Physical Optics

18 March 2021

Dr. Tsung Sheng Kao

Office: TKP building R416B

Department of Photonics & Institute of Electro-Optical Engineering
National Chiao Tung University
 Schedule
Week Date Content
1 02/25 Course Schedule & Introduction of Photonics
Comprehensive quiz; Development of photonics; Introduction of waves
2 03/04 Maxwell’s Equations & Electromagnetic Waves
Maxwell’s equations; Electromagnetic waves; Electromagnetic optics
3 03/11 Electromagnetic Optics
Electromagnetic optics; course review
4 03/18 Geometric Optics
Geometric optics; Matrix optics
5 03/25 1st exam (mainly on Electromagnetic optics)
6 04/01 Gaussian Optics
Fundamentals; ABCD law; Different orders of Gaussian beams
7 04/08 Reflection & Refraction
Thin film optics; Total internal reflection & Brewster angle; Polarization, birefringence and
nonlinear optics
8 04/15 Optical Coherence and Interference
Coherence; Interference
9 04/22 Resonators
Parallel and spherical mirrors; Gaussian waves in resonators, 2D/3D resonators
10 04/29 2nd exam (mainly on Geometric optics, Gaussian optics, Light on interfaces)
 Ray optics
• Light travels in the optical paths/lines. (optical energy flow, perpendicular to wavefront)
• The characteristic of an optical medium is the refractive index 𝑛 (𝑛 = 𝑐0 Τ𝑐 ≥ 1) 𝑑 𝑟Ԧ
tan 𝜃 = = 𝑟′ ≅ 𝜃
• Light travels in a distance of 𝑑, while the required time 𝑡 = 𝑑/𝑐 = 𝑛𝑑/𝑐0 . 𝑟Ԧ 𝑑𝑧
𝐵
• The optical path (𝑛𝑑) is 𝑆 = ‫𝑟 𝑛 ׬‬Ԧ 𝑑𝑠 . B transverse plane
𝐴

𝑑𝑠
A 𝑆Ԧ
Fermat’s principle
• Light takes the shortest optical path between two points. (shortest time)
• The optical path followed by light between two fixed points, A and B, is an extremum.
• The optical path is defined as the physical length multiplied by the refractive index 𝑛.
𝑡1
1 𝑡1 𝑐 𝑑𝑠 1 𝐵 𝐵 𝑏
𝑇 = න 𝑑𝑡 = න 𝑑𝑡 = න 𝑛𝑑𝑠 → 𝑆 = න 𝑛𝑑𝑠 → 𝛿 න 𝑛 𝑟Ԧ 𝑑𝑠 = 0
𝑡0 𝑐 𝑡0 𝑣 𝑑𝑡 𝑐 𝐴 𝐴 𝑎

Fermat’s principle: Law of reflection Fermat’s principle: Law of refraction

𝑂𝑃𝐿𝐴𝐵 = 𝑛 (−𝑥1 )2 +(𝑦2 − 𝑦1 )2 + 𝑛 (−𝑥3 )2 +(𝑦3 − 𝑦2 )2 𝑂𝑃𝐿𝐴𝐵 = 𝑛 (𝑥2 − 𝑥1 )2 +(𝑦1 )2 + 𝑛 (𝑥3 − 𝑥2 )2 +(−𝑦3 )2

fix 𝑥1 , 𝑦1 , 𝑥3 , 𝑦3 fix 𝑥1 , 𝑦1 , 𝑥3 , 𝑦3
1 1 1 1
𝑑𝑂𝑃𝐿𝐴𝐵 𝑛[2(𝑦2 − 𝑦1 )] 𝑛[2(𝑦3 − 𝑦2 )](−1) 𝑑𝑂𝑃𝐿𝐴𝐵 𝑛 [2(𝑥2 − 𝑥1 )] 𝑛 [2(𝑥3 − 𝑥2 )](−1)
=0= 2
+ 2 (𝑥3 , 𝑦3 ) =0= 2 𝑖
+ 2 𝑡
𝑑𝑦2 (𝑥1 )2 +(𝑦2 − 𝑦1 )2 (𝑥3 )2 +(𝑦3 − 𝑦2 )2 𝑑𝑥2 (𝑥2 − 𝑥1 )2 +(𝑦1 )2 (𝑥3 −𝑥2 )2 +(𝑦3 )2

𝑛[(𝑦2 − 𝑦1 )] 𝑛[(𝑦3 − 𝑦2 )] (𝑥3 , 𝑦3 ) 𝑛𝑖 [(𝑥2 − 𝑥1 )] 𝑛𝑡 [(𝑦3 − 𝑦2 )]

0= − 𝜃𝑟 𝜃𝑡 0= −
(𝑥1 )2 +(𝑦2 − 𝑦1 )2 (𝑥3 )2 +(𝑦3 − 𝑦2 )2 (𝑥2 − 𝑥1 )2 +(𝑦1 )2 (𝑥3− 𝑥2 )2 +(𝑦3 )2
𝜃𝑖
(𝑥2 , 𝑦2 )
0 = 𝑛 sin 𝜃𝑖 − 𝑛 sin 𝜃𝑟 → sin 𝜃𝑖 = sin 𝜃𝑟
0 = 𝑛𝑖 sin 𝜃𝑖 = 𝑛𝑡 sin 𝜃𝑟
𝑛𝑖 𝑛𝑡
(𝑥1 , 𝑦1 ) *Prove the path which takes the min. time
 Planar boundaries
Snell’s law: 𝑛1 sin 𝜃𝑖 = 𝑛2 sin 𝜃𝑡
𝑛2
when 𝜃𝑡 = 90° , 𝜃𝑖 = 𝜃𝑐 = sin−1 ,𝑛 > 𝑛2
𝑛1 1

𝜃𝑖 > 𝜃𝑐 , total internal reflection

𝜃3 𝜃5
sin 𝜃 ≅ 𝜃 − + + ⋯ rays for which angles are small enough → sin 𝜃 ~ tan 𝜃 ~𝜃; cos 𝜃 ~1
3! 5!
Paraxial rays

𝜃 − sin 𝜃
when 𝜃 = 15° ≅ 0.2617 𝑟𝑎𝑑. sin(0.2617) ≅ 0.2587 ∴ ≅ 1.1%
sin 𝜃
 Planar boundaries
Snell’s law: 𝑛1 sin 𝜃𝑖 = 𝑛2 sin 𝜃𝑡
𝑛2
when 𝜃𝑡 = 90° , 𝜃𝑖 = 𝜃𝑐 = sin−1 ,𝑛 > 𝑛2
𝑛1 1

𝜃𝑖 > 𝜃𝑐 , total internal reflection

𝜃3 𝜃5
sin 𝜃 ≅ 𝜃 − + + ⋯ rays for which angles are small enough → sin 𝜃 ~ tan 𝜃 ~𝜃; cos 𝜃 ~1
3! 5!
Paraxial rays

𝜃 − sin 𝜃
when 𝜃 = 15° ≅ 0.2617 𝑟𝑎𝑑. sin(0.2617) ≅ 0.2587 ∴ ≅ 1.1%
sin 𝜃
 Spherical mirrors for paraxial rays
Sign convention
𝑅 > 0, for convex boundary 𝜃
𝑅 < 0, for concave boundary 𝜃
𝑦
𝑑𝑦 𝜃 > 0, for upward rays −𝜃2
𝜃=
𝑑𝑧 𝜃 < 0, for downward rays 𝑃1 𝜃1 𝐶 𝜃0 𝑃2 𝐹
𝑓 > 0, for converging lens
𝑓 < 0, for diverging lens
𝑧1 −𝑅 𝑧2 −𝑅/2 0

𝜃1 = 𝜃0 − 𝜃
ቊ → 𝜃1 + (−𝜃2 ) = 2𝜃0
−𝜃2 = 𝜃0 + 𝜃
𝑧𝐿 𝑧𝑅
𝑂 𝑦 𝑦 𝑦 1 1 2 1 −𝑅
+ =2 → + = = , 𝑓=
𝑧1 𝑧2 (−𝑅) 𝑧1 𝑧2 (−𝑅) 𝑓 2
𝑧1 = ∞, far out on the 𝑧 axis (paraxial approximation)
Paraxial lights approximately focus to on point.
Lens center is the origin
Both 𝑧𝐿 and 𝑧𝑅 are positive
 Paraxial ray refraction

(𝑦1 , 𝑧1 ) −𝜃2 + 𝜃0
𝜃1 −𝜃2
𝜃0
𝑦

𝑃1 −𝜃2 + 𝜃0 𝐶 𝑃2
𝑛1 𝑛2

0 −𝑅
(𝑦2 , 𝑧2 )
𝑛1 (𝜃1 + (−𝜃2 ) + 𝜃0 ) = 𝑛2 𝜃0 Imaging

𝑦 𝑦 𝑛1 𝑛2 𝑛2 − 𝑛1
−𝜃2 + 𝜃0 = → 𝜃0 = + 𝜃2 + ≅
𝑅 𝑅 𝑧1 𝑧2 𝑅
𝑛1 𝑧2
𝑦2 = − 𝑦
𝑦 𝑦 𝑛2 𝑧1 1
𝑛1 (𝜃1 + + 𝜃2 ) = 𝑛2 ( + 𝜃2 )
𝑅 𝑅

𝑛1 𝑛2 − 𝑛1 𝑦
→ 𝜃2 ≅ 𝜃1 −
𝑛2 𝑛2 𝑅
Bi-convex spherical lens

1 1 1
= (𝑛 − 1)( − )
𝑓 𝑅1 𝑅2

1 1 1
+ = Imaging equation (thin lens)
𝑧1 𝑧2 𝑓

𝑧2
𝑦2 = − 𝑦 Magnification
𝑧1 1
 Calculus of variation
y

𝑦𝜀 𝑥 = 𝑦 𝑥 + 𝜀ℎ(𝑥)ℎ 𝑎 = ℎ 𝑏 = 0 → 𝑦′𝜀 𝑥 = 𝑦′ 𝑥 + 𝜀ℎ′(𝑥)

𝑦(𝑥)
To find the extreme value of the function 𝑓(𝑦𝜀 𝑥 , 𝑦′𝜀 𝑥 ; 𝑥) in [𝑎, 𝑏]
𝑦𝜀 (𝑥) 𝑑 𝑏 𝑏
𝜕
x න 𝑓 𝑦𝜀 𝑥 , 𝑦′𝜀 𝑥 ; 𝑥 𝑑𝑥 = න 𝑓 𝑦𝜀 𝑥 , 𝑦′𝜀 𝑥 ; 𝑥 𝑑𝑥 = 0
𝑎 𝑏 𝑑𝜀 𝑎 𝑎 𝜕𝜀

By chain rule

𝜕 𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑦𝜀 𝜕𝑓 𝜕𝑦′𝜀 𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑓
𝑓 𝑦𝜀 𝑥 , 𝑦′𝜀 𝑥 ; 𝑥 = + + = + ℎ(𝑥) + ′ ℎ′(𝑥)
𝜕𝜀 𝜕𝑥 𝜕𝜀 𝜕𝑦𝜀 𝜕𝜀 𝜕𝑦′𝜀 𝜕𝜀 𝜕𝑥 𝜕𝜀 𝜕𝑦𝜀 𝜕𝑦 𝜀

𝑏 𝑏
𝜕𝑓 𝜕𝑥 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
න ( + ℎ(𝑥) + ′ ℎ′(𝑥))𝑑𝑥 = 0 → න ( ℎ(𝑥) + ′ ℎ′(𝑥))𝑑𝑥 = 0
𝑎 𝜕𝑥 𝜕𝜀 𝜕𝑦𝜀 𝜕𝑦 𝜀 𝑎 𝜕𝑦𝜀 𝜕𝑦 𝜀

𝑏 𝑏 𝑏 𝜕𝑓 𝑑 𝜕𝑓
𝜕𝑓 𝜕𝑓 𝑑𝑓 𝑏
𝑑 𝜕𝑓
න ℎ′(𝑥)𝑑𝑥 = න 𝑑ℎ(𝑥) = ℎ 𝑥 |𝑎 − න ℎ(𝑥) ( ′ )𝑑𝑥 ∴ − =0
𝑎 𝜕𝑦 ′
𝜀 𝑎 𝜕𝑦 ′
𝜀
𝑑𝑦 ′
𝜀 𝑎 𝑑𝑥 𝜕𝑦 𝜀 𝜕𝑦𝜀 𝑑𝑥 𝜕𝑦 ′ 𝜀
𝑏 𝑛
𝜕𝑓 𝑑 𝜕𝑓
Euler-Lagrangian equation න ෍[ − ] ℎ𝑖 (𝑥) = 0
𝑎 𝑖=1 𝜕𝑦𝑖 𝑑𝑥 𝜕𝑦 ′ 𝑖

𝑑 𝜕𝑓 𝜕𝑓 𝑑 𝜕𝑓 𝜕𝑓 𝑑 𝜕𝑓 𝜕𝑓
− = 0, − = 0, ⋯ − =0
𝑑𝑥 𝜕𝑦 ′1 𝜕𝑦1 𝑑𝑥 𝜕𝑦 ′ 2 𝜕𝑦2 𝑑𝑥 𝜕𝑦 ′ 𝑛 𝜕𝑦𝑛
 Ray equations
𝑏 𝑏 𝑏
Fermat’s principle 𝛿 න 𝑛 𝑟Ԧ 𝑑𝑠 = 0 → 𝛿 න 𝑛 𝑥 𝑠 , 𝑦 𝑠 , 𝑧(𝑠) 𝑑𝑠 = 𝛿 න 𝑛 𝑥, 𝑦, 𝑧, 𝑥 ′ , 𝑦 ′ , 𝑧 ′ ; 𝑠 𝑑𝑠 = 0
𝑎 𝑎 𝑎

𝑑𝑥 2 𝑑𝑦 2 𝑑𝑧 2
𝑑𝑠 = (𝑑𝑥)2 +(𝑑𝑦)2 +(𝑑𝑧)2 = ( ) +( ) +( ) 𝑑𝑠 = 𝑥′2 + 𝑦′2 + 𝑧′2 𝑑𝑠
𝑑𝑠 𝑑𝑠 𝑑𝑠

∴ 𝑛 𝑥, 𝑦, 𝑧, 𝑥 ′ , 𝑦 ′ , 𝑧 ′ ; 𝑠 = 𝑛(𝑥, 𝑦, 𝑧) 𝑥′2 + 𝑦′2 + 𝑧′2

𝜕𝑓 𝜕𝑛 𝜕𝑓 𝑛𝑥′
∴ = 𝑥′2 + 𝑦′2 + 𝑧′2 , =
𝜕𝑥 𝜕𝑥 𝜕𝑥′ 𝑥′2 + 𝑦′2 + 𝑧′2 𝑥′2 + 𝑦′2 + 𝑧′2 = 1

𝑑 𝜕𝑓 𝜕𝑓 𝑑 𝑛𝑥′ 𝜕𝑓 𝜕𝑛
By Euler-Lagrangian equation − =0 → = = 𝑥′2 + 𝑦′2 + 𝑧′2
𝑑𝑠 𝜕𝑥′ 𝜕𝑥 𝑑𝑠 𝑥′2 + 𝑦′2 + 𝑧′2 𝜕𝑥 𝜕𝑥

𝑑 𝑑𝑥 𝜕𝑛 𝑑 𝑑𝑦 𝜕𝑛 𝑑 𝑑𝑧 𝜕𝑛
𝑛 = 𝑛 = 𝑛 =
𝑑𝑠 𝑑𝑠 𝜕𝑥 𝑑𝑠 𝑑𝑠 𝜕𝑦 𝑑𝑠 𝑑𝑠 𝜕𝑧

𝑑 𝑑𝑟Ԧ
Ray equations: 𝑛 = 𝛻𝑛
𝑑𝑠 𝑑𝑠
 Paraxial ray equations
Along the 𝑧 direction, 𝑑𝑥, 𝑑𝑦 ≪ 𝑑𝑧 𝑑𝑠 = (𝑑𝑥)2 +(𝑑𝑦)2 +(𝑑𝑧)2 ≈ 𝑑𝑧 𝑛 𝑥, 𝑦, 𝑧

𝑑 𝑑𝑟Ԧ 𝑑 𝑑𝑥 𝜕𝑛 𝑑 𝑑𝑦 𝜕𝑛
Ray equations: 𝑛 = 𝛻𝑛 → 𝑛 ≈ 𝑛 ≈
𝑑𝑠 𝑑𝑠 𝑑𝑧 𝑑𝑧 𝜕𝑥 𝑑𝑧 𝑑𝑧 𝜕𝑦

𝑑2 𝑥 𝑑2 𝑦
• In homogeneous medium 𝑛 𝑥, 𝑦, 𝑧 = k =0 =0 along the 𝑧 direction
𝑑𝑧 2 𝑑𝑧 2
y
• Graded-index slab (in the y direction)
𝑑 𝑑𝑦 𝑑𝑛 𝑑2 𝑦 1 𝑑𝑛 𝑑𝑛
By ray equation: 𝑛 = → 2= 𝜃(𝑦 + ∆𝑦) 𝑛 𝑦 +
𝑑𝑦
∆𝑦
𝑑𝑧 𝑑𝑧 𝑑𝑦 𝑑𝑧 𝑛 𝑑𝑦
𝜃(𝑦) 𝑛(𝑦)
If 𝑦(0) and 𝑦(0)
ሶ are known, the trace can be obtained. z
𝜃(𝑦): azimuth angle

By Snell’s law: First order in Taylor series

𝑑𝑛 𝑑𝜃
𝑛 𝑦 cos 𝜃(𝑦) = 𝑛 𝑦 + ∆𝑦 cos 𝜃 𝑦 + ∆𝑦 = 𝑛 𝑦 + ∆𝑦 cos 𝜃 𝑦 − ∆𝑦 sin 𝜃 𝑦
𝑑𝑦 𝑑𝑦

𝑑𝑛 𝑑𝜃
→0= ∆𝑦 cos 𝜃 𝑦 − 𝑛 𝑦 ∆𝑦 sin 𝜃 𝑦 − 𝑓(∆𝑦 2 ) by ∆𝑦 → 0
𝑑𝑦 𝑑𝑦
 Graded-index slab
• Graded-index slab (in the y direction)
𝑑 𝑑𝑦 𝑑𝑛 𝑑2 𝑦 1 𝑑𝑛 𝑑𝑛
By ray equation: 𝑛 = → 2= 𝜃(𝑦 + ∆𝑦) 𝑛 𝑦 +
𝑑𝑦
∆𝑦
𝑑𝑧 𝑑𝑧 𝑑𝑦 𝑑𝑧 𝑛 𝑑𝑦
𝜃(𝑦) 𝑛(𝑦)
If 𝑦(0) and 𝑦(0)
ሶ are known, the trace can be obtained. z
𝜃(𝑦): azimuth angle

By Snell’s law: First order in Taylor series

𝑑𝑛 𝑑𝜃
𝑛 𝑦 cos 𝜃(𝑦) = 𝑛 𝑦 + ∆𝑦 cos 𝜃 𝑦 + ∆𝑦 = 𝑛 𝑦 + ∆𝑦 cos 𝜃 𝑦 − ∆𝑦 sin 𝜃 𝑦
𝑑𝑦 𝑑𝑦

𝑑𝑛 𝑑𝜃
→0= ∆𝑦 cos 𝜃 𝑦 − 𝑛 𝑦 ∆𝑦 sin 𝜃 𝑦 − 𝑓(∆𝑦 2 ) by ∆𝑦 → 0
𝑑𝑦 𝑑𝑦

𝑑𝑛 𝑑𝜃 𝑑𝑦 𝑑𝑛 𝑑𝜃 𝑑𝑦 𝑑 𝑑𝑦 𝑑2 𝑦
= 𝑛 tan 𝜃 and tan 𝜃 ≅ 𝜃 = = 𝑛𝜃 =𝑛 =𝑛 2
𝑑𝑦 𝑑𝑦 𝑑𝑧 𝑑𝑦 𝑑𝑦 𝑑𝑧 𝑑𝑦 𝑑𝑧 𝑑𝑧

𝑑2 𝑦 1 𝑑𝑛
compared with =
𝑑𝑧 2 𝑛 𝑑𝑦
 Parabolic GRIN
• Parabolic GRIN (graded refractive index) slab 𝑛2 𝑦 = 𝑛02 (1 − 𝛼 2 𝑦 2 )
1
If the 𝑛 variation is small, 𝑎𝑦 ≪ 1 → 𝑛 𝑦 ≈ 𝑛0 (1 − 𝛼 2 𝑦 2 )
2

𝑑 2 𝑦 1 𝑑𝑛 𝑛0 2 2 𝑑2 𝑦
→ 2
= 2 2
= −( )𝛼 𝑦 ≈ −𝛼 𝑦 → + 𝛼 2 2
𝑦 =0
𝑑𝑧 𝑛 𝑑𝑦 𝑛 𝑑𝑧 2

trial solution: 𝑦 𝑧 = 𝐴 cos 𝛼𝑧 + 𝐵 sin 𝛼𝑧, and 𝑦 0 = 𝑦0 , 𝜃0 = 𝑦′(0)

𝜃0 𝑑𝑦(𝑧)
𝑦 𝑧 = 𝑦0 cos 𝛼𝑧 + sin 𝛼𝑧 𝜃 𝑧 = = −𝑦0 sin 𝛼𝑧 + 𝜃0 cos 𝛼𝑧
𝛼 𝑑𝑧
slope of the trajectory
Period: 2𝜋/𝛼

1
𝜃0 𝑓≅
𝑛0 𝑑𝛼 sin 𝛼
𝑦0

𝜃
𝑦𝑚𝑎𝑥 = 𝑦02 + ( 0 )2 , and 𝜃𝑚𝑎𝑥 = 𝛼𝑦𝑚𝑎𝑥
𝛼
 2D gradient-index slab
𝑛2 𝑥, 𝑦 = 𝑛02 (1 − 𝛼 2 (𝑥 2 + 𝑦 2 )) if the 𝑛 variation is small, 𝛼 2 (𝑥 2 + 𝑦 2 ) ≪ 1

𝑑2 𝑥 2 2
𝑑2 𝑦
→ 2 + 𝛼 𝑥 = 0, and 2
+ 𝛼2𝑦2 = 0
𝑑𝑧 𝑑𝑧

𝜃𝑥0
𝑥 𝑧 = sin 𝛼𝑧
𝛼
ቊ 𝜃𝑥0
𝑦 𝑧 = sin 𝛼𝑧 + 𝑦0 cos 𝛼𝑧
𝛼
helical ray, skew ray
Physical solutions to GRIN media
• By 𝑑𝑛Τ𝑑𝑇: thermal-optical effect

• By Kerr effect: 𝑛 𝑥, 𝑦, 𝑧 = 𝑛0 + 𝑛′ 𝐼 𝑥, 𝑦, 𝑧

• By ion-exchange diffusion
 Eikonal equation
Scalar surface: 𝑆(𝑟),
Ԧ intensity, potential profiles, etc. 𝑑 𝑟Ԧ
tan 𝜃 = = 𝑟′ ≅ 𝜃
𝑟Ԧ 𝑑𝑧
transverse plane
→ 𝛻𝑆(𝑟)
Ԧ can be exploited to describe the optical paths.

Fulfilling with the Fermat’s principle Ԧ wavefront

𝑆(𝑟):

2 𝜕𝑆 𝜕𝑆 𝜕𝑆
𝛻𝑆 2
= 𝑛 (𝑟)
Ԧ 𝑖𝑒. ( )2 +( )2 +( )2 = 𝑛2 (Eikonal equation)
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝑑 𝑑 𝑟Ԧ
Ray equations: 𝑛 = 𝛻𝑛
𝑑𝑠 𝑑𝑠

The optical path is

𝐵 𝐵 𝐵
2
𝑆 𝑟𝐵 − 𝑆 𝑟𝐴 = න 𝛻𝑆 𝑑𝑠 = න 𝑛𝑑𝑠 → න ( 𝛻𝑆 −𝑛2 )𝑑𝑠 = 0
𝐴 𝐴 𝐴

∴ To obtain the optical path, we may need to know the 𝑛 𝑟 or 𝑆(𝑟).

 Ray-Transfer matrix
In geometric optics, the paraxial approximation is a small-angle approximation used in
Gaussian optics and ray tracing of light through an optical system. (tan 𝜃 ≈ sin 𝜃 ≈ 𝜃)

Input (𝑦1 , 𝜃1 ) Output (𝑦2 , 𝜃2 ) • Consider 2D projection

Optical system • Ray can be defined by position and angle
y • Paraxial approximation

𝑦2 𝑦2 𝐴 𝐵 𝑦1
𝜃2 =
𝜃2 𝐶 𝐷 𝜃1
𝜃1
𝐴 𝐵 𝑛1 𝑛
𝑦1 det = ; (𝐴𝐷 − 𝐵𝐶 = 1 )
z
𝐶 𝐷 𝑛2 𝑛2
𝑧1 𝑧2
𝑦 = 𝑦1 + 𝑑 tan 𝜃1 𝐴 𝐵 1 𝑑
• In free space, light travels a distance (𝑑) → ቊ 2 ∴ =
𝜃2 = 𝜃1 𝐶 𝐷 0 1

𝑦2 = 𝑦1 𝑦2 = 𝑦1
• At the flat interface between 𝑛1 and 𝑛2 → ቊ ቊ
𝑛2 sin 𝜃2 = 𝑛1 sin 𝜃1 𝑛2 𝜃2 = 𝑛1 𝜃1

𝐴 𝐵 1 0
∴ =
𝐶 𝐷 0 𝑛1 Τ𝑛2
 Ray-Transfer matrix
• At the spherical interface between 𝑛1 and 𝑛2 →

𝑦2 = 𝑦1 𝑦2 = 𝑦1
ቊ𝑛 sin(𝜃 +𝜃) = 𝑛 sin(𝜃 +𝜃) → ቊ𝑛 (𝜃 +𝜃) = 𝑛 (𝜃 +𝜃)
2 2 1 1 2 2 1 1

𝑦1 𝑦2 = 𝑦1 𝑛1 (𝑛2 − 𝑛1 )
sin 𝜃 ≈ 𝜃 ≈ ൗ𝑅 → ቊ𝑛 (𝜃 + 𝑦 Τ𝑅) = 𝑛 (𝜃 + 𝑦 Τ𝑅) 𝜃2 = 𝜃1 − 𝑦1
2 2 1 1 1 1 𝑛2 𝑛2 𝑅

𝐴 𝐵 1 0
∴ = − 2 −𝑛1 )
(𝑛 𝑛1
• Thin lens 𝐶 𝐷 𝑛2 𝑅 𝑛2
𝐴 𝐵 1 0
At interface I, 𝑛1 → 𝑛2 ∴ = (𝑛2 −𝑛1 ) 𝑛1
𝐶 𝐷 − 𝑛2 𝑅1 𝑛2

𝐴 𝐵 1 0
At interface II, 𝑛2 → 𝑛1 ∴ = (𝑛1 −𝑛2 ) 𝑛2
𝐶 𝐷 𝑛1 𝑅2 𝑛1

𝐴 𝐵 1 0 1 0 𝐴 𝐵 1 0
∴ = (𝑛1 −𝑛2 ) 𝑛2 (𝑛 −𝑛 )
− 2 1
𝑛1 → = (𝑛1 −𝑛2 ) 𝑛 (𝑛 −𝑛 )
− 2 2 1 1
𝐶 𝐷 𝑛1 𝑅2 𝑛1 𝑛2 𝑅1 𝑛2 𝐶 𝐷 𝑛1 𝑅2 𝑛1 𝑛2 𝑅1

1 0
𝐴 𝐵
𝑛2 𝑅1 𝑛1 − 𝑛2 − 𝑛2 𝑅2 (𝑛1 𝑛2 ) 𝑛2 1 1 1 ∴ = −1 1
∴𝐶= =− −1 + =− 𝐶 𝐷
𝑛1 𝑛2 𝑅1 𝑅2 𝑛1 𝑅1 𝑅2 𝑓 𝑓
 Ray-Transfer matrix
𝐴 𝐵 1 𝑑
• In free space, light travels a distance (𝑑) → =
𝐶 𝐷 0 1

𝐴 𝐵 1 0
• At the flat interface between 𝑛1 and 𝑛2 → =
𝐶 𝐷 0 𝑛1 Τ𝑛2

𝐴 𝐵 1 0
=
• At the spherical interface between 𝑛1 and 𝑛2 → 𝐶 𝐷 − 2 −𝑛1 )
(𝑛 𝑛1
𝑛2 𝑅 𝑛2

1 0
𝐴 𝐵
• Thin lens = −1 1
𝐶 𝐷
𝑓

𝐴 𝐵 1 0
• Flat mirror =
𝐶 𝐷 0 1

𝐴 𝐵 1 0
• Spherical mirror =
𝐶 𝐷 2/𝑅 1
 ABCD matrix
Optical system
𝑦1 𝑦2 𝑦2 = 𝐴𝑦1 + 𝐵𝜃1
𝐴 𝐵
𝜃1 𝜃2 𝜃2 = 𝐶𝑦1 + 𝐷𝜃1
𝐶 𝐷

• 𝑨 = 𝟎 𝑦2 = 𝐵𝜃1 , for arbitrary 𝑦1 • 𝑩 = 𝟎 𝑦2 = 𝐴𝑦1 , for arbitrary 𝜃1

𝑦1
𝜃1 𝑦2 Focal plane(2nd or back) 𝑦2 Imaging matrix
• Focusing system • Thin lens
• Parabolic mirror • Elliptical mirror

• 𝑪 = 𝟎 𝜃2 = 𝐷𝜃1 , for arbitrary 𝑦1 • 𝑫 = 𝟎 𝜃2 = 𝐶𝑦1 , for arbitrary 𝜃1

𝜃2 𝜃2
𝜃1 Parallel in & out 𝑦1 Focal plane (1st or front)
• Refraction prism • Focusing system
• Parabolic mirror
Serial connection of optical components

𝑀1 𝑀2 ⋯ 𝑀𝑁 𝑀 = 𝑀𝑁 𝑀𝑁−1 ⋯ 𝑀2 𝑀1

𝑛
𝑛=1 𝑛=1
1 0 1 𝑑 1 0 1 𝑑/𝑛
𝑀= =
0 𝑛 Τ1 0 1 0 1Τ𝑛 0 1
𝑑

𝑛1 𝑛2 𝑛𝑁
𝑛
𝑛=1 ⋯ 𝑛=1 𝑑𝑖
1 ෍
𝑀= 𝑛𝑖
𝑖=1
0 1
𝑑1 𝑑2 𝑑𝑁
 Formula for image formation
𝑓

1 0 1 𝑑 1 𝑑
𝑀= =
𝑑 −1/𝑓 1 0 1 −1/𝑓 1 − 𝑑 Τ𝑓

𝑓
𝑑2 𝑑1 𝑑2
1 0 1− 𝑑1 + 𝑑2 −
𝐴 𝐵 1 𝑑2 1 1 𝑑1 𝑓 𝑓
𝑀= = =
𝑑1 𝑑2 𝐶 𝐷 0 1 −𝑓 1 0 1 1 𝑑1
− 1−
𝑓 𝑓

If point-to-point imaging occurs, the focal point is irrelevant to the angles. →

𝑑1 𝑑2 1 1 1
𝑑1 + 𝑑2 − =0→ + =
𝑓 𝑑1 𝑑2 𝑓
Formula for image formation
 Recursion relation

𝑀1 𝑀2 ⋯ 𝑀𝑚−1 𝑀𝑚 𝑀𝑚+1
(𝑦0 , 𝜃0 ) (𝑦𝑚+1 , 𝜃𝑚+1 )
(𝑦1 , 𝜃1 ) (𝑦𝑚 , 𝜃𝑚 )

𝑦𝑚 𝐴 𝐵 𝑚 𝑦0 𝑦𝑚+1 𝐴 𝐵 𝑦𝑚 𝑦𝑚+1 = 𝐴𝑦𝑚 + 𝐵𝜃𝑚

= 𝜃𝑚+1 = → ቊ
𝜃𝑚 𝐶 𝐷 𝜃0 𝐶 𝐷 𝜃𝑚 𝜃𝑚+1 = 𝐶𝑦𝑚 + 𝐷𝜃𝑚

𝑦𝑚+1 − 𝐴𝑦𝑚 𝑦𝑚+2 − 𝐴𝑦𝑚+1 𝑦𝑚+1 − 𝐴𝑦𝑚

∴ 𝜃𝑚 = → 𝜃𝑚+1 = = 𝐶𝑦𝑚 + 𝐷𝜃𝑚 = 𝐶𝑦𝑚 + 𝐷( )
𝐵 𝐵 𝐵
∴ 𝑦𝑚+2 − 𝐴𝑦𝑚+1 = 𝐵𝐶𝑦𝑚 + 𝐷 𝑦𝑚+1 − 𝐴𝑦𝑚
→ 𝑦𝑚+2 = 𝐴 + 𝐷 𝑦𝑚+1 − 𝐴𝐷 − 𝐵𝐶 𝑦𝑚 = 2𝑏𝑦𝑚+1 − 𝐹 2 𝑦𝑚 Recursion relation

Trial solution: 𝑦𝑚 = 𝑦0 𝑒 𝑖𝑚𝜑 ∴ 𝑒 𝑖2𝜑 = 2𝑏𝑦0 𝑒 𝑖𝜑 − 𝐹 2 → 𝑒 𝑖𝜑 = 𝑏 ± 𝑏 2 − 𝐹 2

𝐴+𝐷
cos 𝜑 + 𝑖 sin 𝜑 = 𝑏 ± 𝑏 2 − 𝐹 2 → cos 𝜑 = 𝑏 = , sin 𝜑 = 𝑏 2 − 𝐹 2 , 𝑖𝑓𝑏 2 − 𝐹 2 ≤ 0
2
𝐴+𝐷
∴ −1 ≤ cos 𝜑 = 𝑏 = ≤ 1, 𝑜𝑟 𝑏 ≤ 1 stable solution (for laser cavity design)
2
 Periodic optical system
𝑓
1 0 1 𝑑
1 𝑑
𝑀 = −1 1 0 = 1 𝑑
𝑑 1 − 1−
𝑓 𝑓 𝑓

𝑓 𝑓 𝑓

𝑑
⋯ ∴ −2 ≤ 2 − ≤ 2 → 0 ≤ 𝑑 ≤ 4𝑓
𝑑 𝑑 𝑑 𝑓
 Laser resonator

𝑑 𝑑
1− 𝑑(2 − )
1 0 1 0 𝑓1 𝑓1
1 𝑑 1 𝑑
𝑀= −1 1 0 −
1
1 0 = 1 𝑑
1 1 − − 1/𝑓1 1 − 𝑑 ൗ𝑓1 − 𝑑/𝑓2 (2 − )
𝑓2 𝑓1 𝑑 𝑓1
𝑓2 1 −
𝑓1
𝑑 𝑑 𝑑 𝑑2 𝑑 𝑑
−1 ≤ 1 − + 1 − − ≤1→0≤ − − +1≤1
𝑓1 𝑓1 𝑓 2 − 𝑑 4𝑓1 𝑓2 2𝑓1 2𝑓2
2 𝑓1
𝑑 𝑑 𝑑 𝑑
0≤ 1− 1− ≤ 1,0 ≤ 1 − 1− ≤1
2𝑓1 2𝑓2 𝑅1 𝑅2
 Eikonal equation
Scalar surface: 𝑆(𝑟),
Ԧ intensity, potential profiles, etc. 𝑑 𝑟Ԧ
tan 𝜃 = = 𝑟′ ≅ 𝜃
𝑟Ԧ 𝑑𝑧
transverse plane
→ 𝛻𝑆(𝑟)
Ԧ can be exploited to describe the optical paths.

Fulfilling with the Fermat’s principle Ԧ wavefront

𝑆(𝑟):

2 𝜕𝑆 𝜕𝑆 𝜕𝑆
𝛻𝑆 2
= 𝑛 (𝑟)
Ԧ 𝑖𝑒. ( )2 +( )2 +( )2 = 𝑛2 (Eikonal equation)
𝜕𝑥 𝜕𝑦 𝜕𝑧 𝑑 𝑑 𝑟Ԧ
Ray equations: 𝑛 = 𝛻𝑛
𝑑𝑠 𝑑𝑠

The optical path is

𝐵 𝐵 𝐵
2
𝑆 𝑟𝐵 − 𝑆 𝑟𝐴 = න 𝛻𝑆 𝑑𝑠 = න 𝑛𝑑𝑠 → න ( 𝛻𝑆 −𝑛2 )𝑑𝑠 = 0
𝐴 𝐴 𝐴

∴ To obtain the optical path, we may need to know the 𝑛 𝑟 or 𝑆(𝑟).

 Wave optics & Ray optics
𝑑 𝑟Ԧ
tan 𝜃 = = 𝑟′ ≅ 𝜃
The limitation of the wave optics is the ray optics. 𝜆Τ𝑎 → 0. 𝑘(𝑟) ∥ 𝛻𝑆 𝑑𝑧

𝑈 𝒓, 𝑡 = 𝑈(𝒓)𝑒𝑥𝑝(𝑖𝜔𝑡), where 𝑈 𝒓 = 𝑎(𝒓)𝑒𝑥𝑝 −𝑖𝜑 𝒓

𝑟Ԧ
𝑆Ԧ
𝑈 𝒓 = 𝑎(𝒓)𝑒𝑥𝑝 −𝑖𝑘0 𝑆 𝒓 𝜑 𝒓 = 𝑘0 𝑆 𝒓 spatial phase distribution

Helmholtz equation (𝛁2 + 𝑘 2 )𝑈 𝒓 = 0

(𝛁 2 + 𝑘 2 )(𝑎 𝒓 𝑒𝑥𝑝 −𝑖𝑘0 𝑆 𝒓 ) = 0

𝑘02 𝑛(𝒓)2 − 𝛁𝑆(𝒓) 2 𝑎 𝒓 + 𝛁2 𝑎 𝒓

− 𝑖𝑘0 [2𝛁𝑆(𝒓) ∙ 𝛁𝑎 𝒓 + 𝑎(𝒓)𝛁2 𝑆(𝒓)] = 0

2 2
𝜆0 2 𝛁 2 𝑎 𝒓
𝛁𝑆(𝒓) = 𝑛(𝒓) + ( ) =0
2𝜋 𝑎
𝜆0 2 𝛁 2 𝑎 𝒓
if ( ) ≪1 𝛁𝑆(𝒓) 2
= 𝑛(𝒓)2 (Eikonal equation)
2𝜋 𝑎