Local mechanical and electromechanical properties of the P(VDF-TrFE)-graphene oxide thin films

Applied Surface Science 421 (2017) 42–51

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Full Length Article

Local mechanical and electromechanical properties of the P(VDF-TrFE)-graphene oxide thin films M.V. Silibin a , V.S. Bystrov b , D.V. Karpinsky a,c , N. Nasani d , G. Goncalves d , I.M. Gavrilin a , A.V. Solnyshkin a,e , P.A.A.P. Marques d , Budhendra Singh d , I.K. Bdikin a,d,∗ a

National Research University of Electronic Technology “MIET”, 124498 Moscow, Russia Institute of Mathematical Problems of Biology, Keldysh Institute of Applied Mathematics RAS, 142290 Pushchino, Moscow Region, Russia c Scientific-Practical Materials Research Centre of NAS of Belarus, 220072 Minsk, Belarus d TEMA-NRD, Mechanical Engineering Department and Aveiro Institute of Nanotechnology (AIN), University of Aveiro, 3810-193 Aveiro, Portugal e Tver State University, 170100 Tver, Russia b

a r t i c l e

i n f o

Article history: Received 15 October 2016 Received in revised form 28 January 2017 Accepted 29 January 2017 Available online 31 January 2017 Keywords: Ferroelectric polymer Graphene oxide Composites Piezoelectricity Piezoresponse force microscopy

a b s t r a c t Recently, many organic materials, including carbon materials such as carbon nanotubes (CNTs) and graphene (single-walled carbon sheet structure) were studied in order to improve their mechanical and electrical properties. In particular, copolymers of poly (vinylidene fluoride) and poly trifluoroethylene [P(VDF-TrFE)] are promising materials, which can be used as probes, sensors, actuators, etc. Composite thin film of the copolymer P(VDF-TrFE) with graphene oxide (GO) were prepared by spin coating. The obtained films were investigated using piezoresponse force microscopy (PFM). The switching behavior, piezoelectric response, dielectric permittivity and mechanical properties of the films were found to depend on the presence of GO. For understanding the mechanism of piezoresponse evolution of the composite we used models of PVDF chain, its behavior in electrical field and computed the data for piezoelectric coefficients using HyperChem software. The summarized models of graphene oxide based on graphene layer from 96 carbon atoms C: with oxygen and OH groups and with COOH groups arranged by hydrogen were used for PVDF/Graphene oxide complex: 1) with H-side (hydrogen atom) connected from PVDF to graphene oxide, 2) with F-side (fluorine atom) connected from PVDF graphene oxide and 3) Graphene Oxide/PVDF with both sides (sandwich type). Experimental results qualitatively correlate with those obtained in the calculations. © 2017 Elsevier B.V. All rights reserved.

1. Introduction In the recent years, significant interest has grown to find novel polar materials having low density, good elasticity, large piezoelectric and pyroelectric activity. Among these objects a special interest is focused on finding a composite material based on polymers and low-dimensional nanostructures of carbon (graphene, carbon nanotubes), as it exhibits an unusual electrical and mechanical characteristic. The study of such systems is dictated by the need to overcome the limitations in mechanical and electrical parameters of ferroelectric polymers. Composites containing nanostructured graphene in polymer matrix are promising materials due to their mechanical elasticity, low acoustic resistance, low dielectric con-

∗ Corresponding author at: National Research University of Electronic Technology “MIET”, 124498 Moscow, Russia. E-mail address: [email protected] (I.K. Bdikin). http://dx.doi.org/10.1016/j.apsusc.2017.01.291 0169-4332/© 2017 Elsevier B.V. All rights reserved.

stant and high piezoelectric and pyroelectric coefficients [1–3]. Furthermore, their properties can be varied by using different polymers and crystalline ferroelectrics, which gives an additional advantage for their use in functional devices such as pyro/piezo sensor, ultrasound transducers etc. In recent past, composites with ferroelectric polymer matrix based on polyvinylidene fluoride or copolymers have been explored by many researchers [4]. This is due to fact that, polar polymers base is characterized by high piezoelectric coefficients as compared to other polymer materials, and at the same time graphene particles embedded in a ferroelectric polymeric matrix is expected to provide highly unipolar state and a large electromechanical and pyroelectric activity. Furthermore, the polymer matrix doped with graphene can exhibit anomalous low coercive field without influencing its piezoelectric properties, which provides an additional degree of freedom for tailoring the properties of composites. In addition, it also allows the use of composite ferroelectric polymer-based materials in piezoelectric and pyroelectric sensor systems, as a relatively high conductivity of

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

43

graphene offsets its low value for the polymer matrix, which in turn leads to significant enhancement in various vital parameters. Among the various methods to synthesize composite materials based on polymers and graphene, the most common ways are the crystallization from solution [5], hot pressing [6], centrifugation [7,8] etc. Many models have been proposed to explain the dielectric characteristics of such composites, which, in addition to the dielectric constant of the phases, also take into account of their concentration and depolarization factor. However, experimental and theoretical data till now did not give clear picture for the polarization behavior and changes of other physical properties of ferroelectric polymers as a function of graphene inclusion into the matrix [9]. Thus, to understand and control the functional properties of such ferroelectric composites, a detailed study regarding correlation of their property their composition, preparation conditions, microstructure must be addressed in detail. Fig. 1. X-ray diffraction pattern of P(VDF-TrFE) and P(VDF-TrFE)-graphene oxide thin films.

2. Experimental and computational details Graphene oxide (GO) was prepared by the chemical exfoliation of graphite (Graphite powder, <45 ␮m, ≥99.99%, Sigma-Aldrich). Briefly, it consists of the reaction of graphite flakes with concentrated H2 SO4 and KMnO4 in order to obtain individual sheets in an oxidized state. The resultant suspension was extensively washed with distilled water by filtration and centrifugation and finally subjected to dialyses to remove ionic contaminants. The resulting GO was dried by lyophilization to avoid aggregation. For the composite sample preparation, poly (vinylidene fluoride-trifluoroethylene) copolymer with an TrFE content of 30% was used. The P(VDF-TrFE) copolymer was chosen because it favours crystallization from a solution or melt directly to the ferroelectric ␤-phase without additional stretching in comparison with the pure PVDF polymer. 4 g weight of the copolymer powder was added to 100 ml of a dimethyl sulfoxide (DMSO) and acetone mixture in 80/20 ratio. The powder was dissolved within 2 h at a temperature of 100 ◦ C. The solution was then carefully filtered to remove impurities. Dried graphene oxide was added to the solution and then stirred magnetically for 30 min at 50 ◦ C. For a homogeneous distribution of GO in the solution, this mixture was subjected to ultrasonic vibration for 1 h and stirred again for 3 h at 50 ◦ C. The resulting composition was used to produce the thin composite films on the base of copolymer P(VDF-TrFE) as the matrix. Composite films of pure P(VDF-TrFE) and with GO (GO concentration of 1 wt%) were prepared by spin coating technique using P(VDF-TrFE)-GO solutions with film thickness of 400–500 nm. Glass with conductive coating of InSbO4 was chosen as a substrate for the films deposition. Before the sample fabrication, the substrates were carefully cleaned with acetone and then dried. The powder X-ray diffraction (XRD) patterns of the films were collected at room temperature in a continuous scanning mode (step 0.02◦ and time 10 s) on a Siemens D500 diffractometer with a secondary monochromator CuK␣ X-radiation in the range 2 = 5◦ –60◦ . The nanoindentation measurements were performed using a three-sided pyramidal Berkovich diamond indenter having a nominal edge radius of 20 nm (faces 65.3◦ from vertical axis) attached to a fully calibrated nanoindenter (TTX-NHT, CSM Instruments). Atomic Force Microscopy (AFM) measurements were carried out using a Veeco AFM Multimode Nanoscope (IV) MMAFM-2, Veeco microscopy. Piezoresponse force microscopy (PFM) was performed using AFM instrument with an external lock-in amplifier (EG&G 5205 Lock-in Amplifier), used to apply ac and dc voltages. Local piezoelectric properties of the films were visualized simultaneously by using AFM in contact mode and piezoresponse force microscopy (PFM) methods [10]. The PFM technique is based on the

detection of the mechanical response of the sample to an applied electric voltage due to converse piezoelectric effect. A conductive Si cantilever (Nanosensors, nominal force constant of 15 N/m) was used to both apply the voltage to the surface and to measure mechanical response of the sample. The voltage applied to the sample was sine wave (Uac sin(␻t)). Under this electrical field the piezoresponse signal z(␻) = d33eff Uac (␻) were detected via vibration of the laser beam position on the photodiode due to sample deformation. The amplitude A( ) and phase difference (␻) of this vibration are measured with a lock-in amplifier. Domains with different orientation polarization obtained by PFM mode have different piezoresponse amplitude A(␻)∼z(Vac ) dependent on the orientation of the polarization and to the existence of difference phase (␻) between the two opposite domain 0 and 180◦ . So, using the function A(␻) cos((␻)) allows the mapping of piezoresponse signals on the sample surface as additional image. The amplitudes of the acquired signals are proportional to the effective vertical piezoelectric coefficients, d33eff . In the poling experiments, the AFM conductive tip is fixed at a predefined position on the surface and the external bias pulse is applied. After poling, this area is scanned repeatedly to image the ferroelectric switching. In this work, several versions of molecular models for PVDF (␤–phase) – graphene oxide ferroelectrics systems were developed and investigated using HyperChem software ver.7.52 and 8.0. Various computational methods were used, including molecular mechanics (MM) methods (such as BIO CHARM), quantum mechanical (QM) self-consistent field (SCF) Hartree-Fock (HF) calculations based on density functional theory (DFT), as well as semi-empirical methods (such as PM3), in both restricted Hartree-Fock (RHF) and unrestricted Hartree-Fock (UHF) approximations. The main approach of both the MM and QM methods used for molecular modeling is to obtain the minimum of the total, or potential energy surface (PES), of a studied molecular system. The optimization of molecular geometry is executed using the Polak–Ribere (conjugate gradient) algorithm, which determines an optimized geometry at the minimum energy point (using PES) [11]. 3. Results and discussion Fig. 1 shows X-ray diffraction (XRD) pattern (2␪ = 5–60◦ scan) for P(VDF-TrFE)-Graphene oxide/InSbO4 , P(VDF-TrFE)-Graphene oxide and InSbO4 films. The strong diffraction peaks corresponding to three main phases of wiz. InSbO4 , P(VDF-TrFE) and Graphene oxide are observed. The XRD pattern of GO shows a broad diffraction peak at 10.5◦ corresponding to the interplaner distance (d200 ) of 8.42 Å. In the mixed sample, three phase related to InSbO4 , P(VDF-TrFE) and GO was observed. In Initial P(VDF-TrFE) sample,

44

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

Fig. 2. AFM images of the P(VDF-TrFE) and P(VDF-TrFE)-graphene oxide thin films. (a and b)—P(VDF-TrFE) and (c and d)—P(VDF-TrFE)-graphene oxide thin films.

0.4 1% 0%

Force (mN)

0.3

0.2

0.1

0.0

0

50

100 150 Distance (nm)

200

Fig. 3. Load–displacement curves of indentations made on the P(VDF-TrFE) and P(VDF-TrFE)-graphene oxide thin films.

two main phase (␣ and ␤) was observed. However, in the XRD patterns of the GO/P(VDF-TrFE) composites, peaks related to only ␤ phase of P(VDF-TrFE) was present. The formation of the ␤ phase can be explained by the adsorption of PVDF chains onto GO sheets [12]. Furthermore, an increase in the interplaner spacing (d200 ) in composite from 8.42 to 8.84 Å was evident when compared to pure GO sample. This result suggests that the incorporation of GO particles in polymer matrix governs the polymer chain movement and consequently affects the molecule polarization. Fig. 2 demonstrates the morphology of P(VDF-TrFE)/GO as observed by AFM measurements. The average thickness of film as

obtained from the AFM analysis was found to be in the range of 200–400 nm. In addition, it is observed that GO particles present on the surface and in the body of the film have a size of less than 1 ␮m. Roughness and average grain size were also found to increase for GO composite. The RMS roughness and average grain size of GO and its composite were found to be 11.8 nm, 83.3 nm and 1.2 ␮m, 3.7 ␮m, respectively. Fig. 3 shows a representative load–displacement curvesfrom nanoindentation measurement performed on P(VDF-TrFE)Graphene oxide films. The standard Oliver and Pharr method was used for the analysis (hardness, H and Young’s modulus, E) of the obtained results [13]. In principle, E should be a constant regardless of the value of hc , unless the indentation load (or the indentation depth) is too large so that the indenter starts to feel the effect of the underlying silicon substrate. The very thin film on the substrate will indeed make the nanoindentation process more challenging [14]. The load–displacement curve acquired at indentation depths of less than 200 nm (0.3 mN maximum loading) showed linear elastic (reversible) loading and unloading behavior (Fig. 3), which was also confirmed by the fact that indentation at such small load did not show any evidence of plastic surface damage. Hardness and Young’s modulus on the films starts out at a value of 358.4 MPa and 32.9 GPa at pure P(VDF-TrFE) film (0.3 mN maximum loading) to 275.5 MPa and 34.0 GPa for P(VDF-TrFE)- GO composite film. It was found similar for GO-polymer composite that employing high GO content which could further lead to a performance enhancement is limited [15]. So, for the GO/PVDF composites, with low concen-

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

45

60

700

H, MPa

E, GPa

600

50

500

400

E, GPa

Hardness, MPa

40

300

30

20 200

10

100

(a )

(b ) 0

0 50

100

150

200

250

50

300

100

150

200

250

300

Depth, nm

Depth, nm 700

E, GPa

60

H, MPa

600

50

40

400

E, GPa

Hardness, MPa

500

300

30

20

200

100

(c)

0

10

(d )

0

50

100

150

200

250

300

Depth, nm

50

100

150

200

250

300

Depth, nm

Fig. 4. Hardness (a and c), Young’s modulus (b and d) of P(VDF-TrFE)-graphene oxide thin films with depth thickness for pure P(VDF-TrFE) (a and b) and P(VDF-TrFE)-graphene oxide thin films (c and d).

trations of just a few per cent very strongly change the mechanical properties of composites similar to MWCNT [16]. The presence of graphene oxide particles in the body of the films clearly attested by modulation parameters measured film hardness (Fig. 4). For pure film, stiffness and hardness parameters were not found to change till an indentation depth of 200 nm. However, in the range of 200–300 nm, stiffness and hardness change smoothly to the substrate parameters (Fig. 4a and b). P(VDF-TrFE)/Graphene oxide film in the range up to 200 nm, a variation of parameters in the range 200–500 MPa was observed. This suggests a statistical distribution of graphene oxide particles in the films (Fig. 4c and d). Furthermore, the nanoindentation measurements results are also correlated with AFM images (Fig. 5). However, a similar surface profile of the film surface after nanoindentation was evident. InP(VDF-TrFE)/Graphene oxide film, as the GO is harder than the polymer, hence during indentation run if indenter falls in the matrix where graphene oxide particle is present, the depth of indentation would decrease sharply, and an increase in the deformation of the surface around the area would be evident as observed in Fig. 6. Obtained local piezoresponse hysteresis loops of P(VDF-TrFE)graphene oxide thin films is shown in Fig. 7a. Decrease in the value of the piezoelectric coefficient d33 for P(VDF-TrFE)- GO composite film was observed. As shown by the nanoindentation, composite deformation is difficult (Young modulus increase for P(VDF-TrFE)graphene oxide film). This may be one of reason for reducing the piezoelectric coefficient d33 . The shifts and broadening of the loops are mainly related to piezoelectric, dielectric properties and self-polarization component. Furthermore, graphene oxide is very anisotropic material with large surface area of individual sheets,

which allows to trap electric charges at the surface and as a result local polarization can be induced. In composite materials the separation of sheets is preferable and the separation effects are more pronounced than in the compressed pure graphene state. It can explain an increase in the coercive field (Fig. 7c). The d33 shift as observed for negative values (Fig. 7d) confirm the imprint effect, which tends to disappear for P(VDF-TrFE)-GO composite film. To investigate macroscopically a piezoelectricity of the films, the domain structure was written in an area of 10 × 5 mm2 (as illustrated in Fig. 8) by applying dc voltage (−50 V/+50 V) more than its coercive voltage with a slow scanning velocity (Vtip = 2.5 ␮m s−1 ). Fig. 9 shows the distribution of the local piezoresponse signal of the films extracted from Fig. 8. Three characteristic peaks were observed. Peak with negative sign corresponds to the area were positive poling (+50 V) voltage was applied. On the other hand, peak with positive signal corresponds to the piezoresponse from negatively poled (−50 V) area. The negative sign of the piezoelectric signal with respect to the polarity of the applied switching voltage corresponds to negative value of piezoelectric coefficients in P(VDF-TrFE). Thus, the distances between the peaks are related to the effective coefficient d33 polar states and their positions can be roughly considered as the mean values in the piezoelectric response. Since the PFM results give an average value of surface vibrations caused by the inhomogeneous distribution of electric field, it is necessary to compare the results with those attributed to a well known piezoelectric as LiNbO3 . Thus, the piezoelectric effect obtained for the samples can be quantified. After callibration we have obtained the next values of d33eff for our films: 38.0 pm/V for pure P(VDF-TrFE) and 30.1 pm/V for P(VDF-TrFE)/GO composite

46

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

Fig. 5. AFM after indentation on P(VDF-TrFE) (a and b) and (c and d) P(VDF-TrFE)-graphene oxide thin films.

Fig. 6. Cross-sections AFM images. (a–c)—A, B and C indentation points on Fig. 6c.

films. It is seen that the pattern of polarization is similar to the local piezoresponse hysteresis loops (Fig. 7). For understanding the mechanism of piezoresponse evolution of the composite we have used our previous models of PVDF chain (Fig. 10a) [17], its behavior in electrical field (Fig. 10b) and computed the data for piezoelectric coefficients [18]. The symmetrized models of graphene oxide based on graphene layer consisting of 96 carbon atoms: with oxygen and OH groups (Fig. 11a), marked by Gr96N2O2H2, and with COOH groups (Fig. 11b), marked shortly. These models are based on the approaches developed and used, e.g., in [19,20]. Additionally, the nitrogen atoms in Gr96 structures models are included, because usually GO contain nitrogen atoms after synthesis due to the vinyl-pyridine resin in the initial compo-

nents or using strong oxidant (such as HNO3 ) and it is important for many practical cases [20]. It is possible to assume several simplest models for PVDF/Graphene oxide complex (Fig. 12) and compute its piezoelectric coefficients by the same calculations algorithms as declared in Refs. [17,18]. We have constructed the models of PVDF/Graphene oxide composites: viz. three variants: 1) with H-side (hydrogen atom) connected from PVDF to graphene oxide, 2) with F-side (fluorine atom) connected from PVDF to graphene oxide (these both first variants show approximately the same values of piezoelectric coefficients) and 3) Graphene Oxide/PVDF with both sides (sandwich type) as show in Fig. 12. For deeper understanding, we also considered two versions of the mutual rotation of the graphene

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

600

0% 1%

47

1200

d33 eff (μV)

300

d33 eff

800 0 -300 -600

400

(b)

(a) -75

0 -50

-25 0 25 Udc ( Volts )

50

75

0.0

0.5

1.0

% GO

40

0

33 eff

o f f set d

20

U

cr

30

10

0.0

-30 -45

(c) 0

-15

0.5

1.0

% GO

(d) 0.0

0.5

1.0

% GO

Fig. 7. Piezoresponse hysteresis loops of P(VDF-TrFE) and P(VDF-TrFE)-graphene oxide thin films. Uac = 3 V, 50 kHz.

Fig. 8. AFM Topography (a and c) and PFM image (b and d) for P(VDF-TrFE) (a and b) and (ca nd d) P(VDF-TrFE)-graphene oxide films.

layers in relation with PVDF chain. The results are presented in Table 1. An optimization strategy similar to that used in the paper [17,18] was used: firstly, the model without external applied electric field has been used, to find the initial optimal position of modeled composite structure (the distance between graphene oxide layer and

PVDF chain was approximately 4 Å) and to determine the initial optimal parameters of PVDF chain heights in its central part (h1, and h2). After that we have applied electric field along Z direction (along the main vector of PVDF polarization) and tried to search optimal geometry for new atomic configuration under action of electric field. Then we have compared the changes of the main parameters

48

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

Table 1 The piezoelectric coefficients d33 calculated for different types of structures of Graphene Oxide with OH and COOH groups and PVDF (electric field Ez ∼ 500 GV/m, for comparison the data from our paper [18] were taken). Type of structure

d33 , pm/V

PVDF12/Gr96N2O2H2 PVDF12/Gr96O Gr96N2O2-H2/PVDF12/Gr96N2O2-H2 Gr96O/PVDF12/Gr96O PVDF12

− 14.6 − 13.5 − 29.8 −18.7 −38.5

Number of points

1500

1000

0% 1%

-50 V

+50 V Fig. 10. Model of PVDF chain: (a)—PVDF and its central part, (b)—PVDF chain deformation under applied electrical field.

500

0 -6

-4

-2 0 2 Piezoresponse, mV

4

6

Fig. 9. Distribution of the local piezoresponse after poling analysis of P(VDF-TrFE) and P(VDF-TrFE)-graphene oxide thin films.

(h1 and h2, Fig. 10) from initial optimal parameters, to determine the deformation h1 and h2, to calculate the corresponding values of voltage U and, finally, to compute the piezoelectric coefficient d33 , using dielectric permittivity value of ␧ = 10 (for simplest case comparable to other data). Comparison with the initial data known for d33 testifies that under the influence of graphene oxide layer the piezoelectric coefficient d33 is decreased and now have three times lower value: d33 = −14.6 pm/V (or pC/N) for the first models (Fig. 12) of GO eight OH groups as compared with the average value of the pure PVDF d33 = −38.5 pm/V (pC/N). It is important to note, that the sign of d33 is negative in all cases as in the initial pure PVDF case. In the case of double sides graphene oxide model (sandwich structure) the piezoelectric coefficient d33 is increased to the value of d33 = −29.8 pm/V (pC/N) for this model (see Table 1). The mentioned value is calculated for the case of the simplest graphene oxide structure and its corresponding orientation. For other graphene oxide compositions and various mutual orientations, the piezoelectric coefficient has somewhat different values, but the main tendency remains the same (e.g. the values noted in Table 1 which are computed for the models of GO with COOH groups

(Fig. 11b)). One interesting peculiarity here is that the mutual orientation influences on the resulted values of the piezoelectric coefficients. The main obtained data and comparison with the data presented in the paper [18] are presented in Table 1, where data for Gr96O/PVDF12/Gr96O model is average value from two different mutual orientations. These are the first averaged modeling data, which are in line with above mentioned experimental PFM data, and further detailed investigations are needed for different mutual orientations, number of layers and the order of the GO and PVDF. Experimental results qualitatively correlate with those obtained in the calculations. We can assume that experimental data obtained for the P(VDF-TrFE)-GO composite film can be associated with the model constructed for the case of PVDF with graphene oxide from one side only. This leads to a reduction in the piezoresponse coefficient. Increasing of GO content most probably corresponds to the model assuming sandwich clusters in the composite. Experimentally, piezo signal is reduced for 1% of GO content due to the statistical disorientation of graphene oxide and PVDF layers, and uncontrolled thickness of the individual layers of graphene oxide and PVDF. Statistical disorder does not yield an exact match with the simulation. Even at low concentrations, there is the effect of molecular order, but for these composites the probability that the sandwich structures are formed is quite significant. In the case of the controlled hetero-structures one can assume much greater effect. Quality manifestation of the effect of the graphene oxide grains on the piezoelectric properties of the composite films can be observed at the boundaries of graphene grains (Fig. 13). Thus

Fig. 11. Graphene oxide layers models: (a) with OH group, (b) with COOH group.

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

Fig. 12. Models of PVDF/graphene oxide structures, H-side. (a)—PVDF/graphene oxide, Type 1, (b)—sandwich graphene oxide/PVDF/graphene oxide, Type 2.

49

graphene oxide grain has a layered structure with separated layers. Separated sheets are more ease at the edges of the sheets. So, the grain boundary will split on the graphene oxide grains and PVDF fills all cavities in the composite. That may lead to the formation of alternating layers of graphene oxide and PVDF on the boundary grains. Based on the results of our simulations the piezoresponse for sandwich structures PVDF/GO must be increased (Table 1). This is qualitatively observed from our PFM measurements (Fig. 13a). Figure shows the contour boundaries of graphene grain. In the grain boundary an increased intensity of piezoresponse was observed (Fig. 13b). This is possibly due to the alternating structure. This suggests that in the heterostructures under controlled alternating layers, this effect will be clearer and they can be controlled. Fig. 14a and b demonstrates the morphology of P(VDF-TrFE) and P(VDF-TrFE)/GO composite films as observed with AFM. From AFM images, it is observed that GO particles located on the film surface have a size of about several microns. Piezoresponse images after point poling in the central point of the selected area show the emergence of the polar domain. It can be seen that the domain size is different for pure P(VDF-TrFE) and P(VDF-TrFE)-GO composite films. Effective diameters of domains are 0.81 and 1.80 ␮m for pure and composite films, correspondently. It should be noted that the domain for composite film very often was not a circle, and reflect sharp border of graphene oxide particles. Fig. 14d shows that the

Fig. 13. PFM image (a), dots—GO grain, (b)—cross-section along line L and (c)—model of P(VDF-TrFE)-GO- P(VDF-TrFE). Arrows indicates the boundary areas.

50

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

Fig. 14. AFM images of the P(VDF-TrFE) (a) and P(VDF-TrFE)-graphene oxide (b) thin films. PFM images (c)—P(VDF-TrFE), (d)—P(VDF-TrFE)-graphene oxide after poling −90 V during 100 s in the center point. (e) cross-sections on PFM images (c) and (d).

domain is split onto two domains, although there was one point of polarization and it located in the center of the larger domain. The result of the analysis shows that for P(VDF-TrFE)/GO composite film has modulated piezoelectric coefficient (Fig. 14e) by local GO grain structure and some areas of the film are characterized by higher piezoresponse than that estimated for the pure film.

obtained considering the PFM images with area poling and justify a decrease of piezo-coefficient d33eff values under the influence of G/GO layers. Modelling of sandwich structures of PVDF/GO and PFM measurements on the graphene oxide grains show the prospects of the multilayer structures of these composites with a significant increase in piezoelectric parameters.

4. Conclusions

Acknowledgement

Nano-indentation and piezoresponse force microscope tests have been used for assessment of the mechanical and electromechanical properties of PVDF/Graphene oxide composite thin films. The summarized models of PVDF/Graphene oxide composite with oxygen and OH groups and with COOH groups arranged by hydrogens were used for PVDF/Graphene oxide complex viz. 1) with H-side (hydrogen atom) connected from PVDF to graphene oxide, 2) with F-side (fluorine atom) connected from PVDF graphene oxide and 3) Graphene Oxide/PVDF with both sides (sandwich type). The results of theoretical modelling show qualitative agreement of the piezoelectric properties with the experimental data

The authors wish to acknowledge the Russian Science Foundation (Grant 16-19-10112) for financial support. Graphene oxide has been synthesized by P.A.A.P. Marques in the frames of the FCT project # UID/EMS/00481/2013. References [1] S.-H. Bae, O. Kahya, B.K. Sharma, J. Kwon, H.J. Cho, B. Özyilmaz, J.-H. Ahn, Graphene-P(VDF-TrFE) multilayer film for flexible applications, ACS Nano 7 (4) (2013) 3130–3138. [2] R. Md Ataur, L. Byung-Chul, P. Duy-Thach, C. Gwiy-Sang, Fabrication and characterization of highly efficient flexible energy harvesters using

M.V. Silibin et al. / Applied Surface Science 421 (2017) 42–51

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

PVDF–graphene nanocomposites, Smart Mater. Struct. 22 (8) (2013) 10–085017. T. Kuilla, S. Bhadra, D. Yao, N.H. Kim, S. Bose, J.H. Lee, Recent advances in graphene based polymer composites, Prog. Polym. Sci. 35 (11) (2010) 1350–1375. P. Martins, A.C. Lopes, S. Lanceros-Mendez, Electroactive phases of poly(vinylidene fluoride): determination, processing and applications, Prog. Polym. Sc. 39 (2014) 683–706. R.K. Layek, A.K. Das, M.J. Park, N.H. Kim, J.H. Lee, Enhancement of physical, mechanical, and gas barrier properties in noncovalently functionalized graphene oxide/poly(vinylidene fluoride) composites, Carbon 81 (2015) 329–338. F. Liu, R. Huo, X. Huang, Q. Lei, P. Jiang, Crystalline properties, dielectric response and thermal stability of insitu reduced graphene oxide/poly(vinylidene fluoride) nanocomposites, dielectrics and electrical insulation, IEEE Trans. 21 (2014) 1446–1454. E. Thangavel, S. Ramasundaram, S. Pitchaimuthu, S.W. Hong, S.Y. Lee, Sh.-S. Yoo, D.-E. Kim, E. Ito, Y.S. Kang, Structural and tribological characteristics of poly(vinylidene fluoride)/functionalized graphene oxide nanocomposite thin films, Compos. Sci. Technol. 90 (2014) 187–192. Ch. Yu, D. Li, W. Wu, Ch. Luo, Y. Zhang, Ch. Pan, Mechanical property enhancement of PVDF/graphene composite based on a high-quality graphene, J. Mater. Sci. 49 (2014) 8311–8316. S. Ansari, E.P. Giannelis, Functionalized graphene sheet- poly (vinylidene fluoride) conductive nanocomposites, J. Polym. Sci. Part B Polym. Phys. 47 (2009) 888–897. N. Balke, I. Bdikin, S.V. Kalinin, A.L. Kholkin, Electromechanical imaging and spectroscopy of ferroelectric and piezoelectric materials: state of the art and prospects for the future, J. Am. Ceram. Soc. 92 (8) (2009) 1629–1647. H. HyperChem(TM) Professional 7.51, Inc., 1115 NW 4th Street, Gainesville, Florida 32601, USA, 2002.

51

[12] L. Huang, C. Lu, F. Wang, L. Wang, Preparation of PVDF/graphene ferroelectric composite films by in situ reduction with hydrobromic acids and their properties, RSC Adv. 4 (2014) 45220–45229. [13] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564–1583. [14] S. Lin, C.T. Chen, I. Bdikin, V. Ball, J. Gracio, M.J. Buehler, Tuning heterogeneous poly(dopamine) structures and mechanics: in silico covalent cross-linking and thin film nanoindentation, Soft Matter 10 (2014) 457–464. [15] M. Ionita, A.M. Pandele, L. Crica, L. Pilan, Improving the thermal and mechanical properties of polysulfone by incorporation of graphene oxide, Compos. Part B 59 (2014) 133–139. [16] O. Eggedi, U. Valiyaneerilakka, M.R. Dar, S. Varghese, Nanoindentation and thermal characterization of poly (vinylidenefluoride)/MWCNT nanocomposites, AIP Adv. 4 (2014) 047102. [17] V.S. Bystrov, N.K. Bystrova, E.V. Paramonova, G. Vizdrik, A.V. Sapronova, M. Kuehn, H. Kliem, A.L. Kholkin, First principle calculations of molecular polarization switching in P(VDF–TrFE) ferroelectric thin Langmuir–Blodgett films, J. Phys. Condens. Matter 19 (45) (2007) 456210. [18] V.S. Bystrov, E.V. Paramonova, I.K. Bdikin, A.V. Bystrova, R.C. Pullar, A.L. Kholkin, Molecular modeling of the piezoelectric effect in the ferroelectric polymer poly(vinylidene fluoride) (PVDF), J. Mol. Model. 19 (9) (2013) 3591–3602. [19] A. Burian, A. Ratuszna, J. Dore, Radial distribution function analysis of the structure of activated carbons, Carbon 36 (1998) 1613–1621. [20] Q. Wei, X. Tong, G. Zhang, J. Qiao, Q. Gong, S. Sun, Nitrogen-doped carbon nanotube and graphene materials for oxygen reduction reactions, Catalysts 5 (2015) 1574–1602.