Silicon oxynitride platform for linear and nonlinear photonics at NIR wavelengths

Gioele Piccoli,1, 2, ∗ Matteo Sanna,2 Massimo Borghi,2, 3 Lorenzo Pavesi,2 and Mher Ghulinyan1
1
Sensors and Devices, Fondazione Bruno Kessler, Trento, Italy
2
Department of Physics, University of Trento, Trento, Italy
3
Present address: Department of Physics, University of Pavia, Pavia, Italy
The development of versatile and novel material platforms for integrated photonics is of prime
importance in the perspective of future applications of photonic integrated circuits for quantum
information and sensing. Here we present a low-loss material platform based on high-refractive
index silicon oxynitride (SiON), which offers significant characteristics for linear and non-linear
optics applications in a wide range of red/near-infrared wavelengths. The demonstrated propagation
loss < 1.5 dB/cm for visible wavelengths enables the realization of long and intricate circuitry for
photon manipulations, as well as the realization of high quality factor resonators. In addition,
the proposed SiON shows a high nonlinear coefficient of 10−19 m2 /W, improving the strength of
arXiv:2206.04388v1 [physics.optics] 9 Jun 2022

nonlinear effects exploitable for on-chip photon generation schemes.

I. INTRODUCTION It is transparent at wavelengths from 400 nm to 2.0µm,

while very recent developments report on operation in
The development of novel platforms for integrated the ultra-violet region [8]. Commercially available plat-
photonics, characterized by a strong versatility in terms forms show propagation losses < 0.1 dB/cm over the
of applications, is of prime importance for the realiza- whole operational range, with passive components that
tion of emerging photonics applications in the fields show similar performance to their SOI counterparts [9–
of sensing, metrology, quantum communications and 11]. Due to its dielectric nature, active components
quantum computing [1, 2]. Several technologies for in SiN technology can relay to only relatively slower
photonic integrated circuits (PICs) have been proposed thermo-optical modulation of the refractive index. Nev-
and developed in the last decades. Silicon photonics ertheless, due to the large optical bandgap of Eg > 4 eV,
emerged first, exploiting the advanced CMOS man- SiN does not suffer EC or TPA losses, while nonlinear
ufacturing techniques optimized for microelectronics. optical generation can still be significant owing to ap-
Nowadays, Si photonic devices, based on the Silicon- preciable third-order nonlinearities. Stoichiometric sili-
on-Insulator (SOI) platform, are already spreading in con nitride (Si3 N4 ) offers superior optical quality, how-
the commercial world, mainly for the telecommuni- ever, the film thicknesses should be kept to < 200 nm to
cation networks (see Ref. [3] and references therein). avoid film cracking due to the large tensile stress. This
State of the art commercial SOI devices, designed for limitation can be overcome using sophisticated tech-
the telecom C-band, achieve propagation losses of less niques, such as photonic Damascene [12] and multilay-
than 1 dB/cm, offering highly efficient passive compo- ered TriPleX [10, 13, 14] processes.
nents distributed within standard product development An alternative approach to reduce film stress con-
kit [4–6]. sists in introducing oxygen into the SiN material, by
Active components are also implemented within the depositing directly a silicon oxynitride (SiON) film [15–
SOI platform, with phase shifters based on p-n junc- 18]. The refractive index of SiON can be continuously
tions realized directly on the Si waveguides, while for tuned between 1.45 (SiO2 ) and 2.00 (Si3 N4 ) by con-
photodetectors and light sources hetero-integration of trolling the relative content of O and N in the film.
germanium or III-V compound semiconductor mate- The material loss can be lower than that of SiN, mean-
rials is employed. A limiting constrain of Si PICs is time maintaining a very low stress for film thicknesses
that their linear operation is acceptable at wavelengths of up to few micrometers. As a drawback, the reduc-
longer than 1.1 µm (energy bandgap Eg ∼ 1.1 eV) tion of the material refractive index causes an increase
where the core material’s absorption is insignificant, in the device footprint, as well as a weakening of the
while for nonlinear optical applications excited-carrier thermo-optical and nonlinear optical characteristics of
(EC) and two-photon absorption (TPA) can still be sig- the guiding components [19].
nificant at photon energies ∼ Eg /2. In this work, we present a novel, low-loss, photonic
Silicon Nitride (SiN) has been introduced as an alter- platform based on SiON channel waveguides, capable to
native dielectric platform for integrated photonics [7]. manage a wide range of VIS/NIR wavelengths for both
linear and nonlinear optics applications. The core ma-
terial consists in a relatively high refractive index SiON
(1.66 at 850nm wavelength), which enables small foot-
∗ [email protected] print PIC designs with propagation losses ∼ 1.5 dB/cm,
 2

Figure 1. Sketch of the fabrication process of the high-index SiON photonic platform. (1) The Si wafer with the bottom
SiO2 cladding and the Si3 N4 film, (2) deposition of the SiON core material, (3) lithography and etching of waveguides,
(4) deposition of the top Si3 N4 film, (5) deposition of the top SiO2 cladding, (6) sputtering of the TiN film, (7) lithography
and etching of the TiN resistors, (8) deposition of an oxidation protective film and (9) optional local opening of the
waveguides top cladding.

improvable by at least a factor of two. The material on the chip, without the risk to underetch the bottom
shows an optical bandgap of 3.8 eV with a relatively SiO2 cladding. The removal of top SiO2 is an optional
strong optical nonlinearity of 1.3 ± 0.6 × 10−19 m2 /W process, which can be used for different purposes, for
close to the TPA absorption edge. This, combined with example, a chemical functionalization of the waveguide
the possibility to remove locally the cladding without surface for sensing applications [17, 21–24] or, as it will
damaging the waveguide, offers large versatility to en- be discussed in Section V, in cases when the refractive
gineer the waveguide dispersion for applications such as index dispersion of waveguides should be engineered for
nonlinear Four-Wave Mixing (FWM) in ring resonator non-linear optics applications [25–30].
devices. Our SiON platform has the potential to be The fabrication process of the proposed platform is
further developed for the monolithic integration of all schematically described in Fig. 1. Starting from a 6 inch
necessary functionalities – photon sources, light manip- silicon wafer, first a 1.6 µm silicon oxide has been grown
ulation circuits and photon detection (recently proven by wet thermal oxidation at 975 ◦ C to form the bottom
in Ref. [20]) – on a single chip, operating at room tem- cladding. On top of it, a 50 nm thick Si3 N4 is deposited
perature, for classical and quantum applications. in a Low-pressure chemical vapor deposition (LPCVD)
The paper is structured as follows: Sec. II presents furnace at 770 ◦ C, followed by a 500 nm SiON film
the technological approach, Sec. III and IV describe the deposition in a plasma-enhanced chemical vapor depo-
linear and nonlinear properties of the SiON platform, sition (PECVD) chamber using SiH4 , N2 O and NH3
respectively. In Sec. V we describe the dispersion en- gas precursors. Next, the photonic devices were defined
gineering of ring resonators for non-linear FWM and by photoresist patterning using an i-line stepper lithog-
the fabrication of these devices. Finally, in Sec. VI we raphy, and the pattern transferred to the SiON/Si3 N4
summarize our results and draw conclusions. layers using reactive ion etching (RIE). Then, the SiON
waveguides were thermally treated at 1050 ◦ C for 90
min in a N2 atmosphere to allow the release of residual
II. PHOTONIC PLATFORM H2 and the improvement of the optical properties of the
SiON film. Next, a second deposition of 50 nm Si3 N4
The photonic platform we introduce is based on the was performed, followed by a deposition of LPCVD
use of high-index SiON for the core material and SiO2 borophosphosilicate glass (BPSG) and PECVD SiO2
claddings, resulting in a relatively large core/cladding films at 640 ◦ C and 300 ◦ C, respectively, to form the
index contrast of ∼ 15%. In addition, the channel top cladding of a total of 1.6 µm.
waveguide is encapsulated between two thin films of In a next step, a multi-stack of 150 nm TiN and
Si3 N4 . These last act as an etch-stop barrier during the 1200 nm of Al was sputtered, patterned, and etched
wet chemical etching, which is used to open windows in with RIE to allow for the realization of metal lines,
the waveguide’s top SiO2 cladding in specific locations contact pads and micro-heaters for the thermo-optical
 3

Wavelength (μm)
(a) (b) 0.8 0.4 0.3 0.2
2 0.03

Extinction coefficient
1.9

Refractive index
0.02

1.8

0.01
1.7

1.6 0
2 4 6
Energy (eV)

Figure 2. (a) Cross-sectional SEM micrograph of an 800 nm × 300 nm SiON waveguide. The different materials of the
multilayer composing the platform are labelled on the figure. (b) Dispersions for the real refractive index and the extinction
coefficient of the SiON core material. The wavelength range of study for this work is highlighted in grey.

tuning of photonic components. The top Al film was This model allows to estimate the optical band gap Eg
selectively removed on top of the micro-heaters via wet of the platform’s core material. The parameters fj , Γj
chemical etching in order to realize efficient TiN micro- and Ej describe, respectively, the amplitude, the spec-
resistances with a sheet resistance of 5 Ω/sq. The wafer tral width and the spectral position of the absorption
was covered with 500 nm PECVD SiO2 protective film, peak in energy units. The obtained dispersions for the
which was then removed from pad positions to allow for real refractive index and the extinction coefficient are
external electrical contact. plotted in Fig. 2b.
Finally, the chip boundaries and the waveguide facets Of particular interest for this work, the refrac-
were defined by RIE of the dielectric multilayer, with tive index of SiON has been engineered to be higher
an additional 140 µm deep etch into the Si substrate than that of typical low-index SiON used in previous
through a Bosch process in order to ease the butt- works [15, 32, 33]. This choice allows to design opti-
coupling between optical fibers and waveguides. Fig- cal components with smaller footprint and larger op-
ure 2a shows the SEM cross-sectional micrograph of the tical mode confinement. Despite the relatively higher
core of an 800 nm × 300 nm SiON waveguide. All the refractive index (1.66 at 850 nm), the estimated band-
different films, including the surrounding 50 nm-thick gap of Eg ≈ 3.8 eV allows the photonic platform to
Si3 N4 film, the bottom and top claddings as well as the be used without material absorption loss in the whole
Si substrate are clearly visible. near-infrared and visible regions.

III. LINEAR PROPERTIES B. Propagation loss

A. Material dispersion Another important property to be optimized for a

photonic platform is the propagation loss of waveguides,
The optical properties of the SiON films were char- that can be attributed to several sources such as: the
acterized by variable-angle spectroscopic ellipsometry material absorption, the radiative loss at bends, the
(VASE). The refractive index was modeled with the loss toward the silicon substrate and the scattering due
New Amorphous model based on the Forouhi-Bloomer to surface roughness. The overall value of propagation
dispersion equations [31] loss, per unit length, can be characterized emulating the
cutback method, by measuring the input/output power
B · (E − Ej ) + C ratio of spiral-like waveguides of different lengths, as
n(ω) = n∞ + , (1) shown in Fig. 3.
(E − Ej )2 + Γ2
This characterization was realized for two different
where n∞ is the refractive index when ω → ∞, Bj and widths of waveguides, namely 1.1 µm and 1.3 µm, re-
Cj contain the material characteristics sulting in average losses of, respectively, 1.8±0.2 dB/cm
and 1.5 ± 0.2 dB/cm in the studied wavelength region
fj 740 nm to 840 nm, with a coupling loss of 3.8 ± 0.3 dB
B= · (Γ2j − (Ej − Eg )2 ), (2) per facet. Note that, at the moment of writing this
Γj
manuscript, an improvement of our etching technique
C = 2fj Γj (Ej − Eg ). (3) has lead to propagation losses of the same SiON waveg-
 4

-6
(a) wg w = 1.1µm
(b)
-8
Transmission [dB] wg w = 1.3µm
-10

-12

-14

-16

-18

-20
0 1 2 3 4 5
Length [cm]

Figure 3. (a) Measurement of the propagation losses α through the cutback method at 780 nm. Two different waveguide
widths (wgw ), 1.1 µm (blue) and 1.3 µm (orange) show α = 1.8 ± 0.1 dB/cm and 1.5 ± 0.1 dB/cm, respectively. Errorbars
and dotted lines represent the standard deviation of the experimental data and the fit, respectively. (b) Optical image of a
2.7 cm long spiral waveguide under laser excitation.

uides as low as 0.8 dB/cm. The pulse-broadening experiment was performed for
different input powers, ranging from 0 dBm to -20 dBm
measured at the output of the waveguide, in order to
IV. NON-LINEAR PROPERTIES verify the intensity dependence of SPM. Considering
that the non-linear effects on the input pulse may oc-
The third-order optical nonlinearities of the SiON cur also in the injection optics, composed of lenses and
photonic platform have been studied by exploiting the fibers, it is important to attenuate the power injected
phenomenon of Self-Phase-Modulation (SPM) [34, 35]. into the waveguides after the external optics. Therefore,
An intense laser pulse, which propagates in a nonlinear the variation of the power coupled to the waveguide
medium, induces a local variation of the refractive index was realized by moving the input-fiber away from the
due to strong light-matter interactions. This variation waveguide’s facet, to decrease the coupling efficiency.
causes a phase shift between the spectral components of Figure 4 shows an example of the experimental data, ob-
the pulse resulting in a modulation of pulse spectrum. tained for two different pulses with central wavelengths
Consequently, by measuring the spectral broadening of at 780 nm and 840 nm.
an ultra-short pulse with known power, one can retrieve In order to estimate the unknown Kerr nonlinearity
the non-linear index of refraction n2 of the material. n2 of the SiON material, we have performed numerical
In particular, following the split-step method described simulations which transform an input pulse spectrum
in Ref. [36], one can simulate the expected SPM effect into a broadened one. Having the knowledge of the
for a given set of parameters, including: the material’s waveguide’s length and propagation loss, of the effective
n2 coefficient, the waveguide geometry and the initial mode area (obtained by numerical simulations) and of
characteristics of the pulse. Then, by comparing the the effective refractive index, the nonlinear Kerr coeffi-
simulated results with the measured SPM spectra, the cient can be then estimated according to the following
nonlinear coefficient of the material can be estimated. procedure. For each set of measurements, the lowest-
The SPM measurements were obtained using a mode- power signal is taken as the reference input pulse-shape.
locked Ti:Sapphire laser, tunable in the wavelength Then, for each power, a set of expected output signals
range 720 nm-840 nm, with a 3-dB pulse-width of are simulated with the split-step method for different
0.2 nm (2 ps) and a repetition rate of 82 MHz. The values of n2 (Fig. 5a). The final value of n2 is thus esti-
laser is directly injected into the SiON waveguide us- mated as the one that minimizes the spectral difference
ing a lensed glass optical fiber. The transmitted pulse between the experimental data and the numerical solu-
is collected with a second identical fiber at the waveg- tion (Fig. 5b). The main source of error in the estima-
uide output and analyzed in an Optical Spectrum An- tion of n2 within this approach is given by the error σα
alyzer with 0.04 nm spectral resolution and sensitivity of the propagation loss imposed in the numerical model.
of -60 dBm. The uncertainty of the estimated n2 was evaluated by
 5

SPM at λ=780nm SPM at λ=840nm

Normalized Power [dB]

Normalized Power [dB]

0 0
(a) (b)
-5

Laser Power [dBm]

-5 -5

-10
-10 -10

-15
-15 -15

-20 -20 -20

-0.5 0 0.5 -0.5 0 0.5
Pulse detuning [nm] Pulse detuning [nm]

Figure 4. Intensity-dependent spectral broadening of an ultra-short laser pulse induced by SPM effect in SiON waveguide
for two different pulse wavelengths centered at (a) 780 nm and (b) 840 nm . The spectra are normalized to their peak
powers. The difference in the input pulses lineshapes (the darkest blue lines) at the two central wavelengths is attributed
to laser and table optics prior to interaction in the waveguides.

(a) (b)

Figure 5. Representation of the numerical fitting procedure to estimate the value of n2 at wavelength of 840 nm. (a)
Starting from the input pulse lineshape (dashed line), a set of output pulse lineshape was simulated for increasing values of
the n2 (blue to red, colorbar in units of 10−20 m2 /W), in order to match the experimental lineshape (black solid line). (b)
Sum of the square residuals between the theoretical and the measured lineshapes. The minimum position corresponds to
the optimal value of n2 that gives the best fit to the experimental data. Estimations are obtained for three different values
of propagation loss α, α + σα and α − σα (red, blue and yellow, respectively), in order to project the error on the measured
loss to the error of estimated n2 .

simulating the spectra at three values of propagation firms the reliability of our analysis. The results are also
loss: α, α + σα and α − σα , and taking the difference consistent with previous works [19, 37, 38], indicating
δn2 = (nmax
2 − nmin
2 )/2 as the estimated error of n2 . that the n2 values of our SiON material are in between
the expected values for pure SiO2 and pure Si3 N4 .
The described numerical SPM simulations method is
based on the experimental inputs, which account for the In the following, we have selected one of the 1100 nm
geometrical dimensions and linear optical properties of wide waveguides (sample A) and performed a spec-
the fabricated waveguides. In order to verify the robust- tral analysis of the variation of n2 with wavelength.
ness of our approach, we applied this method to study Figure 6b shows that the Kerr coefficient strongly in-
a set of four waveguides. In particular, a pair of waveg- creases while reducing the pump wavelength from Near-
uides was studied for two different nominal widths of Infrared to Visible-Red region. This behavior is in
1100 nm and 1300 nm, as described in Table I. The val- accordance with the theoretical model that foresees a
idation test, reported in Fig. 6a, shows that at a fixed maximum in the nonlinear coefficient located close to
wavelength the estimations of n2 are independent on the the TPA edge at Eg /2 [38], corresponding to a wave-
waveguide’s geometry, within error, and therefore con- length of about λTPA ≈ 320 nm for our SiON platform.
 6

waveguide width length n2 (780 nm) n2 (840 nm)

wg A 1.1µm 27mm 13.1 ± 0.5 5.7 ± 0.4
wg B 1.1µm 37mm 13.5 ± 0.5 5.0 ± 0.4
wg C 1.3µm 27mm 12.3 ± 0.7 5.8 ± 0.5
wg D 1.3µm 37mm 13.5 ± 0.7 5.6 ± 0.5

Table I. Geometrical dimensions and estimated n2 coefficients (in units of 10−20 m2 /W), at two different pump wavelengths,
for the different investigated devices.

15 15
(a) 780nm 840nm (b) 14
13

n2 [ 10-20 m2/W ]
n2 [ 10-20 m2/W ]

12
10
11
10
9
5 8
7
6

0 5
wg A wg B wg C wg D 750 800 850
Wavelength [nm]

Figure 6. (a) Results of n2 estimation for different geometries of waveguides (see Table I) at two different laser wavelengths:
780 nm (blue) and 840 nm (orange). The implemented analysis method allows to estimate n2 values which are independent
on the waveguide geometry, ensuring that the geometrical properties of the waveguides are weighted properly in the data
analysis. (b) The measured spectral dispersion of the nonlinear refractive index n2 . An increasing trend towards shorter
wavelengths is clearly visible.

V. DISPERSION-ENGINEERED SION RING interactions, the first two orders β1 and β2 play an im-
RESONATORS FOR GENERATION OF portant role and represent, respectively, the group index
CORRELATED PHOTON PAIRS ng and the Group Velocity dispersion (GVD):
 
The knowledge of the linear and non-linear properties 1 dneff
β1 (ω) = ng /c = neff + ω , (5a)
of the developed SiON photonic platform makes it pos- c dω
sible to engineer the modal refractive index, neff (ω), of dβ1 1

dneff d2 neff

the waveguide to match particular applications. Non- β2 (ω) = = 2 +ω . (5b)
linear schemes of generation of non-classical states of dω c dω dω 2
NIR photons often relay on SiN integrated micropho-
tonic devices [27, 30, 39, 40]. In this section, we de- In a ring resonator, the group velocity β1 defines
scribe our approach for the engineering and realization the spectral separation δc between successive resonant
of SiON-based ring resonator devices for on-chip gener- modes – the free-spectral range (FSR) – following the re-
ation of entangled photon pairs via Spontaneous Four lation δc = 2π/(β1 Lc ), where Lc is the physical length
Wave Mixing (SFWM). of the cavity. A flat dispersion of β1 (ω) over a range
The dependence of neff on the light angular frequency of frequencies provides energy-equidistant cavity reso-
leads to the dispersive nature of the mode’s propagation nances, which is essential to fulfill the energy conserva-
constant β(ω) = neff (ω) · ω/c, which can be expressed tion requirement, for example, in nonlinear FWM and
as a Taylor series around a central frequency ω0 frequency comb generation experiments [30, 39], where
the pump, the signal and the idler should satisfy the
X βm
β(ω) = β0 + (ω − ω0 )m , (4) relation ωi = 2ωp − ωs .
m=1
m! The parameter β2 describes how the different spectral
components of a propagating pulse travel, and in either
where the diverse orders account for different propa- case β2 > 0 (normal dispersion) or β2 < 0 (anomalous
gation phenomena. In particular, for nonlinear wave dispersion) result in temporal broadening of the pulse.
 7

Figure 7. (a) Sketch of the cross-sectional geometry of the SiON waveguide embedded in SiO2 matrix. (b) The same
geometry with the top SiO2 cladding substituted with air. (c) The calculated GVD’s of quasi-TE-modes for the respective
geometries, showing that anomalous GVD is achieved when the top SiO2 cladding is substituted with air. For comparison,
the dispersion of the bulk SiON material is also shown (dashed line). (d) The calculated FSR of two rings evidence the
sufficiently flat δc -trend for the air-cladded device.

The case of β2 = 0 at some frequency – the zero disper- (Fig. 7b). This new configuration turned to be partic-
sion frequency (ZDF) – is of particular interest for non- ularly interesting since the GVD appeared to be much
linear optical applications because around the ZDF dif- more sensitive to the variations of the waveguide width.
ferent spectral components experience largely reduced We found that the GVDs of air-cladded resonators are
second-order dispersion. In SFWM experiments with already anomalous for smaller widths (∼ 1100 nm) and
microring resonators around ZDF the spectral spread turn slowly into a normal one over 1350 nm. In Fig. 7c
of δc is minimal and the flatness of ng provides with we compare the calculated GVDs for 1300 nm waveg-
larger nonlinear generation bandwidth [36, 41]. uide width rings in both configurations with top SiO2
We performed numerical axisymmetric simulations (red) and air-claddings (blue). A zero-dispersion point
based on the Finite Elements Method in order to is observed at around λ ≈ 822 nm for a 1300 nm waveg-
develop proper geometries for a device operating at uide width rings. The air-cladded ring shows sufficiently
NIR wavelengths, supporting single-mode characteris- flat dispersion of the group index and, consequently, a
tics and showing an anomalous GVD over the spectral stable FSR over a large span of wavelengths from 800
range of interest. The radius of the studied ring res- to 850 nm’s (Fig. 7d). The effective nonlinearity γ for
onators was set to 25 µm. This choice is motivated by the 1.3 × 0.5 µm2 ring at the zero-dispersion point of
the necessity to keep the ring radius large enough to 822 nm is estimated to be of the order of 1 m−1 W−1 .
avoid radiative losses but small enough to provide with
δc ∼ 2.5 nm in order to minimize the spectral overlap of
pump pulses with more than one mode around 800 nm
of wavelength. A schematic of the dispersion-engineered resonator,
We first investigated a conventional geometry, in which has been fabricated on the same chip together
which the waveguide of a slightly trapezoidal form is with spiral waveguides, is shown in Fig. 8a. The top
fully embedded within the SiO2 cladding (Fig. 7a). The SiO2 cladding around the resonators was removed via
lateral boundaries of the waveguide have an inclination a selective wet-etching procedure in buffered HF solu-
angle of 86o , which is the typical value according to tion during which the 50 nm Si3 N4 film acts as an ideal
our fabrication process. The height of the SiON waveg- etch-stop film to prevent underetching of waveguiding
uide was fixed to 500 nm. Simulations were performed components. The opening window around the ring was
by varying the ring waveguide width from 900 nm to designed such that the remaining top SiO2 cladding af-
1700 nm and the azimuthal number M of the fundamen- ter wet etching is retracted to a safe distance of at least
tal radial mode from 270 to 330, covering a wavelength 2.5 µm from the waveguide (Fig. 8b). Figure 8c shows
span of 200 nm around λ = 850 nm. The obtained an optical micrograph of arrays of ring resonators both
results for both the quasi transverse electric (TE) and in add-drop and all-pass configurations, while a blowup
magnetic (TM) modes showed that the GVD remains around an air-cladded all-pass ring resonator is shown
normal for all ring widths within the spectral range be- in Fig. 8d. The linear characterization and non-linear
tween 750 nm and 950 nm. sFWM experiments on these devices are currently on-
Next, we performed simulations of the same ring going, however, their discussion is out of the scope of
geometry substituting the top SiO2 cladding with air the current manuscript.
 8

Figure 8. (a) A 3D sketch of the ring resonator with top-air cladding and a TiN phase shifter. (b) Cross-sectional geometry
at the ring-waveguide coupler position. (c) Optical image of the fabricated chip with arrays of ring resonators and (d) the
zoom around a ring resonator in all-pass configuration.

VI. CONCLUSIONS the good linear properties, the promising optical nonlin-
earities for on-chip photon generation and our recently
developed technology for on-chip photon detection [20],
we envision the potential of this platform to achieve, in
In this work, we have demonstrated a new silicon the near future, a full integration of photon generation
oxynitride-based integrated photonic platform for lin- sources, manipulation and detection on a single Silicon
ear and nonlinear application in the VIS-NIR wave- chip, operating at room temperature, for classical and
length range. The fabricated devices show a low prop- quantum photonics applications.
agation loss < 2 dB/cm, comparable to commercially
available devices at 800 nm wavelength and with on-
going improvement of at least a factor two. We have VII. ACKNOWLEDGMENT
demonstrated that, despite the reduction of the re-
fractive index with respect to SiN, the SiON waveg- Funding. European Commission (777222, 899368).
uides preserve a relatively strong optical nonlinearity of Acknowledgments. The authors are thankful to
13 ± 06 × 10−20 m2 /W around the wavelength 780 nm. M. Bernard and G. Pucker for the fruitful discussions
Furthermore, thanks to the possibility to remove locally regarding the design and fabrication of the photonic de-
the cladding without damaging the waveguide, our plat- vices.
form allows for a larger versatility in engineering the Disclosures. The authors declare no conflicts of inter-
waveguide dispersion. This enables to investigate one est.
of the specific and peculiar applications of nonlinear Data availability. Data underlying the results pre-
FWM by properly adjusting the group index and group sented in this paper are not publicly available at this
velocity dispersion in order to enhance nonlinear pho- time but may be obtained from the authors upon rea-
ton pair generation in ring resonators. By combining sonable request.

[1] J. Capmany and D. Pérez, Programmable Integrated [2] R. Osgood jr and X. Meng, Principles of Photonic In-
Photonics (Oxford University Press, 2020).
 9

tegrated Circuits: Materials, Device Physics, Guided [20] M. Bernard, F. Acerbi, G. Paternoster, G. Pic-
Wave Design (Springer Nature, 2021). coli, L. Gemma, D. Brunelli, A. Gola, G. Pucker,
[3] S. Y. Siew, B. Li, F. Gao, H. Y. Zheng, W. Zhang, L. Pancheri, and M. Ghulinyan, Optica 8, 1363 (2021).
P. Guo, S. W. Xie, A. Song, B. Dong, L. W. Luo, et al., [21] A. Densmore, M. Vachon, D.-X. Xu, S. Janz, R. Ma, Y.-
Journal of Lightwave Technology 39, 4374 (2021). H. Li, G. Lopinski, A. Delâge, J. Lapointe, C. Luebbert,
[4] W. Bogaerts and L. Chrostowski, Laser & Photonics et al., Optics letters 34, 3598 (2009).
Reviews 12, 1700237 (2018). [22] A. L. Washburn, L. C. Gunn, and R. C. Bailey, Ana-
[5] N. M. Fahrenkopf, C. McDonough, G. L. Leake, Z. Su, lytical chemistry 81, 9499 (2009).
E. Timurdogan, and D. D. Coolbaugh, IEEE Journal [23] H. Mukundan, A. S. Anderson, W. K. Grace, K. M.
of Selected Topics in Quantum Electronics 25, 1 (2019). Grace, N. Hartman, J. S. Martinez, and B. I. Swanson,
[6] T. Aalto, M. Cherchi, M. Harjanne, S. Bhat, Sensors 9, 5783 (2009).
P. Heimala, F. Sun, M. Kapulainen, T. Hassinen, and [24] R. Heideman, M. Hoekman, and E. Schreuder, IEEE
T. Vehmas, IEEE Journal of selected topics in quantum Journal of Selected topics in quantum electronics 18,
electronics 25, 1 (2019). 1583 (2012).
[7] D. J. Blumenthal, R. Heideman, D. Geuzebroek, [25] A. C. Turner, C. Manolatou, B. S. Schmidt, M. Lipson,
A. Leinse, and C. Roeloffzen, Proceedings of the IEEE M. A. Foster, J. E. Sharping, and A. L. Gaeta, Optics
106, 2209 (2018). express 14, 4357 (2006).
[8] T. J. Morin, L. Chang, W. Jin, C. Li, J. Guo, H. Park, [26] S. Mas, J. Caraquitena, J. V. Galán, P. Sanchis, and
M. A. Tran, T. Komljenovic, and J. E. Bowers, Optica J. Martí, Optics express 18, 20839 (2010).
8, 755 (2021). [27] H. Zhao, B. Kuyken, S. Clemmen, F. Leo, A. Sub-
[9] D. J. Moss, R. Morandotti, A. L. Gaeta, and M. Lipson, ramanian, A. Dhakal, P. Helin, S. Severi, E. Brainis,
Nature photonics 7, 597 (2013). G. Roelkens, et al., Optics letters 40, 2177 (2015).
[10] C. G. Roeloffzen, M. Hoekman, E. J. Klein, L. S. [28] Y. Zhang, J. Zou, and J.-J. He, Optics Express 26,
Wevers, R. B. Timens, D. Marchenko, D. Geskus, 26057 (2018).
R. Dekker, A. Alippi, R. Grootjans, et al., IEEE journal [29] K. Guo, J. B. Christensen, X. Shi, E. N. Christensen,
of selected topics in quantum electronics 24, 1 (2018). L. Lin, Y. Ding, H. Ou, and K. Rottwitt, IEEE Pho-
[11] J. Liu, G. Huang, R. N. Wang, J. He, A. S. Raja, T. Liu, tonics Journal 10, 1 (2018).
N. J. Engelsen, and T. J. Kippenberg, Nature commu- [30] Y. Zhao, X. Ji, B. Y. Kim, P. S. Donvalkar, J. K. Jang,
nications 12, 1 (2021). C. Joshi, M. Yu, C. Joshi, R. R. Domeneguetti, F. A.
[12] M. H. Pfeiffer, A. Kordts, V. Brasch, M. Zervas, Barbosa, et al., Optica 7, 135 (2020).
M. Geiselmann, J. D. Jost, and T. J. Kippenberg, Op- [31] A. Forouhi and I. Bloomer, Physical review B 38, 1865
tica 3, 20 (2016). (1988).
[13] R. Heideman, A. Melloni, M. Hoekman, A. Borreman, [32] G.-L. Bona, R. Germann, and B. J. Offrein, IBM Jour-
A. Leinse, and F. Morichetti, in Proc. of the 9th nal of Research and Development 47, 239 (2003).
IEEE/LEOS Symp. in the Benelux (2005) pp. 71–74. [33] D. Chen, A. Kovach, S. Poust, V. Gambin, and A. M.
[14] F. Morichetti, A. Melloni, M. Martinelli, R. G. Heide- Armani, Applied Physics Letters 115, 051105 (2019).
man, A. Leinse, D. H. Geuzebroek, and A. Borreman, [34] R. H. Stolen and C. Lin, Physical Review A 17, 1448
Journal of Lightwave technology 25, 2579 (2007). (1978).
[15] K. Wörhoff, A. Driessen, P. Lambeck, L. Hilderink, [35] N. Tzoar and M. Jain, Physical Review A 23, 1266
P. W. Linders, and T. J. Popma, Sensors and Actu- (1981).
ators A: Physical 74, 9 (1999). [36] G. P. Agrawal, in Nonlinear Science at the Dawn of the
[16] R. Germann, H. Salemink, R. Beyeler, G. Bona, 21st Century (Springer, 2000) pp. 195–211.
F. Horst, I. Massarek, and B. Offrein, Journal of the [37] K. Ikeda, R. E. Saperstein, N. Alic, and Y. Fainman,
Electrochemical Society 147, 2237 (2000). Optics express 16, 12987 (2008).
[17] A. Samusenko, D. Gandolfi, G. Pucker, T. Chalyan, [38] M. Sheik-Bahae, D. J. Hagan, and E. W. Van Stryland,
R. Guider, M. Ghulinyan, and L. Pavesi, Journal of Physical review letters 65, 96 (1990).
lightwave technology 34, 969 (2016). [39] R. Cernansky, F. Martini, and A. Politi, Optics Letters
[18] D. Chen, A. Kovach, X. Shen, S. Poust, and A. M. 43, 855 (2018).
Armani, ACS Photonics 4, 2376 (2017). [40] X. Lu, Q. Li, D. A. Westly, G. Moille, A. Singh,
[19] A. Trenti, M. Borghi, S. Biasi, M. Ghulinyan, V. Anant, and K. Srinivasan, Nature physics 15, 373
F. Ramiro-Manzano, G. Pucker, and L. Pavesi, AIP (2019).
Advances 8, 025311 (2018). [41] I. H. Agha, Y. Okawachi, and A. L. Gaeta, Optics
Express 17, 16209 (2009).