR
  Detectors

8
InGaAs ?c vs. T (InGaAs needs to be above 100K
for 1.55 micron)
9
Need low k (less than 0.1) with high gain in a
linear mode
10
HgCdTe APD Design
11
K-value engineering
12
Eg and Delta tuning to achieve low excess noise
in the gain region
13
Resonance at 0.928 eV (1.375 micron cutoff used
for gain region)
14
1.60 and 1.38 micron ?c vs. T (Does not change
much with T)
15
APD Test Procedure
APD attached to a Test TIA and signal
g
current resulting from the optical pulse determine
Flux measured using a calibrated diode (W)
g
Dark current is subtracted from light current
g
and the
responsivity
(A/W Determined)
QE measured voltage at which the APD is fully
g
depleted 40V but not showing gain.
Gain is determined as the ratio of
responsivity
g
at high voltage to
responsivity
at 40V
Noise is measured 10MHz
g
and integrated over 42MHz Bandwidth
16
Excess Noise of the APD versus Gain fit to the
McIntyre Model for APDs

17
Quantitative Measurements Use Dedicated APD Test
Facility
Acquisition Software and Data Base
Pulse response/detector rise time
Noise Spectrum
Optics/Display
Calibrated Lasers
Laser Illuminated Wafer Probe
18
HCT APD Gain vs. Vdet (APD operates at low
temperatures)
19
Id/G vs. G _at_100K (100 µm diameter APD Id very low
at 100K)
20
Linear mode 1.55 micron single photon counting is
possible
21
RTIA Simulation (low noise and high speed are
possible)
22
Future Plans

• Test more HgCdTe APDs
• Test APDs in linear mode with RTIA (HgCdTe and
  InGaAs/Si)
• Test APDs in Geiger mode (InGaAs/InP, InGaAs/Si
  and HgCdTe)
• Make a selection of APD and mode
• Build two single photon counting modules

23
Summary and Conclusions

• HgCdTe appears promising in linear mode operation
  to achieve single photon counting. The APD
  achieves high gain at 100K. Id and excess noise
  is low.
• InGaAs/Si also has low noise and may work very
  well in a linear mode and Geiger mode.
• New APD detectors need to be evaluated for Geiger
  and linear mode to meet NIST requirements.
• Innovations in circuitry and APD detectors will
  improve single photon counting at 1.55 microns.
• Linear mode may distinguish between one and two
  photons.