- Email: [email protected]
Carbon nanofibers in polyurethane foams: Experimental evaluation of thermo-hygrometric and mechanical performance
Polymer Testing 67 (2018) 234–245
Contents lists available at ScienceDirect
Polymer Testing journal homepage: www.elsevier.com/locate/polytest
Carbon nanofibers in polyurethane foams: Experimental evaluation of thermo-hygrometric and mechanical performance
T
F. Stazia,∗, F. Tittarellia, F. Saltarellia, G. Chiappinib, A. Morinic, G. Cerric, S. Lencid a
Dipartimento di Scienze e Ingegneria della Materia, dell’Ambiente ed Urbanistica (SIMAU), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy b Dipartimento di Ingegneria Industriale e Scienze Matematiche (DIISM), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy c Dipartimento di Ingegneria dell'informazione (DII), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy d Dipartimento di Ingegneria Civile e dell’Architettura (DICEA), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy
A R T I C L E I N F O
A B S T R A C T
Keywords: Polyurethane foam Carbon nanofibers Functionalization process Magnetic field alignment Microstructure Mechanical behavior Thermo-hygrometric behavior
Polymer nanocomposites synergistically combine the good thermal properties of the hosting polymer matrix with the high mechanical performance of the fillers, providing a new class of materials with superior properties. The present study aims to evaluate in a multidisciplinary way the enhancement in mechanical and thermalhygrometric properties of low and medium density nanophased polyurethane (PUR) foams with either randomly oriented or aligned nanofibers as compared to the neat ones. To this aim, 1% weight of carbon nanofibers (CNFs) were homogeneously dispersed into polyol of PUR foam by an ultrasonic cavitation method. In parallel, a small amount of CNFs was functionalized in advance by a coprecipitation method so as to align them into the polymer matrix through an external low intensity magnetic field. SEM analyses were used to compare the microstructure of the neat and nanophased samples. Results have shown that the addition of carbon nanofibers in the foams products a closer structure with a more uniform size and shape. Moreover, functionalized CNFs play a significant role in regulating cells shape as well as strengthening cells walls. Mechanical test results also demonstrated that CNFs increase both strength and stiffness of the samples. The alignment of carbon nanofibers within medium density nanophased foams determines the highest mechanical properties. However, the more noticeable improvement in samples performance occurred in low density nanophased foams. Finally, carbon nanoparticles decrease the thermal conductivity and increase the resistance against water adsorption.
1. Introduction Nanoscience and nanotechnology consist of understanding and controlling the materials at a nanoscale which is a billion part of the meter (10−9 m). Moreover, the real significance of the nanoscale is that the materials obtain new properties at this level [1]. Polymer nanocomposites have been drawing a great deal of interest due to their high potential to achieve improvements in their performance by adding a small amount of nanoparticles in polymer matrices. In particular, polyurethane foams belong to lightweight and energy efficient materials for buildings [2], but their application is limited because of low mechanical strength, poor surface quality, low thermal
as well as dimensional stability. The introduction of fillers in the formulation of building materials, thanks to a refinement of the microstructure, leads to an increase in durability and mechanical performance [3]. At present, carbon nanofibers (CNFs), carbon nanotubes (CNTs), nanoclay and oxides such as SiO2 and TiO2 are some of the most commonly adopted fillers in the field of polymer nanocomposites. Therefore, a small amount of well-dispersed nanoparticles in polymer matrix may significantly improve a wide variety of properties without sacrificing the lightweight structure of polymer foams. Vapor-grown carbon nanofibers (VGCNFs), commonly named CNFs and used in the present study, are multiwall and highly graphitic fibers
∗
Corresponding author. E-mail addresses: [email protected] (F. Stazi), [email protected] (F. Tittarelli), [email protected] (F. Saltarelli), [email protected] (G. Chiappini), [email protected] (A. Morini), [email protected] (G. Cerri), [email protected] (S. Lenci). https://doi.org/10.1016/j.polymertesting.2018.01.028 Received 18 September 2017; Received in revised form 4 December 2017; Accepted 27 January 2018 Available online 07 February 2018 0142-9418/ © 2018 Elsevier Ltd. All rights reserved.
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
2. Materials
with diameters ranging from 70 to 200 nm and lengths up to a few hundred microns. Due to their high thermal stability, tensile strength and Young's modulus, CNFs have demonstrated their great potential in nanoscale polymer reinforcements. Indeed, nanofibers act as nucleation sites to facilitate the bubble nucleation process, leading to an enhancement in mechanical and thermal performance because of a finer cell structure of the hosting foam [4–7]. However, due to their high surface energy, nanofibers tend to agglomerate, bundle together and entangle. These drawbacks may lead to many defect sites in the composites having detrimental effects on polymer performances and limiting the efficiency of CNFs in polymer matrices. Thence, it is imperative to well disperse nanofibers into the polymer matrix when they are still in liquid phase so that interactions at the molecular level can be achieved in order to produce a material with superior thermal and mechanical properties [8]. In the literature, there are several techniques to improve the dispersion of nanoparticles in polymer matrices, such as by optimum physical blending, in situ polymerization and chemical functionalization. For polymer/CNFs composites, high power dispersion methods, such as ultrasound and high speed shearing, are considered the simplest and most convenient ones to reach a homogeneous dispersion of CNFs [9,10]. Moreover, apart from achieving a uniform dispersion, the alignment of carbon nanofibers along a predetermined direction has been found to produce outstanding improvements in mechanical and thermal performance when compared to their randomly-oriented counterparts [11]. Several techniques with the aim of aligning nanoparticles have been reported in literature [12–15]. Among them the application of an external magnetic field is recognized as one of the most effective and simple methods. However, due to the low magnetic susceptibility of nanofibers, an extremely strong magnetic field (ranging from 25 to 30 T) would be required to achieve a satisfying alignment [16]. Thus, the need of employing such high magnetic field limits the practical application of this method. Therefore, it is crucial to functionalize carbon nanofibers in advance by coating them with magnetic materials to align them in a polymer matrix under a relatively weak magnetic field, ranging from 50 mT to 1 T [17,18]. In summary, to the best of our knowledge, the majority of research available in the present literature regards the experimental evaluation of nanocomposites mechanical performance [4,7,10,18,19]. Less common are multidisciplinary studies which synergistically combine both thermal and mechanical aspects of polymer nanocomposites [5,8,11,16]. Moreover, polymers are generally tested in terms of decomposition temperature [20] without addressing the thermal conductivity issue. Therefore, the present work is intended to quantify the enhancement in mechanical and thermal-hygrometric properties of nanophased foams with low and medium densities, containing both randomly oriented and aligned carbon nanofibers, as compared to neat ones. To this aim, a small amount of CNFs has been functionalized in advance by a co-precipitation method in accordance with a previous study carried out by Shuying Wu and co-workers [18]. 1 wt% of nanofibers were dispersed homogeneously into polyol of PUR foam by an ultrasonic cavitation method. Subsequently, nanophased foam was obtained by adding diisocyanate to the mixture and pouring it in a mould whilst functionalized CNFs were aligned through an external magnetic field of about 90 mT (value measured in the proximity of the magnetic disks). Finally, the properties of the samples were compared in terms of microstructure, mechanical performance, thermal conductivity, hygroscopic adsorption and apparent density.
2.1. Polyurethane foams Polyurethane foams of two different apparent densities were used. The foam consists of two liquid precursors: part-A is a diphenylmethane diisocyanate polymer and part-B is a polyol. It was supplied by Claudio Foresi s.r.l. and has a density of 30 kg/m3. In addition, a higher density PUR foam was obtained from the same precursors by varying the foam production process through the adoption of a closing cap on the top of the mould, thus restraining the foam rise. 2.2. Carbon nanofibers (CNFs) According to material data supplied by Tech-Star s.r.l., the vapor grown carbon nanofibers (VGCNFs) adopted have an average diameter in the range of 30–50 nm and a length of 10–30 μm. 2.3. Magnetic carbon nanofibers (Fe3O4@CNFs) One of the main aims of this study was to demonstrate the improvement in thermal and mechanical properties of PUR foam by the introduction of CNFs aligned by a weak magnetic field. However, since the CNFs have low magnetic susceptibility, their surface requires to be covered in advance with materials characterized by strong magnetic properties such as iron oxide. Therefore, the present section describes the method adopted to functionalize CNFs and the preliminary analysis to evaluate the magnetic properties of Fe3O4@CNFs. The fabrication methods of magnetic CNFs-embedded composites have been extensively investigated in literature. A detailed review could be found in Ref. [9]. For the present study, a simple co-precipitation method, developed by Shuying Wu et al. [18], was adopted to functionalize carbon nanoparticles and align CNFs through a relatively low magnetic field. Indeed, thanks to the functionalization process, fillers may show excellent magnetic properties. The original CNFs were first subjected to an oxidation treatment in the nitric acid to modify the surface in order to achieve a better dispersion. Subsequently, 2 g of as-received CNFs were mixed with 200 mL of concentrated nitric acid whilst stirring vigorously. This mixture was then treated at 100 °C for 6 h under magnetic stirring. Afterwards, the mixture was washed several times using deionized water until reaching a pH value of ∼7. The samples were vacuum filtrated and dried in a vacuum oven. After this acid treatment, the CNFs are expected to possess oxygencontaining functional groups on their surfaces and are denoted by CNFs-OX. The functionalized nanoparticles were fabricated by a co-precipitation method from the CNFs-OX material, prepared as described above. Firstly, 0.40 g of the CNFs-OX were dispersed in 355 mL distilled water by sonication for 15 min and 0.40 g of Fe3O4 was added whilst stirring. Subsequently, the mixture was vigorously stirred for 15 min whilst being heated to 50 °C under a nitrogen (N2) atmosphere. Then 0.32 g of FeSO4⋅7H2O were added with continuous stirring under a N2 atmosphere for 30 min. Next, 27 mL of 8M NH4OH aqueous solution were added drop-wise to precipitate ferric and ferrous salt. The pH of the mixture was kept at ∼10 and the reaction was carried out at 50 °C for 30 min under vigorous magnetic stirring. N2 was continuously purged during the reaction to prevent oxidation.
235
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
longitudinal axis, coincident with the foam rise direction, through a pair of magnetic disks, using the above mentioned methods. Magnets were thus placed at a mutual distance of 20 cm, respectively at the top and at the bottom of the mould. The alignment of the Fe3O4@CNFs was imposed during the foaming process, prior to the hardening of the PUR, using a magnetic field of 90 mT in the proximity of the magnetic disks. Table 1 shows the overall number of specimens (type A and type B), containing various percentages of nanoparticles, which were manufactured in order to carry out mechanical tests. All specimens were stored under ambient laboratory conditions (temperature of 23 ± 2 °C, relative humidity of 50 ± 5%) for 30 days, as recommended by manufactures. Three samples for each type of foams were considered to determine the average value of mechanical strength and modulus. Further samples were also realized to conduct different analysis.
The Fe3O4@CNFs were obtained by magnetic separation, washed with distilled water and ethanol, and finally dried under vacuum at 50 °C.
3. Manufacturing The fabrication of nanophased polyurethane foams was conducted in three steps. In step 1, a pre-calculated amount of nanoparticles was mixed with polyol of PUR foam by a mechanical stirrer in order to have a specific weight ratio. Polyol was selected for infusion of nanoparticles since it is the least reactive part [8]. In this study, 1 wt% of CNFs and 1 wt% of Fe3O4@CNFs were chosen to conduct the experimental tests. Indeed, previous studies have found that the improvement in polymer foam properties may be achieved by infusion of a small percentage of nanoparticles, especially in the range of 0.5–2 wt% [19–21]. Ultrasonic cavitation method was chosen to disperse nanoparticles into polymer since this method was found to be the most effective in producing a homogeneous dispersion of nanoparticles [22]. It involves the formation, growth, pulsating and collapsing of tiny bubbles, producing transient micro hot spots which enable to enhance the wettability of nanoparticles. The ultrasonic wave can break the agglomerating bodies by damaging the Coulomb and Van der Waals's forces between the particles and make them homogeneously dispersed in the liquid [8]. The mixture was thus subjected to ultrasonic cavitation using a high intensity ultrasonic horn (Bandelin Sonoplus) at the amplitude of 40% and for 10 min, as shown in Fig. 1a. In step 2, diisocyanate was added to the mixture at the ratio of 48–52 wt% using a mechanical blender for about 10 s. In step 3, the mixture was poured in a wooden mould and kept inside for 15 min. In particular, in order to obtain specimens with different densities and sizes, two types of moulds were adopted: type A with the size of 20 × 5 × 3 cm3 and type B with the size of 50 × 50 × 7.2 cm3. Type A moulds were closed (Fig. 1b) by a contrasting element on the upper side and compressed by a concrete cube. This technique let us produce foams whose nominal density was 50 kg/m3. Moreover, type A samples were studied to maintain the external surface veil for mechanical tests to reproduce the real conditions of this application in buildings (Fig. 1c-d). Type B moulds were open on the upper side as shown in Fig. 2a-b. Type B samples for each foam typology were cut into the suitable shape, then the external surface veil was removed to conduct the tests. The nominal density of the foams in thus type of mould was 30 kg/m3. Neat and nanophased PUR foams were fabricated for comparison. A stoichiometric amount of Fe3O4@CNFs was only added to high density PUR foams and nanoparticles were aligned along the sample
4. Experimental methods 4.1. Measurement of the magnetic field intensity The magnetic field intensity was measured using the 5170 Gauss Tesla Meter probe, equipped with a hold sensor, which allowed recording both longitudinal (BH) and radial (BR) components of the magnetic flux density. Two neodymium (NdFeB) magnetic disks, covered in Nichel, with a diameter of 60 mm and thickness of 5 mm were used to conduct the tests. They were placed on the edges of a rectangular box at a mutual distance of 5 cm. Preliminary probe calibration was necessary to cut out external interferences. Measurements were carried out using a grid of dots drawn on a graph paper in order to make the method repeatable. Firstly, the magnetic flux density was estimated in the absence of nanoparticles. After that, Fe3O4@CNFs were poured into the box as shown in Fig. 3a-b, then we observed their effect on the variation of the magnetic field intensity. The figure clearly shows that all the nanoparticles have assumed magnetic susceptibility since successfully attracted by magnetic poles. 4.2. Microstructure The overall morphology of neat and nanophased foams was studied using Philips XL20 scanning electron microscope (SEM). Higher magnified analyses upon average cell size and cell wall thickness were conducted with Zeiss Supra 40, which also provided details of single nanoparticles such as average diameter and length. Small amounts of nanoparticles were placed on one side of a doublesided carbon tape (opposite side is attached to the sample holder) for
Fig. 1. Ultrasonic cavitation method (a); type A mould between a couple of magnetic dicks (b); type A 50_neat (c) and 1% Fe3O4@CNFs PUR foam (d), covered with the external surface veil.
236
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 2. Type B mould (a) and resulting 30 kg/m3 PUR foam panel infused with 1wt% CNFs (b).
4.4. Tension test
Table 1 Summary of the tested samples.
30_PUR 50_PUR
Neat
1 wt.% CNFs
1 wt.% Fe3O4@CNFs
9 samples_type B 6 samples_type A
9 samples_type B 6 samples_type A
– 6 samples_type A
Tensile tests were carried out according to the ASTMD 638-89 using Type IV specimens. Specimens were loaded quasi-statically until failure using a Zwick/ Roell test machine with 2.5 kN load cell at a crosshead speed of 2 mm/ min (Fig. 4a). Three specimens for each neat and nanophased foam type were tested and the results were obtained as average values. The sample size was adapted to the specific case since the standard dimensions were too small for the PUR density tested in this study. Therefore, the following dimensions were adopted for the specimens: total length = 200 mm, initial gauge length = 80 mm, narrow section width = 20 mm, overall width = 50 mm and thickness = 30 mm. Both ends of the sample were glued to two aluminium plates so that they were firmly clamped to the testing machine (Fig. 4b). Load-displacement measures were recorded during the experiment using a data acquisition for further analyses of the tensile data. The typical failure of the samples is shown in Fig. 4c. Digital Image Correlation (DIC) was used to determine the contour and the displacement of the samples under load in three dimensions [23,24]. The tensile properties of 30 kg/m3 and 50 kg/m3 neat and nanophased PUR samples were examined and compared in tensile test.
taking SEM micrographs. The samples of neat and nanophased PUR were cut into small pieces and attached to an aluminium SEM stub. Both nanoparticles and foam samples were coated with gold by a plasma sputter to prevent charge build-up by the electron to be adsorbed by the specimens. Subsequently, the samples were moved into the chamber and micrographs were taken at an accelerating voltage of 20 kV using Philips XL20, whereas of 7 kV for Zeiss Supra 40. Morphological analysis was carried out along the perpendicular direction to the foam rise, except for samples infused with aligned carbon nanofibers where both perpendicular and longitudinal directions were taken into account.
4.3. Apparent density Apparent density test was carried out according to UNI EN 1602. Three prismatic PUR specimens for each type of foam were analyzed. They were placed in an air-conditioned chamber with stable temperature of (23 ± 2) °C and relative humidity of (50 ± 5) % until reaching a constant mass. The weight and volume of the samples were estimated with an accuracy equals to 0.5%. Average apparent densities were calculated as the ratio between average weight and volume of each type of samples.
4.5. Compression test Compression tests were carried out according the UNI EN 826 standard. Fig. 5a shows the prismatic specimens with dimension of 50 × 50 × 30 mm3 which were tested using a Zwick/Roell testing system at a crosshead speed respectively of 10 mm/min (Fig. 5b). Three specimens for each neat and nanophased foam type were tested and the results were obtained as average values. Yield point of the material could not be reached due to the high compressibility of the
Fig. 3. Magnetic alignment of functionalized CNFs under an external magnetic field (a-b).
237
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 4. Zwick/Roell test machine used to conduct tensile tests (a); dog bone shape sample equipped with aluminium wings at the both ends (b) and sample failure in correspondence of the section reduction (c).
Strain results were elaborated by Digital Image Correlation (DIC) technique since punching shear, which occurred on the compressive side of the samples, compromised the reliability of deflection values provided by the test machine. Therefore, DIC analysis was focused on the most stressed portion of the samples (Fig. 7), referring the strain calculation to the midpoint of the lower samples side. The maximum stress, which occurred at the midpoint of the outer samples surface, was calculated in accordance with the above cited standard using the following equation:
material. Therefore, in accordance with the specific standard, the compressive strength was assumed as the compressive value corresponding to a strain of 10%. Specimen displacements were recorded from the crosshead movement and the data acquisition system. Nanophased foam specimens with aligned nanofibers were loaded along the same direction of the fiber alignment within PUR matrix. Both 30 kg/m3 and 50 kg/m3 neat foams were tested. The properties of the nanophased PUR samples were also compared with those of the neat ones.
σf = 3PL / 2bd2
4.6. Flexure test
(1)
where: Flexure tests were carried out using the Zwick/Roell test machine with 5 kN load cell at a crosshead speed of 2 mm/min, according to the ASTM D 790-02 standard. Prismatic specimens of dimensions: thickness = 40 mm, width = 30 mm and span length = 173 mm were cut from the foam panel in the in-plain direction using a hot wire cutter and loaded under three-point bending (Fig. 6a-b). Only 30 kg/m3 both neat and nanophased PUR samples were examined in flexure tests. Three specimens for each neat and nanophased foam type were tested and the results were obtained as average values.
σ = stress in the outer fibers at midpoint (kPa); P = load at a given point on the load-deflection curve (N); L = support span (mm); b = width of sample tested (mm); d = depth of sample tested (mm). The tangent modulus of elasticity is the ratio, within the elastic limit, of the stress to corresponding strain. It was calculated by drawing a tangent to the steepest initial straight-line portion of the stress-strain curve.
Fig. 5. Prismatic specimens of neat and nano-phased PUR foams (a) and compressive test in progress using Zwick/Roell test machine (b).
238
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 6. Prismatic sample under load at the initial (a) and advanced (b) stage of the flexure test.
humidity of the environment influences the thermal performance of the foam. The effect of CNFs addition on the hygroscopic sorption curve of 30 kg/m3 the foams was determined on prismatic samples with dimensions of 10 × 10 × 4 cm3, according to UNI EN ISO 12571. The standard method consists of drying samples in M710 thermostatic oven at 40 °C, until reaching a constant weight. Then, specimens were placed in four growing relative humidity environments in the range of 30–95% at constant temperature (T = 20 ± 0.5 °C). Different relative humidity values were obtained by using different saturated salt solutions: NaOH for RH = 8.91 ± 2.40%, MgCl2 for RH = 33.07 ± 0.18%, NH4NO3 for RH = 54.38 ± 0.23%, NaCl for RH = 75.47 ± 0.14%, KNO3 for RH = 94.62 ± 0.66%. After achieving the equilibrium in each environment, the sample weight was measured. Moisture content of each sample was evaluated using the following expression:
4.7. Thermal conductivity The heat flow meter apparatus was employed to test thermal conductivity of a 30 kg/m3 PUR foam panel in accordance with UNI EN12667 [25,26]. A steady state one-dimensional heat flux through the test specimen was established by exploiting the combination of two parallel plates, at constant but different temperatures (respectively 30 °C and 10 °C), which are connected to thermostatically-controlled water baths. A 50 square centimeters in size PUR foam sample with a thickness of 7.2 cm (as cited in section 4, type B) was placed between two plates. The equipment assembly used for the test is schematically shown in Fig. 8. Thermocouples fixed in the plates measured the temperature across the specimen, while thermal flux meters measured the heat flow through the specimen. The guarded area ensured a one-dimensional heat transfer through the measuring area. Indeed, the whole system was insulated in order to avoid thermal radiation and temperature variations due to thermal exchange inside the test environment. The steady state thermal conductivity of the sample can be calculated as follows:
λ = s·q/ΔT
u = (m – m0) / m0
(3)
where: m represents sample weight at each relative humidity value (g); m0 represents dehydrated sample weight (g).
(2)
and reported in a diagram as a function of the relative humidity values (hygroscopic adsorption curves).
where: q = heat flux (W/m2); s = Thickness of the sample (m); ΔT = temperature difference across the specimen (K).
5. Results and discussion 5.1. Measurement of the magnetic flux density
4.8. Hygroscopic adsorption
Summary of longitudinal (BH) and radial (BR) component of the magnetic flux density, generated by a couple of neodymium (NdFeB) magnetic disks at a mutual distance of 5 cm, is shown in Table 2.
The ability of the foam to adsorb water at different relative
Fig. 7. Strain calculation at the midpoint of the lower sample side using DIC technique.
239
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 8. Laminated 30 kg/m3 PUR foam specimen used for the thermal conductivity test.
The response in terms of the variation in magnetic flux density as a result of the introduction of functionalized CNFs is shown within the brackets. Using a gauss meter, the density of the magnetic flux was measured to be approximately 60 mT at the middle distance between the magnets, whereas it turned out to be about 90 mT in the proximity of them where nanoparticles were concentrated. Therefore, the introduction of covered nanofibers led to a 10% increase in magnetic flux density in the zone near the magnetic disks. On the other hand, magnetic flux density decreased by 50% in the zones close to the magnets in correspondence of voids where nanoparticles were absent. In addition, in the zones were nanoparticles were concentrated, an increase in radial component (BR) combined with a decrease in longitudinal component (BH) of the magnetic flux density were observed. This depends on the magnetic flux conservation. The fibers thus behaved as magnets provoking a distortion in the field line.
observe individual nanofibers. The average diameter was measured to be about 50 nm, value consistent with the data sheet. As the size of nanoparticles is small, specific surface area is large and so as surface energy. Thus, the adhesive force between the nanoparticles is strong and they easily agglomerate. Fig. 10a-f shows the scanning electron micrographs of both 30 kg/ m3 and 50 kg/m3 PUR foams modified with increasingly percentages of CNFs either randomly-oriented or aligned through the external magnetic field. The structure of the neat foams of the two densities (Fig. 10 a and c) is characterized by both closed and open cells with variable dimensions, irregular polyhedral shapes and thin walls. In the case of high density foams the structure has smaller cells with thicker walls. The nano-reinforced foams showed a reduction in cell size in conjunction to higher wall thicknesses (Table 3), a more uniform distribution of the cells and a negligible quantity of cells with broken walls (Fig. 10 b and d). The functionalized CNFs played a significant role in regulating cell shape in both transversal (Fig. 10 e) and longitudinal cross sections (Fig. 10 f) as well as strengthening cell walls. The addition of fillers multiplied the nucleation sites for bubble formation, thus increasing cell number and decreasing cell size because
5.2. Microstructure SEM micrographs of carbon nanofibers and magnetic carbon nanofibers, prior to the foaming process, are shown in Fig. 9a-b. Since CNFs have length in the order of micron, it was possible to Table 2 Magnetic flux density before and after the introduction of functionalized CNFs into the box. (mT)
BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR
Magnet
A B C D F G H I L M N
1
2
3
4
5
6
7
8
9
100 (113) 38 (48) 75 (77) 34 (33) 59 (55) 27 (20) 50 (46) 17 (24) 44 (41) 13 (8) 43 (49) −3 (3) 44 (41) −13 (−8) 48 (48) −16 (−58) 55 (56) −35 (−29) 72 (72) 39 (−47) 103 (110) −43 (−62)
93 (100) 19 (23) 82 (81) 17 (20) 66 (64) 20 (15) 56 (54) 9 (9) 51 (49) 9 (4) 49 (47) 0 (0) 51 (49) −9 (−8) 56 (55) −12 (−15) 65 (63) −28 (−25) 78 (80) 28 (−40) 101 (101) −33 (−35)
90 (97) 16 (11) 83 (82) 10 (9) 71 (71) 10 (5) 61 (60) 9 (0) 57 (55) 7 (0) 54 (53) 0 (−2) 56 (55) −9 (−10) 62 (62) −10 (−16) 70 (69) −22 (−19) 82 (81) 20 (−27) 96 (90) −29 (−23)
90 (52) 8 (6) 83 (84) 7 (3) 74 (71) 3 (0) 64 (64) 4 (−3) 59 (58) 0 (−1) 58 (57) 0 (−3) 49 (59) −9 (−9) 64 (65) −10 (−14) 72 (72) −14 (−12) 83 (82) −11 (−20) 93 (89) −20 (−17)
91 (50) 3 (6) 83 (82) 3 (0) 72 (74) 1 (−3) 65 (64) −6 (−4) 60 (60) 0 (−1) 58 (59) −4 (−2) 60 (61) −9 (−2) 66 (67) −8 (−9) 75 (72) −7 (−6) 84 (82) −10 (−13) 91 (90) −9 (−7) Magnet
91 (94) −8 (4) 86 (84) −3 (−5) 74 (74) −8 (−7) 64 (64) −6 (−5) 60 (60) 0 (−6) 58 (59) −3 (0) 60 (61) −6 (−1) 66 (67) −6 (−6) 75 (72) −5 (−5) 84 (84) −4 (−6) 91 (90) −13 (−6)
96 (110) −2 (3) 86 (87) −17 (−12) 73 (74) −15 (−13) 63 (64) −10 (−8) 57 (60) −1 (−6) 56 (57) −2 (−1) 58 (59) −5 (0) 66 (66) 1 (4) 75 (74) 2 (2) 86 (86) 2 (−4) 93 (90) −5 (−4)
105 (103) −20 (−2) 86 (85) −26 (−20) 70 (70) −23 (−25) 58 (62) −10 (−14) 54 (56) −6 (−6) 91 (53) −6 (−1) 54 (56) 0 (1) 64 (61) 6 (11) 72 (73) 9 (8) 89 (88) 15 (6) 100 (94) 9 (−2)
105 (116) −44 (−14) 73 (85) −40 (−23) 69 (67) −32 (−30) 50 (56) −14 (−15) 46 (50) −8 (−8) 48 (47) −5 (−3) 48 (50) 3 (4) 54 (57) 10 (11) 69 (69) 16 (16) 90 (90) 25 (24) 114 (114) 19 (31)
240
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 9. SEM micrographs of carbon nanofibers (a) and magnetic carbon nanofibers (b) prior to the foaming process.
Fig. 10. SEM images: (a) 30_Neat PUR foam; (b) 30_1%CNFs PUR foam; (c) 50_Neat PUR foam; (d) 50_1%CNFs PUR foam; (e) 50_1%Fe3O4@CNFs PUR foam (aligned-functionalized), transversal section and (f) 50_1%Fe3O4@CNFs PUR foam (aligned-functionalized), longitudinal section.
the introduction of CNFs due to a lower cell size and an increase in cell density as observed by SEM images (see 5.2). However, in 50 kg/m3 PUR foams a slight increase in samples density was observed after the addiction of aligned CNFs.
Table 3 Microstructural results for SEM. Material
30_NEAT 30_1% CNFs 50_NEAT 50_1% CNFs 50_1% Fe3O4@CNFs
Cell face thickness (μm)
Cell size, d (μm)
Average ± S.D.
Average ± S.D.
0.33 0.51 0.59 0.68 0.83
431 160 331 260 225
± ± ± ± ±
0.03 0.10 0.02 0.15 0.03
± ± ± ± ±
50 20 40 42 48
5.4. Tensile response Tensile stress-strain curves concerning both neat and nanophased PUR foam are shown in Fig. 11. The curves represent the average value of the three samples for each type studied. Generally, it can be observed that stress-strain curves are almost linear up to about 80% of the failure load indicating that material has an elastic behavior with brittle failure. Tensile strength and Young's modulus largely increase at the increasing of foams density and with the addition of CNFs. Results also point out the key role that the alignment of CNFs along the stress direction has on the mechanical performance of the foams, in particular in terms of stiffness. Therefore, the highest tensile strength and Young's modulus was achieved by 50 kg/m3 foams with 1 wt% of aligned CNFs. By contrast, the worst behavior was shown by the 30 kg/m3 neat samples without the surface veil.
of a reduction in coalescence among the bubble [4]. Moreover, the more bubbles started to nucleate concurrently, the less amount of gas was available for bubble growth, leading to a reduction in cell size. Hence, the decrease in cells size and increase in cells density were achieved by the addition of a small amount of carbon nanoparticles uniformly dispersed within polymer matrix. 5.3. Apparent density Summary of the average apparent densities is shown in Table 4. In 30 kg/m3 PUR foams, the samples density increased by 10% with 241
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Table 4 Summary of the average apparent density of the samples. Samples
30_NEAT 30_1%CNFs 50_NEAT 50_1%CNFs 50_1%Fe3O4@CNFs
Weight (kg)
Volume (m3)
Apparent density (kg/m3)
Average ± S.D.
Average ± S.D.
Average ± S.D.
(3.599 (4.273 (2.673 (3.928 (3.620
± ± ± ± ±
0.52) 0.36) 0.22) 0.26) 0.26)
x x x x x
10−3 10−3 10−3 10−3 10−3
(1.1183 (1.2085 (0.5357 (0.7827 (0.6889
± ± ± ± ±
0.0354) 0.0381) 0.0908) 0.0369) 0.0390)
x x x x x
10−4 10−4 10−4 10−4 10−4
32.13 35.44 49.89 50.18 52.54
± ± ± ± ±
3.88 4.07 4.03 1.09 4.18
tensile strength and Young's modulus by 277% and 308%, respectively. The results were beyond expectation as compared to previous studies [8] in which the introduction of 1 wt% of CNFs in 240 kg/m3 PUR foams had led to an increase in tensile strength and Young's modulus by 35% and 86%, respectively. The more noticeable improvement in the present work could be ascribed to the adoption of foams with lower densities in which the nanoparticles addition has grater effects. The failure generally occurred in correspondence of the samples sections reduction, near the grippings (Fig. 4c). This local weakness is the result of the anisotropy of the specific material tested combined with the typical samples shape adopted for the test. Fig. 12a-b shows the longitudinal deformation of the sample which appears quite non-uniform due to the irregular morphology of the foams. The grid in the central area of the specimen gauge length was developed by Matlab in order to measure the change in gauge length until tensile failure. Therefore, engineering strain (ε) was the result of the ratio between the change in gauge length and the initial gauge length of the samples. 5.5. Compressive response Fig. 11. Tensile stress-strain plot.
Compressive stress-strain curves for the neat and nanophased PUR foams are shown in Fig. 13. The curves represent the average value of the three samples for each type studied. The general trend of the graph is characterized by two stages of deformation: an initial almost linear behavior followed by a wide plateau region which likely corresponds to the material internal instability (e.g. bucking of cells walls) developing by increasing the load until reaching the failure. As shown in previous studies [8], this tendency is more noticeable in high density PUR foams than low density ones which do not have an initial linear behavior due to their microstructure and, thus, correlated repercussions on mechanical properties. The initial slope of each curve was used to determine the foam compressive Young's modulus while the compression strength corresponds to the 10% of the relative foam deformation. Generally, it can be observed that the foam density has a lower incidence on the final performance of the samples under compression stress as compared to the previous results. Similarly, the CNFs alignment determines the highest outcomes among all the tested samples. Summary of the compression test results is shown in Table 6. The data show that compressive performance is higher for nanophased foams than for neat ones. More precisely, the addition of 1 wt% of CNFs into the 50 kg/m3 foam increased its compressive strength and Young's modulus by about 69% and 100%, respectively. The corresponding values for 50 kg/m3 nanophased foam, infused with carbon nanofibers aligned along the direction of the compressive stress, were up to 125% and 178%, respectively. These results point out that carbon nanofibers are very effective to produce significant improvements in the mechanical properties, especially along the direction
The results are consistent with the findings of Saha [8] where the neat foam had the lowest stiffness. However, in this latter study the adopted foams had a higher density (240 kg/m3) characterized by a more compact and closer morphology which allowed to reach a higher tensile strength. The Young's modulus of elasticity is calculated from the slope of the initial linear part of the stress-strain curve, while the maximum stress is taken as the tensile strength. Summary of the tensile test results, which were obtained with the aid of Digital Image Correlation, is shown in Table 5. The tensile-tested samples had demonstrated that the addition of 1 wt% CNFs into the 50 kg/m3 PUR foam increased its tensile strength and Young's modulus by 11% and 34%, respectively. Moreover, the introduction of 1 wt% aligned CNFs led to a 19% increase in tensile strength and 140% increase in tensile Young's modulus. In 30 kg/m3 PUR foam, the addition of 1 wt% CNFs increased the
Table 5 Summary of the tensile test results. Material
30_NEAT 30_1% CNFs 50_NEAT 50_1% CNFs 50_1% Fe3O4@CNFs
Tensile modulus (kPa)
Tensile strength σu (kPa)
Average ± S.D.
Average ± S.D.
693 ± 50 2832 ± 82.3 11873 ± 54.2 15894 ± 58.5 28483 ± 97.4
48 ± 3.3 137 ± 36 372 ± 40.2 413 ± 6.9 441 ± 54.6
242
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 12. Initial (a) and final (b) longitudinal deformation of the sample tested in tensile behavior.
of the alignment when compared to their randomly-oriented counterparts. In 30 kg/m3 PUR foam, the addition of 1 wt% CNFs increased its compressive strength and Young's modulus by 68% and 140%, respectively. Once more, results were beyond expectation as compared to previous studies [8] in which the introduction of 1 wt% of CNFs in 240 kg/ m3 PUR foams had led to an increase in compressive strength and Young's modulus by 57% and 40%, respectively. 5.6. Flexural response Flexure stress-strain curves of the neat and nanophased foams are shown in Fig. 14.
Fig. 13. Compressive stress-strain plot.
Table 6 Summary of the compression test results. Material
30_NEAT 30_1% CNFs 50_NEAT 50_1% CNFs 50_1% Fe3O4@CNFs
Compressive modulus (kPa)
Compressive strength σ10 (kPa)
Average ± S.D.
Average ± S.D.
932 ± 41.5 2245 ± 59 943 ± 42 1893 ± 90 2623 ± 25.5
54 ± 12 85 ± 2.43 91 ± 5.2 144 ± 25.5 192 ± 3.72 Fig. 14. Flexure stress-strain plot.
243
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
improving the gas barrier properties. Other authors demonstrated that the use of nanoparticles leads to a sensible enhancement of thermal insulating properties of the foams as well as a reduction of the thermal aging phenomena [20]. These effects were also influenced by the compatibility between polymer matrix and fillers. Therefore, high compatibility as in the present case leads to a great reduction in nucleation free energy, thus generally favoring the formation of a high number of nucleation sites and, therefore, a fine cell structure. The consequent decrease in cell size leads to an increase in the number of screens that heat flux must pass through. However, positive effects of the filler on the thermal insulating properties may be achieved only when a homogeneous dispersion is reached [5].
Table 7 Summary of the flexure test results. Material
30_NEAT 30_1% CNFs
Flexural modulus (kPa)
Flexural strength (kPa)
Average ± S.D.
Average ± S.D.
3085 ± 71.4 3657 ± 146.4
127 ± 9.81 159 ± 13.92
Table 8 Summary of the thermal conductivity test results. Material
Thermal conductivity λ (W/(mK)
30_NEAT 30_1% CNFs
0.026 0.023
5.8. Hygroscopic adsorption
In general, it can be observed that stress-strain curves are fairly nonlinear and the initial slope is higher for the nanophased foams respect to the neat ones. Summary of the flexure test results is presented in Table 7. The flexural-tested samples had further demonstrated that the addition of 1 wt% CNFs into the 30 kg/m3 PUR foam increased flexure strength and Young's modulus by 25% and 18%, respectively, as compared to neat one. These results are consistent with previous studies which showed an enhancement in flexural strength and Young's modulus of 45% and 40%, respectively [8]. Failure analyses suggested that failure started on the tensile side of nanophased specimens, whereas neat foam did not reach failure due to the elevate punching shear.
Hygroscopic adsorption curves are drawn in Fig. 15. The curves represent the average value of the three samples for each type studied. As can be observed, at low relative humidity values both neat and nanopahased PUR foams show a similar behavior. On the contrary, increasing the relative humidity degree in the test environment, the introduction of CNFs in polymer matrix reduced the moisture content in the samples, which means a better resistance against water adsorption and, thus, better waterproofing properties. The moisture content in modified foam decreased about 1% as compared to neat foam. The slightly improvement was due to the morphology of the samples doped with nanoparticles which is characterized by closer cells with smaller size that prevent the entry of water [5].
5.7. Thermal conductivity Summary of the thermal conductivity test results is shown in Table 8. It could be observed that the addition of only 1 wt% of CNFs into PUR foam decreased its thermal conductivity by about 10%. This is because the nanoparticles act as nucleating agents during foaming process producing the finer cell structure (see 5.2) and
6. Conclusions Neat and nanophased polyurethane foams of two different densities (30 and 50 kg/m3) were manufactured. CNFs were selected as
Fig. 15. Hygroscopic adsorption curves.
244
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
nanofillers and sonication technique was used to disperse them into the polymer foam precursor. Moreover, functionalized CNFs were fabricated by attaching magnetite nanoparticles to CNFs. Co-precipitation method was adopted to enhance the magnetic susceptibility of the nanoparticles so as to be aligned into the polymer matrix using an external weak magnetic field. Both neat and nanophased foams were compared in terms of microstructure, tensile, compressive as well as flexural behavior, thermal conductivity, hygroscopic adsorption and apparent density. The results demonstrated that:
[2]
[3] [4] [5]
[6]
1. in terms of microstructure, the addition of nanofillers multiply the nucleation sites for bubble formation, thus increasing the number of the cells while decreasing cell sizes. Fe3O4@CNFs also play a significant role in regulating cell shapes as well as strengthening cell walls. 2. in terms of tensile response, failure analyses of 50 kg/m3 tested samples demonstrate a 11% and 34% increase in tensile strength and Young's modulus by adding 1 wt% of CNFs. Aligned carbon nanofibers lead to a 19% and 140% enhancement in tensile strength and Young's modulus. Moreover, the introduction of 1 wt% CNFs in 30 kg/m3 tested samples increases their tensile strength and Young's modulus by 277% and 308% as compared to the neat foams. 3. in terms of compressive response, failure analyses of 50 kg/m3 tested samples demonstrate a 69% and 100% increase in compressive strength and Young's modulus by adding 1 wt% CNFs while 1 wt % Fe3O4@CNFs leads to a 125% and 178% increase in compressive strength and Young's modulus. Moreover, the introduction of 1 wt% CNFs in 30 kg/m3 tested samples increases compressive strength and Young's modulus by 68% and 140%, respectively. 4. in terms of flexure response, failure analyses of 30 kg/m3 tested samples demonstrate a 25% and 18% increase in flexure strength and Young's modulus by adding 1 wt% CNFs. 5. in terms of thermo-hygrometric response, the addition of 1 wt% CNFs into 30 kg/m3 tested panel leads to a 10% decrease in thermal conductivity. Moreover, nanoparticles reduce the moisture content of the foam which obtains a better resistance against the water adsorption.
[7] [8]
[9] [10]
[11]
[12] [13]
[14] [15]
[16]
[17]
[18]
[19]
Therefore, the introduction of CNFs is an effective method to improve the foam properties. Moreover, the alignment of functionalized CNFs within a high density PUR foam resulted to be a key factor in achieving the highest compression and tensile performance. Our future studies will be focused on PUR foams with higher densities and infused with other types of nanoparticles.
[20]
[21] [22] [23] [24]
Acknowledgements The authors wish to thank the technicians Orlando Favoni, Eng. Adriano Di Cristoforo and Alfredo De Leo for the support in laboratory tests as well as Simone Rogante (Claudio Foresi Company) for the samples manufacturing.
[25]
[26]
References [1] Dorel Feldman, Polymer nanocomposites in building, construction, J. Macromol.
245
Sci., Part A 51 (3) (2014) 203–209, http://dx.doi.org/10.1080/10601325.2014. 871948. F. Stazi, F. Tittarelli, G. Politi, C. Di Perna, P. Munafò, Assessment of the actual hygrothermal performance of glass mineral wool insulation applied 25 years ago in masonry cavity walls, Energy Build. 68 (2014) 292–304. F. Tittarelli, S.P. Shah, Effect of low dosages of waste GRP dust on fresh and hardened properties of mortars: Part 1, Construct. Build. Mater. 47 (2013) 1532–1538. L. James Lee, Changchun Zeng, Xia Cao, Xiangming Hang, Jiong Shen, Guojun Xu, Polymer nanocomposite foams, Compos. Sci. Technol. 65 (2005) 2344–2363. Caterina Lorusso, Viviana Vergaro, Annagrazia Monteduro, Antonio Saracino, Giuseppe Ciccarella, Paolo Maria Congedo, Barbara Federica Scremin, Characterization of polyurethane foam added with synthesized acetic and oleicmodified TiO2 nanocrystals, Nanomater. Nanotechnol. 5 (2015) 26, http://dx.doi. org/10.5772/61275 I. Gunjan, Maheshwari; Nilanjan, Mallik; Jandro, Abot; Albert, Song; Emily, Head; Vesselin, Shanov; Mark, Schulz; Chaminda, Jayasinghe; Pravahan, Salunke; YeoHeung, Yun; Sergey, Yarmolenko, Powering up nanoparticles: versatile carbon materials for engineering and medicine. Newsroom 10.1117/2.1200809.1305. Juan Lobos, Sachin Velankar, How much do nanoparticle fillers improve the modulus and strength of polymer foams? J. Cell. Plast. 0 (0) (2014) 1–32. M.C. Saha, MdE. Kabir, S. Jeelani, Enhancement in thermal and mechanical properties of polyurethane foam infused with nanoparticles, Mater. Sci. Eng. 479 (2008) 213–222. Xiao-Lin Xie, Yiu-Wing Mai, Xing-Ping Zhou, Dispersion and alignment of carbon nanotubes in polymer matrix: a review, Mater. Sci. Eng. R 49 (2005) 89–112. Hassan Mahfuz, Vijaya K. Rangari, Mohammad S. Islam, Shaik Jeelani, Fabrication, synthesis and mechanical characterization of nanoparticles infused polyurethane foams, Composites Part A 35 (2004) 453–460. Mohammad M. Hasan, Yuanxin Zhou, Shaik Jeelani, Thermal and tensile properties of aligned carbon nanofiber reinforced polypropylene, Mater. Lett. 61 (2007) 1134–1136. Dongzhi Yang, Jianfeng Zhang, Jing Zhang, Jun Nie. Aligned electrospun nanofibers induced by magnetic field. J. Appl. Polym. Sci. DOI 10.1002. Mohammad Chowdhury, George Stylios, Process optimization and alignment of PVA/FeCl3 nano composite fibres by electrospinning, J. Mater. Sci. 46 (2011) 3378–3386, http://dx.doi.org/10.1007/s10853-010-5226-5. Chuncheng Hao, Xiaojiao Li, Guizhen Wang, Magnetic alignment of nickel-coated carbon fibers, Mater. Res. Bull. 46 (2011) 2090–2093. Shahrzad, Rahmani; Mehdi, Rafizadeh; Framarz, Afshar, Taromi, Statistical analysis of nanofibers alignment in magnetic-field-assisted electrospinning including an alignment percentage formula. J. Appl. Polym. Sci. DOI: 10.1002/app.41179. Hassan Mahfuz, Shaik Zainuddin, Martin R. Parker, Tariq Al-Saadi, Vijay K. Rangari, Shaik Jeelani, Reinforcement of SC-5 epoxy with CNT/CNF under high magnetic field: an investigation of mechanical and thermal response, J. Mater. Sci. 44 (2009) 1113–1120. O. Malkina, H. Mahfuz, K. Sorge, A. Rondinone, J. Chen, K. More, S. Reeves, V. Rangari, Magnetic alignment of SWCNT decorated with Fe3O4 to enhance mechanical properties of SC-15 epoxy, AIP Adv. 3 (042104) (2013). Shuying Wu, Raj B. Ladani, Jin Zhang, Anthony J. Kinloch, Zhiheng Zhao, Jun Ma, Xuehua Zhang, Adrian P. Mouritz, Kamran Ghorbani, Chun H. Wang, Epoxy nanocomposites containing magnetite-carbon nanofibers aligned using a weak magnetic field, Polymer 68 (2015) 25–34. H. Mahfuz, et al., Response of sandwich composites with nanophased cores under flexural loading, Composites Part B 35 (2004) 543. Michele, Modesti; Alessandra, Lorenzetti; Stefano, Besco, Influence of nanofillers on thermal insulating properties of polyurethane nanocomposites foams. Polym. Eng. Sci. DOI 10.1002/pen.20819. MdE. Kabir, Effect of ultrasound sonication in carbon nanofibers/polyurethane foam composite, Mater. Sci. Eng. A 459 (2007) 111–116. O. Abramov, Ultrasound in Liquid and Solid Metals, CRC Press, Boca Raton, FL, 1994. D. Amodio, G.B. Broggiato, F. Campana, G.M. Newaz, Digital speckle correlation for strain measurement by image analysis, Exp. Mech. 43 (2003) 396–402. G. Palmieri, G. Chiappini, M. Sasso, S. Papalini, Hyperelastic materials characterization by planar tension tests and full-field strain measurement, Society for Experimental Mechanics - SEM Annual Conference and Exposition on Experimental and Applied Mechanics, 4 2009, pp. 2232–2238. Ulf Hammerschmidt, Jacques Hameury, Radek Strnad, Emese Turzò-Andras, Jiyu Wu, Critical review of industrial techniques for thermal-conductivity measurements of thermal insulation materials, Int. J. Thermophys. 36 (2015) 1530–1544. Cinzia Burratti, Elisa Moretti, Elisa Belloni, Fabrizio Agosti, Development of innovative aerogel based plasters: preliminary thermal and acoustic performance evaluation, Sustainability (2014), http://dx.doi.org/10.3390/su6095839.
Contents lists available at ScienceDirect
Polymer Testing journal homepage: www.elsevier.com/locate/polytest
Carbon nanofibers in polyurethane foams: Experimental evaluation of thermo-hygrometric and mechanical performance
T
F. Stazia,∗, F. Tittarellia, F. Saltarellia, G. Chiappinib, A. Morinic, G. Cerric, S. Lencid a
Dipartimento di Scienze e Ingegneria della Materia, dell’Ambiente ed Urbanistica (SIMAU), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy b Dipartimento di Ingegneria Industriale e Scienze Matematiche (DIISM), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy c Dipartimento di Ingegneria dell'informazione (DII), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy d Dipartimento di Ingegneria Civile e dell’Architettura (DICEA), Facoltà di Ingegneria, Università Politecnica delle Marche, Via Brecce Bianche, 60131 Ancona, Italy
A R T I C L E I N F O
A B S T R A C T
Keywords: Polyurethane foam Carbon nanofibers Functionalization process Magnetic field alignment Microstructure Mechanical behavior Thermo-hygrometric behavior
Polymer nanocomposites synergistically combine the good thermal properties of the hosting polymer matrix with the high mechanical performance of the fillers, providing a new class of materials with superior properties. The present study aims to evaluate in a multidisciplinary way the enhancement in mechanical and thermalhygrometric properties of low and medium density nanophased polyurethane (PUR) foams with either randomly oriented or aligned nanofibers as compared to the neat ones. To this aim, 1% weight of carbon nanofibers (CNFs) were homogeneously dispersed into polyol of PUR foam by an ultrasonic cavitation method. In parallel, a small amount of CNFs was functionalized in advance by a coprecipitation method so as to align them into the polymer matrix through an external low intensity magnetic field. SEM analyses were used to compare the microstructure of the neat and nanophased samples. Results have shown that the addition of carbon nanofibers in the foams products a closer structure with a more uniform size and shape. Moreover, functionalized CNFs play a significant role in regulating cells shape as well as strengthening cells walls. Mechanical test results also demonstrated that CNFs increase both strength and stiffness of the samples. The alignment of carbon nanofibers within medium density nanophased foams determines the highest mechanical properties. However, the more noticeable improvement in samples performance occurred in low density nanophased foams. Finally, carbon nanoparticles decrease the thermal conductivity and increase the resistance against water adsorption.
1. Introduction Nanoscience and nanotechnology consist of understanding and controlling the materials at a nanoscale which is a billion part of the meter (10−9 m). Moreover, the real significance of the nanoscale is that the materials obtain new properties at this level [1]. Polymer nanocomposites have been drawing a great deal of interest due to their high potential to achieve improvements in their performance by adding a small amount of nanoparticles in polymer matrices. In particular, polyurethane foams belong to lightweight and energy efficient materials for buildings [2], but their application is limited because of low mechanical strength, poor surface quality, low thermal
as well as dimensional stability. The introduction of fillers in the formulation of building materials, thanks to a refinement of the microstructure, leads to an increase in durability and mechanical performance [3]. At present, carbon nanofibers (CNFs), carbon nanotubes (CNTs), nanoclay and oxides such as SiO2 and TiO2 are some of the most commonly adopted fillers in the field of polymer nanocomposites. Therefore, a small amount of well-dispersed nanoparticles in polymer matrix may significantly improve a wide variety of properties without sacrificing the lightweight structure of polymer foams. Vapor-grown carbon nanofibers (VGCNFs), commonly named CNFs and used in the present study, are multiwall and highly graphitic fibers
∗
Corresponding author. E-mail addresses: [email protected] (F. Stazi), [email protected] (F. Tittarelli), [email protected] (F. Saltarelli), [email protected] (G. Chiappini), [email protected] (A. Morini), [email protected] (G. Cerri), [email protected] (S. Lenci). https://doi.org/10.1016/j.polymertesting.2018.01.028 Received 18 September 2017; Received in revised form 4 December 2017; Accepted 27 January 2018 Available online 07 February 2018 0142-9418/ © 2018 Elsevier Ltd. All rights reserved.
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
2. Materials
with diameters ranging from 70 to 200 nm and lengths up to a few hundred microns. Due to their high thermal stability, tensile strength and Young's modulus, CNFs have demonstrated their great potential in nanoscale polymer reinforcements. Indeed, nanofibers act as nucleation sites to facilitate the bubble nucleation process, leading to an enhancement in mechanical and thermal performance because of a finer cell structure of the hosting foam [4–7]. However, due to their high surface energy, nanofibers tend to agglomerate, bundle together and entangle. These drawbacks may lead to many defect sites in the composites having detrimental effects on polymer performances and limiting the efficiency of CNFs in polymer matrices. Thence, it is imperative to well disperse nanofibers into the polymer matrix when they are still in liquid phase so that interactions at the molecular level can be achieved in order to produce a material with superior thermal and mechanical properties [8]. In the literature, there are several techniques to improve the dispersion of nanoparticles in polymer matrices, such as by optimum physical blending, in situ polymerization and chemical functionalization. For polymer/CNFs composites, high power dispersion methods, such as ultrasound and high speed shearing, are considered the simplest and most convenient ones to reach a homogeneous dispersion of CNFs [9,10]. Moreover, apart from achieving a uniform dispersion, the alignment of carbon nanofibers along a predetermined direction has been found to produce outstanding improvements in mechanical and thermal performance when compared to their randomly-oriented counterparts [11]. Several techniques with the aim of aligning nanoparticles have been reported in literature [12–15]. Among them the application of an external magnetic field is recognized as one of the most effective and simple methods. However, due to the low magnetic susceptibility of nanofibers, an extremely strong magnetic field (ranging from 25 to 30 T) would be required to achieve a satisfying alignment [16]. Thus, the need of employing such high magnetic field limits the practical application of this method. Therefore, it is crucial to functionalize carbon nanofibers in advance by coating them with magnetic materials to align them in a polymer matrix under a relatively weak magnetic field, ranging from 50 mT to 1 T [17,18]. In summary, to the best of our knowledge, the majority of research available in the present literature regards the experimental evaluation of nanocomposites mechanical performance [4,7,10,18,19]. Less common are multidisciplinary studies which synergistically combine both thermal and mechanical aspects of polymer nanocomposites [5,8,11,16]. Moreover, polymers are generally tested in terms of decomposition temperature [20] without addressing the thermal conductivity issue. Therefore, the present work is intended to quantify the enhancement in mechanical and thermal-hygrometric properties of nanophased foams with low and medium densities, containing both randomly oriented and aligned carbon nanofibers, as compared to neat ones. To this aim, a small amount of CNFs has been functionalized in advance by a co-precipitation method in accordance with a previous study carried out by Shuying Wu and co-workers [18]. 1 wt% of nanofibers were dispersed homogeneously into polyol of PUR foam by an ultrasonic cavitation method. Subsequently, nanophased foam was obtained by adding diisocyanate to the mixture and pouring it in a mould whilst functionalized CNFs were aligned through an external magnetic field of about 90 mT (value measured in the proximity of the magnetic disks). Finally, the properties of the samples were compared in terms of microstructure, mechanical performance, thermal conductivity, hygroscopic adsorption and apparent density.
2.1. Polyurethane foams Polyurethane foams of two different apparent densities were used. The foam consists of two liquid precursors: part-A is a diphenylmethane diisocyanate polymer and part-B is a polyol. It was supplied by Claudio Foresi s.r.l. and has a density of 30 kg/m3. In addition, a higher density PUR foam was obtained from the same precursors by varying the foam production process through the adoption of a closing cap on the top of the mould, thus restraining the foam rise. 2.2. Carbon nanofibers (CNFs) According to material data supplied by Tech-Star s.r.l., the vapor grown carbon nanofibers (VGCNFs) adopted have an average diameter in the range of 30–50 nm and a length of 10–30 μm. 2.3. Magnetic carbon nanofibers (Fe3O4@CNFs) One of the main aims of this study was to demonstrate the improvement in thermal and mechanical properties of PUR foam by the introduction of CNFs aligned by a weak magnetic field. However, since the CNFs have low magnetic susceptibility, their surface requires to be covered in advance with materials characterized by strong magnetic properties such as iron oxide. Therefore, the present section describes the method adopted to functionalize CNFs and the preliminary analysis to evaluate the magnetic properties of Fe3O4@CNFs. The fabrication methods of magnetic CNFs-embedded composites have been extensively investigated in literature. A detailed review could be found in Ref. [9]. For the present study, a simple co-precipitation method, developed by Shuying Wu et al. [18], was adopted to functionalize carbon nanoparticles and align CNFs through a relatively low magnetic field. Indeed, thanks to the functionalization process, fillers may show excellent magnetic properties. The original CNFs were first subjected to an oxidation treatment in the nitric acid to modify the surface in order to achieve a better dispersion. Subsequently, 2 g of as-received CNFs were mixed with 200 mL of concentrated nitric acid whilst stirring vigorously. This mixture was then treated at 100 °C for 6 h under magnetic stirring. Afterwards, the mixture was washed several times using deionized water until reaching a pH value of ∼7. The samples were vacuum filtrated and dried in a vacuum oven. After this acid treatment, the CNFs are expected to possess oxygencontaining functional groups on their surfaces and are denoted by CNFs-OX. The functionalized nanoparticles were fabricated by a co-precipitation method from the CNFs-OX material, prepared as described above. Firstly, 0.40 g of the CNFs-OX were dispersed in 355 mL distilled water by sonication for 15 min and 0.40 g of Fe3O4 was added whilst stirring. Subsequently, the mixture was vigorously stirred for 15 min whilst being heated to 50 °C under a nitrogen (N2) atmosphere. Then 0.32 g of FeSO4⋅7H2O were added with continuous stirring under a N2 atmosphere for 30 min. Next, 27 mL of 8M NH4OH aqueous solution were added drop-wise to precipitate ferric and ferrous salt. The pH of the mixture was kept at ∼10 and the reaction was carried out at 50 °C for 30 min under vigorous magnetic stirring. N2 was continuously purged during the reaction to prevent oxidation.
235
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
longitudinal axis, coincident with the foam rise direction, through a pair of magnetic disks, using the above mentioned methods. Magnets were thus placed at a mutual distance of 20 cm, respectively at the top and at the bottom of the mould. The alignment of the Fe3O4@CNFs was imposed during the foaming process, prior to the hardening of the PUR, using a magnetic field of 90 mT in the proximity of the magnetic disks. Table 1 shows the overall number of specimens (type A and type B), containing various percentages of nanoparticles, which were manufactured in order to carry out mechanical tests. All specimens were stored under ambient laboratory conditions (temperature of 23 ± 2 °C, relative humidity of 50 ± 5%) for 30 days, as recommended by manufactures. Three samples for each type of foams were considered to determine the average value of mechanical strength and modulus. Further samples were also realized to conduct different analysis.
The Fe3O4@CNFs were obtained by magnetic separation, washed with distilled water and ethanol, and finally dried under vacuum at 50 °C.
3. Manufacturing The fabrication of nanophased polyurethane foams was conducted in three steps. In step 1, a pre-calculated amount of nanoparticles was mixed with polyol of PUR foam by a mechanical stirrer in order to have a specific weight ratio. Polyol was selected for infusion of nanoparticles since it is the least reactive part [8]. In this study, 1 wt% of CNFs and 1 wt% of Fe3O4@CNFs were chosen to conduct the experimental tests. Indeed, previous studies have found that the improvement in polymer foam properties may be achieved by infusion of a small percentage of nanoparticles, especially in the range of 0.5–2 wt% [19–21]. Ultrasonic cavitation method was chosen to disperse nanoparticles into polymer since this method was found to be the most effective in producing a homogeneous dispersion of nanoparticles [22]. It involves the formation, growth, pulsating and collapsing of tiny bubbles, producing transient micro hot spots which enable to enhance the wettability of nanoparticles. The ultrasonic wave can break the agglomerating bodies by damaging the Coulomb and Van der Waals's forces between the particles and make them homogeneously dispersed in the liquid [8]. The mixture was thus subjected to ultrasonic cavitation using a high intensity ultrasonic horn (Bandelin Sonoplus) at the amplitude of 40% and for 10 min, as shown in Fig. 1a. In step 2, diisocyanate was added to the mixture at the ratio of 48–52 wt% using a mechanical blender for about 10 s. In step 3, the mixture was poured in a wooden mould and kept inside for 15 min. In particular, in order to obtain specimens with different densities and sizes, two types of moulds were adopted: type A with the size of 20 × 5 × 3 cm3 and type B with the size of 50 × 50 × 7.2 cm3. Type A moulds were closed (Fig. 1b) by a contrasting element on the upper side and compressed by a concrete cube. This technique let us produce foams whose nominal density was 50 kg/m3. Moreover, type A samples were studied to maintain the external surface veil for mechanical tests to reproduce the real conditions of this application in buildings (Fig. 1c-d). Type B moulds were open on the upper side as shown in Fig. 2a-b. Type B samples for each foam typology were cut into the suitable shape, then the external surface veil was removed to conduct the tests. The nominal density of the foams in thus type of mould was 30 kg/m3. Neat and nanophased PUR foams were fabricated for comparison. A stoichiometric amount of Fe3O4@CNFs was only added to high density PUR foams and nanoparticles were aligned along the sample
4. Experimental methods 4.1. Measurement of the magnetic field intensity The magnetic field intensity was measured using the 5170 Gauss Tesla Meter probe, equipped with a hold sensor, which allowed recording both longitudinal (BH) and radial (BR) components of the magnetic flux density. Two neodymium (NdFeB) magnetic disks, covered in Nichel, with a diameter of 60 mm and thickness of 5 mm were used to conduct the tests. They were placed on the edges of a rectangular box at a mutual distance of 5 cm. Preliminary probe calibration was necessary to cut out external interferences. Measurements were carried out using a grid of dots drawn on a graph paper in order to make the method repeatable. Firstly, the magnetic flux density was estimated in the absence of nanoparticles. After that, Fe3O4@CNFs were poured into the box as shown in Fig. 3a-b, then we observed their effect on the variation of the magnetic field intensity. The figure clearly shows that all the nanoparticles have assumed magnetic susceptibility since successfully attracted by magnetic poles. 4.2. Microstructure The overall morphology of neat and nanophased foams was studied using Philips XL20 scanning electron microscope (SEM). Higher magnified analyses upon average cell size and cell wall thickness were conducted with Zeiss Supra 40, which also provided details of single nanoparticles such as average diameter and length. Small amounts of nanoparticles were placed on one side of a doublesided carbon tape (opposite side is attached to the sample holder) for
Fig. 1. Ultrasonic cavitation method (a); type A mould between a couple of magnetic dicks (b); type A 50_neat (c) and 1% Fe3O4@CNFs PUR foam (d), covered with the external surface veil.
236
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 2. Type B mould (a) and resulting 30 kg/m3 PUR foam panel infused with 1wt% CNFs (b).
4.4. Tension test
Table 1 Summary of the tested samples.
30_PUR 50_PUR
Neat
1 wt.% CNFs
1 wt.% Fe3O4@CNFs
9 samples_type B 6 samples_type A
9 samples_type B 6 samples_type A
– 6 samples_type A
Tensile tests were carried out according to the ASTMD 638-89 using Type IV specimens. Specimens were loaded quasi-statically until failure using a Zwick/ Roell test machine with 2.5 kN load cell at a crosshead speed of 2 mm/ min (Fig. 4a). Three specimens for each neat and nanophased foam type were tested and the results were obtained as average values. The sample size was adapted to the specific case since the standard dimensions were too small for the PUR density tested in this study. Therefore, the following dimensions were adopted for the specimens: total length = 200 mm, initial gauge length = 80 mm, narrow section width = 20 mm, overall width = 50 mm and thickness = 30 mm. Both ends of the sample were glued to two aluminium plates so that they were firmly clamped to the testing machine (Fig. 4b). Load-displacement measures were recorded during the experiment using a data acquisition for further analyses of the tensile data. The typical failure of the samples is shown in Fig. 4c. Digital Image Correlation (DIC) was used to determine the contour and the displacement of the samples under load in three dimensions [23,24]. The tensile properties of 30 kg/m3 and 50 kg/m3 neat and nanophased PUR samples were examined and compared in tensile test.
taking SEM micrographs. The samples of neat and nanophased PUR were cut into small pieces and attached to an aluminium SEM stub. Both nanoparticles and foam samples were coated with gold by a plasma sputter to prevent charge build-up by the electron to be adsorbed by the specimens. Subsequently, the samples were moved into the chamber and micrographs were taken at an accelerating voltage of 20 kV using Philips XL20, whereas of 7 kV for Zeiss Supra 40. Morphological analysis was carried out along the perpendicular direction to the foam rise, except for samples infused with aligned carbon nanofibers where both perpendicular and longitudinal directions were taken into account.
4.3. Apparent density Apparent density test was carried out according to UNI EN 1602. Three prismatic PUR specimens for each type of foam were analyzed. They were placed in an air-conditioned chamber with stable temperature of (23 ± 2) °C and relative humidity of (50 ± 5) % until reaching a constant mass. The weight and volume of the samples were estimated with an accuracy equals to 0.5%. Average apparent densities were calculated as the ratio between average weight and volume of each type of samples.
4.5. Compression test Compression tests were carried out according the UNI EN 826 standard. Fig. 5a shows the prismatic specimens with dimension of 50 × 50 × 30 mm3 which were tested using a Zwick/Roell testing system at a crosshead speed respectively of 10 mm/min (Fig. 5b). Three specimens for each neat and nanophased foam type were tested and the results were obtained as average values. Yield point of the material could not be reached due to the high compressibility of the
Fig. 3. Magnetic alignment of functionalized CNFs under an external magnetic field (a-b).
237
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 4. Zwick/Roell test machine used to conduct tensile tests (a); dog bone shape sample equipped with aluminium wings at the both ends (b) and sample failure in correspondence of the section reduction (c).
Strain results were elaborated by Digital Image Correlation (DIC) technique since punching shear, which occurred on the compressive side of the samples, compromised the reliability of deflection values provided by the test machine. Therefore, DIC analysis was focused on the most stressed portion of the samples (Fig. 7), referring the strain calculation to the midpoint of the lower samples side. The maximum stress, which occurred at the midpoint of the outer samples surface, was calculated in accordance with the above cited standard using the following equation:
material. Therefore, in accordance with the specific standard, the compressive strength was assumed as the compressive value corresponding to a strain of 10%. Specimen displacements were recorded from the crosshead movement and the data acquisition system. Nanophased foam specimens with aligned nanofibers were loaded along the same direction of the fiber alignment within PUR matrix. Both 30 kg/m3 and 50 kg/m3 neat foams were tested. The properties of the nanophased PUR samples were also compared with those of the neat ones.
σf = 3PL / 2bd2
4.6. Flexure test
(1)
where: Flexure tests were carried out using the Zwick/Roell test machine with 5 kN load cell at a crosshead speed of 2 mm/min, according to the ASTM D 790-02 standard. Prismatic specimens of dimensions: thickness = 40 mm, width = 30 mm and span length = 173 mm were cut from the foam panel in the in-plain direction using a hot wire cutter and loaded under three-point bending (Fig. 6a-b). Only 30 kg/m3 both neat and nanophased PUR samples were examined in flexure tests. Three specimens for each neat and nanophased foam type were tested and the results were obtained as average values.
σ = stress in the outer fibers at midpoint (kPa); P = load at a given point on the load-deflection curve (N); L = support span (mm); b = width of sample tested (mm); d = depth of sample tested (mm). The tangent modulus of elasticity is the ratio, within the elastic limit, of the stress to corresponding strain. It was calculated by drawing a tangent to the steepest initial straight-line portion of the stress-strain curve.
Fig. 5. Prismatic specimens of neat and nano-phased PUR foams (a) and compressive test in progress using Zwick/Roell test machine (b).
238
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 6. Prismatic sample under load at the initial (a) and advanced (b) stage of the flexure test.
humidity of the environment influences the thermal performance of the foam. The effect of CNFs addition on the hygroscopic sorption curve of 30 kg/m3 the foams was determined on prismatic samples with dimensions of 10 × 10 × 4 cm3, according to UNI EN ISO 12571. The standard method consists of drying samples in M710 thermostatic oven at 40 °C, until reaching a constant weight. Then, specimens were placed in four growing relative humidity environments in the range of 30–95% at constant temperature (T = 20 ± 0.5 °C). Different relative humidity values were obtained by using different saturated salt solutions: NaOH for RH = 8.91 ± 2.40%, MgCl2 for RH = 33.07 ± 0.18%, NH4NO3 for RH = 54.38 ± 0.23%, NaCl for RH = 75.47 ± 0.14%, KNO3 for RH = 94.62 ± 0.66%. After achieving the equilibrium in each environment, the sample weight was measured. Moisture content of each sample was evaluated using the following expression:
4.7. Thermal conductivity The heat flow meter apparatus was employed to test thermal conductivity of a 30 kg/m3 PUR foam panel in accordance with UNI EN12667 [25,26]. A steady state one-dimensional heat flux through the test specimen was established by exploiting the combination of two parallel plates, at constant but different temperatures (respectively 30 °C and 10 °C), which are connected to thermostatically-controlled water baths. A 50 square centimeters in size PUR foam sample with a thickness of 7.2 cm (as cited in section 4, type B) was placed between two plates. The equipment assembly used for the test is schematically shown in Fig. 8. Thermocouples fixed in the plates measured the temperature across the specimen, while thermal flux meters measured the heat flow through the specimen. The guarded area ensured a one-dimensional heat transfer through the measuring area. Indeed, the whole system was insulated in order to avoid thermal radiation and temperature variations due to thermal exchange inside the test environment. The steady state thermal conductivity of the sample can be calculated as follows:
λ = s·q/ΔT
u = (m – m0) / m0
(3)
where: m represents sample weight at each relative humidity value (g); m0 represents dehydrated sample weight (g).
(2)
and reported in a diagram as a function of the relative humidity values (hygroscopic adsorption curves).
where: q = heat flux (W/m2); s = Thickness of the sample (m); ΔT = temperature difference across the specimen (K).
5. Results and discussion 5.1. Measurement of the magnetic flux density
4.8. Hygroscopic adsorption
Summary of longitudinal (BH) and radial (BR) component of the magnetic flux density, generated by a couple of neodymium (NdFeB) magnetic disks at a mutual distance of 5 cm, is shown in Table 2.
The ability of the foam to adsorb water at different relative
Fig. 7. Strain calculation at the midpoint of the lower sample side using DIC technique.
239
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 8. Laminated 30 kg/m3 PUR foam specimen used for the thermal conductivity test.
The response in terms of the variation in magnetic flux density as a result of the introduction of functionalized CNFs is shown within the brackets. Using a gauss meter, the density of the magnetic flux was measured to be approximately 60 mT at the middle distance between the magnets, whereas it turned out to be about 90 mT in the proximity of them where nanoparticles were concentrated. Therefore, the introduction of covered nanofibers led to a 10% increase in magnetic flux density in the zone near the magnetic disks. On the other hand, magnetic flux density decreased by 50% in the zones close to the magnets in correspondence of voids where nanoparticles were absent. In addition, in the zones were nanoparticles were concentrated, an increase in radial component (BR) combined with a decrease in longitudinal component (BH) of the magnetic flux density were observed. This depends on the magnetic flux conservation. The fibers thus behaved as magnets provoking a distortion in the field line.
observe individual nanofibers. The average diameter was measured to be about 50 nm, value consistent with the data sheet. As the size of nanoparticles is small, specific surface area is large and so as surface energy. Thus, the adhesive force between the nanoparticles is strong and they easily agglomerate. Fig. 10a-f shows the scanning electron micrographs of both 30 kg/ m3 and 50 kg/m3 PUR foams modified with increasingly percentages of CNFs either randomly-oriented or aligned through the external magnetic field. The structure of the neat foams of the two densities (Fig. 10 a and c) is characterized by both closed and open cells with variable dimensions, irregular polyhedral shapes and thin walls. In the case of high density foams the structure has smaller cells with thicker walls. The nano-reinforced foams showed a reduction in cell size in conjunction to higher wall thicknesses (Table 3), a more uniform distribution of the cells and a negligible quantity of cells with broken walls (Fig. 10 b and d). The functionalized CNFs played a significant role in regulating cell shape in both transversal (Fig. 10 e) and longitudinal cross sections (Fig. 10 f) as well as strengthening cell walls. The addition of fillers multiplied the nucleation sites for bubble formation, thus increasing cell number and decreasing cell size because
5.2. Microstructure SEM micrographs of carbon nanofibers and magnetic carbon nanofibers, prior to the foaming process, are shown in Fig. 9a-b. Since CNFs have length in the order of micron, it was possible to Table 2 Magnetic flux density before and after the introduction of functionalized CNFs into the box. (mT)
BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR BH BR
Magnet
A B C D F G H I L M N
1
2
3
4
5
6
7
8
9
100 (113) 38 (48) 75 (77) 34 (33) 59 (55) 27 (20) 50 (46) 17 (24) 44 (41) 13 (8) 43 (49) −3 (3) 44 (41) −13 (−8) 48 (48) −16 (−58) 55 (56) −35 (−29) 72 (72) 39 (−47) 103 (110) −43 (−62)
93 (100) 19 (23) 82 (81) 17 (20) 66 (64) 20 (15) 56 (54) 9 (9) 51 (49) 9 (4) 49 (47) 0 (0) 51 (49) −9 (−8) 56 (55) −12 (−15) 65 (63) −28 (−25) 78 (80) 28 (−40) 101 (101) −33 (−35)
90 (97) 16 (11) 83 (82) 10 (9) 71 (71) 10 (5) 61 (60) 9 (0) 57 (55) 7 (0) 54 (53) 0 (−2) 56 (55) −9 (−10) 62 (62) −10 (−16) 70 (69) −22 (−19) 82 (81) 20 (−27) 96 (90) −29 (−23)
90 (52) 8 (6) 83 (84) 7 (3) 74 (71) 3 (0) 64 (64) 4 (−3) 59 (58) 0 (−1) 58 (57) 0 (−3) 49 (59) −9 (−9) 64 (65) −10 (−14) 72 (72) −14 (−12) 83 (82) −11 (−20) 93 (89) −20 (−17)
91 (50) 3 (6) 83 (82) 3 (0) 72 (74) 1 (−3) 65 (64) −6 (−4) 60 (60) 0 (−1) 58 (59) −4 (−2) 60 (61) −9 (−2) 66 (67) −8 (−9) 75 (72) −7 (−6) 84 (82) −10 (−13) 91 (90) −9 (−7) Magnet
91 (94) −8 (4) 86 (84) −3 (−5) 74 (74) −8 (−7) 64 (64) −6 (−5) 60 (60) 0 (−6) 58 (59) −3 (0) 60 (61) −6 (−1) 66 (67) −6 (−6) 75 (72) −5 (−5) 84 (84) −4 (−6) 91 (90) −13 (−6)
96 (110) −2 (3) 86 (87) −17 (−12) 73 (74) −15 (−13) 63 (64) −10 (−8) 57 (60) −1 (−6) 56 (57) −2 (−1) 58 (59) −5 (0) 66 (66) 1 (4) 75 (74) 2 (2) 86 (86) 2 (−4) 93 (90) −5 (−4)
105 (103) −20 (−2) 86 (85) −26 (−20) 70 (70) −23 (−25) 58 (62) −10 (−14) 54 (56) −6 (−6) 91 (53) −6 (−1) 54 (56) 0 (1) 64 (61) 6 (11) 72 (73) 9 (8) 89 (88) 15 (6) 100 (94) 9 (−2)
105 (116) −44 (−14) 73 (85) −40 (−23) 69 (67) −32 (−30) 50 (56) −14 (−15) 46 (50) −8 (−8) 48 (47) −5 (−3) 48 (50) 3 (4) 54 (57) 10 (11) 69 (69) 16 (16) 90 (90) 25 (24) 114 (114) 19 (31)
240
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 9. SEM micrographs of carbon nanofibers (a) and magnetic carbon nanofibers (b) prior to the foaming process.
Fig. 10. SEM images: (a) 30_Neat PUR foam; (b) 30_1%CNFs PUR foam; (c) 50_Neat PUR foam; (d) 50_1%CNFs PUR foam; (e) 50_1%Fe3O4@CNFs PUR foam (aligned-functionalized), transversal section and (f) 50_1%Fe3O4@CNFs PUR foam (aligned-functionalized), longitudinal section.
the introduction of CNFs due to a lower cell size and an increase in cell density as observed by SEM images (see 5.2). However, in 50 kg/m3 PUR foams a slight increase in samples density was observed after the addiction of aligned CNFs.
Table 3 Microstructural results for SEM. Material
30_NEAT 30_1% CNFs 50_NEAT 50_1% CNFs 50_1% Fe3O4@CNFs
Cell face thickness (μm)
Cell size, d (μm)
Average ± S.D.
Average ± S.D.
0.33 0.51 0.59 0.68 0.83
431 160 331 260 225
± ± ± ± ±
0.03 0.10 0.02 0.15 0.03
± ± ± ± ±
50 20 40 42 48
5.4. Tensile response Tensile stress-strain curves concerning both neat and nanophased PUR foam are shown in Fig. 11. The curves represent the average value of the three samples for each type studied. Generally, it can be observed that stress-strain curves are almost linear up to about 80% of the failure load indicating that material has an elastic behavior with brittle failure. Tensile strength and Young's modulus largely increase at the increasing of foams density and with the addition of CNFs. Results also point out the key role that the alignment of CNFs along the stress direction has on the mechanical performance of the foams, in particular in terms of stiffness. Therefore, the highest tensile strength and Young's modulus was achieved by 50 kg/m3 foams with 1 wt% of aligned CNFs. By contrast, the worst behavior was shown by the 30 kg/m3 neat samples without the surface veil.
of a reduction in coalescence among the bubble [4]. Moreover, the more bubbles started to nucleate concurrently, the less amount of gas was available for bubble growth, leading to a reduction in cell size. Hence, the decrease in cells size and increase in cells density were achieved by the addition of a small amount of carbon nanoparticles uniformly dispersed within polymer matrix. 5.3. Apparent density Summary of the average apparent densities is shown in Table 4. In 30 kg/m3 PUR foams, the samples density increased by 10% with 241
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Table 4 Summary of the average apparent density of the samples. Samples
30_NEAT 30_1%CNFs 50_NEAT 50_1%CNFs 50_1%Fe3O4@CNFs
Weight (kg)
Volume (m3)
Apparent density (kg/m3)
Average ± S.D.
Average ± S.D.
Average ± S.D.
(3.599 (4.273 (2.673 (3.928 (3.620
± ± ± ± ±
0.52) 0.36) 0.22) 0.26) 0.26)
x x x x x
10−3 10−3 10−3 10−3 10−3
(1.1183 (1.2085 (0.5357 (0.7827 (0.6889
± ± ± ± ±
0.0354) 0.0381) 0.0908) 0.0369) 0.0390)
x x x x x
10−4 10−4 10−4 10−4 10−4
32.13 35.44 49.89 50.18 52.54
± ± ± ± ±
3.88 4.07 4.03 1.09 4.18
tensile strength and Young's modulus by 277% and 308%, respectively. The results were beyond expectation as compared to previous studies [8] in which the introduction of 1 wt% of CNFs in 240 kg/m3 PUR foams had led to an increase in tensile strength and Young's modulus by 35% and 86%, respectively. The more noticeable improvement in the present work could be ascribed to the adoption of foams with lower densities in which the nanoparticles addition has grater effects. The failure generally occurred in correspondence of the samples sections reduction, near the grippings (Fig. 4c). This local weakness is the result of the anisotropy of the specific material tested combined with the typical samples shape adopted for the test. Fig. 12a-b shows the longitudinal deformation of the sample which appears quite non-uniform due to the irregular morphology of the foams. The grid in the central area of the specimen gauge length was developed by Matlab in order to measure the change in gauge length until tensile failure. Therefore, engineering strain (ε) was the result of the ratio between the change in gauge length and the initial gauge length of the samples. 5.5. Compressive response Fig. 11. Tensile stress-strain plot.
Compressive stress-strain curves for the neat and nanophased PUR foams are shown in Fig. 13. The curves represent the average value of the three samples for each type studied. The general trend of the graph is characterized by two stages of deformation: an initial almost linear behavior followed by a wide plateau region which likely corresponds to the material internal instability (e.g. bucking of cells walls) developing by increasing the load until reaching the failure. As shown in previous studies [8], this tendency is more noticeable in high density PUR foams than low density ones which do not have an initial linear behavior due to their microstructure and, thus, correlated repercussions on mechanical properties. The initial slope of each curve was used to determine the foam compressive Young's modulus while the compression strength corresponds to the 10% of the relative foam deformation. Generally, it can be observed that the foam density has a lower incidence on the final performance of the samples under compression stress as compared to the previous results. Similarly, the CNFs alignment determines the highest outcomes among all the tested samples. Summary of the compression test results is shown in Table 6. The data show that compressive performance is higher for nanophased foams than for neat ones. More precisely, the addition of 1 wt% of CNFs into the 50 kg/m3 foam increased its compressive strength and Young's modulus by about 69% and 100%, respectively. The corresponding values for 50 kg/m3 nanophased foam, infused with carbon nanofibers aligned along the direction of the compressive stress, were up to 125% and 178%, respectively. These results point out that carbon nanofibers are very effective to produce significant improvements in the mechanical properties, especially along the direction
The results are consistent with the findings of Saha [8] where the neat foam had the lowest stiffness. However, in this latter study the adopted foams had a higher density (240 kg/m3) characterized by a more compact and closer morphology which allowed to reach a higher tensile strength. The Young's modulus of elasticity is calculated from the slope of the initial linear part of the stress-strain curve, while the maximum stress is taken as the tensile strength. Summary of the tensile test results, which were obtained with the aid of Digital Image Correlation, is shown in Table 5. The tensile-tested samples had demonstrated that the addition of 1 wt% CNFs into the 50 kg/m3 PUR foam increased its tensile strength and Young's modulus by 11% and 34%, respectively. Moreover, the introduction of 1 wt% aligned CNFs led to a 19% increase in tensile strength and 140% increase in tensile Young's modulus. In 30 kg/m3 PUR foam, the addition of 1 wt% CNFs increased the
Table 5 Summary of the tensile test results. Material
30_NEAT 30_1% CNFs 50_NEAT 50_1% CNFs 50_1% Fe3O4@CNFs
Tensile modulus (kPa)
Tensile strength σu (kPa)
Average ± S.D.
Average ± S.D.
693 ± 50 2832 ± 82.3 11873 ± 54.2 15894 ± 58.5 28483 ± 97.4
48 ± 3.3 137 ± 36 372 ± 40.2 413 ± 6.9 441 ± 54.6
242
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
Fig. 12. Initial (a) and final (b) longitudinal deformation of the sample tested in tensile behavior.
of the alignment when compared to their randomly-oriented counterparts. In 30 kg/m3 PUR foam, the addition of 1 wt% CNFs increased its compressive strength and Young's modulus by 68% and 140%, respectively. Once more, results were beyond expectation as compared to previous studies [8] in which the introduction of 1 wt% of CNFs in 240 kg/ m3 PUR foams had led to an increase in compressive strength and Young's modulus by 57% and 40%, respectively. 5.6. Flexural response Flexure stress-strain curves of the neat and nanophased foams are shown in Fig. 14.
Fig. 13. Compressive stress-strain plot.
Table 6 Summary of the compression test results. Material
30_NEAT 30_1% CNFs 50_NEAT 50_1% CNFs 50_1% Fe3O4@CNFs
Compressive modulus (kPa)
Compressive strength σ10 (kPa)
Average ± S.D.
Average ± S.D.
932 ± 41.5 2245 ± 59 943 ± 42 1893 ± 90 2623 ± 25.5
54 ± 12 85 ± 2.43 91 ± 5.2 144 ± 25.5 192 ± 3.72 Fig. 14. Flexure stress-strain plot.
243
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
improving the gas barrier properties. Other authors demonstrated that the use of nanoparticles leads to a sensible enhancement of thermal insulating properties of the foams as well as a reduction of the thermal aging phenomena [20]. These effects were also influenced by the compatibility between polymer matrix and fillers. Therefore, high compatibility as in the present case leads to a great reduction in nucleation free energy, thus generally favoring the formation of a high number of nucleation sites and, therefore, a fine cell structure. The consequent decrease in cell size leads to an increase in the number of screens that heat flux must pass through. However, positive effects of the filler on the thermal insulating properties may be achieved only when a homogeneous dispersion is reached [5].
Table 7 Summary of the flexure test results. Material
30_NEAT 30_1% CNFs
Flexural modulus (kPa)
Flexural strength (kPa)
Average ± S.D.
Average ± S.D.
3085 ± 71.4 3657 ± 146.4
127 ± 9.81 159 ± 13.92
Table 8 Summary of the thermal conductivity test results. Material
Thermal conductivity λ (W/(mK)
30_NEAT 30_1% CNFs
0.026 0.023
5.8. Hygroscopic adsorption
In general, it can be observed that stress-strain curves are fairly nonlinear and the initial slope is higher for the nanophased foams respect to the neat ones. Summary of the flexure test results is presented in Table 7. The flexural-tested samples had further demonstrated that the addition of 1 wt% CNFs into the 30 kg/m3 PUR foam increased flexure strength and Young's modulus by 25% and 18%, respectively, as compared to neat one. These results are consistent with previous studies which showed an enhancement in flexural strength and Young's modulus of 45% and 40%, respectively [8]. Failure analyses suggested that failure started on the tensile side of nanophased specimens, whereas neat foam did not reach failure due to the elevate punching shear.
Hygroscopic adsorption curves are drawn in Fig. 15. The curves represent the average value of the three samples for each type studied. As can be observed, at low relative humidity values both neat and nanopahased PUR foams show a similar behavior. On the contrary, increasing the relative humidity degree in the test environment, the introduction of CNFs in polymer matrix reduced the moisture content in the samples, which means a better resistance against water adsorption and, thus, better waterproofing properties. The moisture content in modified foam decreased about 1% as compared to neat foam. The slightly improvement was due to the morphology of the samples doped with nanoparticles which is characterized by closer cells with smaller size that prevent the entry of water [5].
5.7. Thermal conductivity Summary of the thermal conductivity test results is shown in Table 8. It could be observed that the addition of only 1 wt% of CNFs into PUR foam decreased its thermal conductivity by about 10%. This is because the nanoparticles act as nucleating agents during foaming process producing the finer cell structure (see 5.2) and
6. Conclusions Neat and nanophased polyurethane foams of two different densities (30 and 50 kg/m3) were manufactured. CNFs were selected as
Fig. 15. Hygroscopic adsorption curves.
244
Polymer Testing 67 (2018) 234–245
F. Stazi et al.
nanofillers and sonication technique was used to disperse them into the polymer foam precursor. Moreover, functionalized CNFs were fabricated by attaching magnetite nanoparticles to CNFs. Co-precipitation method was adopted to enhance the magnetic susceptibility of the nanoparticles so as to be aligned into the polymer matrix using an external weak magnetic field. Both neat and nanophased foams were compared in terms of microstructure, tensile, compressive as well as flexural behavior, thermal conductivity, hygroscopic adsorption and apparent density. The results demonstrated that:
[2]
[3] [4] [5]
[6]
1. in terms of microstructure, the addition of nanofillers multiply the nucleation sites for bubble formation, thus increasing the number of the cells while decreasing cell sizes. Fe3O4@CNFs also play a significant role in regulating cell shapes as well as strengthening cell walls. 2. in terms of tensile response, failure analyses of 50 kg/m3 tested samples demonstrate a 11% and 34% increase in tensile strength and Young's modulus by adding 1 wt% of CNFs. Aligned carbon nanofibers lead to a 19% and 140% enhancement in tensile strength and Young's modulus. Moreover, the introduction of 1 wt% CNFs in 30 kg/m3 tested samples increases their tensile strength and Young's modulus by 277% and 308% as compared to the neat foams. 3. in terms of compressive response, failure analyses of 50 kg/m3 tested samples demonstrate a 69% and 100% increase in compressive strength and Young's modulus by adding 1 wt% CNFs while 1 wt % Fe3O4@CNFs leads to a 125% and 178% increase in compressive strength and Young's modulus. Moreover, the introduction of 1 wt% CNFs in 30 kg/m3 tested samples increases compressive strength and Young's modulus by 68% and 140%, respectively. 4. in terms of flexure response, failure analyses of 30 kg/m3 tested samples demonstrate a 25% and 18% increase in flexure strength and Young's modulus by adding 1 wt% CNFs. 5. in terms of thermo-hygrometric response, the addition of 1 wt% CNFs into 30 kg/m3 tested panel leads to a 10% decrease in thermal conductivity. Moreover, nanoparticles reduce the moisture content of the foam which obtains a better resistance against the water adsorption.
[7] [8]
[9] [10]
[11]
[12] [13]
[14] [15]
[16]
[17]
[18]
[19]
Therefore, the introduction of CNFs is an effective method to improve the foam properties. Moreover, the alignment of functionalized CNFs within a high density PUR foam resulted to be a key factor in achieving the highest compression and tensile performance. Our future studies will be focused on PUR foams with higher densities and infused with other types of nanoparticles.
[20]
[21] [22] [23] [24]
Acknowledgements The authors wish to thank the technicians Orlando Favoni, Eng. Adriano Di Cristoforo and Alfredo De Leo for the support in laboratory tests as well as Simone Rogante (Claudio Foresi Company) for the samples manufacturing.
[25]
[26]
References [1] Dorel Feldman, Polymer nanocomposites in building, construction, J. Macromol.
245
Sci., Part A 51 (3) (2014) 203–209, http://dx.doi.org/10.1080/10601325.2014. 871948. F. Stazi, F. Tittarelli, G. Politi, C. Di Perna, P. Munafò, Assessment of the actual hygrothermal performance of glass mineral wool insulation applied 25 years ago in masonry cavity walls, Energy Build. 68 (2014) 292–304. F. Tittarelli, S.P. Shah, Effect of low dosages of waste GRP dust on fresh and hardened properties of mortars: Part 1, Construct. Build. Mater. 47 (2013) 1532–1538. L. James Lee, Changchun Zeng, Xia Cao, Xiangming Hang, Jiong Shen, Guojun Xu, Polymer nanocomposite foams, Compos. Sci. Technol. 65 (2005) 2344–2363. Caterina Lorusso, Viviana Vergaro, Annagrazia Monteduro, Antonio Saracino, Giuseppe Ciccarella, Paolo Maria Congedo, Barbara Federica Scremin, Characterization of polyurethane foam added with synthesized acetic and oleicmodified TiO2 nanocrystals, Nanomater. Nanotechnol. 5 (2015) 26, http://dx.doi. org/10.5772/61275 I. Gunjan, Maheshwari; Nilanjan, Mallik; Jandro, Abot; Albert, Song; Emily, Head; Vesselin, Shanov; Mark, Schulz; Chaminda, Jayasinghe; Pravahan, Salunke; YeoHeung, Yun; Sergey, Yarmolenko, Powering up nanoparticles: versatile carbon materials for engineering and medicine. Newsroom 10.1117/2.1200809.1305. Juan Lobos, Sachin Velankar, How much do nanoparticle fillers improve the modulus and strength of polymer foams? J. Cell. Plast. 0 (0) (2014) 1–32. M.C. Saha, MdE. Kabir, S. Jeelani, Enhancement in thermal and mechanical properties of polyurethane foam infused with nanoparticles, Mater. Sci. Eng. 479 (2008) 213–222. Xiao-Lin Xie, Yiu-Wing Mai, Xing-Ping Zhou, Dispersion and alignment of carbon nanotubes in polymer matrix: a review, Mater. Sci. Eng. R 49 (2005) 89–112. Hassan Mahfuz, Vijaya K. Rangari, Mohammad S. Islam, Shaik Jeelani, Fabrication, synthesis and mechanical characterization of nanoparticles infused polyurethane foams, Composites Part A 35 (2004) 453–460. Mohammad M. Hasan, Yuanxin Zhou, Shaik Jeelani, Thermal and tensile properties of aligned carbon nanofiber reinforced polypropylene, Mater. Lett. 61 (2007) 1134–1136. Dongzhi Yang, Jianfeng Zhang, Jing Zhang, Jun Nie. Aligned electrospun nanofibers induced by magnetic field. J. Appl. Polym. Sci. DOI 10.1002. Mohammad Chowdhury, George Stylios, Process optimization and alignment of PVA/FeCl3 nano composite fibres by electrospinning, J. Mater. Sci. 46 (2011) 3378–3386, http://dx.doi.org/10.1007/s10853-010-5226-5. Chuncheng Hao, Xiaojiao Li, Guizhen Wang, Magnetic alignment of nickel-coated carbon fibers, Mater. Res. Bull. 46 (2011) 2090–2093. Shahrzad, Rahmani; Mehdi, Rafizadeh; Framarz, Afshar, Taromi, Statistical analysis of nanofibers alignment in magnetic-field-assisted electrospinning including an alignment percentage formula. J. Appl. Polym. Sci. DOI: 10.1002/app.41179. Hassan Mahfuz, Shaik Zainuddin, Martin R. Parker, Tariq Al-Saadi, Vijay K. Rangari, Shaik Jeelani, Reinforcement of SC-5 epoxy with CNT/CNF under high magnetic field: an investigation of mechanical and thermal response, J. Mater. Sci. 44 (2009) 1113–1120. O. Malkina, H. Mahfuz, K. Sorge, A. Rondinone, J. Chen, K. More, S. Reeves, V. Rangari, Magnetic alignment of SWCNT decorated with Fe3O4 to enhance mechanical properties of SC-15 epoxy, AIP Adv. 3 (042104) (2013). Shuying Wu, Raj B. Ladani, Jin Zhang, Anthony J. Kinloch, Zhiheng Zhao, Jun Ma, Xuehua Zhang, Adrian P. Mouritz, Kamran Ghorbani, Chun H. Wang, Epoxy nanocomposites containing magnetite-carbon nanofibers aligned using a weak magnetic field, Polymer 68 (2015) 25–34. H. Mahfuz, et al., Response of sandwich composites with nanophased cores under flexural loading, Composites Part B 35 (2004) 543. Michele, Modesti; Alessandra, Lorenzetti; Stefano, Besco, Influence of nanofillers on thermal insulating properties of polyurethane nanocomposites foams. Polym. Eng. Sci. DOI 10.1002/pen.20819. MdE. Kabir, Effect of ultrasound sonication in carbon nanofibers/polyurethane foam composite, Mater. Sci. Eng. A 459 (2007) 111–116. O. Abramov, Ultrasound in Liquid and Solid Metals, CRC Press, Boca Raton, FL, 1994. D. Amodio, G.B. Broggiato, F. Campana, G.M. Newaz, Digital speckle correlation for strain measurement by image analysis, Exp. Mech. 43 (2003) 396–402. G. Palmieri, G. Chiappini, M. Sasso, S. Papalini, Hyperelastic materials characterization by planar tension tests and full-field strain measurement, Society for Experimental Mechanics - SEM Annual Conference and Exposition on Experimental and Applied Mechanics, 4 2009, pp. 2232–2238. Ulf Hammerschmidt, Jacques Hameury, Radek Strnad, Emese Turzò-Andras, Jiyu Wu, Critical review of industrial techniques for thermal-conductivity measurements of thermal insulation materials, Int. J. Thermophys. 36 (2015) 1530–1544. Cinzia Burratti, Elisa Moretti, Elisa Belloni, Fabrizio Agosti, Development of innovative aerogel based plasters: preliminary thermal and acoustic performance evaluation, Sustainability (2014), http://dx.doi.org/10.3390/su6095839.