We have investigated the formation and the nature of negative electron affinity on p-type GaAs surfaces prepared with Cs and NF₃ as the fluorine carrier. The results have been compared with those obtained from the Cs-O prepared NEA photocathodes. High resolution photoelectron core level and valence band spectroscopy is utilized to reveal the underlying physics and chemistry during the NEA activation process. We have shown clear evidence of dipole formation both at the physics and chemistry during the NEA activation process. We have shown clear evidence of dipole formation both at the Cs/GaAs interface and in the activation layer. We also demonstrate that no chemcial reaction takes place between fluorine and the substrate. Other aspects related to the NEA formation such as the so-called two-stage activation, activation layer stoichiometry, and aging process have been studied. The study of this model system leads us to conclude that the NEA formation can be adequately explained by the double dipole layer model.

A Monte Carlo technique has been used to calculate the electron energy distributions and the escape probabilities (EP) for the negative electron affinity (NEA) GaAs transmission photocathode. The results are compared with experimental data. A simple model, which assumes electrons from the (Gamma) minimum and that ideal conditions exist at the semiconductor-vacuum interface, does not work properly. Different types of scattering in the activation layer and a nonideal interface are considered as possible causes for the discrepancies between calculations and experiments. The results show that a non-ideal interface could be the best candidate for explaining the behavior of the electron angular distribution.

The longitudinal component of the photoelectron energy distribtuion from a transmission mode GaAs photocathode with quantum efficiency up to 18% is accurately measaured by using Generation III image intensifier as a parallel plates retarding field electron energy analyzer. The measurements are performed within the temperature range of 77-300 K with different photon energies for photoelectron excitation above or below the band gap of semiconductor. It is shown that conduction band tails in heavily doped GaAs influence photoelectron transport near the surface during the process of emission into vacuum. The transmission coefficient of a residual surface barrier is measured as a function of electron kinetic energy above the conduction band minima and is found to be much less than unity. The obtained results demonstrate the usefulness of this technique for the investigation of the physical factors which control the quantum efficiency and other parameters of photocathodes with neagitive electron affinity.

We have developed two types of metal package photomultiplier tube (PMT): the R5600 series and the R5900 series. The R5600 series is a compact PMT that incorporates an 8-stage electron multiplier (metal channel dynode) constructed with stacked thin electrodes into a TO- 8 type metal can package. The R5900 series consists of the same thin electrodes as the R5600 series in a 28 mm square 20 mm height metal package. The metal channel dynode is supported by computer simulation of electron trajectory and fine etching technique.

This paper details the status of a program at Intevac ATD to develop high sensitivity transmission photocathodes which function in the 0.95-1.7 micron wavelength range. The goal of the program is to develop this technology for use with both imaging and nonimaging detectors. Sealed tube results are presented. Measured performance characteristics include: cathode spectral response, dark current, linearity, and the effects of cooling. A brief discussion of planned development, potential applications, and simple modeling illustrating the advantages of the proposed detectors are included.

The energy spectra of electrons emitted from transmission-mode negative electron affinity photocathodes have been measured at high resolution using a parllel-plate retarding technique. The spectra from GaAs photocathodes have a basic structure that varies with temperature, activation layer qualitites, cathode thickness, and illuminating wavelength. A FWHM energy spread of approximately 50meV at room temperature has been achieved. Spectra from a GaAsP cathode show a markedly different structure and a much wider energy spread.

Demand for a cheap and simple laser detector in industry has become high due to a growing number of laser sources. Clearly, a rugged, simple, and cheap method of detecting these laser wavelengths had to be devised. The problems inolved were to 'see' the wavelength to recognize and identify it. To do this, some form of optical device for channeling captured light into the detector was devised. This light had to be converted into a digital signal, then to a digital readout giving information on the wavelength received. In turn this information could be used to operate devices to take any appropriate action on incoming wavelengths. The electro-optical methods that were used and devised at DRA Malvern are described including an outline to the history of this project. The working solutions and results are also contained in this report.

The main requirements for an optical detector are low noise and high sensitivity. The old method used was only to cool the photomultiplier tube. This new method of compact cooled- PMT cools only the photocathode. The main noise of the photomultiplier tube is generated from the photocathode. By utilizing this new method of cooling the photocathode, the noise from the photomultiplier tube can be reduced significantly. A small thermo-cooler attached to the photocathode can cool the photomultiplier tube directly. As a result, within five minutes, the photocathode reaches a lower temperature, for example, of 5 degrees C in an ambient temperature of 20 degrees C. The performance of the photomultiplier tube is one tenth of the dark current of ambient temperature. Also, the dark count decreases by a factor of one tenth of the ambient temperature. Equivalent noise inputs wer 2.3 X 10^-17 (at 600 nm).

The fabrication and performance of thin film p-i-n and metal-semiconductor-metal (MSM) photodetectors and the integration of these detectors onto silicon circuitry is presented. The thin film photodetectors are separated from the growth substrate using epitaxial lift off or total substrate removal, and are subsequently bonded to silicon circuits. Performance of the thin film photodetectors is comparable to on-wafer counterparts, and in the cases of resonant cavity p-i-ns and inverted (fingers on the bottom) MSMs, the performance is enhanced through the removal of the substrate. Receiver circuits have been designed, integrated with thin film photodetectors, and tested. Finally, smart pixel arrays of photodetectors have been integrated directly on top of an array of silicon oscillator circuits to demonstrate three dimensionally interconnected image processing systems.

The monolithic integration of a SiGe-optical waveguide with a detector based on SiGe- absorbing layers is presented. A maximum internal quantum efficiency of (eta) equals 40% has been measured at (lambda) equals 1.3 micrometers , which corresponds to an external efficiency of (eta) equals 11%. This device is suitable for 2.5 GBit/s data transmission, the performance is limited by the RC time constant due to a capacitance of C equals 1.7 pF.

Today very compact optical sensor systems are realizable with common CMOS processes. Line cameras for visible and near infrared light may include intelligent image processing and a bus interface on the same chip gaining adaptive, robust, and cost efficient microsystems. Major advantages are the contrast of 10⁷ to 1, the variable frequency up to continuous mode and the combination of analog and digital signal processing. Line sensor systems with up to 512 phototransistors for several applications have been realized. The physical and electrical parameters including the on-chip-signal-processing will be explained. For the positioning of rotating parts in machines the position angle may be required with a resolution of 0.5 inches or better. Such angular sensors may be realized as optical line sensors. Specialized color slides have been used as filters in combination with the line sensors.

In this paper, we discuss the range of applicability of the drift-diffusion and hydrodynamic models as applied to the study of interdigitated metal-semiconductor-metal photodetectors. The hydrodynamic model is an extension of the standard drift-diffusion technique which determines the electon and hole energies in addition to the carrier concentrations and potential. The hydrodynamic method can properly account for energy dependent phenomena such as nonstationary transport phenomena and thermionic emission currents. The key engineering figure of merit, the time response, is calculated and compared using both models for a 1D device design that closely mimics an InGaAs/AlInAs metal-semiconductor-metal device. Structures incorporating heterobarriers and blocking contacts, wherein differences between the energy dependent and independent models are expected to occur, are examined.

In this paper, we present a comparative study of transparent and opaque electrode InAlAs/InGaAs metal-semiconductor-metal photodiodes (MSMPDs) for operation at 1.31 and 1.55 micrometers . The transparent materials are indium-tin-oxide (ITO) and cadmium-tin-oxide (CTO) and the opaque material is Ti-Au. The rf magnetron sputtered films of ITO and CTO, deposited at a substrate temperature of 300 degrees C, exhibited as-deposited resistivities of 5.6(DOT)10^-3 (Omega) (DOT)cm and 1.0(DOT)10^-3 (Omega) (DOT)cm, respectively. The resistivity of the ITO and CTO films dropped to 1.1(DOT)10^-3 (Omega) (DOT)cm and 5.2(DOT)10^-4 (Omega) (DOT)cm, respectively, after a 4 minute 400 degree C anneal in an N₂ ambient. The interdigitated ITO and CTO electrodes were made by etching in a methane:hydrogen (1 to 3) plasma. The responsivity of 1 micrometers finger by 1 micrometers spacing (1 by 1 micrometers ), 50 X 50 micrometers ² active area, MSMPDs was 0.40 A/W for the Ti-Au, 0.66 A/W for the CTO, and 0.69 A/W for the ITO MSMPDs. The Ti-Au, CTO and ITO MSMPDs had 3- dB cut-off frequencies of 14.0 GHz, 7.5 GHz, and 5.0 GHz, respectively, from time-domain measurements performed at 1.3 micrometers and 11.26 GHz, 4.00 GHz, and 2.61 GHz, respectively, from frequency-domain measurements performed at 1.55 micrometers . Discrepencies between the 3-dB cut-off frequency obtained from the time-domain and the frequency-domain measurements are attributed to the time-domain measurement system's inability to accurately resolve low frequency behavior (below 2 GHz) and space charge effects.

Arrays of high speed, high gain avalanche photodiodes (APDs) have been developed for use as high sensitivity optical photon detectors. The 1 mm² area APD pixels yield a maximum avalanche gain of 40,000 and a high signal-to-noise ratio with only moderate cooling (-22 degrees to -43 degrees C). These devices demonstrate 70% detection efficiency for 6 photon optical pulses and 35% detection efficiency for 3 photon optical pulses. The rise time is less than 2 nsec, and the fall time less than 7 nsec. Pixellating the PAD into a monolithic array will significantly reduce the cost per pixel compared to discrete devices. These devices will have great utility in various applications, ranging from high energy physics to biological instrumentation. The measured performance of these APD arrays as optical detectors will be discussed.

We report the first results obtained with a germanium quad-cell sensor operated in Geiger- mode regime. After a quantitative characterization of the single pixel, both in counting and in timing applications, we quantitatively assess the intensity of the optical coupling among the detectors of the cell due to secondary photon emission from hot carriers. This effect, intrinsically related to Geiger-mode operation, has been overcome by sequentially driving the pixels of the cell. A preliminary test demonstrates the tracking capabilites of the sensor. Since the single pixel can detect the arrival time of the photon with a precision better than 100ps FWHM arrays of such devices could be also employed in wavelength and timing resolved luminescence measurements in the near-infrared.

The task of analog few-photon pulse detection and measurement can be solved if two main conditions are carried out: the high quantum efficiency is realized and low noise mechanism of the internal amplification is used. Traditional facilities do not allow to perform both of these conditions in one device. So vacuum photomultiplier has high gain and low noise but its quantum efficiency is not rather high. Another wide spread device-avalanche photodiode having sufficiently high quantum efficiency displays very rapid rise of noise factor with gain increasing. In this report the ability of analog few-photon detection (AFPD) mode realizing by the new type of photodetectors--avalanche with negitive feedback (ANF) photodiodes is appreciated. As known single carrier avalanche process with negative feedback (SC ANF), when negative feedback governs the avalanche process initiated by one separate charge carrier, have the main features similar to those of 'ideal' amplifier: high gain, low noise, and fastness. Considering that the SC ANF photodetector as semi-conductor device has also high quantum efficiency it may be the adequate facility for AFPD in principle. In the report the properties of SC ANF process determined the characteristics and parameters of high sensitive analog photodetector such as gain, amplification fluctuations and efficiency, rise time are demonstrated and discussed. The analog detection of a several-photoelectron pulse is illustrated.

Seven years ago, at the Houston Congress of IMEKO, I tried to show that optics will play a key role in measurement technology and informatics. Already the relatively short time since the Houston Congress has shown that optics gained on importance in many fields of measurement technology. The better understanding of diffraction and how to use optics below the diffraction limit has opened a number of new applications pushing the micro- miniaturization to levels thought a few years ago as physically impossible. At present one of the most fascinating subjects is the exploration of human visual mechanism and the application of the results of this research in different areas of photonics. The human visual system codes the retinal image into a number of sub-images, depending on brightness, color, movement sensation, etc. In electronic image processing we are just starting to learn how such coding could be performed. To understand how the visual image is processed is of utmost importance also in order to be able to present the measurement results on the display. This paper discusses color coding and measurement--how this branch of photonic measurements can gain from vision research. Further, the progress of standardization and measurement in photometry and radiometry is discussed.

We have developed a system for measuring the nonlinearity of optical power meters or detectors over a dynamic range of more than 60 dB at telecommunications wavelengths. This system uses optical fiber components and is designed to accommodate common optical power meters and optical detectors. It is based on the triplet superposition method. The system also measures the range discontinuity between neighboring power ranges or scale settings of the optical power meter. We have developed an algorithm to treat both the nonlinearity and the range discontinuity in a logically consistent manner. Measurements with this system yield correction factors for powers in all ranges. The measurement system is capable of producing results which have standard deviations as low as 0.02%. With slight modification the system can operate over a 90 dB dynamic range at telecommunications wavelengths. This measurement system provides accurate determination of optical power meter or detector nonlinearity; the characterized detectors then can be used for such applications as absolute power and attenuation measurements.

The effective radiant exposures for artificial and natural UV-sources are determined by temporal integration over an 8 h working day. Therefore the spectrally weighted integration of the spectral irradiance from the radiation source in the plane of the exposure is to measure. Such measaurements are made with two different detector systems: measurements of UV radiation according to the integral method should be possible according to a quasi partial filtering method using different individually filtered photodiodes. A spectroradiometer for UV radiation analysis was tested due to its application in field measurements for meteorology, medicin, and occupational safety. The optical part of this compact instrument consists of a cosentrance optic, a monochromator and detector system. A comparison with commercial instruments is described.

LEDs are light- and near-infrared emitting diodes. Their line-like emission spectra causes considerable difficulties, not unlike phosphor emission of color CTRs, when accurate color measurement is required. LEDs exhibit many different spatial distribution characteristics as well, requiring relatively large area, uniform detector surface for photometric and colorimetric evaluation. Discussion of several diffferent type of colorimeters follows, showing trade-off between precision and price as well as suitability in their application for many different situations.

One of the most significant advances in light detection methods in recent years is the development of CCD technology. There is a great difference, however, in merely detecting light versus providing accurate photometric measurements of a field of view. Many factors must be taken into account in controlling CCD and camera characteristics, data acquisition, and subsequent data processing. Calibration techniques are particularly significant if we are to ensure proper photometric analysis.

A cryogenic bolometer has been developed for use as a transfer standard in an ambient temperature infrared detector response comparator facility at the National Institute of Standards and Technology (NIST). Issues affecting calibration of the bolometer response were studied and a calibration procedure was developed. These issues included frequency dependence, stability, repeatability, and spectral flatness. The relative spectral response was determined from the reflectance of the bolometer absorber and the transmittance of the bolometer window. The bolometer calibration has been tied to the NIST primary standard high accuracy cryogenic radiometer using a silicon trap detector and a pyroelectric detector.

This paper presents a comparison of two different methods for the realization of the SI unit of luminous intensity, the candela. A measurement system has been set up at IEN for comparing results obtained by a standard ECAR technique with those based on the work of Ikonen et al. on trap detectors. The paper highlights some of the problems encountered; preliminary results show that the agreement between the two methods is within the declared uncertainty and suggests promising areas for future works.

The theoretical model for the heat diffusion in the case of a high power IR electrically calibrated laser powermeter, developed at the Institute for Atomic Physics in Bucharest, is presented. The IR laser beam falls onto a laser detector, a special design copper disc wafer which absorbs the laser beam, heats its center. A daisy-chain of thermocouple elements having one set of junctions thermally connected to the central region of the disc and the other ones to the disc's boundary is used to detect temperature rise induced by the exposure to the laser beam. For calibration, the copper disc is electrically heated and the electric power that produces the same temperature rise as one induced by an incident laser beam, should equal the laser beam power. The electric heater is designed to provide a uniform heating of the copper disc. The solution for heat diffusion equation was searched as a series of Bessel functions of zero order, the cold junction's temperature was imposed as boundary condition and the heat induced by the laser beam in the disc's center was regarded as input data. To find the correct solutions, there must be taken into account the designing elements of the copper disc: termic material's properties (caloric capacity, termic conductibility), laser detector's geometry, copper's density. The electric power for calibration was injected using a precision power injection circuit which allows a stability of the calibration power, better than 0.1%.

At first, this paper gives a brief description of the cryogenic radiometer used as standard for radiometric measurements at the Institut National de Metrologie (INM). After, it describes the procedure used to calibrate transfer detectors, for the measurements in the field of optical fiber telecommunications. This absolute calibration was performed at the laser diode wavelength of 1550 nm. The first results for three different detectors calibrated by comparison to INM cryogenic radiometer for the above wavelength shows that the uncertainty of the calibration is less than 0.5% at 1 (sigma) level and is about twice as good as the uncertainty given by the former extrapolation method used at INM.

An anomalous performance has been found in studying the optical response of a QED-100 device against the reverse bias voltage applied to the device. After reaching a saturation value, the optical response suddenly decreases at a given voltage. Further increasing of the reverse bias raises the optical response again, but it never reaches the previous saturation value within the voltage interval studied in this work. The response drop disappeared after eliminating the electrical protection diode of the device, but the optical response still decreases for bias voltages larger than that needed to saturate the response. The decrease depends on wavelength and optical power.

This paper describes a noncontact and computer compatible test setup to calibrate the angular motion of a life test model of a scanning mirror mechanism using a lateral-effect position sensitive photodetector (LPSD). Laser based measurements of its movements using the LPSD adds up the suitability of this mechanism to very high resolution radiometer effectively. The description of the scan mirror mechanism and the existing calibration test methods are briefed. The characteristic properties of the LPSD are discussed. Experimental details and the results obtained are presented.

Light coupling from a light emitting diode to an optical system is discussed on the basis of the Hasegawa model. Experimental observations are presented to show the variations in power coupling at different planes in the image space and the limitations of the model.

A number of color absorption and luminescence indicators being dispersed within a porous matrix have been investigated by means of spectral techniques. As a support for the indicator molecules served a mesoporous silica glass of Vycor type (pore size of 7.5 nm) which was transparent to the light and permeable to the ambient gases and vapors. The selected immobilized indicators have revealed the well-defined spectral sensitivity to certain components of an atmosphere or other gas mixture. After a proper chemical and thermal treatment these indicators have gained high selectivity of spectral response on an individual gas appearing in the ambients, the observed spectral changes have demonstrated its reversibility. The indicators immobilized within the porous glass turned out to be mostly advantageous as applied in the colorimetric sensors and anlyzyers, sepcifically in the automated remote air monitoring systems. Some of the essential parameters of the developed portable individual gas analyzers and remote multicomponent monitoring systems with use of fiber optic guides are reported.

Testing has been conducted on 8 by 8 avalanche photo diode (APD) arrays derived from large area (16 mm) APDs, both produced by Advanced Photonics, Inc. The array structure was produced using a novel reverse etching process. Tests have been conducted measuring cross- talk, bandwidth, rise and fall times, gain, effective pixel size, and noise characteristics. Measurements have been made as functions of wavelength, optical intensity, and bias voltage. Cross-talk between pixels was characterized under both CW and pulsed (3 nsec) conditions. The effective pixel size was measured by scanning a very small laser spot (.25 mm) across the pixel under test while monitoring the output current. The measured pixel size was approximately 1 mm. This matched very well with the expected physical pixel size of 1 mm. The pulse response was measured by injecting a 3 nsec laser pulse into the pixel under test. The measured response shows that the signal decays approximately 3 orders of magnitude in 60 nsec. The rise time of the pixel is on the order of 5 nsec. Cross-talk between pixels was measured by injecting an optical signal into a pixel. The current output of an adjacent pixel was measrued as the optical power input was increased. The cross-talk a CW optical input is on the order of 1000 to 1. The pulsed cross-talk is on the order of 100 to 1. The cross talk ratio remains constant with varying optical input intensities. The pulsed wavelength response of the APD was characterized at 440 nm and 700 nm. The APD exhibited no difference between the two wavelengths.

While in the low power range, cryogenic electrical substitution (ESR) is a must when it comes to establish reference radiant power (flux) scales, for powers from 100 mW up, this technique is much less feasible. Therefore, we tried to elaborate a method and an appropriate radiometer design intended to minimize the inherent uncertainties associated with ambient temperature electrical substitution, i.e., those generated by the spatio-temporal nonequivalences. The paper describes the method and the solutions we thought might lead to the minimization of these uncertainties.