TheoremBase

The derivatives come from the difference-quotient estimate for the complex exponential at the point ix, by taking real and imaginary parts with Euler's formula; continuity follows from differentiability; and the power series are the real and imaginary parts of the even partial sums of the exponential series at it.

Proof

Routine facts are used without comment (Routine Facts, Continued: Typed Arithmetic, Maps and Predicativity), and the earlier and the new versions of the definitions involved are read through The Earlier and the New Versions of the Definitions of the Natural Numbers, the Complex Numbers, the Conjugate and the Modulus Agree, and the Two Versions of Real Correspond (ii), The Earlier and the New Versions of the Definitions of the Natural Numbers, the Complex Numbers, the Conjugate and the Modulus Agree, and the Two Versions of Real Correspond (iv), The Earlier and the New Versions of the Definitions of the Derivative on an Interval and of the Riemann Integral Agree (ii), The Earlier and the New Versions of the Definitions of Sequences, Limits, Cauchy Sequences, Series and Iterated Operations Agree (i), The Earlier and the New Versions of the Definitions of Sequences, Limits, Cauchy Sequences, Series and Iterated Operations Agree (ii), The Earlier and the New Versions of the Definitions of Sequences, Limits, Cauchy Sequences, Series and Iterated Operations Agree (iii) and The Earlier and the New Versions of the Definitions of Sequences, Limits, Cauchy Sequences, Series and Iterated Operations Agree (v). Throughout, exp⁑:Cβ†’C\exp:\mathbb{C}\to\mathbb{C} is the complex exponential function, so that cos⁑s=Re⁑exp⁑(is)\cos s=\operatorname{Re}\exp(is) and sin⁑s=Im⁑exp⁑(is)\sin s=\operatorname{Im}\exp(is) for every s∈Rs\in\mathbb{R} (The Real Sine and Cosine Functions (i), The Real Sine and Cosine Functions (ii)).

Two facts about the modulus. As i=0+1β‹…ii=0+1\cdot i, ∣i∣=02+12=1|i|=\sqrt{0^{2}+1^{2}}=1 by The Modulus of a Complex Number (i), and on real numbers the modulus is the absolute value (The Modulus of a Complex Number (ii)); so multiplicativity (Conjugation and Modulus of Complex Numbers: Field Automorphism, Real and Imaginary Parts, Multiplicativity, Inverses, Component Bounds, Triangle Inequalities and Finite Sums (iv)) gives, for every v∈Cv\in\mathbb{C} and s∈Rs\in\mathbb{R},

∣iv∣=∣v∣,∣is∣=∣s∣.(1)|iv|=|v|,\qquad |is|=|s|. \tag{1}

Derivatives. Fix x∈Rx\in\mathbb{R}. By Euler's formula (Euler's Formula: the Complex Exponential on the Imaginary Axis, Its Modulus and Its Conjugate (i)),

iexp⁑(ix)=i(cos⁑x+isin⁑x)=βˆ’sin⁑x+icos⁑x,(2)i\exp(ix)=i(\cos x+i\sin x)=-\sin x+i\cos x, \tag{2}

using i2=βˆ’1i^{2}=-1 (The Complex Numbers, with the Real Numbers Identified with a Subset of the Complex Numbers (vii)).

Let Ξ΅>0\varepsilon>0 be given; then choose

δ=min⁑(1,Ρe),\delta=\min\Bigl(1,\frac{\varepsilon}{e}\Bigr),

which is positive because e=exp⁑R(1)>0e=\exp_{\mathbb{R}}(1)>0 (Complex Analysis: Standing Notation (iv), Basic Properties of the Exponential Function (iii)). Let u∈Ru\in\mathbb{R} with 0<∣uβˆ’x∣<Ξ΄0<|u-x|<\delta, put h=uβˆ’xβ‰ 0h=u-x\neq0, and set

w=exp⁑(iu)βˆ’exp⁑(ix)hβˆ’iexp⁑(ix)=i[exp⁑(ix+ih)βˆ’exp⁑(ix)ihβˆ’exp⁑(ix)],w=\frac{\exp(iu)-\exp(ix)}{h}-i\exp(ix)=i\left[\frac{\exp(ix+ih)-\exp(ix)}{ih}-\exp(ix)\right],

where iu=ix+ihiu=ix+ih and ih≠0ih\neq0. The difference-quotient estimate for the exponential (Estimates for the Complex Exponential and Its Continuity (iii)), applied at the point ixix with the increment ihih, together with (1), gives

∣wβˆ£β‰€βˆ£exp⁑(ix)βˆ£β€‰βˆ£ihβˆ£β€‰exp⁑R(∣ih∣)=∣hβˆ£β€‰exp⁑R(∣h∣),|w|\le|\exp(ix)|\,|ih|\,\exp_{\mathbb{R}}(|ih|)=|h|\,\exp_{\mathbb{R}}(|h|),

since ∣exp⁑(ix)∣=1|\exp(ix)|=1 (Euler's Formula: the Complex Exponential on the Imaginary Axis, Its Modulus and Its Conjugate (ii)). As ∣h∣<1|h|<1, the real exponential being increasing (Basic Properties of the Exponential Function (ix)) gives exp⁑R(∣h∣)<exp⁑R(1)=e\exp_{\mathbb{R}}(|h|)<\exp_{\mathbb{R}}(1)=e, and therefore

∣w∣<eβ€‰βˆ£h∣<eβ‹…Ξ΅e=Ξ΅.|w|<e\,|h|<e\cdot\frac{\varepsilon}{e}=\varepsilon .

Now take real and imaginary parts of ww. For the real number hβ‰ 0h\neq0 and v∈Cv\in\mathbb{C}, Re⁑(v/h)=(Re⁑v)/h\operatorname{Re}(v/h)=(\operatorname{Re}v)/h and Im⁑(v/h)=(Im⁑v)/h\operatorname{Im}(v/h)=(\operatorname{Im}v)/h, and real and imaginary parts are additive (Conjugation and Modulus of Complex Numbers: Field Automorphism, Real and Imaginary Parts, Multiplicativity, Inverses, Component Bounds, Triangle Inequalities and Finite Sums (ii)). Applying this to v=exp⁑(iu)βˆ’exp⁑(ix)v=\exp(iu)-\exp(ix) and using (2),

Im⁑w=sin⁑uβˆ’sin⁑xuβˆ’xβˆ’cos⁑x,Re⁑w=cos⁑uβˆ’cos⁑xuβˆ’x+sin⁑x.\operatorname{Im}w=\frac{\sin u-\sin x}{u-x}-\cos x,\qquad \operatorname{Re}w=\frac{\cos u-\cos x}{u-x}+\sin x .

Both are bounded in absolute value by ∣w∣|w| (Conjugation and Modulus of Complex Numbers: Field Automorphism, Real and Imaginary Parts, Multiplicativity, Inverses, Component Bounds, Triangle Inequalities and Finite Sums (vi)), so

∣sin⁑uβˆ’sin⁑xuβˆ’xβˆ’cos⁑x∣<Ξ΅,∣cos⁑uβˆ’cos⁑xuβˆ’xβˆ’(βˆ’sin⁑x)∣<Ξ΅\Bigl|\frac{\sin u-\sin x}{u-x}-\cos x\Bigr|<\varepsilon,\qquad \Bigl|\frac{\cos u-\cos x}{u-x}-(-\sin x)\Bigr|<\varepsilon

for every u∈Ru\in\mathbb{R} with 0<∣uβˆ’x∣<Ξ΄0<|u-x|<\delta. Thus cos⁑x\cos x is a derivative of sin⁑\sin at xx and βˆ’sin⁑x-\sin x is a derivative of cos⁑\cos at xx (Derivative of a Real Function on an Interval (i)); so sin⁑\sin and cos⁑\cos are differentiable at xx with sin⁑′(x)=cos⁑x\sin'(x)=\cos x and cos⁑′(x)=βˆ’sin⁑x\cos'(x)=-\sin x (Derivative of a Real Function on an Interval (ii)). As xx was arbitrary, sin⁑\sin and cos⁑\cos are differentiable on R\mathbb{R} (Derivative of a Real Function on an Interval (iii)), and their derivatives sin⁑′,cos⁑′:Rβ†’R\sin',\cos':\mathbb{R}\to\mathbb{R} take the values stated in (ii) (Derivative of a Real Function on an Interval (iv)).

Continuity. By (ii), sin⁑\sin and cos⁑\cos are differentiable at every x∈Rx\in\mathbb{R}, so each is continuous at every xx relative to R\mathbb{R}, because differentiability implies continuity (The Increment Form of the Derivative, and Differentiability Implies Continuity (ii)). By the characterization of continuity on a set in The Real Line: Standing Notation and Background for Calculus on Intervals (vi), sin⁑\sin and cos⁑\cos are continuous on R\mathbb{R}. This proves (i).

Power series. Fix t∈Rt\in\mathbb{R}. Let (an)(a_{n}) be the sequence in C\mathbb{C} with

an=(it)nβˆ’1(nβˆ’1)!(n∈N)a_{n}=\frac{(it)^{n-1}}{(n-1)!}\qquad(n\in\mathbb{N})

(The Complex Exponential Series Converges Absolutely, and Its Sum Is a Function on β„‚ (i)), and let sn=βˆ‘j=1najs_{n}=\sum_{j=1}^{n}a_{j} be its partial sums (Series of Complex Numbers: Partial Sums, Convergence, the Sum and Absolute Convergence (i)). For k∈Nk\in\mathbb{N} write

ck=(βˆ’1)kβˆ’1t2kβˆ’2(2kβˆ’2)!,Οƒk=(βˆ’1)kβˆ’1t2kβˆ’1(2kβˆ’1)!,c_{k}=\frac{(-1)^{k-1}t^{2k-2}}{(2k-2)!},\qquad \sigma_{k}=\frac{(-1)^{k-1}t^{2k-1}}{(2k-1)!},

and let Cm=βˆ‘k=1mckC_{m}=\sum_{k=1}^{m}c_{k} and Sm=βˆ‘k=1mΟƒkS_{m}=\sum_{k=1}^{m}\sigma_{k} be the partial sums of these two real sequences (Series of Real Numbers: Partial Sums, Convergence, the Sum and Absolute Convergence (i)).

Since i2=βˆ’1i^{2}=-1 (The Complex Numbers, with the Real Numbers Identified with a Subset of the Complex Numbers (vii)), we have (it)2kβˆ’2=(i2)kβˆ’1t2kβˆ’2=(βˆ’1)kβˆ’1t2kβˆ’2(it)^{2k-2}=(i^{2})^{k-1}t^{2k-2}=(-1)^{k-1}t^{2k-2} and (it)2kβˆ’1=i (βˆ’1)kβˆ’1t2kβˆ’1(it)^{2k-1}=i\,(-1)^{k-1}t^{2k-1}, so

a2kβˆ’1=ck,a2k=iΟƒk(k∈N).(3)a_{2k-1}=c_{k},\qquad a_{2k}=i\sigma_{k}\qquad(k\in\mathbb{N}). \tag{3}

We claim that

s2m=Cm+iSm(m∈N).(4)s_{2m}=C_{m}+iS_{m}\qquad(m\in\mathbb{N}). \tag{4}

For m=1m=1, s2=a1+a2=1+it=C1+iS1s_{2}=a_{1}+a_{2}=1+it=C_{1}+iS_{1}. If (4) holds for mm, then by (3)

s2m+2=s2m+a2m+1+a2m+2=(Cm+cm+1)+i(Sm+Οƒm+1)=Cm+1+iSm+1,s_{2m+2}=s_{2m}+a_{2m+1}+a_{2m+2}=(C_{m}+c_{m+1})+i(S_{m}+\sigma_{m+1})=C_{m+1}+iS_{m+1},

so (4) holds for all mm by induction. As CmC_{m} and SmS_{m} are real, (4) gives

Re⁑s2m=Cm,Im⁑s2m=Sm(5)\operatorname{Re}s_{2m}=C_{m},\qquad \operatorname{Im}s_{2m}=S_{m} \tag{5}

(Real and Imaginary Parts of a Complex Number (i)).

The exponential series at itit converges (The Complex Exponential Series Converges Absolutely, and Its Sum Is a Function on β„‚ (ii)), and its sum is exp⁑(it)\exp(it) (The Complex Exponential Function (i)), so snβ†’exp⁑(it)s_{n}\to\exp(it) (Series of Complex Numbers: Partial Sums, Convergence, the Sum and Absolute Convergence (ii)). Taking real and imaginary parts (Limits of Complex Sequences: Uniqueness, Real and Imaginary Parts, Arithmetic, Real Sequences, Conjugates, Moduli, the Cauchy Criterion and the Modulus Metric (ii)),

Re⁑snβ†’Re⁑exp⁑(it)=cos⁑t,Im⁑snβ†’Im⁑exp⁑(it)=sin⁑t.\operatorname{Re}s_{n}\to\operatorname{Re}\exp(it)=\cos t,\qquad \operatorname{Im}s_{n}\to\operatorname{Im}\exp(it)=\sin t .

The map m↦2mm\mapsto2m from N\mathbb{N} to N\mathbb{N} is strictly increasing, so by (5) the sequences (Cm)(C_{m}) and (Sm)(S_{m}) are subsequences of (Re⁑sn)(\operatorname{Re}s_{n}) and (Im⁑sn)(\operatorname{Im}s_{n}). Every subsequence of a convergent sequence converges to the same limit (Limits of Sequences of Real Numbers: Uniqueness, Boundedness, Constants, Tails, Arithmetic, Quotients, Absolute Values, Finite Sums, Order, Squeezing, Domination and Subsequences (xiii)), hence Cmβ†’cos⁑tC_{m}\to\cos t and Smβ†’sin⁑tS_{m}\to\sin t. That is, the two series of (iii) converge, with sums cos⁑t\cos t and sin⁑t\sin t (Series of Real Numbers: Partial Sums, Convergence, the Sum and Absolute Convergence (ii)).

It remains to show absolute convergence. By (3), (1) and The Modulus of a Complex Number (ii), ∣ck∣=∣a2kβˆ’1∣|c_{k}|=|a_{2k-1}| and βˆ£Οƒk∣=∣iΟƒk∣=∣a2k∣|\sigma_{k}|=|i\sigma_{k}|=|a_{2k}|. The series βˆ‘n=1∞∣an∣\sum_{n=1}^{\infty}|a_{n}| converges (The Complex Exponential Series Converges Absolutely, and Its Sum Is a Function on β„‚ (ii), Series of Complex Numbers: Partial Sums, Convergence, the Sum and Absolute Convergence (iii)); call its sum AA. Its terms are nonnegative, so each of its partial sums is at most AA (Series of Real Numbers with Nonnegative Terms, Comparison, Absolute Convergence and the Geometric Series (i)). Hence, for every m∈Nm\in\mathbb{N},

βˆ‘k=1m∣ck∣+βˆ‘k=1mβˆ£Οƒk∣=βˆ‘k=1m(∣a2kβˆ’1∣+∣a2k∣)=βˆ‘n=12m∣anβˆ£β‰€A.\sum_{k=1}^{m}|c_{k}|+\sum_{k=1}^{m}|\sigma_{k}|=\sum_{k=1}^{m}\bigl(|a_{2k-1}|+|a_{2k}|\bigr)=\sum_{n=1}^{2m}|a_{n}|\le A .

So the partial sums of βˆ‘k=1∞∣ck∣\sum_{k=1}^{\infty}|c_{k}| and of βˆ‘k=1βˆžβˆ£Οƒk∣\sum_{k=1}^{\infty}|\sigma_{k}| are bounded above by AA, and since these series have nonnegative terms they converge (Series of Real Numbers with Nonnegative Terms, Comparison, Absolute Convergence and the Geometric Series (i)). That is, both series of (iii) converge absolutely (Series of Real Numbers: Partial Sums, Convergence, the Sum and Absolute Convergence (iii)), which completes the proof.

Citations

Loading…

Dependencies

Uses0

Loading…

Comments

Log in to comment.

Loading…