Hydrovoltaic energy — electricity from water flowing over a surface — depends on charged surfaces interacting with ions in solution. MoS₂ is a promising material because it's hydrophilic and atomically flat. Both its metallic (1T') and semiconducting (2H) phases are MoS₂ — same atoms, same stoichiometry. You'd expect similar hydrovoltaic performance from both, with maybe a modest edge for one.
Note (2026-11-14; read against the paper's full text): the figures (2.0 milliwatts per square metre for the metallic 1T' phase and 2.4 microwatts per square metre for the 2H phase, a ratio of about 833, with channels of about 6 angstroms) are the paper's, from water-evaporation-based generators made of laminates about 1 micrometre thick. The two samples are not simply the same atoms in different arrangements: the 1T' phase is made by lithium intercalation and chemical exfoliation, the 2H phase by annealing it, and the paper reports lithium in the 1T' phase and not in the 2H phase, whose negative surface charge it attributes mainly to sulfur vacancies, and the two phases also differ in conductivity (metallic versus semiconducting). The paper finds the surface charge and wetting similar (zeta potentials of minus 38.8 and minus 33.0 millivolts, contact angles of about 34 and 37 degrees) and concludes from its theoretical analysis that the generation is mainly electrokinetic and surface diffusion; the essay's account of channels multiplying solid-liquid interfaces and its claim that this is a channel property and not a surface property are stronger than the paper's wording.
Suvigya, Lalita, Nemala Sankar, Capasso, Yeh, and Gopinadhan (arXiv:2501.01739) find an 833× difference. The metallic 1T' phase produces 2.0 mW/m² peak power density; the semiconducting 2H phase produces 2.4 μW/m². Nearly three orders of magnitude separating chemically identical materials.
The mechanism: the 1T' phase forms laminates with well-defined 6-angstrom interlayer channels. The 2H phase doesn't form ordered channels. Water flowing through the 6-angstrom channels experiences surface diffusion and electrokinetic effects at every interface. No channels, no interface multiplication — the 2H phase relies only on its outer surface.
The structural insight is that hydrovoltaic performance is not a surface property but a channel property. The relevant parameter isn't the material's surface charge or hydrophilicity — it's the existence of ordered nanoscale conduits that multiply the number of solid-liquid interfaces per unit area. The same atoms, arranged to create channels versus remaining disordered, produce a factor of 833 in output. The chemistry is necessary but not sufficient. The geometry does the work.