Melt Diffusion-Moderated Crystal Growth and its Effect on Euhedral Crystal Shapes

• Published in: Journal of petrology Vol. 64; no. 8
• Main Authors: Mangler, Martin F, Humphreys, Madeleine C S, Geifman, Eshbal, Iveson, Alexander A, Wadsworth, Fabian B, Brooker, Richard A, Lindoo, Amanda, Hammond, Keiji
• Format: Journal Article
• Language: English
• Published: 01-08-2023
• ISSN: 0022-3530; 1460-2415
• DOI: 10.1093/petrology/egad054

Crystal growth is often described as either interface-controlled or diffusion-controlled. Here, we study crystal growth in an intermediate scenario where reaction rates at the crystal-melt interface are similar to the rates of diffusive transport of ions through the melt to the advancing crystal surface. To this end, we experimentally investigated euhedral plagioclase crystal shapes in dry mafic (basaltic) and hydrous silicic (haplodacitic) melts. Aspect ratios and inferred relative growth rates of the 3D short (S) and intermediate (I) crystal dimensions vary significantly between mafic and silicic melts, with δS:δI = 1:6–1:20 in basalt and 1:2.5–1:8 in hydrous haplodacite. The lower aspect ratios of plagioclase grown in the silicic melt coincide with 10 to 100× lower melt diffusion rates than in the mafic melt. Using an anisotropic growth model, we show that such differences in melt diffusivity can explain the discrepancy in plagioclase aspect ratios: if interface reaction and melt diffusion rates are of similar magnitude, then the growth of a crystal facet with high interfacial reaction rates may be limited by melt diffusion, while another facet of the same crystal with lower interfacial reaction rates may grow uninhibited by melt diffusivity. This selective control of melt diffusion on crystal growth rates results in progressively more equant crystal shapes as diffusivity decreases, consistent with our experimental observations. Importantly, crystals formed in this diffusion-moderated, intermediate growth regime may not show any classical diffusion-controlled growth features. The proposed model was developed for plagioclase microlites but should be generalisable to all anisotropic microlite growth in volcanic rocks.