Defining the limits to long-term nano-indentation creep measurement of viscoelastic materials.

Abstract An ultra-stable instrumented nano-indentation tester (UNHT, Anton Paar) was used to study extremely long (30,000 s) indentation creep of polymers. Total drift rate, measured on fused silica and sapphire samples, was less than 0.2 pm/s for up to 10 h, enabling collection of valid, low-uncertainty, long-term creep data - orders of magnitude longer than previously possible. Fits of the popular N -element Kelvin model to indentation creep data gave values of instant elastic modulus E 0 ∗ and infinite modulus E ∞ ∗ strongly dependent on the time-span of data fitted. Comparison with the long-term experimental data showed that the model was unable to use short term data to predict creep at the much longer times required by industry. A new analysis method is proposed to obtain a better estimate of the true value of infinite modulus, E ∞ ∗ , which is the simplest indicator of maximum dimensional change of polymeric material components subject to long-term stress. Highlights • We have shown that the UNHT system provides extremely good stability when located in a well-controlled environment: displacement drift <0.2 p.m./second • The ultra-stable experimental setup enables the collection of valid, low-uncertainty, long-term creep data (up to 30,000 s) - orders of magnitude longer than previously possible, to investigate the viscoelastic behaviour of polymer • The results set clear limits to the validity of long-term indentation creep data collection • The results set important limits to the prediction of viscoelastic behaviour using Kelvin model • We have proposed and demonstrated a new model to overcome the analysis limitations identified as a pragmatic attempt to predict the future behaviour of a material such as PS-3 from shorter duration test data and to provide results that are more independent of the creep test duration. • We made a number of recommendations to improve indentation creep testing including the selection of indenter geometry and loading cycles [ABSTRACT FROM AUTHOR]