Back to Toolbox

Running a Strategy (with Mosflm)

 

Once you have collected one or two images, it is worth trying to predict a data collection strategy using Mosflm. This will make an initial estimation of your unit cell and space group in order to predict how you can collect the most complete data. By doing this you WILL rapidly speed up your experiment whilst minimise the crystals exposure to the beam (ie giving you better data in the long run).

First you need to open a Unix/Linux shell preferably in the directory where you are saving your images and type:

setccp4

mosflm

You then need to tell Mosflm which image to look at by typing in your image name (something along the lines of):

image ./name_1_001.img

go

This should give you a window that looks like:

mos1

Spot Picking

Next you need to do pick spots for the Autoindex routine. Do this by clicking on "Find Spots" and then following the defaults. Sometimes an error message will pop up complaining that it has found too many spots and wants to "increase the threshold". This is perfectly reasonable so just click "OK". You should see a bunch of red crosses appearing on your diffraction image once it is finished.

If you have collected more than one image (e.g. at 0 and 90 degrees) you will also want to find spots on your second image. To do this click on "Read image" and then paste in the file name of your next image. It should pick up the angle from the header of the image so again just follow the defaults. You then need to pick spots as before, and again follow the defaults until red crosses appear on your second image.

Autoindexing

Once you have your spots picked, you can click on the "Autoindex" button. For the vast majority of data you will want to again accept all the defaults by hitting enter each time it asks you a question. Occassionally, if your unit cell is very large, you can give it a large "maximum cell edge". I would recommend putting in 400 if you think the default of 181 is too small. If you keep hitting enter you will eventually end up with something that look like this:

mos2

Actually choosing the correct spacegroup can depend on a lot of factors, however for a simple case as illustrated above, you want to choose the spacegroup with the highest symmetry but lowest penalty. In this case it would be the monoclinic P2 (NB it gives you a suggested solution). To select the spacegroup you type:

3 p2

Again you should hit enter to accept all the defaults. You will end up back on the screen with all the red crosses on your image. Next you want to check the accuracy of the prediction. To do this click "Clear Spots" and enter "Y". Then click the "Predict" button. This will hopefully produce a lot of blue and yellow boxes around many of your spots. Check to make sure the boxes are actually around your spots by zooming in to different sections of your image (click "Zoom" on the top right of the screen, and then draw a box using the left mouse button around the segment you want to see more clearly). Not all the spots will be boxed due to overlaps etc., but you want to make sure a good number of them are. To cancel the zoom and return back to the image double click on the zoom button.

Setting Resolution and Mosaicity

As the program is currently trying to box spots right out to the edge of the detector, you next need to limit the resolution. To do this click on the "Pick" button and then draw a box in an area where there are few spots. This should give you something that look like:

mos3

Note sometimes the pop-up window appears right at the bottom of the screen.

In the pop-up look to see the average pixel value (in the above case I am guessing about 72). This should be approximately your background value for the image. Enter this value into the "Min" box at the top right of the screen and press enter. The background of the image should get much lighter allowing you to see the spots more clearly. WHATEVER YOU DO DO NOT PRESS THE X IN THE TOP RIGHT OF THE POP-UP as this will kill the whole program. To get rid of the pop-up move it out of the way and then click on the "cancel pick" button.

You can now estimate your highest resolution more clearly by zooming into an area on the detector where the spots look to be fading outing, and clicking on the last couple spots you can see. The resolution at this point will appear in the "output" box (middle bottom of screen). Next click on the word "Resolution" in the "input box" (left of screen, almost exactly in the middle) and enter your maximum resolution. After pressing enter you should see your prediction boxes only extending to this higher resolution.

To estimate mosaicity click on "Estimate Mosaicity". This will tell you that it is "waiting" and then give you a value. Accept this value. On occasion the program complains that it cannot estimate a mosaicity value. You can therefore set one yourself by clicking on the word "Mosaic" in the input box (left of screen, approximately a quarter of the way down) and input the number 0.5. You should see more prediction boxes (probably yellow) appear. Have a look to see if the boxes are around more of the spots. If not increase the mosaicity. I wouldn't recommend a mosaicity value much higher than 1.

Finally - the Strategy!

Once you have finished playing around with the resolution and the mosaicity, click the "Strategy" button. This will give you the following screen:

mos4

Press yes to proceed, and then at the prompt type:

strategy

go

This will take a moment before giving you the following:

mos5

The key bit to notice is the section that says "Optimum rotation gives 100.0% of unique data". Two lines underneath (in this case saying "From 116.0 to 296.0") you have your ideal Phi values.

As I normally collect 0.5 degree images I next simply subtract 296 from 116 and then add 15 degrees extra (for luck!). I then divide this number (195 degrees) by my oscillation angle (0.5 degrees) to tell me I need to collect 390 images in this case. If you have a particularly large unit cell (ie the spots are closer together) OR you have very high mosaicity (above 1) then you would probably be justified in collecting 0.3 degree images. To work this out you simply divide 195 by 0.3 to tell you to collect 690 images.

Next go back to the X-ray software and enter your starting angle (116), oscillation value (0.5) and number of images (690) and away you go!

(NB if you want to close Mosflm at this point type "end" at the strategy prompt and then click on "Save/Exit". Either a box will pop up asking you if you want to save your session (click yes) or alternatively you will be returned to the Mosflm prompt. If the save box HAS NOT popped up type "save" and then exit to return to your shell prompt.)

Back to Toolbox