One of the most basic properties of a metal is the screening of electric fields. A metal will not tolerate a voltage difference.The charges in metals are delocalized and thus free to move until they reach a configuration where the internal electric field is exactly zero. This is in fact the definition of a metal that is given in introductory electromagnetism. An insulator by contrast is a material where the electrons and ions are bound to a specific location by chemical bonds and their motion throughout the body of a solid is blocked.

We performed electric deflection experiments on free clusters made from over 20 elements (Na, Al, V, Mn, Fe, Co, In, Nb, Mo, Ru, Rh, Pd, Ta, W, Au, Y, Pr, Tb Ho, Tm, and Bi) in a search for electric dipole moments. An electric dipole moment implies that charges are seperated inside a cluster and thus there must exist large internal electric fields. The presence of a dipole moment implies a loss of metallicity.

A molecular beam electric deflection experiment is an extremely robust test for the presence of a dipole moment. The deflection experiments were performed at a temperature of 15-20 K which ensures that the overwhelming majority of the clusters in the beam are in their vibrational ground state. A polar cluster with a permanent dipole moment will rapidly precess inside of the electric deflection field and the beam will be defocused or broadened. This broadening of the beam in deflection is the signiture of a permanent electric dipole moment, and allows for an order of magnitude estimate of the size of the moment. (A precision measurement of the dipole moment requires a priori structural information and must be extracted by comparison with molecular dynamics simulations)

Many free clusters are either ferromagnetic or paramagnetic and have nonzero total spin. This spin will will precess when placed into an external magnetic field. The frequency of this precessional motion depends on the total spin and the strength of the magnetic field. If we excite the precessing spin with microwave radiation at this frequency then the orientation of the spin will be changed. These "Rabi oscillations" can be detected by molecular beam deflection experiments. We direct a beam of clusters through a series of three magnets. The first inhomogenous Stern-Gerlach magnet (A) deflects the clusters, defocusing the beam. The (C) field is spatially homogenous and contains a high-Q microwave resonator. If strength of C field is tuned so that the cluster's spin precesses at the cavity's resonant frequency then the cluster's spin will be reversed. The third Stern-Gerlach magnet (B) is identical to the first. The fraction of clusters that were resonantly excited by the (C) field are refocused while the rest of the clusters are deflected out of the beam.

In a cluster, the spin is not completely uncoupled. The spin-orbit and hyperfine interactions couple the spin to the cluster body, as well as to the rotations and nuclear spin moments. Each of these couplings will split the resonances and slightly shift their frequencies. These frequency shifts can be measured to extremely high precision. It is reasonable to expect that the shifts can be determined to 1 part per million. This should be compared with the present level of precision in cluster experiments which is seldom better than 1 percent! It should also be noted that this is among the most sensitive detection schemes ever invented - a molecular beam ESR spectrometer is capable of detecting the flipping of a single quantum spin. This represents an energy change on the order of a few meV!

Cobalt and iron clusters Co_N, Fe_N (20 < N < 150) measured in a cryogenic molecular beam are found to be bistable with magnetic moments per atom both µ_N/N~2µ_B in the ground states and µ_N*/N~µ_B in the metastable excited states (for iron clusters, µ_N ~3Nµ_B and µ_N* ~Nµ_B). This energy gap between the two states vanish for large clusters, which explains the rapid convergence of the magnetic moments to the bulk value and suggests that ground state for the bulk involves a superposition of the two, in line with the fluctuating local orders in the bulk itinerant ferromagnetism.

The Langevin-Debye Susceptibility formula holds for a cold cluster adiabatically entering a magnetic field despite the lack of a heat bath to thermalize the orientation of the spin. This is a consequence of the dense nest of a avoided level crossings in the Zeeman diagram of the cluster.

At low temperatures Nb, V, and Ta clusters acquire large electric dipole moments. The deflection profiles are highly asymmetric suggesting that the ferroelectric polarization is weakly coupled to the cluster body. The ferroelectric fraction of the beam declines at temperatures ~40 K. Laser heating experiments which add only one quantum of angular momentum but raise the temperature of the cluster by a huge factor demonstrate the the vanishing of the dipole moment at high temperatures is not an artifact of the rotational motion of the cluster.

The most revealing effect of all is the strong odd-even alternation of the ferroelectric fraction which begins around N = 20, and persists consistently until N = 100. The even-N clusters show a larger ferroelectric fraction than the odd-N clusters. This odd-even alternation depends only on total number of valence electrons in the cluster. The electronic origin of this effect has been demonstrated by a series of experiments where Nb clusters were doped with an impurity - impurities which donate or oxidize an odd number of valence electrons invert the odd-even effect, while impurities such as oxygen which oxidize an even number of valence electrons leave the odd even effect unaffected. Doping a cluster with a magnetic impurity like manganese quenches the dipole moment. (although Co does not appear to destroy the ferroelectricity).