Analysis of the Effective Refractive Index

of Silicon Waveguides Through the

Constructive and Destructive Interference
in a Mach–Zehnder Interferometer
Volume 3, Number 6, December 2011

Yonathan Dattner, Student Member, IEEE

Orly Yadid-Pecht, Fellow, IEEE

DOI: 10.1109/JPHOT.2011.2171678
1943-0655/$26.00 ©2011 IEEE
IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Analysis of the Effective Refractive Index

of Silicon Waveguides Through the
Constructive and Destructive Interference
in a Mach–Zehnder Interferometer
Yonathan Dattner, Student Member, IEEE, and
Orly Yadid-Pecht, Fellow, IEEE

Electrical and Computer Engineering, University of Calgary, Calgary, AB T2N 1N4, Canada

DOI: 10.1109/JPHOT.2011.2171678
1943-0655/$26.00 Ó2011 IEEE

Manuscript received September 8, 2011; revised October 4, 2011; accepted October 5, 2011. Date of
publication October 13, 2011; date of current version November 22, 2011. Corresponding authors:
Y. Dattner (e-mail: ydattner@ucalgary.ca).

Abstract: This paper introduces a method of measuring the delta between the effective
refractive index of a silicon waveguide and a waveguide with wider dimensions through the
constructive and destructive interference in a Mach–Zehnder interferometer (MZI). The
method consists of a fixed effective refractive index variation incorporated by tapering one
of the arms in the interferometer to a wider waveguide dimension. The MZI consists of a
Y-branch splitter and a multimode interference (MMI) coupler. The Y-branch splitter splits
the input light 50/50 into the two arms, and the MMI is used for recombination of the two
arms. A change in the effective refractive index of one arm in comparison with the other arm
in the interferometer will introduce a phase difference on recombination in the MMI. The MMI
has the following three ports: the top and bottom output ports, which are the antisymmetric
outputs, and the middle port, which is the symmetric output. When the two signals are in
phase, all the light is coupled into the symmetric port, and when the two inputs are
out of
phase, the light is coupled 50/50 into the antisymmetric ports. The interferometer is
designed on a silicon-on-insulator (SOI) wafer and fabricated through IMEC Belgium.
Theoretical, simulation, and measured results are presented and compared.

Index Terms: Mach–Zehnder interferometer (MZI), multimode interference (MMI) coupler

silicon waveguides, silicon-on-insulator (SOI).

1. Introduction
Silicon waveguides on a silicon-on-insulator (SOI) technology have been the subject of much in-
vestigation over the past few years due to their high potentiality in optoelectronic circuits [1]–[3]. The
high contrast between Silicon and Silicon Dioxide make SOI technology suitable for optical wave-
guiding. Silicon is the most popular medium for integrated electronics but is also highly transparent in
the infrared region enabling low-loss optical waveguiding. The integration of photonics with
electronics enables many technologies such as light sources, light guiding, light modulation, and
light detection to be incorporated into a single chip.
One of the key building blocks for the development of silicon microphotonics is high-speed optical
modulators. Several methods have been investigated to get silicon-based modulators. These include
electro-optical effects in strained silicon [4], in bulk silicon-germanium [5], the quantum-confined Stark
effect in SiGe/Ge quantum wells [6], and, most popular today, modulators based on free carrier ab-
sorption, which are also known as the plasma dispersion effect [1], [2], [7]. The effective refractive index

Vol. 3, No. 6, December 2011 Page 1123

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

is important when designing the optical system as it determines the propagation constant of the optical
field but, more importantly, in high-speed communications when dispersion can be a limiting factor.
In this paper, we will incorporate a fixed effective refractive index variation in the Mach–Zehnder
interferometer (MZI) by tapering one of the arms waveguides from 525 nm wide to 625 nm wide and
then back down. The length of the 625-nm-wide waveguide portion relative to the 525-nm-wide
waveguide in the other arm will determine the phase difference on recombination. The tapered
portion will also introduce a phase difference, and it will be theoretically quantified in the next
section. The effective refractive index can be extrapolated through the constructive and destructive
interference patterns at the output of the MZI.
The output of the MZI consists of a multimode interference (MMI) coupler, consisting of two input
ports and three output ports. The MMI coupler is based on the principle of self-imaging in a multimode
waveguide [9]–[11]. The two input ports receive the initial light that was split 50/50 at the input of the
MZI with a Y-branch splitter. For two input signals in phase the symmetric case dominates and light is
coupled into the second output port. For two input ports out of phase by
, the light is coupled 50/50 into
the first and third output port, respectively. MMI design parameters and dimensions will be discussed
more in detail in Section IV. This study will compare the delta effective refractive index variations using
the effective refractive index model [12] to Lumerical finite-difference time-domain (FDTD) simulation
results and finally compare the results with a fabricated chip through IMEC Belgium.
In Section 2, we will review the theory behind calculating the effective refractive index MZI design
for both methods, i.e., with a fixed refractive index variation due to larger waveguide dimensions
and the MZI design to utilize the plasma dispersion effect. In Section 3, we will discuss more in
detail the Kramers–Kronig relations and the expected behavior on our design. In Section 4, the MMI
coupler design will be introduced, and finally, the conclusion is in Section 5.

2. Effective Index Method

For the rectangular core step-index waveguide there is no exact solution of the mode profile. Many
attempts have been made to provide approximation for the mode solution, yielding very good
results. Nevertheless, some of the best results have been obtained from a technique called
effective index method (EIM) [8]. In this section we will present a brief review of the method.
Waveguides work on the concept of total internal reflection and Fig. 1(a) shows the structure of
the rib waveguide, in which a rectangular silicon waveguide is on top of an insulator layer made of
Silicon Dioxide (SiO2 ). The solution shown here will assume that it is surrounded by air, thus leading
to an asymmetric case. The solution for the asymmetric case is more general and a solution for the
symmetric case, i.e., waveguides covered by SiO2 , can be directly derived from the asymmetric
case. To solve the field distribution in the waveguides and extract the modes, first, the transverse
electric (TE) and transverse magnetic (TM) modes for a 2-D waveguide will be analyzed, and the
EIM will be used. It is not possible to solve directly for the modes in a rib waveguide structure and
therefore the EIM is used to derive the properties of the waveguide. The EIM states that first a
solution for the TE modes (or TM modes) is solved when taking a cross section view of the
waveguides, and assuming it is infinitely wide as shown in Fig. 1(c). After solving for the 2-D
structure, an effective refractive index of the structure in Fig. 1(c) is calculated. The next step is to
use the effective refractive index found from Fig. 1(c) and use it instead of the refractive index of
silicon when looking from the top view, as shown in Fig. 1(d). With the new material, having a
refractive index calculated from the first step, the TM mode (or TE mode if initially TM mode was
used) is solved for the structure shown in Fig. 1(d), and the final effective refractive index of the 3-D
waveguide is derived. The propagation constant  of the waveguide is
2
neff
¼ (1)

where neff is the effective refractive index, and  is the wavelength.
The infrared radiation will propagate in the waveguide at a velocity corresponding to the effective
refractive index.

Vol. 3, No. 6, December 2011 Page 1124

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Fig. 1. Silicon waveguide structure. (a) 3-D structure. (b) Axis and boundary conditions used to solve for
the TE and TM modes. (c) 2-D cross section view of the waveguide used for the first step in the effective
refractive index method. (d) 2-D top view of the waveguide used for the second step in the effective
refractive method in which the refractive index of silicon is substituted with the solution from (c).

To solve the TE and TM modes of the 2-D structure of Fig. 1(b) is used. The solution for the TE
modes, i.e., the electric fields is in the y -direction is shown in
8
qx
< Ce

> ; x 0
q

Ey ðx Þ ¼ C cosðhx Þ
h sinðhx Þ ; 0  x 
t
>
:
q
 pðx þt Þ
C cosðht Þ þ sinðht Þ e ; x 
t
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi h qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
q ¼ 2
k02 n12 ; p ¼ 2
k02 n32 ; h ¼ k02 n22
2
k0 ¼ 2
=0 ;  ¼ k0 neff (2)

where C is a constant. Using (2), the mode condition for the TE mode is shown in
pþq
tanðht Þ ¼  : (3)
h 1
pq
h2

The mode condition is the eigenvalue equation for the TE modes of the asymmetric slab wave-
guide, i.e., n1 6¼ n3 . Equation (3) is an implicit relationship which involves the wavelength, refractive
index of the layers and core height as known quantities and the propagation constant  as the only
unknown quantity. There are only discrete values of  that satisfy (3), the discrete solutions of  are
the discrete modes that a waveguide supports. Each solution of  is then used in (2) to solve the
field profile and effective refractive index of the waveguide. The effective refractive index is used in
Fig. 1(d) as the refractive index of the new material instead of silicon, the TM mode is solved using
Fig. 1(b), and
8 h
qx
>
q Ce ; x 0
>
>
< h i
h
Hy ðx Þ ¼ C
q cosðhx Þ þ sinðhx Þ ; 0  x 
t
>
> h i
>
:
C h cosðht Þ þ sinðhtÞ epðx þt Þ ; x 
t
q

Vol. 3, No. 6, December 2011 Page 1125

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Fig. 2. Proposed Y-branch splitter MZI with a tapered waveguide.

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

q¼ 2
k02 n12 ; p¼ 2
k02 n32 ; h¼ k02 n22
2
n22 n22
q¼ q; p¼ p; k0 ¼ 2
=0 ;  ¼ k0 neff (4)
n12 n32
where C is a constant [Fig. 4(d) is rotated 90 to align with the coordinate system in Fig. 4(b)]. Using
(4), the mode condition for the TM mode is shown in
pþq
tanðhtÞ ¼  : (5)
h 1
pq
h 2

Again the discrete solutions of  in (5) are the TM modes of the waveguide in Fig. 1(d). The
 solutions are used in (4) to extract the magnetic field profile and effective refractive index of
the 3-D waveguide. The effective refractive index is a measure of the wavelength velocity in which
the infrared radiation for a particular wavelength will propagate in the waveguide. Using the solu-
tions to (2) and (4), the power flowing in the direction of propagation can be derived using the
complex Poynting vector shown in

~ave ¼ 1 ReðE
S ~H ~Þ (6)
2
~H
where E ~ave is the power flowing across an area, i.e.,
~  is the complex Poynting vector, and S
2
W/m .

3. Mach–Zehnder Interferometer
The EIM can be used to extract the effective refractive index neff and can change rather abruptly
with slight variations in the waveguide dimensions. This is the reason that constant phase control
can be achieved with very short MZI in contrast to using other methods such as free carrier
absorption.
The accumulated phase for light travelling in a waveguide a distance L is given by
ZL
2

ðLÞ ¼ neff ðx Þ dx (7)

0

where  is the wavelength, and neff is the effective refractive index of the waveguide.
Fig. 2 shows the Y-Branch splitter MZI with a tapered waveguide on the bottom arm. The MZI
consists of a Y-branch splitter, a tapered waveguide to 625 nm, an MMI coupler and three output
ports. The Y-branch splitter splits the light 50/50 into the two arms and is wavelength independent.
The tapered waveguide to 625-nm will introduce a phase difference from the above arm, depending
on the length of the 625-nm portion. From (7), one can calculate the phase difference on

Vol. 3, No. 6, December 2011 Page 1126

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Fig. 3. Effective refractive index and mode profile at  ¼ 1:55 m for (a) 525-nm-wide waveguide
(b) and 625-nm-wide waveguide.

recombination by taking the total phase accumulated in the top arm in comparison to the bottom arm
of the MZI. When the two waves entering the MMI coupler are in phase, all of the light will couple to the
middle output port, and when they are
out of phase, the light will couple half and half into the top and
bottom output ports, respectively.
The authors would like to mention that the original layout design was for 220 nm height
waveguides, 600 nm width for the standard waveguide, and 700 nm width for the larger tapered
waveguide. Due to fabrication lithography exposure doses the actual fabricated waveguides where
it is 210 nm in height and 525 nm in width for the standard waveguide and 625 nm in width for the
larger tapered waveguides. The dimensions were derived using a Tencor profilometer for wave-
guide heights and a scanning electron microscope (SEM) for the width.
Using (1)–(6), the effective refractive index of the waveguides were calculated and validated using
Lumerical FDTD simulation software. The input light is TE polarized and the waveguides dimensions
for both the 525-nm and 625-nm portion only support the lowest fundamental mode at  ¼ 1:55 m.
It is important for the waveguides to be single mode because the higher order modes would
propagate with a different effective refractive index. The simulation results for the effective refractive
index of a 525-nm-wide waveguide and 625-nm-wide waveguide at  ¼ 1:55 m are shown in Fig. 3.
The effective refractive index of the waveguide at 525 nm width is 2.381263 and at 625 nm width
is 2.505045. Due to the large variation in the effective refractive index, the waveguides were
tapered to reduce reflections on the interfaces between the 525-nm and 625-nm portions.
When calculating (7), these tapered regions will also affect the phase difference. To calculate the
effect of the tapered regions on the phase difference, neff as a function of the waveguide width (W )
is derived in

neff ðW Þ ¼ P1  W 3 þ P2  W 2 þ P3  W þ P4 (8)
where P1 ¼ 7:95  10
9 , P2 ¼
1:71  10
5 , P3 ¼ 1:299  10
2 , and P4 ¼
0:8786. The equation
was derived using a curve fitting algorithm for waveguides width varying from 525 nm to 625 nm.
It is difficult to derive the initial amount of light that actually entered the waveguide. For this reason
a wavelength dependent 50/50 coupler is fabricated for each MZI under test. The concept is shown
in Fig. 4, in which a wavelength dependent 50/50 coupler is incorporated before the MZI under test.
By measuring the amount of light exiting Output #2, the amount of light that entered the input grating
is known and useful conclusions can be derived for the MZI under test. As discussed earlier, the
MZI is designed for operation at 0 ¼ 1:55 m, and the 50/50 coupler is also designed for the same
wavelength. As the wavelength changes the coupling ratio will also vary and the Lumerical FDTD
simulated results of the 50/50 coupler are shown in Fig. 5.
Due to the high exposure dose when the chip was fabricated, the gap between the waveguides
was wider then designed. To derive the actual coupling ratio there was a test structure on chip that

Vol. 3, No. 6, December 2011 Page 1127

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Fig. 4. Wavelength dependent 50/50 coupler and the MZI test design.

Fig. 5. 50/50 coupler’s transmission (Output #2).

had the 50/50 coupler as a standalone design. When measured the actual coupling ratio was 60/40,
i.e., 60% of the input light went into the MZI under test, and 40% was used to normalize the results
to actual input light.

4. MMI Coupler
The MMI coupler works on the principle of self-imaging effect [9]. An input field profile is reproduced
in single or multiple images at periodic intervals along the propagation direction. This occurs due to
constructive interference between the waveguide modes. The beat length L
is derived by using the
propagation constants between any two lowest order modes [11]

4nr We2
L
¼ ffi (9)
0
1 30
where 0 and 1 are the propagation constants of the two lowest order modes, nr is the refractive
index of the rib waveguide, We is the effective width pf the multimode section of the splitter/
combiner, and 0 is the free space wavelength. Using (9) and Lumerical FDTD mode solutions to
find the two lowest order modes, L
¼ 88:2 m. Fig. 6 shows the FDTD simulation of the field
undergoing self-imaging, giving a result of L
 77 m.
Using the results from the simulation, the design of the MMI is presented in Fig. 7. The MMI has
two input ports coming from the MZI, and it is assumed that each input has half of the total light
intensity. There are three output ports; the P0 output port will have the light coupled into it when the
two inputs are in phase, and the output ports P
will have the light coupled into them half and half
when the two input ports are
out of phase. The output ports are tapered, and their distances apart
from one another are optimized to collect the most amount of light, depending on the phase

Vol. 3, No. 6, December 2011 Page 1128

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Fig. 6. Lumerical FDTD self-imaging simulation.

Fig. 7. Multimode interference (MMI) coupler.

Fig. 8. MMI electric field profile when the two inputs are in (a) phase (b) and out of phase by
the input
light is coupled.

conditions of the inputs. Simulation results show that approximately 5% of the input light is lost due
to scattering in between the waveguides.
The simulation results of the MMI are presented in Fig. 8. When the two inputs are in phase (a),
the electric field is coupled into the middle port, and when the two inputs are
out of phase, the
input light is coupled half and half into the top and bottom output ports. An SEM image of the MMI
coupler is shown in Fig. 9.

5. Results and Discussion

Using (2)–(8), the theoretical transmission of light coupled into the middle output port as a function
of the length of the 625-nm-wide waveguide portion is plotted in Fig. 10 (solid black line). From

Vol. 3, No. 6, December 2011 Page 1129

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

Fig. 9. SEM image of the MMI coupler.

Fig. 10. Theoretical, simulated, and measured transmission of light coupled into the middle output port
in the MMI as a function of the length of the 625-nm-wide waveguide portion.

Lumerical FDTD, it was derived that approximately 5% of the input light to the MMI was lost due to
scattering and nonideal behavior of the MMI, i.e., a small portion light was coupled into the incorrect
port. The Lumerical FDTD simulation results of the in phase port, i.e., middle port, normalized to the
maximum (93% of total light entering the MMI) and minimum (4% of the total light entering the
MMI) is plotted in Fig. 10 (dashed red line). In a similar manner, the measured chip results are
plotted as blue crosses for the available MZI structures on chip.
When extracting the delta effective refractive index between a 525-nm to 625-nm-wide silicon
waveguide, 210 nm in height on a 2-m SiO2 layer from Fig. 10, the tapered regions only introduce
a constant phase shift. To derive the measured delta effective refractive index
neff based on (7),
(9), and the properties of the MMI is used



neff ¼  cos
1 ð2  T
1Þ þ 2
 k ; k ¼ 0; 1; 2; . . . ; 1 (10)
2
 L
where  is the wavelength, L is the distance traveled in the wider 625-nm section, T is the
measured transmission in the output of the middle port of the MMI normalized to the transmission

Vol. 3, No. 6, December 2011 Page 1130

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

without the constant phase introduced by the tapered regions, and k is the number of 2
phase
differences introduced in the wider waveguide in relation to the narrower waveguide when
travelling a distance of L.
Solving for (10) and using Fig. 10, the delta effective refractive index has a measured mean value
of 0.1031 with a standard deviation of 0.0474. The theoretical and simulated delta effective refrac-
tive indexes are highly correlated and give a result of 0.123782. It can be seen that the standard
deviation of the measured delta effective refractive index falls within the theoretical value. In the
measured results, there are many sources of possible errors, some of the more likely ones being
fabrication differences between varying structures tested and differences in alignment of fiber optics
used to collect and deliver the light to the chip. Another source of error is fabrication imperfections in
the tapered waveguide from 525 nm to 625 nm and back down. Imperfections in the tapered
portions cause a nonideal 50/50 split in the two arms of the MZI because of reflections from
effective refractive index variations. In the fabricated chip, there were 19 structures on chip with
varying lengths, and if more structures were used, the standard deviation would also decrease. The
theoretical delta effective refractive index was validated experimentally, and this method for mea-
suring how the effective refractive index changes as a function of the waveguides geometry has
been validated.

6. Conclusion
A method of measuring the delta between the effective refractive index of a silicon waveguide and a
waveguide with wider dimensions through the constructive and destructive interference in an MZI
was introduced. The method consists of a fixed effective refractive index variation incorporated by
tapering one of the arms in the interferometer to a wider waveguide dimension. Theoretical simu-
lated and measured results were shown. The technique was validated with the measured results
mean and standard deviation width falling within the theoretical and simulated result.

Acknowledgment
The authors would like to thank Associate Prof. L. Chrostowski and Prof. N. Jaeger for their
guidance and use of test facilities at the University of British Columbia. The authors would also like
to thank Dr. Hayden and AMIF at the University of Calgary for their assistance with waveguide
geometry measurements.

References
[1] B. Schmidt, Q. Xu, J. Shakya, S. Manipatatruni, and M. Lipson, BCompact electro-optic modulator on silicon-on-
insulator substrates using cavities with ultra-small modal volumes,[ Opt. Exp., vol. 15, no. 16, pp. 3140–3148,
Mar. 2007.
[2] A. Liu, L. Liao, D. Rubin, J. Basak, Y. Chetrit, H. Nguyen, R. Cohen, N. Izhaky, and M. Paniccia, BRecent development
in a high-speed silicon optical modulator based on reverse-biased pn diode in a silicon waveguide,[ Semicond. Sci.
Technol., vol. 23, no. 6, p. 064001, Jun. 2008.
[3] N. Rouger, L. Chrostowski, and R. Vafaei, BTemperature effects on Silicon-On-Insulator (SOI) racetrack resonators: A
coupled analytic and 2D finite difference approach,[ J. Lightw. Technol., vol. 28, no. 9, pp. 1380–1391, May 2010.
[4] R. S. Jacobsen, K. N. Andersen, P. I. Borel, J. Fage-Pedersen, L. H. Frandsen, O. Hansen, M. Kristensen,
A. V. Lavrinenko, G. Moulin, H. Ou, C. Peucheret, B. Zsigri, and A. Bjarklev, BStrained silicon as a new electro-optical
material,[ Nature, vol. 441, no. 7090, pp. 199–202, May 2006.
[5] J. Liu, M. Beals, A. Pomerene, S. Bernardis, R. Sun, J. Cheng, L. C. Kimerling, and J. Michel, BWaveguide-integrated,
ultralow-energy GeSi electro-absorption modulators,[ Nat. Photon., vol. 2, no. 7, pp. 433–437, Jul. 2008.
[6] Y. Kuo, Y. K. Lee, Y. Ge, S. Ren, J. E. Roth, T. I. Kamins, D. A. B. Miller, and J. S. Harris, BStrong quantum-confined
Stark effect in germanium quantum well structures on silicon,[ Nature, vol. 437, no. 7063, pp. 1334–1336, Oct. 2005.
[7] P. Dainesi, A. Kung, M. Chabloz, A. Lagos, P. Fluckiger, A. Ionescu, P. Fazan, M. Declerq, P. Renaud, and P. Robert,
BCMOS compatible fully integrated Mach–Zehnder interferometer in SOI technology,[ IEEE Photon., vol. 12, no. 6,
pp. 660–662, Jun. 2000.
[8] Q. Wang, G. Farrel, and T. Freir, BEffective index method for planar lightwave circuits containing directional couplers,[
Opt. Commun., vol. 259, no. 1, pp. 133–136, 2006.
[9] M. Bachmann, P. A. Besse, and H. Melchior, BGeneral self-imaging properties in NxN multimode interference couplers
including phase relations,[ Appl. Opt., vol. 33, no. 18, pp. 3905–3911, Jun. 1994.

Vol. 3, No. 6, December 2011 Page 1131

IEEE Photonics Journal Effective Refractive Index of Silicon Waveguides

[10] J. Yu, X. Wang, J. Liu, Q. Yan, J. Xia, Z. Fan, Z. Wang, and S. Chen, BMMI Optical Couplers and Switches with SOI
Technology,[ in Proc. Optoelectron., 2003, pp. 251–254.
[11] N. Q. Ngo and S. R. Natarajan, BDesign of multimode interference coupling structures for photonic integrated circuit
applications,[ Proc. SPIE, vol. 4904, pp. 368–374, 2002.
[12] S. T. Peng, H. L. Bertoni, and T. Tamir, BAnalysis of periodic thin-film structures with rectangular profiles,[ Opt.
Commun., vol. 10, no. 1, pp. 91–94, Jan. 1974.
[13] T. S. Moss, G. J. Burrell, and B. Ellis, Semiconductor Opto-Electronics. London, U.K.: Butterworth, 1973.
[14] C. Cocorullo, M. Iodice, I. Rendina, and P. M. Sarro, BSilicon thermo-optical micro-modulator with 700 kHz–3 dB
bandwidth,[ IEEE Photon. Technol. Lett., vol. 7, no. 4, pp. 363–365, Apr. 1995.

Vol. 3, No. 6, December 2011 Page 1132