Fulleride superconductors: Jahn–Teller coupling, Mott physics and driven states

Scope

Alkali-doped fullerides combine molecular vibrations, narrow electronic bands and strong local interactions. Ideal A₃C₆₀ has three donated electrons in a molecular t₁ᵤ manifold with six spin-orbital states: half-filling describes electron counting, not a guarantee of metallic transport or superconductivity. The same nominal Cs₃C₆₀ composition supports fcc and A15 packings with different magnetic and pressure-dependent superconducting behavior. Specify composition, structure, preparation and measurement conditions before comparing temperatures. Pressure-induced equilibrium superconductivity and optically driven responses are separate evidence classes. R5 R4 R7

Nominal composition is not a sample assay. Ganin's fcc-rich powder had refined composition Cs₂.₉₀₁₍₆₎C₆₀ and an fcc fraction of about 86%, with minority phases. Its ambient-pressure magnetic insulating behavior illustrates why three donated electrons alone cannot define the phase. These preparation and phase-fraction qualifications accompany the idealized A₃C₆₀ description. R4

The 1991 potassium discovery already illustrates why one temperature is insufficient: the abstract reports a thin-film resistive onset near 16 K and essentially vanishing resistance by 5 K, while bulk magnetic and microwave measurements indicate an 18 K transition. The subsequent rubidium report gives a maximum of 28 K from bulk magnetic and microwave evidence. These abstract-level checks establish historical observations; they do not constitute a complete reassessment of the original specimens. R1 R2

Mathematical assumptions

One influential model represents each molecule by three effective p-like orbitals with local terms $H_U=U(n-3)^2/2$ and $H_J=J[2\mathbf S^2+\mathbf L^2/2]+5J(n-3)^2/6$. Angular momenta are dimensionless. In this convention ordinary Hund exchange has $J=-J_H<0$; integrating out Jahn–Teller vibrations can instead give $J=-J_H+J_{JT}>0$. This static reduction requires the coherent electronic scale $ZW$ to lie below the relevant phonon energy. It should not be applied across the whole bandwidth range without checking retardation. R5

At exactly three electrons, the model multiplets $^4S$, $^2D$ and $^2P$ have energies $15J/2$, $9J/2$ and $5J/2$, with respective degeneracies 4, 10 and 6. They account for $\binom{6}{3}=20$ local states. Positive effective $J$ favors the low-spin $^2P$ multiplet; negative $J$ favors $^4S$. This is an algebraic check of the effective local Hamiltonian, not a measured molecular spectrum or proof of superconducting order. R5

The heuristic Cooper-channel amplitude $A^*\simeq ZU-10J/3$ illustrates how correlations can reduce repulsion relative to an inverted exchange. For illustrative fixed ratios $U/W=0.5$ and $J/W=0.05$, its sign changes at $Z=1/3$. Simultaneously the coherent bandwidth is $ZW$: decreasing $Z$ does not supply an unlimited transition-temperature gain. The published argument explicitly allows deviations near the Mott boundary; its underlying DMFT model uses an infinite-coordination Bethe lattice. No self-consistent DMFT calculation or measured transition follows from this toy sweep. R5

Experimental criteria

For fcc Cs₃C₆₀, Ganin and colleagues report a pressure-dependent magnetic onset: 25.6 K at 1.3 kbar has only about 0.1% shielding, whereas 29.7 K at 2.2 kbar has about 12%. The maximum is approximately 35 K near 7 kbar, equivalent to 0.7 GPa. Measurements use 20 Oe, zero-field-cooled/field-cooled protocols and a tin pressure reference. A15 Cs₃C₆₀ reaches a separately reported maximum near 38 K; its value must not be assigned to the fcc sample. R3 R4

The fcc study also separates magnetic ordering near 2.2 K from the A15 ordering scale near 46 K. Neither is a superconducting transition. Multiphase powders and small shielding fractions require particular care when locating the first superconducting signal. The magnetometry pressure range ends near 10.6 kbar; the separate diffraction measurements extend to much higher pressures. Combining those ranges would falsely imply that superconductivity was measured throughout the structural experiment. R4

Published online on 9 June 2026, Shchukin and colleagues tune Rb uptake in nanostructured C₆₀ films. Figure 2f labels Rb₂.₇₁C₆₀ with $T_{\mathrm{onset}}=22.4$ K and $T_{\mathrm{zero}}=7.0$ K, and Rb₂.₉₈C₆₀ with 26.4 K and 14.5 K. The former criterion is a local resistivity maximum; the latter uses $d\rho/dT<5\times10^{-4}$ mΩ cm K⁻¹ in the plotted low-temperature window. It is a stated operational criterion, not an instrument-independent definition of exact zero resistance. R8

That experiment uses Raman analysis with relative scattering cross sections to estimate composition, and van der Pauw transport. The 10 μA intercalation current differs from the 1 μA resistance-probe current. Stoichiometry, grain connectivity, thickness and transition criterion must accompany any comparison with bulk rubidium fullerides. The supplement and underlying transport datasets have not been independently reanalyzed here. R8

Limits and disputes

Mitrano's optical study and Budden's longer-lived driven response concern pumped K₃C₆₀. A finite-frequency inductive response and suppressed transient resistance constrain the nonequilibrium state, but do not by themselves establish an equilibrium thermodynamic transition. Budden's final Figure 4 separates two measurements at a 100 K base temperature and approximately 25 mJ cm⁻² pump fluence:

The electrical readout uses a 1 ns voltage probe pulse. The corrected resistance is compatible with zero within the stated calibration uncertainty, and recovers on a tens-of-nanoseconds timescale. This supports the authors' assignment of a metastable superconducting state; it does not turn the 100 K measurement condition into an equilibrium Tc. Neither the pump duration nor the probe duration is the state lifetime. Both records retain transient-response temperatures with no equilibrium Tc. Complete supplement review, raw-data refitting and comparison of alternative conductivity models remain open. R6 R7

The historical hole-doped C₆₀ claim at 52 K was retracted on 6 March 2003. The original paper and its notice are retained together for traceability; the withdrawn value is excluded from experimental transition records. The notice identifies concerns about device-measurement data following an independent investigation. A citation in an older, otherwise valid paper does not restore the withdrawn claim. R9 R10

Discovery implications

Fullerides motivate simultaneous control of bandwidth, local vibronic interactions, molecular valence and disorder. A superconducting dome near a Mott state is compatible with cooperation between phonons and correlations; it does not uniquely determine the pairing kernel. Neither a fully open gap nor proximity to a magnetic insulator proves that one interaction acts alone. Report model assumptions beside the experimental evidence used to constrain them. R4 R5

Current-driven intercalation offers a method for tuning samples rather than a new bulk temperature record. Reproducible optimization needs composition maps, measurement-current controls and comparable transition definitions. Keep pressure polymorphs and optically driven states separate when selecting targets: their conditions cannot be collapsed into a single family-wide Tc ranking. R8 R4 R7

Related topics

Continue with electron–phonon pairing, Eliashberg theory, Mottness, s-wave pairing and zero resistance and driven-state evidence.

Material cases

Values describe individual reports, not a ranking. Unknown means the checked source did not supply the field at this review scope.