Zscan technique using tophat beams

W. Zhao and P. PalffyMuhoray

Citation: Applied Physics Letters 63, 1613 (1993); doi: 10.1063/1.110712

View online: http://dx.doi.org/10.1063/1.110712
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/63/12?ver=pdfcov
Published by the AIP Publishing

This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
150.165.149.104 On: Thu, 14 Nov 2013 18:34:08
 Z-scan technique using top-hat beams
W. Zhao and P. Palffy-Muhoray
Liquid Crystal Institute and Department of Physics, Kent State University, Kent, Ohio 44242
(Received 5 May 1993; accepted for publication 15 July 1993)
Top-hat instead of Gaussian beams are used in Z-scan experiments to measure nonlinear optical
Kerr coefficients of materials. An empirical expression is obtained which allows direct
calculation of the Kerr coefficient from measured peak-valley transmittance differences.
Predictions of the model are compared with Z-scan measurement on CS,. Using top-hat beams,
the sensitivity of Z-scan measurements is a factor of 2.5 greater than for Gaussian beams.

The Z-scan technique developed by the CREOL (Cen-
ter for Research in Electra-Optics and Lasers) group172is a
sensitive yet simple method of accurately measuring
intensity-dependent optical nonlinearities of materials. It is
based on self-focusing or defocusing of an optical beam by
a thin sample. The sample is moved along the propagation
direction (Z axis) of a focused beam, and the variatibn of
the far-field intensity is used to determine the optical non-
linearity. To date, primarily Gaussian beams have been
used in Z-scan measurements. Gaussian beams are not al-
ways readily available, however. Beams with top-hat spa-
tial profile, on the other hand, can be readily obtained by
taking a small portion of the expanded beams from lasers
with arbitrary beam profiles.
We have carried out theoretical and experimental stud-
ies of the Z-scan method using beams with a top-hat spatial
profile, and compared the results with those for Gaussian
beams. Such comparisons provide an estimate of the de-
pendence of Z-scan results on beam profile. The sensitivity
of Z-scan measurements with top-hat beams is about 2.5
times greater than that with Gaussian beams. In the fol-
lowing, we confine our attention to the case where the
sample under study is a nonlinear lossless Kerr medium,
i.e., only the real part of xc3) is nonzero. Such a medium
can be described by an intensity-dependent refractive index
n =nof 71, where I is the light intensity and y is the non-
linear Kerr coefficient to be measured. The value of xc3) is
related to y by x (3) =2&,-gy. The case where xt3) is com-
plex will be considered elsewhere.
A schematic of the Z-scan setup considered is shown
in Fig. 1. The field distribution of a focused top-hat beam
is described by Lommel functions near the focal point z=O
(Ref. 3). If the aperture Al in front of the focusing lens L
has a diameter of d and the focal length of the lens is f,
then the Airy radius of the focused spot of the top-hat
beam is 1.2UF, where F= f/d (see Fig. 1) is the F num-
ber of the system. We choose the waist of the focused
top-hat beam to be w,=/ZF, then the Rayleigh range is
z0=rrw2;l. At the focal plane, the field distribution is
normalized so that the light intensity I(r,z) = 1E(r,z) 12.
Due to its nonlinearity, the sample introduces a phase
modulation proportional to the intensity. If the sample
thickness L is much less than the Rayleigh range z,, the
field distribution at the exit surface of sample, to a good
approximation, is
E,(r,z)=E(r,z)exp[ikL,~lEE(r,z)121, (2)
where z is the position of the sample, k=2n;//Z is the prop-
agation constant of the beam in vacuum, L,,=[l
-exp( --a: L)]/cr is the effective sample thickness, and a: is
the linear absorption coefficient. When the sample is at the
focal point z=O, the nonlinear phase shift at r=O is Q,,
= k&yi,, where I0 is the on-axis intensity at the f&al
point.
To simplify the analysis, we assume that the circular
aperture A2, at distance D from the sample, is far from the
focal point. When the Fresnel number wt/il D is much less
than unity, the field distribution in the far field is propor-
tional to the Fourier transform of field at the exit surface of
the sample, E,( r,z) for values of z of the order of z,. This
allows straightforward calculation of the iield distribution
at the plane of the aperture A2.
Using this scheme, we calculate, for different sample
positions, the relative power transmitted through the aper-
ture A2. The transmittance 7’(z) is shown in Fig. 2. T(z)
is normalized so that T(z) = 1 when the sample is far from
the focal point ( Iz/zol >l) and the nonlinear phase mod-
ulation is negligible. The top-hat beam gives rise to Z-scan
curves which are similar to those with Gaussian beams. In
Fig. 2, Z-scan curves are shown for both beam profiles with
the same value of <PO.As can be seen the peak-valley trans-

(1)
c , To computer
where r is the coordinate perpendicular to the z axis, z is
position of the sample, P is the optical power of the top-hat FIG. 1. Schematic of the Z-scan setup. Al, A2: apertures; L: lens with
beam, and J1 is the’first order Bessel function. E(r,z) is focal length f; PD: photodetector.

This article is copyrighted

Appl. Phys.asLett.
indicated in the
63 (12), 20 abstract.
SeptemberReuse
1993of0003-6951/93/63(12)/1613/3/$6.00
AIP content is subject to the terms at: @
http://scitation.aip.org/termsconditions.
1993 American institute of Physics Downloaded
1613 to IP:
1613
150.165.149.104 On: Thu, 14 Nov 2013 18:34:08
 1.2

1.1

g 1.0

0.9

0.6

-5.0 -2.5 0 2.5 5.0

0.950’ * I 1 I
-5.0 -2.5 0 2.5 5.0
z/z,
FIG. 4. Norm&ed transmittance T(z) as a function of sample position
z with @,=0.49 and S=O.O7. The incident energy is 21.5 PJ.
FIG. 2. Z-scan curves for the top-hat and Gaussian beams.

in diameter and subsequently passing through a 3.175-mm-

mittance difference, Tpv== Tp-- T,, obtained with the top-
hat beam is about 2.5 times greater than that with the
Gaussian beam. This implies that the top-hat beam Z-scan
has a greater sensitivity than a Gaussian beam for measur-
ing y. It also indicates that the peak-valley transmittance
difference, which is the measure of the nonlinearity, can
vary by this factor due to variations in beam profile. The
distance between sample positions corresponding to the
peak and valley in the transmittance is approximately
1.42,. As in the case of Gaussian beams, this distance var-
ies only slightly with Qo.
Another salient feature of the top-hat beam Z scan is
the asymmetry of the relative magnitudes of peak and val-
ley transmittance. For a Gaussian beam, the valley has a
larger magnitude than the peak, that is, Tp - 1> 1 - T,.
The opposite occurs for top-hat beams. This difference is
shown in Fig. 3, where the asymmetry parameter A = ( Tp
- I)/( 1 - T,) is plotted against Q. for the two beams. The
larger peak transmittance may be advantageous in the
measurement of n2 for samples with large nonlinear ab-
sorption.
In the experiment, the beam from a Q-switched,
frequency-doubled Nd:YAG laser was expanded to 20 mm
0.6
0 0.5 1.0 1.5 2.0 2.5
'0
FIG. 3. Asymmetry of the peak-valley transmittance as a function of
a,, for top hat and Gaussian beams.
diam aperture A 1, followed by an f = 146.4 mm lens L.
The resulting beam waist w. at the focal point is 24.5 pm,
and its corresponding Rayleigh range z. is 3.55 mm. Al-
though the dependence of transmittance on the flatness of
the phase front of the top-hat beam across Al has not been
thoroughly examined, we do not expect large variations for
modest distortions of the phase profile. Since the beam
divergence before aperture A 1 is small ( < 0.1 mrad), we
assume the phase protile at Al is essentially constant. The
CS2 sample, mounted on a computer-controlled translation
stage, was 1 mm thick, which is less than one third of the
Rayleigh range. The full width at half-maximum
(FWHM) time T-of the pulse was 6.5 ns. The incident
pulse energy E was measured by a calibrated energy detec-
tor. For a Gaussian temporal profile, the on-axis peak in-
tensity 1, within the sample is lo= ,/~no&/w~r. The
peak power of both the incident and transmitted beams
was measured using fast photodetectors having response
times less than 1 ns. The measured transmittance T(z) is
plotted as a function of z in Fig. 4. The theoretical fit is
obtained by taking the value of y for CS, to be y=3.4
>( lo-l4 cm2/W(n2= 1.2X 10-l’ esu).
To estimate the contribution to n2 from dectrostric-
tion, we note that this contribution is given by
co(ng- l)‘(n~+1)2~/18no(SI) -5X lo-l2 (esu),” where
fi- 10-s Pa-’ is the compressibility and no= 1.5 the re-
fractive index of CS,. The response time is approximately
r,=w&‘,, where vs is the speed of sound. In our experi-
ment 7, -0.2 ps, almost two orders of magnitude longer
than the pulse duration. Therefore, electrostrictive contri-
butions are expected to be less than 2%.
To obtain y from the measured peak-valley transmit-
tance difference Tpu, the relation between <PO,Tpu and the
aperture linear transmittance S must be known. S is the
ratio of power transmitted through the aperture A2 to the
total power incident on the plane of the aperture in the
limit of low intensity.” A simple empirical fit is
3.0
T
$,=2.7 tanh^’ 2.8(1~~)i,14 . (3)
The agreement of this expression with exact numerical re-

Appl.asPhys. Lett.,inVol. W. Zhao and P. Palffy-Muhoray 1614

This article is1614
copyrighted indicated the 63, No. 12,
abstract. 20 September
Reuse 1993is subject to the terms at: http://scitation.aip.org/termsconditions.
of AIP content Downloaded to IP:
150.165.149.104 On: Thu, 14 Nov 2013 18:34:08
 1.5

1.0

0.5

n
"0 0.25 0.50 0.75 1.00 1.25 1.50
Jb

FIG. 5. The nonlinearphase shift @c as function cf the normalized peak-

valley transmittance difference Tp,, with .S=O.O7.
sults is within 1% for S < 1 and 1a0 1 < n-. Once Q,, is
known, y=*dkL,,I,-, can be calculated at once. Values of
@c from Eq. (3) and from exact numerical calculation are
shown in Figs. 5 and 6 for fixed S and Cpe.In these figures
we have used the known value of y for CS, to calculate @e.
In conclusion, we have studied the Z-scan technique
using top-hat beams. An empirical expression to determine
the Kerr coefficient from the peak-valley transmittance dif-
ference is given. Experimental results indicate the possibil-
ity of high accuracy measurements of the nonlinear Kerr
coefficient using top-hat beams. Advantages of this Z-scan
scheme are the relative ease of producing top-hat beams
and the increased sensitivity of the Z-scan method.
FIG. 6. Normalized peak-valley transmittance difference Try as function
of aperture transmittance S with (Pa=O.60.
This work has been supported by DARPA through
DAAB07-88-C-F421 and NSF under ALCOM Grant
DMR-20147. We are grateful to Tai-Huei Wei for helpful
discussions.
‘M. Sheik, A. A. Said, and E. W. Van Stryland, Opt. Lett. 14, 955
(1989).
“M. Sheik, A. A. Said, T. H. Wei, D. J. Hagan, and E. W. Van Stryland,
IEEE J.~Quantum Electron. QE-26, 760 (1990).
‘M. Born and E. Wolf, Principles of Optics, 6th ed. (Pergamon, Oxford,
1980), Sect. 8.8.
4J.-I. Sakai, Phase Conjugate Optics (McGraw-Hill, New York, 1992), p.
135.

This1615
article isAppl.
copyrighted as indicated
Phys. Lett., Vol. 63, inNo.
the12,
abstract. Reuse of1993
20 September AIP content is subject to the terms at: http://scitation.aip.org/termsconditions.
W. Zhao and P. Palffy-Muhoray Downloaded
1615 to IP:
150.165.149.104 On: Thu, 14 Nov 2013 18:34:08