PHYS362 - PowerPoint PPT Presentation


PPT – PHYS362 PowerPoint presentation | free to download - id: 1fe75e-ZDc1Z


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation



PHYS362 Advanced Observational Astronomy – PowerPoint PPT presentation

Number of Views:53
Avg rating:3.0/5.0
Slides: 137
Provided by: car7158
Learn more at:
Tags: phys362 | red | tu


Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: PHYS362

PHYS362 Advanced Observational Astronomy
  • Professor David Carter

Slide set 4
Electronic Imaging Text Books
  • Astrophysical techniques C.R. Kitchen IOP
    Publishing. ISBN 0 7503 0498 7.
  • Electronic Imaging in Astronomy I.S. McLean
    Wiley-Praxis. ISBN 0 471 96972 9.

Definitions for photon sensitive detectors
  • Responsive Quantum Efficiency (RQE) Ratio of
    the number of detected photons to the number of
    incident photons.
  • Detective Quantum Efficiency Square of the
    ratio of the output signal-to-noise ratio to the
    input signal-to-noise ratio
  • DQE is always less than or equal to the RQE
    because of the effect of noise.

Charge Coupled Device (CCD) detectors
  • Ubiquitous in optical and very near infra-red
  • The detector of choice at wavelengths from 400 nm
    to 1µm.
  • Available comparatively cheaply in very large
    pixel formats (a 2048 x 4096 array will cost
    around 40,000).
  • Robust enough to fly on space missions.

Principles of the CCD
  • In a solid crystal electrons in the uppermost
    orbits of atoms, the valence electrons, interact
    to bind the atoms together.
  • The energy levels are shared between atoms, are
    split by small amounts. With a crystal with
    millions of atoms, the levels are spread out into
    a band.
  • Higher levels are split also

Principles of the CCD
  • The lowest band is filled with electrons because
    there is one electron for each atom. This is
    called the valence band.
  • The next level up is empty, and is called the
    conduction band. If electrons do get into this
    they are free to move between atoms.
  • Between the valence band and the conduction band
    is a forbidden energy gap or band gap.

Principles of the CCD
Principles of the CCD
Principles of the CCD
  • In a conductor the valence band and conduction
    bands overlap, allowing electrons to move freely
    between the two.
  • In an insulator there is a large forbidden energy
    gap, so electrons cant get into the conduction
    band to conduct electricity
  • In a semiconductor there is a small forbidden
    energy gap, so some electrons can move into the
    conduction bands.
  • If an electron moves into the conduction band,
    because of thermal effects or interaction with a
    photon, it leaves a hole in the valence band, so
    an electron-hole pair is created

Part of the periodic table. Semiconductors are in
column IV of this table (e.g. Silicon, Germanium)
or compounds of III-V (e.g. Indium Antimonide) or
occasionally II-VI (e.g. Mercury Cadmium
Telluride) elements.
Semiconductor band gap
Band gap EG and corresponding critical wavelength
?c for various semiconductors. The silicon band
gap is less than then energy of optical photons.
InSb and HgCdTe are used in infra-red
applications. Band gap is temperature dependent.
The Charge Coupled Device
  • A CCD is a silicon wafer which is exposed to
  • The picture elements (pixels) of the CCD are
    defined by the electrode structure which is
    applied to this wafer.

Principles of the CCD
  • A series of electrodes is then placed above the
    surface of the silicon, separated from the
    silicon by a thin silicon dioxide insulating
    layer. A positive voltage is applied to the
    electrode , which then attracts and stores the
    electrons generated beneath this electrode

Principles of the CCD
  • As you expose the CCD to radiation, electron-
    hole pairs are generated and the electrons build
    up in the electron storage areas immediately
    below the positive potential electrodes.
  • After a while (seconds to minutes) the CCD
    contains an electrostatic representation of the
    pattern of incident radiation on it.
  • This somehow must be read out and stored in
    digital form.

Principles of the CCD
  • CCD is made out of p-doped silicon. This has
    excess holes compared with pairs. Holes are
    driven away from the positive potential, so a
    depletion layer is formed under the electrode.
    This is the collection phase. Regions away from
    the electrode or under electrodes at lower
    positive potential have excess holes, and form an
    insulating layer. This is the barrier phase

CCD readout
3 phase CCD design. Each picture element (pixel)
has 3 electrodes. During exposure the centre of
the three is held at positive voltage (about 10
V) and the charge builds up under this. Then the
shutter is closed, and the voltage on the
electrode to the right ramps up to 10V as that
on thecentre one ramps down. The charge shifts to
under the right electrode. Repeat twice more, and
the charge has moved one whole pixel.
CCD readout
CCD readout
  • CCD is a two dimensional device. After the
    shutter is closed, the entire image is shifted
    down by one pixel, the bottom pixel is shifted
    into a horizontal readout register.
  • The readout register is shifted to the right by
    one pixel, and the pixel at the bottom right is
    shifted into a readout capacitor.
  • The charge causes an instantaneous change VQ/C
    in the voltage of the input line of the on-chip
    transistor, which in turn causes a change in
    voltage on the input line
  • The voltage change is then amplified and
  • The storage capacitor is then reset.

CCD readout
  • This process is then repeated until each pixel in
    the readout register has been digitised.
  • The image is then shifted down vertically by one
    more pixel, the next row is shifted into the
    readout register, and this is digitised in the
    same way.
  • The whole process is repeated until the entire
    image is read out.
  • For a 2048 x 2048 pixel CCD it takes
    approximately 10 seconds to read out the whole
  • A CCD read out this way is a line transfer CCD.

CCD readout
  • The channel stops between columns are permanent
    as charge does not move horizontally except in
    the readout register.
  • These channel stops are biased to negative
    potential by doping, hence charge cannot leak

CCD output circuit
Buried channel CCD
  • CCDs described before are surface channel CCDs,
    but in these the charge is being shifted along in
    a thin layer just below the oxide insulator.
  • Surface layer has crystal irregularities which
    can trap charge, causing loss of charge and image
  • If there is a layer of n-doped silicon above the
    p-doped layer, and a voltage bias is applied
    between the layers, the storage region will be
    deep within the depletion region
  • This is called a buried-channel CCD, and suffers
    much less from charge trapping.

Buried channel CCD
A single pixel in a buried channel CCD
Buried channel CCD
Thinned back-illuminated CCD
  • As described to now, the CCDs are illuminated
    through the electrodes. Electrodes are
    semi-transparent, but some losses occur, and they
    are non-uniform losses, so the sensitivity will
    vary within one pixel.
  • Solution is to thin the CCD, either by mechanical
    machining or chemical etching, to about 10µm, and
    mount it the other way up, so the light reaches
    it from the back.

Thinned back-illuminated CCD
Thinned back illuminated (left) and thick virtual
phase (right) CCDs
Frame transfer CCDs
  • Instead of reading the CCD out line by line as
    described before, a Frame transfer CCD has half
    of its area masked off to stop light reaching it.
    On readout, the whole CCD is clocked vertically
    so that the image area is transferred to the
    storage area.
  • The image can then be read out from this storage
    area whilst the image area is being exposed

Dark Current
  • Dark current is generated when thermal effects
    cause an electron to move from the valence band
    to the conduction band.
  • The majority of dark current is created near the
    interface between the Si and the SiO2, where
    interface states at energy between the valence
    and conduction bands act as a stepping stone for
  • CCDs are operated at temperatures of around 140K,
    to reduce thermal effects.

Dark Current
  • Multi-Phase Pinned (MPP) CCDs are doped with
    boron to allow the gate potentials to be positive
    with respect to the substrate, which causes holes
    to migrate to the surface area where they fill up
    these interface states.
  • This has the effect of reducing dark current, and
    MPP CCDs can be run at much higher temperatures
    than non-MPP CCDs.
  • Dark current at 140K is typically 10-4
    electrons/s/pixel, i.e. negligible.

Linearity and Saturation
  • CCD pixels have a linear response of measured
    output voltage to a value quite close to the full
    well capacity of the pixel. The number of
    electrons which can be stored is given by
  • Q CV/e
  • V is the voltage, and C is the capacitance of
    the pixel, given approximately by
  • C ? A?e0/d
  • A is the area of the pixel, d is the thickness
    of the SiO2 layer, ? is the dielectric constant
    of SiO2 (about 4.5) and e0 is the permittivity of
    free space.

Linearity and Saturation
  • Typically the full well capacity of a CCD pixel
    25 µm square is 500,000 electrons. If the charge
    in the well exceeds about 80 of this value the
    response will be non-linear. If it exceeds this
    value charge will spread through the barrier
    phase to surrounding pixels.
  • This charge bleeding occurs mainly vertically, as
    there is little horizontal bleeding because of
    the permanent doped channel stops.
  • Readout register pixels are larger, so there is
    less saturation effect in the readout register.

Cosmic rays, X-rays, and particle radiation
  • There are a number of types of radiation which
    can interact with the silicon to produce several
    tens of electron-hole pairs in a cluster, which
    appears as a bright spot (if the radiation is
    normal to the detector) or a streak if it is
    steeply inclined. These radiation events are
  • Secondary muons in cosmic ray air showers.
  • X rays emitted by UV transmitting glass in the
    optics of the instrument.
  • Radioactivity from heavy metal impurities in the
  • These events are identified, classified and
    rejected by splitting the CCD exposure into two
    or more equal parts, the hits dont occur in the
    same place.

Charge Transfer Efficiency
  • When the wells are nearly empty, charge can be
    trapped by impurities in the silicon. So faint
    images can have tails in the vertical direction.
  • Modern CCDs can have a charge transfer efficiency
    per transfer of 0.9999995, so after 2000
    transfers only 0.1 of the charge is lost.

CCD readout noise
  • CCDs suffer from readout noise which has a
    variety of sources
  • The output Field Effect Transistor. This is the
    ulitimate limit to the readout noise, at a level
    of 2-3 electrons.
  • Transfer loss fluctuations. During transfer an
    amount of charge is left behind, but this amount
    varies. Transfer noise is given by str ?
    (2?nN0) where ? 1-CTE is the fraction of charge
    not transferred, n is the number of transfers and
    N0 is the original charge. For faint sources
    (?100 electrons) this noise is less than 1

CCD readout noise
  • Reset noise there is a noise associated with
    recharging the output storage capacitor, given by
    sres ? (kTC) / e where C is the output
    capacitance in Farads. This noise could dominate,
    but it is removed by Correlated Double Sampling,
    where the reset voltage is measured after reset
    and again after readout. The first value is
    subtracted from the second, as this voltage will
    not change.
  • Surface State noise, due to fast interface states
    which absorb and release charges on short
    timescales. This is given by sss ? (2kTn?ssA),
    where n is the number of transfers, ?ss is the
    density of fast interface states, and A is the
    pixel area. In buried channel CCDs, ?ss is very
    low and this source of noise is less than 1

Other noise sources
  • Fixed pattern noise. The sensitivity of pixels is
    not the same, for reasons such as differences in
    thickness, area of electrodes, doping. However
    these differences do not change, and can be
    calibrated out by dividing by a flat field, which
    is an exposure of a uniform light source.
  • Bias noise. The bias voltage applied to the
    substrate causes an offset in the signal, which
    can vary from pixel to pixel. This can be removed
    by subtracting the average of a number of bias
    frames, which are readouts of zero exposure
    frames. Modern CCDs rarely display any fixed
    pattern bias noise

Interference Fringes
  • In thinned CCDs there are interference effects
    caused by multiple reflections within the silicon
    layer, or within the resin which holds the CCD to
    a glass plate to flatten it.
  • These effects are classical thin film
    interference (Newtons rings).
  • Only visible if there is strong line radiation in
    the passband, either in the object or in the sky
  • Visible in the sky at wavelengths gt 700nm.
  • Corrected by subtracting off a scaled exposure of
    blank sky.

Other noise sources
  • Electronic Pickup. Electronic noise sources
    caused by ground loops, or pickup of Radio
    Frequency radiation by cables and wires in the
    system, often dominates over noise sources within
    the CCD itself. This noise is easy to see, it
    forms ripples in the background, particularly in
    bias frames, but it is not fixed pattern so it is
    not easy to remove.

Signal to noise ratio (S/N)
  • Output S/N Input S/N x ?(DQE)
  • To calculate the S/N of an observation we need to
    know the signal, and all sources of noise. These
  • Photon noise (shot noise from the signal).
  • Photon noise from the sky background under the
  • Photon noise from the sky background measurement
    to be subtracted off.
  • Readout noise from all sources.
  • Electronic pickup noise.
  • Fixed pattern noise.
  • Bias noise
  • Dark current noise.

A case study of simple aperture photometry
  • We observe a star on a CCD detector, and process
    the data in the simplest way possible.
  • An area centred on the star is defined to be the
    object area, and is large enough to contain all
    of the photons from that star.
  • An equal area some distance away, which is found
    to be free of stars, is defined as the sky
    background area, and the sky background is
    measured from that.

A case study of simple aperture photometry
  • We will make some assumptions
  • We have eliminated fixed pattern noise by
    dividing the image by a normalised long exposure
    of a uniform light source, this is called a flat
  • Electronic pickup noise is indistinguishable from
    readout noise, and is included in the value given
    for readout noise.
  • Bias noise and dark current noise are negligible,
    as this is a cryogenically cooled, buried
    channel, MPP CCD.

Aperture photometry
  • Pixels are x µm square which equates to y
    arcseconds on the sky.
  • Star is observed in a circular aperture of area a
    square arcseconds which covers npix pixels.
  • Sky background is determined from a circular
    aperture of the same size.
  • Readout/pickup noise is sR electrons.
  • Aperture of telescope is D metres
  • We observe a star of magnitude V in the visual
    (green) passband.

Standard photometric passbands
Standard photometric passbands
Standard photometric passbands
Johnson and Stromgren passbands
(No Transcript)
Signal calculation
  • We start from the number of photons incident upon
    the top of the atmosphere of the earth from this
  • A star of V0, observed at a wavelength of 500nm
  • 1.0 x 108 ?? A photons/second
  • In bands other than V the number which is 1.0
    here is different, in bluer bands it is more, in
    redder bands less.
  • incident upon the top of the atmosphere in
    photometric (clear) conditions
  • ?? is the filter passband in nm
  • A is the telescope collecting area in metres2.

Signal calculation
  • Thats at the top of the atmosphere. There are a
    number of efficiency factors we need to multiply
    by to calculate the signal which reaches the
  • Atmospheric transmission eatm (0.88 for a star
    at the zenith).
  • Telescope reflection efficiency etel (0.92 per
    mirror 0.846 for a Cassegrain telescope)
  • Filter transmission efilt (0.85 for a
    broadband filter)
  • CCD Responsive Quantum Efficiency eCCD (0.75)
  • Cryostat entrance window efficiency ewin (0.95)

Signal calculation
  • There is also a geometric efficiency factor as
    part of the aperture of the telescope is blocked
    by the secondary mirror.
  • For a 2.0 metre aperture telescope with a 0.6
    metre secondary mirror
  • egeom (p 1.02 - p 0.32) / p 1.02
  • 0.91

Signal calculation
  • For a star of magnitude V the number of photons
    which is detected is given by
  • Nstar 1.0 x 108 10(-0.4 V) eatm etel efilt eCCD
    ewin egeom A ?? t
  • t is the exposure time in seconds. ?? 87 nm
    for the V band
  • For example for a star at V23 on a 2 metre
    telescope with the efficiencies we have quoted
  • Nstar 2.5 t

Signal calculation
  • In the absence of sky background and readout
    noise it would be simple, we would integrate for
    1000 seconds, detect 2500 photons, and have a
    signal to noise ratio of 50. But sky and readout
    noise are significant.
  • In dark sky (i.e. the moon below the horizon) at
    a dark site with no reflected street light, then
    the magnitude of a 1 arcsecond patch of sky in
    the V band is approximately VSKY21.5.
  • Every square arcsecond of sky gives
  • Nsky 1.0 x 108 10(-0.4 VSKY) eatm etel efilt
    eCCD ewin egeom A ?? t
  • 9.95t photons

Aperture photometry
  • Assume we have two apertures, one on the star and
    one on sky. Star aperture includes sky as well,
    and our estimate of the star intensity is the
    difference between the two.
  • Signal in aperture B (which is the sky aperture)
  • IB a Nsky
  • Signal in aperture A (which is the star aperture)
  • IA a Nsky Nstar

Noise on the measurements
  • Noise on the measurements has two components,
    photon noise which is given by the square root of
    the number of photons, and readout noise, which
    is determined by the readout noise and by the
    number of pixels in the aperture. The noise
    components add in quadrature
  • sB2 IB npix sR2
  • sA2 IA npix sR2
  • s(A-B)2 IA IB 2 npix sR2
  • s(A-B)2 2 a Nsky Nstar 2 npix sR2

Signal to Noise ratio
  • S/N (IA IB) / s(A-B) Nstar / s(A-B)
  • For example t1000 seconds Nstar 2.5t a
    19.6 sq arcseconds (5 arcsec diameter
    apertures) Nsky 9.95t npix 316 (0.25
    arcsecond pixels) sR 10 electrons.
  • S/N 3.77
  • In this example Sky noise dominates, if exposure
    time is
  • shorter then readout noise will dominate.

Improving the Signal to Noise
  • Smaller object aperture reducing the object
    aperture reduces both sky noise and readout
    noise. However you lose signal. The problem is if
    you are comparing the signal in different images,
    and fluctuations in image size (seeing) cause the
    amount of signal you lose to vary, then this
    introduces systematic errors in the brightness
    measured (photometry).
  • Larger Sky Aperture Increasing the sky aperture
    and scaling it to the size of the object
    aperture, or using several sky apertures and
    averaging them, reduces the noise to
  • s(A-B)2 ? a Nsky Nstar ? npix sR2
  • where ? ( 1 npix1 / npix2), and npix1
    and npix2 are the number of pixels in the star
    and sky apertures respectively. In practice the
    sky aperture is often an annulus around the star
    aperture. Must be careful that stars do not get
    in the sky aperture!

Improving the Signal to Noise
  • On-chip binning it is possible to shift two or
    more rows consecutively into the horizontal
    readout register without reading this register
    out, to bin the charge in the vertical direction.
    Similarly it is possible to read two or more
    pixels of the readout register consecutively into
    the output capacitor. But you only have to read
    the output capacitor out once and you only get
    one lot of readout noise.
  • This way you reduce readout noise at the expense
    of resolution, the resolution. Resolution should
    always be smaller than the characteristic size of
    the star images.

On-chip binning
Profile Fitting
  • Profile fitting is used most commonly in crowded
    fields, where it is difficult or impossible to
    define a sky aperture free of stars (or
  • It does however offer an advantage in precision
    even in sparse fields, because it weights the
    data more correctly.

Profile Fitting
  • Basic assumption is that the intensity profile
    (which is in principle a 2 dimensional function)
    is the same for all stars in a particular CCD
  • Intensity profile is determined by seeing or by
    diffraction, or occasionally by aberrations.
  • If it is determined by aberrations you need to be
    very careful, because the assumption that the
    profile is the same at all positions on the CCD
    may not be correct.

Profile Fitting
  • From a set of isolated, comparatively bright (but
    not saturated) stars in the frame, determine the
    image profile, this is called the Point Spread
    Function (PSF).
  • For ground based data an empirical approximation
    to the PSF is the Moffat function
  • f(r) Ci ( 1 r2/R02)-ß Bi (r lt rmax)
  • f(r) Bi (r gt rmax)
  • R0 is the characteristic radius of the star
    image, r is the distance from the centre of the
    image, ß describes the PSF at large radius, Bi is
    the background in the region of star i, and Ci is
    the relative brightness of star i.
  • Fit this function for each of the stars in the
    image to the data, using a least squares or
    similar technique.
  • For each star determine Bi and Ci. R and ß are
    constant within an image.

Profile Fitting
  • Then we have a set of scaling factors, which can
    be converted to a relative magnitude.
  • We need aperture photometry of one star, either
    from this CCD frame or from another, this can be
    a bright isolated star with high S/N, this gives
    the magnitudes of all of the stars in the frame.
  • The fit gives the correct weighting, rather than
    adding in lots of pixels with very little signal,
    S/N from profile fitting is usually at least a
    factor of 2 higher than from aperture photometry.
  • Profile fitting can cope with fields in which
    stars are close or their images even overlap.

Profile Fitting
  • For ground based data the PSF is determined by
    the seeing, and must be redetermined for each CCD
  • For space based (e.g. Hubble Space Telescope)
    data the PSF is fixed, and is often available as
    part of the standard calibration data produced
    with the observations. It still depends upon the
    passband (filter).

Procedures for photomtery
  • Astronomers work in magnitudes scales, which are
    a logarithmic scale
  • For each star calculate an instrumental
  • minst -2.5 log (Di / t)
  • Di is a measure of the total brightness in
    the star image t is exposure time.
  • We need to compare the instrumental magnitudes of
    stars of known magnitude with their true
    magnitudes, to calculate the offset, and thus to
    calculate the true magnitudes of all of the stars
    in the frame.

Procedures for photomtery
  • If we have standard stars in the CCD field that
    we are observing, then its fairly easy to
    calibrate, as we can just use the Ci values as
    our measure of intensity.
  • If not then we need to observe standard stars in
    separate CCD frames, and as the PSF will vary
    between different frames, we need to find a true
    measure of the brightness of the stars.
  • Di Ci ?0rmax 2pr ( 1 r2/R02)-ß dr
  • rmax is chosen so that we get all of the

Procedures for photomtery
  • Observe a set of standard stars (of known
    magnitude) at different airmass and at different
  • Atmospheric absorption is proportional to the
    airmass, which is sec(z) where z is zenith
    distance. Strictly this assumes a plane parallel
    atmosphere, but this is a good approximation for
    z lt 70 degrees.
  • There is a colour term, caused by the variation
    in spectral profile of the stars and the filter
    response over the passband.

Atmospheric absorption
Absorption ? sec z for a plane parallel atmosphere
Procedures for photomtery
  • minst mtrue C1 C2 sec(z) C3 (B-V) C4
    (B-V) sec(z)
  • (B-V) is the colour of a star, and measures
    the ratio of the intensity in the V band to that
    in the B band. Other colours can be used, e.g.
  • Solve for C1, C2 , C3 , C4 from stars of known
  • Usually C3 is negligible, often C4 is too. In
    this case we can now simply convert the values of
    minst to mtrue using the values of C1 and C2 that
    we solve for, and the value of z for each

Procedures for photomtery
  • However if C and/or D is not zero, we need to
    know (B-V) for the star to calculate the true
    magnitude, and we do not. In this case we must
    observe in two passbands, for instance B and V,
    and use
  • Vinst Vtrue C1 C2 sec(z) C3 (Binst-Vinst)
    C4 (Binst-Vinst) sec(z)
  • Binst Btrue C5 C6 sec(z) C7 (Binst-Vinst)
    C8 (Binst-Vinst) sec(z)

Image Intensifiers
  • An image intensifier is a device which amplifies
    light signals by
  • converting photons to electrons via the
    photoelectric effect at a photocathode,
  • accelerating the electrons them via electrostatic
  • having them impact on an output phosphor
    releasing a shower of photons,
  • recording the output photons using a photographic
    emulsion or some more modern detector (or indeed
    the human eye).

Image Intensifiers
  • The gain of an image intensifier is the ratio of
    the number of output photons to the number of
    input photons.
  • Some means must be used to focus the electron
    beam, i.e. to ensure that there is a one to one
    mapping between the position of impact of the
    incident photon on the photocathode, and the
    position of release of the output shower on the
    phosphor. Image tubes are either
    electrostatically or magnetically focussed.

Image Intensifiers
Image Intensifiers
  • Often several (up to 4) stages of intensifier are
    used, leading to a total gain of order 106.
  • Image intensifiers are now used very little in
    the optical, where CCDs have taken over, because
  • Photocathodes have lower RQE than CCDs.
  • Intensifiers require high voltage supplies, and
    are unreliable for instance in damp conditions.
  • They suffer from saturation effects when used in
    photon counting mode.

Image Intensifiers
  • Image intensifiers remain popular in the
    ultraviolet because
  • RQE of CCDs drops because the electrodes they
    rely on are opaque at UV wavelengths.
  • RQE of Image Intensifiers is higher because
    photocathodes respond more efficiently to higher
    energy photons.
  • Photon rates from astronomical sources are lower,
    so saturation effects are less serious, and CCD
    readout noise becomes more serious (photon
    counting detectors are noise free).

Microchannel Plate intensifiers
  • A microchannel plate is a modern image
  • It consists of a thin disk of lead oxide glass
    with numerous microscopic channels running
    parallel to each other from one face to the
  • A potential of a small number of kiloVolts is
    applied between one face and the other.
  • Each channel acts like a tiny image intensifier.
    electrons hitting the walls eject additional
    electrons resulting in a cascade of electrons.

Microchannel Plate intensifiers
  • Microchannel plate still needs a photocathode and
    an output phosphor.
  • Advantages over conventional image intensifiers
  • Channels or pores confine the electron shower so
    that the resolution is better.
  • Voltages are lower (2 kV as opposed to 30 kV
    for gain of 106.
  • Pores are either slanted in opposite directions
    in a stack, or curved
  • To allow the electrons to hit the walls to
    provide the gain
  • So that positive ions produced from residual gas
    within the tube hit the walls and are absorbed
    before they acquire enough energy to generate a

Microchannel Plate intensifiers
Z Stack of Microchannel plates
Photon Counting Detectors
  • Run an image intensifier at high gain (106), and
    image the output phosphor onto a CCD or similar
    solid state detector.
  • For each photon incident at the photocathode
    there is a large splash of photons at the
  • Read this out and centroid using hard wired logic
    or software on a fast computer.
  • Record in solid state or computer memory a photon
    at the location of the centroid.
  • In this way build up the image photon by photon.

Saturation in Photon Counting Detectors
  • If more than one photon arrives in a particular
    location within the frame time of the detector
    then one or both will be lost.
  • There is a limit to the count rate in a
    particular location
  • In some devices there is also a limit to the
    total count rate in the frame.
  • Unlike CCDs, you cannot remove saturation by
    taking short exposures.
  • Photon counting detectors are therefore most
    useful in the Ultraviolet, where photon rates are
  • Photon counting detectors have no readout noise
    and have a potential advantage for all ultra-low
    light level applications.

The Image Photon Counting System
  • Developed in the 1970s by A. Boksenberg and J.
    Fordham at University College London.
  • Early generations used 4 stage magnetically
    focussed Image Intensifiers, and Plumbicon TV
  • In later generations the Intensifier is replaced
    by a Microchannel plate, and the Plumbicon by a

The Image Photon Counting System
MCP/CCD incarnation of the Image Photon Counting
The Multi-Anode Microchannel Array
  • Developed for space applications (particularly
  • Used a position sensitive anode instead of an
    output phosphor and light sensitive detector.
  • Anode consists of two perpendicular\ sets of
    coding electrodes.

The Multi-Anode Microchannel Array
S/N with a photon counting detector
  • Going back to our case of aperture photometry
  • S/N (IA IB) / s(A-B) Nstar / s(A-B)
  • s(A-B)2 2 a Nsky Nstar 2 npix sR2
  • Nstar1.0 x 108 10(-0.4 mstar) eatm etel efilt
    eDet ewin egeom A ?? t
  • Nsky1.0 x 108 10(-0.4 msky) eatm etel efilt eDet
    ewin egeom A ?? t
  • For a photon counting detector eDet is generally
    lower than for a CCD, as it is given by the
    photocathode efficiency. However sR is 0. If
    mstar and msky are faint, A is small, or ?? is
    narrow, of if t must be short because good time
    resolution is required, then it is possible that
    S/N might be higher with the Photon Counting
    Detector than with the CCD.

Infrared Passbands
  • Infrared passbands are determined by the windows
    of transparency of the atmosphere.

Infrared Passbands
Primary Infrared bands available for ground based
Infrared Array detectors
Infrared semiconductor detectors cannot be made
out of silicon any more as the band gap is too
large. Germanium can be used at the shortest
wavelengths but Indium Antimonide and Mercury
Cadmium Telluride are more widely used.
Infrared Array detectors
  • People have tried making CCDs out of Indium
    Antimonide, but the yield is too small.

Photovoltaic detectors
  • Single pixel infra-red detectors have long used
    the photovoltaic effect.
  • Diode is formed at the junction between a p- and
    n-doped semiconductor.
  • pn junction generates an internal electric field
    to seperate the photon generated electron-hole
  • Migration of holes and electrons changes the
    electric field, hence there is a voltage change
    across the junction which can be measured.

Hybrid Arrays
  • Modern IR arrays are hybrid arrays, formed of a
    sandwich of three layers.
  • Top layer (assuming radiation is coming down) is
    a Indium antimonide or Mercury Cadmiun Telluride,
    doped to act as a photovoltaic detector.
  • Bottom layer is a silicon multiplexer, which can
    be a CCD but if more often an array of tiny
    MOSFET (Metal Oxide Semiconductor Field effect
    Transistor) amplifiers.

Hybrid Arrays
  • In between are Indium bump bonds providing an
    electrical connection between locations on the IR
    detector and the elements of the silicon
    multiplexer. Indium is a good conductor which is
    soft even at low temperature, so the device does
    not crack when cooled.
  • Gaps between the indium bonds are filled with
    epoxy resin to provide mechanical stability.

Hybrid Arrays
Hybrid Arrays
Architecture of an Indium antimonide hybrid array
Performance of Infrared arrays
  • Indium Antimonide and Mercury Cadmium Telluride
    arrays are available in 1024 square formats,
    maybe 2048 square.
  • Quantum efficiency 60 - 80
  • Noise 40 electrons/pixel
  • Cost 300,000 per array!

X-ray detectors
  • X-ray detectors are chosen to have good spatial
    resolution as well as some level of intrinsic
    energy resolution. The three types in serious use
    today are-
  • Proportional Counters
  • Microchannel Plates
  • CCDs

Proportional counters
  • X-rays are detected through their interaction
    with inert gas (e.g Argon) in a windowed chamber.
  • Primary photoelectron is emitted by the
    photoelectric effect.
  • Secondary Auger electrons emitted in a
    localised cloud.
  • Mean number of electrons released is N E/w,
    where E is the energy of the X-ray, and w is the
    ionisation energy of the gas (w 26.2 eV for
    Argon 21.5 eV for Xenon).

Proportional counters
  • There is a variance on the number of electrons
    released, sN2 F N, where F is a property of the
    gas called the Fano Factor. For Argon and Xenon F
  • Position information is obtained using a position
    sensitive (resistive) anode. The pulse is
    extracted from both ends of the anode, and from
    the size of the pulse at each end the location is
    determined. A grid of anodes provides two
    dimensional imaging. Alternatively the position
    can be measured using a pair of crossed cathode
    grids above and below the anode.

Position Sensitive Proportional Counter
Position Sensitive Proportional Counter
  • Cathodes modify the field strength, providing a
    low field drift region where the X-rays are
    absorbed, and a high field avalanche region
    near to the anode.
  • Anticoincidence detector is used to reject
    charged particle events. Clouds caused by X-rays
    will not impact on the anticoincidence grid.
    However charged particles will, so events
    detected on both the main grids and the
    anticoincidence grid are rejected.

X-ray CCDs
  • CCDs become very inefficient in the Ultraviolet,
    primarily because the electrodes they use become
    opaque to radiation.
  • However at soft X-ray wavelengths they become
    efficient again, especially thinned (but not so
    thin as optical) CCDs.
  • CCDs now produce many electron-hole pairs per
    incident photon, and the number they produce
    depends upon the energy of the photon.

X-ray CCDs
  • Process is photoeletric, and is exactly the same
    as in a gas proportional counter.
  • X-ray causes the release of a high energy
    electron, which in turn releases secondary
    electron-hole pairs by a solid state analogue of
    the Auger effect.
  • Number of electrons released N E / w, where w
    is not the band gap energy because the electrons
    released are now mostly bound electrons.
  • For silicon w 3.65 eV

X-ray CCDs
X-ray CCDs
  • As some energy is transferred to the crystal
    lattice, there is a statistical variation in the
    number of electron-hole pairs produced.
  • Thus there is a solid state Fano factor, so
  • sN2 F N
  • For silicon F 0.1
  • Energy resolution is given by
  • ?E 2.36 (FEw)1/2
  • factor 2.36 is because this is expressed as a
    full width half maximum rather than a standard

X-ray CCDs
  • For silicon, w is lower than for the inert gases
    (3.65 eV as opposed to 20) and F is slightly
    lower (0.1 as opposed to 0.17) so CCDs have much
    better intrinsic energy resolution than gas
    proportional counters.

X-ray CCDs
X-ray CCDs
CCD array for the XMM-Newton satellite
Microchannel plates
  • Microchannel plates work at X-ray wavelengths
  • X-rays release photons in the lead glass channel
    walls directly.
  • Walls are coated with a material of high
    photoelectric yield, especially at energies below
    1 keV.
  • At energies higher than 5 keV, X-rays can
    penetrate the channel walls to release
    photoelectrons in neighbouring channels. This
    degrades resolution.

Microchannel plates
X-ray MCP detector flown on CHANDRA
Gamma Ray Detectors
  • Scintillation crystals
  • Solid state Cadmium Zinc Telluride detectors
  • Compton Scattering detectors
  • Pair production detectors
  • Air Cerenkov detectors

Scintillation crystals
  • Crystal converts a gamma ray to a shower of lower
    energy photons, which are detected by a

Scintillator produces very poor directional and
energy resolution.
Compton Scattering Detectors
  • Photon scatters off an electron, transferring
    energy to the electron.
  • Detector consists of two levels. In the upper
    level the photon is scattered.
  • Photon continues to the lower level, but electron
    emits a shower of photons which are measured by
  • Lower level is a scintillation detector, photon
    is absorbed again emitting a shower of low energy
    photons, which are measured by photomultipliers.
  • From the location and energy of the two showers
    of photons, incoming gamma ray direction and
    energy can be calculated.

Compton Scattering Detectors
Pair production detectors
  • Gamma ray with energy gt 30 MeV interacts with a
    material to produce an electron-positron pair.
  • Electron and positron are tracked by spark
    chamber layers.
  • Energy and direction of the electron and positron
    give the energy and direction of the incoming
    Gamma ray by conservation of energy and momentum.

Pair production detectors
Egret pair production detector
Cadmium Zinc Telluride (CZT) detectors
  • Cadmium Zinc Telluride is a high band gap
  • By adding pixel electrodes on one face of a wafer
    of CZT, a two dimensional photoelectric solid
    state detector can be made.
  • These have small number of pixels, so pixels tend
    to be read out individually (rather than charge
    shuffled as in a CCD).

Cadmium Zinc Telluride (CZT) detectors
CZT detector arrays
Coded mask imaging
  • CZT arrays are often used with coded mask imaging
  • A coded mask is an array of transparent and
    opaque (usually lead) elements in an optimised
  • It operates as an array of pinhole cameras.
  • From the pattern on the CZT array you can
    determine the direction of the gamma ray source
    (there will only be one in your field of view!)

Coded mask imaging
Coded mask design for the Integral satellite
Coded mask/CZT array
Coded mask array proposed for the Gamma ray Large
Area Survey telescope (GLAST) project
Laue Diffraction Imaging
Detector in any Laue diffraction imager is also
likely to be a CZT array
Air Cerenkov Detectors
  • Rely on Cerenkov radiation from secondary
    particles created from pair produuction in the
    upper atmosphere.
  • Secondary charged particles travelling faster
    than the local speed of light emit radiation.
  • Detected by ground-based optical telescopes.

Air Cerenkov Detectors
The Quantum Calorimeter
  • A Quantum calorimeter is a bolometer which is so
    sensitive, and has such good energy resolution,
    that it is able to measure the energy deposited
    in an absorber by an individual photon.
  • This is rather easier of course for higher energy
  • Transition Edge Sensors have been used
    successfully as Quantum Calorimeters at X-ray,
    and recently optical, wavelengths.

Quantum Calorimeter
Quantum Calorimeter
Quantum Calorimeter
Quantum Calorimeters
  • Energy resolution of a quantum calorimeter is
    given by
  • ?E ? ? ((kTb2 C) / a)
  • where Tb is the temperature of the heat sink,
    and C is the heat capacity of the absorber at
    that temperature. From the notes on bolometers
  • a (T/R)(dR/dT) -0.5 ? (T0 / T)
  • is the temperature coefficient of resistance.
  • Time constant is t C / G, but electrothermal
    feedback reduces this.

Quantum Calorimeters
  • For a Quantum Calorimeter operating at optical
    wavelengths, we make C and Tb as low as possible,
    and a as large as possible. The time contant
    must be short enough that the array pixel
    recovers from one photon event before the next
    arrives. Electrothermal feedback in TES sensors
    makes the time constant short.

Quantum Calorimeters
  • At X-ray energies there is an extra problem, if
    the energy of a single photon takes the TES out
    of the transition region, then the response to
    energy will be nonlinear.

Quantum Calorimeters
  • Increasing the heat capacity C will overcome
    this, but at the expense of lengthening the time
    constant of the detector, resulting in photon
    pile-up (i.e. when one photon arrives before the
    previous one has been read out).
  • Solution is to make a TES with a higher critical
    temperature Tc. More bias power is now required
    so electrothermal feedback is stronger.

Quantum Calorimeters
  • Electrothermal feedback will now keep the device
    in the transition region, and decrease the time
    constant so that higher incident photon rates can
    be observed.
  • In practice the device is made with a combination
    of two metals, a bilayer film, and by adjusting
    the relative thickness of the layers the critical
    temperature can be tuned in the range between the
    critical temperatures of the two individual

X-ray Quantum Calorimeters
Two Quantum Calorimeter array detectors for X-ray
Superconducting Tunnel Junctions
  • STJ detectors are cryogenic detectors operating
    on a principal more similar to the semiconductor
    detectors we discussed earlier.
  • Consist of two superconducting electrodes
    separated by a thin insulator.
  • There is a small energy gap between the
    superconducting electronic ground state (which
    consists of Cooper pairs), and excited single
    particle states (quasiparticle states).

Superconducting Tunnel Junctions
Superconducting Tunnel Junctions
  • Photons break Cooper pairs and excite many
  • Band gap between ground state and quasiparticle
    states is 1 meV (thats milli electron Volt), so
    there are 1000 times more carriers generated
    than in Silicon detectors where the band gap is
    1.1 eV.

Superconducting Tunnel Junctions
  • Bias voltage is applied across the barrier. To
    suppress the zero illumination Josephson current
    a parallel magnetic field is applied.
  • Carriers then tunnel through the barrier to the
    other electrode, and produce an increased current
    which can be measured.
  • STJs made out of Niobium or Tantalum have been
    used in astronomical applications from soft X-ray
    to mid infra-red wavelengths.

Superconducting Tunnel Junctions
  • Energy resolution of an STJ is limited by an
    equivalent of Fano noise, and in general is not
    as good as TES Quantum Calorimeters.
  • Time resolution is limited by the recombination
    time of the quasiparticle states, which is of
    order 2-10 µs, which is better than TES Quantum
  • RQE of both kinds of detector is in the range

Superconducting Tunnel Junctions
The S-cam (ESA) STJ array (left), and its
complete instrumental setup at the William
Herschel Telescope (right