This PDF file contains the front matter associated with SPIE Proceedings Volume 7229, including the Title Page, Copyright information, Table of Contents, and the Conference Committee listing.

In this paper, we describe high temperature operation of high speed 1.1μm-range oxide-confined vertical-cavity surfaceemitting lasers (VCSELs) for optical interconnection applications. For achieving high speed of over 25 Gbit/s under a high temperature, we applied InGaAs/GaAsP strain-compensated multiple quantum wells (SC-MQWs) as the active layer. The developed device showed 25 Gbit/s error-free operation at 100°C. We also examined reliability of the VCSELs via accelerated life tests. The result showed extremely long MTTF lifetime of about 10 thousand hours under an ambient temperature of 150°C and a bias current of about 19 kA/cm², a reliability that either equals or surpasses that of conventional 850-nm VCSELs with 10 Gbit/s. Moreover, we revealed a typical failure mode of the device; the result of analysis indicated that the failure was caused by <110> dark line defects (DLDs) generated in the n-DBR layers under the current aperture area.

We present recent results on high-power, high-efficiency two-dimensional vertical-cavity surface-emitting laser (VCSEL) arrays emitting around 808nm. Selectively oxidized, top-emitting single VCSEL emitters with 49% power conversion efficiency were developed as the basic building block of these arrays. Because of the strong GaAs absorption at the 808nm wavelength, the traditional bottom-emitting, substrate-emission configuration is not possible for large arrays that require efficient heat dissipation. The processing and packaging challenges are discussed. We demonstrate 3mm x 3mm arrays and 5mm x 5mm arrays with the GaAs substrate completely removed and mounted on diamond submounts. These arrays emit more than 50W and 120W, respectively, and exhibit a maximum powerconversion efficiency of 42%.

Directly modulated 850nm oxide VCSEL is a key enabling technology for short reach, high speed data-communication applications. Current commercially available optical transceiver products operate at data rate up to 10Gb/s per channel, for aggregate data rate of 70Gb/s and beyond, in the case of parallel optical data link. High volume, low cost, over temperature optical modulation speed, spectral width, output power, thermal power budget, large signal electrical interaction with the IC driver, and reliability are some of the key requirements driving the 850nm oxide VCSEL development. In this paper, we discuss some of the engineering issues investigated for developing a viable oxide VCSEL product operating at 10Gb/s per channel and higher data rate.

In this paper we describe progress in moving VCSELs toward production-ready status in several applications, among others including substantially higher modulation speeds (14-25 Gbps, or even higher) than in current production. In addition we describe potential VCSEL failure mechanisms not previously published, as well as the limitations of some reliability testing techniques.

In this paper, we will focus on the new design concept of optical sub-assembly solution for single fiber optical HDMI connector. A Tx-BOSA comprises dual 1310nm/1550nm lasers and one 850nm GaAs receiver, a Rx-BOSA comprises one 850nm laser/InGaAs receiver and another InGaAs receiver. We will present the design, OE characterization and performance at 3.125Gbps data transmission per channel.

A future generation of high-performance low-power atomic systems is expected to require VCSEL linewidths below 10 MHz for compatibility with the natural atomic linewidth (5 MHz for cesium) that is realized with atomic beams, trapped atoms, and trapped ions. This paper describes initial efforts at Sandia to reduce VCSEL linewidth by increasing the effective cavity length of an 850-nm monolithic VCSEL. In particular, two aspects of VCSEL design will be discussed: the Q of the VCSEL cavity, and the linewidth enhancement factor of the active region material. We report a factor of two linewidth reduction, from 50 MHz for our standard oxide-aperture VCSEL to 23 MHz for an extended-cavity VCSEL.

In recent years, research into microfluidic devices has attracted much interest in the fields of biology and medicine, since they promise cheap and fast sample analysis with drastically reduced volume requirements. The combination of various analysis steps on one chip forms a small-sized biomedical system, where handling, fixing, and sorting of particles are major components. Here, it was demonstrated that optical manipulation is an efficient tool; in particular it is accurate, contactless, and biocompatible. However, the commonly required extensive optical setup contradicts the concept of a miniaturized system. We present a novel particle manipulation concept based on vertical-cavity surface-emitting lasers (VCSELs) as light sources. The small dimensions and the low power consumption of these devices enable a direct integration with microfluidic systems. The symmetric geometry of VCSELs leads to a high-quality, circular output beam, which we additionally shape by an etched surface relief in the laser output facet and an integrated photoresist microlens. Thus, a weakly focused output beam with a beam waist of some micrometers is generated in the microfluidic channel. With this configuration we were able to demonstrate particle deflection, trapping, and sorting with a solitary VCSEL with output powers of only 5mW. Furthermore, the surface emission of VCSELs allows a comparatively easy fabrication of two-dimensional laser arrays with arbitrary arrangement of pixels. Smart particle sorting and switching schemes can thus be realized. We have fabricated densely packed VCSEL arrays with center-to-center spacings of only 24 μm. Equipped with integrated microlenses, these arrays are integrated with microfluidic chips based on polydimethylsiloxane (PDMS), enabling ultra-compact particle sorting and fractionation.

We describe the hybrid integration of vertical-cavity surface-emitting lasers with a network of microfluidic channels to form a compact microfluidic microsystem. VCSEL dies, created by standard fabrication techniques, are integrated on a silicon substrate which is merged with a micro-fluidic network of PDMS channels to form an opto-fluidic microsystem. The fabrication and integration process of VCSEL dies, silicon host substrate, and microfluidic network are discussed. Absorption measurements of the laser output power using IR absorbing dyes indicate a detection limit of 13 μM of dye concentration. A future integration scheme using monolithically integrated VCSEL / PIN photodetector dies is proposed.

Results on new 850nm and 1310nm VCSEL products under development at JDSU will be presented with emphasis on reliability criteria, advances in performance, and interconnect design. An update will also be provided on JDSU's effort to introduce 10Gpbs LW VCSEL based components and modules into the marketplace.

980 nm VCSELs based on sub-monolayer growth show for 20 Gbit/s large signal modulation clearly open eyes without adjustment of the driving conditions between 25 and 120 °C. To access the limiting mechanism for the modulation bandwidth, a temperature dependent small signal analysis is carried out on these devices. Single mode devices are limited by damping, whereas multimode devices are limited by thermal effects, preventing higher photon densities in the cavity.

A model to simulate Vertical-Cavity Surface-Emitting Laser (VCSEL) operation above threshold is presented. The power - injected current (PI) curves are computed accounting for mode competition arising from spatial hole-burning and temperature profiles. The latter affect many laser parameters, such as the gain spectra and the optical modes, which change their shapes and wavelengths during operation. This comprehensive model aims to describe the details of VCSEL operation above threshold, in a non-circularly symmetric geometries, preserving at the same time computational efficiency. The optical treatment is vectorial, using the in-house developed VELM code. The model is based on a solution of the dynamical equations for field-carrier interactions. Similarly to the more mature vectorial optical treatment, the numerical efficiency is achieved by expanding in proper basis of simple and analytical functions all the involved 3D variables: carrier densities, temperature and optical fields.

Many VCSEL based applications require optical feedback of the emitted light. E.g. light output monitor functions in transceivers are used to compensate for thermally induced power variation, power degradation, or even breakdown of pixels if logic for redundancy is available. In this case integrated photodiodes offer less complex assembly compared to widely used hybrid solutions, e.g. known in LC-TOSA assemblies. Especially for chip-on-board (COB) assembly and array configurations, integrated monitor diodes offer a simple and compact power monitoring possibility. For 850 nm VCSELs the integrated photodiodes can be placed between substrate and bottom-DBR, on top of the top-DBR, or inbetween the layer sequence of one DBR. Integrated intra-cavity photodiodes offer superior characteristics in terms of reduced sensitivity for spontaneously emitted light [1] and thus are very well suited for power monitoring or even endof- life (EOL) detection. We present an advanced device design for an intra-cavity photodiode and according performance data in comparison with competing approaches.

A tapered hollow waveguide multiplexer is proposed to combine the output of a multi-wavelength VCSEL array. The design of the proposed hollow waveguide is presented based on ray optics. We demonstrated the multiplexing of 4- channel output of a VCSEL array for coupling into a multi-mode fiber with the tapered hollow waveguide. The wavelength of each channel was allocated with adjusting driving current. The proposed hollow waveguide multiplexer is useful for realizing compact and low-cost WDM transceivers with a multi-wavelength VCSEL array for high capacity short reach optical networks.

We report the nano-scale patterning of concentric ring-shaped metal corrugations around a sub-wavelength aperture in Ag deposited on top of a vertical-cavity surface-emitting laser. The presence of the rings results in more than a doubling of collected far-field power and a reduction in far-field angular width from 100° to 28°. These nano-aperture lasers thus have the unique property of both a small near-field spot and a relatively low beam divergence. Finite-difference timedomain simulations confirm the experimental results and show that the far-field pattern is highly sensitive to misalignment of the aperture and to the presence of multiple transverse laser modes.

Slowing light has been attracting much interest for optical buffers, optical memories and so on. Slowing light also enables us to reduce the size of various optical devices. We have already proposed and demonstrated slow light devices with a Bragg reflector waveguide. An important issue is the coupling between conventional waveguides and slow light, since their mode fields and propagation constants are highly mismatched. In this paper, we propose a novel coupling method with a micro-corner-mirror coupler for a slow light waveguide. The numerical modeling is presented and prospects for monolithic integration of slow light devices are discussed.

An integrated optoelectronic device, comprising VCSEL and intracavity electro-absorption modulator within the same epitaxial structure, has been previously developed by several research groups. Such a combination device, despite having relatively weak DC modulation, exhibits strong optical feedback, resulting in strong optoelectronic resonance feature in small-signal modulation response characteristic . At large modulation amplitude, device demonstrates pulsed response. Similar to Q-switching operation, energy accumulated in the gain medium over full modulation cycle is released in a single short pulse once cavity Q-factor is increased. As a result, traditional NRZ amplitude modulation becomes ineffective. We are proposing a phase-pulse modulation approach to drive this device, when strong optical feedback is used for obtaining very fast rise and fall times of short pulses. Such transient times can be on the order of few photon lifetimes, e.g. few picoseconds. Gain medium depletion can be avoided by variation of Q-factor both above and below steady-state value and keeping total emitted energy per cycle at a constant level. Data showing modulation properties (pulse energy >100 fJ, FWHM 40 ps non-controlled pulse length at 4 GHz,) and device characteristics, along with numerical analysis of such device for different modulation waveforms is presented.

Vertical Cavity Surface Emitting Lasers (VCSELs) emitting in the Ultra Violet (UV) wavelength range (300 nm) are of great importance for spectroscopy, storage, or medical applications. GaAlN based material is well appropriate to achieve this objective. We have developed a model that takes into account band-gap, absorption, refractive index of GaAlN material as function of its composition and used it to define the rules for the design of high reflectivity GaAlN based Bragg reflectors. On one hand, the index contrast between the two materials of the Bragg mirror has to be as high as possible to have a limited number of periods and high reflectivity. On the other hand, we are limited by two phenomena to reach the lowest lasing wavelength : the absorption of the material and the lattice mismatch between the two material used in the mirror. In other words, starting with AlN/GaN couple of material that offers the highest index contrast, we have to increase the Ga content in AlN layer to have a better lattice matching with GaN and to increase the content of Al in GaN layer to limit the absorption of the material. This paper will discuss the trade - off that has to be considered to get high reflectivity mirror with the lowest number of period for very short wavelength (300nm).

We present here a 1.55 μm single mode Vertical-Cavity Surface-Emitting Laser (VCSEL) based low phasenoise ring optoelectronic (OEO) oscillator operating at 2.49 GHz for aerospace, avionics and embedded systems applications. Experiments using optical fibers of different lengths have been carried out to obtain optimal results. A phase-noise measurement of -107 dBc/Hz at an offset of 10 kHz from the carrier is obtained. A 3-dB linewidth of 16 Hz for this oscillator signal has been measured. An analysis of lateral mode spacing or Free Spectral Range (FSR) as a function of fiber length has been carried out. A parametric comparison with DFB Laser-based and multimode VCSEL-based opto-electronic oscillators is also presented.

Shannon's information theory is applied to Wavelength Modulation Spectroscopy (WMS) providing quantitative figures of merit such as the measurement precision and a prediction of the optimal detection harmonic order to be used. The amount of information, in bits, that can be extracted in any WMS measurement is calculated. The theory is applied to experimental results we have obtained in WMS experiments in congested spectra with overlapping lines that have highly disparate absorption cross-sections. A key result is that the complexity of signal structure can play a much more important role than the conventional signal to noise ratio. We show that there are some parts (where it exhibits turning points and zero crossings) of the structurally-rich WMS signal that play a larger role in conveying information about the measurement than other parts of the signal. Practical applications follow immediately. We also show that, for a particular noise limitation of the apparatus, there is a finite amount of information that can be transmitted (to the detection equipment) by the probe laser as it samples the probed species. The apparatus is analogous to a Shannon's information channel. Application of the theory developed to our experimental absorption measurements in the Oxygen A-band shows why high detection harmonic orders (up to the 7th or 8th) yield the highest resolution. This is in contrast to statements in the literature, based on conventional signal to noise ratio considerations, that the best results are to be expected with second harmonic detection.