Why diamond for UV sensitive detectors
The performance of UV sensitive detectors
has steadily increased over the last decades in many respects, and astronomical applications benefit from this
evolution. These sensitive and highly linear imagers have made possible the success of recent solar missions such
as SOHO, YOHKOH, TRACE, and others. Nevertheless, CCD's designed for UV observations exhibit a few drawbacks
that are difficult to overcome within silicon technology :
Cooling reduces the dark current and prevents degradations from ionizing radiations, but it is a difficult and
expensive solution in space missions.
The low temperature of the detector transforms it into a cold trap for contaminants. The molecules stick
to the sensitive surface and may even polymerize under the UV flux, degrading the detector perforances
irreversibly.
The presence of ionizing radiations leads to images that are immediately covered by cosmics (bright points and
strikes) that are hard to separate from the UV signal. Besides, the subsequent degradation of the charge transfer
jeopardizes the mission life time.
The existence of an evolving oxide deteriorates the QE, its stability, and its spatial homogeneity, affecting
the calibration reliability.
The minimal size of the silicon UV pixel is limited to circa 10 µm.
The penetration depth of the photons in the silicon determines a pan-chromatic sensitivity which is
deleterious when observing a bright visible source like the Sun : one must add filters that absorb the undesired
photons, but also attenuate the expected ultraviolet signal.
These drawbacks become critical in the context of solar space missions where the highest spatial resolution,
temporal cadence and photometric accuracy are sought after. A significantly better spatial resolution can be achieved
either by going close to the Sun with a "standard" instrument, or by increasing appreciably the aperture and the focal
length of a telescope in an Earth orbit. In the first category (Solar Probe and Solar Orbiter missions), the whole
package (including the detector) is submitted to a high radiative and particle flux. In the second case, the increase
of spatial resolution happens inevitably at the expense of the signal level, especially if the temporal resolution is
to be matched with the smaller observables. For instance, a resolution of tens of kilometers on the Sun (better than
0.1 arcsec) implies exposure times smaller than 1 sec since expected velocities (e.g. Alfven) may be of the
order of 1000 km/s. Diamond imagers would help to circumvent many of the limitations listed above, leading to improved
performances.
The 5.5eV bandgap permits to operate the detector at room temperature, with no need for cooling and reduced
pollution risk. The bandgap also makes the diamond detector ''solar-blind'' (or visible-blind, ie. insensitive
to optical light. As the filters have a thermal role and because of the incomplete solar-blindness, they cannot be
suppressed, but their number can be reduced, thus improving the effective area of UV instruments (and allowing shorter
integration times, improved S/N, etc.).
The compact crystal network of diamond provides rad-hardness.
The absence of oxide can improve the QE and its stability.
The pixel is potentially in the sub-micrometer range, an order of magnitude smaller than achieved today.
Consequently, a diamond imager would open -just like the CCD in the past- new opportunities in the development of solar
telescopes and spectrometers of higher performances. They will be more cost-effective as well by sparing the cooling
hardware.CVD diamond has been identified as an excellent material for UV detection early. Many of the contributors
to BOLD not only have the appropriate testing equipment, but also find an interest in the aimed device. Other teams are
working in similar directions. Until now, the factor which prevented the development of such detectors has been the
material quality. Therefore, the chosen strategy has been to first address meaningful parameters on the best available
samples, using a well-proven technique with synchrotron XUV. This is why -although the a priori wavelength range
of interest is from 100 to 2200Å- the November 1999 measurements were done around the Carbon CK edge
(289 eV ~ 43Å) where the penetration depth of the photons spans from 100nm to 1 µm. A channeltron collected the
photo-electrons produced by the impinging synchrotron beam, in partial (PEY) or total (TEY) electron yield mode,
in order to separate bulk and surface properties. By varying the sample preparation steps (e.g. cleaning) and
the electrodes geometry, some insight has been gained into the physics of the signal collection.
Last update on 15th March 2000