Why Rationalize a Difference Quotient?
A difference quotient can contain a radical difference such as \(\sqrt{x}-\sqrt{a}\). When \(x\) approaches \(a\), both this difference and the denominator \(x-a\) approach zero. Direct substitution then gives an indeterminate form: it does not tell us the limit. Multiplying by a carefully chosen expression, the conjugate, can expose a simpler quotient whose limit is easier to determine.
The same algebra can lead to different outcomes. At a point where the radical is positive, rationalization often leaves a denominator approaching a positive number and produces a finite limit. At a point where the radical is zero, the remaining denominator may also approach zero, and the quotient may grow without bound. The domain of the radical matters throughout.
The Conjugate and the Key Identity
The conjugate of \(\sqrt{u}-\sqrt{v}\) is \(\sqrt{u}+\sqrt{v}\). Their product is a difference of squares:
provided \(u\geq 0\) and \(v\geq 0\). This identity is useful because it replaces a difference of square roots with the difference of their radicands.
For a square-root difference quotient, multiply the numerator and denominator by the conjugate of the numerator. This does not change the quotient where the multiplier is nonzero. It can simplify the expression so that the limit is visible.
Deriving the Basic Simplification
Here is the complete algebra behind the identity. Since \(x\ne a\), the original denominator is nonzero. Also, because \(x\) and \(a\) are nonnegative and are not both zero, \(\sqrt{x}+\sqrt{a}>0\). Thus multiplying by the conjugate does not multiply by zero.
Multiply the numerator and denominator by \(\sqrt{x}+\sqrt{a}\), keeping the quotient unchanged.
The numerator becomes \((\sqrt{x}-\sqrt{a})(\sqrt{x}+\sqrt{a})=x-a\).
Because \(x\ne a\), cancel \(x-a\) from the numerator and denominator to obtain \(1/(\sqrt{x}+\sqrt{a})\).
The cancellation is valid only for \(x\ne a\), which is already required in the definition of the difference quotient. The simplified expression can then be used to take a limit as \(x\) approaches \(a\); the value at \(x=a\) need not be part of the original quotient.
When the Simplified Limit Is Finite
If \(a>0\), then as \(x\) approaches \(a\), the denominator \(\sqrt{x}+\sqrt{a}\) approaches \(2\sqrt{a}\), a positive number. Continuity of the square-root function on its domain therefore gives a finite limit. The identity proves the result directly:
The condition \(a>0\) matters: it ensures that the denominator of the simplified expression does not approach zero. The calculation is a limit of a rationalized difference quotient; the focus here is the algebraic simplification and its limit, not a general rule for differentiating square-root functions.
Worked Example: A Finite Limit at a Positive Radicand
Evaluate \(\displaystyle \lim_{x\to 9}\frac{\sqrt{x}-3}{x-9}\). The quotient is defined for \(x\geq 0\) with \(x\ne 9\) near the point of approach. Direct substitution gives \(0/0\), so rationalize.
The conjugate of \(\sqrt{x}-3\) is \(\sqrt{x}+3\). Multiply the numerator and denominator by this conjugate:
Now take the limit using the simplified expression:
Check the simplification by multiplying back: \((\sqrt{x}-3)(\sqrt{x}+3)=x-9\), so the factor canceled really is \(x-9\). The limit is \(\frac16\), and the simplified expression tends to the same value.
Radicands That Are More Than Just \(x\)
The conjugate method also works when the radicand is an expression in \(x\). If \(g(x)\geq 0\) and \(g(a)\geq 0\), then, wherever the original quotient is defined and the conjugate is nonzero,
This identity follows by multiplying by \(\sqrt{g(x)}+\sqrt{g(a)}\) and applying the difference-of-squares identity. It does not by itself guarantee a finite limit. One must inspect both factors in the new denominator and the behavior of \(g(x)-g(a)\). In particular, a positive limiting value of the sum of square roots makes the remaining limit easier to analyze; a sum approaching zero calls for further care.
Worked Example: A Shifted Linear Radicand
Evaluate \(\displaystyle \lim_{x\to 1}\frac{\sqrt{3x+1}-2}{x-1}\). For \(x\) near \(1\), \(3x+1\) is positive, so the square root is defined. At \(x=1\), the numerator and denominator both approach zero.
The conjugate of \(\sqrt{3x+1}-2\) is \(\sqrt{3x+1}+2\). Multiplying gives
Simplify the numerator, then cancel the nonzero factor \(x-1\), which is nonzero in the original quotient:
Now substitute into the simplified expression to take the limit:
Check the numerator calculation: \((\sqrt{3x+1})^2-2^2=3x+1-4=3x-3=3(x-1)\). Thus the canceled factor and the resulting value are consistent.
When Rationalization Reveals an Unbounded Quotient
At \(a=0\), the basic quotient has a different outcome. The domain condition \(x\geq 0\) means that \(x\) approaches zero through nonnegative values. For \(x>0\), rationalization gives
As \(x\) approaches zero from the domain, \(\sqrt{x}\) is positive and approaches zero. Its reciprocal grows without bound. Thus the quotient has no finite limit; it tends to \(+\infty\) on its domain side. The algebra identifies why: after rationalization, the denominator \(\sqrt{x}+\sqrt{0}=\sqrt{x}\) approaches zero.
Worked Example: An Unbounded Quotient at the Domain Boundary
Determine the behavior of \(\displaystyle \frac{\sqrt{x}-0}{x-0}\) as \(x\to 0\), with \(x\geq 0\) and \(x\ne 0\). Because the square root is defined only for \(x\geq 0\), values in the domain near zero satisfy \(x>0\).
For every such \(x\), rationalization or direct cancellation gives
To verify that this grows without bound, let \(M>0\). If \(0<x<1/M^2\), then \(\sqrt{x}<1/M\), so \(1/\sqrt{x}>M\). Every positive bound \(M\) is therefore exceeded for all sufficiently small positive \(x\). Hence the quotient tends to \(+\infty\) as \(x\to 0\) through its domain.
Check the algebra by multiplying back: \((1/\sqrt{x})x=\sqrt{x}\) for \(x>0\). The simplified expression agrees with the original quotient everywhere it is defined near zero, and its growth confirms the unbounded behavior.
What the Conjugate Does—and Does Not—Guarantee
Rationalization is an algebraic transformation, not an automatic answer to every limit question. It changes the form of the quotient so that its behavior can be examined. The resulting expression still has to be analyzed, and all domain restrictions remain in force.
| Situation | What rationalization shows |
|---|---|
| The limiting sum of square roots is positive | The simplified denominator stays away from zero; a finite limit may be obtained by evaluating the simplified expression. |
| The limiting sum of square roots is zero | The simplified denominator approaches zero; the quotient may be unbounded or may require more algebra to classify. |
| The radicand is negative near the point | The real-valued square root is not defined there, so this real difference quotient is not available on that interval. |
| The original denominator is zero at the point | The quotient is considered only at nearby domain points where the denominator is nonzero; cancellation must not be treated as defining the original quotient at the point. |
A common error is to cancel a factor and then claim that the original quotient is defined at the point where that factor is zero. Cancellation gives an equal expression only for the values where the original quotient was defined. It can simplify a limit, but it does not repair the original expression at the excluded point.
Common Mistake: Treating Cancellation as Pointwise Equality at the Excluded Point
Consider \(\displaystyle \frac{\sqrt{x}-2}{x-4}\). Multiplying by the conjugate gives, for \(x\ne 4\),
The incorrect conclusion is: “At \(x=4\), the original quotient equals \(1/4\).” This is wrong because substituting \(x=4\) into the original quotient gives \(0/0\), so the original expression is not defined there. The corrected statement is that the two expressions are equal for \(x\geq 0\) with \(x\ne 4\), and the limit can be evaluated using the simplified expression:
Check: the conjugate product is \((\sqrt{x}-2)(\sqrt{x}+2)=x-4\), which verifies the cancellation on the allowed values \(x\ne4\). The limit is \(\frac14\), but the original quotient still has no value at \(x=4\).
A Reliable Method
When a difference quotient contains a difference of square roots, use the following sequence. The steps keep the algebra, limit, and domain conditions separate.
Check that each radicand is nonnegative for the values under consideration, and note where the original denominator is nonzero.
For \(\sqrt{u}-\sqrt{v}\), use \(\sqrt{u}+\sqrt{v}\).
Use \((\sqrt{u}-\sqrt{v})(\sqrt{u}+\sqrt{v})=u-v\), then factor the resulting radicand difference if possible.
Cancel factors that are nonzero for the values in the original quotient, and do not assign the original quotient a value at an excluded point.
Check whether the remaining denominator approaches a positive number or zero, and account for the direction allowed by the domain.
Why This Technique Matters
The conjugate method turns a radical difference into an ordinary difference of radicands. That can uncover a factor matching the difference quotient’s denominator, as in the finite-limit examples. It also prevents a premature conclusion from the initial \(0/0\) form: the quotient may simplify to a finite expression, or it may reduce to an expression that grows without bound.
The key is to make no assumptions beyond the algebra. Verify the conjugate product, track where cancellation is valid, and take the limit only after simplification. These habits make radical difference quotients manageable while keeping the distinction between a limit and a value at a point clear.
Check Your Understanding
For each question, pay attention to the domain and to the values at which cancellation is valid.
- What is the conjugate of \(\sqrt{x+5}-3\), and what difference of squares results from multiplying by it?
- Rationalize \(\displaystyle \frac{\sqrt{x}-4}{x-16}\) for \(x\geq0\), \(x\ne16\), and use the simplified expression to find its limit as \(x\to16\).
- Why does \(\displaystyle \frac{\sqrt{x}}{x}\) grow without bound as \(x\to0\) through \(x>0\)?
- For \(\displaystyle \frac{\sqrt{2x+1}-1}{x}\), what conjugate should be used, and what factor can be canceled when \(x\ne0\)?
- After canceling a factor that vanishes at the limiting point, what must you remember about the original quotient’s value at that point?
