This article focuses on some of the applications being developed by the Civil Engineering Department at the Massachusetts Institute of Technology (MIT), for whom Anton Paar developed an innovative humidity-controlled nanoindentation system. This group focuses on the study of geomaterials or natural composites, i.e., materials which have complex heterogeneous structures and those that occur naturally. Examples of such materials are shales (which cover oilfields) and cement pastes (used for concrete).
C-S-H is a nonstoichiometric compound and the average Ca/Si ratio in ordinary hardened cement paste is about 1.7. It has a layered crystal structure similar to that of the mineral tobermorite, and this is the reason why this mineral is frequently used as a model material as it is analogous to the main hydrated phase of cement paste and can be artificially created. The multiscale heterogeneity of concrete eventually determines its in vivo mechanical performance (strength, stiffness) and degradation (damage, failure, fracture). It is possible to divide the microstructure into four levels as shown in Figure 1, from the scale of mortar (10-2 m) down to the C-S-H solid phase (10-10 m) which represents the smallest material length scale that is presently accessible by mechanical testing (nanoindentation). By experimentally investigating the mechanical properties of cement paste at varied length scales provides a means of correlating such microscale properties to macroscale applications.The last decades have seen a gradual improvement of the mechanical properties of cement, but this has been achieved more by trial and error than by an in-depth understanding of what is happening at the micro and nanoscales. The setting of cement is not a drying process, as sometimes believed, but in fact, it is the exact opposite. When cement is mixed with water, it goes through a dissolution reaction generating silicate, calcium and aluminate ions in the interstitial solution. New products (hydrates) then precipitate after reaching their solubility limit and also after a nucleation period. In a common cement, such as Portland cement, this dissolution-diffusion-precipitation process produces calcium hydroxide (Portlandite) and calcium silicate hydrate (C-SH). During hydration, the slurry coagulates after the cement is mixed with water, after which setting occurs. Some proportion of the anhydrous cement is converted into C-S-H and other hydrates.
The need for in-situ analysis of chemically complex phases obviates conventional mechanical testing of bigger specimens representative of these material components. It is thus possible to use nanoindentation as a 2D mapping tool for examining the properties of constituent phases independently of each other.
Figure 1. Four-level microstructure of cement-based composite materials
Grid Indentation Technique
A novel grid-indentation technique can be used to attain significant mechanical properties of heterogeneous materials (such as C-S-H) at a specific length scale, and it also provides access to the volume fractions of independent phases. Consider a material to be composed of two phases of varied mechanical properties and characterized by a length scale D, as shown in Figure 2. If the indentation depth is much smaller than the equal to the surface fraction occupied by the two phases on the indentation surface. On the other hand, an indentation test performed to a maximum indentation depth that is much bigger than the characteristic size of the individual phases, h > D, senses the average response of the composite material, and the properties extracted from such an indentation experiment are representative in a statistical sense of the average properties of the composite material.
These large matrices of indentations offer a statistical analysis comprising of distributions and their derivatives (e.g. frequency diagrams or histograms) of mechanical properties determined by a large number of indentation experiments at a particular scale of material observation defined by the indentation depth. Hardened concrete always comprises of a major fraction of liquid water from the capillary condensation of water vapor in the intergranular pores, hence it is of particular interest to carry out grid-indentations with accurate Relative Humidity (RH) control in order to quantify the influence of water fraction on mechanical properties. This indeed was the motivation for designing a completely-automated nanoindentation instrument, capable of making hundreds to thousands of indentations while maintaining the ambient humidity to an accuracy of +/- 0.1 % RH.
Figure 3 shows the basic configuration of the humidity control system. The Nano Hardness Tester (NHT), XYZ sample displacement stage and video optical microscope are housed within the inner chamber which is controlled by a closed loop air conditioning system and can be sealed at any moment with valves V2 and V1. These valves are opened to permit the inner chamber to acclimatize to a predefined humidity level, after which they are closed to seal the chamber during the nanoindentation load-depth cycle and prevent any movement of air which could disturb the ultra-sensitive experiment. The disadvantage of switching-off the flow of controlled air is that the humidity level will tend to drift during the test cycle, owing to losses via the chamber walls.
Figure 2. Schematic of the principle of the grid indentation technique for heterogeneous materials. Small indentation depths allow the determination of phase properties, while larger indentation depths lead to the response of the homogenized medium.
Figure 3. Schematic of the humidity system with independently controlled inner chamber and outer air jacket. The system operates over the ranges 10 – 40 °C and 30 – 90 % RH.
In order to prevent such losses, the inner chamber is surrounded by an outer air jacket which is independently controlled by temperature through valves V3 and V4. Maintaining a layer of air with stable temperature means that a predefined RH value can be kept constant for sufficient time periods (several minutes), sufficient for carrying out an indentation test.
The control valves V1 and V2 are operated between each indentation cycle, when running a large matrix of indentations, thus allowing hundreds of indentations to be made under environmental conditions that are perfectly controlled. The system works over the ranges 10 – 40 °C and 30 – 90 % RH as summarized in Figure 4. The actual dewpoint in the inner chamber is accurately monitored at all times in order to prevent condensation which could damage the most sensitive parts of the nanoindentation and XYZ positioning system. In addition, it is essential to prevent condensation as this would prevent high- quality imaging through the integrated optical microscope.
Figure 4. Operating range for the system shown in Fig. 3, showing the relationship between the temperature and humidity.
Figure 5. Inner chamber (a) showing the nanoindentation head (left) and the integrated optical video microscope (right) mounted over the automated XYZ sample displacement stage. The humidity and temperature controller is shown in (b).
The actual layout of the system is shown in Figure 5 and the main door is equipped with a sealed window for inspection and a glovebox for manipulating samples whilst maintaining controlled conditions. The XYZ sample displacement stage and the optical video microscope are both seperately enshrouded by water-resistant plastic housings in order to prevent any effect of condensation at high humidity levels.
Figure 6. Typical nanoindentation measurement file showing temperature and humidity signals superimposed over the load-unload cycle. This example shows an indentation on fused silica at 33 °C and 65% RH with maximum load 50 mN.
Figure 6 shows a typical nanoindentation measurement with corresponding temperature (T) and humidity signals recorded over the measurement period (65 seconds), establishing that both RH and T are perfectly maintained during the test cycle.
Sample preparation can be an issue with cementitious materials which comprise of extremely different constituent phases. Obviously, the indentation penetration depth must remain significantly smaller than the phase size for measuring the surface mechanical properties of individual phases. In reality, with depths typically < 200 nm, it is essential for the surface roughness of the polished section to be extremely low otherwise there will be significant standard deviation in the results.
Figure 7. Indentation hardness and modulus plotted as a function of relative humidity (35 – 90%) for 2 tobermorite samples.
Many concrete mixtures also contain air voids from air which gets trapped during mixing. These microscopic voids offer free space to ease hydraulic pressure when concrete freezes, otherwise it may crack. It may be essential to maintain the true void structure throughout the preparation process. Some investigators may actually impregnate the surface with a low viscosity epoxy so that the epoxy will readily penetrate the cracks and voids, thus preventing collapse of the void walls during polishing.
The actual polishing method will be determined by the constituent phases present and their relative wear rates which help in determining which abrasive particle size is most suited. If the abrasive paper, liquid medium and speed of polishing are not controlled sufficiently, then preferential polishing of softer phases will frequently occur, causing excessive roughness which will lead to errors in the mechanical properties measured.
Figure 8. Indentation hardness and modulus density groups for a tobermorite sample, summarized from a matrix of 300 nanoindentations (max load 5 mN). Note the presence of 3 phases.
Figures 7 and 8 show some typical nanoindentation results on a tobermorite sample; the former showing modulus and hardness as a function of relative humidity, the latter showing the Probability Density plots (histograms) of modulus and hardness. Note that three distinct phases are clearly distinguishable (a 3-Gaussian fit) from a nanoindentation matrix comprising of 300 equispaced indents.
Hydrated cement pastes can be confidentially divided into at least three significant mechanical phases: two hydrated phases and the remaining unhydrated clinker. The two hydrated phases are a low density C-S-H phase and a high density C-S-H phase. In Figure 8, the first peak has a Modulus (E) of 41 GPa and a Hardness (H) of 1.2 GPa, the second peak has H = 2.3 GPa and E = 62 GPa and the third peak has H = 7.8 GPa and E = 120 GPa.
The grid indentation technique together with the ability to measure properties at the nanoscale in a controlled humidity environment has been shown to offer a wealth of information in cementitious materials:
- The mechanical properties (H, E, packing density, volume fraction, etc.).
- Homogenized indentation modulus (calculated by integrating the surface volume fractions of each phase with their respective moduli).
- The porosity distribution within a sample.
Additional optimization of the technique will allow extraction of in situ mechanical properties at the micro and nanoscales, providing a suitable method of correlating individual phase properties with bulk response. For a more detailed explanation of the grid indentation technique and some examples on other types of materials (including metal composites), see Ref. 7.
Figure 9. Elastic modulus map (a) on a C-S-H sample for a 100 indent nanoindentation matrix with spacing between indents of 10 mm. A similar magnification Scanning Electron Microscope (SEM) image is shown in (b) for comparison.
Figure 10. Scanning Force Microscope (SFM) image of a nanoindentation in a C-S-H phase made to a depth of 500 nm using a Berkovich diamond indenter.
Prof. Franz-Josef Ulm (MIT Civil Eng.), Matthieu Vandamme (Ecole des Ponts, Paris) and Chris Bobko (North Carolina State Uni.) are acknowledged for sharing their interesting results.
1. R. J. –M Pellenq and H Van Damme, MRS Bulletin, (May 2004) 319 – 323
2. G. Constantinides, F. J. Ulm and K. J. Van Vliet, Materials and Structures, 36 (2003) 191 - 196
3. F. J. Ulm, G. Constantinides and F. H. Heukamp, Concrete Science and Engineering, 37 (2004) 43 – 58
4. G. Constantinides and F. J. Ulm, Cement and Concrete Research, 34 (2004) 67 – 80
5. G. Constantinides, K. S. Ravi Chandran, F. J. Ulm and K. J. Van Vliet, Materials Science and Engineering, A 430 (2006) 189 - 202
6. G. Constantinides and F. J. Ulm, Journal of Mechanics and Physics of Solids, 55 (2007) 64 – 90
7. N. X. Randall, M. Vandamme and F. J. Ulm, J. Mater. Res., Vol. 24, No. 3 (March 2009) 679 - 690
8. M. Miller, C. Bobko, M. Vandamme and F.-J. Ulm, Cement and Concrete Research 38 (4) (2008) 467-476
This information has been sourced, reviewed and adapted from materials provided by Anton Paar TriTec SA.
For more information on this source, please visit Anton Paar.