Beruflich Dokumente
Kultur Dokumente
WEI ZHANG
Abstract. Here are the notes I took for Wei Zhangs course on modular forms offered at
Columbia University in Spring 2013 (MATH G4657: Algebraic Number Theory). Hopefully
these notes will appear in a more complete form during Fall 2014. I recommend that you
visit my website from time to time for the most updated version.
Due to my own lack of understanding of the materials, I have inevitably introduced
both mathematical and typographical errors in these notes. Please send corrections and
comments to phlee@math.columbia.edu.
Contents
Classical Modular Forms
1. Lecture 1 (January 23, 2013) 3
1.1. Introduction 3
1.2. Basic Definitions 3
1.3. Eisenstein series 5
2. Lecture 2 (January 28, 2013) 6
2.1. Eisenstein series 6
3. Lecture 3 (January 30, 2013) 9
3.1. Modular Curves 9
4. Lecture 4 (February 4, 2013) 13
4.1. Modular Curves 13
4.2. Elliptic curves with level structures 15
5. Lecture 5 (February 6, 2013) 16
5.1. Heegner points 16
5.2. Dimension formulas 17
6. Lecture 6 (February 11, 2013) 19
6.1. Hecke operators 19
7. Lecture 7 (February 13, 2013) 21
7.1. Hecke operators for level 1 21
7.2. Hecke operators for level N 23
7.3. Petersson inner product 25
8. Lecture 8 (February 18, 2013) 25
8.1. Petersson inner product 25
8.2. L-functions 27
9. Lecture 9 (February 20, 2013) 29
2
1. Lecture 1 (January 23, 2013)
1.1. Introduction. Last year this course was about class field theory. This year we will
focus on modular forms. The two main topics are:
(1) classical (holomorphic) modular forms for the full modular group SL2 (Z) and its
congruence subgroups. We will be interested in both the analytic and arithmetic
theory.
(2) p-adic properties and p-adic modular forms.
We will not follow any specific textbook closely, but a list of references includes:
(1) Diamond, Shurman, A First Course in Modular Forms. Homework will mainly be
assigned from this book (among other resources).
(2) Serre, A Course in Arithmetic. The last part of this book has a concise explanation
of modular forms for SL2 (Z).
(3) Shimura, Introduction to the Arithmetic Theory of Automorphic Functions.
(4) Lang, Introduction to Modular Forms.
1.2. Basic Definitions. We begin with some basic definitions.
Definition 1.1. The upper half plane is H = { C : Im( ) > 0}.
We will reserve the
letter zfor any complex number not necessarily contained in H.
a b
Any SL2 (R) = : a, b, c, d R, det = 1 acts on H via
c d
a + b
= .
c + d
Check that this defines a group action. The key point is
Im( )
Im( ) =
|c + d|2
using the fact that det() = 1. Moreover this action defines a biholomorphic automorphism
of H. The set of biholomorphic automorphisms of H is
Aut(H)
= PSL2 (R) = SL2 (R)/{I}.
We will be interested in discrete subgroups of SL2 (Z). More specifically,
Definition 1.2. Congruence subgroups are subgroups SL2 (Z) that contain (N ) for
some positive integer N , where (N ) is the principal congruence subgroup of level N defined
by
a b a b 1 0
(N ) = SL2 (Z) : (mod N ) .
c d c d 0 1
We can define various other subgroups that have finite index in SL2 (Z), for example, by
changing the condition into
a b a b
0 (N ) = SL2 (Z) : (mod N ) .
c d c d 0
Definition 1.3. A weakly modular function of weight k is a function f : H C such that:
(1) (meromorphicity) f is meromorphic on H;
3
a + b k a b
(2) (transformation rule) f ( ) = f = (c + d) f ( ) for all =
c + d c d
SL2 (Z).
The group SL2 (Z) is generated by two elements, which are the matrices corresponding to
translation 7 + 1 and inversion 7 1 .
1 1 0 1
Lemma 1.4. SL2 (Z) = , .
0 1 1 0
With this we can show
Lemma 1.5. A fundamental domain of H modulo SL2 (Z) is
1 1 1
D = H : Re( ) < , | | > 1 H : Re( ) 0, | | = 1
2 2 2
i.e. under SL2 (Z)-transformation, every point in H is equivalent to exactly one point in D.
(Diagram here)
A proof can be found in Serres book. The key idea is to maximize Im( ). Maximum
is achieved precisely when it lies in the fundamental domain.
Since I acts trivially on H, we consider PSL2 (Z) = SL2 (Z)/{I}.
Suppose
f is a weakly modular function that is holomorphic on H. Transformation by
1 1
= shows that f satisfies f ( + 1) = f ( ), i.e. f has period 1. We can then consider
0 1
its Fourier expansion: for any = x + iy, viewing f as a function of x gives
X
f (x + iy) = an (y)e2inx
nZ
So far we have not used the holomorphicity Rof f . By contour integration over the rectangle
1+iy
with vertices iy, 1+iy, iy 0 , 1+iy 0 , we have that 0+iy f ( )e2in d is independent of y because
the integral over the two vertical edges cancel. Hence we can write
X
f ( ) = an e2in .
nZ
2i
It is convenient to introduce q = e . Then the Fourier expansion of f is given by
X
f ( ) = an q n .
nZ
Geometrically, the map 7 q = e2i sends the upper half plane H into D {0} where
D = {z C : |z| < 1} is the open unit disk. In fact this induces a biholomorphism
H/Z = D {0}.
Here we see that the singularity around the origin is important. A priori we could have any
Fourier expansion.
4
Definition 1.6. f is meromorphic at = if there exists M Z such that an = 0 for all
n < M . f is holomorphic at = if an = 0 for all n < 0.
This is equivalent to saying that f (q) extends to a meromorphic (holomorphic) function
on the unit disk D.
Definition 1.7. A modular form of weight k is a function f : H C such that:
(1) f is holomorphic on H{} (i.e. the Fourier expansion only has nonnegative powers
of q);
a b
(2) f ( ) = (c + d)k f ( ) for all = SL2 (Z).
c d
Let us look at some examples, otherwise we would just be doing function theory.
Gk ( ) = (c + d)k Gk ( ).
Gk is identically zero when k is odd, so it is only interesting when k is even.
We have not addressed convergence issue yet.
Lemma 1.8. Gk ( ) is absolutely convergent if k 4, and uniformly convergent on any
compact subset of H, so it defines a holomorphic function on H.
Proof (Sketch). To prove convergence, we can assume D is in the fundamental domain.
Then
D
|w|2 = |m + n|2 = m2 + 2mn Re( ) + n2 m2 mn + n2 = |m + n|2
(where is a cube root of unity). Therefore we can bound
X0 X0
|Gk ( )| |w|k = |m + n|k
m,n m,n
which is convergent when k 3. This proves convergence and uniform convergence over
compact subsets.
By absolute convergence, we can evaluate limits term-wise.
X0
Lemma 1.9. lim Gk ( ) = nk = 2(k) where is the Riemann zeta function.
nZ
Corollary 1.10. For k 4 even, Gk is a modular form of weight k, called the Eisenstein
series of weight k.
5
Later we will see that these are essentially the only modular forms.
Next we consider the Fourier expansion of Gk ( ). It is well-known that
1 X 1 1 cos z
+ + = cot z =
z d=1 z + d z d sin z
which has simple
P poles at the integers and is holomorphic elsewhere. Be careful not to write
1
the sum as dZ z+d because this is not convergent!
For a proof of this identity, Diamond-Shurman gives the hint
Y z2
sin z = z 1 2
n1
n
but this is like cheating because the two identities are equivalent the identity we want is
just the logarithmic derivative of this one. Instead we invoke a general theorem.
Lemma 1.11 (Special case of Mittag-Leffler). If f : C C is a meromorphic function
having simple poles a1 , a2 , , ai , with residues b1 , b2 , , bi , respectively such that
(1) the sequence 0 < |a1 | |a2 | |ai | goes to ;
(2) there exist closed contours Cm with length lm and distance Rm to 0 such that lm /Rm <
N1 , Rm , and |f |Cm | < N2 for given positive numbers N1 , N2 , then
X 1 1
f (z) = f (0) + bi +
i=1
z ai ai
The proof uses the Cauchy integral formula and is left as an exercise.
1The left hand side is written this way for simplicity. There will not be any convergence issues after
differentiation.
7
X
E6 = 1 504 5 (n)q n Z[[q]].
n=1
In general we only have E2k ( ) Q[[q]]. The fact that E2k has rational coefficients is
important in the second half of the course when we discuss p-adic modular forms.
X (d)
We can generalize modular forms for SL2 (Z) to (N ), for example .
(c + d)k
c0 (mod N )
We will prove that, in a certain sense, the Eisenstein series generate all the modular forms.
If f is a weakly modular function of weight k for SL2 (Z) that is meromorphic everywhere
on H {}, we define its order of vanishing v (f ) at H {} as follows. At H, this
is just the routine definition for complex-valued functions. For = , consider the Fourier
expansion !
X X
f ( ) = an q n = q n0 an0 + an q n
n n>n0
where an0 6= 0. Then the order of vanishing is defined to be
v (f ) = n0 .
We can check that v depends only on the SL2 (Z)-orbit of .
Lemma 2.2.
X k
m v (f ) = (1)
12
D{}
where 1
if = i
2
1
m = if =
3
1 if = or D\{i, }
Remark. The weights are present because i and have non-trivial stabilizers. Later we will
give a more geometric proof using modular curves and Riemann-Roch.
Proof (Sketch). Use contour integration along the boundary of D truncated by a horizontal
segment which bounds all the zeroes and poles in D, detouring around , i, + 1 along
small circular arcs. We first assume there are no zeroes or poles on the boundary. By the
argument principle, Z 0
1 f ( ) X
d = v (f ).
2i f ( )
D\{i,}
1
Around = , + 1, the contribution from the two arcs is 2( 2 v (f )) = 31 v (f ).
3
Similarly, around = i, the contribution is 12 vi (f ).
Under 7 1 , f ( 1 ) = k f ( ). The arc on the unit circle has contribution 12 k
.
2i
Under 7 q = e , the horizontal path gives v (f ).
If there are zeroes or poles on the boundary, we detour along small circular arcs such that
each equivalence class of zeroes and poles is counted exactly once inside the contour.
This handy formula can help us find the dimension and basis of modular forms for SL2 (Z).
Theorem 2.3.
8
(1) The graded algebra of all modular forms for SL2 (Z) is
M
Mk = C[E4 , E6 ]
k=0
where the latter B e is the corresponding subgroup of GL2 (Z/N ). It is a standard exercise to
find the order of GL2 (Z/N ).
For N = p a prime, the order is (p2 1)(p2 p), the number of bases of (Z/p)2 .
For N = pk a prime power, the order is ( Np )4 (p2 1)(p2 p) = N 4 (1 p1 )(1 p2 ).
In general, | GL2 (Z/N )| = N 4 p|N (1 p1 )(1 p2 ).
Q
So we have proved
d 1 1 1
Theorem 3.8. g(X ) = 1 + e2 e3 e .
12 4 3 2
Next time we will compute this X0 (N ), somewhat the most useful case. In this case,
the elliptic points are interesting. We will apply the so-called moduli interpretation using
elliptic curves.
In general it is hard to calculate eh . It is easy if is a normal subgroup, e.g. (N ), because
every point in a fiber will either be simultaneously elliptic or non-elliptic. But 0 (N ) is not
normal.
(
0 if 2 | N or 9 | N,
e3 (0 (N )) = Q
3
p|N 1+ p
otherwise,
(
0 if (Q( 3), N ) does not satisfy the Heegner condition,
=
2#{odd p|N } if (Q( 3), N ) satisfies the Heegner condition,
13
4 3
= 0; otherwise p is the Legendre symbol, and
where we define 2
= 0 and 3
X
N
e (0 (N )) = gcd d, ,
d
d|N
P
where (n) = c|n 1.
Definition 4.2. The Heegner condition for (K, N ) means for every p | N , either p is split
in K or p is ramified with p || N .
Remark. This condition is important for the development of the Gross-Zagier formula.
To prove these formulas it is better to use the moduli interpretation of Y0 (N ), at least
as a set.
For N = 1, we can give meaning to SL2 (Z)\H as the moduli of elliptic curves over C, i.e.
the set of 1-dimensional complex tori C/ for some lattice modulo isomorphism classes as
Riemann surfaces. Equivalently, this is the set of lattices C modulo homothety, namely
1 2 if and only if there exists C such that 1 = 2 .
Indeed, we can establish
a bijection by sending SL2 (Z)\H to = Z + Z . Replacing
a b
by where SL2 (Z), the lattice becomes
c d
a + b 1 1
Z+Z = (Z(c + d) + Z(a + d)) = (Z + Z ),
c + d c + d c + d
i.e.
1
=
c + d
and so the map is well-defined. Bijectivity can then be verified easily.
Remark. A function f on H satisfying the weight k condition
f ( ) = (c + d)k f ( )
can be interpreted as a function g on the set of lattices of homogeneous degree k
g() = k g().
Given g, we can define f on H by f ( ) = g( ). Then
1
f ( ) = g( ) = g( ) = (c + d)k g( ) = (c + d)k f ( )
c + d
and vice versa.
We can further interpret this as a function h on the set of equivalence classes (E, ) of
elliptic curves with a 1-form 0 6= H 0 (E, 1E/C ) such that
h(E, ) = k h(E, ).
Given a function h, how do we get f ? To any given H we associate the elliptic curve
E = C/ and the canonical 1-form dz, which descends to E because it is invariant un-
der translation, and define f ( ) = h(E, dz). Then C/ c1+d is isomorphic to C/ via
multiplication by c + d, so
1
h(C/ , dz) = h(C/ , dz) = h(C/ , (c + d)dz) = (c + d)k h(C/ , dz),
c + d
14
i.e.
f ( ) = (c + d)k f ( ).
Note that we have not put any holomorphic conditions on H or at the cusps. We are just
trying to interpret the weight k condition by considering the points of the modular curves
as a set.
4.2. Elliptic curves with level structures. By definition, an elliptic curves with (N )-
level structure ( = 0, 1 or nothing) is a pair (E, ) where
cyclic subgroup of order N (of E[N ])
if = 0,
3
= order N point (of E[N ]) if = 1,
a Z/N -basis of E[N ]: (, ) E[N ]2 with Z/N + Z/N = E[N ] if = nothing,
where E[N ] is the subgroup of N -torsion. If E = C/ with group structure induced by the
complex numbers, then E[N ] = N1 / = (Z/N )2 as abelian groups.
The equivalence relation is given in the obvious way: two pairs (E1 , G1 ) and (E2 , G2 ) are
isomorphic if there is an isomorphism E1 E2 with (G1 ) = G2 .
Lemma 4.3. There is a bijection between Y (N ) and {(E, )} modulo equivalence given by
(N ) 7 (C/ , )
where
1
N +
if = 0,
= N1 + if = 1,
( 1 , )
if = nothing.
N N
1
(For E = C/ , E[N ] = /
N
= ( N1 Z + 1
N
Z )/ .)
Proof (Sketch). Let us check this
is well-defined
for the 0 (N )-level structure, which is the
a b
only case we will use. Let = 0 (N ). We have
c d
1 1 c +d c + d d 1
C/ , C/ , = C/ , = C/ ,
c + d N N N N
since N | c and gcd(d, N ) = 1. In fact the same calculation proves injectivity as well: if
and go to the same pair, then must be in 0 (N ). Surjectivity can be proved by
changing basis.
Now we return to the interpretation of elliptic points of X0 (N ) or Y0 (N ). The correspon-
dence above implies that the stabilizer group of any point in the orbit 0 (N ) is isomorphic
to
0 (N )
= Aut(E, G)
where G is an order N cyclic subgroup. In the case N = 1, this corresponds to the fact that
Aut(E) can only be Z/2, Z/4, Z/6, with the last two cases occurring when E has complex
multiplication structure by the Gaussian integers Z[i] or the Eisenstein integers Z[e2i/3 ]
respectively. In general we have
3A (N )-level structure is also called the full level structure. In fact the basis (, ) is required to have
Weil pairing equal to a fixed root of unity. See Section 1.5 of Diamond-Shurman.
15
Theorem 4.4. There is a bijection between the set of elliptic points of X0 (N ) of order
h = 2 and the set of ideals N Z[i] such that Z[i]/N = Z/N , and a bijection between the
set of elliptic points of X0 (N ) of order h = 3 and the set of ideals N Z[e2i/3 ] such that
Z[e2i/3 ]/N = Z/N .
Thus the number of elliptic points is equal to the number of ideals with certain properties,
which are easy to count since Z[i] and Z[e2i/3 ] are Dedekind domains (even PID!). Assuming
this, we can give a proof of Theorem 4.1.
Proof of Theorem 4.1 (Sketch). It suffices by the Chinese Remainder Theorem to consider
N = pa . For a quadratic field K, we want to find ideals N such that OK /N = Z/pa is cyclic.
There are three situations, depending on whether p is ramified, inert or split.
If p is ramified, then p = p2 and N = pb . If b 2, then p | N and OK /N cannot be cyclic,
since OK /(p) is (Z/p)2 . Thus N = p. Now p must exactly divide N because OK /p = Z/p.
2
If p is inert, then pOK is a prime ideal with OK /(p) = (Z/p) , so OK /N cannot be cyclic.
If p is split, then pOK = pp. Then the only choice of N is either pa or pa , because if
N = pi pj then it is contained in (p)min(i,j) , but Ok /(p) is not cyclic.
This is how we get precisely the factor 1 + ( dp ), where d is the discriminant of K.
The cusps of P1 (Q) are counted by brute-force.
Now we explain how to get the bijectivity in Theorem 4.4. The key point is that the extra
automorphisms of the pair (E, G) force certain lattices to be fractional ideals.
Proof of Theorem 4.4 (Sketch). In the order 2 case, the lattice is = Z[i] = OK for the
quadratic field K = Q(i). Our goal is to count the number of subgroups G C/OK such
that G is cyclic of order N and the pair (C/OK , G) admits non-trivial automorphisms.
Since the automorphism group of the lattice OK is generated by multiplication by i (iOK =
OK = Z[i]), we want [i] : C/OK C/OK to preserve G. Writing G = M/OK , where
M K = Q OK is contained in N1 OK , the condition is translated as iM = M. Hence
M is a Z[i]-module of K, i.e. a fractional ideal. Therefore, it suffices to count the number
of ideals N = M1 OK such that OK /N = M/OK = Z/N .
Corollary 4.5. If N = p is a prime, then
(
b p+1
12
c 1 if p 1 (mod 12),
g(X0 (p)) = p+1
b 12 c otherwise.
Note
the congruence condition on p precisely depends on its decomposition in Q(i) and
Q( 3).
Corollary 4.6. g(X0 (N )) = 1 if and only if N {11, 14, 15, 17, 19, 20, 21, 24, 27, 32, 36, 49}.
These are important because they are both modular curves and elliptic curves.
This immediately implies the second requirement that every double coset has a represen-
tative fixed by .
Note defines an action on C[G//H] as well by (g) = (g ), and is an anti-involution
( ) = .
Since each double coset has an invariant element, acts trivially on C[G//H]. Combining
these, we conclude that = , so C[G//H] is commutative.
4The hypothesis on (G, H) implies every double coset HgH is a finite union of cosets in H\G. See Lemma
5.1.2 of Diamond-Shurman.
19
Remark. (G, H) is called a Gelfand pair, i.e. for all irreducible representations of G, the
space HomH (, C) is at most 1-dimensional.
How do we deal with Example 6.1 where (G, H) = (GL2 (Q)+ , SL2 (Z))? We will only
consider the subalgebra generated by functions supported in M2 (Z) GL2 (Q)+ . This set is
certainly bi-H-invariant.
Theorem 6.5. This subalgebra is commutative.
Lemma 6.6. For n 1, M [n] = M2 (Z)[n] = { M2 (Z) : det = n} is equal to the
disjoint unions
a 0 a n/ x
=
0 0
|>0 |n
=n 0x<
where = SL2 (Z). In particular, we have
X
#\M [n] = = 1 (n).
|n
(In order to find a double coset representative for any matrix in M [n], note that will be
the gcd of the four entries.)
Proof (Sketch). The same proof works by taking to be transposition again.
Consider the characteristic function 1M [n] for n 1.
Theorem 6.7.
(1) 1M [n] is multiplicative, i.e. 1M [n] 1M [m] = 1M [nm] if gcd(n, m) = 1.
(2) For p prime, 1M [p] 1M [pn ] = 1M [pn+1 ] + p1pM [pn1 ] as equality in C[G//H].
Proof.
(1) We can write
M [nm] = M [n] M [m]
in an essentially unique way: if = n m = n0 m 0
, then n01 n = m 0 1
m has deter-
1 1
minant 1 and is contained in n M2 (Z) m M2 (Z) = M2 (Z), hence is in = SL2 (Z).
0
Thus m m differ by an element in .
(2) We can easily check 1M [p] 1M [pn ] 1M [pn+1 ] + p1pM [pn1 ] , so it is enough to prove
they have the same integral, i.e.
vol(M [pn ]) vol(M [p]) = vol(M [pn+1 ]) + p vol(M [pn1 ])
1 (pn )1 (p) = 1 (pn+1 ) + p1 (pn1 )
pn+1 1 p2 1 pn+2 1 pn 1
= +p .
p1 p1 p1 p1
Today we will study Mk and Sk for the full modular group = SL2 (Z). For GL2 (Q)+ ,
define the weight-k -operator
f []k ( ) = (f |, k)( ) = (det )k1 j(, )k f ( )
a b
where for = , j(, ) = c + d is the automorphy factor. Using this notation, the
c d
weight-k condition for SL2 (Z) means f []k = f . We can verify that f []k = (f []k )[]k .
20
The underlying fact is that the automorphy factor satisfies the cocycle condition j(, ) =
j(, )j(, ).
For C[GL2 (Q)+ //], define
X
f []k = f []k ()
\ GL2 (Q)+
Z
which is a finite sum. Again this can be interpreted as a formal integral f []k ()d,
GL2 (Q)+
by putting the counting measure on \ GL2 (Q)+ .
Lemma 6.8. []k defines operators Mk Mk and Sk Sk .
Proof (Sketch). If f []k = f for all , then f []k satisfies the same property. We can
also check holomphicity at the cusp.
Definition 6.9. The Hecke operator Tn is defined to be the operator []k for = 1M [n] .
Thus Tn is contained in End(Mk ) and End(Sk ). Note that = 1pM [pn1 ] corresponds to
k2
p Tpn1 because
f [p]k = det(p)k1 j(p, )k f ( ) = p2(k1) pk f []k
since and p act on H in the same way. With this, Theorem 6.7 translates into
Theorem 6.10.
(1) Tn Tm = Tnm for gcd(n, m) = 1.
(2) For p prime, Tp Tpn = Tpn+1 + pk1 Tpn1 .
Lemma 6.11 (Effect on q-expansion). If f = n
P
n=0 an (f )q Mk , then Tm f Mk with
n-th Fourier coefficient given by
X
an (Tm f ) = dk1 anm/d2 (f ).
d|(n,m)
Hence X
Tm (f )( ) = mk1 j(, )k f (( ))
a b a +b
where runs through the coset representatives . Since ( ) = , e2in( ) =
0 d d
a +b
e2in d and j(, ) = d, we have
d1
a b
X X X
Tm (f )( ) = mk1 dk an e2in( d + d )
ad=m n=0 b=0
a
X X
k1 k
=m d dan e2in( d )
ad=m d|n
X X m k1 0
= adn0 e2in a
ad=m n0 =0
d
X X
= qn ak1 adn0 .
n=0 n0 a=n
ad=m
Example 7.3 (Eisenstein series). Recall that the Eisenstein series of weight k is given by
X
k (2i)k X
Gk ( ) = w = 2(k) + k1 (n)q n
(k 1)! n=1
w {0}
22
where = Z + Z . It turns out to be an eigenform for Tm , i.e.
Lemma 7.4. Tm (Gk ) = k1 (m)Gk .
Proof. It is clear that if Gk is indeed an eigenform, then the eigenvalue must be k1 (m) by
comparing the constant terms. In fact we can directly check the formula
X nm
dk1 k1 = k1 (m)k1 (n).
d2
d|(n,m)
Remark. Recall that functions on H satisfying the weight-k condition can be interpreted
as functions on lattices of homogeneous of degree k. With this, we can give a simpler
definition of Hecke operator
X
(Tn f )() = nk1 f (0 ).
0
[:0 ]=n
(2) Equivalently, in the language of elliptic curves, viewing f : {(E, G, )}/ C gives
X
Tn (f )(E, G, ) = nk1 f (E/H, (G + H)/H, H )
HE subgroup
of order n
To address convergence issues, note that we have the following trivial bounds on Fourier
coefficients.
Lemma 8.6. If f is a cusp form, then |an | Cnk/2 for some constant C. In general, we
have |an | Cnk .
Proof. For simplicity, assume = (N ) where = 0, 1, so the local parameter is given by
q = e2i (for general congruence subgroups we need to take into account the width of at
). Then
Z Z 1
1 n dq
an = f (q)q = f (x + iy)e2in(x+iy) dx
2i q 0
is independent of y > 0 by holomorphicity. Choosing y = n1 , we have
Z 1
i
an = e 2
f x+ e2inx dx
0 n
27
Recall that |f ( )|2 | Im( )|k is invariant under . If f is cuspidal, |f ( )|2 | Im( )|k C, so
|f (x + ni )| C( n1 )k/2 = Cnk/2 and so
|an | Cnk/2 .
For the Eisenstein series Ek for SL2 (Z), we have
X
an = k1 (n) = dk1 Cnk .
d|n
Since any modular form for SL2 (Z) can be written as a sum of an Eisenstein series and a
cusp form, we obtain the trivial bound.
We omit the proof for general congruence subgroups.
(
1 + k2 if f is cuspidal and = 0, 1,
Corollary 8.7. L(s, f ) converges for Re(s) >
1 + k if f Mk ().
We can consider the Euler product for L(s, f ), just like for the Riemann zeta-function.
By definition, we say f is normalized if a1 = 1.
Theorem 8.8. For f Sk (SL2 (Z)), the following are equivalent:
(1) f is a normalized eigenform;
(2) L(s, f ) has an Euler product
Y
L(s, f ) = (1 ap ps + pk12s )1 .
p prime
The property of having an Euler product is obviously not stable under addition. Being an
eigenform means having a simple spectrum. Euler product thus reflects the spectrum.
Proof. (1) (2). Suppose f is an eigenform with a1 = 1. For Tm where m Z1 , recall
that X
an (Tm (f )) = dk1 anm/d2 .
d|(m,n)
This implies that the eigenvalue of Tm is am . Since Tm Tn = Tmn for (m, n) = 1 and
Tpi+1 = Tp Tpi pk1 Tpi1 , we have the same relations
(
am an = amn if (m, n) = 1,
k1
(2)
api+1 = ap api p api1 if i 1.
Hence,
!
Y X
L(s, f ) = apk pks .
p prime k=1
pi
The sequence bi = a has recursive relation bi+1 = ap bi pk1 bi1 , hence characteristic
polynomial T 2 ap T + pk1 . Thus
X
apk pks = (1 ap ps + pk12s )1 .
k=1
28
(2) (1). The same proof goes backwards. The Euler product implies that the Fourier
coefficients an satisfy the relations (2). It suffices to verify Tp f = ap f for primes p (since Tp
generate the Hecke algebra), which is equivalent to
an (Tp f ) = an ap
for all n, i.e.
X
dk1 anp/d2 = an ap .
d|(n,p)
(
anp if (n, p) = 1,
But the left hand side is just which is equal to an ap by (2).
anp + pk1 an/p if p | n,
We use the same recursive relations as above to define Tn , and define hdi = 0 if (d, N ) > 1.
Remark. All the above relations can be combined into
X Y Y
Tn ns = (1 Tp ps + hpipk12s )1 (1 Tp ps )1 .
n=1 p-N p|N
29
Therefore, we have defined Tn and hni for all n 1, which are endomorphisms of Sk ().
The Hecke algebra T generated by them is still a commutative C-algebra, but not necessarily
semisimple.
Lemma 9.1. Let (n, N ) = 1. Then the adjoint of Tn with respect to the Petersson inner
product is
Tn = hni1 Tn .
In particular, Tn is a normal operator if (n, N ) = 1.
This5 suggests we should look at the subalgebra T0 T given by
T0 = C[Tn , hni](n,N )=1 ,
which is semisimple by the lemma.
What happens to Tp = Up when p | N ? We can still find a formula for its adjoint. Recall
1 0
that Tp is associated to the double coset , so Tp is associated to the double coset
0 p
p 0
. These two are in general not the same double coset. Let us introduce one more
0 1
operator.
For = (N ), define the normalizer
N () = NGL2 (R)+ () = { GL2 (R)+ : 1 = }.
The quotient N ()/ is interesting because it induces automorphisms on the modular curve,
i.e. we have a map
N ()/ Aut(\H ).
It is non-trivial to find a normalizer, but here is one. Let
0 1 1 0 0 1
wN = =
N 0 0 N 1 0
which is contained in N (), so that the double coset wN is just a coset wN . Then we
can check that f 7 f [wN ]k maps into the same space of cusp forms, so we can consider
wN = [wN ]k End(Sk ()).
Since wN2
= N , we see that wN is an involution on the modular curve \H , but it is
not quite an involution on Sk ().
Lemma 9.2.
2
(1) wN acts by (1)k N k2 on Sk ().
(2) For all n, we have (
1
Tn = wN Tn wN ,
1
hni = wN hniwN .
Proof.
2
(1) wN = [wN2
]k = (N 2 )k1 (N )k = (1) k k2
N .
1 0 p 0
(2) The key point is that the matrices and differ by conjugation by wN .
0 p 0 1
5We can also check that hni = hni1 for (n, N ) = 1, so hni is indeed normal.
30
Corollary 9.3. For = 0 (N ) and (n, N ) = 1, we have
1
Tn = Tn = wN Tn wN .
Thus wN commutes with Tn for (n, N ) = 1.
9.2. Atkin-Lehner theory. To understand the relationship between T0 and T, we need the
Atkin-Lehner theory (also known as the newform theory), and later we will give an example
where T is not semisimple. The idea is to study the newforms of Sk (1 (N )), which are the
primitive forms in some sense.
When M | N , we always
have the inclusion Sk (1 (M )) , Sk (1 (N )). Consider the span
d 0 N
of f (d ) = d1k f where f Sk (1 (M )), M | N and d | M . In fact it is enough
0 1 k
to consider forms of level dividing N/p where p | N is a prime:
Skold (1 (N )) = span{f (d ) : f Sk (1 (N/p)), d | p, p | N } Sk (1 (N ))
and define the newspace
Sknew (1 (N )) = Skold (1 (N )) .
Note that we are not calling this the space of newforms yet. Newform will have a specific
meaning it is an eigenform.
Atkin and Lehner proved that the newspace behaves like the level 1 case, where we just
need to look at T0 . Before we make this more precise, let us prove a few lemmas.
Lemma 9.4. The spaces Skold and Sknew are T-stable, where T is the full Hecke algebra.
Proof (Sketch). Introduce T , the algebra generated by the adjoints Tn and hni . It suffices
1
to prove that Skold is both T- and T -stable. But T = wN TwN , so it is enough to prove Skold
is T-stable and wN -stable.
Define the operator Vp : Sk (1 (N/p)) P Sk (1 (N )) by f 7 f (p ). This is a one-sided
inverse of the Up operator. Indeed, for f = n
n=0 an q ,
X
X 1 j 1
X +j
Up (f ) = f =p f = an q n/p
0 p k p
0jp1 0jp1 p|n
whereas
X
Vp (f ) = an q pn .
n=0
It is clear that Up Vp = Id.
It remains to show that the sum of the images of Sk (1 (N/p)) , Sk (1 (N )) and Vp :
Sk (1 (N/p)) Sk (1 (N )) is stable under the operators Tq , hqi and w for all primes q.
For example, consider the diagram (note the two Tp s are different!)
Vp
Sk (1 (N/p)) / Sk (1 (N ))
Tp Tp
Vp
Sk (1 (N/p)) / Sk (1 (N ))
which is not quite commutative. Since p | N , the Tp on the right is equal to Up and so
Tp Vp (f ) = f for all f Sk (1 (N/p)). This shows Tp preserves Skold .
31
The other operators for other primes can be checked similarly6.
Example 9.5. For p - N , Up acts on Sk (0 (p3 N )). Let f Sk (0 (N )) be an eigenform
of Tp . Consider the space spanned by fi = f (pi ) for 0 i 3. We claim that Up acts
non-semisimply on it.
Since Vp fi = fi+1 for 0 i 2, we have Up fi = fi1 for 1 i 3. Using the formula
X
Tp (f ) = an q n/p + pk1 hpif (p )
p|N
and the fact that f = f0 is an eigenform, we see that there exists C such that7
f0 = Tp f0 = Up f0 + pk1 f1 ,
i.e.
Up f0 = f0 pk1 f1 .
1
pk1 0 1
Thus Up has matrix , which is not semisimple8.
0 1
0
Theorem 9.6 (Atkin-Lehner). If f Sknew (1 (N )) is an eigenform for T0 , then f is also
an eigenform for the full Hecke algebra T.
We are not going to prove this, since the proof is just some combinatorics and not inspi-
rational. There is a more conceptual proof using representation theory.
Definition 9.7. f Sk (1 (N )) is a newform if it is contained in Sknew (1 (N )), an eigenform
for T and normalized such that a1 (f ) = 1.
Newforms have a rigid meaning. In particular, newforms do not form a vector space, just
a set!
Corollary 9.8 (Multiplicity one). If f is a newform with eigenvalue : T C, then
dim{g Sk (1 (N )) : T g = (T )g for all T T} = 1.
and
1k p 0 X
Vp f = p f = an q np
0 1 k
n1
so that Up is a left inverse to Vp .
The guiding example is as follows. For p - N , let f 6= 0 be an eigenform for Tp of level N ,
with Tp f = f . Consider the span of ei = (Vp )i f where i = 0, 1, , d, which are linearly
independent. Then Up , as a Hecke operator of level N pd , is given by
(
Up ei = ei1 for 1 i d,
k1
Up e0 = e0 p e1 ,
so its matrix is of the form
1
pk1 0 1
Up,d = .
0 1
0
If d 3, then this matrix is not diagonalizable, i.e. the action of Up is not semisimple. This
follows from an easy exercise in linear algebra.
Exercise. The minimal polynomial of Up,d is X d1 (X 2 X + pk1 ).
The Atkin-Lehner theory takes into account these operators Up .
Lemma 10.1. The map T Sk () C defined by (T, f ) 7 a1 (T f ) is a perfect pairing.
Proof. Recall that a1 (Tm f ) = am (f ).
If (T, f ) = 0 for all T , then in particular (Tn , f ) = 0, so an (f ) = a1 (Tn f ) = 0 for all n and
f = 0.
If (T0 , f ) = 0 for all f Sk (), we want to show T0 = 0, i.e. T0 f = 0. Indeed, for all n,
an (T0 f ) = a1 (Tn T0 f ) = a1 (T0 Tn f ) = (T0 , Tn f ) = 0.
If f Sk () is an eigenform of T, then T f = f (T )f satisfies f (Tn Tm ) = f (Tn )f (Tm )
and so defines a character f : T C. Define Sk ()[f ] to be the f -eigenspace, i.e.
Sk ()[f ] = {g Sk () : T g = f (T )g for all T T}.
Lemma 10.2 (Multiplicity one). If f Sk () is an eigenform of T, then dim Sk ()[f ] = 1.
Proof. For g Sk ()[f ], we have an (g) = a1 (Tn g) = f (Tn )a1 (g) for all n 1.
Theorem 10.3 (Atkin-Lehner).
(1) If f Sknew () is an eigenform of T0 , then it is an eigenform of T.
(2) Let f be a newform of level Nf dividing N , and Sf be the span of the linearly inde-
pendent elements Vd f = f (d ) where d | N/Nf . Then
Sf = {g Sk ( (N )) : T g = f (T )g for all T T0 }.
In particular, Sf is stable under T.
33
(3) There is a decomposition
M M
Sk ( (N )) = Sf
M |N f newform
of level M
as Hecke modules.
Remark.
P By Lemma 10.2, if M = N , then dim Sf = 1. In general, Sf has dimension
d|N/Nf 1.
and similarly
Mk,R () = {f Mk,C () : an (f ) R for all n}.
Today let us only consider the level 1 case = SL2 (Z). Recall the graded algebra
M
Mk,C = C[E4 , E6 ]
kZ
k
P
where Ek = 1 Bk n=1 k1 (n)q n for k 4 even. Since Ek has rational coefficients, we
immediately get
The basis element E4a E6b is contained in Mk,Q , and this space is stable under TQ since
X
an (Tm f ) = dk1 anm/d2 (f ).
d|(n,m)
Thus TQ is a subalgebra of EndQ (Mk,Q ), hence finitely generated over Q, i.e. a finite-
dimensional Q-vector space.
Similarly, we have Sk,Q C
= Sk,C , and TQ Sk,Q Q is a perfect pairing.
10.3. Plan. The plan before spring break is to study the rational and integral structures of
spaces of modular forms, and prove a result of Shimura on the algebraicity of special values
of L-functions. We will start discussing p-adic modular forms after the break.
36