Why changing basis in Heun equations needs a phase
Heun-type equations — the Heun equation together with its confluent, doubly confluent and reduced forms — appear in two settings the paper treats as one problem. In gauge theory they describe the quantisation of Seiberg-Witten differentials for N=2 supersymmetric Yang-Mills in the Nekrasov-Shatashvili background, with different confluences covering different flavour numbers. In gravity they appear when the Einstein equations for a perturbed black hole metric are separated, with the radial part of the Teukolsky or Regge-Wheeler equations taking a confluent Heun form.
Two natural bases of solutions live on these equations. The in/up basis used in the ODE/Integrable Models correspondence consists of a solution that is incoming at the horizon and one that is outgoing at infinity; in integrability parameters both decay at their respective singular points. The Floquet solutions instead satisfy ψ±(y+2πi)=e^{±2πik}ψ±(y), so they are periodic up to a known multiplier.
Passing from one basis to the other is not a relabelling. The connection coefficients depend on a phase φ that the Floquet solutions acquire when moving between singularities, and that phase is the object the paper sets out to compute. The paper identifies φ as a further exponent of the Floquet solutions, alongside the index k and the Wronskian, and notes that these ingredients together allow the parallel transport or Fenchel-Nielsen coordinates to be computed.
The kink method turns quantum momentum into series
The computational device is the logarithmic derivative Π(y)=d/dy ln ψ(y), called the quantum momentum, which obeys a Riccati equation associated with the Heun-type equation. The difficulty is that the limits θ→−∞ and y→±∞ conflict: a naive expansion in the small instanton parameter cannot be taken while also sitting at a singular point.
The paper resolves this by introducing two kink Riccati equations, one for the region y>0 and one for y<0, with explicit instances for the modified Mathieu equation, the doubly confluent Heun equation, the Heun equation and their confluent limits. The solutions of these kink equations are written as convergent series and found through a systematic recursive procedure, so the quantum momentum of the Floquet solutions is obtained explicitly at all orders rather than only at leading order.
Because the Floquet solutions can then be integrated term by term, the acquired phase φ follows as a series as well, together with its two regional contributions φ< and φ>. The paper reports explicit first-order formulae for each equation class, and also shows that the quantum momentum of the in/up solutions can be written as a single convergent series, for example in the modified Mathieu case and for the incoming and outgoing confluent Heun solutions.
The phase matches the dual gauge period A_D
The gauge-side test is a comparison, not an assumption. The quantity φ for the modified Mathieu equation, and its symmetrised version Φ for the doubly confluent Heun, confluent Heun and Heun equations, are computed at various orders and then conjectured to coincide with A_D/ℏ, the dual gauge period of N=2 supersymmetric Yang-Mills in the Nekrasov-Shatashvili background, for N_f=0, 2, 3 and 4 respectively.
The status of this identification differs by case, and the paper is explicit about it. For the modified Mathieu equation the rigorous proof is attributed to separate work, and the paper states that the extension to the other equations is left to an incoming publication. What this paper establishes is the order-by-order computation and the matching, which the authors describe as a non-trivial check against instanton expansions of gauge theory.
The same discussion records an identity relating the derivative of P² with respect to k to the derivative of φ with respect to θ. The paper presents this as the irrotationality condition for the vector (P², −φ/2), implying that this vector is the gradient of a potential whose rescaled form coincides with the gauge theory prepotential. The paper states that this extension has been achieved order by order at small e^θ for the doubly confluent, confluent and full Heun equations, while a functional proof is deferred.
Gravity results agree with standard black hole methods
On the gravity side the paper compares its explicit series with established treatments of the Teukolsky and Regge-Wheeler equations. The comparison is not merely qualitative: in the confluent Heun case the authors reproduce a known approximation for the incoming solution, and they note specific differences. Their expression differs from the earlier result by an extra prefactor of 1/2, which they suggest is missing in the earlier work, and it improves the arguments of the hypergeometric and gamma functions by resumming all instantons.
The paper also stresses where the Floquet approximation stops being valid. The approximation holds when e^{θ3} is much smaller than e^{2y'}, which in gravity variables means ωr ≪ M/r, and it cannot be extended close to the horizon at r=2M. Near the horizon the exact behaviour is instead reproduced by a specific linear combination of the two Floquet functions, and the paper provides a separate single convergent series in powers of e^{θ3} that works around the horizon.
A concrete gravity application is the dynamical tidal Love number of a Schwarzschild black hole. The paper separates the radial perturbation into a response part and a tidal part, defines the dynamical Love number as the ratio of their coefficients, and computes it to first order in Mω. The static value, obtained by setting ω=0, vanishes for physical integer ℓ≥2 because gamma functions in the expression have simple poles, so the leading contribution is the O(Mω) term. The resulting expression agrees with previously reported results.
What the method replaces and what it does not yet settle
The paper positions the kink method as an effective alternative to instanton and gravity computations: instead of summing instantons or solving the gravitational perturbation problem by standard techniques, one solves the ODE and its monodromy and connection problem directly. The authors also report that the wave-function expansions reproduce recently found Floquet results and, more surprisingly, expansions of the MST ansatz.
The scope conditions are set by the expansion parameter. The series are convergent expansions in e^θ, the small instanton coupling of the gauge theory, and the paper states that the identifications and constructions are expected to remain valid in the strong-coupling regime of large Λ/ℏ and in the strong-field regime of large Mω, where methods from earlier work can be applied. Those strong-coupling statements are expectations about applicability, not results demonstrated here.
The paper also frames its own reach modestly. It describes itself as marking initial steps toward a dictionary between integrability and gravitational quantities, and it leaves the functional proof of the prepotential relation beyond the modified Mathieu case, as well as derivatives with respect to masses, to future work.
Where the paper changes the comparison
The most decision-relevant result is not a single number but a change in what has to be computed. For the modified Mathieu equation, the phase φ and the dual period A_D are tied by an order-by-order series that the paper can generate systematically; for the doubly confluent, confluent and full Heun equations, the same matching is presented as a conjecture supported by computed orders rather than a proof. A reader comparing methods should therefore treat the N_f=0 case as the established benchmark and the N_f=2, 3, 4 cases as the paper's extrapolation of that pattern.
The gravity side carries a similar split. The agreement with standard Teukolsky and Regge-Wheeler analyses, and the first-order-in-Mω dynamical Love number matching earlier results, are checks against known quantities. The genuinely new offer is the systematic all-order series for the phase and for the quantum momentum, including single-series forms for the in/up solutions, which the paper presents as a route to quantities otherwise obtained from instanton or gravity computations.
The practical priority that follows is to use the kink-method series where the expansion parameter e^θ is small and the point of interest is not too close to the horizon or to infinity, and to switch to the dedicated single-series representations near those boundaries. The paper's own framing makes the gauge-period identification for N_f=0 the firmest anchor, with the remaining flavour cases and the strong-coupling regime awaiting the proofs and extensions it announces.
Sources used in this answer
Regular and Floquet bases for gauge and gravity theories: a non perturbative approach
A kink method computes the quantum momentum of Heun-type equations as convergent series at all orders, yielding the Floquet acquired phase as a series that matches the dual gauge period A_D and agrees with standard black hole perturbation results.