Ti3C2 MXene co-catalyst on metal sulfide photo-absorbers for enhanced visible-light photocatalytic hydrogen production

Platinum is wonderful for splitting water, and painfully rare. So, there has been a long-standing hunt for something common, cheap, and just as good at the job. That’s the promise behind a class of materials called MXenes, and in particular one called Ti3C2. It’s metallic, it carries charge extraordinarily well, and its surface can be chemically tuned. In other words, you can engineer not just the light-absorbing semiconductor, but also the surface it hands electrons to. Jingrun Ran and colleagues took that idea and pushed it hard: could Ti3C2 stand in for platinum as the co-catalyst that actually does the hydrogen-making, and do it without sacrificing speed? Here’s the headline answer. When they coupled cadmium sulfide, a classic visible-light absorber, with just 2.5 weight percent of Ti3C2, the composite produced hydrogen at 14,342 millimoles per hour per gram. That’s with an apparent quantum efficiency — the fraction of photons that become hydrogen — of 40.1 percent at 420 nanometers. For comparison, the same amount of platinum on cadmium sulfide produced 10,978, and bare cadmium sulfide limped along at roughly 105. So Ti3C2 didn’t just keep up; it beat platinum in that head-to-head test under the same conditions. That’s a big deal because in solar hydrogen, speed is survival. Why does this weirdly named material help so much? It comes down to two things: how it holds a hydrogen atom on its surface, and how its electronic energy lines up with cadmium sulfide. Let’s start with the surface chemistry. In electrochemistry, there’s a simple yardstick for the hydrogen evolution reaction: the free energy of hydrogen adsorption, usually written as delta G H star. If that number is near zero, the surface holds hydrogen just right — not too weakly to fall off, not too strongly to get stuck. Ran’s team used density functional theory, a quantum-chemical workhorse, to calculate how hydrogen binds to Ti3C2 when the surface is capped with different terminal groups. Pure Ti3C2 grabbed hydrogen far too tightly, with delta G H star around negative 0.93 electron volts. Fluorine-capped Ti3C2 was the opposite — basically disinterested, at plus 2.0. Oxygen-capped Ti3C2, though, sat right on the Goldilocks spot: at half a monolayer coverage, it landed at about 0.0028 electron volts in absolute value. That’s closer to ideal than the numbers they cited for benchmark catalysts like platinum at roughly 0.09, molybdenum disulfide at about 0.08, and tungsten disulfide at 0.22. At that half-coverage, a unit cell could host four adsorbed hydrogens, so the surface isn’t quickly saturated. Below half-coverage, the binding stayed favorable; above it, activity tailed off. This is surface thermodynamics speaking clearly: put oxygen on Ti3C2 and it becomes an intrinsically eager site for hydrogen evolution. The second lever is the electronics — literally where the electrons want to go. Ti3C2 is metallic, so it has a Fermi level, the energy that marks the highest filled electronic state. You want that level to be positioned so that electrons photoexcited in cadmium sulfide can spill downhill into the co-catalyst. Among the terminations they studied, oxygen-capped Ti3C2 had the most positive Fermi level, about 1.88 volts on the standard hydrogen electrode scale. That makes it an excellent electron sink. Cadmium sulfide’s conduction band edge, by contrast, sits around minus 0.79 to minus 0.91 volts on that same scale. Couple the two, and you set up what’s called a Schottky junction: a metal-semiconductor contact that traps electrons on the metal side and slows their undesired march back to recombine with holes. In the optimized composite, the measured conduction band of cadmium sulfide settled near minus 0.79 volts, consistent with that downhill electron transfer toward Ti3C2 and the hydrogen reduction potential. Now, there’s a clever synthetic trick here. As-synthesized Ti3C2 usually carries a mix of surface groups: hydroxyls, oxygens, and fluorines, with that last one being a problem for hydrogen evolution. Ran’s group fixed that in the same step that built the junction. They started from the so-called MAX phase, Ti3AlC2, and etched out the aluminum to make layered Ti3C2. Sonication broke those layers into nanoparticles. Then, during a hydrothermal growth of cadmium sulfide, the hot aqueous conditions helped swap fluorine terminations for oxygen and hydroxyls right on the Ti3C2 surface. You can see it in the composition: the fluorine content dropped from about 20.6 percent in the starting nanoparticles to roughly 8 percent in the hydrothermally treated ones. Cadmium ions latch onto those oxygen sites, thiourea supplies sulfide, and out grow sub-micrometer, cauliflower-like spheres — around four to five hundred nanometers across — where cadmium sulfide and Ti3C2 are in tight contact. On the lattice scale, the interface showed a Ti3C2 layer spacing of about one nanometer and a cadmium sulfide plane spacing of roughly 0.36 nanometers, stitched together. The chemistry is doing two jobs at once: creating the right terminations for catalysis and welding the catalyst to the light absorber so electrons have a short, fast path. It’s worth pausing on what doesn’t happen. Ti3C2 on its own isn’t a photocatalyst under visible light; it needs the semiconductor to supply the excited electrons. Load too much of it, and it starts to smother the cadmium sulfide surface and clump, which cuts down the active area. You can see that in the activity curve: 2.5 percent by weight is the sweet spot, but by 7.5 percent, the rate drops to about 2,707. And these things aren’t fragile. Across seven back-to-back reaction cycles, the 2.5 percent composite held its activity, and its used form looked essentially the same as before testing. That kind of durability is crucial if you’re going to claim “platinum-level” performance without platinum. How do we know electrons are actually taking the intended route? The team threw just about every probe at the interface. Steady-state and time-resolved photoluminescence, which lights up when electrons and holes recombine, went quiet and slowed down in the composite, meaning carriers were being separated and living longer. The emission near 560 nanometers — characteristic of cadmium sulfide recombination — was strongly quenched. Electrochemical impedance spectroscopy, which reads as a semicircle whose size tells you about charge-transfer resistance, shrank dramatically once Ti3C2 was on board, signaling an easier interfacial hop for electrons. Transient photocurrent spikes were larger, and when they added sulfide and sulfite as hole scavengers, the signals grew further, a classic sign that the electrons were being drawn off efficiently while holes were safely consumed. Put together with the band positions and the surface energetics, the mechanistic picture holds: light kicks electrons into the cadmium sulfide conduction band, the Schottky contact funnels them into oxygen-terminated Ti3C2, those electrons meet protons at sites that bind hydrogen just right, and molecular hydrogen bubbles off. There’s also a simple but telling variation on the theme. If Ti3C2 is really acting as a universal electron sink and hydrogen-evolution site, it shouldn’t only work with cadmium sulfide. And it didn’t. Even when they didn’t bother to chemically bond it, just mixing Ti3C2 nanoparticles with zinc sulfide or a zinc-cadmium sulfide alloy lifted both systems substantially. Activities jumped by about 217 percent for zinc sulfide and 386 percent for the zinc-cadmium alloy compared with their pristine forms. That’s not a delicate, one-time-only pairing. It’s a co-catalyst doing its job across different hosts. Could they squeeze more out of the cadmium sulfide system? Yes, by splitting the labor for holes as well as electrons. Nickel sulfide is a p-type semiconductor — it naturally carries holes. When Ran and colleagues co-loaded a small amount of nickel sulfide along with 2.5 percent Ti3C2 on cadmium sulfide, they created a p-n junction for holes to flow into nickel sulfide, while electrons kept streaming to Ti3C2. That three-part assembly hit 18,560 millimoles per hour per gram. The logic is elegant: drive electrons and holes apart in opposite directions, give each of them a surface where the respective reaction is fast, and you minimize the chance they meet and cancel each other out. A few fine points from the calculations help tie the bow. The favorable hydrogen adsorption on oxygen-terminated Ti3C2 emerges at a specific surface coverage — about half of the possible sites — and one unit cell can host four hydrogens at that point. That means the active surface won’t clog instantly. The electronic structure stays metallic with either oxygen or fluorine terminations, so conductivity isn’t sacrificed as you tune the surface. And when cadmium sulfide and Ti3C2 make intimate contact, the measured conduction band of cadmium sulfide shifts slightly, settling near minus 0.79 volts versus the hydrogen couple, while Ti3C2’s Fermi level moves upward. Those shifts are exactly what you expect when charges flow to equilibrate the junction and set up a barrier that keeps electrons on the Ti3C2 side where they can do work. Step back, and the message is surprisingly practical. If you want to replace platinum in visible-light hydrogen production, don’t just search for a different metal. Look for a platform whose surface chemistry you can dial in and whose electronic level lines up to vacuum electrons out of your light absorber. Ti3C2 MXene hits both notes. Oxygen terminations give it nearly ideal binding to hydrogen, its Fermi level sits where cadmium sulfide can feed it, and its high conductivity whisks electrons to abundant, durable sites. The result is platinum-class performance without platinum, a record 40.1 percent photon-to-hydrogen efficiency at a visible wavelength, and stability that survives repeated use. It also hints at something bigger. MXenes are a family, not a single material. If one member, with the right surface, can rival or outdo a precious metal, that suggests a whole landscape of terminations, supports, and junctions to explore for solar fuels. That’s not a promise of miracles. It’s a blueprint. Engineer the energetics, build the contact, and let the chemistry do the rest. Ran and colleagues have shown the path. The next steps are about walking it across other absorbers and into devices that can harvest sunlight at scale, using the periodic table we can afford.