Blake A.J.(ed.) Crystal Structure Analysis

Подождите немного. Документ загружается.

322 The crystallographic information ﬁle (CIF)

(b) Each data name must have its corresponding data item, but the

latter need not contain real information. Sometimes a placeholder

such as ? or . is used as in

_chemical_name_common ?

These placeholders are used within loop structures when some

data items are not relevant to every line of the loop. In the follow-

ing example the fourth data name in the loop applies only in the

second line.

loop_

_geom_bond_atom_site_label_1

_geom_bond_atom_site_label_2

_geom_bond_distance

_geom_bond_site_symmetry_2

_geom_bond_publ_flag

Ni N1 2.036(2) . Yes

Ni N1 2.054(2) 2_555 Yes

Ni S2 2.421(10) . Yes

C S2 1.637(3) . Yes

N1 C1 1.327(3) . ?

N1 C5 1.358(4) . ?

N2 C12 1.309(3) . ?

Some items are mandatory for certain CIF applications: for exam-

ple, the list of data items required for a submission to Acta

Crystallographica Sections C and E is given in these journals’ Notes

for Authors.

must be used whereverpossible. For example, there arenow seven

standard codes for the treatment of H atoms during reﬁnement

associated with the data name

_refine_ls_hydrogen_treatment

and these include refall (all H parameters reﬁned) and constr

(e.g. a riding model). If the standard codes are inappropriate or

inadequate then a fuller explanation can be given as part of an

experimental section.

(d) There are now a large number of ancillary programs available for

preparing, checking, manipulating and extracting CIF data. Some

of these will be mentioned later and others are given on the IUCr

website (www.iucr.org).

(e) Additional tables can be created within the CIF format. The

most common of these contain hydrogen-bonding parameters

and there are now standard geom_hbond_ data items to facil-

itate their input. However, it is possible to set up tables of

non-standard parameters by deﬁning additional data items; the

following example sets up a comparison table of molecular

geometry parameters.

20.5 Some practicalities 323

loop_

_publ_manuscript_incl_extra_item

’_geom_extra_tableA_col_1’

’_geom_extra_tableA_col_2’

’_geom_extra_tableA_col_3’

’_geom_extra_tableA_col_4’

’_geom_extra_tableA_col_5’

’_geom_extra_tableA_col_6’ # up to 14 columns allowed

’_geom_extra_table_head_A’ # for table heading

’_geom_table_headnote_A’ # for headnote if needed

’_geom_table_footnote_A’ # for footnote if needed

_geom_extra_table_head_A

;

Table 3.

Comparison of molecular geometry parameters (\%A,\%) for

1,3-dioxolan-2-ones

;

loop_

_geom_extra_tableA_col_1

_geom_extra_tableA_col_2

_geom_extra_tableA_col_3

_geom_extra_tableA_col_4

_geom_extra_tableA_col_5

_geom_extra_tableA_col_6

Parameterˆaˆ (I) (II) (III) (IV) (V)

O1---C2 1.33 1.327(2) 1.316(6) 1.34(2) 1.323(5)

"C2\\db O2" 1.15 1.207(2) 1.192(6) 1.21(2) 1.200(6)

C2---O3 1.33 1.341(2) 1.316(6) 1.28(2) 1.348(6)

O3---C4 1.40 1.447(2) 1.443(5) 1.42(2) 1.460(6)

C4---C5 1.52 1.531(2) 1.498(7) 1.53(2) 1.527(6)

O1---C5 1.40 1.448(2) 1.420(6) 1.46(2) 1.456(5)

O1---C2---O3 111 112.7(1) 111.9(4) 113(1) 112.0(4)

_geom_table_footnote_A

;

(I) 1,3-dioxolan-2-one (Brown, 1954)

(II) D-erythronic acid 3,4-carbonate (Moen, 1982)

(III) 4-p-chlorophenyloxymethyl-1,3-dioxolan-2-one

(Katzhendler et al.,1989)

(IV) 4-(5-(2-iodo-1-hydroxyethyl)-5-methyl-tetrahydro-2-

furyl)-4-methyl-1,3-dioxolan-2-one (Wuts, D’Costa &

Butler, 1984)

(V) 1,6-bis(1,3-dioxolan-2-one)-2,5-dithiahexane (this

work)

ˆaˆ Atom numbering scheme has been standardised as for (V)

;

20.5 Some practicalities

20.5.1 Strings

The correct handling of strings throughout the CIF is vital. There are

three ways to supply the information in these and examples of each

follow.

324 The crystallographic information ﬁle (CIF)

(a) Delimitation by blanks – the data item is effectively a word or a

number without any spaces within it. The data item cannot extend

beyond the end of a line. Examples are:

_publ_contact_author_email J.O-Groats@north.ac.uk

_cell_length_a 10.446(3)

_diffrn_standards_number 3

Note that

J. O-Groats@north.ac.uk

and

10.446 (3)

are not allowed because they contain spaces.

(b) Delimitation by quotation marks (single or double) – the data item

may now contain spaces. It is limited to one line, but it can be on

the line following the data name if required. For example:

_exptl_crystal_density_method ’not measured’

_chemical_formula_moiety ’C12 H24 S6 Cu 2+, 2(P F6 -)’

_publ_section_acknowledgements

"We thank EPSRC for support (to J.O’G.)."

is necessary for blocks of text that exceed one line in length. For

example:

_publ_section_abstract

;

In the title compound C˜12˜H˜22˜O˜7˜, (1), molecules

occur exclusively as the cis geometric isomer and are

linked by hydrogen bonding to form helices running

parallel to the crystallographic c direction.

;

20.5.2 Text

(a) When preparing a data ﬁle for publication, most effort will be

devoted to the textual sections, in particular _publ_section_

abstract, _publ_section_comment and _publ_section_

references. Much of this may appear as normal text, but

certain special character codes are commonly required. Sub-

scripted and superscripted text is delimited by pairs of tilde (∼)

and caret (ˆ) characters, respectively; for example, if you want

[Cu(H

]

to appear in your paper, you need to enter it

as [Cu(H∼2∼O)∼4∼]ˆ2+ˆ. Note that these must not be used

in _chemical_formula_moiety, etc. In discussing molecular

geometry you will need the symbols Å and

◦

, the codes for which

are \%A and \% (not ˆoˆ), respectively. You will occasionally need

20.5 Some practicalities 325

other codes – see the appropriate page inthe IUCrjournalswebsite

for the full list.

(b) Certain trivial errors occur frequently and can cause a great deal

of annoyance because the CIF processing software does exactly

what you tell it to, rather than what you want. CIF checking

software (see below) will strive to return helpful reports on the

location of errors in the data ﬁle, but sometimes the results of the

fault are so pervasive, or appear so far removed from the orig-

inal error, that a manual search is necessary. Some of the most

common and irritating faults arise from the simplest of causes,

such as the failure to have matching subscript and superscript

codes: forgetting to ‘switch off’ these features means that subse-

quent information in the CIF is misinterpreted. Another frequent

mistake is not terminating text strings or text blocks correctly.

publication depends on the setting of the _geom_type_publ_

flag for each parameter, where type is bond, angle or

torsion. Setting a ﬂag to Yes (or y) indicates that the corre-

sponding parameter should be published: anything else (No, n or

?, for example) means that it will not. Make sure that this editing

does not disrupt the number of data items (including placehold-

ers), as doing so will create problems for any program attempting

to read the CIF.

20.5.3 Checking the CIF

Before a CIF is used, for whatever purpose, it is essential to sub-

mit it for automated checking. The checkCIF procedure tests for valid

CIF data names, correct syntax, missing IUCr Journals Commission

requirements, consistency of crystal data, correct space group, unusual

atomic displacement parameter values, completeness of diffraction

data, etc. It is available through the IUCr website (www.iucr.org), and

the current version returns reports within a few seconds. Depending on

the intended purpose of the CIF, failure to satisfy certain of these tests

(e.g. missing Journals Commission requirements) may not be impor-

tant as they are designed primarily as a check on a data ﬁle before

electronic submission to Acta Crystallographica. If you have a way of

reading PDF ﬁles, such as the free program Adobe Reader, you should

take advantage of another utility: printCIF, also available through the

IUCr website, returns a preprint of your paper for checking. This is

valuable because some errors, especially formatting ones, may not be

detected by checkCIF but they are usually horribly obvious on a preprint.

The IUCr program publCIF (Westrip, 2009) includes the same function-

ality as printCIF but, like enCIFer, it is interactive, offering an HTML

representation of the required CIF publication data. You can also carry

out useful additional checks with programs such as PLATON (Spek,

2003) running on your own computer: these include not only numerical

checks but also visual checks of conformation, ellipsoid plots and other

326 The crystallographic information ﬁle (CIF)

features. It is worthwhile checking whether chemically equivalent geo-

metric parameters are equal within the relevant standard uncertainty

limits; for example, in a tertiary butyl substituent are all three C–CH

bond distances close to 1.52 Å, and is there tetrahedral geometry around

the central carbon atom? If not, you should check for disorder or other

problems.

References

Allen, F. H., Johnson, O., Shields, G. P., Smith, B. R. and Towler, M. (2004).

J. Appl. Crystallogr. 37, 335–338.

Hall, S. R. (1991). J. Chem. Inf. Comput. Sci. 31, 326–333.

Hall, S. R., Allen, F. H. and Brown, I. D. (1991). Acta Crystallogr.

A47, 655–685. Reprints are available from the International Union of

Crystallography, 5 Abbey Square, Chester CH1 2HU, England.

Spek, A. L. (2003). J. Appl. Crystallogr. 36, 7–13.

Westrip, S. P. (2009). In preparation.

In 2006 the IUCr published Volume G of International Tables for Crystal-

lography (S.R. Hall and B. McMahon (eds.); ISBN 1-4020-3138-6). This

deals with the deﬁnition and exchange of crystallographic data by

means of the CIF format.

Crystallographic

databases

Jacqueline Cole

21.1 What is a database?

A database is a collection of related data along with tools for amending,

updating and adding records and selectively extracting information.

Crystallographic databases generally contain basic crystallographic

data (unit cell dimensions, space group, atomic co-ordinates and per-

haps atomic displacement parameters) and may also carry derived

information on connectivity or atom and bond properties. Bibliographic

information such as author names, the journal and year of publication

will be stored. A compound’s formula and systematic name, abso-

lute conﬁguration, polymorphic form and pharmaceutical or biological

activity may be indicated where applicable. Experimental details such

as the temperature and radiation used in the experiment may be avail-

able. Entries may carry ﬂags indicating the level of precision for each

structure, the presence of disorder, and there may even be comments

about problems or unresolved queries regarding the structure. The exact

contents depend on the database.

21.2 What types of search are possible?

Depending on the database, it may be possible to carry out searches

based on a structure fragment, compound name, compound formula,

compound properties, experimental conditions, space group, unit cell

dimensions or bibliographic criteria – or some combination of these.

For molecular compounds the ability to sketch or deﬁne a structure

fragment whose occurrence within the database can then be probed is a

powerful and intuitive tool for chemists. Whereas details of the internal

structure of a database are not generally of interest, a good knowledge

of the search facilities available is essential in order to best exploit the

stored information.

327

328 Crystallographic databases

21.3 What information can you get out?

• bibliographic data,

• space group and cell dimensions,

•

atomic co-ordinates,

• atomic displacement parameters (maybe),

• molecular geometry parameters,

• intermolecular geometry parameters,

• analyses of the above, etc.

21.4 What can you use databases for?

• ﬁnding out what has been done before – surveying the ﬁeld,

• determining if speciﬁc compounds have been reported,

• checking if a unit cell is known – diagnostic for the phase,

• obtaining parameters to assist structure solution or reﬁnement

(e.g. geometrical restraints),

• validation and comparison of your structure against published

ones,

• deriving typical bonds lengths or other parameters,

• as a source of parameters for calculations or simulations,

• as a research tool for structure correlation, identifying trends or

relationships, data mining,

• etc.

21.5 What are the limitations?

You need to be aware that there can be a considerable delay between

the appearance of a structure in the literature and its inclusion in a

database, although the availability of data in standard electronic format

should reduce this delay. Such delays may result from a number of fac-

tors; perhaps the required data have not been (automatically) sent to the

database, or database updates may not be distributed frequently. Struc-

tures that have been determined but neither published nor deposited

will obviously not be included, but structures that form part of confer-

ence proceedings, including poster presentations, may be represented

only by very basic data. For these reasons you should also check non-

crystallographic databases such as Chemical Abstracts, which are more

current (but may contain little or no crystallographic data).

21.6 Short descriptions of crystallographic

databases

The Cambridge Structural Database (CSD) is the comprehensive collec-

tion of small-molecule organic and organometallic crystal structures. It

does not contain structures of inorganic compounds like NaCl, PtS or

21.6 Short descriptions of crystallographic databases 329

Fig. 21.1 ConQuest (CSD)

CuSO

·5H

O; metals or their alloys; or macromolecular structures such

as proteins or nucleic acids. The ConQuest (and earlier Quest) software

has been developed for the search, retrieval, display and analysis of CSD

information and its particular strength is the ability to search for struc-

tures on the basis of a chemical diagram, although text-based searches

are also possible (Fig. 21.1). At the beginning of 2009 the CSD con-

tained crystal structure data for over 460 000 organic and organometallic

compounds. The CSD is updated through a full release every year,

with approximately quarterly interim updates via the internet for reg-

istered users. There are a number of related programs such as Vista

(graphical display of results, statistical analysis), Isostar (CSD-derived

library of non-bonded contacts), Mogul (library of molecular geome-

tries) and Mercury (advanced graphical visualization of structures). See

http://www.ccdc.cam.ac.uk for further details.

The Inorganic Crystal Structure Database (ICSD) provides fast

retrieval of structural and bibliographic information, allows logical

manipulation of retrieved material and display of results, and is

complementary to the CSD (Fig. 21.2). By late 2008, the ICSD contained

over 108 000 entries. It is searchable by using either a command-line or

330 Crystallographic databases

Fig. 21.2 ICSD

a web interface and contains two types of data: (1) bibliographic infor-

mation about each entry, giving authors, journal reference, compound

name, formula, mineral name (if any), etc.; (2) numeric information from

the crystal structure analysis (if it is available), giving cell parameters,

spacegroup,atomicco-ordinatesanddisplacement parameters. Because

of the nature of the structures, connectivity searching is not appropriate.

See http://cds.dl.ac.uk/cds/datasets/crys/icsd/llicsd.html for further infor-

mation.

The Metals Data File or CrystMet (MDF) provides fast retrieval of

structural and bibliographic information for metals and alloys. For

searching, the MDF uses a set of instructions similar to ICSD. It con-

tains two types of data: (1) bibliographic information about each entry,

giving authors, journal reference, compound name, formula, mineral

name (if any) etc.; (2) numeric data from the crystal structure analysis (if

it is available), giving cell parameters, space group, atomic co-ordinates

and displacement parameters. The MDF currently contains around

100 000 entries for metals, alloys and intermetallics.

See http://cds.dl.ac.uk/cds/datasets/crys/mdf/llmdf.html for further infor-

mation.

21.6 Short descriptions of crystallographic databases 331

CDIF was an online retrieval package using the National Institute of

Standards and Technology (NIST, Washington) Crystal Data Identiﬁca-

tion File. Entries in this ﬁle comprised unit cell data for some 237 000

organic, inorganic and metal crystal structures, some 72 000 of which

did not appear in any other database. Each entry gave details of cell

dimensions, crystal class, name and formula of the substance, and jour-

nal reference. CDIF has been superseded by the Daresbury CrystalWeb

interface that allows searches of the CSD, the ICSD, CrystMet and CDIF

on the basis of bibliographic, unit cell, reduced cell, formula, database

code or combined queries.

See http://cds.dl.ac.uk/cweb/ for information on CrystalWeb.

The Protein DataBank (PDB) contains bibliographic and co-ordinate

details for proteins and other biological macromolecules. At the begin-

ning of 2009, the PDB contained over 56 000 entries, of which most

were derived from X-ray studies and others from NMR work. See

http://www.rcsb.org/pdb/ for further information. There is also a compan-

ion nucleic acid database.

With the exception of the PDB, these databases are available free of

charge to UK academic researchers through the Chemical Database Ser-

vice (CDS) at STFC Daresbury Laboratory. The CDS homepage is at

http://cds.dl.ac.uk/. This Service has also provided access to a range of

databases for organic chemistry, physical chemistry and spectroscopy

but some of these services have been discontinued.Access to CSD, ICSD

and DETHERM (a thermophysical properties database) is currently

guaranteed until 2011. See http://cds.dl.ac.uk for up-to-date information.