DENSIFIED ULTRA-LIGHT CEMENT-BASED MATERIALS. A FUTURE LOW CARBON CEMENT TECHNOLOGY BY SUPERABSORBENT POLYMERS

DENSIFIED ULTRA-LIGHT CEMENT-BASED MATERIALS A FUTURE LOW CARBON CEMENT TECHNOLOGY BY SUPERABSORBENT POLYMERS

Luis Pedro Esteves1* 1. Aalborg University, Sofiendalsvej 9-11, 9200 Aalborg, Denmark Abstract Densified cement systems were developed in the early 1980s, about three decades past. The research led to historical developments in cement and concrete research, forming the baseline for the design of modern cement systems, the socalled high-performance and ultra-high performance concrete. Cement production comprehends one of the relevant carbon emission footprints in the world. The substitution of cement by supplementary cementitious additions encompasses several other health hazards, risks and also technical difficulties such as limited or incoherent pozzolanic activity. Superabsorbent polymers can be used as a “clean technology” in the production of cement-based materials for structural applications with a low carbon footprint. This paper describes the principles of this concept coupled with experimental results on the basic properties of this enhanced type of cement-based materials with combined dense solid skeleton and yet low carbon cement technology. Originality This paper describes an alternative path to current cement technologies aiming at the development of low carbon cements. The use of superabsorbent polymers is a “clean” and sustainable alternative to inexistent or insufficient supplementary cementitious additions. The use of superabsorbent polymers as a low carbon agent is original by itself. This paper discloses several unseen properties of cement-based materials with superabsorbent polymers, and it defines U-DSP cement binders, a new class of cements with lower density than ordinary Portland cement and relatively high strength, which can have wide application in cement and concrete technology. Keywords: functional cement, superabsorbent polymers, low carbon cement, green concrete technology

1

Corresponding author: [email protected]

1. Introduction In the early 1980s, Hans H. Bache (1981) shocked the world with his basic principles on densified cement ultra-fine particle-based materials (DSP), concept known today as ultra-high performance concrete. Through his work, he proposed a new class of binders, which included the use of very fine silica powder – “Mikrosilica”. With the help of the newly developed dispersion agent “Mighty”, it was possible to mix powder blends into viscous pastes with very low water amounts (0.13 to 0.18 by weight of powder). This led to new cement-based materials with unseen mechanical properties. The use of water-entrainment by means of superabsorbent polymers was thoroughly explored in high performance (low water) cement-based materials to avoid early-age cracking through mitigation of self-desiccation, having its origin in the work by Ole M. Jensen (2001). An “inconvenient” of the use of superabsorbent polymers is the formation of empty porosity. This is a result of a sequence of events starting with the formation of polymer-water spheres, hydration of cement, internal curing and subsequent drying. This apparent “inconvenience” in terms of material strength may be a great advantage in terms of durability, and in regions of the world where frost resistance is required, studies show that the use of SAP is a smart possibility to design for durability of cement under freeze/thaw action (Mönnig 2009). In the FIBCEM project, research funded by the EU FP7 European 7th Framework Programme, SAP was used for the generation of low density cement foams in the production of extruded fibre-cement sandwich composites. SAP versatility in terms of pore agent makes it unique in the sense that pores can be designed with any specific size, and shape. More importantly, the addition of SAP results in closed porosity, which may be beneficial for holding good mechanical properties of cement-based materials. In addition to the potential enhancements in the microstructure, using SAP as pore agent leads to significant improvements in the hydration of densely-packed cement pastes (Justs et al 2014, Esteves 2014). Simultaneously with these developments, there are substantial efforts in the scientific community for finding alternatives to using cement as construction material. The more popular approach is the substitution of cement by supplementary cementitious materials – SCMs (Lothenbach et al 2011, Meyer 2009). This approach is, however, turning into an “obsession” judging by the development of publications indexed to the topic “supplementary cementitious materials” through search in Web of Science, with 700% increase from 2005 to 2012. Karen Scrivener (2012), a lead researcher in the field stated recently, that “Even fly ash, we´ve all heard about fly ash, and it is probably the most widely available SCM, it is produced in very small quantities. There simply isn´t enough fly ash worldwide to replace cement in any big amount…”. Being expected that the cement production may duplicate in the next three decades, alternatives to using SCMs as cement replacement will be critical in terms of demand in the near future (Schneider et al 2011). In other words, an alternative research path is urgently required. I believe that it is possible to act differently on the material design of cement-based materials. In this paper, a new role for the use of superabsorbent polymers in cement-based materials is explored – the possibility of creative design porosity of cement itself, optimising its design properties with focus on its technical and environmental performance. As a few grams of SAP can generate a high volume of water spheres, its applicability as a low carbon component in cement makes sense in the present economical context. This can be a new path to develop alternative low carbon cements for structural applications. The following will describe the principles of this conceptual approach, more holistic in terms of the performance of modern cement-based materials.

2. General principles The term DSP refers to “Densified Systems containing homogenously arranged, ultra-fine Particles”. It seems natural to introduce a term related to the empty space occupied by the optimised pore structure of the composite material – “Ultra-light”, from where results the acronym U-DSP. The principle to design densified ultra-light cement-based materials consists in proportioning a blend of a pore agent, e.g. superabsorbent polymers, and a densified low-water and low-porosity cement matrix – a high-strength solid fraction. Design of low carbon cement pastes via introduction of superabsorbent depends directly on the concentration and absorbency of the superabsorbent polymer used in the mix proportioning. Both the strength class and carbon equivalent of the blend are a function of the designed porosity. The design of the pore structure with this concept may be performed theoretically at any scale. Both sub-microscopic cell porosity and microscopic pores (inclusions) can be thought. In a DSP, the size of the particles arranged homogenously is within 0.1 and 100 µm. In a U-DSP, the predominant size range is a function of the particle size distribution of SAP. It is possible to idealise several different models of particles, and unimodal or bimodal distributions can be straightforwardly obtained. Fig.1 shows pore design as idealized for a range between 1 and 10 µm.

a

10 mm

b

10 mm

Figure 1. Design of U-DSP cement and illustration of the new concept with SAP leading to optimized porosity in the hydrates of the densely packed binder. a) DSP cement paste, b) an idealized U-DSP cement paste with close sub-microscopic cell porosity.

The packing density of pores is controlled in a similar way as the mix proportioning of filler or aggregate particles, in the sense that it is a particle-based system. In line with this method, it requires an exact measure of the swollen state of the SAP particles. The design through swollen particle size distribution of SAP can be conceptually performed via absorbency measurements as elaborated in Esteves (2014). An example of a specific SAP used in the production of porous cements is given in Fig.2. The packing density of water spheres may be studied in the same manner as for aggregate particles (see Esteves 2009). For a tri-component material based on SAP dispersed in cement and microsilica blends, the maximum density of the predominant phase will have as upper limit the postulate by Kepler, which refers to the maximum packing density of single spheres in an infinite

Euclidean space. According to this principle, a SAP with a mode of 75 µm may generate a maximum amount of 64% in pore volume, from where its absolute maximum proportion can be derived. The following will explore the design of low carbon cements taking SAP with particle diameters above the average size of cement grains. The chemistry in the hydration reaction of cement is assumed to consist of the same basic reaction products – cement hydrates, with predominance of calcium silicates. Volume [%] 20

15

10

5

0 1

10

100

317

1000

10000 Diameter [μm]

Figure 2. Volumetric PSD of a SAP material at its collapsed (orange) and swollen (blue) states. The measurement of the swollen PSD is performed in a relevant chemical environment by a laser diffraction technique, specially developed for application of SAP in cement-based materials. Absorbency (k) is derived according to (Esteves 2014).

3. Structure and properties of the hardened material The microstructure of low carbon hardened U-DSP consists of micro porosity generated by SAP and densely-packed cement-silica paste. The pore structure primarily depends on the concentration and particle size distribution of SAP. Fig.3a shows the microstructure of cement paste with superabsorbent polymers with a mode around 300 µm (swollen state). This value is confirmed by image analysis performed on data obtained in cement pastes by computer tomography CT. The mix composition of the cement blend governs the microstructure of the solid fraction. The case refers to a microstructure of a modified silica fume cement paste, but it may be extended to other types of cement blends, e.g. fly ash-cement blends. The microstructure of these systems consists in wellknown stable hydrated phases, and it will be treated as a homogenous solid fraction comprehending non-hydrated cement grains and hydrated cement compounds up to its maximum hydration degree. The total porosity of the solid phase can be estimated after Powers (1948), as elaborated in (Esteves and Jensen 2012)2. In the example given in Fig.3b, the pore structure of the dense-packed cement paste is characterized by non-hydrated parts corresponding to about 40% of the total amount of cement, with approximate density of 3.025 g cm-1, and amorphous silicate hydrates with approximate density of 2.331 g cm-3, taking Jennite as the reference C-S-H (Balonis and Glasser 2009). The total porosity of these hydrates is 0.059 cm3 g-1, as measured by nitrogen absorption, with pores situated around a

2

Internal curing may have a significant effect on the specific porosity of the hydrated solid phases.

mode value of 70 nm. This corresponds a surface area (BET) of 3.182 m2 g-1, which matches well with previous research on basic cement (Roessler and Odler 1985, Feldman 1985).

Figure 3. Microstructure of hydrated U-DSP cement. The cement paste is composed by cement with w/c ratio of 0.25. SAP was added at a rate of 1% by cement weight. a) Image refers to a computerized tomography scanned slice on a low water cement paste with superabsorbent polymers. The cylindrical sample has approximately 28mm diameter, consisting of 20483 voxels (which matches an edge length of 17 µm). Scale bar =1 mm. b) Image refers to the solid fraction of U-DSP taken by SEM with a BSE detector. Scale bar =10 µm.

3. Mechanical Properties 3.1. Compressive strength Blended cement pastes composed by cement and silica fume added at a rate of 10 to 20 % by cement weight, with w/c ratio of 0.20 - 0.25 (water to powder ratio of 0.16 - 0.18), superplasticizer and SAP at

a rate of 1% by cement mass lead to compressive strengths in the range of 40 - 60 MPa. Fig. 4 shows results taken on initial batches to analyze the strength development of the new porous cement pastes at an early-age. For a density of approximately 1700 kg m-3, measured at equilibrium in an atmosphere at 20 °C, the compressive strength of the composite is approximately of 50 MPa. This is far from the strength level achievable with DSP materials. With no SAP, the strength of the reference DSP cured in sealed conditions and prepared with standard vibration procedures is of 105 MPa. With appropriate curing conditions, a similar cement-silica blend can achieve higher values in compressive strength (see for example results in Lura et al. 2006, with strength values approaching 150 MPa at the same hydration degree3). However, these strength levels are difficult to achieve in practice for real-size concrete elements, without internal curing procedures. Compressive Strength [MPa] 100

80

60

U-DSP 50,6

40

48,6

44,4

24,6

20

0 1

3

7

28 time [days]

Figure 4. Compressive strength of U-DSP cement measured on cylinders with 10 cm diameter and 20 cm high (average of three samples), cured in sealed conditions at 40°C. Specimens were vibrated for 10-20 sec. at 30-50 Hz. Basic water to cement ratio w/c is 0.25. Entrained water to cement ratio (w/c)e is 0.20, calculated by multiplying the absorbency of SAP by the its mass. Absorbency of SAP can be consulted in Fig.2. Composition in kg m-3: 4

Portland cement Silica fume Water SAP water-entrained 5 “Mighty”

3

1140 228 285 11.4 228 14

This result was obtained in relatively small cylindrical specimens, with 20 mm diameter and 40 mm height. Therefore, they should not be directly compared with the values reported in this paper. 4 Portland cement refers to standard rapid hardening cement marked as CEM I 52.5N, acquired in Denmark. The basic chemical composition can be consulted from Aalborg Portland for the first half of 2012. 5 Refers to a standard superplasticizer based in polyether ester, acquired in Denmark.

3.2. Modulus of elasticity The U-DSP cement paste is relatively less brittle than a usual high-strength cement paste. The stressstrain curve is stable until failure, which means that most of the mechanical work may be considered elastic. In other words, the pores left by SAP can actively participate by absorbing elastic energy, which may include energy released during micro cracking, without compromising the mechanical stability of the material. In general, this is usually achieved by the aggregate phase in the case of structural concrete. In a standard DSP material, fragile fracture is often observed due to the state of the material (undergoing early-age cracking), and crack propagation during loading of the samples.

Stress [MPa] 100 DS

P

80

60

40 S U-D

P

20

0 0

0,05

0,1

0,15

0,2

0,25

0,3

Strain [%]

Figure 5. Stress-strain diagrams of U-DSP cement, in comparison with a standard DSP cement paste. The dotted red line indicates a stress-strain region of high probability for cracking and fragile structural collapse.

3.3. Fracture mechanics The fracture mechanics of cement is, as any other ceramic glass, characterised by brittle failure, in the sense that the initiation of a crack may lead to sudden failure and total structural collapse. Internal crack stabilizers such as aggregate particles are required to control this behaviour, but so far, without the presence of fibres or other high toughness reinforcement, cement-based materials cannot deform in a plastic regime without collapsing. Fig.6 shows how is possible, through material design, to change the stress-strain relationship (in bending) of U-DSP cements, so ductility is obtained. With about 2% PVA fibres by volume, strain-hardening behaviour is obtained and the material shows a relatively high toughness (in the range of 800 to 1200 N mm-1).

Flexural stress [MPa] 14

12

5 mm 200 mm

10

8

6

4

2

0

5

10

15

20

25

30

Deflection [mm]

Figure 6. Stress-strain diagram (in bending) of U-DSP cement with 2 % by volume 6 mm PVA fibers. The specimen was casted as single element with 240 x 60 mm and 5 mm thickness. Fig.4 gives the basic mix proportion of the material.

3.4. Volume changes It is well known that low water and low porosity cements show problematic early-age shrinkage, socalled autogenous shrinkage (Bentz and Jensen 2004). Cements in such physical systems are sensible to cracking if the deformation is impeded, and the simple presence of aggregate particles may result in substantial internal stress leading to microcracking (Esteves 2009). This has an effect on the mechanical properties of DSP materials. Volume changes of a U-DSP cement paste are dramatically different, as the system is composed by an even higher concentration of internal curing sources, compared with the suggested amount to balance the chemical shrinkage of cement hydration (Jensen and Hansen 2002). Although autogenous shrinkage and early-age microcracking of U-DSP materials is absolutely controlled, the drying shrinkage of this material may offer some challenges. In concrete designed with internal curing by SAP, it was found that the drying shrinkage rate decreases, when higher values in shrinkage-strain are obtained (Assmann and Reinhardt 2013). The impact of this behaviour on cracking was discussed on the basis of restrained shrinkage tests (Igarashi et al 2000) and well-established failure criteria. It was concluded that internal curing agents alone have only a mild effect on the prevention of early-age cracking (Zhutovsky et al 2013).

4. Environmental Performance The environmental performance of cement-based materials is typically evaluated through analysis of its life cycle. There are many factors that can determine the life span of cementitious materials. It is not only a matter of its mix design, but also of its function and performance in service, following its more or less early disposal. The main environmental impact of concrete is attributed to the CO2 emission related with the production of cement. Therefore, the CO2 footprint of concrete is largely influenced by the amount of cement used in its composition (Purnell and Black 2012). One of the ways to reduce the CO2 incorporated in concrete is to use mineral additions as replacement of Portland cement. In principle, SCMs used as by-products from other industrial developments contain lower amounts of incorporated CO2, so the resultant binder will have, in average, a lower amount of

embodied CO2 - eCO2. Fig.7 shows the eCO2 as function of strength for the relevant U-DSP cement, compared with the high-strength DSP, along with a reference representing the so-called green cement – a cement blend that includes high amounts of SCMs, viz. low amount of clinker. The results should be compared with some precaution, in the sense that compressive strength is not an absolute measure to evaluate environmental performance, as there may be different structural requirements for each of the model cements. For the same structural performance – e.g. same volume of concrete to produce a specific structure, U-DSP materials may reduce the environmental impact of cement 25% in relation to the DSP materials. At an equivalent strength, the green cement shows a potential for additional improvement of 10%, neglecting the favourable effect of U-DSP in relation to consumption of natural resources, energy used in transport and mass of the material. This is a promising result that requires further elaboration. eCO2 [kg CO2 per m3] 1,50

DSP 1,20

U-DSP

0,90

FA [1]

0,60

0,30

0 0

50

100

150

Strength [MPa]

Figure 7. Environmental performance (expressed by the value of eCO2 to the material strength) of U-DSP materials in comparison with standard DSP and green cement based on fly ash replacement. eCO2 is elaborated from the mix design of each system, weight-averaging of individual carbon loads. DSP and U-DSP refer to a blend of cement and silica fume made with basic w/c ratio of 0.25 (see basic mix proportion in Fig.4). G-CEM refers to a blend of cement and fly ash with basic w/c ratio of 0.85. Strength of G-CEM was adapted from (Proske et al 2013), by multiplying the cube strength by a factor of 0.8, in order to convert to cylinder strength.

5. Durability With appropriate design and production methods, with relevance for the curing methods, the durability of high-strength cement-based materials is outstanding when compared to the durability of standard cements used in concrete structures (Aïtcin 1998). The fine microstructure obtained with the fumed silica results in an impervious skin, which is also present in U-DSP cements. It is expected that the durability of this cement is comparable or higher to that obtained in dense-packed cements. There are two main reasons for this. Firstly, the introduction of SAP mitigates autogenous shrinkage and, therefore, microcracking present in DSP cement, especially those subjected to either internal or external restraints. Secondly, the internal curing has as result a refinement of porosity in C-S-H due to extended hydration of cement compounds. This means that the transport of specific substances through

the relatively less permeable solid skeleton will be far more difficult. Basic research on the transport mechanisms of these materials is needed. A few examples of applied research directed to the understanding of durability of concrete with superabsorbent polymers may be found in the literature. Freeze-thaw resistance of concrete made with small amounts of SAP used as air-entrainment (5 to 10% by volume) is significantly improved (Lautzen et al 2008). The addition of relatively higher amounts of SAP in the mix design of U-DSP cements has as microstructural consequence, a decrease in the pore spacing factor, which is beneficial in controlling the scaling of the material due to exposure to freeze-thaw atmospheres (Zhutovsky et al 2011). Chloride induced corrosion is the major cause of deterioration in structural concrete. The level of chloride is pointed as the critical factor, and it is governed by transport mechanisms in cement (Angst et al 2009). Chloride migration coefficient determined by rapid migration test on concrete composed by a standard DSP cement with addition of a small amount of SAP showed a lower value when compared to the basic concrete formulation (Reinhardt and Assmann 2012).

6. Applications Many applications can be foreseen for U-DSP materials. In principle, mix compositions can be developed for both structural and non-structural concrete elements. So far, only non-structural applications were experimentally validated. The design of structural elements through this conceptual approach requires further investigation. 6.1. H CEM The use of SAP with sub-microscopic size will lead to optimized spherical porosity in the C-S-H system. In the relevant size fraction, SAP may act as catalyzer in the hydration of cement compounds, as it minimizes Gibbs free energy in the so-called heterogeneous nucleation. It is reasonable to assume that SAP should be designed with a size higher than the critical cluster size, which is defined as the size at which Gibbs free energy G reaches its maximum value. The size range of this type of particle, or pore, can be within 1 and 10µm. SAP is typically cross-linked by 20% monomer by mass. Taking absorbency as 10 ml g-1, as measured in the relevant chemical environment, the previous elaboration for SAP with 1µm means that the new cement contains water nucleus of around 2.5µm. This 8x10-7µg particle will be delivering approximately 2.6x1017 water molecules useful to sustain hydration of C-SH, particularly (but not limited) in low water and low porosity cement systems. 6.2. Structural Concrete Structural concrete is used in most of the main civil infra-structure all over the world. This includes bridges, tunnels and most of the modern building developments, either in situ casted or precast concrete elements. U-DSP materials are particularly interesting for the production of sustainable concrete structures, due to its high specific strength, despite the relatively lighter material in comparison with DSP materials. Different mix proportioning can be set up for this purpose. Lightweight bridge-decks and structural building elements can be produced with remarkably lower density. Design loads (dead loads) in structural concrete can be substantially reduced, without compromising the typical mechanical specifications. The Great Belt Bridge in Denmark was built with about 1.1 billion cubic meters of concrete with an average specific mass of 2300 kg m-3. This results in a total mass over 2.5x106 tones. It is reasonable to foresee green concrete produced with U-DSP cements, with 1800 kg m-3 and the same mechanical performance. This means that the total dead load of the bridge can be reduced up to about 20% with the current technology (other speculations will be explored). 6.3. Fibre-cement U-DSP concept was successfully used in the FIBCEM (see Fig.8). The project aimed at the production of extruded foam cement for sandwich panels with improved mechanical and durability properties. Cement foams with densities lower than 1000 Kg m-3 were produced with relatively high compressive

strength (Esteves and Jensen 2012). Single extrusion of low density U-DSP was successfully obtained. The foam cement based on SAP can be used in the core of the FIBCEM material, a high-performance cement sandwich, resulting in a lower density material compared to existing fiber-reinforced cements FRC. In addition, the configuration of the core material leads to improvements in thermal and acoustic properties of the cement panels. FIBCEM will be produced by a low energy multilayer extrusion process in which both foam cement core and fiber reinforced skins can be simultaneously formed such that no discontinuity is formed between them. By using a foam core and replacing part of the cement with materials such as fly ash and silica fume, the CO2 footprint of the material can be significantly reduced compared to existing FRC solutions.

Figure 8. FIBCEM - Nanotechnology Enhanced Extruded Fibre Reinforced Foam Cement Based Environmentally Friendly Sandwich Material for Building Applications. This image is reprinted from the memorandum agreement, and was kindly authorized by Cembrit Holding Denmark.

7. Concluding remarks The development of clean cement technologies for modern construction materials is required for our future as a society. In a context where environmental constrains are increasingly demanding with regards to the efficient use of natural resources, to carbon footprint and use of energy, alternative design methods for cement-based materials are urgently required for the production of low carbon concrete. The use of superabsorbent polymers as replacement of cement can have a huge impact on the environmental performance of cement-based materials. This can be achieved through design of the pore structure of cement itself at both microscopic and submicroscopic scales coupled with the strengthening of its solid fraction. This paper describes the potential applications of SAP as a strong low carbon agent candidate, to be used in the future of cement-based materials, minimising the problem of scarcity of raw material sources such as supplementary cementitious materials. Structures with enhanced durability can be produced with this type of material, and huge savings in cement can be foreseen. This paper reports an un-explored research area that requires more and better attention by the research society working with the cement. SAP was “born” to prevent early-age cracking in high-strength cements, but it has the potential to be “re-born” in many other cement applications. Young scientists may find it interesting to help generating the necessary critical mass into this new path.

Acknowledgements The inventive step towards H CEM, described in this manuscript is built on prior art resultant from the EU FP7 project FIBCEM - Nanotechnology Enhanced Extruded Fibre Reinforced Foam Cement Based Environmentally Friendly Sandwich Material for Building Applications (Grant agreement no. 262954). The selection to do this work by Ole M. Jensen is appreciated. Thanks to Anja M. Bache for inspiration, discussion and trust, where also emerged many time-travelled thoughts by Hans H. Bache. References ! Aïtcin, P. C. High-Performance Concrete. (E & F.N. Spon, 1998). ! Angst, U., Elsener, B., Larsen, C. K. & Vennesland, Ø. Critical chloride content in reinforced concrete — A review. Cem. Concr. Res. 39, 1122–1138 (2009). ! Assmann, A. & Reinhardt, H. W. The use of superabsorbent polymers to mitigate shrinkage of concrete. in Mech. Phys. Creep, Shrinkage, Durab. Conrete a Tribut. To Zdenek P. Bazant - Proc. 9th Int. Conf. Creep, Shrinkage, Durab. Mech. CONCREEP 2013 301–307 (American Society of Civil Engineers (ASCE), 2013). ! Bache, H. H. Densified cement ultra-fine particle-based materials. (1981). ! Balonis, M. & Glasser, F. P. The density of cement phases. Cem. Concr. Res. 39, 733–739 (2009). ! Bentz, D. P. & Jensen, O. M. Mitigation strategies for autogenous shrinkage cracking. Cem. Concr. Compos. 26, 677–685 (2004). ! Esteves, L. P. Internal curing in cement-based materials. PhD thesis. Portugal: Aveiro University. (2009). ! Esteves, L. P. and Jensen, O.M., Design of porous cement-based materials – Defining a research context. Technical University of Denmark, 2012. (working paper) ! Esteves, L. P. Recommended method for measurement of absorbency of superabsorbent polymers in cementbased materials. Mater. Struct. (2014). doi:10.1617/s11527-014-0324-5 ! Esteves, L. P., Lukošiūtė, I. & Čėsnienė, J. Hydration of cement with superabsorbent polymers. J. Therm. Anal. Calorim. 118, 1385-1393 (2014). ! Feldman, R. F. Properties of Portland cement-silicate fume pastes. I. Porosity and surface properties. Cem. Concr. Res. 15, 765 – 774 (1985). ! Igarashi, S., Bentur, A. & Kovler, K. Autogenous shrinkage and induced restraining stresses in high-strength concretes. Cem. Concr. Res. 30, 1701–1707 (2000). ! Jensen, O. M., Hansen, P. F. Water-entrained cement-based materials - I. Principles and theoretical background. Cem. Concr. Res. 31, 647 – 654 (2001). ! Jensen, O. M., Hansen, P. F. Water-entrained cement-based materials II. Experimental observations. Cem. Concr. Res. 32, 973–978 (2002). ! Justs, J., Wyrzykowski, M., Winnefeld, F., Bajare, D. & Lura, P. Influence of superabsorbent polymers on hydration of cement pastes with low water-to-binder ratio. J. Therm. Anal. Calorim. 115, 425–432 (2014). ! Laustsen, S., Hasholt, Marianne T., Jensen, O. M. A new technology for air entrainment of concrete. in Microstruct. Relat. Durab. Cem. Compos. 61, 1223 – 1230 (RILEM, 2008). ! Lothenbach, B., Scrivener, K. & Hooton, R. D. Supplementary cementitious materials. Cem. Concr. Res. 41, 1244–1256 (2011). ! Lura, P., Durand, F., Loukili, A., Kovler, K. & Jensen, O. M. Compressive strength of cement pastes and mortars with superabsorbent polymers. Vol. Chang. hardening Concr. 117–726 (2006). ! Meyer, C. The greening of the concrete industry. Cem. Concr. Compos. 31, 601–605 (2009). ! Mönnig, S. Superabsorbing additions in concrete&: applications, modelling and comparison of different internal water sources. (2009). at http://elib.uni-stuttgart.de/opus/volltexte/2009/4781 ! Powers, T. C. & Brownyard, T. L. Studies of the physical properties of hardened Portland cement paste, Bulletin 22. Res. Lab. Portl. Cem. Assoc. (1948). ! Proske, T., Hainer, S., Rezvani, M. & Graubner, C.-A. Eco-friendly concretes with reduced water and cement contents — Mix design principles and laboratory tests. Cem. Concr. Res. 51, 38–46 (2013). ! Purnell, P. & Black, L. Embodied carbon dioxide in concrete: Variation with common mix design parameters. Cem. Concr. Res. 42, 874–877 (2012). ! Reinhardt, H.-W. & Assmann, A. in Appl. Super Absorbent Polym. Concr. Constr. (Mechtcherine, V. & Reinhardt, H.-W.) 2, 115–135 (Springer Netherlands, 2012). ! Roessler, M. & Odler, I. Investigations on the relationship between porosity, structure and strength of hydrated portland cement pastes - i. Effect of porosity. Cem. Concr. Res. 15, 320–330 (1985).

! Schneider, M., Romer, M., Tschudin, M. & Bolio, H. Sustainable cement production—present and future. Cem. Concr. Res. 41, 642–650 (2011). ! Wray, P., Scrivener, K. Straight talk with Karen Scrivener on cements, CO2 and sustainable development. Am. Ceram. Soc. Bull. 91, 47–50 (2012). ! Zhutovsky, S., Kovler, K. & Bentur, A. Revisiting the protected paste volume concept for internal curing of high-strength concretes. Cem. Concr. Res. 41, 981–986 (2011). ! Zhutovsky, S., Kovler, K. & Bentur, A. Effect of hybrid curing on cracking potential of high-performance concrete. Cem. Concr. Res. 54, 36–42 (2013).