Ray Optics

Chandan Sengupta
Optics Intro
Geometric Optics
Raytracing
 Reflection
• We describe the path of light as straight-line rays
• “geometrical optics” approach
• Reflection off a flat surface follows a simple rule:
• angle in (incidence) equals angle out
• angles measured from surface “normal”
(perpendicular)
surface normal
same exit ray
incident ray angle

Winter 2012 UCSD: Physics 121; 2012 3

 Reflection, continued
• Also consistent with “principle of least time”
• If going from point A to point B, reflecting off a mirror,
the path traveled is also the most expedient (shortest)
route

A
shortest path;
too long equal angles B

Winter 2012 UCSD: Physics 121; 2012 4

 Hall Mirror
• Useful to think in terms of images

“real” you

mirror only “image” you

needs to be half as
high as you are tall. Your
image will be twice as far from you
Winter 2012
as the mirror.
UCSD: Physics 121; 2012 5
Curved mirrors
• What if the mirror isn’t flat?
• light still follows the same rules, with local surface
normal
• Parabolic mirrors have exact focus
• used in telescopes, backyard satellite dishes, etc.
• also forms virtual image

Winter 2012 UCSD: Physics 121; 2012 6

Refraction
• Light also goes through some things
• glass, water, eyeball, air
• The presence of material slows light’s progress
• interactions with electrical properties of atoms
• The “light slowing factor” is called the index of
refraction
• glass has n = 1.52, meaning that light travels about 1.5
times slower in glass than in vacuum
• water has n = 1.33
• air has n = 1.00028
• vacuum is n = 1.00000 (speed of light at full capacity)

Winter 2012 UCSD: Physics 121; 2012 7

 Refraction at a plane surface
• Light bends at interface between refractive indices
• bends more the larger the difference in refractive index
• can be effectively viewed as a “least time” behavior
• get from A to B faster if you spend less time in the slow medium

Snell’s Law:
1 n1sin1 = n2sin2
n1 = 1.0
n2 = 1.5

2

B
Winter 2012 UCSD: Physics 121; 2012 8
 Driving Analogy
• Let’s say your house is 12 furlongs off the road in
the middle of a huge field of dirt
• you can travel 5 furlongs per minute on the road, but
only 3 furlongs per minute on the dirt
• this means “refractive index” of the dirt is 5/3 = 1.667
• Starting from point A, you want to find the quickest
route:
• straight across (AD)—don’t mess with the road
• right-angle turnoff (ACD)—stay on road as long as possible
• Aangled
B turnoff
C (ABD)—compromise
leg between
dist. the twot@3
t@5
road
AB 5 1 —
dirt AC 16 3.2 —
AD 20 — 6.67
D (house) BD 15 — 5
AD: 6.67 minutes CD 12 — 4
ABD: 6.0 minutes: the optimal path is a “refracted” one
Winter 2012
ACD: 7.2 minutes UCSD: Physics 121; 2012 9
Note: both right triangles in figure are 3-4-5
 Total Internal Reflection
• At critical angle, refraction no longer occurs
• thereafter, you get total internal reflection
n2sin2 = n1sin1  crit = sin1(n1/n2)
• for glass, the critical internal angle is 42°
• for water, it’s 49°
• a ray within the higher index medium cannot escape at
shallower angles (look at sky from underwater…)
incoming ray hugs surface n1 = 1.0
n2 = 1.5
42°

Winter 2012 UCSD: Physics 121; 2012 10

 Refraction in Suburbia
• Think of refraction as a pair of wheels on an axle
going from sidewalk onto grass
• wheel moves slower in grass, so the direction changes

Note that the wheels

move faster (bigger space)
on the sidewalk, slower
(closer) in the grass

Winter 2012 UCSD: Physics 121; 2012 11

 Even gets Total Internal Reflection
Right
• Moreover, this analogy is mathematically
equivalent to the actual refraction phenomenon
• can recover Snell’s law: n1sin1 = n2sin2

Wheel that hits sidewalk starts to go faster,

which turns the axle, until the upper wheel
re-enters the grass and goes straight again

Winter 2012 UCSD: Physics 121; 2012 12

 Reflections, Refractive offset
• Let’s consider a thick piece of glass (n = 1.5), and
the light paths associated with it
• reflection fraction = [(n1 – n2)/(n1 + n2)]2
• using n1 = 1.5, n2 = 1.0 (air), R = (0.5/2.5)2 = 0.04 = 4%

n1 = 1.5 n2 = 1.0
incoming ray
(100%)
image looks displaced
due to jog
96%
8% reflected in two
reflections (front & back)
4%
92% transmitted
4% 0.16%
Winter 2012 UCSD: Physics 121; 2012 13
 Let’s get focused…
• Just as with mirrors, curved lenses follow same
rules as flat interfaces, using local surface normal
A lens, with front and back curved surfaces, bends
light twice, each diverting incoming ray towards
centerline.

Follows laws of refraction at each surface.

Parallel rays, coming, for instance from a specific

direction (like a distant bird) are focused by a convex
(positive) lens to a focal point.

Placing film at this point would record an image of

the distant bird at a very specific spot on the film.
Lenses map incoming angles into positions in the
focal plane.
Winter 2012 UCSD: Physics 121; 2012 14
Cameras, in brief
object pinhole
image at
film plane

In a pinhole camera, the hole is so small that light hitting any particular point
on the film plane must have come from a particular direction outside the camera

object image at
film plane

lens

In a camera with a lens, the same applies: that a point on the film plane
more-or-less corresponds to a direction outside the camera. Lenses have
the important advantage of collecting more light than the pinhole admits
Winter 2012 UCSD: Physics 121; 2012 15
Positive Lenses
• Thicker in middle
• Bend rays toward axis
• Form real focus

Winter 2012 UCSD: Physics 121; 2012 16

Negative Lenses
• Thinner in middle
• Bend rays away from the axis
• Form virtual focus

Winter 2012 UCSD: Physics 121; 2012 17

Raytracing made easier
• In principle, to trace a ray, one must calculate the intersection of each
ray with the complex lens surface, compute the surface normal here,
then propagate to the next surface
• computationally very cumbersome
• We can make things easy on ourselves by making the following
assumptions:
• all rays are in the plane (2-d)
• each lens is thin: height does not change across lens
• each lens has a focal length (real or virtual) that is the same in both directions

Winter 2012 UCSD: Physics 121; 2012 18

Thin Lens Benefits
• If the lens is thin, we can say that a ray through the lens center is
undeflected
• real story not far from this, in fact: direction almost identical, just a jog
• the jog gets smaller as the lens gets thinner

Winter 2012 UCSD: Physics 121; 2012 19

Using the focus condition
real foci virtual foci

s=∞ s=∞
s’ = f s’ = f

s=f s = f
s’ = ∞ s’ = ∞

s=∞ s=∞
s’ = f s’ = f

Winter 2012 UCSD: Physics 121; 2012 20

Tracing an arbitrary ray (positive lens)

1. draw an arbitrary ray toward lens

2. stop ray at middle of lens
3. note intersection of ray with focal plane
4. from intersection, draw guiding (helper) ray
straight through center of lens (thus undeflected)
5. original ray leaves lens parallel to helper
why? because parallel rays on one side of lens meet each
other at the focal plane on the other side

Winter 2012 UCSD: Physics 121; 2012 21

Tracing an arbitrary ray (negative lens)

1. draw an arbitrary ray toward lens

2. stop ray at middle of lens
3. draw helper ray through lens center (thus undeflected) parallel to
the incident ray
4. note intersection of helper with focal plane
5. emerging ray will appear to come from this (virtual) focal point
why? parallel rays into a negative lens appear to diverge from the same
virtual focus on the input side

Winter 2012 UCSD: Physics 121; 2012 22

Image Formation

• Place arrow (object) on left, trace through image:

• 1) along optical axis (no defl.); 2) parallel to axis, goes
through far focus with optical axis ray; 3) through lens
center; 4) through near-side focus, emerges parallel to
optical axis; 5) arbitrary ray with helper
• Note convergence at image position (smaller arrow)
• could run backwards just as well

Winter 2012 UCSD: Physics 121; 2012 23

Notes on Image Formation

• Note the following:

• image is inverted
• image size proportional to the associated s-value: ray 3
proves it
• both s and s’ are larger than f (s = 120; s’ = 80; f = 48)
• Gaussian lens formula (simple form):
Winter 2012 UCSD: Physics 121; 2012 24
Virtual Images

• If the object is inside the focal length (s < f):

• a virtual (and larger) image is formed
• non-inverted
• Ray numbers are same procedure as previous
• This time s’ is negative:
• s = 40; f = 60; s’ = 120
• negative image distances indicate virtual images

Winter 2012 UCSD: Physics 121; 2012 25

The lens-maker’s formula
• We saw the Gaussian lens formula before:

• f is positive for positive lenses, negative for negative lenses

• s is positive on left, s’ is positive on right
• But in terms of the surface properties:

• R1 is for the left surface (pos. if center of curvature to right)

• R2 is for right surface (pos. if center of curvature to right)
• bi-convex (as in prev. examples) has R1 > 0; R2 < 0
• n is the refractive index of the material (assume in air/vac)

Winter 2012 UCSD: Physics 121; 2012 26

Deriving Gaussian Formula from Rays

• Object has height, h; image height = h’

• tangent of ray 3 angle is h/s, so h’ = h(s’/s)
• ray 2 angle is h/f, so h’ = (h/f)(s’  f)
• set the two expressions for h’ equal, and divide by hs’
• the result will pop out
• can do the same trick using virtual images too
Winter 2012 UCSD: Physics 121; 2012 27
Lenses map directions into displacements

• Two objects at infinity an angle  apart produce

distinct spots separated by 
• following geometry,  = f·tan  f· for small 
• hint: look at central rays
• so lens turns angle () into displacement ()

Winter 2012 UCSD: Physics 121; 2012 28

Telescope

• A telescope has an “objective” lens and an

eyepiece
• sharing a focal plane; giving the eye the parallel light it
wants
• Everything goes as ratio of focal lengths: f1/f2
• magnification is just M = 2/1 = f1/f2
• after all: magnification is how much bigger things look
• displacement at focal plane,  = f11 = f22  relation above
• ratio of collimated beam (pupil) sizes: P1/P2 = f1/f2 = M

Winter 2012 UCSD: Physics 121; 2012 29

 Reflector/Refractor Analogy

• For the purposes of understanding a reflecting system, one may

replace with lenses (which we know how to trace/analyze)
• focal length and aperture the same; rays on other side
• for a reflector, f = R/2 [compare to 1/f = (n  1)(1/R1  1/R2) for lens]
• for n = 1.5, R2 = R1 (symmetric lens), f = R
• so glass lens needs twice the curvature of a mirror
Winter 2012 UCSD: Physics 121; 2012 30
Parabolic Example
Take the parabola:
y = x2

Slope is y’ = 2x

Curvature is y’’ = 2

So R = 1/y’’ = 0.5

Slope is 1 (45) at:

x = 0.5; y = 0.25

So focus is at 0.25:
f = R/2

Note that pathlength to focus is the same for depicted ray and one along x = 0
Winter 2012 UCSD: Physics 121; 2012 31
Cassegrain Telescope

• A Cassegrain telescope can be modeled as as positive and negative lens

• eyepiece not shown: only up to focus
• Final focus depends on placement of negative lens
• if |s| = |f2|, light is collimated; if |s| > |f2|, light will diverge
• both s and f2 are negative
• For the Apache Point 3.5 meter telescope, for example:
• f1 = 6.12 m; f2 = 1.60 m; d12 = 4.8 m; s = d12  f1 = 1.32 m
• yields s’ = 7.5 m using 1/s + 1/s’ = 1/f2

Winter 2012 UCSD: Physics 121; 2012 32

Cassegrain focus

• Abstracting mirrors as lenses, then lenses as sticks:

• trace central ray with angle 1
• figure out 2 and then focal length given s’ and d12
• y2 = d121 (adopt convention where 1 is negative as drawn)
• y1 = f21 (f2 is negative: negative lens)
• 2 = (y1  y2)/f2 = 1(f2  d12)/f2
• yf = y2 + 2s’ = 1(d12 + s’(f2  d12)/f2)
• feff = d12 + s’(f2  d12)/f2 = f1s’/s after lots of algebra
• for Apache Point 3.5 meter, this comes out to 35 meters

Winter 2012 UCSD: Physics 121; 2012 33

f-numbers f = D f = 4D
D D

f/1 beam: “fast” f/4 beam: “slow”

• The f-number is a useful characteristic of a lens or system of

lenses/mirrors
• Simply  = f/D
• where f is the focal length, and D is the aperture (diameter)
• “fast” converging beams (low f-number) are optically demanding to
make without aberrations
• “slow” converging beams (large f-number) are easier to make
• aberrations are proportional to 1/2
• so pay the price for going “fast”

Winter 2012 UCSD: Physics 121; 2012 34

f-numbers, compared

• Lens curvature to scale for n = 1.5

• obviously slow lenses are easier to fabricate: less
curvature

Winter 2012 UCSD: Physics 121; 2012 35

Pupils

• Consider two “field points” on the focal plane

• e.g., two stars some angle apart
• The rays obviously all overlap at the aperture
• called the entrance pupil
• The rays are separate at the focus (completely distinct)
• Then overlap again at exit pupil, behind eyepiece
• want your pupil here
• just an image of the entrance pupil satisfying 1/s’ + 1/(f1 + f2) = 1/f2
• size is smaller than entrance pupil by magnification factor
• M = f1/f2; in this picture, f1 = 48; f2 = 12; M = 4; s’ = 15

Winter 2012 UCSD: Physics 121; 2012 36

Pupils within Pupils

• Looking at three stars (red, green, blue) through telescope, eye

position is important
• So is pupil size compared to eye pupil
• dark adapted pupil up to 7 mm diameter (2–3 mm in daylight)
• sets limit on minimum magnification (if you want to use the full aperture)
• 210 mm aperture telescope must have M > 30
• for f/5 scope, means f2 < 35 mm; f/10 scope means f2 < 70 mm
• 3.5-m scope means M > 500; at f/10, f2 < 70 mm

Winter 2012 UCSD: Physics 121; 2012 37

Vignetting
• Rays that don’t make it through an optical system are said to be
vignetted (shadowed)
• maybe a lens isn’t big enough
• maybe your eye’s pupil isn’t big enough, or is improperly placed
• Often appears as a gradual darkening as a function of distance from
the field center
• the farther out you go, the bigger your lenses need to be
• every optical system has a limited (unvignetted) field of view
• beyond this, throughput goes down

Winter 2012 UCSD: Physics 121; 2012 38

Infrared Cold Stop

• An infrared detector is very sensitive to terrestrial heat

• so want to keep off of detector
• if detector located at primary focal plane, it is inundated with emission
from surroundings and telescope structure
• note black lines intersecting primary focal plane
• Putting a “cold” stop at a pupil plane eliminates stray emission
• cool to LN2; image of primary objective onto cold stop
• only light from the primary passes through; detector focal plane then
limits field of view to interesting bit
• Also the right place for filters, who prefer collimated light
Winter 2012 UCSD: Physics 121; 2012 39
Raytrace Simulations
• In Google, type in: phet
• top link is one to University of Colorado physics education page
• on this page, click: go to simulations
• on the left-hand bar, go to: light and radiation
• then click the geometric optics simulation link (picture)
• Can play with lots of parameters
• real and virtual images
• lens radius of curvature, diameter, and refractive index
• see principle rays (ones you’d use to raytrace)
• see marginal rays
• use a light source and screen
• see the effect of two sources

Winter 2012 UCSD: Physics 121; 2012 40

Aberrations: the real world
• Lenses are thick, sin  
• sin    3/6 + 5/120  7/5040 + …
• tan   + 3/3 + 25/15 + 177/315 + …
• Different types of aberration (imperfection)
• spherical aberration
• all spherical lenses possess; parabolic reflector does not
• coma
• off-axis ailment: even aspheric elements have this
• chromatic aberration
• in refractive systems only: refractive index is function of 
• astigmatism
• if on axis, then lens asymmetry; but can arise off-axis in any system
• field curvature/distortion
• detectors are flat: want to eliminate significant field curvature

Winter 2012 UCSD: Physics 121; 2012 41

Spherical Aberration
• Rays at different heights focus at different
points
• Makes for a mushy focus, with a halo
• Positive spherical lenses have positive S.A.,
where exterior rays focus closer to lens
• Negative lenses have negative S.A., as do
plates of glass in a converging beam
• “Overcorrecting” a positive lens (going too neg. S.A.
far in making asphere) results in neg. S.A.

lens side zero S.A.

pos. S.A.

Winter 2012 UCSD: Physics 121; 2012 42

Coma
• Off-axis rays meet at different places
depending on ray height
• Leads to asymmetric image, looking
something like a comet (with
nucleus and flared tail)
• thus the name coma
• As with all aberrations, gets worse
with “faster” lenses
• Exists in parabolic reflectors, even if
no spherical aberration

Winter 2012 UCSD: Physics 121; 2012 43

Chromatic Aberration
• Glass has slightly different
refractive index as a function of
wavelength
• so not all colors will come to
focus at the same place
• leads to colored blur
• why a prism works
• Fixed by pairing glasses with
different dispersions (dn/d)
• typically a positive lens of one
flavor paired with a negative lens
of the other
• can get cancellation of aberration
• also helps spherical aberration to
have multiple surfaces (more
design freedom)

Winter 2012 UCSD: Physics 121; 2012 44

Optical Alignment Techniques
• The performance of an optical system often depends vitally on careful
positioning of the optical elements
• A step-wise approach is best, if possible: aligning as the system is built up
• if using a laser, first make sure the beam is level on the table, and going straight
along the table
• install each element in sequence, first centering the incident beam on the element
• often reflections from optical faces can be used to judge orientation (usually should roughly
go back toward source)
• a lens converts position to direction, so careful translation cross-wise to beam is
important
• orientation is a second-order concern
• Whenever possible, use a little telescope to look through system: the eye is
an excellent judge
Winter 2012 UCSD: Physics 121; 2012 45
Zemax Examples

Winter 2012 UCSD: Physics 121; 2012 46

Lab 4: Raytracing
• While it may not be Zemax, I’ve cobbled together a C-program to do
raytracing of any number of lenses
• restricted to the following conditions:
• ray path is sequential: hitting surfaces in order defined
• ray path is left-to-right only: no backing up
• elements are flat or have conic surfaces
• refractive index is constant, and ignorant of dispersion
• We will use this package to:
• analyze simple lens configurations
• look at aberrations
• build lens systems (beam expanders, telescopes)
• learn how to compile and run C programs (and modify?)
• in conjunction with some geometrical design
Winter 2012 UCSD: Physics 121; 2012 47
Raytracing Algorithm
• Detailed math available on website under Lab Info
• Basically, compute intersection of ray with surface, then apply Snell’s Law

• Can have as many surfaces as you want!

• Must only take care in defining physical systems
• e.g., make sure lens is thick enough for the diameter you need

Winter 2012 UCSD: Physics 121; 2012 48

References and Assignments
• Optics, by Eugene Hecht
• a most excellent book: great pictures, clear, complete
• Text reading (see assignments page for sections):
• Ray Tracing; Paraxial Ray Tracing; other topics of interest
• Apertures, Stops, Pupils; Vignetting
• Geometrical Aberrations & skim 5 types thereof
• Simple and Gal. Telescopes; Laser beam expanders & spatial filters; Lens
aberrations
• Flip through rest of chapter 4 to learn what’s there
• Lab Prep: read raytrace.pdf on raytrace algorithm

Winter 2012 UCSD: Physics 121; 2012 49