Figure 4(a). Diagram of a typical detector spectral comparator instrument employed at NIST to characterize optical radiation detectors and detector systems. The major components of the instrument include appropriate light sources, a high quality monochromator, entrance and exit optics to convey the light to its intended points, and a mechanical arrangement such as translation stages to position the TSD or working standard detector and test devices in the beam.(b) Typical characterization results of a selection of photodetectors. The vertical scale is absolute responsivity in units of A/W and the horizontal scale is wavelength in nm. Other types of photodetectors are used in the UV, Far UV and the IR.
The typical result of DSC measurements are curves such as shown in Fig. 4(b). These are representative plots of absolute responsivity in units of amperes/watt for several of the solid state detectors in widespread use. The data shown in Fig. 4(b) was taken with the UV and visible to near IR instruments at NIST. The chain of calibration has resulted in electrical power measurements in the HACR being transferred by a series of steps to determine the electrical response of a solid state device to optical power. As a result, the response of the photodetector is determined in terms of the electrical watt maintained by NIST in the form of voltage, current, and resistance standards [31]. Some of the NIST DSCs have a stage for translation of the detector vertically in the plane perpendicular to the incident light. This feature is labeled x/y scanning carriage in Fig. 4(a). The spot size of the incident beam is usually on the order of a millimeter but can be varied for other purposes by the focusing optics. The small size allows measurement of the spatial uniformity of a detector system which is important in circumstances where the detector system will be employed in an underfilled mode of operation and hence possibly show a sensitivity to the incident beam position on the detector. Additionally, the spatial uniformity may be a function of wavelength in certain semiconductor devices, and if left uncharacterized can cause unwanted irregularity in system calibrations. NIST routinely supplies this information to calibration customers requiring spatial performance characteristics for detectors.
A DSC is a versatile optical instrument which can be used for spectral
transmittance measurements and other characterizations requiring a known
optical beam. Important to the establishment of detector based optical
units is the characterization of a photodetector coupled with an optical
filter restricting the wavelength interval of transmittance. Such a device
is often called a Filter Radiometer (FR). An FR can range from a simple
silicon photodiode with a colored glass broadband filter to a sophisticated
temperature controlled device designed for precision measurements. An
example of the latter type of device is shown in Fig. 5
[32]. The FR system consists of a
precision aperture to define the amount of light entering the system, a
thermoelectric (TE) temperature controlled filter, and appropriate absorbing
glasses to protect the system against unwanted radiation and environmental
effects. (Depending upon use, the precision aperture may or may not be the
limiting aperture in the optical system.) After propagating through the beam
definition optics the radiation is collected by an appropriate detector. The
device shown in Fig. 5 is modular and can support a wide range of filters
and detectors by a simple mechanical interchange. Where necessary the detector
module can be temperature controlled. Devices like this require considerable
care in design and assembly to provide elimination of unwanted reflections and
scatter of the radiation entering in the system.
Figure 5. Cross section diagram of a typical filter radiometer system consisting of a precision aperture, filter, and sensor system. The NIST design is modular to provide the opportunity to use various filters and detector systems with some economy of components.
The absolute spectral responsivity of the resultant sensor system can be calibrated using the DSC facility. Figure 6 shows some typical examples of FRs calibrated at NIST for use in photometry and for implementing a detector based spectral radiance and irradiance unit. FR#2 with a peak response around 550 nm is a photometer and the 380 nm and 910 nm systems were designed for radiance temperature measurement research. FRs can be sensitive to beam divergence and polarization as well as exhibiting interference anomalies if a coherent radiation source is used. Consequently considerable care, calibration, and study must be directed toward the use of a FR in radiometry and photometry. Depending upon the light source to be measured by the FR, the out of band rejection of the filter system must be characterized in detail and with sufficient accuracy to provide the measurement accuracy desired. As an example, if the system shown as FR#2 in Fig. 6 has a poor near IR rejection where a silicon detector has good sensitivity, large errors can result when used with an incandescent source which has large near IR output.
Figure 6. Examples of the calibration of filter detector systems in the DSC instrument. The vertical scale is absolute responsivity in A/W and the horizontal scale is wavelength in nm. These examples represent a UV filter, FR#1, a photopically corrected filter radiometer, FR#2, and a near IR system, FR#6.
A calibration scheme based upon the HACR has allowed the accuracy of DSC performed calibrations to be significantly improved over that reported in Ref. 30 when the instrument was first constructed [25]. The relative combined standard uncertainty for the detector response measurements now is in the 0.1% range in the visible (400 nm to 900 nm) using silicon detectors and 0.2% to 1.3% in the near IR for germanium devices. In the UV region (200 nm to 400 nm), the relative expanded uncertainty is 1.75% to 0.1% with the higher uncertainty in the shorter wavelength region. As our understanding of the physics of these devices improves, the uncertainties associated with interpolating the spectral response between the laser wavelengths will continue to diminish with a corresponding decrease in the uncertainty of calibrated detectors that NIST provides.*
An alternate method of calibrating an FR was developed by Schaefer and his collaborators at NIST [33]. This technique utilized a dye laser system which featured wavelength tunability and provided the capability of scanning over the wavelength range of the filter's transmittance. The laser power could be directly determined with a calibrated detector and the absolute response of the FR inferred. It was found that care had to be exercised to account for artifacts of the measurement introduced by the coherent properties of the laser radiation. While the coherence aspect adds an extra dimension to the practical utilization of the FR calibrated using a dye laser system, it is a manageable matter and this technique offers the possibility of high accuracy direct calibration of FR systems based upon a cryogenic radiometer. Schaefer and his colleagues reported a relative combined standard uncertainty of 0.18% in their system using an absolute silicon detector system for the reference.
