The X-ray energy selected by the monochromator on X31 is subject to fluctuations arising from two major sources. Firstly thermal expansion/contraction of the monochromator crystal implies a change in the lattice spacing and therefore a change in monochromatic beam energy. Secondly, a change in the angle of the electron beam orbit at the tangent point gives a change in Bragg angle for the reflection resulting also in a change of the monochromatic energy.
The behaviour of the monochromator crystal under thermal stress is generally in the hands of the user and with adequate temperature control of the crystal the former of the two effects can be minimised. However, for most synchrotron sources the electron orbit position and angle are out of the control of the user. This will mean that a beam line like X31 is prone to X-ray energy shifts even after the monochromator performance has been controlled.
The presence of an X-ray energy sensitive device in the experiment is crucial when
MAD experiments are performed at energies where the 
and 
correction terms are quickly varying with energy.
The calibrator allows an X-ray energy from a fluorescence spectrum to be accurately
selected and monitored over the course of a diffraction experiment.
The calibrator is most sensitive to changes in the monochromatic X-ray energy
on the rising and falling edges in energy space of calibrator diffraction lines.
Reflections which are very narrow in energy will be most sensitive in this case.
Accurate assessment of the X-ray energy stability can best be achieved via this method by
using two calibration lines, one in each channel, which are slightly split in energy.
In this way the falling edge of the lower energy reflection is used as well as the rising
edge of the higher energy reflection. The reference X-ray energy is set to be when the
calibrator signals from the two channels are identical and deviations of the energy from the
reference point are observed as a `see-sawing' of the two signals. Fig. 
demonstrates the technique and the effect of an energy shift on the two calibrator signals.
Figure: The technique used to monitor the X-ray energy produced by
the monochromator. At the reference energy the signal from the two calibration channels is
equal. However a small decrease in the X-ray energy has the effect of increasing the signal
in calibrator channel 1 and decreasing the signal in channel 2. The narrowest reflections
produced by the calibrator have widths comparable to the energy resolution of the monochromator
and along with a very low intrinsic noise they enable the calibrator to be sensitive to
energy shifts of 
in the best cases.
It is most convenient to use 
and 
reflections when
setting up a reference energy. For such reflections it is possible to adjust the extent
of their splitting, i.e. their relative separations, by varying the 
angle. Adjusting
the 
angle shifts both reflections in tandem, up or down in energy. In this way it
is a simple matter to monitor any energy over a continuous range limited by the calibrators
response and the beam line geometry being used. In practice there is some interdependence of
the 
and angles due to the fact that the 
reciprocal lattice vector
of the calibration crystal is not perfectly aligned with respect to the goniometers
vertical rotation axis. The means that some iteration is usually necessary before the two
angles are correctly set. This procedure may at most require ten short range scans over the
absorption edge of interest to be made for each reference energy which is set up.
For the X31 beam line this amounts to a total of 
minutes exposure for the sample
in question.
The nature of synchrotron radiation means that the calibrator signals observed at a particular energy decay steadily with time. This necessitates the use of two calibration lines. A condition attached to the use of this method is that the response of both calibrator channels must be similar over the period of a synchrotron radiation run. When the apparatus is initially set up for monitoring a particular energy the baselines on which the diffraction lines sit are at the same height, that is the background scattering in each channel is equal. If the responses of the two channels are dissimilar then as the storage ring current decays the two signals will drop at different rates. This implies that the reference points between the two calibration lines shifts giving a false impression of the X-ray energy.
To overcome this a manual method was employed to ensure that the reference point remained at the
desired position. By tuning the energy to a point away from the two calibration
lines (approx 
above or below) it
was possible to readjust the amount of background scattering in one of the channels until
it matches the other by fine tuning the photomultiplier high voltage supply. This procedure
corrects for any shift in the reference position and ensures that the energy being maintained
is the desired one. Since the X-ray detector used on this line is an on-line image
plate scanner
there is always a 
minute gap between exposures. Correcting for the shift in the
reference position takes about 
minute and can easily performed manually while
the image plate is busy. It is important to note here that the presence of a local beam shutter
immediately before the sample crystal means that the calibrator may still be used when the sample
is not being exposed.
For this particular calibrator and for typical electron current behaviour at DORIS III it is
adequate to make the above correction every 
minutes and suffer a shift
in the reference point of less than 
, although this once again depends on the width
of the calibration lines being used. The narrower the line profile in energy, the less
susceptible the reference point is to drift.
Manual correction of the X-ray energy during data collection and in between exposures is
achieved using a hand operated 5-phase stepper motor driver which is connected directly to
the monochromator `Bragg-angle' axis. In this mode the axis can be controlled with a precision
of 
steps/deg. By applying successive single step pulses in the appropriate direction to
the driver it is possible to correct the Bragg angle of the monochromator reflection and
compensate for energy drift due the thermal instability of the crystal and electron-orbit
instabilities occuring on a time scale of the order of seconds.
Reflections which are diffracted horizontally with low Bragg angles into the calibrator detector
(generally those reflections where differ significantly) have
diffraction profiles which are largely dependent on the horizontal profile of the X-ray beam reflected
off the mirror. Misalignments of the mirror appear as distortions of the line profile away from
Gaussian. For example, if the mirror were positioned to low such that the centre portion of the horizontal
fan of radiation passed unperturbed over the mirror surface, leaving only the outer regions of the fan to
reach the focus, the observed diffraction profile would exactly reflect this appearing as a camel hump -
the central part of the diffraction profile missing. Given such clues it is possible to make deductions
about the cause of any problems resulting from mirror misalignment and corrections can be made.