th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

EFFECTS OF DEWATERING ON LONG TERM MOVEMENT

CHARACTERISTICS OF LIME MORTARS
Kioy, Stella1; Walker, Peter2; Ball, Richard3; Fodde, Enrico4; Peter, Ulrike5;
Lesueur, Didier6
1
2

Student, University of Bath, Architecture and Civil Engineering Department, S.M.Kioy@bath.ac.uk

PhD, Professor, University of Bath, Architecture and Civil Engineering Department, P.Walker@bath.ac.uk
3

PhD, Lecturer, University of Bath, Architecture and Civil Engineering Department, R.J.Ball@bath.ac.uk

PhD, Lecturer, University of Bath, Architecture and Civil Engineering Department, E.Fodde@bath.ac.uk

PhD, R&D Engineer Building Construction, Lhoist Recherche et Dveloppement, Ulrike.Peter@lhoist.com

6

PhD, Materials R&D Manager, Lhoist Recherche et Dveloppement, Didier.Lesueur@lhoist.com

The lower strength and stiffness of lime mortars, compared to cement mortars, can be
regarded as beneficial for the movement characteristics of the masonry. Traditional solid lime
mortared masonry typically have much fewer, if any, movement joints compared to higher
strength (but thinner) cement mortared walls. The modern revival in lime mortar use offers
the potential to revise specifications of movement joints in modern masonry construction.
However, there are presently little experimental data on which to base any design
recommendations. This paper summarises findings from an experimental study on the effects
of dewatering, by the brick during construction, on the movement characteristics of lime
mortar specimens. The specimens have been prepared for simulating dewatering effects of
bricks on mortar development using prismatic (40mm x 40mm x 160mm) and cylindrical
specimens (18mm diameter x 36mm long). The effects of dewatering on strength
development and creep performance are reported and supported by microstructural
development of materials. This research is part of a wider programme aimed at improving
understanding of movement characteristics of lime mortars and masonry looking at the effects
of different brick properties and mortar mix designs. In order to improve the understanding of
movement in lime mortars, the creep testing method is compared to those documented by
other researchers.
Keywords: hydrated lime, mortars, dewatering, creep

INTRODUCTION
Until recently little research had been carried out on lime based masonry mortars which has
coincided with recent increase in their use for new build and restoration work. Lime mortars
offer a lower impact alternative to cement based mortars. Building limes, produced by
burning limestone at around 900C, generally have less embodied carbon dioxide than
Portland cements, Hammond (2008). However, their performance characteristics compared to
cement mortars, such as improved permeability and beneficial low strength and stiffness, are
increasingly recognised as primary reasons for specifying lime mortars. Although lime

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

mortars acquire strength at a slower rate than cement mortars, their reported ability to
accommodate movement of structures could have benefits such as reduced number of
movement joints for example.
This paper presents results from a study on the effect of fresh mortar dewatering on the
strength and creep properties of lime mortar. Figure 1 illustrates a typical creep curve for
masonry which is characterised in three stages. Primary creep occurs upon application of the
load which then levels off to a much slower rate of deformation with no further increase in
load during the secondary creep stage. Eventually, the deformations will become significant
enough to cause load redistribution and failure as is characterised in the tertiary creep stage.
The whole creep process is a complex function of a number of varying mechanisms occurring
simultaneously, Ball et al. (2009), making creep difficult to predict. Masonry structures are
designed to remain in the secondary creep stage for their entire design life. It is therefore
essential to understand how much movement can be accommodated before failure occurs.

Figure 1: Typical creep progression curve for masonry

Previous studies on creep and shrinkage of mortars by Taves (1994), investigated the effects
of varying mortar composition, concluding that creep increased as the proportion of lime in a
mix increased. Ball et al. (2011) investigated the effects of carbonation on movement of lime
mortars, reporting that creep and shrinkage deformations occur in the lime matrix, first by
dissolution of calcium hydroxide crystals under stress then later by precipitation of calcium
carbonate in stress-free areas of the matrix.
EXPERIMENTAL PROCEDURE
A test programme was developed to investigate the creep and drying shrinkage of dewatered
and non-dewatered lime mortar cylinders. Testing was carried out on mortars produced using
a natural hydraulic lime (NHL3.5). The aggregate used was medium Binnegar mortar sand
supplied by Lime Technology with the particle size distribution is shown in Figure 2. A 1:3
(NHL3.5: sand) mortar mix was selected for the study.
Water sufficient to provide a flow of between 160 - 180mm was added to a mixture of dry
sand and lime, BS EN 1015-3:1999. All mixes were carried out in a mixer conforming to BS
EN 196-1:2005. The mixing procedure is based on that outlined for cement mortar mixes but

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

modified to suit lime mortars, with a total mix time of 5 minutes . Mortar cylinders (Figure 3)
were cast inside plastic moulds 18mm diameter by 36mm high.

Figure 2: Binnegar sand particle size distribution

Figure 3: Cylindrical Specimen

The moulds were filled in two layers, with each layer tamped five times with a 6mm diameter
Perspex tamper. Prisms, each measuring 40x40x160mm, were cast in steel moulds for the
non-dewatered specimens and in phenolic plywood moulds for the dewatered specimens. The
moulds were filled in two layers, each layer receiving 25 tamps with a Perspex tamper. To
dewater the mortar specimens a slice from an Autoclaved Aerated Concrete (AAC) block,
with an initial rate of suction of 1.5kg/m2.min, was placed on the top and bottom of the
moulds for 30 minutes (Figure 4). Japanese tissue paper was placed between the mortar and
masonry unit to prevent adhesion. This is a thin, strong paper that will transmit moisture
without absorbing it thus ensuring the dewatering of the mortar is only due to contact with the
block. After dewatering the base of the cylindrical moulds was then replaced; the plywood
moulds were placed onto oiled glass plates.

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

Figure 4: Dewatering of specimens

The specimens were allowed to harden in the moulds for 7 days under laboratory conditions
of 202C and 655% relative humidity. Following de-moulding specimens remained under
these conditions until testing. Cylinder specimens were capped using a thin layer of gypsum
based dental plaster, applied to the top and bottom to provide parallel and flat surfaces thereby
minimising localised stress concentrations during loading. Compressive strength was
determined using a 50kN Instron 3369 testing machine (Figure 5). Measurements were taken
at 14, 28, 56 and 91 days, with loading applied at rate of 0.5mm/sec. Six 40 mm cubes and
five 18 mm diameter cylinders were tested in each series.

Figure 5: Compression test set-up

The creep rig shown in Figure 6, designed for testing up to 8 specimens simultaneously, was
used throughout. A constant load is applied through a lever arm to each specimen and
Solatron transducers connected to a computer continually measure and record the movement.
Readings were initially taken at 60 second intervals up to 42 days, subsequently reduced to 10
minute intervals.

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

Figure 6: Creep test set-up

Eight specimens were placed in the creep rig at 16 days after casting. Four specimens were
loaded with 40% of their 14 day average compressive strength, whilst the other four were left
un-loaded. Details of loads applied are given in Table 1.
Table 1: Loads applied to specimens in creep rig
Specimen Number
Dewatered
Load Applied (N/mm2)
1
2
3
4
5
6
7
8

NO
NO
NO
NO
YES
YES
YES
YES

0.29
0
0.29
0
0.52
0
0.52
0

A Micrometics Autopore III was used for Mercury Intrusion Porosimetry (MIP) testing and a
JEOL JSM6480LV for Scanning Electron Microscopy (SEM) testing. The two techniques
were used to compare the pore developments in both types of specimens at different ages.
RESULTS
Dewatering of the fresh mortars with the AAC block produced strong adhesion within 30
minutes. Upon removal, the unit adhered to the mortar despite the use of Japanese tissue
paper, Figure 7.

Figure 7: Masonry unit after dewatering for 30 minutes

For both the prismatic and cylindrical specimens, the average compressive strength of the
dewatered specimens was higher than that of the non-dewatered specimens, as shown in
Figure 8. The average strength of the dewatered specimens was 1.4/mm2 at 14 days,
increasing to 1.7N/mm2, 2.0N/mm2 and 2.1N/mm2 at 28, 56 and 91 days respectively. For the

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

non-dewatered specimens averages were 0.5N/mm2 at 14 and 28 days increasing to 0.6N/mm2

and 0.7N/mm2 at 56 and 91 days respectively.

Figure 8: Average compressive strengths and carbonation front progression

The intensity and extent of the phenolphthalein staining decreased with the age of the
specimens indicating a reduction in the pH and therefore the proportion of calcium hydroxide.
A boundary between stained and unstained fracture surface regions, shown in the figure as
insets, indicate the extent of the carbonation front progressing from the surface to the centre
of the specimen. For dewatered specimens the carbonation front had progressed 2mm, 6mm
and 12mm at 14, 28 and 56 days respectively. Non-dewatered specimen carbonation front was
2mm, 7mm and 14mm at similar time intervals, as shown in Figure 8. For the cylindrical
specimens, the intensity of phenolphthalein staining decreased with age, however there was
no clear indication of the carbonation front.
The dewatered mortar cylinder specimens exhibited less shrinkage deformation when
unloaded, whilst under load they experienced more deformation compared to the nondewatered specimens, as shown in Figure 9(a). The non-dewatered specimens have a similar
strain progression in the loaded and unloaded condition with the two curves appearing
parallel. For the dewatered specimens, the two gradients were initially parallel but at around
14 weeks the gradient of the unloaded curve levelled off.

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

(a)

(b)
Figure 9: (a) Strain-Time and (b) Creep -Time graph of NHL3.5 mortar specimens

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

In the creep-time curves shown in Figure 9(b) the non-dewatered specimens exhibited
secondary creep for a longer period. The gradient of the non-dewatered creep curve is lower
than that for the dewatered specimens.

Figure 10: MIP results on mortars at 28 days after mixing

In Figure 10 the MIP results for 28-day old dewatered and non-dewatered mortar specimens
are compared. The dewatered specimens have a reduced volume of pores compared to nondewatered ones. The peak pore sizes in the non-dewatered specimens were ten times larger
than those in dewatered specimens (Figure 10).

Figure 11: SEM images (a) non-dewatered and (b) dewatered mortars at 56 days after
mixing
SEM images shown in Figure 11 corroborate MIP results with a higher number of pores
recorded on the non-dewatered specimens when assessed by eye. Pores in the dewatered
specimen show less interconnectivity whereas those in the non-dewatered specimens form
extended pore channels along the edge of the aggregate particles.

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

DISCUSSION
The effect of dewatering on NHL3.5 mortar has been observed to increase compressive
failure stress. A similar effect on mortars after dewatering was also observed by El-Turki et
al. (2010). Dewatering however does not appear to improve the rate of carbonation of a
mortar with the carbonation depths measured for non-dewatered being slightly more than
those on dewatered specimens. From SEM images dewatered specimens have a reduced
number of interconnected pores. This could account for a reduced rate of carbonation
observed. However, due to the effect of dewatering, carbonation can commence sooner as
mix water is drawn out of the mortar, freeing pores spaces for the absorption of carbon
dioxide. Saturated pores slow down the absorption of carbon dioxide, which is 10,000 times
slower in water than in air, Van Balen (2005).
The increased strength observed due to dewatering can be attributed to changes in pore
structure of the mortars. Pore size as well as pore volume decrease when a mortar is
dewatered. Consolidation of the mortar matrix has been thought to account for the improved
strength by El-Turki et al. (2010). Aggregate particles transfer load through a mortar
therefore by consolidating the mortar matrix, the aggregate articles are brought closer together
and thus are capable of sustaining higher loads than in non-dewatered specimens where a
larger proportion of the load must go through the binder accounting for the lower strength
capability.
Dewatered specimens were observed to deform more under load than the non-dewatered
specimens. As these specimens attained higher compressive failure strengths, this movement
can be accounted for by the higher stress on the specimens. Chemically unbound pore water
is readily lost to the atmosphere by diffusion until equilibrium is reached between the pore
water content and atmospheric humidity levels. In the unloaded condition, non-dewatered
specimens shrink more than the dewatered ones as they have more chemically unbound pore
water. At 12 weeks diffusion equilibrium appears to have been reached in the unloaded
dewatered specimens but not in the non-dewatered specimens. Forth et al (2000) found that
the interaction of brick units with mortar in model wall tests, reduced shrinkage in the mortar
joints.
Both dewatered and non-dewatered specimens have a similar duration of the primary creep
stage with the dewatered specimens experiencing a shorter and faster secondary stage
demonstrated by the higher gradient. As applied stress increases, the gradient of the creep
curve also increases as illustrated in Figure 1. Similar results were observed by Ball et al
(2010) where increasing stress levels on a specimen increased the percentage strain recorded.
CONCLUSION
The results presented demonstrate that dewatering increases the compressive failure stress of
lime mortars by reducing the number and size of pores, consolidating the binder and bringing
aggregate particles closer together.
In the dewatered state mortar is capable of
accommodating three times more load-dependent deformation without visible cracking. This
might lead to conservative design if non-dewatered performance data are used. Nondewatered specimens do not accurately portray the in-situ conditions of a mortar and therefore
lead to underestimation of mortar properties.

th

15 International Brick and Block

Masonry Conference
Florianpolis Brazil 2012

ACKNOWLEDGEMENTS
The authors wish to acknowledge Lhoist Research and Development SA for their financial
support and the University of Bath staff, in particular Brian Purnell, Will Bazeley, Graham
Mott, Sophie Hayward and Chris Arnold for their technical assistance.
REFERENCES
Ball, Richard J., El-Turki, Adel, and Allen, Geoffrey C. (2011), 'Influence of carbonation on
the load dependent deformation of hydraulic lime mortars', Materials Science and
Engineering: A, 528 (7-8), 3193-99.
Ball, Richard J., El-Turki, Adel, Allen, William J., Nicholson, John A. and Allen, Geoffrey
C. (2009), 'Deformation of NHL3.5 and CL90/PC hybrid mortars', Proceedings of the
Institution of Civil Engineers Construction Materials (162), 29-35.
Ball, Richard James and Allen, Geoffrey Charles (2010), 'Load-dependent deformation and
shrinkage in hydraulic lime mortars', International Journal of Sustainable Engineering, 3 (1),
40-46.
British Standards Institute (1999), 'BS EN 1015-3:1999 Methods of test for mortar for
masonry Part 3: Determination of consistence of fresh mortar (by flow table)'.
British Standards Institute (2005), 'BS EN 196-1:2005 Methods of testing cement Part 1:
Determination of strength'.
El-Turki, Adel, Ball, Richard J., Carter, Margaret A., Wilson, Moira A., Ince, Ceren, Allen,
Geoffrey C. (2010), 'Effect of Dewatering on the Strength of Lime and Cement Mortars',
Journal of the American Ceramic Society.
Forth, J. P., Brooks, J. J., and Tapsir, S. H. (2000), 'The effect of unit water absorption on
long-term movements of masonry', Cement & Concrete Composites, 22 (4), 273-80.
Hammond, G. (2008), Inventory of Carbon & Energy: ICE.
Taves, L. B. S. (1994), 'An Analysis of Creep and Shrinkage in Mortars', Journal of Testing
and Evaluation, 22 (6), 548-55.
Van Balen, K. (2005), 'Carbonation reaction of lime, kinetics at ambient temperature', Cement
and Concrete Research, 35 (4), 647-57.