RADEMACHER SUMS, HECKE OPERATORS AND MOONSHINE
by
PAUL BRUNO
Submitted in partial fulfillment of the requirements For the degree of Doctor of Philosophy
Thesis Advisor: Dr. John F. R. Duncan
Department of Mathematics, Applied Mathematics and Statistics
CASE WESTERN RESERVE UNIVERSITY
May 2016 CASE WESTERN RESERVE UNIVERSITY
SCHOOL OF GRADUATE STUDIES
We hereby approve the thesis/dissertation of
Paul Bruno
candidate for the Doctor of Philosophy degree *
Committee Chair: Dr. John F. R. Duncan Dissertation Advisor Assistant Professor, Department of Mathematics, Applied Mathematics and Statistics Department of Mathematics and Computer Science, Emory University
Committee: Dr. David A. Singer Professor and Interim Chair, Department of Mathematics, Applied Mathematics and Statistics
Committee: Dr. Elisabeth M. Werner Professor, Department of Mathematics, Applied Mathematics and Statistics
Committee: Dr. Colin McLarty Truman P. Handy Professor of Philosophy, Department of Philosophy
May 2016
*We also certify that written approval has been obtained for any proprietary material contained therein. DEDICATION
For Margaret
iii Table of Contents
Dedication ...... iii Table of Contents ...... iv Acknowledgements ...... v Abstract ...... vi
1 Introduction and Background 1 1.1 Introduction and Overview of Results ...... 1 1.2 Modular Group ...... 3 1.2.1 Modular Functions and Lattice Functions ...... 5 1.2.2 Eisenstein Series ...... 9 1.2.3 Space of Modular Forms ...... 11 1.3 Hecke Operators ...... 13 1.4 Rademacher Sums ...... 17 1.5 Monstrous Moonshine ...... 20
2 Rademacher Sums 22 2.0 Matrices and Cosets ...... 22 2.1 Generalization of Rademacher Sums ...... 25 2.2 Coefficients ...... 30
3 Rademacher Sums and Hecke Operators 43
4 Rademacher Sums, Hecke Operators and Moonshine 56
5 Conclusions 70 Bibliography ...... 72
iv ACKNOWLEDGEMENTS
I would like to thank my advisor, Dr. John F. R. Duncan, for his mentoring over the past four years. He has been a constant role model and a true friend. I would also like to thank the Department of Mathematics, Applied Mathematics and Statistics at Case Western Reserve University for their support and guidance. I am particularly thankful for Dr. David A. Singer, Dr. Elisabeth M. Werner and Dr. Colin McLarty for their service on my thesis defense committee. Thank you to my fellow graduate students for making our department a family. Thanks to Richard Lartey for putting up with my music in the office. Thank you to my friends and family for having faith in me throughout the years. Thank you to Margaret for being the perfect partner.
v Rademacher Sums, Hecke Operators and Moonshine
Abstract
by
PAUL BRUNO
In this thesis, we analyze the actions of Hecke operators on Rademacher sums. To investigate the problem, we consider the coefficients of the Fourier expansion of a Rademacher sum. We derive explicit formulae for these coefficients and prove a number of relationships between them. Using this analysis, we identify distinguished families of Hecke operators for a large class of discrete groups of isometries of the hyperbolic plane. These groups coincide with the groups appearing in monstrous moonshine.
vi Chapter 1
Introduction and Background
1.1 Introduction and Overview of Results
The theory of modular forms has been an important area of study for mathematicians for the past two centuries. One of the most important modular functions is the elliptic modular invariant, which defines a correspondence between isomorphism classes of complex elliptic curves and points on the complex plane. In 1938, Hans Rademacher derived an explicit formula for the coefficients of the Fourier series expansion of the modular invariant. Rademacher’s method was applied to certain groups other than the modular group by Knopp in the 1960s. Now, functions constructed in this way are called Rademacher sums. The primary goal of this thesis is to find the Hecke operators for certain subgroups of the modular group that generalize the following result from Duncan and Frenkel.
(−k) Let T (k) denote the kth Hecke operator for the modular group and let RΓ (z) denote the order k Rademacher sum associated to the modular group. (For full definitions,
1 see Sections 1.3 and 1.4.) Then,
1 T (k)R(−1) (z) = R(−k)(z). (1.1.1) Γ k Γ
We are able to define modified Hecke operators for positive integers k and N, denoted
Tb(k, N), for certain subgroups of the modular group that give the result
(−1) 1 (−k) Tb(k, N)R (z) = R (z) (1.1.2) Γ0(N) k Γ0(N) where R(−k) (z) is the order k Rademacher sum associated to a particular subgroup Γ0(N) of the modular group, denoted Γ0(N). A full description of these Hecke operators is found in Chapter 3. We apply these Hecke operators to more general groups in Chapter 4. The monstrous moonshine conjecture associates with each conjugacy class of the monster group a modular function. Thanks to the work of Duncan and Frenkel, we know that these modular functions are equal to normalized Rademacher sums. For instance, the modular function associated to the conjugacy class of the identity element of the monster group is equal to the normalized Rademacher sum associated to the modular group. Note that this particular Rademacher sum coincides with the elliptic modular invariant with constant term set equal to zero. This example illustrates some of the remarkable relationships between the monster group and modular functions. In particular, the coefficients of the aforementioned normalized Rademacher sums turn out to be sums of the dimensions of irreducible representations of the monster group. Our analysis of Hecke operators and Rademacher sums has significance for mon- strous moonshine, as well. We show that the Hecke operators in equation (1.1.2) act
2 on the modular functions associated to certain conjugcay classes of the monster in the following way.
(−1) 1 (−k) Tb(k, N)R (z) = R (z) (1.1.3) Γg k Γg
For the remainder of Chapter 1, we give a survey of basic results in the theory of modular forms. Then, we briefly describe Hecke operators, Rademacher sums and monstrous moonshine. In Chapter 2, we generalize the Rademacher sums of [DF11] and derive formulae for their Fourier series coefficients. We will prove a number of relationships between these coefficients. In Chapter 3, we apply Hecke operators to Rademacher sums. We will see that for particular Hecke operators acting on Rademacher sums associated to certain discrete subgroups of the modular group, the result is another Rademacher sum. In Chapter 4, we investigate the implications of the results of the previous chapters with respect to monstrous moonshine.
1.2 Modular Group a b ˆ The group SL2 (R) = | a, b, c, d ∈ R, ad − bc = 1 acts on C = C∪{∞} c d a b in the following way. For γ = ∈ SL ( ) and z ∈ , set γ · z = az+b , and for 2 R C cz+d c d a z = ∞, set γ · z = c when c 6= 0, and γ · z = ∞ otherwise. Denote the upper half-plane by H = {z ∈ C|=(z) > 0} . We can verify that H is a b stable under the above action of SL2 (R) . Take γ = ∈ SL2 (R) and z ∈ H. c d
3 Then,
az + b =(γ · z) = = cz + d 1 az + b az¯ + b = − 2i cz + d cz¯ + d (1.2.1) 1 z(ad − bc) − z¯(ad − bc) = |cz + d|2 2i =(z) = . |cz + d|2
Hence, if z ∈ H, then γ · z ∈ H. Denote
a b SL2 (Z) = | a, b, c, d ∈ Z, ad − bc = 1 . (1.2.2) c d
We call this group the modular group. We now investigate properties of the modular group found in [Ser73]. Denote 0 −1 D = {z ∈ H | |z| ≥ 1, |<(z)| ≤ 1/2} . Denote S = ∈ SL2 (Z) and 1 0 1 1 T = ∈ SL2 (Z) . Then we have S · z = −1/z and T · z = z + 1. One can show 0 1 that SL2 (Z) = hS,T i and that D is a fundamental domain for the action of SL2 (Z) on H. Further, we can conclude that the stabilizer of any point in D is the identity except in the following cases. The stabilizer of z = i is the group of order 2 generated by S. The stabilizer of z = e(1/3) is the group of order 3 generated by ST and the stabilizer of z = e(1/6) is the group of order 3 generated by TS, where e(z) represent the quantity e2πiz.
4 1.2.1 Modular Functions and Lattice Functions
A complex-valued function is meromorphic on an open set if it is holomorphic every- where in that set except for a set of isolated points, the poles of the function. That is, the function has no essential singularities. Every meromorphic function is expressible as a fraction of two holomorphic functions. Hence, the isolated points are the roots of the holomorphic function of the denominator. For example, the quotient of two polynomials is a meromorphic function so long as the denominator is not identically zero.
A function, f(z), is weakly modular of weight 2k if f(z) is meromorphic on H a b and satisfies the following condition for any γ ∈ SL2 (Z) . For γ = , set c d jac(γ, z) = (cz + d)−2.
f(z) = jac (γ, z)k f (γ · z) (1.2.3)
Since SL2 (Z) is generated by S and T, we can rewrite the relationship of (1.2.3) with the following pair of equations.
f(z) = f(z + 1) (1.2.4) f(z) = z−2kf(−1/z)
The first condition of (1.2.4) allows us to write weakly modular functions as functions of e(z). In particular, we can rewrite f(z) as a sum over n as follows.
∞ X f(z) = cne(nz) (1.2.5) −∞
5 A weakly modular function is simply a modular function if it is meromorphic at infinity. In which case (1.2.5) becomes the following. For some n0 ∈ Z,
∞ X f(z) = cne(nz). (1.2.6)
n=n0
A modular function is a modular form if it is holomorphic everywhere, including at infinity. Then, we can rewrite (1.2.6) with n0 = 0. In particular,
∞ X f(z) = cne(nz). (1.2.7) n=0
A modular form that vanishes at infinity is called a cusp form. In this case we have c0 = 0 in (1.2.7). If f(z) is a modular form of weight 2k and f 0(z) is a modular form of weight 2k0, then the product ff 0(z) is a modular form of weight 2(k + k0). A holomorphic function on the upper half-plane h(z) is a weight zero mock modular form for Γ0(N) if it satisfies the following conditions [DMZ].
1. There exists a weight two modular form g(z) for Γ0(N), called the shadow of
∗ ∗ h(z), such that h(z) + g (z) is invariant under the action of Γ0(N) where g (z) solves
∂g∗(z) = −2πig(z). (1.2.8) ∂z¯