This PDF file contains the front matter associated with SPIE Proceedings Volume 7606, including the Title Page, Copyright information, Table of Contents, and the Conference Committee listing.

Integration of III-V materials on silicon wafer for active photonic devices have previously been achieved by growing thick III-V layers on top of silicon or by bonding the III-V stack layers onto a silicon wafer. Another way is the epitaxial lateral overgrowth (ELOG) of a thin III-V material from a seed layer directly on the silicon wafer, which can be used as a platform for the growth of active devices. As a prestudy, we have investigated lateral overgrowth of InP by Hydride Vapor Phase Epitaxy (HVPE) over SiO₂ masks of different thickness on InP substrates from openings in the mask. Openings which varied in direction, width and separation were made with E-beam lithography allowing a good dimension control even for nano-sized openings (down to 100 nm wide). This mimics overgrowth of InP on top of SiO₂/Si waveguides. By optimizing the growth conditions in terms of growth temperature and partial pressure of the source gases with respect to the opening direction, separation and width, we show that a thin (~200 nm) layer of InP with good morphology and crystalline quality can be grown laterally on top of SiO₂. Due to the thin grown InP layer, amplification structures on top of it can be well integrated with the underlying silicon waveguides. The proposed ELOG technology provides a promising integration platform for hybrid InP/silicon active devices.

We present the results of an optical investigation of a series of Er doped silicon nanocrystal (Si-NC) samples which were fabricated via ion implantation into SiO₂ on sapphire substrates, followed by a range of rapid thermal processing. The photoluminescence spectra of the Si-NC emission revealed an increase in luminescence intensity and a red-shift of the peak wavelength as a function of annealing conditions. We attribute the former effect to the reduction of implantation induced defects with increasing annealing temperature/duration. Measurements of the rate of decay of photoluminescence intensity at room temperature show a corresponding increase in the carrier lifetimes which is also an indication of a reduced contribution from non-radiative centers. The red-shift of the peak Si-NC intensity is ascribed to an increasing mean Si-NC size as a function of the annealing conditions. Also presented is an estimation of the relative Er sensitization which reveals that the smallest Si-NC size distribution leads to the greatest sensitization ratio. Further investigation in the form of excitation spectroscopy was used to show that Er ions are sensitized not only by energy transfer from the Si-NCs, but also, crucially, from defect states in the SiO₂.

Single photon Si detectors were fabricated by STMicroelectronics and fully characterized in standard operation conditions and after irradiations. Both single cells and arrays, of dimensions ranging from 5x5 up to 64x64, were electrically tested. The devices operation was studied as a function of the temperature from -25°C to 65°C varying the voltage over breakdown, from 5% up to 20% of the breakdown voltage before and after irradiation using both light ions, 10 MeV B ions to doses in the range 3×10⁷ - 5×10¹⁰ cm^-2, and X-rays irradiations in the range 0.5 - 20 krad(Si). Optical characterization was performed using a laser at 659 nm and opportunely chosen filters to vary the optical power. A strong difference in the radiation damage effect is observed for the two different irradiation sources. Ion irradiation, or better implantation, produces a damage preferentially sitting in the active device region, hence even at the lowest irradiation dose the device functionality is compromised, while at the highest dose the device is completely blind. On the other hand, X-rays produce damage in a low concentration, in fact it does not significantly affect the device dark current, only an increase in the leakage current under breakdown is observed. Hence the device functionality is preserved to doses up to 20 Krad(Si).

While commercial flow cytometers are sophisticated analytical instruments extensively used in research and clinical laboratories, they do not meet the challenging practical requirements for point-of-care (POC) diagnostics. In this paper we will describe and illustrate a new detection technique that will enable a compact, microfluidic-based flow cytometer that satisfies POC specifications for performance, robustness, compactness, cost, reagent consumption, and ease of use. The technology has been demonstrated with CD4 counts in whole blood.

Despite decades of gains in biotechnology, currently much remains unknown about the molecular mechanisms of a number of deadly diseases. New tools are needed to speed discovery at the molecular level and to mitigate the complexity of studying the systems biology of disease. In the last decade, a number of advancements have been made in photonic biosensing aimed at developing a new generation of technologies for the study and early detection of disease. Label-free photonic biosensing techniques such as surface plasmon resonance imaging (SPRI), nanoporous silicon waveguides, interferometry, and optical resonators have achieved detection limits on par with conventional techniques while decreasing the size of the biosensing element and reducing the number of steps involved in the biosensing procedure. In addition, surface enhanced Raman spectroscopy (SERS) has been shown to detect molecules in extremely low numbers, even single molecules. In parallel to the development of a new generation of photonic biosensing techniques, advanced microfluidic technologies have emerged. The last decade has seen the development of automated micro-valves and micro-pumps, particle sorting, concentration gradient generation, and sample mixing. Also, photonic components, including waveguides and lenses, have been fabricated from soft-lithography materials. Synthesizing advanced microfluidic techniques with the new generation of photonic biosensing is likely to be the key to increasing throughput by increasing automation, decreasing equipment costs, and enabling multiplexing. This report will review examples of recent work in optofluidic biosensors and will discuss the opportunities for advancing research related to disease at the molecular level utilizing optofluidic biosensing technologies.

Integrated liquid core AntiResonant Reflecting Optical Waveguide (ARROW) are used as basic component for the realization of complex optofluidic devices. Liquid core ARROW waveguides permit to confine the light in a low refractive index liquid core, by means of two high refractive index cladding layers designed to form a high reflectivity Fabry-Perot antiresonant cavity. This arrangement allows to realize liquid core waveguides that can be very useful in optofluidic applications. We report the fabrication and the characterization different optofluidic devices based on hollow core ARROW waveguide like tuneable couplers and Mach-Zehnder interferometers. The proposed devices have been realized by silicon technology. The channels have been realized by etching the silicon wafer, while the two claddings have been deposited on both wafers by LPCVD or ALD depositions.

Fluorescence resonance energy transfer (FRET) is one of the powerful tools used to study the dynamics of biomolecules. By monitoring the energy transferred from a donor fluorophore to an acceptor fluorophore, one can determine the spatial proximity between the fluorophores on the nanometer scale and thus extract information regarding interactions between biomolecules. Here we demonstrate a novel way of measuring FRET from oligonucleotides using an integrated optofluidic chip containing a planar liquid-core waveguide that can guide liquid and light simultaneously. FRET experiments were carried out using fluorescein and Cy3 labeled oligonucleotides FRET pairs. An excitation laser was fiber-coupled to a solid-core waveguide perpendicular to the chip's liquid-core channel. A FRET efficiency of 50% was measured in good agreement with bulk microscopy experiments. By photobleaching the acceptors and manipulating the fluidic flow, we also demonstrated controllable FRET events: an increase in donor signal, a decrease in acceptor signal and the recovery of FRET due to the influx of new FRET pairs. The flexibility of our chip design also allows for improvements such as separate donor and acceptor detection at either chip end using integrated filters.

We present an SOI biosensor microarray chip that allows multiple molecular binding reactions to be simultaneously monitored. The individual biosensors are formed using 0.26 × 0.45 μm² silicon photonic wire waveguides, which are arranged in compact Mach-Zehnder interferometer geometries with near temperature independent response. The sharp bend radius permitted by the photonic wires is exploited to form dense spiral waveguide structures that provide several millimeters of path length in a compact 130 μm diameter circular area. This design provides the high sensitivity of a long waveguide, while maintaining compatibility with commercial microarray spotting tools. For low volume analyte delivery the sensor array chip contains a monolithically integrated microfluidic channel formed in an SU-8 overlayer. Multiple antibody-antigen reactions are observed in real-time by using an infrared camera to monitor the optical powers emerging from the sensor array output waveguides.

Silicon photonic biosensors using SiN_x/SiO₂ waveguides fabricated in a commercial CMOS process are read out with an integrated polysilicon photodetector array buried under the waveguide. Increased surface refractive index, such as that generated by the specific binding of target molecules, locally decreases evanescent coupling to the detector array resulting in reduced photocurrent from elements under the higher index region. The operating principle of this local evanescent array coupled (LEAC) sensor differs from those of other waveguide sensors and has been explored using numerical simulations. Experimental studies have been performed using adsorbed bovine serum albumin (BSA) nanofilms as well as antigen-antibody pairs.

In this work, we have fabricated a porous silicon (PSi) Bragg reflectors microarray using a proper technological process based on photolithography and electrochemical anodization of silicon. Each element of the array is characterized by a diameter of 200 μm. The PSi structures have been used as platform to immobilize label-free DNA probe and a simple optical method has been employed to investigate the interaction between probe-DNA and its complementary target. In order to confirm the specificity of the DNA hybridization, we have also verified that the reaction of probe-DNA with non-complementary DNA did not occur.

We discuss several applications of both resonant and non-resonant subwavelength gratings (SWGs) for silicon photonics. We present results of evanescent field molecular sensing using the transverse magnetic mode of a 0.22 μm thick silicon slab waveguide with a resonant SWG, which couples a free space laser beam to the silicon waveguide mode. The optical readout of this configuration is almost identical to the established surface plasmon resonance sensing technology. Using calibrated sucrose solutions, we demonstrate a bulk refractive index sensitivity of 111 nm/RIU in good agreement with rigorous coupled wave analysis calculations. The binding of a monolayer of streptavidin protein on the waveguide surface is monitored in real time with a signal-to-noise ratio of ~500. In another application, non-resonant SWGs are used to create effective dielectric media with a refractive index that can be tuned between the values of silicon (3.48) and SU-8 polymer used for the cladding (1.58). For example, we present SWG waveguides with an effective core index of approximately 2.65, which exhibit lower propagation loss than photonic wire waveguides of similar dimensions. We use these SWG waveguides to demonstrate highly efficient fiber-chip couplers.

Integrated Bragg gratings are an interesting candidate for waveguide coupling, telecommunication applications, and for the fabrication of integrated photonic sensors. These devices have a high potential for optical integration and are compatible with CMOS processing techniques if compared to their optical fibre counterpart. In this work we present design, fabrication, and testing of Germanium ion implanted Bragg gratings in silicon on insulator (SOI). A periodic refractive index modulation is produced in a 1μm wide SOI rib waveguide by implanting Germanium ions through an SiO₂ hardmask. The implantation conditions have been analysed by 3D ion implantation modelling and the induced refractive index change has been investigated on implanted samples by Rutherford Backscattering Spectroscopy (RBS) and ellipsometry analysis. An extinction ratio of up to 30dB in transmission, around the 1.55μm wavelength, has been demonstrated for Germanium implanted gratings on SOI waveguides.

Previously we have reported the effects of silicon ion irradiation on free carrier lifetime and propagation loss in silicon rib waveguides, and simulated net Raman gain based on experimental results. We further extend this work by reporting the effects of thermally treating a silicon irradiated sample with a higher dose and energy than previously reported, which produced a poor trade-off between free carrier lifetime and excess optical absorption prior to thermal treatment. Excess losses greater than 80dB/cm were recorded prior to annealing. After thermal treatment, the sample demonstrated characteristics of excess loss and free carrier lifetime recorded previously in much lower energy and dose silicon ion irradiated samples, suggesting that thermally treating samples could enhance the trade-off between free carrier lifetime and excess loss introduced to the rib waveguides. Raman gain simulations based on the new experimental data are reported and show an increase in net gain over previously reported data, suggesting that higher dose, shallow silicon ion implantation is the most efficient way of optimising the trade-off between lifetime reduction and excess optical absorption in silicon rib waveguides, a proposal in our earlier work. The effects of thermally treating low temperature oxide clad waveguides with respect to free carrier lifetime are also reported. Results show that thermally treating a low temperature oxide clad waveguide can vary the intrinsic lifetime. The results of this investigation as well as a discussion into the possible origin of the lifetime change are given.

Cantilever couplers for fiber coupling to silicon photonic integrated circuits are presented. The couplers consist of silicon inverse width tapers embedded within silicon dioxide cantilevers that are deflected out-of plane by residual stress. Coupling efficiencies of 1.6 dB per connection for TE polarization and 2 dB per connection for TM polarization are achieved. Finite difference time domain simulations are conducted to assess coupling loss as a function of cantilever geometry and fiber-to-cantilever misalignment. The cantilever couplers enable high efficiency propagation of light from optical fibers to photonic integrated circuits directly on a chip surface without dicing or cleaving.

The ability to control the porosity and hence the refractive index of porous silicon makes it an interesting material for photonic applications. Layers with refractive indices as low as 1.5 up to that of bulk crystalline silicon can be easily fabricated by varying the electrochemical etching parameters during anodization. This ability to control the refractive index makes it possible to design waveguides that more closely match the properties of silica based optical fiber, thus reducing insertion loss. In this paper we explore the possibility of using a focused laser in order to create waveguiding regions in porous silicon substrates comprising of multiple layers. The direct write process can be used to locally oxidize the porous material forming micron sized channels that can be used for waveguiding. Various designs are simulated using a finite element mode solver in order to optimize the design parameters for single mode waveguiding. Experimental results showing the effect of laser irradiation on multilayered structures are also presented.

We discuss a new effect of optical forces, namely, the self-alignment or instability of two parts of a waveguide broken by an offset and a gap. Using a mode matching technique, we investigate the case where there is no gap and only an offset. It is shown that for a TE incident mode, self-alignment occurs for small values of the waveguide thickness, whereas for large values, the system is unstable. For a TM incident mode, however, the situation is reversed and the two parts tend to self-align for large values of the waveguide thickness. It is also shown that the forces are due to the presence of both radiation modes, generated by the discontinuity, and the guided mode. For a TM incident mode, polarization surface charges are formed, that may cause strong self-alignment forces in both the transverse and longitudinal directions.

In this work, we describe the use of a combination of proton beam irradiation and electrochemical etching to fabricate high index-contrast waveguides directly in silicon without the need for silicon-on-insulator substrate. Various types of waveguides with air or porous silicon cladding have been demonstrated. We show that porous silicon (PS) is a flexible cladding material due to the tunability of its refractive index and thickness. The Si/PS waveguide system also possesses better transmittance in the ranges of 1.2-9 and 23-200 μm, compared to Si/SiO₂ waveguides. This is potentially important for mid and far-IR applications. Since it is compatible with conventional CMOS technology, this process can be used for fabrication of integrated optoelectronics circuits.

We discuss the performance of slow-light enhanced optical switches and modulators fabricated in silicon. The switch is based on photonic crystal waveguides in a directional coupler geometry, and the dispersion of the device is engineered to allow a switching length as short as 5 μm and rerouting of optical signals within 3 ps. The 3 ps switching time is demonstrated using free carriers in the silicon generated by the absorption of a femtosecond pump pulse. The modulator is based on a Mach-Zehnder interferometer configuration, with photonic crystal waveguides in each arm to act as phase-shifters. A flat-band slow-light region has been engineered in the phase-shifters to provide an extinction ratio in excess of 15 dB over the entire 11 nm bandwidth of the modulator device.

Silicon optical modulators have generated an increasing interest in the recent years, as their performances are crucial to achieve high speed optical links. Their evolutions are presented, from the first demonstrations to the latest developments on optical modulators integrated in waveguides. Among possibilities to achieve optical modulation in silicon-based materials, the carrier depletion effect has demonstrated good performances. Recent results on high speed and low loss silicon optical modulator are presented. Free carrier-concentration-variation effect is used in PIN diode. Carriers are depleted when the diode is reverse biased, leading to refractive index variations. This device presents a low capacitance which allows the reduction of RC time constant and power consumption. A Mach-Zehnder interferometer with few mm-long phase shifter is used to convert the effective index variation into intensity modulation. Performance comparison between PN, PIN and PIPIN diodes are discussed and characteristics of PIPIN diodes are presented.

We calculate the π-power and rise time for several 1×1 y-branch MZI SOI thermo-optic switches as a function of device size. These switches consist of waveguide core thickness ranging from 10 to 0.22 μm. Upon scaling the core thickness, the best power and speed performance occurs at 0.7 μm. Further miniaturization results in an increased power and reduced speed, mainly due to the mode expansion from the core into the cladding. We show that varying the MZI arm separation is an effective approach to improve the performance of miniaturized thermo-optic switches.

A simulation technique for modeling optical absorption in Ge/SiGe multiple quantum well (MQW) heterostructures is described, based on a combined 6 × 6 k • p hole wave-function a one-band effective mass electron wavefunction calculation. Using this model, we employ strain engineering to target a specific applications-oriented wavelength, namely 1310 nm, and arrive at a design for a MQW structure to modulate light at this wavelength. The modal confinement in a proposed device is then found using finite-element modeling, and we estimate the performance of a proposed waveguide-integrated electroabsorption modulator.

Although switching techniques based on charge injection in silicon have progressed greatly in recent years, switching energies are still above 10 fJ/bit, which is considered the threshold for practical implementation in on-chip optical interconnects. This is due primarily to silicon's relatively weak electro-optic response, as well as the large physical extent of existing switching geometries, both of which increase the energy required to achieve switching. By using a resonant approach in which the optical mode is spatially tightly confined, however, the volume of active material is decreased, resulting in reduced switching energy. In this paper we report on the use of a thin MOS capacitor to inject charge into a resonator based on a photonic crystal microcavity. By injecting charge only into the volume in which the optical mode is localized, switching energy can be reduced below 1 fJ/bit. The index shift available (Δn ~ 0.001) allows the use of a relatively low-Q resonator (Q ~ 550), enabling high optical bandwidth of 100 Gbps with a device footprint below 25 μm².

Silicon photonics has generated a growing interest with impressive results on active devices like optical modulators and photodetectors in the last few years. In the framework of the European project HELIOS, several research groups and industrial partners work on the main building blocks to make high-speed optical links based on either silicon-based materials or III-V components bonded on silicon. Here, we present an overview of the main achievements on PN and PIPIN optical modulators based on carrier depletion and on germanium and III-V photodetectors integrated with silicon waveguides.

Silicon photonic wire waveguides, featuring very strong optical confinement and compatibility with silicon electronics, provide a compact photonic platform on which passive, dynamic, and active photonic devices can be integrated. We have already developed a low-loss waveguide platform and integrated various photonic devices. For passive devices, we have developed polarization-independent wavelength filters using a monolithically integrated polarization diversity circuit, in which waveguide-based polarization manipulation devices are implemented. The polarization-dependent loss of a ring resonator wavelength filter with polarization diversity is less than 1 dB. For dynamic devices, we have developed compact carrier-injection-type variable optical attenuators (VOAs). The length of the device is less than one millimeter, and the response time is nanosecond order. The device has already been made polarization independent. We have recently monolithically integrated these fast VOAs with low-dark-current germanium photodiodes and achieved synchronized operation of these devices. For nonlinear devices, a free-carrier extraction structure using a PIN junction implemented in the waveguide can increase the efficiency of nonlinear functions. For example, in a wavelength conversion based on the Four-wave-mixing effect, the conversion efficiency can be increased by 6 dB.

Crystalline semiconductor nanomembranes (NMs), which are transferable, stackable, bondable and manufacturable, offer unprecedented opportunities for unique and novel device applications. We report here nanophotonic devices based on stacked semiconductor NMs on Si, glass and flexible PET substrates. Photonic crystal Fano resonance-based surface-normal optical filters and broadband reflectors have been demonstrated with unique angle and polarization properties. Flexible photodetectors and solar cells have also been developed based on the NM stacking processes. Such NM stacking process can lead to a paradigm shift on silicon photonic integration and inorganic flexible photonics.

Nanomembranes (NM) are crystalline semiconductor materials (Si, GaAs, SiGe, etc) that have been released from their substrates and redeposited on foreign, flexible or flat substrates enabling the best features of both materials. Although they are in fact crystalline in nature and possess the electronic/photonic properties of bulk material, they are flexible, deformable, and conformable. An obvious choice is silicon-on-insulator (SOI). SOI provides, beyond its application in the Si industry, the ultimate platform for exploring novel science and technological advancements in this class of nanomaterial. In SOI, a SiO2 layer is interspersed between a thin crystalline top Si layer and the bottom Si wafer; the ability to etch this buried oxide selectively creates the nanomembranes. When released from the oxide, this layer can form extremely flexible strain-engineered thin nanomembranes with thicknesses from several hundred nanometers to less than 10 nm, and in various shapes. Photonic devices originally structured in an SOI substrate can now be transferred and stacked on new substrates, rigid and flexible, to increase optical interconnect densities.

In this paper we present the rationale behind photonic integration of optical buffers for packet switching using optical delay lines. Our work on integrated buffers consisting of Indium-Phosphide-based gate matrix switches and silica-based delay lines is reviewed. Successful implementation of this hybridly integrated chip in a synchronously loaded optical packet buffer is shown. Integration of an optical buffer on the hybrid silicon platform, where active switches can be combined with a low-loss silicon delay line, is discussed and the results for a fully integrated 1.1-ns delay buffer are presented. We report on our current efforts to reduce the silicon waveguide loss to scale the hybrid silicon buffer to be able to store 40-byte packets.

We overview the recent progress on optofluidic ring resonator (OFRR) dye lasers developed in our research group. The fluidics and laser cavity design can be divided into three categories: capillary optofluidic ring resonator (COFRR), integrated cylindrical optofluidic ring resonator (ICOFRR), and coupled optofluidic ring resonator (CpOFRR). The COFRR dye laser is based on a micro-sized glass capillary with a wall thickness of a few micrometers. The capillary circular cross-section forms the ring resonator and supports the whispering gallery modes (WGMs) that interact evanescently with the gain medium in the core. The laser cavity structure is versatile to adapt to the gain medium of any refractive index. Owing to the high Q-factor (>10⁹), the lasing threshold of 25 nJ/mm² is achieved. Besides directly pump the dye molecules, lasing through fluorescence resonance energy transfer (FRET) between the donor and acceptor dye molecules is also studied in COFRR laser. The energy transfer process can be further controlled by designed DNA scaffold labeled with donor/acceptor molecules. The ICOFRR dye laser is based on a cylindrical ring resonator fused onto the inner surface of a thick walled glass capillary. The structure has robust mechanical strength to sustain rapid gain medium circulation. The CpOFRR utilizes a cylindrical ring resonator fused on the inner surface of the COFRR capillary. Since the capillary wall is thin, the individual WGMs of the cylindrical ring resonator and the COFRR couples strongly and forms Vernier effect, which provides a way to generate a single mode dye laser.

Optical properties of circular grating resonators in a silicon-on-insulator system are investigated numerically. These structures are of interest as building-blocks of integrated photonic devices. We use a time-harmonic 3D finite-element solver for studying transmission of waveguide modes through the system. We compare numerical results to experimental findings from the literature.

In this invited paper, we first address the differences between the uses of coherent anti-Stokes Raman scattering (CARS) as light generating mechanism in Raman devices and as optical analysis tool in spectroscopy. Next, with respect to the light generating functionality of CARS, we briefly review the latest progress on CARS-based Raman wavelength converters in silicon-on-insulator (SOI) technology and explain a new approach for boosting the efficiency of these converters. Finally, we show that the CARS process is also able to extract heat from a Raman-active medium and that therefore CARS can be used for reducing the heat dissipation in SOI-based Raman lasers.

For CMOS silicon-based light emitting devices to become practical the external power efficiency must be increased. In this paper a reach-through technique is described whereby the external power efficiency can be increased as a result of three phenomena: i) increase in internal quantum efficiency, ii) increase in light extraction efficiency, and iii) lower operating voltage. The three techniques are discussed and the factor 7 improvement in external power efficiency will be described in terms of the electrical characteristics as well as the external radiation patterns.

Emission levels in the 450-750nm range of about 80-100 fold higher than that emitted by single junction avalanche LEDs, has been obtained. CMOS Si LED p⁺-i-np⁺ structures were modeled in order to investigate the effect of various depletion layer profiles and defect engineering on the photonic transitions in the 1.4 to 2.8 eV, 450-750nnm regime. Modeling and device simulation results showed that by utilizing a short lowly doped layer in between a highly doped p+ layer and n layer can enhance the photonic yields by orders of magnitude through an increase in the dynamic carrier densities in the device and favoring enhanced lateral multiplication processes. The electric field profile should be of the order of 5 x 10⁵ V.cm^-1 and about 0.5 micron long. Injecting of carriers of opposite charge type from an opposing forward bias junction further enhance the photonic yield. These models and interpretations is confirmed by analyses of device designs as realized in 1.2 μm and 0.35 CMOS technology. The device design involved normal CMOS design and processing procedures with no excessive micro-dimensioning. The current devices operated in the 8-10V, 1uA - 2mA regime and yield emission intensities of up to 100 nW.μm^-2. The current emission levels are about three orders higher than the low frequency detectability limit of Si CMOS p-n detectors of corresponding area. The particular design favors higher emission levels towards the 750nm wavelength region. This makes diverse electro-optical applications possible such as optical communication on chip, diverse optical signal processing and wave-guiding. It also enables realization of on chip Micro-Optical-Electro-Mechanical Sensors (MOEMS), which could lead to the development of so-called "smart chips" utilizing standard CMOS integrated circuitry.

Development of a silicon-based on-chip light source could be facilitated by the incorporation of nanocrystalline silicon (nc-Si) into a multislot waveguide structure, using erbium embedded in silicon oxide as a luminescence source. The multislot waveguide confines TM polarized light in the oxide (low-index) layers, thus reducing the loss caused by interaction with free carriers in the nc-Si layers. Here we demonstrate a lateral electrical injection scheme using a p-i-n junction embedded into the multislot, allowing much more efficient charge injection than alternative vertical injection approaches which have been limited by the highly insulating oxide layers. By exploiting the difference in the mode profiles of TE and TM light, we were able to gauge the injection of free carriers as a function of applied voltage, by measuring the polarization-dependent optical loss for light transmitted through the multislot waveguide. Experimental measurements are well-predicted by numerical computations using both FDTD and the transfer matrix method.

Optofluidic devices exploit the characteristics of liquids to achieve a dynamic adaptation of their optical properties. The use of liquids allows for functionalities of optical elements to be created, reconfigured or tuned. We present an overview of our work on fluid-control of optical elements and highlight the benefits of an optofluidic approach, focusing on optofluidic cavities created in silicon photonic crystal (PhC) waveguide platforms. These cavities can be spatially and spectrally reconfigured, thus allowing a dynamic control of their optical characteristics. PhC cavities are major building blocks in many applications, from microlasers and biomedical sensor systems to optical switches and integrated circuits. In this paper, we show that PhC microcavities can be formed by infusing a liquid into a selected section of a uniform PhC waveguide and that the optical properties of these cavities can be tuned and adapted. By taking advantage of the negative thermo-optic coefficient of liquids, we describe a method which renders PhC cavities insensitive to temperature changes in the environment. This is only one example where the fluid-control of optical elements results in a functionality that would be very hard to realize with other methods and techniques.

We have investigated microcavities in Silicon-on-Insolator (SOI) waveguides. The rectangular waveguides with 500 nm width are fabricated in the 220 nm silicon device layer. The microcavities are formed by one-dimensional photonic crystals in Fabry-Perot structure directly written in the waveguides. The SOI photonic structures are produced in a CMOS environment using 248 nm DUV lithography, where the waveguides as well as the photonic crystals are created in the same step using a single mask. In order to achieve a desired spectral shape of the filter function capable for several applications, a number of different cavities were investigated, e.g. single cavities of first and higher order as well as multi-cavity filters. The experimental results are compared with simulations of photonic crystal microcavities in strip waveguides. The spectral transmission function of such filters dependent on the design parameters are calculated by an analysis based on Finite-Difference-Time-Domain (FDTD) method.

Primary challenges in Si science and technology are centered on active device engineering issues with the impact made at their photonic characteristics and applications. Low cost and mature Si technology is driving force toward making optoelectronic devices like LEDs, lasers, modulator, wavelength converter, to name a few. However, to fully exploit tremendous potential of these Si-based devices, which are not as efficient as their III-V-based counterparts, external cooling units are required. In this work, we present a first step towards experimental realizing radiative cooling in Si. The fundamentals behind this approach are in the light down-conversion process through thermal emission that releases the energy from the cooling element in the form of multiple intraband red-shifted photons when it is pumped with interband incoherent light. We demonstrate that a large-area Si wafer kept in an evacuated chamber becomes net cooled by >3.0 K starting from 470 K or demonstrates power conversion efficiency of >100% when pumped with 1.06-μm light. One of other attractive attributes of the device is that it can be easy integrated with other silicon photonic components. We also discuss the way toward further improvement of the performance of these devices.

Silicon nanophotonic circuits can exhibit a very high level of functional integration that can be added directly into IC layers. Such an integration is typically performed by wafer to wafer bonding: the first wafer having the optical and optoelectronic layers, the second wafer having the microelectronic layers. After the substrate removing at the photonic wafer side, the optical structures appear then face down when compared with the initial fabrication. This structure flip may be detrimental to the performance of fibre grating couplers as there may be no more optimized layers below as with gratings on SOI. We present an efficient fibre grating coupler with a bottom mirror that anticipates the structure flip. In this approach the mirror is first coated above the encapsulation layer of the fibre grating coupler in order to appear below it after the optical layers integration onto an IC wafer. This way the thickness between the grating and the bottom mirror, which is a very sensitive parameter, remains under control whatever the wafer to wafer bonding process tolerances. For this design we can compare the efficiency of various bottom mirror structures using single or multilayer silicon/silica stacks. We show then that this wafer-bonding compatible design increases the fibre grating coupler efficiency from 50% to more than 75%.

A novel Si LED with three terminals has been designed and fabricated compatible completely with standard n-well CMOS technology. It is composed by a combination of a forward biased p⁺-n junction controlling by gate voltage and a poly-Si/ultrathin oxide/Si tunneling junction. The experimental results demonstrates that: (1)The optical emitting power of LED increases with both forward p⁺-n junction current increasing and positive gate voltage increasing; (2) The optical emitting power of LED still increases with gate Voltage increasing while p⁺-n junction forward current at zero; (3) The spectra of the optical output power on wavelength λ occurs a peak near 1000nm. The results can be explained from the enhancement on the p-n junction forward current by the gate voltage induced barrier lowering effect and the S(poly-Si)OS tunneling junction theory.

In this paper, we investigate athermal and low propagation loss silicon-on-insulator (SOI) rib waveguides. Propagation losses have been modeled for different dimensions of ridge waveguides achieving good agreement with experimental measurements. At certain waveguide widths, it is possible to obtain low propagation losses for both TE (transverse electric) and TM (transverse magnetic) modes. Racetrack ring resonator structures based on ridge waveguides covered by a polymer have been fabricated, aiming for an athermal design and therefore, a very small temperature dependent wavelength shift. Design guidelines for temperature insensitive and small propagation loss ridge waveguides are presented in this paper together with experimental data.

Novel multimode-interference (MMI)-based crossings integrated with miniaturized tapers are numerically presented for Silicon wire waveguides. These miniature tapers function as field expanders to reduce transition loss between the input/output waveguide and the MMI region and the crosstalk in the crossing section. As a consequence, the lengths of MMI sections reduce to less than twice of the beat length. Using finite different time domain method, we demonstrate that the MMI-based waveguide crossing embedded in the quadratic tapers has the size of 5800x5800 nm², the insertion loss of 0.15 dB and the crosstalk of -42 dB at the wavelength of 1550 nm and broad transmission spectrum ranging from 1500 nm to 1600 nm.

We have systematically studied Bragg Grating Filters in Si photonic-wire waveguides. Highly uniform samples have been fabricated in a specialized foundry service for Si photonics. The reflected wavelength has been controlled over an 180 nm range. By changing the design parameters we controlled the photonic band-gap to be between 6 to 16 nm. Thermal tuning of the reflected wavelength was also measured, resulting in a wavelength shift of more than 3 nm.

In this paper, we propose an all optical logic gate based on a photonic crystals multimode interference (PC-MMI). The all optical logic gate with multifunctional performance has been designed theoretically in two-dimensional photonic crystals (PhCs) structure using multimode interference (MMI) principle. The PhCs consist of periodical air holes in silicon-on-insulator structure. The MMI using line defect waveguide is a 3×3 structure on the PhCs. By switching the optical signal to different input waveguide ports, the device can operate as NAND, NOT, NOR, and OR gates simultaneously or individually. In our optimum design, the width of MMI section is 4.8 μm and the length is 64 μm. The spacing between two line-defect waveguides is 1.2 μm. The total length of the device is 70 μm with an input length of 3.2 μm. It is 1/10 smaller size than conventional MMI optical logic device with nearby 1 mm. Therefore, switching speed is also enhanced by the reduced device size. Our proposed structure would be very useful for construction of optical circuit, optical computer, and future Si-based optical integrated circuits.

We study filling slot waveguide structures with atomic layer deposition. We present our recent progress on the growth of slot waveguide structures with oxide materials and their nanocomposites. The effects of the slot geometry and the optical parameters of the filling material on the mode properties are theoretically studied.

e macro size pyramidal horn probe such as klystron horn antenna has been used to provide the excellent focusing capabilities in microwave region. In the similar way, the pyramidal probe with the micron size mirror (pyramidal horn probe) has been fabricated with a nano-size aperture with diameter ranging from 30 to 330 nm on its apex. Light transmission through the micro-fabricated pyramidal horn probe has been measured to enhance the light transmission due to resonant effects between the cavity mode and the slit modes in the probe, along with directionality of the transmitted beam. The resonant tunneling between two standing waves in the input groove and in the output groove can provide the transmission enhancements. Below ~170 nm, the output power normalized to the input power (ratio) has been increased with decreasing diameter. On the other hand, for the diameter ranging from 330 to 170 nm, the ratio has been decreased with decreasing diameter. The transmission (T) is measured to be inversely proportional to the area (A), and TA value for input wavelength 532 nm is found to be constant, 0.136 for the diameter below 160 nm, and to be 0.053 for diameter greater than 160 nm.

The low cost and mass integration potential of CMOS integrated circuits create an attractive opportunity for investigating CMOS as an optical platform. Although silicon, as an indirect band gap material, is known for inefficient electroluminescence, silicon-based optical transmission is still a much sought after capability. This paper shows the potential of an all silicon transmission system for both clock and data transmission. By utilizing silicon light emitting diodes operating in avalanche, it is shown that a switching speed of above hundred megahertz is possible. The transmitter consists of an array of light sources, with metal light directors for improved external quantum efficiency. The array is pulsed across an optical fibre and received by an avalanche photodiode and amplifier module. Spectral results of the received signal confirm an optical component in excess of 100 MHz, were the off-chip driver circuitry and the photodiode receiver currently limit the bandwidth of the system. As the requirements for wideband data transmission are more stringent than for a narrow band clock signal, the transmission system was tested as a baseband digital communication system, with transmission speeds of up to 176 kbps. We also present eye diagrams of the received signal to prove the success of the transmission system, where transmission speed is limited to detectable optical levels versus allowable in-band noise. A refinement on these principles might lead to CMOS as a contender in high speed clock transmission as well as an alternative to III-V devices for low cost optical transmission systems.

We present a combined analysis using cross-sectional transmission electron microscopy (X-TEM) and Raman spectroscopy to study the early formation dynamics of Si-nanocrystals, formed in SiO₂ thin films after Si⁺ implantation and rapid thermal processing (RTP). We obtained values for the diffusion coefficient of Si in thermally grown SiO₂ and the activation energy to precipitate formation in the first 100 seconds of high temperature annealing. These values indicate that the formation of Si-nanocrystals in implanted oxides proceeds much more efficiently than purely via a self diffusion process. We propose that the nanocrystal formation is assisted by the presence of both oxygen vacancies and SiO molecular species, presumably generated by the ion irradiation. Microscopy images reveal the ensemble nanocrystal population to be most accurately represented by a lognormal distribution function with characteristic values for the mean particle diameter, d and variance, σ. The evolution of the silicon nanocrystals with annealing was also investigated by measuring the Raman scattering signal associated with the TO phonon mode arising from Si-Si bonds in Si-rich oxides grown on transparent (Al₂O₃) substrates. This greatly simplifies the experimental observation of the Raman spectra from Si-nanocrystals as compared to previous studies of nanocrystals in oxide films on silicon substrates.

The application of silicon photonic technologies to optical telecommunications requires the development of near-infrared detectors monolithically integrated to the Si platform. Recently, efforts in this area have focused on developing detectors from pure-Ge grown epitaxially on Si substrates. Much effort has been spent on achieving growth of high quality, relaxed Ge films for device structures, but low temperature growth and processing compatible with complementary-metal-oxide-semiconductor (CMOS) technology has yet to be achieved. In this paper, we report on p-i-n heterostructure photodiodes fabricated from Ge films grown directly on Si substrates using a low temperature chemical vapor deposition (CVD) process. The heterostructures were grown on arsenic-doped (n-type) Si(100) with resistivity 0.003 Ω-cm. A 350nm thick layer of intrinsic Ge was deposited first as the active region, followed by 100nm of boron-doped (p-type) Ge. Ohmic contacts were formed by evaporation of Cr and Au. The diodes were characterized with respect to their dark currents and responsivities in the near-IR. For a 60-μm-diameter device at room temperature, the dark current densities were on the order of 10^-2 A/cm² and 10³ A/cm² at -1V and 1V, respectively, the "turn-on" voltage was found to be 0.26 V, and the ideality factor n was found to be 1.2. The external quantum efficiency of the devices was measured at room temperature over the range 1-1.8 μm. The responsivities at 1.3 and 1.55 μm were found to be 0.26 and 0.11, respectively.