Furthermore, for η
2
0, 0 θ ≤ C
2
and τ
N
≤ T a stopping time, for all 1 ≤ j ≤ J,
P Z
τ
N
+θ τ
N
D η
N ,
ω s
, L
ν
N
s
ϕ
j
E ds
≥ η
2
≤ 1
η
2 2
E
Z
τ
N
+θ τ
N
D η
N ,
ω s
, L
ν
N
s
ϕ
j
E ds
2
,
≤ C
2
η
2 2
E
Z
τ
N
+θ τ
N
D η
N ,
ω s
, L
ν
N
s
ϕ
j
E
2
ds
,
≤ C
2
η
2 2
Z
T
E D
η
N ,
ω s
, L
ν
N
s
ϕ
j
E
2
ds, ≤
C C
2
η
2 2
Z
T
E
h η
N s
2 −3,2α
i ds,
≤ C T C
2
η
2 2
A
N
, cf. 25. And,
P
M
N τ
N
+θ
ϕ
j
− M
N τ
N
ϕ
j
η
2
≤
C C
2
η
2 2
1 N
N
X
i= 1
1 + |ω
i
|
4 α
. So, for all j
≥ 1, by definition of C
2
, P
η
N ,
ω τ
N
+θ
ϕ
j
− η
N ,
ω τ
N
ϕ
j
≥ η
2
≤
η
1
A ǫ
A
N
+ 1
N
N
X
i= 1
1 + |ω
i
|
4 α
. Consequently,
Θ
N
K
ǫ 2
φ
1
, . . . , ϕ
J
≥ P A
N
+ 1
N
N
X
i= 1
1 + |ω
i
|
4 α
Aǫ .
Letting J → ∞, we get lim sup
N
Θ
N
S
J
K
ǫ 2
ϕ
1
, . . . , ϕ
J c
≤ ǫ. Eq. 33 is proved.
4.3.3 Identification of the limit
The proof of the fluctuations result will be complete when we identify any possible limit.
Proposition 4.13 Identification of the initial value.
The random variable ω 7→ L
η
N ,
ω
converges in law to the random variable ω 7→ L X ω,
where for all ω, X ω = Cω + Y , with Y a centered Gaussian process with covariance Γ
1
. Moreover ω 7→ Cω is a Gaussian process with covariance Γ
2
, where Γ
1
and Γ
2
are defined in 8 and 9.
Proof. For simplicity, we only identify here the law of
D η
N ,
ω
, ϕ
E for all
ϕ. The same proof works for the law of finite-dimensional distributions
D η
N ,
ω
, ϕ
1
E , . . . ,
D η
N ,
ω
, ϕ
p
E , p
≥ 1. We write
820
Γ
i
for Γ
i
ϕ, ϕ, i = 1, 2. One has: D
η
N ,
ω
, ϕ
E =
1 p
N
N
X
i= 1
ϕξ
i
, ω
i
− Z
S
1
ϕx, ω
i
λ dx
+ 1
p N
N
X
i= 1
Z
S
1
ϕx, ω
i
λ dx − ν
, ϕ
,
=: A
N ,
ω
+ B
N ,
ω
. It is easy to see that B
N ,
ω
converges in law to Z
2
∼ N 0, Γ
2
. Moreover, for P-almost every ω, A
N ,
ω
converges in law to Z
1
∼ N 0, Γ
1
see Billingsley [5], Th. 27.3 p. 362. That means that for all u
∈ R, ψ
A
N
u := E
λ
e
iuA
N ,
ω
converges to ψ
Y
u := e
−
u2 2Γ1
. But, then, for all F ∈ C
b
R, E
F E
λ
e
iu D
η
N ,
ω
, ϕ
E
= E h
F E
λ
h e
iuA
N ,
ω
+B
N ,
ω
ii = E
h F
e
iuB
N ,
ω
ψ
N
u i
. Since
ψ
N
u converges almost surely to a constant, the limit of the expression above exists Slutsky’s theorem and is equal to E
F e
iuZ
2
−
u2 2Γ1
.
Proposition 4.14 Identification of the martingale part. For P-almost every ω, the sequence M
N ,
ω
converges in law in C [0, T ], S
′
to a Gaussian process W with covariance defined in
7. Proof.
For fixed ω, M
N ,
ω
is a sequence of uniformly square-integrable continuous martin- gales cf. Remark 4.7, which is tight in
C [0, T ], S
′
. Let W
1
and W
2
be two accumulation points continuous square-integrable martingales which a priori depend on
ω and M
φN ,ω
and M
ψN ,ω
be two subsequences converging to W
1
and W
2
, respectively. Note that we can suppose that
φN ≤ ψN for all N. For all ϕ ∈ S , lim
N →∞
¬ M
φN ,ω
ϕ , M
ψN ,ω
ϕ ¶
t
= W
1
ϕ , W
2
ϕ
t
, for all t, and ¬
M
φN ,ω
ϕ , M
ψN ,ω
ϕ ¶
t
= Z
t
¬ ν
φN s
, ϕ
′ 2
¶ ds.
We now have to identify the limit: we already know that for P-almost every realization of the disorder
ω, ν
N ,
ω
converges in law to P. But, the latter expression, seen as a function of ν, is continuous. So
W
1
ϕ , W
2
ϕ
t
= R
t
¬ P
s
, ϕ
′ 2
¶ . So W
1
− W
2
is a continuous square integrable martingale whose Doob-Meyer process is 0. So W
1
= W
2
and is characterized as the Gaussian process with covariance given in 7. The convergence follows.
Proof of the independence of W and X . We prove more : the triple Y, C, W is independent. For sake
of simplicity, we only consider the case of Y ϕ, Cϕ, W
t
ϕ for fixed t and ϕ. Let us first recall some notations: let A
N ,
ω
, B
N ,
ω
and M
N ,
ω t
ϕ be the random variables defined in the proof of Proposition 4.13 and 4.14 and let
ψ
A
N
u := E e
iuA
N ,
ω
, ψ
B
N
v := E e
i vB
N ,
ω
, ψ
M
N
w := E e
iw M
N ,
ω t
ϕ
be their characteristic functions u, v, w ∈ R. We know that, for almost
821
every ω, ψ
A
N
u converges to ψ
Y
u = e
−
u2 2Γ1
and that ψ
M
N
w converges to the deterministic function
ψ
W
w := E
e
iwW
t
ϕ
. But, if
ψ
C
v = E
e
iwC
, then, for all u, v, w
∈ R, using the
independence of the Brownian with the initial conditions, E
E e
iuA
N ,
ω
+i vB
N ,
ω
+iw M
N ,
ω t
ϕ
− e
i vB
N ,
ω
ψ
A
N
uψ
M
N
w = 0.
Using Slutsky’s theorem, we see that any limit couple Y, C, W satisfies E
E
e
iuY +i vC+iwW
t
ϕ
= ψ
Y
uψ
C
vψ
W
w. which is the desired result.
We recall that the limit second order differential operator L
s
is defined by L
s
ϕ y, π := 1
2 ϕ
′′
y, π + ϕ
′
y, πb[ y, P
s
] + c y, π + P
s
, ϕ
′
·, ·b·, y, π .
As in Lemma 4.5, we can prove the following:
Lemma 4.15. Assume
H
b ,c
. Then for every N , s ≤ T , ω, the operator L
s
and L
ν
N
s
are linear continuous from S
to S and
L
s
ϕ
6, α
≤ C ϕ
8, α
, L
ν
N
s
ϕ
6, α
≤ C ϕ
8, α
. We are now in position to prove Theorem 2.10:
Proof of Theorem 2.10. Let Θ be an accumulation point of Θ
N
. Thus, for a certain subsequence which will be also denoted as N for notations purpose, the random variable
ω 7→ H
N ,
ω
con- verges in law to a random variable
H with values in M
1
C [0, T ], S
′
with law Θ. Applying Skorohod’s representation theorem, there exists some probability space Ω
1
, P
1
, F
1
and ran- dom variables defined on Ω
1
, ω
1
7→ H
N
ω
1
and ω
1
7→ Hω
1
such that H
N
has the same law as ω 7→ H
N ,
ω
, H has the same law as
H , and for P
1
-almost every ω
1
∈ Ω
1
, H
N
ω
1
converges to H
ω
1
in M
1
C [0, T ], S
′
.
An easy application of Proposition 4.8 and Borel-Cantelli’s Lemma shows that P
1
-almost surely,
E sup
t ≤T
η
ω
1
t −6,α
∞. Then we know from Lemma 4.15 that the integral term R
t
L
∗ s
η
ω
1
s
ds makes sense as a Bochner’s integral in W
−8,α
⊆ S
′
. Let
η
N ,
ω
1
with law H
N
ω
1
; η
N ,
ω
1
converges in law to some η
ω
1
with law H ω
1
. By uniqueness in law convergence, using Propositions 4.13 and 4.14, we see that
η
ω
1
, W as the same law as X ω
1
, W . For fixed ϕ ∈ S , we define F
ϕ
from C [0, T ], S
′
into R by F
ϕ
γ := γ
t
, ϕ
− γ
, ϕ
− R
t
γ
s
, L
s
ϕ ds. The function F
ϕ
is continuous and since η
N ,
ω
1
converges in law to η
ω
1
, the sequence F
ϕ
η
N ,
ω
1
converges in law to F
ϕ
η
ω
1
. To prove the theorem, it remains to show that the law of the term
R
t
D η
N ,
ω
1
s
, L
ν
N
s
ϕ − L
s
ϕ E
ds converges in law to 0. We show that there
822
is convergence in probability: For all ǫ 0, for all A 0, using Proposition 4.8, Lemma 4.15, and
Cauchy-Schwarz’s inequality, U
N ,
ǫ
:= P
1
E
Z
t
D η
N ,
ω
1
s
, L
ν
N
s
− L
s
ϕ E
ds
ǫ
, = P
E
Z
t
D η
N ,
ω s
, L
ν
N
s
− L
s
ϕ E
ds
ǫ
, ≤ P
Z
t
E
η
N ,
ω s
2 −6,α
1 2
E L
ν
N
s
− L
s
ϕ
2 6,
α 1
2
ds ǫ
,
≤ P
C
N
ω
1
, . . . , ω
N 1
2
Z
t
E L
ν
N
s
− L
s
ϕ
2 6,
α 1
2
ds ǫ
cf. Prop 4.8,
≤ P Z
t
E L
ν
N
s
− L
s
ϕ
2 6,
α 1
2
ds ǫ
p A
+ P C
N
A .
Using 4.8, it suffices to prove that, for all ǫ 0,
lim sup
N →∞
P Z
t
E L
ν
N
s
− L
s
ϕ
2 6,
α 1
2
ds ǫ
= 0.
34 Indeed, for every
ϕ ∈ S , U
N s
ϕ y, π := L
ν
N
s
− L
s
ϕ y, π = ϕ
′
y, πb[ y, ν
N s
] − b[ y, P
s
]. An analogous calculation as in Lemma 4.5 shows that, using Lipschitz assumptions on b, and Propo-
sition 4.3:
E
sup
s ≤t
U
N s
ϕ
2 6,
α
≤ ϕ
2 8,
α
CN + D
N
ω
1
, . . . , ω
N
, with the property that lim
A →∞
lim sup
N
PN D
N
A = 0. Equation 34 is a direct consequence. Since there is uniqueness in law in 10, Θ is perfectly defined, and thus, unique. The convergence
follows.
5 Proofs for the fluctuations of the order parameters
We end by the proofs of paragraph 2.3.3.
5.1 Proof of Proposition 2.13