Cell determination using Denzo
In a cell determination experiment there are three buttons in the
experiment window. You can either run through all 3 of them in sequence,
or just press the "determine cell" button which will run through the
whole process automatically. Each button will turn green when the
associated action has been completed.
A run-through for an index experiment.
For this example, we will go through a cell determination in an empty
directory. If there are already files in the directory, some default
values might be changed, or questions might pop up in a different order.
Click on "Determine cell".
Click on "Strategy".
The strategy for determining the unit cell is very simple: we need to make a
single short "rotation". The parameters can be filled in on the window that
appears next:
The two parameters are:
- The crystal to detector distance (dx)
-
A crystal to detector distance of 40 should be OK
for cells with axis lengths upto about 40 A. If there are longer
axes, it will be difficult to separate the spots. For determining
the unit cell it is more important to be able to separate all
spots than it is to have many spots, hence it is not so bad to be
a bit further away than strictly necessary. The default can be
changed in the configuration parameters.
- Total scan length
-
For a small molecule structure a rotation of 10 degrees will normally
give enough reflections to find the unit cell. For minerals or other
extremely small unit cells, a larger value might be needed. For
proteins and other large structures 1-2 degrees might be sufficient.
The "More options" button gives one more question:
- Datacollection Strategy
-
Normally the cell can be determined from a phi scan. If an
unfortunate mount of a problematic crystal makes it impossible to
find the unit cell from a phi scan, a "strategy" consisting of an
omega scan at kappa=60 degrees can be chosen instead. You might
have better luck there... Under cryogenic conditions, this strategy
can be used as well to prevent icing of the phi motor.
Press "Ok" to calculate the appropriate scan.
Another dialog box will pop up, requesting the host name of the
Windows PC that controls the goniometer. A connection to the hardware
is required at this moment to let "collect" make sure that the scan it
proposes can be made within the limitations of the hardware.
If you have only one goniometer, the default machine name should be
correct, and all that is required is to press the button labelled "OK"
to make the connection. The default name can be changed by setting the
"ccdhostname" variable in the configuration file
Press "OK" to make a connection to the hardware, and to start calculating the
scan. The resulting scan will show up in the scan-set window.
Press "OK" to accept the scan. [Please note that although it is possible
to add more scans to the scan set now, this will confuse the collect program
later on. It is safe to use "Add operations" for any other operations like
those controlling the cryostat and generator].
The scan parameters window will appear next:
- Scan angle per frame
-
This is set by default to 1.0 degrees.
If you expect a very small unit cell, this can be increased to
2.0 degrees (or even larger, but with greater care. Not much
is gained by using larger scan angles).
- Integration time
-
-
This will be set by default to 20 seconds per degree.
This can be increased in a second attempt if the diffraction
is too weak to find the unit cell.
Under the "More options" button, you will find:
- Filename pattern
-
The filename pattern should contain two sets of "#" characters (at
least 2 # characters in the first set, and 3 in the second). The first
set of # will contain the scan number from the scan set, the second
set will contain the frame number. The default value is "i##f###.kcd";
the "i" indicating that the images are part of an "index"
experiment. If "i" images already exist in the current directory, the
default will change to "j" (if required to "k", etc) automatically.
- Number of iterations
-
The number of times each image is measured. Normally this value is 2,
for so-called de-zingering of images. If the exposure time is very
long (more than an hour), a triple image might be needed to get rid of
all zinger pixels completely (but please note that triple images can
not be used in denzo). For an index experiment do not lower this value
to 1, because it might confuse the peak searching process.
- Number of the first scan
-
The first scan number (the number on the first set of ## marks in the file
name pattern) is set to 1 by default. It should not be changed for an
index experiment.
Data collection for an index scan will generally take only a few minutes...
The buttons at the bottom of the main application window will turn
green when each step of the process has been successfully completed.
The resulting cell will be displayed on the screen for your information.
If the process failed, a message to that effect will be shown. The best solution
is to run the HKL gui, and try to find the proper cell manually.
If the Cambridge Structural Database (CSD) is installed on your computer, and
collect knows about it, the "Find
CSD Hits" button will run the csdcell program
for you:
