Lecture 11-IR arrays and hybrid detectors

Download Report

Transcript Lecture 11-IR arrays and hybrid detectors

Astronomical Observational Techniques
and Instrumentation
RIT Course Number 1060-771
Professor Don Figer
IR array/hybrid detectors
1
Aims for this lecture
•
describe modern infrared hybrid array technology used in
Astronomy
2
Lecture Outline
•
•
•
•
•
•
Scientific value of infrared arrays
Hybrid architecture
Light-sensitive materials
ReadOut Integrated Circuit (ROIC)
Array characterization and performance
Common devices in the field today
3
Scientific Value of Infrared Arrays
4
History
•
•
•
•
•
•
•
•
Herschel’s detection of IR from Sun in 1800
Johnson’s IR photometry of stars (PbS) mid 60’s
Neugebauer & Leighton: 2um Sky Survey (PbS), late 60’s
Development of bolometer (Low) late 60’s
Development of InSb (mainly military) early 70’s
IRAS 1983
Arrays (InSb, HgCdTe, Si:As IBCs) mid-80’s
NICMOS, 2MASS, IRTF, UKIRT, KAO, common-user
instruments, Gemini, etc.
• JWST and the search for cosmic origins
5
Historical motivation
• Exploration & discovery
– Neugebauer, Leighton, Low, Fazio, Townes
• Technological opportunities
–
–
–
–
Bolometer (Low)
PbS (Neugebauer)
balloons (Fazio)
IR lasers & interferometry (Townes)
• A few, key problems
– Bolometric luminosities (Herschel, Johnson)
– The Galactic Center (Becklin)
– Star formation
6
Current interest in infrared
• High redshift objects
lobs = l0 (1+z)
5000 Å  >1 mm for z > 1
– Classical problems require infrared data
• Obscuration by dust
– Al ~ l-1.9  A2.2mm ~ 0.1 AV (Mathis 1990, ARAA, 28, 37)
– Now important for:
• Galactic nuclei, esp. AGN (unified model)
• Starburst galaxies
• Young stars
• Very low mass objects & extrasolar planets
– Tplanet ~ 50 to 500 K
– TBD ~ 900 – 2000 K
lpeak ~ 5 – 50 mm
lpeak ~ 1 – 5 mm
7
Extinction by dust
J. S. Mathis 1990, ARAA, 28, 37.
l
A(l)/A(J)
l
A(l)/A(J)
l
A(l)/A(J)
( m m)
( m m) R v= 3.1 R v= 5.0
( m m) R v= 3.1 R v= 5.0
250
0.0015
5
0.095
0.095
0.24
9.03
5.13
100
0.0041
3.4
0.182
0.182 0.218
11.29
6.03
60
0.0071
2.2
0.382
0.382
0.2
10.08
5.32
35
0.013
1.65
0.624
0.624
0.18
8.93
4.66
25
0.048
1.25
1.00
1.00
0.15
9.44
4.57
20
0.075
0.9
1.70
1.70
0.13
11.09
4.89
18
0.083
0.7
2.66
2.43
0.12
12.71
5.32
15
0.053
0.55
3.55
3.06 0.091
17.2
-12
0.098
0.44
4.7
3.67 0.073
19.1
-10
0.192 0.365
5.53
4.07 0.041
9.15
-9.7
0.208
0.33
5.87
4.12 0.023
7.31
-9
0.157
0.28
6.9
4.34 0.004
3.39
-7
0.07
0.26
7.63
4.59 0.002
1.35
8
Hybrid Architecture
9
Infrared Hybrid Array
• Infrared arrays use light-sensitive material that can detect
infrared photons, wavelengths beyond ~1um.
• Therefore, silicon cannot be used as the light-sensitive layer.
• This poses a problem because the readout circuit is most easily
implemented in silicon.
• Therefore, infrared arrays are “hybrids” – they use one
material to detect light and silicon for the readout circuit.
10
Pixel-level Cross Section: InSb
11
Pixel-level Cross Section: HgCdTe
12
NICMOS
13
Si PIN Architecture
• Si PIN devices are architecturally similar to infrared arrays, so
we discuss them in this lecture even though they do not
detector infrared light (they use silicon!).
• A PIN is a sandwich of p-type/instrinsic/n-type layers.
• These devices are usually implemented for improved response
at longer wavelengths, i.e. they are thick.
• In order to ensure that photogenerated charge makes it to the
integrating node, there needs to be a strong electric field that
depletes the intrinsic layer.
PIN diode
14
Detector Size
15
Light-sensitive Materials
16
Valence & Conduction Bands in
Semiconductors
• When atoms (a) come together to
form a crystal, the outer energy
levels overlap and blend to create
bands (b).
• The outermost filled band is
called the valence band (c).
• Above the valence band, one
finds a forbidden energy gap -the
“band gap”, and (at higher
energies) conduction bands
populated by thermally excited
electrons.
• In metals, the valence and
conduction bands overlap
resulting in conduction. In
insulators, the band gap is wider
resulting in very poor conduction.
17
Periodic Table
• Semiconductors occupy column IV of the Periodic Table
• Outer shells have four empty valence states
• An outer shell electron can leave the shell if it absorbs enough18
energy
Periodic Table Continued
• The column number gives the number of valence electrons per
atom. Primary semiconductors have 4.
• Compounds including elements from neighboring columns can
be formed. These alloys have semiconductor properties as well
(e.g. HgCdTe & InSb).
• Mercury-cadmium-telluride (HgCdTe; used in NICMOS) and
indium-antimonide (InSb; used in SIRTF) are the dominant
detector technologies in the near-IR.
19
The Band Gap Determines the Red Limit
E G  hc 
hc
lc
.
(1)
20
ROIC aka MUX
21
NICMOS MUX
22
Pixel Capacitance and Gain
Q  C V
23
Characterization and Performance
24
Dark Current
• Like a CCD, IR arrays are affected by dark current and a
variety of noise mechanisms.
• Dark current is the signal that is seen in the absence of any
light. For the near-IR (1-2.5 μm), the dominant components
are diffusion, thermal generation-recombination (G-R) of
charges within the semiconductor, and leakage currents.
Combining these, it can be shown that
1
2
idark
kT
2kT  V 
eV / kT
exp  1 
1   exp eV / 2 kT  1  ileak , (2)

eR0diff
eR0GR  Vbi 
where V is the voltage across the detector and R0diff and R0GR
are the detector impendences at zero bias for diffusion and G25
R.
Dark Current Example
26
Dark Current vs. Temperature
27
kTC Noise
• “kTC” noise occurs in both CCDs and IR arrays when the detector
capacitor is recharged. As we will see shortly, this goes as
 kTC
kT C

electrons
e
where k is the Boltzmann constant, T is temperature, C is capacitance,
and e is the elementary charge.
28
1/f Noise
• 1/f noise in the multiplexer’s output field effect transistor
(FET).
– In the mid-1980s, readout noise ~300 electrons was common.
Better output FETs increased the number of microvolts per
electron and thereby helped to reduce readout noise to <10
electrons seen today.
• Other 1/f components are seen. e.g.,
– The “pedestal effect” seen in NICMOS whereby the entire
exposure’s bias is seen to vary by ~75 electrons. There is some
consensus that these 1/f components are thermally driven.
29
Noise Reduction Through Multiple Sampling
• By being a bit clever about reading out the array, one can
minimize or eliminate some of these noise modes.
• During an exposure, typically each pixel is sampled several
times.
• The most common approaches are correlated double sampling
(CDS), multiple non-destructive reads (aka “Fowler
Sampling”), & fitting a line (aka “up the ramp”). We discuss
these briefly.
30
Fowler Sampling
1 N
1 N
V   Vi j   Vf j
N j 1
N j 1
31
“Up the Ramp”
•
•
•
•
Fit best line to multiple non-destructive samples.
Sample spacing does not need to be uniform.
Not clear whether this or Fowler sampling is best.
This is what is done in NICMOS MULTIACCUM mode.
32
Read Noise Example
33
QE Example
Relative QE for Bepi-Colombo device, at 293 K, of RVS
MCT array after substrate removal. High response in
the UV region is due to charge gain. The QE reduces
by ~8% at 190 K.
34
Well Depth and Non-linearity
35
Persistence
• Persistence is produced by charge traps.
• A trap can be modelled as a square well with a lip.
• One robust conclusion is that the trap will decay with an
exponential timescale.
36
Persistence Continued
37
Pixel-to-pixel Crosstalk
• Crosstalk occurs when a pixel value is influenced by its
neighbors.
• Charge diffusion is an important crosstalk mechanism in IR
arrays and CCDs.
• Once charge carriers are created, their motion is governed by
charge diffusion.
• We discuss some of his results.
38
IPC
• Interpixel capacitance (IPC) is another form of crosstalk.
• In this case, charge in a pixel induces a voltage change in a
neighbor, just like the behavior between parallel plates in a
capacitor.
• The effect is to blur the point spread function.
• The induced voltage does not have noise.
39
IPC
• In this example, IPC is very large for the H4RG SiPIN device
(10 um pixel size).
40
Arrays in Use Today
41
Teledyne Family Arrays
H2RG HgCdTe array
42
Teledyne H4RG Si PIN
• This is an image of the H4RG device in the Rochester
Imagingn Detector Laboratory.
43
Raytheon Family Arrays
44
Potential Future Innovations
45
MCT on Si
Design of the light-sensitive part of the pixel for the
RVS SB-390 MCT on Silicon (Bornfreund et al. 2005).
46
MCT on Si
5
Cutoff Wavelength (um)
293 K
150 K
77 K
40 K
4
3
2
1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
x
Egap  -0.3021.93x  5.35(10-4 )T(1-2x) -0.810x2  0.832x3.
47
Operation of Avalanche Photodiode
Linear
on
Geiger
mode
mode
on
Linear
Geiger
quench
mode
mode
avalanche
Current
off
off
arm
Vdc + DV
Vbr
Voltage
Avalanche Diode Architecture
-V
hν
Quartz substrate
p+ implant (collects holes)
low E-field
10 µm
p+ implant
high E-field
n+ implant (collects electrons)
metal
metal
metal
bump bond
ROIC
+V
0.5 µm
GM-APD
50
Performance Parameters
 Photon detection
efficiency (PDE)
 The probability that a
single incident photon
initiates a current pulse
that registers in a
digital counter
 Dark count rate (DCR)
 The probability that a
count is triggered by
dark current
Single photon input
APD output
time
Discriminator
level
Digital comparator output
time
time
Successful
single photon
detection
Photon absorbed
but insufficient
gain – missed
count
Dark count –
from dark
current
Zero Read Noise Detector ROIC
metal bump bond
pad
2 pixels, 50 mm
core
(active
quench,
discriminator,
APD latch)
counters (4 pixels)
counter
rollover latch
2 pixels, 50 mm
(left) Floorplan of the unit cell (2×2 pixels) for a previously-designed
256×256 pixel CMOS ROIC. (right) Photograph of this ROIC.
Figure 1.
52
GMAPD Project Hardware
LM-APD
54
Applications
Imaging
Both linear mode and Geiger mode APDs have strengths and weaknesses when it
comes to imaging.
–
–
–
–
GM-APDs can be implemented with zero read noise (often the limiting factor in state-of-the-art
imaging systems) when designed with an in-pixel counter. This makes the entire detection process
digital.
GM-APDs cannot distinguish between one or more photons arriving simultaneously, while that
information is maintained in a LM-APD.
LM-APDs have effectively no dead time, which this is a limiting factor in GM-APD performance.
LM-APDs must be reset periodically to remove the charge accumulated by leakage current or the
detector quickly becomes saturated, while this is not a problem with GM-APDs.
0 and 10 electrons read noise for
images with simulated shot noise
55
Applications
Imaging (single photon counting)
Figures Courtesy of Don Hall (University of Hawaii)
56
Applications
Imaging (GM-APD)
This image taken with a
prototype GM-APD imager by
researchers at Lincoln
Laboratory (MIT).
Image courtesy of Don Figer.
57