CUBIC Calibration at the SRC

A lab version of the CUBIC instrument, affectionately known as "Baby CUBIC", has been calibrated at the Aladdin light source of the University of Wisconsin's Synchrotron Radiation Center (SRC). Baby CUBIC has a single CCD identical to the CUBIC flight CCDs, and uses flight spare electronics to provide as close an approximation to the flight camera as possible. A team including David Burrows, George Chartas, Steve Deiker (Univ. of Wis.), Ian Hutchinson (Leicester Univ.), Doug Kelly, Eugene Moskalenko, Kaori Nishikida, and Wilton Sanders (Univ. of Wis.) performed a series of calibrations of the energy resolution and quantum efficiency of the CCD in August 1996. (Five countries are represented by the team members: USA, UK, Japan, Australia, and Russia.)

This image shows the beam incident on our CCD from the SRC's Mark V Grasshopper monochromator, operating at an energy of 542 eV. The Mark V monochromator covers an energy range of about 100 - 1000 eV with resolution of 0.05 Angstroms, corresponding to 0.5 - 4 eV between 300 and 900 eV. We used the monochromator to provide narrow X-ray lines of known energy to illuminate the CCD. In order to achieve the low fluxes needed for an instrument designed for astronomical applications, the Aladdin beam had to be run in "micro-beam" mode, with the intensity about a factor of a millionth of its normal intensity. The image shown here is a false-color representation, in which the intensity of the light in a pixel is represented by its color. The outline of the circular aperture is clearly visible. The beam has bright bands produced by the monochromator optics and therefore does not uniformly illuminate the aperture. The square grid pattern is produced by absorption in the nickel mesh that supports our Al/Ti optical blocking filter.

This beam was used to calibrate the quantum efficiency of the CCD/filter combination against a NIST-calibrated XUV photodiode. A raw histogram of pixel values from a CCD exposure to a 530 eV beam in this mode is shown below.

This image shows the effects of multiple-photon events: the first peak (from left) is produced by pixels containing a single photon, the second peak is produced by pixels hit by two photons, the third peak is due to three photons per pixel, etc. These data allow us to count the total number of photons detected by the instrument in this over-exposed image. By comparing this with the incident beam flux, as determined by a well-calibrated photodiode, we can directly compute the quantum efficiency of the CCD at this energy. However, such a measurement is of little or no use in determining the spectral point spread function (energy resolution) of the instrument.

The beam intensity was reduced by an additional factor of 10 to permit energy resolution calibrations. At this intensity there are very few pixels containing charge from more than one photon, allowing us to accurately determine the amount of charge produced in the CCD by each photon. Some of the results of these calibrations are shown below.

The figure above shows the pulse-height spectrum measured by Baby CUBIC for an input line of 500 eV energy. The horizontal axis represents the energy channels of the instrument, which have a gain of 3.03 eV per channel (DN). The strong peak at channel 160 is the 500 eV line, while the weak peak at about channel 320 is a combination of 1000 eV photons in second order and some events which recorded two 500 eV photons in the same pixel. The energy resolution at 500 eV is about 57 eV.

The complete energy range of the instrument is shown in this spectrum of our Fe-55 calibration source. The source produces X-ray lines at 5.89 and 6.4 keV. These two lines are in channels 1920 and 2120, respectively. The energy resolution at 5.9 keV is about 135 eV (but depends somewhat on the exact details of the data processing). The small bump at about channel 1360 is an escape peak produced in the detector by the loss of a Si K-alpha photon, and has energy 4.11 keV. The peak at about channel 3850 is from bright pixels that saturated the 12 bit A/D converter, while the small peak at the left end of the plot is from the background noise level of the device.

CUBIC Home Page Ask Us