Characterization and comparison of hygric properties of rape straw concrete and hemp concrete

Construction and Building Materials 102 (2016) 679–687

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Characterization and comparison of hygric properties of rape straw concrete and hemp concrete M. Rahim ⇑, O. Douzane, A.D. Tran Le, G. Promis, T. Langlet Laboratoire des Technologies Innovantes (LTI), University of Picardie Jules Verne, Avenue des Facultés – Le Bailly, 80 025 Amiens Cedex 1, France

h i g h l i g h t s
New bio-based material called ‘‘straw lime concrete” (SLC).
Experimental protocol validation (hemp lime concrete (HLC); the reference material).
Moisture properties characterization.
The both materials (SLC and HLC) exhibit an ‘‘excellent” moisture buffer capacity.

a r t i c l e

i n f o

Article history: Received 11 February 2015 Received in revised form 7 October 2015 Accepted 5 November 2015 Available online 13 November 2015 Keywords: Straw lime concrete Hemp lime concrete Hygric properties Moisture buffer capacity Moisture penetration depth

a b s t r a c t The use of plant particles as aggregates in construction materials provide a very good hygric performance, and offers a real contribution to sustainable buildings. This paper focuses on the hygric properties of a new bio-based material called ‘‘rape straw lime concrete” in comparison with a ‘‘hemp lime concrete” which is used to validate the experimental protocol of this work. The tests are conducted on the sorption isotherm, the water vapor permeability, the capillary water absorption and the moisture buffer capacity which is estimated from the moisture buffer value determined under equilibrium condition. The results showed that the materials have very interesting hygric properties and exhibit an excellent moisture buffer capacity. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction The building sector strongly affects the environment and generates waste and emissions of pollutants acting on the environment and indoor air quality. In this context, there has been an increase in the use of vegetable particles as building material aggregates. Among them, different natural building materials have been studied such as: straw-clay [1], flax shives [2,3], diss [4], cork [5], sugarcane [6,7], straw bales [8,9]. Currently, hemp lime concrete (HLC) has been extensively studied in many researches [10–13]. This material suggests a low overall environmental impact [14,15]. The hemp concrete enables to store approximately 14–35 kg of CO2 per square meter of wall built with a thickness of 25 cm over a period of 100 years [16]. Due to its insufficient mechanical performance [10,17], hemp concrete is mainly used to cover the walls of masonry or to fill the timber-framed structures in new and old buildings. Likewise, ⇑ Corresponding author. E-mail address: [email protected] (M. Rahim). http://dx.doi.org/10.1016/j.conbuildmat.2015.11.021 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.

it can be used to isolate the roofs or floors. Compared to the standard building materials, hemp concrete has a low conductivity (0.1 W/(mK)) which reduces heat diffusion and hence reduce energy consumption [17–19]. In addition, HLC has a good moisture buffer capacity [3,12,20,21], according to the classification proposed by the Nordtest project [22]. At the whole building level, numerical and experimental results showed that hemp concrete can reduce the indoor RH variation and leads to reduce energy consumption [23,24]. Regarding its thermal conductivity, this one depends on its density, its formulation and its water content. The thermal conductivity increases less than 15–20% in a wide range of relative humidity, and the water content has a lower effect on thermal conductivity than the density [18]. Referring to a study of Benfratello [25], the dry thermal conductivity increases from 0.09 to 0.14 W/(mK) – i.e. 55% – in a density range from 370 to 600 kg/m3 at 15 °C. From the hygric point of view, hemp concrete has a high moisture diffusion coefficient; its water vapor permeability is approximately 2.3  1011 kg/(Pams) and is nearly constant for a low or middle relative humidity [3,11]. In addition, the hemp lime

680

A Ac Cp dp,1% Dw dw du

e gv hc Le m MBV Ps

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687

surface (m2) water absorption coefficient (kg/(m2s1/2)) heat capacity (J/(kgK)) penetration depth (m) moisture diffusivity (m2/s) specific hygric capacity (kg/m3) thickness (m) density of moisture flow (kg/(m2s)) convective heat transfer (W/(m2K)) Lewis number (–) mass (kg) moisture buffer value (kg/(m2%RH)) saturation vapor pressure (Pa)

concrete has a high moisture buffer capacity of the order of 2 g/(m2%RH) [3,11]. At whole building level, numerical and experimental results showed that hemp concrete can decrease the daily indoor relative humidity variations and reduce energy consumption [23,24]. This article presents a new material called ‘‘rape straw lime concrete” by combining rape straw with a lime-based binder. Rape is a plant cultivated for the production of edible oil. Otherwise, its straw is considered as a waste product resulting from the harvest of the seed part. The choice of other resources such as rapeseed straw for aggregates is related to its wide availability. In France, the rapeseed cultivated area amounts to 1,555,000 ha, and the yield of straw is on average about 34.5 t/ha. It is the fourth most cultivated species in the country after wheat, maize and barley. Indeed, using straw as green material will help to promote new economic prospect for rapeseed straw. However, moisture properties are necessary to understand the hygrothermal behavior of materials under real climate conditions. Therefore, this article presents the results of moisture properties of two bio-based materials: hemp lime concrete and rape straw lime concrete. Hemp lime concrete is used as the reference material in order to validate the experimental protocol by comparing our results to the ones obtained by other authors. Also, the comparison of these two materials is intended to provide more information about the rape straw concrete which is a new bio-based material; that is the main subject of this study. These materials were tested for moisture properties (sorption isotherm, water vapor permeability and capillary water absorption) and moisture buffer values inspired by the Nordtest project [22], which represent their ability to dampen the indoor relative humidity variations thanks to their moisture buffering capacity [23]. 2. Materials 2.1. Manufacturing The hemp lime and rape straw lime concretes are comprised of water, limebased binder and plant particles such as hemp shives for hemp concrete (HLC) case and rape straw for rape straw concrete (SLC). Hemp shives are coming from the inner woody core of the plant, rape straw is the woody heart of stems. Hemp shives and rape straw have a relatively constant size distribution (5  5  15 mm for hemp shives and 0.5  3  15 mm for rape straw case). In this present work, the binder used for the formulation is Tradical PF70 which contains around 75% of rich lime (Ca(OH)2), 15% of hydraulic binder and 10% of puzzolon binders. The proportion of lime binder, hemp/rape straw and water by mass (%) is presented in Table 1. The concretes were manufactured by using a concrete-mixer. The aggregates and the binder were combined together, after water was added and the material was mixed for a few minutes until complete homogeneity. Every mould was manually filled and then packed down in order to avoid empty zones. Once filled, each mould has been placed in a room conditioned at 90% of relative humidity and 18 °C of temperature. Three days later, the concrete samples were removed from the mold, and in order to ensure full carbonation of concrete, they were stored in the same room for 9 months as recommended in [17].

Pv Rv tp V w bp u da dp k

water vapor pressure (Pa) gas constant for vapor (J/(kgK)) period (s) volume (m3) water content (kg/m3) mass transfer coefficient (kg/(Pam2kg)) relative humidity (–) air vapor permeability (kg/(Pams)) water vapor permeability (kg/(Pams)) thermal conductivity (W/(mK)) vapor resistance factor (–) air velocity (m/s) mass density (kg/m3)

l m q

2.2. Density and porosity The measurements of dry density, matrix density and porosity of concretes and aggregates were carried out; the results are presented in Table 2. The matrix density is obtained experimentally according to the pycnometer method which is to fill the pores with toluene. The total porosity is calculated from the matrix density and dry bulk density of material. Concerning the open porosity, it is determined according to the vacuum saturation method. The results show that both hemp shives and rape straw aggregates present an important porosity, which is of 90% and 89% respectively. Concerning the concretes, the total porosities are very close (76.4% and 75.1%, respectively for HLC and SLC). The slightly lower total porosity of SLC compared to HLC is related to its low matrix density (1954 kg/m3 compared to 2030 kg/m3 respectively for SLC and HLC). Nevertheless, the open porosity it is a bit higher for the case of SLC.

3. Theory and method 3.1. Sorption isotherm The sorption isotherm or the hygroscopic curve describes the equilibrium between the water content and relative humidity. These curves can be measured under quasi-equilibrium conditions according to continuous or discontinuous methods [26]. The samples used for the test are cylinders with 10 cm diameter and 3 cm thickness. The specimens were placed in desiccators containing salts solutions in order to keep relative humidity at desired levels, which have been chosen: 0%, 33%, 51%, 81% and 95%. To control the temperature, the desiccators were stored in a test room where the indoor temperature was maintained at 23 °C by an air conditioner. For each measuring point, three samples were used. Before testing, the samples were dried in an oven at 70 °C until the weight loss between two successive measurements, with a time interval of at least 24 h, remained less than 0.1%. Then, for each specimen, the water content was weighted using a balance with a 0.1 mg of accuracy. The specimen was considered to be in steady-state when the mass variation at three consecutive intervals of 24 h (or more) is less than 0.1% of the mass of the wet sample. The water content was calculated from following equation:

w¼

m  m0 V

ð1Þ

Several mathematical models have been suggested to represent the relationship between equilibrium moisture content and relative humidity (u) [27–30]. Among the existing models, BET and GAB models can be cited. An experimental study comparing two

Table 1 Proportion of lime binder, hemp/straw and water by mass for concretes. Material

Granular part

Lime

Water

HLC SLC

16 14

36 36

48 50

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687 Table 2 Density and porosity of studied concretes and aggregates. Dry density [kg/m3]

Matrix density [kg/m3]

Open porosity [%]

Total porosity [%]

478 ± 7 125 ± 9 487 ± 6 125 ± 5

2030 ± 30 1259 ± 21 1954 ± 36 1140 ± 32

49.9 ± 2 – 54.9 ± 2 –

76.4 ± 0.1 90.1 ± 0.5 75.1 ± 0.7 89.0 ± 0.5

The water vapor pressure is calculated from temperature and relative humidity measured during the test from the usual equation: 4042:9

HLC Hemp shives SLC Rape straw

Pv ¼ u  exp23:5771T37:58

models showed that, in the case of a porous medium, the BET model is quite accurate, but is limited for low relative humidity. Whilst, the GAB model has much wider application, it also allows to find the results of BET model [31]. Therefore, the GAB model was chosen and it is expressed as:

w¼

u

ð2Þ

a  u2 þ b  u þ c

3.2. Water vapor permeability test

ð3Þ

where e: thickness of sample (m), g v : density of moisture flow (kg/(m2s)) and DPv : water vapor pressure gradient (Pa). The vapor resistance factor (l) can also be used instead of the water vapor permeability:

l¼

da dp

where T is the temperature (K). However, the Eq. (3) does not take into account the water vapor resistances in the air between the salt solution and the inner face of the sample, and on both internal and external sides of the sample. Therefore, in order to take into account the resistance of the boundary layers Zp,in and Zp,out, and the air layer in the cup Zp,air, Peuhkuri [33] proposed to modify the Eq. (3) as follows:

dp ¼ DPv gv

Z p;air ¼

The water vapor permeability characterizes the material ability to transport water vapor under a vapor pressure gradient once the steady state is reached. In this study, it has been determined in accordance with standard cup-test [32]. For the both types of concrete, six samples, with 16 mm thickness and 9 cm in diameter, have been tested. Each sample was sealed over the cup (Fig. 1) containing desiccant (silica gel) that keeps the relative humidity of the atmosphere in the cup to 0–2% RH (dry state at the lower surface of specimen). The air layer between the specimen and the silica gel is about 13 mm. The sample and cup assembly was then placed into the device (brand Gravitest Gintronic CA10-006C). This device is equipped with a scale for measuring the mass of specimens with a resolution of 1 mg and a room temperature and humidity adjustable with an accuracy of ±0.5 °C and 3%RH. For this test, the temperature and relative humidity were maintained at 23 °C and 50%, respectively. At the steady state, the water vapor permeability (dp) can be determined, assuming that the water vapor diffusion process can be described by Fick’s law as follows:

e DP v

ð5Þ

e  ðZ p;in þ Z p;out þ Z p;air Þ

ð6Þ

The resistance of the air layer between the sample and salt solution is given by:

where a, b and c are GAB model constants.

dp ¼ g v 

681

ð4Þ

where da is the water vapor permeability of air (21010 kg/(msPa)).

ea da

where ea is the thickness of air layer in the cup. Zp,in and Zp,out ((Pam2s)/kg) are the resistance factors which can be estimated from the convective mass transfer coefficients (bp).

ZP ¼

1 bp

bp can be estimated using the Lewis relation, which is given in Eq. (7).

Rv  T  ðqC p Þair 1 ¼  Le3=4 bp hc

ð7Þ

where q is the air density (125 kg m3), Cp: heat capacity of air (1005 J/(kgK)), Rv: gas constant for vapor (461.5 (J/(kgK)), T: temperature (20 °C), hc: the convective heat transfer coefficient (W/(m2K)) and Le is the Lewis number. It is defined as the ratio between the thermal diffusivity to vapor diffusivity of air and the volumetric heat capacity of air. The exponent ¾ is recommended for interior surfaces in buildings [34]. According to Pedersen [35], the convective heat transfer (hc) can be estimated from the measures of air velocity (m) by the following relation:

hc ¼ 5:82 þ 3:96  m The hc,out is determined from the average velocity measured of air in the climate chamber which is equal to 2 m/s. Whereas the velocity taken into account for the estimation of hc,in is 0 m/s (the surface is in contact with air within the cup).

Fig. 1. Standard Cup-Method for determination of water vapor permeability.

682

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687

3.3. Moisture diffusivity Moisture diffusivity is widely used for investigating the moisture transport within building materials under a vapor pressure gradient. This transport property is calculated from the vapor permeability measured experimentally and specific hygric capacity, which is the slope of moisture retention curve. Moisture diffusivity is expressed as follows:

Dw ¼

dp  Ps dw du

ð8Þ

where dw is the specific hygric capacity (kg/m3), Ps is the saturation du vapor pressure (Pa) and dp is the water vapor permeability (kg/Pams)). 3.4. Capillary water absorption The standard parameter for describing the capillary suction characteristics of building materials in contact with water is the water absorption coefficient (Ac). This experimental test was conducted in accordance to standard DIN 52617/87 [36]. Three samples of each material were placed in contact with water at a depth of approximately 3 mm and maintained at a constant level by using a tank water system (Fig. 2). To restrict diffusion to a one-dimensional direction, the samples, previously dried (70 °C), were sealed hermetically using a plastic film. The test has been conducted in the room where the indoor temperature was maintained at 23 °C by an air conditioner. The water absorption coefficient was derived by measuring the specific mass of water absorbed per unit area as a function of the square root of time. The liquid transfer coefficient Ac is the slope of the mass uptake curve, when expressed with the square root of time, as described by Eq. (9):

Ac ¼

Dm pffiffi AD t

ð9Þ

where Ac (kg/(m2s1/2)) is the water absorption coefficient, Dm (kg) is the mass of water absorbed, A (m2) is the surface area of sample in contact with water and t (s) is the time interval absorption. 3.5. Moisture buffering capacity The indoor relative humidity is one of most important factors to consider for evaluating the hygrothermal comfort and building performance [37,38]. High indoor humidity can cause the degradation of a building structure and significantly shorten its life [39,40]. Many authors showed that the hygroscopic materials can

contribute to moderate the indoor relative humidity variations, and so improve comfort for occupants, and reduce energy consumption [23,24,22]. Up to now, the moisture properties of materials which are measured under steady state and equilibrium conditions of relative humidity and temperature have been presented. However, it is more realistic to study the hygric behavior of materials under dynamic conditions. The Nordtest project [22] proposed an experimental protocol to evaluate the moisture buffer capacity of hygroscopic materials via the definition of moisture buffer value (MBV). A similar method has been proposed in Japanese Industrial Standard (JIS) [41]. The both methods differ in time scheme, levels of imposed signal, thickness of specimen and the report data [42]. It is noticed that the MBV value takes into account the mass transfer resistances at the boundary, so this is not an intrinsic property of the material. The aim of this part is to determine the MBV values of studied materials and to experimentally investigate the impact of specimen thickness on the moisture buffer capacity. 3.5.1. Moisture penetration depth and ideal MBV The penetration depth is defined as the depth where the amplitude of moisture content variations does not exceed 1% of the surface amplitude. Further details can be found in [22]. The penetration depth (dp,1%) provides a referential value which can be used to evaluate the thickness of the sample for measuring the MBV value. The moisture penetration depth is calculated using the following:

dp;1% ¼ 4:61

ð10Þ

p

Nordtest defined also the ideal moisture buffer value (MBVideal) which is calculated based on the Fourier analysis when relative humidity variation at boundary of material is a sinusoidal form. The MBVideal value is derived from the material properties and is given by:

MBVideal  0:00568  Ps  bm 

pffiffiffiffi tp

ð11Þ

Moisture effusivity bm allows to describe the ability of a material to absorb or release moisture and can theoretically express the rate of moisture absorbed by a material when it is subjected to a sudden increase in surface humidity [22]. Moisture effusivity is expressed as follows expression:

bm ¼

sffiffiffiffiffiffiffiffiffiffiffiffiffi dp  @w @u ps

ð12Þ

3.5.2. MBV experimental protocol The practical MBV, as defined in Nordtest project, indicates the amount of water vapor uptake or release by material per open surface area, during a certain period of time when it is subjected to variations in relative humidity of the surrounding air. Experimentally, the sample is subjected to cyclic step-changes in relative humidity between 75%RH during 8 h and 33%RH during 16 h respectively at a constant temperature of 23 °C. The MBV value is calculated at steady-state by the following equation:

MBV ¼

Fig. 2. Schematic of experimental device of capillary water absorption test.

rffiffiffiffiffiffiffiffiffiffiffiffiffiffi Dw  tp

Dm A  ðHRhigh  HRlow Þ

ð13Þ

where Dm is the moisture uptake/release during the period, RHhigh/low the high/low relative humidity level and A is the exposed surface. The MBV value is determined in steady state, which is supposed reached when, between three consecutive cycles, the test satisfies the following two conditions [22]:

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687

683

- The change in mass Dm [g] is less than 5% between the last three cycles. Dm is the average between the weight gain during the moisture uptake and the weight loss during drying. - The difference between the weight gain and weight loss within each cycle should be less than 5% of Dm. The moisture buffer values of HLC (for validation of experimental setup) and SLC are experimentally measured according to the Nordtest protocol [22]. In order for this to be conducted, a climate chamber (CL2 Biaclimatic Type-25) was used (see Fig. 3). This device can control the temperature with an accuracy of ±0.3 °C, and the relative humidity with an accuracy of ±2%RH, in the range of 10–98%RH. The temperature is kept constant at 23 °C. The change of the relative humidity (75%, 33%) in the climate chamber, according to the 8/16 h scheme, is carried out manually. The regulation of temperature and relative humidity can be verified using the sensor of the climate chamber. Based upon the penetration depths calculated above, the thicknesses of samples (with the area of 100 cm2) are respectively 3, 5 and 7 cm, and for each thickness, three samples were prepared. The exposured areas of the surface materials are higher than the value recommended in the Nordtest protocol. After being sealed on the back and edges with an aluminum tape as shown in Fig. 3, these were placed in the climatic chamber. The gravimetrical determination of the moisture absorption/desorption was made outside the climate chamber using a balance with 0–750 g and 0.001 g resolution. To minimize the perturbation of temperature and relative humidity in the climate chamber, the specimens were weighed three times during the absorption period and two times during desorption period.

Fig. 4. Kinetics of adsorption and desorption of hemp lime concrete (HLC).

4. Results and discussion 4.1. Sorption isotherm Figs. 4 and 5 show the kinetics of sorption curves for HLC and SLC. To reduce the experimental time for the SLC case, the measurements for 95%RH were performed on samples stabilized at 81%. As it can be seen from these figures, the sorption process was very slow, particularly under high relative humidity levels. It required more than 400 and 200 days respectively for HLC and SLC concretes to reach equilibrium when the relative humidity of atmosphere is at 95%. Moreover, as can be seen in Fig. 4, unexpected behaviors of the adsorption kinetics were observed for HLC, at 90 days in the case of 75%RH and at 150 and 250 days for

Fig. 5. Kinetics of adsorption and desorption of straw lime concrete (SLC).

95%, which could be due to a disruption of relative humidity regulation in the desiccators. Concerning the kinetics curves of desorption for both materials, the water content decreases rapidly at the beginning of exposure and then slowly until the equilibrium state, which is reached after about 200 and 300 days for HLC and SLC,

Fig. 3. Experimental setup and specimens.

684

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687

has a finer and more homogeneous porosity compared to hemp shives. The pores are of different sizes with diameters respectively of 10–40 lm and 10–30 lm for the hemp shives and straw rape. Regarding the desorption isotherm curves of HLC and SLC, they are similar. Furthermore, the moisture content during adsorption is smaller than the one during desorption. The average of moisture content difference between adsorption and desorption curves at 33, 51, 75 and 81%RH are about 20 kg/m3 for both HLC and SLC. Common physical explanations for the existence of hysteresis are: capillary condensation, contact angle hysteresis and the inkbottle effect due to pores with small entry [44]. So, after drying out, the desorption curve did not return to the initial condition, the moisture content at 0–2%RH for HLC and SLC are 9.2 kg/m3 and 10 kg/m3, respectively. This residual moisture content was also observed in [11] and it may be due to the carbonation occurred during the sorption test. The sorption isotherms of HLC can be compared to the one obtained by Collet [45] and Evrard [12]. Compared to the present study, the curves found by these authors have a similar shape, however the moisture content is slightly lower.

Fig. 6. Sorption isotherms of HLC and SLC.

Table 3 GAB model constants.

HLC SLC

a

b

c

0.085 0.083

0.058 0.056

0.030 0.028

respectively. Note, however, that no mold was detected in the samples during the experiment, but care was taken to prevent due of the duration of exposure. Fig. 6 shows the sorption isotherms obtained from experimental adsorption and fitting curves with GAB model (GAB model constants are given in Table 3). The GAB model equation leads to satisfactory prediction of the adsorption equilibrium moisture content of both materials. The curves are providing good fits to the type II, according to the IUPAC classification of the adsorption isotherms [43]. This type of curve characterizes the mesoporous and macroporous materials. It can be observed that the adsorption isotherms of the two materials display a similar pattern, which is to be expected because they have a similar microstructure. As the relative humidity increases, the moisture content increases, due to adsorption of more water vapor. But when relative humidity exceeds 75%, moisture content of SLC increases more rapidly and higher than the one of HLC. At 95%RH, the moisture content of SLC is 160 kg/m3 compared to 95 kg/m3 for HLC. This difference may be associated with the microstructure of aggregates of lime shives and rape straw. Indeed, it is apparent from scanning Electron Microscopy (SEM) analysis of hemp and straw particles (see Fig. 7) that rape straw

4.2. Water vapor diffusion The resistance factors of the boundary layers and of the inside cup air, the water vapor permeability and the vapor resistance factor (with and without Zp resistance factors) for HLC and SLC are given in Table 4. The results show that neglect these resistances increases vapor resistance factor; for example, for the SLC case, its value decreases from 9.36 to 7.87 when resistance factors are considered (thus a difference of 19%). Indeed, SLC is a bit more permeable to water vapor (with l value of 7.87) compared to HLC (with l value of 8.98), which can be explained by the higher open porosity of SLC. Concerning the water vapor permeability of HLC, the value reported in the present study (2.23 1011 kg/(Pams)) is similar to the one obtained by Collet and al [11]. The moisture diffusivity as a function of water content, calculated from Eq. (8) where specific hygric capacity was derived from GAB model curve, is presented in Fig. 8. It can be seen that curve shapes are in accordance with the one proposed by De Vries [46] theory and for HLC case the shape is similar to the one obtained by Collet [45]. When moisture content is low (w < 6.84 kg/m3 for HLC and w < 6.07 kg/m3 for SLC), the moisture transfer in vapor phase predominant, therefore, the moisture diffusivity increases up to its maximum value. Following this, its value decreases due to capillary condensation; the moisture transport in liquid phase is more significant compared to vapor phase. Indeed, the results show that the moisture diffusivities of both materials are quite

Fig. 7. (A) SEM Image of hemp particle. (B) SEM Image of straw rape particle.

685

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687 Table 4 Water vapor permeability and vapor resistance factor of HLC and SLC. HLC Resistance factor

Without Zp factors

7

2

Zp,out [10 (Pam s)/kg] Zp,in [107 (Pam2s)/kg] Zp,air [107 (Pam2s)/kg]

2.18 2.47 6.50

dp [1011 kg/(Pams)]

1.89 ± 0.09 10.58 ± 0.46

2.14 ± 0.14 9.36 ± 0.64

2.23 ± 0.012 8.98 ± 0.47

2.56 ± 0.23 7.87 ± 0.74

l With Zp factors

SLC

dp [1011 kg/(Pams)]

l

Fig. 9. Water uptake of HLC and SLS as a function of square root of time.

Table 5 MBVideal, dp,1%, Dw, bm of HLC and SLC. Concrete

Dw [m2/s]

bm [g/(m2Pas1/2]

MBVideal [g/(m2%RH)]

dp,1% [cm]

HLC SLC

1.54109 1.69109

5.67107 6.23107

2.66 2.92

3.01 3.14

Fig. 8. Moisture diffusivity as a function of water content for HLC and SLC.

similar. This can be explain by the similar slopes of the sorption isotherms calculated according to GAB model. 4.3. Capillary water absorption coefficient The mass of water uptake per unit area versus the square root of time for HLC and SLC is displayed in Fig. 9. The mass curves appear linear for the first two hours of measurement. The water absorption coefficient of HLC and SLC concretes calculated from the slope of curves are 0.062 and 0.267 kg/(m2s1/2), respectively. The capillary water absorption of SLC is four times more than the one of HLC concrete. This leads to the assumption that the open pores (accessible to water) in case of SLC concrete, is more important and the size of porosity is more adapted to the capillary diffusion of water. 4.4. Moisture buffering capacity of HLC and SLC 4.4.1. Moisture buffer value and classification of materials The ideal MBV value and penetration depth, calculated from the experimental results, are presented in Table 5. It can be seen that the ideal MBV of SLC is 2.92 g/(m2%RH) and higher than the one of HLC which is equal to 2.66 g/(m2%RH), because SLC has a greater moisture effusivity. Since moisture diffuses more rapidly in SLC than in HLC (moisture diffusivity of SLC is 1.69 109 m2/s compared to 1.54 109 m2/s for HLC), the penetration depth is greater for straw lime concrete case. The values calculated and reported in Table 5 are giving only a referential thickness of specimen because the boundary conditions are different in the Nordtest protocol; this will be further discussed in the next section. Fig. 10 shows the moisture uptake and release for the SLC sample at steady state, RH and T in the climate chamber. The amounts of moisture uptake and release are very close. For the HLC case, the MBV obtained is equal to 2.02 g/(m2%RH), which is very close to the values in the range of 1.75–2.15 g/(m2%RH)) given by [20,23,47]. These results validate the experimentation done in this

Fig. 10. Moisture uptake and release for the SLC sample at steady state.

work. Compared to HLC, the MBV value of the SLC is higher (MBV is equal to 2.59 g/(m2%RH), so that the moisture buffering capacity of straw lime concrete is better. To compare with other common building materials, the classification proposed by Rode [22], which defines five ranges from negligible to excellent for practical moisture buffer values, has been used. Fig. 11 shows the classification of studied materials (obtained with samples thickness of 7 cm) in comparison with other materials tested in Nordtest project. The first noticed point is that the concrete and brick fall into the lowest category. Plaster presents a moderate moisture buffer capacity performance and cellular concrete is classified like a good one. The SLC shows the highest moisture buffering capacity and it is followed by the HLC. Both bio-based materials exhibit an ‘‘excellent” moisture buffer performance. 4.4.2. Effect of the specimen thickness According to the Nordtest protocol, the specimen thickness should be like expected in its use or at least equal to the moisture penetration depth (1% definition) for daily moisture variation. As discussed above, the penetration depth has been calculated

Excellent

Good

Lmited

Moderate

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687

Negligible

686

SLC HLC Cellular concrete Plaster Brick Concrete

0

0.5

1

1.5

2

2.5

3

MBV (g/(m².%RH)) Fig. 11. Moisture buffer value classes of studied materials in comparison with other materials tested in Nordtest project.

Table 6 MBV value of SLC and HLC for different thickness of the samples. Concrete

e [cm]

MBV [g/(m2%RH)]

MBVideal [g/(m2%RH)]

HLC

3 5 7

1.84 1.86 2.02

2.66

SLC

3 5 7

2.29 2.45 2.59

2.92

material called ‘‘rape straw lime concrete” and the second one is a hemp lime concrete, which is used to validate the experimental protocol of this work. The two materials are highly porous with interconnected porosity which is expressed by both a high absorption capacity and a high diffusion of water vapor. The results show that the sorption curves of the two materials display a similar pattern due to their similar microstructure. For a low relative humidity, moisture content is weak and increases when the relative humidity become higher due to the trigger of capillary condensation mechanisms. So, compared to SLC, the moisture content of HLC is lower. Moreover, the fitting curves obtained with GAB models give a good approximation. The water vapor permeability of the two materials leads to a high vapor diffusion capacity, even if HLC concrete is less permeable to water vapor than SLC concrete. Concerning the liquid transfer, both concretes show a high sensitivity to the water capillary absorption, especially for straw concrete (its water capillary absorption coefficient is four times more than the one of hemp concrete). Finally, hygric characterization under dynamic conditions has been done in accordance with the Nordtest protocol. It is shown that the MBV values for HLC and SLC concretes are respectively 2.02 and 2.59 g/(m2%RH), therefore the both materials exhibit an ‘‘excellent” moisture buffer capacity. In addition, the influence of the thickness of sample on the MBV value has been examined, and the results showed that the choice of the sample thickness based only on moisture penetration depth for determining MBV may lead to non representative results. Acknowledgements

assuming sinusoidal variation, while the Nordtest protocol uses exposures by stepwise change. The moisture penetration depth can be only considered as an approximation. Numerical studies showed that penetration depth like used in Nordtest project effects the MBV value [48]. Therefore, the effect of sample thickness on the MBV values was experimentally investigated and the MBV values as a function of specimen thickness are given in Table 6. For HLC case, the MBV value increases from 1.84 to 2.02 g/(m2%RH) (a variance of 9.7%) while thickness is varying from 3 cm to 7 cm, and from 2.29 to 2.59 g/(m2%RH) (a difference of 13.1%) for SLC case. So, the MBV value increases with the increasing thickness of sample, as confirmed by previous studies [48,42]. It can be concluded that the choice of the sample thickness based on moisture penetration depth for determining MBV may lead to non-representative results. The theoretical penetration depth is based on steady state data (constants of sorption curve and water vapor permeability). It is shown that the use of steady state data to characterize dynamic behavior hampers the reliability of calculation. In addition, only one value is taken into account for vapor permeability, when it has been shown that the water vapor permeability of hemp concrete increases with moisture [17]. Therefore, the calculated penetration depth is an approximation, and the variation of MBV with the thickness of the specimen must be due to the fact that the actual penetration depth is higher than the calculated one. Besides, the MBV value measured is significantly less than the MBVideal because ideal experimental conditions rarely exist, and the film resistance due to boundary air layer on the exchange surface of sample are neglected. 5. Conclusion This article focuses on the characterization of the moisture properties of two materials: the first one is a new bio-based

This study was done within the project ParATE1 (in partnership with CODEM Picardie and University of Reims), which is funded as part of the Picardie Region and the European Regional Development Fund (ERDF). The authors wish to thank them. References [1] D. Alejandro, Energy and carbon embodied in straw and clay wall blocks producedlocally in the Andean Patagonia, Energy Build. 70 (2014) 15–22. [2] M. Rahim, O. Douzane, B. laidoudi, G. Promis and T. Langlet, Caracterisations mecanique et thermique de bétons agro-sourcés, 2 eme édition du congrés ecobat sciences & techniques, Paris, France, 2013, pp. 179–186. [3] M. Rahim, O. Douzane, A. Tran Le, G. Promis, B. Laidoudi, A. Crigny, B. Dupre, T. Langlet, Characterization of flax lime and hemp lime concretes: hygric properties and moisture buffer capacity, Energy Build. 88 (2015) 91–99. [4] M. Sellami, M. Merzoud, S. Amziane, Improvement of mechanical properties of green concrete by treatment of the vegetals fibers, Constr. Build. Mater. 47 (2013) 1117–1124. [5] S. Silva, M. Sabino, E. Fernandas, V. Correlo, L. Boesel, R. Reis, Cork: properties, capabilities and applications, Int. Mater. Rev. 50 (6) (2005) 345–365. [6] Z. Kodah, M. Jarrah, N. Shanshal, Thermal characterization of foam-cane (Quseab) as an insulant material, Energy Convers. Manage. 40 (4) (1999) 349– 367. [7] K. Manohar, J. Ramroopsingh, D. Yarbrough, Use of sugarcane fiber as building insulation, Am. Soc. Test. Mater. 1426 (2002) 299–313. [8] T. Ashour, H. Georg, W. Wu, Performance of straw bale wall: a case of study, Energy Build. 43 (8) (2011) 1960–1967. [9] S. Goodhew, J. Carfrae, P. De Wilde, Briefing: challenges related to straw bale construction, Proc. Inst. Civil Eng.: Eng. Sustain. 163 (4) (2010) 185–189. [10] A. Laurent, G. Etienne, Experimental study of parameters influencing mechanical properties of hemp concretes, Constr. Build. Mater. 28 (2012) 50–56. [11] F. Collet, J. Chamoin, S. Pretot, C. Lanos, Comparison of the hygric behaviour of three hemp concretes, Energy Build. 62 (2013) 294–303. [12] A. Evrard, A. De Herde, Hygrothermal performance of lime hemp wall assemblies, J. Build. Phys. 34 (2010) 15–25. [13] M. Lawrence, E. Fodde, K. Paine, P. Walker, Hygrothermal performance of an experimental hemp-lime building, University de Bath. Key Eng. Mater. 517 (2012) 413–421. [14] S. Pretot, F. Collet, C. Garnier, Life cycle assessment of a hemp concrete wall: impact of thickness and coating, Build. Environ. 72 (2014) 223–231.

M. Rahim et al. / Construction and Building Materials 102 (2016) 679–687 [15] P. Daly, P. Ronchetti, T. Woolley, Hemp lime bio-composite as a building material in irish construction, STRIVE Report 2009-ET-DS-2-S2, The Environment Protection Agency, Irland, 2012. [16] M. Boutin, C. Flamin, S. Quinton and G. G, Analyse du cycle de vie-Compounds thermoplastique chargés fibres de chanvre et mur en béton de chanvre banché sur ossature bois, Rapport d’étude, INRA, Lille, 2005. [17] V. Cerezo, Propriétés mécanique, thermique et acoustique d’un matériau à base de particules végétales: Approche expérimentale et modélisation thèorique (Ph.D. thesis), MEGA-ENTPE.INSA Lyon, France, 2005. [18] F. Collet, S. Pretot, Thermal conductivity of hemp concretes: variation with formulation, density and water content, Constr. Build. Mater. 65 (2014) 612– 619. [19] T. Pierre, T. Colinart, P. Glouannec, Measurement of thermal properties of biosourced building materials, Int. J. Thermophys. (2013) 1–21. [20] F. Collet, S. Pretot, Experimental investigation of moisture buffering capacity of sprayed hemp concrete, Constr. Build. Mater. 36 (2012) 58–65. [21] E. Latif, M. Lawrence, A. Shea, P. Walker, Moisture buffer potential of experimental wall assemblies incorporating formulated hemp-lime, Build. Environ. 93 (2015) 199–209. [22] C. Rode, R. Peuhkuri, H. Lone, B. Time, A. Gustavsen, T. Ojanen, J. Ahonen, K. Svennberg, Moisture buffering of building materials, Nordic Innovation Centre Report: BYG-DTU R-126, ISSN, 2005, pp. 1601–2917. [23] A. Tran Le, C. Malouf, T. Mai, E. Wurtz, F. Collet, Transient hygrothermal behaviour of a hemp concrete building envelope, Energy Build. 42 (2010) 1797–1806. [24] A. Shea, M. Lawrence, P. Walker, Hygrothermal performance of an experimental hemp-lime building, Constr. Build. Mater. 36 (2012) 270–275. [25] S. Benfratello, C. Capitano, G. Peri, G. Rizzo, G. Scaccianoce, G. Sorrentino, Thermal and structural properties of a hemp-lime biocomposite, Constr. Build. Mater. 48 (2013) 745–754. [26] EN ISO 12572, Hygrothermal performance of building materials and productsdetermination of water vapour transmission properties, European Committee for Standardization, Brussels, Belgium, 2001. [27] S. Brunauer, P. Emmet, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (1938) 309–319. [28] E. Barett, L. Joyner, P. Halenda, The determination of pore volume and area distributions in porous substances. Computations from nitrogen isotherms, J. Am. Chem. Soc. 73 (1951) 373–380. [29] E. Guggenheim, Application of Statistical Mechanics, Clarendon Press, Oxford, 1966. Chapter 11. [30] R. Anderson, W. Hall, Modifications of the brunauer, emmett and teller, equation II, J. Am. Chem. Soc. 70 (5) (1948) 1727–1734.

687

[31] J. De Boer, The Dynamical Character of Adsorption, Clarendon Press, Oxford, 1953. [32] NF EN 12086, Thermal insulating products for building applications. Determination of the transmission properties of water vapor. [33] R. Peuhkuri, Moisture dynamics in building envelopes (Ph.D. thesis), Department of Civil Engineering, Technical University of Denmark, 2003. [34] P. Sandberg, Byggnadsdelars fuktbalans i naturligt klimat, Institutionen för byggnadsateknik, Tekniska högskolan i Lund. Rapport 43, 1973 [35] C. Pedersen, Combined Heat and Moisture Transfer in Building Constructions, Thèse de doctorat (Rapport n°214), Thermal Insulation Laboratory, Technical University of Denmark, 1990. [36] DIN52617/87, Determination of the water absorption coefficient of building materials, Deutsches Institut für Normung, Berlin, Germany, 1987. [37] T. Kalamees, J. Vinha, Hygrothermal calculations and laboratory tests on timber-framed wall structures, Build. Environ. 38 (2003) 689–697. [38] J. Lstiburek, Moisture, building enclosures, and mold, HPAC Heat. Pip. Air Condition. Eng. 74 (2002) 77–80. [39] J. Toftum, A. Jorgensen, P. Fanger, Upper limits of air humidity for preventing warm respiratory discomfort, Energy Build. 28 (1998) 15–23. [40] C. Roulet, Santé et qualité de l’environnement intérieur dans les bâtiments, Polytechnic Press and University of Romandes, Lausanne, 2004. [41] JIS A 1470-1, Test Method of Adsorption/Desorption Efficiency for Building Materials to Regulate an indoor Humidity – Part 1: Response Method of Humidity, 2002. [42] R. Staf, J. Hans, A comparison of the nordtest and japanese test methods for the moisture buffering performance of building materials, J. Build. Phys. 30 (2) (2006) 137–161. [43] International Union of Pure and Applied Chemistry, Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity, Int. Union Pure Appl. Chem 57 (4) (1986) 603–619. [44] H. Derluyn, D. Derome, J. Carmeliet, E. Stora, R. Barbarulo, Hysteretic moisture behavior of concrete: modelling and analysis, Cem. Concr. Res. 42 (2012) 1379–1388. [45] F. Collet, M. Bart, L. Serres, J. Miriel, Porous structure and water vapour sorption of hemp-based materials, Constr. Build. Mater. 22 (2008) 1271–1280. [46] D. De Vries, The theory of heat and mass transfert in porous media revisited, J. Heat Mass Transf. 30 (7) (1987) 1343–1350. [47] S. Dubois, A. Evrard, F. Lebeau, Modeling the hygrothermal behavior of biobased construction materials, J. Build. Phys. (2013). 1744259113489810. [48] O. Osanyintola, P. Talukdar, C. Simonson, Effect of initial conditions, boundary conditions and thickness on the moisture buffering capacity of spruce plywood, Energy Build. 38 (2006) 1270–1282.