Valorization of seashell by-products in pervious concrete pavers

Construction and Building Materials 49 (2013) 151–160

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Valorization of seashell by-products in pervious concrete pavers

0950-0618/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.conbuildmat.2013.08.017

⇑ Corresponding author. Tel.: +33 2 31 46 23 00; fax: +33 2 31 43 89 74.E-mail address: [email protected] (N. Sebaibi).

Dang Hanh Nguyen a,b, Mohamed Boutouil a, Nassim Sebaibi a,⇑, Lydia Leleyter b, Fabienne Baraud b

a Ecole Supérieure d’Ingénieurs des Travaux de la Construction de Caen (ESITC-Caen), 1, rue Pierre et Marie Curie, 14610 Epron, Franceb Normandie Univ., France, UCBN UR ABTE EA 4651, QALEA, F-14032 Caen, France

h i g h l i g h t s

� Seashell By-Products (SBP) size highly influences the granular arrangement of matrix.� The SBP can effectively replace gravel 2/6 mm for pervious concrete pavers.� SBP 2/4 mm provides a good compromise for the overall composition of the matrix.� SBP 4/6.3 mm decrease in mechanical strength and promoting the water infiltration.� The compaction pressure into a single layer of 7.4 kPa is optimal with good strength.

a r t i c l e i n f o

Article history:Received 29 May 2013Received in revised form 8 August 2013Accepted 12 August 2013

Keywords:ValorizationSeashell by-productsPervious concrete paversPermeability

a b s t r a c t

Seashell By-Products (SBP) are produced in an important quantity in France and are considered as waste.This paper studies their use as a partial replacement of aggregates in pervious concrete pavers consideredas an environmentally friendly building material. After designing the control pervious concrete pavers byinvestigating the energy and the pressure compaction, the coarse aggregate fraction were partially (20%or 40% by mass) replaced by SBP obtained from the Crepidula shell. The crushed Crepidula seashell of 2/4 mm and 4/6.3 mm were used to make new seashell by-products based pavers. In this paper, themechanical and hydrologic properties of both pervious concrete pavers were determined. Results showthat the seashell by-products have the potential to be used as aggregate. The mix design allows achievingboth a compressive strength of 16 and 15 MPa for respectively the control pervious concrete pavers andthe seashell by-products based pavers and a permeability coefficient in the range of 3–8 mm s�1.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

According to the French Union of Aggregate Producers [1], inFrance, almost 400 million tons of aggregates are consumed eachyear, i.e. 6 tons per inhabitant. However, most of the aggregatesused are natural aggregates which are usually excavated from riverbeds, quarries, dredged from sand or shingle banks under the sea.In order to reduce the environmental impact of building materialsand specially natural resource consumption, the reuse of waste andby-products is one of the solutions. For aggregates, many studies[2–6] have been carried out to investigate the replacement of nat-ural aggregates in concrete by recycled aggregates, slag aggregates,and recently seashells.

Regarding the seashells, it was reported [4–6] that they are ahard material that can produce good quality concrete, however, ahigher cement content may be required. Moreover, due to theangularity of the shells, additional cement paste is required to ob-tain the desired workability. Aggregate containing complete shells

(uncrushed) should be avoided as their presence may result invoids in the concrete and lower the compressive strength.

In Europe, France has an important fishing and shellfish farmingindustry that produces nearly 200,000 tons of shells from shellfishbreeding and nearly 50,000 tons of shellfish per year from fishing[7]. These activities generate thousands of tons of seashell by-products (empty shells) to be discharged, as they are consideredas waste. For the moment, some attempts has been made in Franceto recycle them as soil conditioner [8] or animals food [9] but noneof these attempts gave satisfaction in terms of viable and added va-lue recycling.

In this study, Seashell By-Products (SBP) from the French westcoast were prepared and used to partially replace natural aggre-gates to make a specific concrete: pervious concrete pavers. Pervi-ous concrete is an environmentally friendly material and aneffective means to meet growing environmental demands. Pervi-ous concrete is used to prevent from flooding during heavy rainand to increase the water infiltration into the soil [10,11].

Pervious concrete uses the same materials as conventional con-crete, with the exceptions that the fine aggregate is nearly or en-tirely eliminated, and the size distribution of the coarse

Table 1Physical and chemical properties of cement CEM I 52.5 R.

Chemical analysis (%) Physical properties

CaO 63.4 Specify gravity (kg m�3) 3140SiO2 19.2 Specific surface Blaine (cm2 g�1) 4900Al2O3 4.5 Compressive strength (MPa)Fe2O3 3.9 2 days 39MgO 1.1 7 days 53SO3 3.5 28 days 64K2O 0.90 Initial setting time (min) 170N2O 0.07Loss on ignition 2.6

Phase composition C3S C2S C3A C4AF68% 9% 6% 13%

Gravel 4/6.3 Sand 0/4

Crepidula 4/6.3 Crepidula 2/4

1 2 3 4 5 6 7 80

10

20

30

40

50

60

70

80

90

100

Perc

ent p

assi

ng (

%)

Grain size (mm)

Fig. 1. Grain size distribution of natural aggregate and crushed crepidula shell.

Fig. 2. Crepidula; (a) Crude crepidula; (b) Crepi

Table 2Physical and chemical properties of the crushed crepidula and aggregates.

Characteristics Crushed crepidula

D 2/4 mm D 4/6.3

Specific gravity (kg m�3) 2716 2729Water absorption (%) 2.25 2.02Chloride ion content (%) 0.065 0.065Organic matter content (%) 1.86 1.86

152 D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160

aggregate is kept narrow, allowing for relatively little particlepacking. A system of interconnected voids (15–35%) is createdresulting in a highly permeable concrete that drains very quickly[10–12]. The compressive strength of the material ranges from 2to 28 MPa [10,11]. The draining rate of pervious concrete will varywith aggregate size and density of the mixture, but will generallyfall within the range of 1.35–12.2 mm/s [11]. For parking lots, a de-sign compressive strength of about 13.8 MPa (2000 psi) is desired,and even lower strengths may be acceptable when the concrete re-ceive light vehicular loads [13].

2. Experimental program

2.1. Materials

2.1.1. Cement and aggregateThe cement used in this study is an Ordinary Porland Cement (OPC) CEM I 52.5

R. The chemical and physical properties of this cement are summarized in Table 1.This cement contains small quantities of C3A (see Table 1) that reduces its water de-mand [14] and increases the compressive strength at 7 days approximately 80–90%value at 28 days.

The alluvial quartz sand with a grain size 0/4 mm was used. This sand presentsa specific gravity of 2620 kg m�3, an absorption coefficient of 0.50% and a finenessmodulus of 2.81.

To ensure the infiltration capacity of pervious concrete, the selection of aggre-gate monogranular (single-sized aggregates) is critical to achieve the interconnec-tion of the porous system [10,11]. The size distribution of gravel and sand isgiven in Fig. 1.

The monogranular angular aggregate fraction 4/6.3 mm was employed with aspecified gravity of 2740 kg m�3 and water absorption of 0.48%.

2.1.2. Seashell by-productsTo evaluate the possible use of Seashell By-Products (SBP) as aggregate in per-

vious concrete pavers, crepidula seashell was chosen. This seashell is very abundant[15] on the on the Normandy and Brittany coasts.

Crepidula (Fig. 2a) were subjected to different preparations such as washing,grinding and screening (Fig. 2b and c) to obtain different fraction and grain size dis-tributions. In this research, we crushed the crepidula by one crusher laboratory.Effectively, crushing shells provides granular particle >63 lm and fine fraction

dula 4/6.3 mm; and (c) Crepidula 2/4 mm.

Aggregates

mm Coarse aggregate 4/6.3 mm Sand 0/4

2740 26200.48 0,50 00 0

Table 3Composition of different mixes proposed.

ID Cement(kg/m3)

Watera

(kg/m3)Gravel 4/6.3 mm(kg/m3)

Sand (kg/m3)

Crepidula 4/6.3 mm(kg/m3)

Crepidula 2/4 mm(kg/m3)

Specify density(kg/m3)

Ratebshell Theoretical initial

porosity (%)

CPCP 309.0 92.7 1452.3 72.6 – – 1926.3 0 34.4SPCP1 309.0 92.7 1161.8 72.6 290.5 – 1926.3 20 34.4SPCP2 309.0 92.7 871.4 72.6 580.9 – 1926.3 40 34.4SPCP3 309.0 92.7 1161.8 72.6 – 290.5 1926.3 20 34.3SPCP4 309.0 92.7 871.4 72.6 – 580.9 1926.3 40 34.3SPCP5 309.0 92.7 871.4 72.6 290.5 290.5 1926.3 40 34.3

CPCP: Control Pervious Concrete Pavers.SPCP: Seashell by-product based Pervious Concrete Pavers.

a Efficient water.b Weight percentage of shell on the weight of the aggregate in the control pervious concrete pavers.

D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160 153

<63 lm. In this research, a sieving was performed to separate the fine and granularparticle. The aim of the washing phase is to remove the impurities and to limit theorganic matter and chloride ions content.

The particle size distribution of crepidula is given also in Fig. 1 while Table 2summarizes the physical and chemical properties of Crepidula SBP. The specificgravity of SBP is similar to that of the natural aggregate. SBP are mainly made fromcalcium carbonate Moreover, due to its porous structure [4], the SBP absorbs morewater than natural aggregate.

2.2. Mix composition and specimen preparation

The pervious concrete strength, porosity and permeability depend, for a givenmix proportion, on the placement method. Static and vibrating methods are usedto compact pervious concrete. In one hand, if the concrete is under-compactedthe aggregates will not bond well and if, on the other hand, the concrete is over-compacted, the surface will be sealed and the pavement will not be permeable [16].

In this study, the target performances are a compressive strength of 15 MPa at28 days and a total porosity of 25% and the mix design was carried out into twosteps. Step I was dedicated to design the Control Pervious Concrete Pavers (CPCP)and to investigate the effect of the placement method on the strength. Step IIwas to investigate the effects of the incorporation of the Crepidula SBP aggregateson the properties of the Seashell by-product based Pervious Concrete Pavers (SPCP).

The CPCP and SPCP concrete mix proportions are summarized in Table 3. Forthese mixtures, cement water and sand contents and w/c ratio are kept constantwhile the gravel content varies with the incorporation of the Crepidula aggregatefractions 2/4 mm and 4/6.3 mm. The theoretical initial porosities (volume of airand water) of all the mixes (with and without the shell) are almost identical be-cause the specific gravity of the seashell is similar to the aggregates’ one. An amountof water corresponding to the absorption of SBP was added. The water added variedbetween 5% and 10% of the water efficient according to the shells content. This wasdone to ensure that the SBP will not absorb a large amount of water during castingand interfering with the w/c ratio.

All the mixtures were used to make cubic pervious concrete pavers specimens(150 � 150 � 150 mm).

As mentioned above, Step I included investigations on the placement method toobtain the specimens by studying different compaction methods. The compactionsystem is shown on Fig. 3. The cases of compaction are as following:

Fig. 3. Compaction system.

� Case no. 1: Filling in one layer, then vibration without pressure P = 0 kPa for15 s;� Case no. 2: Filling in one layer, then vibration with pressure P = 2.5 kPa for 15 s;� Case no. 3: Filling in one layer, then vibration with pressure P = 10 kPa for 15 s;� Case no. 4: Filling in one layer, then vibration with pressure P = 15 kPa for 15 s;� Case no. 5: Filling in one layer, then vibration with pressure P = 10 kPa for 30 s;� Case no. 6: Filling in three consecutive layers, then vibration for 5 s after each

layer without pressure (P = 0 kPa).

The Cases 2–5 implementation methods simulates the industrial procedureused in the precast concrete industry.

According to the variation of the 7 days compressive strength (see Section 3.2)with the different placement cases, Case no. 6 was chosen to make the laboratorypavers. Immediately after casting, all specimens were stored during 24 h in a con-trolled room maintained at 20 �C ± 2 �C and 95% ± 5% relative humidity. After 24 h,the specimens were removed from the molds and kept at the same conditions for 7and 28 days curing periods.

2.3. Test methods

2.3.1. Compressive testThe compressive strength is measured on cubic 15 � 15 � 15 cm specimens in

accordance with the European Standard EN 12390 [17]. The compressive strengthwas performed at 7 and 28 days using a constant rate loading of 0.06 MPa s�1.The reported result is an average of three to five tests.

2.3.2. Splitting testAccording to the European Standard EN 1338 [18], the splitting test was per-

formed on cubic specimens 15 � 15 � 15 cm. This test was performed at constantload rate of 0.05 MPa s�1 (Fig. 4).

2.3.3. Porosity and densityThe air void content of pervious concrete has been measured, using ‘the exper-

imental procedure’ recommended by French Association of Civil Engineering [19].The value obtained by this method is water-accessible porosity. Then, the bulk den-sity of concrete can be calculated.

The image analysis was employed to assess the pores diameter distribution andto determine the pore size distribution of control pervious concrete [20–22].

Rigidbearers

Concretespecimen

Packingpieces

F

F

Fig. 4. Scheme of splitting tension test.

100

100

100 Valve to

keep theconstantwater level

Graduated transparentpipe with an insidediameter of 99.4 mm

Inlet valve

Outletvalve

Drain pipe40

Perviousconcretespecimenenclosed inmold

Fig. 5. Device for measuring the permeability coefficient of the pervious concretepavers.

Fig. 6. Consistency of fresh concrete.

Table 5Seven day compressive strength of control pervious concrete pavers prepared withdifferent compaction methods.

Case No. 1 No. 2 No. 3 No. 4 No. 5 No. 6

Layer 1 1 1 1 1 3Pressure (kPa) 0 2.5 10 15 10 0Vibration time/layer (s) 15 15 15 15 30 5Compressive strength at

7 days (MPa)13.39 14.98 15.37 13.38 15.14 16.19

Standard deviation 0.66 0.94 0.74 0.66 0.59 0.33

0 2 4 6 8 10 12 14 1613.0

13.5

14.0

14.5

15.0

15.5

16.0

Case n° 3

Case n° 4

Case n° 2

Com

pres

sive

str

engt

h at

7 d

ays

(MPa

)

Applied pressure (kPa)

y = -0.0419x2+0.675x+13.509

R2= 0.98

7.4 kPaCase n° 1

Fig. 7. Effect of applied pressure on the compressive strength at 7 days of CPCP.

Coarseaggregate

Voids

Pastethickness

Fig. 8. Paste thickness.

154 D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160

2.3.4. Water permeability testThe water permeability of a material is the ability of water to pass through it

under the effect of a pressure gradient. It is expressed by Darcy’s relationship whichis valid in laminar flow regime [23]. Water permeability test aims to determine thecoefficient permeability K by constant and falling head method. Constant head per-meability is measured with various water levels (100 mm, 200 mm and 300 mm).This test is the most applicable for materials with relatively high permeability, nor-mally K greater than 10�2 mm s�1 and the permeability is calculated using the flow-ing formula:

K ¼ Lh� Q

Aðt2 � t1Þð1Þ

Table 4Fresh concrete.

Mix Bach type Compaction method S

CPCP Control concrete Case no. 1 – 1 layer/P = 0 kPa/15 s 0Case no. 2 – 1 layer/P = 2.5 kPa/15 s 0Case no. 3 – 1 layer/P = 10 kPa/15 s 0Case no. 4 – 1 layer/P = 15 kPa/15 s 0Case no. 5 – 1 layer/P = 10 kPa/30 s 0Case no. 6 – 3 layer/P = 0 kPa/5 s per layer 0

SPCP1 20%SBP 4/6.3 3 layers/P = 0 kPa/5 s per layer (as case n� 6) 0SPCP2 40%SBP 4/6.3 3 layers/P = 0 kPa/5 s per layer (as Case no. 6) 0SPCP3 20%SBP 2/4 3 layers/P = 0 kPa/5 s per layer (as Case no. 6) 0SPCP4 40%SBP 2/4 3 layers/P = 0 kPa/5 s per layer (as Case no. 6) 0SPCP5 40%SBP 2/6.3 3 layers/P = 0 kPa/5 s per layer (as Case no. 6) 0

where K (mm s�1) is the permeability coefficient at 20 �C, L (mm) the sample length,Q (mm3) the volume of water collected, h (mm) the head of water, t2 � t1 (s) theduration of water collection, and A (mm2) is the cross-sectional area of specimen.

Falling head permeability is measured with an initial water level h1 = 140 mm,240 mm, 340 mm and final height h2 = 50 mm. The falling head permeability test issuitable for the materials with a permeability coefficient less than 1 mm s�1.

In that case, the permeability coefficient is calculated as follows:

K ¼ Atube � LA � t � ln h1

h2

� �ð2Þ

lump (mm) Bulk density (kg m�3) Initial porosity of fresh concrete (%)

1886.37 35.751894.81 35.471902.81 35.191902.22 35.211876.15 36.101916.83 34.72

1867.70 36.401810.77 38.341888.74 35.621897.96 35.361907.26 35.00

(a) (b) (c)Fig. 9. Interaction between particle gravel under different energy compaction (a)low, (b) reasonable, and (c) excessive.

CPCP SPCP1 SPCP2 SPCP3 SPCP4 SPCP58

9

10

11

12

13

14

15

16

17

18

Com

pres

sive

str

engt

h at

28

days

(M

Pa)

Mix

Fig. 10. Compressive strength of pavers 15 � 15 � 15 cm at 28 days.

Fig. 12. SEM image of a crushed crepidula shell [24].

D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160 155

where K (mm s�1) is the water permeability, A (mm2) the cross-sectional area of thespecimen, Atube (mm2) the cross-sectional area of the standing pipe, L (mm) thelength of the specimen, t (s) the time required for the water level to fall from initialwater head to the final water head, h1 (mm) the initial height of water, and h2 (mm)is the final height of water.

The permeability coefficient of pavers is unknown, thus constant and fallingpermeability test are implemented. It is noteworthy that several authors[11,21,24] prefer to use the falling head tests than constant head load test even withpermeable materials. Fig. 5 shows, the test device for the water permeability. Dur-ing both tests, concrete samples were kept in saturated state.

3. Results and discussion

3.1. Fresh concrete

Fresh pervious concrete is characterized by having a very lowslump, even zero slumps (Fig. 6). The fresh concretes are very stiffsince they contain less water and cement paste. This characteristicmakes it suitable for products prefabrication as pavers.

Table 4 presents the test results of bulk density and initialporosity of all pervious concrete mixture at the fresh state,

Pores Gravel Crushed crepidula Pores

1cm

(a)Fig. 11. Structural arrangement of the concrete matrix with SBP (a) and schemati

according to the compaction method and the SBP content. Averagebulk density values ranged from 1810 to 1917 kg m�3. This rangeof values is comparable to the values reported in papers researchon pervious concrete, which ranged from 1600 to 2000 kg�3 [10].The American Concrete Institute (ACI) specifies that the bulk den-sity of concrete should be within 80 kg m�3 of a specified density[11]. Table 4 also shows that all the mixtures were within the80 kg m�3 of the specific density (Table 3). However, the mix con-taining 40%SBP 4/6.3 mm had unit weight that were significantlylower than 80 kg m�3 of the target value for the mix design, andtherefore did not meet ACI specification.

The initial porosity is also deduced from the bulk density of theconcrete in the fresh state. It is noted that the initial porosity variesaccording to the method of compaction and the level of substitu-tion. This porosity is comparable to that of the pervious concreteat 28 days curing (Section 3.4.2). This result seems to indicate thatone of the best measures for quality control of pervious concretemixtures is the test of weight [10].

3.2. Influence of compaction methods on compressive strength

Compaction is one of the main factors for pervious concretemanufacture. High compaction can reduce the air voids of perviousconcrete, and result in a low permeability. However, inadequatecompaction gives a loose matrix, less durable and more likely tobe detached from the surface.

The influence of the studied cases on the compressive strengthis presented in Table 5. It can be noted that the compressivestrength is higher in Case no. 6: Filling in three layers, then vibra-tion for 5 s after each layer without pressure (P = 0 kPa). The fol-lowing hypotheses could explain this phenomenon:

Cement paste

Pores

Crushed crepidula

Gravel

(b) (c)c model of this arrangement before (b) and after (c) the incorporation of SBP.

-17% -16%

-7%

+2.5%

--32%

CPCP SPCP1 SPCP2 SPCP3 SPCP4 SPCP5

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

Split

ting

tens

ile s

tren

gth

at 2

8 da

ys (

MPa

)

Mix

10 11 12 13 14 15 16 17 18

1,6

1,8

2,0

2,2

2,4

2,6

2,8

3,0

Split

ting

tens

ile s

tren

gth

at 2

8 da

ys (

MPa

)

Compressive strength at 28 days (MPa)

Rt = 0.137Rc+ 0.253

R2= 0.85

(a) (b)Fig. 13. Tensile strength by splitting test at 28 days (a) and the relationship between tensile strength and compressive strength (b).

CPCP SPCP1 SPCP2 SPCP3 SPCP4 SPCP51600

1650

1700

1750

1800

1850

1900

1950

Bul

k de

nsity

(kg

.m-3

)

Mix

Fig. 14. Bulk density for all mixtures.

156 D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160

� For the Case nos. 2–5 where the fresh concrete is filled in moldinto a single thick layer, the natural arrangement of aggregatesbecomes difficult under an important self weight of overlyinggravels, and the granular packing is not optimized.

� In case of filling by three layers, the thickness of each layer isless important. Therefore, the fresh concrete can displace easilyunder the vibration, and the air-void is easily removed to thesurface to have more compact matrix. In addition, the fresh con-crete can approach into the corners of the mold to ensure onehomogeneous matrix.

Although casting with single layer (Case nos. 2–5) results in alower compressive strength compared to Case no. 6), it has one pri-

30

31

32

33

34

35

36

Poro

sity

acc

essi

ble

to w

ater

(%

)

MixCPCP SPCP1 SPCP2 SPCP3 SPCP4 SPCP5

(a)Fig. 15. Porosity accessible to water (a) and relationsh

mordial advantage. Fig. 7 shows the effect of compaction energy onthe compressive strength of CPCP for Case nos. 1–4. A pressure of7.4 kPa seems to be the optimum pressure to obtain the highestcompressive strength. This value is about 10 times less than theminimum value recommended by experts for in situ pervious con-crete (7.4 kPa vs 69 kPa) [25]. This difference could come from thenon vibration for the in situ pervious concrete.

Generally, under a weak compaction, the packing density ofgranular matrix is low, gravel grains interlock poorly one to an-other. As a result, the contact area between the gravel is small,and therefore, their connection is small. Besides, with an incom-plete compaction, there are large pores in the concrete structure.Consequently, the pervious concrete pavers have a low compres-sive strength.

However, with excessively high pressure, the gravel grains dis-place with difficultly to get an optimum arrangement, especiallywhen the cement paste content in the CPCP is not enough to lubri-cate these grains. In fact, the volume of the cement paste in thecomposition of control pervious concrete pavers is 19.5% of the to-tal volume of concrete compared with about 25–40% in the ordin-ary concrete. Therefore, according to Section 3.1, the cementitiouspaste prepared with w/c ratio of 0.3 is very stiff and therefore it isdifficult to efficiently full lubricate the gravel. In addition, in thepervious concrete matrix, aggregates are coated by a thin layer ofcementitious paste or mortar [26]. The thickness of the paste layercan be initially calculated using one method in which we haveapproximated each aggregate and paste coating as a sphericaland the volume of compact paste is assumed zero (Fig. 8). Follow-ing this approach, the paste layer thickness of CPCP is about

30 31 32 33 34 35 36 8

9

10

11

12

13

14

15

16

17

18

Com

pres

sive

str

engt

h at

28

days

(M

Pa)

Porosity accesible to water (%)

Rc= -1.6769P

t+ 70.248

R2= 0.87

(b)ip between porosity and compressive strength (b).

Fig. 16. Representative images of 2D sections of CPCP, (a) scanned and cropped image and (b) processed image (the red areas are the pores). (For interpretation of thereferences to colour in this figure legend, the reader is referred to the web version of this article.)

2 3 4 5 60

5

10

15

20

25

30

35

40

45

50

Equivalent pore size (mm)

Freq

uenc

y di

stri

btio

n (%

)

0

10

20

30

40

50

60

70

80

90

100

Cum

ulative frequency (%)

Fig. 17. Frequency distribution of pore sizes.

D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160 157

0.32 mm. Under strong pressure, gravels penetrate easily this thinlayer and get in touch directly one with the other (Fig. 9). The ab-sence of cement paste at the contact point causes a weak bindingand thereby, a less solid matrix.

In terms of vibration time, the compressive strength of Case no.3 (when mixture is filling in one layer, then vibration with pressureP = 10 kPa for 15 s) is slightly reduced compared to that of Case no.5 (when the mixture is filling in one layer, then vibration withpressure P = 10 kPa for 30 s). Increasing the vibration time reducesthe distance between aggregates by eliminating the thin layer ofcement paste between them; consequently the concrete presentslow strength.

CPCP SPCP1 SPCP2 SPCP3 SPCP4 SPCP5 0

1

2

3

4

5

6

7

8

9

Wat

er p

erm

eabi

lity

coef

fici

ent (

mm

.s-1

)

Mix

Head H = 100 mm Head H = 200 mm Head H = 300 mm

Fig. 18. Permeability coefficient by the falling head test.

In the rest of the study, the pervious concrete pavers with orwithout seashell by product are casted in three layers and vibratedfor 5 s /layer.

3.3. Mechanical properties

3.3.1. Compressive strengthResults of mechanical test of CPCP show that the compressive

strength after 7 days reached 98.5% of that at 28 days (Rc-

7d = 16.19 MPa and Rc-28d = 16.44 MPa). This is due to the type ofcement and the low w/c ratio. Ghafori and Sebaibi reported thesame finding [27,28]. The rapid increasing of performance at earlyage allows some studies on pervious concrete to be conducted at7 days instead of 28 days of curing as for ordinary concrete [29].

The SPCP have a lower compressive strength than CPCP (seeFig. 10). Six hypotheses may explain this result:

The crushed crepidula shell is fragile than the natural gravel. Infact, the Los Angeles coefficient [30] of natural aggregate is lowerthan that of SBP, respectively 11 and 28.

The substitution of gravel with SBP can increase of the totalporosity. In fact, natural gravel has a round shape that allows anoptimum packing density of mixture. By contrast, the SBP have aflat shape, once incorporated, they play a role as a wall, will pre-vent the approach of natural aggregates and disturb the granulararrangement, thus reduce the compactness (Fig. 11). In addition,the SBP size ranging from 2 to 6.3 mm, and cannot be easily

CPCP SPCP1 SPCP2 SPCP3 SPCP4 SPCP50

1

2

3

4

5

6

7

8

9

Wat

er p

erm

eabi

lity

coef

fici

ent (

mm

.s-1)

Mix

Head H = 140 mm Head H = 240 mm Head H = 340 mm

Fig. 19. Permeability coefficient by constant head test.

30 31 32 33 34 35 362

3

4

5

6

7

8

9

Wat

er p

erm

eabi

lity

coef

fici

ent (

mm

.s-1

)

2

3

4

5

6

7

8

9

Wat

er p

erm

eabi

lity

coef

fici

ent (

mm

.s-1

)

Porosity accessible to water (%)30 31 32 33 34 35 36

Porosity accessible to water (%)

K c= 1.077Pt - 29.88

R2= 0.83 K c= 0.564Pt - 15.20

R2= 0.91

(a) (b)Fig. 20. Relationship between the porosity accessible to water Pt and the water permeability K obtained by the falling head test (a) and the constant head test (b).

158 D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160

inserted in inter granular pores; those characterized by a meandiameter of 3.10 mm (see Section 3.4.2).

The crushed crepidula shells are flats, the flakiness index of nat-ural aggregates 4/6.3 mm is 20.1 instead of 73 for the case of SBP 4/6.3 mm. These values were determined through the Europeanstandard EN 933-03 [31]. Hence the surface area of SBP is greaterthan that of natural aggregate. As a consequence, with the sameamount of cement paste, the coating of cement paste around thegrains of natural gravel and SBP for SPCP is more slight, probablythe gravel are not fully covered. One incorrect paste thicknesscauses a weak bond of the matrix and therefore, the strengthwas worse.

The high content of organic matter and chlorides can disruptthe hydration of cement or cause adhesion defects between aggre-gate and paste, which can affect the strength of concrete.

Porous structure of SBP. The porosity of SBP is important, theporosity of fraction 2/4 mm is 6.12% and fraction 4/6.3 mm is5.51%. Scanning Electron Microscopy (SEM) observations of thecrushed crepidula shell show a heterogeneous structure (Fig. 12)[4].

The crushed crepidula shells have a strong absorption capacity.Despite the supplementary water added to take into account thischaracteristic, but a distribution of water to cement, gravel andcrushed crepidula shells according to their absorption coefficientis impractical, the heterogeneous distribution of total mixing wateris possibly, this causes a heterogeneous matrix.

The pervious concrete pavers based on SBP 4/6.3 mm have low-er compressive strength than those made from the SBP 2/4 mm.Firstly, the flat shape of SBP 4/6.3 mm results in a more fragile

10 11 12 13 14 15 16 17 182

3

4

5

6

7

8

9

Wat

er p

erm

eabi

lity

coef

fici

ent (

mm

.s-1

)

Compressive strength at 28 days (MPa)

Kv = -0.3384Rc + 8.4657

R2= 0.96

(a)Fig. 21. Relationship between the compressive strength Rc and the water permea

matrix. Secondly, the SBP 4/6.3 mm is more curved; the hollowside of the crepidula is difficult to fill with cementitious pastewhereas the amount of cement paste in concrete is limited.

It can be also noted that the incorporation of hybrid fraction (2/4 mm and 4/6.3 mm) gives a compressive strength almost equiva-lent to control pervious concrete pavers.

3.3.2. Tensile strengthThe splitting tensile strength test was conducted on three cubic

samples at 28 days curing from each mixture and the results areshown in Fig. 13. The tensile strength of pavers varies from 1.78to 2.56 MPa.

The tensile strength is proportional to the compressive strength.The relationship between the resistance Rt and Rc for all pavers isgiven by Rt = 0.137Rc + 0.253 (Fig. 13b).

The influence of the crushed crepidula size on the tensilestrength seems to be similar to that on the compressive strengthof the pavers: the bigger the crushed crepidula shell is, the lowerthe tensile strength is.

3.4. Physical properties

3.4.1. Bulk densityThe bulk density measurement shows that the concrete with or

without SBP at 28 days is lightweight with a bulk density rangingfrom 1780 kg m�3 to 1868 kg m�3 (Fig. 14). Therefore, these con-cretes are classified as lightweight concretes.

From Fig. 14, it can be seen that there is no significant differencein the bulk density between the CPCP and SPCP. This is probably

2

3

4

5

6

7

8

9

Wat

er p

erm

eabi

lity

coef

fici

ent (

mm

.s-1

)

10 11 12 13 14 15 16 17 18

Compressive strength at 28 days (MPa)

Kc= -0.6235Rc+ 14.98

R2= 0.89

(b)bility K obtained by the falling head test (a) and the constant head test (b).

D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160 159

due to the fact that the specific gravity of the SBP is similar to thatof the natural gravel (see Table 2). A slight change in density mightbe due to the variation of concrete porosity.

3.4.2. PorosityThe porosity was determined using the volumetric method. In

total, 18 samples were investigated for the six mixtures, whichrepresent three analyses for each mixture.

Fig. 15 shows the porosity for all mixtures. It can be noted thatthe porosity are very high, vary from 31.8% to 34.9%. The volume ofvoid increases with the increasing of the replacement content ofSBP 4/6.3 mm. The replacement of gravel by SBP disturbs the gran-ular arrangement of concrete. The highest porosity is obtained inthe case of 4/6.3 mm SBP.

A linear relationship between porosity and compressivestrength of the concrete can be deduced from both porosity andcompressive strength results (Fig. 15b): Rc = �1.6769 � P + 70.248where P is the porosity of the sample.

The pore size was estimated using the analysis images. Accord-ing to Fig. 16, the pore space in pervious concretes is seldom reg-ular. In this method, the pores were considered as the ellipsesand the maximum and minimum diameters of the selected poreswere recorded.

Following the result of image analysis, the characteristic poresize of control pervious concrete is 3.1 mm (see Fig. 17). This valuecomplies with the linear relationship found by Neithalath betweenthe aggregate size (5.15 mm in this research) and characteristicpores size Dp: Dp = 1.44 + 0.36 � Daggregate [21]. This characteristicpore size value shows that the SBP 4/6.3 mm cannot insert intothese pores. On the contrary, a part of SBP 2/4 mm can fit into thesepores. This could be an explanation of the better behavior of thepervious concrete made with seashell by-products 2/4 mm.

3.5. Water permeability

The main purpose of pervious concrete is to achieve an ade-quate porosity and a continuous network so that water can easilypass through. As prescribed in the technical specification PTV122 [32] the permeability of pavers must be at least5.4 � 10�2 mm/s.

Figs. 18 and 19 show the permeability coefficient obtained bythe falling head test and the constant for all mixes of Table 3.According to these figures, the permeability increases with in-crease of the amount of SBP introduced. These results can be inter-preted by the porosity of the material which increases with thepercentage of SBP introduced (Fig. 20). The infiltration rate of con-crete is also connected to the porous network and pore size. Thesecomponents of the porous structure will be examined in futurestudies.

Permeability is also related to the compressive strength. Indeed,the variation of the compressive strength of pervious concrete isinversely proportional to the permeability (Fig. 21).

An important difference of the permeability coefficient K ob-tained by two methods was observed. Indeed, the falling head testis suitable for materials with low permeability; while the concretestudied in this research have a high permeability.

The permeability coefficients are consistent with porosity, rang-ing from 3 mm/s to 8.4 mm/s. Other authors show similar resultswith typical permeability ranging from 2 mm/s to 12 mm/s[10,12,28,33]. For example, from the falling head test, Tenniset al. [10] reported one permeability coefficient of 6 mm/s and33% for porosity. According to Fig. 19 for the case SPCP3 and SPCP4,our results show the same value which is around 6 mm/s. How-ever, our order of magnitude differs from the result reported bysome other researchers. According to Zouaghi et al. [24] and Mein-inger [34], with a porosity of 35%, permeability coefficient ranges

from 35 mm/s [34] to 40 mm/s [24]. However, the total porosityis not the only factor influencing the permeability. Zouaghi andMeininger used large gravel (5–40 mm) for their pervious concretewith larger pore size than in our work.

The value of permeability coefficient is influenced by thechange of head loss in the case of the constant head test(Fig. 19). Clearly, permeability coefficient decreases with increas-ing water head of 100 mm to 300 mm whatever the SBP quantity.This is probably caused by change in water flow pattern from stea-dy to turbulent when the piezometric head is important. The localcircular currents develop and increase greatly the flow resistance.Zouaghi and Aoki had shown the same phenomenon [24,35]. Inaddition, by one synthesis of permeability coefficient variation, itis recommended that laboratory permeability test be performedat head loss of 150 mm since at this level it was ease to maintaina constant water flow.

4. Conclusions

Experimental tests performed on pervious concrete paversbased seashell by-products confirm the feasibility of valorizationof this waste as a replacement in the composition of concrete.Moreover, the diameter of Seashell By-Products ranging (SBP) be-tween 2 mm and 6 mm, these SBP are located in the granular rangeof gravel (2/6 mm). Thus, the incorporation of the SBP is achievedby substituting by weight a part of gravel by crushed shell. Two dif-ferent fractions of crushed crepidula shell were used (2/4 mm and4/6.3 mm). Then, several mixtures were selected at different quan-tity of crushed crepidula shell.

The main results of this work are:

� Pervious concrete pavers based on SBP have a mechanicalstrength comparable to that of raw pervious concrete paverswithout SBP.

� SBP size strongly influences the granular arrangement of con-crete matrix and consequently the compressive and tensilestrength. Apparently, SBP 2/4 mm provides a good compromisefor the overall composition of the matrix. In fact, with a smallersize, one part of SBP 2/4 mm can insert into the pore in decreas-ing the porosity. By contrast, category 4/6.3 mm disrupts thegranular arrangement resulting in a decrease in mechanicalstrength and promoting the water infiltration.

� The compaction methods are a major factor that influences themechanical strength of pavers. A very low or too importantcompaction decreases strongly the concrete performance. Forthe case of compaction with pressure into a single layer, a pres-sure of 7.4 kPa is optimal to have good strength. However, thefilling into three-layer consecutive gives better strength.

� In this research, pervious concrete pavers have high water per-meability due to the presence of an interconnected porous sys-tem. Water permeability varies from 3 mm/s to 8.4 mm/s. Theseresults are close to those of other authors previouslymentioned.

� The method of measurement of permeability of pervious con-crete pavers influences permeability coefficient. For the con-stant head method in laboratory, a head loss of 150 mm isrecommended since at this level it was easy to maintain a con-stant water flow.

� Porosity studied in this research is two times higher than that ofordinary concrete, which makes the concrete lighter (density1700–1800 kg m�3) with low mechanical strength.

With a very high permeability, it is possible to rework the mix-ture proportions to optimize the compactness of the material whilemaintaining permeability complies fully with the requirementsimposed by the technical specification PTV 122.

160 D.H. Nguyen et al. / Construction and Building Materials 49 (2013) 151–160

Acknowledgements

The work presented in this article is part of a collaborative pro-ject VECOP cofinanced by ERDF and the Regions of Basse-Norman-die et Bretagne (France). The authors wish to thank the co-financiers and all project partners for their support.

References

[1] National Union of Industry Careers and Building Materials. <http://www.unicem.fr> [accessed 19.12.11].

[2] Thomas C, Setién J, Polanco JA, Alaejos P, Sánchez de Juan M. Durability ofrecycled aggregate concrete. Constr Build Mater 2013;40:1054–65.

[3] Gesoglu M, Güneyisi E, Mahmood SF, Öz HÖ, Mermerdas� K. Recycling groundgranulated blast furnace slag as cold bonded artificial aggregate partially usedin self-compacting concrete. J Hazard Mater 2012;235–236:352–8.

[4] Robillard C. Etude de valorisation des coproduits marins coquilliers sous formed’additions et de granulats dans les matériaux cimentaires. Ph D thesis.University of Caen-Basse Normandie; 2012. 246p.

[5] Sugiyama M. The compressive strength of concrete containing tile chips,crushed scallop shells, or crushed roofing tiles, Gakuen Ronshu. J Hokkai-Gakuen Univ 2005;124(2005):61–9.

[6] Yang EI, Kim MY, Park HG, Yi ST. Effect of partial replacement of sand with dryoyster shell on the long-term performance of concrete. Constr Build Mater2010;24(5):758–65.

[7] France Agri Mer. Les filières pêche et aquaculture en France. Edition 2011;2011. 36p.

[8] SRC Normandie – Mer du nord. Gestion des déchets et des coproduits de laconchyliculture; 2009.

[9] Ovive. <http://www.ovive-sa.com/> [accessed 10.04.13].[10] Tennis PD, Leming ML, Akers DJ. Pervious concrete pavements. Technical

report, EB302.02. Portland Cement Association, Skokie, Illinois, and NationalReady Mixed Concrete Association. Maryland: Silver Spring; 2004. 32p.

[11] ACI 522R-10. Report on pervious concrete. Farmington Hills, Michigan:American Concrete Institute; 2010. 38p.

[12] Montes F, Haselbach LH. Measuring hydraulic conductivity in perviousconcrete. Environ Eng Sci 2006;23:960–9.

[13] Crouch LK, Sparkman A, Dunn TR, Hewitt R, Mittlesteadt W, Byard B, Pitt J.Estimating pervious PCC pavement design inputs with compressive strengthand effective void content. In: Concrete technology forum, proceedings,national ready mixed concrete association. Maryland: Silver Spring; 2006. 15p.

[14] Wuest J. Comportement structural des bétons de fibres ultra performants entraction dans des éléments composés. Ph D thesis. Ecole PolytechniqueFédérale de Lausanne; 2007. 244p.

[15] Ifremer. Programme LITEAU: La Crépidule: identifier les mécanismes de saprolifération et caractériser ses effets sur le milieu pour envisager sa gestion.Rapport final Liteau 1ère tranche, Août; 2002, 76p.

[16] Delatte N, Dan Miller D, Mrkajic A. Portland cement pervious concretepavement: field performance investigation on parking lot and roadwaypavements. Report to the RMC research & education foundation; 2007. 76 p.

[17] NF EN12390-3. Testing hardened concrete—Part 3: Compressive strength oftest specimens; 2003. 18p.

[18] NF EN1338. Concrete paving blocks – requirements and test methods; 2010.68p.

[19] AFGC–AFREM. Proceedings of technical meeting AFPC–AFREM. Toulouse,France; 1997. 238p.

[20] Huang B, Cao J, Chen X, Shu X. Laboratory and analytical study of permeabilityand strength properties of pervious concrete. National ready mix concreteassociation concrete technology forum: focus on pervious concrete. Nashville,TN; May 24–25, 2006.

[21] Neithalath N. Development and characterization of acoustically efficientcementitious materials. Ph D thesis. West Lafayette, Indiana: PurdueUniversity; 2003. 242p.

[22] Sebaibi N, Benzerzoura M, Abriaka NE, Binetruy C. Mechanical and physicalproperties of a cement matrix through the recycling of thermoset composites.Constr Build Mater 2012;34:226–35.

[23] Das BM. Principles of geotechnical engineering. 7th ed. Wadsworth, Inc.; 2010.p. 683.

[24] Zouaghi A, Nakazawa T, Imai F, Shinnishi N. Permeability of no-fines concrete.Trans Jpn Concr Inst 1998;20:31–8.

[25] Colorado Ready Mixed Concrete Association. Specifier’s guide for perviousconcrete pavement design. Version 1.2; 2009. 24p.

[26] Portland Cement Association. <http://www.cement.org> [accessed 18.12.12].[27] Ghafoori N, Dutta S. Laboratory investigation of compacted no fines concrete

for pavings materials. J Mater Civ Eng 1995;7:183–91.[28] Sebaibi N. Valorisation des composites thermodurcissables issus du recyclage

dans une matrice cimentaire: application aux betons a ultra-hauteperformance. Ph D thesis. Ecole Des Mines de Douai, France; 2011. 191p.

[29] Schaefer VR, Wang K, Suleiman MT, Kevern J. Mix design development forpervious concrete in cold climates. Technical report. National concretepavement technology center. Ames, Iowa: Iowa State Univ.; 2006.

[30] NF EN 1097-2. Tests for mechanical and physical properties of aggregates. Part2: Methods for the determination of resistance to fragmentation; 2010. 38p.

[31] NF EN 933-3. Tests for geometrical properties of aggregates. Part 3:Determination of particle shape – flakiness index; 2012. 15p.

[32] PTV122. Technical specifications for permeable concrete paving blocks andslabs. 2nd ed.; 2005.

[33] Wang K, Schaefer VR, Kevern JT, Suleiman MT. Development of mix proportionfor functional and durable pervious concrete. NRMCA concrete technologyforum: focus on pervious concrete. Nashville, TN; May 24–25, 2006.

[34] Meininger R. No-fines pervious concrete for paving. Concr Int 1988;175:20–7.[35] Aoki Y. Development of pervious concrete, master of science thesis. Sydney:

University of Technology; 2009. 138p.