The inverse tangent has one of the neatest derivatives in calculus, and that single result unlocks its integral, its power series and a famous formula for π.
The derivative
Result: d/dx arctan x = 1 / (1 + x²)
Proof. Let y = arctan x, so tan y = x. Differentiate both sides with respect to x:
- sec² y · dy/dx = 1
- dy/dx = 1 / sec² y
- Since sec² y = 1 + tan² y = 1 + x², we get dy/dx = 1 / (1 + x²).
With the chain rule: d/dx arctan(u) = u′ / (1 + u²). For example d/dx arctan(3x) = 3 / (1 + 9x²).
The integral of 1/(1 + x²)
Reading the derivative backwards gives one of the most used integrals:
∫ 1 / (1 + x²) dx = arctan x + C
More generally, ∫ 1 / (a² + x²) dx = (1/a) · arctan(x / a) + C.
The integral of arctan x
Integrate by parts with u = arctan x and dv = dx:
- du = dx / (1 + x²), v = x
- ∫ arctan x dx = x · arctan x − ∫ x / (1 + x²) dx
- The remaining integral is ½ ln(1 + x²).
∫ arctan x dx = x · arctan x − ½ ln(1 + x²) + C
The Taylor series
For |x| < 1, 1 / (1 + x²) is a geometric series: 1 − x² + x⁴ − x⁶ + … Integrating term by term:
arctan x = x − x³/3 + x⁵/5 − x⁷/7 + …
The series also converges at x = ±1, although slowly.
The Leibniz formula for π
Put x = 1 into the series. Because arctan 1 = π/4:
π/4 = 1 − 1/3 + 1/5 − 1/7 + 1/9 − …
It is beautiful but converges very slowly. Faster formulas, such as Machin’s π/4 = 4 arctan(1/5) − arctan(1/239), were used to compute π by hand to over 100 digits.
Quick reference
| Expression | Result |
|---|---|
| d/dx arctan x | 1 / (1 + x²) |
| ∫ 1/(1 + x²) dx | arctan x + C |
| ∫ arctan x dx | x arctan x − ½ ln(1 + x²) + C |
| lim x→∞ arctan x | π/2 |
Try it yourself: open the inverse tangent calculator to check any value from this article with full step-by-step working.