232 CHAPTER 10. IMPROPER INTEGRALS

= e(n)+E (n)+∫ bn

0

11+ t2 dt−

∫ bn

0e−bnt (cosbn)+ t (sinbn)

1+ t2 dt

= e(n)+E (n)+∫ bn

0

11+ t2 dt−

∫ bn

0e−bnt 1√

1+ t2cos(bn−φ t)dt

where φ t is a phase shift. This is because(1√

1+ t2,

t√1+ t2

)= (cosφ t ,sinφ t) some φ t

so the formula for cos(bn−φ) is used. This last integral satisfies∣∣∣∣∫ bn

0e−bnt 1√

1+ t2cos(bn−φ t)dt

∣∣∣∣≤ ∫ bn

0e−bntdt ≤ 1

bn

Therefore, include it in E (n) and it follows that∫ bn

an

sinxx

dx = e(n)+E (n)+∫ bn

0

11+ t2 dt

where limn→∞ (e(n)+E (n)) = 0. Taking a limit,

limn→∞

∫ bn

an

sinxx

dx = limn→∞

(e(n)+E (n)+

∫ bn

0

11+ t2 dt

)= π/2 (10.2)

A much shorter way to verify this identity is in the exercises but it depends on a theoremwhich has not been discussed yet and to use it, you need to know the existence of theDirichlet integral which is obtained here as part of the argument.

For I an interval let XI (t)≡{

1 if t ∈ I0 if t /∈ I .

Lemma 10.1.2 Suppose f is Riemann integrable on [a,b] . Then for each ε > 0, thereis a step function s which satisfies |s(x)| ≤ | f (x)| and∫ b

a| f (x)− s(x)|dx < ε

Also there exists a continuous function h which is 0 at a and b such that |h| ≤ | f | and∫ b

a| f (x)−h(x)|2 dx < ε

Proof: First suppose f (x)≥ 0. Then by Theorem 9.3.10, there is a lower sum L( f ,P)such that ∣∣∣∣∫ b

af dt−L( f ,P)

∣∣∣∣≤ (U ( f ,P)−L( f ,P))< ε

Let s correspond to this lower sum. That is, if L( f ,P) = ∑mk=1 mk (xk− xk−1) , you let

s(x)≡ ∑mk=1 mkXIk (x) where I1 = [x0,x1] , Ik = (xk−1,xk] for k ≥ 2. Then∫ b

a| f (x)− s(x)|dx =

∫ b

a( f (x)− s(x))dx =

∫ b

af dt−L( f ,P)< ε

232 CHAPTER 10. IMPROPER INTEGRALSbn bn (cos by) +t (sin by)= E —dt — Pat " “dte(n) + n+ f 1+? i ° 1+?= e(n) +B (n+ [ nae [or cos —@,)dt7 0 1+f7 0 V1l4+e somwhere @, is a phase shift. This is because(= ==) = (cos @,,sing,) some 9,so the formula for cos (b, — @) is used. This last integral satisfiesDn 1 bn 1—b,t —Dnten cos (by — a < [ e "dt < —if V14e On ~ $1) at) s 0 ~ bnTherefore, include it in FE (n) and it follows thatbn sinx bn—dx = E —— dt[, mel El) +/f 1+?where lim,-;.. (e (m) + E (n)) = 0. Taking a limit,. bn sinx . bn]A much shorter way to verify this identity is in the exercises but it depends on a theoremwhich has not been discussed yet and to use it, you need to know the existence of theDirichlet integral which is obtained here as part of the argument.. _ f liftelFor J an interval let 27 (t) = { Oifrgr”Lemma 10.1.2 Suppose f is Riemann integrable on |a,b]. Then for each € > 0, thereis a step function s which satisfies |s (x)| < |f (x)| andb[ lf) -s@lar<eaAlso there exists a continuous function h which is 0 at a and b such that |h| < |f| andp 2[lf -he@Par<eaProof: First suppose f (x) > 0. Then by Theorem 9.3.10, there is a lower sum L(f, P)such that[ ta—uirr) <(U(f,P)-L(,P)) <€Let s correspond to this lower sum. That is, if L(f,P) = D7, mx (xe —xXe-1), you lets(x) = YL, me 21, (x) where 1, = [x0,.x1] Ak = (Xe-1,%x] for k > 2. Thenba[ires-sepiac= [re)-seyacs [far—L,") <e