Data Cards in MCNP

After the blank line that closes the surface block, everything else in an MCNP deck is a data card. Geometry has already named the regions and the surfaces that bound them; data cards say what fills those regions, which particles to follow, how the run ends, and what gets written out. This page covers physics control, execution, temperature, and output. Composition belongs on Material Cards; tallies and sources have their own pages later in the path.

The cards fall into a few practical groups — materials (M, MT, TMP), physics (PHYS, CUT, MODE), problem control (NPS, CTME, PRDMP, and for eigenvalues KCODE / KSRC), and output (PRINT, PTRAC, DBCN). Start with a thin set — usually MODE, a termination card, and the materials the cells already reference — and add the rest when the problem needs them.

These cards decide which particles exist in the run and which interactions are allowed. Omit MODE and MCNP transports neutrons only; name photons (or electrons) when the problem actually needs them.

c Particle transport modes
MODE  n p                 $ Transport neutrons and photons

c Physics parameters
PHYS:N  100  0  0  -1  -1  $ Neutron physics to 100 MeV
PHYS:P  100  0  1          $ Photon physics to 100 MeV

c Energy cutoffs (J = jump, meaning use default for that entry)
CUT:N   J  0.001           $ Default time, 1 keV energy cutoff
CUT:P   J  0.01            $ Default time, 10 keV energy cutoff

On a PHYS card the first entry is the maximum energy for that particle type; later entries toggle specific models. CUT sets the energy (and optionally time) below which a particle is killed. Cut too high and you lose the soft part of the spectrum; cut too low and you spend time tracking particles that no longer matter. The J on those cards means jump — leave that positional parameter at its default and keep reading. The same letter shows up on PRDMP, DBCN, and other multi-entry cards throughout this page.

c Criticality-specific controls
TOTNU    NO                $ Use prompt nu-bar only (no delayed neutrons)
ACT      FISSION=ALL      $ Produce all delayed particles from fission

c Variance reduction
IMP:N    1 1 0 2 4        $ Neutron importance by cell
IMP:P    1 1 0 2 4        $ Photon importance by cell
FCL:N    1 1 1            $ Force collisions in cells 1-3

TOTNU chooses whether delayed neutrons contribute to ν-bar in a criticality calculation. IMP sets cell importances for variance reduction, and FCL forces collisions in named cells. None of these are required for a first pin-cell eigenvalue; they appear here so the names are familiar when a later page needs them.

How the run stops depends on the problem type. A fixed-source deck counts particle histories; a criticality deck counts fission cycles. Mixing the two termination styles — or forgetting a terminator entirely — is a common way to get a run that never finishes for a reason that is hard to see in the geometry.

Fixed-source problems terminate on history count or wall-clock time:

c Basic execution control
NPS    1e6                 $ Run 1 million particle histories
CTME   60                  $ Time limit: 60 minutes

c Statistical checks
DBCN   17J 1              $ Debug control card
PRDMP  J  1e6  1  2       $ Dump RUNTPE every 1e6 histories; write MCTAL

c Random number control
RAND   GEN=2  SEED=12345  $ Random number generator

NPS is the primary stop. CTME is a safety net in minutes — useful on a cluster queue with a hard time limit. PRDMP writes restart and MCTAL files so a long job can be resumed or post-processed. RAND pins the generator and seed when you need bit-reproducible results. MCNP also runs ten statistical checks on each tally; if they fail, the fix is usually more histories, not a different card spelling.

Criticality problems use KCODE instead of NPS. Hover each line below for the field-by-field reading.

c Criticality calculation setup
KCODE  10000  1.0  50  250
KSRC   0 0 0               $ Initial source point
       10 0 0              $ Additional points for coverage
 
c Criticality-specific controls
KOPTS  BLOCKSIZE=10        $ Batch size for statistics
CTME   120                 $ Time limit: 2 hours
PRDMP  J  25  1  2         $ Dump RUNTPE every 25 cycles

KCODE sets particles per cycle, an initial k guess, inactive cycles for source convergence, and the total cycle count. KSRC places the starting fission points inside fissile cells — never on a surface, never in a void. KOPTS groups cycles into blocks for better uncertainty estimates. The same CTME and PRDMP cards work here as a wall-clock stop and a restart cadence.

Material composition — nuclides, fraction signs, library suffixes, and mt thermal tables — lives on Material Cards. Density signs belong on the cell card and are covered on Cell Cards. What remains here is how temperature reaches the transport physics through TMP.

c Water at room temperature
m1    1001.80c  2.0        $ Hydrogen
      8016.80c  1.0        $ Oxygen
mt1   lwtr.20t             $ Thermal scattering data (293.6 K)

c Fuel at operating temperature
m2    92235.80c  0.05      $ U-235
      92238.80c  0.95      $ U-238
      8016.80c   2.0       $ Oxygen

c Graphite moderator
m3    6000.80c  1.0        $ Carbon
mt3   grph.20t             $ Graphite thermal treatment (293.6 K)

c Temperature card: one value per cell (kT in MeV)
c Assumes cells use materials 1, 2, 3 in order
TMP   2.5301e-8  6.6918e-8  8.6173e-8

TMP takes one entry per cell in the problem, in cell order, with units of MeV (kT = K × 8.617×10⁻¹¹). It is a cell property, not a material property. It adjusts only the free-gas thermal scattering treatment — full Doppler broadening of resonances requires cross-section libraries evaluated at the right temperature (for example .81c / .82c for hot fuel). A zero entry on a non-void cell is a fatal error. Leave the card off and MCNP adjusts free-gas scattering back toward room temperature; resonances stay at whatever the library suffix baked in, so a hot-suffix deck with no TMP is inconsistent. The mt cards in the block above supply the bound-atom kernels for the moderators; those details, and the failure mode when a table has no matching target, are on the materials page.

These cards control what MCNP writes. Default output is verbose; thinning it is usually the first edit once a deck runs cleanly.

c Print table control
PRINT  30 40 50 110 126 -85 -86 -87  $ Print tables (negative suppresses)

c Particle tracking
PTRAC  BUFFER=1000  FILE=ptrac.txt  $ Particle trace file
       WRITE=POS,CEL,MAT,NPS       $ Track position, cell, material

c Performance monitoring
PRDMP  2J  1  2           $ Write MCTAL; keep 2 RUNTPE dumps

On PRINT, a positive table number enables that table and a negative number suppresses it. PTRAC writes a particle-by-particle trace for debugging geometry and source placement. PRDMP again controls MCTAL and RUNTPE — the same card that appeared under calculation control, reused here as an output switch.

c Detailed physics settings
PHYS:N  100  0  0  -1  -1  J  $ J = use default for the next entry
PHYS:P  100  0  1  0   0      $ 3rd entry 1 = coherent scattering off

c Cross-section information
PRINT  104                 $ Cross-section tables
c The XSDIR file location is set via the DATAPATH
c environment variable, not via an input card.

c Debug diagnostics
DBCN   17J 1               $ Extended diagnostics

PRINT 104 dumps cross-section table summaries. The directory of those tables is not an input card — it comes from the DATAPATH environment variable. DBCN turns on extended diagnostics when a run misbehaves in a way the standard tables do not explain.

The running example for this path is a PWR pin cell on Example: Pin Cell. Its data block is four materials (written on Material Cards), a criticality pair, tallies, and a short dump card — not a fixed-source NPS run and not an SDEF source.

The eigenvalue setup is kcode 5000 1.0 50 250 with ksrc 0 0 0.5 — mid-slice, inside the fuel, away from the reflecting planes. prdmp j 300 1 2 writes an MCTAL at the end and keeps two RUNTPE dumps. Those numbers match mcnp-examples/example_pin/pin_cell.i line for line.

There is no MODE card on the pin because neutrons alone are the default, and there is no TMP card because the deck stays on room-temperature libraries. When you later move the fuel to a hot suffix, that is the moment TMP has to appear as well — the materials page walks through why.