Insights into the long-standing controversy over sound velocities in lizardite
Serpentine minerals are abundant H-bearing phases in the upper mantle and play a central role in water transport in subduction zones. Lizardite is the low-temperature polymorph expected in cold slabs and shallow serpentinized mantle. Using an ab initio thermoelastic framework combining density-functional theory with the r2SCAN meta-GGA and the quasiharmonic approximation, we calculate its compression curve, elastic tensor, and acoustic velocities at 300 K. r2SCAN reproduces the measured compression curve and equilibrium volume to within 0.2%. Weak interlayer hydrogen bonding gives rise to pronounced axial softening, with $C_{33} \simeq 0.21C_{11}$ and $C_{44} \simeq 0.13C_{66}$ at 0 GPa. Solving the Christoffel equation over all propagation directions, we construct the velocity density of states (VelDOS), which reveals a high density of slow acoustic modes and other directional features obscured by Voigt-Reuss-Hill (VRH) aggregate averages. Reported experimental velocities preferentially sample the slower portion of this distribution rather than clustering near the VRH values, a correspondence consistent with subtle crystallographic texture or preferential orientation not resolved experimentally. Thus, the appropriate seismic-velocity endmember for lizardite depends on its orientational state, and reliance on VRH velocities alone may lead to overestimation of the degree of serpentinization when seismic observations preferentially sample slow crystallographic directions. Direction-resolved single-crystal measurements and aggregate measurements with independently quantified texture would provide direct tests of this interpretation.