Presentation on theme: "Juan Estrada, Gustavo Cancelo"— Presentation transcript:
1 Juan Estrada, Gustavo Cancelo Research Techniques Seminar Advances in CCD Applications with Sub Electron Noise TechniquesJuan Estrada, Gustavo Cancelo07/19/2011
2 R&D Projects General CCD R&D will benefit the following projects: DAMIC.Neutrino coherent scattering experiment (νCCD).Neutron imaging.Spectroscopy for astronomy.CCD fast readout.
3 Recent papers1. arXiv: [pdf] Title: Deep sub electron noise readout in CCD systems using digital filtering techniquesAuthors: Gustavo Cancelo, Juan Estrada, Guillermo Fernandez Moroni, Ken Treptow, Ted Zmuda, Tom Diehl ), to be submitted to Experimental Astronomy.2. arXiv: [pdf, ps, other] Title: Achieving sub-electron readout noise in Skipper CCDsAuthors: Guillermo Fernandez Moroni, Juan Estrada, Eduardo E. Paolini, Gustavo Cancelo, Stephen E. Holland, H. Thomas Diehl ), to be submitted to Experimental Astronomy.3. arXiv: [pdf, other] Title: Direct Search for Low Mass Dark Matter Particles with CCDsAuthors: J. Barreto, H. Cease, H.T. Diehl, J. Estrada, B. Flaugher, N. Harrison, J. Jones, B. Kilminster, J. Molina, J. Smith, T. Shwarz, A. Sonnenschein4. arXiv: [pdf, other] Title: Plasma effect in Silicon Charge Coupled Devices (CCDs)Authors: Juan Estrada (Fermilab, USA), Jorge Molina (FIUNA, Paraguay), J. Blostein (CAB, Argentina), G. Fernandez (UNS, Argentina), submitted to NIM.5. arXiv: [pdf] Title: Direct Dark Matter Search using CCDsAuthors: Juan EstradaIn preparation:12 channel readout results.Full well problem in DECam CCDs resultsMany conference presentations.
4 CollaboratorsFermilab: J. Estrada, T. Diehl, H. Cease, D. Kubik, G. Derilo, K. Kuk, K. Schultz, A. ?, W. Struemer T. Shaw (PPD team) G. Cancelo, T. Zmuda, K. Treptow, N. Wilcer, J. Chramowicz, (CD team) Vic Scarpine (AD) B. Kilminster, J. Smith, T. Schwarz, A. Sonnenschein (DAMIC) J. Molina Facultad de Ingeniería, Universidad Nacional de Asunción, Asunción, Paraguay J. Blostein, others Centro Atómico Bariloche and Instituto Balseiro, Comisión Nacional de Energía Atómica, Universidad Nacional de Cuyo, Bariloche, Argentina G. Fernández, E. Paolini Universidad Nacional del Sur, Bahía Blanca, Argentina. Javier Castilla, CIEMAT, Madrid Spain. N. Harrison: Naperville North High school, now freshman at University of Chicago S. Wagner: Naperville North High school. J. Jones: Batavia High School, now at Purdue. Jacob Johansen University of Chicago Possible collaborators: Juan Carlos D’Olivo , Alexis Aguiler (UNAM, Mexico) – interest in reactor experiment with CCDs in Mexico Helio da Motta, Carla Bonifazi, Martin Makler (CBPF, UFRJ) interest in reactor experiment with CCDs in Angra. Requested a grant to CONICET, Argentina (Mayosky, Cancelo)
5 Outline Introduction to CCDs. Sub electron noise results. Fast readout results.Applications:Neutron imager.DAMIC.Neutrino coherent scattering.Spectroscopy for astronomy.
6 Charge Coupled Devices (CCD) Potential wellCharacteristics:Properly biased CCDs store charge in a potential well.Very low noise detectors => high dynamic range.1e- of noise RMS => 3.6eV ionization energy.High spatial resolution: 15 x 15 micron pitch for DeCam CCDs.High density: 8Mpix for DeCam CCDs.
7 Optical characteristics of LBNL CCDs used for DES DECam: High resistivity, 250µ thick. Fully depleted!Photon Transfer Curve (PTC)High QE in near infrared. Z>11g of mass, good for direct DM search.p-channel, better than n-channel for space telescopesPhoton Transfer Curve:Full wellReadout gainPixel and dark current non uniformitymore...Low resistivity CCDs
8 Dark Energy Camera (DECam) New wide field imager for the Blanco telescope (largest focal plane in the southern hemisphere)Largest CCD project at FNAL.DECam is being built at FNAL including CCD packaging, full characterization, readout electronics.CCD facilities at SiDet and 5+ years of experience positions FNAL as a leader for this task.Blanco 4m TelescopeCerro Tololo, ChileMechanical Interface of DECam Project to the BlancoCCDReadoutFiltersShutterHexapodOpticalLensesFocal plane with 74 CCDs (~600 Mpix).All the scientific detectors in hand, packaged and characterized at FNAL.
9 Low noise is critical in spectroscopy DECam estimates redshift from the colors of the objects. DeCam used 4 filtersDeSPEC spectrograph proposal:Lower signal to noise ratio.X000fibers4 DES filterscolors change as galaxy moves in zseveral spectrographs
10 Low noise CCD readoutA low noise CCD based readout system will greatly benefit projects such as:Direct search for Dark MatterNeutrino coherent scattering.Spectroscopy.Two Low noise reduction techniques:Reduction of Pixel to pixel correlated noise using fast sampling, precision A/D conversion and digital filtering.Skipper CCDs.
11 Outline Introduction to CCDs. Sub electron noise results. Fast readout results.Applications:Neutron imager.DAMIC.Neutrino coherent scattering.Spectroscopy for astronomy.
12 CCD Images Reset pulses are ~ 50,000 e- FITS image: HORIZONTAL REGISTERH1 H2 H3e-VERTICAL CLOCKSVIDEOAMPLIFIER3-PHASEARRAY CCDFITS image:Each pixel is a n-bit digital representation of the pixel charge.CsLoadResistorVideo Out.OutputJFETResetCGSe-VREFVDDSummingWellTransferGatePixelperiodchargePedestal
13 CCD noise: single video transistor and system noise Red trace: CCD noise measured by the LBNL designers using a test board.1/f noise larger than WGN up to 50 KHz.WGN about 10nV/√Hz.Black trace: FNAL 24 bit ADC based system.x3 lower noise than the Monsoon system used for DeCam (DES).Despite power supply and EMI noise reduction the system still shows some 60Hz and high frequency resonances.
14 Correlated Double Sampling (CDS) Integration intervals: t4-t3 = t2-t1 = Tx(n) = s(n) + n(n) + w(n)Video fragment: Npix pixels and Npix pedestals long.Pedestal iPixel iPedestal i+1Pixel i+1s(n) pedestals and pixelsn(n) correlated noise (LFN)w(n) white Gaussian noiseFor the white and Gaussian noise w ~ N(0,σ²), the CDS is the optimum estimator.butIt actually grows for longer T because the 1/f noise grows exponentially as f->0.
15 CDS transfer functionThe CDS filters very low frequency noise close to DC.Minimum noise rejection at f~0.4/TPix.Nulls at f=k/TPix, where k=1,2,3,…Better filtering for higher frequencies.Transfer function maximums follow a |sin(x)/x| decay.Tpix is a “free” parameter.In the analog CDS we adjust Tpix for the minimum noise where the 1/f contribution is small.But short Tpix limit WGN reduction.So far analog CDS techniques achieve ~2e- of noise at Tpix of ~20useconds.
16 Estimator and digital CDS Digital sample the video signal.Estimate the correlated noise of a string of pixels.Subtract the correlated noise from the original video.Perform the digital CDS of the filtered signal.χ2 estimator, because it does not assume a particular noise model:Inversion of a large matrix. Only one time and can be done off-line.Linear model is not orthogonal. Ill-posed problem.Goal:Implement the estimator and the digital CDS in an FPGA. Create FITS image.We can eliminate the pedestal and pixel values si from the estimation problem.where <x(n)> is the average signal+noise value in each pixel (step function)New linear model:where θ is a px1 vectory(n)x(n)
17 How many modes?Noise spectrum weighted by the CDS transfer functionCumulative noise spectrum weighted by the CDS transfer function200 modes account for ~85% of the low frequency correlated noise.If parameter estimation could be done with zero error.
18 Estimator and digital CDS Results (DeCam CCD) 0.5e- of noise achieved (consistently) for Tpix of 70useconds.
19 Estimator and digital CDS Results (LBNL 12 channel CCD) The plot displays the average noise of 100 data sets and 1-σ error bars.0.4e- at 120 μs.It is also interesting that the 1-σ error bars of the estimator processed data are 4 times smaller than the ones for the unprocessed data.
20 Noise spectrum comparison Compares the noise power spectrum of the unfiltered signal and the filtered signal after the low frequency estimation of 200 modes has been subtracted.On average, the LFC noise has been reduced by almost an order of magnitude on average.
21 FPGA implementation of the estimator and digital CDS X-ray image using a 55FE sourceCCD OverscanThe implementation of the estimator in the FPGA is on going.In this image the FPGA is performing the digital CDS with noise results very similar to the off-line results.
22 skipper CCD mounted on one of our testing dewars.
23 Skipper CCDthese type of detectors allow multiple readout of the charge in each pixel, with a noise reduction 1/sqrt(N). I was invented in 1990, but not really used much because it takes more time to readout.in a recent R&D run LBNL produced thick detectors like this, and sent them to us for testing. They are very interesting for low threshold experiments...
24 Skipper CCD Frequency response for: Tʃ = 110 µs and N = 5 (a), N = 10 (c) and N = 20 (d),T ʃ = 160 µs, and N = 10 (b).Left axis: CCD noise PSD (a).Standard CDS Tʃ = 110 µs (b)Skipper CDS Tʃ = 110 µs and N = 10 (c).
25 what have we done with the skipper: > built the cables to readout the skipper> programmed the controller to operate this new CCD> started tests> this is the main topic of the work that Jacob Johansen is doing as part of his graduate class at UCx-ray exposureCCD x-ray hits look like small horizontal bars in the image because the same pixel is readout many times.Skipper CCD e− RMS as a function of the number of averaged samplesN. Continuous line: e− RMS measured from images.Dashed line: theoretical 1/√N decay of WGN.
26 Low-energy X-ray detection experiment with the Skipper CCD. First time this peak is seen with a CCDFe55 X-ray source and a teﬂon targetA shielding is placed between the source and the CCD for stopping direct X-rays.The Fe55 X-ray source produces two diﬀerent energy rays from Mn: 5832 eV (K α ) and 6412 eV (K β ). Both X-rays hit the teﬂon with carbon and ﬂuorine, which emits lower energy X-ray by ﬂuorescence.Each atom has a precise energy pattern of emitting photons, the most probable emitted X-rays are at energies of 277 eV and 677 eV, respectively.These low-energy X-rays are detected by the CCD together with some high energy X-rayscoming directly from the source that get to cross the shielding.
27 Skipper paper: http://arxiv. org/PS_cache/arxiv/pdf/1106/1106. 1839v2 what is next with skipper:1. Optimization of readout for operation with extremely low readout noise2. We need to get more detectors packaged with low background package and start DM run at minos.
28 Summary and future work for CCD low noise readout The estimator and digital CDS reduce the CCD noise deep into the sub-electron region.New avenues for HEP experiments and telescopes are open:The price to pay for lower noise is a more sophisticated readout system.The estimator and digital CDS is being implemented in an FPGA with good success.
29 12 channel LBNL CCD6This CCD would be a good candidate for telescopes that require high pixel bandwidth such as LSST.
30 12 channel LBNL CCD12 channel CCD detector arrived on October 3 from LBNLstage 1: readout the CCD using the DECam electronics with minimal modifications . Show that we can do it and use results to motivate FNAL management to support this R&D effort. Done on Octber 13!stage 2: make necessary modifications to readout the 12 channels with DECam performance (250 kpix/sec and 8 e- of noise). Use the DECam 12 channel monsoon board and a new cable for the 12 channel CCD (with 12 JFETs and 12 preamps). Required design of preamp board and flex cable. Completed in January 2011.stage 3: works towards more aggressive performance thinking in LSST requirement. Study the high readout rate limitations of the devices. Readout rate could be increase by digitizing during the horizontal clocking. Started in mid-January. Javier Castilla, from CIEMAT (DECam collaborator) expert on monsoon system came to FNAL work on this for two months.FNAL R&D effort support this now.Javier Castilla will return to Fermilab in December using funds from his home institution.
31 stage 1: Installation of 12-channel CCD in Cube. Simple flat ribbon cable used to connect to the CCD. No JFET installed.
32 Results stage 1:first readout. 12 channels connected to the DECam 12 channel readout board without any conditioning for the signal. For DECam we use JFET next to the CCDs and preamps inside the dewar. We got 11 out of the 12 channels to work. Very noisy readout low gain.second readout. 2 channels connected to the DECam 8 channel prototype readout board with preamps outside the dewar. The readout worked and got ~20e of noise at 250 kpix/sec.
33 flatfield of 2 channel readout using prototype decam boards. nice cosmetics!
34 we obtain a gain of 0.47ADU/e. channel 1 (left)photon transfer:we obtain a gain of 0.47ADU/e.(typical gain for DECam CCDs with JFET and preamp is 0.8 ADU/e).fullwell at 89,361e-.(consistent with area reduction with respect to DECam pixels)
38 At slow readout the noise performance is comparable with DeCam CCDs
39 Outline Introduction to CCDs. Sub electron noise results. Fast readout results.Applications:Neutron imager.DAMIC.Neutrino coherent scattering.Spectroscopy for astronomy.
40 Neutron imager why neutrons? Potential SRF accelerator application? Key tool for Basic Energy Science.Hydrides may be a major player in Rres and Q0. (A. Romanenko, FNAL)
41 Example of a beam of collimated low energy (thermal) neutrons (NIST CNR facility). The image resolution is a function of many parameters, some of which are improved by the 10B-CCD detector.We need high quality beams to make full use of high resolution detectors.The plasma experiment (a slide coming next) shows that 1um resolution is achievable.
42 the CCDs would allow us to achieve micron level resolution state of the art10B-CCDthe CCDs would allow us to achieve micron level resolutionand at the same time 1 frame per second....
43 our scheme...CCDMotivated by conversations with neutron physics group in argentina during ICFA 10.Thanks to the R&D funding we were able to do the preliminary test at FNAL inviting Dr. Jeronimo Blostein.
44 Image taken with neutron 252Cf source, 5um film of 10Boron and CCD First test with Am-241 source. The CCDs are really nice alpha detectorsalphasbig circlesx-rayssmall dotsImage taken with neutron 252Cf source, 5um film of 10Boron and CCDalphasbig circles
48 Plasma effect measurement Our dataBest current measurement 2008Energy dependence of the cluster size for α particles in both regions of energy.Red points are for the exposition ofthe CCD to the 241Am source, while black points are for α particles coming from the (n,α) reactionPlasma eﬀect in Silicon Charge Coupled Devices (CCDs), J. Estrada, J. Molina, J. Blostein , G. Fernández.
49 2nd test, July 2011 (work by Sebastian Wagner, sophomore student at Naperville North high school) 2nd exposed surface1st exposed surfaceBottom half of open crossA factor of 4 better resolution
51 Next for the Neutron Imager: Current limitation is the neutron beam. Until now tests with CF-252 source and this week we are trying with Am-Be with x100 higher neutrons flux. (Thanks to the people in Site 40!)plans for new beam:starting in September we have been invited to try our detector in a neutron imaging facility in Argentina (Bariloche). One of our collaborators works in this facility.Vic Scarpine is trying to set us up with a real beam. He is doing some simulations to understand what we need in order to be able to use the beam at the neutron therapy facility at FNAL. Also considering other US facilities.plans for detector:the next step with the detector is to learn how to deposit the layer of Boron-10 directly on the surface of the CCD. This could get us micron resolution.
52 Outline Introduction to CCDs. Sub electron noise results. Fast readout results.Applications:Neutron imager.DAMIC.Neutrino coherent scattering.Spectroscopy for astronomy.
53 DAMIC (Dark Matter in CCDs) 7.2 eV noise ➪ low threshod (~0.036 keVee)250 μm thick ➪ reasonable mass (a few grams detector)muons, electrons and diffusion limited hits.nuclear recoils will produce diffusion limited hits
55 self shieldingcalibration to nuclear recoils still missing
56 thanks to our low threshold of 40eV we are competitive with 107 g-day exposure.
57 Next for DAMIC:We will continue running DAMIC at NuMI near detector for a few more months to understand the stability of the background.Move to SNOLAB:We are trying to move to SNOLAB in April Writing the proposal these days.We should be able to reduce our background by x10 if we eliminate the cosmogenic neutrons. At NuMI we have no muon veto.Plan to increase the mass by a factor of 8. By using a few thicker detectors. We have developed a new package that would allow us to use more CCDs in the same dewar.If things go well, we should be able to also reduce threshold by using either skipper CCD or digital filtering of 1/f.
58 Outline Introduction to CCDs. Sub electron noise results. Fast readout results.Applications:Neutron imager.DAMIC.Neutrino coherent scattering.Spectroscopy for astronomy.
59 Detection of Neutrino Coherent Scattering with CCDs Texono arXiv:
60 Two groups are interested on trying this at different reactors Two groups are interested on trying this at different reactors. One group is from Mexico (UNAM) and the other Group is from Brazil(Rio).Proposal to CNPQ in August (Helio da Motta, Martin Makler, Carla Bonifazi). The idea is to install something very similar to DAMIC with a skipper CCD.CD is going to get the equipment (vacuum vessel, cryocooler, pump).We hope to start measuring backgrounds in Rio this year.ν/fission/MeVmeasured antineutrino spectrum at ILLExpected recoil spectrum at 30mof 3GW reactor (Texono)
61 Alexis Aguilar-Arevalo UNAM Mexican group is starting to design a 100CCD system for a neutrino/dark matter experiment.Juan Carlos D’OlivoAlexis Aguilar-ArevaloUNAMWe will collaborate on this very ambitious project.They also want to have this ready for a large CCD DM search in the ANDES underground lab.mwe
62 Outline Introduction to CCDs. Sub electron noise results. Fast readout results.Applications:Neutron imager.DAMIC.Neutrino coherent scattering.Spectroscopy for astronomy.
63 What is the next detector technology We believe that the detectors that will replace the CCDs in astronomy are the MKIDS.Magnetic Kinetic Inductance Detectors:highly multiplexed readoutphoton counting UV-VIS-IR100 usec timing for each photonspectroscopy with R=400 should be possible (now R~20)pixel size ~100 umfirst implementation of a 32x32 pix array on sky in August (Ben Mazin UCSB)Ben would like to collaborate with FNAL in building the next prototype x10 largerfuture plan : giga-Z to follow up LSST objectthe only problem is that they work at mK temperatures
67 One good reason to look for low mass dark matter : The DAMA/LIBRA result Bernabei et al, 2008>8 σ detection of annual modulation consistent with the phase and period expected for a low mass dark matter particle (~7 GeV) consistent with recent COGENT results.
68 DAMIC experiment at FNAL Number of recoils exponentially increase al low energies.COGENT1 gram of SiDAMICDAMAw chan.CRESSTDAMAw/o chanLow noise limited:thanks to our low noisewe have the best result in the worldand we are reaching the DAMA regionMass limited:Need bigger detector.
69 Particle detection with CCDs muons, electrons and diffusion limited hits.nuclear recoils will produce diffusion limited hits
70 Neutrino coherent scattering Toward Coherent Neutrino Detection Using Low-Background Micropattern Gas Detectors P. S. Barbeau, J. I. Collar, J. Miyamoto, and I. Shipsey, IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 50, NO. 5, OCTOBER 2003 Abstract: The detection of low energy neutrinos (few tens of MeV) via coherent nuclear scattering remains a holy grail of sorts in neutrino physics. This uncontroversial mode of interaction is expected to profit from a sizeable increase in cross section proportional to neutron number squared in the target nucleus, an advantageous feature in view of the small probability of interaction via all other channels in this energy region. A coherent neutrino detector would open the door to many new applications, ranging from the study of fundamental neutrino properties to true “neutrino technology.” Unfortunately, present-day radiation detectors of sufficiently large mass (~1 kg) are not sensitive to sub keV nuclear recoils like those expected from this channel.☚ we are here ~ eVrecoil energy (keV)event rate ( cpd/keV/kg)lower noise in this case means that we can still see the recoils in a larger background.Expected rate for nuclear recoils ~30m from the center of a 3GW reactor. (From the Texono collaboration)☚ goal ~ 103 to to 5eVnoisereductionSM cross section is relatively large, the challenges are detector sensitivity and background control.SM and new physics.Help understand how to study supernovae neutrinos.Monitoring of nuclear reactors.
78 DECam has allowed us to build at FNAL a powerful CCD lab closely monitoring production of dies for more than 2 years, giving quick feedback on performancedeveloped CCD package for focal plane that meets scientific requirementsproduced/tested 240+ CCDs like an efficient factorydesigned and buildreadout electronicsfor a large focal planethe experience building silicon trackers transferred nicely to this project. The work in this talk has been possible thanks to this CCD lab.
79 Thin Al “window” for dewar Cf-252 source with moderatorto produce a lot of thermal muons
80 Next for the neutron imager: Detector:The 10B-CCD detector is being characterized with a high intensity source at FNAL site 40. We will need more detectors packed in this way or with some improvements that we will learn from this experiment.Beam:We are considering going to a good beam facility. Maybe a reactor, or a spallation source.We are working in simulations to design a moderator.Independently the group in Argentina is trying to get funds to buy get a system (CCD/dewar/electronics) of their own to do this in their experimental reactor.