EXFOR Basics
I. A Quick Guide to EXFOR

(table of contents)

What is EXFOR?

EXFOR is the library and format for the collection, storage, exchange and retrieval of experimental nuclear reaction data. The library is the product of a worldwide co-operation, namely the international Network of Nuclear Reaction Data Centres (NRDC) which is co-ordinated by the IAEA Nuclear Data Section (NDS).

At present (May 2008), the EXFOR database contains about 17,000 works with around 129,000 data tables, representing

• a "complete" compilation of low-energy experimental neutron-induced reaction data,
• a less complete compilation of charged-particle-induced reaction data,
• a selected compilation of photon-induced, heavy-ion-induced, and high energy neutron-induced reaction data.

At present, compilation efforts concentrate on complete coverage of newly published data as well as, depending on the available manpower, on filling gaps in old measurements which are important for certain applications.

Selective retrievals from the database are available in various formats from the web sites of the IAEA Nuclear Data Section and other cooperating NRDC centres. The output formats include the original EXFOR (Exchange) format (described in detail in Chapter II) as well as various other "user" formats which may differ at various data centres (examples of several user output formats are given in Chapter III).

The basic unit of EXFOR is an entry, which corresponds to one experiment which is usually described in one or more bibliographic references (journal articles, laboratory reports, conference proceedings etc.). An entry contains the numerical data and their definition as measured by the authors, along with the related bibliographic information and a brief description of the experimental method. An entry is typically divided in several subentries containing the various data tables resulting from the experiment.

EXFOR is, unlike bibliographic systems, primarily work-oriented, not publication-oriented, and contains many data which have never been published in numerical form, and it is regularly updated (e.g. when authors revise their data after publication).

Principles of EXFOR

• EXFOR is not a bibliographic system but contains numerical nuclear data with cross-references to pertinent publications
• EXFOR contains many data that have never been published in numerical form. It is therefore a publication medium supplementary to conventional, formal publications.
• EXFOR data are currently updated. When authors revise their data after publication, the EXFOR files are kept up-to-date accordingly.
• The numerical data in EXFOR are supplemented by explanatory text giving essential information on meaning and quality of the data including summaries on measurement techniques, corrections and error analysis, standard reference values used, etc.
• An EXFOR "entry" represents the results of a work performed at a given laboratory at a given time; an EXFOR "entry" does not necessarily correspond to the information found in one particular publication. Very often, a "work" is reported in several - formal or informal - publications, typically in one or more progress reports, a conference paper with preliminary results, a lab report, and a final but often less detailed article in a refereed international journal. The EXFOR compiler extracts the essential information from all these sources and, in addition, contacts the author in order to obtain additional information (e.g. details on the error analysis) and/or tabular data for results published only in graphical form, and to verify that the data compiled are the author's final results. This makes EXFOR entirely different from any bibliographic system, which may or may not include all the relevant publications, but will never tell the user whether they describe actually the same work or not. (The only exception to this is CINDA, a bibliography also maintained by the NRDC and closely related to EXFOR, where publications describing the same experiment are listed together in one block).
• EXFOR contains also numerical data which were digitized from results published only in graphical form, and where the original tabulated data could not be obtained from the authors.
• An EXFOR "entry" is identified by an accession number and a date (meaning the date of compilation or of the last revision of the entry). If an entry is revised, nature and reason of the revision are documented within the entry.
• Each EXFOR entry is divided into a number of subentries (data sets) containing the data tables from this particular work. A subentry is identified by a subaccession number, consisting of the accession number and a subentry number.
• EXFOR is designed for flexibility, to meet the diverse needs of the nuclear reaction data centers and to allow the compilation of very diverse type of quantities while making computerized processing of the data possible.
• Compilations are done following as much as possible the author's representations of the quantities and the originally published data units, to avoid mistakes during data input and to facilitate comparison with the original publication. Computerized processing and plotting of the data therefore do not use the basic EXFOR exchange format but one of the available computational formats which are offered as additional output options in the data centres's retrieval systems.
• EXFOR may include also preliminary data (labelled as such) or (with consent of the author) pre-publication data. Preliminary data will be routinely replaced by the final data once they become available, and any new bibliographic references describing the results will be added.
• EXFOR is a compilation of the author's original published experimental data. While the format allows the inclusion of data renormalized to up-to-date standard values (with proper documentation), this task is normally left to data evaluators who systematically review the experimental works.
• EXFOR is not a collection of recommended values for each reaction but will usually contain results from different authors for the same reaction which may or may not be in agreement. The task of recommending best evaluated cross section data is the task of data evaluators. Their work is largely based on the experimental data from EXFOR and their results are collected in various evaluated data files such as ENDF/B, JEFF, JENDL, etc., most of them also available from the IAEA Nuclear Data Section. The retrieval software allows to make comparison plots of evaluated data with experiments retrieved from EXFOR.

Data types included in EXFOR

Quantities

All types of microscopic cross sections and related data, in particular:

• Integral and partial cross sections (including excitation functions, spectrum-averaged data, ratios etc.)
• Differential cross sections of many types, including angular distributions and Legendre coefficients, secondary particle spectra, double-differential cross sections, polarization data, etc.
• Resonance parameters
• Fission product yields, ( Nu-bar), fission quantities
• Product yields and thick target yields
• Reaction rates, resonance integrals

Projectiles

• Regular compilation for projectile energies up to 1 GeV; selected data for higher energies may be included also
• Neutrons (Note that the neutron files include also some data with projectile '0', e.g. fission product yields from spontaneous fission)
• Charged particles (regular compilation up to A=12)
• Heavy ions with A>12 (selected coverage)
• Photons

Basic structure of the EXFOR format

EXFOR (EXchange FORmat for experimental numerical nuclear reaction data) presents in a convenient compact form numerical data as well as physical information necessary to understand the experiment and interpret the data.

Keywords and codes make the information computer intelligible, while English "free text" gives additional information for the human user.

The basic unit of EXFOR is an entry, which corresponds to one experiment which is usually described in one or more bibliographic references (journal articles, laboratory reports, conference proceedings etc.). As the results may consist of several data tables (e.g. cross sections σ(E) for several nuclides), each entry is divided into a number of subentries (data sets). Each entry is assigned an accession number; each subentry is assigned a subaccession number (the accession number plus a subentry number). The accession numbers are associated with a particular work throughout the life of the EXFOR system.

The subentries are further divided into:

• a text part containing bibliographic, descriptive and bookkeeping information (usually called "BIB" information),
• common data that applies to all data throughout the subentry (called "COMMON" section), and
• a data table ("DATA" section).

The first subentry contains the information which is common to the whole entry, i.e. usually the bibliographic reference(s), essentials about the experimental method, and any common data (common parameters for the whole work). The first subentry does not contain any numerical results.

The following subentries contain the actual experimental results (tables), plus any bibliographic and experimental information or common parameters specific to the individual subentry (table). Therefore, any entry consists of at least two subentries: the first subentry with the general information, and one or more following subentries with data tables.

The text part (bibliographic, experimental and bookkeeping information) is specified in variable length fields whose content is defined by keywords. An entry contains only those keywords relevant for the particular work. The information attached to a keyword may consist of "codes" (standard abbreviations taken from a "dictionary") and/or unstructured English "free text".

The table below gives a list of the information identifier keywords. For many of these there is a specific dictionary of permitted standard abbreviations (codes). These dictionaries are open ended in the sense that new codes may be added whenever need arises. The keywords given in bold in the table below will appear in every entry. Some of the keywords describing the experiment (such as FACILITY, METHOD, DETECTOR) will also appear in every entry.

ADD-RES	DETECTOR	INSTITUTE	REFERENCE
ANALYSIS	EN-SEC	LEVEL-PROP	REL-REF
ASSUMED	ERR-ANALYS	METHOD	RESULT
AUTHOR	EXP-YEAR	MISC-COL	SAMPLE
COMMENT	FACILITY	MOM-SEC	STATUS
CORRECTION	FLAG	MONIT-REF	TITLE
COVARIANCE	HALF-LIFE	MONITOR
CRITIQUE	HISTORY	PART-DET
DECAY-DATA	INC-SOURCE	RAD-DET
DECAY-MON	INC-SPECT	REACTION

Data definition

The data of each table in a subentry are defined in the keyword REACTION which defines both the nuclear reaction as such (e.g. neutron-induced fission on ²³⁵U, or the microscopic production cross section of a certain radionuclide by bombarding a lead target with protons) and the quantity measured (e.g. integral or differential cross section, fission product yield, resonance parameters, etc.) The REACTION code therefore consists of up to 9 subfields, subfield 1 through 4 for defining the nuclear reaction, and subfields 5 through 9 for a detailed definition of the quantity measured and its representation.

Subfields 1 through 4 are usually self-explanatory:

1-H-1(N,G)1-H-2	=1H(n,γ)2H
92-U-235(N,F)	=235U(n,f)
26-FE-56(N,INL)26-FE-56	=56Fe(n,n')
28-NI-0(P,X)11-NA-24	=natNi(p,x)24Na production

Subfields 5 through 8 define the quantity measured. They are separated by commas; only subfield 6 is always present. Simple examples for complete REACTION codes are:

REACTION (92-U-235(N,F),,SIG)	= fission cross section σn,f for 235U
REACTION (28-NI-60(N,P)29-CU-60,,DA)	= dσ/dΩ for the reaction 60Ni(n,p)60Cu

The numerical part of a subentry consists of the data table itself (DATA section) and, most often, of one or more constant parameters given in the COMMON section. In both cases, it is structured in columns with a constant field length of 11 characters. All numerical columns are headed and defined by

• column headings, for example
• EN for incident particle energy
• DATA for the actual data defined under the keyword REACTION
• DATA-ERR for the uncertainty of the data, etc.
• data units, such as
• EV for electron-Volts
• MB for millibarns, etc.

EXFOR examples

The following pages show examples of two EXFOR entries. The examples are given in two formats, the basic EXFOR Exchange format, and the Interpreted "Exfor+" format, where many of the abbreviations used for bibliographic references, quantities, detectors,etc., are expanded to full English text. These explanatory lines added by the "Interepreted EXFOR" output program are labelled with a # sign at the beginning.

(For examples of other user output formats see Chapter III.)

Both example entries contain constant parameters in COMMON sections. Note that a COMMON parameter given in subentry 2 is valid only for this particular subentry, while a COMMON parameter given in subentry 1 is valid for all of the following subentries.

In any retrieval, always subentry 1, which contains the information common to the whole entry, is output together with all those other subentries which satisfy the retrieval criteria. One of the examples shown here (accession number 31439) contains subentries 1, 5, and 6, which is sufficient to demonstrate a complete set of information output for a typical retrieval.

Subentry 31439.006 has a more complex structure in two respects:

• The table has data for 3 reactions (quantities). (This is possible if the data are integrally related to each other and depend on the same independent variables; here it is the cross sections for production of the ground state and for the metastable state of the same product, and the related isomeric ratio). In such cases, the 3 REACTION codes are linked with the appropriate DATA columns by means of "pointers", in this case the flags "1", "2" and "3". Such pointers may appear in EXFOR also in other places to link related peaces of information.
• The data table has 8 columns (the data for all 3 reactions and the monitor reaction, each with its related uncertainty). Since the Exchange format allows only 6 physical columns per line, the "logical lines" are broken - in this case into a first line of 6 columns and a second line with the remaining 2 columns. This is inconvenient for the human reader, and the Interpreted EXFOR+ format outputs all 8 columns side by side.

• Example 1 (Subentry C1582.002): Exchange format / EXFOR+ format
• Example 2 (Subentry 31439.006): Exchange format / EXFOR+ format

(EXFOR+ format: Lines starting with # are explanatory lines added by the "Interpreted EXFOR" output program)

History of EXFOR

Systematic collection of experimental neutron nuclear data started in the 1960s at four data centres, each using its own data storage and retrieval system:

• Brookhaven National Laboratory (BNL), USA (formerly Sigma Center, now NNDC - National Nuclear Data Center), using "SCISRS";
• OECD Nuclear Energy Agency, France (formerly Neutron Nuclear Data Centre, Saclay, now NEA Data Bank, Gif-sur-Yvette), using "NEUDADA";
• International Atomic Energy Agency (IAEA), Vienna, Austria (formerly Nuclear Data Unit, now Nuclear Data Section), using "DASTAR";
• Fiziko-Energeticheskij Institut (IPPE) Obninsk, Russia, (Center Jadernykh Dannykh), using a USSR computer system incompatible to Western computers.

It became obvious that these activities required coordination, and through discussions held between software experts and physicists from Saclay, Vienna, Livermore and Brookhaven, a joint nuclear data exchange format "EXFOR" was formulated in its initial form at a panel meeting in Brookhaven in February 1969. It was accepted at an IAEA Consultant's Meeting in Moscow in November 1969, and in 1970 the system went into operation, including the Obninsk Center, which solved the compatibility problem to USSR computers. Thus an East-West information exchange on magnetic tapes was initiated for the first time. Data compiled at any one of the centres were speedily transmitted to the other centres, making them available to the fast increasing community of data users throughout the world.

As the name suggests, EXFOR was designed to be the format of data exchange between centres. The centres were free to use different formats for internal storage and/or for output retrievals made available to users.

Subsequently, data compiled earlier were converted to EXFOR, and in the 1970s, the scope was widened to include also charged-particle induced nuclear data and photonuclear data. For this purpose, the original format was modified, and additional nuclear data centres joined the network. A list of the cooperating data centres is given on the following page.

The data retrieval services to users were, for many years, done centrally by data centre staff who received specific data retrieval requests by mail and in turn mailed output listings on paper or magnetic tapes back to the users. From the late 1980s, online retrieval systems, using Telnet or other remote access systems, started to be used for nuclear data in parallel to the central retrieval services, including EXFOR (1988 - NNDC, 1992 - IAEA-NDS). In the late 1990s, web-based retrieval systems were introduced (starting 1997 at IAEA-NDS) and are continuously being refined.

While the network of nuclear reaction data centres has been coordinated by the IAEA Nuclear Data Section since its start, each participating centre independently maintained and updated its EXFOR master file. This sometimes led to occasional differences in the database contents due to differences in the synchronisation of updates or other reasons. Therefore, a common global EXFOR master file has been introduced in 2005 which is maintained centrally by IAEA-NDS and is the basis for the worldwide EXFOR operations and services.

The Network of Nuclear Reaction Data Centres

National, regional and specialized nuclear reaction data centres, coordinated by the International Atomic Energy Agency, cooperate in the compilation, exchange and dissemination of nuclear reaction data, in order to meet the requirements of nuclear data users in all countries. At present, the following data centres participate in the network:

NNDC	US National Nuclear Data Center, Brookhaven, USA
NEADB	OECD/NEA Nuclear Data Bank, Issy-les-Moulineaux, France
NDS	IAEA Nuclear Data Section
CJD	Russian Nuclear Data Center, Obninsk, Russia
CNDC	China Nuclear Data Center, China Institute of Atomic Energy, Beijing, China
ATOMKI	Charged-Particle Nuclear Reaction Data Group, Institute for Nuclear Research (ATOMKI), Debrecen, Hungary
NDPCI	Nuclear Data Physics Centre of India, Mumbai, India
JAEA	Nuclear Data Center, Japan Atomic Energy Agency, Tokai-Mura, Japan
JCPRG	Nuclear Reaction Data Centre, Hokkaido University, Sapporo, Japan
KNDC	Korea Nuclear Data Center, Korea Atomic Energy Research Institute, Yuseong, Daejeon, Republic of Korea
CDFE	Centre for Photonuclear Experiments Data, Moscow State University, Moscow, Russia
CNPD	Center of Nuclear Physics Data, All Russian Scientific Research Institute of Experimental Physics, Sarov, Russia
UkrNDC	Ukrainian Nuclear Data Center, Institute for Nuclear Research, Kyiv, Ukraine

A detailed description of the objectives of the network and the contributions of each Centre to these activities are given in INDC(NDS)-0401 (Rev.5), "International Network of Nuclear Reaction Data Centres".

A summary of the network's objectives, and hyperlinks to all participating centres can be found on the NRDC web page, see http://www-nds.iaea.org/nrdc/.

EXFOR Retrieval Websites

NNDC (Brookhaven, USA):	http://www.nndc.bnl.gov/exfor/
NEADB (Issy-les-Moulineaux, France):	http://www.oecd-nea.org/janisweb/search/exfor/
NDS (Vienna, Austria, maintains global master file):	http://www-nds.iaea.org/exfor/
JCPRG (Sapporo, Japan):	https://www.jcprg.org/exfor/
JAEA (Tokai, Japan):	http://spes.jaea.go.jp/
CDFE (Moscow, Russia):	http://cdfe.sinp.msu.ru/exfor/

Citation Guidelines

When citing data extracted from EXFOR, always both, the original reference, and the EXFOR dataset with its retrieval source should be cited.

Example:

A.B. Author, et al., J. Nucl. Phys. 12, 345 (1979). Data taken from the EXFOR database, file EXFOR 12345.002 dated April 5, 1980, retrieved from the IAEA Nuclear Data Services website.

To reference the EXFOR database in general, the present report (IAEA-NDS-206, June 2008) may be used.