Impact of bacterial admixtures on the compressive and tensile strengths, permeability, and pore structure of ternary mortars: Comparative study of ureolytic and phototrophic bacteria

The addition of bacterial biomass to cementitious materials can improve strength and permeability properties by altering the pore structure. Photoautotrophic bacteria are understudied mortar bio‐additives that do not produce unwanted by‐products compared to commonly studied ureolytic species. This study directly compares the impact of the addition of heterotrophic Bacillus subtilis to photoautotrophic Synechocystis sp. PCC6803 on mortar properties and microstructure. Cellulose fibers were used as a bacteria carrier. A commercial concrete healing agent composed of dormant bacterial spores was also tested. Strength, water absorption tests, mercury intrusion porosimetry, differential scanning calorimetry, thermogravimetric analysis, and scanning electron microscopy were applied to experimental mortar properties. The photoautotrophic modifications had a stronger positive impact on mortar strength and permeability properties than sporulated heterotrophic modifications due to differences in surface properties and production of exopolysaccharides. The findings provide support for photoautotrophic species as additives for mortars to move away from ammonia‐generating species.

harmful metabolic by-products.[11][12] However, no studies exist so far that compare heterotrophic and photoautotrophic microbial species for mortar and cement biotechnology.
The viability of cells is a driving factor in optimizing cement properties.Bacterial cells directly added to cement can be damaged by mechanical stresses during mixing, the high pH of cement, and be crushed over time by shrinking pores.The use of spore-forming bacteria has been favored to increase viability, though requires activation by crack ingress water and the addition of nutrients [13] .The viability of cells can be increased by incorporating them in a "carrier" material, [13,14] which protects the bacteria from the stresses of the cement environment. [15]Carriers must be inexpensive, have no negative impacts on cement properties, and be easy to scale up commercially.
Existing research directions on mortars with bacteria additions (biomortars) can be split into two categories based on the desired impact of bacteria: (1) mortar property improvement, for example, strength; (2) intensification of the self-healing process.When adding bacteria for the improvement of mortar properties, the biomass and carbonates plug the pores of the cement matrix, but still allow some permeability of water. [16]The transportation of water from the surface to the inner matrix completes the cement hydration reaction for optimum cement strength. [17]Therefore, the optimum concentration of bacteria must be low enough to not completely block the surface pores.However, higher bacterial concentrations are required for self-healing.
Regarding abiotic self-healing, various techniques are used to improve the cement microstructure by enhancing the cement hydration reaction, calcium hydroxide carbonation, and mineral precipitation to seal small cracks. [18,19]Biotic self-healing approaches optimize microbialinduced CaCO 3 precipitation (MICP) through microbial metabolic activity. [13,20]High concentrations of bacteria rapidly precipitate carbonate across the surface of the cement, sealing pores and preventing the transportation of water to the interior matrix. [21]For both mortar property improvement and self-healing intensification, infilling pores and cracks are the targets of MICP application and require the determination of an optimum number of bacteria.
Approximately 84% of species used in bio-mortar studies are from the genus Bacillus, with a focus on species that use the ureolytic metabolic pathway. [22]Since bacterial strains differ in their biochemical functions and metabolism, the rates of calcium carbonate precipitation differ and can be linked to bacterial abundances, surface properties, exopolysaccharide (EPS) production and the purpose of MICP. [7]Commercial self-healing additives based on heterotrophic bacteria for MICP-based bio-mortars already exist, one of which was patented by the Green-Basilisk BV company in 2017. [23] the few studies using photoautotrophs as additions to biomortars, most demonstrate success in improving properties.Synechocystis, Synechococcus, and Gloeocapsa cyanobacteria species were able to adhere to mortar surfaces and successfully precipitated carbonates in both live and UV-killed states when incubated with a calcium and a carbonate source. [9,10]Photosynthetic organisms shift the calcium carbonate equilibrium to calcium carbonate production because of photosynthesis.Bacterial cells can increase the local calcium oversaturation through binding of Ca 2+ on the cell surface and on the EPS. [24,25]Thus, cell surfaces and EPS can nucleate CaCO 3 crystals. [9] is still unknown how the addition of different phototrophic species impact the mortar properties, and how efficient they are compared to heterotrophic species.This study aims to evaluate the impact of the additions of hetero-and photoautotrophic microorganisms immobilized on cellulose fibers on ternary mortar properties.The focus is on how microbial additions to the mortar impacts the properties to prevent cracking.Ureolytic studies were not included as the goal is to find alternatives to ammonia-generating species and explore photoautotrophic impacts.A model heterotroph Bacillus subtilis (B.subtilis) was used in a vegetative and sporulated state.A model photoautotroph Synechocystis sp.PCC6803 (Syn.PCC6803) was used in a live and UV-killed state.A commercial self-healing additive produced by Green-Basilisk BV was also used.The application of cellulose fibers alone was performed as a reference experiment.Mechanical strength was characterized through compressive and tensile strength tests; permeability properties and pore structure were evaluated by water absorption and mercury porosimetry, respectively; and mortar composition including carbonate content was estimated through differential scanning calorimetry (DSC), X-Ray diffraction (XRD), and thermogravimetric analysis (TGA).The morphological properties of the cells and mortar samples were examined using scanning electron microscopy (SEM).

Bacterial modifications, fibers, and mortar mixtures
Spore-forming B. subtilis strain UAMH 6930 was purchased from the UAMH Centre for Global Microfungal Biodiversity and cultured in a Yeast Extract-Peptone-Urea nutrient medium.Endospores were prepared using agar plates of Shaeffer's sporulation medium. [26]Syn.PCC6803 was received from the Pasteur Culture Collection and cultured in BG11 media (Supplementary information).UV-killed Syn.PCC6803 cells were prepared by subjecting cells to UV-radiation for 1 h.Aliquots of cell cultures were diluted to 10 mL at concentrations of 10 7 cells/mL using sterile deionized H 2 O and incubated with cellulose fibers for 24 h.
The cellulose fibers (Ultra Fiber 500 by Solomon Colors, Inc) were used as a bacteria carrier [27] (Table S1).The commercial bio-based additive consists of alkaliphilic spore-forming Bacilli species, nutrient salts, calcium lactate, and polylactic acid.The additive was added to the mix at a ratio of 2%.
The ternary mortar materials were composed of a mixture of cement, fly ash, calcined clay, sand, and chemical admixtures.The different modifications were comprised of the addition of the cellulose fibers, commercial product, and cellulose fiber-immobilized bacteria (Table S2).The final cell concentrations in each individual mortar sample are estimated to be 5 × 10 6 cells per cube and 8 × 10 6 cells per cylinder.The mortars were mixed based on Section 8 of ASTM C305-20.Details for the methods and experimental setup are specified in the Supplementary information (Figures S1,S2, Table S2).

Testing procedure
The

RESULTS
The mortar modifications were divided into two categories for the analyses: experiments with and without the bacterial modifications.
The control mortar is used as a reference for the first category, called the "Reference Group" and consists of the control (Cntl), commercial additive (Com), and the cellulose fiber addition (Cf) mortars.The second category is called the "Bacterial Group" and consists of the cellulose fiber addition (Cf), the vegetative B. subtilis additive (BS1), the sporulated B. subtilis additive (BS2), the live Syn.PCC6803 additive (PCC1), and the UV-killed Syn.PCC6803 additive (PCC2).

Physical and mechanical properties of the mortars
The compressive strength of the Reference Group mortars were not statistically different at 3 days.At 7 days, both Com and Cf mortars were statistically weaker than the Cntl.By day 28, the Cntl, Com, and Cf mortars achieved strengths of 87, 82, and 72 MPa, respectively.
The Cf mortar was statistically weaker than the Cntl and Com mortars (Figure 1A, Tables S3A,S4).At 3 days, the strength of the Cf, BS1, and PCC1 mortars in the Bacterial Group were statistically similar.The BS2 and PCC2 mortars were statistically weaker than the Cf mortar by 9.35% and 11.32%, respectively.There were no statistical differences at 7 days.At 28 days, Cf, BS1, PCC1, and PCC2 mortar compressive strengths were statistically similar.The BS2 mortar was statistically weaker than the PCC1 and PCC2 modifications by 13.70% and 12.41%, respectively (Figure 1B, Tables S3B,S5).There were no statistically significant differences between the split tensile strengths of the mortar cylinders in either the Reference or Bacterial groups (Figure S3, Tables S6-S8).
In the Reference Group, the Cntl and the Cf mortars absorbed statistically similar water volume.The Com mortar absorbed significantly more water than the Cntl and Cf mortars by 28.57% and 20.69% (Figure 2A, Tables S9A,S10).
In the Reference group, the Cntl and Cf mortars had statistically similar initial sorptivity at 4.4 and 3.1 × 10 −3 mm/s 1/2 but were both statistically lower than the Com mortar at 5.2 × 10 −3 mm/s 1/2 .The secondary sorptivity between the mortars in this group were all statistically different.The Cf mortar had the highest secondary sorptivity and the Cntl mortar had the lowest, respectively (Tables S12A,S13, Figure 3A).
In the Bacterial group, the mortars with bacteria additions all had statistically higher initial sorptivity than the Cf mortar.The only statistical similarities for initial sorptivity were between the BS1 and PCC1 mortar, and BS1 and PCC2 mortar.The BS1 and BS2 mortars had the highest initial sorptivity with 7.8 and 7.7 × 10 −3 mm/s 1/2 .The PCC1 and PCC2 mortar initial sorptivity were lower at 6.8 and 6.0 × 10 −3 mm/s 1/2 .All secondary sorptivity measurements were significantly different in this group, and bacterial modifications had lower secondary sorptivity than the Cf mortar.BS2 had the lowest secondary sorptivity of the bacterial mortars followed by PCC1, PCC2, and then BS1(Tables S12B,S14, Figure 3B).
In the Reference Group, total pore volume was highest in the Cf mortar at 0.16%, which also had the highest capillary pore volume at 0.12%.The Cntl and Com mortars had the same pore volume, and all mortars had the same gel pore volume.The critical pore diameter for the Cf mortar, 9.6 µm, was three orders of magnitude larger than the  S15A).

Mortar composition and proportional CaCO 3 content
The heat flow of the samples followed the same relative pattern but differed in magnitude at various peaks (Figures S4-S6, Table S16).Peak  1A).The overall carbonate mineral content was highest in the Com mortar, then the Cntl, then the Cf mortar.The Com mortar contains the highest CaCO 3 content, followed by the Cf and then the Cntl mortars (Table 1A).
In the Bacterial group, the temperatures at peak 5 for the BS1, BS2, PCC1, and PCC2 mortars were higher than the temperature of the Cf mortar peak.The heat absorbed by the BS1 mortar at peak 5 was higher than the other mortars in the Bacterial group, while the others precipitated comparable amounts (Table 1B).The BS1 mortar contains the  mortars (Table 1B).The PCC1 and PCC2 mortars contained the highest calcium carbonate content followed by the BS1, Cf, and BS2 mortars.

The relationship between mortar variables and variations between modifications
The Details of the specific PCA relationships and correlations for both Reference and Bacterial Groups are in the Supplementary information (Tables S17-S21).

Mortar performance is affected by changes in the pore structure
Changes to microstructure properties such as capillary pore intrusion, gel pore intrusion, and critical pore diameter strongly impact the strength and water absorption. [28,29]Pores can be classified as gel (<0.01 µm) and capillary (0.01-10 µm) and make up the total pore volume. [28]Large volumes of capillary pores and large critical pore diameters act as structural weak points under compressive stress. [29]gher gel pore volume is associated with the development of hydration products and shrinking of capillary pore volume, [28] increasing compressive strength. [30]Hydration products fill in voids to decrease the size of pores and decrease points of stress in the mortar.Secondary sorptivity is a measure of the diffusion of air out of porous materials and is an indirect criterium of the air content within the sample. [31] increase in air content reduces compressive strength. [32]Capil-lary pores are responsible for permeability transport processes, and large critical pore diameters are found in the capillary pore range. [33] increase in large capillary pores and large critical pore diameters are related to an increase in initial sorptivity, and linked to water absorption. [34]1.1 The commercial additive and cellulose fibers impact the water permeability and compressive strength by modifying cement hydration Hydration delay in the Com mortars increased water absorption while improved cement hydration and pore refinement in the Cf mortars decreased water permeability.The Com mortar had a higher initial sorptivity and water absorption despite having a higher CaCO 3 content.The permeability increase can be linked to the presence of yeast extract in the healing admixture.Yeast extract inhibits the cement microstructure development by absorbing to particle surfaces, preventing water contact, delaying the cement hydration reaction. [35]A cement hydration delay decreases the formation of hydration products, changing the pore structure, leading to an increase in permeability. [36]e MIP analysis (Figure 4A) shows a higher abundance of 0.01-10 µm capillary pores for the Com mortar compared to the Cntl mortar, increasing mortar permeability.Strength was not affected due to similar overall pore volume.There were no high correlations between CaCO 3 content and compressive strength, so there was insufficient precipitation to impact the strength properties.The lack of impact in compressive strength highlights the main purpose of the commercial additive as a crack healing agent, rather than a strength enhancer or permeability reducer. [23] I G U R E 3 Initial and secondary sorptivity for the Reference group (A) and Bacterial group (B) mortars at day 28.
The decrease in strength, initial sorptivity and water absorption for the Cf mortar is attributed to the drying and shrinking of watersaturated fibers.The additional water in the fibers prolonged the cement hydration reaction, reducing connectivity of the channels and reducing the permeability of the mortar. [37]The abundance of gel pores <0.003 µm and lower proportion of 0.01-0.3µm capillary pores of the Cf mortar compared to the Cntl and Com mortars are evidence of continued hydration.However, the addition of cellulose fibers produced capillary pores >0.01 µm due to the shrinking of swollen fibers.As the fibers dry, they shrink and increase the transition zone space between the fiber and mortar, creating local pore channels that function as structural weak points. [14,38]Furthermore, the fibers were observed to stay in clumps and did not sufficiently break down into individual fibers.The impacts of cellulose fibers should be more positive when they are well distributed in mortars.Thus, the shrinking of cellulose fiber clumps that were heterogeneously distributed in the matrix are responsible for the critical pore diameter peak at 9.6 µm in the MIP analysis (Figure 4A, Table S15A).A decrease in mortar cube compressive strength was also observed in a study on the addition of cellulose fibers to a sand-based mortar. [39]Thus, the presence of the fibers can partly explain the decrease in compressive strength.

Microbial interaction with the mortar structure varies between bacterial species
The concentration of bacteria used in this study matches previous studies [17] (Table S2).Since the bacteria were added at similar cell concentrations, the impact of the modifications on the pore structure reflects the specific characteristics of the selected strains and behavior in the mortar.
The viability of the bacterial cells was reduced due to insufficient protection from the mortar environment by cellulose fibers. [13]The optimal adsorption of cells to porous carriers occurs when the pore is 2-5 times the size of the microbe to accommodate the cells, which was not observed with the smooth fiber surface (Figures S7,S8). [40]nsequently, the cells were distributed in the mortar matrix.A F I G U R E 5 Principal component analysis of combined mechanical strength, permeability, and CaCO 3 precipitation properties, for the Reference group mortars (A) and the bacterial mortars (B).similar bio-mortar study used 2.5-5x the cellulose fiber volume content to immobilize B. subtilis 168 and observed visible carbonate formation at 5x fiber content but not 2.5x. [41]Therefore, the number of viable carbonate-forming cells in cellulose fibers decreases with decreasing bacteria content.
Notably, photoautotrophic bacterial additions demonstrated significantly stronger compressive strength and reduced permeability compared to the sporulated heterotrophic addition.The difference between the BS2 and PCC2 mortar is significant considering mortar treated with PCC2 has a higher porosity and capillary pore volume than those with BS2.However, all mortars with bacterial additions had higher initial sorptivity and lower secondary sorptivity than the Cf mortar, suggesting that the cells filled in air voids but increased pore channel connectivity.The pore-plugging of capillary pores by biomass is the main factor that influenced the differences between bacterial mortar properties, indicated by the lack of strong capillary pore relationships to expected mortar parameters in the PCA (Table S21, Figure 5).Differences between bacterial modifications are instead, a result of how either species interacts with the mortar 3D structure.The CaCO 3 mass percent had strong correlations to compressive strength, however, the range of CaCO 3 content between the bacterial mortars compared to the statistically similar strengths indicate that the CaCO 3 -strength relationship is not causative.Thus, it is concluded that there was insufficient CaCO 3 precipitated by the bacterial modifications to significantly affect the mortar properties.
A differential adhesion of spores to mortar pore channels may be responsible for the increased permeability and decreased compressive strength of the BS2 mortar.B. subtilis spores have different surface proteins from their vegetative states that may reduce adhesion to other surfaces. [42]Cell adhesion is considered an important factor for interactions with mortar surfaces and for efficient plugging of pores. [43]Decreased adhesion prevents efficient pore plugging through interfacial transition zones (ITZs), increasing pore connectivity and mortar permeability. [44]The reduced adhesion at the ITZs may reduce the bond strength between the materials, which reduces compressive strength. [45]Similar compressive strength between BS1 to BS2 suggest more pore-plugging differences between B. subtilis and

Syn. PCC6803.
The PCC2 mortar has the highest porosity yet performs better than the BS2 mortar.One explanation is additional pore-plugging by EPS produced because of exposure to UV. UV-radiation triggers the production of EPS by cyanobacteria as a UV-defense mechanism. [46]lleable substances in mortars like EPS can be deformed during high pressures of mercury intrusion and be detected as a pore space.Thus, the space taken up by EPS contributed to total porosity and prevented significant later age strength loss or increase in permeability. [35]re-plugging by biomass resulted in lower capillary pore volume and smaller critical pore diameter patterns in the BS1 mortars and PCC1 than the Cf mortar.However, pore-plugging by biomass did not improve 28-day mortar strength and permeability properties due to cell decomposition.Exposed cells contribute to strength by filling in pores at initial stages (Table S3B), decompose and leave behind pores at later ages that act as structural weak points [47] .Increased capillary pores volume <0.5 µm in BS1 and PCC1 mortars counteracted potential benefits of plugged pores >0.5 µm, resulting in similarities to the Cf mortar.BS1 and PCC1 MIP peaks <0.004 µm are interpreted as pores formed by decomposed cell fragments.MIP peaks <0.004 µm were not observed likely due to spore durability [42] and EPS protection (Figure 6). [9,10]In contrast, Singh and Gupta [41] detected a decrease in compressive strength up to 39.67% compared to the control due to the higher cellulose fiber volume fraction compared to this study.Homogenously distributed fiber additions with immobilized Bacillus species have been found to improve compressive strength in concrete. [48]Pore-plugging by sufficiently protected bacteria should also decrease water permeability. [49]Thus, in this case, the cells are insufficiently protected by the cellulose fibers as there is no strong impact on the strength and permeability properties of the mortar.
The photoautotrophic species perform better than sporulated heterotrophic species, concluding that the interactions of the cyanobacteria with the mortar surfaces have a more positive impact on mortar properties than the spores of heterotrophic cells.Considering that photoautotrophic species have the benefit of carbon sequestration for a more sustainable alternative to ammonia generation, their application can be preferable.

Microbial carbonate precipitation differs between photoautotrophic and heterotrophic species, vegetative and sporulated cell states
Differences in surface proteins between the bacterial modifications [42] affect their ability to facilitate the formation of minerals, which are reflected through variations in crystal properties. [50]CaCO 3 precipitation in biotic modifications can be attributed to MICP as evidenced through the CaCO 3 decarboxylation peak temperature (Table S16).
Microbial-induced minerals are more thermally stable than abiotic minerals, as observed in this study. [51]Biomass incorporated into the crystal structure affects anisotropy and heat flow and differed between modifications as seen in the heat absorption [54] (Table 1).
Microbial carbonates precipitated at early ages may have been formed via metabolic pathways that consume CO 2 : Syn.PCC6803 uses photosynthesis and vegetative B. subtilis produces carbonic anhydrase. [52]The main MICP mechanism at later mortar ages is likely the carbonation of Ca(OH) 2 on biological surfaces and varies in efficiency between species.Cell walls are known to accelerate Ca(OH) 2 carbonation and carbonate formation. [53]e varying CaCO 3 content between modifications supports differences in precipitation between species and cell-states. [22]The PCC1 and PCC2 mortars precipitated the most CaCO 3 , suggesting that the photoautotrophic bacteria surface is better at precipitating calcite than the heterotrophic surface.It can be concluded that photoautotrophic bacteria precipitate more carbonate than sporu-lated heterotrophic cells, skipping activation by crack ingress water. [13]us, differences in cell metabolisms and surfaces are sufficient to impact the crystal properties.The variations in microbial carbonate size and shape can impact the pore-plugging ability and should be investigated further when precipitated by photoautotrophs at various cell concentrations.Future studies should also investigate the impact of photoautotrophic MICP in bio-mortars while using carriers other than cellulose fibers as they have been observed here to be inefficient at protecting cells due to lack of porous void space. [40]e current study demonstrated that the photoautotrophic bacterial additions had higher strength and were less permeable than the sporulated heterotrophic bacterial modifications at the concentrations used.Variations in the biotic carbonate crystal properties depend on the species and can potentially affect the quality of pore plugging. [35]Carbonate precipitation by photoautotrophic species at higher concentrations can potentially improve mortar properties more than sporulated heterotrophic bacterial additions.The critical benefit of CO 2 sequestration without producing ammonia further emphasizes the advantage of photoautotrophs for bio-mortar applications.(Figure6).

CONCLUSION
The impacts of photoautotrophic species on mortar properties are comparable to those by vegetative heterotrophic species.In comparison to the sporulated heterotrophic modification, both photoautotrophic modifications had higher compressive strength by 13.70% and 12.41% and absorbed less water by 42.11%.All additives influence the pore structure, which affects cement hydration, strength, and permeability properties.Although the cellulose fiber shrinkage resulted in a critical pore diameter at 9.6 µm and decreased compressive strength by 18.9%, the additional water prolonged the cement hydration reaction.The commercial additive materials delayed the cement hydration, increasing the initial sorptivity and the water absorbed by 28.57%.
The findings showed that bacterial cell surfaces determine how the cells interact with the mortar, influencing pore size distribution, mechanical strengths, and permeability properties.These findings suggest that surface properties that reduce cell adhesion negatively impact strength and permeability.The bacterial production of EPS can increase pore space while contributing to strength and permeability properties.The PCC1 and PCC2 mortars precipitated more calcite compared to the BS1 and BS2 mortars.The more positive impact of photoautotrophic bacteria compared to sporulated heterotrophic cells demonstrates that photoautotrophic species perform better than spores at improving mortar properties.Thus, this study demonstrates that photoautotrophic bacteria are feasible environmentally-friendly, CO 2 -sequestering alternatives to sporulated heterotrophic bacteria.
primary critical pores of the Cntl and Com mortars.The Com MIP curve shows more capillary pores than the Cntl mortar in the 0.01-10 µm range.The Cf MIP curve shows more capillary pores >1 µm and less capillary pores in the 0.01-0.03µm range than the Cntl and Com mortars.There are more Cf mortar gel pores <0.004 µm than the Com mortar (Figure 4A, Table BS1 and PCC1 modifications had critical pore diameters <0.1 µm, at 8.0 and 7.3 × 10 −3 µm.Both BS1 and PCC1 also had secondary minor critical pore diameters at 1.1 µm and 4.0 µm.No secondary critical pore F I G U R E 1 Compressive strengths of the Reference group (A) and Bacterial group (B) mortar cubes.diameters were detected for Cf, BS2 and PCC2.The BS2 and PCC2 samples contained the largest critical pore diameters (Figure 4B, Table

5 ,
represents the decarboxylation of a calcium carbonate polymorph.Details for the DSC analysis, other peaks, and temperatures of reaction are in the SI.In the Reference group, the Cntl mortar absorbed the least heat at 801.6 • C, while the Cf mortar absorbed more heat at the same temperature.The Com mortar absorbed the most heat, 197.4 J g −1 , and at a higher temperature of 812.4 • C (Table highest carbonate content followed by the PCC2, PCC1, Cf, and BS2 F I G U R E 2 Total water absorbed by the day 28 Reference group (A) and Bacterial group (B) mortar cubes after 168 h of contact with water by one face of the cube.TA B L E 1 Temperatures of calcium carbonate decarboxylation during the DSC analysis, and carbonate quantities for from the XRD and TGA analysis for the Reference group (A) and Bacterial group (B) mortars.

F I G U R E 4
Mercury intrusion porosimetry results for the Reference group (A) and Bacterial group (B) mortars at day 28.The labeled peaks indicate critical pore diameters.

F I G U R E 6
Impact mechanisms of the (A) vegetative B. subtilis, (B) sporulated B. subtilis, (C) live Syn.PCC6803, and (D) UV-killed Syn.PCC6803 with the mortar.