Construction and Building Materials 113 (2016) 732–740

Contents lists available at ScienceDirect

Construction and Building Materials

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

A novel lightweight gypsum composite with diatomite

and polypropylene fibers
Osman Gencel a,⇑, Juan Jose del Coz Diaz b, Mucahit Sutcu c, Fuat Koksal d, Felipe Pedro Álvarez Rabanal b,
Gonzalo Martínez-Barrera e
a
Civil Engineering Department, Faculty of Engineering, Bartin University, 74100 Bartin, Turkey
b
Department of Construction, University of Oviedo, 33204 Gijón, Spain
c
Department of Materials Science and Engineering, Faculty of Engineering and Architecture, Izmir Kâtip Celebi University, 35620 Izmir, Turkey
d
Civil Engineering Department, Faculty of Engineering and Architecture, Bozok University, 66100 Yozgat, Turkey
e
Laboratorio de Investigación y Desarrollo de Materiales Avanzados (LIDMA), Facultad de Química, Universidad Autónoma del Estado de México, Km.12 de la carretera
Toluca-Atlacomulco, San Cayetano 50200, Mexico

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

Article history: Plaster and composites made up of gypsum have been widely used in building for different purposes. In
Received 3 February 2016 this work, a new gypsum composite containing diatomite and polypropylene fibers was produced.
Accepted 22 March 2016 Diatomite, fossilized of microscopic plants, was used up to 20% while fiber up to 1%. Samples with
120
120
20 mm and 40
40
160 mm size were produced to do tests. In addition to XRF, XRD,
TGA and SEM investigations of diatomite, physical, mechanical and thermal test were done on produced
Keywords: samples. Also, Analysis of Variance (ANOVA) test were done in order to see the effect of diatomite and
Gypsum
fibers studied on the quadratic model. It was observed that the usage of diatomite increases porosity that
Diatomite
Fibers
results in decreasing unit weight and thermal conductivity coefficient. Fibers increase mechanical
Composite properties.
Plaster Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction strengths, fibers can be introduced into the gypsum composites

Gypsum has been used as a construction material since ancient
times. The oldest applications of gypsum plaster or gypsum binder
were dated by archeologists to 7000 years BCE [1]. Gypsum has
been an important subject among many researchers due to abun-
dant availability of gypsum in nature, its ecological and technolog-
ical properties, low energy consumption of manufacturing and
many other positive properties [2]. Today, gypsum and gypsum
derived materials such as plaster, board and composite have been
increasingly used for many purposes in building construction
[3–17] since they presents a good fire resistance, thermal and
sound isolation properties etc. Gypsum products as construction
materials can be improved in point of physical, mechanical, ther-
mal and sound insulation properties. In order to improve physical,
thermal and sound insulation properties of gypsum composites,
their porosity can be used by using pore making agents such as
expanded silica gel granules and vermiculite [10,17]. To improve
their mechanical properties such as bending and compressive
⇑ Corresponding author.
E-mail address: osmangencel@gmail.com (O. Gencel).
since fibers such as glass fiber, carbon fiber, polypropylene fiber,
steel fiber, polyamide fiber, polyester fiber have been extensively
used in construction technology for their specific advantages
[3–7,18–20]. Thus, usage fiber reinforcement in gypsum compos-
ites can increase the materials strength.
Diatomite is a siliceous sedimentary rock, which consists prin-
cipally of fossilized skeletal remains of diatoms, unicellular aquatic
plants related to algae. These microscopic plants have the unique
capability of extracting silica from water to produce their skeletal
structure. When diatoms die, their skeletons settle to form a diato-
mite deposit. It is a chalk-like, soft, friable, earthy, very fine grained
rock, usually light in color. It is very finely porous and very low in
density. As the diatoms are composed of an amorphous form of
silica containing a small amount of microcrystalline material, dia-
tomite is chemically stable and inert. They are generally amor-
phous, hydrated or opaline silica (SiO2nH2O) and sometimes clay
minerals, calcite and organic components [21]. Due to the open
structure of diatom skeletons, it is a lightweight rock with enough
hardness. Diatomite is commonly found in volcanic environments,
with the deposits forming in lakes in volcanically active areas.
Turkey has abundant of diatomaceous earth deposits in Afyon
region [22]. Thus, diatomite composed of amorphous opaline silica

http://dx.doi.org/10.1016/j.conbuildmat.2016.03.125
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740 733

can be transformed into other mineral phases such as disordered Table 2

cristobalite/tridymite or quartz [21]. Diatomite has a low density Chemical composition of the diatomite (XRF analysis, wt.%).

and porous structure which is desired for thermal performance, Compounds Diatomite
fire resistance and sound absorption, that makes it attractive for SiO2 75.1
use as a lightweight construction material. Due to its low density, Al2O3 8.37
diatomite can be used as a constituent of concretes and plasters. CaO 2.35
The use of diatomite in gypsum mix should allow a reduction of MgO 0.55
Fe2O3 1.93
the density and thermal conductivity values of the gypsum com- K2O 0.58
posites. Existing studies use different fiber type, while no results Na2O 0.03
have been reported for using both diatomite and polymeric fibers TiO2 0.32
in gypsum composites. Thus, in this study, gypsum composites Loss on ignition 10.5
containing diatomite and polypropylene fibers were produced to
investigate their thermal conductivities, physical and mechanical
properties. 2.4. Preparation and testing of the gypsum composites

Experimental set and mix proportions of the gypsum composite are presented
in Table 4. The gypsum mixtures containing with diatomite and PP fibers were pre-
2. Experimental studies
pared in a mixer. The gypsum and additives were first put in the mixer and mixed
for 5 min in order to achieve a homogeneous dry mixture. Water was then added to
2.1. Gypsum
the dry mixture and mixed for 3 min so as to obtain the plaster slurry. The ratio of
plaster mix/water used was 900 g/585 ml. The water used during these experi-
A commercially available gypsum plaster mix characterized according to TS EN
ments was the room temperature of about 20 °C. Two different shaped molds with
13279-1 [23] was used in this study. The properties of gypsum used were given in
dimensions of 120
120
20 mm and 40
40
160 mm were prepared to pro-
Table 1.
duce test samples. Rectangular (40
40
160 mm) and plate shape
(120
120
20 mm) samples were formed by slurry casting for the mechanical
tests and thermal conductivity determination, respectively. For each test, three ser-
2.2. Diatomite
ies of specimens were produced. In the meantime, the slurry was compacted by a
shaker for 10 s in order to ensure complete filling of the mold. All the specimens
In this study, diatomite for gypsum composite production obtained from Selko
were kept in molds for 24 h at room temperature, and then removed from molds.
A.S. manufacturer in Bartin, Turkey was used. It was characterized for chemical
The specimens were left to dry in ambient conditions for 28 days, further dried in
(X-ray fluorescence analysis), mineralogical (X-ray diffraction analysis), thermal
an oven maintained at 40 °C for 48 h. Dry unit weight values of the samples were
(thermogravimetric analysis) and microstructural (scanning electron microscopy
measured after drying steps. Then porosity and water absorption values, mechani-
analysis) properties.
cal properties like compressive and bending strengths and thermal conductivity
The diatomite particles smaller than sieving intervals of 100-mesh (149 lm)
were determined (Table 5).
were used for mixes. The chemical and physical properties of diatomite are pre-
Compressive and bending strengths, of the gypsum composite samples were
sented in Table 2. Diatomite was added to the mixtures at ratios of 10% and 20%.
determined according to TS EN 13279-2 [24]. The dry unit weight, porosity and
The amount of diatomite listed is the ratio of the diatomite mass to the total dry
water absorption values were determined by Archimedes method (soaking period
mass.
of 24 h in water at room temperature). In order to determine the thermal conduc-
The microstructure of diatomaceous earth is shown in Fig. 1. They have differ-
tivity of the samples a TCi Thermal Conductivity Analyzer has been used. This
ent morphological structures such as cylindrical and others. Their particle sizes are
instrument has been developed by C-Therm Technologies and it is based on the
smaller than 25 lm. The individual particles commonly include submicron sized
Modified Transient Plane Source (MTPS) method, a non-destructive technique
pores. The surface area of diatomite was measured as 13.925 m2/g by a B.E.T. sur-
allowing us to obtain the thermal conductivity and effusivity of the samples tested
face area analyzer. Semi-quantitative chemical analysis of diatomite grains were
[25,26].
determined by the energy-dispersive X-ray spectroscopy (EDS). According to EDS
analysis, the chemical composition of diatomite is 92.7 ± 0.5% SiO2, 6.09 ± 0.4%
Al2O3 and 1.2 ± 0.1% CaO. Also, X-ray fluorescence (XRF) analysis of diatomite was
performed. The analysis result is given in Table 2. The diatomite includes mainly sil-
3. Results and discussion
icon oxide, also the oxides of aluminum, calcium, iron and others. Loss on ignition of
diatomite is about 10.5% by weight. 3.1. Unit weight
Fig. 2 shows X-ray diffraction analysis of the diatomite. The broad hump in the
vicinity of the cristobalite and quartz peaks (from 20° to 40°) is typical for diato-
The unit weights of produced gypsum composites were pre-
mite. This case indicates an amorphous phase (opaline silica) as well as crystalline
silica phases. sented in Fig. 4. As seen from Fig. 4, the unit weight values of the
The TGA curve of diatomite is given in the Fig. 3 where a total weight loss of composites varied between 1007.3 and 881.5 kg/m3 by depending
about 11% for 1100 °C. In the first weight loss, the dehydration of the diatomite on composition. The highest value was observed in reference gyp-
takes place between 20 and 200 °C which corresponds to the loss of free water sum mixture (A1) as 1007.3 kg/m3. The lowest value belongs to the
physically absorbed on the surfaces and in the pores of diatomite particles. The
mass losses in the range of 200–500 °C can be attributed to the removal of chemi-
mixture C3 including 20% diatomite and 1% fiber. As clearly under-
cally absorbed water molecules of amorphous opaline silica (SiO2nH2O). stood from Fig. 4, unit weight decreases in the increment of diato-
mite and fiber concentration in the mixture. The effect of diatomite
is dominant than that of fibers. The effect of fibers is only around
2.3. Polypropylene fibers
1%. The reason behind decreasing unit weight is that diatomite
Polypropylene fibers with 22 lm diameter and 12 mm length were used, their has a lower density due to its porous structure and its concentra-
properties listed in Table 3. The fibers were added into the mixtures in the ratios of tion in the mixture. When diatomite and fibers are used in the ratio
0.5 wt.% and 1 wt.%. The amount of fibers is counted as a ratio to the total dry mass. of 20% and 1%, respectively, unit weight decreases 12.5%. The rela-
tion between unit weight and porosity is presented in Fig. 5.
The ANOVA tests also were done for the unit weight data based
Table 1
The properties of gypsum used.
on the quadratic model. The design referred to quadratic model
with R-squared of 0.9983 and standard deviation of 2.94. The
Workability time (min) 60–90 ANOVA test results for the dry unit weight data are given in Table 6.
Final setting time (min) 150
Compressive strength (MPa) 2.5
The sum of squares is used as a measure of overall variability in the
Flexural strength (MPa) 1 data. Mean square values are obtained by dividing the sum of
Dry density (kg/m3) 650–1000 squares by the degrees of freedom. The model F-value of 357.07
1000 lm (% passing) 100 implies the model is significant. There is only a 0.02% chance that
150 lm (% passing) 60
a Model F-Value this large could occur due to noise. Values of
734 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740

Fig. 1. SEM images of diatomite powder used.

Table 4
Mixture proportions.

Mix code Diatomite (%) Fiber (%) Gypsum (%)

A1 0 0 100
A2 0 0.5 99.5
A3 0 1 99
B1 10 0 90
B2 10 0.5 89.5
B3 10 1 89
C1 20 0 80
C2 20 0.5 79.5
C3 20 1 79

‘‘Prob > F” less than 0.05 indicate model terms are significant. In
this case, A, B, AB and A2 are significant model terms.
Final equation in terms of coded factors for dry unit weight
(UW), Eq. (1):
 
kg
UW ¼ þ958:7  48:67  A  12:83  B  5:15  A  B
m3
Fig. 2. XRD pattern of the diatomite.  8:40  A2  0:8  B2 ð1Þ

3.2. Porosity

The porosity values of gypsum composites produced range from

%35.9 to %47.8 as presented in Fig. 6. The highest value was
observed in mixture C3 with %47.8. The lowest value belongs to
the mixture A2 including 0% diatomite and 0.5% fiber. As clearly
understood from Fig. 6, porosity increases in increment diatomite
concentration in the mixture. When diatomite is used at 20% ratio
porosity increases about %32.4. The effect of diatomite is dominant.
There is no effect of fibers in the mixture on porosity. The reason
behind increasing porosity is that diatomite has porous structure
and its concentration in the mixture. In fact, increment in porosity
results in strength loss. But, in point of thermal performance high
porosity is desired.
The design referred to quadratic model with R-squared of
0.9897 and standard deviation of 0.85. The ANOVA test results
for the porosity data are given in Table 7. The model F-value of
57.67 implies the model is significant. There is only a 0.35% chance
Fig. 3. The thermo-gravimetric analysis (TGA) of the diatomite. that a Model F-Value this large could occur due to noise. Values of
‘‘Prob > F” less than 0.05 indicate model terms are significant. In
this case, A and A2 are significant model terms. Values greater than
Table 3
0.1 indicate the model terms are not significant.
Properties of polypropylene fiber.
Final equation in terms of coded factors for porosity (P), Eq. (2):
Length Diameter Aspect ratio Density Tensile strength
(mm) (lm) (length/diameter) (g/cm3) (N/mm2) Pð%Þ ¼ þ44:31 þ 5:63  A þ 0:55  B þ 0:05  A  B  2:77
12 22 545 0.91 300–400  A2 þ 0:28  B2 ð2Þ
 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740 735

Table 5
Experimental results of the gypsum based composite samples.

Mix code Unit weight Porosity (%) Water Ultrasound (m/s) Bending Compressive Thermal
(kg/m3) absorption (wt.%) strength (MPa) strength (MPa) conductivity (W/m K)
A2 999.0 35.9 39.2 579.7 0.8 1.1 0.494
A3 989.0 36.3 39.6 575.9 1.0 1.4 0.495
B1 967.7 43.1 43.8 542.8 0.5 1.4 0.464
B2 958.9 44.6 44.8 540.9 0.6 1.6 0.450
B3 947.9 45.8 45.2 538.4 0.7 1.8 0.455
C1 920.4 47.4 50.4 508.4 0.8 1.6 0.390
C2 901.4 46.9 51.3 509.4 0.9 1.9 0.387
C3 881.5 47.8 52.1 508.1 1.2 2.3 0.384

Fig. 4. Response surfaces for the model given in Eq. (1) for the unit weight.

value was observed in mixture C3 with %52.1. The lowest value

Fig. 5. The relation between unit weight and porosity.
3.3. Water absorption
The water absorption values of gypsum composites produced
vary between %39.2 and %52.1 as presented in Fig. 7. The highest
belongs to the mixture A2. Use of diatomite at %20 ratio in the
mixture increases water absorption up to %29.9. This can be
attributed to the porous structure of diatomite and porosity.
There is a relation between water absorption and porosity. While
the porosity increases the water absorption increases as seen
from Fig. 8. There is no effect of fibers in the mixture on water
absorption of composite as seen clearly ANOVA analysis given
in Table 8.
The design referred to quadratic model with R-squared of
0.9974 and standard deviation of 0.42. The ANOVA test results
for the water absorption data are given in Table 8. The model
F-value of 231.21 implies the model is significant. There is only a
0.04% chance that a Model F-Value this large could occur due to
noise. Values of ‘‘Prob > F” less than 0.05 indicate model terms
are significant. In this case, A is significant model term. Values
greater than 0.1 indicate the model terms are not significant.
Final equation in terms of coded factors for water absorption
(WA), Eq. (3):

Table 6
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the dry unit weights with R2 = 0.9983.

Source Sum of squares df Mean square F value p-value Prob > F

Model 15447.32 5 3089.46 357.07 0.0002 Significant
A: Diatomite (%) 14210.67 1 14210.67 1642.43 <0.0001
B: PP fibers(%) 988.17 1 988.17 114.21 0.0018
AB 106.09 1 106.09 12.26 0.0394
A2 141.12 1 141.12 16.31 0.0273
B2 1.28 1 1.28 0.15 0.7262
Residual 25.96 3 8.65
Cor total 15473.28 8
736 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740

Fig. 6. Response surfaces for the model given in Eq. (2) for the porosity.

Table 7
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the porosities with R2 = 0.9897.

Source Sum of squares df Mean square F value p-value Prob > F

Model 207.70 5 41.54 57.67 0.0035 Significant
A: Diatomite (%) 190.41 1 190.41 264.32 0.0005
B: PP fibers(%) 1.81 1 1.81 2.52 0.2106
AB 0.01 1 0.01 0.014 0.9137
A2 15.31 1 15.31 21.25 0.0192
B2 0.16 1 0.16 0.22 0.6691
Residual 2.16 3 0.72
Cor Total 209.86 8

WAð%Þ ¼ þ44:53 þ 5:82  A þ 0:43  B þ 0:55  A  B þ 0:85 3.5. Bending strength

 A2 þ 0:1  B2
3.4. Ultrasound
As seen in Fig. 9, ultrasound values range from 508.1 m/s to
579.7 m/s. The A2 had the highest value and addition of diatomite
decreases the velocity. When diatomite is used at %20 ratio, ultra-
sound changes about %11.8 compare to one with no diatomite.
Ultrasound pulse velocity is related to the porosity of media in
which pulse travels. If the media has a high porosity the pulse
takes the distance in long time. The diatomite addition increases
the porosity. In this respect, diatomite is an important factor on
ultrasound.
The design referred to quadratic model with R-squared of
0.9987 and standard deviation of 1.77. The ANOVA test results
for the ultrasound data are given in Table 9. The model F-value
of 453.61 implies the model is significant. There is only a 0.02%
chance that a Model F-Value this large could occur due to noise.
Values of ‘‘Prob > F” less than 0.05 indicate model terms are
significant. In this case, A is significant model term. Values greater
than 0.1 indicate the model terms are not significant.
Final equation in terms of coded factors for ultrasound (U),
Eq. (4):
Uðm=sÞ ¼ þ541:82  34:33  A  0:85  B þ 0:025  A  B
þ 2:27  A2  1:68  B2
ð3Þ
Polypropylene (PP) fibers are widely used in concrete technol-
ogy to increase strength and durability. However, the stress
transfer between matrix and fibers is complicated and models
dealing with effects of the different fiber ratios on the mechanical
properties of gypsum-reinforced composite were not available. PP
fibers were added to the gypsum composite mixtures to improve
the mechanical properties. As seen clearly from Fig. 10, PP fibers
addition increases bending strength of gypsum composite from
0.4 MPa to 1.2 MPa which is corresponding %200 increments
depending on fiber content. The local response of fiber/matrix
interface during fracture is of great importance. The most impor-
tant effect of fibers is to act as bridging ligaments in the crack plane
in order to limit crack propagation and opening, increasing the
deformation energy needed to reach fracture [27]. As shown in
Anova analysis, effect of diatomite on bending strength of gypsum
composites is not significant.
The design referred to quadratic model with R-squared of
0.9133 and standard deviation of 0.12. The ANOVA test results
for the bending strength data are given in Table 10. The model
F-value of 6.32 implies the model is significant. There is only a
8.00% chance that a Model F-Value this large could occur due to
noise. Values of ‘‘Prob > F” less than 0.05 indicate model terms
are significant. In this case, B is significant model term. Values
greater than 0.1 indicate the model terms are not significant.
Final equation in terms of coded factors for bending strength
ð4Þ (BS), Eq. (5):
 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740 737

Fig. 7. Response surfaces for the model given in Eq. (3) for the water absorption.

of diatomite at 20% ratio and fibers at 0.5% and 1% ratios, compres-

sive strength increases 111.1% and 155.6%.
The design referred to quadratic model with R-squared of 0.9898
and standard deviation of 0.069. The ANOVA test results for the
compressive strength data are given in Table 11. The model
F-value of 58.48 implies the model is significant. There is only a
0.34% chance that a Model F-Value this large could occur due to
noise. Values of ‘‘Prob > F” less than 0.05 indicate model terms are
significant. In this case, A and B are significant model terms. Values
greater than 0.1 indicate the model terms are not significant.
Final equation in terms of coded factors for compressive
strength (CS), Eq. (6):
Fig. 8. The relation between water absorption and porosity.
CSðMPaÞ ¼ þ1:58 þ 0:4  A þ 0:27  B þ 0:05  A  B  0:067

BSðMPaÞ ¼ þ0:6 þ 0:12  A þ 0:2  B  0:05  A  B þ 0:25  A 2

ð5Þ  A2 þ 0:033  B2 ð6Þ

3.6. Compressive strength 3.7. Thermal conductivity

The compressive strength values of gypsum composites pro-
duced vary between 0.9 MPa and 2.3 MPa as presented in Fig. 11.
The highest value was observed in mixture C3 with 2.3 MPa. The
lowest value belongs to the mixture A1. Use of diatomite at %10
ratio in the mixture increases compressive strength up to %55.6.
Use of diatomite at %20 ratio in the mixture increases compressive
strength up to %77.8. This can be attributed to pozzolanic activity
between Ca(OH)2 and diatomite. Also fibers have an effect on com-
pressive strength of gypsum composites. When fibers are used at %
0.5 and 1% ratios, compressive strength of gypsum composites
without diatomite increases 22.2% and 55.6%. When looked at
combined effect of diatomite at10% ratio and fibers at 0.5% and
1% ratios compressive strength increases 77.8% and 100%. Effect
Thermal conductivity values of gypsum composites are pre-
sented in Fig. 12. Thermal conductivity value decreases depending
on diatomite concentration in the mixture. This means that
thermal performance of composites increases up to 63.8%. The
highest value was obtained as 0.497 W/mK which belongs to A1.
The lowest value was seen in C3 as 0.384 W/mK. The reason behind
it is that density decreases significantly with increasing diatomite
in the mixture. Also diatomite increases porosity. The increment in
porosity results in decreasing thermal conductivity value. This
relation is shown in Fig. 13. This shows that diatomite can be used
to produce plaster and gypsum block element for heat isolation in
building constructions. There is no effect of fibers on thermal con-
ductivity of composites.

Table 8
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the water absorption with R2 = 0.9974.

Source Sum of squares df Mean square F value p-value Prob > F

Model 206.80 5 41.36 231.21 0.0004 Significant
A: Diatomite (%) 203.0 1 203.0 1134.79 <0.0001
B: PP fibers(%) 1.13 1 1.13 6.30 0.0870
AB 1.21 1 1.21 6.76 0.0803
A2 1.45 1 1.45 8.08 0.0655
B2 0.02 1 0.02 0.11 0.7601
Residual 0.54 3 0.18
Cor Total 207.34 8
738 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740

Fig. 9. Response surfaces for the model given in Eq. (4) for the ultrasound.

Table 9
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the ultrasound test with R2 = 0.9987.

Source Sum of squares df Mean square F value p-value Prob > F

Model 7092.95 5 1418.59 453.61 0.0002 Significant
A: Diatomite (%) 7072.67 1 7072.67 2261.58 <0.0001
B: PP fibers(%) 4.34 1 4.34 1.39 0.3240
AB 0.0025 1 0.0025 0.000799 0.9792
A2 10.28 1 10.28 3.29 0.1675
B2 5.67 1 5.67 1.81 0.2709
Residual 9.38 3 3.13
Cor Total 7102.33 8

Fig. 10. Response surfaces for the model given in Eq. (5) for the bending strength.

Table 10
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the bending strength with R2 = 0.9133.

Source Sum of squares df Mean square F value p-value Prob > F

Model 0.46 5 0.091 6.32 0.08 Not Significant
A: Diatomite (%) 0.082 1 0.082 5.65 0.0978
B: PP fibers(%) 0.24 1 0.24 16.62 0.0267
AB 0.01 1 0.01 0.69 0.4664
A2 0.13 1 0.13 8.65 0.0604
B2 5.55E-17 1 5.55E-17 3.84E-15 1.0000
Residual 0.043 3 0.014
Cor Total 0.5 8
 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740 739

Fig. 11. Response surfaces for the model given in Eq. (6) for the compressive strength.

Table 11
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the compressive strength with R2 = 0.9898.

Source Sum of squares df Mean square F value p-value Prob > F

Model 1.41 5 0.28 58.48 0.0034 Significant
A: Diatomite (%) 0.96 1 0.96 199.38 0.0008
B: PP fibers(%) 0.43 1 0.43 88.62 0.0025
AB 0.01 1 0.01 2.08 0.2452
A2 0.0088 1 0.0088 1.85 0.2674
B2 0.0022 1 0.0022 0.46 0.5456
Residual 0.014 3 0.0048
Cor Total 1.42 8

Fig. 12. Response surfaces for the model given in Eq. (7) for the thermal conductivity.

The design referred to quadratic model with R-squared of

0.9977 and standard deviation of 0.0037. The ANOVA test results
for the thermal conductivity data are given in Table 12. The model
F-value of 260.6 implies the model is significant. There is only a
0.04% chance that a Model F-Value this large could occur due to
noise. Values of ‘‘Prob > F” less than 0.05 indicate model terms
are significant. In this case, A and A2 are significant model terms.
Values greater than 0.1 indicate the model terms are not
significant.
Final equation in terms of coded factors for thermal conductiv-
ity (TC), Eq. (7):
Fig. 13. The relation between thermal conductivity and porosity.
740 O. Gencel et al. / Construction and Building Materials 113 (2016) 732–740

Table 12
Analysis of Variance (ANOVA) for Response Surface Quadratic Model for the thermal conductivity with R2 = 0.9977.

Source Sum of squares df Mean square F value p-value Prob > F

Model 0.018 5 0.00363 260.6 0.0004 Significant
A: Diatomite (%) 0.018 1 0.018 1264.13 <0.0001
B: PP fibers(%) 4.8E5 1 4.8E5 3.46 0.1599
AB 4.0E6 1 4.0E6 0.29 0.6292
A2 4.6E4 1 4.6E4 33.04 0.0105
B2 2.94E5 1 2.94E5 2.11 0.2422
Residual 4.17E5 3 1.39E5
Cor Total 0.018 8

TCðW=mKÞ ¼ þ0:45  0:054  A  0:00283  B  0:001  A
2 2
 B  0:015  A þ 0:00383  B
4. Conclusion
The key findings of the using diatomite and polypropylene
fibers in production of gypsum composites are as follows;
The unit weight of gypsum composites with diatomite
decreased significantly. Compressive strength of composites
increases depending on increment in diatomite content in the
mixture.
Fibers play an important role to modify mechanical properties
such as compressive and bending strength. When the tensile stress
is transferred to fibers, the transfer can arrest the crack propagat-
ing and opening and substantially improve the strength. Crack
control plays a crucial role in the performance life of composites.
Effect of fibers is high to increase bending strength of composites.
But, there is no effect of diatomite for bending strength. Addition of
diatomite increase porosity which is desirable in point of thermal
performance. When diatomite content is increased in gypsum
composite the thermal conductivity coefficient of composite
decreases. This means the thermal insulation performance of com-
posite increases.
References
[1] Alena Vimmrova, Martin Keppert, Lubos Svoboda, Robert Cerny, Lightweight
gypsum composites: design strategies for multi-functionality, Cement Concr.
Compos. 33 (2011) 84–89.
[2] Juris Skujans, Andris Vulans, Uldis Iljins, Aivars Aboltins, Measurements of heat
transfer of multi-layered wall construction with foam gypsum, Appl. Therm.
Eng. 27 (2007) 1219–1224.
[3] S. Eve, M. Gomina, A. Gmouh, A. Samdi, R. Moussa, G. Orange, Microstructural
and mechanical behaviour of polyamide fibre-reinforced plaster composites, J.
Eur. Ceram. Soc. 22 (2002) 2269–2275.
[4] A. Çolak, Physical and mechanical properties of polymer–plaster composites,
Mater. Lett. 60 (2006) 1977–1982.
[5] S. Gutiérrez-Gonzalez, J. Gadea, A. Rodríguez, M.T. Blanco-Varela, V. Calderon,
Compatibility between gypsum and polyamide powder waste to produce
lightweight plaster with enhanced thermal properties, Constr. Build. Mater. 34
(2012) 179–185.
[6] S. Gutierrez-Gonzalez, J. Gadea, A. Rodriguez, C. Junco, V. Calderon, Lightweight
plaster materials with enhanced thermal properties made with polyurethane
foam wastes, Constr. Build. Mater. 28 (2012) 653–658.
[7] M.H.P. Gazineu, V.A. Dos Santos, C.A. Hazin, W.E. de Vasconcelos, C.C. Dantas,
Production of polymer-plaster composite by gamma irradiation, Prog. Nucl.
Energy 53 (2011) 1140–1144.
[8] M. Murat, A. Attari, Modification of some physical properties of gypsum plaster
by addition of clay minerals, Cem. Concr. Res. 2 (1991) 378–387.
ð7Þ [9] A. Vimmrova, M. Keppert, L. Svoboda, R. Cerny, Lightweight gypsum
composites: design strategies for multi-functionality, Cement Concr. Compos.
33 (2011) 84–89.
[10] M.S. Baspınar, E. Kahraman, Modifications in the properties of gypsum
construction element via addition of expanded macroporous silica granules,
Constr. Build. Mater. 25 (2011) 3327–3333.
[11] T. Toppi, L. Mazzarella, Gypsum based composite materials with micro-
encapsulated PCM: experimental correlations for thermal properties
estimation on the basis of the composition, Energy Build. 57 (2013) 227–236.
[12] C.M. Lai, R.H. Chen, C.Y. Lin, Heat transfer and thermal storage behaviour of
gypsum boards incorporating micro-encapsulated PCM, Energy Build. 42
(2010) 1259–1266.
[13] D. Heim, J.A. Clarke, Numerical modelling and thermal simulation of PCM–
gypsum composites with ESP-r, Energy Build. 36 (2004) 795–805.
[14] A. Oliver, Thermal characterization of gypsum boards with PCM included:
thermal energy storage in buildings through latent heat, Energy Build. 48
(2012) 1–7.
[15] M. Li, Z. Wu, M. Chen, Preparation and properties of gypsum-based heat
storage and preservation material, Energy Build. 43 (2011) 2314–2319.
[16] A. Sarı, A. Bicer, Preparation and thermal energy storage properties of building
material-based composites as novel form-stable PCMs, Energy Build. 51 (2012)
73–83.
[17] O. Gencel, J.J. del Coz Diaz, M. Sutcu, F. Koksal, F.P. Alvarez Rabanal, G.
Martinez-Barrera, W. Brostow, Properties of gypsum composites containing
vermiculite and polypropylene fibers: numerical and experimental results,
Energy Build. 70 (2014) 135–144.
[18] J.A. Mohandesi, A. Sangghaleh, A. Nazari, N. Pourjavad, Analytical modeling of
strength in randomly oriented PP and PPTA short fiber reinforced gypsum
composites, Comput. Mater. Sci. 50 (2011) 1619–1624.
[19] E. Tazawa, Effect of self stress on flexural strength of gypsum-polymer
composites, Adv. Cem. Based Mater. 7 (1998) 1–7.
[20] Y.F. Wu, The effect of longitudinal reinforcement on the cyclic shear behavior
of glass fiber reinforced gypsum wall panels: tests, Eng. Struct. 26 (2004)
1633–1646.
[21] D.M. Harwood, Diatomite, 2nd ed., in: John P. Smol, Eugene F. Stoermer (Eds.),
The Diatoms: Applications for the Environmental and Earth Sciences,
Cambridge University Press, 2010.
[22] O. Unal, T. Uygunoglu, A. Yildiz, Investigation of properties of low-strength
lightweight concrete for thermal insulation, Build. Environ. 42 (2007) 584–
590.
[23] TS EN 13279-1, Gypsum binders and gypsum plasters – Part 1: Definitions and
requirements, Ankara, 2009.
[24] TS EN 13279-2, Gypsum Binders and Gypsum Plasters – Part 2: Test Methods,
Ankara, 2009.
[25] J. Cha, J. Seo, S. Kim, Building materials thermal conductivity measurement and
correlation with heat flow meter, laser flash analysis and TCi, J. Therm. Anal.
Calorim. 109 (1) (2012) 295–300.
[26] D. Mikulic, B. Milovanovic, I. Gabrijel, Analysis of Thermal Properties of
Cement Paste during Setting and Hardening Nondestructive Testing of
Materials and Structures (RILEM Book Series), vol. 6, Springer, Netherlands,
2013. 465–47.
[27] F. Koksal, O. Gencel, B. Unal, M.Y. Durgun, Durability properties of concrete
reinforced with steel-polypropylene hybrid fibers, Sci. Eng. Compos. Mater. 19
(1) (2012) 19–27.