The need for low cost micro optical devices is increasing thru all application fields like tele- and data-communication, industrial automation, displays, automobile, sensor applications etc. Polymer technologies can follow this demand due to the possibility of mass fabrication by replication techniques. Various technologies have been developed in the past to fulfill the demanding requests given by the use of micro structures in optical applications. Part of them are already used in industrial manufacturing. Also demanding products are introduced into the market. In the paper we will give an overview of the relevant techniques and demonstrate their possibilities by a few product examples.

VCSELs (Vertical-Cavity Surface-Emitting Lasers) emit circularly symmetric beams vertical to the substrate; the small footprint of the active area (around 400 um²) enables the simultaneous fabrication of several thousand devices on a single wafer. Micro-optical components can modify the free-space optical properties of VCSELs for applications such as fiber-coupling in transceiver modules, illumination purposes, or beam profiling in sensing applications. However, the alignment of a laser towards a lens, for example, is expensive when performed separately for each device. Here we demonstrate a wafer-scale replication process to realise microlenses directly on top of the undiced VCSEL wafers. The process combines uv-casting and lithography to achieve material-free bonding pads and dicing lines. Several examples of lenses and gratings are given. An organically modified sol-gel material (ORMOCER) has been used as lens material. The micro-optical components on the wafer show good stability while sawing and bonding, where temperatures up to 220°C may occur. We have compared refractive lenses on top of the VCSELs with lenses on glass substrates. The lenses on the glass wafers were illuminated from the back-side by a planar wave. Spot diameters around 1.2 um and focal lengths of 30 um to 100 um were measured depending on the radii of curvature. On the VCSELs the lenses showed a strong influence on the transversal mode behaviour.

In this paper we review the most important fabrication techniques for glass and plastic refractive microlenses and we quantitatively characterize in a systematic way the corresponding state-of-the-art microlenses which we obtained from selected research groups. For all our measurements we rely on three optical instruments: a non-contact optical profiler, a transmission Mach-Zehnder interferometer and a Twyman-Green interferometer. To conclude we survey and discuss the different fabrication techniques by comparing the geometrical and optical characteristics of the microlenses.

Recent results of our studies into optical effects where sub-micron length scales play a pivotal role are presented. We start with a discussion of fine optical features produced by relatively large objects, and then move on to consider the big effects that can be produced by sub-micron structures. Topics covered include fine structure in the optical field of microlenses and gratings, and then further down in length scale from microstructured surfaces to resonant filters, photonic crystal waveguides and metallic nanoparticles. For each step we demonstrate potential applications in which such a length scale can present important advantages, as well as discussing some of the disadvantages and challenges in the design and fabrication of such elements. We particularly highlight the sensitivity of many of the structures to small variations in optical situation (e.g. geometry, orientation, material, polarization) leading significant optical effects for small-scale changes. Methods for the characterization of optical fields produced by objects at these smaller dimensions are also presented.

In this paper, we review several different design techniques for the creation of diffractive optical elements (DOEs). We compare the performance speed of these disparate methods against the efficiency and fidelity of the grating output. In addition, we investigate the mechanisms behind observed deviations of the actual element output from both the desired and simulated DOE outputs. This investigation allows the relative importance of the different fabrication error mechanisms to be assessed and some conclusions regarding modification to the fabrication process to be reached.

Free-space optical interconnects have a wide range of potential applications in the field of telecommunications, computing, and data storage. Of particular interest is their use at the printed circuit board-to-printed circuit board level to either augment or replace a conventional electronic backplane. Although it is possible to design a free-space interconnect capable of operating at data rates equal to, or greater than conventional electronic backplanes, acceptance of free-space technology has been slow due to the perceived difficulties in aligning the system, and in maintaining alignment during the operational cycle of a product. In this paper we describe the implementation of a free space optical interconnect that uses a programmable diffractive element for active beam steering in order to maintain the link between the two printed circuit boards. Moreover, by using a modal wave front sensor, the diffractive element corrects for static aberrations. These static aberrations, such as field curvature and coma, are due to intrinsic imperfections in the optical system related to the lenses used. In addition, aberrations introduced by misalignments can also be compensated for. The diffractive structure is displayed on a ferroelectric liquid crystal spatial light modulator. Simulations and theoretical discussions of the performance of the system are shown and analysed.

Different types of planar waveguide microlenses are fabricated with PLC technologies from a variety of optical materials such as silica, photo-definable epoxy resins, and a number of other optical polymers. Hybrid microlenses are also fabricated in which the base of the lens, with a double concave gap, is formed from silica and the gap is filled with an optical polymer. The optimized lens structures provide the maximum coupling efficiencies between the input and output channels at distances up to 100 mm with a minimum channel pitch of 0.5-0.7 mm. Experimental and theoretical studies provide results on collimation and focusing properties of single and double microlenses made of silica, polymer, and silica/polymer. The evaluation of the temperature and wavelength effects on the collimation characteristics of the lenses demonstrate that the single lenses are more stable and, thus, more suitable for operations under varying conditions. Examples of the planar waveguide microlens applications are presented. In one application the microlens arrays are integrated in fast electrooptic photonic switching modules. In the other application the microlenses are embedded in the backplanes with nonblocking optical interconnects.

A wide variety of integration strategies for micro-optical systems have been employed. Here we review some of these and comment on their relative strengths and weaknesses. In particular we compare approaches that are based on monolithic fabrication with those that make use of discrete components. As applications we consider free-space optical interconnects, telecommunications optical space switches and radiation mode interconnects for optical waveguides.

Downscaling the CMOS technology is expected to encounter a metallic interconnect bottleneck in the near future due to the increasing delays of global on-chip interconnects, problems of signal integrity and timing uncertainty (skew and jitter) as well as power consumption. The possible silicon-compatible monolithic integration of optical on-chip interconnects is described as an alternative solution. It is shown that integrated optics using SOI single-mode waveguides, Si-based modulators, and Ge photodetectors offers a feasible way to distribute global signals such as the global clock across a chip. Taking into account the photodetectors followed by a CMOS-inverter-based transimpedance front-end amplifier with additional gain stages to ensure sufficient voltage swing, optical interconnects characteristics are compared with the performances of future metallic global interconnects recently published in the literature. The main advantages brought by optics include signal propagation with negligible distortion over cm-long distances, reduction of total chip power consumption, reduced delay, skew and jitter if compared with electrical repeated lines, and a lower sensitivity to temperature variations.

Optical switches based on liquid crystal SLM (Spatial Light Modulators) have traditionally been considered unsuitable for packet switching due to slow reconfiguration speed. In this paper we investigate the constraint of reconfiguration time in an optically interconnected packet switch. A system architecture based on the established knockout principle and input/output buffers is simulated with self-similar traffic patterns and packet length statistics obtained from NLANR. Analysis includes packet delay distribution, queue length growth. A physical realisation of the system will use VCSEL arrays, detector arrays and multi-mode ribbon fibre. Data granularity of the system is chosen to match the specification of modern line cards used in routers. It is found that a reconfiguration time in the order of micro seconds is sufficient for an acceptable delay and loss rate. Relationships between required reconfiguration time and system parameters are established.

We present a system for direct parallel optical data communication between integrated circuits on neighboured printed circuit boards based on a monolithic integrated CMOS smart pixel array, fibre arrays, and VCSELs. The advantage of our system versus backplane systems is the direct data transfer through the space avoiding planar and area consuming interconnections. The detector chip allows a data rate of 625 Mbit/s per link and is cycled by an optical clock. A simulation of the chip layout showed 260 % more performance versus electrical off-chip interconnects. In principle an 8'8 data transfer is feasible allowing a data rate of 40 Gbit/s. The detector combines an optical receiver array with a digital processor array which executes image processing algorithms. The optical receiver is formed by a PIN photodiode with a diameter of 40 µm, a transimpedance amplifier (TIA) and a decision-making postamplifier. The measured responsivity of the photodiode without antireflection coating is R=0.382 A/W at an optical wavelength of 670 nm. The TIA consists of a CMOS inverter and a PMOS transistor forming the feedback resistor. Together with the postamplifier, formed by a chain of five CMOS inverters and attaining digital CMOS levels, a data rate of 625 Mbit/s is achieved.

This paper describes a complete technology family for parallel optical interconnect systems. Key features are the two-dimensional on-chip optical access and the development of a complete optical pathway. This covers both chip-to-chip links on a single boards, chip-to-chip links over an optical backpanel, and even system-to-system interconnects. Therefore it is a scalable technology. The design of all parts of the link, and the integration of parallel optical interconnect systems in the design flow of electronic systems is presented in this paper.

The development of optical interconnects in printed circuit boards (PCBs) is driven by the increasing bandwidth requirements in servers, supercomputers and switch routers. At higher data rates, electrical connections exhibit an increase in crosstalk and attenuation; which limits channel density and leads to high power dissipation. Optical interconnects may overcome these drawbacks, although open questions still need to be resolved. We have realized multimode acrylate-polymer-based waveguides on PCBs that have propagation losses below 0.04 dB/cm at a wavelength of 850 nm and 0.12 dB/cm at 980 nm. Transmission measurements at a data rate of 12.5 Gb/s over a 1-m-long waveguide show good eye openings, independent of the incoupling conditions. In the interconnect system, the transmitter and receiver arrays are flip-chip-positioned on the top of the board with turning mirrors to redirect the light. The coupling concept is based on the collimated-beam approach with microlenses in front of the waveguides and the optoelectronic components. As we aim for large two-dimensional waveguide arrays, optical crosstalk is an important parameter to be understood. Accordingly, we have measured optical crosstalk for a linear array of 12 optical channels at a pitch of 250 um. The influence of misalignment at the transmitter and the receiver side on optical crosstalk will be presented as a function of the distance between waveguide and transmitter/receiver.

This paper presents the potential characteristics of germanium photodetectors integrated in silicon-on-insulator optical waveguides for intra-chip optical interconnects. Experimental results on the optical absorption, from 1.1 μm to 1.7 μm of Ge layers epitaxially grown on Si are reported, as well as the measured responsivity of an interdigited MSM Ge photodetector. Light coupling from a rib SOI waveguide to a Ge photodetector is studied for two possible configurations: butt coupling or Ge deposition on top of the waveguide. Comparisons between MSM and PIN Ge detectors are carried out by estimating the dark current, capacitance and time response.

Integration of high-speed parallel optical interconnects into printed wiring boards (PWB) is studied. The aim is a hybrid optical-electrical board including both electrical wiring and embedded polymer waveguides. Robust optical coupling between the waveguide and the emitter/detector should be achieved by the use of automated pick-and-place assembly. Different coupling schemes were analyzed by combining non-sequential ray tracing with Monte-Carlo tolerance simulation of misalignments. A modular demonstrator was designed based on three different kind of optical coupling schemes: butt-coupling and couplings based on microlens arrays and on micro ball lenses. The optical front-ends were implemented with PIN and flip-chip-VCSEL arrays as well as 10-Gb/s/channel electronics onto LTCC-based (low-temperature co-fired ceramic) transmitter and receiver modules, which were surface mounted on high-speed PWBs. An electrical simulation model was developed for the design of a VCSEL-based transmitter circuit. Polymer waveguides were fabricated on separate FR-4 boards to allow characterization of alignment tolerances with different waveguides. Optical and adhesion properties of several potential waveguide materials were characterized. The simulations and experiments suggest that, with optimized optomechanical structures and with low loss waveguides, it is possible to achieve acceptable total path loss and yield with the accuracy of automated assembly.

The development of printed circuit boards (PCB) with integrated layers for optical data transfer was pushed during the last few years. Solutions with optical fibers or planar waveguides fabricated from plastics or glass will soon be available on the market. Nevertheless the low loss coupling of functional optical components as connectors, transmitters and receivers to these new generations of PCBs still is open. The packaging of otical transceivers or connectors actually is based mainly on single device solutions or active coupling concepts. On the other side the connectors of external optical data lines or of daughter cards to the main boards and the coupling of transmitter and receiver modules to optical PCBs do need linear array concepts. And the coupling efficiency should not decrease during reflow process. Actual concepts using mulit-mode connectors or a direct waveguide coupling of receivers suffer under high optical losses. However the use of micro-optical functional elements allows the realization of coupling concepts with teh lowest losses possible. The total losses for optical lines from the transmitter to the waveguide and back to the receiver can be reduced below 4 dB. For cost reduction even symmetric optical set-up can be used. The transmission rate can be as high as 40 Gb/s. With this concept error tolerant systems for the optical interconnection are possible. We report about the modeling, the design and the characterization of micro-optical interconnect modules for high efficient contacts to the optical layer in PCBs. For the assembly of the modules we use the new concept of a desk-top factory with miniaturized tools for handling, assembly, and inspection. This concept increases the flexibility and reduces the manufacturing costs.

We have developed technologies that bring functional redundancy to flip-chip mounted 850nm backside-emitting high-speed VCSEL arrays with 250μm channel pitch. A self-aligned dry-etch process results in vertical mesa sidewalls facilitating extremely narrow gaps as small as 2μm between adjacent mesas, and even smaller gaps seem possible with this technology. Very dense intra-cell arrangements of oxide-confined circular and pie-shaped VCSELs were fabricated with minimal current aperture center-to-center distances of 20μm and 17μm, respectively. The paper also gives a first theoretical approximation of the additional coupling loss introduced by the inevitable radial VCSEL--fiber displacement for a 10μm VCSEL coupling into a standard 50μm fiber channel under varying offset launch conditions. We also present 4x8 and 8x8 regular arrays directly hybridized onto silicon carrier chips by means of an indium solder based flip-chip technology. The elimination of thermal bottlenecks by direct mesa bonding cuts the thermal resistance by half to about 1.3K/mW for 10μm devices as compared to offset-bonded devices.

We report on advances in the fabrication and performance of monolithic 850 nm, linearly polarized vertical-cavity surface-emitting lasers (VCSELs) incorporating a semiconductor surface grating at the outcoupling facet. Depending on the grating parameters, the light is polarized either parallel or perpendicular to the grating grooves. Deep-etched gratings enable complete polarization pinning even in directions that are 45 degrees off the preferred crystal axes. On the other hand, such devices can show strong side-lobes in the far-field which may limit the available output power for some applications. Shallow-etched VCSELs with almost undistorted far-fields deliver output powers as high as 29 mW with about 12 dB orthogonal polarization suppression ratio. A combination of surface relief and grating is used to increase the transverse single-mode output power while maintaining polarization stability.

We present a comprehensive study of coupled-cavity VCSEL structure (or CC-VCSEL), i.e. two optical cavities that share a coupling mirror and that can be independently biased using three electrical contacts. Each cavity contains multi-quantum-wells (MQW) serving as a gain or absorbing medium. We perform a thorough theoretical modeling of the CC-VCSELs taking into account the modification of the MQW optical absorption in a perpendicular external electric field, i.e. the Quantum Confined Stark Effect. Our model considers the situation of single and double wavelength lasing in forward biased cavities. We also consider the case when the light generated in one cavity optically pumps the second one with a strength that depends on the applied reverse bias to the second cavity. On this basis we develop a rate equation model that properly considers the optical field distribution of the two longitudinal modes of the coupled-cavity VCSEL. The number of modes drastically increases when one takes into account the polarization of light and the transverse mode properties. We compare our results with recently obtained experimental results on CC-VCSELs. Finally, we exploit the possibilities to realize high-speed optical switching devices such as optical modulators and field-effect QW lasers based on the QCSE in CC-VCSELs for applications in large capacity optical communications systems.

VCSELs (Vertical Cavity Surface Emitting Lasers) are nowadays more and more exploited in optoelectronic applications, monitoring their lasing power in a compact and low cost manner becomes crucial. To collect and control the output light, an external photodetector associated with an optical microlens array can be used. Integrated solutions based on the use of a bulk or QW photodetection section added in single-or double-cavity structures have also been proposed. Here, we have investigated a simpler solution based on a standard VCSEL array. Light emitted by a VCSEL has been electrically detected by adjacent VCSELs located in the same array, using in plane optical waveguiding of spontaneous emission in the intrinsic central zone of the devices. We show that the detected photocurrent can be related to the power of the emitting VCSEL. Signal intensity has been studied as a function of VCSELs distance. This method could lead to a more efficient way to monitor VCSEL emission.

Vertical-cavity surface-emitting lasers (VCSELs) have now entered the stage of implementation in optical devices and networks. Besides the number of properties superior to EEL, VCSELs have the drawback of less stabilized polarization. As any semiconductor laser, VCSELs are sensitive to Optical Feedback (OF) due to parasitic reflections inevitable in practical applications. Here we experimentally and numerically investigate the spectral and polarization properties of VCSELs subject to polarization insensitive optical feedback from an extremely short external cavity -ESEC (few microns). We observe that the total output power and the wavelength of laser emission are sinusoidally, in-phase modulated with the external cavity (EC) length and with a period equal to half the wavelength of the VCSEL. Moreover the currents at which the switches between the two linearly polarized (LP) modes happen and the hysteresis width are also sinusoidally modulated with the same period. When biasing the VCSEL at the switching current polarization switching between two orthogonal linearly polarized states happens at different EC lengths. This polarization switching happens through a hysteresis when decreasing and increasing the EC length. We have also develop a two modes rate equation model that proves and explains the changes of all VCSEL characteristics observed in the experimental part, showing very good agreement with the experimental results.

We describe the development of long-wavelength InGaAs/GaAs vertical-cavity surface emitting lasers (VCSELs). Using highly strained double-quantum wells (DQWs) in combination with negative gain-cavity detuning we have been able to realise such VCSELs with emission wavelength up to 1300 nm. High-performance device characteristics include mW-range output power, mA-range threshold currents, 10 Gbit/s data transmission and very good temperature stability with continuous-wave operation up to at least 140°C. Singlemode emission is realised using an integrated mode filter consisting of a patterned silicon layer on the out-coupling mirror surface, yielding output power and threshold currents for 1270-nm devices of 1.2 - 0.5 mW and 2.3 - 0.6 mA, respectively, over a temperature interval of 10 - 140°C. Multimode devices have been found to deliver more than 2 mW at 1290 nm. Preliminary lifetime measurements do not reveal any intrinsic reliability problems related to the highly strained quantum wells.

In this paper, a novel two-chip-concept for an electrically pumped and micro-mechanically tunable vertical-cavity surface-emitting laser (VCSEL) operating in the 1.55 μm wavelength range is presented. One chip contains the active region with 5 quantum wells based on the material system AlGaInAs/InP and a buried tunnel junction (BTJ) to provide current confinement and waveguiding. A dielectric mirror forms the back reflector. The second chip consists of a curved mirror membrane that can be displaced by electro-thermal heating. The main advantage of this approach is that both parts can be optimized separately. Packaged laser devices show continuous-wave operation at room temperature with an output power of up to 200 μW and very good side mode suppression in the order of 45 dB. Single-mode operation was observed across a tuning range of more than 30 nm.

We report the experimental observation of transverse optical patterns in optically injected phase-coupled VCSEL arrays. The devices consist of a single VCSEL with a metallic top contact defining a 8x8 pixels array. The different semiconductor/air - metal/air index step determines the periodic variation of the cavity transmission, that enables the selection of a single laser supermode above threshold. The patterns were observed in the near field at driving currents below the laser threshold, while injecting a nearly resonant optical field. At large frequency detuning hexagonal patterns were observed, whereas rolls appeared closer to the cavity resonance. The experimental observations are explained according to a well established model (Spinelli, Tissoni, Brambilla, Prati, Lugiato, Phys.Rev.A 58, (1998) 2542) accounting for diffraction and nonlinear effects in optically injected VCSELs, and modified to allow for the spatially modulated cavity transmission.

A comprehensive numerical model has been applied to simulate single and multimode vertical-cavity surface-emitting lasers (VCSELs) with pseudorandom modulation of the current at a rate of 10 Gb/s. Eye diagrams, probability density functions of the power at the decision time, averaged turn-on delay and timing jitter are analysed for different values of the on and off-state currents. Extensive simulations have been performed to obtain the Bit Error Rate (BER). We find that the BER performance of single-mode VCSELs is better than the one obtained with multimode VCSELs when the off-state current is smaller than the threshold current and when the off-state current is larger than the threshold value, providing that the on-state current is large enough. However BER in single-mode VCSELs is greater than in multimode VCSELs when the off-state current is equal to the threshold current and when the off-state current is larger than the threshold value, if the on-state current is small enough

A demonstrator system for a free space adaptive optical interconnect is reported in which beam position at the receiver plane is maintained using feedback control, given disturbances to the alignment of the optical system. A Liquid Crystal Spatial Light Modulator (SLM) displaying binary phase gratings is used to steer the beam. The design and implementation of the controller is presented, and shown in experimental results to completely correct for step disturbances within one to two sample periods. Extension of the system to tolerate random vibrations is also considered, both in terms of the required SLM technology, and design of the control algorithms. The discussion also highlights some of the trade offs in choosing components when designing for tolerance to different levels of vibration.

Refracting surfaces of spherical and rod microlenses have been made with the melting method. Size of spherical lenses ranged between tens of microns to hundreds of them. Length of rod lenses ranged between hundreds of microns to about some millimeters. Diameter of these last lenses ranged between hundreds of microns to about 1.2 mm. The testing of the refracting surfaces has been done by means of a Scanning Electron Microscope and an Atomic Force Microscope. An USAF test target has been used to find the resolution of the lenses. Theoretical behavior of the lenses has been obtained through ray tracing. Applications of the microlenses in combination with VCSELs and in other field are presented.

In this paper we present the state-of-the-art of Deep Lithography with Protons (DLP), a technology that we have adopted and optimized to rapidly prototype three-dimensional micro-optical components and high-aspect ratio micro-mechanical structures in Poly(Methyl MethAcrylate). In particular we focus on the fabrication of individual plastic refractive microlenses featuring a wide range of numerical apertures, diameters and pitches and their 2D arrays. We give a detailed description of the microlens fabrication technique and the calibration procedure that goes along with it. We highlight the quantitative geometrical and optical characteristics of these DLP microlenses and we demonstrate the reproducibility of their fabrication process. We also illustrate the prototyping flexibility of DLP by making arrays featuring microlenses with different sags, pitches and diameters.

A SiGe/Si Modulation-Doped Multiple Quantum Well modulator, integrated in a rib waveguide is presented. Refractive index variation is achieved by depletion of free carriers initially present in the wells. Numerical simulations are used to design the modulator. An optimal structure is defined: it is formed by 3 SiGe QW, 10 nm-thick, and 4 Si P+ layers, 5 nm-thick. The predicted electrorefractive effect is 1.7.10^-4 under a -6V bias voltage, and the associate absorption variation is 0.7 cm^-1. Intensity modulation is obtained by including this active region inside an interferometer structure. Using resonant cavities of a few hundred of µm long, modulation depth of more than 60% is achieved with losses in the ON state below 12 dB. Mach-Zehnder interferometer of 3.8 mm-long leads to modulation depth of more than 95%, with losses in the ON state lower than 8 dB.

We discuss the possible implementation of a complete set of all-optical logical gates using second harmonic generation and parametric down-conversion. We also experimentally tested an all-optical single-bit half-adder made of a XOR and an AND gate, operating on binary stripes encoded as amplitude modulation of the transverse pattern of laser beam.