Current progress of Lamp / CCD calibration

Abstract

We have begun an effort to characterize the interpixel variation of HgCdTe Infrared devices. To ramp up for such a measurement, we begin with a simple experiment to measure the color of calibrated lamp using a Si photodiode. The photodiode was purchased from Hamamatsu and has QE calibrated to 5% accuracy. In order to measure the color of the lamp, we use narrow band filters and sample the spectrum of the near-blackbody lamp. We ratio these measurements to our measurement at 500 nm and then compare this ratio to that of a blackbody spectrum.
With the color measurement in hand, we will then check the absolute calibration of a NIST calibrated lamp. Once the calibration is verified, we will be able to calibrate other lamps and detectors in house. The calibration and experience gained from this study will be enable us to begin QE measurements of the NIR devices for SNAP.

Figures

Figure 1 Irradiance ratios before and after the bias correction. The pink line demostrates how the bias affects the low detected signal at shorter wavelength. The blue line shows that a simple subtraction of the bias helps. Corrections for filter bandpasses and detector spectral response have also been made.

Figure 2: Irradiance ratios using 1mm, 3mm, and 5mm pinholes after the bias correction. A large linear systematic is found with errors up to 30%

Figure 3: The irradiance output of the calibration lamp. With a blackbody peak of 900 nm, I find the lamp temperature is about 3220 K. I then use the best fit temperature and scattered light components from minimum chi^2 and calculate the ratios relative to 500 nm. I use these calculated ratios instead of the given lamp values.

Equation 1: To fit the ratios, I use the following formula for an unnormalized chi^2 fit. B_lambda(T) is the blackbody function where T is a fit parameter for temperature, "C" is a correction for the pinhole aperture, "a" is a fit parameter for white scattered light, and "b" is a fit parameter for wavelength-dependent scattered light.

Figure 4: The irradiance ratios for the all pinholes with only the blackbody temperature correction. Error bars are constructed by adding in quadrature the repeatability error at a given wavelength to the 5% detector calibration error.

Figure 5: The irradiance ratios for the all pinholes with a blackbody temperature correction and scattered light correction. The fit value for the scattered white light term 'a' is negative which is an unphysical value. No benefit comes from adding the scattered light terms to the correction

Figure 6: The irradiance ratios for the all pinholes with a blackbody temperature correction for a new, uncalibrated lamp. Notice that large differences between pinholes remain in spite of the new lamp. This indicates the real problem is with our setup, not the lamp.

Figure 7: Recently measured irradiance ratios of the uncalibrated lamp using an integrating sphere. Here the fitted blackbody temperature is 3126 K, which is much closer to the expected lamp temperature. The systematic dependence on pinhole size has also subsided. Unfortunately, ratios at wavelengths shortward of 500nm have large systematic errors due to the low light levels and the bias subtraction.

Conclusions

We have made color measurements in the visible regime of a near-blackbody lamp. Using an integrating sphere, these irradiance ratios are consistent with the lamp's irradiance values and our detector's QE calibration except at the lowest light levels.

Measurements are sensitive to the angle of incident light on the interference filters and the internal system reflections. However, we have designed our setup such that scattered light does not seem to be a major factor at the 5% level.

The corrected ratios still show a persistent structure. We think this is due to the QE calibration of the detector or imperfections on filter surfaces.

Future

Measure the absolute irradiance at a single wavelength and use the color information to extrapolate to other wavelengths. Validate with a NIST calibrated lamp.

Calibrate other QTH lamps and Si photodiode.

Repeat the color and absolute irradiance measurements with a calibrated InGaAs device and a local dewar.

Design and build a Lambertian source setup using lamps and detectors calibrated from the visible to NIR.

Test Lambertian setup with QE measurements of an off-the-shelf CCD camera

Provide and install calibrated diodes in Michigan NIR dewar system. Make QE measurements of HgCdTe devices