The solubility of hydroxyl in coesite was investigated in multianvil experi
ments performed at 1200 degreesC over the nominal pressure range 5-10 GPa,
at an f(O2) close to the Ni-NiO buffer. The starting material for each expe
riment was a cylinder of pure silica glass plus talc, which dehydrates at h
igh P and T to provide a source of water and hydrogen (plus enstatite and e
xcess SiO2). Fourier-transform infrared (FTIR) spectra of the recovered coe
site crystals show five sharp bands at 3606, 3573, 3523, 3459, and 3299 cm(
-1), indicative of structurally bonded hydrogen (hydroxyl). The concentrati
on of hydrogen increases with pressure from 285 H/10(6) Si (at 5 GPa) to 14
15 H/10(6) Si (at 10 GPa). Assuming a model of incorporation by (4H)(Si) de
fects, the data are fit well by the equation C-OH = Af(H2O)(2)exp(-P DeltaV
/RT), with A = 4.38 H/10(6) Si/GPa, and DeltaV = 20.6 x 10(-6) m(3) mol(-1)
. An alternative model entailing association of hydrogen with cation substi
tution can also be used to fit the data. These results show that the solubi
lity of hydroxyl in coesite is approximately an order of magnitude lower th
an in olivines and pyroxenes, but comparable to that in pyropic garnet. How
ever, FTIR investigations on a variety of ultrahigh pressure metamorphic ro
cks have failed in all cases to detect the presence of water or hydrogen in
coesite, indicating either that it grew in dry environments or lost its hy
drogen during partial transformation to quartz. On the other hand, micro-FT
IR investigations of quartz crystals replacing coesite show that they conta
in varying amounts of H2O. These results support the hypothesis that preser
vation of coesite is not necessarily linked to fast exhumation rates but is
crucially dependent on limited fluid infiltration during exhumation.