1 Article

2 Introduction

With over 1010 m3 per year, concrete is the most widely used construction material [1]. A main component of concrete is cement which is produced from limestone [2]. During this process, high temperatures of 1450 °C [3] are used and the energy required for this process accounts for about 2.6% of the global energy demand [4]. Microbiologically induced calcium carbonate precipitation (MICP) is a process in which calcium carbonate is produced by microbiological activity. If the calcium carbonate is produced in the cavities of mineral particles it can form bridges between particles and solidify them. With 20–50 °C the optimal temperatures for MICP are lower than the temperatures for producing cement. Because of that, MICP has the potential to produce construction material that has a lower energy demand than concrete. Previous studies showed that MICP can improve the strength of construction material like sandstone and cement mortar [5, 6], to repair cracks in these materials [7, 8] and to produce construction material like concrete [9,10,11]. Furthermore, MICP can be used to improve the resistance of soils to earthquake-induced liquefaction [12, 13] and can be utilized for the mitigation of wind erosion of soil [14]. The ureolytic hydrolysation of urea is the most common used mechanism for MICP [15, 16] because it is easy to control [17]. During this mechanism one mol of ammonia is hydrolysed by urease (EC 3.5.1.5) [18] into two mol ammonia and one mol carbonic acid (Eqs. 1, 2). Both products hydrolyse which results in a pH increase and the formation of carbonate ions (Eqs. 3, 4). In the presence of soluble calcium ions, the carbonate ions are precipitated as calcium carbonate (Eq. 5)

$$\mathrm{CO}{\left({\mathrm{NH}}_{2}\right)}_{2}+{\mathrm{H}}_{2}\mathrm{O}\to {\mathrm{NH}}_{2}\mathrm{COOH}+{\mathrm{NH}}_{3}$$
(1)
$${\mathrm{NH}}_{2}\mathrm{COOH}+{\mathrm{H}}_{2}\mathrm{O}\to {\mathrm{H}}_{2}{\mathrm{CO}}_{3}+{\mathrm{NH}}_{3}$$
(2)
$$2{\mathrm{NH}}_{3}+2{\mathrm{H}}_{2}\mathrm{O}\leftrightarrow 2{\mathrm{NH}}^{4+}+2{\mathrm{OH}}^{-}$$
(3)
$$2{\mathrm{OH}}^{-}+{\mathrm{H}}_{2}{\mathrm{CO}}_{3}\leftrightarrow {\mathrm{CO}}_{3}^{2-}+2{\mathrm{H}}_{2}\mathrm{O}$$
(4)
$${\mathrm{Ca}}^{2+}+{\mathrm{CO}}_{3}^{2-}\to {\mathrm{CaCO}}_{3}$$
(5)

Besides the production of carbonate ions, the ureolytic microorganisms are discussed to have a further effect on the precipitation of carbonate. Positive charged calcium ions are attracted by negative charged carboxy- and phosphoryl-groups on the cell surface through electrostatic interactions [19,20,21]. The negatively charged cell walls act as nucleation sites for the formation of calcium carbonate crystals. For the consolidation of sand using ureolytic MICP, specimens are treated sequentially with cell suspension and a cementation solution containing urea and a calcium source. Previous studies have often used a solution with equimolar concentrations of urea (1 M) and a calcium salt (1 M) [22,23,24,25]. However, some studies have shown that MICP is more efficient for different ratios of urea and calcium [26,27,28,29]. These results suggest that there is an optimal ratio between urea and calcium ions which provides the highest consolidation of quartz sand for a given amount of cells in the specimen. The one factor at a time method is a time consuming process for the optimization of a system with multiple variables and often ignores the alternative effects between components based on the literature MICP is a process that is dependent on a lot of factors that can also influence each other. For example the concentration of urea and calcium in the cementation solution, the concentration of cells, urease activity, pH, temperature, and more [30]. During RSM the required number of experiments is reduced by proper experimental design. Furthermore it allows to investigate the interactions of the variables [31]. Therefore RSM can help to mitigate the disadvantages of the one factor at a time method for MICP. The optimization of MICP in terms of compressive strength was studied by Sotoudehfar et al. [29]. They described optimal concentrations of 3000 mM urea and 1500 mM calcium for MICP. This optimum was based on a study using the Taguchi method, in which three concentration levels were investigated. Due to the nature of the method, concentrations between these levels were not considered. Up to now, no study has investigated the efficiency of MICP in terms of compressive strength of consolidated sand using response surface methodology (RSM). For this reason, a central composite design (CCD) was used in this study to determine the optimum concentrations of urea and calcium chloride and volume of cell suspension that maximize the compressive strength of consolidated sand. The next section explains the experimental setup of the optimization process. In Sect. 3.1 the results of the CCD are shown and discussed in detail. In Sect. 3.2 a comparison of the optimized protocol and common literature protocols is given. Section 4 is a short summary and conclusion of the study.

3 Material and methods

3.1 Bacteria and cultivation procedure

Sporosarcina pasteurii (ATCC 11859) has been used as urease producing organism during this study. NH4-YE medium was used as culture medium according to ATCC recommendation. The medium contained 15.75 g TRIS-Buffer, 10 g ammonium sulfate and 20 g of yeast extract. For media preparation the TRIS-Buffer solution was adjusted to pH 9.2 and divided into two portions. Ammonium sulfate and yeast extract were then added in each part. The solutions were autoclaved separately at 121 °C for 20 min. After cooling, the parts were combined under aseptically conditions. For the preculture, 100 mL of NH4-YE medium was placed in 250 mL Erlenmeyer flasks and inoculated with 20 μL of a cryo-culture (15 v/v % glycerol) and incubated to an optical density at 600 nm (OD600) of 0.3 at 120 rpm at 30 °C (Multitron S-000115689, Infors HT, Swiss). For the main culture 200 mL of NH4-YE medium was placed in 500 mL Erlenmeyer flasks and inoculated with 3 v/v % of the preculture and incubated at 120 rpm at 30 °C until an OD600 of 1.6 was reached.

3.2 Determination of urease activity

For the determination of urease activity the conductivity method was used [9, 32]. 1 mL of cell suspension was added to 19 mL of a solution containing 1.053 M urea and 10 mM Tris buffer. The change in conductivity was recorded for 5 min with a conductivity probe (Qcond 2200, VWR International GmbH, Germany). To correlate the change in conductivity (mS/cm/min) with the degraded urea, urea standards (50–250 mM) were hydrolysed with urease from Jackbean (Carl Roth GmbH) and the change in conductivity was measured after complete degradation of urea.

3.3 Preparation of sand columns

The sand used in this study was quartz sand from a deposit in Haltern, Germany. It consists of 98% quartz and the particle size ranges from 125 to 500 μm (see Table 1).

Table 1 Particle size distribution of Quartz sand H33
Full size table

50 mL reaction tubes (Greiner centrifuge tubes, Sigma-Aldrich) were used as moulds. For this, they were cut off at the bottom and split in half. A hole (6 mm) was drilled into the cap to allow suspended effluent to exit the mould. Before packing the mould, the hole was covered with a sheet of cellulose fibres. 42 g of sand were packed loosely into the tubes which resulted in a height of the specimen of 50 mm and a diameter of 27 mm. The top part of the packed sand was also covered with a sheet of cellulose fibres to prevent a dislocation of the upper particles during treatment (see Fig. 1).

Fig. 1
figure 1

a packed mould for the MICP b components of the mould before assembly

Full size image

3.4 MICP treatment

For the MICP treatment the sand columns were injected sequentially with cell suspension and a cementation solution. Cell suspension of S. pasteurii was used directly after cultivation (OD600 = 1.6, 10.6 mM urea/min). Between each cycle of the treatment the suspension was stored at 7 °C to prevent further growth. The cementation solution contained various concentrations of urea and calcium chloride (see Table 2). Each solution had its pH adjusted to pH 7. For treatment, a defined volume of cell suspension was injected into the column by a peristaltic pump with a constant flow of 1 mL/min (see Fig. 2).

Table 2 Experimental range and levels of independent variables
Full size table
Fig. 2
figure 2

Experimental setup during MICP. A Storage of cell suspension/calcination solution, B Peristaltic pump, C Moulds for MICP as described in Fig. 1

Full size image

After a period of 30 min the cementation solution was injected into the column to achieve a total volume of cell suspension and cementation solution of 15 mL which corresponds to an injection pore volume of 1. The specimens were incubated for 24 h at 30 °C. This procedure was repeated twice for a total of three treatment cycles. After the third incubation period the specimen was oven dried at 60 °C for 72 h and, after cooling at room temperature for 24 h, the compressive strength was measured.

3.5 Experimental design

The software ‘Design Expert´ (Version 12.0.12.0, Stat-Ease, Minneapolis, USA) was used to determine a CCD. The variables utilized in the design were concentration of urea (A), concentration of calcium chloride (B) and the volume of cell suspension during treatment (C) (see Tables 2, 3). These variables were chosen based on a previous screening experiment (Data not shown). The design had a total of 20 experiments, containing 6 center points and 14 non-center points. Five levels of variables were considered for the design: negative star (− α), minimum (− 1), center (0), maximum (+ 1) and positive star (+ α). The compressive strength was considered as the main response.

Table 3 Design matrix of the CCD used in this study
Full size table

3.6 Determination of compressive strength

The compressive strength was measured by means of a concrete penetrometer (B19082, Form + Test Prüfsysteme, Germany) with a measuring range of 0–5000 kPa as suggested by Al-Thawadi [33]. The penetrometer was fixed with a tripod clamp and a lifting platform was positioned underneath, on which the specimen was measured. If there were any unevenness’s in the surface of the specimen, these were ground off with a file so that the measuring head of the penetrometer rested even on the specimen. For the measurement itself, the specimen was positioned so that the measuring head of the penetrometer rested centrally on the specimen. The lifting platform was then moved upwards consistently (200 kPa per 5 s). As soon as the specimen broke, the applied force was read on the penetrometer.

4 Results and discussion

4.1 Optimization of compressive strength

In this study an experimental design was conducted to optimize the compressive strength of quartz sand treated with MICP. Concentrations of urea (A), calcium chloride (B) and volume of cell suspension (C) were chosen as the variables for the CCD. The observed compressive strengths (Y) ranged from 0 to 3300 kPa (see Table 3).

For experiment number 10 and 16 the specimen could not be removed from the moulds without breakage due to no visible consolidation. Experiment number 14 and 18 could be removed from the moulds but showed no measurable consolidation. Further, these specimens showed signs of humidity after cooling to room temperature. All broken samples, as well as samples for which no measurable consolidation was achieved, have in common that they were treated with low amounts of cell suspension (< 5 mL) and high CaCl2 concentrations (> 1500 mM). These findings are probably due to unreacted calcium chloride. The poor solidification of the samples could be attributed to a limitation of urease activity due to the high calcium concentrations. Various studies have demonstrated an inhibitory effect of calcium salts on the activity of ureases from different sources. For example, Ferrer et al. describe an inhibition of calcium carbonate precipitation by Deleya Halophila above a calcium concentration of 10 mM [34]. Gorospe et al. observed inhibition of S. pasteurii urease activity for different calcium salts, including calcium chloride, for a concentration of 50 mM. [35]. However, while these studies only investigated the effect of a low concentration of calcium ions, Whiffin demonstrated a steady decrease in the urease activity of S. pasteurii by increasing the concentration of calcium nitrate [36]. She found that at a concentration of 2000 mM the urease activity of S. pasteurii stopped completely. By substituting a solution containing pure calcium nitrate to a mixture of calcium nitrate and calcium chloride (1:1) she found that the inhibition of urease activity was lower for this mixture than a solution of pure calcium nitrate. Therefore, it is not clear that the inhibition of urease activity was due to calcium or nitrate ions. In addition to a decrease in urease activity, the restriction of various other metabolic pathways of S. pasteurii could also result in the absence of calcium carbonate precipitation at high calcium concentrations.

The highest strength was calculated for CaCl2 concentrations of 1000 to 1600 mM and a cell suspension fraction of 5.5–8 mL (see Fig. 3, Supplement 1 and Supplement 2). These results are consistent with expectations that there must be an optimal ratio between the reactants available in the specimen and the cells. If cell concentrations are too low, limitation of calcium carbonate precipitation may occur due to cell encapsulation during the process of MICP. The calcium carbonate precipitaded in close proximity of the cell surface forms a layer around the cell which limitates mass transfer of urea into the cell. Once the cell is fully encapsulated, the cell dies and MICP comes to a halt [16]. Muynck and Verbeken investigated the influence of different concentrations of urea and CaCl2 on the improvement of strength parameters of sandstone [26]. They also describe that for a certain number of cells in the system, there must be an optimum level of urea and CaCl2, above which the nutrients are not further degraded and negative effects, such as the accumulation of salt in the sample, occur. For a fixed volume of cell suspension (C) of 7.5 mL calcium chloride concentration (B) and urea concentration (A) are more likely independent influence factors on the compressive strength (S1). The same seems to be the case for the factors volume of cell suspension (C) and urea concentration (A) at a constant calcium chloride concentration (C) of 1500 mM (Supplement 2). This is supported by the fact that the only statistically significant model term containing two independent factors for Eq. 6 is “AB” (see Table 4).

Fig. 3
figure 3

a contour plot and b 3D-Surface plot of the interactive effects of calcium chloride concentration (B) and cell suspension (OD 1.6, 10.6 mM urea/min) (C) on the compressive strength of consolidated sand at constant urea concentration of 1500 mM

Full size image
Table 4 Design analysis for the predictive equation by the Design expert software
Full size table

The regression analysis of the experimental data found that the response and the test variables are related using the following second order polynomial equation:

$$\sqrt{Y}=+48.45-0.1128A-6.04B+3.79C-0.503AB-0.0235AC+10.80BC-5.33{A}^{2}-6.23{B}^{2}-12.97{C}^{2}$$
(6)

R2 should be above 0.6 and the larger this value, the higher the statistical significance of the model. The R2 (0.9815) and adj. R2 (0.8269) of Eq. 6 indicate a high correlation between the observed and predicted values of compressive strength by Design Expert. The F-value of the model of 10.02 implies that the model is significant and that there is only a 0.18% chance that an F-value this large could occur due to noise (see Table 4). A p-value < 0.0500 indicates that the model term is significant for a 5% significance level. P-values > 0.1000 indicate the model term is not significant. The p-value of the model was 0.0018 and therefore found to be adequate for prediction. In this case BC, A2 and C2 are significant model terms (see Table 4). The mutual effects of the three variables urea concentration, calcium chloride concentration and volume of cell suspension were analysed by Design Expert software. The model predicted that the optimal urea concentration is 1492 mM, calcium chloride concentration is 1391 mM and the volume of cell suspension is 7.47 mL. For both concentrations, this optimum is above the values described in the literature for optimum calcium carbonate precipitation in terms of enzymatic activity and calcium carbonate precipitation rate [27, 28]. In order to validate the model created to optimize the compressive strength, 5–10 verification specimens must be examined (Jensen 2016). For this purpose, the cementation was performed according to the determined optimum and the results were used to validate the model by Design Expert. The PIMean (Prediction Interval Mean) approach is used as the method of validation. In this approach, the mean value of the validation runs is compared with the prediction interval in a 95% confidence interval [37]. A total of 6 experimental runs were conducted using the optimized parameters from the mathematical model. The obtained mean value of 1876.85 kPa lies within the 95% confidence interval (see Table 5). Thus, it can be assumed that the model is valid.

Table 5 Results of the validation runs at urea concentration 1492 mM, calcium chloride concentration 1391 mM and volume of cell suspension 7.47 mL
Full size table

These results indicate that calcium chloride is the limiting component in the cementation solution in terms of compressive strength and that for a given amounts of cells, an optimum amount of urea and calcium chloride exists for the MICP. Muynck et al. also describe that for a certain amount of cells an optimum urea and calcium dosage exists [16]. Some studies have investigated the MICP regarding the concentration of urea, calcium ions and the cell concentration. However, the target parameters considered varied, as MICP has a wide range of applications and the same target parameters are not considered optimal for every application (see Table 6).

Table 6 Overview of studies on the optimization of MICP by variation of the concentration of urea and CaCl2
Full size table

While studies optimizing MICP for the rate of calcium carbonate formation use a significant excess of urea [27, 28] optimization for the rate of MICP is not a useful parameter for sand consolidation because low concentrations of calcium-ions stoichiometrically result in low amounts of calcium carbonate (see Eq. 5). Various studies have shown that the strength of sand after MICP treatment correlates with the mass fraction of precipitated calcium carbonate [39, 40]. A high concentration of calcium-ions in cementation solution and therefore also a high resulting mass fraction of calcium carbonate is desirable. Studies that investigated MICP for a post MICP strength parameter achieved optimum results for nearly equimolar ratios of urea and CaCl2 [26, 38, 39]. Only Sotoudehfar et al. [29] found an optimum for a ratio of 2:1 urea:CaCl2. Since a comparability of these studies is difficult due to different experimental setups and bacterial strains, the protocol optimized by RSM was compared with two literature protocols in this study under the same treatment conditions to ensure a comparable comparison with literature values.

4.2 Comparison of three cementation solutions

The optimized and validated model was additionally compared with two protocols from the literature (see Table 7). Cementation solution 1 was chosen because it is utilized in many studies concerning MICP and solution 3 is based on the optimized protocol by Sotoudehfar et al. [29]. In addition to the absolute compressive strength, the efficiency of cementation in relation to the nutrients used was investigated. For this purpose, the quotient of the compressive strength and the total concentration of nutrients was formed.

Table 7 Overview of cementation solutions from this study in comparison to chosen references
Full size table

For this comparison, cementation was performed with 7.47 mL portion of cell suspension for all protocols in order to limit the comparison of the methods to the composition of the cementation solution. The lowest compressive strength (771 ± 149 kPa) and compressive strength per mol (0.39 ± 0.08 kPa/mM) was obtained with cementation solution 1 (see Fig. 4).

Fig. 4
figure 4

Comparison of three cementation protocols from this study and with literature values. Proportion of cell suspension 7.47 mL, reaction temperature: 30 °C, reaction time 24 h, measurement with penetrometer, n = 6

Full size image

The second highest compressive strength (1877 ± 240 kPa) was achieved with cementation solution 2, which was optimized in this work. Thus, by increasing the amount of urea and CaCl2 used by 49% and 39%, respectively, an increase in the compressive strength of cementation solution 1 to 2 of 144.1 ± 34.5% was achieved. The cementation solution 1, which is frequently used in the literature, is consequently not optimal for consolidating quartz sand under the conditions carried out in this study. The highest absolute compressive strength was achieved with solution 3 (2471 ± 368 kPa) according to Sotoudehfar et al. [29]. However, solution 2 obtained a higher compressive strength per mol (0.65 ± 0.08 kPa/mM) than solution 3 (0.55 ± 0.08 kPa/mM). due to the use of 101% more urea and 7.8% more CaCl2 which results in a 15.9% lower compressive strength per mol cementation solution. Although solution 3 achieved a higher absolute compressive strength the efficiency of the consolidation regarding the nutrients is lower. Therefore, the efficiency for MICP using solution 2 is higher. It should be stressed that while solution 3 achieved a high compressive strength the borders of the CCD used in this study were lower than the urea concentration used in solution 3. Thus, the model could not find an optimum concentration in the range of the concentrations used in solution 3. The use of the high urea concentrations in solution 3 should be considered critically. Due to the stoichiometric ratios in calcium carbonate formation (Eqs. 1–5), a maximum of 1500 mM CaCO3 could be formed at concentrations of 3000 mM urea and 1500 mM CaCl2. Excess urea could only be metabolized by S. pasteurii for growth. Regardless of whether the excess urea is completely degraded, or metabolized by S. pasteurii in the samples, additional use of urea is associated with additional costs and additional CO2 emissions during production, due to the energy costs in urea production [43]. However, if urea was to be completely degraded, this would cause a significantly higher concentration of ammonium ions in the samples, which is harmful due to its toxicity especially to water organisms. Any excess urea leads to the formation of extra ammonium, which can outgas as toxic ammonia under high pH conditions. By knowing the optimum ratio of urea to calcium ions, an excess of urea can be avoided. There are also various approaches in the literature for handling the ammonium formed. For example, the efflux after MICP in the Basarov reaction can be converted to urea after MICP in the Basarov reaction [44, 45] or the ammonium oxidized during wastewater treatment [45, 46]. Also, studies have been conducted to precipitate ammonium from MICP as struvite [47]. The use of zeolites [48] in MICP has also been discussed to solve this problem. Hence, a ratio of urea and CaCl2 close to equimolar is desirable, while obtaining a high compressive strength and to avoid the production of more excess ammonia than necessary. The optimized solution shown in this study is a helpful approach to identify the optimal concentrations in cementation solution for the consolidation of sand. This optimized protocol might help to improve the results of research investigating MICP for production of novel construction materials based on MICP. The results could also be of interest for other fields of application of MICP like the granulometric stabilization of soil for unsealed road construction [49] or the improvement of liquefaction resistance [12, 13]. For that, further parameter like stiffness, erosion resistance, permeability and shear strength after treatment with this optimized protocol need to be evaluated and compared to literature protocols.

5 Conclusion

In this study the response surface methodology was used to optimize the compressive strength of quartz sand treated with MICP. It could be shown that the concentrations of urea and calcium chloride as well as the volume of cell suspension have the greatest impact on consolidation of quartz sand. Finally, the model predicted that the optimal urea concentration is 1492 mM, calcium chloride concentration is 1391 mM and the volume of cell suspension is 7.47 ml (OD600 1.6, 10.6 mM urea/min) with a predicted compressive strength of 2.495 kPa. The predicted compressive strength could be verified (n = 6). Thus, it can be assumed that the model is valid. Additionally, the achieved results were compared to standard methods used in literature (Solution 1: 1000 mM Urea, 1000 mM CaCl2) as well to another optimized protocol (Solution 3: 3000 mM Urea, 1500 mM CaCl2). The compressive strength could be increased by 2.5 times compared to standard methods. Although solution 3 achieved a higher absolute compressive strength the efficiency of the consolidation regarding the nutrients is lower. Therefore, the efficiency for MICP using the method optimized in this study is higher and less excess ammonium is produced. Further research is necessary to optimize the conditions for MICP regarding compressive strength of consolidated sand.

Short summary:

  • Calcium concentration has negative impact on MICP

  • Amounts of Urea, Calcium and cell suspension have the highest impact on compressive strength of quartz sand

  • Predicted compressive strength could be verified

  • Compressive strength could be increased 2.5 times compared to standard methods described in literature using the optimal concentrations predicted by the model