238 CHAPTER 10. IMPROPER INTEGRALS

Now consider g(0+) I2. It equals g(0+) 2π

∫ rh0

sin(u)u du so its limit as r→ ∞ is g(0+).

It is just the Dirichlet integral again.Finally, consider I3. For t ≥ h, g(t)

t is in L1 (h,a) and so, the 0 extension off [h,∞) is inL1 ([0,a)) . By the Riemann Lebesgue lemma, of Theorem 10.2.1, this integral I3 convergesto 0 as r→ ∞.

With this, here is a different version of Theorem 10.2.5.

Corollary 10.2.8 Suppose that g ∈ L1 (R) and that at some x, g is of finite total varia-tion on [x−δ ,x+δ ] for some δ > 0. Then

limr→∞

∫∞

0

sin(ur)u

(g(x−u)+g(x+u)

2

)du =

g(x+)+g(x−)2

.

Proof: This follows from Lemma 10.2.7 applied to u→ g(x−u)+g(x+u)2 .

10.3 The Gamma FunctionRecall the definition of an improper integral specialized to (0,∞). You let an ↓ 0,bn ↑ ∞

and∫

0 f (t)dt = limn→∞

∫ bnan

f (t)dt.

Definition 10.3.1 Whenever α > 0, Γ(α)≡∫

0 e−ttα−1 dt.

Lemma 10.3.2 The improper integral∫

0 e−ttα−1 dt exists for each α > 0.

Proof: Let f (t) ={

tα−1 if t ≤ 1Ce−t/2 if t > 1

where C is chosen large enough that for t >

1,Ce−t/2 > e−ttα−1. Then obviously f ∈ L1 (0,∞) and f (t) ≥ e−ttα−1. Also, if [α,β ] ⊆(0,∞) , ∫

β

α

e−ttα−1 dt ≤∫

0f (t)dt < ∞

and so t → e−ttα−1 is in L1 (0,∞) so the improper integral exists as claimed thanks toProposition 10.0.2.

This gamma function has some fundamental properties described in the following pro-position. In case the improper integral exists, we can obviously compute it in the formlimδ→0+

∫ 1/δ

δf (t)dt which is used in what follows. Thus also the usual algebraic proper-

ties of the Riemann integral are inherited by the improper integral.

Proposition 10.3.3 For n a positive integer, n! = Γ(n+1). In general, one has thefollowing identity: Γ(1) = 1,Γ(α +1) = αΓ(α)

Proof: First of all, Γ(1) = limδ→0∫

δ−1

δe−tdt = limδ→0

(e−δ − e−(δ

−1))= 1. Next,

for α > 0,

Γ(α +1) = limδ→0

∫δ−1

δ

e−ttα dt = limδ→0

[−e−ttα |δ

−1

δ+α

∫δ−1

δ

e−ttα−1dt

]

= limδ→0

(e−δ

δα − e−(δ

−1)δ−α +α

∫δ−1

δ

e−ttα−1dt

)= αΓ(α)

238 CHAPTER 10. IMPROPER INTEGRALSrh sin(u)Now consider g (0+) h. It equals g (0+) 2 {j" =It is just the Dirichlet integral again.Finally, consider /3. For t > h, al is in L' (h,a) and so, the 0 extension off [,-) is inL' ({0,a)). By the Riemann Lebesgue lemma, of Theorem 10.2.1, this integral J; convergestoOasr>o0. &With this, here is a different version of Theorem 10.2.5.du so its limit as r + is g(0+).Corollary 10.2.8 Suppose that g € L' (R) and that at some x, g is of finite total varia-tion on [x — 6,x+ 6] for some 6 > 0. Then_ 2 (@sin(ur) (g(—u)+g(x+u) g(xt+) +8 (x-)lim — ( 5) ) ay = SEroo TT JO uProof: This follows from Lemma 10.2.7 applied to u > aenu) ralety) |10.3. The Gamma FunctionRecall the definition of an improper integral specialized to (0,00). You let a, | 0,b, + ©and [5° f (t)dt = lino J?" f (t) dt.Definition 10.3.1 Whenever a > 0, T(a)= fo et t™! dt.Lemma 10.3.2 The improper integral {y e~'t®—! dt exists for each a > 0.tel ift<1Proof: Let f(t) = { Ce“ if t > 11,Ce/? > et, Then obviously f € L! (0,00) and f (t) > et*!. Also, if [a, B] C(0, ee),where C is chosen large enough that for t >B co| etre at < | f (t)dt <0a 0and so t + e~‘t®~! is in L' (0,00) so the improper integral exists as claimed thanks toProposition 10.0.2.This gamma function has some fundamental properties described in the following pro-position. In case the improper integral exists, we can obviously compute it in the formlims_504 ; /6 f (t) dt which is used in what follows. Thus also the usual algebraic proper-ties of the Riemann integral are inherited by the improper integral.Proposition 10.3.3 For n a positive integer, n! =T(n+1). In general, one has thefollowing identity: T (1) = 1,1 (a@+1) = al (a)Proof: First of all, P(1) = limg_49[® e~tdt = lims_40 (68 -e (@)) = |. Next,for a > 0,t) _ 5T(a+1) = lim e't%dt = lim ee ‘+a | cut60/5 50 J6- o= lim Gua sea | cue ta = al (a)56-0