Materials and Design 52 (2013) 514–523

Contents lists available at SciVerse ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Compressive strength development of binary and ternary lime–pozzolan

mortars
E.R. Grist a,b,⇑, K.A. Paine b, A. Heath b, J. Norman c, H. Pinder a
a
Ramboll, 40 Queen Square, Bristol BS1 4QP, UK
b
University of Bath, Claverton Down, Bath BA2 7AY, UK
c
University of Bristol, Bristol BS8 1TR, UK

a r t i c l e i n f o a b s t r a c t

Article history: This study considers the compressive strength development of broad range of hydraulic lime mortars
Received 20 February 2013 prepared with a range of commercially available alumino-silicate by-products and modern pozzolanic
Accepted 4 May 2013 additions. Specifically this paper considers the effect of mineral addition selection, binary and ternary
Available online 15 May 2013
combinations, pozzolan content and the effect of curing conditions on the compressive strength develop-
ment of hydraulic lime based mortars. The study was undertaken as the initial phase of a broader inves-
Keywords: tigation considering the feasibility of producing modern, sustainable hydraulic lime–pozzolan concretes
Hydraulic lime
with comparable strengths to Portland cement based concretes. The aim of the initial phase was to iden-
Mortars
Pozzolanic efficacy
tify a small number of additions, and combinations thereof, which would result in a structural strength
Ternary combinations lime–concrete when scaled up from mortars to concretes.
Compressive strength In the absence of a definitive source of information on the mechanical properties of hydraulic-lime
Curing mortars prepared with binary and ternary combinations of alumino-silicate by-products, 22 combina-
tions consisting of Natural Hydraulic Lime (NHL5) and a range of possible additions, identified from his-
torical and current practice, were prepared. The results have shown that combining an eminently-
hydraulic NHL5 with silica fume and ground granulated blastfurnace slag can produce mortars with a
28-day compressive cube strength of around 28 N/mm2, at a water-to-binder (w/b) ratio of 0.5. This is
eight times the strength of an equivalent mortar prepared with NHL5 alone and broadly speaking com-
parable with that of low-heat cementitious mortars. The contribution of the pozzolanic reaction to the
strength of hydraulic lime mortars is discussed for a range of alumina-silicious materials and combina-
tions thereof.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction eral by-products, amongst others, have been classified as Type II

Concern about the harmful environmental impact of global ce-
ment manufacture has prompted extensive research into the
search for and development of more sustainable binders for con-
struction. As a result there has been a resurgence of interest in
building limes, which when produced at a large enough scale with
the same production efficiencies as cement can, and in some cases
do [1], demand less energy and emit less carbon dioxide in
manufacture.
In looking to minimise the environmental impact of specified
concretes, current best practise promotes the use of alumino-sili-
cious by-products such as, silica fume, fly-ash and ground granu-
lated blastfurnace slag, as partial clinker replacement materials
or as additives to improve aspects of performance [2]. These min-
⇑ Corresponding author at: Ramboll, 40 Queen Square, Bristol BS1 4QP, UK. Tel.:
+44 (0) 1179 295 200.
E-mail address: ellen.grist@ramboll.co.uk (E.R. Grist).
additions, having been shown to enhance the properties of Port-
land-cement based concretes due to their pozzolanic or latent-
hydraulic properties [3–5].
The use of pozzolanic minerals to enhance the properties of
cementitious binders long pre-dates the invention of Portland ce-
ment. Ancient civilisations are known to have used naturally
occurring pozzolanic materials in conjunction with lime and this
was, until the 1850s, the principal binder for use in mortars and
concretes.
Analysis of ancient mortars, dated to 7000BC, shows evidence of
the inclusion of pozzolanic materials, but chemical analysis alone
cannot conclude whether or not these materials were added delib-
erately in the knowledge of their beneficial effects [6]. The art of
‘‘gauging’’ lime mortars with natural pozzolanic materials was cer-
tainly appreciated by the ancient Greeks, who created hydraulic
lime-mortars by utilising the volcanic ash that was deposited on
the Island of Santorini by the eruption of Thera in 1500BC [7].
Lime–pozzolan technology was subsequently widely disseminated

0261-3069/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.matdes.2013.05.006
 E.R. Grist et al. / Materials and Design 52 (2013) 514–523 515

across the known-world under Roman occupation. Following the
collapse of the Roman Empire in the fifth-century AD, knowledge
of lime technology in Western Europe was gradually lost until its
renaissance in 1755–1759 when English civil engineer John Sme-
aton, conducted extensive research into the hydraulic properties
of lime–pozzolan binders.
Smeaton’s work was shortly followed by relatively rapid ad-
vances in binder technology, which led to the development of Port-
land cements, which almost entirely replaced lime-based binders
due to their superior strength and speed of set. Only relatively re-
cently has damage to the fabric of historic buildings ‘repaired’ with
cement-based building materials highlighted the incompatibility
of impermeable, rigid, cement-based materials with their historic
lime-based counterparts [8]. Furthermore it is recognised that
the high strengths attained by Portland-cement based materials
are unnecessary in many applications. Instead it is considered
appropriate in modern practice to mitigate the environmental im-
pact of concretes by specifying Portland-composite cements, in
which a substantial proportion of the Portland cement is substi-
tuted by secondary materials, despite this in some cases resulting
in reduced strengths [2].
Unlike the majority of current research into lime-mortars for
masonry and render applications, the guiding philosophy of this
research project was on the future potential of lime-binder tech-
nologies in concretes. Consciously deviating from the current wis-
dom on the appropriate specification of flexible and breathable
lime-technologies, in both restoration and new-build construction
projects, the aim of this experimental investigation was a compres-
sive strength comparable to that of Portland cement based mixes.
As a result this work is unique in considering the effect of combin-
ing NHL5 with modern Type II additions familiar in modern con-
crete technology. It is intended that by creating a modern, high-
performance lime–pozzolan binder, and pushing the boundaries
of lime technology, a broader range of lime-based materials could
then be developed, with further reduced environmental
performance.
The primary purpose of this study was to investigate the feasi-
bility of producing a modern structural-strength lime–pozzolan
concrete. Within this broader project context this study had three
specific aims:
(i) To investigate the range of compressive strengths attainable
by hydraulic lime-mortars prepared with modern alumina-
siliceous additions, or combinations thereof.
(ii) To identify a small number of hydraulic lime–pozzolan com-
binations with the potential capability to achieve 28-day
compressive strengths of 30 N/mm2, or above, when scaled
up to lime–pozzolan concretes.
(iii) To investigate the sensitivity of hydraulic lime–pozzolan
binders to curing conditions.
Although cement technologists have demonstrated that ternary
combinations of alumnio-silicate materials in combination with
Portland cement, can have a synergistic effect on resulting proper-
ties [9], it is acknowledged that further research is needed to
understand the mechanisms by which these materials interact
[7]. Given that many of these supplementary cementitious materi-
als are activated by the calcium hydroxide, Ca(OH)2, in Portland-
cement based systems, their activity with hydraulic lime is of
broader interest.
2. Experimental method
The experimental programme comprised the production, curing
and compressive strength testing of hydraulic-lime mortars.
2.1. Materials
The NHL5 used was a natural hydraulic lime manufactured in
France and supplied by a specialist lime-building merchant in the
UK. Eight alumino-silicate mineral additions were used in this
study. Silica Fume (SF), conforming to BS EN 13263 [10], was ob-
tained in the form of a slurry, with a SF:water ratio of 50:50 by
mass. The Ground Granulated Blastfurnace Slag (GGBS) and Fly
ash (FA) conformed to BS EN 15167 [11] and to BS EN 450 [12]
respectively. Three forms of metakaolin, with varying pozzolanic
indices and surface areas, were used and these were designated
MK1, MK2 and MK3. Finally, for comparison, two red-brick dusts
were supplied by a stockist of traditional building materials; one
an English Red Brick Dust (ERBD) and the other an Italian Red Brick
Dust (IRBD). The oxide compositions of these materials are shown
in Table 1.
Two fine aggregates were used in this investigation. All except
one of the mortars were prepared with Guiting dust, limestone
fines from Guiting Quarry, in Gloucestershire. The particle size dis-
tribution (PSD) of this fine aggregate was determined in accor-
dance with BS 933-1:2012 and is shown in Fig. 1 [13]. One
mortar was also prepared using a CEN standard sand as a point
of reference, when considering the resulting mortar strengths.
The PSD of the CEN standard sand, in accordance with BS EN 196
[14], is also shown for comparison.
Alongside the binary and ternary lime combinations, three con-
trol samples were prepared, against which the resulting mechani-
cal properties could be benchmarked. These control mortars were:
 NHL5 only, prepared with a CEN Standard Sand in accordance
with BS EN 196-1 [14].
 NHL5 only, prepared with Guiting dust.
 NHL-PC, a commercially available formulated lime mortar con-
sisting of NHL and Portland cement (PC).
2.2. Sample preparation
Mortars corresponding to each of the 22 combinations were
prepared in a paddle mixer in the ratio 2 parts binder:3 parts
sand:1 part water (450 g binder, 1350 g sand, 225 g water), in
accordance with BS EN-196-1 [14], as shown in Table 2. Despite
the demand for water varying depending on the water absorption
characteristics of the materials in each combination, the mass of
water was kept constant according to BS EN-196-1. This allowed

Table 1
Oxide analysis of constituent mineral additions.

NHL5 NHL5–PC SF GGBS FA MK1 MK2 MK3 IRBD ERBD

Oxide analysis (weight %)
SiO2 15.0 16.0 94.5 33.0 53.0 55.0 55.0 55.0 23.7 38.4
Al2O3 1.9 2.1 0.3 14.0 30.0 41.0 40.0 39.0 7.7 7.3
K2O + Na2O 0.3 0.3 1.3 0.8 0.7 2.0 0.8 1.0 3.3 1.6
Fe2O3 0.6 1.1 0.3 0.4 7.0 0.6 1.4 1.8 5.7 8.5
TiO2 0.2 0.2 0.0 0.0 1.5 0.0 1.5 1.5 0.0 0.8
CaO + MgO 60.0 60.9 0.8 47.0 4.0 0.2 0.3 0.6 10.1 3.0
516 E.R. Grist et al. / Materials and Design 52 (2013) 514–523

Fig. 1. Particle size distribution of the two fine aggregates.

Table 2
Mortar mix details.

Mix description NHL5 Addition Fine aggregate Water

g g g g
100% NHL5 450 0 1350 225
100% NHL5 (CEN) 450 0 1350a 225
100% NHL–PC <450b 0 1350 225
Additions at 15% (IRBD, ERBD, FA, MK1, MK2, MK3, GGBS, SF) 382.5 67.5 1350 225
Additions at 30% (IRBD, MK1, SF) 315 135 1350 225
15% MK1 and 15% FA 315 67.5 + 67.5 1350 225
15% MK1 and 15% GGBS 315 67.5 + 67.5 1350 225
15% SF and 15% FA 315 67.5 + 67.5 1350 225
15% SF and 15 % GGBS 315 67.5 + 67.5 1350 225
25% MK1 and 25% FA 225 112.5 + 112.5 1350 225
25% MK1 and 25% GGBS 225 112.5 + 112.5 1350 225
25 SF and 25% FA 225 112.5 + 112.5 1350 225
25% SF and 25% GGBS 225 112.5 + 112.5 1350 225
a
All the fine aggregate was Guiting dust except which was CEN standard sand.
b
A formulated lime mortar containing PC in accordance with BSEN 459-1 [15].

the resulting compressive strengths of the combinations to be
compared at a single water-to-binder (w/b) ratio of 0.5, although
it resulted in mortars of varying consistence. Each mortar sample
was cast into metal moulds, producing three identical prisms with
dimensions of 40 mm  40 mm  160 mm. The moulds were pre-
pared with a light coating of mineral oil to facilitate de-moulding.
90 days and the results are in Table 3 for the water-cured speci-
mens. A few of the key results are summarised in Fig. 2 to emphasis
the range of observed strengths.
3.1. Pozzolanic activity

Three main properties describe the activity of pozzolanic

2.3. Curing
materials:
Specimens were cured both in air and in water to assess the
(i) how soon the pozzolanic material reacts with the calcium
influence of curing on compressive strength development. In both
hydroxide (Ca(OH)2)
cases the mortar samples were cured in a sealed polythene bag for
(ii) the rate of the reaction with Ca(OH)2
24 h before demoulding. Then the lime–mortar specimens were di-
(iii) the total consumption of Ca(OH)2
vided into two groups for curing. Air-cured mortar bars were cured
in a conditioning lab maintained at 20 ± 0.5 °C and 60–65% RH.
There are a number of standard tests that seek to define and
Water-cured mortar bars were immersed in a water bath main-
quantify the pozzolanic activity of various natural and man-made
tained at 20 °C.
pozzolans. The tests vary not only in scope and duration, but also
conceptually – considering pozzolanicity as a property, a process
3. Results and discussion and/or an outcome. The Chapelle test is a material-property test
that is used to determine the overall consumption of Ca(OH)2,
Compressive strength was measured in accordance with BSEN measured in mg Ca(OH)2/g. This 16-h test is accelerated by elevat-
196-1 [14] for these standard sized samples at 7, 28, 56 and ing the temperature of the reaction. This figure is quoted by some
 E.R. Grist et al. / Materials and Design 52 (2013) 514–523 517

Table 3
Compressive strength development of water-cured lime–pozzolan mortars.

Compressive strength (N/mm2)

Constituents Age 7 28 56 90
100% NHL5 1.6 3.4 4.8 6.4
100% NHL5 (CEN) 2.0 4.9 7.9 9.7
100% NHL–PC 10.8 12.7 13.2 12.8
85% NHL5
15% IRBD 1.0 3.2 5.3 7.3
15% ERBD 1.3 3.5 5.3 6.1
15% FA 1.6 4.7 7.9 10.3
15% MK1 3.2 5.9 8.5 7.2
15% MK2 2.9 7.0 9.2 10.4
15% MK3 6.3 7.5 8.7 10.9
15% GGBS 2.4 9.3 11.1 13.5
15% SF 4.9 13.3 18.5 19.7
70% NHL5
30% IRBD 1.0 3.2 5.3 7.3
30% MK1 2.1 4.5 5.0 6.9
Fig. 2. Development of compressive strength for key water-cured mortars.
15% MK1 and 15% FA 6.7 13.4 15.5 18.2
15% MK1 and 15% GGBS 3.8 14.3 15.7 20.3
15% SF and 15% FA 4.5 15.1 20.6 24.5
was the approach adopted in BS 3892: Part 1 (1993), which defined
15% SF and 15% GGBS 7.7 20.8 24.6 31.0
30% SF 6.5 28.7 34.3 37.6 the Strength Factor of a type II addition as the ratio of the 28-day
strength of a mortar containing a fixed proportion of addition, with
50% NHL5
25% MK1 and 25% FA 4.8 11.4 10.7 14.0 that of a control mortar [17]. Using this simple analysis approach
25% MK1 and 25% GGBS 7.2 13.2 14.8 18.8 on the results in Table 3 showed that the replacement of 50%
25% SF and 25% FA 6.1 19.9 25.3 29.8 NHL5 with a combination of 25% SF and 25% GGBS, resulted in a
25% SF and 25% GGBS 6.9 28.0 31.5 38.2 mortar with a 28-day compressive strength around 8 times as
strong as a mortar prepared with NHL5 only.
A similar approach was adopted by Yu et al. who defined a
suppliers of pozzolanic materials as a measure of pozzolanicity. number of formulas for calculating the contribution of the pozzo-
Alternatively continuous electrical-impedance measurements have lanic effect to the overall strength of cement mortars [18]. The
been used to monitor the progression of pozzolanic reactions by same outcome-orientated approach has been adopted in this study,
direct activation with calcium hydroxide [16] and this process-ori- which considers the (beneficial) effect of the range of pozzolanic
entated approach is used to model early age reaction kinetics. Such materials, and combinations thereof, on the compressive strength
tests are good at describing the onset and rate of reactions and development of the hydraulic lime mortars.
MacCarter and Tran present a formula for combining these two fac- The effectiveness of each of the pozzolanic materials has been
tors into a pozzolanicity index (PI) [16]. analysed by calculating the Pozzolanic Index, PI (%), which repre-
An arguably more holistic approach is to consider the outcome sents the favourable contribution of the pozzolan to compressive
of inclusion of a pozzolanic material on a cementitious system. This strength, relative to the binder. This study has adopted the

Table 4
Pozzolanic efficacy (%) of a range of additions and combinations at 28-days.

Lime Mix Mortars cured in water Mortars cured in air

% constituents
Average 28d Specific PE (%) Specific strength Average 28d Specific PE Specific strength
compressive strength strength of the (water) of the addition, S compressive strength strength of the (%) of the addition, S
(N/mm2) mortar, K (N/mm2) mortar, K (air)
100 100% NHL5 3.43 1.00 0 n/a 3.23 1.00 0 n/a
85 IRBD 15% 3.29 1.13 11 0.7 2.29 0.83 20 0.9
ERBD 15% 3.47 1.19 16 1.1 3.41 1.24 19 0.9
FA 15% 4.65 1.59 37 3.4 4.14 1.51 34 1.4
MK1 15% 5.88 2.02 50 5.8 3.20 1.17 14 2.9
MK2 15% 6.96 2.39 58 7.9 4.60 1.67 40 3.8
MK3 15% 7.55 2.59 61 9.0 8.01 2.92 66 7.6
GGBS 15% 9.34 3.20 69 12.5 6.42 2.34 57 10.9
SF 15% 13.26 4.55 78 20.1 9.01 3.28 70 12.9
70 IRBD 30% 3.19 1.33 25 0.8 3.11 1.37 27 0.1
MK1 30% 4.48 1.86 46 2.0 2.32 1.03 3 0.9
MK1 15% and FA 15% 13.44 5.60 82 10.7 7.01 3.10 68 4.9
MK1 15% and GGBS 15% 14.33 5.97 83 11.6 10.71 4.74 79 5.3
SF 15% and FA 15% 15.13 6.30 84 12.4 7.35 3.25 69 8.5
SF 15% and GGBS 15% 20.81 8.66 88 17.9 10.48 4.64 78 8.7
SF 30% 28.70 11.94 92 25.5 14.58 6.45 84 12.7
50 MK1 25% and FA 25% 11.36 6.62 85 5.6 4.65 2.88 65 1.9
MK1 25% and GGBS 25% 13.15 7.66 87 6.7 9.49 5.87 83 4.2
SF 25% and FA 25% 19.91 11.60 91 10.6 8.48 5.25 81 4.9
SF 25% and GGBS 25% 28.04 16.34 94 15.3 11.48 7.11 86 6.1

Note the mixes have been ordered within each % dosage by PE (%) in water.
518 E.R. Grist et al. / Materials and Design 52 (2013) 514–523

definition of Pozzolanic Index used by Yu et al. [18], although it is
suggested that this property might more clearly be referred to as
‘Pozzolanic Efficacy’, PE (%), as it describes the ability of the pozzo-
lan to produce the desired result and also differentiates it from the
material property resulting from the Chapelle test. This proposed
terminology has been adopted throughout the rest of this paper.
Note that this definition of pozzolanic efficacy, describes only the
effectiveness of the pozzolanic addition in increasing the mechan-
ical strength of the resulting mortar. It is recognised that addition
of pozzolanic materials may result in other ‘beneficial’ mortar
properties, but that these are beyond the scope of this paper. Given
that the Pozzolanic Efficacy (%) also varies through time as the
reaction progresses, the test-age should also be designated when-
ever this figure is stated e.g. PE(%)28d.
Following Yu et al.’s nomenclature, Rc is defined as the com-
pressive strength of the control mortar prepared with 100% pri-
mary binder (in this case NHL5). The measured compressive
strength of the mortars prepared with mineral additions (fc)
are divided by the proportion of primary binder in the mix to
give a comparative mortar strength Rm, (Rm = fc/%lime), thus Rm
is a scaled measurement of the mortar compressive strength,
in which the contribution of the lime to the overall compressive
strength is equal to Rc in every case. Hence the contribution of
the pozzolanic effect, due to the mineral admixture, to the total
strength can be calculated from Rp = Rm  Rc and this can be ex-
pressed as a percentage of the total strength using PE(%) = Rp/
Rm  100. The Pozzolanic Efficacy (%) of the tested mortars is
presented in Table 4. An alternative, but equivalent method of
comparing the mortars is to use K, the ‘specific strength’ defined
by Yu et al. as Rm/Rc.
The specific strength of the addition, denoted S, has also been
defined. This is the ratio of the strength due to the pozzolanic

Fig. 3. Variation in the Pozzolanic Efficacy (%) of a range of alumina-siliceous additions.

 E.R. Grist et al. / Materials and Design 52 (2013) 514–523
action of the mineral addition relative to the strength of the
equivalent proportion of primary binder it replaces. For example
the water-cured mortar prepared with 15% FA has a specific
strength of 3.4. This means that the 15% of the FA in the mortar
has contributed 3.4 times the strength gain that the equivalent
15% lime would have done, were it not indeed replaced. If addi-
tions, at a given % dosage, result in a compressive strength less
than that of the control mortar then the specific strength of the
addition is less than 1.0, as in the case of IRBD at both 15% and
30% in water-cured samples. If the specific strength of the addition
is less than 0.0 then its inclusion has an adverse impact on the
strength gain of the lime. This is the case of the use of 15% IRBD
in the dry-cured lime mortar.
In Table 4 the mortars are grouped according to the percent-
age dosage of alumina-silicate additions: 0%, 15%, 30% and 50%.
Within each grouping the mortar mixes are ordered according
to the relative effectiveness of the binary and ternary combina-
tions as described by their pozzolanic efficacy when cured in
water, as shown in Table 3. Note that a different pozzolanic effi-
cacy, and thus order, results from the analysis of air cured mortar
strengths.
The results show that, of the additions used, IRBD is the least
and silica fume the most effective pozzolanic addition, for both
air and water cured hydraulic lime mortars. This result is consis-
tent with industry understanding. Previous studies have shown
that not all brickdusts can be classified as pozzolanic; rather poz-
zolanicity has been shown to depend on the clay type, firing tem-
perature and fineness of the ground brick [19,20]. Silica fume at the
other end of the scale is widely appreciated to be a highly reactive
pozzolanic material within the concrete industry [5,21]. When
hydraulic lime mortars are cured in water the order of preference,
based on pozzolanic efficacy, for the addition of alumino-silicate
by-products, has been shown to be:
SF ! GGBS ! MK ! FA ! BRICKDUST
High efficacy Low efficacy
The proprietary metakaolin MK3 consistently outperformed
MK1 and MK2. While both MK2 and MK3 are dehydroxylated alu-
minium silicates, produced from calcination of kaolinitic clay from
the Charentes basin in France, MK3 has a higher surface area and
Pozzolanic Index than MK2. The reported surface area of MK2
and MK3 are 17,000 m2/kg and 19,000 m2/kg respectively, and
the reported Pozzolanic Indexes of the two metakaolins, measured
by the Chapelle test, are 1100 and 1400 (mg Ca(OH)2/g) respec-
tively [22], this explains the improved performance of MK3 as a
pozzolanic addition. Consistent with this analysis, MK1 was ob-
served to have the lowest pozzolanic efficacy at 28-days, it also
had the lowest surface area (14,000 m2/kg) and Pozzolanic Index
(1000 (mg Ca(OH)2/g)) [22].
At 28-days the contribution of the pozzolanic reaction to the
overall mortar strength was greater when samples were cured in
water than when cured in air. With the pozzolanic reaction known
to be dependent on the presence of moisture it is clear that an
inadequate supply of water is going to have a detrimental effect
on its overall contribution.
The results in Table 4 present the calculated Pozzolanic Efficacy
(%) at a single age (28-days). Plotting the variation of the Pozzola-
nic efficacy, over time gives a graphical insight into the relative
contribution of the different pozzolanic additions as the reaction
progresses, as shown in Fig. 3a–f.
Considering Fig. 3a, IRBD is seen to have a detrimental effect on
the 7-day hydraulic lime mortar strength having with a negative
Pozzolanic Efficacy (%) at 7-days. By 28-days the inclusion of the
brickdust is observed to improve the compressive strength. Within
3 months the contribution of the pozzolan reaches a maximum le-
vel at which it plateaus. The same shaped curve is seen at a dosage
519
of 15% and 30% but the higher dosage attains a higher total
contribution.
This analysis demonstrates that silica fume makes a very sub-
stantial contribution to compressive strength at an early age, great-
er than 70% of the 7-day strength at a dosage of 15% and greater
than 80% of the strength at a dosage of 30%.
MK1, MK2 and MK3 all exhibit a substantial early contribution
to compressive strength (48–80%), but this contribution then re-
duces relative to the contribution of the lime until 3-months, at
which point its contribution tends to plateau, in the case of MK1
at around 25%. This high initial reactivity is consistent with the
work of others [23,24], where it is attributed to the hydration of
the large proportion of alumina (Al2O3) phases present in MK as
seen in Table 1. Increasing the dosage of MK1, from 15% to 30%,
is seen in (c) to have a detrimental effect on its contribution to
strength at early ages.
3.2. Ternary combinations
Fig. 4 looks at the effect of using two alumino-silicate additions
in a ternary combination with hydraulic lime.
All four of these examples demonstrate a complimentary effect
when pozzolans are used in ternary combinations. MK is shown to
be beneficial in combination with GGBS and with FA. In both cases
the MK can be seen to contribute to the early age performance of
these ternary combinations. The complimentary effect is specifi-
cally evident in the case of a ternary combination of MK and FA,
Fig. 4a. Increased compressive strengths resulting from ternary
combinations of pozzolanic materials in PC-systems are the result
of the complex interaction of chemical and physical effects [25,26].
A ternary combination of SF and GGBS has been shown to result
in the greatest overall pozzolanic efficacy (%), attaining a maxi-
mum of 94% at 28-days, see Fig. 4d. On the strength of these results
it was decided to undertake a subsequent study (beyond the scope
of this paper) to investigate the properties of lime–pozzolan con-
cretes based on the following ternary combinations:
 NHL5 + SF + GGBS
 NHL5 + SF + FA
 NHL5 + FA + MK
3.3. Effect of the total content of a range of ternary combinations
Fig. 5 presents the variation of Pozzolanic Efficacy (%) when the
proportion of pozzolan in ternary combination is increased from
30% to 50% of the total binder.
It can be seen in Fig. 5 part (a) that increasing the dosage of the
ternary combination of MK + FA from 30% to 50% of the total binder
was seen to have no effect on its overall Pozzolanic Efficacy (%) at
any stage. An increased dosage will reduce the environmental im-
pact as both MK and FA have a lower embodied CO2 than NHL5. In
all of the other three cases, this increase in dosage is seen to mar-
ginally increase the Pozzolanic Efficacy (%).
The above demonstrates the potential for lime–pozzolan bind-
ers utilising a substantial proportion of secondary materials. The
favourable compressive strength results attained by lime–pozzolan
mortars containing a substantial dosage of silica fume (30% by
mass), informed the subsequent investigation into the properties
of hydraulic lime–silica fume concretes.
3.4. Effect of curing
These results have demonstrated a substantial difference in the
compressive strength of air-cured and water-cured hydraulic lime
mortars. The strength development of hydraulic-lime–pozzolans is
the result of four chemical reactions:
520 E.R. Grist et al. / Materials and Design 52 (2013) 514–523

Fig. 4. Variation in the pozzolanic efficacy (%) of four ternary combinations.

Fig. 5. Variation in the Pozzolanic Efficacy (%) with total content of additions.
 E.R. Grist et al. / Materials and Design 52 (2013) 514–523 521

(1) Pozzolanic reaction of the alumina-silicates with the initial Table 5

free lime (calcium hydroxide). Compressive strength development of dry-cured lime–pozzolan mortars.

(2) Hydration of the calcium silicates (predominantly belite) Dry cured Dry/wet
and calcium aluminates in the hydraulic lime, which pro- (20 C, 65% RH)
duces additional calcium hydroxide. Age 7 28 56 7 28 56
(3) Pozzolanic reaction of the alumina-silicates with the addi- Composition
tional calcium hydroxide produced. 100% NHL5 3.3 3.2 5.8 2.1 0.9 1.2
(4) Carbonation of the remaining free lime. 100% NHL5 (CEN) 2.4 2.3 4.0 1.2 0.5 0.5
100% NHL–PC 17.3 15.1 18.6 1.6 1.2 1.4

The degree and order of these reactions within the binder sys- 85% NHL5
15% IRBD 2.9 2.3 4.8 2.6 0.7 1.0
tem depends on the hydraulicity of the lime, the reactivity of the 15% ERBD 3.1 3.4 4.5 2.4 1.0 0.9
pozzolan, the binder composition and the curing conditions [27]. 15% FA 4.3 4.1 5.6 2.7 0.9 0.7
Cizer et al. concluded that the pozzolanic reaction dominates car- 15% MK1 3.3 3.2 6.8 1.0 0.5 0.8
bonation, independent of curing conditions, in the presence of 15% MK2 4.6 4.6 8.1 1.6 0.7 0.9
15% MK3 8.4 8.0 12.4 1.3 1.1 1.4
reactive pozzolanic materials [28]. Furthermore they suggest that 15% GGBS 5.3 6.4 7.4 2.2 0.7 0.7
even in dry conditions (RH = 60–65%) that carbonation of mortars 15% SF 6.0 9.0 8.5 1.2 0.7 0.5
only proceeds gradually from 28-days. Both the hydration of the 70% NHL5
hydraulic components and the pozzolanic reaction are dependent 30% IRBD 2.4 3.1 4.5 2.3 1.0 0.9
on the availability of water, so it is understandable that in dry con- 30% MK1 2.3 2.3 4.0 1.1 0.5 0.8
15% MK1 and 15% FA 6.9 7.0 11.3 1.0 0.5 0.7
ditions, in which the free water evaporates from capillary pores, 15% MK1 and 15% GGBS 5.9 10.7 13.5 1.5 0.7 0.9
both hydration and pozzolanic reactions are hampered. 15% SF and 15% FA 6.0 7.4 8.8 1.3 0.5 0.4
In case of Portland-cement based materials a number of param- 15% SF and 15% GGBS 6.2 10.5 11.7 0.8 0.5 0.5
30% SF 10.3 14.6 16.9 1.6 0.5 0.5
eters have been found to influence the sensitivity of the binder sys-
tem to different curing conditions: water to binder ratio [27,29]; 50% NHL5
25% MK1 and 25% FA 4.3 4.7 7.6 0.9 0.4 0.7
binder content [27,29]; nature of alumina-silicious addition
25% MK1 and 25% GGBS 5.8 9.5 11.2 0.8 0.7 0.8
[30,31] and replacement level [27,29,30]. These mix parameters 25% SF and 25% FA 6.4 8.5 8.3 1.1 0.4 0.3
are shown to be similarly important in the adequate curing of 25% SF and 25% GGBS 9.1 11.5 14.4 1.3 0.4 0.5
hydraulic lime-based binders.
Table 5, gives the compressive strength (N/mm2) of dry-cured
mortars (20 °C and 60–65% RH) as well as the ratio of dry to higher water demand, is thought to contribute to an increased sen-
wet-cured strengths. The shaded figures highlight specific cases sitivity to curing conditions in the case of the hydration of lime.
in which the dry-cured strengths were greater than the wet cured At 7 days the compressive strength of almost all the samples
strengths (dry/wet >1.0). can be seen to be higher in dry-cured samples than in water cured
It is shown that different alumino-silicate minerals have a vary- equivalents. This is attributed to secondary forces caused by the
ing impact on the sensitivity of lime–pozzolan mortars to different distribution of moisture in the mortar matrix. Two separate phe-
curing conditions. The compressive strength of hydraulic-lime nomena are discussed here, dewatering and moisture gradients:
mortar samples prepared with 15% SF and then dry cured, were When hydraulic lime mortars are used in practice the fresh
around 30% lower at 28-days and around 50% lower at 56-days, mortar is laid onto masonry units. Depending the sorptivity of
in comparison to identical specimens cured in water. When the the masonry substrate and the desorptivity of the fresh mortar,
dosage of SF was increased to 30% the reduction was more pro- free water is typically pulled out of the mortar by capillary suction,
nounced at 28-days, attaining only around 50% of the compressive a process known as dewatering. Studies have shown that dewater-
strength at both 28 and 56 days. The magnitude of the adverse ef- ing has a beneficial effect on the compressive strength of hydraulic
fect of dry-curing on the strength development of the lime–pozzo- lime mortars [32]. This has been attributed to a localised decrease
lan mortars was observed to vary depending on the specific in the w/b ratio leading to a denser microstructure as well as to an
pozzolanic addition. Dry curing was seen to be most detrimental increased rate of crystallisation of silicate phases, which has been
to ternary hydraulic-lime mortars prepared with 25% SF and 25% observed by Scanning Electron Microscopy of the microstructure
FA. of dewatered mortars. At 7-days evaporation of free water from
The inclusion of alumino-silicate minerals in blended cements the surface of the mortars would result in a similar loss of water
has previously been shown to have an impact on the sensitivity at the surface. It is noteworthy that this phenomenon means that
of Portland-cement based binders to different curing conditions. mortars cast in steel moulds, in accordance with the standard BS
In general the strength reductions observed in this study are great- EN 1015 [33], leads to lab-results that underestimate the strength
er than those reported for Portland-cement systems in literature. of mortars in practise [32].
Atis et al. showed that the compressive strength of silica fume con- The compressive strength of each dry-cured mortar bar was
cretes that were cured in dry conditions, (65% RH and 20 °C) were tested immediately after removing it from the storage atmosphere
on average 13% lower than those cured in water (at 20 °C) [29]. of 20 °C and 60–65% RH without soaking it. Similarly the wet-cured
Similar results have been reported by Ramezanianpour and Malho- samples were tested immediately after removing them from the
tra [30]. The compressive strength of concretes prepared with water bath without allowing time to dry out. In none of the previ-
GGBS have previously been observed as being on average 15% low- ous studies reported in literature, nor in this study, was the mois-
er when dry cured than those cured in water [27]. ture condition of the sample at the time testing specifically
Portland cement, which consists of spherical cement grains, has considered. Higher compressive strengths in dry-cured mortars
a typical surface area of 350 m2/kg [3]. NHL consists of hexagonal at 7 days has, however, identified the effect of the moisture condi-
plate-like grains, and the NHL5 used in this project had a surface tion at point of testing, a phenomenon that was investigated in de-
area of 800 m2/kg. The increased surface area, leading to an in- tail by Popovics [34], but that appears to have received little
crease of surface forces between lime particles, results in a higher attention since.
overall water demand and substantially increased cohesiveness at Popovics’ work, investigating the effect of moisture condition
low water/binder (w/b) ratios. The higher surface area, and thus on the resulting compressive strength of concretes, showed that
522 E.R. Grist et al. / Materials and Design 52 (2013) 514–523
the distribution of moisture through a concrete sample could have
a considerable effect on its strength. Mortar samples cured in dry
conditions for 7 days would experience a moisture gradient, with
the surface of the mortar being drier than the core. Popovics pur-
ported that drying of the surface of concrete specimens, results
in localised shrinkage, which restrains, or pre-compresses, the in-
ner portion, putting it into lateral biaxial compression, increasing
the global uniaxial compressive strength [34].
Given that each tested mortar sample constituted a third of a
mortar bar, cut with a water-lubricated diamond saw immediately
before testing, the moisture gradient through the samples was not
uniform and very unlikely to be in a state of true biaxial compres-
sion. However the relative degree of any strength effects associated
with secondary forces in the dry-cured mortar system would have
been most apparent at 7-days, when the strength associated with
hydration was lowest. In the case of the dry-cured samples on-
going evaporation would have reduced the moisture gradient
through each sample over time, reducing both the absolute and
the relative strength gain associated with the uneven moisture dis-
tribution. However the overall moisture condition of the dry and
wet-cured mortars might have continued to have an impact on
the measured compressive strengths, specifically when an external
load is applied to a saturated concrete the increase in pore water
pressure can lead to crack propagation, and thus lower compres-
sive strengths [34]. This phenomenon would have resulted in low-
er measured compressive strengths in the saturated wet-cured
mortars and minimised the difference between the reported wet
and dry cured mortar strengths.
Despite evidence of the effect of moisture on the strength of
mortars, the standard testing procedures do not explicitly state
the moisture condition in which samples should be tested. BS EN
196-1 presupposes continuous moist curing and testing of satu-
rated specimens [14]. BS EN 1015-11:1999 on the other hand pre-
sumes curing at 65% RH for 3 weeks prior to testing and requires
that specimens are tested immediately after removal from this
environment [33]. Neither standard details the preparation of mor-
tars immediately prior to mechanical testing and therefore neither
facilitates a direct comparison of wet and dry cured mortar
strengths. Specimen preparation is also passed over in the stan-
dards for testing the compressive strength of concrete specimens
[35,36]. This is a change from the earlier British Standard BS
1881-116, published in 1983, which explicitly stated that samples
not cured in water should be re-wetted prior to testing [37]. Many
authors investigating the effect of curing conditions on the com-
pressive strength of concretes, fail to describe the moisture condi-
tion of samples at the point of testing [27,29,30]. For this reason
the observed effect of curing conditions can only be analysed by
comparing the resulting strengths of mortars within this test set.
Considering the cases in which the dry-cured strengths were
greater than the wet cured strengths, other than the results at 7-
days which have been discussed, a couple of other observations
are interesting. The compressive strength of the proprietary formu-
lated-lime mortar (NHL5–PC) was greater for dry-cured specimens
than for water-cured specimens at all ages. The same was found for
hydraulic lime-mortars prepared with 15% MK3, contrary to the
two other grades of metakaolin. This effect has not been investi-
gated in detail for this particular project.
3.5. Effect of replacement level
Plotting the best-fit linear relationships between 28-day
strengths of wet and dry-cured groups of mortars prepared with
different replacement levels of additions (Fig. 6) clearly demon-
strates the increased sensitivity of samples with higher replace-
ment levels. This is consistent with the finding of
Ramezanianpour and Malhotra [30] and Atis et al. [29], who ob-
Fig. 6. Relationship between dry and wet cured mortar strengths.
served that the sensitivity to curing conditions increased with an
increased replacement level of GGBS and SF respectively.
4. Conclusions
The results of this investigation challenge the broadly accepted
view that lime-based building materials are notoriously weak and
slow to set and cannot compete with Portland cement based alter-
natives for these reasons. Rather these results demonstrate that
hydraulic lime–pozzolan mortars prepared in the lab using modern
pozzolanic Type II additions can attain 28-day compressive cube
strengths of over 25.0 N/mm2. An analysis of the contribution of
the pozzolanic addition to the strength gain has shown that a ter-
nary combination of silica fume and ggbs can have a Pozzolanic
efficacy of 94% and result in a mortar eight times as strong as a
mortar prepared with NHL5 alone.
This work has also demonstrated that alumino-silicate addi-
tions can be combined with a complimentary effect. For example
the use of metakaolin was shown to be beneficial in combination
with both GGBS and FA as it facilitated an earlier pozzolanic reac-
tion contributing to early strength gain in the lime–pozzolan mor-
tar system. A ternary combination of NHL5, SF and GGBS was
shown to produce mortars with the highest 28-day compressive
strengths when cured in water or in air. Further it has been shown
that the composite blend of additions not only affects the mortars
compressive strength but also its sensitivity to curing conditions.
In an extension of the work of Yu et al. this work has shown that
plotting Pozzolanic Index, PI (%) against time is a powerful graph-
ical method for describing the relative contribution of the pozzola-
nic addition to the process of strength development in lime–
pozzolan cementitious systems [18]. However it is suggested that
the term Pozzolanic Efficacy (PE%) is adopted as opposed to Pozzo-
lanic Index (PI%) to differentiate this outcome-orientated approach
from the material property measured by the Chapelle test.
This experimental investigation was successful in identifying a
small number of lime–pozzolan combinations that were deemed
promising for scaling up for the subsequent production and testing
of lime–pozzolan concretes. Three of the four lime–concrete mixes
identified were based on the complimentary performance of ter-
nary combinations of alumino-silicate additions. The fourth was
based on the high strengths attained with a high dosage of silica
fume alone. Observed results also successfully differentiated be-
tween the three metakaolins used in this investigation, on which
basis MK3 was adopted for the subsequent investigation.
The dry-consistence of hydraulic-lime pozzolan mortars at a w/
b ratio of 0.5 highlighted that a suitable Water Reducing Admixture
 E.R. Grist et al. / Materials and Design 52 (2013) 514–523
(WRA) would need to be identified to control the rheology of the
fresh lime–concretes. An appropriate WRA would also allow the
hydraulic lime–pozzolan concretes to be produced at a lower
water/binder ratio and thus maximise the compressive strength.
It is notable that this work has highlighted a limitation in the
mortar testing standards with regards to curing in a range of con-
ditions and appropriate preparation of the mortars for mechanical
testing. Specifically it is has been noted that neither BS EN 196-1
[14] nor BS EN 1015-11:1999 [33] explicitly deal with the moisture
condition at which compressive strength tests should be
performed.
Acknowledgements
This work was supported by the EPSRC funded Industrial Doc-
torate Centre in Systems (Grant EP/G037353/1), Ramboll and the
Universities of Bath and Bristol.
References
[1] CESA [Internet]. CO2 emissions of various binders: St. Astier Natural Hydraulic
Limes (NHL). <http://www.stastier.co.uk/nhl/testres/co2emissions.htm>.
[updated 2006, accessed 2012]
[2] Mineral Product Association. Specifying sustainable concrete. The Concrete
Centre; 2011.
[3] Bye GC. Portland cement. 3rd ed. ICE Publishing; 2011.
_ Karakurt C, Sarıdemir M. Predicting the strength development of
[4] Topçu IB,
cements produced with different pozzolans by neural network and fuzzy logic.
Mater Des 2008;29(10):1986–91.
[5] Lee C-L, Huang R, Lin W-T, Weng T-L. Establishment of the durability indices
for cement-based composite containing supplementary cementitious
materials. Mater Des 2012;37:28–39.
[6] P. Ellis [Internet]. Gauging lime mortars. <www.buildingconservation.com/
articles/limeguaging.htm>. [updated 2001, accessed 2013]
[7] Torgal FP, Shasavand A. Eco-efficient concrete using industrial wastes: a
review. Mater Sci Forum 2013;730–732:581–6.
[8] Sepulcre-Aguilar A, Hernández-Olivares F. Assessment of phase formation in
lime-based mortars with added metakaolin, Portland cement and sepiolite, for
grouting of historic masonry. Cem Concr Res 2010;40(1):66–76.
[9] Sadique M, Al Nageim H, Atherton W, Seton L, Dempster N. A new composite
cementitious material for construction. Constr Build Mater 2012;35:846–55.
[10] BSEN 13263-1. Silica fume for concrete: definitions, requirements and
conformity criteria. BSI; 2005.
[11] BSEN 15167-1. Ground granulated blast furnace slag for use in concrete,
mortar and grout –Part 1: Definitions, specifications and conformity criteria.
BSI; 2006.
[12] BSEN 450-1. Fly ash for concrete: definition, specifications and conformity
criteria. BSI; 2012.
[13] BS 933-1. Tests for geometrical properties of aggregates: determination of
particle size distribution – Sieving method. BSI; 2012.
[14] BSEN 196-1. Methods of testing cement: determination of strength. BSI; 2005.
[15] BSEN 459-1. Building Lime – definitions, specifications and conformity criteria.
BSI; 2010.
523
[16] McCarter WJ, Tran D. Monitoring pozzolanic activity by direct activation with
calcium hydroxide. Constr Build Mater 1996;10(3):179–84.
[17] BS 3892-1. Specification for pulverized-fuel ash for use with Portland cement.
BSI. 1993.
[18] Yu LH, Ou H, Lee LL. Investigation on pozzolanic effect of perlite powder in
concrete. Cem Concr Res 2003;33(1):73–6.
[19] Baronio G, Binda L. Study of the pozzolanicity of some bricks and clays. Constr
Build Mater 1997;11(1):41–6.
[20] Walker R, Pavía S. Physical properties and reactivity of pozzolans, and their
influence on the properties of lime–pozzolan pastes. Mater Struct
2011;44(6):1139–50.
[21] Siddique R, Chahal N. Use of silicon and ferrosilicon industry by-products
(silica fume) in cement paste and mortar. Resour Conserv Recycl
2011;55(8):739–44.
[22] AGS MINERAUX [Internet]. Product data sheet – metakaolin. <http://
www.imerys-perfmins.com/eu/markets/building-construction.htm>.
[updated 2013, accessed 2013].
[23] Poon C-S, Lam L, Kou SC, Wong Y-L, Wong R. Rate of pozzolanic reaction of
metakaolin in high-performance cement pastes. Cem Concr Res
2001;31(9):1301–6.
[24] Cassagnabère F, Escadeillas G, Mouret M. Study of the reactivity of cement/
metakaolin binders at early age for specific use in steam cured precast
concrete. Constr Build Mater 2009;23(2):775–84.
[25] Isaia GC, Gastaldini ALG, Moraes R. Physical and pozzolanic action of mineral
additions on the mechanical strength of high-performance concrete. Cement
Concr Compos 2003;25(1):69–76.
[26] Radlinski M, Olek J. Investigation into the synergistic effects in ternary
cementitious systems containing Portland cement, fly ash and silica fume.
Cement Concr Compos 2012;34(4):451–9.
[27] Duran Atis C, Bilim C. Wet and dry cured compressive strength of concrete
containing ground granulated blast-furnace slag. Build Environ
2007;42(8):3060–5.
[28] Cizer Ö, Van Balen K, Van Gemert D. Competition between hydration and
carbonation in hydraulic lime and lime–pozzolana mortars. Adv Mater Res
2010;133:241–6.
[29] Atisß CD, Özcan F, Kılıç A, Karahan O, Bilim C, Severcan MH. Influence of dry and
wet curing conditions on compressive strength of silica fume concrete. Build
Environ 2005;40(12):1678–83.
[30] Ramezanianpour AA, Malhotra VM. Effect of curing on the compressive
strength, resistance to chloride-ion penetration and porosity of concretes
incorporating slag, fly ash or silica fume. Cement Concr Compos
1995;17(2):125–33.
[31] Ozer B, Ozkul MH. The influence of initial water curing on the strength
development of ordinary Portland and pozzolanic cement concretes. Cem
Concr Res 2004;34(1):13–8.
[32] El-Turki A, Ball RJ, Carter MA, Wilson MA, Ince C, Allen GC. Effect of dewatering
on the strength of lime and cement mortars. J Am Ceram Soc
2010;93(7):2074–81.
[33] BSEN 1015-11. Methods of test for mortar for masonry: determination of
flexural and compressive strength of hardened mortar. BSI; 1999.
[34] Popovics S. Effect of curing method and final moisture condition on
compressive strength of concrete. ACI J Proc 1986;83(4):650–7.
[35] BS 12390-2. Testing hardened concrete – making and curing specimens for
strength tests. BSI; 2009.
[36] BS 12390-3. Testing hardened concrete – compressive strength test of
specimens. BSI; 2009.
[37] BS 1881-116. Testing concrete: method for determination of compressive
strength of concrete cubes. BSI; 1983.