1.7 - Differentiating powers
Chord gradients explain what a derivative means; rules make it quick to find one. You will differentiate rational powers, rewrite roots and fractions before differentiating, and find second derivatives. Keeping the original domain in view will prevent an algebraic answer being used where the function is undefined.
The power rule
How can you find the gradient of without building a new chord table at every point? The same rule that gives for works for rational powers, including fractions and negative numbers.
For positive integer powers, expanding a nearby value explains the pattern. For example, the chord gradient for at is ; the terms containing vanish in the limit. The result is . The rule below extends to rational exponents on intervals where the function is defined and differentiable.
The power rule
For a constant rational number ,
Use the rule where the real power and its derivative are defined. A constant function has derivative zero.
In the formula booklet.
Multiply by the old exponent, then subtract from that exponent. For instance, gives , and gives . A constant such as has a horizontal graph, so its derivative is .
Worked example 1: a fractional power
Question. Differentiate for , and find the gradient at .
Plan. Apply the power rule, then substitute into the derivative.
Answer. , with gradient at .
Check. The nearby chord from to has gradient , close to .
Change just the exponent's sign. A negative exponent makes the function fall on the positive domain.
Worked example 2: a negative fractional power
Question. Differentiate for , and find the gradient at .
Plan. Keep the negative coefficient and subtract from .
Answer. , with gradient at .
Check. decreases as positive grows, agreeing with the negative derivative. The nearby chord to has gradient approximately , close to .
The domain matters. A reciprocal power excludes . For , the real function allows , but the formula gives a finite derivative only for . At , the curve has a vertical direction, not a finite gradient. Odd roots can allow negative inputs: for .
Watch out
, not . Subtracting makes a negative exponent more negative. Write it with a common denominator if needed.
Apply that same change to a new negative fractional power.
Sums and multiples
A root or a fraction may hide a simple power. Rewrite it before applying the rule: , , and . Differentiate each term and keep any constant multiplier.
Why can terms be treated separately? A chord's total change in is three times the change in minus twice the change in . Taking the limiting gradient preserves that combination.
Sums and constant multiples
Where and are differentiable and are constants,
The derivative of a constant is zero.
Not stated separately in the formula booklet: learn these rules.
Keep rewriting and differentiation as separate lines. This makes the exponents visible before they change.
Worked example 3: a root in a sum
Question. Differentiate , for .
Plan. Rewrite the root as a power and differentiate term by term.
Answer. .
Check. Adding only moves the graph upwards, so it should not alter any gradient. Its disappearance is appropriate.
Move only the root into the denominator. This changes its exponent and therefore the sign of its derivative contribution.
Worked example 4: a reciprocal root in a sum
Question. Differentiate , for .
Plan. Write the reciprocal root as , then keep the two negative signs separate.
Answer. .
Check. At , . The chord from to has gradient approximately , consistent with this value.
Algebra may also remove an apparent quotient. For , , so its derivative is on the original domain. Cancelling does not put the excluded input back into the original function.
Watch out
, not . The original coefficient multiplies the exponent. Also, differentiating a constant gives , not that constant.
Rewrite first, then differentiate.
Repeated differentiation
The first derivative is another function, so apply the same rules to it to find the second derivative. Differentiate the complete first derivative before substituting a particular value of . Substituting too soon turns it into a number and loses how it changes.
Evaluating a derivative
means find the gradient function , then evaluate it at . Similarly, means evaluate the second derivative there. These are different from , the function's height.
Use one line for each differentiation so that each change of exponent can be checked.
Worked example 5: two polynomial derivatives
Question. For , find , , and both derivatives at .
Plan. Differentiate twice, then substitute into the two resulting functions.
Answer. First derivative ; second derivative . Their values at are and .
Check. The gradient changes from to . Its change divided by is approximately , matching the second derivative.
Change only the exponent in the middle term from to . Track its signs through both derivatives.
Worked example 6: differentiating a reciprocal twice
Question. For , , find the first and second derivatives and their values at .
Plan. Apply the power rule separately at each differentiation.
Answer. and , for . At , they are and .
Check. A centred change in the first derivative from to , divided by , is approximately , close to .
The opening function now needs only one line: its gradient is for . The limiting meaning remains the same even though the calculation is quicker.
Watch out
For , the second derivative is , not . Differentiate the first derivative; do not square it.
Keep until both differentiation steps are complete.
Quick Recap
- For a rational exponent, multiply by and reduce the exponent by .
- Rewrite roots and reciprocal terms as powers before differentiating.
- Keep constant multipliers; the derivative of a constant is zero.
- Differentiate again for a second derivative, then evaluate at the required input.
- Preserve the original domain and check where the derivative is finite.
For , the tangent gradient is on .