This PDF file contains the front matter associated with SPIE Proceedings Volume 7279, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and the Conference Committee listing

Growth of III-nitrides on silicon is investigated to realize optical devices on silicon substrate. The emphasis is focused on the growth of highly qualified GaN micro-crystals on patterned Si substrate so that we may achieve GaN-Si integration. Because of the differences in the crystal structure and the thermal expansion coefficient, the epitaxial layer is subject to defects/cracks, which are overcome by selective area growth (SAG) on patterned Si substrate. By using the anisotropy etching of the Si, we can make {111} facet of Si on the substrate surface. The SAG of the GaN is achieved on the thus prepared {111} facet. As a result, growth of non-polar or semi-polar GaN has been achieved in self-organizing manner. The growth of (1-101), (11-22) and (11-20)GaN are demonstrated on (001), (113) and (110) Si substrate, respectively. Fabrication of an optical waveguide and light emitting diode are demonstrated on Si substrate.

A₂S₈ glass films are shown to undergo changes in the refractive index under ultraviolet (UV) illumination and annealing treatment, respectively. Under UV illumination, the maximum change in the refractive index of the as-deposited A₂S₈ is ~0.056; while under annealing treatment, it lies on and increases with increase of the annealing temperature. In addition, the changes of the refractive index in the well-illuminated A₂S₈ glass films after annealing are not reversible until the annealing temperature is up to 160°C.The A2S8 channel waveguide fabricated by exposure to UV illumination still exists after annealing of 130~160°C, and the refractive index differences between the areas of the channel waveguide and the planar waveguide in this waveguide are 0.0120~0.0081. Light propagation experiment shows that this A₂S₈ channel waveguide has a light guide character with good quality and a low propagation loss of ~4dB/cm.

We propose and demonstrate true demodulation of 40Gbit/s NRZ-DPSK signals using cascaded second-order nonlinearities in a PPLN waveguide. All-optical 40Gbit/s ODB-to-NRZ and AMI-to-RZ format conversions are achieved in the experiment.

Ba₃CaSi_1.7Al_0.3O₈: Eu, Mn white-emitting phosphor was prepared by solid reaction and X-ray powder diffraction (XRD) analysis confirmed the formation of Ba₃CaSi_1.7Al_0.3O₈: 0.1Eu, 0.1Mn. It was found experimentally that, its three-color emission peaks situated at 621, 491 and 435 nm respectively under excitation of 350-450nm irradiation. The 435nm and 491nm emission bands are attributed to 4f to 5d transition of Eu²⁺ ions. The emission band peaking at 621nm originates from the ⁴T₁-⁶A₁ transition of Mn²⁺ ions after the energy transferring from Eu²⁺ to Mn²⁺ ions. When the concentrations of the Al³⁺, Eu²⁺ and Mn²⁺ ions were 0.8 mol, 0.1mol and 0.1mol respectively, the sample presented intense white emitting. The addition of Al ion to the host leads to a substantial change of intensity ratio between blue to green emission.

A novel TE mode transmission waveguide polarizer has been designed based on SiO2/Si waveguide birefringence effect and based on the coupling mode theory. After numerical simulation by 3D FD-BPM method combined with transparent boundary condition, the typical polarizer can achieve a high extinction ratio over 50dB with the device length of 8mm at the wavelength of 1.55 μm. Without increasing the complexity of waveguide manufacturing, this structure can be directly used as a polarizer, and also can be integrated easily with other waveguide devices.

A complex-coupled DFB Laser with the sampled grating has been designed and fabricated. The +1^st order reflection of the sampled grating is utilized for laser single mode operation, which is 1.5387μm in the experiment. The typical threshold current of the device is 30mA, and the optical output power is about 10mW at the injected current of 100mA.

A new fiber coupled AOM for high power Q-switched fiber laser is proposed and experimentally presented. Each end of the AOM has a fiber collimator. For its particular structure, amplified spontaneous emission (ASE) and other deleterious nonlinear processes such as stimulated Raman scattering (SRS) and stimulated Brillouin scattering (SBS) in high power pulsed fiber laser can be easily suppressed. Based on our proposed acoustic-optic modulator (AOM), the 3dB line width of the fiber laser can be narrowed to only 0.1nm, which made it suitable to be used as the seed in master-oscillator power amplifier (MOPA) configuration based all-fiber pulsed fiber laser. When the repetition rate is 20 kHz, 2W output from the fiber laser can be achieved. Full width at half maximum (FWHM) of the fiber laser's pulse is 56ns and the peak power of the fiber laser is nearly 1.8kW. When the fiber laser is used as the seed of a high power fiber laser, 20W with 20 kHz repetition rate pulsed fiber laser has been obtained. The pulse width is nearly 100ns and the peak power can reach as high as 10kW. Based on the proposed technique, mature product of the pulsed fiber laser with this configuration has been successfully implemented.

There are many methods to improve the qualities of free optical communication (FSO) .It is one of the effective ways to improve anti-interference and noise faculty. Disadvantage factors and receiver background light in the laser communication result in interference. Noise is mostly in the receiver. The paper introduces in the FSO system, the laser signal which transmits in the atmosphere random channel is effected by atmosphere absorption, scattering, atmospheric turbulence and interference by the building sloshing; when received, it is effected by the ground light interference. Accordingly, we compare some proposed method with others. Based on the sources of noise, we detail the receiver noise and give some advices on designing photoelectric detective circuit. This study on FSO has a theoretical reference value.

Unlike laser, spontaneous emission from bulk thermal sources is usually perceived to be incoherent and broadband. However, when subwavelength microstructures are fabricated on the surface of bulk thermal sources, the radiative characteristics can be greatly altered. In this paper, a vertical resonant-cavity narrowband thermal emitter, which is based on the theory of surface wave existing in the forbid band of the truncated one-dimensional photonic crystals, is proposed. The central wavelength, bandwidth (FWHM) and emittance at its central wavelength are 2μm, 4.3nm and nearly 100%, respectively. Theory explanation is given, and two types of numerical simulation methods (transfer matrix method (TMM) and finite difference time domain method (FDTD)) are adopted to simulate this structure. The numerical results obtained by these two different methods are consistent; therefore, the feasibility of this devised emitter is confirmed.

Vanadium dioxide as an optoelectronic material exhibits a phase transition between semiconductor and metallic phases at a temperature around 68°C, which has attracted the attention of researchers for many decades. A proper theoretical model was actually presented to analyze the infrared optical properties of vanadium dioxide film by using the theory of multilayer matrix calculation and a mended Sellmeier dispersion model. According to the theoretical model, the transmittances of the vanadium dioxide films were calculated at the temperature from 20°C to 100°C and the wavelength ranged from 400nm to 2500nm. The vanadium dioxide films with different thickness were prepared by magnetron sputtering on glass substrate, and the optical property measurement was carried out in the experiment. The simulated calculation of the theoretical model and experimental curves fitted very well. Study on the vanadium dioxide thin film thickness dependence of the optical properties, the transmittance of vanadium dioxide film reduced with the increasing in the thickness of film, and thermal hysteresis loop became wide and the phase transition temperature rose with the increasing in the thickness of the film. The thickness of the VO₂ film had significant effect on its optical properties.

The impact of Mach-Zehnder modulator (MZM) chirp on carrier suppressed double sideband (CS-DSB) radio over fiber (RoF) link is analyzed and simulated. Finite extinction ratio and drive signal unbalanced ratio of nonideal MZM are considered, power variation of targeted microwave signal and harmonic suppression are calculated and evaluated. The results show that drive signal unbalanced ratio could be optimized to minimize the power variation owing to the finite extinction ratio in the CS-DSB RoF link. However, harmonic suppress ratio declined rapidly with the decrease of extinction ratio, although the unbalanced ratio could also be adjusted to improve the harmonic suppression. For a nonideal modulator with low extinction ratio, unbalanced ratio could be optimized to overcome power penalty caused by finite extinction ration, while an additional filter is required to suppress the unwanted harmonic components in the CS-DSB RoF link.

An improved stable single-longitudinal-mode (SLM) erbium-doped fiber (EDF) laser with multiple-linear short cavity is demonstrated. Three fiber Bragg gratings (FBGs) with the same parameters directly written in a homemade photosensitive EDF (PEDF) in a single step are used as the wavelength-selective and mode-selective component in a 14 cm long linear laser cavity. The optical signal-to-noise ratio (OSNR) is over 50 dB. The amplitude variation in nearly one hour is less than 0.3 dB. The proposed laser has the advantages such as simple fabrication and compact all-optical fiber configuration.

Switchable dual-wavelength with orthogonal polarizations fiber laser based on semiconductor optical amplifier (SOA) and polarization-maintaining fiber Bragg grating (PMFBG) at room temperature is proposed. Owing to the polarization dependent loss of the PMFBG, the laser can be designed to operate in stable dual-wavelength or wavelength-switching modes with a wavelength spacing of 0.336 nm at room temperature by adjusting the polarization controller (PC). The amplitude variation in nearly half an hour is less than 0.1 dB for both wavelengths, which is more stable than that of erbium doped fiber (EDF)-based laser with similar configuration.

We demonstrate a multiwavelength semiconductor optical amplifier (SOA) fiber ring laser with a dual- pass Mach- Zehnder interferometer (MZI) filter. Two SOAs provide enough gain and wider gain spectrum for more wavelength lasing. The dual-pass MZI with a high extinction ratio (ER) serves as comb filter. 82 wavelengths within a power deviation of 5 dB with the signal-to-noise ratio of 30 dB oscillate simultaneously.

We present a multiwavelength mode-locked fiber ring laser incorporating two semiconductor optical amplifiers (SOAs) and a segment of dispersion-compensating fiber (DCF). One SOA is mode-locked by an injected external optical signal and act as loss modulator and assistant gain medium. The other SOA is not modulated by injected signal and only severs as gain medium. Two SOAs synchronously provide enough cavity gain for high repetition rate multiwavelength oscillation. A segment of commercial DCF is employed to introduce intracavity dispersion and functions as a self-tuned comb filter. Our laser source can generate 16 synchronized wavelength channels, each mode-locked at 10 GHz. 16 wavelengths from 1581.96 to 1597.52 nm is with a power deviation of about 4 dB and the mode-locked pulse average width is 39 ps. Oscillation wavelengths can be smoothly tuned to appointed wavelengths and the output power is rather stable.

In this paper, we propose and theoretically demonstrate two-input arbitrary Boolean logic gates using single semiconductor optical amplifier (SOA) and optical filtering at 10Gb/s. The system consists of an SOA and a specialdesigned optical filter. Before the SOA, there are two input data signals and a probe signal. The data signals should be ultrashort pulse sequence. Because different input power can cause different frequency shifts of the probe signal, different logic gates can be obtained by different filter detuning. Assisted by single Gaussian optical filter, we obtained logic NOR, OR, AND, XOR. Assisted by special filter with double peaks, we obtained logic XNOR and NAND. All the logic gates are obtained by changing the consequent filter only. The bit error rate (BER) measurement is analyzed for all logic gates. Our scheme is quite simple and robust for multifunctional logic gates. These logic gates can be used to the advanced complex logic circuits.

A new optical interconnection module for high coupling efficiency on Electro-optical printed circuit board (EOPCB) is presented in this paper. Polymer light waveguide which is embedded in common PCB has been used to construct EOPCB. The waveguide is fabricated by the method called Doctor-blading technology. A new 90°-deflection element which contains a MT-compatible coupling module with sphere interface is used to connect the VCSEL and waveguide. The MT-compatible coupling element is fabricated using high NA fiber which is placed very close to the vertical cavity surface emitter laser (VCSEL), which can make the loss between the VCSEL and coupling element ignored. On the other hand, the sphere interface at the end of the coupling element can provide a higher coupling efficiency and a larger misalignment tolerance between the coupling element and waveguide. The simulation result shows that the coupling efficiency can reach a peak value of 97.1% when the radius of the sphere interface is about 500μm. The horizontal misalignment tolerance and the vertical misalignment tolerance amount to ±25μm and ±32μm for 3dB-loss respectively.

Powder CsI crystal has been deposited with vacuum thermal evaporation on three different kinds of substrates: Si, SiO2/Si and Pt/Si. We have analyzed and observed these CsI films with different depth and various preparation conditions by XRD measurement. Through analyzing, we find that in such process condition the crystal state of CsI film has a strong relationship with the crystal structure of substrate, and non-crystal substrate goes against crystallization. By contrasting standard XRD diagram of CsI(Tl), we discover that with the influence of the surface structure of substrate, CsI crystal film has a preferred orientation in (200) crystal face. We also notice that the preferred orientation of CsI film has a close relation with the depth of the film: the preferred orientation has been weakened as the depth of film turning from 70μm to 100μm.

With density functional theory (DFT), the structures and stability of B_n clusters with n=2-8 have been studied. By using the all electron basis, all the geometries have been globally optimized without any symmetry constraint. It is found that all the small B_n (n=2-8) clusters prefer to form planar structures with sp² bonds, which are in good agreement with others' related studies. B_n and B_n^- are also compared. In contrast with the neutral Bn clusters, although B-B distances in Bn - have slight differences, but addition of one electron does not change their structures significantly. As for energies, all the anions are lying lower than their corresponding neutral clusters. In addition, calculations of energetic and electronic properties for all the neutral clusters have been presented. Both of these two properties show that in B_n (n=2-8), B₃ and B₅ are more stable than others. Vibrational spectra of B_n (n=3-8) clusters have also been discussed. In each spectrum, intensity peaks which are associated with the vibration of boron clusters related to B-B bond stretching can be observed and they are highest. However, among all the B_n clusters, such peaks of B₃ and B₅ show lower intensity than others. This results suggest that B₃ and B₅ are relatively more stable, which further demonstrates the conclusion above.

An all-optical filter structure for interference suppression of microwave signals is presented. The filter is based on a recirculating delay line (RDL) loop consisting of a semiconductor optical amplifier (SOA) followed by a tunable narrowband optical filter, and a fiber Bragg grating connected after the RDL loop. Negative tap is generated in wavelength conversion process based on cross-gain modulation of amplified spontaneous emission spectrum of the SOA. A narrow passband filter with negative coefficients and a broadband all-pass filter are synthesized to achieve a narrow notch filter with flat passband which can excise interference with minimal impact on the wanted signal over a wide microwave range. Experimental results show that measured and theoretical frequency responses agree well and the filter is tunable.

In this paper, a novel approach for loadable and erasable optical memory unit based on dual microring optical integrators coupled via 3×3 couplers is proposed and studied. The optical integrator, which can generate an optical step function for data storing, is synthesized using active media for loss compensation and a tunable phase shifter for data reading at any time. The input data into the memory is return-to-zero (RZ) signal, and the output data read out from the memory is also RZ format with a narrower pulse width, which is determined by the phase shift pulse. A great number of such memory units is promised to be densely integrated on a photonic circuit for future large scale data storage and buffer.

The electrically pumped InGaAsP/InP active microring has been theoretically analyzed and numerically simulated based on subsection model, with carrier rate equation and amplified spontaneous emission taken into account. After the subsection model is introduced in detail, the spectrum characteristics are numerically investigated. The simulation results show that spectrum characteristics will shift under different pump current and injected light power. The performances of an optical on/off switching utilizing pump-probe method based on InGaAsP/InP active microring are investigated. The results show that if we choose probe light off resonation initially, and pump light on resonation, only several hundred of microwatt power of pump light can realize the probe light on/off conversion. Such a potential will be applied in many other optical applications with ultra low power consumption. However, carrier lifetime will be the main factor restricting the characteristics of electrically pumped active microring due to the fact that the refractive index change is induced by carrier density change.

Enhancement of light extraction in a GaN light-emitting diode (LED) employing a binary blazed grating reflector (BBGR) is presented. The BBGR consists of asymmetrically periodic structure etched on the GaN layer. Rigorous coupled-wave analysis (RCWA) is adopted to calculate the reflectivity of the BBGR, which shows that it has the characteristics of broadband reflection spectrum and very high reflectivity. The result of high angle-averaged reflectivity up to 94% from 300nm to 450nm predicts the potential enhancement of light-extraction efficiency of GaN LEDs with BBGR.

We proposed a novel extended semiconductor optical amplifier (SOA) model for applications in femtosecond pulse transmission, in which the various ultra-fast nonlinear effects, such as gain dispersion, group velocity dispersion (GVD) were taken into account. We use the extended SOA model to analyze the refractive index in response to variation of the input current and the length of SOA. In addition, the impact of carrier density pulsation and carrier heating on the refractive index dynamics of SOA was discussed respectively. The refraction index dynamics that occur in bulk InGaAsP SOA in response to femtosecond pulse (Full width at half maximum is 200fs) are studied numerically. We use the alpha factors (α-factors) to calculate the phase shift from the gain variations, which are related to the refractive index changes. We have considered the effects of carrier heating (CH), and spectrum hole burning (SHB) to analyze the impact of intra-band carrier dynamics on nonlinear refraction index. We also consider the impact of the length of SOA and operation condition on the refraction dynamics.

Through the detail analysis of inter-satellites links in modern small satellite constellations, a new optical filters-based four-channel wavelength division multiplexer was designed and fabricated. This WDM device working at 0.85, 1.06, 1.31 and 1.55μm is designed to build an optical inter-satellites link, which provide two data streams at 0.85 and 1.55μm, an inter-satellite tracking channel at 1.06μm, and an different orbit communication channel at 1.31μm. Through experiments, we got the channel download efficiencies at 0.85, 1.06, 1.31 and 1.55μm are 49.7%, 42.3%, 33.4% and 25.1% and the channel interval is 5mm. The features of small size and light weight make this WDM device quite attractive for space borne applications and the big channel interval could also accommodate the wavelength shifts due to Doppler Effect, temperature variations and radiation effects in space.

The spatially modulated polarization interference imaging spectroscopy (SMPIIS) is based on the polarization interferometer.The core of the polarization interferometer is the Savart polariscope. When light passing through polarizer,the natural light becomes linearly polarized in the direction of the polarization orientation of a linear polarizer. The linearly polarized ray is split into two polarized components by a Savart polariscope. After going through the analyzer,the two rays get linearly polarized in the polarized orientation of the analyzer and are focused by the imaging lens to form the interferogram and target image. From analyzing the mechanism of formation the interferogram and target image,we find the discrete between ordinary ray (o-ray) and extraordinary ray (e-ray) greatly influence the intensities and the results of the interferogram and target image. The wave vector of e-ray is obtained by the extension of Snell's law in the Savart polariscope. According to the relation between the wave vector of e-ray and e-ray,the direction of e-ray is solved. After the directions of o-ray and e-ray are given, the formula of the discrete between o-ray and e-ray can be got by calculating.

PbO-B₂O₃ glasses were prepared by traditional melting method. XRD and optical spectra measurements were done to check properties. Second-harmonic generation was observed in the glasses induced by thermal treatment and thermally poling process, respectively. But second-harmonic generation intensity was normally larger for the later, and it increased with increasing PbO content, poling voltage, temperature and time. The mechanism of second-harmonic generation in the PbO-B₂O₃ glasses was also discussed. A promising material for all-optical switching devices is expected.

In order to enlarge the dynamic range of image intensifier, and CCD can work normally under the strong light (105lx), based on the generation II image intensifier, we have brought forward an novel equipment of double lens optical system including a liquid crystal panel (twisted nematic liquid crystal panel LCX023-CMT) and a photometric CCD which detects the illuminance real-time. The light intensity is controlled by liquid crystal automatically and the partial gating can be realized. The principle of auto-controlling light intensity is introduced detailedly in this paper. And the transmittance of liquid crystal has been simulated. Experimental result displays that the optical transmittance changed from 1.28% to 25.60% with the gray levels from 0 to 250. The high linearity relation of transmittance to gray levels is obtained when the gray levels change from 90 to 220. The feasibility of partial light intensity auto-controlled by liquid crystal has been further proved by the two obvious contrast pictures that one is photographed with the liquid crystal panel and another without.

We report growth of InGaN/GaN multi-quantum-wells (MQWs) structures and GaN layers on silicon-on-insulator (SOI) substrates by metalorganic chemical vapor deposition (MOCVD). The growth conditions were tuned to realize blue-green emission peaks centered around 420-495 nm from such MQWs on SOI. X-ray diffraction, atomic force microscopy, scanning electron microscopy, and photoluminescence techniques were used to characterize the MQWs. Using a combination of selective dry etching techniques, GaN micromechanical structures are demonstrated on silicon-on-insulator (SOI) substrates. The dry releasing technique employs a controlled gas phase pulse etching with non-plasma xenon difluoride (XeF2), which selectively etches the Si overlayer of SOI, thus undercutting the GaN material on top. The mechanical properties of these released microstructures are characterized by micro-Raman spectroscopy. Such approach to realize multi-color light-emitting InGaN/GaN MQW structures and GaN micromechanical structures on SOI substrates is suitable for the integration of InGaN/GaN-based optoelectronic structures on SOI-based micro-opto-electromechanical systems (MOEMS) and sensors.

It is well known that the performance of modern photodiodes is determined by three basic parameters: sensitivity, response time and noise equivalent power (NEP). In practical situations it is almost always necessary to achieve the maximum product of sensitivity and reciprocal response time (bandwidth). At low Continuous Wave (CW) light incidence, a photodiode gives a signal proportional to optical intensity. But with increasing light power, photocurrent deviates from a linear behavior. High-linear performance is becoming increasingly important for photodiodes because high photocurrents directly translate into an increased dynamic range and reduced noise figure. This is crucial for many photonic systems, including photonic analog-to-digital converters and high-bit-rate digital receivers. The increase of photocurrent depends on two primary factors: one is space-charge limitations, which are influenced by device physical dimensions, structure type, illumination conditions, maximum electric field, and the other is thermal considerations. From the view of long-term reliability of photodiodes, thermal effects are crucial because it results in device failure due to dark current runaway. As known to us, there have many reports on the thermal effect originating from InGaAs intrinsic region. However, nonlinearity originating from both InGaAs intrinsic region and contact resistance is still unclear. In some cases, the contact resistance between metal electrode and semiconductor is not negligible. In this report, N type heavily doped InP single crystal is used as a substrate, on which a buffer layer of n doped InP is grown. Then the intrinsic absorbing layer of InGaAs, and finally a transparent InP cap layer are deposited with MOCVD technique. A circuit model has been developed and the nolinearities of PIN photodiodes has also been discussed.

The time delay of the light reflected from and transmitted through the uniform fiber Bragg grating (FBG) are computed and shown to be equal. The physical meaning of the time delay is discussed based on tunneling effect and energy storage function of the FBG. The energy density of the optical field distributed along the FBG and the total energy stored in the FBG are calculated based on the coupled mode theory (CMT). The time needed for light to form the stable energy density distribution is computed and shown to be equal to the time delay of the FBG.

Using first-principles calculations, we investigated the structure and electronic properties of Ag-related defects in ZnO. The calculation results indicate that AgZn behaves as acceptor. Simultaneously, by comparing the formation energy and electronic structure of Ag-related defects in ZnO, Oi-AgZn behaves as acceptor in Ag-doped ZnO and it is better to gain p-type ZnO. However, Hi-AgZn complex has the lowest formation energy. Thus, the formation of the other point defects is greatly suppressed by the formation of Hi in Ag-doped ZnO. Moreover, the H atoms can be easily dissociated from hydrogen-passivated complexes by post-annealing at moderate temperatures, thus, codoping Ag with H may be a good method to achieve p-type in Ag-doped ZnO.

Two fiber Bragg gratings with ~35mm spatial interval are written in the self-made photosensitive Erbium-doped fiber (PEDF) to form a fiber Bragg grating Fabry-Perot (FBG F-P) cavity. A distributed Bragg reflector (DBR) fiber laser is built up based on the FBG F-P cavity. 1.8mW laser output is achieved with 63mW pump power. By calculating the mode-spacing using the effective length of the FBG F-P cavity, the output of the DBR fiber laser is proved to be singlelongitudinal- mode (SLM).

Based on the electro-optical properties of liquid crystal, we have designed a novel partial gating detector. Liquid crystal can be taken to change its own transmission according to the light intensity outside. Every single pixel of the image is real-time modulated by liquid crystal, thus the strong light is weakened and low light goes through the detector normally .The purpose of partial-gating strong light (>10⁵lx) can be achieved by this detector. The modulation transfer function (MTF) equations of the main optical sub-systems are calculated in this paper, they are liquid crystal panels, linear fiber panel and CCD array detector. According to the relevant size, the MTF value of this system is fitted out. The result is MTF= 0.518 at Nyquist frequency.

A liquid crystal panel module coupled in the image intensifier which has been coupled the magnetic mirror has been proposed. The normal imaging under strong light and the part gating of light intensity control can be achieved by changing the transmittance of these pixel points which is changed by the voltage imposed on pixel points. Influence of the response time and transmittance of the liquid crystal to the entire system have been researched, that liquid crystal can achieve part gating function has been confirmed through experiments. Liquid crystal display module (LCP)'s response time is 22ms by using the 650nm, 4mw SZ-04 laser and the transmittance of LCP ranges from 1.28 to 25.60%.It has been proved that the liquid crystal selected in this article fit for the entire LCP. Photography effect graphs achieved by part gating of light intensity have been got.

In order to enhance the dynamic range of digital imaging system, we have brought forward processing image by the Cyclone II chip and controlling HTPS's high-density liquid crystal unit. The system in this paper proposed a kind of low cost solution, the luminous intensity entering each pixel of CCD can be adjusted directly accurately. Meanwhile, the attenuation luminance coefficients set by system algorithm and precision parameters can be adjusted artificially on the physical level to satisfy different needs. It is proving that the function of partial gating image is realized by the image processing experiment in which the details of the final image are more obvious and clear than the original. Then the system latitude problem is solved fundamentally. The image of partial gating is obtained through the experiment, and then the anticipated goal has achieved successfully.

In this paper, we theoretically and numerically demonstrate a two-dimensional Metal-Dielectric-Metal (MDM) waveguide based on finite-difference time-domain simulation of the propagation characteristics of surface plasmon polaritons (SPPs). For practical applications, we propose a plasmonic Y-branch waveguide based on MDM structure for high integration. The simulation results show that the Y-branch waveguide proposed here makes optical splitter with large branching angle (~180 degree) come true. We also introduce a finite array of periodic tooth structure on one surface of the MDM waveguide which is in a similar way as FBGs or Bragg reflectors, potentially as filters for WDM applications. Our results show that the novel structure not only can realize filtering function of wavelength with a high transmittance over 92%, but also with an ultra-compact size in the length of a few hundred nanometers, in comparison with other grating-like SPPs filters. The MDM waveguide splitters and filters could be utilized to achieve ultra-compact photonic filtering devices for high integration in SPPs-based flat metallic surfaces.

This paper firstly presents the fundamental principles of the microwave photonic filters. According to the number of the taps of the microwave signal is finite either infinite, a microwave photonic filter is classified into finite impulse response (FIR) filter and infinite impulse response (IIR) filter. As an example to explain how to implement a microwave photonic filter, a specific FIR configuration is illustrated. Next, the theory and methods around how to gain high Q value is discussed. Then, divided into FIR filter, one-order IIR filter, and multi-order IIR filter, several novel approaches to gain high Q value are discussed and compared. The technical difficulties in multi-order IIR filter are analyzed concretely in this part. In the end, we make a summary for the microwave photonic filters with high Q value.

The significant effect factor of slow light transmission in photonic crystal coupled resonator optical waveguide has been investigated. The change tendency of guided mode band, group velocity and group delay with the tuning of distance between neighboring cavities have been discussed in detail. The guided mode is flattened clearly and the group velocity decreases sharply with the increasing of cavities distance. When the interval between contiguous cavities is 7 lattice constants of photonic crystal, the ultra small group velocity below 2×10^-4c has been obtained. This is roughly two orders of magnitude smaller than that in line defect PC waveguide. The maximal group delay through CROW of 13.23μm with supercell 7×7 exceeds 400ps, which are four orders of magnitude larger than that in traditional waveguide with same length.

Large positive and negative Goos-Hänchen(GH) shifts on an symmetrical metal-cladding waveguide (SMCW) were investigated both theoretically and experimentally. The SMCW is simply a glass slab with several hundred micrometers sandwiched between two gold films. A recently developed free-space coupling technique was used to transfer the light energy into the ultrahigh order modes of the SMCW. It is shown that during the resonant excitation, and the intrinsic damping coincides with the radiative damping of the modes, the lateral shifts of reflected beam are enhanced to as large as hundreds of micrometers. It is also known that sign of the GH shift is totally dependent on the difference between the radiative damping and the intrinsic damping. GH shifts with sub-millimeter scale were experimentally observed. The experimental results show good agreement with theoretical prediction.

In this paper we get the rate equation of photogenerated carriers by solving the rate equation of carriers in a certain condition. By simulation we draw the general rule of variation of photogenerated carriers' density which depends on trigger light pulses as well as their width. Also by experimental tests we achieve the effects of photogenerated carriers on the response speed of Semi-insulating GaAs Photoconductive Switches (PCSS's) and analyze the reasons for the changes of the rising edge of output electrical pulses.

Thin film growth on plane structure surface was discussed by mostly articles, but thin film growth on the microstructure surface like sub-wave grating was little reported. The growth of 100 nm Au film on the surface of sub-wave grating with sine groove shape and 80 nm groove depth was investigated in this article. It was shown by experiments that the thin film growth was the profiling grow, the grating sine grooved characteristic and periodic were almost unchanged, but after coating, some phenomena like the increasing of sine duty factor, the worse uniformity of grating groove depth and the decreasing of groove depth were shown by experiments. The theory of thin film growth shadow effect which is brought by the grating microstructure was proposed to carry on the theoretical analysis in this article. The judging method of film dispersion pattern was also brought up, which was based on the thickness of thin film and the groove depth of grating before coating. When the thickness of thin film was lower than the groove depth, film dispersion pattern was defined as "A" model. The opposite was defined as "B" model.

Within the framework of light-semiconductor interaction and the Lutting-Kohn model for the semiconductor band structure, we theoretically studied the mechanism of the polarization switch with coherent transient features. The result shows, in the coherent limitation, when the pump pulse E_p and signal pulse E_s are orthogonal polarized, the third order polarization P⁽³⁾, with first order of E_s and second order of |E_p|, approach to zero, due to the destructive quantum interference, which results in an ultrafast pump-signal relative polarization dependent P⁽³⁾, therewith a polarization switch. The study also shows that, when the light interacts with semiconductors coherently, the polarization state of the light is transferred to the relative phase of the electron spin, rather than the electron spin itself. The population of the polarization eigenstate excited by the pump pulse, do not origin from the Coulomb force of electrons, and should not thought phenomenologically as the exciton characterized by light polarization. This population will dephase to real carriers population, i.e. electron with exact spin |↑> or |↓>, by losing phase coefficients of the polarization eigenstates, in an ultrashort quantum dephase time.

A novel multi-electrode SOA is proposed to use as all-optical wavelength converter based on XGM, dynamic frequency chirp characteristics are analyzed with a multi-section model, and the peak to peak chirp is simulated as a function of total injection current, signal and probe light power and bit rate. The numerical results show that the carrier non-uniform distribution in MSOAs can be controlled by selecting different segment length and current, thus the frequency chirp can be flexibly manipulated. Therefore, Multi-electrode is an effective and promising technology to improve SOA's dynamic performance in optical communication.

A theoretical study is presented to solve the semiconductor optical amplifier (SOA) polarization dependent problem. We take InAs quantum dot (QD) coupled with tensile-strained InyGa1-yAsP quantum well (QW) structure as SOA active region. A theoretical model which describes the carrier transport process in QD, QW and barrier is established and the carrier rate equations and optical propagation equations are achieved based on the above model. The performance of QD coupled QW SOA is simulated. Optical gain spectrums of transverse electronic (TE) mode and transverse magnetic (TM) mode are presented and the polarization power ratio between TE mode and TM mode as the function of injection current and input optical signal wavelength is also achieved. By optimal designing of the structural parameters of SOA and operation conditions, we can get polarization independent QD coupled QW SOA in certain ranges of input optical wavelength.

In this work, pulsed atomic-layer epitaxy (PALE) technique has been used to grow a-plane GaN films on r-plane sapphire substrates. During growth, the supply of N atoms was alternatively turned on and off while Ga atoms were continuously supplied with a constant flow rate. By optimizing the on/off periods of N source at 20/10 secs, pit-free GaN films have been obtained with significantly reduced full width at half maximum of X-ray rocking curve and brighter characteristic emission of stacking faults, as compared to the case using conventional metal-organic chemicalvapor deposition. The improved epitaxy quality and optical properties in PALE is due to the enhanced the Ga-adatom surface migration and improved the lateral-growth rate, which results in larger grain size and longer stacking faults.

Carbon-based micro/nano materials and structures such as carbon nanofibers (CNFs) and carbon nanotubes (CNTs) have attracted a wide spread attention and been widely studied since it was invented because of its excellent electrical properties, remarkable mechanical strength, thermal conductivity, chemical inertness and special optical properties. However, it still remains a crucial problem in applying and characterizing of carbon-based micro/nano optoelectronic devices, since it is difficult to realize interconnection of these micro/nanostructures such as CNTs for device integration. Recently, it was reported that CNFs and CNTs can be fabricated and integrated with 3-dimensional carbon electrodes through modified photolithography and pyrolyzed process. Following this work, a novel approach is introduced to fabricate suspended micro/nano wires between electrodes which can be a promising method as interconnects for CNTbased devices. This paper briefly presents the process, experimental results and discussions for fabrication of suspended 3-dimensional carbon interconnects. By dispersion of special nanoparticles with thick film SU-8 photoresist, suspended carbon interconnects can be obtained between 3-dimensional adjacent carbon posts during photolithography process with masking effect. Then, further work can be continued with fabrication of CNTs or CNFs over the structures through controlled pyrolysis process.

The effective reflectivity of planar tunable filter fiber Fabry-Perot (TF-FFP) caused by coupling losses would led to the change of tuning range and the full width half maximum (FWHM) in the process of gap distance tuning. The general transfer function expression of FFP is developed by Gaussian beams model and coupling losses. A method that is by appropriate unsymmetrical reflectivity of fiber end can increase the tuning range and reduce the FWHM is proposed, the obtained parameters may be useful for the analysis and design of other TF-FFP. At last the initial gap position and reflectivities of both ends of TF-FFP are designed theoretically for C-band and C+L-band respectively.

In this work, we have investigated the effect of Al composition in AlGaN low-temperature buffer layer (BL) on the crystal quality of a-pane GaN thin films grown on r-plane sapphire substrates. GaN films grown using AlGaN BL with 15% Al exhibit smoothest surface morphology, lowest value of full width at half maximum of X-ray rocking curve and least impurity incorporation, as compared to those films grown using GaN and AlN BLs. AlN BLs result in the worst crystal quality of GaN films due to the large lattice match between GaN and AlN BL, while GaN BLs lead to intermediary quality of GaN films due to large lattice match between GaN BL and r-sapphire substrate. Adding 15% AlN in to GaN BLs can significantly reduce the lattice mismatch between BL and r-sapphire substrate, while still keep the lattice mismatch between BL and GaN films relatively small, and thus optimizes the crystal quality of a-GaN films grown on r-sapphire substrates.

A Free Space Optical Communication (FSO) system transmits modulated light through atmospheric media. Because of the uneven distribution of refractive index result from atmospheric turbulence, the phase distribution of light is changed leading to distortion of wavefront and requiring reconstruction at the receiver. However, current wavefront compensation relies on channel modeling which has difficulties in extracting channel information from highly random turbulent atmosphere. In this paper, a wavefront reconstruction system based on neural network algorithm is constructed. The neural network requires little channel information but predicts distortion by past experience. Then, distorted phase distribution is adaptively revised when light passes through a piezoelectric ceramic deformable mirror controlled by neural network. Dynamic study factors are added to neural network algorithm as improvement which adjusts the study speed of the system according to turbulence intensity providing best result between respond time and reconstruction accuracy. In addition, light transmitted in atmospheric channel is studied.

We numerically investigate a transparent wavelength conversion for a high-speed orthogonal frequency-shift keying and intensity-modulated (FSK/IM) modulation format based on four-wave-mixing (FWM) in semiconductor optical amplifier (SOA). The inter-band and intra-band carrier dynamics, such as gain saturation, carrier density pulsation (CDP), carrier heating (CH) and spectral hole burning (SHB) have all been taken into account in our in-house SOA model. Simulation results show that about 3.5-dB and 1.5-dB power penalty at BER of 1e-9 can be achieved for the 40-Gb/s FSK payload and 2.5-Gb/s IM label after conversion. With fixed signal wavelength, the 1-dB sensitivity penalty range of the pump wavelength for the payload and label can be observed at the value of 4-nm and 5.6-nm respectively. Additionally, the dynamic range of IM modulation depth from 0.3 to 0.8 can also be observed after conversion.

A monolithic white light-emitting diode (LED) with blue and yellow light active regions has been designed and studied. With the Al_xGa_1-xN / In_yGa_1-yN distributed Bragg reflector (DBR) resonant-cavity, the extraction efficiency and power of the yellow light are enhanced so that high quality white light can be obtained.

Inclusions and micro-defects revealed by etch pits are the main defects of CdZnTe crystal, which can not be eliminated thoroughly at present. These defects will affect the photoelectric characteristic of CdZnTe and the quality of the MCT epilayer used for the infrared detectors. Many technologies are employed to suppress or eliminate the defects of CdZnTe materials, and one of them is ampoule coating technology. In the paper, ordinary carbon coating, improved carbon coating and BN (Boron Nitride) coating technology were used to study their effects on defects in CdZnTe crystal. After carbon coating technology was improved, the size of inclusion in the crystal had an obvious decrease from more than 30 μm to less than 15 μm, the density of inclusion reduced from 3~6×10⁴cm^-3 to 2×10⁴cm^-3, and etch pits density(EPD) reduced from 1×10⁵cm^-2 to less than 5×10⁴cm^-2. This meant carbon film breaking off from the ampoule was major factors to form inclusion with high density and large size, and high EPD might be relative with impurities in the material. The size of inclusions and the densities of inclusion and etch pits could be further decreased to less than 5 μm, 4×10³cm^-3 and 1×10⁴cm^-2 respectively after taking BN coating technology, while cell structure of etch pits found in the previous CdZnTe wafers disappeared. This result further showed that high density of inclusion and EPD originated from carbon impurities in the materials.

In this paper, we report growth of AlN epilayers by pulsed atomic-layer epitaxy (PALE) method, by which the group III element sources and NH3 were alternatively transported into the reactor. We have systematically investigated the effects of growth conditions of PALE on the crystal quality of AlN epilayers. By optimizing the PALE growth conditions, the root mean square (rms) of AlN layers was 1.319 nm and the full width at half-maximum (FWHM) was 0.18 arcmin, room-temperature band edge absorbing peak at 198 nm was easily achieved, indicating the small mosaicity and low dislocation density of the films.

High power Ultraviolet Light-Emitting Diode (UV-LED) is currently in high demand for a variety of applications including lighting, printing, and polymer curing, with its' advantages of durability, reliability, non-hazardous and safety. Recently, the technology of Semiconductor Light Matrix (SLM) by multiple individual LEDs mounted on panels was put forward to obtain higher power for curing application. However, the illumination uniformity of high power LED arrays still remains an issue. On the other hand, it has been also predicted previously that the lifetime of a device decays exponentially as the temperature increases. Therefore, the thermal management of high power LEDs is also a critical factor to the development of high-power UV-LED array light source. This paper presents our latest investigations of illumination uniformity and thermal management to satisfy the requirements of packaging UV-LED arrays for curing application.

In this work, we have grown a-plane ZnO films on a-plane GaN/r-sapphire templates by pulsed laser deposition. The aplane GaN of the templates is aimed to mitigate the large lattice mismatch between ZnO and sapphire, and was grown by metal organic chemical vapor deposition. The grown a-plane ZnO films have been analyzed by various techniques such as high resolution X-ray diffraction, photoluminescence. It shows that high quality a-plane ZnO films have been achieved by our growth method.

We theoretically studied a two cascaded stages of microring coupled-resonator optical waveguides with corresponding analytical model developed. Simulations show that multiple electromagnetically induced transparency-like spectra appear in this system, which can be ascribed to the combined effects of mode splittings and instructive interferences. Applications of such multiple electromagnetically induced transparency-like spectra are promising to be tailored for multi-channel delaying lines.

A high-speed current driver circuit board for wavelength control of SG-DBR laser has been designed. Experiments have been performed to show performances of the driver circuit board and characteristics of wavelength switching of the SGDBR laser.

We proposed an all-optical tunable delay line based on wavelength conversion in semiconductor optical amplifiers (SOAs), and group-velocity dispersion (GVD) in single-mode fiber (SMF). The system operates in 1.5 μm window, with a nonreturn-to-zero (NRZ) pattern at 10 Gb/s. The maximal time delay up to 2600 ps is obtained. The scheme achieves continuous control of a wide range of delays, nearly no pulse broadening and very little spectral distortion.

In certain species of moths and butterflies iridescent colors arise from sub-wavelength diffractive surface corrugation of the wing-scales. The optical properties of such structures depend strongly on the wavelength, the incidence angle, the polarization of illuminating radiation, and the index of ambient medium. In this paper, after getting the SEM picture of the dorsal scales of the Morpho didius butterfly, we construct a bionic two dimension model, whose ridge contains a certain quasi-periodic arrangement of tree-like sub-wavelength microstructures. Then using a multilayer rigorous coupled wave analysis method in two dimensions, we study the reflection spectra of the wings of Morpho didius butterfly by simulating the multilayer model of a transverse cross-section comprised of the ground scale. Here we assume that the structure is made of a slightly lossy dielectric material and analyzed the polarization, the incidence angle and the index of ambient medium which affect the reflection spectra strongly. The results got, have revealed the natural phenomenon of iridescent colors and color-changed in essence, and the simulation results enable an artificial microsensor which discriminate vapor or component by reflective efficiency spectra.

In the photonic crystal (PhC) slab, a waveguide is typically created by introducing a line defect into the perfect periodic lattice, and the waveguide modes will appear within the photonic band gap. An practical waveguide should possess sufficient single-mode bandwidth and display minimal group velocity dispersion within the bandwidth. By adjusting the air holes adjacent to the line defect, the dispersion curves of waveguide modes can be tailored and the requests above are reached. In this way, an air-bridge slab waveguide gets a single-mode bandwidth of 0.025 c/a, where c is the light speed in vacuum and a is the lattice constant of the photonic crystal. In addition, an SOI (Silicon-on-Insulator) slab waveguide receives a single-mode bandwidth of 0.007 c/a and a frequency window of 0.003 c/a with a nearly constant group velocity of 0.082 c.

A novel uni-traveling-carrier photodiode based on materials of ultra-short carrier lifetime (ST-UTC-PD) is designed and numerically analyzed. The ST-UTC-PD solves the problem of high dark current of short carrier lifetime photoconductors, while takes advantage of the ultra-high speed response of short lifetime materials. By using ultra-short carrier lifetime materials for the absorption layer of UTC-PD, the limitation of speed by electron diffusion time is overcome, which yields ultrahigh speed response of ST-UTC-PD. The activities of electrons in the absorption layer are analyzed and the two deciding factors to the speed of ST-UTC-PD are discussed.

The miniature soft X-ray detector (M-XRD) is the key part of the miniature soft X-ray spectrometer, which was developed and used in laser-produced plasma interaction experiments on Shenguang III Laser Facility. The results of the analyze shown that the significant schemes was to get well-proportioned electric field distributions, which can obtain the strongest intensity of electric field, by using the ultra-broad band air transmission line to transmit the signal to the measure port in less aberration. Adopting the finite element theory, the simulation of the X-ray detecting diode and the design of ultra-broad band air coaxial transmission line by HFSS has been carried out. Compared with the traditional XRD, the size of M-XRD reduced to 1/3. The experimental results indicated that the response time was excelled to the traditional XRD and the response time was less than 80 ps.

Optical buffer is a key component in all-optical information processing systems. Slow light at room temperature via four-wave mixing in semiconductor an optical amplifier (SOA) is experimentally investigated. Time delay of 0.4 ns is achieved for a sinusoidal modulation signal at 0.1GHz, corresponding to a delay-bandwidth product (DBP) of 0.04. Factors that affect the experimental results are discussed. It is found out that the variable optical delay via four-wave mixing in SOA can be controlled either electrically by changing the SOA bias, or optically by varying the pump power or pump-probe detuning.

We propose a novel and simple non-return-to-zero differential phase shift keying (NRZ-DPSK) wavelength division multiplexing (WDM) system which can simultaneously demultiplex and demodulate multiple wavelengths. The phaseto- intensity demodulation principle is based on detuned filtering, which is achieved by using a single commercial array waveguide grating (AWG) in our scheme. By properly choosing appropriate AWG channels at the transmitter, the AWG at the receiver can act as both the de-multiplexer and the demodulator of the DPSK signals. Simulations at 10, 20 and 40Gb/s show good flexibility and performance for the proposed system.

The compensation of the characteristic deviations of Guided mode resonant filters (GMRFs) induced by over-etching is obtained by the cover layer of GMRFs. The investigations of GMRFs are reported widely, but there are few reports about the preparation and commercial application of GMRFs. The reason for this is that the characteristics of GMRFs are strict with the error of preparation. The data of this paper shows the linear relationship between the grating depth and the resonant wavelength in the grating depth error range of 0-15 nm. Through choosing material and adjusting the thickness of cover layer, 10 nm deviations of GMRF resonant value induced by 12 nm etching errors is adjusted almost. It shows that within the minor fabrication errors range, the grating layer and waveguide layer has different degree influence on the resonant peak value of GMRF. This method of adjusting the preparing error is very useful in the preparing of GMRFs.

Hydrogenated amorphous silicon (a-Si:H) thin films have been prepared by radio frequency magnetron sputtering. The effects of hydrogen pressure, substrate temperature and sputtering power on deposition rate of a-Si:H thin films and optical characteristics have been investigated. The films are studied by ultraviolet-visible spectrophotometer and NKD7000w thin film analysis system. The results show that the hydrogen pressure, substrate temperature and sputtering power will affect the deposition rate respectively; The refractive index, extinctive index and optical bandgap of hydrogenated silicon thin films vary regularly with the change of one of the deposition parameters. The optical bandgap, absorption coefficient and extinctive index of the films are evaluated. The absorption coefficient values range from 1.1×10⁴cm^-1 to 8.3×10⁴cm^-1 and the corresponding extinctive index vary from 0.23 to 0.35 at the wavelength of 400nm, and the optical bandgap of a-Si:H vary from 1.77 to 1.89eV.

The liquid crystal (LC) variable focus lens combining LC Fresnel Zone Plate with spherical lens (FZP-S) presented in the paper is an electro-optical system using the applied voltages to control the variation of its focal length. Theoretical study for the mathematic model of the hybrid FZP-S lens shows that the focus length of the hybrid lens can use the similar method to that of the system with conventional refractive lens. The LC FZP with the electrodes widths vary as a FZP which was fabricated on a glass substrate with two layers of transparent electrodes in our experiment. The minimum of the electrodes width is 1.2um.The advantage of such system is that the focal length of which can be changed electrically without mechanical movement. The effect of the focus length variation was verified by the experiment results.

In order to enhance the civil digital photography system's latitude, a kind of new structure is put forward which is made the liquid crystal plate coupled to CCD/CMOS (Charge Coupled Device/Complementary Metal Oxide Semiconductor) sensitive chip in this paper. We call this structure chip is LCCCD (liquid crystal CCD). This new system's key is proximity coupled a high precision HTPS (Low Temperature Poly-Silicon) liquid crystal plate on the foundation of current CCD/CMOS chip; each pixel on the liquid crystal plate corresponds with the pixel on CCD/CMOS one by one. They compose the new sensitive photosensitive chip that can control the each pixel's ratio of photoelectric conversion by changing each liquid crystal unit's transmittance. This paper expounds this system's structure as well as various modules' cooperation mode and the process of achieving the aim. Through analysis original image by common camera and the same image by LCCCD camera, these results are obtained: the image is changed directly on the physical level by the new system is much better than produces by post treatment, moreover the system can rectify images by 10bit precision, and the imaging latitude has been enhanced more than 5EV (EV is exposure level).

A novel imaging system has been designed in order to enhance the dynamic range of image intensifier, and make it image normally at high illumination based on the principle of liquid crystal for partial gating detection. This system is made up of lens, HTPS TFT-LCD (proximity to the image intensifier), photometric CCD, main CCD and control circuit. Under the high illumination environment, main control chip FPGA extract the output signals of photometric CCD as feedback signals, and get the environmental absolute intensity after some algorithm processing. Then the chip renews the color signals' value of YUV formats. Lastly, the HTPS TFT-LCD drive board receives the processed color signal, controls the transmittance of every pixel on the HTPS TFT-LCD. So the ultra bright partial of image can be weakened. Under the normal or weak light condition, LCD will be moved away by mechanical devices because its transmittance is about 30%. This paper gives the control principle and structure of the system and puts forward the CCD' photometric principle.

An application of the genetic algorithm in designing omnidirectional optical filters is reported in this paper. Concerning different periodic numbers and thickness ratios in the photonic crystal, we gave some optimization examples and finally achieved a photonic heterostructure with very broad omnidirectional filter bandwidth as well as two very narrow transmission windows. And they are found that when the normal incident beam is tilted at a negligibly small angle, the perfect transmittance peak will vanish. Hence, this heterostructure can be regarded as an omnidirectional high-precision two-channel filters with potential application in optical filters, optical switches, and many other optical telecommunication areas.

Due to the Multipoint-to-Point nature of the uplink, the upstream data transmission in a GPON system is burst-mode, and both the guard time and preamble time are much shorter(32 bits and 44 bits respectively).This burst-mode nature of the GPON uplink brings many challenges for the design of the burst-mode receiver. This paper presents an improved design of the main amplifier which is fit for the AC-coupled burst-mode optical receiver with short time constant. Simulation analysis of this scheme at the aspect of the pattern dependent jitter and the data pattern jitter is also presented. Finally, simulation results are provided to show the feasibility of the scheme.