Effect of Hemp ’ S Soluble Components on the Physical Properties of Hemp Concrete

Lime hemp concrete is a sustainable building material made with hemp shiv and lime. In this research, pozzolans are added to the lime binder. Hemp shiv contains water soluble constituents including pectic polysaccharides which can dissolve in the mixing water. Sugars and pectins are known to retard clinker hydration in portland cement. This research studies whether the hemp soluble components impact lime-pozzolan paste reactions thus affecting the properties of hemp concrete with lime-pozzolan binder. Physical tests and microstructural analyses were undertaken to determine how the water soluble hemp constituents influence the hydration reactions responsible for setting and hardening of lime-pozzolan pastes. The effect of hemp particles on a lime-pozzolan binder was also investigated using microstructural analysis. Physical testing evidenced that the hemp’s soluble constituents do not alter ultimate strength however, they delay pozzolanic reaction retarding setting and early strength development. The microstructural analysis of lime-pozzolan pastes agreed with the above, evidencing that the hemp delays the formation of hydrates, in particular calcium silica hydrates. The halt of the pozzolanic reaction by the hemp particles was clear at the hemp interface, where few hydrates were evident. This suggests that the properties of the hemp particles including their high suction (which removes water at the interface) and their high porosity (which assists CO2 diffusion) contribute to the halt of the pozzolanic reaction.


Introduction
Lime hemp concrete is a sustainable, carbon negative building material that can replace high embodied energy materials in certain applications thus making construction more sustainable.Lime hemp composites have been used in France since the 1990s and are gaining popularity in Europe.They are made with hemp shiv, which is the woody interior of the hemp stalk, and a lime-based binder.Commercial binders currently available and binders designed on site by building practitioners typically include hydraulic lime, portland cement (PC) and additions.The purpose of the hydraulic lime and PC is to improve early age properties such as setting and strength development.This paper is part of a research programme which aims at formulating a binder using calcium lime and pozzolans.It is intended that the pozzolans will eliminate the need for hydraulic lime and PC so that the material can have a lower environmental impact.
Pozzolans are materials with an amorphous siliceous or siliceous and aluminous content that react with portlandite (CH=Ca(OH) 2 ) in the presence of water to form hydrates (similar to cement hydration) thereby accelerating the slow hardening (carbonation) of calcium limes by imparting a hydraulic set.They have been used to enhance the properties of composites since antiquity.Previous research by the authors indicated two suitable pozzolans for use in lime hemp composites: Ground Granulated Blastfurnace Slag (GGBS) and metakaolin; on account of their fast setting and high reactivity (Walker & Pavía, 2010;Walker & Pavía, 2011).In this paper, Rice Husk Ash (RHA) was also included in several tests to study the effect of the pectins on a fully siliceous pozzolan.

Effect of Organic Compounds on Material Properties
The amount and type of hydrates produced upon binder hydration, and the time at which they appear, impact the physical properties of cement and lime composites, including setting and early strength development.Sugars and  (Walker & Pavía, 2010;Walker & Pavía, 2011), therefore, it is likely that aluminates also contribute to setting.

Materials
A hydrated commercial lime (CL90s) complying with EN 459-1 was used.As aforementioned, two pozzolans, metakaolin and GGBS, were identified as having potential for use in the lime hemp composite on account of their fast setting times and high reactivity (Walker & Pavía, 2010;Walker & Pavía, 2011).The programme of research concentrates on these two pozzolans however RHA was included in several tests to study the effect of the pectins on a fully siliceous pozzolan.The pozzolan composition, rate of amorphousness and surface area are included in Table 1.
The chemical composition was determined by XRF using a Quant'X EDX Spectrometer and UniQuant analysis package.The mineralogy and rate of amorphousness were analysed by X-Ray diffraction (XRD), using a Phillips PW1720 XRD with a PW1050/80 goniometer and a PW3313/20 Cu k-alpha anode tube at 40 kV and 20 mA.The specific surface area was measured using a Quantachrome Nova 4200e and the BET method, a model isotherm based on adsorption of gas on a surface.Industrial hemp shiv was supplied by La Chanvrière De L'aube in central France.As shown in Table 1, the chemical and mineral composition, rate of amorphouseness and surface area of pozzolans (Walker & Pavía, 2010;Walker & Pavía, 2011).

Setting Time
It is not possible to determine the setting time of composites such as hemp concrete using the Vicat test as the organic aggregate impedes needle penetration and absorbs water; hastening the drying of the paste and giving inaccurate results.Therefore, the hemp aggregate was replaced by hemp water; and the effect of hemp on setting determined by comparing the Vicat test results of pastes made with water and hemp water.Mixing was in accordance with EN 459-2 except for the addition of the pozzolan (added after 1 minute and the mixing stopped for 30 seconds).
The hemp water was prepared by soaking the shiv for 45 minutes so that it releases its water soluble constituents including pectins.The concentration of soluble hemp compounds in the mixing water present in the lime hemp concrete during curing was not established.Longer soaking times would yield a higher concentration of soluble compounds, however it was felt that long soaking times in a water excess were not representative of the actual conditions in hemp concrete, thus soaking the shiv for 45 minutes was deemed an adequate solution.The influence of varying amounts of soluble hemp compounds on setting was measured by varying the hemp water content (0, 25, 50 and 100%).

Water Content for Setting
Consistency in the water content is of paramount importance as it affects the kinetics of the binder reactions and the final properties of the material: a small water increase significantly delays setting of lime-pozzolan pastes (Walker & Pavía, 2010).A standard consistency for the lime-pozzolan pastes was achieved by adding the amount of water necessary to produce a paste with a 165 mm initial flow diameter.The water required in order to produce such paste was measured according to EN 459-2.The quantity of water required by different pozzolans to produce a paste of the same standard consistency (necessary for the Vicat test) varies: GGBS has a much lower water demand than lime while, in contrast, lime and metakaolin have a similar water demand (Walker & Pavía, 2011).For the lime-GGBS pastes, the binder: water ratio of 1:0.69 and 1:0.64 for the 30% and 37.5 % (Figure 2) GGBS content respectively.As hydrated lime and metakaolin have a similar water demand, the water content of the lime-metakaolin pastes was determined at 37.5% pozzolan content and fixed at a lime+pozzolan:water ratio of 1:0.93 (by weight) for both pozzolan contents.

Compressive and Flexural Strength
Strength was measured at 30% pozzolan content with the water content adjusted to produce a 165 mm initial flow diameter.The prisms were demoulded after 1 day and stored in a curing room at 20°C±3°C and RH 60%±10%.The unconfined compressive strength was measured according to EN196-1 and EN 459-2 at 28 days.
Flexural strength tests were conducted according to EN196-1 and EN 459-2 using the center-point loading method.Here, the prisms were placed on fixed supports with the longitudinal axis normal to the applied load.
After breaking the prism, the following equation was used to determine the flexural strength: F f load applied to the middle of the prism at fracture (N) b side of the square section of the prism (mm) l distance between the supports (mm) The values reported are the mean of 6 compressive and 3 flexural strength tests.

Analysis of Hydration Products With Scanning Electron Microscopy (SEM)
SEM and XRD were used to determine whether water soluble constituents of hemp were responsible for a difference in the nature and/or amount of hydrates.The microstructure of the binder and the formation of hydrates were investigated using a Tescan MIRA Field Emission SEM.The samples were fractured and covered with a gold coating in an 'Emscope SC500' plasma coating unit.Lime-pozzolan pastes (30% pozzolan by weight) were investigated at 1, 3, 7 and 28 days.Pastes with 2% hemp content (untreated and washed hemp in which the hemp was immersed in water for 24 hours and then dried prior to mixing) were investigated at 28 days.15 random areas were analyzed in each sample.
XRD was undertaken on the lime-pozzolan pastes, at different ages up to 6 months, using a Phillips PW1720 XRD with a PW1050/80 goniometer and a PW3313/20 Cu k-alpha anode tube at 40 kV and 20 mA.

Testing Regime
The testing regime of the different pastes is as set out in Table 2.

Effect of Water-Soluble Hemp Components on Setting Time
The pozzolanic reaction is significantly controlled by the pozzolan's active silica and alumina content (Massazza, 2007).The pozzolans have a different composition ranging from the clinker-bearing GGBS to the alumina-rich metakaolin and the silica-rich RHA.As aforementioned, RHA was introduced in order to ascertain the effect of the hemp on an almost fully siliceous pozzolan.A 100% calcium lime paste was included as a control sample.
The results show that hemp's soluble constituents do not alter the setting time of lime: both the lime and the lime/hemp-water pastes display a similar set, with the final set at 110 hours and a maximum deviation of 9 hours (Figure 2).This indicates that the hemp does not alter mechanical processes such as flocculation, drying and early carbonation responsible for the initial hardening of lime.In contrast, the hemp water retards the setting of lime-pozzolan pastes.RHA is retarded the most followed by GGBS and metakaolin: 39, 23 and 10 hours respectively for a setting equivalent to a depth of penetration of the Vicat needle of 25 mm (Figure 2).Therefore, the hemp delays pozzolanic reaction and the amount of retardation is partially determined by the chemical composition of the pozzolan.It could be argued that the retardation of the set is due to the water soluble hemp constituents retaining water in the binder rather than delaying pozzolanic reaction.However, the samples were continuously weighed during drying, and the similar drying rates of all samples (with and without hemp water) indicate that there is no water retention triggered by the water soluble hemp constituents.Therefore the set retardation is due to the soluble hemp constituents interfering with the lime-pozzolan reactions responsible for setting.
The setting of the most siliceous pozzolan (RHA) is retarded the most; the initial set starts late (at approximately 80 hours).In contrast, the setting of the alumina-rich pozzolan paste (metakaolin) is delayed the least, despite the fact that it contains more soluble hemp compounds due to its greater water demand (section 2.3).This suggests that the hydration of silica is retarded to a greater extent than that of alumina and agrees with previous authors stating that retardation by pectins is greater for low C 3 A cement (Peschard, Govin, Grosseau, Guilhot & Guyonnet, 2004).
In order to assess the influence of increasing amounts of soluble hemp compounds on setting, the Vicat test was repeated with varying hemp water concentrations (ranging from 0% to 100%) for the GGBS and metakaolin pozzolans.It is evident from the results (Figure 3) that increasing hemp water content increases the delay in setting.This was true in all cases except for the metakaolin paste with hemp water content over 50% and this may be due to the presence of a surplus of soluble hemp constituents on account of the higher water content in the lime-metakaolin pastes.It was also evidenced that a small amount of hemp water (25%) significantly delays setting. .Influence of increasing soluble hemp compounds on the setting of lime-pozzolan pastes.G-GGBS; Mmetakaolin; W=100% water; H1/4=25% hemp water and 75% water; H1/2=50% hemp water; H=100% hemp water.All 30% pozzolan by weight

Effect of Water-Soluble Hemp Components on Compressive Strength
The impact of the soluble constituents of hemp on compressive strength was investigated in samples of 30% pozzolan content (by weight) made with water and hemp water, at 5, 7, 14 and 28 days (Figure 4).The lime samples at 5 days were still soft and compressive strength was therefore only investigated at 7, 14 and 28 days.
The low coefficients of variation (Table 3) indicate that the results are reliable.
The results indicate that the hemp does not affect the strength development of lime pastes however, it delays the strength development of lime-pozzolan pastes (Figure 4); in particular that of the metakaolin.The pozzolanic reaction contributes to strength development and it is evident that the hemp water delays this reaction.This agrees with the setting results indicating that hemp delays pozzolanic reaction.It differs from the setting results in that metakaolin is delayed further than GGBS.However, as strength development is largely attributed to the formation of silica hydrates, it is clear that the hemp water is delaying their formation.
The magnitude of strength reduction gradually decreases over time and, at 28 days, the pozzolan pastes made with water and hemp water show similar strength.Therefore, the soluble hemp constituents undermine early strength but do not significantly affect ultimate strength.In order to confirm that the soluble hemp components produce a delay in compressive strength development, the compresive strength test was repeated using aluminium sulphate as an additive.Aluminium sulphate precipitates pectins and should therefore reduce the effect of pectin on the lime-pozzolan reaction.The results (Figure 5) show that aluminium sulphate has no impact on the strength development of pastes made with water but improves the strength of pastes made with hemp water (in particular that of the metakaolin paste whose strength was lowered the most).This further confirms that the soluble hemp components (in particular pectin) cause a delay in compressive strength development.

Conclusion
The water-soluble hemp components delay hydration.As a result, they retard the setting and early strength development of lime-pozzolan pastes however, they do not affect their ultimate strength.In contrast, the soluble hemp compounds do not impact the setting, hardening and initial strength development of calcium lime pastes.
The results also revealed that the water-soluble hemp components delay the formation of silica hydrates further than alumina hydrates.
It was noted that a small amount of hemp water significantly halts hydration.The effect of hemp water does not increase linearly with concentration; at high hemp water contents, increasing concentration has a smaller effect.This is likely on account of a surplus of active hemp components in solution.
The SEM analysis agreed with the setting and strength results indicating that hemp water delays the formation of hydrates.This was clearly evident in the lime-RHA pastes in where a delay in the formation of CSH was readily observed.
The halt of the pozzolanic reaction by the hemp particles was clear at the hemp interface, where few hydration products appeared and their quantity slightly increased when the hemp particles were washed to remove their water soluble constituents.This paper concludes that the properties of the hemp particles, including their high suction removing water at the interface and their high porosity assisting CO 2 diffusion, contribute to the halt of the pozzolanic reaction.
This research also evidenced that pozzolanic reaction takes place early in lime-pozzolan pastes (significant hydrates were observed at 24 hours).An evolution in the morphology of the pozzolanic hydrates over time, from predominantly needle-shaped at very early ages (24 hours); into fibrous, sponge and gel-like hydrates at later stages, was determined.

Figure 5 .
Figure 5. Influence of soluble hemp components on the early compressive strength of lime-pozzolan pastes with and without aluminium sulphate additive at 7 days.G-GGBS; M-metakaolin; L-lime; H-hemp water; A-aluminium sulphate Figure unwashe should be due to the formation of CSH.A previous study which compared reactivity of a number of pozzolans found that pozzolans with the highest Al 2 O 3 content set the fastest

Table 2 .
Matrix of testing As shown in Table2, the testing undertaken in this research.

Table 3 .
Coefficient of variation of the compressive strength results in Figure4