Construction and Building Materials 170 (2018) 26–39

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Stabilisation of construction and demolition waste with a high fines

content using alkali activated fly ash
Nuno Cristelo a,⇑, Ana Fernández-Jiménez b, Castorina Vieira c, Tiago Miranda d, Ángel Palomo b
a
CQ-VR, Centro de Química – Vila real, Department of Engineering, University of Trás-os-Montes e Alto Douro, Apartado 1013, 5001-801 Vila Real, Portugal
b
Instituto Eduardo Torroja (IETcc), CSIC, C/Serrano Galvache 4, 28033 Madrid, Spain
c
CONSTRUCT, Faculty of Engineering of the University of Porto, Porto, Portugal
d
ISISE, School of Engineering of the University of Minho, Guimarães, Portugal

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Feasibility of an industrial waste Industrial Waste 1

(CDW) for large scale industrial
applications. Extraction of Substituted by Large scale
 Performance similar to natural, high- natural application
quality material (soil). aggregates of CDW
 Possible combination with another
industrial waste (fly ash) very Very performant
advantageous. material
 Type of material formed is a
combination of NAAASAH and Industrial
Waste 2
CAAASAH gels.

Fly Alkali
Ash Activated

a r t i c l e i n f o a b s t r a c t

Article history: This paper deals with the stabilisation of the fines portion of Construction and Demolition Waste (CDW)
Received 4 January 2018 produced by a Portuguese recycling company. The recycled aggregate was analysed with and without the
Received in revised form 3 March 2018 addition of a well-known precursor – fly ash (FA), considering different CDW/FA weight ratios. The result-
Accepted 6 March 2018
ing mixtures were activated with a solution based on sodium hydroxide and sodium silicate, and their
Available online 12 March 2018
microstructure and mechanical behaviour were extensively characterised, using XRD, FTIR, SEM, BSEM/
EDX and uniaxial compression strength (UCS) tests. Results obtained showed that the presence of fly
Keywords:
ash improves compression strength and elasticity module, reaching UCS values higher than 8 MPa, after
Alkaline activation
Construction and demolition waste
28 days of curing at ambient temperature and relative humidity. Such values are above any kind of
Recycled aggregates threshold considered for geotechnical works such as road and railway foundations or infrastructure
Sustainability embankments, which are regarded as effective applications for recycled CDW. The products resulting
Mechanical performance from the alkali activation of the CDW/FA mixtures were characterised as a mixture of the gels
NAAASAH and CAAASAH, which was attributed to the initial high content of amorphous Si, Al and Ca
of these mixtures.
Ó 2018 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.
E-mail addresses: ncristel@utad.pt (N. Cristelo), anafj@ietcc.es (A. Fernández-Jiménez), cvieira@fe.up.pt (C. Vieira), tmiranda@civil.uminho.pt (T. Miranda), palomo@ietcc.
es (Á. Palomo).

https://doi.org/10.1016/j.conbuildmat.2018.03.057
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 27

1. Introduction The inclusion of this recycled material in alkali activated mix-

tures constitutes a recent research approach. However, some
Over the last years, the environmental sustainability has been studies have been produced in the last few years, either related
demanding the recycling and reuse of wastes in the most varied with the use of the CDW as a precursor or related with its use
sectors of industry, including the construction industry. Until as a filler and/or courser aggregate in mortar and concrete.
recently, in most European countries, the construction sector was Regarding the use of CDW in a precursor role, Lampris et al.
very conservative and reluctant to apply wastes as a substitute, [22] showed that the fines (mostly silt) resulting from the wash-
or even as a partial replacement, of traditional materials. However, ing of the CDW have the capacity to be activated, producing max-
nowadays, the use of industrial wastes is an imperative need and a imum compressive strengths of 18 MPa and 22 MPa (when mixed
compulsory way for a more sustainable world. with 20% FA), after only 7 days curing at 25 °C. Such results were
According to the Environment Directorate-General of the Euro- possible due to the extensive work developed to optimise the
pean Commission, Construction and Demolition waste (CDW) is mixtures, in terms of solids/liquids ratio, soluble Si concentration,
one of the heaviest and most voluminous waste streams, repre- NaOH concentration and curing conditions. For comparison pur-
senting 25%–30% of all waste generated in the European Union poses and considering only the results that were obtained with
(EU). CDW are usually defined as the residues from construction, parameters similar to those used in the present paper, a clear sim-
reconstruction, maintenance and demolition of buildings and other ilarity can be observed, i.e. compression strength results between
civil infrastructures. These wastes are very heterogeneous and can 8 and 12 MPa.
contain any material that is part of a building or infrastructure, as Reig et al. [23] have also shown the precursor potential of a
well as any other materials used during the construction works ceramic CDW, reaching compressive strengths of up to 41 MPa
(e.g. concrete, bricks, gypsum, wood, glass, asphalt, excavated soil and 29 MPa, with mortars prepared from red clay bricks and porce-
and many others). lain stoneware, respectively. Such significant differences relatively
Due to the large volumes of CDW generated annually, as well as to the values obtained in the present research can be attributed to
the high potential for reuse and recycling of some of its compo- different curing conditions (7 days at 60 °C) and to the presence of
nents, CDW was identified by the European Commission as a prior- a granular phase formed by siliceous sand with a maximum parti-
ity stream [1]. On the one hand, the reuse of CDW reduces the cle size of 2 mm.
consumption of natural, non-renewable resources, while, on the Komnitsas et al. [24] studied the potential, for geopolymer syn-
other hand, reduces the landfilling of these inert materials coming thesis, of concrete, red clay bricks and ceramic tiles. The CDW was
from the construction industry. Despite these significant advan- submitted to previous pulverisation, and the alkaline activated
tages of recycling CDW, some member-states of the EU still have pastes were cured for 7 days at 60, 80 or 90 °C. The tile-based mix-
low recycling rates, far below the minimum threshold of 70% stip- tures reached a maximum of 58 MPa, while the brick-based and
ulated by the Waste Framework Directive of the European Parlia- concrete-based mixtures showed considerably lower strength val-
ment, to be achieved in 2020 [2]. This can be partially explained ues of 35 MPa and 13 MPa, respectively, thus concluding that the
by possible concerns regarding their environmental behaviour. CDW originated from concrete showed a lower precursor potential
Such concerns are justified and constitute an important section than the ceramic residues. However, it should be noted that such
of the national standards designed to establish the conditions nec- compressive strength values were only possible using curing tem-
essary for the application of recycled aggregates. Indeed, for any peratures of 80 °C or 90 °C, and after extensive pulverisation of the
specific application, a set of mandatory requirements are standard- initial wastes. Pathak et al. [25] confirmed this trend through
ised in European and national specifications (e.g. density, sulphates calorimetric analysis, obtaining higher reaction rates from brick
content, chlorides content, organic material content, among dust than concrete power.
others). Apart from the environmental requirements, established Robayo et al. [26] evaluated the mechanical and microstructural
in the European Council Decision of 19 December 2002 [3], which properties of red clay brick waste (RCBW) activated with sodium
regulate the minimum thresholds demanded for each class of hydroxide (SH) and sodium silicate (SS), either single or in combi-
landfill, there are several national standards regulating additional nation with up to 20% of Portland cement. The addition of SS to the
physical and mechanical properties required for each type of original SH activator represented a strength increase of 626% (from
application. For instance, the Portuguese standard (E 474) sets 7.49 MPa to 54.38 MPa) for the RCBW precursor, cured at 25 °C for
the minimum requirements for pavement construction applica- 28 days. The addition of 20% cement to the SH + SS activator
tions, regulating the waste in terms of constituents (severe limita- yielded a further strength increase, to 102.6 MPa. The addition of
tion of the floating and unidentified particles), nature and only 10% cement to the original SH activator produced a compres-
geometry (fines quality, content and maximum dimension), sion strength of 41.1 MPa, for the same curing conditions. The
mechanical performance (resistance to fragmentation and wear) authors concluded that a Ca source represents a major contribution
and chemical behaviour (soluble sulphates content and classifica- towards the performance of the final material.
tion as inert waste). Vásquez et al. [27], using an activator composed of sodium
Numerous studies have been carried out on the use of recycled hydroxide and sodium silicate, and curing for 28 days at room tem-
aggregates from CDW in base and subbase layers of roadways perature, obtained compressive strengths of about 25 MPa, with a
[4–11], in masonry construction products [12] and in concrete pro- 100% CDW geopolymer; 33 MPa, with a hybrid geopolymer (70%
duction [13–18] showing the high potential of these materials. CDW + 30% Portland Cement); and 46 MPa, with a binary geopoly-
Other studies tackled the use of CDW fine fractions as an alterna- mer (90% CDW + 10% metakaolin). The higher strength values of
tive pipe backfilling material [19] and also as a filling material the hybrid and binary systems, when compared with the results
for geosynthetic reinforced structures [20,21]. Albeit these positive obtained in the present study, are not surprising, given the addi-
results, the fine grain portion of the CDW, due to the high fine con- tion of cement and metakaolin. Regarding the value obtained for
tent and scattering of its constituents (soil, glass, concrete, mortars, the 100% CDW mixture, also higher than the values achieved in this
clay masonry units, among others), is frequently considered inap- study, it is probable that such difference is due to the quality and
propriate for the above-mentioned applications. The present work origin of each CDW (which, in the reported study, was 100% demol-
is thus oriented towards the development of an outlet for the fine ished concrete).
fraction of CDW, based on its stabilisation through alkali activation, Research targeting the use of CDW as an aggregate in a geopoly-
either alone or mixed with fly ash. meric matrix is even scarcer. Mohammadinia et al. [28] mixed
28 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39

Fig. 1. General view of: (a) CDW and (b) NGS, after breaking of the clods.

crushed bricks, crushed aggregates and asphalt pavement with a (identified as material mass ‘M4’, in the mentioned standard), only
precursor content of 4% (fly ash or slag), and the resulting mixture 15% of the total CDW mass was analysed. It should also be noted
was then activated with a solution made from sodium hydroxide that, in the present study, the CDW was used as received in the
and sodium silicate. Maximum strength values of approximately laboratory, i.e. with fractions M1 and M4 both included.
1.8 MPa, 4.5 MPa and 2.1 MPa were obtained for the crushed
bricks, crushed aggregate and asphalt pavement, respectively.
This paper aims to establish the viability of stabilising a CDW Table 2
batch, with a high fines content, resulting mostly from the demo- Chemical composition of the FA, CDW and NGS (% wt).

lition of single houses, using alkali activated binary systems. The Element FA CDW NGS
CDW was mixed with fly ash (FA), using CDW/FA weight ratios Na2O 1.18 0.31 0.85
between 100/00 and 60/40, which was then activated with a com- SiO2 45.2 64.25 65.0
bination of sodium hydroxide and sodium silicate. Mechanical Al2O3 28.5 9.87 21.9
behaviour was assessed through uniaxial compression strength MgO 1.07 0.79 0.21
K2O 1.56 2.91 8.67
tests, and the reaction products were characterized by XRD, FTIR,
CaO 6.04 17.06 0.11
SEM and BSEM. Also, a natural residual granitic soil (NGS) was used TiO2 2.48 0.41 0.17
to create threshold values from a high-quality natural aggregate. Fe2O3 10.7 2.25 2.49
ZnO 0.04 0.02 0.01
PbO 0.04 0.01 0.01
2. Materials and methods Cl – 0.03 –
CO2 – 0.52 –
2.1. Materials Cr2O3 0.05 0.04 0.06
MnO 0.08 0.03 0.07
P2O5 0.91 0.17 0.31
The CDW (Fig. 1a), regarded as an aggregate needing stabilisa- SO3 1.16 1.22 0.04
tion, was provided by a Portuguese recycling company. Based on ZrO2 0.09 0.02 0.01
the technical sheet, it is part of a specific set mainly originated BaO 0.38 – 0.03
from the demolition of single family houses and recovering of ille- CuO 0.05 – 0.02
Rb2O 0.01 – 0.04
gal deposits – and is thus considered a ‘‘Selected C&D waste” (EWC Others 0.46 0.09 –
code 1701 07), according to [3]. The constituents of the CDW pre- L.O.I. 6.47 – –
sented in Table 1 were determined based on the contents of the Si/Al 2.53 6.51 2.97
standard BSi EN 933-11 (2009) [29]. It should be noted that,
according with the procedure indicated, only the fraction collected
on the 4 mm sieve (identified as ‘M1’ material mass in the
mentioned standard) was submitted to a constituent analysis. This
means that, since 85 wt% of the material was under 4 mm

Table 1
Distribution of the constituents of the CDW (material mass M1).

Constituents wt% cm3/kg

Concrete, concrete products, mortar, concrete 25 –
masonry units, Rc
Unbound aggregate, natural stone, 32 –
hydraulically bound aggregate, Ru
Clay and calcium silicate masonry units, 12 –
aerated non-floating concrete, Rb
Bituminous materials, Ra 3 –
Glass, Rg 2 –
Other materials (including soils), X 26 –
Floating particles, FL – 28.5
Fig. 2. Particle size distribution of the FA, CDW and NGS.
 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 29

The total recycled CDW aggregate was mixed with different The chemical composition of these three materials, obtained by
percentages of fly ash, type F [30], which was provided by a Por- X-ray fluorescence (XRF), on a PHILIPS PW-1004 X-ray spectrome-
tuguese thermo-electric power plant. A natural residual granitic ter, is presented in Table 2, together with the mass loss on ignition
soil (NGS) was used as an alternative aggregate (in substitution (L.O.I.) up to 1000 °C. The main components of the FA are silica
of the CDW), to fabricate and test twin specimens. These dupli- (54.8%) and alumina (21.6%), with a calcium content of just 1.7%.
cated tests, on samples fabricated with a well-known, high- The CDW shows a significant calcium content (17.1%), while the
quality material, were developed to establish threshold values for NGS showed the higher silica content (65%) of the three materials,
the assessment of the quality of the CDW-based mixtures. This and only traces of calcium. The soluble sulphur content of the CDW
NGS was collected in the northwest region of Portugal (Fig. 1b). was determined to be 3.97%, following the procedure described in
BSi EN 1744-1 (2009) [31].
Fig. 2 presents the particle size distribution of the CDW and
Table 3
Main geotechnical properties of the FA, CDW and NGS.
NGS, determined using sieves and sedimentation (for particles
below 0.063 mm), while some geotechnical properties are pre-
Property FA CDW NGS
sented in Table 3. Particle size distribution of the fly ash was also
Specific gravity (kN/m3) 21.75 25.8 26.2 determined, although in this case by laser granulometry, on a Coul-
D50 (mm) 0.013 0.39 0.42 ter LS 130 analyser, able to measure particles from 0.05 to 900 lm,
Fines fraction (sieve N° 200) (%) 98.1 21.2 22.8
Maximum dry unit weight (kN/m3)* – 20.2 18.5
and using Fraunhofer diffraction theory.
Optimum water content (%)* – 8.33 12.3 From the mineralogical point of view, the XRD pattern of the FA
Minimum void ratio – 0.467 – showed the main presence of quartz and mullite (Fig. 3). The halo
Maximum void ratio – 0.819 – visible in the FA, approximately between the 17 °2h and 33 °2h
*
Modified Proctor test. angles, is an evidence of the presence of a vitreous phase, essential

Fig. 3. XRD patterns of the original FA, CDW and NGS (c-calcite: CaCO3 PDF 05-0586); g-gypsum: CaSO4
2H2O PDF 33-311; i-illite: (KH3O)(AlMgFe)2(SiAl)4O10[(OH)2(H2O)]; k-
kaolinite: Al2Si2O5(OH)4 PDF 14-164; mu-mullite: Al6Si2O13 PDF 15-776; m-muscovite: KAl3SiO10(OH)2 PDF 75-0948; o-orthoclase: KAlSi3O8 PDF 31-0966; q-quartz: SiO2 PDF79-
1910).

Table 4
Identification and general characterisation of the mixtures fabricated.

Mixture type Aggregate/ash ratio Dry unit weight*(kN/m3) Water content* (wt%) Activator content** (wt%)
100% aggregate 100/00 20.2 8.33 11.28
90% aggregate + 10% FA 90/10 20.0 8.90 12.04
80% aggregate + 20% FA 80/20 19.3 9.36 12.48
70% aggregate + 30% FA 70/30 18.5 10.58 14.72
60% aggregate + 40% FA 60/40 17.4 11.52 17.04
*
Obtained from Modified Proctor tests.
**
Activator/solids weight ratio.
30 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39

in alkali activation reactions. The diffractogram of the CDW the presence of concrete residues, as shown in Table 1. The CDW
revealed several mineral phases usually found in soils – quartz, pattern also showed a halo in a similar range, although less pro-
illite, muscovite, orthoclase – which is not surprising given the nounced. The XRD of the NGS revealed the presence of quartz,
magnitude of the CDW fraction below 4 mm (85%). Gypsum and kaolinite, muscovite and orthoclase, all very common in this type
calcite were also detected in the CDW, which is associated with of soil.

Fig. 4. General view of the moulded specimens (a) and broken specimens (b).

Fig. 5. Unconfined compressive strength (UCS) of the aggregate-ash mixtures.

Fig. 6. Young’s Module obtained at 50% of the peak stress mixtures.

 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 31

The activator solution used was a mixture of sodium hydrox- 15 m SH concentrations. The sodium silicate had a unit weight
ide (SH), with a concentration of 10 mol/kg (10 m), and sodium of 1.464 g/cm3 at 20 °C, a SiO2/Na2O weight ratio of 2.0 (molar
silicate (SS), with a SH/SS weight ratio of 2. The SH concentra- oxide ratio of 2.063) and a Na2O concentration in the solution
tion was defined based on the results of preliminary tests, of 13.0%. Deionised water was used throughout the entire exper-
performed on CDW specimens activated with 5 m, 10 m and imental work.

Fig. 7. XRD patterns of the (a) CDW-FA mixtures and (b) NGS-FA mixtures, cured for 28 days (c-calcite: CaCO3 PDF 05-0586); g-gypsum: CaSO4
2H2O PDF 33-311; i-illite: KH3O.
AlMgFe.2(SiAl)4.O10[(OH)2.(H2O)]; k-kaolinite: Al2Si2O5(OH)4 PDF 14-164; mu-mullite: Al6Si2O13 PDF 15-776; m-muscovite: KAl3SiO10(OH)2 PDF 75-0948; o-orthoclase: KAlSi3O8
PDF 31-0966; q-quartz: SiO2 PDF79-1910).
32 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39

2.2. Preparation of the binders and mechanical testing the uniaxial compression strength tests. These tests were carried
out under monotonic displacement control, at a rate of 0.2 mm/
The composition of the mixtures was based on the variation of min, and both the peak load and Young’s module were obtained
the ratio between the aggregate (natural or recycled) and the fly from each test result. A pair of linear variable differential trans-
ash (Table 4). A Modified Proctor test was performed for each formers was installed between the two compression plates, for
CDW/FA ratio, to define the dry unit weight and water content increased strain measurement accuracy. Results (see Fig. 4b)
(results also presented in Table 4). It is clear that an increase in showed in this paper represent the average of three tested
fly ash content represented an increase in the optimum water con- specimens.
tent and a decrease in the maximum dry density. Due to the sim- The precursors and alkali-activated pastes were characterized
ilarity between the CDW and the NGS particle size distribution by Scanning Electron Microscopy (SEM), Back-scattered Electron
curves, the dry unit weight and water content values defined for Microscopy (BSEM), X-ray Diffraction (XRD) and Fourier Transform
the CDW/FA mixtures were also considered for the NGS/FA mix- Infrared Spectroscopy (FTIR). The SEM and BSEM analyses were
tures. Note that the total amount of water present in the solution performed on a JEOL JSM 5400 scanning electron microscope (20
was always equal to the optimum water content defined by the kV), in low vacuum mode (1.3 mbar), avoiding the deposition of
corresponding Proctor test. a conductive layer. The device was fitted with a solid-state BSE
Preparation of the CDW and NGS consisted on drying and de- detector and a X-ray energy dispersive analyser (EDX) – Oxford
flocculation by hand (not mechanically ground). The fly ash was LINK-ISIS – using a ZAF correction model for quantitative chemical
received with water content lower than 3%, and thus was consid- analysis. The XRD diffractograms were obtained on a PANalytical
ered ready for use. The solid components (CDW/FA or NGS/FA) X’Pert Pro diffractometer, using CuKa radiation. The FTIR spectra
were dry mixed for 5 min in a Hobart counter mixer. After this ini- were acquired using a Thermo Scientific Nicolet iS50 FTIR spec-
tial homogenization period, the activator, which was prepared 4 to trometer, resorting to an ATR (Attenuated Total Reflectance) acces-
6 h before, was slowly added, and an additional minimum period sory with a diamond crystal. The instrument was controlled by the
of 5 min followed, in order to guarantee homogenisation. The acti- software package Omnic, version 9.2.28, from Thermo Fisher Scien-
vator temperature at the time of the mixing was 20 °C ± 1 °C. tific Inc. A spectrum range of 1600 to 400 cm 1 was defined, with
Immediately after the mixing process, the homogenized paste a resolution of 4 cm 1 and 128 accumulations per spectrum. The
was compacted inside a cylindrical stainless-steel mold, with nom- spectra were corrected for background interferences, using a base-
inal height and diameter of 140 mm and 70 mm, respectively. The line to ‘‘normalize” the data before comparison.
weight of the compacted mixture was defined so that the desired
unit weight was obtained. Moulding of the specimens for uniaxial
compression strength tests was substantially based on BSi EN 3. Results
13286-53 (2004) [32]. The general aspect of the moulded speci-
mens can be seen in Fig. 4a. 3.1. Compression strength
After compaction, the specimens were wrapped in cling film
and left to cure in a controlled environment (20 °C and 80% RH) Fig. 5 presents the uniaxial compression strength (UCS) results
for 7, 14 or 28 days. Each specimen was weighted and measured obtained by all the mixtures considered in this study, after 7, 14
shortly before being tested. and 28 days curing. Results obtained with the natural aggregate
A servo-hydraulic testing machine, fitted either with a 25 kN or (NGS) are also included. The CDW/FA samples showed higher
50 kN capacity load cell, with a resolution of 0.006 kN, was used for strength values than the NGS/FA, at 7 and 14 days. However, after

Fig. 8. IR spectra of the original materials (FA, CDW, NGS) and their respective combinations – CDW-based mixtures (a) and NGS-based mixtures (b).
 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 33

28 days curing, the NGS/FA specimens were clearly more perform- activator. These changes, more evident in the NGS100/00 mixture,
ing than the CDW/FA specimens. Therefore, the curing time did not indicate that some Si and Al from the kaolinite phase present in the
increase the strength of the CDW samples at the same rate as the NGS did participate in the reactions.
NGS. The fly ash has also shown the bands associated with quartz and
A similar trend was observed for the elasticity module (Fig. 6), mullite (band between 560 1 and 550 cm 1 is associated with the
i.e. a higher value was obtained for the CDW specimens after 7 octahedral aluminium in mullite; while the band associated with
and 14 days curing, after which a substantial increase of the NGS tetrahedral aluminium appears between 830 and 920 cm 1). How-
strength, between 14 and 28 days, allowed this mixture to show ever, the most important band in the fly ash appears around the
higher UCS values than the CDW mixtures. 1080 cm 1 wavelength, associated to the vibrations of TAO bonds
(T@Al or Si). This band is an overlap of the bands resulting from the
3.2. Mineralogical and microstructural characterization vibration of quartz, mullite and the vitreous phase of the fly ash. As
a result of the alkaline activation reactions in CDW60/40 and
For the characterization of the reaction products, mixtures
CDW100/00 and NGS100/00, and their respective M60/40 binary
systems (CDW60/40 and NGS60/40), were selected and analysed,
after curing for 28 days. Fig. 7 shows the XRD diffractograms of
the original materials (CDW, NGS and FA) together with those of
the mixtures 100/00 and 60/40. The materials resulting from the
activation reactions (100/00 and 60/40) didn’t present any new
peaks that could be attributed to the formation of new crystalline
phases. Basically, the same crystalline phases present in the pre-
cursors were also detected in the hydrated mixtures. However,
the intensity of the orthoclase, muscovite, gypsum and calcite
phases was a little lower in the CSW100/00 mixture, thus indicat-
ing small reactions of these phases. A similar trend was observed in
the CSW60/40 mixture, although with a less significant intensity
reduction, which can be attributed to the dilution factor (from
the introduction of FA).
As mentioned above, the diffractogram of the FA showed a
hump in the 17°–33° (2h) angle range, typical of structurally disor-
dered compounds, with a high amorphisation level. The diffrac-
tograms of the activated mixtures showed smaller versions of
this hump, with the mixtures with more fly ash (60/40) showing
a higher hump, after activation, than the mixtures without fly
ash (100/00). The alkaline activation reaction was also responsible
for the slight translation of these humps to the right, towards 2h
angle positions approximately between 25° and 32°. This transla-
tion represents the formation of the amorphous sodium alumi-
nosilicate gel (which is considered to be a zeolitic precursor, yet
no zeolites could be detected at this stage of the curing process).
Fig. 8 shows the FTIR spectra for the original materials and for
samples 100/00 and 60/40. The assignments of the bands proved
to be a complicated task, due to several cases of overlapping.
In the CDW-based samples (Fig. 8a), the bands centred at 1420,
873 and 711 cm 1 are originated by the presence of calcite in the
original CDW, and thus its intensity decreases with the decrease
in CDW content. Bands at 1030, 531 and 467 cm 1 are associated
with the vibration of TAO (T@Si or Al) bonds, originated by the
presence of orthoclase. Some of these bands overlap with the
vibration of the SiAO and AlAO bonds (bands 1023, 531 and 477
cm 1) present in the muscovite aluminosilicate phase, also
detected by XRD analysis. Bands at 1166 and 1084, the double
band at 798–780, and the bands at 697, 512 and 462 cm 1, are
all assigned to quartz, the crystalline phase in the original materi-
als (FA, CDW and NGS). This signal didn’t show significant changes,
suggesting that, as expected, this crystalline phase wasn’t affected
by the alkaline reactions, and thus its respective Si and Al ions
didn’t contribute to the formation of the gel. Other examples are
the main band centred at 1008 cm 1, corresponding to the stretch-
ing vibrations of the SiAOASi and SiAOAAl bonds of illite (from the
significant soil portion of the CDW), or the small peak at 915 cm 1,
typical of the vibrations of the Al(IV)AOH bonds (usually found in
several clay minerals). These bands also appear in the NGS
(Fig. 8b) – although in the NGS the most significant band corre-
sponds to the SiAO stretching, at 1010 cm 1 (belonging to the Fig. 9. SEM images of mixtures CDW100/00 (a), CDW60/40 (b) and NGS60/40 (c),
kaolinite), and shift to lower frequencies with the addition of cured for 28 days under ambient conditions.
34 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39

NGS60/40, these TAO asymmetric stretch bands shift towards product (points A in Fig. 9b). Several unreacted ash particles, of dif-
lower wavelengths, i.e. 990 cm 1 (CDW60/40) and 980 cm 1 ferent sizes, were also detected, as well as partially reacted parti-
(NGS60/40), thus indicating the formation of the alkaline alumi- cles, sometimes covered with small portions of reaction product
nosilicate hydrate gel type NAAASAH [33,34]. This behaviour (points B in Fig. 9b). In general, the presence of a large number
was also observed in both 100/00 mixtures. Such trend can be a of unreacted particles is a consequence of a moderate reaction
consequence of the formation of gels with different compositions, degree. Finally, the NGS60/40 sample (Fig. 9c) showed a higher
which would explain the different strength values obtained by the particle dissolution and bulkier phase than the CDW-based mix-
CDW-based and NGS-based pastes, after 28 days curing. tures. Thus, the aggregate particles are clearly more difficult to
The peak centred at the 465 cm 1 frequency, detected in the detect in the NGS-ash mixtures than in the CDW-ash mixtures.
original CDW, shifted to the 458 cm 1 frequency in the Considering the more compacted appearance of the mixtures with
CDW100/00 and CDW60/40 mixtures (although more intense in fly ash, compared with mixture CDW100/00, and based also on the
the former). This band is assigned to SiAO and AlAO bending vibra- more evenly and continuous distribution of the respective gel, it
tions, and provides an indication of the ‘‘amorphisation” stage of appears that a higher reaction degree was reached when fly ash
the material since, according with [34], the intensity of this peak was added to the CDW, implying that the fly ash has a higher vit-
doesn’t depend of the crystallisation degree. reous content than the CDW. However, even in the mixtures con-
In short, most bands detected in the CDW-based mixtures were taining ash (either CDW-based or NGS-based), several unreacted
also detected in the NGS-based mixtures, except those bands ash particles, sometimes partially covered with reaction products,
assigned to the presence of carbonates, which is not surprising were easily detected (points B in Fig. 9b points C in Fig. 9c).
given the fact that the NGS is a siliceous aggregate. BSEM/EDX was used to perform a quantitative analysis of the
Mixtures CDW100/00, CDW60/40 and NGS60/40, cured at mixtures and their respective reaction products. In the sample
ambient temperature for 28 days, were also analysed by micro- CDW100/00 (Fig. 10), the resulting product that surrounds and
scopy. Fig. 9 presents SEM images of selected samples, enabling binds the silica particles (from the soil fraction of the recycled
an understanding of the mixtures’ overall morphology. Sample aggregate), is a mixture of CAAASAH gel (points 1 and 4), and
CDW100/00 (Fig. 9a) showed less compacted and very porous NAAASAH gel (points 2 and 3). The aluminium is evenly dis-
CDW pastes, which can help explain the lower mechanical strength tributed throughout the gel volume, as well as the silicon
of this mixture, when compared with mixture NGS100/00, after the (although, naturally, the latter presented a higher concentration
same curing period. Sample CDW60/40 showed a significant in the unreacted SiO2 particles). However, both the sodium and
number of aggregate particles covered with amorphous reaction the calcium elements are circumscribed to specific areas.

Fig. 10. BSEM/EDX analysis of the CDW100/00 sample cured for 28 days under ambient conditions.
 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 35

Sample CDW60/40 (Fig. 11) presented a gel distribution similar inferior to natural aggregates. However, such drawback is compen-
to that observed with the CDW100/00, i.e. gels CAAASAH (points 1 sated by its higher availability (it is produced in significant vol-
and 3) and NAAASAH (point 4) were formed around unreacted sil- umes worldwide), which inclusively generates the constant need
ica particles, but on perfectly defined areas, with a clear interface to find alternatives to landfilling. Nevertheless, it does poses some
between the two types of gel. There was a prevalence of Ca on concerns regarding its environmental behaviour, highlighted by
one area, and of aluminosilicates on the other. However, in this the fact that, in some cases, and depending on the type of CDW,
case, and contrary to what was observed in the CDW100/00 sam- it needs to be deposited in non-hazardous or even hazardous land-
ple, there was a homogeneous distribution of the sodium through- fills. The aim of this study was to assess the feasibility of stabilising
out the NAAASAH and CAAASAH areas, while the calcium was a CDW batch with a high fines content, obtained from several Por-
mostly circumscribed to the CAAASAH area, although the tuguese operators, with alkali activated fly ash. As well as with
NAAASAH gel detected in point 4 has taken up a certain amount most natural aggregates, stabilisation is a necessary step in
of calcium [35–37]. geotechnical applications, to ensure its structural adequacy. In this
The analysis of sample NGS60/40 (Fig. 12) supported the con- case, and considering the origin of both the CDW and the FA, the
clusion that the type of precursor had a significant influence on stabilising agent needs to address not only the structural, but also
the chemical composition of the gels formed around the silica par- the environmental performance.
ticles (point 1). The Ca content in the gel resulting from the activa- Based on the strength results, and even if the 28th day overall
tion of the NGS-ash mixture is significantly lower than in the CDW- performance was inferior to that obtained with the natural aggre-
based mixtures (Figs. 10 and 11). Therefore, a straight-forward gate, the response of the CDW was acceptable, showing UCS values
NAAASAH type gel was formed, with practically no calcium pre- above 4 MPa). Although the CDW strength values, for every fly ash
sent and significant Al contents (points 2–4). content, were higher than the soil strength values at 7 and 14 days
curing, the measurement of the strength performance after 28 days
should be used as a reference, since it yields a more accurate idea
4. Discussion of the long-term relative strength of both types of mixture. Fur-
thermore, a longer curing period would probably increase the
Construction and demolition waste has a high potential for strength difference between the high Ca system (CAAASAH gel),
reutilisation as a filling structural material, in applications like road which has a tendency to stagnate after a few weeks; and the low
and railway base and sub-base layers or embankments. Its perfor- Ca system (NAAASAH gel), which has been found to increase over
mance, from the mechanical point of view, is know to be slightly periods of up to 1 year [38–40].

Fig. 11. BSEM/EDX analysis of the CDW60/40 sample cured for 28 days under ambient conditions.
36 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39

Fig. 12. BSEM/EDX analysis of the NGS60/40 sample cured for 28 days under ambient conditions.

The assessment of the CDW adequacy as an aggregate for alkali by other authors (see Section 1). Such gap is explained by the dif-
activated binders becomes less straight-forward after concluding ferences in processing, especially those regarding the curing condi-
that the CDW itself reactes with the activator. However, a rough tions (time, humidity and temperature), type and concentration of
estimation can be made considering that the strength ratio the activator, and even specimen preparation. In the present work,
between the CDW (0.17 MPa) and NGS (0.31 MPa), when moulded the preparation of the specimens followed classic procedures
simply with water, is 0.55, which is in line with identical strength usually applied to soils and granular materials (e,g, Proctor com-
ratios obtained for the activated mixtures, after 28 days: 0.67, 0.52, paction test, mixtures cured at ambient temperature); while, in
0.65, 0.69 and 0.94 (for an ash contents of 0%, 10%, 20%, 30% and
40%, respectively). This suggests that the CDW isn’t responsible
for any performance decrease during or after the alkali activation
reactions, and thus can be regarded as an adequate aggregate to
be used in such applications.
The increase, with the addition of FA, of the CDW compressive
strength and rigidity was clearly achived (Fig. 13). The progressive
substitution of the aggregate (CDW or NGS) by fly ash increases the
similarity, in terms of overall chemical composition of the precur-
sor, between the CDW/FA and NGS/FA mixtures, while also increas-
ing the volume of the reaction products. Therefore, the
compressive strength increases with the progressive addition of
fly ash, either in the CDW or NGS-based mixtures. This is a conse-
quence of the higher vitreous content of the FA, when compared
with the aggregates. Due to its chemical constitution, this particu-
lar fly ash produces a NAAASAH gel type as the main reaction
product, which was clearly detected during the BSEM/EDX analy-
ses of samples CDW100/40 and NGS60/40.
At this point, it is important to highlight that, although the
strength development, at 28 days curing, reached meaningful val- Fig. 13. Evolution of the strength and elasticity ratios between the CDW and the
ues (6–8 Mpa), these are still slightly lower than those obtained NGS mixtures.
 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 37

the studies mentioned in the literature, the method of kneading- lower strength values, after 7 and 14 days, and higher values, after
cement was used, together with curing temperatures above 60 °C. 28 days, than the CDW mixtures. The FTIR band shifting from 1010
In any case, the values obtained are enough to comply with Por- cm 1 to 980 cm 1 confirms such reaction.
tuguese and European standards regarding geotechnical applica- Fig. 14a) shows the composition distribution of the gels
tions. Portuguese standard LNEC E 474 [41] states that, in order obtained from the CDW100/00 and CDW60/40 pastes. Although
to be accepted as a structural landfill or pavement layer material, the Ca content is significantly lower in the latter, it still presented
recycled aggregates need to comply with the threshold levels 36.4% of BSEM points with a Ca content above 0.60, while the
required to obtain ‘inert landfill’ status, established in [3]. This CDW60/40 presented only a marginally superior value (38.5%). It
was achived within the framework of this study. appears that the presence of the fly ash, with its high vitreous con-
The microstructural characterization shows clear differences in tent, enhanced the reaction degree achieved by the CDW60/40
the chemical composition of each of the two main types of reaction paste, allowing the dissolution of a higher quantity of calcium ions,
products, which are a consequence of the type of aggregate, since even if the original Ca content of the CDW60/40 mixture is lower
Ca was almost absent in the NGS-based samples, while it is very than the original Ca content of the CDW100/00 mixture. This
significant in the CDW-based samples. The presence of Ca in the higher Ca content detected in the CDW60/40 paste was responsible
CDW samples accelerates the initial stages of the reactions, for the CAAASAH gel, which was formed in higher quantities than
enabling the formation of a CAAASAH type gel, possibly due to the NAAASAH gel, judging by the lower Na content showed in
the reaction between the Ca (from the CDW) and the soluble silica Fig. 14b). Also interesting is the fact that the CDW60/40 presented
(from the activator). Such mechanism explains the higher strength a higher Al content than the CDW100/00, which points towards the
presented by all the CDW mixtures at 7 and 14 days curing. How- development, in the latter, of a CA(A)ASAH gel type, instead of a
ever, when this primary gel is formed, the subsequent evolution of fully matured CAAASAH gel.
the reactions is significantly hindered, resulting in strength values, Fig. 15a) shows a tendency of the CDW60/40 paste to present
at 28 days, very similar to those obtained after 14 days. FTIR spec- higher Ca contents than the NGS60/40, which was expected, based
tra showed only slight variations between the original materials on the higher original Ca content of the CDW aggregate. However, a
and the reaction products (Fig. 8), while the latter presented a significant number of CDW points show a Ca content below 20%.
poorly compacted matrix during the microscopy analysis (Fig. 9). These two sets – high Ca content (lower left area) and low Ca con-
In both these analyses, CAAASAH was the predominant gel type. tent (upper right area), were previously detected and associated
On the other hand, the NGS aggregate, used as a reference in this with CAAASAH and NAAASAH gels, respectively [35], which is
study, possesses only a residual Ca content. Nevertheless, it also corroborated by the BSEM analysis of the CDW60/40 paste pre-
reacts with the activator, although more slowly, which results in sented in Fig. 11. At the same time, the 5.29% average Na content

Fig. 14. Chemical composition of the gels (qualitative data) formed with each CDW/FA ratio.

Fig. 15. Chemical composition (EDX) of the gels (qualitative data) formed in the 60/40 mixtures.
38 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39

found in the NGS-based gel (Fig. 15b) was higher than the average [7] I. Vegas, J.A. Ibañez, J.T. San José, A. Urzelai, Construction demolition, waelz
slag and MSWI bottom ash: a comparative technical analysis as material for
value found in the CDW-based gel (3.78%). It is also possible to
road construction, Waste Manag. 28 (2008) 565–574.
observe that the NGS-based paste presented higher Si and Al per- [8] C.S. Vieira, P.M. Pereira, Damage induced by recycled Construction and
centages than the CDW-based paste, even if both species are Demolition Wastes on the short-term tensile behaviour of two
included in the gel mainly via fly ash, which is present in the same geosynthetics, Transp. Geotech. 4 (2015) 64–75.
[9] D.X. Xuan, A.A.A. Molenaar, L.J.M. Houben, Evaluation of cement treatment of
proportion in both mixtures. This higher Na and Al and lower Ca reclaimed construction and demolition waste as road bases, J. Clean. Prod. 100
and Si contents in the NGS60/40 paste explains the absence of a (2015) 77–83.
CAAASAH gel, as shown in Fig. 12. Such differences, in the content [10] A. Mohammadinia, A. Arulrajah, J. Sanjayan, M. Disfani, W.B. Myint, S.
Darmawan, Laboratory evaluation of the use of cement-treated construction
of the most important ions (Si, Al, Na and Ca), should certainly play and demolition materials in pavement base and subbase applications, J. Mater.
a significant role in the development of the significant compressive Civ. Eng. 27 (2015) 1–12.
strength differences between the two types of aggregate (Fig. 5). [11] A. Arulrajah, E. Yaghoubi, Y.C. Wong, S. Horpibulsuk, Recycled plastic granules
and demolition wastes as construction materials: resilient moduli and
However, the precise level of influence is difficult to define, not strength characteristics, Constr. Build. Mater. 147 (2017) 639–647.
only because the exact strength of the gel can only be fully deter- [12] R.A. Robayo-Salazar, J.F. Rivera, R. Mejía de Gutiérrez, Alkali-activated building
mined with micro indentation tests, but also because, as stated materials made with recycled construction and demolition wastes, Constr.
Build. Mater. 149 (2017) 130–138.
above, the morphology and the depth of the reactions have also a [13] M. Bravo, J. de Brito, J. Pontes, L. Evangelista, Mechanical performance of
significant influence on mechanical performance. concrete made with aggregates from construction and demolition waste
recycling plants, J. Clean. Prod. 99 (2015) 59–74.
[14] C. Medina, W. Zhu, T. Howind, M.I. Sánchez de Rojas, M. Frías, Influence of
5. Conclusions mixed recycled aggregate on the physical – mechanical properties of recycled
concrete, J. Clean. Prod. 68 (2014) 216–225.
[15] H. Mefteh, O. Kebaïli, H. Oucief, L. Berredjem, N. Arabi, Influence of moisture
The aim of this research was to assess the efficiency of a con- conditioning of recycled aggregates on the properties of fresh and hardened
struction and demolition waste with a high fines content, resulting concrete, J. Clean. Prod. 54 (2013) 282–288.
[16] R.V. Silva, J. de Brito, R.K. Dhir, Establishing a relationship between modulus of
from a Portuguese licenced operator, as an aggregate in alkali acti- elasticity and compressive strength of recycled aggregate concrete, J. Clean.
vated composites. Different CDW/FA ratios were prepared and Prod. 112 (2016) 2171–2186.
studied, and every test was duplicated with a residual granite soil, [17] A. Rao, K.N. Jha, S. Misra, Use of aggregates from recycled construction and
demolition waste in concrete, Resour. Conserv. Recycl. 50 (2007) 71–81.
replacing the CDW. The efficiency of the composite materials was [18] R.V. Silva, J. de Brito, R.K. Dhir, Properties and composition of recycled
determined by mechanical behaviour, namely using uniaxial com- aggregates from construction and demolition waste suitable for concrete
pressive strength. The elemental constitution of the CDW allowed production, Constr. Build. Mater. 65 (2014) 201–217.
[19] M.A. Rahman, M. Imteaz, A. Arulrajah, M.M. Disfani, Suitability of recycled
the formation of a CAAASAH gel in the early stages of the alkaline construction and demolition aggregates as alternative pipe backfilling
activation reactions, while the progressive addition of FA promoted materials, J. Clean. Prod. 66 (2014) 75–84.
the co-existence of this gel with a NAAASAH type gel, especially [20] E.C.G. Santos, E.M. Palmeira, R.J. Bathurst, Performance of two geosynthetic
reinforced walls with recycled construction waste backfill and constructed on
after 4 weeks curing. The NGS/FA mixtures showed only
collapsible ground, Geosynth. Int. 21 (2014) 256–269.
NAAASAH gel, responsible for the higher UCS of the NGS mixtures [21] C.S. Vieira, P.M. Pereira, M. de L. Lopes, Recycled Construction and Demolition
after 28 days, while the CDW mixtures were more effective after 7 Wastes as filling material for geosynthetic reinforced structures. Interface
and even 14 days. This particular CDW batch proved to constitute properties, J. Clean. Prod. 124 (2016) 299–311.
[22] C. Lampris, R. Lupo, C.R. Cheeseman, Geopolymerisation of silt generated from
an effective substitute for natural aggregates, acting not only as an construction and demolition waste washing plants, Waste Manag. 29 (2009)
aggregate, but also contributing to the total volume of amorphous 368–373.
material available for the alkaline reactions. Due to the significant [23] L. Reig, M.M. Tashima, L. Soriano, M.V. Borrachero, J. Monzó, J. Payá, Alkaline
activation of ceramic waste materials, Waste Biomass Valorization 4 (2013)
Ca content in the CDW, two types of aluminosilicate gel were 729–736.
detected, namely NAAASAH and CAAASAH, contrary to what [24] K. Komnitsas, D. Zaharaki, A. Vlachou, G. Bartzas, M. Galetakis, Effect of
was observed in the soil-based mixtures, in which case only the synthesis parameters on the quality of construction and demolition wastes
(CDW) geopolymers, Adv. Powder Technol. 26 (2015) 368–376.
former was detected. [25] A. Pathak, S. Kumar, V.K. Jha, Development of building material from
geopolymerization of construction and demolition waste (CDW), Trans.
Indian Ceram. Soc. 73 (2014) 133–137.
Acknowledgments [26] R.A. Robayo, A. Mulford, J. Munera, R. Mejía de Gutiérrez, Alternative cements
based on alkali-activated red clay brick waste, Constr. Build. Mater. 128 (2016)
163–169.
The authors would like to acknowledge the contribution of the [27] A. Vásquez, V. Cárdenas, R.A. Robayo, R.M. de Gutiérrez, Geopolymer based on
company RCD SA, for the supply of the construction and demolition concrete demolition waste, Adv. Powder Technol. 27 (2016) 1173–1179.
waste; and the company PEGOP – Energia Elétrica SA, for the sup- [28] A. Mohammadinia, A. Arulrajah, J. Sanjayan, M.M. Disfani, M. Win Bo, S.
Darmawan, Stabilization of demolition materials for pavement base/subbase
ply of the fly ash.
applications using fly ash and slag geopolymers: laboratory investigation, J.
Mater. Civ. Eng. 28 (2016) 4016033.
[29] BSi EN 933-11, Tests for geometrical properties of aggregates. Classification
References test for the constituents of coarse recycled aggregate, Br. Stand. Institution,
London. (2009).
[1] C.S. Vieira, P.M. Pereira, Use of recycled construction and demolition materials [30] ASTM C618, Standard specification for coal fly ash and raw or calcined natural
in geotechnical applications: a review, Resour. Conserv. Recycl. 103 (2015) pozzolan for use in concrete, ASTM Int. Annu. B. Stand. (2012) 1–5.
192–204. [31] BSi EN 1744-1, Tests for chemical properties of aggregates, Part 1: Chemical
[2] EU, Directive 2008/98/EC of the European Parliament and of the Council of 19 analysis, Br. Stand. Institution, London. (2009).
November 2008 on waste, Official Journal of the European Union, 2008. [32] BSi EN 13286-53, Unbound and hydraulically bound mixtures. Method for the
[3] EU, Council Decision 2003/33/EC of 19 December 2002, Official Journal of the manufacture of test specimens of hydraulically bound mixtures using axial
European Union, 2003. compression, Br. Stand. Institution, London. (2004).
[4] A. Arulrajah, J. Piratheepan, M.M. Disfani, M.W. Bo, Geotechnical and [33] I. García-Lodeiro, N. Cherfa, F. Zibouche, A. Fernández-Jiménez, A. Palomo, The
geoenvironmental properties of recycled construction and demolition role of aluminium in alkali-activated bentonites, Mater. Struct. 48 (2015) 585–
materials in pavement subbase applications, J. Mater. Civ. Eng. 25 (2013) 597.
1077–1088. [34] A. Fernández-Jiménez, A. Palomo, Mid-infrared spectroscopic studies of alkali-
[5] A. Barbudo, F. Agrela, J. Ayuso, J.R. Jiménez, C.S. Poon, Statistical analysis of activated fly ash structure, Microporous Mesoporous Mater. 86 (2005) 207–
recycled aggregates derived from different sources for sub-base applications, 214.
Constr. Build. Mater. 28 (2012) 129–138. [35] I. García-Lodeiro, A. Fernández-Jiménez, A. Palomo, Variation in hybrid
[6] C.S. Poon, D. Chan, Feasible use of recycled concrete aggregates and crushed cements over time. Alkaline activation of fly ash–portland cement blends,
clay brick as unbound road sub-base, Constr. Build. Mater. 20 (2006) 578–585. Cem. Concr. Res. 52 (2013) 112–122.
 N. Cristelo et al. / Construction and Building Materials 170 (2018) 26–39 39

[36] I. García-Lodeiro, A. Fernández-Jiménez, A. Palomo, D.E. MacPhee, Effect of [39] S. Rios, N. Cristelo, T. Miranda, N. Araújo, J. Oliveira, E. Lucas, Increasing the
calcium additions on N-A-S-H cementitious gels, J. Am. Ceram. Soc. 93 (2010) reaction kinetics of alkali-activated fly ash binders for stabilisation of a silty
1934–1940. sand pavement sub-base, Road Mater. Pavement Des. 19 (2018) 201–222.
[37] I. García-Lodeiro, A. Palomo, A. Fernández-Jiménez, D.E. MacPhee, [40] N. Cristelo, S. Glendinning, L. Fernandes, A.T. Pinto, Effects of alkaline-activated
Compatibility studies between N-A-S-H and C-A-S-H gels. Study in the fly ash and Portland cement on soft soil stabilisation, Acta Geotech. 8 (2013)
ternary diagram Na2O-CaO-Al2O3-SiO 2–H2O, Cem. Concr. Res. 41 (2011) 395–405.
923–931. [41] LNEC E 474, Guide for the Use of Recycled Materials Coming from Construction
[38] N. Cristelo, S. Glendinning, A.T. Pinto, Deep soft soil improvement by alkaline and Demolition Waste in Embankment and Capping Layer of Transport
activation, Proc. Inst. Civ. Eng. Gr. Improv. 164 (2011) 73–82. Infrastructures, (2009).