Tight requirements on the Large Synoptic Survey Telescope point spread function (PSF) demand sensor contribution
to PSF be both small and well characterized. The sensor PSF is determined by the lateral charge
diffusion on the drift path from the photon conversion point to the gates. The maximum drift path occurs
for photons converted at the window, for blue optical photons in particular. Charges generated at the window
surface undergo "worst case" charge spreading and the blue optical PSF is used to characterize the sensor's PSF.
Different techniques for charge diffusion characterization have been developed, each with its own systematics
and measurement difficulties. A new way to measure charge diffusion using an X-ray source is presented. We
demonstrate the effectiveness and limitations of our technique and discuss relation of charge diffusion value
obtained with X-ray measurements to sensor PSF.
Future wide field astronomical surveys, like Large Synoptic Survey Telescope (LSST), require photometric precision
on the percent level. The accuracy of sensor calibration procedures should match these requirements. Pixel
size variations found in CCDs from different manufacturers are the source of systematic errors in the flat field
calibration procedure. To achieve the calibration accuracy required to meet the most demanding science goals
this effect should be taken into account.
The study of pixel area variations was performed for fully depleted, thick CCDs produced in a technology
study for LSST. These are n-channel, 100μm thick devices.
We find pixel size variations in both row and column directions. The size variation magnitude is smaller in
the row direction. In addition, diffusion is found to smooth out electron density variations. It is shown that the
characteristic diffusion width can be extracted from the flat field data.
Results on pixel area variations and diffusion, data features, analysis technique and modeling technique are
presented and discussed.
Knowledge of the point spread function (PSF) of the sensors to be used in the Large Synoptic Survey Telescope (LSST)
camera is essential for optimal extraction of subtle galaxy shape distortions caused by gravitational weak lensing. We
have developed a number of techniques for measuring the PSF of candidate CCD sensors to be used in the LSST camera,
each with its own strengths and weaknesses. The two main optical PSF measurement techniques that we use are the
direct Virtual Knife Edge (VKE) scan as developed by Karcher, et al.<sup>1</sup> and the indirect interference fringe method after
Andersen and Sorensen<sup>2</sup> that measures the modulation transfer function (MTF) directly. The PSF is derived from the
MTF by Fourier transform. Other non-optical PSF measurement techniques that we employ include <sup>55</sup>Fe x-ray cluster
image size measurements and statistical distribution analysis, and cosmic ray muon track size measurements, but are not
The VKE technique utilizes a diffraction-limited spot produced by a Point-Projection Microscope (PPM) that is scanned
across the sensor with sub-pixel resolution. This technique closely simulates the actual operating condition of the sensor
in the telescope with the source spot size having an f/# close to the actual telescope design value. The interference fringe
method uses a simple equal-optical-path Michelson-type interferometer with a single-mode fiber source that produces
interference fringes with 100% contrast over a wide spatial frequency range sufficient to measure the MTF of the sensor
directly. The merits of each measurement technique and results from the various measurement techniques on prototype
LSST sensors are presented and compared.
We present characterization methods and results on a number of new devices produced specifically to address LSST's
performance goals, including flatness, QE, PSF, dark current, read noise, CTE, cosmetics, and crosstalk. The results
indicate that commercially produced, thick n-channel over-depleted CCDs with excellent red response can achieve tight
PSF at moderate applied substrate bias with no evidence of persistent image artifacts. We will also report ongoing
studies of mosaic assembly techniques to achieve chip-to-chip
co-planarity, high fill factor, and thermal stability.
The LSST camera is a wide-field optical (0.35-1μm) imager designed to provide a 3.5 degree FOV with 0.2
arcsecond/pixel sampling. The detector format will be a circular mosaic providing approximately 3.2 Gigapixels per
image. The camera includes a filter mechanism and shuttering capability. It is positioned in the middle of the telescope
where cross-sectional area is constrained by optical vignetting and where heat dissipation must be controlled to limit
thermal gradients in the optical beam. The fast f/1.2 beam will require tight tolerances on the focal plane mechanical
assembly. The focal plane array operates at a temperature of approximately -100°C to achieve desired detector performance. The
focal plane array is contained within a cryostat which incorporates detector front-end electronics and thermal control.
The cryostat lens serves as an entrance window and vacuum seal for the cryostat. Similarly, the camera body lens serves
as an entrance window and gas seal for the camera housing, which is filled with a suitable gas to provide the operating
environment for the shutter and filter change mechanisms. The filter carousel accommodates 5 filters, each 75 cm in diameter, for rapid exchange without external intervention.
The LSST camera focal plane array will consist of individual Si sensor modules, each 42×42mm<sup>2</sup> in size, that are
assembled into 3×3 "raft" structures, which are then assembled into the final focal plane array. It is our responsibility at
Brookhaven National Lab (BNL) to insure that the individual sensors provided by the manufacturer meet the flatness
requirement of 5 μm PV and that the assembled raft structure be within the 6.5 μm PV flatness tolerance. These
tolerances must be measured with the detectors operating in a cryogenic environment at -100C in a face-down
configuration. Conventional interferometric techniques for flatness testing are inadequate to insure that edge
discontinuities between detector elements are within the tolerances because of the quarter-wave phase ambiguity
problem. For this reason we have chosen a combination of metrology techniques to solve the discontinuity ambiguity
problem that include both a full aperture interferometer and a scanning confocal distance microscope. We will discuss
concepts for performing flatness metrology testing with these instruments under these conditions and will present
preliminary results of measurement sensitivity and repeatability from tests performed on step height artifacts.
The LSST camera is a wide-field optical (0.35-1um) imager designed to provide a 3.5 degree FOV with better than 0.2 arcsecond sampling. The detector format will be a circular mosaic providing approximately 3.2 Gigapixels per image. The camera includes a filter mechanism and, shuttering capability. It is positioned in the middle of the telescope where cross-sectional area is constrained by optical vignetting and heat dissipation must be controlled to limit thermal gradients in the optical beam. The fast, f/1.2 beam will require tight tolerances on the focal plane mechanical assembly.
The focal plane array operates at a temperature of approximately -100°C to achieve desired detector performance. The focal plane array is contained within an evacuated cryostat, which incorporates detector front-end electronics and thermal control. The cryostat lens serves as an entrance window and vacuum seal for the cryostat. Similarly, the camera body lens serves as an entrance window and gas seal for the camera housing, which is filled with a suitable gas to provide the operating environment for the shutter and filter change mechanisms. The filter carousel can accommodate 5 filters, each 75 cm in diameter, for rapid exchange without external intervention.
The LSST project has embarked on an aggressive new program to develop the next generation of silicon imagers for the visible and near-IR spectral regions. In order to better understand and solve some of the technology issues prior to development and mass-production for the huge LSST focal plane, a number of contracts have been written to imager firms to explore specific areas of technology uncertainty. We expect that these study contracts will do much toward reducing risk and uncertainty going into the next phase of development, the prototype production of the final large LSST imager.
Sensors for the LSST camera require high quantum efficiency (QE) extending into the near-infrared. A relatively large thickness of silicon is needed to achieve this extended red response. However, thick sensors degrade the point spread function (PSF) due to diffusion and to the divergence of the fast f/1.25 beam. In this study we examine the tradeoff of QE and PSF as a function of thickness, wavelength, temperature, and applied electric field for fully-depleted sensors. In addition we show that for weakly absorbed long-wavelength light, optimum focus is achieved when the beam waist is positioned slightly <i>inside</i> the silicon.
Advances in neutron scattering studies will be given a large boost with the advent of new spallation and reactor sources at present under consideration or construction. An important element for future experiments is a commensurate improvement in neutron detection techniques. At Brookhaven, a development program is under way for greatly increasing the angular coverage, rate capability and resolution of detectors for scattering studies. For example, a curved detector with angular coverage of 120° by 15° has recently been developed for protein crystallography at a spallation source. Based on neutron detection using <sup>3</sup>He, the detector has the following major, new attributes: eight identical proportional wire segments operating in parallel, a single gas volume with seamless readout at segment boundaries, parallax errors eliminated in the horizontal plane by the detector's appropriate radius of curvature, high-throughput front-end electronics, position decoding based on high performance digital signal processing. The detector has a global rate capability greater than 1 million per second, position resolution less than 1.5 mm FWHM, timing resolution about 1 μs, efficiency of 50% and 90% at 1Å and 4 Å respectively, and an active area 1.5 m x 20 cm.
PN-CCDs are being developed as focal plane detectors for ESA's X-ray Multi-Mirror satellite mission (XMM), to be launched at the end of this century. As a part of the European Photon Imaging Camera (EPIC) the pn-CCDs will convert the incoming X-ray radiation with high quantum efficiency, low readout noise, excellent background rejection, timing in the microsec regime, radiation tolerance up to several hundreds of krads and a position resolution tailored according to the angular resolution of the telescope. The goal of our laboratorial efforts for this mission is to fabricate a monolithic pn-CCD of an active area of 6 x 6 sq cm having 768 on-chip JFET amplifiers located at the end of each CCD line. It is the aim of this contribution to report on the ongoing work of the pn-CCD system. This article focuses on the position resolution capabilities of fully depleted pn-CCDs, some recent results in the noise analysis and preliminary results on 10 MeV proton damage.
Recent results on the on-chip electronics, transfer properties, and radiation entrance window of pn-CCDs are presented. With recently fabricated devices, an improved charge transfer efficiency per pixel of 0.9995 and an energy resolution of the CCD output stage of 5 e(-) rms have been measured. This performance is achieved without a degradation of other characteristics of the devices, such as an X-ray efficiency of 90 percent at 10 keV, more than a factor of 1000 better time resolution in the full frame mode in comparison with all other CCD concepts, and a one-dimensional spatial resolution of 24 microsec in the timing mode. The use of pn-junctions instead of MOS structures makes the devices intrinsically radiation resistant.