XMM-Newton SOC Glossary

Below is a glossary of commonly used terms in the SOC documentation.

sigma parameter
The Gaussian function is determined by the normalization, the mean (i.e. the position of the peak), and its sigma. The sigma is proportional to the width in that the FWHM=2*sigma*sqrt(2*ln(2))=2.354*sigma. The contents holds the following +/-1 sigma 68.33% +/-2 sigma 95.45% +/-3 sigma 99.73%
sigma parameter in the context of event thresholding
The zero level peak is used to estimate the noise of the system by fitting a Gaussian. The sigma of this Gaussian may be multiplied by an integer of real and this may then be used as threshold to distinguish valid events from noise. Two levels of threshold may be used: one higher for the main absorption site, and a lower one for surrounding pixels, which contain split charge.
Absolute measurement accuracy (AMA)
The absolute measurement accuracy is the angular separation between the actual direction and the reconstructed direction (a posteriori) of an axis.
Absolute pointing error (APE)
The absolute Pointing Error is the angular separation between the actual direction and the intended direction of an axis.
Absolute pointing drift (APD)
The absolute pointing drift is the change of the angular separation between the actual direction and the intended direction of an axis over a given period in time.
Absorption feature
Every material absorbs light with increased efficiency around its characteristic edges. For example, the spectrum of white light seen through a hot gas shows some notches in intensity around the edges of the materials contained in this gas. These notches are called absorption features. Absorption features on XMM may be observed from astrophysical processes (similar to the one described above), and from materials contained in the components on the path of light.
Background Rate
Background rate is a general term for the count rate which has no origin in celestial sources and which are registered by the instrument. The background rate/background events constitute out of different components, such as internally (e.g. thermally) generated events, infalling cosmic particles triggering the detection chain. Some large scale celestial sources (diffuse X-ray background, optical zodiacal light, etc.) are in the most cases also considered as contributors to the background rate.
Bad Pixel
General term for CCD Pixels or Image Pixel exhibiting a abnormal behaviour which make them useless for scientific evaluation. Depending on the nature of the defect, bad pixel are either insensitive (dead pixel) or mimic a signal (hot or warm pixel).
Boresight offset
Boresight is a synonym for the optical axis. Offset refers to differences of optical axis of the instruments.
Bright Star Tracking
OM will track guide star to perform its onboard drift correction, which prevents image smearing induced by s/c jitter. There are two different ways of tracking. Bright Star tracking provides information (centroid and peak coordinates, integrated and peak counts and parameters describing the window used to delineate the stellar image and moments) on the N brightest stars in the reference frame throughout an exposure, whereas the Threshold tracking provides information on the brightest stars in each tracking frame.
Centroiding Algorithm
In order to achieve high resolution with the 256x256 pixel CCD the spatial position of the centre of gravity (centroid) of each photoevent is determined in hardware to sub-pixel accuracy (1/8 of a pixel both in X and Y). The event centroid is calculated from the relative intensity distribution in the 3 CCD pixels in an algorithm applied independently in X and Y direction.
Charge spectrum
The charges contained in the CCD pixels are accumulated in a histogram.
In a CCD, the process of transferring charge from pixel-to-pixel is not 100% efficient. This causes a degradation in correspondence between detected charge measured in the CCD output, and the original X-ray energy. The charge transfer loss and noise is therefore spatially and temporally changing. Charge transfer efficiency (CTE) is the term to describe this less than perfect efficiency (measured per pixel normally).
Charge Transfer Inefficiency: CTI= (1- CTE)
Convolution in a strict mathematical sense is the multiplication of two functions and caluclating the integral over the range of validity. This term is often used for the multiplication of the output from an astrophysical model with the response matrix of instruments.
Cross dispersion direction
The cross dispersion direction is the direction normal to the dispersion direction.
Cycle Time
The time period of a complete cycle in a readout sequence of one or more detectors. In the example of the EPIC PN detector in Full-Frame mode, it is the time interval between two consecutive reads of pixel 1 of CCD 0 (all 12 CCDs have been read out in the meantime).
Dark Noise
Electrons or light thermally generated in the detection device can be registered as valid events by the instrument and contribute to the background. The levels of dark noise in the instruments are determined by special measurements with the filterwheel in blocked position.

For the OM MIC detector, dark noise from the photocathode and the microchannel plate appear as low charge level events in the centroiding CCD detector. If they have enough charge to appear as a valid event, the signals will be centroided and the data will be registered as a photon, indistinguishable from normal sky background light.

For EPIC and RGS, the dark noise has a different effect: Through cooling the CCDs, the total number of dark current electrons in a pixel in one CCD frame should be < or order 1. However, whatever the mean level is, this is taken as the reference level against which true signal charge levels are measured. Thus if the CCD temperature were to increase, the mean level of dark current would likewise. The major degrading effects on performance are: (i) the random shot noise on dark charges collected per pixel per frame degrades energy resolution, and (ii) if the increase in mean level is not detected, a false zero reference level would be assumed, and energy scale and event patterns not registered properly
Optical element originally designed to be reflective for shorter wavelength and transmitting for longer wavelength to split the infalling beam into the red and blue beam. For the selection of between the two blue detectors the dichroic is rotated by 180 degree. Now after the demise of the red detector the dichroic is replaced by a pure reflecting element, which changes the direction of the beam onto the blue detector. The rotation by 180 degree for the selection between the two detectors is still possible.
Difference between pn and MOS CCDs
The two CCDs used in EPIC are fabricated from different technology. The EPIC MOS (and RGS) CCDs are produced by conventional MetalOxideSemiconductor processes. That is the silicon substrate is overlaid with an oxide insulator followed by a metal (actually polycrystalline silicon) gate electrode. The photolithographic definition of these gate electrodes produces the pixellised structure. The pn-type CCD is produced by implants of p- and n-type material in the silicon, and the connection to, and definition of these implant structures defines the pixels.

Beyond this process definition there are a multitude of minor factors that need to be borne in mind in comparing the details of operation w.r.t. EPIC. However, the most important difference is that the pn-CCD has been designed with relatively large pixels especially to match the XMM spatial resolution, and with individual amplifiers associated with every column of these pixels. A much higher access rate to the pixels can be achieved with the multi-parallel readout. In addition, much simpler event selection is performed in off-chip electronics on the pn.

The MOS CCDs are read out to only one amplifier, so that rows are fed into an output register, and then serially the pixels are transferred through the register to an output node. The data are thus accessed more slowly, and in order to reconstruct events spread across pixel boundaries, the data have to be operated on by off-chip electronics. (For background details on the use of CCDs for X-ray astronomy, see the paper by Lumb et al (Exptl. Astronomy v2 p179 1991)
Dispersion in the context of gratings refers to the fact that photons originating geometrically from the same incident beam are exiting the grating at angles, which are dependent on their wavelength. Dispersion occurs only in the direction normal to the grid lines of the grating. Photons with longer wavelengths (lower energies) are dispersed more, those with shorter wavelengths are dispersed less.

Dispersion also is a periodic process, such that dispersion towards larger angles is possible (i.e. higher orders). The dispersion direction in such a case is the same as that for a photon with an integer multiple of the wavelength. Provided that the orders can be separated, the energy of photons can be determined very accurately from the angle of dispersion.
Dispersion Direction
The direction dispersion occurs.
Dispersion Scales
Energy determination of dispersed photons can only be carried out, if the hypotetical zero of the wavelength scale is known. The zero of the wavelength scale refers to the position of a beam of zero wavelength, infinite energy, which is not dispersed. Different scales, although all having the same wavelength zero, arise from higher orders. The dispersion scale may also be shifted by different angles of incidence on the dispersive element (i.e. the RGA).
Effective Area
Not every photon falling into the telescope is finally detected and registered as event. The effective area is the fraction of events finally registered multiplied with the collection area of the telescope, thus it is a virtual area, where all infalling photons are detected and registered.
Generic term for a signal occurring in the detector. For the X-ray cameras, the deposition of an X-ray in the CCD produces signal charge, which is eventually registered when the matrix of pixel charges are read out to external electronics. Anything registered above the threshold level of these electronics might also be considered an event, for example the detection of a cosmic ray or proton, gamma ray, or other noise source. In the OM, the desirable signal event is of course the detection of an optical photon. In the science documents, event is used synonymously with "photon" and "count"
Flexible Image Transport System (FITS)
FITS is a general self descriptive file format, where the contents of each file is described by keywords within the same file. Almost any file format can be constructed. Standards have evolved in astronomy to describe photon lists, spectral, and image data which are used by major packages. GSFC defined the OGIP standard (see K.M.Smale and N.White: "Legacy The Journal of the High Energy Astrophysics Science Archive Research Center, Number 2, November 1992, for copies send e-mail to request@legacy.gsfc.nasa.gov) which is suggested to be used for data storage and distribution by future projects.
Flat Field
Ideally an image obtained via an observation with a homogeneous illumination of the whole field of view. Non uniformities in the data-image are caused by the differing sensitivity of the detection chain from image pixel to pixel. The Flat Field can be used to correct for these difference and to normalize the detection chain to uniform detection sensitivity. As a uniform illumination of the whole field of view can't be achieved in the most cases in orbit, different techniques are utilized either to remove the inhomogeneity of the illumination or to compare and reference non uniform images of a time stable source illumination the whole field of view.
Frame time
In the context of the instrument detectors, a data set generated by a detector during a quantised period, such as a single CCD readout or an interim OM photon-counting image accumulated in a memory is termed a "FRAME". For the X-ray cameras the frames will be identified by a changing time stamp associated with a CCD readout. OM has BLUE tracking FRAMES, which are accumulated in memory. Again these are not telemetered as individual data units. Thus we derive a definition of "FRAME TIME" as the time for collection of photons within one CCD (irrespective of clocking state) or other sort of (eg. tracking) FRAME. For a CCD the Frame Time is thus equal to the Integration Time + Readout Time (or Transfer Time in some cases, depending on the operatingmode)
The grism is an optical element in the filter wheel, which allows to generate spectra of point like objects. The grism disperses the infalling photon beam into one direction, i.e. the image of a point like source is a strip, where each position in the strip corresponds to a particular wavelength.
Guide Star recognition
OM employs an onboard algorithm to perform tracking. The stars used in the tracking process have to be have to be determined ("recognised") for every tracking frame, in order to calculate the offset in the image within the last frametime.
Heartbeat (DPU)
There are actually two type of heartbeats, the SSI and the SCI heartbeats. The SSI (the interface between the DPU and the ICU) heartbeat is a signal sent from the DPU to the ICU saying, "We are here!" It is sent every 10 seconds (time interval set during compilation) whenever the DPU is powered, The SCI (interface between the DSPs) heartbeat is a signal sent by the Blue 1, Blue 2, to the White DSP indicating that they are operational. Every 1.0 seconds (time interval set during compilation), the White DSP asks each of the other DSPs, in turn via SCI, if it is alive. Each must respond with the "I am alive" signal written to global memory (when that DSP has access to the global address bus), which the White DSP checks the following bus cycle. If White does not read the "I am alive" signal from global memory, it issues an alert on the SSI to the ICU
Hot Pixels
Synonymously used with terms "dark spikes", "bad pixels" etc.. Any pixel which has a permanently high signal such that it is detected in every frame of a CCD is called a "hot pixel". In the CCD, extra dark current generating states or other defects cause unwanted signal charges to be gathered during the frame. For the X-ray cameras, such sites appear like isolated X-rays. Particularly after radiation damage, a large number of such pixels would occupy valuable telemetry bandwidth. Also, due to the requirement to examine data for events splitting between pixels, the hot pixel invalidates signals from the 8 surrounding pixels. The simplest approach is to have a bad pixel look-up table in the on-board electronics (and/or ground software) to blank out the effect of these pixels. (see also pixel flickering).

Similar effects will be manifested in the OM. Individual microchannel plate pores may repeatedly produce a signal level which is centroided in the off-chip electronics as a real event. Again the affects also surrounding pixels (up to 24 x 24 defined by the centroiding process). However, because the OM images are stored in memory, and spacecraft drift is small, this area is only read down through telemetry only once per 1000 sec image.
Hot Column
Occasionally, a hot pixel generates so much charge, that (especially if it is located in the store section (MOS CCDs) of the device, it emits charge into every pixel within a column, as these pixels are clocked past the generation site. Every pixel in the column thus has an anomalously high signal level.
Hot Spot
Synonym for hot pixel.
Individual chip rotations
There are 7 EPIC MOS CCDs which are assembled into the final focal plane array. They are carried on a chip carrier, and after the assembly is performed, the location of the devices will be checked optically. Not only will the gaps between active pixel areas be measured, and the position with respect to the X-ray axis determined, but the rotation of the row/column axes must be determined. Rotation of one part in 1000 away from parallelism may be enough for an outer pixel to be displaced by one in X or Y direction from its expected position. This may be important for astrometric determination of the position of serendipitously detected sources in the sky.
Integration Time
Time interval during which the detector accumulates photons, whithout clocking. Events
incident during the Integration Time are thus not smeared (as opposed to those incident during Readout Time).
Intensity Transfer Curve
Diagram/curve which describes the conversion from measured counts (i.e. photons) in intensity (i.e. photons) infalling into the telescope aperture.
Isolated Event
If an X-ray photon is absorbed sufficiently close to the center of a pixel, all the generated charge is contained within this pixel. Since integration times are selected such that it is highly unlikely that another photon is absorbed in a neighboring pixel within the same frame, the neighboring pixels will not contain charges above threshold. Events with this pattern type are called isolated events (see also split events).
Latch-up event
Also Soft Event Upset. Highly ionising event in CMOS devices can produce bit flipping, or in worst case a run-away current surge which permanently turns a transistor pair on/off.
Light curve phase folding
The intensity of some sources (e.g. binaries) is changing in time with a fixed period. Periods are generally analyzed by Fourier transformation. The identification of the power of each period is called light curve phase folding.
Light Leaks
In the ideal case only light out of the direction of the FOV should be detected by the instruments. However if there is an unforeseen opening somewhere in the telescope, light from another direction than the FOV can shine thorough and fall on the detectors (either directly or after reflections indirectly). These unforeseen opening/holes are called light leaks.
Line profile
The term line profile is used for the two dimensional image, which is generated on the RFC for an incident mono-energetic beam. The components of this image are: the PSF, which contributes to the width of the line in the dispersive direction and in the cross-dispersive direction; the scattering from the RGA.
Line spread function (LSF)
The LSF of a dispersive instrument (like the RGS) is defined as the integral of the two dimensional diffraction image (see line profile) along the cross-dispersive direction.
Mean level/value of zero pixels
(see dark noise). Due to effects of dark current or other constant background signals, every CCD pixel will have a mean level of charge which will be the reference against which X-ray signals are measured. In any case, due to the noise distribution around the mean level, the Analogue-to-Digital Converters must be arranged that a valid AD unit is produced even for the most negative of noise excursions, thus the mean value for pixels with no apparent signal will be set at some positive ADU value.

For OM it is not expected that there is a true zero value, due to sky background light. However in the presence of a blank shutter, a short exposure MIGHT result in some pixels with no detected signal. Then there WOULD be a true zero value stored in the appropriate memory locations.
Multi Channel Plate (MCP)
A MCP is a photon detection device. A photon kicks an electron out of a photocathode. This electron is amplified via a three stage channel plate with a voltage step between each plate. Each channel plate is a bundle of very tiny pipes, so that the position information is preserved. After the third plate the signal is converted back to photons via an output phosphor. The outcoming photons fall as a big splash onto a CCD and are registered as one event.
Observation target coordination frame
Off-axis event
Sources which are not in the center of the field of view, have a different focus. The focus has the same focal length and is along line which follows from the source through the point of intersection of the optical axis at the center plane of the mirrors. See also on-axis focus.
On-axis focus
The focus per definition is the point of convergence of a parallel incident beam. The optical axis is the symmetry axis in the center of the optical system.
On-board IC background rejection
Known sources of background like hot pixels can be rejected on-board by the IC and are thus not consuming bandwidth. This rejection is mainly carried out by a cross-check of the data against a lookup table. Another type of background may be event patterns, high charges (on the assumption thah they are generated by minimum ionizing events), etc.
On-chip binning
(see pixel binning). By collecting more than one CCD row of pixels in the output register, and more than one pixel in the output node before measuring the signals, The signals from a number of pixels can be combined on-chip. E.g. collection of two rows in the register, and pairs of pixels in the node leads to 2x2 pixel binning. The advantage of performing this on-chip is that it is in principle not subject to the quadrature summation of noise that is imposed by the amplifier reading process if the summation is performed in software.
Optical Axis
The symmetry axis of the instrument. The optical axis should be coaligned with the pointing direction of the instrument (z-axis).
Partial Event Floor
X-ray events liberate charge into a near point-like distribution within the original pixel. If the signal charge subsequently diffuses to neighbouring pixels before collection, or it recombines partially at some CCD interface structure, the total amount of charge detected suffers some deficit. Some fraction of all events are therefore manifested with "partial" signals, and in an energy spectrum will be seen as a collection of events with lower magnitude than the peak distribution of events seen at the fully collected charge level.
Pileup is generated by two or more X-rays being absorbed close enough that the charges they generate either end up in the same pixel, or in neighboring pixels such that the signature of these charges is indistinguishable from a split event. In all these cases the observed energy is the sum of the energies of the photons.
PHA Channel
The detection process of a photon results in a analog signal. This signal is converted to a digital signal via an analog-to-digital converted (ADC). The digital signal is registered by a pulse-height analyser (PHA). The amplitude of the signal (typically in the range of 1-512 or 1-1024) corresponds to the PHA Channel.
Photometric Calibration
The photometric calibration is essentially described the by transfer function. It describes how the measured brightness (measured counts) of an object converts into magnitudes.
Photon Pulse Width distribution
Diagnostic data to describe the MCP characteristic. It is no longer foreseen.
Photon Pulse Height Distribution
Normally OM only adds an event per signal above threshold to the image currently accumulated in memory. As diagnostic tool, the pulse height of the signals can be accumulated to a spectrum and this spectrum is called photon pulse height distribution.
pixel= picture element. The pixels of the X-ray CCDs are uniquely defined by the photolithographic manufacturing process. The OM pixels are logical pixels defined by a centroiding mechanism, such that one CCD readout pixel is associated with 8 logical pixels. Logical pixels can also be further defined according to any binning processes. For the OM in particular, the detector pixels are normally co-added 2x2 in memory before transmission to ground.
Pixel Flickering
(see hot pixels). In addition to pixels with permanently high levels of signal charge, some dark current generation states are meta-stable and result in excess signals, whose magnitude may vary from CCD frame-time to the next. The range of flickering timescales can be very great, so that treating these pixels via simple look-up tables is not straight-forward.
Pixel Binning
See on-chip binning. The binning of signals from pixels can occur on- or off- chip. This deliberate reduction in spatial resolution may be useful for reducing data content.
Plate Scale
The plate scale describes which coverage of sky one pixel at the FOV center has. The units are arcsec/pixel.
Point Spread Function (PSF)
Function which describes the transformation of the real sky into the data space of the measurement. I.e. if the sky in the FOV is known, the data measured by the instrument can be calculated by convolving the "input sky" with the PSF. The PSF can be split into different parts, e.g. the RGS PSF can be split into contributions of the mirror, the grating array and the detectors.
Priority Window Data
Telemetry packet which is issued with a high priority, i.e. shortly after the commencement of each exposure. It will contain parameters describing the detector windows assigned for sciencee data and tracking.
Priority Tracking Report
Telemetry packet which is issued with a high priority, i.e. shortly (5 seconds) after the end of a tracking frame. It will contain drift and roll information.
Priority field acquisition
Telemetry packet which is issued with a high priority, i.e. shortly after the start of every pointing. It will contain position information about the targets within the field of view.
Pulse Height Amplitude (PHA)
The detection process of a photon results in a analog signal. This signal is converted to a digital signal via an analog-to-digital converted (ADC). The digital signal (number) increases with increasing analog signal. The derived (by hardware) digital signal is the pulse height amplitude.
RA and Dec (2000)
Equatorial co-ordinate system, precessed to equinox 2000. Right ascension and declination.
Readout Time
Time it takes to read out a detector or a selected window (region of interest) thereof. Photons incident during this time interval are smeared out (usually termed "out-of-time" events).
Reference frame
The frame used as the star zero-drift position reference for the tracking process. The reference frame is a frame taken at the beginning of an exposure using the same filter wheel position as used for the scientific data.
Relationship between eV and Angstrom
Let E be the energy of a photon in eV, L its wavelength in Angstrom, f its frequency, c the speed of light, and h Planck's quantum, then with E=h*f and c=f*L, follows E = h*c/L = 12398.54/L
Relative pointing error (RPE)
The relative pointing error is the angular separation between the actual direction of an axis and a reference axis specified for a given time interval.
Rigidity Value
A term used to describe the position within the earth's magnetic field with respect to shielding against solar protons. In low earth orbit this may be used as a selection criteria for "good data " periods. There may be some equivalent measure for the XMM position within the megnetosphere.
Scatter plot
A scatter plot is a three dimensional histogram. It can be explained on the example of an image where the intensity of each pixel is shown in the vertial direction. Scatter plots can be visualized by e.g. dark scale images, LEGO plots, surface plots.
Single event upset
See latch-up event
During accumulation of each frame the CCD clocks could be shifted such that the charge of each pixel is moved to and fro from a neighboring pixel instead of keeping the charges physically stable underneath the same CCD pixel. Sloshing is used to reduce pixel noise.
In the CCDs, X-ray events may still fall on the image registers while charge from a previous imageframe is being read out. This will be particularly true for the pn CCD in which there is no shielded data store section. Thus bright point sources will produce excess events in the same CCD columns (also called "out of time events")
Solar protons
Protons from the sun may be a major source of charged particle background, and potentially of damage during flares.
Source spectral parameters
neutral hydrogen column density. Depending on galactic co-ordinate, and source spectrum may be absorbed by a varying amount of galactic material. It is well-correlated with the column density measured in this direction.
the normalization of a spectrum in photons/keV/cm²/s at 1 keV
the temperature of the hot X-ray emitting gas.
power law spectrum in energy domain can be characterised by a slope, i.e. Flux = K Energy ** (Alpha)
The histogram of number of events versus X-ray energy bin is generally termed the spectrum of the observed targets.
Split Event
X-rays being absorbed close to the border of a pixel or sufficiently deep in the CCD, may generate charge in two or more pixels. These are called split events. Split events can be classified according to the pattern of pixels which are occupied by charge from the same X-ray. See also isolated events.
Split Event Reconstruction
Split events are recognized by neiboring pixels with charges above threshold. These charges are added and the position of the event is reconstructed by a weighted mean. See also isolated event, split event.
X-ray and optical light from sources outside the nominal field of view can be scattered into the focal plane detector.
Temporal Resolution
Usually means the detector frame time, i.e. the highest resolution in time of any photon detection is limited by the finite CCD frame collection time.
Tracking Frame
Raw image collected in a tracking frame time stored in either the small word memory of the DPU. The tracking frame consists of photon counts for all detector pixels within memory windows. whether they are user requested or designated for guide stars
Tracking Frame Time
The time over which events are accumulated in the small word memory of the DPU to form a tracking frame. The frame time is constrained by the time required by the DPU to calculate a tracking solution and perform the shift-and-add. This is not a feature of the detector but of the DPU and is not user selectable.
Transient Submode/Tracking
The DPU calculates the position and intensity of the 10 (TBD) brightest sources in the defined memory windows for each of the tracking frames
Transfer Time
Time in which a selected window of a CCD is rapidly clocked to the output node(s) to be read out. It is, e.g., the time of one frame-transfer of the EPIC MOS CCD in full frame mode. Photons incident onto the CCD during this time will be smeared ("out-of-time" events).
Trapped protons
Protons trapped in the earth's magnetic field in the radiation belts will be a major contributor to radiation damage in experiment CCDs.
Zero pixel
Pixels containing no signal charge. They may in fact contain signal from dark current and other background sources such as a constant low level of light, but then this level becomes the zero reference for the measurement of desired X-ray generated charge.
Zero level peak
All the zero pixels are digitised to provide a ADU value (greater than 0). Due to noise in the systems, there will be a fluctuation about the mean level. A histogram of these values is termed the zero level peak

C.Erd, D.Lumb, R.Much
April, 1997