Int. J. Engng Ed. Vol. 22, No. 4, pp.

876881, 2006
Printed in Great Britain.

0949-149X/91 $3.00+0.00
# 2006 TEMPUS Publications.

The CD Lens for Optical Engineering

WEI LEE
Department of Physics, Chung Yuan Christian University, Chung-Li, Taiwan 32023, Republic of China.
E-mail: wlee@phys.cycu.edu.tw
PO-CHUN LIN and CHUN-KAI TSENG*
Taipei Jing-Xing Junior High School, Taipei, Taiwan 11656, Republic of China
Compact disks as a device for the teaching of optics have uses other than traditional diffraction
experiments. Like a Fresnel zone plate that can focus light and form images of distant objects, the
circular grating-like structure of a compact disk (CD) renders itself a chromatic lens. This paper
suggests an interesting experimental activity that can be easily reproduced for early physics classes.
The optical converging nature and chromatic characteristics of focal lengths of the CD lens are
experimentally presented and theoretically interpreted.

tions. The former suggested a distinct use of a

transparent CD, enabling one to observe the
wavelength-dependent focusing of sunlight [5].
Reporting his rough observation of exploiting the
whole disk as a lens, Gluck pointed out that
sharper focusing may be achieved by masking the
disk in order to restrict the range of illuminated
radii [5]. Inspired by Gluck, we carried out this
project as a way of doing hands-on science during
the winter break in the academic year 20022003.
The activity entails both outdoor quantitative
observations and data analyses. We found the
experiment evolved to be not only educational
but also intriguing because of its use of everyday
materials. A brief report on this project in
layman's language, focusing on the operational
aspect with some preliminary results, has been
published in a recent issue of Physics Education
as a Frontline note [15]. In this relatively more indepth paper, we provide more detailed technical
information and more physical insights on our
investigation of the chromatic properties of a
transparent CDR with a given annular opening
of a mask. The averaged wavelengths of major
colour bands of white light are deduced from the
observed data of corresponding focal lengths and
compared with the generally accepted spectral
ranges of colours.

INTRODUCTION
The use of interference-based optical devices is
becoming a very basic and important part of the
optics industry. Solid-state lasers, photonic bandgap devices, chip-scale spectrometers and fiberoptic communications components are just some
examples of modern devices that use interference
to solve specific technological needs. It is highly
desirable that the engineering students can be
provided with related information, ideas, and
materials for improving and updating their background and their understanding of the fundamental optics, and that the optical engineers, as
educators, can be assisted in their teaching and in
their effectiveness in developing the talents of
students.
Since about a decade after the invention of the
first type of compact disks (CDs); i.e., compact
disksdigital audio, numerous science instructors
have devoted increasing attention to the use of a
CD for classroom demonstrations or laboratory
activities on diffraction patterns of a laser beam or
optical spectroscopy of various light sources
[114]. Treated as a reflecting diffraction grating
of 625 lines/mm, a standard CD that has an
angular resolution of 0.16 arc min for the firstorder diffraction [7] provides a simple, as well as
inexpensive, approach for effective displays of
visible emission spectra. Among the above cited
works, Lee and co-workers presented an interesting comparison between diffraction patterns of a
normally or obliquely incident heliumneon laser
beam obtained with a CDread-only memory
(CDROM), unused and data-written CDrecordable (CDR), and a digital versatile disk (DVD
ROM) [3]. Recent articles by Gluck [5] and Knauer
[7] have individually explored the transmission
mode of CD gratings used in optical demonstra-

THEORETICAL BACKGROUND
A grating, framework or latticework having an
even arrangement of long narrow objects with
interstices between them, can be used to disperse
light by interference between wave trains from the
interstices. Let us consider a diffraction grating of
transmission type consisting of a large number of
identical parallel fine grooves each separated by a
constant distance d. If a light beam is incident to
the grating, a multi-order diffraction pattern can

* Accepted August 2005.

876

CD Lens for Optical Engineering

877

be seen in the half-space behind the grating plane.

Principal maxima of the diffracted beams in the far
field originating from a normally incident beam of
wavelength  are found to occur at different angles
 in accordance with the grating equation or
formula [16]

diffracted rays of wavelength  from the impinging

light at radius R form a divergent frustum of conic
shell and a right circular cone shell of radius R
whose apex is located at a common point; i.e., the
focal point, of superposition of diffracted light on
the optical axis. The focal length f is given by

d sin  n;

 dR f 2 R2 1=2

f d=2 11=2 R:

where the integer n specifies the order of diffraction. It is obvious that the angle of diffraction
increases with the increase of the wavelength for a
given order. Note that the relation between wavelength and angle of diffraction sets a constraint on
the highest allowed order to be observed due to the
requirement that sin  n/d  1, indicating that
the shorter the wavelength, the greater the possibility for higher-order diffraction to occur. If the
grating spacing, also known as the grating
constant, d 1.6 m 1600 nm, for instance, the
diffracted red light up to the second order can take
place, while diffraction up to the fourth order of
violet light may be visible.
A standard CD has 20 625 turns of a single
spiral of total length over 5 m, each track spaced
by 1.6 m [17]. The grooves on the CD are, therefore, nearly circular, allowing one to regard the
disk as a circular diffraction grating. Now consider
an expanded and collimated light beam passing
normally through a transparent CD whose surface
is masked with a black, matt paper disk from
which a concentric annular zone with an adequate
width has been cut out to provide an entry for the
light beam. The CD will bring the diffracted light
of a given wavelength to several foci located at
certain distances that are determined by the
allowed orders of diffraction. See Fig. 1.
Based upon the grating formula, the first-order
diffraction is governed by the equation
d sin  :

Because the penetrating light has a circularly

symmetric distribution on the CD, the first-order

Fig. 1. Geometry for the derivation of the focal-length formula;

i.e., Equation (3), used in locating images. This figure illustrates
image formation of diffracted rays of a single component of
normally incident white light.

or
Equation (4) is the focal-length formula for a
masked CD acting as a chromatic lens. The focal
length varies linearly with the radius of the annular
opening on the mask. One can also see that the
focal length is dependent on wavelength, suggesting that the CD possesses obviously chromatic
aberration. The CD that is acting as a lens actually
has the property of forming a real inverted image
of a point on the axis, as does a typical converging
lens, but by a process of diffraction instead of
refraction. It is also due to focusing by wavelength-dependent diffraction that the chromatic
aberration of the CD lens is much more severe
than a traditional lens made of optical glass. While
the focal length of a typical glass lens is slightly
shorter for light of a shorter wavelength, the focal
length of the CD lens decreases as the wavelength
is increased for a fixed radius in use. Because
/d < 1, it is of interest to examine the binomial
expansion of Equation (4). Some simple steps of
manipulation lead to
f Rd=1 =d 2 1=2
d= =d=2 =d3 =8 . . . R; 5
indicating that the CD lens is even more chromatic
than a Fresnel zone plate, which can focus light
through diffraction and whose focal length is
inversely proportional to the wavelength [16].

Fig. 2. Chromaticity of a CD lens from 380 to 780 nm. The

solid curve shows the wavelength-dependent focal length given
by Equation (4) and the dashed curve represents the approximation of order 1 of the binomial expansion of the focallength formula.

878

W. Lee, P.-C. Lin and C.-K. Tseng

Figure 2 displays the theoretical curves of the focal

length in accordance with Equation (4) and the
first-term (namely, order 1) approximation in the
expansion based on Equation (5).
EXPERIMENT
The primary items of which an efficient apparatus can be made for this project include a 25-cmlong hollow cylinder with a radius approximately
equal to that of a CD; i.e., 6 cm, a ruler stick with a
length of at least 20 cm, a white screen attached to
a flattened loop that can slide along the ruler, and
some transparent CDRs, each covered with a
distinct, annularly opened black mask. Figure 3
depicts the complete apparatus we constructed and
named CDscope. A black unglossy inner surface of
the circular cylinder is preferred. We marked a
cross at the optical axis, defined by the axis of the
cylinder, on the upright screen that is movable
inside the cylinder. Although transparent CDs
can be obtained with some extra effort by removing the coatings on the regular CDRs, as
described in the article by Byrne [2], we elected to
use those `protective' disks without the metalized
layer available from bulk packs of CDRs [7].
Each masked CD ought to be easily mounted to
and removed from the front end of the cylinder.
When mounting a CD onto the cylinder, try to
make sure that the center of the CD coincides with
the optical axis. We simply used a tape to fix a CD
temporarily on the cylinder. It is worth mentioning
that, to facilitate an observation, a mechanism
allowing this device to be aimed at the Sun is of
help because the sunlight needs to enter in parallel
along the axis of the cylinder in the experiment. We
constructed the apparatus with a pair of hinges,
allowing us to adjust the angle according to the
altitude of the Sun (see Fig. 3). In addition, with
the aid of some binder clips, we used a piece of
thick black cloth to block the stray light when an
experiment was being performed.
We collected raw data of focal lengths on the flat
roof of a building on campus on four sunny days in
February 2003. Gradually sliding the screen to

Fig. 4. Formation of foci of colour components. Distinct patterns are revealed at different positions of the movable screen.

alter its distance from the CD, observers very

much enjoyed watching the splendid `real-time'
pattern. A typical view yields multiple rings of
continuous spectrum of white light, as shown in
Fig. 4. Expectedly from Fig. 4, various patterns
with distinct orders of colour rings can be observed
at different locations of the movable screen. The
effect of the width of an annular mask opening, b,
was examined. We tried masks with b 0.1, 0.2,
and 1.0 cm and decided to use those with b 0.2
cm as they gave the best images observed. The
contrast of the colourful images depends on the
total throughput of the incident light, which, in
turn, depends on the total area of the annular
window of the mask, while the sharpness of a
focused spot depends on the narrowness of the
width of the annular zone. Although it is desirable
to employ a larger b in order to exploit its greater
light-gathering power, this is, as a matter of fact, at
the expense of the sharpness of the images formed
by constructive interference of the diffracted rays.
All observing runs for R 2.0, 2.5, 3.0, 3.5, 4.0,
4.5, and 5.0 cm were performed between 10:00 a.m.
and 2:00 p. m. With b fixed to be 0.2 cm, patterns
were clearly seen to turn brighter for openings with
greater radii owing to the intensive throughput of
total optical power resulting from a mask with a
greater opening area of 2Rb. The data used for
analysis were acquired on the last day of the
experiment to ensure the reliability after the prior
three-day exercise.

RESULTS AND DISCUSSION

Fig. 3. Schematic of the apparatus, nicknamed CDscope, for

observation of wavelength-dependent focusing of sunlight by a
CD mounted on the front of the tube.

The experimental data each taken as an average

of four different measurements for a given experimental condition are displayed in the first row for
each diameter ring in Table 1. For comparison,
listed in the displaced rows are the accepted ranges
of focal lengths for various colours of light and
values of R according to Equation (4). This
requires the wavelength boundaries of each light
colour to be known. These, adapted from reference
18, are presented in the last column in Table 2. One

CD Lens for Optical Engineering

879

Table 1. Averaged values each taken from four distinct measurements (first row) in comparison with theoretically predicted ranges
of focal lengths for various colour components from normally incident white light (second row). The subscripts R, O, Y, G, B, and
V stand for red, orange, yellow, green, blue, and violet, respectively. Numerical data underlined are those beyond the predicted
lower bound. All units are in cm
Focal lengths of different colours

Radius of
opening R
2.0
2.5
3.0
3.5
4.0
4.5
5.0

fO

fR
4.4
3.7
4.6
5.6
6.5
7.4
8.4

fY

4.5
4.7

5.6

6.0

5.8

6.5

7.0

7.0

7.0

7.7

8.2

7.9

8.9

9.3

9.6

10.2

10.5

11.0

9.3

11.5

11.7

fG

4.8
5.0
6.2
7.5
8.7
10.0
11.2

6.6
7.5
8.5
10.0
10.5
13.0

12.5

fB

5.5
5.6
6.7
8.0
9.4
10.7
12.0
13.4

7.5
8.0
8.7
10.7
11.9
14.0

fV

6.0
6.2

8.5

7.8

9.0

9.3

9.6

10.9

11.5

12.4

13.0

14.0

16.0

15.5

6.5
7.2
9.0
10.8
12.5
14.3
16.1
17.9

8.9
10.1
11.2
14.0
17.0
18.0

8.2
10.2
12.3
14.3
16.4
18.4
20.4

Table 2. Observed wavelengths obs deduced from the fR plot

Color
Red
Orange
Yellow
Green
Blue
Violet

Slope
2.120
2.274
2.486
2.701
3.001
3.528

0.038
0.021
0.043
0.051
0.077
0.077

Standard
deviation
0.368
0.207
0.412
0.495
0.743
0.739

can see from Table 1 that the experimental data are

generally consistent with the theoretical results.
Among the 42 data, there are 20 falling behind
the predictions and none over the predicted ranges.
With the distributions of raw data of measurements taken into account, we attribute this result
to the systematic error caused by either the device
itself or the common operational process. During
the experiment, the observers found that they
seemed to have trouble discerning orange light
from red and that blue and violet were not easily
distinguishable as the rings shrank to little spots.
These problems appear consequently in the table
where the experimental data associated with the
orange, blue, and violet light are not very ideal.
Figure 5 illustrates the relation between the
observed focal length and the radius of the annular
opening according to the experimental data given
in Table 1. Each straight line in the figure, expected
from Equation (4) for a constant wavelength,
represents a least-squares fit to the experimental
data. Note that Equation (4) implies the line to
pass the origin. The relevant results of the fits,
including the slopes, errors, correlation coefficients, and standard deviations, are presented in
Table 2. By use of Equation (3), two of the authors
(PCL and CKT) calculated the observed wavelength obs for each light colour via averaging
seven wavelength values derived from the focal
lengths listed for each column in Table 1. Instead

obs (nm)
683
644
597
555
506
436

10
5
9
9
12
9

Accepted range of
wavelength [18]
630760
595630
560595
490560
430490
380430

of finding experimentally the average of wavelength of each observed colour by substitution

into Equation (3) of the experimental data, the
other author (WL) performed an independent
approach toward the wavelengths. He simply
took full advantage of Fig. 5, which yields the

Fig. 5. Focal length, f, as a function of the radius R of the annular

opening on a mask for various colour components of white light.
Each line associated with a set of seven data points represents the
results of a linear least-squares analysis with a coefficient of
determination r2 > 0.95. Observed wavelengths can be computed
from the slopes in the diagram via Equation (4).

880

W. Lee, P.-C. Lin and C.-K. Tseng

numerical values of slopes with correlation coefficients r > 0.974, as shown in the second column of
Table 2. The results for obs agreed to within their
mutual errors. From the slope of each line, c, the
weighted average of wavelength can be deduced
using
 dc2 11=2 1600c2 11=2 nm,
given by Equation (4). The results are listed in
the obs column in Table 2. The obtained wavelength associated with orange yields a precision of
0.8% and any of the others a precision of 2%. For
comparison, listed also are the accepted ranges of
each major colour in the visible spectrum. They are
found in the last column. It is obvious from Table
2 that the experimental values, within their uncertainties, are in agreement with the accepted results
in general. The observed wavelengths, however, all
seem to suffer an appreciably systematic red shift.
Adopting the midpoint of each wavelength range
as the accepted value, one can see that the estimated wavelengths are 310% higher than the
accepted values except for the red one, which is
2% below. The calculated wavelengths for the
orange and blue are apparently outside the
accepted ranges and the deduced value for the
orange differs by at least 10 nm in comparison
with the upper limit of the wavelengths of orange
light. It is worth mentioning again that all of the
observers, when carrying out the experiment,
encountered a common difficulty of picking the
orange focus out accurately and that they all
suffered from a difficulty in discerning blue and
violet foci. Their similar procedures; i.e., moving
the screen that was, at the beginning, just behind
the CD further away to locate an image, may have
added to the contribution of the apparent tendency
of the shift.
We qualitatively observed second-order focused
images at points closer to the disk. A primary
obstacle, in addition to the images being fainter,
existed that hindered us from taking the measurements effectively. This was due to the third-order
images of shorter wavelengths merging with the

second-order images of longer wavelengths, as

predicted theoretically by the grating formula.
The primary effort in the demonstration is to
determine the wavelength dependence of the focal
length. To emphasize the value of everyday-life
materials in the activity, we elected to rely upon
subjective judgment of the wavelength based on
the perception of colour. Improvements can be
made without changing any existing work or
results. For example, one can make the use of
several narrow-band optical filters that allow a
small number of wavelengths to pass: use `red,'
`green,' or `blue' filters in front of the CD or screen
to measure the location of the focal length.
CONCLUSIONS
We have described an educational demonstration using readily available and inexpensive materials to illustrate how interference may be used to
create a lens. We have also demonstrated that a
CD acts as a chromatic lens whose focal length is
very sensitive to the wavelength. In particular,
assuming that obs constant for all the eyes
judging the major colours, we detail a simplified
analysis which produces an estimated value for
each representative wavelength. There exhibits a
fairly good agreement between the experimental
results and the accepted values, although the
experimental results seem to suffer a systematic
error, giving rise to a bathochromic shift. Two sets
of rings of continuous spectra of white light were
qualitatively observed.
This project is certainly in coherence with the
missions of hands-on science. We emphasize that
the use of everyday materials and outdoor observation, such as those used in the present experiment, has the positive effect of capturing
engineering students' interest. This demonstration
could be used in its existing form or be easily
modified for the classroom.

REFERENCES
1. H. Brouwer, Line spectra using a CD disc, Journal of Chemical Education, 69 (1992), p. 829.
2. A. Byrne, Compact disk spectroscopes revisited! The Physics Teacher, 41 (2003), pp. 144145.
3. H.-Y. Chen, K.-X. Yang and W. Lee, Diffraction experiments with compact disks, Chinese Physics
Education (Taipei), 5 (2001), pp. 4651.
4. M. G. Cornwell, CD means colourful diffraction, Physics Education, 28 (1993), pp. 1214.
5. P. Gluck, Compact disk optics, The Physics Teacher, 40 (2002), pp. 468469.
6. J. E. Kettler, The compact disk as a diffraction grating, American Journal of Physics, 59 (1991),
pp. 367368.
7. T. Knauer, A compact disk transmission spectroscope, The Physics Teacher, 40 (2002),
pp. 466467.
8. H. Kruglak, The compact disc as a diffraction grating, Physics Education, 26 (1991), pp. 255256.
9. H. Kruglak, Diffraction demonstration with a compact disc, The Physics Teacher, 31 (1993), p. 104.
10. R. C. Mebane and T. R. Rybolt, Atomic spectroscopy with a compact disc, Journal of Chemical
Education, 69 (1992), p. 401.
11. C. Noldeke, Compact disc diffraction, The Physics Teacher, 28 (1990), pp. 484485.
12. J. Tellinghuisen, Exploring the diffraction grating using a HeNe laser and a CDROM, Journal of
Chemical Education, 79 (2002), pp. 703704.
13. F. Wakabayashi, K. Hamada and K. Sone, CDROM spectroscope: A simple and inexpensive tool
for classroom demonstrations on chemical spectroscopy, Journal of Chemical Education, 75 (1998),
pp. 15691970.

CD Lens for Optical Engineering

14. V. Zanetti and J. Harris, Spectra of three light sources with a CD, The Physics Teacher, 31 (1993),
pp. 8283.
15. W. Lee, P.-C. Lin and C.-K. Tseng, Compact disc doubles as a chromatic lens, Physics Education,
39 (2004), pp. 384386.
16. E. Hecht, Optics, 4th edn, Addison Wesley, New York (2002), pp. 466 and 488.
17. J. A. Cope, The physics of the compact disc, Physics Education, 28 (1993), pp. 1521.
18. G. J. Shugar and J. T. Ballinger, Chemical Technicians' Ready Reference Handbook, 3rd edn,
McGraw-Hill, New York (1990), p. 715.

Wei Lee earned his M.Sc. degree in electro-optical engineering from National Chiao Tung
University, Taiwan, in 1987 and his Ph.D. degree in physics from the University of
Alabama at Birmingham, AL, in 1993. He was a Visiting Assistant Professor of physics
and astronomy at the University of Toledo, OH, between 1994 and 1997. In 1997, he joined
the faculty at Chung Yuan Christian University, where he currently holds the positions of
Professor of Physics and Director of the Section for Research Promotion at the university's
Office of Research and Development. He has worked in the areas of liquid-crystal
photonics, flat panel displays, laboratory astrophysics, structural vibration analysis and
physics education since 1985. Dr. Lee serves as a councilor of the ROC Taiwan Liquid
Crystal Society and the sole editor of the TLCS Briefs magazine, which is published online
biannually.

Po-Chun Lin was a student at Taipei Jing-Xing Junior High School in the Wen-Shan
District of Taipei City at the time of conducting this study. She is now a student at Taipei
Jing-Mei Girls High School, Taiwan.

Chun-Kai Tseng was a student at Taipei Jing-Xing Junior High School, Taipei, Taiwan at
the time of conducting this study. He is now a senior high school student at Taipei
Municipal Muzha Vocational High School in Taiwan.

881