Intuition What this page is for
The parent note proved d x d x n = n x n − 1 for every kind of exponent. This page drills it. We will hit every case the rule can throw at you — a positive integer, a negative integer, a fraction, a root, an irrational exponent, a case where you must rewrite before you differentiate , a real-world rate, and an exam trap that looks like the power rule but isn't. Nobody should meet a scenario in a test that wasn't rehearsed here.
Before symbols, one reminder of what every symbol means — earn them all:
Definition The symbols we reuse everywhere
x — the input , a number we are free to nudge (the horizontal position on a graph).
x n — "x multiplied by itself, in the pattern set by n ". If n = 3 it is x ⋅ x ⋅ x ; if n = − 2 it is x ⋅ x 1 ; if n = 2 1 it is x .
h — a tiny nudge we add to x . The slope is measured by comparing the height at x + h with the height at x , then shrinking h toward 0 (this is the Limit definition of the derivative ). Whenever you see "h → 0 ", picture the nudge vanishing.
d x d ( ) — "the slope of the graph of whatever is inside, at position x ". A slope is rise over run : how many units up the curve climbs for one unit right. A big slope = steep hill; slope 0 = flat ground; negative slope = downhill.
The power rule is a shortcut for that slope: drop the exponent to the front, then lower the exponent by one.
Intuition One caveat we will keep repeating
A slope formula only means anything where the original function exists and is smooth . So for each example we also ask: for which x is this actually valid? Where the curve has a hole, a vertical tangent, or a sharp corner, the formula must be read with care.
Definition What is a "cusp"? (we meet one in Ex 4)
A cusp is a sharp point on a curve where two pieces of the graph meet but the tangent line suddenly flips direction — like the pointed tip of a heart or the bottom of the letter "V". Approaching from the left, the tangent leans one way; approaching from the right, it leans the opposite way; and right at the tip the two directions disagree, so no single slope exists there . Picture the sharp corner of a mountain peak: there is no one "steepness" at the exact tip. Contrast this with a smooth curve, where the tangent turns gradually and always has one clear direction. (See the figure in Ex 4 for the tip of x 2/3 .)
Every problem in this whole topic is one of the cells below. The examples that follow are tagged with the cell they cover.
Cell
Case class
What's tricky about it
Example
A
Positive integer exponent
none — the "easy" base case
Ex 1
B
Exponent = 0 / a plain constant
slope of a flat line is 0
Ex 2
C
Negative integer exponent
answer's sign flips; undefined at x = 0
Ex 3
D
Rational exponent / root
fraction arithmetic; cusp at x = 0
Ex 4
E
"Hidden" power — needs algebra first
3 x , x 5 1 , x x 2 must be rewritten as x n
Ex 5
F
Evaluate the slope at a point (sign of slope)
is the curve going up or down there?
Ex 6
G
Degenerate / undefined input
slope blows up as x → 0
Ex 7
H
Real-world rate (with units)
a word problem; keep the units honest
Ex 8
I
Irrational exponent
does the shortcut still work for x π ?
Ex 9
J
Exam twist — looks like power rule but isn't
2 x is exponential, not a power
Ex 10
Worked example Differentiate
f ( x ) = x 7 .
Forecast: cover the answer and guess. Drop the 7 , lower the exponent... what do you get?
Step 1 — identify n . Here n = 7 , a positive whole number.
Why this step? The rule d x d x n = n x n − 1 needs us to name n first, so we know what to drop and subtract.
Step 2 — drop the exponent to the front. That gives a factor of 7 .
Why this step? This is the first "Drop" of the mnemonic.
Step 3 — lower the exponent by one. 7 − 1 = 6 .
f ′ ( x ) = 7 x 6 .
Why this step? Recall the nudge h : when you expand ( x + h ) 7 − x 7 and divide by h , every term except 7 x 6 still carries a factor of h , so those terms shrink to 0 as the nudge vanishes. Only 7 x 6 survives.
Valid where? Everywhere — a polynomial is smooth for all x .
Verify: at x = 1 , f ′ ( 1 ) = 7 . Slope check from first principles with a tiny h = 0.001 : 0.001 1.00 1 7 − 1 ≈ 7.02 — close to 7 . ✓
Worked example Differentiate
g ( x ) = x 0 , and separately c ( x ) = 5 .
Forecast: x 0 equals 1 (any nonzero number to the zero power is 1 ). The graph of y = 1 is a flat horizontal line. What is the slope of flat ground?
Step 1 — recognise both are constants. x 0 = 1 and 5 are both just numbers; their graphs are horizontal lines.
Why this step? A horizontal line never rises, so its rise-over-run is 0 everywhere.
Step 2 — apply the rule to x 0 . With n = 0 : 0 ⋅ x − 1 = 0.
g ′ ( x ) = 0 , c ′ ( x ) = 0.
Why this step? The factor we "drop to the front" is 0 , which zeroes the whole thing — matching the flat-line intuition.
Valid where? For c ( x ) = 5 , everywhere. For g ( x ) = x 0 , everywhere except x = 0 (since 0 0 is not defined) — but the slope is 0 at every point where it is defined.
Verify: slope of any horizontal line is 0 . ✓ (No units, no motion — nothing changes.)
Worked example Differentiate
f ( x ) = x − 5 .
Forecast: the answer will have a negative number out front. Guess: drop − 5 , then − 5 − 1 = ?
Step 1 — identify n . n = − 5 .
Why this step? The rule holds for any real n ; the parent note proved the negative-integer case by writing x − 5 = 1/ x 5 and taking a common denominator, so we may just apply the shortcut.
Step 2 — drop n to the front. Factor of − 5 .
Step 3 — lower the exponent by one. − 5 − 1 = − 6 . The trap is to add: it is subtract, so it goes more negative .
f ′ ( x ) = − 5 x − 6 = x 6 − 5 .
Why this step? The general pattern for a negative-integer exponent is d x d x − k = − k x − k − 1 (here the whole number k = 5 ).
Valid where? For all x = 0 . The original function x − 5 = 1/ x 5 itself is undefined at x = 0 (you cannot divide by zero), so neither the function nor its slope exists there — the formula − 5/ x 6 correctly blows up as x → 0 .
Verify: f = 1/ x 5 is positive and decreasing for x > 0 (bigger x ⇒ smaller value), so the slope should be negative — and − 5/ x 6 is indeed negative there. ✓
Worked example Differentiate
f ( x ) = x 2/3 .
Forecast: n = 3 2 . The front factor is 3 2 ; the new exponent is 3 2 − 1 . What is 3 2 − 1 ?
Step 1 — identify n = 3 2 (this is 3 x 2 , the cube root of x 2 ).
Why this step? We can't expand x 2/3 with a nudge h the way we did for whole powers, so the parent note used a different route: set y = x 2/3 , cube both sides to get y 3 = x 2 — now both exponents are whole numbers , which we already know how to differentiate. Differentiating that relation (this is Implicit differentiation , because y secretly depends on x ) recovers exactly the shortcut n x n − 1 . So the same "drop and lower" works for fractions.
Step 2 — drop 3 2 to the front.
Step 3 — subtract 1 from the exponent. Write 1 = 3 3 , so 3 2 − 3 3 = − 3 1 .
f ′ ( x ) = 3 2 x − 1/3 = 3 3 x 2 .
Why this step? The common mistake is fumbling 3 2 − 1 ; always turn the 1 into 3 3 first.
Valid where? Because the cube root is defined for negative numbers too, f ( x ) = x 2/3 exists for all x — but the slope 3 3 x 2 is undefined at x = 0 . There the curve has a cusp (defined above — a sharp point where the tangent flips): coming in from the left the tangent points steeply up-left, from the right steeply up-right, and the two don't agree, so no single slope exists at x = 0 (look at the sharp tip in the figure). For x < 0 the formula still applies (e.g. at x = − 8 , 3 − 8 = − 2 , so f ′ ( − 8 ) = 3 ⋅ ( − 2 ) 2 = − 3 1 : the curve descends on the left branch).
Verify: at x = 8 : f ′ ( 8 ) = 3 2 ⋅ 8 − 1/3 = 3 2 ⋅ 2 1 = 3 1 . Numerically 0.001 8.00 1 2/3 − 8 2/3 ≈ 0.333 . And at x = − 8 : f ′ ( − 8 ) = − 3 1 . ✓
Worked example Differentiate
f ( x ) = x x 2 .
Forecast: you cannot apply the power rule to a fraction directly. First guess: can we rewrite this as a single x something ?
Step 1 — rewrite as one power. x = x 1/2 , and dividing subtracts exponents:
x 1/2 x 2 = x 2 − 1/2 = x 3/2 .
Why this step? The power rule only sees x n . Turning the expression into a single clean x 3/2 is the whole battle; the differentiation is then trivial.
Step 2 — apply the rule with n = 2 3 . Drop 2 3 ; new exponent 2 3 − 1 = 2 1 .
f ′ ( x ) = 2 3 x 1/2 = 2 3 x .
Why this step? This is where "rewrite before you differentiate" saves you from the Quotient rule , which would be slower and error-prone here.
Valid where? For x > 0 (the x in the original needs a non-negative input, and the division needs x = 0 ).
Verify: at x = 4 : f ′ ( 4 ) = 2 3 4 = 2 3 ⋅ 2 = 3 . Numeric slope 0.001 f ( 4.001 ) − f ( 4 ) ≈ 3.0 . ✓
f ( x ) = x 3 , find the slope at x = − 2 and at x = 2 . Is the curve going up or down at each?
Forecast: x 3 is one connected curve. Guess: does it ever go downhill?
Step 1 — differentiate. f ′ ( x ) = 3 x 2 (Cell A: n = 3 ).
Step 2 — plug in both points.
f ′ ( − 2 ) = 3 ( − 2 ) 2 = 3 ⋅ 4 = 12 , f ′ ( 2 ) = 3 ( 2 ) 2 = 12.
Why this step? ( − 2 ) 2 = + 4 — squaring kills the sign. So the slope is + 12 at both points. The tangent lines are equally steep and both point uphill .
Step 3 — read the geometry (look at the figure below: the red tangent lines both climb).
At both x = ± 2 : slope = 12 > 0 ⇒ curve is rising.
Why this step? Since 3 x 2 ≥ 0 for all x , the cubic never goes downhill — a fact you would miss without checking the sign.
Valid where? Everywhere — smooth polynomial.
Verify: f ′ ( x ) = 3 x 2 is zero only at x = 0 and positive elsewhere; slopes at ± 2 are both 12 . ✓
f ( x ) = x = x 1/2 , what happens to the slope as x → 0 + ?
Forecast: near x = 0 the x curve looks almost vertical. Guess: does the slope stay finite or shoot to infinity?
Step 1 — differentiate. n = 2 1 , so f ′ ( x ) = 2 1 x − 1/2 = 2 x 1 (Cell D).
Step 2 — probe the limit. As x → 0 + , x → 0 , so 2 x 1 → + ∞ .
lim x → 0 + f ′ ( x ) = + ∞.
Why this step? The formula stays valid, but at x = 0 itself the slope is undefined — the tangent is vertical.
Step 3 — the other end. As x → ∞ , 2 x 1 → 0 : the curve flattens out.
Why this step? Covering both limiting directions is the point of this cell — steep near 0 , flat far out (see the figure).
Valid where? For x > 0 . At x = 0 the tangent is vertical (no finite slope); x is not defined for x < 0 .
Verify: f ′ ( 0.01 ) = 2 0.01 1 = 0.2 1 = 5 (large); f ′ ( 100 ) = 2 ⋅ 10 1 = 0.05 (tiny). ✓
Worked example A balloon's volume is
V ( r ) = 3 4 π r 3 (cm3 , with radius r in cm). How fast does volume grow per cm of radius when r = 5 cm?
Forecast: "how fast volume grows per cm of radius" is a slope: d r d V . Guess which power of r survives.
Step 1 — differentiate, treating 3 4 π as a constant multiplier. With n = 3 : d r d r 3 = 3 r 2 .
d r d V = 3 4 π ⋅ 3 r 2 = 4 π r 2 .
Why this step? A constant out front just rides along — the power rule acts only on the r 3 . (Notice 4 π r 2 is the surface area — a lovely sanity signpost.)
Step 2 — evaluate at r = 5 .
d r d V r = 5 = 4 π ( 25 ) = 100 π ≈ 314.16 cm 3 per cm .
Why this step? The question asked for a specific radius; plug in last, after the algebra.
Valid where? For r > 0 (a physical radius is positive).
Verify: units are cm cm 3 = cm 2 (an area rate — matches surface area). Numerically 0.001 V ( 5.001 ) − V ( 5 ) ≈ 314.2 . ✓
Worked example Differentiate
f ( x ) = x π (yes, the exponent is the number π ≈ 3.14159 ).
Forecast: does "drop and lower" still work when the exponent isn't a fraction at all?
Step 1 — check the rule still applies. π is a real number, and the power rule was proved for every real exponent — a fraction like 7 22 obeys it, and π is the limit of such fractions, so the shortcut carries over unchanged.
Why this step? We are not allowed to expand ( x + h ) π with a nudge (no finite expansion exists for a non-whole power), so we lean on the fact that the rule extends to all real n — you simply treat π as "just a number".
Step 2 — drop π to the front, lower the exponent by one.
f ′ ( x ) = π x π − 1 .
Why this step? Same mechanical move as always; the only oddity is that the new exponent π − 1 ≈ 2.14159 is itself irrational — that is fine.
Valid where? For x > 0 (a general real power like x π is only defined for positive bases).
Verify: at x = 1 the exact answer is f ′ ( 1 ) = π ⋅ 1 π − 1 = π ≈ 3.1416 . To confirm this with the Limit definition of the derivative , use a tiny nudge so the secant slope (rise over run between two nearby points) approaches the tangent slope: h f ( 1 + h ) − f ( 1 ) with h = 0.001 gives 0.001 1.00 1 π − 1 π = 0.001 1.003146 − 1 ≈ 3.146 . Why this step? As h → 0 this secant slope converges to the true derivative π ≈ 3.1416 , and 3.146 is already close (a bigger h would overshoot because the curve bends). ✓
Worked example Differentiate
f ( x ) = 2 x . (Careful — is this a power rule problem?)
Forecast: it is tempting to write x ⋅ 2 x − 1 . Pause: which is the variable — the base or the exponent?
Step 1 — diagnose the structure. In x n the base is the variable , the exponent is fixed. In 2 x the base is fixed (2 ) and the exponent is the variable . These are opposite creatures.
Why this step? The power rule was proved for a variable base, constant exponent . It does not apply here at all — using it is a category error.
Step 2 — use the exponential rule instead (see Derivative of exponential functions ):
d x d a x = a x ln a ⇒ d x d 2 x = 2 x ln 2.
Why this step? This is the correct tool; ln 2 ≈ 0.693 is the "growth constant" of base 2 .
Valid where? Everywhere — 2 x is defined and smooth for all real x .
Verify: at x = 3 : correct answer 2 3 ln 2 = 8 ⋅ 0.693 ≈ 5.545 ; the wrong power-rule guess x 2 x − 1 = 3 ⋅ 4 = 12 — different, confirming they are not the same rule. ✓
Common mistake The traps these examples defuse
Subtracting wrong for negatives (Ex 3): − 5 − 1 = − 6 , not − 4 .
Fraction arithmetic in n − 1 (Ex 4): turn 1 into 3 3 before subtracting.
Forgetting the domain (Ex 3, 4, 7): a slope formula is only valid where the original function exists and is smooth — watch x = 0 (division, vertical tangent, cusp).
Differentiating a fraction directly (Ex 5): rewrite as a single x n first.
Confusing x n with n x (Ex 10): variable base vs variable exponent.
Recall Self-test (predict, then reveal)
Differentiate 3 x 1 from scratch, and state where it is valid.
::: Rewrite as x − 1/3 (Cell E). Then n = − 3 1 : f ′ = − 3 1 x − 1/3 − 1 = − 3 1 x − 4/3 = 3 3 x 4 − 1 . Valid for x = 0 (undefined at x = 0 ). ✓
d x d x 7 ?7 x 6 .
d x d x 0 and d x d 5 ?Both 0 (flat lines have zero slope).
d x d x − 5 , and where is it valid?− 5 x − 6 = − 5/ x 6 ; valid for x = 0 .
d x d x 2/3 , and what happens at x = 0 ?3 2 x − 1/3 ; at x = 0 there is a cusp, so no slope exists there.
What is a cusp? A sharp point on a curve where the tangent flips direction, so no single slope exists there.
Before differentiating x x 2 , what must you do? Rewrite as one power: x 3/2 .
Why is the slope of x 3 never negative? f ′ = 3 x 2 ≥ 0 for all x (squares are non-negative).
What happens to d x d x as x → 0 + ? It goes to + ∞ (vertical tangent; undefined at x = 0 ).
d x d x π ?π x π − 1 (the rule works for every real exponent; valid for x > 0 ).
d r d V for V = 3 4 π r 3 ?4 π r 2 (the surface area).
Is d x d 2 x = x 2 x − 1 ? No — 2 x is exponential; answer is 2 x ln 2 .