15th International Congress on the Chemistry of Cement

Prague, Czech Republic, September 16–20, 2019

Revealing the impact of synthetic C-S-H onto cement hydration

Elisabeth John a, Dietmar Stephanb

Building Materials and Constructional Chemistry, TU-Berlin, Berlin, Germany

[email protected]
[email protected]

ABSTRACT
The development of green cements to reduce CO2 emissions often results in a reduced hydration
activity especially at early times, which collides with economic interests and process requirements.
Besides pozzolans like nano-silica, the performance of calcium silicate hydrate (C-S-H) nanoparticles
recently came to focus, due to their outstanding ability to accelerate the cement hydration without
compromising the long-term strength of the seeded cement (Brau 2012, Archieves of toxicology).
Many properties of C-S-H were found to influence their accelerating performance, the particle size was
unanimously identified as the most crucial parameter (Kanchanson 2015, ICCC). The impact of the
crystallographic phase of the C-S-H onto cement hydration is less intensively studied even though
already in 1969 Greening and Seligmann reported a significant influence of the crystalline C-S-H
phase afwillite onto the product formation in seeded C3S pastes (Seligmann 1969, ISCC). Regarding
the Ca/Si of C-S-H gel controversial results were published. While Alizadeh et al. found that the
hydration of C3S was accelerated more when C-S-H seeds richer in silicon where applied, Land et al.
observed that seeds richer in calcium were better accelerators. In both works the particle size,
respectively surface area was not kept constant (Alizadeh 2009, J. Mater. Chem., Land 2012,
Hipermat). Additionally, other properties like the water to silicon ratio or the crystallinity of the C-S-H
are not investigated so far.
The current work aims towards a systematic investigation of the influence of the C-S-H chemical and
physical properties, and towards an explanation of the observed effects.
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

1. INTRODUCTION

The so called C-S-H seeding describes a method to accelerate cement hydration by the addition of
synthetic calcium silicate hydrates that are assumed to act as nucleation seeds for the hydration
products in the paste (Nicoleau 2011; John et al. 2018). Compared to conventional chemical
accelerators C-S-H seeding offers the advantage of being non-toxic, non-corrosive and it does not
compromise the long-term strength of the seeded cementitious products (Brau et al. 2012; Bost et al.
2016). Various authors proved that synthetic calcium silicate hydrates show great potential as
accelerators, one day strength gains up to 270% were reported, and calorimetry experiments show that
the dormant period can be eliminated almost completely (Reichenbach-Klinke & Nicoleau 2011; Land &
Stephan 2018).

Nevertheless, the reported performance values obtained from calorimetry or compressive strength
testing vary over a large range (Table 1). While it is generally reported that a higher C-S-H dosage
enhances the acceleration Owens et al. reported a reduction in performance with more C-S-H, and Ling-
fei et al. could not accelerate the hydration any further by doubling the dosage. The published
performance values range from 9% with 4wt.-% C-S-H to a 270% increased one-day strength with only
0.3 wt.-% C-S-H. The C-S-H synthesis method plays a decisive role for the performance as accelerator.
It was found that the resulting particle size, agglomeration and total surface are important factors,
additionally the calcium to silicon ratio is known to alter the effectiveness of the additive (John & Stephan
2018a; Land & Stephan 2015; Kanchanason & Plank 2018). Some authors try to stabilize the synthetic
hydrates with polymer additives to ensure a controlled and small particle size, the PCE used
nevertheless expresses an impact onto hydration itself (Kanchanason & Plank 2015; Jiang Ling-fei
2015).

Table 1: One day strength gain of seeded pastes relative to unseeded pastes with varying
amount of C-S-H prepared by different synthesis methods, originally published in (John et al.
2018). SG – Sol-gel synthesis, CP – Coprecipitation

The currently used synthesis methods for C-S-H seeds include co-precipitation from sodium silicate and
calcium nitrate or chloride in aqueous solution, a sol-gel synthesis from tetra ethylene ortho silicate and
calcium ethylate in organic solvents or the synthesis from silicon dioxide and calcium hydroxide in water
(Land & Stephan 2015). The preparation of polymer/C-S-H composites is usually done by co-
precipitation in the presence of the polymer (Plank et al. 2018).

In this paper parameters altering the performance of C-S-H seeds were investigated systematically with
the aim to gain more information on the mechanisms of the acceleration process in correlation with the
properties of the calcium silicate hydrate. The results shall help to unify the varying results published to
date, and generate a better understanding of C-S-H seeding in general.
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

2. METHODS AND MATERIALS

2.1 Raw materials

The raw materials for the direct synthesis were calcium oxide (>92M.-%, VWR), a colloidal silicon dioxide
suspension with an average particle size of 7 nm and 30wt.-% solid content (CWK, Köstrosol 0730) and
deionized water. For sol-gel synthesis ethanol (99.5wt.-%, 1wt.-% MEK, VWR) was used as solvent,
tetra ethyl ortho silicate was supplied from Wacker (TEOS, Wacker Chemie) and metallic granular
calcium from Merck (98wt.-%). The cement used for compressive strength, vicat, ultra sound and
rheology experiments was supplied from CEMEX (CEM I 52.5 R), for calorimetry experiments a
CEM I 42.5 R (CEMEX) cement was used.

2.2 Analytical methods

CALORIMETRY data was obtained from externally mixed pastes containing 10g cement and 5g water
(w/s = 0.5) with 2wt.-% C-S-H (per cement mass). C-S-H was suspended as prepared in the water using
an ultrasonic rod. Data points were recorded every 10 s at 20°C (Isothermal heat flow calorimeter MC-
CAL100, C3 Analysentechnik). COMPRESSIVE STRENGTH tests were conducted at 20·20·20mm3
cubes (Type 2060, TONI Technik GmbH), the results were calculated as average from 6 measurements
each. The initial and final setting was determined according to EN 196-3 with an automatic VICAT testing
device (ToniSET, Toni Technik). The setting and hardening of the pastes were additionally monitored
by ULTRASONIC SOUND SPEED measurement (IP-8, Ultratest). The pastes for the compressive
strength, vicat and ultrasound measurements were prepared as described for calorimetry.
RHEOLOGICAL investigations were conducted with pastes containing 1wt.-% C-S-H, for the
experiment the basket geometry of the viscomat NT (Schleibinger Geräte) was used. The rotation speed
was raised from 0 to 150 rpm in 5 steps, 30 rpm each. Each step was kept at constant speed for 60s,
the 150 rpm plateau was kept constant for 180 s, after the maximum the speed was lowered in the same
manner stepwise to zero. Viscosity and shear rate were calculated with at least 20 constant values at
the end of each step. PARTICLE SIZE ANALYSIS was carried out in water (Mastersizer 2000, Malvern
Instruments), the C-S-H suspension was dispersed with an ultrasonic rod for one minute before
measuring. A reflective index of 1.68 for C-S-H and 1.33 for water was used for the calculations. Freeze
dried C-S-H samples were analyses by XRD (PANalytical Empyrian, PIXcel 1D Detector, Cu K α,
Panalytical) and 29SI-MAS-NMR (Avance 400, 400 MHz, Bruker).

2.3 Synthesis methods

For the direct synthesis the starting materials were mixed in the desired calcium to silicon (C/S) and
water to solid ratio (w/s). The suspension was transferred into plastic bottles and kept under constant
mixing in a shaker. The calcium ethylate precursor for sol-gel synthesis is obtained by reacting
elementary calcium with ethanol, the silicon precursor is then added to the stirred solution and
condensation is being induced by the addition of water. A gel forms in a time range of a few minutes to
a few hours, depending on the reaction conditions. The gel was stirred for 24 h. Residuals are removed
from the product by repeatedly washing the gel with ethanol and water. For the analytical methods the
C-S-H was freeze-dried. Hydrothermal reactions were carried out with pre-prepared C-S-H gel of the
desired C/S in stainless steel autoclaves. Tobermorite was synthesized from pre-prepared C-S-H with
a calcium to silicon ratio of 0.85 at 100°C for 4 weeks.

3. RESULTS AND DISCUSION

For the current investigation C-S-H was prepared by three different approaches, a direct reaction of
CaO, SiO2 and H2O with and without mechanical treatment and by sol-gel synthesis. The calcium silicate
hydrate prepared by direct reaction equals C-S-H I (ICDD 033-0306) as described by Taylor 1950, broad
reflexes were found at [16.7 (0.04), 24.8 (0.36), 29.2 (1.0), 32.0 (0.21), 49.7 (0.32), 55.2 (0.14) and 60.5
(0.03)] °2Θ (Cu Kα). The basal spacing was found to be reduced with the calcium content from 12.3 Å
(C/S=0.8) to 10.0 Å (C/S=1.2) and the mean chain length, indicated by the ratio of Q 1/Q2 species rises
with the silicon content, which is both in accordance with literature (Taylor 1950; Kunther et al. 2017).
SEM images show a bulky material without structural order. More details about the characterization of
the C-S-H can be found elsewhere (John & Stephan 2018b). The C-S-H prepared by sol-gel synthesis
is comprised of a platelet-like structure comparable to the one reported by Land et. al (Figure 1, left)
(Land & Stephan 2018). It is almost X-ray amorphous but shows a broad hump at approximately 27 °2Θ,
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

which indicates the presence of amorphous silica. The 29Si-MAS-NMR confirms this assumption, since
no peaks assignable to the typical C-S-H Q1 or Q2 usually found between -70 to -90 ppm (Yu et al. 1999)
were detected (Figure 1, right). Sol-gel C-S-H is reported to show a very good performance in
accelerating cement hydration, but since it does not resemble the natural C-S-H characteristics in
several points it will not be focus of further investigations in this paper.

SiO2

C-S-H (Sol-Gel)

1 µm 29Si chemical shift [ppm]

Figure 1: SEM image (left) and 29Si-MAS-NMR spectrum of C-S-H prepared by sol-gel synthesis
in comparison with the spectrum of amorphous silica (right).

Several factors were identified to influence the performance of C-S-H prepared by direct reaction,
parameters that can be tuned easily during synthesis include time and temperature. The contact time of
calcium oxide and silicon dioxide during synthesis was investigated in in-situ XRD experiments to
determine the time needed for a complete turnover. After 7 h the calcium hydroxide reflex completely
disappeared, and the C-S-H reflex area remained constant. Further aging of the hydrate for more than
7 h is not necessary for product formation but was expected to alter its agglomerate size. Figure 2 (left)
shows the result of the laser granulometry investigations done with different aged C-S-H, the mean
particle size of a freshly prepared C-S-H is enhanced from 2.3 µm to 3.7 and finally to 6.4 µm during
aging for approximately 200 d. For seeding applications, the particle size and agglomeration are crucial
factors, hence it was expected that aged calcium silicate hydrates will have less impact onto cement
hydration. In calorimetry the difference between cement pastes seeded with seven and fourteen-day
aged C-S-H is only minor (Figure 2 right), nevertheless the older C-S-H led to a higher heat evaluation
during the dormant period and the younger C-S-H to a higher hydration maximum.

Figure 2: Particle size of calcium silicate hydrates prepared with different synthesis durations
(left) and the influence onto the performance in calorimetry compared to a reference cement
(right).

C-S-H aged for different durations was also used to prepare small cubes for compressive strength
testing. Additionally, to 1, 7 and approximately 365-day old C-S-H, one batch of cubes was prepared
with a freshly mixed calcium oxide and silicon dioxide suspension to evaluate the effect of the starting
material mix. Figure 3 shows the results of compressive strength testing from 1 to 28 d of hydration. The
findings from calorimetry were confirmed. The 7-day aged C-S-H shows, expect for day one, a
significantly higher compressive strength than the 1-day old C-S-H. It appears that an optimum of aging
time exists since the oldest C-S-H does not show a better strength then the 7-day old C-S-H which might
be due to the formation of too large C-S-H aggregates that compromise the positive effect of aging.
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

Figure 3: Compressive strength of 20·20·20 mm³ cement cubes prepared with 3wt.-% C-S-H that
was aged for 1, 7, 365 days or directly applied as prepared (0 days).

A higher reaction temperature was expected to alter the performance comparably to a longer reaction
time, since the reaction should be faster at higher temperatures due to the enhanced solubility of the
silicate species. Figure 4 presents the results from isothermal heat flow calorimetry experiments with C-
S-H prepared with C/S = 1 at 20, 30 and 60°C with a reaction time of one week. As expected, the highest
temperature resulted in the strongest effect, but the dormant period was not changed significantly while
the maximum heat release was enhanced which was not the case for longer reaction times. The reason
why C-S-H prepared at elevated temperatures is a better accelerator is not clear yet, a possible
explanation is the formation of small quantities of tobermorite during reaction. The crystalline calcium
silicate hydrate, that has a calcium to silicon ratio of approximately 0.83 can be formed under relatively
mild hydrothermal conditions. Viehland et. al report the formation of 14Å tobermorite at 80°C (Viehland
et al. 1997). For the preparation of C-S-H at different reaction temperatures the C/S was 1, the formation
of a tobermorite-like C-S-H phase would hence lead to the release of calcium hydroxide into the
suspension, which is known to accelerate cement hydration (Ludwig & Neumann 2005). In 2015 Land
et al. presented the results of a systematic performance optimization of C-S-H seeds prepared by
mechanochemical synthesis. They found that seeds prepared with an initial C/S of 2 were most efficient
(Land & Stephan 2015). At this calcium to silicon ratio under the given reaction condition free calcium
hydroxide had to be present, too. To test whether this theory might be valid tobermorite was synthesized
hydrothermally and used in calorimetry experiments to see whether it alters cement hydration (Figure
4). It was confirmed that tobermorite does accelerate cement hydration significantly, and that with higher
quantities both dormant period and the maximum of acceleration period are shifted to earlier times and
higher values. In additional hydrothermal experiments with C-S-H of a higher C/S it was confirmed that
a significant amount of calcium hydroxide is formed over time although no crystalline C-S-H could be
detected by XRD at that time.

Figure 4: Isothermal heat flow calorimetry of cements accelerated with 1wt.-% C-S-H prepared
at 20, 30 and 60 °C (left) and 1 and 3wt.-% tobermorite (right).
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

The addition of synthetic calcium silicate hydrates was found to enhance the viscosity of cement pastes
drastically. With a water to cement ratio of 0.5 the plastic viscosity was more than doubled from an
average of 0.44 Pa∙s without C-S-H to more than 1.18 Pa∙s with 1 wt.-% C-S-H prepared by direct
reaction. The viscosity was raised to higher levels with low calcium C-S-H (~ 1.33 Pa∙s, C/ = 0.8 and 1)
than with C-S-H containing more calcium (~ 1.04 Pa∙s, C/S = 1.2 and 1.4). The stronger rheology altering
effect of low calcium C-S-H is assumed to be due to their higher specific surface area. Beaudoin et. al
pointed out the similarity to clay minerals, and although calcium silicate hydrates do not swell they can
incorporate small molecules into their interlayer spaces which might reduce the free water content of
the cement paste and increase the viscosity (John & Stephan 2018b; Beaudoin et al. 2008).

The compressive strength of a cement paste at a given time of hydration is primarily determined by the
clinker characteristics, mixing and curing conditions, and the water to cement ratio, whereas latter is
strongly correlated to the porosity of the hardened paste (Taylor 1997; Feldman & Beaudoin 1976).
Figure 5 shows results of compressive strength tests on pastes with different w/c (0.4 to 0.5) with and
without C-S-H added. As expected, the lower water content pastes develop a higher compressive
strength than the corresponding high water to cement ratio pastes. The addition of C-S-H does enhance
the one-day strength of all three samples but after 3 days the strength of the lowest w/c paste was not
improved and after 7 days the seeding resulted in a lower compressive strength for this paste compared
to the unseeded specimen. After 28 days an improvement can only be found for the paste with the
highest water content, no great difference was found for w/c = 0.45 and the strength was strongly
reduced for the paste with the lowest water content. The compressive strength of all three seeded
cement pastes is almost equal after 7 respectively 28 days. This effect can probably be explained by
the principle of C-S-H seeding, the synthetic hydrate is assumed to drive the product formation of cement
hydration away from the clinker grains towards the pore space which might compensate for the
enhanced porosity caused by the higher water content. It is reported that C-S-H seeds reduce the
capillary porosity and generally a denser and more homogenous microstructure was found (Thomas et
al. 2009; Jiang Ling-fei 2015). The dense microstructure might also be a reason for the decreased long-
term strength in low-water pastes since the degree of hydration is limited to the available mobile water
in the pore structure. Further experiments towards the changes in pore structure development induced
by C-S-H seeding are expected to give more support to this theory.

Figure 5: Compressive strength of cement with and without 2wt.-% C-S-H and with a water to
cement ratio ranging from 0.4 to 0.5.
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

4. SUMMARY

It was found that C-S-H prepared by sol-gel synthesis does not show the characteristics of neither
natural C-S-H formed in concrete nor synthetic C-S-H described in literature. XRD and solid-state NMR
confirm that the product of the sol-gel synthesis resembles amorphous silica.

Various synthesis and experimental conditions were identified to affect the performance of C-S-H seeds
in an unexpected manner. Although longer reaction times during direct reaction increase the
agglomerate size of the calcium silicate hydrates seeds up to a certain age, they were found to be more
effective accelerators than younger seeds. Additionally, an elevated reaction temperature does not
induce the same effect as an elongated reaction time. In compressive strength tests it was demonstrated
that the water to cement ratio drastically alters the performance of C-S-H and that even a strength
reduced by 12% can be recorded in low-water pastes with the same seeds that accelerate a high-water
paste by 25%.

Synthetic calcium silicate hydrate enhances the water demand of pastes as it was observed during paste
preparation and indicated by rheological measurements. Low-calcium C-S-H had a stronger impact onto
rheology which is explained by its higher specific surface area.

5. ACKNOWLEDGEMENT

This research was supported by the Deutsche Forschungsgemeinschaft, DFG, project number STE
1086/15-1. The authors would like to thank Dr. Christian Lehmann and Dr. Jan Epping for taking the
SEM images and recording NMR spectra. The compressive strength, vicat and ultra sonic sound tests
were carried out by Maxim Ghanadri during his work for his bachelor thesis.

6. REFERENCES

Beaudoin JJ, Dramé H, Raki L, Alizadeh R (2008) Formation and characterization of calcium silicate
hydrate–hexadecyltrimethylammonium nanostructure. J. Mater. Res. 23: 2804–2815.

Bost P, Regnier M, Horgnies M (2016) Comparison of the Accelerating Effect of various Additions on
the early Hydration of Portland Cement. Construction and Building Materials 113: 290–296.

Brau M, Ma-Hock L, Hesse C, Nicoleau L, Strauss V, Treumann S, Wiench K, Landsiedel R,

Wohlleben W (2012) Nanostructured Calcium Silicate Hydrate Seeds accelerate Concrete Hardening:
a combined Assessment of Benefits and Risks. Archives of toxicology 86: 1077–1087.

Feldman RF, Beaudoin JJ (1976) Microstructure and strength of hydrated cement. Cement and
Concrete Research 6: 389–400.

Jiang Ling-fei KX-m (2015) Influences of nano-sized C-S-H Particles on Cement Hydration of OPC in
the Presence of Fly Ash or Polycarboxylate Superplasticizer, ICCC.

John E, Matschei T, Stephan D (2018). Cement and Concrete Research 113: 74–85.

John E, Stephan D (Hrsg. 2018a) C-S-H Keime zur Beschleunigung der Zementhydratation.
Deutschland. F.A. Finger-Inst. für Baustoffkunde Bauhaus-Univ. Weimar, Weimar.

John E, Stephan D (Hrsg. 2018b) Properties of calcium silicate hydrate with varying stoichiometry.

Kanchanason V, Plank J (2015) C-S-H - PCE Nanocomposites for Enhancement of Early Strength of
Portland Cement, ICCC.

Kanchanason V, Plank J (2018) Effectiveness of a calcium silicate hydrate – Polycarboxylate ether (C-
S-H–PCE) nanocomposite on early strength development of fly ash cement. Construction and Building
Materials 169: 20–27.
 15th International Congress on the Chemistry of Cement
Prague, Czech Republic, September 16–20, 2019

Kunther W, Ferreiro S, Skibsted J (2017) Influence of the Ca/Si Ratio on the Compressive Strength of
Cementitious Calcium-Silicate-Hydrate Binders. J. Mater. Chem. A: 17401–17412.

Land G, Stephan D (2015) The Synthesis of C-S-H Seeds Methods, Variables and their Impact on the
Ability to accelerate Cement Hydration, ICCC.

Land G, Stephan D (2018) The effect of synthesis conditions on the efficiency of C-S-H seeds to
accelerate cement hydration. Cement and Concrete Composites 87: 73–78.

Ludwig H-M, Neumann T (2005) Hydraulisches Bindemittel.

Nicoleau L (2011) Accelerated growth of calcium silicate hydrates. Experiments and simulations.
Cement and Concrete Research 41: 1339–1348.

Plank J, Schönlein M, Kanchanason V (2018) Study on the early crystallization of calcium silicate
hydrate (C-S-H) in the presence of polycarboxylate superplasticizers. Journal of Organometallic
Chemistry.

Reichenbach-Klinke R, Nicoleau L (2011) Use of csh suspensions in well cementing. Google Patents.

Taylor HFW (1950) Hydrated Calcium Silicates Part I Compound Formation at ordinary Temperatures.
Journal of the chemical society: 3682–3690.

Taylor HFW (1997) Cement Chemistry. 2 Thomas Telford.

Thomas JJ, Jennings HM, Chen JJ (2009) Influence of Nucleation Seeding on the Hydration
Mechanisms of Tricalcium Silicate and Cement. The Journal of Physical Chemistry C 113: 4327–4334.

Viehland D, Yuan LJ, Z. Xu (1997) Structural Studies of Jennite and 1.4 nm Tobermorite: Disordered
Layering along the [100] of Jennite. J American Ceramic Society: 3021–3028.

Yu P, R. Kirkpatrick J, Poe B, McMillan PF, Cong X (1999) Structure of Calcium Silicate Hydrate (C-S-
H): Near-, Mid-, and Far-Infrared Spectroscopy. J American Ceramic Society 82: 742–748.