Conservative Forces and Potential Energy

Picture a marble rolling around the inside of a smooth bowl. Let it go from the rim and it swings down, races through the bottom, climbs the far side to exactly the height you dropped it from, and comes back — over and over, never quite escaping, never quite dying away. Now tip a ball into a mountain valley: it rolls down into the hollow and, given no friction, would oscillate between the two hillsides forever. In both cases the shape of the landscape alone tells you where the object speeds up, where it slows, how far it can climb, and where it will finally come to rest. You never had to track the force step by step.

That is the astonishing gift of a conservative force. When a force is conservative, all of its bookkeeping can be packed into a single scalar map of the terrain — a potential energy U — and the motion becomes a story of a ball sliding on that map. This page builds the idea from the ground up: what makes a force conservative, how the map U is born from it, how energy is conserved, and how to read an object's entire fate straight off the curve of U(x).

What "conservative" really means: the path doesn't matter

Recall from work and line integrals that the work a force \vec F does as an object moves along a path C is the line integral

W = \int_{C} \vec F \cdot d\vec r.

For most forces this depends on the whole route: take a longer or wigglier path and you get a different number. A force is conservative when it does not — when the work depends only on where you start and where you end, and not one bit on the path taken in between. Climb a hill by the gentle switchback or the sheer cliff face: gravity does the same work against you either way, because only your change in height counts.

There are three equivalent ways to say "path-independent," and it is worth seeing that they are the same statement wearing three hats:

Path independence. The work \int_C \vec F\cdot d\vec r between any two points is the same for every path joining them.
Zero work around a loop. Equivalently, the work done around any closed loop vanishes: \oint \vec F \cdot d\vec r = 0.
Curl-free field. Equivalently (for a field defined on a simply-connected region), the force has no swirl: \nabla \times \vec F = \vec 0.

Why are the first two the same? Take two paths from A to B. Go out along one and back along the other: that is a closed loop. If the round trip does zero net work, the two one-way trips must have done equal work — path independence. Run the argument backwards and you recover the loop test. The curl condition is the local, point-by-point version of the loop condition, and it is usually the quickest way to check whether a given force field is conservative.

Where potential energy comes from

Here is the pay-off. Because a conservative force's work depends only on the endpoints, we can pin a single number to each point in space and let the work between two points be just the difference of those numbers. That number is the potential energy U. Define it so that the work done by the force as the object goes from A to B is the drop in potential:

W_{A\to B} = U(A) - U(B) = -\Delta U.

A force can only be summarised this way because it is conservative — for a path-dependent force like friction there is no consistent number to assign, so no potential energy exists. Reading the relationship the other way round, in one dimension where W = \int F\,dx = -\Delta U, the force is the negative slope of the potential:

F(x) = -\frac{dU}{dx}.

In three dimensions the slope becomes the gradient, which points in the direction of steepest increase of U. The force points the opposite way — downhill, toward lower potential energy:

A conservative force is the negative gradient of its potential energy: \vec F = -\nabla U = -\left(\frac{\partial U}{\partial x},\, \frac{\partial U}{\partial y},\, \frac{\partial U}{\partial z}\right).
The minus sign means the force always pushes the object toward lower potential energy — a ball rolls downhill, a stretched spring pulls back toward its rest length.
In one dimension this collapses to F = -\dfrac{dU}{dx}: the force is minus the slope of the U(x) curve.

Two potentials you will meet everywhere

Gravity near the ground. Close to Earth the gravitational force on a mass m is F = -mg (pointing down, taking up as positive h). Integrating F = -dU/dh gives the famous linear ramp

U(h) = mgh.

Its slope is dU/dh = mg, so F = -mg — the pull is constant and downward, exactly as it should be. Every metre you lift a 2\ \text{kg} book stores another 2 \times 9.8 \approx 19.6\ \text{J}.

A spring. A spring obeys Hooke's law F = -kx, pulling back toward its natural length with a force proportional to the stretch x. Integrating F = -dU/dx gives a parabolic well:

U(x) = \tfrac{1}{2}kx^{2}.

Check the slope: dU/dx = kx, so F = -kx. The deeper you push into either side of the parabola, the harder the spring shoves you back toward the bottom at x = 0.

Conservation of mechanical energy

Now the reason it is called a conservative force. Suppose the only force acting is conservative. The work–energy theorem says the work done equals the change in kinetic energy, W = \Delta KE. But we just saw W = -\Delta U. Setting them equal, \Delta KE = -\Delta U, i.e. \Delta(KE + U) = 0. The sum never changes:

E = KE + U = \tfrac{1}{2}mv^{2} + U(x) = \text{constant}.

This single conserved quantity E, the total mechanical energy, is the engine behind countless shortcuts. Drop the marble from rest at height h: at the top all its energy is potential, E = mgh; at the bottom all of it is kinetic, E = \tfrac{1}{2}mv^{2}. Equate them and the mass cancels:

\tfrac{1}{2}mv^{2} = mgh \quad\Longrightarrow\quad v = \sqrt{2gh}.

No forces resolved, no time integrated — energy in equals energy out, and the speed at the bottom falls straight out. That v = \sqrt{2gh} is the same whether the marble drops straight down or slides down a curved, frictionless ramp, precisely because gravity is conservative.

Reading motion off the landscape

Here is the whole payoff made visible. The blue curve is a potential-energy landscape U(x) = x^{4} - a\,x^{2} — a double well with two valleys and a hill between them. The orange line is the constant total energy E. Imagine a ball sliding on the blue curve: wherever the ball is, KE = E - U is the vertical gap between the orange line and the blue curve. Drag the sliders and watch the physics unfold.

Three things to notice, and they are the whole art of reading a potential curve:

Turning points. Where the orange E line crosses the blue curve, E = U, so KE = 0 — the ball momentarily stops and reverses. The ball is confined to the regions where E \ge U (the curve dips below the line); it can never enter a region where U > E, since that would demand negative kinetic energy.
Speed. The vertical gap E - U is the kinetic energy, so the ball moves fastest at the bottom of a well (biggest gap) and slows as it climbs toward a turning point.
Equilibria. Flat spots, where dU/dx = 0, are the places the force F = -dU/dx vanishes — equilibrium points. A minimum (a valley bottom) is a stable equilibrium: nudge the ball and the slope pushes it back. A maximum (the central hilltop) is unstable: the tiniest nudge and the slope sends it away.

Worked example: analysing a well

Take the double well with a = 2: U(x) = x^{4} - 2x^{2}. Let us find and classify every equilibrium, using nothing but calculus.

Step 1 — find the flat spots. Equilibria sit where the slope is zero:

\frac{dU}{dx} = 4x^{3} - 4x = 4x\,(x^{2} - 1) = 0 \quad\Longrightarrow\quad x = -1,\; 0,\; +1.

Step 2 — classify each with the second derivative. The sign of U'' = 12x^{2} - 4 tells us the concavity: positive means a minimum (stable), negative means a maximum (unstable).

At x = 0: U'' = -4 < 0 — a maximum, the central hilltop, an unstable equilibrium (U = 0 there).
At x = \pm 1: U'' = 12 - 4 = 8 > 0 — two minima, the valley bottoms, stable equilibria (U = 1 - 2 = -1 there).

Step 3 — read off the motion. A ball released with total energy E = -0.5 (below the central hilltop at U = 0) is trapped in whichever valley it starts in, oscillating between the two turning points where x^{4} - 2x^{2} = -0.5. Give it enough energy to clear the hill (E > 0) and it roams freely across both wells. That is the entire life story of the particle, extracted from one curve.

Watch out! — this is the classic trap. It is tempting to think every force gets its own stored energy, but friction (and air drag) are the textbook non-conservative forces: they have no potential energy, and mechanical energy is not conserved when they act.

The reason is exactly the path-dependence test. Slide a box across the floor from A to B in a straight line, then again by a long winding route: friction does far more negative work on the longer path, because it always opposes motion and the longer road has more motion to oppose. The work depends on the route, not just the endpoints — so \oint \vec F\cdot d\vec r \neq 0, and there is no consistent number U you could assign to each point. The "lost" mechanical energy hasn't vanished; it has turned into heat, which is a different bookkeeping. Only conservative forces earn a potential energy.

There isn't one — and that is fine. Potential energy has no absolute value; only differences in U have physical meaning. Look back at how it was defined: everything came through W = -\Delta U and F = -dU/dx, and both involve a change or a slope. Add any constant to U(x) everywhere and neither the force nor the work changes one bit.

That freedom is why you get to choose where U = 0. For gravity we usually set U = 0 at the ground and write U = mgh, but the floor of a valley, or sea level, or the ceiling would all do — the physics of the falling ball is identical. When you read a landscape, it is the shape of U(x) — where the hills and valleys are — that matters, never the height of the zero line.