NIST single photon detector workshop 2003

1
Workshop on Single-Photon Detectors. NIST,
Gaithersburg, MD April 1,2003
Room temperature IR InGaAs/InP APD photon
counters for quantum optics
Paul Voss, Marco Fiorentino, Kahraman Koprulu,
Jay Sharping, Xiaoying Li, and Prem
Kumar Center for Photonic Communication and
Computing Department of Electrical and Computer
Engineering Northwestern University, Evanston, IL
60208-3118 Tel [phone] Fax (847)
491-4455E-mail voss_at_ece.northwestern.edu Funded
by NASA GSRP, ONR, and ARO MURI
2
Outline

• Motivation
• Quantum Optics Experiments with all fiber source
• Quantum correlated photon pair production in
  fiber
• All fiber polarization entangled source
• 600 kHz coincidence detectors results
• 14 MHz coincidence detectors results
• Conclusion

3
Motivation

• Study four-wave mixing in optical fibers at
  1550nm at single photon level
• Picosecond pulsed light ? gated mode detection
• High pump pulse rate (75 MHz) ? GHz source
  possible
• Entanglement experiments take lots of time ?
  only NIST has equipment and environments that
  are stable forever
• Why room temperature?
• Afterpulsing much smaller at room temperature
• Higher dark count rate tolerable with gated mode
  operation

4
Four-Wave Mixing (FWM)in Optical Fiber

• Signal and idler photons are created in pairs
• They exhibit entanglement properties similar to
  signal and idler photons created in c(2)
  parametric down-conversion

5
Photon Counting of Parametric Fluorescence
output spectral filter
dispersion-shifted fiber Sagnac loop
Coherent OPO spectral slicer for pump signal
6
Photon Counting of Parametric Fluorescence
7
Coincidence Counting Results
Accidental Coincidences
Fiorentino, Voss, Sharping, and Kumar, IEEE
Photon. Technol. Lett. 14, 983 (2002)

• We measure coincidences for signal and idler
  photons generated by one pump pulse
• Coincidence rate is greater than that measured
  for two adjacent pulses
• The latter fit well with the theory for two
  independent pump sources

8
Polarization-Entangled Photon Pairs
9
Observation of Two-Photon Interference

• System Parameters
• Pump pulse separation 30ps
• Pump pulse width 4ps
• Pump power in each arm 0.39mW average(approx.
  1.3W peak)
• Down-conversion probability 0.12 pair/pulse
• Probability of getting 2 photons/pulse 10.7
• PM fiber length 20m
• Total efficiency 10 and 7
• Filter bandwidth 0.6nm

• Bi-photon Fringe Visibility gt 90

10
Outline

• Motivation
• Quantum Optics Experiments with all fiber source
• Quantum correlated photon pair production in
  fiber
• All fiber polarization entangled source
• 600 kHz coincidence detectors results
• 14 MHz coincidence detectors results
• Conclusion

11
600 kHz schematic
Trigger or photodiode input (75.3 MHz)
Photons to be detected
600 kHz
Avalanche Photodiode
Comparator (AD8611)
50 MHz Low Pass Filter (Minicircuits)
?128 down counter (74AC4040)
40 dB Amplifier (AD8009))
To PC
Digital Delay (DG535)
Analog Discriminator (AD8611)
Pulse Generator 8V 1 ns FWHM (Avtech AVPP1)
Pulse Shaper (74123)
To PC
Bias voltage (48 to 60 Volts)
Detection Thresholding
Synchronization
12
600 kHz detection thresholding
13
Photon Counting at 1550 nm

• InGaAs/InP, fiber coupled, APD (Epitaxx,
  EPM239BA) used in gated Geiger mode
• Diodes have high efficiency at 1550 nm, but also
  high dark-count rate
• To limit after-pulsing, diodes are gated once
  every 128 laser pulses

• 25, PD 0.2
• NEP 1.010-15 W/Hz1/2

• 20, PD 0.3
• NEP 1.610-15 W/Hz1/2

• Best results at room temperature to date

Thanks to Scott Endicter and Mark Itzler of
Epitaxx / JDS Uniphase, Inc.
14
Higher rate operation (14 MHz)

• Initial setup below 1 MHz because
• Delay Generator DG535 rate limit (1 MHz)
• Data Acquisition card (NI PCI6010E)
• Avtech Pulse Generator
• Upgrade includes
• Delay line built on phase shifter at 14 MHz
• Avtech Pulse Generator (25 MHz)
• Wanted fast A/D to sample photocurrent (AD9033)
• Want digital thresholding for auto-recalibration
• Acquisition to 20 MHz with NI PCI5102

15
14 MHz schematic
Detection Thresholding
Synchronization
Optical or Electric 14 MHz mode locked input (50?)
Photons to be detected (14 MHz mode locked fiber
laser)
G11 Amplifier (AD8009 non-inverting)
21.5 MHz LPF (Minicircuits)
Avalanche Photodiode
Coarse Fine Adjust Pot.
180 Phase Shifter (Minicircuits SPH-16)
Low Pass Filter (100 MHz)
16 MHz LPF (Homemade)
Amplifier (10 dB)
Duty Cycle Adjust Pot.
Trig In
Comparator (AD8611)
DAQ Scan Clock to PC
Analog/Digital Converter
Result (carry bit) out to PC
3X 4bit Adder (74ACT283)
Inverter (74AC04)
3X S/P Converter (74AC164)
Serial Threshold Loaded From PC
Mechanical Switch
Main output
AVC1 Avtech Pulser (1ns FWHM 10 V)
Bias voltage (-48.5 V)
?10 Monitor output
Duty Cycle Adjust Pot.
Amplifier (G20 Inverting AD8009))
Comparator (AD8611)
16
14 MHz setup
17
14 MHz results

• Threshold scanned for each bias 47-51 V
• 800 ps pulse 12 V high
• Q.E. 10-20 for dark count rate of 0.2
• Inconsistency due to timing drift in delay line

18
14 MHz results (longer pulse)

• Threshold scanned for each bias 49-53 V
• 1 ns pulse 10 V high
• Q.E. 8 for dark count rate of 0.2
• Inconsistency due to slow timing drift

19
Threshold determination

• Dark counts and light counts have different
  voltage distribution
• Similar to PMT
• There is an optimum bias voltage and threshold

20
After pulsing at room temperature

• 10 V
• 1 ns pulse

21
Conclusions

• High speed IR detectors crucial for our
  experiments
• Built coincidence detectors at 600 kHz and 14
  MHz
• Quantum efficiency 10-20.
• dark count rates 0.2
• afterpulsing rate negligible at 600 kHz, not at
  14 MHz
• Future work ? clean up timing drift