Refracted near-field measurements of refractive index and geometry
Refracted near-field measurements of refractive index
and geometry of silica-on-silicon
integrated optical waveguides
Philippe Oberson, Bernard Gisin, Bruno Huttner, and Nicolas Gisin
The standard refracted near-field technique for measuring the refractive-index profile of optical fibers
cannot be directly used for silica-on-silicon integrated optical waveguides because of the opacity of silicon.
A modified method is thus presented to characterize this kind of waveguide. The resolution it gives, both
spatially and in the refracted index, is practically as good as that obtained with the standard technique
for measuring optical fibers. © 1998 Optical Society of America
OCIS codes: 120.5710, 230.7370.
2. Measurement Principle and Setup
Integrated optics on silicon has reached a high level
of performance in passive optical components and
circuits,1–5 with applications for telecommunications
as well as for sensors. The basic physical characteristic of optical waveguides is given by the distribution
of the refractive index. From this distribution all
local parameters can be computed, such as the mode
profile that determines the coupling losses. Measurement of the refractive-index profiles also provides useful quality control during the development
and production of the components. The refracted
near-field ~RNF! technique is the most advanced
technique used to measure refractive-index profiles of
optical fibers6,7 and glass-integrated optical
waveguides.8 –10 However, because of the presence
of silicon, an opaque material that creates reflections
at the interface, the standard RNF technique cannot
be directly used for silica-on-silicon integrated optical
waveguides. In this paper we present an adaptation
of the RNF technique for the measurement of
refractive-index profiles and geometry of silica-onsilicon integrated optical waveguides. Section 2
gives details of the technique used for these measurements, and Section 3 presents the results.
The measurement principle and setup are quite similar to those presented in Refs. 6, 7, and 9. A laser
beam is focused on the sample, and the intensity of
the refracted light is measured to determine the refractive index. The measurement cell is modified to
accommodate the new type of waveguide. Figure 1
illustrates the new cell. It resembles the one that
we used for glass-integrated optical waveguides,9
though a great deal of additional effort has been put
into it to simplify the handling. The component can
easily be placed in the cell, and the area to be analyzed can be centered as required. This measurement cell is a vessel containing an index-matching
oil. On the bottom surface of the vessel is an aperture that is closed by a diopter made out of thin glass
~0.17 mm!. A light detector is situated 11 mm above
the diopter. The integrated-optics component of silica on silicon is placed vertically, with the end face in
contact with this diopter. It is necessary that the
end face of the component should be well cleaved
~perpendicularly to the guide! and that the refractiveindex profile should be invariant under vertical
translation ~Z direction in Fig. 1! over approximately
1 mm from this end face. Under the measurement
cell is a 5-mm-diameter parallel laser beam ~wavelength, 675 nm! focused by an immersion microscope
objective of 1.25 numerical aperture ~NA! on the
integrated-optics component. This microscope objective is different from the one used for standard
methods; the reasons for this change are explained
below. Three quarters of the parallel beam are
blocked in such a way as to prevent any light from
The authors are with the Group of Applied Physics, University of
Geneva, 1211 Geneva 4, Switzerland.
Received 10 April 1998; revised manuscript received 8 June
© 1998 Optical Society of America
APPLIED OPTICS y Vol. 37, No. 31 y 1 November 1998
Fig. 1. RNF technique adapted for silica-on-silicon integrated
optical waveguides. The incident beam is shaped as a quarter
disk and focused through a diopter by a microscope objective ~NA,
1.25! onto the end face of the component under test. On a detector
a half-disk-shaped mask is centered around the optical axis. Consequently, only the refracted light rays that have an angle larger
than umin are collected by the detector. These correspond to incident rays that have an angle larger than wmin, where wmin and umin
are related by Eq. ~2!. The detected intensity I~x, y! is therefore
a function of n~x, y!.
being reflected on the vertical face of the silicon. A
drop of index-matching oil is inserted between the
microscope objective and the diopter. The measurement cell can be moved with the help of a motorized XYZ microdisplacement system in steps of
0.1 mm. The Z displacement, parallel to the optical
axes, allows one to adjust the position of the focal
point on the section plane of the integrated
waveguide. For each ~ x, y! position, the detector
signal caused by the refracted light intensity I~ x, y!
is collected, its value being sufficient to determine
the refractive index n~ x, y!, as shown in the following paragraph.
The starting point is Snell’s law ~written with vectors!:
v1 z s 5 v2 z s,
where vi 5 nic represents both the direction of a light
ray ~optical geometry! and its velocity in medium i; ni
is the refractive index; and s is any vector tangent to
the interface. By applying Eq. ~1! to the horizontal
interface and the vertical interfaces and using s 5 ẑ,
the vertical unitary vector for all the vertical interfaces, one can show that
sin2 w 5
2 cos2 u,
where w and u are the angles between the Z direction
and a light ray entering and leaving the component,
respectively, and nl is the index of oil. Equation ~2!
is the fundamental relation of the RNF technique.
This way of deriving Eq. ~2! shows that it is valid in
full generality, without any assumption of circular
symmetry or of axial rays. In particular, note that
the projection of the ray on the XY plane does not
necessarily follow a straight line ~lens effect!. In
practice, the deviation in this XY plane does not exceed 20 –30°. For this reason only a quarter of a
cylindrical beam is used to illuminate the end face of
the waveguide, and a half-disk detector collects the
refracted light. A circular mask, centered on the
optical axis, covers the central part of the detector.
This ensures that only refracted light rays with an
angle larger than umin are collected. These correspond to incident rays on the component with an
angle larger than wmin, where wmin is given by Eq. ~2!.
Since wmin increases with the refractive index n~x, y!,
the intensity I~x, y! collected at the detector decreases
for increasing n~ x, y!. Hence, after calibration,
n~ x, y! can be recovered from the collected intensity
Note that incident rays with an angle larger than
wmin should reach the detector, so they should not be
guided by the component. This imposes a value of
umin given by
umin $ arccos
~2n2min 2 n2max!1y2
where nmin and nmax are the minimum and maximum
refractive indices that can be found in the component,
Movement of the cell will cause variations in umin of
the order of 0.1° and thus a slight change in the
intensity @in addition to the variation due to the profile of n~ x, y!#. For the standard technique, the first
order of the development of intensity I~ x, y! in the
neighborhood of the central position is zero, because
of the circular symmetry of the cell. With this setup
the first order of x is not zero, because all the refracted light escapes in the direction of negative x.
These changes in intensity can easily be corrected
during calibration, since they can be measured in the
oil @when n~x, y! 5 nl is constant#. Note that a firstorder correction is sufficient.
The silicon part of the component causes other
problems that do not appear with transparent materials such as optical fibers. Indeed, for transparent
materials, any weak parasitic light from outside the
beam creates only a constant additional intensity,
whereas with opaque materials the collected intensity varies according to the position of the opaque
obstacle and causes a perturbation of the signal.
Therefore special care has to be taken to avoid all
spurious reflections. In particular, with a standard
objective, the Fresnel reflection on the measurement
cell window followed by a reflection on the microscope
objective distorted the measurement. This problem
was satisfactorily solved by use of the immersion ob1 November 1998 y Vol. 37, No. 31 y APPLIED OPTICS
Fig. 2. Refractive-index profile of a NPL calibrated optical fiber
glued onto a silica-on-silicon substrate. This fiber was measured
with the setup of Fig. 1.
jective, which decreased the index steps causing the
reflections. Moreover, with the standard objective
~0.85 NA, corrected for a 0.17-mm diopter! a spherical
~focal! aberration was observed when the laser beam
covered the entire aperture. This aberration decreased with the immersion objective, mainly because the large NA ~1.25! allows one to use only the
central part of the beam.
Before measuring the integrated waveguides, we
tested the setup by measuring a calibrated optical
fiber provided by the National Physical Laboratory
~NPL!, Teddigton, UK. This fiber was glued to a
silica-on-silicon substrate. The results are presented in Fig. 2. The values provided by NPL and
the oil refractive indices are used for the calibration
of n~ x, y! as a function of I~ x, y!. A simple linear
calibration is used because the variation of the angle
wmin is slight. Agreement with the NPL data within
5 3 1024 was obtained. Note that with this measurement method the silicon appears to have an infinitely high refractive index.
Figures 3 and 5 present calibrated raster scans of
two different silica-on-silicon waveguides. The
guides can be seen clearly, as well as the highrefractive-index lines. These are due to the manufacturing process: The silicon is covered by two
silica layers, one with a low index and one with a high
index. The high-index layer is chemically removed
around the guiding region, and a final low-index layer
is deposited on it. From this measurement both the
geometry and the refractive-index profile of the guide
can be determined with an accuracy similar to that
obtained for optical fibers, i.e., a few tenths of a micrometer. Note that, because of the manufacturing
process, the outer surface of these components is not
Fig. 3. Raster scan of a silica-on-silicon integrated optical waveguide.
APPLIED OPTICS y Vol. 37, No. 31 y 1 November 1998
Fig. 4. Refractive-index profiles of the component under test in
Fig. 3. The solid and dashed curves correspond to a scan perpendicular ~x scan! and parallel ~ y scan!, respectively, to the component’s surface.
Fig. 6. Refractive-index profiles of the component under test in
Fig. 4. The solid and dashed curves correspond to a scan perpendicular ~x scan! and parallel ~ y scan!, respectively, to the component’s surface.
flat. Finally, Figs. 4 and 6 show some calibrated
X–Y profiles through these waveguides. The
refractive-index step and waveguide size match the
manufacturer’s data well. The aberrations ~elevation of the index near the silicon! caused by the presence of silicon are present in a region of only a few
micrometers around the junction.
The width of the transitions visible on the figures is
sometimes less than 0.5 mm, a satisfactory result
when one considers that this is the minimum theo-
retical diameter of the spot ~calculated from the beam
aperture and the laser wavelength!. The smallest
refractive-index differences observable on the figures
are less than 5 3 1024.
Refractive-index profiling and geometry characterization of silica-on-silicon integrated waveguides
have been obtained by a modification of the standard
refracted near-field technique. Problems due to
Fig. 5. Raster scan of a second silica-on-silicon integrated optical waveguide.
1 November 1998 y Vol. 37, No. 31 y APPLIED OPTICS
the presence of opaque silicon have been overcome
by use of an immersion objective and only a quarter
of the laser beam. The estimated accuracy is similar to that obtained for optical fibers: a few tenths
of a micrometer for the geometry and approximately 5 3 1024 for the refractive index. Special
attention has been paid to the simplification of handling procedures. This technique does not require
any sample preparation; moreover, it is nondestructive and provides waveguide characterization in a
few minutes. An industrial application should be
a straightforward matter, as it would require only a
change of objective on the measuring instrument
and the use of a different measurement cell.
Special thanks are due to Gilles Grand for stimulating discussions. The integrated optical components were kindly provided by the Laboratoires
d’Electronique, de Technologie et d’Instrumentation,
1. P. Mottier, “Integrated optics at the Laboratoires d’Electronique, de Technologie et d’Instrumentation,” Int. J. Optoelectron. 9, 125–134 ~1994!; J. Magerand, G. Grand, P. Pouteau, and
P. Philippe, “Integrated polarization insensitive 1.3y1.55 micrometer duplexer on silica-based technology,” presented at the
International Symposium on Integrated Optics, Lindau, Germany, 11–15 April 1994; V. Delisle, G. Grand, A. Fournier, and
P. Mottier, “Reduce-size low crosstalk PECUD silica phasar using widened continuous bends,” in Digest of the Eighth European
APPLIED OPTICS y Vol. 37, No. 31 y 1 November 1998
Conference on Integrated Optics (Optical Society of America,
Washington, D.C., 1997), pp. 72–75.
M. Kawachi, “Silica waveguides on silicon and their application to integrated-optic components,” Opt. Quantum Electron.
22, 391– 416 ~1990!.
C. Dragone, C. A. Edwards, and R. C. Kistler, “Integrated
optics N 3 N multiplexer on silicon,” IEEE Photonics Technol.
Lett. 3, 896 – 899 ~1991!.
W. J. Stewart, “A new technique for measuring the refractive
index profiles of graded optical fibers,” in Technical Digest 1006
of the 1977 International Conference on Integrated Optics and
Optical Fiber Communication (Institute of Electronics and Communication Engineers of Japan, Tokyo, 1977!, pp. 395–398.
K. I. White, “Practical application of the refracted near-field
technique for the measurement of optical fiber refractive index
profiles,” Opt. Quantum Electron. 11, 185–196 ~1979!.
K. W. Raine, J. G. N. Baines, and D. E. Putland, “Refractive
index profiling—state of the art,” IEEE J. Lightwave Technol.
7, 1162–1169 ~1989!.
N. Gisin, R. Passy, and B. Perny, “Optical fiber characterization
by simultaneous measurement of the transmitted and refracted
near field,” IEEE J. Lightwave Technol. 11, 1875–1883 ~1993!.
R. Göring and M. Rothhardt, “Application of the refracted
near-field technique to multimode planar and channel
waveguides in glass,” J. Opt. Commun. 7~3!, 82– 85 ~1986!.
N. Gisin, J. P. Pellaux, P. Stamp, N. Hori, and N. Masuyama,
“Alternative configuration of refracted near-field measurements of refractive index on glass-integrated-optics
waveguides,” Appl. Opt. 31, 7108 –7112 ~1992!.
L. Goldberg, “Interferometrique method for measuring diffused channel waveguide-index profile,” Appl. Opt. 20, 3580 –