Understanding the Integral of x²eˣ²: A Deep Dive into Non-Elementary Functions
The integral of $x^2 e^{x^2}$ represents a fascinating crossroads in calculus where standard integration techniques reach their limit, forcing mathematicians to define entirely new classes of functions. Unlike the integral of $x e^{x^2}$, which yields neatly to a simple $u$-substitution, the presence of the $x^2$ term outside the exponential prevents an elementary antiderivative. This article explores why this integral resists standard methods, how to express its solution using the imaginary error function (erfi), and the series expansions that allow for practical computation The details matter here..
Why Standard Techniques Fail
When students first encounter $\int x^2 e^{x^2} , dx$, the instinct is to reach for the standard toolkit: $u$-substitution or integration by parts.
Attempting U-Substitution If we set $u = x^2$, then $du = 2x , dx$, implying $dx = \frac{du}{2x} = \frac{du}{2\sqrt{u}}$. Substituting this into the integral transforms it into: $ \int u e^u \frac{du}{2\sqrt{u}} = \frac{1}{2} \int \sqrt{u} , e^u , du $ This new integral, $\int u^{1/2} e^u , du$, is arguably more complicated than the original. It involves a fractional power inside an exponential integral, which has no elementary solution It's one of those things that adds up. And it works..
Attempting Integration by Parts The LIATE rule (Logarithmic, Inverse trig, Algebraic, Trigonometric, Exponential) suggests setting $u = x$ (Algebraic) and $dv = x e^{x^2} dx$ (Exponential).
- $u = x \implies du = dx$
- $dv = x e^{x^2} dx \implies v = \frac{1}{2}e^{x^2}$
Applying the formula $\int u , dv = uv - \int v , du$: $ \int x^2 e^{x^2} dx = x \left( \frac{1}{2}e^{x^2} \right) - \int \frac{1}{2}e^{x^2} dx $ $ = \frac{1}{2}x e^{x^2} - \frac{1}{2} \int e^{x^2} dx $
This reduces the problem to finding $\int e^{x^2} dx$. It cannot be expressed using a finite combination of polynomials, exponentials, logarithms, or trigonometric functions. This is the famous Gaussian integral (indefinite form), which is the canonical example of a non-elementary integral. This dead end confirms that $\int x^2 e^{x^2} dx$ is also non-elementary.
The Solution: The Imaginary Error Function (erfi)
Since the integral cannot be written with elementary functions, mathematicians defined a special function to handle it: the Imaginary Error Function, denoted as $\text{erfi}(x)$.
Definition and Relationship to erf
The standard Error Function is defined as: $ \text{erf}(x) = \frac{2}{\sqrt{\pi}} \int_0^x e^{-t^2} dt $ It is used extensively in probability, statistics, and partial differential equations (heat equation).
The Imaginary Error Function handles the positive exponent $e^{+x^2}$. It is defined by the relationship: $ \text{erfi}(x) = -i , \text{erf}(ix) = \frac{2}{\sqrt{\pi}} \int_0^x e^{t^2} dt $ Where $i = \sqrt{-1}$. Despite the name "imaginary," for real inputs $x$, $\text{erfi}(x)$ returns purely real values Worth knowing..
This changes depending on context. Keep that in mind.
The Final Closed Form
Using the definition of $\text{erfi}(x)$, we can solve the remaining integral from our integration by parts step: $ \int e^{x^2} dx = \frac{\sqrt{\pi}}{2} \text{erfi}(x) + C $
Substituting this back into our integration by parts result: $ \int x^2 e^{x^2} dx = \frac{1}{2}x e^{x^2} - \frac{1}{2} \left( \frac{\sqrt{\pi}}{2} \text{erfi}(x) \right) + C $ $ \boxed{\int x^2 e^{x^2} dx = \frac{1}{2}x e^{x^2} - \frac{\sqrt{\pi}}{4} \text{erfi}(x) + C} $
This is the exact, closed-form solution. In modern computational environments like Mathematica, Maple, Python (SciPy), and MATLAB, erfi is a built-in standard function, making this expression just as "usable" as $\sin(x)$ or $\ln(x)$.
Alternative Representation: The Dawson Function
In some physics and engineering contexts—particularly plasma physics and optics—the Dawson Function $F(x)$ is preferred over $\text{erfi}(x)$ because it avoids the rapid exponential growth of $e^{x^2}$ for large arguments, offering better numerical stability.
The Dawson Function is defined as: $ F(x) = e^{-x^2} \int_0^x e^{t^2} dt = \frac{\sqrt{\pi}}{2} e^{-x^2} \text{erfi}(x) $
Rearranging for $\text{erfi}(x)$: $ \text{erfi}(x) = \frac{2}{\sqrt{\pi}} e^{x^2} F(x) $
Substituting this into our integral result: $ \int x^2 e^{x^2} dx = \frac{1}{2}x e^{x^2} - \frac{\sqrt{\pi}}{4} \left( \frac{2}{\sqrt{\pi}} e^{x^2} F(x) \right) + C $ $ = \frac{1}{2}e^{x^2} \left( x - F(x) \right) + C $
This form, $\frac{1}{2}e^{x^2} (x - F(x))$, is often the preferred output in specialized scientific libraries.
Series Expansion: Computing by Hand
If you do not have access to special function libraries, or if you need to understand the behavior of the function near $x=0$, the Maclaurin Series (Taylor Series at 0) is the standard approach.
Recall the series for $e^u$: $ e^u = \sum_{n=0}^{\infty} \frac
