Composites Science and Technology 83 (2013) 22–26

Contents lists available at SciVerse ScienceDirect

Composites Science and Technology

journal homepage: www.elsevier .com/ locate /compsci tech

The glass fiber–polymer matrix interface/interphase characterizedby nanoscale imaging techniques

0266-3538/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.compscitech.2013.04.014

⇑ Corresponding author. Tel.: +420 541149304; fax: +420 541149361.E-mail address: cech@fch.vutbr.cz (V. Cech).

V. Cech a,⇑, E. Palesch a, J. Lukes b

a Institute of Materials Chemistry, Faculty of Chemistry, Brno University of Technology, Purkynova 118, CZ-612 00 Brno, Czech Republicb Department of Mechanics, Biomechanics and Mechatronics, Czech Technical University, Technicka 4, CZ-166 07 Prague 6, Czech Republic

a r t i c l e i n f o

Article history:Received 11 December 2012Received in revised form 5 March 2013Accepted 11 April 2013Available online 26 April 2013

Keywords:A. Glass fiberA. Polymer–matrix composites (PMCs)B. InterfaceB. InterphaseD. Atomic force microscopy (AFM)

a b s t r a c t

Atomic force microscopy (surface topography, phase imaging, lateral forces), atomic force acousticmicroscopy, and dynamic mechanical analysis (modulus mapping) were used to characterize the inter-phase region of unsized, industrially sized, and plasma coated glass fibers (GF) in GF/polyester composite.The nanoscale imaging techniques revealed the sharp changes in mechanical properties within 0.1 lm ofthe interlayer/fiber and matrix/interlayer interfaces for plasma polymer coated fibers, or at the matrix/fiber interface in the case of unsized fibers. A region of modified matrix with a thickness of 0.5 lmwas determined near the fiber surface in the case of industrially sized fibers.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The interphase is a three-dimensional region between the bulkfiber and bulk matrix according to Drzal’s conception [1]. It in-cludes not only the two-dimensional (2D) area of contact (inter-face) between the fiber and the matrix, but also a region of somefinite thickness extending on both sides of the interface in boththe fiber and matrix. A schematic illustration of the compositeinterphase is given in Fig. 1 (adapted from Ref. [1]) with across-section of fiber–reinforced composite (left) and a detail ofthe region at the fiber surface (right). The interphase comprisesthe functional interlayer and the part of the matrix affected bythe presence of the coated fiber. The coated interlayer should im-prove compatibility between the fiber and the matrix by forminga strong but tough link between both phases [2]. We can distin-guish three interfaces in the region, i.e., three 2D contact areas, be-tween the composite phases. It is well known that the compositeinterphase control the mechanical properties of a composite [1].Details on interface/interphase properties in polymer–matrix com-posites and their characterization were discussed by Kim and Mai–[2,3] Jesson and Watts in the recent review [4].

Many attempts have been made to reveal the interphase regionin specific composite systems and evaluate its thickness, anddetermined interphase thicknesses varied depending on the tech-nique employed for evaluation. Ikuta et al. [5] investigated the

interphase region in a glass fiber (GF) reinforced vinylester com-posite system using microscopic FTIR spectroscopy. Scanning theregion while recording spectra, which were compared with thatof the bulk vinylester resin, they evaluated the interphase thick-ness as 80 lm. The interface region in GF/polymer compositeswas investigated using nanoindentation and nanoscratch tech-niques, which seemed to be more efficient and sensitive for suchinvestigations [6]. The nanoindentation measurements revealedan interphase thickness of 2 lm for GF/polyester and 6 lm forGF/phenolic composite systems. The interphase thickness wasevaluated as the thickness of the transition zone, where the matrixhardness increased and the friction coefficient decreased close tothe fiber surface. Kim et al. [7] used nanoindentation, nanoscratch,and thermal capacity jump measurements for fine investigation ofthe interphase region in GF/vinylester composites. They found thatthe measured interphase thickness varied between 0.8 and 1.5 lmdepending on the type and concentration of silane coupling. Thetrend of decreasing interphase thicknesses determined by increas-ingly precise techniques is evident. The greater resolution of cur-rent analytic techniques enables to characterize the interphaseregion at nanoscale. However, the resolution of nanoscale imagingtechniques still has limits due to measurement artifacts as dis-cussed by Young et al. [8].

In the current study, atomic force microscopy (AFM), atomicforce acoustic microscopy (AFAM), and dynamic mechanical anal-ysis (nanoDMA) were used for mapping of the surface topography,phase shift, lateral forces, AFAM signal, and viscoelastic propertiesof a cross-section of fiber–reinforced composite. These techniques

Fig. 1. Schematic illustration of a composite interphase. Adapted from Ref. [1].

V. Cech et al. / Composites Science and Technology 83 (2013) 22–26 23

were employed to characterize the interphase region and distin-guish the interfaces of unsized, industrially sized, and plasma poly-mer coated glass fibers in GF/polyester composite. We tried toidentify if the interface is sharp or diffusional as a sharp interfacecould be of critical importance for composite performance.

2. Experimental

Unsized and industrially sized glass fiber bundles (E-glass, 1200tex, and mean diameter 19 lm) were supplied by Saint-GobainAdfors CZ, Czech Republic. The unsized fibers were plasma polymercoated using tetravinylsilane, Si–(CH = CH2)4 (TVS, purity 97%, Sig-ma Aldrich) as the monomer. Plasma-polymerized tetravinylsilanefilms were prepared by plasma-enhanced chemical vapor deposi-tion (PECVD) employing an RF helical coupling system [9], oper-ated in a pulsed regime. The fibers were pretreated by O2 plasma(5 sccm, 4 Pa, 25 W) for 10 min to improve film adhesion. Plasmapolymer film, from a few nanometers to 10 lm thick, was depos-ited on bundles of unsized glass fibers at an effective power of2.5 W. Pulsed plasma was operated at a monomer flow rate of0.80 sccm and a corresponding pressure of 1.3 Pa. The free radicals(plasma species) diffuse into the central part of the bundle, forminga thin film even on the surface of central fibers during the plasmapolymer deposition. However, the deposition rate decreases radi-ally into the fiber bundle due to a shadowing effect of the sur-rounding fibers; thus, the film coating on the central fibers isthinner than on the fibers at the bundle edge.

Unsized, industrially sized, and plasma polymer coated glass fi-bers were embedded in unsaturated polyester resin Viapal HP 349F (Sirca S. p. A., Italy) to form a GF/polyester composite. A bundle offibers was impregnated with the resin and extra resin was carefullywiped from the bundle. The impregnated bundle was positionedaxially in a silicon rubber mold, which was filled with resin andcured at 140 �C to form a polymer disk 14 mm in diameter and5 mm in height. The disk was embedded in a metallographic spec-imen mount with the fibers normal to the specimen surface, andthe surface was polished using conventional metallographic tech-niques. Using nanoscale imaging techniques, we investigated thefiber–polymer matrix interface/interphase of a polished cross-section of composite samples.

The surface characteristics of the samples were observed usingan NTegra Prima Scanning Probe Microscope (SPM) (NT-MDT). Sur-face topography and phase imaging were measured in semicontact(tapping) mode, under ambient conditions, using silicon probesNSG03 (Rc < 10 nm, NT-MDT) of resonant frequency 90 kHz andforce constant 1.1 N m�1. The polished cross-section of compositesamples was also investigated in contact mode under a constantforce, using lateral force (LF) and atomic force acoustic microscopy(AFAM) measurements. Silicon probes CSG10 (Rc < 10 nm, NT-MDT) of resonant frequency 20 kHz and force constant 0.1 N m�1

were used in this case.

A Hysitron TI 950 TriboIndenter nanomechanical test instru-ment with a nanoDMA III (Dynamic Mechanical Analysis) packagewas employed to assess the viscoelastic properties of the fiber–polymer matrix interface/interphase. The Modulus Mapping pack-age combines the in situ SPM imaging capability of Hysitron’snanomechanical testing instruments with the ability to performnanoDMA III tests. During a test, the indenter is raster-scannedacross the surface to image the topography of the sample. Whilescanning, the probe is oscillated sinusoidally with a specified fre-quency and load amplitude. The resulting signal from the trans-ducer is then analyzed to determine the displacement amplitudeand phase lag of the oscillation at each pixel in the image. The soft-ware plots these measurements in separate image files, and oncethe scan is completed, the images can be analyzed to determinethe storage and loss modulus at each point. A Berkovich diamondindenter with a radius of curvature of about 100 nm (Rc) was used.A dynamic force amplitude of 5 lN at an oscillation frequency of220 Hz was superimposed on the normal contact force of 15 lNto keep the dynamic displacement amplitude at 4 nm while scan-ning the interface surface in an in situ SPM regime.

3. Results and discussion

The polished cross-section of composite samples with unsized,industrially sized, and plasma polymer coated fibers was investi-gated using different AFM modes and modulus mapping to distin-guish the composite phases and evaluate the interphase region onthe basis of surface topography and/or the difference in mechanicalproperties of the phases.

The surface topography (Fig. 2a) and phase distribution (Fig. 2b)of the composite with plasma polymer coated fibers were mea-sured simultaneously using semicontact mode. The interlayer(plasma polymer film) can be seen as a white annular ring sur-rounding the fiber (Fig. 2a). Both the interfaces between the inter-layer and the fiber and between the matrix and the interlayer areevident. The surface topography clearly distinguishes the compos-ite phases due to their different mechanical properties, influencingthe shape of the polished cross-section. The Young’s modulus was3.3 GPa and 60 GPa for polyester resin and the single glass fiber,respectively, as determined by tensile tests. The plasma polymerfilm deposited on planar glass substrate was characterized by sta-tic nanoindentation measurements; here the Young’s modulus wasevaluated as 17 GPa. When scanning in the phase imaging mode,the tip of the oscillating probe periodically comes into contact withthe sample surface. Its behavior is affected by the surface proper-ties of the sample. This can affect both the oscillation amplitudeand phase. If mechanical properties of the sample surface are inho-mogeneous, some shift of the phase occurs. Thus, the phase shiftdistribution over the sample surface can characterize the distribu-tion of mechanical properties, and could distinguish between stif-fer (fiber) and more pliant (interlayer, matrix) phases. In the phasedistribution image (Fig. 2b), a sharp interlayer/fiber interface and aless clear matrix/interlayer interface can be seen. The height(dashed line) and phase (solid line) profiles across the interphaseregion are given in Fig. 2c. The position of the profile is markedby a short white line in Fig. 2a and b. The height profile delimitsthe interlayer position, highlighted by dashed-and-dotted lines inFig. 2c. The phase distribution at the matrix/interlayer interfaceis somewhat too noisy to clearly distinguish the interface, as thephase for the interlayer and matrix are similar. And at the inter-layer/fiber interface there is a phase shift, likely due to a greaterdifference in Young’s modulus between the interlayer and the fi-ber. The phase shift is quite abrupt, occurring within a distanceof 100 nm.

The composite phases could be distinguished on the basis of dif-ferent frictional characteristics resulting from their different

Fig. 2. AFM images of cross-sectioned GF/polyester composite with plasmapolymer coated fibers using semicontact mode: (a) Surface topography; (b) Phasedistribution (the short white line indicates the position of the signal profile); (c)Height (dashed line) and phase (solid line) profiles across the interphase region.

Fig. 3. AFM images of cross-sectioned GF/polyester composite with plasmapolymer coated fibers using contact mode: (a) Surface topography; (b) Distributionof lateral forces (the short white line indicates the position of the signal profile); (c)Height profile (dashed line) and current profile (solid line) representing the lateralforces across the interphase region.

24 V. Cech et al. / Composites Science and Technology 83 (2013) 22–26

mechanical properties. Therefore, lateral force microscopy, usingcontact mode AFM, was employed to investigate the cross-sectioned composite sample with plasma polymer coated fibers.

A feedback loop maintained a constant force on the sample byadjusting the height of the cantilever to compensate for topograph-ical features of the surface, resulting in a surface topography image(Fig. 3a). Simultaneously, the torsion of the cantilever supporting

Fig. 4. AFM images of cross-sectioned GF/polyester composite with unsized fibersusing AFAM mode: (a) Surface topography; (b) Distribution of AFAM signal.

Fig. 5. Modulus mapping of cross-sectioned GF/polyester composite with plasmapolymer coated fibers: (a) Distribution of storage modulus (the modulus scale is inGPa; the black line indicates the position of the modulus profile); (b) Height(dashed line) and storage modulus (solid line) profiles across the interphase region.

Fig. 6. Height (dashed line) and storage modulus (solid line) profiles across theinterphase region using modulus mapping for cross-sectioned GF/polyester com-posite with unsized fibers.

V. Cech et al. / Composites Science and Technology 83 (2013) 22–26 25

the probe depended on the frictional characteristics of the surface,resulting in the distribution of lateral forces given in Fig. 3b. In thiscase, a sharp matrix/interlayer interface and a less clear interlayer/fiber interface can be observed in the distribution of lateral forces.The height profile (dashed line) and the current profile (solid line),representing the lateral force, are plotted in Fig. 3c. The profilesacross the interphase region correspond to the position marked(short white line) in Fig. 3a and b. We can see that the torsion ofthe cantilever changed sharply, within a distance of 80 nm, at thematrix/interlayer interface (marked by a dashed-and-dotted line)but changed only slightly at the interlayer/fiber interface(dashed-and-dotted line), which can be identified from the heightprofile in Fig. 3c. The frictional characteristics of the interlayer andfiber thus seem to be similar.

Atomic force acoustic microscopy is a combination of ultrason-ics with atomic force microscopy, where the vibration behavior ofthe AFM cantilever in contact with the composite sample surface issensitive to its local elastic properties [10]. In the AFAM arrange-ment, the composite sample is coupled to a piezoelectric trans-ducer. The ultrasonic transducer emits longitudinal waves intothe sample, which cause ultrasonic vibrations of the sample sur-face. These vibrations are transmitted via the tip into the cantileverand excite it to bending vibrations. The high-frequency vibrationsare detected by a four-sectioned photodiode, as used for topogra-

phy measurements, and evaluated by a lock-in amplifier. This set-up was used to record the amplitude change of the tip-samplesystem along the polished surface of GF/polyester composite withunsized fibers. The contact-mode topography image (Fig. 4a) wasacquired at the same time as the acoustic image (Fig. 4b). The scanarea was 10 � 10 lm2 in this case. The contrast between the fibersand the matrix in the AFAM image (Fig. 4b) is due to elastic anisot-

Fig. 7. Height (dashed line) and storage modulus (solid line) profiles across theinterphase region using modulus mapping for cross-sectioned GF/polyester com-posite with industrially sized fibers. The modified matrix is indicated by an arrow.

26 V. Cech et al. / Composites Science and Technology 83 (2013) 22–26

ropy. We can see that regions with greater stiffness (glass fiber)correspond to darker areas in the AFAM image. The matrix/fiberinterface is sharp in the AFAM image, although the correspondingsurface topography does not exhibit an abrupt step in height(diffused interface in Fig. 4a), which could indicate a potentialinfluence on the acoustic image. A sharp matrix/fiber interfacecan be expected in AFAM images in the case of unsized glass fibersdue to low interfacial adhesion.

Modulus mapping is well suited for samples with modulus vari-ations at nanoscale, such as multi-phase materials like polymercomposites. Mapping of the storage modulus enables one to easilydistinguish the composite phases, as shown in Fig. 5a for GF/poly-ester composite with plasma polymer coated fibers. However,some artifacts, due to deficient feedback compliance caused by sur-face topography, are found at the interlayer region. A profile of thestorage modulus in the direction indicated by the solid line(Fig. 5a) is plotted in Fig. 5b. The storage modulus of the matrixcan be simply distinguished from that corresponding to the inter-layer delimited by the height profile. Some of the spikes in themodulus profile of the interlayer are the artifacts mentioned above.Both interfaces in the interlayer region are relatively sharp. Noregion of modified matrix characterized by increased matrix mod-ulus near the matrix/interlayer interface or modified interlayernear the interlayer/fiber interface can be found at micron scale.Maps of the loss modulus were too noisy to correctly characterizethe interphase region.

Similarly, the profile of the storage modulus for the compositesample with unsized fibers (Fig. 6) does not indicate a modifiedmatrix of a thickness greater than 0.1 lm close to the matrix/fiberinterface. Only a slight increase of the storage modulus from 6 to8 GPa, within a distance of about 0.5 lm followed by a steepincrease within a few tens of nanometers (20–40 nm), can be foundat the matrix/fiber interface for industrially sized fibers embeddedin polyester resin (Fig. 7). The region, where the matrix modulusincreased from a bulk value marked by a dotted line, could be as-signed to the modified matrix affected by sized fiber due to a stron-ger interfacial adhesion.

4. Conclusion

The fiber–polymer matrix interface/interphase in GF/polyestercomposites with unsized, industrially sized, and plasma polymercoated fibers was characterized by surface topography, phaseimaging, distribution of lateral forces, acoustic imaging (AFAM),and distribution of storage modulus using AFM and nanoindenta-tion (modulus mapping) measurements. The composite phasescould be distinguished by surface topography mode due to the spe-

cific shape of the polished cross-sectioned composite samples. Thephase imaging mode was more suitable for characterization of thefiber/interlayer interface than the matrix/interlayer interface,likely due to a greater difference in the Young’s modulus betweenthe fiber and the interlayer. However, the lateral force mode wasmore suitable for characterizing the matrix/interlayer interfacethan the fiber/interlayer interface, probably due to a greater differ-ence in the friction coefficient between the matrix and the inter-layer. We can summarize all the nanoscale measurements andreport that all the composite interfaces (interlayer/fiber andmatrix/interlayer for plasma polymer coated fibers, or matrix/fiberin the case of unsized fibers) were relatively sharp with respect to asteep change in mechanical properties within a distance of 0.1 lm.Modulus mapping revealed a region of modified matrix with athickness of 0.5 lm in the case of industrially sized fibers. There-fore, the interphase thickness in GF/polyester composites wasdetermined by the interlayer thickness for plasma polymer coatedfibers and by the thickness of the modified matrix for industriallysized fibers, with an accuracy of 0.1 lm. The sharp interfaces inmulti-phase materials like composites and nanocompositesstrongly influence their performance, and a more gradual inter-phase could increase their performance [11]. The plasmachemicaltechnology (PECVD) is a perspective technique for construction ofgradient interlayers required for novel conception of compositeswithout interfaces [12]. Future nanoscale imaging techniques withincreased resolution will give more detailed insight into compositeinterface/interphases.

Acknowledgements

This work was supported in part by the Czech Science Founda-tion, Grant Nos. P106/11/0738 and P205/12/J058, and the Technol-ogy Agency of the Czech Republic, Grant No. TA01010796. Theauthors would like to thank Brian Rook (Composite Materials andStructures Center, Michigan State University, MI) for his kind helpwith sample polishing and Dr. Miroslav Sirovy and Saint-GobainAdfors CZ s.r.o. for providing glass fibers.

References

[1] Drzal LT, Rich MJ, Lloyd PF. Adhesion of graphite fibers to epoxy matrices: I. therole of fiber surface treatment. J Adhes 1983;16:1–30.

[2] Kim JK, Mai YW. High strength, high fracture toughness fiber composites withinterface control – a review. Compos Sci Technol 1991;41:333–78.

[3] Kim JK, Mai YW. Engineered interfaces in fiber reinforcedcomposites. Amsterdam: Elsevier; 1998.

[4] Jesson DA, Watts JF. The interface and interphase in polymer matrixcomposites: Effect on mechanical properties and methods for identification.Polymer Rev 2012;52:321–54.

[5] Ikuta N, Yanagawa A, Suzuki Y, Ochiai S. Investigation on resin interphaseproduced near silane-treated glass fiber in vinyl ester resin. Compos Interfaces2000;7:511–5.

[6] Hodzic A, Stachurski ZH, Kim JK. Nano-indentation and Nano-scratch ofpolymer/glass interface. Part I: Experimental and mechanical analysis. Polymer2000;41:6895–905.

[7] Kim JK, Sham ML, Wu J. Nanoscale characterization of interphase in silanetreated glass fiber composites. Composites: Part A 2001;32:607–18.

[8] Young TJ, Monclus M, Broughton WR, Ogin SL, Smith PA. Observations oninterphase characterisation in polymer composites by nano-scale indentationand the use of AFM cantilever torsion to identify measurement artefacts. PlastRubber Compos 2012;41:240–6.

[9] Cech V. Plasma-polymerized organosilicones as engineered interlayers in glassfiber/polyester composites. Compos Interfaces 2007;14:321–34.

[10] Kopycinska-Muller M, Caron A, Hirsekorn S, Rabe U, Natter H, Hempelmann R,Birringer R, Arnold W. Quantitative evaluation of elastic properties of nano-crystalline nickel using atomic force acoustic microscopy. Z Phys Chem2008;222:471–98.

[11] Cech V. Plasma polymer film as model interlayer for polymer composites. IEEETrans Plasma Sci 2006;34:1148–55.

[12] Cech V. Plasma polymer films: from nanoscale synthesis to macroscalefunctionality. In: Zhang S, editor. Nanostructured thin films and coatings,1. Boca Raton: CRC Press; 2010. p. 481–527.