Linear optical effects

in Z-scan measurements of thin films

Brian M. Patterson, W. Roc White, Todd A. Robbins, and Randall J. Knize

We show that it is possible, in Z-scan measurements of thin films, to obtain data that closely resemble
typical results for nonlinear optical materials, but which actually arise from linear optical effects caused
by sample damage. Z-scan measurements on a silica-based thin film yielded the expected peak–valley
signature of Z-scan data, but subsequent analysis and microscopic examination of the film indicated that
the data resulted from an ablation hole produced in the film when it was near the laser focus. The
resulting spatial variation of the linear refractive index of the film produced a lensing effect that
mimicked the typical Z-scan response. Scalar diffraction theory was used to model the effects of a
spatially varying refractive index and gave results that qualitatively agreed with the Z-scan measure-
ments. © 1998 Optical Society of America
OCIS codes: 310.6860, 190.4400.

The Z-scan technique, developed in 1989 by Sheik-
Bahae and co-workers,1,2 is a simple single-beam
method for characterizing the third-order response of
a nonlinear optical material. Since that time, the
technique has been used to analyze a variety of ma-
terials. The Z-scan measurement provides informa-
tion on both the sign and the magnitude of the
nonlinear refractive index n2 and offers the advan-
tages of simplicity and high sensitivity. Care must
be taken, however, to ensure that the laser-pulse en-
ergy is below the damage threshold of the sample.
This is particularly important for thin films in which
the absorbing medium can be confined to a thickness
of less than a micrometer. In this paper we use the
Z-scan technique to measure the optical response of a
thin copper-doped silica film and obtain data that are
similar to those obtained from bulk nonlinear optical
materials. We show that the Z-scan data from the
thin film actually arise from linear optical effects
caused by laser ablation and not from nonlinear op-
tical effects.
The film used in this study was prepared with a rf
sputtering technique that has been described else-
where.3 Briefly, the film was deposited onto a fused-
The authors are with the Laser and Optics Research Center,
Department of Physics, U.S. Air Force Academy, Colorado 80840.
Received 28 July 1997; revised manuscript received 4 November
1997.
0003-6935y98y101854-04$15.00y0
© 1998 Optical Society of America
silica substrate by use of a composite Cu:SiO2
sputtering target with 10% Cu atomic concentration.
A film thickness of 130~20! nm was determined by
profilometer measurements carried out at the edges
of the deposited film. Separate measurements indi-
cated a linear optical absorption of approximately 1%
at the 532-nm wavelength used in the subsequent
Z-scan measurements. Although we did not carry
out a detailed microstructural analysis, the deposi-
tion technique used to prepare these films typically
produces fine metallic particles dispersed in an amor-
phous SiO2 matrix for low metal concentrations.4
For the case of nanometer-sized particles, these types
of granular materials are of current scientific interest
because of reported enhancements in their nonlinear
optical properties.5–7 Our results, however, are in-
dependent of the exact nature of the microstructure
and should apply to any thin film.
The Z-scan experimental setup is shown schemat-
ically in Fig. 1. For these measurements, we used a
frequency-doubled Coherent Infinity Nd:YAG laser ~l
5 532 nm! with 3-ns pulses operated at a 50-Hz
repetition rate. The laser had Gaussian temporal
and spatial profiles. We produced a top-hat beam
profile by expanding the beam and passing it through
a 6-mm aperture ~not shown!. We focused the top-
hat beam to a 65-mm-diameter focal spot, correspond-
ing to the first Airy pattern minimum, using a 30-cm
focal length lens. For the Z-scan measurements, the
sample was translated along the optical axis of the
beam, and the transmitted light intensity was re-
corded as a function of sample position z with use of

1854 APPLIED OPTICS y Vol. 37, No. 10 y 1 April 1998

Fig. 1. Experimental arrangement for the Z-scan measurement.
The laser beam is focused with a lens of focal length f, and the
detector is located a distance zD away from the focus. The thin
film is moved a distance z relative to the lens focal point, and the
light transmitted through an aperture is measured.
an apertured detector in the far field. Measure-
ments were carried out over a pulse energy range of
5 to 50 mJ. The resulting maximum on-axis inten-
sities therefore varied from approximately 1.5 to 15 3
1012 Wm22 at the beam focus.
Because of self-focusing or self-defocusing effects in
a nonlinear sample, the transmitted light intensity
varies as the sample is moved relative to the focus.
The value of the nonlinear index can be extracted
from this variation in transmission. For our mea-
surements, we calculated the value of n2 using an
analysis developed by Zhao and Palffy-Muhoray for
top-hat beam profiles.8 We determined the nonlin-
ear index in electrostatic units ~esu’s! from the Z-scan
data using2,8
cln0F0
n2 5 , (1)
80p2LeffI0
where n0 is the intensity-independent linear refrac-
tive index, l is the laser wavelength, and I0 is the
maximum on-axis intensity at the focus. The effec-
tive sample thickness Leff is defined in terms of the
linear absorption coefficient a as Leff 5 ~eaL 2 1!ya,
where L is the actual sample thickness. We deter-
mined the phase shift F0 introduced by the nonlinear
sample using8
S D
Tpv
F0 5 2.7 tanh21 , (2)
2.8~1 2 S!1.14
where Tpv is the transmittance difference between
the peak ~maximum! and valley ~minimum! of the
Z-scan data ~see, for example, Fig. 2! and S is the
aperture linear transmittance. The aperture trans-
mittance is the fraction of the incident power that is
transmitted through the detector aperture with the
sample removed. Results of a test run using a 1-mm
cuvette of CS2 are shown in Fig. 2~a!. We measured
n2 for CS2 to be 11.2 ~0.4! 3 10211 esu, which agrees
with the value of 11.2 3 10211 esu reported for the
reorientational Kerr effect in CS2.2,8 The positive
nonlinearity is confirmed by the valley–peak config-
uration of the data.
Z-scan data for the Cu:SiO2 film are shown in Fig.
Fig. 2. Normalized Z-scan data for ~a! the 1-mm cuvette of CS2
and ~b! the Cu:SiO2 thin film. The transmission was normalized
with data obtained at large values of z. The error for the trans-
mission values is approximately twice the size of the data symbol.
2~b! for an incident laser-pulse energy of 30 mJ. The
pattern is inverted relative to the CS2 data indicating
a negative nonlinearity. When the data were ana-
lyzed as described above, a value of 21.5 3 1027 esu
was obtained for n2, in apparent agreement with re-
ports for similar granular materials.5 It is impor-
tant to note, however, that for measurements carried
out on a pristine region of the film, no Z-scan re-
sponse was observed for pulse energies less than a
threshold of approximately 20 mJ. Subsequent
measurements at low pulse energies, carried out af-
ter exposure to energies greater than this threshold,
did show a response and suggested that the film had
been modified permanently by the laser interaction.
Further analysis, described below, indicated that the
transmittance variation observed in these Z-scan
measurements was not a nonlinear optical effect but
resulted from a hole ablated into the film when the
sample was near the beam focus. The resulting spa-
tial dependence of the linear refractive index n0 pro-
duced an intensity variation that mimicked the third-
order nonlinear optical effect.
We carried out additional measurements for both
the CS2 and the Cu:SiO2 film using a range of laser
intensities. The results are summarized in Fig. 3.
Because n2 is a material parameter that does not
depend on laser intensity for low powers, each Z-scan
measurement should yield the same value for n2.
1 April 1998 y Vol. 37, No. 10 y APPLIED OPTICS 1855
Fig. 3. Intensity variation of the magnitude of the nonlinear re-
fractive index n2 calculated from the Z-scan data for CS2 ~filled
circles! and the Cu:SiO2 thin film ~open squares!. The dashed
curve represents an I21 dependence.
For CS2, the magnitude of n2 was independent of
laser intensity to within the precision of the measure-
ment. The film data, however, showed a strong de-
pendence on intensity and thus did not involve a
purely nonlinear optical response. The intensity de-
pendence of these data can be explained by an effec-
tive linear refractive-index change Dn0 resulting from
the presence of an ablation hole. The overall
refractive-index n 5 n0 1 ~n2y2! uEu2 can be written in
terms of Dn0 as
S D
n2 Dn0
n 5 n0 1 1 uEu2.
2 uEu2
The effective nonlinear index of refraction n2eff 5
n2 1 2Dn0yuEu2 contains a true optical nonlinearity
and a linear optical term that depends inversely on
the intensity, which is in qualitative agreement with
the film data shown in Fig. 3.
To confirm the presence of ablation holes, pro-
filometer measurements were carried out on the film
surface in the areas used for the Z-scan measure-
ments. The results for one profilometer measure-
ment are shown in Fig. 4. The figure clearly
indicates a hole in the film resulting from ablation of
Fig. 4. Profilometer scan showing a laser ablation hole in the
Cu:SiO2 thin film.
1856 APPLIED OPTICS y Vol. 37, No. 10 y 1 April 1998
Fig. 5. ~a! Incident field amplitude fa~r! for various positions ~z!
of the film plane. ~b! Calculated power received by the detector as
a function of sample position.
(3)
film material. For the six ablation holes examined,
the hole diameter ranged from approximately 25 to
80 mm, roughly the spot size of the focused laser
beam. The holes were all approximately 130 nm
deep, suggesting that they extend through the entire
thickness of the film.
Ablation of the Cu:SiO2 film results in a lensing
effect similar to a diverging lens, which can produce
a Z-scan signal similar to a material with a negative
n2. This is shown with scalar diffraction theory9 to
follow the propagation of the top-hat beam through
the experimental setup. Assuming azimuthal sym-
metry, the field amplitude fa~r, z! at the film plane
can be determined from the initial top-hat profile
c0~s! before the focusing lens:
k exp F ikr2
G
fa~r, z! 5
2~ f 1 z!
i~ f 1 z! * c0~s! L~s!
S D F
3 J0 k
rs
f1z
exp
iks2
2~ f 1 z!
sds, G (4)
where z is the position of the film relative to the focal
point and L~s! 5 exp~2iks2y2f ! is the phase shift
induced by the initial f 5 30-cm focal length converg-
ing lens. Figure 5~a! shows the resulting field am-
plitude for various positions of the film plane
determined from a frequency-doubled Nd:YAG top-
hat beam with a 6-mm diameter. The top-hat profile
produces several higher-order Fourier components,
which explains the secondary maxima and minima.
In the film plane, the wave undergoes a phase shift
with a radial dependence. For a film with thickness
d, the transmission function T~s! was represented as
H
1 s,a
T~s! 5 , (5)
exp~iu! s $ a
where u 5 ~n 2 1!kd and a is the radius of the hole in
the ablated film. This acts similar to a negative lens
because the index of refraction n ' 1.5 is higher for
the film ~s . a! than for the ablated region ~s , a!
where n 5 1. The wave propagation to the detector
plane yields the field amplitude fD:
*
k sensitivity single beam n2 measurement,” Opt. Lett. 14, 955–
fD~r! 5 fa~s!T~s!
~zD 2 z!
S D F G
rs iks2
3 J0 k exp sds. (6)
zD 2 z 2~zD 2 z!
Here zD represents the distance to the detector, and
the initial constant phase term has been dropped.
Integrating ufD~r!u2 over the detector aperture yields
the transmission. Figure 5~b! shows the results of
calculations by use of experimental parameters, with
a hole diameter of 34 mm, a film thickness of 130 nm,
and the detector at 85 cm. A comparison of the cal-
culated results shown in Fig. 5~b! to the experimental
data shown in Fig. 2~b! suggests that these input
parameters reasonably describe the data.
In conclusion, we have shown that it is possible to
obtain data in Z-scan measurements of thin films
that closely resemble the results observed in thicker
nonlinear optical materials, but which arise from the
modification of the linear refractive index through
the creation of an ablation hole. This conclusion is
supported by both the intensity dependence of the n2
value calculated from the data and the profilometer
measurements of the ablation holes. The response
observed in the Z-scan measurements is in good
agreement with scalar diffraction calculations based
on the presence of an ablation hole. Other than the
cautionary note for Z-scan measurements of thin
films, the creation of this type of microlens could
potentially be useful in data storage or optical pro-
cessing applications. The difference between maxi-
mum and minimum transmission in the actual data
is quite significant considering that it is due to a
linear effect and is thus present even for low-incident
signal levels.
We gratefully acknowledge K. M. Unruh and I.
Weerasekera of the University of Delaware for fabri-
cation of the Cu:SiO2 thin film and the financial sup-
port of the Air Force Office of Scientific Research.
References
1. M. Sheik-Bahae, A. A. Said, and E. W. Van Stryland, “High
957 ~1989!.
2. M. Sheik-Bahae, A. A. Said, T.-H. Wei, D. J. Hagan, and E. W.
Van Stryland, “Sensitive measurements of optical nonlineari-
ties using a single beam,” IEEE J. Quantum Electron. 26, 760 –
769 ~1990!.
3. K. M. Unruh, B. M. Patterson, and S. I. Shah, “Melting behavior
of Snx~SiO2!1002x granular metal films,” J. Mater. Res. 7, 214 –
218 ~1992!.
4. B. Abeles, “Granular metal films,” in Applied Solid State Sci-
ence, R. Wolfe, ed. ~Academic, New York, 1976!, Vol. 6, pp.
1–117.
5. R. H. Magruder, III, R. F. Haglund, Jr., L. Yang, J. E. Wittig,
and R. A. Zuhr, “Physical and optical properties of Cu nanoclus-
ters fabricated by ion implantation in fused silica,” J. Appl.
Phys. 76, 708 –715 ~1994!.
6. K. Fukumi, A. Chayahara, K. Kadono, T. Sakaguchi, Y. Horino,
M. Miya, K. Fujii, J. Hayakawa, and M. Satou, “Gold nanopar-
ticles ion implanted in glass with enhanced nonlinear optical
properties,” J. Appl. Phys. 75, 3075–3080 ~1994!.
7. T. Tokizaki, A. Nakamura, S. Kaneko, K. Uchida, S. Omi, H.
Tanji, and Y. Asahara, “Subpicosecond time response of third-
order optical nonlinearity of small copper particles in glass,”
Appl. Phys. Lett. 65, 941–943 ~1994!.
8. W. Zhao and P. Palffy-Muhoray, “Z-scan technique using top-
hat beams,” Appl. Phys. Lett. 63, 1613–1615 ~1993!.
9. J. W. Goodman, Introduction to Fourier Optics ~McGraw-Hill,
San Francisco, 1968!.
1 April 1998 y Vol. 37, No. 10 y APPLIED OPTICS 1857