Composites: Part A 101 (2017) 254–264

Contents lists available at ScienceDirect

Composites: Part A

journal homepage: www.elsevier .com/locate /composi tesa

Graphene for flame-retarding elastomeric composite foams havingstrong interface

http://dx.doi.org/10.1016/j.compositesa.2017.06.0221359-835X/� 2017 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: School of Engineering and Future Industries Institute,University of South Australia, SA 5095, Australia.

E-mail address: Jun.Ma@unisa.edu.au (J. Ma).

Sherif Araby a,c, Jihui Li b, Ge Shi a, Zheng Ma b, Jun Ma a,b,⇑a School of Engineering and Future Industries Institute, University of South Australia, SA 5095, AustraliabCollege of Chemistry and Material Science, Hebei Normal University, Shijiazhuang 050024, PR ChinacDepartment of Mechanical Engineering, Benha Faculty of Engineering, Benha University, Egypt

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

Article history:Received 2 April 2017Received in revised form 12 June 2017Accepted 18 June 2017Available online 19 June 2017

Keywords:GrapheneCompositesPolyvinyl alcoholFoams

It is a great challenge to make elastomeric polymer foams antistatic, flame-retardant and mechanicallyrobust. The challenge was addressed herein by in situ polymerizing polyvinyl alcohol, formaldehydeand graphene sheets. The graphene sheets – each in average being �5 nm thick – had a carbon to oxygenatomic ratio of 9.8 and a Raman ID/IG of 0.03. The sheets proved to react with formaldehyde building up astrong interface for the composites, and the reaction promoted the exfoliation and dispersion of graphenesheets in the matrix. They were found to create a large number of fine pores to the composites. Graphenesheets at 0.12 vol% increased the foam water retention rate from 346% to 784%. These composites had apercolation threshold of electrical conductivity at 0.023 vol%. The composites reached a limiting oxygenindex of 59.4, implying an exceptional self-extinguishing performance.

� 2017 Elsevier Ltd. All rights reserved.

1. Introduction

Of all engineering materials, polymers have seen the largestincrease in industrial applications over the past decades becauseof their relatively high specific strength and low manufacturingcosts. However, polymers are limited by high flammability [1]and inability to dissipating static electric charges [2] that relateto catastrophic explosions in underground mining and chemicalenvironments [3–10]. Therefore, it is of significance to developanti-static, flame-retarding polymers. Commercial polymers havesurface resistivity 1012–1017 X/square [2,11], and it requires 106–1012 X/square to dissipate static charge [2,11]. Utilizing functionalfillers is the most cost-effective method to make polymers anti-static and flame-retardant.

Polymeric foams are produced by introducingmicroscopic voidsand/or gas bubbles through polymerization. They are increasinglymore used in construction and packaging industries [12–15],because of favourable properties – light weight, thermal andacoustic insulation, energy absorption and cost effectiveness [16–18]. Practical applications such as packaging require these foamsto be mechanically strong, anti-static and flame-retarding; thisposes a formidable challenge.

Polyvinyl formaldehyde (PVFMH) is a product of polyvinyl alco-hol and formaldehyde by condensation. PVFMH-based foams arerelatively new in comparison with those foams based on polyur-ethane, but they are more advantageous in terms of environmentalfriendliness. Neat PVFMH [19,20] and its composite foams contain-ing alginate [21], chitosan [22] have shown promise in biomedicalindustry. A chelating PVFMH foam [23] was reported for environ-mental protection, where the foam surface was modified by amethacrylate and hydroxylamine solution via grafting polymeriza-tion; the modified foam showed adsorption capacity 73.23 mg/gfor Cu2+ from water [23].

These studies are respectable, but they are associated withthree limitations (i) low rate of water absorption, (ii) unsatisfactorymechanical robustness – a neat foam just disintegrates itself afterabsorbing and removing water for 30 cycles, and (iii) lack of anti-static and flame-retarding performance; these severely limit theapplication. A hypothesis made herein was that compoundingPVFMH with graphene sheets should address these limitations,because the sheets are exceptionally stiff and strong, have highelectrical conductivity and flame-retarding property, and wouldreact with formaldehyde likely creating a strong interface for thecomposites. CVD-grown graphene has been recently reported tocreate foams for various applications such as supercapacitors,absorbents and energy harvesting [24,25]; but its costs preventthe application in industry.

S. Araby et al. / Composites: Part A 101 (2017) 254–264 255

We in this study develop PVFMH/graphene composite foamsstarting from graphene sheets, polyvinyl alcohol and formaldehydevia in situ polymerization. The foams not only demonstrate anti-static performance and a high rate of water absorption, but remainremarkable structural integrity even through 150 cycles of absorb-ing and removing water. More importantly, limiting oxygen indexmeasurement proves the composite foams to be highly flame-retardant.

2. Materials and methods

2.1. Raw materials

Flake graphite of �160 mm in lateral dimension was purchasedfrom Baoding Carbon Co. Pty Ltd, China. Potassium dichromate(90.0 wt%), concentrated sulfuric acid (98.0 wt%), peroxyaceticacid, formaldehyde (37.0 wt%), polyvinyl alcohol (99.0 wt%) andn-pentane (95.0 wt%) were ordered from Tianjin HongguangChemical Co. Pty Ltd.

A graphite intercalation compound (GIC) was prepared accord-ing to previous studies [26,27]. In brief, 1.30 g of potassium dichro-mate, 18.40 g of sulfuric acid and 5.00 g of peroxyacetic acid werecarefully moved into a 250-ml flask that sat in a water bath withmagnetic stirring; then 5.0 g of flake graphite was added into themixture, with the temperature controlled at 45 �C for 30 min. Themixture was subsequently washed through filtration until beingneutral, followed by drying at 60 �C for 30 min; this produced agraphite intercalation compound.

The compound was thermally treated in a normal atmospheric-pressure furnace as detailed below. A crucible was preheated in thefurnace at 800 �C for 10 min. Then 1 g of GIC was transferred intothe crucible and treated in the furnace for 1 min, which created aloose pile of graphene sheets. The crucible was taken out and leftin a fume cupboard to allow the product to cool down. The processshould be conducted in a fume cupboard, and the operator mustwear safety glasses, specific heat resistant gloves, protection clothand closed shoes.

2.2. Fabrication of neat polyvinyl formaldehyde foam and its graphenecomposite foam

Polyvinyl formaldehyde (PVFMH) was synthesized takingadvantage of the reaction between the hydroxyl groups of PVAand the carbonyl groups of formaldehyde in Scheme 1; this reac-tion was reported and proved before [28]. In specific, 10.0 g of

Scheme 1. Synthesis of polyvinyl formaldehyde, where both

polyvinyl alcohol was dissolved in 90.0 g of distilled water by mag-netic stirring at 300 rpm; the polymer was observed to dissolvecompletely after 60 min. A mixture of 5.0 g concentrated sulfuricacid and 34.5 g formaldehyde solution was added into the PVAsolution, followed by 30 min of stirring at 65 �C; this produced aPVFMH solution for the following.

A PVFMH foam was fabricated by using n-pentane as a foamingagent, because n-pentane has low cost and boiling temperature.4.50 ml of n-pentane was added into the PVFMH solution withmagnetic stirring at 300 rpm and then the mixture was immedi-ately poured into a mold, followed by heat treatment in an ovenat 60 �C for 90 min. This produced a disk-like foam of �12.0 cmin diameter.

A series of PVFMH/graphene composites were prepared byin situ polymerization. As described above, 10.0 g of PVA was dis-solved in 90.0 g of distilled water, into which 5.0 g concentratedsulfuric acid, 34.5 g formaldehyde solution and a desired quantityof graphene were added, followed by 30 min of stirring at 65 �C;this produced PVFMH/graphene composites. These composite weretransformed into disk-like foams by using a similar process to thefabrication of neat PVFMH foam. Through varying the ratio of gra-phene to PVFMH, we prepared composite foams at grapheneweight fractions 0.45–2.63 wt%. In order to obtain volume fractions(Vf), Eq. (1) was adopted:

Vf ¼ qmWf

qf ð1�Wf Þ þ qmWfð1Þ

where Wf, qf and qm respectively refer to weight fraction, matrixdensity and filler density. The density values of graphene sheetsand foam were taken as 2.2 and 0.114 g/cm3. The density of neatfoam was measured based on weight and volume of three replicas.

2.3. Characterization

X-ray photoelectron spectroscopy (XPS) analysis was performedby a SPECS SAGE XPS system with a Phoibos 150 analyzer and aMCD-9 detector. Non-monochromated Mg Ka radiation was usedand set at 10 kV and 200W. A circular spot of 3 mm in diameterwas investigated. Ratios of C/O were determined by calculatingthe peak area of C1s and O1s via OriginLab� software. Raman spec-tra were obtained using a Renishaw inVia Raman microspectrom-eter by a laser excitation of 633 nm in wavelength.

Atomic force microscopy was performed to obtain the thicknessof graphene sheets by an NT-MDTSPM instrument with an NSG03non-contact golden cantilever. Graphene sheets were immersed in

intermolecular and intramolecular crosslinking occur.

256 S. Araby et al. / Composites: Part A 101 (2017) 254–264

acetone to form 0.1 wt% suspension after 30-min ultrasonication.When the mixture was diluted to 0.0001 wt%, it was dropped ontoa silicon wafer and dried at 80 �C, followed by cooling formeasurement.

X-ray diffraction analysis was performed by a laboratory-basedRigaku D/MAX Rapid II micro-diffractometer with Co Ka radiation(wavelength 1.7902 Å; accelerating voltage and filament current:40 kV and 15 mA) at a fixed incident angle of 25 deg. Since mostof the literature perform XRD using Cu Ka radiation, the obtaineddata were converted into Cu Ka–XRD patterns.

The internal structures of neat foam and its composite at0.023 vol% graphene were visualized by a scanning electron micro-scope (JSM-7800F, made in Japan). After a typical sample wasimmersed in liquid nitrogen for 1 min, it was taken out and imme-diately broken into halves by using pairs of tweezers; this processprevented any ductile deformation to the pores. The fracture sur-face was coated with platinum and observed by JSM-7800F.

Differential scanning calorimetry was conducted using a TAinstrument DSC Q10 (V9.9 Build 303). Test was performed as tem-perature increased at 10 �C/min up to 300 �C under nitrogen atmo-sphere. The onset and peak softening temperature and enthalpywere determined during the testing for all samples. Three repli-cates were tested to calculate the average property.

Water absorption capacity was tested by calculating the waterretention rate of neat foam and its composites. Dried samples weresoaked in 1000 ml of distilled water for 5 min to reach a constantweight. The wet samples were gently tapped by a soft tissue toremove any excess water, and then weighed on a high-resolutionbalance (0.1 mg). The water retention rate (Wr) was determinedusing Eq. (2).

Wr ð%Þ ¼ Ww �Wd

Wd� 100 ð2Þ

where Ww and Wd are the sample mass in wet and dry conditions.The recorded data are the average of three measurements for eachfraction.

Immediately after the foams were dried by vacuum at 80 �C for4 h, we investigated their surface area by N2 adsorption measure-ments using a micromeritics ASAP2020.

The foam surface resistivity was measured using a four-pointprobe technique – Fig. 1 – by ACL Staticide, Digital Megohmmeter(model 800, Accuracy ±10%, USA) at room temperature accordingto ASTM D257 standard. Input voltage was set at 100 V. Surfaceresistance (Rs) is the electrical resistance of a material surface irre-spective to its thickness; its measurement unit is X/square [29–31]. Basically, it is determined by the following formula [29,30]:

Rs ¼ pln 2

VI

ð3Þ

where V and I are the applied electric potential and current. Rs isrelated to the bulk resistivity (R, whose unit is X):

I

V

Composite

4-Point probe

1 23 4

Fig. 1. A four-point probe attached to the composite surface to measure surfaceresistance. (For interpretation of the references to colour in this figure legend, thereader is referred to the web version of this article.)

R ¼ Rs � t ð4Þ

Here t refers to the testing sample thickness.A sample was wiped by tissue and acetone to remove any pol-

lutants on the surface and left overnight at 60 �C to dry. Then a Dig-ital Megohmmeter was simply attached to the workpiece surfaceas schematically presented in Fig. 1. Three measurements werereported for each fraction to obtain an average value.

The limiting oxygen indexes were determined as below. Adesired shape (10 � 1 � 1 cm) was made for each foam. Each sam-ple was installed in the burning barrel of an oxygen index mea-surement instrument (XWR-2406, Changzhou Jilu ManufacturingPty Ltd.), where the ratio of oxygen to nitrogen was adjusted to findout the minimum oxygen concentration required to sustain thesample combustion. The oxygen index is often expressed as

n ð%Þ ¼ CO2

CO2 þ CN2

� 100 ð5Þ

where CO2 and CN2 are respectively the minimum concentration ofoxygen and nitrogen in the inflow gas.

Compressive testing of neat PVFMH foam and its compositeswas carried out using an Instron 5567. The test was performed ata cross-head speed of 5 mm/min [32–34] and stopped at 75% com-pression. All of compressive Young’s moduli were calculated at astrain range 19–22%. At least three samples were tested to reportaverage values.

3. Results and discussion

3.1. Graphene sheets thickness

Graphene is a single layer of graphite with a theoretical thick-ness of 0.35 nm equivalent to the carbon atom diameter. But thethickness was experimentally measured to be 0.4–1.7 nm[35,36]; the discrepancy would be caused by the selection of mea-surement methods and the thickness change during graphene fab-rication. For example, thermal shock and oxidative treatmentcaused surface corrugation increasing the thickness up to 1 nm[37,38].

The current work started from preparing a graphite intercala-tion compound. We then thermally treated the compound toobtain a loose worm-like structure demonstrating a volume expan-sion ratio over 200 times. Atomic force microscopy was conductedto measure the thickness of graphene sheets that had been dis-persed in solvent by sonication. Fig. 2a–c presents images andstatistics of graphene sheets deposited on a silicon wafer. Mostof graphene sheets each are �5 nm thick corresponding to 4–5 lay-ers of graphene. These results are in agreement with previous stud-ies [39–41], where graphite intercalation compounds werereported to yield graphene sheets, each of which consisted of 2–5 graphene layers or even more.

It is noteworthy that the Kim’s group pioneered using thermallyexpanded graphene for epoxy resins [42], where graphene sheetswere named graphite nanoplatelets whose thickness was not mea-sured. For a given volume fraction of a polymer/sheet composite,lower sheet thickness relates to a higher number of sheets thatshould have more interface with the matrix for reinforcing, tough-ening or adding functionalities to the matrix. According to ourmodelling [43], the sheets would have an obvious reinforcing ortoughening effect on polymers when their thickness is below5 nm; and vice versa. A 5-nm thick graphene sheet would consistof approximately 5 graphene layers, and its conductivity andmechanical strength should not differ significantly from single-layer graphene [44]. The lower the thickness, the higher themechanical and functional properties are. However, the cost of

(a)

(b)

(c)

0-5 5-15 15-25 25+0

5

10

15

20

25

Cou

nts

Thickness (nm)

Fig. 2. Characterization of graphene sheets by atomic force microscopy: (a) arepresentative image of graphene sheets on a silicon wafer, (b) thickness profile ofparticles selected by the green line in the image, (c) histogram of sheet number vsthickness. (For interpretation of the references to colour in this figure legend, thereader is referred to the web version of this article.)

100 200 300 400 500 600 700

Graphene sheets

C/O= 9.8

O1sXPS

cou

nts

Binding energy (eV)

C1s

(b)

100 200 300 400 500 600 700

C/O= 4.6

m-Graphene sheets

O1s

XPS

cou

nts

Binding energy (eV)

C1s

(c)

(a)

S. Araby et al. / Composites: Part A 101 (2017) 254–264 257

single-layer graphene prevents its application in polymercomposites.

Formaldehyde bridging graphene sheets with polymer matrix

Fig. 3. (a and b) X-ray photoelectron spectroscopy of pristine and modifiedgraphene, and (c) schematic of formaldehyde bridging graphene with PVFMH. (Forinterpretation of the references to colour in this figure legend, the reader is referredto the web version of this article.)

3.2. X-ray photoelectron spectroscopy (XPS)

Graphene sheets made by the expansion were analyzed by XPS.High carbon to oxygen ratio relates to a low degree of surface

defects and high structural integrity and electrical conductivity[30]. In Fig. 3, the sheets have a carbon to oxygen ratio of 9.8, cor-responding to far higher structural integrity than graphene oxide[30,45]. Graphene sheets made by expansion may carry hydroxylgroups that should react with formaldehyde. The reaction was ver-

(a)

5 10 15 20 25 30 35 40

Graphite

Graphite interclated compounds

Inte

nsity

2θ (°)

Expanded product

(b)

0 5 10 15 20 25 30 35 40

0.0 0.5 1.0 1.5 2.0 2.5 3.0Inte

nsity

2θ (°)

Neat foam

CompositeIn

tens

ity

2θ (°)

Composite

Neat foam

Fig. 5. X-ray diffraction patterns of (a) graphite, graphite intercalated compoundsand graphene sheets, and (b) neat PVFMH and its graphene composite foam at0.023 vol%. (For interpretation of the references to colour in this figure legend, the

258 S. Araby et al. / Composites: Part A 101 (2017) 254–264

ified herein by a purpose-designed experiment. In specific, thesheets were treated with formaldehyde using the same parametersas described in the Experimental for the reaction between polyvi-nyl alcohol and formaldehyde; the modified graphene sheets werethoroughly washed and filtrated to remove unreacted formalde-hyde. In Fig. 3a and b, the modified sheets have a carbon to oxygenratio of 4.6 that is far lower than that of unmodified sheets; this isbecause more oxygen-containing formaldehyde molecules graftedto graphene through the reaction between formaldehyde and thehydroxyl groups of graphene. Since formaldehyde is known toreact with polyvinyl alcohol, it is able to bridge polyvinyl alcoholwith graphene sheets during the fabrication, possibly promotingthe exfoliation and dispersion of graphene sheets in the matrix(Fig. 3c).

3.3. Raman spectroscopy

Fig. 4 shows Raman spectra of pristine and modified graphenesheets, both of which had been thoroughly washed prior to themeasurement. Both have absorptions at three locations – 1343,1586 and 2690 cm�1 respectively corresponding to D, G and 2Dbands [46]. In this case D-band refers to the vibrations of defectivesp2-hybridized carbon atoms while G-band corresponds to the in-plane stretching motion of normal sp2-hybridized atoms. Theintensity ratio of D- to G-band (ID/IG) is often used to measurethe structure integrity of graphene sheets. Our graphene sheetsshow an ID/IG of 0.03 implying high structural integrity. The chem-ical modification introduced organic groups onto graphene, leadingto an ID/IG ratio of 0.10; this provides a solid evidence again for thereaction of formaldehyde with graphene, in agreement with thepreceding XPS analysis in Fig. 3. In the current study most gra-phene sheets each contain over four layers of graphene, and thismay explain why all samples have similar 2D-bands [47].

3.4. X-ray diffraction (XRD)

In Fig. 5a, graphite shows a characteristic diffraction at �27�assigned to the graphitic plane (002); through intercalation, thediffraction moves slightly leftwards to a sharp peak at 26.5� corre-sponding to basal spacing 0.34 nm. This means that the interca-lated chemicals – potassium dichromate, sulfuric acid and

1000 1250 1500 1750 2000 2250 2500 2750 3000

Graphene sheets

Modifiedgraphene sheets

Raman shift (cm-1)

Inte

nsity

(a.u

.)

GD 2D

Fig. 4. Raman spectra of graphene sheets and modified graphene sheets. (Forinterpretation of the references to colour in this figure legend, the reader is referredto the web version of this article.)

reader is referred to the web version of this article.)

peroxyacetic acid – pose little effect on the basal spacing; thesemay intercalate to produce stages, each of which contains a fewgraphene layers. The thermal expansion at 800 �C for 1 min wasobserved to create a loose pile of black powder. In comparison withthe intensive diffraction observed for the graphite intercalationcompound, the expanded product indicates a nearly invisiblediffraction, which is explained below. The intercalated compoundsoften release immense gasses with pressure 40–100 MPa [38,48],causing the exfoliation of stages. Each stage herein is treated as agraphene sheet consisting of a few graphene layers; these stageswould randomly arrange themselves upon thermal shock, leadingto a nearly invisible diffraction [49]. The expanded sheets can read-ily suspend in various solvent [48]; a high degree of oxidationwould help to suspend in water [50].

Table 1X-ray diffraction analysis for neat and composite foam.

Sample Peak angle(�)

Peak start(�)

Peak end(�)

Heightreduction

Neat foam 0.75 0.38 1.81 As benchmark0.023 vol%

composite0.85 0.43 2.05 70%

40 60 80 100 120 140 160 180 200 220 240 260 280 300

62.55°C131.2 J/g

66.67°C141.8 J/g

109.12°C

Composite

Hea

t flo

w

Temperature (°C)

Neat foam

103.85°C

Fig. 6. DSC of neat foam and the 0.023 vol% composite. (For interpretation of thereferences to colour in this figure legend, the reader is referred to the web version ofthis article.)

1mm

(a)

1mm

(d)

15μm

(b)

15μm

(e)

1μm

(g)

Fig. 7. SEM images of neat foam (a–c) and its 0.023 vol% composite (d–g). (For interpretatversion of this article.)

S. Araby et al. / Composites: Part A 101 (2017) 254–264 259

Neat foam indicates a narrow, strong diffraction at �0.75� inFig. 5b relating to the molecular configuration of PVFMH. Throughcompounding with 0.023 vol% graphene sheets, the diffractionslightly moves rightwards to 0.85�, and the intensity is signifi-cantly reduced by �70% as detailed in Table 1. The reduced inten-sity would be caused by the introduction of graphene into thefoam, implying that graphene has changed the molecular configu-ration of PVFMH. By using software OriginLab�, we found out thatthe peaks of neat and composite foams start and end at differentdiffraction angles as presented in Table 1. The change in peak char-acteristics must relate to the graphene added, and thus these sam-ples were further investigated by differential scanning calorimetry.

3.5. Differential scanning calorimetry (DSC) analysis

Both neat foam and the composite are crosslinked. Their glasstransition temperatures Tgs were recorded from second dynamicscans of DSC. In Fig. 6, the composite has a Tg of 109.0 �C, 5% higherthan neat foam 103.8�; this means that graphene increases thethermal stability of PVFMH. The improvement is attributed to (i)graphene – being exceptional stiff and strong with high meltingtemperature – can significantly stiffen polymers and increase theirTg and Tm, and (ii) graphene sheets might work as pore initiatingsites for foaming, creating a wide range of pore size (the followingSEM Section) that might pose an effect on the thermal properties.

5μm

(c)

5μm

(f)

ion of the references to colour in this figure legend, the reader is referred to the web

(a)

0.00 0.03 0.06 0.09 0.12 0.153.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

8.5

Wat

er re

tent

ion

rate

, g/g

Graphene sheets, vol%(b)

0.00 0.03 0.06 0.09 0.12 0.150.08

0.10

0.12

0.14

0.16

0.18

0.20

0.22

0.24

BET

sura

fce

area

, m2 /g

Graphene sheets, vol%

Fig. 8. Water retention rate and BET surface area of neat foam and its composites.(For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

260 S. Araby et al. / Composites: Part A 101 (2017) 254–264

3.6. Foam morphology

Polyvinyl formaldehyde (PVFMH) foam and its 0.023 vol% gra-phene composite were examined by SEM. The samples were frac-tured at �150 �C to ensure the preparation posing no effect onthe pore size, because PVFMH is not ductile any more at such atemperature. Both samples have macron-sized pores visible tonaked eyes. Fig. 7a–c contains the micrographs of neat foam. Atlow magnification, the wall thickness of inter-connected pores var-ies much. Smooth surface is observed at high magnification inFig. 7c.

Fig. 9. Macrostructure of neat foam and its 0.023 vol% composites before (a) and after (bthis figure legend, the reader is referred to the web version of this article.)

Although the composite in Fig. 7d reveals a similar morphologyto the neat foam in Fig. 6a, far more pores are seen at high magni-fication in Fig. 7e, where pores as small as �0.5 lm in diameter areobserved as pointed by yellow arrows. At higher magnification inFig. 7f, a higher degree of deformation is seen than neat foam inFig. 7c; this means more energy consumed to fracture the compos-ite sample during preparation due to the stiffing and strengtheningeffect of graphene. In comparison with neat foam in Fig. 7c, a num-ber of fine particles are observed in Fig. 7f and g; these particleswould be produced when the composite sample was fractured at�150 �C. The difference in the morphology between neat foamand the composite affects the mechanical and functional propertiesof bulk samples, which is presented in the following sections.

3.7. Water retention and BET surface area

Polymeric foams can be either stiff or elastomeric, depending ona number of factors including polymer type, microstructure (cellu-lar morphology), glass transition temperature, degree of crys-tallinity or crosslink density [12,16]. Stiff foams hold strongcellular structure and are not squeezable; therefore they are notfavourable in liquid absorption. Elastomeric foams have spring-like cell walls that enable absorption and desorption. PVFMHfoams are elastomeric, and thus water absorption is of practicalimportance.

In Fig. 8a, neat PVFMH foam shows a water retention rate of346%, and the addition of graphene increases the rate obviously.At 0.12 vol% graphene, the rate increases to 784% representing a127% improvement over neat foam, likely because (i) the additionof graphene caused more pores as illustrated in Fig. 7 and (ii) gra-phene improved the stiffness and strength of pore walls, and thispromoted absorption. At a relatively high fraction 0.14 vol%, gra-phene sheets may stack themselves, and the stacked sheets arenot as effective as those exfoliated sheets in terms of creating morepores and stiffening and reinforcing the matrix, which resulted inslightly reduced water retention rate. It is well-known that the fil-ler dispersion is crucial in determining functional properties andreinforcement efficiency [51–53]. Previous studies [40,51] showedclearly the processing history influenced the dispersion quality andconsequently the functional and mechanical properties of thecomposites.

As the addition of graphene caused more pores to the matrix,we proceeded to measure the total surface area of the samplesby N2 adsorption measurements using a micromeritics ASAP2020,which is known as Brunauer, Emmett and Teller (BET) measure-ment. In Fig. 8b, the addition of graphene sheets into PVFMHmark-edly improves the surface area. The area initially appears toincrease with the graphene fractions linearly until 0.12 vol% whereit starts to slow down. After reaching a peak of 0.21 m2/g at0.12 vol%, it reduces to 0.11 at 0.14 vol%. The evolution of the sur-face area is in agreement with the water absorption in Fig. 8a. Itimplies that surface area plays a key role in water absorption,and the composite at 0.12 vol% graphene sheets would be idealfor practical applications.

) absorption and desorption cycles. (For interpretation of the references to colour in

0.00 0.03 0.06 0.09 0.12 0.15 0.181E7

1E8

1E9

1E10

ϕt=0.023 vol%Surf

ace

resi

stiv

ity (R

s), Ω

/squ

are

Graphene sheets, vol%

-1.6 -1.5 -1.4 -1.3 -1.2 -1.1 -1.0

-7.86

-7.80

-7.74

-7.68

-7.62

log(ϕ−ϕt)

log

( σC)

Slope, t = 0.35R2= 94%

Fig. 10. Surface resistivity of neat foam and composites; the inset is the fitting ofthe results into the power law. (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.)

S. Araby et al. / Composites: Part A 101 (2017) 254–264 261

It is noteworthy that the composite foam not only demonstratesa high rate of water absorption, but remains remarkable structuralintegrity even after 150 cycles of absorbing and desorbing water.By contrast, neat foam cannot survive over 30 cycles – it disinte-grated after the cycles in Fig. 9b. The difference is caused by thereinforcement of graphene.

0.00 0.03 0.06 0.09 0.12 0.1540

42

44

46

48

50

52

54

56

58

60

62

Graphene sheets , vol%

Lim

iting

oxy

gen

inde

x,%

Fig. 12. Limiting oxygen index (LOI) of composite foams. (For interpretation of thereferences to colour in this figure legend, the reader is referred to the web version ofthis article.)

3.8. Anti-static property of composite foams

Neat foam is inherently insulating, and cannot be used in manyapplications such as electronic packaging, aerospace and automo-tive industries. The most cost-effective solution is to add into thefoam conductive fillers, i.e. metal fibres [54], carbon nanotubes[55] and graphene [51]. Graphene – known for high electrical con-ductivity 6000 S/cm [56] has been utilized to develop polymercomposites having low percolation thresholds of electrical conduc-tivity [39,48,57]. The graphene sheets that we used have a highcarbon to oxygen ratio of 9.8, and thus they should have high struc-tural integrity and electrical conductivity [30]. The surface resistiv-ity of neat PVFMH foam was measured as 1.6 � 1010X/square inFig. 10. It dropped dramatically to 8.29 � 107 X/square by addinga low volume fraction of graphene 0.023 vol%.

The surface resistivity was significantly lowered for three rea-sons: (i) the high contrast in resistivity between neat foam and gra-phene sheets; (ii) the sheets were able to react with formaldehydeduring the fabrication, as discussed previously in the XPS andRaman sections, and this should help to exfoliate graphene sheetsand to promote their dispersion in PVFMH; and (iii) thanks to theporous structure, graphene sheets may disperse through the porewalls connecting with each other to form continuous conductivepaths, as schematically shown in Fig. 11.

Neat foam

Fig. 11. Schematic of neat foam and its composite where graphene sheets are dispersed ain this figure legend, the reader is referred to the web version of this article.)

To have better understanding with regards to a minimum gra-phene fraction to obtain anti-static foam, we have fitted the resultsto a power law model [57], which relates to the inset in Fig. 10:

rc / ðu�utÞt ð6Þ

where rc, ut and t are the composite surface conductivity (recipro-cal of the surface resistivity), the percolation threshold and the uni-versal critical exponent, respectively. As presented in Fig. 10 inset,appropriate fitting reveals a percolation threshold at 0.023 vol%; itshows a universal exponent t = 0.35 with a regression coefficientR2 = 94%. According to literature [58,59], a universal exponentrelates to filler dispersion: t � 2 for randomly 3D dispersion andt < 2 for randomly 2D dispersion. The fitting implies that graphenesheets are randomly dispersed in the matrix. The measurementproved that the composite foams would suffice anti-static applica-tions [2,7].

3.9. Flame-resistant property of composite foams

The foam flame-retarding performance was measured by limit-ing oxygen index which is the minimum concentration of oxygento support the foam combustion; the measurement was under-taken through flowing a mixture of oxygen and nitrogen over aburning foam, and reducing the oxygen level until a minimumlevel is reached to sustain the combustion. The higher the limitingoxygen index, the more flame-retarding the sample is. In Fig. 12,neat PVFMH foam shows an index of 42.4, and it increases nearly

Foam/graphene composite

nd connected through the pore walls. (For interpretation of the references to colour

20μm 2μm

(a) (b)

Fig. 13. SEMmicrographs of 0.023 vol% composite’s residue after burning. (For interpretation of the references to colour in this figure legend, the reader is referred to the webversion of this article.)

262 S. Araby et al. / Composites: Part A 101 (2017) 254–264

linearly with addition of graphene sheets. At 0.14 vol%, the com-posite reaches a limiting oxygen index of 59.4, implying an excep-tional self-extinguishing behavior for such a compressible, light-weight foam.

The foam matrix itself has a satisfactory limiting oxygen indexof 42.4, and it can be highly improved to 59.4 through incorporat-ing graphene sheets. There might be three reasons for the improve-ment. Firstly, high aspect-ratio graphene sheets may act asphysical barriers that formed on the composite surface insulatingthe underlying material from the heat flux of the flame [60]. Sec-ondly, as discussed before, formaldehyde crosslinked polyvinylalcohol macromolecules, and it also bridged graphene sheets withthe matrix. Thus the composites’ crosslinked nature may prevent(i) the migration of polymer molecules onto the burning surfaceand thus remove the fuel supply and (ii) the structure collapse dur-ing burning. Thirdly, the modified graphene sheets having a carbonto oxygen atomic ratio of 4.6 may shield the composite from flameattack [61]. Upon heating, the crosslinked matrix would never meltbut soften and then degrade to yield a carbonaceous residue that –driven by the gaseous products and graphene sheets – would forma layer acting as a barrier to heat, oxygen and other pyrolysis prod-ucts [14,61].

SEM analysis was carried out to further investigate the residueof the burnt 0.023 vol% composite. At low magnification, Fig. 13ashows piles of residues that may contain degraded foam and gra-phene sheets. During the burning process, the top surface wouldbe flamed first. Polymers consists of organic molecules and thus

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80.0

0.1

0.2

0.3

0.4

0.5

Com

pres

sive

stre

ss, M

Pa

Strain, mm/mm

Neat foam0.023 vol% Composite

Fig. 14. Compressive stress-strain graphs of neat foam and its 0.023 vol% compos-ite. (For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

these should be readily burned and consumed; by contrast, gra-phene sheets are far less flammable and their two dimensionalgeometry would form shielding layers that may prevent oxygenand flammable gases transmitting to and from the hot surface. Inthis case, the matrix would work as a binder to produce a protec-tive dense-layer as shown in Fig. 13b. It is noteworthy that neatfoam was completely burnt leaving no residues.

3.10. Mechanical properties of composites

The elastic moduli of neat PVFMH foam and its graphene com-posite foam were measured by compression. In Fig. 14, compres-sion stress increases with strain for both neat foam andcomposite foam, and it increases more for the neat foam. Compres-sive modulus was determined at 19–22% strain. This range waschosen for two reasons: (i) to remove the initial unstable regioncaused by the self-adjustment of the Instron’s crosshead and (ii)to clearly understand the compressive moduli of neat and compos-ite foams as in Fig. 14 both graphs almost overlap at low strain.

The composite modulus was calculated as 0.121 ± 0.01 MPa,�133% lower than that for neat foam, 0.282 ± 0.06 MPa. This formsa strong contrast with previous studies where inorganic additiveswere often reported to significantly increase the polymer stiffness[62–68]. Since it is clear that graphene stiffs and strengths poly-mers, the reduction in modulus must be caused by the foam den-sities. So we proceeded to carefully measure the density of neatfoam and its composite. By randomly picking up six simples foreach, neat foam has a density of 0.114 ± 0.010 g/cm3 in comparisonwith 0.104 ± 0.009 g/cm3 for the composite. The composite has lesssolid content, and this explains why it has lower modulus.

4. Conclusions

Taking advantage of the high structural integrity of our gra-phene and its reactivity to formaldehyde, we developed polyvinylformaldehyde/graphene composite foams. The reaction built up astrong interface for the composites, and contributed to the exfoli-ation and good dispersion of graphene sheets in the matrix. Whileneat foam showed a water retention rate of 346%, the addition of0.12 vol% graphene sheets increased it up to 784%. At 0.023 vol%graphene, the foam surface resistivity lowered from 1.6 � 1010 X/square to 8.29 � 107. At 0.14 vol%, the composite reached a highlimiting oxygen index of 59.4, implying an exceptional fire-retarding performance for such a compressible, light-weight foam.

Acknowledgements

The authors appreciate financial support by the AustralianResearch Council (LP140100605). JM thanks Dr Andrew Michel-more for contribution of AFM micrographs.

S. Araby et al. / Composites: Part A 101 (2017) 254–264 263

References

[1] Tang Z, Guo B. Flame retardant polymer-halloysite nanocomposites. Functionalpolymer composites with nanoclays. The Royal Society of Chemistry; 2017. p.245–70 [chapter 9].

[2] Li Y, Wang X. Intrinsically conducting polymers and their composites foranticorrosion and antistatic applications. In: Yang X, editor. Semiconductingpolymer composites. Weinheim: Wiley-VCH Verlag GmbH & Co. KGaA; 2012.p. 269–98.

[3] Wichman IS. Material flammability, combustion, toxicity and fire hazard intransportation. Prog Energy Combust Sci 2003;29(3):247–99.

[4] Singh H, Jain AK. Ignition, combustion, toxicity, and fire retardancy ofpolyurethane foams: a comprehensive review. J Appl Polym Sci 2009;111(2):1115–43.

[5] Sang B, Li Z-w, Li X-h, Yu L-g, Zhang Z-j. Graphene-based flame retardants: areview. J Mater Sci 2016;51(18):8271–95.

[6] Bar M, Alagirusamy R, Das A. Flame retardant polymer composites. FibersPolym 2015;16(4):705–17.

[7] Irvine DJ, McCluskey JA, Robinson IM. Fire hazards and some commonpolymers. Polym Degrad Stab 2000;67(3):383–96.

[8] Liu YG, Ren YL, Dong EY. Development of environmental friendly flameretardants for wood plastic composites (WPC). In: Qian XM, Liu HW, editors.Advanced textile materials, pts 1–3. Durnten-Zurich: Trans Tech PublicationsLtd; 2011. p. 1880–3.

[9] Ying F, Cui YH, Xue GX, Qian HQ, Li AH, Wang X, et al. Preparation andproperties of an antistatic UV-curable coating modified by multi-walledcarbon nanotubes. Polym Bull 2016;73(10):2815–30.

[10] Maki N, Nakano S, Sasaki H. Development of a packaging material using non-bleed-type antistatic ionomer. Pack Technol Sci 2004;17(5):249–56.

[11] <http://www.aclstaticide.com/manuals/800manual_0813.pdf> [accessed Jan2017].

[12] Khemani KC. Polymeric foams: an overview. American Chemical Society; 1997.p. 1–7.

[13] Notario B, Pinto J, Rodriguez-Perez MA. Nanoporous polymeric materials: anew class of materials with enhanced properties. Prog Mater Sci 2016;78–79:93–139.

[14] Kind DJ, Hull TR. A review of candidate fire retardants for polyisoprene. PolymDegrad Stab 2012;97(3):201–13.

[15] Liu PS, Chen GF. Applications of polymer foams. Porousmaterials. Boston: Butterworth-Heinemann; 2014. p. 383–410 [chapter 8].

[16] Khemani KC. Polymeric foams. American Chemical Society; 1997.[17] Eaves D. Handbook of polymer foams. Rapra Technology; 2004.[18] Lee ST, Park CB, Ramesh NS. Polymeric foams: science and technology. CRC

Press; 2006.[19] Lewin-Epstein J. Polyvinyl sponge (Ivalon) as a scaffold for bone. Brit J Oral

Surg 1964;2:115–9.[20] Rádestad A, Christensen NJ. Magnesium sulphate and cervical ripening.

Contraception 1989;39(3):253–63.[21] Kirdponpattara S, Phisalaphong M. Bacterial cellulose–alginate composite

sponge as a yeast cell carrier for ethanol production. Biochem Eng J2013;77:103–9.

[22] Chen C, Liu L, Huang T, Wang Q, Ye Fang. Bubble template fabrication ofchitosan/poly(vinyl alcohol) sponges for wound dressing applications. Int JBiol Macromol 2013;62:188–93.

[23] Cheng C, Wang J-N, Yang X. Preparation of novel chelating sponge as anadsorbent for the removal of Cu2+ from water. Chin Chem Lett 2013;24(11):997–1000.

[24] Qiu Y, Liu J, Lu Y, Zhang R, Cao W, Hu P. Hierarchical assembly of tungstenspheres and epoxy composites in three-dimensional graphene foam and itsenhanced acoustic performance as a backing material. ACS Appl MaterInterfaces 2016;8(28):18496–504.

[25] Vahidifar A, Nouri Khorasani S, Park CB, Naguib HE, Khonakdar HA. Fabricationand characterization of closed-cell rubber foams based on naturalrubber/carbon black by one-step foam processing. Ind Eng Chem Res2016;55(8):2407–16.

[26] Li J-h, Feng L-l, Jia Z-x. Preparation of expanded graphite with 160 lm mesh offine flake graphite. Mater Lett 2006;60(6):746–9.

[27] Li J-h, Liu Q, Da H-f. Preparation of sulfur-free exfoliated graphite at a lowexfoliation temperature. Mater Lett 2007;61(8–9):1832–4.

[28] Yang JM, Su WY, Leu TL, Yang MC. Evaluation of chitosan/PVA blendedhydrogel membranes. J Membr Sci 2004;236(1–2):39–51.

[29] Blythe AR. Electrical resistivity measurements of polymer materials. PolymTesting 1984;4(2–4):195–209.

[30] Pei S, Cheng H-M. The reduction of graphene oxide. Carbon 2012;50(9):3210–28.

[31] Zhang L, Hu N, Yang C, Wei H, Yang Z, Wang Y, et al. Free-standing functionalgraphene reinforced carbon films with excellent mechanical properties andsuperhydrophobic characteristic. Compos A Appl Sci Manuf 2015;74:96–106.

[32] Liu D, Ma Z, Wang Z, Tian H, Gu M. Biodegradable poly(vinyl alcohol) foamssupported by cellulose nanofibrils: processing, structure, and properties.Langmuir 2014;30(31):9544–50.

[33] Pan Y, WangW, Peng C, Shi K, Luo Y, Ji X. Novel hydrophobic polyvinyl alcohol-formaldehyde foams for organic solvents absorption and effective separation.RSC Adv 2014;4(2):660–9.

[34] Espert A, Vilaplana F, Karlsson S. Comparison of water absorption in naturalcellulosic fibres from wood and one-year crops in polypropylene compositesand its influence on their mechanical properties. Compos A Appl Sci Manuf2004;35(11):1267–76.

[35] Ishigami M, Chen JH, Cullen WG, Fuhrer MS, Williams ED. Atomic structure ofgraphene on SiO2. Nano Lett 2007;7(6):1643–8.

[36] Novoselov KS, Geim AK, Morozov SV, Jiang D, Zhang Y, Dubonos SV, et al.Electric field effect in atomically thin carbon films. Science 2004;306(5696):666–9.

[37] Meyer JC, Geim AK, Katsnelson MI, Novoselov KS, Booth TJ, Roth S. Thestructure of suspended graphene sheets. Nature 2007;446(7131):60–3.

[38] McAllister MJ, Li J-L, Adamson DH, Schniepp HC, Abdala AA, Liu J, et al. Singlesheet functionalized graphene by oxidation and thermal expansion ofgraphite. Chem Mater 2007;19(18):4396–404.

[39] Meng Q, Kuan H-C, Araby S, Kawashima N, Saber N, Wang CH, et al. Effect ofinterface modification on PMMA/graphene nanocomposites. J Mater Sci2014;49(17):5838–49.

[40] Araby S, Meng Q, Zhang L, Kang H, Majewski P, Tang Y, et al. Electrically andthermally conductive elastomer/graphene nanocomposites by solutionmixing. Polymer 2014;55(1):201–10.

[41] Araby S, Zhang L, Kuan H-C, Dai J-B, Majewski P, Ma J. A novel approach toelectrically and thermally conductive elastomers using graphene. Polymer2013;54(14):3663–70.

[42] Li J, Kim J-K, Lung Sham M. Conductive graphite nanoplatelet/epoxynanocomposites: effects of exfoliation and UV/ozone treatment of graphite.Scripta Mater 2005;53(2):235–40.

[43] Sherif A, Izzuddin Z, Qingshi M, Nobuyuki K, Andrew M, Hsu-Chiang K, et al.Melt compounding with graphene to develop functional, high-performanceelastomers. Nanotechnology 2013;24(16):165601.

[44] Lee C, Wei X, Li Q, Carpick R, Kysar JW, Hone J. Elastic and frictional propertiesof graphene. Phys Status Solidi (b) 2009;246(11–12):2562–7.

[45] Yang Z, Kuang W, Weng P, Tang Z, Guo B. Graphene oxide/rhodanine redoxchemistry and its application in designing high-performanceelastomer/graphene composites. RSC Adv 2015;5(103):84398–405.

[46] Tang Z, Zhang L, FengW, Guo B, Liu F, Jia D. Rational design of graphene surfacechemistry for high-performance rubber/graphene composites.Macromolecules 2014;47(24):8663–73.

[47] Ferrari AC. Raman spectroscopy of graphene and graphite: disorder, electron–phonon coupling, doping and nonadiabatic effects. Solid State Commun2007;143(1–2):47–57.

[48] Zaman I, Kuan H-C, Meng Q, Michelmore A, Kawashima N, Pitt T, et al. A facileapproach to chemically modified graphene and its polymer nanocomposites.Adv Func Mater 2012;22(13):2735–43.

[49] Naebe M, Wang J, Amini A, Khayyam H, Hameed N, Li LH, et al. Mechanicalproperty and structure of covalent functionalised graphene/epoxynanocomposites. Sci Rep 2014;4:4375.

[50] Shi G, Meng Q, Zhao Z, Kuan H-C, Michelmore A, Ma J. Facile fabrication ofgraphene membranes with readily tunable structures. ACS Appl MaterInterfaces 2015;7(25):13745–57.

[51] Xu H, Gong L-X, Wang X, Zhao L, Pei Y-B, Wang G, et al. Influence of processingconditions on dispersion, electrical and mechanical properties of graphene-filled-silicone rubber composites. Compos A Appl Sci Manuf 2016;91(Part1):53–64.

[52] Zang J, Wan Y-J, Zhao L, Tang L-C. Fracture behaviors of TRGO-filled epoxynanocomposites with different dispersion/interface levels. Macromol MaterEng 2015;300(7):737–49.

[53] Gong L-X, Pei Y-B, Han Q-Y, Zhao L, Wu L-B, Jiang J-X, et al. Polymer graftedreduced graphene oxide sheets for improving stress transfer in polymercomposites. Compos Sci Technol 2016;134:144–52.

[54] Jalali M, Dauterstedt S, Michaud A, Wuthrich R. Electromagnetic shielding ofpolymer–matrix composites with metallic nanoparticles. Compos B Eng2011;42(6):1420–6.

[55] Li W, Yuan J, Dichiara A, Lin Y, Bai J. The use of vertically aligned carbonnanotubes grown on SiC for in situ sensing of elastic and plastic deformation inelectrically percolative epoxy composites. Carbon 2012;50(11):4298–301.

[56] Araby S, Meng Q, Zhang L, Zaman I, Majewski P, Ma J. Elastomeric compositesbased on carbon nanomaterials. Nanotechnology 2015;26(11):112001.

[57] Stankovich S, Dikin DA, Dommett GHB, Kohlhaas KM, Zimney EJ, Stach EA,et al. Graphene-based composite materials. Nature 2006;442(7100):282–6.

[58] Mahmoud WE. Morphology and physical properties of poly(ethylene oxide)loaded graphene nanocomposites prepared by two different techniques. EurPolymer J 2011;47(8):1534–40.

[59] Derrida B, Stauffer D, Herrmann HJ, Vannimenus J. Transfer matrix calculationof conductivity in three-dimensional random resistor networks at percolationthreshold. J Phys Lett 1983;44(17):701–6.

[60] Khobragade PS, Hansora DP, Naik JB, Chatterjee A. Flame retardingperformance of elastomeric nanocomposites: a review. Polym Degrad Stab2016;130:194–244.

[61] Han Y, Wu Y, Shen M, Huang X, Zhu J, Zhang X. Preparation and properties ofpolystyrene nanocomposites with graphite oxide and graphene as flameretardants. J Mater Sci 2013;48(12):4214–22.

[62] Kumar S, Doshi H, Srinivasarao M, Park JO, Schiraldi DA. Fibers frompolypropylene/nano carbon fiber composites. Polymer 2002;43(5):1701–3.

[63] Antunes M, Velasco JI. Multifunctional polymer foams with carbonnanoparticles. Prog Polym Sci 2014;39(3):486–509.

264 S. Araby et al. / Composites: Part A 101 (2017) 254–264

[64] Li W, Dichiara A, Bai J. Carbon nanotube–graphene nanoplatelet hybrids ashigh-performance multifunctional reinforcements in epoxy composites.Compos Sci Technol 2013;74:221–7.

[65] Wan Y-J, Gong L-X, Tang L-C, Wu L-B, Jiang J-X. Mechanical properties of epoxycomposites filled with silane-functionalized graphene oxide. Compos A ApplSci Manuf 2014;64:79–89.

[66] Wang X, Gong L-X, Tang L-C, Peng K, Pei Y-B, Zhao L, et al. Temperaturedependence of creep and recovery behaviors of polymer composites filled with

chemically reduced graphene oxide. Compos A Appl Sci Manuf2015;69:288–98.

[67] Liu S, Li A, He S, Xuan P. Cyclic compression behavior and energy dissipation ofaluminum foam–polyurethane interpenetrating phase composites. Compos AAppl Sci Manuf 2015;78:35–41.

[68] Alonso MV, Auad ML, Nutt S. Short-fiber-reinforced epoxy foams. Compos AAppl Sci Manuf 2006;37(11):1952–60.