This PDF file contains the front matter associated with SPIE Proceedings Volume 10108, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and Conference Committee listing.

Ring resonator modulators reach high modulation efficiencies, are very compact and can be electrically driven as lumped elements. However, their limited optical bandwidth requires temperature stabilization, limiting their power efficiency. A novel ring assisted Mach-Zehnder modulator (MZM) aggressively reduces power consumption. Moreover, an integration scheme passively sets the 3 dB point during attachment of the input fiber relative to a multimode grating coupler used as the first splitter element of the interferometer. Straight phase shifters are replaced by arrays of highly overcoupled resonators maintaining a sufficiently high finesse and a substantial resonant enhancement while minimizing the excess losses at the resonator to waveguide junctions. A large resonance bandwidth compatible with thermal operation over 50 °C without dynamic compensation is obtained together with a factor larger than four in the reduction of power consumption relative to a conventional MZM.

In this presentation we will report on our recent work on new materials that can be monolithically integrated on high-index contrast silicon or silicon nitride photonic ICs to enhance their functionality. This includes graphene and other 2D-materials for realizing compact electro-absorption modulators and non-linear devices, ferroelectric materials for realizing phase modulators and adiabatic couplers for realizing bistable switches.

Silicon microring resonator modulators are versatile active on-chip devices capable of high-speed modulation with low energy consumption. However, the effects of PN junction alignment variance for different doping concentrations during fabrication have not been looked into. In this work, we numerically demonstrate and analyse the optimisation of the silicon microring resonator modulator based on the carrier depletion mechanism for high extinction ratio and low energy consumption at the communication wavelength of 1550 nm. A range of carrier doping concentrations and offset of the PN junction to the waveguide centre can be used to optimise the modulation efficiency, energy consumption and insertion losses of the microring modulator. In particular, the effects of the offset of the PN junction are analysed for three cases in the carrier-depletion silicon phase shifter: (i) p-type doping < n-type doping, (ii) p-type doping = n-type doping, and (iii) p-type doping > n-type doping. Subsequently, three types of microring ring modulator architecture – the all-pass microring resonator, the add-drop microring resonator, and the all-pass dual uncoupled microring resonator – are realised and analysed. Our results suggests that doping concentration between 2 × 10¹⁷ cm^-3 to 5 × 10¹⁷ cm^-3, with the p-doping concentration lower than the n-doping concentration, should be employed in order to achieve a tunability of > 16 pm/V and extinction ratio of > 8 dB.

In this paper, we propose and demonstrate an all-optical switch using graphene oxide infiltrated subwavelength grating waveguide. Benefiting from the extremely large Kerr coefficient of graphene oxide (four orders of magnitude larger than conventional materials) and large mode volume overlap factor of the subwavelength grating waveguide (4~10 times larger than conventional strip waveguides), the switch is capable of achieving THz speed with less than 1 fJ energy consumption per bit, which is more than three orders of magnitude smaller than THz switches reported so far.

Silicon-based high-speed optical modulators, fabricated using CMOS compatible nanofabrication technology, are the key components for integrated photonics, especially the high-speed intra- and inter-chip optical interconnects. In this paper, we propose a high speed, low power consumption electro-optic (EO) modulator based on the EO polymer/silicon hybrid subwavelength grating (SWG) waveguide ring resonator. The core of the SWG waveguide consists periodically arranged silicon pillars along the light propagation direction. EO polymer (SEO125) is used as the top cladding. An intriguing advantage that the SWG waveguide has over the conventional silicon strip waveguide is its large mode volume overlap with EO polymer. Besides, compared with the plasma dispersion effect, electro-optic polymers have a large electro-optic coefficient and ultrafast response speed. Furthermore, among the different modulator structures, ring resonator is one of the most promising structures as it has a small footprint which is the key for VLSI (Very Large Scale Integration), and it allows complex optical functionalities monolithically integrated with advanced electronics at a competitive cost. Thus, the proposed EO polymer infiltrated SWG waveguide ring resonator based modulator is a very promising candidate for low cost, small size, light weight, and low power consumption (CSWaP) optical interconnect.

Silicon photonics enables the development of optical components on a chip with the potential for large-scale optical integrated circuits that can be fabricated at the wafer-scale using foundries similar to those used in the electronics industry. Although silicon is a passive optical material with an indirect bandgap, reconfigurable devices have been demonstrated using thermo-optic effects (large phase shifts, but relatively slow with large power consumption) and carrier plasma dispersion effects (high-speed, but small phase shifts). We recently demonstrated a low-power approach for inducing large phase shifts (>2π) using a technique that we call micro-opto-electro-mechanical index perturbation (MOEM-IP). In this initial work we characterized silicon nitride waveguides in which the propagating optical mode’s evanescent field is vertically coupled to silicon nitride microbridges. This interaction leads to an effective index tuning that is a strong function of the waveguide-microbridge separation. We now extend our MOEM-IP approach to different configurations (i.e. in-plane coupling) and material systems (i.e. silicon-oninsulator). Mode perturbation simulations indicate that the MOEM-IP approach is widely applicable to many configurations and material systems enabling large effective index tuning (Δn_effective>0.1) requiring microbridge displacements of only a few hundred nanometers. We also examine several device applications that take advantage of MOEM-IP. These include tunable optical filters using high-Q microring cavities and optical phased arrays that enable chip-scale beam steering in two-dimensions using low-power phase shifting enabled by MOEM-IP.

We have successfully fabricated and measured our silicon bridge waveguide polarization beam splitter (PBS). Our proposed PBS is based on a bend directional coupler with a bend bridge waveguide and is experimentally realized using silicon waveguide thickness of 220 nm and 250 nm, which are the commonly used silicon thickness for silicon photonics manufacturing. Our experimental results demonstrated high extinction ratio of > 20 dB for the TE-like mode, and > 15 dB for the TM-like mode across a broad bandwidth of 90 nm that covers the entire C-band with a small footprint of ~18×9 μm². On-chip high performance PBS is important for polarization diversity in integrated photonics, and for communication applications such as dual-polarization quadrature phase-shift keying (DP-QPSK) modulation.

We show theoretically and experimentally how a flat-top second-order response can be achieved with a self-coupled single add-drop ring resonator based on two couplers with different splitting ratios. The resulting device is a 1x1 filter, reflecting light back in the input waveguide at resonating wavelengths in the passbands, and transmitting light in the output waveguide at all other non-resonating wavelengths. Different implementations of the filter have been designed and fabricated on a micron-scale silicon photonics platform. They are based on compact Euler bends - either U-bends or Lbends - and Multi-Mode Interferometers as splitters for the ring resonators. Different finesse values have been achieved by using either 50:50 MMIs in conjunction with 85:15 MMIs or 85:15 MMIs in conjunction with 95:05 double MMIs. Unlike ordinary lowest order directional couplers, the MMIs couple most of the power in the cross-port which make them particularly suitable for the topology of the self-coupled ring, which would otherwise require a waveguide crossing. Experimental results are presented, showing good agreement with simulations. The proposed devices can find applications as wavelength-selective reflectors for relatively broad-band lasers or used as 2x2 add-drop filters when two exact replicas of the device are placed on the arms of a Mach-Zehnder interferometer.

The demonstration of a CMOS compatible laser working at room temperature has been eagerly sought since the beginning of silicon photonics. Although bulk Germanium (Ge) is an indirect bandgap material, Tin (Sn) can be incorporated into it to turn the resulting alloy into a direct band-gap semiconductor. Recently, lasing was demonstrated at cryogenic temperatures using thick GeSn layers with Sn contents of 8.5% and above. Optical micro-cavities were later added to reduce the laser threshold. Here, an under-etching of thick GeSn layers selectively with regard to Ge confines optical modes and relaxes the compressive strain built inside the layers, resulting in more direct band-gaps behavior. Such photonic components rely on technological processes dedicated to GeSn. In this paper, we present our recent developments on (i) anisotropic etching of GeSn and (ii) isotropic etching of Ge selective with regard to GeSn. Even for GeSn with a Sn content as low as 6%, the etching selectivity is of 57. For 8% Sn content, the selectivity reaches 433. We used these processes to fabricate micro-disk optical cavities in thick GeSn layers. Under continuous wave pumping, optical modes were detected from photoluminescence spectra.

To enlarge the tensile strain in Ge light emission diodes (s-Ge LED) we applied a GeSn virtual substrate (VS) on Si (001) with a Sn content of 4.5 %, to produce s-Ge LEDs. The LED stack was grown by molecular beam epitaxy. Electroluminescence investigations of the s-Ge LED show a major direct Ge peak and a minor peak at lower energy, which is formed by the GeSn-VS and the s-Ge indirect transition. The main peak of a 100 nm thick s-Ge LED is red-shifted as compared to the Ge peak of an unstrained reference Ge LED grown on Ge-VS. At a temperature of T = 80 K the increased tensile strain, produced by the GeSn-VS, causes a redshift of the direct Ge peak from 0.809 eV to 0.745 and 0.769 eV, namely for the s-Ge LED with a 100 and 200 nm thick active layer. At T = 300 K the direct Ge peak is shifted from 0.777 eV of the reference Ge LED to 0.725 eV (for 100 nm) and 0.743 eV (for 200 nm). The peak positions do not differ much between the 50 and 100 nm thick s-Ge LEDs. The intensities of the direct Ge peak increase with the s-Ge layer thickness. Moreover, the intensity of the 50 nm thick s-Ge sample is found to be larger than that of the 100 nm thick reference Ge LED.

In this work we study Ge structures grown on silicon substrates. We use photoluminescence and photoreflectance to determine both direct and indirect gap of Ge under tensile strain. The strain is induced by growing the Ge on an InGaAs buffer layer with variable In content. The band energy levels are modeled by a 30 band k·p model based on first principles calculations. Characterization techniques show very good agreement with the calculated energy values.

To harness the advanced fabrication capabilities and high yields of the electronics industry for photonics, monolithic growth and CMOS compatibility are required. One promising candidate which fulfils these conditions is GeSn. Introducing Sn lowers the energy of the direct Γ valley relative to the indirect L valley. The movement of the conduction band valleys with Sn concentration is critical for the design of efficient devices; however, a large discrepancy exists in the literature for the Sn concentration at which GeSn becomes a direct band gap. We investigate the bandgap character of GeSn using hydrostatic pressure which reversibility modifies the bandstructure. In this work we determine the movement of the band-edge under pressure using photocurrent measurements. For a pure Ge sample, the movement of the band-edge is dominated by the indirect L valley with a measured pressure coefficient of 4.26±0.05 meV/kbar. With increasing Sn concentration there is evidence of band mixing effects, with values of 9.4±0.3 meV/kbar and 11.1±0.2 meV/kbar measured for 6% and 8% Sn samples. For a 10% Sn sample the pressure coefficient of 13±0.5 meV/kbar is close to the movement of the direct bandgap of Ge, indicating predominately direct Γ-like character for this GeSn alloy. This further suggests a gradual transition from indirect to direct like behaviour in the alloy as also evidenced from theoretical calculations. The implications of this in terms of optimising device performance will be discussed in further detail at the conference.

Optical interconnects made of silicon are viewed as emerging efficient solutions for addressing the communication bottlenecks that plague high-performance computing systems and big-data centers. Due to large index contrast and optical nonlinearity of silicon, waveguides and active devices based on silicon can be scaled down to sub-wavelength size, making silicon photonics an ideal platform towards integrated on-chip photonic circuits. In order for this potential to be fulfilled, one needs to understand the factors that affect the quality of optical signals propagating in silicon optical interconnects, namely the bit-error ratio (BER), as well as the relationship between the parameters characterizing the optical signal and the BER. In this work, an accurate approach to calculate the BER in single-channel silicon optical interconnects utilizing arbitrarily-shaped pulsed signals is presented. The optical interconnects consist of either strip single-mode silicon photonic waveguides (Si-PhWs) or silicon photonic crystal (PhC) waveguides (Si-PhCWs), and are linked to a direct-detection receiver. The optical signal consists of a superposition of Gaussian pulses and white noise. The signal dynamics in the silicon waveguides is modelled using a modified nonlinear Schrodinger equation, whereas the Karhunen-Loeve series expansion method is employed to calculate the system BER. Our analysis reveals that in the case of the Si-PhWs the pulse width is the main parameter that determines the BER, whereas in the case of Si-PhCWs the BER is mostly affected by the waveguide properties via the pulse group-velocity. A good system performance is achieved in centimeter-long Si-PhWs whereas similar system performance is obtained using 100× and 200× shorter Si-PhCWs operating in the fast- and slow-light regimes, respectively.

High-brightness lasers are widely used in fields such as spectroscopy, infrared countermeasures, free-space communication, and industrial manufacturing. Integration of a broad-band, multi-spectral laser is made possible by heterogeneously integrating multiple gain materials on one silicon (Si) substrate chip. A single multi-spectral output with high beam quality can be achieved by wavelength beam combining in multiple stages: within the gain bandwidth of each laser material and then coarsely combining each spectral band to a single output waveguide. To make power scaling feasible with this system, heterogeneously integrated lasers spanning the near- to the mid-infrared with corresponding low-loss wavelength beam combining elements on chip must be demonstrated. In this work, a review of multi-spectral lasers integrated on Si is presented and various waveguide materials are discussed for spanning the visible to the mid-infrared. Recent work integrating 2.0μm diode and 4.8μm quantum cascade lasers on Si extend the previously available 1.3μm and 1.5μm diode lasers on Si to the mid-infrared. Spectral beam combining elements for spanning the visible to the mid-infrared with low loss are discussed.

A simple, easy-to-use and physically meaningful analytical (mathematical) model has been developed for the prediction of thermal stresses in an elastic bonded elongated cylindrical tri-material body of finite length. The model has been developed in application to an optical silica fiber embedded (soldered) into Silicon. The body is fabricated at an elevated (soldering, curing) temperature and is subsequently cooled down to a low (room, operation or testing) temperature. Thermal stresses arise because of the thermal contraction mismatch of the dissimilar materials in the body. The addressed stresses include radial, tangential and axial normal stresses acting in the body’s mid-portion, as well as the interfacial shearing stresses that concentrate at the body’s ends. The numerical example is carried out for the case of a “soft” (silver-tin) solder in application to structures, in which an optical silica fiber is soldered into a silicon chip (silicon photonics technology). It is concluded that the appropriate solder material and its thickness should be selected based on the predictions obtained on the basis of the developed model, so that the radial and the longitudinal interfacial stresses in the solder ring are sufficiently low.

The Multi-Pixel Photon Counter (MPPC), which is also called a silicon photomultiplier (SiPM)^1,2, is one promising candidate for automotive light detection and ranging (LIDAR)³. Due to high internal gain around 10⁶, photon counting is possible and satisfies long range measurement. Compared to photo diodes (PDs) and avalanche photo diodes (APDs), read-out circuits for MPPCs are very simple because no external amplifier is needed. Conventional MPPCs have been developed for targeting blue scintillation light around 400 nm for positron emission tomography (PET) and high energy physics experiments. In this paper we report new near-infrared (NIR)-enhanced MPPCs whose development targets include 905 nm laser light for automotive LIDAR systems. Conventional MPPCs have a p-on-n structure and show 2% photon detection efficiency (PDE) at 905 nm. Our newly developed n-on-p MPPC achieved 7% PDE without greatly changing the impurity concentration profile of the depletion layer. This n-on-p MPPC has been released as an NIRenhanced MPPC: S13720-1325CS. For further improvement of NIR sensitivity, we tried several silicon wafers and process conditions of p-n junction profiles. Even though dark noise and the voltage range have to be modified, the latest sample shows 11% PDE, suggesting potential for further sensitivity improvement.

Recently, CMOS time-resolved imaging devices are being widely used for scientific and medical applications. A fluorescence lifetime imaging microscopy (FLIM), which is a powerful analysis tool in fundamental physics as well as in the life science, is a typical application for the time-resolved imaging devices. For better time-resolution in the lock-in pixel design, a multi-tap pixel architecture is very effective and useful. In this paper, we have proposed an 8-tap CMOS lock-in pixel with lateral electric field charge modulator (LEFM) and demonstrated the effectiveness of designed pixel by CAD simulation. The proposed pixel makes possible to measure the highly time-resolved images with a high signal to noise ratio (SNR) and to observe various images of cells even if a sample has a multi-lifetime component. An 8-tap time-resolved CMOS image sensor chip is developed by 0.11μm 1P4M CIS process technology.

Due to the exponential growth of the bandwidth requirement for wireless communication systems, new frequency bands need to be utilized. For future 5G wireless networks, frequencies of 30 GHz to 90 GHz are considered, while for satellite and aircraft communications the sub-terahertz frequencies are considered. However, with increasing millimeter-wave frequencies (30 GHz – 300 GHz), high-speed electronic solutions become energy-inefficient, and alternative solutions are required.

Photonics offers the bandwidth and a potentially seamless integration with the fiber-wireless technology (Fi-Wi) for 5G communications. Commercially available terahertz generators are often based on photonics, i.e., lasers, too. One particularly promising technique to generate the microwave or sub-terahertz signal is to use the comb generated by modulating a continuous-wave laser signal. By filtering two non-adjacent comb lines, a beat signal is generated that has a frequency that is an integer multiple of the electrical modulator driving signal. In this way, frequency multiplication is achieved using microwave photonics. Photodetectors and/or photomixers can then be used to convert the beat signal to a millimeter-wave. However, the energy-efficiency of these techniques – and how they compare to all-electronic solutions – has not been analyzed yet.

In this paper we will present this energy-efficiency analysis, based on a silicon photonics implementation. Silicon photonics has the potential to miniaturize such systems, for ubiquitous and low-cost implementation. Silicon-based modulators, however, are not ideal phase modulators, and simulation tools need to incorporate this. The regimes, in terms of signal power and frequency, where photonics compares favorably over electronics, will be discussed.

Infrared (IR) thermal emitters are widely used in monitoring applications. For autonomous systems, miniaturized devices with low power consumption are needed. We have designed, fabricated and tested a novel device design, packaged on the wafer level by Al-Al thermo-compression bonding. 80 μm wide Aluminium frames on device and cap wafers were bonded in vacuum at 550°C, applying a force of 25 kN for 1 hour. The bond force translated to a bond pressure of 39 MPa. Subsequent device operation showed that the seals were hermetic, and that the emitters were encapsulated in an inert atmosphere.

The emitters were optimized for radiation at λ=3.5 μm. Emission spectra by Fourier Transform Infrared Spectroscopy showed high emissivity in the wavelength range 3 – 10 μm at 35 mA driving current and 5.7 V bias, i.e. 200 mW power consumption. The emitter temperature was around 700 °C. The rise and fall times of the emitters were below 8 and 3 ms, respectively. The low thermal mass indicates that pulsed operation at frequencies around 100 Hz could be realized with about 90 % modulation depth. The measured characteristics were in good agreement with COMSOL simulations. Thus, the presented devices have lower power consumption, an order of magnitude higher modulation frequency, and a production cost reduced by 40 – 60%^1-4 compared to available, individually packaged devices. The patented device sealing provides through-silicon conductors and enables direct surface mounting of the components.

We present a theoretical model for two high-throughput optical logic methodologies, using voltage-induced free-carrier dispersion and stimulated Raman scattering based Zeno switching. Increased computational throughput is achieved by accessing higher switching speeds, optimizing the use of space, and by using multiple wavelengths for parallel processing. The condition of CMOS compatibility is maintained to take advantage of the high-volume, low-cost manufacturing potential of the industry and to help lower each design's spatial footprint (enabled by the high refractive index contrast of silicon-on-insulator waveguides and resonators). Each design is made with the potential of higher-order operations in mind; for their use must not only stand alone, but must also have the ability to incorporate into future all-optical or optoelectronic computational devices.

Developing secure communications is a research area of growing interest. During the past years, several cryptographic schemes have been developed, with Quantum cryptography being a promising scheme due to the use of quantum effects, which make very difficult for an eavesdropper to intercept the communication. However, practical quantum key distribution methods have encountered several limitations; current experimental realizations, in fact, fail to scale up on long distances, as well as in providing unconditional security and speed comparable to classical optical communications channels. Here we propose a new, low cost and ultra-fast cryptographic system based on a fully classical optical channel. Our cryptographic scheme exploits the complex synchronization of two different random systems (one on the side of the sender and another on the side of the receiver) to realize a “physical” one paid system. The random medium is created by an optical chip fabricated through electron beam lithography on a Silicon On Insulator (SOI) substrate. We present experiments with ps lasers and commercial fibers, showing the ultrafast distribution of a random key between two users (Alice and Bob), with absolute no possibility for a passive/active eavesdropper to intercept the communication. Remarkably, this system enables the same security of quantum cryptography, but with the use of a classical communication channel. Our system exploits a unique synchronization that exists between two different random systems, and at such is extremely versatile and can enable safe communications among different users in standards telecommunications channels.

We present a sidewall patterned shifted Bragg grating based on an add-drop filter in silicon-on-insulator platform with a coating of amorphous titanium dioxide. This particular waveguide grating is equivalent to two identical gratings written across either sides of a waveguide with a longitudinal offset of half of a period. The add-drop operation occurs on the basis of mode conversion due to shifted sidewall structure followed by mode splitting with asymmetric Y-junction. A signal launched through the wide arm (single mode) of an asymmetric Y-junction generates the fundamental mode at the stem of the Y-branch. First order mode is generated at the stem if the signal is launched through the narrow arm. Thus, an asymmetric Y-branch is used as a mode splitter fulfilling proper limiting condition for an adiabatic operation. A signal at the Bragg wavelength launched through the wide arm of asymmetric Y-junction generates fundamental mode at the stem. The fundamental mode converted to first order upon reflection from the shifted Bragg grating. The reflected mode couples into the narrow arm of the Y-junction. The bandwidth of the reflected signal depends on the grating strength. We used 80 nm grating amplitude for 800 nm wide waveguide. The height of the guiding layer is 220 nm. The TiO2 thickness is set to 180 nm. A reflection bandwidth of 2.2 nm with 14 dB extinction ratio is obtained at 1552.5 nm for 300 µm long grating. We further demonstrate the potential of TiO2 recoating with atomic layer deposition as a method of fine tuning the spectrum.

Integrated circuits based on micron-scale silicon waveguides have the clear advantage of being tolerant to fabrication errors, thanks to the high mode confinement within the guiding core. Here we show how flat-top interleavers can be achieved on a micron-scale silicon photonics platform based on ring-loaded Mach-Zehnder Interferometers (MZIs), without the need for any thermal tuning. Robust designs are also guaranteed by resorting to Multi-Mode Interferometers (MMIs) as power splitters in both the MZIs and the ring resonators. A trade-off between in-band ripple and roll-off can be achieved by changing the ring splitting ratios. In particular rings with different finesse based on MMIs with 50:50, 72:28, and 85:15 splitting ratios have been designed, fabricated and successfully tested. In-band ripples as low as 0.2 dB and extinction ratios exceeding 15 dB have been measured from the fabricated samples. Repeatability of the performances from chip to chip and wafer to wafer is presented to show the tolerance of the devices to fabrication errors. Even though these particular devices have been designed for TE polarization only, polarization insensitive designs can be also achieved. All designs are based on strip waveguides and compact Euler-bends, leading to footprints in the order of 700x300 μm², also thanks to an optimized configuration. They can find applications as interleavers as such or as stages in cascades of N interleavers to achieve flat-top 1x2^N (de)multiplexers.

This paper presents a novel analytical model for the analysis of the electromagnetic field radiation in grating couplers. As will be shown, the radiation pattern of the grating couplers can be described with appropriate accuracy as periodic structures. The obtained field distribution of the coupler can be modeled as a sequence of Fourier series for particular distance values, periodicities and wavelengths. This is compatible with the Floquet-Bloch theory of periodic structures. With this model all relevant parameters for the radiation pattern can be investigated. The results of the proposed analytical model are compared with simulation results for a wavelength of 1550 nm. The model can be used for any periodic structure.

Recently we discovered a signal amplification mechanism to amplify photocurrent with high efficiency and low noise. Unlike conventional impact ionization used in avalanche photodetectors, the new amplification mechanism can produce high (>1000) gain with very low excess noise factor (<2 for Si) under very low bias voltage (3V). The new amplification mechanism offers a promising solution for light detection for Si-photonics, imaging, and sensing. Physics of this mechanism lies in two subsequent processes i) Auger excitation between mobile and highly localized electrons and ii) electron-phonon coupling. In this paper, experimental results are supported by the proposed physical model using simulations within density functional theory (DFT) framework.

This paper proposes a single-photon avalanche diode (SPAD) sensor array comprised of a hybrid structure which can maximize the fill factor of the active area and be compatible with the other detector layer optimized for various demands. In order to implement the hybrid structure, a 100μm pitch through silicon via (TSV) implementation method has been developed to access the back surface of the sensor layer. The achieved fill factor is up to 60%, thus, photon detection efficiency can be reached 35%. A 32×32 SPAD array and a dedicated application specific IC has been designed. We have proved the concept structure can work successfully through the characterization of the hybridized chip. On the other hand, we realized multi-event detection capability should be considered when we apply the photon counting image sensor to a time-of-flight application in high background intensity, and the new concept of a SiPM-based pixel structure has been considered. In order to prove the concept, fundamental experiments have been performed by using the new SiPMs which have extended sensitivity in the near infrared region, and a current mode front-end ROIC which can mark a time-of-arrival and distinguish a photon quantity. A walk error has been studied and found the plot of the time-of-arrival and the photon quantity can be utilized for the measurement compensation.

Photonic Integrated Circuits (PIC) will change the fundamental paradigms for the design of multi-color laser engines for life sciences. Exemplified with flow cytometry (FCM), integrated optical technology for visible wavelengths will be shown to open a new spectrum of possibilities to control flow cell illumination patterns, such as the number of output spots, the spot size, and even complex patterns generated by interferometry. Integration of additional optical functions such as variable optical attenuation, wavelength division multiplexing or fast shutters adds value to the PIC. TOPTICA is demonstrating integration of PICs in present Multi-color Laser Engine (MLE) architectures. Multiple wavelengths (405nm, 488nm, 561nm, 640nm) are coupled free space into the chip, leveraging its beam steering COOLAC (Constant Optical Output Level Auto Calibration) technology for automatic realignment, thus overcoming the need of fiber input delivery. Once in the waveguide, the light can be redirected and shaped to a desired output pattern and pitch, reducing the need of discrete optical components. In this work, we will discuss the implementation of various building blocks in PIC technology for MLEs and analyze the advantages over current macroscopic counterparts.

We report here experimental result of refractive index (RI) sensing using the singly degenerate high quality (Q) factor mode in photonic crystal slab (PCS) with the incident beam close to normal incidence. Q = 3.2×10⁴ was measured with high extinction ratio (ER > 10 dB) for the PCS in air. Q = 1.8×10⁴ and spectral sensitivity (S) of 94.5 nm/RIU were measured in liquid for the PCS. A detection limit (DL) of 10^-5 RIU has been achieved with our device. The high-Q for the singly degenerate mode close to normal incidence with relatively high S is very promising to achieve a lower DL for RI sensing.

In this paper, we propose a new design for Mach-Zehnder interferometer with different cross-sectional area than that for conventional silicon waveguide. The structure has a high sensitivity towards the surrounding media as compared to conventional silicon waveguides. This enabled using our design in different applications including optical sensing and modulation.

The extension of silicon photonics towards the mid infrared (mid-IR) spectral range has recently attracted a lot of attention. The development of photonic devices operating at these wavelengths is crucial for many applications including environmental and chemical sensing, astronomy and medicine. Recent works regarding the development of Ge-rich SiGe waveguides on graded buffer layers will be presented. It will be shown that these waveguides demonstrate low loss and strong mode confinement for a large range of wavelengths and that they have a good potential for being a major building block of mid-infrared photonic integrated circuits.

The GeSn alloy with Sn composition of 11% has been grown using an industry standard reduced-pressure chemical vapor deposition reactor in a single run epitaxy. Low-cost commercially available GeH₄ and SnCl₄ were used as Ge and Sn precursors, respectively. The material characterization showed that the threading dislocations were trapped in the Ge/GeSn interface and do not propagate to the GeSn layer, resulting in high quality material. The temperature-dependent photoluminescence study revealed that the direct bandgap GeSn alloy was achieved, as the emission intensity significantly increased at low temperature. The sample was than fabricated into photoconductive detectors and waveguide lasers. For the photodetector, the spectral response wavelength cutoff at 3.0 μm was observed. The specific detectivity of 3.5×10¹⁰ cm•Hz^1/2W^-1 was achieved, which is close to that of market dominating InGaAs photodetectors that are operating in the same wavelength range; For the waveguide laser, the lasing threshold pumping power density of 86.5 kW/cm² at 10 K and the highest operating temperature of 110 K were obtained. Furthermore, the characteristic temperature was evaluated as 65 K.

Micro-ring modulators for use in high-speed telecommunication transceivers designed for silicon-on-insulator (SOI) for 2 μm wavelength operation are described and simulated with comparison to 1.55 μm. Device simulations show improved DC modulation performance due to the free-carrier effect described in the plasma dispersion relations which is stronger for longer wavelengths. WDM applications are described and simulated. Micro-ring modulator devices were designed and fabricated at A*STAR IME and are pending measurement.

We report our advances in development of subwavelength engineered silicon photonic devices for near- and mid-infrared applications. By periodically patterning Si with a pitch small enough to suppress diffraction, we synthesize an effective photonic medium with refractive index between those of Si and the cladding material. This technique releases new degrees of freedom in engineering of light-matter interaction, chromatic dispersion and light propagation in Si photonic waveguides. We present an overview of our recent results in the realization of novel devices including filters and waveguides for near- and mid-infrared wavelength range.

We report the successful fabrication of low-loss sub-micrometric Silicon-On-Insulator strip waveguides for on-chips links. Several strategies including post-lithography treatment, and post-Silicon smoothening techniques such as thermal oxidation and hydrogen annealing have been investigated to smoothen the waveguide sidewalls, as roughness is the major source of transmission losses. An extremely low silicon line edge roughness of 0.75nm is obtained with the optimized process flow combining resist mask Si patterning and hydrogen annealing at 850°C. As a result, record low optical losses of less than 0.5dB/cm are measured at 1310nm for waveguide dimensions superior to 500nm. They range from 2dB/cm to 0.8dB/cm for 300-400nm wide waveguides. Those results are to our knowledge the best ever published for a 1310nm wavelength.

A crucial component of any large scale manufacturing line is the development of autonomous testing at the wafer scale. This work offers a solution through the fabrication of grating couplers in the silicon-on-insulator platform via ion implantation. The grating is subsequently erased after testing using laser annealing without affecting the optical performance of the photonic circuit. Experimental results show the possibility for the realisation of low loss, compact solutions which may revolutionise photonic wafer-scale testing. The process is CMOS compatible and can be implemented in other platforms to realise more complex systems such as multilayer photonics or programmable optical circuits.

Germanium-On-Insulator based photonics is a promising technological platform. As part of the initial development of a germanium photonic platform, optical losses in passive structures and electrical injection for active components have been studied on Germanium-On-Insulator (GeOI) substrates fabricated using Smart-Cut™ technology [1]. The low threading dislocation density of the germanium (Ge) film is expected to reduce unwanted carrier recombination leading to improved performance of the passive and active Ge components. Fiber-couplers and 500nm square waveguides have been fabricated from 200-mm GeOI substrates allowing optical loss measurements in the TM mode at a wavelength of 2.3µm. Propagation losses were evaluated at 1.5dB/cm. Circular-shaped radius bends and evanescent couplers/splitters have also been simulated, fabricated and characterized to determine the bending losses and the coupling coefficient for ring resonators. Furthermore, innovative bend shapes showed lower bending losses than those with circular-shaped radii, which would allow higher component densities. Finally, the electrical injection is of prime importance in fabricating efficient active components on GeOI. For this, the rectifying or ohmic behavior at the metal-semiconductor contact and its dependence as a function of the Ge doping or metal type has to be precisely known. In particular, we examined the challenging question of the contact resistance between metal and n-type Ge. A study has been conducted using Transfer Length Model (TLM) structures to determine the most suitable metal to contact n-doped and p-doped germanium. [1] Reboud, V. et al. Proc. SPIE 9367 (2015). [2] Kang, J. et al. Opt. Express 24, 11855-11864 (2016).

We report a 200 mm silicon photonic platform integrating a set of devices dedicated on HPC applications. PiN microring modulator layout and process are optimized together. Active tuning through heating section is investigated using either doped silicon or metal resistors. This technology is supported by a dedicated process design kit (PDK) compatible with conventional CMOS EDA tools. The PDK includes optical device models that will be described and compared with experimental results. A focus will be done on the PiN micro-ring modulator models which covering a wide range of geometries. DC mode and RF behaviors are supported.

The realization of efficient III-V lasers directly grown on Si substrates is highly desirable for large-scale and low-cost silicon based optoelectronic integrated circuits. However, it has been hindered by the high threading dislocation (TD) density generated at the interface between III-V compounds and Si substrates. Introducing dislocation filter layers (DFLs) to suppress the TD propagation into the active region becomes a key factor for realising lasers with advanced performance. In this paper, optimization of InGaAs/GaAs DFLs in III-V quantum dot (QD) lasers on Si is demonstrated. Based on these optimized DFLs and other strategies, we have achieved a high performance electrically pumped QD laser on a Si substrate with threshold current density of 62.5 A cm^-2, over 105 mW output power, maximum operation temperature of 120 °C and over 100,158 h of extrapolated lifetime.

The realization of efficient laser sources compatible with the microelectronics industry is currently one of the main challenges for silicon photonics. As Ge is CMOS compatible, the interest of using tensile strain or n-type doping to improve its light emission properties has significantly increased over the last few years. Theoretically, it has been predicted that the Ge bandgap becomes direct at around 4% strain for uniaxial tensile stress or 2% strain for bi-axial tensile stress. Several methods to induce such extreme levels of strain are currently investigated. The highest value of strain has been reached with Ge micro-bridges fabricated from Ge-On-Insulator (GeOI) substrates in a controllable and reproducible way. In this work we have first of all investigated the material properties of 200-mm GeOI wafers. Very high crystallographic quality is demonstrated at the micron-scale using Raman spectroscopy and synchrotron based Laue micro-diffraction performed at BM32-ESRF. We give then optimized designs of micro-bridge by comparing suspended and landed micro-bridges on different materials. We theoretically show that the thermal management is strongly improved in landed micro-bridges. Finally, we have developed specific processing for landing Ge micro-bridges on Si or SiO₂, the photoluminescence measurements performed on landed micro-bridges shows an improvement of the Ge light emission with strain.

Here we describe a uniform diameter, direct bandgap Ge_1-xSn_x alloy nanowires, with a Sn incorporation up to 9%, the fabricated through a conventional catalytic bottom-up growth paradigm employing innovative catalysts and precursors. Optical characterization by means of temperature dependent photoluminescence is used to identify transition point from indirect to direct badgap of GeSn nanowires.

Two ultra-compact silicon bandpass filters are proposed and partly experimentally presented. Both of them have wide bandwidth tunability. Based on the first filter (filter-I), the second filter (filter-II) was designed and has large Free Spectrum Range (FSR). Two filters share the same architecture (matrix architecture), consisting two groups of micro-ring resonator-cascade structures (simply called as micro-ring resonator in this letter). Using this matrix architecture, a wide bandwidth tunability from 75 to 300 GHz can be achieved in filter-I. Based on matrix architecture, double micro-ring resonator (MRR) were adopted and Vernier effect was used in design. It is showed both in simulation and experiment that the FSR of filter-II exceeded 35 nm (around 40 nm in simulation), which is much larger than the FSR of single MRR. Filter-II’s tunability of center wavelength in simulation covers most wavelength from 1530 to 1570 nm. The comparison of bandwidth tunability between filter-I and II reveals that adding paths in matrix architecture may be more effective than adopting high-order micro-ring resonators.

While optical interconnects is expected in the near future to provide the most definitive answer to the current bottleneck in further scale down of the electrical interconnects in VLSI circuits by replacing electrical interconnects altogether, it is currently hindered by the fact that traditional optical interconnect would usually require waveguides that are at least an order of magnitude larger than its electrical interconnect counterpart with a separation distance of few microns to avoid undesirable coupling. Plasmonics offer a solution to the waveguide dimension problem as the guiding mechanism in plasmonic waveguide depends on the coupling between electrons and photons and allow for using waveguides with sub-wavelength dimensions on the expense of greater losses. By using silicon with high concentration of excess carriers as the material of choice, we can acquire plasmonic mode in the near and mid infrared. In this paper we use slot waveguides with both intrinsic silicon with and without high excess carrier’s materials and investigate their transmission effectiveness over 90 degree bends. For silicon with high excess carrier concentration, the modes are plasmonic and allow for excellent performance in transmission through 90 degree bends. This enables dynamic control of routing over 90 degree bends by manipulating the number of free carriers through light excitation. The fact that the slot waveguide is used makes the optical interconnect has dimensions in the same order of magnitude as current electrical interconnects dimension.

We present a diffusion doping based plasma dispersion optical modulator in Silicon-On-Insulator platform. To the best of our knowledge this is the first demonstration of a diffusion-doped (Boron and Phosphorus) modulator in a compact Silicon waveguide. However, it results in a graded, isotropic doping profile where lateral diffusion length of the dopants is a critical parameter. We use a micro-ring resonator and proximity doping with varying offset besides the waveguide to experimentally measure the lateral diffusion length. The lateral diffusion is characterized from the change in the extinction of the ring resonator. Experimental measurement shows a lateral diffusion length of 1600 nm in a 220 nm thick Si device layer, which agrees well with the theoretical calculation. With the lateral dopant diffusion length, we have designed and fabricated a 1 mm long pn MZI modulator. Fabrication was done using a combination of optical and e-beam lithography. The MZI waveguides were defined with 160 nm etch in a 220 nm device layer with a waveguide width of 450 nm. As an initial demonstration, we show plasma dispersion based spectral blue shift of 1.5 nm with a reverse-bias voltage of 5 V.

In this paper a 2D Photonic Crystal array in SOI platform having hexagonal periodicity with a ring defect incorporated along with two bus waveguides is conceptualized and realized for various applications of optical communication, sensing etc. The ring structure filters out a resonant wavelength from the spectrum carried to it through the line defect where the resonated peak is determined by the effective ring radius. The hexagonal architecture enables more coupling length than an ideal ring structure which helps in better intensity accumulation. The resonant peak exhibited at 1554nm in simulation, which is observed in the optical characterization at 1543nm. This is attributed to the fabrication tolerance.

Localized Surface Plasmon Resonance (LSPR) that occur in plasmonic nanoparticles due to interaction with electromagnetic waves at wavelengths larger than the nanoparticles themselves has been exploited in many application like solar cells, cancer treatment and spectroscopy due to the enhanced scattering and absorption cross sections that LSPR provides. Being able to control the resonance peaks of scattering in real time using light can be a valuable tool for sensing-related applications as well especially if it happens in the near and Mid-IR spectrum where most of the biological molecules can be sensed as such spectrum contains strong characteristic vibrational transitions of many important molecules . In this work presented here, we used silicon nanoparticles and increased the concentration of free excess carriers in the nanoparticles by light generation until the free carrier concentration was large enough to cause LSPR similar to what we get with nanoparticles made of Noble metals. The LSPR generated by Si nanoparticles with high concentration of free carriers caused the resonance peak to happen in near and mid IR. Depending on the level of carrier concentration which can be changed dynamically in real time, we can control the scattering resonance peak characteristics and position as shown in our work. Successful fabrication of the Silicon nanosphere is demonstrated as well.