August 1, 2004 / Vol. 29, No.

15 / OPTICS LETTERS

1721

Consequences of the KramersKronig relations for light

diffraction on thick gratings
Michael Hckel and Lothar Kador
Institute of Physics and Bayreuther Institut fr Makromoleklforschung, University of Bayreuth, D-95440 Bayreuth, Germany

Carsten Frenz and Hans-Werner Schmidt

Macromolecular Chemistry I, University of Bayreuth, D-95440 Bayreuth, Germany
Received March 16, 2004
Diffraction of monochromatic light on a grating leads to the attenuation of the transmitted beam of diffraction
order zero. In the case of a thick grating the diffraction efficiency, and hence the effective attenuation coefficient, is a fast-varying function of the Bragg mismatch angle. According to Kramers Kronig theory, the transmitted beam encounters a phase shift that also depends on the mismatch angle. This phase shift is measured
for holographic gratings in a photoaddressable block copolymer and compared with analytical calculations.
2004 Optical Society of America
OCIS codes: 050.7330, 050.5080.

The diffraction of monochromatic light on thick phase

and absorption gratings has been studied in detail
both theoretically1,2 and experimentally.3 However,
little attention has been paid to the properties of the
transmitted beam (i.e., a beam of diffraction order
zero). In the case of thick gratings there is only one
additional diffraction order, e.g., order one. Hence,
the diffraction to the first order near the Bragg
angle leads to an angle-dependent attenuation of the
zero-order beam.
According to Kramers Kronig theory, the attenuation (or amplif ication) of light is always connected
with a phase shift. The interaction between a light
wave and matter can be described with a complex susceptibility of the material, whose real and imaginary
parts are connected with the phase shift and the amplitude variation of the wave, respectively. Since the
susceptibility must be an analytical function, its real
and imaginary parts are not independent and are related by the Kramers Kronig relations.
Usually the susceptibility is investigated as a function of the optical frequency, and strong variations
occur near the excitation frequencies (e.g., electronic
or vibrational) of the material. The Kramers Kronig
relations are more general, however, since their origin is purely mathematical and is not related to a specific effect in physics. Therefore, they can be used
with any variable as long as the susceptibility as a
function of this variable fulfills the well-known mathematical conditions. In polar crystals, e.g., which give
rise to second-harmonic generation (SHG) of strong
light fields, the fundamental wave encounters a phase
shift that varies with the SHG phase-mismatch angle.
The physical origins of this effect are the cascaded
x 2 processes of SHG and difference-frequency generation.4 The phase shift can be calculated analytically
with the Kramers Kronig relations when the attenuation of the fundamental wave as a result of SHG is
expressed as a function of the mismatch angle.5,6
In this Letter we investigate a third example,
the phase shift of the transmitted beam in a thick
0146-9592/04/151721-03$15.00/0

holographic grating near the Bragg mismatch angle.

The analytical expression that follows from Kramers
Kronig theory is compared with experimental data
measured on a photoaddressable block copolymer
containing azobenzene chromophores.
According to Kogelnik,1 the diffraction eff iciency of a
thick homogeneous refractive-index grating is given by
h

sinn 2 1 j 2 12 2 ,
1 1 j 2 n 2

(1)

with
n

pdn1
,
lcS cosu0 1 Du12

(2)

pdDu sinc 2 u0 ,
LcS

(3)

cS cosu0 1 Du 2

l cos c .
n0 L

(4)

Here d denotes the grating thickness, n0 is the average refractive index of the material, n1 is its first spatial Fourier component in the grating, l is the vacuum
wavelength of the light, and L is the grating period.
u0 denotes the angle of incidence at Bragg condition,
Du is the Bragg mismatch angle (both measured inside
the material), and c is the angle between the grating
vector and the surface normal. For small diffraction
efficiencies (n ,, 1), Eq. (1) simplif ies to
h n2

sin2 j .
j2

(5)

In this limit n determines the diffraction eff iciency of

the grating, whereas the parameter j, which will be
the variable in the following, is related to the angular
width of the Bragg peak. For thick gratings with high
angular selectivity (i.e., jDuj ,, 1) and u0 0, j depends linearly on Du and n is constant to lowest order.
2004 Optical Society of America

1722

OPTICS LETTERS / Vol. 29, No. 15 / August 1, 2004

The propagation of light in transparent materials

can be described by a complex transmission function,

Fj
fj 2 idj ,

(6)

which relates the electric f ield strengths of the light

wave entering and exiting the sample according to

Et E0 texp2i Fj
.

(7)

The attenuation of the zero-order beam is determined

by diffraction eff iciency h,
dj 2ln1 2 hj12 ,

(8)

which for h ,, 1 can be approximated by

dj

1
hj .
2

(9)

The imaginary part of Fj

then reads as
dj

n 2 sin2 j .
2
j2

(10)

Equation (10) describes the same functional dependence as in the case of cascaded x 2 processes. From
the Kramers Kronig relations the corresponding real
part is calculated as5

n 2 sin2j
fj
21 ,
(11)
2j
2j
and the full complex transmission function reads as6
in 2

exp22ij 1 2ij 2 1 .
Fj
2
2ij2

(12)

It has no singularities and tends to zero for jjj ! `

in the lower complex half-plane; hence the Kramers
Kronig relations are applicable.
The maximum phase shift occurs for j 6p2; it
is related to the maximum diffraction efficiency (at
j 0) by
fmax 6

hmax .
p

(13)

A phase shift of the transmitted beam is also contained

in Kogelniks theory, but it was not calculated explicitly
in his paper.1
The experiments were conducted on an injectionmolded sample of 1.1-mm thickness and a refractive
index of 1.60 6 0.05. It consisted of a polymer
blend composed of 10.9 wt. % of the photoaddressable block copolymer and 89.1 wt. % of polystyrene.
The block copolymer was polystyrene-block-polyhydroxyethyl methacrylate with methoxyazobenzene
side groups and an azo content of 11.3 wt. %.7
Details of the experimental setup for inscribing the
gratings have been published elsewhere.8 Brief ly,
the green light of an Ar1 -ion laser (514.5 nm) was
divided into two components that were superimposed
on the sample under an angle of 30, resulting in a
grating period of 1 mm. The intensity in each of the

writing beams was 125 mWcm2 , and writing times

of 5 20 min were used. Refractive-index changes of
the polymer are caused by cis trans isomerization of
the azobenzene moieties after repeated excitation
relaxation cycles.9
For hologram readout and measurement of the phase
shift we used the interferometer shown in Fig. 1. The
radiation of a He Ne laser (632.8 nm) was split into
two parts, one of which was focused onto the sample,
whereas the second component was directed around the
sample and acted as a reference beam. Lens L1 had a
focal length of 1 m and was required for generating a
plane wave front at the sample.
For the angle-dependent measurement of the attenuation, the intensity of the transmitted beam (order 0)
was recorded with a photodiode. For measuring
the phase shift this photodiode was removed and
the transmitted beam was superimposed with the
reference beam on a CCD camera (Sanyo VCB-3512P)
to yield a bulls eye interference pattern. Lens L2
ensured that both beams had the same spot size
but slightly different divergence on the camera chip.
Equal intensities were adjusted with a half-wave plate
(l2) in the path of the reference beam. Polarizers
P1 P3 ensured vertical polarization in all parts of the
setup. The sample was mounted on a rotating stage
driven by a stepper motor.
The phase shift of the transmitted beam was
obtained from f its of the recorded interference pattern. We assumed Gaussian cross sections of both
the transmitted and the reference beam and allowed
for slightly different center positions, widths, and
intensities. The fringes were well described by a
radial variation of the path difference ~ sin r 2 . To
minimize the computing time we did not fit the whole
two-dimensional image but only a linear section
through the center. The error of the f its was much
smaller than the f luctuations caused by air f lows and
temperature changes.
Rotating a thick sample always causes a phase
shift, since the path length of the light in the material
changes. To minimize this effect we inscribed the
grating such that the Bragg angle was perpendicular
to the surface (i.e., u0 0 and 2L cos c ln0 ).
The background phase shift was measured before
generating the hologram and was subtracted from the
data with the hologram. The signal-to-noise ratio
was improved by averaging 200 measurements for the
background and 50 measurements for the signal each.

Fig. 1. Interferometer for measuring the phase shift of

the zero-order beam.

August 1, 2004 / Vol. 29, No. 15 / OPTICS LETTERS

Fig. 2. (a) Phase shift of the transmitted light and

(b) diffraction eff iciency as a function of the Bragg mismatch angle outside the sample [arcsinn0 sin Du] for
three different holographic gratings.

The resulting error of the phase shift was 60.25 for

every data point.
We measured the phase shift as a function of the
Bragg mismatch angle for three holographic gratings
of different diffraction efficiencies. The results are
shown in Fig. 2, together with the corresponding eff iciencies. The two gratings with the lower eff iciencies
have angular widths of 0.13 (FWHM), corresponding
to an effective hologram thickness of 0.56 mm [see
Eq. (1)]. This is smaller than the sample thickness,
owing to the optical density of 0.5 at the writing wavelength. The eff iciency curve of the strongest grating
is slightly narrower, indicating a larger penetration
depth of the writing beams after prolonged irradiation. For maximum diffraction efficiencies of 7%,
18%, and 24%, maximum phase shifts of 61.3, 63.3,
and 64.4, respectively, are expected from Eq. (13);
the phase shifts should occur at 60.07. The angular
positions of the experimental maxima in Fig. 2(a) are
in good agreement with this value, the curve of the
strongest grating again being slightly narrower. The
magnitude of the phase shifts, however, is somewhat
smaller than expected. This probably is due to
sample inhomogeneities and wave-front imperfections
of the writing and (or) reading laser. In the case of

1723

the higher diffraction efficiencies deviations from the

approximation of Eq. (9) may also contribute.
We have experimentally demonstrated that diffraction of light on a thick grating is connected with a
phase shift of the transmitted beam (of diffraction order zero). This phase shift is caused by the interaction
of the light with the grating and is intimately related
with the attenuation of the zero-order beam as a result
of diffraction into the f irst order. The phase shift was
measured as a function of the Bragg mismatch angle
on thick holographic gratings in a photoaddressable
block copolymer and was compared with an analytical expression derived from Kramers Kronig theory.
To our knowledge, this is the first time that an angledependent phase shift in a thick diffraction grating
was observed. This phase shift provides a third example of the general applicability of Kramers Kronig
theory. Other examples are the well-known dispersion behavior near an absorption line (as a function
of frequency) and the intensity-dependent phase shift
of the fundamental wave in a x 2 crystal near the
phase-matching condition (as a function of the phasemismatch angle).
Financial support from the Deutsche Forschungsgemeinschaft (project SFB 481) is gratefully acknowledged. M. Hckels e-mail address is michael.
haeckel@uni-bayreuth.de.
References
1. H. Kogelnik, Bell Syst. Tech. J. 48, 2909 (1969).
2. G. Montemezzani and M. Zgonik, Phys. Rev. E 55, 1035
(1997).
3. H. I. Bjelkhagen and H. J. Caulf ield, eds., Selected
Papers on Fundamental Techniques in Holography,
Vol. 171 of SPIE Milestone Series (SPIE, Bellingham,
Wash., 2001).
4. R. DeSalvo, D. J. Hagan, M. Sheik-Bahae, G. Stegeman,
E. W. Van Stryland, and H. Vanherzeele, Opt. Lett. 17,
28 (1992).
5. L. Kador, Appl. Phys. Lett. 66, 2938 (1995).
6. L. Kador, Mol. Cryst. Liq. Cryst. 283, 179 (1996).
7. C. Frenz, Diblockcopolymere mit photoadressierbaren
Chromophoren zur holographischen Datenspeicherung,
Ph.D. dissertation (University of Bayreuth, Bayreuth,
Germany, 2003).
8. T. Bieringer, R. Wuttke, and D. Haarer, Macromol.
Chem. Phys. 196, 1375 (1995).
9. A. Natansohn, Azobenzene-Containing Materials,
Vol. 137 of Macromolecular Symposia (Wiley, New York,
1999).