International
Tables for
Crystallography
Volume G
Definition and exchange of crystallographic data
Edited by S. R. Hall and B. McMahon

International Tables for Crystallography (2006). Vol. G, ch. 3.3, pp. 117-130
https://doi.org/10.1107/97809553602060000735

Chapter 3.3. Classification and use of powder diffraction data

B. H. Tobya*

aNIST Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899-8562, USA
Correspondence e-mail: brian.toby@nist.gov

Powder diffraction creates new demands for versatility in CIF. Unlike single-crystal experiments, where raw data can be automatically reduced to an instrument-independent form (structure factors), powder-diffraction data will always be instrument dependent, and there is a wide array of instrumentation: (i) X-ray: conventional, synchrotron, energy-dispersive; (ii) neutron: CW, TOF; (iii) and detection: single, multiple, linear position-sensitive, area, film. pdCIF has other special needs: (i) multiple crystallographic phases may be present (requires inter-block links); (ii) multiple sets of data may be used for a single model (inter-block links); (iii) and calibration data may appear in other CIFs (inter-file links). Since powder diffraction may be used for noncrystallographic purposes, such as phase identification, the data themselves may be the result. For classification purposes, the pdCIF dictionary is divided in sections. Few pdCIFs will have entries defined for all sections. The use of each section and the component data items is presented on a section-by-section basis, describing: the experimental configuration; sample preparation, characterization and mounting details; references to calibration information; the experimental (raw) diffraction data set; the processed diffraction data set; calculated diffractogram; peak table; reflection assignments and intensities; and derived structural results. The use of the different data items is shown through examples.

3.3.1. Introduction

| top | pdf |

Powder diffraction is used for many purposes. One important use is the determination of crystallographic models (crystal structures), particularly when single crystals are not available. Other common uses of powder diffraction include the identification of unknown materials; determining the amounts of different crystalline phases in a mixture; studies of phase transitions; and measuring changes in lattice constants with composition, pressure or temperature. Residual stress measurements are also frequently made using powder diffraction. Measurements of the degree of preferred orientation of grains in a processed material (texture or pole-figure measurements) are also carried out using powder diffraction; these are useful as they can be used to relate the engineering properties of a material to its processing conditions.

A wide range of instrumentation is used for powder diffraction. Synchrotron, sealed-tube and rotating-anode X-ray sources and both spallation and reactor neutron sources are used. When monochromatic radiation is used, diffraction intensities are measured as a function of the detector setting angle, [2\theta]. When polychromatic radiation is used, however, energy-dispersive detectors at fixed angles are used. For a pulsed neutron source, the neutron time-of-flight (TOF) is measured to provide energy-dispersive detection. Energy-dispersive X-ray diffraction is not common, but is used for certain specialized measurements such as in situ high-pressure studies. Most laboratory powder diffractometers have a single discrete detector, but a few use position-sensitive detectors (PSDs). Multiple detectors or PSDs are common for specialist instruments, for example at large national user facilities.

Historically, X-ray sensitive film in a diffraction camera was used to measure the intensities. At present, diffractometers are much more common than cameras, but film is an inexpensive method of area detection and cameras continue to have some advantages over diffractometers. X-ray film, however, has a nonlinear sensitivity and a limited dynamic range, so electronic area detectors and image plates are replacing film for many applications.

3.3.2. Dictionary design considerations

| top | pdf |

The many applications of powder diffraction and the wide range of different instruments used dictate the design of several aspects of the powder diffraction CIF (pdCIF) data dictionary. In contrast to single-crystal measurements, where observations are reduced to instrument-independent structure factors, in powder diffraction it is not possible to interpret the data without a detailed knowledge of the type of instrument used. Another difference is that while single-crystal measurements are rarely made for any purpose other than structure determination, powder diffraction data have many non-crystallographic applications (e.g. identifying the phases in a mixture of unknown composition). In a single-crystal diffraction experiment, one sample yields one crystal structure. In a powder diffraction experiment, however, it is common for the sample to be a mixture of phases. Analysis of the diffraction data may result in multiple crystal structure models. It is also common to use multiple data sets, for example to fit one crystallographic model to both X-ray and neutron data simultaneously.

The powder dictionary was developed with the following basic objectives in mind. A pdCIF:

(i) should record the data-collection experiment as completely as possible and document any data analysis;

(ii) should be appropriate for the exchange of unprocessed measurements, so that it can be used in national laboratories and other shared facilities;

(iii) may contain more than one data set and/or crystal structure through the use of multiple data blocks;

(iv) may accommodate references between data blocks; and

(v) should recognize and accommodate data from as many different types of instruments as possible.

The CIF syntax, as opposed to the STAR File syntax, is not well suited to handling large multi-dimensional data sets. For some two-dimensional image formats, this deficiency was addressed by the development of imgCIF (see Chapters 2.3[link] and 3.7[link] ). It is also true that the CIF syntax is not well suited to storing unprocessed powder-diffraction measurements from the many instruments that use area detectors (particularly for the case of the three-dimensional data structures needed for modern TOF instruments). Even in these cases, however, diffraction intensities are commonly reduced to simpler representations, such as might be input to a Rietveld refinement program. The pdCIF definitions are intended for use with these reduced data sets.

3.3.3. pdCIF dictionary sections

| top | pdf |

At the time when the powder CIF dictionary was created, it was recommended practice to give data items in other dictionaries names that distinguished them from items in the core CIF dictionary. This recommendation proved cumbersome and was subsequently withdrawn, but pdCIF data items are all named in a way that distinguishes them clearly from core CIF dictionary data items: data names in the pdCIF dictionary all begin with the string _pd_ and, likewise, the data categories in the pdCIF dictionary all begin with PD_.

The data items in the pdCIF dictionary are named according to the type of information they contain, rather than the way they will be used in a data structure. For example, data items relating to observed data have names beginning with _pd_meas_, even though some data items beginning with _pd_meas_ may be assigned to the same category as items in other parts of the dictionary. The formal category assignments (see Section 3.1.10.1[link] ) were dictated by the aim to make pdCIF as versatile and simple to use as possible. As an example, one might want to put both the observed and calculated intensities in a single table (loop), as would be the case for a Rietveld refinement. This requires that the calculated and observed intensity data items are assigned to the same category. Note that this does not require that these items always appear in a single loop; calculated and observed intensity values could appear in separate loops, for example, if the increment between data points differs.

This need to contain diverse items in a common `looped' list has led the pdCIF dictionary to use category names in a different way from the other CIF dictionaries, in which CIF data items are usually named according to their category. In the pdCIF dictionary, data items that might appear in the loop for diffraction intensities are assigned to the category PD_DATA. Only one data item is named using this category as prefix, _pd_data_point_id. Another departure from the convention used in other dictionaries is that several _pd_refln_* data names are assigned to the category REFLN so that these items may be included in a loop with _refln_* items defined in the core CIF dictionary.

Table 3.3.3.1[link] summarizes the category groups in the pdCIF dictionary; the individual categories are listed alphabetically in Appendix 3.3.1[link]. The appendix also lists for each category the section of this chapter in which the category is described.

Table 3.3.3.1| top | pdf |
Category groups defined in the powder CIF dictionary

The groups are listed in the order in which they are described in this chapter.

SectionCategory groupSubject covered
(a) Experimental measurements
3.3.4.1[link] PD_CHAR Characterization of a sample
3.3.4.1[link] PD_PREP Preparation of a sample
3.3.4.2[link] PD_SPEC Specimen used in an experiment
3.3.4.3[link] PD_CALIB Calibration information
3.3.4.3[link] PD_INSTR The experimental instrument
3.3.4.4[link] PD_MEAS Raw measurements and instrumental settings
     
(b) Analysis
3.3.5.1[link] PD_PROC Processed settings
3.3.5.2[link] PD_CALC Simulated settings
3.3.5.3[link] PD_PEAK Diffraction peak table
3.3.5.4[link] REFLN Reflection assignments and intensities
     
(c) Atomicity, chemistry and structure
3.3.6.1[link] PD_PHASE Phases present
     
(d) File metadata
3.3.7.1[link] PD_BLOCK Relationships between data blocks
3.3.4.3[link], 3.3.4.4[link], 3.3.5.1[link], 3.3.5.2[link] PD_DATA Measured and simulated intensities

The order in which the categories are discussed follows the scheme of Table 3.1.10.1[link] , so that the contents of the dictionary are summarized under the headings Experimental measurements (Section 3.3.4[link]), Analysis (Section 3.3.5[link]), Atomicity, chemistry and structure (Section 3.3.6[link]) and File metadata (Section 3.3.7[link]). The pdCIF dictionary does not contribute any new data items relevant to publication beyond those already in the core CIF dictionary.

The data items in each category are listed below. Category keys, if specified, are listed first and are marked by a bullet ([\bullet]); the remaining data items in each category are listed alphabetically. Note that the category PD_DATA is discussed in several different sections.

3.3.4. Experimental measurements

| top | pdf |

The categories in the powder CIF dictionary relating to the crystallographic experiment are as follows:

Characterization and preparation of the sample (§3.3.4.1[link])
PD_CHAR group
 PD_CHAR
PD_PREP group
 PD_PREP
Description of the specimen (§3.3.4.2[link])
PD_SPEC group
 PD_SPEC
Instrument calibration and design (§3.3.4.3[link])
PD_CALIB group
 PD_CALIB
 PD_CALIBRATION
PD_INSTR group
 PD_INSTR
Observations and measurement conditions (§3.3.4.4[link])
PD_DATA group
 PD_DATA (items beginning with _pd_meas_*)
PD_MEAS group
 PD_MEAS_INFO
 PD_MEAS_METHOD

The pdCIF dictionary differentiates between the terms sample and specimen. The terms are often treated as interchangeable, but they have quite distinct meanings. The term sample refers to a batch of material, while the term specimen refers to the particular portion of the sample that was used for a measurement. In some cases, the specimen is modified before it is used for data collection. For example, it may be mixed with an internal standard, dried, hydrated or pressed into a pellet.

3.3.4.1. Characterization and preparation of the sample

| top | pdf |

The data items in these categories are as follows:

(a) PD_CHAR [Scheme scheme1]

(b) PD_PREP [Scheme scheme2]

The PD_CHAR data items describe information known about the sample from observation and chemical analysis. For example, a description of the sample morphology can be specified using _pd_char_particle_morphology. Note that there are data items in the core dictionary that are appropriate for use with powder diffraction. For example, _atom_type_analytical_mass_% can be used for chemical analysis results and _chemical_melting_point for the melting point. Several similar data items occur in the pdCIF and core dictionaries. _exptl_crystal_colour and _pd_char_colour both describe the sample colour, but _pd_char_colour is more systematic. Also, _pd_char_atten_coef_mu_calc and _exptl_absorpt_coefficient_mu describe similar properties, but _pd_char_atten_coef_mu_calc is adjusted for the sample packing fraction, so it can be compared with the experimental value, _pd_char_atten_coef_mu_obs, when a direct measurement is made.

The PD_PREP data items describe how the sample was collected or prepared. For example, _pd_prep_pressure and _pd_prep_temperature describe the pressure and temperature used to prepare the sample. Note that these will probably differ from the pressure and temperature conditions at which diffraction measurements are made. Measurement conditions are recorded in _diffrn_ambient_pressure and _diffrn_ambient_temperature.

3.3.4.2. Description of the specimen used in the experiment

| top | pdf |

The data items in this category are as follows:

PD_SPEC [Scheme scheme3]

The PD_SPEC data items describe the specimen used to measure the diffraction data. The data item _pd_spec_preparation describes how the specimen that was used to measure the diffraction data was treated, not how the sample was prepared (PD_PREP) or characterized ( PD_CHAR).

The PD_SPEC data items are also used to describe how the specimen was mounted for the diffraction experiment. For example, _pd_spec_mount_mode and _pd_spec_orientation describe the measurement geometry, while _pd_spec_shape and _pd_spec_size_* describe the specimen shape and size.

3.3.4.3. Instrument calibration and description

| top | pdf |

The data items in these categories are as follows:

(a) PD_CALIB [Scheme scheme4]

(b) PD_CALIBRATION [Scheme scheme5]

(c) PD_INSTR [Scheme scheme6]

(d) Part of PD_DATA [Scheme scheme7]

The bullet ([\bullet]) indicates a category key.

Calibration information can be placed in the PD_CALIB and PD_CALIBRATION categories. The _pd_calibration_* data items are descriptive and will not appear in a loop. The _pd_calib_* items may be looped to describe multiple detectors. Correction values for [2\theta] can be given using the _pd_calib_2theta_offset and _pd_calib_2theta_off_* data items. A calibration equation can be given using _pd_calibration_conversion_eqn. When multiple detectors are used, _pd_calib_detector_response is used to indicate the relative performance of each detector. The detector deadtime is specified using the core data item _diffrn_detector_dtime (which cannot be looped by detector).

If an internal standard is added to the sample for calibration, this information is specified using _pd_calib_std_internal_name to specify the material added and _pd_calib_std_internal_mass_% to specify the amount.

When a set of calibration intensities is measured using an external standard, it is possible to include the measurements and the derived results in a separate CIF block. A data block would then use _pd_calib_std_external_block_id to link to the block containing the calibration information. See Section 3.3.7[link] for a discussion of block pointers and block IDs. Note that the use of a unique name for the block ID allows the calibration information to be stored in a separate file, so that the calibration CIF need not be repeated in every CIF that references it.

The PD_INSTR section of the pdCIF dictionary contains terms that describe the instrument used. For example, the instrument or laboratory location is given using _pd_instr_location. The instrument type can be indicated using _pd_instr_geometry. The instrument geometry can be described in much greater detail using several data items. The geometry is described in terms of four regions of the experiment: radiation source to monochromator (src/mono); monochromator to specimen (mono/spec); specimen to analyser (spec/anal); and analyser to detector (anal/detc). If no monochromator is present, the first two regions are combined into radiation source to specimen (src/spec). If no analyser is present, the last two regions are combined into specimen to detector (spec/detc). Thus two, three or four sets of values describe the dimensions of the instrument and the collimation. For example, _pd_instr_dist_src/mono would be used to specify the distance between the radiation source and the monochromator. Alternatively, _pd_instr_dist_src/spec would be used to specify the distance between the radiation source and the specimen if no monochromator was present.

Two methods may be used to describe the slits limiting the divergence in the equatorial plane. The angular divergence allowed by the slits in degrees can be specified using the _pd_instr_divg_eq_* data items. Alternatively, the dimensions of the slits in the equatorial direction in millimetres may be specified using _pd_instr_slit_eq_*. The dimensions of these slits in the axial direction, i.e. the direction perpendicular to the equatorial plane and containing the incident or diffracted beam as appropriate, are specified in millimetres using _pd_instr_slit_ax_*. The axial slit lengths, along with the _pd_instr_dist_* distances, are useful for estimating the low-angle peak asymmetry (Finger et al., 1994[link]). Note that angular divergence in the axial plane is not a well defined concept for line-focus instruments, but can be specified, where appropriate, using _pd_instr_divg_ax_*.

The axial and equatorial directions are shown schematically in Fig. 3.3.4.1[link]. The equatorial plane contains the equatorial direction vectors, as well as the incident beam, the diffracted beam and the scattering vector. The axial plane is perpendicular to the equatorial plane and contains the sample centre, which is the point where the incident and diffracted beams meet. For area-detection instruments, the designations of axial and equatorial directions may be arbitrary.

[Figure 3.3.4.1]

Figure 3.3.4.1 | top | pdf |

The axial and equatorial directions in a powder-diffraction experiment.

Soller collimators are described using _pd_instr_soller_eq_* data items rather than _pd_instr_divg_eq_* data items. It is common practice to specify the Soller collimation in arc-minutes (e.g. 30′). However, pdCIF defines these items to have units of degrees, so 30′ would be recorded in the CIF as 0.5. It is not usual to limit the axial divergence, except to reduce low-angle asymmetry, but if this is done, the _pd_instr_soller_ax_* data items can be used to define this.

For constant-wavelength instruments, it is common to have a monochromator or filter either before the sample, or after the sample (an analyser), or sometimes both. This is described using _pd_instr_monochr_pre_spec and _pd_instr_monochr_post_spec. It is rare, but possible, to have both a filter and a monochromator in the same location. The BT-1 neutron powder diffractometer at NIST uses both a Cu(311) monochromator and a graphite filter to attenuate the [\lambda/2] component. In this case, the two elements would be placed in a loop: [Scheme scheme8]

Note that the monochromator and analyser takeoff angles are given using _pd_instr_2theta_monochr_pre and _pd_instr_2theta_monochr_post. It is useful to record these values for X-ray studies, as they are needed for proper polarization corrections.

In a conventional Bragg–Brentano diffractometer, the divergence slits limit the illumination area at the sample. However, since the [\varphi] axis (the sample [\theta] axis) is usually set to bisect the [2\theta] angle of the detector, the actual length of the area of the sample that is illuminated changes with [2\theta]. One should choose divergence slits so that the beam does not illuminate areas outside the sample at the lowest diffraction angle used. An alternative method for data collection is to have a divergence slit that opens as [2\theta] increases, so that a constant area of the sample is illuminated. This is known as a [\theta]-compensating slit. Using a [\theta]-compensating slit provides a better signal-to-noise ratio at larger [2\theta] values, but means that the diffraction intensities have to be normalized to compensate for the change in illumination. It also introduces greater optical aberrations with [2\theta]; a flat plate becomes an increasingly worse approximation to the curved sample geometry in the true Bragg–Brentano geometry.

The use of a variable divergence slit can be recorded in the form: [Scheme scheme9] Note that if _pd_instr_cons_illum_flag is not specified, the value is assumed to be no, indicating that a fixed-width divergence slit has been used.

The beam size can be specified in two different ways: as the size at the source, using _pd_instr_source_size_ax and _pd_instr_source_size_eq, or as the size at the sample position, using _pd_instr_beam_size_ax and _pd_instr_beam_size_eq. Note that the size of the beam at the sample differs from the illumination length described above except when the sample is perpendicular to the beam. When a variable-divergence slit is in use, the beam size at the sample changes with [2\theta], so if this size is known directly, the _pd_instr_beam_size_eq data item can be included in the loop containing the diffraction intensities. Similarly, in a constant-divergence instrument, where the illumination length changes with [2\theta], the illumination length can be specified in the loop using _pd_instr_var_illum_len.

There are also several data items in the core CIF dictionary that should be present in the description of the instrument in a pdCIF. Use _diffrn_radiation_probe and _diffrn_radiation_type to specify the type of radiation used and _diffrn_detector_type to specify the detection type.

3.3.4.4. Observations and measurement conditions

| top | pdf |

The data items in these categories are as follows:

(a) Part of PD_DATA [Scheme scheme10]

(b) PD_MEAS_INFO [Scheme scheme11]

(c) PD_MEAS_METHOD [Scheme scheme12]

The arrow ([\rightarrow]) is a reference to a parent data item.

The item _pd_data_point_id identifies each entry in the list of measured, processed or simulated intensities. It is the only item in the PD_DATA category that actually begins with the string _pd_data_, and is included here for convenience. If the list of intensities is split across several distinct loops, the role of this identifier may be adopted by other identifiers, such as _pd_meas_point_id in an isolated list of measured intensities.

The _pd_meas_* data items contain unprocessed measurements and documentation on the instrumental settings used for the measurements. Note that the choice of the data items used to represent this information is determined by the type of diffraction instrument, as well as how the measurement was conducted. This will be discussed further in Section 3.3.8[link]. However, some _pd_meas_* data items are independent of the instrument type. For example, the use of _pd_meas_datetime_initiated is good practice, as is use of the _pd_meas_info_author_* data items. It is probably good practice to record the number of data points in _pd_meas_number_of_points for the benefit of people who might read the CIF, but there is no requirement that this item be present. This means that software should determine the number of data points directly when reading the CIF, rather than relying on the presence of a value for _pd_meas_number_of_points.

3.3.5. Analysis

| top | pdf |

The categories relating to the information derived from the measurements are as follows:

Processed intensities, positions and data processing (§3.3.5.1[link])
PD_DATA group
 PD_DATA (items beginning with _pd_proc_*)
PD_PROC group
 PD_PROC_INFO
 PD_PROC_LS
Simulated intensities and their positions (§3.3.5.2[link])
PD_DATA group
 PD_DATA (items beginning with _pd_calc_*)
PD_CALC group
 PD_CALC
Diffraction peak table (§3.3.5.3[link])
PD_PEAK group
 PD_PEAK
 PD_PEAK_METHOD
Reflection assignments and intensities (§3.3.5.4[link])
 REFLN

In Rietveld and other studies, processed or simulated intensities are presented alongside measured values. This leads to the presence of both derived and measured values in the same category (PD_DATA). However, the purposes of the data items that refer to processed and simulated data points are made clear by the way they are named. Overall descriptions of processed and simulated intensity data are covered by the categories PD_PROC_INFO, PD_PROC_LS and PD_CALC. The two categories PD_PEAK and PD_PEAK_METHOD are used to describe lists of peak positions, which would typically be used to search and match powder profiles. Some additional data items relevant to the table of Bragg reflections are defined as additions to the existing REFLN category in the core CIF dictionary.

3.3.5.1. Processed intensities, their positions and processing information

| top | pdf |

The data items in these categories are as follows:

(a) Part of PD_DATA [Scheme scheme13]

(b) PD_PROC_INFO [Scheme scheme14]

(c) PD_PROC_LS [Scheme scheme15]

The pdCIF dictionary distinguishes between values that are measured directly and values that are derived from these observations. For example, in a constant-wavelength instrument, diffraction intensities are recorded as a function of [2\theta]. One may derive d-space values from the [2\theta] values using the value of the wavelength and corrections for the [2\theta] zero-point error and the sample displacement. One may also derive a new set of data points from the observations, for example by summing adjacent data points when the increment between the data points is much smaller than is warranted by the peak widths. For peak searching and other non-quantitative purposes, the diffraction intensities may be smoothed or otherwise modified. Note that the unprocessed measurement values are retained using the data items _pd_meas_*. Since the original measurements are still available, modifications like these do not result in the loss of the original data. In fact, by placing processed values in multiple blocks, a single CIF may contain measurements that have been processed in more than one way.

It is good practice to use the _pd_proc_info_author_* and _pd_proc_info_datetime data items. It is also a good idea to describe how the measurements were processed using _pd_proc_info_data_reduction.

The _pd_proc_* data items in this list may be used to calibrate the [2\theta] or energy values of the data. These are defined in the items _pd_proc_2theta_corrected, _pd_proc_2theta_range_*, _pd_proc_d_spacing, _pd_proc_energy_*, _pd_proc_recip_len_Q and _pd_proc_wavelength.

When corrections, scaling or other processing, such as averaging or smoothing, are applied to the intensities, the results are stored using the _pd_proc_intensity_* data items. Note that if the number of data points does not change, it might be most convenient to include the processed intensities in the same loop as the observed values. This is not always possible, so these items can be placed in a separate loop if there is no longer a one-to-one correspondence between the [2\theta] or energy positions for the _pd_proc_intensity_* values and the _pd_meas_counts_* or _pd_meas_intensity_* values.

For energy-dispersive measurements, the incident spectrum must be determined for normalization. This can be recorded using _pd_proc_intensity_incident. For other types of normalization, _pd_proc_intensity_norm should be used.

For full-pattern fitting, there is a series of _pd_proc_ls_* data items for recording settings and results. For example, agreement factors can be recorded using the _pd_proc_ls_prof_*_factor data items. Some data items may be included in the loop(s) containing the measured or the processed data: _pd_proc_ls_weight specifies the weight assigned to each point and _pd_proc_intensity_bkg_calc specifies the fitted background. Note that background values are usually generated by extrapolation from fixed values set during the refinement or are determined from a function that is fitted to the observations, and occasionally both are used together. The function that has been fitted can be described using _pd_proc_ls_background_function, while fixed points are listed using _pd_proc_intensity_bkg_fix. If sections of the pattern are not fitted, this is indicated using _pd_proc_info_excluded_regions.

3.3.5.2. Simulated data

| top | pdf |

The data items in these categories are as follows:

(a) Part of PD_DATA [Scheme scheme16]

(b) PD_CALC [Scheme scheme17]

It is common to calculate powder-diffraction intensities from a crystallographic model. This is necessary for Rietveld refinements, where the model is fitted to the experimentally observed intensities. It is also used to simulate the diffraction pattern of a material for which the structure is known, perhaps for comparison with a measured diffraction pattern.

A crystallographic model can be described in CIF using data items from the core CIF dictionary, as described in Chapter 3.2[link] . To record the results of the simulation, the data items _pd_calc_intensity_net or _pd_calc_intensity_total are used. The difference between these two data items depends on the treatment of background. If the pattern is simulated with a fitted background added to it, _pd_calc_intensity_total is used; otherwise _pd_calc_intensity_net is used. The values will typically be placed in a loop with the processed (_pd_proc_*) data items or the observed (_pd_meas_*) data items. If neither observed nor processed data are present (e.g. for a simulation), or if, for some reason, the simulation has been performed with a different [2\theta] range or step size, the appropriate _pd_proc_* data items are used to define the [2\theta] values etc. used for the simulation.

3.3.5.3. Diffraction peak table

| top | pdf |

The data items in these categories are as follows:

(a) PD_PEAK [Scheme scheme18]

(b) PD_PEAK_METHOD [Scheme scheme19]

The bullet ([\bullet]) indicates a category key. The arrow ([\rightarrow]) is a reference to a parent data item. Items in italics are defined in the core CIF dictionary.

When diffraction intensities are first measured, particularly when attempting to identify unknown phases in a material, the first step in the analysis is often to compile a list of peak positions. These peak positions are commonly used to search the Powder Diffraction File, which contains lists of peak heights and positions for approximately 100 000 materials (International Centre for Diffraction Data, 2004[link]).

Information on diffraction peaks is recorded in the PD_PEAK section of the pdCIF. Peak positions are recorded using _pd_peak_2theta_maximum or _pd_peak_2theta_centroid, for positions determined from the intensity maxima or from the peak centroids, respectively. It is also possible to record peak positions using _pd_peak_d_spacing. Peak intensities are recorded using _pd_peak_intensity and _pd_peak_pk_height, for the integrated peak area or the intensity value at the peak maximum, respectively. Peak widths are recorded using _pd_peak_width_2theta and _pd_peak_width_d_spacing.

A separate loop is used to list reflections, as will be discussed in Section 3.3.5.4[link]. To link reflections to peaks (one peak may consist of many reflections), each peak is assigned a unique code using _pd_peak_id, which is then referenced in the reflection table using _pd_refln_peak_id.

When intensities are measured using radiation with more than one wavelength, for example when both Cu Kα1 and Kα2 radiation are used or when a monochromator passes both [\lambda] and [\lambda/2] radiation, peaks may be assigned a wavelength symbol using _pd_peak_wavelength_id, where the wavelength symbol is defined in a separate _diffrn_radiation_wavelength_id loop. However, for many experiments, the assignment of wavelengths to peaks will be impractical owing to reflection overlap. It is usually better practice to specify wavelength labels in the reflection table using _pd_refln_wavelength_id.

3.3.5.4. Reflection assignments and intensities

| top | pdf |

In addition to the REFLN data items defined in the core CIF dictionary, the following items are defined:

REFLN [Scheme scheme20]

The arrow ([\rightarrow]) is a reference to a parent data item. The dagger ([\dagger]) indicates a deprecated item, which should not be used in the creation of new CIFs. Items in italics are defined in the core CIF dictionary.

In a single-crystal experiment, a reflection table contains the initial experimental observations for structural analysis. In contrast, the reflection table for a powder-diffraction experiment is a derived result that depends on the model used to apportion intensity between overlapping reflections. Another difference is that in a single-crystal experiment, the reflection list will refer to only one phase (one hopes), while it is common to have reflections from more than one phase in a powder-diffraction reflection list.

A list of reflections in a powder-diffraction pattern is commonly generated by Rietveld analysis, where Hugo Rietveld's algorithm (Rietveld, 1967[link], 1969[link]) is used to estimate the intensity of each reflection. Alternatively, when the structure of one or more phases is not known, it is possible to use full-pattern intensity-extraction methods such as the algorithms developed by Pawley (1981[link]) or Le Bail et al. (1988[link]). In fact, intensity information obtained by full-pattern intensity extraction is often used for ab initio structure determination.

Most of the information in the reflection table will be defined using data items from the core CIF dictionary (see Section 3.2.2.2[link] and Chapter 4.1[link] ). For example, _refln_index_h, _refln_index_k and _refln_index_l will be used for the indices. The structure factors and reflection intensities are specified using _refln_intensity_calc, _refln_intensity_meas, _refln_F_squared_calc and _refln_F_squared_meas; reflection positions are defined using _refln_d_spacing. To link a reflection with a powder-diffraction peak, the pdCIF data item _pd_refln_peak_id is used. The value for _pd_refln_peak_id serves as a pointer to an entry in the peak table which has been labelled, using the data name _pd_peak_id, with the same symbol. Likewise, to link a reflection to a phase, the pdCIF data item _pd_refln_phase_id points to a phase defined using _pd_phase_id in the phase table. Since a single reflection may be observed with more than one wavelength, for example, with [\lambda/2] or Kα2 wavelengths, the pdCIF dictionary defines a wavelength link, _pd_refln_wavelength_id, that defines a wavelength label. However, since version 2.1, the core CIF dictionary defines _refln_wavelength_id and this should be used in preference to _pd_refln_wavelength_id. The data items _refln_wavelength_id and _pd_refln_wavelength_id both point to a wavelength label defined using _diffrn_radiation_wavelength_id.

The International Centre for Diffraction Data abstracts peak positions and heights for inclusion in the Powder Diffraction File. This information would be found in the _pd_peak section of a pdCIF. However, in many studies, particularly in Rietveld refinements, peak tables are never generated. In principle, it should be possible to calculate peak positions and peak heights (or better still, peak areas) from the information in a reflection table. An algorithm for this would be very useful.

3.3.6. Atomicity, chemistry and structure

| top | pdf |

The structural model of a compound determined by powder-diffraction methods can be described by the data items in the core CIF dictionary. However, for a powder-diffraction study of a mixture of phases, the PD_PHASE category is used to list the phases present. This is the only category in the pdCIF dictionary that extends the description of the structural model beyond that covered by items in the core CIF dictionary.

3.3.6.1. Table of phases

| top | pdf |

The data items in this category are as follows:

PD_PHASE [Scheme scheme21]

The bullet ([\bullet]) indicates a category key.

When a sample contains more than one phase, the PD_PHASE data items are used to create a table describing the phases present. For example, the name and abundance of each phase can be specified using _pd_phase_name and _pd_phase_mass_%, respectively.

Two types of pointers can also be defined:

(i) Since the crystallographic description of each phase must be incorporated in a separate data block, _pd_phase_block_id contains the unique block ID (see Section 3.3.7[link]) pointing to the block containing the data for the phase.

(ii) An arbitrary label is assigned to every phase using _pd_phase_id so that reflections can be assigned to a phase using _pd_refln_phase_id. This is discussed further in Section 3.3.5.4[link].

3.3.7. File metadata

| top | pdf |

The many data items in the core dictionary that decribe file auditing and history cover most of the metadata requirements of a pdCIF, but two new data items in the pdCIF category PD_BLOCK are introduced to provide a specific mechanism for identifying and relating individual data blocks.

Data items in this category are as follows:

PD_BLOCK [Scheme scheme22]

The data item _pd_block_id is used to define a unique name for each data block. This name is used so that one data block may reference another data block. Since CIF blocks may be separated into different files, or many CIFs from different sources may be grouped into a single file, the block ID provides a robust mechanism for maintaining references between blocks, independent of how CIF blocks have been arranged between files. The intent is that a site that archives pdCIFs will construct an index to _pd_block_id names that can be used to resolve block ID references.

The definition for _pd_block_id gives a procedure for creating a _pd_block_id name that is extremely unlikely to be duplicated. Other mechanisms for creating unique names can also be used: for example, using a web page name (URL) could be appropriate if care is taken never to reuse the URL.

The need for the block ID/block pointer mechanism is demonstrated by the following example. Consider a case where a neutron powder diffraction data set and an X-ray powder diffraction data set have been used together to determine a single structural model for a single crystalline phase. CIF does not allow the two data sets to be placed in a single block, since this would require two independent loops of observations where each loop uses some of the same data names. One can create a CIF with two blocks and include the structural model in the block that contains either of the two data sets. However, if this is done, a logical link is needed between the two blocks to make it clear that the structural model was derived from both data sets. It is better practice to place the structural model in a third data block, as this emphasizes the fact that the model is derived from both data sets. Again, logical links to the data sets are needed.

In both these cases, the data item _pd_block_diffractogram_id would be included in the data block containing the structural model and will point to _pd_block_id values assigned in the data blocks containing the diffraction data to establish the connection between the data sets and the structural model. The presence of more than one value for _pd_block_diffractogram_id, through use of a loop, indicates that multiple data sets were used and thus these structural results are from a combined refinement. Sometimes, powder and single-crystal diffraction data are used together (most commonly to team X-ray single-crystal diffraction data with neutron powder diffraction data). In this case, _pd_block_diffractogram_id will point to two _pd_block_id values, where one is assigned to the single-crystal data set.

In contrast to the example above, in which block pointers are used to link a single structural model to multiple data sets, another application for these pointers is for describing materials that contain more than one phase. In this case, _pd_phase_block_id is placed in the data block containing the data set to link it to the blocks defining the phases.

In summary, three types of links between data blocks are defined.

(i) _pd_block_diffractogram_id connects a phase to one or more data-set blocks;

(ii) _pd_phase_block_id connects a data set to one or more phase blocks;

(iii) _pd_calib_std_external_block_id connects a block to measurements used to provide calibration constants used in the block.

It is good practice to use both _pd_block_diffractogram_id and _pd_phase_block_id in a pdCIF with multiple blocks.

3.3.7.1. Use of block pointers

| top | pdf |

More complex link structures will be needed when multiple data sets and multiple phases occur together. Example 3.3.7.1[link] outlines a pdCIF reporting the results of a TOF powder-diffraction study of a physical mixture of nickel and silicon powders in which two separate diffraction banks, measured at two different Bragg angles, were used. In this case, five CIF blocks are used. The first CIF block reports the overall and publication details. The next two CIF blocks report crystallographic information for each phase and the last two blocks report the observed, processed and calculated diffraction intensities and reflection tables.

Example 3.3.7.1. A CIF with multiple data blocks, demonstrating a suitable construction when multiple data sets and multiple phases occur together.

[Scheme scheme23] [Scheme scheme24] [Scheme scheme25] [Scheme scheme26]

A second purpose for _pd_block_id is to provide a mechanism for tracking successive modifications to a CIF. Consider the case where a data set is obtained at a user facility and the resulting measurements are distributed as a CIF. In this file, a value is supplied for _pd_block_id based on the time when the measurements were made. At a later time, when these observations are analysed, a new CIF is created, containing both the original measurements and the results from the analysis. Rather than replace the original value for _pd_block_id, the data item can be placed in a loop and another value, defining a second block ID, can be added. This will indicate the connection to the initial CIF, since the original block ID is retained.

A potential future use for block pointers may be to reference non-CIF data files that contain large two- and three-dimensional data structures. This is expected to become increasingly important as neutron and synchrotron instruments are constructed that cover increasing ranges of solid angle. As mentioned in Section 3.3.2[link], CIF is not well suited to these complex, large and possibly irregular measurement arrays. The NeXus format has been developed by a consortium of synchrotron and neutron laboratories to address these concerns and is currently being used for a variety of scattering applications (NeXus, 1999[link]). The NeXus format is based on the platform-independent HDF binary standard (HDF, 1998[link]). The use of block pointers to resolve references to non-CIF documents will require additional definitions.

3.3.8. pdCIF for storing unprocessed measurements

| top | pdf |

While many researchers prepare a CIF only when a project is complete, there are good reasons for preparing a pdCIF when the diffraction data are measured, as this is the best time to document how the measurement was performed. Much of the instrumental information will remain unchanged for all pdCIFs from a given diffraction instrument, so it is a good idea to prepare a file that describes each of the common settings for an instrument. This file will probably contain some of the following data items and their associated values:

(i) The _pd_instr_* items, such as the instrument type in _pd_instr_geometry, the size of the instrument and the collimation in _pd_instr_dist_* and _pd_instr_divg_*, and monochromatization in _pd_instr_monochr_* (see Section 3.3.4.3[link])

(ii) Depending on how the calibration is performed, it may be appropriate to include _pd_calib_* items.

(iii) Information about the radiation source should be specified using the _diffrn_radiation_* and _diffrn_source_* data items.

(iv) Detector information should be specified using _diffrn_detector_* items, for example, the detector type in _diffrn_detector_type and perhaps calibration values such as the deadtime (in _diffrn_detector_dtime).

A second section of the pdCIF will contain information specific to the experiment, such as the diffraction conditions (i.e. pressure and temperature) recorded using the _diffrn_ambient_* data items. Sample and specimen information will appear in the _pd_prep_*, _pd_spec_* and _pd_char_* data items.

A third section of the pdCIF contains the observations. The data items used to specify the unprocessed observations will vary with the type of instrument used, as described in Sections 3.3.8.1[link] to 3.3.8.10 below.

3.3.8.1. Single pulse-counting detectors

| top | pdf |

In the most common measurement method, where a single pulse-counting detector is scanned over a range of [2\theta], the _pd_meas_* entries (see Section 3.3.4.4[link]) will be of the form shown in Example 3.3.8.1[link]. If the data were scanned using a variable step size, the observations might be given as shown in Example 3.3.8.2[link]. Note that when _pd_meas_counts_* is used, the values given must be counts, so that the standard uncertainty will be the square root of the intensity values. This means that the intensity values must not be scaled, for example if the values were counts per second; otherwise the statistical uncertainty estimates will be incorrect.

Example 3.3.8.1. Measurements from a single pulse-counting detector with constant-step scan.

[Scheme scheme27]

Example 3.3.8.2. Measurements from a single pulse-counting detector with variable-step scan.

[Scheme scheme28]

3.3.8.2. Detectors that do not count pulses

| top | pdf |

When the method used to detect intensities does not count individual quanta as they hit the detector, for example, the digitization of intensities recorded on film or on an imaging plate, or even with data recorded using a detector having a built-in deadtime correction, the standard-uncertainty values are not the square root of the intensities. [Note that when the actual deadtime correction is known, it is best to incorporate this scaling into the monitor value (see _pd_meas_counts_monitor in Section 3.3.4.4[link]) or else save the uncorrected measurements and create a second set of corrected intensity values as _pd_proc_intensity_net (see Section 3.3.5.1[link]).] The _pd_meas entries for an experiment using non-pulse-counting detection will look like the examples given in Section 3.3.8.1[link], except that the data loop will be in the form [Scheme scheme29] or [Scheme scheme30] If standard uncertainties for the intensity values are known, they can be given using the conventional notation [Scheme scheme31] Note that when _pd_meas_intensity_* is used, it is best to specify _pd_meas_units_of_intensity as well.

3.3.8.3. Multiple detectors

| top | pdf |

At present, CIF does not offer the ability to construct true multi-dimen­sional data structures. However, many instruments with multiple detectors produce reasonably tractable numbers of data points. For such instruments, it is possible to include an additional data item, _pd_meas_detector_id, in the loop with the data to indicate the detector that made the observation.

In Example 3.3.8.3[link], four detectors placed 20° apart are referenced with arbitrarily chosen labels A, B, C and D. Note that the detector characteristics will typically be specified in a separate calibration loop containing terms such as _pd_calib_detector_id, _pd_calib_detector_response and _pd_calib_2theta_offset. The labels given for _pd_calib_detector_id should match those in _pd_meas_detector_id.

Example 3.3.8.3. Identifying intensities from multiple detectors.

[Scheme scheme32]

3.3.8.4. Energy-dispersive X-ray detection

| top | pdf |

For energy-dispersive X-ray diffraction, an X-ray detector is placed at a fixed value of [2\theta] and a diffractogram is measured on a multichannel analyser. The channel number is then calibrated to yield photon energies. From the energy and [2\theta] angle, a d-spacing or Q value [(Q = 4 \pi\sin\theta/\lambda)] is calculated for each diffraction point. Note that energy, d spacing or Q are not the experimental independent variable. Rather, they result from processing, since calibration information is required. The calibration equation should be described in _pd_calibration_conversion_eqn.

In Example 3.3.8.4[link], the nominal [2\theta] setting is 6.5°, but the actual position (determined by prior calibration) is 6.6071°, so the difference is indicated using a _pd_calib_2theta_offset value (see Section 3.3.4.3[link]).

Example 3.3.8.4. Measurements from an energy-dispersive X-ray diffraction experiment.

[Scheme scheme33]

3.3.8.5. Neutron time-of-flight detection

| top | pdf |

Neutron time-of-flight (TOF) detection in theory should be no different from energy-dispersive X-ray detection, but TOF instruments record complex three-dimensional data structures, where diffraction intensities are recorded as a function of time for as many as several hundred detectors. For some instruments, both the position along the detector and the time of flight are recorded, so there may be effectively thousands of detectors. To add even further complexity, the data may be binned in different time steps for detectors at different [2\theta] values. CIF is likely to be cumbersome for the storage of unprocessed measurements from TOF instruments, owing to the one-dimensional nature of CIF, but it could be useful to translate files from one binary format to another using CIF as a common intermediate. To do this, a single loop is used for all data points, where each detector (or detector section, in the case of a position-sensitive detector) is assigned a detector ID. In a second loop, the detector ID values are defined. In addition to [2\theta], _pd_meas_angle_omega and _pd_meas_angle_chi are defined where needed (Example 3.3.8.5[link]).

Example 3.3.8.5. Measurements from a neutron time-of-flight diffraction experiment.

[Scheme scheme34]

TOF data are usually reduced to a small number of `banks' consisting of intensity as a function of d space or Q, where multiple detectors are summed. Data in this form can be recorded using a loop containing _pd_proc_d_spacing and _pd_proc_intensity_net. A data block is needed for each bank.

3.3.8.6. Digitized film and image plates

| top | pdf |

To record intensities from digitized X-ray film or from image plates properly requires the storing of two-dimensional data structures, which in some cases can be accommodated through imgCIF (see Chapters 2.3[link] and 3.7[link] ). However, it is possible to record a one-dimensional scan using _pd_meas_position and _pd_meas_intensity_total (not _pd_meas_counts_total!). _pd_proc_2theta_corrected values can then be assigned using calibration information, and they can then be included in the same loop, as in Section 3.3.8.4[link].

3.3.8.7. Direct background measurements

| top | pdf |

For some diffraction experiments, particularly for the determination of radial distribution functions, measurements are made for background scattering from the diffraction instrument and from the sample container. When this is done, the values can be included in a single loop using _pd_meas_counts_background, _pd_meas_counts_container and _pd_meas_counts_total.

3.3.8.8. Noting sample orientation

| top | pdf |

For texture measurements, intensity measurements can be made as a function of different sample setting angles. These setting angles can be specified using _pd_meas_angle_chi, _pd_meas_angle_omega and _pd_meas_angle_phi. The change in these values may be specified by including these data items in the loop with the diffraction intensities. In some cases, it may be more convenient to separate measurements with different setting angles into different blocks. In this case, the values for the setting angle(s) that are invariant will be set outside of a loop.

It is common in powder diffraction to reduce preferred orientation and improve crystallite averaging by rocking or rotating the sample. This is indicated by specifying the axis used for rocking, usually [\varphi] for capillary specimens or [\chi] for flat-plate specimens, as _pd_meas_rocking_axis. The data item _pd_meas_rocking_angle is used to record the angular range through which the sample is rocked, where 360 indicates that one complete revolution occurs during each counting period. Numbers greater than 360 are possible.

3.3.8.9. Use of an incident-intensity monitor

| top | pdf |

For radiation sources for which the intensity may vary, such as synchrotron-radiation sources, the intensity of the incident radiation is measured using an incident-intensity monitor. This value may be specified for every data point using _pd_meas_counts_monitor (or _pd_meas_intensity_monitor) by including this data item in the loop with the diffraction intensities. For some instruments, counting times are set so that the same number of monitor counts are measured for each data point. If this is the case, _pd_meas_counts_monitor will be the same for every data point and need not be included in the loop.

3.3.8.10. Recording detector livetime

| top | pdf |

The detector deadtime is often more a function of the counting electronics than of the intrinsic properties of the detector. In these circumstances, the counting circuit may provide a gating signal that indicates when the electronics are processing an event versus when the circuit is idle and waiting for an event to process. From this gating signal, a detector livetime signal can be generated. Livetime is a better way to correct intensities than applying a deadtime correction, because if appreciable numbers of events are processed but are not counted (for example, counts due to fluor­escence), the actual deadtime can be quite high, even though the recorded number of counts can be quite low. To use the livetime signal, the count time can be multiplied by the livetime or the livetime can be treated as a monitor (see Section 3.3.8.9[link]). If an incident-intensity monitor and a livetime are both available, the _pd_meas_intensity_monitor value can contain the incident intensity times the livetime.

3.3.9. Use of pdCIF for Rietveld refinement results

| top | pdf |

One of the major aims of the development of the pdCIF definitions was to be able to communicate the results of Rietveld refinements, and this is expected to be the most common use for pdCIF. To aid the development of software that prepares pdCIF output from Rietveld refinements, this section describes the blocks and loops to be found in a pdCIF, noting variations due to the type of Rietveld refinement. Programmers may also wish to look at the GSAS2CIF program, which creates CIFs for a wide range of types of diffraction data, and for multiple data sets and phases (Toby et al., 2003[link]).

It is valuable for the CIF to contain the structural model(s), the observed powder-diffraction intensities and the calculated powder-diffraction intensities so that the fit of the model to the observed diffraction pattern can be viewed graphically. It is the present author's firm belief that it is impossible to judge the quality of a Rietveld refinement by R factors or any other numerical metric, since these values describe not just how well the structural model fits the measurements, but also how well the background and peak shape are fitted as well. Very poor models can have good R factors and [\chi^2] values if there is a significant amount of non-Bragg scattering that has been well fitted. On the other hand, with high-resolution observations measured to excellent precision, even trivial imperfections in the peak shapes can result in poor agreement factors. There is no substitute for the visual examination of a plot of the observed and calculated patterns, optimally at more than one magnification level. The program pdCIFplot (Toby, 2003[link]) plots the observed and calculated powder-diffraction intensities in a pdCIF and allows the fit to be examined in more detail than can be provided by a figure showing the whole profile at once.

3.3.9.1. A single phase

| top | pdf |

When a single set of diffraction measurements is used to model a single phase, a pdCIF will usually contain only one block. There will be several important loops present.

One loop will contain atomic parameters, such as coordinates. The unit cell must also be specified.

A second loop will contain the reflection table.

A third loop will contain the observed (or processed) diffraction measurements and the simulated pattern. Other items that should be included in this loop are the least-squares weights (usually σ−2, where σ is the standard uncertainty) so that it is possible to determine the quality of the fit in individual regions. Weight values of zero can also be used to indicate that data points have been excluded from the refinement. Since background fitting is quite important in Rietveld analysis, it is also valuable to include the background values. Thus, this loop should specify:

(i) the ordinate of the Rietveld plot, using one or more of: _pd_meas_2theta_scan, _pd_meas_time_of_flight, _pd_proc_2theta_corrected, _pd_proc_d_spacing or _pd_proc_recip_len_Q; alternatively the ordinate can be specified using either _pd_meas_2theta_range_* or _pd_proc_2theta_range_*, where _* is _min, _max and _inc outside the loop.

It is recommended that all CIFs describing the results of a Rietveld refinement include either _pd_proc_d_spacing or _pd_proc_recip_len_Q.

(ii) The observed (or processed) intensity values, using the items _pd_meas_counts_total, _pd_meas_intensity_total, _pd_proc_intensity_total or _pd_proc_intensity_net.

(iii) The background, using the item _pd_proc_intensity_bkg_calc.

(iv) The least-squares weights, using the item _pd_proc_ls_weight. If these weights are not specified, then it must be presumed that all points have been used in the refinement and that the weights are the reciprocal of the intensity values (if _pd_meas_counts_total was used) or the reciprocal of the intensity standard uncertainties, if specified.

(v) The calculated pattern should appear using either _pd_calc_intensity_net or _pd_calc_intensity_total.

It is good practice always to include at least one data item from each entry in the list above.

Apart from the information contained in these loops, information from almost all sections of the pdCIF dictionary can be valuable. Such items include data items that define how the diffraction measurements were made, how the sample was prepared and characterized, how the refinement was performed, and least-squares parameters and R factors. A template and an example pdCIF showing the combined use of pdCIF and core data items form part of the Acta Crystallographica instructions for authors at http://journals.iucr.org/services/cif/powder.html .

3.3.9.2. Multiple phases

| top | pdf |

When more than one phase is present, multiple CIF blocks are needed. The resulting CIF will contain much the same information as would be found in a single-phase pdCIF, as described in Section 3.3.9.1[link]. However, there will be a separate block for each phase containing information specific to that phase, such as the unit cell and the loop containing the atomic parameters.

The CIF will usually (see Section 3.3.7[link]) contain one additional block with the observed and calculated pattern and a reflection table, as well as the other data items that define how the diffraction measurements were made, how the refinement was performed etc. While reflection tables for each phase can be placed in each phase block, it is better to include a single reflection table in the block that contains the diffraction data. This block will also contain a phase table that uses the block pointer _pd_block_diffractogram_id to link to the phase blocks. The phase blocks can also be linked to the data block using the block pointer _pd_phase_block_id. For most Rietveld refinements, each phase is allowed to have different profile parameters, so _pd_proc_ls_profile_function should also be included in the phase-table loop.

3.3.9.3. One phase, multiple sets of measurements

| top | pdf |

It is fairly common to use more than one diffraction data set to determine a model for a single phase. Some examples include: combined refinement using both neutron and X-ray powder diffraction; use of multiple X-ray wavelengths to make use of anomalous dispersion; and the use of single-crystal X-ray and powder neutron diffraction data in a single refinement. For these cases, there will be a CIF block for each data set. Each of these blocks will contain a reflection table and a loop with the observed and calculated diffraction intensities, as described in Section 3.3.9.1[link].

As explained in Section 3.3.7[link], the resulting structural parameters could be placed in a block with one of the sets of diffraction data. However, it is better practice to create one additional block for these parameters, as it then becomes clear that the result is from a combined refinement. This is indicated by linking the phase block and the data-set blocks using a loop of _pd_block_diffractogram_id values in the phase block. The data-set blocks can also have a link to the phase information using the block pointer _pd_phase_block_id.

3.3.9.4. Multiple sets of measurements and phases

| top | pdf |

Multiple data sets may be used for mixtures as well as single phases. This is becoming increasingly common as more complex materials are studied using powder diffraction. The treatment of this case follows logically from that of Sections 3.3.9.2[link] and 3.3.9.3[link]. If there are M diffraction data sets and P phases, there will be P blocks containing the crystallographic parameters for each phase. There will be M blocks with the observed and calculated diffraction intensities, as well as reflection tables. Depending on the Rietveld software, there may be M × P sets of some parameters, for example phase fractions and profile descriptions. These parameters may be placed in the phase table loop within the data-set block(s).

Ideally, the same specimen, or at least the same sample, will be used for all measurements. Sometimes, however, different samples are used for combined refinements to extend the number of observations, despite the possibility that the samples might have slightly different structures or compositions. If there are S samples, there will be an additional S blocks that record the sample and specimen preparation and characterization information. Thus, in this case there will be a total of M + P + S blocks.

As before, the phase blocks will use the block pointers _pd_block_diffractogram_id to link to the data-set blocks. Likewise, the data-set blocks will have phase tables with _pd_phase_block_id values that link to the phase blocks. The sample blocks can use both _pd_block_diffractogram_id values and _pd_phase_block_id values to link to the the diffraction data and the analysis results. This is shown in the CIF in Example 3.3.7.1[link]. The program GSAS2CIF (Toby et al., 2003[link]) can create CIFs for multiple sets of measurements and phases.

3.3.10. Other pdCIF applications

| top | pdf |

As mentioned above, there are other applications for pdCIF than the storage of unprocessed measurements and the reporting of the results of a Rietveld refinement. This section describes the use of data items in other common pdCIF applications.

3.3.10.1. Simulated intensities

| top | pdf |

It is common to simulate a diffraction pattern from a known or hypothetical structural model. The structural model is recorded in CIF using core data items, such as _atom_site_label, _atom_site_fract_x, _atom_site_fract_y, _atom_site_fract_z, _atom_site_U_iso_or_equiv and _atom_site_occupancy, as well as the unit cell in _cell_length_* and _cell_angle_*. Calculated reflection intensities can be recorded using _refln_index_* and _refln_F_squared_calc, as described in Section 3.3.5.4[link]. The simulated pattern can be recorded using _pd_calc_* data items, as described in Section 3.3.5.2[link].

The simulated diffraction pattern will be determined not only by the structural parameters, but also by the type of experiment that is being simulated. For example, it is good practice to define data items to specify the type of radiation in _diffrn_radiation_probe, the wavelength in _diffrn_radiation_wavelength and the profile in _pd_proc_ls_profile_function.

3.3.10.2. Phase identification and indexing

| top | pdf |

For phase identification, a CIF will include unprocessed measurements, as described in Section 3.3.8[link]. Sample characterization information, for example chemical analysis information, can often aid phase determination. Characterization information is described in Section 3.3.4.1[link]. Similarly, sample preparation information can also be quite valuable (see Section 3.3.4.1[link]). Since preferred orientation or other artifacts of the measurement can make phase identification more difficult, it is a good idea to include specimen preparation and mounting information, as described in Section 3.3.4.2[link].

Peaks will be located and documented in the peak table, as discussed in Section 3.3.5.3[link]. In this case, it can be very helpful to specify _pd_peak_wavelength_id for peaks that are clearly a Kα2 component. Similarly, recording a peak width can also be helpful for some autoindexing programs.

To identify peaks by phase in the case where at least one phase in a material is known but other peaks remain unidentified requires the use of a reflection table and a phase table, as shown in Example 3.3.10.1[link]. This example shows a diffraction pattern with eight peaks, of which five have been identified as arising from a phase where the unit cell has not yet been determined. The peaks labelled B1, B2 and B3 are referenced in the peak table, but are not defined in the reflection loop. This implies that they arise from an unknown phase or phases. The remaining peaks A1 to A5 are referenced in the peak table. Note that the reflection table must include the _refln_index_* data items, even though no reflection indices are assigned. This is because CIF rules require that the _refln_index_* data items be present in this loop, as noted in the pdCIF dictionary definitions for _pd_refln_peak_id and _pd_refln_phase_id. The place holder ? is used to indicate that the reflection indices are not yet known.

Example 3.3.10.1. Phase identification using a reflection table and a phase table.

[Scheme scheme35]

Appendix A3.3.1

A3.3.1. Category structure of the powder CIF dictionary

| top | pdf |

Table A3.3.1.1[link] provides an overview of the structure of the powder CIF dictionary by informal category group.

Table A3.3.1.1| top | pdf |
Categories in the powder CIF dictionary

Numbers in parentheses refer to the section of this chapter where each category is described in detail.

PD_BLOCK group (§3.3.7[link]) PD_MEAS group (§3.3.4.4[link])
 PD_BLOCK (§3.3.7[link])  PD_DATA (part of category) (§3.3.4.4[link](a))
PD_CALC group (§3.3.5.2[link])  PD_MEAS_INFO (§3.3.4.4[link](b))
 PD_CALC (§3.3.5.2[link](b))  PD_MEAS_METHOD (§3.3.4.4[link](c))
 PD_DATA (part of category) (§3.3.5.2[link](a)) PD_PEAK group (§3.3.5.3[link])
PD_CALIB group (§3.3.4.3[link])  PD_PEAK (§3.3.5.3[link](a))
 PD_CALIB (§3.3.4.3[link](a))  PD_PEAK_METHOD (§3.3.5.3[link](b))
 PD_CALIBRATION (§3.3.4.3[link](b)) PD_PHASE group (§3.3.6.1[link])
PD_CHAR group (§3.3.4.1[link])  PD_PHASE (§3.3.6.1[link])
 PD_CHAR (§3.3.4.1[link](a)) PD_PREP group (§3.3.4.1[link])
PD_DATA group  PD_PREP (§3.3.4.1[link](b))
 _pd_calc_* items (§3.3.5.2[link](a)) PD_PROC group (§3.3.5.1[link])
 _pd_instr_var_illum_len  PD_DATA (part of category) (§3.3.5.1[link](a))
  (§3.3.4.3[link](d))  PD_PROC_INFO (§3.3.5.1[link](b))
 _pd_meas_* items (§3.3.4.4[link](a))  PD_PROC_LS (§3.3.5.1[link](c))
 _pd_proc_* items (§3.3.5.1[link](a)) PD_SPEC group (§3.3.4.2[link])
PD_INSTR group (§3.3.4.3[link])  PD_SPEC (§3.3.4.2[link])
 PD_DATA (part of category) (§3.3.4.3[link](d)) REFLN group (§3.3.5.4[link])
 PD_INSTR (§3.3.4.3[link](c))  REFLN (§3.3.5.4[link])

Acknowledgements

Many people contributed to the creation of this dictionary, most notably Syd Hall, I. David Brown and Brian McMahon. Ian Langford helped considerably with the evolution of the dictionary. This project also benefited from the efforts of many members of the PDF-3 Database Format Subcommittee of the International Centre for Diffraction Data; I am particularly indebted to Walter Schriener for useful concepts and discussions. My initial involvement with powder diffraction file formats was prompted by the encouragement of Richard Harlow; I may someday forgive him.

References

Finger, L. W., Cox, D. E. & Jephcoat, A. P. (1994). A correction for powder diffraction peak asymmetry due to axial divergence. J. Appl. Cryst. 27, 892–900.
HDF (1998). NCSA HDF home page. http://www.hdfgroup.org/ .
International Centre for Diffraction Data (2004). Powder Diffraction File. International Centre for Diffraction Data, 12 Campus Boulevard, Newtown Square, Pennsylvania, USA. http://www.icdd.com .
Le Bail, A., Duroy, H. & Fourquet, J. L. (1988). Ab initio structure determination of LiSbWO6 by X-ray powder diffraction. Mater. Res. Bull. 23, 447–452.
NeXus (1999). NeXus data format. http://www.nexusformat.org .
Pawley, G. S. (1981). Unit-cell refinement from powder diffraction scans. J. Appl. Cryst. 14, 357–361.
Rietveld, H. M. (1967). Line profiles of neutron powder-diffraction peaks for structure refinement. Acta Cryst. 22, 151–152.
Rietveld, H. M. (1969). A profile refinement method for nuclear and magnetic structures. J. Appl. Cryst. 2, 65–71.
Toby, B. H. (2003). CIF applications. XII. Inspecting Rietveld fits from pdCIF: pdCIFplot. J. Appl. Cryst. 36, 1285–1287.
Toby, B. H., Von Dreele, R. B. & Larson, A. C. (2003). CIF applications. XIV. Reporting of Rietveld results using pdCIF: GSAS2CIF. J. Appl. Cryst. 36, 1290–1294.








































to end of page
to top of page