The derivative $\dfrac{d}{dx}\left(e^x\right)=e^x$ can be proven a number of ways. I typically like to teach my students to prove the derivative using logarithmic differentiation: \begin{align} y&=e^x\\ \log y&=x \\ \frac{d}{dx}\left(\log y\right) &= \frac{d}{dx} \left(x\right) \\ \frac{1}{y}\cdot y'&=1 \\ y'&=y\\ y'&=e^x \end{align}

Recently I have been thinking about how to prove the derivative using the first principles definition of the derivative: $$f'(x)=\lim_{h\to0}\frac{f(x+h)-f(x)}h$$ and in doing so, we need to be able to prove $$\lim_{h\to0}\frac{e^h-1}h=1.$$

I understand that one definition of $e$ is that it is the only number that satisfies the above limit of it is used as the base of the exponential function, and in my years of study I have chosen to simply accept it as a definition. However, I would like to be able to prove it upon request, using basic properties of limits and another definition of $e$: $$e=\lim_{n\to0}\left(1+n\right)^{\frac1n}.$$ Is my proof below acceptable? Is there anything about it I could improve?

Statement: $\lim\limits_{h\to0}\frac{b^h-1}{h}=1 \iff b=e$.

Proof. Let $u=b^h-1$. Then $h=\log_b(u+1)$, and $h\to0\implies u\to0$. We then have \begin{align} \lim_{h\to0}\frac{b^h-1}h&=\lim_{u\to0}\frac{u}{\log_b(u+1)}\\ &=\lim_{u\to0}\frac1{\frac1u\log_b(u+1)}\\ &=\lim_{u\to0}\frac1{\log_b(u+1)^{\frac1u}}\\ &=\frac1{\log_b\left(\lim\limits_{u\to0}(1+u)^{\frac1u}\right)}\\ \lim_{h\to0}\frac{b^h-1}h&=\frac1{\log_be} \end{align} Then, solving the equation, we have $$\frac1{\log_be}=1\iff b=e.$$

  • 1
    $\begingroup$ Related article on proofwiki: proofwiki.org/wiki/Derivative_of_Exponential_at_Zero $\endgroup$ – Maximilian Janisch Feb 12 '20 at 15:39
  • $\begingroup$ In $h\to0\implies u\to0$ you are hiding that $e^x$ is continuous at $x=0$ and that $e^0=1$ and a similar property for $\log_b$ when the limit switches with it. $\endgroup$ – user748968 Feb 12 '20 at 15:54
  • $\begingroup$ It is quite easy to prove that $$\lim_{x\to0}\dfrac{b^x-1}{x}=\lim_{x\to0}\dfrac{b^x\log b}{1}=\log b\space \text {where the logarithm is neperian }$$ and $$\log b=1\iff b=e^1$$ $\endgroup$ – Piquito Feb 12 '20 at 15:59
  • $\begingroup$ You need to justify the exchange of the limit and the logarithm. $\endgroup$ – Yves Daoust Feb 12 '20 at 16:31

Your first approach is simpler and the preferred one. It is based on the following assumption:

There exists a function $\log:(0, \infty) \to\mathbb{R} $ such that $\log 1=0$ and $$\dfrac{d} {dx} \log x=\dfrac{1}{x},\,\forall x \in(0,\infty) $$ Further the symbol $e^x$ is defined by $$y=e^ x\iff x=\log y$$

The above assumption is easily justified by using the definition $$\log x=\int_{1}^{x}\frac{dt}{t}$$

The second approach you have chosen is difficult. It involves defining the symbol $a^b, a>0,b\in\mathbb {R} $ without the use of logarithm. And then one analyzes the limit $$f(a) =\lim_{h\to 0}\frac{a^h-1}{h}$$ and shows that it exists for every $a>0$ and hence defines a function $f:(0, \infty) \to\mathbb {R} $.

Further one establishes that $f$ defined above is strictly increasing, continuous and the range of $f$ is $\mathbb {R} $ and $$f(1)=0,f(xy) =f(x) +f(y), f'(x) =1/x$$ Hence there is a unique number $e>1$ such that $f(e) =1$.

Once you have reached this point, it is easy to show that $$e=\lim_{n\to \infty} \left(1+\frac{1}{n}\right) ^n$$ We have $$f((1+(1/n))^n)=nf(1+(1/n))=\dfrac{f(1+(1/n))-f(1)}{1/n}\to f'(1)=1$$ as $n\to\infty $. Let $g$ be the inverse of $f$ so that $g$ is also continuous and $g(1)=e$. Clearly we have $$g(f((1+(1/n))^n))\to g(1)=e$$ or $$(1+(1/n))^n\to e$$ and we are done.


Hint: The derivative of $b^x$ is $\ln b\cdot b^x$.

Further hint: $(b^x)'(0)=\lim_{h\to0}\dfrac{b^h-1}h$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.