Sunteți pe pagina 1din 8

Bond Strength of Normal-to-Lightweight Concrete Interfaces

Hugo COSTA
Research Assistant
Polytechnic Institute of Coimbra
Coimbra, Portugal

Pedro SANTOS
Adjunct Professor
Polytechnic Institute of Leiria
Leiria, Portugal

Eduardo JLIO
Full Professor
ICIST & DECivil, IST-UTL
Lisbon, Portugal

hcosta@isec.pt

pedro.santos@ipleiria.pt

ejulio@civil.ist.utl.pt

Hugo Costa, born 1977, civil


engineering degree (2002), MSc degree
(2008) at Univ. of Coimbra. PhD
student at Univ. of Coimbra, since
2008. His research area is related to
mixture design and structural behaviour
of lightweight aggregate concrete.

Pedro Santos, born 1977, civil


engineering degree (2001), MSc
degree (2005), PhD degree (2009) at
University of Coimbra. His research
interests include pre-casting and
strengthening and repair of RC
structures.

E. Jlio is a full professor at the


Instituto Superior Tcnico of the
Technical University of Lisbon. His
research interests are in the field of
Structural Concrete. He is member
of IABSE WC 3, fib Commission 5
and ACI 364 Committee.

Summary
Structural lightweight aggregate concrete (LWAC) can be used with advantage in new structures,
especially in pre-cast components, as well as in the rehabilitation of existing structures. In both
situations, the result is a composite structural element, with parts in LWAC and others in normal
weight concrete (NWC). Besides material properties, it is also important to characterize the
behaviour of LWAC-to-NWC interfaces. With this aim, an experimental study was conducted to
characterize the shear strength and the tensile strength of NWC-to-NWC, LWAC-to-NWC and
LWAC-to-LWAC interfaces and to compare results with codes predictions. A NWC and three
different LWAC were considered. Six different types of surface roughness were produced on the
concrete substrate and characterized with roughness parameters assessed using the 2D-LRA method,
developed by the authors. Slant shear and splitting tests were adopted to quantify the bond strength
of the interface in shear and in tension, respectively.
Keywords: lightweight aggregate concrete (LWAC); interface; bond strength; roughness
parameters; codes.

1.

Introduction

Structural LWAC presents low density and high performance in terms of both strength and
durability. Therefore, this is a competitive material not only for new structures, whenever reduction
in dead-weight is aimed, e.g. for the pre-casting industry, but also for the rehabilitation of existing
structures. All these situations can give rise to composite structural members, consisting of parts in
LWAC and parts in NWC, casted at different ages. In these cases, it is important to assess not only
the material properties but also the behaviour of the LWAC-to-NWC interface.
The design expressions proposed by codes for concrete structures to predict the shear strength of
concrete-to-concrete interfaces are all based on the shear-friction theory [1] and depend mainly on
four parameters: (i) compressive strength of the weakest concrete; (ii) normal stress at the interface;
(iii) shear reinforcement crossing the interface; and (iv) roughness of the substrate surface. Santos
[2] characterized the influence of other parameters on the behaviour of NWC-to-NWC interfaces,
namely the differential shrinkage and Youngs modulus, that are not considered by current design
codes, particularly by Eurocode 2 (EC2) [3]. Differential shrinkage may not be a relevant parameter
in LWAC-to-NWC interfaces (in the case LWAC is the added layer), due to improved internal
curing provided by the moisture in the interior of the lightweight aggregates (LWA) leading to
reduced shrinkage [4]. However, since the Youngs modulus of LWAC can be quite different from
that of NWC [5], depending on the strength and on the density of the adopted LWAC, differential
stiffness may influence significantly the behaviour of the interface. The strength of the binding
matrix is usually higher than the respective LWAC strength, due to the reduced strength of LWA.
Nevertheless, this is also ignored by codes to predict the interface strength.

In order to characterize the shear and tensile strength of NWC-to-NWC, NWC-to-LWAC and
LWAC-to-LWAC interfaces, an experimental study was conducted to evaluate the influence of
substrate roughness, differential stiffness and LWAC strengths of both LWA and binding matrix.
First, the following concrete mixtures were designed, produced and characterized: a single NWC
mixture, with a compressive strength of 50 MPa; and three LWAC mixtures, with densities and
compressive strengths ranging between 1.5 and 1.9 and 45 and 75 MPa, respectively. Then, six
different texture conditions were considered for the substrate surface: (i) smooth surface (SS);
(ii) surface treated with steel wire-brush (WB); (iii) surface left free after vibration (SF); (iv)
surface treated by shot-blasting (SB); (v) surface prepared by hand-scrubbing/raking (HS); (vi)
surface treated by chemical deactivation (CD). These were all characterized using roughness
parameters assessed with the 2D - laser roughness analyser (2D-LRA) method, developed by the
authors [1]. Slant shear and splitting tests were adopted to quantify the bond strength of the
interface, in shear and in tension, respectively.

2.

Concrete mixture design and production

The design method used for NWC and LWAC mixtures is based on the methodology initially
proposed by Lourenco et al. [6] and further developed by Costa et al. [7], which includes the
following steps: (i) prediction of the binding paste strength, through the Ferets expression;
(ii) adjustment of the mixtures curve to the Faurys reference curve and definition of the wanted
density; (iii) computation of the LWAC strength, considering the strength reduction taking into
account the types and dosages of lightweight aggregates.
2.1
Materials
In the study herein described, the following constituents were adopted for the binding paste: cement
CEM II-A/L 42.5 R; fly ash addition (in the LWAC mixtures); third generation superplasticizer,
polycarboxylates based; water.
The following types of aggregates were used to produce the LWAC used in the study: (i) three types
of coarse lightweight expanded clay aggregates (Leca) - structural Leca 2/4 mm (HD2/4), structural
Leca 4/12 mm (HD4/12) and structural Leca 4/10 mm (MD); (ii) two types of siliceous fine
aggregates - fine sand 0/2 mm (FS) and medium sand 0/4 mm (MS); and (iii) one type of fine Leca
sand 0.5/3 mm (XS). In relation to the NWC mixture, besides FS and MS, crushed limestone 6/12
mm (CL) and gravel 4/8 mm (Gr) were also used.
Granulometric analysis was performed
Table 1: Characterized properties of Leca aggregates.
for all aggregates and the following
Type of
0
HP
AS
AN
fCr
P0
properties were characterized for Leca
3
3
LWA (kg/dm ) (kg/dm ) (%) (%) (%) (MPa) aggregates (Table 1): dry particle density,
HD4/12
1,37
0,76
3,0
9,5
5,6
11,7
P0; dry bulk density, 0; interior
moisture, HP; absorption from natural
HD2/4
1,34
0,74
9,5 11,9 1,5
11,0
state, AN; total absorption until saturated
MD
0,89
0,49
3,1 11,0 7,5
4,0
state, AS; and crushing strength, fCr.
XS

1,02

0,58

4,2

10,2

5,1

5,3

2.2
Produced concrete
Different binding pastes were defined for the NWC and for the LWAC (Table 2), with different
dosages of cement (C), fly ash (FA) and admixture. Thus, these binding paste matrices resulted in
different values of the concrete compactness and water / binder (W/B) mass ratio. Consequently, the
NWC and the LWAC mixtures have different predicted values for the strength of the binding paste
matrix, fbp,p, 50 and 90 MPa, respectively.
In the LWAC mixtures, the granulometric adjustment was performed and combined with the
selection and pre-blending of fine and of coarse aggregates, aiming to achieve the desired
densities, , and compressive strengths, fc,p, between 45 and 76 MPa.

Table 2: Concrete references, aggregates arrangement and mixtures parameters.


Concrete
Fine
Coarse
C
FA
W/B
fbp,p
reference
aggregate
aggregate
(kg/m3) (kg/m3)
(MPa)
N2.4
H1.9
H1.7
H1.5

FS+MS
FS+MS
FS+XS
FS+XS

CL+Gr
HD4/12+HD2/4
HD4/12+HD2/4
MD+HD2/4

340
420
420
420

0
84
84
84

0,50
0,27
0,27
0,27

50
90
90
90

fc,p
(MPa)
50
76
67
45

(kg/m3)
2350
1900
1700
1500

Typical failure surfaces of the adopted concrete are presented in Fig. 1, being possible to identify
the type and distribution of the aggregates in the concrete matrix.

a) N2.4

b) H1.9
c) H1.7
Fig. 1: Failure surface of the produced concrete.

d) H1.5

fcm (MPa)

2.3
Mechanical properties
For all concrete mixtures, the average value of compressive strength (fcm) was characterized in
cubic specimens of 150 mm edge [8], at 7, 28 and 90 days of age. The results are presented in Fig. 2,
together with the corresponding hardening curve proposed by EC2 [3]. The following properties
were also characterized at 28 days of age (Table 3): i) average tensile strength (fctm), assessed with
splitting tests using standard cylindrical specimens [8]; ii) average Youngs modulus [9] (Ecm),
determined on prismatic specimens of 150150600 mm3. For each situation, three specimens were
tested.
90
80
70
60
50
40
30
20
10
0

N2.4
H1.9
H1.7
H1.5
EC2
0

7 14 21 28 35 42 49 56 63 70 77 84 91

Age (days)

Table 3: Measured values, at 28 days, of


tensile strength and of Youngs modulus.
Concrete
fctm
Ecm
reference
(MPa)
(GPa)
N2.4

3,89

33,4

H1.9

4,82

27,1

H1.7

3,70

20,8

H1.5

2,55

14,6

Fig. 2: Evolution of fcm with age.

3.

Interface strength

3.1
Surface roughness
The bond strength of the interface was determined in shear, with slant shear tests, and in tension,
with splitting tests, at 28 days of age of the added layer. For this purpose, two types of concrete
substrate, H1.5 and N2.4, were used. On the specimens with concrete substrate N2.4, four types of
concrete (N2.4, H1.9, H1.7 and H1.5) were added and five different roughness conditions of the
substrate surface (Fig. 3) were adopted: SS, WB, SB, HS and CD. For specimens produced with
concrete substrate H1.5, two types of concrete were added (H1.7 and H1.9) and three different
roughness conditions of the substrate surface (Fig. 3) were considered: SS, SF and CD. For each
situation, three specimens were tested. These were cured in water at 20 C.

The EC2 [3] proposes a qualitative assessment of the roughness of the substrate surface, only based
on a visual inspection. This leads to a classification in four classes: (1) very smooth, (2) smooth,
(3) rough and (4) indented. This qualitative approach is subjective since it depends on the
inspectors judgment. Santos et al. [10] proposed a quantitative classification, and therefore more
rigorous, of the roughness of the substrate, using roughness parameters, namely [10]: maximum
height of peak (Rp) average height of peak (Rpm) maximum depth of the valley (Rv) average depth
of the valley (Rvm). These parameters can be assessed using a specific method and equipment, the
2D- LRA, developed by Santos and Jlio [1].

a) N2.4-SS.

e) N2.4-CD.

b) N2.4-WB.

c) N2.4-SB.

d) N2.4-HS.

f) H1.5-SS.
g) H1.5-SF.
Fig. 3: NWC and LWAC substrate surface roughness.

h) H1.5-CD.

The 2D-LRA method gets the texture profile of the surface and computes the corresponding
roughness parameters. These values were obtained for each surface condition considered in the
study as the average of ten readings (Table 4).
Table 4: Roughness parameters of the considered surface conditions.
N2.4
N2.4
N2.4
N2.4
N2.4
H1.5
Parameter H1.5
(SS)
(WB)
(SB)
(CD)
(HS)
(SF)
Rp (mm)
0,57
1,59
2,15
5,11
8,22
3,36
Rpm (mm)
0,23
0,86
1,38
4,21
4,29
1,97
Rv (mm)
2,25
2,56
4,00
4,90
9,54
2,31
Rvm (mm)
0,63
1,24
2,31
3,56
6,88
1,36

H1.5
(CD)
4,37
3,56
3,71
2,83

3.2
Interface shear strength
The expression defined by EC2 [3] to determine the design value of the longitudinal shear strength
of the interface between concrete layers cast at different ages, Rdi, is given by:

Rdi = cf ctd + n + f yd ( sin + cos ) 0,5f cd

(1)

being c and the cohesion and friction coefficients, respectively. According to EC2, the surfaces of
Fig. 3, as well as the corresponding values of c and , are shown in Table 5.

The interface shear strength was experimentally assessed with slant shear tests [11], using prismatic
specimens of 150150450 mm3 with the interface at 30 degrees with the vertical (Fig. 4a). The
adopted test rate was 5,0 kN/s.
Table 5: Roughness classification and coefficients, of EC2, for the used surfaces.
N2.4
N2.4
N2.4
N2.4
N2.4
H1.5
H1.5
Substrate
H1.5
surface
(SS)
(WB)
(SB)
(CD)
(HS)
(SF)
(CD)
EC2
Very
Smooth Rough Rough Rough Smooth Rough
Classification smooth
c (EC2)
0,25
0,35
0,45
0,45
0,45
0,35
0,45
(EC2)
0,5
0,6
0,7
0,7
0,7
0,6
0,7

Two types of failure were observed: (1) cohesive (monolithic failure of the weakest concrete),
Fig. 4.b; and (2) adhesive (debonding of the interface), Fig 4.c. The first one occurred mostly in
specimens with higher interface roughness (HS, CD, SB) and the second took place mainly in
specimens with lower roughness of the interface (SS, WB, SF).

a) Test specimen.

b) Cohesive failure.
Fig. 4. Slant shear test.

c) Adhesive failure.

The obtained average values of the shear strength are presented in Fig. 5, for each substrate
concrete, each roughness of the interface and each added concrete.

20
15

25

N2.4
H1.9
H1.7
H1.5

Shear strength (MPa)

Shear strength (MPa)

25

10
5
0

H1.9
20

H1.7

15
10
5
0

SS

WB

SB

HS

Interface roughness

CD

SS

SF

CD

Interface roughness

a) Substrate N2.4
b) Substrate H1.5
Fig. 5: Average values of shear strength of the interface.

It was observed that the shear strength increases with the interface roughness. For N2.4 substrate,
the use of LWAC with higher strength of the binding paste matrix results, generally, in higher
interface strength when compared to NWC. It was also observed that reducing the density of added
LWAC, and hence its strength and Youngs modulus, the shear strength of the interface also
decreases.
It was also analyzed the correlation between the roughness parameters and the shear strength of the
interface. In Fig. 6, the correlation with Rvm is presented. This adjusts well to a power function,

although the amplitude depends on the concrete strength. Thus, the coefficients of cohesion and
friction, c and , can also be estimated using this function [2, 9].
25

Shear strength (MPa)

Shear strength (MPa)

25
20
15
N2.4

10

H1.9
H1.7

20
15
10

H1.9
H1.7

H1.5

0
0

Rvm (mm)

Rvm (mm)

a) Substrate N2.4
b) Substrate H1.5
Fig.6: Comparison between the interface shear strength and the parameter Rvm.

Shear strength (MPa)

25,0

When comparing the values of the interface shear


strength, experimentally obtained, with the
corresponding EC2 prediction, Ri, according to
(1), for nominal values and considering the
coefficients of Table 5, it can be observed that
this approach leads to too conservative values. In
fact, experimental results are approximately
twice the predicted values (Fig. 7).

SS
WB
SB
SF
HS
CD

20,0
15,0
10,0
5,0
0,0
0,0

2,5

5,0

7,5

10,0

12,5

Ri (MPa)

Fig.7: Comparison between the shear


strength and the EC2 prediction, Ri

3.3
Interface tensile strength
The tensile strength of the interface, between concretes cast at different ages, was characterized
with splitting tests [7], using cylindrical specimens of 150 mm of diameter and 300 mm of height
(Fig. 8.a), tested at a rate of 2,0 kN/s. The occurrence of two distinct failure modes was observed, as
mentioned before: cohesive (Fig. 8.b) and adhesive (Fig. 8.c).
The cohesive failure was observed on specimens with N2.4 concrete substrate, with high surface
roughness and with H1.5 added concrete. This also occurred on specimens with H1.5 concrete
substrate, with high surface roughness and with H1.9 added concrete. Interface failure was observed
in the remaining situations.

a) Test specimen.

b) Cohesive failure.
Fig. 8: Splitting test.

c) Adhesive failure.

In Fig. 9, the average values of tensile strength of the interface, for both types of concrete substrate,
are presented, for each roughness condition of the interface surface and for each type of added
concrete.
4,5

4,0

N2.4

3,5

H1.9

3,0

H1.7

2,5

H1.5

Tensile strength (MPa)

Tensile strength (MPa)

4,5

2,0
1,5
1,0
0,5

4,0

H1.9

3,5

H1.7

3,0
2,5
2,0
1,5
1,0
0,5
0,0

0,0
SS

WB

SB

HS

CD

SS

Interface roughness

SF

CD

Interface roughness

a) Substrate N2.4.
b) Substrate H1.5.
Fig.9: Average values of the tensile strength of the interface.

Results revealed an increase of the tensile strength with the interface roughness. On specimens with
N2.4 substrate, the use of LWAC with higher strength of the binding paste matrix results on higher
tensile strength of the interface, when compared to NWC, similarly to what was observed with the
shear strength. Results also show that a density reduction of the added LWAC leads, in general, to a
reduction of the tensile strength of the interface.
Comparing the roughness parameters with the tensile strength of the interface, it can be concluded
that Rpm presents a better correlation. This can again be adjusted to a power function (Fig. 10),
although the amplitude also depends on the concrete strength.
5

Tensile strength (MPa)

Tensile strength (MPa)

5
4
3
N2.4

H1.9
H1.7

4
3
2

H1.9
H1.7

H1.5

0
0

4
Rpm (mm)

Rpm (mm)

a) Substrate N2.4.
b) Substrate H1.5.
Fig.10: Comparison between the interface tensile strength and the parameter Rpm.

4.

Conclusions

Both the slant shear and the splitting tests proved to be adequate to evaluate the strength of NWCto-LWAC interfaces, respectively in shear and in tension.
Two failure modes were observed, cohesive (at the weakest concrete) and adhesive (debonding of
the interface). For slant shear tests, the failure mode was mainly monolithic, for very rough surfaces
(HS and CD), and mainly adhesive, for very smooth and smooth surfaces (SS and WB). For rough
surfaces, SF and SB, both failure modes were observed. For splitting tests, the failure mode was
mainly adhesive, excepting for the specimens with a very rough substrate and with significant
differences in both density and Youngs modulus between the substrate and the added concrete
layers.

The interface strength, both in shear and in tension, increases with the roughness of the substrate
surface. Roughness parameters revealed a good correlation with shear and tensile strengths,
adopting power functions. However, it seems that an increase of the substrate roughness above
certain values, does not contribute to significantly increase the interface strength.
An added LWAC layer with high strength of the binding matrix leads to higher interface strength,
when compared to NWC. On the contrary, the addition of a LWAC layer with lower density, hence
with lower strength, leads to a lower strength of the interface.
Finally, it was concluded that the LWAC-to-NWC shear strength, predicted by EC2, is too
conservative, since measured values were nearly the double.

Acknowledgements
The authors acknowledge the financial support of the Portuguese Science and Technology
Foundation (FCT) through the PhD Grant number SFRH/BD/44217/2008. This study was
developed under the research project Intelligent Super Skin - Enhanced Durability for Concrete
Members funded by the Portuguese Science and Technology Foundation (FCT) with reference
PTDC/ECM/098497/2008.

References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]

SANTOS P., JLIO E., Development of a laser roughness analyser to predict in situ the
bond strength of concrete-to-concrete interfaces, Magazine of Concrete Research, Vol. 60,
No. 5, pp 329-337, 2008.
SANTOS P., Assessment of the Shear Strength between Concrete Layers. PhD Thesis.
University of Coimbra, Portugal, 2009.
EN 1992-1-1, Eurocode 2: Design of concrete structures - Part 1.1: General rules and rules
for buildings, CEN, 225 p, 2004.
COSTA H., JLIO E., and LOURENO J., Lightweight Aggregate Concrete codes
review and needed corrections, Codes in Structural Engineering - Developments and Needs
for International Practice, IABSE-fib, Dubrovnik, 2010.
COSTA H., Mixture Design and Mechanical Characterization of Lightweight Aggregate
Concrete, MSc Thesis, University of Coimbra, Portugal, 2008. (In Portuguese).
LOURENO J., JLIO E., and MARANHA P., Lightweight Concrete of Expanded Clay
Aggregates, APEB, Lisbon, 2004. (In Portuguese).
COSTA H., JLIO E., and LOURENO J., A New Mixture Design Method for Structural
Lightweight Aggregate Concrete. 8th fib PhD Symposium, Lyngby, Denmark, 2010.
EN 12390, Testing Hardened Concrete, CEN, 2004.
Specification E 397-1993, Concrete-Testing the Youngs modulus, LNEC, Lisbon, 1993.
SANTOS P.M.D., JLIO E.N.B.S., SILVA V.D., "Correlation between concrete-to-concrete
bond strength and the roughness of the substrate surface", Construction and Building
Materials, Vol. 21(8), pp. 1688-95, 2007.
EN 12615, Products and systems for the protection and repair of concrete structures. Test
methods - Determination of slant shear strength, CEN, 1999.

S-ar putea să vă placă și