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

Подождите немного. Документ загружается.

312 The presentation of results

may be calculated elsewhere but not transferred automatically. You may

even want to include non-existent ‘bonds’, for example to demonstrate

that two atoms are not close enough to interact. These values and their

s.u.s should be calculated by the reﬁnement program.

19.12 The format of tables

Journals tend to have their own requirements for tables that you must

follow or risk objections from the referees or editors. The precision to

which results are required does vary: Acta Crystallographica prefers s.u.s

in the range 2–19, while Dalton Transactions have preferred 2–14 in the

past, and some referees and editors object to s.u.s of 1. This can mean

allowing more signiﬁcant ﬁgures for the co-ordinates of heavier ele-

ments. Make sure the s.u.s look sensible and that any redundant data

(such as the U

components for atoms on special positions other than

inversion centres) have the correct relationships between their values

(and among their s.u.s).

Ensure that the table headings are informative and correct: are the

powers of ten quoted there actually those used in the table? Are the

displacementparameters correctlyidentiﬁed asU or B?Aretheheadings

on the structure-factor tables correct, with any reﬂections not used in the

reﬁnement ﬂagged? If it is possible, I suggest having a compound code

or other identiﬁer on every page of tables so that structures cannot be

mixed up.

Whileoneprogrammight producegeometric parameters based on the

order in which atoms occur in the reﬁnement model, another might give

the bonds in ascending order of length, and so on. It is worth looking

at the tables to see whether this can be improved. In my opinion, the

clarity of this table of selected bond lengths (in Å):

Pd--N6 1.996(8) Pd--N2 2.017(7)

Pd--N4 2.001(6) Pd--N5 2.035(6)

Pd--N1 2.008(7) Pd--N3 2.057(7)

is much improved by re-ordering to give:

Pd--N1 2.008(7) Pd--N4 2.001(6)

Pd--N2 2.017(7) Pd--N5 2.035(6)

Pd--N3 2.057(7) Pd--N6 1.996(8)

19.13 Hints on presentation

19.13.1 In research journals

This has mostly been covered already. Follow the instructions to authors

for submitting papers, including experimental data, tables, ﬁgures and

supplementary data. What is the policy on colour plots? The range of

formats for literature references can appear overwhelming, and if you

19.13 Hints on presentation 313

submit to a wide range of journals the use of reference management

software may be worthwhile. If you regularly use the same references

in the same format they could simply be stored in a standard ASCII or

word processor ﬁle.

19.13.2 In theses and reports

Here you have much more freedom, but you have to be careful that the

result is appropriate in style and length to the purpose in hand: a thesis

that runs to 300 pages may be acceptable but an interim report of that

size is ridiculous. Fortunately, guidelines are normally available at each

institution, so consult them before you write a word.

Don’t overdo the tables: it is seldom necessary to include structure-

factor tables, even in an appendix. However, you could put co-ordinate

data and full molecular geometry tables in appendices and retain only

selected information in the main body: this will cause less disruption to

the ﬂow of your report. These appendices do not necessarily even have

to be in hard copy: does your institutionallow you to include appendices

on a CD or DVD?

You can be more generous with diagrams than when publishing in

a journal, but remember that there must always be a good reason for

including any diagram.

19.13.3 On posters

Select the most important points you want to get across. Save yourself

time by planning what you are going to present – there is no point in

producing material you don’t have space for. (Do you know the size

and orientation of the poster display area available to you?) Text must

be readable from a distance of one to two metres – you may not always

attract a crowd but the poster will make a greater impact if it can be

viewed from a comfortable distance. Keep any tabulated information

short and relevant. (Are the crystal data really needed on the poster, or is

it sufﬁcient to have these to hand in case someone asks?) Posters are one

placewherecolour can be exploited to thefull,andnotonlyinﬁguresbut

in text, backing material and surrounds. It is harderto overdo it here, but

still possible! You should keep the information density relatively low:

this has the additional advantage that you will have something more to

tell those who express an interest in your work.

19.13.4 As oral presentations

Many of the points mentioned in respect of posters apply here too: avoid

high-density slides or overhead transparencies that nobody will have

time to read. Don’t be tempted to use that convenient table prepared for

publication if it consists mostly of values that are not relevant to your

lecture.

314 The presentation of results

An important function of your visual aids can be to remind you what

to say next, so they must be ‘in phase’ with your talk; they must not

let the audience see your ﬁnal results while you are still outlining the

problem! If you need to refer to the same slide or transparency at two

points in your talk, it is better to make two copies than to waste time

rummaging around for the single copy you last saw ten minutes before.

If you have just covered the material outlined on a slide and made a

point that requires the audience to absorb information from the slide,

do not immediately proceed to the next one. This point is particularly

relevant if the slide contains a crystal structure diagram.

Compose your visual aids carefully. Try to ﬁnd out the size of the

auditorium and about its facilities. Mixing slides and transparencies

requires some planning to ensure you don’t lose track of what you are

saying. If you are inexperienced it is safer to have all your visuals in

the same format if at all possible. Colour can be extremely powerful,

especially when used in bold, simple illustrations.

Do not try to cover too much material. Your audience will be less

familiar with your material than you are and it will not help if you

speak too quickly. Time your talk in advance–apractice session with a

sympathetic (and constructively critical) audience is a good idea if you

are unused to speaking in public. You should assess the composition of

your audience in advance. If they are not experts in your own ﬁeld, you

will need to give more background information so that they understand

the context, before you begin to describe your own work and its results.

Humour can be a good way to engage the audience’s attention but

it needs to be used carefully and sparingly. In some circumstances it is

wholly inappropriate. If in any doubt, avoid it.

19.13.5 On the web

The web is an excellent medium for disseminating research results but

it has its own special requirements. The speed at which data can be

transferred is limited: as faster networks are installed these are required

to carry ever more demanding applications. Combined with a constant

increase in the number of users this ensures that bandwidth is always

restrictedin some way.As a result, the mosteffectiveweb pages areoften

those that do not involve large-scale data transfer in order to be useful.

Authors need to bear in mind that many of their potential audience may

be using modest hardware and not-so-recent software, and they must

ensure that using new features in the latest version of their web author-

ing software does not restrict this audience. On a related point, any web

page should at the very least be viewable with the most common web

browsers such as Microsoft Internet Explorer or Mozilla Firefox.

It might seem obvious that any website should be designed so that

the visitor can ﬁnd information easily, yet many impressive-looking

corporate sites are so poorly structured that ﬁnding what you want

is time consuming, inefﬁcient and frustrating. If you have an extensive

website you should give serious thought to how it is constructed, and

19.14 Archiving of results 315

in particular whether your homepage allows a visitor easily to begin

navigating it. As well as hyperlinked text and graphics, web publishing

offers the possibility of illustrations that can be rotated or otherwise

moved, either according to your pre-programmed instructions or in

response to user input. This allows the visualization of complex struc-

tures and packing diagrams, for example. These features are usually

implemented by means of Virtual Reality Modelling Language (VRML)

extensions to your browser.

Publishing results on the web is quite different from displaying them

on a poster that you can take down at the end of a conference. Once

placed on the web, material assumes a substantial degree of permanence

as it is accessed, stored in various caches, copied to other formats and

printed. On the other hand, such results can be more ephemeral that

those printed in a journal, as it is possible to remove, modify or update

them. The copyright implications of publishing your results on the web

are often far from clear, but these are likely to become more serious with

the spread of electronic publishing. As with other forms of publication,

placing results on the web should be done only with the knowledge

and consent of all those contributing to the work, and only when the

consequences of doing so have been fully explored.

19.14 Archiving of results

Although some of the results of your structure determinations will end

up in a database after publication, you must keep your own copies of

all relevant ﬁles and other information safely and in an accessible form.

Most crystallographers know the frustration of setting out to prepare a

structure for publication, only to ﬁnd that some experimental parameter

such as the colour of the crystal or the type of diffractometer used is not

immediately available and has to be ferreted out. In the not-so-distant

pastthe only safe way seemed to be to keepevery piece ofhard-copy out-

put ever generated for a structure, but now archiving and transmission

tools such as the CIF format allow this to be done much more concisely.

The ‘paperless ofﬁce’ once promised by the advocates of information

technology may have provedillusory elsewhere, but in the modern crys-

tallography laboratory it has largely materialized. The CIF and its uses

are described in detail in the next chapter.

When archiving data the main considerations are safety and accessi-

bility. To address the ﬁrst, you need to keep backup copies of your ﬁles,

possibly on tape, including a set that will survive ﬁre, ﬂood and theft

at your workplace. This could involve a ﬁreproof safe, but keeping a

backup in a safe place at home is probably as reliable. Keeping all the

ﬁles for one structure together aids organisation, and utilities such as

PKZIP, WinZIP, PowerArchiver or WinRAR allow these to be compressed

within a single archive ﬁle, with the bonus of a considerable saving on

disk space. For accessibility, you need some form of indexing so that the

structures you require can be quickly and uniquely identiﬁed. Before

316 The presentation of results

starting to rely on it, you must check the backup procedure works by

restoring some typical ﬁles from the archive. Most backup devices come

not only with software to drive them but also with documentation that

includes advice on how to implement a suitable and effective backup

regime. This documentation should include explanations of how to use

both full and incremental backups, the latter referring to the procedure

whereby only those ﬁles that have changed since the last backup are

transferred to the archive.

There are two distinct aspects to backing up a particular computer.

The ﬁrst requires an archive medium capacious enough to allow you

to back up everything on that computer, including operating system,

applications and data. This would allow you to re-create your working

environment in the event of the computer or its hard disk failing totally,

andrequires a backup medium with the same capacityas your hard disk.

The best solution used to be some kind of tape drive, and models with

capacities of up to many tens of gigabytes are currently available, but

there is an increasing view that the only effective, convenient backup

medium for a hard disk is another hard drive. An additional level of

security could be provided by backing up your working hard disk to

another physical hard disk (not a different logical drive on the same disk)

on the computer. Such a backup will survive any failure of the working

disk, and most disasters short of theft or outright destruction of your

computer. If the computer is automatically backed up over a network

you may be content to rely on this, but make sure that the frequency of

backup is appropriate and that the backup ﬁles are accessible. Remem-

ber to update the backup ﬁles whenever you make signiﬁcant changes

to the computer, for example after a major new application has been

installed and conﬁgured.

The second aspect is the regular backup of new data and the media

that can be used will depend on the volume of these data. The essential

ﬁles to be kept at the end of a structure analysis may well amount to only

a few hundred kilobytes after compression and several structures could

be archived on a standard 3.5” ﬂoppy disk. In contrast, the frames for

one data collection using an area-detector diffractometer occupy several

hundred Mbytes and, while archiving to CD-ROM is a possibility, each

CD will hold only two or three sets of frames at most. However, with

DVD writers costing from less than £50 per unit, and each disk holding

15–40 sets, this seems a sensible backup medium. Two rival formats

(Blu-ray and HD DVD, see Table 19.1) were in competition to succeed

DVD, but in 2008 it became clear that Blu-ray had won. For data transfer,

solid-state drives are now available with sufﬁcient capacities (e.g. 8 Gb)

and have the advantage of simplicity.

When you are planning a backup regime, ease of use is an important

factor. You are unlikely to regularly use any method that is cumbersome

or time consuming.Archive media can change and develop as rapidly as

other aspects of computer hardware, and factors such as capacity, cost,

convenience and durability need to be considered. It is not necessarily

best to adopt the latest technology: in fact it may be safer to select one

References 317

Table 19.1. Data storage capacity.

Medium Capacity

3.5" disk 1.44 Mb

super-ﬂoppy 120 Mb

cartridges 100 Mb–2 Gb

solid state drives 1–128 Gb

CD-ROM 550 Mb

DVD 4.7–8.5 Gb

HD DVD 30 Gb

Blu-ray disk 50 Gb

tape 400 Mb–120+ Gb

portable hard disk 40–500 Gb

that has gained reasonably wide acceptance so that consumables such

as tapes are likely to remain available for the useful life of the computer.

Table 19.1 gives (sometimes rather approximate) current capacities for

various storage media.

References

Farrugia, L. J. (1999). J. Appl. Crystallogr. 32, 837–838.

Johnson, C, K. (1965). ORTEP. Report ORNL-3794. Oak Ridge National

Laboratory, Oak Ridge, Tennessee, USA.

Johnson, C. K. (1976). ORTEPII. Report ORNL-5138. Oak Ridge National

Laboratory, Oak Ridge, Tennessee, USA.

Keller, E. (1989). J. Appl. Crystallogr. 22, 19–22.

Macrae, C. F., Edgington, P. R., McCabe, P., Pidcock, E., Shields, G. P.,

Taylor, R., Towler, M. and van de Streek, J. (2006). J. Appl. Crystallogr.

39, 453–457.

Pearce, L. J., Watkin, D. J. and Prout, C. K. (1996). CAMERON. Chemical

Crystallography Laboratory, University of Oxford, UK.

Sheldrick, G. M. (2001). SHELXTL XP graphics module: various ver-

sions, and for different platforms (e.g. PC, Unix, Linux).

This page intentionally left blank

The crystallographic

information ﬁle (CIF)

Alexander Blake

20.1 Introduction

The crystallographic information ﬁle (CIF) is an archive ﬁle for the trans-

mission of crystallographic data: this transfer can be between different

laboratories or computer programs, or to a journal or database. The ﬁle

is free-format, ﬂexible and designed to be read by both computer pro-

grams and humans (the latter require a little practice at the start). The

speciﬁcation of the CIFstandardhas been published and thesame article

providesinformationon its evolution(Hall et al.,1991). Itisbased around

the self-deﬁning text archive and retrieval (STAR) procedure (Hall,

1991), and consists of data names and the corresponding data items with

a loop facility tohandle repeated items such as the author/addresslist or

the fractional co-ordinates. The format is extensible, so that data names

covering new developments such as area detectors can easily be accom-

modated. However, once a data name is included it is never removed,

otherwise portions of those archives written in the interim would be

undeﬁned.

20.2 Basics

The CIF is an ASCII ﬁle, such that only the following characters are

allowed:

abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ

0123456789 !@#$%ˆ&*()_+{}:"˜<>?|\-=[];’‘,/.

Any others that you may want to include in a manuscript (such

as Å,

◦

, é, ø, subscripts and superscripts, Greek letters, mathemati-

cal symbols such as ±, and chemical multiple bonds) require spe-

cial codes that are detailed in the Notes for Authors for Acta Crys-

tallographica, Section C. Do not attempt to show italicized, bold or

underlined text (e.g. in space group symbols) as these attributes

319

320 The crystallographic information ﬁle (CIF)

should be added automatically. Many data names have implicit units

(e.g. Å for _cell_length_a;Å

for _cell_volume; minutes for

_diffrn_standards_interval_time) and these units must not be

appended; thus

_cell_volume 2367.5(8)

is correct but

_cell_volume 2367.5(8)\%Aˆ3ˆ

is not.

If you prepare your CIF using a word processor rather than a text

editor you must make sure that the output ﬁle is ASCII, that there

are no (hidden?) embedded codes and that lines do not exceed 80

characters in length. If you are exporting an ASCII ﬁle from a word

processor it is best to have used a large ﬁxed font (e.g. Courier 12

point or larger) so that any lines that are part of blocks of text do

not overrun upon conversion. Do not include any non-ASCII char-

acters in the word processor ﬁle as these will be lost or corrupted

upon writing the ASCII ﬁle and, even if they survived unchanged,

the CIF processing software will not recognize them correctly. It is

much safer to use a dedicated CIF editor such as enCIFer (Allen

et al., 2004; see http://www.ccdc.cam.ac.uk/free_services/encifer/

for the enCIFer home page) or publCIF (Westrip, 2009; see http://

journals.iucr.org/services/cif/publcif), because these programs also

check the CIF for syntax errors and whether the data names and items

correspond to valid CIF dictionary entries. EnCIFer can also display the

structure graphically.

The following CIF terminology is used.

text string a string of characters delimited by blanks, quotes, or

semi-colons (;) as the ﬁrst character on the line

data name a text string starting with an underline (_) character

data item a text string not starting with an underline, but preceded

by a data name

data loop a list of data names, preceded by _loop and followed by a

repeated list of data items

data block a collection of data names and data items (which may be

looped) preceded by a data_code statement and terminated

by another data_ statement or the end of the ﬁle. A data

name may only occur once within any one data block.

data ﬁle a collection of data blocks: no two data block codes may

have the same name

CIF data name and data block code deﬁnitions are restricted to a max-

imum of 76 characters, but their hierarchical construction and careful

design mean that they are largely self-explanatory: e.g.

20.4 Some properties of the CIF format 321

_publ_author_name

_exptl_crystal_colour

_computing_structure_solution.

The term ‘CIF’ has acquired a range of informal meanings: it is used to

describe the format, the data output by a control or reﬁnement program,

as well as the data ﬁle (comprising two or more data blocks) submitted

as a manuscript for electronic publication.

20.3 Uses of CIF

(a) Your own local archive. A CIF produced by a reﬁnement program

upon convergence of your structure can be edited and augmented

so that all the relevant results and details of procedures can be

stored. As with other uses, the degree of manual editing required

will depend on the extent to which your data collection and

reduction programs produce relevant output in CIF format.

(b) A standard method of transmitting data between crystallographic

programs (an increasing number of which read ﬁles in CIF format)

or to colleagues in other laboratories.

containing crystal-structure determinations.

(d) A standard route for deposition into structural databases.

(e) A route to standard printed tables (e.g. via the SHELXL ancillary

program CIFTAB or XCIF).

(f) Direct electronic submission of manuscripts to journals such

as Sections C and E of Acta Crystallographica or Zeitschrift für

Kristallographie. The required data ﬁle will consist of a num-

ber of data blocks. One, perhaps called data_global and

possibly prepared by manual editing of a template, will con-

tain author contact information, submission information, an

author/afﬁliation/address list, a title, a synopsis, an abstract, a

comment (discussion) section, experimental details, references,

ﬁgure captions and acknowledgements. This block will be fol-

lowed by one data block for each structure to be included, perhaps

called data_compound1, data_compound2, etc.

20.4 Some properties of the CIF format

(a) Within any data block, the ordering of the associated pairs of

data names and data items is not important, as ﬁle integrity does

not depend on ﬁnding these in a particular sequence. However,

humanreaderswill ﬁnd it easier to readthe CIF if they are grouped

logically. Furthermore, there are no restrictions on the ordering of

the data blocks.