automatic solution of crystal structures : SIR92 & SIRPOW92
Hi,
To all the crystallographers and others, I am a very satisfied
user of two programs (SIR92 and SIRPOW92)for automatic solution
of crystal structures by direct methods, with a very nice graphic
interface.
The program for single crystal data produces (often with all the
atoms assigned and R<10%) input for two commonly used programs
for structure refinement SHELX or CRYSTALS
For more information see below.
Davide Proserpio
To obtain a copy please contact Dr Cascarano providing your ftp
address.
Dr. Gianluca Cascarano
Istituto di Ricerca per lo Sviluppo di Metodologie
Cristallografiche CNR, c/o Dipartimento Geomineralogico,
Campus Universitario, 70124 Bari, Italy
phone +39-80-242624, FAX +39-80-242591
e-mail cryst -AatT- arba.ba.cnr.it
*****************************************************************
SIR92 - A PROGRAM FOR AUTOMATIC SOLUTION OF CRYSTAL STRUCTURES BY
DIRECT METHODS.
A. ALTOMARE, G. CASCARANO, C. GIACOVAZZO, A. GUAGLIARDI
Istituto di Ricerca per lo Sviluppo di Metodologie
Cristallografiche CNR, c/o Dipartimento Geomineralogico, Campus
Universitario, 70124 Bari, Italy,
M.C. BURLA, G. POLIDORI
Dipartimento di Scienze della Terra, Universita', 06100
Perugia, Italy,
M. CAMALLI
Ist. di Strutturistica Chimica "G. Giacomello", CNR, Via
Salaria Km 29,200, 00016 Monterotondo stazione (Roma),
Italy.
THE CRYSTALLOGRAPHIC PROBLEM: Our recent developments in direct
methods theory and in crystal structure refinement persuaded us
to develop a fully automated structure-determination program
which is able to solve the phase problem and to refine atomic
parameters in a single computer run.
METHOD OF SOLUTION: The method for the automatic structure
determination is based on the representation theory (Giacovazzo,
1977, 1980). 1-phase structure seminvariants and 3-phase
structure invariants are estimated via their second
representation, 2-phase seminvariants and 4-phase invariants via
their first representation. SIR92 is the heir of SIR88 (Burla,
Camalli, Cascarano, Giacovazzo, Polidori, Spagna & Viterbo, 1989)
of which it retains most of the features (f.e., treatment of
pseudotranslational symmetry, use of prior information, tangent
weighting schemes, etc.). In SIR88 1-,2-phase seminvariants and
positive estimated 3-phase invariants could be actively used in
the phasing process. Negative estimated triplet and quartet
invariants were used as a FOM together with psi-zero triplets. In
SIR92 negative estimated triplet and quartet invariants and
psi-zero triplets are actively used in the phasing process, which
proved to be much more robust and efficient. The various trial
solutions can be obtained either by magic integer permutation
(Main, 1978) or by random phases (Baggio, Woolfson, Declercq &
Germain, 1978). The best trial solution, selected by powerful
FOM's, is automatically processed through a cyclic procedure
combining structure factor calculation - least squares - 2Fo-Fc
Fourier synthesis. The final outcome is a set of refined atomic
parameters (x,y,z and isotropic B values) associated with
suitable atomic species. The residual R value is calculated for
user usefulness (final values usually between 0.08 and 0.15). The
program runs in default when supplied with a minimum amount of
information (space group symbol, cell parameters, unit cell
chemical content and reflections) but documentation has been
produced to allow the user to change default values. If a graphic
device is available the user can follow structure solution and
refinement on the screen. In the final stage a menu-driven
interface is available in order to study molecular geometry and
restart refinement.
SOFTWARE ENVIRONMENT: The program has been written in standard
Fortran77. A module written in C language is supplied in order to
interface the program with X-window or DEC-Window terminals for
graphics. Therefore, besides the C compiler, an Xlib library is
also needed. Two ASCII files are associated to the program: the
first contains coefficients for calculating scattering factors,
the second, necessary for graphics, contains on-line help.
HARDWARE ENVIRONMENT: The program runs on UNIX and DEC
workstations, on mainframe and on personal computers (for this
last one 4 Mbytes of RAM and a VGA monitor are needed).
PROGRAM SPECIFICATION: The source code of the program consists of
41.000 FORTRAN lines and of 2000 C lines (including comments).
DOCUMENTATION: Users instructions and program description (about
1500 lines) are supplied as an ASCII text file.
AVAILABILITY: The program is available free of charge from the
teams in Bari and in Perugia. A licence agreement has to be
signed.
REFERENCES
Burla, M.C., Camalli, M., Cascarano, G., Giacovazzo, C.,
Polidori, G., Spagna, R. & Viterbo, D. (1989). J. Appl. Cryst.
22,389-393.
Baggio, R., Woolfson, M.M., Declercq, J.P. & Germain, G. (1978).
Acta Cryst. A34,883-892.
Giacovazzo, C. (1977). Acta Cryst. A33,933-944.
Giacovazzo, C. (1980). Acta Cryst. A36,362-372.
Main, P. (1978). Acta Cryst. A34,31-38.
*****************************************************************
SIRPOW.92 - A PROGRAM FOR AUTOMATIC SOLUTION OF CRYSTAL
STRUCTURES BY DIRECT METHODS OPTIMIZED FOR POWDER DATA.
A. ALTOMARE, G. CASCARANO, C. GIACOVAZZO, A. GUAGLIARDI
Istituto di Ricerca per lo Sviluppo di Metodologie
Cristallografiche CNR, c/o Dipartimento Geomineralogico, Campus
Universitario, 70124 Bari, Italy,
M.C. BURLA, G. POLIDORI
Dipartimento di Scienze della Terra, Universita', 06100
Perugia, Italy,
M. CAMALLI
Ist. di Strutturistica Chimica "G. Giacomello", CNR, Via
Salaria Km 29,200, 00016 Monterotondo stazione (Roma),
Italy.
THE CRYSTALLOGRAPHIC PROBLEM: Solving crystal structures from
powder data is not a routinary job: peak overlapping, preferred
orientation and difficulties in background estimation make
uncertain the evaluation of the diffraction intensities for the
various reflections. Consequently both Patterson and direct
methods lose efficiency and structure solution is not
straightforward. The first aim of this program is to optimize
direct methods for powder data.
METHOD OF SOLUTION: The standard input of SIRPOW.92 consists of
lattice constants, space group, chemical content of the unit
cell, and a file of reflections as provided by current peak
decomposition programs. Neutron data as well as X-ray data can be
managed. The overlapping conditions, chosen by the program or
fixed by the user, are stored for subsequent use. For example,
/E/'s of some "important" overlapping reflections can be suitably
varied in order to generate different phasing pathways. As in
SIR92 (Altomare, Cascarano, Giacovazzo, Guagliardi, Burla,
Polidori & Camalli, 1994) 1-phase and 2-phase structure
seminvariants and 3- and 4-phase structure invariants are
estimated via proper representations (Giacovazzo, 1977, 1980).
Trial solutions are obtained by the magic integer approach (Main,
1978) or by random phases (Baggio, Woolfson, Declercq & Germain,
1978). The best solution, selected by proper FOM's, is
automatically processed through a cyclic procedure integrating
structure factor calculation - least squares - 2Fo-Fc Fourier
synthesis. Due to reflection overlap the observations in the
least squares routine are the total intensities of groups of
reflections, while intensities of single reflections constitute
single observations only when they do not overlap with any other.
The final outcome is a set of refined atomic parameters (x,y,z)
associated with atomic species. When neutron data are processed
parameters of atomic species with negative scattering length can
also be determined. If a graphic device is available the user can
follow structure solution and refinement on the screen. In the
final stage a menu-driven interface is available in order to
study molecular geometry and restart refinement. Due to the small
ratio "number of observations / number of parameters" the
residual R values must be carefully considered by the user
(usually final values between 0.06 and 0.20).
SOFTWARE ENVIRONMENT: The program has been written in standard
Fortran77. A module written in C language is supplied in order to
interface the program with X-window or DEC-Window terminal for
graphics. Therefore, besides a C compiler, an Xlib library is
also needed. Two ASCII files are associated to the program: the
first contains coefficients for calculating scattering factors,
the second, necessary for graphics, contains on-line help.
HARDWARE ENVIRONMENT: The program runs on UNIX and DEC
workstations, on mainframe and on personal computers (for this
last one 4 Mbytes of RAM and a VGA monitor are needed).
PROGRAM SPECIFICATION: The source code of the program consists of
44.000 FORTRAN lines and of 2000 C lines (including comments).
DOCUMENTATION: Users instructions and program description (about
1400 lines) are supplied as an ASCII text file.
AVAILABILITY: The program is available free of charge from the
teams in Bari and in Perugia. A licence agreement has to be
signed.
REFERENCES
Altomare, A., Cascarano, G., Giacovazzo, C., Guagliardi, A.,
Burla, M.C., Polidori, G. & Camalli, M. (1994). J. Appl.
Cryst. 27,000-000.
Baggio, R., Woolfson, M.M., Declercq, J.P. & Germain, G. (1978).
Acta Cryst. A34,883-892.
Giacovazzo, C. (1977). Acta Cryst. A33,933-944.
Giacovazzo, C. (1980). Acta Cryst. A36,362-372.
Main, P. (1978). Acta Cryst. A34,31-38.