How A Charge Coupled Device (CCD) Image Sensor Works

How a Charge Coupled Device(CCD) Image Sensor Works50 YEARS of the CCD

TELEDYNE IMAGINGHow a Charge Coupled Device(CCD) Image Sensor Worksthe area defined by one of the pixels willbe converted into one (or more) electronsand the number of electrons collected willbe directly proportional to the intensityof the scene at each pixel. When the CCDis clocked out, the number of electrons ineach pixel are measured and the scene canbe reconstructed.(Above) Teledyne e2vCCD47-20 backilluminated 1024 x 1024pixel 13.3 µm pixel size.2TELEDYNEIMAGING.COMWillard Boyle and George E.Smith invented the chargecoupled device (CCD) in 1969in the United States at AT&TBell Labs. In 1970, Boyle and Smith submitted a paper on their invention of the CCDto the Bell System Technical Journal. Theiroriginal ideas for the CCD was to createa memory device. However, with thepublication of Boyle and Smith’s research in1970, other scientists began experimentingwith the technology on a range of applications. Astronomers discovered that theycould produce high-resolution images ofdistant objects, because CCDs offered aphotosensitivity one hundred times greaterthan film.A charge coupled device is a highly sensitive photon detector. The CCD is dividedup into a large number of light-sensitivesmall areas (known as pixels) which can beused to build up an image of the scene ofinterest. A photon of light which falls within

HOW A CHARGE COUPLED DEVICE (CCD) IMAGE SENSOR WORKS3

TELEDYNE IMAGINGSection of a CCD Array0VElectrode or Gate0V10 VChannel-stopCHARGE COLLECTIONInsulator SiO2about 0.1 μm thick0VBuried Channel(n-type) about1 μm deepSubstrate (p-type)typically 600 μm thickStored ElectronsSubstrate Connection,SS 0V (ground)Section of a CCD Array with Electrodesand Channel ‘Columns’Incident photonsIncident photonsIncident photonsPolysiliconElectrodesSignal generation by the‘photoelectric effect’0VSubstrate connection (SS)TELEDYNEIMAGING.COMThe charge will collect in this region witha positive applied voltage. In practice, thecharge is stored in a buried channel regionto keep it away from contact with the surface and one image column is separatedfrom the next with a channel stop. Thestructure shown is a single CCD pixel.Large numbers of pixels are combined toform imaging devices, for example theTeledyne e2v CCD290 has 81 million pixels.The diagram below shows a 3 x 3 arrayof pixels.The electrodes or gates used to controlthe electrons within the CCD are madeof polysilicon rather than metal as this isreasonably transparent to incoming lightat wavelengths above about 400 nm andthe fact that all pixels are identical and areread though a single output port gives auniform image.READOUTChannel‘columns’4During integration, charge is collectedas clouds or buckets of electrons underbiased electrodes. At least two electrodesare required per pixel to control this chargecollection although typically four electrodesare used for scientific devices to optimisethe peak signal.For most of the CCD, the electrodes ineach pixel are arranged so that the chargeis transferred downwards along the columns.Hence, during the CCD clocking operation,rows are transferred downwards to the finalrow (the readout register) which is used totransfer the charge in each pixel out of theCCD so it can be measured.In the read out register, the electrodesare arranged so that the charge is transferred in the horizontal direction, along thereadout register.When charge has been collected and isto be read-out it is transferred one chargepacket at a time to an output amplifier

HOW A CHARGE COUPLED DEVICE (CCD) IMAGE SENSOR WORKSPrinciple of Charge Transferwhere the charge is converted to a voltage.During the readout operation, electrodesare alternately biased with high and lowvoltages to transfer charge down the array.A single transfer step is shown below. For a4 phase pixel 4 of these steps are requiredto transfer the charge one pixel down thearray.0V0V0VStep 10VAll of the electrodes in the image areaof a given phase are connected togetherso that only four clocks are needed totransfer the charge down the image areato a readout register. One row of chargeis transferred into the readout register at atime. The readout register is then clockedin the same way and read one pixel at atime at the output node by converting thecharge to a voltage. In the figure below a4 x 4 pixel three-phase device is shown:Because the entire array is read througha single amplifier the output can be highlyoptimised to give very low noise andextremely high dynamic range. CCDs canhave over 100dB dynamic range with lessthan 2e of noise.A CCD camera or instrument will consistof the CCD chip, and associated electronics,which is used at this point to amplify thesmall voltage on the CCD, remove noisecomponents, digitise the pixel values andoutput the values of each pixel for exampleto a processor. The CCD is an analoguedevice, and the analogue voltage valuesare converted into a digital form by thecamera electronics. 10 V 10 V 10 V0VStep 20V0V 10 V0VStep 3Readout RegisterIØ1IØ2IØ3Single pixelElectrodesImageSectionCharge-to-voltageOutput circuitBuried channelOutputBias etcRØ3RØ2RØ15

Teledyne Imaging’s Sensor CapabilitiesSpan the Complete Spectrum fromDetectorMaterialX-Rayto Very Long Wave InfraredCapabilitiesTeledyne Imaging’s sensor capabilities enable ourAerospace, Science and Defense partners to seefarther, from X-ray to Very Long Wave Infrared.0.001 nm0.1 nmHard X-Ray0.01 μmSoft X-Ray0.2 μmVacuum UV0.4 μmUV0.9 μmVisible2.5 μmSWIR5 μmMWIR10 μmLWIR100 μmVLWIRSiliconInAs/GaSb (T2SL)HgCdTe (MCT)Substrate-removed HgCdTe (MCT)InSbInGaAsGermaniumInAsPbSPbSeSiAsVOxWe are able to provide the righttechnology solution according tothe application using a range ofcompound semiconductor materialstaken from the imaging engineersperiodic table.Our image sensor technologiesrange from CCD to CMOS throughto hybrid infrared ROIC arrays andmicro-bolometers, and more.www.teledyneimaging.com

HOW A CHARGE COUPLED DEVICE (CCD) IMAGE SENSOR WORKSKey CCDParametersThe percentage of photons that are detectedis known as the Quantum Efficiency (QE)of a detector. The human eye has a QE ofapproximately 20%, photographic filmhas a QE of around 10%, and modern dayCCDs achieve a QE of over 90%. Quantumefficiency varies with wavelength and canbe extended across these wavelengthsthrough innovations such as backthinning,back-illumination, anti-reflective coatingsand high resistivity silicon.and the pixel is saturated. For a typicalscientific CCD this may occur at around150,000 electrons or so.The minimum signal that can be detectedis not necessarily one electron (corresponding to one photon at visible wavelengths).In fact, there is a minimum amount ofelectronic noise which is associated withthe physical structure of the CCD and isusually around 2 to 4 electrons for eachpixel. Thus, the minimum signal that canbe detected is determined by this readoutnoise. Single electron or Electron Multiplication CCDs (EMCCDs) are very low noisesensors designed to be highly sensitivedetectors where very few photons orelectrons are to be detected.WAVELENGTH RANGELINEARITYCCDs can have a wide wavelength rangeranging from about 0.1nm (soft x-ray) to400 nm (blue visible) through to about1000 nm (near infrared) with a peak sensitivity at around 700 nm. Detection in theshorter x-ray and ultraviolet wavelengthsis made possible through back illuminationand increased sensitivity in the longer nearinfrared wavelengths through lower noiseand high resistivity silicon.An important consideration in a detector isits ability to respond linearly to any imageit views. If the CCD detects 100 photons itwill convert these to 100 electrons (assuming100% QE). In such a situation, the detectorhas a linear response. A linear responseis useful as there is no need for additionalprocessing on the image to determine thereal or true intensity of for example differentobjects in an image.DYNAMIC RANGENOISEThe ability to view bright and faint sourcescorrectly in the same image is a very usefulproperty of a detector. The differencebetween a brightest possible source and thefaintest possible source that the detectorcan accurately see in the same image isknown as the dynamic range.When light falls onto a CCD the photonsare converted into electrons. The dynamicrange of a CCD is usually discussed in termsof the minimum and maximum numberof electrons that can be imaged. As morelight falls onto the CCD, more and moreelectrons are collected in a potential well.Eventually no more electrons can beaccommodated within the potential wellThere are a number of contributions to thenoise performance of a CCD.QUANTUM EFFICIENCYDARK CURRENTDark current is thermally generated noise.At room temperature, the noise performanceof a CCD can be thousands of electrons perpixel per second. In this situation, the full wellcapacity of each pixel will be reached in afew seconds and the CCD will be saturated.Dark current can be reduced by coolingthe detector with a system such as a Peltiercooler or even a cryo-cooler which canreduce the noise performance of the CCDto only tens of electrons per pixel persecond at -40 C.7

Credit: European Space Agency (ESA), XMM-NewtonTELEDYNE IMAGINGREADOUT NOISEPOWERThe readout noise originates from theconversion of the electrons in each pixelto a voltage on the CCD output node. Themagnitude of this noise depends on thesize of the output node. Advances havebeen made in reducing CCD readout noiseand continues to be an important part ofcurrent and future CCD development.Readout noise determines the dynamicrange and should be as low as possible,especially for detecting very faint energysources. For example detecting photons atx-ray energies such as in ESA’s XMM-Newtonspace observatory.CCDs themselves consume very little power.The key consideration is the electronicsrequired to operate the CCD and processthe images.www.teledyneimaging.comFor further information on CCDs pleasevisit www.teledyneimaging.com.Copyright 2020 Teledyne Imaging All Rights Reserved 20201022ESA’s XMM-Newtonspace observatory,celebrated 20 yearsin 2019.

A CCD camera or instrument will consist of the CCD chip, and associated electronics, which is used at this point to amplify the small voltage on the CCD, remove noise components, digitise the pixel values and output the values of each pixel for example to a processor. The CCD is an analogue device, and the analogue voltage values