High resistivity CCDs --------------------- E. Hoeg 1 Feb. 1999 SAG_CUO_57 ABSTRACT: A new type of CCD is being developed with high red response and other advantages. They would be interesting for the NBP Paschen observations and for the RVS, but probably not for the Astro instrument. The high resistivity CCDs are described on http://pdg.lbl.gov/~deg/ccd.html e.g. in the paper by Stover, Wei et al 1998 where Table.1 shows QE.Hi: lambda QE.Hi QE.MMS 300 - 0.48 320 .21 0.55 350 .32 0.67 400 .67 0.84 500 .86 0.83 900 .71 0.22 950 .60 0.13 1000 .50 - QE.MMS is from a recent message from MP (27 Jan. 1999). I give below some comments by my colleague Anton Norup Soerensen. Introduction ------------ Anton Norup Soerensen, CUO, writes to me on 29 Jan 1999. I had given him the report of 15 Dec 1998 from MMS (4 pages fax from MACP) on CCDs for GAIA. Anton notes that the most interesting he found in the report is the analysis of MTF versus level of illumination. He has never before seen numbers on that before. Anton then gives some facts on the Hi-Res CCDs: Resistivitet is 10-12KOhm/cm QE at 950nm 60% Pixel size 15um 300um substrat, fully depleted In spite of a rather small pixel to thickness ratio (1:20) the authors claim a good MTF. Further info ------------ Anton then gives some mails from the CCD-world mailing list on these devices: [Short version of SAG_CUO_57 is truncated here by EH, see full version on: http://www.astro.ku.dk/~erik/gaia] >From richard@ucolick.org Fri Jan 15 15:26:55 1999 Received: from niu.cfht.hawaii.edu (niu.cfht.hawaii.edu [128.171.80.58]) by ursa.astro.ku.dk (8.8.8/8.8.8) with ESMTP id PAA28623; Fri, 15 Jan 1999 15:26:49 +0100 (MET) Received: from uwila.cfht.hawaii.edu ([128.171.80.50]) by niu.cfht.hawaii.edu (Post.Office MTA v3.1.2 release (PO203-101c) ID# 0-44996U300L100S0) with ESMTP id AAA6378 for ; Thu, 14 Jan 1999 12:33:31 -1000 Received: from ucolick.org (lick.ucolick.org [128.114.23.139]) by uwila.cfht.hawaii.edu (8.8.8/8.8.8) with ESMTP id MAA23869 for ; Thu, 14 Jan 1999 12:33:30 -1000 (HST) Received: from gardiner.ucolick.org (gardiner2.ucolick.org [128.114.23.38]) by ucolick.org (lick) (8.9.0/LICK-GATEv8) id OAA15631; Thu, 14 Jan 1999 14:33:28 -0800 (PST) Received: by gardiner.ucolick.org (4.1/SMI-4.1) id AA00426; Thu, 14 Jan 99 14:33:27 PST Date: Thu, 14 Jan 99 14:33:27 PST From: richard@ucolick.org (Richard Stover) Subject: High resistivity CCDs To: ccd-world@cfht.hawaii.edu Message-Id: Reply-To: CCD-world@cfht.hawaii.edu Content-Length: 3252 Status: RO Posted to CCD-world: -+-+-+- I've received numerous responses to my last inquiry regarding wafer runs. Keep those cards and letters coming. Perhaps we should also have a discussion on the broadcast list regarding whether or not wafer runs are even a good idea. I'm not sure everyone understands what a wafer run is, what the limitations are, what sort of risks are involved, and what has to happen to minimize risk. There's probably little agreement on what constitutes success. But that's not why I write today. I want to guage interest in another possible development effort. UCO/Lick has been working with Lawrence Berkeley National Laboratory for a couple years on very high resistivity CCDs. We need to identify possible funding sources to keep this effort going. If you don't know about high resistivity CCDs you should check out the high resistivity home page: http://pdg.lbl.gov/~deg/ccd.html There are links to a number of papers and research notes on these devices. These devices promise some really spectacular performance because they are 300um thick and are sensitive throughout that entire volume. This means they have superior red response and NO FRINGING out to 1000nm. Thinning is not required, so the expensive and difficult thinning processing is eliminated. Holes, instead of electons, are the signal carriers so special backside treatments to enhance QE aren't needed. This means they have good blue response too. We just tested the first 2Kx2K high resistivity CCD. Serial CTE was indistinguishable from 1.0. The LBL lab director has been funding this CCD development activity from internal funds, but this cannot continue. I'd like to find out if there is interest in the astronomical community to help fund the continued development in this area. I think LBL needs something like $300K per year for the next couple years to make these CCDs generally available. Nobody can come up with this kind of money, but perhaps a group of us can join together to do it. The payoffs could be great for direct imagers and suitably designed spectrographs that need the best possible fringe-free red response. This is just a preliminary inquiry to assess interest. Take a look at the home page and let me know if you are interested in joining in on this effort. You should probably send me private email to avoid clogging the broadcast. No design is perfect and even these CCDs have their limitations. (Some are spelled out in the papers found at the home page.) But I can tell you without a doubt that I've had more fun working on these CCDs than anything else in years. Nobody makes anything else like these CCDs. They are amazing. Richard Stover -------------------------------------------------------------------------- Richard Stover email: richard@ucolick.org Detector Development Laboratory http://gardiner.ucolick.org/~ccdev UCO/Lick Observatory Voice: 831-459-2139 University of California FAX: 831-459-5244 Santa Cruz, CA 95064 USA FAX: 831-426-3115 (Alternate) -------------------------------------------------------------------------- -+-+-+- For information about CCD-world, send email to owner-CCD-world@cfht.hawaii.edu. >From richard@ucolick.org Thu Jan 21 09:57:10 1999 Received: from niu.cfht.hawaii.edu (niu.cfht.hawaii.edu [128.171.80.58]) by ursa.astro.ku.dk (8.8.8/8.8.8) with ESMTP id JAA02962; Thu, 21 Jan 1999 09:57:06 +0100 (MET) Received: from uwila.cfht.hawaii.edu ([128.171.80.50]) by niu.cfht.hawaii.edu (Post.Office MTA v3.1.2 release (PO203-101c) ID# 0-44996U300L100S0) with ESMTP id AAA5451 for ; Wed, 20 Jan 1999 08:51:59 -1000 Received: from ucolick.org (lick.ucolick.org [128.114.23.139]) by uwila.cfht.hawaii.edu (8.8.8/8.8.8) with ESMTP id IAA27894 for ; Wed, 20 Jan 1999 08:51:58 -1000 (HST) Received: from gardiner.ucolick.org (gardiner2.ucolick.org [128.114.23.38]) by ucolick.org (lick) (8.9.0/LICK-GATEv8) id KAA03308; Wed, 20 Jan 1999 10:51:55 -0800 (PST) Received: by gardiner.ucolick.org (4.1/SMI-4.1) id AA20868; Wed, 20 Jan 99 10:51:54 PST Date: Wed, 20 Jan 99 10:51:54 PST From: richard@ucolick.org (Richard Stover) Subject: RE: High resistivity CCDs To: ccd-world@cfht.hawaii.edu, Paul.Jorden@eev.com Message-Id: Reply-To: CCD-world@cfht.hawaii.edu Content-Length: 4226 Status: RO Posted to CCD-world: -+-+-+- Paul, >> I recognise the advantages, and would hope that EEV could make more such >> variants in the future. LBNL has been looking for someone to commercialize these devices. If no manufacturer in the U.S. has the wisdom to do it, perhaps EEV would be interested. They could contact the LBNL folks any time. >> I would be interested in seeing a good summary of >> the pros and cons of these devices- do you know of one? This is discussed a bit in the papers found at the LBNL web site. The cons people have worried about the most are CTE, charge diffusion, defocus at long wavelength, and cosmic rays. We've now shown that CTE can be excellent and charge diffusion is no worse than a standard thinned epi CCD. Our modeling suggests that defocus is not a problem, and we will demonstrate this when we put one of the CCDs in a spectrograph in about two months. Cosmic rays have 300um of silicon for depositing charge. So cosmic rays can be big and ugly. But they are easily recognized and don't really cover a lot of area. Pros are, in no particular order: 1. High QE from 350nm to 1000nm. 2. Stable QE wrt temperature and other environmental factors. 3. Spatially uniform response. 4. Easier to make - potentially cheaper than thinned epi CCDs if we can ever get production out of the research lab. 5. Large full well. 6. Virtually no fringing to 1000nm. 7. Mechanically rugged and naturally flat. 8. Excellent CTE. 9. Radiation hard (good for space applications) >> (a) direct imaging. High QE is especially useful for good S/N. With high >> resistivity, the depletion well would be expected to collapse more quickly >> as charge collects. Do you have any figures on full-well capacity? Full well depends on implant concentrations. These devices have a full well of about 300K e-. >> >> (b) Spectroscopy. An improvement in far-red throughput really helps. I >> guess most spectrographs have ~ f/3 cameras. Do you know how good a spatial >> resolution (ie spectral resolution) can be achieved? I still can't see how >> a 300 um thick device can be used with effectively a large far red depth of >> focus. It seems to me that in this application spectrographs are normally >> focussed with a line width of order one pixel (~15 um), and minimum >> broadening or scattering would be important. Modeling suggests that defocus will be about 1/2pixel (15 um pixels) in an f/1.5 beam at 1000nm wavelength. Silicon helps us here because index of refraction is greater than 3.5. So a fast beam outside the CCD becomes a much slower beam inside the silicon. Absorption depth at 1000nm is still less than 200um, so most of the light doesn't penetrate to this depth. As mentioned above, we're going to put a CCD into one of our spectrographs with an f1.7 camera. That spectrograph also has standard epi CCDs of both the thinned and unthinned variety, so we'll be able to make direct comparisons. An f/3 camera is no problem at all. >> >> Also, how do you find general quality of your deep depletion devices- eg >> cosmetics, dark current, blemishes or spots? I realise that it is early >> days yet, but any info. is valuable. Spots and other blemishs that arise from thinning variations or backside QE treatments are absent. We have demonstrated dark current as low as a standard epi CCD. This is discussed in one of the papers posted at the LBNL web site. We do tend to see a little charge pickup around the edge of the CCD. This appears as a bright edge. These are all early results. Things can only get better unless the program is allowed to die for lack of funds. Regards, Richard -------------------------------------------------------------------------- Richard Stover email: richard@ucolick.org Detector Development Laboratory http://gardiner.ucolick.org/~ccdev UCO/Lick Observatory Voice: 831-459-2139 University of California FAX: 831-459-5244 Santa Cruz, CA 95064 USA FAX: 831-426-3115 (Alternate) -------------------------------------------------------------------------- -+-+-+- For information about CCD-world, send email to owner-CCD-world@cfht.hawaii.edu.