Creep hydrolytically weakened synthetic quartz crystals oriented to promote {2-1-10} < 0001 > slip : a brief summary of work to date

Abstract

Over 80 creep experiments have been carried out on hydrothermally grown single crystals of synthetic quartz with 50 to 4000 OH- per 106 Si atoms and oriented with the compression direction at 450 to a and c. The creep tests were performed at temperatures between 400 and 800° C, at applied stresses between 434 and 1621 bars., and were compressed to strains of typically 4.5 + 1.0%. All of the creep curves were sigmoidal in shape, with an initial incubation stage of accelerating creep rates followed by a hardening stage in which creep rates decreased continuously with time. Cardinal rate parameters for each test are the maximum strain rate, the incubation time t{ (the zero strain intercept of the segment of the creep curve with the highest strain rate) , and the strain rate at a specific creep strain, usually 2, 3, or 4 %. The strain rates so chosen follow thermally activated power law rate equations. The stress exponent n is, within experimental uncertainty, independent of OH-concentration and unaffected by the alpha-beta phase transformation and takes on values between 3.0 and 5.5. The activation energy for creep Ec* does not appear to vary systematically with OH- concentration but is strongly affected by the alpha-beta phase transformation, with E,.* ranging between 20 and 40 kcal/mole in the alpha quartz stability field and 8 to 18 kcal/mole in the beta quartz field. Optical and transmission electron microscopy on the creep specimens have proven that {2110} < 0001 > is the dominant slip system. Dislocation density rapidly increases in the incubation stage. A dislocation multiplication model based on exponential growth quantitatively fits creep strain variations with time and the observed rate of increase of dislocation density with creep strain. At strains sufficiently low that dislocation interactions can be neglected, the inverse of the incubation time \jtt is proportional to the dislocation velocity. The effects of composition, temperature, and stress on dislocation velocity can then be inferred from the effects of these parameters on 1 jtf. By this new technique, we show that the dislocation velocity follows a thermally activated power law with stress exponent and activation energy relatively insensitive to variations in OH-concentration. Preliminary experiments suggest that dis¬ location velocity increases with increasing OH-concentration. In the hardening stage, the dislocation density is essentially constant, and the volume fraction of precipitated molecular water increases with time and strain. It is believed that the hardening stage is associated with the precipitation of molecular water into bubbles, but we have not been successful in quantitatively fitting the observed hardening rates to any precipitation hardening model. We have shown that the hardening effects associated with the obstruction of dislocation movement by bubbles are small relative to possible effects associated with the loss of structurally bound water during precipitation. The evidence derived from the kinetics in the incubation stage suggests that the steady state flow law for synthetic quartz in the absence of precipitation hardening effect is of the form : è = A σn exp (— EC*/RT) where n = 3.5 + 0.5, Ec* = 39 ± 5 kcal/mole and A increases with increasing OH-concentration. The parameters n and Ec* from the higher strain creep rate data deviate from these values, and we believe that the discrepancies stem from the effects of hardening by precipitation of molecular water and from the effects of strain-induced microfracturing in some of the tests.