OPTICAL FIBER COMMUNICATIONS 2018

UNIT-IV
PHOTO DETECTORS –
PHYSICAL PRINCIPLES OF PHOTO DIODES,
PHOTO DETECTOR NOISE: SNR, DETECTOR
RESPONSE TIME, AVALANCHE
MULTIPLICATION NOISE, COMPARISONS OF
PHOTO DETECTORS.
FUNDAMENTAL RECEIVER OPERATION –
PREAMPLIFIERS,
ERROR SOURCES,
RECEIVER CONFIGURATION –
PROBABILITY ERROR,
QUANTUM LIMIT.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 1
 OPTICAL FIBER COMMUNICATIONS 2018

PHOTO DETECTORS:
The Photo Detector senses the luminescent power falling and converts the
variation of the optical power into a corresponding varying electric current. The
Photo Detector must meet very high performance requirements those are with high
response or sensitivity in the emission wavelength.

In OFC system there must be a receiving device which interprets the

information contained in the optical signal. The first element of this receiver is a
photodetector. Different types of photodetectors are in existence. They are Photo
multipliers, Pyroelectric detectors, Photo diodes.

Photomultipliers: Photomultipliers consists of a photocathode and an electron

multiplier packaged in a vaccum tube that are capable of providing very high
gain and very low noise.

Pyroelectic Detectors: Pyroelectric photo detectors involve the conversion of

photons to heat. They give rise to variation in the dielectric constant which is
usually measured as a capacitance change.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 2
 OPTICAL FIBER COMMUNICATIONS 2018

Photodiode: Photodiode is used almost for all fiber optic systems because of its
small size, suitable material, high sensitivity & fast response time.

Two types of photo diodes are (1) PIN photo detector (2) Avalanche Photo
diode [APD]

THE PIN PHOTODETECTOR:

The most commonly used photo detector is the photodiode. The device
structure consists of p & n regions separated by a very lightly n-doped intrinsic
region. In normal operation a sufficiently large reverse bias voltage is applied across
the device so that the intrinsic region is fully depleted of carriers. So, the intrinsic n-
carrier and p-carrier concentration is negligibly small and it is compared with
impurity concentration.
When an incident photon has energy greater than or equal to the band-gap
energy, the photon can give up its energy and exit an electron from valance band
to the conduction band.

Figure 4.1: Schematic representation of a pin photodiode circuit with an applied

reverse bias.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 3
 OPTICAL FIBER COMMUNICATIONS 2018
This process generates free electron-hole pairs which are known as photo
carriers (or) Photon-generated charge carriers. The carriers are generated in
depletion region where most of incident light is observed. The high electric field
present in the depletion region causes the carrier to separate and collected across the
reverse bias junction.
This gives a current flow in an external circuit, with electron flow for
every carrier pair generated. This current flow is known as photo current. As the
charge carrier flow through the material, some electron-hole will recombine &
hence disappear.
The charge carrier travels at distance Ln or Lp for electrons and holes
respectively. This distance is known as diffusion length. The time it takes for an
electron or hole to recombination is known as carrier life time & is represented by
Ƭn & Ƭp respectively.
Life times & diffusion lengths are related by
Ln= (DnƬn)1/2 & Lp=(Dp Ƭp)1/2
where Dn & Dp are electron and hole diffusion coefficients.
The optical radiation is absorbed as P(x) = Po(1- e(-αs(λ)x)) where αs(λ)
is absorption coefficient at a wavelength λ, Po = the incident optical power level
P(x)=optical power absorbed in a distance x
If the depletion region has a width w, then the total power absorbed in the
distance w is P(w)=Po(1- e(- αsw))
The primary photocurrent Ip= (q/hv)*Po(1- e (- αsw))(1-Rf)
Where Po=optical power incident on Photodetector, q=electron charge,
Rf=refractivity, hv=photon energy
The quantum efficiency Ƞ is the no. of electron-hole carrier pairs generated
per incident photon of energy hv is given by
Ƞ = no. of electron-hole pairs generated / no. of incident photons
= [Ip/q] / [Po/hv]
where Ip=average photo current generated by a steady-state average optical
power Po incident on the photodetector.
Problem: In a 100ns pulse 6x10(6) photons at a wavelength of 1300nm on an
InGaAs photodetector. On the average 3.9x10(6) electron hole pair is
generated. The quantum efficiency is ---------
Ans. Ƞ=3.9x106 / 6x106 = 0.65
Thus the quantum efficiency at 1300nm is 65%.
UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 4
 OPTICAL FIBER COMMUNICATIONS 2018

AVALANCHE PHOTODIODES:
APDs internally multiply the primary signal before it enters the input
circuitry of the following amplifier. In order for carrier multiplication to take place,
the photo generated carriers must traverse a region where very high electric field is
present. The carrier multiplication mechanism is known as impact ionization.
The newly created carriers are also accelerated with high electric field, thus
gaining enough energy to cause further impact ionization. This phenomenon is
called as Avalanche effect.
A commonly used structure for achieving carrier multiplication with the
little excess noise is the reach through construction. RT photo diode is composed of
high resistive p type material deposited as an epitaxial layer on a p+[ heavily doped
p-type ] substrate.
A p-type diffusion or ion implant is then made in the high-resistive material
followed by construction of an n+[ heavily doped ] layer. This configuration is
referred to as a p+Пpn+ reach through structure.

Figure 4.2: Reach-through avalanche photodiode structure and electric fields in the
depletion and multiplication regions.

The П layer is basically an intrinsic material. In normal usage RAPD is

operated in the fully depleted mode. Light enters the device through p+ region and
is absorbed in П material, which acts as the collection region for photo generated
carriers.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 5
 OPTICAL FIBER COMMUNICATIONS 2018

The average no. of electron hole pairs created by a carrier per unit distance
traveled is called the ionization rate. The multiplication M for all carriers generated
in the photodiode is
M=Im/Ip where Im= average value of the total multiplied output current.
Ip= the primary unmultiplied photo current.

PHOTODETECTOR NOISE & SNR:

In fiber optic communication systems the photodiode is generally required
to detect very weak optical signals. Detection of the weakest possible optical
signals requires that the photodetector and its following amplification circuit to
optimize so that a given SNR is maintained.

The SNR at the output of an optical receiver is defined by

[S/N] = signal power from photocurrent ------ [1]
photodetector noise power + amplifier noise power

To achieve a high SNR, the following conditions should met:

• The photodetector must have a high quantum efficiency to generate a large

signal power.
• The photodetector and amplifier noises should be kept as low as possible.
The sensitivity of a photodetector in an optical fiber communication system
is describable in terms of the minimum detectable optical power. This is
optical power necessary to produce a photocurrent of the same magnitude as
the root mean square of the total noise current.

The interrelationship of the different types of noises affecting the SNR,

consider a simple receiver model and its equivalent circuit shown in figure 4-2. The
photo diode has small series resistance RS, total capacitance Cd consisting of
junction and packaging capacitances, and bias (or load) resistor RL. The amplifier
following the photodiode has an input capacitance Cd & a resistance Ra. For
practical purposes, RS is much smaller than the load resistance RL and can be
neglected.

If a modulated signal of optical power P(t) falls on the detector, the primary
photo current iph(t) generated is

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 6
 OPTICAL FIBER COMMUNICATIONS 2018

Figure 4.3: (a) simple model of a photodetector receiver (b) its equivalent circuit.

This primary current consists of a dc value Ip, which is the average

photocurrent due to the signal power and a signal component i p[t]. For pin
photodiodes the mean square signal current < is2> is
< is2> = < ip2[t]> → [3a]

Whereas for avalanche photodetectors, < is2> is

< is2> = < ip2[t]>M2 → [3b]
where M is the average of the statistically varying avalanche gain.

For a sinusoidal varying input signal of modulation index m, the signal

component < ip2> is of the form
< ip2[t]> = [m2/2]Ip2 →[4] where m is modulation index.

The principal noises associated with quantum noise, dark-current noise

generated in the bulk material of the photodiode and surface leakage current noise.

The quantum or shot noise arises from the statistical nature of the production
and collection of photoelectrons when an optical signal is incident on a
photodetector. The quantum noise current has a mean square value in a bandwidth
B which is proportional to the average value of the photocurrent Ip

<iQ2> = 2qIpBM2F(M) →[5]

where F(M) is a noise figure associated with the random nature of the
avalanche process.
The photodiode dark current is the current that continues to flow through the
bias circuit of the device when no light is incident on the photodiode. This is the
combination of bulk and surface currents. The bulk dark current iDB arises from the
electrons or holes which are thermally generated in the pn junction of the
photodiode. The mean square value of this current is given by
<iDB2> = 2qIDM2F(M)B →[6]
where ID is the primary detector bulk current.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 7
 OPTICAL FIBER COMMUNICATIONS 2018

The surface dark current is also referred to as a surface leakage current of

simply the leakage current. It is dependent on surface defects, cleanliness, bias
voltage & surface area. The mean square value of the surface dark current is given
by
<iDS2> = 2qILB →[7]
Since the dark currents and the signal current are uncorrelated, the total mean
square photodetector noise current

<iN2> = <iQ2> + <iDB2> + <iDS2>

= 2q(Ip + ID)M2F(M)B + 2qILB →[8]

The photodetector load resistor contributes a mean square (Johnson) thermal

noise current
<iT2> = [4KBTB]/RL →[9]
where KB is Boltzmann’s constant and T is the absolute temperature.
𝑆
= _______________________<ip2>M2_____________________ →[10]
𝑁
2q(Ip+ID)M2F(M)B + 2qILB + 4KBTB/RL

DETECTOR RESPONSE TIME:

The response time of photo diode together with its output circuit depends
mainly on the allowing 3 factors:
[a] The transit time of the photo carrier in the depletion region.
[b] The diffusion time of the photo carriers generated outside the depletion
region.
[c] The RC time constant of the photodiode and its associated circuit.
Photo diode parameters responsible for these 3 factors are absorption
coefficient αs, depletion region width w, photodiode junction & package
capacitance, the amplifier capacitance, the detector load resistance, the amplifier
input resistance and photo diode series resistance.

Transit time of photocarriers in the depletion region:

The transit time depends on carrier drift velocity & depletion region width.
td = w/vd where w = depletion region width, vd = carrier drift velocity
Diffusion time can be seen by considering the photodiodes response time.
This response time is described by rise time & fall time of detector output when
detector is illuminated by a step input of optical radiation. The rise time measured
from 10-90%, fall time is measured from 90-10%. For fully depleted photo diode
rise time & fall time are same. They can be different at low bias levels where the
photo diode is not fully depleted. The response time of partially depleted photodiode
is shown in fig. the fast carriers allow the device output to rise to 50 percent of its
maximum value in approximately 1 nsec, but the slow carriers cause a relatively
long delay before the output reaches its maximum value.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 8
 OPTICAL FIBER COMMUNICATIONS 2018

Figure 4.4: Photodiode response to an optical input pulse showing the 10 to

90% rise time and 90 to 10% fall time

Figure 4.5: Response time of photo diode that is not fully depleted.

To achieve a high quantum efficiency, the depletion region width must be much
larger than 1/ αs i.e., w >> 1/ αs
If the depletion layer is too narrow, any carriers created in the undepleted
material would have to diffuse back into the depletion region before they could be
collected. Devices with very thin depletion regions show distinct slow and fast
response components as shown in figure 4.6.
The fast component in the rise time is due to carriers generated in the
depletion region, where as slow component arises from diffusion of carriers that
are created with a distance Ln from the edge of depletion region.
The diffusion of carriers that are with in a distance Ln of the depletion region
edge appears as the slowly decaying tail at the end of the pulse. Also, if w is too
thin, the junction capacitance will become excessive.
This excessiveness will then give rise to a large RC time constant, which
limits the detector response time.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 9
 OPTICAL FIBER COMMUNICATIONS 2018

Figure 4.6: Photodiode pulse response under various detector parameters.

AVALANCHE MULTIPLICATION NOISE:

The avalanche process is statistical in nature, since not every photo
generated carrier pair undergoes the same multiplication. If ‘m’ denotes the
statistically varying, then <m2 > = M2 . where <> symbols denote an ensemble
average.
The noise created in the avalanche process depends on the mean square
gain <m2 >, the noise in an avalanche photodiode can be relatively high.
<m2 > ≅ M2+x where x varies between 0 & 1 depending
on the Photodiode material and structure.
The ratio of the actual noise generated in a avalanche photodiode to the
noise that would exist if all carrier pairs were multiplied by M is called excess
noise factor F
F= <m2 > / M2
For injected electrons and hole, the excess noise factors are

The weighted ionization rate ratio K1 & K2 take into account the non-
uniformity of the gain and the carrier ionization rates in the avalanche region.
They are given by

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 10
 OPTICAL FIBER COMMUNICATIONS 2018

Normally, to first approximation K1 & K2 do not change much with variation in

gain and can be considered constant and equal.

for electron injection, and

for hole injection, where the effective ionization rate ratios are

Figure 4.7: Variation of the electron excess noise factor Fe as a function of

the electron gain for various values of effective ionization rate ratio.

From the empirical relationship for the mean-square gain given by F = Mx

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 11
 OPTICAL FIBER COMMUNICATIONS 2018

Temperature Effect on Avalanche Gain:

The mechanism of Avalanche Photodiode is very temperature-sensitive because the

temperature depends on the electron and hole ionization rates. Temperature
dependence is particularly critical at high bias voltages and if temperature changes
it cause large variations in gain. Temperature-dependent expression is given as:

Where, VB=Breakdown voltage at which M goes to infinity, N varies between 2.5

and 7 depending on the material, V=Va-IMRM, Va=reverse bias voltage applied to
the detector, IM=Multiplied photocurrent, RM=Photodiode series & detector load
resistance
Break down voltage (VB) interms of temperature is
VB(T) = VB(T0)[1+a(T-T0)]
The temperature dependence of the avalnce gain can be given as
n(T) = n(T0)[1+b(T-T0)]
Where, The constants a and b are positive for reach-through avalanche photodiodes

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 12
 OPTICAL FIBER COMMUNICATIONS 2018

COMPARISION OF PHOTODETECTORS:

Some generic operating characteristics of Si, Ge & InGaAs photo

diodes are summarized as follows.

For pin photo diodes:

Parameter Si Ge InGaAs

WL range [nm] 400-1110 800-1650 1100-1700

Responsivity [A/W] 0.4-0.6 0.4-0.5 0.75-0.95

Rise time[ns] 0.5-1 0.1-0.5 0.05-0.5

BW[GHz] 0.3-0.7 0.5-3 1-2

Bias voltage[v] 5 5-10 5

For APD:

Parameter Si Ge InGaAs

WL range [nm] 400-1110 800-1650 1100-1700

Avalanche gain 20-400 50-200 10-40

Rise time[ns] 0.1-2 0.5-0.8 0.1-0.5

Gain-BW[GHz] 100-400 2-10 20-250

Bias voltage[v] 150-400 20-40 20-30

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 13
 OPTICAL FIBER COMMUNICATIONS 2018

FUNDAMENTAL RECEIVER OPERATION:

The design of optical receiver is much more complicated then optical
transmitter because of receiver must be able to detect weak, distorted signal and
make decision. OFC uses Intensity modulated direct detection system [IM-DDS]
that use a binary on-off keyed [OOK] digital signal.
DIGITAL TRANSMISSION SYSTEM:
The function of optical transmitter is to convert the electrical signal to an
optical signal. The optical signal emerging from the LED or LASER transmitter, 1
is represented by a pulse of optical power of duration Tb, whereas 0 is the absence
of any light.
The optical signal that is coupled from the light source to the fiber becomes
attenuated and distorted as it propagates along the fiber waveguide. Upon arriving
at the end of fiber, a receiver converts the optical signal back to an electrical
format. The first element is either a pin or APD, which produces an electrical
current that is proportional to received optical power level. Since this electrical
current is very weak, a front-end amplifier boosts it to a level that can be used by
following electronics.
After amplification signal passes through a low-pass filter to reduce the noise
that is outside of signal bandwidth. In final optical receiver module, a sampling
circuit and decision circuit samples the signal level at the midpoint of each time
slot and compares it with a certain reference voltage known as threshold level. If
the received signal is greater then threshold level, 1 is said to have been received,
other wise 0.

Figure 4.8: Signal path through an optical data link.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 14
 OPTICAL FIBER COMMUNICATIONS 2018

ERROR SOURCES:
Errors in detection mechanism can arise from various noises and
disturbances associated with the signal detection system. The noise sources can be
either external to the system or internal to the system. The noise is caused by the
spontaneous fluctuations of current in electric circuits. Example for this fluctuations
are quantum noise and thermal noise.
Another noise source is photodiode dark current, which is the current
that flow through bias circuit of device when no light is incident on the
photodiode. This is a combination of bulk and surface dark current.
Summary of photodetector noise sources

Figure 4.9: Noise Sources and disturbances in the optical pulse detection mechanism

If detector is illuminated by an optical signal P(t), then the no. of electron -

hole pair generated in a time τ is

Where η is detector quantum efficiency, hv is the photo energy, E is energy received

FRONT-END AMPLIFIER:
Noise sources at the front end of a receiver dominate the sensitivity and
bandwidth the goals generally are to maximize the receiver sensitivity while
maintaining a suitable bandwidth. To achieve these goals, a basic concern in front
end design is what load resistor RL to choose because this parameter affects both the
bandwidth and the noise performance.
UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 15
 OPTICAL FIBER COMMUNICATIONS 2018
Front-end amplifiers used in OFC systems can be classified into 3 categories
• LOW-IMPEDANCE
• HIGH-IMPEDANCE
• TRANSIMPEDANCE
LOW-IMPEDANCE [LZ] PREAMPLIFIER is most straightforward, but not
the optimum preamplifier design. The Photo diode is configured into a LZ
PAMP. Effective input resistance Ra, a bias resistance or load resistance Rb is
used to match the amplifier impedance.
The total load resistance RL = RaRb⁄[Ra+Rb]
The value of bias resistance in conjunction with amplifier input capacitance
decides BW of amp. A small load resistance yields a large BW. But thermal noise
dominates. Thus LZ amplifier can operate over a wide BW, they do not provide
high receiver sensitivity because only a small signal voltage can developed across
the total input impedance. This limits the use of these preamplifiers to special
short-distance applications in which receiver sensitivity is not a major concern.

HIGH-IMPEDANCE [HZ] PREAMPLIFIERS: The thermal noise is inversely

proportional to load resistance .The BW is inversely proportional to load resistance.
Thus for high impedance front end, a high load resistance results in low noise but
gives low receiver band width. Although equalizers sometimes can be implemented
to increase BW, if the BW is much less than the bit rate, then such a front end
amplifier cannot be used.

Figure 4.10: High Input-Impedance Preamplifier

TRANS-IMPEDANCE [TZ] PREAMPLIFIERS:

It is used to largely over comes the drawbacks of high impedance amplifier.
In this case, RL is used as a negative feedback resistor around an inverting
amplifier. Now RL can be large since the negative feedback reduces the effective
resistance seen by the photodiode by a factor of G, so that Rp = R L/[G+1], where G
UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 16
 OPTICAL FIBER COMMUNICATIONS 2018
is the gain of the amplifier. This means that compared to the high impedance design
the trans impedance BW increases by a factor G+1 for the same load resistance.

Figure 4.10: Transimpedance Preamplifier Equivalent Circuit

PROBABILITY OF ERROR:
Bit error rate is defined as ratio of no. of errors Ne occurring over a
certain time interval t by the number of pulses Nt transmitted during this
interval.
𝑁𝑒
BER =
𝑁𝑡
This error rate depends on the SNR at the receiver. To compare the BER
at the receiver we have to know the probability distribution of the signal at the
equalizer output.

Figure 4.11: Probability distributions for 0 and 1 signal levels

The shapes of 2 probability distributions are shown in above figure 4.11. These are

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 17
 OPTICAL FIBER COMMUNICATIONS 2018

Which is probability that the equalizer output voltage is less than v when a logical 1
pulse is sent, and

which is the probability that the output voltage exceeds v when a logical 0 is
transmitted.
If the threshold voltage is vth then the error probability Pe is defined as

The weighting factors a and b are determined by priori distribution of the data.
The probability of error P0(v) is the chance that the output voltage v(t) will fall some
where between vth and α, then

Where the subscript 0 denotes the presence of a 0 bit.

Similarly, probability of error that a transmitted 1 is misinterpreted as a 0

Where subscript 1 denotes the presence of a 1 bit.

If the probabilities of 0 & 1 are equally likely then

THE QUANTUM LIMIT:

In designing an optical system, it is useful to know what is fundamental physical
bounds are on the system performance. Let us see what is bound for photo detector
process. Suppose we have Ideal PD with unity quantum efficiency and no dark
current. Given this condition, it is possible to find minimum received optical power
required for a specific bit error rate performance in a digital system. This minimum
optical power level is known as QUANTUM LIMIT.

UNIT – 4 R.PRATAP SINGH & A.SRINIVASA RAO, E.C.E DEPT., VTA. Page 18