1

L2‐1. Beam Optics
(Chs. 1&3  in Saleh & Teich)
 2

Plane waves in homogeneous media

Propagations of rays through homogeneous and inhomogeneous (but slow varying) media,
such as thin lenses, gradient index lenses, dielectric interfaces and mirrors.

- Ray optics
A ray is an idealized narrow beam of light.
A light ray is a line or curve that is perpendicular to the light's wave-fronts
(or called surfaces of constant scalar S(r)).

Boundary of each optical theory

Plane wave Spherical wave
 3

- Position and direction of a ‘paraxial’ ray

In most optical systems that involve lenses made of
homogeneous optical materials with spherical surfaces, the
axis of symmetry (z-axis) is often called the optical axis.

‘Paraxial rays’ are defined as those rays whose angle with the
optical axis  is small enough that sin   tan    .
 Simply we can represent a paraxial ray with a vector,

r  dr
 r  , where r and r = are the position and direction(slope) of the ray.
  dz

- Ray matrices
In paraxial approximation, when all angles are
sufficiently small so that  ≈ sin , the relation
between the input ray and the output ray is
r
linear and can generally be written in the form
 rin , rin 
 rout   A B   rin   rout , rout 
 r    C D   r  
 out     in 
 4

- Ray matrices of simple optical components

1) Free-space propagation 2) Refraction at a planar interface

 rout   rin  rin d  1 d   rin   rout  1 0   rin 

 r     r    0 1   r    r    0 n1  
 
 out   in     in   out   n2   rin

3) Refraction at a spherical interface 4) Transmission through a ‘thin’ lens

 rout   1 0   rin   rout   1 0   rin 

 r      n2  n1   
 r    1 1   rin 
 out    n2 R
n1
  rin 
n2   out   f

5) Reflection from a planar mirror 6) Reflection from a spherical mirror

 rout  1 0   rin   rout   1 0   rin 

 r    0 1  r   r    2 1   rin 
 out     in   out   R
 5

- Matrices of cascaded optical components

A cascade of N optical components or systems whose ray-transfer matrices are M1,
M2, …, MN is equivalent to a single optical system of ray transfer matrix

Example:
1    rin 
N di
A set of parallel transparent plates M i 1 ni
 
r
0 1   in 
Note: does not consider interference

- Matrices of periodic optical components

m
 rm   A B   r0   A B   rm 1 
 r    C D   r    C D   r  
 m    0    m 1 
r  Arm
Combining two equations, rm 1  Arm  Brm  rm  m 1 and rm 1  Crm  Drm ,
we can get a recurrence relation as following: B
A D
rm  2  2brm 1  F 2 rm where b  and F 2  AD  BC  det M
2
 6

- Solution of the recurrence relation

A D
rm  2  2brm 1  F 2 rm where b  and F 2  AD  BC  det M
2
Let’s use a trial solution, rm  r0 h
m

 h 2  2bh  F 2  0  h  b  i F 2  b 2  F exp  i  where   cos 1  b / F 

 rm  ro F m exp  im 
In general, rm  Aro F exp  im   Bro F exp  im 
m m

- Condition for a stable trajectory

In order not to make the ray position diverge,  should be real.
On the other hands, b F  cos   1 . Note that, in many cases, F = det|M|=1.

Unstable trajectory with b/F >1

Stable trajectory
with  cos-1(b/F) = 6/11  period of 11
 7

- A sequence of equally spaced identical lenses

 1 0  1 d   1 d 
M  1   det M  1
 f 1  0 1    1f 1  df 

d
For being a stable waveguide, b / F  b  1  1  0  d  4 f
2f
d 2f d f

- An optical resonators consisting of two spherical mirrors

 1 0  1 d   1 0  1 d  1  2Rd1 B 
M  2     det M  1
  R2 1  0 1    R21 1  0 1   C 1  2d
 4d
 4d 2
R1R2 
 R1 R2 

For being a stable resonant mode,

1  2d 2d  2d   2d    d  d 
1  1    1   1     1  0  1    1    1
2 R1 R2  R1   R2    R1   R2 

Here, we can consider a sequence of a pairs of two lenses.

 d  d 
0  1  1   1
 2 f1  2 f 2 
 8

- Skew rays
For three-dimensional trajectory along the optical axis,
m m
 xm   A B   x0   ym   A B   y0 
m
 rm   A B   r0 
 r    C D   r    x     C D   x  ,  y     C D   y  
 m    0  m    0  m    0
and rm  ro F m exp  im   x m  xo eix F m exp  im 
and 
 y m  yo e F exp  im 
i y m

- Reentrant rays: well defined (not chaotic) resonant or guide modes

m
A B 1 0 
C D   0 1  and m  2n
   

Well-defined mode Chaotic mode

 9

- Wave equation with a complex amplitude light field (convenient!)

1  2U
 U  2 2 0
2
U  r, t   a (r ) exp  j  r   exp  j 2 vt 
c t
- Helmholtz equation: the time-independent form of the wave equation (separation of variables)

Substituting U  r, t   U  r  exp  j 2 vt  into the wave equation gives

2 v 
 2U  k 2U  0 , where k  
c c
- (paraxial) Helmholtz equation: Assumes that light fields vary slowly w.r.t. z-axis

One way of constructing a paraxial wave is to start with a plane wave A exp   jkz  , regarding it
as a “carrier” wave, and modify or “modulate” its complex envelope A, making it a slowly
varying function of position, A(r), so that the complex amplitude of the modulated wave becomes

U  r   A  r  exp   jkz 

A
T2 A  j 2k  0 , where T2   2 x 2   2 y 2
z
 10

- The paraxial Helmholtz equation

The assumption that the field amplitude,  varies slowly with respect to z signifies
that within a distance z = , the change  is much smaller than itself: A
 kA
z

- A simple solution to the paraxial Helmholtz equation: paraboloidal wave


Note that exp ik x 2  y 2  z 2 
  x2  y 2 
 exp  ik  z   ...   when x 2  y 2  z 2
  2z 

Parapoloidal wave: An approximation of the spherical wave

 11

- General solution to the paraxial Helmholtz equation: Gaussian beam

 Ao  x2  y 2 
Ar   exp  ik  , q  z   z  izo
q z  2q  z  
Here, q and zo are called as the q-parameter of the beam and the Rayleigh range.
In order to separate the amplitude and phase of the complex envelop,
we can write the function 1/q(z) as following:
1 1 
 i
q  z  R  z  W 2  z 
- Gaussian beam parameters
  Wo  x2  y 2   x2  y 2 
E  r   aE
ˆ o exp   2  exp  ikz  ik  i  z  
W z  W z   2R  z  
Envelop term Phase term

2
  zo  2  The parameters are functions
W ( z )  Wo 1   z  R( z )  z 1  
z   of zo and  !!!
 zo     In addition, zo can obtained
from the smallest radius (the
Wo 
 zo  ( z )  tan 1  z z  waist radius) Wo !!!
  o
 12

- Beam width of Gaussian beam

  Wo  x2  y 2   x2  y2 
E  r   aEo
ˆ exp   2  exp  ikz  ik  i  z  
W  z  W   
z 2 R  
z 

2
The beam width (beam radius) W ( z )  Wo 1   z 
 zo 
Wo is known as the waist radius.

- Beam divergence
Wo z W  
For z>>zo, W ( z)    o z thus  o  o  
zo zo  Wo 2
Ex) A Gaussian beam laser with a beam spot size of 1 mm and a wavelength of 1.064 um
(Nd.YAG laser) has a beam divergence angle of 0.02o and W(10 m)~3.5 mm.
 13

- Intensity of Gaussian beam

   2 
I  r   E  r   I o 
W 
2
 2  x 2
 y 2
 
 exp   
o

W  z   W  z 
2
 

- The ratio of the power carried within a circle of radius ro

1  2ro2 
P  ro   I  r , z  2 rdr  1  exp   2
ro

P   
 W z
0


~86% of the total power is contained within a circle of the beam radius W2(z)
 14

- Depth of focus

The axial distance within which the beam width is no greater than a factor of √2
times its minimum value, so than the beam area is within a factor 2 of the minimum:
2 Wo2
2 zo 


Ex) at = 632.8nm (He-Ne laser), a spot size of 2Wo = 2 cm  2zo~1 km

and 2Wo = 20 m  2zo ~ 1 mm

- Beam quality
The deviation of the profile of an optical beam from the Gaussian form, Wo and o
2Wm  2 m
M2  M 2   m  o , when Wm  Wo
2Wo  2 o

M2 ≥ 1, the smallest possible divergence angle is preferred for most cases

Commonly available He-Ne laser: M2 < 1.1
High-energy multimode laser: M2 ~ 3-4
15
15
 16

- Wavefronts and curvature radius of the surfaces of constant phase

   x2  y 2 
E  r   aˆ exp  ikz  ik  i  z   Curvature radius in the phase term
 2R  z  

  zo  2 
R  z   z 1    
  z   z = 0 : Planar wavefront and infinite R
z  infinite : RGaussian≈Rspherical

- Accumulated phase retardation

  Wo  x2  y 2   x2  y 2 
E  r   aE
ˆ o exp   2  exp  ikz  ik  i   
z
W z  W   
z 2 R  
z 
At r=0,   0, z   kz    z   kz  tan 1  z / zo 

Accumulated phase difference,

comparing to the plane or the
Guoy effect spherical waves
 17
17

If a Gaussian beam is transmitted through a set of circularly symmetric optical components

aligned with the beam axis, the Gaussian beam remains a Gaussian beam

- Transmission through a thin lens

(i)(ii)(iii)

(i) The phase of the incident wave at the plane of the lens:
(ii) The complex amplitude transmittance of a thin lens of focal length f :
(iii) Then, the phase of the emerging wave becomes:

In general, this kind of calculations can be simplified using the ABCD law (next page)
18
18
 19
19

- The ABCD law for Gaussian beam

The Gaussian beam is obtained from the “paraxial” Helmholtz equation.
 The q-parameters of the incident and transmitted Gaussian beams through a
“paraxial” optical system can be related using the components, A, B, C, and D, of
the ray-transfer matrix M as following:

Aq1  B
q2  , q  z  izo
Cq1  D

When considering the propagation of a Gaussian beam through two media of M1

and M2, we can obtain the following result:

A1q1  B1
A2  B2
q3 
A2 q2  B2

C1q1  D1

 A2 A1  B2C1  q1   A2 B1  B2 D1  AT q1  BT

C2 q2  D2 C A1q1  B1  D  2 1 2 1  1  2 1 2 1  CT q1  DT
C A  D C q  C B  D D
2 2
C1q1  D1
Using this relation, we can
 AT BT   A2 B2   A1 B1 
 MT     M 2 M1 obtain the q-parameter of
CT DT  C2 D2  C1 D1  the transmitted beam.
 20
20

- Examples of the ABCD law: Gaussian beam though a thin lens (revisited)

Let the waist of the incident beam be at distance d1 in front of the lens and
the waist of the output beam at distance d2 behind the lens.
From q ( z )  z  izo , q1  0  iz1o  iz1o and q2  iz2 o
Aiz1o  B ACz1o  BD  iz1o
2
Aq1  B
q2   iz2 o  
Cq1  D Ciz1o  D C 2 z12o  D 2

   
 ACz12o  BD  i z1o  Az2 o C 2 z12o  D 2   0

 ACz12o  BD  0 and z1o  Az2 o C 2 z12o D 0
2

A
 z2 o  z1o where AD  BC  1
D Continue to the next page…
 21
21

- Examples of the ABCD law: Gaussian beam though a thin lens

The transfer matrix is given as

1 d 2   1 0  1 d1  1  d 2 f d1  d 2   d1d 2  f 
M       
0 1   1 f 1  0 1   1 f 1  d1 f 
A d2  f d2  f 2
 z2 o  z1o  z1o or W2 o 
2
W1o  M 2 W12o
D d1  f d1  f
magnification
Several limiting cases:
(a) Ray optics assumption (a zero spot size, z1o=0) B=0  d1  d 2   d1d 2  f  0
1 d1  1 d 2  1 f : well known lens fomular
(b) A plane wave beam: z1o=infinite  d1=infinite  d2 = f
 22
22

- Beam focusing

When a lens is placed at the waist of the incident

beam and thus d1 is zero,
f2
W  2
2
2o W1
2

z1o  f 2
o

When z1o is much lager than the focal length of the lens,

By Taylor series,

D 1 f
W2 o  W1o  W1o
1   f z1o 
2
z1o

 W12o f
When using z1o  , W2 o 
  W1o
 23
23

- Numerical aperture and the waist radius

In many applications using laser focusing, such as

laser scanning, laser printing, compact disk reading
D
and so on, it is desired to generate the smallest
possible spot size.
 shortest , thickest incident beam, shortest f

D 4 f
When taking W1o  , the focus diameter 2W2 o 
2  D
The numerical aperture (NA) of an optical system is a dimensionless number that
characterizes the range of angles over which the system can accept or emit light.
D Here n is the refractive index of
N . A.  n sin   n tan   n
2f media out of the lens.

2 n
 2W2 o 
 N . A.
 The larger the lens diameter, the larger the N.A. and the smaller the focus diameter
 At a given wavelength , we cannot get 2Wo smaller than 2/,
since N.A. cannot be larger than the refractive index of environmental media.
 24
24

Laguerre-Gaussian modes Hermite-Gaussian modes

 25
25

In a laser with cylindrical symmetry, the transverse mode patterns are described by a
combination of a Gaussian beam profile with a Laguerre polynomial. The modes are denoted
TEMpl where p and l are integers labeling the radial and angular mode orders, respectively.

In many lasers, the symmetry of the optical resonator is restricted by polarizing elements such
as Brewster's angle windows. In these lasers, transverse modes with rectangular symmetry
are formed. These modes are designated TEMmn with m and n being the horizontal and vertical
orders of the pattern.

Cylindrical transverse mode patterns TEM(pl) Rectangular transverse mode patterns TEM(mn)
 26
26

- Hermite-Gaussian beams

The Gaussian mode treated in the previous sections has a field variation that depends
only on the axial distance z and the radial distance r. if we do not impose the condition
∂/∂=0 (where  is the azimuthal angle) and we consider spatially constant , the
solution of the paraxial wave equation has solutions in the form of

Wo  x   y 
Elm  x, y , z   Eo Hl  2  H m  2 
W  z   W  z    W  
z Hermite polynomial

 x2  y2 
 exp  ik  ikz  i  i  m  1  z  
 2 q  
z 

Intensity distribution of several low-order Hermite-

Gaussian beams in the transverse plane
Ho  x  1
H1  x   2 x
H 2  x   4x2  2
H 3  x   8 x 3  12 x
...
27
27
28
28