A Force That Only Notices You If You're Moving

For more than two thousand years, electricity and magnetism were studied as two entirely separate subjects, until a single, almost accidental observation in the summer of 1820 revealed they were intimately connected. During a lecture demonstration, the Danish physicist Hans Christian Oersted noticed that a compass needle placed near a wire deflected the moment current flowed through that wire, aligning itself tangent to an imaginary circle centred on the wire; reverse the current's direction, and the needle's deflection reversed too. Oersted's simple, careful observation led to a sweeping conclusion: moving charges, or currents, produce a magnetic field in the surrounding space, exactly the way static charges produce an electric field. Just as an electric field E exerts a force qE on a charge q placed in it, a magnetic field B exerts a force on a moving charge, though this force behaves quite differently from its electric counterpart. The complete force on a charge q, moving with velocity v, in the presence of both an electric field E and a magnetic field B, is called the Lorentz force: F = q[E(r) + v x B(r)]. The magnetic part of this force, q(v x B), has three distinctive features worth holding onto. It depends on the charge's velocity, not just its position, so a charge at rest feels no magnetic force at all, however strong the field. Being a cross product, it vanishes whenever velocity and field are parallel or antiparallel, and otherwise always acts sideways, exactly perpendicular to both v and B at once, its precise direction given by the right-hand rule. And because this magnetic force is always perpendicular to the velocity, it can never do any work on the charge, or change its speed; it can only ever change the DIRECTION of motion, never the charge's kinetic energy. This last fact, force without work, shapes everything else in this chapter. The same reasoning extends naturally from a single moving charge to a whole current-carrying wire, since a current is nothing but many charges drifting together: a straight rod of length l carrying current I in an external field B experiences a force F = I(l x B), where l is a vector pointing along the direction of the current, with magnitude equal to the rod's own length.

Since the magnetic force never does work, a charge moving through a uniform magnetic field keeps a constant SPEED throughout, however curved its path becomes; only the direction of its velocity changes, continuously. Consider first the simplest case, a charge moving exactly perpendicular to a uniform field B: the magnetic force, always perpendicular to velocity, now points constantly toward one fixed centre, playing exactly the role of a centripetal force, and the charge is forced into a perfect circle. Equating the magnetic force qvB to the centripetal force mv²/r gives the radius of this circle, r = mv/qB, showing that a more energetic (faster or heavier) charge sweeps a wider circle for the same field and charge. The angular frequency of this circular motion, w = qB/m, called the cyclotron frequency, has a genuinely remarkable property: it does not depend on the charge's speed or the radius of its circle at all, only on the charge, mass, and field strength; a slow charge and a fast charge of the same type, in the same field, complete one full circle in exactly the same time, the fast one simply tracing a proportionally bigger circle to do it. This speed-independence is the whole reason a machine called the cyclotron can work, repeatedly accelerating a charged particle using a field that only needs to alternate at one single, fixed frequency, regardless of how fast the particle has already become. If the charge's velocity is not purely perpendicular to B, but has some additional component running parallel to the field, that parallel component is entirely unaffected by the magnetic force, and continues on unchanged, while the perpendicular component still traces its circle as before; the combined motion is a helix, a corkscrew path that circles around the field direction while steadily advancing along it, covering a fixed distance, called the pitch, with every complete turn.

Just as Coulomb's law gives the electric field due to a point charge, the Biot-Savart law gives the magnetic field due to a small current element, the basic building block from which every magnetic field, of any shape of wire, can in principle be assembled. Consider a tiny element of a current-carrying wire, of length dl, carrying current I; the magnetic field dB it produces at a point P, a displacement r away, is proportional to the current, to the element's own length, and inversely proportional to the square of the distance, exactly like Coulomb's law so far. But unlike Coulomb's law, direction here is not simply along the line joining source and point; instead, dB = (m0/4p) x I(dl x r)/r³, a genuine cross product, meaning the field is always perpendicular to the plane containing both the current element and the displacement vector r, and vanishes entirely along the current element's own direction (where the angle between dl and r is zero). The constant m0/4p, exactly 10⁻⁷ tesla-metre/ampere by definition, introduces m0, the permeability of free space, magnetism's direct counterpart to electrostatics' e0. In fact, e0 and m0 are not independent constants at all; their product is fixed by the speed of light itself, e0m0 = 1/c², a genuinely deep connection that becomes central once electromagnetic waves are studied. The Biot-Savart law and Coulomb's law share the same inverse-square dependence and the same superposition principle, but differ in a few fundamental ways: Coulomb's law's source is a scalar, electric charge, while the Biot-Savart law's source is inherently a vector, the current element Idl; and while an electric field points directly along the displacement vector from its source, a magnetic field is always perpendicular to it.

Putting the Biot-Savart law to work on a genuinely useful shape, a circular loop of radius R carrying a steady current I, reveals a field along the loop's own axis that has real structural similarities to the field of an electric dipole. Each small current element around the loop contributes a field dB perpendicular to the plane containing itself and the displacement vector to the axial point; by the loop's symmetry, every component of these contributions running perpendicular to the axis exactly cancels against the diametrically opposite element, leaving only the components running along the axis to survive and add. Carrying out this sum around the full loop gives the field at a point on the axis at distance x from the centre: B = m0IR²/2(x²+R²)^(3/2), a vector directed straight along the axis. Setting x=0 gives the field exactly at the loop's own centre, a clean, simple result: B0 = m0I/2R. Far from the loop, at distances x much larger than R, this expression simplifies to B = m0IA/2px³, where A=pR² is the loop's own area; strikingly, this is essentially identical in form to the far-field of an electric dipole, provided m0 stands in for 1/e0 and the loop's magnetic moment m = IA stands in for the electric dipole moment. This is not a coincidence: at large distances, a current loop behaves exactly like a tiny magnetic dipole. There is, however, one profound difference underneath this similarity: an electric dipole is genuinely built from two separate elementary pieces, positive and negative charges, which can in principle exist entirely on their own; a magnetic dipole cannot be taken apart this way -- no isolated magnetic monopole, the magnetic equivalent of a lone electric charge, has ever been found. The current loop itself, not some pair of hidden magnetic charges, is the truly elementary magnetic object.

The Biot-Savart law, while completely general, can involve genuinely difficult integration for anything but the simplest shapes; Ampere's circuital law offers an alternative, exactly the same relationship between current and field but expressed differently, and it becomes a genuine shortcut whenever a current distribution has enough symmetry, in precisely the same way Gauss's law is a shortcut for electric fields with the right symmetry. Ampere's law states that the closed-loop integral of B.dl, taken all the way around any closed path (an Amperian loop), equals m0 times the total current I passing through that loop. Whenever a loop can be chosen along which B is either tangential and constant in magnitude, normal, or exactly zero, this reduces to simple algebra, BL = m0 Ie, where L is the length of the loop along which B is tangential, and Ie is the current actually enclosed. Applied to a long, straight, current-carrying wire, choosing a circular Amperian loop of radius r centred on the wire (where B is tangential and, by symmetry, the same size all the way around) gives B(2pr) = m0I, so B = m0I/2pr: the field circles the wire, falling off as 1/r, and forms closed loops rather than radiating outward the way an electric field does, a genuinely fundamental difference between electric and magnetic field lines that traces back to the non-existence of magnetic monopoles. Ampere's law is not new physics beyond the Biot-Savart law; it can be derived from it, exactly as Gauss's law can be derived from Coulomb's law -- it is simply a different, often more convenient, way of expressing the very same relationship between current and field, genuinely useful whenever symmetry cooperates, but not always able to extract a result on its own even when it strictly still holds true, since knowing an integral's total value does not always reveal the function's value at every point unless symmetry guarantees that value stays constant around the chosen loop.

A solenoid, a long wire wound into a tightly-spaced helix, is one of the most useful applications of Ampere's law, precisely because it is the standard, practical way to produce a strong, remarkably uniform magnetic field over an extended region, something a single loop or a straight wire cannot easily do. Picture the solenoid stretched long enough that its length is much greater than its radius: between any two neighbouring turns, the fields of adjacent loops act to cancel each other almost entirely, while outside the solenoid altogether, the field falls close enough to zero to treat as exactly zero in the ideal case; inside, however, the field builds up, becoming strong, uniform, and directed straight along the solenoid's own axis. Applying Ampere's law with a rectangular Amperian loop, drawn with one side inside the solenoid, parallel to the axis, and the opposite side outside it (where B is taken to be zero), only the inside segment contributes to the tangential integral: B x h = m0 x (n h) x I, where n is the number of turns per unit length and h is the chosen length of that inside segment, giving simply B = m0 n I, remarkably independent of the solenoid's own radius, or of exactly where inside the field is measured, as long as it is comfortably away from the two ends. This clean, controllable relationship, more turns per unit length or more current both directly increase the field, makes the solenoid the standard tool whenever a real experiment or device needs a dependable, uniform magnetic field, from laboratory equipment to the electromagnets used in MRI scanners and particle accelerators; a soft iron core slipped inside the solenoid, as the next chapter explores, can boost that field considerably further still.

Since a current-carrying wire both produces a magnetic field and, separately, feels a force from an external magnetic field, it follows immediately that two current-carrying wires placed near each other must exert forces on one another: the field from one acts on the current in the other. Consider two long, straight, parallel wires, separated by distance d, carrying currents Ia and Ib. Wire a produces a field at wire b's location of magnitude Ba = m0Ia/2pd; this field then exerts a force on wire b's own current, of magnitude Fba = IbLBa = m0IaIbL/2pd on a length L of wire b, directed straight toward wire a. By Newton's third law, an equal and opposite force acts on wire a due to wire b, and this is borne out exactly by the Biot-Savart law and Lorentz force together, a reassuring, non-obvious consistency check. Working out the direction carefully reveals a rule that is the precise opposite of the electrostatic case: parallel currents, flowing the same way, attract each other, while antiparallel currents, flowing opposite ways, repel; recall that in electrostatics, it is exactly the opposite, LIKE charges repel and unlike charges attract. This force per unit length between two parallel wires, fba = m0IaIb/2pd, is genuinely fundamental enough that it defines the SI unit of current itself: one ampere is the steady current which, flowing in each of two infinitely long, straight, parallel wires exactly one metre apart in vacuum, produces a force of exactly 2x10⁻⁷ newtons on every metre of each wire's length. The coulomb, in turn, is defined from the ampere: the charge that flows past a point in one second, when a steady current of one ampere is maintained.

Place a rectangular current loop inside a uniform magnetic field, and something genuinely elegant happens: the NET force on the whole loop works out to exactly zero, yet the loop still experiences a real, non-zero torque, tending to rotate it. Consider the simplest case, where the field B lies in the same plane as the loop. Two of the loop's four sides feel no force at all, since current and field run parallel along them; the other two sides, each of length b, feel equal and opposite forces of magnitude IbB, but since these two forces act along two DIFFERENT parallel lines, separated by the loop's other side-length a, they form a couple, a torque of magnitude t = IbB x a = IAB, where A=ab is the loop's own area. In general, defining theta as the angle between the field B and the NORMAL to the loop's plane (this simplest case, field lying in the loop's plane, corresponds to theta=90 degrees), the torque becomes t = IAB sinq, reaching its maximum, IAB, when the normal is perpendicular to the field (theta=90°, the field lies in the loop's plane) and vanishing entirely when the normal is parallel to the field (theta=0°, the field is perpendicular to the loop's plane). Defining the loop's magnetic moment as a vector m = IA, of magnitude IA and direction given by the right-hand thumb rule (curl the fingers along the current's direction; the thumb gives the direction of m, along the normal), both cases combine into one single, compact vector equation: t = m x B, precisely analogous to the torque on an electric dipole, t = p x E. This torque vanishes only when m is exactly parallel or exactly antiparallel to B, marking two equilibrium orientations: parallel is a STABLE equilibrium, since any small rotation away from it produces a restoring torque pulling the loop back; antiparallel is an UNSTABLE equilibrium, since any small rotation away from it produces a torque that pushes the loop further away still. This is exactly why a small compass needle, itself a magnetic dipole, reliably swings around to align with a magnetic field rather than sitting at any other angle, and why, given N tightly-wound turns rather than a single loop, the magnetic moment simply scales up to m = NIA.

The torque a current loop feels in a magnetic field is not just a theoretical curiosity, it is the exact working principle behind the moving coil galvanometer, the standard instrument for detecting and measuring small currents and voltages. A coil of N turns and area A, free to rotate, sits inside a carefully shaped RADIAL magnetic field (arranged, with the help of a soft iron core, so the field is always exactly along the coil's own plane, keeping sinq = 1 throughout the coil's swing), and a spring provides a restoring torque, kf, proportional to the twist angle f, that grows to exactly balance the magnetic torque NIAB at some final, steady deflection: kf = NIAB, or f = (NAB/k) x I. Since the bracketed quantity is fixed for any one particular galvanometer, the deflection f is directly proportional to the current I, letting a calibrated scale read off current directly. A bare galvanometer, however, cannot be used directly as an ammeter: it is far too sensitive, giving a full deflection for currents of only a few microamps, and its own resistance is large enough to noticeably disturb the very current it is trying to measure. The fix is a shunt resistance, a small resistor rs connected in PARALLEL with the coil, diverting most of the current around the galvanometer and leaving the combination's overall resistance small enough not to disturb the circuit. To use the same device as a voltmeter instead, the opposite fix is needed: a large resistance R connected in SERIES with the coil, ensuring the combination draws only a very small current and disturbs the voltage being measured as little as possible. A subtlety worth noticing: doubling a galvanometer's number of turns N doubles its CURRENT sensitivity, f/I = NAB/k, but since more turns also roughly double the coil's own resistance, and voltage sensitivity is f/V = (NAB/k)(1/R), the VOLTAGE sensitivity can end up entirely unchanged, a reminder that current sensitivity and voltage sensitivity are genuinely different properties, and improving one does not automatically improve the other.