Construction and Building Materials xxx (2018) xxx–xxx

Contents lists available at ScienceDirect

Construction and Building Materials

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

Further evidence of interfacial adhesive bond strength enhancement

through fiber reinforcement in repairs
Cristina Zanotti a,⇑, Giulia Rostagno b, Brian Tingley a
a
Department of Civil Engineering, University of British Columbia, 6250 Applied Science Lane, Vancouver, BC V6T 1Z4, Canada
b
Department of Civil, Architectural, Environmental, Land Planning Engineering and Mathematics (DICATAM), University of Brescia, via Branze 43, 25123 Brescia, Italy

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

Adding steel or PVA fibers to a repair

increases cohesive bond to the
substrate.

With 0.5% & 1% fiber volume
fractions, cohesion strength is
increased of up to 100%.

Fibers’ enhancement of substrate-
repair bond requires proper interface
roughness.

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

Article history: While intensive structural and non-structural repair is required worldwide to compensate the current
Received 3 March 2017 infrastructure deficit, repair effectiveness is jeopardized by poor durability, compatibility, and bond.
Received in revised form 13 September Benefits of fiber reinforcement in concrete repairs are remarkable as durability is enhanced and the effect
2017
of poor compatibility can be mitigated. Furthermore, fibers’ potential to improve concrete-concrete bond,
Accepted 21 December 2017
Available online xxxx
a crucial property in repair applications, has been demonstrated. Only a few studies, however, are avail-
able on the effect of fibers on substrate-repair bond and additional analysis form different bond tests,
fiber reinforcements, and substrate treatments are required to fully utilize such benefits in repair appli-
Keywords:
Concrete repair
cations.
Substrate-repair bond In this study, substrate-repair shear bond strength in fiber reinforced repair mortars is investigated.
Adhesion strength Based on previous encouraging results on 8 mm long Poly-Vinyl-Alcohol (PVA) fibers, PVA fibers with dif-
Cohesion ferent lengths (8 and 12 mm) and 13 mm long steel fibers are compared. Two fiber volume fractions, equal
Fiber reinforced concrete to 0.5% and 1% are applied beyond the control condition (plain mortar). While results currently available
Surface roughness are focused on roughened substrates, sandblasted substrates and substrates left as-cast are considered in
these experiments. Substrate-repair bond strength is assessed through Modified Slant Shear Cylinder
(MSSC) test with different bond plane inclinations, corresponding to different normal-shear stress ratios.
Adhesion strength and friction coefficient, two parameters inherently characterizing substrate-repair
bond, are assessed. The bond enhancing mechanisms offered by the different types of fibers and their cor-
relation to surface treatment are discussed. Variations of indirectly determined coefficients are statisti-
cally validated through a permutation technique applied to the 170 samples tested overall.
Ó 2017 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.
E-mail address: zanottic@mail.ubc.ca (C. Zanotti).

https://doi.org/10.1016/j.conbuildmat.2017.12.140
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
2 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx
1. Introduction
Several countries are facing a remarkable infrastructure deficit,
requiring extensive budget investment for maintenance, repair,
rehabilitation, and retrofit. Lack of durability has not only been
observed in concrete infrastructure, but also in repair interventions
affected by substrate-repair debonding and poor compatibility,
which intensify repair vulnerability to environmental attack and
internal chemo-physical deteriorative processes [1,2]. Fiber rein-
forcement represents a viable way to enhance concrete durability.
Fibers can improve crack growth, impact and fatigue resistance
[3,4], control permeability under stress [5], and reduce bleeding
[6], as well as shrinkage and thermal cracking [7–9]. Sporadic stud-
ies have shown that also substrate-repair bond is improved by the
addition of fiber reinforcement to the repair material [10–14].
With substrate-repair interface representing the weakest link of
repair systems, the ability of fibers to enhance not only mechanical
properties and durability, but also interfacial bond, has generated
increasing interest in their implementation for repair applications.
Banthia and Dubeau [10] tested the tensile bond between con-
crete and cement-based composites reinforced with carbon and
steel micro-fibers, to assess their suitability as thin repairs. The
authors performed closed-loop tensile tests on the interface and
observed adhesion enhancement from 0.8 MPa (plain repair) up
to 1.4 MPa when a 3% volume fraction of steel fibers was added
to the mortar matrix. Lim and Li [11] studied the bond between
concrete substrates and overlays made with Engineered Cementi-
tious Composites (ECC); bending tests were performed, which
resulted in a combined state of shear and tension at the interface.
Higher bond strength was observed when ECC was employed,
compared to other cementitious overlays; the improvement was
attributed to the development of a crack trapping mechanism that
arrested crack growth in ECC when the interfacial crack was devi-
ated (kinked [11,15]) out of the bond plane. Zanotti et al. [12]
tested the crack growth resistance of sandblasted interfaces in
Mode-I (pure tension) and found that the addition of 8 mm Poly-
Vinyl-Alcohol (PVA) fibers to the repair mortar enhanced interfa-
cial crack nucleation and resistance to growth. Micro-crack devia-
tion outside the interface was observed and the interfacial crack
improvement was attributed to crack blunting as well as reduced
shrinkage and operational damage. Wagner et al. [13] performed
mechanical characterization of interfaces between concrete sub-
grade and strain hardening cementitious repair layers reinforced
with 8 mm PVA fibers. The authors tested the interfaces in closed
loop under wedge splitting and various combinations of shear
and tensile stresses. In the wedge splitting test (tension), it was
found that PVA fibers were effective only when surface roughness
and bond strength were sufficiently high to allow crack formation
in the fiber reinforced concrete. For combined shear and tension,
conclusive result discussion was referred to ensuing inverse analy-
sis. In this regard, Sajdlová and Kabele [16] have recently offered
novel insights on the stability of Mode II crack propagation
between two or more layers of materials with variable composition
and microstructure.
It is evident that, while the studies presented above highlighted
that fiber reinforcement can enhance interfacial adhesive bond,
knowledge gaps remain and additional studies comparing different
fiber reinforcements, repair matrices, and substrate treatments, as
well as different interfacial stresses and repair size, are required
before the repair sector can comprehensively master the bond
advantages offered by fibers.
In particular, the results available focus on interfaces subjected
to tension (pure or combined to shear), while the effect of fiber
reinforcement on shear bond strength and correlated adhesive
shear bond remains mostly unknown. Zanotti et al. [14,17] pre-
Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
sented a study where interfacial bond was tested with a Modified
Slant Shear Cylinder (MSSC) test under different combinations of
shear and compressive stresses and cohesion and friction coeffi-
cients were extrapolated based on the Mohr-Coulomb approach.
The research demonstrated that 8 mm PVA fibers enhance shear
bond strength and adhesion in sandblasted interfaces [14]; along-
side, additional questions were raised: what would be the effect of
different fiber reinforcements? To what extent may sandblasting
affect the adhesive shear bond amelioration conferred by fibers?
A new study is presented where the same testing technique
adopted in [14], based on MSSC test, Mohr-Coulomb approach,
and statistical validation through permutation analysis, was
applied to compare the effects of three types of fibers, namely: 8
mm long Poly-Vinyl-Alcohol (PVA) fibers, 12 mm long PVA fibers,
and 13 mm long steel fibers. The effect of PVA fiber length was
assessed by comparing 8 mm and 12 mm long fibers. Steel fibers
with similar length (13 mm) were also considered. Steel fibers
and PVA fibers have rather different mechanical, physical, and
chemical properties as well as different adhesion characteristics
to the cement matrix. Therefore, different substrate-repair bond
strengths and different substrate-repair slip resisting mechanisms
can be expected. Beyond the plain mortar as control condition, 0.5%
and 1% fiber volume fractions were employed.
Each repair mortar was applied on two types of surfaces, that is,
sandblasted substrates and substrates left as cast (smooth inter-
face). While there are several studies available on the effect of
sandblasting on substrate-repair bond (such as [17–22] among
others), their comparison within this context demonstrated the
key-role played by substrate treatment in activating the shear
bond enhancing mechanisms offered by fibers.
2. Materials and methods
2.1. Materials
In order to maximize compatibility and consistently with the
previous investigation on 8 mm PVA fibers and sandblasted inter-
faces [14], the mix proportions shown in Table 1 were adopted. The
effects of the following three types of fiber reinforcement were
compared: Poly-Vinyl-Alcohol (PVA) fibers with two different
lengths (8 mm and 12 mm) and 13 mm long steel fibers. Technical
information on PVA and steel fibers is provided in Table 2.
2.2. Test set-up
The slant shear test consists of applying a compressive load to a
cylinder where repair material and substrate are bonded together
at a standard 30° inclination [23] (Fig. 1a), resulting in a combination
of shear and normal stresses along the substrate-repair interface
(Fig. 1b). Compared to the standard slant shear test, a modified ver-
sion previously developed was adopted in this study [24]. This exact
same testing procedure was adopted in [14], so that the previous
results obtained with 8 mm PVA fibers on sandblasted interfaces
can be compared. The main features of the test are summarized here-
after, while more details are provided in [24]. Three different bond
plane inclinations, a, equal to 30°, 25°, and 20° were applied
(Fig. 1c and d) rather than the standard 30° angle only [23]. The cylin-
der diameters were varied to keep similar interfacial areas and min-
imize size effects as the slant angle, a, changes. The distance
between the edges of the cylinder and the bond plane was kept equal
to the diameter to minimize the effect of loading plate-cylinder fric-
tional stresses on the substrate-repair interfacial stresses.
Substrate-repair interfacial shear stresses are given by:
1
sn ¼ r0 sinð2aÞ ð1Þ
2
 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx 3

Table 1
Mix Proportions: ratio to cement content [/c], content per unit volume [kg/m3], and volume fraction of fibers (Vf).

Cement Fly ash Sand 10 mm Aggregate Water Fibers (Vf)

3 3 3 3 3
/c kg/m /c kg/m /c kg/m /c kg/m /c kg/m
Substrate 1 840 0.25 210 2 1680 0.48 400 0.5 420 –
Repair 1 840 0.25 210 2 1680 – – 0.5 420 0%, 0.5%, 1%

Table 2
Technical properties of the fibers employed.

Type of fiber Diameter [mm] Cut length [mm] Tensile strength [MPa] Young’s modulus [GPa] Specific gravity [–]
8 mm PVA 0.04 8 1600 40 1.3
12 mm PVA 0.10 12 1100 28 1.3
13 mm steel 0.20 13 2160 210 7.85

Fig. 1. Slant shear bond test (a), stresses developed at the interface (b), and geometry of cylinders employed in this study (c–e) [All dimensions in mm].

rn ¼ r0 sin2 a ð2Þ of 1 min and up to 2 min depending on when the aggregate would
be visible on the substrate surface. In order to avoid oil spreading
where sn is the shear stress acting parallel to the bond plane, ro is over the contact surface, the concrete was positioned into pre-oiled
the applied axial stress that produced failure along the bond plane, molds. Finally, the new mortar was poured on substrates in opti-
and rn is the normal stress acting perpendicular to the bond plane. mum Saturated Surface Dry (SSD) condition [20] and eventually
A Mohr-Coulomb approach is commonly accepted to character- cured as for plain concrete. Overall, two surface conditions and
ize shear bond strength as a function of normal interfacial stresses seven repair mortars were considered. Twelve specimens (four
and fundamental interfacial properties [22,25]. This approach is replicates for each of the three slant angles) were tested for each
also adopted by design codes [26,27]: of the fourteen combinations of repair mortar and surface treat-
sn ¼ c þ rn tanðUÞ ð3Þ ment. Bond tests were supported by standard compressive test
on cylinders [28].
where c is the adhesion strength (or cohesion) and U is the internal
angle of friction.
3. Results
2.3. Specimen and test preparation
3.1. Failure modes
Concrete substrate was cast first. Pre-formed molds were
All slant shear cylinders exhibited a bond failure; however, in
adopted to cast only half of the final specimen. Samples were cov-
few of the sandblasted cylinders (6 of the 85 sandblasted cylinders
ered with plastic sheets soon after casting, demolded 24 h later,
and cured in standard conditions in a moist/temperature con- Table 3
trolled room. Afterwards, the concrete halves were allowed to Percent product passing ASTM E11 [28] sieves for the two grades of glass abrasive
dry for 24 h; for half of the specimens, the contact surfaces were used for sandblasting.
sandblasted, while for the other half, surfaces were left as cast Sieve Size [mm] GA 30–60 GA 50–100
before being cleaned from dust and put back in the molds. Sand-
0.85 99.9% –
blasting was operated with a pressure washer at 4000 PSI (27.6 0.6 99.7% –
MPa). A commercially available glass abrasive product with less 0.425 63.6% 99.8%
than 1% silica content was used. A 50%-50% combination of two 0.3 15.9% 98%
size ranges (Table 3) was employed. The nozzle was held at a dis- 0.212 3.6% 63.2%
0.15 0.2% 19.7%
tance of 10–20 cm from each surface portion for a minimum time

Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
4 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx
tested), partial material and bond failure was observed; this was
associated with cracks propagating from a different area of the
specimen and eventually reaching the interface in mixed bond-
material failure. The fact that mixed bond-material failure was
observed in some of the sandblasted specimens and not in the
non-treated ones can be explained by the higher bond strength
of the former (as discussed in Section 3.3).
3.2. Compressive strength
The average compressive strength [30] of the fiber reinforced
repair mortar is equal to 66–71 MPa for the 8 mm long PVA fibers
with Vf = 0.5% and 1% respectively, 64–67 MPa with 12 mm long
PVA fibers and 65 MPa with 13 mm steel fibers. The average com-
pressive strength of the plain mortar is 64 MPa and that of plain
concrete is equal to 66 MPa.
3.3. Bond strength
Average shear bond strength, sn,av, obtained with different
quantities and types of fiber reinforcement and in case of both
sandblasted and ‘‘as cast” interfaces, are shown in Fig. 2. Interfacial
shear stresses were calculated through Eq. (1). As expected, surface
treatment had a major impact on bond strength, which was signif-
icantly higher in case of sandblasting, compared to ‘‘as cast” sur-
faces (Fig. 2).
Shear bond strength of repair mortars reinforced with 8 mm
long PVA fiber ranges between 23 and 25 MPa, depending on the
fiber volume fraction (Vf), in case of sandblasted interface and for
bond plane inclination a = 30°. The shear-normal stress ratio
increases for decreasing bond angles and, therefore, bond strength
decreases up to 14–21 MPa for a = 25° and 14 MPa for a = 20°.
Shear bond strength is higher for longer PVA fibers (12 mm) and
is equal to 14–18 MPa (depending on Vf) for a = 20°, 18–20 MPa
for a = 25°, and 24–25 MPa for a = 30°. Overall, shear bond
strength is increased by increasing the PVA fiber content in the
repair mortar (Fig. 2).
A different trend was observed with 13 mm long steel fibers.
Bond strength ranges between 29 and 26 MPa for a = 30°, 23–20
MPa for a = 25° and 18–15 MPa for a = 20°, in case of Vf = 0.5%
and 1%, respectively (Fig. 2). In other words, while the addition
Fig. 2. Substrate-repair shear bond strength, sn, plotted as a function of the fiber volume fraction in the repair mortar (Vf = 0%, 0.5%, 1%), the bond plane inclination (a = 30°,
25°, 20°), the substrate treatment (sandblasted or ‘‘as cast”), and the type of fiber reinforcement: PVA (8 mm), PVA (12 mm), steel (13 mm).
Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
of steel fibers in the repair enhances bond strength (higher than
the one of the plain repair mortar), one will notice that mortars
with Vf = 0.5% exhibit higher bond strength than the ones with
Vf = 1% (this aspect is further discussed in the next sections).
In conclusion, substrate-repair shear bond strength is signifi-
cantly affected by the inclusion of PVA and steel fibers, in case of
sandblasted substrate. Conversely, if substrates are left as cast
(i.e., not treated) the addition of fiber reinforcement does not sig-
nificantly affect bond strength in most instances (Fig. 3). For
smooth interfaces, more insightful information on the effect of
fibers can be drawn by the following analysis of interaction dia-
grams, cohesion variations, and their statistical significance.
3.4. Interaction diagrams
While average shear bond strength values were discussed
above, shear-normal stress data points from single slant tests are
plotted in Fig. 3. A different combination of fiber type and content
is presented in each of the seven charts. Interfacial shear and nor-
mal stresses were determined through Eqs. (1) and (2). Shear-
normal stress interaction diagrams were obtained through linear
fitting of experimental data points; such fitting is relevant because
three different slant inclinations, corresponding to different shear-
normal stress ratios, were applied. Based on Mohr-Coulomb failure
criterion (Eq. (3)), cohesion, c, and friction angle, U, were extrapo-
lated from linear fitting. The linear fitting resulted suitable to sand-
blasted interfaces (R2 = 0.90–0.99); however, lower R2 values were
obtained for smooth (as cast) interfaces (R2 = 0.69–0.97). Hence,
before analysing the variations of c and U among species, their sta-
tistical confidence needs to be assessed.
3.5. Statistical validation
The statistical confidence of cohesion and friction angle varia-
tions among different species was assessed through a statistical
analysis technique developed when the effect of 8 mm PVA fibers
was first analysed [14] and run through the software Matlab. This
technique is required as the bond coefficients (cohesion and fric-
tion angle) are indirectly determined from fitting of experimental
data points and therefore.
 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx 5

Fig. 3. Interfacial shear-normal stress (sn-rn) interaction points at failure and linear fitting for plain repair mortar (a), and for mortars reinforced with 0.5% (b, d and f) and 1%
(c, e and g) volume fractions of 8 mm PVA fibers (b and c), 12 mm PVA fibers (d and e), and 13 mm steel fibers (c).

Individual test data in terms of vertical stress at failure, ro,v,
were permuted for each of the 14 species considered (i.e., for each
fiber type, volume fraction and substrate treatment). Pairs of data
within a species were obtained from results with different slant
angles. Cohesion and friction values were then extrapolated by lin-
ear fitting of each pair of experimental data through Eqs. (1)–(3).
Since three (n = 3) slant angle tests were performed on four repli-
cates (m = 4), a total of k = mn = 64 values of cohesion and friction
angle were obtained for each species. Sets of results for two species
that differ for one of the three factors considered (fiber volume
fraction, fiber type, and surface treatment) were compared by run-
ning a statistical T-test between the rows of cohesion and friction
values obtained with permutation. Statistical confidence is gener-
ally fulfilled when the probability, P, that the null hypothesis is

Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
6 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx
true (that is, no statistical evidence that the values being compared
are significantly different) is less than 5%. Comparisons where P >
5% are considered not significant, comparisons around P = 5% are
considered marginal.
T-test results on c and U variations in sandblasted interfaces
with different types or amounts of fiber reinforcement in the repair
mortars are plotted in Fig. 4. One can note that P is significantly
lower than 5% (confidence threshold, Fig. 4) in most comparisons.
Cohesion variations observed with the inclusion of different
amounts of PVA or steel fiber reinforcement in a repair mortar
are statistically significant when the substrate is sandblasted. Both
fiber content and fiber type are influencing factors in this regard.
Cohesion variations result marginal rather than significant only
in the following two instances: when the 8 mm PVA fiber volume
fraction is increased from 0.5% to 1% and when 8 mm and 12 mm
PVA fibers are compared in case of Vf = 0.5%. Accordingly, similar
cohesion values were obtained in those cases, as shown in the next
Section.
Friction variations among sandblasted specimens are less statis-
tically significant than cohesion variations (Fig. 4). With surface
treatment being the most important factor when it comes to fric-
tion, fiber reinforcement has a second order effect on interfacial
friction angle. As a result, some specimens exhibit similar friction
angles, with marginal or non-significant deviations (13S-05 vs
13S-1, 8P-0.5 vs 8P-1, 8P-0.5 vs 12P-0.5, 8P-1vs 12P-1, and 12P-
0.5 vs 13s-0.5, Fig. 3).
Overall, lower statistical significance was obtained in smooth
(non-sandblasted) interfaces, compared to the sandblasted ones
(Fig. 4). Such decrease in overall statistical confidence is related
to the lower bond strength of non-sandblasted specimens and,
especially, to the lack of ductile slip resisting mechanisms avail-
able, as thoroughly discussed in Section 4. Only 3 of 13 cohesion
variations can be confidently considered significant in case of
Fig. 4. Probability, P, that the null hypothesis is true from T-test on pairs of rows of data (for both cohesion and friction angle) belonging to different repair mortars, in case of
either sandblasted interface or interface left as cast (smooth).
Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
smooth interface and they are obtained with inclusion of 12 mm
PVA fibers (Fig. 4). Non-significant friction angle variations are
obtained in smooth interfaces (Fig. 4). Similarly, also the effect of
substrate roughness (sandblasted vs smooth) is not significant in
this study. While it is well recognized that surface roughness
affects the friction angle, reasons for this statistical outcome lie
in the low bond strength of non-sandblasted specimens, whose
experimental shear normal-stress data points are located near
the origin of the interaction diagrams (Fig. 4). This, combined to
the larger strength scatter in smooth interfaces (resulting from
the lack of more ductile slip resisting mechanisms such as mechan-
ical interlock - see Chapter 4), leads to non-statistically-significant
results in terms of friction variations. Therefore, while cohesion
variations will be discussed for both sandblasted and smooth spec-
imens, the friction angle will be analyzed only for sandblasted
specimens.
When comparing the sandblasted and smooth interfaces statis-
tically (Fig. 5), it is found that all cohesion values rejected the null
hypothesis (P  5%) and, therefore, cohesion variations obtained as
a function of surface treatment are significant.
3.6. Substrate-repair pure shear strength (cohesion)
Cohesion values are plotted in Fig. 6 for varying the fiber vol-
ume fraction in the repair mortar and for different types of fiber
reinforcement and surface treatments.
A steady cohesion enhancement is observed for the sandblasted
substrates, by increasing the content of PVA fibers in the repair
mortar (Fig. 6); cohesion is increased up to 103% with PVA fibers.
For the Vf range considered in this study (0–1%), a linear c-Vf trend
is noted with both 8 mm and 12 mm long PVA fibers. 12 mm long
fibers offer a better performance than 8 mm long PVA fibers, with
10–20% higher c values, depending on Vf.. Cohesion trends are in
 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx 7

Fig. 5. Probability, P, that the null hypothesis is true from T-test on pairs of rows of data (for both cohesion and friction angle) belonging to different interface treatments
(sandblasted or not) when the same repair mortar is applied.

Fig. 6. Substrate-repair interfacial cohesion, c, obtained varying the fiber volume fraction, Vf, in the repair mortar for different types of fiber and interface treatment
(sandblasted or as cast).

agreement with the shear bond strength variations discussed in
Section 3.3.
A considerable cohesion increase (98%, Fig. 6) is obtained by
introducing a 0.5% volume fraction of steel fibers in the plain
matrix; 0.5% steel fibers provide about the same cohesion strength
achieved with twice the volume fraction of PVA fibers (Vf = 1%,
Figs. 3 and 6). However, only a 62% cohesion increase is obtained
by increasing Vf from 0% to 1%. In other words, increasing Vf from
0.5% to 1% causes a drop in cohesion. This reversed variation con-
cerns only steel fibers but not PVA fibers, and could be attributed
to the stiffer nature of the steel fibers and their larger diameter
as compared to the expected size of the interface. It is worth men-
tioning that the test was repeated three times on the 12 specimens
(4 replicates for each of the 3 slant angles) with 1% Vf of steel fibers
and different amounts of superplasticizers and the same cohesion
value was obtained. The effect of higher steel fiber contents on
cohesion strength and interfacial failure mechanisms is currently
under investigation.
Much lower cohesion values are obtained for interfaces left as
cast, in comparison with sandblasted interfaces. In addition, fiber
reinforcement has a small influence on cohesion in this case, as val-
idated by the statistical analysis (Fig. 4). The only remarkable
increase in the cohesion is obtained by introducing a 1% volume
fraction of 12 mm long PVA fibers (Fig. 6).

Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
8 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx
4. Discussion
In this session, the bond enhancing mechanisms behind the
cohesion variations observed are discussed. Since cohesion is the
equivalent of the pure interfacial shear strength (sn) in absence
of normal stresses (rn = 0), only the resisting mechanisms involved
when rn = 0 are considered hereafter. Compared to smooth inter-
faces left as cast, sandblasted interfaces exhibited an increase not
only in cohesion but also in statistical significance, due the capabil-
ity of interfacial roughness to promote deviation of the fracture
plane from the bond plane [29]. Such deviation entails a failure
mode shift from bond failure to mixed bond and localized material
failure (in both substrate and repair, Fig. 7) [31] and thereby
enhances shear resistance by means of additional shear resisting
mechanisms that are negligible in smooth interfaces. With refer-
ence to a micro-mechanics approach [33], the following three
interfacial shear resisting mechanisms are identified [31–35]:
interlocking, overriding, and fracturing. Interlocking and overrid-
ing require interfacial roughness to be activated; in addition, local-
ized compressive crushing occurring with fracturing is negligible
in smooth interfaces as crack roughness is minimized (while, con-
versely, it is promoted in case of roughened interface). In other
words, when the interface is smooth, the only slip resisting mech-
anism available under pure shear stresses is interfacial fracture
resistance with minimized effect of localized crushing [33]. There-
Fig. 7. Shear bond failure mechanisms of a sandblasted surface [31] and qualitative prediction of steel fiber orientation based on surface roughness profile (from laser
scanning technique) and direction of flow at casting.
Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
fore, in absence of other mechanisms that can mitigate micro-crack
propagation, not only overall shear strength and cohesion are
much lower (Figs. 2,3 and 6), but also the sensitivity to micro-
cracks and defects is much higher for smooth interfaces. This gen-
erates larger strength fluctuations, resulting in lower statistical
confidence, as observed for both linear fitting of experimental data
points (lower R2, Fig. 3) and for cohesion coefficient variations
among different species (higher P, Fig. 4).
Those same mechanisms are helpful to understand why fiber
reinforcement has a different impact on shear strength of sand-
blasted and smooth interfaces, and to visualize the additional bond
enhancing mechanisms involved in presence of fiber
reinforcement.
With the sandblasted surfaces, one could expect that the larger
depressions in the interface profile could encourage the steel fibers
to run non-parallel to the overall shear plane during casting. To
help visualize this effect, a qualitative prediction of the steel fiber
orientation as a function of direction of flow, substrate roughness,
and substrate inclination is provided in Fig. 7. Steel fibers are
known to tend to lie within the matrix in the direction of flow
[31]; the effect of the surface roughness profile on fiber orientation
is clearly affected by the relative size of fibers and concavities. In
these experiments, steel fibers are 13 mm long and the deepest
gaps reach about 4 mm depth, as shown in Fig. 7, where the profile
curve obtained with laser scanning technique of a representative
 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx
sandblasted substrate is plotted along the substrate-repair bond
plane. Steel fibers are also sketched considering their actual
dimensional proportion to the interface roughness profile. Eventu-
ally, localized grouping of steel fibers were observed on the failure
surfaces of both substrate and repair halves of the sandblasted
specimens (Fig. 8). From those pictures, one can conclude that steel
fibers settled inside the larger concavities within the substrate sur-
face and created a ‘‘dowel effect” as they intersected the shear/
fracture plane.
These findings are further validated by juxtaposition of the
cohesion increments obtained for sandblasted and smooth inter-
faces (Fig. 6). No significant cohesion enhancement was recorded
when steel fibers reinforcement was applied in combination to a
smooth interface (Fig. 4). The cut section of a steel fiber reinforced
repair mortar on a smooth interface after failure is shown in Fig. 9.
Steel fibers lied mostly parallel to the bond plane and were not
affected by the failure plane. Therefore, their contribution to bond
was null, as quantitatively reported (Figs. 4 and 6). In addition,
although steel fibers may still be beneficial in reducing shrinkage
cracking and delamination of the repair, their grouping together
nearby and parallel to the bond plane generates concern as regions
of weakness can form that negatively impact the substrate-repair
bond strength.
Significant cohesion enhancement was obtained also with PVA
fibers, for sandblasted specimens. PVA fibers are much more
deformable than steel fibers and a dowel effect is not possible in
this case; on the other hand, PVA fibers can still be engaged in
the failure process through different slip resisting mechanisms.
PVA fibers reduce hygro-thermal damage [36,37], such as repair
cracking and delamination induced by restrained shrinkage, and
mitigate operational damage by enhancing damage tolerance and
promoting interfacial soundness while the bond is being devel-
oped. In addition, PVA fibers enhance interfacial failure resistance
through micro-crack deviation and blunting [12,14]. Like steel
Fig. 8. Fracture surface of a sandblasted substrate where steel fibers are bound to repair mortar after failure (a) and fiber-concrete details from optical microscope (b and c).
Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
9
fibers, also PVA fibers were engaged during failure of sandblasted
specimens. Fibers bound to repair mortar residues were observed
along the fracture surface of the concrete substrate, as shown in
Fig. 10. The presence of coarse aggregate in the picture is reassur-
ing about the fact that those fracture surfaces belong to the con-
crete substrate rather than the repair mortar. Bond enhancing
mechanisms offered by PVA fibers are more effective with 12
mm fiber length than with 8 mm fiber length, given that 10–20%
higher cohesion was achieved with the former in case of sand-
blasted specimens. This trend is in agreement with previous results
on the effect of PVA fiber length on concrete crack bridging [4].
In sandblasted specimens, cohesion enhancements due to
fiber reinforcement are accompanied by small though steady
friction reductions (Fig. 3). Although friction variations are less
significant (10%) and should be regarded as second order effects,
they were statistically validated (Fig. 4). Based on the visual
observation of repair residues and fibers along the substrate
fracture plane of sandblasted specimens, it is reasonable to con-
clude that fiber reinforcement, by increasing interfacial cohesive-
ness, slightly increases the deviation of the fracture plane from
the bond plane (Fig. 7) and that such deviation will slightly
decrease roughness, and, hence, friction along the shear fracture
plane.
The most notable cohesion enhancement observed in non-
sandblasted specimens is the one obtained with 1% Vf of 12 mm
PVA fibers (Fig. 6), where cohesion is increased from 0.72 MPa to
1.77 MPa. Marginal enhancements are obtained with smaller
amounts of PVA fibers (Figs. 3 and 6). Although cohesion is still
very low, this outcome outlines the ability of PVA fibers to enhance
interfacial cohesion even in smooth or quasi-smooth interfaces, as
opposed to the behavior of steel fibers, which have no impact on
cohesive bond of smooth interfaces (Figs. 3 and 6). These different
performances are mostly attributed to the different stiffness and
diameter of PVA and steel fibers (Table 2).
10 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx

Fig. 9. Cut section of a steel fiber reinforced repair mortar after failure in case of smooth substrate (a) and detail of a steel fiber parallel to the fracture surface from optical
microscope (b).

Fig. 10. Fracture surface of sandblasted substrate where PVA fibers are bound to repair mortar residues after failure (a) and fiber-concrete details from optical microscope (b–
d).

5. Concluding remarks
A study on the effect of various fiber reinforcements on
substrate-repair shear bond strength was presented. 0%, 0.5%,
and 1% volume fractions (Vf) of 8 mm long PVA fibers, 12 mm long
PVA fibers, and 13 mm long steel fibers were compared. Two dif-
ferent substrate treatments (sandblasted and ‘‘as cast”) were con-
sidered to investigate the role of interfacial roughness on the
activation of the bond enhancing mechanisms offered by fibers.
Major conclusions can be summarized as follows:

Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
 C. Zanotti et al. / Construction and Building Materials xxx (2018) xxx–xxx
- Sandblasted interfaces exhibited higher pure shear strength
than interfaces left as cast. In addition, results on non-
sandblasted interfaces had higher variability and lower statisti-
cal confidence overall, due to their higher sensitivity to crack
propagation, without additional mechanisms available (i.e.,
interlocking, overriding, or localized crushing during
fracturing).
- Steel fibers increased adhesion strength up to 98.5% in sand-
blasted interfaces, where a dowel effect was observed.
- In sandblasted interfaces, PVA fibers offered a substantial
increase of interfacial adhesive bond strength (from 5 MPa up
to 11 MPa), resulting from micro-crack deviation and blunting,
as well as from minimization of shrinkage cracking and opera-
tional damage [36,37].
- 12 mm PVA fibers performed better than 8 mm PVA fibers over-
all. Although steel fibers performed better than 12 mm PVA
fibers for 0.5% Vf, their performance was lower at 1% Vf. Such
variations are affected by the stiffer nature of the steel fibers
and their larger diameter as compared to the expected size of
the interface.
- Enhanced substrate roughness achieved with sandblasting
played a key-role in allowing deviation of the fracture plane
from the bond plane and therefore promoting the activation
of the bond enhancing mechanisms offered by fibers. In case
of steel fibers, sandblasting also promoted fiber orientation
non-parallel to the shear plane nearby the interface, which
made the dowel effect possible.
Acknowledgement
The support of the Natural Science and Engineering Research
Council of Canada, NSERC, is gratefully acknowledged (grant num-
ber RGPIN-2016-03863).
References
[1] P.H. Emmons, A.M. Vaysburd, J.E. McDonald, A rational approach to durable
concrete repairs, Concr. Int. 18 (3) (1993) 34–38.
[2] C. Zanotti, S. Talukdar, N. Banthia, A state-of-the-art on concrete repairs and
some thoughts on ways to achieve durability in repairs. In: Infrastructure
corrosion and durability- a sustainability study, 2014.
[3] A. Bentur, S. Mindess, Fiber Reinforced Cementitious Composites, Taylor and
Francis, New York, 2007.
[4] N. Banthia, K. Chokri, Y. Ohama, S. Mindess, Fiber-reinforced cement based
composites under tensile impact, J. Adv. Cem. Based Compos. 1 (1994) 131–
141.
[5] N. Banthia, C. Zanotti, M. Sappakittipakorn, Sustainable fiber reinforced
concrete for repair applications, Constr. Build. Mater. 67 (2014) 405–412.
[6] C. Qi, J. Weiss, J. Olek, Characterization of plastic shrinkage cracking in fiber
reinforced concrete using image analysis and a modified Weibull function,
Mater. Struct. 36 (2003) 386–395.
[7] N. Banthia, R. Gupta, Influence of polypropylene fiber geometry on plastic
shrinkage cracking in concrete, Cem. Concr. Res. 36 (7) (2006) 1263–1267.
[8] A. Reggia, S. Sgobba, F. Macobatti, C. Zanotti, F. Minelli, G.A. Plizzari,
Strengthening of a bridge pier with HPC: modeling of restrained shrinkage
cracking, Key Eng. Mater. 711 (2016) 1027–1034.
[9] N. Banthia, V. Bindiganavile, F. Azhari, C. Zanotti, Curling control in concrete
slabs using fiber reinforcement, J. Test. Eval. 42 (2) (2014) 1–8.
Please cite this article in press as: C. Zanotti et al., Further evidence of interfacial adhesive bond strength enhancement through fiber reinforcement in
repairs, Constr. Build. Mater. (2018), https://doi.org/10.1016/j.conbuildmat.2017.12.140
11
[10] N. Banthia, S. Dubeau, Carbon and steel microfiber-reinforced cement-based
composites for thin repairs, J. Mater. Civil Eng. 6 (1) (1994) 88–99.
[11] Y.M. Lim, V.C. Li, Durable repair of aged infrastructures using trapping
mechanism of Engineered Cementitious Composites, Cem. Concr. Comp. 19
(1997) 373–385.
[12] C. Zanotti, N. Banthia, G. Plizzari, Towards sustainable repairs: substrate-repair
interface Mode-I fracture analysis, Int. J. Sustain. Mater. Struct. Syst. 1 (3)
(2014) 265–281.
[13] C. Wagner, N. Bretschneider, V. Slowik, Characterization of the interface
between strain hardening cementitious repair layers and concrete subgrade.
FraMCoS-8, Toledo (Spain), 2013.
[14] C. Zanotti, N. Banthia, G. Plizzari, A study of some factors affecting bond in
cementitious fiber reinforced repairs, Cem. Concr. Res. 63 (2014) 117–126.
[15] M.Y. He, J.W. Hutchinson, Kinking of a crack out of an interface, J. Appl. Mech.
56 (1989) 270–278.
[16] T. Sajdlová, K. Petr, Analysis of stability of Mode II crack growth, Appl. Mech.
Mater. 825 (2016) 161–164.
[17] C. Zanotti, P.H.R. Borges, A. Bhutta, N. Banthia, Bond strength between concrete
substrate and metakaolin geopolymer repair mortar: effect of curing regime
and PVA fiber reinforcement, Cem. Concr. Comp. (2017). In Press, Accepted
Manuscript, Available online 20 February 2017.
[18] International Concrete Repair Institute, Selecting and specifying concrete
surface preparation for sealers, coating, and polymer overlays. Technical
Guideline n 03732 (Reapproved 2002), 1997.
[19] E.N.B. Jùlio, F.A.B. Branco, V.D. Silva, Concrete-to-concrete Bond Strength.
Influence of the roughness of the substrate surface, Constr. Build. Mater. 18
(2004) 675–681.
[20] A. Momayez, M.R. Ehsani, A.A. Ramezanianpour, H. Rajaie, Comparison of
methods for evaluating bond strength between concrete substrate and repair
materials, Cem. Concr. Res. 35 (2005) 748–757.
[21] D.R. Morgan, Compatibility of concrete repair materials and systems, Constr.
Build. Mater. 10 (1) (1996) 57–67.
[22] A.I. Abu-Tair, S.R. Rigden, E. Burley, Testing the bond between repair materials
and concrete substrate, ACI Mater. J. 93 (6) (1996) 553–558.
[23] S. Austin, P. Robins, P. Youguang, Shear bond testing of concrete repairs, Cem.
Concr. Res. 29 (1999) 1067–1076.
[24] A.S.T.M. Standard C 882, 1978, Standard Test Method for Bond Strength of
Epoxy-Resin Systems Used With Concrete By Slant Shear, ASTM International,
West Conshohocken, PA, 2005.
[25] C. Zanotti, N. Banthia, Modified Slant Shear Cylinder (MSSC) test for inherent
characterization of bond in concrete repairs, Indian Concr. J. 90 (8) (2016).
[26] M.E. Mohamad, I.S. Ibrahim, R. Abdullah, Friction and cohesion coefficients of
composite concrete-to-concrete bond, Cem. Concr. Comp. 56 (2015) 1–14.
[27] EN B. 1-1:2004, Eurocode 2: design of concrete structures. General rules and
rules for buildings, 1992.
[28] A.S.T.M. Standard E11–17, Standard specification for Woven Wire Test Sieve
Cloth and Test Sieves, ASTM International, West Conshohocken, PA, 2017.
[29] Model Code 2010 (2010) First complete draft – vol 1. Comité Euro-
International du Béton, Secretariat permanent, Case Postale 88, CH-1015
Lausanne, Switzerland 318.
[30] A.S.T.M. Standard C39, Standard Test Method for Compressive Strength of
Cylindrical Concrete Specimens, ASTM International, West Conshohocken (PA),
2010.
[31] S. Austin, P. Robins, Y. Pan, Shear bond testing of concrete repair, Cem. Concr.
Res. 29 (1999) 1067–1076.
[32] N. Randl, Design recommendations for interface shear transfer in fib Model
Code 2010, Struct. Concr. 14 (3) (2013) 230–241.
[33] J.C. Walraven, E. Vos, H.W. Reinhardt, Theory and experiments on the
mechanical behavior of cracks in plane and reinforced concrete subjected to
shear loading, Heron (Netherlands) 26 (1A) (1981) 68.
[34] P.M.D. Santos, E.N.B.S. Júlio, Interface shear transfer on composite concrete
members, ACI Struct. J. 111 (1) (2014) 113–122.
[35] M.P. Divakar, A. Fafltis, Micromechanics-based constitutive model for interface
shear, J. Eng. Mech.-ASCE 118 (7) (1992) 1317–1337.
[36] D.Y. Yoo, N. Banthia, Mechanical properties of ultra-high-performance fiber-
reinforced concrete: a review, Cem. Concr. Comp. 73 (2016) 267–280.
[37] T. Kanda, V.C. Li, Interface property and apparent strength of a high strength
hydrophilic fiber in cement matrix, ASCE J. Mater. Civil Eng. 10 (1) (1998) 5–
13.