Reason: Initial published proof of the basic properties of the exponential function; approved by Aaron.
Proof
Claim 1.exp(0)=1 directly from the series of The Real Exponential Function with the convention 00=1. For multiplicativity, form the Cauchy product of the two absolutely convergent series: the family ujvk/(j!k!) over pairs (j,k) is absolutely summable (its absolute sums are bounded by the product of the two absolute-value series), so it may be summed by diagonals β for series of nonnegative terms every rearrangement has the same supremum of finite partial sums, and the general case follows by splitting into positive and negative parts. Grouping by j+k=l and using the binomial expansion (u+v)l=βjβ€lβ(jlβ)ujvlβj with (jlβ)=l!/(j!(lβj)!) (Factorial of a Natural Number) gives exp(u)exp(v)=βlβ(u+v)l/l!=exp(u+v).
Claim 2. By claim 1, exp(u)exp(βu)=exp(0)=1, so exp(u)ξ =0 and exp(βu)=1/exp(u). For uβ₯0 the series has nonnegative terms and constant term 1, so exp(u)β₯1>0; for u<0, exp(u)=1/exp(βu)>0.
so (exp(s)β1)/sβ1 as sβ0. By claim 1, (exp(u+s)βexp(u))/s=exp(u)(exp(s)β1)/sβexp(u), so exp has derivativeexpβ²=exp at every point. Iterating, all higher derivatives exist and equal exp, which is continuous (differentiability at a point implies continuity there, directly from the definition of the derivative and limit arithmetic); hence exp is a smooth map on R=R1.
Claim 4.exp(u)β₯1+u for uβ₯0: all omitted series terms are nonnegative. Strict monotonicity: for u<v, claims 1 and 2 give exp(v)βexp(u)=exp(u)(exp(vβu)β1)β₯exp(u)(vβu)>0. Decay at ββ: given Ξ΅>0 choose M>0 with 1/(1+M)<Ξ΅; for u<βM, exp(βu)β₯1+(βu)>1+M, so exp(u)=1/exp(βu)<Ξ΅.
Claim 5. Injectivity is strict monotonicity (claim 4). For surjectivity onto (0,β), let tβ₯1 (the case 0<t<1 follows by applying the result to 1/t and using claim 2). The set S={uβ[0,t]:exp(u)β€t} contains 0 and is bounded above by t, so it has a least upper bound s by Least Upper Bound Property of the Real Numbers. Since exp is continuous at s (claim 3): if exp(s)<t, then s<t (as exp(t)β₯1+t>t) and exp<t on a neighborhood of s, contradicting that s is an upper bound of S; if exp(s)>t, then exp>t on a neighborhood of s, contradicting that s is the least upper bound (points of S would be excluded from that neighborhood, giving a smaller upper bound). Hence exp(s)=t, and exp is a bijection from R onto (0,β) (the image is contained in (0,β) by claim 2). β