FOUR MAGNETS, ONE PHONE, THREE LAWS OF NATURE
Far from a small magnet, its field weakens with distance as 1/d³: double the distance and the field drops eightfold. That is the field of a dipole, a standard experiment in introductory physics labs. But physics does not stop there. The field of any compact source can be written as a sum of ever finer terms — dipole, quadrupole, octupole — and each successive term fades one power of distance faster: 1/d³, then 1/d⁴, then 1/d⁵.
This “multipole expansion” is everywhere. The reference model of the Earth’s magnetic field starts with a dipole, a quadrupole and an octupole. The magnets that steer particle beams in accelerators are named dipoles, quadrupoles, sextupoles and octupoles. Tuning the magnets of MRI scanners means cancelling unwanted multipole terms one by one. Yet students almost never measure anything beyond the dipole: higher-order fields are weak, die away fast, and the probes needed to map them are expensive and fiddly.
Cancel the dipole, reveal the next term
Since there are no magnetic monopoles, the leading term of any magnet is its dipole. To see what comes next, you must cancel it. Join two identical magnets with opposite orientations, and their dipoles cancel: what is left is a quadrupole. Join two quadrupoles in opposition, and the quadrupole terms cancel too: an octupole remains.
Martín Monteiro, Cecilia Stari and Arturo Martí in Montevideo, with Enrique Arribas, Isabel Escobar, Raquel Ramirez-Vazquez and Augusto Beléndez in Spain, show that four identical neodymium cube magnets — 6.35 millimetres on a side — are enough to build not one but three different octupoles: four magnets stacked in a line, four in a row side by side, and a 2 × 2 square. With the same magnets, they also build a dipole and two quadrupoles as references.
A simple set of formulae, based on where each magnet sits, predicts which component of the field survives at each measuring point, and how fast it decays. One case catches students out: four magnets side by side with alternating poles — plus, minus, plus, minus — look like an octupole, but are not: the real side-by-side octupole runs plus, minus, minus, plus (plus and minus giving the direction each magnet points), two quadrupoles in opposition. Their dipole cancels, their quadrupole does not, and their field fades only as 1/d⁴.
The sensor in your pocket
The measuring instrument is a smartphone. Many phones contain a three-axis magnetometer, readable with free apps; here, a Samsung Galaxy S10e with the phyphox app. The phone lies flat on graph paper, with a line drawn through its sensor. The first step is to find where the chip is: slide a small magnet over the screen and watch for the highest reading. The Earth’s field and the phone’s own field are recorded first, then subtracted.
The magnets are then moved along the line, from about 3 to 16 centimetres from the sensor, and each reading averaged over five seconds.
To check the numbers, the team measured the strength of the magnets independently. Four cubes stacked into a bar, swinging freely in the Earth’s horizontal field — 21.5 microteslas, measured with the same phone — oscillate with a period of 0.89 seconds — the stronger the magnet, the faster it swings. That gives a magnetic moment, the standard measure of a magnet’s strength, of 0.252 ampere-square-metres per cube.
Three, four, five
For a single magnet, the field follows the 1/d³ law over a fourfold range of distances, with fitted exponents of 2.948 and 2.994. On the magnet’s axis, the fitted strength — the field multiplied by the cube of the distance, a number that stays constant if the 1/d³ law holds — is 50.95 nanotesla-cubic-metres, against 50 predicted from the measured moment; to the side, 24.64 against 25. The ratio of the two, 2.07, reproduces the factor of two between the two positions that theory predicts.
The four quadrupole measurements give exponents between 3.80 and 4.18, clustered around 4. The octupoles give between 4.66 and 5.2, around 5, with deviations of up to 0.34. Their strength factors vary tenfold from one arrangement to another, although the same four magnets are used each time: the octupole is very sensitive to the geometry. Inserting a spacer two magnets wide in one arrangement roughly doubles its field, as the formula predicts.
Why the octupole is hard
The octupole is squeezed from both sides. Too close, the next, finer terms of the series still matter — up to 28% at 3 centimetres in one arrangement. Too far, any small imperfection leaves a slower-fading residue that ends up dominating: one magnet 5% stronger than the others leaves a dipole, one shifted by half a millimetre a quadrupole. The useful window is narrow, and beyond 10 to 12 centimetres the octupole sinks into the sensor’s noise. The position of the chip inside the phone, known to within about 2 millimetres, adds its own uncertainty.
The authors propose the activity as a single three-hour lab session or a short project: dipole first, then quadrupoles, then octupoles. Most of the learning, they write, happens while planning — deciding which component to read, how to align the magnets and over which distances the measurement makes sense. Apart from the phone, the whole kit is four small magnets and a sheet of graph paper.
