A small collection of derivative formulas covers the vast majority of differentiation work. Polynomials, trigonometric functions, exponentials, and logarithms appear constantly, and their derivatives should be available from memory. Every more complex derivative—a composition, a product, an implicitly defined relation—ultimately reduces to these building blocks combined through differentiation rules.
Each formula below is provable from the limit definition. For some, the proof is a direct algebraic manipulation of the difference quotient. For others, it requires special limits such as limh→0hsinh=1. The proofs confirm the formulas; the formulas replace the proofs in daily computation.
Key Terms
Derivative— the formulas on this page give f′(x) for standard functions
The graph of a constant function is a horizontal line. Every secant line through two points on it has slope zero, so the tangent line at every point also has slope zero. From the limit definition, the difference quotient hc−c=0 for all h=0, and the limit is 0.
Under differentiation, constant terms vanish. In any sum f(x)+c, the constant contributes nothing to the derivative: dxd[f(x)+c]=f′(x).
The rule applies uniformly across exponent types. Positive integers: dxd[x5]=5x4. Negative integers: dxd[x−3]=−3x−4. Fractions: dxd[x1/2]=21x−1/2=2x1. Irrational exponents: dxd[xπ]=πxπ−1.
For positive integer n, the proof expands (x+h)n using the binomial theorem. The leading term after cancellation is nxn−1h, and dividing by h and taking the limit gives nxn−1. All higher-order terms contain h as a factor and vanish in the limit.
Pitfall: Lowering a negative exponent toward zero
Computing −3+1 instead of −3−1 turns dxd[x−3] into −3x−2, which is −0.75 at x=2. The rule always subtracts 1: the derivative is −3x−4, which is −0.1875 at x=2.
Extension to negative and fractional exponents uses the quotient rule and the chain rule respectively, or alternatively logarithmic differentiation: writing xn=enlnx and applying the chain rule gives dxd[xn]=enlnx⋅xn=xn⋅xn=nxn−1.
Exponent type
Function
Derivative
Positive integer
x⁵
5x⁴
Negative integer
x⁻³ = 1/x³
−3x⁻⁴
Positive fraction
x1/2 = √x
(1/2)x−1/2 = 1/(2√x)
Negative fraction
x−1/2 = 1/√x
−(1/2)x−3/2
Irrational
xπ
π · xπ−1
f(x) = x² and f′(x) = 2x, marked at x = 1
The upper curve is the function, the lower one its derivative, and at x = 1 the marker reads f′(1) = 2, the slope of the parabola there. The derivative is a straight line because bringing the exponent down leaves degree one: that is the power rule drawn out. Move the marker and watch the slope of the parabola track the height of the line on the derivative explorer.
Every polynomial derivative is this picture repeated term by term.
Polynomials
A polynomial p(x)=anxn+an−1xn−1+⋯+a1x+a0 is differentiated term by term using the power rule, constant multiple rule, and sum rule:
Polynomials
p′(x)=nanxn−1+(n−1)an−1xn−2+⋯+a1
Each term drops its degree by one. The constant term a0 disappears. A polynomial of degree n has a derivative of degree n−1.
Polynomials are differentiable at every real number. Their derivatives are again polynomials, so the process can be repeated indefinitely. After n differentiations, a degree-n polynomial becomes a constant n!⋅an. One more differentiation yields zero, and all subsequent derivatives remain zero. This termination property distinguishes polynomials from transcendental functions, whose derivatives cycle or persist indefinitely.
Trigonometric Functions — Sine and Cosine
The two fundamental trigonometric derivatives are:
The sine derivative is proved from the limit definition. Expanding sin(x+h) using the angle addition formula gives sinxcosh+cosxsinh. The difference quotient becomes
Trigonometric Functions — Sine and Cosine
hsinx(cosh−1)+cosxsinh=sinx⋅hcosh−1+cosx⋅hsinh
The special limitslimh→0hsinh=1 and limh→0hcosh−1=0 yield the result: cosx.
Fact: Degrees would spoil the formula
Measured in degrees, dxd[sinx∘]=180πcosx∘≈0.01745cosx∘, because hsinh tends to 1 only when h is in radians. Radians are the one unit that makes the extra factor exactly 1.
All the trigonometric formulas on this page measure x in radians. On the unit circle, an angle of h radians cuts off an arc of length exactly h, and for small h the height sinh of the arc's endpoint is almost the arc itself. That is the geometric content of hsinh→1.
The cosine derivative follows similarly, or by differentiating cosx=sin(π/2−x) using the chain rule. The negative sign in (cosx)′=−sinx is essential and a frequent source of error.
Repeated differentiation cycles with period four: sinx→cosx→−sinx→−cosx→sinx. This periodicity extends to higher-order derivatives: dxndn[sinx]=sin(x+nπ/2).
f(x) = sin x and f′(x) = cos x, marked at x = 1
The derivative curve peaks where the sine climbs fastest, at x = 0, and crosses zero where the sine tops out, at π/2: the cosine is the sine’s slope read off point by point. At x = 1 the marker reads cos 1 ≈ 0.54. Slide the marker to a peak of the sine and watch the derivative hit zero on the derivative explorer.
The quarter-period shift between a function and its derivative is the whole content of these two formulas.
For tanx=cosxsinx, the quotient rule gives cos2xcosx⋅cosx−sinx⋅(−sinx)=cos2xcos2x+sin2x=cos2x1=sec2x.
A pattern runs through the six derivatives: the cofunctions—cos, cot, csc—each carry a negative sign in their derivatives, while sin, tan, sec do not. This sign pattern is worth noting as a memory aid rather than a coincidence.
Each derivative is valid on the domain of the original function. The derivatives of tanx and secx are undefined at x=π/2+kπ (where cosx=0). The derivatives of cotx and cscx are undefined at x=kπ (where sinx=0).
Function
Derivative
Sign
Where derivative is undefined
sin x
cos x
+
nowhere (defined on all ℝ)
cos x
−sin x
−
nowhere (defined on all ℝ)
tan x
sec² x
+
x = π/2 + kπ
cot x
−csc² x
−
x = kπ
sec x
sec x · tan x
+
x = π/2 + kπ
csc x
−csc x · cot x
−
x = kπ
Every trigonometric derivative comes from those of sine and cosine, through the quotient rule.
Exponential Functions
The natural exponential function has the unique property of being its own derivative:
No other function satisfies f′(x)=f(x) except constant multiples Cex.
This self-replicating property is what makes e the natural base for exponential functions.
The proof from the limit definition uses the special limitlimh→0heh−1=1. The difference quotient for ex factors as ex⋅heh−1, and the limit gives ex⋅1=ex.
Bringing the exponent down turns dxd[2x] into x⋅2x−1, which is 0 at x=0. The variable is in the exponent, so the derivative is 2xln2, which is ln2≈0.693 at x=0.
This follows from rewriting ax=exlna and applying the chain rule. The factor lna is constant—it scales the derivative. When a=e, lna=1 and the factor disappears, confirming the special role of base e.
f(x) = eˣ and f′(x) = eˣ coincide
The two curves lie exactly on top of each other: at every x the slope of eˣ equals its height, so at x = 1 both readings are e ≈ 2.72. No other function has this property, which is why e is the natural base. Check any point and see the two readouts agree on the derivative explorer.
Every other exponential differs from its derivative only by the constant factor ln a.
This can be derived by implicit differentiation. If y=lnx, then ey=x. Differentiating both sides: eydxdy=1, so dxdy=ey1=x1.
Alternatively, the limit definition gives limh→0hln(x+h)−lnx=limh→0h1ln(xx+h)=limh→0h1ln(1+xh), which evaluates to x1 using the limit definition of e.
This follows from the change of base formula logax=lnalnx and the constant multiple rule. The natural logarithm gives the simplest derivative—another reason e is the preferred base.
RULE
The function ln∣x∣ extends the domain to all x=0, and its derivative is x1 for both positive and negative x.
This extended form appears frequently in integration.
f(x) = ln x and f′(x) = 1/x, marked at x = 1
The logarithm climbs steeply near zero and flattens as x grows, and its derivative curve 1/x records exactly that: large on the left, small on the right, equal to 1 at x = 1 where the marker sits. Move the marker toward zero and watch the slope reading run away on the derivative explorer.
The reciprocal derivative is what makes the logarithm the antiderivative of 1/x.
Summary of Common Derivatives
The complete set of common derivative formulas:
Summary of Common Derivatives
dxd[c]=0dxd[xn]=nxn−1
Summary of Common Derivatives
dxd[sinx]=cosxdxd[cosx]=−sinx
Summary of Common Derivatives
dxd[tanx]=sec2xdxd[cotx]=−csc2x
Summary of Common Derivatives
dxd[secx]=secxtanxdxd[cscx]=−cscxcotx
Summary of Common Derivatives
dxd[ex]=exdxd[ax]=axlna
Summary of Common Derivatives
dxd[lnx]=x1dxd[logax]=xlna1
These twelve formulas, combined with the differentiation rules, handle every explicit function built from powers, trigonometric functions, exponentials, and logarithms. Functions involving inverse trigonometric, hyperbolic, or piecewise definitions require the additional formulas collected in derivatives of special functions. The master reference card below collects everything in one place, grouped by function family for quick lookup.
Five mistakes account for most wrong answers when these formulas are applied. Four of them are marked where they happen in the sections above; the sign of the cosine derivative is the one the sine and cosine section warns against. Here each is worked through in full, with the wrong and the right working side by side.
Applying the power rule to a constant
A constant written as a power, such as e3, π2 or 54, is differentiated as if its base were the variable.
Wrong working: dxd[e3]=3e2, by the power rule.
Right working: e3≈20.09 is a fixed number; nothing in it depends on x. By the constant rule, dxd[e3]=0.
Why it happens: the expression has the shape "something to a power", and the power rule is matched by shape. The rule is about xn, where the base is the variable and the exponent is fixed.
How to check: look for x. If the expression contains no x, its graph is a horizontal line and the derivative is 0.
Lowering a negative exponent toward zero
With a negative exponent, 1 is added to it instead of subtracted, so the new exponent moves toward zero.
Wrong working: dxd[x−3]=−3x−3+1=−3x−2. At x=2 this gives −43=−0.75.
Right working: the rule gives n−1, and −3−1=−4. So dxd[x−3]=−3x−4, and at x=2 this is −163=−0.1875.
Why it happens: for positive exponents, subtracting 1 makes the exponent smaller in size, and that habit is carried over. For negative exponents, subtracting 1 makes the exponent larger in size.
How to check: compute a difference quotient. For f(x)=x−3, 0.001f(2.001)−f(2)≈−0.1873, close to −0.1875 and far from −0.75.
Dropping the minus sign on the cosine derivative
The derivative of cosx is written as sinx, without the minus sign, by analogy with dxd[sinx]=cosx.
Wrong working: dxd[cosx]=sinx, so the slope of cosx at x=2π is sin2π=1.
Right working: dxd[cosx]=−sinx, so the slope at x=2π is −1.
Why it happens: the two formulas are learned as a pair, and the pair looks symmetric. It is not: going from sine to cosine keeps the sign, going from cosine to sine changes it.
How to check: think of the graph. Between 0 and π the cosine falls from 1 to −1, so its slope there must be negative, and −sinx is negative on that interval.
Using the power rule on an exponential
In 2x the variable is the exponent, but the power rule is applied as if it were xn: the exponent is brought down and lowered by one.
Wrong working: dxd[2x]=x⋅2x−1. At x=0 this gives 0.
Right working: the exponential rule applies. dxd[2x]=2xln2, and at x=0 this is ln2≈0.693.
Why it happens: x2 and 2x use the same symbols, and the power rule is the first rule learned, so it is applied by habit.
How to check: compute a difference quotient at x=0. 0.00120.001−20≈0.6934, which is not 0. An exponential with a base above 1 is always increasing, so its slope cannot be 0 anywhere.
Dropping the factor ln a
The derivative of logax is written as x1, the formula for the natural logarithm, for every base.
Wrong working: dxd[log10x]=x1, so the slope at x=1 is 1.
Right working: log10x=ln10lnx, so dxd[log10x]=xln101. At x=1 this is ln101≈0.434.
Why it happens: x1 is the logarithm derivative used most often, and the base is easy to overlook.
How to check: compute a difference quotient at x=1. 0.001log101.001−log101≈0.4341, close to 0.434 and not 1. Only for base e, where lne=1, is the derivative exactly x1.
Common Derivatives FAQ
What is the derivative of e^x?
+
It is e^x — the natural exponential is its own derivative. No other function has this property except constant multiples Ce^x, and that self-replication is what makes e the natural base. The proof factors the difference quotient as e^x · (e^h − 1)/h and applies the special limit (e^h − 1)/h → 1.Read more →
What is the derivative of ln x?
+
It is 1/x. The quickest derivation is implicit: if y = ln x then e^y = x, and differentiating gives e^y · dy/dx = 1, so dy/dx = 1/e^y = 1/x. The extended form ln|x| has the same derivative 1/x for every x ≠ 0, which is why it appears throughout integration.Read more →
What is the derivative of a^x?
+
For any base a > 0 with a ≠ 1, the derivative is a^x · ln a. It follows from rewriting a^x as e^(x ln a) and applying the chain rule, which leaves the constant factor ln a scaling the result. When a = e the factor becomes ln e = 1 and disappears — the reason base e is treated as the natural choice.Read more →
What is the derivative of tan x?
+
It is sec²x. Writing tan x as sin x / cos x and applying the quotient rule gives (cos x · cos x − sin x · (−sin x)) / cos²x, whose numerator collapses to cos²x + sin²x = 1 by the Pythagorean identity, leaving 1/cos²x = sec²x. The companion results are cot′ = −csc²x, sec′ = sec x tan x and csc′ = −csc x cot x.Read more →