Construction and Building Materials 61 (2014) 340–348

Contents lists available at ScienceDirect

Construction and Building Materials

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

Mechanical properties and durability of hemp-lime concretes

R. Walker a,⇑, S. Pavia a, R. Mitchell b
a
Dept. of Civil Engineering, Trinity College, Dublin, Ireland
b
School of Engineering and Applied Biology, Harvard University, Cambridge, MA 02138, USA

h i g h l i g h t s

 The effect of binder type on mechanical strength and durability is investigated.

 At the hemp interface, the commercial binder has abundant hydrates while lime:pozzolan binders are mainly carbonated.
 Binder hydraulicity contributes to strength development but has a lesser effect at 1 year when other factors also contribute.
 Hemp concretes made with lime:pozzolan binders are more sensitive to freeze:thaw action than more hydraulic binders.
 Short-term salt exposure resulted in the precipitation of salt layers but did not impact compressive strength at 1 year.

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

Article history: Hemp-lime concrete is a sustainable and carbon negative construction material. This paper investigates
Received 15 November 2013 the effect of binder type on mechanical strength and durability (resistance to freeze–thaw, salt exposure
Received in revised form 24 February 2014 and biodeterioration). It compares hemp-lime concretes made with a hydrated lime and pozzolan binder
Accepted 25 February 2014
to those including hydraulic lime and cement.
Available online 9 April 2014
SEM analysis revealed abundant hydrates at the hemp interface of the strongly hydraulic commercial
binder while the lime:pozzolan binders were mostly carbonated. Increasing binder hydraulicity enhances
Keywords:
early strength development however, all concretes achieved similar compressive strengths at 1 year irre-
Hemp-lime concrete
Pozzolans
spective of the binder type. The concretes with lime:pozzolan binders are more sensitive to freeze:thaw
Strength action than those with more hydraulic binders. Salt exposure resulted in the precipitation of salt layers in
Freeze:thaw action the concrete however, this did not have a detrimental impact on the compressive strength of the concrete
Salt exposure at 1 year. The results evidenced that hemp concrete is resistant to biodeterioration (7 month exposure).
Finally, the addition of water retainer improved the early strength development and freeze:thaw resis-
tance of the concrete with lime–pozzolan binder.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction One of its most outstanding environmental qualities of the con-

crete is that it is carbon negative. CO2 is emitted during the produc-
The built environment is responsible for 40% of primary energy tion of lime however this is offset by the carbon sequestering of the
use and 36% of energy related CO2 emissions [1]. Therefore, it is hemp. As hemp grows it absorbs CO2 and the carbon is integrated
important to develop low-embodied energy, carbon-negative, sus- into the structure of the plant. The carbon therefore remains
tainable construction materials to replace cement-based products. trapped in the hemp-lime walls. Boutin et al. (2006) determined
Hemp-lime concrete is a mix of a lime-based binder and hemp. It that 1 m2 of hemp-lime wall (260 mm thick) requires 370–
was developed in the late 1980s/early 1990s in France as an alter- 394 MJ of energy for production and sequesters 14–35 kg of CO2
native to wattle and daub for the restoration of historic buildings over its 100 year life span [2]. In addition to its sustainable creden-
and to lighten portland cement (PC) concrete. Its potential as an tials, hemp-lime concretes exhibit an excellent thermal perfor-
environmentally sustainable construction material was quickly mance; a high thermal capacity coupled with a medium density
realised and it has since been used in the construction and thermal and a low thermal conductivity that provides the concretes with
upgrading of hundreds of buildings in Europe. good insulation capability.
The aim of this research is to investigate the effect of the type of
⇑ Corresponding author. binder on the mechanical strength and durability (resistance to
E-mail addresses: [email protected] (R. Walker), [email protected] (S. Pavia), freeze–thaw, salt exposure and biodeterioration) of hemp-lime
[email protected] (R. Mitchell). concrete. Currently, PC and hydraulic lime are added to hemp-lime

http://dx.doi.org/10.1016/j.conbuildmat.2014.02.065
0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.
 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348 341

concrete to speed up setting and hardening however, using pozzo- The principal concerns of PC concrete durability include freeze–
lans instead should lower environmental impact. This paper com- thaw action, sulphate attack, alkali-aggregate reaction, chloride
pares hemp-lime concretes made with hydrated lime and penetration and carbonation. Considering the lack of tri-calcium
pozzolans to those including hydraulic lime and cement. Two aluminate, silica based aggregate or steel reinforcement in the
pozzolans, metakaolin and GGBS, were identified as having poten- hemp concrete, it was considered that freeze–thaw action, salt
tial for use in hemp-lime concrete on account of their fast setting crystallization and biodeterioration were amongst the most likely
and high reactivity [3,4]. GGBS is a by-product of the iron and steel aggressors for the hemp concrete that have not been widely
manufacturing process. GGBS is typically a latent hydraulic mate- researched.
rial and hydrates in the presence of water, and as such it is some- In relation to freeze:thaw, the ability of a mortar to resist free-
times not considered a true pozzolan. The self hydration of GGBS ze:thaw action partly depends on its pore structure being able to
however is very slow and lime acts as an activator [5]. The hydra- withstand strain caused by volume changes brought about by
tion reaction of GGBS is accompanied by the slower lime:GGBS water freezing. Botas et al. (2010) reviewed the freeze:thaw resis-
pozzolanic reaction: the amorphous silica and alumina in the slag tance of lime and lime:cement mortars and noted that, as lime
react with lime forming additional hydrates. The replacement of PC hydraulicity increased, resistance to freeze:thaw decreased likely
by GGBS requires lower energy consumption and reduced CO2 due to the presence of smaller pores [22]. They also observed that,
emissions [6]. GGBS is created by a polluting industry however, it despite increasing open porosity, air entraining agents did not im-
is a waste product that would otherwise be disposed of in landfill. prove mortar resistence to freeze:thaw. The lime:cement mortars
Metakaolin is a calcined kaolin clay that reacts with lime and forms however displayed a greater resistence, as the effect of their supe-
hydrates: calcium silica hydrate (CSH) and calcium aluminosili- rior mechanical strength supercedes the pore structure character-
cates. Metakaolin is a less energy intensive processed material than istics. A ‘‘wall mixture’’ hemp-lime concrete resisted 20 cycles of
cement [7]. severe freeze–thaw (less than 1% loss in apparent volume) but it
The hemp aggregate absorbs large quantities of water (325% of was noted that overly-hydraulic or unsuitable limes altered after
its own weight at 24 h [8]) and this can undermine hydration two cycles [9 referring to 23]. Further research indicates that 25
therefore, some of the concretes investigated include a water freeze–thaw cycles does not have a negative influence on the com-
retainer. pressive strength of the concrete [23].
Hemp concrete is largely a non-load bearing material that is Soluble salts greatly limit the durability of porous building
typically used with a load bearing frame. However, its compressive materials. The performance of hemp-lime concrete following salt
strength is important as it is the most commonly measured prop- exposure has not yet been reported. This paper investigates the
erty and allows comparison between the different binders. Fur- performance of hemp-lime concrete following exposure to sodium
thermore, the early compressive strength development of the chloride (NaCl), one of the most common salts found in buildings
concrete is important as the hemp concrete must support its and one of the main causes of material decay. Several theories exist
own self weight when wet (which can be over double its dry on the mechanism of salt damage in inorganic porous materials
weight). The early strength of the hemp concrete may influence [24]. Lubelli et al. (2006, 2010) studied NaCl in mortar and showed
when shuttering is removed during construction and the spacing crystallization of salt layers on pore walls [25,26]. However, under
of the supporting structural timber frame. Typical compressive the salt layer, most of the distinguishable pores appear empty.
strengths for 2:1 (binder:hemp by weight) mixes range from 0.2 Benavente et al. (2004) observed that NaCl crystallizes heteroge-
to 0.12 MPa [9–14] mainly depending on density, binder type neously on pore walls (of porous stone), and that there is a strong
and age. Flexural strength is also low and previous research has interaction between the salt and the pore wall [27].
determined strengths between 0.06 and 1.2 MPa [12,15–17]. The
low strength of the concrete is attributed by Bouloc et al. (2006) 2. Materials and methods
to the ductile nature of the hemp particles and their disordered
arrangement [18]. Nguyen (2010) believes that the high porosity 2.1. Materials
of the shiv gives the concrete a lower mechanical strength com-
A hydrated lime (CL90s—calcium lime) and a hydraulic lime NHL 3.5 complying
pared to other lightweight concretes [16]. with EN 459-1 [28] and Portland cement (CEM I) complying with EN197-1:2011
The contribution of the binder’s hydraulicity to the strength of [29] were used (only the builder’s mix includes PC). The ‘‘commercial mix’’ includes
the concrete has yielded varying opinions. Hirst et al. (2010) found a proprietary, lime-based binder with hydraulic additions developed for use with
that strength does not increase with the binder’s strength [13]. hemp. Its composition is not disclosed by the manufacturers for commercial rea-
sons. Two pozzolans: metakaolin and GGBS; were identified as having potential
However, Nguyen (2010) claims that stronger binders increase
for use in hemp-lime concrete on account of their fast setting and high reactivity
strength provided that their hydraulicity is not compromised by [3,4]. The pozzolans’ chemical composition, amorphousness and surface area are in-
the absorption of water by the hemp [16]. Murphy et al. 2010 also cluded in Table 1. The chemical composition was assessed by XRF using a Quant’X
observed that concretes made with a hydraulic commercial binder EDX Spectrometer and UniQuant analysis package. The degree of amorphousness
showed higher ultimate compressive and flexural strengths than was indicated by X-ray diffraction (XRD), using a Phillips PW1720 XRD with a
PW1050/80 goniometer and a PW3313/20 Cu ka anode tube at 40 kV and 20 mA.
those fabricated with hydrated lime, and that the rate of strength The specific surface area was measured using a Quantachrome Nova 4200e and
development depends on the hydraulicity of the binder, with the BET method, a model isotherm based on adsorption of gas on a surface. The
hydraulic binders gaining strength faster [12]. De Bruijn et al. water retainer investigated is modified hydroxypropyl methyl cellulose.
(2009) also showed that higher compressive strengths are obtained Industrial hemp shiv was supplied by La Chanvrière De L’aube in central France.
Hemp properties vary with growing conditions and harvesting, and this influences
for cement-rich binders [19].
the properties of the concrete. Therefore, hemp from the same consignment, stored
Little research has been undertaken on the durability of the in the same conditions was used in all concretes to ensure that variability of hemp
hemp concrete. However, lime has been used since antiquity often did not influence the results. The water content of the hemp depends on the relative
including organic material such as timber (lath and beams), straw humidity and also impacts the properties of the concrete and was measured as
or hair, in structures that have shown great long-term durability. In 12.4% prior to mixing.

addition, hemp-lime houses have been built in France for over

20 years without any serious reported weathering or durability 2.2. Composition of concrete

failures [20]. Empirically, the material is widely regarded as being Six mixes were studied only differing in the binder composition as set out in
resistant to mould and insects on account of the alkalinity of the Table 2. As each binder has a different water demand which depends on its compo-
lime [21]. sition, the water content could not be kept constant. Therefore, it was based on
342 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348

Table 1 2.6. Flexural strength

Chemical composition, amorphousness and surface area of pozzolans.
The flexural strength was guided by EN 196-1 [32]. Adjustments were required
Composition, amorphousness and surface area GGBS Metakaolin to accommodate the behaviour of the concrete. The load was applied normal to the
SiO2 34.14% 51.37% direction of compaction at a rate of 10 N/s. Large specimens were tested to ensure
Al2O3 13.85% 45.26% that the behaviour was representative of the concrete and not unduly affected by
CaO 39.27% the larger particle size of the hemp compared to mineral aggregate.
Fe2O3 0.41% 0.52%
SO3 2.43%
MgO 8.63% 0.55% 2.7. Resistance to freeze–thaw
Rate of amorphousness Totally Mostly
Surface area (m2/g) 2.65 18.3 Currently, there are no standards to measure the resistence of lime mortars or
hemp concretes to freeze:thaw action therefore, the test was guided by EN
15304:2010 [34]. Nine month concretes were subject to 10 freeze–thaw cycles be-
tween 15 and 20 °C.
workability which was consistent in all concretes. As no workability test currently Near saturation conditions provide the severest conditions for freeze:thaw ac-
exists, the water content was determined by the expertise of a skilled building prac- tion as the effect of expanding ice will be most detrimental. Therefore, the samples
titioner, Henry Thompson, who has built hemp-lime houses for over 10 years. were soaked for 48 h prior to freezing, time by which they had absorpted 90% of
According to Evrard (2003) experience is the best guarantee for a good mixture [9]. their total water at saturation. Abundant water condensed in the freezer therefore,
to ensure that water content remained near saturation, the samples were immersed
for 12 h after cycles 4 and 8.
Four specimens of each mix were tested. After cycling, the samples were al-
2.3. Mixing, moulding and curing
lowed to dry for 2.5 months and the weight loss calculated. The 2.5 month limit
was selected so that samples were fully dry and testing was undertaken at 1 year.
The mixing sequence in lime hemp concrete has not yet been established. Some
Finally, the compressive strength was determined and compared to those of refer-
authors wet the hemp prior to adding the binder [10,16] while others form a slurry
ence samples and 1 year concretes. The reference samples were soaked for 48 h and
with water and binder before adding the hemp [13,14]. A preliminary investigation
stored in polythene bags during testing to protect them against drying (EN
revealed that, in the lime:pozzolan concretes, wetting the hemp before the addition
15304:2010); they are used to determine whether deterioration is due to free-
of binder increases water demand and does not impart significant benefits to the
ze:thaw action or to water saturation during testing.
properties measured [8] therefore, prewetting the hemp was not considered.
Mixing was done in a large pan mixer with 2 batches per mix (total mixing time
7 min). The dry binder was premixed by hand and 3=4 of the total mixing water was 2.8. Resistance to salt exposure
then added and mixed for 2.5 min to form a slurry. The hemp and remaining water
were then gradually included. No standards currently exist to test hemp-lime concrete on salt exposure. 9-
An amount of concrete was weighted to ensure a dry density of c. 360 kg/m3. month old concretes were subject to salt weathering in a SC1000 Open Lid Salt
The density was closely controlled due to its significant effect on concrete proper- Fog Chamber. A 5% (or 0.86 M) NaCl solution was evenly dispersed as small droplets
ties. The concrete was placed into cling-film lined timber moulds in a single layer through nozzles into the chamber for 12 h. This was followed by 12 h of drying. The
and gently pressed generating a density similar to that of a typical wall construc- concretes underwent 4 weeks of cycles (the first two weeks at 20 °C and the
tion. The mould was removed and the samples transferred to a curing room at remaining time at 40 °C). They were then allowed to dry for 2 months and their
16 °C ± 3 °C temperature and 55% ± 10% relative humidity. The effect of the curing compressive strength determined. Four specimens of each concrete were tested.
conditions on the concrete is discussed in other research [30]. 100 mm cubes were In order to evaluate the resistance of the concrete to salt exposure, the compressive
moulded for compressive strength, freeze/thaw cycling and salt exposure and 100 * strength after salt exposure was compared to that of the reference samples and the
100 * 400 prisms for flexural strength. concrete values at 1 year. Comparison with the reference sample was used to deter-
mine whether deterioration is due to salt action or to water saturation during
testing.
2.4. Microstructure

The microstructure of the concretes was investigated using a Tescan MIRA Field 2.9. Biodeterioration
Emission Scanning Electron Microscope. The binder coating of the hemp particles
was investigated for all concretes at 6 months in order to inform on adhesion at All the lime hemp concretes investigated were inoculated repeatedly with a
the interface, a vital area in relation to strength and durability. In addition, as the high concentration of a mixed culture of microorganisms at 2 years. The microor-
rate of hydration and pozzolanic reaction affect setting and strength development, ganisms in the inoculum were typical organisms found in soil and in the air includ-
the formation of hydrates in lime:pozzolan pastes was studied at 1, 3, 7, 14 and ing the fungi Aspergillus and Penicillium and the bacterium Bacillus. The concretes
28 days. were stored in a humidity chamber at 30 degrees centigrade and 80% relative
humidity for 7 months. Intermittently the concretes dried out and were
remoistened.

2.5. Compressive strength

As there is no standard procedure for the measurement of unconfined compres- 3. Results: strength and durability of hemp-lime concretes
sive strength of hemp-lime concrete, the test was guided by EN 459-2 [31] and EN
196-1 [32] with a loading rate of 50 N/s. Typically, the concrete does not break but 3.1. Microstructure
continuously deforms therefore, the ultimate strength was set as the stress at which
the stress/strain curve departs from linear (Fig. 6). Testing time was typically be-
tween 1 and 2 min. Glouannec et al. (2011) observed that geometry of the samples
In all concretes, SEM analysis evidenced good adhesion at the
clearly influences the specimen’s behaviour however cylinders and cubes exhibit interface showing hemp particles well coated with scalehohedral
similar ultimate compressive strength [33]. calcium carbonate (carbonated lime) and/or hydrates.

Table 2
Composition of the hemp concretes.

Name Notation Binder composition (% by weight) Binder:hemp:water ratio (by weight)

B H W
Builder’s mix BM 70% calcium Lime, 20% NHL3, 10% PC 2 1 3.1
Commercial mix CM 100% commercial binder 2 1 2.9
GGBS G 70% calcium lime, 30% GGBS 2 1 3.1
GGBS + WR G + WR 70% calcium lime, 30% GGBS, 0.5% methyl cellulose 2 1 3.1
Metakaolin M 80% calcium lime, 20% Metakaolin 2 1 3.3
Metakaolin + WR M + WR 80% calcium lime, 20% metakaolin, 0.5% methyl cellulose 2 1 3.1
 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348 343

Fig. 1. Hemp interface in a commercial binder concrete at 6 months including Fig. 3. Hemp interface in a metakaolin binder concrete at 6 months which appears
abundant hydrates. mainly carbonated.

There are significant microstructural differences at the hemp

interface in the different concretes at 6 months: the commercial
binder includes abundant needle-shaped hydrates (Fig. 1) while
the builder’s mix shows a smaller amount of hydrates (Fig. 2)
and the lime:pozzolan binders are largely carbonated with a very
small amount of pozzolanic hydrates (Figs. 3 and 4). This indicates
that the commercial binder has a significant hydraulic content. The
methyl cellulose water retainer slightly increased the amount of
hydrates in lime:pozzolan concretes; is discussed in other research
[8] and suggests that retaining water by the water retainer has en-
hanced pozzolanic hydration.
It was noted that, in lime:pozzolan pastes, hydrates appear
early (at 24 h of curing) and their morphology changed over time,
from predominantly needle-shape at early ages (24 h) to sponge
and gel types at later ages [35]. As aforementioned, the hemp inter-
face in lime:pozzolan concretes is largely carbonated with little or
no hydrates, while hydrates are found in the bulk paste. This was
attributed to both the high water suction of the hemp and the
water soluble hemp constituents inhibiting hydration of the Fig. 4. Hemp interface in a metakaolin binder with methyl cellulose concrete at
6 months featuring strong carbonation and some needle-shaped hydrates.
lime:pozzolan binder at the interface [35].
A wide distribution of pore sizes was evident in all binders. At a
30,000 magnification, most pores ranged from c. 200 to 2 lm. In the commercial binder, needle-like hydrates were evidenced grow-
ing into pores thus reducing pore size.

3.2. Mechanical behaviour of hemp-lime concrete under compression

The mechanical behaviour of the hemp concrete is similar to

that of timber [36,37] and other cellular solids [38]. The concrete
continuously compresses on load application (Fig. 5), and the
deformation is similar to that shown by [17] whereby the stress–
strain curve can be divided into three regions: linear, plateau and
densification (Fig. 6). Initially, the curve is linear and the concrete
behaviour is quasi-elastic [14]. Later, a cracking noise and binder
powder appear indicating that the binder is failing. The behaviour
then departs from linear and deformation significantly increases
for a small stress increase (plateau region). During this phase, the
hemp cells begin to collapse and the intensity of the cracking noise
grows indicating that binder failure is becoming greater. Following
the plateau region, there is a strain densification where the stress
increases rapidly in relation to the strain. This is attributed to the
collapse of the hemp particles [17] which increases the stiffness
Fig. 2. Partially carbonated, hemp interface in a builder’s binder concrete at of the concrete as the contact between cell walls provides addi-
6 months including some hydrates. tional mechanical strength. The cube specimens deform in a differ-
344 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348

Fig. 7. Compressive strength of hemp-lime concretes over time. 80% of COVs < 20%.
BM – builder’s mix; CM – commercial mix; G – GGBS mix; G + WR – GGBS binder
with water retainer; M – metakaolin mix; M + WR – metakaolin with water
retainer.

Fig. 5. Typical failure of hemp-lime concrete under an axial compressive load. lower than that of the other concretes. As expected, the commer-
cial mix gains strength quickly due to the early formation of hy-
ent manner to cylinders which typically show a decrease in stress drates. However, its strength development after 28 days is slow.
rather than strain densification in the final stage. In this research, Similarly, Nguyen (2010) and Hirst et al. (2010) found that strength
the point at which the mechanical behaviour departs from a linear in commercial (hydraulic) binders did not increase significantly
stress/strain curve is considered as the ultimate strength, similarly after 28 days [13,16]. In this research, commercial concretes which
to [17]. were immersed in water after curing nearly doubled their strength
(reaching 0.63 MPa at 1 year), indicating that a low mixing water
3.3. Compressive strength content/dry curing conditions inhibited binder hydration halting
strength development [30].
Compressive strength varied between 0.02 and 0.04 MPa at The results suggest that the strength of hemp-lime concrete is
5 days and 0.29 and 0.39 MPa at 1 year (Fig. 7) falling within the not the sole function of binder hydration. Factors such as carbon-
range observed by the aforementioned authors. Most strength ation contribute towards strength at later ages [10]. The results
develops between 5 and 28 days mainly due to drying and agree with Ngoyen (2010) whereby natural hydraulic lime binders
hydration. (NHL2/3.5) were found to reach higher compressive strengths than
The metakaolin concrete shows the highest strengths at 3 and commercial binders of greater hydraulicity at 90 days [16].
6 months. A slight strength reduction is observed between
3 months and 1 year (metakaolin concrete) and 6 months and 3.4. Mechanical behaviour and flexural strength of hemp-lime concrete
1 year (metakaolin with water retainer). Similarly, strength loss in bending
has been reported in lime:metakaolin pastes [39,40], lime:metaka-
olin hemp concretes [41] and natural hydraulic lime mortars [42] As the flexural load is applied, the concrete bar bends as shown
being attributed to changes in the morphology of the hydrates. in Fig. 8. Following the maximum flexural strength, a crack be-
The GGBS concrete achieves the greatest compressive strength tween the underside of the sample and prism’s main axis develops.
(not significant, p < 0.1) at 1 year despite showing a slower rate The flexural strength of the concretes at 3 months and 1 year
of strength gain at earlier ages. The strength drop at 6 months is can be found in Table 3. The results are within the range previously
likely an error as the overall trend is increasing strength between reported by the aforementioned authors. At 3 months, the flexural
3 months and 1 year. During the first 6 months, the water retainers strength of the builder’s, commercial and lime:pozzolan with
increase the compressive strength of the pozzolan concretes. This water retainer concretes were broadly similar ranging between
is likely due to extra hydration resulting from the water retainer 0.11 and 0.13 MPa. The lime:pozzolan concrete yielded lower val-
holding a greater amount of water in the binder [8]. However, at ues ranging between 0.06 and 0.09 MPa.
1 year, the strength enhancement by the water retainer is no long- Between 3 months and 1 year, the flexural strength increase is
er evident. larger in the commercial binder concrete. Similarly to compressive
Despite its cement content, the builder’s mix is the slowest to strength, water retainers increase the flexural strength of hemp
develop strength and, at 1 year, its compressive strength is slightly concretes with lime:pozzolan binders. The flexural strength in-

Fig. 6. Two representative diagrams of the typical behaviour of hemp-lime concrete under a compressive axial load. The left graph has a more distinct plateau region (large
deformation for small increase in stress) than the right graph.
 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348 345

(including exposure to freeze:thaw conditions) as a hydraulic lime

mortar [45]. It is therefore evident that carbonation should con-
tribute towards durability of the concrete. Furthermore, a reduc-
tion in the porosity of the binder due to the increase in size of
CaCO3 compared to Ca(OH)2 [46] should reduce the quantity of
water absorbed and thereby minimise vulnerability to freeze–thaw
action.
As it can be seen from the results, freezing conditions have a
stronger impact on lime:pozzolan concretes, with a greater reduc-
tion in compressive strength and a higher weight loss for both
pozzolans, in particular metakaolin (Tables 4 and 5). In relation
to mass loss, the metakaolin and GGBS concretes show the highest
weight loss while the builder concrete shows a negligible value
and the commercial concrete increases weight due to additional
hydration (Table 4). The water retainer reduced weight loss.
The commercial concrete substantially increases strength
(60.56%) during freeze–thaw cycling relative to the 1 year old con-
crete (Table 4). This is due to additional hydration, taking place
during the immersion episodes of cycling, increasing the amount
Fig. 8. Typical deflection of hemp-lime concrete in bending.
of hydrates [30]. Slight compressive strength reductions are ob-
served in the freeze:thaw samples when compared to the reference
samples, suggesting that freeze:thaw action has a small effect on
Table 3
the commercial binder.
Flexural Strength of concrete at 3 months and 1 year.
The GGBS and metakaolin concretes show significant strength
Name Flexural strength @ COV Flexural strength COV losses after freeze:thaw action. The addition of water retainer im-
3 months @ 1 year
proves resistance to freeze:thaw action and the reduced strength
Builder’s mix 0.12 12.46 0.13 40.70 loss is likely due to enhanced hydration (Table 4). Water retainers
Commercial mix 0.11 10.69 0.19 17.56
increase the quantity of small pores which are believed to under-
GGBS 0.09 35.36 0.14 12.86
GGBS + WR 0.13 8.06 0.20 14.24 mine freeze:thaw resistance therefore, the improvement observed
Metakaolin 0.06 18.25 0.10 33.06 is probably due to the water retainer lowering water absorption
Metakaolin + WR 0.11 3.44 0.14 24.05 and enhancing hydration.
According to the results, the commercial binder concrete and, to
a lesser extent, the builder’s mix are more freeze:thaw resistant
than the lime:pozzolan concretes. This is likely due to the higher
crease between 3 months (0.06–0.12Mpa) and 1 year (0.1– mechanical strength of their hydraulic binders and their lower
0.2 MPa) is proportionally greater than the compressive strength water absorption (they contain smaller, hydrate-filled pores)
increase. Similar trends have been found in lime mortars [43] resulting in less water available to induce freeze:thaw damage.
and lime:cement mortars [44].

3.6. Durability: resistance to salt exposure

3.5. Durability: resistance to freeze–thaw

Sodium chloride exposure during one month does not appear to

No cracks appeared in any of the concretes following freeze–
damage the compressive strength of any concrete (Table 5). The
thaw action however, they were soft prior to drying and the
metakaolin with water retainer concrete (showing a strength de-
lime:pozzolan concretes were friable after drying (in particular
crease) is an exception however, this is likely an inconsistency as
the metakaolin ones). SEM analyses did not reveal any visual
all the other pozzolan concretes are consistently undamaged.
changes in the microstructure of the binder resulting from freez-
Cracks were not visible following salt exposure, although the con-
ing. Strength and mass were measured following freeze:thaw dam-
cretes became soft when wet due to water immersion and hard-
age, and the values compared to those of 1 year old and reference
ened again when dry. Similarly to freeze–thaw, the commercial
concretes (Table 4). Carbonation of lime should confer similar
concrete shows a significant increase in compressive strength fol-
properties to the concrete as in lime mortar; Waldum (1999),
lowing salt exposure and this is attributed to additional hydration
noted that a lime mortar achieved equivalent weather resistance
due to the presence of water.
All concretes show an increase in weight due to the crystalliza-
Table 4 tion of salt within their pores. The salt content of the lime:pozzolan
Weight and compressive strength after freeze:thaw action. concrete is the lowest (c. 6% weight increase for the GGBS and
Sample % weight % compressive strength loss metakaolin concretes). In contrast, the concretes with smaller
loss pores (commercial, builder’s and lime:pozzolan with water retai-
Relative to 1 year Relative to reference
concrete samples ner) show higher salt contents (c. 8–9% weight increase for the
builder and pozzolan with water retainer concretes). The commer-
BM 0.23 10.94b 14.30b
CM +7.2 60.56a 9.31 cial binder exhibits the highest weight gain of 16% which is due to
G 2.6 34.90b 31.87%b both the presence of salt and additional hydration. The weight gain
G + WR 1.6 24.84b 4.70 results suggest that small pores facilitate salt crystallization. This
M 4 59b 60b agrees with [47] who found that chlorides preferentially crystal-
M + WR 2 19.02 1.10
lized in small pores ranging between 0.1 and 1 lm.
a
P < 0.1. SEM/EDX analyses revealed that the salt mainly crystallises in
b
P < 0.3. layers, and that the salt does not penetrate into pores under the
346 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348

Table 5
Compressive strength loss of sample due to freeze/thaw and salt exposure cycling.

Sample Concrete @ 1 year (Mpa) Soaked concrete (Mpa) Freeze/thaw concrete (Mpa) Salt exposure concrete (Mpa)
COV COV COV COV
BM 0.37 10.1 0.39 6 0.33 16.9 0.43 6.8
CM 0.36 9.3 0.63 20.4 0.57 13 0.81 13.8
G 0.41 15.9 0.39 15.9 0.26 26.8 0.37 32.0
G + WR 0.39 17.7 0.28 20.9 0.30 17.6 0.37 6.0
M 0.34 6.3 0.35 21.4 0.29 29.5 0.40 8.2
M + WR 0.32 2.5 0.38 16.2 0.38 3.3 0.28 45.5

layers (Fig. 9). This agrees with mortar and porous stone results by The lack of salt crystallization damage in the concrete can be
Lubelli and coworkers [25,27]. due to the short-term salt exposure. However, the high ductility
The salt layers of NaCl, varying in extent and approximately of the concrete, accommodating the stresses imposed by the salt,
2 lm thick were identified with EDS (Fig. 10). Their morphology is probably responsible for the lack of damage by salt crystalliza-
ranges from smooth to an agglomeration of regularly shaped cubic tion. In addition, the hemp concrete pores may have a weaker
crystals (Fig. 11). Less frequently, single cubic or needle shaped affinity to salt than stone and mortar pores due to the organic
crystals are also observed (Fig. 12). No difference in salt crystalliza- nature of the aggregate, and this may lower the impact of salt
tion morphology was noted for the different binders. action.

Fig. 9. Left image: typical salt layer on the binder surface. Right image: side view of salt layer crack showing that the salt does not penetrate into the binder.

Fig. 10. Salt layer and its composition (EDS) (large peak resulting from gold coating).

Fig. 11. SEM micrographs of salt layers with varying morphology. Left image: smooth salt layer. Right image: agglomeration of cubic NaCl crystals forming a layer.
 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348 347

Fig. 12. Isolated NaCl crystals on the binder surface.

3.7. Durability: resistance to biological deterioration The freeze:thaw resistance of hemp concretes is a function of
the hydraulicity of the binder: the binders of greatest hydraulicity
As aforementioned, hemp concrete is widely regarded as being (commercial and builder) show superior freeze:thaw resistance
resistant to mould and insects on account of the alkalinity of the despite containing smaller pores. This suggests that the superior
lime [21]. However, there are many microorganisms that grow mechanical strength of the hydrates supersedes the pore structure
well in alkaline conditions. Following inoculation and despite re- characteristics of the binder with regard to freeze:thaw resistance.
peated efforts to keep them moist, the microorganisms frequently The concretes with the most hydraulic binders may also benefit
dried out and died during the seven months of testing. Within two from reduced water absorption resulting in less water available
months of the first inoculation, the microorganisms had died. None to induce freeze:thaw damage. Water retainers were found to im-
of the concretes showed any sign of deterioration despite a heavy prove freeze:thaw resistance of the lime:pozzolan binder.
inoculation of microorganisms and repeated inoculations after Despite the high concrete porosity (which ensured near satura-
populations died off. This is due to insufficient available nutrients tion conditions during testing) and the salt growth determined by
to support growth and/or unsuitable environmental conditions. SEM, the concretes did not suffer significant deterioration in a salt
These initial results suggest that hemp concrete is resistant to bio- environment following 1 month exposure. The lack of salt damage
deterioration when initial high humidity alternates with periods of of hemp-lime concrete is partially attributed to the high ductility
drying, conditions which relate well to those on site in Europe. Re- of the pore walls accommodating expansive salt crystallization
search is ongoing into microbial attack in extreme conditions. pressures. Salt crystallization was greater in the binders with smal-
ler pore sizes.
The resistance to repeated heavy microbial innoculations indi-
cates that hemp concrete is resistant to biodeterioration in envi-
4. Conclusion ronmental conditions close to those on site.
The improvements in strength development and resistance to
This paper investigates the effect of binder type on mechanical freeze:thaw action of the binders with water retainers highlight
strength and durability (resistance to freeze–thaw, salt exposure the potential of additives to improve strength and durability of
and biodeterioration). It compares hemp-lime concretes made hemp concrete with lime:pozzolan binder.
with hydrated lime and pozzolans (GGBS and metakaolin) to those
including hydraulic lime and cement (commercial or site binders
containing PC). Acknowledgements
There are significant microstructural differences at the hemp
interface in the different concretes at 6 months; the commercial The authors wish to thank the Environmental Protection
binder displays abundant needle-shaped hydrates while the Agency for funding this research; and the Traditional Lime Com-
builder’s mix shows a smaller amount of hydrates and the pany, Clogrennane Lime and Ecocem for the provision of materials.
lime:pozzolan binders are largely carbonated with a small amount We are grateful to Dr. Heath Bagshaw (Centre for Microscopy and
of pozzolanic hydrates. There is a small increase in hydration prod- Analysis) and Dr. Robbie Goodhue (Geology Department) for their
ucts evident in the lime:pozzolan binders with water retainer com- help with the SEM and XRD respectively.
pared to those without.
The compressive strength results indicate that at early ages, the References
commercial binder with greatest binder hydraulicity displays the
highest strength. However, as the concrete ages, other factors such [1] Nässén J, Holmberg J, Wadeskog A, Nyman M. Direct and indirect energy use
as carbonation contribute towards strength and the lime:pozzolan and carbon emissions in the production phase of buildings: an input–output
analysis. Energy 2007;32(9):1593–602.
concretes achieve similar strength to those including hydraulic [2] Boutin M-P, Flamin C, Quinton S, Gosse G, Inra L. Étude des caracteristiques
lime and cement at 1 year. Water retainers improved early environnementales du chanvre par l’analyse de son cycle de vie [Study of
strength of lime:pozzolan hemp concrete. This is attributed to an environmental characteristics of hemp by means of an analysis of its lifecycle].
Ministère de l’Agriculture et de la Pêche [Ministry of Agriculture and Fishery],
increased amount of mixing water retained in the binder resulting
2005.
in enhanced hydration. [3] Walker R, Pavia S. Behaviour and properties of lime–pozzolan pastes. In:
In relation to flexural strength, the concrete with lime:pozzolan Proceedings of the 8th international masonry conference. Dresden, Germany;
binders achieved lower strengths than the more hydraulic com- 2010: 353–362.
[4] Walker R, Pavia S. Physical properties of pozzolans for lime and cement
mercial binder at 3 months and 1 year; and water retainers im- composites. In: Proceedings of the BCRI bridge infrastructure concrete research
proved the flexural strength of the lime:pozzolan concrete. Ireland, Univ. of Cork, Ireland, 2010 457–68.
348 R. Walker et al. / Construction and Building Materials 61 (2014) 340–348

[5] Moranville-Regourd M. Cements made from blastfurnace slag. Lea’s chemistry [25] Lubelli B, van Hees RPJ, Huinink HP, Groot CJWP. Irreversible dilation of NaCl
of cement and concrete. London: Elsevier Ltd.; 1988. contaminated lime–cement mortar due to crystallization cycles. Cem Concr
[6] Prusinski JR, Marceau ML., VanGeem MG. Life cycle inventory of slag cement Res 2006;36(4):678–87.
concrete. In: The 8th CANMET/ACI international conference on fly ash, silica [26] Lubelli B, Nijland TG, van Hees RPJ, Hacquebord A. Effect of mixed in
fume, slag and natural pozzolans in concrete host committee supplementary crystallization inhibitor on resistance of lime–cement mortar against NaCl
papers. American Concrete Institute, 2004. crystallization. Constr Build Mater 2010;24(12):2466–72.
[7] Sabir BB, Wild S, Bai J. Metakaolin and calcined clays as pozzolans for concrete: [27] Benavente D, Garcı MA, Garci J, Sánchez-Morald S, Ordóñez S. Role of pore
a review. Cement Concr Compos 2001;23(6):441–54. structure in salt crystallisation in unsaturated porous stone. J Cryst Growth
[8] Walker R, Pavia S. Impact of water retainers in the strength, drying and setting 2004;260(3–4):532–44.
of lime hemp concrete. In: Proceedings of the BCRI bridge infrastructure [28] EN 459-1:2010 Building lime. Definitions, specifications and conformity
concrete research ireland, Trinity College and DIT, Dublin, Ireland, 2012. p. criteria.
355–60. [29] EN 197-1:2011 Cement. Composition, specifications and conformity criteria
[9] Evrard A. Bétons de chanvre – synthèse des propriétés physiques Saint for common cements.
Valérien. Construire en Chanvre, 2003. [30] Walker R, Pavia S. Hydration of lime hemp concretes made with commercial
[10] Cerezo V. Propriétés mécaniques, thermiques et acoustiques d’un matériau à binder. Ireland: Civil Engineering Research in Ireland. Queen’s University
base de particules végétales: approche éxpérimentale et modélisation Belfast; 2014.
théorique Saint Valérien. L’Institut National des Sciences Appliquées de [31] EN 459-2:2010 Building lime. Test methods.
Lyon; 2005. [32] EN 196-1-2005 Methods of testing cement. Determination of strength.
[11] de Bruijn P. Hemp concretes - mechanical properties using both shives and [33] Glouannec P, Collet F, Lanos P, Mounanga C, Pierre T, Poullain P. Propriétés
fibres. Licentiate Thesis. Sweden: Swedish University of Agricultural Sciences; physiques de bétons de chanvre. Mater Tech 2011;99:657–65.
2008. [34] EN 15304:2010 Determination of the freeze–thaw resistance of autoclaved
[12] Murphy F, Pavía S, Walker R. An assessment of some physical properties of aerated concrete.
hemp-lime concrete. In: Proceedings of the BCRI bridge infrastructure concrete [35] Walker R, Pavia S. Effect of Hemp’s soluble components on the physical
research ireland, Univ. of Cork, Ireland, 2010. p. 431–38. properties of hemp concrete. J Mater Sci Res 2014;3(2) [in press].
[13] Hirst E, Walker P, Paine K, Yates T. Characterisation of low density hemp-lime [36] Reid SR, Peng C. Dynamic uniaxial crushing of wood. Int J Impact Eng
composite building materials under compression loading. In: The 2nd 1997;19(5–6):531–70.
international conference on sustainable construction materials and [37] Gibson L, Ashby M. Cellular solids: structures and properties. Oxford: Pergamon
technologies, Italy, 2010. Press; 1988.
[14] Gourlay E, Arnaud L. Comportement hygrothermique des murs de béton de [38] Daxner T. Finite element modelling of foams. Udine, Italy: CISM International
chanvre. In: Proceedings of the Actes du congrès SFT, France, Le Touquet, Centre for Mechanical Sciences; 2010.
2010. [39] Cizer O, Van Balen K, Van Gemert D, Elsen J. Competition between carbonation
[15] Bütschi PY. Utilisation du chanvre pour la préfabrication d’éléments de and hydration on the hardening of calcium hydroxide and calcium silicate
construction. PhD thesis, Université de Moncton, Canada, 2003. binders [extract from PhD thesis]. Katholieke Universiteit Leuven, Belgium;
[16] Nguyen T. Contribution à l’étude de la formulation et du procédé de fabrication 2009.
d’éléments de construction en béton de chanvre, PhD thesis. Université de [40] Donchev I, Nivov J, Doykov I, Petrova N, Dimova L. On the formation of cement
Bretagne Sud, France, 2010. phases in the course of the interaction of kaolinite with portlandite. J Univ
[17] Elfordy S, Lucas F, Tancret F, Scudeller Y, Goudet L. Mechanical and thermal Chem Technol Metall 2010;45(4):391–6.
properties of lime and hemp concrete (‘‘hempcrete’’) manufactured by a [41] Eires R, Nunes JP, Fangueiro R, Jalali S, Camões A. New eco-friendly hybrid
projection process. Constr Build Mater 2008;22:2116–23. composite materials for civil construction. In: Proceedings of the European
[18] Bouloc P, Allegret S, Arnaud L. Le chanvre industriel: production et conference on composite materials, Biarritz, France, 2006.
utilisations. France: France Agricole; 2006. [42] Hanley R, Pavía S. A study of the workability of natural hydraulic lime mortars
[19] de Bruijn PB, Jeppssona KH, Sandin K, Nilssona C. Mechanical properties of and its influence on strength. Mater Struct 2008;41(7):373–81.
lime–hemp concrete containing shives and fibres. Biosyst Eng 2009;103(4): [43] Costigan A, Pavía S. Influence of the mechanical properties of lime mortar on
474–9. the strength of brick masonry. Historic mortars: characterisation, assessment
[20] Daly P, Ronchetti P, Woolley T. Hemp lime bio-composite as a construction and repair. RILEM book series, vol. 7. 2012. p. 349–59.
material. Ireland: Environmental Protection Agency; 2010. [44] Arandigoyen M, Alvarez JI. Pore structure and mechanical properties of
[21] Bevan R, Woolley T. Hemp lime construction – a guide to building with hemp cement–lime mortars. Cem Concr Res 2007;37(5):767–75.
lime composites. Bracknell, (UK): IHS BRE Press; 2008. [45] Waldum A. Durability of lime-based mortars in a severe climate - results from
[22] Botas S, Rato VM, Faria P. Testing the freeze/thaw cycles in lime mortar. In: field and artificial ageing tests. In: 12th international RILEM workshop on
Proceedings of the second historic mortar conference. Prague, Czech Republic; historic mortars: characteristics and tests, Scotland, 1999.
2010. p. 417–25. [46] Thomson M, Lindqvest JE, Elson J, Groot CJWP. Porosity of mortars.
[23] Chagneaud B, Coquillat A. Utilisation de matériaux renouvelables - Etude de characterisation of old mortars with respect to their repair. France: RILEM;
quatre bétons de chanvre. France: Centre Expérimental de Recherches et 2004.
d’Études du Bâtiment et des Travaux Publics; 2002. [47] Rodriguez-Navarro C, Doehne E. Salt weathering: influence of evaporation
[24] Charola AE. Salts in the deterioration of porous materials: an overview. J Am rate, supersaturation and crystallization patter. Earth Surf Processes Landform
Inst Conserv 2000;39(3):327–43. 1999;24:191–204.