Bug in pymatgen.analysis.reaction_calculator.ComputedReaction

[1azyking]

Python version

Python 3.10.6

Pymatgen version

2024.8.8

Operating system version

Windows

Current behavior

rxn: 12 LiCuS -> Cu + Cu7S4 + 4 Li3CuS2
rxn_e: 0.003
reactants: ['H2O', 'LiCuS']
products: ['H2S', 'Cu', 'Cu7S4', 'Li3CuS2']
--------------after normalization to H2O---------------
rxn: 6e+15 LiCuS + 1.489 H2S -> H2O + 5e+14 Cu + 5e+14 Cu7S4 + 2e+15 Li3CuS2
rxn_e: 1499976386498.000
reactants: ['H2O', 'LiCuS']
products: ['H2S', 'Cu', 'Cu7S4', 'Li3CuS2']

version of pymatgen: 2024.8.8

When I take Cu, Cu7S4, and Li3CuS2 as the products of the hydrolysis reaction of LiCuS, it reports a decomposition from LiCuS to Cu, Cu7S4, and Li3CuS2. However, the H2O and H2S are also listed in reactant and product entries, respectively. Moreover, when I normalize the decomposition to H2O or H2S, it reports a wrong hydrolysis reaction with strange stoichiometric numbers.

Expected Behavior

The reaction with imbalanced reactant and product entries is expected to be reported as None. And I wonder why the error normalization occured.

Minimal example

from pymatgen.analysis.reaction_calculator import ComputedReaction
from pymatgen.entries.computed_entries import ComputedEntry
from pymatgen.core import Composition
from pymatgen import core

entry_h2o = ComputedEntry(Composition("H2O"), -3.440)
entry_h2s = ComputedEntry(Composition("H2S"), -1.049)
entry_licus = ComputedEntry(Composition("LiCuS"), -2.592)
entry_cu = ComputedEntry(Composition("Cu"), 0.0)
entry_cu7s4 = ComputedEntry(Composition("Cu7S4"), -2.541)
entry_li3cus2 = ComputedEntry(Composition("Li3CuS2"), -7.14)

reactants = [entry_h2o, entry_licus]
products = [entry_h2s, entry_cu, entry_cu7s4, entry_li3cus2]

rxn = ComputedReaction(reactants, products)
print(f"rxn: {rxn}")
print(f"rxn_e: {rxn.calculated_reaction_energy:.3f}")
print(f"reactants: {[e.name for e in rxn._reactant_entries]}")
print(f"products: {[e.name for e in rxn._product_entries]}")
print("--------------after normalization to H2O---------------")
rxn.normalize_to(entry_h2o.composition, 1.0)
print(f"rxn: {rxn}")
print(f"rxn_e: {rxn.calculated_reaction_energy:.3f}")
print(f"reactants: {[e.name for e in rxn._reactant_entries]}")
print(f"products: {[e.name for e in rxn._product_entries]}")
print()
print(f"version of pymatgen: {core.__version__}")

Relevant files to reproduce this bug

No response