Keywords

1 Introduction

The interest for this research to go for the “unstabilized” rammed earth (URE) lies in the latter adopting sustainable construction material inherent in a minimal embodied energy while enhancing the recycling potential [1, 2]. Yet, moisture ingress [3, 4] negatively affects compressive strength while causing walls’ deformability due to the air’s relative humidity that induces transient hygroscopic transfers into the compacted specimens [5]. The hygroscopic state is in command over suction [6] distribution which has an impact on the mechanical response of URE construction, particularly in association with stiffness and/or cracking failure. The importance of URE sensitivity to water grabs attention for investigation in this research because it exhibits a crucial issue that affects URE material’s mechanical strength to the extent that its neglect results in URE material being coined with a relatively low strength test, as the next subsection of this article suggests.

1.1 Previous Research

To address the relatively-low strength test of the “unstabilized” technique, the literature shows two strategies. One is internal suction [7,8,9,10] by the latter reinforcing the contact to occur between the soil particles resulting in a dried-earthen material that in turn increases stiffness and/or strength. A particular attempt [11], has been initiated to quantify the structural behavior while considering hygroscopic conditions, which suggests excessive time necessary to establish earthen material’s stability due to the latter’s material slowly drying effect. Another technique is soil “particle size distribution” (PSD) [12, 13], or soil ingredient composition. PSD’s importance outlines the rate at which humidity may well ingress into the compacted earthen material as a result of the capillary suction [14, 15]. PSD also affects “soil porosity and inter-particle friction/interlock” [15] by primarily relying upon dynamic compaction to maximize density, therefore a lower percentage of fine aggregates and clay content might be needed. Besides this strategy’s physical stabilization, the literature studies the chemical bonding’s effect [16] by holding on to the additives’ strategy in an attempt to intervene in the “controlled modification of soil texture, structure, or physic-mechanical properties” [9]. For URE, clay (8–14%, by mass) remains the bonding substance [17]. A slight rise in URE’s mechanical strength emerges, ranging between 1.27 MPa and 1.42 MPa [18] – common mechanical strengths for URE are 1–2.5 MPa [19, 20].

The relatively low compressive strength has encouraged many studies [21, 22] to investigate other bond elements; other than clay, particularly cement, or lime, to stem the pitfall in URE’s strength. Their conclusions, albeit variations, agree on the excellent performance of the stabilized specimens, while the moisture content upon compaction and throughout the curing stage did not record a significant rise within the cement-stabilized walls. With an increase in the cement content from 2.5% to 7.5%, the wet-to-dry strength shows a rise in a strength ratio from 0.51 to 0.85 [23], while noting that the acceptable ratio starts from 0.35 [24]. Burroughs [25] has in particular studied the suitability of cement and lime for stabilizing varied soil types by investigating the soil’s linear shrinkage and its plasticity index. Central to Burroughs’s assessment is grading neatly associated with finished texture and friability. The literature concerned with grading bounds its analyses [9] with a correlation being investigated not only between type and %wet of stabilization but also between shrinkage and plasticity while holding a focus on soil’s physical characteristics (or, grading) [27].

Both lime and/or cement nevertheless harm the environment through their materials’ pH, largely because both materials are in nature industry-produced stabilizers that not only glue the earthen materials’ grains through a chemical reaction [27] but also duplicate the “greenhouse gas emissions” [3]. Recent studies [15], therefore, have turned to environment-friendly additives. Despite the latter contributing to a rise in strength tests such altitude arguably remains limited. As this study explains, environment-friendly studies utilize ramming [28] at a moisture content to achieve workability rather than an optimum moisture content (OMC) for the maximum dry density (MDD). Indeed, Ciancio et al. [29] suggest the OMC being +1%–2%, while the New Zealand Standard NZS4298 [32] recommends that the moisture content upon compaction falls below 3% of the OMC; however, not exceeding 4% dry or 6% wet of the optimum. Furthermore, the concern in the previous studies (cited in [31]) has been either with the impact that rain has on the durability of earthen material or addressing how the wet conditions might influence the mechanical properties of earthen material specimens. A particular issue emerges in these studies which concerns a change in relative humidity on the internal suction. This issue has been well recognized in previous studies; nevertheless, less unaddressed, as this research suggests. A research gap emerges in this respect.

1.2 Research Question

To address the research gap identified in the previous subsection of this article – the gap is concerned with stemming a moisture ingress that occurs once completing the compaction process and goes throughout the curing stage, analyses in this article turn away from the strategies narrowed down by either investigating the wide-ranging choices of chemical bond substances or their effect while primarily relying on ramming to ensure workability – as the previous research suggests. Despite this article recognizing such strategies’ vital importance, it instead adopts a two-fold strategy. Laboratory analyses undertaken for this article first seek the contribution of natural infill materials to physically interlock the compacted grains and, secondly, look into the effect of a combined, natural, and locally available bond substance. The strategy responds to the question for addressing in this research which concerns how a moisture ingress; arising as a result of internal humidity or rainfall, for example, and occurring upon completing the compaction process and thereafter, might be minimized.

2 Research Methodology

The methodology adopted here concerns laboratory research undertaken within a sandy limestone case study, chosen here as a particular material. The properties of URE specimens are examined to evaluate their suitability for URE construction in accordance with universal recommendations. The PSD of the limestone material has been carefully observed and described. Concerning the mechanical performance of the URE, compression and water absorption erosion tests have been undertaken. The next subsections of this article describe the stages of the methodology program employed here: 1) materials selection; 2) specimens’ preparation; 3) specimens’ manufacturing; and lastly 4) testing of the specimens. The methodology is described as follows:

2.1 Raw Material

The material is characterized by the means of expenditure (chemical analysis, drop test, and dry strength test) and the laboratory test (PSD analysis). The sandy limestone, or sedimentary, rocks have been transported from the superficial revelations that exist along the Gaza Strip’s Mediterranean coastal line. Notably, the rocks’ soil quarry appears to be unsuitable for cultivation due to being arguably located on the outer crust of a non-agricultural area. The choice of sandy limestone, therefore, endorses the argument that its use as a material for construction would not negatively affect our Earth’s crust. Sandy limestone has further benefits, for example, it inheres to a greater consistency in mineralogy while depicting in its micrograph low-fine particles (especially, the sand ones), which might in turn interfere with cement hydration (or, “hydration water”) [33].

2.2 Specimen’s Preparation

The experiment tools include two pneumatic rammers; sieve trays (0.0150–4.75 mm), a sieve shaker, a wooden wrench, and five plastic piles. The limestone material necessary to manufacture the specimens falls within a man-made earthen soil mixture, which is predominantly composed of sand particles (55%, by mass). Initially, a realistic wooden formwork (90 cm × 45 cm in plan and 60 cm height) and three wooden cubes (10 cm × 10 cm in plan and 10 cm height) for testing have to be constructed. Three realistic sample walls and about 115 specimens have been constructed in this research. The specimens’ preparation concerns setting the sandy limestone mixture; including the grain sizes and the latter’s quantities.

At the start, the moistened sedimentary rocks brought from the coastal line have been laid exposed to the sun’s heat to dry out for, at least, a week. Once the rocks have been dried, the specimens’ preparation started with breaking down the sedimentary rocks using an iron hammer (1.689 kg) into small gravels (10–20 mm in diameter) while manually excluding the gravels <10 mm to avoid a reduction in a strength test, when the gravel size increases. Table 1 shows limestone particle quantities. Two notes worth noting in this table. Once concerns cockle shells/marble/natural stone that acts as infill materials. Another note is clay content of 5% combined with 3% burning residue of olive and orange branches which serves as an additive. Furthermore, the ranges, shown in Table 1, fall in agreement with the PSD envelope recommended by Houben and Gillard [32] and also with the material mixture adopted by most authors working within URE literature [15, 33,34,35].

Table 1. Limestone’s mixture quantities.
Full size table

Worth noting, that ordinary lime has not been perceived as a bond substance. Rather, it has been added to the mixture quantities shown in Table 1 at 7% (by mass) not only to compensate for the sandy limestone’s natural loss of lime due to acidic rain but also to decrease the earthen material’s sensibility to water while protecting the latter from an insect attack, or rot, for example. Chemical reactions in this vein occur due to the rainfall hitting the limey outer veneer of the sedimentary rocks due to the more reactive minerals that existed within the limestone than in the sandier beds [29].

The sieving stage has followed to set the grain sizes bypassing the small gravel ˂10 mm through the sieving nets to evict both the gravel (4.75–2.36 mm in diameter) and the sand grains in three levels (0.6–0.85mm, 0.150–0.6 mm, and ˂0.150 mm). The cockle shells have been fractured followed by sieving to produce the cockle shell particles (0.85–0.6 mm, in diameter). The latter’s particles, waste marble powder, and natural stone have been ground to powder ˂0.6 mm using a manual grain mill.

2.3 Specimen’s Manufacturing

This stage includes mixing, compaction and curing. The particle-size quantities have been thoroughly mixed dry four times for five minutes to enable a sufficient interlocking of the particles throughout the material mixture. A known amount of water w% has been added and the material mixture has been also mixed four times for a further five minutes. A “drop test” technique ‒ recommended by rammed earth (RE) guidelines ‒ ensures that the material mixture’s “wet state” reaches the OMC. The “drop test” starts by dropping from a height of 1.5 m a moistened and compacted material mixture by hand onto a hard, and/or flat, surface. Once the material mixture is too dry the moistened ball fragments into many segments. By contrast, when the moisture content is nearing the OMC the ball fragments into a few segments. In case the moistened ball is too wet, the ball falls in one segment onto the surface.

On the site, the material mixture has been not moistened at once, but rather in several stages to release the energy embodied within the mixture which is necessary to initiate a chemical reaction, each time when water is sprayed over the dried mixture. When satisfactory, according to the “drop test”, the moistened quantities have been poured loose into the sample formwork in layers that each has arisen about 20 cm in height. For the laboratory tests, the moistened mixture has been poured loose into the cubes (numbered 4B, 7, C, 5, 7a, and 5A) (10 cm × 10 cm × 10 cm) in five layers of 3 cm height; each layer has been compacted using both a wooden wrench and a 1.689 kg flat-based rammer at a height of 2 cm. The layers borrow the same sandy limestone-moistened mixture of its sample wall. The next stage concerns compacting each layer; namely being undertaken into three phases. The first uses an iron pestle to physically interlock the moistened mixture’s particles. Secondly, ramming uses a timber pestle that enables a much tighter interlocking. The last phase uses a timber wrench that fine-tunes the edges, followed by repeating the second phase of compaction to work out the surface before proceeding. Each layer has been sacrificed before the start with the next layer to ensure a good bonding surface between the layers. The compaction phase has resulted in 4 layers (Fig. 1); each about 14 cm in height, and 5 layers for each cube. The sample walls and the cubes have been left to cure for at least 3 days before demolding (Fig. 1). Following molding, the sample walls and the cubes have been wrapped with a nylon film for 7 days to allow each sample and its cubes to hydrate (Fig. 1).

Fig. 1.
figure 1

Molding and wrapping the specimens with a nylon film.

Full size image

Once wrapped, the specimens have been cured or hydrated. The curing periods of 7, 28, and 90 days are adopted here following ACI (American Concrete Institute) standards; commonly used for curing the concrete, to monitor the development of strength recorded in association with curing time before undertaking the compressive strength tests. By adopting the same set of curing periods applied for concrete in this research, a comparison has been established between the hydration of cement and the sandy limestone mixture. The ACI in this vein recommends the minimum curing time that corresponds to concrete achieving 70% of its characteristic compressive strength. The 70% characteristic compressive strength would be accordingly achieved at 7 days upon casting the concrete. Therefore, a minimum curing period of 7 days has been set in this research. The curing period of 7 days has been sprayed with 7% water (by mass) early in the morning, namely to avoid spraying the cold water on the sun’s heated surfaces. The other two curing periods follow this procedure, however, have been cured for 28- and 90-days and sprayed with water being set at 8% and 8.5%, respectively.

2.4 Testing

A rise in the sandy limestone mixture’s unconfined compressive strength (UCS) has been sought in this research by not only intervening in the sandy limestone mixture’s quantities but also altering the physical properties. Two laboratory tests have been conducted; one is a strength test while the other concerns porosity.

The first test measures the UCS: each specimen has been accordingly tested upon accomplishing its curing period and also being left for a day within the outdoor environment to dry out prior to testing (Fig. 2). The effect of age of curing on the UCS is shown in Fig. 8. Two notes are worth noting. One is a slight variation to be found between measured UCS of the specimens and even that existed within the same batch (Fig. 8). In comparison, as per Indian Standard (IS-456), concrete gains its strength rapidly in the initial days upon casting, namely a 65% compressive strength in 7 days and 99% in 28 days. It also gains 90% strength in 14 days and the remaining 9% strength is gained in 28 days. Hence, one may suggest that after 14 days of the curing period, concrete’s strength slowly increases. It means that the process of hydration of cement sharply increases during the first 14 days upon casting, whilst it sharply drops down thereafter. It is therefore advisable that concrete should be cured for a minimum of 14 days upon casting to achieve durability-related properties.

Fig. 2.
figure 2

Specimen’s fracture failure.

Full size image

Furthermore, the specimens’ strength in this research, in contrast to concrete, slowly increases, and this has been clearly noticeable all through the set curing periods. This note suggests that the process of hydration remains ineffective primarily due to the weak chemical reaction occurring. An example is cube 4B (Fig. 3). The relative UCS obtained in this example, upon curing periods of 28 and 90 days, has arisen by a 14.08% and a 37.08%, respectively. The curing periods of 28 and 90 days have therefore arisen steadily in strength than that already recorded within 7 days of the curing period. Upon 90 days, the UCS gained has arisen by a 20.01% in comparison with those reported within 28 days. The UCS’s percentages (Fig. 3) witness an increase of 6.02% in 28 days; when multiplying the curing period of 7 days. These also show an increase of 17.07% in 90 days; when multiplying the curing period from about a month to almost two months. What the UCS’s percentages further suggest is that the specimens require a much longer curing period than concrete specimens would do to attain durability-related properties ‒ primarily blaming the slowly occurring hydration process. In essence, a three-time increase in the UCS has been scored; namely from 6.02% to 17.07%, by multiplying the curing period from nearly a month, in 28 days, to two months, in 90 days, of curing, instead of multiplying the initial curing period of 7 days in weeks, namely in 28 days of curing period. The rate of the UCS increment ranges from 0.3 to 1.8 (megapascal) MPa per every 7 days of the curing period.

Fig. 3.
figure 3

UCS tests.

Full size image

Maximum dry density (MDD), OMC, and water ingress: a dry density has been specified in this research by weighing the specimen dry and wet; that is, by weighing that occurs before and upon soaking the dried specimen into a water basin for 24 h (Fig. 4). Before testing, all specimens have been left within the outdoor environment for five days to dry properly. Specimens’ porosity emerges as a result of subtracting the wet volume from the dry one, followed by dividing the outcome by the wet volume (Fig. 5). As is shown in Fig. 5, the moisture content decreases in all specimens tested here as the dry density increases. Or, it may be suggested that increasing the moisture content accompanies a decrease in the strength of compacted sandy sub-mixture.

Fig. 4.
figure 4

Porosity test.

Full size image
Fig. 5.
figure 5

Dry density to OMC.

Full size image

A porous structure marks the sedimentary limestone rock suggesting a water capillary movement to easily sneak into its pore spaces. To address this issue, the strategy adopted here enhances the hydration process of the sandy limestone’s compacted sub-mixture by leaning upon a filler effect, rather than an impact occurring through chemical bonding. It on the one hand uses in doing so sand particles (55%, by mass), less in size than the limestone’s gravels, as infill material. This strategy benefits from the limestone’s characteristic concerned with slowly dissolving in water; thereby the sand particles arguably hinder the spread of water to move into the pore spaces that exist between both the gravels and also within the gravels’ porous.

On the other, the strategy turns to cockle shells considered here as an infill substance between the limestone’s gravels which have been characterized by their greater surface area, or a “surface chemistry” [8, 32] that circumscribes a arise of moisture ingress [9, 27]. What their morphology structure suggests is a neatly bonded cockleshell’s composition marked by a smaller size of the grain, and aggregates, offering “slender needle-like shapes” [35, p. 133] forming a morphology of structured mixture and small in size pores. The result is an increase in bending strengths. However, the sandy limestone sub-mixture’s composition uses in this research as particle degrees, instead of ground cockle shells, due to the em manual working procedure. What remains yet unresolved are the pore spaces left unfilled between the limestone gravels. And this might explain why the specimens’ moisture contents upon compaction remain relatively high (Fig. 4) than the porosity rate ranging between 5.5% to 10.2% (by mass).

3 Tests’ Discussion

Two tests have been undertaken for this research which their analyses suggest being correlated. One concern measuring a rise in the UCS tests. URE, in this vein, has been labeled with low UCS tests, as it has been reported earlier in this article, namely not falling below 1.4 MPa while not exceeding 2.5 MPa. As shown in Fig. 2, the specimens have a UCS > 2 MPa ‒ Uniform Building Code Standard 71-1 (UBS) specifies a satisfactory UCS starting from 2 MPa. An optimum UCS has been reached in this research upon 90 days of curing, namely a strength test of 2.92 MPa, which arguably lies within the highest ranges of UCS scores recorded for URE, and these range between 2–3 MPa [6]. To explain, a granular gradient strategy in this study contributes to such a UCS score by a “soil porosity and interparticle friction/interlock” [9]. The strategy employed in this research nevertheless does not rely upon a chemical reaction supported by sufficient compaction. Rather, it targets minimizing pore spaces by interlocking the mixture particles. Two methods are adopted in doing so. One concerns the particle degrees marked by a limestone density of 2.71 g/cm3 to score a rise in the UCS tests. The second method echoes the effect of fine particles seen as an infill substance.

Yet, a granular gradient strategy has been arguably insufficient, and this has been witnessed in the relatively low score of 2.92 MPa, particularly when being compared to stabilized rammed earth’s (SRE) UCS: 3 MPa–20 MPa. Clearly, a remarkable gap in the UCS tests remains. To overcome the difference between SRE’s and URE’s bending strengths, the UCS tests in this research increase when multiplying the curing time and, therefore, the argument put forward here advocates having a long curing period towards maximizing the mechanical strength. However, the argument may appear impractical. Explanation blames the mineral calcite of the limestone and the cockle shells, which their grains slowly dissolve in water (about 10–15%) as they are in need of acid and/or extensive heat in doing so to not only release Cao particles, necessary to chemically react with H2O but also to produce the calcium hydroxide (Ca (OH)2). In this vein, the samples have been in this research cured within a natural Mediterranean environment of about 30 Celsius throughout the summer period, thereby suggesting a slow process of calcination that requires temperatures ranging between 1070–1270 Celsius. The result has arguably been a low degree of bonding betweenthe particles, evident in a maximum UCS sore of 2,92 MPa. What remarkably remains a central theme for the specimens tested in this research, that has yet to be addressed, is the pore spaces.

Such a challenge becomes evident in the second test undertaken in this research which concerns measuring a moister content upon compaction. The literature in this respect suggests suction that endorses a decrease in the UCS tests when a moisture content upon compaction arises. And this is true in this research. The moisture content upon compaction remains relatively high (Fig. 4) despite the 90 days of curing; consequently, the specimens have scored a relatively low UCS test of 2,92 MPa. Critical, as analyses in this research suggest, has been the degree of saturation being measured through the lens of porosity which has not been yet minimized in this study, and this degree not surprisingly records a relatively high OMC (Fig. 4) despite the specimens showing an increase in the dry density. The problem, as this research further suggests, lies in particle sizes [24]. Cockle shells have been blamed because their particles have arguably resulted in an ineffective filler effect. This is because cockle shell particles have left infilled pore spaces between the limestone gravels suggesting an increase in suction, thereby a decrease in the UCS tests – as this research suggests. From the perspective of a capillary movement, pore spaces being left unfilled remain at stake.

Figure 4 shows a slight decrease in moisture content despite an increase in dry density. The pore structure of the limestone gravels and the cockle shells’ particle sizes are arguably blamed for laying a template of network channels to be occupied by a moister content shown by a relatively high OMC when conducting the porosity test. What appears to suggest is the filler effect of cockle shells remains ineffective in this research. The cockle shell particle degrees pose a further problem by the limestone mixture composition necessitating a long curing period when a comparison is made with the ground cockle shells marked by a “surface chemistry”. An explanation for a long time, as analyses in this research suggest, lies in a moisture content that sneaks into the pore spaces and, when evaporating, the particle size distribution points to the yet unfilled pore spaces, thereby scoring a low UCS test of 2,92 MPa.

4 Conclusions

The aim of this research is with scoring a rise in the UCS tests of URE. Key findings emerge in this research as follows:

  • Granular gradient strategy has been effective, yet it remains insufficient on its own;

  • In comparison with those suggested in previous research, a slight rise in the UCS tests of the URE has been achieved in this research, yet this rise remains relatively low in comparison with the stabilized soil mixture;

  • Material strength has been arguably associated in this research with minimizing the pore spaces or creating less pore structure, necessary to penetrate the capillary movement of moisture content.

  • A shortage concerned with the infill particles’ effect has been evident in this research which arguably explains the relatively low rise in the UCS tests. The shortage blames the use of cockle shell degree particles, instead of employing powder, thereby not benefiting from the bonding effect of “surface chemistry”;

  • To obtain durability-related properties, this research’s analyses suggest that URE does arguably need a long curing period in comparison with concrete specimens. And this appears to be impractical.

  • This research’s analyses endorse the contribution of an infill effect towards maximizing the material strength of URE, however, a chemical reaction to bond the fine particles residing between the gravels remains vital.