Why do two gases with magnetic electrons respond differently?
Oxygen and nitric oxide both respond to a magnetic field, yet simply counting magnetic electrons did not explain their measured susceptibilities, the amount of magnetization induced by a weak field. The puzzle became sharper when spectroscopy identified closely separated quantum states in nitric oxide, different allowed arrangements and energies of the molecule’s electrons. A field could act on those states in ways that a classical picture of freely rotating permanent magnets missed.
Van Vleck’s work connected magnetic response with the same energy levels used to interpret spectra. His 1928 paper calculated nitric oxide’s susceptibility using a spectroscopic separation of about one hundred twenty-one inverse centimetres and predicted a detectable departure from Curie’s inverse-temperature law. The important move was methodological: the magnetic experiment and the spectroscopic experiment had to describe one molecule with one set of quantum states. The moment and its temperature dependence now followed from the molecule’s internal structure.
Sources: [1][2]
A susceptibility curve is a carefully normalized slope
Magnetic susceptibility describes the small-field change in magnetization produced by an applied field. In the usual SI volume convention, magnetization and magnetic-field strength are measured in amperes per metre, making their ratio dimensionless. A measurement can use a magnetic force, an induced signal or a magnetometer. The response must be divided by the appropriate sample amount before temperatures or different materials are compared.
For a gas, changing temperature at fixed pressure also changes the number of molecules per volume. A raw signal can therefore vary even before its magnetic response per molecule changes. Sample holders, windows and impurities contribute backgrounds; some are diamagnetic and oppose the applied field. A useful plot specifies whether susceptibility is per volume, per mole or per mass. Otherwise an apparent failure of a microscopic model can begin as an unnoticed change of units or density.
Sources: [1][2]
Curie’s law has assumptions that can be tested
For independent, isotropically oriented moments in a weak field, the field slightly favours orientations with lower magnetic energy. Thermal motion opposes that preference. The resulting Curie law says that susceptibility grows as temperature falls. Its coefficient depends on the effective quantum moment and on how many magnetic objects the sample contains.
In this formula, susceptibility is dimensionless in the SI volume convention, temperature is in kelvin, and the Curie constant has units of kelvin. Its value grows with the number of independent moments per volume and with the square of their effective moment. That effective moment includes the quantum angular-momentum factor, which differs from simply taking its largest projection along an axis. Multiplying measured susceptibility by temperature makes an ideal Curie contribution constant. Curvature in that product points to changing level populations, interactions or an additional response that varies little with temperature.
Sources: [1]
Nitric oxide put a spectroscopic energy scale into the prediction
The low-lying states of nitric oxide are separated by an energy that is neither negligible nor overwhelmingly large over an ordinary laboratory temperature range. Their Boltzmann populations therefore change appreciably as the gas is heated or cooled. The magnetic response must average the different states with those temperature-dependent weights, including the allowed magnetic couplings between them.
For scale, one hundred twenty-one inverse centimetres corresponds to roughly one hundred seventy-four kelvin after dividing the energy by Boltzmann’s constant. In a hypothetical two-level estimate before degeneracy factors, the upper-to-lower population ratio would be about three percent at fifty kelvin and more than one half at three hundred kelvin. Those numbers illustrate why replacing all internal structure by one fixed moment misses a temperature dependence. A full molecular calculation adds the actual number of states at each energy and the strength with which a field connects them.
Sources: [1]
A nonmagnetic ground state can still acquire a moment
Quantum mechanics supplies another route to a positive response. A ground state, the lowest allowed quantum state, can be unique and have zero permanent moment yet mix slightly with excited states when a magnetic field is applied. For a weak enough field, that small mixing lowers the ground-state energy by an amount proportional to field squared. The corresponding induced moment grows in direct proportion to the field.
For weak magnetic induction, measured in tesla, this energy shift grows with the square of the field. Its size depends on how strongly the magnetic moment connects the states and decreases as their energy separation grows. The approximation applies while the excited-state admixture is small. This is a change of the quantum state itself; thermal population remains small when temperature is far below the excitation energy. Further cooling then changes little, giving the approximately temperature-independent positive response associated with Van Vleck’s name.
Sources: [1][3]
A crystal changes the states before the magnetometer measures them
A magnetic ion sits inside an electric environment produced by its neighbours. That crystal field can split orbital levels, change the ground-state degeneracy and alter the matrix elements that a magnetic field couples. The same chemical element can consequently give different susceptibility curves in different compounds. Counting its unpaired electrons is a starting clue, not a complete material prediction.
A later experimental example is thulium aluminium garnet. Awaka and colleagues measured its susceptibility from two to three hundred kelvin with a SQUID magnetometer, a superconducting instrument sensitive to tiny magnetic signals, and reported a nearly temperature-independent low-temperature response below about eight kelvin. Their analysis used a nonmagnetic crystal-field ground state and excited states. They also stated that a simplified cubic field did not capture the ion’s full local symmetry. The example shows both the power of level-based reasoning and why a fitted energy scheme needs independent structural or spectroscopic checks.
Sources: [3]
Change the temperature and the level spacing for different tests
The two mechanisms give different predictions even before a complicated material calculation. In an ideal Curie contribution, halving the temperature doubles susceptibility. For an isolated nonmagnetic ground state far below its excited levels, the mixing contribution instead changes little. Temperature variation therefore asks whether real thermal populations are controlling the response.
A second thought experiment holds the strength of the magnetic coupling between the states fixed and doubles the energy separation to the only relevant excited level. The weak-field mixing contribution is then halved: the distant state is harder for a weak field to mix into the ground state. This is an illustrative comparison of energy scales. In a real crystal, a change of composition or pressure may alter coupling strengths and several gaps at once. Comparing magnetic data with spectroscopy along more than one crystal direction helps distinguish those possibilities. The measurements test a level scheme through more than its ability to fit one temperature curve.
Sources: [1][3]
A flat low-temperature curve is the beginning of an explanation
A nearly constant susceptibility is not by itself a unique fingerprint of Van Vleck mixing. Core-electron diamagnetism and itinerant-electron contributions can also be weakly temperature dependent. A small concentration of magnetic impurities can instead add a rising low-temperature tail. Separating these terms requires field dependence, temperature range and knowledge of the sample, not just a straight segment on one graph.
Van Vleck’s lasting change was to make magnetic response a question about internal quantum states. A magnetic field can rearrange their populations or gently mix them, and spectroscopy tells us which states are available. Comparing those measurements in the same molecule or crystal turns magnetism from a count of permanent moments into a probe of how its electrons are bound and coupled.
Sources: [1][3]