Effectiveness of Atomic Force Microscopy (AFM)

polymers

Article Carbon-Based Aeronautical Epoxy Nanocomposites: Eﬀectiveness of Atomic Force Microscopy (AFM) in Investigating the Dispersion of Diﬀerent Carbonaceous Nanoparticles

Marialuigia Raimondo 1,* , Carlo Naddeo 1 , Luigi Vertuccio 1, Khalid Lafdi 2, Andrea Sorrentino 3 and Liberata Guadagno 1

1 Department of Industrial Engineering, University of Salerno, Via Giovanni Paolo II, 132 – 84084 Fisciano (SA), Italy; [email protected] (C.N.); [email protected] (L.V.); [email protected] (L.G.) 2 University of Dayton, 300 College Park, Dayton, OH 45440, USA; [email protected] 3 Institute for Polymers, Composites and Biomaterials (IPCB-CNR), via Previati n. 1/E, 23900 Lecco, Italy; [email protected] * Correspondence: [email protected]; Tel.: +39-089-964019

Received: 21 March 2019; Accepted: 2 May 2019; Published: 8 May 2019

Abstract: The capability of Atomic Force Microscopy (AFM) to characterize composite material interfaces can help in the design of new carbon-based nanocomposites by providing useful information on the structure–property relationship. In this paper, the potentiality of AFM is explored to investigate the dispersion and the morphological features of aeronautical epoxy resins loaded with several carbon nanostructured fillers. Fourier Transform Infrared Spectroscopy (FTIR) and thermal investigations of the formulated samples have also been performed. The FTIR results show that, among the examined nanoparticles, exfoliated graphite (EG) with a predominantly two-dimensional (2D) shape favors the hardening process of the epoxy matrix, increasing its reaction rate. As evidenced by the FTIR 1 signal related to the epoxy stretching frequency (907 cm− ), the accelerating effect of the EG sample increases as the filler concentration increases. This effect, already observable for curing treatment of 60 min conducted at the low temperature of 125 ◦C, suggests a very fast opening of epoxy groups at the beginning of the cross-linking process. For all the analyzed samples, the percentage of the curing degree (DC) goes beyond 90%, reaching up to 100% for the EG-based nanocomposites. Besides, the addition of the exfoliated graphite enhances the thermostability of the samples up to about 370 ◦C, even in the case of very low EG percentages (0.05% by weight).

Keywords: epoxy nanocomposites; carbonaceous nanoﬁllers; FTIR analysis; Thermosetting resins; exfoliated graphite; surface analysis; Atomic Force Microscopy (AFM)

1. Introduction Atomic force microscopy (AFM) is one of the most powerful characterization techniques for morphological and structural analysis of polymeric nanocomposites thanks to its ability to produce three-dimensional (3D) topographic images with a resolution in the order of nanometers [1,2]. These peculiar characteristics allow for the gathering of information about the morphology by drawing the heights and position of very small features present on the sample surface. The AFM also gives the possibility of carrying out a variety of surface characterizations through the controlled interaction between the microscopic probes and the sample surface. Moreover, the AFM is advantageous because it can be used in ambient conditions with minimum sample preparation; in fact, the sample does not have to be conductive and does not require the metallization process before being submitted to

Polymers 2019, 11, 832; doi:10.3390/polym11050832 www.mdpi.com/journal/polymers Polymers 2019, 11, 832 2 of 20 morphological investigation. These capabilities make this technique an extraordinary tool for the direct characterization of a variety of samples having complex morphological organizations. In recent years, nanostructured polymeric composites have attracted considerable interest from both the academy and industry. The possibility to manipulate these materials on a nanometric scale allows one to obtain multifunctional materials with completely new properties. For example, the use of carbonaceous nanoparticles such as carbon nanotubes, carbon nanofibers and graphene nanoplatelets as filler in polymeric composites is a well-established and effective strategy aimed at improving mechanical, electrical, and thermal performance [3–12]. In aeronautical and aerospace fields, the continuous and pressing request for high performance materials combining facile manufacturing process and superior mechanical and thermal properties have stimulated the research in different ways [13]. From this point of view, high performance fiber composites characterized by epoxy-based matrices with very high glass transition temperature and impressive thermal and chemical resistance have demonstrated compelling prospectives [14]. The low density and the good physical resistance make these materials very interesting in structural-functional engineering applications. High performance continuous fiber-reinforced (e.g. carbon fiber, glass fiber) composites with complex geometries can be obtained with a simple manufacturing process and with relative low cost. However, the complete replacement of metallic structure is still impossible due to some limitations related to the polymer matrix properties. The low electrical conductivity and the poor barrier properties are important limits still to be overcome. The addition of functional filler can improve these properties without affecting the processability and the cost of these composites. In this paper, different types of one-dimensional (1D) carbon nanofillers such as multi-walled carbon nanotubes (MWCNTs), as-received and heat-treated carbon nanofibers (CNFs), and two-dimensional (2D) predominant shape exfoliated graphite nanoparticles (EG) have been embedded in the same epoxy matrix for obtaining functional structural materials. In particular, the graphite nanoplatelets embedded in polymeric matrices can replace carbon nanotubes both because they are low-priced and because they offer guarantees in terms of wished performance [15]. A good degree of nanoparticle dispersion in the system generally yields the sought properties in an efficient way. The main limitation in the widespread use of these nanocomposites is represented by the need to avoid the formation of filler particle agglomerates during the dispersion stage within the polymeric matrix. The poor control on the particle dispersion is also related to the limited knowledge of the particles-matrix interactions. The inadequate control of the interface formation at nanometric level also limits the possibility of thermomechanical improvements allowed by the use of these nanoparticles [16]. In this paper, AFM is used to examine the dispersion state of carbonaceous nanofillers within epoxy resin. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) are used to characterize the thermal behavior of both the epoxy matrix and the corresponding nanocomposites. FTIR, used to monitor the curing process of the samples at different temperatures and filler concentrations, shows that, among the analyzed carbonaceous nanoparticles, EG plays the role 1 of accelerating agent allowing the reduction of the band at 907 cm− due to the presence of oxirane rings as the nanofiller concentration increases. The results of FTIR investigation are in good agreement with DSC results. In fact, as verified through FTIR analysis, DSC data show that, for conversions up to 20%, the conversion of EG-based nanocomposites is shifted at temperature values lower than 10 ◦C (i.e. TBD + 6.5%EG sample) compared to the epoxy matrix TBD. In conclusion, AFM morphological investigation, as well as FTIR ad DSC tests, demonstrate that the good levels of carbon nanofiller dispersion into the epoxy mixture not only improve the properties of the nanocomposites, but also affect the curing process. Polymers 2019, 11, 832 3 of 20

2. Experimental

2.1. Materials and Manufacturing of Thermosetting Specimens Polymers 2019, 11, x FOR PEER REVIEW 3 of 21 In this work, the unﬁlled formulation TBD contains the epoxy precursor T, namely tetraglycidyl methyleneIn this dianiline work, the (TGMDA), unfilled formulation the reactive TB diluentD contains B, namely the epoxy 1-4 butanedioldiglycidylprecursor T, namely tetraglycidyl ether (BDE) mixedmethylene at 80% anddianiline 20% by(TGMDA) weight,,respectively the reactive anddiluent the B, hardener namely agent1-4 butanedioldiglycidyl D, namely 4,4-diaminodiphenyl ether (BDE) sulfonemixed (DDS), at 80% added and to the 20% TB epoxyby weight mixture, respectively in stoichiometric and the quantity. hardener The diluent agent BDED, namely (B) was 4,4 used- to adjustdiaminodiphenyl the viscosity sulfone of the mixture,(DDS), added in order to to the facilitate TB epoxy the mixture nanoparticle in stoichiometric dispersion in quantity. the resulting The thermosettingdiluent BDE composites, (B) was used and to adjust better the control viscosity the of degassing the mixture process, in order during to facilitate the curing the nanoparticle stage [17,18 ]. In addition,dispersion BDE in the (B) resulting was found thermosetting to be particularly composites, e ﬀandective better in control increasing the dega thessing degree process of curing during of nanochargedthe curing epoxystage [17 resins,18]. [In19 addition,]. In Figure BDE1, the(B) was chemical found formulas to be particularly of all the effective components in increasing (purchased the fromdegree Sigma-Aldrich) of curing of are nanocharged displayed. epoxy resins [19]. In Figure 1, the chemical formulas of all the components (purchased from Sigma-Aldrich) are displayed.

FigureFigure 1. Chemical 1. Chemical formulas formula ofs of tetraglycidyl tetraglycidyl methylene methylene dianiline dianiline (TGMDA)(TGMDA) (T) (T),, butanedioldiglycidyl butanedioldiglycidyl etherether (BDE) (BDE (B),) (B), and and 4,4-diaminodiphenyl 4,4-diaminodiphenyl sulfone sulfone (DDS) (DDS) (D). (D).

ThreeThree diff differenterent carbon carbon nanostructured nanostructured particles particles (MWCNTs, (MWCNTs, CNFs,CNFs, EG)EG) were used used as as filler filler in in the the epoxyepoxy matrix. matrix. In particular,In particular, MWCNTs MWCNTs (3100 (3100 Grade) Grade) were were acquired acquired fromfrom Nanocyl S.A. S.A. and and analyzed analyzed by transmissionby transmission electron electron microscopy microscopy (TEM). (TEM) The. The outer outer diameterdiameter andand the length of of these these MWCNTs MWCNTs rangerange from from 10 to10 30to 30 and an fromd from 100 100 to to 1000 1000 nm, nm respectively., respectively. Brunauer–Emmett–TellerBrunauer–Emmett–Teller method method has has allowedallowed us tous determineto determine the the specific specific surface surface area area of of MWCNTs MWCNTs whosewhose value is is about about 250 250–300–300 m m2/g.2/g. TheThe thermogravimetric thermogravimetric analysis analysis showed showed a carbona carbon purity purity greater greater than than 95%.95%. CNFsCNFs (PR25XTPS1100, (PR25XTPS110 Pyrograf0, Pyrograf III—Applied III—Applied Sciences Sciences Inc., Cedarville, Inc., Cedarville, OH, USA) OH, were USA produced) were by pyrolyticallyproduced by strippingpyrolytically at 1100 stripping◦C (here at named1100 °CCNFs (here as-received)named CNFs and as-received) subsequently and heat-treatedsubsequently at heat-treated at 2500 °C (here named CNFs heat-treated) to improve the electrical properties [3]. 2500 ◦C (here named CNFs heat-treated) to improve the electrical properties [3]. TheThe exfoliated exfoliated graphite graphite (EG) (EG) was was prepared prepared by by mixing mixing aa sulphonitratingsulphonitrating mixture with with natural natural graphitegraphite (Asbury (Asbury graphite graphite grade grade 3759, 3759, Asbury Asbury Carbons, Carbons, NJ), NJ), accordingaccording to the procedure procedure described described in in our previous paper [6]. EG particles are characterized by an average diameter of 500 µm and a degree our previous paper [6]. EG particles are characterized by an average diameter of 500 µm and a degree of exfoliated phase equal to 56% [6]. of exfoliated phase equal to 56% [6]. TGMDA (T), BDE (B) and DDS (D), giving the TBD sample, were mixed at 120 °C and then TGMDA (T), BDE (B) and DDS (D), giving the TBD sample, were mixed at 120 C and then added with 1D and 2D carbonaceous nanofillers at different load levels from 0.05 to 6.5%◦ by weight, added with 1D and 2D carbonaceous nanofillers at different load levels from 0.05 to 6.5% by weight, depending on the nanoparticle type. The mixtures were sonicated (Hielscher model UP200S) for 20 min at 24 kHz and 200W. Process conditions were chosen by considering the mechanical and electrical conductivity of the resulting composites [3,5,6]. The curing process was carried out in two isothermal stages: a first stage at 125 °C for 1 h followed by a second stage at higher temperatures reaching up to 180 and 200 °C for 3 h. This

Polymers 2019, 11, 832 4 of 20 depending on the nanoparticle type. The mixtures were sonicated (Hielscher model UP200S) for 20 min at 24 kHz and 200W. Process conditions were chosen by considering the mechanical and electrical conductivity of the resulting composites [3,5,6]. The curing process was carried out in two isothermal stages: a first stage at 125 ◦C for 1 h followed by a second stage at higher temperatures reaching up to 180 and 200 ◦C for 3 h. This procedure contemplates, for the first curing step, times and temperatures lower than those required for the second step. It corresponds to the common industrial conditions aimed at ensuring an optimal impregnation of the carbon fibers at lower temperature before the resin solidification at higher temperature takes place. The nanocomposites were coded as TBD + X(nanofiller), where X is the nanofiller percentage.

2.2. Characterizations Infrared spectroscopy (FTIR) tests were carried out with a Bruker Vertex70 FTIR-spectrophotometer 1 1 in the range of 4000-400 cm− , with a resolution of 2 cm− (32 scans collected). The infrared spectra were obtained in absorbance. Thermogravimetric analysis (TGA) was carried out with a Mettler TGA/SDTA 851 thermal 1 analyzer, in the range 25–1000 ◦C at a heating rate of 10 ◦C min− under air flow. Differential thermal analysis (DSC) was carried out with a Mettler DSC 822 differential scanning calorimeter in a flowing nitrogen atmosphere, in the temperature range of 30–300 ◦C with a scan rate 1 of 10 ◦C min− . Morphological investigation by atomic force microscopy (AFM) was carried out with a Dimension 3100 coupled with a Bruker NanoScope V multimode AFM (Digital Instruments, Santa Barbara, CA, USA). The image acquisitions were performed in tapping mode and ambient atmosphere. Commercial 1 probe tips with nominal spring constants of 20–100 N m− , resonance frequencies of 200–400 kHz, and tip radius of 5–10 nm were used. The images were analyzed using the Bruker software Nanoscope Analysis 1.80 (Build R1.126200). Sample slices were cut from samples by a sledge microtome and etched with potassium permanganate in a solution mixture of 95 mL sulfuric acid (95%–97%) and 48 mL orthophosphoric acid (85%) for 36 h. The samples were then washed in water and hydrogen peroxide and dried under vacuum for 5 days. Concerning the working principle of Atomic Force Microscopy (AFM), the AFM topographic images are achieved with a very sharp probe on the whole surface of the sample. The attractive force that develops when the tip approaches the surface, determines the deflection of the cantilever. The changes in sample topography produce a cantilever deflection towards or away from the surface. This deflection is detected by a laser beam, whose reflection moves on the surface of the position-sensitive photodiode. As the cantilever moves through its interaction with the surface, the power of the photodiode signal also changes. Thus, the photodiode registers the cantilever deflection and the subsequent change in direction of the reflected beam when an AFM tip crosses a lifted-up surface feature. The visualization of the topographic image can be obtained by scanning the cantilever on a portion of interest of the sample surface. The cantilever deflection, monitored by the position sensing photodiode, is strongly affected by the higher and lower morphological features appearing on the investigated sample parts. The AFM can create a well-defined topographic map of the surface characteristics through a feedback loop aimed at controlling the height of the tip above the surface [20–22]. Figure2 shows the schematic setup of a typical atomic force microscope AFM. The AFM measurements can be operated in three different modes: non-contact, contact and tapping mode. AFM tapping mode has been used to characterize polymeric surfaces with nanoscale resolution and minimum sample damage [23,24]. SEM micrographs were obtained using a field emission scanning electron microscope (FESEM, mod. LEO 1525, Carl Zeiss SMT AG, Oberkochen, Germany). Transmission electron microscopy (TEM) was carried out on a Philips CM-20 model equipped with a 200 kV accelerating voltage and a high brightness LaB6 gun for high coherence and a small probe. Polymers 2019, 11, 832 5 of 20 Polymers 2019, 11, x FOR PEER REVIEW 5 of 21

FigureFigure 2. 2.Schematic Schematic drawing of of an an atomic atomic force force microscope microscope (AFM) (AFM):: a sharp a sharp tip scans tip scansover the over the specimenspecimen surface, surface, the the piezoelectricpiezoelectric scanner scanner moves moves the the specimen specimen (or the (or probe the probe tip) intip) xyz indirections. xyz directions. A feedback-loopA feedback-loop controls controls thethe distance between between tip tip and and specimen, specimen, which which is evaluated is evaluated by a computer by a computer andand results results in in a a visual visualpicture picture of the sample sample topography. topography.

3. ResultsThe and AFM Discussion measurements can be operated in three different modes: non-contact, contact and tapping mode. AFM tapping mode has been used to characterize polymeric surfaces with nanoscale resolutionFTIR spectra and minimum of neatand sample nanocomposite damage [23,24 systems]. (0.32% by wt of nanofiller) cured up to 180 ◦C for 3 h areSEM shown micrographs in Figure were3. obtained As evident using from a field the emission figure, thescanning absorbing electron bands microscope of the O–H(FESEM, stretch at 1 3200–3650mod. LEO cm 1525,− increase Carl Zeiss as SMT the AG, reaction Oberkochen, proceeds. Germany). At the same time, the curing reaction produces 1 the disappearingTransmission of bothelectron asymmetric microscopy ring-stretching (TEM) was carried band out of on the a Philips epoxy ringCM-20 (907 model cm− equipped) and the N–H 1 stretchingwith a 200 vibration kV accelerating bands ofvoltage DDS (D)and (3060–3500a high brightness cm− ).LaB6 When gun thefor curinghigh coherence process and is complete, a small only veryprobe. little C–O–C band of the epoxy ring and N–H stretch bands are detectable. This means that, substantially, the totality of the amine and epoxy groups react giving rise to the crosslinked network. 3. Results and Discussion Besides, FTIR analysis confirms the effectiveness of the BDE (B) in determining an improvement in the FTIR spectra of neat and nanocomposite systems (0.32% by wt of nanofiller) cured up to 180°C curingPolymers degree 2019, 11 of, x epoxyFOR PEER nanocomposites REVIEW [19]. 6 of 21 for 3 h are shown in Figure 3. As evident from the figure, the absorbing bands of the O–H stretch at 3200–3650 cm-1 increase as the reaction proceeds. At the same time, the curing reaction produces the disappearing of both asymmetric ring-stretching band of the epoxy ring (907 cm−1) and the N–H stretching vibration bands of DDS (D) (3060–3500 cm−1). When the curing process is complete, only very little C–O–C band of the epoxy ring and N–H stretch bands are detectable. This means that, substantially, the totality of the amine and epoxy groups react giving rise to the crosslinked network. Besides, FTIR analysis confirms the effectiveness of the BDE (B) in determining an improvement in the curing degree of epoxy nanocomposites [19]. In this regard, Figure 3 clearly shows that the rate of epoxy/amine polymerization of TBD unfilled sample (with diluent), estimated by the progressive decrease of the epoxide band intensity, is higher than that of TD unfilled sample (without diluent).

FigureFigure 3. 3.FTIR FTIR spectraspectra of of neat neat and and nanocomposite nanocomposite systems systems cured cured up to up180°C to 180for 3◦ Ch. for 3 h.

The decrease of the epoxide group content of the TBD sample with respect to the TD sample is strongly evident also after an intermediate step of the curing process, at the temperature of 125°C for 1 h (see Figure 4). Among the different nanofillers, EG nanoparticles speed up the curing process of the epoxy resin with consequent decrease of the epoxy stretching frequency (907 cm−1) as it is evident from the comparison between epoxy samples filled with 0.64% by wt of MWCNTs, CNFs as-received, CNFs heat-treated and EG nanoparticles cured up to 125°C for 1 h (see Figure 4).

Figure 4. FTIR spectra of neat and nanocomposite systems cured up to 125°C for 1 h.

The effect of the reactive diluent BDE (B) on the viscosity of the epoxy matrix has been studied in previous papers [17,18]. As stated above, when it is mixed with the TGMDA (T) epoxy precursor, it decreases the viscosity of the epoxy matrix. The decrease of the amount of the epoxy groups of the TBD sample with respect to the TD sample is most likely due to the reduction of the viscosity of the

Polymers 2019, 11, x FOR PEER REVIEW 6 of 21

Polymers 2019, 11, 832 6 of 20

In this regard, Figure3 clearly shows that the rate of epoxy /amine polymerization of TBD unfilled sample (with diluent), estimated by the progressive decrease of the epoxide band intensity, is higher Figure 3. FTIR spectra of neat and nanocomposite systems cured up to 180°C for 3 h. than that of TD unfilled sample (without diluent). The decrease of the epoxide group content of the TBD sample with respect to the TD sample is The decrease of the epoxide group content of the TBD sample with respect to the TD sample is strongly evident also after an intermediate step of the curing process, at the temperature of 125 C for strongly evident also after an intermediate step of the curing process, at the temperature of 125°C◦ for 1 h (see Figure4). Among the di fferent nanofillers, EG nanoparticles speed up the curing process of the 1 h (see Figure 4). Among the different nanofillers, EG nanoparticles speed up the curing process of epoxy resin with consequent decrease of the epoxy stretching frequency (907 cm 1) as it is evident the epoxy resin with consequent decrease of the epoxy stretching frequency (907 cm−−1) as it is evident from the comparison between epoxy samples filled with 0.64% by wt of MWCNTs, CNFs as-received, from the comparison between epoxy samples filled with 0.64% by wt of MWCNTs, CNFs as-received, CNFs heat-treated and EG nanoparticles cured up to 125 C for 1 h (see Figure4). CNFs heat-treated and EG nanoparticles cured up to 125°C◦ for 1 h (see Figure 4).

Figure 4. FTIR spectra of neat and nanocomposite systems cured up to 125 ◦C for 1 h. Figure 4. FTIR spectra of neat and nanocomposite systems cured up to 125°C for 1 h. The effect of the reactive diluent BDE (B) on the viscosity of the epoxy matrix has been studied in previousThe effect papers of the [17 reactive,18]. As diluent stated above,BDE (B) when on the it isviscosity mixed withof the the epoxy TGMDA matrix (T) has epoxy been precursor, studied init decreasesprevious paper the viscositys [17,18]. of As the stated epoxy above, matrix. when The it decrease is mixed of with the amountthe TGMDA of the (T) epoxy epoxy groups precursor, of the itTBD decreases sample the with viscosity respect of to the the epoxy TD sample matrix. is The most decrease likely due of the to theamount reduction of the of epoxy the viscosity groups of of the the TBDepoxy sample matrix. with To betterrespect evaluate to the TD the sample curing is degree most oflikely the developeddue to the nanocomposites,reduction of the viscosity the FTIR of bands the characteristic of the epoxy group (in the range 860 cm 1 and 920 cm 1) and the strong asymmetric and − − 1 1 1 1 symmetric SO2 stretching peaks at 1143 cm− and 1105 cm− (in the range 1000 cm− and 1200 cm− ) were decomposed into their different components using a complex fitting in which a Lorentzian and Gaussian contribution were considered in the form: " # x x 2 H f (x) = (1 L)H exp 0 (4 ln 2) + L (1) − 2 − − w x x0 4 −w + 1 where x0 is the peak position, h is the height, w is the width at half-height and L is the Lorentzian component. The results of this deconvolution procedure for the sample TBD in the above-mentioned ranges of wavenumbers is shown in Figure5. Polymers 2019, 11, x FOR PEER REVIEW 7 of 21

epoxy matrix. To better evaluate the curing degree of the developed nanocomposites, the FTIR bands characteristic of the epoxy group (in the range 860 cm−1 and 920 cm−1) and the strong asymmetric and symmetric SO2 stretching peaks at 1143 cm−1 and 1105 cm−1 (in the range 1000 cm−1 and 1200 cm−1) were decomposed into their different components using a complex fitting in which a Lorentzian and Gaussian contribution were considered in the form:  2   x  x0  H f (x)  1 LH exp   4ln 2  L 2  w    x  x  (1)   4 0  1  w 

where x0 is the peak position, H is the height, w is the width at half-height and L is the Lorentzian

Polymerscomponent.2019, 11 ,The 832 results of this deconvolution procedure for the sample TBD in the above-mentioned7 of 20 ranges of wavenumbers is shown in Figure 5.

Figure 5. Figure 5.Deconvolution Deconvolution of of the the FTIR FTIR spectrum spectrum of of the the sample sample TBD TBD collected collected in in the the region region of of the the epoxy epoxy group (b) and the SO2 stretching signals (a). group (b) and the SO2 stretching signals (a). It is worth noting that the curing process does not influence the intensity of the signals It is worth noting that the curing process does not influence the intensity of the 1 signals corresponding to the strong asymmetric and symmetric SO2 stretching peaks at 1143 cm− and corresponding1 to the strong asymmetric and symmetric SO2 stretching peaks at 1143 cm−1 and 1105 1105 cm− . This allows us to use one of these bands as an internal reference to analyze the change cm−1. This allows us to use one of these bands as an internal1 reference to analyze the change in the in the intensity of the band due to the epoxy ring at 907 cm− , which is diagnostic for the evaluation intensity of the band due to the epoxy ring at 907 cm−1, which is diagnostic for the evaluation of the of the extent of the curing reactions and therefore it is directly related to the curing degree of the extent of the curing reactions and therefore it is directly related to the curing degree of the sample. sample. The described deconvolution procedure, shown for the sample TBD, has been applied to all The described deconvolution procedure, shown for the sample TBD, has been1 applied to all the1 the formulated nanocomposites. The ratio between the areas of band at 907 cm− and at 1105 cm− −1 −1 formulated nanocomposites. The ratio between the areas of band at 907 cm and at 1105 cm (1A (A (907)/A (1105)) and the ratio of the intensities (in terms of height of the signals) of band at 907 cm− −1 (907)/A (1105))1 and the ratio of the intensities (in terms of height of the signals) of band at 907 cm and at 1105 cm− (I (907)/I (1105)) are shown in Table1 for all the formulated nanocomposites. and at 1105 cm−1 (I (907)/I (1105)) are shown in Table 1 for all the formulated nanocomposites. 1 1 Table 1. Ratios between the areas and the intensities of the peaks at 907 cm− and 1105 cm− for the unfilled samples and the nanocomposites containing a percentage of 0.64 wt % of nanofiller.

Sample A (907)/A (1105) I (907)/I (1105) (cured up to 125 ◦C for 1 h) TD 0.149 0.413 TBD 0.055 0.126 TBD+0.64%EG 0.048 0.119 TBD+0.64%MWCNTs 0.100 0.250 TBD+0.64%CNFs 0.109 0.236 heat-treated TBD+0.64%CNFs 0.063 0.166 as-received

From data reported in Table1, it is evident that the presence of the reactive diluent BDE (B) determines a strong increase in the curing degree of the matrix. In fact, both the ratios, between the areas (A (907)/A (1105)) and the intensities (I (907)/I (1105)) of the peaks show a relevant reduction going from 0.149 to 0.055 for the ratio between the areas and from 0.413 to 0.126 for the ratio between the intensities. All the formulated nanocomposites, for the same curing treatment, highlight a higher Polymers 2019, 11, x FOR PEER REVIEW 8 of 21

Table 1. Ratios between the areas and the intensities of the peaks at 907 cm−1 and 1105 cm−1 for the unfilled samples and the nanocomposites containing a percentage of 0.64 wt % of nanofiller.

Sample A (907)/A (1105) I (907)/I (1105) (cured up to 125°C for 1 h) TD 0.149 0.413 TBD 0.055 0.126 TBD+0.64%EG 0.048 0.119 TBD+0.64%MWCNTs 0.100 0.250 TBD+0.64%CNFs heat-treated 0.109 0.236 TBD+0.64%CNFs as-received 0.063 0.166 From data reported in Table 1, it is evident that the presence of the reactive diluent BDE (B) determines a strong increase in the curing degree of the matrix. In fact, both the ratios, between the Polymersareas (A2019 (907)/, 11, 832A (1105)) and the intensities (I (907)/I (1105)) of the peaks show a relevant reduction8 of 20 going from 0.149 to 0.055 for the ratio between the areas and from 0.413 to 0.126 for the ratio between curingthe intensities. degree with All the respect formulated to the initial nanocomposites, sample TD (without for the same reactive curing diluent treatment, BDE (B)). highlight This proves a higher that thecuring reactive degree diluent with BDErespect (B) wellto the counterbalances initial sample theTD adverse(without e ffreactiveect of the diluent viscosity BDE increase, (B)). This due proves to the presencethat the reactive of nanofiller, diluent on BDE the (B) curing well processcounterbalances of the formulated the adverse nanocomposites. effect of the viscosity From increase, Table1, it due is alsoto the evident presence that of the nanofiller, sample EG, on the in percentage curing process 0.64 wtof the %, showsformulated a slight nanocomposites. increase in the curingFrom T degreeable 1, alsoit is withalso evident respect tothat the the sample sample TBD. EG, CNFs in percentage heat-treated, 0.64 MWCNTswt %, shows seem a slight to cause increase a slight in decrease the curing in thedegree curing also degree with respect with respect to the to sample the sample TBD. TBD; CNFs whereas heat-treated no remarkable, MWCNTs changes seem to are cause observed a slight for thedecrease sample in containingthe curing CNFsdegree (0.64 with wt respect %) as-received, to the sample which TBD; shows wher onlyeas ano very remarkable small decrease changes in theare curingobserved degree. for the In anysample case, containing also for MWCNTs, CNFs (0.64 CNFs wt %) heat-treated as-received, and which as-received, shows foronly a percentage a very small of nanofillerdecrease in of the 0.64 curing wt %, degree. the curing In any degree case, is also consistently for MWCNTs, higher CNFs than heat the sample-treated TD. and as-received, for a percentageFor the EG-based of nanofiller samples, of 0.64 an wt advancing %, the curing decrease degree of is the consistently epoxy stretching higher bandthan the with sample increasing TD. For the EG-based samples, an advancing decrease of the epoxy stretching band with increasing concentration is observed (see Figure6) for the same curing treatment (at 125 ◦C for 1h). concentration is observed (see Figure 6) for the same curing treatment (at 125°C for 1h).

Figure 6. FTIR spectra of EG-based nanocomposites. Figure 6. FTIR spectra of EG-based nanocomposites. 1 It can easily be seen in Figure6 that, during the curing process, the peaks at 1143 cm − and 1 −1 1105 cmIt can− attributableeasily be seen to in the Figure strong 6 asymmetricthat, during and the symmetriccuring process, SO2 stretchingsthe peaks at [ 251143] of cm the hardenerand 1105 DDScm−1 attributable (D) do not changeto the strong and also asymmetric the baseline andis symmetric the same. SO This2 stretching allows uss [25 a direct] of the visualization hardener DDS of the(D) trenddo not related change to and the also intensity the baseline of the epoxyis the same. group This as the allows concentration us a direct of visualization the nanofiller of increases.the trend However,related to thethe sameintensity deconvolution of the epoxy procedure group as performed the concentration for the FTIR of the spectra nanofiller of the increases. samples of However, Figure4, hasthe beensame applieddeconvolution to some procedure of the samples performed shown for in Figurethe FTI6.R For spectra instance, of the Figure samples7 shows of Figure the results 4, has obtained for the sample TBD + 6.5%EG. The results of the deconvolution procedure obtained for some of these samples (see Table2) confirm the trend observed in Figure6. The calculation of the ratios between the areas and the 1 1 intensities of the peaks at 907 cm− and 1105 cm− for these samples shows a reduction of the values by increasing the EG percentage; going from 0.048 (for the percentage of nanofiller 0.64 wt %) to 0.039 (for the percentage of nanofiller of 6.5 wt %). Although the differences in the numbered values of the ratios are not very high, the observed trend seems to confirm a progressive and slight increase in the curing degree with increasing the EG percentage. The study of the curing behavior of the nanocomposites during an intermediate stage (in our case after a curing treatment of 125 ◦C for 1 h) is relevant from an industrial point of view. In fact, in the case of structural panels, for example, Carbon Fiber-Reinforced Panels (CFRPs) impregnated with nanofilled resins, a two-steps curing cycle is frequently adopted. This allows, during the first step at lower temperature, a better impregnation of the carbon fabric. Together with the study of the Polymers 2019, 11, 832 9 of 20

resin behavior after the ﬁrst step of the curing cycle (1 h at 125 ◦C), it is also important the evaluation ofPolymers the curing 2019, 11 degree, x FOR ofPEER the REVIEW formulated nanocomposite after the complete curing cycle. DC values9 of of 21 EG-based nanocomposites after the complete curing cycle (a ﬁrst stage at 125 ◦C for 1 h followed by a been applied to some of the samples shown in Figure 6. For instance, Figure 7 shows the results second stage at higher temperatures reaching up to 180 and 200 ◦C for 3 h) are shown in Table3. obtained for the sample TBD + 6.5%EG.

Figure 7. Deconvolution of the FTIR spectrum of the sample TBD + 6.5 %EG collected in the region of Figure 7. Deconvolution of the FTIR spectrum of the sample TBD + 6.5 %EG collected in the region of the epoxy group (b) and the SO2 stretching signals (a). the epoxy group (b) and the SO2 stretching signals (a). 1 1 Table 2. Ratios between the areas and the intensities of the peaks at 907 cm− and 1105 cm− for samples containingThe results diﬀ oferent the percentages deconvolution of EG nanoparticles.procedure obtained for some of these samples (see Table 2) confirm the trend observed in Figure 6. The calculation of the ratios between the areas and the Sample intensities of the peaks at 907 cm−1 and 1105 cmA−1 (907) for /theseA (1105) samples I show (907)/Is (1105)a reduction of the values (cured up to 125 ◦C for 1 h) by increasing the EG percentage;TBD+0.64%EG going from 0.048 (for0.048 the percentage of0.119 nanofiller 0.64 wt %) to 0.039 (for the percentage of nanofillerTBD+3%EG of 6.5 wt %). Although0.042 the differences in 0.112 the numbered values of the ratios are not very high,TBD the+ observed6.5%EG trend seems to0.039 confirm a progressive0.062 and slight increase in the curing degree with increasing the EG percentage. Table 3. DC values of EG-based nanocomposites. -1 -1 Table 2. Ratios between the areas and the intensities of theCure peaks Degree at 907 (%) cm and 1105 cm for samples containing different percentagesEpoxy Formulation of EG nanoparticles. Oven cure (200 ◦C) TBD 91 Sample TBD+0.05%EG A (907)/A (1105)100 I (907)/I (1105) (cured up to 125°C for 1 h) TBD+0.64%EG 96 TBDTBD++0.641.5%EG%EG 0.048 100 0.119 TBD+33%EG%EG 0.042 97.5 0.112 TBDTBD++6.55%EG%EG 0.039 97.3 0.062 TBD+6%EG 98 The study of the curing behavior of the nanocomposites during an intermediate stage (in our caseData after shown a curing in Tabletreatment3 highlight of 125°C that for the 1 curingh) is relevant degree from of the an EG-based industrial nanocomposites point of view. isIn higher fact, in thanthe case 95% of for structural all the investigated panels, for ﬁller example percentages,, Carbon proving Fiber-Reinforced a good potentiality Panels (CFRPs) for eventual impregnated applications. with nanofilledIn agreement resins, witha two current-steps curing literature, cycle the is frequently curing degree adopted. (DC) hasThis been allows estimated, during from the first DSC step tests at assuminglower temperature that the exothermic, a better impregnation heat evolved during of the curecarbon is proportionalfabric. Together to the with extent the ofstudy reaction of the [26 r,27esin]. behavior after the first step of the curing cycle (1 h at 125°C), it is also important the evaluation of the curing degree of the formulated nanocomposite after the complete curing cycle. DC values of EG- based nanocomposites after the complete curing cycle (a first stage at 125°C for 1 h followed by a second stage at higher temperatures reaching up to 180 and 200° C for 3 h) are shown in Table 3.

Polymers 2019, 11, x FOR PEER REVIEW 10 of 21

Table 3. DC values of EG-based nanocomposites.

Cure Degree (%) Epoxy formulation Oven cure (200°C) TBD 91 TBD+0.05%EG 100 TBD+0.64%EG 96 TBD+1.5%EG 100 TBD+3%EG 97.5 TBD+5%EG 97.3 TBD+6%EG 98

Data shown in Table 3 highlight that the curing degree of the EG-based nanocomposites is higher than 95% for all the investigated filler percentages, proving a good potentiality for eventual Polymers 2019, 11, 832 10 of 20 applications. In agreement with current literature, the curing degree (DC) has been estimated from DSC tests assumingFigure8a shows that the the exothermic non-isothermal heat DSC evolved test results during carried cure is out proportional on epoxy resin, to the while extent Figure of8 b reaction shows [26,27the variation]. Figure of8a theshows fractional the non conversion-isothermal (DSCα) as test a function results carried of temperature out on epoxy for theresin epoxy., while A Figure linear 8baselineb shows hasthe variation been used of to the fit fractional the enthalpic conversion curve before() as a and function after the of temperature exothermic peak.for the The epoxy. curing A linearreaction baseline is considered has been complete used to whenfit the the enthalpic signal leveled curve obeforeff to the and baseline. after the The exothermic fractional conversionpeak. The curing(α) as areaction function is of considered temperature complete can be calculatedwhen the assignal follow: leveled off to the baseline. The fractional conversion () as a function of temperature can be calculated as follow: ∆H(T) α(T) =H(T) (2) (T)  ∆HTOT (2) H TOT where ∆H(T) is the heat released at temperature T and ∆HTOT is the total heat released during reaction. wBothhere quantitiesH(T) is the are heat calculated released integrating at temperature the area T and between HTOT theis the exothermic total heat curvereleased and during the baseline. reaction. Both quantities are calculated integrating the area between the exothermic curve and the baseline. a 1.5 b 1.0 1.0

[ ] 0.8

)



( 0.5 0.6

0.4 0.0

Conversion Conversion 0.2

Heat flow [W/g] exo -0.5 0.0 50 100 150 200 250 300 100 125 150 175 200 225 250 275 300 Temperature [°C] Temperature [°C]

Figure 8. Non-isothermal DSC enthalpy for the epoxy matrix (a); variation of the fractional conversion Figure 8. Non-isothermal DSC enthalpy for the epoxy matrix (a); variation of the fractional conversion (α) as a function of temperature for the epoxy matrix (b). () as a function of temperature for the epoxy matrix (b). The application of equation 2 allows for evaluation of the evolution of the fractional conversion as The application of equation 2 allows for evaluation of the evolution of the fractional conversion a function of the temperature. as a function of the temperature. For reactions carried out in isothermal conditions, the curing degree (DC) is most eﬀective for For reactions carried out in isothermal conditions, the curing degree (DC) is most effective for characterizing the sample properties. It can be determined by performing a series of isothermal characterizing the sample properties. It can be determined by performing a series of isothermal experiments at various temperatures. Thus, DC can be determined from the following equation: experiments at various temperatures. Thus, DC can be determined from the following equation:

∆Hiso(T) DC =∆퐻푖푠표(푇) 100 (3) 퐷퐶 = ∆HTOT ×× 100 (3) ∆퐻푇푂푇 where ∆Hiso(T) is the total heat of curing reaction at temperature T. where ΔHiso(T) is the total heat of curing reaction at temperature T. The curing degree (DC) for EG-filled samples cured at 200 ◦C is higher than that observed for unfilled epoxy resin, reaching 100% at the end of the test (see Table3). In any case, the DC values are higher than 90% for all the formulations cured at 200 ◦C. As verified in FTIR analysis of Figure6, the DSC analysis (see Figure9) show that, for conversions up to 20%, the conversion of the EG systems is shifted of about 10 ◦C at lower temperatures (i.e. TBD + 6.5%EG) with respect to unfilled epoxy resin TBD. For high degree of conversions, the differences obtained are minimal because the crosslinking phenomena of the matrix are governed exclusively by the curing temperature. Several authors found that, in the early curing stage, the acceleration effect is due to characteristic of the filler, such as functionalization, specific surface area, presence of catalyst and this effect is evident only for high concentrations [28]. In a previous paper [6], the presence of functional groups was found responsible for the acceleration of the curing reaction. Although the percentage by weight of functional groups present in the filler is low (oxygen content of 0.4 wt %) [6], the acceleration effect becomes relevant for high filler concentrations (6.5 wt %). Polymers 2019, 11, x FOR PEER REVIEW 11 of 21

The curing degree (DC) for EG-filled samples cured at 200°C is higher than that observed for unfilled epoxy resin, reaching 100% at the end of the test (see Table 3). In any case, the DC values are higher than 90% for all the formulations cured at 200°C. As verified in FTIR analysis of Figure 6, the DSC analysis (see Figure 9) show that, for conversions up to 20%, the conversion of the EG systems is shifted of about 10°C at lower temperatures (i.e. TBD + 6.5%EG) with respect to unfilled epoxy resin TBD. For high degree of conversions, the differences obtained are minimal because the crosslinking phenomena of the matrix are governed exclusively by the curing temperature. Several authors found that, in the early curing stage, the acceleration effect is due to characteristic of the filler, such as functionalization, specific surface area, presence of catalyst and this effect is evident only for high concentrations [28]. In a previous paper [6], the presence of functional groups was found responsible for the acceleration of the curing reaction. Although the percentage by Polymersweight2019 , of11 , functional 832 groups present in the filler is low (oxygen content of 0.4 wt %)11 [ of6], 20 the acceleration effect becomes relevant for high filler concentrations (6.5 wt %).

FigureFigure 9. Conversion 9. Conversion vs temperaturevs temperature for for EG-based EG-based epoxy epoxy nanocomposites. nanocomposites.

AllAll the the nanocomposites nanocomposites show show a higher a higher thermal thermal stability stability compared compared to unfilledto unfilled epoxy epoxy matrix matrix TBD TBD (see(see Figure Figure 10). 10 Thermogravimetric). Thermogravimetric analysis analysis of theof the nanocomposites nanocomposites filled filled with with 0.64 0.64 wt wt % of% carbonof carbon Polymersnanoparticlesnanoparticles 2019, 11, x FOR highlights highlights PEER REVIEW a stabilizing a stabilizing effect effect of theof the nanocomposites nanocomposites inthe in the first first degradation degradation stage.12 stage of 21 .

FigureFigure 10. 10.TGATGA curves curves in air in airof neat of neat and and nanocomposite nanocomposite systems. systems.

FigureFigure 11 11shows shows the the thermogravimetric thermogravimetric curves curves in inair air of of neat neat and and EG EG-based-based nano nanocomposites,composites, wherewhere a considerable a considerable increase increase in inthe the thermal thermal stability stability up up to to370°C 370 ◦isC observed is observed even even for fora very a very low low percentagepercentage of ofEG EG (0.05 (0.05 wt wt %). %). EG EG filled ﬁlled samples demonstratedemonstrate thethe advantages advantages that that can can be be obtained obtained from fromgraphene-based graphene-based materials materials [29 –[2933–].33].

Figure 11. TGA curves in air of neat and EG based nanocomposites.

AFM microscopy allows us to analyze the morphology and the properties of the sample surface up to the atomic scale [34,35]. Figures 12, 13 and 14 show the AFM micrographs and the corresponding 3D profiles of the fracture surface of the TBD+3%EG, TBD + 0.64%MWCNTs, and TBD + 1%CNFs heat-treated epoxy samples, respectively. We have chosen to investigate the morphology of these samples because the amount of nanofiller in the mentioned formulations is well above the electric percolation threshold [3,5,6] and therefore represents a good reference to evaluate the effectiveness of the nanofiller dispersion in the resin. In order to obtain reproducible results, the whole surface of the epoxy nanocomposites has been analyzed. In particular, several points of the

Polymers 2019, 11, x FOR PEER REVIEW 12 of 21

Figure 10. TGA curves in air of neat and nanocomposite systems.

Figure 11 shows the thermogravimetric curves in air of neat and EG-based nanocomposites, where a considerable increase in the thermal stability up to 370°C is observed even for a very low Polymers 2019, 11, 832 12 of 20 percentage of EG (0.05 wt %). EG filled samples demonstrate the advantages that can be obtained from graphene-based materials [29–33].

FigureFigure 1 11.1. TGATGA curves curves in in air air of of neat neat and and EG EG based based nanocomposites. nanocomposites.

AFM microscopy allows us to analyze the morphology morphology and and the the properties properties of the sample surface surface up to to the the atomic atomic scale scale [34 [,34,3535].]. Figures Figures 12 –12,14 show13 and the 14 AFM show micrographs the AFM and micrographs the corresponding and the corresponding3D profiles of the3D fractureprofiles of surface the fracture of the surface TBD+3%EG, of the TBD+3%EG, TBD + 0.64%MWCNTs, TBD + 0.64%MWCNTs and TBD +, and1%CNFs TBD +heat-treated 1%CNFs heat epoxy-treated samples, epoxyrespectively. samples, respectively. We have chosenWe have to chosen investigate to investigate the morphology the morphology of these ofsamples these samples because because the amount the amou of nanofillernt of nanofiller in the mentioned in the mentioned formulations formulations is well aboveis well the above electric the electricpercolation percolation threshold threshold [3,5,6] and [3,5,6 therefore] and represents therefore a represent good references a good to evaluate reference the to e ff evaluateectiveness the of effectivenessthe nanofiller of dispersion the nanofiller in the dispersion resin. In order in the to obtain resin. reproducibleIn order to obtain results, reproducible the whole surface results, of the the wholeepoxy nanocompositessurface of the epoxy has beennanocomposites analyzed. In has particular, been analyzed. several pointsIn particular, of the specimen several points surface of have the been scanned to assure the images to be representative. Since the AFM characterization was carried out on etched samples, the carbon nanofillers are well evident and separated from the surrounding epoxy matrix. For each sample, we report three AFM image types: height (or topography) amplitude and phase. The height (or topography) images are obtained by keeping constant the distance between the tip and the sample. These images are very useful as they allow us to estimate both lateral (xy) and height (z) dimensions of the sample features. Nevertheless, the reason why other types of image are usually reported, as in our case, is that such “height profiles” do not always resemble the object under investigation, or more precisely a certain shape may appear very different from what it would have in electron microscopy. This could have the consequence that, for an observer not particularly expert, such images do not easily show the shape of the sample features. The ways in which this happens include shading the image and, most habitually, creating a pseudo-3D image from height data. However, the amplitude (tapping mode) or deflection (contact mode) images represent an effective alternative, since they, being equivalent to a sample slope map, often show the sample shape more readily. It is important, however, to bear in mind that the z-scale in deflection or amplitude is completely meaningless in terms of sample structure. In fact, all that it shows is how the tip folded when it lightly taps the surface of the sample with consequent reduction of the oscillation amplitude. This change in amplitude is used by AFM to trace the surface topography. It is worth noting that the best images are obtained when the amplitude (or deflection) signals are minimized, since the amplitude (or deflection) images are the error signals in AFM technique. The phase images, obtainable in tapping mode, represent a map of how the phase of cantilever oscillation is influenced by its interaction with the surface. In addition to topographic information, the phase image provides insights on the local stiffness due to the direct relation to the material density and elastic modulus [11] thus resulting very Polymers 2019, 11, 832 13 of 20 Polymers 2019, 11, x FOR PEER REVIEW 14 of 21 sensitivethat the to TEM surface images stiffness of the/softness CNFs sample and adhesion (see Figure between 17) clearly the tip show and surface.the changes In fact, in morphology besides to of amplitudethe carbon variations, nanofibers when duea sharp to heatprobe-trea approachestment. In near fact, the the surface, carbon there nanofibers are also, phasebefore variations the thermal determinedtreatment by are a vertical characterized oscillation by ofthe the peculiar probe nearmorphology its mechanical called resonance “Dixie Cup” frequency. [3] where In particular,the CNFs as- the changesreceived of have the phasean orientation signal are analogous detectable to when that of the a probeset of meetsstacked areas Dixie of cups the sample with a withhollow diff coreerent (see composition.red ellipse) Phase. At changeslow temperatures, are recorded carbon as luminous manifests and only dusky local regions molecular in phase ordering images,. comparableAs soon as the withtemperature color contrasts of 2500 caused °C byis reached height changes, the warp that carbon are viewable graphene in height layers images. become In flattened, general, giving the phase rise to imagea straight is particularly structure useful, where in analyzing a minimum the surfaceinterlayer of mixedspacing (heterogeneous) was reached for samples the CNFs where, heat in- thistreated case,sample. it favors As explicit shown resolutionby a TEM micrograph of the various in materialthe Figure phases 17, the highlighted layers within by the a strong “Dixie contrast Cup” carbon of colorsnanofiber between have domains coalesced because following phase heat imaging-treatment. allows Th fore chemicalrigid and mapping aligned morphology of surfaces based of the on heat- thesetreated material carbon differences. nanofibers AFM is technique,distinctly visible through in thethe capturehigh-resolution of the height AFM (topography), micrographs amplitudeof the fracture andsurface phase data of the simultaneously, TBD + 1%CNFs allows heat surface-treated structure sample (see and Figure material 17 domains). to be directly compared.

Figure 12. Tapping mode AFM micrographs of the fracture surface of the TBD+3%EG sample (see on the top)Figure and 12. the Tapping corresponding mode AFM 3D profiles micrographs (see on of the the bottom). fracture surface of the TBD+3%EG sample (see on the top) and the corresponding 3D profiles (see on the bottom). Atomic force microscopy proved to be an excellent method for assessing the level of dispersion of carbon nanofillers within the polymer matrix for all formulated nanocomposites. The AFM micrographs allow us, in fact, to observe how the different carbon-based nanoparticles are effectively dispersed in the resin. Similar results are obtained for lower concentrations of carbon nanofillers inside the epoxy matrix. They uniformly cover the entire surface of the samples under investigation on which they appear distinctly with the peculiar morphological characteristics, thus also allowing us to irrefutably discriminate between the different type of nanofiller. It is worth noting that the nanoparticles are clearly visible on the surface thanks to the etching procedure, which consumed part of the resin from which they were covered, thus demonstrating the effective design of the nanocomposites that were able to resist the chemical action of the strongly oxidizing solution. In particular, EG-based samples show a very interesting morphological feature (see Figure 12). In this regard, on the fracture surface, we can observe folded graphene sheets that look like a draped fabric. High-resolution AFM micrographs (height and amplitude profiles) of the fracture surface of the TBD + 3%EG, TBD + 0.64%MWCNTs, and TBD+1%CNFs heat-treated epoxy samples, together with the TEM micrographs collected for the individual carbonaceous nanofillers EG, MWCNTs, and CNFs heat-treated are shown in Figure 15, Figure 16, and Figure 17, respectively. It is possible to visualize the morphological peculiarities of the nanoparticles more clearly and directly thanks to the TEM images that provide a valid support in distinguishing their presence inside the epoxy resin. It is worth noting that the TEM images of the CNFs sample (see Figure 17) clearly show the changes in morphology of the carbon nanofibers due to heat-treatment. In fact, the carbon nanofibers, before the thermal treatment

Figure 13. Tapping mode AFM micrographs of the fracture surface of the TBD+0.64%MWCNTs sample (see on the top) and the corresponding 3D profiles (see on the bottom).

Polymers 2019, 11, x FOR PEER REVIEW 14 of 21

that the TEM images of the CNFs sample (see Figure 17) clearly show the changes in morphology of the carbon nanofibers due to heat-treatment. In fact, the carbon nanofibers, before the thermal treatment are characterized by the peculiar morphology called “Dixie Cup” [3] where the CNFs as- received have an orientation analogous to that of a set of stacked Dixie cups with a hollow core (see red ellipse). At low temperatures, carbon manifests only local molecular ordering. As soon as the temperature of 2500 °C is reached, the warp carbon graphene layers become ﬂattened, giving rise to a straight structure, where a minimum interlayer spacing was reached for the CNFs heat-treated sample. As shown by a TEM micrograph in the Figure 17, the layers within the “Dixie Cup” carbon nanofiber have coalesced following heat-treatment. The rigid and aligned morphology of the heat- treated carbon nanofibers is distinctly visible in the high-resolution AFM micrographs of the fracture surface of the TBD + 1%CNFs heat-treated sample (see Figure 17).

Polymers 2019, 11, 832 14 of 20 are characterized by the peculiar morphology called “Dixie Cup” [3] where the CNFs as-received have an orientation analogous to that of a set of stacked Dixie cups with a hollow core (see red ellipse). At low temperatures, carbon manifests only local molecular ordering. As soon as the temperature of 2500 ◦C is reached, the warp carbon graphene layers become ﬂattened, giving rise to a straight structure, where a minimum interlayer spacing was reached for the CNFs heat-treated sample. As shown by a

TEM micrograph in the Figure 17, the layers within the “Dixie Cup” carbon nanoﬁber have coalesced following heat-treatment. The rigid and aligned morphology of the heat-treated carbon nanoﬁbers is distinctlyFigure visible 12. Tapping in the high-resolutionmode AFM micrographs AFM micrographs of the fracture of surface the fracture of the surface TBD+3%EG of the sample TBD +(see1%CNFs on heat-treatedthe top) sampleand the (seecorresponding Figure 17 ).3D profiles (see on the bottom).

FigureFigure 13. 13.Tapping Tapping mode mode AFM AFM micrographs micrographs of the of fracture the fracture surface surface of the TBD of the+0.64%MWCNTs TBD+0.64%MWCNTs sample Polymers(seesample 2019 on, the11 (see, x top) FOR on andthe PEER top) the REVIEW corresponding and the corresponding 3D proﬁles 3D (see profile on thes (see bottom). on the bottom). 15 of 21

Figure 14. Tapping mode AFM micrographs of the fracture surface of the TBD+1%CNFs heat-treated sampleFigure 14. (see Tapping on the top) mode and AFM the correspondingmicrographs of 3D the proﬁles fracture (see surface on the of bottom).the TBD+1%CNFs heat-treated sample (see on the top) and the corresponding 3D profiles (see on the bottom).

Figure 15. Tapping mode high-resolution AFM micrographs of the fracture surface of the TBD+3%EG sample (see on the left: (a) height profile, (b) amplitude profile) and the TEM micrograph of EG nanoparticles (see on the right).

Polymers 2019, 11, x FOR PEER REVIEW 15 of 21

Polymers 2019, 11, 832 15 of 20 Figure 14. Tapping mode AFM micrographs of the fracture surface of the TBD+1%CNFs heat-treated sample (see on the top) and the corresponding 3D profiles (see on the bottom).

FigureFigure 15. Tapping 15. Tapping mode mode high-resolution high-resolution AFM AFM micrographs micrographs of the of fracture the fracture surface surfaceof the TBD+3%EG of the TBD +3%EG sample (see on the left: (a) height profile, (b) amplitude profile) and the TEM micrograph of EG samplenanoparticles (see on the(see on left: the (righta) height). proﬁle, (b) amplitude proﬁle) and the TEM micrograph of EG nanoparticlesPolymers 2019, 11, (seex FORon PEER the REVIEW right). 16 of 21

FigureFigure 16. Tapping 16. Tapping mode mode high-resolutionhigh-resolution AFM AFM micrographs micrographs of the fracture of the surface fracture of the surface TBD+0.64% of the TBD+0.64% MWCNTsMWCNTs sample sample (see (se one on the the left:left: ( (aa) )height height profile, proﬁle, (b) amplitude (b) amplitude profile) proﬁle)and the TEM and micrograph the TEM micrograph of MWCNTsof MWCNTs nanoparticles nanoparticles (see (see on on the the right).right).

Polymers 2019, 11, 832 16 of 20

Polymers 2019, 11, x FOR PEER REVIEW 17 of 21

FigureFigure 17. 17Tapping. Tapping mode mode high-resolution high-resolution AFM AFM micrographs micrographs of ofthe the fracture fracture surface surface of the of TBD+1%CNFs the TBD+1%CNFs heat-treatedheat-treated sample sample (see (see on on thethe left: ( (aa) )height height profile, profile, (b) (amplitudeb) amplitude profile) profile) and the and TEM the micrograph TEM micrographss of CNFs nanoparticles, before and after heat-treatment (see on the right). of CNFs nanoparticles, before and after heat-treatment (see on the right). In order to analyze the homogeneity of the nanofiller dispersion in the polymeric matrix, the epInoxy order nanocomposites to analyze the with homogeneity EG, MWCNT ofs and the nanofillerCNFs heat-treated dispersion were ininvestigated the polymeric by means matrix, of the epoxy nanocompositesSEM. The analysis with was EG, carried MWCNTs out on etched and CNFssurface heat-treatedfracture of the werenanofilled investigated samples. Figure by meanss 18, of SEM. The analysis19 and 20 wasshow carried the SEM out images on etched of nanof surfaceilled epoxy fracture resins of at the the nanofilled loading rate samples. of 3% by Figureswt of EG, 18 –20 show the SEM0.64% imagesby wt of MWCNTs, of nanofilled and 1% epoxy by wt resins of CNFs at heat the-treated, loading respectively. rate of 3% by wt of EG, 0.64% by wt of MWCNTs,Painstaking and 1% observation by wt of CNFs highlights heat-treated, a homogeneous respectively. structure for all the samples, in which the carbonaceousPolymers 2019 nanofillers, 11, x FOR PEER are REVIEW uniformly distributed in the epoxy matrix, thus confirming18 of the 21 morphological results obtained by AFM technique. Besides, it should be pointed out that the structural features of all the nanofillers within the host matrix appear evident, as it is possible to detect by comparing them with the TEM images of the EG, MWCNTs, and CNFs heat-treated nanoparticles shown in the Figures 15, 16 and 17, respectively.

FigureFigure 18. Fracture 18. Fracture surface surface SEM SEM image ofof the the TBD TBD + 3%EG+ 3%EG sample. sample.

Figure 19. Fracture surface SEM image of the TBD+0.64%MWCNTs sample.

Polymers 2019, 11, x FOR PEER REVIEW 18 of 21

Polymers 2019, 11, 832 17 of 20 Figure 18. Fracture surface SEM image of the TBD + 3%EG sample.

Polymers 2019, 11, x FigureFOR PEER 19. REVIEWFractureFracture surface SEM image of the TBD+0.64%MWCNTs TBD+0.64%MWCNTs sample. 19 of 21

Figure 20. Fracture surface SEM image of the TBDTBD+1%CNFs+1%CNFs heat heat-treated-treated sample.

4. ConclusionsPainstaking observation highlights a homogeneous structure for all the samples, in which the carbonaceous nanofillers are uniformly distributed in the epoxy matrix, thus confirming the morphologicalThe analytical results potentiality obtained by of AFM AFM technique. technique Besides, in examining it should betopographical pointed out that and the nanometric structural resolutionfeatures of imaging all the nanofillers and its effectiveness within the host in matrix investigating appear the evident, dispe asrsion it is of possible different to detect types by of nanofillers are explored. Different epoxy nanocomposites have been obtained by adding carbon nanotubes (MWCNTs), carbon nanofibers (CNFs), and exfoliated graphite nanoparticles (EG) to an aeronautic epoxy resin. AFM, TEM, and SEM images highlight good levels of carbon nanofiller dispersion into the epoxy mixture, also allowing us to obtain information about the morphological feature of the nanostructured particles. Thermal and spectroscopic characterizations of these nanocomposites are also considered. FTIR characterization shows that the EG nanoparticles accelerate the curing process of the epoxy resin determining the progressive decrease of the epoxy stretching frequency (907 cm−1) with increasing nanofiller concentration. The accelerating effect is in good agreement with DSC results. The curing degree (DC) for the EG-based nanocomposites cured up to 200°C has been found very high compared to unfilled epoxy resin, reaching up to 100%. A substantial increase in the thermal stability up to 370 °C is also observed even for a very low EG percentage (0.05 wt %). A consequential result of this investigation lies in the possibility to exploit these promising properties together with the already assessed benefits related to the mechanical and electrical properties. In light of these results, the potential of graphene sheets as promising nanofiller for high- performance nanocomposites in the aeronautical field is well evident.

Author Contributions: Marialuigia Raimondo and Liberata Guadagno conceptualized and designed the experiments; Khalid Lafdi carried out the preparation of carbon-based nanoparticles; Marialuigia Raimondo carried out the preparation of the investigated epoxy samples, FTIR and AFM characterizations and related data analysis; Carlo Naddeo and Luigi Vertuccio performed the DSC and TGA characterizations and related graphic data processing; Liberata Guadagno has done data analysis; Andrea Sorrentino supervised the final manuscript and contributed to the interpretation of the experimental data; Marialuigia Raimondo wrote the paper. All authors read and approved the final manuscript.

Funding: The research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 760940 – MASTRO.

Polymers 2019, 11, 832 18 of 20 comparing them with the TEM images of the EG, MWCNTs, and CNFs heat-treated nanoparticles shown in the Figures 15–17, respectively.

4. Conclusions The analytical potentiality of AFM technique in examining topographical and nanometric resolution imaging and its effectiveness in investigating the dispersion of different types of nanofillers are explored. Different epoxy nanocomposites have been obtained by adding carbon nanotubes (MWCNTs), carbon nanofibers (CNFs), and exfoliated graphite nanoparticles (EG) to an aeronautic epoxy resin. AFM, TEM, and SEM images highlight good levels of carbon nanofiller dispersion into the epoxy mixture, also allowing us to obtain information about the morphological feature of the nanostructured particles. Thermal and spectroscopic characterizations of these nanocomposites are also considered. FTIR characterization shows that the EG nanoparticles accelerate the curing process 1 of the epoxy resin determining the progressive decrease of the epoxy stretching frequency (907 cm− ) with increasing nanofiller concentration. The accelerating effect is in good agreement with DSC results. The curing degree (DC) for the EG-based nanocomposites cured up to 200 ◦C has been found very high compared to unfilled epoxy resin, reaching up to 100%. A substantial increase in the thermal stability up to 370 ◦C is also observed even for a very low EG percentage (0.05 wt %). A consequential result of this investigation lies in the possibility to exploit these promising properties together with the already assessed benefits related to the mechanical and electrical properties. In light of these results, the potential of graphene sheets as promising nanofiller for high-performance nanocomposites in the aeronautical field is well evident.

Author Contributions: M.R. and L.G. conceptualized and designed the experiments; K.L. carried out the preparation of carbon-based nanoparticles; M.R. carried out the preparation of the investigated epoxy samples, FTIR and AFM characterizations and related data analysis; C.N. and L.V. performed the DSC and TGA characterizations and related graphic data processing; L.G. has done data analysis; A.S. supervised the final manuscript and contributed to the interpretation of the experimental data; M.R. wrote the paper. All authors read and approved the final manuscript. Funding: The research leading to these results has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 760940 – MASTRO. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Binnig, G.; Rohrer, H.; Gerber, Ch.; Weibel, E. Surface Studies by Scanning Tunneling Microscopy. Phys. Rev. Lett. 1982, 49, 57–60. [CrossRef] 2. Friedbacher, G.; Prohaska, T.; Grasserbauer, M. Surface Analysis with Atomic Force Microscopy through Measurement in Air and Under Liquids. Mikrochim. Acta 1994, 113, 179–202. [CrossRef] 3. Guadagno, L.; Raimondo, M.; Vittoria, V.; Vertuccio, L.; Lafdi, K.; De Vivo, B.; Lamberti, P.; Spinelli, G.; Tucci, V. The role of carbon nanofiber defects on the electrical and mechanical properties of CNF-based resins. Nanotechnology 2013, 24, 305704. [CrossRef][PubMed] 4. Zhou, Y.X.; Wu, P.X.; Cheng, Z.Y.; Ingram, J.; Jeelani, S. Improvement in electrical, thermal and mechanical properties of epoxy by filling carbon nanotube. Express Polym. Lett. 2008, 2, 40–48. [CrossRef] 5. Guadagno, L.; Naddeo, C.; Raimondo, M.; Barra, G.; Vertuccio, L.; Russo, S.; Lafdi, K.; Tucci, V.; Spinelli, G.; Lamberti, P. Influence of carbon nanoparticles/epoxy matrix interaction on mechanical, electrical and transport properties of structural advanced materials. Nanotechnology 2017, 28, 094001. [CrossRef] 6. Guadagno, L.; Raimondo, M.; Vertuccio, L.; Mauro, M.; Guerra, G.; Lafdi, K.; De Vivo, B.; Lamberti, P.; Spinelli, G.; Tucci, V. Optimization of graphene-based materials outperforming host epoxy matrices. RSC Adv. 2015, 5, 36969–36978. [CrossRef] 7. Guadagno, L.; Vietri, U.; Raimondo, M.; Vertuccio, L.; Barra, G.; De Vivo, B.; Lamberti, P.; Spinelli, G.; Tucci, V.; De Nicola, F.; et al. Correlation between electrical conductivity and manufacturing processes of nanofilled carbon fiber reinforced composites. Compos. Part B Eng. 2015, 80, 7–14. [CrossRef] Polymers 2019, 11, 832 19 of 20

8. Guadagno, L.; Naddeo, C.; Raimondo, M.; Barra, G.; Vertuccio, L.; Sorrentino, A.; Binder, W.H.; Kadlec, M. Development of self-healing multifunctional materials. Compos. Part B Eng. 2017, 128, 30–38. [CrossRef] 9. Gojny, F.H.; Wichmann, M.H.G.; Fiedler, B.; Kinloch, I.A.; Bauhofer, W.; Windle, A.H.; Schulte, K. Evaluation and identification of electrical and thermal conduction mechanisms in carbon nanotube/epoxy composites. Polymer 2006, 47, 2036–2045. [CrossRef] 10. Gallego, M.M.; Verdejo, R.; Khajet, M.; Ortiz de Zarate, J.M.; Essalhi, M.; Lopez-Manchado, M.A. Thermal conductivity of carbon nanotubes and graphene in epoxy nanofluids and nanocomposites. Nanoscale Res. Lett. 2011, 6, 610. [CrossRef] 11. Raimondo, M.; Guadagno, L.; Speranza, V.; Bonnaud, L.; Dubois, P.; Lafdi, K. Multifunctional graphene/POSS epoxy resin tailored for aircraft lightning strike protection. Compos. Part B Eng. 2018, 140, 44–56. [CrossRef] 12. Raimondo, M.; Guadagno, L.; Vertuccio, L.; Naddeo, C.; Barra, G.; Spinelli, G.; Lamberti, P.; Tucci, V.; Lafdi, K. Electrical conductivity of carbon nanofiber reinforced resins: Potentiality of Tunneling Atomic Force Microscopy (TUNA) technique. Compos. Part B Eng. 2018, 143, 148–160. [CrossRef] 13. Meenakshi, K.S.; Sudhan, E.P.J. Development of novel TGDDM epoxy nanocomposites for aerospace and high performance applications—Study of their thermal and electrical behaviour. Arab. J. Chem. 2016, 9, 79–85. [CrossRef] 14. May, C.A. Epoxy Resins, Chemistry and Technology, 2nd ed.; Marcel Dekker: New York, NY, USA, 1988. 15. Yasmin, A.; Daniel, I.M. Mechanical and thermal properties of graphite platelet/epoxy composites. Polymer 2004, 45, 8211–8219. [CrossRef] 16. Šupová, M.; Martynková, G.S.; Barabaszová, K. Effect of Nanofillers Dispersion in Polymer Matrices: A Review. Sci. Adv. Mater. 2011, 3, 1–25. [CrossRef] 17. Guadagno, L.; Raimondo, M.; Lafdi, K.; Fierro, A.; Rosolia, S.; Nobile, M.R. Influence of nanofiller morphology on the viscoelastic properties of CNF/epoxy resins. AIP Conf. Proc. 2014, 1599, 386–389. [CrossRef] 18. Nobile, M.R.; Raimondo, M.; Lafdi, K.; Fierro, A.; Rosolia, S.; Guadagno, L. Relationships between nanofiller morphology and viscoelastic properties in CNF/epoxy resins. Polym. Compos. 2015, 36, 1152–1160. [CrossRef] 19. Guadagno, L.; Raimondo, M.; Vittoria, V.; Vertuccio, L.; Naddeo, C.; Russo, S.; De Vivo, B.; Lamberti, P.; Spinelli, G.; Tucci, V. Development of epoxy mixtures for application in aeronautics and aerospace. RSC Adv. 2014, 4, 15474–15488. [CrossRef] 20. De Sousa, F.D.B.; Scuracchio, C.H. The Use of Atomic Force Microscopy as an Important Technique to Analyze the Dispersion of Nanometric Fillers and Morphology in Nanocomposites and Polymer Blends Based on Elastomers. Polímeros 2014, 24, 661–672. [CrossRef] 21. Palacio, M.L.B.; Bhushan, B. Normal and Lateral Force Calibration Techniques for AFM Cantilevers. Crit. Rev. Solid State Mater. Sci. 2010, 35, 73–104. [CrossRef] 22. Sawyer, L.C.; Grubb, D.T.; Meyers, G.F. Polymer Microscopy; Springer: New York, NY, USA, 2008. 23. Jalili, N.; Laxminarayana, K. A review of atomic force microscopy imaging systems: Application to molecular metrology and biological sciences. Mechatronics 2004, 14, 907–945. [CrossRef] 24. Igarashi, T.; Fujinami, S.; Nishi, T.; Asao, N.; Nakajima, K. Nanorheological Mapping of Rubbers by Atomic Force Microscopy. Macromolecules 2013, 46, 1916–1922. [CrossRef] 25. Pretsch, E.; Bühlmann, P.; Badertscher, M. Structure Determination of Organic Compounds “Tables of Spectral Data Fourth, Revised and Enlarged Edition”; Springer: Berlin/Heidelberg, Germany, 2009. 26. Horie, K.; Hiura, H.; Sawada, M.; Mita, I.; Kambe, H. Calorimetric investigation of polymerization reactions. III. Curing reaction of epoxides with amines. J. Polym. Sci. A Polym. Chem. 1970, 8, 1357–1372. [CrossRef] 27. Edwards, G.D.; Ng, Q.Y. Elution behavior of model compounds in gel permeation chromatography. J. Polym. Sci. Polym. Symp. 1968, 21, 105–117. [CrossRef] 28. Vertuccio, L.; Russo, S.; Raimondo, M.; Lafdi, K.; Guadagno, L. Influence of carbon nanofillers on the curing kinetics of epoxy-amine resin. RSC Adv. 2015, 5, 90437–90450. [CrossRef] 29. Guadagno, L.; Sarno, M.; Vietri, U.; Raimondo, M.; Cirillo, C.; Ciambelli, P. Graphene-based structural adhesive to enhance adhesion performance. RSC Adv. 2015, 5, 27874–27886. [CrossRef] 30. Yan, N.; Buonocore, G.; Lavorgna, M.; Kaciulis, S.; Balijepalli, S.K.; Zhan, Y.; Xia, H.; Ambrosio, L. The role of reduced graphene oxide on chemical, mechanical and barrier properties of natural rubber composites. Compos. Sci. Technol. 2014, 102, 74–81. [CrossRef] 31. Scherillo, G.; Lavorgna, M.; Buonocore, G.G.; Zhan, Y.H.; Xia, H.S.; Mensitieri, G.; Ambrosio, L. Tailoring assembly of reduced graphene oxide nanosheets to control gas barrier properties of natural rubber nanocomposites. ACS Appl. Mater. Interfaces 2014, 6, 2230–2234. [CrossRef] Polymers 2019, 11, 832 20 of 20

32. Geim, A.K.; Novoselov, K.S. The rise of graphene. Nat. Mater. 2007, 6, 183–191. [CrossRef] 33. Dikin, D.A.; Stankovich, S.; Zimney, E.J.; Piner, R.D.; Dommett, G.H.B.; Evmenenko, G.; Nguyen, S.T.; Ruoﬀ, R.S. Preparation and characterization of graphene oxide paper. Nature 2007, 448, 457–460. [CrossRef] 34. Calheiros Gomes, L.; Mergulhão, F.J. SEM Analysis of Surface Impact on Bioﬁlm Antibiotic Treatment. Scanning 2017, 2017, 2960194. 35. Zhang, H.; Li, X.; Chen, Y.; Park, J.; Li, A.P.; Zhang, X.G. Postprocessing Algorithm for Driving Conventional Scanning Tunneling Microscope at Fast Scan Rates. Scanning 2017, 2017, 1097142. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).