- Email: [email protected]
Polypropylene hybrid composites reinforced with long glass fibres and particulate filler
COMPOSITES SCIENCE AND TECHNOLOGY Composites Science and Technology 65 (2005) 257–267 www.elsevier.com/locate/compscitech
Polypropylene hybrid composites reinforced with long glass fibres and particulate filler J. Hartikainen
a,*
, P. Hine b, J.S. Szabo´ c, M. Lindner a, T. Harmia a, R.A. Duckett b, K. Friedrich d
a
FACT Future Advanced Composites and Technology GmbH, Hertelsbrunnenring 9, D-67657 Kaiserslautern, Germany IRC in Polymer Science and Technology, School of Physics and Astronomy, University of Leeds, Leeds LS2 9JT, UK Department of Polymer Engineering, Faculty of Mechanical Engineering, Budapest University of Technology and Economics, H-1111 Budapest, Mu¨egyetem rkp. 3., Hungary d Institut fu¨r Verbundwerkstoffe GmbH, Erwin-Schro¨dinger-Straße Geb. 58, D-67663 Kaiserslautern, Germany b
c
Received 10 May 2004; received in revised form 6 July 2004; accepted 11 July 2004 Available online 3 September 2004
Abstract Structure and fracture behaviour of long glass fibre (LGF) reinforced polypropylene (LGF PP) composites including calcium carbonate (CaCO3) as a filler were studied. Fibre orientation, fibre length distribution and mechanical properties of LGF PP/CaCO3 hybrid composites, as well as the crystallinity changes of polypropylene upon filler addition are reported. Furthermore, an acoustic emission (AE) analysis was applied to the fracture mechanical test, in order to get information about the fracture modes during the loading. It was found out that the filler addition had little effect on the fibre orientation and crystallisation behaviour of LGF PP, but the average fibre length decreased. AE analysis showed that the addition of filler caused early stage debonding of the LGFs, when the samples were subjected to low speed tensile loading (1 mm/min). These observations may explain the changes of the mechanical properties of LGF PP upon filler addition. 2004 Elsevier Ltd. All rights reserved. Keywords: A. Polymer-matrix composites; A. Hybrid compounds; B. Fracture; C. Anisotrophy; D. Acoustic emission
1. Introduction Polypropylene (PP) is one of the most successful thermoplastic polymers ever, characterised by its excellent cost-to-performance ratio. Often the mechanical properties of PP are modified by adding various mineral fillers, the most studied filler types being talc and calcium carbonate [1,2]. It has been repeatedly shown that such filler particles increase YoungÕs modulus of PP, yet causing the decrease of the strength and the toughness. PP can also be reinforced with short glass fibres (SGF) to improve the stiffness and the fracture *
Corresponding author. E-mail address: hartikainen@fact-kunststoffe.de (J. Hartikainen).
0266-3538/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2004.07.010
toughness. However, long glass fibres (LGF) are used more and more often as a reinforcement, since it is known that longer fibres with the same fibre diameter (i.e. with higher fibre aspect ratio) provide higher stiffness, tensile strength and toughness compared to shorter ones [3–6]. Recently, it has been observed that by incorporating filler particles into the matrix of fibre reinforced composites, synergistic effects may be achieved in the form of higher modulus and reduced material costs, yet accompanied with decreased strength and impact toughness [2,7]. Such multicomponent hybrid composites based on poly(butylene terephtalate) (PBT) [8], acrylonitrile–butadiene–styrene co-polymer (ABS) [9,10] and PP [7,11] matrices have been studied.
258
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
However, hybrid reinforced composites form a complex system and there is inadequate data available about phenomena behind the property changes due to the addition of particulate fillers to the fibre reinforced thermoplastic composites. Thus, the present paper is an attempt to clarify the mechanical behaviour of hybrid composites based on the PP matrix, LGF reinforcement and calcium carbonate particulate filler, by a systematic study on flow-induced fibre alignment, fibre length degradation, matrix crystallinity and fracture mechanisms. Fibre orientation produced by injection moulding of plate shaped sample specimens was studied using a purpose built image analysis facility [12,13], to establish whether the addition of the calcium carbonate affects the orientation of the moulded components. The average fibre length in the moulded components was also determined, again to assess any effect of the filler on glass fibre properties. Tensile properties, as well as fracture toughness accompanied with an acoustic emission (AE) analysis, were determined to get information on failure mechanisms of filler filled, fibre reinforced and hybrid reinforced PP composites. Furthermore, morphology changes due to the particle addition were analysed by differential scanning calorimetry (DSC). The microstructure and failure modes of such hybrid composites are discussed on the basis of these experimental results.
2.2. Sample preparation Firstly, calcium carbonate concentrate was dry blended with PP to 40 wt% filler concentration and compounded by using a Haake Bucher Rheocord System 40 laboratory scale twin screw extruder. Temperature profile was set to 180–280 C and the used screw speed was 40 rpm. The extruded material was granulated into pellets with length of 11 mm. Also neat PP was processed into 11 mm pellets, in order to facilitate the proper mixing during the injection moulding and also to harmonize processing histories of the components. Next, the samples were prepared by mechanically mixing the compounded materials with LFT pellets in different ratios of fillers, glass fibres and PP (see Table 1 for sample denotion and mixing ratios). The bone shaped specimens (geometry Type 2 – Fig. 1, according to standard ISO 3167) were then injection moulded using Arburg Allrounder 320 C 600-250 injection moulding machine with maximum clamping force of 600 kN. The processing temperature was 225–270 C and the mould temperature was 50 C. Similarly, 2 mm thick plate shaped specimens (Geometry Type 1 – Fig. 1) were injection moulded by the company Werner GmbH (Austria). The used melt temperature during the moulding of the plate specimens was 220–240 C and the mould temperature was 15 C. 2.3. Fibre orientation and fibre length analysis
2. Experimental 2.1. Materials LGF reinforced PP pellets were provided by FACT GmbH (Germany). The material was of the type FACTOR PP LGF 60 N 11, a LGF reinforced PP with glass fibre content of 60 wt%. Fibre length in the pellets was the same as the pellet length of 11 mm. Calcium carbonate was used as a polyolefin based concentrate including olefin based particle coating (87 wt% filler content) and it was provided by Omya GmbH (Germany). The type of the concentrate was Omyalene 102 with a mean particle size of 2.0 lm. Neat PP was provided by Basell (homopolymer with an MFI of 45).
The system used for the fibre orientation and length characterisation (only Geometry Type 1) is an in-house image analysis facility developed at University of Leeds [12]. For fibre orientation investigations, the chosen method is one of optical reflection microscopy of polished 2D sections taken from the areas of interest of the composite. Each fibre that meets the 2D section is seen as an elliptical footprint, and measuring the ellipticity of these images allows the two polar angles, h and /, that specify the orientation of each fibre to be determined: h is the angle the fibre makes with the sectioned surface normal (1) and / is the angle the fibre makes with the two axis when projected into the 23 plane (Fig. 2(a)). Fig. 2(b) shows a typical elliptical fibre foot-
Table 1 Sample denotion and mechanical properties of LGF PP/CaCO3 hybrid composites Formulation
Fibre loading (vol%)
CaCO3 loading (vol%)
Tensile strength (MPa)
Tensile modulus (GPa)
Fracture toughness (MPa
A B C D E F
0 0 3.7 4.2 12.9 14.6
0 7.7 0 7.4 0 6.6
38a 27a 64 58 112 96
1.68 1.76 3.34 3.89 7.27 7.96
1.62 1.52 2.95 2.54 4.49 3.96
a
Yield stress.
p
m)
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
259
2.4. Differential scanning calorimetry Mettler Toledo DSC 821 device was used to study the effect of the used CaCO3 filler on crystallisation behaviour of PP. The samples were first heated to 190 C and then cooled back to room temperature to equalize thermal histories. Crystallisation peak temperature Tc was determined from the second cooling scan. The rate of the heating and cooling was 10 C/min. 2.5. Mechanical properties
Fig. 1. Schematic drawing describing sample geometries.
A Zwick UPM 1485 universal testing apparatus equipped with 250 kN load cell, a mechanical clamping device and an extensiometer was used to measure the tensile strength and fracture toughness of the samples. Tensile strength test was carried out according to the standard ISO 527 for a dumbbell shaped standard specimen by using a constant crosshead speed of 2 mm/min. Fracture properties were analysed by using single edge notched tensile (SEN-T) specimens with a notch depth of 2 mm. In fracture mechanical test, a constant crosshead speed of 1 mm/min was used. Critical stress intensity factor Kc was determined according to the following equation [14]: pffiffiffi K c ¼ r c Y a; ð1Þ where rc is the gross fracture stress, a is the length of the notch and Y is the geometrical factor. The following equation was used for geometrical factor Y [15]: pffiffiffi 5 p Y ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð2Þ 20 13ða=W Þ 7ða=W Þ2 where W is the specimen width. Although the formulas (1) and (2) strictly hold for homogeneous and isotropic materials, they are often adopted for inhomogeneous, non-isotropic materials, including discontinuous fibre reinforced injection moulded composites [16,17].
Fig. 2. Definition of the orientation angles h and /.
2.6. Acoustic emission print, and indicates that h is given by the inverse cosine of the ratio of the semi-minor to semi-major axis of each elliptical image and / by the angle between the major axis and the two axis. An XY stage allows a large area to be scanned, allowing a determination of fibre orientation over a large area. For further details of the system and the analysis routines see [13]. Fibre length characterisation is achieved by taking a chosen section of the sample, burning off the PP matrix in a furnace for 8 h at 450 C, and then spreading the resulting glass fibres thinly on a glass plate. In this instance the fibres are viewed in transmitted light and image analysis routines are used to determine the length of each fibre shadow image.
In order to get a deeper understanding of the failure behaviour, AE of the notched dumbbell specimens was recorded in situ (during the loading), by using a miniature sensor (10 mm diameter) attached to the specimen surface and coupled to a Defektophone NEZ 220 device (AEKI, Hungary). The acoustic events were picked up by a wide bandwidth heat-proof transducer in the frequency range 100–6001 kHz (peak sensitivity ±70 dB/ V/lbar, type Micro-30D of Dunegan Co., USA). The output signal of the transducer was amplified logarithmically. The transfer function for the whole measuring system (including the logarithmic amplifier and acquisition unit) was
260
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
Peak amplitude ¼ 100 þ 20 lgðU inp =0:4Þ ½dB referred to 1V:
ð3Þ
During the tests the following primary AE signals were measured, calculated and stored: elapsed time, number of events, peak amplitude, AE energy, event width and rise time [18]. 2.7. Scanning electron microscopy The fracture surfaces after LEFM test were analysed using JEOL JSM 5400 scanning electron microscope. Before the analysis, the fracture surfaces were first cautiously cleaned with air blow and then coated with a thin layer of gold–palladium alloy (ratio 5/1).
3. Results and discussion 3.1. Fibre orientation and length analysis (Geometry Type 1 – Fig. 1) The main aim of this section of the study was to assess the effect of the addition of the CaCO3 filler on the orientation and length of the glass fibres. This was carried out using the Type 1 injection moulded plates (Fig. 1), and using samples E (LGF/CaCO3 = 12.9/0 vol%) and F (LGF/CaCO3 = 14.6/6.6 vol%). 3.1.1. Fibre orientation studies Fibre orientation in moulded components was compared at two areas, A and B, as shown in Fig. 1. Area A was located in the centre of the hexagon shaped section of the component. Here three sections were investigated, all through thickness sections (XZ plane). Each section was 30 mm long and 15 mm apart from each others. Fig. 3 shows a typical set of image analysis results, in this case for the section nearest the arrow on Fig. 1. During scanning, h and / of each individual fibre image recognised are stored so that after scanning it is possible to reconstruct the scanned area. Fig. 3(a) shows a reconstruction of half of the scanned section (2 mm thick and 15 mm long). In common with the majority of injection moulded plates, the sample shows the location of central core region where the fibres are aligned perpendicular to the injection direction and outer shell regions where the fibres are more preferentially aligned along the injection (X) direction. An instructive way to display this data quantitatively is to calculate values of the second order orientation averages hcos2hXi, hcos2hYi and hcos2hZi in strips across the sample: the higher the value of each of these averages, the greater the orientation along that axis and in addition the three averages always add up to 1. Fig. 3(b) shows values of these parameters by dividing the sample into strips across the thickness direction. Here we see clearly the outer
Fig. 3. Fibre orientation at scanning area A: (a) typical reconstituted image scan, (b) typical values of the second order orientation average hcos2hi i across the sample thickness and (c) a comparison of the histogram of h for samples E (LGF/CaCO3 = 12.9/0 vol%) and F (LGF/CaCO3 = 14.6/6.6 vol%).
shell layers and central core layers: the core occupies approximately 20% of the sample thickness. Interestingly the maximum level of orientation in the two layers is roughly similar albeit with respect to different axes. Finally we can see that the level of out-of-plane orientation (Z axis) is very low, which would be expected for a 2 mm thick plate and the long fibres used here. For the sample material E (no CaCO3) analysis of the three sections showed very little difference, suggesting that the fibre orientation was homogeneous over a 30 mm square region at the centre of the hexagon shaped section. A similar analysis for sample F (+CaCO3) showed the same good agreement. Fig. 3(c) therefore shows a comparison of the fibre orientation at the same section position (again the section nearest the arrow on Fig. 1) for the two samples. In this case we have presented the data as a histogram of the angle h, but any other comparison technique would give the same result,
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
namely that the fibre orientation state for the two sample formulations is not identical but very similar. A similar comparison was then carried out at the second scanning area, B. This section was again taken through the sample thickness (XZ plane). The section was taken 4 mm behind the hole, and it was 35 mm in width and centrally located. Fig. 4(a) shows a comparison of the fibre orientation for the two samples E and F. In this instance we have displayed the data as an area map of the second order orientation average hcos2hXi: here a white colour denotes high orientation along the X axis. It is seen that qualitatively, the addition of the CaCO3 filler has little effect of the fibre orientation at this position. Again we have used a strip analysis to produce quantitative data, this time with strips parallel to the Z axis and moving along the sample width. Fig. 4(b) (for the sample Type E) shows this analysis. It can be seen that there is a region (which is directly located behind the hole in this sample) where the fibres
261
are highly aligned in the flow direction. Fig. 3(c) compares the strip analyses for hcos2hXi for the two samples, and it seen that both the shape and quantitative values are very similar, confirming that the incorporation of the CaCO3 filler has little effect of the orientation of the glass fibres. 3.1.2. Fibre length studies Fibre length studies were carried out on pieces taken from the narrow central section of the two sample Types (E and F). As described above, the samples were placed into a furnace to burn away the PP matrix and then placed onto a glass plate for analysis. Fig. 5 shows the length distributions obtained from this analysis (1300 fibres) for each sample. The calculated number average lengths for these two distributions were as follows: sample E (LGF/CaCO3 = 12.9/0 vol%) – 887 lm: sample F (LGF/CaCO3 = 14.6/6.6 vol%) – 648 lm. It appears that while the CaCO3 filler does not significantly affect the fibre orientation state, it does lead to a reduction in the average fibre length. 3.2. Crystallinity changes The most prominent effect of particulate fillers on the crystalline structure of semi-crystalline thermoplastics is their ability to work as a nucleation agent [2]. Though there has been some debate over the effects of crystalline changes on macromechanical properties, it seems clear that by increasing crystallinity, the modulus of PP increases and the strength and deformability decrease [2,19]. However, the nucleation effect differs strongly by filler type, particle size (i.e. surface area) and filler surface treatment. For example, it has been observed that talc is an active filler, affecting strongly on the crystalline morphology of PP [20]. Calcium carbonate, instead, has typically little effect on crystallinity and can be considered as an inactive filler in sense of nucleation effect [20–22]. By modifying the surface chemistry of the
Fig. 4. Fibre orientation at scanning area B: (a) area maps of the second order orientation average Æcos2hX æ for the two materials and (b) strip analysis along the scanned width of sample E and (c) a comparison of the strip analysis for the two materials.
Fig. 5. Fibre length distribution in plate shaped specimen comparing sample E (LGF/CaCO3 = 12.9/0 vol%) and sample F (LGF/ CaCO3 = 14.6/6.6 vol%).
262
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
filler, nucleation effect may be unaffected, increased or decreased [21–24]. On the other hand, smaller particle size has been observed to increase the activity of CaCO3 as a nucleation agent, possibly due to aggregate formation [21]. In this work the effect of the used filler on crystallinity of PP was analysed by DSC. Fig. 6 depicts DSC thermograms at 100–150 C region of the second cooling cycle for pure PP (sample A) and PP including 7.7 vol% of calcium carbonate (sample B). Crystallisation exotherm appears as a single, distinct peak in both cases and the Tc temperature is almost unchanged (125 C for neat PP and 123 C for filled PP). Thus, it seems that the filler used in our study is inactive in sense of nucleation effect, probably due to the facts that (1) the average particle size of the used filler was relatively high (2 lm), and (2) the filler particles were surface treated to decrease the surface energy, which both are known to reduce the ‘‘activity’’ of calcium carbonate.
Fig. 7. Effect of filler addition to the shape of tensile stress–strain curves at different LGF/CaCO3 ratios (vol%): 3.7/0 (sample C), 4.2/7.4 (sample D), 12.9/0 (sample E) and 14.6/6.6 (sample F).
3.3. Tensile properties Tensile stress–strain curves of LGF PP/CaCO3 composites at different levels of fibre and filler contents are shown in Figs. 7 and 8 which present the averaged tensile strength and tensile modulus values. The values are also collected in Table 1. It can be seen that both the strength and modulus are distinctly increased as the LGF content increases. On the other hand, the elongation at break is simultaneously reduced, which can be attributed to the fact that reinforcing fibres strongly restrain the deformation of the matrix polymer as demonstrated in several previous studies [10,25]. The shape of the curves remains essentially unchanged as calcium carbonate filler is added to the matrix, but both the maximum strength and strain decrease. However, the modulus values (Fig. 8(b)) show that by filler addition the stiffness of the LGF composite may be further en-
Fig. 8. Tensile strength and tensile modulus of LGF PP/CaCO3 composites. In the cases of samples A and B yield stress is given.
Fig. 6. DSC scans (second cooling cycle) showing the crystallinity exotherms for neat PP (sample A) and PP with 7.7 vol% of CaCO3 (sample B).
hanced. In the case of calcium carbonate filled PP (sample B) it can be noticed that the modulus increases about 5% upon filler addition, which is in accordance with the values reported elsewhere [20,26]. On the other hand, in the cases of LGF reinforced composites the modulus increase is more distinct i.e. of the order of 10–15%. Thus, it seems that a synergy effect takes place by incorporating particulate filler into the matrix, leading to a higher stiffness than would be expected on the basis of change of the matrix modulus. Note that the amount of filler in the matrix was kept constant (7.7 vol%, or 20 wt%)
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
in all samples, and that the crystallinity effects can be excluded, as was evidenced by the DSC analysis. 3.4. Fracture behaviour Linear elastic fracture mechanics offers a useful way of getting information about different fracture mechanisms which take place during the loading of composite specimens, especially if the mechanical testing is accompanied by other methods like AE or optical microscopy [3,27–29]. In the present work we used an LEFM approach to compare fracture behaviour of the LGF reinforced PP composites to that of the filler filled PP and LGF PP/CaCO3 hybrid composites. The fracture toughness values obtained are shown in a diagram of Fig. 9. It can be observed that the critical stress intensity factor Kc increases considerably when the materials are reinforced with LGF. As calcium carbonate is added to the matrix, Kc remains almost unchanged in the case of unreinforced PP, but decreases in the case of hybrid composites. It is widely accepted that the fracture toughness of PP can be increased by reinforcing with LGFs [1]. On the other hand, it has been reported that filler inclusions may lead to the decreased fracture toughness when added to thermoplastics like PP, especially if large particle agglomerates are present. For example, it has been observed that the fracture toughness of PP decreases sharply above a 5 vol% addition of calcium carbonate decreasing to half of the original value at 30 vol% filler content [30]. In this study, the amount of calcium carbonate was only 7.7 vol% of the matrix, and particle agglomeration was minimised by surface treatment of the filler and by the separate compounding step. Consequently, the fracture toughness of neat PP was almost unaffected by the filler addition. However, in hybrid composites Kc values decreased, which suggests that there are effects that cannot be explained purely by the filler-matrix related mechanisms. In order to get a deeper understanding on the failure mode and sequence, an AE analysis was carried out during the
Fig. 9. Fracture toughness of LGF PP/CaCO3 composites.
263
fracture mechanical test. In the case of unreinforced PP, the main mode of energy absorption during the fracture mechanical test is matrix deformation. However, as PP is reinforced with LGFs, several additional failure mechanisms are observed to occur in the fracture process; such as fibre debonding, fibre breakage, and pull-out of fibres and fibre bundles [3]. These different failure mechanisms can be distinguished by AE, because they give acoustic signals at certain characteristic amplitudes. Based on numerous AE studies performed with the same experimental setup on discontinuous fibre-reinforced composites, the following correlation exists between the AE amplitude ranges and individual failure events [1,31]: fibre/matrix debonding (<35 dB) <fibre pull-out (35–55 dB) <fibre fracture (>55 dB). The cumulative numbers of the AE events registered during loading of SEN-T specimens are summarized in Table 2. The loading curve was sectioned as follows: range I ends at 40% of Fmax, range II extends from 40% of Fmax to 80% of Fmax, range III extends from 80% of Fmax to Fmax, and range IV covers the section from Fmax downwards (Fig. 10). A notable feature in the values of Table 2 is that the neat PP and its CaCO3 modified version were acoustically much less active than LGF reinforced composites. Furthermore, the signals are of low amplitude, i.e. close to the threshold level (10 dB), which reflect the matrix related failure mechanisms [3]. The related signals may arise from matrix deformation (neat PP) and particle/matrix debonding (calcium carbonate filled PP). As the LGF content increases, so does the cumulative number of AE signals. The AE amplitude distribution remains practically the same in samples including LGF reinforcement, in each loading section (I. . .IV), independent of fibre content. This means that the same type of failure mechanism occurs in the respective loading sections. In previous studies on fracture behaviour of LGF PP composites it has been observed that crack initiation takes place before the maximum load is achieved [3]. Similar conclusion was reached using AE in the present study. Fig. 11 shows a distribution of the amplitudes at different stages of force-elongation curve of hybrid composite including 14.6 vol% of LGF and 6.6 vol% of calcium carbonate. It can be seen that there is some AE activity already during the stage I of failure sequence. This indicates that some fibre debonding and related matrix deformation may take place at early stage of the loading. The activity increases considerably at stage II, indicating the start of the crack propagation. At stage II higher amplitude signals are emitted, which suggests that fibre pullouts take place. In stage III fibre debonding and pullouts are even more evident, and additionally fracture of fibres occurs as suggested by acoustic signals emitted with amplitudes over 55 dB. Finally the activity decreases as the fracture terminates during the last stage IV. Thus, it can be
264
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
Table 2 Relative distribution of amplitude during the AE analysis Relative distribution of amplitude (%)
Formulation A: PP/LGF/CaCO3 100/0/0 vol%
Formulation B: PP/LGF/CaCO3 92.3/0/7.7 vol%
Formulation C: PP/LGF/CaCO3 96.3/3.7/0 vol%
Formulation D: PP/LGF/CaCO3 88.4/4.2/7.4 vol%
Formulation E: PP/LGF/CaCO3 87.1/12.9/0 vol%
Formulation F: PP/LGF/CaCO3 78.8/14.6/6.6 vol%
Amplitude interval (dB) 31–40
Number of signals
11–20
21–30
41–50
51–60
61–70
71–80
81–90
I II III IV
100.0 87.5 100.0
12.5
I II III IV
100.0 100.0 69.7
21.2
9.1
I II III IV
72.0 31.9 24.2 62.1
12.0 15.9 10.0 16.7
16.0 11.6 11.1 1.5
15.9 30.8 1.5
22.0 22.5 10.6
2.7 1.4 7.6
1097
I II III IV
70.9 73.8 7.4 16.5
21.8 17.3 6.2 6.4
7.3 3.1 19.3 13.8
2.7 56.0 52.1
2.5 11.1 11.2
0.6
833
I II III IV
82.8 73.8 51.3 38.3
14.8 17.3 12.5 25.2
2.4 3.1 6.3 13.6
2.7 12.6 9.9
2.5 14.4 10.1
0.6 2.8 2.9
I II III IV
76.5 64.2 28.3 41.3
18.5 17.2 11.5 31.9
5.0 4.9 12.2 15.9
5.9 20.0 7.2
5.8 21.1 2.2
1.7 6.6 1.5
11
36
2592 0.1
0.3 0.3
1717
Sectioning of the force–elongation curve: I, 0–0.4 Fmax; II, 0.4–0.8 Fmax; III, 0.8–1 Fmax; IV, 1–0 Fmax.
Fig. 10. Load–displacement curve and cumulative AE signals of the SEN-T specimen (LGF/ CaCO3 = 12.9/0). Sectioning of the load– displacement curve is shown with Roman numerals I–IV.
concluded that fibre related mechanisms dominate the fracture behaviour of studied composites. It can be noted from the values in Table 2 that the addition of CaCO3 to LGF reinforced composites results in a pronounced decrease in AE activity. Secondly, there are interesting changes in amplitude distribution upon addition of filler, especially evident at lower
Fig. 11. Distribution of CaCO3 = 14.6/6.6 vol%).
amplitude
in
AE
analysis
(LGF/
LGF contents. Firstly, the amplitude distribution at stages I and II indicates more frequent matrix deformation and fibre debonding related fracture incidents as fil-
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
265
ler is added to the LGF PP composite, compared to samples without a filler. In the case of sample D (PP including 4.2 vol% of LGF and 7.4 vol% of calcium carbonate), more than 90% of the signals at stage II are emitted at amplitudes below 30 dB. However, in the case of sample C (PP with 3.7 vol% of LGF) these were less than 50%. This means that filler particles increase the incidents of matrix deformation and fibre debonding [1,31]. Secondly, at later phases of the failure (stages III and IV) the majority of the incidents, in the case of hybrid samples, take place at amplitudes that suggest fibre debonding. Thus, it seems that filler particles induce particle debonding, which may explain the decreased mechanical properties like strength and fracture toughness. This view has already been suggested in other studies of SGF reinforced ABS/CaCO3 [10] and glass mat reinforced PP/mica [11] hybrid composites. These conclude that fillers may cause premature fibre debonding under loading, due to void formation around particles near fibres, or due to direct contact between particles
and fibres. It should be noted that the increase of low amplitude signals at stages I and II in hybrid samples cannot be explained only by filler related events, since the number of AE arising from matrix deformation was only slightly increased as calcium carbonate was added to neat PP. Thus, it can be concluded that LGF-related failure events dominate also in hybrid systems, but the number of the acoustic signals is less than in the case of LGF composites without a filler. This can be explained by the fact that pullouts occur more easily due to reduced fibre–matrix adhesion (originating from the strong debonding at early stage of failure sequence). Finally, the fracture surfaces after fracture mechanical testing were studied by scanning electron microscopy (SEM). In the case of LGF PP composite with 12.9 vol% of fibres (Fig. 12(a)) there is clear evidence of ductile matrix deformation, which is a result of the fibre debonding and pullout processes. On the other hand, a more irregular fracture surface is observed in the case of corresponding hybrid sample
Fig. 12. SEM micrographs of fracture surfaces after fracture mechanical test for PP hybrid composites at 350 times magnification: (a) LGF/ CaCO3 = 12.9/0 vol% and (b) LGF/CaCO3 = 14.6/6.6 vol%. Crack growth direction is from left to right in both cases.
Fig. 13. SEM micrographs of fracture surfaces after fracture mechanical test for PP hybrid composites: (a) LGF/CaCO3 = 12.9/0 vol%, 750· magnification and (b) LGF/CaCO3 = 14.6/6.6 vol%, 1500· magnification. Crack growth direction is from left to right.
266
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
(Fig. 12(b)). In addition, there are more marks of pullout events as indicated by the high amount of voids originating from the disconnected fibres. Fig. 13 shows SEM micrographs at higher magnification, demonstrating the difference in matrix deformation of LGF reinforced composites and hybrid materials. As calcium carbonate is added to the matrix, void formation around the filler particles take place during the loading. Fig. 13(b) presents the fracture surface at a site where fibre has been pulled out. It can be seen that the particles and the voids may come into contact with fibres and thus, induce early stage fibre debonding. Furthermore, it can be noted from Figs. 12 and 13 that ductility is reduced in filler filled samples as evidenced by the smaller deformation patterns compared to unfilled LGF composites. It can be concluded that this ductility reduction of the matrix also contributes to the decrease of fracture toughness when calcium carbonate is added to LGF PP. These SEM observations are in good agreement with the mechanical test results and AE analysis.
4. Conclusions The present work showed that the properties of LGF PP composites can be modified by adding calcium carbonate. It was shown that the addition of CaCO3 in LGF PP increased the stiffness. On the other hand, the strength and toughness were decreased. Various analysis methods were used to investigate the mechanisms behind the property changes. Optical image analysis for the plate shaped specimens showed that the addition of CaCO3 did not considerably affect the fibre orientation. However, the average length of the fibres was 25% lower in the case of hybrid composites compared with LGF PP. The fracture mechanical test accompanied with AE analysis was used to study the fracture modes. It was shown that the fracture toughness of unreinforced PP remained unaffected as calcium carbonate was added. However, the fracture toughness of LGF reinforced PP decreased upon filler addition. The AE of the samples showed that the acoustic activity of LGF PP/ CaCO3 hybrid samples was much higher compared to LGF PP without a filler. Secondly, the distribution of the amplitudes suggested that filler addition results in debonding of the fibres at early stage of the loading. This results in fibre pull-out mechanism at earlier phase of the fracture sequence compared to unfilled samples. Furthermore, SEM analysis for the fracture surfaces showed decreased ductility in the case of filler filled hybrid materials. This may have a contributory influence on decrease of fracture toughness values upon CaCO3 addition.
Acknowledgements HTP Fohnsdorf GmbH (Austria) is acknowledged for the permission to use the mould of the plate shaped specimens. Prof. J. Karger-Kocsis is gratefully acknowledged for the numerous discussions and support. J. Hartikainen wishes to thank the European Commission for financial support (Marie Curie Fellowships program, contract number G5TR-CT2001-00052).
References [1] Karger-Kocsis J. Microstructural aspects of fracture in polypropylene and in its filled, chopped fiber and fiber mat reinforced composites. In: Karger-Kocsis J, editor. Polypropylene: Structure, blends and composites. London: Chapman & Hall; 1995. p. 142–201. [2] Pukanszky B. Particulate filled polypropylene: structure and properties. In: Karger-Kocsis J, editor. Polypropylene: Structure, blends and composites. London: Chapman & Hall; 1995. p. 1–70. [3] Karger-Kocsis J, Harmia T, Cziga´ny T. Comparison of the fracture and failure behavior of polypropylene composites reinforced by long glass fibers and by glass mats. Compos Sci Technol 1995;54(3):287–98. [4] Karger-Kocsis J. Swirl mat- and long discontinuous fiber matreinforced polypropylene composites-status and future trends. Polym Compos 2000;21:514–22. [5] Stokes VK, Inzinna LP, Liang EW, Trantina GG, Woods JT. A phenomenological study of the mechanical properties of longfiber filled injection-molded thermoplastic composites. Polym Compos 2000;21(5):696–710. [6] Thomason JL. The influence of fibre length and concentration on the properties of glass fibre reinforced polypropylene: 5. Injection moulded long and short fibre PP. Composites: Part A 2002;33:1641–52. [7] Acosta JL, Morales E, Ojeda MC, Linares A. Effect of addition of sepiolite on the mechanical properties of glass fiber reinforced polypropylene. Angew Makromol Chem 1986;138:103–10. [8] Hargarter N, Friedrich K, Catsman P. Mechanical properties of glass fiber/talc/polybutylene-terephtalate composites as processed by the radlite technique. Compos Sci Technol 1993;46:229–44. [9] Fu S-Y, Lauke B. Fracture resistance of unfilled and calciteparticle-filled ABS composites reinforced by short glass fibers (SGF) under impact load. Composites: Part A 1998;29:631–41. [10] Fu S-Y, Lauke B. Characterization of tensile behaviour of hybrid short glass fibre/calcite particle/ABS composites. Composites: Part A 1998;29:575–83. [11] Zhao R, Huang J, Sun B, Dai G. Study of the mechanical properties of mica-filled polypropylene-based GMT composite. J Appl Polym Sci 2001;82:2719–28. [12] Hine PJ, Davidson N, Duckett RA, Ward IM. Measuring the fibre orientation and modelling the elastic properties of injectionmoulded long-glass-fibre-reinforced Nylon. Compos Sci Technol 1995;53:125–31. [13] Hine PJ, Davidson N, Duckett RA, Clarke AR, Ward IM. Hydrostatically extruded glass-fiber-reinforced polyoxymethylene. I: The development of fiber and matrix orientation. Polym Compos 1996;17:720–9. [14] Anderson TL. Fracture mechanics – fundamentals and applications. Boca Raton: CRC Press Inc.; 1995. [15] Harris DO. Stress intensity factors for hollow circumferentially notched round bars. J Basic Eng 1967;89(1):49–54.
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267 [16] Hashemi S, Koohgilani M. Fracture toughness of injection molded glass fiber reinforced polypropylene. Polym Eng Sci 1995;35(13):1124–32. [17] Pegoretti A, Ricco T. Fatigue crack propagation in polypropylene reinforced with short glass fibres. Compos Sci Technol 1999;59:1055–62. [18] Szabo´ JS, Cziga´ny T. Investigation of static and dynamic fracture toughness on short ceramic fiber reinforced polypropylene composites. J Macromol Sci, Phys B 2002;B41:1191–204. [19] Wright DGM, Dunk R, Bouvart D, Autran M. The effect of crystallinity on the properties of injection moulded polypropylene and polyacetal. Polymer 1988;29(5):793–6. [20] Pukanszky B, Belina K, Rockenbauer A, Maurer FHJ. Effect of nucleation, filler anisotropy and orientation on the properties of PP composites. Composites 1994;25(3):205–14. [21] Kowalewski T, Galeski A. Influence of chalk and its surface treatment on crystallization of filled polypropylene. J Appl Polym Sci 1986;32:2919–34. [22] Rybnika´r F. Interactions in the system isotactic polypropylenecalcite. J Appl Polym Sci 1991;42(10):2727–37. [23] Garton A, Kim SW, Wiles DM. Modification of the interface morphology in mica-reinforced polypropylene. J Polym Sci, Polym Lett Ed 1982;20:273–8. [24] Zuiderduin WCJ, Westzaan C, Huetink J, Gaymans RJ. Toughening of polypropylene with calcium carbonate particles. Polymer 2003;44:261–75.
267
[25] Thomason JL, Vlug MA, Schipper G, Krikor HGLT. Influence of fibre length and concentration on the properties of glass fibrereinforced polypropylene: Part 3. Strength and strain at failure. Composites: Part A 1996;27:1075–84. [26] Demjen Z, Pukanszky B, Nagy J. Evaluation of interfacial interaction in polypropylene/surface treated CaCO3 composites. Composites: Part A 1998;29:323–9. [27] Cziga´ny T, Karger-Kocsis J. Comparison of the failure mode in short and long glass fiber-reinforced injection-molded polypropylene composites by acoustic emission. Polym Bull 1993;31: 495–501. [28] Cziga´ny T, Mohd Ishak ZA, Karger-Kocsis J. On the failure mode in dry and hygrothermally aged short fiber-reinforced injection-molded polyarylamide composites by acoustic emission. Appl Compos Mater 1995;2:313–26. [29] Cziga´ny T, Marosfalvi J, Karger-Kocsis J. An acoustic emission study of the temperature-dependent fracture behavior of polypropylene composites reinforced by continuous and discontinuous fiber mats. Compos Sci Technol 2000;60(8):1203–12. [30] Levita G, Marchetti A, Lazzeri A. Fracture of ultrafine calcium carbonate/polypropylene composites. Polym Compos 1989;10: 39–43. [31] Benevolenski O, Karger-Kocsis J. Fracture and failure behavior of partially consolidated discontinuous glass fiber mat-reinforced polypropylene composites (Azdel SuperLite). Macromol Symp 2001;170:165–79.
Polypropylene hybrid composites reinforced with long glass fibres and particulate filler J. Hartikainen
a,*
, P. Hine b, J.S. Szabo´ c, M. Lindner a, T. Harmia a, R.A. Duckett b, K. Friedrich d
a
FACT Future Advanced Composites and Technology GmbH, Hertelsbrunnenring 9, D-67657 Kaiserslautern, Germany IRC in Polymer Science and Technology, School of Physics and Astronomy, University of Leeds, Leeds LS2 9JT, UK Department of Polymer Engineering, Faculty of Mechanical Engineering, Budapest University of Technology and Economics, H-1111 Budapest, Mu¨egyetem rkp. 3., Hungary d Institut fu¨r Verbundwerkstoffe GmbH, Erwin-Schro¨dinger-Straße Geb. 58, D-67663 Kaiserslautern, Germany b
c
Received 10 May 2004; received in revised form 6 July 2004; accepted 11 July 2004 Available online 3 September 2004
Abstract Structure and fracture behaviour of long glass fibre (LGF) reinforced polypropylene (LGF PP) composites including calcium carbonate (CaCO3) as a filler were studied. Fibre orientation, fibre length distribution and mechanical properties of LGF PP/CaCO3 hybrid composites, as well as the crystallinity changes of polypropylene upon filler addition are reported. Furthermore, an acoustic emission (AE) analysis was applied to the fracture mechanical test, in order to get information about the fracture modes during the loading. It was found out that the filler addition had little effect on the fibre orientation and crystallisation behaviour of LGF PP, but the average fibre length decreased. AE analysis showed that the addition of filler caused early stage debonding of the LGFs, when the samples were subjected to low speed tensile loading (1 mm/min). These observations may explain the changes of the mechanical properties of LGF PP upon filler addition. 2004 Elsevier Ltd. All rights reserved. Keywords: A. Polymer-matrix composites; A. Hybrid compounds; B. Fracture; C. Anisotrophy; D. Acoustic emission
1. Introduction Polypropylene (PP) is one of the most successful thermoplastic polymers ever, characterised by its excellent cost-to-performance ratio. Often the mechanical properties of PP are modified by adding various mineral fillers, the most studied filler types being talc and calcium carbonate [1,2]. It has been repeatedly shown that such filler particles increase YoungÕs modulus of PP, yet causing the decrease of the strength and the toughness. PP can also be reinforced with short glass fibres (SGF) to improve the stiffness and the fracture *
Corresponding author. E-mail address: hartikainen@fact-kunststoffe.de (J. Hartikainen).
0266-3538/$ - see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.compscitech.2004.07.010
toughness. However, long glass fibres (LGF) are used more and more often as a reinforcement, since it is known that longer fibres with the same fibre diameter (i.e. with higher fibre aspect ratio) provide higher stiffness, tensile strength and toughness compared to shorter ones [3–6]. Recently, it has been observed that by incorporating filler particles into the matrix of fibre reinforced composites, synergistic effects may be achieved in the form of higher modulus and reduced material costs, yet accompanied with decreased strength and impact toughness [2,7]. Such multicomponent hybrid composites based on poly(butylene terephtalate) (PBT) [8], acrylonitrile–butadiene–styrene co-polymer (ABS) [9,10] and PP [7,11] matrices have been studied.
258
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
However, hybrid reinforced composites form a complex system and there is inadequate data available about phenomena behind the property changes due to the addition of particulate fillers to the fibre reinforced thermoplastic composites. Thus, the present paper is an attempt to clarify the mechanical behaviour of hybrid composites based on the PP matrix, LGF reinforcement and calcium carbonate particulate filler, by a systematic study on flow-induced fibre alignment, fibre length degradation, matrix crystallinity and fracture mechanisms. Fibre orientation produced by injection moulding of plate shaped sample specimens was studied using a purpose built image analysis facility [12,13], to establish whether the addition of the calcium carbonate affects the orientation of the moulded components. The average fibre length in the moulded components was also determined, again to assess any effect of the filler on glass fibre properties. Tensile properties, as well as fracture toughness accompanied with an acoustic emission (AE) analysis, were determined to get information on failure mechanisms of filler filled, fibre reinforced and hybrid reinforced PP composites. Furthermore, morphology changes due to the particle addition were analysed by differential scanning calorimetry (DSC). The microstructure and failure modes of such hybrid composites are discussed on the basis of these experimental results.
2.2. Sample preparation Firstly, calcium carbonate concentrate was dry blended with PP to 40 wt% filler concentration and compounded by using a Haake Bucher Rheocord System 40 laboratory scale twin screw extruder. Temperature profile was set to 180–280 C and the used screw speed was 40 rpm. The extruded material was granulated into pellets with length of 11 mm. Also neat PP was processed into 11 mm pellets, in order to facilitate the proper mixing during the injection moulding and also to harmonize processing histories of the components. Next, the samples were prepared by mechanically mixing the compounded materials with LFT pellets in different ratios of fillers, glass fibres and PP (see Table 1 for sample denotion and mixing ratios). The bone shaped specimens (geometry Type 2 – Fig. 1, according to standard ISO 3167) were then injection moulded using Arburg Allrounder 320 C 600-250 injection moulding machine with maximum clamping force of 600 kN. The processing temperature was 225–270 C and the mould temperature was 50 C. Similarly, 2 mm thick plate shaped specimens (Geometry Type 1 – Fig. 1) were injection moulded by the company Werner GmbH (Austria). The used melt temperature during the moulding of the plate specimens was 220–240 C and the mould temperature was 15 C. 2.3. Fibre orientation and fibre length analysis
2. Experimental 2.1. Materials LGF reinforced PP pellets were provided by FACT GmbH (Germany). The material was of the type FACTOR PP LGF 60 N 11, a LGF reinforced PP with glass fibre content of 60 wt%. Fibre length in the pellets was the same as the pellet length of 11 mm. Calcium carbonate was used as a polyolefin based concentrate including olefin based particle coating (87 wt% filler content) and it was provided by Omya GmbH (Germany). The type of the concentrate was Omyalene 102 with a mean particle size of 2.0 lm. Neat PP was provided by Basell (homopolymer with an MFI of 45).
The system used for the fibre orientation and length characterisation (only Geometry Type 1) is an in-house image analysis facility developed at University of Leeds [12]. For fibre orientation investigations, the chosen method is one of optical reflection microscopy of polished 2D sections taken from the areas of interest of the composite. Each fibre that meets the 2D section is seen as an elliptical footprint, and measuring the ellipticity of these images allows the two polar angles, h and /, that specify the orientation of each fibre to be determined: h is the angle the fibre makes with the sectioned surface normal (1) and / is the angle the fibre makes with the two axis when projected into the 23 plane (Fig. 2(a)). Fig. 2(b) shows a typical elliptical fibre foot-
Table 1 Sample denotion and mechanical properties of LGF PP/CaCO3 hybrid composites Formulation
Fibre loading (vol%)
CaCO3 loading (vol%)
Tensile strength (MPa)
Tensile modulus (GPa)
Fracture toughness (MPa
A B C D E F
0 0 3.7 4.2 12.9 14.6
0 7.7 0 7.4 0 6.6
38a 27a 64 58 112 96
1.68 1.76 3.34 3.89 7.27 7.96
1.62 1.52 2.95 2.54 4.49 3.96
a
Yield stress.
p
m)
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
259
2.4. Differential scanning calorimetry Mettler Toledo DSC 821 device was used to study the effect of the used CaCO3 filler on crystallisation behaviour of PP. The samples were first heated to 190 C and then cooled back to room temperature to equalize thermal histories. Crystallisation peak temperature Tc was determined from the second cooling scan. The rate of the heating and cooling was 10 C/min. 2.5. Mechanical properties
Fig. 1. Schematic drawing describing sample geometries.
A Zwick UPM 1485 universal testing apparatus equipped with 250 kN load cell, a mechanical clamping device and an extensiometer was used to measure the tensile strength and fracture toughness of the samples. Tensile strength test was carried out according to the standard ISO 527 for a dumbbell shaped standard specimen by using a constant crosshead speed of 2 mm/min. Fracture properties were analysed by using single edge notched tensile (SEN-T) specimens with a notch depth of 2 mm. In fracture mechanical test, a constant crosshead speed of 1 mm/min was used. Critical stress intensity factor Kc was determined according to the following equation [14]: pffiffiffi K c ¼ r c Y a; ð1Þ where rc is the gross fracture stress, a is the length of the notch and Y is the geometrical factor. The following equation was used for geometrical factor Y [15]: pffiffiffi 5 p Y ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ; ð2Þ 20 13ða=W Þ 7ða=W Þ2 where W is the specimen width. Although the formulas (1) and (2) strictly hold for homogeneous and isotropic materials, they are often adopted for inhomogeneous, non-isotropic materials, including discontinuous fibre reinforced injection moulded composites [16,17].
Fig. 2. Definition of the orientation angles h and /.
2.6. Acoustic emission print, and indicates that h is given by the inverse cosine of the ratio of the semi-minor to semi-major axis of each elliptical image and / by the angle between the major axis and the two axis. An XY stage allows a large area to be scanned, allowing a determination of fibre orientation over a large area. For further details of the system and the analysis routines see [13]. Fibre length characterisation is achieved by taking a chosen section of the sample, burning off the PP matrix in a furnace for 8 h at 450 C, and then spreading the resulting glass fibres thinly on a glass plate. In this instance the fibres are viewed in transmitted light and image analysis routines are used to determine the length of each fibre shadow image.
In order to get a deeper understanding of the failure behaviour, AE of the notched dumbbell specimens was recorded in situ (during the loading), by using a miniature sensor (10 mm diameter) attached to the specimen surface and coupled to a Defektophone NEZ 220 device (AEKI, Hungary). The acoustic events were picked up by a wide bandwidth heat-proof transducer in the frequency range 100–6001 kHz (peak sensitivity ±70 dB/ V/lbar, type Micro-30D of Dunegan Co., USA). The output signal of the transducer was amplified logarithmically. The transfer function for the whole measuring system (including the logarithmic amplifier and acquisition unit) was
260
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
Peak amplitude ¼ 100 þ 20 lgðU inp =0:4Þ ½dB referred to 1V:
ð3Þ
During the tests the following primary AE signals were measured, calculated and stored: elapsed time, number of events, peak amplitude, AE energy, event width and rise time [18]. 2.7. Scanning electron microscopy The fracture surfaces after LEFM test were analysed using JEOL JSM 5400 scanning electron microscope. Before the analysis, the fracture surfaces were first cautiously cleaned with air blow and then coated with a thin layer of gold–palladium alloy (ratio 5/1).
3. Results and discussion 3.1. Fibre orientation and length analysis (Geometry Type 1 – Fig. 1) The main aim of this section of the study was to assess the effect of the addition of the CaCO3 filler on the orientation and length of the glass fibres. This was carried out using the Type 1 injection moulded plates (Fig. 1), and using samples E (LGF/CaCO3 = 12.9/0 vol%) and F (LGF/CaCO3 = 14.6/6.6 vol%). 3.1.1. Fibre orientation studies Fibre orientation in moulded components was compared at two areas, A and B, as shown in Fig. 1. Area A was located in the centre of the hexagon shaped section of the component. Here three sections were investigated, all through thickness sections (XZ plane). Each section was 30 mm long and 15 mm apart from each others. Fig. 3 shows a typical set of image analysis results, in this case for the section nearest the arrow on Fig. 1. During scanning, h and / of each individual fibre image recognised are stored so that after scanning it is possible to reconstruct the scanned area. Fig. 3(a) shows a reconstruction of half of the scanned section (2 mm thick and 15 mm long). In common with the majority of injection moulded plates, the sample shows the location of central core region where the fibres are aligned perpendicular to the injection direction and outer shell regions where the fibres are more preferentially aligned along the injection (X) direction. An instructive way to display this data quantitatively is to calculate values of the second order orientation averages hcos2hXi, hcos2hYi and hcos2hZi in strips across the sample: the higher the value of each of these averages, the greater the orientation along that axis and in addition the three averages always add up to 1. Fig. 3(b) shows values of these parameters by dividing the sample into strips across the thickness direction. Here we see clearly the outer
Fig. 3. Fibre orientation at scanning area A: (a) typical reconstituted image scan, (b) typical values of the second order orientation average hcos2hi i across the sample thickness and (c) a comparison of the histogram of h for samples E (LGF/CaCO3 = 12.9/0 vol%) and F (LGF/CaCO3 = 14.6/6.6 vol%).
shell layers and central core layers: the core occupies approximately 20% of the sample thickness. Interestingly the maximum level of orientation in the two layers is roughly similar albeit with respect to different axes. Finally we can see that the level of out-of-plane orientation (Z axis) is very low, which would be expected for a 2 mm thick plate and the long fibres used here. For the sample material E (no CaCO3) analysis of the three sections showed very little difference, suggesting that the fibre orientation was homogeneous over a 30 mm square region at the centre of the hexagon shaped section. A similar analysis for sample F (+CaCO3) showed the same good agreement. Fig. 3(c) therefore shows a comparison of the fibre orientation at the same section position (again the section nearest the arrow on Fig. 1) for the two samples. In this case we have presented the data as a histogram of the angle h, but any other comparison technique would give the same result,
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
namely that the fibre orientation state for the two sample formulations is not identical but very similar. A similar comparison was then carried out at the second scanning area, B. This section was again taken through the sample thickness (XZ plane). The section was taken 4 mm behind the hole, and it was 35 mm in width and centrally located. Fig. 4(a) shows a comparison of the fibre orientation for the two samples E and F. In this instance we have displayed the data as an area map of the second order orientation average hcos2hXi: here a white colour denotes high orientation along the X axis. It is seen that qualitatively, the addition of the CaCO3 filler has little effect of the fibre orientation at this position. Again we have used a strip analysis to produce quantitative data, this time with strips parallel to the Z axis and moving along the sample width. Fig. 4(b) (for the sample Type E) shows this analysis. It can be seen that there is a region (which is directly located behind the hole in this sample) where the fibres
261
are highly aligned in the flow direction. Fig. 3(c) compares the strip analyses for hcos2hXi for the two samples, and it seen that both the shape and quantitative values are very similar, confirming that the incorporation of the CaCO3 filler has little effect of the orientation of the glass fibres. 3.1.2. Fibre length studies Fibre length studies were carried out on pieces taken from the narrow central section of the two sample Types (E and F). As described above, the samples were placed into a furnace to burn away the PP matrix and then placed onto a glass plate for analysis. Fig. 5 shows the length distributions obtained from this analysis (1300 fibres) for each sample. The calculated number average lengths for these two distributions were as follows: sample E (LGF/CaCO3 = 12.9/0 vol%) – 887 lm: sample F (LGF/CaCO3 = 14.6/6.6 vol%) – 648 lm. It appears that while the CaCO3 filler does not significantly affect the fibre orientation state, it does lead to a reduction in the average fibre length. 3.2. Crystallinity changes The most prominent effect of particulate fillers on the crystalline structure of semi-crystalline thermoplastics is their ability to work as a nucleation agent [2]. Though there has been some debate over the effects of crystalline changes on macromechanical properties, it seems clear that by increasing crystallinity, the modulus of PP increases and the strength and deformability decrease [2,19]. However, the nucleation effect differs strongly by filler type, particle size (i.e. surface area) and filler surface treatment. For example, it has been observed that talc is an active filler, affecting strongly on the crystalline morphology of PP [20]. Calcium carbonate, instead, has typically little effect on crystallinity and can be considered as an inactive filler in sense of nucleation effect [20–22]. By modifying the surface chemistry of the
Fig. 4. Fibre orientation at scanning area B: (a) area maps of the second order orientation average Æcos2hX æ for the two materials and (b) strip analysis along the scanned width of sample E and (c) a comparison of the strip analysis for the two materials.
Fig. 5. Fibre length distribution in plate shaped specimen comparing sample E (LGF/CaCO3 = 12.9/0 vol%) and sample F (LGF/ CaCO3 = 14.6/6.6 vol%).
262
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
filler, nucleation effect may be unaffected, increased or decreased [21–24]. On the other hand, smaller particle size has been observed to increase the activity of CaCO3 as a nucleation agent, possibly due to aggregate formation [21]. In this work the effect of the used filler on crystallinity of PP was analysed by DSC. Fig. 6 depicts DSC thermograms at 100–150 C region of the second cooling cycle for pure PP (sample A) and PP including 7.7 vol% of calcium carbonate (sample B). Crystallisation exotherm appears as a single, distinct peak in both cases and the Tc temperature is almost unchanged (125 C for neat PP and 123 C for filled PP). Thus, it seems that the filler used in our study is inactive in sense of nucleation effect, probably due to the facts that (1) the average particle size of the used filler was relatively high (2 lm), and (2) the filler particles were surface treated to decrease the surface energy, which both are known to reduce the ‘‘activity’’ of calcium carbonate.
Fig. 7. Effect of filler addition to the shape of tensile stress–strain curves at different LGF/CaCO3 ratios (vol%): 3.7/0 (sample C), 4.2/7.4 (sample D), 12.9/0 (sample E) and 14.6/6.6 (sample F).
3.3. Tensile properties Tensile stress–strain curves of LGF PP/CaCO3 composites at different levels of fibre and filler contents are shown in Figs. 7 and 8 which present the averaged tensile strength and tensile modulus values. The values are also collected in Table 1. It can be seen that both the strength and modulus are distinctly increased as the LGF content increases. On the other hand, the elongation at break is simultaneously reduced, which can be attributed to the fact that reinforcing fibres strongly restrain the deformation of the matrix polymer as demonstrated in several previous studies [10,25]. The shape of the curves remains essentially unchanged as calcium carbonate filler is added to the matrix, but both the maximum strength and strain decrease. However, the modulus values (Fig. 8(b)) show that by filler addition the stiffness of the LGF composite may be further en-
Fig. 8. Tensile strength and tensile modulus of LGF PP/CaCO3 composites. In the cases of samples A and B yield stress is given.
Fig. 6. DSC scans (second cooling cycle) showing the crystallinity exotherms for neat PP (sample A) and PP with 7.7 vol% of CaCO3 (sample B).
hanced. In the case of calcium carbonate filled PP (sample B) it can be noticed that the modulus increases about 5% upon filler addition, which is in accordance with the values reported elsewhere [20,26]. On the other hand, in the cases of LGF reinforced composites the modulus increase is more distinct i.e. of the order of 10–15%. Thus, it seems that a synergy effect takes place by incorporating particulate filler into the matrix, leading to a higher stiffness than would be expected on the basis of change of the matrix modulus. Note that the amount of filler in the matrix was kept constant (7.7 vol%, or 20 wt%)
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
in all samples, and that the crystallinity effects can be excluded, as was evidenced by the DSC analysis. 3.4. Fracture behaviour Linear elastic fracture mechanics offers a useful way of getting information about different fracture mechanisms which take place during the loading of composite specimens, especially if the mechanical testing is accompanied by other methods like AE or optical microscopy [3,27–29]. In the present work we used an LEFM approach to compare fracture behaviour of the LGF reinforced PP composites to that of the filler filled PP and LGF PP/CaCO3 hybrid composites. The fracture toughness values obtained are shown in a diagram of Fig. 9. It can be observed that the critical stress intensity factor Kc increases considerably when the materials are reinforced with LGF. As calcium carbonate is added to the matrix, Kc remains almost unchanged in the case of unreinforced PP, but decreases in the case of hybrid composites. It is widely accepted that the fracture toughness of PP can be increased by reinforcing with LGFs [1]. On the other hand, it has been reported that filler inclusions may lead to the decreased fracture toughness when added to thermoplastics like PP, especially if large particle agglomerates are present. For example, it has been observed that the fracture toughness of PP decreases sharply above a 5 vol% addition of calcium carbonate decreasing to half of the original value at 30 vol% filler content [30]. In this study, the amount of calcium carbonate was only 7.7 vol% of the matrix, and particle agglomeration was minimised by surface treatment of the filler and by the separate compounding step. Consequently, the fracture toughness of neat PP was almost unaffected by the filler addition. However, in hybrid composites Kc values decreased, which suggests that there are effects that cannot be explained purely by the filler-matrix related mechanisms. In order to get a deeper understanding on the failure mode and sequence, an AE analysis was carried out during the
Fig. 9. Fracture toughness of LGF PP/CaCO3 composites.
263
fracture mechanical test. In the case of unreinforced PP, the main mode of energy absorption during the fracture mechanical test is matrix deformation. However, as PP is reinforced with LGFs, several additional failure mechanisms are observed to occur in the fracture process; such as fibre debonding, fibre breakage, and pull-out of fibres and fibre bundles [3]. These different failure mechanisms can be distinguished by AE, because they give acoustic signals at certain characteristic amplitudes. Based on numerous AE studies performed with the same experimental setup on discontinuous fibre-reinforced composites, the following correlation exists between the AE amplitude ranges and individual failure events [1,31]: fibre/matrix debonding (<35 dB) <fibre pull-out (35–55 dB) <fibre fracture (>55 dB). The cumulative numbers of the AE events registered during loading of SEN-T specimens are summarized in Table 2. The loading curve was sectioned as follows: range I ends at 40% of Fmax, range II extends from 40% of Fmax to 80% of Fmax, range III extends from 80% of Fmax to Fmax, and range IV covers the section from Fmax downwards (Fig. 10). A notable feature in the values of Table 2 is that the neat PP and its CaCO3 modified version were acoustically much less active than LGF reinforced composites. Furthermore, the signals are of low amplitude, i.e. close to the threshold level (10 dB), which reflect the matrix related failure mechanisms [3]. The related signals may arise from matrix deformation (neat PP) and particle/matrix debonding (calcium carbonate filled PP). As the LGF content increases, so does the cumulative number of AE signals. The AE amplitude distribution remains practically the same in samples including LGF reinforcement, in each loading section (I. . .IV), independent of fibre content. This means that the same type of failure mechanism occurs in the respective loading sections. In previous studies on fracture behaviour of LGF PP composites it has been observed that crack initiation takes place before the maximum load is achieved [3]. Similar conclusion was reached using AE in the present study. Fig. 11 shows a distribution of the amplitudes at different stages of force-elongation curve of hybrid composite including 14.6 vol% of LGF and 6.6 vol% of calcium carbonate. It can be seen that there is some AE activity already during the stage I of failure sequence. This indicates that some fibre debonding and related matrix deformation may take place at early stage of the loading. The activity increases considerably at stage II, indicating the start of the crack propagation. At stage II higher amplitude signals are emitted, which suggests that fibre pullouts take place. In stage III fibre debonding and pullouts are even more evident, and additionally fracture of fibres occurs as suggested by acoustic signals emitted with amplitudes over 55 dB. Finally the activity decreases as the fracture terminates during the last stage IV. Thus, it can be
264
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
Table 2 Relative distribution of amplitude during the AE analysis Relative distribution of amplitude (%)
Formulation A: PP/LGF/CaCO3 100/0/0 vol%
Formulation B: PP/LGF/CaCO3 92.3/0/7.7 vol%
Formulation C: PP/LGF/CaCO3 96.3/3.7/0 vol%
Formulation D: PP/LGF/CaCO3 88.4/4.2/7.4 vol%
Formulation E: PP/LGF/CaCO3 87.1/12.9/0 vol%
Formulation F: PP/LGF/CaCO3 78.8/14.6/6.6 vol%
Amplitude interval (dB) 31–40
Number of signals
11–20
21–30
41–50
51–60
61–70
71–80
81–90
I II III IV
100.0 87.5 100.0
12.5
I II III IV
100.0 100.0 69.7
21.2
9.1
I II III IV
72.0 31.9 24.2 62.1
12.0 15.9 10.0 16.7
16.0 11.6 11.1 1.5
15.9 30.8 1.5
22.0 22.5 10.6
2.7 1.4 7.6
1097
I II III IV
70.9 73.8 7.4 16.5
21.8 17.3 6.2 6.4
7.3 3.1 19.3 13.8
2.7 56.0 52.1
2.5 11.1 11.2
0.6
833
I II III IV
82.8 73.8 51.3 38.3
14.8 17.3 12.5 25.2
2.4 3.1 6.3 13.6
2.7 12.6 9.9
2.5 14.4 10.1
0.6 2.8 2.9
I II III IV
76.5 64.2 28.3 41.3
18.5 17.2 11.5 31.9
5.0 4.9 12.2 15.9
5.9 20.0 7.2
5.8 21.1 2.2
1.7 6.6 1.5
11
36
2592 0.1
0.3 0.3
1717
Sectioning of the force–elongation curve: I, 0–0.4 Fmax; II, 0.4–0.8 Fmax; III, 0.8–1 Fmax; IV, 1–0 Fmax.
Fig. 10. Load–displacement curve and cumulative AE signals of the SEN-T specimen (LGF/ CaCO3 = 12.9/0). Sectioning of the load– displacement curve is shown with Roman numerals I–IV.
concluded that fibre related mechanisms dominate the fracture behaviour of studied composites. It can be noted from the values in Table 2 that the addition of CaCO3 to LGF reinforced composites results in a pronounced decrease in AE activity. Secondly, there are interesting changes in amplitude distribution upon addition of filler, especially evident at lower
Fig. 11. Distribution of CaCO3 = 14.6/6.6 vol%).
amplitude
in
AE
analysis
(LGF/
LGF contents. Firstly, the amplitude distribution at stages I and II indicates more frequent matrix deformation and fibre debonding related fracture incidents as fil-
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
265
ler is added to the LGF PP composite, compared to samples without a filler. In the case of sample D (PP including 4.2 vol% of LGF and 7.4 vol% of calcium carbonate), more than 90% of the signals at stage II are emitted at amplitudes below 30 dB. However, in the case of sample C (PP with 3.7 vol% of LGF) these were less than 50%. This means that filler particles increase the incidents of matrix deformation and fibre debonding [1,31]. Secondly, at later phases of the failure (stages III and IV) the majority of the incidents, in the case of hybrid samples, take place at amplitudes that suggest fibre debonding. Thus, it seems that filler particles induce particle debonding, which may explain the decreased mechanical properties like strength and fracture toughness. This view has already been suggested in other studies of SGF reinforced ABS/CaCO3 [10] and glass mat reinforced PP/mica [11] hybrid composites. These conclude that fillers may cause premature fibre debonding under loading, due to void formation around particles near fibres, or due to direct contact between particles
and fibres. It should be noted that the increase of low amplitude signals at stages I and II in hybrid samples cannot be explained only by filler related events, since the number of AE arising from matrix deformation was only slightly increased as calcium carbonate was added to neat PP. Thus, it can be concluded that LGF-related failure events dominate also in hybrid systems, but the number of the acoustic signals is less than in the case of LGF composites without a filler. This can be explained by the fact that pullouts occur more easily due to reduced fibre–matrix adhesion (originating from the strong debonding at early stage of failure sequence). Finally, the fracture surfaces after fracture mechanical testing were studied by scanning electron microscopy (SEM). In the case of LGF PP composite with 12.9 vol% of fibres (Fig. 12(a)) there is clear evidence of ductile matrix deformation, which is a result of the fibre debonding and pullout processes. On the other hand, a more irregular fracture surface is observed in the case of corresponding hybrid sample
Fig. 12. SEM micrographs of fracture surfaces after fracture mechanical test for PP hybrid composites at 350 times magnification: (a) LGF/ CaCO3 = 12.9/0 vol% and (b) LGF/CaCO3 = 14.6/6.6 vol%. Crack growth direction is from left to right in both cases.
Fig. 13. SEM micrographs of fracture surfaces after fracture mechanical test for PP hybrid composites: (a) LGF/CaCO3 = 12.9/0 vol%, 750· magnification and (b) LGF/CaCO3 = 14.6/6.6 vol%, 1500· magnification. Crack growth direction is from left to right.
266
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267
(Fig. 12(b)). In addition, there are more marks of pullout events as indicated by the high amount of voids originating from the disconnected fibres. Fig. 13 shows SEM micrographs at higher magnification, demonstrating the difference in matrix deformation of LGF reinforced composites and hybrid materials. As calcium carbonate is added to the matrix, void formation around the filler particles take place during the loading. Fig. 13(b) presents the fracture surface at a site where fibre has been pulled out. It can be seen that the particles and the voids may come into contact with fibres and thus, induce early stage fibre debonding. Furthermore, it can be noted from Figs. 12 and 13 that ductility is reduced in filler filled samples as evidenced by the smaller deformation patterns compared to unfilled LGF composites. It can be concluded that this ductility reduction of the matrix also contributes to the decrease of fracture toughness when calcium carbonate is added to LGF PP. These SEM observations are in good agreement with the mechanical test results and AE analysis.
4. Conclusions The present work showed that the properties of LGF PP composites can be modified by adding calcium carbonate. It was shown that the addition of CaCO3 in LGF PP increased the stiffness. On the other hand, the strength and toughness were decreased. Various analysis methods were used to investigate the mechanisms behind the property changes. Optical image analysis for the plate shaped specimens showed that the addition of CaCO3 did not considerably affect the fibre orientation. However, the average length of the fibres was 25% lower in the case of hybrid composites compared with LGF PP. The fracture mechanical test accompanied with AE analysis was used to study the fracture modes. It was shown that the fracture toughness of unreinforced PP remained unaffected as calcium carbonate was added. However, the fracture toughness of LGF reinforced PP decreased upon filler addition. The AE of the samples showed that the acoustic activity of LGF PP/ CaCO3 hybrid samples was much higher compared to LGF PP without a filler. Secondly, the distribution of the amplitudes suggested that filler addition results in debonding of the fibres at early stage of the loading. This results in fibre pull-out mechanism at earlier phase of the fracture sequence compared to unfilled samples. Furthermore, SEM analysis for the fracture surfaces showed decreased ductility in the case of filler filled hybrid materials. This may have a contributory influence on decrease of fracture toughness values upon CaCO3 addition.
Acknowledgements HTP Fohnsdorf GmbH (Austria) is acknowledged for the permission to use the mould of the plate shaped specimens. Prof. J. Karger-Kocsis is gratefully acknowledged for the numerous discussions and support. J. Hartikainen wishes to thank the European Commission for financial support (Marie Curie Fellowships program, contract number G5TR-CT2001-00052).
References [1] Karger-Kocsis J. Microstructural aspects of fracture in polypropylene and in its filled, chopped fiber and fiber mat reinforced composites. In: Karger-Kocsis J, editor. Polypropylene: Structure, blends and composites. London: Chapman & Hall; 1995. p. 142–201. [2] Pukanszky B. Particulate filled polypropylene: structure and properties. In: Karger-Kocsis J, editor. Polypropylene: Structure, blends and composites. London: Chapman & Hall; 1995. p. 1–70. [3] Karger-Kocsis J, Harmia T, Cziga´ny T. Comparison of the fracture and failure behavior of polypropylene composites reinforced by long glass fibers and by glass mats. Compos Sci Technol 1995;54(3):287–98. [4] Karger-Kocsis J. Swirl mat- and long discontinuous fiber matreinforced polypropylene composites-status and future trends. Polym Compos 2000;21:514–22. [5] Stokes VK, Inzinna LP, Liang EW, Trantina GG, Woods JT. A phenomenological study of the mechanical properties of longfiber filled injection-molded thermoplastic composites. Polym Compos 2000;21(5):696–710. [6] Thomason JL. The influence of fibre length and concentration on the properties of glass fibre reinforced polypropylene: 5. Injection moulded long and short fibre PP. Composites: Part A 2002;33:1641–52. [7] Acosta JL, Morales E, Ojeda MC, Linares A. Effect of addition of sepiolite on the mechanical properties of glass fiber reinforced polypropylene. Angew Makromol Chem 1986;138:103–10. [8] Hargarter N, Friedrich K, Catsman P. Mechanical properties of glass fiber/talc/polybutylene-terephtalate composites as processed by the radlite technique. Compos Sci Technol 1993;46:229–44. [9] Fu S-Y, Lauke B. Fracture resistance of unfilled and calciteparticle-filled ABS composites reinforced by short glass fibers (SGF) under impact load. Composites: Part A 1998;29:631–41. [10] Fu S-Y, Lauke B. Characterization of tensile behaviour of hybrid short glass fibre/calcite particle/ABS composites. Composites: Part A 1998;29:575–83. [11] Zhao R, Huang J, Sun B, Dai G. Study of the mechanical properties of mica-filled polypropylene-based GMT composite. J Appl Polym Sci 2001;82:2719–28. [12] Hine PJ, Davidson N, Duckett RA, Ward IM. Measuring the fibre orientation and modelling the elastic properties of injectionmoulded long-glass-fibre-reinforced Nylon. Compos Sci Technol 1995;53:125–31. [13] Hine PJ, Davidson N, Duckett RA, Clarke AR, Ward IM. Hydrostatically extruded glass-fiber-reinforced polyoxymethylene. I: The development of fiber and matrix orientation. Polym Compos 1996;17:720–9. [14] Anderson TL. Fracture mechanics – fundamentals and applications. Boca Raton: CRC Press Inc.; 1995. [15] Harris DO. Stress intensity factors for hollow circumferentially notched round bars. J Basic Eng 1967;89(1):49–54.
J. Hartikainen et al. / Composites Science and Technology 65 (2005) 257–267 [16] Hashemi S, Koohgilani M. Fracture toughness of injection molded glass fiber reinforced polypropylene. Polym Eng Sci 1995;35(13):1124–32. [17] Pegoretti A, Ricco T. Fatigue crack propagation in polypropylene reinforced with short glass fibres. Compos Sci Technol 1999;59:1055–62. [18] Szabo´ JS, Cziga´ny T. Investigation of static and dynamic fracture toughness on short ceramic fiber reinforced polypropylene composites. J Macromol Sci, Phys B 2002;B41:1191–204. [19] Wright DGM, Dunk R, Bouvart D, Autran M. The effect of crystallinity on the properties of injection moulded polypropylene and polyacetal. Polymer 1988;29(5):793–6. [20] Pukanszky B, Belina K, Rockenbauer A, Maurer FHJ. Effect of nucleation, filler anisotropy and orientation on the properties of PP composites. Composites 1994;25(3):205–14. [21] Kowalewski T, Galeski A. Influence of chalk and its surface treatment on crystallization of filled polypropylene. J Appl Polym Sci 1986;32:2919–34. [22] Rybnika´r F. Interactions in the system isotactic polypropylenecalcite. J Appl Polym Sci 1991;42(10):2727–37. [23] Garton A, Kim SW, Wiles DM. Modification of the interface morphology in mica-reinforced polypropylene. J Polym Sci, Polym Lett Ed 1982;20:273–8. [24] Zuiderduin WCJ, Westzaan C, Huetink J, Gaymans RJ. Toughening of polypropylene with calcium carbonate particles. Polymer 2003;44:261–75.
267
[25] Thomason JL, Vlug MA, Schipper G, Krikor HGLT. Influence of fibre length and concentration on the properties of glass fibrereinforced polypropylene: Part 3. Strength and strain at failure. Composites: Part A 1996;27:1075–84. [26] Demjen Z, Pukanszky B, Nagy J. Evaluation of interfacial interaction in polypropylene/surface treated CaCO3 composites. Composites: Part A 1998;29:323–9. [27] Cziga´ny T, Karger-Kocsis J. Comparison of the failure mode in short and long glass fiber-reinforced injection-molded polypropylene composites by acoustic emission. Polym Bull 1993;31: 495–501. [28] Cziga´ny T, Mohd Ishak ZA, Karger-Kocsis J. On the failure mode in dry and hygrothermally aged short fiber-reinforced injection-molded polyarylamide composites by acoustic emission. Appl Compos Mater 1995;2:313–26. [29] Cziga´ny T, Marosfalvi J, Karger-Kocsis J. An acoustic emission study of the temperature-dependent fracture behavior of polypropylene composites reinforced by continuous and discontinuous fiber mats. Compos Sci Technol 2000;60(8):1203–12. [30] Levita G, Marchetti A, Lazzeri A. Fracture of ultrafine calcium carbonate/polypropylene composites. Polym Compos 1989;10: 39–43. [31] Benevolenski O, Karger-Kocsis J. Fracture and failure behavior of partially consolidated discontinuous glass fiber mat-reinforced polypropylene composites (Azdel SuperLite). Macromol Symp 2001;170:165–79.