Construction and Building Materials 95 (2015) 850–859

Contents lists available at ScienceDirect

Construction and Building Materials

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

The effect of the cementitious paste thickness on the performance

of pervious concrete
Anthony Torres a,⇑, Jiong Hu b, Amy Ramos a
a
Texas State University, 601 University Dr., San Marcos, TX 78666, United States
b
University of Nebraska-Lincoln, The Peter Kiewit Institute, Omaha, NE 68182, United States

h i g h l i g h t s

 Developed a new method to characterize cement paste thickness of pervious concrete.

 Related cement paste thickness with key performance criteria of pervious concrete.
 Validated the effect of design variables on the cementitious paste thickness.

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

Article history: The performance of pervious concrete is dictated by the void structure features such as the size of coarse
Received 11 May 2015 aggregate, void volume, void size, void distribution, and the amount of cement used. Pervious concrete
Received in revised form 24 July 2015 consists of coarse aggregate surrounded by a thin layer of Portland cement paste. The amount of cement
Accepted 28 July 2015
used affects the aggregate coating thickness, which has an effect on the porosity and other mechanical
Available online 30 July 2015
properties of the concrete. The void size is not purely governed by the size of coarse aggregate, but the
aggregate size, gradation, and the thickness of cement paste. A higher porosity of the material does
Keywords:
not ensure a higher permeability, as the permeability is a function of the void surface area, void size,
Pervious concrete
Cementitious paste
and distribution. In this study a correlation is made between key pervious concrete properties such as
Porosity the porosity, permeability, compressive and tensile strength versus the cement paste thickness. A thicker
Compaction cement coating will lead to a lower percolation rate by reducing the porosity, however, this could have a
Performance positive effect on other desirable mechanical properties, such as compressive/tensile strength and perme-
Runoff ability. This research focus on proportioning the coarse aggregate, cement content, and the void content
in order to correlate the cement paste thickness to the key properties of pervious concrete. This was com-
pleted through a controlled concrete mix-design, a controlled laboratory preparation technique and an
analysis of the hardened concrete cross-section.
Published by Elsevier Ltd.

1. Introduction
1.1. Background
Pervious concrete is a sustainable alternative to conventional
paving materials as it facilitates water infiltration while maintain-
ing its structural performance. The ability to allow water to pene-
trate through this material allows for a superior storm-water
run-off control and control of nonpoint source pollution (NPS).
NPS pollution results from land runoff, precipitation, drainage
and seepage and has been identified as a cause of receiving water
⇑ Corresponding author.
E-mail addresses: anthony.torres@txstate.edu (A. Torres), jionghu@hotmail.com
(J. Hu), ajr93@txstate.edu (A. Ramos).
quality degradation in the United States [1]. Most of this runoff is
due to impervious pavements in large parking areas, roadways,
and roofs. Pervious concrete is a sustainable alternative to
impervious pavements that can greatly influence the control of
runoff and NPS pollution. It is documented that pervious concrete
helps in groundwater recharge.
Pervious concrete consists of coarse aggregate surrounded by a
thin layer of Portland cement paste. The amount of cement used
affects the aggregate coating thickness, which has an effect on
the porosity and other mechanical properties of the concrete. The
infiltration performance of pervious concrete is dictated by the
void structure features such as the size of coarse aggregate, void
volume, void size, void distribution, and the amount of cement
used [2]. A higher porosity of the material does not ensure a higher
permeability, as the permeability is a function of the void surface

http://dx.doi.org/10.1016/j.conbuildmat.2015.07.187
0950-0618/Published by Elsevier Ltd.
 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859
area and void size [3]. The void size is not purely governed by the
size of coarse aggregate, but both the aggregate size and the thick-
ness of cement paste. In this study a correlation will be made
between the porosity versus the cement paste thickness. A thicker
cement coating will lead to a lower percolation rate by reducing
the porosity, however, this could have an effect on other desirable
mechanical properties, such as strength, absorption, freeze/thaw
capabilities, durability, etc. Reports have shown that pervious con-
crete is to have a void content in the range of 11–35% [4,5]. The
work by Kevern et al. [6] describes that pervious concrete contains
two types of void systems: (i) intentionally designed porosity that
facilitates water penetration and (ii) entrained air in the mortar
(cementitious material) that surrounds the aggregate. Montes
et al. define the porosity of pervious concrete as the percent vol-
ume of voids with respect to the total volume of the specimen. It
is further articulated by describing the ‘‘effective porosity’’. The
effective porosity is described as the portion of the total void space
that is active in the fluid permeability process, which excludes the
‘‘inactive porosity’’ sights. Montes et al. point out that not all the
porous spaces are effective in holding and available for fluid flow,
which is grouped as the ‘‘inactive porosity’’. Some porous spaces
can be isolated (closed off) from other void spaces and will not
be able to transmit fluid. This behavior is expected in highly com-
pact or small coarse aggregate (4.75 mm) pervious concrete as
smaller voids have the ability to retain the trapped fluid due to sur-
face tension effects and capillary action. Wimberly et al. [7] report
the effective porosity as the fraction of total porosity that was
allowed to drain in 30 min, called the ‘‘rapid flow’’ porosity, in
which a the fluid would rapidly flow, or drain, from the pervious
concrete sample. However, this method has an increase in variabil-
ity due to the high possibility of unknowns.
Despite the efforts outlined above, describing the characteriza-
tion of the void systems of pervious concrete, there is still uncer-
tainty in how different pervious concrete mixtures will affect the
void systems within the concrete. The coarse aggregate in pervious
concrete is typically single-sized [8], in which the coarse aggregate
is all the same size. In this case, the remaining variables in the
design mixture will be the amount of cement, the amount of water
and the amount of compaction energy. Adjusting any of these vari-
ables will inherently affect the paste thickness surrounding the
aggregate, therefore affecting the void system. Thicker paste thick-
ness surrounding the aggregate results in a lower void percentage
and thus a lower permeability (performance). These factors are
only attributed to the mixture design and other factors such as
placement (compaction) techniques are not considered.
1.2. Objectives
The objective of this study is to develop a correlation between
the cementitious paste thickness to the performance of pervious
concrete. This was completed through limiting the coarse aggre-
gate size (two sizes), proportioning the cement content (low, med-
ium, and high), and controlled laboratory compaction technique.
The proportioning was coordinated through a refined concrete
Table 1
Physical properties of aggregates.
Property Standard
Unit weight ASTM C29
Water absorption ASTM C127
Bulk specific gravityssda ASTM C127
Bulk specific gravityodb ASTM C127
Voids ASTM C29
a
ssd, saturated surface dry condition.
b
od, oven dried condition.
851
mix-design and the laboratory compaction technique was limited
to three compaction energy levels. The paste thickness was ana-
lyzed by specimen cross-sectional analysis of the hardened con-
crete and hand measurements. This study aims to establish new
understanding of pervious concrete through a unique investigation
of the material to help guide future mixture designs. This was mea-
sured by investigating the effect of the cementitious paste thick-
ness on the performance of pervious concrete, such as strength,
permeability, and porosity.
2. Experimental program
2.1. Materials
Two sizes of aggregate were used in this study in order to deter-
mine how the paste thickness correlates with varying sizes of
aggregate. The two sizes are 9.54-mm (3/8-in.) and 6.35-mm
(1/4-in.) limestone, which were obtained from local quarries in
Hays County, Texas. The two sizes were chosen to reflect typical
pervious concrete coarse aggregate sizes [8]. Type I/II cement
was used, which was obtained locally. The specific gravity, water
absorption, voids, and unit weight of each aggregate size is shown
in Table 1.
In order to focus this study on the cementitious paste thickness,
other variables were minimized as much as possible, such as
aggregate type, cement type, water-to-cement ratio, sample
size, admixtures, supplementary cementitious materials, etc.
Therefore, only a single aggregate type, a single cement type, and
no admixtures were used for this study.
2.2. Mixture proportions
Six mixtures with a water-cement ratio of 0.33 were propor-
tioned based on the methods described in ACI 522 R-10 [8]. The
experimental design consisted of two limestone aggregate sizes
and three compaction levels as described in Section 2.3. Due to
the three compaction levels and two aggregate sizes the total vary-
ing mixtures amounted to 18 sample groups, as shown in Table 2.
Mixtures were prepared using a rotating drum mixer and mixed in
accordance to ASTM C192-15 [9].
2.3. Experimental methods
2.3.1. Laboratory compaction techniques
Three compaction levels were used in this study to provide a
range of porosity quantities that would mimic in-place pervious
concrete systems. In order to produce a broad spectrum of com-
paction while simultaneously managing test variables the three
compaction levels were (i) uncompacted, which provides the low
end of the compaction spectrum as the concrete was simply placed
into the molds, (ii) two layers compacted with a 9.54-mm (3/8-in.)
tamping rod five times per layer, which provides an approximate
middle compaction level, and (iii) three layers compacted with a
2.5-kg (5.5-lb) standard proctor hammer having a height of fall
Unit Limestone (9.54 m) Limestone (6.35 mm)
3 3
kg/m (lb/ft ) 1442 (90.0) 1458 (91.0)
% 2.33 3.54
– 2.57 2.58
– 2.51 2.52
% 29.36 30.15
852 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859

Table 2
Mixture proportions.

Size of limestone mm (in.) Mix name Aggregate kg/m3 (lb/yd3) Cement kg/m3 (lb/yd3) Water kg/m3 (lb/yd3)
9.54 (3/8) Unc 9.54 (HP) 1404.8 (2367.7) 388.1 (654.1) 128.1 (215.9)
Unc 9.54 (MP) 1465.3 (2469.8) 351.5 (592.5) 116 (195.5)
Unc 9.54 (LP) 1531.4 (2581.1) 311.7 (525.4) 102.9 (173.4)
2-5R 9.54 (HP) 1404.8 (2367.7) 388.1 (654.1) 128.1 (215.9)
2-5R 9.54 (MP) 1465.3 (2469.8) 351.5 (592.5) 116 (195.5)
2-5R 9.54 (LP) 1531.4 (2581.1) 311.7 (525.4) 102.9 (173.4)
3-10PH 9.54 (HP) 1404.8 (2367.7) 388.1 (654.1) 128.1 (215.9)
3-10PH 9.54 (MP) 1465.3 (2469.8) 351.5 (592.5) 116 (195.5)
3-10PH 9.54 (LP) 1531.4 (2581.1) 311.7 (525.4) 102.9 (173.4)
6.35 (1/4) Unc 6.35 (HP) 1459.6 (2460.1) 355 (598.4) 117.2 (197.5)
Unc 6.35 (MP) 1532.2 (2528.4) 311.2 (534.6) 102.7 (173.1)
Unc 6.35 (LP) 1632.9 (2752.3) 250.4 (422.1) 82.6 (139.3)
2-5R 6.35 (HP) 1459.6 (2460.1) 355 (598.4) 117.2 (197.5)
2-5R 6.35 (MP) 1532.2 (2528.4) 311.2 (534.6) 102.7 (173.1)
2-5R 6.35 (LP) 1632.9 (2752.3) 250.4 (422.1) 82.6 (139.3)
3-10PH 6.35 (HP) 1459.6 (2460.1) 355 (598.4) 117.2 (197.5)
3-10PH 6.35 (MP) 1532.2 (2528.4) 311.2 (534.6) 102.7 (173.1)
3-10PH 6.35 (LP) 1632.9 (2752.3) 250.4 (422.1) 82.6 (139.3)

Note: Unc = uncompacted; 2-5R = 2 lifts 5 rods/lift; 3-10PH = 3 lifts 10 proctor hammer blows/lift; HP = high paste content, MP = medium paste content, LP = low paste
content.

of 300 mm (12-in.), that distributes the concrete with a 100-mm
(4-in.) head, which provides an approximate high compaction
level. ASTM C192-14 was used as a guide for selecting the desired
compaction levels of the cylinders and the size of cylinder molds.
The concrete specimens were cast and compacted in plastic cylin-
ders of 200-mm (8-in.) lengths by 100-mm (4-in.) diameter molds.
For the uncompacted specimens, the fresh concrete was collected
from the mixer using the plastic mold as a ‘scoop’ and the opera-
tor’s gloved hand was used to ensure the mold was full. This was
done to ensure the compaction energy was minimized. After cast-
ing and compacting, the samples were left at room temperature
(22 °C) for 24 h. After that period of time, all the samples were
demolded and stored in a concrete curing room at >98% humidity
and a temperature of ±72°F (±22.2 °C) for 28 days in accordance to
ASTM C192-14.
2.3.2. Test procedures
2.3.2.1. Cementitious paste thickness. There is no developed
standard for determining the cementitious paste thickness of a
pervious concrete sample; therefore a novel procedure was estab-
lished for this study. First, this study defines the ‘‘cementitious
paste thickness’’ as the average thickness of cementitious paste
that surrounds each individual aggregate in a pervious concrete
system. That is the maximum length from the edge of an individual
aggregate to the outermost edge of the cement paste, which sur-
rounds that individual piece of aggregate. It should be pointed
out that this study generalizes the paste as cementitious, which
could include cement and/or other supplementary cementitious
materials (SCMs). This study only utilizes cement, but the results
could be adaptable to other studies with the same overall percent
of cementitious materials.
In order to quantify the cementitious paste thickness, the sam-
ples were cut widthwise into cross-sections that were 100-mm
(4-in.) wide. This step is already required in the operating proce-
dures of the falling head permeameter [10]; therefore these cross
sections were utilized in order to correlate the data between the
same samples in which other tests were completed. An example
of the cross sections taken from the samples is shown in Fig. 1.
Once all cross-sections were produced a method for determin-
ing the cementitious paste thickness was implemented. There are
two options for completing this analysis: software based or manual
hand measurements. Both provide their own benefits and draw-
backs, which are addressed here. Software based analysis may
not be able to distinguish between extraneous features, such as

Fig. 1. Example cross-section of pervious concrete samples.

 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859 853

void spaces, paste that is below the analysis plane, shadows and off
color aggregate. Software based analysis requires each sample to
be scanned by a high-resolution scanner, such that all features
are visible. Next, the image needs to be processed and imported
into a software package capable of sophisticated measurements.
Each image needs to have a scale bar incorporated into the image
such that the software can reference each measurement. Lastly, the
software may not understand where to make the measurements
from, in regards to each edge of the aggregate or edge of the paste.
An operator could possibly overcome most of these drawbacks, but
due to the uniqueness of each cross-section the operation would
have to be adjusted for each sample. Due to these obstacles a man-
ual method was developed that would require approximately the
same amount of time required to produce the unique computa-
tional adjustments required for the alternative.
First, a systematic pattern was developed such that there would
be consistency between each sample. A two-dimensional linear
grid was produced in AutoCAD that consisted of vertical and hori-
zontal lines in a 100-mm (4-in.) circle, which is the diameter of the
cross sectional pieces. Two separate grids were produced with the
spacing of both the vertical and horizontal lines at 9.54-mm
(3/8-in.) and 6.35-mm (1/4-in.), which is the size of the two
aggregates used in this study. The grids were printed on a transpar-
ent plastic film, which are often used for overhead projector
slides. The grids were than placed on top of each cross-sections
as shown in Fig. 2.
The two grid systems were developed in order to establish con-
sistent measuring point across each sample. The line spacing was
selected due to the aggregate size in the respective samples, such
that the appropriately sized grid will be overlaid on a
cross-section with the same aggregate size. Once the respective
grid was overlaid and secured in place, a digital measuring caliper
with a resolution of 0.05-mm ± 0.05-mm (0.001-in ± 0.001-in) was
used to measure the cementitious paste thickness while measuring
across each vertical and horizontal line. The operator began with
the first horizontal line at the top of the sample and working from
left to right measuring across each horizontal line. As the operator
measured across the line, the digital caliper was used to measure
the length (along the gridline) of the paste until a void or piece
of aggregate was encountered, at which point the process was
repeated each time a cementitious paste edge was observed.
Finally, the length of each past thickness and the corresponding
line were recorded. This processes was repeated for all horizontal
and vertical lines. Fig. 3 demonstrates a single measurement being
taken along a solitary horizontal line.
The grid system incorporates 22 (both vertical and horizontal)
measurable lines with the 9.54-mm (3/800 -in.) grid and 30 (both
vertical and horizontal) measurable lines with the 6.35-mm
(1/4-in.) grid. Each line may have approximately 3-15 measure-
ments per line. The drawbacks to this system are the amount of
time required and the human error element. However, the opera-
tor’s understanding of what to disregard and what to measure

Fig. 2. Example of 9.54-mm (3/8-in.) [left] and 6.35-mm (1/4-in.) [right] grids overlaid on cross-sections of pervious concrete samples.

Fig. 3. An example of a paste thickness measurement using a grid and digital measurement calipers.
854 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859
counteracts the downside of the human error element. Due to the
amount of measurements taken, the human error element will
ultimately be averaged out. Lastly, an equivalent amount of time
is necessary for the alternative software based method.
Therefore, the by-hand method developed was used throughout
this study.
Utilizing the grid system introduces a slight oversight in mea-
surement gathering, such that the measurement taken along the
line may not always result in the largest paste thickness dimension
between aggregates. An example of this issue is depicted in Fig. 4.
It is shown in Fig. 4, the aggregates and paste do not align per-
fectly such that the measured distance is the largest paste thick-
ness between the aggregate. The depiction on the right (Fig. 4B)
shows a more realistic, random, distribution of aggregate and
paste. The depiction in Fig. 4 only demonstrates a representation
of the concrete mixture and only identifies measurements that
spans from aggregate to aggregate. It can be seen from Fig. 4B
(realistic case) that the realistic paste thickness measurements
(dashed line) could result in larger or smaller lengths than what
would be measured with the grid system established for this study.
It should be noted that there are more possible ‘measurement
lines’ than the ones listed in Fig. 4, such that the angle of the
dashed line can be altered plus/minus any degree resulting in a
completely different ‘measurement line’ with a different length.
Therefore, to correct for this instance a correction factor has been
developed such that the perpendicular (vertical/horizontal) lines
from the respective grids are the only measurements necessary.
A basic trigonometry relationship can be used to determine
the actual length of the cementitious paste thickness as shown in
Eq. (1).
measured thickness
Actual thickness ¼
cos h
Due to the difficulty of measuring each h angle an average of h
angles from 0° to 90°, at a 1° interval, were produced to develop
the correction factor shown in Eq. (2),
P90
cos h
h¼0
b¼
91
where b is the correction factor. Computing Eq. (2) results in a cor-
rection factor of 0.63, which will be applied to all measurements
taken via the developed grid method. Using the developed correc-
tion factor for all measurements will produce more realistic paste
thickness measurements.
To confirm the bias of the measurement method, the
single-operator repeatability was investigated by fully measuring
Fig. 4. Depiction of cementitious measurements vs. realistic measurements.
one cross section (as outlined above) three times by the same oper-
ator. ASTM C670-13 [12] was used as a guide when preparing the
precision and bias measurements. The maximum average paste
measurement difference was found to be 0.08-mm (0.0031-in.)
and the minimum average paste measurement difference was
found to be 0.07-mm (0.0027-in.) with an average standard devi-
ation of 0.04. The reproducibility was also investigated by fully
measuring the paste thickness of one cross section three times
by three different operators and the respective averages were com-
pared for variability. The reproducibility investigation revealed a
maximum average paste measurement difference of 0.11-mm
(0.0043-in.) and the minimum average paste measurement differ-
ence was found to be 0.09-mm (0.0035-in.) with an average stan-
dard deviation of 0.05. Based off these findings and the precision of
the caliper used, the results will have a ± 0.16-mm (0.0062-in.)
variability after the correction factor (b) is applied, which is the
maximum result from the repeatability/reproducibility investiga-
tion plus the precision of the calipers. Therefore, error bars are
used when reporting the cementitious paste thickness with the
value of ±0.16-mm (0.0062-in.).
2.3.2.2. Porosity and permeability. The porosity of the pervious con-
crete samples was determined using the method developed by
Montes et al. [3], which was the basis for ASTM C1754-12 [11].
ASTM C1754-12 was followed in order to guarantee accuracy with
the underwater weight of the specimens.
A falling-head permeameter was used to measure the perme-
ability of each sample as outlined by Neithalath et al. [10]. The pro-
cedures, developed by Neithalath et al [10], were followed to
determine the permeability of the pervious concrete samples.
Three specimens were tested per mix name (e.g., Unc 9.54 (MP))
ð1Þ
and the results were averaged.
2.3.2.3. Compressive strength and splitting tensile strength. The com-
pressive strength tests were performed in accordance with ASTM
C39-15a [13]. The specimens were capped on the ends using a sul-
ð2Þ fur compound to provide plane surfaces and ensure an even distri-
bution of the compressive force, which was completed in
accordance to ASTM C617-15 [14]. The compressive force was
applied in accordance to ASTM C39-15a [13] until the specimen
displayed a well-defined fracture pattern. The compressive
strength of the specimens was determined at the age of 28-day.
The splitting tensile strength tests were performed in accor-
dance with ASTM C496-11 [15]. The specimens were tested using
a constant load until the specimens displayed a well-defined frac-
ture pattern. The splitting tensile strength of the specimens was
determined for 28-day strengths. Three specimens were tested
per mix name (e.g., Unc 9.54 (MP)) and the results were averaged
and reported. The sample variability is presented in the results
using error bars.
3. Analysis of results
3.1. Cementitious paste thickness
The process outlined in Section 2.3.2.1 to determine the cemen-
titious paste thickness of each sample results in many measure-
ments per sample, due to each horizontal and vertical line
containing multiple measurements as well as each sample contain-
ing multiple cross-sections. Therefore, a data processing and anal-
ysis technique was developed to produce a single solitary number
to represent each concrete sample. First, all the measurements
from each cross section were recorded on a table similar to that
of Tables 3 and 4.
 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859 855

Table 3
Example cementitious measurement table for an uncompacted 9.54-mm (HP) sample.

Unc. 9.54-mm (HP) 1

Line designation H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 H11
Measurements (mm) 1.77 1.02 4.13 1.28 1.22 1.91 3.16 1.94 1.29 1.73 4.46
– 1.40 5.69 2.62 1.57 0.86 5.00 2.91 4.49 2.48 1.15
– 2.63 0.80 0.68 0.82 3.04 1.65 5.07 1.47 1.15 –
– – 2.08 – 2.43 1.65 2.00 1.00 3.50 – –
– – 0.50 – 0.61 2.71 1.87 – – – –
– – 1.45 – – 0.88 1.60 – – – –
– – – – – 0.49 – – – – –
Average (mm) 1.77 1.68 2.44 1.53 1.33 1.65 2.55 2.73 2.69 1.79 2.81
Maximum (mm) 1.77 2.63 5.69 2.62 2.43 3.04 5.00 5.07 4.49 2.48 4.46
Minimum (mm) 1.77 1.02 0.50 0.68 0.61 0.49 1.60 1.00 1.29 1.15 1.15

Line designation V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11

Measurements (mm) 1.77 1.02 4.13 1.28 1.22 1.91 0.73 0.84 0.98 1.11 1.88
– 1.40 5.69 2.62 1.57 0.86 3.80 1.36 0.63 0.60 –
– 2.63 0.80 0.68 0.82 3.04 2.90 0.56 5.66 – –
– – – 2.08 2.43 1.65 1.96 4.22 4.48 – –
– – – – 0.61 2.71 0.40 3.83 – – –
– – – – – 0.88 1.80 – – – –
– – – – – – 1.11 – – – –
Average (mm) 1.77 1.68 3.54 1.67 1.33 1.84 1.81 2.16 2.94 0.86 1.88
Maximum (mm) 1.77 2.63 5.69 2.62 2.43 3.04 3.80 4.22 5.66 1.11 1.88
Minimum (mm) 1.77 1.02 0.80 0.68 0.61 0.86 0.40 0.56 0.63 0.60 1.88

Table 4
Example cementitious measurement table for 3-10PH 6.35-mm (LP) sample.

3-10PH 6.35-mm (LP) 1

Line Designation H1 H2 H3 H4 H5 H6 H7 H8 H9 H10 H11 H12 H13 H14 H15
Measurements (mm) 0.90 1.45 2.08 4.50 5.60 0.88 1.63 5.30 1.34 2.35 0.89 0.00 0.85 0.00 1.90
– 0.45 0.50 0.64 0.30 0.48 0.55 2.00 3.10 0.74 – – 1.19 – –
– 1.40 1.45 – 1.23 1.60 – 0.80 – 0.85 – – – – –
– – – – – – – – – 1.01 – – – – –
Average (mm) 0.90 1.10 1.34 2.57 2.38 0.99 1.09 2.70 2.22 1.24 0.89 0.00 1.02 0.00 1.90
Maximum (mm) 0.90 1.45 2.08 4.50 5.60 1.60 1.63 5.30 3.10 2.35 0.89 0.00 1.19 0.00 1.90
Minimum (mm) 0.90 0.45 0.50 0.64 0.30 0.48 0.55 0.80 1.34 0.74 0.89 0.00 0.85 0.00 1.90

Line designation V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12 V13 V14 V15

Measurements (mm) 2.28 0.00 1.38 0.46 1.48 2.71 0.42 2.20 1.65 0.63 0.49 1.05 0.32 0.37 1.22
– – 0.55 – 0.25 0.88 0.40 3.83 0.45 – 0.61 – 0.67 0.26 –
– – – – 2.40 – 1.80 – 1.08 – – – – – –
– – – – – – 1.34 – – – – – – – –
Average (mm) 2.28 0.00 0.97 0.46 1.38 1.80 0.99 3.02 1.06 0.63 0.55 1.05 0.50 0.32 1.22
Maximum (mm) 2.28 0.00 1.38 0.46 2.40 2.71 1.80 3.83 1.65 0.63 0.61 1.05 0.67 0.37 1.22
Minimum (mm) 2.28 0.00 0.55 0.46 0.25 0.88 0.40 2.20 0.45 0.63 0.49 1.05 0.32 0.26 1.22

Table 3 shows example measurements taken from a single cross
section that was produced from a 9.54-mm (3/8-in.) uncompacted
(high paste) sample. This table represents actual measurements
taken for that particular cross section with both the horizontal
and vertical lines measured. Shown on Table 3 is the line designa-
tion area, which counts the horizontal lines from left to right as
well as the vertical lines on the grid from top to bottom. The fol-
lowing columns are labeled H1–11 and V1–11, which correspond
to the horizontal and vertical lines as described in
Section 2.3.2.1. That section mentions a total of 22 vertical and hor-
izontal lines in the 9.54-mm (3/8-in.) grids. Table 4 represents a
6.35-mm (1/4-in.) sample and corresponding grid overlay. As
opposed to the 9.54-mm (3/8-in.) grid, this grid contains 15 hori-
zontal and vertical lines for a total of 30 lines. As with Table 3,
the line designation area counts the measurement lines from left
to right and top to bottom with the labels of H1–15 and V1–15.
By inspecting Table 3, it can be seen first that the shorter the
line length (H1/H11/V1/V11) the less measurements per line.
This is intuitive by nature, but it should be pointed out that there
are cases in which a longer measurement line could result in fewer
measurements due to a tighter compaction. For example, Table 4
represents the highest compaction level and lowest paste level
with overall lower measurements taken than Table 3.
Measurement line H8 (Table 4), which is found at the center of
the sample, has a longer measurement length than any other line
but contains only three measureable cementitious paste thick-
nesses. In fact, H2, in this case, has the same measureable informa-
tion as H8, although H2 is the second shortest line on the grid. The
trend overall, however, does favor longer lines resulting in more
measureable content. Once all measureable information has been
extracted from the cross sections, the data per line was averaged
and the maximum and minimum was produced. To further consol-
idate this information the average of the averages, maximum of the
maximums, and minimum of the minimums was produced. This
process was completed for each cross section produced per sample
and an example of the information can be seen in Fig. 5.
856 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859

Fig. 5. Cementitious paste thickness of Unc. 9.54-mm (HP) sample. Fig. 6. Relationship between the average cementitious paste thickness and the
porosity for 9.54-mm (3/8-in.) aggregate samples.

As shown in Fig. 5, the cementitious paste thickness quantifica-

tion process results in a solitary number per cross section per min,
max, and average parameter. The last step used to produce a single
representative number per concrete sample was to take a final
average of the four averages produced from each cross section.
As seen in Fig. 5, the four average values have a small variation
between themselves and taking a final average of these values pro-
duces a representative cementitious paste thickness value for each
cylinder. The final values for the cementitious paste thickness for
all samples can be found in Table 5.

3.2. Porosity

3.2.1. Porosity vs. cementitious paste thickness

As the cementitious paste thickness increases the active poros-
ity sites will get smaller, which results in the overall porosity of the
samples decreasing. Figs. 6 and 7 shows the relationship between
porosity and permeability for the three compaction levels and
the two different aggregate sizes.
As seen in both Figs. 5 and 6 the uncompacted samples result in
a lower paste thickness as well as a correspondingly higher poros-
ity. This result is as expected due to the lack of compaction on the
samples. It can be seen that the middle compaction level (2 lifts
and 5 rods per lift; 2-5R) results in a higher cementitious paste
thickness and consequently lower porosity. Following this trend
Fig. 7. Relationship between the average cementitious paste thickness and the
porosity for 6.35-mm (1/4-in.) aggregate samples.
is the highest compacted samples (3 lifts and 10 proctor hammer
blows per lift; 3-10PH) with the majority of the samples having
the thickest paste and the lowest porosity. It can be gleaned from
these graphs and Table 4 that the cement content plays a role in
the porosity of the samples. For all six samples displayed in the
graph the highest paste samples are on the lower right, resulting

Table 5
Data summary.

Mix name Cementitious paste thickness, mm (in.) Porosity, % Permeability, k cm/s (in/s) Compressive strength, Mpa (psi) Tensile strength, Mpa (psi)
Unc 9.54 (HP) 3.79 (0.14) 25.60 1.05 (0.41) 12.5 (1812) 2.24 (325)
Unc 9.54 (MP) 3.24 (0.12) 28.20 1.25 (0.49) 8.3 (1204) 1.89 (274)
Unc 9.54 (LP) 2.57 (0.10) 31.20 1.69 (0.67) 3.2 (464) 1.09 (158)
2-5R 9.54 (HP) 4.69 (0.18) 22.30 0.72 (0.28) 14.8 (2146) 2.58 (374)
2-5R 9.54 (MP) 4.00 (0.15) 24.90 0.98 (0.39) 13.2 (1914) 2.44 (354)
2-5R 9.54 (LP) 3.43 (0.13) 27.10 1.19 (0.47) 10.4 (1508) 2.21 (320)
3-10PH 9.54 (HP) 5.61 (0.22) 17.98 0.42 (0.17) 17.1 (2479.5) 2.88 (418)
3-10PH 9.54 (MP) 4.87 (0.19) 20.13 0.59 (0.23) 16.4 (2378) 2.79 (405)
3-10PH 9.54 (LP) 4.19 (0.16) 22.89 0.69 (0.27) 15.1 (2190) 2.68 (389)
Unc 6.35 (HP) 4.02 (0.15) 26.41 1.15 (0.45) 12.4 (1798) 2.44 (354)
Unc 6.35 (MP) 3.64 (0.14) 28.90 1.26 (0.50) 9.9 (1435.5) 2.24 (325)
Unc 6.35 (LP) 2.98 (0.11) 31.34 1.43 (0.56) 6.5 (943) 1.83 (265)
2-5R 6.35 (HP) 4.76 (0.18) 22.89 0.72 (0.28) 16.4 (2378) 2.88 (418)
2-5R 6.35 (MP) 4.24 (0.16) 25.89 1.03 (0.41) 15.7 (2277) 2.8 (406)
2-5R 6.35 (LP) 3.79 (0.14) 27.89 1.21 (0.48) 11.9 (1726) 2.49 (361)
3-10PH 6.35 (HP) 5.59 (0.22) 18.90 0.53 (0.21) 18.6 (2697) 3.09 (448)
3-10PH 6.35 (MP) 4.91 (0.19) 20.45 0.63 (0.25) 17.9 (2595.5) 3.01 (436)
3-10PH 6.35 (LP) 4.46 (0.17) 22.27 0.74 (0.29) 16.6 (2407) 2.89 (419)
 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859 857

in higher cementitious paste thickness values and corresponding

lower porosity levels. The low cement content mixtures result in
the highest value for each data group, which results in low paste
thickness and high porosity. Noticed from the graphs and Table 5
is that the smaller aggregate 6.35-mm (1/4-in.) with higher paste
content show similar porosities to that of the larger aggregate,
9.54-mm (3/8-in) with lower paste content. An example of this
case is the Unc 6.35 (HP) sample with a porosity of 26.41 percent
and the 2-5R 9.54 (LP) sample with a porosity of 27.10 percent.
This is noticed within the aggregate size group, such as the com-
paction energy variable and the paste content produce similar
porosities (e.g., 3-10PH 9.54 (LP) and 2-5R 9.54 (HP)). This trend
shows that there is a correlation between the cement content in
a mixture and the compaction technique used. However, it should
be pointed out that the paste thickness is a function of the surface
area of the aggregate used and the volume of paste used as articu-
Fig. 9. Relationship between the average cementitious paste thickness and the
lated in reference [17,18]. The compaction energy simply reduces
permeability for 6.35-mm (1/4-in.) aggregate samples.
the porosity and causes more ‘bridging’ of the paste between the
aggregate particles, as shown in the realistic case of Fig. 4.
paste bridging from aggregate to aggregate from the compaction
Therefore, the trend is an artifact of the measuring technique, as
energy. It can be seen that the cement content corresponds similar
large paste thicknesses are being recorded due to the bridging of
to the porosity results, such that the higher cement content results
the paste from the induced compaction energies. Lastly, for both
in higher paste thickness and lower porosity, which is expected. It
aggregate sizes the majority of the 3-10PH samples resulted in
is observed that the low cement content results in low paste thick-
lower porosity values, which is due to the extra energy exerted
ness and high porosity. Other similarities are noticed, such as the
from the proctor hammer blows. Therefore, this result is expected
approximate results between higher cement content and lower
and explained.
compaction levels and that the smaller aggregate tests result in
3.3. Permeability generally higher permeability at similar cement paste thicknesses,
due to larger active void sites.
3.3.1. Permeability vs. cementitious paste thickness
As the cementitious paste thickness increases the permeability 3.4. Compressive strength
of the pervious concrete samples decreased. Figs. 8 and 9 demon-
strate the relationship between the cementitious paste thickness 3.4.1. Compressive strength vs. cementitious paste thickness
and the permeability of the samples. With an increase in cementitious paste thickness, an increase in
Similar to the porosity, the permeability decreases with larger compressive strength is observed. Figs. 10 and 11 display the rela-
cementitious paste thickness. This result is due to the active void tionship between the cementitious paste thickness and the 28-day
sites getting smaller as the cement paste gets larger. This result compressive strength of the samples.
is as expected as the documented trend with porosity versus per- It is shown that the least compacted and lowest cement content
meability is inversely proportional [16], therefore a similar trend samples result in the lowest compressive strength samples and the
is expected for cementitious paste thickness and permeability. highest compacted and highest cement content samples result in
Similar to the porosity results, the uncompacted results show the higher compressive strength. This result is expected as the
majority of the data having the lowest paste thickness increased compaction results in a higher density sample and supe-
(2.25-mm) and the highest permeability (1.25 cm/s). The same rior aggregate/paste interlock for higher compressive strengths.
results are noticed with the 3-10PH data, such that the majority of The higher cement content results in superior bond and coating
the data points have the highest paste thickness (3.2-mm) and of the aggregates, which results in a higher compressive strength.
consequently the lowest permeability (0.50 cm/s). As discussed Similarities continue to carry over from the porosity and perme-
previously, the higher recorded paste thickness is a result of the ability analysis, such that the cement content and compaction

Fig. 8. Relationship between the average cementitious paste thickness and the Fig. 10. Relationship between the average cementitious paste thickness and the 28-
permeability for 9.54-mm (3/8-in.) aggregate samples. day compressive strength for 9.54-mm (3/8-in.) aggregate samples.
858 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859
Fig. 11. Relationship between the average cementitious paste thickness and the 28-
day compressive strength for 6.35-mm (1/4-in.) aggregate samples.
methods are relatable. It is gleaned from the figures that the smal-
ler aggregate generally has higher compressive strength values as
expected. It can be seen from the graphs that the data points seem
to be reaching an asymptote at approximately 16 MPa for the
9.54-mm (3/8-in.) samples and approximately 18 MPa for the
6.35-mm (1/4-in.) samples. A result such as this is expected, as a
linear-proportionate trend would suggest that a pervious concrete
mixture could be developed with larger cementitious paste thick-
ness to achieve higher compressive strengths. However, this can-
not be the case while still maintaining a pervious concrete
mixture. The cementitious paste thickness will reach a point where
it spans from aggregate to aggregate without voids to permit water
matriculation, which defeats the purpose of this material.
3.5. Splitting tensile strength
3.5.1. Splitting tensile strength vs. cementitious paste thickness
An increase of splitting tensile strength is recorded as the
cementitious paste thickness increases. Figs. 12 and 13 show the
relationship between the cementitious paste thickness and the
28-day splitting tensile strength for each aggregate size.
It can be seen in Figs. 12 and 13 that the splitting tensile
strength follows a similar trend to that of the compressive strength
versus cementitious paste thickness. Both trends are expected; as
the paste gets thicker the strengths should increase. It is observed
that the splitting tensile relationship data approaches an asymp-
tote at approximately 3 MPa for the 9.54-mm (3/8-in.) aggregate
Fig. 12. Relationship between the average cementitious paste thickness and the 28-
day splitting tensile strength for 9.54-mm (3/8-in.) aggregate samples.
Fig. 13. Relationship between the average cementitious paste thickness and the 28-
day splitting tensile strength for 6.35-mm (1/4-in.) aggregate samples.
and approximately 3.1 MPa for the 6.35-mm (1/4-in.) aggregate.
As with the compressive strength trend, this trend is expected, as
the cement paste gets larger. It is observed that the smaller aggre-
gate samples (6.35-mm [1/4-in.]) have a slightly higher strength
than the larger aggregate samples (9.54-mm. [3/8-in.]), which is
an anticipated trend with much documentation showing that
smaller aggregate leads to stronger concrete [4].
3.6. Data summary
In addition to the graphs produced and discussed in the afore-
mentioned sections, a table summarizing all the data collected
for this study has been developed for easier readability, which
can be seen in Table 5. It should be noted that three samples were
tested and averaged to produce each data point (i.e. three uncom-
pacted, 9.54-mm (HP), samples were tested and average to pro-
duce a single data point).
4. Conclusions
A novel characterization method has been implemented on per-
vious concrete that takes into account the effect of the amount of
cement in a mixture, the size of the aggregate, and the compaction
method to relate the cement paste thickness to the performance of
pervious concrete.
Based on the data obtained in this study, the following conclu-
sions may be drawn:
 The cement content has an effect on the cementitious paste
thickness, such that more cement content results in thicker
paste surrounding the aggregate. A relationship exists between
the aggregate size and the paste content, such that smaller
aggregate with higher paste content show similar porosities to
that of the larger aggregate with lower paste content.
 The porosity of the pervious concrete samples decreases with
an increase in cementitious paste thickness. When the paste
becomes larger, the active voids size will decrease and reduce
the porosity of the sample.
 The permeability of the pervious concrete samples decreased
with an increase in cementitious paste thickness. The larger
the paste thickness, the smaller the void space.
 The 28-day compressive strength and splitting tensile strength
increased with an increase in cementitious paste thickness up
to a point of approximately 3.5-mm (0.14-in.) for both sized
aggregates. If the paste thickness becomes too thick, the void
size is consumed resulting in a void-less concrete, which defeats
the purpose of pervious concrete.
 A. Torres et al. / Construction and Building Materials 95 (2015) 850–859
Acknowledgements
The authors gratefully acknowledge the Concrete Industry
Management (CIM) National Steering Committee (NSC) for their
invaluable financial support.
References
[1] M.E. Barrett Jr., J.F. Malina, R.J. Charbeneau, Characterization of Highway
Runoff in Austin, Texas, Area, J. Environ. Eng. 124 (2) (1998) 131–139.
[2] American Concrete Institute. Pervious Concrete. ACI 522 Committee Report.
Farmington Hills, MI, 2010.
[3] F. Montes, S. Valavala, M. Haselbach, A new test method for porosity
measurements of Portland cement pervious concrete, J. ASTM Int. 2 (1)
(2005). Paper ID JAI12931.
[4] P.D. Tennis, M.L. Leming, D.J. Akers, Pervious concrete pavements. Technical
report, EB30202, Portland Cement Association, Skokie, Illinios, and National
Ready Mixed Concrete Associated, Silver Spring, Maryland, 2004.
[5] V.R. Schaefer, K. Wang, M.T. Suleiman, J. Kevern, Mix design development for
pervious concrete in cold climates, Technical report. National Concrete
Pavement Technology Center, Iowa State Univ., Ames, Iowa, 2006.
[6] J. Kevern, K. Wang, V. Schaefer, Test methods for characterizing air void
systems in Portland cement pervious concrete, J. ASTM Int. 6 (9) (2009). Paper
ID JAI102451.
[7] J.D. Wimberly, M.L. Leming, R.A. Nunez, Evaluation of mechanical and
hydrological properties of high voids pervious concrete, North Carolina State
University, Raleigh, NC, 2001.
[8] American Concrete Institute, Pervious Concrete. ACI 522 Committee Report,
Farmington Hills, MI, 2010.
859
[9] ASTM Standard C192/C192M, Standard Practice for Making and Curing
Concrete Test Specimens in the Laboratory, Annual Book of ASTM Standards,
vol. 09.49, ASTM International, West Conshohocken, PA, 2014.
[10] N. Neithalath, M. Sumanasooriya, O. Deo, Characterizing void volume, size, and
connectivity in pervious concretes for permeability prediction, Mater. Charact.
61 (May 2010) 802–813.
[11] ASTM Standard C1754/C1754M, Standard Test Method for Density and Void
Content of Hardened Pervious Concrete, Annual Book of ASTM Standards, Vol.
09.49, ASTM International, West Conshohocken, PA, 2012.
[12] ASTM Standard C670, Standard Practice for Preparing Precision and
Bias Statements for Test Methods for Construction Materials, Annual Book of
ASTM Standards, vol. 09.49, ASTM International, West Conshohocken, PA,
2013.
[13] ASTM Standard C39/C39M, Standard Test Method for Compressive Strength of
Cylindrical Concrete Specimens, Annual Book of ASTM Standards, vol. 09.49,
ASTM International, West Conshohocken, PA, 2015a.
[14] ASTM Standard C617/C617M, Standard Test Method for Capping Cylindrical
Concrete Specimens, Annual Book of ASTM Standards, vol. 09.49, ASTM
International, West Conshohocken, PA, 2015a.
[15] ASTM Standard C496/C496M, Standard Test Method for Splitting Tensile
Strength of Cylindrical Concrete Specimens, Annual Book of ASTM Standards,
vol. 09.49, ASTM International, West Conshohocken, PA, 2011.
[16] C. Geadicke, A. Marines, F. Miankodila, A method for comparing cores and cast
cylinders in virgin and recycled aggregate pervious concrete, Const. Build.
Mater. 52 (2014) 494–503.
[17] A.M. Neville, Properties of Aggregate, Prop Concrete, fifth ed., Pearson
Education, New York, 2011. pp. 108–182 [chapter 3].
[18] Albert K.H. Kwan, Leo G. Li, Combined effects of water film, paste film and
mortar film thicknesses on fresh properties of concrete, Constr. Build. Mater.
50 (2014) 598–608.