Complex moments of the derivative of the Riemann zeta function
Christopher Hughes, Andrew Pearce-Crump · 2025
Read Complex moments of the derivative of the Riemann zeta function by Christopher Hughes, Andrew Pearce-Crump (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-07788.txt.
Introduction
Christopher Hughes and 1 co-author conjecture formulas in 2025 for the complex moments of the derivative of the Riemann zeta function at its zeros. The zeta function sits at the heart of number theory, and its nontrivial zeros encode the distribution of prime numbers. Moments, averages of powers, of its derivative sampled at those zeros are expected to follow patterns first seen in random matrix theory.
Two computations support the conjecture. The first finds an exact formula for the matching moments of characteristic polynomials of unitary matrices, averaged over the Haar measure, using Selberg's integral. The second follows the hybrid approach for zeta proposed by Gonek, Hughes, and Keating. Agreement between independent routes is the standard evidence offered for such conjectures, which guide later attempts at proof. The results are presented as conjectures, not theorems.
Catalog record
| CREATOR | Christopher Hughes, Andrew Pearce-Crump |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 40 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-07788
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Complex moments of the derivative of the Riemann zeta function Christopher Hughes, Andrew Pearce-Crump Abstract . We conjecture results about the complex moments of the derivative of the Riemann zeta function, evaluated at the non-trivial zeros of the Riemann zeta function. We do this via two different random matrix computations. In the first, we find an exact formula for the complex moments of the derivative of the characteristic polynomials of unitary matrices averaged over Haar measure using Selberg’s integral. In the second, we consider the hybrid approach for zeta, first proposed by Gonek, Hughes and Keating. 1. Introduction Let \zeta(s) be the Riemann zeta function. Throughout this paper we assume the Riemann Hypothesis, that is, that all non-trivial zeros can be written as \rho=\tfrac{1}{2}+i\gamma with \gamma real. The purpose of this paper is to make the following conjecture on discrete moments of the derivative of the Riemann zeta function, which we formulate by developing the philosophy of Keating and Snaith [ 11 ] , Hughes, Keating and O’Connell [ 9 ] , and Gonek, Hughes and Keating [ 7 ] from random matrix theory to discrete moments the zeta function. Conjecture 1 . For \Re(k)>-3 \frac{1}{N(T)}\sum_{0<\gamma\leq T}\zeta^{\prime}\left(\tfrac{1}{2}+i\gamma\right)^{k}\sim\frac{1}{\Gamma(k+2)}\left(\log\frac{T}{2\pi}\right)^{k}, as T\to\infty , where \Gamma(z) is the Gamma function and where N(T) is the number of non-trivial zeros in the critical strip, with 0<\gamma\leq T , given by N(T)=\frac{T}{2\pi}\log\frac{T}{2\pi e}+O(\log T). We prove a number of random matrix results that lead to this conjecture, and they are detailed in Section 3 . It is worth highlighting here that in order to take complex powers of complex numbers, we need to define the appropriate branch of the logarithm. The way the branch was chosen (discussed in Section 3 ) is harder to directly apply here without considering the Weierstrass product for zeta. We note that the appropriate choice of branch is not the same as that which comes from continuously varying \log\zeta^{\prime}(s) . We note that this conjecture agrees with the two previously known cases (ignoring the trivial k=0 case): • The case k=1 was proven by Conrey, Ghosh, and Gonek in [ 3 ] , unconditionally. • The case k=-1 can be found as equation (3.9) in a paper of Garaev and Sankaranarayanan [ 5 ] , under the assumption of simplicity of zeros. It was previously known under the additional assumption of the Riemann Hypothesis, and can be found in various sources, including [ 10 , 14 , 13 ] . Before forming this general conjecture, no other cases were known or even conjectured. Remark . We make a final remark about the special case when k=-2 : the right hand side of Conjecture 1 is zero. Indeed, in the random matrix calculations that lead to this conjecture, the equivalent calculation equals zero exactly. In the case of zeta, what we mean is the left hand side is bounded by some error terms with no leading-order asymptotic. An educated guess suggests the error terms will be O(T^{-1/2+\varepsilon}) . Remark . In a future paper [ 8 ] , we will consider the moments of mixed derivatives case, that is, when we consider moments of the form \sum_{0<\gamma\leq T}\zeta^{(n_{1})}\left(\tfrac{1}{2}+i\gamma\right)\dots\zeta^{(n_{k})}\left(\tfrac{1}{2}+i\gamma\right) for n_{1},\dots,n_{k}\in\mathbb{N} , as T\rightarrow\infty . In particular, in this integer moment case, we will also give full asymptotic formulae for such sums. As a special case, we can set all derivative to be equal and consider integer moments of any derivative of the Riemann zeta function given by \sum_{0<\gamma\leq T}\zeta^{(n)}\left(\tfrac{1}{2}+i\gamma\right)^{k} as T\rightarrow\infty for any k\in\mathbb{N} . 2. Philosophy of predicting moments of the zeta function through random matrix theory In this section we describe the philosophy of Keating and Snaith [ 11 ] and Hughes, Keating and O’Connell [ 9 ] from random matrix theory to moments of the zeta function. The model proposed by Keating and Snaith [ 11 ] is that the characteristic polynomial of a large random unitary matrix can be used to model the value distribution of the Riemann zeta function near a large height T . They used the characteristic polynomial (2.1) Z(\theta,A)=\det(I-Ae^{-i\theta})=\prod_{n=1}^{N}(1-e^{i(\theta_{n}-\theta)}), where the \theta_{n} are the eigenangles of an N\times N random unitary matrix A , to model \zeta(s) . The motivation for studying this is that it has been conjectured that the limiting distribution of the non-trivial zeros of \zeta(s) , on the scale of their mean spacing, is asymptotically the same as that of the eigenangles \theta_{n} of matrices chosen according to Haar measure, in the limit as N\to\infty . Equating the mean densities yields the connection between matrix size and height up the critical line N=\log\frac{T}{2\pi}. Averaging over all Haar-distributed N\times N unitary matrices, Keating and Snaith [ 11 ] calculated the moments of |Z(\theta)| and found that for fixed real \theta , if \Re(k)>-1/2 , as N\to\infty \mathbb{E}_{N}\left[|Z(\theta,A)|^{2k}\right]\sim\frac{G^{2}(k+1)}{G(2k+1)}N^{k^{2}} where \mathbb{E}_{N} denotes expectation with respect to Haar measure on N\times N unitary matrices, and G(z) is the Barnes G -function. Comparing with all known and conjectured results for continuous moments of zeta, they were lead to conjecture the following result. Conjecture 2 (Keating–Snaith) . For any fixed k with \Re(k)>-1/2 , \frac{1}{T}\int_{0}^{T}\left|\zeta\left(\tfrac{1}{2}+it\right)\right|^{2k}\ dt\sim a(k)\frac{G^{2}(k+1)}{G(2k+1)}\left(\log\frac{T}{2\pi}\right)^{k^{2}}, as T\rightarrow\infty , and (2.2) a(k)=\prod_{p\text{ prime}}\left(1-\frac{1}{p}\right)^{k^{2}}\sum_{m=0}^{\infty}\left(\frac{\Gamma(m+k)}{m!\,\Gamma(k)}\right)^{2}p^{-m} is an arithmetic factor which is discussed below. The Keating and Snaith result is for a continuous moment, but random matrix theory has also been used to predict discrete moments of \zeta(s) . Hughes, Keating and O’Connell [ 9 ] followed a similar approach and calculated the discrete moments of |Z^{\prime}(\theta,A)| (where the prime denotes differentiation with respect to \theta ) and found that for \Re(k)>-3/2 , as N\to\infty (2.3) \mathbb{E}_{N}\left[\frac{1}{N}\sum_{n=1}^{N}|Z^{\prime}(\theta_{n},A)|^{2k}\right]\sim\frac{G^{2}(k+2)}{G(2k+3)}N^{k(k+2)} where the sum is taken over all the eigenangles of the matrix A . As in the continuous case above, motivated by this result they made a conjecture for zeta which agreed with all known and previously conjectured results and is believed to hold in general. Conjecture 3 (Hughes–Keating–O’Connell) . For any fixed k with \Re(k)>-3/2 , \frac{1}{N(T)}\sum_{0<\gamma\leq T}\left|\zeta^{\prime}\left(\tfrac{1}{2}+i\gamma\right)\right|^{2k}\sim a(k)\frac{G^{2}(k+2)}{G(2k+3)}\left(\log\frac{T}{2\pi}\right)^{k(k+2)}, where a(k) is the same arithmetic factor as in the previous conjecture, given in ( 2.2 ). What wasn’t clear at the time was why the same arithmetic factor ( 2.2 ) should appear in these two conjectures. Indeed, the arithmetic factor itself was added in an ad hoc manner after the random matrix theory calculations were computed. This additional factor has been shown to be included naturally in the hybrid model developed by Gonek, Hughes and Keating [ 7 ] for continuous moments and of Bui, Gonek and Milinovich [ 2 ] for discrete moments. This hybrid approach gives equal weighting to Z_{X} , a partial Hadamard product over the zeros of zeta, and P_{X} , a partial Euler product over the primes; it is stated exactly in Theorem 5 in the next section. One strength of the hybrid approach is that it explained why the same arithmetic factor should appear in both conjectures. 3. Results justifying the formulation of Conjecture 1 The previous two problems where random matrix theory was used both concerned moments of a real function, where an absolute value of zeta was taken. For our problem, we consider the complex moments of the complex function. In this paper we attack our problem in two ways. First,