Transcript No Slide Title

The Paradigm of Optoelectronics
optics
optical
SOURCE
experiment
medium, or
optical circuit
DETECTOR
signal out
signal in
information
optoelectronics
processing
electronics
INTRODUCTION
1
a good photodetector is like a good source
Preceived
S/N = (atten) . P source / P noise
system
quality
factor
decreasing detector
noise by K is the
same as increasing
source power
by K .!.!
INTRODUCTION
2
Milestones in Photodetection
1829-33: Nobili(I) and Macedonio Melloni (I) invent the thermopile
1873: W. Smith (UK) discovers photoconductivity in Selenium
1905: A.Einstein explains photoemission by the quanta hypotesis
1910s: first S-1 photocathodes and vacuum-photodiodes
1919: J. Slepian (USA) invents the photomultiplier
1930: V. Zworikyn and G. Morton (USA) demonstrate television
1930: Image converter tubes and streak-camera tubes developed
1940s: The ‘Sniperscope’ image converter tubes used as night-vision aid
1950s: Solid-state theory developed, first Ge-photodiodes
1965: Planar Si-photodiodes, EBS camera tubes, InSb IR-detectors
1967: Apollo 11 send images of the moon taken with EBS
1967: Avalanche photodiodes invented
1970: first CCD image-pickup devices
1970: The Apollo 15 Lunar Ranging Experiment
1975: Compound semiconductor photodiodes cover all the IR bands
1980: Camcorders with CCDs are mass produced
1995: room-temperature Thermovisions
199 6: Hubble Space Telescope is equipped with a CCD camera
INTRODUCTION
3
Photodetectors and their Spectral Ranges
SINGLE ELEMENT
- photoemission devices
(or external
photoelectric devices)
vacuum photodiode
gas photodiode
photomultiplier
- internal photoelectric
devices
semiconductor photodiode
avalanche photodiode
phototransistor (BJT, FET)
photoresistance
- thermal detectors
- weak interaction
detectors
IMAGE
pickup tubes
image intensifiers
and converters
CCDs
vidicon
thermocouple (or photopile)
thermistor (or bolometer)
uncooled IR FPA
pyroelectric
IR vidicon
photon drag, Golay cell
photoelectromagnetic
point contact diode
0.1mm
1mm
10mm
100mm
(l)
___|_____________|_____________|_____________|_____________|_
__photoemission ____
____internal photoelectric effect _____
_____________________thermal________________________________
Detectors based on Photoelectric effect
P
DET
Power collected P = hn F
is a flux F of photons of energy hn
I
V
bb
C
R
V
Output current I = e F’
is a flux F’ of electrons of charge e
Then, current is proportional to power,
I/P= s = e F’ / hn F = h (e /hn)
where h = F’/F is quantum efficiency (electrons-to-photons)
and
s = I/P= h (le /hc) = h (l /1.24) [A/W]
is spectral sensitivity (current out -to-power in)
INTRODUCTION
5
Detectors based on Photoelectric effect 2
To trade photons for electrons we need a material
requiring an energy not larger than the photon
energy, so hn≥Ecc , where energy Ecc for the charge
carrier generation is EW (work function) in external
and EG(bandgap) in internal photoemission.
This is the threshold condition: hc/l≥Ecc or
l ≤ lt = hc/eEcc = 1.24 / Ecc (eV)
In alkaline antimonides, EW ≈1.2-3.0 eV,
and lt ≈1-0.4 mm (blue to NIR)
ternaries (InGaAs) EG ≈0.75 eV, lt ≈1.8 mm
InSb
EG ≈0.25 eV, lt ≈5 mm (MIR)
HgCdTe EG ≈0.08
eV, l ≈16 mm (FIR)
INTRODUCTION t
6
Detectors based on Photoelectric effect 3
s = I/P
(A/W)
threshold
h=1
1.0
h=0.5
real response
1.24
l
l
t
general response curve of a quantum detector:
at P=cons, current increases linearly with l, then
sharply decreases to 0 at the photoelectric threshold
a real detector has a curve rather than a triangle
INTRODUCTION
7
Detectors based on Photoelectric effect 4
Once produced, we shall remove charge carriers
fast, so we need very thin layers to cross or a
favorable electric field
helping collection
photocathodes
pn junction in a diode
base-collector junct.of BJT
gate-drain junct in a FET
depleted layer in a MOS
3rd junct in a SCR
applied field in a resistance
INTRODUCTION
8
Quantum detectors: S/N vs Iph
10
2
170
10
160
(S/N)
1
√Iph0/2eB
S/N (dB re
1mA,1Hz)
F
150
-1
10
140
thermal regime
-2
10
quantum regime
130
10-2
10
-1
1
10
10 2
10
3
I ph / I ph0
INTRODUCTION
9
Time measurements
coax line
R c = 50W
3R
R
R
R
1.5 R
- HV
- best with n=12 stages (G=107 - 108) for weak signals
- output terminated on Rc=50 W for best bandwidth, first (3R)
and last dynodes with more voltage (3R, 6R)
impulse response: SER-limited (typ. duration Dt =2 ns),
intrinsic limit of accuracy: sT={st02+[g/(g-1)g1]sti2}1/2
(typ.)=0.58 ns for 1 photoelectron
PMT APPLICATIONS
10
Time measurements
Time resolution of a
typical 12-stage
PMT with Dt=3 ns
followed by a
constant fraction
timing (CFT). Data
for a I(t)=d(t) light
pulse; timing
threshold is set at a
fractional level S0
=C/R of the total
collected charge R.
R=1
TIME ACCURACY s
T (ps)
1000
10
100
100
1000
10
0
0. 05
0. 1
0. 15
0. 2
threshold l evel (at frac tion of charge C/R)
PMT APPLICATIONS
11
Photon counting
F
DELAY t r
LINEAR
GATE
50W
SHAPER
AL
DISCR.
S1
COUNTER
reset
0-T
DISCR.
S2
INTEGRATE
AND DUMP
DISPLAY
90
PHOTON C OUNTS
-V
E
N s+ N d
60
N
d
30
basic functional scheme of
photon counting with PMTs
(top), and an example of a
measurement, showing
evolution of signal plus dark,
dark only, and result after
dark subtraction (bottom);
vertical bars on DN indicate
the ±0.5sDn standard
deviation confidence
intervals
DN
0
1
2
3
4
COUNTING TIME T
PMT APPLICATIONS
12
Photon counting (cont’d)
Requirements for PMTs in a photocounting regime:
 a SER amplitude in the mA range (to get ≈100mV in circuits),
i.e., G≈107-108 and n≈12 dynodes
 a high first dynode gain for a good discrimination efficiency
 a voltage divider adequate to have a short Dt of SER.
- maximum photon rate acquired for the photocounting:
F=1/Tr
(set by the integrator recovery time, Tr =3-10 Dt typ. )
- dynamic range (in power):
P=e/Trs
(for s=20mA/W, Tr=10ns it is P=0.8 nW)
PMT APPLICATIONS
13
Analysis of the photon counting
Total counting N= Ns+Nd, (signal plus dark) has a mean value:
N= Ns+Nd = h.hp F T +(hdId /e)T
and, following Poisson statistics, variance is:
sN2 = Ns+ Nd
whence
(S/N)2 = Ns2 / [Ns+Nd]
noise figure NF2 of the photocounting process:
NF2 = (S/N)2Nd=0 /(S/N)2 =1+ Nd/Ns= 1+ (hdId /e)/h.hp F
The minimum measurable radiant power, at dark counting rates Id/e of a few
electrons/cm2.s, (at h. h p=0.1, near the peak of photocathode response) is:
P = Fhn/h.hp 10-18 W
a performance unsurpassed by any other kind of photodetector
PMT APPLICATIONS
14
Photon counting with dark subtraction
Subtracting dark from signal plus dark (with the same T)gives:
DN= N1-N2=Ns+Nd-Nd=Ns
The variance of DN is the sum of the variances, so that:
sDN2 = Ns+2Nd
and
(S/N)2 =Ns2/ [Ns+2Nd]
NF2 = 1+ 2Nd/Ns = 1+ (2hdId /e)/h.hp F
letting S/N=1 and for weak signals (Ns<<Nd), it is:
Nsmin = √[2Nd] = √(2hdId T/e)
or, the minimum detectable signal, being Ns=hhp FT, is:
Fmin = √(2hdId /eT)/hhp
PMT APPLICATIONS
15
Photon counting with dark subtraction: an example
With a PMT having:
- a S-24 response, with h=0.35 at l=400nm,
- a 1cm2 area, with a dark current rate Id/e=1cm-2
- a first dynode gain g1=3
- thresholds Q1/Ge=0.8 and Q2/Ge=2.5 (hp=0.7, hd=0.8)
the minimum detectable rate is: Fmin=√(2.0.7.1/T)/0.35.0.7
=4.83/√T photon/s.
for a 8-hour integration period this yields:
Fmin=4.83/√28800=0.028 photon/s, or
Pmin=hnFmin= 1.3 .10-20 W,
i.e., the power collected from a m= 28th magnitude star with
a 1m2 telescope aperture.
PMT APPLICATIONS
16
Nuclear spectrometry with scintillation counters
CHARGE
AMPLIFIER
SCINTILLATOR
A
B
50W
PULSE
SHAPER
C
in
-VAL
DISCRIMINATOR
enable
MULTICHANNEL
ANALYZER
A
DISPLAY
B
t
C
Cs 137
K 40
COUNTS

662 KeV
functional scheme of
energy spectrometry
measurements (top) and
waveforms (middle). The
typical energy spectrum
obtained with the
scintillation detector
(bottom) reveals
radionuclides species
(energy signature) and
their concentrations
(counts intensity)
1.46 MeV
AMPLITUDE
PMT APPLICATIONS
17
Dating with scintillation counters
Carbon isotope 14C, initially absorbed from atmosphere during
specimen life, decays with a half-life T1/2=5700 years emitting
160 keV electrons.
Dating is performed by looking at the detection rate of 160 keV
electrons (peak amplitude of the MCA+scintillator spectrum),
which is proportional to the residual content of 14C in it.
Measurable concentrations are down to c=10-3 of the initial
value c0, according to the simple expression
T=T1/2 log2 (c/c0),
with a practical limit in dating to about 40 000 years.
Other radionuclides, such as 87Rb, 232Th, 235U and 40K, are used
to cover much larger time spans (up to 106-109 years) and in
inorganic samples where 14C is absent
PMT APPLICATIONS
18