A method of discriminating Doppler shift using Fizeau interferometer and linear PMT array for a direct-detection wind Doppler lidar is proposed. At first, its working principle is introduced. According to the spectral character of Fizeau interferometer, optimum design of frequency shift discrimination interferometer is presented, and the designed Fizeau interferometer has the cavity length of 150 mm and reflectivity of 0.755 from plate. The numerical simulation for wind speed measurement using the Fizeau interferometer and PMT array detector is also presented. The effect of molecular backscattering taken as background noise is analyzed. The error of wind speed measurement using system with Fizeau interferometer and linear PMT array is calculated. The results show that wind velocity error can be less than 0.56 m/s in the 5 km range under 30 s average time using the proposed system.

Aerosol detecting using lidar is an important part of atmosphere monitoring. Many aerosol properties are gained on the base of the aerosol extincition coefficient received through lidar equation. The traditional method restricts the precision of the aerosol extinction coefficient because of many hypotheses. A new procedure was developed, based on the Madaline artificial neural network theory to retrieve the aerosol extinction coefficient. Through the trained network, the aerosol extinction coefficient could be retrieved from the lidar echo signal directly, consequently the new method dispensed with those hypotheses in the traditional method. According to the comparison of the different test results, the new method improved the precision of retrieval greatly and good result was obtained.

The sensitivity of a laser optical particle counter (L-OPC) measuring to the refractive index of particles is analysed, and the dependence of response on ni, the imaginary part of the refractive index is discused. Sensitivity function fi is defined. A scattering angle less than 20° is chosen and another is between 40° and 60° according to analysis of theory. For example, scattering angles system ψ=9°, β=5°, λ=0.65 μm and ψ=50°, β=20°,λ=0.65 μm, the one is slightly not effacted by ni, the other is greatly. By useing different sensitivity of different scattering angles to ni, the value of ni can be deduced from the measurement by a laser optical particle counter of double scattering system. The simulation method will be used to deduce ni.

The relatively low diffraction efficiency of photorefractive holographic recording, especially that of the two-center holographic recording with 90° geometry, is studied, based on the theory of local diffraction. It shows that the diffraction efficiency in 90° geometry can be comparable with that in transmission geometry only when the amplitude of refractive index gratings is more than 10-4 with the same refractive index change and beam width of 2 mm. According to the effects of fringe vibration caused by environmental perturbation on holographic recording, a concept of effective modulation is proposed. Narrow grating space in 90° recording geometry makes it more sensitive to ambient disturbs and leads to low refractive index change. To obtain high diffraction efficiency in experiments with 90° geometry, an active fringe stabilization system is necessary. Furthermore, in two-center holographic recording, the effects of microcosmic optical parameters on diffraction efficiency are also important.

A new type of symmetrical demodulation method is proposed for the distributed fiber Raman temperature sensor. The anti-Stokes scattering optical time domain reflectometer (OTDR) curves are demodulated by Rayleigh optical time domain reflectometer curves. Firstly, the measured temperature scope is divided into several regions; secondly, when the temperature is obtained by the traditional demodulation method, it will be located on the relative region; then the temperature is demodulated by the region's demarcated temperature, even-number values of temperature are obtained, and the average value is taken as the measured value. The temperature precision and stability of the sensor is improved, with temperature precision error in ±0.05 ℃ and spatial resolution of 1 m. The experimental results agree well with the theoretical ones.

A novel approach for measurement of refractive index based on analysis of surface plasma resonance (SPR) spectrum is proposed. It is found that the variety of liquid characteristics will lead to the change of refractive index sensitivity of SPR spectrum. According to the different range of refractive index, resonance wavelength and resonance intensity can serve as measuring parameters, respectively. The refractive index resolution can reach 10-5 theoretically by using this method. Based on theoretical analysis and experiment, a reflective optical fiber SPR sensor system based on resonance light intensity interrogation is designed and demonstrated, and the ratio of sensing signal and reference signal is used as the measurement value of refractive index. In the refractive index range of 1.3325～1.3991, the measurement value decreases with the increase of refractive index, with a linearity of 0.9983. In addition, the coupling coefficient, which can be used to estimate intensity and change trend of SPR effect, is researched.

Guided-wave optics method is used to determine the temperature of the crosslinking polymer films. In the parameters of polymer films are determined according as the guided modes peaks in the attenuated total reflection spectrum depend strongly on the refractive index of polymer films. The refractive index change of the crosslinkable polyurethane is investigated with the ascending temperature. Experimental results show that the curve for the change of refradive index with temperature dn/dT evolues uncontinously, and there exist two intersection points. The temperature range between the two intersection points is regarded as the polymeric crosslinking temperature range. The first and second intersection point are regarded as the beginning temperature of crosslinking and the completing temperature of crosslinking, respectively. The crosslinking temperature range of the polyurethane, namely optimal poling temperature range, is from 85 ℃ to 122 ℃. The measured result is examined and convinced by the measurement of electro-optic coefficient of the poled polyurethane film with several temperatures in the range of crosslinking temperature at which the poling process starts.

All-optical regeneration based on self-phase modulation in a highly nonlinear microstructured fiber is proposed. The dispersion and nonlinearity properties of a series of microstructured fibers are analyzed and the results show that the nonlinearity coefficient is closely related to the structure of the fiber. The nonlinear coefficient is increased by reducing the effective mode area and a highly nonlinear microstructured fiber with large air-filling ratio is used as the nonlinear medium in optical regeneration. The process of all-optical regeneration based on self-phase modulation in the fiber is investigated. The numerical results show a good optical regeneration result can be obtained by using a relatively short fiber length due to the high nonlinearity of the fiber. The input peak power launched into the microstructured fiber and the parameters of the filter have much influence on optical regeneration. In order to achieve good optical regeneration results, those parameters need to meet certain requirements. Furthermore, the transmission characteristic of the regenerator is also discussed. With adjustable input peak power and filter parameters, the regenerator can deal with the input pulses with different pulse widths.

Optical coherence tomography (OCT) is a recently developed imaging technology. With OCT technology one could obtain clear tomography pictures of retina and macula lutea. For the purpose of clinical diagnosis and retina image recognition, a preprocessing technique of retina OCT image has been developed and demonstrated. The edge and contour were extracted automatically through image segmentation and enhancement. It opened the door for the further recognition and analysis of OCT image. Experiments were done on cystoid macular edema, macular hole and normal retina OCT image. The effectiveness of the proposed methods is demonstrated by practical experiments.

In order to analyze and evaluate the spectrum restoration property of image restoration algorithms, a new method based on Gaussian function is proposed. The optical transfer function H and degraded image spectrum function G are assumed as Gaussian functions, and the variance and variance ratio are used as indexes of the spectrum width to analyze and evaluate the spectrum restoration property of the image restoration algorithms quantificationally. The curves of H and G are enacted in two groups of different variances, and the variance and variance ratio of the image spectrum restored by constrained least squares restoration (CLS) and maximum likelihood estimation (PML) are calculated in no-noise and with-noise cases in analysis. The two algorithms are analyzed and evaluated by the calculation results, and the analysis proves that PML is better than CLS in spectrum extrapolation and noise restraint. The experimental results show that the method is easy and effective.

Based on the imaging characteristics of the spatially modulated imaging interferometer, the new compression method for interference spectral images, weighted rate-distortion optimization for set partitioning in hierarchical trees (SPIHT), is proposed. According to the image features in time and frequency domain, it is deduced by the proposed method that, with the increase of the optical path difference in time domain, the effect of interference spectral images on the spectral curve profile in frequency domain is increasing. In the optical path difference direction, the slope ratio of rate distortion is lifted increasingly, and the protection for images is enhanced, which not only makes up the deficiency of code rate distribution in SPIHT, but also protects the spectral information in frequency domain efficiently. The experimental results show that the proposed algorithm achieves improved performance over the conventional algorithm in time and frequency domain.

To reduce the influence of factitious factors on triangle orientation discrimination (TOD) threshold measurement and so enhance the measurement efficiency, a machine vision system to perform TOD instead of human observers was studied, and an objective discrimination model (ODM) for TOD was developed. Based on the main effect and course of signal processing of human visual system (HVS), ODM uses oriented filters to extract essential information from image-signal and discriminate the orientation of a triangle by comparing the visual character vectors with standard ones. With the triangle images generated by the virtual staring imaging system, experiments were held to compare the performance of ODM with human observers' one. Three groups of experimental results of different triangles size were given and the results show that ODM can predict the observers' performance well.

For the forward problem of excitation photon and emission photon in the fluorescent molecular tomography, which is usually solved in the sequential manner, a parallel algorithm which can simultaneously solve two forward diffusion equations is proposed. The algorithm realizes parallel computation through introducing multiplier matrix and decoupling the coupled equations. Two-dimensional numerical simulation is presented with the finite element method. Then the algorithm proposed is compared with the sequential method and the parallel algorithm proposed by Ralf B. Schulz et al.. Experimental results show that both the speed of the forward computation and precision of the forward solutions can be improved by using the proposed algorithm, which is adaptive to any Stokes shift and has more broad application. As a result, the whole process of fluorescent molecular tomography will also benefit from it.

To register multiple range images, we present a method for the pose estimation of multiple viewpoints for three-dimensional digital imaging system based on fringe projection. This method illuminates the test object with two orthogonal fringe sequences, respectively, from at least two different viewpoints, then reconstructs Cartesian coordinates of the object in three-dimensional space from phase map with help of fringe projection and phase reconstruction techniques. As the projector can conceptually be regarded as camera acting in reverse, the relation between projection plane of the projector and imaging plane of the camera is fixed. Taking epipolar geometry into account, this approach can further realize the pose estimation of three-dimensional active vision system based on fringe projection. We build up a non linear equation to calculate the parameters of orientation even if the input data are influenced by noise. Mathematical formulation of the proposed method is given and the effectiveness of this approach has been demonstrated by the designed experiments.

A method of optical radiometric calibration based on correlated photons was proposed to improve calibration precision. It was independent of the methods based on radiometric source or detectors. With 351 nm laser to pump β-Barium Borate(BBO)crystal, a radiometric calibration system based on correlated photons had been set up, and its principle and experimental setup were analyzed. Coincident metrical system was given, and the experimental result is discussed. The results showed that the calibrated quantum efficiency of photomultiplier (R2949) at 702 nm was (5.96±0.003)% , the relative error of experimental result was about 1.0% with respect to the typical quantum efficiency of photomultiplier (R2949)at this wavelength, and relative combined uncertainty of the results was about 2.61%.

A phase-unwrapping method based on dual-frequency colorful fringe projection was proposed to improve measurement speed. Only one frame image was collected to realize phase-unwrapping in high-speed measurement and dynamic object contour measurement. The dual-frequency phase measurement and the principle of unwrapping with variational precision were discussed, and then factors affecting measurement precision were analysed in detail. A dual-frequency dual-color sinusoidal fringe pattern encoded by computer was projected by a liquid crystal digital projector. And two monochroic fringe patterns were analysed by Fourier transform to obtain height information of the measured object with high and low precisions and then carry out the unwrapping with variational precision. The results showed that the measurement speed and the unwrapping precision were improved. The maximal absolute error was +1.413～-1.582 mm, and the standard deviation was 0.363 mm.

We demonstrate an all-fiber pulse shaping system that can generate 2.2 ns pulses with arbitrary pulse shapes by stacking a set of 100 ps chirped pulses. To form the standard 100 ps chirped Gaussian pulses, stable pulses generated from an Yb3+-doped mode-locked fiber oscillator is stretched and filtered by a chirped fiber Bragg grating (CFBG) and a 1 nm Gaussian spectral filter. The standard pulse is selected and stacked in an all-fiber pulse stacker composed of 32 time delay lines, and a 2.2 ns arbitrarily shaped pulse with 32 bit shaping precision operating at 1 Hz rate is obtained. The characteristics of the stacked pulse are studied, and the analysis shows that the wide spectrum of shaped pulses which periodically sweeps from long wavelength to short wavelength may optimize the beam smoothing in inertial confined fusion (ICF). Experimental results shows the shaped 2.2 ns pulse has a rise time less than 50 ps, and it is precisely synchronized with the 100 ps pulse with time jitter less than 4 ps.

Unlike traditional coherent laser beams, the beam quality evaluation of partially coherent beams involves not only its far-field radiation characteristic but also partial coherence of the light source itself. The generalized expression for the phase-space product Q of a beam generated from a partially coherent source is presented by using coherent-mode representation of partially coherent beams. Compared with the result for the fully coherent light source obtained in previous works, the derived expression for Q includes not only contributions directly from the linear combination of each individual elementary fully coherent modes but also interactions between different elementary modes.

Atmospheric pressure metal-organic chemical-vapour deposition (MOCVD) has been used to infiltrate the voids within synthetic opals with InP to modify the natural photonic behaviour of these materials. Experiment is carried out for given growth conditions to study the effect of growth cycles. Morphologies and reflectivity spectrum properties of the artificial opal and InP-infiltrated artificial opal were researched by SEM and UV-Vis spectrometer. The results show that an increase in the InP infiltrationl of the pore volume is achieved if the total time of each reactant over the opal is constant while the total number of cycles is increased. The extent of InP infill within the voids increasing in turn increases the extent of refractive index contrast between the silica spheres and the void, and it is possible to modify the opal photonic band gap in a controllable manner. It is also found that the periodic growth is in favor of InP infiltration the InP grown shows a high quality by SEM. The study provides a scientific basis for manufacturing three-dimensional InP photonic crystals.

White organic electroluminescence devices (OLEDs) were fabricated, with BCzVB as blue luminescence dye and Btp2Ir(acac) as red phosphorescence dye doped into CBP substrate as the light-emitting layer. The reason for the system color-coordinate shift is explored. For the device with the doping mass fraction of 6% BCzVB and 0.2% Btp2Ir(acac), the color-coordinate shifts slightly from (0.340, 0.273)to(0.308, 0.273), when the driving current varies from 4 to 200 mA/cm2. The reason of the slight blue shift is derived by the luminescence spectrum and luminance-current density curve. It is found that internal physical processes, such as the complete energy transfer from host CBP to dopant Btp2Ir(acac) and incomplete energy transfer from blue dye of BCzVB to red dye of Btp2Ir(acac), and luminesence saturation of phosphorescent dye of Btp2Ir(acac) determine the slight blue shift.

To investigate the influence of dry etching on strained multiple quantum wells (SMQWs), we etch the cap layer of metalorganic chemical vapor phase deposition (MOCVD) grown InGaN/AlGaN SMQWs about 95 nm by inductively coupled plasma (ICP). Photoluminescence (PL) measurements show that, the photoluminescence intensity of the quantum wells is enhanced about 3 times after dry etching. The surface of the quantum well becomes fluctuant, the degree of roughness becomes larger, and it allows the photons generated within the quantum wells to escape surface more easily by multiple scattering at the rough surface. The theoretic calculation result shows that the pattern change of the quantum well surface causes the enhancement of the photoluminescence intensity about 1.3 times. On the other side, because the inductively coupled plasma power is small, the damage depth almost cannot reach the well layer, but the tunneling of Ar+ inside the quantum well structure may bring new luminescence centers, and enhances the photoluminescence intensity.

An advanced microbe chip testing method was put forward, and a new digital imaging scanning system for testing biochips was developed. A novel CCD scanning system with laser induced fluorescence for biochip detection was proposed, with a high numerical aperture NA=0.72 for collecting fluorescence, long working distance 3.22 mm and high sensitivity less than 1 fluorescent molecule /μm2. Microbe chip was designed and tested by the above novel CCD scanning system, which provided a whole solution scheme for efficient food safty inspection. An example of testing escherichia coli and xanthomonas campestris was shown and proved the stability and creditability of the system. The experimental results obtained by the developed digital imaging scanning system indicated the same correct identification of microbe as the commercial confocal scanner.

The axial resolution of optical coherence tomography (OCT) depends on not only the bandwidth of the light source but also the dispersion balance between the sample arm and reference arm. If the dispersion between two arms is not matched accurately, the axial resolution of optical coherence tomography system will degrade. Dispersion problem is very important especially in ultrahigh resolution optical coherence tomography system. A method for dispersion compensation in optical coherence tomography by a double-grating based rapid scanning optical delay line (RSOD) is put forward. The independent variable of the distance between gratings in the added gratings, along with the adjustable distance between grating and Fourier transform lens in a conventional rapid scanning optical delay line can be used to compensate the group velocity dispersion (GVD) and third order dispersion (TOD) of optical coherence tomography system simultaneously. The dispersion characteristics as well as the dispersion management scheme are discussed. Example of dispersion compensation in a typical optical coherence tomography system is provided.

Spectroscopy of polarized light scattering is a novel technique prospected to diagnose precancerous lesions. However, a perfect theoretical method to study the mechanism of this technique is absent. With the parallel computing and Stokes-Müller formalism to describe the polarization of photon and media, respectively, a Monte-Carlo program, which can simulate the spectra of polarized light scattering, is developed. The program is designed with a master-slave architecture based on message passing interface (MPI), and a dynamic load balance is obtained via taskpool scheme. It has been validated by the comparison of polarized light backscattering spectra in turbid media between simulation and experiments. Furthermore, results show that this program can provide excellent scalability of task size and slaver number. For long-time required simulations, the computational time of using n slavers can be reduced as almost 1/n times of using 1 slaver.

The quantitative measurements of layer tissues, such as eyeground, in optical coherence tomographic (OCT) imaging depend on the extraction of layer structures. The second derivative item is introduced to control diffusion along coherence orientation in coherence enhancing diffusion algorithm (CED). The method is expected to preprocess the original optical coherence tomographic images to remove noise and speckle, enhance and protect the layer structure, and realize more accurate quantitative measurement for the significant layer structure of images. The coherence enhancing diffusion algorithm with a second derivative item is used in optical coherence tomographic images for different samples. The quantitative measurement for the thickness of significant layer structures in optical coherence tomographic images is done, by combining the location of the layer structure and sample refractivity information in the preprocessed images. Experimental results show that it is helpful to preprocess the optical coherence tomographic images with the modified algorithm for more accurate measurement of significant layer structures.

The feasibility and sensitivity of optical waveguide lightmode spectroscopy (OWLS) in studying DNA-protein interaction are dissused. In the label-free DNA and DNA-protein interaction detection assay based on optical waveguide lightmode spectroscopy (OWLS), the probe was located online onto the amino silanized sensor chip of optical waveguide lightmode spectroscopy by glutaraldehyde covalent coupling; the DNA-protein complex, combination of NF-κB p50 homodimer and oligonucleotide with binding site to NF-κB p50 homodimer and binding sequence to probe, was used as the detect molecule. The different concentrations of p50 in detection samples and affinity of oligonucleotides with different binding site sequence of NF-κB to p50 were tested in this system. With this detection system, as low as 0.33 nmol/l of NF-κB p50 in the samples was detected, and the affinity of oligonucleotides with different binding site sequence to NF-κB is quite different. The label-free studying system of DNA-protein interaction based on the optical waveguide lightmode spectroscopy technique is feasible and high sensitivity, and this system can be used in the research targeted at the interaction of DNA and DNA-binding proteins and detection of the affinity of oligonucleotides to binding protein.

Thiazone, a kind of new enhancer, is introduced to investigate the effect of optical clearing on skin tissue. The samples including skin with or without stratum corneum (SC), are treated with pure polyethyleneglycol 400 and mixed solution (thiazone in combination with polyethyleneglycol 400) respectively. The transmittance of light with wavelength 632.8 nm is measured dynamically by using an integrating sphere system, and then photos of the samples are taken by CCD. Experimental results show that mixed solutions have optical clearing effect on both kinds of skin, The change in attenuation coefficient of skins without SC is 19 times of that with SC. Single PEG400 can only do some effect on the sample without SC, and the change in attenuation coefficient of samples is just 2/3 of that treated by mixed solution. It is illustrated that thiazone can increase the effect of optical clearing on skin, because it is an efficient enhancer and has synergistic effects with PEG400.

Compared with event-related potential (ERP), which is widely used in psychological research, functional near-infrared imaging (fNIRI) is a new technique, and can supply hemodynamic information related with brain activity, excepting for electrophysiological signals. We conducted a series of experiments with the classic paradigm of ocular nonselective attention, monitored responses if brain with fNIRI and ERP respectively, and then analyzed the data. The results showed us that ocular nonselective attention companied activation was on the prefrontal lobe, which had the same results as its related P3a in ERP. In fact, the fNIRI results demonstrated a smaller area of the source than ERP, which was at the middle part of Brodmann 46, and the phenomenon that the hemodynamic activation occurred before P3a and strengthened just after P3a, supported the possibility of further study in brain mechanism. Therefore, fNIRI is a potential technique in advanced function research of brain.

Depolarizer is a key device in polarization sensitive optical systems, by which a completely or partially polarized light is converted into unpolarized one. An artificial polarization-dependent delay is employed substitute for polarization maintaining fiber, a compact depolarizer for linearly polarized light with any direction of vibration is proposed. Before being depolarized, the incident linearly polarized light is converted into an elliptical polarized light by a quarter-wave plate and is divided into two orthogonal polarized parts always with equal intensity, so that the depolarizer is independent of the optical field direction of incident light. The depolarization performance evolution with the order of quarter-wave plate, center wavelength and optical field direction of incident light is calculated. Low DOP (2.6%), low insertion loss (0.6 dB) and small insertion loss fluctuation (0.11 dB) are obtained for a linearly polarized light with 0.13 nm spectral width. Theoretical analyses and experimental results show that proposed depolarizer has low DOP, low insertion loss and small insertion loss fluctuation, and can be operated in a wide spectral band.

The energy from the source was rearranged through the reflection by a freeform reflector, and the desired illumination on a certain plane was got, which simplified the illumination system in the projectors and provided a choice to minimize the device further. A partial differential equations set was presented according to the character of the light source and the desired energy distribution on the illumination plane. The numerical results of the partial differential equations set were investigated to obtain a freeform reflector. LED with a Lambertian light emitting surface of 1 mm×1 mm and a viewing angle of 120° was applied as the light source, and the desired illumination on certain face was a 4∶3 rectangular. Two kinds of free reflectors were designed, and the simulated results showed a uniform illumination near to 90%. The projection lengths on the optical axis of both free form reflectors are less than 25 mm, and the projection lengths on the axis perpendicular to the optical axis are both less than 22 mm. The total lengths of the illumination system are both less than 40 mm, so this system is very compact and suitable for the portable LED projectors.

The current injection is a daunting problem in designing and fabricating electrically driven single-cell photonic crystal microcavity laser, and the current path often decreases the quality factor of the microcavity. The electrically driven single-cell photonic crystal microcavity laser is studied with three-dimensional finite-difference time-domain. The eigen mode distribution of the microcavity is given, and accordingly the monopole mode is chosen and investigated.Numerical simulation shows that, under the condition of C6V symmetry, the positions micro-tuning of the six poles nearest the cavity center will result in slow change of the optical field of the monopole mode in the cavity marigin. The quality factor and Purcell factor are raised 7 times, and the laser threshold decreases to one seventh of the original value.

Based on coupled-mode theory for linearly polarized (LP) modes, spun dual-lobe stress region fibers (i. e. spun linear birefringence fibers) and spun 4-lobe stress region fibers are theoretically analyzed. The expression of the anisotropic dielectric tensor of the additional stress-induced dielectric constant at the core region of the fiber is presented. By the expression, the coupled-mode equations for the linearly polarized modes propagating in multi-lobe stress region fibers with rotation angle are derived. Then, super-local modes in multi-lobe stress region spun fibers can be obtained by using mode transform. Finally, the fiber birefringence is decided from the phase constants of super-local modes. It is found that high elliptical birefringence can be obtained in spun dual-lobe stress region fibers with ideal symmetry, and there is no birefringence in spun 4-lobe stress region fibers if four lobes are perfectly symmetrical. The same result is valid for spun fibers with 8-lobe and 16-lobe stress region, both of which have zero birefringence.

The cavity field spectrum of an atom interacting with two-mode fields in the presence of Stark shift is investigated. The cavity field spectrum expression of the system with the cavity fields in photon number state is given. Some numerical results are presented. The effects of Stark shift on the frequency and intensity of the spectrum structure are obvious, when the initial fields are weak. The symmetry of the spectrum was broken by the Stark shift and a complex spectrum structure appears. When the two-mode fields are both strong, the effect of Stark shift on the spectrum is weak. When the field I is vacuum, and the initial intensity of field Ⅱ is increasing, the higher-frequency peak of field Ⅱ is heavily depressed by Stark effect. While, lower-freqnency peak is depressed for weak initial field, and enhanced for strong initial field. When field I is vacuum and field Ⅱ is much stronger, mode Ⅱ spectrum show a single classical resonant fluroscent spectrum.

Ti-doped WO3 films were prepared by mid-frequency dual-target magnetron sputtering method. The Ti-doped WO3 films were characterized with X-ray diffraction (XRD), Raman spectroscopy, ultraviolet spectrophotometer, chronoamperometry and atomic force microscopy (AFM) .The results indicate that after doping Ti, its crystallinity decreases at the same annealing temperature, and its surface roughness decreases dramatically with the increase of holes for ion injection and extraction, Accordingly, the response speed of the WO3 film doped with fraction of atom number 0.051 of Ti is greatly improved compared with the undoped film, and its cycle life is increased more than four times, but the spectral transmittance decreases after doping.

The reasons leading to polarization for optical coatings at oblique incidence were analyzed. Analysis was made for the equivalent admittance and phase of different polarization states, and the equivalent indices and phase thickness for both polarizations of a symmetrical structure at 45° incidence were computed. Then a non-polarizing broadband antireflection coating with good optical performance within 600～900 nm was designed by using equivalent layer method. A sample was fabricated through electron beam evaporation technique, and its optical performance absolutely satisfied the requirements. From 600 to 900 nm, the average reflectance was less than 1.38%, and the polarization of reflectance was lower than 0.89%. In addition, after analyzing its theoretical and experimental performance, angle sensitivity and layer-thickness error sensitivity, we drew a conclusion that symmetrical layer combination was a practical way to design a non-polarizing broadband antireflection coating at oblique incidence.