11.04 Strength of stone

Categories: Stone Cladding

Introduction
Strength can vary by several orders of magnitude between different types of stone as outlined in the table below, and to a lesser extent between stones of the same type.  Even within a single stone, measured strength can vary by a factor of 2 or more, depending on such variables as:

• Moisture content
• Sample orientation
• Type of test

Strength can also vary with time.
 

Note: This table shows a typical range of values of flexural strength of stones on buildings.  Stones of lower strength may be used from time to time.
 

Purpose of testing
This section deals specifically with strength testing.  However, it is important to bear in mind that the strength testing criteria are interdependent with durability and dimensional stability and all need to be considered together.

The requirement for strength testing may present itself in a number of different ways:

• An architect may wish to use a new and untried stone, for which there is no available test information
• Two or more stones may be visually suitable, and a decision on procurement may rest on selecting the one with the best engineering behaviour
• A particular fixing system, panel size and/or location on a building may impose unusually high loads on the stone; the long term strength of the stone will need to be verified in order to confirm its suitability
• Strength performance data provided by the supplier, often relate to the stone cut to a particular orientation;  for visual reasons the architect may wish to use the stone in a different orientation and the strength in this orientation should be checked
• The stone on an existing building appears to have failed by cracking at the fixing points; in order to establish the cause of failure, the strength of the stone on the building has to be checked
• A particular application demands a consistent high strength stone; strength testing may need to form part of the quality assurance check during production.

In considering the various ways in which the stone can be important to building construction, it is helpful to have in mind the diagram show here.  This shows a typical frequency distribution of the strength that generally occurs when a large number of identical tests are performed on the same stone.  Such variations result from intrinsic differences in physical and chemical bonding within the material and are influenced by moisture content, temperature, type of test and sample orientation.  The small number of recorded lower strength results highlights the significant loss of strength that can occur when discontinuities or weathering influence the test results.

It is important to bear in mind that published performance data on stone issued by suppliers will generally tend to accentuate the higher strength range of the material.  For design purposes, it is more appropriate to work to a lower strength and this is the methodology adopted within this document.

Where substantial and frequent low strength results are obtained a stone will normally be rejected, both by the procurer because it does not comply with specified performance criteria and by the supplier because it will be susceptible to breakage during production.

More problematical will be the intermittent, visually difficult to detect flaws, which may manifest themselves as an occasional low strength value in a set of otherwise acceptable test results.  It is often tempting to dismiss such low strength as spurious and irrelevant.  However, it should be always be borne in mind that the principal aim in performance testing is not to prove the fundamental strength of the material, but rather to establish its weakness.  On the building it will be primarily the weaknesses which gives rise to failures and which may only show up some years after the building is completed.

Accepting that it will rarely be practical or economically feasible to test a truly representative sample of the total area of stone on a building, and that it will be difficult to test the stone under true working conditions, the engineer is faced with the decisions of what type of tests to perform, when to undertake the testing and how many tests are needed.  Section 11.02 outlines a regime of preliminary and production tests and gives guidance on the number of tests to perform. Section 11.05 describes ways of analysing strength test results to give design values.
 

Index testing (preliminary testing)
Index strength tests include such tests as scratch hardness and geological hammer impact, which can be useful for differentiating between different types of strength category of the stone.

These tests are of particular value at an early stage in the selection of the stone, when carried out as part of an initial geological assessment at the quarry.  Many of the more significant features that impact on strength can be identified and tested at this stage.

When index testing is carried out under more controlled conditions in the laboratory, it can also be used as a means of rapid and inexpensive quality testing prior to or during production.  For example, where core samples can conveniently be taken prior to production at a particular quarry, point load testing on the core can be used to obtain an advanced measure of strength consistency.

If a correlation can be established between, for example, MOR strength and Brazil tensile strength, the latter can be used as proxy for the former during production quality testing.
 

Routine (standard) laboratory testing
Uniaxial compressive strength
Because of its simplicity, this was the first test to be widely applied for strength measurements on stone (first ASTM Standard 1941).  The test has also found wide application in rock mechanics (ISRM 1972) and in concrete testing (BS 1881:Part 118).
Over the years, many of the idiosyncrasies of the test method, in relation to sample dimensions and platen affects, have been well documented, including:

• Physical size of test sample (smaller samples generally give higher strength values;  size can be particularly important in relation to grain size/ crystal size)
• Platen end effects (the roughness of the loading surfaces and their parallelism will have an impact on the results)

Less well documented and understood is the apparently large variation in compressive strength of certain stone types, particularly sandstones, when tested wet and dry.  Wet compressive strength may be less than half of the dry strength, although the corresponding influence of moisture variation on MOR and flexural strength, in the same stone, may be more reserved.

The table below shows the range of compressive strengths of various natural stones that can be expected, and should be used as a guide when considering stone as an external cladding material.
 
 

Recent compressive strength testing on a range of sandstones, where stress and strain measurements have also been made, have revealed that the mode of failure appears to change, depending on the degree of saturation.  In wet conditions, failure occurs by vertical splitting, and is accompanied by an increase in Poisson’s Ratio.  In the dry condition, failure occurs by internal shearing on conical failure surfaces.
 

Modulus of Rupture Strength (MOR)
The MOR test essentially measures the tensile strength of a deep beam section in three point loading, image.  It is a commonly used strength test for cladding stone and the results are often used as the basis for design on modern stone cladding projects.  The standard test (ASTM C 99) uses a fixed size sample, the test being carried out wet and dry and parallel and perpendicular to the rift. There is also provision in the test standard for testing samples cut in different orientations to check for strength anisotropy, image.

It is a presumption of the test and of the determination of the MOR strength, that failure takes place beneath the central loading point along the neutral axis, where the applied stress is a maximum.  In uniform isotropic materials this would normally be expected to be the case.  However, many types of stone experience strength anisotropy (rift) and strength inconsistency.  Strength anisotropy can occur as a result of:

• Preferred orientation of minerals in igneous and metamorphic rocks, or of grains in sedimentary rocks
• Deposition layering in sedimentary (and some igneous) rocks where weak and strong minerals alternate in repetitive fashion
• Microstrain cleavage fabric, where rocks (particularly igneous and metamorphic) subjected to large tectonic stresses have suffered internal micro-cracking.

Strength inconsistencies include:

• Localised weathering (often of particular minerals)
• Intermediate incipient joints.

It is of interest to note that although the central-point loading MOR test was previously used for assessing the tensile strength of concrete both in the UK (BS 1881:Part 117) and USA (ASTM C 99) it has since been discontinued in favour of a four point loading arrangement (BS 1881:Part 118, ASTM C 880), similar to the flexural strength for stone.
 

Flexural strength
The flexural strength test (ASTM C 880) measures the strength of the slender beam in four-point loading as shown in this image.  The standard size of the specimen is 38mm wide, by 25mm thick, by 300mm long, and there is provision in the test for measuring strength on wet and dry specimens and on specimens cut parallel and perpendicular to the bedding or rift.

It is a feature of a standard specimen that the grain size of the stone can have an undue influence on the measured strength.  For instance, many coarse grained granites have large individual crystals (phenocrysts) which can occupy much of the total sectional area of the specimen.  It is now accepted practice when testing stone for use in cladding of a particular thickness, to match the thickness of the specimen to that of the cladding. The ASTM specification has recently been modified to allow for this, as well as for testing of specimens of different geometry.

The very nature of the flexural strength test produces a constant bending moment between the central loading points.  This loading arrangement inherently produces a greater probability of a weak element being subjected to the critical stress than when a central load acts (i.e. in the MOR test).  The flexural strength test is therefore more appropriate for testing stones which may have intermittent discontinuities or weaknesses.  A feature of very strong stones, such as granites, is that they are very stiff and brittle.  It requires very little deflection of larger panels or slender elements, for them to crack.  This compares to weaker and more deformable stones, such as porous sandstones, which can undergo considerable deflection with no sign of structural failure.
 

Specialist testing
There are other specialist laboratory tests that might be considered as a means for extending the data on strength for a particular stone.  These are more often used in rock mechanics and geotechnical testing, but their application is no less valid.
 

Deformation characteristics in uniaxial compression
This test involves measuring the stress-strain behaviour, Young’s Modulus and Poisson’s Ratio of a specimen subjected to uniaxial compression.  A test procedure is described by the ISRM (1972).

The test data can be of value in trying to understand the failure mechanisms and mode of failure in uniaxial compression, particularly in sedimentary rocks where moisture content, clay content and matrix content are likely to be important.

The standard ASTM flexural strength test (C 880) provides a suitable basis for measuring elastic modulus, provided that a dial gauge is used to measure the deflection at the mid-point for each increment of load.  ASTM C 120 gives a procedure.
 

Tensile strength
Indirect measurement of tensile strength may be made with the Brazil ‘split cylinder’ test and also with the MOR test.  However, for certain uses, for instance, where a slender stone element is to be rigidly fixed across a structural gap or to a structural member which may be liable to expansion, it may well be appropriate to measure tensile strength directly.  A suitable test procedure is described by the ISRM (1972).

It has to be recognised that stone is at its weakest in tension and that discontinuities such as joints, bedding, cleavage as well as grain and crystal orientations will have a significant influence over the measured strength on specimens cut with different orientations.  The test can therefore be used to determine the effect of these features on the strength of the stone.  However, because of the inherent difficulties in performing a direct tensile test, it will generally be more appropriate to use the MOR or Brazil ‘split cylinder’ test.
 

Shear strength
For most non-load-bearing applications, estimation of the shear strength from compressive strength data, or pull-out tests, will be adequate.  For certain load bearing applications, where there are discontinuities or planes of weakness, direct measurement of shear strength may be necessary.

A suitable test procedure is described by the ISRM (1972).
 

Creep
The process of creep can be defined as the gradually increasing permanent deformation of a material under stress.  A number of stone types, marble in particular and sedimentary rocks in general, exhibit a propensity for creep under load.  Marble and other limestones are well known for their tendency to deform under relatively low differential stress induced by, for example, moisture and temperature variations, self weight and wind loading.

Measurement of these effects can be undertaken by suitably modifying the test procedures previously described.
 

Fatigue and ageing behaviour
Many rock types, granites included, are known to suffer loss of strength through a combination of repeated stress fatigue and long term environmental exposure (ageing).  Thus in addition to the standard tests, other specific tests should be made to establish:

• The residual static flexural strength after the application of a sequence of fatiguing loads
• The static residual resistance of the stone around the fixing points after the application of a sequence of fatiguing loads

These tests are significant since the wind is variable and can therefore cause gradual damage to the stone over time.
 

Non-standard strength testing
Where assessment of strength-controlled performance under a variety of working conditions is required, it may be appropriate to undertake purpose design tests. Such tests might comprise:

• Pull-out tests on epoxy-cemented pin fixings
• Pull-out test on undercut anchors
• Lateral load tests on side-pin fixings
• Lateral load tests on kerf fixings
• Soft/hard body impact testing.

As with all such tests, it is important to conduct the tests at the working scale, on representative samples of stone which are sufficiently large to avoid edge effects. Nevertheless, care will still be required in interpreting the results and relating them to the working conditions.