LECTURE NOTES ON OSCILLATORY INTEGRALS

Atelier d’Analyse Harmonique 2019

A large part of modern harmonic analysis is concerned with understanding can-

cellation phenomena happening between different contributions to a sum or inte-
gral. Loosely speaking, we want to know how much better we can do than if we had
taken absolute values everywhere. A prototypical example of this is the oscillatory
integral of the form
ˆ
eiφλ (x) ψ(x)dx

(solutions of dispersive PDEs often take this form). Here ψ, called the amplitude,
is usually understood to be “slowly varying” with respect to the real-valued φλ ,
called the phase, and thus the oscillatory behaviour is given mainly by the complex
exponential eiφλ (x) , where λ denotes a parameter or list of parameters and generally
φ0λ gets larger as λ grows, for example φλ (x) = λφ(x). Expressions of this form
arise quite naturally in several problems, as we will see in Section 1, and typically
one seeks to provide an upperbound on the absolute value of the integral above
in terms of the parameters λ. Intuitively, as λ gets larger the phase φλ changes
faster and therefore eiφλ oscillates faster, producing more cancellation between the
contributions of different intervals to the integral. We expect then the integral to
decay as λ grows larger, and usually seek upperbounds of the form |λ|−α . Notice
that if you take absolute values inside the integral above you just obtain kψkL1 , a
bound that does not decay in λ at all.
The main tool we will use is simply integration by parts. In the exercises you will
also use a little basic complex analysis to obtain more precise information on certain
special oscillatory integrals.
Notation: We will use A . B or A = O(B) to denote the estimate A ≤ CB where C > 0 is some
absolute constant, and A ∼ B to denote the fact that A . B . A. If the constant C depends on
a list of parameters L we will write A .L B or A = OL (B).
We will further denote by Cck the space of k-times differentiable functions with compact support.

1. Motivation
In this section we shall showcase the appearance of oscillatory integrals in analysis
with a couple of examples. You can find other interesting examples in the exercises,
see particularly Exercise 8, Exercise 14 part viii), Exercise 15 and Exercise
21.

1.1. Fourier transform of radial functions. Let f : Rd → C be a radially

symmetric function, that is there exists a function f0 : R → C such that f (x) =
f0 (|x|) for every x ∈ Rd . Let’s suppose for simplicity that f ∈ L1 (Rd ) (equivalently,
that f0 ∈ L1 (R, rd−1 dr)), so that it has a well-defined Fourier transform. It is easy
to see1 that fb must also be radially symmetric, that is there must exist g : R → C
such that fb(ξ) = g(|ξ|); we want to understand its relationship with f0 . Therefore

1By composing f with a rotation and using a change of variable in the integral defining fb.

1
2 OSCILLATORY INTEGRALS

we write using polar coordinates

ˆ
fb(ξ) = f (x)e−2πix·ξ dx
Rd
ˆ ∞ˆ
= f0 (r)e−2πirω·ξ rd−1 dσd−1 (ω)dr,
Sd−1
ˆ ∞
0
ˆ 
= f0 (r)r d−1
e−2πirω·ξ dσd−1 (ω) dr
0 Sd−1

where dσd−1 denotes the surface measure on the unit (d − 1)-dimensional sphere
Sd−1 induced by the Lebesgue measure on the ambient space Rd . By inspection,
we see that the integral in brackets above is radially symmetric in ξ, and so if we
define ˆ
J(t) := e−2πitω·e1 dσd−1 (ω),
Sd−1
with e1 = (1, 0, . . . , 0), we have
ˆ ∞
fb(ξ) = g(|ξ|) = f0 (r)rd−1 J(r|ξ|)dr. (1.1)
0

This is the relationship we were looking for: it allows one to calculate the Fourier
transform of f directly from the radial information f0 . You might also notice that
J is simply the Fourier transform of the spherical measure dσd−1 .
Now we claim that the function J is an example of oscillatory integral of the type
mentioned at the beginning. Indeed, observe that the inner product ω · e1 depends
only on the first component of ω; thus we write ω = (cos θ, ω 0 sin θ), with θ the
angle between ω and e1 and ω 0 ∈ Sd−2 . By factorising the spherical measure dσd−1
along the e1 axis, we can use this change of variables to write
ˆ ˆ π
J(t) = e−2πit cos θ (sin θ)d−2 dθdσd−2 (ω 0 )
d−2
S
ˆ π0
=cd−2 e−2πit cos θ (sin θ)d−2 dθ,
0
´
because the integrand does not depend on ω 0 (here cd−2 = Sd−2 dσd−2 (ω 0 ) =
σd−2 (Sd−2 )). It is now trivial to match the last expression to the one for an oscil-
latory integral where the parameter is t. Later on we will see how to estimate it in
terms of |t| and show that J(t) is bounded by a constant multiple of (1+|t|)−(d−1)/2 .
The information we get from how fast J(t) decays in terms of |t| tells us of
something interesting. Indeed, recall that if f ∈ L1 (Rd ) then its Fourier transform
fb is continuous. In general, when 1 < p ≤ 2 and f ∈ Lp (Rd ), we know by Hausdorff-
0
Young inequality that fb is in Lp (Rd ), but we don’t know whether it is continuous
or not. You can easily see this is generally not the case when p = 2. However, for
radial functions one can show continuity of the Fourier transform if the exponent
p is sufficiently close to 1! In particular
2d
Proposition 1.1. If 1 ≤ p < d+1 then the Fourier transform fb of any radial
p d
function f ∈ L (R ) is continuous away from 0.
This follows purely from the decay of the oscillatory integral J and you will prove
it in Exercise 1.

1.2. Counting the number of ways in which a number can be written

as the sum of two squares. Consider the following diophantine-type problem:
when can we write an integer n as the the sum of two squares, n = x2 + y 2 ? If you
 OSCILLATORY INTEGRALS 3

fiddle around with examples a bit you might find that certain numbers have such
a representation, say
5641 = 752 + 42 ,
while others, such as the successor 5642, have no such representation. Studying
this problem more in depth you might discover the Fibonacci identity
(a2 + b2 )(c2 + d2 ) = (ac − bd)2 + (ad + bc)2 ,
and deduce as a consequence that it suffices to study the case where n is prime. If
you went further down this road, you might learn that there is a way to write p
prime as x2 + y 2 if and only if p ≡ 1 mod 4, and this gives the full answer for the
problem: n = x2 + y 2 has a solution if and only if in the prime factorisation of n
the prime factors that are congruent to 3 mod 4 appear an even number of times
each.
However, after this, you might notice that some numbers have even more than one
representation as a sum of two squares: for example,
5645 = 742 + 132 = 672 + 342 .
The next step would then be to try and count how many distinct solutions there
are. It turns out that this can be done in terms of the exponents in the prime
factorisation of n (and the answer involves the Gaussian integers Z[i]); however,
this is not too convenient, because factorising large integers is knowingly hard. Let
us take a different, less algebraic approach instead.
We can generalise the question a little by considering sums of more than two
squares. Letting k ∈ N be fixed and given an integer n ∈ N, we define the k-th
sum-of-squares function to be
rk (n) := #{(x1 , . . . , xk ) ∈ Zk such that x21 + · · · + x2k = n};
in words, rk (n) is the number of ways in which n can be expressed as the sum of
k squares. Then we want to study the behaviour of the function rk (n) as n grows.
In the following we will consider only k = 2.
If you plot the graph of r2 you will quickly realise it is not a very regular function2.
A natural approach to deal with this irregularity is to study the averaged sequence
instead, hoping it will behave better. This should help us answer the question “in
how many ways can we expect to be able to write an arbitrary number n as a
sum of two squares on average?”. In formulas, we want to study the behaviour of
PN
n=1 r2 (n)/N as N tends to ∞. The limit of the expression above is easy to find:
the key is to notice that n = x2 + y 2 means that the point (x, y) ∈ Z2 belongs to
the circle of radius n1/2 centred at (0, 0), and therefore to the ball of radius N 1/2
with the same centre. Therefore
XN
r2 (n) = #{(x, y) ∈ Z2 such that x2 + y 2 ≤ N },
n=1
PN
and a simple geometric argument shows that the limit of N1 n=1 r2 (n) is π (see
Exercise 2). The geometric argument actually says something more: namely, we
can also give an upperbound on the error, that is
N
X
r2 (n) = πN + O(N 1/2 ).
n=1
The upperbound on the error is significant because it is of smaller order than the
main term (which grows like O(N )). The limit of the averaged sequence was very
easy to find, so now we ask a more sophisticated question: what is the behaviour
2Sequence A004018 in OEIS, https://oeis.org/A004018/graph.
4 OSCILLATORY INTEGRALS

of the error term? How much can the sum deviate from πN ? O(N 1/2 ) is what one
would expect from a uniformly random behaviour, but it turns out that it is far
from optimal, and we can do better. However, we do not know yet how much better
we can do! It is still an open problem3 to establish the true asymptotic behaviour
PN
of the error term E(N ) := n=1 r2 (n) − πN . Using oscillatory integral techniques,
we can show the improved estimate that follows.
Proposition 1.2. For N large enough, we have the error term estimate
|E(N )| = O(N 1/3 ).
If you stop for a moment to think about it, it is almost absurd that techniques
developed to study oscillatory integrals can say something highly non-trivial about
questions concerning the integers.
Proof. The idea is to use the Poisson Summation formula to reveal some interesting
oscillation hidden in the problem. In the end, the proof will again rely on a good
decay estimate for a certain oscillatory integral related to J.
Recall that, for f ∈ S(R2 ), that is a Schwartz function, the Poisson Summation
formula says that
X X
f (n) = fb(n).
n∈Z2 n∈Z2
P
The quantity we are interested in is Θ(N ) := n∈Z2 1B(0,N 1/2 ) (n), but 1B(0,N 1/2 )
(the characteristic function of a ball of radius N 1/2 ) is not in S(R2 ) - thus we have
to regularise it. Let δ > 0 be a small parameter to be chosen later ´ and let ϕ be a
bump function in C ∞ , supported in the unit ball and such´that R2 ϕ(x)dx = 1; we
let ϕδ denote the rescaled function δ −2 ϕ(δ −1 x) (so that ϕδ = 1 too). Then we
define
χN,δ := 1B(0,N 1/2 ) ∗ ϕδ ,
which is in S(R2 ) and is an approximation to 1B(0,N 1/2 ) . Applying the Poisson
Summation formula to this function, we obtain
X X
Θδ (N ) := χN,δ (n) = 1
bB(0,N 1/2 ) (0)ϕ(0)
b + 1
bB(0,N 1/2 ) (n)ϕ(δn);
b
n∈Z2 n6=0∈Z2

the first term is easily evaluated to be exactly πN , our main term! Thus the second
term is an error we have to control. Observe that, since 1B(0,N 1/2 ) is radial, we
have by (1.1) that
ˆ N 1/2
1
bB(0,N 1/2 ) (n) = rJ(r|n|)dr.
0
The function J is not only an oscillatory integral, but it is itself oscillating, and
with the techniques of the next section you will be able to show in Exercise 9 that
ˆ R
rJ(r)dr = O(R1/2 ). (1.2)
0
bB(0,N 1/2 ) (n)| . N 1/4 |n|−3/2 .
After a change of variables, estimate (1.2) shows that |1
−1
In the range |n| < δ we have |ϕ(δn)|b . 1 and therefore the above estimate gives
us
X X
|1
bB(0,N 1/2 ) (n)ϕ(δn)|
b . N 1/4 |n|−3/2 ∼ N 1/4 δ −1/2 .
|n|<δ −1 ,n6=0 |n|<δ −1 ,n6=0

3It is conjectured that for any  > 0 the error E(N ) is bounded by C N 1/4+ for some constant

C > 0. The estimate is certainly false with  = 0.
 OSCILLATORY INTEGRALS 5

In the range |n| ≥ δ −1 things are even better, since ϕ

b decays fast, being in C ∞ ; in
−1
particular, we have |ϕ(δn)|
b . (δ|n|) and thus, by the above same argument, this
range also contributes O(N 1/4 δ −1/2 ). Summarising, we have proven that
Θδ (N ) = πN + O(N 1/4 δ −1/2 ).
Now, observe that
Θδ ((N 1/2 − δ)2 ) ≤ Θ(N ) ≤ Θδ ((N 1/2 + δ)2 ); (1.3)
this is because, as you can easily see by expanding the convolution,
χ(N 1/2 −δ)2 ,δ ≤ 1B(0,N 1/2 ) ≤ χ(N 1/2 +δ)2 ,δ .
Therefore we have, thanks to (1.3) and the fact that (N 1/2 ± δ)2 ≈ N ± 2N 1/2 δ,
Θ(N ) − πN = E(N ) = O(N 1/4 δ −1/2 ) + O(N 1/2 δ),
and if we optimize by choosing δ = N −1/6 we see that both terms above are O(N 1/3 )
and we are done. 

2. Oscillatory integrals in one variable

In this section we present some techniques that allow one to estimate oscillatory
integrals when one has a lowerbound for some of the derivatives of the phase. We
will analyse objects of the form
ˆ b
I(λ) := eiλφ(x) dx
a
and more in general
ˆ b
Iψ (λ) := eiλφ(x) ψ(x)dx.
a
Before we start, let’s consider the following heuristic. Suppose that we have func-
tions f, g and we know that g is “slowly varying” - that is, the derivative g 0 is small.
´b
In order to find a good upperbound for an expression of the form | a f (t)g(t)dt|,
we could take advantage of the fact that g 0 is small by using integration by parts to
´b
estimate instead the expression − a F (t)g 0 (t)dt + (boundary terms), where F is a
primitive of f . What we gain is that now we are working with an integral for which
we know one of the factors of the integrand, g 0 , is small. We will use this idea and
variations thereof over and over.
2.1. Principle of non-stationary phase. We begin by considering the following
case. Let ψ ∈ C ∞ (R) have compact support in (a, b) (so that in particular ψ(a) =
ψ(b) = 0) and suppose that the phase φ ∈ C ∞ satisfies φ0 (t) 6= 0 for all t ∈ [a, b].
We claim that in this case the integral Iψ (λ) decreases very fast in λ, in particular
Proposition 2.1 (Principle of non-stationary phase). Let ψ, φ be as above, that is
ψ ∈ Cc∞ ((a, b)) and φ ∈ C ∞ is such that φ0 6= 0 on [a, b]. Then for every N > 0 we
have
|Iψ (λ)| .N,ψ,φ |λ|−N .
Remark 1. Notice that the bound given by the proposition above is only interest-
ing when |λ| is large. Indeed, when |λ| ≤ 1 we can simply bound |Iψ (λ)| ≤ kψkL1 .
1 by taking the absolute value inside the integral.
Proof. The proof is a simple integration-by-parts argument.
We want to use integration by parts a number of times. Notice that (eiλφ )0 =
iλφ0 eiλφ , so if we define the differential operator D by
1 df
Df (t) := 0
iφ (t) dt
6 OSCILLATORY INTEGRALS

we have λ−1 D(eiλφ ) = eiλφ . Notice D is well-defined because φ0 6= 0. Using

integration by parts we then have
ˆ b ˆ b
e iλφ(t)
ψ(t)dt = λ−1 D(eiλφ(t) )ψ(t)dt
a a
b ˆ b
eiλφ ψ 
= + eiλφ(t) λ−1 Dt ψ(t)dt
iλφ0 a a
ˆ b
−1
=λ eiλφ(t) Dt ψ(t)dt,
a

 ψ;
where the boundary term vanishes by the hypothesis on the support of  here
t 4 t 1 d f
D denotes the transpose of the operator D, namely D f (t) = − i dt φ0 . By
repeating the argument N times we get
ˆ b
Iψ (λ) = λ−N eiλφ (Dt )N (ψ)dt,
a

and we conclude by taking absolute values (the resulting integral is finite). 

We can interpret the principle of non-stationary phase as saying that, for a

generic phase φ, the behaviour of the oscillatory integrals Iψ (λ) is determined by
the points where φ0 = 0, because away from those points the integral contributes
at most an error term ON (|λ|−N ). In Section 2.3 below we will show how this
heuristic can be made precise.

2.2. Van der Corput’s lemma. We will now consider more general situations
than the non-stationary phase one.

We start by considering I(λ), that is the case where ψ ≡ 1, and we assume that
we have |φ0 (t)| > 1 for all t ∈ (a, b). We make a further important assumption, that
is we also assume that φ0 is monotonic5. Then we have
Proposition 2.2. If φ is such that φ0 is monotonic and |φ0 | > 1 on [a, b], we have
|I(λ)| ≤ C|λ|−1
for an absolute constant C > 0.
Some observations are in order, before we proceed to the proof:
i) First of all, the assumption that φ0 is monotonic is fundamental: indeed, the
statement is false otherwise! (prove this in Exercise 4; see also Exercise 5.)
ii) Secondly, we cannot get a decay in λ better than |λ|−1 , as the example φ(t) = t
shows (you are cordially invited to do the calculation).
iii) Thirdly, by a simple rescaling of the phase, we can see that the proposition is
more general than it looks: if the lowerbound on φ0 becomes more generally
|φ0 | > µ, then the estimate becomes |I(λ)| ≤ C(µ|λ|)−1 .
iv) Finally, notice that the constant C in the statement above depends neither
on the phase φ nor on the interval [a, b]! In particular, if we allowed the
constant to depend arbitrarily on the interval, we would trivially have that
|I(λ)| ≤ |a − b|, which is rather uninteresting.

4Formally, the operator such that hDf, gi = hf, D t gi for all f, g ∈ C ∞ .

c
5Recall that a function f : R → R is monotonic non-decreasing if x < y ⇒ f (x) ≤ f (y),
and monotonic non-increasing if x < y ⇒ f (x) ≥ f (y). A function is monotonic if it is either
monotonic non-decreasing or monotonic non-increasing.
 OSCILLATORY INTEGRALS 7

Observations ii)-iv) are actually related, thanks to the scaling behaviour of the
inequality. Indeed, if we ask for which values of α we can have the inequality
|I(λ)| ≤ Cφ |λ|−α
hold with a constant Cφ that is independent 6 of (a, b), then the answer is that
necessarily it must be α = 1. Prove this in Exercise 6.
Now we proceed with the proof.

Proof. We repeat the same integration-by-parts argument as before, except this

time the boundary terms do not vanish. Thus we have, with D the same differential
operator as before, that
ˆ b  eiλφ(b) ˆ b
eiλφ(a) 
eiλφ(t) dt = 0 (b)
− 0 (a)
+ eiλφ(t) λ−1 Dt (1)dt.
a iλφ iλφ a

Since |φ0 | > 1, the boundary term is bounded by 2/|λ| simply by triangle inequality.
As for the other term, λ−1 Dt (1) = −(iλ)−1 d/dt(1/φ0 ), and by taking absolute
values inside the integral we have that it is bounded by
ˆ b   
1 d 1 
|λ| a dt φ0
 dt.

However, φ0 is monotonic and therefore so is 1/φ0 ; this means that in the last
integral the derivative is single-signed, and therefore we can take the absolute value
outside (something normally prohibited!) and obtain that it equals
ˆ
1  b d  1  
dt.
|λ| a dt φ0

 
By the Fundamental Theorem of Calculus this is equal to |λ|−1  φ01(b) − φ01(a) , which
is bounded by 1/|λ|, thanks to the monotonicity assumption. Thus we have proven
the theorem with C = 3. 

The natural step after this is to investigate what happens when the condition
|φ0 | > 1 is violated but we still have lowerbounds for some higher derivatives.
Indeed, we have
Theorem 2.3 (Van der Corput’s lemma). Let k ≥ 2 and let φ ∈ C k be such that
|φ(k) | > 1 on (a, b). Then
|I(λ)| ≤ Ck |λ|−1/k ,
where Ck is an absolute constant depending only on k.
Notice that |λ|−1/k decays slower than |λ|−1 as λ → ∞. This is to be expected,
since the phase “slows down” near the zeros of φ0 and hence there is less overall
cancellation. It is indeed sharp, as the example φ(t) = tk over [0, 1] shows (again,
you are cordially invited to do the calculation - you can use complex integration
k
and regularize the integral by introducing a factor of e−εt , then take ε → 0 at the
end; see also Exercise 13).
Again, it goes without saying that if we have |φ(k) | > µ instead, then the inequality
becomes |I(λ)| ≤ Ck (µ|λ|)−1/k . Moreover, the exponent 1/k is necessary for the
inequality to hold with Ck independent of the interval (a, b) (see again Exercise
6).

6As the subscript indicates, the constant might depend on φ in principle. However, this will
not be the case for us, at least for now.
8 OSCILLATORY INTEGRALS

Remark 2. You should have noticed that this time we are not making an explicit
monotonicity assumption such as in Proposition 2.2. However, we implicitely still
are! Indeed, the fact that |φ(k) | > 1 for some k ≥ 2 implies that φ(k−1) attains
value zero in at most one point of (a, b), and thus φ(k−2) attains value zero in at
most two points, and so on; iterating, we see that φ0 has at most k − 2 changes of
monotonicity, or in other words we can partition (a, b) into at most k − 1 intervals
such that φ0 is monotonic on each interval.
Proof. The proof proceeds by induction on k. The case k = 1 (with the additional
assumption that φ0 be monotonic) has been proven in Proposition 2.2. Notice that
if |φ00 | > 1 then φ0 is monotonic.
Assume now that the statement is true for k − 1. Since |φ(k) | > 1, the (k − 1)-th
derivative φ(k−1) can have at most one zero in (a, b). Denote this zero by t0 and
split the integral as
ˆ t0 −δ ˆ t0 +δ ˆ b
+ +
a t0 −δ t0 +δ
for some δ > 0 to be chosen later. In the interval (a, t0 − δ) the function φ(k−1)
is never zero, but moreover we have by the assumption on its derivative φ(k) that
|φ(k−1) | > δ; similarly on the interval (t0 + δ, b). By inductive hypothesis we
therefore have that
 ˆ t0 −δ  ˆ b 
e iλφ(t) 
dt+ eiλφ(t) dt ≤ 2Ck−1 (δ|λ|)−1/(k−1) .
  

a t0 +δ
As for the remaining integral, we just estimate trivially
 ˆ t0 +δ 
eiλφ(t) dt ≤ 2δ.
 

t0 −δ

Putting everything together, we have shown that |I(λ)| ≤ 2Ck−1 (δ|λ|)−1/(k−1) +2δ,
and by choosing δ = |λ|−1/k this gives
|I(λ)| ≤ (2Ck−1 + 2)|λ|−1/k ,
proving the induction (with Ck = 2Ck−1 + 2). 
Although the constant Ck obtained with the above argument suffices for most (if
not all) applications, it’s interesting to notice it is not optimal in k. See Exercise
11 if you are interested in determining the correct behaviour of the optimal constant
in k.
It is a simple matter of integrating by parts to extend Van der Corput’s lemma
to the oscillatory integrals Iψ (λ) as well. We obtain
Corollary 2.4. Let ψ ∈ C 1 and let the phase φ satisfy |φ(k) | > 1 in (a, b) for some
k ≥ 1 (if k = 1, assume additionally that φ0 is monotonic). Then the inequality
h ˆ b i
|Iψ (λ)| ≤ Ck0 |ψ(b)| + |ψ 0 (t)|dt · |λ|−1/k
a
holds, with Ck0 > 0 an absolute constant depending only on k.
Thus the overall constant will depend on a, b through the amplitude ψ (though
notice that if we assume kψkL∞ + kψ 0 kL1 is finite, it is independent of a, b), but we
have isolated its dependence in the term in square brackets. The constant is still
independent of the phase φ.
´t
Proof. Let Φ(t) := a eiλφ(s) ds, so that
ˆ b
Iψ (λ) = Φ0 (t)ψ(t)dt.
a
 OSCILLATORY INTEGRALS 9

Integrating by parts and taking absolute values we have that |Iψ (λ)| is bounded by
ˆ b
|Φ(b)ψ(b)| + |Φ(t)||ψ 0 (t)|dt.
a

For the first term, by Theorem 2.3 we can estimate |Φ(b)| .k |λ|−1/k . Similarly,
´b
the second term can then be estimated by Ck |λ|−1/k a |ψ 0 (t)|dt. Summing these
contributions gives the stated bound. 

Notice the integration-by-parts trick we used here is different from the one used
for Propositions 2.1, 2.2.

2.3. Method of stationary phase. Now we go back to the observation made at

the end of Section 2.1. Recall that the portions of the integral Iψ (λ) where the
phase satisfies φ0 6= 0 contribute at most Oψ,N (λ−N ) for arbitrary N > 0 and
thus can tipically be treated as an error term. The behaviour of Iψ (λ) is therefore
determined by the zeros of φ0 and higher derivatives of the phase. In particular, one
can perform a full asymptotic expansion of the oscillatory integral Iψ (λ). This is
known as the method of stationary phase; it’s particularly useful in physics, where it
prominently appears in the semi-classical approximation to Quantum Field Theory.
We can state (one version of) the method as follows.

Theorem 2.5 (Method of stationary phase). Assume that φ ∈ C ∞ . Let k ≥ 2 and

assume that

φ0 (t0 ) = . . . = φ(k−1) (t0 ) = 0 and that φ(k) (t0 ) 6= 0.

If ψ ∈ Cc∞ is supported in a sufficiently small neighbourhood of t0 , then there exist

coefficients aj for j ∈ N (each depending only on finitely many derivatives of φ and
ψ at t0 ) such that
X
Iψ (λ) ' eiλφ(t0 ) λ−1/k aj λ−j/k , (2.1)
j∈N

where by ' we mean that for all n > 0 we have for the sum truncated at n − 1 that
 n−1
X 
Iψ (λ) − eiλφ(t0 ) λ−1/k −j/k  −n/k
 
 aj λ  .ψ,φ,n |λ|
j=0

as λ → ∞, and moreover for all n, ` > 0 we have

  h n−1 i
 d ` iλφ(t0 ) −1/k
X
−j/k  −`−n/k

 dλ Iψ (λ) − e λ aj λ  .ψ,φ,`,n |λ| .
j=0

The coefficients aj can be determined explicitely - for example, when k = 2 an

explicit calculation of a0 shows that the main term in Iψ (λ) is
 2π 1/2
eiλφ(t0 ) ψ(t0 ).
−iλφ00 (t0 )
This is excellent for a quick-and-dirty estimate of complicated integrals.
We will not prove the above theorem here since it would make these notes unnec-
essarily long, but if you fancy you can prove it yourself in Exercise 13 following
the guidelines provided there.
10 OSCILLATORY INTEGRALS

3. Oscillatory integrals in several variables

In the previous section we have analysed the situation for single variable phases,
that is for integrals over (intervals of) R. In this section, we will instead study
the higher dimensional situation, when the phase is a function of several variables.
Things are unfortunately generally not as nice as in the single variable case, as you
will see.

In order to avoid having to worry about connected open sets of Rd (see Exercise
18 for the sort of issues that arise in trying to deal with general open sets of Rd ),
in this section we will study mainly objects of the form
ˆ
Iψ (λ) := eiλu(x) ψ(x)dx,
Rd
where ψ has compact support. We have switched to u for the phase to remind the
reader of the fact that it is a function of several variables now.
3.1. Principle of non-stationary phase - several variables. The principle of
non-stationary phase we saw in Section 2.1 continues to hold in the several variables
case.
Given a phase u, we say that x0 is a critical point of u if
∇u(x0 ) = (0, . . . , 0).
Proposition 3.1 (Principle of non-stationary phase - several variables). Let ψ ∈
Cc∞ (Rd ) (that is, smooth and compactly supported) and let the phase u ∈ C ∞ be
such that u does not have critical points in the support of ψ. Then for any N > 0
we have
|Iψ (λ)| .N,ψ,u |λ|−N .
Proof. The proof is simply a reduction to the single variable case of Proposition
2.1. Indeed, by assumption there exists a c > 0 such that |∇u(x)| > 2c for x in the
support of ψ, and therefore for any such x there exists a small ball Bx , centered at x
and a unit vector ξx such that ξx · ∇u(y) > c for every y ∈ Bx . By compactness we
can find a finite collection of such balls, {Bj }j , that covers the support of ψ;Pwith
a partition of unity associated to this collection, we can therefore write ψ = j ψj
with ψj supported in Bj . Since the number of balls is finite (depending only on
ψ, u), it suffices to prove the claimed bound for a single Iψj (λ). By a change of
coordinates, we can assume ξj = (1, 0, . . . , 0) and since
ˆ ˆ ˆ 
eiλu(x) ψj (x)dx = eiλu(x1 ,x2 ,...,xd ) ψj (x1 , x2 , . . . , xd )dx1 dx2 · · · dxd ,
Rd Rd−1 R
we can conclude by applying Proposition 2.1 to the inner integral and then inte-
grating in the remaining variables (over the projection of Bj ). 
3.2. Van der Corput’s lemma in several variables. Proceeding by analogy to
what was done in Section 2, we now want to study what happens when we have
a lowerbound on some (possibly mixed) derivative of the phase u. We look for
inequalities that share the same scaling behaviour as the single variable ones in
Proposition 2.2 and Theorem 2.3.

We will denote by α = (α1 , . . . , αd ) a multi-index, that is an element of Nd , and

by |α| the sum α1 + . . . + αd . Then ∂ α denotes the partial derivative of order |α|
∂ α = ∂xα11 · · · ∂xαdd .
We have the following several-variables version of Van der Corput’s lemma (more
precisely, of Corollary 2.4).
 OSCILLATORY INTEGRALS 11

Theorem 3.2 (Van der Corput’s lemma in several variables). Let ψ ∈ Cc1 (B(0, 1))
(that is, compactly supported in the unit ball). Assume that for every x in the
support of ψ the phase u satisfies
|∂ α u(x)| > 1
for a certain fixed multi-index α, with |α| ≥ 1. Furthermore we assume u ∈ C |α|+1 .
Then we have
|Iψ (λ)| ≤ C|α|,u kψkL∞ + k∇ψkL1 · |λ|−1/|α| ,
 

where the constant C|α|,u depends only on |α| and on the phase u (in particular, on
the derivatives of u up to order |α| + 1).
One can appreciate the superficial similarity between Corollary 2.4 and Theorem
3.2. We reiterate that if the hypothesis on the phase becomes |∂ α u(x)| > µ then
the inequality holds with |λ|−1/|α| replaced by (µ|λ|)−1/|α| .
However, there are several key differences with Corollary 2.4 which make the above
theorem weak:
i) unfortunately, the constant now depends on the phase too, which can some-
times be an issue when one has to deal with multiple phases at once;
ii) as you will see from the proof, the constant also depends on the size of the
support of ψ if one removes the extra hypothesis that ψ ∈ Cc1 (B(0, 1)), and
thus the inequality does not have the nice scaling properties that it had in the
single variable case;
iii) the estimate is no longer tight: indeed, if we take a phase like u(x, y) = xy,
so that |∂x ∂y u| = 1, and an amplitude that is just a bump function supported
near the origin, the theorem gives us a bound of O(|λ|−1/2 ), but in reality one
can prove the better estimate7 O(|λ|−1 ) (see Section 3.3 or Exercise 19).

Proof. The proof will again work by reducing to the single variable case (and this
is the source of the inefficiencies mentioned above).
Let k = |α|. In order to reduce to the single variable case, we need to show that the
assumption |∂ α u| > 1 implies a similar lowerbound for some derivative of the form
d


(ξ · ∇)k u - that is, a k-th derivative in a fixed direction, (ξ · ∇)f (x) = dt f (x + tξ) ,
so that we can apply single variable results. Indeed, the real vector space of partial
derivatives of order k admits a basis of the above form: there exist unit vectors
ξ (1) , . . . , ξ (m) (where m = #{β ∈ Nd s.t. |β| = k}) such that any partial derivative
of order k can be expressed as a linear combination of derivatives (ξ (j) · ∇)k . Prove
this in Exercise 16.
Since ∂ α can be expressed as a linear combination of (ξ (j) · ∇)k , for each x in the
support of ψ there exists a unit vector ξx such that |(ξx · ∇)k u(x)| &k 1; moreover,
since we have assumed that u ∈ C k+1 , we have that kukC k+1 is bounded on the
support of ψ, and therefore for a small ball Bx (of radius . kuk−1 C k+1
) centered at
x we have actually
|(ξx · ∇)k u(y)| &k 1 for all y ∈ Bx
(with a worse constant than the one implicit in the previous lowerbound). As in
the proof of Proposition 3.1, we can find a finite covering of the support of ψ by
P balls and use them to create a partition of unity that allows us to split ψ as
such
j ψj with each ψj supported in one of such balls. It suffices to estimate Iψj (λ)
separately (ψ being supported in B(0, 1), the number of such balls only depends

7For a suggestive calculation: estimate ´ exp(iλxy)dxdy where Ω = {(x, y) : |x| + |y| ≤ 1}

Ω
(hint: rotate coordinates and use single variable Van der Corput’s lemma).
12 OSCILLATORY INTEGRALS

on kukC k+1 ). After a change of variables, we can assume that ξ = (1, 0, . . . , 0) for
the ball under examination, and therefore we write again (with x0 = (x2 , . . . , xd ))
ˆ ˆ ˆ 
0
e iλu(x)
ψj (x)dx = eiλu(x1 ,x ) ψj (x1 , x0 )dx1 dx0 .
B(0,1) Rd−1 ∩B(0,1)

Applying Corollary 2.4 to the inner integral we get that it is bounded by

h ˆ i
.k kψkL∞ + |∂x1 ψ(x1 , x0 )|dx1 · |λ|−1/k ;

integrating in the remaining variables gives us the estimate we want. 

3.3. Method of stationary phase in several variables - non-degenerate
phases. When |α| = 2, we can obtain better estimates than the ones given by
Theorem 3.2, provided that we have some extra information about the phase being
well-behaved. In particular, if x0 is a critical point for phase u, we say that it is a
non-degenerate critical point if
det(Hess(u)(x0 )) 6= 0,
 2 
where Hess(u) is the Hessian matrix ∂x∂i ∂x
u
j
.
ij
If the phase has only non-degenerate critical points, we can assert an analogue of
Theorem 2.5 for phases of several variables.
Theorem 3.3. Assume that ψ ∈ Cc∞ (Rd ) and that the phase u ∈ C ∞ (Rd ) has
x0 as a non-degenerate critical point. If ψ is supported in a sufficiently small
neighbourhood of x0 (in particular, there are no other critical points of u in it),
then there exist coefficients aj for j ∈ N (each depending only on finitely many
derivatives of u and ψ at x0 ) such that
X
Iψ (λ) ' eiλu(x0 ) λ−d/2 aj λ−j , (3.1)
j∈N

where ' is in the same sense as in Theorem 2.5. In particular, |Iψ (λ)| .φ,ψ |λ|−d/2 .
When u(x, y) = xy, we see that this gives Iψ (λ) = O(λ−1 ) as claimed before.
We do not prove this theorem either (but if you have proven Theorem 2.5 in
Exercise 13, you can prove this one as well in Exercise 20 following the same
strategy with little extra effort), but rather make a remark.
Remark 3. The theorem above is limited to the case |α| = 2. Indeed, observe
that in the single variable case the fact that φ(k) (t0 ) 6= 0 and φ(t0 ) = φ0 (t0 ) =
. . . = φ(k−1) (t0 ) = 0 implies that we can put the phase in the simple canonical
form φ(t) = tk (after a change of variable); however, in several variables there is no
simple analogue of this, except in the case |α| = 2, in which we can put the phase
u in a standard quadratic form - hence our limitation.
3.4. Estimates independent of the phase. As observed above, one of the down-
sides of Van der Corput’s lemma in several variables is that the constant ends up
depending on the phase u as well, which is sometimes a problem. It is also known
that, if we consider the case ψ ≡ 1 and thus look at oscillatory integrals of the form
´ iλu(x)
Ω
e dx with Ω a connected (bounded) open set, the constant cannot be made
independent of Ω (prove this in Exercise 18).
It is reasonable to look for estimates with constants that are independent of as many
parameters as possible. The above observation forces us to give up independence in
Ω, and therefore we concentrate on independence on the phase u. It turns out that
it is possible to obtain oscillatory integral estimates with constant independent of
the phase, but it seems that one has to pay the price somehow (see the log in the
 OSCILLATORY INTEGRALS 13

estimate below). Such estimates form a rich theory, but here we will only present
an example to give you a taste of what kind of results and techniques are involved.
Theorem 3.4. If u : [0, 1]2 → R is such that
 ∂2u 
(x, y) > 1 ∀(x, y) ∈ [0, 1]2
 
∂x∂y


and furthermore
∂3u
(x, y) 6= 0 ∀(x, y) ∈ [0, 1]2 ,
∂ 2 x∂y
we have ˆ 
eiλu(x,y) dxdy  . (log(2 + |λ|))1/2 |λ|−1/2 .
 

[0,1]2
In particular, the constant is independent of u.
Obviously, there is an analogous statement for ∂x ∂y2 u in place of ∂x2 ∂y u. No-
tice that the assumption on this third mixed derivative is nothing more than a
monotonicity assumption (which turns out to be necessary).
Proof. Assume λ > 1 for convenience. By Cauchy-Schwarz and Fubini, we have
ˆ 2 ˆ 1  ˆ 1 2 ˆ 1
iλu(x,y) iλu(x,y)
e dxdy ≤ e dy dx · 1dx
   
   
[0,1]2 0 0 0
ˆ 1 ˆ 1ˆ 1 
0
= eiλ(u(x,y)−u(x,y )) dx dydy 0 .
0 0 0
Letting φy,y0 (x) := u(x, y) − u(x, y 0 ) we have that
ˆ y
φ0y,y0 (x) = ∂x u(x, y) − ∂x u(x, y 0 ) = ∂x ∂y u(x, t)dt,
y0

which by the hypotheses on u implies that φ0y,y0 is monotonic (since φ00y,y0 is single-
signed) and that |φ0y,y0 (x)| > |y − y 0 |. By Proposition 2.2 it follows that
ˆ 1 0

eiλ(u(x,y)−u(x,y )) dx . min{1, (λ|y − y 0 |)−1 }.
 

0
Thus the above is bounded by
ˆ 1ˆ 1
min{1, (λ|y − y 0 |)−1 }dydy 0
0 0
¨ ¨
−1 0 −1 0
=λ |y − y | dydy + 1dydy 0 . λ−1 log λ + λ−1 ,
|y−y 0 |>λ−1 |y−y 0 |<λ−1
which gives the claim. 
It is not known whether the estimate holds without the logarithmic term.
 OSCILLATORY INTEGRALS 15

Exercises
This is a (long) list of exercises meant to improve your grasp on the topic and
provide you with useful ideas for the future - do the ones you like. Those that require
a bit more work are marked by F’s - they are not necessarily harder, just longer
maybe. The ones unmarked are probably more important for a basic understanding
though. In any case, hints are given in the next section. Since the exercises are
meant to be a complement to the lectures, don’t be afraid to have a look at the
hints - you are actually encouraged to do so.
Exercise 1. (F) Let p be such that 1 ≤ p < 2d/(d + 1). Show, by using the
decay estimate J(t) . (1 + |t|)−(d−1)/2 , that if f is a radial function in Lp (Rd )
and Rd 3 ξ 6= 0, then the Fourier transform fb is continuous at ξ. (See hints for a
walk-through)
Exercise 2. Show, using the geometric interpretation given in Section 1.2 of solu-
tions to n = x2 + y 2 as points in a disk, that there exists some constant C > 0 such
that for all N > 1
N
X
r2 (n) − πN  ≤ CN 1/2


n=1
holds.
Exercise 3. If D is a (linear) differential operator, its transpose is given by the
linear operator Dt that satisfies
hDf, gi = hf, Dt gi
for all test functions f, g ∈ Cc∞ . While D satisfies Leibniz’s rule
D(f g) = gDf + f Dg,
show that Dt in general does not.
Exercise 4. Show that Proposition 2.2 is false if the assumption that φ0 is mono-
tonic is dropped. In other words, untangling the quantifiers, construct a family
of phases {φλ }λ such that |φλ 0 | > 1 on (a, b) for each λ, but |λ||I(λ; φλ )| is not
bounded as λ → +∞.

[To destroy cancellation in I(λ), you should aim to make φλ 0 oscillate at the scale
|λ|−1 . Indeed, consider the following: looking only at the real part of the integral,
´1
we are trying to make 0 cos(2πλφ 1 λ (t))dt  large (and positive, say). The function
1
cos(2πλx) is positive for x in − 4λ , 4λ + λ1 Z, and negative otherwise. If you look
at the graph of φλ (and I encourage you to make a drawing of the argument that
follows),
 you want
 it1 to
spend as much time as possible in the horizontal bands
1 1
R × − 4λ , 4λ + λ Z and  as little time as possible in the complement, that is
1 3
+ λ1 Z , so that the positive contribution outweighs the


in the bands R × 4λ , 4λ
negative one. To achieve this, φλ 0 should be small when φλ is in the former bands
(but not too small, since we still want |φλ 0 | > 1), and quite large when in the latter
bands. In particular, φλ 0 will not be monotone and it will oscillate between two
behaviours over an interval of length ∼ λ−1 .]
Exercise 5. Show the following weaker version of Van der Corput’s lemma for
k = 1: assume that |φ0 | > µ but, instead of assuming that φ0 is monotonic, we only
assume that |φ00 | < M on (a, b); prove that
1 M (b − a)  −1
|I(λ)| . + |λ| .
µ µ2
16 OSCILLATORY INTEGRALS

Exercise 6. Let φ be a phase that satisfies |φ(k) | > 1 on (a, b) for some k ≥ 1.
Show that a necessary condition for inequality
|I(λ)| ≤ Cφ |λ|−α
to hold with Cφ independent of [a, b] is that α = 1/k.
Exercise 7. Show, using integration by parts (as many times as necessary) and
Van der Corput’s lemma, that we have the estimate
|J(t)| .d (1 + |t|)−(d−1)/2 ,
where J is the function introduced in Section 1.
Exercise 8. The Airy equation is the dispersive PDE
(
∂t u + ∂x3 u = 0,
u(0, x) = f (x).
By taking a Fourier transform in the spatial variable x, it is not hard to see that
the solution to the Airy equation can be written formally as the convolution
ˆ x − y
1 1
u(t, x) = f (y) 1/3 Ai 1/3 dy,
2π t t
where Ai(x) denotes the Airy function
ˆ
3
Ai(x) := ei(ξ +xξ)
dξ,
R

which is clearly an oscillatory integral (provided the integral exists!). You will show
that it satisfies the following estimates:
1) For x > 1 we have superpolynomial decay, that is for every N > 0 we have
|Ai(x)| . |x|−N .
2) For −1 ≤ x ≤ 1 we have |Ai(x)| . 1.
3) For x < −1 we have |Ai(x)| . |x|−1/4 .
Follow these steps:
i) It will be useful to smoothly split R dyadically. Let ϕ be a C ∞ bump function
supported in [−2, 2], with ϕ ≡ 1 on [−1,
P 1]; define ψ(ξ) := ϕ(ξ/2) − ϕ(ξ), and
ψj (ξ) := ψ(2−j ξ). Show that ϕ(ξ) + j∈N ψj (ξ) = 1 for every ξ and that ψj
is supported in 2j < |ξ| < 2j+2 . We can make sense of the integral defining
Ai(x) as the limit as n → ∞ of
ˆ n ˆ
3 X 3
ei(ξ +xξ) ϕ(ξ)dξ + ei(ξ +xξ) ψj (ξ)dξ.
j=0

ii) Splitting the integral using the decomposition above, show 1) by the principle
of non-stationary phase. In doing this, you will also prove that the limit above
indeed exists pointwise.
iii) Show 2) adapting the argument you just used for ii).
iv) Show 3) by Corollary 2.4 (and the splitting).
v) Show the simple dispersive estimate
ku(·, t)kL∞ (R,dx) . |t|−1/3 kf kL1 (R) .
[The integration in the definition of Ai(x) is over all of R instead of a finite interval
as in Section 2; that is why we introduced the splitting. More importantly, notice
that the phase xξ + ξ 3 is not of the form λφ, or at least so it appears at first sight...]
 OSCILLATORY INTEGRALS 17

Exercise 9. Show estimate (1.2) using integration by parts, Van der Corput’s
lemma and splitting the integrals where necessary; that is, when d = 2 (thus J(r) =
´ π −2πir cos θ
0
e dθ) show that we have for large R
ˆ R 
rJ(r)dr . R1/2 .
 

0
´R
The point is that 0 rJ(r)dr is itself an oscillatory integral (expanding J and using
Fubini shows this is the case).
[Notice this is a much better estimate than the one you would get from just plugging
in the estimate |J(r)| . |r|−1/2 , which would give you the much larger O(R3/2 ).
There is therefore additional cancellation to be exploited.]
Exercise 10. An estimate of the form
|{t ∈ (a, b) s.t. |φ(t)| < λ}| = O(λα )
for some α > 0 is called a sublevel-set estimate. Notice that it is only interesting for
small λ. These estimates are in many ways related to oscillatory integral estimates,
as this exercise will show:
i) Show that Van der Corput’s lemma implies that, under the same hypotheses
on φ, for every λ > 0 we have
|{t ∈ (a, b) s.t. |φ(t)| < λ}| .k λ1/k . (♠)
We stress that the constant only depends on k.
ii) Show conversely that if we assume that (♠) holds, we can prove Van der
Corput’s lemma by splitting the interval (a, b) into {t ∈ (a, b) s.t. |φ0 (t)| < θ}
and its complement, where θ is a parameter to be optimized (Remark 2 might
come in handy).
iii) Now prove (♠) directly by induction in k (thus providing another proof of Van
der Corput’s lemma, when combined with ii)).
iv) (Only if you know about p-adics, otherwise ignore) Let p be a prime and let
Zp be the ring of p-adic integers, with its non-archimedean valuation | · |p . If
P is a polynomial in Z[X], what does the sublevel-set
{x ∈ Zp s.t. |P (x)|p < p−s }
correspond to, in simpler terms? And what would an estimate like (♠) mean
in this context?
Exercise 11. (F) The proof of Van der Corput’s lemma we have presented in
these notes gave us a constant that is exponential in k; in particular, you can easily
see that it gives Ck ∼ 2k (Theorem 2.3). This behaviour is not sharp. Indeed, if
Bk denotes the smallest constant for which the inequality
|I(λ)| ≤ Bk |λ|−1/k
holds (under the hypotheses of Theorem 2.3), then the constant actually grows
linearly in k, namely Bk ∼ k.
In this exercise you will obtain the optimal behaviour of the constant, showing that
|I(λ)| ≤ Ck|λ|−1/k . This will be achieved by using the same strategy as in ii)-iii)
of Exercise 10, but with an improved dependency on k for estimate (♠).
i) Show that, if E is a measurable subset of R with |E| > 0, for any k one can
find x0 , . . . , xk ∈ E such that for any j ∈ {0, 1, . . . , k}
Y  |E| k
|xi − xj | ≥ .
2e
i : i6=j

[hint: squeeze E into an interval of length |E|, then do the obvious thing.]
18 OSCILLATORY INTEGRALS

ii) Show the following generalisation of the mean-value theorem: let x0 < x1 <
· · · < xk and f ∈ C k ([x0 , xk ]); then there exists y ∈ (x0 , xk ) such that
k
X f (xj )
f (k) (y) = (−1)k k! Q .
j=0 i : i6=j (xi − xj )

[hint: use Lagrange interpolation at the xj ’s to get a polynomial approximant,

then apply Rolle’s theorem to the difference k times.]
iii) Show that if φ is such that φ(k) (t) > 1 for all t ∈ R, then we have the sublevel-
set estimate (see Exercise 10)
|{t ∈ R s.t. |φ(t)| < λ}| ≤ (2e)((k + 1)!)1/k λ1/k .
In particular, use i) and ii) above applied to E = {t ∈ R s.t. |φ(t)| < λ} and
f = φ.
iv) Use Stirling’s approximation to show that ((k + 1)!)1/k ∼ k.
v) Prove Van der Corput’s lemma using the same proof as in ii) of Exercise
10 (using the sublevel-set estimate above that has an improved constant) to
conclude that
|I(λ)| ≤ Ck|λ|−1/k .
vi) Show that Bk cannot grow slower than k [check a canonical example and
use complex integration techniques; don’t expect a straightforward calculation
though.]
Exercise 12. You might have noticed that the statement of Theorem 2.5 (the
method of stationary phase) omits the case k = 1. Indeed, in this case there can
be no expansion in terms of powers of λ−1 , since the integral is ON (λ−N ) for every
N > 0 (by non-stationary phase principle). However, in the case where the support
of ψ is not strictly contained in (a, b) (and thus non-stationary phase does not
apply), a similar statement holds. Prove an asymptotic expansion of the form
X 
λ−1 aj eiλφ(a) + bj eiλφ(b) λ−j
j∈N

for Iψ (λ), under the hypotheses that φ0 6= 0 in (a, b) and that φ, ψ ∈ C ∞ (in
particular, in general ψ(a), ψ(b) 6= 0). Calculate explicitely the first few coefficients
aj , bj .
[hint: just use integration by parts repeatedly. ]
Exercise 13. (FF) In this exercise you will prove the method of stationary phase
as stated in Theorem 2.5. The proof will proceed in stages, and we can assume
t0 = 0 and φ(0) = 0 for simplicity.
i) Begin with the case k = 2, which will already contain all the main ideas.
2
Take φ(t) = t2 and ψ(t) = tm e−t for some m > 0 (which is not compactly
supported, but still very concentrated around the origin); let a = −∞ and
b = +∞. Show, using standard complex integration techniques, that Iψ (λ) =
´ +∞ iλt2 m −t2
−∞
e t e dt equals (fixing the principal branch of z −(m+1)/2 )
ˆ +∞
2
−(m+1)/2
(1 − iλ) tm e−t dt,
−∞

and argue by a power series expansion in λ−1 that therefore it satisfies (2.1).
ii) Now we keep the quadratic phase φ(t) = t2 but replace the gaussian factor
2
e−t directly with a compactly supported function η. Thus, let η ∈ Cc∞ and
let ψ(t) = tm η(t). Show, by splitting the region of integration smoothly into
 OSCILLATORY INTEGRALS 19

one close to 0 and one away from 0 (according to a well chosen parameter),
that in this case
|Itm η (λ)| .η,m |λ|−(m+1)/2 .
More precisely, let ϕ denote a smooth bump function supported in [−1, 1] and
decompose 1 = ϕ(t/δ) + (1 − ϕ(t/δ)). For the region close to 0 you can take the
absolute value inside the integral and give a trivial estimate; for the region away
from 0, apply the integration by parts argument as in the proof of Proposition
2.1 N times, where 2N is bigger than m+1, then take the absolute value inside
and estimate the size of this integral. Finally, optimise in the parameter δ.
iii) Show, using the above same arguments, that if g ∈ S(Rd ) (that is, g is a
Schwartz function) is such that g ≡ 0 in a neighbourhood of 0, then for any
N > 0 one has |Ig (λ)| = ON,g (|λ|−N ). (The phase is still φ(t) = t2 .)
iv) Now, still in the case φ(t) = t2 , we tackle the general ψ case. Write
ˆ ˆ
2 2 2 2
eiλt ψ(t)dt = eiλt e−t et ψ(t) η(t)dt,

where η ∈ Cc∞ is such that η(t) = 1 for t in the support of ψ. Perform a Taylor
2
expansion of et ψ(t) to degree n and substitute into Iψ (λ). Then use i)-ii)-iii)
to argue that (2.1) holds for each term you get out of this procedure.
v) Now consider the general k = 2 case. The idea is to deform the phase to turn it
into t2 again. Find a diffeomorphism from a sufficiently small neighbourhood
U of 0 that, by a change of variables, achieves this, and then conclude that
(2.1) holds by iv).
vi) Generalise the above to k > 2. The substitute for i) will be the identity
ˆ ∞
k k
eiλt e−t tm dt = Ck,m (1 − iλ)−(m+1)/k ,
0
which is similarly proven by standard complex integration techniques.
Exercise 14. In these notes we have only considered amplitudes ψ that are ev-
erywhere C ∞ . In applications however (see viii) below) more singular cases arise,
and it is instructive to see that some of those cases can still be treated using the
techniques here developed. This exercise will moreover show you that sometimes
some cancellation can also arise from the amplitude.
Pd
Let P (t) = j=1 cj tj be a polynomial of degree d. You will show that for arbitrary
0 < ε < R it holds ˆ dt 
eiP (t)  ≤ Cd

t

ε<|t|<R
with a constant that depends only on d (and not on the coefficients of P !). Notice
that the amplitude 1/t is only C ∞ away from 0, and has a bad singularity there.
In the following, it will be really important that integration is over a set symmetric
with respect to the origin such as {t ∈ R : ε < |t| < R} = (−R, −ε) ∪ (ε, R) above.
i) Show by a change of variable that we can assume that cd = 1 (at the price of
changing ε, R; but these are arbitrary).
ii) Show, using Corollary 2.4, that the portion of the integral over 1 < |t| < R
is bounded by Od (1), so that it suffices to concentrate on the remaining part
ε < |t| < 1. (Notice that even in the range |t| > 1 we cannot estimate the
integral by taking absolute values, since the resulting quantity grows like log R;
we really need oscillatory integral estimates!).
Pk
iii) Let Pk (t) := j=1 cj tj (thus Pd = P ). Using the trivial estimate |eiθ − 1| . |θ|
and taking absolute values, show that
ˆ ˆ
iP (t) dt dt 
e − eiPd−1 (t)  .d 1.

t t

ε<|t|<1 ε<|t|<1
20 OSCILLATORY INTEGRALS

´
iv) Use steps i)-iii) repeatedly to reduce to the case of integral eic1 t dt
t .
´
ε<|t|<1
v) It remains to show that the above integral is Od (1). Show that a<|t|<b dt t =0
for any a, b > 0 (this is a form of cancellation!). Therefore, you can arbitrarily
ic1 t
subtract 1 from the exponential
´ ´ e
term above...
vi) Split the integral as ε<|t|<1/c1 + 1/c1 <|t|<1 and apply again Corollary 2.4 to
the second integral to show ´ that it is Od (1).
vii) It remains to estimate ε<|t|<1/c1 eic1 t dt t . Using v), show that it is the same
´ ic1 t dt
as ε<|t|<1/c1 (e − 1) t and estimate the latter by taking the absolute value
and using the trivial estimate for |eiθ − 1|. This concludes the proof.
viii) Consider the operator
ˆ
dt
f 7→ lim f (x − t, y − t2 ) ;
ε→0, ε<|t|<R t
R→∞

this is known as the Hilbert transform along a parabola. Show that this oper-
ator is L2 (R2 ) → L2 (R2 ) bounded. [hint: use Plancherel and take a Fourier
transform of the operator above; after some calculations, you should recognise
a certain object... ]
Exercise 15. (F) The discrete counterparts to oscillatory integrals, which we
haven’t discussed in these notes, are exponential sums, that is sums of the form
N
X
e2πif (n)
n=1

(sometimes they are normalised by dividing by N ). Such sums are ubiquitous in

number theory, where they are used to encode all sorts of information, from the
number of solutions to diophantine equations to the behaviour of the Riemann Zeta
function.
It should not be too surprising that some of the techniques developed for oscillatory
integrals transfer over to the treatment of exponential sums. Indeed, integration by
parts has a very direct discrete counterpart in summation by parts. In this exercise
you will prove some estimates for exponential sums in the spirit of the ones studied
in these lectures. We will use a notational convention from number theory, namely
we introduce the function e(x) := e2πix so that e2πif (n) = e(f (n)).
You will prove an analogue of Proposition 2.2:
i) Show the trigonometric identity
1 1 i
iθ
= + cot(θ/2).
1−e 2 2
e(f (n))
ii) Let g(n) = e(f (n))−e(f (n−1)) ; show by summation by parts that
N −1
X   NX
e(f (n)) = e(f (N ))g(N ) − e(f (0))g(1) − e(f (n))(g(n + 1) − g(n)).
n=1 n=1

iii) Using i), show that

 
g(n + 1) − g(n) = (i/2) cot(π(f (n + 1) − f (n))) − cot(π(f (n) − f (n − 1))) .
iv) So far we just did a bunch of algebra; now we start doing analysis. Assume
that f 0 is monotonic in the interval [0, N ] and that for some 1/2 > δ > 0
we have dist(f 0 (x), Z) > δ for all x ∈ [0, N ] (that is, the derivative always
stays within two consecutive integers, never getting any closer than δ to any
 OSCILLATORY INTEGRALS 21

of them). Show that the sequence n 7→ f (n) − f (n − 1) is also monotonic and

that for some fixed integer k we have for all n ∈ {1, . . . , N }
k + δ ≤ f (n) − f (n − 1) ≤ k + 1 − δ.
v) Since cot is also monotonic in intervals of the form [πk, π(k + 1)], show that
under the hypotheses in iv) on f we have
−1
 NX 
e(f (n))(g(n + 1) − g(n)) ≤ |g(N ) − g(0)|,
 

n=1

and deduce that the above is bounded by . δ −1 . [hint: monotonicity allows

you to take the absolute values from inside of a sum to outside of it; also,
cot(x) ≤ 1/x for x ∈ (0, π/2].]
vi) Put everything together (check those remaining terms) and conclude that you
have shown: if f 0 is monotonic on interval I and dist(f 0 (x), Z) > δ for all
x ∈ I, then  X 
e(f (n)) . δ −1 . (♦1 )
 

n∈I∩Z
Compare this with Proposition 2.2 and appreciate the similarities between
the respective proofs.
Now we move on to the 2nd derivative. You will prove an analogue of Van der
Corput’s lemma, that is: if f ∈ C 2 is such that for all x ∈ I we have
λ < |f 00 (x)| < Cλ
(for some C > 1), then
 X 
e(f (n)) . C|I|λ1/2 + λ−1/2 . (♦2 )
 

n∈I∩Z

The strategy will be to partition I into intervals where (♦1 ) above applies, and
where it doesn’t to use trivial estimates, and finally optimize between the two.
a) Assume for convenience that f 00 is positive and let I = [a, b]. Show that the
range of f 0 is J := [f 0 (a), f 0 (b)] and show that |J| ≤ Cλ|I|.
b) Partition J into intervals close to integers and intervals that avoid integers. More
precisely, let 1/2 > δ > 0 be a parameter to be chosen and define
Ak :=[k − δ, k + δ] ∩ J,
Bk :=[k + δ, k + 1 − δ] ∩ J.
Show that Ak and Bk are not empty for at most O(Cλ|I| + 1) values of k and
that they partition J.
c) Since f 0 is monotonic, we can partition I into a disjoint union of intervals
(f 0 )−1 (Ak ) and (f 0 )−1 (Bk ). Show that on the first ones we have trivially
 X 
e(f (n)) . δ/λ
 

n∈(f 0 )−1 (Ak )∩Z

and on the second ones by (♦1 ) we have

 X 
e(f (n)) . δ −1 .
 

n∈(f 0 )−1 (Bk )∩Z

d) Now put everything together to show that we have

 X  δ 1
e(f (n)) . (Cλ|I| + 1) + ,
 
λ δ

n∈I∩Z
22 OSCILLATORY INTEGRALS

and conclude estimate (♦2 ) by choosing an optimal δ (i.e. minimise δ/λ + 1/δ).
Again, you should appreciate the similarities between this proof and the proof
of Van der Corput’s lemma we have given in the notes.
e) An
P estimate is only interesting if it beats the trivial one, which in this case is
| n∈I∩Z e(f (n))| ≤ |I| (by triangle inequality; here |I| > 1). For which range
of values of λ is estimate (♦2 ) interesting?
f) Consider truncations of the Riemann Zeta function along the critical line, that
PB
is sums of the form n=A n−1/2−it . Argue that in order to bound such sums,
PB 0
by summation by parts one can reduce to study sums of the form n=A0 n−it
instead. As n−it = e(−(t/2π) log n), verify that the estimates (♦1 ), (♦2 ) that
you proved above are well suited to treat

 these last oscillatory sums. This is
how bounds on the growth of ζ 21 + it  as t → ∞ are usually proven.
Exercise 16. Let V denote the real vector space of partial derivatives of order k,
that is the space of differential operators of the form
X
cα ∂ α .
α :|α|=k

Show that V admits a basis of vectors all of the form (ξ · ∇)k , where the ξ’s are
unit vectors.
Equivalently, you can identify V with the real vector space of homogeneous poly-
nomials of degree k in d variables.
[hint: Using the identification with homogeneous polynomials, it suffices to find an
inner product on V such that if a polynomial is orthogonal to all polynomials of
the form (ξ · X)k , with X = (X1 , . . . , Xd ), then it has to be the zero polynomial.]
Exercise 17. Show again the estimate
|J(t)| .d (1 + |t|)−(d−1)/2
as in Exercise 7, but this time using the method of stationary phase in several
variables as in Theorem 3.3. Use the expression for J in terms of an integral over
the sphere, then express the half-sphere as the graph of a function.
Relatedly, recall that if µ is a finite measure on Rd we can define its Fourier trans-
form dµ
c as
ˆ
dµ(ξ)
c := e−2πix·ξ dµ(x).
Rd
Calculate the Fourier transform dσ(ξ)
c of the spherical measure dσ on the (d − 1)-
d−1
dimensional sphere S and deduce its decay properties from the above. Oscilla-
tory integrals provide a great way to obtain such estimates.
Finally, consider the flat surface in Rd given by [−1, 1]d−1 × {0} and let dµ be its
surface measure. Show that there is a direction in which dµ c does not decay at all.
This should convince you that the source of the extra cancellation that makes dσ c
decay is the curvature of the underlying surface.
Exercise 18. In this exercise you will show that in dimension greater than 1 it is
impossible to have estimates of the form
ˆ 
eiλu(x) dx ≤ C|λ|−1/|α|
 

Ω
with constant C independent of Ω (here we are assuming |∂ α u| > 1).
i) Repeat the proof you gave for i) of Exercise 10 to show that the inequality
above would imply
{x ∈ Ω s.t. |u(x)| < λ} ≤ C 0 λ1/|α|
 OSCILLATORY INTEGRALS 23

with C 0 independent of Ω. Thus it will suffice to disprove the sublevel-set

estimates.
ii) First we show that if Ω is not bounded, there is no hope to have an estimate
like the above. Consider Ω = [−R, R]2 and u(x, y) = 12 (x + y)2 . We have
∂x ∂y u = 1. Show that, however,
|{(x, y) ∈ Ω s.t. |u(x, y)| < λ}| ∼ Rλ1/2 ,
so there is an unavoidable dependence on diam(Ω).
iii) Now we show that even if we take Ω contained in a bounded set (say [0, 1]2 ),
the constant will still depend on properties of Ω. Here the trick will be to
consider a set Ω that looks a bit like a comb with N teeth. In particular,
define
Ωj := [1/2, 1] × [j/N, (j + 1/2)/N ]
for j = 0, . . . , N − 1 (this will be the j-th “tooth”) and
−1
N[
Ω := [0, 1/2] × [0, 1] ∪ Ωj .
j=0

We define the phase u piecewise. Let ϕ be a smooth positive function such

that ϕ ≡ 0 on [0, 5/8] and ϕ ≡ 1 on [3/4, 1]. Then on each tooth we prescribe
u to behave the same way: for (x, y) ∈ Ωj , u(x, y) = y − jϕ(x)/N . On the rest
of the comb, we prescribe: for (x, y) ∈ [0, 1/2] × [0, 1], u(x, y) = y. Show that
|∂y u| = 1 on Ω and show that for λ sufficiently smaller than N −1 we have
|{(x, y) ∈ Ω s.t. |u(x, y)| < λ}| & N λ,
so that the constant necessarily depends on Ω.
Exercise 19. (F) In this exercise you will prove a result that is weaker than the full
asymptotics provided in Theorem 3.3 but which is sufficient for most applications.
In particular, under the same hypotheses (in particular, the non-degeneracy of the
critical point of u), you will prove the estimate
|Iψ (λ)| .u,ψ,d |λ|−d/2 .
The proof is quite similar to that of Theorem 3.4. For starters, let u, ψ satisfy the
hypotheses of Theorem 3.3. We expand |Iψ (λ)|2 into a double integral in dxdy and
change variables to (y, z) by introducing z = x − y. We obtain
ˆ ˆ
|Iψ (λ)|2 = eiλ(u(y+z)−u(y)) ψ(y + z)ψ(y)dydz.
Rd Rd

i) Verify the calculations above. ´

ii) The expression we have is of the form Rd Jλ (z)dz with Jλ an oscillatory
integral. You will show that for any N > 0 it is |Jλ (z)| .N (1 + λ|z|)−N .
Show that this will imply that |Iψ (λ)|2 . |λ|−d , hence concluding the estimate
we want.
iii) Now you will prove the claimed |Jλ (z)| .N (1 + λ|z|)−N using essentially
the same proof given for Proposition 2.1. The oscillatory term of Jλ (z) is
eiλ(u(y+z)−u(y)) ; calculate its ∇y gradient and use this information to craft a
differential operator D such that D(eiλ(u(y+z)−u(y)) ) = eiλ(u(y+z)−u(y)) .
iv) Your differential operator should have the form D = (iλ)−1 θz · ∇y for some
explicit Rd -valued function θz (y). Show by simple Taylor expansion arguments
that for any multi-index α one has
k∂yα θz (y)k .α |z|−1 .
24 OSCILLATORY INTEGRALS

v) Now use integration by parts N times to show that

ˆ
Jλ (z) = eiλ(u(y+z)−u(y)) (Dt )N (ψ(y + z)ψ(y))dy

and, using iv) above, conclude the estimate (Dt being the transpose of D, which
you will have to work out explicitely).
Exercise 20. (FF) In this exercise you will prove the method of stationary
phase in several variables, that is Theorem 3.3, following the same strategy used
in Exercise 13 (which you should attempt before attempting this one). Assume
x0 = 0.
αd
Let x = (x1 , . . . , xd ), and if α ∈ Nd is a multi-index let xα := xα
1 · . . . · xd .
2

P`
i) Assume the phase is purely a non-degenerate quadratic form Q(x) = j=1 x2j −
−kxk2 α
Pd 2
j=`+1 xj . Take ψ(x) = e x and show that (3.1) holds, that is
ˆ
2 X `,α
eiλQ(x) e−kxk xα dx ' λ−d/2−|α|/2 aj λ−j .
Rd j∈N

[hint: the integral factorises, then you can repeat what you did in Exercise
13.]
ii) Now we need to prove a result for a compactly supported ψ. Indeed, let η(x) a
smooth function compactly supported around the origin and let ψ(x) = xα η(x).
You will show that, with the same purely quadratic phase Q as above, we have
|Iψ (λ)| .α,η,Q |λ|−(d+|α|)/2 .
To do so, we will need to use integration by parts at some point with respect
to some ∂xk , and thus it will be convenient to localise in that direction. We
achieve this as follows: we roughly divide Rd into the cones
Γk := {x ∈ Rd s.t. kxk2 /d ≤ |xk |2 }
for k ∈ {1, . . . , d}; you can see that Γk is a cone with axis the xk -axis. Show
that
[d
Rd = Γk .
k=1
Using slightly larger cones
e k := {x ∈ Rd s.t. kxk2 /(2d) ≤ |xk |2 },
Γ
argue by a partition of unity that it suffices to reduce to the case where the
amplitude is ψ(x) · Gk (x) with Gk a function homogeneous of degree 0 (that
is, Gk (x) = Gk (x/kxk)), smooth away from the origin, supported in Γ e k and
identically 1 in a conical neighbourhood of the xk -axis (say, points x such that
(9/10)kxk2 ≤ |xk |2 ).
iii) This step will be the exact analogue of step ii) in Exercise 13. With phase
Q and amplitude xα η(x)Gk (x), split the region of integration smoothly into
kxk . δ and the complement. For the first part use a trivial estimate, and for
the second part use repeated integration by parts with respect to the differential
operator
1
Dk f (x) := ± ∂x .
2iλxk k
Optimizing in δ, conclude the estimate in ii). [hint: notice that Dk applied to
eiλQ returns exactly eiλQ .]
 OSCILLATORY INTEGRALS 25

iv) Using the same arguments as above, show that if g ∈ S(Rd ) is identically 0
near the origin, then for any N > 0
|Ig (λ)| .N,g |λ|−N .
v) For a generic amplitude ψ and phase Q as above, write
ˆ ˆ
2 2
iλQ(x)
ψ(x)dx = eiλQ(x) e−kxk ekxk ψ(x) η(x)dx


e

with η compactly supported and identically 1 on the support of ψ. Perform a

2
Taylor expansion of ekxk ψ(x) to some finite degree, then use i)-iv) above to
argue that (3.1) holds for this oscillatory integral.
vi) Finally, if u is a generic phase satisfying the hypotheses of Theorem 3.3, argue
by Morse’s Lemma that there is a smooth change of variables that lets us put
u in the form Q above (provided ψ has small enough support). Therefore,
conclude by v). [If you do not know about Morse’s lemma, it says more or less
precisely what we claimed; look it up, it is a very convenient tool.]
Exercise 21. (FFF) Oscillatory integrals are prototype solutions for many par-
tial differential equations:
1) Show that the oscillatory integral
ˆ
2
u(x, t) := ei(t|ξ| +x·ξ) ψ(ξ)dξ

is a solution for the Schrödinger equation i∂t u − ∆u = 0.

2) Show that the oscillatory integral
ˆ
u(x, t) := ei(t|ξ|+x·ξ) ψ(ξ)dξ

is a solution for the wave equation ∂t2 u − ∆u = 0.

3) Show that the oscillatory integral
ˆ
u(x, t) := ei(tmhξ/mi+x·ξ) ψ(ξ)dξ,

where hξi (the so-called japanese bracket) denotes (1 + |ξ|2 )1/2 , is a solution for
the Klein-Gordon equation ∂t2 u − ∆u + m2 u = 0 (m > 0 is a constant - the
mass).
Now consider the wave equation in Rd above. You will prove some decay estimates
for a packet of waves of frequency approximately 2k .
Let ψ(ξ) be a C ∞ function supported in the annulus 1 < |ξ| < 2 and radially
symmetric, for simplicity. Fix an integer k ∈ Z and consider the solution to the
wave equation ˆ ξ 
Φk (x, t) := ei(x·ξ+t|ξ|) ψ k dξ,
2
k k+1
which is frequency localised in 2 < |ξ| < 2 . You will show that for any N > 0
the above solution obeys the estimate
|Φk (x, t)| .N 2kd (1 + 2k ||x| − |t||)−N (1 + 2k |t|)−(d−1)/2 . (♣)
The proof will boil down to a number of applications of the non-stationary phase
principle, which will reduce matters to estimating the contribution of the critical
region of integration where the phase t|ξ| + x · ξ is stationary.
i) As a warm-up, give a physical interpretation of (♣) when t is large. For the
sake of the interpretation, replace the fast-decaying term (1 + 2k ||x| − |t||)−N
by the characteristic function 1[−1,1] (2k ||x| − |t||) and draw its support in a
space-time diagram. Show that the estimate is compatible with the energy
conservation law for waves.
26 OSCILLATORY INTEGRALS

ii) Show by rescaling that it suffices to prove (♣) for the case k = 0.
iii) Show that the symmetries of the problem allow one to assume that t > 0 and
x = (x1 , 0, . . . , 0) with x1 ≥ 0.
iv) Now you have an oscillatory integral with phase φ(ξ) = t|ξ| + x1 ξ1 . Show
by using the non-stationary phase principle in ξ1 that when t . 1 one has
|Φ0 (x1 , t)| .N (1 + x1 )−N for any N > 0. This proves (♣) in the small
time regime. Notice that the phase is not of the form λφ and therefore one
cannot use the non-stationary phase principle (Proposition 2.1) out of the box;
however, the proof adapts easily.
v) Assume then that t & 1. We concentrate on proving the localization in |x| − |t|.
Show, again by non-stationary phase principle in ξ1 , that if t > 4x1 or 41 x1 > t
then one has |Φ0 (x1 , t)| .N (1 + |x1 − t|)−N for any N > 0. Show that this
proves (♣) in this regime.
vi) Assume therefore that t & 1 and that 4x1 > t > 41 x1 . Show that the phase
φ(ξ) = t|ξ| + x1 ξ1 has critical points only if t = x1 and the set of such critical
points is contained in (−∞, 0]e1 . We expect that the main contribution to
the integral defining Φ0 comes from these critical points; therefore, smoothly
excise a thin cone with axis Re1 , say Γ := {ξ = (ξ1 , ξ 0 ) ∈ R × Rd−1 s.t. |ξ1 | >
(99/100)|ξ|}, from the support of ψ and show by non-stationary phase principle
in ξ 0 that the remaining integral (outside the cone) is dominated by ON ((1 +
t)−N ) (which is acceptable since it is dominated by (♣) in this regime).
vii) Show by non-stationary phase principle in ξ1 that one can do away with the
portion of the cone Γ where ξ1 > 0. Indeed, in that region |∇φ| & x1 ; show
that this contribution is bounded by ON ((1+x1 )−N ), which is again acceptable
in this regime.
Supp(ψ)

Critical region Γ

viii) Finally, we come to the critical region (the portion of the cone where |ξ| ∼ 1
and ξ1 < 0). You have shown that if x1 = t and ξ = (−ξ1 , 0) then ∇φ = 0. Now
show that, even when x1 6= t, if |ξ 0 | . t−1/2 then |φ(−ξ1 , ξ 0 )−φ(−ξ1 , 0)| ≤ 1/10
(this means that the phase is barely changing). Taking advantage of this
information, split (smoothly) the critical region dyadically into a sub-region
where |ξ 0 | . t−1/2 and sub-regions where |ξ 0 | ∼ 2j t−1/2 , where j ∈ N (and
2j t−1/2 . 1 because |ξ 0 | . 1 on the critical region). For the former, show
that ∂ξ1 φ(ξ) = x1 − t + O(1). This is good when |x1 − t| & t, but if x1 − t is
small then the integration by parts argument behind the non-stationary phase
principle accumulates some unfortunate powers of t (coming from the fact that
|∂ξj1 φ| ∼ t for j > 2). To compensate for this, observe that there is oscillation
in the ξ 0 directions as well, and thus you should apply a non-stationary phase
argument both in ξ1 and ξ 0 . Show therefore that this contribution is bounded
exactly by (♣). [If it helps, consider first the totally critical case x1 = t. ]
ix) For the remaining sub-regions, show that for each j the quantity ∂ξ1 φ(ξ) −
(x1 − t) is positive and has size ∼ 22j (recall 22j . t). Use once again the non-
stationary phase principle in ξ1 , ξ 0 to estimate the contribution of each region.
Finally, sum up in j and show that the result is controlled by (♣) again.
 OSCILLATORY INTEGRALS 27

Hints
Hints for Exercise 1: Take the estimate
|J(t)| . (1 + |t|)−(d−1)/2 (♥)
for granted, and follow these steps:
(i) Split the function f ∈ Lp (Rd ) as f (x) = f1 (x) + f2 (x) := f (x)1B(0,1) (x) +
f (x)(1 − 1B(0,1) (x)), where 1B(0,1) denotes the characteristic function of the
ball of radius 1 centered at 0. Argue that f1 is in L1 (Rd ) and therefore the
contribution to fb given by fb1 is continuous at ξ 6= 0.
(ii) Let ρ = |ξ| and write the radial Fourier transform g(ρ) of f2 according to
equation (1.1); then in the integral use a change of variable and argue that it
will suffice to show the continuity of the function Φ(ρ) := ρd g(ρ).
(iii) It suffices to show continuity of Φ from above and from below, so take ρ0 > ρ
and show that you can write (abusing notation a little, we write f2 for the
radial part of f2 (x) as well)
ˆ ρ0  
0 s d−1
Φ(ρ) − Φ(ρ ) = f2 s J(s)ds (3.2)
ρ ρ
ˆ +∞   
s  s 
+ f2 − f2 0 sd−1 J(s)ds. (3.3)
ρ ρ ρ
(iv) Fixing ρ, use Hölder’s inequality and (♥) (and the hypothesis on p) to show
´ +∞  s  d−1
that the integral ρ |f2 ρ |s |J(s)|ds is finite; hence deduce by standard
arguments that term (3.2) tends to 0 as ρ0 → ρ+ .
(v) Using again Hölder’s inequality and (♥), argue that term (3.3) also tends to 0
as ρ0 → ρ+ (you might want to do a change of variables s = ρt and then argue
by the continuity the Lp norm with respect to dilations). Putting everything
together, the proof is concluded.
PN
Hints for Exercise 2: Prove upper- and lower-bounds for n=1 r2 (n); tile the
plane into squares with centers in the Z2 lattice.
d
Hints for Exercise 3: Try with D = a(x) dx .
Hints for Exercise 4: Let a = 0, b = 1 for simplicity. Construct phases φλ
with oscillatory first derivatives where the oscillations happen at scale λ−1 . A good
choice would be φλ (t) = 2πt + λ−1 θ(λt) with θ a smooth periodic function of period
1. Show that ˆ 1
λI(λ) = λ ei(2πt+θ(t)) dt + O(1),
0
and furthermore show that you can choose θ such that the integral above is 6= 0
and such that |φλ 0 | > 1.
Hints for Exercise 5: Just adapt the proof of Proposition 2.2.
Hints for Exercise 6: Do a change of variable t = θs in I(λ), with θ a
parameter. If the inequality is to be invariant with respect to this transformation,
a certain condition involving α and k will have to be met.
Hints for Exercise 7: Let Jd be the function
ˆ π
Jd (t) := e−2πit cos θ (sin θ)d−2 dθ
0
and show by integration by parts that we have the recurrence relation
(2πit)2 Jd (t) = (d − 3)(d − 5)Jd−4 (t) − (d − 3)(d − 4)Jd−2 (t);
iterate this to reduce to the case d = 3, 2. For J3 , compute the integral directly and
get a bound of O(|t|−1 ); for J2 , show by Van der Corput’s lemma that we have a
28 OSCILLATORY INTEGRALS

bound of O(|t|−1/2 ) (besides the trivial bound of O(1)). Notice that you will need
´ π/4 ´ 3π/4 ´ π
to split the integral in 3 parts (say, 0 + π/4 + 3π/4 ) in order to apply Van der
Corput, since there is not a uniform lowerbound on φ0 or φ00 over the entire interval
[0, π].
Hints for Exercise 8: Find a rescaling of the ξ variable that allows you to
rewrite the phase as |x|3/2 φ(ξ) = |x|3/2 (sgn(x)ξ + ξ 3 ), which is of the form we
studied.
´
(ii) When x > 1, φ0 6= 0. Show that this implies | exp(i|x|3/2 φ(ξ))ϕ(ξ)dξ| .N
min{1, |x|−3N/2 } by non-stationary phase. For the ψj (ξ) terms instead, apply
the proof of the non-stationary phase principle to each term, with D the
differential operator such that D(eiφ(ξ) ) = eiφ(ξ) , to show that it contributes
ˆ
|x|−3N/2 |(Dt )N ψj (ξ)|dξ.

To show that this is summable in j, show that one will have

N
X
(Dt )N ψj (ξ) = cφ,k,N (ξ)2−jk ψ (k) (2−j ξ)
k=0

for some (integrable) functions cφ,k,N depending on φ0 that one can compute
explicitely, and conclude using the fact that j∈N 2−jk |ψ (k) (2−j ξ)| .k 1.
P

If you cannot convince yourself that (Dt )N ψj can be put in that form, try this
other way: show by induction in N that every term that appears in the expan-
sion of (Dt )N ψj is of the form (ψj )(k) (ξ)ξ ` (3ξ 2 + 1)−m ´(up to some constant)
and that moreover k − ` + 2m > 1. Then show that |(ψj )(k) (ξ)||ξ|` (3ξ 2 +
1)−m dξ ∼ 2j 2−(k−`+2m)j , which is summable in j.
(iii) When −1 < x < 1, φ might have critical points (if x < 0), but only in the
support of ϕ. Bound the ϕ contribution by O(1) trivially, and the contribu-
tion of the ψj by the same argument above (since the non-stationary phase
principle applies to them again). √
(iv) When x < −1, φ certainly has two critical points (±1/ 3) inside the support
of ϕ. For the ψj terms, argue as above that they contribute ON (|x|−3N/2 )
and can thus be ignored. For ϕ, isolate the critical points by splitting ϕ(ξ) =
ϕ(100ξ) + (ϕ(ξ) − ϕ(100ξ)) and similarly split the integral. Notice the support
of ϕ(100ξ) is [−1/50,
´ 1/50] and thus does not contain any critical points. For
the contribution exp(i|x|3/2 φ(ξ))(ϕ(ξ) − ϕ(100ξ))dξ, notice that |φ00 | & 1
and use Corollary 2.4 with k = 2.
Hints for Exercise 9: Here one needs to be a little clever: the point is to notice
that the phase 2πr cos θ has first derivative non-zero around the problematic point
θ = π/2. If ϕ denotes a smooth bump function supported in [−π/4, π/4], split J(r)
as J1 (r) + J2 (r) where
ˆ
J1 (r) := e−2πir cos θ ϕ(θ − π/2)dθ.

Argue by the non-stationary phase principle that |J1 (r)| .N (1 + r)−N and that
´R
therefore 0 rJ1 (r)dr = O(1). For J2 , use Fubini and integration by parts to
show that
ˆ R ˆ π  −2πiR cos θ
Re e−2πiR cos θ 1 
rJ2 (r)dr = − + − (1−ϕ(θ−π/2))dθ.
0 0 2πi cos θ (2πi cos θ)2 (2πi cos θ)2
Where 1 − ϕ(θ − π/2) does not vanish, one has that the 2nd derivative of the phase
cos θ is bounded from below; conclude therefore using Corollary 2.4 term by term.
Hints for Exercise 10:
 OSCILLATORY INTEGRALS 29

i) You need to rewrite the measure of the sublevel set as a nice oscillatory integral.
´b
Show that |{t ∈ (a, b) : |φ(t)| < λ}| = a 1(|φ(t)|/λ < 1)dt. Choose some
smooth non-negative bump function ϕ such that 1(|φ(t)|/λ < 1) ≤ ϕ(φ(t)/λ)
for every t and consequently replace the latter in the integral. Apply Fourier
inversion to the term ϕ(φ(t)/λ), followed by Fubini, and apply Van der Corput’s
lemma to the resulting oscillatory integral. ´
ii) Splitting (a, b) as suggested, (♠) takes care of | {|φ0 |<θ} exp(iλφ(t))dt| by tak-
ing the absolute value inside and {|φ0 | ≥ θ} consists of a bounded number of
intervals (how many? See Remark 2), for each of which one can apply Propo-
sition 2.2 to the associated oscillatory integral. Finally, choose θ in order to
optimize the resulting bound (that is, make it as small as possible).
iii) The base case k = 1 is easy. For the general k case, use the same strategy as
in ii): split (a, b) into those points where |φ0 (t)| < θ, to which the inductive
hypothesis applies, and those where |φ0 (t)| ≥ θ, which consist of boundedly
many intervals on each of which you can apply the base case.
iv) What does it mean for x ∈ Zp to have |x|p ≤ p−1 ? And what does it mean for
it to have |x|p ≤ p−2 ? etc.
Hints for Exercise 11:
i) Squeezing E into an interval can only make the left hand side of the inequality
smaller, so it is fine to do. Once E is an interval, just take uniformly distributed
points inside it as your xj ’s and perform the simple calculation that results.
ii) By Lagrange interpolation, you can find a polynomial P (X) of degree k such
that P (xj ) = f (xj ) for all j ∈ {0, . . . , k}. Argue that P (k) − f (k) must have
a zero inside (x0 , xk ). Then notice that P (k) (X) must be a constant, precisely
k! times the coefficient of X k . Calculate that coefficient explicitely (the con-
straints P (xj ) = f (xj ) give you a linear system of equations in the coefficients
of P , with a nice Vandermonde matrix appearing; solving the system with
Cramer’s rule is a breeze).
iii) Part i) gives you the points xj and for each j reversing the inequality one has
Q −1
i : i6=j |x i − xj | ≤ (2e)k |E|−k . Then use ii) together with this inequality,
the fact that |φ(k) | > 1 and that |φ(xj )| < λ for each j.
v) Remember Remark 2: how many intervals is {|φ0 | > θ} made of at most?
vi) Use (a, b) = (0, 1) and φ(t) = tk . To do this you will need complex integration.
´1 k k
Use a regularisation: evaluate instead 0 eit e−t dt and take  → 0 at the
end. To evaluate the integral, see it as the integral of holomorphic function
k
ez along the path Γ0 := {(i − )1/k t s.t. t ∈ [0, 1]} (fixing a branch of the k-th
root). Complete the path Γ0 to a well-chosen closed path Γ that includes part
of the real line and conclude using Cauchy’s integral theorem and some simple
k
estimates on the decay of |ez | along certain directions of the complex plane.
Hints for Exercise 12: The first step is as in the proof of Proposition 2.2, and
one has
ˆ b  eiλφ(b) ψ(b) eiλφ(a) ψ(a)  ˆ b  ψ 0
iλφ(t) iλφ(t)
e ψ(t)dt = − − e (t)dt.
a iλφ0 (b) iλφ0 (a) a iλφ0
Iterating this on the integral on the right hand side shows that it is of size O(λ−2 )
and therefore the above gives you a0 , b0 . Keep applying integration by parts until
necessary.
Hints for Exercise 13:
i) You should see the integral as the complex integral of (a multiple of) the
2
holomorphic function z m e−z along the path Γ0 := (1−iλ)1/2 R ⊂ C. Truncate
30 OSCILLATORY INTEGRALS

to Γ+ 0 := {(1 − iλ)
1/2
t : t ∈ [0, R]} and complete this path to a closed
path Γ that includes the real interval [0, R]. Use Cauchy’s integral theorem
2
and simple estimates on |z m e−z | to conclude that as R → ∞ this (plus an
equal contribution from a similarly defined Γ− 0 ) gives precisely what is claimed.
The power series expansion of (1 − iλ)−(m+1)/2 is standard, just write it as
λ−(m+1)/2 (λ−1 − i)−(m+1)/2 and expand the function (z − i)−(m+1)/2 in the z
variable. ´
ii) It is straightforward to show that |t|m |η(t)|ϕ(t/δ)dt .m,η δ m+1 . The other
term looks a bit nasty, but luckily one does not have to make all the compu-
tations implied. Let η(t)(1 − ϕ(t/δ)) =: ωδ (t) for convenience. Show that the
support of ωδ (t) and all its derivatives is contained in δ . |t| . 1. The inte-
gration by parts´ argument in the proof of Proposition 2.1 reduces matters to
estimating λ−N |(Dt )N (tm ωδ (t))|dt with Df (t) := (2it)−1 f 0 (t). (Dt )N is too
horrible to evaluate precisely (this is due to the fact that Dt is not a derivation,
that is it does not satisfy Leibniz’s rule; see Exercise 3), but we do not need
to. Show by induction on N that every term in the expansion of (Dt )N f is
of the form f (j) t−(2N −j) for j ∈ {0, . . . , N }. Apply this to function tm ωδ (t)
together with the general Leibniz’s rule (the binomial expansion of (f g)(j) ) to
conclude that (Dt )N (tm ωδ (t)) is of the form
m X
N
(j−k)
X
cj,k,N ωδ (t)t−(2N −m)+(j−k)
k=0 j=k

for some constants cj,k,N whose precise value does not concern us. Show that,
(j−k)
by the support of ωδ and the fact that 2N > m + 1, the integral of each
(j−k)
term is at most λ−N δ −(2N −m)+1 (careful with ωδ , when expanded this
−`
contains factors of δ for ` ≤ j − k). Therefore the quantity to be estimated
is controlled by δ m+1 + λ−N δ −(2N −m)+1 , and it remains for you to choose a
good δ that minimises this.
iii) Just repeat the second part of the above with δ ∼ 1.
2
iv) Taylor expansion gives you e−t ψ(t) = P (t) + tn+1 Rn (t) for some polynomial
P
´ of degree n and a remainder Rn (t) that is well-behaved. Write the term
exp(iλt2 ) exp(−t2 )P (t)η(t)dt as
ˆ ˆ
iλt2 −t2 2 2
− e e P (t)dt − eiλt e−t P (t)(1 − η(t))dt;

use i) on each monomial of P (t) to deal with the first integral, and use iii) on
the second integral to show it is an error term (show that exp(−t2 )P (t)(1−η(t))
is a Schwartz
´ function).
For exp(iλt2 )tn+1 (exp(−t2 )Rn (t)η(t))dt, use ii) to show that it is also an
error term. Choosing n large enough gives the result.
v) Simply show that under the given hypotheses s = |φ(t)|1/2 defines a diffeo-
morphism. Finally, if t = θ(s) denotes the inverse diffeomorphism, show that
ψ(θ(s))θ0 (s) satisfies the same properties as ψ. Performing the change of vari-
able t 7→ s concludes the argument.
vi) Repeat the proof you gave for i) basically word by word and adapt the rest.
Hints for Exercise 14:
i) The important thing is that dt/t is invariant with respect to changes of variables
of the form t 7→ αt.
ii) Since P is a polynomial, we have P (d) (t) = d!. With ψ(t) = 1/t, the constant
in Corollary 2.4 is ∼d 1/R = Od (1).
d ´1 d
iii) eiP (t) − eiPd−1 (t) = eiPd−1 (t) eit − 1 and 0 |t|


|t| dt = Od (1).
 OSCILLATORY INTEGRALS 31

iv) Steps i)-ii)-iii) together have the effect of removing the top monomial from P .
Repeat until there is only the last monomial c1 t.
vi) Here λ−1 = |c1 |−1 , but with ψ(t) = 1/t the constant in Corollary 2.4 evaluates
to |c1 |.
vii) We can add or subtract 0 for free to any quantity and bear in mind this is only
relevant when c1 > 1.
viii) You can consider ε, R fixed for the sake of the exercise. Calculate the Fourier
transform of the expression and show by Fubini that it equals
ˆ
2 dt
fb(ξ, η) e−2πi(ξt+ηt ) .
ε<|t|<R t
The phase ξt + ηt2 is a polynomial, so apply what you just proved to get an
L∞ bound for it. Conclude using Plancherel.
Hints for Exercise 15:
i) Just a Pboring high-school
P trigonometry exercise.
ii) Write n e(f (n)) as n g(n)(e(f (n)) − e(f (n − 1))) and check that every term
that appears in this sum appears exactly once in the right hand side (with the
correct sign) and viceversa.´n
iv) Write f (n) − f (n − 1) = n−1 f 0 (s)ds.
v) Apply triangle inequality to the left hand side, thus erasing the e(f (n)) fac-
tors. Then appeal to monotonicity to take the absolute values outside of the
summation and observe that the resulting expression is a telescopic sum, thus
only the first and last terms survive. Finally, use the trigonometric identity in
i) to evaluate g(N ), g(0), combine with periodicity of cot and iv).
vi) The remaining terms are bounded by |g(N )| + |g(1)| and are dealt with as in
the final part of v).
a) The bound on the length of J is an immediate consequence of the mean value
theorem.
b) Ak ∪ Bk has length 1 and they are all disjoint.
c) The first sum runs over those n such that |f 0 (n) − k| < δ. Since |f 00 | > λ, there
are at most O(δ/λ) such values of n.
For the second sum, verify that the hypotheses of (♦1 ) as in vi) apply.
d) Summing over k gives the factor of Cλ|J| + 1 by part b).
e) It must be the case that both C|I|λ1/2 < |I| and λ−1/2 < |I|...
Hints P for Exercise 16: Try with inner
α
P product αhP, Qi := [P (∇)]Q, where if
P (X) = α:|α|=k aα X then P (∇) = α:|α|=k aα ∂ . What does it mean that
k
h(ξ · X) , Qi = 0?
Hints for Exercise 17: Decompose the sphere smoothly in a bounded number
of pieces (see Exercise 20 part ii) for an idea how). Do a change of variables: a
point ω ∈ Sd−1 is given by ω = (x, φ(x)) where x ∈ Rd−1 and φ(x) = 1 − x21 − . . . −
´
x2d−1 )1/2 . Performing the change of variables, one has that Sd−1 exp(−2πitω1 )dσ(ω)
is a sum of integrals of the form
ˆ
e−2πitφ(x) S(x)ψ(x)dx,
Rd−1
where ψ is a smooth bump function supported around the origin and S(x)dx gives
the surface measure of Sd−1 in the x-coordinates. Justify all of the above, check
the Hessian of φ and apply Theorem 3.3 directly.
Hints for Exercise 18:
ii) just solve |u(x, y)| < λ for x, y.
iii) Show that for x close to 1 the function u(x, y) is constant in x and periodic
in y with period 1/N . Thus concentrate on a single tooth and show that for
32 OSCILLATORY INTEGRALS

these x the set of y in this tooth such that |u(x, y)| < λ has measure bounded
from below by & λ.
Hints for Exercise 19:
´
ii) Just perform the integration Rd (1 + λ|z|)−N dz with N large. If you prefer,
you can split Rd into dyadic annuli {z ∈ Rd : 2j < |z| ≤ 2j+1 }, estimate the
integral over each annulus and sum in j ∈ Z.
iii) ∇y eiλ(u(y+z)−u(y)) = iλeiλ(u(y+z)−u(y)) ∇y (u(y + z) − u(y)), so take the inner
∇ (u(y+z)−u(y))
product of this vector by vector (iλ)−1 |∇yy(u(y+z)−u(y))|2 to get the pure expo-
nential factor back.
∇ (u(y+z)−u(y))
iv) From the previous point, θz (y) is the vector |∇yy(u(y+z)−u(y))|2 . For the numer-
ator, observe that by hypothesis we can control |∂yβ u| .β,u,ψ 1 for any multi-
index β; show by the Fundamental Theorem of Calculus that ∂yα (u(y + z) −
´1
u(y)) = 0 ∇y (∂yα u(y+tz))·zdt and deduce that |∂yα (u(y+z)−u(y))| .α,u,ψ |z|.
For the denominator, use the hypothesis that ψ has small support to argue that
by Taylor expansion ∇y (u(y + z) − u(y)) is approximately [Hess(u)](y)z, and
use the hypothesis on Hess(u) to argue that this is &u,ψ |z| uniformly in y.
Finally, combine the above information and the generalized Leibniz’s rule to
estimate ∂yα θz .
v) Show that Dt f = (iλ)−1 ∇ · (f θz ) (that is, the divergence of the Rd -valued
function f θz ). Let Ψ(y, z) := ψ(y + z)ψ(y) for simplicity. Show by induction
that (Dt )N Ψ contains only terms of the form ∂jα00 Ψ · ∂jα11 θz · . . . · ∂jαNN θz with
α0 + α1 + . . . + αN = N , and deduce by iv) that |(Dt )N Ψ| is thus controlled
by a sum of terms of the form |z|−N |∂jα00 Ψ|.
Hints for Exercise 20:
i) The integral factorises into a product of integrals that are exactly like the ones
in i) of Exercise 13.
ii) The decomposition of Rd into cones follows from a trivial pigeonholing argu-
Pd
ment. Building the partition of unity 1 = k=1 Gk (x) is also a standard
exercise: simply take non-negative bump functions ρk on the sphere Sd−1
with support contained in Γ̃k ∩ Sd−1 and identically 1 on Γk , then define
Pd
−1
Gk (x) := ρk (x/kxk) j=1 ρj (x/kxk) and show this is C ∞ and satisfies
the other conditions.
iii) See the hints for ii) of Exercise 13, for this is extremely
´ similar. The trivial
estimate near the origin will give you a contribution |xα |φ(x/δ)dx . δ d+|α| .
Then you will end up having to estimate the integral of (Dkt )N (xα ωδ (x)Gk (x)),
where ωδ (x) = η(x)(1 − ϕ(x/δ)). You can again appeal to induction to prove
−(2N −j) j
that the terms of this expression are of the form xk ∂k (xα ωδ (x)Gk (x)),
j
then apply Leibniz’s rule to expand the ∂k derivative and conclude that since
all the terms contain factors of ∂k` Gk they are supported in Γ̃k , where crucially
it holds |xk | ∼ kxk. Performing the integration will give a total contribution of
O(λ−N δ −(2N −|α|)+d ), which combined with the trivial estimate and optimised
in δ will give you the desired estimate |Ixα η (λ)| . |λ|−(d+|α|)/2 .
iv) Again, repeat the second part of the proof you gave for step iii) above, but
with δ ∼ 1.
v) See the hints for part iv) of Exercise 13 and follow them closely.
vi) Morse’s lemma says precisely what we want: if f ∈ C ∞ (Rd ) is such that
f (0) = 0, ∇f (0) = 0 but det(Hess(f )) 6= 0 (where Hess(f ) denotes the Hessian
of f ), then there exists a neighbourhood U of 0 and a C ∞ -diffeomorphism
Φ : Rd → Rd such that Φ(0) = 0 and f ◦Φ−1 (y) = x21 +. . .+x2k −x2k+1 −. . .−x2d
for all y ∈ Φ(U ). Moreover, (k, d − k) is the signature of Hess(f ).
 OSCILLATORY INTEGRALS 33

Use the change of variable provided by the diffeomorphism Φ to turn the phase
u(x) into a quadratic phase and conclude by appealing to the previous steps.
Hints for Exercise 21:
i) At time t the wave solution is concentrated in a shell of width ∼ 1 and radius
∼ t, and it has amplitude ∼ t−(d−1)/2 . Multiply the square of the amplitude
by the volume occupied by the wave to get an estimate for the energy.
ξ
iv) The gradient of the phase is ∇φ(ξ) = t |ξ| + x1 e1 , so if t . 1 and x1 & 1 one
has ∂ξ1 φ is bounded from below by & x1 . Show that for higher derivatives
one has |∂ξk1 φ| . 1. Apply the integration by parts argument of Proposi-
tion 2.1 to the integral in ξ1 with respect to the differential operator D1 =
(i∂ξ1 φ(ξ))−1 ∂ξ1 . Show by induction in N that |(D1t )N ψ| .N,ψ x−N 1 (by now
you should know how to proceed: show that (D1t )N ψ consists of terms of the
Qm
form (∂ξ1 φ)−` ∂ξβ11 ψ k=2 (∂ξk1 φ)βk where ` ≥ N ...). Finish by integrating in the
remaining variables ξ2 , . . . , ξd .
v) Show that one always has |∂ξ1 φ(ξ)| & |x1 − t| & t in this regime. Then repeat
the analysis done in iv), but be more careful this time: now you can only
say (show this) that |∂ξk1 φ| . t. In order not to lose the decay, you will have
Qm
to show by induction that the terms of the form (∂ξ1 φ)−` ∂ξβ11 ψ k=2 (∂ξk1 φ)βk
t N
P m
Pm in the expansion of (D1 ) ψ actually satisfy ` − k=1 kβk−N
appearing = 0 and
β1 + k=2 (k − 1)βk = N , thus yielding the correct decay of |x1 − t| .
vi) For the smooth excision of a cone, refer to the hints for parts ii)-iii) of Exercise
20. For the estimate, observe that ignoring the ξ1 component one has |∇φ(ξ)| &
t|ξ 0 |/|ξ| and outside the cone one has |ξ 0 | & |ξ|. Upon a further smooth partition
of unity into cones with axes the coordinate directions ξ2 , . . . , ξd , one can reduce
to the case where |ξj | & |ξ| and thus |∂ξj φ(ξ)| & t. Then apply the same
argument as in v) above with differential operator Dj = (i∂ξj φ(ξ))−1 ∂ξj to
conclude.
vii) Same reasoning as in vi).
viii) You are allowed to use Taylor expansion of |ξ| = ξ1 (1 + (|ξ 0 |2 /ξ12 ))1/2 in terms
of |ξ 0 |2 /ξ12 , both for estimating the difference in the phases and for estimat-
ing ∂ξ1 φ. The non-stationary phase principle in both ξ1 , ξj (after once again
using a smooth partition of unity to reduce to´ the case where |ξj | & |ξ|)
will result in having to estimate (for example) |(Djt D1t )N ψ(ξ)|dξ. This is
quite excruciating, but you can again use an inductive argument to show that
0
the integrand consists entirely of terms of the form (∂j φ)−` (∂1 φ)−` · ∂jα ∂1β ψ ·
k k0 γ(k,k0 )
Q
k+k0 ≥2 (∂j ∂1 φ) where the various parameters involved satisfy certain
favourable contraints. Then one can see that |∂1 φ| & |x1 − t|, |∂12 φ| . 1 (this
0
because of the restriction to |ξ 0 | . t−1/2 ), |∂j φ| ∼ t and in general |∂jk ∂1k φ| . t
for all k + k 0 ≥ 2. Using these facts and the induction one can verify that
|(Djt D1t )N ψ(ξ)| . |x1 − t|−N . The factor of t−(d−1)/2 will come from the fact
that the region of integration is approximately a cylinder of length 1 and radius
t−1/2 .
ix) This part is a repetition of part viii) above essentially, except for the fact that
now we can assume ξj ∼ 2j t−1/2 instead and consequently one has to update
the upper- and lower-bounds on the derivatives of the phase accordingly.

If you find typos, mistakes or generally have questions please send an email to
[email protected]