intro.tex

\section{INTRODUCTION}
\label{sINTRO}

The NJOY nuclear data processing system\cite{NJ1,NJ2,NJOY91,NJ10,
NJ12,njoytutorial}
is a comprehensive computer code package for producing pointwise and
multigroup nuclear cross sections and related quantities from
evaluated nuclear data in the ENDF format.  The U.S. Evaluated
Nuclear Data Files (ENDF) have progressed through a number of
versions, notably ENDF/B-III, ENDF/B-IV, ENDF/B-V, ENDF/B-VI,
and ENDF/B-VII\cite{ENDF7,ENDF71}.  The ENDF format has also evolved
through many versions.  Variations of the format called
``ENDF-6'' were used for ENDF/B-VI and ENDF/B-VII, and will be\
used in ENDF/B-VIII.  The latest version of the format is
described in the ENDF-102 document\cite{ENDF102,endftutorial}.
\index{ENDF}
\index{ENDF!ENDF/B-III}
\index{ENDF!ENDF/B-IV}
\index{ENDF!ENDF/B-V}
\index{ENDF!ENDF/B-VI}
\index{ENDF!ENDF/B-VII}
The ENDF format is also used in other nuclear data
libraries such as the JEFF libraries in Europe and the JENDL libraries
in Japan, or in specialized libraries distributed through the Nuclear
Data Section of the International Atomic Energy Agency (IAEA).
\index{JEFF}
\index{JENDL}
These libraries represent the underlying nuclear data from a physics
viewpoint, but practical calculations usually require special
libraries for particle transport codes or reactor core physics codes.
This is the mission of NJOY --- to take the basic data from the nuclear
data library and convert it into the forms needed for applications.

\subsection{The Modules of NJOY}
\label{ssINTRO_Modules_of_NJOY}

The NJOY code consists of a set of main modules, each performing a
well-defined processing task.  Each of these main modules is
essentially a separate computer program.  They are linked to one
another by input and output files.  The main modules are supported
by a number of subsidiary modules providing things like physics
constants, utility routines, and mathematics subroutines that can
be ``used'' by the main modules.  The NJOY modules are grouped as
follows:
\begin{singlespace}
\begin{description}
\item[\hyperlink{sNJOYhy}{NJOY}] directs the flow of data through
the other modules.
  Subsidiary modules for locale, ENDF formats, physics constants,
  utility routines, and math routines are grouped with the NJOY
  module for descriptive purposes.
\index{NJOY}
\index{data flow}

\item[\hyperlink{sRECONRhy}{RECONR}] reconstructs pointwise (energy-dependent)
  cross sections from ENDF resonance parameters and interpolation schemes.
\index{RECONR}
\index{resonance reconstruction}

\item[\hyperlink{sBROADRhy}{BROADR}] Doppler-broadens and thins pointwise
 cross sections.
\index{BROADR}
\index{Doppler-broadening}

\item[\hyperlink{sUNRESRhy}{UNRESR}] computes effective self-shielded pointwise
  cross sections in the unresolved energy range.
\index{UNRESR}
\index{self-shielding}
\index{unresolved resonance range}

\item[\hyperlink{sHEATRhy}{HEATR}] generates pointwise heat production cross
 sections (neutron KERMA factors) and radiation damage production cross
 sections.
\index{HEATR}
\index{heating}
\index{damage}

\item[\hyperlink{sTHERMRhy}{THERMR}] produces cross sections and
 energy-to-energy matrices for free or bound scatterers in the thermal
 energy range.
\index{THERMR}

\item[\hyperlink{sGROUPRhy}{GROUPR}] generates self-shielded multigroup cross
  sections, group-to-group scattering matrices, photon production matrices,
  and charged-particle multigroup cross sections from pointwise input.
\index{GROUPR}

\item[\hyperlink{sGAMINRhy}{GAMINR}] calculates multigroup photoatomic cross
 sections, photon KERMA factors, and group-to-group photon scattering matrices.
\index{GAMINR}

\item[\hyperlink{sERRORRhy}{ERRORR}] computes multigroup covariance matrices
  from pointwise covariance data.
\index{ERRORR}
\index{covariances}

\item[\hyperlink{sCOVRhy}{COVR}] reads the output of ERRORR and performs
 covariance plotting and output formatting operations.
\index{COVR}

\item[\hyperlink{sMODERhy}{MODER}] converts ENDF ``tapes'' back and forth
 between formatted (that is, ASCII) and blocked binary modes.
\index{MODER}

\item[\hyperlink{sDTFRhy}{DTFR}] formats multigroup data for transport codes
 that use formats based on the DTF-IV code.
\index{DTFR}
\index{DTF-IV}

\item[\hyperlink{sCCCCRhy}{CCCCR}] formats multigroup data for the CCCC standard
  files ISOTXS, BRKOXS, and DLAYXS.
\index{CCCCR}
\index{ISOTXS}
\index{BRKOXS}
\index{DLAYXS}

\item[\hyperlink{sMATXSRhy}{MATXSR}] formats multigroup data for the newer
 MATXS cross-section interface file, which works with the TRANSX code
 to make libraries for many particle transport codes.
\index{MATXSR}
\index{MATXS format}
\index{TRANSX}

\item[\hyperlink{sRESXSRhy}{RESXSR}] prepares pointwise cross sections in a
 CCCC-like format for thermal flux calculators.
\index{RESXSR}

\item[\hyperlink{sACERhy}{ACER}] prepares libraries in ACE format for the
  Los Alamos continuous-energy Monte Carlo MCNP and MCNPX codes.  The ACER
  module is supported by subsidiary modules for the different classes
  of the ACE format.
\index{ACER}
\index{MCNP}
\index{Monte Carlo}

\item[\hyperlink{sPOWRhy}{POWR}] prepares libraries for the EPRI-CELL
  and EPRI-CPM codes.
\index{POWR}

\item[\hyperlink{sWIMSRhy}{WIMSR}] prepares libraries for the thermal reactor
  assembly codes WIMS-D and WIMS\nobreakdash-E.
\index{WIMSR}
\index{WIMS!WIMS-D}
\index{WIMS!WIMS-E}

\item[\hyperlink{sPLOTRhy}{PLOTR}] makes plots of cross sections and
  perspective plots of distributions for both pointwise and multigroup
  data by generating input for the VIEWR module.
\index{PLOTR}

\item[\hyperlink{sVIEWRhy}{VIEWR}] converts plotting files produced by the
  other modules into high-quality color Postscript plots.
\index{VIEWR}

\item[\hyperlink{sMIXRhy}{MIXR}] is used to combine cross sections into
  elements or mixtures, mainly for plotting.
\index{MIXR}

\item[\hyperlink{sPURRhy}{PURR}] is used to prepare unresolved-region
  probability tables for the MCNP continuous-energy Monte Carlo code.
\index{PURR}
\index{unresolved resonance range}
\index{probability tables}

\item[\hyperlink{sLEAPRhy}{LEAPR}] produces thermal scattering data in ENDF-6
  File 7 format that can be processed using the THERMR module.
\index{LEAPR}
\index{thermal scattering}

\item[\hyperlink{sGASPRhy}{GASPR}] generates gas-production cross sections
  in the pointwise PENDF format from basic ENDF cross sections.
\index{GASPR}
\index{gas production}
\end{description}
\end{singlespace}

\noindent
The methods used in these modules and instructions on how to use them
are given in subsequent sections of this report.  The sections on the
modules are followed by an additional section on NJOY maintenance
and testing.

\subsection{Data Flow in NJOY}
\label{ssINTRO_dataflow}

The modules of NJOY can be linked in a number of different ways to
prepare libraries for various nuclear applications.  The following
brief summary illustrates the general flow of data in the code.
\index{data flow}

\hyperlink{sRECONRhy}{RECONR}\index{RECONR}
reads an ENDF\index{ENDF} file
and produces a common energy grid for all reactions (the union grid)
\index{union grid} such that all cross sections can be obtained to
within a specified tolerance by linear interpolation.  Resonance
cross sections are reconstructed\index{resonance reconstruction}
using a method of choosing the energy grid that incorporates
control over the number of significant figures\index{significant figures}
generated and a resonance integral criterion to reduce the number
of grid points generated for some materials.  Summation cross sections
\index{summation cross sections} (for example, total, inelastic) are
reconstructed from their parts.  The resulting pointwise cross sections
\index{pointwise cross sections} are written onto a ``point-ENDF''
(PENDF) file for future use.  \hyperlink{sBROADRhy}{BROADR}
\index{BROADR} reads a
PENDF\index{PENDF} file and Doppler-broadens\index{Doppler-broadening}
the data using the accurate point-kernel method.\index{kernel broadening}
The union grid allows all resonance reactions to be broadened
simultaneously, resulting in a saving of processing time.  After
broadening and thinning, the summation cross sections are again
reconstructed from their parts.  The results are written out on a
new PENDF file for future use.  \hyperlink{sUNRESRhy}{UNRESR}
\index{UNRESR} produces effective
self-shielded pointwise cross sections,\index{self-shielding}
versus energy and background cross section, in the unresolved
\index{unresolved resonance range} range.  This is done
for each temperature produced by \hyperlink{sBROADRhy}{BROADR},
using the average resonance parameters from the ENDF
evaluation.  The results are added to the PENDF file using a
special format.

If desired, additional special kinds of data can be added to the
PENDF file.  \hyperlink{sHEATRhy}{HEATR}\index{HEATR}
computes energy-balance heating,
\index{heating} KERMA,\index{KERMA} and damage energy\index{damage}
using reaction kinematics or by applying conservation of energy.  The
ENDF photon production files can be used in this step, when available.
Comparisons of momentum and energy calculations with the photon files
can be used to find energy-balance errors in the evaluations.
\index{energy-balance checks} For ENDF-6 formatted data,
charged-particle distributions in File 6\index{File 6} are used
directly to compute accurate heating and damage parameters.  The
energy-balance heating\index{energy-balance heating} results are
added to the PENDF file using ENDF reaction numbers in the 300 series;
kinematic KERMA\index{KERMA!kinematic KERMA} uses the special identifier
443,\index{MT!MT443} and damage results use the special identifier
444.\index{MT!MT444}  \hyperlink{sTHERMRhy}{THERMR}
\index{THERMR} produces pointwise cross
sections in the thermal range.  Energy-to-energy incoherent inelastic
scattering\index{incoherent inelastic} matrices can be computed for
free-gas scattering\index{free gas} or for bound scattering using
a precomputed scattering law S($\alpha$,$\beta$)\index{$S(\alpha,\beta)$}
in ENDF format.  The secondary angle and energy grids are
determined adaptively so as to represent the function to a
desired precision by linear interpolation;  the angular
representation is then converted to one based on equally probable
angles.  See the chapter on \hyperlink{sTHERMRhy}{THERMR}
for another possible representation.
Coherent-elastic scattering\index{coherent elastic} from crystalline
materials can be computed using internal lattice information, or
for ENDF-6 formatted files, using data from the evaluation directly.
The scheme used provides a detailed representation of the delta
functions in energy and angle.  Incoherent-elastic scattering
\index{incoherent elastic} for hydrogenous materials is represented
using equally probable angles computed analytically or using
ENDF-6 parameters.  The results for all the processes are added to
the PENDF file using special formats and special reaction numbers
between 221 and 250.\index{MT!MT221 -- 250}  Additional reactions describing
the production of the gases \index{gas production} $^1$H, $^2$H,
$^3$H, $^3$He, and $^4$He can be added to the PENDF file with
MT=203 -- 207\index{MT!MT203 -- 207} using the
\hyperlink{sGASPRhy}{GASPR}\index{GASPR} module.

\hyperlink{sGROUPRhy}{GROUPR}
\index{GROUPR} processes the pointwise cross sections
\index{pointwise cross sections} produced by the modules described
above into multigroup form\index{multigroup cross sections}.  The
weighting function \index{weight functions} for group averaging
\index{group structures} can be taken to be the Bondarenko
\index{Bondarenko method} form, or it can be computed from the
slowing-down equation for a heavy absorber in a light moderator.
\index{flux calculator}  Self-shielded cross sections,
scattering matrices, photon production matrices,
charged-particle matrices, and photonuclear matrices are all
averaged in a unified way, the only difference being in the
function that describes the ``feed'' into a secondary group $g'$
with Legendre order $\ell$ from initial energy $E$.  The feed
function\index{feed function} for two-body scattering is computed
using a center-of-mass (CM) Gaussian integration scheme, which
provides high accuracy even for small Legendre components of
the scattering matrix.  Special features are included for delayed
neutrons,\index{delayed-neutrons} coupled energy-angle distributions
\index{coupled energy-angle distributions} (either from
\hyperlink{sTHERMRhy}{THERMR}\index{THERMR}
or from ENDF-6 evaluations using File 6),
discrete scattering angles arising from thermal coherent reactions,
and charged-particle elastic scattering.  Prompt fission is treated
with a full group-to-group matrix\index{group-to-group fission}.
The results are written in a special ``groupwise-ENDF'' format
(GENDF)\index{GENDF} for use by the output formatting modules.
\hyperlink{sGAMINRhy}{GAMINR}\index{GAMINR} uses
a specialized version of \hyperlink{sGROUPRhy}{GROUPR} to
compute photoatomic\index{photoatomic} cross sections and
group-to-group matrices.  Coherent and incoherent atomic form
factors\index{atomic form factors} are processed in order to extend
the useful range of the results to lower energies.  Photon heat
production cross sections are also generated.  The results are
saved on a GENDF file.

The covariance module, \hyperlink{sERRORRhy}{ERRORR},
\index{ERRORR} can either produce its
own multigroup cross sections using the methods of
\hyperlink{sGROUPRhy}{GROUPR} or start
from a precomputed set.  The cross sections and ENDF covariance
data are combined in a way that includes the effects of deriving
one cross section from several others.  Special features are
included to process covariances for data given as resonance
parameters or ratios (for example, fission $\bar{\nu}$).  It
is also possible to process covariances for the P$_1$
component of an angular distribution, a secondary-energy
distribution, and radionuclide production.  The
\hyperlink{sCOVRhy}{COVR}\index{COVR}
module uses the output from \hyperlink{sERRORRhy}{ERRORR}
together with the \hyperlink{sVIEWRhy}{VIEWR}
\index{VIEWR} module to make publication-quality plots of
covariance data; it also provides output in the efficient BOXER
\index{BOXER format} format, and it provides a site for user-supplied
routines to prepare covariance libraries\index{covariance libraries}
for various sensitivity systems.\index{sensitivity systems}

\hyperlink{sMODERhy}{MODER}\index{MODER} is
often used at the beginning of an NJOY job
to convert ENDF library files into binary mode for calculational
efficiency, or at the end of a job to obtain a printable version
of a result from ENDF, PENDF, GENDF or ERRORR input.  It can also
be used to extract desired materials from a multimaterial library, or to
combine several materials into new ENDF, PENDF, GENDF or
ERRORR files.
\hyperlink{sDTFRhy}{DTFR} \index{DTFR} is
a simple reformatting code that produces
cross-section tables acceptable to many discrete-ordinates transport
codes.\index{discrete-ordinates transport}  It also converts the
GROUPR fission matrix to $\chi$ and $\bar{\nu}\sigma_f$ and prepares
a photon-production matrix, if desired.  The user can define
edit cross sections that are any linear combination of the
cross sections on the GENDF file.  This makes complex edits
such as gas production possible.  \hyperlink{sDTFRhy}{DTFR}
also produces plotting files for \hyperlink{sVIEWRhy}{VIEWR}
to use in making routine plots for the
cross sections, P$_0$ scattering matrix, and photon production
matrix.  This module has become somewhat obsolete with the advent
of the MATXS/TRANSX system.\index{MATXS/TRANSX}

A number of other interface file formats are available from
NJOY.  The \hyperlink{sCCCCRhy}{CCCCR}\index{CCCCR}
module is a straightforward
reformatting code that supports all the options of the
CCCC-IV\cite{CCCC4}\index{CCCC format!CCCC-IV} file specification.  In the
cross-section file (ISOTXS)\index{ISOTXS}, the user can choose
either isotope $\chi$ matrices or isotope $\chi$ vectors collapsed
using any specified flux.  The BRKOXS\index{BRKOXS} file includes
the normal self-shielding factors plus self-shielding factors for
elastic removal.  The DLAYXS\index{DLAYXS} provides delayed-neutron
data for reactor kinetics codes.  Note that some of the
cross sections producible with NJOY are not defined in the CCCC
files.  For that reason, we have introduced the new CCCC-type
material cross section file MATXS\index{MATXS}.  The
\hyperlink{sMATXSRhy}{MATXSR}
\index{MATXSR} module reformats GENDF data for neutrons, photons,
and charged particles into the MATXS format, which is suitable
for input to the TRANSX (transport cross section)
code\cite{TRANSX}.\index{TRANSX}  TRANSX can produce libraries
for a variety of particle transport codes that have been
used over the years, such as ANISN\cite{ANISN}\index{ANISN},
ONEDANT\cite{ONEDANT}\index{ONEDANT},
TWODANT\cite{TWODANT}\index{TWODANT}, and
DIF3D\cite{DIF3D}.\index{DIF3D}  TRANSX has special features
to support the latest S$_{\rm N}$ transport code from Los Alamos,
PARTISN\cite{PARTISN}.\index{PARTISN}  The MATXS format uses
efficient packing techniques and flexible naming conventions
that allow it to store all NJOY data types.  A companion module,
\hyperlink{sRESXSRhy}{RESXSR}, formats \index{RESXSR} pointwise data into
a CCCC-like format for use in thermal flux calculators.

Pointwise data can also be fed directly into the
\hyperlink{sACERhy}{ACER}\index{ACER}
module.  This module prepares cross sections and scattering laws
in ACE format\index{ACE format} (A Compact ENDF) for the MCNP
code\cite{MCNP}.\index{MCNP}  All the cross sections are
represented on a union grid\index{union grid} for linear interpolation
by taking advantage of the representation used in
\hyperlink{sRECONRhy}{RECONR} and
\hyperlink{sBROADRhy}{BROADR}.  ``Laws'' for
describing scattering and photon production
are very detailed, providing a faithful representation of the
ENDF-format evaluation with few approximations.  The data are
organized for random access for purposes of efficiency.   MCNP
handles self-shielding in the unresolved-energy range using
probability tables.  The \hyperlink{sPURRhy}{PURR}
\index{PURR} module of NJOY can
be used to prepare these tables and add them to a PENDF file
for reading by the \hyperlink{sACERhy}{ACER} module.

Another alternate path for multigroup data is to use the
\hyperlink{sPOWRhy}{POWR}\index{POWR} module
to produce libraries for the power reactor
codes EPRI-CELL\index{EPRI!EPRI-CELL} or EPRI-CPM\index{EPRI!EPRI-CPM}
\footnote{EPRI-CELL and EPRI-CPM are proprietary products
of the Electric Power Research Institute (EPRI), 3420
Hillview Avenue, Palo Alto, CA 94304.}.  Similarly, the
\hyperlink{sWIMSRhy}{WIMSR}\index{WIMSR} module
can be used to prepare libraries
for the thermal reactor assembly codes WIMS-D\index{WIMS!WIMS-D}
and WIMS-E\cite{WIMS}.\index{WIMS!WIMS-E}

At the end of any NJOY run, the \hyperlink{sPLOTRhy}{PLOTR}
\index{PLOTR} and \hyperlink{sVIEWRhy}{VIEWR}
\index{VIEWR} modules can be used to view the results
or the original ENDF data.  \hyperlink{sPLOTRhy}{PLOTR}
can prepare 2-D plots with
the normal combinations of linear and log axes, including
legend blocks or curve tags, titles, and so on.  Several curves
can be compared on one plot (for example, pointwise data can
be compared with multigroup results), and experimental data
points with error bars can be superimposed, if desired.
Plots can be prepared showing the ratio or percent difference
of two cross sections.  \hyperlink{sPLOTRhy}{PLOTR}
can also produce 3-D perspective
plots of ENDF and GENDF angle or energy distributions
or thermal S($\alpha,\beta$) tables.   The output of
\hyperlink{sPLOTRhy}{PLOTR} is passed to
\hyperlink{sVIEWRhy}{VIEWR}, which renders the plotting commands into
high-quality color Postscript graphics for printing or for viewing.
The \hyperlink{sHEATRhy}{HEATR}, \hyperlink{sCOVRhy}{COVR},
\hyperlink{sDTFRhy}{DTFR}, and \hyperlink{sACERhy}{ACER}
modules also produce plotting files in \hyperlink{sVIEWRhy}{VIEWR}
format that are useful for quality reviews of data-processing
results.  The \hyperlink{sMIXRhy}{MIXR}\index{MIXR} module can be used
to combine isotopes into elements for plotting and other purposes.
It only works for simple cross sections at the present time.

\subsection{Computer Implementation}
\label{ssINTRO_implementation}

NJOY2016 is written in a modern subset of Fortran-90 and
later, stressing the
use of modules.  The modules
help to enforce information hiding.  Thus, for example, the
\hyperlink{sRECONRhy}{RECONR} module only
makes the single subroutine \cword{reconr} public; all the internal
routines and data structures of the module are protected.  The
modules also help to promote logical structuring.  For example,
all of the routines and data structures for working with the ENDF
formats, such as subroutine \cword{contio} and the \cword{math,mfh,mth}
values are made public by the \cword{endf} module, and they can be
easily accessed wherever they are need by the \cword{use endf}
statement.  As another example, what could be a very
large \hyperlink{sACERhy}{ACER}
module is made more structured by providing subsidiary modules for
handling each of the different ``classes'' of the ACE format ({\it e.g.}
\cword{acefc} for continuous class ``c'' data, or \cword{acepn}
for ``u'' class photonuclear data).

The advanced capabilities that modern Fortran provides for typing
variables (the ``kind'' property) have enabled us to remove
all of the complex short-word/long-word controls from NJOY.
Almost all of the internal data in NJOY are handled using
8-byte kinds for real and integer words.  These properties
are set up in the \cword{locale} module, and they can be changed,
if necessary, without touching the balance of the code.  An
exception is made for the CCCC modules
\hyperlink{sCCCCRhy}{CCCCR}, \hyperlink{sMATXSRhy}{MATXSR},
and \hyperlink{sRESXSRhy}{RESXSR},
where 4-byte variables and equivalencing are still used to
construct records with mixed real, integer, and Hollerith values.

Modern Fortran also provides a capability for the dynamic allocation
\index{dynamic storage allocation} of memory for data structures.
In the past, NJOY used its own
\cword{storag} system for this purpose;
this system has now been abandoned for the readability, consistency,
and exportability provided by the new Fortran standard.  We have
limited ourselves to the use of ``allocatable arrays''.  This
relieves us of some storage limitations in NJOY, and introduces others.

The earliest versions of NJOY used their own free-form input routine,
\index{free-form input} \cword{FREE}, which was developed long
before Fortran supported equivalent capabilities.  More recently we
have abandoned \cword{FREE} in favor of the standard Fortran READ*
method.  In doing this, we have lost some capabilities (such as the
repeat field), but we have gained in transportability.  Test
fields that were previously delimited with the star character now
must be delimited with the single-quote character.

For consistency and convenience, the NJOY modules in previous
versions of the code made use of a set of common functions and
subroutines located in the \hyperlink{sNJOYhy}{NJOY} module.  Beginning
with NJOY2012, these utility routines have been repackaged into
Fortran-90 modules.  They include \cword{locale} for
localization variables,
\cword{physics} for physics constants, \cword{mainio} for
input, output, and scratch units, \cword{util} for utility
routines like time and date, \cword{endf} for ENDF processing
routines and variables, and \cword{math} for mathematics
routines.  Each of these modules only makes public the
minimum set of routines and variables needed by the other
modules that ``use'' them.  These modules are described in detail
in the \hyperlink{sNJOYhy}{NJOY} chapter of this report.

NJOY is heavily commented.\index{commenting}  Each module starts
with a long block of comment cards that gives a brief description
of the module and then gives the current user input instructions.
\index{input instructions} Users should always check the input
instructions in the current version of the NJOY source code rather
than the instructions summarized in this manual --- changes may have
been made.  For the convenience of users, the input instructions
are also available on the web\cite{webinclude}.  Furthermore, each function
or subroutine starts with a block of comment cards that describes
its function and special requirements.  Additional lines of
comment cards are used inside each procedure to block off its
major components.

Typography conventions for Fortran differ from place to place.
On most machines at Los Alamos National Laboratory (LANL), Fortran
text is given in lowercase.  In
order to avoid translation problems, previous
versions of NJOY avoided using mixed-case text
for comments or for labels on graphs.  We are gradually moving away
from this limiting convention.  In this report, Fortran text and
variable names are printed using a lowercase \cword{typewriter font}.
\index{typewriter font}  ENDF-related formats and variables
are given using an uppercase \cword{TYPEWRITER FONT}.   CCCC formats were
traditionally given using uppercase characters,
but the usage in this report is mixed.

\subsection{History and Acknowledgments}
\label{ssINTRO_history}

NJOY was started as a successor to MINX\cite{MINX}\index{MINX}
(A Multigroup Interpretation of Nuclear X-sections) late
in 1973 (it was called MINX-II then).  The current name was
chosen in late 1974 to be evocative of ``MINX plus'' and to
eliminate the reference to ``multigroup.''  The first
goals were to add a photon production capability like that in
LAPHAN0\cite{LAPHAN0}\index{LAPHAN0}, to add a photon interaction
capability like GAMLEG\cite{GAMLEG}\index{GAMLEG}, to provide an
easy link to the Los Alamos 30-group libraries of the day using
DTF\cite{DTF}\index{DTF} format, and to merge in the capabilities
of ETOPL\cite{ETOPL}\index{ETOPL} to produce libraries for the MCN
Monte Carlo code (the ancestor of MCNP).\index{MCNP}  Most of the work was
done by MacFarlane\index{MacFarlane}; Rosemary Boicourt\index{Boicourt}
joined the project in 1975.  First, the RESEND\cite{RESEND}\index{RESEND}
and SIGMA1\cite{SIGMA1}\index{SIGMA1} modules of MINX were converted
to use union grids\index{union grid}, and a new method of resonance
reconstruction\index{resonance reconstruction} was developed.  These
steps led to \hyperlink{sRECONRhy}{RECONR}\index{RECONR}
and \hyperlink{sBROADRhy}{BROADR}\index{BROADR}.
\hyperlink{sUNRESRhy}{UNRESR}\index{UNRESR}, which was based on methods
from ETOX\cite{ETOX}\index{ETOX}, was moved over from MINX with
only a few changes.  Next, a completely new multigroup averaging
program, \hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR},
was developed around the unifying
concept of the ``feed function,'' which handled neutron- and
photon-production cross sections in a parallel manner.  The
CM Gaussian integration for discrete two-body scattering was
developed.  \hyperlink{sDTFRhy}{DTFR}\index{DTFR}
was developed as the first NJOY output module.  The
first versions of the \hyperlink{sNJOYhy}{NJOY} utility codes were
introduced; the new concepts of ``structured programming''
\index{structured programming} inspired some of the features
of the new NJOY code.

Major influences during this period included Don Harris\index{Harris},
Raphe LaBauve\index{LaBauve}, Bob Seamon\index{Seamon}, and Pat
Soran\index{Soran} at Los Alamos, and Chuck Weisbin\index{Weisbin}
at the Oak Ridge National Laboratory\index{Oak Ridge National Laboratory!ORNL}
(ORNL).  Odelli Ozer\index{Ozer} at the Brookhaven National Laboratory
\index{Brookhaven National Laboratory!BNL} (and later EPRI\index{EPRI})
helped with RESEND, and Red Cullen at Lawrence Livermore National
Laboratory\index{Lawrence Livermore National Laboratory!LLNL} and John
Hancock\index{Hancock} at Los Alamos\index{Los Alamos National
Laboratory!LANL} helped with the Doppler broadening module.  In those days,
the development of NJOY was supported by the U.S. Fast Breeder
Reactor\index{Fast Breeder Reactor Program} and Weapons
Programs.\index{Weapons Program}

Code development continued during 1975.  The \hyperlink{sERRORRhy}{ERRORR}
\index{ERRORR} module was added for calculating covariances
\index{covariances} from ENDF/B files.  The \hyperlink{sACERhy}{ACER}
\index{ACER} module was created by borrowing heavily from
ETOPL\index{ETOPL} and Chuck Forrest's\index{Forrest}
MCPOINT \index{MCPOINT} code.  Rich Barrett\index{Barrett}
joined the project, and he did most of the work in creating a new
\hyperlink{sCCCCRhy}{CCCCR}\index{CCCC} module for NJOY
that had several advances over the MINX version and met the
CCCC-III\index{CCCC format!CCCC-III} standards\cite{CCCC3}.  By
the end of the year, \hyperlink{sHEATRhy}{HEATR}\index{HEATR}
had also been added to the code (with ideas from Doug Muir\index{Muir}).
\hyperlink{sHEATRhy}{HEATR} gave NJOY most of the capabilities of
the original KERMA factor\index{KERMA} code, MACK
\cite{MACK}.\index{MACK}

During 1976, free-form input\index{free-form input}
and dynamic data storage were\index{dynamic data storage}
added to NJOY.  \hyperlink{sGAMINRhy}{GAMINR}\index{GAMINR}
was written to complete the original NJOY goal of processing
photon interaction cross sections, and the \hyperlink{sMATXSRhy}{MATXSR}
\index{MATXSR} module was designed and written,
primarily by Rich Barrett.  This completed the capability to
construct fully coupled cross sections for neutron-photon heating
problems.  A major new effort was writing the
\hyperlink{sTHERMRhy}{THERMR}\index{THERMR}
module to improve upon the thermal moderator scattering cross sections
then produced using the FLANGE-II\cite{FLANGE}\index{FLANGE-II}
and HEXSCAT\cite{HEXSCAT}\index{HEXSCAT} codes, and starting the
\hyperlink{sPOWRhy}{POWR}\index{POWR} module to produce cross
sections for the EPRI-CELL\index{EPRI!EPRI-CELL} and EPRI-CPM
\index{EPRI!EPRI-CPM} codes used by the U.S. electric utility
companies.  This work was funded by the Electric Power Research
Institute (EPRI).

The first release of NJOY to what was then called the Radiation
Shielding Information Center (RSIC) at Oak Ridge and to the
National Energy Software Center (NESC) at Argonne was
NJOY77\index{NJOY 77} in the summer of 1977.  This
version was tested and converted for IBM machines by
R. Q. Wright\index{Wright} (ORNL).  Also, TRANSX\index{TRANSX}
was developed during 1977, the MATXS1\index{MATXS1} 30x12 library
\index{30x12 library} was produced based on ENDF/B-IV\index{ENDF!ENDF/B-IV},
the flux calculator\index{flux calculator} was added to NJOY
to support the EPRI library work, and the first version of the
EPRI-CELL\index{EPRI!EPRI-CELL} library was generated and used.

A second release of NJOY called NJOY78\index{NJOY78} was made
in 1978\cite{NJ78}.  In addition, further improvements were made
for preparing EPRI cross sections, the MATXS/TRANSX\index{MATXS/TRANSX}
system was improved, and a thermal capability was added to the
MCNP\index{MCNP} Monte Carlo\index{Monte Carlo} code using
cross sections from \hyperlink{sTHERMRhy}{THERMR}\index{THERMR}
as processed by \hyperlink{sACERhy}{ACER}.\index{ACER}

In 1979, the radiation damage\index{radiation damage} calculation
was added to \hyperlink{sHEATRhy}{HEATR}\index{HEATR}, and
the \hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR}
flux calculator\index{flux calculator} was further improved.
In 1980, a plotting option\index{covariance plotting} was added
to \hyperlink{sERRORRhy}{ERRORR}\index{ERRORR}.  During this period,
NJOY had become more stable.  Changes usually consisted of small
improvements or bug fixes instead of major new capabilities.
Starting in this period, NJOY received some support from the U.S.
Magnetic Fusion Energy Program\index{Magnetic Fusion Energy Program},
mostly for covariance work and TRANSX related library support.

In 1981 and 1982, improvements included the momentum-conservation
method for radiative capture in \hyperlink{sHEATRhy}{HEATR}
\index{HEATR}. Analytic $\psi\chi$ broadening\index{$\psi\chi$
broadening} was added to \hyperlink{sRECONRhy}{RECONR}\index{RECONR}
for some cases, and the integral criteria for resonance reconstruction
with significant-figure control were installed in
\hyperlink{sRECONRhy}{RECONR}.  Several new capabilities
were added to \hyperlink{sERRORRhy}{ERRORR}\index{ERRORR}, and the
\hyperlink{sCOVRhy}{COVR}\index{COVR} module was added to NJOY to handle
both \hyperlink{sERRORRhy}{ERRORR} plotting and covariance library
output. Much of this
covariance-related work was done by Doug Muir\index{Muir}. Wiley
Davidson (LANL) and B.H. Broadhead and R. W. Peele at ORNL were
helpful.  Some support for the documentation work came from
the Paul Scherer Institute (PSI) in Switzerland and OECD Nuclear
Energy agency during visits to those institutions.  In addition,
\hyperlink{sCCCCRhy}{CCCCR}\index{CCCCR} was updated to the
CCCC-IV\index{CCCC format!CCCC-IV} standards.  A new release,
NJOY (10/81),\index{NJOY 10/81} was made to the
code centers, and the first two volumes of a new
NJOY report were written and published.  European users began
to make important contributions about this time.  Enrico
Sartori\index{Sartori} of the NEA Data Bank\index{NEA Data Bank},
then at Saclay in France, Margarete Mattes\index{Mattes}
of the University of Stuttgart in Germany, and Sandro
Pelloni\index{Pelloni} of the Paul Scherer Institute
\index{Paul Scherer Institute} in Switzerland deserve mention.

Another major release, NJOY 6/83\index{NJOY 6/83}, was made in 1983.
By this time, NJOY was in use in at least 20 laboratories in the United States
and around the world.  Small improvements continued, such as the kinematic
KERMA calculation\index{KERMA!kinematic KERMA} in
\hyperlink{sHEATRhy}{HEATR}\index{HEATR}.  The
temperature dependence of the \hyperlink{sBROADRhy}{BROADR}
energy grid was introduced early in 1984 based on an
observation by Ganesan (India).  Volume IV of the NJOY report was
published in 1985, and Volume III appeared in 1987.

The next big set of improvements in NJOY was associated with the
introduction of the ENDF-6 format\index{ENDF!ENDF format!ENDF-6 format}.  This
required significant changes in \hyperlink{sRECONRhy}{RECONR}
\index{RECONR} to support new resonance formats like
Reich-Moore\index{Reich-Moore} and Hybrid R-Function
\index{Hybrid R-Function} (implemented with help from Charlie
Dunford\index{Dunford} of Brookhaven), in
\hyperlink{sHEATRhy}{HEATR}\index{HEATR} to implement
direct calculations of KERMA\index{KERMA!direct KERMA} and
damage from charged-particle and recoil distributions in
File 6\index{File 6}, in \hyperlink{sTHERMRhy}{THERMR}\index{THERMR}
to support new formats for File 7\index{File 7}, and in
\hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR} to
support the group-to-group transfer matrices using
energy-angle\index{energy-angle data} data from
File 6.  The \hyperlink{sPLOTRhy}{PLOTR}\index{PLOTR}
module was also developed during
this period.  The result was the release of NJOY89\cite{NJ89}
\index{NJOY89} in time for processing the new
ENDF/B-VI\index{ENDF!ENDF/B-VI} library and the
JEF-2 \index{JEF-2} library (which was also in ENDF-6 format).

During 1989 and 1990, initial processing of ENDF/B-VI and JEF-2
exposed a number of small problems that had to be fixed.  In
addition, the \hyperlink{sACERhy}{ACER}\index{ACER} module
was rewritten to clean it up,
to add capabilities to produce ACE dosimetry\index{ACE dosimetry data}
and photoatomic libraries,\index{ACE photoatomic data} and
to provide for convenient generation of files in several
different formats for users away from LANL.  A
\hyperlink{sMIXRhy}{MIXR} module\index{MIXR} was added
to NJOY, mostly to allow elemental cross sections to be reconstructed
from ENDF/B-VI isotopes for plotting purposes.  A new technique
was introduced into \hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR}
and all the output modules for
multigroup data that provided for more efficient
processing of fission and photon production matrices with lots of
low-energy groups.  Major revisions were made to the
MATXS format\index{MATXS format} to allow for charged-particle
cross sections,\index{charged-particle cross sections}
to pack matrices with lots of low-energy groups more efficiently,
and to make inserting and extracting new materials easier.  The
\hyperlink{sWIMSRhy}{WIMSR}\index{WIMSR} module, which
had been under development for
a number of years in cooperation with WIMS\index{WIMS} users
in Canada and Mexico, was introduced into NJOY.  Finally, the
\hyperlink{sPURRhy}{PURR}\index{PURR} module for generating
unresolved-region \index{unresolved resonance range} probability
tables\index{probability tables} for use with MCNP\index{MCNP},
which had also been under development
for many years, was formally added to the code.  The result of
this year-and-a-half of work was NJOY91.\index{NJOY91}

During the balance of 1991 and 1992, a number of changes were made
in response to problems identified by users as ENDF-6 evaluations
began to be used in earnest.  The first of a number of attempts to
handle laboratory-frame Legendre data in File 6 in MCNP was
made --- namely, the attempt to convert such sections to use Kalbach
systematics.  Treatments for phase-space and angle-energy versions
of File 6 were added in 1993, as well as improvements in
\hyperlink{sWIMSRhy}{WIMSR}\index{WIMSR} (with contributions of
Fortunato Aguilar\index{Aguilar}, ININ, Mexico).  Plotting was added for
$S(\alpha,\beta)$ curves from ENDF-6 File 7 and for 2-D
and 3-D plots from GENDF data.  \hyperlink{sHEATRhy}{HEATR}\index{HEATR}
was modified to include incident-energy effects on the fission Q values
in 1993.  In 1994, we made another attempt to handle the cases of
File 6 using laboratory-frame Legendre expansions for MCNP by
converting them into the LAW=7 angle-energy format.  A fix to
the Kalbach option for energy-angle distributions was made based
on work by Bob Seamon\index{Seamon} (LANL).  Also, work was
done in \hyperlink{sHEATRhy}{HEATR}\index{HEATR} to properly
handle the damage energy \index{damage energy} cutoff at low
energies.  Quite a bit of work was done during this period to
improve the portability of the code by installing it on a variety
of systems and using codes like cflint.  Other people providing
suggestions during this period included Margarete Mattes\index{Mattes}
(Stuttgart), Piet De Leege\index{DeLeege} (Delft), John White
\index{White} (RSIC), and the Petten users.  The final version of
the NJOY91 series was 91.118 in November 1994, and it was
able to process all of the ENDF/B-VI evaluations that had been
tried up to that date.  A new user manual for NJOY91\cite{NJOY91}
was released in 1994.  This document was the
primary reference for NJOY until the release of
NJOY2012 manual\cite{NJ12}.

NJOY94\index{NJOY94} was issued at the end of 1994 to clean up
91.118 after 3 years of changes.  It also provided a new
direct-to-Postscript plotting\index{Postscript plots} system
by splitting the old PLOTR module into
\hyperlink{sPLOTRhy}{PLOTR}\index{PLOTR} and
\hyperlink{sVIEWRhy}{VIEWR}\index{VIEWR}, an updated
version of the \hyperlink{sPURRhy}{PURR}\index{PURR}
module, and a new \hyperlink{sLEAPRhy}{LEAPR}\index{LEAPR}
module for computing thermal
scattering functions.  In 1995, a new capability was added to
handle nuclide production based on a proposed extension
to the ENDF-6 File 8, and the new \hyperlink{sGASPRhy}{GASPR}
\index{GASPR} module was added to handle gas production.  Early
in 1996, a capability to
pass damage cross sections into MCNP\index{MCNP} was added.
A number of changes were made based on suggestions from users,
and based on processing experience with the new 150-MeV evaluations
becoming available during this period.  A change in the erfc
function was made to improve the consistency of $1/v$ cross
sections (problem observed by Cecil Lubitz\index{Lubitz}, KAPL).
Late in the year, the plotting system was upgraded to promote
color Postscript plotting.  In mid 1997, a large update to the
\hyperlink{sWIMSRhy}{WIMSR}\index{WIMSR} module was
made based on the work of Andrej Trkov\index{Trkov}
(Slovenia).  Problems with calculational accuracy (especially for
\hyperlink{sBROADRhy}{BROADR}) led to a major upgrade of the
math routines.  Routines from the SLATEC library were adopted
and converted to NJOY style. A new capability was added to
\hyperlink{sBROADRhy}{BROADR}\index{BROADR}
to compute and display some standard thermal quantities, such
as thermal cross sections, g factors, and ratio integrals ($\alpha$, K1).
 After some additional bug patches and portability improvements,
the NJOY94\index{NJOY94} series was concluded with the issuance
of 94.105 in July of 1997.

NJOY 97.0\index{NJOY97} was released in October of 1997.  The
major change was to move to using 8-byte words throughout the code.
Modern resonance evaluations push the limit of 6 significant figures
available using 32-bit words, and obtaining consistency between people
using 32-bit machines and people on 64-bit machines was becoming more
and more desirable.  Many changes were associated with this,
including taking great care with all literal constants in the
code, removing Hollerith constants, changing from the \cword{FREE}
input routine to a more standard READ* method, and developing
the techniques to use either 7-digits or 9-digits in resonance
grids, as needed.  SAVE statements were added to support stack-based
compilers.  The code now automatically determined the version of
ENDF data on the input files, which changed the input a little.
Finally, some steps were made to remove statement numbers and move
part way to Fortran-90 style without abandoning Fortran-77
compatibility.

A number of changes were made during 1998 to cement the features
of the new version.  The first new capability added was new
conditional probabilities in \hyperlink{sPURRhy}{PURR}\index{PURR}
for heating.  This feature requires running \hyperlink{sHEATRhy}{HEATR}
\index{HEATR} with partial cross sections for
elastic, fission, and capture.  Corresponding changes were also
made in \hyperlink{sACERhy}{ACER}\index{ACER}.  The tolerances in
\hyperlink{sRECONRhy}{RECONR}\index{RECONR}
and \hyperlink{sBROADRhy}{BROADR}\index{BROADR} were
modified to use a tighter tolerance in the thermal range to help
preserve the 0.0253-eV cross section and other thermal parameters
better (Lubitz\index{Lubitz}).  Piet DeLeege\index{DeLeege} (Delft),
Sandro Pelloni\index{Pelloni} (PSI), and Andrej Trkov\index{Trkov}
(IJS/Slovenia) were helpful in finding problems during this
period.  A large number of changes were made based on
issues raised by testing the code with the ftnchek program.  These
help for portability between different computer systems.  This
work was motivated after Giancarlo Panini\index{Panini} (Italy)
questioned small differences seen using different compilers.  Early
in 1999, the method for setting the damage threshold in
\hyperlink{sHEATRhy}{HEATR}\index{HEATR} was
revised.  An internal table of default values was provided, and
a new option to allows the user to enter a value for $E_d$ was added.
This change affects the damage near the threshold for elastic
displacements.  John White\index{White} (RSIC) helped with this.
Based on some code comparisons, Nancy Larson\index{Larson} (ORNL)
promoted updating the basic constants in the various codes to
enhance compatibility, and we followed suit in NJOY.  More work
was done to enhance activation processing and to provide an
automatic loop over all the production reactions for
\hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR}.  Finally,
a new option to properly handle
channel spin in Reich-Moore resonance evaluations was added using
code contributed by ORNL.  The NJOY97\index{NJOY97} series was
frozen in September of 2000.

During the last part of the life of NJOY97\index{NJOY97}, we also
had NJOY99\index{NJOY99} available (dated 31 December 1999).  It is
a cleaned up version of NJOY97\index{NJOY97} that moves further
toward using block constructs and eliminating statement numbers
(but lots of statement numbers are still left).  It is compatible
with both Fortran-77 and Fortran-90 compilers.  Physical constants
were moved into a few common blocks and standardized on the 1987
CODATA values from NIST.  The NJOY Y2K problem was fixed.  The bulk
of the changes are in the \hyperlink{sACERhy}{ACER}\index{ACER}
module to support new MCNP\index{MCNP} features, including
high-energy data, incident charged particles, and photonuclear
data.  Before release, this version was tested for a large number of
different machine/compiler combinations, including tests on X86
PCs.  Compatibility has improved to the point where the unix diff
function can be used to compare files and only a small number of
very small differences are found.  This version was able to
process all the materials from ENDF/B-VI Release 5 into a library
using the new ACE formats (cumulative elastic distributions,
LAW=61, and charged-particle production sections) that were
planned to come out with MCNP4c.

We began making patches to NJOY99\index{NJOY99} in the spring
of 2000, starting with fixing a problem with the cold hydrogen
and deuterium calculations in
\hyperlink{sLEAPRhy}{LEAPR}.\index{LEAPR}  This was
followed by a number of other small patches.  A capability
was added for processing anisotropic charged particle emission
in \hyperlink{sACERhy}{ACER}.\index{ACER}  In 2001 the
series of MT numbers from 875 -- 891
was installed to represent levels in the (n,2n) reaction
(needed for a new European $^{9}$Be evaluation).  We added a capability
to include delayed neutron data in the ACE files to feed a new
capability in MCNP\index{MCNP}. A photonuclear capability was added
to \hyperlink{sMATXSRhy}{MATXSR}\index{MATXSR} --- this enables
the TRANSX\index{TRANSX} code to generate fully coupled sets
for n\nobreakdash-$\gamma$ transport.  Some
coding was added to generate fluorescence data for MCNP\index{MCNP}
using the existing format with new numbers coming from the ENDF/B-VI
atomic data.  This work does not completely support all the
atomic data now available in ENDF/B-VI.  The
\hyperlink{sACERhy}{ACER}\index{ACER}
consistency checks were upgraded to include delayed neutrons, and
plots for $\bar{\nu}$ and the delayed neutron spectra were added.  In
addition, delayed neutron processing was generalized to allow
for 8 time bins as used by the JEFF evaluations.

\hyperlink{sPLOTRhy}{PLOTR}\index{PLOTR} was modified
to allow ratio and difference plots using the right-hand scale.  The
default tolerances used in \hyperlink{sRECONRhy}{RECONR}\index{RECONR}
and \hyperlink{sBROADRhy}{BROADR}\index{BROADR} were tightened
up a bit.  Some changes to the heating for photoatomic data were
provided by Morgan White\index{White} (LANL).  A change was made
to \hyperlink{sHEATRhy}{HEATR}\index{HEATR} to provide the photon
contribution to heating using a special MT number
(442)\index{MT!MT442}.  When passed to
MCNP\index{MCNP}, it allows the code to get good answers for
heating even when photons are not being transported.  Late
in 2002, \hyperlink{sLEAPRhy}{LEAPR}\index{LEAPR} was updated
to include coherent elastic scattering for FCC and BCC crystalline
lattices.  In 2004, some extensions to the energy grid used for
incoherent inelastic scattering were made.  A few additional
smaller patches were also made during this period.

In 2005, changes proposed for \hyperlink{sTHERMRhy}{THERMR}
\index{THERMR} and \hyperlink{sLEAPRhy}{LEAPR}\index{LEAPR}
by Margarete Mattes\index{Mattes} (IKE/ Stuttgart)
to support the new IAEA-sponsored evaluations for thermal scattering
in water, heavy water, and ZrH were installed.  Some additional
group structures used in Europe were added.  Through this period,
we were always increasing the storage space allowed as we adapted
to newer and larger evaluations coming out for ENDF/B-VII.
During the summer, a number of changes to covariance processing
were provided by Andrej Trkov\index{Trkov} (working at IAEA).  Early
in 2006, a new sampling scheme for thermal scattering was developed
for MCNP\index{MCNP} that uses continuous distributions for
secondary energy instead of the previous discrete values.  This
removed unsightly artifacts in computed fluxes at low energies, and
it alleviated some problems that the cold-neutron-source people were
having.  More code improvements were made based on detailed compiler
checking.  Some errors in the treatment of energy-dependent
fission Q were fixed based on a review of the work of Dave
Madland\index{Madland} (LANL).  An approximate treatment for the
relativistic gamma in the ENDF/B-VII evaluation for n+$^1$H was added
in \hyperlink{sHEATRhy}{HEATR}\index{HEATR} and
\hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR}  based on theoretical
work from Gerry Hale\index{Hale} (LANL).  We added a new plot to the
\hyperlink{sACERhy}{ACER}\index{ACER} set that shows the recoil part
of the heating.  This is a sensitive test of energy-balance.

In the summer of 2007, we added some smoothing options to make
the low-energy shape of neutron distributions look more like
the theoretical shape, namely ${\rm sqrt}(E')$.  Similar changes
were provided for delayed neutron spectra.  Additional smoothing
was provided for some of the fission spectra at energies above
10 MeV using an exponential shape. A change was made in
\hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR} to override
Cartesian interpolation in favor of unit-base interpolation for
scattering distributions.  This gives smoother scattering source
functions and is consistent with what MCNP\index{MCNP} does.

In 2007, a big change was made to covariance processing by
replacing the original NJOY ERRORR\index{ERRORR} module with
ERRORJ\cite{ERRORJ}\index{ERRORJ} as contributed to the NJOY project
by Japan (Go Chiba\index{Chiba}).  This new module added covariance
capabilities for the more modern resolved-resonance representations,
angular distributions, and secondary-energy distributions.  A series
of additional changes to the new \hyperlink{sERRORRhy}{ERRORR} were
made over the next couple of years.  Go Chiba\index{Chiba}, Andrej
Trkov\index{Trkov} (IAEA and IJS Slovenia), and Ramon
Arcilla\index{Arcilla} (BNL) were involved in this.  A capability
to handle energy-dependent scattering radius data in the unresolved
range was added.  In 2008, ERRORJ\index{ERRORJ} work continued.
Some work was done in \hyperlink{sPLOTRhy}{PLOTR}\index{PLOTR}
to implement graphs of \hyperlink{sGROUPRhy}{GROUPR}\index{GROUPR}
emission spectra.  In 2009, changes were made
for unresolved resonance cross sections to force log-log interpolation
to better represent $1/v$ cross sections.  The parameter that
looked for steps in the unresolved-range energy grids that were
too large was changed from its former value of 3 to 1.26.  Some
of the steps in ENDF/B-VII are unreasonably large when representing
$1/v$ cross sections.  The default energy grid used in the
unresolved range was upgraded to one using about 13 points per
decade.  Additional pages were added to the \hyperlink{sACERhy}{ACER}
\index{ACER} plots to display the unresolved-range self-shielded
cross sections.  This work was influenced by Red Cullen\index{Cullen}
(LLNL).  In a related effort, some changes were made in the binning
logic for \hyperlink{sPURRhy}{PURR}\index{PURR} to improve
unresolved-range results.  In 2010, changes were made to support
the processing of the IRDF international dosimetry file, including the
addition of many new reaction MT numbers.  Some work was done on the
photonuclear options to support the TENDL-2009 library.  More
\hyperlink{sERRORRhy}{ERRORR} changes from Trkov\index{Trkov}
were implemented, and a capability to process the uncertainty
in the scattering radius was added.  Finally, several updates were
submitted by the Japan Atomic Energy Agency (JAEA) related to
processing discrete photon data from file 6 into ACE files.  NJOY 99.364
was released in February of 2011.

Concurrent with the maintenance of NJOY99\index{NJOY99}, we worked
on a new Fortran-90 version of NJOY that took advantage of the
module system, the numerical precision system, and the dynamic storage
system of modern Fortran.  This enabled us to get rid of COMMON statements
(always problematic in Fortran) and to obtain a more readable
scheme for allocated storage.  Additional work was done to upgrade
to block structures, but some statement numbers still remain.
That new version was called NJOY2010\index{NJOY2010}, but was only
released to a few in-house (LANL) users, selected external users
at Brookhaven, Oak Ridge and Argonne National Laboratories plus the
Naval Reactors laboratories, Bettis and KAPL in the United States and to
the Atomic Weapons Establishment in the UK.

Further code revisions
lead to NJOY2012\cite{NJ12}\index{NJOY2012}.  For most
applications NJOY2012 is compatible
with NJOY99.  However, there were some new capabilities in NJOY2012
that are not supported by NJOY99.  We borrowed coding from the
SAMMY\cite{SAMMY} code (thanks are due to Nancy Larson\index{Larson},
ORNL) to handle the new Reich-Moore-Limited resonance parameter
representation now included in the ENDF-6 format.  This coding
is used in \hyperlink{sRECONRhy}{RECONR}\index{RECONR} to
generate pointwise resonance cross sections and angular
distributions.  It is used in \hyperlink{sERRORRhy}{ERRORR}
\index{ERRORR} to generate the sensitivity of resonance
cross sections to the resonance parameters for use in
calculating cross section uncertainties and covariances in the
resonance range.  In \hyperlink{sTHERMRhy}{THERMR}\index{THERMR},
we added an option to construct thermal cross sections using ($E,\mu,E'$)
ordering in addition to the normal ($E,E',\mu$) ordering.  This
ordering is convenient for comparing to experiment, and it is used
for thermal sampling in some Monte Carlo codes.

With NJOY2016 we now have an ``open source'' code release.  The
current version of the source code is now freely available, and can be
downloaded from \href{http://njoy.lanl.gov}{http://njoy.lanl.gov}.

\cleardoublepage