Principal value and integral of 1/z

  • Thread starter Thread starter wasia
  • Start date Start date
  • Tags Tags
    Integral Value
Click For Summary
SUMMARY

The discussion centers on the principal value of the integral of 1/z, specifically the expression P ∫ 1/z = iπ, derived using the Cauchy integral theorem. The user attempts to express this principal value explicitly through the limit process involving integrals from -R to -ε and from ε to R. However, the confusion arises from the path dependence of the logarithm function ln(z), which complicates the evaluation of the integral and leads to an incorrect conclusion of 0. The key takeaway is the necessity of a consistent definition of ln(z) to properly evaluate the integral.

PREREQUISITES
  • Understanding of complex analysis, particularly contour integration.
  • Familiarity with the Cauchy integral theorem and its applications.
  • Knowledge of the logarithmic function in the complex plane.
  • Experience with limits and improper integrals.
NEXT STEPS
  • Study the properties of complex logarithms and their branch cuts.
  • Explore advanced topics in contour integration, focusing on path dependence.
  • Learn about the residue theorem and its applications in evaluating integrals.
  • Investigate the implications of principal value integrals in complex analysis.
USEFUL FOR

Students and professionals in mathematics, particularly those specializing in complex analysis, as well as physicists and engineers dealing with integrals in the complex domain.

wasia
Messages
50
Reaction score
0
I am slightly confused about the definition of principle value. If we have an integral
[tex]\int 1/z,[/tex]
where the integration from [tex]-\infty[/tex] to [tex]\infty[/tex] is implied, then by Cauchy integral theorem we know that the principle value
[tex]P \int 1/z=i\pi.[/tex]

However, I would like to write down this principle value explicitly. My best shot is
[tex]\lim_{\epsilon\rightarrow0}\lim_{R\rightarrow\infty}\int_{-R}^{-\epsilon}1/z+\int_{\epsilon}^{R}1/z.[/tex]

Assuming that this is correct (is it?) I can (can I?) calculate the integrals first and take limits afterwards. I get

[tex]\lim_{\epsilon\rightarrow0}\lim_{R\rightarrow\infty} \ln\left(-\frac{\epsilon}{\epsilon}\right) + \ln\left(-\frac{R}{R}\right)=2\ln(-1)=0.[/tex]

Can you tell me what am I doing wrong?
 
Physics news on Phys.org
You need to think about how to define the function ln(z). There is no unique definition, precisely because of the path dependence of the integral of 1/z.
 

Similar threads

Replies
6
Views
3K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 6 ·
Replies
6
Views
2K
Replies
17
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 16 ·
Replies
16
Views
3K
Replies
5
Views
2K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 9 ·
Replies
9
Views
2K