J Ind Microbiol Biotechnol (2012) 39:567–577

DOI 10.1007/s10295-011-1037-1

BIOTECHNOLOGY METHODS

Diatomaceous earth as a protective vehicle for bacteria applied

for self-healing concrete
J. Y. Wang • N. De Belie • W. Verstraete

Received: 10 June 2011 / Accepted: 27 August 2011 / Published online: 17 September 2011
Ó Society for Industrial Microbiology 2011

Abstract Crack repair is crucial since cracks are the main The images showed that cracks with a width ranging from
cause for the decreased service life of concrete structures. 0.15 to 0.17 mm in the specimens containing DE immo-
An original and promising way to repair cracks is to pre- bilized bacteria were completely filled by the precipitation.
incorporate healing agents inside the concrete matrix to Scanning electron microscopy (SEM) and energy disper-
heal cracks the moment they appear. Thus, the concrete sive spectrometry (EDS) were used to characterize the
obtains self-healing properties. The goal of our research is precipitation around the crack wall, which was confirmed
to apply bacterially precipitated CaCO3 to heal cracks in to be calcium carbonate. The result from a capillary water
concrete since the microbial calcium carbonate is more absorption test showed that the specimens with DE
compatible with the concrete matrix and more environ- immobilized bacteria had the lowest water absorption (30%
mentally friendly relative to the normally used polymeric of the reference ones), which indicated that the precipita-
materials. Diatomaceous earth (DE) was used in this study tion inside the cracks increased the water penetration
to protect bacteria from the high-pH environment of con- resistance of the cracked specimens.
crete. The experimental results showed that DE had a very
good protective effect for bacteria. DE immobilized bac- Keywords Bacillus sphaericus  Ureolytic activity 
teria had much higher ureolytic activity (12–17 g/l urea Microbial CaCO3  Carrier  Crack repair
was decomposed within 3 days) than that of un-immobi-
lized bacteria (less than 1 g/l urea was decomposed within
the same time span) in cement slurry. The optimal con- Introduction
centration of DE for immobilization was 60% (w/v, weight
of DE/volume of bacterial suspension). Self-healing in Biomineralization is a ubiquitous complex phenomenon in
cracked specimens was visualized under light microscopy. natural environments. It integrates biologically induced
precipitation in local microenvironments created by
organisms under certain conditions that allow visible
J. Y. Wang (&)  N. De Belie extracellular chemical precipitation of mineral phases [10].
Magnel Laboratory for Concrete Research, Bacterially induced carbonate precipitation is also related
Faculty of Engineering and Architecture, to biomineralization. Some bacteria can produce or induce
Ghent University, Technologiepark Zwijnaarde 904,
bio-minerals during their growth and metabolism. Under
9052 Ghent, Belgium
e-mail: [email protected] suitable conditions, most bacteria are capable of inducing
carbonate precipitation [2, 11, 21, 23]. Nowadays, micro-
N. De Belie
e-mail: [email protected] bial carbonate precipitation is being widely investigated in
the field of civil engineering, such as for surface protection
J. Y. Wang  W. Verstraete [4], cementation of sand [27], and concrete crack repair
Laboratory of Microbial Ecology and Technology (LabMET),
[5, 19].
Faculty of Bioscience Engineering, Ghent University,
Coupure Links 653, 9000 Ghent, Belgium Concrete is the most widely used building material.
e-mail: [email protected] However, its inherent heterogeneous nature, low tensile

123
568 J Ind Microbiol Biotechnol (2012) 39:567–577

strength, and non-ideal service environments make it sus- squeeze the bacteria inside the pores [15]. In this work,
ceptible to cracking. In principle, there are two ways to diatomaceous earth (DE) was used as a carrier for the
repair cracks, i.e., a passive way and an active way. The bacteria.
passive way follows the procedure of detecting, monitor- DE is a natural soft siliceous sedimentation. It has a
ing, and repairing. The repair work will be performed after particle size ranging from less than 1 lm to more than
the cracks are detected. Healing agents are applied from the 1 mm, but is typically from 10 to 200 lm. DE consists of
outside and have to penetrate inside the cracks. It is a fossilized remains of diatoms. The diatoms skeletons are
method that operates from the outside to the inside. The highly porous, light in weight, and chemically stable and
active way is that the repair work will be done by the inert. DE has been mainly used as a filtration agent and
concrete itself when cracks appear, through a self-healing functional fillers for paints and plastics. Meanwhile, it can
process. Cracks will be healed by certain kinds of healing also be used in fireproof cement, insulation materials, and
agents released from the concrete when cracking happens. as an absorbent in explosives manufacture because of its
It is a method from the inside to the outside. Recently, self- great resistance to heat and chemical action [3].
healing concrete has become an important research topic Furthermore, DE can be used as a bacterial carrier.
because it holds the potential to be more economical and Immobilization of Pseudomonas species on DE resulted
convenient [24]. Self-healing properties in concrete may be in an increased efficiency of removing 3,5,6-trichloro-2-
obtained by secondary hydration of unhydrated cement pyridinol from industrial wastewater [9]. The oxidation rate
[12], encapsulation of polymers [7, 8], addition of expan- of ferrous ions by DE supported Thiobacillus ferrooxidans
sive agents [16], and so on. Another alternative self-healing was increased by about 20% (10–15 h shorter than that of
method is microbial carbonation precipitation. Use of free cells) and the bacterial specific growth rate also
microbial carbonation to repair concrete cracks was first increased 1.2 times from 0.077 to 0.09 h-1 [28]. DE is also
reported by Bang et al. [1]. They first immobilized bacteria used as a carrier of non-pathogenic microbes inside a
into polyurethane foam and then put the strips of PU foam system for in situ bioremediation of contaminated soil and
inside the simulated cracks on mortar specimens. Subse- ground water [22]. The porous cells of DE pellets can
quently, the specimens were immersed into a medium provide a home both for microbes and for oxygen, water,
containing urea and CaCl2. Compressive strength was and nutrients to help sustain the life of the augmented
increased about 10% in the cracked specimens with colonies of microbes introduced in the pellets.
PU-immobilized bacteria compared to the ones without The aim of this study was to use DE as a carrier or a
PU-immobilized bacteria. However, in this method, the protective vehicle for the bacteria, which were added into
active agents are applied from the outside. Hence, it cannot the high-pH concrete environment as a self-healing agent.
be considered as self-healing. The research on self-healing First, the experiments were performed in a mimicked
concrete by implementing bacterially precipitated carbon- concrete environment (a kind of cement slurry, made of
ation started in recent years [14, 20, 26]. cement and water, 20 g/l). In a second series of experi-
Bacillus sphaericus was used in our study, which is a ments, the DE-immobilized bacteria were incorporated into
ureolytic strain that is able to precipitate calcium carbonate mortar specimens, in which realistic cracks were made, to
(CaCO3) in its microenvironment by the decomposition of investigate the potential use of DE-immobilized bacteria in
urea (CO(NH2)2) into ammonium (NHþ the self-healing of cracks. Crack filling was visualized by
4 ) and carbonate
(CO2 ). The biogenic CaCO has a good potential to be means of a microscopy and the healing effect was evalu-
3 3
used to heal concrete cracks because it is natural, envi- ated by testing the capillary water absorption.
ronmentally friendly, and compatible with the concrete
matrix. For the aim of self-healing concrete cracks, the Materials and methods
bacteria should be added to concrete during the process of
mixing. When cracks appear, bacteria inside the concrete For each series of experiments, there were three replicates.
would be activated by water and oxygen penetrating inside
the cracks, and then precipitate CaCO3 to fill the cracks. Bacteria and growth conditions
However, bacterial cells could not be added to concrete
directly based on previous research. On the one hand, Bacillus sphaericus LMG 22557 (Belgian Co-ordinated
bacterial activity will be greatly decreased when they are Collections of Micro-organisms, Ghent) was used in the
added to an environment with pH higher than 12; on the study based on previous research [6]. The medium used to
other hand, bacterial cells might be destroyed during the grow B. sphaericus consisted of yeast extract and urea
process of hydration because the micropores become (VWR International, Belgium). The yeast extract medium
smaller and smaller during hydration, and this might was first autoclaved for 20 min at 120°C and then the

123
J Ind Microbiol Biotechnol (2012) 39:567–577 569

Fig. 1 Morphology of the DE

powders

sterilized urea solution was added, which was obtained by sterile urea medium, which consisted of urea and yeast
means of filtration through a sterile 0.22-lm Millipore filter extract. The medium was then put on the shaker (100 rpm)
(Millipore, USA). The final concentrations of yeast extract again for 3 days. As a control, 20 ml BS not mixed with
and urea in the growth medium were both 20 g/l. For all DE was added to the same urea medium. The initial con-
experiments, B. sphaericus cultures were obtained after centrations were 20 g/l or 40 g/l for urea and 1 g/l for yeast
two times subsequent culturing (1% inoculums) from a extract, respectively.
-80°C stock culture. Cultures were incubated at 28°C on a The urea media with different pH values were made as
shaker at 100 rpm for 24 h. Bacterial cells were harvested follows. The urea medium with a neutral pH was obtained
by centrifuging the 24-h old grown culture (5,000 g, 7 min) by adjusting the pH of the medium to 7.0 using a 1-M
and were re-suspended in a physiological solution (NaCl, NaOH solution. The urea medium with a moderate alka-
8.5 g/l). The concentration of bacterial cells in the sus- linity was obtained by autoclaving. A small amount of urea
pension was 109 cells/ml. (1 g/l) was decomposed into ammonia and carbonate ions
during the process of autoclaving, through which the pH of
Diatomaceous earth (DE) the medium increased to about 9.1.
To mimic a high-pH concrete environment, a cement
The DE used in the experiments was provided by the suspension was made by adding cement powder (Ordinary
company Dicalite Europe NV, Belgium. The specific sur- Portland Cement CEM I 52.5 N, 20 g/l) to the urea med-
face area of DE was 29 m2/g, determined by nitrogen gas ium. The cement suspension was first put on the shaker
adsorption (BELSPORP-Mini II). The morphology of DE (100 rpm) for 1 day to make sure that the cement powder
was observed by a scanning electron microscope (FEI reacted with water completely and the pH reached a stable
QUANTA 200F SEM), shown in Fig. 1. The DE particles value (about 12.5). After 1 day, the pre-mixture was
used in the experiments were raw materials without any transferred to the cement suspension. The mixture con-
treatment. They had different kinds of irregular shapes taining pre-mixture (DE-immobilized bacteria) and cement
(Fig. 1a). The particles size distribution was not homoge- suspension is referred to as cement slurry. The cement
nous, ranging from 4 to 20 lm (Fig. 1b). A large amount of slurry was also put on the same shaker for 3 days. The
tiny surface pores are present on the DE particles. The size detailed information about the experimental arrangement
of these pores ranged from 0.1 to 0.5 lm. Some particles can be seen in Table 1.
had hollow inner structures (Fig. 1c). The ureolytic activity of the bacteria was indicated by
the amount of urea decomposed by bacteria in the urea
Activity of immobilized bacteria under different media, which was determined by the total ammonium
pH conditions nitrogen (TAN) in the urea media. One mole of urea
(CO(NH2)2) produces 2 mol of NHþ 4 . The amount of NH4
þ

The ureolytic activity of the bacteria with and without can hence indicate the amount of urea decomposed. TAN
immobilization onto DE was examined in different kinds of concentrations were measured calorimetrically by the
pH environments. Bacterial suspension (BS, 109 cells/ml) method of Nessler [13]. The amount of urea decomposed
obtained from the same procedure as described in ‘‘Bac- after 1 day and 3 days was calculated based on the TAN
teria and growth conditions’’ was mixed with sterile DE values measured in the urea media.
powders in a 50-ml falcon tube (20 ml of BS was mixed
with 4 g DE in each falcon tube). Subsequently, the falcon Optimization of immobilization methodology
tube was put on a shaker at 100 rpm for 1 h. After 1 h, the
20-ml mixture (pre-mixture) containing DE immobilized Different concentrations of DE (w/v), 4 g DE ? 20 ml BS
bacteria in the falcon tube was transferred to 80 ml of (20%), 8 g DE ? 20 ml BS (40%), 10 g DE ? 20 ml BS

123
570 J Ind Microbiol Biotechnol (2012) 39:567–577

Table 1 Composition of urea media with different pH values

Series Pre-mixture (20 ml) Urea medium (80 ml) Initial pH of the
final mixture
DE (g) BS (ml) Urea (g) Yeast (g) Cement (g)

Free cells, U20, 7 0 20 2 0.1 0 7

Free cells, U40, 7 0 20 4 0.1 0 7
DE?BS, U20, 7 4 20 2 0.1 0 7
DE?BS, U40, 7 4 20 4 0.1 0 7
Free cells, U20, 9.1 0 20 2 0.1 0 9.1
Free cells, U40, 9.1 0 20 4 0.1 0 9.1
DE?BS, U20, 9.1 4 20 2 0.1 0 9.1
DE?BS, U40, 9.1 4 20 4 0.1 0 9.1
Free cells, U20, 12.5 0 20 2 0.1 2 12.5
Free cells, U40, 12.5 0 20 4 0.1 2 12.5
DE?BS, U20, 12.5 4 20 2 0.1 2 12.5
DE?BS, U40, 12.5 4 20 4 0.1 2 12.5

Fig. 2 Digital photos of the

mixture of DE and BS at
different concentrations of DE
(DE concentration in a, b, c and
d were 40, 50, 60, and 70%,
respectively)

(50%), 12 g DE ? 20 ml BS (60%), 14 g DE ? 20 ml BS DE-immobilized bacteria applied in mortar specimens

(70%), were used in the pre-mixtures to investigate the
protective effect of DE for bacteria in the high-pH cement Bacteria immobilized into DE were added to mortar
slurry. When the concentration of DE was higher than specimens during the process of mixing to examine their
50%, the workability of the pre-mixtures decreased sig- effect on self-healing of cracks. Two series of mortar
nificantly (Fig. 2). Therefore, they were not put on the specimens (40 9 40 9 160 mm) were made with a water-
shaker but were kept still for 1 h at the same temperature of to-cement ratio of 0.5 and a sand-to-cement ratio of 3. The
the shaker. After 1 h, the pre-mixtures were added to the components of the specimens are shown in Table 3.
cement suspensions, which were made by the same During the process of mixing, nutrients (yeast extract,
method as described in ‘‘Activity of immobilized bacteria urea, and Ca(NO3)2.4H2O) were firstly dissolved in the
under different pH conditions’’. The concentrations of water. The nutrients solution was mixed with cement, sand,
urea and yeast extract in cement slurry were 20 and 1 g/l, and DE. When making specimens with DE-immobilized
respectively. bacteria, the pre-mixture was first made by adding 45 g DE
to 225 ml BS. Then the pre-mixture was put on the shaker
Effect of nutrient (100 rpm, 28°C) for 1 h. Subsequently, the pre-mixture
containing DE-immobilized bacteria were mixed with
The influence of the nutrient (yeast extract) on bacterial cement, sand, and the nutrient solution. The BS used in the
ureolytic activity in the cement suspension was also specimens was the same as in ‘‘Activity of immobilized
investigated. The detailed experimental arrangement is bacteria under different pH conditions’’ and ‘‘Optimization
shown in Table 2. At a certain DE concentration (40, 50, of immobilization methodology’’.
and 60%), during the process of immobilization, 0 or 10 g/l In each series, six prisms were made. Reinforcements
yeast extract was used. The concentration of yeast extract were added to the mortar specimens to control the
in cement slurry was 1 or 10 g/l. There were three repli- crack width. To make specimens with reinforcements, a
cates in each series. The initial concentration of urea in 10-mm mortar layer was firstly added into the molds. After
cement slurry was 20 g/l. this layer was compacted by means of vibration, two

123
J Ind Microbiol Biotechnol (2012) 39:567–577 571

Table 2 Experimental
Series DE used Concentration of yeast Concentration of yeast Total concentration of
arrangement for investigation of
(w/v) [%] extract (g/l) in pre- extract (g/l) in cement yeast extract in cement
the effect of the nutrient
mixture (20 ml) suspension (80 ml) slurry (100 ml)

Y0, Y1.25 40 0 1.25 1

Y0, Y12.5 40 0 12.5 10
Y10, Y1.25 40 10 1.25 3
Y10, Y12.5 40 10 12.5 11
Y0, Y1.25 50 0 1.25 1
Y0, Y12.5 50 0 12.5 10
Y10, Y1.25 50 10 1.25 3
Y10, Y12.5 50 10 12.5 11
Y0, Y1.25 60 0 1.25 1
Y0, Y12.5 60 0 12.5 10
Y10, Y1.25 60 10 1.25 3
Y10, Y12.5 60 10 12.5 11

Table 3 Components of mortar specimens in each series crack width ranged from 0.15 to 0.17 mm. Afterwards,
Series Cement Sand Water DE BS Nutrients three of the cracked specimens were immersed into water
(g) (g) (g) (g) (ml) (g) and three were immersed into a deposition medium (made
of urea and Ca(NO3)2, 0.2 M) for 40 days.
DE 900 2,700 450 45 0 69.75
After 40 days, the specimens were taken out of the
DE?BS 900 2,700 225 45 225 69.75
water or the deposition medium and were rinsed gently
Nutrients included 2.25 g yeast extract, 22.5 g urea and 45 g with tap water to remove the particles that were not firmly
Ca(NO3)24H2O attached to the surface of the specimens. The specimens
were then stored at room temperature for 3 days to let the
reinforcement bars (D = 2 mm, L = 14 cm) were placed surfaces dry.
on top of it (Fig. 3a). Afterwards, the molds were com-
pletely filled with mortar and vibrated. All of the molds Visualization of crack filling
were put in an air-conditioned room with a temperature of
20°C and a relative humidity of more than 90% for 24 h. After the surfaces became dry, the specimens were sub-
After demolding, the mortar specimens were placed again jected to a light microscopy analysis. The microscope
in the air-conditioned room. (Leica S8AP0) was connected to a camera (Leica DFC295)
After 14 days, specimens were taken out of the air- to examine the crack filling. The cracks were viewed and
conditioned room and cracks were created by means of a photographed under a specific magnification.
crack width controlled three-point bending test, as shown
in Fig. 3b. Crack width was measured by means of a linear Capillary water absorption
variable differential transformer (LVDT) that was attached
at the bottom of the specimens. The crack width was A modified capillary water absorption test based on
increased with a velocity of 0.0005 mm/s until a crack of RILEM 25 PEM II-6 [18] was performed to measure
0.3 mm was obtained. After unloading, the remaining the water penetration resistance of the cracked mortar

Fig. 3 Mortar specimens and

the methodology for making a
realistic crack. a Mold and
reinforcement used for making
prisms. b Setup of three-point
bending test to create cracks in
the mortar prisms

123
572 J Ind Microbiol Biotechnol (2012) 39:567–577

specimens with or without bacteria. The obtained water 20 kV. Secondary electron imaging (SEI) was used for
absorption coefficient was used to compare the crack- electron micrography. Samples were completely dried in an
healing efficiency between the test series. The specimens oven at 40°C for 2 days and were gold coated by a Baltec
were dried at 40°C in an oven until weight changes came to SCD030 Sputter Coater before examination. An energy
less than 0.1% at 24-h intervals. Before the test, the dispersive spectrometer (EDAX, America) connected with
specimens were coated with a waterproof paint (SikaCorÒ SEM was used simultaneously to detect the components of
277) at four sides adjacent to the bottom surface where the the precipitation.
crack existed. There was only one crack on the bottom
surface. In order to protrude the effect of the crack filling
and decrease the influence of other surface flaws on the Results
capillary water absorption, part of the bottom surface was
also coated with the paint (Fig. 4). An area with Bacterial ureolytic activity at different pH conditions
40 9 20 mm around the crack was kept uncoated. The
coated specimens were weighed before their exposure to a As shown in Fig. 5, both free bacterial cells and immobi-
water bath (initial weight) and then were immersed to a lized cells showed a high ureolytic activity in neutral pH
depth of 5 ± 1 mm of tap water with the bottom surface and moderate alkaline environment, in which about 95%
facing downwards. The test was carried out in an atmo- (in 20 g/l urea media) and 90% (in 40 g/l urea media) of
sphere of 20°C and relative humidity of 60%. At regular urea was decomposed. There was no difference in ureolytic
time intervals (1, 2, 4, 8, 12, 24, 48, 72, 120, 240, and activity between un-immobilized and immobilized bacte-
360 h), the specimens were taken out of the water ria. However, the amount of urea decomposed by the free
and weighed after wiping the surface with a wet towel. bacterial cells in the high pH cement slurry was greatly
After the weighing, the specimens were immediately decreased, from 95% to less than 5%. The immobilized
put back into the water. The water absorption coefficient bacteria kept a much higher ureolytic activity than free
K (g cm-2 h-1/2) was determined by Eq. (1). bacterial cells. About 60% of the urea was decomposed by
Q pffi DE-immobilized bacteria in the cement slurry. The values
¼K t ð1Þ of decomposed urea measured on the third day were
S
slightly lower than the values after 1 day; this might be due
in which Q is the weight of water absorbed at different time to volatilization losses.
intervals (g); S is the uncoated area of the bottom facing to
the water (cm2); t is the time (h). K can be obtained from Influence of the DE concentration
the slope of a plot of water absorbed per cm2 against the
square root of time. It can be seen from Fig. 6 that upon increasing the amount
of DE from 20 to 70%, the amount of urea decomposed by
Characterization of precipitation bacteria in cement slurry with initial urea concentration of
20 g/l was increased from about 60% (12 g/l) to 85%
After the water-absorption test, the specimens with obvious (17 g/l). The highest bacterial ureolytic activity was
precipitation in cracks were broken along the original obtained by using 60 or 70% DE. There was an obvious
cracks (also by means of a three-point bending) and sam- increase in the amount of decomposed urea when the
ples were taken from the fracture surface to characterize amount of DE was increased from 20 to 50%: the amount
the precipitation along the crack. of decomposed urea was increased from about 60% (12 g/l)
The morphology of the precipitation was studied in a to 80% (16 g/l), but there was no significant difference
FEI QUANTA 200F SEM at accelerating voltage of between using 60% DE and 70% DE. A maximum of about
85% of the urea was decomposed in these two series. The
values of decomposed urea measured on the third day were
also slightly lower than the values on the first day due to
ammonia losses.

Influence of nutrient

The nutrient used (yeast extract) had a positive effect on the

bacterial ureolytic activity during immobilization. As shown
Fig. 4 Part of the bottom surface of the prism coated by the
waterproof paint (a crack was in the uncoated area, at the position of in Fig. 7, in the series with a concentration of DE equal to
the arrow) 40%, there was an increase in the amount of decomposed urea

123
J Ind Microbiol Biotechnol (2012) 39:567–577 573

Fig. 5 Ureolytic activity of pH 7 pH 9.1 pH 12.5

40
un-immobilized (free BS) and

Urea decomposed (g/L)

DE-immobilized bacteria
1d
(DE?BS) at different urea 30
3d
concentration and pH [codes
such as U20 refer to urea
concentration (e.g., 20 g/l)] 20

10

0
free BS, free BS, DE + BS, DE + BS, free BS, free BS, DE + BS, DE + BS, free BS, free BS, DE + BS, DE + BS,
U 20 U 40 U 20 U 40 U 20 U 40 U 20 U 40 U 20 U 40 U 20 U 40

20
filled (Fig. 8d). Similar to Fig. 8b, there were also some
Urea decomposed (g/L)

white deposits at the surface, but the amounts were much

15 higher in Fig. 8d. Some pores in the surface were entirely
filled by the white precipitation.
10
Capillary water absorption
5
The precipitation in cracks profoundly decreased the water
0 absorption of the cracked specimens, especially in the first
20% 40% 50% 60% 70%
24 h. It can be seen from Fig. 9 that the speed of water
Fig. 6 Ureolytic activity of immobilized bacteria in high-pH cement absorption in the specimens without bacteria was much
slurry when using different concentrations of DE faster than in the ones with bacteria. The sequence was
DEBS M \ DEBS W \ DE M and DE W. After 24 h, the
in the cement slurries that contained more yeast extract (by speed of water absorption gradually decreased. Cracks of
comparing series 40%Y0Y1.25 and 40%Y0Y12.5, and the specimens with DE-immobilized bacteria immersed in
series 40%Y10Y1.25 and 40%Y10Y12.5). However, no the deposition medium were completely filled by the pre-
obvious difference in the amount of urea decomposed was cipitation and showed the lowest water absorption. The
observed between series 40%Y0Y1.25 and 40%Y10Y1.25, ones with partly filled cracks showed more water absorp-
40%Y0Y12.5 and 40%Y10Y12.5. In the series where the tion, but less water absorption than those without bacteria.
concentration of DE was 50 and 60%, more urea was Statistically, there was no significant difference among the
decomposed when using yeast extract during immobilization. specimens only with DE (without bacteria). Whether they
The overall trend in Fig. 7 is that the amount of decomposed were immersed in water or in the deposition medium, they
urea increased as the concentration of DE increased, which all had higher water absorption than the specimens that
confirmed the results in ‘‘Influence of the DE concentration’’. were added with DE-immobilized bacteria. The water
absorption in DEBS M and DEBS W was about one-third
Visualization of cracking filling and 50% of that in DE W, respectively.

The specimens in the same series showed similar crack- Morphology of the precipitation in cracks
filling states. Therefore, one representative image of each
series is shown in Fig. 8. There was almost no precipitation Precipitation from different series of specimens showed
formed in the cracks of the specimens with DE that were different morphology. Figure 10a provides the image of
immersed in water (Fig. 8a). It was noticed that small the precipitation in the crack of the specimen with
amounts of white crystals were formed at the surface and in bacteria, which was immersed in water. Small particles
the cracks of the specimens with DE that were immersed in with a size of about 5 lm agglomerated into flower-
the deposition medium (Fig. 8b), yet the specimens with shape grains (15–20 lm). These grains were attached to
DE-immobilized bacteria had a quite different appearance. the crack wall. The result of the element analysis from
Cracks in the ones immersed in water were partly filled by EDS confirmed that these particles were calcium car-
the precipitation (Fig. 8c). Much more white precipitation bonate. Similarly, a large amount of particles was found
formed in cracks of the specimens immersed in deposition in the crack of the specimen with bacteria which had
medium. For those specimens, the cracks were completely been immersed in the deposition medium. The particles

123
574 J Ind Microbiol Biotechnol (2012) 39:567–577

DE 40% DE 50% DE 60%

20

18

Urea decomposed (g/L)

16

14

12

10
Y0, Y0, Y10, Y10, Y0, Y0, Y10, Y10, Y0, Y0, Y10, Y10,
Y1.25 Y12.5 Y1.25 Y12.5 Y1.25 Y12.5 Y1.25 Y12.5 Y1.25 Y12.5 Y1.25 Y12.5

Fig. 7 Ureolytic activity of immobilized bacteria (with different concentrations of DE) under different concentrations of yeast extract

Fig. 8 Light microscopy images of the cracks in different specimens (a, b: specimen only with DE, immersed in water, and in the deposition
medium, respectively; c, d: specimens with DE-immobilized bacteria, immersed in water, and in the deposition medium, respectively)

had irregular round shapes and most of them had a size of urea or Ca(NO3)2 along the crack wall. This needs to be
5–10 lm. A few grains had a larger size of 20–25 lm. further explored.
These particles were also calcium carbonate based on the
result of EDS. Except for irregular round particles, some
kind of long needle-like materials were also observed in Discussion
the cracks. They were distributed around the calcium
carbonate particles (indicated by the black arrow in In this study, it was shown that DE powders had a profound
Fig. 10b). It could be that they were crystals of un-reacted protective effect on the bacteria, particularly in the high-pH

123
J Ind Microbiol Biotechnol (2012) 39:567–577 575

1 50%Y0Y12.5 and 60%Y0Y12.5 (10 g/l). In this case, the

DE,W
nutrient had a more obvious impact on the process of
DE,M
Water absorped (g/cm2 )

0.8 immobilization than in the cement suspension. The pre-

DEBS,W
DEBS,M mixture made of DE and BS was a kind of easily flowable
0.6 suspension when using low concentrations of DE and the
nutrient was distributed homogeneously in the pre-mixture.
0.4
At higher DE concentrations (60 and 70%), the pre-mixture
became un-flowable paste (Fig. 2c, d). The nutrient was
0.2
incorporated inside the paste together with the bacterial
cells. Therefore, the bacteria could get more nutrients
0
0 100 200 300 400 around them from the microenvironment in the un-flowable
Time (h) paste than in the flowable suspension. After the pre-mix-
tures were added to the cement suspensions, the amount of
Fig. 9 Capillary water absorption of the specimens (in different
series) as a function of time
nutrients in the cement suspension had less of an impact on
the bacteria in the paste than in the flowable suspension,
where the bacteria did not get enough nutrients from the
cement slurry, which was made to mimic the really high microenvironment during the process of immobilization.
pH environment inside concrete. The more DE was used, In view of the crack filling in different kinds of speci-
the higher the protective effect. DE particles have a porous mens, it can be concluded that there was a self-healing
structure. Most of the pores are at the nano-scale. That is behavior in the specimens with DE-immobilized bacteria.
the reason why DE has such a high specific surface area. The healing was more effective when the specimens were
Only some particles have hollow inner structures (Fig. 1c), immersed in the medium with urea and Ca2?. Bacteria can
which can shelter the bacteria that may be absorbed inside. not only decompose urea from the crack wall of the
Most of the pores are only 0.1–0.5 lm, but the size of the specimens (incorporated during casting), but also use the
bacteria is about 1–2 lm. Therefore, bacterial cells were urea and Ca2? from the deposition medium to produce
mainly sorbed on the surface of the particles. As shown in more calcium carbonate. That is why more precipitation
Fig. 5, free bacterial cells had a high ureolytic activity in formed and completely filled the cracks in the specimens
neutral and moderate alkaline environment. Yet, this (with bacteria) immersed in the medium. Based on the
activity decreased rapidly in the high-pH cement slurry, SEM images and EDS results of the precipitation, it can be
which means that the unprotected bacteria can only show seen that the precipitation in the partly filled cracks (in the
ureolytic activity in the environments of moderate pH. specimens with bacteria that were immersed in water) was
Therefore, the possible reason why DE immobilized bac- calcium carbonate, but the precipitation in the completely
teria could still keep a certain degree of ureolytic activity in filled crack contained both calcium carbonate and some
the extremely high environment is that DE particles pro- other kind of material, which had a needle shape. Since the
vided a kind of microenvironment for bacteria, in which only difference between the specimens DEBS W and
the local pH around the bacteria was not as high as that in DEBS M was the immersion media, the extra long needle
the cement slurry [25]. A higher ureolytic activity (indi- particles were mainly due to the re-crystallization of
cated by more urea decomposed in cement slurry) was un-reacted urea or Ca(NO3)2 from the deposition medium.
obtained when using a higher amount of DE. This is due to From the result of capillary water absorption, it can be seen
the fact that the same amount of bacteria was surrounded that the specimens with completely filled cracks had the
by more DE particles, which could form a thicker protec- highest water penetration resistance due to the precipitation
tive layer for the microenvironment and the bacteria. inside the cracks.
The nutrient (yeast extract) had a positive effect on It can be seen from Figs. 5 and 6 that the amount of urea
bacterial activity during the stage of immobilization and in decomposed at the first day was almost the same as that at
cement suspension. The more nutrient used, the more urea the third day, which indicated that bacterial ureolytic
was decomposed. At lower concentrations of DE (20 and activity under this extremely high pH might not last beyond
40%), the nutrient in the cement suspension had more of an 1 day. However, the DE-immobilized bacteria showed a
effect on bacterial ureolytic activity than in the pre-mixture different behavior after being added into real cement
during immobilization. The situation was different when specimens. They still hold ureolytic activity after 2 weeks
using a higher concentration of DE. It can be seen that because the cracks, made after 2 weeks, were completely
more urea was decomposed in series 50%Y10Y1.25 and or partly filled by the precipitation produced by the bac-
60%Y10Y1.25, though the total concentration of yeast teria. The reason could be due to the difference between the
extract of these series (3 g/l) was lower than that in series cement slurry and real cement specimens although they

123
576 J Ind Microbiol Biotechnol (2012) 39:567–577

Fig. 10 Morphology of the precipitation in cracks of the specimens specimen with DE-immobilized bacteria immersed in water. b Pre-
with DE-immobilized bacteria (the black arrow in b indicated the cipitation in the crack of the specimen with DE-immobilized bacteria
long needle-like material). a Precipitation in the crack of the immersed in the deposition medium

almost had the same pH. The cement slurry was kept being be increased by 20% because of these supplements. How-
stirred on a shaker (100 rpm, 28°C) during the whole ever, in terms of the added value generated by the putative
period of testing, which means more severe exposure to self-healing process, it is still very promising for the
bacteria in the slurry. In mortar specimens, the mixing only practical use since the costs for the later on maintenance
lasts 2 min before casting. After that, there was no con- would be completely saved or at least greatly decreased.
tinuous mixing in the specimens since they started to Further research needs to be done on how to supply
solidify. The DE-immobilized bacteria might transfer to more urea and Ca2? from the inside of the concrete matrix
the dormant condition under the environment with high pH for the bacteria to precipitate enough calcium carbonate to
and no oxygen. It has been reported in US Patent 3898132 completely fill the cracks. The supply should come from
[17] that DE was used to obtain reversible dormancy of the inside (for example by only immersing into water or
active microorganisms by mixing a high amount of DE sprayed with water) because this constitutes real self-
with microorganisms in a closed environment. After healing.
cracking occurred in the specimens, the bacteria around the
crack were activated by the oxygen and water and then
showed the activity to precipitate CaCO3. Conclusions
It is noted that although the optimal concentration of DE
for immobilizing bacteria was 60%, practical difficulties Diatomaceous earth (DE) was found to have a protective
were encountered during the casting of the specimens. DE effect for the bacteria in a high-pH cement environment.
powders have a strong capacity to absorb water because of The possible mechanism is that DE particles have a strong
the high specific surface area. If the addition of DE is more capacity to sorb bacterial cells on the surfaces due to their
than 5% of the cement by mass, the mortar paste would high specific surface area. After sorption, DE provided a
become very dry and the workability decreases a lot. For kind of microenvironment around the bacteria, in which the
the latter application, if a larger amount of DE is needed to local pH was less aggressive than that in the whole
improve the protective effect and obtain more precipita- cementitious environment and thus bacteria could still
tion, some amount of sands could be considered to be decompose urea. The more DE that was used, the more
replaced by DE at the prerequisite that the strength of the urea that was decomposed, which indicated a higher ure-
specimens will not be decreased. Overall, taking the opti- olytic activity. The optimal concentration of DE for
mal combination of DE, BS, nutrients and mortar speci- immobilization was 60% (w/v). Cracks with a width of
mens, one comes to the equivalent weight of 11 kg DE 0.15–0.17 mm in mortar specimens were partly or com-
empowered with 6 g bacterial dry mass, 5.5 kg urea, 11 kg pletely filled by the aid of DE-immobilized bacteria
Ca(NO3)2, and 0.55 kg yeast extract, i.e., a total of about depending on the immersion media. The precipitation in
28 kg additives per 1,000 kg of plain mortar (made of the cracks was mainly composed of calcium carbonate
cement, sand, and water). The costs of the concrete could with a small amount of excessive urea or Ca(NO3)2 crystals

123
J Ind Microbiol Biotechnol (2012) 39:567–577 577

(if the specimens were immersed in the deposition med- 13. Ivanov VM, Figurovskaya VN, Barbalat Yu A, Ershova NI
ium). The capillary water absorption in the specimens with (2005) Chromaticity characteristics of NH2Hg2I3 and I2: molec-
ular Iodine as a test form alternative to Nessler’s reagent. J Anal
bacteria was about 50% (cracks were partly filled) or 70% Chem 60(7):707–710
(cracks were completely filled) lower than in the specimens 14. Jonkers HM, Schlangen E (2007) Self-healing of cracked con-
without bacteria. crete: a bacterial approach. In: Proceedings of FRACOS6: frac-
ture mechanics of concrete and concrete structures. Catania, Italy,
Acknowledgments The authors appreciate the financial support pp 1821–1826
from the Research Foundation Flanders (FWO-Vlaanderen) for this 15. Jonkers HM, Thijssen A (2010) Bacteria mediated remediation of
study (Project No. G.0157.08). The authors express their thanks to the concrete structures. In: Proceedings of the second international
Department of Inorganic Chemistry and the Department of Physics symposium on service life design for infrastructures. Delft, The
for providing BET and SEM analysis. Netherlands, pp 833–840
16. Keisuke Y, Akira H, Toshiharu K, Shinichirou N (2007) Crack
self-healing properties of expansive concrete with various
cements and admixtures. In: Proceedings of the first international
References conference on self-healing materials. Noordwijk aan Zee, The
Netherlands
1. Bang SS, Galinat JK, Ramakrisshnan V (2001) Calcite precipi- 17. La Verne AH (1975) Method of preparing stowable dormant
tation induced by polyurethane-immobilized Bacillus pasteurii. bacteria. United States Patent 3898132
Enzyme Microb Tech 28:404–409 18. NBN B 24-213 (1976) Belgische norm: proeven op metselste-
2. Castanier S, Le Metayer-Levrel G, Perthuisot J (1999) Carbon- nen—Wateropslorping onder vacuum
ates precipitation and limestone genesis—the microbiogeologist 19. Ramachandran SK, Ramakrishnan V, Bang SS (2001) Remedi-
point of view. Sediment Geol 126:9–23 ation of concrete using micro-organisms. ACI Mater J 98:3–9
3. Degirmenci N, Yilmaz A (2009) Use of diatomite as partial 20. Ramakrishnan V (2007) Performance characteristics of bacterial
replacement for Portland cement in cement mortars. Constr Build concrete: a smart biomaterial. In: Proceedings of the first inter-
Mater 23:284–288 national conference on recent advances in concrete technology.
4. De Muynck W, Cox K, De Belie N, Verstraete W (2008) Bac- Washington DC, USA, pp 67–78
terial carbonate precipitation as an alternative surface treatment 21. Rodriguez-Navarro C, Rodriguez-Gallego M, Ben Chekroun K,
for concrete. Constr Build Mater 22:93–103 Gonzalez-Munoz MT (2003) Conservation of ornamental stone
5. De Muynck W, De Belie N, Verstraete W (2010) Microbial by Myxococcus xanthus-induced carbonate biomineralization.
carbonate precipitation in construction materials: a review. Ecol Appl Environ Microbiol 69(4):2182–2193
Eng 36(2):118–136 22. Seth C Hunt (1996) Method and system for bioremediation of
6. Dick J, De Windt W, De Graef B, Saveyn H, Van Der Meeren P, contaminated soil using inoculated diatomaceous earth. United
De Belie N et al (2006) Bio-deposition of a calcium carbonate States patent 5570973
layer on degraded limestone by Bacillus species. Biodegradation 23. Stocks-Fischer S, Galinat JK, Bang SS (1999) Microbiological
17:357–367 precipitation of CaCO3. Soil Biol Biochem 126(1–4):25–34
7. Dry C (1994) Matrix cracking repair and filling using active and 24. Van Breugel K (2007) Is there a market for self-healing cement
passive modes for smart timed release of chemicals from fibers based materials? In: Proceedings of the first international con-
into cement matrices. Smart Mater Struct 3:118–123 ference on self-healing materials. Noordwijk aan Zee, The
8. Dry C, McMillan W (1996) Three-part methylmethacrylate Netherlands
adhesive system as an internal delivery system for smart 25. Vandamme EJ, De Baets S, Vanbaelen A, Joris K, De Wulf P
responsive concrete. Smart Mater Struct 5:297–300 (1998) Improved production of bacterial cellulose and its appli-
9. Feng Y, Racke KD, Bollag JM (1997) Use of immobilized cation potential. Polym Degrad Stab 59:93–99
bacteria to treat industrial wastewater containing a chlorinated 26. Wang JY, Van Tittelboom K, De Belie N, Verstraete W (2010)
pyridinol. Appl Microbiol Biotechnol 47:73–77 Potential of applying bacteria to heal cracks in concrete. In:
10. Hamilton WA (2003) Microbially influenced corrosion as a Proceedings of the second international conference on sustainable
model system for the study of metal microbe interactions: a construction materials and technologies. Ancona, Italy, pp 1807–
unifying electron transfer hypothesis. Biofouling 19(1):65–76 1818
11. Hammes F, Boon N, de Villiers J, Verstraete W, Siciliano SD 27. Whiffin VS (2004) Microbial CaCO3 precipitation for the pro-
(2003) Strain-specific ureolytic microbial carbonate precipitation. duction of biocement. School of Biological Sciences and Bio-
Appl Environ Microbiol 69(8):4901–4909 technology, Murdoch University, Perth
12. He H, Guo ZQ, Stroeven P, Stroeven M, Johannes Sluys L (2007) 28. Yoshishige K, Koichi S, Chihiro I, Tadashi C (1999) Enhance-
Self-healing capacity of concrete-computer simulation study of ment of the specific growth rate of Thiobacillus ferrooxidans by
unhydrated cement structure. Image Anal Stereol 26:137–143 diatomaceous earth. J Biosci Bioeng 88(4):374–379

123