Mitigating shrinkage of alkali activated slag with biofilm

As an emerging alternative cementitious binder, alkali activated slag (AAS) is gaining great attention, but considerable shrinkage caused by alkali activation and drying limit its potential applications. Herein, we demonstrate that the addition of an environmentally benign biofilm, cultured from B. subtilis, mitigates both the autogenous and drying shrinkage of AAS. The influences of the biofilm on the hydration kinetics, water absorption and strengths are investigated. Results show the addition of the biofilm increases the hydrophobicity of the pore wall, which in turn decreases the capillary tension. The hydrophobic modification by the biofilm significantly reduces the water loss from the AAS to its direct environment (up to 86% at 35 d exposure). Consequently, both autogenous and drying shrinkage of AAS are dramatically reduced. Moreover, a new mechanism is proposed to explain the mitigation of AAS shrinkage, which takes into account the increase in internal RH and reduction in capillary pressure.


Introduction
With the enormous demand for cementitious materials in construction and increasing global concerns over sustainable development, alternative binders for cement are sought to reduce the environmental impact of cement, in particular with respect to CO 2 -emission and energy consumption [1][2][3][4][5][6]. Alkali activated slag (AAS) is a promising alternative cementitious binder with lower CO 2 emission than Ordinary Portland Cement (OPC) during production, excellent fire resistance and adequate mechanical properties [2,6,7]. Notwithstanding, the large scale application of AAS remains limited due to its high rate and extent of shrinkage and consequently high risk of cracking [6,[8][9][10]. The autogenous and drying shrinkage of AAS are 10-fold and 3-to 6-fold larger than that of OPC, and up to now there are no promising methods to solve this issue [10][11][12].
Generally, the shrinking of AAS can be attributed to moisture loss through two mechanisms: (a) internal water loss (i.e., hydration), or (b) moisture loss to the environment due to evaporation [9,10,[13][14][15]. The former results in autogenous shrinkage by the continuous consumption of water during the reaction of slag, which decreases the internal relative humidity and generates a water-air meniscus inside AAS [9,10,15]. Drying shrinkage originates from a lower environmental humidity than the internal humidity of AAS. This causes evaporation of water and a reduced humidity in the pores, which also leads to capillary stress as water menisci begin to develop in the capillary pores [11,13]. AAS exhibits considerably higher drying shrinkage than OPC, which hinders its large scale application in construction industry [11,13].
Various strategies have been developed to overcome the high shrinkage of AAS, with the general principles of either compensating with water or providing extra volume to mitigate the volumetric reduction. Among these, the most commonly applied methods are the addition of internal curing agents [14,16,17] or shrinkage reducing admixtures [10,15,18]. Internal curing agents, such as lightweight aggregates or superabsorbent polymers (SAP), supply extra water to compensate for the water consumption [14,16,17]. In this way, the capillary pressure induced by the drop of relative humidity is mitigated [19][20][21]. However, the addition of internal curing agents could deteriorate the mechanical properties of AAS, especially for the systems with lightweight aggregates due to their high porosity [10,16,22,23]. The addition of shrinkage reducing admixtures (SRA) can reduce the https://doi.org/10.1016/j.cemconres.2020.106234 Received 7 June 2020; Received in revised form 23 September 2020; Accepted 24 September 2020 T surface tension of the pore solution, which reduces the capillary pressure (according to the Kelvin-Laplace equation), so that the relative humidity is maintained at a higher level [16,18,19,21,24,25]. Previous researches reveal that not all the SRA that are widely used in OPC-based systems work efficiently in AAS and SRA are less efficient in mitigation drying shrinkage than autogenous shrinkage as SRA do not reduce water evaporation [2,10,24]. It is therefore important to develop new types of multifunctional chemicals to efficiently mitigate both the autogenous and drying shrinkage of AAS.
Another important parameter which also has a large influence on the capillary pressure generated in porous media is the wettability of the pore wall. The hydrophobic modification can dramatically decrease the capillary forces in the porous medium, therefore influencing other properties such as liquid transport in the pores and the evaporation rate [26][27][28]. Shahidzadeh-Bonn et al. [26] reported that the capillary rise in a porous medium composed of hydrophobic glass beads is only one tenth of the capillary rise when hydrophilic beads are employed. Accordingly, the evaporation rate can be decreased by 75% under the same environmental conditions. Addition of hydrophobic agents into cementitious materials is already applied to limit water ingress, which results in the improvement of durability [29][30][31]. However, there is no related work reported to mitigate the shrinkage of the AAS through the hydrophobic modification [29][30][31].
Materials can be 'hydrophobized' by addition of hydrophobic agents such as fatty acids, (polymeric) hydrocarbons or silanes and siloxanes [31][32][33][34]. Environmental concerns inspired researchers to explore biotechnological methods for the sustainable production of hydrophobic agents considering for instance the benign features or natural products [24,29]. A fascinating approach, which was recently introduced is the addition of a biofilm into concrete to change its wetting behaviour [35]. The group of Lieleg applied B. subtilis 3610 biofilms in mortar and found that the ingress of water by capillary forces was decreased, which was attributed to the enhanced hydrophobicity of the mortar [35]. Their further research resulted in a US patent on an engineering hybrid cement-based composition with increased wetting resistance [36]. B. subtilis colonies were found to produce bio-surfactants during growth, which lowered the air-water surface tension to allow for colony expansion [37][38][39]. Bio-surfactants are biologically surface-active chemicals produced by microorganisms [40,41]. Lipopeptide bio-surfactants produced by Bacillus strains have been reported as one of the most efficient bio-surfactants, in terms of minimal surface tensions and critical micelle concentration [40,41].
Inspired by the recent biotechnological approaches to hydrophobize mortar, we here study whether addition of a biofilm to AAS acts as hydrophobic agent to enhance its resistance to water ingress and mitigate autogenous and drying shrinkage. Bacillus subtilis 3610 is chosen because of its generation of amphiphilic protein Bacillus surface layer protein A on its cell surface which present rather hydrophobic properties by forming a hydrophobic surface layer and increasing the microroughness of the biofilm surface [42]. A new shrinkage mitigation mechanism through the hydrophobic modification is proposed, which sheds light on the large-scale application of AAS. A direct comparison of the performance of the biofilm and a traditional polyether-type SRA reveals that the biofilm is not only efficient in decreasing the autogenous shrinkage, but also in reducing the drying shrinkage, as the water loss of AAS is dramatically decreased after the hydrophobic modification. The influences of the biofilm on hydration kinetics, reaction products, surface tension of pore solution, internal relative humidity, water absorption and water loss of the AAS are also investigated. Our results show that both the autogenous and drying shrinkage of AAS can be effectively reduced by the addition of biofilm, which is environmentally friendly and valuable in widening the commercial acceptance of AAS.

Materials
Ground granulated blast furnace slag (GGBS) is obtained from ENCI (the Netherlands). The chemical composition of the GGBS was determined by X-ray fluorescence and is given in Table 1

Preparation of the (biofilm-modified) alkali activated slag (AAS)
Biofilms are formed by bacteria on various surfaces via the synthesis and secretion of a cohesive and protective extracellular matrix that helps them tolerate harsh or nutrient-poor environments. Natural scaffolds for bacterial biofilm formation often consist of amyloid fibers, proteins that self-assemble to form various cross-beta nano-architectures. The preparation process of biofilm modified alkali activated slag (AAS) is shown in Fig. 1. Firstly, frozen B. subtilis bacteria is inoculated overnight in 15 ml liquid Luria/Miller LB-Medium at 37°C. Next, the cultures are transferred to a shaking incubator (250 rpm) to grow for 12 h, after which the liquid culture is plated on agar plates enriched with LB Plus medium, to allow the B. subtilis liquid cultures to generate a biofilm on the agar plates. After 24 h of growth, the biofilms were harvested by scraping by rubber spatulas (rubber policemen) from the plates. Optical tensiometry of a sessile water droplet on the biofilms on agar yielded a contact angle of 143°, indicating that the B. subtilis biofilm is superhydrophobic (Fig. 2).
Next, the alkali activator is prepared by mixing sodium hydroxide pellets into a sodium silicate solution. The activator was firstly cooled down to room temperature prior to mixing with the extracted biofilm. Finally, the biofilm-contained activator is added to the raw material GGBS to prepare the (biofilm-modified) AAS using a laboratory mixer. The AAS samples are prepared with mass ratios of biofilm to dry slag of 1%, 2%, and 3% and an equivalent Na 2 O content (activator) of 6.2% by the mass of slag. An overview of the five specimens examined is given in Table 2. The activator modulus and water/binder ratio are kept constant as 1.2 and 0.45, respectively. To compare the performance of biofilm modified AAS, we prepare a negative and positive control sample. The former is prepared by mixing GGBS with an activator to which no extra components are added. The latter is prepared by addition to GGBS of an activator to which a polyether-type shrinkage reducing admixture (SRA) is added to the activator. Herein the ratio of SRA mass to AAS binder dry mass is fixed at 3%.

Characterization of the AAS samples
The influence of the biofilm on the reaction kinetics of the AAS is investigated by isothermal calorimetry using an isothermal calorimeter (TAM Air, Thermometric). To this end, specimens are first prepared by mixing the activator with the GGBS for about 1 min and then the prepared paste is immediately transferred into the calorimeter to monitor the heat release over a time period of 80 h at a fixed temperature of 20°C. The results are provided in heat release normalized by the mass Table 1 The chemical compositions of GGBS used in this study (mass% as oxide). SiO  of GGBS. The water contact angle of the AAS cured for 3 days is determined in triplicate using the sessile drop method on a goniometer (DataPhysics SCA20). Pore solution of AAS cured for 1 and 3 days are obtained by squeezing the cubic specimens under a pressure of 70 MPa [4,43]. Surface tension analyser is used to analyse the surface tension and the surface tension of distilled water 72.22 mN/m by this analyser. The X-ray diffractometric (XRD) analysis was performed on powders of 3d cured AAS pastes by using a Cu tube (40 kV, 30 mA) with a scanning range from 5°to 65°2θ, applying a step 0.02 and 5 s/step measuring time. The FT-IR spectra of the reaction products (3d) were collected using a PerkinElmer FrontierTM MIR/FIR Spectrometer using the attenuated total reflection (ATR) method (GladiATR). All spectra were scanned 48 times from 4000 to 400 cm −1 at a resolution of 4 cm −1 . Xray photoelectron spectroscopic (XPS) is performed on powders of 3d cured AAS pastes and biofilms scraped off the agar plates to confirm the presence of biofilm in AAS using an X-ray photoelectron spectrometer (ThermoScientific K-Alpha) and the spectra is fitted by CasaXPS software.

Mechanical properties under drying condition
The influence of the biofilm on the mechanical properties of AAS under drying condition is investigated as concrete structures are always exposed to drying environment which can deteriorate the mechanical performances due to the water loss [44]. The AAS specimens are kept under laboratory conditions at a temperature of 20°C and a relative humidity (RH) of 50%, which results in drying of the samples. The flexural and compressive strengths of the AAS with 1, 7 and 28d age are determined according to EN 196-1 [45].

Water absorption
The ingress of water in the AAS specimens is determined in a capillary water absorption test performed in accordance with EN 480-5 [46]. AAS samples are firstly cured for 3 days during which the specimens are vertically stored in a chamber with a RH of 95% at 20°C. Next, the samples are immersed in water with a depth of about 3 mm for 43 days, during which the change in AAS mass is periodically measured.

Internal humidity
The internal relative humidity (RH) of AAS was measured by a Hygropin moisture meter (Newa Techniek B.V.) with in-situ RH sensor probes. The measuring sleeves with 6 cm depth were embedded in the fresh AAS paste after which the surfaces of the cast samples were sealed with plastic wrap. The internal RH was tested daily during a period of 28 days. Three replicate measurements were recorded to calculate the average internal RH.

Autogenous and drying shrinkage
Based on the ASTM C 1698-09 [47], autogenous shrinkage is examined on samples cast in corrugated polyethylene tubes (Φ 29 mm × 420 mm) under the constant temperature of 20°C. Drying shrinkage was studied of 40 × 40 × 160 mm 3 AAS samples held in a climate chamber with a fixed temperature of 20°C and relative humidity of 50%, respectively. During the drying shrinkage test, the mass loss of the AAS samples is measured simultaneously.

Reaction kinetics and reaction products
The heat evolution and accumulated heat release of AAS samples with 3% biofilm (BF), with 3% polyether-type shrinkage reducing admixture (SRA), and neither (Ref) are shown in Fig. 3(a) and (b). Five distinct periods including pre-induction, induction, acceleration, deceleration and steady state diffusion stages can be observed in the curves, which is similar as OPC and other alkali activated materials [15,48]. Both the addition of 3% biofilm and SRA lead to a retardation effect in the reaction of AAS, which would consequently help with the reduction of autogenous shrinkage at early ages, as will be discussed in the following sections. It has been reported that the addition of SRA decreases the polarity that results in the reduction of alkalinity of the pore solution. Hence the reaction process is slowed down, and as a result the formation of the C-A-S-H gel is delayed [18,49]. For the biofilm-containing AAS, the reason may be attributed to the accumulation of bacterial end-products induced by the metabolic activity of B. subtilis that affects the release and availability of Ca 2+ for alkali activation, which has been reported to have a negative function on the hydration of OPC. It should be noted that during the growth of B. subtilis, bio-surfactants is produced to weaken the water surface tension which has a same function as the SRA to reduce the polarity and then alkalinity of the pore solution. This also can be used to explain the more prominent retarding effect of the biofilm than SRA at same dosage.
The FTIR patterns of the AAS are shown in Fig. 3(c). All the reference and biofilm modified samples present similar infrared spectra.  Fig. 4(d). We can conclude that the products composition of the AAS is not changed by the addition of biofilm. The main phases are C-A-S-H type gels.

Influence of the biofilm on the wetting, microstructure and water absorption of AAS
XPS experiments were performed on an AAS specimen with 3% biofilm, a reference AAS specimen without the biofilm and on a pristine biofilm to investigate whether the biofilm is present at the surface of the biofilm-modified AAS (Fig. 4g). The presence of the highly hydrophobic biofilm on the surface of the hydration products is evident from the unique nitrogen peak in the XPS patterns belonging to the biofilm, which is absent in the XPS spectrum of the reference AAS (Fig. 4g). Next, WCA experiments were performed to study whether the biofilm renders the AAS surface more hydrophobic. As expected, the addition of the B. subtilis biofilm increases the wetting resistance of the AAS (Fig. 4a) as the water contact angle (WCA) of the pores is increased upon the addition of the hydrophobic agent [29,31]. The WCA of AAS increases with the biofilm content from a rather low value of 45.7°for AAS with 1% biofilm to a much higher value of 78.5°for the 2% specimen. There is a small increase of 3°in WCA to a maximal value of 81.5°upon a further increase in biofilm content from 2 to 3%. This indicates that the WCA of biofilm-modified AAS would not reach the very high WCA value of 143°observed for the pristine biofilm on agar. This is in contrast with previous work that the addition of a biofilm can enhance the WCA by increasing the multi-scale roughness of OPC [35].
To examine whether the biofilm alters the microstructure of the AAS, specimens were studied by scanning electron microscopy ( Fig. 4(c)-(f)). Surprisingly, we find that the microstructure of biofilmcontaining AAS is flatter than that of the reference specimen without the biofilm, which displays a rougher microstructure. This clearly demonstrates that the biofilm has an influence on the morphology of the hydration products of AAS. The observed flattening contrasts sharply with the effect of a biofilm addition on OPC, which was previously reported to enhance its multi-scale roughness [35]. These findings explain why the highly hydrophobic biofilm with a contact angle of 143 o unfortunately does not render the AAS with very high hydrophobicity. Two prerequisites must be met to create highly hydrophobic materials reminiscent of the lotus leaf: the microstructure must display multiscale roughness (at the micro and nano length scales) and the surface must have a low surface energy [35,50]. Clearly, biofilm-modified AAS does not meet these two requirements simultaneously as the microstructure does not display multiscale roughness. Consequently, the highly hydrophobic biofilm on the surface of the hydration products does raise the WCA, but not to values above 90°, i.e. being hydrophobic.
Interestingly, the biofilm does significantly lower water absorption (Fig. 4b) despite the WCA < 90°, which is still classified as 'hydrophilic' [35,51] which is in line with the findings by Ertelt et al. who focused on cement mortar [52]. The wetting performance experiments clearly show an effect of the biofilm on both the extent and rate of water absorption. With higher dosages of the biofilm, the water absorption is significantly reduced. At 2% and 3% addition of biofilm, the water absorption is decreased by 80% and 85%. This can be attributed to the increase of the hydrophobicity of the pores of AAS. Therefore, we can conclude that the addition of biofilm enhances the difficulty of water to penetrate the AAS through the empowered hydrophobicity, which results in the improved durability of AAS.

Influence of the biofilm on the surface tension of the pore solution and the internal relative humidity of the hybrid AAS
The effects of biofilm and SRA on the surface tension of the pore solution of the AAS are shown in Fig. 5(a). The surface tension of the 3 days' pore solution is a little lower than that of 1 day, which can be attributed to the continuous hydration of slag and more ions in the pore solution precipitate to form C-A-S-H gel [10]. It should be noted that the pore solution of the reference presents a relatively high value of surface tension (around 71 mN/m at 1d) which is close to the value of distilled water (72.22 mN/m). This can result in a high capillary stress which increases the shrinkage of AAS [10,53]. At 1d, the addition of the biofilm decreases the surface tension of the pore solution of AAS by 14.2%, 20.7% and 29.8% when 1%, 2% and 3% biofilm by mass of the slag is added, respectively. This is partially due to the retardation effect when biofilm is added into the AAS [54], as shown in Fig. 3(a), the biofilm decreases the reaction rate and less ions dissolve in the pore solution. Another reason may be attributed to the bio-surfactants produced during the growth of B. subtilis which weakens the surface tension of water to allow expansion of the biofilm [38,55,56]. The addition of 3% SRA remarkably reduces the surface tension of AAS by around 50% which is similar to the function as reported in OPC or other alkali activated materials [15,18]. The reduction of the surface tension of the pore solution will result in lower capillary stress in the pores of AAS [10,57].
The addition of biofilm also influences the internal humidity of the AAS which is shown in Fig. 5(b). For the reference AAS, the internal humidity decreases quickly in the first 3 days due to the formation of menisci in the capillary pores [57]. The internal RH of the AAS with SRA decreases much slower in the first 7 days and then reaches stable around 86% which is 20% higher than the reference sample. The AAS samples contained 2% and 3% biofilm present similar decreasing pattern of internal RH. It has been known that reduced pore solution surface tension resulted in the decrease of the internal equilibrium RH according to Kelvin equation [10,18]. However, it should be noted that even with a higher surface tension, AAS with 2% and 3% biofilm presents almost the same internal RH changes as SRA contained AAS. This may be attributed to the enhanced hydrophobic property of the pore wall of the AAS. It reduces the potential for the pores to be wetted by pore solution, in turn generating a less curved meniscus with higher internal RH [14,58]. However, at this point of research, it is not clear yet whether biofilms with hydrophobic surface properties are required to obtain the effects observed here, and further investigation is still needed. Fig. 6 shows the autogenous shrinkage of the AAS with 3% SRA content and different contents of biofilm. It can be seen that the increase of the biofilm addition dramatically reduces the autogenous shrinkage of the AAS. With 3% biofilm content, the autogenous shrinkage of hybrid AAS decreases by about 70% at 56 days compared with the reference. It should be noted that there is a sharp decrease of autogenous shrinkage from 1% to 2% addition of biofilm which shows similar pattern as the changes of wetting resistance as shown in Fig. 4(a). This indicates that the wetting behaviour of the AAS shows a key influence on the mitigation of autogenous shrinkage. With 3% SRA addition, the autogenous shrinkage of AAS at 56 days is reduced by about 62% compared with the reference while lower than the AAS with 3% biofilm. As reported in [15,18], reduced surface tension results in the decrease of capillary pressure which is beneficial to mitigate the autogenous shrinkage of AAS. It also can be seen that the use of biofilm at the same weight ratio to mitigate autogenous shrinkage is more efficient than SRA.

Proposal of a new mechanism for the mitigation of the autogenous shrinkage
Concrete shrinks due to the moisture loss as the hydration reaction and drying proceeds. According to Young-Laplace equation (Eq. (1)) and Kelvin equation (Eq. (2)), the drop of the relative humidity leads to an increase in the tensile stress of the pore solution in capillary pores, which is the main driving force for the high rate of autogenous shrinkage of AAS [10,15].
where, σ is the capillary pressure in MPa, γ is the surface tension in N/ m, θ is the water contact angle of the wall of the capillary pore, r is the radius of the pore in cm, RH K is the internal relative humidity because of menisci formation, V w is the molar volume of pore solution in m 3 / mol, R is the universal gas constant which is 8.314 J/(mol·K) and T is the absolute temperature which is 293.15 K. RH is the experimentally measured internal humidity, RHs is relative humidity due to dissolved salts and assuming that the internal humidity before setting is the due to the dissolved salts and it keeps constant during the hydration. The water in the pore system of AAS is consumed following the order of bigger pores to smaller pores [19,59], which results in the formation of menisci between liquid and gas in the capillary pores and then capillary pressure rises as shown in Fig. 7(a).
When applying Eq. (1) or Eq. (2) in cementitious materials, the water contact angle of the pore wall is always taken as 0 and cos θ = 1 as cementitious materials are hydrophilic [10,15,18,19,21]. In this situation, the pore radius is equal to the Kelvin radius which is shown in Fig. 7(a). As the shape of the meniscus in the pore is concave, negative capillary pressure is generated, which is the driving force for the shrinkage of cementitious materials [12,15]. Wetting behaviour plays an important role to influence the internal RH and capillary pressure with water contact angle values varying according to the nature of the solid matrix [12,26,27,58]. However, the influence of the wetting property to the shrinkage of cementitious materials has not been investigated systematically, as such, a parametric investigation is performed in this study to see how the water contact angle influences the internal RH and capillary pressure.
A parametric study is carried out to study the influence of biofilm contents on RH and capillary pressure as a function of pore radius at 3d, as presented in Fig. 8(a) and (b). It should be noted that the influence of contact angles on internal humidity caused by menisci formation and capillary pressure is mainly performed on the pores with a radius less   than 50 nm which belongs to the capillary pores. This is consistent with the other researches [11,19], finer pores generate higher capillary stress resulting higher autogenous or drying shrinkage. In this range of pore sizes, the curvature of meniscus decreases with increasing hydrophobicity of the pore wall ( Fig. 7(b)). The internal RH rises and the capillary pressure drops with the increase of the water contact angle, which is confirmed by the experimental observation on the evolution of internal RH (Fig. 5(b)). The decrease of the capillary pressure mitigates the autogenous shrinkage of AAS.

Drying shrinkage and mass loss
When exposed to a low humidity environment, AAS suffers shrinkage due to not only the self-desiccation, but also the water evaporating from the capillary pores and gel [12,60,61]. Unlike OPC, more water is not chemically bound in AAS and some free water maintains as interstitial water [12]. As a result, the moisture loss rate of AAS is high due to the relatively large amount of non-chemically bound water which is prone to evaporate to the drying environment, leading to high capillary stress between the wet and dry areas of the capillary pores [54,62]. The development of drying shrinkage and mass loss of the AAS with SRA and different B. subtilis biofilm contents are shown in Fig. 9.
The drying shrinkage and moisture loss present a strong interconnected relation, namely the water evaporation causes a volumetric instability, which is in line with the results in the previous studies [54,62]. The influence of the biofilm content on drying shrinkage and mass loss shows a similar pattern like the wetting property (see Fig. 4a), with an obvious increase from 1% to 2% addition of biofilm but much less prominent with the further increased addition content. With the increase of the biofilm content, the drying shrinkage of the AAS is dramatically reduced. At 2% and 3% addition dosages, the drying shrinkage is decreased by 67% and 78% compared to the reference, respectively, while mixing 3% SRA decreases drying shrinkage by 58.7%. Meanwhile, the addition of biofilm decreases the moisture loss of the AAS dramatically. As shown in Fig. 9(b), the moisture loss at 35 d is decreased by 73% with 3% biofilm. This can be attributed to the enhanced hydrophobicity of the AAS induced by the biofilm, resulting in the decrease of the capillary stress, which reduces liquid transport inside the AAS. The increased difficulty of water supplement breaks the hydraulic connection between the saturated internal space and evaporation zone, therefore, the evaporation rate of a hydrophobic porous medium is lower than that of a hydrophilic porous medium [26,27,58]. Since less water evaporates, less menisci form due to evaporation, which in turn lowers the capillary stress [54,62]. The addition of SRA   decreases the moisture loss of the AAS by 23%, which is also attributed to the reduction of capillary pressure as SRA efficiently decreases the surface tension of the pore solution [54]. The addition of the B. subtilis biofilm can generate menisci with large radii and reduce moisture loss to the environment, which therefore results in a dramatic mitigation in drying shrinkage. Consequently, as presented here, the hydrophilic matrix of AAS shows dramatic positive effects on properties water absorption and shrinkage. It is of interest to explore whether biofilms generated from other bacteria also have a favorable impact, which will be investigated in our future research.

Mechanical properties
The mechanical properties of the AAS are shown in Fig. 10. It can be seen that at 1d, both the compressive strength and the flexural strength are lower compared to the reference. This is attributed to the delay of the reaction which slows down the strength development due to the retardation effect of the biofilm and SRA. However, at 7d and 28d, the compressive strength of biofilm modified AAS are higher or similar to the reference. Previous research has attributed the strengthening effect to the bio-mineralization of bacterial materials in OPC [63,64]. However, the direct proof such as enhanced calcium carbonate precipitation in AAS is not found from the XRD or the FTIR results. One possible explanation could be the decrease of the water loss of the biofilm contained AAS in the drying environment which is helpful for sufficient ion exchange between the activator and the slag while quick drying results in absence of water which stops ion exchange [65]. A similar strengthening effect can also be observed in the flexural strength at 7d and 28d which is beneficial for AAS which has been reported to be strong under compressive load, but relatively weak under tensile stress. The addition of biofilm and/or SRA only increases the flexural strength of AAS, while the effect on the compressive strength is limited compared to the reference sample, especially at later curing stages. This may be because the flexural strength is more sensitive to microcracks, which occur less in the presence of the additives, because they mitigate the shrinkage [10]. Furthermore, the low flexural strength of the reference sample is mostly due to cracks induced by thermal stress due to very high initial reaction rates, which can be seen in Fig. 3a. This negative effect can also be reflected by the limited increase of the flexural strength of the reference beyond 1-day, due to the generated microcracks.

Conclusions
Despite its promising mechanical performance and eco-friendly characteristics, the cementitious binder AAS has not yet been applied on large-scale due to the considerable shrinkage issues. In this study, we aim to mitigate the autogenous and drying shrinkage of alkaline activated slag (AAS) by the addition of biofilm and understand the involved mitigation mechanism. The influences of the addition contents of biofilm on the wetting performance and water absorption of AAS are explored. Based on the experimental results above, the following conclusions can be drawn: • A new mechanism to mitigate the autogenous shrinkage of alkali activated slag through the hydrophobic modification of the pore wall is proposed. With enhanced hydrophobicity by the addition of biofilm, the internal RH rises and the capillary pressure drops which decrease the autogenous shrinkage of the AAS. The addition of the biofilm at 2% and 3% dosage dramatically decreases the autogenous shrinkage of the AAS by 60% and 70% at 56 d, respectively.
• Biofilm significantly reduces the drying shrinkage which is much more efficient than conventional shrinkage reducing admixture as a combined result of decrease in capillary tension and moisture loss of the AAS. The addition of the biofilm decreases the water movement to the exposed surface which in turn reduces the evaporation rate of the pore solution.
• The addition of the biofilm enhances the hydrophobicity of the AAS, resulting in the reduction of water absorption. With a WCA of 81.5°, the water absorption of AAS is decreased by 85% under the 3% addition content of biofilm. The enhanced water resistance demonstrates that biofilm is a promising admixture to increase the durability of AAS.
• The addition of biofilm delays the hydration of AAS which is helpful to improve the workability for casting as the setting time of AAS is often very short. The 1-day mechanical property of AAS with biofilm is weakened due to the delay of the hydration. However, at later age, the compressive and flexural strength of AAS with biofilm are higher than the reference.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.