The MATXS format was designed to generalize and simplify the
existing files. The first design principle was that all
information would be identified using lists of Hollerith names.
As an example, if the list of reactions included in the file
contains entries such as NF, NG, and
N2N,
it is trivial to add additional reactions such as KERMA
or DPA. The second design principle was that the file
would be designed to hold sets of vectors and rectangular
matrices and that the same format would be used regardless of the
contents of the vectors and matrices. As a consequence, once a
code is able to handle $n{\leftarrow}n$ matrices, it can also
handle $\gamma{\leftarrow}\gamma$ data; once a code can handle
$\gamma{\leftarrow}n$, it can also handle $n{\leftarrow}\gamma,
\beta {\leftarrow}n$, or even $p{\leftarrow}d$. It was originally
decided to divide the file into ``data types'' identified by its
input and output particles. As an example, $\gamma{\leftarrow}n$
is a data type characterized by input particle equals neutron and
output particle equals photon. The matrices contain cross
sections for producing photons in photon group $n_g$ due to
reactions of neutrons in group $n_n$. The vectors, if any, would
contain photon production cross sections versus neutron group.
Then, each data type was divided into materials and submaterials.
A material might be a particular isotope or mixture (e.g.,
U238 or SS304). Each submaterial could then
represent a particular temperature and background cross section
in a basic library, or perhaps a particular region for a
reactor. Starting with NJOY 91.0, this organization was changed
to have the material as the outermost loop. Each material was
then divided into submaterials, where each submaterial could
belong to a given data type and have a given temperature and
background cross section. This organization makes it much easier
to add, delete, or extract materials from MATXS files. The MATXS
library format that resulted from following these general
principles is described here. The standard
CCCC form is used.
As in all CCCC-type files, a MATXS library starts with a
file identification record that gives the name of the file, a
user identification string, and a version number. This is
followed by the file control record; here NPART
refers to the list of particles found in the file (e.g., N,
G, B); each particle has its own distinctive
group structure. NTYPE is the count of data type names,
and NHOLL is the count of words in a general description
of the contents of the file, processing methods, accuracy
criteria, and so on. NMAT is the number of materials
given in the library. MAXW is the maximum size of a
record to be found in the file. It helps in establishing storage
requirements. Finally, LENGTH gives the total number of
records in the library.
The file data record gives the particle and data type names
(see the two tables below), the names of the
materials, and the number of groups for each particle. It also
characterizes each data type by identifying its input and output
particles. Finally, it gives the number of submaterials for each
material and the LOCM vector. The LOCM vector
makes it easy to skip directly to a desired material without
reading all the intervening data.
Particle Types ------------------------------------- Name Definition ------------------------------------- N Neutron G Photon (gamma or x-ray) P Proton D Deuteron T Triton H 3He A Alpha (4He) -------------------------------------
Data Types -------------------------------------- Name Definition -------------------------------------- NSCAT Neutron scatting GSCAT Photon scattering NG Photon production NTHERM Thermal neutron scattering PSCAT Proton scattering PN Proton-to-neutron matrix NP Neutron-to-proton matrix ... etc. --------------------------------------Each material starts with a material control record, which gives the name of the material and information on the submaterials. Once again, it is possible to jump directly to a desired submaterial by using the
LOCS vector.
The first record for each submaterial is the vector control
record. The vector data are organized by reactions with
Hollerith names (NLGNG,
NA, or
N2N). The symbols NWTO, NWT1,
NTOTO, and NTOT1 refer to the flux- and
current-weighted components of the library weight function and
the neutron total cross section. Fission may be split into the
partial reactions NF, NNF, and N2NF;
in any case, NFTOT is the total fission reaction.
Discrete-inelastic scattering in ENDF is denoted by the level
plus 50; thus, MT51 is $(n,n^{\prime}_1)$. On the MATXS library,
this is given as N01. If the level decays by particle
emission, the particle identifiers are tacked on after the number
(e.g., N62DAA). Emission of $^3$He is denoted by
H.
The neutron files also contain some special quantities.
HEAT and DAME are the local heat production and
damage energy production cross sections. NUBAR,
XI, GAMMA are the continuous slowing-down
parameters. INVEL is the group-averaged reciprocal
velocity. Finally, NUD and CHID
are for delayed fission as described in Sec.~II-B.
In the thermal data type, reaction names identify the binding
condition; FREE for free gas, H2O for light
water, D2O for heavy water, POLY for
polyethylene, BENZ for benzene, GRAPH for
graphite, ZRHYD for
zirconium, and BEO for beryllium oxide. The elastic part
is distinguished from the inelastic part by a \$ sign on the end
of the name.
In the photon interaction data type, the reactions are
GWTO, GTOTO, GHEAT,
and GDAME
analogous to the neutron names, GCOH for coherent
scattering, GINCH for incoherent scattering,
GPAIR
for pair production, and GABS for photoelectric
absorption.
After the vector blocks comes the matrix control record. The
representation used for cross-section data is similar to ISOTXS
in that it packs the data as bands of source groups that scatter
into each given sink (final) group.
The matrix names used in the current libraries are the same as
for the corresponding vectors. All the partial scattering
matrices are kept separate for maximum flexibility. If groups of
these partial matrices were to be combined by an auxiliary
program, new names would have to be defined such as TOTAL
or INELAS.
Summary descriptions of several existing MATXS libraries will be found here. These libraries were derived from ENDF/B-V and ENDF/B-VI; they use neutron group structures with 30 to 187 groups and photon structures with 12 or 24 groups.
TRANSX HyperText Manual
TRANSX HyperText Help Package
T-2 Nuclear Information Service