A Guide to Using the SHELXTL Crystallographic Software Package
XSHELL version
Gregory S. Girolami
Department of Chemistry
University of Illinois at Urbana-Champaign
25 January 2004
1. SHELXTL programs and files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. General Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3. Preparing for X-ray work-up. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
4. Setting up for the refinement – XPREP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
5. Solving by direct methods – XS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
6. Assessing the solution – XSHELL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
7. Solving by Patterson methods – XS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
8. What to do when you can’t find a solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
9. Inspecting the .ins file. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
10. Least squares refinement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
11. Checking for missed crystallographic symmetry – PLATON . . . . . . . . . . . . . . . . 24
12. Turning atoms anisotropic and adding hydrogen atoms . . . . . . . . . . . . . . . . . . . . 25
13. Correcting for absorption – XPREP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
14. Extinction, weighting schemes, and absolute configuration . . . . . . . . . . . . . . . . . 37
15. Final details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
16. Generating graphical displays – XP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
17. Generating graphical displays – XSHELL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
18. Printing diagrams generated by XP using CorelDraw. . . . . . . . . . . . . . . . . . . . . . 45
Appendix 1. Sample structure report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Acknowledgments: I thank Scott R. Wilson, Vera V. Mainz, Julia L. Brumaghim, James G. Priepot, and
Jon P. Goveia for their help and advice.
Copyright © 2003 by G. S. Girolami.
Reproduction or translation of any part of this work beyond that permitted by Section 107 or 108 of the
1976 United States Copyright Act without the permission of the copyright owner is not permitted.
2
1. SHELXTL programs and files
1.1) SHELXTL programs (click on Help – About to find the SHELXTL version number)
SHELXTL is a software package that is useful for solving and refining single-crystal X-
ray diffraction data sets. SHELXTL consists of four major programs, which are:
XPREP – Space group determination, absorption corrections, unit cell
transformations, and reciprocal space plots. XPREP reads the raw data file
filename.raw and the parameter file filename.p4p written by the diffractometer
control program, and writes the instruction file filename.ins and reflection data file
filename.hkl for use by the programs XS and XL.
XS – Structure solution by direct methods or Patterson methods. XS reads
filename.hkl and filename.ins (generated by XPREP) and writes the best solution (a
list of atomic coordinates plus other information) to the results file filename.res. A
detailed listing of the program’s activities are written to the file filename.lst.
XL – Least-squares structure refinement. XL reads filename.ins (obtained by
renaming the filename.res file generated by XS or by a previous XL refinement
cycle) and writes a new results file filename.res and listing file filename.lst.
XSHELL – Graphical user interface for structure refinement, interactive molecular
graphics, and publication-quality diagrams. XSHELL reads the filename.res from
XS or XL and (if so desired) can (re)write the filename.ins file for the next series
of refinements.
1.2) SHELXTL files
SHELXTL uses files of the type ‘filename.ext’, where filename is a name of up to eight
alphanumeric characters, and ext is a three character extension. Unless you change them,
files for a given problem will always have the same filename but different extensions. The
extensions define the function of each file and are generated or recognized automatically
by the SHELXTL programs. The file extensions used by SHELXTL are:
.raw – raw reflection data; used as input file for XPREP
._ls – statistical analysis of raw reflection data
.prp – a listing of what was done while running XPREP
.p4p – data collection parameter file; used as input file for XPREP
.hkl – reflection intensity table created by XPREP.
.ins – instruction file containing cell parameters, etc. created by XPREP;
input file for XS, XL, and XSHELL
.res – results file created by running XS or XL; updated form of .ins file
.lst – a listing of what was done in the last computer run.
1.3) Solving a crystal structure
The flowchart on the following page briefly outlines the steps necessary to solve an X-ray
crystal structure by the methods discussed in this handout. It includes everything from
X-ray data collection to printing out a plot of the newly characterized molecule.
This handout does not, however, discuss how to handle disorder, non-integral site
occupancy factors, or restraints. A separate handout is available that deals with these
issues. Consult an expert for help if you encounter these problems.
3
Data Collection X-ray detector
original.raw, original.p4p, original._ls
XPREP original
filename.ins and filename.hkl
XS filename XS filename
XL filename
Final cycles of least squares
XP absfile.res
Space Group Determination
Direct Methods (TREF) Patterson Heavy Atom Method (PATT)
Direct Methods
Fail
Patterson Methods
Fail
Least Squares Refinement
Further Refinements
- anisotropic refinement
- H atom refinements
Absorption Correction
Plot Generation
XPREP original
4
2. General Procedures
2.1) Almost all commands that you type in must be followed by hitting the
key. In this
handout, this is implicit; you will not always be reminded that hitting the key is
necessary.
2.2) NOTE: The SHELXTL software is unforgiving of mistakes in file handling – it is very
easy to corrupt the .ins file and undo all the work you have done. It is useful to make
regular backup copies of key files (especially .ins files), so you will easily be able to
recover easily from data handling mishaps. Use files given names such as filename.bak1,
filename.bak2, etc. The filename must be eight characters or shorter and must consist
only of letters and numbers.
2.3) Some editing of text files will be necessary. This guide will assume that your text editor is
Wordpad; below is a summary of some of the keyboard-activated commands used in this
text editor:
↑, ↓, ←, and → keys Moves cursor around the document
and Deletes characters
f Opens File pull-down menu; commands listed below:
o Open file
s Save file
x Exit file
p Print file
e Opens Edit pull-down menu; commands listed below:
u Undo last action
a Select all text
t Cut selected text
c Copy selected text
p Paste selected text
l Delete selected text
To highlight a certain portion of text to edit, hold down the key and use the ↑, ↓,
←, and → keys to move the cursor. Alternatively, you may use a mouse.
3. Preparing for X-ray work-up
3.1) This guide assumes that the diffraction data you will solve has already been transferred to
your computer, and that all files are stored in a folder on the desktop called Data. To
view a list of files in your folder, double click on the Data icon.
3.2) Look in your Data folder and make sure that the original.raw and original.p4p files are
present, where original is a name that has been assigned to the files by the
crystallographer who collected the data (e.g., w25r).
The .raw file is a list of the reflection intensities sorted by Miller index. The .p4p file (and
._ls file, if present) contain summaries of the data collection parameters and include the
unit cell constants.
3.3) Next, you must tell the software which file contains your data. You do this by defining a
“project”. First, open the SHELXTL window by double clicking the SHELXTL icon.
3.4) Start a new project by clicking Project, dragging to New, and then releasing.
5
3.5) In the window labeled “Look in:”, navigate through the folders (if your Data folder isn’t
already selected) until you see the original.raw file displayed in the window.
3.6) Type in a name for the project and write it down - we recommend using the same name as
the name of the .raw file.
3.7) Then highlight the original.raw file with your mouse, and click Open.
3.8) You should now see a window that has your project name and path.
(If in the future you wish to see this window, click Project in the SHELXTL window, drag
to Open, and then release. Highlight the appropriate project, and then click Open.)
3.9) Go to Section 4.
4. Setting up for the refinement – XPREP
4.1) Now you are ready to run XPREP, which is a program that will help you decide on a
space group and set up the files necessary to solve your structure. With the SHELXTL
window open, click on XPREP.
4.2) If XPREP asks you to enter the cell parameters (a b c α β γ), simply type the appropriate
values for a b c α β γ with spaces in between.
Normally, however, XPREP finds these values in the .p4p file and you do not need to type
them in. Go to step 4.3.
4.3) XPREP will next read and analyze the diffraction data in the .raw file. The first analysis is
a table of Lattice Exceptions. The table is headed P A B C I F Obv Rev All,
where P = primitive cell, A B C are end-centered cells, I = body-centered cell, F = face-
centered cell, and Obv and Rev are for trigonal and hexagonal cells.
4.4) Look at the bottom number in each column, which gives the average value of the
intensities for that lattice condition divided by the error (mean int/sigma). If, for any
column, the value in the bottom row is 3 or less (or significantly smaller than the values in
all the other columns), the condition is a likely one for your crystal (the smaller the
number, the better).
4.5) Look to see what XPREP has written after the words “Select option.” If one of the letters
P, A, B, C, I, or F is enclosed in brackets, write down the lattice condition for future
reference. Then hit to accept XPREP’s choice. Skip to step 4.7.
4.6) If a question mark is enclosed in brackets [?], this means that XPREP is having trouble
recommending a lattice condition. If this happens, look to see which lattice condition has
the smallest number in the last line (smallest value for I/σ). Write down this choice for
future reference.
Type the letter (P, A, B, C, etc.) of the column that has the smallest mean int/sigma
value, unless you know from an earlier attempt to solve the structure that this choice may
not be correct. Then go to step 4.7.
4.7) You should now be back in the XPREP main menu, and the suggested option should be to
look for higher symmetry [H]. Hit to accept this option.
6
4.8) XPREP will first print out some information that you can ignore, and then it will display
one or more options for the crystal system; each choice is printed in between ------------
dashed lines. (The seven possible crystal systems are triclinic, monoclinic, orthorhombic,
tetragonal, trigonal, hexagonal, and cubic).
4.9) If there is only one option between the dashed lines, write down the crystal system of that
option for future reference. Then skip to step 4.11.
4.10) If there is more than one option between the dashed lines, record the two options with the
lowest FOM values (FOM = figure of merit; note: triclinic always has FOM = 0). Ignore
the comment below the table about the original cell.
4.11) If XPREP makes a recommendation for a crystal system, the recommendation will be
enclosed in brackets, e.g. [A], which stands for choice A. Write down the crystal system
for that choice for future reference. Hit to accept its choice unless you know it to
be wrong, and then skip to step 4.13.
4.12) If XPREP has difficulty choosing a crystal system for the unit cell (it will show [?] instead
of a letter), it is best to pick the highest symmetry crystal system listed among the possible
options. For example, pick orthorhombic over monoclinic, and monoclinic over triclinic,
unless you know that the higher symmetry choice does not lead to a solution of the data.
Write down the crystal system of your choice for future reference. Then type in the
appropriate letter of the best option, hit , and go to step 4.13.
4.13) After the search for higher symmetry is done, you will be back in the main menu and the
suggested option will be [S] to enter the space group determination routine. Hit .
4.14) The next suggested option will also be [S] to determine the space group. Hit .
4.15) XPREP will then suggest a crystal system (triclinic, etc.). Choose the same crystal system
as you did in step 4.11 or 4.12.
4.16) XPREP will then go through the lattice condition routine and make a suggestion
(primitive, etc.). Choose the same option as you did in step 4.5 or 4.6.
4.17) Next, XPREP will give a table of possible space groups. If at least one space group
appears in the table, skip to step 4.20. If no space groups appear in the table, go to step
4.18.
4.18) If no space groups appear in the table, this means that XPREP has been unable to find any
space group that agrees with the systematic absences it finds. If you have previously
changed the tolerances, go to step 4.19.
If you have not previously changed the tolerances, do the following:
Return to the XPREP main menu by hitting .
Type T to change the tolerances.
Type G and make the minimum I/sigma gap half of its current size.
Type A and make the maximum mean I/sigma twice its current size.
Type E to exit to the XPREP main menu.
Go to step 4.7.
4.19) In some cases, XPREP will refuse to identify a space group even after adjusting the
tolerances, and you will have to decide yourself what the systematic absences are.
7
Look at the bottom two entries in each column of the systematic absence exceptions table.
A systematic absence is likely if mean intensity or mean int/sigma is significantly less
than the values in the other columns.
From the probably systematic absences, you should be able to guess a space group.
Consult an expert if in doubt.
Enter the space group manually by typing its symbol; use minus signs and parentheses to
indicate rotoinversions and screw axes, respectively: thus P-1 or P2(1)/c.
Then go to step 4.26.
4.20) In the table of possible space groups, the last number in each row is the CFOM number
(which stands for combined figure of merit) for each space group. Usually, the smaller the
CFOM, the better.
If CFOM is 2 or less, the space group choice is a good one (although not necessarily the
right one!). If it is 12 or greater, you will probably have problems solving the data set in
that space group.
4.21) Write down the three or four space groups that have the lowest values of CFOM (if there
are that many). Record the CFOM values for each.
4.22) Look at the space group with the lowest value of CFOM. If it is a centrosymmetric space
group (i.e., if it is labeled ‘centro’ in the Type column), then skip to step 4.26.
4.23) If the space group with the lowest value of CFOM is chiral or non-centrosymmetric, then
answer the following two questions:
a) Of the choices available in the table on the screen, is there a centrosymmetric
space group whose CFOM is within 3 units of the CFOM of the “best” space
group?
b) Does your crystal contain heavy scatterers (for Mo radiation, atoms with Z
greater than 18)?
4.24) If the answer to either question in step 4.23 is “no”, your space group is probably chiral or
non-centrosymmetric. Skip to step 4.26.
4.25) If the answer to both questions in step 4.23 is “yes”, it is likely that the actual space group
is not chiral or non-centrosymmetric, but is the centrosymmetric space group with the
lowest CFOM.
Skip to step 4.28 and override XPREP’s choice with the centrosymmetric space group. If
it proves impossible to find a solution in the centrosymmetric space group, you can always
re-run XPREP to choose a different space group.
4.26) Above the table of systematic absence exceptions is a number, labeled Mean |E*E-1|,
which has been calculated from the data in the .raw file. Answer the following question:
is the value of Mean |E*E-1| closer to the value calculated for centrosymmetric space
groups or for non-centrosymmetric space groups?
4.27) If the space group with the lowest CFOM is centrosymmetric, and the Mean |E*E-1|
value is closer to that expected for a centrosymmetric space group, then your space group
is almost certainly centrosymmetric. Skip to step 4.30.
8
4.28) If the space group with the lowest CFOM is chiral or noncentrosymmetric, and the Mean |E*E-1| value is closer to that expected for a noncentrosymmetric space group, then your
space group is almost certainly chiral or noncentrosymmetric. Skip to step 4.30.
4.29) If neither situation in steps 4.27 or 4.28 is true, your space group is probably
centrosymmetric.
For your information, the Mean |E*E-1| value becomes an unreliable indicator of whether
the cell is centrosymmetric or not in the following cases: (a) heavy scatterers (atoms with
Z > 18) are present, or (b) when your data crystal is twinned.
Be prepared to consult an expert if you have trouble solving the structure. For now,
however, go to step 4.30.
4.30) If XPREP shows a question mark in brackets [?] as its Select Option, skip to step 4.32. If
it gives a recommended space group in brackets: e.g. [F] for option F, go to step 4.31.
4.31) If you like XPREP’s recommended choice, accept it by hitting and go to step
4.33.
4.32) If you wish to choose a different space group than the one recommended by XPREP (or if
XPREP can’t decide and shows [?] as its choice), type the letter corresponding to the
option you think is best, then hit .
4.33) After you have chosen a space group, a screen will come up with a table of information
about the current data set.
Write down the space group, and from the line that reads “current cell”, write down the
values of a, b, c, α, β, γ, and the volume of the unit cell V, which appear in that order.
4.34) XPREP will next suggest option D to merge data sets. Override this choice by typing C.
(If you accidentally hit and accept option D, type E to return to the main menu
and then type C.)
4.35) The program will then display a chemical formula for the compound or ask you to provide
one.
If the formula is not present, enter it manually (the program understands abbreviations
such as Me, Et, Ph, etc.).
If the formula is incorrect, type F and enter the correct formula.
4.36) Look