Section 6.2 Modular Curves Background I (John)
¶ Main references are lecture notes by Darmon and Weinstein “introduction to modular forms”.
Definition 6.2.1 .
Let
\begin{equation*}
\Gamma (N) = \left\{\begin{pmatrix} a \amp b \\ c \amp d \end{pmatrix}\in \SL_2(\ZZ) : \begin{pmatrix} a \amp b \\ c \amp d \end{pmatrix} \equiv \begin{pmatrix} 1 \amp 0 \\ 0 \amp 1 \end{pmatrix} \pmod N\right\}\text{.}
\end{equation*}
\(\Gamma\subseteq \SL_2(\ZZ)\) is a congruence subgroup if it contains \(\Gamma (N)\) for some \(N\text{.}\) Some important examples are
\begin{equation*}
\Gamma_1 (N) = \left\{\begin{pmatrix} a \amp b \\ c \amp d \end{pmatrix}\in \SL_2(\ZZ) : \begin{pmatrix} a \amp b \\ c \amp d \end{pmatrix} \equiv \begin{pmatrix} 1 \amp \ast \\ 0 \amp 1 \end{pmatrix} \pmod N\right\}
\end{equation*}
\begin{equation*}
\Gamma_0 (N) = \left\{\begin{pmatrix} a \amp b \\ c \amp d \end{pmatrix}\in \SL_2(\ZZ) : \begin{pmatrix} a \amp b \\ c \amp d \end{pmatrix} \equiv \begin{pmatrix} \ast \amp \ast \\ 0 \amp \ast \end{pmatrix} \pmod N\right\}
\end{equation*}
Definition 6.2.2 .
\begin{equation*}
f\colon \HH \to \CC
\end{equation*}
is a modular form of weight \(2k\) for \(\Gamma \) (with character \(\epsilon \)) if
\(f\) is holomorphic on \(\HH\text{.}\)
\(f\) is holomorphic at infinity.
\begin{equation*}
f|_{2k} \gamma (z) = f(z)\,\forall\gamma \in \Gamma
\end{equation*}
where
\begin{equation*}
f|_{2k} \gamma (z) = (cz+d)^{-2k}f(\gamma z)\epsilon (d)
\end{equation*}
Example 6.2.3 .
For \(\Gamma = \SL_2(\ZZ)\) from now on.
\begin{equation*}
f(z+1) = f(z)
\end{equation*}
\begin{equation*}
f\left(-\frac 1z\right) = z^{2k}f(z)
\end{equation*}
Using this we can write
\begin{equation*}
f(z) = f(q),\,q= e^{2\pi i z}\text{,}
\end{equation*}
where \(q\) is a parameter at infinity.
\begin{equation*}
f(q) = \sum_{n=0}^\infty a_n q^n\text{.}
\end{equation*}
Definition 6.2.4 .
A modular form is a cusp form if \(a_0 = 0\text{.}\)
Definition 6.2.5 .
\(M_k(\Gamma )\) is the space of weight \(k\) modular forms. \(S_k(\Gamma )\) is the space of weight \(k\) cusp forms.
Example 6.2.6 .
\begin{equation*}
G_{2k}(z) = \sum_{m,n\in \ZZ}' \frac{1}{(mz+n)^{2k}}
\end{equation*}
\begin{equation*}
g_{2} = G_4(z)/2\zeta (4)
\end{equation*}
\begin{equation*}
g_{3} = G_6(z)/2\zeta (6)
\end{equation*}
then
\begin{equation*}
\Delta = \frac{g_2^3 - g_3^2}{1728}
\end{equation*}
is a cusp form of weight 12.
Theorem 6.2.7 .
\begin{equation*}
D\colon M_k \to S_{k+12}
\end{equation*}
\begin{equation*}
f\mapsto \Delta f
\end{equation*}
is an isomorphism of vector spaces.
\begin{equation*}
\forall k \lt 0,\, M_k = 0
\end{equation*}
\begin{equation*}
k=2,\, M_k = 0
\end{equation*}
\(k\) odd, \(M_k = 0\text{.}\)
\begin{equation*}
M_k = S_k + G_k \CC,\,\forall k \in 2\ZZ^+
\end{equation*}
Proof.
\begin{equation*}
(f\Delta) (-1/z) = (cz+d)^{-k}f(z) (cz+d)^{-12} \Delta (z)
\end{equation*}
\begin{equation*}
= (cz+d)^{-(k+12)}(f\Delta ) (z)\text{.}
\end{equation*}
If \(k \lt 0\) , \(f \in M_k\) have \(f^{12}\Delta ^k \in S_0 = 0\text{.}\) \(M_0 =\CC\) corresponds to holomorphic functions on \(\SL_2(\ZZ) \backslash \HH\text{.}\)
\begin{equation*}
M_k \to \CC
\end{equation*}
\begin{equation*}
f(q) = a_0 + a_1 q + \cdots \mapsto a_0
\end{equation*}
\begin{equation*}
\dim(M_k / \ker) \le 1
\end{equation*}
\begin{equation*}
M_k = S_k + G_k \CC
\end{equation*}
we get
\(n\)
\(\dim M_k\)
\(\dim S_k\)
\(\lt 0\)
\(0\)
\(0\)
\(0\)
\(1\)
\(0\)
\(2\)
\(0\)
\(0\)
\(4\)
\(1\)
\(0\)
\(6\)
\(1\)
\(0\)
\(8\)
\(1\)
\(0\)
\(10\)
\(1\)
\(0\)
\(12\)
\(2\)
\(1\)
\(14\)
\(1\)
\(0\)
\(16\)
\(2\)
\(1\)
\(18\)
\(2\)
\(1\)
\(20\)
\(2\)
\(1\)
\(22\)
\(2\)
\(1\)
Table 6.2.8. dimensions
Hecke operators.
Definition 6.2.9 .
\(\Lambda \) is a lattice if it is a rank 2 \(\ZZ\)-module in \(\CC\) s.t. \(\CC/\Lambda \) is compact.
\begin{equation*}
\Lambda = \tau _1\ZZ + \tau _2 \ZZ
\end{equation*}
\begin{equation*}
\dim_\RR(\tau _1 \RR + \tau _2\RR) = 2
\end{equation*}
\begin{equation*}
\Im (\tau _2 / \tau _1 ) \gt 0
\end{equation*}
\(F\) is a homogeneous lattice function of weight \(k\) if it is
\begin{equation*}
F\colon R \to \CC
\end{equation*}
where \(R\) is the set of lattices, such that
\begin{equation*}
F(\lambda \Lambda ) = \lambda ^{-k}F(\Lambda )
\end{equation*}
\(F\) is holomorphic if \(f\colon \HH \to \CC\)
\begin{equation*}
f(\tau ) = F(\ZZ + \tau \ZZ)
\end{equation*}
is holomorphic on \(\HH\text{.}\)
\begin{equation*}
\{\text{holo. homog. wt. }k\text{ lattice fns.}\}\leftrightarrow \{\text{wt. }k\text{ mod. fms.}\}
\end{equation*}
\begin{equation*}
F\mapsto f_F\colon \tau \mapsto F(\ZZ+ \tau \ZZ)
\end{equation*}
\begin{equation*}
F_f \mapsfrom f
\end{equation*}
\begin{equation*}
F_f (\tau _1 \ZZ+ \tau _2 \ZZ) = f(\tau _2/\tau _1)\text{.}
\end{equation*}
Definition 6.2.10 .
\(F\) is a homogeneous holomorphic weight \(k\) lattice function then
\begin{equation*}
T_{n,k} F(\Delta ) = n^{k-1} \sum_{\Lambda ' \subseteq \Lambda ,\,[\Lambda : \Lambda '] = n} F(\Lambda ')
\end{equation*}
\begin{equation*}
T_{n,k}f = f_{T_{n,k}(F_f)}
\end{equation*}
\begin{equation*}
T_{n,k} f(z) = n^{k-1} \sum_{\gamma \in \SL_2(\ZZ)\backslash M_n } f(\gamma z)(cz+d)^{-k}
\end{equation*}
where \(M_n \subseteq M_2(\ZZ)\) is the set of integer matrices of determinant \(n\text{.}\)
\begin{equation*}
f|\alpha \beta = (f|\alpha )|\beta \text{.}
\end{equation*}
The fourier expansions of these are given by
\begin{equation*}
f(q) = a_0 + a_1 q + \cdots
\end{equation*}
\begin{equation*}
T_{n,k}f(q) = \sum_{m = 0}^\infty \sum_{d|(m,n)} d^{k-1} a_{nm/d^2} q^m
\end{equation*}
fixing \(k\) we have if \((a,b) = 1\)
\begin{equation*}
T_aT_b = T_{ab}
\end{equation*}
for \(p \) prime
\begin{equation*}
T_p T_{p^t} = T_{p^{k+1}} + p^{k-1} T_{p^{t-1}}\text{.}
\end{equation*}
Figure 6.2.11. The fundamental domain These \(T_{n,k}\) operate on \(M_k(\SL_2(\ZZ))\) and we can define:
Definition 6.2.12 .
Let \(f\in M_k(\SL_2(\ZZ))\) is an eigenform if it is a simultaneous eigenvector for \(\{T_n\}_{n=1}^\infty \text{.}\) We have
\begin{equation*}
\pair - - \colon S_k \times S_k\to \CC
\end{equation*}
\begin{equation*}
\pair fg = \int_F f(z) \overline{g(z)} y^k \frac{\diff x \diff y}{y^2}\text{.}
\end{equation*}
Proposition 6.2.13 .
\(T_n\) is self-adjoint in \(S_k\)
\begin{equation*}
\pair{T_n f}{g}= \pair {f}{T_ng}
\end{equation*}
\begin{equation*}
S_k = \bigoplus f_i \CC
\end{equation*}
an orthogonal basis of eigenforms. \(G_k\) is an eigenform and
\begin{equation*}
M_k = G_k \CC + \bigoplus f_i \CC\text{.}
\end{equation*}
Definition 6.2.14 .
\(f\) is a normalized eigenform if
\begin{equation*}
a_1(f) =1
\end{equation*}
\begin{equation*}
a_n(f) = a_1(T_nf) = a_1(\lambda _n f) = \lambda _na_1(f) = \lambda _n \text{.}
\end{equation*}
Proposition 6.2.15 .
If \(f\in S_k\) is a normalized eigenform then
\begin{equation*}
\QQ(a_1(f), a_2(f) , \ldots)
\end{equation*}
is a finite totally real extension of \(\QQ\) with degree \(\le \dim S_k = d\text{.}\)
Proof. \(G_k/\zeta (k)\) and \(\Delta \) have rational coefficients.
\begin{equation*}
M_k(\QQ) = f_1 \QQ+ \cdots + f_d \QQ
\end{equation*}
\(T_n\) operates on \(M_k(\QQ)\) so
\begin{equation*}
T_n \hookrightarrow \Mat_d(\QQ)
\end{equation*}
\begin{equation*}
\mathbf T_k =\QQ[T_1,T_2, \ldots] \hookrightarrow \Mat_d(\QQ)
\end{equation*}
\(\forall \phi \in \Hom(\mathbf T_{k}, \overline \QQ)\) have \(\im (\phi )\) lies in a degree \(\le d\) extension. this is totally real because \(T_n\) are self adjoint w.r.t. a Hermitian inner product.
Generalisation.
If \(\Gamma \) is any congruence subgroup
\begin{equation*}
M_k(\Gamma ) = \operatorname{Eis}_k(\Gamma ) + S_k(\Gamma )
\end{equation*}
\begin{equation*}
M_k(\SL_2(\ZZ)) \subseteq M_k(\Gamma )
\end{equation*}
Given \(d|N\) we have
\begin{equation*}
f \in M_k(\Gamma (d))
\end{equation*}
can define dilation
\begin{equation*}
g(z) = f(N/d z) \in M_k(\Gamma )
\end{equation*}
for large enough \(k\) we can find a basis of \(M_k(\Gamma (N))\) by taking dilations of products of \(M_a(\Gamma (N))\) for \(a \lt k\) and Hecke operators.
For \((n,N) = 1\text{,}\) \(f\in M_k(\Gamma (N))\text{.}\)
\begin{equation*}
T_{n,k} f = n^{k-1} \sum_{\gamma \in \Gamma (N) \backslash M_n} (f|\gamma )(z)
\end{equation*}
\(M_n\) is integer upper triangular with determinant \(n\text{.}\) For primes \(l\)
\begin{equation*}
l\nmid N,\, T_l f(q) = \sum a_n l q^n + l^{k-1} \sum a_n \langle l \rangle f q^{nl}
\end{equation*}
\begin{equation*}
l | N,\, T_lf(q) = \sum a_{nl} q^n
\end{equation*}
\begin{equation*}
\langle l \rangle f(z) = \left(f|_k \begin{pmatrix} l \amp 0 \\ 0 \amp l\inv \end{pmatrix}\right) (z)\text{.}
\end{equation*}
Definition 6.2.16 .
\begin{equation*}
g \in S_k(\Gamma (N))
\end{equation*}
s.t.
\begin{equation*}
g(z) = f(d z)
\end{equation*}
for some \(f \in S_k(\Gamma (N/d))\) is called an oldform. Newforms are \(f\in S_k(\Gamma (N))\) s.t.
\begin{equation*}
\pair fg = 0
\end{equation*}
for all \(g\in S_k(\Gamma (N))^{old}\text{.}\)