Section 24.3 From Riemann to Dirichlet and Euler
ΒΆ
βdβ£nΞΌ(d)d=βpβ£n(1β1p)
xxxxxxxxxx
def _(e=(1,[0..3]),f=(2,[0..3]),g=(0,[0..3])):
n = 2^e*3^f*5^g
pretty_print(html("You picked $%s=2^{%s}3^{%s}5^{%s}$"%(n,e,f,g)))
str = '$$'+'+'.join([r'\frac{%s}{%s}'%(moebius(d),d) for d in divisors(n)])+'=%s$$'%sum([moebius(d)/d for d in divisors(n)])
str2 = '$$'+''.join([r'\left(1-\frac{1}{%s}\right)'%p for (p,e) in factor(n)])+'=%s$$'%prod([1-1/p for (p,e) in factor(n)])
pretty_print(html(str))
pretty_print(html("compare to "+str2))
ββn=1ΞΌ(n)ns.
Subsection 24.3.1 Dirichlet series
ΒΆWe give such series a name. The following definition is purely formal, considered without considering issues such as convergence. (See [C.2.8, Chapter 4.6] for an interesting formal viewpoint on the set of these series.)Definition 24.3.1.
In general, for an arithmetic function f(n), its Dirichlet series is
F(s)=ββn=1f(n)ns.
For what arithmetic function is the Riemann zeta function the Dirichlet series?
What would the Dirichlet series of N be?
What about the Dirichlet series of I?
Subsection 24.3.2 Euler products
ΒΆFor our purposes, the very important thing to note about such series is that they often can be expanded as infinite products.Definition 24.3.2.
In general, for an arithmetic function f(n), its Dirichlet series is said to have an Euler product if the series can be written as an infinite product in the following manner:
ββn=1f(n)ns=βp( a formula involving f(p) and ps).
Example 24.3.3. Euler product for Riemann zeta function.
We have already suggested one for the zeta function:
ΞΆ(s)=ββn=11ns=βp(11βpβs).
ββn=1ΞΌ(n)ns=βp(1β1ps)=βp(1βpβs)
At least, we can consider this wherever it makes sense. See [C.4.6, Chapter 11.5] or [C.2.1, Chapter 9.8] for some criteria, or simply below at Theorem 24.5.4.
In the next section, we justify more of this discussion, and connect our wonderful results about Dirichlet products of finite arithmetic functions to deep properties of their Dirichlet series.