Crystals Primer

Chapter 2: Example Of A Simple Structure

2.1: Script mode
2.2: Starting CRYSTALS
2.3: Command mode


CRYSTALS can be run in two modes:

COMMAND-LINE in which the user enters commands in the order he chooses.
SCRIPT, in which CRYSTALS tries to help by asking questions.

This example will illustrate how to take reflections and a trial structure from another system and get them into CRYSTALS, and then start refining the structure. The data is on the distribution media in a subdirectory NKET. You should create a subdirectory and then copy the data into it. The directory NKET contains files for a simple structure determination. Some of them are for Command mode, others for Script mode.


[Top] [Index] Manuals generated on Wednesday 27 April 2011

2.1: Script mode

The file to be used is:

      NKET.DOC      A brief description of the structure
      NKET.REF      Pre-processed reflection in SHELX format

The general information about the anaylsis is:

      Colourless crystal, .2x.2x.2 mm
      Molecular formula C8 H8 N1 O4
      Space Group P 41, 4 formula units in cell

          OH\        /OH
              Ch    Ch
            /  \  /    \
          /      CH     CHH
      OH-CH      |      |
          \      N      CHH
            \ /     \ /
             CHH     C=O

In the following example, the left column summarises the question CRYSTALS asks, the middle column is the users response, and the right column are comments. User responses shown in square brackets [YES] indicate a response chosen from a menu list. DO NOT include the '[]'. <cr> represents 'ENTER' or 'RETURN'.


[Top] [Index] Manuals generated on Wednesday 27 April 2011

2.2: Starting CRYSTALS

On a PC, type CRYSTALS or double click the CRYSTALS icon.

Summary of Screen             User's response         Comments

The CRYSTALS banner, an
invitation to use SCRIPTS     [YES]
Type of analysis              [NON-ROUTINE]
The main menu                 [NEWSTART]     Starting a new structure.
Structure title               'Nket at 0 C'  Some text as a title.
The newstructure menu         [QUICKSTART]   Startup for SHELX reflections.
Input LIST 30                 [YES]          These are general items.
A series of general questions                If there is a value in [] at
                              <cr>           the end of a question, this is
                                .            the default value. Use 'ENTER'
                                .            to accept default, or enter a
                                .            real value.
                              <cr>           If the line ends [-], there is
                                             no default. You MUST give an
                                             answer to continue.
Number of formula units [-]   4              This question must be answered.
Colour                        Clear
Shape                         Cube
Cell Parameters [-]           7.533          No defaults available.
                              7.533 15.78    You can often insert all the
                                             answers at once.
Cell angles [90]              <cr>           Accept default.

Space group                   P 41           Separate each component by <space>.

Experimental conditions file  [NO]
Radiation                     [CU]
Friedels Law                  [YES]
Crystal Orientation           [NO]
Cell Contents                 C 32           4 formula units in this case.
                              H 32           Don't forget the spaces.
                              N 4 O 16       Type ahead.
                              END            Use 'END' to end input.
Refinement against Fsquared?  [NO]           You could choose YES.
Input Cell esds, List 31      [YES]
Cell esds [.001]              <cr>           Accept the defaults, or insert
                               .             real values.
Reflection file name  [-]     nket.ref
Data type                     [FSQ]          F-squared data.
Change any LISTS              [NO]
Free Format data?             [NO]
Standard format?              [YES]
Sigma values                  [YES]
Format [(3F4.0,2F8.0)]        <cr>           If you change it, don't forget
                                             the ( ).
Merge?                        [YES]
Sigma threshold               3.0            All reflections are stored, and
                                             the level can be reset later
                                             if required.
Sort/Merge [YES]              <cr>           Remove systematic absences and
                                             merge equivalents.
SHELX atoms [NO]              <cr>
Solve [YES]                   <cr>
Method [SIR92]                <cr>
First try [YES]               <cr>           If the structure doesn't solve,
                                             choose [STRUCTURE] from main menu,
                                             and when you get here, try [NO].
                                             CRYSTALS will prepare a more
                                             intensive job


Crystals now closes down. To run SIR92, type SIR92. When SIR shows a menu, click QUIT until it closes down. Restart CRYSTALS.

Use SCRIPTS                   [YES]          Return to SCRIPT mode
Structure type                [NON-ROUTINE]
Main Menu                     [STRUCTURE]
Structure Sub-menu            [INEMAP]       Input results of direct methods
How solved [SIR]              [SIR]
Version                       [SIR92]        The SIR92 atoms are read in
See the structure?            [YES]          In graphics, edit the structure
                                             using 'rena(me)' to change atoms.
                                             Remember each instruction needs
                                             two <cr>. Use 'END' to end.
Save changes [YES]            [YES]
Structure fragmented?         [NO]
Resequence [YES]              [YES]          Renumber atoms so that adjacent
                                             atoms have sequential numbers.
Compute over all scale        [YES]

                              [BACK]


This is the end of this branch of the process. A message suggests the next step. If all atoms were revealed in the e-map the next step is least squares, other wise try Fourier Maps. In this case the structure is OK, so you can jump directly to the refinement section below.

                              [MAPS]
                              [AUTO]         CRYSTALS chooses the 'best' map
                                             The atoms may be resequenced to
                                             include newly-found ones, and
                                             assembled into a molecule.
Reject by rho                 [YES]
Resequence atom numbers       [NO]
Collect into molecule         [NEW]          Only apply symmetry to newly found
                                             peaks.

Graphical mode                [YES]
                                             Relabel Q to real atoms, etc. END
                                             to end.
Save the changes              [YES]
Repeat Fourier cycle          [NO]
                              [BACK]         Back to main control menu.


Refinement. If the structure is routine, least squares will finish the job.

                              [LEASTSQ]
                              [AUTO]         CRYSTALS chooses the next task on
                                             the basis of R factor and
                                             previous history.
R factor intermediate
Refine x's & Uiso online?     [YES]
How many cycles               [5]            If 5 is too many, CRYSTALS will
                                             stop anyway.
Refine x's Uaniso online
How many cycles               [5]

Locate H atoms                [YES]
Info or continue              [NO]
Delete existing H atoms       [YES]          It won't matter if there are none.
                                             CRYSTALS computes structure factors,
                                             computes potential H sites and does
                                             a difference map. The computed H
                                             should be near found ones.
Graphics                      [YES]
                                             Tidy up the map. Delete computed if
                                             you prefer the found, otherwise
                                             ignore the found. They will be
                                             eliminated later. Change any found
                                             peaks to H if there is no
                                             equivalent computed (eg H bonds).
                                             Some OH H should appear.
                                             'END' to end.
Save changes                  [YES]
Keep Geometric                [YES]
How many cycles                [5]
Continue                      [YES]
Locate Hydrogens              [YES]          If there still are some missing.
Delete existing H atoms       [NO]           We are looking for additional ones.
Graphics                      [YES]          Try to find H amongst soup of peaks.
                                             Try:
                                    DIST (click O and q) and rename
                                 or PACK CELL COMPLETE INCL CELL VIEW
                                    ALONG AXIS A VIEW
                                    CONNECT HBOND O 0.0 3.0 VIEW
                                    EXCLUDE AREA <cr><cr>
                                             (use mouse to exclude some
                                             of the diagram - left button to
                                             mark corner of polygon)
                                    VIEW
                                    EXCLUDE ....
                                    UNPACK   (otherwise CRYSTALS will get all)
                                    END
Save                          [YES]
Geometric                     [YES]
Refine                        [5]
Locate H atoms                [NO]           We have them all.
Check for exinction           [YES]          The program explains how.
Check refinement weighting    [NO]           No, since we just included
                                             extinction.
Number of cycles              [5]

Locate hydrogen atoms         [NO]           If they are not all found by now,
                                             ask an expert.
Check refinement weighting    [YES]          The program explains how.
Type                          [OPTIMAL]
Change the final parameter    [NO]           The 3 parameter scheme is fine.
How many cycles               [5]
Locate Hydrogen               [NO]
Cycles                        [5]
Continue refinement           [NO]           Now choose CYCLE off LSQ menu.
See the structure             [YES]
                                             Use ELLIPS TYPE ALL 4 VIEW to
                                             check the adps.

Continue refinement           [NO]           Now choose CYCLE off LSQ menu.
                              [CYCLE]        Cycle until full convergence.
                              [BACK]         To main menu.

                              [PUBLISH]      Select what you want, eg CIF.
                              [FINISH]


Very few real analyses will run exacly like this.

Potential Problems

Fourier recycling does not complete the structure. Try some least squares before more Fourier recycling.

Some atoms are 'physically unreasonable'. Check the atomic type assignments. Many atoms are physically unreasonable. Incorrect data reduction, serious absorption.


[Top] [Index] Manuals generated on Wednesday 27 April 2011

2.3: Command mode

The files to be used in Command mode are:

NKET.REF    Reflections in SHELX format
NKET.QCK    Initial data, including a set of trial coordinates
NKET.LSQ    Isotropic refinement and molecule assembly
NKET.ANI    Anisotropic refienement
NKET.FOU    Difference Fourier


In Command mode, the user can enter commands and data directly from the keyboard, or read them in from pre-prepared files.
Start CRYSTALS as above, but answer NO in reply to the question about SCRIPT mode. You will get the CRYSTALS prompt '!'or '>'. Enter the following commands in the order given each time you are prompted. Note that interactive commands can be mixed with the use of pre-prepared files, and that individual SCRIPTS can be activated. You can print a listing file (132 characters wide) when you exit from CRYSTALS. The file will have the extension .LIS or .LIn.

      !\USE NKET.QCK
      !\CAMERON
      !END            (In CAMERON, type MENU ON <cr><cr>
                        Type END <cr> to end the graphics)
      !\USE NKET.LSQ
      !\USE NKET.ANI
      !\USE NKET.FOU
      !\SCRIPT PLOT  (Rename some of the Q to H atoms)
      !\PERHYDRO
      >END
      !\LIST 12
      >BLOCK SCALE O(1,X'S,U'S) UNTIL C(9)
      >END
      !\SFLS
      >REFINE
      >REFINE
      >END
      !\FINISH
                                          Note that CRYSTALS commands are
                                          the same for all computers


The display on the terminal will look something like:


      Issue  10 for PC486 under DOS 6.0   September 1996
       Copyright Chemical Crystallogaphy Laboratory, Oxford

        To get help, type \HELP HELP
        To end, type      \FINISH

 ************************************************************
  A new direct access file has been created on unit 1 -- Data
 ************************************************************

     CCCC  RRRR   Y   Y   SSSS  TTTTT   AAA   L       SSSS
    C      R   R  Y   Y  S        T    A   A  L      S
    C      R   R  Y   Y  S        T    A   A  L      S
    C      RRRR    YYY    SSS     T    AAAAA  L       SSS
    C      R R      Y        S    T    A   A  L          S
    C      R  R     Y        S    T    A   A  L          S
     CCCC  R    R   Y    SSSS     T    A   A  LLLLLL SSSS

( YES NO )
 Do you want to use SCRIPTS? (YES or NO) [ NO ] : no           <==
 !\USE NKET.QCK                                                <==

    0001*  \USE NKET.QCK
    0002*  \QUICKSTART
    0003*  CELL 7.5330  7.5336 15.7802
    0004*  SPACEGROUP P 41
    0005*  CONTENT C 32 H 32 N 4 O 16
    0006*  DATA 1.5418 FSQ
    0007*  FILE CRDIR:NKET.REF
    0008*  FORMAT (3F4.0, 2F8.2 )
    0009*  END

Quick startup ends            Version  1.03
Disc file extended by   0 record(s) (      0 integer words )
**** Disc file on unit   1 extended by     1 records


The two line with the arrow, <== , contain the user input. Other input is drawn from the specified files. The data in those files could have been typed into CRYSTALS directly at the '!' prompt, so that depending on the amount of input and the users confidence, either or both input methods may be used. The following pages explain the contents of the files.

NKET.QCK
This contains the basic data, the name of the file containing the reflections and a trial model obtained from somewhere else.
 \ This is the startup file for NKET
 \QUICKSTART
 CELL 7.5330  7.5336 15.7802
 SPACEGROUP P 41
 CONTENT C 32 H 32 N 4 O 16
 DATA 1.5418 FSQ
 FILE CRDIR:NKET.REF
 FORMAT (3F4.0, 2F8.2 )
 END
 \LIST      5
 READ NATOM =     13
 ATOM C       1 X=0.819200   0.697300   0.118000
 ATOM C       2 X=0.502300   0.661800   0.030100
 ATOM C       3 X=0.809300   0.322200   0.053100
 ATOM C       4 X=0.739400   0.046200   0.091800
 ATOM C       5 X=0.936400   0.463600   0.025800
 ATOM C       6 X=1.059800   0.461400   0.171800
 ATOM C       7 X=0.836400   0.629600   0.033900
 ATOM C       8 X=0.634800   0.560100   0.004300
 ATOM C       9 X=1.106300   0.435500   0.091400
 ATOM C      10 X=1.180100   0.263500   0.070100
 ATOM C      11 X=1.048600   0.109000   0.096000
 ATOM C      12 X=0.857800   0.160100   0.077100
 ATOM C      13 X=0.630000   0.381500   0.040500
 END
 \SFLS
 SCALE
 END
 \ANALYSE
 END


The ANALYSE instruction gives an analysis of residuals, and shows the R factor (20%) for the model after Fo has been scaled to Fc.
 

NKET.LSQ
This file sets up isotropic refinement, and requests several cycles of least squares. The COLLECT assembles the atoms into a molecule.

 \LIST 12
 FULL X'S U[ISO]
 END
 \SFLS
 REFINE
 REFINE
 CALC
 END
 \COLLECT
 END
 \EDIT
 SORT U[ISO]
 END
 \DISPLAY HIGH
 END
 \MOLAX
 ATOM FIRST UNTIL LAST
 PLOT
 END


The sort on U[iso] will help you decide which atoms to change from carbon, and the MOLAX (molecular axes) produces a simple plot on the terminal.
 

NKET.ANISO
This file starts by changing the selected atoms, and then doing anisotropic refinement.

 \EDIT
 RENAME C(1) O(1), C(2) O(2), C(4) O(4), C(6) O(6)
 REN C(3) N(3)
 EXECUTE
 END
 \LIST 12
 FULL X'S U'S
 END
 \SFLS
 SCALE
 REFINE
 REFINE
 END



 

NKET.FOUR
This file computes a difference Fourier, and then collects the new peaks to chemically suitable positions.

 \FOURIER
 REFLECT  CALC=YES
 MAP TYP=DIFF
 PEAK HEIGHT=2
 END
 \PEAK
 END
 \COLLECT
 SEL TYPE=PEAK
 END
 \MOLAX
 ATOM FIRST UNTIL LAST
 PLOT
 END



 



© Copyright Chemical Crystallography Laboratory, Oxford, 2011. Comments or queries to Richard Cooper - richard.cooper@chem.ox.ac.uk Telephone +44 1865 285019. This page last changed on Wednesday 27 April 2011.