Sie sind auf Seite 1von 8

Eur. J. Wood Prod.

DOI 10.1007/s00107-010-0434-x


Physical and mechanical properties of gypsum particleboard

reinforced with Portland cement
Raúl Espinoza-Herrera · Alain Cloutier

Received: 7 January 2009

© Springer-Verlag 2010

Abstract High thickness swelling, high water absorption Zusammenfassung Große Dickenquellung, hohe Wasser-
and low mechanical properties of gypsum particleboard aufnahme und schlechte mechanische Eigenschaften gipsge-
limit its utilization in building construction. Gypsum par- bundener Spanplatten schränken deren Verwendung im Bau-
ticleboard reinforced with Portland cement could result in wesen ein. Gipsgebundene Spanplatten verstärkt mit Port-
a product with higher mechanical properties and an accept- landzement könnten ein Produkt mit höheren mechanischen
able resistance to moisture. Physical and mechanical prop- Eigenschaften und akzeptabler Feuchtebeständigkeit erge-
erties of gypsum-cement particleboards were analyzed for ben. Untersucht wurden die physikalischen und mechani-
specimens previously conditioned at 20 °C and 60% rela- schen Eigenschaften von mit Zement verstärkten Gipsspan-
tive humidity and then soaked in water for 24 hours. The re- platten anhand von Prüfkörpern, die bei 20 °C und 60% rel.
sults showed that Portland cement incorporation increased Lf. klimatisiert worden waren und anschließend für 24 Stun-
the mechanical resistance of the boards. In the dry state, den in Wasser gelagert wurden. Die Ergebnisse zeigten, dass
Portland cement addition generated a modulus of rupture in- die mechanischen Eigenschaften der Platten durch Zugabe
crease ratio of 53% and an internal bond strength increase von Portlandzement verbessert wurden. Durch Zugabe von
ratio of 206%. Higher increase ratios were obtained after Portlandzement wurde die Biegefestigkeit im trockenen Zu-
24 hours water soaking. An increase ratio of 642% was ob- stand um 53% und die Querzugfestigkeit um 206% erhöht.
Nach 24-stündiger Wasserlagerung erhöhten sich die Wer-
tained for the modulus of rupture and 97% for hardness.
te prozentual noch stärker. Die Biegefestigkeit nahm um
Furthermore, the addition of Portland cement resulted in a
642% zu und die Härte um 97%. Des Weiteren verringer-
reduction ratio of 21% for water absorption after 2 hours
te sich durch Zugabe von Portlandzement die Wasserauf-
water soaking and 26% after 24 hours water soaking. More-
nahme nach 2-stündiger Wasserlagerung um 21% und nach
over, reduction ratios of 43% and 61% in thickness swelling
24-stündiger Wasserlagerung um 26%. Außerdem wurden
and 33% and 46% in linear variation were observed after 2
nach 2- und 24-stündiger Wasserlagerung eine Abnahme
and 24 hours water soaking. It can be concluded that Port-
der Dickenquellung um 43% bzw. 61% und der Quellung
land cement is a suitable reinforcing material for improving in Plattenebene um 33 bzw. 46% beobachtet. Daraus kann
the performance of gypsum particleboard. geschlossen werden, dass die Zugabe von Portlandzement
geeignet ist, die Eigenschaften gipsgebundener Spanplatten
Physikalische und mechanische Eigenschaften zu verbessern.
von mit Portlandzement verstärkten gipsgebundenen
1 Introduction

R. Espinoza-Herrera · A. Cloutier () Gypsum particleboard (GPB) is an inorganic-bonded panel

Centre de recherche sur le bois, Département des sciences du bois manufactured from wood particles and using gypsum as
et de la forêt 2425, rue de la Terrasse, Université Laval, Québec,
QC, G1V 0A6, Canada
adhesive. It has been primarily used for residential con-
e-mail: struction, such as wall and roof sheathing. It does not emit
Eur. J. Wood Prod.

formaldehyde, and gypsum and cement protect wood par- Considering that gypsum is used to control cement set-
ticles against fire. However, there are some drawbacks with ting and taking into account the results obtained for CPB
GPB such as high thickness swelling, high water absorption, (Okino et al. 2004, 2005), one can hypothesize that Port-
and low mechanical properties compared with those of ce- land cement could improve the moisture resistance of gyp-
ment particleboard. These factors hinder its use as building sum. This can be a good option for the elaboration of wood-
construction material (Deng and Furuno 2001). inorganic composite boards. On the basis of the character-
In order to improve the properties of GPB, some authors istics of each constituent (gypsum, cement, and wood) the
have proposed the addition of reinforcing materials. Deng development of gypsum-cement particleboards should re-
and Furuno (2001, 2002) added polypropylene (PP) fibers, sult in a product of higher mechanical properties and better
jute fibers and coconut fibers to GPB. In a posterior work, moisture resistance. However, the manufacturing of wood-
Deng et al. (2006) studied the influence of an organosil- inorganic materials composite boards could be limited by
icon agent on the physical and mechanical properties of the compatibility between wood and inorganic materials.
the boards. However, the results obtained by these authors Compatibility analysis is particularly important since ex-
were not conclusive, mainly regarding the water resistance tractive compounds can delay the hydration of the inorganic
of GPB. materials thereby affecting their adhesion to wood and the
The incorporation of cement to particleboard has been mechanical properties of the composite (Ahn and Moslemi
examined by Okino et al. (2004, 2005). They obtained bet- 1980; Simatupang et al. 1988).
ter physical and mechanical properties of cement particle- The hydration characteristics of inorganic substances
boards (CPB) than those reported by Frick (1988) for wood- have been commonly used to determine their compatibil-
gypsum composite boards. The CPB elaborated by Okino ity with wood by different methods. One of these meth-
et al. (2004, 2005) presented lower water absorption, better ods involves the use of differential scanning calorimeter
dimensional stability and higher mechanical strength than (DSC) to study the compatibility of wood-cement mixtures
for GPB, gypsum fiberboard, gypsum cardboard and gyp- (Defo et al. 2004), wood-gypsum and wood-gypsum-cement
sum flakeboard as reported by Frick (1988). mixtures (Espinoza-Herrera and Cloutier 2008). As a re-
The utilization of Portland cement in gypsum particle- sult of these studies, jack pine (Pinus banksiana) was found
boards has not been proposed yet as an alternative. Port- to be highly compatible with gypsum, Portland cement
land cement and gypsum are two substances that harden and gypsum-cement mixture (Defo et al. 2004; Espinoza-
when reacting with water. The main component of gypsum Herrera and Cloutier 2008). For that reason, jack pine was
is calcium sulfate hemihydrate (CaSO4 · 12 H2 O). Portland retained for the manufacturing of boards in the present study.
cement is mainly composed of tricalcium silicate (C3 S), di- The aim of this study was to determine whether the use
calcium silicate (C2 S), tricalcium aluminate (C3 A), tetra- of Portland cement as a reinforcing material could improve
calcium aluminoferrite (C4 AF) and calcium sulfate in low the physical and mechanical properties of gypsum particle-
proportions (Sha et al. 1999). Calcium sulfate (gypsum) boards.
is added to Portland cement in order to delay its setting.
Gypsum reacts with tricalcium aluminate to form ettringite
(Ca6 Al3 (SO4 )3 (OH)12 ·26H2 O) (Minard et al. 2007). 2 Materials and methods
While gypsum presents a low moisture resistance, Port-
land cement has low water solubility and is considered to 2.1 Materials
be moisture resistant. This characteristic makes it an impor-
tant material for building construction. Nevertheless, Port- The gypsum used for this study included at least 80% cal-
land cement is relatively expensive and it presents higher cium sulfate hemihydrate (CaSO4 · 12 H2 O) and was provided
hardening time than gypsum. Thus, if used in combination by BPB Canada Inc, Montreal, Canada. The composition of
with gypsum for panel manufacturing, the Portland cement type 10 commercial Portland cement used in this study is
proportion should be kept low. displayed in Table 1. The wood used was jack pine (Pinus

Table 1 Chemical composition

of type 10 Portland cement Chemical name Chemical Main compound Mass (%)
Tab. 1 Chemische
Zusammensetzung von Tricalcium silicate 3CaO·SiO2 C3 S 50
Portlandzement des Typs 10 Dicalcium silicate 2CaO·SiO2 C2 S 25
Tricalcium aluminate 3CaO·Al2 O3 C3 A 12
Tetrecalcium aluminoferrrite 4Ca.Al2 O3 ·Fe2 O3 C4 AF 8
Calcium sulfate CaSO4 ·2H2 O CSH2 3.5
Eur. J. Wood Prod.

banksiana) since previous work has shown its high compat- swelling, linear variation and water absorption according to
ibility with Portland cement (Defo et al. 2004), with gyp- the ASTM D1037 (2002) standard. Measurements of thick-
sum and with gypsum-cement mixture (Espinoza-Herrera ness swelling (TS) were made in the center of the specimen
and Cloutier 2008). face (c) with a digital micrometer accurate at ±0.001 mm;
whereas the linear variation (LV) was determined from the
2.2 Methods width (a) and length (b) of the specimens measured with a
digital calliper accurate at ±0.01 mm. After water soaking,
2.2.1 Boards manufacturing the specimens were dried in an oven at 103 ± 2 °C in order
to obtain the oven-dry weight. Then, the moisture content
The boards were made from 53% inorganic material, 22% was calculated from the weight after conditioning and after
jack pine particles (0.5–2.6 mm) at 10% moisture content 2 and 24 hours soaking. Thickness swelling was determined
(MC) (oven-dry basis) and 25% liquid water based on the as follows:
humid mass of the components mixture (gypsum/wood or tss − tcs
gypsum-cement/wood ratios are 2.3:1 based on oven dry TS = × 100 (1)
weight). The mats were formed from the wet furnish and
pressed at 60 °C under a pressure of 5–7 MPa during 30 min where TS: thickness swelling (%); tss : thickness of water
to a thickness of 11 mm. Two types of boards were made. soaked specimen after 2 and 24 hours (mm); tcs : thickness
The first one contained jack pine particles plus gypsum as of conditioned specimen at 20 °C and 60% RH before water
inorganic material. The second one was a mixture of gyp- soaking (mm).
sum and Portland cement in a proportion of 70:30 by dry The LV was determined as follows:
mass of powder in order to increase the mechanical proper- (a + b)ss − (a + b)cs
LV = × 100 (2)
ties and moisture resistance of the boards. The proportions (a + b)cs
were based on the results of preliminary density and hydra- where LV: linear variation (%); a: width of specimens (mm);
tion tests. Before testing, the boards were stored for at least b: length of specimens (mm); ss: water soaked specimen af-
28 days in a climate chamber at 20 °C and 60% relative hu- ter 2 and 24 hours; cs: conditioned specimen at 20 °C and
midity (RH). As gypsum-cement particleboard is a poten- 60% RH before water soaking.
tial substitute for gypsum board (GB) also known as dry-
wall panel used for the finishing of interior walls and ceil- 2.2.3 Determination of mechanical properties
ings, commercial GB laminated with paper on both faces
was used as a control in the tests performed. Gypsum board
The mechanical properties of the boards were determined
basically consists of calcium sulfate in the form of dihydrate
according to the ASTM D1037 standard. For static bending,
with overlay paper on both sides.
test specimens of GCPB, GPB and GB (76×314 mm2 ) were
used in different conditions. Specimens of the first group
2.2.2 Determination of physical properties
were conditioned in a controlled room at 20 °C and 60% RH
until they reached equilibrium moisture content. Specimens
Specimens of gypsum-cement particleboard (GCPB), GPB
of the second group were submerged in water at 20 °C for
and GB (50 × 50 mm2 ; Fig. 1) were soaked in water for
24 hours. Specimens of the third group were stored in the
2 and 24 hours at 20 °C in order to determine thickness
conditions described for the first group but paper of the type
used for gypsum board was applied to each face of GCPB
and GPB. The testing speed was 5.3 mm/min and the span
was 264 mm. According to Cramer et al. (2003) the strength
and bending stiffness of gypsum board are significantly dif-
ferent depending on the orientation of the specimen in the
board. Therefore, gypsum board specimens were oriented
parallel to the long direction of the board and perpendicular
to that direction.
Specimens of GCPB, GPB and GB (75 × 150 mm2 ) were
used to determine board Janka hardness. Two penetrations
Fig. 1 Specimen of 50 × 50 mm2 used for the determination of thick- on each of two flat faces of a specimen made from two lam-
ness swelling, linear variation and water absorption. Specimen cen- inated boards were performed at 6 mm/min. Before testing,
tre (c), width (a) and length (b)
Abb. 1 50 × 50 mm2 Prüfkörper für die Bestimmung der Dickenquel- specimens of the first group were stored at 20 °C and 60%
lung, der Quellung in Plattenebene und der Wasseraufnahme. (c) Mitte RH and those of the second group were submerged in water
des Prüfkörpers, (a) Breite, (b) Länge at 20 °C for 24 hours.
Eur. J. Wood Prod.

Table 2 Physical properties of the boards after 2 and 24 h water soaking

Tab. 2 Physikalische Eigenschaften der Platten nach 2- und 24-stündiger Wasserlagerung

Water absorption (%) Moisture content (%) Thickness swelling (%) Linear variation (%) Apparent density (kg/m3 )

GB2 h 73 A CS SS GPB24 h 5.4 A GPB24 h 1.3 A GCPBss24 h 1273 A

GB24 h 73 A 2h 24 h GPB2 h 3.3 B GPB2 h 0.9 B GCPBss2 h 1248 A
GPB24 h 50 B GB 27 120 119 GCPB24 h 2.1 C GCPB24 h 0.7 C GPBss24 h 1180 B
GPB2 h 43 C GPB 25 74 85 GB24 h 1.9 C GCPB2 h 0.6 C GPBss2 h 1160 B
GCPB24 h 37 D GCPB 18 60 64 GCPB2 h 1.9 C GB2 h −0.6 D GBss24 h 1039 C
GCPB2 h 34 E GB2 h 1.7 C GB24 h −3.2 E GBss2 h 989 D
GCPBcs 965 D
GPBcs 854 E
GBcs 573 F

Means with different letters are significantly different (p = 0.05). GPB: Gypsum particleboard; GCPB: Gypsum-cement particleboard; GB:
Gypsum-board laminated with paper on both faces; SS: specimen after 2 and 24 h water soaking; CS: specimen conditioned at 20 °C and 60%
relative humidity

Fig. 2 Reduction ratio (%) in water absorption, thickness swelling and linear variation of gypsum-cement particleboards compared to gypsum
particleboards after 2 and 24 h water soaking
Abb. 2 Prozentuale Abnahme der Wasseraufnahme, der Dickenquellung und der Quellung in Plattenebene von mit Zement verstärkten gipsge-
bundenen Spanplatten im Vergleich zu gipsgebundenen Spanplatten nach 2- und 24-stündiger Wasserlagerung

The resistance of GCPB, GPB and GB to screw with- where Zi represents the MOR, IB, Janka hardness, SW, WA,
drawal in a direction perpendicular to the face was deter- TS or LV of boards without addition of Portland cement, and
mined. Two pieces (76 × 102 mm2 ) were glued up and holes Zx is the value of the properties of the boards with Portland
of 3.2 mm in diameter were predrilled. Number 10 type AB cement.
screws were used. The testing speed was 15 mm/min. An analysis of variance was performed on the results
For evaluation of the internal bond strength of the boards, obtained for each property tested. When appropriate, the
50 × 50 mm2 specimens were used. They were bonded to means were classified using a Duncan’s multiple range test
two blocks of aluminum with an epoxy resin. The testing by means of the statistical analysis system program SAS.
speed was 0.64 mm/min.
The increase ratios (K) of modulus of rupture (MOR), in-
ternal bond strength (IB), hardness, screw withdrawal (SW), 3 Results and discussion
and the reduction ratios of water absorption (WA) thickness
swelling (TS) and linear variation (LV) for boards contain- 3.1 Water absorption (WA) and moisture content (MC)
ing Portland cement, compared to ones without Portland ce-
A significant difference in water absorption was observed
ment, were calculated using the following formula:
  for the three types of board after 2 and 24 hours of immer-
K =  − 1 × 100 (%) (3)
sion (Table 2). Gypsum board showed the highest water ab-
Z i sorption. These values were mainly due to the high porosity
Eur. J. Wood Prod.

of the board core. No differences between 2 and 24 hours 3.3 Linear variation (LV)
were observed in GB. Gypsum particleboard showed inter-
mediate values of water absorption which increased with As shown in Table 2, GPB presented higher LV than the
immersion time. Gypsum-cement particleboard showed the other composite boards studied. This parameter was sig-
lower values of water absorption. nificantly different compared to GCPB and GB. Gypsum-
After addition of Portland cement to GPB, a reduction cement particleboard had a low linear variation without sig-
ratio of 20.9% and 26.0% in water absorption after 2 hours nificant changes with soaking time, and GB presented a neg-
and 24 hours of water soaking, respectively, was observed ative linear variation due to board degradation. After the in-
(Fig. 2). These results are higher than those obtained by corporation of Portland cement to GPB, reduction ratios in
Deng and Furuno (2001, 2002) and Deng et al. (2006) who linear variation of 33.3% and 46.1% after 2 hours and 24
proposed the addition of jute fibers, polypropylene fibers hours water soaking, respectively, were obtained (Fig. 2).
(PP) and 3–5% of an organosilicon agent to GPB. After 24 In general, GCPB exhibited a better dimensional stability
hours of water soaking, these authors found a maximum re- than GB and GPB. This result is in agreement with the study
duction in WA of 3.1% for boards with jute fibers, 4.6% by Adeola and Olajide (1993) who found a better dimen-
for boards with polypropylene fibers, and 2.0% for boards sional stability as a result of an increasing cement propor-
with an organosilicon agent. Overall, cement presented bet- tion in wood-cement particleboard. The dimensional stabil-
ter characteristics as waterproof agent in GPB than these ity of GCPB could be explained on the basis of the cement
substances. chemical properties. Cement has a low solubility in water.
Cement mixed with wood particles was shown to be a Therefore, it can cover the wood particles, reduce water ab-
suitable material for the manufacturing of water resistant sorption and prevent their expansion.
boards. Indeed, Okino et al. (2005) reported that water ab-
sorption of cement particleboards after 2 and 24 hours of 3.4 Apparent density
water soaking was only 9 and 12%, respectively.
Table 2 presents boards MC as a function of water soak- Table 2 shows the apparent density values of boards at dif-
ing time. After 2 hours water soaking, GB was completely
ferent MCs. Significant differences in the apparent density
saturated and it presented a slightly higher MC than that ob-
values between GCPB, GPB, and GB were found in the air
served after 24 hours water soaking. At the end of the exper-
dry state and after 2 and 24 hours water soaking. Gypsum-
iments, GB had the highest MC, GPB presented an interme-
cement particleboard had the highest apparent density val-
diate MC and GPCB exhibited the lowest MC.
ues in the air dry state and after water soaking followed
by GPB. In contrast, GB exhibited the lowest density. The
3.2 Thickness swelling (TS)
lower apparent density of GB compared to those of GCPB
As observed in Table 2, GPB showed the higher values of and GPB is due to the manufacturing process resulting in a
TS after 2 and 24 hours water soaking. The TS values ob- higher porosity. A continuous band-type press at low pres-
tained for GCPB and GB were not significantly different sure is used for the formation of GB. In contrast, pressures
but were significantly lower than those obtained for GPB. of 5–7 MPa in a hot press were used for GCPB and GPB
These results show that Portland cement is a better water- manufacturing.
proof agent than the organosilicon and jute fibers used by All boards increased their apparent density values af-
Deng et al. (2006) and Deng and Furuno (2002). In fact, ter water soaking. However, the apparent density values
the TS reduction ratio after 2 and 24 hours water soaking of GCPB and GPB after water soaking did not present
due to the addition of cement was 43.2% and 61.3%, re- differences at 2 and 24 hours. Considering that the hu-
spectively, whereas the organosilicon agent resulted only mid masses of GPB and GCPB were the same, differences
in a maximum reduction of 5.0% (Deng et al. 2006). The in the apparent density values could be due to the differ-
addition of jute fibers reduced TS but this reduction was ent hydration products of gypsum and Portland cement.
small (Deng and Furuno 2002). Moreover, if one consid- Gypsum hardens when it reacts with water generating cal-
ers only TS, the values obtained for GCPB could be com- cium sulfate dihydrate (CaSO4 ·2H2 O; Maya 1974), whereas
pared to those reported by Okino et al. (2005) for cement Portland cement mainly generates calcium silicate hydrate
particleboard (CPB). Thus, after 2 hours and 24 hours water (3CaO·2SiO2 ·4H2 O) and calcium hydroxide (Ca(OH)2 ; Sha
soaking, the TS for GCPB were 1.9 and 2.1%, respectively, et al. 1999). The higher amount of products of hydration of
whereas CPB presented lower TS values of 1.3% and 1.8%, the cement and its different molecular weights in relation to
respectively. Moreover, TS obtained for GCPB after 2 hours hydration products of gypsum could explain the higher ap-
of immersion was lower than that reported by Frick (1988) parent density of GCPB in relation to GPB.
for GPB (8%), gypsum fiberboard (3%), gypsum cardboard The apparent density value obtained for GCPB
(3%), and gypsum flakeboard (2–5%). (965 kg/m3 ) while using a gypsum-cement/wood ratio of
Eur. J. Wood Prod.

Table 3 Mechanical properties of the boards

Tab. 3 Mechanische Eigenschaften der Platten

MOR (MPa) Hardness (N) Screw withdrawal (N) Internal bond (kPa)

GCPB(p) 7.3 A GCPBcs 2349 A GCPBcs 328 A GCPBcs 215 A

GB(=) 6.9 B GPBcs 2312 A GPBcs 283 A GPBcs 70 B
GPB(p) 6.0 C GCPBss24 h 818 B GBcs 211 B
GB⊥ 2.4 D GBcs 756 B
GCPBcs 1.0 E GPBss24 h 415 C
GCPBss24 h 0.9 E F GBss24 h 136 D
GPBcs 0.6 G F
GB(=)ss24 h 0.6 G
GB(⊥)ss24 h 0.5 G
GPBss24 h 0.1 H

Means with different letters are significantly different (p = 0.05). GPB: Gypsum particleboard; GCPB: Gypsum-cement particleboard; GB:
Gypsum-board laminated with paper on both faces; p: with paper on both faces; =: Tested specimens parallel to the main orientation of the board;
⊥: Tested specimens perpendicular to the main orientation of the board; ss: Specimen after 2 and 24 h water soaking; cs: Specimen conditioned at
20 °C and 60% relative humidity

Fig. 3 Increase ratios (%) for static bending modulus of rupture (MOR), internal bond strength (IB), Janka hardness and screw withdrawal (SW)
after Portland cement addition to gypsum particleboard. SS: soaked specimen. CS: air conditioned specimen
Abb. 3 Prozentuale Zunahme der statischen Biegefestigkeit (MOR), der Querzugfestigkeit (IB), der Janka Härte und des Schraubenauszugswi-
derstands (SW) nach Verstärkung der gipsgebundenen Spanplatten mit Portlandzement. SS: in Wasser gelagerter Prüfkörper, CS: bei 20 °C/60%
rel. Lf. klimatisierter Prüfkörper

2.3:1 (based on oven dry weight) is lower than that of CPB by Deng and Furuno (2001, 2002) (1200 kg/m3 with a gyp-
(1300 kg/m3 ) made from cypress particles (Cupressus ssp.) sum/wood ratio of 4:1) and GPB reported by Lempfer et
using a cement/wood ratio of 4:1 by weight (Okino et al. al. (1990) (1100 to 1200 kg/m3 with a gypsum/wood ra-
2005). However, this value was similar to that of CPB (800– tio of 4:1) its apparent density is lower. Thus, it can be
1000 kg/m3 ) elaborated from a tropical wood (Terminalia observed that the apparent density varies as a function of
superba) as reported by Adeola and Olajide (1993) (ce- the inorganic-binder/wood ratio, the manufacturing process,
ment/wood ratio 1.5:1 by weight). On the other hand, the ap- and the target density.
parent density of GCPB made in the current study is higher
than that of cement-bonded excelsior boards (513 kg/m3 ) 3.5 Static bending
elaborated from cypress (Taxodium distichum) and south-
ern pine (Pinus palustris) as reported by Lee (1985) (ce- As shown in Table 3, GCPB with paper on each face
ment/wood ratios 2:1 by weight). If one compares the den- (GCPB(p) ) presented the highest MOR in static bending fol-
sity of GCPB to those of gypsum particleboards obtained lowed by GB evaluated in the parallel direction (GB(=) ),
Eur. J. Wood Prod.

GPB with paper (GPB(p) ), and GB evaluated in the perpen- 3.7 Screw withdrawal (SW)
dicular direction (GB(⊥) ). The differences between GB(=)
and GB(⊥) can be attributed primarily to machine and cross Table 3 presents the results obtained for SW resistance of
direction properties of the face paper. It is well established the boards. GCPB and GPB presented a higher SW resis-
that the machine direction of paper is significantly stronger tance than GB. This is most likely due to the wood particles
and stiffer than the cross direction (Cramer et al. 2003). present in the boards. No significant differences were found
Gypsum-cement particleboard presented a better surface in this parameter between GCPB and GPB.
quality than GPB thereby improving paper adhesion to the Resistance to screw withdrawal of GCPB (328 N) can
board, which contributed to a higher mechanical resistance be compared to values reported by Frick (1988) for boards
of GCPB. These results show the importance of paper and elaborated with gypsum. Thus, GCPB screw withdrawal was
its orientation on the mechanical properties of the boards. higher than that of gypsum cardboard (60–80 N), similar
For raw boards without paper, GCPB presented a higher to that of gypsum flakeboard (300–400 N) and lower than
MOR than GPB. Cement addition to the boards resulted in that of gypsum fiberboard (400 N) and GPB (400–800 N).
an increase ratio of 53% as depicted in Fig. 3. Similar re- In contrast, screw withdrawal resistance of GCPB and GPB
sults were obtained by Deng and Furuno (2001, 2002) who from the current study were lower than that of CPB (1500–
added 12% of PP fibers (12 mm length) and 15% of jute 2368 N; Okino et al. 2004, 2005). These results suggest that
fibers (12 mm length) to GPB. Indeed, these boards had an an increase in cement proportion could improve SW resis-
increase ratio of MOR of 66 and 55%, respectively. tance of the boards.
If one analyzes the MOR values of GCPB(p) (7.3 MPa)
3.8 Internal bond strength (IB)
and GPB(p) (6.0 MPa) obtained in this study and those re-
ported by Deng and Furuno (2001, 2002) for GPB (5.0 MPa)
The addition of cement to boards resulted in a better co-
and boards with PP fibers (8.0 MPa) and jute fibers (7.5 MPa)
hesion of wood particles than the sole utilization of gyp-
it can be observed that our results were quite similar.
sum, therefore increasing the IB values (Fig. 3). After ce-
These values are also similar to those of wood-cement
ment incorporation, a high increase ratio of internal bond
particleboard (4–10 MPa) reported by Adeola and Olajide
strength (206.4%) was observed. Deng and Furuno (2001,
(1993). However, Lee (1985) reported lower MOR values
2002) obtained lower IB increases after addition of PP fibers
for cement-bonded excelsior made from cypress (2.5 MPa)
(39.2%) and jute fibers (90%). Moreover, when they added
and cement-bonded excelsior made from southern pine an organosilicon agent to GPB (Deng et al. 2006) the IB
(2.6 MPa). increase ratio was lower (45%) than that obtained after Port-
As shown in Table 3, after 24 h of immersion in water, land cement addition. On the other hand, the IB values for
GCPB was the material showing the highest strength. In GCPB (215 kPa) and GPB (70 kPa; Table 3) were lower than
fact, immersion did not generate significant differences as that of CPB (400 kPa) obtained by Okino et al. (2005).
to static bending MOR values of GCPB. In contrast, GPB In general, the addition of Portland cement to gypsum
was highly reduced by water immersion. The increase ra- particleboards has a better impact on the mechanical proper-
tio of static bending MOR due to the addition of cement to ties of the boards than the addition of polypropylene fibers,
gypsum particleboard was 642% (Fig. 3). Gypsum boards jute fibers, and a waterproof organosilicon agent (Deng and
lost considerably their static bending MOR in both direc- Furuno 2001, 2002; Deng et al. 2006).
tions when soaked in water (Table 3).

3.6 Janka hardness 4 Conclusion

As shown in Table 3, the Janka hardness of GCPB and GPB The results obtained in this study show that the use of Port-
conditioned to 20 °C and 60% RH (GCPBcs and GPBcs ) land cement as a reinforcing material significantly improves
were not significantly different. In these conditions, GCPBcs the physical and mechanical properties of gypsum particle-
and GPBcs presented higher Janka hardness than that of board in the following ways:
GBcs . After immersion in water for 24 h, the Janka hard- • Lower water absorption;
ness of all boards was reduced. Gypsum-cement particle- • Lower thickness swelling;
board Janka hardness after 24 h immersion in water was • Lower linear variation;
twice as much the value obtained for GPB after 24 h immer- • Better dimensional stability, since wood particles were
sion. Gypsum board presented the lowest Janka hardness. covered by cement, avoiding board degradation as a re-
The Janka hardness of GCPB after 24 h immersion can be sult of water immersion;
compared to dry GB conditioned at 20 °C and 60% RH. • Higher apparent density (increase from 854 to 965 kg/m3 );
Eur. J. Wood Prod.

• Higher modulus of rupture (53% increase) in dry condi- Deng Y, Furuno T (2001) Properties of gypsum particleboard rein-
tions; forced with polypropylene fiber. J Wood Sci 47(4):445–450
Deng Y, Furuno T (2002) Study on gypsum-bonded particleboard re-
• Higher modulus of rupture (642% increase) after water
inforced with jute fibers. Holzforschung 56(4):440–445
immersion during 24 h; Deng Y, Xuan L, Feng Q (2006) Effect of waterproof agent on gypsum
• Higher internal bond strength (206% increase); particleboard properties. Holzforschung 60(3):318–321
• Higher hardness, after immersion in water for 24 h. Espinoza-Herrera R, Cloutier A (2008) Compatibility of four eastern
Canadian woods with gypsum and gypsum-cement binders by
Although the addition of 30% Portland cement to gypsum isothermal calorimetry. Maderas, Cienc Tecnol 10(3):275–288
particleboard generally improved the mechanical proper- Frick E (1988) The bison system for the production of wood gyp-
ties of the boards, the screw withdrawal resistance did not sum particleboards. In: Moslemi AA, Hamel MP (eds) Interna-
tional conference on fiber and particleboard bonded with inor-
change significantly. Their higher water resistance favors the
ganic binder. Idaho, USA, pp 98–102
use of gypsum-cement particleboards. Lee AWC (1985) Bending and thermal insulation properties of cement-
bonded cypress excelsior board. For Prod J 35(11/12):57–58
Acknowledgements The authors are very grateful to the National Lempfer K, Hilbert T, Günzerodt H (1990) Development of gypsum-
Council for Science and Technology (CONACYT) of Mexico for a bonded particleboard manufacture in Europe. For Prod J
Ph.D. scholarship to Raúl Espinoza Herrera. 40(6):37–40
Maya GG (1974) Materiales de construcción. McGraw-Hill, México.
212 pp (in Spanish)
References Minard H, Garrault S, Regnaud L, Nonat A (2007) Mechanisms and
parameters controlling the tricalcium aluminate reactivity in the
Adeola FJ, Olajide A (1993) Bending strength and dimensional stabil- presence of gypsum. Cem Concr Res 37(10):1418–1426
ity of tropical wood-cement particleboard. Bioresource Technol Okino EYA, de Souza MR, Santana MAE, da Alves MV, de Sousa
44(1):77–79 ME, Teixeira DE (2004) Cement-bonded wood particleboard
Ahn WY, Moslemi AA (1980) SEM examination of wood-Portland with a mixture of eucalypt and rubberwood. Cem Concr Comp
cement bonds. Wood Sci 13(2):77–82 26(6):729–734
American Society for Testing and Materials (2002) ASTM D 1037- Okino EYA, de Souza MR, Santana MAE da Alves MV, de Sousa ME,
93 Standard test method for evaluating properties of wood-base Teixeira DE (2005) Physico-mechanical properties and decay re-
fiber and particle panel materials. In: Annual Book of ASTM, sistance of Cupressus spp. cement-bonded particleboards. Cem
vol 04.09. ASTM, Philadelphia Concr Comp 27(3):333–338
Cramer SM, Friday OM, White R, Sriprutkiat G (2003) Mechani- Sha W, O’Neill EA, Guo Z (1999) Differential scanning calorimetry
cal properties of gypsum board at elevated temperatures. In: Fire study of ordinary Portland cement. Cem Concr Res 29(9):1487–
and materials 8th international conference, January 27–38, Inter- 1489
science Communications Limited, San Francisco, CA, London, Simatupang MH, Lange H, Kasim A, Seddig N (1988) Influence of
pp 33–42 wood species on the setting of cement and gypsum. In: Moslemi
Defo M, Cloutier A, Riedl B (2004) Wood-cement compatibility of AA, Hamel MP (eds) (1989) International conference on fiber and
some Eastern Canadian woods by isothermal calorimetry. Forest particleboard bonded with inorganic binder. Idaho, USA, pp 33–
Prod J 54(10):49–56 42