REVIEW 2 major objections 4 minor 17 references
Surface states in defect-free polyatomic lattices described by a tight-binding model
T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read Tamm-type surface states appear in defect-free polyatomic lattices once the unit cell has at least three atoms, with the effect entering at second order in the hopping.
desk verdict The paper's core claim is too broad: with alternating hoppings, a binary chain already has an exact Tamm-type surface state, so the 'minimum basis three' result requires a uniform-hopping assumption. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the second-order effective Hamiltonian obtained by degenerate perturbation theory on the tight-binding model $\hat H = \hat H_0 + t\hat V$. Starting from the $L$-fold degenerate subspace of sites with on-site energy $\epsilon_1$, the second-order correction produces an $(M+1)\times(M+1)$ matrix $H$ whose diagonal entries give the renormalized surface and bulk energies and whose off-diagonal entries $H_{m',m'\pm 2/b}$ are nonzero only for $b=2$; for $b>2$ they vanish, leaving the two surface sites with distinct energy shifts. This matrix, defined by Eqs. (7)-(8), is what carries the argument: its zero off-diagonal structure for $b>2$ is exactly why surface states can split off from the band.
What would settle it
Diagonalize the tight-binding Hamiltonian exactly for a finite binary chain with two atoms per unit cell (sites alternating A and B, both ends A), with $\epsilon_A \neq \epsilon_B$ and uniform hopping $t$, and check whether any eigenstate has probability density decaying exponentially away from an end. The paper's claim predicts no such surface states for $b=2$; finding one would disprove the lower bound of three atoms per cell.
Extended reading notes
Core claim
The central claim is that in a finite tight-binding chain with $b$ atoms per unit cell and nearest-neighbor hopping, Tamm-type surface states exist without any surface defect if and only if $b \ge 3$, provided the on-site energy $\epsilon_1$ of the surface atom differs from the other species in the cell. At second order in the hopping $t$, the $L$ degenerate states built from the $\epsilon_1$ sites split into two surface levels with energies $E_1 \simeq \epsilon_1 - \alpha_1^2 t^2/(\epsilon_2-\epsilon_1)$ and $E_N \simeq \epsilon_1 - \alpha_b^2 t^2/(\epsilon_b-\epsilon_1)$, separated from the remaining bulk levels $E \simeq \epsilon_1 - (\alpha_b^2/(\epsilon_b-\epsilon_1)+\alpha_1^2/(\epsilon_2-\epsilon_1)) t^2$. For $b=2$, off-diagonal matrix elements of the effective Hamiltonian hybridize the bulk levels, and no surface states appear; for $b>2$ the off-diagonal elements vanish, so the surface states split off. In two dimensions, the same condition yields fourfold-degenerate corner states with energy shift $-2t^2/(\epsilon_B-\epsilon_A)$ and edge states with shift $-3t^2/(\epsilon_B-\epsilon_A)$, consistent with the different coordination numbers of corners, edges, and bulk.
Load-bearing premise
The argument assumes the finite chain is terminated on the special surface species at both ends, so both end sites carry the same on-site energy $\epsilon_1$; if the chain ends on a different atom, the surface-state criterion may change or disappear.
Editorial extensions
If this is right
- Any one-dimensional polyatomic chain with at least three atoms per unit cell, terminated on the low-energy species, will show two exponentially localized end states even though the lattice is periodic and defect-free.
- The surface-state energies are captured quantitatively by the second-order formulas and vary as $t^2$, so the states persist at weak hopping and can be predicted before full diagonalization.
- In two-dimensional arrays, the same three-atom-per-direction criterion produces two distinct classes of surface localization, corner states and edge states, with different energies set by coordination number.
- The mechanism works for both electrons (tight-binding solids) and light (photonic crystals and waveguide arrays) because it relies only on the structure of the hopping and on-site energies.
- For a two-atom basis there are no surface states in this model, so the transition from $b=2$ to $b=3$ is sharp and testable.
Reading between the lines
- The paper's perturbative picture is built on chains terminated at both ends on the special species; for other terminations the surface-state condition may change, so the abstract's unqualified statement likely needs that extra condition.
- The same coordination-number argument suggests three-dimensional lattices will show face, edge, and corner states with energy shifts set by the number of missing nearest neighbors, a case the paper only conjectures.
- One could test the mechanism directly in coupled waveguide arrays with three waveguides per cell: the two split-off modes should appear at the second-order energies and remain localized at the ends over long propagation distances.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies single-particle tight-binding chains and lattices with a periodic polyatomic basis, terminated on the A species. It claims a lower bound of three atoms per unit cell for Tamm-type surface states under nearest-neighbor hopping, provided the local energy of the surface atom differs from the rest of the unit cell. The argument uses degenerate second-order perturbation theory and numerical diagonalization for ABA (b=2) and ABBA (b=3) chains, and it asserts an extension to two dimensions with corner and edge surface states. The paper presents no fitted parameters and reports numerical spectra that match the second-order energies for the specific equal-hopping chains shown.
Significance. If the claimed lower bound were true for general nearest-neighbor tight-binding models, it would be a simple and broadly applicable design principle for defect-free surface localization. The paper is self-contained, and the equal-hopping examples in one dimension are supported by a transparent perturbative calculation and independent numerical diagonalization. However, the central claim as stated is false: a defect-free binary chain with alternating hoppings and A-terminated ends supports an exact Tamm-type surface state at E=epsilon_A, contradicting the asserted minimum basis size of three. The two-dimensional section is also asserted without derivation or numerical verification. The significance of the reported mechanism is therefore limited to a uniform-hopping special case unless the claims are substantially revised.
major comments (2)
- [Abstract and §3, after Eq. (11)] The central claim that no surface states exist for a two-atom basis under nearest-neighbor hopping is false for the general model defined by Eqs. (1)-(3). Consider a b=2 chain with on-site energies epsilon_A and epsilon_B, alternating hoppings t1 and t2, and A atoms at both ends as in the paper. The state with amplitudes only on A sites, a_m = (-t1/t2)^{m-1}, and zero B-site amplitudes has energy epsilon_A and satisfies every Schrödinger equation, because the B-site equation reduces to -t1 a_m - t2 a_{m+1} = 0. For |t1|<|t2| this state decays exponentially from the left A surface, is normalizable in the thermodynamic limit, and lies in the band gap for epsilon_A != epsilon_B. This is exactly a Tamm-type surface state in a periodic defect-free lattice with basis size two. The same conclusion already follows from the paper's own second-order effective Hamiltonian: Eqs. (7)-(8) give an end-site potential step of t^2 alpha_2^2/(epsilon_2-epsilon_1) relative to the bulk and a hopping t^2 alpha_1 alpha_2/(epsilon_2-epsilon_1), which supports a bound state when alpha_2 > alpha_1. Thus the uniform-hopping assumption used in the figures, but absent from the abstract and from Eq. (3), is load-bearing; the claimed lower bound of three is incorrect for general nearest-neighbor hopping.
- [§4, 'Surface states in two-dimensions'] The two-dimensional section consists entirely of assertions: the corner-state energy E_corner, the edge-state energy E_edge, and the four-fold degeneracies are stated without a derivation, without a definition of the finite two-dimensional lattice and its termination, and without numerical spectra. Furthermore, the claimed degeneracy of the edge states is ambiguous and likely incorrect for a finite square lattice, where each of the four edges hosts multiple A sites and should give more than one edge state per edge when L > 3. Since the abstract and conclusion advertise 'other kinds of surface states' identified in two dimensions, this section needs a real perturbative calculation or numerical confirmation before the claims can be accepted.
minor comments (4)
- [Figure 4 caption and text] The curve labels in Figure 4 do not match Eqs. (9)-(11). For epsilon_A=-2, epsilon_B=0 and alpha_r=1, Eq. (9) gives E = epsilon_A + t^2/epsilon_A, while Eq. (11) gives E = epsilon_A + 2t^2/epsilon_A; the label 'E = epsilon_A + t^2/(2 epsilon_A)' corresponds to neither. The labels should be corrected so the comparison to the perturbation formulas is unambiguous.
- [Figure 3 caption] The caption of Figure 3 reads 'Energy spectrum of the binary BH chain with two bosons vs eigenstate index,' which describes neither the two-dimensional lattice geometry shown in the figure nor the content of the paper; it appears to be a copy-paste error and must be replaced.
- [Throughout] The phrase 'degenerated perturbation theory' should be 'degenerate perturbation theory,' and the reference to 'Shokley' should be 'Shockley.'
- [Abstract and Introduction] The termination condition is not stated: the mechanism relies on both ends of the chain being A atoms, and the abstract should explicitly say that the lattice is terminated on the surface species A. As written, the condition 'provided the local energy of the surface atom is different from the rest' is insufficient without specifying which species terminates the finite lattice.
Circularity Check
No circularity: the surface-state energies and the minimum-basis criterion are derived from the stated tight-binding Hamiltonian, with numerical diagonalization as an independent check.
full rationale
The paper's derivation is self-contained. The central results, Eqs. (9)-(11), follow from degenerate perturbation theory applied directly to the tight-binding Hamiltonian (1)-(3), with no fitted parameters and no quantity defined in terms of the target result. The predicted surface-state energies are then compared with numerical diagonalization of the same Hamiltonian (Fig. 4), which is an independent check rather than a renamed input. The claimed lower bound of three atoms per unit cell is obtained from the structure of the second-order effective Hamiltonian: for b > 2 the off-diagonal elements vanish and the surface-site diagonal elements differ from the bulk-site diagonal elements, whereas for b = 2 the off-diagonal elements lift the degeneracy. No circular step is present. The only self-citation, reference [17], is used as an analogy ('These results are similar to those obtained by Pinto et al'), not as load-bearing justification for the present derivation. The limitation that the proof assumes the chain is terminated on the special species and that the general-hopping case may admit exceptions is a correctness or scope concern, not a circularity concern, and therefore does not raise the circularity score.
Assumptions & free parameters
assumptions (4)
- domain assumption The system is described by a nearest-neighbor tight-binding Hamiltonian with one orbital per site and no longer-range hopping.
- domain assumption The on-site energy of the A species differs from all other species in the unit cell, epsilon_{r != 1} != epsilon_1.
- domain assumption The finite chain is terminated on the A species at both ends.
- ad hoc to paper For the two-dimensional extension, the second-order self-energy counting by coordination number remains valid.
Cite this review
Pith. "Pith review of Surface states in defect-free polyatomic lattices described by a tight-binding model." pith.science (2026). https://pith.science/paper/K3LZIS4P
@misc{pith2026190803976,
author = {Pith},
title = {Pith review of: Surface states in defect-free polyatomic lattices described by a tight-binding model},
year = {2026},
howpublished = {\url{https://pith.science/paper/K3LZIS4P}},
note = {Machine review of arXiv:1908.03976}
}
read the original abstract
We report about a mechanism for surface localization, present in finite defect-free polyatomic lattices described by a tight binding model. Numerical diagonalization and degenerated perturbation theory show that there is a minimum number of atoms within each unit cell in the lattice for which surface states may exist, provided the local energy of the surface atom is different from the rest in the unit cell. It is shown that the appearance of surface states is a second-order effect in the hopping parameter. Other kinds of surface states are identified in the two-dimensional case.
Figures
Figures from the paper (1 more)
Reference graph
Works this paper leans on
- [1]
-
[2]
R. H. Fowler, Proc. R. Soc. Londno, Serv. A 141, 56 (1933); S. Rijanow, Z. Phys. 89, 806 (1934)
work page 1933
-
[3]
E. T. Goodwin, Proc. Camb. Phil. Soc. 35, 205 (1939); 35, 221 (1939); 35, 232 (1939)
work page 1939
-
[4]
S. G. Davison and J. Grindlay, Surf. Sci. 11, 99 (1968); S. G. Davison, T. Y. Ling, and U. S. Ghosh, J. Phys. Chem. Solids 28, 1921 (1967)
work page 1968
- [5]
-
[6]
Zak, Phys
J. Zak, Phys. Rev. B 32, 2218 (1985)
1985
-
[7]
E-Ni Foo, How-Sen Wong, Phys. Rev. B 9, 1857 (1974)
work page 1974
- [8]
Show all 17 references
-
[9]
J. D. Johannopoulus, R. D. Meade, J. N. Winn, Photonic crystals, molding the flow of light , (Lawrenceville, NJ, 1995)
1995
-
[10]
Kossel, J
D. Kossel, J. Opt. Soc. Am. 56, 1434 (1966); J. A. Ar- naud and A. A. M. Saleh, Appl. Opt. 13, 2343 (1974)
1966
-
[11]
P. Yeh, A. Yariv, and Chi-Shain Hong, J. Opt. Soc. Am. 67, 424 (1977). P. Yeh, A. Yariv, A. Y. Cho, Appl. Phys. Lett. 32, 104 (1978)
1977
-
[12]
K. G. Makris, S. Suntsov, D. N. Christodoulides, G. I. Stegeman, and A. Hach´ e, Opt. Lett. 30, 2466 (2005)
2005
-
[13]
M. I. Molina, R. A. Vicencio, Yu. S. Kivshar, Opt. Lett. 31, 1693 (2006)
2006
-
[14]
Suntsov, K
S. Suntsov, K. G. Makris, D. N. Christodoulides, G. I. Stegeman, A. Hach´ e, R. Morandotti, H. Yang, G. Salamo, M. Sorel, Phys. Rev. Lett. 96, 063901 (2006); E. Smirnov, M. Stepi´ c, C. E. R¨ uter, D. Kip, V. Shan- darov, Opt. Lett. 31, 2338 (2006); C. R. Rosberg, D. 5 N. Nesh...
2006
-
[15]
I. L. Garanovich, A. A. Sukhorukov, and Yu. S. Kivshar, Phys. Rev. Lett. 100, 203904 (2008)
2008
-
[16]
The band edges for the ABBA -chain in figure 2-b were obtained by numerically finding the roots of the cubic equation which gives the dispersion relation E(k), for k = 0 and ka = π , at each value of the hopping parameter t
-
[17]
Pinto, Masudul Haque, and Sergej Flach, Phys
Ricardo A. Pinto, Masudul Haque, and Sergej Flach, Phys. Rev. A 79, 052118 (2009)
2009
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.