Additive pulse mode-locking theory is applied to an Er-doped fiber laser and to a Cr⁴⁺:YAG laser. Both are mode- locked with a multiple quantum well saturable absorber. It is found that the theory explains their operation quite well. The Cr⁴⁺:YAG laser was found to lose mode-locking effectiveness due to a band edge of the quantum well being in the operating range of the laser. The equations also help model the Er-doped fiber laser operation, but more work is required to verify its efficacy.

We have demonstrated self-starting passive mode-locking of a chromium-doped:YAG (Cr⁴⁺:YAG) laser using a saturable absorber mirror (SAM) structure. Highly stable femtosecond pulses tunable from 1488 to 1535 nm were generated. Average TEM₀₀ output powers ranged from 40 to 80 mW with a minimum pulse width of 120 fs measured at 1488 nm. The generation of ultrashort pulses in solid-state lasers using the Kerr lens modelocking (KLM) technique has been the center of much attention in recent years. Sub 100 fs pulses have been produced using many different laser systems. However, the KLM process is very sensitive to cavity alignment and is easily perturbed by mechanical vibrations and pump power fluctuations. A more reliable process makes use of a saturable absorber to start and stabilize the soliton formation process. The saturable absorber eliminates the need for critical cavity alignment. The femtosecond pulse train produced by the SAM structure in the Cr⁴⁺:YAG laser system is highly stable over very long periods of time. The wide tunability of the Cr⁴⁺:YAG laser throughout the 1.5 micrometer transmission window of optical fiber makes it an ideal spectroscopic source for the characterization and development of novel materials and devices for ultrafast optical interconnects.

Injection seeding of a passively mode-locked fiber laser by an actively mode-locked fiber laser source is described. The passively mode-locked laser employs a multiple quantum well saturable absorber to establish pulsed operation. Mode-locked synchronized operation was maintained with average injection powers as low as 1.3 mW. Stable synchronized pulses were observed with pulse widths as narrow as 10 ps.

Optical pulse sources with repetition rate approaching terahertz are very important for many photonics applications including ultra-high speed optical communication and generation of sub-mm waves. Both active and passive mode locked fiber lasers are the appropriate choice for this purpose because of the availability of erbium doped fiber amplifier. In general, the mode locking occurs with a repetition rate of nf₀, where n is an integer and f₀ is the longitudinal mode frequency spacing. This is called harmonic mode locking. In the case of rational harmonic mode locking, the repetition rate is (np plus 1) f₀ where p is also another integer. For the case of active mode locking, this is obtained when the modulation frequency to the amplitude or phase modulator used for mode locking is given by (n plus 1/p) f₀. For the case of passive mode-locking, the rational harmonic mode-locking occurs when the saturable absorber in a ring laser is offset by a fraction p/L: from the center where L is the length of the cavity. We have developed a theory of the rational mode locked fiber laser. The results of the theory are compared with experimental results obtained from a 1.5 (mu) fiber laser actively mode-locked with a LiNbO₃ electro-optic phase modulator.

A new and highly promising optical memory technology based on a proprietary metal oxide material has been developed in an on-going Rome Laboratory/Laser Chemical Corporation effort. Metal oxide materials are being evaluated for use as optical disk media and optically addressable electronic memory. Simultaneously presented Nd:YAG laser beams at 1.06 micrometer and doubled YAG at 532 nm are used to write blue bits on the metal oxide material. The 532 nm radiation causes electron transfer from oxygen to metal atoms. This weakens the metal- oxygen bonds, allowing lattice oxygen to be driven out. This additional electron then can move from one metal atom in the written spot to another. This process, called 'intervalence transfer,' involves absorption of visible light, making the material appear a different color then the unwritten material. The color change is permanent until the material is reheated to a temperature greater than 350 degrees Celsius either in an oven for global erase, or by a concentrated infrared laser beam for spot erase. Multiwavelength schemes are incorporated into optically reading the bits. Design, fabrication and evaluation of WO₃ thin films is discussed.

Thulium- (Tm-)doped fluorozirconate (ZBLAN) fiber can be pumped by infrared (IR) light that is resonant with the 4f electronic energy levels. With the long lifetimes of these levels, upconversion processes are possible. Accompanied by the appropriate IR wavelengths and optical feedback, upconversion can result in lasing near 480 nm with powers over 100 mW. We discuss the parameters of our laser experiment and the expected results.

Semiconductor laser diodes may play an important role in synchronous optical networks as sources of optical modelocked pulse trains with robust timing stability. In this paper, it is demonstrated how injection modelocked semiconductor lasers can be used in all-optical clock recovery systems. Experimental measurement of clock recovery dynamics shows that these devices offer robust clock recovery with low injected data power (less than 6 uW), large repetition rate locking bandwidth (2.9 X 10^-3 fractional bandwidth), and small timing jitter (less than 93 fs). Clocking from an injected data stream at 1/50th the clocking repetition rate is demonstrated.

Multiwavelength high speed optical signal sources will play an important role in novel wavelength-division-multiplexed and time-division-multiplexed (WDM-TDM) optical networks. These lasers will serve as sources of ultrashort optical pulses, with low absolute timing jitter and high correlated jitter between each wavelength channel. In this paper, a single- stripe GaAs/AlGaAs semiconductor traveling wave amplifier has been used to generate four modelocked WDM channels simultaneously, each transmitting 18 ps pulses at 2.5 GHz for an aggregate data transmission rate of 10 Gbit/s. The spectral separation of four wavelengths can be continuously tuned from 0.8 nm to 2.1 nm. Also, the four wavelengths can be varied over 18 nm while keeping the wavelength separation constant. We observed that by actively modelocking the laser diode, we can take advantage of the transient unsaturated gain to achieve stable multiwavelength lasing, in contrast to gain competition in cw lasers, which tends to prevent simultaneous laser operation at multiple wavelengths. In addition, enhanced phase correlation between pulse trains at each wavelength is observed owing to the use of single-stripe devices. This technology is critical in novel photonic networks utilizing combined WDM-TDM data formats.

In this paper, experimental results are reported on optoelectronic feedback sustained pulsation in multi-quantum well InGaAsP laser diodes at 1300 nm and AlGaAs injection laser diodes at 780 nm. The feedback intensity plays an important role in feedback sustained pulsation in these two different kinds of laser diodes. When the feedback exceeds a certain level, the frequency of feedback sustained pulsation exhibits continuous, but staircase-like, dependence on the drive current with two characteristic slopes. The values of these slopes depend on the feedback level. Further, at appropriate feedback levels, there exists a bistability between the oscillation frequency and the drive current. The amplitude of feedback sustained pulsation can be modulated for subcarrier multiplexing applications. By further increasing the feedback level, period doubling phenomena was also observed in this dynamical system.

The phase fluctuations in semiconductor lasers have shown to be an important source of sensitivity degradation in coherent lightwave systems. In this paper, we analyze the performance of amplifier based synchronous heterodyne binary ASK systems by taking into account the combined effects of the noise arising due to the optical amplifiers, receiver and the phase fluctuations of the transmitter and local oscillator lasers. Special attention is paid towards the study of the influence of optimum threshold on the bit error rate of binary ASK systems both in absence and presence of optical amplifiers, for different amounts of laser phase noise.

An optically controlled phase array antenna was characterized at Rome Laboratories Newport Test Site. The wideband antenna has a 500 MHz bandwidth and a scan angle of minus 60 to plus 60 degrees. Five bit photonic time delay modules are the main reason for the large bandwidth capability. A conventional phased array antenna using electronic phase shifters would have on the order of a 50 MHz bandwidth. A wide bandwidth antenna such as this could be used to combine radar, communications and electronic warfare antennas into a single phased array antenna.

Several relatively new photonic components have recently come to the forefront in applications related to the photonic processing of broadband microwave signals. Included in this set of photonic elements are the wavelength tunable laser, high dispersion optical fiber, and the Bragg reflection grating. Because of the efficiency and potential cost effectiveness offered by these components they hold the promise of revolutionizing the field of broadband photonic processing. Practical commercial photonic processors are likely to become viable in the near term as a result of these new components. This paper presents a brief overview of the performance characteristics of these components and shows how they can be used for two broadband phased array beamforming applications. A technical overview is presented for each system with experimental data provided to specifically demonstrate the performance of a discrete true time delay beamformer. The second configuration introduced has the capability of providing continuous beamsteering; an essential characteristic for null steering applications. Both systems shown take maximum advantage of component reuse and fully integrate the transmit and receive modes in one efficient hardware compressive topology.

To obtain uniform illumination of photonic reconfigurable antennas, a waveguide grating with a nonuniform grating profile may be used. Theoretical studies using approximate models indicate that the grating profile should have a hyperbolic spatial variation along the length of the coupler. This yields a spatially varying diffraction efficiency that compensates for the loss of light as it is diffracted out of the waveguide. Utilizing a holographic interferometer with a computer controlled shutter in one arm, gratings with appropriate spatial profile variation have been recorded in photoresist and transferred to produce photopolymer waveguide gratings. These planar couplers are integrated with optical fiber bundles for input light delivery. The grating periods are chosen to produce orthogonally propagating output waves. A dielectric mirror arrangement is used to reflect the parasitic diffracted order back onto the antenna element. The best devices obtained to date exhibit output uniformity of plus or minus 6% over a coupler length of 20 mm with total efficiency exceeding 50%.

A version of an image consisting of multiple wavelet scales allows for more flexible feature extraction when compared to the use of one wavelet scale. We proposed an imaging system based on a multiple-input joint-transform correlator, that could be used for multiple wavelet-scale analysis of an input image. Given a single input image and wavelet, for m wavelet scales, m versions of the wavelet and m copies of the input image were generated using conventional optics that are used as inputs to a joint wavelet-transform correlator. The output consisted of 4 m - 1 correlation results, one of which is the desired output. The space-bandwidth product of the system is the same as for a conventional two-input joint-transform correlator.

In this paper, we discuss new optical information processing techniques for encryption, security and anticounterfeiting which have been shown to perform well for a variety of problems and applications.

A Franz-Keldysh effect InGaAsP electroabsorption waveguide device is utilized as the high-frequency, high-linear dynamic range modulator and photodetector. The dual-function modulator/photodetector can be useful in compact transmit/receive front end antenna architectures. Adjusting the electrical bias to the reverse-biased p-i-n diode, either efficient optical modulation or detection is demonstrated. As an electroabsorption modulator, a fiber optic link with a minus 17.4 dB rf loss and a 124 dB-Hz^4/5 sub-octave spurious-free dynamic range is obtained with electrical biases in the 2 to 3 V range. As a waveguide photodetector, a 0.47 A/W fiber coupled responsivity, photocurrents up to 20 mA, and an output second-order intercept of plus 34.5 dBm are achieved at 7 V electrical bias. Supporting measurements on additional test devices show a trend toward larger intercept point with longer device lengths.

We demonstrate all-optical switching in an integrated Y- junction device fabricated using conventional contact photolithography, photo-resist masking and wet chemical etching on a GaAs/AlGaAs multiple quantum well structures. Two different mechanisms that can be used to cause the ultrafast switching are compared. They are the carrier induced nonlinearity and the nonlinearities due to the optical Stark effect. In the first case, the device geometry is exploited by a two-pulse switching process whereby one control pulse turns the device on and a subsequent control pulse turns it off. Here the relatively long relaxation time of the carriers does not impose a limit on the speed of operation and time gating of signal pulses within a 9 ps window was realized in our experiment. In the second case, the nonlinearity is ultrafast and hence switching and recovery takes place during the control pulse evolution. Consecutive switching events spaced 1.7 ps apart has been achieved.

Importance of acousto-optical devices in photonics as a modulator, Q-switch, passive-mode locker is well known. In these devices the acoustic signal or ultrasound is generated in a solid medium using an electronic signal applied to a transducer bonded to the acousto-optic medium. However, it is well-known that ultrasound can also be generated using high power light pulses incident on the medium. This method not only avoids the use of transducer and electronics associated with it but also has the potential of performing acousto-optic interaction in the terahertz regime using femto-sec pulses. Some preliminary experiments have been performed using LiNiO₃ SAW delay line and a beam splitter. Results indicate that once optimized, these devices are useful and efficient.

When two beams that are different colors are propagating in a nonlinear medium the refractive index for one beam can be changed by the intensity of the other beam. This effect falls under the general category known as cross phase modulation, which is a mutual coupling between the two beams. This coupling does not necessarily involve power transfer but can cause a redistribution of energy between the two beams. The effect of this induced change in the index of refraction has been called induced focusing, thermal lensing and induced waveguiding. But it all amounts to the same thing, the index of refraction is changed by a pumping beam causing a probe beam propagates differently. Table 1 is a short list of some of the work previously done with this effect. The remainder of this paper consists of a series of experiments demonstrating the effects of pumping beams upon probe beams in semiconductor and semiconductor doped glasses.

A number of CdSSe-doped fibers have successfully been fabricated and their linear properties characterized. The fiber was fabricated using Schott glass RG630, with a CdS_0.5Se_0.5 composition as the core material, and RG6 as the cladding material. The diameters of the fiber core range between 3 - 8 micrometers, and the outer diameter of the fiber range between 150 - 350 micrometer. The fibers were designed for single-mode operation in the 1300 nm and 1550 nm communication wavelength region. In this paper we report the linear characteristics of the fiber, including its linear absorption coefficient and transmittance.

We report here an experimental result for the non-linear refractive index of a CdSSe-doped fiber measured at 1313 nm. The semiconductor is in the form of nano-spherical particles randomly scattered in the glass host. The linear properties of the fiber have been reported. The approach for the measurement of the nonlinear refractive index follows the Z-scan method developed by Sheik-Bahae et al. The set-up measures the transmittance of the sample as a function of the position of the sample as it traverses through the foal point of the illuminating 1313 nm laser radiation.

We have fabricated and tested fibers having a thin AlCu alloy film strip covering a 15 degree of arc along the core cladding boundary. The fabrication of these fibers had a dual purpose. In order for the metal film to deform smoothly during the fiber pulling process from a coating on the 1 mm diameter core rod in the preform to a coating of the 10 micrometer core in the fiber the viscosity of the metal must be less than the viscosity of the surrounding glass. The viscosity of the AlCu film can be conveniently changed by varying the composition of the alloy. These fibers can, eventually, be used as polarizers. Light polarized parallel to the metal strip is absorbed more than light polarized perpendicular to the strip. We tested both the spectral and polarization transmission properties of these fibers.

The Abel inversion is a tomographic problem which we meet in a wide range of areas from astronomy, plasma and jet research to cell microscopy. The Abel inversion restores a radial distribution of refractive index, absorption index or other values for axial symmetric objects. It uses derivative operation that is why the theoretical solving is incorrect. Complex mathematical methods are usually used in practice. We apply an interferometric technique to provide the derivative procedure. The final calculation is approximately twenty times faster than usual techniques. A noise/signal ratio reduces too. An example of biological application is described.

In some cases, target recognition method must satisfy both the consistent requirements of the foreground and background. In this paper, a recognition scheme based on mathematical morphology is proposed to meet these requirements. An error function derived from morphological transform is used to locate the probable location of target. Distribution function of structure element center is introduced to realized size invariance for the target recognition, and a new dredging operation and training procedure are also adopted to reach this aim. A serial x-ray photograph is used to simulate this recognition scheme, results indicate that the algorithm attains a high performance.