A 1978 paper by N. F. Mott proposed that metal-insulator transitions (MITs) should generally be first order in crystals but can be continuous in disordered systems, and it conjectured qualitative criteria — tied to disorder strength, band-tail density-of-states exponents, and dielectric screening — for when the transition is discontinuous with a two-phase coexistence region and a critical point. The paper's stated aim was "to enquire under what conditions a discontinuous transition can occur in non-crystalline systems", and as part of that enquiry it discussed four then-unresolved issues: the apparent discontinuity in metal-rare-gas (Pb-Ar) films, the missing MIT-type critical point in expanded liquid mercury, the possible ordered non-metallic nature of condensed electron-hole phases, and the quantitative establishment of the disorder-and-screening criteria in real materials. The material-specific instances are now substantially resolved: Pb-Ar films are two-phase percolative systems rather than single-phase discontinuous Mott systems; expanded mercury undergoes a continuous band-crossing metal-nonmetal crossover, while a separate first-order electronic component near its liquid-vapor critical region remains debated; and the condensed electron-hole phases in Si and Ge are metallic liquids with measured phase diagrams, not Wigner-crystal-like non-metals. What survives as genuinely open is the general quantitative question: under what conditions, expressed through disorder strength, band-tail exponent s, screening kappa and their evolution near the transition, is a non-crystalline MIT discontinuous rather than continuous. Modern theory (Anderson-Mott scaling, Mott-Anderson mean field, DMFT with disorder, screening-tuned interaction order) supplies frameworks and partial results but no validated criterion.

Background

A metal-insulator transition (MIT) is a change of state in which a material's electrical conductivity collapses from metallic to insulating values. Two mechanisms drive it. The Mott transition is interaction-driven: strong Coulomb repulsion between electrons localizes them even in a perfect crystal. Anderson localization is disorder-driven: quantum interference between electron waves scattered from random potentials localizes the electronic states, starting at the band edges. The transition is called discontinuous (first order) when conducting and insulating phases coexist over a range of densities, so that the conductivity jumps, there is a latent heat, and a critical point terminates the two-phase coexistence line; it is called continuous when the conductivity vanishes smoothly at a critical density.

Several empirical benchmarks existed by the late 1970s. Mott had argued that a metal cannot sustain a conductivity below a minimum scale σmin⁡∼0.05 e2/(ℏa)\sigma_{\min}\sim 0.05\,e^2/(\hbar a), with aa an inter-atomic distance, so that a discontinuous transition should jump from σmin⁡\sigma_{\min} to nearly zero; a disordered insulator with a finite density of states at the Fermi level but all states localized is called a Fermi glass. In doped semiconductors such as Si:P (phosphorus-doped silicon) the observed transition was continuous but sharp, with conductivities falling below σmin⁡\sigma_{\min}. In clean crystals such as V2_2O3_3 the Mott transition was known to be first order. Photoexcited silicon and germanium showed electrons and holes condensing into liquid droplets (the electron-hole liquid), with a debated denser 'plasma' phase. Expanding a liquid metal such as mercury or cesium along its liquid-vapor coexistence line toward the critical point drives a metal-nonmetal (MNM) transition within the one-phase liquid, whose relation to the liquid-vapor critical point was unclear.

In the source paper, Mott (1978) proposed the synthesis that in crystals most MITs are expected to be first order while in disordered systems they need not be, and conjectured qualitative criteria for discontinuity tied to disorder, screening, and structure. Disorder smears the band edges into tails with a density of states N(E)=AEsN(E)=A E^s, the degree of disorder being quantified by the localized-tail weight W=∫0EcAEs dEW=\int_0^{E_c} A E^s\,dE up to the mobility edge EcE_c (equivalently, by the tail coefficient AA); combining the Coulomb correlation energy −(e2/κ)n1/3-(e^2/\kappa)n^{1/3} (with κ\kappa the dielectric constant and nn the carrier density) with the band-tail contribution, Mott's energy-balance argument gives E(n)=−(e2/κ)n1/3+Cn1/(s+1)E(n)=-(e^2/\kappa)n^{1/3}+C n^{1/(s+1)}, so that for tail exponents s>2s>2 no minimum at finite nn exists and a first-order jump into a Fermi glass is precluded. He also contrasted expanded cesium, whose liquid-vapor critical behavior appeared entangled with the electronic transition, with expanded mercury, where it did not. The paper discussed the following as open questions: (i) whether the reported sharp, apparently single-phase conductivity drop in metal-rare-gas films (notably Pb-Ar) is a genuine discontinuous Mott-like transition — a true thermodynamic discontinuity rather than percolation, i.e. transport governed by the geometry of connected metallic clusters in a two-phase mixture — given its dependence on vacuum quality and traces of other gases; (ii) whether liquid mercury lacks an MIT-type critical point because such a point lies below the melting point and is unobservable, or because disorder prevents discontinuity; (iii) whether the condensed electron-hole 'plasma' phase is an ordered, non-metallic phase — possibly Wigner-crystal-like (a lattice of electrons stabilized by Coulomb repulsion dominating kinetic energy), possibly with modulated electron and hole densities but no net charge modulation — and where its two-phase region and critical point lie; and (iv) the quantitative establishment of the conjectured disorder-and-screening criteria, including how the band-tail exponent ss, the degree of disorder, and κ\kappa (which varies only logarithmically with the gap ΔE\Delta E) actually evolve near the transition in real materials.

Problem Statement

Mott (1978) conjectured qualitative criteria — tied to disorder (through band-tail density-of-states exponents), to screening, and to structure — for when a metal-insulator transition outside a perfect crystal is discontinuous rather than continuous; the paper's stated aim was "to enquire under what conditions a discontinuous transition can occur in non-crystalline systems". The unresolved general question from that paper is: establish quantitatively the conditions under which a metal-insulator transition in a disordered or non-crystalline system is discontinuous (first order, with a two-phase coexistence region and an associated critical point) rather than continuous. Concretely: (i) how do the ingredients of the conjectured criteria — the band-tail exponent ss, the degree of disorder, and the dielectric screening constant κ\kappa — behave and couple near the transition in real systems, given the formula-level specifics of the conjecture (the band-tail form N(E)=AEsN(E)=A E^s, the energy balance E(n)=−(e2/κ)n1/3+Cn1/(s+1)E(n)=-(e^2/\kappa)n^{1/3}+C n^{1/(s+1)} with the threshold s>2s>2 for precluding a first-order jump, and the approximately logarithmic variation of κ\kappa with the gap near the transition); and (ii) what mechanism, or combination of disorder, screening, and correlation, actually determines the order of the transition in a given disordered or non-crystalline system. The quantitative question is whether, and if so under what quantitative conditions expressed through the band-tail exponent ss, the degree of disorder, and the screening constant κ\kappa, the order of the transition is fixed by these ingredients: whether a quantitative criterion for the order of the transition — following from the mechanism identified in (ii) and making a falsifiable prediction for at least one disordered or non-crystalline system outside the material classes represented by the limiting behaviors listed below — exists, or whether no such criterion fixed by these ingredients alone exists. It must address both (i), for a specified class of real systems, and (ii), which are separate and jointly required. For orientation, the post-1978 experimental characterizations — continuous transitions in doped semiconductors such as Si:P with conductivities below σmin⁡\sigma_{\min}, first-order transitions in clean crystals such as V2_2O3_3 and the quasi-two-dimensional organic conductors, the continuous band-crossing crossover in expanded fluid mercury (valence and conduction bands overlapping gradually as the liquid expands, with a first-order electronic component near the liquid-vapor critical region still debated), and purely percolative behavior in Pb-Ar films — are known limiting behaviors that any proposed criterion must be consistent with. The material-specific instances posed in the source (Pb-Ar films, the nature of condensed electron-hole phases) have since been resolved experimentally; the surviving open core is this general quantitative criterion, which the later literature supplies frameworks for but does not answer.

References

  1. N. F. Mott, "Continuous and discontinuous metal-insulator transitions", Philosophical Magazine B 37, 377-386 (1978). DOI 10.1080/01418637808227677, https://doi.org/10.1080/01418637808227677 (source paper)
  2. N. F. Mott, "Review lecture: Metal–insulator transitions", Proceedings of the Royal Society of London A 382, 1-24 (1982). DOI 10.1098/rspa.1982.0086, https://doi.org/10.1098/rspa.1982.0086
  3. R. V. Aldridge, J. Quinn, S. J. Raeburn, "An investigation of the atomic arrangements in Pb-Ar films formed by codeposition at 4K", Journal of Physics C: Solid State Physics 12, 3621-3633 (1979). DOI 10.1088/0022-3719/12/17/032, https://doi.org/10.1088/0022-3719/12/17/032
  4. H. Nakajima, R. Hanawa, K. Sajiki, "Metal-Nonmetal Transition and Annealing Effects on Electrical Conduction in the Mercury-Argon System", Progress of Theoretical Physics Supplement 72, 150-155 (1982). DOI 10.1143/ptps.72.150, https://doi.org/10.1143/ptps.72.150
  5. J. R. Franz, "Metal-Nonmetal Transition in Expanded Liquid Mercury", Physical Review Letters 57, 889-892 (1986). DOI 10.1103/PhysRevLett.57.889, https://doi.org/10.1103/PhysRevLett.57.889
  6. S. Jüngst, B. Knuth, F. Hensel, "Observation of Singular Diameters in the Coexistence Curves of Metals", Physical Review Letters 55, 2160-2163 (1985). DOI 10.1103/PhysRevLett.55.2160, https://doi.org/10.1103/PhysRevLett.55.2160
  7. G. Kresse, J. Hafner, "Ab initio simulation of the metal/nonmetal transition in expanded fluid mercury", Physical Review B 55, 7539-7548 (1997). DOI 10.1103/PhysRevB.55.7539, https://doi.org/10.1103/PhysRevB.55.7539
  8. F. Hensel, M. Stolz, G. Hohl, R. Winter, W. Götzlaff, "Critical phenomena and the metal-nonmetal transition in liquid metals", Journal de Physique IV 1, C5-191-C5-205 (1991). DOI 10.1051/jp4:1991523, https://doi.org/10.1051/jp4:1991523
  9. F. Hensel, W. W. Warren, "Fluid Metals: The Liquid-Vapor Transition of Metals", Princeton University Press (1999). DOI 10.1515/9781400865000, https://doi.org/10.1515/9781400865000
  10. F. Hensel, "The Metal–Nonmetal Transition in Fluid Mercury: Landau–Zeldovich Revisited", in "Metal-to-Nonmetal Transitions", Springer Series in Materials Science, pp. 23-35 (2010). DOI 10.1007/978-3-642-03953-9_2, https://doi.org/10.1007/978-3-642-03953-9_2
  11. L. Calderin, L. E. González, D. J. González, "Expanded fluid mercury in the metal-nonmetal transition range. An ab-initio MD study", The European Physical Journal Special Topics 196, 27-34 (2011). DOI 10.1140/epjst/e2011-01415-2, https://doi.org/10.1140/epjst/e2011-01415-2
  12. L. J. Schowalter, F. M. Steranka, M. B. Salamon, J. P. Wolfe, "Separate Mott and liquid-gas transitions in a coulomb gas: Photoexcited germanium", Solid State Communications 44, 795-799 (1982). DOI 10.1016/0038-1098(82)90276-9, https://doi.org/10.1016/0038-1098(82)90276-9
  13. L. M. Smith, J. P. Wolfe, "Time-resolved study of electron-hole plasmas near the liquid-gas critical point in Si: Evidence for a second condensed phase", Physical Review B 51, 7521-7543 (1995). DOI 10.1103/PhysRevB.51.7521, https://doi.org/10.1103/PhysRevB.51.7521
  14. M. A. Paalanen, T. F. Rosenbaum, G. A. Thomas, R. N. Bhatt, "Stress Tuning of the Metal-Insulator Transition at Millikelvin Temperatures", Physical Review Letters 48, 1284-1287 (1982). DOI 10.1103/PhysRevLett.48.1284, https://doi.org/10.1103/PhysRevLett.48.1284
  15. T. F. Rosenbaum, R. F. Milligan, M. A. Paalanen, G. A. Thomas, R. N. Bhatt, W. Lin, "Metal-insulator transition in a doped semiconductor", Physical Review B 27, 7509-7523 (1983). DOI 10.1103/PhysRevB.27.7509, https://doi.org/10.1103/PhysRevB.27.7509
  16. P. Limelette, A. Georges, D. Jérôme, P. Wzietek, P. Metcalf, J. M. Honig, "Universality and Critical Behavior at the Mott Transition", Science 302, 89-92 (2003). DOI 10.1126/science.1088386, https://doi.org/10.1126/science.1088386
  17. F. Kagawa, K. Miyagawa, K. Kanoda, "Unconventional critical behaviour in a quasi-two-dimensional organic conductor", Nature 436, 534-537 (2005). DOI 10.1038/nature03806, https://doi.org/10.1038/nature03806
  18. T. Furukawa, K. Miyagawa, H. Taniguchi, R. Kato, K. Kanoda, "Quantum criticality of Mott transition in organic materials", Nature Physics 11, 221-224 (2015). DOI 10.1038/nphys3235, https://doi.org/10.1038/nphys3235
  19. D. Belitz, T. R. Kirkpatrick, "The Anderson-Mott transition", Reviews of Modern Physics 66, 261-380 (1994). DOI 10.1103/RevModPhys.66.261, https://doi.org/10.1103/RevModPhys.66.261
  20. V. Dobrosavljević, G. Kotliar, "Mean Field Theory of the Mott-Anderson Transition", Physical Review Letters 78, 3943-3946 (1997). DOI 10.1103/PhysRevLett.78.3943, https://doi.org/10.1103/PhysRevLett.78.3943
  21. K. Byczuk, W. Hofstetter, D. Vollhardt, "Mott-Hubbard Transition versus Anderson Localization in Correlated Electron Systems with Disorder", Physical Review Letters 94, 056404 (2005). DOI 10.1103/PhysRevLett.94.056404, https://doi.org/10.1103/PhysRevLett.94.056404
  22. M. Imada, A. Fujimori, Y. Tokura, "Metal-insulator transitions", Reviews of Modern Physics 70, 1039-1263 (1998). DOI 10.1103/RevModPhys.70.1039, https://doi.org/10.1103/RevModPhys.70.1039
  23. M. Schüler, E. G. C. P. van Loon, M. I. Katsnelson, T. O. Wehling, "First-order metal-insulator transitions in the extended Hubbard model due to self-consistent screening of the effective interaction", Physical Review B 97, 165135 (2018). DOI 10.1103/PhysRevB.97.165135, https://doi.org/10.1103/PhysRevB.97.165135
  24. I. Z. Kostadinov, B. R. Patton, "Second Critical Point in First Order Metal-Insulator Transitions", Physical Review Letters 101, 226407 (2008). DOI 10.1103/PhysRevLett.101.226407, https://doi.org/10.1103/PhysRevLett.101.226407
  25. P. P. Poduval, S. Das Sarma, "Anderson localization in doped semiconductors", Physical Review B 107, 174204 (2023). DOI 10.1103/PhysRevB.107.174204, https://doi.org/10.1103/PhysRevB.107.174204
  26. S. V. Kravchenko, M. P. Sarachik, "Metal–insulator transition in two-dimensional electron systems", Reports on Progress in Physics 67, 1-44 (2004). DOI 10.1088/0034-4885/67/1/R01, https://doi.org/10.1088/0034-4885/67/1/R01