The MX group are in the process of updating beamlines with Eiger2 X detectors. This page provides useful information on the detectors, the file format we are using here at Diamond and how to use, view and analyse the data.
Please note there are some adjustments required for Data Processing from i03 Eiger Data from mid 2021.
|Detector||Eiger2 XE 16M||Eiger2 X 4M|
|Active area (w x h, mm)||311.2 x 327.8||155.2 x 162.5|
|Pixels (w x h)||4148 x 4362||2068 x 2162|
|Pixel size (µm2)||75 x 75||75 x 75|
|Count rate capability (ph/s/pixel)||107||107|
|Point-spread function (pixel)||1||1|
|Silicon sensor thickness (µm)||450||450|
|Data format||HDF5 / NeXus||HDF5 / NeXus|
|Frame rate (Hz)||560||560|
With Eiger2 detectors the file format and handling has changed compared with Pilatus detectors. The Eiger detectors do not write individual cbf image files, instead use the HDF5 format which is well suited to handling large data sets at high data rates.
Diamond has developed it's own file structure optimised for our network to optimise detector performance here. The structure of the data is similar to that produced by the Dectris internal file writer, but is not the same. To this end you will need to follow the instructions below to view and process data.
An example of the files produced for a 3600 frame acquisition (Given the ID of thermolysin_xtal1):
The Pilatus3 6M detector on I03 was replaced by an Eiger2 XE 16M detector in April 2019. The table below highlights the main differences between the 2 detectors.
|Parameters||Eiger2X 16M||Pilatus3 6M||Effects|
(w x h, mm)
|311.2 x 327.8||423.6 x 434.6||The Eiger is smaller than the Pilatus and the sample to detector distance will need to be reduced to achieve the same resolution. If you need better than 1.4 Å data, you will need to increase beamline energy. Use the MX Calculators to determine required energy|
Pixels (w x h)
|4148 x 4362||2463 x 2527|
|Pixel size (µm2)||75 x 75||172 x 172|
|Max Frame Rate (Hz)||560||100||Faster data collections. N.B. if you have weakly diffracting crystals and you previously needed to use long exposure times with 100% transmission with the Pilatus then you should use the same exposure times with the Eiger.|
|Data Format||HDF5/NeXus||CBF||Image viewing and processing has changed. See details below.|
Currently all data collection options are available (grid scans and rotation for MR, SAD and MAD, including line scans) with the exception of inverse beam SAD and wedge MAD.
We highly recommend fine slicing (omega 0.1o) with careful attention paid to speed and transmission (this is sample dependent - diffraction quality and radiation sensitivity and beamline dependent). On I03:
These suggestions for data collection settings only apply to the default beamline energy of 12.7 keV, or very close by. Transmission should be reduced substantially at lower energies and increased at higher to reflect the varying interaction of photons with the sample. The settings were derived with the 100 micron aperture. If you have a smaller crystal and are using the 50 or 20 micron apertures, then you should increase the transmission if you have issues with insufficient spots found for indexing. It is important to remember to reduce the total exposure time accordingly to avoid radiation damage. A smaller aperture does not reduce radiation damage to the illuminated volume, so consider using multiple crystals.
Collecting highly redundant, low dose data can be an effective way to collect high quality data, particularly if used in conjunction with random reorientation of the crystal using the multi-axis SmarGon goniometer. In this case using 1/4, 1/2, 1 X your usual transmission for 3 x 360o sweeps of a randomly reorientated crystal (chi=0/10/20) can be a useful strategy for a wide variety of use cases.
A typical example of this might be:
Examine the 3 datasets for signs of radiation damage and truncate as appropriate. You can calculate the transmission to use for a single sweep once you have done this and use that for future collections. For more information on this approach, please see How best to use photons
Using the 'Data Collection/Screening' tab in GDA, collect three images 45 deg apart in omega with the following default settings:
Results of strategies from EDNA and mosflm will be available in ISPyB and can be applied via the strategies button in the Data Collection tab of the Data Collection view. The XOalign results are also displayed in the case where a crystal alignment along a symmetry axis is possible with the SmarGon multiaxis goniometer.
Images can be viewed when at or connected to Diamond via NX with the following software. We do not recommend you make decisions for data collections from viewing the compressed jpeg images you find in ISPyB. Please make use of available strategy information or by using the full image viewers below when making plans for data collections:
DIALS image viewer
From the DLS_Launchers folder on your desktop double click the Dials Image Viewer icon to open the last collected data set.
Alternatively to look at specific data sets use the Load File option from an already open viewer or, on the command line:
> module load dials
> dials.image_viewer /path/to/image_master.h5
See DIALS image viewer manual here
The ADXV Autoload icon in the DLS Launchers has changed for the Eiger. It now launches a control window, through which you can cycle through your images, pause and unpause, and turn on/off features like resolution rings. It will automatically load image headers correctly, which won't happen through the normal adxv program (see below).
The commands are listed in the Control window, make sure the Control window is selected before pressing a key. By default, it will go to your latest data collection, and scroll through up to 30 images, at 3s intervals, with rings and weak data mode on. It will only find the latest dataset after it has finished collecting, so be patient.
On the command line:
> module load adxv
Browse to image folder and open the e.g. image_header.cbf file (in the figure below Thau_4_1_header.cbf) and then the image_000001.h5 file you wish to view (see figure), not the image_master.h5.
The image header contains all the information about the energy,distance and beam centre for your frames and needs to be reloaded whenever you open an h5 file with different settings.
Zooming in to 100% shows the image with good contrast settings automatically. Be very careful viewing images at 25 or 50% as the scaling can make small or weak spots vanish, don't judge diffraction resolution at these zoom levels.
With adxv as well as viewing individual frames, you can combine frames together using the slab fields. See figure below or page 29 of the adxv manual.
The frame 'number' is the middle "Slab" field above. In order to scroll through your dataset with the large arrow keys, you need to have the "+Slabs" box checked. The right hand side "Slabs" counter tells you how many stacked images are being displayed at a time, for example you can view a 0.5 degree image from 0.1 degree frames by setting this box to 5. Set to 1 to view individual frames.
On the command line:
> module load albula
or double click on the Albula icon in Desktop_Launchers
Open master file.
Use tools:histogram to adjust contrast (both lower and upper levels) - autocontrast currently not working. Good starting values are: Background -4.0, Foreground 4.0
GDA Image Viewer
We are working on this.
Viewing results from Grid Scans
Results of the DIALs analysis are overlaid on the crystal image in both GDA and ISPyB. To view the diffraction images click on the box in the grid scan results tab or use your preferred image viewer from the list above. We aim to have the image viewing capabilities in ISPyB as soon as possible.
Autoprocessing with fast_dp, xia2 DIALS and xia2 XDS and autoPROC all work and results are available in ISPyB.
Reprocessing your data via ISPyB works also. Instructions are here.
For manual processing with DIALS, xia2 is probably easiest -
> module load xia2
> xia2 pipeline=dials image=/path/to/master.h5
If you have a massive data set i.e. multiple turns it can be substantially quicker to process this in blocks i.e.
> xia2 pipeline=dials image=/path/to/master.h5:1:9600:1200
to process in 8 x 1200 image blocks (relates to how DIALS refinement works).
DIALS Regular manual processing pipeline as shown in the tutorial will work fine:
For manual processing with XDS xia2 is probably easiest:
> module load hdf5/1.10
> module load xia2
> xia2 pipeline=3dii image=/path/to/master.h5 plugin=durin-plugin.so
You can also generate an XDS.INP input file automatically using eiger2xds as follows (do not currently use generate_XDS - this needs updating with this new detector):
> module load dials
> module load hdf5/1.10
> eiger2xds path/to/master.h5
If you wish to process with XDS at home then use the eiger2xds command as outlined above at Diamond and then copy home the generated XDS.INP file and update
and you will need to locally install the durin-plugin - see https://github.com/DiamondLightSource/durin and adjust LIB path appropriately in your XDS.INP.
If manually creating an XDS.INP file, the specific lines needed for the EIGER detector are:
DETECTOR=EIGER MINIMUM_VALID_PIXEL_VALUE=0 OVERLOAD=65535
NX=4148 NY=4362 QX=0.0750 QY=0.0750
Example XDS.INP for I04:
Diamond Light Source is the UK's national synchrotron science facility, located at the Harwell Science and Innovation Campus in Oxfordshire.
Copyright © 2020 Diamond Light Source
Diamond Light Source Ltd
Harwell Science & Innovation Campus
Diamond Light Source® and the Diamond logo are registered trademarks of Diamond Light Source Ltd
Registered in England and Wales at Diamond House, Harwell Science and Innovation Campus, Didcot, Oxfordshire, OX11 0DE, United Kingdom. Company number: 4375679. VAT number: 287 461 957. Economic Operators Registration and Identification (EORI) number: GB287461957003.