Analytic Number Theory¶

Job: extract arithmetic information (primes, $L$-values, sieve counts) from complex-analytic properties of Dirichlet series and $L$-functions.

Core¶

Analytic number theory trades arithmetic sums for contour integrals of meromorphic generating functions. The classical spine is $\zeta(s)$ (and Dirichlet $L$-functions): Euler product $\leftrightarrow$ primes; poles/zeros $\leftrightarrow$ main terms and error terms via Perron and explicit formulae.

Page	Focus
Foundations	Dirichlet series, $\zeta$, Euler products, PNT + zero-free region sketch, $L$-functions, functional equation pointer
L-functions	Dirichlet/Hecke $L$, analytic continuation, GRH, explicit formulae

Theorem spine¶

Euler product for $\zeta$ (absolutely convergent $\operatorname{Re}s>1$) $\Rightarrow$ infinitude of primes via pole at $s=1$.
Analytic continuation + functional equation $\Rightarrow$ global control of $\zeta'/\zeta$.
Zero-free region near $\operatorname{Re}s=1$ $\Rightarrow$ prime number theorem $\psi(x)\sim x$.
Explicit formula: primes $\leftrightarrow$ nontrivial zeros of $\zeta$ (or of $L(s,\chi)$).

Interfaces¶

Number theory: elementary primes, congruences, multiplicative functions
Complex analysis: residues, contour shift, Phragmén–Lindelöf
Algebraic number theory: Dedekind $\zeta_K$, Hecke $L$, Artin $L$
Fourier analysis / Poisson summation: functional equations of $\theta$-series type

See it¶

Chebyshev function via Perron. For $c>1$ and $x\notin\mathbb{Z}$, $$ \psi(x)=\frac{1}{2\pi i}\int_{c-i\infty}^{c+i\infty}-\frac{\zeta'}{\zeta}(s)\frac{x^s}{s}\,ds. $$ Shift left, pick residue $x$ at $s=1$ and residues $-x^\rho/\rho$ at nontrivial zeros $\rho$: error in PNT is governed by zero-free width.

Check¶

Why does $\zeta(s)\neq 0$ on $\operatorname{Re}s=1$ imply $\psi(x)\sim x$?

Nonvanishing on the line + growth estimates let one push the Perron contour to $\operatorname{Re}s=1-\delta$ (or use Ikehara/Wiener–Ikehara tauberian theorems). The pole of $-\zeta'/\zeta$ at $s=1$ supplies the main term $x$; absence of other poles on $\operatorname{Re}s=1$ kills competing $x^{1+it}$ oscillations of full strength.

What arithmetic fails if one only knows $\zeta$ meromorphic of finite order, without an Euler product?

Location of zeros need not encode primes; explicit formulae and positivity (e.g. $\zeta(s)>0$ for $s>1$ real) use the Euler product / Dirichlet series with non-negative coefficients.

Pitfalls¶

Conditional vs unconditional: RH-quality zero-free region is still open; classical zero-free regions are narrow near $s=1$.
Dirichlet density $\neq$ natural density without extra tauberian/input; PNT needs genuine analytic work.
Trivial zeros of $\zeta$ (negative even integers) are from the $\Gamma$-factor, not from primes.

Refs¶

Davenport, Multiplicative Number Theory
Iwaniec–Kowalski, Analytic Number Theory
Titchmarsh, The Theory of the Riemann Zeta-Function
Montgomery–Vaughan, Multiplicative Number Theory I

Next¶

Foundations · L-functions · Number theory · Complex analysis · Algebraic number theory