Cement & Concrete Composites 31 (2009) 29–38

Contents lists available at ScienceDirect

Cement & Concrete Composites

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

Corrosion resistance of self-consolidating concrete in full-scale reinforced beams

A.A.A. Hassan, K.M.A. Hossain, M. Lachemi *
Department of Civil Engineering, Ryerson University, 350 Victoria St, Toronto, ON, Canada M5B 2K3

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

Article history: The corrosion of steel reinforcement embedded in full-scale self-consolidating concrete (SCC) beams was
Received 5 January 2008 investigated compared to normal concrete (NC). 400 mm width  363 mm depth  2340 mm length
Received in revised form 23 October 2008 beams containing epoxy- and non-epoxy-coated stirrups were monitored under an accelerated corrosion
Accepted 23 October 2008
test. The corrosion performance of NC/SCC beams was evaluated based on the results of current measure-
Available online 13 November 2008
ment, half-cell potential tests, chloride ion content, mass loss and bar diameter degradation. The inves-
tigation also included the effect of admixture type and the size of specimen on corrosion performance.
Keywords:
In general, SCC beams showed superior performance compared to their NC counterparts in terms of
Corrosion
Self-consolidating concrete
corrosion cracking, corrosion development rate, half-cell potential values, rebar mass loss and rebar
Chloride diameter reduction. However, SCC beams showed localized corrosion with concrete spalling due to
Durability non-uniform concrete properties along the length, which was a result of the casting technique. The
results also showed that the difference between SCC and NC mixes in terms of corrosion was more pro-
nounced in large-scale beams, and that types of admixture used in SCC have no influence on corrosion
performance.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction properties only; information on testing the variation of durability

properties along the length of the element was missing.
The problem of casting concrete in heavily reinforced sections Corrosion of steel reinforcement in concrete is a major aspect
has been a major topic of interest in construction. Proper consoli- affecting concrete durability. When concrete is subjected to a chlo-
dation and placement of concrete in such heavy reinforced sections ride-rich environment, the chloride ions can penetrate and diffuse
require adequate compaction by internal or external mechanical through the body of the concrete, ultimately reaching the steel bars
vibrators operated by skilled workers. However, when passing and causing corrosion. Concrete that has low permeability and
through narrow areas, excessive vibration can lead to segregation, dense microstructure is believed to obstruct chloride diffusion
bleeding and blockage of concrete particles. The use of self-consol- through the concrete body, which helps reduce the rate of corro-
idating concrete (SCC) is one way to reduce the intensive labor de- sion of embedded reinforcing steel.
mand for vibration and to alleviate the problems arising from SCC is expected to have a dense and less permeable microstruc-
bleeding and segregation. SCC is a highly flowable concrete; it is ture because of its superior resistance to bleeding and segregation.
usually poured from one side of the formwork and spreads under In addition, segregation and bleeding accompanying excessive
its own weight, filling all spaces and corners until it reaches the vibration are not a factor in SCC. The production of SCC usually in-
other side and levels itself without segregation, and without the volves the use of high range water reducer (HRWR), superplasticiz-
use of vibrators for consolidation. er and/or supplementary cementing materials (SCM). SCM has
A number of researchers have investigated the uniformity of been proven to increase the concrete corrosion resistance [5–7],
in situ SCC mixtures in relatively small and large specimens and while HRWR helps to disperse cement particles in the mix and re-
compared the results with that of normal concrete (NC) mixtures duce overall concrete permeability [8,9].
[1–4]. Their studies involved testing cores, pull-out of pre-embed- The durability of SCC has been investigated by a number of
ded inserts and/or rebound hammer to evaluate the in situ researchers [10–13] and proved to be excellent. Their investiga-
strength of beams or columns in different locations. Although their tions involved rapid chloride permeability, water permeation, sul-
conclusion indicated that the in situ SCC mixtures were consider- fate resistance/expansion, freezing and thawing and/or deicing salt
ably uniform, their results were based on testing of mechanical scaling resistance tests. However, their investigation did not ad-
dress the durability of SCC in terms of corrosion performance
and cracking. Moreover, the investigated SCC samples/specimens
* Corresponding author. Tel.: +1 416 979 5000; fax: +1 416 979 5122. in their studies were always relatively small and the durability
E-mail address: mlachemi@ryerson.ca (M. Lachemi). properties of those samples may not reflect the in situ durability

0958-9465/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cemconcomp.2008.10.005
30 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38

and uniformity performance of large SCC structural elements such test, as well as chloride ion content near the bar surface, crack pat-
as beams. In general, the effect of segregation and bleeding (which terns and widths, mass loss of rebar/stirrup and reduction of the
influences durability performance) is more pronounced in large- longitudinal bar diameter at the end of the test.
scale beams when compared to small concrete samples [14]. In
addition, the rebar position and casting condition/technique are 2.2. Second stage: studying the corrosion performance in small
different than those in the small concrete samples. In large-scale cylinder specimens
beams, the quality of concrete below lower horizontal bars is ex-
pected to be weak and porous due to insufficient compaction and This stage focused on the corrosion of reinforcement in small
restraint from the horizontal bars in this area [15]. This is likely cylinder specimens made with the same NC/SCC mixes used for
to occur in full-scale beams rather than in small laboratory sam- the first stage. In addition, two other SCC mixtures – similar to
ples and would definitely affect concrete durability and rebar cor- the SCC used in the first stage but with different types of high
rosion. Therefore, testing the durability of SCC/NC in full-scale range water reducers – were also used. Accelerated corrosion test-
beams is essential, since SCC mixes are likely to present distinct ing monitored corrosion initiation, corrosion rate, crack patterns
advantages over NC in such situations. and widths during the study period. The primary objective of this
The objective of this paper is to investigate the durability per- stage was to verify the influence of bleeding, segregation and cast-
formance of full-scale SCC/NC beams in terms of corrosion resis- ing technique (which is manifested in full-scale beams rather than
tance. This paper also compares the corrosion performance in small laboratory cylinders) in enhancing SCC durability and corro-
full-scale beams with that observed in small laboratory cast cylin- sion protection. The secondary objective was to investigate the ef-
ders to manifest the effect of bleeding, segregation and rebar cast- fect of different types of high range water reducers in SCC
ing position (that occur in full-size beams rather than small corrosion protection (if any).
cylinders) on corrosion performance. The corrosion test was car-
ried out using an accelerated corrosion test, in which the concrete
samples were partially immersed in sodium chloride solution, and 3. Experimental procedure
the corrosion rate was monitored by the amount of current passing
with time. 3.1. Specimen preparation

For the first stage, a total of four concrete beams were used; two
2. Research program made with SCC and two with NC. The four beams were divided into
two sets:
The research program described herein was divided into two
main stages:  The first set contained one SCC beam and its NC counterpart.
Both beams had non-epoxy-coated stirrups in certain locations
 First stage: studying the corrosion performance in full-scale (three on both sides and one in the middle) and were designed
beams. for moderate corrosion levels.
 Second stage: studying the corrosion performance in small cyl-  The second set contained another SCC beam and its NC counter-
inder specimens. part with epoxy-coated stirrups (same number and locations as
the first set) and were designed for severe corrosion levels.

2.1. First stage: studying the corrosion performance in full-scale beams All beams were 400 mm wide, 363 mm deep and 2340 mm
long. The four beams contained three 25 M longitudinal bars at
This stage was designed to study the corrosion of reinforcement the bottom and two 15 M bars at the top. The longitudinal bar cov-
(longitudinal steel and stirrups) and corrosion-induced cracking in er was 40 mm, while the cover below stirrups was 30 mm. The
full-scale SCC beams compared to representative NC beams. The beam designation included a combination of letters: SCC or NC to
variation of corrosion performance along the beam length/perime- indicate the concrete type, and E or N to indicate the epoxy- and
ter was examined to assess the durability of SCC beams. Since cor- non-epoxy-coated stirrups. For example, a SCC containing epoxy-
rosion performance is greatly affected by type of admixture, both coated stirrups is designated as SCC-E.
SCC and NC mixes were made from the same types of admixture For the second stage, a total of eight concrete cylinders
and materials. The only differences between the two mixes were (100 mm in diameter and 200 mm in height), reinforced axially
the mix design and the addition of high range water reducer in with a single 25 M bar at the centre, were used. The cylinders were
the SCC in order to obtain the required workability. divided into four groups of four concrete types. Each group con-
The beams were divided into two sets: the first set contained tained two cylinders from each concrete type. The four types were
non-epoxy-coated stirrup beams investigated until they reached NC and SCC mixes used in the first stage and two other SCC mixes
a moderate corrosion level. The second set contained epoxy-coated made with two different high range water reducers.
stirrup beams investigated until they reached a severe corrosion
level. The epoxy-coated stirrups were chosen to help study the 3.2. Materials
degradation of beam strength due to longitudinal bar corrosion;
that study is not included in this paper. The test for the moderate Four mixes (one NC and three SCCs) were designed to achieve
corrosion level was terminated after either of the two beams (SCC similar compressive strength. NC and SCC mixes were similar in
or NC) reached 10% theoretical mass loss of rebar/stirrup. For the terms of materials used but had different mix proportions. The
severe corrosion level, the test was terminated when either of main difference between NC and SCC mixes was the coarse aggre-
the two beams reached 30% theoretical rebar/stirrup mass loss. gate content; SCC had a lower coarse aggregate content (900 kg/
The corrosion investigation was conducted under accelerated m3) compared to NC (1130 kg/m3). In addition, SCC mixes had dif-
corrosion testing and the corrosion rates were monitored by mea- ferent high range water reducers while NC had none. Details of SCC
suring the current passing with time. The corrosion characteristics and NC mix proportions are presented in Table 1.
along the beam length/perimeter was studied at different locations Type GU Canadian cement similar to ASTM Type I and slag ce-
by taking periodical half-cell potential measurements during the ment were used as cementitious materials for both SCC and NC
 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38 31

Table 1
Mixture proportions for SCC and NC mixtures.

Concrete type Type GU cement Slag cement 10-mm coarse aggregate Fine aggregate Water HRWR mL/100 kg of binder WR mL/100 kg of binder
(kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3)
SCC 315 135 900 930 180 Variable to obtain similar slump flow 0
NC 300 100 1130 725 160 0 300

mixtures. The chemical and physical properties of cement and slag for cylinders and the bottom reinforcement of the beams was
are presented in Tables 2 and 3, respectively. 25 mm with an average yield strength of 480 MPa and an average
Natural sand and 10 mm maximum size stone were used as a tensile strength of 725 MPa.
fine and coarse aggregates, respectively. High range water reducers
similar to Type F of ASTM C 494 [16] and a water reducer (WR) 3.3. Casting of beam/cylinder specimens
similar to Type A of ASTM C 494 [16] were used to adjust the flow-
ability and cohesiveness of SCC and NC mixtures, respectively. The two concrete mixtures (SCC and NC) used in the first stage
Table 4 presents the fresh properties and the strength of NC and were delivered to Ryerson University Structures Laboratory in six-
SCC mixtures. cubic-meter trucks by Dufferin Concrete Group, Toronto, Canada.
The traditional slump test, according to ASTM C 143/C 143/M The delivered SCC mixture was very similar to that successfully
[17], was conducted for the NC mixture. The slump flow test [18] used in the Pearson International Airport project in Toronto in
was conducted to evaluate the viscosity and flowability of SCC 2000 [23].
mixture while the V-funnel [19] and L-box [20] tests were con- Immediately after concrete delivery, concrete fresh properties
ducted to evaluate the stability and the passing ability of SCC, tests were carried out and beams were cast in prepared wooden
respectively. 100  200 mm control cylinders were used to deter- forms. SCC beams were cast without consolidation; the concrete
mine the compressive strength (fc0 ) and the indirect tensile strength was poured in the formwork from one end until it flowed and
(fct0 ) as per ASTM standards C 39/C 39 M and C 39/C 39 M [21,22] reached the other end. Visual observation showed that the SCC
for both NC and SCC mixtures. The diameter of deformed bars used properly filled the forms with ease of movement around reinforc-
ing bars. On the other hand, NC beams were consolidated using
electrical vibrators, and trowel-finished for smooth top surfaces.
Table 2 The placement of NC beams was labor-intensive and the time re-
Chemical and physical properties of cement. quired to cast and finish each beam element was much longer than
Chemical analysis (%) Physical analysis
that required for SCC beams. Formworks were removed after 24 h
of casting and the beams were moist-cured for 4 days, then air
SiO2 19.64 Residue45 lm (%) 8.42
Al2O3 5.48 Blaine fineness (m2/kg) 410
cured until the date of testing.
Fe2O3 2.38 Air content (%) 7.78 The four concrete mixtures used in the second stage were cast
CaO 62.44 Initial set (min) 103 in the laboratory using a rotating planetary mixer. The concrete
MgO 2.48 Auto. expansion (%) 0.14 materials were delivered from the same companies as in the first
SO3 4.32 Sulf. expansion (%) (prev. month) 0.013
stage to ensure fair comparison between the two stages. The NC
Total alkali 0.97 – –
Free lime 1.03 Compressive strength (MPa) – mixture was compacted using a vibrating table, while the SCC mix-
LOI 2.05 1 day 19.23 tures did not have any compaction. The concrete cylinders were
C3S 52.34 3 days 29.12 cured in a similar manner as those in the first stage.
C2S 16.83 7 days 33.82
C3A 10.50 28 days 41.45
C4AF 7.24 – –
3.4. Accelerated corrosion setup and current measurements

Accelerated corrosion tests have been used successfully to

determine the susceptibility of reinforcing and other forms of
structural steel to corrosion [24]. The accelerated corrosion setup
Table 3
Chemical and physical properties of slag cement. used in this investigation consisted of plastic tanks, electrolytic
solution (5% sodium chloride (NaC1) by the weight of water) and
Chemical analysis (%) Physical analysis
steel mesh placed at the bottom of each tank. Two tanks
SiO2 40.3 Residue 45 lm (%) 3.14 (2800 mm in length, 1500 mm in width and 500 mm in depth)
Al2O3 8.4 Blaine fineness (m2/kg) 422
were used to test the four concrete beams, while the small con-
Fe2O3 0.5 – –
CaO 38.71 Compressive strength (MPa) – crete cylinders were placed in a smaller (700 mm length,
MgO 11.06 (50:50 cement:slag) – 700 mm width and 500 mm depth) tank. Each beam was partially
SO3 0.57 7 days 23.75 immersed in the electrolytic solution up to one-third of its height
K2O 0.37 28 days 41.96 (to avoid the corrosion of the top longitudinal bars). The small cyl-
Na2O – – –
LOI 0.65 Slag activity index (% of 28 days control) 99.9
inders were immersed up to two-thirds of their total height. To
eliminate any variance in the concentration of the NaC1 and pH

Table 4
Fresh and hardened properties of SCC and NC mixtures.

Concrete type Slump (mm) Slump flow V-funnel flow time (s) L-box 28-d fc0 (MPa) 28-d fct0 (MPa)
Flow (mm) T500 (s) Index (%) T200 (s) T400 (s)
SCC – 700 30 5.5 90 1.5 2 45 3.8
NC 80 – – – – – – 47 4

T500: flow time to achieve 500-mm slump flow; T200 and T400: times taken by concrete to travel a horizontal distance of 200 mm and 400 mm, respectively.
32 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38

of the solution, the electrolyte solution was changed on a weekly

basis.
The specimen’s steel bars and the bottom steel mesh in each
tank were connected to a 12 V DC power supply. The direction of
the current was arranged so that the steel mesh served as a cath-
ode while the specimen bars served as anodes (Fig. 1). The acceler-
ated corrosion test continued until two degrees of corrosion
occured. After the power supply was turned on, the current flowing
through the system was recorded at one-hour intervals using a
computer-controlled data-acquisition system (Fig. 2). Based on
Faraday’s Eq. (1), the amount of corrosion is related to the electrical
energy consumed and is a function of ampere and time. Therefore,
the amount of corrosion can be estimated by using Eq. (1) as
follows
tiM
Mass loss ¼ ð1Þ
zF
where t = the time passed (s), i = the current passed (amperes),
M = atomic weight (for iron M = 55.847 g/mol), z = ion charge (two
moles of electrons) and F = Faraday’s constant which is the amount
Fig. 2. Corrosion current monitoring by data-acquisition system.
of electrical charge in one mole of electron (F = 96,487).

3.5. Half-cell potential measurements

This method was first developed in the late 1950s [25] and was
adopted in 1977 as an ASTM method C876 [26]. It is based on mea-
suring the electrochemical potential of reinforcement against a ref-
erence CSE (copper–copper sulfate electrode) placed on the
concrete surface (Fig. 3). The instrument outputs a range of values;
more negative values indicate a higher probability of corrosion
while positive ones indicate a very low probability of corrosion.
Half-cell potential measurements were taken periodically at 25
different points/locations along the beam length/perimeter. Fig. 4
shows a typical layout of the tested points on each beam. Half-cell
potential measurements are usually affected by a number of fac-
tors [27], including the moisture condition of the concrete cover
and its contamination by carbonation and/or chlorides. The oxygen
access strongly determines the half-cell reading; low oxygen con-
tent results from a wet surface show low potentials and give higher
negative values. Therefore, bi-weekly readings were taken for all
beams at the same time and in the same moist surface conditions
in order to ensure fair comparison between all beams. Fig. 3. Half-cell potential measurements.

3.6. Chloride ion measurements

measure the chloride content of concrete powders in the extraction
Chloride ion (Cl) determination was conducted using a high solution as a percentage of concrete mass. Five standard samples
impedance electrometer with a calibrated chloride electrode to containing chloride ranging from 0.005% to 0.5% Cl were used to

Power Supply
12 VDC
Steel Bar Connected to
the Mesh (Cathode)

5% NaCl by
weight

Specimens (Anode)

Steel Mesh (Cathode) Plastic Tank

Fig. 1. Schematic representation of the accelerated corrosion test.

 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38 33

Bottom view of the beam

Front view of the beam

Back view of the beam

Fig. 4. Typical layout of the tested points along the beam length/perimeter.

calibrate the instrument before use. The concrete powder samples sure they were free of any adhering concrete or corrosion products,
were taken from the same locations near the bar surface using an then soaked in a chemical solution (1:1 of HCl and water) accord-
electrical drill (Fig. 4). The concrete powders were then weighed, ing to ASTM Standards G1-03 method [28]. The clean bars were
poured into small plastic vials containing extraction liquid, and left then weighed and the percentage mass loss for each bar was calcu-
overnight before testing. lated based on Eq. (2):

ðinitial weight  final weightÞ

3.7. Measurement of mass loss and reduction of bar diameter % of mass loss ¼  100 ð2Þ
initial weight
After the beams were corroded and the measurement of half- In order to investigate the reduction of bar diameter and mass
cell, chloride ion content and crack widths were taken, the beams loss along the length of the beam, each bottom longitudinal bar
were jackhammered to remove the corroded stirrups and longitu- was cut into five segments to measure mass loss and bar diameter
dinal bars. The corroded bars were cleaned with a wire brush to en- along each segment.

Fig. 5. Corrosion performance, crack pattern and widths (mm) in each tested beam.
34 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38

4. Experimental results and analysis 3.5

NC-N
3 SCC-N
4.1. General cracking observation
2.5

Current (A)
Fig. 5 shows the crack patterns and crack widths of each tested 2
beam after the completion of corrosion tests. SCC beams exhibited
less cracking in terms of number and width than NC beams for 1.5

both moderate and severe corrosion levels. For example, the SCC- 1
N beam did not have any longitudinal cracking and had minimal
0.5
transverse cracking at the stirrup location, while the NC-N beam
had a number of longitudinal and transverse cracks. Also, the aver- 0
age crack width for the SCC-E beam was around 0.4 mm, while the 0 4 8 12 16 20 24 28 32 36 40 44 48
Time (Days)
average crack width for the NC-E beam was around 2.4 mm. This
result can be attributed to the higher durability and the superior Fig. 6. Current–time history for non-epoxy-coated stirrup beams in moderate
performance of SCC mixture in rebar corrosion resistance due to corrosion level.
its dense and enhanced microstructure.
Fig. 5 also shows that the cracking patterns and widths were not
uniformly distributed in SCC beams as they were in NC beams.
9
SCC-N had one broken part at the corner located farthest away NC-E
from the casting point, while NC-N had uniform cracking patterns 8 SCC-E
and widths along the beam length. Similarly, SCC-E had a big 7
spalled cover at the corner of the beam, which was located far 6

Current (A)
away from the casting point, and the crack width at the bottom 5
of the beam was increasing towards the direction of that corner. 4
Meanwhile, NC-E showed a uniform cracking pattern and had no
3
spalled parts along the beam length. This difference can be attrib-
2
uted to the fact that SCC beams were cast from one end, allowing
the concrete to reach the other end under its own weight. The 1

weight of the falling concrete during casting was enough to com- 0

0 8 16 24 32 40 48 56 64 72 80 88 96 104
pact the corner of the beam below the casting point, but not en- Time (Days)
ough to compact it well at the far end. Corners of beams usually
require enough compaction due to the restraint of concrete move- Fig. 7. Current–time history for epoxy-coated stirrup beams in severe corrosion
ment from the formworks in three sides. Therefore, SCC beams level.

showed weak and porous concrete at the corners located far away
from the casting points.
The results also indicated that SCC beams were easier to break period. The lower current passing through the concrete specimens
at the aforementioned corners compared to NC beams (which did is an indication of the higher resistivity of the concrete. Permeabil-
not break even at corners, with higher crack widths). The reason ity of the concrete is the main factor influencing the concrete resis-
could be attributed to the fact that the NC mixture contained tivity [30]. The high flowability and superior resistance to bleeding
25% more coarse aggregate content than the SCC mixture. The and segregation of SCC beams were thought to be the main factors
higher volume of coarse aggregate in the NC mixture caused higher that improved the permeability and enhanced the quality of con-
crack arresting and prevented the concrete from spalling even at crete, especially below longitudinal bars.
corners with higher crack widths. The point of first increase of the slope in the time–current curve
indicates the corrosion initiation, and the slope of the curve repre-
sents the rate of corrosion. NC beams showed earlier corrosion ini-
4.2. Time-dependent corrosion tests results tiation and a higher corrosion rate than SCC beams. The corrosion
initiation in NC beams started after 3 and 9 days in NC-N and NC-E,
4.2.1. Current results respectively, compared to 13 and 30 days in SCC-N and SCC-E
During the accelerated corrosion test, the corrosion rate was beams. The increase of the corrosion initiation time in epoxy-
monitored by a computer-controlled data-acquisition system coated compared to non-epoxy-coated stirrup beams was related
which recorded the current at 1-h intervals. Figs. 6 and 7 show to the stirrup corrosion, which occurred before longitudinal bar
the relationship between the current in mA and the immersion corrosion due to lower cover thickness.
time in days for SCC/NC beams with epoxy- and non-epoxy-coated SCC beams exhibited sudden jump in the time–current curve
stirrups tested for severe and moderate corrosion levels, respec- after relatively steady slope compared to NC beams which
tively. In general, the current–immersion time relationship for all showed relatively gradual increase of the current with the time.
beams showed an initial decrease in the current, followed by a The sudden jump in SCC time–current curves was the indication
gradual increase. The decrease of the current in the first few days of the concrete cover spalling in SCC beams at the end corners
is an indication of the formation of the passive film around the (Fig. 5).
reinforcing steel bar, which protects the steel from corrosion.
When depassivation of the steel occurs, corrosion starts and then 4.2.2. Half-cell potential measurements
the rate of corrosion increases significantly [29]. Fig. 8 presents the variation of average cross-section half-cell
NC beams demonstrated higher current values in the early potential readings along the length of the beams with age. NC
stages of the test, approximately 1.8 and 0.9 mA in NC-N and NC- showed more negative values compared to SCC in both epoxy-
E, respectively, compared to SCC beams which demonstrated 0.6 and non-epoxy-coated stirrup beams (moderate and severe corro-
and 0.4 mA in SCC-N and SCC-E, respectively. Also, the current in sion levels). The more negative values in NC beams are an indica-
NC beams was higher than that of SCC beams during the whole test tion of the higher probability of corrosion for NC beams compared
 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38 35

0.7 Time (Days)

Point 1
0.6 Point 2 0
Point 3
Half cell potential (V,CSE)
0.5 0 10 20 30 40 50
Point 4

Half cell potential (V,CSE)

0.4 Point 5 -0.1
Point 1
0.3 Point 2
-0.2 Point 3
0.2 Point 4
Point 5
0.1 -0.3
0
-0.1 0 10 20 30 40 50 -0.4
-0.2
Time (Days) -0.5
-0.3
-0.4 SCC-N -0.6 NC-N

0.8 Time (Days)

Point 1
Point 2 0
0.6 Point 3 -0.05 0 20 40 60 80 100
Half cell potential (V,CSE)

Half cell potential (V,CSE)

Point 4
0.4 Point 5 -0.1 Point 1
Point 2
-0.15 Point 3
0.2 Point 4
-0.2 Point 5
0
-0.25
0 20 40 60 80 100
-0.2 -0.3

-0.4 Time (Days) -0.35

-0.4
-0.6
SCC-E -0.45 NC-E

Fig. 8. Average cross-section half-cell reading at each point along the beam length.

to SCC beams. This expectation was verified and confirmed at the 4.3. Test results after corrosion
end of the test by the other tests’ results.
Both SCC and NC beams showed no significant differences in the 4.3.1. Results of chloride content, mass loss and rebar’s diameter
half-cell readings around the beam cross-sections as it showed reduction
along the beam length (Fig. 8). In addition, the readings along the Each beam was checked at 25 points/locations along the beam’s
length of the beam for both SCC and NC were dependent on the length/perimeter (Fig. 4). After the completion of the two corrosion
cover thickness, regardless if the steel underneath was epoxy- or levels, the chloride ion content, measurements of crack widths, re-
non-epoxy-coated. For example, points 1, 3 and 5 located below bar mass loss and the reduction of the rebar diameter were taken
stirrups (on any beam, either epoxy- or non-epoxy-coated) that at each location.
had a 30 mm cover thickness showed more negative values than Fig. 9 presents the results of the chloride ion content, crack
points 2 and 4, which were not at stirrup location and had a width, mass loss and diameter reduction at each point located on
40 mm cover thickness. This result matched other researchers’ re- the beam length/perimeter. In general, the chloride ion contents
sults [27] indicating that the concrete cover is a big factor in deter- near the longitudinal bar surface at all the points on NC-N and
mining the half-cell potential reading. Also, point 3, located at the NC-E beams were higher compared to those of SCC-N and SCC-E
middle of any beam, showed less negative value compared to tested points, respectively. This result confirms the finding of the
points 1 and 5. This is because point 3 was located below only half-cell and the current monitoring results, which indicated the
one stirrup compared to points 1 and 5, which were below three superior performance of the SCC mixture in protecting the steel
stirrups. bars from corrosion. The chloride ion content was also increased
In both epoxy-(E) and non-epoxy-(N) coated stirrup beams, the with the increase of the corrosion level (Fig. 9). The corrosion level
readings along the length of SCC beams were varied based on the was indicated by the rebar mass loss or diameter reduction, or by
distance from the casting point, while the readings along the the crack width located close to the tested points. This finding is
length of NC beams did not show any significant differences be- similar to that found by other researchers [31,32]. It is important
tween the two ends of the beam (Fig. 8). The curves for points 1 to note that, in both NC-N and SCC-N, the chloride ion content
and 5 in NC-N and NC-E beams were close to each other, indicating was higher at points located below stirrups than at points located
no significant differences between the two ends of the beam. On below longitudinal bars where the concrete cover is bigger. This re-
the other hand, the curves for these two points shifted away in sult also confirms other researchers’ results [32,33] which indicate
the SCC-N and SCC-E beams, indicating a significant difference be- that the chloride concentration decreases from the concrete sur-
tween the two beam ends. This result indicates an inferior quality face to the interior in the vicinity of the steel surface.
of concrete and a higher probability of corrosion at points far away Fig. 9 revealed the lesser quality of SCC beams at points located
from casting points in SCC beams. far away from the casting point (point 1) compared to better qual-
Fig. 8 also shows that the reading of the half-cell potential ity of concrete below the casting point (point 5). Chloride ion con-
test presented closed values at late testing age (high degree of tent, crack width, rebar mass loss and reduction of bar diameter
corrosion) compared to early testing age (initial corrosion stage). were high at the beam’s corner located far away from the casting
This result indicates that the half-cell potential test used can point, compared to that at the corner located below the casting
only represent the probability of corrosion for uncorroded beams, point (Fig. 9). This result explained the spalling of the concrete cov-
but may not give good indication for beams that are already er at the mentioned corners and matched the results of the half-
corroded. cell potential test during the corrosion-time monitoring.
36 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38

1.60 NC-E 0.70 SCC-E

1.40 External Bar 0.60 External Bar
1.20 0.50
1.00
0.40
0.80
0.30
0.60
0.40 0.20

0.20 0.10
0.00 0.00
1-1&1-2 2-1&2-2 3-1&3-2 4-1&4-2 5-1&5-2 1-1&1-2 2-1&2-2 3-1&3-2 4-1&4-2 5-1&5-2
Point Number Point Number

0.35 NC-E 0.40 SCC-E

0.30 Internal Bar 0.35 Internal Bar
0.30
0.25
0.25
0.20
0.20
0.15
0.15
0.10 0.10
0.05 0.05
0.00 0.00
1-3 2-3 3-3 4-3 5-3 1-3 2-3 3-3 4-3 5-3
Point Number Point Number

0.60 NC-E 0.40 SCC-E

External Bar 0.35 External Bar
0.50
0.30
0.40
0.25
0.30 0.20

0.20 0.15
0.10
0.10
0.05
0.00 0.00
1-4&1-5 2-4&2-5 3-4&3-5 4-4&4-5 5-4&5-5 1-4&1-5 2-4&2-5 3-4&3-5 4-4&4-5 5-4&5-5
Point Number Point Number

Chloride Content (%)

Mass Loss (Kg)

Maximum Total Crack Widths (mm)

Reduction of Bar Diameter (% of original diameter)

Fig. 9. Chloride content, rebar mass loss, maximum crack width and the reduction of bar diameter along external and internal bars of NC-E and SCC-E.

SCC-N demonstrated no longitudinal cracks, rebar mass loss or calculated mass loss was compared with the actual mass loss for
diameter reduction at any of its points compared to NC-N, which each of the tested beams after computing the total actual metal
showed a number of cracks and rebar mass loss and diameter loss in the longitudinal bars and stirrups (Fig. 10). The results show
reduction. Also, SCC-E had lower crack widths and less rebar mass
loss and diameter reduction in most of its points compared to NC-
E. This is another indication of the superior performance of SCC in 9000
rebar corrosion protection.
8000 Theoritical mass loss
The crack width increased with the increase of the corrosion le-
7000
vel in all beam types (Fig. 9). The corrosion products accumulated Actual mass loss
Mass loss (gm)

around the bar surface occupied more space and exerted pressure 6000
on the concrete cover, causing cracking. The crack widths increased 5000
with the increase of rebar mass loss, except for the middle part of 4000
the external bar of NC-E, which showed lower crack width at rela-
3000
tively high mass loss (Fig. 9). This is due to the fact that the entire
2000
bar was corroded and diminished at this point.
1000

4.4. Comparison of theoretical and actual corrosion mass loss 0

NC-N SCC-N NC-E SCC-E
Beam type
The corrosion mass loss was computed using Faraday’s Eq. (1)
based on the amount electrical energy sent through the bar. The Fig. 10. Comparison of theoretical and actual mass loss in all tested beams.
 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38 37

that the actual mass loss was less than the theoretical mass loss for concrete confinement to embedded bar was ensured compared to
all tested beams. The percentages of actual to theoretical mass the bottom longitudinal bars in the full-scale beams. This is why
losses were 97%, 91%, 83% and 74% in NC-N, SCC-N, NC-E and the small concrete cylinders did not manifest the difference be-
SCC-E, respectively. As indicated in previous studies [34,35], when tween SCC and NC in rebar corrosion protection. In addition, the
a current passes through a bar suspended in salt solution, the cor- close results of the corrosion performance in all SCC mixtures indi-
relation between actual and predicted mass loss is almost perfect, cate that the HRWR types do not have any chemical effect on cor-
while for bars embedded in concrete, the mass loss based on Fara- rosion protection.
day’s law overestimates the actual mass. This is attributed to the
fact that some of the passing currents do not contribute to corro-
sion but are consumed while passing through the concrete cover. 5. Conclusions
The results also indicated that the percentage of actual to theo-
retical mass loss was higher in NC compared to SCC for all tested The corrosion performance and cracking behavior of self-con-
beams. NC beams corroded faster and developed earlier cracks solidating concrete (SCC) were described and compared with nor-
than SCC beams; these cracks decrease the concrete resistance, mal concrete (NC) based on test results of full-scale experimental
resulting in corrosion that is closer to predicted levels [34,35]. beams and small-scale concrete cylinders. The current measure-
ments, half-cell potential readings, crack pattern and widths, chlo-
ride ion content, rebar mass loss and diameter reduction were
4.5. Results of the small concrete cylinder specimens
critically analyzed to study the influence of SCC mixtures in rebar
corrosion protection and durability. Based on the results presented
After evaluating the results of the full-scale beams and confirm-
in this paper, the following conclusions were warranted:
ing the superior performance of SCC compared to NC, it was essen-
tial to examine their performance in small concrete cylinders to
 Based on overall performance of the full-scale tested beams, SCC
manifest the effect of bleeding and segregation on durability deg-
mixture exhibited superior rebar corrosion protection compared
radation, and to investigate the effect of different types of HRWR
to its NC counterpart. Distinct advantages of SCC over NC in
(used in SCC) on the protection of rebar corrosion.
terms of corrosion protection were revealed from the results
The accelerated corrosion test for the concrete cylinders was
of current measurements with time, crack widths and patterns,
terminated after 25 days, when all samples indicated similar corro-
half-cell potential measurements, chloride ion contents near the
sion behavior as observed from the crack widths and the current
bar surface and the rebar mass loss/diameter reduction.
measurement history. Fig. 11 demonstrates one corroded concrete
 The cracks in SCC beams were easily propagated and extended
cylinder from each concrete type. As concluded from the current–
compared to NC beams. SCC beams exhibited breaking and
time measurements, the corrosion initiation time in all concrete
spalling of concrete cover, even at locations which had lower
types was very close. Also, the rate of corrosion after corrosion ini-
crack widths compared to NC beams. This inferior quality could
tiation was similar in all four concrete mixes. The corrosion initia-
be attributed to the presence of a lower volume of coarse aggre-
tion time in SCC and NC mixes used in the full-scale concrete
gate in SCC beams (25% less than NC), causing lower crack-
beams (first stage) commenced after 13 days of testing, while the
arresting capacity that induces concrete spalling even at loca-
other two SCC mixes exhibited corrosion initiation time 2 and 3
tions with lower crack widths.
days earlier. The crack pattern and crack widths were also similar
 The SCC mixture showed non-uniform concrete properties along
in the four concrete mixes. All concrete cylinders exhibited one
the length of the full-scale concrete beams when casting
longitudinal crack of a 1 mm width along the length of the embed-
occurred from one end, causing lesser quality concrete at the
ded bar (Fig. 11).
far end due to improper compaction and distribution. As a result,
As expected, bleeding and segregation (associated with large
severe corrosion and spalling of concrete cover was observed at
concrete volume and adopted casting/placing/vibrating tech-
corners located far away from the casting point. The results of
niques) were the factors affecting the concrete performance in
half-cell measurements, crack widths, chloride ion contents,
the full-scale concrete beams where SCC showed superior perfor-
rebar mass loss and rebar diameter reduction confirmed these
mance over NC. This was not observed in small cylinder specimens
findings. Therefore, when casting SCC beams, it is recommended
where the bleeding and segregation were minimized and better
that the casting point be moved along the beam length (partic-
ularly if the beam is long, shallow and narrow) to ensure uni-
form compaction, especially at corners.
 Strong correlation between the predicted rebar mass loss by Far-
aday’s equation and experiments suggests that the theoretical
estimates can be used to examine the effect of corrosion over
time.
 The types of admixture used in SCC mixes have no effect on cor-
rosion performance in terms of corrosion initiation, corrosion
rate and crack patterns and widths.

In terms of corrosion performance, the difference between SCC

and NC mixes was only pronounced in large-scale concrete beams.
No such difference was observed in small-scale cylinder speci-
mens. This difference is due to the fact that the effect of bleeding
and segregation (which is less in SCC) was greatly reduced in
small-scale cylinders. In addition, the casting technique in small-
scale cylinders yields good quality of concrete, especially around
the embedded steel bars. This is in comparison to the weaker, more
porous layer of concrete (enhanced when using SCC) below the
Fig. 11. Corrosion performance in the small concrete cylinders. longitudinal bars in the large-scale beams.
38 A.A.A. Hassan et al. / Cement & Concrete Composites 31 (2009) 29–38

Acknowledgements [16] ASTM C 494-99a. Standard specification for chemical admixtures for concrete.
American Society for Testing and Materials; 2001.
[17] ASTM C 143. Standard test method for slump of hydraulic-cement concrete.
The authors gratefully acknowledge the financial assistance of American Society for Testing and Materials; 2001.
the Natural Sciences and Engineering Research Council (NSERC) [18] Nagataki S, Fujiwara H. Self-compacting property of highly flowable concrete.
ACI SP 154. Las Vegas (USA): ACI; 1995. p. 301–314.
of Canada and the Canada Research Chair Program. Thanks to Mr.
[19] Ozawa K, Sakata N, Iwai M. Evaluation of self-compactibility of fresh concrete
John Pontarollo and Mr. Dennis Baker of St. Lawrence Cement, using funnel test. Proc JSCE 1994;490:V-23.
Canada, for their great support. [20] Sonebi M, Batros P, Zhu W, Gibbs J, Tamimi A. Properties of hardened concrete,
task 4, final report (2000). Advance Concrete Masonry Center, University of
Paisley, Scotland, UK; 2000. p. 6–73.
References [21] ASTM C 39. Standard test method for compressive strength of cylindrical
concrete specimens. American Society for Testing and Materials; 2001.
[1] Khayat KH, Manai K, Trudel A. In situ mechanical properties of wall elements [22] ASTM C 496-96. Standard test method for splitting tensile strength of
cast using self-consolidating concrete. ACI Mater J 1997;94(6):491–500. cylindrical concrete specimens. American Society for Testing and Materials;
[2] Khayat KH, Tremblay S, Paultre P. Structural response of self-consolidating 2001.
concrete columns. In: Skarendahl A, Petersson O, editors. Proceedings of the [23] Lessard M, Talbot C, Phelan WS, Baker D. Self-consolidating concrete solves
RILEM symposium on self-compacting concrete. Stockholm: RILEM challenging placement problems at the Pearson International Airport in
Publications S.A.R.L.; 1999. p. 291–306. Toronto, Canada. In: First North American Conference on the design and use
[3] Sonebi M, Tamimi AK, Bartos PJM. Performance and cracking behavior of of self-consolidating concrete (SCC), Rosemont, Illinois; 2002. p. 12–13.
reinforced beams cast with self-consolidating concrete. ACI Mater J [24] Amleh L. Bond deterioration of reinforcing steel in concrete due to corrosion. A
2003;100(6):492–500. thesis submitted to the Faculty of Graduate Studies and Research in partial
[4] Zhu W, Gibbs JC, Bartos PJM. Uniformity of in situ properties of self- fulfillment of requirements for the degree of Doctor of Philosophy, McGill
compacting concrete in full scale structural elements. Cem Concr Compos University, Montreal, Canada; 2000.
2001:57–64. [25] Stratfull RF. The corrosion of steel in a reinforced concrete bridge. Corrosion
[5] Hussain S, Rasheeduzzafar. Corrosion resistance performance of fly ash 1957:173–9.
blended cement concrete. ACI Mater J 1994;91(3):264–72. [26] ASTM C 876. Standard test method for half-cell potentials of uncoated
[6] Al-Amoudi O, Rasheeduzzafar, Maslehuddin M, Al-Mana A. Prediction of long- reinforcing steel in concrete. Annual Book of ASTM Standards, American
term corrosion resistance of plain and blended cement concretes. ACI Mater J Society for Testing and Materials; 1991.
1993;90(6):564–70. [27] Klinghoffer O. In situ monitoring of reinforcement corrosion by means of
[7] Rasheeduzzafar, Al-Saadoun SS, Al-Gahtani AS. Reinforcement corrosion- electrochemical methods. Nordic seminar, February 1995.
resisting characteristics of silica-fume blended-cement concrete. ACI Mater J [28] ASTM G1. Standard practice for preparing, cleaning, and evaluating corrosion
1992;89(4):337–44. test specimens. American Society for Testing and Materials; 2003.
[8] Haque MN, Kayyali OA. Aspects of chloride ion determination in concrete. ACI [29] Cornet I, Ishikawa T, Bresler B. The mechanism of steel corrosion in concrete
Mater J 1995;92(5):532–41. structure. Mater Prot 1968;7(3):44–7.
[9] McCurrich LH. Reduction in permeability and chloride diffusion with [30] Hope B, Alan K. Corrosion of steel in concrete made with slag cement. ACI
superplasticizers. Concrete 1986;20(8):9–10. Mater J 1987;84(5):525–31.
[10] Khayat KH. Optimization and performance of air-entrained, self-consolidating [31] Rasheeduzzafar, Ehtesham HS, Al-Saadoun SS. Effect of tricalcium aluminate
concrete. ACI Mater J 2000;97(5):526–35. content of cement on chloride binding and corrosion of reinforcing steel in
[11] Zhu W, Bartos P. Permeation properties of self-compacting concrete. Cem concrete. ACI Mater J 1992;89(1):3–13.
Concr Res 2003:921–6. [32] Amleh L, Mirza S. Corrosion influence on bond between steel and concrete. ACI
[12] Persson B. Internal frost resistance and salt frost scaling of self-compacting Struct J 1999;96(3):415–23.
concrete. Cem Concr Res 2003;33:373–9. [33] Montemor MF, Cunha MP, Ferreira MG, Simoes AM. Corrosion behavior of
[13] Nehdi M, Pardhan M, Koshowski S. Durability of self-consolidating concrete rebars in fly ash mortar exposed to carbon dioxide and chlorides. Cem Concr
incorporating high volume replacement composite cement. Cem Concr Res Compos 2002(24):45–53.
2004:2103–12. [34] Auyeung Y, Balaguru P, Chung L. Bond behavior of corroded reinforcement
[14] Hoshino M. Relationships between bleeding, coarse aggregate, and specimen bars. ACI Mater J 2000;97(2):214–20.
height of concrete. ACI Mater J 1989;86(2):185–90. [35] Spainhour LK, Wootton IA. Corrosion process and abatement in reinforced
[15] Park R, Paulay T. Reinforced concrete structure. Ultimate strength design of concrete wrapped by fiber reinforced polymer. Cem Concr Compos
reinforced concrete structures, printed by the University of Canterbury for 2008;30(6):535–43.
extension study seminars conducted for practicing structural engineers in
New Zealand; 1975:V-1.