Deriving The Secant-Power Reduction Formula

Why Reduce a Secant Power?

An integral containing a large power of secant can be difficult to handle directly. A useful strategy is to relate it to an integral with a smaller power. Repeating that reduction eventually reaches a power whose antiderivative is already known.

For an integer $n\geq 2$, the secant-power reduction formula expresses an antiderivative of $\sec^n x$ using a boundary term and an integral of $\sec^{n-2}x$. The drop in exponent is two, and the coefficient of the reduced integral is essential: it is $(n-2)/(n-1)$.

Key idea: Use $\sec^n x=\sec^{n-2}x\,\sec^2 x$, then integrate by parts with $\sec^2 x\,dx=d(\tan x)$. The derivative of the remaining factor creates the lower-power integral.

The Reduction Formula and Its Conditions

Theorem: Let $n$ be an integer with $n\geq2$, and let $I$ be an interval on which $\cos x\neq0$. Then, on $I$, $$ \int \sec^n x\,dx = \frac{\sec^{n-2}x\tan x}{n-1} + \frac{n-2}{n-1}\int \sec^{n-2}x\,dx. $$ The equality is understood up to an arbitrary additive constant. The interval condition ensures that $\sec x$ and $\tan x$ are defined throughout $I$.

The exponent must be an integer at least two for this recurrence as stated. The condition on the interval matters because secant and tangent have singularities where $\cos x=0$. Antiderivatives are considered on intervals that do not cross those points; a formula valid on one such interval need not define an antiderivative across a singularity.

The recurrence needs two base cases. At exponent zero, the integrand is $1$. At exponent one, a standard antiderivative is expressed using a logarithm:

\int \sec^0 x\,dx=x+C, \qquad \int \sec x\,dx=\ln|\sec x+\tan x|+C.

For the second identity, $\sec x+\tan x$ is nonzero wherever it is defined: $(\sec x+\tan x)(\sec x-\tan x)=1$. Thus the logarithm is defined with an absolute value, and differentiation gives

\frac{d}{dx}\ln|\sec x+\tan x| = \frac{\sec x\tan x+\sec^2 x}{\sec x+\tan x} = \sec x.

For exponent two, the recurrence itself reduces directly to $\int\sec^2 x\,dx=\tan x+C$. These base cases show how repeated reductions can terminate: even powers eventually reach exponent zero, while odd powers eventually reach exponent one. The focus here is the derivation and structure of the recurrence.

Deriving the Formula

Write $I_n=\int\sec^n x\,dx$. Since $\sec^n x=\sec^{n-2}x\sec^2 x$, choose $u=\sec^{n-2}x$ and $dv=\sec^2 x\,dx$. Then $v=\tan x$, while

du=(n-2)\sec^{n-2}x\tan x\,dx.

Integration by parts, $\int u\,dv=uv-\int v\,du$, gives

\begin{aligned} I_n &=\sec^{n-2}x\tan x -(n-2)\int \sec^{n-2}x\tan^2 x\,dx. \end{aligned}

Now use the identity $\tan^2 x=\sec^2 x-1$. The resulting integral contains $I_n$ again, which is the feature that makes it possible to solve for $I_n$.

Separate two powers.
Write $\sec^n x=\sec^{n-2}x\sec^2 x$, so $\sec^2 x\,dx$ has the antiderivative $\tan x$.

Integrate by parts.
The derivative of $\sec^{n-2}x$ introduces $(n-2)\sec^{n-2}x\tan x$, leaving an integral involving $\tan^2 x$.

Replace the tangent square.
Substitute $\tan^2 x=\sec^2 x-1$, so the integral separates into $I_n$ and $I_{n-2}$.

Collect the $I_n$ terms.
Move the resulting multiple of $I_n$ to the left and divide by $n-1$.

In detail, the substitution gives

\begin{aligned} I_n &=\sec^{n-2}x\tan x -(n-2)\int \sec^{n-2}x(\sec^2 x-1)\,dx\\ &=\sec^{n-2}x\tan x-(n-2)I_n+(n-2)I_{n-2}. \end{aligned}

Add $(n-2)I_n$ to both sides. This produces $(n-1)I_n$ on the left, so division by $n-1$, which is nonzero because $n\geq2$, yields

I_n = \frac{\sec^{n-2}x\tan x}{n-1} + \frac{n-2}{n-1}I_{n-2}.

This derivation explains both coefficients. The boundary term is divided by $n-1$, and the reduced integral has coefficient $(n-2)/(n-1)$. The latter is not an extra factor from the trigonometric identity; it results from collecting $I_n$ on one side.

Worked Examples

Worked Example: Reduce the Second Power

Apply the formula with $n=2$. Its boundary term has $\sec^{0}x=1$, and the coefficient on the reduced integral is zero:

\begin{aligned} \int\sec^2 x\,dx &=\frac{\sec^0 x\tan x}{2-1} +\frac{2-2}{2-1}\int\sec^0 x\,dx\\ &=\tan x+0\\ &=\tan x+C. \end{aligned}

Check by differentiation: $\frac{d}{dx}(\tan x+C)=\sec^2 x$, as required. The exponent-zero term is harmless because its coefficient is zero.

Worked Example: Reduce the Fourth Power

For $n=4$, the recurrence gives

\begin{aligned} \int\sec^4 x\,dx &=\frac{\sec^2 x\tan x}{3} +\frac{2}{3}\int\sec^2 x\,dx\\ &=\frac{1}{3}\sec^2 x\tan x+\frac{2}{3}\tan x+C. \end{aligned}

To check, differentiate the proposed antiderivative. By the product rule,

\begin{aligned} \frac{d}{dx}\left(\frac13\sec^2 x\tan x+\frac23\tan x\right) &=\frac13(2\sec^2 x\tan^2 x+\sec^4 x)+\frac23\sec^2 x\\ &=\frac23\sec^2 x(\sec^2 x-1)+\frac13\sec^4 x+\frac23\sec^2 x\\ &=\frac23\sec^4 x-\frac23\sec^2 x+\frac13\sec^4 x+\frac23\sec^2 x\\ &=\sec^4 x. \end{aligned}

The second line uses $\tan^2 x=\sec^2 x-1$; the terms in $\sec^2 x$ cancel.

Worked Example: Reduce the Sixth Power

For $n=6$, first reduce to the fourth power, then substitute the antiderivative just obtained:

\begin{aligned} \int\sec^6 x\,dx &=\frac15\sec^4 x\tan x+\frac45\int\sec^4 x\,dx\\ &=\frac15\sec^4 x\tan x +\frac45\left(\frac13\sec^2 x\tan x+\frac23\tan x\right)+C\\ &=\frac15\sec^4 x\tan x+\frac4{15}\sec^2 x\tan x+\frac8{15}\tan x+C. \end{aligned}

Check by differentiating each term and replacing $\tan^2 x$ with $\sec^2 x-1$:

\begin{aligned} \frac{d}{dx}\left(\frac15\sec^4 x\tan x+\frac4{15}\sec^2 x\tan x+\frac8{15}\tan x\right) &=\frac15(4\sec^4 x\tan^2 x+\sec^6 x)\\ &\quad+\frac4{15}(2\sec^2 x\tan^2 x+\sec^4 x)+\frac8{15}\sec^2 x\\ &=\sec^6 x-\frac45\sec^4 x+\frac45\sec^4 x-\frac8{15}\sec^2 x+\frac8{15}\sec^2 x\\ &=\sec^6 x. \end{aligned}

Thus the differentiated expression equals the original integrand.

A Common Coefficient Error

A frequent mistake is to collect the $I_n$ term with the wrong sign. For example, after integration by parts one might incorrectly write

I_n=\sec^{n-2}x\tan x+(n-2)I_n-(n-2)I_{n-2},

then arrive at the incorrect coefficient $1/(3-n)$ for the boundary term. The error is the sign in front of the integral produced by integration by parts: it must be $-\int v\,du$. The correct line is

I_n=\sec^{n-2}x\tan x-(n-2)I_n+(n-2)I_{n-2}.

A quick check at $n=2$ is useful. The correct recurrence gives $\int\sec^2 x\,dx=\tan x+C$. A proposed formula that fails this basic case cannot be correct.

When the Recurrence Is Useful

Each application replaces an exponent $n$ with $n-2$. Consequently, the exponent decreases in steps of two, and the base case reached depends on whether the starting integer is even or odd. This predictable decrease is why the recurrence is an effective method for systematically reducing secant powers.

The formula is an identity between antiderivative expressions on any interval where $\cos x\neq0$. It does not remove the singularities of the original integrand, and it does not justify joining antiderivatives across a point where secant is undefined. When applying it, keep the interval in view, use the correct exponent in both coefficients, and retain an arbitrary constant in the final antiderivative.

Takeaway: For integer $n\geq2$, integrate $\sec^n x$ by parts after splitting off $\sec^2 x$. The resulting recurrence is $$ \int \sec^n x\,dx = \frac{\sec^{n-2}x\tan x}{n-1} + \frac{n-2}{n-1}\int \sec^{n-2}x\,dx, $$ valid on intervals where $\cos x\neq0$.

Check Your Understanding

Use the derivation and conditions above to answer these questions.

Which factor is chosen as $dv$ when deriving the recurrence, and why is it convenient?
Where does the coefficient $(n-2)/(n-1)$ come from?
What is the result of applying the recurrence at $n=2$? Verify it by differentiation.
State the interval condition required for the formula to apply.
Identify the sign error in the expression $I_n=\sec^{n-2}x\tan x+(n-2)I_n-(n-2)I_{n-2}$.

Deriving The Secant-Power Reduction Formula

What You'll Learn

Why Reduce a Secant Power?

The Reduction Formula and Its Conditions

Deriving the Formula