Optical Resonators

An optical cavity or optical resonator has two reflecting mirrors. One of the mirrors is made totally
reflecting, while the other side is partially reflective, so that radiation is emitted from the other side.
Two parallel mirrors placed around the gain medium. Light is reflected by the mirrors back into the
medium and is amplified.

For laser action to occur not only laser amplification is necessary, but also a resonant system where
modes can reach such high intensity that remarkable output can be obtained.

Ring resonator

Optical amplifier

Reflecting mirror

Figure 13 : Ring resonator

The light travels through the optical amplifier and gathers intensity at each roundup. With sufficient
amplification, the losses through the mirrors are compensated and the laser action occurs.

In some cases, only one mode is desired from the system. This is achieved using reflectors that show
strong wavelength dependence. Normally only in small region around the desired wavelength the
resonator exhibits high reflectivity thus supporting modes with that wavelength.

To prevent parasitic modes that can rise at high amplification levels, parasitic reflections are eliminated.
This can be achieved through

(i) The mirrors are placed on the ends of the amplifying region

(ii) The output facets of the amplifier are anti-reflection coated

(iii) The facets are tilled to the Brewster angle which supports one polarization. The other polarization
will be allowed to travel out of the resonators

Optical amplifier
  //

Figure 14: Elimination of one polarization

in the laser with a Brewster-window

Where the sign  illustrate orthogonal and // parallel polarization of the beams

Since only one polarization is reflected, it exhibits lower losses and thus that polarization is dominant in
lasers with Brewster window inside the resonator.

Fabry-Perot Resonator

Optical resonators play an important role in amplifying laser beam. Fabry-Perot resonators (etalons) are
optical cavities which are used for the purpose.

They consist of smooth parallel walls perpendicular to the junction with reflectances R1 and R2 and
separated by a length L . The optical resonators are so placed that the distance between them (L) is
related to the wavelength of the emitted laser (𝜆), given by

𝑚 is an integer and 𝑛𝑟 is the refractive index of the lasing medium.

A Fabry-Perot resonator model is as shown in figure 15.

Input wave from left-hand side is multiplie reflected and amplified. The output on the right-hand side is
collected coherently i.e. we take not only amplitudes but also the phase of each output wave into
account.

For simplicity, consider an electromagnetic wave E i of amplitude incident normally on one of the faces
of the etalon.

Starting from the incident ray or light as shown in figure 16, the amount of light that is transmitted
through the left-hand side mirror is t 1E i where t 1 is transmission coefficient of the left-hand side
L
mirror. It then travels through the etalon between the reflectors gathering some phase   k dz 
0
(where k is the wave number in the etalon or the wave vector of the field ). On the whole way
L
1
absorption occurs given by     i dz (where  is the loss constant inside the etalon) and inside
20
L
1
2 0
the amplification zone, some gain   g dz exists.

Laser amplifier
Figure 15: Fabry-Perot cavity
 t1t 2 r1 r2 Ei e  jk (3l )
'

t1r1r2 Ei e  jk (3l )
'

t1 r1 r2 Ei e  jk ( 2l )
'

t1 r2 Ei e  jk ( 2l )
'

t1r2 Ei e  jk l
'

t1t 2 Ei e  jk l
'

t1 Ei e  jk l
'
Incident ray t1 Ei

L
output rays
Figure 16: Detailed section of Fabry-Perot cavity

Now taking all values of k,  i and 𝑔 as constants and assuming an effective length l , the amplitude on
 l 
the left hand side is changed by a factor exp (   i )  j (k  k )l  .
 2 
On reaching at the right-hand reflector, some portion of the wave passes through the reflector to form
part of the output ray while the other portion is reflected back and the process continues as shown in
figure 16.

Thus, the total output ray is

'

Etot  t1t 2 Ei e  jk l 1  r1 r2 e  jk 2l  r12 r22 e  jk 4l  ...
' '
 (1)

Since equation is a geometrical series, it can be written as

 t1t 2 e  jk l 
'

Etot  Ei   jk ' 2 l
 (2)
1  r1r2 e 

  
Where k '  k  k  j   (3)
 2 

x ' ( )
And where k  k (4)
2n 2

x '' ( ) c2
And   k  ( N 2  N1 ) g (v ) (5)
n2 8v 2 spont
With x being susceptibility [usually in complex form]

Thus expanding equation (2) becomes

  j ( k  k ) l
(   )
l

 
2
t1t 2 e e
Etot  Ei   j 2 ( k  k ) l (  ) l 
(6)
1  r1 r2 e e
 

For a laser medium, it is possible to create a laser action such that

Etot  Ei

If self-oscillation is to be achieved (i.e. at resonance conditions), then from equation (6), the term

1  r1r2 e  j 2( k k )l e (  )l  0 (7)

Or

r1r2 e  j 2( k k )l e (  )l  1
The above equation can be split into two terms i.e. for amplitude and phase

For amplitude,

r1r2 e ( t ( ) )l  1 (8)

Where  t ( ) is the threshold gain coefficient

For phase we have

2k  k ()l  2m where m  1,2,3,... (9)

The amplitude condition of equation (8) can be written as

l
 t ( )    ln(r1 r2 ) (10)
2
Thus, this is the minimum gain required for self-oscillation to occur. To sustain oscillation,

 t ( )   t ( )

Now

c2
  ( N 2  N1 ) g (v )
8v 2 spont

Then the density of inversion distribution at threshold oscillation is

 8v 2 spont 8v 2 spont  1 
N t  ( N 2  N1 ) t     ln(r1 r2 ) (11)
c 2 g (v ) c g (v ) 
2
l 

8 spont  1 
   ln(r1 r2 ) (12)
 g (v ) 
2
l 
Oscillation frequency

From equation (9),

2k  k ()l  2m

Substituting for k from equation (4) have,

 x ' ( )   x ' ( ) 
2 k  k l  2m or kl 1    m (13)
 2n 2   2n 2 

c
Let the frequency v m  m , which is the mth resonant frequency. This is obtained from
2l

 λ c 
l  2 m with λ  v 
 
then l  c m or v  c m
 2v 2l 

2(v0  v) ''
Now x ' ( )  x ( ) where v 0 is the centre frequency of spectrum distribution function
v

kx'' (ω) kx'' (v)

And γ(ω)   or γ(v)  
n2 n2

(v)n 2
Thus x (v)  
''
and
k

2(v0  v)    (v)n 2   v  v 0    (v ) n 2 
x ' (v )     2  
v  k   v   k 

Hence equation (13) becomes

  v  v0   (v)n 2 
 2   
  v  k    m or kl 1   v  v0   (v)   m
kl 1     
 2n 2    v  k 
 
 
 2 2v
But k  or
 c
Then substituting for k in above equation, have

2v   v  v0   (v)    v  v0   (v)  cm

l 1      m or v 1    
c   v  k    v  k  2l

c
But v m  m
2l

  v  v 0   (v ) 
 v 1      vm (14)
  v  k 
Equation (14) can be written as

 (v m )c
v  v m  (v m  v 0 )
2v
Where v m - mth resonant frequency

v 0 -centre frequency of the spectrum distribution function

v - Oscillation frequency

 1 
v - the gain line width   
 g (v ) 
Or

v 12
v  v m  (v m  v 0 )
v
Where

c(1  R)
v 12  with R being reflection coefficient of the resonator and l is the length of the resonator.
2l