Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

1 © Springer 2007

CRACK REPAIR BY CONCRETE-IMMOBILIZED BACTERIA

Henk M. Jonkers and Erik Schlangen

Delft University of Technology, Faculty of Civil Engineering and GeoSciences / Microlab, Stevinweg 1, 2628 CN Delft, The Netherlands

Tel: +31 (0)15 2782 313

Fax: +31 (0)15 2786 383

e-mail: h.m.jonkers@tudelft.nl

Cracks in concrete allow water and chemicals to enter, a process that may lead eventually to the unwanted corrosion of the steel reinforcement and the deterioration of the concrete structure. Within the framework of the Self Healing Materials Research Project of the Delft Center for Materials, the possible application of bacteria to extend the lifetime of concrete is studied. The goal of this project is to incorporate dormant but viable bacteria in the concrete matrix which will contribute to the concretes self-healing potential. Water entering freshly formed cracks will activate the dormant bacteria which in turn will seal these cracks through the process of metabolically mediated calcium carbonate precipitation. Concrete, however, is due to its high internal pH (>12), relative dryness and lack of nutrients needed for growth, a rather hostile environment for common bacteria. Yet, certain extremophilic bacteria may be able to endure this artificial environment. In this study we tested the applicability of alkaliphilic spore-forming bacteria of the genus Bacillus as self-healing agent in concrete. We found that incorporation of high numbers of bacteria (109 cm-3) as well as some suitable organic growth substrates in concrete did not negatively affect compressive- and flexural tensile strength. ESEM analysis revealed furthermore the self-healing potential of immobilized cells, as bacterial- but not control cement stone samples were found to deposit a new layer of calcium carbonate minerals on its surface. We therefore conclude that these specific bacteria are promising candidates to act as self-healing agent in concrete structures.

Keywords: Self healing, concrete, bacteria

1 Introduction The durability of steel reinforced concrete structures is substantially affected by cracking. Cracks in concrete occur due to various mechanisms such as shrinkage, freeze-thaw reactions and mechanical compressive- and tensile forces. Cracking of the concrete surface may enhance the deterioration of embedded steel rebars as ingression rate of corrosive chemicals such as water and chloride ions into the concrete structure is increased. An active and rapid crack-healing mechanism can therefore be expected to substantially decrease chemical ingression and thus rebar corrosion, resulting in a significant increase in lifetime of concrete structures. Such an autonomous repair or self-healing mechanism would also be beneficial for economical reasons as manual inspection and repair of large structures is costly. However, the self-healing mechanism, or the nature of the self-healing agent in concrete should be such that it would not negatively affect the original structural- and mechanical characteristics of the material. One possible repair agent could be non-reacted or not fully hydrated cement particles. Water ingress through cracks would result in secondary hydration reactions of these particles what could result in sealing-, or plugging, of the cracks.

Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

2 © Springer 2007

A disadvantage of application of cement particles is that their healing properties are limited to single events as the agent itself will be consumed in the process. Moreover, application of larger quantities to improve structures self-healing potential will negatively affect concrete characteristics as the initial structure may be weaker and more brittle. In this study we test the application of an alternative healing agent, i.e. bacteria which have the potential to seal cracks through the production and precipitation of calcium carbonate minerals. The advantage of application of bacteria over cement particles as healing agent is that the former only mediate the self-healing process and are thus not converted in the process themselves. This characteristic therefore potentially allows a virtually endless repetition of the self-healing process. The application of bacteria for the repair or maintenance of various materials is not new. In previous studies the potential of bacteria to clean concrete surfaces (DeGraef et al 2005), improve the strength of cement-sand mortar (Dick et al 2006; Ghosh et al 2005), repair of degraded limestone and ornament stone surfaces (Rodriguez-Navarro et al 2003) and crack repair on surfaces of concrete structures (Bang et al 2001; Ramachandran et al 2001) was investigated. One methodological limitation of these studies, however, was that bacteria with a relatively short lifetime and temporal activity (days rather than weeks or months) were used. For this reason the bacteria had to be manually applied on the surface of the target area and such an approach can therefore be considered as a natural, but not as a truly self-healing mechanism. In the present study we therefore investigated whether specialized bacteria characterized by a long-term viability can be integrated in the concrete matrix and act there as potent self-healing agent for autonomous crack repair. We tested for the application in concrete the suitability of some spore-forming alkali-resistant bacteria of the genus Bacillus as self-healing agent. Bacterial spores are specialized cells which can endure extreme mechanical- and chemical stresses and spores of this specific genus are known to remain viable for up to 200 years (Schlegel 1993). We hypothesize that dormant but viable bacterial spores immobilized in the concrete matrix will become metabolically active when revived by water entering freshly formed cracks. These cracks will subsequently be rapidly plugged and sealed through metabolically mediated microbial calcium carbonate precipitation, hampering further ingression of water and other chemicals. As revived bacteria also need a suitable substrate that can metabolically be converted to calcium carbonate, this also needs to be part of the concrete matrix. Thus, besides bacterial spores, additional substrates need to be tested for concrete compatibility, i.e. whether they will not negatively affect initial concrete mechanical and structural qualities.

2 Materials and methods

2.1 Bacterial strains and spore formation

Starter cultures of alkaliphilic (i.e. alkali-resistant) spore-forming bacteria were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ), Braunschweig, Germany. These cultures were initially cultivated according to the suppliers' recommendations in yeast extract based medium (medium DSMZ-2 for Sporosarcina pasteurii DSM 33 and DSMZ-31 for Bacillus cohnii DSM 6307; Bacillus halodurans DSM 497 and Bacillus pseudofirmus DSM 8715). Subsequently, growth and spore-forming potential (sporulation) was further tested in mineral medium amended with different organic carbon sources (6 g Na-citric acid or 5 g peptone plus 3 g yeast extract per liter).

Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

3 © Springer 2007

Mineral medium contained per liter of Milli-Q ultra pure water: 0.2g NH4Cl, 0.02g KH2PO4, 0.225g CaCl2, 0.2g KCl, 0.2g MgCl2.6H2O, 1 ml per liter trace elements solution SL12B, 0.1g yeast extract and 8.4g sodium bicarbonate. The pH of this medium was 9.2. Aerobic batch cultures were incubated in 2-l Erlenmeyer flasks on a shaker table at 150 rpm. Growth was monitored by microscopy and cell numbers and percentage of sporulating cells were quantified by microscopy using a Burger-Turk counting chamber.

2.2 Concrete compatibility of bacteria and organic substrates

A series of tests were performed in order to investigate whether the incorporation of bacteria or organic substrates (needed for bacterial calcium carbonate formation) in the concrete matrix do not negatively affect strength characteristics. Therefore concrete bars with and without (control) bacteria or organic substrates were prepared for flexural tensile- and compressive strength determination. For the preparation of bacterial concrete, a dense culture of S.pasteurii, grown in DSMZ-2 medium, was obtained and total cell number was quantified by microscopy using a Burger-Turk counting chamber. Cells were harvested after a double washing step by centrifugation (20 min x 10000g) and suspension of the cell pellet in tap water. The washed cells were finally suspended in a 20-ml aliquot of tap water which was used as part of the needed make up water for concrete bar preparation. The quantity of the tested organic compounds was 0.5% of cement weight. The individual organic compounds (Na-aspartate, Na-glutamate, Na-polyacrylate, Na-gluconate, Na-citric acid and Na-ascorbic acid) were firstly dissolved in the needed make up water, prior to concrete bar preparation. Bacterial-, organic compound- and control sets of concrete bars for flexural tensile- and compressive strength testing were prepared. Each set consisted of three replicate bars (dimensions 16 x 4 x 4 cm) made from ordinary portland cement (ENCI CEMI 32.5R), make up water and aggregates. Aggregate (gravel and sand) composition and quantities of used components are listed in Table 1. The concrete bars were uncased after an initial curing period of 24 hours and were subsequently further cured in tap water-filled separate plastic containers at room temperature. The sets of three replicate bars were tested for flexural tensile- and compressive strength after 28 days curing, following the procedure according to EN 196-1 Standard Norm. Table 1: Cement, water and aggregate composition needed for the production of 3 concrete bars of dimensions 16 x 4 x 4 cm used for flexural tensile- and compressive strength characterization of bacterial-, organic compound-, and control (no bacteria added) concrete Compound: Weight (g): Cement (ENCI CEMI 32.5) 390 Water 195 Organics (see text) 19.5 Aggregate (gravel and sand): 4 - 8 mm 562 2 - 4 mm 378 1 - 2 mm 283 0.5 - 1 mm 283 0.25 - 0.5 mm 243 0.125 - 0.25 mm 132

Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

2.3 Self-healing bacterial concrete

An experiment to test the hypothesis, that concrete immobilized bacterial spores are able to revive and start biomineral production in cracked- and water-entrained concrete, was set up. A dense culture of Bacillus pseudofirmus grown in mineral medium amended with 6 g/l Na-citrate was harvested after sporulation had occurred. The number of spores in the cell suspension, after a double washing step in tap water, was quantified by microscopic counting. Cement stone samples (with and without bacteria) were prepared using the cell suspension as part of the make up water for the bacterial samples. Spore density in the bacterial samples amounted to 109 cm-3. Ordinary portland cement (ENCI CEMI 32.5R) and a water-cement ratio of 0.5 was used for the preparation of cement stone samples (disks with dimensions of 4 cm diameter and 1 cm height), cast in plastic vials closed with a plastic lid. After 24 hours curing at room temperature, disks were further cured in tap water at room temperature. After ten days curing disks were removed from their plastic molds and chipped to pieces with a chisel. Chips of bacterial- and control samples were subsequently incubated in rich medium (mineral medium amended with 3 g yeast extract and 5 g peptone per liter) after pasteurizing for 30 min at 70ºC. Pasteurization of bacterial- and control cement stone chips before incubation was done to inactivate bacteria that potentially came into contact with the cement stone samples during the non-sterile handling of the cement stone samples and chips. As bacterial spores are not killed by the pasteurization procedure, this treatment ensured that potential differences between control- and bacterial samples after incubation were mediated by added bacteria and not by accidentally introduced contaminants. Individual chips were incubated aerobically in 100-ml medium aliquots on a shaker table at 100 rpm at 25ºC for 12 days. The chips were subsequently rinsed with tap water and stored wet in closed plastic vials until ESEM analysis what done within two days after sampling. For ESEM analysis, chips were mounted on a 1-cm2 metal support and kept in place with adhesive tape and observed with a Philips XL30 Series Environmental Scanning Electron Microscope without any further sample treatment.

3 Results

3.1 Cultivation of alkaliphilic spore-forming bacteria

The four selected species of alkali-resistant strains grew well in rich (yeast extract and peptone amended) medium. However, spore formation appeared much better in mineral medium with Na-citrate as growth substrate (Figure 1).

Figure 1: Spore formation by Bacillus cohnii grown in mineral medium with sodium citrate as growth substrate. Bright dots are spores formed by the vegetative cells (dark rods)

4 © Springer 2007

Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

3.2 Strength characteristics of bacterial-, organic compound- and control concrete

Concrete bars made for flexural tensile- and compressive strength testing contained 109 cm-3 S.pasteurii cells (bacterial concrete) or 19.5 g (0.5% of cement weight) organic carbon compounds. No additions were made to control concrete bars. After 28 days curing, no significant difference was found in flexural tensile- and compressive strength between control-, bacterial- and amino acid (aspartic acid and glutamic acid)-containing concrete bars (Table 2). Concrete to which polyacrylic acid and citric acid was added suffered significant strength loss, while gluconate- and ascobic acid amended concrete did not develop any strength during a 28 days curing period.

Table 2: Flexural tensile- and compressive strength characteristics of control-, bacteria-, and organic carbon amended concrete bars after a 28 days curing period Type of concrete: Tensile strength

(N/mm2): Compressive strength (N/mm2):

Control 7.78 ± 0.38 31.92 ± 1.98 S. pasteurii 7.45 ± 0.45 34.78 ± 1.52 Na-aspartate 7.33 ± 0.37 33.69 ± 1.89 Na-glutamate 7.16 ± 0.19 28.52 ± 3.56 Na-polyacrylate 6.42 ± 0.47 20.53 ± 4.50 Na-citrate 3.48 ± 1.72 12.68 ± 1.82 Na-gluconate 0 0 Na-ascorbate 0 0

3.3 Biomineral production by bacterial concrete

Cement stone samples with immobilized B.pseudofirmus spores produced copious amounts of mineral crystals on its medium-exposed surface as was revealed by ESEM analysis (Figure 2). No crystal formation was observed on control samples which were incubated under identical conditions in peptone- and yeast extract-containing medium.

Figure 2: Copious formation of minerals on the surface of cement stone with B. pseudofirmus-immobilized spores

5 © Springer 2007

Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

6 © Springer 2007

4 Discussion and conclusions The main objective of this study was to investigate whether bacteria can potentially act as a self-healing agent in concrete. The bacteria tested are known to be alkali-resistant, i.e. they grow in natural environments characterized by a relatively high pH (10-11). In addition, these strains can produce spores which are resting cells with sturdy cell walls that protect them against extreme environmental mechanical- and chemical stresses (Schlegel 1993). Therefore these specific bacteria may have the potential to resist the high internal concrete pH values (12-13 for portland cement-based concrete), and remain viable for a long time as well, as spore viability for up to 200 years is documented (Schlegel 1993). We hypothesized that concrete-immobilized spores of such bacteria may be able to seal cracks by biomineral formation after being revived by water and growth nutrients entering freshly formed cracks. The experimental data presented support our hypothesis, as cement stone samples with immobilized bacteria but not control samples precipitated minerals on surfaces exposed to growth medium. Although the exact nature of the produced minerals still needs to be clarified, they appear morphologically related to calcite precipitates. The mechanism of bacterially-mediated calcite production likely proceeds via organic carbon respiration with oxygen what results in carbonate ion production under alkaline conditions. The produced carbonate ions which can locally reach high concentrations at bacterially active 'hot spots' precipitate with excess calcium ions leaking out of the concrete matrix. This microbial calcium carbonate precipitation mechanism is well studied and occurs worldwide in natural systems such as oceans, biofilms, microbial mats and stromatolites (Krumbein et al 1977; Riding 2000; Arp et al 2001; Ludwig et al 2005). For an autonomous self-healing mechanism all needed reaction components, or self-healing agents, must be present in the material matrix to ensure minimal externally needed triggers. In the biomineral precipitation experiment, however, the organic substrate needed for bacterial carbonate ion production was still supplied as part of the incubation medium. As it should ideally also be part of the concrete matrix we tested the compatibility of bacteria as well as different organic components, which can act as bacterial growth substrates. The obtained flexural tensile- and compressive strength data indicate that the incorporation of high numbers of bacteria (109 cells cm-3) and the amino acids aspartate and glutamate (0.5% of cement weight) in the concrete matrix, do not result in a significant loss of strength after a 28 days curing period. However, the same experiment also revealed that apparently only specific organic components are suitable for incorporation, as some others resulted in dramatic strength loss. Another aspect that was not considered in this study but what is of major importance for the long term self-healing potential of bacterial concrete is the long term viability of concrete-immobilized bacterial spores. As most concrete structures are build to last for 50 years or more the viability of immobilized spores should keep up with that. One advantage of application of bacteria as self-healing agent is that a healing event not only revives bacterial cells but also potentially results in the production of fresh spores what resets the viability status. To conclude we can state that the application of bacteria as a self-healing agent in concrete appears promising. We have demonstrated that concrete-immobilized bacterial spores revive and produce copious minerals after stimulation by suitable medium, i.e. water containing an organic growth substrate. To further improve the autonomous bacterial self-healing mechanism current research focuses on long term viability and selection of best adapted bacterial species to the concrete environment as well as the incorporation of compatible biomineral-producing organic substrates to the concrete matrix.

Proceedings of the First International Conference on Self Healing Materials 18-20 April 2007, Noordwijk aan Zee, The Netherlands Henk M. Jonkers et al.

7 © Springer 2007

ACKNOWLEDGEMENTS

We would like to thank Arjan Thijssen for help with ESEM analysis and for providing ESEM photographs. Financial support from the Delft Center for Materials (DCMat: www.dcmat.tudelft.nl) for this study is gratefully acknowledged

REFERENCES

1. Arp G, Reimer A, Reitner J (2001) Photosynthesis-induced biofilm calcification and calcium concentrations in Phanerozoic oceans. Science 292:1701–1704

2. Bang SS, Galinat JK, Ramakrishnan V (2001) Calcite precipitation induced by polyurethane-immobilized Bacillus pasteurii. Enzyme and Microbial Technology 28:404-409

3. DeGraef B, DeWindt W, Dick J, Verstraete W, DeBelie N (2005) Cleaning of concrete fouled by lichens with the aid of Thiobacilli. Materials and Structures 38(284):875-882

4. Dick J, DeWindt W, DeGraef B, Saveyn H, VanderMeeren P, DeBelie N, Verstraete W (2006) Bio-deposition of a calcium carbonate layer on degraded limestone by Bacillus species. Biodegradation 17:357-367

5. Ghosh P, Mandal S, Chattopadhyay BD, Pal S (2005) Use of microorganism to improve the strength of cement mortar. Cement and Concrete Research 35(10):1980-1983

6. Krumbein WE, Cohen Y, Shilo M (1977) Solar Lake (Sinai). 4. Stromatolitic cyanobacterial mats. Limnol Oceanogr 22:635–656

7. Ludwig R, Al-Horani FA, DeBeer D, Jonkers HM (2005) Photosynthesis-controlled calcification in a hypersaline microbial mat. Limnol Oceanogr 50:1836-1843

8. Ramachandran SK, Ramakrishnan V, Bang SS (2001). Remediation of concrete using micro-organisms. ACI Materials Journal 98:3-9

9. Riding R (2000) Microbial carbonates: The geological record of calcified bacterial–algal mats and biofilms. Sedimentology 47:179–214

10. Rodriguez-Navarro C, Rodriguez-Gallego M, BenChekroun K, Gonzalez-Munoz MT (2003) Conservation of ornamental stone by Myxococcus xanthus-induced carbonate biomineralization. Appl Environm Microbiol 69:2182-2193

11. Schlegel HG (1993) General microbiology, 7th edition, Cambridge University Press.