Investigation of the microstructural evolution of ...

UNIVERSITA’ DEL PIEMONTE ORIENTALE “A. AVOGADRO”

Department of Science and Technological Innovation (DISIT)

PhD in Chemical Sciences XXVIII cycle

Advanced materials in cement industry Study and development of multifunctional cement-based materials

Sara Irico

Scientific Supervisor: Dr. Enrico Boccaleri Buzzi Unicem Supervisors: Dr. Fulvio Canonico, Dr. Daniela Gastaldi Coordinator: Prof. Domenico Osella

INDEX

GENERAL INTRODUCTION ................................................................................ 1 INTRODUCTION – MULTIFUNCTIONAL CEMENTIOUS MATERIALS 1

2

MULTIFUNCTIONALITY IN CEMENT-BASED MATERIALS ............. 5 1.1

Definition of multifunctional material ......................................................... 5

1.2

Multifunctionality applied to building materials .......................................... 7

1.3

Building towards a sustainable and resilient future ..................................... 8

1.4

Objectives of the research .......................................................................... 11

THE CHEMISTRY AND MINERALOGY OF THE CEMENT ............... 12 2.1

Introduction to cement chemistry ............................................................... 12

2.2

Composition of Portland clinker ................................................................ 13

2.3

Hydration of Portland cement .................................................................... 14

2.4

Influence of physicochemical properties on multifunctionality ................. 17

FIRST PART - PHOTOCATALYTIC CEMENT-BASED MATERIALS 3

4

5

TiO2 ASSISTED PHOTOCATALYIS .......................................................... 23 3.1

NOx oxidation mechanism ......................................................................... 25

3.2

Self-cleaning ability and Photo-induced hydrophilicity............................. 27

PHOTOCATALYSIS IN THE CONSTRUCTION INDUSTRY ............... 31 4.1

State of the art ............................................................................................ 31

4.2

Standardization and certification assisting commercialization .................. 34

EXPERIMENTAL PART I: NOX REDUCTION ........................................ 39 5.1

NOx removal test set-up ............................................................................. 39

5.2 Pilot scale production of a photoactive cement with functionalized floating n-TiO2 ................................................................................................................... 42 5.2.1

Effect of the mix design on phocatalytic properties ........................... 43

5.2.2

Effect of environmental conditions on photocatalytic properties ....... 46

5.2.3

Costs assessment analysis of different building materials .................. 49

5.3

Toward a micro-TiO2 assisted photocatalysis in cement ........................... 51

5.3.1

Efficiency of the m-TiO2: experimental considerations ..................... 52

5.3.2

Photocatalytic functionalized demolding agents ................................ 53

5.3.3 Mechanochemical doping of m-TiO2: an industrial scale potential technique ........................................................................................................... 56

5.4

6

5.3.3.1

Transition metal doping “Cu-doped TiO2” .................................. 59

5.3.3.2

Non-metal doping “N-doped TiO2” ............................................. 63

Field test: photocatalytic demonstrator ...................................................... 65

5.4.1

Photocatalytic demonstrator set-up and Computational modelling .... 67

5.4.2

Preliminary results .............................................................................. 72

EXPERIMENTAL PART II: SELF-CLEANING ABILITY ..................... 74 6.1

Rhodamine test: method and considerations .............................................. 74

6.2

A case study of self-cleaning concrete ....................................................... 77

6.3

Self-cleaning vs air purification ability ...................................................... 79

Conclusions .............................................................................................................. 80 SECOND PART - SELF-HEALING CEMENTITIOUS MATERIALS 7

SELF-HEALING IN CONSTRUCTION MATERIALS ............................ 83 7.1

Self-healing phenomena ............................................................................. 83

7.2

Why self-healing in cement-based materials? ............................................ 84

7.3

Evaluation of the healing effect ................................................................. 85

7.4

State of the art of self-healing in cementitious material ............................. 88

7.4.1

Recovery against environmental action .............................................. 89

7.4.2

Recovery against mechanical action ................................................... 92

7.4.3

Encapsulation approach ...................................................................... 93

8 THE “DualCEM” PROJECT: ENCAPSULATION OF HEALING AGENT .................................................................................................................... 99 8.1

Healing agent characterization ................................................................. 102

8.1.1 Sodium silicate as healing agent: insight of the reaction with Portland cement..............................................................................................................104 8.2

Extruded Cementitious Hollow tubes containers ..................................... 112

8.3

Spherical capsules by mechanical dripping method ................................ 121

8.3.1

Sodium/Calcium Alginate capsules .................................................. 123

8.3.2 Assessment of encapsulation synthesis by a preliminary chemometric approach .......................................................................................................... 134 Conclusions ............................................................................................................ 142

GENERAL CONCLUSIONS AND IMPACTS OF THE RESEARCH ON UNIVERSITY-INDUSTRY COLLABORATION ............................................ 144 An example of knowledge sharing between UPO University and Buzzi Unicem on the development of sustainable cement .................................................................. 149 Activities and congress participations..................................................................... 150 Bibliography ........................................................................................................... 152 Web site references ................................................................................................. 159 Acknowledgements ................................................................................................. 160 APPENDIX ............................................................................................................. 161

GENERAL INTRODUCTION

W

ith the dawn of twenty-first century, the world has entered into an era of sustainable development. The main challenge for the concrete industry is to combine two relevant aspects of the society, the attention

to environmental sustainability, on one hand and, - on the other hand - meeting the infrastructural requirements of the growing population. First of all, the concrete industry is one of the largest consumers of natural resources and Portland cement, the cement of the modern infrastructures, is not environmentally friendly. In the latest decades, more attention has been paid to the development of advanced cementitious materials, which focuses on such issues as environmental sustainability, reduction of clinker content, reduction of CO2 emission during the production, new clinker development and increased use of recycled materials. Moreover, an increasing number of studies have been carried out in order to improve the mechanical performance and durability aspects, both also related to an improvement of concrete as well. The new challenges in cement research are nowadays comprised of the development of advanced multi-functional features, such as self-cleaning and photocatalysis, self-healing, self-monitoring and thermal isolation, to name a few. Looking at all these aspects, the potential for innovation in cement materials science and in construction technology is far wider. This PhD research was financed by the R&D division of Buzzi Unicem, a leading international company in the production of cement and concrete. This research is based on a long-term collaboration between Buzzi Unicem and the Department of Science and Technological Innovation of the University of Piemonte Orientale. The research activities were carried out in the R&D laboratories of Buzzi Unicem in Trino (VC), in the University laboratories in Alessandria (AL) and the laboratories of materials science in Vercelli (VC). Additionally, some of the activities were performed with the collaboration of industrial partners and academic institutes, such as the University of Milano and Politecnico of Torino.

1

The main objective of the PhD research consists of increasing the basic knowledge and comprehension tools highly innovative research fields, and transferring the knowledge and the technology from academic research to potential industrial applications. In this PhD research, different multifunctional items have been investigated: photocatalytic building materials and self-healing cementitious materials. The evaluation and assessment of new functionality has been explored, including environmental, durability and aesthetic targets. The multifunctional advanced materials are designed to meet specific requirements through tailored properties. Smart materials can be considered as multifunctional ones that have the ability to react upon an external stimulus, simulating in this way, the behavior of nature’s materials. Currently, photocatalytic cements have gained increased attention due to their ability to reduce air pollutants and maintain their aesthetic surface properties, but a lot of work is needed to drive these products into the market, such as the reduction of costs, the improvement in the durability and efficiency checks in a real scale. In the same way, research on self-healing cementitious materials has shown an increasing number of publications. The technologies of self-healing are based on the principle that a material is able to repair itself without external intervention of maintenance. Although this functionality is an appealing research field in cement materials science, barriers to its implementation in real applications are the main challenges of the scientific community. The financial support of Buzzi Unicem in this research, allowed me to be involved in several research activities carried out in the company with the collaboration of the Department of Science and Technological Innovation. Some of these activities are the investigation of the potentiality of thermoporometry technique for microstructural characterization of cement paste, investigation of the effect of carbonation attack on cementitious materials, evaluation of sulfate attack on structural properties, quality control tests in special cement production. An example of sustainable use of hydrated cement waste (HCW) in the production cycle of Portland cement was also investigated. Recycling is one issue of sustainability. Hydrated cement waste obtained as by-product of efficient separation 2

of fine recycled concrete aggregates (FRCA) could be of great interest as recycled material in the cement industry.

3

INTRODUCTION

MULTIFUNCTIONAL CEMENTITIOUS MATERIALS

4

Introduction – Multifunctional cementitious materials

1

1.1

MULTIFUNCTIONALITY IN CEMENT-BASED MATERIALS Definition of multifunctional material

M

ultifunctional materials can be defined as those presenting specific desirable electronic, magnetic, optical, thermal or other properties able to satisfy previously unachievable performance features.

In science and technology, ideas often come from observing nature, or in other cases crossing knowledge over materials science fields. The bio-inspiration is the driveline to stimulate research in non-biological sciences toward smart materials. The introduction of these materials into production will significantly modify the design philosophy itself in the direction of sustainability (Figure 1.1). In other cases, relevant drivelines for innovation are the findings in fields not directly concerned with the specific research areas; for instance, the growth of scientific work on nanostructured semiconductors and on the synthesis of inorganic nanoparticles has stimulated the development of innovative photocatalytic and nanostructured building materials. The multifunctional properties are combined with each other or with specific mechanical properties including stiffness, ductility or strength aimed at developing new emerging applications and mixtures in the background of developing multifunctional composite structures (Figure 1.2).

Figure 1.1 Smart material towards sustainability

5


Nevertheless, the use of such materials is limited by the processing techniques available. In general, the costs of designing and producing a novel multifunctional material can be considered high, and the risk of this investment is significant. Therefore, the economic fostering of such an attempt depends largely on acceptance of this type of material, both by the industry and customers. The costs of production are often still high and special care is required so that the functionality and properties of the raw materials are not disturbed. Furthermore, it must be take into account the strict international regulations and certifications standards. Three major families of multifunctional materials have been identified and are considered1: 

Structural composite materials



Smart or intelligent materials



Nanostructured materials

Figure 1.2 Synthesis of multifunctional composite structures [K. Salonitis et al. (2010)].

Intrinsically, smart structural composites are multifunctional structural materials, which can perform additional functions such as sensing strain, stress, damage or temperature, damage (crack) propagation prevention, thermoelectric energy generation,

electromagnetic

interference

(EMI)

shielding,

electric

current

rectification, internal self-healing, and active noise and vibration control.

1

K. Salonitis, J. Pandremos, J. Paralikas, G. Chryssolouris, Multifunctional materials: engineering applications and processing challenges, Int J Adv Manuf Technol 49 (2010) 803–826.

6


In this research, self-healing materials are investigated as an example of smart materials. The smart or “intelligent” materials respond to environmental stimuli with particular changes in some of their variables. Applications of self-healing materials already exist in buildings2, in asphalts, for coatings in automotive sector (Nissan Motor Co., 2005), aeronautics, polymeric industry, to name a few. Nano (or micro) science and engineering came into common usage to develop nanostructured materials. Concrete can be nano-engineered by the incorporation of nanosized building blocks or objects (e.g., nanoparticles and nanotubes) to control material behavior and add novel properties. In the case of photocatalytic and selfcleaning materials, the addition of n-TiO2 or m-TiO2 provide a surface functionality, which can be tailored to promote specific interfacial interactions. In biological systems, the self-cleaning effect is present in lotus leaves, the legs of the water strider, cicada’s wings, gecko feet and butterflies wings3.

1.2

Multifunctionality applied to building materials

Why are we talking about a traditional material like cement as an innovative and smart material? In the past few years, scientific research in the field of building materials has directed more and more resources to the development and the optimization of advanced and multifunctional cement-based materials. The main purpose is to increase the potentiality of application of these materials responding to specific use requirements, combining new physico-chemical properties with the characteristic features of common building materials (i.e. structural properties, performances, durability, etc.) to achieve innovative properties and performance. The physico-chemical properties of cementitious materials can be modified and tailored to specific uses to obtain materials with multifunctional properties and characteristics that had never been achieved, nor considered, in the past. 2

C. Joseph, D. Garder, T. Jefferson, B. Isaacs, B. Lark, Self-healing in cementitious materials: a review of recent work, Construction Material 164 (2011). 3 P. Ragesh,V. Anand Ganesh, S.V. Nair, A. Sreekumaran Nair, A review on ‘self-cleaning and multifunctional materials’, J Mater Chem A 2 (2014) 14775.

7


In the cement-based materials science field, very interesting from the scientific and applicative points of view are, for example: (i) self-healing materials which have the intrinsic ability to repair themselves when damaged, extending the life cycle; (ii) selfcleaning cement-based materials that are not soiled by atmospheric smog; (iii) photocatalytic manufactured elements and buildings able to reduce air pollutants and purify the air of our cities, etc.. In the future, the use of such materials is considered one of the key strategies to enable the development of smart cities (Figure 1.3) that combine innovation, environment and quality of life by rebuilding, in terms of sustainable development. Increased resources in the research and development of multifunctional cement-based materials will maintain a high level of competition in the cement industry and contribute actively to the development of science and technology of sustainable building materials.

Figure 1.3 Schematic illustration of the smart city applications

1.3

Building towards a sustainable and resilient future

Sustainability is generally associated with the availability and a prudent use of natural resources, and consideration for both present and future generations. Sustainability also means control and following conditions that can fulfil the social, economic, environmental aspects and other requirements of the present society and future generations. Figure 1.4 shows the three pillars and intersecting areas of synergistic opportunities.

8


Figure 1.4 The functional definitions that align with the “three pillars” of sustainability, or the triple bottom line, where the “Three E’s” of environment, economics and efficiency overlap. The theoretical framework for sustainable decision-making is used to place more or less emphasis on each of the pillars, or to provide synergistic opportunities between the them. [J. Elkington (1997)4].

In building materials research, recycling is one issue for sustainable development. In Europe, about 180 million tons of concrete demolition waste (CDW) are produced every year, corresponding annually to 500 kg for each citizen5: this amount represents around 31% of all the waste produced in the European Union6. In recent years concrete recycling has gained more and more interest, due to increasing attention towards environmental protection and conservation of natural resources. The use of recycled concrete aggregates (RCA) is a well-established practice in Belgium, Denmark and the Netherlands. In these countries, recycling rates are 80%7. Recently, the application of recycled aggregates in precast industrial production, even in the presence of fine aggregates, has been gaining increasing attention with satisfying results.

4

J. Elkington, Cannibals with Forks: The triple bottom line of 21st century business, Capstone: Oxford, (1997). 5 Symonds Group Ltd 469697. Construction and demolition waste management practice, and their economic impact. Final Report to DGXI, European Commission; February 1996. 6 C. Fisher, M. Werge EU as a recycling society – present recycling levels of municipal waste and construction & demolition waste in the EU. ETC/SPC working paper 2. (2009) Copenhagen. 7 V. Corinaldesi, G. Moriconi, Behaviour of cementitious mortars containing different kinds of recycled aggregate, Constr Build Mater 23 (2009) 289–94.

9


In Germany, only the coarser part of the recycled concrete can be used in the production of new concrete.

Sustainability and resilience are fundamental design components that should be inextricably linked when making investment, design, planning, and community support decisions8. The Figure 1.5 shows the relationship between traditional sustainable design, disaster resilience, and sustainability.

Figure 1.5 Relationship between traditional sustainable design, disaster resilience and sustainability [Buffenberger (2015)].

A sustainable perspective must rely on a synergistic approach based on materials with limited impact and a reliable use of them through a careful design. Alongside the transition to a low carbon society, increasing resilience in response to climate-related, natural and manmade disturbances is the key to assuring security, reduction of property damage, and enabling businesses. Commitment to advanced building stock resilience must include concrete structures with extended service lives, functionality and adaptability, reduced operations and maintenance, environmental sustainability, and increased public health and safety. Improvements in sustainable and resilient project delivery are possible through a multidisciplinary integration of material and design selection based upon life cycle analysis measurements, implementation of life cycle cost analysis versus lowest cost economics, use of innovative materials and technologies; and collaborative platforms during design and construction projects.

8

J. K. Buffenbarger, Building toward a sustainable and resilient future, proceeding of DSCS2015, Bologna.

10


Hence, enhancing the resilience of buildings and infrastructures through designed robustness, durability, longevity, disaster resistance and safety should be a priority for every sustainable community stakeholder9.

1.4

Objectives of the research

Concrete is the most applied material in building structures. The main construction objectives are related to normative requirements in construction, such as compressive strength resistances, exposure classes (defined for specific regional durability concerns), durability aspects and desired setting time, to name a few. Nevertheless, there are many possibilities to give concrete/concrete surface new functionalities that could add new properties, never considered in the past, to an “old fashioned” material. The objectives of this PhD research are the evaluation and assessment of new functionalities with the following main targets: 

Pollution reduction by a photocatalytic cementitious surface – environmental target;



Self-cleaning (also called photocatalytic activated concrete surface) – durability and aesthetic target;



Self-healing cementitious material – durability target;



Economic considerations;



Application of the technologies to real application case studies.

11


2 2.1

THE CHEMISTRY AND MINERALOGY OF THE CEMENT Introduction to cement chemistry

T

he fundamental concept of “cement” is taken from the Technical Commission for the Building Standardization as follows: “The cements are binders able to become hard in air conditions or in water,

and after hardening, to resist water. The cements are a combination of calcium oxides with silicate oxides, aluminum oxides and iron oxides, able to satisfy the standard requirements for the cement, such as compressive resistance and volume stability. The raw meal, or the main component, must be burned until the synthetization occurs”. Portland cement is the most well-known and used cement. Portland cement is produced by burning a mix of limestone (CaCO3), quartz sand and clays at 1450°C in air conditions. Limestone is used as a precursor of calcium oxide - further sources are gypsum, and other slag containing calcium and dolomite (MgCa(CO3)2). The clays are a source of aluminum and silicon, and might contain clay minerals, such as kaolinite, montmorillonite, illite, chlorites, mica. During the burning process, which occurs around 1450°C, the partial fusion of the oxides leads to the synthesis of a product called “clinker”. The chemical composition of the clinker in terms of oxides, is as follows: 63-70% CaO, 19-24% SiO2 , 3-7% Al2O3, 0-5% Fe2O3 and other minor components9. These oxides are present in the clinker mainly as four mineralogical components: alite (tricalciumsilicate), belite (dicalciumsilicate), tricalcium aluminate, tricalcium aluminoferrite, described as follows in details10. Other types of cement are produced in the cement industry. These include sulfoaluminate cement (CSA), with a clinker based on calcium sulfoaluminate phases instead of silica phases. These phases are produced through the combustion of gypsum and lime in a rotary kiln at 1300 °C. Another example is the calcium aluminate cement (CA), is a cement consisting predominantly of hydraulic calcium aluminates, the main

9

F. Locher, (2000) Zement, Grundlagen der Herstellung und Verwendung. Düsseldorf: Verlag und Technik. 10 H.F.W. Taylor (1997) Cement Chemistry, 2° ed. Thomas Telford, London.

12


constituent is monocalcium aluminate. This cement is produced from limestone and low-silica bauxite, and is used in a number of small-scale, specialized applications.

2.2

Composition of Portland clinker

Alite is the main phase in a Portland cement. This mineral phase is a tricalcium silicate 3CaO·SiO2 (Ca3SiO5) that can contain a small amount of sodium, magnesium, aluminium and iron. The alite is the calcium silicate phase that reacts faster with water, and is the main component responsible for the hardening of cement after 28 days. Belite is a dicalcium silicate 2CaO·SiO2 (Ca2SiO4) mainly present in cement as “β polymorph”, stabilized thanks to the presence of foreign ions. The reaction of belite is quite low, but this phase is responsable for the increase of compressive strength beyond 28 days. Tricalcium aluminate 3CaO·Al2O3 (Ca3Al2O6) has a very fast reaction, sometimes unwanted, with water, and for this reason gypsum is added (CaSO4·2H2O) to control the hardening process. Browmillerite 4CaO·Al2O3·Fe2O3 (Ca4Al2Fe2O10), has a behaviour similar to tricalcium aluminate. According to the convention of mixed oxides, widely used in the cement industry, the following abbreviations are used: 

C = CaO



S = SiO2



A = Al2O3



F = Fe2O3



H = H2O



S = SO3



K = K2O

The chemical formulae of the clinker components are shortened as follows: 

alite C3S;



belite C2S;



tricalcium aluminate C3A;

13




tetracalcium ferroaluminate C4AF.

After the quenching process, the resulting clinker is milled and ground into a fine powder. At this stage various other materials, such as limestone, gypsum, pozzolanic materials (main components according to EN 197-1), can be added to the cement powder (latent hydraulic and pozzolanic properties, inert for optimization of e.g. packing density). When the cement is mixed with sand and water, the hardened material is called cement mortar. In the majority of cases the cement is used in concrete, in this case the cement is mixed with water, sand and coarse aggregates (>4 mm), fibers, additives, etc. The amount of cement used in a standard concrete is 300-350 Kg/m3.

2.3

Hydration of Portland cement

The cement powder is mixed with water, and then poured for the desired application. The final hydrated cement product generally contains about 30-40% water after hydration (38-40% as chemical phases and gel pores, more water is stores in so-called capillaries). This value varies little in the composition of the cement clinker. The hydration of cement consists of a complex series of chemical reactions, which are still not completely understood11. Cement hydration can be affected by a variety of factors, such as water-to-cement (w/c) ratio, the curing conditions in terms of temperature and humidity, the specific surface and the presence of additives. The reaction with water leads to the achievement of the main physical, chemical and mechanical properties of the cement products. The hydration of silicates can be summarized as follows: C2S + H2O  CxSHy + nCH C3S + H2O  CxSHy + mCH In both reactions calcium hydroxide (portlandite) and hydrated calcium silicate (conventionally known as C-S-H) are formed. While portlandite is characterized by a well-known crystal structure, C-S-H is a phase having variable stoichiometry and a porous microstructure - it can be defined as a rigid gel. The C-S-H gel is often referred 11

D. C. MacLaren, M. A. White, Cement: Its chemistry and properties, J. Chem. Educ. (2003) 80(6) 623.

14


to in literature as Tobermorite, a crystalline/semi-crystalline gel phase. Other calcium aluminium silicate hydrates (C-A-S-H) can be also formed. Jennings (2008)12, proposed a second generation of model structures, a nanometric frame of C-S-H where water molecules take places in the interlayers.

(a)

(b)

Figure 2.1 (a) C-S-H globules, (b) aging of C-S-H structure [Jennings (2008)].

The hydration of aluminate phases occurs quickly in the first hours of hydration, with a considerable heat development reaction: C4AF + H2O  C(A,F)H C3A + H2O  CAH Calcium aluminate hydrates (C-A-H) are the reaction products of aluminates phase hydration, C-(A,F,)-H indicates the presence of iron ions in the crystalline lattice. The reaction of the aluminates is not responsible for the mechanical strength beyond 28 days. The rapid reaction leads to a fast loss in the workability (fast setting), that must be controlled by additions of calcium sulfate and additives to guarantee the workability of the hydrated cement. The morphology of the C-A-H phase is based on hexagonal platy like crystals and cubic crystals. Contrary to the fibrous structure of C-S-H, the morphology of C-A-H crystals is not favorable for improving mechanical strength resistance.

12

H. M. Jennings, Refinements to colloid model of CSH in cement: CM-II, Cem Concr Res, 38 (2008)

275-289.

15


To overcome to the problem of fast setting time ( 20% ;

75

First Part – Photocatalytic cement-based materials

𝑅ℎ =

𝑎∗ (0) − 𝑎∗ (ℎ) 𝑥 100 𝑎∗ (0)

Sample preparation. Standard mortars are prepared according to EN-196/1, w/c 0.5. A silicon mold is used with sample dimension 10x10x2cm (squared sample allows a better homogeneous irradiation). The cement used is the white cement CEM II/B-LL 32.5 R, the photocatalytic additive used is the commercial TiO2 CM added in the amount of 4% with respect to the cement content. Hereafter, the results obtained on two analyzed samples are reported: i.

a standard mortar with 4% of TiO2 CM (CEM Active);

ii.

a not active standard mortar without photocatalysts, as reference (CEM TQ).

The results reveal that the sample CEM Active reaches R4 equal to 84.1% (Figure 6.2) this means that is a photocatalytic material with high self-cleaning ability; on the contrary the reference sample shows no self-cleaning ability (Figure 6.3). Some additional treatments, not included in the norm procedure were carried out to evaluate the durability and the behavior of the self-cleaning surfaces. After the irradiation of 4 h the sample were exposed to the following different treatments: (i)

12 hour in dark condition: the a* value increase from 7 to 15,6 (increasing in a* value means higher amount of rhodamine);

(ii)

12 hour at 60°C without UV radiation: the a* increased from 7 to 15,5;

(iii)

the washing process: the a* increased from 7 to 11,2.

When the sample is irradiated again for 4 hour the a* value return to be very low, namely 6.3. This means that the rhodamine molecules are still present in the pore structure and without the UV exposition (enhanced by the treatment), they will be evident on the surface.

76


Figure 6.2 a* value of the rhodamine test. Evaluation of the effects of the treatments.

Figure 6.3 Rhodamine discoloration test on white cementitious mortar. The cement non-active at the left side, CEM TQ (a) before irradiation and (c) after irradiation. The photoactive cement at the right side, (a) before irradiation and (b) after irradiation.

6.2

A case study of self-cleaning concrete

Two photocatalytic concrete tiles were produced in large scale (60x60x10cm), in order to evaluated the ability to maintain aesthetical properties in real condition of exposure: (i) with photocatalytic white cement CEM II/B-LL 32.5R with 3% of TiO2 CM and (ii) without photocatalysts - a standard concrete used as reference. The tiles were exposed horizontally to environment in the plant of Buzzi Unicem in Trino (VC) until 2 years. The measure of the brightness by means of colorimetric

77


CIELAB instrument reveals that, although at the beginning the non-photocatlytic tile had higher L* value, after two year a significant decrease in brightness is present. On the contrary the photocatalytic one maintain the L* value69. It is possible to appreciate the qualitative difference in the pictures in Figure 6.5.

Figure 6.4 Photocatalytic concrete tiles exposed to environmental condition in the Buzzi Unicem plant of Trino. On the left the reference tile (not photocatalytic active), while on the right the photocatalytic concrete tile.

Figure 6.5 The trend of the L* value of CIELAB (lightness index) measured on the surface of the photocatalytic concrete tile. Black line = photocatalytic concrete tile – Red line = reference concrete not active. It is evident that the photocatalytic concrete remained whiter than the reference one after two years of exposure.

69

Before measure

78


6.3

Self-cleaning vs air purification ability

The abilities of NOx reduction and the self-cleaning responds to specific surface properties of the cementitious materials. As observed, high surface area and very rough surface increase the pollutant reduction ability, due to the higher active TiO2 on the surface. We cannot state the same when we want to explore the self-clean effect of a concrete surface. In the case, high porosity and roughness go in the opposite direction. This because soils or dyes can penetrate inside the porosity of the material, in which the UV light do not reach the TiO2 molecules and subsequently the photocatalytic degradation, does not occur. The cleaning effect is promoted if the materials has smooth surface with low porosity, thus low surface area the cleaning effect is promoted due to low surface permeability (Figure 6.6).

Figure 6.6 Comparison of the features, normative and requirements in the NOx abatement ability and self-cleaning effect.

Figure 6.7 Characterization techniques for the main photocatalytic activities.

79


Conclusions

I

n this research, different photocatalytic building materials were developed and produced in order to investigate the air purification ability in degradation of NOx pollutants and the capacity of self-cleaning by concrete surfaces. The concepts

of basic research were transferred to real application case studies. In addition, the potentialities and barriers of the technology has been evaluated. The effectiveness of new functionalized floating nano-TiO2 photocatalysts, produced at industrial scale, was measured in term of NOx abatement. The concrete mix design revealed to be a crucial aspect in the efficiency of photoactive surfaces regarding potential applications. It was demonstrated that high surface roughness and high surface area improved significantly the NOx reduction ability. Nevertheless, this aspect is in contrast with the low porous and smooth surfaces suitable for self-cleaning applications. The micro-structured photocatalysts in cement is an effective alternative to the more used nano-TiO2. It was demonstrated, that the photocatalytic efficiency of cementitious materials is influenced by several experimental considerations, such as curing treatment, water/cement ratio, surface scraping and superplasticizer addition. Concepts of basic academic disciplines were applied as innovative supports: -

an experimental design was used to evaluate the influence of experimental conditions (flow rate and NOx concentration) on the pollutants abatement;

-

computational modelling was used for the developing and testing the photocatalytic demonstrator;

-

mechanochemistry approach was applied for the synthesis of Vis-active photocatalysts, as an industrial potential technique to produce Vis-active additives at lower costs and using equipment potentially present in the cement industry.

An economical assessment on the produced photocatalytic cementitious materials was performed through the costs assessment index (CAI). This index, especially developed on this purpose, permits the evaluation of the additional costs due to photocatalysts addition, related to the performance and specific applications. The most convenient products from economical point of view are pervious concretes and mortars one with the commercial micro-structured pigment (CM pigment). 80


Efforts are still needed to measure the ability to reduce pollutant in real scale, and the most important aspects is the need of normative to transfer such products into the market. The self-cleaning ability seems to be the most appealing features for potential customers. The ability to maintain the aesthetic properties is a more tangible and measurable feature, than air purification. The tailoring of the surface is in this case the challenge of the scientific research. The SWOT matrix70 (Strengths-Weakness-Opportunities-Threats analysis) below, resumes the impacts of the research, describing the potentiality and barriers to the implementation of the technology after this research findings.

70

The SWOT matrix (Strengths-Weakness-Opportunities-Threats) is a strategic analysis tool for the evaluation of potentialities and barriers in research and development project.

81

SECOND PART

SELF-HEALING CEMENTITIOUS MATERIALS

82

Second Part – Self-healing cementitious materials

7 7.1

SELF-HEALING IN CONSTRUCTION MATERIALS Self-healing phenomena

T

he self-healing ability, commonly present in biological system, in living plants and animals (i.e. the restoration of bones after a fracture or the regeneration of the skin after a superficial shear), is generally not found in

man-made materials. Strength, in more general terms, deals with the ability of a material to sustain a high load without damaging and forming new surfaces. Often, the plastic deformation of a material, can lead to internal defects, which can grow into cracks and finally cause the degradation of the product. Recently, materials scientists, inspired from the observation of the biological mechanism, started to change the design philosophy and to develop materials based on the damage prevention, able to repair themselves when damage occurs. The design concept of damage prevention has proven to be a very useful productive concept. However, the formation of damage during use can never be completely excluded; it also means that structures made of current material invariably need periodic inspection to monitor possible damage development. Furthermore, any observed damage requires actions and costs, sooner or later. In self-healing materials approach, the alternative concept of damage management is introduced. The concept is based on the notion that the formation of damage is not problematic as long as it is counteracted by a subsequent process of “recovery” or “healing” of the damage. The final material’s performance of self-healing materials depends on two contingent processes: the rate of damage formation versus the rate of damage recovery or healing. In order to transform the concept of self-healing materials from idea to practice, it is worthwhile to discuss some crucial steps. A first step is the acceptance of damage. Failure of a material is not regarded as a total failure state, but much more as the gradual process is often the reality. A next step is the necessity of a trigger mechanism. When a damage occurs by an action, a recovery reaction against this action should be triggered at some point to start the healing process.

83


Another issue, is the need trigger mechanism has initiated a healing process, some sort of transport is required; this means that the healing agent must move to the location of the damage in order to perform its healing function. The healing agent must fill, replace and/or react with the materials in order to counteract the damage in the material. This self-healing phenomenon is present in all the self-healing materials. The majority of the studies addressed to the development of self-healing materials are in the embryonal state. Nevertheless, some of these technologies are already a commercial reality such as the self-healing coating developed by the automotive sector (Nissan Motor Co. 2005), by the polymeric materials industries (self-healing tires) and for aerospace applications. All the self-healing materials can be included in the definition of “smart materials”. Smart materials are engineered materials, which are able to provide a unique beneficial response, when a particular change occurs in the surrounding environment71. The smart materials are an example of resilience material technologies. Examples of smart materials include piezoelectric materials, magnetostrictive materials, shape memory materials, temperature-responsive polymers (able to change color with the temperature), and smart gel able to shrink or swell by factor of up to 1000 in response to chemical or physical stimuli.

7.2

Why self-healing in cement-based materials?

Historically, the construction materials have been designed with fixed features and the degradation of the materials has always been considered inevitable. The durability of the construction material is a crucial parameter for the product development. An enhanced service life of concrete structure will reduce the demand for new structure. This result in the use of less raw materials and an associated reduction in pollution, energy consumption and CO2 production. Statistics report an enormous amount of money spent by the society due to the lack of the quality and durability of concrete structures. The design life of a building structure is estimated

71

M. de Rooij, K. Van Tittelboom, N. De Belie, N. Schlangen, Self-healing Phenomena in Cement-based materials, State-of-the-Art Report of RILEM Technical Committee 221-SHC.

84


to be around 50 years. Statistical analysis in the UK demonstrated that around 45% of the costs in the construction sector derive from the maintenance. Apart from saving direct costs for maintenance and repair, the saving due to reduction of the indirect costs are generally most welcomed by the owner. The initial costs will be substantially higher than that of structures made by with traditional concrete mixture. The absence of maintenance and repair costs, however, could finally results in a financially positive situation for the owner.

The cracks that can occurs in a concrete are due to several factors such as compressive structural load, plastic shrinkage, drying shrinkage, thermal effects, catastrophic events, etc. The microcracks due to degradative phenomenon are commonly diffused. The chemical or mechanical degradative effects could improve the capillary porosity or open channels, leading to the entry of seawater, acid rain or gases as CO2. Considering the heterogeneity of concrete uses in civil engineering, the environmental conditions, the chemical composition of the infiltrated agents, and the different degradative effects occurring severely threaten the durability of the structure. Infrastructures such as high ways or tunnels need a continuous maintenance, which in certain case means intensive maintenance works. Furthermore, the costs for the diagnostic and restoring process could be prohibitive. In several circumstances, a self-restoring of the mechanical properties by means of self-healing of microcracks without human intervention is an actual chance to improve the quality standard and the safety of the civil infrastructures. In the latest years, several technological approaches were studied and proposed in the research field of self-healing construction materials, nevertheless additional efforts must be done in order to transfer the know-how from the scientific research to the industrial applications.

7.3

Evaluation of the healing effect

Until now, standards normative to evaluate the self-healing effect in building materials do not exist. The characterization techniques are strongly correlated to the type of self-healing mechanism.

85


Specific techniques have been used in self-sealing and self-healing research, in order to verify cracks sealing (recovery against environmental actions) or crack healing (recovery against mechanical actions). For recovery against environmental actions, it turns out that permeability tests are often applied, while for recovery against mechanical actions the regain in strength by mechanical testing is the preferred choice. Traditionally, concrete quality has been evaluated by both non-destructive and destructive mechanical tests. The non-destructive testing (NDT) and monitoring methods are potential techniques to characterize the effect of self-healing mechanisms, both in lab-scale tests and up to real constructions requires. Examples of NDT72 methods are: ultrasonic measurement, microwaves and radar techniques, tomography, active infrared thermography, etc. Among the techniques used to examine crack healing also microscopic methods are used to characterize the damage and assess the degree of self-healing (fluorescence microscopy, polarized light microscopy, optical and electron microscopies). In order to verify the recovery against environmental action, water permeability, air permeability, capillary water adsorption, ultrasonic measurements, microscopies and computed tomography are often applied. To measure the recovery against mechanical actions (which means a regain in strength and stiffness) of the healed crack, cracks have to be made in the concrete in controlled way. For this purpose a three-point-bending tests, four-point-bending tests and compressive resistances are often chosen and performed on prismatic concrete specimens produced on purpose.

Three-point-bending test. In this experimental research, the three-point-bending test was applied for evaluating the self-healing effect in mortar sample. For this reason, the test is hereafter described in details. The deformation is measured with two linear variable differential transformer fixed to the bottom of the specimen (in the centre). i.

the deformation measured with these transformer gives a bending strain at bottom of the specimen. This value is a measure of the crack opening. In the

72

C. U. Grosse, K. van Tittelboom and N. De Belie, Non-destructive testing techniques for the observation of healing effects in cementitious materials – an introduction, Proceeding of fourth International Conference on Self-healing Materials (2013) 60.

86


test, the loading is stopped at the predefined crack width. These specimens are loaded in deformation control up to a predefined displacement, Crack Mouth Opening Displacement (CMOD) control mode. ii.

Then, the specimen is unloaded and taken out of the machine. In order to measure the recovery of the healed crack, the specimens are tested again in three-point-bending after a certain period. The mechanical recovery have to be compared with a cracks free sample. The references sample must be stored in the same environmental conditions.

iii.

The samples are tested again, after the healing time recovery, under tensile deformation. The maximum load reached in the second stage of the test is taken to be maximum flexural strength stress for an unhealed crack.

The regain in mechanical properties in the three-point-bending test after a healing period, serves as a quantification of the healing. The self-healing effect was evaluated after each re-loading stage through some performance recovery indices, by analogy with a common approach reported in the literature73. In particular, a Load Recovery Index LRI was defined as:

LRI % 

Pr  Pu  100 Pp  Pu

where Pr is the peak load obtained during the re-loading stage, Pp is the peak load reached during the pre-loading stage (i.e. during crack creation in the intact material, corresponding to the material maximum strength) and Pu is the residual load obtained at the moment of unloading preceding the re-loading stage. In addition, a Stiffness Recovery Index SRI was calculated as:

SRI % 

Sr  100 Sp

where Sr is the stiffness of the specimen during the re-loading stage and Sp is the stiffness of the intact specimen (i.e. the stiffness recorded in the pre-loading stage, prior to crack creation). In all cases, the stiffness was defined as the slope of the least-square linear fitting curve of the portion of the Load vs. CMOD data between 7,5% and 75% of the corresponding peak load.

73

D. Homma, H. Mihashi, T. Nishiwaki, Self healing capability of fiber reinforced cementitious composites, Journal of Advanced Concrete Technology 7, (2009) 2: 217-228.

87


Figure 7.1 Three-point-bending test scheme illustration. The difference between the maximum load (red line) and the re-loaded sample strength (blue line) defines the load recovery. The difference in the green area illustrated is related to the stiffness recovery.

Figure 7.2 Flexural strength of a pre-cracked mortar sample tested by three-point-bending test.

7.4

State of the art of self-healing in cementitious material

The first association with the word “damage” is often “broken”, especially in the field of science with very brittle materials like cement-based materials. This has resulted in a wave of research on cracks and fractures. There are many types of damage that benefit from healing. Taking over in the self-healing concept, damage can occur when the applied load during the service life is higher than the design load, or when damage is due to environmental actions attacks. The consequence is, in both cases, that the resistance of some sort of the recovery or self-healing is necessary. When the regain is provided through the mechanical design concept itself, it is defined as self-healing. Although concrete is a common material with a long history, with regard to autogenous healing in term of engineering evolution, we are just standing at the beginning of quantitative approaches with the aim of designing concrete materials and structure that have appropriate healing functions.

88


There are many approaches to healing in concrete; some of them are often covered by one technical term whereas they are different from each other in the mechanism. Although, the definition of self-healing should not be too detailed, thereby making the necessary definition too complex. This result in the following definitions of RILEM: 

Self-healing: any process operated by the material itself involving the recovery and hence improvement of a performance after an earlier action that had reduced the intrinsic performance of the materials; o Autogenic: the self-healing process is autogenic when the recovery process uses materials components that could otherwise also be present and not specifically designed for self-healing (own generic materials); o Autonomic: the self-healing process is autonomic when the recovery process uses materials components that would otherwise not be founded in the materials (engineered addition).

At the national level, applications on industrials scale of self-healing construction material are not known. In this contest, the development of self-healing technologies on building structures is a highly innovative research field.

7.4.1

Recovery against environmental action

Autogenic self-healing. In the recovery against environmental action it is possible to distinguish between physical, chemical, and mechanical causes (Figure 7.3). The physical healing phenomenon occurs when water is adsorbed by the hydrated cement pastes and reaches the space between the constituents of the hydrated particles, causing the swelling of the matrix near the crack faces. In the chemical process the continued hydration of the anhydrous particles leads to new reaction products in the free space. The continued hydration cannot be responsible for a complete self-closing of cracks, but, assuming a small crack width of 0.1 mm and assuming a simultaneous action of swelling and hydration, a partial but relevant self-healing of the fracture can be envisaged. Additionally, the formation of calcium carbonates and the growth of crystals on the crack-free faces is a further chemical healing process. Calcium ions from the pore water of concrete (Ca2+) react with the carbonate ions in water (or in

89


air) in the cracks (CO32-) and combine to CaCO3, which precipitates in the crack. This reaction depends on temperature, pH, and concentration of the reactants74. The mechanical causes can be due to the presence of fine particles in the water which leaks through the cracks or/to the fracturing of smalls concrete particles from the cracks surfaces. Concrete is composed by hydrated products (C-S-H, CAH, CAH, calcium hydroxide, etc.), additions (i.e. additives, filler), coarse aggregates and very fine materials from aggregates. C-S-H is slightly soluble and will supplies calcium ions when the ions from calcium hydroxide (soluble in water) are exhausted. The supply of calcium ions from the hydrated phases is higher in Portland cements than in other types of cement. A known example of autogenic healing is the case of the Amsterdam 18th century bridges, which shows healing effect due to calcite formation between the bricks; in this case, the high water supply as a promoter of hydration and particle filling of the crack was the cause of such high self-healing process75. A drawback on this healing process consists in the fact that without water in the crack no healing process is started (water is needed for the chemical reactions and for transports of fine particles).

Figure 7.3 Different causes that can lead to autogenic self-healing [M. de Rooij et al (Eds.): Selfhealing phenomena in cement-based materials, RILEM 11, pp. 65-117. DOI: 10.1007/978-94-0076624-2_3].

74

C. K. Edvarsen, Water permeability and self-healing of through-cracks in concrete, DAfStb Bull 455 (1996). 75 S. Qian, J. Zhou, M.R. de Rooij, E. Schlangen, G. Yea, K. van Breugel, Self-healing behavior of strain hardening cementitious composites incorporating local waste materials, Cement & Concrete Composites 31 (2009) 613–621.

90


Autonomic self-healing. In this section, two examples of autonomic healing processes are reported. The first example consists in the cracks sealing, enhanced by using superabsorbent polymers (SAPs). These highly hydrophilic cross-linked polymers have the ability to absorb a significant amount of water from surrounding environment and to retain liquids within their structure. The main advantage is the autogenous shrinkage reduction; the drawbacks are (i) the decreasing of resistances and (ii) the effect of SAP as shrinkage reducer is strongly dependent from curing conditions. The SAPs are also applied in epoxy coating protective films for steel reinforced against corrosion, in order to absorb water avoiding the contact with the metal. This research is a part of the “SheMat”76 project. The second example is the application of microbial induced carbonate precipitation (MICP) for cracks sealing in concrete (bio-concrete). The biomineralization process of calcium carbonate is “induced” by the bacteria metabolic process. Different type of bacteria, as well as abiotic factors seem to contribute, in a variety of ways, to the precipitation of calcium carbonate in a wide range of environment. In this case, the efficiency of CaCO3 precipitation is evaluated by the consumption of oxygen. The application of bacteria as self-healing agent is a part of larger research program at TUDelft. In the last November 2014, Green-Basilisk, a spinoff of the TUDelft, has been created aimed to improve the commercial potentiality of this system. Bacterias and their nutrients are usually encapsulated in order to protect them from basic pH of cement. The first generation approach consisted in encapsulate the bacteria in light weight aggregates (LWA); up to now the second generation application, optimized inside the “HealCON” project77, consists in creating pellets of compressed powder together with bacteria78. An application in a real scale of the bio concrete (and natural fibres) was developed in the production of irrigation concrete channel in Ecuador in 2013, using bacteria encapsulated in LWA79. In 2014 V. Wiktor and H. M. Jonkers, presented a field performance in a parking garage of the bacteria-based repair system for concrete. The bacteria-based system 76

www.shemat.eu/ www.healcon.ugent.be/ 78 V. Wiktor, Recent advances on the development of bacteria based self-healing concrete for fullscale outdoor application, proceeding in “Self-healing materials-from concepts to the market” EMRS2015 Fall Meeting 2015. 79 www.youtube.com/watch?v=aJpusHsssJQ 77

91


has been sprayed onto the surface of cracks and on concrete pavement. The cracksealing efficiency and improvement of frost salt scaling were assessed by water permeability and freeze/thaw resistance tests respectively. The results revealed that as only cracks that had not been treated with the bacteria-based repair system were still heavily leaking80. The bio-based healing system is a considered an economic, suitable healing system for concrete, nevertheless scarce information on bacterial durability, and the influence of bacteria encapsulated on mechanical properties are nowadays available. The water supply is a fundamental requirement for the healing effect. The water activates the metabolic reaction and the transport of nutrients. This aspect limits the use of bacteria healing system to applications with a continuous water supplying.

7.4.2

Recovery against mechanical action

Autogenic self-healing. The main mechanism of autogenic self-healing of cracks is the production of calcium silicate hydrated (C-S-H), which provides further strength to hydrated cement paste and consequent regain of mechanical properties. During cement hydration, some grains of cement containing alite and belite do no fully react, resulting in unhydrated cores surrounded by hydrated C-S-H and CH material (a natural encapsulation of reactive minerals). During cracking, these naturally encapsulated particles start to hydrate when exposed to water. This curing causes a volumetric expansion capable of completely closing microcracks. The continued hydration of unreacted cement is present in all binders based on Portland cement. For cementitious systems containing fly ash, blast furnace slag, silicate fume or clay, the pozzolanic reaction may also provide a degree of self-healing capacity. In fact, in alkaline environments, silicate species can dissolve creating silicic acid (H4SiO4) deprotonated units (silicate ions). Silicic acid can react with portlandite developing CS-H. The C-S-H produced during the pozzolanic reaction can then heal the cracks in the same way as the C-S-H produced from the hydration of anhydrous phases81. It is fundamental to highlight that without water, autogenic self-healing cannot occurs.

80

V. Wiktor, H. M. Jonkers, Field performences of bacteria-based repair system: pilot study in a parking garage, Case Studies in Construction Materials (2015). 81

M. de Rooij, K. Van Tittelboom, N. De Belie, N. Schlangen, Self-healing Phenomena in Cement-based materials, State-of-the-Art Report of RILEM Technical Committee 221-SHC.

92


Other examples of autogenic healing occurs with the addition of fibres such as (Polyvinil Acetate, Polypropilene, Polyethylene are some example of polymeric fibres). Fibres are usually used in concrete to control cracking due shrinkage. The fibre-reinforced concrete (FRC) contains fibrous materials, which increase its structural integrity and adding specific properties to concrete82. L. Ferrara et al (2015)83 used innovative natural fibres (sisal fibres) as promoters of autogenous healing in high performance fibres reinforced concrete. Thanks to their porous hierarchical microstructure, natural fibres can adsorb water, either in a pre-saturation stage or in the mixing one or even when into a cement based matrix. The natural fibres are able to create a porous network through which the moisture can be distributed throughout the cementitious matrix and activate the delayed hydration reactions, which, together with carbonation ones, can be responsible of the autogeneous healing of cracks84. Autonomic self-healing. In the autonomic healing phenomena the main technique used is based on the approach of encapsulation of the healing agent before addition to the cement matrix. In this case, the healing agent is released inside the cracks by the breakage of the capsules, repairing the damage. The release of the healing agent can occurs by mechanical activation, that means for instance by cracking the capsules by mechanical impact. However, the trigger mechanism may also be a physical action, such as release of heat. Other methods use vascular systems as an external supply of the healing agent. Typically, glass tubes or polymeric brittle materials are used on this purpose.

7.4.3

Encapsulation approach

The encapsulation approach of healing agents was for the first time published by the Prof. Scott White from University of Illinois in 200185. The encapsulated systems are

82

Fibers in a concrete can increase some features such as the freeze-thaw resistance, improve ductility, improve resistance to explosive spalling in case of sever fire improve abrasive resistance, improve structural strength and reduce steel reinforcement requirements. 83 L. Ferrara, S. R. Ferreira, V. Krelani, M. della Torre, F. Silva, R. T. Filho, Natural fibre as promoters of autogenous healing in HPFRCCS: results from on on-going Brazil-Italy cooperation, in proceeding of DSCS2015, Bologna. 84 “EnCoRe” Project (FP7-3 PEOPLE-2011-IRSES n. 295283; www.encore-fp7.unisa.it). 85 E. Schlangen http:/www.rilem.net/gene/main.php?base=8750&gp_id=228.

93


formed on capsules (tubular or spherical shaped) in which the core is composed by a healing agent and an external shell. Hollow fibers, or glass tube, containing a polymeric healing agent, are also widely present in literature86-87. In this methodology (in an ideal self-healing system), the damage or crack must be able to trigger the release of the encapsulated healing agent. This means that the wall of the capsules should be able to prevent the inner agent from undesirable release and on the other hand be sensitive enough to the pressure generated by the damage or cracks, which could ensure the rupture of the wall88. Hollow fibres (pipette or tubes) and microencapsulated capsules both share the same concept and belong to the same encapsulation family, of which the only difference is the healing agent carrying system. The encapsulation approach for healing agent delivery is widely treated in literature; nevertheless, it is worth comparing the advantages of this approach to the other healing strategies present in literature (Figure 7.4). In the use of hollow tubes or capsules, the delivery system can respond to fracture at many different locations as the capsules are dispersed in the matrix. However, a successful manufacture of desirable microcapsules for the application in cementitious materials is often not so straightforward. The bond strength at the interface between the microcapsules and the matrix is often a concern. If the strength of the shell is higher than the interface with cement matrix, the microcapsules would not crack after the initialization of the cracks. Then, no healing agent will be released. It should also be noted that the possible negative effect of mechanical properties of the construction material with the introduction of hollow fibres or capsules, whose dimension and fraction matters, can occur.

86

M. M. Escobar, S. Vago, A. Vázquez, Self-healing mortars based on hollow glass tubes and epoxy–amine systems, Composites: Part B 55 (2013) 203–207. 87 K. V. Tittelboom, N. De Belie, D. Van Loo, P. Jacobs, Self-healing efficiency of cementitious materials containing tubular capsules filled with healing agent, Cem Concr Compos 33 (2011) 497– 505 88 M. Wua, B. Johannesson, M. Geiker, A review: Self-healing in cementitious materials and engineered cementitious composite as a self-healing material, Con. Build. Mat. 28 (2012) 571-583.

94


Figure 7.4 A summarized comparison between different strategies.

The synthesis of an encapsulated delivery system can be theoretically done with different techniques and approaches such as water-oil emulsions, dripping, extrusion process, catalytic synthesis, copolymerization of the shell, spray deposition, evaporation of the solvent, methods inspired from pharmaceutical sector, etc. Each method have some advantages and drawbacks that will be not be discussed here for sake of brevity. The majority of these methods are optimized and performed in the lab scale; for this reason, efforts must be done to address the research to the large scale in order to test the efficiency of the delivery systems produced in concrete prototypes. The applied approach and the synthesis method depend also on the healing agent used. In the literature, among the most used healing agents it is possible to include epoxy resins, cyanoacrylates, methyl methacrylate, bi-component resins, polyurethane and sodium silicate. 95


In the majority of cases, single-component polymeric resins, such as cyanoacrylates and acrylics, are preferred to multi-component healing agent, for example epoxy resins80, or polyurethane foams81, in which the polymerization occurs only when all components react together simultaneously inside a crack, adding a further complexity to the system89. Some of the materials used for the encapsulation (i.e. for the shell formation) are polyuria-formaldehyde, melamine-formaldehyde, polyurethane, epoxy resins, paraffin, glass or polymeric tubes.

The microencapsulation has been constantly improved, modified and adapted for a variety of purposes and uses not only restricted to self-healing technology. It has become an example of a knowledge-intensive and dynamic technology, characterized by a rapid growth of patent applications, reflecting industrial research and development, as well as by an increasing number of scientific articles, deriving from the basic research. In addition to the graphic and printing industries, microcapsules have been used for pharmaceutical (i.e. drug delivery) and medical purposes, in cosmetic and food products, agricultural formulations, as well as in the chemical, textile and construction materials industries, biotechnology and waste treatment90

89

K. V. Tittelboom, D. Snoeck, J. Wang, N. De Belie, Most recent advances in the field of selfhealing cementitious materials, In: Proceeding of fourth International Conference on Self-healing Materials (2013) 406. 90 B. Boh, B. Šumiga, Microencapsulation technology and its applications in building construction materials, RMZ – Materials and Geoenvironment 55-3 (2008) 329-344.

96


Figure 7.5 Growth of new patent documents and scientific articles on microencapsulation (CA Plus database) [Boh et al. (2008)].

Looking at scientific literature and patents, there are numerous possibilities of adding microencapsulated active ingredients into construction materials, such as cement, lime, mortar, concrete, sealant, paints etc. A summary of applications is reported in Figure 7.5.

97


Figure 7.6 Applications of microcapsules in building construction materials. Elaboration from B. Boh et al (2008).

98


8

THE “DualCEM” PROJECT: ENCAPSULATION OF HEALING AGENT

T

he self-healing in concrete and its complexity as described above were approached in the framework of the research project “DualCEM: Development of self-healing cementitious materials with high durability”.

The DualCEM project, ended in 2014, was financed by Regione Piemonte91. The project was carried out through a fruitful cooperation between Politecnico di Torino, as an academic subject, and several companies (see Figure 8.1) as additional Research & Development subjects or end-users. The project aimed to investigate two key technological factors of a self-healing concrete system: on one side the efficiency of different healing agents and, on the other side, the potential of the encapsulation technology approach. The experimental activities carried out within the project can be summarized as follow: 

research and development on resins and chemical agents to be used as healing agent in construction materials;



characterization under the mechanical and chemical point of view of the efficiency of the healing agents used;



development of encapsulated systems: extruded cementitious hollow tubes and spherical macrocapsules;



development and realization of automatic and large scale set up for the synthesis of encapsulated systems;



evaluation of environmental impact;



development of a self-healing concrete prototype.

91

POR FESR 07/13 Misura I.1.3 “Poli d’innovazione” III.

99


Figure 8.1 DualCEM Project partners.

Concerning the first aspect, different healing agents such as sodium and potassium silicate, acrylic resins and natural tannins were tested and evaluated in terms of adhesion and mechanical strength properties. All these agents are mono-components and are able to cure in presence of moisture (sodium and potassium silicate) or reacting with cement components (tannins). In particular, a specifically synthesized tannin resin (supplied by Silva Team s.r.l.92) was evaluated as an innovative healing agent for cementitious materials. This optimized resin is a solution with tannins extracts93, water and methyl-ethylen glycole (MEG). The resin is able to polymerize in presence of the basic pH of cement, acting as a potential healing agent. The efficiency of the healing agent was previously evaluated in term of “adhesion property” (in collaboration with the Politecnico of Torino research group) by means of the three-point bending tests, and it will be discussed later.

In the DualCEM project, the distribution of the healing agents was basically focused on encapsulation technologies. Both spherical and tubular shape capsules were synthesized and the ability as suitable healing agent containers was evaluated looking at the potentialities and the barriers in construction applications.

92

Silvateam is an international leading company in the production of tannins extracts. The tannin resins are extracts of acacia or quebracho plants. The addition of mSiO2 improve the polymerization of the tannin polymers in basic pH condition. This aspect is interesting for new applications, nevertheless the real chemical mechanism of the polymerization in not fully explained in literature [A. Masson, A. Merlin, A. Pizzi, Comparative Kinetics of Induced Radical Autocondensation of Polyflavonoid Tannins. I. Modified and Nonmodified Tannins, J Appl Pol Sc Vol 60 263-269]. Applications of tannins are i.e. in the in renewable resources adhesives, as rust converter for iron or iron alloys, as natural additions in red and white wine and food industry and in the tan of leathers. 93

100


Two different respective technologies of healing agent containers preparation were adopted. The first technology is based on the encapsulation by a dripping process. As it will be discussed more in detail in the following paragraphs, the spherical beads (millimeter size) were produced through a technical equipment specifically designed for this purpose, developed with the collaboration of Varnish Painting Technology. In the second technology, extruded cementitious hollow tubes (with internal diameter of 7.5 mm) as containers and vehicles for healing agents were synthesized by an extrusion process of an optimized cement paste by the Department of Applied Science and Technology (DISAT) of the Politecnico of Torino. In the laboratories of the Politecnico of Torino, preliminary tests were performed, in order to define the main features of the tubes and to verify their durability. The ability of crack healing was evaluated in term of recovery of mechanical properties by means of the three-point bending tests. In the Figure 8.2, the organization of the project and the main milestones are illustrated.

101


Figure 8.2 Research activity in the DualCEM project.

8.1

Healing agent characterization

Adhesion property. In order to compare the efficiency of the healing agents used in this research project, an adhesion test was developed on this purpose. In this test, the flexural strength of cement paste sample94 (CEM I 52.5R, supplied by Buzzi Unicem), was measured by three point bending test until the complete failure of the prism, using an electrochemical testing system (MTS Insight 1 kN, produced by MTS System Corporation). After the complete failure, the healing agents were brushed by hand on the fracture surfaces and the two fragments were kept joined again with an adhesive tape for 7 days in order to restore the original prims. Then the tape was removed and the flexural strength ability was tested again to evaluate the binding ability of the

94

The cement paste (7,5x2x2 cm) was produced with a water/cement ratio of 0,4 using 1% of plasticizer (Dynamon SP1). The samples were cured 7 day at 90%RH and 7 days in lab environment.

102


healing agent. Depending on their capacity to restore (part of) the original strength of the prism, the healing agent has been considered suitable (or not) for a concrete repairing system. In this test, the binding ability, and consequently the recovery of mechanical properties, can be influenced by some factors such as (i) the amount of agent brushed on the surface (difficult to control), (ii) the porosity and the sorption capacity of the surface,(iii) the viscosity of the solution and (iv) the curing conditions of the samples and the chemical reactivity between the agent and the cementitious matrix. The results revealed that sodium silicate has the higher bending strength recovery in comparison to the copolymer of ethyl acrylate and methyl methacrylate (EA/MMA), tannin resin and potassium silicate has the higher bending strength recovery95. Furthermore, the reassembled samples with sodium silicate showed a higher bending strength than the original sample, nearly double on average (see Figure 8.3). In the light of these results, sodium silicate solution was selected as the referencehealing agent and promoted to the subsequent encapsulation step.

Figure 8.3Results of the adhesion test by means of three point bending test of a Portland (CEM I 52.5 R) using sodium silicate [Formia et al. (2015)] 96.

95

A. Formia, S. Irico, F. Bertola, F. Canonico, P. Antonaci, N. Pugno, J-M. Tulliani, Experimental analysis of the self-healing cement-based materials incorporating extruded cementitious hollow tubes, J Intell Mater Syst Struct (2015). 96 A. Formia, P. Antonaci, S. Irico, F. Canonico, J-M- Tulliani, Extruded cementitious hollow tubes for healing agent delivery, in proceeding of DSCS 2015, Bologna.

103


8.1.1

Sodium silicate as healing agent: insight of the reaction with Portland cement

Sodium silicate, unlike other healing agents or sealants, which repel water (i.e. silanes, silicones, etc.) or act as physical barrier (i.e. epoxies, vinyl, polyurethane, etc.), is adsorbed by concrete surface and is supposed to react with calcium hydroxide (portlandite) to form C-S-H gel. Several literature works consider sodium silicate as a suitable healing agent in a cement based material97-98-99-100. Pelletier et al. (2011), have been demonstrated that a self-healing concrete with polyurethane encapsulating sodium silicate inside, shows mechanical properties retrieval around 26% of the original strength after flexural test, versus a recovery of ordinary concrete around 10% (water curing). Moreover, it has been observed that samples with the encapsulated sodium silicate had a significant amount of corrosion inhibition, compared to reference samples. In this case, sodium silicate forms a passivated layer able to protect the steel debars over time. Dong et al. (2013), proposed sodium silicate powder encapsulated in a polystyrene shell as delivery system in an engineered cementitious composite (ECC). This approach needs more supplementary water or moisture available to dissolve sodium silicate and promote the reaction with the cement matrix. Huang at al. (2011)101, investigate the mechanism of self-healing by sodium silicate solution. In this study sodium silicate was encapsulated and embedded in an ECC. Three point bending tests were carried out to evaluate the efficiency of the healing system in term of load recovery index. The self-healing phenomenon was observed by SEM and X-ray microanalysis (EDS) to investigate the reaction and the chemical species involved. They observe that when the samples cracked, the sodium silicate solution is released to promote the self-healing. The recovery efficiency reduces as

97

M. Pelletier, R. Brown, A. Shukla, A. Bose, Self-healing concrete with a microencapsulated healing agent. Technical report. University of Rhode Island, (2011) Kingston, RI. 98 H. Dong, H. Huang and G. Ye, Inorganic powder encapsulated in brittle polymer particles for selfhealing cement-based materials, In: Proceeding of fourth International Conference on Self-healing Materials, 2013, p. 123. 99 M. Pellietier, A. Bose, Patent Application Publication No. US2011/0316189 A1. 100 A. Formia, S. Terranova, P. Antonaci, N. Pugno, J-M Tulliani, Setup of extruded cementitious hollow tubes as containing/releasing devices in self-healing systems, Materials 8 (2015) 1897-1923. 101 H. Huang, G. Ye, Application of sodium silicate solution as self-healing agent in cementitious materials, International RILEM Conference on Advances in Construction Materials Through Science and Engineering, (2011).

104


far as the concentration of the solution decreases; this can be attributed to the amount and cohesive property of the healing products. Based on the analysis of the Ca/Si and Na/Si ratio by means of EDS analysis, they concluded that the healing products formed in the cracks are the composites of C-S-H and sodium silicate.

Soluble sodium silicate solutions have already been used in several different aspects of cement industry. They were used in the past as set accelerators shotcreting applications102; they can be employed as moisture reducers in the wet kiln process of clinker production; they can be incorporated into cementitious waste forms and they are concrete sealer103. They are often used as alkali activator in geo-polymeric systems.

Even though the mechanism of sodium silicate reaction with a Portland cement paste has been partially expected and evaluated mainly by SEM observations and EDS analyses, the definition of all the reaction products involved, the time-scale of the kinetic and how these reactions leads to mechanical performances is still lacking in literature. During this PhD research, the insights on the reactivity between sodium silicate and a Portland cement matrix are experimentally studied. The binding ability was quantitatively verified the binding ability by means of a “healing agent compressive strength test”, while the chemical species involved in the reactions, the timescale of the reaction and the chemical nature of the products were investigated through solidstate nuclear magnetic resonance (SS-NMR) and in-situ X-ray powder diffraction (XRPD) techniques104. These two techniques are complementarily useful as SS-NMR can assess local changes in the chemical and mineralogical composition of Si and Al phases regardless of their crystallinity, while XRPD, despite its sensitivity only for structurally ordered

102

In special applications, sodium silicate is substituted nowadays by alkali-free shotcreting solutions. The shotcrete is a concrete conveyed through a hose and pneumatically projected at high velocity onto surface, used frequently for tunneling construction technique. 103 J. LaRosa Thompson, M.R. Silshee, P.M. Gill, and B.E. Scheetzt, Characterization of sodium silicate sealer on concrete, Cem Concr Res 27 (1997) 1561-1567. 104 A. Bovio, Ricerca ed applicazione di strategie per l’autoriparazione di materiali cementizi, Master Thesis, Università del Piemonte Orientale (2012).

105


species, is much faster in the detection of phase compositional changes due to hydration and chemical reactions.

Binding ability. In order to quantify the interaction of sodium silicate with cement matrix and evaluate the relation between the sodium silicate amount and the adhesion strength a “healing agent compressive strength” was specifically designed on purpose. A CEM I 52.5R standard mortar105 was ground in a laboratory jaw crusher until all fragments were smaller than 2 mm and the obtained crushed materials was used as a support material for the healing agent strength test. The crushed hydrated material was mixed with the healing agent (sodium silicate106) and water in the ratios reported below and shaped in a mortar mold. Three different sodium silicate water solutions has been used, with fixed ratio between the sodium silicate solution and the mortar (or reference water) to 0.5: 

Sodium silicate (sample A)



Sodium silicate : water (85 : 15) (sample B)



Sodium silicate : water (70 : 30) (sample C)



Water (used as a reference) (sample D)

The samples (40x40x40 mm) were cured at RH 95% and 20°C for 2, 7 and 28 days, until they were tested for compressive strength by means of a compressive press (VT200 Bombardieri).

Figure 8.4 Mixing of the crushed mortar with sodium silicate and casting of the mixtures in 40x40x40 mm prisms.

The results of the healing agent compressive strength test are summarized in the chart of Figure 8.5. The reference samples, prepared by mixing the hydrated grounded

105 106

Prepared according to UNI/EN196, and stored for 28 days at RH 95% and 20°C. The sodium silicate solution used is supplied by Panreac aqueous solution 40%wt..

106


mortar with water showed no binding effect at all; it was even not possible to demold them due to their scarce consistency. The development of compressive strength observed when sodium silicate solutions (a)-(b)-(c) are used, is a clear indication of an interaction between the solution and the cement particles. The reaction highlights a quite fast reactivity, with strong adhesive properties providing significant strength resistances. The Rc values are relevant since 2 days of curing, and increase during ageing, up to 10.9 MPa in the sample with pure sodium silicate (a) at 28 days. Dilution of the sodium silicate solution results in a significant decrease of the Rc in the samples (b) and (c), being lower the amount of sodium silicate available for the reaction. This aspect is in agreement with the results described by Huang at al. (2011) about the load recovery index by encapsulated sodium silicate with different concentrations. This means that less sodium silicate is reacting with the cement in comparison to the sample (a), the reactions products are in these two cases in a smaller amount and it is clearly related with Rc. The compressive strength results reveal that sodium silicate solutions are indeed potentially able to recover the integrity of a building structure after damage by an efficient and widespread crack healing.

Figure 8.5 Compressive strength resistances of the hydrated fragments mixed with the sodium silicate solutions.

107


Chemical reactivity. The chemical reactivity between the healing agent and hydrated cement paste107 (CEM I 52.5R) was investigated in order to define the chemical species involved in the restoring of mechanical performances and the time scale of a potential healing effect. The Portland cement standard paste was produced and stored for 28 days at RH 90% and 20°C, to guarantee a fully hydration degree. After curing, the hydrated paste was ground and the obtained powder was mixed with the sodium silicate solution108 (paste/liquid as 0,5). The evolution of the reaction with time was then followed by means of in situ XRPD109 and SS-NMR110 observing in general 1H,

27

Al and

29

Si

nuclei. For in-situ XRPD analyses, the sample paste was poured in a holder and sealed with a kapton film (Polyimide Film 7,6 μm supplied by LGC standards) to prevent any contact with the environment until 7 days. X-ray diffraction pattern collection was performed after 0, 2, 5, 8 hours and 7 days. At the same time in the SS-NMR measurement, in order to avoid the removal of water in the samples, and to perform experiments actually on the same specimen, cement and water/sodium silicate mixtures were packed in a proper plastic sample holder and inserted in the zirconia rotor. From the mineralogical characterization it is possible to observe that a decrease of the intensities of the peaks of Ca(OH)2 is detected (Figure 8.6); then, a huge modification of the shape of the diffraction pattern in the region around 30° 2θ is recorded. At the In contrast with the “healing agent compressive strength”, in this case a cement paste was used instead of mortar to facilitate the characterization of the hydration species by XPRD and SSNMR avoiding the presence of sand (quartz). 108 The water based solution of sodium silicate as the ratio of sodium silicate: water as 70:30. Higher concentrations have a very fast reactivity and prevent a suitable sample preparation. 109 Mineralogical investigations were performed using a Bruker AXS D4 Endeavor diffractometer working in Bragg-Brentano geometry, equipped with a ceramic X-ray tube KFF (Cu Kα radiation) and a “Lynx Eye” dispersive detector. Mineralogical phase identification was performed through the software EVA 2.0 commercially supplied by Bruker AXS), allowing the comparison to the PDF database. 110 Solid-state NMR spectra were acquired on a Bruker Avance III 500 spectrometer and a wide bore 11,7 Tesla magnet with operational frequencies for 29Si and 27Al of 99.35 and 130.33 MHz, respectively. A 4 mm triple resonance probe with MAS was employed in all the experiments and the samples were packed on a Zirconia rotor and spun at different MAS rate. The 27Al 1D MAS spectra have been acquired on large sweep width with small pulse angle (π/12) to ensure quantitative interpretation. All the 29Si MAS NMR spectra were recorded under high-power proton decoupling conditions. The relaxation delays, d1, between accumulations were 120 and 1s for 29Si and 27Al NMR, respectively. All chemical shifts are reported using δ scale and are externally referenced to TMS for 29Si and to 1.0 M AlCl3 solution for 27Al. 29Si MAS NMR spectra were fitted with DMFIT functions for quantitative deconvolution of overlapping peaks. 107

108


beginning of the test (0 hours), two well defined peaks are identifiable, respectively ascribable to portlandite at 29° 2θ (ICDD 44-1481) and Calcite at 30° 2θ (ICDD 050586). At longer contact times, a hump starts appearing below the calcite peak and increases up to 7 days. According to the literature (ICDD 86-2275), in this region the main peaks of crystalline C-S-H (tobermorite) are observed.

Figure 8.6 XRPD diffraction patterns of the ground cement paste treated with sodium silicate solution diluted 70% wt with water, XRD allowed to define early reaction time scale for self-healing effect.

On the basis of these evidences, it is possible to state that portlandite reacts with sodium silicate: as a result, a pseudo-crystalline C-S-H (characterized by wide and overlapped diffraction peaks) is formed and, as more important feature, the reaction already takes place after few hours that silicate and cement paste are put in contact. The characterization of the reaction with SSNMR proved that the chemical interactions with the cement paste is more complicated than theoretically expected. In fact, the combined study with XRPD/SSNMR study highlighted, aside the reaction with calcium hydroxide, the involvement of Aluminium containing phases, i.e. ettringite and Aluminum monosulfate (commonly named AFt/AFm phases) and unreacted C3S/C2S phases developing crystalline C-S-H/C-A-S-H (tobermorite).

The SS-NMR is a valid support in the investigation of non-crystalline hydrated phases in cement, SS-NMR is able to detect the nuclei of 29Si and 27Al ions. The Figure 8.7 reports a characterization of a Portland cement: the tetraedrical silicates phases and octaedrical aluminates.

109


Figure 8.7 NMR applied in cement materials science. NMR spectra of a Portland cement during the hydration process: (a) 29Si NMR spectra and (b) 27Al spectra.

Figure 8.8 27Al MAS and 29Si MAS SS-NMR spectra of a CEM I 52.5R, sodium silicate treated for 1 days (Sample S_1d), 7 days (Sample S_7d) and 15 days (Sample S_15d) samples. E = Ettringite; M = Monosulfate; TAH = third aluminate hydrate; C-S-H = calcium silicate hydrate. SS-NMR is informative on the local reactivity of silicate species.

27

Al MAS NMR spectra of ground cement paste treated with sodium silicate solution

are shown in Figure 8.8. After one-day treatment, the spectra show: -

an increase in the tetrahedral resonances due to Al substitution in the C-S-H to form C-A-S-H;

-

an increase in the octahedral Al species such as ettringite and monosulfate ;

-

a decrease in the intensity of the calcium aluminate hydrates (TAH).

After long-term treatment (7 and 15 days) with sodium silicate, dramatic reductions in both tetrahedral and octahedral aluminum resonances are observed. Moreover, a new resonance at around 75 ppm started to appear and is attributed to 27Al in aluminate phases. A progressive destruction of the AFt and AFm calcium aluminate phases upon 110


treatment leads to the conclusion that the healing agent attacks the crystalline and semi-crystalline phases. From the 29Si MAS NMR spectrum, it is evident that sodium silicate reaction leads to the formation of new C-S-H/C-A-S-H phases (tobermorite). This is probably due to the consumption of unhydrated C2S and C3S phases by sodium silicate. The line width of resonances due to new C-S-H/C-A-S-H is much narrower and further corroborated the XRPD data in the formation of tobermorite phase. However, sodium silicate further attacks the crystalline and semi-crystalline aluminate AFt, AFm and TAH phases to extract probably calcium and/or aluminum ions for the formation of new C-S-H/C-A-S-H. Consequently, new aluminate phase is formed and its intensity is increased upon treatment for longer duration. Hence, it is possible to state that the treatment of cement paste with sodium silicate leads to the complete hydration of anhydrous phases and accelerates the formation of crystalline/semicrystalline C-S-H phase tobermorite. The development of new C-S-H/ C-A-S-H phase, promoted by the reaction of sodium silicate with Ca ions, is probably the responsible for the recovery of the mechanical flexural strength previously described. The investigation on the reactivity of hydrated cement phases with sodium silicate allowed to define the reaction time scale for the healing effect. Sodium silicate reacts with the cement matrix after few hours that they were put in contact, making it further promising, because of the rapidity of setting, for self-healing purpose. This information is complementary to those obtained by SS-NMR, which, though extremely informative on the local reactivity of silicate species and their quantitative determination, cannot provide results in the time scale of hours due to the low responsivity of 29Si nucleus. In Figure 8.9, the proposed mechanism of reaction between sodium silicate and Portland cement in a crack, where a full filling of cracks is operated by sodium silicate reaction with almost all the mineral phases present along the cement crack surface.

111


Figure 8.9 Proposed mechanism of sodium silicate reaction with Portland cement in a crack.

8.2

Extruded Cementitious Hollow tubes containers

Several types of reservoir suitable for a healing agent have been used in literature: microcapsules, ceramic tubes, hollow porous fibres, hollow glass tube111 and polymeric hollow tubes112. However, up to now, glass tubes have been mostly used to store healing agents. The glass tube break upon crack appearance but they cannot survive the concrete mixing process due to their brittleness. Ceramic hollow tubes, a very brittle materials, were used by Van Tittelboom who observed a better bond strength with cement matrix than with glass tubes. B. Hilloulin et al. (2015) investigated the potentiality in the use of polymeric tubes synthetized by extrusion process such as Poly(lactic acid) (PLA), Polystyrene

(PS)

and

Poly(methyl-methacrylate/n-butyl

methacrylate)

P(MMA/nBMA) to prevent the breakage of the tubes during the mixing.

The DISAT department of Politecnico of Torino developed and produced cementitious hollow tubes (CHT) as delivery system. Cementitious hollow tubes were obtained by extrusion of a specifically designed cement paste113, in accordance with the materials and processes (developed and then published in Formia et al 2015. A low-pressure extruding device composed of a screw extruder and replaceable dies was used to extrude the fresh cement paste. The used

111

M. M. Escobar, S. Vago, A. Vázquez, Self-healing mortars based on hollow glass tubes and epoxy–amine systems, Composites: Part B 55 (2013) 203–207. 112 B. Hilloulin, K., Van Tittelboom, E. Gruyaert, N. De Belie, A. Loukili, Design of polymeric capsules for self-healing concrete, Cem Concr Res 55 (2015) 298–307. 113 A. Formia, S. Terranova, P. Antonaci, N. Pugno, J-M Tulliani, Setup of extruded cementitious hollow tubea as containing/releasing devices in self-healing systems, Materials 8 (2015) 1897-1923.

112


die has a ring shape, with an external diameter of 10 mm and an internal hole of 7.5 mm. After extrusion, the hollow tubes were left in a moist environment for 7 days and later in air for complete curing as suggested in literature for polymer-modified cementitious mortars114. In order to verify the reproducibility of the extrusion process the flexural strength of the tubes of 6 different set of samples was measured (Figure 8.10). The results confirmed the good reproducibility of the production system and a good flexural resistance of the hollow tubes, average 1,7 MPa.

Figure 8.10 The flexural strength of the CHT of different set of specimens production, measured by means of three point bending test. The results evidence the good reproducibility of the production technology. On the left side is reported an empty cementitious hollow tube. By courtesy of Politecnico of Torino.

After the extrusion process, in order to increase to healing agent lifetime and to improve shock resistance of the tube shells, the outer and inner surfaces of the CHT were coated with sodium silicate. The extruded cylinders were cut into tubes measuring about 5 cm in length, so that the volume of each element was approximately of 4 cm3. Sodium silicate used as healing agent was poured inside the tubes with a syringe, and the extremities were sealed with wax. The CHT were cured in lab environment until use. The ability of the CHT to break after a crack recovering mechanical performances in a mortar, was evaluated by means of the three point bending test in crack mouth opening displacement (CMOD) control.

114

L.K. Aggarwal, P.C Thapliyal, S.R. Karade, Properties of polymer modified mortars using epoxy and acrylic emulsions, Constr Build Mater 21 (2007) 21 379–383.

113


The amount of initial crack and the time of the healing are the two main influencing factors in the experimental measurement of the healing effect. The test consists in: i.

preparing standard mortars (CEM I 52,5R) with a single hollow tube manually placed in the center of the prism (perpendicular to the crack opening);

ii.

testing the flexural strength of the mortars until a CMOD 800 μm115;

iii.

exposing the cracked samples at room conditions for a healing time of 1 month;

iv.

evaluating the flexural strength in terms of load recovery index and the stiffness recovery index.

Three series of test were prepared: -

the MS series contain 1 CHT with the sodium silicate inside;

-

the MI series was filled with sand;

-

the TQ series are respectively filled with sand and empty tube for a comparison with reference samples.

The LRI% and SRI%116 of mortar sample with 1 CHT containing sodium silicate reveal that an healing effect occurs (Figure 8.11). Whilst, no recovery has been observed in the reference samples (MI and TQ series).

Figure 8.11 Load recovery index and the stiffness recovery index of mortar sample containing a single CHT.

The CMOD value of 400 μm can be regarded as a threshold discriminating between acceptable crack width in the serviceability limit state (cracks smaller than this do not need a specific control in a variety of structure and exposed classes) as reported in the European Standard EN 1992-1-1: Eurocode 2: Design of concrete structure. 116 Load recovery index and stiffness recovery index % are the parameters used for evaluate the mechanical strength recovery. 115

114


A digital image analysis (Figure 8.12) of the samples containing the hollow tubes with sodium silicate after complete failure was also conducted117 to correlate the bending strength and stiffness recovery to the area covered by the diffusion of the healing agent upon cracking. The area covered by the healing agent, the position of the tube with respect to the notch, the average diameter and the aspect of the tube were considered. A direct correlation between these parameters and the mechanical properties recovery was not found. However, it could be observed that if the healing agent is able to move backward with respect to the direction of propagation of the crack, the mechanical recovery indices are higher.

Figure 8.12 Image analysis on two examples of the self-healing mortar prototypes containing one CHT embedded in the samples submitted to complete failure by three point bending test. The sodium silicate delivered is evidenced in red by using a software image elaboration. By courtesy of Politecnico di Torino.

Compressive strength influence on concrete. A preliminary verification of the suitability of the CHT container for concrete applications was conducted considering their ability of self-healing concrete to preserve its mechanical properties, with no significant alteration due to the presence of the hollow tubes. For this purpose, compressive tests were carried out on cubic concrete specimens with a side length of 10 cm. They were made of self-compacting concrete (SCC) using CEM type I 52.5R supplied by Buzzi Unicem S.p.A. One concrete sample with no CHT added, was prepared and used as a reference, while three concrete samples incorporating four empty CHT per liter in a self-compacting concrete (4CHT/L), were

117

The images were acquired with a digital camera and the images were processed with DxO Optics Pro 6.5.5 and elaborated using Photoshop CS. The area covered by the healing agent was determined by Image-Pro Plus 4.0 by MediaCybernetics Inc., USA.

115


used to evaluate the influence of the presence of the cementitious hollow tubes on the concrete compressive strength118. In a precautionary perspective, they were not filled with any substance to avoid any possible positive contribution to the compressive strength and also to simulate the condition of empty capsules, consequent to the release of the healing agent after damage. The CHT were added in a proportion of four tubes per liter, corresponding to a volume ratio with respect to concrete of about 1.6%. The CHT were manually placed in specimens, in such a way to distribute them horizontally, vertically and randomly distributed with respect to the load direction; a self-compacting concrete mix design was preferred for this test, in order to avoid the vibration process of the concrete, thus assuring the right CHT orientation. The samples were cured for 28 days at 90% RH until they were tested for compressive strength by means of a compressive press (MCC8 Controls). The results obtained are reported in Figure 8.12 and highlight that the addition of the CHT in a self-compacting concrete (4CHT/L) does not significantly influence the compressive strength.

Figure 8.13 Compressive strength resistances of a concrete sample containing the 4CHT/L. On the right, a picture of a broken concrete sample containing the tubes. The position of the CHT remains unaltered during aging and cracking.

A minor effect related to the orientation of the CHT with respect to the load direction was observed, as the samples with randomly dispersed CHT showed a small strength decrease on average. Such a reduction can be ascribed to the fact the tubular capsules, in this case were incorporated by pressing them into the fresh concrete mix during the casting in order to guarantee a random orientation in space and this could have contributed to worsen the interfacial properties between concrete matrix and CHT. However, this variation is negligible, as confirmed by the Standard UNI EN 196-1

118

Tubes are characterized by a diameter of 1 cm and length of about 5 cm so each element has approximately a volume of 15.7cm3.

116


Methods of testing cement - Part 1: Determination of strength. Moreover, no slipping or de-bonding of the tubes was observed at the end of the compressive tests, each tube being perfectly adherent to the concrete matrix even after final failure (see Figure 8.13). Ability to deliver the healing agent under crack. The ability to deliver the healing agent under cracking is usually evaluated by three point bending test with a controlled cracks opening, in this case, a load pressure is continuously applied on the sample containing the tubes. In this research, the ability of the CHT to be broken upon crack formation, without any additional loads was also tested on self-compacting (SCC) concretes containing the hollow tubes exposed to restrained shrinkage cracking. On this purpose, rings tests119 were performed on self-healing concrete prototypes. The ring test simulates a damage situation frequently occurring in full-scale structures (i.e. restrained shrinkage cracking), in view of a potential scale-up process of the proposed self-healing system. The instrumental setup consisted of a steel ring that partly restrained the shrinkage of an annular concrete specimen, which was cast around the ring (Figure 8.14). Due to the drying shrinkage, this type of restrain gives rise to a radial pressure on the steel ring and induces tensile stresses in the concrete. In the present setup, this stress was generated at 20°C and 60% RH and was measured by a strain gauge placed on the inner side of the steel ring. The concrete composition was the same as for the compressive tests previously described. A SCC was chosen in order to enhance the drying shrinkage and hence the tensile stresses in the sample. The mix was poured into a mold that contained a Vshape indentation, which was introduced in such a way to create a notch in the annular concrete specimen (Figure 8.14). In this way it was possible to induce macro-crack formation in a known region of the specimen.

119

The ring tests were performed in accordance with the ASTM Standard C1581-04: Standard Test Method for Determining Age at Cracking and Induced Tensile Stress Characteristics of Mortar and Concrete under Restrained Shrinkage.

117


Figure 8.14 Ring test setup. Formation of a microcrack (left) and complete failure macrocrack (right). By courtesy of Buzzi Unicem.

Four equally spaced CHT were manually positioned in this region, perpendicularly to the notch (Figure 8.15). The following samples containing CHT were considered: i.

SCC sample with CHT containing a sodium silicate solution (SCC_CHT_S);

ii.

SCC sample with CHT containing an aqueous solution of a colorant, Rhodamine B (SCC_CHT_W), to highlight the diffusion of the liquid into the crack120.

iii.

Finally, a SCC sample without CHT was used as a reference (SCC_ref).

Figure 8.15 shows the results of the average strain measured at the inner surface of the restrained steel ring at early age.

120

A limit of this test could be the different viscosity of the aqueous solution with the dye in comparison to sodium silicate. I was preferred to use the aqueous solution with the dye, instead of use sodium silicate with dye to avoid an additional complexity into the reaction with the cement, and to ensure the healing agent stability the inside the tubes.

118


Figure 8.15 Ring tests results: plain concrete (SCC_ref); concrete with hollow tubes filled with sodium silicate solution (SCC_CHT_S ); concrete with hollow tubes filled with Rhodamine-colored water (SCC_CHT_W).

The strain values are shown after the beginning of the exposure to drying (24 hours after casting). The time when complete cracking occurs is captured by a sudden strain release. Comparing the reference sample (SCC_ref) and the sample containing CHT filled with sodium silicate (SCC_CHT_S) it is possible to affirm that both samples manifested macro-cracks substantially at the same time. The sample SCC_CHT_S displayed also an additional sudden strain release 4 days after the appearance of the first macro-crack, which was due to the formation of another macro-crack at the diametrically opposite to the first one in the ring. This result suggests that, as expected, the tubes themselves were not able to counteract the shrinkage macro-cracking phenomena; on the other hand, they broke after the damage occurrence, successfully releasing the sodium silicate inside the crack path, as revealed by visual inspection (Figure 8.15).

119


Figure 8.16 Ring tests images: macro-cracking leading to the diffusion of Rhodamine B water solution up to the external surface; and crack surfaces with Rhodamine-colored water diffusion around the hollow tubes.

The Rhodamine B highlighted the diffusion of the fluid into the macro-crack by means of a visual inspection. The rupture of two hollow tubes as a consequence of cracking is clearly visible from the colored spots at two different points along the crack (Figure 8.16). This demonstrates that the liquid inside the tubes was spread out along the fracture surface with a radius of about 5 cm. The sample with the CHT containing aqueous solution of Rhodamine B (SCC_CHT_W) displayed the occurrence of a macro-crack at a later age. Indeed, the time of complete cracking detected by the strain release was delayed, probably be due to a higher degree of micro-cracking and damage accumulation in this sample. This innovative use of the ring test is a powerful method for detecting cracking in restrained ring concrete specimens containing tubes/capsules for self-healing. Indeed, it is easy to perform and does not require any other mechanical load to evaluate the capability of the capsules to be broken upon crack occurrence in the concrete matrix.

However, the repairing effect cannot be measured in this experimental setup due to the different time scales of the processes involved, the self-healing process being governed by the time for the chemical reactions to be substantially complete, while the strain release process upon restrained shrinkage macro-crack formation is instantaneous.

120


8.3

Spherical capsules by mechanical dripping method

As previously mentioned, the parameters influencing the efficiency of the healing are many such as the capsules morphology, the materials of the shell, the survival to the mixing process, the interface between the capsules shell and the cement matrix, the stability of the healing agent until crack heal, the chemical interaction of the healing agent with the matrix, the experimental set-up for evaluating the efficiency, to name a few. The cylindrical capsules present offer better chances to break upon crack appearance and release the agent more efficiently. On the other hand, spherical shape capsules can be more easily mixed, better dispersed and the effects due to the orientation of the capsules inside a concrete are avoided. In some studies, computational simulations of the behavior of the capsules inside cementitious matrix were done in order to simulate the mechanical phenomena involved. By the computational science, it is also possible to evaluate the bond strength interface between the capsules and the matrix, the behavior under different mechanical loads and the ability to release the healing agent. The modelling can be a useful tool to design suitable self-healing systems optimizing the experimental activities. Nevertheless, the parameters involved in such a complex systems can be various and difficult to manage simultaneously. Furthermore, the complexity of the system increase in the real scale when parameters such as storage, mixing, curing of the concrete are not controlled as in laboratory scale.

Inside the DualCEM project, a simulation of the flexural behavior of a cementitious prototype containing spherical capsules with wax shell was done using the Finite Element Methodology121 (FEM) method by the Department of Structural, Geotechnical and Building Engineering, Laboratory of Bio-Inspired Nano-Mechanics “Giuseppe M. Pugno” of the Politecnico di Torino and described in the Master Thesis

121

Finite element methods are numerical methods for approximating the solutions of mathematical problems that are usually formulated to precisely state an idea of some aspect of physical reality. A finite element method is characterized by a formulation with different variables, a discretization strategy, one or more solution algorithms and post-processing procedures. The method has since been generalized for the numerical modeling of physical systems in a wide variety of engineering disciplines such as electromagnetism, heat transfer, and fluid dynamics.

121


of M. Benci122. Some works in literature, present wax capsules shell. Nevertheless, the scarce interaction with the cement matrix and difficulties in the synthesis processes are still a concern. The study revealed that a capsule 5 mm sized and wax shell (0,1-0,4 mm in width), is able to break under an elastic deformation of the matrix, releasing the healing agent inside a crack. The interaction with the cement matrix and the shell of the capsule is one of the main problematic aspect to consider in the development of the capsules. By tailoring the roughness of the shell or the geometry of the capsule, the interface bond between the cement matrix could be improved.

Experimental activities has been hence performed, in this PhD research, in order to confirm the modeling observations. Sodium silicate was encapsulated in an impermeable wax shell. Spherical capsules were synthetized by a manually dripping process and immersed in cold water to obtain the hardened shell. The structural and morphologic features of the spherical capsules were evaluated through stereomicroscopy (Discovery V20 Zeiss)123. In order to characterize the durability of the capsules, they were added into a standard mortar mix design124 (2% wt with respect to the amount of cement), the mortars with capsules inside were cured at 95% RH at 20°C and broken until failure after 28 days from hydration. The survival of the capsules to the mixing process was evaluated by stereomicroscopy. In the mixing process, the capsules were subjected to strong destructive stress forces induced mainly by the compressive strength of the mortar mixer and friction with the aggregates. A visual examination of the self-healing mortar prism after the compressive test confirmed that, even in a mortar mixture, the capsules survived until breaking and their content was still liquid after 28 days. Moreover, fragments of broken capsules were observed after the complete rupture of the mortar, releasing the healing agent (Figure 8.17). This indicates that the mechanical resistance of the wax

122

M. Benci, Studio e simulazione FEM del comportamento in flessione di un prototipo di un cemento self-healing micro-incapsulato, Master Thesis, Politecnico di Torino (2013). 123 A. Formia, S. Irico, The DualCEM project evaluation of healing agents and development of encapsulation technologies for self-healing concrete, Cement International 13 (2015). 124 produced following UNI/EN196.

122


shell is sufficient for the capsules to survive the mixing, but is also suitable for the healing agent to be released, as a consequence of a structural stress.

Figure 8.17 A capsule of wax broken inside a self-healing mortar prototype after a compressive load damage. In this picture, it is possible to observe to release of sodium silicate healing agent after the breakage of the wax shell.

The main problem in this system consists in the methodology of synthesis. The manual dripping process produce a small amount of capsules for some laboratory scale test and the reproducibility of the synthesis is scarce. Parameters as the thickness of the wax shell and the volume of sodium silicate encapsulated are not controlled. For this reason, looking at methods able to produce capsules in large scale and inspired from the pharmaceutical encapsulation approach, the potentiality in using sodium alginates in the development of healing capsules were investigated.

8.3.1

Sodium/Calcium Alginate capsules

Sodium alginate applications. Alginate compounds are products derived from seaweeds and widely used in several biomedical applications, including the encapsulation of drugs or for cell delivery systems125-126, wound dressing, plant tissue

125

T. D. Dang, S. W. Joo, Preparation of tapped shaped calcium alginate microparticles with sphericity control, Colloids and surfaces B: Biointerfaces 102, 2013, 766-771. 126 L. Liu, F. Wua, X. Ju , R. Xie, W. Wang, C. Hui Niu, L. Chu, Preparation of monodisperse calcium alginate microcapsules via internal gelation in microfluidic-generated double emulsions, J Colloid Interf Sci 404 (2013) 85–90.

123


cultures, for immobilizing enzymes, in the release of proteins for bone repair application127. Alginates are anionic polysaccharides able to form hydrogels in a solution with multivalent cations such as Ca2+, Ba2+, Mg2+ and Na2+, widely used in biomaterials science to obtain gel form capsules to encapsulate functional molecules presented in an aqueous solution with tailored dimensions and morphologies128-129. The preparation of different material forms of alginates, especially calcium alginate microparticles has been intensively investigated in recent years.

Figure 8.18 a) Mannuronic Acid 1,4 ß-D- b) α-L-gluronic acid, c) Sodium Alginate.

Sodium alginates found recent applications as delivery systems in different selfhealing materials. In the work of S. van der Zwaag et al130alginate fibres were synthetized and applied for self-healing reinforced polymer composites. The selfhealing composite system consisted of alginate fibres containing unconnected compartments embedded in a polymethylmethacrylate (PMMA) matrix. The healing agent

in

the

compartments

polymethylmethacrylate

mixture

was

a

dichlorobenzene,

(DCB/DBB/PMMA).

In

dibromobenzene, this

case,

the

compartmented fibres were spun from an emulsion of ortho-dichclorobenzene in a

127

C. Wu, W. Fanm M. Gelinsky, Y. Xiao, J. Chang, T. Friis, G. Cuniberti, In situ preparation and protein delivery of silicate-alginate composite microsphere with core-shell structure, J. R. Soc. Interface (2011) 8, 1804-1814. 128 T.D. Danga,b, S.W. Jooa, Preparation of tadpole-shaped calcium alginate microparticles with sphericity control, Colloids and Surfaces B: Biointerfaces 102 (2013) 766– 771. 129 E. Mele, D. Fragouli, R. Ruffilli, G. L. De Gregorio, R. Cingolani, A. Anthanassiou, Complex architectures formed by alginate drops floating on liquid surfaces, Soft Matter 9 (2013) 6338. 130 S. van der Zwaag, A. M. Grande, W. Post, S. J. Garcia, T. C. Bor, Review of current strategies to induce self-healing behaviour in fibre reinforced polymer based composites, Mat Sci Tech-Lond 30 (2014) 1633.

124


water solution of sodium alginate into a CaCl2 solution gelation bath to form the healing agent filled compartment fibres. The diameter of the fibres and the volume of healing agent are controlled in the extrusion process and the fiber were finally wound onto a pre-heated (40°C) plastic bobbin. Micro-computed tomography combined with flexural bending test on poly(methylmethacrylate) model composites containing compartmented fibres proved that compartmented fibres offer a high potential for multiple locally distributed release of healing agent into the damaged polymer matrix in fibre reinforced polymer composites. A. Tabakovic et al (2015)131, produced alginate fibres as a new method for incorporating rejuvenators into asphalt pavement mixtures. They state that this approach offer advantages over spherical microcapsules, such as: 

alginate is an organic material and include poses no environmental/leaching risks;



due to the high aspect ratio of the compartments in the alginate compartment fibres there is a higher potential that the containers will encourage a fracture while at the same time locally releasing higher amount of healing agent (i.e. rejuvenator);



if not opened via fracture contact, alginate fibres will degrade over time, releasing the healing agent rejuvenator and presenting secondary self-healing trigger mechanisms.

The research findings support the potential for compartmented sodium alginate fibres to contribute to the further development of self-healing asphalt pavement systems. In cement materials science, sodium alginates were recently applied as a novel biobased curing compound for preventing the evaporation of water during the hydration process and the drying shrinkage132. The most common method to prevent the evaporation of the water is accomplished by covering the surface with a plastic sheets or by spraying it with a curing compound (polymers solutions/emulsions) creating a film that hinders water evaporation. The large amount of calcium produced by cement

131

A. Tabaković, W. Post, S. Garcia, E. Schlangen, Use of compartmented Sodium Alginate fibres as a healing agent delivery system for asphalt pavements, in proceedings of EMRS 2015, Warsaw. 132 J. Slopasa, E.A.B. Koenders, S.J.Picken, A novel bio-based curing compound for cement-based materials Application of superadsorbent Polymers and Other New admistrures in Concrete Construction, RILEM proceedings PRO 95 (2014).

125


hydration, and freely available on the surface of a cementitious compound, can react by a cross-link reaction with sodium alginate forming a non-water soluble calcium alginate gel and later a film. In this article, the possibility to use sodium alginate as an external curing compound was investigated by SEM observation of the microstructure and chloride permeability measurement.

Sodium/Calcium alginate capsules synthesis. The calcium alginate capsules were formed through the gelation of alginate droplets according to the following ion exchange reaction: 2Na − alginate + CaCl → Ca − alginate + 2NaCl Since the calcium cation has a much smaller size than the large polymer molecule, it can diffuse into the alginate solution, and according to the egg-box model (Figure 8.18), this leads to the cross linking process. After the gelation water insoluble gel is formed. The capsules of calcium alginate are prepared by dripping processes of a solution of sodium alginate and healing agent in a calcium ion batch. An illustration of the dripping process in reported in Figure 8.20.

Figure 8.19 Egg-box structure of a calcium alginate gel.

126


Figure 8.20 Schematic diagram and image of tadpole-shaped calcium alginate particles produced by a microfluidic flow-focusing device [T.D. Dang et al (2013)] 133.

The experimental phases of the synthesis process can be summarized as follows: Step 1. The first step consists in the preparation of the solution containing the sodium silicate and sodium alginate134 compounds. Sodium silicate was diluted in a solution of water and ethylene glycol (this last was added in order to avoid water evaporation during the process and facilitate the dissolution of alginates in water solution). The percentage of mixing components was optimized after several experimental attempts (as fully described and presented in the bachelor thesis of M. Lopresti) 135 and is expressed in Figure 20. The solution was mixed with a magnetic stirred for some hours to obtain a homogeneous solution. This step is very important for a good reproducibility during the dripping step. It was observed that increasing the temperature of the solution helps in the dissolution of alginates. Afterwards, the gelation bath is, a water solution of CaCl2 10% wt is prepared.

133

T.D. Danga, S.W. Jooa, Preparation of tadpole-shaped calcium alginate microparticles with sphericity control, Colloids and surfaces B: Biointerphaces 102 (2013) 766-771. 134 Alginic acid solution salt from Sigma-Aldrich. 135 M. Lopresti, Messa a punto di sistemi per l’autoriparazione di materiali cementizi, Bachelor Thesis , Università del Piemonte Orientale (2013).

127


Component in 100 ml Sodium silicate

40 ml

Water

54 ml

Ethylene Glycol

3 ml

Sodium alginate

1g

Figure 8.21 Optimized standard composition of the capsules in percentage and on the left different capsules morphologies obtained with a different amount of glycol.

Step 2. A technical equipment for an automated and repeatable synthesis of the capsules was specifically designed by Varnish s.r.l. (Figure 8.22), within the DualCEM project. The instrument is essentially based in on electronic dripper device, able to produce capsules with tailored features on large scale. When the sodium silicate/alginate solution is dripped in the gelation bath, a calcium alginate shell constituted and of a sodium silicate/sodium alginate water solution core, is formed. By changing the composition of the solution and the instrumental set-up (volume of the drops, dripping velocity, the distance between the needle and the surface of the gelation bath, CaCl2 concentration etc.) it is possible to obtain capsules with different morphologies and characteristics. Step 3. In the last step, the capsules are manually extracted from the gelation bath after 60 sec. This reaction time was optimized in order to obtain capsules with a good mechanical resistance136. The capsules are then stored in a closed pan.

136

The thickness of the shell is proportional to the time reaction into the gelation bath. The more the capsules react with CaCl2 the more the sodium alginate react and the shell became thick.

128


Figure 8.22 Schematic illustration of the dripping process. The dripper device produced by Varnish. A section of a capsules obtained by optical microscopy.

The synthesis with the automatized dripper allows to produce capsules in large scale amount: the device produces 89 capsules/min with a very high reproducibility. By means of a life cycle assessment analysis (LCA) performed by NovaRes s.r.l137. Some relevant environmental impact parameters were obtained, such as the energetic consumption equal to 0,0215 Wh/capsule and the Global Warming Potential (GWP) is equal to 13,5 mg CO2 eq/capsule138.

Unfortunately, these capsules constituted by a shell of calcium alginate and sodium silicate solution as encapsulated healing agent139 are not able to survive in a cement matrix, due to the very high pH and the high amount of calcium ions available in the matrix that contribute to the reaction with calcium alginate. The healing agent must be stabilized for the use in a cement matrix. On this purpose, several experimental activities were conducted with the aim to create an additional impermeable cover to the capsules. 137

Nova Res was born in 2009 as a spin-off company aiming at becoming a partner for the start-up, setup and optimization of new applicative ideas involving innovative materials. 138 Factor of conversion used is 0,627 Kg CO2/kWh (low voltage – Italy). Data obtained from Ecoinvent (www.ecoinvent.ch). GWP calculated on the materials used is 69,56 g CO2 eq/capsule. 139 In the core, the solution is constituted by a solution of glycol, water, sodium silicate, sodium alginate.

129


Healing agent stabilization. In order to create an additional impermeable cover protection to the capsules four techniques were used and here briefly presented: i.

Polymerization on the surface. In this case, the capsules are functionalized on the surface with a reagent (i.e. Fenton solution140) able to create radical group on the surface and subsequently promote a radical polymerization of some organic monomers such as styrene and methacrylate. The optimization of the Fenton reaction is not easy to control. The difficulty, in this approach is the stabilization of the OH radical grous at the surface of the capsules, in order to promote di polymerization at the interface;

ii.

Covering with epoxy resins. The capsules shell with epoxy resins are quite often described in literature: in this case commercial bi-component resins was tested and the capsules were manually covered;

iii.

Polymer precipitation. As a first step, the capsules were immersed in concentrated organic solution (chloroform) of several polymers such as poly methyl methacrylate (PMMA), polystyrene (PS), polyvinyl alcohol (PVA), polyvinyl butyral (PVB) and polyvinyl difluoride (PVDF). As second step, methanol is used for enhancing the precipitation of the polymer141. After this step: a polymeric film precipitates on the surface of the capsules. In this research, in order to use this method in large scale, the chloroform solution of polymer and methanol (as a non-solvent) were putted together in the same backer (they are immiscible) and the capsule were dropped inside. It happens that in the first step, the capsules are covered by the monomers and subsequently they fall down into the methanol, when the precipitation of the polymer occurs. This method was called “one-step covering” and is illustrated in Figure 8.23).

iv.

Tumbling – powder covering. This approach is used in food industries (i.e. candies production) to cover food with powders; in our case, different types of powder are used such as Arabic gum and powder of wax142. The idea

The Fenton reaction Fe2+ + H2O2 → Fe3+ + OH* + OH− or Fe3+ + H2O2 → Fe2+ + OOH* + H+ The chloroform is a chlorinated aliphatic hydrocarbon solvent, while the methanol is the nonsolvent. 142 This method was also used in order to evaluate the use of the wax without thermal treatment. To heat up the wax and use it in liquid form leads to several difficulties in the control of the wax hardening around the capsules. From technological point of view is not suitable. 140 141

130


consists to create the capsules shell with a mechanical covering of powders. In this case, the capsules must survive the mechanical compressive and frictional resistances during the tumbling. In Table 8-1, the critical aspects of the methods and the main results obstained during the experimental activity are summarized.

Figure 8.23 The precipitation of the polymer on the capsules surface in the “one-step covering” method.

Figure 8.24 A sodium/calcium alginate capsule with a PS shell by using the polymer precipitation technique.

131


Strategy

Methodology

Critical aspects - Difficulties in the control of the Fenton reaction

Polymerization on the surface

OH radical polymerization promoted by Fenton reaction

- Stabilization of OH groups on the capsules surface - Not suitable for large scale - low adhesion - inhibition by oxygen

- Difficulties in handling the capsules (very fragile)

Covering with epoxy resins

Polymer precipitation

Manual cover of the capsule with a bicomponent resin

Covering of the capsules with a polymer dispersed in chloroform and subsequently precipitation of it on capsules surface by a non-solvent

- The viscosity of the resins was too high - The thickness of the shell is not easily controllable

Covering of the capsules with powders by tumbling methods

- Comparison of a polymeric film on the capsules but not uniform (microscopy observation) - Agglomeration of capsules and polymeric film - The capsules were not stable - Capsules with a strong shell of epoxy resin but a not controllable dimension - The healing agent inside the capsules reacts with the epoxy

- Not suitable for large scale

- The capsules were not stable

- Compact film around the capsules

- Capsules stability only 24h

- Necessary to create more layers on the capsules

- Difficult to handle the capsules after the onestep method

- Creation of a layer at the interphase of the polymer solution and non-solvent

- Capsules need to have compressive resistances to survive the tumbling Tumbling – powder covering

Results

- Need to have impermeable film from powder tumbling - Not easy to control from the technological set-up available in laboratory

- The polymer layer is not always homogenous

- No capsules with a homogenous layer are produced - Agglomeration of the capsules during the tumbling process.

Table 8-1 Capsules stabilization strategy and analysis of results and critical aspects.

132


On the purpose to create an additional shell on the capsules, Varnish Technology S.p.A.143 designed and produced a second module of the electronic device. The device has two slides working like two waterfalls (Figure 8.25). In the first, CaCl2 solution is flowing and sodium/calcium alginate capsules are formed by dripping the alginate solution at the head of the slide: in this step, the dimension of the capsules is controlled. After that, capsules are stored in a first tank and from which they fall down into the second slide were they are covered with a polymeric material. The system, although having high potential and automated, has some barrier to the implementation. First, the methods discussed previously are not successfully suitable for alginate capsules. The main critical aspects are (i) the scarce bond interface with polymers and calcium alginate shell, (ii) the difficulties in handle the capsules during the processes, (iii) the polymerization condition must be carefully controlled, (iv) the need for adequate instruments and technologies for the covering realization and (vi) the presence of volatile solvents, which can be noxious and flammable.

Figure 8.25 Second module of the automated dripping system developed and produced for the additional covering of the capsules.

143

Varnish is an Italian leading company in painting technology for general industrial application and automotive sector.

133


8.3.2

Assessment of encapsulation synthesis by a preliminary chemometric approach

The experimental design approach is an innovative tool in experimental materials science and is one of the most applied field of chemometrics technique, also used in the optimization of industrial experimental activities. The experimental designs are useful to solve multivariate problems, this happens when several experimental factors or variables may influence the results of a research. In this case, screening experiments are usually performed in order to determine the experimental variables and interactions that have significant relevance on the results. When the experimental variables and the responses144 have been defined, the experiment can be planned and performed in such a way that a maximum of information is gained from a minimum number of experiments. In this research, a factorial design method is used to investigate the influence of, (i) the content of sodium silicate, (ii) the amount of water, (iii) the addition of ethylene glycol and (iv) the sodium alginate content, on some features of a sodium/calcium alginate capsules in order to optimize the synthesis process. The responses selected, as the output of the experimental design, were chosen looking at possible applications of the capsules as healing agent containers and are: (a) capsules morphology, (b) compressive mechanical resistance, (d) weight of the capsules, and (d) their stability in time (measured by the residual weight). In a factorial design, the influences of all experimental variable, factors, and interaction effects on the response are investigated simultaneously. If the combinations of k factors are evaluated at two levels, a factorial design will consists of 2k experiments. In Figure 8.27, the factorial designs for 2, 3 and 4 experimental variables are shown. A zero-level is also included, namely a central experiment, in which all variables are set at their mid value. Three or four center experiments should always be included in factorial designs, for the following reasons: -

the risk of missing non-linear relationships in the middle of the intervals is minimized, and

-

repetitions allows for determination of confidence intervals.

“Responses” are defined as the measured value of the result(s) from experiments. They are the output of the experimental design. 144

134


The levels of the factors are given by – (minus) for low level and + (plus) for high level. What – and + should correspond for each variable is defined from what is assumed to be a reasonable variation to investigate. In this way the size of the experimental domain has been settled. The factorial design used in the assessment of the capsules features is shown in Figure 8.26.

Figure 8.26 The minimum, maximum and central parameters used in the factorial experimental design.

Figure 8.27 Factorial design. Note that all variables are changed simultaneously in a controlled way, to ensure that every experiment in each design is a unique combination of variable levels145 [Lundstedt et al (1998)].

145

T. Lundstedt, E. Seifert, L. Abramo, B. Thelin, A. Nystrom, J. Pettersen, R. Bergman, Experimental design and optimization, Chemometrics and Intelligent Laboratory Systems 42 (1998) 3–40.

135


Figure 8.28 Experimental design used for the optimization and control of capsules production.

The responses measured are presented and described as follows: Capsules morphology. The change in the parameters affects the morphology of the capsules (see Figure 8.29). A scale of optical comparison on the capsules morphology was specifically created. This method is based on the observation of drops by means of a mask in which are illustrated different capsules morphologies (Figure 8.30). Each capsules, is associated to a specific morphology index; from level one, which represents the spherical approximation, to level nine, in which the morphology is associated to “flat capsules”.

136


Figure 8.29 Example of different capsules morphology obtained during the experimental design.

Figure 8.30 Capsules morphology scale.

Compressive mechanical resistance. The compressive resistance gives information on the ability of the capsules to survive the compressive pressure. A dedicated setup was designed on this purpose. It consists in a vertical column in which at the bottom there is a slide glass, the pressure increases when fine sand is gradually introduced in the column. The total pressure on a single capsule consists in the sum of the weight of the column and sand added inside. The total pressure on a single capsule is measured. The capsules resulting from each experiment showed different mechanical behavior, a sequence a progressive index level as described in Table 8-2.

137


Level 1

Description The capsules cannot be handled; their compressive resistance is so low that it is even impossible to extract them from the gelation bath. The capsules are very fragile and it is very difficult to extract them from

2

the gelation bath. The range in compressive resistances is between 2 and 5 g each capsule (g/cap). The capsules are fragile but they have a sufficient compressive resistance

3

to extract them from the gelation bath. The range in compressive resistances is between 5 and 15 g/cap.

4

5

The capsules are less fragile and have more elastic behavior; under the compressive load, the capsules seep the liquid in the core until the rupture. The capsules are able to reach compressive resistances of 40 g/cap but they lose all the liquid core from the capsules shell membrane. The capsules are fragile, the range in compressive resistances is between

6

15 and 40 g/cap. In this case the capsule membrane doesn’t lose the liquid core. The capsules have the highest compressive resistances (> 40g/cap) and

7

also in this case the capsules shell breaks with fragile behavior, without losing the liquid core.

Table 8-2 Progressive index level of the compressive resistance of the capsules.

Loss of weight. The weight of the capsules is measured as supplementary characterization immediately after the removal from gelation bath. The stability of the capsules at 20°C and 75% RH is evaluated by the residual weight in percentage with respect to the initial weight. 𝑟𝑒𝑠𝑖𝑑𝑢𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡 = (𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑤𝑒𝑖𝑔ℎ𝑡 − 𝑤𝑒𝑖𝑔ℎ𝑡 𝑎𝑓𝑡𝑒𝑟 𝑥 ) × 100 Where x represents 30, 60, 90, 120 minutes and 24 hours of exposure. The results obtained from the experimental design, are all resumed in the Table 8-3.

138


Table 8-3 Responses of the experimental design: Morphology index, Mechanical index, Initial weight and Residual weight.

The analysis of the results of a factorial experimental design is often complex and related to mathematical and statistical assumptions. The correlation between the variables is one of the most common and useful statistics. A correlation is a single number that describes the degree of relationship between the variables, it refers to a broad class of statistical relationship involving dependence. The software “Statistica” (StatSoft) was used to analyze the correlation between the analyzed variables. For the sake of brevity, several results are omitted in the text, but findings are explained following. The results of the correlation, revealed that the mechanical resistance and the morphology index have a higher degree of correlation. This means that a relationship between these two variables is present. The scatter plot of the two-variables in Figure 8.31 shows that they are linearly dependent, the higher is the morphology index, and the higher is the compressive resistance. In the scatter plot, two clusters of data are highlighted based on the amount of sodium alginate in the dripped solution. The higher is the sodium alginate content and the lower is the morphology index (far from spherical capsule) and vice versa. No

139


cluster division is observed for the other variables (sodium silicate, water and ethylene glycol). The presence of three overlapped points (S2H2G2A2), which represent the central points of the experimental design, indicates a good reproducibility of the measurements.

Figure 8.31 Scatterplot of Compressive strength resistances index against morphology index.

A further investigation on the relationship between the initial weight and the mechanical resistances relationships highlights that a correlation is present (probably not linear). Also in this case, is possible to divide the data in two clusters based on the sodium alginate content. The higher is the sodium alginates amount, the higher is the mechanical resistance.

Figure 8.32 Scatterplot of Compressive strength resistances index against initial weight.

140


From the correlation between the other variables it was difficult to extract some relevant information due to the high dispersion of data. Nevertheless, the initial weight of the capsules is related to the residual weight at long time of exposure at 20°C – 75%RH (at 24 hours). The residual weights at 30, 60, 90, 120 minutes and at 24 hours are correlated between each other, as expected. From the results of the experimental design analysis, is possible to state that: 

the sodium alginate content is the most crucial parameter for tailoring the capsules features during the synthesis process;



the morphology of the capsules and mechanical resistance are in linear correlation;



in order to obtain capsules with spherical morphology and higher mechanical resistance, high amount of sodium alginate is needed;



a correlation between the initial weight and the mechanical resistances is observed, revealing that also in this case a higher amount of sodium alginate allows higher compressive resistances;



less information were obtained about the capsules stability due to the low correlation between the other variables with the residual weight;



the experimental design allowed to individuate in the center experiment (S2H2G2A2) the optimal composition of the dripping solution to obtain capsules with high mechanical compressive resistance (> 40g/cap and index 7) and spherical morphology (morphology index 2).

The experimental design approach used in this work allowed to have a control on the synthesis method of capsules and on the characterization of their behaviors, evaluating the reproducibility of the proposed solution.

141


Conclusions

I

n this research, the self-healing in building science was explored as innovative functionality for concrete structures. The encapsulation of healing agents was used as technical approach. Inside the DualCEM project, two different strategies

of encapsulation of sodium silicate (as healing agent) were investigated: -

cementitious hollow tubes, produced by a extrusion process of a special developed cement paste;

-

sodium/capsule alginate spherical capsules, synthetized by dripping process through an automatized dripper, especially produced on this purpose inside the project.

The cementitious hollow tubes revealed to be a suitable container for sodium silicate, with a good stability in the cement matrix. Flexural strength recovery, measured by the three point bending test in CMOD control, demonstrated the ability of a concrete containing cementitious hollow tube, to re-gain mechanical performance. The main experimental considerations, on the encapsulation strategies explored, are summarized in the Table below.

Concepts from the academic disciplines were applied: -

the reactivity of sodium silicate with a Portland cement matrix was assessed by a combined approach based on SS-NMR and XRD powder diffraction. The reactivity of sodium silicate revealed to be more complex than expected. In addition, it was quantitative demonstrated the ability of sodium silicate to develop mechanical strengths in reaction with hydrated Portland cement;

-

an experimental design was used, an a innovative tool, for the assessment of the capsules synthesis. The experimental design allowed to understand the influence of the reagents involved in the synthesis reaction on the final capsules features 142


(morphology, stability, compressive strength resistances). Nevertheless, this capsules are not stable in cement matrix, and an impermeable covering is still needed with strength resistance to the mixing process. The synthesis of containers in large scale and stability are crucial aspects for applicative case studies. In addition, the survival to the mixing process is still a critical issue. Although several self-healing approaches are already presents in literature, it is possible to state that, every self-healing method must be finalized for a specific application. In the light of the complexity of such systems the final application and the specific research approach must be intertwined. In addition, a high knowledge about encapsulation technology for possible applications in building materials has been achieved.

The SWOT matrix (Strengths, Weaknesses, Opportunities, Treats) below resumes the potentialities and the limits to the implementation of self-healing in building science after this research findings.

143

General Conclusion and Impacts of the research

GENERAL CONCLUSIONS AND IMPACTS OF THE RESEARCH ON UNIVERSITY-INDUSTRY COLLABORATION

T

wo fields of research on innovative multifunctional materials have been successfully extended to the chemistry of cement: i.

ii.

photocatalytic materials able to reduce environmental pollution and characterized by self-cleaning ability; self-healing materials able to repair themselves after an externally induced damage.

In both cases, significant contributions to the basic knowledge of multifunctional systems within a cement matrix have been obtained. Moreover, and more importantly, the know-how transfer from laboratory scale to large scale in-field has been introduced, supplying new ideas and giving a new impulse for innovating the “old concrete world”. This PhD research is characterized by an interdisciplinary approach in which chemical sciences were combined with different academic disciplines (i.e. environmental sciences, materials science and engineering, structural engineering, polymeric science and technology and computational science) and industry research experience in cementitious materials. The collaboration and teaming between universities and industries was essential to integrate data, methodologies, perspectives, and concepts in order to advance fundamental understanding in the multifunctional cement based materials studied, including the evaluation of potentialities and barriers to the implementation in real applications. The concepts of basic chemistry research, on the explored topics, were transferred to the applied knowledge in advanced cement based materials as a transmission channel for industry research and development. Innovative tools, such as computer modeling and chemometric approaches were also used and evaluated as innovative supports for the experimental activities. A photocatalytic demonstrator was created to evaluate, in real scale, the potentiality of developed photoactive concrete tiles to reduce smog pollutants. Nevertheless, here

144


some efforts are needed to increase the understanding of the performances in real applications (as concluded in several European projects), this is a crucial need in order to put such materials into a potential market. Looking at market potentialities, an economic assessment of the photocatalytic products was performed. The commercialization of photocatalytic concrete for air purification is limited by an inadequate, and scarce uniformity on a legal level and by scientific doubts about the impacts on real scale application, in which the environmental conditions play a relevant role. A case study on self-cleaning ability of a white photocatalytic concrete, exposed to atmospheric conditions for two years, was performed, revealing a good conservation of aesthetic surface properties. This behavior could probably be the most appealing functionality for potential customers. Photocatalytic building materials can be a marketing tool for commercialize innovative and high value products, but would need a dedicated sales, marketing, and technical support approach. The encapsulation of a healing agent was the approach used in the investigation on self-healing concretes. The sodium silicate has proved to be an effective healing agent by the characterization of chemical reactivity with cement matrix and structural tests. Prototypes of self-healing mortars and concretes were produced, and their ability to work under a restrained shrinkage crack failure was evaluated in a real damage application. The expertise and know-how obtained in the encapsulation technology of engineering additions to building materials could leads to future research opportunities for other technological applications. Self-healing concrete is a very interesting material, but it will probably remain a niche application for the next years. Healing agent are available and it was clearly demonstrated that they can play a role in the achievement of a resilient strengths, however the industrial use of such additions has not yet been optimized for the concrete industry’s world. Most of the self-healing technologies are not able to survive the mixing process of concrete or they cannot guarantee the durability over the lifetime of a concrete structure (30-100 years). Further research is required before that self-healing materials become part of common use in concrete, and the participation of the building chemical industry is necessary for the development and commercialization of this new products.

145


For both of the explored issues, it is possible to state that improvements in sustainable and resilient project delivery are possible through a multidisciplinary integration of material and design selection, based upon life cycle analysis measurements; implementation of life cycle costs analysis versus lowest cost economics, use of innovative materials and technologies and collaborative platforms during project design and construction.

Beyond experimental results, the PhD financial support provided by Buzzi Unicem has led to several impacts on this industrial and academic research activity. The impact of a research work based on university-industry collaborations, can be evaluated in multiple aspects, such as productivity, direction of research, development of skills and experience for scientists and end-users, development of a social networking and technological transfer. There has been a rapid rise in commercial knowledge transfers from university to endusers in recent years, through licensing agreements, joint research joint ventures and start-ups. One of the most important university-industry technology transfer (UITT), is the informal transfer of know-how through more informal interaction between university scientists and managers/entrepreneurs in the private sector. This is not something that could immediately be put into a product, but it might be something that increases knowledge, and that will help the industry in the development of innovative research projects. Buzzi Unicem increased its know-how in innovative research fields, like multifunctionality applied to building materials, as well as the role and the potentialities of innovative characterization (i.e. SS-NMR, thermoporometry) and experimental (i.e. chemometrics) techniques. This includes the evaluation of the main barriers to implementation, and at the same time, the exploration of perspectives for future creative research on innovative technologies. It is important to consider UITT from a strategic perspective, which means increasing the ability to identify the commercial viability of new sciences and technologies. An effective UITT is possible when all the stakeholders (university scientists, university technological managers who discover new technologies, firms and managers who commercialize university-based technologies) have common points of view about research goals/culture/constraints. The mutual articulation of well-defined 146


goals and objectives is a crucial way to achieve an understanding of each group’s values and needs, and reach a common ground.

The value of personal relationships and social networks is an important aspect in UITT processes. In this PhD research, collaborations with industrial partners and academic researchers on the financed research project (DualCEM) and the co-working in experimental theses, resulted in fruitful cooperation and multidisciplinary teamwork. The participation in international conferences leads, in addition to the dissemination of scientific results, to development of international academic networking. In this research, the built social networks allowed the knowledge transfer in multifunctional materials technologies, and gave the chance to explore new research perspectives in other research issues. Through university-industry collaboration, besides the technology transfer, the construction of professional skill development is an important aspect for graduated students and for the stakeholders. This includes improvement in theoretical and technical knowledge of instrumental techniques, the evaluation of innovative scientific tools to support experimental research (i.e. computational science, chemometric approaches), the organization of a research project and the development of personal soft skills. Interacting with industry, scientists can conduct “better” basic research as they can give their experiments/results a wider role, refine and tune the experiments, and sometimes have different perspectives on a problem to solve146. This also means the evaluation of the existing technologies and studies in literature from a critical point of view, identifying the potentialities and the obstacles, as discussed in this research thesis. An interesting study by D.S. Siegel et al.42 analyzed the UITT processes and their outcomes. Based on 98 structured interviews of key UITT stakeholders (university administrators, academics, industry scientists, business managers and entrepreneurs) at five research universities in US, they concluded that these stakeholders have

146

D.S Siegel et al, Commercial knowledge transfers from universities to firm: improving the effectiveness of university-industry collaboration, Journal of High Technology Management Research 14 (2003) 111-133.

147


different perspectives on the desired outputs of UITT. Numerous barrier to an effective UITT are identified, including culture clashes, bureaucratic inflexibility, poorly designed reward systems, and ineffective management of university technology transfer. Based on this qualitative evidence, they provide several recommendations for improving UITT. Recommendations for university-based improvements include the need to improve their understanding of the requests of companies that can potentially commercialized their technologies, devote additional resources to technological transfer and patenting, and recognize the value of personal relationships and social networks, involving scientists, graduate students and alumni. On the other hand, industry-based improvements to the UITT process need proactive efforts to bridge the cultural gap with academia and technology managers with university experience, and to improve UITT social networks. They conclude that universities and industries often have different perspectives and goals with respect to intellectual property. However, the UITT may be one process that can improve the convergence between different organizations. In the US, extensive collaboration between universities and industry and the ensuing transfer of scientific knowledge, has been viewed as one of the main contributors to the successful technological innovation and economic growth of the past three decades147.

Hall, B., 2004, “University-Industry Research Partnerships in the United States”, in Contzen, Jean-Pierre, David Gibson, and Manuel V. Heitor (eds.), Rethinking Science Systems and Innovation Policies, Proceedings of the 6th International Conference on Technology Policy and Innovation, Purdue University Press. 147

148


An example of knowledge sharing between UPO University and Buzzi Unicem on the development of sustainable cement In the light of the knowledge developed during the scientific collaboration between the University of Piemonte Orientale (UPO) and Buzzi Unicem, and the know-how resulting from this PhD research, the possible recovery of hydrated cement extracted from concrete demolition waste (CDW) has been explored as an application of sustainable technology. In this research, a pure sample of hydrated cement waste (HCW) was prepared and used as a raw material in combination with limestone and schist for the production of new clinker. Hydrated cement consists of amorphous calcium silicate hydrates, calcium hydroxide and minor amount of calcium/magnesium carbonate; its chemical composition makes it appealing for the re-use as raw material in the cement production chain. To investigate the potentialities of this application, two powder formulations were produced by using 30 and 55% of HCW as a replacement of a reference meal for Portland clinker production. The production and characterization of the experimental clinker is widely described in the publication “An investigation on the recycling of hydrated cement from concrete” by D. Gastaldi et al (2015)148 reported in Appendix and the technology was patented in 2015149. The results revealed that the hydrated cement extracted from CDW demonstrated itself to be a resource instead of a waste material. Its composition (in particular CaO/SiO2 ratio) makes it appealing as a raw component, substituting natural quarried resources. When a higher amount of HCW is used, a non-Portland clinker is produced with valuable properties as supplementary cementitious material. The environmental advantages in using recycled cement include: -

reducing CO2 emissions associated to the clinker/cement production;

-

the use of recycled hydrated cement reduces the consumption of natural resources.

148

D. Gastaldi, F. Canonico, L. Capelli, L. Buzzi, E. Boccaleri, S. Irico, An investigation on the recycling of hydrated cement from concrete, Cem Con Compos 61 (2015) 29-35. 149 EP2878586 - Cementitious products obtainable from disposed concrete (2015), patent of UPO University and Buzzi Unicem.

149


Activities and congress participations The participation to congresses during the PhD activity, as well as the involvement in research development activities of the company, are included into the technological transfer between the academic and industrial research. In this paragraph, it will be reported the main participations and activities.

-

International Conference on Self-Healing Materials (ICSHM2013), Ghent, Belgium, 2013. This is one of the most relevant conference in the field of selfhealing materials, which allowed improving the knowledge in this field.

-

The International Conference on Photocatalysis – Standardization and Certification Assisting Commercialization, Prague, Czech Republic, 2014. The conference allowed having an overview on the main barriers to the implementation, commercialization and standardization of photocatalytic materials for air purification and self-cleaning purpose. It has been also a chance to improve the scientific networking in this topic.

-

Fall Meeting 2015 (EMRS2015), Self-Healing Materials: from concept to the market, Warsaw, Poland 2015. This conference has been a chance for the dissemination of the results of DualCem project research. The poster presentation “SS-NMR and XRD investigation on binding mechanism for self-healing cementitious materials design: the assessment of the reactivity of sodium silicate” is reported in Appendix.

-

International Workshop on Durability and Sustainability of Concrete Structures, Bologna, Italy, 2015. During this conference, it was presented an oral presentation on some results achieved in the DualCem project in collaboration with the Politecnico of Torino. It has been also a chance to improve the scientific networking. The abstract “Extruded cementitious hollow tubes for healing agent delivery” is reported in Appendix.

150


-

Two poster presentations was also presented, and reported in Appendix, at the XLI Congresso

Nazionale

di

Chimica

Fisica

2013,

Alessandria

(AL),

“Thermoporometry: an innovative method for the study of microstructural evolution of cement” and at the conference Giornate Italo Francesi della Chimica 2014, Torino (TO), “Innovative materials in the cement field: from structural to multifunctional applications”.

-

Research stay at the Wihelm Dyckerhoff Institute (WDI), Wiesbaden, Germany in 2014. I was involved in the research on photocatalytic materials and the optimization of self-cleaning tests for photocatalytic cementitious materials.

-

I was involved in the development of a quality control test in the production chain of the bentonite-cement by means of Near Infrared Spectroscopy (NIR), in the Buzzi Unicem plant of Settimello (FI).

151

Bibliography

Bibliography Multifunctionality in cement-based materials

K. Salonitis, J. Pandremos, J. Paralikas, G. Chryssolouris, Multifunctional materials: engineering applications and processing challenges, Int J Adv Manuf Technol 49 (2010) 803–826. C. Joseph, D. Garder, T. Jefferson, B. Isaacs, B. Lark, Self-healing in cementitious materials: a review of recent work, Construction Material 164 (2011). P. Ragesh,V. Anand Ganesh, S.V. Nair, A. Sreekumaran Nair, A review on ‘selfcleaning and multifunctional materials’, J Mater Chem A 2 (2014) 14775. J. Elkington, Cannibals with Forks: The triple bottom line of 21st century business, Capstone: Oxford, (1997). Symonds Group Ltd 469697, Construction and demolition waste management practice, and their economic impact. Final Report to DGXI, European Commission; February 1996. C. Fisher, M. Werge EU as a recycling society – present recycling levels of municipal waste and construction & demolition waste in the EU. ETC/SPC working paper 2. (2009) Copenhagen. V. Corinaldesi, G. Moriconi, Behaviour of cementitious mortars containing different kinds of recycled aggregate, Constr Build Mater 23 (2009) 289–94. J. K. Buffenbarger, Building toward a sustainable and resilient future, proceeding of DSCS2015, Bologna.

The chemistry and mineralogy of the cement F. Locher, (2000) Zement, Grundlagen der Herstellung und Verwendung. Düsseldorf: Verlag und Technik. H.F.W. Taylor (1997) Cement Chemistry, 2° ed. Thomas Telford, London. D. C. MacLaren, M. A. White, Cement: Its chemistry and properties, J. Chem. Educ. (2003) 80(6) 623. H. M. Jennings, Refinements to colloid model of CSH in cement: CM-II, Cem Concr Res, 38 (2008) 275-289.

152

Bibliography

D. Gastaldi, F. Canonico, S. Irico, D. Pellerej, M.C. Paganini, Near-infrared spectroscopy investigation on the hydration degree of a cement paste, J Mater Sci (2010) 45 3169-3174. F. Sanchez, K. Sobolev, Nanotechnology in concrete – A review, Construct Build Mater, 24 (2010) 2060-2071. R. Sugrañez, J.I. Álvarez, M. Cruz-Yusta, I. Mármolc, J. Morales, J. Vila, L. Sánchez, Enhanced photocatalytic degradation of NOx gases by regulating the microstructure of mortar cement modified with titanium dioxide, Build Env 69 (2013) 55-63. M. Brun, A. Lallemande, J. Quinson, C. Eyraud, A new method for the simultaneous determination of the size and the shape of pores: the thermoporometry, Thermochim. Acta, 21 (1977) 59-88. S. Irico, D. Gastaldi, F. Canonico, G. Magnacca, Investigation of the microstructural evolution of calcium sulfoaluminate cements by thermoporometry, Cem Concr Res 53 (2013) 239-247. M. Horgnies, I. Dubois-Brugger, E.M. Gartner, NOx de-pollution by hardened concrete and the influence of activated charcoal additions, Cem Concr Res 42 (2012) 1348-1355.

TiO2 assisted photocatalyis D. Pellerej, Studies on self-cleaning and air-cleaning properties of photocatalytic cement-based materials, PhD thesis 2010. K. Hashimoto, H. Irie, A. Fujishima, TiO2 Photocatalysis: An historical overview and future prospects, Jap Journal Appl Physic 44 no.12 (2005) 8269-8285. J. Zhang, Y. Nosaka, Mechanism of the OH radical generation in photocatalysis with TiO2 of different crystalline type, J Phys Chem C 118 (20) (2014) 10824-10832. K. Ya Kondratyev, A. P. Cracknell, Observing Global Climate Change (1998). N. J. Krou, I. Batonneau-Gener, T. Belin, S. Mignard, M. Horgnies, I Dubois-Brugger, Mechanisms of NOx entrapment into hydrated cement paste containing activated carbon - Influences of the temperature and carbonation, Cem Concr Res 53 (2013) 5158. L. Zhang, R. Dillert, D. Bahnemann, M. Vormoor, Photo-induced hydrophilicity and self-cleaning: models and reality, Energy Environ Sci 2 (2012)7491. S. Nishimoto, Bharat Bhushan, Bioinspired self-cleaning surfaces with superhydrophobicity, superoleophobicity, and superhydrophilicity, RSC Adv, 3 (2013) 671-690. 153

Bibliography

B. Tryba, M. Piszcz, A.W. Morawski, Photocatalytic and Self-Cleaning Properties of Ag-Doped TiO2, The Open Materials Science Journal 4 (2010) 5-8.

Photocatalysis in the construction industry M.V. Diamanti, M. Ormellese, M. Pedeferri, Characterization of photocatalytic and superhydrophylic properties of mortars containing titanium dioxide, 38 (2008) 13491353. G. L. Guerrini, T. De Marco, L. Bottalico, The situation of Photocatalysis in Europe and other regions, in proceeding of i.lab seminar Recent Issue and Trends in Photocatalysis (2015), Bergamo. C. Minero. Studio di materiali fotocatalitici per la pavimentazione stradale. Technical report, Università degli Studi di Torino – Regione Piemonte - Città di Torino, 2008. Bewertung der NO-Abbauleistung von photokatalytisch aktiven Produkten, Fachverband angewandte Photokatalyse, 2014. Experimental part I: NOx reduction E. Jimenez-Relinque, J.R. Rodriguez-Garcia, A. Castillo, M. Castellote, Influence of Chemical Composition of Cementitious Matrix, Surface Roughness and Microstructure on Photocatalytic activity of TiO2 Mortars, ICCC, Beijing 2015. S. Irico, D. Gastaldi, F. Canonico, L. Capelli, M. Bianchi, G. Ricchiardi, E. Boccaleri, Photocatalysis in building materials: the effect of MIX DESIGN in cement based products on the air pollutant reduction activity, SPEA7, Oporto 2012. J.V. Staub de Melo, G. Trichês, Evaluation of the influence of environmental conditions on the efficiency of photocatalytic coatings in the degradation of nitrogen oxides (NOx), Build Environ 49 (2012) 117-123. J.K. Sikkema, S.K. Ong, J.E. Alleman, Photocatalytic concrete pavements: Laboratory investigation of NO oxidation rate under varied environmental conditions, Constr Build Mater 100 (2015) 305-314. C. L. Bianchi, C. Pirrola, F. Galli, G. Cerrato, S. Morandi, V. Cappucci, Pigmentary TiO2: A challenge fot its use as photocatalyst in NOx air purification, Chem Eng J 261 (2015) 76-82. A. Folli, C. Pade, T. B. Hansen, T. De Marco, D. E. Macphee, TiO2 photocatalysis in cementitous systems: Insights into self-cleaning and depollution chemistry, Cem Concr Res 42 (2012) 539-548.

154

Bibliography

V. Etacheri, c. Di Valentin, J. Schneider, D. Bahneman, Visible-light activation of TiO2 photocatalysts: Advances in theory and experiments, J Photochem Photobiol C: Photochemistry Reviews 25 (2015) 1-29. A. Zaleska, Doped-TiO2: A review, Recent Patent on Engineering 2 (2008) 157-164. L. Gomathi, K. Eraiah Rajashekhar, Mechanochemical reaction of TiO2 with balanine or the preparation of visible light active nitrogen doped titania: adsorption and kinetic studies, J Sol-Gel Technol 60 (2011) 144-158. E. Borgarello, J. Kiwi, M. Gratzel, E. Pelizzetti, M. Visca, Visible light induced water cleaverage in colloidal solutions of chromium-doped titanium dioxide particle, J Am Chem Soc 104 (1982) 2996. D. Di Battista, Utilizzo di metodi meccanochimici per il drogaggio di semiconduttori, Bachelor Thesis, (2015) Università del Piemonte Orientale. A. Gaber, M. A. Abdel- Rahim, A. Y. Abdel-Latief, Mahmoud. N. Abdel-Sala, Influence of calcination temperature on the structure and porosity of nanocrystalline SnO2 synthesized by a conventional precipitation method, Int. J Electrochem Sci, 9 (2014) 81-95. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Visible-light photocatalysis in nitrogen-doped titanium oxides, Science 293 (2001) 269. Self-healing in construction materials M. de Rooij, K. Van Tittelboom, N. De Belie, N. Schlangen, Self-healing Phenomena in Cement-based materials, State-of-the-Art Report of RILEM Technical Committee 221-SHC. C. U. Grosse, K. van Tittelboom and N. De Belie, Non-destructive testing techniques for the observation of healing effects in cementitious materials – an introduction, Proceeding of fourth International Conference on Self-healing Materials (2013) 60. D. Homma, H. Mihashi, T. Nishiwaki, Self-healing capability of fiber reinforced cementitious composites, Journal of Advanced Concrete Technology 7, (2009) 2: 217228. M. de Rooij et al (Eds.): Self-healing phenomena in cement-based materials, RILEM 11, pp. 65-117. C. K. Edvarsen, Water permeability and self-healing of through-cracks in concrete, DAfStb Bull 455 (1996). S. Qian, J. Zhou, M.R. de Rooij, E. Schlangen, G. Yea, K. van Breugel, Self-healing behavior of strain hardening cementitious composites incorporating local waste materials, Cement & Concrete Composites 31 (2009) 613–621.

155

Bibliography

V. Wiktor, Recent advances on the development of bacteria based self-healing concrete for full-scale outdoor application, proceeding in “Self-healing materialsfrom concepts to the market” EMRS2015 Fall Meeting 2015. V. Wiktor, H. M. Jonkers, Field performences of bacteria-based repair system: pilot study in a parking garage, Case Studies in Construction Materials (2015). L. Ferrara, S. R. Ferreira, V. Krelani, M. della Torre, F. Silva, R. T. Filho, Natural fibre as promoters of autogenous healing in HPFRCCS: results from on on-going Brazil-Italy cooperation, in proceeding of DSCS2015, Bologna. M. M. Escobar, S. Vago, A. Vázquez, Self-healing mortars based on hollow glass tubes and epoxy–amine systems, Composites: Part B 55 (2013) 203–207. K. V. Tittelboom, N. De Belie, D. Van Loo, P. Jacobs, Self-healing efficiency of cementitious materials containing tubular capsules filled with healing agent, Cem Concr Compos 33 (2011) 497–505. M. Wua, B. Johannesson, M. Geiker, A review: Self-healing in cementitious materials and engineered cementitious composite as a self-healing material, Con. Build. Mat. 28 (2012) 571-583. K. V. Tittelboom, D. Snoeck, J. Wang, N. De Belie, Most recent advances in the field of self-healing cementitious materials, In: Proceeding of fourth International Conference on Self-healing Materials (2013) 406. B. Boh, B. Šumiga, Microencapsulation technology and its applications in building construction materials, RMZ – Materials and Geoenvironment 55-3 (2008) 329-344. A. Masson, A. Merlin, A. Pizzi, Comparative Kinetics of Induced Radical Autocondensation of Polyflavonoid Tannins. I. Modified and Nonmodified Tannins, J Appl Pol Sc Vol 60 263-269.

The “DualCEM” project: encapsulation of healing agent A. Masson, A. Merlin, A. Pizzi, Comparative Kinetics of Induced Radical Autocondensation of Polyflavonoid Tannins. I. Modified and Nonmodified Tannins, J Appl Pol Sc Vol 60 263-269. A. Formia, S. Irico, F. Bertola, F. Canonico, P. Antonaci, N. Pugno, J-M. Tulliani, Experimental analysis of the self-healing cement-based materials incorporating extruded cementitious hollow tubes, accepted-under revision in J Intell Mater Syst Struct (2015). A. Formia, P. Antonaci, S. Irico, F. Canonico, J-M- Tulliani, Extruded cementitious hollow tubes for healing agent delivery, in proceeding of DSCS 2015, Bologna.

156

Bibliography

M. Pelletier, R. Brown, A. Shukla, A. Bose, Self-healing concrete with a microencapsulated healing agent. Technical report. University of Rhode Island, (2011) Kingston, RI. H. Dong, H. Huang and G. Ye, Inorganic powder encapsulated in brittle polymer particles for self-healing cement-based materials, In: Proceeding of fourth International Conference on Self-healing Materials, 2013, p. 123. M. Pellietier, A. Bose, Patent Application Publication No. US2011/0316189 A1. A. Formia, S. Terranova, P. Antonaci, N. Pugno, J-M Tulliani, Setup of extruded cementitious hollow tubes as containing/releasing devices in self-healing systems, Materials 8 (2015) 1897-1923. H. Huang, G. Ye, Application of sodium silicate solution as self-healing agent in cementitious materials, International RILEM Conference on Advances in Construction Materials Through Science and Engineering (2011). J. LaRosa Thompson, M.R. Silshee, P.M. Gill, and B.E. Scheetzt, Characterization of sodium silicate sealer on concrete, Cem Concr Res 27 (1997) 1561-1567. A. Bovio, Ricerca ed applicazione di strategie per l’autoriparazione di materiali cementizi, Master Thesis, Università del Piemonte Orientale (2012). M. Benci, Studio e simulazione FEM del comportamento in flessione di un prototipo di un cemento self-healing micro-incapsulato, Master Thesis, Politecnico di Torino (2013). A. Formia, S. Irico, The DualCEM project evaluation of healing agents and development of encapsulation technologies for self-healing concrete, Cement International 13 (2015). T. D. Dang, S. W. Joo, Preparation of tapped shaped calcium alginate microparticles with sphericity control, Colloids and surfaces B: Biointerfaces 102, 2013, 766-771. L. Liu, F. Wua, X. Ju , R. Xie, W. Wang, C. Hui Niu, L. Chu, Preparation of monodisperse calcium alginate microcapsules via internal gelation in microfluidicgenerated double emulsions, J Colloid Interf Sci 404 (2013). C. Wu, W. Fanm M. Gelinsky, Y. Xiao, J. Chang, T. Friis, G. Cuniberti, In situ preparation and protein delivery of silicate-alginate composite microsphere with coreshell structure, J. R. Soc. Interface (2011) 8, 1804-1814.5–90. T.D. Danga, S.W. Jooa, Preparation of tadpole-shaped calcium alginate microparticles with sphericity control, Colloids and Surfaces B: Biointerfaces 102 (2013) 766– 771.

157

Bibliography

E. Mele, D. Fragouli, R. Ruffilli, G. L. De Gregorio, R. Cingolani, A. Anthanassiou, Complex architectures formed by alginate drops floating on liquid surfaces, Soft Matter 9 (2013) 6338. S. van der Zwaag, A. M. Grande, W. Post, S. J. Garcia, T. C. Bor, Review of current strategies to induce self-healing behaviour in fibre reinforced polymer based composites, Mat Sci Tech-Lond 30 (2014) 1633. A. Tabaković, W. Post, S. Garcia, E. Schlangen, Use of compartmented Sodium Alginate fibres as a healing agent delivery system for asphalt pavements, in proceedings of EMRS 2015, Warsaw. J. Slopasa, E.A.B. Koenders, S.J.Picken, A novel bio-based curing compound for cement-based materials, Application of superadsorbent Polymers and Other New admistrures in Concrete Construction, RILEM proceedings PRO 95 (2014). T. Lundstedt, E. Seifert, L. Abramo, B. Thelin, A. Nystrom, J. Pettersen, R. Bergman, Experimental design and optimization, Chemometrics and Intelligent Laboratory Systems 42 (1998) 3–40.

General conclusions and impacts

D.S Siegel et al, Commercial knowledge transfers from universities to firm: improving the effectiveness of university-industry collaboration, Journal of High Technology Management Research 14 (2003) 111-133. Hall, B., 2004, “University-Industry Research Partnerships in the United States”, in Contzen, Jean-Pierre, David Gibson, and Manuel V. Heitor (eds.), Rethinking Science Systems and Innovation Policies, Proceedings of the 6th International Conference on Technology Policy and Innovation, Purdue University Press. D. Gastaldi, F. Canonico, L. Capelli, L. Buzzi, E. Boccaleri, S. Irico, An investigation on the recycling of hydrated cement from concrete, Cem Con Compos 61 (2015) 2935. EP2878586 - Cementitious products obtainable from disposed concrete (2015), patent of UPO University and Buzzi Unicem.

158

Bibliography

Web site references www.titanpe.com/library/kb2503.htm, basic functions of titanium oxide description. www.fp/-intec.eu, INTEC Project web site. www.picada-project.com, PICADA Project web site. www.lifeminoxstreet.com/life, MINOx STREET Project web site. http://photopaq3.ircelyon.univ-lyon1.fr/, PHOTOPAQ Project web site. www.photocatalysis-federation.eu/, European Photocatalysis Federation. www.piaj.gr.jp/roller/en/, Photocatalysis Industry Association of Japan. www.fotokatalyza.org/, Czech Association of Applied Photocatalysis. www.shemat.eu/, SHEMAT Project web site. www.healcon.ugent.be/, HEALCON Project web site. www.youtube.com/watch?v=aJpusHsssJQ, Self-healing bio-based concrete channels in Ecuador in 2013. http:/www.rilem.net/gene/main.php?base=8750&gp_id=228, Encapsulation approach by University of Illinois in 2001.

159

Acknowledgements I would like to express my special appreciation and thanks to my supervisor Dr. Enrico Boccaleri. I would like to thank you for encouraging my PhD research and for allowed me to grow as a research scientist. I would like to thank my advisors of Buzzi Unicem, Dr. Fulvio Canonico and Dr. Daniela Gastaldi, which provided the vision, encouragement and advices on both research as well as on my career. I would also include all the research and development staff of Buzzi Unicem laboratories of Trino (VC) for taking part in this work. I wish to tank Dr. Klaus Droll, Wilhelm Dyckerhoff Institute, for his helpful discussions on photocatalysis and his precious support in revising my thesis, which was much appreciated. I would like to sincerely thank Mattia Lopresti who have helped me in in the study on self-healing materials, from experimental problems to laboratory support. I wish to thank Dr. Geo Paul for providing the SS-NMR spectra measurements and Dr. Valentina Gianotti for her contribution to experimental design study. A special thanks to my colleagues in “Università del Piemonte Orientale”, and particularly Vittoria, Mina, friends before colleagues, Francesco, Valentina, Luca and Chiara. I would like to tank Ing. Luigi Buzzi and Buzzi Unicem S.p.A. for the financial support of this PhD research. Words cannot be express how grateful I am to my mother and all my family for all they did for supporting me in these years and for what you have always made on my behalf.

Sara Irico 160

APPENDIX

161

Cement and Concrete Research 53 (2013) 239–247

Contents lists available at ScienceDirect

Cement and Concrete Research journal homepage: http://ees.elsevier.com/CEMCON/default.asp

Investigation of the microstructural evolution of calcium sulfoaluminate cements by thermoporometry Sara Irico a,⁎, Daniela Gastaldi b, Fulvio Canonico b, Giuliana Magnacca c,d a

Department of Chemistry, Università degli Studi di Torino, Via Giuria 7, 10125 Torino, Italy Buzzi Unicem S.p.A., Via Luigi Buzzi 6, 15033 Casale Monferrato, AL, Italy Department of Chemistry, Università degli Studi di Torino, Via Giuria 7, 10125 Torino, TO, Italy d NIS Centre of Excellence, Università degli Studi di Torino, Via Giuria 7, 10125 Torino, TO, Italy b c

a r t i c l e

i n f o

Article history: Received 14 June 2012 Accepted 25 June 2013 Keywords: Thermal analysis (B) Microstructure (B) Characterization (B) Sulfoaluminate (D) Hydration (A)

a b s t r a c t Thermoporometry was applied to the investigation of the microstructural evolution of cementing systems. A pure calcium sulfoaluminate cement – CSA – and one mixed CSA–Portland system, together with a reference Portland cement were considered. Specimen preparation was carefully optimized in order to minimize any structural damage and the repeatability of results was checked through the utilization of inorganic standard. Nitrogen adsorption/desorption was used for comparison. A wide set of information could be acquired regarding the microstructure of the investigated materials: a) both the CSA and the mixed CSA–Portland system mainly revealed ink-bottle pores; b) a much more rapid development of hydrated structures was observed for all CSA cements than for the Portland cement; c) melting and freezing curves allow to gain information about the pore size distribution and the presence of pore entries of preferential size, about tortuosity and connectivity of the cement microstructure, and about the existence of isolated pores. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction The investigation of the microstructure of a cementing system is actually one of the most important research activities in the science of cement, as far as it influences long term physical and mechanical performances of mortar and concrete and affects the mass transport properties and, subsequently, the durability features. Materials with the same total pore volume can even show very different performances depending on their different pore size distributions or on peculiar pore network and connectivity [1]. Moreover, many different production features can modify the pore structure of cement based materials, such as water–cement ratio, curing conditions, type and amount of filler [2–5], and can be used for producing materials with a tailored microstructure. The investigation of the hydrated cement microstructure is also important as it supplies a realistic picture of the progress of the hydration process. The characterization of the microstructure of cement paste is complicated by the fact that pores can have different shapes, sizes and connectivity; moreover, most of the techniques require pretreatment of the specimen which can modify, or even damage the microstructure of cement paste, especially at early hydration stages. A wide range of pore sizes can be present in a mortar or in a concrete and a defined classification is not possible because the distribution of ⁎ Corresponding author. Tel.: +39 3495458760. E-mail address: [email protected] (S. Irico). 0008-8846/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.cemconres.2013.06.012

porosities is a continuum [1]. In these cement-based materials the difference between gel and capillary pores is arbitrary. A differentiation is generally done between [1]: gel pores (Ø b 10 nm), intrinsic of C–S–H1 structure; capillary pores (2 nm b Ø b 10 μm), mainly consisting of empty channels initially occupied by hydration water. Concerning a hydrated Portland cement paste, the most recent C–S–H pore structure descriptions were supplied by Jennings and McDonald [6,7]. As far as the hydration time increases, the amount and distribution of gel and capillary pore change: the volume of capillary pores decreases when they are filled with the products of hydration whereas the gel pore volume increases according to the development of C–S–H. Winslow and Diamond [8] observed that the pores with pore sizes between 10 and 100 nm are substantially present at all stages of aging. A wide variety of techniques has been used for the characterization of the pore structure for cement pastes. According to the classification proposed by Aligizaky [1], they can be divided into: a) indirect methods (such as mercury intrusion porosimetry — MIP, nitrogen adsorption/desorption at 77 K — NAD, nuclear magnetic resonance — NMR), in which the pore structure is deduced from properties such as the adsorptive capacity, density, etc.; b) direct methods (such as

1 Note that standard cement nomenclature is followed here, whereby C = CaO, S = SiO2, A = Al2O3, F = Fe2O3, S = SO3 and H = H2O.

240

S. Irico et al. / Cement and Concrete Research 53 (2013) 239–247

microscopy and image analyses) based on direct observation of the specimen. The more commonly applied characterization technique, in the study of cement paste microstructure, is the MIP, which allows the measurement of pore radii in a range between 2.5 nm and 100 μm. This technique relates the Hg pressure necessary to fill the pores of the cement paste to the pore size. Its application in the cement field is quite wide [8–11] and it is considered, in the cement laboratories, the reference technique. Nevertheless, it has been the subject of several discussions in the scientific community: some authors [12] state that, in cement-based systems, most of the mercury enters the pore structure only when the threshold diameter is reached and, subsequently, large and small pores are filled indiscriminately (the so-called ink-bottle effect [12,13]). On the other hand, the threshold pore width may provide a satisfying indicator of material durability, as it has an important influence on diffusion and permeability features of hydrated cement pastes [11]. Moreover, a recent study [14] involving multicycle MIP and Wood's metal impregnation offers an interesting possibility to overcome some of the limitation related to the ink-bottle effect. A valid alternative technique is represented by NAD (77 K). Its use for the study of cement paste microstructure and pore volume was firstly carried out by Blaine and Valis [15]. The technique is based on the fact that when a porous solid is exposed to a gas of a certain volume and under a finite pressure, it begins to adsorb gas molecules on the outside surface and inside its pores. The adsorption process goes through the following steps: 1) firstly, a monomolecular gas layer (monolayer) on the internal walls of the solid forms; 2) this monolayer attracts further gas molecules forming a multilayer; 3) when the volume of the gas inside the pore approximates the volume of the pore itself, the capillary condensation of the gas takes place. When the gas pressure is reduced, desorption of the gas starts. Information about pore volume, pore size distribution and the specific surface of the investigated material can be deduced when analyzing the adsorption/ desorption curve as a function of the gas relative pressure: many data analysis methods and equations have been developed with different assumptions. For example, the shape of the adsorption hysteresis can be related to the nature of the pore structure [16,17] which means cylindrical, spheroidal, ink-bottle shape, etc.; in the following, the Barret– Joyner–Halenda (BJH) method [18] will be applied for determining the pore size distribution. One of the most promising techniques to approach the study of the microstructure of cement paste is thermoporometry (TPM), even if, nowadays, it is not commonly used. TPM is a technique based on the following phenomenon: when a liquid (water) is confined in a mesopore, it will freeze at a lower temperature than in bulk, and its freezing temperature depends on the size of the mesopore itself [19]. The temperature phase transition is followed by low-temperature differential scanning calorimetry: through a mathematical equation it is possible to convert the DSC pattern into a pore size distribution. Few studies are present in the literature regarding different applications of low temperature micro-calorimetry and low temperature differential scanning calorimetry: in the first case [20] the ice formation and melting in aged hydrated cement pastes have been investigated in order to gain information about the freeze/thaw behavior of cement/concrete; in the second case [21], the LT-DSC has been applied to the study of the evolution of hydrated cement pastes through the evaluation of the Free Water Index (FWI) calculated from the enthalpy of fusion of the residual non-reacted water. The relationship between low temperature behavior of saturated porous building materials and pore structure has been known since the 70s [22], even if the method still needs to be fully understood. With respect to other techniques, TPM shows the undoubted advantage that the stress exerted on the pore surface by water ice is ten times lower than mercury in MIP [23], thus resulting in lower risk of damaging the sample.

As described below, through TPM it is possible to obtain the distribution of the pore interior from the melting thermal cycle and the distribution of pore entries from the freezing thermal cycle [24]: this is one of the main advantages of TPM, i.e. it is possible to obtain porosimetric distribution, morphological information and also to make assumptions regarding the pore connectivity in one single and relatively simple measurement. The pore volume is directly related to the amount of ice forming at a specific temperature [25]. The application of the thermoporometric technique in cement based material science is limited to a few academic examples and has been rarely used before for comparing the microstructural evolution of different cement pastes. In this paper an applicative utilization of TPM concerning the study of the microstructural evolution of hydration of different types of cement is presented. The combined utilization of the TPM and NAD allowed similarities and differences in the hydration mechanism and pore structure development of three different cementing systems to be highlighted: one reference ordinary Portland cement (OPC) and two mixed cements based on an low CO2-footprint calcium sulfoaluminate binder [26] whose mechanical enhanced properties and environmentally friendly production features have recently promoted extended research interest. The study of these materials is particularly appealing as their microstructure is very different from that of OPC due to the peculiar chemistry of sulfoaluminate system. Even if some studies have been done to put microstructural features in relation to macroscopic behavior [27–29], no structural models are present in the literature. 2. Materials and methods/experimental Three cement samples, supplied by Buzzi Unicem S.p.A., were investigated: one calcium sulfoaluminate cement identified by the commercial code SAcement SR03 (CSA SR03), one CSA–Portland mixed cement (CSA MIX) and an ordinary Portland cement (OPC) CEM II/A-LL 42.5 R (Portland limestone cement), used as a reference. CSA SR03 is pure CSA cement composed of 83% CSA clinker and 17% natural gypsum; the CSA MIX sample is a ternary mixture consisting of CSA clinker, natural gypsum and CEM II/A-LL 42.5 R in the mass ratio 28:12:60. The hydrated cement pastes were prepared using demineralised water in a water-to-cement ratio (w/c) 0.5 and mixed with a Vortex shaker in apposite vials for 1 min. The pastes were then poured in 80 × 80 mm vessels and shaped into thin layers (a few millimeters) in order to facilitate the preparation of the specimen for the following analyses; after 24 h curing in humid chamber (RH N 90% and T = 20 °C), they were finally stored in demineralised water at room temperature until use (1, 7, 28, 56, 90 days). Mineralogical investigations were performed utilising XRD analyses, using a Bruker AXS D4 Endeavor diffractometer working in Bragg–Brentano geometry, equipped with a ceramic X-ray tube KFF (Cu Kα radiation) and a “Lynx Eye” dispersive detector. Refinement for semi-quantitative analyses was conducted by the Rietveld method using the Topas 2.0 package (commercially supplied by Bruker AXS): structural models for all phases were taken from the software database. Mineralogical compositions of the obtained cement are summarized in Table 1. Chemical analyses were performed by dispersive X-ray fluorescence, using a Panalytical Axios spectrometer on pressed powder: the element content, expressed as a percentage of the corresponding metal oxide, is summarized in Table 2. Thermogravimetric analysis (TG) investigation of hydrated cement pastes has been carried out by means of a Mettler Toledo TG/DSC1 using 70 μl alumina crucible and heating the samples up to 950 °C (10 °C/min) in air flow (80 ml/min). NAD analyses (77 K) were conducted with the gas volumetric analyzer Micromeritics ASAP 2020. The cement pastes and a mesoporus


241

pores [24,25]. The conversion from DSC data to thermoporometric distribution is performed according to the Gibbs–Thomson equation:

Table 1 Mineralogical composition of the investigated cement samples. Mineralogical phase

ICDD

CEM II/A-LL 42.5 R

CSA SR03

CSA MIX

C3 S C2 S C3 A C4AF Yeelimite Calcite Gypsum Minor phases

31-0331 33-0302 38-1429 30-0256 30-0256 05-0586 36-0432 –

62.5 6.2 3.0 11.1 – 11.6 3.6 1.7

– 17.7 10.1 3.2 41.1 1.6 16.6 2.0

44.9 10.9 2.4 7.8 9.3 2.3 12.4 6.7

inorganic standard (SiO2/Al2O3 supplied by Micromeritics used as reference) were pretreated under vacuum (residual pressure 10–2 mbar) at 60 °C for 24 h. To measure the porosity of the samples by TPM, the samples were immersed in isopropanol for 12 h prior to the measurement in order to replace the hydration water (a solution rich in ions such as Na+, 2+ K+, SO2− , OH−, etc.) with alcohol and stop the hydration 4 , Ca reactions. The samples were then dried in an oven at 40 °C for 2 h. Specimen preparation for thermoporometric measurements was performed according to the procedure proposed by Sun and Scherer [24]: first of all, sample fragments (sizes smaller than 5 mm) were evacuated in a glass reactor for 2 h (KNF Neuberger vacuum pump, series N 026.32, 20 mbar final vacuum); subsequently, they underwent backfilling with demineralised water for 30 min. This procedure is essential to obtain accurate thermoporometric analyses: in this way, it is assumed that the sample saturation occurs rapidly through capillarity and the saturation water has low ion concentration. The saturated samples were finally placed in a sealed aluminium DSC pan of 40 μl; a drop of kerosene was placed on the bottom of the pan in order to ensure good thermal contact. Thermal analyses were performed by using a Mettler Toledo LT-DSC (DSC1 model, equipped with the full-range FRS5 sensor, allowing digital resolution of the measurement signal lower than 0.04 μW); the temperature cycles were set through the DSC control software (STARe Software) and regulated by an electro valve; and liquid nitrogen was used as refrigerating liquid. The thermal DSC program consists of four steps: 1. Rapid quenching from 25 °C to 5 °C at a rate of −10 °C/min followed by an isotherm at 5 °C for 10 min in order to obtain good thermal stability inside the measuring cell; 2. Quenching from 5 °C to − 50 °C at a rate of − 2 °C/min: during this step both bulk and confined water freeze; 3. Melting cycle from −50 °C to − 0.2 °C at a rate of 0.5 °C/min: only the ice crystals in the mesopores melt, the macroscopic ice persists; this thermodynamic process depends only on the interior size of pores; 4. Freezing cycle from − 0.2 °C to − 50 °C at a rate of −0.5 °C/min: this slow quenching allows the ice nucleation and crystal growth inside the mesopores; this thermodynamic process depends on pore network and on pore entry radii.

1 1 T M ΔSf ¼ ∫ dT Rp 2γ sl T M ð∞Þ vl

ð1Þ

where ΔSf is the solidification entropy of the liquid phase, vl is the molar volume of liquid, γsl is the interface energy of solid/liquid phase, TM is the melting temperature of a crystal inside a pore with a Rp radius and TM(∞) is the melting temperature of a macroscopic crystal (i.e. with infinite radius) — approximate to 0 °C. This method was expanded by Brun et al. [19] using pure water to arrive at this simple expression: Rp ðnmÞ ¼ −

64:67 −0:23 þ δ ΔT

ð2Þ

where ΔT = TM − TM(∞) represents the under-cooling of water confined in mesopores with respect to the macroscopic system, whereas δ is the thickness of the non-freezable water layer (in cement based materials this value is 1.2 nm [24]). Eq. (2) has been directly used for conversion of data from DSC to thermoporometric distribution: each TM value corresponds to one Rp value. In this way, Watt/g vs T distribution is converted into Watt/g vs Rp. No baseline correction was required since the signal of a reference empty pan was simultaneously acquired during the measurement. To test the accuracy and repeatability of the measurements, a sample of SiO2/Al2O3 was used as an inorganic standard in TPM and NAD. The thermoporometric distribution obtained from this reference sample, extracted from the freezing cycle, is compared to the BJH distribution, obtained from the NAD desorption curve: in Fig. 1 it is possible to appreciate the close agreement between the two distributions. 3. Discussion about the effect of sample pretreatment and TPM procedure When porous structure is investigated, it is important to preserve as much as possible the microstructure of the materials during sample pretreatment. Some considerations regarding the effect of drying, water saturation, as well as TPM thermal cycles, on mineralogy and microstructure of cement pastes are hereafter discussed, in order to support the accuracy of the obtained results.

The thermodynamic processes are evaluated in Watt/g vs. T, and the amount of released heat is directly dependent on the volume of Table 2 Chemical composition of the investigated cement standard samples (expressed as % in weight). Oxide

CEM II/A-LL 42.5 R

CSA SR03

CSA MIX

SiO2 Al2O3 Fe2O3 CaO SO3 Other

18.1 4.2 3.2 62.4 2.8 9.0

8.3 26.0 2.4 42.0 12.7 7.1

11.9 10.3 2.9 53.2 10.9 10.6

Fig. 1. Comparison of thermoporometric pore size distribution of SiO2/Al2O3 standard derived from DSC freezing cycle and BJH distribution extracted from the nitrogen desorption curve.

242


3.1. Effect of drying Most of the available techniques for the evaluation of the microstructure of cement pastes require dry samples to be investigated, therefore a pretreatment to arrest the hydration is in any case necessary, especially when fresh pastes are considered. Many methods have been proposed for the drying of cement paste samples: oven drying, vacuum drying, freeze drying and solvent exchange are the most widely known and applied. It is a diffused opinion that solvent replacement is the technique which guarantees the better preservation of the pore structure in the finest pore region [30]. Among the available solvents, isopropanol revealed to be the best one, being a relatively small molecule, miscible in water and easily evaporable in air [31].

3.2. Effect of water saturation The TPM vacuum treatment, preliminary to water saturation, is performed by means of a laboratory vacuum pump: the residual pressure is higher than 20 mbar, almost three orders of magnitude higher than that required during preparation of samples for MIP or NAD analyses; we can therefore assume that the alteration of the hydrated layer in cement paste is less relevant in TPM than in the other characterization techniques. A combined XRD/TG characterization of fragments of cement pastes before and after pretreatment (drying and water saturation) has been carried out on two samples (CEM II/A-LL 42.5 R and CSA SR03 15 days aged), in order to verify any alteration of the overall mineralogical composition. XRD patterns are shown in Fig. 2, revealing no significant differences between the mineralogical composition of the hydrated cement paste and of the corresponding saturated sample. The amount of calcium carbonate was evaluated by means of TG analysis, after measuring the weight loss due to decarbonation of calcite around 750 °C: in the CEM II/A-LL 42.5 R samples, the amount of calcium carbonate was 16.9 ± 1 and 17.6 ± 1% in weight, before and after the saturation respectively; in the CSA SR03 samples,

5.3 ± 1 and 4.4 ± 1% in weight of calcium carbonate were found before and after the saturation. In both cases, the carbonation of the samples during the saturation pretreatment is minimized. 3.3. Effect of TPM thermal cycles This effect was evaluated on CEM II/A-LL 42.5 R and CSA SR03 samples by means of NAD analysis. NAD measurements performed on cement paste samples before and after TPM thermal cycles indicate that no substantial modifications are induced on sample mesoporosity (results not reported for the sake of brevity). 4. Results and discussion The microstructural evolution of cement pastes was measured at different ages of hydration (1, 7, 28, 56, 90 days). It was observed that all the DSC patterns of the samples analyzed in this study showed a large hysteresis, Fig. 3, between the melting and freezing temperatures of water confined in mesopores: this observation is in agreement with literature data [19] and indicates the presence of cylindrical pore shape. Analyzing the shape of the hysteresis loop shown by the NAD isotherm, it is possible to evaluate the morphology of the pores according to the IUPAC classification [32] and to obtain BJH distributions to support the interpretation of thermoporometric data. The overall elaboration carried out from gas-volumetric data was performed on the adsorption branch of the isotherm, since the sudden closing of the loop occurring in the desorption branch at 0.42 of relative pressure p/p° (Figs. 4a, 5a, 6a) is not due to the morphology of the sample but is due to the physical behavior of nitrogen (the tensile strength causes the emptying of the pores): thus the desorption branch data cannot be considered as a reliable description of the samples. It is expected that the BJH distributions obtained from the adsorption curve agree with the thermoporometric distribution obtained by the melting curve [24], as both models consider the interior pore size distribution.

Fig. 2. XRD patterns of hydrated cement pastes before and after the sample pretreatment obtained from CEM II/A-LL 42.5 R (below) and from CSA SR03 (above) samples (E = ettringite; P = calcium hydroxide; C = calcite; S = anhydrous silicates).


243

Fig. 3. DSC patterns of: (a) CEM II/A-LL 42.5 R after 28 days from hydration; (b) CSA SR03 after 28 days from hydration. Both diagrams show a large hysteresis between melting and freezing temperatures, this phenomenon indicates the presence of cylindrical pores.

4.1. Ordinary Portland cement type CEM II/A-LL 42.5 R The microstructural evolution of an OPC is strongly related to the development of the amorphous calcium silicate hydrate gel: C2 S þ H2 O→Cx SHy þ nCH

ðiÞ

C3 S þ H2 O→Cx SHy þ mCH:

ðiiÞ

In the CxSHy, CaO/SiO2 ratio can range between 1.6 and 2.0, while in a fully saturated cement paste, H2O/SiO2 = 4 is generally assumed [33]. According to a recently developed structural model [6,7], the C–S− H is constituted of globular units resulting from the aggregation of solid mono-layers. The NAD isotherms obtained from the hydrated CEM II/A-LL 42.5 R pastes between 1 and 90 days are compared in Fig. 4a. In all of them a large hysteresis between the adsorption and desorption curves indicates the presence of ink-bottle mesopores: this is in agreement with previous morphological/structural studies (Baroghel using water adsorption/desorption [34], Moro with intrusion/extrusion of mercury [13]). The BJH distributions, reported in Fig. 4b, show the presence of mesopores in a wide range between 1 nm and 100 nm, related to capillary pores. As far as hydration goes on, the aggregation of C–S–H globules allows the development of small gel pores (SGP), with a radius between 1 and 4 nm, in agreement with the literature data [6]; this is particularly evident for aging between 56 and 90 days. The widening in the amplitude of the loop at 90 days does indeed suggest a considerable increase in cumulative pore volume, in contrast to the results normally obtained by MIP analysis [11,29], affected by the limitation due to its lower limit of detection of 2.5 nm.

Fig. 4. (a) NAD isotherms at 77 K and (b) BJH pore distributions of CEM II/A-LL 42.5 R paste.

The thermoporometric distributions (Fig. 7) show that in the first 7 days of hydration the microstructure of cement pastes is mainly characterized by capillary pores. The interior pore size distribution obtained from the melting curve (Fig. 7a) reveals the presence of pores larger than 10 nm: the absence of a maximum in the distribution up to 7 days indicates that pores larger than 100 nm are also present, that cannot be discriminated from bulk water. In this case the ice confined in these pores persists during the whole melting step and acts as a crystallization seed during thermal inversion: due to this incomplete fusion, the re-crystallization of water happens instantaneously. This phenomenon is also evident in the thermoporometric distribution of pore entry size (see Fig. 7b) in which a wide distribution between 10 and 100 nm is detectable. Starting from 28 days of hydration the pore entry size distribution exhibits pore entries between 8 and 20 nm. These pores arise from the aggregation of different globular structures of C–S–H and are called large gel pores (LGP), according to literature data [6]. They act as entries to larger pores (rp about 60 nm reported in Fig. 7a) and this is further confirmation of the ink-bottle pore morphology.

244


Fig. 6. (a) NAD isotherm at 77 K and (b) BJH pore distributions of CSA MIX paste. Fig. 5. (a) NAD isotherms at 77 K and (b) BJH pore distributions of CSA SR03 paste.

4.2. Sulfoaluminate cement SR03 The main product of hydration of a CSA cement is ettringite, a crystalline mineral consisting of elongated hexagonal prismatic structures; amorphous Al(OH)3 forms as a by-product according to the following reaction [35]: C4 A3 S þ 2CS þ 38H→C6 AS 3 H32 þ 2AH3 :

ðiiiÞ

Aluminium hydroxide has a gel structure characterized by a large amount of water molecules in the surrounding: it is known that porosimetric characterization of gels is difficult as their soft structure is strongly affected by external induced strains such as mercury pressure in MIP or ice growing and nucleation in TPM [23]. Even though Al(OH)3 is a secondary hydration product, its amount in a CSA cement paste could be around 15% [36]: this aspect must be therefore kept into account. The NAD experiments, performed in order to evidence possible modifications induced on the microstructure of CSA SR03 during TPM analysis (paragraph 3), revealed that no substantial alteration of mesoporosity occurred.

The crystal growth of ettringite occurs rapidly (most of the ettringite in a CSA cement has already developed after 24 h of hydration) and is known to be thermodynamically stable [37]. These crystals do not undergo an intrinsic structural evolution during aging: they are generated directly in their final crystalline structure providing high stability and subsequently initial high-performances in concrete. Mehta observed that ettringite crystals in these cements were about 6–8 μm long and 1/2 μm wide at 24 h aging [38]. On the basis of these considerations, a large part of the porosity in these cements should be attributed to pores created among the ettringite crystals. The hysteresis loops of the NAD isotherms (Fig. 5a) are clearly different from those observed for the OPC samples, being wider and not ascribable to a specific pore shape based on IUPAC classification. The thermoporometric distributions show that after 24 h of hydration the microstructure is fully developed with pore entry size between 20 nm and 30 nm (see Fig. 8b) and a bimodal distribution of interior pore size around 10 nm and 100 nm (Fig. 8a). This porosity is linked to the formation of the ettringite crystals in the early steps of hydration.


Fig. 7. Thermoporosimetric distributions of CEM II/A-LL 42.5 R paste: (a) pore interior distributions, obtained from melting DSC cycle; (b) pore entry size distributions, obtained from freezing DSC cycle.

As far as the cement paste hydration goes on, even the calcium silicates (C2S) start to dissolve and, according to Ref. [36], to react with Al(OH)3 gel forming strätlingite (C2ASH8): C2 S þ AH3 þ 5H→C2 ASH8 :

ðivÞ

Strätlingite fills up the pores within ettringite crystals: experimentally, this results in the disappearance of the 20 nm peak in the pore entry size distribution. A further modification of the global microstructure with the development of new pore entries between 20 and 30 nm up to 56 and 90 days from the hydration is observed: this can be ascribed to a reorganization of internal porosity of hydrated phases, similar to what observed in the CEM II/A-LL 42.5 R, more than to the formation of new hydrated products. The narrow peak around 2.5 nm in the pore entry size distribution is supposed to be not indicative of a realistic pore size: as a matter of fact, this peak is a calorimetric effect due to the formation of a crystallization seed (through homogeneous nucleation) below the critical

245

Fig. 8. Thermoporometric distributions of CSA SR03 paste: (a) pore interior distributions, obtained from melting DSC cycle; (b) pore entry size distributions, obtained from freezing DSC cycle.

temperature of −38 °C [39] in isolated pores: in this case the freezing temperature of water cannot be univocally related to pore size. The thermoporometric distributions obtained from melting cycles are in clear agreement with the BJH distributions extracted from the adsorption curve (Fig. 5b), both showing a mesopore distribution around 10 nm.

4.3. Sulfoaluminate cement MIX In the CSA MIX sample, the hydration of C4A3S and calcium sulfate occurs in the presence of lime (namely, calcium hydroxide generated by the OPC component hydration), according to the following reaction: C4 A3 S þ 8CS þ 6C þ 96H→3C6 AS 3 H32 Differently from the pure CSA SR03 system, no Al(OH)3 gel forms and, subsequently, no strätlingite [40].

246


According to some authors [37,38] a different kind of ettringite is formed in the presence of lime, being colloidal and not formed by long lath-like crystals as in the previous system. This kind of ettringite is characterized by high specific surface and this, together with its negative charge, is considered responsible for attracting a large number of water molecules around the crystal itself: the repulsion of these charged particles from each other probably causes an expansion of the system, without any change in the crystal lattice of ettringite. Mehta [38] observed that the ettringite crystals that formed in the presence of lime after 24 h from hydration were about 1 μm long and 1–4 μm wide. The shape of NAD loops (in Fig. 6a) reveals the presence of ink-bottle pores, similar to those described for the CEM II/A-LL 42.5 R. The thermoporometric distributions obtained at 24 h and 7 days from hydration, reported in Fig. 9a and b, show that the CSA MIX microstructure is very similar to that observed for CEM II/A-LL 42.5 R, with capillary pores in the region of 20 nm up to more than 100 nm. The expansive behavior of this material [37] is probably responsible for the increment of pore volume that is detectable in the BJH distribution after 28–56 days of hydration (Fig. 6b).

This system is also constituted by a consistent fraction of C3S and C2S whose hydration is not significantly affected by the presence of the sulfoaluminate component. Silicate hydration leads to the formation of C–S–H, whose structural reorganization, previously discussed, is responsible for the formation of new pore entries of around 20– 40 nm observed at 56 and 90 days in the thermoporometric distribution (Fig. 9b). After 90 days the BJH distribution of CSA MIX again exhibits a considerable increase of pore volume, in particular a large amount of SGP pores with a radius between 1 and 4 nm develops and the overall distribution becomes very similar to that described for the CEM II/A-LL 42.5 R (Fig. 6b). 5. Conclusions In this paper, the thermoporometric technique was proposed to study the micro-structural evolution of low CO2 footprint binders based on CSA cement, in comparison to an ordinary Portland cement. From the experimental point of view, the main results that were obtained concern: a) the development and the validation of TPM as a promising method for cement paste characterization, in comparison to NAD; b) the evaluation of the influence of sample pretreatment and TPM analysis on the microstructure of cement pastes; c) the investigation of the microstructural evolution of three types of cement: two calcium sulfoaluminate based binders and one CEM II/A-LL 42.5 R. The most relevant conclusions about the thermoporometric technique are: − the accuracy and repeatability of the measurements were demonstrated by the good agreement with the thermoporometric distribution and the NAD BJH pore distribution on a SiO2/Al2O3 reference standard; − it was verified that the sample pretreatment, as it was performed, was suitable for minimizing the microstructural damaging of hydrated cement paste; − although ice growing occurring during TPM thermal cycles could induce modification in the microstructure of cement pastes, especially when hydrated gel phases are formed (as in the case of CSA SR03 sample), no significative alteration in the mesoporous framework are induced; − TPM is a powerful experimental method for characterizing the microstructures of cement-based material; − the use of TPM, supported by NAD analyses, allows the better understanding of the cement microstructure evolution and an investigation into the depth of the microstructure behavior during hydration; − the main advantage of TPM probably consists of recovering information regarding the pore size, morphology and connectivity of pores through the simultaneous analyses of pore entry size and pore interior size distributions. The micro-structural investigation demonstrated important differences among the three investigated cements. The results that were obtained highlight that:

Fig. 9. Thermoporometric distributions of CSA MIX paste: (a) pore interior distributions, obtained from melting DSC cycle; (b) pore entry size distributions, obtained from freezing DSC cycle.

− the mesopore distribution in the CEM II/A-LL 42.5 R paste is in gradual and continuous evolution up to 90 days from hydration. The microstructure of this material is strictly related to the intrinsic modifications of the C–S–H nanostructure during hydration (the formation of globules, their aggregation and rearrangement, etc.); − the rapid growth of ettringite crystals in the CSA SR03 allows the development of a well-defined microstructure even during the first 24 h of hydration; the modification of mesopore distribution


observed as far as hydration continues, particularly regarding the pore entries, is mainly due to the formation of secondary hydration products (mainly strätlingite forming from the reaction of aluminium hydroxide and C2S); − the CSA MIX is characterized by a behavior which is intermediate between the two other cement samples: the pore size distribution is, on the whole, similar to that observed for the OPC, but welldefined entries develop after 56 and 90 days as in the CSA SR03 sample; moreover, its different hydration features and its expansive behavior causes a significant increase of the overall mesopore amount with aging time. Finally, the work demonstrates that thermoporometric techniques can be successfully and reproducibly used for studying the microstructure development of cement pastes but also for obtaining additional and complementary information such as morphology and the connectivity of pores, isolated pores, pore entry size, not directly achievable with a conventional approach based on MIP or other porosimetric features. Acknowledgments The authors wish to thank G. W. Scherer and Z. Sun (Princeton University, NJ – USA) for the helpful discussion regarding sample treatment and data interpretations. References [1] K.K. Aligizaki, Pore Structure in Cement Materials, Taylor & Francis, New York, 2006. [2] P. Parcevaux, Pore size distribution of Portland cement slurries at very early stages of hydration (influence of curing temperature and pressure), Cem. Concr. Res. 14 (1984) 419–430. [3] P. Pipilikaki, M. Beazi-Katsioti, The assessment of porosity and pore size distribution of limestone Portland cement pastes, Constr. Build. Mater. 23 (2009) 1966–1970. [4] H.N. Atahan, O.N. Oktar, M.A. Taşdemir, Effects of water–cement ratio and curing time on the critical pore width of hardened cement paste, Constr. Build. Mater. 23 (2009) 1196–1200. [5] M. Frías, J. Cabrera, Pore size distribution and degree of hydration of metakaolin– cement pastes, Cem. Concr. Res. 30 (2000) 561–569. [6] H.M. Jennings, Refinements to colloid model of C–S–H in cement: CM-II, Cem. Concr. Res. 38 (2008) 275–289. [7] P.J. McDonald, V. Rodin, A. Valori, Characterization of intra- and inter-C–S–H gel pore water in white cement based on an analysis of NMR signal amplitude as a function of water content, Cem. Concr. Res. 40 (2010) 1656–1663. [8] D.N. Winslow, S. Diamond, A mercury porosimetry study of the evolution of porosity in Portland Cement, J. Mater. 5 (3) (1970) 564–585. [9] G. Bernardo, L. Buzzi, F. Canonico, M. Paris, A. Telesca, G.L. Valenti, Microstructural Investigations on Hydrated High-Performance Cements Based on Calcium Sulfoaluminate, International Congress on the Chemistry of Cement, Montreal, 2007. [10] A.B. Abell, K.L. Willis, D.A. Lange, Mercury intrusion porosimetry and image analysis of cement-based materials, J. Colloid Interface Sci. 211 (1999) 39–44. [11] R.A. Cook, K.C. Hover, Mercury porosimetry of hardened cement pastes, Cem. Concr. Res. 29 (1999) 933–943. [12] S. Diamond, Mercury porosimetry: an inappropriate method for the measurement of pore size distributions in cement-based materials, Cem. Concr. Res. 30 (2000) 1517–1525. [13] F. Moro, H. Böhni, Ink-bottle effect in mercury intrusion porosimetry of cement-based materials, J. Colloid Interface Sci. 246 (2002) 135–149. [14] J. Kaufmann, Pore space analysis of cement-based materials by combined nitrogen sorption — Wood's metal impregnation and multi-cycle mercury intrusion, Cem. Concr. Res. 32 (2010) 514–522.

247

[15] R.L. Blaine, H.J. Valis, Surface available to nitrogen in hydrated Portland cements, JRNSB 42 (1949) 257–267. [16] J.H. De Boer, The structure and properties of porous materials, in: D.H. Everett, F.S. Stone (Eds.), Proceeding of the Tenth Symposium, Colston Research Society University of Bristol, Butterworths Science Pubblications, London, 1958, pp. 68–94. [17] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquérol, T. Siemieniewska, Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity, Pure Appl. Chem. 57 (4) (1985) 603–619. [18] E.P. Barret, L.G. Joyner, P.P. Halenda, The determination of pore volume and area distribution in porous substances. I. Computations from nitrogen isotherm, J. Am. Chem. Soc. 73 (1) (1951) 373–380. [19] M. Brun, A. Lallemande, J. Quinson, C. Eyraud, A new method for the simultaneous determination of the size and the shape of pores: the thermoporometry, Thermochim. Acta 21 (1977) 59–88. [20] C. Le Sage De Fontenay, E.J. Sellevold, Ice formation in hardened cement paste — I, DBMC, ASTM STP 691, in: P.J. Sereda, G.G. Litvan (Eds.), American Society for Testing and Materials, 1980, pp. 425–438. [21] A. Damasceni, L. Dei, E. Fratini, F. Ridi, S. Chen, P. Baglioni, A novel approach based on differential scanning calorimetry applied to the study of tri-calcium silicate hydration kinetics, J. Phys. Chem. B 106–44 (2002) 11572–11578. [22] G. Fagerlund, Determination of pore size distribution from freezing point depression, J. Mater. Struct. 6 (33) (1973) 215–225; E.J. Sellevold, D.H. Bager, Low temperature calorimetry as a pore structure probe, 7th ICCC, vol. IV — Paris, 1980. [23] W.G. Scherer, D.M. Smith, D. Stein, Deformation of aerogels during characterization, J. Non-Cryst. Solid 186 (1995) 309–315. [24] Z. Sun, G.W. Scherer, Pore size and shape in mortar by thermoporometry, Cem. Concr. Res. 40 (2010) 740–751. [25] D.P. Bentz, Capillary porosity depercolation/repercolation in hydrating cement pastes via low-temperature calorimetry measurements and CEMHYD3D modeling, J. Am. Ceram. Soc. 89 (2006) 2606–2611. [26] Établissement Poliet & Chausson, Brevet d'invention No 780747. [27] G. Bernardo, A. Telesca, G.L. Valenti, A porosimetric study of calcium sulfoaluminate cement pastes cured at early ages, Cem. Concr. Res. 36 (2006) 1042–1047. [28] G. Bernardo, L. Buzzi, F. Canonico, F. Paris, M. Telesca, G.L. Valenti, Microstructural investigations on hydrated high-performance cements based on calcium sulfoaluminate, in: J.J. Beaudoin, J.M. Makar, L. Raki (Eds.), 12th International Congress on the Chemistry of Cement, Montreal, 2007, p. W3 11.4. [29] D. Gastaldi, F. Canonico, F. Capelli, M. Bianchi, M.L. Pace, A. Telesca, G.L. Valenti, Hydraulic behaviour of calcium sulfoaluminate cement alone and in mixture with Portland cement, International Congress on the Chemistry of Cement, Madrid, 2011. [30] L. Konecny, S.J. Naqvi, The effect of different drying techniques on the pore size distribution of blended cement mortars, Cem. Concr. Res. 23 (1993) 1223–1228. [31] J. Zhang, G.W. Scherer, Comparison of methods for arresting hydration of cement, Cem. Concr. Res. 41 (2011) 1024–1036. [32] Iupac, Reporting physisorption data for the gas/solid system with special reference to the determination of surface area and porosity, Pure Appl. Chem. 57 (4) (1985) 603–619. [33] H.F.W. Taylor, Cement Chemistry, 2nd ed. Thomas Telford, London, 1997. [34] V. Baroghel-Bouny, Water vapour sorption experiments on hardened cementitious materials: part 1: essential tool for analysis of hygral behavior and its relation to pore structure, Cem. Concr. Res. 37 (2007) 414–437. [35] M. Bianchi, F. Canonico, L. Capelli, M.L. Pace, A. Telesca, G.L. Valenti, Hydration properties of calcium sulfoaluminate–Portland cement blends, Aci International Conference on Recent Advances in Concrete Technology and Sustainability Issues, Seville, 2009. [36] F. Winnefeld, B. Lothenbach, Hydration of calcium sulfoaluminate cements — experimental findings and thermodynamic modelling, Cem. Concr. Res. 40 (2010) 1239–1247. [37] D. Gastaldi, F. Canonico, L. Capelli, M. Bianchi, M.L. Pace, A. Telesca, G.L. Valenti, Hydraulic behavior of calcium sulfoaluminate cement alone and in mixture with Portland cement, XIII International Congress on the Chemistry of Cement, Madrid, 2011. [38] P.K. Mehta, Mechanism of expansion associated with ettringite formation, Cem. Concr. Res. 3 (1973) 1–6. [39] D.H. Rasmussen, A.P. Mackenzie, Clustering in supercooled water, Chem. Phys. 59 (9) (1973) 5003–5013. [40] G. Le Saoût, B. Lothenbach, A. Hori, T. Higuchi, F. Winnefeld, Hydration of Portland cement with additions of calcium sulfoaluminates, Cem. Concr. Res. 43 (2013) 81–94.

Cement & Concrete Composites 61 (2015) 29–35

Contents lists available at ScienceDirect

Cement & Concrete Composites journal homepage: www.elsevier.com/locate/cemconcomp

An investigation on the recycling of hydrated cement from concrete demolition waste D. Gastaldi a,⇑, F. Canonico a, L. Capelli a, L. Buzzi a, E. Boccaleri b,⇑, S. Irico b a b

Buzzi Unicem S.p.A., Via Luigi Buzzi 6, 15033 Casale Monferrato, Alessandria, Italy Science and Technological Innovation Department, University of Piemonte Orientale, Viale T. Michel 11, 15121 Alessandria, Italy

a r t i c l e

i n f o

Article history: Received 30 May 2014 Received in revised form 1 April 2015 Accepted 11 April 2015 Available online 24 April 2015 Keywords: Hydrated cement waste Recycled aggregate Concrete demolition waste Fine fraction CO2 reduction

a b s t r a c t Construction and demolition waste (CDW) recycling is generally limited to the use of the coarser fraction as aggregate for new concrete. The recovery of fine aggregates requires a cleaning by removing the hydrated cement waste (HCW). In this paper, the possibility to use HCW extracted from CDW as alternative component for the production of new clinker is explored. A pure HCW sample was prepared and used in partial replacement of natural materials in raw admixtures for new clinker production. At a replacement degree of 30%, a new Portland clinker containing almost 50% of C3S could be produced with a huge spare in the release of CO2 (about 1/3 less). At higher HCW dosage a non-Portland clinker containing almost 80% of C2S has been obtained: its use as supplementary cementing material in blended cements revealed satisfying long term performances. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction Recycling is the issue for sustainable development. Main streams in the construction industry are constituted by concrete, brickworks, pieces of plaster and gypsum board characterized by extremely wide chemical composition, color, mechanical properties. In Europe about of 180 million tons of concrete demolition waste (CDW) are produced every year, corresponding annually to 500 kg for each citizen [1]: this amount represents around 31% of all the waste produced in the European Union [2]. For long time concrete and brickworks waste have only been used as a filling material or disposed to landfill. Nevertheless, in the late 20th century concrete recycling gained more and more importance, due to the increasing attention toward environmental protection and to the progressively reducing landfill capacity [3]. Current concrete recycling consists of crushing waste concrete and use it again as aggregate for new concrete [4], according to specifications which are based on local regulations in different countries. The production of recycled concrete aggregates (RCA) is a well established practice in Belgium, Denmark and the Nederlands, where recycling rates raise 80% [5], while it is less common in Southern Europe. ⇑ Corresponding authors. Tel.: +39 0161 809740 (D. Gastaldi), +39 0131 360264 (E. Boccaleri). E-mail addresses: [email protected] (D. Gastaldi), enrico.boccaleri@ uniupo.it (E. Boccaleri). http://dx.doi.org/10.1016/j.cemconcomp.2015.04.010 0958-9465/Ó 2015 Elsevier Ltd. All rights reserved.

The quality of RCA is generally lower than that of natural aggregates, due to presence of residual mortar [6]: for this reason, when dealing with concrete recycling, a differentiation between coarse (nominal size >5 mm) and fine aggregates (maximum size