The Periodic Table Just Reopened: 2D Metals and the New Age of Matter

In March 2025, a team led by Guangyu Zhang at the Institute of Physics, Chinese Academy of Sciences, published a paper in Nature that won Physics World's Breakthrough of the Year. The title was understated: "Realization of 2D metals at the angstrom thickness limit."

The result was not.

They had created sheets of metal - bismuth, tin, lead, indium, gallium - that were a single atom thick. Not thin films. Not nanoparticles. Monolayers. True two-dimensional metals at the angstrom scale, with properties that do not exist in bulk metal and were widely assumed to be physically impossible.

This is not an incremental advance. It reopens the periodic table.

Why this was supposed to be impossible

To understand why this matters, you need to understand why nobody expected it to work.

Graphene - the 2D carbon sheet that won Geim and Novoselov their Nobel Prize in 2010 - was possible because graphite has a natural layered structure. Strong covalent bonds hold each carbon sheet together in-plane. Weak van der Waals forces hold the sheets together. You can literally peel them apart with tape. The "weak plane" is built into the crystal structure.

Metals have no weak plane. Metallic bonding is isotropic - equally strong in all three dimensions. There is no direction you can pull that preferentially separates one layer from the rest. The atoms are close-packed, like stacked billiard balls, with strong bonds in every direction.

The thermodynamics are also hostile. A freestanding 2D metal is energetically unfavorable - it wants to ball up into a 3D cluster to minimize surface energy. And even if you could force metal into a monolayer, it would oxidize instantly in air.

For two decades after graphene, the assumption was clear: you can make 2D semiconductors (MoS2, WSe2), 2D insulators (hexagonal boron nitride), and 2D semimetals (graphene). But 2D metals? The physics says no.

Zhang's team did not accept this.

Van der Waals squeezing: the atomic anvil press

The method is beautifully direct. It is, at its core, a very sophisticated anvil.

Two sapphire substrates are each coated with a single-crystalline monolayer of MoS2, grown epitaxially. These surfaces are atomically flat, chemically inert (no dangling bonds to react with the metal), and extremely rigid - sapphire has a Young's modulus of 430 GPa, MoS2 exceeds 300 GPa. They are the hardest, flattest, most chemically boring surfaces you can build.

Pure metal powder is placed between the two MoS2/sapphire anvils. The assembly is heated until the metal melts into a liquid droplet. Then roughly 200 MPa of pressure is applied - about double the pressure at the bottom of the Mariana Trench - squeezing the liquid metal between the two atomically flat surfaces.

The system is cooled under pressure. The metal solidifies in the confined space. What remains is a monolayer of metal, fully encapsulated between two MoS2 sheets that protect it from oxidation.

The samples are stable in air for over a year. By adjusting pressure, the team can produce monolayer, bilayer, or trilayer metals with atomic precision. The method is described as universal - applicable to any metal that can be melted.

Zhang spent a decade getting here. He started by literally trying to hammer metal thin. What he ended up building was an atomic-precision manufacturing technique that forces 3D materials into 2D, overriding their thermodynamic preferences through confinement.

What happens when you make metal two-dimensional

When you confine electrons to two dimensions, quantum mechanics takes over. The properties of 2D metals are not just "thinner versions" of bulk metals. They are fundamentally different materials.

Monolayer bismuth shows enhanced electrical conductivity, a notable field effect (the ability to tune conductivity with a gate voltage - something bulk metals cannot do because they have too many free carriers), a new phonon mode that does not exist in bulk bismuth, and a large nonlinear Hall conductivity at room temperature driven by Berry curvature effects.

The field effect point deserves emphasis. Bulk metals cannot be gated. Their sea of free electrons drowns out any external field. At the monolayer limit, carrier density drops enough that an electric field can modulate conductivity. This opens the door to metal-based transistors - a concept that did not exist before this work.

Here is what each of the five metals becomes at the 2D limit:

Three of these - bismuth, tin, and lead - become topological insulators. Their bulk becomes insulating, but their edges conduct electricity with zero dissipation, protected by the fundamental symmetries of quantum mechanics. This is not a theoretical prediction. Bismuthene on SiC was demonstrated as a quantum spin Hall insulator in 2017 by the Wurzburg group, with a massive 0.8 eV gap - twenty-six times larger than the previous best (HgTe at 30 meV).

The 0.8 eV gap matters enormously. It means these topological states survive at room temperature. Previous topological materials required cryogenic cooling to function. Bismuthene does not. Lead monolayers push this even further, with gaps reaching 1.34 eV when chemically decorated.

Gallium goes in a different direction: it becomes superconducting, with a critical temperature exceeding bulk gallium. Quantum confinement enhances the electron-phonon coupling that drives superconductivity.

Why this changes electronics

The semiconductor industry is running out of road.

TSMC's N2 process (2 nm gate length) entered production in late 2025. At this scale, the channel between source and drain is roughly 10 atoms wide. Making it smaller runs into quantum tunneling - electrons leak through the barrier regardless of gate voltage. This is not an engineering problem. It is a physics wall.

The industry's roadmap for the post-silicon era explicitly includes 2D materials. Samsung, TSMC, Intel, and IMEC have all incorporated 2D semiconductor transistors into their scaling plans for the mid-2030s and beyond. Imec demonstrated functional stacked nanosheet FETs with monolayer MoS2 channels at the 2024 IEDM conference.

But 2D transistors need 2D contacts. You cannot attach a bulk metal electrode to an atomically thin semiconductor channel without interface problems - Schottky barriers, Fermi level pinning, orbital mismatch. The contact resistance dominates device performance at these scales.

2D metals solve this. Atomically thin metal contacts can interface cleanly with atomically thin semiconductor channels. The entire device - channel, contacts, electrodes - operates at the same dimensional scale. This is not a marginal improvement. It is the missing piece for 2D electronics.

CDimension, an MIT spinout, is already producing wafer-scale (8" and 12") uniform single-layer 2D materials for semiconductor fabs. Seoul National University published a full roadmap for 2D gate stack technology in Nature Electronics in September 2025. The infrastructure is being built.

The connection to quantum computing

Lead monolayers on cobalt-silicon islands on silicon substrates show evidence of 2D topological superconductivity with dispersive in-gap edge states. These are candidate platforms for Majorana fermions - exotic quasiparticles that are their own antiparticles and could serve as the basis for topological quantum computing.

Topological qubits are the holy grail of quantum computing because they are inherently error-resistant. The quantum information is stored in the topology of the system, not in any local property that can be disturbed by noise. Microsoft has bet their entire quantum program on this approach. The bottleneck has been finding a material platform that reliably hosts Majorana fermions.

2D metals - specifically lead monolayers with strong Rashba spin-orbit coupling on superconducting substrates - are among the most promising candidates. Bismuth ultrathin films on NbSe2 show coexistence of topological edge states and proximity-induced superconductivity. These are not theoretical predictions. These are measured experimental results.

The connection to neuromorphic computing

I have written extensively about neuromorphic computing and spiking neural networks as the next paradigm in AI hardware. 2D metals connect to this field directly.

Memristors - devices that mimic biological synapses by changing resistance based on history - are fundamental building blocks of neuromorphic hardware. The best memristors use 2D material stacks: a switching layer (MoS2, MoTe2) sandwiched between metal electrodes. At the monolayer limit, these devices become small enough to stack in three dimensions, approaching the density of biological neural tissue.

2D metals serve as the electrode layers in these stacks. Their atomically thin profile enables vertical integration - thousands of memristive layers, each only a few atoms thick, stacked on a single chip. The contact between 2D metal electrodes and 2D switching materials is clean at the atomic level, reducing variability and improving the analog precision that neuromorphic computing requires.

Gallium in particular is interesting here: its ability to induce semiconducting-to-metallic phase transitions in adjacent 2D semiconductors creates a natural switching mechanism that could form the basis for new synaptic devices.

The broader pattern: from discovery to design

Graphene was found by accident. Andre Geim and Konstantin Novoselov were experimenting with Scotch tape on a Friday afternoon. They did not set out to create a new material. They stumbled into one.

2D metals were engineered. Zhang spent a decade developing a deliberate, controllable manufacturing process. The van der Waals squeezing method is not a lucky break. It is a universal fabrication technique applicable to any metal in the periodic table, plus alloys and amorphous compounds.

This distinction matters. It marks a civilizational transition from discovering materials to designing them.

The periodic table has roughly 90 metals. Each now has a potential 2D form with properties that do not exist in bulk. Many of these properties - topological states, enhanced superconductivity, field-effect tunability - are exactly what the next generation of electronics, quantum computers, and neuromorphic hardware needs.

A follow-up paper in 2025 (Tran, Wu, Chen et al. in npj 2D Materials and Applications) extended the concept further: "van der Waals injection-molded crystals" - molten metal injected between van der Waals layers within shaped molds, producing single crystals of bismuth, tin, and indium in arbitrary geometries: Hall bars, rings, nanowires, with controlled thickness from 10-100 nm.

We are not just making new materials. We are learning to shape matter at the atomic level into whatever geometry we need.

Timeline and honest assessment

Geim himself has cautioned that moving from basic research to consumer products takes "at least one generation." Graphene was isolated in 2004 and is only now, 22 years later, reaching early commercial applications. 2D metals benefit from graphene's infrastructure - the fabs, characterization tools, and integration knowledge all carry over - but they face additional manufacturing challenges. Van der Waals squeezing is not yet wafer-scale compatible.

My honest timeline:

• 2025-2028: Fundamental physics exploration. More metals, alloys, amorphous 2D materials. Expanding the catalog of properties.
• 2028-2035: Integration with semiconductor manufacturing. 2D metal contacts for 2D transistors. Early quantum device prototypes using topological 2D metal platforms.
• 2035+: Full 2D electronics - channels, contacts, electrodes, and dielectrics all at the atomic layer. Topological quantum computing on 2D metal platforms. Room-temperature topological electronics using bismuthene and plumbene.

This is a 10-15 year horizon for commercial impact. That is long by startup standards and short by materials science standards. For context, the transistor was invented in 1947 and the integrated circuit in 1958 - eleven years. The gap between Zhang's 2025 paper and commercial 2D metal devices in the mid-2030s is comparable.

Why I am writing about this

I write about technologies that change the substrate on which civilization operates. Energy is one substrate - I have spent thousands of words on thorium and geothermal. Computation is another. Materials are the deepest substrate of all, because both energy systems and computing systems are ultimately constrained by what materials can do.

For 60 years, the answer to "how do we build better electronics" has been "make silicon smaller." That road is ending. The next answer will not come from a faster process node. It will come from a different material platform - one where quantum mechanics is not the enemy to be fought but the property to be exploited.

2D metals are the beginning of that platform. Not the end - there is a decade of engineering ahead. But the physics is no longer in question. The atoms can be made flat. Their quantum properties can be measured. Their integration path with existing semiconductor infrastructure is being mapped.

The periodic table is not a list of known elements. It is a list of building blocks whose 2D forms we are only now learning to create. Ninety metals, each with quantum properties at the monolayer limit that do not exist in bulk. A universal fabrication method to create them. An industry roadmap that needs exactly what they provide.

The age of discovering materials is ending. The age of designing them has begun.