Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Given the logarithmic derivative of the zeta function $\dfrac{\zeta^\prime (s)}{\zeta(s)}$ how does it behave near $s=1$?

I mean if for $s=1$ the Laurent series for the logarithmic derivative becomes

$$\frac{\zeta^\prime (s)}{\zeta(s)}= A+ (s-1)^{-1}$$

where $A$ is a real number constant.

share|improve this question
3  
The $A$ you have there is the Euler-Mascheroni constant, $\gamma$. Actually, your series goes like $\frac{\zeta^\prime (s)}{\zeta(s)}=-\frac1{s-1}+\gamma-(\gamma^2+2\gamma_1)(s-1)+\cdots$ where $\gamma_1$ is the first Stieltjes constant. –  J. M. Oct 9 '11 at 12:52

2 Answers 2

Let's generalize and look at all of the constants in the Laurent expansion. Consider $$-\frac{\zeta^\prime}{\zeta}(s+1)=\frac{1}{s}+\sum_{j=0}^\infty \frac{(-1)^j}{j!}\eta_j s^j.$$ (I put in a negative sign so it corresponds to $\sum\limits_{n=1}^\infty \Lambda(n)n^{-s}$.) I claim that $\eta_0 =-\gamma_0$, Euler's Constant, and that in general $\eta_j$ will be a sum of products of the Stieltjes constants.

Constants arising from a limit: It is interesting to note that in a way analogous to the fact that the Stieltjes constants are given by the limit $$\gamma_n=\lim_{n\to\infty}{\left(\left(\sum_{k = 1}^m \frac{(\log k)^n}{k}\right) - \frac{(\log m)^{n+1}}{n+1}\right)},$$ we have that $$\eta_n=\lim_{n\to\infty}{\left(\left(\sum_{k = 1}^m \frac{(\log k)^{n}\Lambda(k)}{k}\right) - \frac{(\log m)^{n+1}}{n+1}\right)}$$ where $\Lambda(n)$ is the von Mangoldt Lambda Function. This can be proven by using exponential generating series, and then the Dirichlet series for $-\frac{\zeta^\prime}{\zeta}(s).$

Computing the Constants: We can compute $\eta_j$ in terms of the Stieltjes constants by using the relation $$\log\left(1+\sum_{i=0}^\infty \frac{(-1)^j}{j!}\gamma_j s^{j+1}\right)= \log (s\zeta(s+1))=-\sum_{j=0}^\infty \frac{(-1)^j}{j!}\eta_j s^j$$

and then expanding by using the identity $$\log(1+x)=x-\frac{x^2}{2}+\frac{x^3}{3}-\frac{x^4}{4}+\cdots$$

From this we get the power series $$\gamma_0 s-\left(2\gamma_1+\gamma_0^2\right)\frac{s^2}{2}+\left(3\gamma_2+6\gamma_1\gamma_0+2\gamma_0^3\right)\frac{s^3}{6}+\cdots$$ so that $$\eta_0 =-\gamma, \ \ \ \eta_1 =-(2\gamma_1 +\gamma_0^2), \ \ \ \eta_2 =-(3\gamma_2 +6\gamma_1\gamma_0 +2\gamma_0^3) \qquad \text{et cetera}.$$

share|improve this answer
    
THANK YOU very much to the people who answered :) –  Jose Garcia Oct 9 '11 at 13:15

As an addendum to Eric's answer, Coffey gives a recursion relation for the coefficients $\dfrac{\eta_j}{j!}$ appearing in the expansion of the $\zeta$ function's logarithmic derivative. Letting $\gamma_k$ be the Stieltjes constants, and with the initial condition $\eta_0=-\gamma_0=-\gamma$, we have the recursion relation

$$\frac{\eta_j}{j!}=(-1)^{j+1}\left(\frac{j+1}{j!}\gamma_j+\sum_{k=0}^{j-1} \frac{(-1)^{k-1}}{(j-k-1)!}\frac{\eta_k}{k!}\gamma_{j-k-1}\right)$$

A similar formula also appears in Choudhury's paper.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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