Construction and Building Materials 290 (2021) 123272

Contents lists available at ScienceDirect

Construction and Building Materials

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

Mechanical and physical performance of natural hydraulic lime mortars

Maria Apostolopoulou, Asterios Bakolas ⇑, Meletis Kotsainas
Laboratory of Materials Science & Engineering, School of Chemical Engineering, National Technical University of Athens, 9 Iroon Polytechniou St., NTUA Campus, 15780,
Zografou, Athens, Greece

h i g h l i g h t s

B/A ratio affects the physical and mechanical properties of NHL2 and NHL3.5 mortars.

The influence of B/A is different according to the binder type.

20% binder leads to a mortar of lower mechanical performance in both cases.

25% binder seems to be the optimum percentage in the case of NHL2.

30% binder seems to be the optimum percentage in the case of NHL3.5.

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

Article history: Natural hydraulic lime (NHL) is a material which has received increased interest in the past decade as a
Received 20 November 2020 binder for the production of mortars. The renewed interest of researchers and builders in natural hydrau-
Received in revised form 16 February 2021 lic lime is attributed to its enhanced compatibility with many traditional materials, as well as the fact
Accepted 6 April 2021
that it is an environmentally friendly material, especially in relation to modern cement. In the present
study, two different types of natural hydraulic lime mortars are studied, using NHL2 and NHL3.5 as bin-
der (in accordance to the new classification EN 459-1:2010) and the same river sand, while the influence
Keywords:
of different binder to aggregate ratios on mortar characteristics is also investigated. Fresh mortar charac-
Natural hydraulic lime
NHL2
teristics are examined immediately after mixing, while hardened mortar characteristics are evaluated at
NHL3.5 different mortar ages. Specifically, the mechanical performance of the mortars is evaluated through flex-
Pore network structure ural and compressive strength tests, as well as the dynamic modulus of elasticity, which is estimated
Fresh state properties through ultrasound pulse velocity measurements. Physical performance of the mortars is investigated
Mechanical properties through the examination of the pore network structure characteristics, via mercury intrusion porosime-
Dynamic modulus of elasticity try measurements, and through the examination of hygric characteristics, via water absorption tests. The
Water absorption tests binder to aggregate ratio seems to play an important role in the development of the mechanical and
physical characteristics of the mortars, while it influences mortar properties differently according to
the type of the binder.
Ó 2021 Elsevier Ltd. All rights reserved.

1. Introduction [1,5]. The calcination of marly limestones was especially important

in regions where pozzolans were not available.
Natural hydraulic lime (NHL) has been used as a binder for mor- Although the use of NHL was substituted by the extensive use of
tar production since antiquity [1,2]. It is produced through the rel- cement mortars in the 20th century [6], in recent years, it has, once
atively low-temperature calcination and subsequent slaking of again, become a material of interest, both for restoration purposes
marly limestones (limestones naturally containing silicate and alu- (in cases where long-term compatibility with the historic fabric is
minate compounds) [3,4]. In antiquity its use was either on pur- ensured [7]), as well as a binding material in new eco-friendly con-
pose (areas of high humidity or cases where faster setting and structions [8–10]. This renewed interest is on account of (i) its
higher strength, in relation to aerial lime, were demanded) or by higher compatibility with traditional materials, in relation to
chance, due to the characteristics of the locally available limestone cement mortars [3,11,12], (ii) its enhanced behavior in relation to
aerial lime mortars (faster and higher acquisition of mechanical
strength, earlier carbonation due to lower lime content, ability of
⇑ Corresponding author. hydraulic lime to harden in moist conditions or underwater, lower
E-mail address: abakolas@mail.ntua.gr (A. Bakolas). dimensional changes, etc, [11,13–15]), but also on account of

https://doi.org/10.1016/j.conbuildmat.2021.123272
0950-0618/Ó 2021 Elsevier Ltd. All rights reserved.
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

(iii) its enhanced environmental impact, both in relation to its and ~ 0.25 by weight [4,8,11,15,28,29,31,32,34], while only a few
production process [16,17], as well as in relation to its ability to investigate mortars with B/A ratio 0.33 [3,9,33]; furthermore, with
absorb CO2 from its environment during setting and hardening the exception of a few studies [3,9,33], usually a relatively high W/
on account of lime carbonation [13,18]. Additionally, the presence B ratio is applied, close to (or even higher than) 1. It should be
of calcium hydroxide also offers an important potential for self- noted that important research has also been conducted regarding
healing in the case of micro-cracking occurrence, in the appropri- the improvement of NHL mortar characteristics with the addition
ate environment [13,19,20], however at the same time it may be of fine aggregates [30,32], as well as pozzolans [26,35].
associated with leaching phenomena in certain climates and condi- In the present study, different mortars are produced, using two
tions [21,22]. different types of natural hydraulic lime, NHL2 and NHL3.5, while
In 2010 (revised in 2015), the standard EN-459-1, which pro- using the same aggregate and maintaining a constant consistency
vides the requirements that a material must fulfill in order to be for each type of NHL mortar. The aim is to examine the influence
classified as an NHL, was substantially updated [23–25]. A new of the binder type (in accordance to the updated 2010 standard
hydraulic lime class, formulated lime (FM), was introduced, and [24]) and content on physical and mechanical characteristics,
stricter criteria were set, related to chemical and physical proper- within a specific range of binder percentages, 20 to 30% (corre-
ties, as well as the manufacturing process, thus resulting in some sponding to 0.25–0.43 B/A ratios), up to 12 months mortar age,
binders, formerly classified as NHLs, to be now classified as while maintaining a low, appropriate consistency. The results will
hydraulic lime or formulated lime binders [26]; for example, the provide further insight regarding NHL2 and NHL3.5 mortar
NHL-Z class is now classified as formulated lime [24]. According behavior.
to the current standard, in NHLs the hydraulic properties exclu-
sively result from special chemical composition of the natural 2. Materials and methods
raw material, grinding agents are allowed up to 0,1%, and NHLs
do not contain any other additions [24,25]. New chemical require- 2.1. Raw materials
ments are introduced, as each type of NHL binder must present: (i)
an appropriate amount of available lime (Ca(OH)2), higher than The mortars were produced with the use of two types of natural
demanded by the previous 2001 standard, and specifically  35% hydraulic lime, NHL2 and NHL3.5 (Chaux Blanche), procured from
in the case of NHL2, 25% for NHL3.5 and  15% for NHL5 binders, Lafarge, and in accordance to EN 459–1:2010 [24]. The two binders
while (ii) SO3 must now be  2% for all NHL types, in contrast to 3%, presented bulk density 0.60 g/cm3 (NHL2) and 0.64 g/cm3
which was the limit value in the 2001 standard [23–25]. Addition- (NHL3.5). Blaine fineness values were 10,150 cm2/g and
ally to chemical requirements, certain physical requirement crite- 11,000 cm2/g for NHL2 and NHL3.5, respectively. Chemical and
ria have also been modified: (i) particle size limitations, as to the mineralogical analysis results of the two binders is presented in
upper limit of retained material at 0.2 mm, is now lower, 2% in Table 1.
relation to 5% in the previous 2001 standard; (ii) air content must For the production of the mortars, the same natural river sand
be  5% for all NHL mortars, while in the previous standard the (0–4 mm) was used as aggregate, in order to obliterate the effect
limit was 20%; (iii) final setting time of NHL mortars is now differ- of the aggregate’s characteristics on the mortar properties. The
ent for each NHL type and specifically it must be  40 h for NHL2 sand was selected as to have an appropriate gradation (Fig. 1,
mortars, 30 h for NHL3.5 mortars and  15 h for NHL5 mortars estimated through the sieving method in accordance to
[23–25]. Regarding compressive strength, the tolerance values for EN 933–1 [36]) and low presence of fine material passing the
the different types of NHL also present some slight modifications: 0.063 mm sieve (1.7%), characteristics which were evaluated in
(i) there is no limit value for the 7-days compressive strength of accordance to EN 13139:2002 [37]. The sand presented a sand
NHL3.5 mortars in the new standard; (ii) NHL5 mortars must pre- equivalent of 98%, as measured according to EN 933–8:2012 [38].
sent compressive strength up to 15 MPa at 28 days, regardless of Bulk density and voids (%) of the sand were determined in accor-
the binder’s bulk density [23–25]. dance to EN 1097–3:1998 [39] at 1.50 g/cm3 and 40% respectively.
Due to the substantial differentiation of NHL binders after 2010, XRD analysis of the river sand showed quartz as principal miner-
there is an increased need to further study NHL mortar character- alogical component, while albite and muscovite were detected as
istics. Thus, a number of researches have been conducted to this secondary components.
purpose during the last decade. Although NHL5 mortars have
received more attention than NHL3.5 and NHL2, on account of
2.2. Mortar design, mixing and curing
the higher mechanical strength achieved [3,27], the latter binders
are acknowledged to exhibit a higher degree of compatibility with
The binders were used in three different percentages, namely
traditional materials [28]. In the present study only NHL2 and
20%, 25% and 30% by weight (Table 2), aiming to evaluate the effect
NHL3.5 mortars are examined, aiming to provide additional infor-
mation regarding their mechanical and physical properties at dif-
ferent ages. Table 1
Review of the relevant literature highlights the fact that NHL Chemical and mineralogical analysis of NHL2 and NHL3.5 used in the current study.
mortars’ characteristics are dependent on: (i) the type of NHL used Oxide composition NHL2 NHL3.5
[3,4,27,29]; (ii) the ratio of binder (NHL) to aggregate (sand), B/A
SiO2 (%) 8.01 13.02
[27,30]; (iii) the amount of water added to the dry mortar during Al2O3 (%) 0.53 0.6
mixing in relation to the amount of binder used in the mix, W/B, Fe2O3 (%) 0.25 0.2
in order to achieve the desired consistency [10,27,30,31]; (iv) char- CaO (%) 69 67
acteristics related to the aggregates (origin, composition, morphol- MgO (%) 0.66 0.51
SO3 (%) 0.76 0.81
ogy, grain size distribution) [30–33]; as well as (v) curing
K2O (%) 0.12 0.11
conditions [11,31,33]. Researchers highlight the fact, that by alter- Na2O (%) 0.07 0.05
ing these influencing parameters, an NHL mortar can be tailored to LOI (%) 20 17
exhibit desired characteristics [12,18,27,30]. Regarding NHL2 and Free Water (%) 0.6 0.7
Mineralogical Portlandite, calcite, belite, Portlandite, calcite,
NHL3.5 mortars, in the relevant literature, it is noticed that the
Composition larnite, quartz belite, quartz
majority of research is conducted on B/A ratios between 0.13
2
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

Fig. 1. Gradation curve of used aggregate.

Table 2
Mix design of mortar mixes.

Sample code Type of binder Binder percentage Aggregate percentage Binder to aggregate Binder to aggregate
(%) by weight (%)by weight ratio (B/A) by weight ratio (B/A) by volume
NHL2_20 NHL2 20% 80% 1/4 = 0.25 0.63 (~2/3)
NHL2_25 NHL2 25% 75% 1/30.33 0.83 (~4/5)
NHL2_30 NHL2 30% 70% 3/70.43 1.07 (~1/1)
NHL3.5_20 NHL3.5 20% 80% 1/4 = 0.25 0.59 (~1/2)
NHL3.5_25 NHL3.5 25% 75% 1/30.33 0.78 (~3/4)
NHL3.5_30 NHL3.5 30% 70% 3/70.43 1.00 (~1/1)

of binder content on the characteristics of the produced mortars, 16cmx4cmx4cm), in order to examine the hardened mortar char-
for each NHL type. The B/A range was selected in accordance to acteristics. The mixing vessel was 5 l and each batch was produced
usual practice, while also taking into account the characterization with 2 kg dry mortar, in accordance to the mix designs of Table 2.
of historical mortars [1,3,9,10,28–30,33,40], as well as the areas of Four batches were produced and used for a total of 21 prismatic
mortar design not thoroughly examined in the literature specimens, for each mortar mix design.
(e.g. 0.43 B/A ratio by weight). All mortars were cured in the same curing conditions, in accor-
In order for the results to be comparable, care was taken to keep dance to EN 1015-11 [42], thus obliterating the effect of different
a constant consistency in fresh state for each mortar type. In par- curing conditions on the development of their mechanical and
ticular, for the mortars produced with NHL2, a flow of 14 cm was physical characteristics. Specifically, the prismatic mortar speci-
selected as the optimum consistency, in terms of workability of mens were stored in high humidity conditions (RH = 95 ± 5%,
the produced mortar, close to the values also proposed by other T = 20 ± 2 °C), demolded after two days (in the case of NHL2, the
researchers for NHL2 mortars [3,31]. For the mortars produced specimens were demolded after five days, as setting was not com-
with NHL3.5, a 15 cm flow was selected as optimum, similar to plete at two days), and kept in the same conditions for up to seven
the consistency applied by other researchers for NHL3.5 mortars days since mixing and casting. This initial period of high humidity
[31–34]. This difference in consistency among different types of is considered necessary in order to ensure proper setting and hard-
NHL, in order to achieve the same workability, has been reported ening of hydraulic mortars and avoid the development of micro-
by other researchers as well, who also discuss that different types cracks, due to sudden moisture loss [11]. After a total of seven days
of NHL present optimum workability at different levels of consis- in the high humidity chamber, all specimens were stored in stan-
tency, as measured through the flow table; in fact, the higher the dard laboratory conditions (RH = 65 ± 5%, T = 20 ± 2 °C) up to
hydraulicity of the binder, the higher the level of consistency nec- the date of testing.
essary for an optimum workability of the mortar [e.g. 11]. Even in
the precision data table for the physical test methods, the total
mean flow values of NHL2 and NHL3.5 mortars are quite differen-
tiated [41]. 2.3. Methods and techniques
The raw materials were mixed with the appropriate amount of
water to achieve the desired consistency. The binder was added to Fresh state characteristics were examined for three different
the calculated water content and mixed for about thirty seconds, at batches of each mortar mix design, immediately after mixing.
which time the sand started to be gradually mixed in. Total mixing Specifically, the mortars were examined regarding their consis-
time amounted to 5 min. Fresh state mortar characteristics were tency (EN 1015-3 [43], flow table), bulk density (EN 1015-6
examined, while the remaining mortar was poured into steel [44]), air content (EN 1015-7 [45]) and ability to retain water
molds, in accordance to EN 1015-11 [42] (prisms with dimensions (EN 1015-8 [46]).
3
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

The mechanical and physical performance of the hardened mor- order to eliminate air between the specimens and the transducers.
tars (prismatic specimens) was investigated, at different mortar The pulse velocity of longitudinal stress waves was taken into
ages. account in order to estimate the dynamic modulus of elasticity,
The physical properties of the mortars were examined at through the following equation:
12 months mortar age. Mercury intrusion porosimentry (MIP)
ð1 þ v Þ  ð1  2mÞ
was employed [47] (Pascal 140 and 400 Thermo-Electronics- Ed ¼ q  V 2us 
Corporation) in order to study the main pore network structure ð1  m Þ
characteristics of the mortars (open porosity (%), bulk density, where, Ed: dynamic modulus of elasticity, Vus: ultrasonic pulse
average pore radius, specific pore volume, specific surface area), velocity of longitudinal stress waves, q: density, v: dynamic Pois-
as well as the pore size distributions of the mortars. MIP tests were son’s ratio. The v value was taken from the literature for mortar
conducted on three different specimens for each mortar mix (v = 0.25) [52].
design; the specimens were selected from the prismatic specimens
and shaped manually. They were of irregular shape (resembling a
cylinder), with approximately ~ 0.9 cm diameter and ~ 2 cm length. 3. Results and discussion
Water absorption tests were conducted, through capillarity and
total immersion, in accordance to EN 15801 [48] and Normal 07/81 3.1. Fresh mortar characteristics
[49], respectively. Each test was conducted on three different spec-
imens, for each mortar mix design, until saturation. Each specimen As soon as mixing was complete, fresh state characteristics of
was cubic, with a 4 cm edge, and cut from different prismatic spec- the produced mortars, and in particular consistency, bulk density,
imens. Water absorption was conducted by placing the fractured air content and retained water, were evaluated (Table 3). All three
surface in contact with the water front. The water absorption batches examined for each mortar mix design, presented similar
through capillarity test allowed for the determination of the capil- values regarding fresh state characteristics (W/B and consistency
lary rise coefficient (C.R.C.), through the one tangent method, as were constant, bulk density, air content and retained water pre-
well as the absorbed water percentage by weight (Wc%), specific sented standard deviation equal or lower than 0.01, 0.1 and 0.25,
pore volume (SPVc) and open porosity (Pc%) accessible to water respectively, for all mortar mix designs).
through capillaries, while the water absorption through immersion Regarding both types of mortars, it is obvious that the increase
test allowed for the determination of the absorbed water percent- in binder content (B/A) results in a decrease in water demand in
age by weight (Wi%), the specific pore volume (SPVi) and open relation to binder (W/B), for the respective consistency (for a more
porosity (Pi%) accessible to water by immersion. Porosity, absorbed detailed analysis of consistency see section 2.2.). The decrease in
water percentage and specific pore volume are calculated, for each W/B is less intense as the B/A increases, especially in the case of
type of test, through the following equations: the NHL3.5 mortars. This observation can be explained through
considering the role of water in the mortar mix. As the binder per-

e%=(volume of total absorbed water)/(apparent volume of centage increases, more water is necessary for the binder fraction
specimen) to achieve the desired consistency, however, less water is

W%=(mass of total absorbed water) *100/(mass of dry demanded to facilitate rolling of the aggregates, as the aggregate
specimen) fraction decreases [53]; additionally, when the binder proportion

SPV=(volume of total absorbed water)/(mass of dry specimen) increases, beyond a certain point, the distance between the aggre-
gate particles is increased, therefore reducing the interaction
Mechanical tests were conducted in accordance to EN 1015-11 among them, while at the same time the binder fraction assists
[42], at 1, 3, 6 and 12 months, for the determination of three-point in reducing friction among the aggregate grains [13].
flexural strength (ToniTechnik-D-70804, loading rate up to The combined effect of the aforementioned mechanisms leads
0.05 KN/s) and uniaxial compressive strength (ToniTechnik to the decrease of W/B ratio as B/A increases, as well as to a less
DKD-K-23301, loading rate up to 0.01 KN/s). Flexural strength intense decrease of W/B demand for higher B/A ratios. This trend
measurements were conducted on three prismatic specimens for has been noticed for other binding systems as well, to different
each mortar mix design, while the resulting halves were then used extents, according to the characteristics of the binder (e.g. chemical
for the compressive strength tests, resulting in six specimens for composition, morphology, specific surface) and the used aggregate
each mortar mix design. (e.g. particle surface area, apparent volume, void content, segrega-
The dynamic modulus of elasticity was evaluated at 12 months tion possibility, fines content, geometry, uniformity, granulometry)
mortar age, using three different prismatic specimens for each [13,18,40,53]. The NHL3.5 mortars exhibit a higher water demand
mortar mix design. In particular, ultrasonic pulse velocity mea- in relation to the NHL2 mortars, on account of the NHL3.5 binder’s
surements [50] were conducted applying the direct method, along higher fineness.
the prisms (16 cm path length), in accordance to ASTM C597:2002 Bulk densities of the mortars are similar; the NHL2 mortars
[51] (CNS Farnell-Pundit 6, transducers frequency: 54 KHz, probe seem to present similar or slightly higher bulk density values in
diameter 20 mm), using a coupling agent (water soluble gel) in relation to the NHL3.5 mortars, for each binder percentage, on

Table 3
Fresh mortar characteristics of the examined mortars.

Mortar code W/B (g/g) Water (%)* Consistency (cm) Bulk density (g/cm3) Air content (%) Retained water (%)
NHL2_20 0.79 13.64 14 2.08 3.1 80.67
NHL2_25 0.61 13.23 14 2.11 4.0 85.68
NHL2_30 0.51 13.27 14 2.12 3.2 85.90
NHL3.5_20 0.85 14.53 15 2.08 2.8 83.30
NHL3.5_25 0.69 14.71 15 2.07 3.5 86.80
NHL3.5_30 0.63 15.90 15 2.05 2.8 88.77

*Water (%): Water percentage in relation to total fresh mortar: binder, sand and water.

4
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

account of the different chemical and physical characteristics, as NHL3.5 mortars present average pore radius values within a wider
well as water demand of the binders. The air content of the range in relation to the NHL2 mortars. Porosity values are similar
NHL2 mortars lays in higher levels than the air content of the among the mortar mixes, with the NHL3.5 mortars presenting
NHL3.5 mortars, however all mortars examined present similar slightly lower values for each binder percentage in relation to the
values ranging from 2.8 to 4%, as expected for NHL mortars with respective NHL2 mortars. Bulk density of the NHL3.5 mortars is
no additives and in accordance to the limit set by EN 459-1 equal or lower than the NHL2 mortars.
[25,41]. It is interesting that in the case of both NHL2 and In the case of both NHL2 and NHL3.5, the mortars with the low-
NHL3.5 mortars, maximum air content is noticed in the case of est binder percentage present the highest value of specific pore
the intermediate B/A ratio (25% binder content, Table 3). volume, the lowest specific surface area, the highest average pore
Retained water (%) is above 80% for all mortars, both NHL2 and radius and the highest porosity, indicating the development of a
NHL3.5, indicating an adequate behavior in relation to water reten- less dense pore network structure. Total open porosity values of
tion capacity. Retained water increases as binder percentage the NHL2 mortars are similar to the lower range of porosity values
increases, and furthermore, in both mortar types, this increase is presented for NHL2 mortars in the literature [30], while total open
more intense between the mortar with 20% binder and the mortar porosity and average pore radius values presented in this study are
with 25% binder than between the mortar with 25% binder and the in coherence with those measured for NHL3.5 mortars by other
one with 30% binder. The NHL3.5 mortars present higher retained researchers [31–33].
water percentages in relation to the NHL2 mortars, even though These results indicate that the new classification of NHLs, pro-
they were produced with higher water to binder ratios. This is par- vided that no extreme B/A or W/B ratios are selected, leads to
tially on account of the higher fineness exhibited by NHL3.5, while the production of mortars with a porous system compatible with
it also indicates the higher demand of water required by the binder traditional building materials, especially in comparison to other
of higher hydraulicity. binders [40,54–56], while in both cases, the 20% binder percentage
leads to a more porous pore network structure.
3.2. Physical and mechanical characteristics of hardened mortars
3.2.1.2. Pore structure of hardened mortars. Porosity of a mortar is
considered to be influenced by (i) the choice of binder, (ii) the bin-
Natural hydraulic lime mortars harden through two different
der/aggregate and water/binder ratio and (iii) volume, size distri-
processes: hydration of calcium silicates and aluminates, as well
bution and mineralogical nature of the aggregates, while the
as carbonation of calcium hydroxide. The first process demands
development of cracks during setting and hardening and the pres-
water, while for the second process, CO2 must diffuse through
ence of air voids also affect its value. In turn the detailed pore
the pore structure in order to react with calcium hydroxide and
structure influences mechanical properties, as well as transfer phe-
produce calcite; hydration has been proven to occur mostly at
nomena [57,58]. Thus, the pore size distribution of a mortar (PSD,
early ages, while carbonation is a slower process, which continues
the distribution of pore volume with respect to pore size [59],
at later ages [3,19,28]. In the next subsections, the physical and
Fig. 2) is an important parameter which should be taken into
mechanical characteristics of the investigated natural hydraulic
account.
lime mortars are presented.
The NHL2 mortars present relatively similar PSD curves; the
PSD curve of mortar NHL2_25, the intermediate binder percentage,
3.2.1. Physical performance of hardened mortars is slightly shifted towards lower pore radii in relation to the other
3.2.1.1. Main pore network structure characteristics of hardened two percentages. The NHL3.5 mortars present relatively similar
mortars. The main pore network structure characteristics of the PSD curves to the NHL2 mortars, except for the mortar with the
produced mortars, as measured through MIP at 12 months mortars lowest binder percentage, which presents a PSD shifted towards
age, are summarized in Table 4. higher pore radii.
The effect of binder type and content on the main pore network All mortars present a slightly bimodal distribution, while the
structure characteristics of the mortars, within the range exam- most frequent pore radii are differentiated amongst mortar mixes.
ined, seems to be slight; SPV ranges between 117.7 and The most frequent pore radii of mortar NHL2_20 is noticed at
125.2 mm3/g, SSA ranges between 3.08–4.74 m2/g, APR ranges 0.75 lm, of mortar NHL2_30 at 0.70 lm and of mortar NHL2_25,
between 0.48 and 0.79 lm, within the same size magnitude, the mortar with the intermediate binder percentage, at even nar-
P ranges between 22.9 and 24.5% and BD between 1.93 and rower pores, at 0.60 lm. Regarding the NHL3.5 mortars, in the case
1.97 g/cm3. of NHL3.5_20, the most frequent pore radii is noticed at 0.80 lm
Regarding the specific pore volume, the NHL2 mortars present a (higher even than the NHL2 mortars), in the case of NHL3.5_25 it
wider distribution, while the NHL3.5 mortars present values inter- is noticed at 0.60 lm, while in the case of NHL3.5_30, the most fre-
mediate the NHL2 mortars. The NHL3.5 mortars present higher quent pore radii are noticed at even narrower pores, at 0.50 lm.
specific surface area values in relation to the NHL2 mortars. The Thus, it is indicated that increase of the binder, in the case of
NHL3.5, leads to a shift of the pore size distribution towards nar-
rower pores, while in the case of NHL2 this is noticed for the inter-
mediate binder percentage.
Table 4
Main pore network structure characteristics of mortars at 12 months mortar age The classification of pores sizes, when dealing with mortars, is
though MIP. an issue which is still under investigation and there is not yet a
common agreement on the boundary points of the different kind
Mortar mix SPV (mm3/g) SSA (m2/g) APR (lm) P (%) BD (g/cm3)
of pores in mortars [60,61]. The only official classification of pores
NHL2_20 125.2 3.08 0.73 24.5 1.96
is the one presented by IUPAC, which categorizes pores according
NHL2_25 117.7 3.40 0.62 23.2 1.97
NHL2_30 121.4 3.13 0.62 23.6 1.95
to their diameter into micropores (<0.002 lm), mesopores
NHL3.5_20 123.9 3.66 0.79 23.9 1.93 (0.002–0.05 lm) and macropores (>0.05 lm) [60]. However, this
NHL3.5_25 118.4 4.74 0.62 23.1 1.95 classification is not very useful for lime-based mortars, as the
NHL3.5_30 118.2 4.10 0.48 22.9 1.94 majority of pores fall within the macropores range. In the current
*SPV: Specific Pore Volume, SSA: Specific Surface Area, APR: Average Pore Radius, P: investigation, relative pore volume distribution (%) of the exam-
Total Open Porosity, BD: Bulk Density ined mortars is assessed for specific pore radius ranges (Table 5).
5
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

Fig. 2. Pore size distribution of mortars at 12 months.

Table 5
Distribution of relative pore volume (%) in different ranges of pore radius.

Pore radius range % Relative pore volume

NHL2_20 NHL2_25 NHL2_30 NHL3.5_20 NHL3.5_25 NHL3.5_30
>1 lm 20.67 12.14 13.58 27.32 11.03 3.13
0.5–1.0 mm 23.95 22.65 23.28 36.23 30.15 34.52
0.1–0.5 lm 26.22 31.04 39.33 20.86 31.55 38.12
0.05–0.1 mm 11.94 11.63 6.76 2.43 5.28 4.15
0.01–0.05 lm 15.49 20.62 14.69 7.83 16.20 15.73
<0.01 mm 1.73 1.92 2.36 5.33 5.79 4.35
Total 100 100 100 100 100 100

The pore radius ranges were selected by taking into account the On the other hand, pores with a radius lower than 1 lm are con-
ranges proposed by other researchers dealing with lime-based sidered to be interlinked with the binder [63].
mortars [50,57,62]. A high relative pore volume in the pore range 0.1–1 lm is
Pores with a radius higher than 1 lm, however lower than highly associated with lime-based mortars and researchers attri-
50 lm, are mainly attributed to the ITZ (interfacial transition zone bute pores of this size to carbonation and hydration reactions
between paste and aggregates) [50,63]. In mortars, the ITZ is [57], while in aerial lime mortars, pores of ~ 0.5 lm radius are
caused by the heterogeneous component system (binder paste linked to the transformation of portlandite to calcite [50]. Both in
and aggregate) [60]. NHL3.5_20 presents the highest relative pore the case of NHL2 mortars and NHL3.5 mortars, the relative per-
volume in this range, followed by NHL2_20; thus, the mortars with centage of volume in this range, increases linearly as the binder
the lowest binder percentage, in both cases, present the highest ITZ content increases, while the NHL3.5 mortars present a higher per-
volume. The other mortar mixes, with higher binder percentages, centage of relative pore volume (%) in the aforementioned range
present much lower values [64], mostly on account of the smaller for each binder percentage. All examined mortars present the high-
relative amount of aggregates [62]. In the case of NHL3.5, relative est amount of porosity in this pore range (over 50% in all cases),
volume in this pore radius range decreases as binder content indicating possible compatibility with historical mortars, which
increases. In the case of NHL2, the lowest value is noticed for the present the highest amount of pore volume in the pore
intermediate binder percentage, indicating that at NHL2 binder range > 0.1 lm [40]. Within this range, however, the distribution
percentages higher than 25% the ITZ is not affected only by the of porosity is different among the mortar mixes; the NHL3.5 mor-
presence of aggregates, but it is affected in a more complex man- tars present a higher amount of relative pore volume in relation to
ner. NHL3.5 seems to be affected more intensely by the binder per- the NHL2 mortars in pore radii from 0.5 to 1 lm.
centage; the NHL3.5 mortar with 20% binder percentage, presents a Pores with a radius under 0.1 lm are considered to be mainly
quite higher relative volume in this range than the respective NHL2 generated during the hydration process (C-S-H formation) [60].
mortar, however, as the binder increases, the opposite is noticed. It No obvious relationship is noticed between the relative volume
should be noted that only the NHL2 mortar with the lowest binder of pores in this range (<0.1 lm) and binder type or binder percent-
percentage (NHL2_20) presents larger pores, with a pore radius age; however, it is noticed that the NHL3.5 mortar with the lowest
over 10 lm (with a relative pore volume of 1.55% in the range binder percentage, presents the lowest value in this range, while
10–50 lm). No mortar presents coarse pores with a radius higher the NHL2 mortar with the intermediate binder percentage presents
than 50 lm, thus indicating sufficient compaction, the absence of the highest value in this range, amongst all mortars.
entrapped air during mixing and the absence of cracks due to Within this range, pores with a radius lower than 0.05 lm are
shrinkage [57,62]. called sorption pores [62]; these are associated with the hydration
6
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

of hydraulic compounds and are usually not noticed in pure aerial ical characteristics, which will be investigated in the following
lime mortars. Water is bound tightly in these pores and they are sections. It seems that the pore structure is not dependent on the
considered to not participate in water transfer reactions [57]. In binder content alone, but develops through a more complex mech-
the case of NHL2, the highest value is noticed for the intermediate anism; increase of the binder percentage leads to a denser pore
binder percentage. Regarding the NHL3.5 mortars, the relative vol- network structure and a shift towards lower pore radii in the case
ume in this range is similar between the 25% and 30% binder per- of NHL3.5, while in the case of NHL2 this is noticed when binder
centage mortars, however much lower for the lowest binder increases from 20 to 25%, while the opposite is noticed when the
percentage. binder increases from 25 to 30%.
The NHL3.5 mortars seem to present a larger percentage of gel
pores (pore radii < 0.01 lm [17,62]) than the NHL2 mortars; this is
expected, as gel pores are strongly associated with the formation of 3.2.2. Hygric characteristics of hardened mortars
hydration products and NHL3.5 is considered a higher hydraulicity Aiming to examine the water absorption behavior of the hard-
binder [17,62]. Specifically, the NHL2 mortars present relative pore ened mortars, water absorption tests were conducted, by capillar-
volume percentages in the gel pore radius range, 1.73% to 2.36%, ity, as well as by immersion. The results are summarized in Table 6,
increasing with binder percentage increase, while the NHL3.5 mor- while the experimental curves expressing water uptake of the
tars present higher values, ranging from 4.35 to 5.79%, without mortars through capillarity are presented in Fig. 3.
exhibiting a relation to the B/A ratio. The capillary rise coefficient is an important building material
It is interesting to note that the relative pore volume of the mor- characteristic, as it is interlinked with hygric behavior and inhabi-
tars in the pore radius range from 0.05 to 1 lm seems to be linearly tants’ comfort, while, in the long term, it is interlinked with the
affected by the binder content for each type of NHL. However, this durability of a structure [61].
is not always the case when examining the relative volume in pore Among the NHL2 mortars, the mortar with the lowest binder
radius ranges above 1 lm and below 0.05 lm, where pores possi- percentage, presents the highest water capillary rise coefficient,
bly develop in a more complex manner. followed by the mortar with the highest binder percentage. The
Through the above investigation, it is concluded that, the main mortar with the intermediate binder percentage presents the low-
pore network structure characteristics seems to range within sim- est capillary rise coefficient among the NHL2 mortars. Thus, in the
ilar values for the different mortar mixes. Both the type and the case of NHL2 mortars, C.R.C. is not directly controlled by the binder
content of binder affect the pore structure of the developed mor- percentage; the lowest value is noticed for the intermediate binder
tars, and, therefore, are expected to affect their hygric and mechan- percentage.

Table 6
Physical properties of examined mortars (total immersion and capillary rise tests).

Mortar mix Capillary water uptake test Water immersion uptake test
2 1/2 3
C.R.C. (kg/m s ) Wc (%) PVc (mm /g) Pc (%) Wi (%) PVi (mm3/g) Pi (%)
NHL2_20 0.18 12.01 120.42 22.14 12.26 122.81 22.52
NHL2_25 0.11 11.59 116.22 21.60 11.63 116.54 21.66
NHL2_30 0.16 11.98 120.08 22.22 12.05 120.68 22.33
NHL3.5_20 0.11 11.99 120.24 22.14 12.08 121.05 22.29
NHL3.5_25 0.08 11.75 117.82 22.14 11.79 118.07 22.18
NHL3.5_30 0.07 11.47 115.04 21.19 11.69 117.12 21.58

C.R.C.: capillary rise coefficient, Wc: percentage of absorbed water by capillaries, PVC: specific pore volume by capillary rise test, Pc: Open porosity accessible to water through
capillaries; Wi: percentage of absorbed water by immersion test, PVi: specific pore volume by immersion test, Pi: Open porosity accessible to water by immersion test.

Fig. 3. Water uptake of hardened mortars through capillarity.

7
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

The NH3.5 mortars present a slower rate of water absorption In contrast to the C.R.C., the percentage of absorbed water (by
through capillary rise in comparison to the NHL2 mortars, with weight) through capillarity (Wc) does not present intense differ-
the exception of NHL3.5_20, which presents a similar behavior to ences among mortar mixes; it seems that although in some cases
NHL2_25. In the case of NHL3.5, it seems that the C.R.C. decreases, water uptake is faster, in the long run, the amount of absorbed
as the B/A ratio increases. The values in the present study are sim- water in relation to the mass of the specimen is similar. The
ilar to those measured in other researches regarding NHL3.5 mor- NHL3.5 mortars present slightly lower Wc (%) in relation to the
tars [8,31], or slightly lower, due to the denser pore network NHL2 mortars, for each binder percentage. Again, the intermediate
structure developed, as a result of higher binder content applied binder percentage presents the lowest value among the NHL2 mor-
and lower consistency. tars, while in the case of the NHL3.5 mortars, Wc (%) decreases as
Of course the different trends noticed are highly related to the the binder percentage increases.
developed pore network structure, which in turn is related to the Regarding the percentage of water absorbed by immersion and
binder percentage of each mortar mix design. The exact relation- by capillarity, as expected, the former presents slightly lower val-
ship of the pore network structure and the absorption rate of a ues, as water is absorbed only through capillary pores. The small
mortar through capillary rise is an issue which is still under inves- differences noticed are indicative of the presence of mostly capil-
tigation; the difficulty in elucidating this relationship lays mostly lary pores in the pore network structure. The differences among
in the complexity of characterizing the pore structure, on account NHL2 mortars and NHL3.5 mortars are extremely slight regarding
of the existence of pores having different shapes and on account of these characteristics, a result which is in coherence with the slight
the different connectivity between pores [65]. differences they present in terms of main pore network structure
The exact pore size range of capillary pores is an issue still characteristics. The same trends are noticed regarding specific pore
under dispute [61]. Most researchers place pores which participate volume and open porosity of the different mortars, as measured
in water capillary rise in the range from 0.05 lm-50 lm [57,62,63]; through immersion and capillary rise.
this is logical, especially taking into consideration the fact that As expected, total open porosity measured through MIP is
even under vacuum water cannot penetrate pores with a radius higher than the open porosity accessible to water through total
lower than 0.05 lm [34]. However, no correlation was noticed immersion, while through capillary rise tests, where only capillary
between the relative pore volume of the aforementioned range pores participate, the lowest value of open porosity is calculated.
and the C.R.C. of the examined mortars. This can be explained These differences are expected on account of the different measur-
through the examination of the capillary rise mechanism: during able pore sizes of the techniques [50] and highlight the importance
capillary rise, it is generally accepted that water is first absorbed of the measuring technique applied for the evaluation of a mortar’s
in the larger pores (through the smaller capillary pores) [17]. Cap- porosity. The same is noticed regarding the calculation of specific
illary pores of a higher diameter, thus, act as a reservoir, drastically pore volume by the different techniques.
increasing the available water to be moved into the mortar [57].
Thus, the uptake rate, which corresponds to the linear first stage 3.2.3. Mechanical performance of hardened mortars
of absorption, is highly affected by the relative volume of larger 3.2.3.1. Flexural strength of hardened mortars. Fig. 4 presents the
capillary pores, in addition to the amount and radius of smaller values of flexural strength of the hardened mortars at different
capillary pores [17]. With this in mind, and taking into account that mortar ages. The average value for each age is presented along with
the binder type and consistency are constant for each type of mor- error bars, corresponding to the standard deviation of each test.
tar, the C.R.C. of the mortars was investigated in relation to the rel- Regarding the NHL2 mortars, the mortar with the lowest binder
ative pore volume of larger capillary pores, over 0.5 lm. The percentage (NHL2_20), presents the lowest flexural strength values
investigation revealed a linear dependence of the C.R.C. on relative at all ages, with an extremely slight increase of flexural strength
pore volume in this pore radius range (>0.5 lm). The same was from one month to twelve months (~13%); in fact, flexural strength
noticed when comparing the C.R.C. with the relative pore of pores can even be considered constant during these ages, taking standard
with radii > 1 lm. deviation into account.

Fig. 4. Flexural strength evolution in relation to mortar age for all mortar mixes.

8
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

The mortar with the highest percentage of binder (NHL2_30) other NHL3.5 mortar mixes, as found in the relative literature
presents intermediate flexural strength values among the three [8,15,31–34], at all ages, while NHL3.5_20 presents intermediate
NHL2 mortar mixes. The increase of flexural strength with mortar values, at least up to 3 months were data is available. It should
age, is higher (~45%), and is noticed mostly from 6 to 12 months be noted that the mortars found in the literature were fabricated
mortar age. with NHL3.5 and sand, however of different provenance, and
The mortar with 25% binder (NHL2_25) presents the highest applying different mortar mix designs, and, in some cases, curing
flexural strength values of the three NHL2 mortar mixes. Specifi- conditions.
cally, up to 6 months, NHL2_25 and NHL2_30, present similar flex-
ural strength values, especially taking standard deviation into 3.2.3.2. Compressive strength of hardened mortars. Fig. 7 presents
account. However, at 12 months NHL2_25 outperforms the other the values of compressive strength of the hardened mortars at dif-
two NHL2 mortar mixes, on account of a high increase in flexural ferent mortar ages. The average value for each age is presented
strength from 6 to 12 months (~80%). This is indicative that the along with error bars, corresponding to the standard deviation of
25% binder percentage leads to a NHL2 mortar with enhanced each test.
mechanical performance, while the intense increase in flexural As a general comment, it is obvious that all NHL2 mortars pre-
strength highlights the necessity of examining mechanical values sent compressive strength values above 2 MPa at 28 days mortar
at higher ages before concluding on the appropriate mortar mix age, while compressive strength does not exceed 7 MPa at any
design parameters for each case scenario. age; thus, all mortar mixes are within the tolerance ranges stated
Regarding the NHL3.5 mortars’ flexural strength, again the mor- by the respective standard [25].
tar with the lowest binder percentage (NHL3.5_20) presents the The NHL2 mortar with the lowest binder percentage (NHL2_20)
lowest flexural strength values in this category, for all mortar ages. presents the lowest values of compressive strength at all ages, sim-
Although the aforementioned mortar mix design presents higher ilarly to flexural strength. As in the case of flexural strength, the
flexural strength values than all NHL2 mortars up to six months mortar with the highest percentage of binder (NHL2_30) presents
mortar age, it is obvious that at 12 months, flexural strength of intermediate compressive strength values of the three NHL2 mor-
NHL2_25 surpasses flexural strength of NHL3.5_20. It should be tars, at all ages, while the mortar with 25% binder (NHL2_25) pre-
noted that the increase of NHL3.5_200 s flexural strength is negligi- sents the highest compressive strength values at all ages, thus
ble up to 6 months mortar age and extremely slight from 6 to indicating that the 25% binder percentage corresponds to an
12 months (altogether ~ 36% from 1 to 12 months). NHL2 mortar with enhanced mechanical performance. It should
NHL3.5_25 and NHL3.5_30 present distinctly higher flexural be noted that although compressive strength of NHL2_25 and
strength values than all NHL2 mortars, as well as NHL3.5_20, at NHL2_30 increases significantly with mortar age (101% and 92%
all ages. The NHL3.5 mortars with 25% binder (NHL3.5_25) and respectively, from 1 to 12 months mortar age), this is not the case
with 30% binder (NHL3.5_30) present similar flexural strength val- in NHL2_20, where the highest value is reached at 3 months, from
ues, taking into account the standard deviation of the measure- which point compressive strength is relatively constant, taking
ments. NHL3.5_30 seems to present higher flexural strength, standard deviation into account. The slight increase of compressive
marginally above NHL3.5_25. The increase of NHL3.5_25 and strength at higher mortar ages in the lowest binder percentage
NHL3.5_30 from 1 to 12 months is 29% and 18% respectively, thus NHL2 mortar (~34% from 1 to 12 months) is perhaps associated
indicating early acquisition of flexural strength. with the lower percentage of binder and the easier and faster evo-
Comparing the flexural strength of the NHL2 mortars investi- lution of carbonation and hydration processes [2].
gated in this study with the respective values presented in relevant The NHL3.5 mortars present distinctly higher compressive
literature [3,11,15,28,29], it is noticed that all fabricated mortars strength values than the NHL2 mortars at all ages, except for
present intermediate values to those found in the literature NHL3.5_20, the mortar with the lowest binder percentage. The lat-
(Fig. 5). Regarding the NHL3.5 mortars (Fig. 6), it is noticed that ter presents lower compressive strength values than NHL2_25, at
NHL3.5_25 and NHL3.5_30 present higher flexural strength than all ages, and even lower than NHL2_30, at ages higher than

Fig. 5. Comparison of flexural strength of NHL2 mortars found in the literature and flexural strength measured in the current study for different mortar ages.

9
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

Fig. 6. Comparison of flexural strength of NHL3.5 mortars found in the literature and flexural strength measured in the current study for different mortar ages.

Fig. 7. Compressive strength evolution in relation to mortar age for all mortar mixes.

3 months. NHL3.5_20 presents compressive strength values mar- (with the exception of NHL2_20). Specifically, NHL3.5_20,
ginally lower than 3.5 MPa at 28 days, while at higher ages it NHL3.5_25 and NHL3.5_30 present an increase in compressive
exceeds this value. This is an indication that a 20% binder is insuf- strength of 43%, 35% and 42% respectively, indicating earlier acqui-
ficient for NHL3.5 mortars, at least for a 15 cm consistency. sition of mechanical strength, perhaps linked to the higher
As in the case of flexural strength, NHL3.5_25 and NHL3.5_30 hydraulicity of the binder type and the fact that hydration occurs
present higher compressive strength values than all other mortar mostly at early ages [3,18,29].
mixes, above 3.5 MPa at all ages. Again, the two mortar mixes pre- Comparing compressive strength (Fig. 8) of the NHL2 mortars
sent similar values, at all ages, especially taking standard deviation investigated in this study with relevant research results found in
into account; however, NHL3.5_30, seems to present a slightly the literature [3,9,11,15,28,29], it is indicated that the NHL2 mortar
enhanced mechanical performance at most ages, especially at produced with 25% NHL2 presents an enhanced mechanical perfor-
12 months. This is a further indication that by selecting a binder mance, followed by NHL2_30, while the mortar with 20% NHL2
percentage of 30% by weight an NHL3.5 mortar with enhanced presents values intermediate the values found in the literature.
mechanical performance is achieved, in contrast to the NHL2 mor- Regarding the NHL3.5 mortars (Fig. 9), NHL3.5_30 demonstrates
tars, where the best mechanical performance was clearly noticed an enhanced performance in relation to the values found in the
for the mortar mix with 25% binder. The NHL3.5 mortars present respective literature [4,8,15,29,31–34], for all mortar ages, fol-
a lower increase of compressive strength when mortar age lowed by NHL3.5_25. Again the mortar with the lowest binder per-
increases from 1 to 12 months in relation to the NHL2 mortars centage, NHL3.5_20 presents intermediate compressive strength
10
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

Fig. 8. Comparison of compressive strength of NHL2 mortars found in the literature and compressive strength measured in the current study for different mortar ages.

Fig. 9. Comparison of compressive strength of NHL3.5 mortars found in the literature and compressive strength measured in the current study for different mortar ages.

values in relation to those found in the literature for NHL3.5 mor- Research related to aerial lime mortars [66], has indicated an
tars. It should be noted that the mortars found in the literature increase of mechanical strength as binder percentage increases;
were fabricated with NHL (NHL2 and NHL3.5, respectively) and however, this increase is noticed up to a certain B/A ratio (limit)
sand, of different provenance, and applying different mortar mix and thereafter a decline in compressive strength occurs. This has
designs, and, in some cases, curing conditions. been attributed to the production of internal and surface cracks
Overall, regarding both mortar types, it is noticed that the because of binder increment beyond a certain limit [66]. Further-
increase in mechanical strength values in relation to mortar age more, it has been proposed that additionally a certain portion of
is affected by the mix design parameters. The strength increment the lime fraction abandons its binder function and acts as filler
noticed at later ages for higher binder percentages (intensely in [66]. The same trend was also noticed in a similar study regarding
the case of NHL2), has also been noticed by other researchers [2] hydraulic lime mortars (HL5) [2]. In the current study, the afore-
and has been attributed to later C2S hydration and portlandite car- mentioned trend is noticed in the case of NHL2 mortars, and 25%
bonation on account of the higher presence of NHL in the mortar binder percentage seems to be the limit value of binder percentage.
mix design. Although surface cracks were not detected on any of the specimens
It is worth noting that NHL3.5 mortars present an increase of macroscopically, perhaps the increment beyond 25% in the case of
compressive strength values (as measured at 12 months mortar NHL2 mortars, leads to the development of non-discernible and/or
age) as the binder content increases. However, this is not the case internal micro-cracks and the behavior of a certain part of the bin-
for the NHL2 mortars, where the highest compressive strength is der as filler; the above would modify the development of pores
achieved for an intermediate B/A ratio. This is highly interlinked higher than 1 lm and lower than 0.05 lm and negatively affect
with the developed pore network structure of each mortar type. mechanical performance. In the case of NHL3.5 it is supposed that
11
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

Fig. 10. Correlation of total open porosity (%) and compressive strength (MPa) at 12 months mortar age.

the same limit could be higher, 30%, on account of its higher 3.2.3.3. Dynamic modulus of elasticity of hardened mortars. The spec-
hydraulicity as a binder, and, thus, the same trend is not noticed imens were examined in relation to their dynamic modulus of elas-
in the case of NHL3.5, on account of the B/A range applied in the ticity, through ultrasonic pulse velocity measurements (Table 7).
current study. The ultrasonic pulse velocity values are similar to values reported
In order to examine the relationship between pore structure in the literature for NHL3.5 mortars [33] and are expected and sat-
and mechanical performance, the 12 months compressive strength isfactory values for natural hydraulic lime mortars [40]. The
value of the mortars is taken into account. It is interesting to note dynamic modulus of elasticity presents a wide distribution
that total open porosity seems to present an inverse linear rela- amongst mortar mixes; it is noticed that mortars with higher com-
tionship with the NHL mortars’ compressive strength (Fig. 10). pressive strength values, present higher values of dynamic modu-
The same trend is noticed regarding flexural strength; however, a lus of elasticity; this is interlinked of course with the mortars’ pore
separate inverse linear relationship is noticed for NHL2 mortars network structure, as a more compact system leads to higher com-
and NHL3.5 mortars, according to the type of binder. The increase pressive strength, however a less ductile material.
of compressive strength, with decrease of porosity, as a trend
regarding porous materials, is also stated by other researchers
(e.g. [57,67]). 4. Conclusions
Regarding the pore structure developed, it is noticed that com-
pressive strength is inversely affected by the most frequent pore In the present study, different mortar mixes, produced with two
radii; as the pore structure shifts towards narrower pore radii different types of natural hydraulic lime (NHL2, NHL3.5), in accor-
(Fig. 2), compressive strength increases (Fig. 7). dance to the updated EN 459–1 standard, and different binder con-
Additionally, larger pores, are considered to have a negative tents, ranging from 20 to 30%, were examined regarding their
effect on mechanical performance. Especially the ITZ zone, when physical and mechanical performance, while maintaining a con-
it is weaker than the binder, may act as a plane of weakness, from stant consistency for the mortars of each type.
which possible crack propagation initiates [67]. Thus, it is not sur- The main pore network structure characteristics of the different
prising that the mortars’ compressive strength (Fig. 7) seems to mortars are found within a relatively narrow range, however pore
present a strong inverse linear relationship with the relative pore size distribution presents differentiations. The least dense pore
volume in the pore radius range > 1 lm (Table 5). This relationship network structure is noticed in both cases for the 20% binder per-
is different according to the binder type, while the NHL2 mortars centage. In the case of NHL3.5, increase of the binder percentage
seem to be more intensely affected perhaps on account of the shifts the pore size distribution towards narrower pores, while in
lower strength binder. the case of NHL2 this is noticed only when binder increases from
20 to 25%.
The capillary rise coefficient of the mortars, in the case of the
Table 7
Ultrasonic pulse velocity and dynamic modulus of elasticity of the examined mortars NHL3.5 mortars, decreases as the binder percentage of the mortar
at 12 months mortar age. mix design increases. This is not the case however for the NHL2
mortars, where the lowest value of capillary rise coefficient is
Mortar Mix Ultrasonic Standard Dynamic Standard
pulse Deviation modulus of Deviation noticed for the intermediate binder percentage, 25%. The different
velocity (m/s) elasticity (GPa) effect of binder percentage on the water absorption rate is inter-
NHL2_20 2,003 8 6.29 0.04 linked with the different pore network and distribution developed;
NHL2_25 2,279 9 8.47 0.05 a higher relative volume of larger capillary pores (radii > 0.5 lm)
NHL2_30 2,144 12 7.65 0.09 seems to increase C.R.C., a trend perhaps interlinked with the role
NHL3.5_20 2,098 14 7.00 0.04 of larger capillary pores, to accumulate water and act as reservoirs
NHL3.5_25 2,564 17 10.51 0.10
in the first, linear stage of absorption. This effect is different
NHL3.5_30 2,597 10 10.82 0.02
according to the binder type.
12
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

The evaluation of the mechanical performance of the mortars, [4] C. Figueiredo M. Lawrence R. Ball Mechanical properties of standard and
commonly formulated NHL mortars used for retrofitting 2016 University of
demonstrated that different mechanical strength values can be
Bath (p. 68).
achieved, by altering the binder percentage and the type of NHL. [5] A. Moropoulou, A. Bakolas, S. Anagnostopoulou, Composite materials in
In both NHL types, the 20% binder percentage leads to a lower ancient structures, Cem. Concr. Compos. 27 (2) (2005) 295–300.
mechanical performance. Mechanical performance of the NHL3.5 [6] B.A. Silva, A.P. Ferreira Pinto, A. Gomes, Influence of natural hydraulic lime
content on the properties of aerial lime-based mortars, Constr. Build. Mater. 72
mortars is enhanced as the binder percentage increases, while in (2014) 208–218.
the case of the NHL2 mortars the best mechanical performance is [7] A. Henry, Hot-mixed mortars: the new lime revival, Context Institure Hist
noticed for the intermediate binder percentage, 25%, indicating Build Conserv 154 (2018) 30–33.
[8] P. Faria, V. Silva, T. Madeira. Natural hydraulic lime mortars with ceramic
that this is perhaps a limit binder percentage. NHL3.5 mortars residues for masonry In 9th International Masonry Conference 2014 (pp. CD-
seem to acquire their mechanical strength at earlier ages than ROM). Universida de do Minho/Instituto para Sustentabilidade em Engenharia
NHL2 mortars, a trend interlinked with the different hydraulicity Estrutural (ISISE)/International Masonry Society (IMS).
[9] K. Luo, J. Li, Z. Lu, J. Jiang, Y. Niu, Effect of nano-SiO2 on early hydration of
of the different binders. The dynamic modulus of elasticity seems natural hydraulic lime, Constr. Build. Mater. 216 (2019) 119–127.
to be dependent on the developed compressive strength of the dif- [10] R. Hanley, S. Pavía, A study of the workability of natural hydraulic lime mortars
ferent mortars. An inverse linear relationship in noticed between and its influence on strength, Mater. Struct. 41 (2) (2008) 373–381.
[11] J.S. Pozo-Antonio, Evolution of mechanical properties and drying shrinkage in
total open porosity and compressive strength. Furthermore, it lime-based and lime cement-based mortars with pure limestone aggregate,
seems that a higher amount of pores with a radius  1 lm, nega- Constr. Build. Mater. 77 (2015) 472–478.
tively affects compressive strength, with a different intensity [12] A. Isebaert, W. De Boever, V. Cnudde, L. Van Parys, An empirical method for the
estimation of permeability in natural hydraulic lime mortars, Mater. Struct. 49
according to the binder type.
(11) (2016) 4853–4865.
As both hygric and mechanical performance are interlinked [13] M. Amenta, I. Karatasios, P. Maravelaki- Kalaitzaki, V. Kilikoglou, The role of
with the developed pore network structure, it is not surprising that aggregate characteristics on the performance optimization of high
NHL mortars which develop higher strength also present lower hydraulicity restoration mortars, Constr. Build. Mater. 153 (2017) 527–534.
[14] M. Aly, S. Pavia, Mechanical and hygric properties of natural hydraulic lime
C.R.C. values. NHL3.5 mortars tend to present higher mechanical (NHL) mortars with pozzolans, Assessment and Rehabilitation of Civil
strength values and lower C.R.C. values in relation to NHL2 mor- Structures, SMAR, 2015.
tars; however, when the binder percentage is insufficient (20%), [15] S. Barr, W.J. McCarter, B. Suryanto, Bond-strength performance of hydraulic
lime and natural cement mortared sandstone masonry, Constr. Build. Mater.
the resulting pore network structure can even lead to a reversal 84 (2015) 128–135.
of this observation. [16] J.-S. Cho, K.-Y. Moon, M.-K. Choi, K.-H. Cho, J.-W. Ahn, K.-S. Yeon, Performance
This study highlights the importance of selecting an appropriate improvement of local Korean natural hydraulic lime-based mortar using
inorganic by-products, Korean J. Chem. Eng. 34 (5) (2017) 1385–1392.
binder to aggregate ratio, consistency and type of NHL, in order to [17] C. Giosuè, Q.L. Yu, M.L. Ruello, F. Tittarelli, H.J.H. Brouwers, Effect of pore
achieve the desired mechanical and physical performance of a structure on the performance of photocatalytic lightweight lime-based
mortar, while it reveals the different influence of the binder to finishing mortar, Constr. Build. Mater. 171 (2018) 232–242.
[18] M. Apostolopoulou, P.G. Asteris, D.J. Armaghani, M.G. Douvika, P.B. Lourenço, L.
aggregate ratio on the mechanical and physical characteristics of Cavaleri, A. Bakolas, A. Moropoulou, Mapping and holistic design of natural
a mortar according to the type of NHL. hydraulic lime mortars, Cem. Concr. Res. 136 (2020) 106167, https://doi.org/
10.1016/j.cemconres.2020.106167.
[19] A. Kalagri, I. Karatasios, V. Kilikoglou, The effect of aggregate size and type of
CRediT authorship contribution statement binder on microstructure and mechanical properties of NHL mortars, Constr.
Build. Mater. 53 (2014) 467–474.
[20] C. De Nardi, C. Brito de Carvalho Bello, L. Ferrara, A. Cecchi, (2019). Self-healing
Maria Apostolopoulou: Investigation, Data curation, Writing - lime mortars: an asset for restoration of heritage buildings. In Sustainable
original draft, Writing - review & editing. Asterios Bakolas: Materials Systems and Structures SMSS 2019 (Vol. 128, pp. 652-659). FRA.
Conceptualization, Investigation, Methodology, Supervision, [21] A.M. Forster, E.M. Szadurski, P.F.G. Banfill, Deterioration of natural hydraulic
lime mortars, I: Effects of chemically accelerated leaching on physical and
Validation, Writing - review & editing. Meletis Kotsainas: mechanical properties of uncarbonated materials, Constr. Build. Mater. 72
Investigation, Writing - original draft. (2014) 199–207.
[22] P.F.G. Banfill, E.M. Szadurski, A.M. Forster, Deterioration of natural hydraulic
lime mortars, II: Effects of chemically accelerated leaching on physical and
Declaration of Competing Interest mechanical properties of carbonated materials, Constr. Build. Mater. 111
(2016) 182–190.
[23] European Committee for Standardization (CEN). (2001). Building limes, part 1:
The authors declare that they have no known competing finan-
Definitions, specifications and conformity criteria. EN 459-1.
cial interests or personal relationships that could have appeared [24] CEN, En 459–1:2010—Building Lime. Part 1: Definitions, Specifications and
to influence the work reported in this paper. Conformity Criteria 2010 CEN, Brussels, Belgium.
[25] CEN, En 459–1:2015—Building Lime. Part 1: Definitions, Specifications and
Conformity Criteria 2015 CEN, Brussels, Belgium.
Acknowledgements [26] J. Grilo, P. Faria, R. Veiga, A. Santos Silva, V. Silva, A. Velosa, New natural
hydraulic lime mortars–physical and microstructural properties in different
curing conditions, Constr. Build. Mater. 54 (2014) 378–384.
The authors would like to thank TITAN S.A. for their assistance [27] M. Apostolopoulou, D.J. Armaghani, A. Bakolas, M.G. Douvika, A. Moropoulou,
in the production of the mortars, as well as in studying fresh mor- P.G. Asteris, Compressive strength of natural hydraulic lime mortars using soft
tar and mechanical characteristics, as well as Dalkafouki Oikos Ltd computing techniques, Procedia Structural Integrity 17 (2019) 914–923.
[28] C. Figueiredo, M. Lawrence, R.J. Ball, (2016). Chemical and physical
for the provision of the raw materials used for the production of characterisation of three NHL 2 binders and the relationship with the mortar
the mortars. properties. In REHABEND 2016, Euro-Americal Congress: Construction
Pathology, Rehabilitation Technology and Heritage Management. Santander
Spain.
References [29] C. Figueiredo, R.J. Ball, M. Lawrence, (2016). Is BS EN 459-1: 2015 fit for
purpose in the context of conservation?. In: The Journal of the Building Limes
[1] P. Maravelaki-Kalaitzaki, A. Bakolas, A. Moropoulou, Physico-chemical study of Forum (Vol. 23, pp. 46-52). Building Limes Forum.
Cretan ancient mortars, Cem. Concr. Res. 33 (5) (2003) 651–661. [30] M. Vyšvařil, T. Žižlavský, Š. Zimmermann, P. Bayer, (2017). Effect of Aggregate
[2] J. Lanas, J.L. Pérez Bernal, M.A. Bello, J.I. Alvarez Galindo, Mechanical properties Type on Properties of Natural Hydraulic Lime-Based Mortars. In: Materials
of natural hydraulic lime-based mortars, Cem. Concr. Res. 34 (12) (2004) Science Forum (Vol. 908, pp. 35-39). Trans Tech Publications Ltd.
2191–2201. [31] L. Garijo, X. Zhang, G. Ruiz, J.J. Ortega, Z. Wu, The effects of dosage and
[3] D. Zhang, J. Zhao, D. Wang, C. Xu, M. Zhai, X. Ma, Comparative study on the production process on the mechanical and physical properties of natural
properties of three hydraulic lime mortar systems: Natural hydraulic lime hydraulic lime mortars, Constr. Build. Mater. 169 (2018) 325–334.
mortar, cement-aerial lime-based mortar and slag-aerial lime-based mortar, [32] P. Faria, V. Silva, in: Historic Mortars, Springer International Publishing, Cham,
Constr. Build. Mater. 186 (2018) 42–52. 2019, pp. 185–199, https://doi.org/10.1007/978-3-319-91606-4_14.

13
M. Apostolopoulou, A. Bakolas and M. Kotsainas Construction and Building Materials 290 (2021) 123272

[33] A. Arizzi, G. Martinez-Huerga, E. Sebastián-Pardo, G. Cultrone, Mineralogical, [50] E. Aggelakopoulou, A. Bakolas, A. Moropoulou, Lime putty versus hydrated
textural and physical-mechanical study of hydraulic lime mortars cured under lime powder: Physicochemical and mechanical characteristics of lime based
different moisture conditions, Materiales de Construccíon 65 (318) (2015) mortars, Constr. Build. Mater. 225 (2019) 633–641.
e053, https://doi.org/10.3989/mc.2015.v65.i31810.3989/mc.2015.03514. [51] ASTM C597-83. (1989). Standard Test Method for Pulse Velocity through
[34] L. Garijo, M. Azenha, M. Ramesh, P.B. Lourenço, G. Ruiz, Stiffness evolution of Concrete. In Annual Book of ASTM Standards, American Society for Testing
natural hydraulic lime mortars at early ages measured through EMM-ARM, Materials, Philadelphia.
Constr. Build. Mater. 216 (2019) 405–415. [52] T.E. Boothby, B.T. Rosson, Elasto-plastic Hardening and Shakedown of Masonry
[35] D. Zhang, J. Zhao, D. Wang, Y. Wang, X. Ma, Influence of pozzolanic materials Arch Joints, Meccanica 34 (2) (1999) 71–84.
on the properties of natural hydraulic lime based mortars, Constr. Build. Mater. [53] C. Giordani, A.B. Masuero, Blended mortars: Influence of the constituents
244 (2020) 118360, https://doi.org/10.1016/j.conbuildmat.2020.118360. and proportioning in the fresh state, Constr. Build. Mater. 210 (2019)
[36] EN 933-1: 2012: Tests for geometrical properties of aggregates. Determination 574–587.
of particle size distribution. Sieving method; European Commitee for [54] I. Papayianni, M. Stefanidou, Mortars for intervention in monuments and
Standardisation: Brussels, Belgium, 2012. historical buildings, WIT Trans. Built Environ. 66 (2003).
[37] EN 13139:2002: Aggregates for mortar; European Commitee for [55] A. Isebaert, L. Van Parys, V. Cnudde, T. De Kock, J.M. Baele, in: Historic Mortars,
Standardisation: Brussels, Belgium, 2002. Springer International Publishing, Cham, 2019, pp. 169–184, https://doi.org/
[38] EN 933-8:2012: Tests for Geometrical Properties of Aggregates. Assessment of 10.1007/978-3-319-91606-4_13.
Fines. Sand Equivalent Test; European Commitee for Standardisation: Brussels, [56] M. Apostolopoulou, V. Keramidas, N. Galanaki, M. Kalofonou, C. Skoula, M.
Belgium, 2012. Karoglou, E.T. Delegou, C. Mouzakis, A. Bakolas, A. Moropoulou, M. Pikoula, A.
[39] EN 1097-3:1998: Tests for mechanical and physical properties of aggregates - Kalagri, E. Farmakidou, M. Michailidou, A study on the historical materials of
Part 3: Determination of loose bulk density and voids; European Commitee for the Apollo Pythios Temple in Rhodes and the evaluation of potential
Standardisation: Brussels, Belgium, 1998. restoration materials, Heritage 2 (1) (2019) 988–1022.
[40] Schafer, J., Hilsdorf, H. K. (1993, June). 74. Ancient and new lime mortars–the [57] M. Thomson, J.E. Lindqvist, J. Elsen, C.J.W.P. Groot, (2004). 2.5 Porosity of
correlation between their composition, structure and properties. mortars. In Characterisation of Old Mortars with Respect to their Repair-Final
In Conservation of stone and other materials: proceedings of the Report of RILEM TC 167-COM, 77-106.
international RILEM/UNESCO congress held at the UNESCO headquarters, [58] A.R. Santos, M.do.Rosário. Veiga, A. Santos Silva, J. de Brito, José.I. Álvarez,
Paris, June 29-July 1, 1993 (pp. 605-612). Evolution of the microstructure of lime based mortars and influence on the
[41] CEN. (2010). EN 459-2: 2010, Building lime–Part 2: Test methods, CEN, mechanical behaviour: the role of the aggregates, Constr. Build. Mater. 187
Brussels, Belgium, 2010. (2018) 907–922.
[42] EN1015-11:2007: Methods of Test for Mortar for Masonry—Part 11. [59] Burwell Jr, R. L. (1977). Manual of symbols and terminology for
Determination of Flexural and Compressive Strength of Hardened Mortars; physicochemical quantities and units—appendix II heterogeneous catalysis.
European Commitee for Standardisation: Brussels, Belgium, 2007. In Advances in Catalysis (Vol. 26, pp. 351-392). Academic Press.
[43] EN1015-3:2007: Methods of Test for Mortar for Masonry—Part 3. [60] F. Gong, D. Zhang, E. Sicat, T. Ueda, Empirical estimation of pore size
Determination of Consistency of Fresh Mortar (by Flow Table); European distribution in cement, mortar, and concrete, J. Mater. Civ. Eng. 26 (7) (2014)
Commitee for Standardisation: Brussels, Belgium, 2007. 04014023, https://doi.org/10.1061/(ASCE)MT.1943-5533.0000945.
[44] EN1015-6:2007: Methods of Test for Mortar for Masonry—Part 6: [61] N. Karagiannis, M. Karoglou, A. Bakolas, A. Moropoulou, Building materials
Determination of Bulk Density of Fresh Mortar; European Commitee for capillary rise coefficient: Concepts, determination and parameters involved,
Standardisation: Brussels, Belgium, 2007. in: New Approaches to Building Pathology and Durability, Springer, Singapore,
[45] EN 1015-7: 2007. Methods of test for mortar masonry – part 7: determination 2016, pp. 27–44.
of air content of fresh mortar; European Commitee for Standardisation: [62] B.A. Silva, A.P. Ferreira Pinto, A. Gomes, Natural hydraulic lime versus cement
Brussels, Belgium, 2007. for blended lime mortars for restoration works, Constr. Build. Mater. 94 (2015)
[46] EN1015-8:2007: Methods of Test for Mortar for Masonry—Part 8. 346–360.
Determination of Retained Water of Fresh Mortar; European Commitee for [63] R. Nogueira, A.P.F. Pinto, A. Gomes, Design and behavior of traditional lime-
Standardisation: Brussels, Belgium, 2007. based plasters and renders. Review and critical appraisal of strengths and
[47] CNR-ICR, Normal 4/80: Distribution of Pore Volume as a Function of Pore weaknesses, Cem. Concr. Compos. 89 (2018) 192–204.
Diameter C.N.R.-I.C.R Consiglio Nazionale delle Ricerche and Istituto Centrale [64] A. Bentur, M.G. Alexander, A review of the work of the RILEM TC 159-ETC:
del Restauro (CNR-ICR) 1980 Rome, Italy Engineering of the interfacial transition zone in cementitious composites,
[48] EN15801:2010: Conservation of Cultural Property-Test Methods- Mater. Struct. 33 (2) (2000) 82–87.
Determination of Water Absorption by Capillarity; European Commitee for [65] K.K. Aligizaki, Pore structure of cement-based materials: testing,
Standardisation: Brussels, Belgium, 2010. interpretation and requirements, Crc Press, 2005.
[49] Normal 07–81. Assorbimento d’ acqua per immersion totale–Capacita di [66] J. Lanas, José.I. Alvarez-Galindo, Masonry repair lime-based mortars: factors
imbibizione. CNR Centri di Studio di Milano e Roma sulle cause di deperimento affecting the mechanical behavior, Cem. Concr. Res. 33 (11) (2003) 1867–1876.
e sui metodi di conservazione delle opere d’arte–ICR Istituto Centrale per il [67] A.M. Neville, Properties of concrete, Vol. 4, Longman, London, 1995.
Restauro 1981 Rome, Italy

14