P1.4 - Differentiation

Syllabus
2019
Topic
P1.4
Level
AS

Learning objectives

A derivative is the limiting gradient at one point

For y=f(x)y=f(x), the derivative f(x)f'(x) or dydx\dfrac{dy}{dx} gives the instantaneous rate of change of yy with respect to xx. Geometrically, its value at x=ax=a is the gradient of the tangent to the curve there.

f(a)=limh0f(a+h)f(a)hf'(a)=\lim_{h\to0}\frac{f(a+h)-f(a)}{h}

The quotient is the gradient of the secant through inputs aa and a+ha+h. As hh tends to zero, the second point approaches the first and the secant gradients approach the tangent gradient. This limit is why a derivative describes change at one point rather than only an average over an interval.

For f(x)=x2f(x)=x^2, f(a+h)f(a)h=(a+h)2a2h=2a+h\dfrac{f(a+h)-f(a)}h=\dfrac{(a+h)^2-a^2}{h}=2a+h. Taking h0h\to0 gives f(a)=2af'(a)=2a, so the tangent gradient at x=3x=3 is 66.

Differentiating again gives f(x)=d2ydx2f''(x)=\dfrac{d^2y}{dx^2}, the rate at which the first derivative changes. If yy is displacement and xx is time, for example, dy/dxdy/dx is velocity and d2y/dx2d^2y/dx^2 is acceleration; the units change accordingly.

A derivative is a limit of nearby secant gradients, not the gradient between two fixed points. An instantaneous rate is obtained by evaluating the derivative at the required input.

Rewrite first, then differentiate each power of x

ddx(xn)=nxn1,ddx(c)=0\frac{d}{dx}(x^n)=nx^{n-1},\qquad \frac{d}{dx}(c)=0

The power rule applies term by term to sums, differences and constant multiples. Before using it, expand products and rewrite roots or variables in denominators as powers of xx; this exposes every term in the required form axnax^n.

Original structure Preparation
(2x+5)(x1)(2x+5)(x-1) expand to 2x2+3x52x^2+3x-5
x\sqrt{x} write as x1/2x^{1/2}
1/xp1/x^p write as xpx^{-p}
a quotient over x\sqrt{x} divide each numerator term by x1/2x^{1/2}

For y=x2+5x33xy=\dfrac{x^2+5x-3}{3\sqrt{x}}, first write y=13x3/2+53x1/2x1/2y=\tfrac13x^{3/2}+\tfrac53x^{1/2}-x^{-1/2}. Therefore dydx=12x1/2+56x1/2+12x3/2\dfrac{dy}{dx}=\tfrac12x^{1/2}+\tfrac56x^{-1/2}+\tfrac12x^{-3/2}. Each coefficient is multiplied by its old power, then the power is reduced by one.

A constant differentiates to zero: expanding (2x+5)(x1)(2x+5)(x-1) gives derivative 4x+34x+3. Preserve any original domain restriction; the quotient example still requires x>0x>0 even after it is rewritten.

Turn a derivative value into a tangent or normal

At a point on a curve, differentiation supplies the tangent gradient. The normal is the straight line through the same point that is perpendicular to that tangent.

Step Action
locate the point use the given coordinates, or substitute its xx-value into the curve
tangent gradient find f(x)f'(x) and evaluate mT=f(a)m_T=f'(a)
normal gradient use mN=1/mTm_N=-1/m_T when mT0m_T\ne0
equation substitute the point into yy1=m(xx1)y-y_1=m(x-x_1)

For y=x2+2xy=x^2+2x at x=1x=1, the point is (1,3)(1,3) and dy/dx=2x+2dy/dx=2x+2, so mT=4m_T=4. The tangent is y3=4(x1)y-3=4(x-1). The normal gradient is 14-\tfrac14, so the normal is y3=14(x1)y-3=-\tfrac14(x-1). Substitution of (1,3)(1,3) checks both equations pass through the correct point.

If a tangent must be parallel to a line of gradient mm, solve f(x)=mf'(x)=m for every valid xx, then recover the corresponding point on the curve. For a perpendicular tangent, equate f(x)f'(x) to the negative reciprocal of the given line's gradient.

Do not use f(x)f'(x) as a line equation: it supplies a gradient only after evaluation at the point. When mT=0m_T=0, the tangent is horizontal and its normal is the vertical line x=ax=a, so the negative-reciprocal formula is not finite.