Lecture - 25

Injection
Electroluminescence
Basic definitions
Electroluminescence generates light in response to an electric current passing through
some material

Cathodoluminescence generates light by electron impact or cathode ray

Radioluminescence generates light by other fast particles or high energy radiation

Photoluminescence generates light by excitation from optical radiation

Semiconductor light sources
Electroluminescence refers to luminescence in the presence of an electric field, i.e.,
generation of light in the presence of an electric

Injection electroluminescence refers to generation of light in the presence of an injection

current

We are primarily concerned with injection currents in a forward biased pn junction

Of interest is the active region because under forward bias, Quasi Fermi-levels develop, the
separation between the Quasi-Fermi levels can be controlled by forward biasing
Semiconductor light sources
This will determine or enhance the emission efficiency, or if you recall, the probability of
emission:

rsp(v) = A’ρ(ν) Pe(ν)

By choosing appropriate separation (Efc−Efv) one can change rsp(v), which is the number of
emissions per unit time per unit volume from the material

This parameter can be increased by orders of magnitude to achieve significant emission of

light

Light generation comes from the recombination of electrons and holes in a semiconductor
Rates of generation and recombination
Consider a semiconductor at thermal equilibrium:
Let n0 and p0 be the steady state carrier conc. in the semiconductor

n0= ni + ND ni is the intrinsic conc., ND is the donor ion conc. (assume full
p or n doped ionization)
Similarly for p0= pi + pA ; pi=ni and NA is the acceptor ion conc.

The rate of generation G0 ∝ n0 p0

The rate of recombination R0 is the same as rate of generation, because it is in steady state

Therefore we can write:

G0 ∝ R0 ∝ n0 p0
Rates of generation and recombination
Carriers are continuously generated, making upward and downward transitions, i.e.,
electrons continually recombine with holes

At steady state, in thermal equilibrium, the generation and rate of recombination are equal

The proportionality constant can be written as ‘r’, which is the rate constant

‘r’ depends on whether the material has a direct or indirect bandgap, defect density, traps
in the material, etc.

G0 = R0 = rn0 p0
Quasi-equilibrium
Additional electrons and holes are
beam of light, hv p or n doped created due to hv
quasi equilibrium

‘r’ is the rate of injection, i.e., generation of carriers

Why do I use the word injection? Normally these additional carriers are injected through a
forward biased pn junction
pn diode
Previously we have seen how the carrier concentration profile changes
Equilibrium Under forward bias

Under forward bias more carriers are injected into the junction region, which I have
shown in the R.H.S plot
pn diode

Forward bias
Equilibrium
Forward bias
Equilibrium

Under forward bias, the bands on the n side rise-up

So, in the junction region we have an excess carrier concentration

So, we have n = n0+Δn and p = p0+ Δp
Generation and Recombination
If you think of light incident, then for every electron generated here, one hole will be left
behind in the valence band

Similarly, when you inject current, for every electron entering from the negative side, a hole
will be released

Therefore, Δn =Δp always, whether it is by illumination or by injection current

And we now have the rate of generation as G0 + R

Rate of generation due to thermal energy was G0, and R is the rate of injection of carriers

Therefore, G0+R is equal to the rate constant ‘r’ * np;

G0+R=rnp=r( n0+ Δn)(p0+ Δp)=r n0p0+r(n0+ p0+ Δn)Δn

Carrier injection rate
Where I have used the fact that Δn = Δp

So we get R=r( n0+ p0+ Δn) Δn = Δn/τ; where 1/τ = r( n0+ p0+ Δn)

Rate of injection R is number of carriers injected per unit time per unit volume

Unit analysis: R=r( n0+ p0+ Δn) Δn

#/(cc. sec.) (1/cc)

(1/cc)
‘r’ is in units of cc/sec R = Δn/τ ;where τ is the excess carrier
recombination time
The unit of τ is second
Carrier lifetime
If the injection rate is relatively small, so Δn≪n0 , then this can be written as:

1
τ =
r( n0+ p0)

We are interested in τ because lifetime of carriers is a measurable parameter

τ is material dependent parameter

If Δn is very large (high injection), then the excess carrier recombination time τ will depend
on Δn also
For moderate or low injection rates, it is almost independent
Carrier injection rate
‘r’ is the rate constant for recombination, and there are two types of recombinations possible

Radiative or non-radiative recombination

r = rr + rnr So, recombination rate: rnp=rrnp + rnrnp

1
From before we know: τ ∝
r

1 1
Therefore we can define carrier lifetime as τr ∝ r and τnr ∝ r
r nr
Carrier lifetime
Δn0
Δn(t)

3
2
1
Δn0 1
𝑒

t=0 τ1 τ2 τ3 time

At t = 0, you inject a certain carrier concentration Δn0 and Δp0 , by instantaneous illumination
Carriers recombine, so Δn(t) drops down with time
Carrier recombination rate = 1 > 2 > 3
Carrier lifetime
d (Δn) Δn
The rate of recombination R is =
dt τ
𝑡
−τ
Integration gives: Δn(t)=Δn0𝑒

So τ is the time where the ‘Δn0 ’ excess carrier concentration falls to 1/e

If lifetime τ is very small, this means the recombination rate is very fast, and if lifetime is
very large, the recombination rate is very slow

We have defined radiative and a non-radiative recombination lifetime

In certain materials, these two are indeed of the same order and in others they are highly
different
Carrier lifetime For example, if a particular material has
variation like below:
Δn0

1
Δn0
𝑒

t=0 τnr τr time

τr is very large, τ nr is very small

This means carriers recombine very quickly mostly by non radiative transitions
The contribution to radiative transitions will be minimal in the case of a material which has a
very small τnr
Example
Consider a material where n0 and p0 are ~1017 cm-3.
In silicon τr is of the order of 10 ms and τnr is of the order of 100 ns

In GaAs τr and τnr are of the order of 100 ns

For GaN, τr is 10 ns and τnr is 0.1 ns

For GaAs both radiative and non radiative transitions are equally probable

For silicon, τnr is much smaller, which means most recombination events will be non radiative

We have discussed lifetime, because we are most interested in Internal

Quantum Efficiency (IQE)
Photoluminescence carrier lifetime test
Beware…
Photoconductive decay method Photoconductivity decay method by
Stevenson and Keyes

n-type sample under constant illumination (a),

decay of photoexcited carriers with time (b), and
the experimental setup used for measurements
of the lifetime (c).

Pn = pno + τp G
At time t= T1, light is turned off
This shows hole lifetime in n-type
T1 material

Oscilloscope
Internal Quantum Efficiency
In all semiconductor sources, IQE is the most important parameter which determines the
efficiency of generation of light

It is the ratio of radiative to the total recombinations, per second. It is denoted by ηi

We can therefore write:

τ is the constant r multiplied by p
and n. Factor np is common in all
the places, so they cancel
ηi for common materials
1Τ
100
GaAs: ηi = 1Τ 1 = 0.5
100+ Τ100

10−7
Silicon: ηi = = 10−5
10−2

For silicon this means that for 100,000 recombinations, only one photon will be generated

Similarly you can find ηi for GaN

In general, direct bandgap materials have an ηi in the range 0.2 to 0.9

Indirect band gap materials in general have very low IQE’s, from 10-3 to 10-6
What does this tell us?
If we want to realize a source, one should choose a material which has a direct bandgap,
and has a large ηi

Gallium nitride is also a direct band gap material, but does not have a large ηi

For blue LEDs, InGaN is used, which has ηi of approximately 0.3

For optical sources ηi , the IQE, should be as large as possible, so that there are more
radiative transitions which lead to the generation of photons
Injection electroluminescence
For injection electroluminescence we must consider current ‘i'

We need to relate this to the power generated in the device

From a device engineer’s perspective

For a current ‘i’, what is the optical power generated?

What device and material should we use for an LED?

The pn junction
We use a pn diode because it is very easy to inject current

There is a current i which flows through the diode, so what is the power generated?

The rate of recombination is equal to the rate of injection

The recombination rate per unit time per unit volume is R=Δn/τ

i is current which is charge per unit time, so the number of carriers injected per unit time is
i/q
The pn junction
𝑖 Δn
= 𝑉 where V = volume
𝑞 𝜏

Our interest is to determine the optical power generated

Every radiative recombination will produce one photon but not every transition is radiative

If nph is the number of photons generated per unit time, the power generated is nph times
hν . Number of photons generated is:

Δn
ηi 𝑅𝑉 = ηi 𝑉
𝜏
Optical power generated
𝑜𝑝𝑡 𝑖
𝑃𝑔𝑒𝑛 = ηi𝑅𝑉hv = ηi hv P is measured in watts or j/sec
𝑞

Show that P = joules/sec

This expression can be further simplified: h is constant, and ν =c / λ

ℎ𝑐
We have seen that = 1240 when λ is in nm
𝑞

𝑜𝑝𝑡 ℎ𝑣 𝑜𝑝𝑡 1240

𝑃𝑔𝑒𝑛 = ηi i 𝑃𝑔𝑒𝑛 = η𝑖 i
𝑞 λ

We have seen that ℎ𝑣 is the bandgap energy, i is the applied current

B= material constant