Unit 1 differentiated y=f(x). Here the function may be tangled up with y, or defined through a parameter, or exponential — and each case has its own technique.
Exponential and logarithmic derivatives
dxd(ex)=ex
The exponential function is its own derivative — the property that defines e.
With the chain rule:
dxd(ekx)=kekxdxd(ef(x))=f′(x)ef(x)
dxd(e3x)=3e3xdxd(ex2)=2xex2
For logarithms:
dxd(lnx)=x1dxd(lnf(x))=f(x)f′(x)
dxdln(3x2+1)=3x2+16x
And for a general base:
dxd(ax)=axlna
Combining rules. For y=xlnx, the product rule gives
dxdy=(1)(lnx)+(x)(x1)=lnx+1
Implicit differentiation
Some curves cannot be written as y=f(x) — a circle fails the vertical line test. Differentiate every term with respect to x, using the chain rule whenever y appears:
dxd(y2)=2ydxdydxd(y3)=3y2dxdy
Differentiate x2+y2=25 implicitly.
2x+2ydxdy=0⇒dxdy=−yx
At (3,4) the gradient is −43. Check geometrically: the radius to that point has gradient 34, and a tangent is perpendicular to the radius ✓
Notice that the first two differ only in sign, and that none of the three contains a trigonometric function — which is why they turn up as answers to integration questions involving surds and quadratics.
With the chain rule:
dxdtan−1(2x)=1+4x22
Higher derivatives
Differentiate repeatedly:
y=e2x⇒dxdy=2e2x⇒dx2d2y=4e2x
Each differentiation of ekx brings down another factor of k, so dxndny=knekx.
For y=lnx:
dxdy=x−1dx2d2y=−x−2dx3d3y=2x−3
Applications
Everything from Unit 1 still applies — the new derivatives simply widen the range of curves.
Stationary points. For y=xe−x:
dxdy=e−x−xe−x=e−x(1−x)=0
Since e−x is never zero, x=1, giving y=e−1. The second derivative is e−x(x−2), which at x=1 is −e−1<0 — a maximum at (1,e1).
Tangents to implicit curves. Find dxdy implicitly, substitute the point's coordinates (both of them), and use y−y1=m(x−x1) as usual.