Bourbaki

Lemma 11

Tag 031F, permanent. Cite it as [Bourbaki, Tag 031F].

Théories spectrales, chapter IV, § 3. Distributions et distributions tempérées

Let φ∈S(Rn)\varphi \in \mathscr{S}(\mathbf{R}^n). The function F(φ)\mathscr{F}(\varphi ) is indefinitely differentiable on Rn\mathbf{R}^n and one has

F(Xαφ)=(−2iπ)−∣α∣∂α(F(φ)) \mathscr{F}(X^{\alpha}\varphi ) = (-2i\pi )^{-|\alpha|}\partial^{\alpha}(\mathscr{F}(\varphi ))
F(∂αφ)=(2iπ)∣α∣XαF(φ) \mathscr{F}(\partial^{\alpha}\varphi ) = (2i\pi )^{|\alpha|}X^{\alpha}\mathscr{F}(\varphi )

for every α\alpha in Nn\mathbf{N}^n.

One may suppose that n⩾1n\geqslant 1. Let φ∈S(Rn)\varphi \in \mathscr{S}(\mathbf{R}^n). The function defined by (x,y)↦φ(x)(x, y)\mapsto \varphi (x) exp(2iπx⋅yi\pi x\cdot y) from Rn×Rn\mathbf{R}^n\times \mathbf{R}^n into C\mathbf{C} satisfies the hypotheses of Corollary 1 of IV, p. 198 for every integer kk. The Fourier transform of φ\varphi is therefore indefinitely differentiable and satisfies

∂α(Fφ)(y)=(−2iπ)∣α∣∫Rnxαφ(x)\partial^{\alpha}(\mathscr{F}\varphi )(y) = (-2i\pi )^{|\alpha|}\int_{\mathbf{R}^n}x^{\alpha}\varphi (x) exp(−2iπx⋅y-2i\pi x\cdot y)dμ(x)d\mu(x)

for every y∈Rny\in \mathbf{R}^n, which implies the first formula.

Let us prove the second formula. By induction on ∣α∣|\alpha |, it is enough to do so when ∣α∣=1|\alpha |= 1, and one is easily reduced to the case α=(1,0,...,0)\alpha = (1,0, . . . ,0). Write every x∈Rnx\in \mathbf{R}^n in the form x=(x1,x′)x= (x_1, x') with x′∈Rn−1x'\in \mathbf{R}^{n-1}, and denote by μ1\mu_1 (resp. μ′\mu') the Lebesgue measure on Rn−1\mathbf{R}^{n-1} (resp. R\mathbf{R}). By the Lebesgue-Fubini theorem (INT, V, p. 96, § 8, no4^o4, th. 1), for every y=(y1,y′)∈R×Rn−1y= (y_1, y')\in \mathbf{R}\times \mathbf{R}^{n-1}, we get

F(∂1φ)(y)=∫\mathscr{F}(\partial_1\varphi )(y) =\int exp(−2iπx′⋅y′-2i\pi x'\cdot y')

×Rn(∫R−1(∂1φ)(x1,x′)\times^{\mathbf{R}^n}(\int^{^{-1}}_{\mathbf{R}}(\partial_1\varphi )(x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)dμ1(x1))dμ′(x′)d\mu_1(x_1))d\mu'(x').

For every compact interval [a,b][a, b] in R\mathbf{R} and every x′∈Rn−1x'\in \mathbf{R}^{n-1}, one has

∫ab(∂1φ)(x1,x′)\int_a^b(\partial_1\varphi )(x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)dμ1(x1)=d\mu_1(x_1) =

[φ(x1,x′)[\varphi (x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)]ab]^b_a

+2iπy1∫abφ(x1,x′)+ 2i\pi y_1\int_a^b\varphi (x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)dμ1(x1)d\mu_1(x_1) by integration by parts (FVR, II, p. 10, formula (10)). When aa tends to −∞-\infty and bb tends to +∞+\infty, the first term of the second member converges to 0 since φ∈S(Rn)\varphi \in \mathscr{S}(\mathbf{R}^n). The second term converges, by the Lebesgue theorem (INT, IV, p. 137, § 3, no7^o7, th. 6), to

2iπy1∫Rφ(x1,x′)2i\pi y_1\int_{\mathbf{R}}\varphi (x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)dμ1(x1)d\mu_1(x_1),

since the mapping x1↦φ(x1,x′)x_1\mapsto \varphi (x_1, x') is integrable on R\mathbf{R}. As x1↦∂1φ(x1,x′)x_1\mapsto \partial_1\varphi (x_1, x') is also integrable on R\mathbf{R}, one deduces that

∫R∂1φ(x1,x′)\int_{\mathbf{R}}\partial_1\varphi (x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)dμ1(x1)d\mu_1(x_1)

=2iπy1∫Rφ(x1,x′)= 2i\pi y_1\int_{\mathbf{R}}\varphi (x_1, x') exp(−2iπx1y1-2i\pi x_1y_1)dμ1(x1)d\mu_1(x_1) and finally, applying the Lebesgue-Fubini theorem again, one concludes that

F(∂1φ)(y)=2iπy1∫Rnφ(x)\mathscr{F}(\partial_1\varphi )(y) = 2i\pi y_1\int_{\mathbf{R}^n}\varphi (x) exp(−2iπx⋅y-2i\pi x\cdot y)dμ(x)d\mu(x),

as desired.

Bourbaki sets the proof directly after the statement, and the corpus does not separate them, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Nothing.

Machine translation of the English, by gpt-5-mini, gpt-5.4, against glossary version 34. Not checked by a person. That model is a cut down version of the one the rest of this language was written by.