Calculus › Section M · Vector-valued functions › Topic M.2

Differentiating and integrating vector functions

Differentiation and integration were built for numbers, yet a moving point outputs a vector. The rescue is the same one that carried limits and continuity: split the vector into components and do the ordinary calculus you already know, three times in a row.

The derivative, defined honestly

For an ordinary function, the derivative measures the limit of slopes. For a curve traced by \(\mathbf{r}(t)\), copy the same definition:

\mathbf{r}'(t) = \lim_{\Delta t \to 0}\frac{\mathbf{r}(t + \Delta t) - \mathbf{r}(t)}{\Delta t}

The numerator is the displacement vector from one point of the curve to another, so the difference quotient points along the secant line joining them. As \(\Delta t\) shrinks, the secant turns into the tangent, and the limit is a vector that points along the curve in the direction of increasing \(t\). Where that limit exists, \(\mathbf{r}\) is differentiable at \(t\).

Differentiate component by component

Limits of vector functions split componentwise, so the derivative does too: if \(\mathbf{r}(t) = f(t)\mathbf{i} + g(t)\mathbf{j} + h(t)\mathbf{k}\), then

\mathbf{r}'(t) = f'(t)\mathbf{i} + g'(t)\mathbf{j} + h'(t)\mathbf{k}

Every derivative rule you learned for ordinary functions applies to each component: power rule, chain rule, product with a scalar, all of it. For \(\mathbf{r}(t) = \sin t\,\mathbf{i} + t^2\,\mathbf{j} + e^t\,\mathbf{k}\), the derivative is \(\cos t\,\mathbf{i} + 2t\,\mathbf{j} + e^t\,\mathbf{k}\). The magnitude \(\|\mathbf{r}'(t)\|\) is the speed along the curve, and \(\mathbf{r}'(t)\) itself is the velocity vector; the next topic puts those words to work.

Worked example: a tangent vector

For \(\mathbf{r}(t) = 2t\,\mathbf{i} + t^2\,\mathbf{j}\), find \(\mathbf{r}'(t)\) and the tangent vector at \(t = 1\). Differentiating each component gives \(\mathbf{r}'(t) = 2\mathbf{i} + 2t\,\mathbf{j}\), so \(\mathbf{r}'(1) = 2\mathbf{i} + 2\mathbf{j}\), drawn with its tail at the point \(\mathbf{r}(1) = 2\mathbf{i} + \mathbf{j}\). Check the geometry: the curve is \(x = 2t\), \(y = t^2\), that is \(y = \frac{x^2}{4}\), whose slope at \(x = 2\) is \(\frac{x}{2} = 1\). The tangent vector \(2\mathbf{i} + 2\mathbf{j}\) rises 2 for every 2 across, slope 1. The derivative vector really is tangent, and it points the way the curve is travelling.

Higher derivatives

Nothing stops you from differentiating again. The second derivative \(\mathbf{r}''(t)\) is the derivative of the velocity vector, the acceleration, and third and higher derivatives follow the same pattern: differentiate each component as many times as asked. For the helix \(\mathbf{r}(t) = \cos t\,\mathbf{i} + \sin t\,\mathbf{j} + 2t\,\mathbf{k}\): first \(\mathbf{r}'(t) = -\sin t\,\mathbf{i} + \cos t\,\mathbf{j} + 2\mathbf{k}\), then \(\mathbf{r}''(t) = -\cos t\,\mathbf{i} - \sin t\,\mathbf{j}\). A quick check: \(\mathbf{r}'(t) \cdot \mathbf{r}''(t) = \sin t\cos t - \cos t\sin t + 0 = 0\), so on this helix the acceleration is always perpendicular to the velocity.

Smooth curves

A parametrisation is smooth on an interval when \(\mathbf{r}'(t)\) is continuous and never the zero vector there. The zero-vector condition is the geometric one: \(\mathbf{r}'(t) = \mathbf{0}\) means the moving point stops dead, and a stopped point can turn a corner. The curve \(\mathbf{r}(t) = t^2\,\mathbf{i} + t^3\,\mathbf{j}\) has \(\mathbf{r}'(t) = 2t\,\mathbf{i} + 3t^2\,\mathbf{j}\), which vanishes only at \(t = 0\); there the curve, whose equation is \(y^2 = x^3\), has a sharp cusp at the origin. Away from that instant the motion glides, so the curve is smooth on \((0, \infty)\) and on \((-\infty, 0)\), but not across 0.

Product rules, three of them

Derivatives interact with the three ways of multiplying vectors, and each pairing keeps the ordinary product-rule shape: first-times-derivative-of-second plus derivative-of-first-times-second. With \(w(t)\) a scalar function:

\frac{d}{dt}\big[w(t)\mathbf{r}(t)\big] = w'(t)\mathbf{r}(t) + w(t)\mathbf{r}'(t)

For dot products, \(\frac{d}{dt}[\mathbf{r} \cdot \mathbf{u}] = \mathbf{r}' \cdot \mathbf{u} + \mathbf{r} \cdot \mathbf{u}'\), and for cross products \(\frac{d}{dt}[\mathbf{r} \times \mathbf{u}] = \mathbf{r}' \times \mathbf{u} + \mathbf{r} \times \mathbf{u}'\), where the order of the cross must be preserved because crossing is anticommutative.

Worked example: circular motion stays perpendicular

A point moves on the unit circle: \(\mathbf{r}(t) = \cos t\,\mathbf{i} + \sin t\,\mathbf{j}\). Its distance from the origin is constantly 1, and the product rules reveal what that forces. Since \(\mathbf{r} \cdot \mathbf{r} = \cos^2 t + \sin^2 t = 1\) is constant,

\frac{d}{dt}[\mathbf{r} \cdot \mathbf{r}] = 2\,\mathbf{r} \cdot \mathbf{r}' = 0

so \(\mathbf{r}'(t) = -\sin t\,\mathbf{i} + \cos t\,\mathbf{j}\) is perpendicular to \(\mathbf{r}(t)\) at every instant: check \(\cos t(-\sin t) + \sin t\cos t = 0\). The reasoning runs backwards too: whenever \(\|\mathbf{r}\|\) is constant, \(\mathbf{r} \cdot \mathbf{r}' = 0\), and whenever \(\mathbf{r} \cdot \mathbf{r}' = 0\) at all times, the length never changes. Fixed length means the velocity never points in or out, only around.

Integration, component by component

Integration follows the same splitting principle. The indefinite integral of \(\mathbf{r}(t) = f(t)\mathbf{i} + g(t)\mathbf{j} + h(t)\mathbf{k}\) is

\int \mathbf{r}(t)\,dt = \left(\int f(t)\,dt\right)\mathbf{i} + \left(\int g(t)\,dt\right)\mathbf{j} + \left(\int h(t)\,dt\right)\mathbf{k}

Three antiderivatives means three constants, gathered into one constant vector \(\mathbf{C}\). Definite integrals work the same way and produce a vector:

\int_0^{\pi}(\cos t\,\mathbf{i} + \sin t\,\mathbf{j})\,dt = \big[\sin t\big]_0^{\pi}\mathbf{i} + \big[-\cos t\big]_0^{\pi}\mathbf{j} = 0\,\mathbf{i} + 2\,\mathbf{j}

Worked example: recovering a drone's path

A drone's velocity is recorded as \(\mathbf{r}'(t) = 3t^2\,\mathbf{i} + 2t\,\mathbf{j} + \mathbf{k}\) metres per second, and at \(t = 0\) it lifted off from the origin. Where is it at \(t = 2\)? Integrate the velocity: \(\mathbf{r}(t) = t^3\,\mathbf{i} + t^2\,\mathbf{j} + t\,\mathbf{k} + \mathbf{C}\). The initial condition \(\mathbf{r}(0) = \mathbf{0}\) forces \(\mathbf{C} = \mathbf{0}\), so \(\mathbf{r}(2) = 8\mathbf{i} + 4\mathbf{j} + 2\mathbf{k}\), and its distance from the pad is \(\sqrt{64 + 16 + 4} = \sqrt{84} \approx 9.2\) metres. The constant vector is not decoration: had the drone started 5 metres up, the same velocity would end 5 metres higher, and only \(\mathbf{C}\) records that.

Common mistakes to avoid

Differentiating the magnitude \(\|\mathbf{r}\|\) instead of the vector itself loses all direction information; \(\|\mathbf{r}\|\) is a scalar whose derivative is not \(\|\mathbf{r}'\|\) in either direction. Second, an indefinite integral of a vector function needs a constant vector, not a single scalar constant: three components were integrated, so three constants were created. Third, in the cross-product rule keep the order of each pairing; writing \(\mathbf{u}' \times \mathbf{r}\) for \(\mathbf{r}' \times \mathbf{u}\) flips a sign. Fourth, a smooth curve requires \(\mathbf{r}' \ne \mathbf{0}\), not merely that \(\mathbf{r}'\) exists: existence alone permits cusps where the motion halts.

Key takeaways

The derivative \(\mathbf{r}'(t) = \lim_{\Delta t \to 0}\frac{\mathbf{r}(t + \Delta t) - \mathbf{r}(t)}{\Delta t}\) exists and is computed componentwise, giving a vector tangent to the curve in the direction of travel; its magnitude is the speed. Second and higher derivatives repeat the process. A curve is smooth where \(\mathbf{r}'\) is continuous and nonzero; a zero derivative flags a cusp. The scalar, dot and cross products each satisfy a product rule, and from \(\frac{d}{dt}[\mathbf{r} \cdot \mathbf{r}] = 2\mathbf{r} \cdot \mathbf{r}'\) comes the circular-motion fact: constant length means velocity perpendicular to position. Integrals are componentwise, with a constant vector \(\mathbf{C}\) fixed by an initial condition.

Flashcards

Flashcard

Derivative definition

r′(t) = lim Δt→0 of [r(t + Δt) − r(t)]/Δt

Secant displacements turning into a tangent.

Test your understanding

Question 1 of 5

What is the derivative of r(t) = sin t i + t² j?

Question 2 of 5

What is r′(t) for r(t) = cos t i + sin t j + t k?

Question 3 of 5

What is d/dt of r(t) · u(t)?

Question 4 of 5

Evaluate ∫ (2t i + 3 j) dt.

Question 5 of 5

At which value of t does the curve r(t) = t² i + t³ j fail to be smooth?

Standards: 9 (BC)

← Previous topicNext topic →