A new method of relaxation oscillation locking is demonstrated. It is based on injection of a pulsed periodical radiation into fiber laser at the wavelength which is not equal to the fiber laser wavelength. It must be in the amplification range of active ions. The new method was experimentally verified by using a Yb-Er fiber laser whose wavelength was 1535 nm. Stable pulse-periodic emission was observed in case of injection of diode laser radiation at the wavelength 1550 nm. The pulse repetition frequency varied from 20 to 100 KHz depending on the pump power.

Stable rational harmonic mode-locked pulses with a repetition rate of 40 GHz were generated from an Er-doped fiber laser with a semiconductor optical amplifier in the cavity. By tuning the current of the semiconductor optical amplifier, significantly improved rational harmonic mode- locked pulses were obtained and the supermode noise in the RF spectrum of mode locked laser was reduced for certain current range of the semiconductor optical amplifier.

We present an experimental study of the Yb-doped double-clad fiber laser operating in the 1.08 micrometers wavelength. The fiber is side-pumped with a high power laser diode using the v-groove technique. Various experimental configurations are performed from the linear cavity to an all-fiber tunable unidirectional ring cavity.

Ytterbium doped double clad fibers (YDCF) are largely used in the telecom field as high power single mode pump lasers. We present a simple model that permits to describe a double clad pumping scheme. The YDCF is modeled by a sequence of N doped sections. Each section is described by an equation and leads to a set of N coupled equations. It permits to work out the mean fraction of excited ions over each section. The determination of the mean fraction of excited ions is sufficient to evaluate the signal, pump and ASE progress along the fiber. We use this model to calculate the evolution of the amplified spontaneous emission along the fiber, and find out the optimum fiber length for laser operation. We compare this result to experimental results obtained with a homemade fiber.

Highly phosphorus doped (7 - 17 mol%) single-mode fibers for the application in Raman laser have been manufactured. It has been established that with increasing the P₂O₅ concentration level, both optical losses and the fiber Raman gain coefficient increase. Using the fiber technology developed, the maximum efficiency of a single-cascaded Raman laser is achieved at a phosphorous pentoxide doping level of 12 - 14 mol% P₂O₅.

Flat gain EDFA are key devices in long haul DWDM systems. A flat gain is needed to ensure an adequate signal to noise ratio evolution between channels during transmission. The dissipative filter introduced between two sections of doped fiber inside the amplifier can have a small residual reflective contribution at a particular wavelength. We have investigated the impact of such a back reflection and shown it is negligible for EDFA operation.

We present the analysis of reshaping capabilities of an all- optical threshold circuit. The device is based on SOA's XGM with two amplifiers in a feedback configuration. Simulations and experimentations of the threshold behavior are proposed and commented.

One of the major problems associated with high-speed optical time division multiplexed soliton transmission systems is the timing jitter error associated with individual soliton pulses. The two physical effects that cause soliton jitters are the amplified spontaneous emission (ASE) noise in in- line optical amplifiers and soliton interaction. While soliton interaction is negligible when the duty cycle of transmitted pulses is small enough, the root-mean-square (rms) soliton jitter due to culminated ASE noise is a fundamental factor that limits the bit-rate distance product in high-speed long-haul data transmission systems. Soliton jitter is detrimental to all-optical time division demultiplexing as it would manifest itself into a relative intensity noise in the data recovery process at the optical receiver. This paper investigates the possibility of minimizing the ASE-induced soliton jitter by changing the spacing between in-line optical amplifiers. It is found that the rms soliton jitter decreases with the amplifier spacing and the minimum soliton jitter can be attained in the case of distributed amplification. These results have significant implications to all-optical time division demultiplexing in the sense of taking into account the amplifier spacing in the process of optimizing system parameters to achieve minimum power penalty of the all-optical demultiplexer. The optical demultiplexer used in the analysis is a non-linear optical loop mirror.

We propose a method based on the Schur algorithm for the identification of an Nth-order recirculating fiber-optic lattice structure. We can obtain the coupling coefficients and delay time of the Nth-order recirculating fiber-optic lattice structure using the Schur algorithm from the input and the reflected data of the overall system measured by oscilloscope. To verify the validity of proposed method, we have composed a second order recirculating fiber-optic lattice structure with the coupling coefficients 0.1 and 0.3687. We have observed that the coupling coefficients computed by the proposed algorithm match very closely to the corresponding true values.

In recent years, distributed feedback (DFB) or distributed- Bragg-reflector (DBR) lasers, with surface gratings situated on ridge waveguides, have been presented. Good properties of the latter, for instance, improved optical coupling and lowered threshold current, have been achieved for this type of laser. Based upon the stopband width measurement, the coupling coefficient of DFB lasers can be estimated with reasonable good accuracy. However, to characterize DBR lasers, many researchers have used the calculated coupling coefficient, since the stopband width from the reflectivity profile of the lasers is more difficult to clarify for this estimation compared to that of DFB lasers. It is especially true when grating losses due to scattering and absorption are large and difficult to measure for the surface-grating devices. In this case, estimation of the coupling coefficient becomes more problematic or even erroneous. This work used a novel method that involves implementing the relationship between the real and effective grating lengths and mode spacing under subthreshold to determine the coupling coefficient and grating losses of DBR lasers. Through a self-consistent approach, for the DBR lasers with a geometry of La equals 420 micrometers and Lg equals 380 micrometers , where La and Lg are the active-section and grating-section lengths, respectively, a coupling coefficient of 50 cm^-1 and grating losses of 38 cm^-1 at the lasing wavelength of 833 nm were estimated with a relatively good uncertainty of 10%.

Maker fringes measurements performed with hemispherical lenses pressed against a planar sample permit the propagation of light at very high internal propagation angle, and therefore the observation of many more fringes. The thickness of the active nonlinear region can be determined with accuracy. Results from optical measurements are compared to what is predicted by a charge migration model and by observed currents to the electrodes during poling. This comparison is relevant to the identification of the origin of second-order nonlinearities in silica glasses.

In an electro-optic poled-silica switch, modulation arises from both the electro-optic effect and electrostriction. To investigate the magnitude of these two contributions, we measured the phase shift induced in a thin slab of unpoled silica subjected to a strong dc plus a small, low-frequency (0 - 19 MHz) ac voltage. The frequency responses for light polarized parallel and perpendicular to the applied field exhibit a constant term due to electronic dc Kerr, and strong peaks due to electrostriction enhanced by mechanical resonances of the slab. A theoretical model is presented that gives good quantitative agreement with these observations. From this comparison we infer the values of the dc Kerr and electrostrictive constants of silica for each polarization. For the perpendicular polarization, electrostriction largely dominates in the frequency range under study. A potential electrostrictive modulator and the impact of electrostriction on the polarization dependence of poled-silica devices are discussed.

Silica glass can be poled either thermally or with UV exposure during application of a strong electric field. Such treatment allows electret formation. So normally isotropic glass can become anisotropic via formation of a frozen-in field. This produces non-zero second-order nonlinearity in glass. After such poling treatment a change in the third- order nonlinearity has been observed. In this paper we examine if modification of the third-order nonlinearity is real or some artifact. To do this the DC third-order nonlinearity was measured before poling, after poling and then after erasure of the second-order nonlinearity. It was found that modification of the third-order nonlinearity remains after erasure of the frozen-in field. The reason for modification of the third-order nonlinearity is still not understood. It may be due to some kind of structural modification of the glass. It is known that impurity ionic species are moved through the glass during poling. This movement of ions under the high field may be sufficient to modify the glass structure. From our results, it is clear that the second-order nonlinearity is predominantly caused by formation of a frozen-in field. The increase of the third-order nonlinearity is independent of existence of a frozen-in field after poling.

A detailed experimental and theoretical analysis of multi- channel XPM-based wavelength conversion in standard communication fibers. Dispersion impact is taken into account and comparison with FWM based conversion technique is also offered.

It is well known that 3 X 3 optical fiber coupler can take two general types of configurations one of them is a triangular structure and other is a flat form. This kind of structure can be fabricated by placing three fibers side by side. The three fibers are fused and pulled from both sides through the translation in opposite directions of two movable platforms, to which the fibers are secured. We present a device has potential for use as a dense wavelength division multiplexing/demultiplexing but which uses only one coupler. Leading back one of the output port to the same input port forms this device. A 3 X 3 optical fiber flat fused coupler with coupling coefficient Kl of about 35 degree and an excess los and loop loss of about 1 dB was fabricated. The video channel of 6 MHz spacing and separation between channels of about 3 MHz was generated by modulating the loop length 34 meters with PZ element. The extinction ration was about 30 dB. Thus the device may have a potential for use as WDM/WDD.

A photosensitive organically modified silica (ormosil) material has been developed for integrated optics applications. In this material, prepared via a simple one- step sol-gel process, the silica backbone is substituted with methacrylate groups to confer photosensitivity, and with phenyl and/or methyl groups to adjust the refractive index from 1.47 to 1.52. Ormosil films are spin-coated onto silica-on-silicon substrates, and waveguides are defined in the films using a 325 nm HeCd laser direct write system. The unexposed regions are then dissolved with isopropanol to leave the desired pattern of ridge waveguides, before baking. The width of the waveguides increases with the UV dose and the concentrations of the methacrylate groups and the photo-initiator. Waveguides with widths >= 5 micrometers have been written with UV doses of around 0.4 J/cm², and baked at up to 200 degree(s)C without cracking. In the 1310 and 1550 nm communications windows, the major sources of optical loss are vibration overtones of OH and CH groups. The intensities of these absorption bands have been measured with infrared spectroscopy and photothermal deflection spectroscopy, and the effects of some strategies to reduce the concentration of CH and/or OH groups, including fluorination and higher baking temperatures, are assessed.

The influence of the fluorine-containing precursor type in the MCVD process on the size of soot particles is investigated by the method of laser ultramicroscope. The results of these measurements have shown that Freon-113 decreases the content of large soot particles, while SiF₄ does not. These results suggest an explanation of the optical loss reduction in highly GeO₂- or P₂O₅- doped single-mode fibers due to co-doping of the fiber core with fluorine.

We show how to design long period fiber gratings (of two structures) modeled as the multi-port lattice using the simulated annealing and steepest descent methods for the equalization of the erbium-doped fiber amplifiers (EDFA). We have obtained design parameters of piecewise-uniform long period fiber gratings which fit the inverted erbium gain spectrum of a commercially available EDFA over the entire 1525 nm - 1570 nm range. We analyzed the sensitivity of designed structures by Monte-Carlo simulations.

We present necessary conditions of detuning factor in the coupled mode equations for the synthesis of optical fiber gratings using the Gel'fand-Levitan-Marchenko inverse scattering method. From the presented conditions in the coupled mode equations for the short and long period fiber gratings, we show that the use of Gel'fand-Levitan-Marchenko inverse scattering method for the long period fiber gratings is very restrictive. We also propose some methods to synthesize short period fiber gratings using the Gel'fand- Levitan-Marchenko inverse scattering method.

An H^(infinity )-optimized equalizer is presented to realize an optimal causal and stable equalizer in optical fiber systems. With an ordinary (non-adaptive) equalizer, uncertainty due to the length of an optical fiber, the number of users and temperature may cause degradation to the system performance. The proposed non-adaptive equalizer still gives satisfactory performance because it uses the H^(infinity )-criterion, i.e., it is designed to cope with the worst-case. Our algorithm is robust while the channel characteristics and the statistics of input and external noise are uncertain with a bound. Moreover, while the conventional H^(infinity ) approach is to minimize the peak error gain over the entire frequency range, we propose a modified H^(infinity ) equalizer to focus on the limited used frequency region by using a weight function.

GLAS is a satellite laser altimeter designed to measure ice- sheet topography and cloud and atmospheric properties. The onboard laser is an Nd:YAG with doubler, emitting 1064 nm and 532 nm light. The reflected light from the Earth and atmosphere is collected by a telescope and analyzed to produce the altimetry results. Fiber optics are used to route timing and calibration signals within the instrument as well as to provide calibrated delays. The fiber optic system includes free space to fiber coupling optics, cables and connectors, diode laser and LED fiber pigtails and a 2 kilometer fiber optic delay line. Since these signals are received and processed as analog values, it is necessary to preserve the magnitude, timing and pulse shape as much as possible in the thermal, mechanical and radiation environment of space. This is in contrast to many onboard fiber optic systems which are used for digital communication and can accept wide variations in these parameters as long as the thresholds for accurate detection of a bit aren't exceeded. The requirements include a stability of up to 1% in signal magnitude and 20 ps timing stability in the timing signals and fiber optic delay line over the whole mission, including all environmental effects. Meeting these requirements with the spacecraft resources available has proven challenging. Fiber optic components have been chosen and tested to accomplish this as best as possible. The system design, components selected and testing performed will be discussed in this paper.

Periodically, commercially available (commercial off the shelf, COTS) optical fiber cable assemblies are characterized for space flight usage under the NASA Electronic Parts and Packaging Program. The purpose of this is to provide a family of optical fiber cable options to a variety of different harsh environments typical to space flight missions. The optical fiber cables under test are evaluated to bring out known failure mechanisms that are expected to occur during a typical mission. The tests used to characterize COTS cables include: vacuum exposure, thermal cycling and radiation exposure. Presented here are the results of the testing conducted at NASA Goddard Space Flight Center on COTS optical fiber cables over this past year. Several optical fiber cables were characterized for their thermal stability both during and after thermal cycling. The results show how much preconditioning is necessary for a variety of available cables to remain thermally stable in a space flight environment. Several optical fibers of dimensions 100/140/172 microns were characterized for their radiation effects at -125 degree(s)C using the dose rate requirements of International Space Station. One optical fiber cable in particular was tested for outgassing to verify whether an acrylate coated fiber could be used in a space flight optical cable configuration.

The STS-99 Shuttle Radar Topography Mission (SRTM) employed radar interferometry to gather high resolution imagery used to generate the most detailed 3D map of the earth's surface ever produced. Such a map has a broad range of both military and commercial uses. This 11-day mission of the Space Shuttle Endeavour took place from February 11 to 22, 2000, and covered 80% of the earth's surface. The SRTM project gathered 12.3 Terabytes of imaging data, which is equivalent to more than 20,418 compact disks, and approximately equal to the entire contents of the Library of Congress.

Traditional fiber optic cable materials affect the well- known optical stability of a bare fiber. An expanded PTFE buffering system is described with effectively decouples an optical fiber from other cable elements. The optical performance of the resulting cable closely approximates that of a bare optical fiber even under thermal or mechanical stresses.

We theoretically model the polarization properties of an optical fiber by the Jones matrix of an elliptical birefringent plate. The properties of this model are investigated and lead to develop experimental methods to extract the parameters of the model for a real fiber. A magneto-optical method that measures the beat length of the fiber is also presented and gives a more complete description of the fiber. Wavelength dependence of the parameters characterizing the fiber is finally experimentally investigated.

We have fabricated fibers with an a few nm thick Cd₃P₂ semiconductor layer at the clear glass core glass cladding boundary. We have measured a gain of approximately 7.1 dB in a 4 mm long piece of this Semiconductor Cylinder Fiber (SCF) at a wavelength of 1550 nm. The fiber section was pumped from the side with a 38 mW laser operating at a wavelength of 980 nm. We have reason to believe that the test wavelength of 1550 nm is near the short wavelength end of about a few hundred nm wide gain curve. The SCFs have applications as broad band Fiber Light Amplifiers.

The major use of tapered fibers is the in-line all-fiber spectral filters. Tapered filters are fabricated from a single piece of single-mode optical fiber that is heated and tapered until a given non-uniform longitudinal profile is obtained. This profile creates modes coupling with only forward propagating cladding modes. It is shown that tapered fibers induce low loss, low dispersion and low polarization dependence. The temperature sensitivity can be controlled by an appropriate packaging technique. Their main applications are the gain flattening filters, ASE noise suppression filters and spectrum correctors.