Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

230 K ice/brine front frames

Direct-coexistence trajectories of ice Ih against TIP4P/2005 water with Madrid-2019 scaled-charge NaCl, at 230 K. Three loadings, replica r0: 0, 20 and 40 NaCl pairs. Production is 5 ns (2,500,000 steps of 2 fs) after a 100 ps melt at 400 K and a 100 ps quench to 230 K, with the ice half held during melt and quench.

The deck is the campaign input for results/brine/T230_m{0,20,40}_r0. GenIce builds a hydrogen-disordered ice Ih block (1h, replication 4 4 6, 1536 waters, TIP4P geometry). The upper half along z is liquid. NaCl pairs replace liquid waters. The M site is implicit in pair_style lj/cut/tip4p/long. Charges are O -1.1128, H 0.5564, Na +0.85, Cl -0.85. Production is NPT along z at 1 bar (fix npt, Tdamp 200 fs, Pdamp 2000 fs).

Versions and seeds

  • LAMMPS 22 Jul 2025, Update 4, OpenMP package.
  • GenIce2 2.2.13.3 with genice-core 1.2.1. The deck calls genice2 1h --rep 4 4 6 --water tip4p and does not set GenIce's seed, so GenIce uses its default seed 1000.
  • Replica r0. Ion sites are drawn with numpy seed 0.
  • Velocity seed 03 ({replica}3 with replica 0), used for the 400 K melt and again when the ice is released.

Loadings

tag NaCl pairs waters Na Cl atoms
T230_m0_r0 0 1536 0 0 4608
T230_m20_r0 20 1496 20 20 4528
T230_m40_r0 40 1456 40 40 4448

The box starts near 31.29 x 29.41 x 54.22 angstrom. Production NPT moves the z length.

Files

data/frames/T230_m{0,20,40}_r0_{t0,t1,t2,tend}.xyz are extended XYZ frames. Columns are species, position, and the d-SEAMS CHILL+ label on each oxygen (- on H, Na and Cl). The four times are the production steps 0, 830000, 1670000 and 2500000 (0, 1.66 ns, 3.34 ns and 5.00 ns).

data/restarts/ holds LAMMPS restart files, one directory per loading.

  • step-2500000.restart is the 5 ns production state.
  • end.restart is the restart written at the end of the last chunk.
  • restart.a and restart.b are the alternating pair from that chunk.
  • step-1255000.restart through step-2380000.restart (and the matching names for 20 and 40 pairs) are checkpoints every 250 ps (125000 steps) from about 2.5 ns to 5 ns. There are 42 of these production files.
  • step-0.restart, step-830000.restart and step-1670000.restart are the other three frames the Snakemake rule reads. They were rebuilt from the trajectory with deck/in.fromdump.lmp so in.resume.lmp can dump them with run 0. A rerun of in.brine.lmp writes those three steps, and step 2500000, itself.

deck/in.resume.lmp and regenerate.py are the inputs rule brine_frame passes to LAMMPS and to the labeller. Snakefile is that rule. config.yaml is its config: dataset: d-SEAMS/brine-frames, with restarts in data/restarts/ and frames in data/frames/. SHA256SUMS lists every other file in this tree.

Regenerate the twelve frames

From a seams-core checkout that contains paper/data/brine-frames/, place this dataset at d-SEAMS/brine-frames in the repo root (the default in config.yaml). lmp, seams and libyodaLib.so must be on PATH, or pass them with --config.

snakemake -s paper/data/brine-frames/Snakefile -j1 \
  paper/data/brine-frames/out/T230_m0_r0_t0.xyz \
  paper/data/brine-frames/out/T230_m0_r0_t1.xyz \
  paper/data/brine-frames/out/T230_m0_r0_t2.xyz \
  paper/data/brine-frames/out/T230_m0_r0_tend.xyz \
  paper/data/brine-frames/out/T230_m20_r0_t0.xyz \
  paper/data/brine-frames/out/T230_m20_r0_t1.xyz \
  paper/data/brine-frames/out/T230_m20_r0_t2.xyz \
  paper/data/brine-frames/out/T230_m20_r0_tend.xyz \
  paper/data/brine-frames/out/T230_m40_r0_t0.xyz \
  paper/data/brine-frames/out/T230_m40_r0_t1.xyz \
  paper/data/brine-frames/out/T230_m40_r0_t2.xyz \
  paper/data/brine-frames/out/T230_m40_r0_tend.xyz

brine_frame reads step-{0,830000,1670000,2500000}.restart, runs deck/in.resume.lmp for 0 steps, and writes the CHILL+ label with regenerate.py frame. The same command with paper/tex/figures/brine_front.pdf also builds the two-panel figure from the 20-pair start and 5 ns frames. That PDF is not in this dataset.

Downloads last month
4,343