Experimental Deformation of Rocksalt

  1. B.E. Hobbs and
  2. H.C. Heard
  1. J. Handin,
  2. J. E. Russell and
  3. N. L. Carter

Published Online: 18 MAR 2013

DOI: 10.1029/GM036p0117

Mineral and Rock Deformation: Laboratory Studies: The Paterson Volume

Mineral and Rock Deformation: Laboratory Studies: The Paterson Volume

How to Cite

Handin, J., Russell, J. E. and Carter, N. L. (1986) Experimental Deformation of Rocksalt, in Mineral and Rock Deformation: Laboratory Studies: The Paterson Volume (eds B.E. Hobbs and H.C. Heard), American Geophysical Union, Washington, D. C.. doi: 10.1029/GM036p0117

Author Information

  1. Center for Tectonophysics, Texas A&M University, College Station, Texas 77843

Publication History

  1. Published Online: 18 MAR 2013
  2. Published Print: 1 JAN 1986

ISBN Information

Print ISBN: 9780875900629

Online ISBN: 9781118664353



  • Rocks—Testing—Addresses, essays, lectures;
  • Rock deformation—Addresses, essays, lectures


Using newly designed apparatus for triaxial-compression testing of 10 by 20-cm cores of Avery Island rocksalt at constant strain-rates between 10−4 and 10−6/s, temperatures between 100° and 200°C, and confining pressures of 3.4 and 20 MPa, comparing our data with those of other workers on the same material, and observing natural deformations of rocksalt, we find that (1) constant-strain-rate and quasi-constant stress-rate tests (both often called quasi-static compression tests) yield essentially similar stress-strain relations, and these depend strongly on strain rate and temperature, but not confining pressure; (2) fracture excluded, the deformation mechanisms observed for differential stresses between 0.5 and 20 MPa are intracrystal-line slip (dislocation glide and cross-slip) and polygonization (dislocation glide and climb by ion-vacancy pipe diffusion); (3) the same steady-state strain rate $$({\rm \dot \varepsilon }_{{\rm SS}} )$$ and flow stress are reached at the same temperature in both constant-strain-rate and constant-stress (creep) tests, but the strain-time data from transient creep tests do not match the strain-hardening data unless the initial strain, ε0 (time-dependent in rocksalt) is accounted for; in creep tests the clock is not started until the desired constant stress is reached; (4) because the stress-strain curve contains the entire history of the deformation, the constant-strain-rate test rather than the creep test may well be preferred as the source of constitutive data; (5) furthermore, if the stress or temperature of the creep test is too low to achieve the steady state in laboratory time, one cannot predict the steady-state flow stress or strain rate from the transient response alone, whereas we can estimate them rather well from constant-strain-rate data even when strain rates are too high or temperatures too low to reach the steady state within a few hours; (6) the so-called “baseline creep law”, giving creep strain, $${\rm \varepsilon }\,{\rm = }\,{\rm e}_{\rm a} [1\, - \,\exp ( - {\rm \xi t})]\, + \,{\rm \dot \varepsilon }_{{\rm SS}} \,{\rm t}$$, where ea ξ, and $${\rm \dot \varepsilon }_{{\rm SS}}$$ are regarded as material properties as well as fitting parameters, can be valid, if at all, only over intervals of stress and temperature where the same deformation mechanisms operate and only if it is independent of structural changes, that is of loading path, and it poorly predicts constant-stress-rate response in triaxial-compression tests and long-term, low-stress response from data taken over short time at high stress; (7) a potentially more useful, semi-empirical constitutive model, incorporating stress (σ), strain (ε), strain rate $${\rm (\dot \varepsilon )}$$, and absolute temperature (T), and capable of matching at least limited constant-strain-rate, constant-stress-rate, constant-stress (creep), and relaxation (nearly constant strain) data even though constant structure is assumed, is $${\rm \sigma }\,{\rm = }\,{\rm K\dot \varepsilon }ˆ{\rm q} \,\exp \,({\rm B/T})\,[1\, - \,\{ \exp ( - {\rm r}_1 {\rm \varepsilon )}]\, + \,\exp ( - {\rm r}_2 {\rm \varepsilon )\} /2]}$$, where K, q, B, r1, and r2 are to be treated as fitting parameters until their physical significance is better understood; (8) however, we doubt that any single, perfectly general constitutive equation can be written to satisfy all conditions pertinent to repository design, say 25° ≤ T ≤ 300°C and 1 ≤ σ ≤ 20 MPa, and also to be workable in numerical modeling; (9) hence, no matter how abundant and precise site-specific laboratory data may become, one can expect only to approximate the rheological behavior of the prototype.