NASA’s Voyager Mission is also known as the Voyage of Discovery to the outer planets Jupiter, Saturn, Uranus and Neptune . Carl Sagan and many other Space Scientists contributed to the design and management of Voyager. Planetary interiors are ultracondensed matter at finite temperatures T. When the Voyager 1 and 2 spacecraft departed Earth in 1977, the centers of Uranus and Neptune (U/N) were estimated to be at pressure P of 20×106 bar (2 TPa) and T of 7000 K . In 1977 no one yet knew that many Voyager discoveries would be substantially different than imagined prior to launch, nor that spacecraft lifetimes would be so long. Forty years after that launch, the New York Times published articles about the lives of engineers who had maintained connections with the Voyagers and technological advances made over that time [3,4]. Today in 2018, long after it completed its visit to Neptune in 1989, Voyager continues to transmit magnetic-field data measured in interstellar space as it zips along at ~16 km/s.
Most of the diagnostics on Voyager characterized surfaces and local environments of those objects in Space. In contrast, the magnetometer and gravitometer on Voyager 2 measured the magnetic and gravitational fields of U/N, which are sensitive to materials in the deep interiors rather than surface regions. The gravity measurements detected two layers in each planet, a relatively low-density outer region composed mostly of hydrogen and a much higher-density inner region  probably composed mostly of planetary “Ice”, a fluid composed mostly of water (H2O), ammonia (NH3), and methane (CH4).
The magnetometer measured magnetic-field intensities of U/N, which are comparable to those of Earth. However, the measured spatial dependences of the magnetic fields of U/N are non-dipolar and non-axisymmetric, completely unlike Earth’s nearly dipolar and nearly axisymmetric magnetic field. If the measured magnetic fields of U/N are force-fit to effective dipolar fields, those effective magnetic axes are tilted 59° and 47° from their respective rotational axes and their effective dipole centers are offset by 33% and 55% of their respective radii. The magnetic fields of U/N measured by Voyager 2 are illustrated in Fig. 1.
The question of why the magnetic-fields of Earth and of U/N differ so greatly has been a major unresolved question in planetary physics ever since Voyager 2. Planetary magnetic fields are made by dynamos: convective paths of electrically conducting fluids across lines of magnetic force, which electrical currents generate magnetic field . To determine why magnetic fields of U/N are so different from that of Earth, one must consider the cause of planetary magnetic fields, properties of likely constituent materials at pressures P and temperatures T at which planetary magnetic fields are made, and possible interior mechanical coupling between layers of materials that might have a significant affect on convective flows and thus on generation of planetary magnetic fields.
Analogous to Voyager’s trip to U/N to measure planetary magnetic fields, electrical conductivities of likely convecting fluids at likely P/T should be measured on Earth to estimate likely radii in U/N at which external magnetic fields are made, which might also explain their non-dipolar nature. That is, the closer a magnetic field is made to a planetary surface, the more likely it is that the magnetic field will have a non-dipolar contribution . Laboratory measurements are also necessary to determine if molecular “Ice” actually exists in molecular forms of H2O, NH3 and CH4 in the interiors of U/N as commonly assumed. “Icy” molecules that accrete during planet formation decompose at high interior P/T above ~100 GPa or less. If such decomposition occurs, then a substantial compostion of fluid H is available to form MFH, in addition to H provided by accretion of nebular H2.
Near the centers of U/N, today P/T are estimated to be about 700 GPa and 7000 K (Hubbard, private communication, 2011). To make measurements at both high P and T, relevant thermodynamic states were achieved by dynamic compression, which is achieved via a supersonic matter wave. Electrical conductivities of many likely planetary fluids at likely P/T in U/N were measured up to 180 GPa (1.8 x 106 bar) and several 1000 K achieved by dynamic compression generated with a two-stage light-gas gun at Lawrence Livermore National Laboratory. Those P/Ts correspond to regions in U/N in which external magnetic fields of U/N are made. Those extreme conditions were achieved by impact of a metal plate at velocities up to 8 km/s onto small cryostats of liquid H2 at 20 K, for example. Experimental lifetimes were 100 ns. H2O, NH3, CH4, an “Icy” mixture called Synthetic Uranus, CO, CO2, and hydrocarbons were also investigated over a wide range of pressures, densities and temperatures achieved by shock and quasi-isentropic compression. Measured electrical conductivities indicate fluid H is semiconducting from 93 to 140 GPa and a poor metal from 140 to 180 GPa with measured electrical conductivity of 2000/(ohm-cm), which is Mott’s minimum metallic conductivity. Corresponding calculated Ts range from 1700 to 2600 K at pressures from 93 to 140 GPa and then up to 2900 K at 180 GPa.
A key question about U/N is the cause of the apparent decoupling of planetary rotational motion from convective motions of fluids in which planetary magnetic fields are generated. As seen in Fig. 1, the effective magnetic axes of U/N are tilted substantially (~50°) from their respective rotational axes. In contrast Earth’s magnetic axis is tilted only ~10° from its rotational axis. Materials in the two-layer structures and natures of the interfaces between the two layers of U/N and Earth are quite different, which suggests substantially different coupling of planetary rotational motions with convective fluid flows in the two cases, which is probably the case as discussed below.
Earth has a strong, rock mantle which rotates at nearly constant angular velocity, which is expected to cause a tendency to form current loops in convecting Fe-rich fluid such that magnetization generated tends to align parallel or anti-parallel to the axis of rotation of Earth. Possible interactions that might cause coupling between Earth’s strong, rock mantle across the relatively sharp boundary with its fluid Fe-rich Outer Core include an estimated surface roughness of less than 0.5 km on the inner radius of the Earth’s mantle (3500 km) . In contrast, U/N have weak fluid H-He envelops, probably with diffuse crossovers to their weak fluid Cores. In this case local convective dynamo motions of the fluids that produce the magnetic fields of U/N are expected to be weakly coupled to global rotational motions of U/N, as implied by Fig. 1. If true, then the dynamos of U/N would be relatively free to wander as local conditions dictate. In this case effective tilt angles and effective center-offsets of their magnetic fields would be expected to vary slowly and unconstrained over the age of the Solar System. “Polar wander” is probably a better term for the time dependence of magnetic fields of U/N than “polar reversal” as in the case of Earth.
Based on a substantial database measured over three decades for numerous representative planetary fluids, the non-dipolar non-axisymmetric magnetic fields of Uranus and Neptune (U/N) are (i) made primarily by degenerate metallic fluid H (MFH) at or near crossovers from H-He envelopes to “Ice” cores at ~100 GPa (Mbar) pressures and ~90% the radii of U/N; (ii) numerous likely planetary fluids investigated experimentally decompose at P/T above ~100 GPa or less and few 1000 K; (iii) ironically there probably is little molecular nebular “Ice” in the Ice Giants; (iv) electrical conductivity of MFH is up to a factor of 100 larger than conductivity of “Ices” thought previously to make the magnetic fields of U/N (20/(ohm-cm)); (v) because those magnetic fields are made close to outer surfaces, non-dipolar magnetic fields can be expected as observed; (vi) “Ice” cores are a heterogeneous fluid mixture of nebular Ice and Rock that accreted, sank below the H-He envelopes into the cores in which nebular materials decomposed at high pressures and temperatures and re-reacted to form new chemical species; (vii) those magnetic fields are probably non-axisymmetric because rotational motions of U/N are weakly coupled to convective motions that make their magnetic fields by dynamos. For U/N “polar wander” is probably a better descriptor for variations of magnetic field over time than “polar reversal” as for Earth.
 U. S. National Aeronautics and Space Agency Voyager Program website.
 W. B. Hubbard and J. J. MacFarlane. (1980). Structure and Evolution of Uranus and Neptune. Journal of Geophysical Research, 85, 225-234.
 K. Tingley, The Loyal Engineers Steering NASA’s Voyager Probes Across the Universe, New York Times, Aug. 3, 2017.
 D. Overbye, A Reverie for the Voyager Probes, Humanity’s Calling Cards, New York Times, Aug. 21, 2017.
 R. Helled, J. D. Anderson, M. Podolak and G. Schubert (2011). Interior models of Uranus and Neptune. Astrophysical Journal, 726, 15-1-15-7.
 D. J. Stevenson (1983). Planetary Magnetic Fields, Reports on Progress in Physics, 46, 555-620.
 K. D. Granzow. (1983). Spherical harmonic representation of the magnetic field in the presence of a current density. Geophysical Journal of the Royal Astronomical Society, 74, 489-505.
 R. Hide, R. W. Clayton, B. H. Hager, M. A. Spieth and C. V. Voorhdes. (2013). Topographic core-mantle coupling and fluctuations in the Earth’s rotation. In Relating Geophysical Structures and Processes: The Jeffrey Volume, Geophysical Monograph 76, Vol. 16, eds. K. Aki and R. Dmowska, IUGG, pp. 107-120.