PA66 microfibrillar composites

Accepted Manuscript Complex effect of graphite nanoplatelets on performance of HDPE/PA66 microfibrillar composites Ivan Kelnar, Ümitcan Bal, Alexander Zhigunov, Ludmila Kaprálková, Ivan Fortelný, Sabina Krejčíková, Jana Kredatusová PII:

S1359-8368(18)30275-0

DOI:

10.1016/j.compositesb.2018.03.006

Reference:

JCOMB 5565

To appear in:

Composites Part B

Received Date: 23 January 2018 Revised Date:

23 February 2018

Accepted Date: 3 March 2018

Please cite this article as: Kelnar I, Bal Ü, Zhigunov A, Kaprálková L, Fortelný I, Krejčíková S, Kredatusová J, Complex effect of graphite nanoplatelets on performance of HDPE/PA66 microfibrillar composites, Composites Part B (2018), doi: 10.1016/j.compositesb.2018.03.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Complex effect of graphite nanoplatelets on performance of HDPE/PA66 microfibrillar composites Ivan Kelnar*, Ümitcan Bal, Alexander Zhigunov, Ludmila Kaprálková, Ivan Fortelný, Sabina Krejčíková, Jana Kredatusová

*

Corresponding author: [email protected]

SC

ABSTRACT

RI PT

Institute of Macromolecular Chemistry, Academy of Sciences of the Czech Republic, Heyrovského nám. 2, 162 06 Praha, Czech Republic

The effect of nanofillers (NF) on parameters of polymer blends in microfibrillar composites

M AN U

(MFC) is complex due the effect of NF on melt drawing. This work concerns HDPE/PA66 modified with graphite nanoplatelets (GNP) prepared by different mixing protocols. GNP influence the dispersed phase size in the original blend negligibly and mostly lead to finer high-aspect ratio fibrils, i.e. GNP rather support elongation of inclusions than coalescence in

TE D

the course of drawing. Favourable mechanical behaviour, exceeding predicted one, was found with low GNP content using the PA66 masterbatch. MFC with similar structure and GNP localization in PA66 show marked differences in properties depending on mixing protocol.

EP

Antagonistic effects found for the HDPE masterbatch indicate high effect of GNP migration between the components which affects the interphase by variation of crystallinity. The results

AC C

confirm complex effect of GNP and dominancy of other GNP-induced effects over dual reinforcement with GNP and PA66 microfibrils. Keywords: Polymer-matrix composites; Interface/Interphase; Mechanical testing; Extrusion

ACCEPTED MANUSCRIPT Introduction Microfibrillar composites (MFC) are advantageous polymer-polymer composites prepared by melt or cold drawing of suitable polymer blends [1-3]. The reinforcing fibrils forming a minority phase, usually mechanically stronger than the matrix, must have sufficiently higher

RI PT

processing/melting temperature. This is necessary for subsequent effective processing, preferably using injection moulding [4]. Obvious limitation of MFC arising from polymeric reinforcement can be eliminated using NF [5-10]. In MFC, in addition to reinforcement, the

SC

well-known structure-directing activity [11-13] of NF is also of increased importance. This consists in altering thermodynamic and/or kinetic effects governing morphology of a polymer

M AN U

blend with crucial role of the localization of nanoparticles. Moreover, in some recent works even more complex role of NF in MFC in comparison with analogous undrawn systems was found [7-10]. In this respect, rather contradictory effect of NF on dispersed phase size in the course of shear mixing and dimensions of fibrils formed in the course of elongational flow

TE D

must also be considered. In the first case, NF influence dynamic phase behaviour by acting as an active interfacial modifier with effectivity similar to “classical” compatibilizers; reduction of coalescence and/or affecting of viscosity ratio mostly leads to marked refinement of

EP

dispersed phase size. On the contrary, especially presence of NF at the interface slows down fibre extension and can lead to an increase in attractive forces between the dispersed domains

AC C

in elongational flow [14].

As a result, important effect of NF on fibrils formation, consisting in support of coalescence in the course of drawing, can be utilized [14-16]. At the same time, this complex effect leads to both synergistic and antagonistic effects, most probably originating in affecting the interphase [17,18]. Therefore, respective NF effects must be harmonized in order to fully exploit the advantages of dual reinforcement and NF-induced structural changes [7-10]. The recent works indicate that potential of NF for MFC modification has not yet been fully

ACCEPTED MANUSCRIPT explored. Crucial role of NF was found in PCL/PLA [8-10]; this system, based on various combinations of commercial polymers, was characteristic by quite unstable extrusion making any drawing impossible. We have recently found substantial improvement of extrusion stability and melt drawing using high-AR nanoparticles, such as oMMT, HNT and GNP. The

RI PT

best efficiency was found for GNP; their fair effect on MFC formation (drawability) even allowed preparation of “classical” melt-drawn fibres [19], which was quite impossible for the unmodified blend. Based on the above results and the fact that the complex favourable effect

SC

of GNP has not yet been studied in MFC composed of conventional engineering polymers, the present work deals with GNP modification of HDPE/PA66–based MFC with the goal to

M AN U

utilize both higher reinforcing and structure-directing efficiency of carbon-based NF [20-24] in polymer systems in comparison with inorganic NF.

Materials

TE D

Experimental

High density polyethylene (HDPE) HYA 800, MFI 0.7g/10 min (Exxon Mobil); polyamide 66 (PA66) Zytel E55 NC10 (DuPont); HDPE/15% GNP masterbatch heXo-HDPE-15W;

EP

PA66/15% GNP masterbatch heXo-PA66-15W (NanoXplore Inc.); GNP is few-layer

AC C

graphene, heXo-G V (NanoXplore Inc.). All components were used as received

MFC preparation

Prior to mixing, PA66 was dried in a vacuum oven at 85 °C for 12 h. The mixing proceeded in a co-rotating segmented twin-screw extruder (L/D 40) Brabender TSE 20 at 400 rpm, and temperatures of the respective zones of 240, 260, 260, 275, 275, and 270 °C. The extruded bristle was melt-drawn using an adjustable take-up device. The draw ratio is the ratio between

ACCEPTED MANUSCRIPT the velocity of the take-up rolls and the initial velocity of the extruded bristle. Dog-bone specimens (gauge length 40 mm) were prepared in a laboratory micro-injection moulding machine (DSM). The barrel and the mould temperature was 200 °C and 70 °C, respectively.

RI PT

GNP addition protocols applied: a) application of pre-made PA66/GNP nanocomposite by “dilution” of concentrate in extruder, temperatures 240, 260, 260, 275, 275, and 270 °C; b) application of analogous HDPE/GNP nanocomposite (temperature of all zones 200°C); c) combination of a) and b). Example of sample composition: “HDPE+2 /PA66+2” means that

SC

MFC consists of HDPE pre-blended with 2 phr GNP and PA66 pre-blended with 2 phr GNP.

M AN U

The ratio of HDPE/PA66 was 80/20 w/w in all systems studied.

Testing

Tensile tests were carried out using an Instron 5800 apparatus at 22 °C and crosshead speed of 20 mm/min. At least eight specimens were tested for each sample. The Young modulus (E),

TE D

maximum stress (σm), and elongation at break (εb) were evaluated; the corresponding variation coefficients did not exceed 10 %, 2 % and 20 %, respectively. Tensile impact strength, at, was measured with one-side notched specimens using a Zwick

EP

hammer with energy of 4 J (variation coefficient 10–15 %). The reported values are averages of twelve individual measurements.

AC C

Dynamic mechanical analysis (DMA) was performed in single-cantilever mode using a DMA DX04T apparatus at 1 Hz and heating rate of 1 °C/min from -120 to 250 °C. The Differential Scanning Calorimetry (DSC) analysis was carried out using a Perkin-Elmer 8500 DSC apparatus. Samples of 5 ‒ 10 mg were heated from 50°C to 250°C at the heating rate of 10°C/min. The melting temperature Tm was identified as the melting endotherm maximum. The crystallinity was calculated using the values 293.6 and 226.0 J/g for the heat of melting of 100 %-crystalline HDPE and PA 66, respectively.

ACCEPTED MANUSCRIPT Characterization of structure The structure of the fibrils was examined using scanning electron microscopy (SEM) with a Vega (Tescan) microscope; the HDPE matrix was removed using a Soxhlet extraction apparatus with boiling xylene for 10 hours. The structure of undrawn blend was observed

RI PT

using SEM on cryo-fractured samples (liquid nitrogen) etched with formic acid for 30 min. The size of the dispersed elastomer particles was investigated by a MINI MOP image analyzer (Kontron Co., Germany). For the transmission electron microscope (Tecnai)

SC

observations, ultrathin (60 nm) sections were prepared under liquid nitrogen using an Ultracut UCT (Leica) ultramicrotome.

M AN U

Small-angle X-ray scattering (SAXS) and wide-angle X-ray scattering (WAXS) experiments were performed using a pinhole camera (Older Rigaku SMAX2000 upgraded by SAXSLAB/Xenocs) attached to a microfocused X-ray beam generator (Rigaku MicroMax 003) operating at 50 kV and 0.6 mA (30 W). The camera was equipped with a vacuum

TE D

version of Pilatus 300K detector. Experimental setup covering q range of 0.004 – 4 Å-1. Scattering vector, q, is defined as: q = (4π/λ)sinθ, where λ is wavelength and 2θ is scattering angle. Calibration of primary beam position and sample-to-detector distances was performed

AC C

EP

using Si powder sample in case of WAXS and AgBehenate powder in the SAXS region.

Results and Discussion

Effect of GNP and their localization on structure The SEM images (Fig. 1) indicate negligible affecting of dispersed PA66 particle size (~ 2.9 µm) by GNP. Minor decrease (to ~2.5 µm) was found in the case of HDPE/GNP pre-blend in combination with PA66, i.e. in the case of expected higher matrix viscosity, at least in the early stage of mixing (see GNP localization below). This indicates that the expected compatibilizing effect arising from possible partial interfacial localization and higher matrix

ACCEPTED MANUSCRIPT viscosity (in the case of HDPE/GNP pre-blend) [25] is eliminated by migration of GNP into the PA66 inclusions. This is obvious from the TEM observations (Fig. 2) indicating dominant final presence of GNP in the PA66 phase also in the case of the HDPE/GNP masterbatch application. As a result, the expected negative effect of increased dispersed phase viscosity

RI PT

limiting break-up of the dispersed PA66 inclusions [25] dominates. This is confirmed by

AC C

EP

TE D

M AN U

SC

presence of larger (~ 3.5 µm) inclusions in the case of PA66 + GNP pre-blend application.

Fig. 1. SEM images of undrawn blends a) HDPE+ 2/PA66 b) HDPE/PA6+2 c) HDPE/PA66

SC

RI PT

ACCEPTED MANUSCRIPT

M AN U

Fig. 2. TEM images of HDPE+2/PA66 microfibrillar composite

The mentioned predominant localization of GNP in PA66 is confirmed by TEM, TGA of the PA66 fibrils and, indirectly, by rheological behaviour (see below). This dominant GNP presence in PA66, due to migration from HDPE, is in contradiction with that predicted by

TE D

wetting coefficient value ~1.2 [26] indicating slight preference of their localization in the HDPE phase. Therefore, the final dominant localization inside PA66 seems to be a

EP

consequence of other dominant thermodynamic and kinetic factors [27]. Most important seems to be expected linking of amino groups of PA66 with epoxy functionality of GNP [28].

AC C

Further possible effective mechanism is shear-induced movements of GNP causing collisions with dispersed drops combined with possible trapping of GNP aggregates during droplet– droplet coalescence [29]. At the same time, possible hindering of this process (migration from HDPE) by relatively higher viscosity of HDPE seems to be of low importance [30].

b)

c)

d)

TE D

M AN U

SC

a)

RI PT

ACCEPTED MANUSCRIPT

Fig. 3. SEM images of PA66 fibrils (HDPE matrix extracted by boiling xylene) a)

cases

EP

HDPE+2/PA66. b) HDPE/PA66+2, c) HDPE+2/PA66+2, d) HDPE/PA66, draw ratio 6 in all

AC C

The SEM images of the fibrils (Fig. 3) show their relatively similar dimensions in all the systems studied, i.e. average diameter < 1 µm (range ~0.5 - ~1.5 µm) and length exceeding 50 µm (Table 1). Length was not evaluated due to certain tangling and crossing of long fibres. Moreover, increase of AR over 20 has negligible effect on mechanical parameters in the studied system [7]. The GNP addition, especially as the PA66/GNP masterbatch (i.e. without migration between the components) leads to predominantly slightly lower diameter of fibres with comparable length. (Table 1). Comparison of Figs 1 and 3 clearly indicates markedly higher volume of fibrils compared with original inclusions. This indicates coalescence in the

ACCEPTED MANUSCRIPT course of drawing [14-16]. The fibrils observed after GNP modification are mostly finer in comparison with unmodified MFC. This indicates higher drawability caused by GNP (also found in the analogous PCL/PLA system) at the expense of coalescence [19]. The only exception is MFC with 0.5 % GNP content at low draw ratio (DR), where higher diameters of

RI PT

fibres with more marked polydispersity occur; but, marked refinement of fibrils with higher extent of drawing was also found. The low effect of GNP on fibre dimensions and predominant fibre refinement are in contrast with the former HDPE/PA6/MMT system [7],

SC

where especially simultaneous addition of MMT led to significant increase in fibre diameter. In this system, MMT, particularly localized at and/or migrating across the interface [14]

M AN U

apparently supports coalescence. In this way, certain increase in fibre diameter (~0,9 µm) in MFC using the HDPE/GNP masterbatch, i.e. with significant GNP migration (and thus their presence at the interface), can be explained. However, it is obvious that also in this case the GNP-induced higher drawability prevails over the effect on coalescence. Finally, the fibrils

TE D

length increases up to DR ~ 8 in all cases. Surprisingly, the highest drawability was found with high GNP content, i.e. for the combination of the HDPE and PA/66 masterbatches. At the same time, in some MFCs (irrespective of mixing protocols and GNP content) the highest

EP

DR leads to reduction of fibre length (as confirmed by rheological behaviour below). This indicates that the predominant positive effect of GNP on fibrils formation is more complex

AC C

and exceeds the scope of this work. Tentatively, this feature corresponds to the fact that increase in DR also causes higher rate of drawing. The related higher extent of stretching attained within comparable time and distance leads to a finer formed bristle which is more easily cooled and solidifies faster than thicker one. This undoubtedly influences both coalescence and drawing of PA66, especially in presence of GNP [31]. Table 1. Properties of HDPE/PA66 80/20 blends and related MFC Composition

DR

E

Max. stress

ɛb

at

d

ACCEPTED MANUSCRIPT (MPa)

(MPa)

(%)

(kJ.m-2)

(µm)

1

1205±23

31.4±0.5

10.5±1.7

26.4±2.1

-

HDPE/PA66

7

1345±53

44.6±1.2

10.2±0.9

34.1±4

0.85

HDPE/PA66+1

1

1345±230

35.1±0.6

10.4±1.2

29.5±3.5

-

HDPE/PA66+1

7

1735±170

56.9±1.7

10.4±0.8

41±4

-

HDPE/PA66+2

1

1515±150

34.1±0.6

10.0±1.0

27,5±2.5

HDPE/PA66+2

8

1485±140

50.6±2.6

10.5±0.8

37±3.5

0.8

HDPE/PA66+3

1

1190±110

33.3±0.4

9.8±0.7

30±2.5

-

HDPE/PA66+3

6

1730±220

44.9±0.9

10.0±1.0

38±3

-

HDPE+1/PA66

1

1425±105

31.1±0.4

10.9±1.6

29±3

-

HDPE+1/PA66

6

1611±128

43.6±0.5

11.3±0.9

38±2.5

-

HDPE+2/PA66

1

1295±205

31.2±0.5

9.7±1

26± 3

-

HDPE+2/PA66

7

1475±470

49.1±1.6

11.3±1.2

42 ± 4

0.9

HDPE+2.5/PA66

1

1265±135

29.6±0.6

11.3±1.2

20±2.5

-

HDPE+2.5/PA66

8

1465±210

42.4±0.9

10.7±1.2

36±2.5

-

HDPE+1 /PAP66+1

1

1359±210

31.6±0.4

8.8±0.8

24.5 ± 2

HDPE+1 /PAP66+1

9

1460±80

40.7±0.7

10.3±0.6

35 ± 2,5

-

HDPE+2 /PAP66+2

1

1480±110

30.8±0.6

9.2±1.0

24.5 ± 2.3

-

HDPE+2 /PAP66+2

8

1580±120

37.1±0.4

10.7±0.4

32± 2.5

0.85

HDPE+2/

1

1375±120

30.8±0.2

9.8±0.7

19.5+3.5

-

1580±120

46±1

9.7±0.6

40±3

-

HDPE+2/

SC

M AN U

EP

PA66+0.5

7

TE D

PA66+0.5

RI PT

HDPE/PA66

DSC

AC C

Effect of GNP and drawing on crystallinity

The evaluation of crystallinity and melting points of both polymer components in the asprepared (1st run) samples indicates negligible differences in these parameters in all samples prepared (Table 2). The only exception is reduced crystallinity in the sample with the highest GNP content (2 phr in both components), due to hindering effect of GNP [32], and, in some cases, also with higher DR. Although this lower content of the crystalline phase occurs in samples with reduced mechanical parameters (see below), due to relatively insignificant

ACCEPTED MANUSCRIPT changes found (Table 2), it cannot be considered the main (important) reason of this drop. This is confirmed by the fact that some samples with the best mechanical performance have rather slightly lower crystallinity. Generally, it is obvious from DSC that neither GNP nor drawing has any marked effect on crystallinity of the components. Therefore, changes in

RI PT

mechanical behaviour are probably induced by other effects discussed below, e.g. crystallinity at the interface affected by GNP migration. Therefore, XRD and PLM have also been applied.

∆Hm1 (J/g)

Xc1 (%)

Tm2 (°C)

∆Hm2 (J/g)

Xc2 (%)

HDPE/PA66+2 DR7

133

189

64

261

60

27

HDPE/PA66 DR1

133

195

66

260

65

29

HDPE/PA66 DR7

133

HDPE/PA66+1 DR1

135

HDPE/PA66+1 DR5.5

137

HDPE/PA66+1 DR7

132

HDPE/PA66+2 DR7

134

HDPE/PA66+3 DR5,5

133

188

64

260

60

27

189

64

261

55

24

179

61

261

45

20

191

65

257

70

31

194

66

260

65

29

196

67

260

70

31

TE D

HDPE/PA66+3 DR7

SC

Tm1 (°C)

M AN U

Table 2. DSC analysis of the HDPE/PA66 80/20 blends and MFC

134

200

68

260

55

24

131

193

66

260

60

27

134

185

63

260

65

29

134

185

63

261

60

27

HDPE+2,5/PA66 DR7

132

190

65

260

65

29

HDPE+2/PA66+0,5 DR1

132

181

62

261

65

29

HDPE+2/PA66+0,5 DR6

132

181

62

261

70

31

HDPE+2/PA66+0,5 DR7

132

171

58

259

65

29

HDPE+2/PA66+0,5 DR8

133

190

65

261

50

22

HDPE+2/PA66+2 DR7

133

190

65

260

50

22

HDPE/PA66+2 DR8 HDPE+1/PA66 DR5.5

AC C

EP

HDPE+1/PA66 DR7

XRD

ACCEPTED MANUSCRIPT The XRD patterns presented in Fig. 4 indicate that all the samples are identical, showing same crystalline structure and degree of crystallinity, namely ~ 57 %. In addition, no differences in degree of orientation around 62 % were found. Nevertheless, minor differences occur in the small angle X-ray scattering profiles in Fig. 5. According to these results, samples could be

RI PT

split into two groups, based on peak position in q-range 0.045 – 0.1 Å-1. Interestingly, the second peak at q = 0.063 Å-1, which corresponds to correlation distance of 10 nm, was only found in the samples showing the best (HDPE/PA66+1, DR 8) and the worst

SC

(HDPE+2/PA66+2, DR 7) mechanical properties. The remaining samples have this peak shifted towards higher angle with q = 0.072 Å-1, which corresponds to d = 87 Å. Finally, the

M AN U

most intense peak for all the samples is found at q ~ 0.025-0.026 Å-1, and is usually assigned to lamellar periodicities. In average, it led to distance of 245 Å.

The fact that the only dimensional changes found are practically identical for the best and the worst MFC obviously indicates that this parameter has negligible effect on mechanical

TE D

performance. This also confirms that changes in mechanical behaviour are induced by other effects.

d, A

EP

17,65 8,83 5,90 4,43 3,56 2,98 2,56 2,25 2,01 1,82 1,67 4

Intensity

AC C

3

WAXS HDPE/PA66+2 DR8 HDPE/PA66+2 DR6 HDPE+2/PA66+0,5 DR8 HDPE+2/PA66+0,5 DR7 HDPE+2/PA66+0,5 DR6 HDPE/PA66+1 DR7 HDPE+2/PA66+2 DR8

2

1

0 5

10

15

20

25

30

35

2Θ, degrees

Fig. 4. X-ray diffraction patterns of MFC

40

45

50

55

ACCEPTED MANUSCRIPT

2

10

HDPE/PA66+2 DR8 HDPE/PA66+2 DR6 HDPE+2/PA66+0,5 DR6 HDPE+2/PA66+0,5 DR7 HDPE+2/PA66+0,5 DR8 HDPE/PA66+1 DR7 HDPE+2/PA66+2 DR7

1

0

10

-2

10

-3

10

-2

10

-1

q, A

-1

10

Polarized light microscopy

M AN U

Fig. 5. Small angle X-ray scattering profiles of MFC

RI PT

-1

10

SC

Intensity, arb.u.

10

From the observations of the thin layer samples in Figs 6 a,b, it can be speculated about lower content of spherulites near the fibres [18] in the sample HDPE+2/PA66+2,DR 7 showing the

TE D

worst mechanical parameters. This may indicate presence of a “soft interphase” [33] with negative effect mainly on stiffness as indicated by the recent finite element analysis FEA study [18]. Unfortunately, accuracy (reliability) of this observation is limited by lower

AC C

EP

visibility of spherulites due to high GNP content.

a)

b)

Fig. 6. Polarized light microscopy images of a)HDPE/PA66+2, b)HDPE+2/PA66+2

ACCEPTED MANUSCRIPT

Dynamic mechanical analysis Temperature dependences of loss moduli in Figs 7 a,b show increase in Tg of PA66 with DR in all the systems studied, but most marked in the GNP-free system (~6 °C) . At the same

RI PT

time, GNP increased Tg of the undrawn systems (~2-5 °C), similarly to the effect on single PA66 (~2.5 °C increase by addition of 2 % GNP). As a result, Tg of the drawn GNP-modified systems is comparable to that of unmodified MFC. The relatively lower increase in Tg by

SC

drawing in presence of GNP was also found in similar (HDPE/PA6) oMMT-modified MFC [7] and indicate that usual increase in Tg by added nanofillers [34] is partly eliminated by

M AN U

drawing-induced negative effects. This may include e.g. drawing-induced destruction of GNP stacks and related increase in molecular mobility by presence of relatively large nanoplatelets and/or their elongated stacks with size (length) exceeding gyration radius of neighbouring (attached) polymer chains [35]. This is most probably combined with possible GNP-induced

TE D

changes in parameters of the interfacial area discussed below. Moreover, this behaviour further confirms the still not very well understood effect of different NF on chain dynamics

EP

[36].

AC C

Loss modulus (MPa)

150

100

50

DR1 DR 5.5 DR 8

0

-100

0

Temperature (°C)

a)

100

200

ACCEPTED MANUSCRIPT

100

50

0

DR 8,5 DR 7 DR 1 DR 5,5 DR 6,5 -100

0

100

RI PT

Loss Modulus (MPa)

150

200

Temperature (°C)

b)

M AN U

SC

Fig. 7. Temperature dependences of loss modulus a) HDPE/PA66 b) HDPE+1/PA66+1

Effect of GNP and fibrils formation on mechanical properties

The results of mechanical testing in Table 1 demonstrate low correspondence of mechanical properties to reinforcement of respective phases (Table 3) and practically the same dimensions of the PA66 fibrils. This is confirmed especially by the fact that the best

TE D

behaviour was achieved with low GNP content applied as the pre-blend in the PA66 phase, i.e in a system with absence of GNP migration (see the above mentioned GNP localization in

EP

PA66). In this case, the E value markedly exceeds the one predicted by Halpin-Tsai model (<1400 MPa) [37], which indicates synergistic effects, like drawing–induced hardening,

AC C

transcrystallinity, etc. Surprisingly, rather slight decrease in parameters with higher 2 and 3 phr GNP content occurs. Moreover, in these MFC, decrease of parameters with the highest DR was found (Fig 8), but practically without reduction of fibre length (see below). Additionally, variation of AR over its value of ~25 has low effect on properties [7]. Moreover, shorter fibrils may be rather beneficial due to possible reduction of entanglements [38]. Therefore, in accord with other NF-modified MFC [7,17,18], the reason may be negative effect of higher GNP content in PA66 and thus at the interface on crystallinity of the HDPE matrix in this area. Moreover, this may be supported by higher interfacial interactions of

ACCEPTED MANUSCRIPT oxidized carbon–based nanofillers [23.24]. The recent FEA [18] of semicrystalline-matrix MFC showed dramatic effect of reduced amount of rigid spherulites at the fibres surface on modulus. At the same time, it follows from Table 1 that even lower parameters were found in all systems with GNP pre-blended in HDPE either in combination with neat PA66 or

RI PT

PA66/GNP pre-blend. In all these systems, the total content of GNP was higher in comparison with the mentioned best MFC, but also with dominant final presence of GNP inside PA66 (see above). E.g. TGA of separated fibrils indicates migration of majority of GNP into MFC

SC

consisting of HDPE+2phr GNP pre-blend/PA66, leading to ~7 phr GNP content in PA66. Although this undoubtedly leads to enhanced parameters of the PA66 fibrils due to the

M AN U

simultaneous lower reinforcement of the matrix (reduction of GNP content in the matrix), changed GNP localization in the respective polymer phases has relatively low impact on properties. This is demonstrated on the model MFC with composition of either HDPE+1%GNP/PA66 or HDPE/PA66+4%GNP. Application of Halpin-Tsai model [37]

TE D

using experimental values of respective component parameters (partly shown in Tab. 3) indicates practically identical modulus of both systems. Nevertheless, GNP migration from the matrix to PA66 should enhance stiffness due to expected alignment and thus higher

EP

parameters of GNP-modified fibres. In MFC with higher GNP content in PA66, especially combination of PA66+GNP and HDPE+GNP pre-blends, lower parameters of the PA66 phase

AC C

caused by limited reinforcing effect of higher GNP content (due to restacking) [39] must also be considered. This is in agreement with the worst properties of MFC with the highest GNP content (HDPE+2/PA66+2).

ACCEPTED MANUSCRIPT

1600

1200

HDPE/PA66 HDPE/PA66+1 HDPE/PA66+2 HDPE+2/PA66 HDPE+2/PA66+2 HDPE+2/PA66+0,5

1000 800 600 0

2

4

6

8

RI PT

E (MPa)

1400

10

Draw ratio

SC

a) 60 55

45 40 35 30

HDPE/PA66 HDPE/PA66+1 HDPE/PA66+2 HDPE+2/PA66 HDPE+2/PA66+2 HDPE+2/PA66+0,5

25 20 15 10 5 0 0

2

4

6

8

10

TE D

Draw ratio

b) 45 40

EP

35 30 25 20

HDPE/PA66 HDPE/PA66+1 HDPE/PA66+2 HDPE+2/PA66 HDPE+2/PA66+2 HDPE+2/PA66+05

15

AC C

Toughness (kJ.m-2)

M AN U

Stress at break (MPa)

50

10 5 0

0

2

4

6

8

10

Draw ratio

c) Fig. 8. Effect of drawing ratio on mechanical properties of MFC a) modulus b) stress at break c) tensile impact strength

Table 3. Properties of components and related nanocomposites

ACCEPTED MANUSCRIPT at (kJ.m-2)

E

Max. stress

ɛb

(MPa)

(MPa)

(%)

HDPE

1040±95

32.6±0,4

25.9±2,5

54±4

HDPE+2 GNP

1231±120

32.1±1

27.4±2.9

50,5 ± 4

PA66

2278±210

75.9±2.5

55.0±19.9

-

PA66+2GNP

2726±2215

82.5±0.5

11.7±1.8

-

RI PT

Composition (phr)

To summarize the above mentioned results, explanation of relatively low parameters in spite

SC

of higher GNP content, comparable fibrils dimensions and GNP localization most probably consist in the negative effect of high GNP content in PA66 (~7% in the case of HDPE+2GNP

M AN U

masterbatch) on HDPE crystallinity near the fibres. This seems to be supported by higher GNP content at the interface caused by migration. In the case of higher DR, these negative effects seem to be also stimulated by changed solidification of the thinner bristle due to its faster cooling.

TE D

80 70 60

PA66+2 HDPE+2 PA66 HDPE HDPE/PA66+1 DR1 HDPE/PA66+1 DR7

50 40

EP

Tensile Strength (MPa)

90

30 20

AC C

10 0

0

10

20

30

40

50

Strain (%)

Fig 9. Stress-strain curves of polymer components and MFC Fig. 9 demonstrates the effect of reinforcement by GNP and fibrils formation on stress-strain behaviour of the components and MFC. In the case of PA66, GNP reduces yielding and strain hardening. Only slightly lower extent of plastic deformation occurs with HDPE+2GNP. The curves of all HDPE/PA66 MFC indicate that their deformational behaviour is similar to that

ACCEPTED MANUSCRIPT of the HDPE matrix; reinforcement with both GNP and fibrils increases E and strength with negligible effect on elongation. Finally, important feature of all MFC consists in the fact that in-situ fibrils formation leads to increased toughness (Tab. 1), which is rather supported by GNP modification. The thorough study of fracture behaviour (J integral) is in progress. As a

RI PT

result, application of GNP and harmonization of respective GNP–induced effects is a tool to

Effect of GNP on Rheology of Components and MFC

SC

design polymeric materials with enhanced well-balanced mechanical parameters.

Figs. 10 a and 10 b show relatively marked effect of GNP on increase of viscosity of both

M AN U

single components accompanied by shear thinning. This clearly corresponds to presence of fairly dispersed high-aspect-ratio nanoplatelets. On the other hand, Fig 10 c shows significantly lower increase of viscosity by GNP in the 80/20 blend. If we take the above mentioned dominant localization of GNP inside PA66 (at the expense of HDPE) into account,

TE D

we can explain the low increase of viscosity in the blend as a consequence of this phenomenon. The reason consists in the fact that impact of GNP localized in the dispersed PA66 on the system viscosity is less marked due to much lower shear deformation of the

EP

dispersed phase (moreover with higher GNP-caused rigidity) in comparison with the GNPeffect on the matrix viscosity. In other words, viscosity of a blend is proportional to viscosity

AC C

of the matrix but contribution of viscosity of dispersed droplets to the blend viscosity is limited only [40]. This undoubtedly leads to lower increase in viscosity of the blend in comparison with single components with analogous GNP content. In turn, this behaviour “indirectly” confirms the predominant GNP localization in PA66.

ACCEPTED MANUSCRIPT

1000

100

HDPE HDPE+1GNP HDPE+2GNP

10

1 0,1

1

10

100

Freq (rad/s)

SC

100000

1000

PA66 PA66+1GNP PA66+2GNP

100

10 0,1

1

10

Freq (rad/s)

100

TE D

100000

M AN U

Eta* (Pa-s)

10000

10000

HDPE+2/PA66 HDPE/PA66+2 HDPE+2/PA66+2 HDPE/PA66 HDPE+1/PA66+1

1000

EP

Eta* (Pa-s)

RI PT

Eta* (Pa-s)

10000

AC C

100 0,1

1

Freq (rad/s)

10

100

Fig 10. Effect GNP on viscosity of components (a,b) and HDPE/PA66 80/20 blend at 280°C (c)

Figs 11 a,b show that at 200°C, i. e. in the case of short fibre-composite (without melting of PA66 fibrils), increase in viscosity with draw ratio accompanied by more significant shear thinning mostly occurs. This confirms formation of fibres and their prolonging with higher DR [41].

ACCEPTED MANUSCRIPT Rather unexpected finding is decrease in viscosity with GNP addition shown in Fig. 12. Most probably, this also indicates certain shortening of fibrils by GNP addition. We consider the mentioned higher drawability (due to GNP) causing reduction of fibre diameter to be probably accompanied by reduced coalescence causing fibre shortening [16]. Unfortunately, the

RI PT

expected decrease in fibrils length cannot be reliably evaluated from the extracted fibrils observed by SEM due to tangling and crossing of fine long fibrils. Some effect on the observed decrease in viscosity may arise from higher rigidity of GNP-containing fibres which

SC

can be more easily aligned in shear flow, as documented by more marked shear thinning in

M AN U

comparison with the unmodified sample.

100000

1000

DR 1 DR 5.5 DR 6,5 DR 8 DR 9

100

10 0,1

1

TE D

Eta* (Pa-s)

10000

10

100

Freq (rad/s)

EP

a)

AC C

Eta* (Pa-s)

100000

10000

DR1 DR 5.5 DR 6.6 DR 7 DR 8 DR 9,5 DR 11

1000

100 0,1

1

10

100

Freq (rad/s)

b) Fig 11. Effect of draw ratio on complex viscosity of MFC at 200°C in in dependence on GNP addition ratio a) HDPE/PA66, b) HDPE+2/PA66+2

ACCEPTED MANUSCRIPT

100000

1000

HDPE/PA66+2 HDPE/PA66 HDPE+2/PA66+2 HDPE+1/PA66+1 HDPE+2/PA66

100

10 0,1

1

10

100

Freq (rad/s)

RI PT

Eta* (Pa-s)

10000

SC

Fig 12. Effect of GNP addition method and content on complex viscosity of MFC at 200°C,

M AN U

draw ratio 8. Conclusions

The results indicate good potential of GNP to upgrade MFC together with more complex effect of GNP in drawn systems in comparison with analogous blends. Consequently, more rigid and strong material with enhanced toughness can be obtained. In contrast to

TE D

nanosilicates, GNP support drawing of the PA66 inclusions at the expense of coalescence, which leads to finer fibres with high AR. Typical feature of GNP-modified MFC is low

EP

correspondence of mechanical properties to dual reinforcement with GNP and in-situ formed fibrils. In dependence on the method of GNP addition and content, both positive and negative

AC C

deviations from predicted values were found. This behaviour was affected by degree of drawing as well and it occurs in spite of similar structure (dimensions of fibrils) and predominant GNP localization inside the PA66 phase in all MFC. The fact that the best behaviour was achieved with low amount of GNP premixed in PA66 indicates negative effect of high GNP content inside fibrils on parameters of the interface, most probably determined by variation in HDPE crystallinity. This is apparently supported by predominant presence of GNP at the interface by migration during drawing and solidification.

ACCEPTED MANUSCRIPT Acknowledgement

AC C

EP

TE D

M AN U

SC

RI PT

This work was supported by Czech Science Foundation (Grant No 16-03194S)

ACCEPTED MANUSCRIPT References [1] K. Friedrich, M. Evstatiev, S. Fakirov, O. Evstatiev, M. Ishii, M. Harrass, Microfibrillar reinforced composites from PET/PP blends: processing, morphology and mechanical properties, Comp. Sci. Technol. 65 (2005) 107–116.

RI PT

[2] Jayanarayan K, Thomas S, Kuruvilla J. Morphology, static and dynamic mechanical properties of in situ microfibrillar composites based on polypropylene/poly (ethylene terephthalate) blends, Compos A Appl Sci Manuf 2008; 39(2);164-175.

SC

[3] Z.-M. Li, L.-B. Li, K.-Z. Shen, W. Yang, R. Huang, M.-B. Yang, Transcrystalline

M AN U

morphology of an in situ microfibrillar poly(ethylene terephthalate)/poly(propylene) blend fabricated through a slit extrusion hot stretching-quenching process, Macromol. Rapid Commun. 25 (2004) 553–558.

[4] R.J.T. Lin, D. Bhattacharyya, S. Fakirov, Innovative manufacturing of carbon nanotube-

TE D

loaded fibrillar polymer composites, Int. J. Mod. Phys. B 24 (2010) 2459–2465. [5] S. Yesil, G. Bayram, Effect of carbon nanotube surface treatment on the morphology, electrical, and mechanical properties of the microfiber-reinforced polyethylene/poly(ethylene

EP

terephthalate)/carbon nanotubecomposites, J. Appl. Polym. Sci. 127 (2013) 982–991.

AC C

[6] W.J. Li, A.K. Schlarb, M. Evstatiev, Study of PET/PP/TiO2 microfibrillar-structured composites, J. Appl. Polym. Sci. 113 (2009) 3300–3306. [7] I. Kelnar, L. Kaprálková, J. Kratochvíl, J. Kotek, L. Kobera, J. Rotrekl, J. Hromádková, Effect of nanofiller on the behaviour of a melt-drawn HDPE/PA6 microfibrillar composite, J. Appl. Polym. Sci. 132 (2015) 41868. [8] I. Kelnar, J. Kratochvíl, I. Fortelný, L. Kaprálková, A. Zhigunov, V. Khunová, M. Nevoralová, Effect I of halloysite on structure and properties of melt-drawn PCL/PLA microfibrillar composites, eXpress Polym. Lett. 10 (2016) 381–393.

ACCEPTED MANUSCRIPT [9] I. Kelnar, I. Fortelný, L. Kaprálková, J. Kratochvíl, M. Nevoralová, Effect of layered silicates on fibril formation and properties of PCL/PLA microfibrillar composites, J. Appl. Polym. Sci. 133 (2016) 43061. [10] I. Kelnar, J. Kratochvíl, I. Fortelný, L. Kaprálková, A. Zhigunov, M. Nevoralová, Effect

composites, Polym. Compos. (2017). ; DOI: 10.1002/pc.24322.

RI PT

of graphite nanoplatelets on melt drawing and properties of PCL/PLA microfibrillar

[11] M.Y. Gelfer, H.H. Song, L. Liu, B.S. Hsiao, B. Chu, M. Rafailovich, M. Si, V. Zaitsev,

SC

Effects of organoclays on morphology and thermal and rheological properties of polystyrene

M AN U

and poly(methyl methacrylate) blends, J. Polym. Sci. Pt. B-Polym. Phys. 41 (2003) 44–54. [12] I. Kelnar, V. Sukhanov, J. Rotrekl, L. Kaprálková, Toughening of recycled poly(ethylene terephthalate) with clay-compatibilized rubber phase, J. Appl. Polym. Sci. 116 (2010) 3621– 3628.

TE D

[13] I. Kelnar, J. Rotrekl, L. Kaprálková, J. Hromádková, Effect of poly(oxyalkylene)amines on structure and properties of epoxide nanocomposites, J. Appl. Polym. Sci. 125 (2012) 2755–2763.

EP

[14] G. Filippone, D. Acierno, Clustering of coated droplets in clay-filled polymer blends,

AC C

Macromol. Mater. Eng. 297 (2012) 923–928. [15] S. Fakirov, D. Bhattacharyya, R.J.T. Lin, C. Fuchs, K. Friedrich, Contribution of coalescence to microfibril formation in polymer blends during cold drawing, J. Macromol. Sci. Part B-Phys. 46 (2007) 183–193. [16] I. Kelnar, I. Fortelný, L. Kaprálková, J. Hromádková, Effect of nanofiller on fibril formation in melt-drawn HDPE/PA6 microfibrillar composite, Polym. Eng. Sci. 55 (2015) 2133–2139.

ACCEPTED MANUSCRIPT [17] I. Kelnar, L. Kaprálková, J. Kratochvíl, Z. Padovec, M. Růžička, J. Hromádková, Effect of layered silicates and reactive compatibilization on structure and properties of melt-drawn HDPE/PA6 microfibrillar composites, Polym. Bull. 73 (2016) 1673–1688. [18] I. Kelnar, J. Kratochvíl, L. Kaprálková, Z. Padovec, M. Růžička A. Zhigunov,

RI PT

M. Nevoralová, Antagonistic effects on mechanical properties of polymer composites with dual reinforcement: Explanation by FEA model of soft interface, J. Appl. Polym. Sci. 134 (2017) 44712.

SC

[19] I. Kelnar, A. Zhigunov, L. Kaprálková, I. Fortelný, J. Dybal, J. Kratochvíl, M.

M AN U

Nevoralová, M. Hricová, V. Khunová, Facile preparation of unique biocompatible poly (lactic acid)-reinforced poly(ɛ-caprolactone) fibers via graphite nanoplatelets -aided melt spinning, ACS Appl. Mater. Interfaces, submitted.

[20] J.R. Potts, D.R. Dreyer, C.W. Bielavski, R.S. Ruoff, Graphene-based polymer

TE D

nanocomposites, Polymer 52 (2011) 5–25.

[21] Y. Cao, J. Zhang, J. Feng, P. Wu, Compatibilization of immiscible polymer blends using graphene oxide sheets, ACS Nano 5 (2011) 5920–5927.

EP

[22] T.D. Thanh, L. Kaprálková, J. Hromádková, I. Kelnar, Effect of graphite nanoplatelets

AC C

on the structure and properties of PA6-elastomer nanocomposites, Eur. Polym. J. 50 (2014) 39–45.

[23] H. Lu, F. Liang, Y. Yao. J. Gou, D. Hui, Self-assembled multi-layered carbon nanofiber nanopaper for significantly improving electrical actuation of shape memory polymer nanocomposite. Compos B Eng. 59 (2014) 191-195. [24] H. Lu, Y. Yao, W.M. Huang, Noncovalently functionalized carbon fiber by grafted selfassembled graphene oxide and the synergistic effect on polymeric shape memory nanocomposites. Compos B Eng. 67 (2014) 290-295.

ACCEPTED MANUSCRIPT [25] I. Fortelný, Theoretical Aspect of Phase Morphology Development. In: Micro- and Nanostructured Multiphase Polymer Blends Systems; C. Harrats, S. Thomas, G. Groeninckx, Eds. Taylor and Francis, Boca Raton, 2006, pp. 43–90. [26] J. Dai, G. Wang, L. Ma, C. Wu, Study of the surface energies and dispersibility of

RI PT

graphene oxide and its derivatives, J. Mater. Sci. 50 (2015) 3895–3907.

[27] L. Zonder, S. Mccarthy, F. Rios, A. Ophir, S. Kenig, Viscosity ratio and interfacial

tension as carbon nanotubes distributing factors in melt-mixed blends of polyamide 12 and

SC

high-density polyethylene, Adv. Polym. Technol. 33 (2014) 21427.

M AN U

[28] H.K. Jeon, B.J. Feist, S.B. Koh, K. Chang, C.W. Macosko, R.P. Dion, Reactively formed block and graft copolymers as compatibilizers for polyamide 66/PS blends, Polymer 45 (2004) 197–206.

[29] L. Elias, F. Fenouillot, J.-C. Majesté, G. Martin, P. Cassagnau, Migration of nanosilica

TE D

particles in polymer blends, J. Polym. Sci. Pt. B-Polym. Phys. 46 (2008) 1976–1983. [30] F. Gubbels, R. Jerome, E. Vanlathem, R. Deltour, S. Blacher, F. Brouers, Kinetic and Thermodynamic control of the selective localization of carbon black at the interface of

EP

immiscible polymer blends, Chem. Mat. 10 (1998) 1227–1235.

AC C

[31] I. Kelnar, J. Mikešová, J. Kratochvíl, L. Kaprálková, Shear flow effects on the γ crystallinity formation in PA6/Montmorillonite nanocomposite, J. Appl. Polym. Sci. 106 (2007) 3387–3393.

[32] J. Kratochvil, I. Kelnar, Non-isothermal kinetics of cold crystallization in multicomponent PLA/thermoplastic polyurethane/nanofiller system, J. Therm. Anal. Calorim. 130 (2017) 1043–1052.

ACCEPTED MANUSCRIPT [33] W. Xu, F. Wu, Y. Jiao, M. Liu, A general micromechanical framework of effective moduli for the design of nonspherical nano- and micro-particle reinforced composites with interface properties, Mater. Des. 127 (2017) 162–172. [34] J. Brus, M. Urbanová, I. Kelnar, J. Kotek, Glass-transition and segmental dynamics of

NMR, Macromolecules 39 (2006) 5400–5409.

RI PT

polymer chains in semicrystalline nylon-6/clay nanocomposites investigated by solid-state

[35] G. Coativy, C. Chevigny, A. Rolland-Sabaté, E. Leroy, D. Lourdin, Interphase vs

SC

confinement in starch-clay bionanocomposites, Carbohydr. Polym. 117 (2015) 746–752.

M AN U

[36] Y. Wang, W. Wang, Z. Zhang, L. Xu, P. Li, Study of the glass transition temperature and the mechanical properties of PET/modified silica nanocomposite by molecular dynamics simulation, Eur. Polym. J. 75 (2016) 36–45.

[37] J.C. Halpin, J.L. Kardos, The Halpin-Tsai equations: a review. Polym. Eng. Sci. 16

TE D

(1976) 344–352.

[38] M. Samir, F. Alloin, M. Paillet, A. Dufresne, Tangling effect in fibrillated cellulose reinforced nanocomposites, Macromolecules 37 (2004) 4313–4316.

AC C

Munich, 2008.

EP

[39] S.N. Bhattacharya, R.K. Gupta, M.R. Kamal, Polymeric Nanocomposites, Hanser,

[40] L.A. Utracki, Rheology of Polymer Blends. In: Encyclopedia of Polymer Blends, Vol.: Processing, A.I. Isayev, Ed. Wiley-VCH, Weinheim, Germany, 2011, pp. 27–107. [41] G.V. Vinogradov, A.Y. Malkin, Rheology of Polymers, Mir Publishers, Moscow, 1980, p 380.

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

.