Logarithmic Integral — Definition, Formula & Examples
The logarithmic integral, denoted , is a special function defined as the integral of from 0 to . It arises naturally in calculus and is famous for approximating the number of primes up to a given value.
For real , , the logarithmic integral is defined as , where the singularity at is handled as a Cauchy principal value. A common variant, the offset logarithmic integral, is , which avoids the singularity entirely.
Key Formula
Where:
- = A positive real number (with x ≠ 1 for li)
- = Variable of integration
- = Natural logarithm of t
How It Works
The integrand has no elementary antiderivative, so cannot be expressed in terms of standard functions. You evaluate it numerically or via series expansions. The offset form is preferred in number theory because it starts counting from 2, matching the smallest prime. The prime number theorem states that as , where counts the primes up to .
Worked Example
Problem: Estimate the offset logarithmic integral Li(10) and compare it to the actual count of primes up to 10.
Set up the integral: Write the offset logarithmic integral with upper limit 10.
Evaluate numerically: Since there is no closed-form antiderivative, use numerical integration (e.g., Simpson's rule or a calculator). The result is approximately:
Compare to the prime count: The primes up to 10 are 2, 3, 5, and 7, so π(10) = 4. The approximation Li(10) ≈ 5.12 is reasonably close, and the accuracy improves dramatically for larger x.
Answer: , which approximates . The estimate becomes increasingly accurate for large .
Why It Matters
The logarithmic integral is central to analytic number theory: the prime number theorem asserts that is the best simple approximation to the prime counting function . It also appears in physics and engineering contexts involving exponential-integral-type functions.
Common Mistakes
Mistake: Confusing li(x) with Li(x).
Correction: The notation li(x) integrates from 0, while Li(x) integrates from 2. Most number theory results use Li(x), the offset version, which differs from li(x) by the constant li(2) ≈ 1.045.
