A Hill for Charge, and a Jar to Trap It

Lift a stone up a hill, and however winding the path taken, the work done against gravity depends only on how high the stone finally ends up, never on the route: this is what makes gravity a conservative force, and it is what allows the entire messy business of forces and paths to be replaced with a single, simpler number at every point, gravitational potential energy, depending only on position. The electrostatic force, remarkably, shares this exact same property: it too is conservative, which means the work done in moving a charge from one point to another against an electric field also depends only on the two points, never on the path taken between them. This fact licenses the same trick used for gravity: define an electrostatic potential energy, and read off the work done as simply the difference in this quantity between the start and end points. Potential itself is a slightly different, closely related idea: it strips away the specific size of the charge being moved, and asks only about the work done per unit charge. The electrostatic potential at a point P is defined as the work done, by an external agency, in bringing a small unit positive test charge from infinity (where the potential is defined to be zero) to the point P, with no change in its kinetic energy along the way. Because the reference point, infinity, is a free choice, the potential at any single point does not carry an absolute meaning by itself, and an arbitrary constant could technically be added everywhere without changing any physics; what is genuinely, physically significant is always the potential DIFFERENCE between two points, V(P) - V(R), since that difference is exactly the work done per unit charge in moving between those two specific points, regardless of which reference was chosen.

For a single point charge Q, integrating the field along a radial path from infinity to a point at distance r gives a remarkably clean result: V(r) = Q/4πε₀r, positive if Q is positive, negative if Q is negative, falling off as 1/r, one power of distance slower than the field itself, which falls off as 1/r². For an electric dipole of moment p, the potential at a point far from the dipole (at distance r much greater than the dipole's own size) works out to V(r) = p.r̂/4πε₀r², where r̂ is the unit vector toward that point; this expression depends on direction, unlike a point charge's potential, since it matters whether the point lies along the dipole's axis, in its equatorial plane, or somewhere in between (the equatorial value comes out to exactly zero, since points there are equidistant from both charges of the dipole, which then exactly cancel). Notice something easy to mix up: a dipole's POTENTIAL falls off as 1/r², one power faster than a single charge's potential, while a dipole's FIELD falls off as 1/r³, one power faster still; potential and field are related but distinct quantities, and they do not share the same fall-off rate. For any system of several point charges, the potential at a point P is simply the sum of the potentials due to each charge individually, treating each as if the others were not there: V = (1/4πε₀)(q1/r1P + q2/r2P + ... + qn/rnP). This is the same superposition principle used for forces and fields, but with a genuine practical advantage: potential is a scalar, a plain number with no direction, so this sum is ordinary arithmetic, none of the angle-tracking and component-splitting that vector addition of fields demands. Applying this to a uniformly charged spherical shell of total charge q and radius R gives a particularly instructive result. Outside the shell, at any r ≥ R, the potential is exactly V = q/4πε₀r, identical to a point charge q sitting at the shell's centre. Inside the shell, at any r < R, the field is zero everywhere, as already established, but the potential is NOT zero; it stays constant throughout the interior, equal to the value right at the surface, q/4πε₀R. This is not a contradiction: a zero field simply means the potential is not changing from point to point inside, not that its value must itself be zero, in exactly the same way a perfectly flat plateau at the top of a hill has zero slope everywhere on it without needing to be at sea level.

A surface over which the potential has exactly the same value at every point is called an equipotential surface. For an isolated point charge, these surfaces are concentric spheres centred on the charge, since potential depends only on distance r; for a uniform field, they are flat planes perpendicular to the field direction. Two properties of equipotential surfaces follow directly from what potential already means. First, moving a charge between any two points on the same equipotential surface takes zero work, since work done is q times the potential difference between the two points, and that difference is zero by definition of the surface. Second, and less obvious: the electric field at any point must be exactly perpendicular to the equipotential surface passing through that point. If the field had any component lying along the surface itself, that component would do work moving a charge along the surface, which was just shown to be impossible; the only way to guarantee zero work in every direction along the surface is for the field to have no component along it at all, leaving only a component perpendicular to it. This perpendicularity gives a genuinely practical way to connect the two central ideas of this chapter, field and potential: E = -dV/dl, the field in any direction equals the negative rate at which potential changes with distance in that direction. The minus sign matters and has a clear physical meaning: the field points in the direction of the steepest DECREASE in potential, from high potential toward low, exactly the way a ball placed on a slope rolls toward lower height, not higher. Closely spaced equipotential surfaces mean potential is changing rapidly over a short distance, which by this relation means a strong field there; widely spaced equipotential surfaces mean a weak field, the electrical equivalent of reading a contour map, where closely packed contour lines mean a steep slope and widely spaced ones mean nearly flat ground.

Assembling a group of charges from far apart into their final positions takes work, and that work is stored as the potential energy of the system, the same idea already used for a single charge, now applied to a whole configuration. Bringing a single charge q1 from infinity to some point costs no work, since there is no other charge yet to push against; bringing a second charge q2 from infinity to a distance r12 from q1, however, means working against q1's field the entire way, and that work comes out to exactly U = q1q2/4πε₀r12, positive (work must be done) if the charges share a sign, negative (work is released, or would need to be done to pull them apart) if they have opposite signs. For three or more charges, the total potential energy is the sum over every distinct pair, calculated exactly as if each pair existed alone; a genuinely reassuring feature of this result is that it comes out the same regardless of the order in which the charges are imagined to be brought in, since the electrostatic force is conservative. A closely related but distinct idea is the potential energy of a single charge q sitting in a field produced entirely by OTHER, external charges: U = qV(r), where V(r) is the potential due to those external charges alone at the point where q sits, deliberately excluding any contribution from q's own field, which would make the expression meaningless (a charge's potential at its own location is undefined, infinite). At the very small scales where atoms and nuclei live, the joule is an inconveniently large unit for energy, so physicists commonly use the electron volt (eV) instead: the energy gained by a single electron accelerating through a potential difference of exactly one volt, equal to 1.6 x 10⁻¹⁹ J. Finally, an electric dipole of moment p, placed at angle θ to a uniform external field E, has potential energy U(θ) = -pE cosθ; by convention, the reference angle is chosen at θ = 90 degrees, where U is defined to be zero, the natural mirror of choosing potential itself to be zero at infinity. This expression immediately reveals the dipole's two special orientations: at θ=0, aligned with the field, U reaches its most negative, most stable value; at θ=180 degrees, pointing directly against the field, U reaches its most positive, least stable value, an unstable equilibrium that the slightest disturbance tips out of.

Inside the actual conducting material of a conductor in electrostatic equilibrium, the free charge carriers, electrons in a metal, are completely free to move; if there were any electric field inside, they would keep moving in response to it, which by definition means equilibrium has not yet been reached. It follows immediately that, in equilibrium, the field inside a conductor's own material must be exactly zero everywhere. From this single starting fact, four further properties follow, together giving a complete picture. Since the field inside is zero and potential is related to field by E=-dV/dl, the potential must be exactly constant throughout the interior, and by continuity, equal to that same constant value on the surface too; a conductor is, in this sense, always an equipotential body, inside and out. Any excess charge placed on a conductor must reside entirely on its outer surface, none of it in the interior, a direct consequence of Gauss's law applied to a Gaussian surface drawn just inside the conductor's own material (enclosing zero field means enclosing zero net charge). Just outside the surface, the field is not zero, but it must point exactly along the outward normal to the surface, with no component running along the surface itself, since any such component would again mean charges on the surface still experience a force and are not yet in equilibrium. And the magnitude of that just-outside field is directly tied to the local surface charge density by E=σ/ε₀, a result obtained using a small pillbox-shaped Gaussian surface straddling the surface. A remarkable extension of these ideas is electrostatic shielding: if a conductor has a hollow cavity inside it, with no charges placed in the cavity itself, the field inside that cavity is exactly zero, regardless of the conductor's own shape, regardless of how much charge is on it, and regardless of whatever external fields exist outside the conductor entirely. This is precisely why sensitive electronic equipment is often housed inside a metal enclosure, a Faraday cage, to shield it from outside electrical interference, and it is part of why a bird can perch safely on a bare high-voltage wire (its body never spans a potential difference) while a person standing on the ground, touching the very same wire, completes a path between two very different potentials and receives a genuinely dangerous shock.

Dielectrics, unlike conductors, are non-conducting substances with no free charge carriers able to drift through the material; whatever charge exists is bound to individual molecules and cannot simply flow away. Placed in an external field, a dielectric responds very differently from a conductor. Some molecules are non-polar to begin with, meaning the centres of their positive and negative charge already coincide exactly, giving zero permanent dipole moment (oxygen and hydrogen molecules are common examples); an external field stretches these molecules apart very slightly, inducing a small dipole moment in the direction of the field. Other molecules are polar, meaning their positive and negative charge centres are permanently offset even with no external field present (water is the standard example); with no field applied, these permanent dipoles point in random directions due to thermal agitation, cancelling out to zero net effect, but an external field causes them to line up, at least partially, producing a net alignment in the field's direction. Either way, whether by inducing new dipoles or aligning existing ones, the dielectric develops an overall dipole moment per unit volume, called polarisation and denoted P; for the common case of a linear, isotropic dielectric, polarisation is directly proportional to the applied field, P = ε₀χₑE, where χₑ, the electric susceptibility, is a constant characteristic of the particular material. The consequence, at the level of a dielectric slab placed between two charged plates, is that the polarised dielectric develops a net bound surface charge on the two faces normal to the field, positive on one face, negative on the other, even though no net charge appears anywhere inside its volume (adjacent molecular dipoles' charges cancel everywhere except right at the two outer faces). This induced surface charge produces its own field, opposing the external field, exactly parallel to how induced charge behaves in a conductor -- but with one crucial difference: in a conductor, the induced charge's opposing field exactly cancels the external field inside, reducing it all the way to zero, while in a dielectric, the opposing field only partially cancels the external field, weakening it but never reducing it all the way to zero, since a dielectric's bound charges, unlike a conductor's free charges, cannot move far enough to fully neutralise the field.

A capacitor is nothing more exotic than a system of two conductors, separated by an insulator, carrying equal and opposite charges +Q and -Q with some potential difference V between them (a single conductor can even be treated as a capacitor, by imagining the second conductor pushed all the way out to infinity). Since the field between the two conductors is directly proportional to Q, and V is the work done per unit charge moving between them, V is proportional to Q as well, which means their ratio, Q/V, is a genuine constant for a given capacitor, called its capacitance: C = Q/V. Capacitance depends only on the geometry, shape, size and separation of the two conductors (and, as the next section shows, on whatever insulating material fills the gap between them) -- never on Q or V themselves, exactly the way a cup's volume does not depend on how much water is actually poured into it. The practical value of a large capacitance is direct: for a fixed amount of stored charge Q, a bigger C means a smaller V is needed, and since very high potential differences create very strong fields capable of ionising the surrounding air and letting the stored charge leak away (a dielectric's maximum safe field is called its dielectric strength), a capacitor built to store real charge without leaking needs enough capacitance to keep its working voltage comfortably below that breakdown limit. In practice, the farad, the SI unit of capacitance, turns out to be an enormous unit: a parallel plate capacitor with plates just one square metre in area and one millimetre apart works out to less than ten billionths of a farad, and reaching a full farad with a 1 cm gap would require plates roughly 30 kilometres on a side, which is exactly why real capacitors are almost always specified in microfarads, nanofarads or picofarads instead. The simplest capacitor to actually build and analyse is the parallel plate capacitor: two identical flat conducting plates of area A, separated by a small distance d, carrying charges +Q and -Q. Treating each plate as an infinite charged sheet (a good approximation when d is much smaller than the plates' own size) and adding the two plates' fields by superposition shows the field is exactly zero outside the plates entirely, and a uniform E = σ/ε₀ = Q/ε₀A in the region between them. Since this field is uniform, the potential difference is simply V = Ed = Qd/ε₀A, giving a capacitance of C₀ = ε₀A/d for vacuum (or, to a very close approximation, air) between the plates -- depending, exactly as promised, only on the geometry of the two plates.

Slide a slab of dielectric material into the gap of a parallel plate capacitor, filling it completely, and the polarisation effects of the previous section come into play directly. The dielectric develops its own induced surface charge density, σp, on the two faces touching the plates, opposing the field set up by the plates' own free charge, σ. The net field inside the dielectric is therefore reduced, to E = (σ - σp)/ε₀, rather than the full σ/ε₀ that vacuum alone would give. For a linear dielectric, this reduction turns out to be a clean, constant factor: σ - σp = σ/K, where K, always greater than 1, is a number characteristic of the particular dielectric material, called its dielectric constant. Carrying this reduced field through the same steps as before gives a new capacitance, C = ε₀KA/d = KC₀, exactly K times larger than the same capacitor's capacitance with vacuum (or air) alone. This single relationship, C = KC₀, is actually the cleanest working definition of dielectric constant: it is simply the factor by which a material increases a capacitor's capacitance when it completely fills the gap between the plates, and while it was derived here specifically for a parallel plate capacitor, the result holds for capacitors of any shape at all. It is also useful to define the permittivity of the medium itself, ε = ε₀K, so that K, the dielectric constant, is just the ratio ε/ε₀, the medium's permittivity relative to that of a vacuum -- for vacuum itself, K is exactly 1, since there is no dielectric material there at all to polarise. Materials with a high dielectric constant, such as mica or various ceramics, are exactly what capacitor manufacturers reach for when they need to pack a large capacitance into a small physical size, since K multiplies the capacitance directly without requiring larger plates or a narrower gap, both of which come with their own practical limits.

Real circuits often need a capacitance value that no single off-the-shelf capacitor provides, so capacitors get wired together, and the two simplest ways of combining them, series and parallel, behave in a way that surprises almost everyone meeting it for the first time: the rules are exactly swapped compared to resistors. Capacitors in series (one after another, in a single chain) are forced, by charge conservation on the connecting plates between them, to carry the exact same charge Q on every capacitor in the chain, while the total potential difference is the sum of each individual capacitor's own voltage drop, V=V1+V2+...; working through the algebra gives 1/C = 1/C1 + 1/C2 + ..., meaning the combined capacitance is always smaller than the smallest individual capacitor in the chain, precisely mirroring how resistors behave in PARALLEL. Capacitors in parallel (side by side, sharing the same two connection points) instead share the exact same voltage V across every one of them, while the total charge is the sum of each individual capacitor's own stored charge, Q=Q1+Q2+...; this gives simply C=C1+C2+..., meaning the combined capacitance is always larger than the largest individual capacitor, mirroring how resistors behave in SERIES. The energy stored while charging any capacitor can be found by imagining the charging process happening gradually, one tiny increment of charge dQ' at a time, moved from one plate to the other against an ever-increasing potential difference V'=Q'/C; the small amount of work needed for each increment is dW=V'dQ', and adding up (integrating) all these small contributions from an initially uncharged capacitor up to its final charge Q gives a total stored energy of U = Q²/2C, which can equally well be written as U = CV²/2 or U = QV/2, three algebraically equivalent forms of the exact same number. This stored energy can be pictured as living in the capacitor's own field itself, filling the volume between its plates, with an energy density (energy stored per unit volume) of u = ε₀E²/2 -- a result derived here for a parallel plate capacitor specifically, but one that turns out to hold generally, for the electric field due to any configuration of charges whatsoever, a genuinely deep hint that the field itself, not just the charges that created it, is where real physical energy resides. This also finally answers a puzzle worth sitting with: connect a charged capacitor to an identical uncharged one, and even though total charge is conserved perfectly, the combined final energy comes out to only half the original energy, not the full amount -- the missing half was never destroyed, it was simply carried away as heat and electromagnetic radiation during the brief, transient rush of current as the charge redistributed itself between the two capacitors.