b'A particularly surprising37discovery was that thetwo isotopes of hydrogen behaved very differentlyin the melt structure.Typically, water in silicate melts is thought to either dissolve in the melt as water or react with the melt, converting the silicon bond Si-O-Si to silanol, Si-OH. For decades it has been understood that water-melt reactions depend on the amount of water. Variations in the chemistry of the melt were thought to have minimal effect. Focusing on1 H and2 H, Cody and colleagues showed that melt chemistry actually has aCarolyn Beaumont, a summer research intern working with very strong control on what water does.staff scientists George Cody and Bjrn Mysen, prepares alkali silicate oxidewater samples for melting under high pressures and temperatures. Critical to these studies is the ability to seal A particularly surprising discovery was that the twothe starting materials in platinum reactors. Beaumont is using a isotopes of hydrogen behaved very differently in themicroscope and a pulsed DC arc welder to make high-precision melt structure. One of the great mysteries of Earthsmicro-welds that will ultimately seal and retain water in the silicate melts when heated to temperatures as high as 2550F chemistry is why the ratio of deuterium to hydrogen(1400C) and pressures of 15,000 times normal atmospheric of Earths water is significantly larger than that ofpressure using a piston-cylinder solid medium pressure device. the Sun and the universe. Earths oceans evolvedImage courtesy George Codyfrom a primitive magma ocean, and the fact that the hydrogens behave so differently in the melt structure could help explain why Earths oceans have the deuterium to hydrogen ratio that is observed today.This graphic shows the difference of heavy hydrogen called deuterium ( 2 H or D) and regular hydrogen ( 1 H) in hydrated silicate glasses that were quickly cooled from melts. From a chemical perspective, D 2 O and H 2 O are expected to behave very similarly; however, nuclear magnetic resonance (NMR) results clearly show that deuterium (red) and hydrogen (blue)2 H NMRhave strong preference for different molecular environments in the amorphous glass. Classical isotopic fractionation theory, which explains separation processes, cannot explain1 H NMR this significant preference. This high degree of preference may have left a distinct isotopic signature, recording the process when liquid rock and water separated, leading to the60 40 20 0 -20 -40condensation of Earths oceans during the earliest history of the planet. Frequency (in ppm Hz)Image courtesy George Cody'