Construction and Building Materials 126 (2016) 875–885

Contents lists available at ScienceDirect

Construction and Building Materials

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

Using carbon fiber composites for reinforcing pervious concrete

Milena Rangelov, Somayeh Nassiri ⇑, Liv Haselbach, Karl Englund
Washington State University (WSU), United States

h i g h l i g h t s

Cured carbon fiber composite material (CCFCM) pieces were mechanically refined and processed to certain shape and size fraction for incorporation in
pervious concrete (PC).

CCFCM incorporation resulted in improved workability as evident in lower and more consistent porosity of the reinforced PC (rPC), compared to the
control (plain PC).

Incorporation of CCFCM yielded in improved mechanical properties of PC.

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

Article history: Pervious concrete (PC) pavement applications are growing in popularity due to the environmental
Received 14 March 2016 and stormwater management benefits that PC can offer. However, relatively low mechanical prop-
Received in revised form 10 June 2016 erties and durability of PC, comparing to conventional Portland cement concrete (PCC), limit its use
Accepted 13 June 2016
for vehicular applications. In this study, different size fractions of cured carbon fiber composite
material (CCFCM) pieces were incorporated into a PC mixture (rPC) in three volume fractions.
The goal was to determine the physical and mechanical properties of rPC in comparison to the
Keywords:
corresponding properties of plain PC (control). Seven mixture designs were prepared in order to
Pervious concrete
Carbon fiber composite
investigate the effect of CCFCM volume fractions as well as CCFCM particle sizes. The test results
Mass loss indicated that CCFCM addition enhanced the workability of the PC mixtures. rPC mixtures presented
Cantabro higher average infiltration rates when compared to the control mixture. Improvements in mechan-
Surface abrasion ical properties were seen on 28-day compressive strength (4–11%), 7-day tensile strength (11–46%)
Compressive strength and in modulus of elasticity (6–45%). In terms of resistance to mass loss in Cantabro and surface
abrasion, rPC mixtures presented various behaviors, with one mixture containing the highest
volume fraction and a combined size fraction, outperforming the control in both tests. Overall,
the results of this study indicate that incorporation of CCFCM is promising in improving physical
and mechanical properties of PC.
Ó 2016 Published by Elsevier Ltd.

1. Introduction restoration and stormwater runoff reduction. When used as a

Pervious concrete (PC) pavement is one pavement out of the
suite of permeable pavements that simultaneously serves
stormwater runoff management and supporting vehicular or
pedestrian traffic. PC is growing in popularity among municipali-
ties and transportation agencies for applications such as bike lanes,
pedestrian walkways, sidewalks, parking lots, low-volume road-
ways and others. The increased application is mainly due to PC’s
environmental benefits, such as underground water system
⇑ Corresponding author at: Washington State University, 248 PACCAR Environ-
mental Technology Building, 2001 East Grimes Way, Pullman, WA 99164-5815,
United States.
E-mail address: snassiri@wsu.edu (S. Nassiri).
pavement surface course, PC can also mitigate traffic noise [1]
and potentially reduce the heat island effect [2,3].
PC is prepared with similar ingredients as conventional Port-
land cement concrete (PCC), however with none or small amounts
of fine aggregate and a lower amount of water. The elimination of
fines together with the gap (open) gradation of coarse aggregate,
typical for PC, provides an interconnected system of air voids
(15–35 percent) that facilitates the flow of water [4]. Due to the
lack of fine aggregate, the coarse aggregate grains in PC are
bounded solely by a thin layer of cement paste, resulting in lower
mechanical properties of PC comparing to conventional PCC, where
coarse aggregate is embedded in the matrix. Typical values of 28-
day compressive strength for PC range from 2.8 to 28 MPa [4,5]
as opposed to 20–40 MPa for traditional PCC.

http://dx.doi.org/10.1016/j.conbuildmat.2016.06.035
0950-0618/Ó 2016 Published by Elsevier Ltd.
876 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885

Despite the gain in popularity, PC is relatively young in the Table 1

application, research, and development. Current initiatives for PC Mixture composition and proportioning.

improving PC properties include enhanced surface resistance to Material Amount per m3 of PC

abrasion, increased mechanical properties, improved freeze-thaw
(F-T) durability and the resistance to deicer salts when used in
regions with cold climate, as well as better workability for more
consistent placing at different environmental conditions. Lately,
numerous improvements have been introduced to the PC mixture
design to address the above-stated objectives by the addition of
various reinforcing elements or modifying additives. Examples of
studies that focused on improving the mechanical properties
include the implementation of various fibers in PC, which resulted
in increases in compressive, tensile and flexural strength [6–8].
Rehder et al. (2014) [9] reported that polypropylene fibers intro-
duced improvements in residual flexural strength of PC.
In terms of durability, the addition of polypropylene fibers was
found to enhance the F-T durability of PC [10,11]. The latter study
showed that 56-mm long fibers at the dosage of 3 kg/m3 were par-
ticularly beneficial in terms of increasing F-T durability [11]. Other
studies focused on improving durability to abrasion and reported
that the use of polypropylene [11,12] and cellulose fibers [8]
increased the surface abrasion resistance of PC. It was also found
in another study that the addition of styrene butadiene rubber
(SBR) latex to the PC mixture can improve compressive strength
and abrasion resistance [12]. According to a study by Yang
(2011), partial replacement of cement with silica fume tends to
increase the F-T durability of PC, as well as the resistance against
deicer salt [10]. Further, the use of sand as a partial replacement
of coarse aggregate (five to 10 percent by the mass) was shown
to improve the F-T durability of PC [13].
In this study, cured carbon fiber composite materials (CCFCM)
in various particles size fractions and volume fractions were incor-
porated into the PC mixture to potentially improve various proper-
ties. Currently, there are limited opportunities to reutilize excess
material and the end-of-life products generated during the produc-
tion of synthetic composites, such as carbon fiber composites
(CFC). One procedure is to isolate the fibers from the composite
by a chemical solvent or through pyrolisis. However, costs associ-
ated with these thermal or chemical processes can be significant
and result in a low-quality and high-price carbon fiber product.
CCFCM used in this study were processed using a low-energy
mechanical procedure, aiming to preserve the original composite
structure of fiber materials and cured resin and thereby the prop-
erties of the original CCFCM.
The main objective of this study was to evaluate the feasibility
of incorporating refined CCFCM elements into PC mixtures (here-
after reinforced PC (rPC)), while examining mixability, workability,
and compaction of the PC. As part of the study, the physical,
hydraulic and mechanical properties of rPC were also established
and compared to the corresponding properties of plain PC.
2. Materials
SSD coarse aggregate [kg] 1632.0
Fine aggregate [kg] 0.0
Water [kg] 95.7
Type I/II cement [kg] 337.2
Class F fly ash [kg] 59.5
Admixture [ml] 3888.0
CCFCM elements [kg] 8.0 (0.5% volume)
15.6 (1% volume)
23.4 (1.5% volume)
delay the setting, providing more workability time. Implemented
admixture dosage was determined based on the producer’s speci-
fications as 980 ml per 100 kg of cementitious material [14]. PC
mixture was designed following the mixture design procedure
available in ACI 522-R-10 [4], using a target porosity of 27 percent.
The final PC mixture design and the proportioning of the PC con-
stituents are provided in Table 1.
2.2. Processing of cured carbon fiber composite materials
CCFCM were processed at the Composite Materials and Engi-
neering Center (CMEC) at Washington State University (WSU).
The materials were first shredded and then hammer-milled
through a 25.4-mm screen to separate the coarsest particles. Sub-
sequently, the CCFCM elements were differentiated into four frac-
tions by further mechanical screening: (C) combined: particles
smaller than 3.35 mm, (L) large: particles smaller than 3.35 mm
and larger than 2 mm, (M) medium: particles smaller than 2 mm
and larger than 0.841 mm, and (S) small: particles smaller than
0.841 mm (retained on the pan). Particle size distribution is pre-
sented in Fig. 1. Fig. 2 presents four different CCFCM fractions. As
seen in Fig. 2, coarse and flaky CCFCM elements were contained
in C and L fractions, while the S and M mainly contained particles
in the form of fibers. It is noteworthy that CCFCM elements
obtained by hammer-milling cannot be referred to as ‘‘carbon
fibers”, but rather as carbon fiber composite elements. Each ‘‘ele-
ment” contains a dense network of carbon fibers bonded by the
thermoset resin used in the original composite.
3. Experimental design
An experiment was designed to investigate the effect of differ-
ent CCFCM volume fractions, as well as different CCFCM element
particle size fractions on PC properties. Therefore, the experiment
included one control PC mixture, three rPC mixtures containing
100
90
80
Cumulave passing [%]

2.1. Pervious concrete mixture design 70

60
C < 3.35 mm
PC mixtures were prepared with Type I/II ordinary Portland 50
cement, and crushed basalt coarse aggregate. Coarse aggregate 40 2mm < L < 3.35 mm
had a nominal maximum size of 9.5 mm, the specific gravity of 30 0.841mm < M < 2mm
3.102, and 3.11 percent water absorption. Coarse aggregate was 20
S <0.841mm
incorporated into the mixture in saturated surface dry (SSD) condi-
10
tion. No fine aggregate was included in the PC mixture. Fifteen per-
0
cent of the Portland cement by mass was replaced with Type F fly 0. 0 1 0. 1 1 10
ash. Desirable workability was obtained with water to cementi- Sieve size [mm]
tious materials ratio (w/cm) of 0.24. Rheology-modifying chemical
admixture, V-MAR VSC500 from W.R. Grace & Co. was used to Fig. 1. Particle size distribution of CCFCM elements.
 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885 877

Fig. 3. CCFCM particles are distributed evenly throughout the fresh (left), and
hardened (right) PC mixture.

The compaction method for the cylinders was selected to result

in uniformly compacted specimens for strength testing, while the
slabs were compacted to represent field placement and com-
paction procedure. The amount of fresh PC placed in each specimen
mold was pre-determined according to the designed density. Small
and large cylindrical samples were filled with the required PC
amount in two and three lifts, respectively. Lower lifts were com-
pacted with 15 and 20 blows of a standard Proctor hammer, with
Fig. 2. Four different sizes of CCFCM elements included in the experiment: C-
2.5-kg weight and 30.5-cm drop distance, for small and large cylin-
combined, L-large, S-small, M-medium fraction.
ders, respectively. The final lift was placed by filling the mold to
the top and compacting with the needed number of Proctor ham-
Table 2 mer blows to fit the pre-determined mass of the mixture in the
Naming scheme used for the seven PC mixtures in the experiment. mold. Slab molds were filled with fresh PC in one lift and com-
Mixture ID CCFCM element size CCFCM element content pacted with about 33 blows of the standard Proctor hammer. Sub-
[% volume of the PC mixture] sequently, slabs were compacted using a hydraulic compression
Control 00 Not applicable Not applicable
testing machine, applying the load of about 3.1 kN, which corre-
C0.5 Combined 0.5 sponds to a Bunyan roller compaction used for compacting PC in
C1 Combined 1 the field. To make sure the mixture filled the mold consistently,
C1.5 Combined 1.5 the molds were hit with a plastic mallet on the side all around
S1.5 Small 1.5
the specimen about five times per lift for small cylinders and 10
M1.5 Medium 1.5
L1.5 Large 1.5 times per lift for large cylinders and slabs.
All specimens were cured in closed (capped) molds for seven
days in the laboratory conditions, with the ambient temperature
maintained around 21 °C.
three volume fractions of the same size fraction and four rPC mix-
tures containing four different size fractions of the processed 4. Results and discussion
CCFCM. The seven mixtures and their designated naming system
are listed in Table 2. For each mixture, the first letter represents 4.1. Porosity, density, and infiltration rate
the CCFCM element size fraction, (C, L, M, and S), followed by a
number that represents the CCFCM element volume fraction in Upon demolding the specimens at 7-day age, porosity and dry
percentage, 0.5, 1.0, and 1.5. In the case of the control mixture, density were determined in accordance with ASTM C1754 [16].
the letter and the number that describe the CCFCM element size The volume of the specimens was determined based on the aver-
and volume fraction were replaced with 00. age dimensions, namely diameter and height for cylinders, and
width and height for the slabs, obtained from three measurements
3.1. Specimen preparation taken using a caliper. Dry mass of the specimens was recorded.
Specimens were submerged in water for 30 min, after which their
The PC was mixed in accordance with the American Society for submerged mass was recorded. The volume of solids was esti-
Testing and Materials (ASTM) C192 [15]. Prior to mixing, CCFCM mated as the difference between the mass of dry and submerged
and the admixture were dispersed in the total water for the batch. specimen. Porosity was calculated based on the Eq. (1).
Three types of specimens were cast for this experiment: small  
Md  Mw
cylindrical specimens (100 mm in diameter by 200 mm in height), Porosity ¼ 1  ð1Þ
qw  V
prepared for 7- and 28-day compressive strength and Cantabro
tests, large cylindrical specimens (150 mm in diameter by where Mw is the mass of submerged specimen in g, Md is the mass of
300 mm in height) for 7-day split tensile strength test, and slabs dry specimen in g, qw is density of water in g/cm3, and V is volume
(286 by 286 by 83 mm) for mass loss in surface abrasion tests. Dur- of the specimen in cm3. Porosity is determined as the ratio of total
ing mixing, CCFCM dispersed evenly and without clumping volume of voids (void content) to the total volume of the specimen
throughout the fresh PC mixture, as shown in Fig. 3 (left). Disper- (solids and voids).
sion of the CCFCM particles throughout the cement matrix is also Fig. 4 presents the average porosity in three data series, repre-
evident in the hardened specimens, as seen in Fig. 3 (right). senting small, large cylinders, and slabs. As evident in Fig. 4, all
878 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885
Fig. 4. Average porosity based on specimen type; number of specimens: 17 small cylinders, five large cylinders and one slab per mixture.
the rPC mixtures present lower average porosities (22–24%) when
compared to the control PC (28%). It is also evident from the stan-
dard deviation (whisker bars in Fig. 4) that the rPC specimens from
the same mixture were compacted significantly more consistently
than the control specimens.
To establish the significance of the effect of the CCFCM on rPC
porosity, Pearson t-tests were conducted on small cylinders from
each mixture and the control mixture. The sample for the t-test
included 17 specimens for each mixture. The test results show that
the difference in the porosity of rPC mixtures and control mixture
is statistically significant at 95 percent confidence interval (p-value
is zero for each rPC mixture when compared to the control). From a
practical point of view, the significantly lower porosity achieved
for rPC with the same compaction effort implies that the addition
of CCFCM to the PC increases the mixture’s workability. This is a
significant improvement considering the low slump and workabil-
ity of PC, especially when placing in the field in hot weather
condition.
As opposed to conventional PCC, compressive strength is not
typically the prevalent quality control method for PC. Instead, mea-
surements of density can provide a better means for quality control
[17]. Therefore, dry density was estimated for each specimen as
part of the procedure described in the previous section to establish
porosity. The correlation between porosity and dry density for
small cylindrical specimens from all mixtures is presented in
Fig. 5. It is apparent that all rPC mixtures show a strong linear
relationship between porosity and dry density (R2 = 0.97), while
Fig. 5. The correlation between porosity and density for small cylinders from all PC mixtures.
porosity of the control mixture shows a great disparity in the rel-
atively small range of density. The relatively consistent density
for the control samples despite the clear change in porosity is most
likely due to variability in the volume of the specimens. The strong
linear correlation between density and porosity for the rPC
specimens indicates again that CCFCM elements resulted in higher
workability and therefore uniform volumes among all rPC
specimens.
Infiltration rate is one of the properties of PC that is critical for
stormwater management applications. Therefore, the infiltration
rate of the PC specimens was determined based on the procedure
outlined in ASTM C1701 [18] at 7-day age. Cylindrical specimens
were wrapped on the sides with shrink-wrap, which enabled the
water to be poured from the top and exfiltrated through the bot-
tom of the specimens without loss on the sides. Infiltration rate
of the slab specimens was determined by the usage of plastic infil-
tration ring, fastened to the slab by plumber’s putty, as shown in
Fig. 6. In both cases, tests were based on measuring the time in
which the designated volume of water infiltrates through the spec-
imen. For both specimen types, infiltration rates (I) were deter-
mined based on the Eq. (2):
4V
I¼ ð2Þ
D 2 pt
where V is the volume of infiltrated water in cm3, D is the diameter
of the specimen in the case of cylindrical specimens in cm, and the
diameter of the infiltration ring in the case of slab specimens in cm,
 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885
Fig. 6. Infiltration test on the slab specimen.
and t is the time in seconds required for the designated volume of
water to infiltrate through PC. Infiltration test was performed with
one liter of water for small cylinders and slabs, and two liters for
large cylinders. The infiltration rate of cylinders was reported as
the average of two measurements. Infiltration rates for slabs were
reported as the average of the measurements from four different
locations on each slab. It is noteworthy that all infiltration tests
were preceded with pre-wetting of specimen with the same
amount of water used in infiltration test.
Fig. 7 shows the infiltration rates for each specimen category. As
seen in Fig. 7, rPC specimens generally present higher infiltration
rates than the control mixture. The increase in infiltration ranges
from 4 to 32 percent for small cylinders, from 14 to 55 percent
for large cylinders and from 11 to 96 percent for slabs. Also, as seen
previously in Fig. 4, rPC specimens present lower porosity com-
pared to the control mixture. Higher infiltration rates despite lower
porosity for the rPC specimens can be an evidence of the CCFCM
elements influencing the connectivity of the voids and facilitating
the flow of water through the voids. It should be noted that the
average values of infiltration for all the cylindrical specimens are
well within the typical range for PC (750–5000 cm/h) [4].
Fig. 7. Average infiltration rate based on specimen type.
879
Fig. 7 shows that out of the different specimen categories, the
slabs present the highest infiltration rates for all the mixture types.
High values for the infiltration of the slabs are probably a result of
the lower compaction energy in comparison with the cylindrical
specimens.
As seen in Fig. 7, CCFCM dosage does not present a significant
influence on the average infiltration rate in the case of cylindrical
specimens, while the infiltration of the slabs is the lowest at
mid-range CCFCM content. When different CCFCM element types
are compared, it can be seen that S and L CCFCM fractions are asso-
ciated with higher values of infiltration in the case of the slabs and
small cylinders. The infiltration rate of large cylinders is relatively
consistent for all rPC mixture containing 1.5 percent CCFCM.
4.2. Compressive strength
Compressive strength (f0 c) tests were performed on small cylin-
drical specimens at 7- and 28-day ages, according to ASTM C39
[19]. Table 3 shows the average 7- and 28-day f0 c with correspond-
ing standard deviations for all mixtures. Furthermore, Pearson sta-
tistical t-test for two samples at 95 percent confidence interval was
conducted to determine whether the mechanical properties of rPC
and control mixture differ significantly. The sample for t-test con-
sisted of four specimens per mixture for each mechanical test. P-
values are reported in Table 3.
Based on Table 3, only rPC mixture C1 outperforms the control
mixture in terms of 7-day f0 c. Conversely, on 28-day tests, five out
of the six rPC mixtures present higher f0 c than the control mixture
(by 4–11 percent), however, the increase is not statistically signif-
icant. The mixture with the lowest f0 c on both 7- and 28-day tests is
S1.5. Overall, although no significant increase is gained in average
28-day f0 c compared to the control, the average 28-day f0 c for all the
rPC mixtures surpasses 20 MPa, which is a typical value for PC
[4,5].
When evaluating f0 c test results for PC, porosity is an influential
factor that cannot be ignored. Specimens with higher porosity gen-
erally tend to present lower strengths. To consider the effect of
porosity on the f0 c test results, Fig. 8 shows the 7- and 28-day f0 c
results versus the corresponding porosity for each tested specimen.
As seen in Fig. 8, rPC specimens are generally characterized by
lower porosities and higher values of f0 c, comparing to the control
mixture. The increase in 28-day f0 c for almost all rPC mixtures with
reference to the control mixture is still evident in Fig. 8. When dif-
ferent CCFCM dosages are compared, it can be seen that C1 yields
the highest f0 c on both test days. When different CCFCM element
types are compared, it is observed that the coarser CCFCM fractions
generally result in higher 28-day f0 c.
880 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885

Table 3
Mechanical properties of all PC mixtures: 7- and 28-day compressive, tensile strength and modulus of elasticity.

Test/Mixture Control C0.5 C1 C1.5 S1.5 M1.5 L1.5

0
Average 7-day f c [MPa] 15.5 14.0 16.6 11.0 9.5 12.0 12.4
St. dev. [MPa] 1.2 2.7 1.5 2.5 4.7 2.7 1.9
p-value – 0.388 0.287 0.03 0.092 0.082 0.053
Average 28-day f0 c [MPa] 19.5 21.2 21.6 20.6 16.0 20.2 21.4
St. dev. [MPa] 3.4 4.4 1.9 1.8 3.7 2.1 2.8
p-value – 0.550 0.360 0.585 0.222 0.713 0.422
Average 7-day f0 t [MPa] 1.8 2.0 1.8 2.6 2.2 2.6 2.2
St. dev. [MPa] 0.4 0.3 0.2 0.2 0.2 0.3 0.3
p-value – 0.434 0.96 0.02 0.108 0.016 0.146
Average Ec [MPa] 21.9 31.7 29.6 24.5 31.0 23.4 32.1
St. dev. [MPa] 4.8 10.1 3.5 2.1 6.3 5.6 10.6
p-value – 0.603 0.898 0.229 0.380 0.344 0.495

Fig. 8. Compressive strength (f0 c) on (top) 7- and (bottom) 28-day versus porosity.

When gains in f0 c from 7- and 28-day are compared, it is
observed that all rPC mixtures, except C1, underwent more signif-
icant increase in f0 c than the control mixture. This behavior shows
that the CCFCM elements are likely to have hindered or slowed
down the hydration progress. All of the rPC mixtures with 1.5 per-
cent CCFCM had f0 c gains higher than 60 percent, which is substan-
tially higher than that of the control mixture (26 percent). The
effect of CCFCM addition on the hydration progress needs more
investigation in future studies.
To further examine the effect of CCFCM on f0 c, the failure modes
in all specimens were investigated. Eight major failure types are
identified for the 7- and 28-day compressive strength tests and
are presented with photographs and descriptions in Table 4. These
failure types are described based on ASTM C39 [19] for conven-
tional PCC. Additionally, three frequently observed failure types:
bottom-up, and top-down columnar cracking, and cone with shear
were added to those defined in ASTM C39.
Fig. 9 shows the occurrence of the different failure types on
7- and 28-day tests, respectively, for the four specimens tested
on each test day. Six different failure types can be observed in
Fig. 9 for the 7-day test. Columnar cracking (TD and BU) and
columnar cracking with partially formed cones are relatively
frequent failure types. The crushing of top or bottom of the
specimen is another recurrent failure type, typically associated
with low f0 c. Shear and side cracking are less frequent failure
types on 7-day f0 c tests. On 28-day f0 c testing specimens present
four different failure types. The most common failure type is
shear, followed by the combination of cone and shear. Speci-
mens with highest 28-day f0 c demonstrate cone failure type,
while the specimens with lower 28-day f0 c typically fail by side
cracking. Based on the results in Figs. 8 and 9, it can be
concluded that strength gains of PC mixtures were associated
with changes in structural integrity, resulting in a change of
prevalent failure types.
 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885
Table 4
Failure types observed on 7- and 28-day f0 c tests.
Failure type name Bottom-up (BU)
columnar cracks
Failure type from ASTM C39 Not Available (NA)
Description Columnar cracks propagate
from the bottom upwards
evenly around
Photograph
Failure type name Shear
Failure type from ASTM C39
Type 4
Description Diagonal fracture
throughout, without
cracking on top/bottom
Photograph
4.3. Indirect split tensile strength
Indirect split tensile strength (f0 t) test was performed according
to ASTM C496 [20] on four large cylindrical specimens at 7-day
age. The average 7-day f0 t for all PC mixtures with their correspond-
ing standard deviations are presented in Table 3. Similar to f0 c test
results, Pearson statistical t-test at 95 percent confidence interval
was performed on test results and p-values are listed in Table 3.
As seen in Table 3, five out of the six rPC mixtures outperform
the control mixture in terms of 7-day f0 t of the PC by 11–46 percent.
The increase in f0 t is statistically significant for the mixtures C1.5
and M1.5. Seven-day f0 t of mixture C1 was slightly lower (approx-
imately three percent) than that of the control mixture, which is
not statistically significant. All of the mixtures with 1.5 percent
CCFCM present relatively high values of f0 t, beyond 2.2 MPa. Over-
all, the average 7-day f0 t for all rPC mixtures is 2 MPa, which is
approximately 26 percent higher than that of the control mixture
at 1.8 MPa. As Table 3 shows, most 7-day f0 t values for rPC exceed
the typical range for PC reported in other studies (from 1.4 to
2 MPa [7,21]).
Similar to f0 c, porosity influences f0 t of PC mixtures. Fig. 10 pre-
sents 7-day f0 t versus the corresponding porosity for each of the
specimens. As seen in Fig. 10, all rPC mixtures are characterized
by higher 7-day f0 t and lower porosity, comparing to the control
mixture. When different CCFCM dosages are compared, it can be
seen that 1.5 percent of CCFCM elements results in a highest value
of f0 t. When different CCFCM particle sizes are compared, it can be
concluded that combined and medium CCFCM element fractions
yield in highest value of f0 t associated with the mid-range porosity.
In terms of failure types, it was observed that the specimens
that presented higher strength generally had full split failures
and lower occurrence of failures at the lift locations.
4.4. Modulus of elasticity
Load-displacement measurements during testing were
available for the 28-day f0 c tests. The modulus of elasticity (Ec)
881
Top-down (TD) Columnar cracks with Cone
columnar cracks partially formed cones
NA
Type 3 Type 1
Columnar cracks propagate Columnar cracking from Well-formed cones on both
from the top downwards both ends with partially ends, caps intact after failure
evenly around formed cones at one end
Crushing of top or bottom Cone and shear Side fractures
NA
Type 6 Type 5
Severe cracking at Well-formed cone at one Side fractures at top/bottom;
top/bottom, resulting in end, prominent diagonal resembles shear failure, with
crushing of top/bottom (shear) cracks on another cracks on the top/bottom
was estimated as the slope of the linear trend-line used to
approximate the linear portion of stress-strain curves. Initial
stresses (up to 0.35 MPa) were considered as the seating period.
The average Ec with the corresponding standard deviation for
every mixture is given in Table 3. As compressive and tensile
strength test results, the Pearson’s statistical tests were conducted
on the values of Ec and corresponding p-values are provided in
Table 3. As seen in Table 3, four out of six rPC mixtures outperform
the control in terms of Ec (by 4–46 percent). Mixtures C1.5 and
M1.5 present lower average Ec than the control mixture, while
the mixture L1.5 shows the highest Ec, 32 MPa on average. The
changes in Ec, however, are not statistically significant, comparing
to the control mixture. Fig. 11 shows Ec values versus the corre-
sponding porosities of the specimens. When different CCFCM
dosages are compared, it can be seen that higher dosages yield
lower Ec. However, when different CCFCM particle sizes are com-
pared, it can be seen that small and large CCFCM fractions at 1.5
percent volumetric dosage result in relatively high values of Ec.
As already stated, control mixture was characterized by signifi-
cantly higher porosities and lower Ec on average, comparing to rPC.
4.5. Mass loss by Cantabro and surface abrasion tests
Cantabro test was performed according to ASTM C1747 [22] on
four small cylindrical specimens per mixture, obtained by cutting
the regular 100 by 200 mm cylinders in half. Specimens were
tested on Cantabro 50 days after casting. The test was conducted
in a Los Angeles (LA) abrasion test machine without the steel ball
charges. Fig. 12 shows the degradation of one rPC specimen after
every 50 cycles in the LA abrasion machine.
Abrasion tests with the rotational cutter were performed
according to ASTM C944 [23], with a 98-N load on the slabs. The
test was performed at four different locations on each slab, for a
total time of 14 min for each location. This test is more representa-
tive of abrasion under vehicular traffic than the Cantabro test. The
experimental setup for the surface abrasion test is presented in
882 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885

Fig. 9. Occurrence of different failure types on (top) 7- and (bottom) 28-day f0 c tests.

Fig. 10. Tensile strength (f0 t) for 7-day tests versus porosity.

Fig. 13 (left). Fig. 13 (right) shows all the slab specimens after the
surface abrasion test was conducted.
Fig. 14 presents the average mass loss from Cantabro and sur-
face abrasion test for all PC mixtures. Mixture M1.5 was the only
rPC mixture that outperformed the control in terms of average
mass loss on Cantabro test. Mixture L1.5 has the highest average
mass loss on Cantabro test, approximately 40 percent after 500
cycles. Mixture C0.5 presents lower mass loss on the surface abra-
sion test than the control mixture. Mass loss of mixtures C1.5 and
M1.5 is very close to that of the control mixture. Mixture L1.5 pre-
 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885 883

Fig. 11. Elastic modulus (Ec) for 28-day tests versus porosity.

resistance during Cantabro test, as well as relatively high 7-day

Fig. 12. Progression of degradation during Cantabro test. Number of cycles is
provided in the bottom right corner.
sent the highest mass loss on the abrasion test (0.66 percent on the
average), while the average mass loss of the control mixture is
approximately 0.2 percent. The relatively high abrasion resistance
of the mixture M1.5 is correspondent to its relatively high
tensile strength. The mixture L1.5 is characterized with the highest
mass losses on both tests. Mixtures C0.5 and C1.5 demonstrate
high abrasion resistance, followed with mid-range mass losses on
Cantabro. Mixtures C1, S1.5, and L1.5 are characterized by rela-
tively low abrasion resistance at both abrasion and Cantabro test.
In order to determine which mechanical properties (7-day f0 c,
28-day f0 t and Ec) best correspond to the resistance to degradation,
based on Cantabro test, and surface abrasion, Pearson correlation
factors with corresponding p-values were calculated for four tested
specimens from each mixture and presented in Table 5. As seen in
Table 5, resistance to degradation based on Cantabro test correlates
relatively well with 28-day f0 c and 7-day f0 t in the case of mixtures
with combined (C) CCFCM element fraction. In the case of mixture
C0.5, resistance to degradation correlates well with all three exam-
ined mechanical parameters. For mixtures M1.5, L1.5, and the con-
trol mixture, significant linear correlation between the degradation
resistance and mechanical properties cannot be established. A
strong correlation between the surface abrasion resistance and
mechanical properties cannot be defined. Moreover, as Table 5
shows, most correlation factors are negative, which indicates that
higher mechanical properties are associated with lower perfor-
mance on surface abrasion test. However, it is noteworthy that
losses on surface abrasion tests are relatively low (below 0.6%)
and the values of the remaining mass percentage were dispersed
around the relatively narrow range.

Fig. 13. (left) Experimental setup for the surface abrasion test; (right) All slab specimens after the surface abrasion test. The set-up slab was used to set up the machine.
884 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885

Fig. 14. Average mass loss by (left) Cantabro and (right) surface abrasion for all PC mixtures.

Table 5
Correlation factors for the results of the resistance on Cantabro and surface abrasion tests versus mechanical properties (28-day f0 c, 7-day f0 t and Ec), with corresponding p-values.

Control C0.5 C1 C1.5 S1.5 M1.5 L1.5

Cantabro 28-day f0 c, 0.799 0.916 0.649 0.927 NA 0.524 0.384
p-value 0.201 0.084 0.351 0.073 NA 0.476 0.616
7-day f0 t 0.21 0.967 0.49 0.508 NA 0.419 0.035
p-value 0.79 0.033 0.51 0.492 NA 0.419 0.035
Ec 0.308 0.875 0.361 0.696 NA 0.211 0.119
p-value 0.692 0.125 0.639 0.304 NA 0.789 0.881
Surface abrasion 28-day f0 c, 0.153 0.554 0.441 0.904 0.890 0.197 0.898
p-value 0.847 0.446 0.559 0.096 0.11 0.803 0.102
7-day f0 t 0.353 0.285 0.553 0.455 0.821 0.058 0.392
p-value 0.647 0.715 0.447 0.545 0.179 0.942 0.608
Ec 0.769 0.768 0.753 0.224 0.804 0.517 0.518
p-value 0.231 0.232 0.247 0.776 0.196 0.483 0.482

5. Conclusions as the control mixture. All mixtures showed high resistance

An experimental study was designed and executed in order to
investigate the ease and feasibility of mixing pervious concrete
(PC) with cured carbon fiber composite materials (CCFCM). Main
conclusions are summarized below:
– The use of CCFCM in the PC mixture resulted in significantly
improved workability, associated with lower and more consis-
tent porosity compared to the control mixture. Increased work-
ability can ease the placement, finishing, and compaction in the
field, and provide more time for placement, especially during
hot paving days.
– Despite the lower porosity, the infiltration rates of rPCs were
increased in the same specimen. Appropriate compaction pro-
cedure needs to be developed in the future to achieve the
desired infiltration in the field.
– As opposed to relatively low results for 7-day f0 c, on the 28-day
test, five out of six rPC mixtures outperform the control mixture
by 4–11 percent. rPC mixtures are characterized by higher gains
in f0 c comparing to the control mixture, which indicates that
CCFCM elements may influence the hydration process. Average
7-day split tensile strength of five out of six rPC mixtures was
higher than that of the control by 11–46 percent. Modulus of
elasticity of four out of six rPC mixtures outperforms that of
the control mixture by 4–46 percent.
– On Cantabro test, mixture M1.5 was the most resistant to degra-
dation and outperformed the control mixture. On surface abra-
sion test, the only rPC mixture that outperformed the control
mixture was C0.5, although mixtures M1.5 and C1.5
demonstrated similar behavior in terms of abrasion resistance
against surface abrasion.
The results of this preliminary study indicate that the inclusion
of CCFCM has a promising potential to improve physical and
mechanical properties of PC. However, more research is required
to determine the optimum volume and size fractions that would
result in the desired performance. Mechanical tests need to be
repeated at 1, 3, 7, 28, and 90 days to investigate the short- and
long-term rPC properties. Besides, more refined CCFCM fractions
and dosages can be examined for the purpose of the mixture design
optimization. Assessment of CCFCM effects on PC behavior when
exposed to freezing and thawing and/or deicer salts is also
recommended.
Acknowledgements
The authors gratefully acknowledge The Boeing Company for
providing the support and funding for this project. Authors express
their gratitude to the group of graduate, and undergraduate stu-
dents at WSU, and the staff at Composite Materials & Engineering
Center (CMEC), who helped with the extensive laboratory work in
this study.
References
[1] A. Marolf, N. Neithalath, E. Sell, K. Wegner, J. Weiss, J. Olek, Influence of
aggregate size and gradation on acoustic absorption of enhanced porosity
concrete, ACI Mater. J. 101 (1) (2004) 82–91.
[2] L. Haselbach, M. Boyer, J. Kevern, V. Schaefer, Cyclic heat island impacts on
traditional versus pervious concrete pavement systems, Transp. Res. Rec. J.
Transp. Res. Board 2240 (2011) 107–115.
 M. Rangelov et al. / Construction and Building Materials 126 (2016) 875–885
[3] J.T. Kevern, L. Haselbach, V.R. Schaefer, Hot weather comparative heat balances
in pervious concrete and impervious concrete pavement systems, J. Heat Isl.
Inst. Int. 7 (2) (2012).
[4] ACI. 522R-10, Report on Pervious Concrete ACI Committee, 522, 2010, p. 38.
[5] P.D. Tennis, M.L. Leming, D.J. AkersNational Ready Mixed Concrete Association,
Pervious Concrete Pavements, Portland Cement Association, Skokie, IL, 2004.
[6] J.T. Kevern, V.R. Schaefer, K. Wang, M.T. Suleiman, Pervious concrete mixture
proportions for improved freeze-thaw durability, J. ASTM Int. 5 (2) (2008) 12.
[7] S. Hesami, S. Ahmadi, M. Nematzadeh, Effects of rice husk ash and fiber on
mechanical properties of pervious concrete pavement, Constr. Build. Mater. 53
(2014) 680–691.
[8] A.M. Amde, S. Rogge, Development of High Quality Pervious Concrete
Specifications for Maryland Conditions, 2013. Final Report, MD-13-SP009B4F.
[9] B. Rehder, K. Banh, N. Neithalath, Fracture behavior of pervious concretes: the
effects of pore structure and fibers, Eng. Fract. Mech. 118 (2014) 1–16.
[10] Z. Yang, Freezing-and-thawing durability of pervious concrete under
simulated field conditions, ACI Mater. J. 108 (2) (2011).
[11] J.T. Kevern, D. Biddle, Q. Cao, Effects of macrosynthetic fibers on pervious
concrete properties, J. Mater. Civ. Eng. (2014) 06014031.
[12] Q. Dong, H. Wu, B. Huang, X. Shu, K. Wang, Investigation into laboratory
abrasion test methods for pervious concrete, J. Mater. Civ. Eng. 25 (7) (2012)
886–892.
[13] J.T. Kevern, K. Wang, V.R. Schaefer, The effect of aggregate type on the freeze-
thaw durability of pervious concrete, Advancements in Pervious Concrete
Technology, 19, 2008.
[14] G.C. Products, V-MAR VSC500, Rheology Modifying Admixture for Pervious
Concrete, 2010. Available from https://grace.com/construction/en-
us/concrete-technology/V-MAR-VSC500.
885
[15] ASTM C192/C192M-15, Standard Practice for Making and Curing Concrete Test
Specimens in the Laboratory, ASTM International, West Conshohocken, PA,
2015, http://dx.doi.org/10.1520/C0192_C0192M15.
[16] ASTM C1754/C1754M-12, Standard Test Method for Density and Air Void
Content of Pervious Concrete, ASTM International, West Conshohocken, PA,
2012, http://dx.doi.org/10.1520/C1754_C1754M12.
[17] NRMCA, Pervious in Practice: Methods, Materials and Mixtures, National
Ready Mixed Concrete Association (NRMCA), SS, MD. Available from <www.
nrmca.org>.
[18] ASTM C1701/C1701M, Standard Test Method for Infiltration Rate of In Place
Pervious Concrete, ASTM International, West Conshohocken, PA, 2009, http://
dx.doi.org/10.1520/C1701_C1701M-09.
[19] ASTM C39/C39M-05, Standard Test Method for Compressive Strength of
Cylindrical Concrete Specimens, ASTM International, West Conshohocken, PA,
2005, http://dx.doi.org/10.1520/C0039_C0039M-05E01.
[20] ASTM C496/C496M-11, Standard Test Method for Splitting Tensile Strength of
Cylindrical Concrete Specimens, ASTM International, West Conshohocken, PA,
2011, http://dx.doi.org/10.1520/C0496_C0496M-11.
[21] Development of mix proportion for functional and durable pervious concrete,
in: K. Wang, V.R. Schaefer, J.T. Kevern, M.T. Suleiman (Eds.), NRMCA Concrete
Technology Forum: Focus on Pervious Concrete, 2006.
[22] ASTM C1747/C1747M-13, Standard Test Method for Determining Potential
Resistance to Degradation of Pervious concrete by Impact and Abrasion, ASTM
International, West Conshohocken, PA, 2013, http://dx.doi.org/10.1520/
C1747_C1747M-13.
[23] ASTM C944/C944M-12, Abrasion Resistance of Concrete or Mortar Rotating-
Cutter Method, ASTM International, West Conshohocken, PA, 2012, http://dx.
doi.org/10.1520/C0944_C0944M-12.