Comparison of laser receivers

for use at one micron wavelength

Michael A. Krainak, Guangning Yang, Xiaoli Sun, Wei Lu
NASA Goddard Space Flight Center

Xiaogang Bai, Ping Yuan, Paul McDonald, Joseph Boisvert,

Robyn Woo, Kam Wan, Rengarajan Sudharsanan
Boeing Spectrolab

Verle Aebi, Derek Sykora, Ken Costello

Intevac Inc.

Abstract-  Space-based lidar instruments must be able to detect instrument for topography mapping of the Earth’s surface.
extremely weak laser return signals from orbital distance. The These are the Ice, Cloud, and Land Elevation Satellite-II
signals have a wide dynamic range caused by the variability in (ICESat-II), Deformation, Ecosystem Structure and
atmospheric transmission and surface reflectance under a fast Dynamics of Ice (DESDynI) as tier-1 missions and Lidar
moving spacecraft. Ideally, lidar detectors should be able to
detect laser signal return pulses at the single photon level and
Surface Topography (LIST) as a third tier mission. These
produce linear output for multiple photon events. They should three missions could all possibly benefit from a very low-
have high quantum efficiency in the near-infrared wavelength noise (near single-photon sensitive), high-bandwidth (1
region where high-pulse-energy space-qualified lasers are GHz), near-infrared optical receiver. Improvements in
available. Silicon avalanche photodiode (APD) detectors have detector quantum efficiency translate directly to reduced
been used in most space lidar receivers to date. Their laser energy requirements for active laser instruments. This
sensitivity is typically hundreds of photons per pulse at 1064 minimizes spacecraft resource requirements (mass, power,
nm, and is limited by the quantum efficiency, APD gain noise, volume) and greatly enhances laser and system reliability.
dark current, and preamplifier noise. NASA is developing
photon-sensitive near-infrared detectors with linear response
for possible use on the next generation direct-detection space
All of NASA’s space-based laser altimetry missions to date
lidars. We present our measurement results and a comparison have used versions of the near-infrared enhanced silicon
of their performance. avalanche photodiode (APD) detector manufactured by
PerkinElmer Opto-Electronics Canada (a.k.a. EG&G and
I. INTRODUCTION formerly RCA) for detecting laser signal returns at 1064 nm
wavelength. These missions include NEAR,
In 2007 the National Research Council (NRC), responded to CLEMENTINE, MOLA, ICESAT/GLAS, CALIPSO, MLA
requests from the National Aeronautics and Space and LOLA. The spaceflight optical receivers are based on
Administration (NASA) Office of Earth Science, the custom-built improved versions of the commercial Model
National Oceanic and Atmospheric Administration (NOAA) C30954E APD. Its RF bandwidth of 140 MHz is matched
National Environmental Satellite Data and Information to the laser transmitter pulse bandwidth. The Si APDs
Service (NESDIS), and the U.S. Geological Survey (USGS) exhibit low excess noise factor and low noise (NEP of
Geography Division to conduct a decadal survey to generate 40~50 fW/rt(Hz) over a bandwidth of 140 MHz). This
consensus recommendations from the Earth and silicon APD-based optical receiver is described in detail in
environmental science and applications communities recent publications [2, 3]. A recent result [4] was an NEP of
regarding a systems approach to space-based and ancillary 5 fW/rt(Hz) with 200 MHz bandwidth.
observations that encompasses the research programs of
The laser altimeter system RF signal bandwidth is designed
NASA; the related operational programs of NOAA; and
[5] for a specific spatial resolution and average surface
associated programs such as Landsat, a joint initiative of
topography roughness. For the space-based ICESat-GLAS
USGS and NASA. The NRC recommended a total of
instrument, these parameters were 70-meter diameter
fifteen missions1 with benefits range from information for
ground area, surface slope varying from 0-30 degrees and
short-term needs, such as weather forecasts and warnings
140 MHz RF signal bandwidth (corresponding to the (-20
for protection of life and property, to the longer-term
dB) bandwidth for a 6 ns laser transmitter pulse width).
scientific understanding necessary for future applications
NASA’s future laser altimeters plan to have a higher spatial
that will benefit society in ways still to be realized. Three
resolution. A topographic mapping spatial resolution of 5
out of the fifteen recommended missions are laser based

1  
 
meters is planned for the LIST mission. This increases the product) are determined by k value. For low noise and high
RF (-20 dB) signal bandwidth to greater than 1 GHz and speed applications, a small k value is critical. Common
reduces the corresponding laser pulse width to less than 1 semiconductor materials excess noise factors are
ns. The 1 ns pulse width also provides higher height summarized in Figure 1.
resolution and increased probability of ground location                                      
detection under trees for reasonably flat (< 3 degree surface
slope) regions. Si photoreceivers do not meet these
requirements. At 1.06 µm, in order to get reasonable
absorption, a very thick absorber (several tens of microns) is
necessary. The thick absorber limits the Si APD bandwidth
to less than a few hundred MHz. On the other hand, III-V
compound semiconductor based APDs, like InGaAs APDs,
show high quantum efficiency in the 1 to 1.6 µm
wavelength region and fast response [2]; bandwidths greater
than several GHz are commonly achieved with these APDs.
However, current InGaAs APDs have NEP higher than >
500 fW/√Hz due to the use of high excess noise InP layer as
a multiplier.
There are several reasons that make it difficult to extend the
near-infrared-enhanced-silicon APD RF bandwidth to 1
GHz. First, a minimum diameter is required to match the  
laser ground footprint diameter image size for a practical Figure 1. Excess noise factors for commonly
telescope optics f-number. This leads to a larger device used semiconductor materials.
capacitance that contributes to the limiting the maximum
There are several approaches to reduce the k value in
bandwidth. Second, increasing the device thickness to
InGaAs APDs [3]. Spectrolab has been developing InGaAs
possibly further enhance the absorption length is not fruitful
APDs with InAlAs material as a multiplication layer and
because it increases the noise current and it increases the
InAlAs has a lower k value (0.22) [2], than commonly used
electron drift time. Therefore a new detector is required for
InP (0.5). In order to reduce the k-value further by
future NASA lidar missions that require 1 GHz RF
engineering the impact ionization values of InAlAs based
bandwidth.
heterojunction multiplication layers and these structures
were called I2E structures. In this approach, the multiplier
In 2009, we proposed and were awarded a NASA Earth
layers of the APD consist of two different semiconductors
Science Technology Organization (ESTO) Advanced
with different ionization threshold energies. It turns out that
Component Technology (ACT) program to pursue new
the probability of avalanche events is higher in the
detector development. Our idea is to use the impact
semiconductor with lower ionization threshold energy. This
ionization engineered (I2E) material system in Indium
preference translates into low noise because the avalanche
Aluminum Arsenide (InAlAs) to simultaneously achieve
events become less random. k value of 0.1 has been
high near-infrared quantum efficiency (75%), large diameter
demonstrated [4].
(200 mm), > 1 GHz bandwidth and high sensitivity (low-
noise – 300 fW/√Hz). In 2007, we began work on an
Instrument Incubator Program (IIP) to advance and
demonstrate needed technologies for the LIST mission with
the Airborne LIST Simulator (A-LISTS) pathfinder
instrument [6]. A-LISTS is a micropulse, waveform
recording system that is based on a new, and highly efficient
laser measurement approach utilizing emerging laser
transmitter and the I2E APD and intensified photodiode
(IPD) detector technologies described here.

1. LOW EXCESS NOISE APDS

APDs convert light signals into electrical signals with a gain
from an avalanche process. Both photo generated holes and
electrons could initiate an avalanche process. The excess Figure 2. Excess noise data of I2E APD
noise could be characterized by k=α/β or β/α (k<1 always), devices.
where α and β are electron and hole ionization coefficients.
APD excess noise and response speed (gain bandwidth

2  
 
Excess noise values were measured on I2E APDs and the (188 ps FWHM, 78 ps one-sigma) measurement results
data are shown in figure 2. The solid lines correspond to k for our 0.8 µm thick InGaAsP photocathode. We
value of 0.1, 0.15, and 0.2. The Spectrolab I2E APD shows k measured the timing jitter using two independent
value less than 0.1 at low gain regime due to the dead space instruments 1) the Picoquant HydraHarp 400
effect. At high gain regime, the k value tends to fall on a multichannel scaler providing 188 ps FWHM
value of 0.15, which is considerably lower than the InP (0.5) measurement and an Agilent Model DSA91304A
and the InAlAs (0.22). The measurement results are really (13GHz BW) oscilloscope that calculated the one sigma
encouraging for I2E approach. One advantage of I2E method jitter directly from the waveform histogram as 78 ps.
is that it does not rely on the dead space effect, which could
This is an upper limit on the timing jitter because it
yield low k value with multiplier less than 0.1 µm thick. The
includes the photon timing uncertainty associated with
thin multiplier APDs tend to suffer from high dark current
the 100 ps pulse width of the experimental test laser. For
from tunneling. Also, it is difficult to grow uniform thin
multiplier APD arrays. The Spectrolab I2E APDs have a further context, a visible single-channel HPD with a 1 mm
much thicker multiplier, which show very low dark current diameter photocathode has 28 ps timing jitter with 46% QE
and good uniformity. at 500 nm wavelength [3].

One dimensional APD arrays were fabricated to characterize

the wafer uniformity. The arrays had a pitch of 250 µm and
each individual APD had a mesa diameter of 75 µm. For 16
APDs, with a span of 4 mm, the breakdown voltage
variation was less than 200 mV across the array. The
variations on the photo current were the result of different
incident light power on each APD during the measurement.
The APDs showed low dark currents; at gain of 10, the
average dark current of the APD array was 1.34 nA with a
standard deviation of 0.034 nA. Devices also showed high
optical gain; an optical gain over 50 was achieved.
At 1.06 µm, the APD quantum efficiency is 70% and the
Fig. 3. Single-photon timing jitter measurement results for
quantum efficiency can be further improved by depositing IPD with 0.8 µm InGaAsP layer thickness. FWHM = 188
gold metal layer on the backside of APDs and we estimate ps. Dashed line is Gaussian fit with σ = 78 ps.
quantum efficiency will increase to about 85% at 1.06 µm.
Photoreceivers were built by integrating I2E APD devices 3. Additional 16-channel IPD measured characteristics
with a low noise GHz transimpedance amplifier (TIA) and
photoreceivers showed a bandwidth of 1GHz and an noise 0.25
equivalent power (NEP) of less than 150 fW/rt(Hz) at room
temperature. We believe the performance could be 0.2
Experiment

significantly improved by using a custom transimpedance Theory

amplifier rather than the present commercial low-cost TIA. 0.15

Histogram

2. INTENSIFIED PHOTODIODES 0.1

The intensified photodiode (IPD) detector, (a.k.a hybrid

photomultiplier tube (HPMT) or hybrid photodetector 0.05

(HPD)), provides high-gain with low-noise suitable for

single-photon sensitivity via electron bombardment of an 0
0 1 2 3 4 5 6 7 8 9 10
avalanche diode. Advantages of the IPD include large Photon number (k)
detector area (1 mm for a single channel device), high Fig. 4. Measured scaled histogram of the pulse height distribution
maximum count rates (> 200 Mcps), high bandwidth (GHz), and Poisson theory for λ=3.3
low afterpulsing, and near room temperature operation.
Further features include low dark noise, large dynamic The single photon impulse response has a pulse width of
range for full analog pulse received waveform preservation 550 ps. The adjacent pixel cross talk was than 1.1% for any
and photon number resolution. Operating principles and pair of adjacent pixels. The internal gain is >104 for each
performance results of previous IPDs have been reported for pixel. We measured reduction of dark counts per channel
use at visible [7], and near-infrared [8, 9, 10], wavelengths. with decreasing temperature and predict less than 10 kcps at
In recent work, we produced a 16-channel IPD with an -20 C.
InGaAsP photocathode for use at 1064 nm wavelength.
Each channel has a 159 µm x 159 µm sensitive detection An excess noise factor of 1.2 was measured from the pulse
area at the photocathode. Figure 3 shows the timing jitter height amplitude distribution. Figure 4 shows excellent

3  
 
 agreement for two independent sets of measured photon
number resolution (scaled pulse height distribution
histogram) and the Poisson probability mass function
λk e − λ
f (k, λ ) = theory with λ=3.3.
k!
3. LASER RECEIVER SYSTEM COMPARISON
We plan to test both the 16 channel APD and IPD
€ photoreceivers on the ALISTS instrument airborne flights.
In preparation we have modeled the performance for the
ALIST Swath Mapper instrument using a 10 kHz repetition
rate 1030 nm Yb:YAG laser with 1 ns pulsewidth at 10 km
altitude with a 5 m ground spot diameter, 3 degree surface
slope, 7” diameter receiver telescope and 80 degree Sun
angle.
  Fig. 6. ALIST Swath Mapper Performance with I2E APD
receiver.

ACKNOWLEDGMENT
We greatly appreciate the support of the NASA ESTO ACT
program that is funding this work.
REFERENCES
 
[1 ] Earth Science and Applications from Space: National Imperatives for
the Next Decade and Beyond, National Research Council of the
National academies, The National Academies Pres, Washington D.C.,
(2007).
[2] M.A. Krainak, X. Sun, G. Yang, L.R. Miko, and J.B. Abshire,
“Photon detectors with large dynamic range and at near-infrared
wavelength for direct detection space lidars” Proc. SPIE 7320,
732005 (2009).
[3] M.A. Krainak, X. Sun, G. Yang, and W. Lu, “Comparison of linear-
mode avalanche photodiode lidar receivers for use at one-micron
Fig. 5. ALIST Swath Mapper Performance with IPD wavelength” Proc. SPIE 7681, 76810Y (2010).
receiver. [4 ] F. Laforce “Low noise optical receiver using Si APD” Proc. SPIE
7212, 721210 (2009).
[5] C.S. Gardner, “Target signatures for laser altimeters: an analysis”
Applied Optics Vol. 21, No. 3 pp. 448-453 February 1982.
Figures 5 and 6 show the numerical model result of the [6] A. W. Yu, D. J. Harding, M. Krainak, J. B. Abshire, X. Sun, J. F
Cavanaugh, S. R. Valett, L. Ramos-Izquierdo, “Efficient Swath
probability of detection (left ordinate – solid curves) and the mapping Laser Altimeter Instrument Incubator Program
RMS ranging error (right ordinate – dashed curves) vs. the American Geophysical Union, , Abstract #EP44B-01 (2010).
apparent surface reflectivity for the IPD (with a 0.1 mJ laser [7] A. Fukasawa, J. Haba, A. Kageyama, H. Nakazawa, and M. Suyama
energy and 7 shot average)) and I2E APD (with 1.6 mJ laser “High Speed HPD for Photon Counting,” IEEE Transactions On
Nuclear Science 55, 758-762 (2008).
energy and 250 shot average)) respectively. The system [8] X. Sun, M. A. Krainak, W. E. Hasselbrack, W. E., D. F. Sykora, R.
operating points of interest for comparison are circled with La Rue, “Single photon counting at 950 to 1300 nm using InGaAsP
arrows indicating the appropriate axis. The IPD shows far photocathode – GaAs avalanche photodiode hybrid photomultiplier
superior laser receiver performance. However, the I2E APD tubes,” Journal of Modern Optics, 56, 284-295, (2009).
[9] R. A. La Rue, K. A. Costello, G. A. Davis, J. P. Edgecumbe, V. W.
receiver at present has far superior component reliability. Aebi, “Photon Counting III–V Hybrid Photomultipliers Using
Both laser receivers can achieve the same desired Transmission Mode Photocathodes,” IEEE Trans. Electron Devices
performance with the associated laser energies and 44, 672–678 (1997).
averaging. [10] R. A. La Rue, G. A. Davis, D. Pudvay, K. A. Costello, V. W. Aebi,
“Photon Counting 1060-nm Hybrid Photomultiplier with High
Quantum Efficiency,” IEEE Electron Device Lett. 20, 126–128
(1999).

4