## 1. Precise statement and assumptions The claim is true as an **existence statement about an idealized family of equilibrium models**, with “window” defined by a nonzero affinity or occupancy threshold. A precise version is: > Fix a temperature \(T_0>0\), an integer \(m\ge 0\), a relative-affinity threshold \(0<\rho<1\), and a desired width \(\varepsilon>0\). There exists a finite-state, single-ligand-site equilibrium model such that: > \[ > K(T_0)=K_{\max}>0,\qquad > \left.\frac{d^j\Delta H}{dT^j}\right|_{T_0}=0 > \quad(j=0,\ldots,m), > \] > while the set > \[ > \{T>0:K(T)\ge \rho K_{\max}\} > \] > is an interval containing \(T_0\) whose temperature width is less than \(\varepsilon\). Here \(j=0\) means the binding enthalpy itself. The assumptions needed are: - \(K\) is a **dimensionless standard equilibrium association constant**, and \[ \Delta H=-\frac{\partial\log K}{\partial\beta}, \qquad \beta=\frac1{RT}. \] Free-ligand standard-state contributions must either be included or absorbed into the assigned binding thermodynamics. - The state enthalpies and entropy weights are temperature-independent. Parameters such as \(E\) and \(\beta_0\) are held fixed when differentiating. - The construction permits arbitrarily large energy gaps and entropy differences. Increasing \(E\) also requires changing the specified entropy weights. - All relevant states equilibrate; no kinetic limitation is imposed. - “Single-site” means at most one ligand is bound, **not** that the receptor has only two conformational states. - “Center” means the affinity maximum \(T_0\). The threshold window is exactly symmetric in inverse temperature, not generally in temperature. - “Window” does not mean exact zero binding outside a compact interval. These qualifications are important for physical interpretation, but do not obstruct the mathematical result. ## 2. Checking the supporting argument Let \(A>0\) be the common multiplier for bound-state weights, and take an even integer \(N\ge2\) and \(E>0\). Put \[ \delta=\beta-\beta_0,\qquad y=e^{-E\delta}. \] The partition sums are \[ Z_U=\sum_{k\ {\rm even}}\binom Nk y^k, \qquad Z_B=A\sum_{k\ {\rm odd}}\binom Nk y^k. \] The entropy factor \(e^{\beta_0kE}\) has correctly combined with the Boltzmann factor \(e^{-\beta kE}\) to produce \(y^k\). ### Partition sums and affinity The binomial theorem gives \[ Z_U=\frac{(1+y)^N+(1-y)^N}{2}, \] \[ Z_B=A\frac{(1+y)^N-(1-y)^N}{2}. \] At \(\beta=\beta_0\), \(y=1\), so \(K(T_0)=Z_B/Z_U=A\). Also, \[ \frac{1-y}{1+y} =\tanh\!\left(\frac{E\delta}{2}\right)=z. \] Consequently, \[ \boxed{\frac{K(T)}{K(T_0)}=\frac{1-z^N}{1+z^N}.} \] This step is correct. ### Maximum and local flatness For even \(N\), \(0\le z^N<1\) at every finite positive temperature. Thus \(K\le K(T_0)\), with equality only at \(T_0\). Moreover, \[ \log\frac{K}{K(T_0)} =-2\operatorname{artanh}(z^N). \] Using \(z=E\delta/2+O(\delta^3)\), \[ \log\frac{K}{K(T_0)} =-2\left(\frac E2\right)^N\delta^N +O(\delta^{N+2}). \] Therefore \[ \Delta H =2N\left(\frac E2\right)^N\delta^{N-1} +O(\delta^{N+1}). \] Since \(\beta(T)-\beta_0\) has a simple zero at \(T_0\), \(\Delta H(T)\) has a zero of exactly order \(N-1\). Hence \[ \boxed{\left.\frac{d^j\Delta H}{dT^j}\right|_{T_0}=0 \quad\text{for }j=0,\ldots,N-2.} \] The stated derivative count is correct. Choosing even \(N\ge m+2\) handles any prescribed finite number \(m\) of derivatives. An exact check is \[ \Delta H= \frac{NE\,z^{N-1}(1-z^2)}{1-z^{2N}}, \] which has precisely the claimed zero at \(z=0\). ### Arbitrarily narrow threshold window Fix \(0<\rho<1\), and define \[ c_{\rho,N} =2\operatorname{artanh} \left[\left(\frac{1-\rho}{1+\rho}\right)^{1/N}\right]. \] Solving the threshold inequality gives \[ \frac{K(T)}{K(T_0)}\ge\rho \quad\Longleftrightarrow\quad |\beta-\beta_0|\le \frac{c_{\rho,N}}E. \] Write \(b=c_{\rho,N}/E\). Once \(b<\beta_0\), the temperature endpoints are \[ T_{\rm low}=\frac1{R(\beta_0+b)},\qquad T_{\rm high}=\frac1{R(\beta_0-b)}. \] Their separation is \[ \boxed{ \Delta T=\frac{2b}{R(\beta_0^2-b^2)} \longrightarrow0\quad(E\to\infty). } \] This proves the arbitrarily narrow-window assertion in ordinary temperature, not just inverse temperature. ## 3. Independent derivation There is a useful probabilistic derivation that also explains the cancellations. Introduce \(N\) auxiliary binary variables \(x_i\in\{0,1\}\), and let \[ k=\sum_i x_i. \] These variables label internal conformations; they are not ligand-binding sites. Give a configuration weight \(y^k\). Under the normalized product distribution, \[ P(x_i=1)=\frac{y}{1+y}. \] The parity expectation is \[ \mathbb E[(-1)^k] =\prod_i\mathbb E[(-1)^{x_i}] =\left(\frac{1-y}{1+y}\right)^N=z^N. \] Therefore \[ P(k\text{ even})=\frac{1+z^N}{2},\qquad P(k\text{ odd})=\frac{1-z^N}{2}. \] Assigning even parity to the unbound pool and odd parity to the bound pool immediately yields the claimed affinity ratio. Independently, the high-order flatness follows from the classical finite-difference identity \[ \sum_{k=0}^N(-1)^k\binom Nk k^j=0 \qquad(j0\) at every finite positive temperature. Also, at fixed finite \(E\), \[ \lim_{T\to\infty}\frac{K(T)}{K(T_0)} = \frac{1-\tanh^N(E\beta_0/2)} {1+\tanh^N(E\beta_0/2)} >0. \] Thus “outside it, \(K\) approaches zero” needs the following interpretation: \[ \frac{K(T)}{K(T_0)}\longrightarrow0 \quad\text{as }E\to\infty \quad\text{for each fixed }T\ne T_0. \] It is not an exact off-window zero, nor generally a zero high-temperature limit for a fixed receptor. ### Affinity versus actual binding At fixed ligand activity \(a\), occupancy is \[ \theta(T)=\frac{aK(T)}{1+aK(T)}. \] It is monotone in \(K\), so any fixed occupancy threshold below \(\theta(T_0)\) likewise defines an arbitrarily narrow window. A narrow affinity peak therefore can produce a narrow binding window, but appreciable central occupancy requires an appropriate affinity scale and ligand activity. The ligand binding polynomial is only linear in \(a\): \[ Z_{\rm total}=Z_U+aZ_B. \] Accordingly, binding is single-site and Langmuir-like at each temperature. The construction does, however, require elaborate internal conformational thermodynamics. ### No hidden thermodynamic instability is required For either pool, the assumed partition model gives \[ C_p=\frac{\operatorname{Var}(H)}{RT^2}\ge0. \] Thus the construction does not require negative heat capacities or negative statistical weights. It can require enormous *individual* heat capacities whose differences cancel. For \(N\ge4\), at \(T_0\), \[ C_{p,U}=C_{p,B}=\frac{NE^2}{4RT_0^2}. \] Null binding heat-capacity information is therefore not the same as small underlying thermal fluctuations. ### “Finite number” is essential For every finite \(N\), the next derivative is nonzero. This model is analytic around \(T_0\); if **all** derivatives of \(\Delta H\) vanished there, analyticity would force \(\Delta H\) to vanish locally and \(K\) to be locally constant. So the result cannot simply be strengthened to infinitely many null derivatives. **Conclusion of the attempted falsification:** the finite-order mathematical claim survives. No proof gap remains under the assumptions stated above. Its physical generality is limited by the unrestricted energy and entropy scales and exact parameter matching. ## 5. Novelty The mathematical mechanism is a **variant of known results**, principally: 1. **Even/odd binomial projection**, producing the two partition sums. 2. **Binomial finite-difference moment cancellation**: \[ \sum_k(-1)^k\binom Nk k^j=0\quad(j