- Email: [email protected]
metal composites subjected to cold rolling
Composites Science and Technology 72 (2012) 1344–1351
Contents lists available at SciVerse ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Deformation and fracture of polymer/metal composites subjected to cold rolling B.O. Calcagno a,⇑, K.R. Hart b, J.C. Springmann b, G.G. Antoun b, W.C. Crone a,c a
Materials Science Program, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, United States Engineering Mechanics Program, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, United States c Department of Engineering Physics, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, United States b
a r t i c l e
i n f o
Article history: Received 22 December 2011 Accepted 3 May 2012 Available online 22 May 2012 Keywords: A. Polymer–matrix composites (PMCs) B. Fracture C. Stress transfer C. Sandwich structures Cold rolling
a b s t r a c t Cold rolling of a sandwich composite with a metallic strip inclusion in a polymeric matrix can produce a range of outcomes, including deformation and fracture of the inclusion. Using different material combinations under the same processing parameters, the results ranged from minor deformation, to folding and/or loss of adhesion, and fracture of the inclusion, with fracture particles varying in size and shape. Comparisons are made between the resulting structures after cold rolling of the polymer/metal composite to geological formations. In particular, the fracture particles obtained resemble rock structures known as boudins. The phenomena of boudinage and folding encountered in the cold rolling of polymer/metal composites is similar to that seen in geology formations although the time and size scale of these events are several orders of magnitude apart. The experimental results reported show that cold rolling applied to a polymer/metal sandwiched composite induces deformation and fracture behaviors that depend on the mechanical properties of the constituents, deformation behavior of the polymeric matrices, interfacial adhesion, and process parameters such as rollers speed and nip-gap. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Metal inclusions in polymeric composite materials under a process of cold rolling and folding can exhibit fracture under certain conditions. The deformation and fracture patterns observed are similar to those found in some rock structures [1,2]. Similarities are also found with the microstructure of laminate ceramic composites subjected to rolling and folding [3], multilayer ceramic tubes that have undergone co-extrusion [4], as well as some metal/metal laminate subjected to rolling processes [5,6]. The common factor among these processes is the deformation of two contiguous phases with different mechanical properties, which could produce necking and/or fracture of one of the composite constituents. Our interest in this phenomena began with the use of cold rolling (also called calendering) for the purposes of grain refinement in shape memory alloys [5,7], later it was expanded to cold rolling of multilayer materials for the purposes of particle production [8]. This research, as well as that of others [9,10], has shown that a particle disperse composite can be produced through plastic deformation using cold rolling of a sandwich composite made of layers of dissimilar materials. The work presented here is focused on two
⇑ Corresponding author. Address: Department of General Engineering, University of Puerto Rico-Mayagüez, Call Box 9000, Mayagüez, PR 00681, United States. Tel.: +1 787 832 4040x3069; fax: +1 787 265 3816. E-mail address: [email protected] (B.O. Calcagno). 0266-3538/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compscitech.2012.05.004
layers of polymer and metal or alloy strips between them as the initial condition. In the experimental work reported below, sandwich composites were made using polypropylene (PP), polycarbonate (PC), or highdensity polyethylene (HDPE) as the matrix phase, and strips of pseudoelastic nickel titanium alloy or titanium as the inclusion phase. A sampling of the various fracture and folding patterns obtained for the different composites is shown in Fig. 1. The fracture particles varied in size, shape, and in some cases, the inclusion did not break but rather folding and/or loss of adhesion was observed. The range of materials investigated provides an interesting window into the phenomenon of cold rolling. 2. Background Under certain conditions the fracture patterns observed after cold rolling a metal inclusion in a polymeric matrix sandwich resemble fragments of rock layers called boudins. There is substantial literature on these formations. These geological structures occur when a competent rock layer surrounded by a less competent rock matrix is subjected to layer-parallel extension or layer-normal compression. The term competent is used by geologists to identify the stronger or stiffer layer, while the less competent or incompetent layer is the weaker or ductile layer [11–19]. Pollard and Fletcher [20] point out that these terms ‘‘do not refer to an explicit type of material behavior.’’ In those studies where the layers of rocks have been considered as viscous fluids, competency has been
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
1345
Fig. 1. Inclusion fracture patterns obtained after metal/polymer composite fabrication process. (a) NiTi particles in PP matrix after 1 rolling cycle; particle size ranges from 0.1 to 0.5 cm. (b) NiTi strips in HDPE matrix; no break after 1 rolling cycle. (c) Ti strips in PC matrix; no break after 1 pass; the folds measured approximately 0.18 cm. (d) Ti particles in HDPE after 1 rolling cycle; particles average size was 0.4 cm. (One rolling cycle = several passes through the rollers reducing the nip-gap 0.05 cm each pass until a nip-gap of 0.076 cm was reached).
more rigorously defined as a function of the viscosity of the layer, where a more viscous layer is the competent layer and the layer with a lower viscosity or ability to flow represents the incompetent layer [21–24]. When layer-parallel extension or layer-normal compression are applied to the competent layer, flow in the less competent matrix parallel to the layer will occur, generating tensile stresses, extension, and finally fracture of the layer [11,25]. The shape of the boudin is determined by the properties of the layer, where strong brittle layers would produce rectangular boudins with sharp corners as those shown in Fig. 2a [26]. More ductile layers would allow plastic flowage and lateral elongation producing structures termed pinch-and-swell which result from necking. These structures may eventually fracture producing more oblong or lensshape called lenticular boudins, or barrel-shape boudins [11,14]. Fig. 2b shows a barrel-shaped boudin with concave end that is believed to be the result of high plastic deformation imposed on the
originally rectangular boudin. This kind of rock structure is also known as ‘‘fish-mouth boudins’’ [23]. Other rock structures known as folds occur when competent viscous layers in a less stiff matrix buckle when subjected to layer-parallel compression [15]. Structures such as the pinch-and-swell, and folding were explored by Smith’s theory on formations of folds, boudinage, and mullions in Newtonian and non-Newtonian materials [21,22]. Using a general approach, Smith hypothesized that folding and boudinage are related structures that occur when a single layer is squeezed between two thick layers of different viscosity, establishing a state of motion with unstable small disturbances. Fig. 3 portrays a summary of the layer instabilities, where the stronger layer is depicted as the darker region, and the more ductile as the lighter region. The disturbances could grow into flow, fracture boudinage, folding, or mullions. The parallels between the boudin and fold structures and the development of fracture and folding in a metal-alloy/polymer
Fig. 2. Images of typical boudin fracture pattern in rock structures. (a) Rectangular boudin: Tourmaline grains in a quartz matrix from Western Australia [26]. (b) Shapedbarrel boudin with ‘fish-mouth’ shape ends: Boudin in calc-silica layer in marble, Khan Gorge, Namibia [23]. Coin shown for scale.
1346
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 3. Four classifications of dynamic instability cases in a single layer. The more competent layer is depicted in darker gray. Figure adapted from [21].
composite during cold rolling motivated a more detailed investigation. We sought to identify the roles that rolling process, material properties, interfacial adhesion, and load transfer play in the deformation and fracture of an inclusion material sandwiched between layers of dissimilar material. We discuss here the results of two of the experimental techniques used in our work, cold rolling and pullout tests.
3. Experimental methods The process investigated is called cold rolling (or calendering) of a sandwich composite. Fig. 4a shows a schematic of the rolling process. The sandwich composite is introduced to the rollers which run at a constant speed. The nip-gap (distance between the rollers) is fixed at a dimension smaller than the sandwich thickness. The sandwich composites were made using polypropylene (PP), polycarbonate (PC), or high-density polyethylene (HDPE) (McMaster, #8742K131, #8574K24, #8619K421, size 1200 1200 1/1600 ) as the matrix phase, and strips of pseudoelastic nickel titanium alloy (Nitinol Devices & Components, NiTi foil, Alloy type SE 508, pseudoelastic at room temperature, 55.6 at.% Ni, 0.0127 mm thick) or titanium (Alfa Aesar, Ti foil, 99%, 0.127 mm thick) as the inclusion phase [1,27]. The surface of the NiTi and Ti strips were etched using a solution of 1HF + 4HNO3 + 5H2O. The composite sandwich samples were made by placing strips of metal or alloy between two layers of polymers and heating them while under 5 lbs of compressive load. The molding temperature for the PP composites was 197 °C, 230 °C for the HDPE, and 290 °C for the PC.
Two types of cold rolling experiments were performed. In the repeated rolling experiments, a metal/polymer or alloy/polymer sandwich composite containing several strips of metal/alloy was subjected to cold rolling several times. With each rolling pass the nip-gap (the spacing between the rollers) was reduced in order to decrease the sandwich composite thickness and induce plastic deformation. All of the repeated rolling experiments reported here were performed at the same roller speed with a Stanat rolling mill (model TA-315). The sandwich composites were made with 8 cm 8 cm polymer squares and 2.5 cm 0.5 cm metal/alloy strips [2]. A second set of experiments to produce one fracture in the inclusion was performed to investigate the process parameters that were relevant during the cold rolling. Samples with only one metal/alloy strip between polymer layers were prepared and subjected to the minimum number of rolling passes to fracture the inclusion during their passage through the rollers [1,27]. These tests were performed with a 5471 150 New Durston rolling mill (model D2-120). The strip width was 5 mm and the length varied from 5 mm to 20 mm [2]. In order to quantify the strength of the bonding between the metal and the polymeric matrix, pullout tests on the composites were performed where a tensile force was applied to the inclusion to produce debonding between the two constituents of the sandwich composite [27,28]. A double overlapping configuration following that of Al-Sheyyab et al. [29] was used with a 5 mm wide embedded strip. Testing was conducted with an Instron 5566 using a constant crosshead speed of 5 mm/min. The embedded length of the strips ranged between 10 and 12 mm [27]. The surface of all strips was subjected to the same etching treatment applied to the cold rolling strips. Fig. 4b shows a schematic of the pullout test performed. All the experiments described were done at room temperature.
4. Results The repeated rolling experiments produced a range of behaviors depending on the material combination and rolling parameters used (e.g. roller speed, nip-gap, and sample dimensions). Under the same rolling conditions, NiTi in PP fractured producing particles with rectangular corners (see Figs. 1a and 5), similar to the behavior exhibited by Ti in PP. In contrast, NiTi in HDPE did not fracture (Fig. 1b), and Ti in a HDPE produced particles with distorted fracture surfaces (Fig. 1d). Under other conditions, fracture was not observed, and instead delamination of the composite constituents occurred and the strip acquired out-of-plane
Fig. 4. (a) Schematic of the cold rolling test for one pass. (b) Lateral and frontal view of a pullout sample.
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 5. Image of NiTi strip in PP matrix. The NiTi fractured producing two particles with rectangular and sharp corners after four rolling passes at a roller speed of 11.2 rpm. The original sample thickness was 1.55 mm and it was rolled at the following nip gaps: 1.5, 1.0, 0.8, and 0.6 mm. The strip was 20.17 mm long, and 5.65 mm wide; the thickness varied between 0.11 and 0.12 mm. Ruler in millimeter scale.
deformations. For example, for Ti in PC after only one rolling pass the Ti strip exhibited periodically repeating folds (Fig. 1c) and loss of adhesion as shown in Fig. 6. The three polymers were specifically chosen for the difference in their materials properties [30]. Under rolling conditions in the sandwich composite, the PC displayed brittle behavior, while the PP matrix deformed plastically and the HDPE matrix deformed extensively. Although the NiTi and Ti strips were prepared identically and expected to have the same surface oxide, differences in composite behavior were consistently observed between the pseudoelastic NiTi and ductile Ti. Under conditions that produce fracture for a particular composite combination, multiple rolling passes can be made to produce additional fractures either by folding the composite to double its thickness or decreasing the nip-gap. In general, fracture of the strip initiated at the strip edge and propagated perpendicular to the direction of rolling. Fractures in multiple directions can be achieved by rotating the sample as shown in Fig. 7 for NiTi in PC. Fig. 8 shows two samples subjected to ten cycles of cold rolling, folding and heating. This repeated rolling and folding produces a composite with a particle distribution. The size of particles was smaller for Ti in PP compared with the NiTi in PP for the same number of passes. The set of rolling experiments designed to initiate one to two fractures in the metal or alloy strip within the minimum number
1347
Fig. 7. Particles of NiTi in PC matrix after 1 rolling cycle, 90° rotation, and an additional rolling pass; particles measured approximately 0.2 cm.
of rolling passes provided additional information about the relationship between rolling parameters and fracture. Parameters such as rolling speed, nip-gap, sample thickness, strip length, and stretch of the composite were studied. Prior to deformation, lines were drawn on the exterior of the sample 1 cm from each end and outside the location of the metal strip ends. The distance between these lines were again measured after cold rolling in order to calculate the composite stretch, (Dllf/Dll0). The nip-gap roller separation, 2h0, was also compared to the original composite thickness, T0. The key relationship found was between the overall composite stretch (Dllf/Dll0) and the dimensionless nip-gap (2h0/T0) [1,2]. Fig. 9a shows results for one roller speed and Fig. 9b shows for all roller speeds and all composite material combinations. The results for the pull out tests are summarized in Fig. 10 [28]. The use of embedded strips (rather than wires) in the pull out testing was chosen to better emulate the sample configuration in cold rolling and also avoided issues of having the metal/alloy undergo plastic deformation or transformation prior to debonding. The tests applied load to the strips until the maximum force needed to induce failure of the polymer/strip interface was attained. After this point, only friction opposes the sliding at the interface. Adhesion strength is defined as the maximum force needed to produce debonding over the embedded strip surface area. Results were discarded for samples where torsion of the strip occurred due to incorrect placing of the sample in the grips and for samples containing voids in the matrix material that were produced during manufacture of the samples. 5. Discussion
Fig. 6. Images of a Ti/PC composite showing loss of adhesion during cold rolling. (a) Top view of entire sample. (b) Side view of central buckled regions. (c) Side view of strip end. Ruler in millimeter scale.
The cold rolling of sandwich composites with a metallic strip embedded in a polymeric matrix produces different fracture patterns for different combinations of constituents for the same processing parameters as portrayed in Fig. 1. The polymers used as matrices were two semicrystalline polymers, PP and HDPE, and one amorphous, PC. PP exhibits high dynamic loading capacity, and a higher strength and stiffness than HDPE. HDPE is a tough polymer with low strength and stiffness, and high elongation at break; also less dense than other polymers. PC is a transparent polymer with a high strength and stiffness, a great resistance to impact, and susceptible to crack under stress [31,32]. Titanium and pseudoelastic nickel titanium strips were used as inclusions. Titanium is a transition metal, of high strength
1348
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 8. Images of particle dispersed composites obtained by cold rolling and folding. (a) NiTi particles in PP matrix, (b) Ti particles in PP matrix. The original samples measured 8 cm 8 cm and contained the same number of metal strips 2.5 cm 0.5 cm, with thicknesses between 0.28 and 0.29 mm. They were subjected to 10 rolling cycles aligned with the length of the inclusion. Ruler in millimeter scale.
Fig. 9. Composite stretch (Dllf/Dllo) plotted against dimensionless nip gap (2ho/To) for samples whose strip inclusion fractured in one or two locations within three or fewer rolling passes. (a) Data for samples that fractured at a roller speed of 3.7 rpm. The line given is a logarithmic fit curve (R = 0.97). (b) Data for the six composite combinations tested and the roller speeds: 3.7, 11.2, 18.3, 22, 25, and 29 rpm [1].
and toughness, low density, nonmagnetic, and highly resistant to corrosion [33]. Pseudoelastic nickel titanium is a shape memory alloy (SMA), which can be deformed to strains up to 10% when a load is applied and recover its original shape upon unloading. It is nonmagnetic, dense, of high strength, and moderate impact resistance [34]. Both materials used as inclusions, titanium and nickel titanium alloy, oxidize to form a resistant titanium oxide layer on the surface [35,36]. Table 1 contains relevant properties of the composite constituents used in this work. When comparing the rolled composites with the rock structures, the metallic inclusions represented the competent surrounded by less competent polymeric matrices. During the cold rolling process the NiTi in PP fractured producing particles with rectangular corners as shown in Fig. 1a and Fig. 5 similar to the boudins shown in Fig. 2a; Ti in PP (not shown in this paper) exhibited the same behavior. However, a range of behaviors was observed with cold rolling of the polymer/metal combinations described here. Under the same processing parameters (roller speed, nip-gap, and sample dimensions), NiTi embedded in HDPE did not fracture (Fig. 1b), while the cold rolling of Ti in a HDPE matrix (Fig. 1d) produced particles with oblong shapes similar to the shaped-barrel
boudin of Fig. 2b. The Ti strip exhibited elongation of approximately 10%. In both of the HDPE composites the matrix exhibited extensive plastic deformation. In the case of NiTi in PC the metal strip fractured resembling rectangular boudins with minimal plastic deformation of the matrix (see Fig. 7). That was not the case of the Ti in PC, which exhibited folds that repeated periodically along the metal strip after only one rolling pass (Fig. 1c). These folds are very much like the rock fold structures that occur due to buckling of the more competent layer when subjected to layer-parallel compression. However in the experimental case the loading applied was the same for all the samples. Fig. 1c also shows loss of adhesion between the Ti and the PC matrix at both sides of the metal strips. The debonding between the two constituents starts from the strip end and increases as the folds grow as displayed in detail in Fig. 6 where the side view of the strip is portrayed. In order to identify the role of the parameters involved in the cold rolling process, a second cold rolling test was performed. The goal of this test was to get one or two fractures of the inclusion after the minimum number of passes of the composite sample through the rollers. Varying the nip-gap and the speed of the rollers it was possible to fracture all the metallic inclusions in all
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 10. Adhesion strength for pullout samples of nickel titanium (NiTi) and titanium (Ti) strips embedded within polypropylene (PP), polycarbonate (PC), or high-density polyethylene (HDPE) polymers [28].
Table 1 Properties for composite constituents showing: density, Young modulus, ultimate strength, strain at fracture, and yield strength.
a b
Constituent
q (g/cm3)
E (GPa)
rU (MPa)
ef (%)
ry (MPa)
NiTia Tia PPb PCb HDPEb
6.5 4.5 0.3 1.2 0.94
16 ± 2 23 ± 5 1.1–1.3 2.1–2.4 0.7–1.4
743 ± 11 305 ± 5 21–37 56–67 18–35
22 ± 3 35 ± 3 20–800 100–130 150
324 ± 9 270 ± 12
Mechanical properties from reference [27]. Properties from reference [31].
polymeric matrices even in cases such as the NiTi in HDPE (Fig. 1b) and Ti in PC (Fig. 1c) samples. The nip-gap needed to induce fracture depended on the sample initial thickness and the speed; and was different for all inclusion/matrix combinations. For the same initial sample thickness and roller speed, the NiTi embedded in PC matrix fractured at larger nip-gaps than in PP. In order to fracture the Ti in PP, under the same parameters, a smaller nip-gap than the NiTi in PP was needed. Fracture was also obtained at the same roller speed in the case of the Ti in PC but with extremely small nip-gap. For NiTi in HDPE, the speed of the rollers had to be increased more than six times to fracture the inclusion. In those cases where fracture did not occur, the inclusion folded. Dimensional analysis was used to find the relationship between the variables involved in the cold rolling process that induced fracture. As illustrated by Fig. 9a, the composite stretch was shown to have a logarithmic relationship with the dimensionless nip-gap, which was defined as the spacing between the rollers divided by the initial sample thickness. This relation is given by Dllf/Dll0 = 0.97–1.28 log (2h0/T0) [1,27]. It is notable that each inclusion/matrix combination is clustered in a different region of the curve. This trend is maintained when the data for all roller speeds for the six composites is plotted as shown in Fig. 9b. The transfer of the load applied by the rollers to the composite and thus to the metal strip depends on the interfacial adhesion between the two constituents. The strength of this interfacial bonding was measured using pullout tests. The results of these tests, summarized in Fig. 10, show that the adhesion strength of PC and NiTi is similar to that of PC and Ti [27,28]. However, although the preparation of these composites was the same, the cold rolling
1349
tests gave a different behavior for these two material combinations. Figs. 1c and 6, show images of a Ti/PC sample where loss of adhesion occurred between the Ti strip and the PC after an attempt to produce a particle disperse composite using cold rolling. For the same cold rolling parameters, the NiTi in PC fractured. To obtain fracture of the metal strip in Ti/PC samples with only one rolling pass, an extremely small spacing between rollers was needed as proved with the one-rolling-pass tests, producing also cracking of the matrix of PC. For the PP composites, the interfacial bonding between the polymer and the Ti strips was higher than the value obtained for the NiTi and PP. After subjecting these composites to ten cycles of cold rolling, folding and heating, a composite with smaller and more uniform particle size distribution was obtained for the Ti/PP compared with the NiTi/PP composites as shown in Fig. 8 [2]. It would seem that a stronger interfacial adhesion could provide a more efficient load transfer between the Ti strip and the PP matrix. However, the comparison of the adhesion strength values for the HDPE composites show a nearly equivalent value for the NiTi and Ti, while under the same rolling parameters the Ti in HDPE fractured but the NiTi in HDPE did not. Loss of adhesion was also observed in Ti/PP and NiTi/PP composites where fracture was easily obtained. Fig. 11 displays images of a Ti/PP sample where folding occurred and the interfacial bonding between matrix and inclusion was lost during cold rolling. When the process parameters were analyzed, it was found that the nip-gap imposed during the cold rolling was larger than the necessary nip-gap predicted by the relation obtained with onerolling-pass-test [1,27]. For a smaller nip-gap, fracture after one pass through the rollers was obtained in Ti in PP matrix as shown in Fig. 12. The 4.82 mm long and 0.09 mm thick Ti strip fractured in two locations, in one of them the segments appear overlapped while in the second they are completely separated from each other. Although NiTi has pseudoelastic capability, the main differences relevant to fracture behavior between the two inclusions types (NiTi and Ti) are the ultimate strength and elongation at break. In comparison, the ultimate strength of NiTi is significantly higher while the elongation at break for Ti is significantly greater. For the polymers, the ultimate strength of PC is higher while the elongation at break for PP is significantly greater. No obvious explanation for the behavior of these material combinations under cold rolling can be provided by a simple comparison of the mechanical
Fig. 11. Images of a Ti/PP composite (49.80 mm 36.72 mm 2.41 mm) showing loss of adhesion during cold rolling at a speed of 3.7 rpm and a nip gap of 1.91 mm. (a) Top view of entire sample. (b) Side view of strip end. (c) Side view of central buckled region.
1350
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
The phenomena of boudinage and folding encountered in the cold rolling of metal/polymer composites is observed to be similar to geology formations, although the time and size scale of these events are several orders of magnitude apart. These observations warrant further study due to their applicability to both geological problems and composite material processing. Acknowledgments
Fig. 12. Images of a Ti/PP sample showing fracture that occurred during cold rolling at a speed of 3.7 rpm and a nip gap of 0.25 mm. (a) Top view of sample. (b) Side view of overlapped fractured segments. (c) Side view of fractured segments separated by a distance.
properties of the constituents. Clearly there is a more complex interplay between the materials properties of the constituents, their adhesion behavior in the combinations explored, and the rolling parameters. The experimental results obtained in this work show that cold rolling can be used to study the phenomena of boudinage and folding using a metal strip as the competent layer sandwiched between layers of a polymer as the less competent layer. While the state of stress to which the composite is subjected when going through the rollers is complex and may differ from the layer-parallel extension or layer-normal compression cited in the geology references, it produces similar structures. Furthermore cold rolling provides a convenient and simple method to obtain different particle shapes just varying the composite constituents and the processing parameters at room temperature. 6. Conclusions In this research we have shown that the cold rolling process of a composite material consisting of a metallic inclusion sandwiched between layers of a polymeric matrix induces deformation and fracture behaviors that depend not only on the mechanical properties of the constituents, deformation behavior of the polymeric matrices, and interfacial adhesion, but also on processing parameters such as roller speed and nip-gap. In the experimental work described, composites with polypropylene (PP), polycarbonate (PC), or high-density polyethylene (HDPE) as the matrix phase, and strips of pseudoelastic nickel titanium alloy or titanium as the inclusion phase were subjected to cold rolling. Additionally, pullout tests on the composites were performed to quantify the strength of the bonding between the metal and the polymeric matrix. The cold rolling experiments aimed to investigate which process parameters and constituent properties were relevant in the inclusion fracture event. Results show that under the same rolling conditions, NiTi and Ti in PP fractured producing particles with rectangular corners, similar to boudins. In contrast, NiTi in HDPE did not fracture, and Ti in a HDPE produced particles with distorted fracture surfaces that resembled fish mouth boudins. In the PC composites, the Ti strip exhibited periodically repeating folds without fracturing while NiTi fractured producing rectangular particles. The adhesion strength results showed similar values for the interfacial bonding between PC and NiTi and PC and Ti, while a greater strength was obtained for PP and Ti compared to PP and NiTi. Neither adhesion test results or the cold rolling results obtained with the one rolling pass explain the difference in fracture patterns obtained with different combinations of constituents. A more complex analysis which also includes materials properties must be undertaken. However, a logarithmic relationship was observed between the composite stretch and the dimensionless nip-gap in those cases where fracture occurred with the minimum number of rolling passes, regardless of the material combination.
The authors are grateful to Prof. Basil Tikoff for sharing his expertise on geological structures and geophysics. The authors acknowledge funding support from the AFOSR Grant (FA 955004-1-0109), the National Defense Science and Engineering Graduate fellowship (NDSEG), and the University of Puerto Rico at Mayaguez, from where B. Calcagno was on leave of absence. References [1] Calcagno BO, Rinaldi C, Osswald T, Crone WC. Fracture induced by cold rolling in metal/polymer composites. In: Society for experimental mechanics – SEM annual conference and exposition on experimental and applied mechanics, vol. 4; 2009. p. 2357. [2] Calcagno BO. Fracture of a shape memory alloy in polymer composites induced by cold rolling. Master thesis. University of Wisconsin-Madison; 2008. [3] Menon M, Chen I-W. Bimaterial composites via colloidal rolling techniques: I, microstructure evolution during rolling. J Am Ceram Soc 1999;82:3413. [4] Liang Z, Blackburn S. Analysis of crack development during processing of laminated ceramic tubes. J Mater Sci 2002;37:4227. [5] Crone WC, Yahya AN, Perepezko JH. Bulk shape memory NiTi with refined grain size synthesized by mechanical alloying. Metastable, mechanically alloyed and nanocrystalline materials, vol. 386. Zurich-Uetikon: Trans Tech Publications Ltd.; 2002. p. 597. [6] Yazar Ö, Ediz T, Öztürk T. Control of macrostructure in deformation processing of metal/metal laminates. Acta Mater 2005;53:375. [7] Crone WC, Wu D, Perepezko JH. Pseudoelastic behavior of nickel–titanium melt-spun ribbon. Mater Sci Eng A-Struct Mater Prop Microstruct Process 2004;375–77:1177. [8] Antoun GG. Improving to adhesion between a NiTi wire and polymer matrix for composite applications and mechanical fabrication of nanometer-sized shape memory alloy NiTi particles from bulk by rolling and folding. Master science thesis. Madison: Wisconsin; 2003. [9] Perepezko JH, Hebert RJ, Wilde G. Synthesis of nanostructures from amorphous and crystalline phases. Mater Sci Eng A-Struct Mater Prop Microstruct Process 2004;375–77:171. [10] Hebert RJ, Perepezko JH. Deformation-induced synthesis and structural transformations of metallic multilayers. Scr Mater 2004;50:807. [11] Ramberg H. Natural and experimental boudinage and pinch-and-swell structures. J Geol 1955;63:512. [12] Ramberg H. Fluid dynamics of viscous buckling applicable to folding of layered rocks. Bull Am Assoc Pet Geol 1963;47:484. [13] Lloyd GE, Ferguson CC, Reading K. A stress-transfer model for the development of extension fracture boudinage. J Struct Geol 1982;4:355. [14] Schmalholz SM, Schmid DW, Fletcher RC. Evolution of pinch-and-swell structures in a power-law layer. J Struct Geol 2008;30:649. [15] Hudleston PJ, Treagus SH. Information from folds: a review. J Struct Geol 2010;32:2042. [16] Zhao PL, Ji SC. Refinements of shear-lag model and its applications. Tectonophysics 1997;279:37. [17] Ji S, Zhu Z, Wang Z. Relationship between joint spacing and bed thickness in sedimentary rocks: effects of interbed slip. Geol Mag 1998;135:637. [18] Mandal N, Deb K, Khan D. Evidence for a non-linear relationship between fracture spacing and layer. J Struct Geol 1994;16:1275. [19] Price NJ, Cosgrove JW. Analysis of geological structures. Cambridge (New York): University Press; 1990. p. 405. [20] Pollard DD, Fletcher RC. Fundamentals of structural geology. Cambridge (UK, New York): Cambridge University Press; 2005. p. 169. [21] Smith RB. Unified theory of onset of folding, boudinage, and mullion structure. Geol Soc Am Bull 1975;86:1601. [22] Smith RB. Formation of folds, boudinage, and mullions in non-newtonian materials. Geol Soc Am Bull 1977;88:312. [23] Treagus SH, Lan LB. Deformation of square objects and boudins. J Struct Geol 2004;26:1361. [24] Mengong ME, Zulauf G. Coeval folding and boudinage under plane strain with the axis of no change perpendicular to the layer. Int J Earth Sci 2006;95:178. [25] Kidan TW, Cosgrove JW. The deformation of multilayers by layer-normal compression; an experimental investigation. J Struct Geol 1996;18:461. [26] Masuda T, Kimura N, Hara Y. Progress in microboudin method for palaeostress analysis of metamorphic tectonites: application of mathematically refined expression. Tectonophysics 2003;364:1.
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351 [27] Calcagno BO. Deformation and fracture in multilayer systems during calendering processes. University of Wisconsin-Madison. PhD. Thesis; 2010. [28] Calcagno BO, Hart KR, Crone WC. Adhesion strength in metal/polymer composites. Proceedings of the 2010 annual conference on experimental and applied mechanics 2011;vol. 6:p. 49. [29] Al-Sheyyab A, Kuehnert I, Schmachtenberg E. Insert coating as a preprocessing approach for improvement of adhesive bonding in plastic–metal hybrid structures. In: Society of plastics engineers annual technical conference. plastics encounter at ANTEC 2007, conference proceedings, vol. 2; 2007. p. 996. [30] Calcagno BO, Roderic L, Tim O, Crone WC. Comparison of viscoplastic properties for polycarbonate, polypropylene, and high density polyethylene. J Plast Technol 2010;6:229.
1351
[31] Osswald T, Baur E, Brinkmann S, Oberbach K, Schmachtenberg E. International plastics handbook. Hanser Publishers; 2006. p. 522. [32] Osswald T, Hernández-Ortíz JP. Polymer processing: modeling and simulation. Hanser Publishers; 2006. p. 30. [33] Callister WD. Materials science and engineering, an introduction, 6th edition. John Wiley and Sons, Inc.; 2003. [34] Neuking K, Abu-Zarifa A, Youcheu-Kemtchou S, Eggeler G. Polymer/NiTicomposites: fundamental aspects, processing and properties. Adv Eng Mater 2005;7:1014. [35] Chan CM, Trigwell S, Duerig T. Oxidation of an NiTi alloy. Surf Interface Anal 1990;15:349. [36] Pelton AR, DiCello J, Miyazaki S. Optimisation of processing and properties of medical grade nitinol wire. Minim Invasive Ther Allied Technol 2000;9:107.
Contents lists available at SciVerse ScienceDirect
Composites Science and Technology journal homepage: www.elsevier.com/locate/compscitech
Deformation and fracture of polymer/metal composites subjected to cold rolling B.O. Calcagno a,⇑, K.R. Hart b, J.C. Springmann b, G.G. Antoun b, W.C. Crone a,c a
Materials Science Program, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, United States Engineering Mechanics Program, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, United States c Department of Engineering Physics, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53706, United States b
a r t i c l e
i n f o
Article history: Received 22 December 2011 Accepted 3 May 2012 Available online 22 May 2012 Keywords: A. Polymer–matrix composites (PMCs) B. Fracture C. Stress transfer C. Sandwich structures Cold rolling
a b s t r a c t Cold rolling of a sandwich composite with a metallic strip inclusion in a polymeric matrix can produce a range of outcomes, including deformation and fracture of the inclusion. Using different material combinations under the same processing parameters, the results ranged from minor deformation, to folding and/or loss of adhesion, and fracture of the inclusion, with fracture particles varying in size and shape. Comparisons are made between the resulting structures after cold rolling of the polymer/metal composite to geological formations. In particular, the fracture particles obtained resemble rock structures known as boudins. The phenomena of boudinage and folding encountered in the cold rolling of polymer/metal composites is similar to that seen in geology formations although the time and size scale of these events are several orders of magnitude apart. The experimental results reported show that cold rolling applied to a polymer/metal sandwiched composite induces deformation and fracture behaviors that depend on the mechanical properties of the constituents, deformation behavior of the polymeric matrices, interfacial adhesion, and process parameters such as rollers speed and nip-gap. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Metal inclusions in polymeric composite materials under a process of cold rolling and folding can exhibit fracture under certain conditions. The deformation and fracture patterns observed are similar to those found in some rock structures [1,2]. Similarities are also found with the microstructure of laminate ceramic composites subjected to rolling and folding [3], multilayer ceramic tubes that have undergone co-extrusion [4], as well as some metal/metal laminate subjected to rolling processes [5,6]. The common factor among these processes is the deformation of two contiguous phases with different mechanical properties, which could produce necking and/or fracture of one of the composite constituents. Our interest in this phenomena began with the use of cold rolling (also called calendering) for the purposes of grain refinement in shape memory alloys [5,7], later it was expanded to cold rolling of multilayer materials for the purposes of particle production [8]. This research, as well as that of others [9,10], has shown that a particle disperse composite can be produced through plastic deformation using cold rolling of a sandwich composite made of layers of dissimilar materials. The work presented here is focused on two
⇑ Corresponding author. Address: Department of General Engineering, University of Puerto Rico-Mayagüez, Call Box 9000, Mayagüez, PR 00681, United States. Tel.: +1 787 832 4040x3069; fax: +1 787 265 3816. E-mail address: [email protected] (B.O. Calcagno). 0266-3538/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compscitech.2012.05.004
layers of polymer and metal or alloy strips between them as the initial condition. In the experimental work reported below, sandwich composites were made using polypropylene (PP), polycarbonate (PC), or highdensity polyethylene (HDPE) as the matrix phase, and strips of pseudoelastic nickel titanium alloy or titanium as the inclusion phase. A sampling of the various fracture and folding patterns obtained for the different composites is shown in Fig. 1. The fracture particles varied in size, shape, and in some cases, the inclusion did not break but rather folding and/or loss of adhesion was observed. The range of materials investigated provides an interesting window into the phenomenon of cold rolling. 2. Background Under certain conditions the fracture patterns observed after cold rolling a metal inclusion in a polymeric matrix sandwich resemble fragments of rock layers called boudins. There is substantial literature on these formations. These geological structures occur when a competent rock layer surrounded by a less competent rock matrix is subjected to layer-parallel extension or layer-normal compression. The term competent is used by geologists to identify the stronger or stiffer layer, while the less competent or incompetent layer is the weaker or ductile layer [11–19]. Pollard and Fletcher [20] point out that these terms ‘‘do not refer to an explicit type of material behavior.’’ In those studies where the layers of rocks have been considered as viscous fluids, competency has been
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
1345
Fig. 1. Inclusion fracture patterns obtained after metal/polymer composite fabrication process. (a) NiTi particles in PP matrix after 1 rolling cycle; particle size ranges from 0.1 to 0.5 cm. (b) NiTi strips in HDPE matrix; no break after 1 rolling cycle. (c) Ti strips in PC matrix; no break after 1 pass; the folds measured approximately 0.18 cm. (d) Ti particles in HDPE after 1 rolling cycle; particles average size was 0.4 cm. (One rolling cycle = several passes through the rollers reducing the nip-gap 0.05 cm each pass until a nip-gap of 0.076 cm was reached).
more rigorously defined as a function of the viscosity of the layer, where a more viscous layer is the competent layer and the layer with a lower viscosity or ability to flow represents the incompetent layer [21–24]. When layer-parallel extension or layer-normal compression are applied to the competent layer, flow in the less competent matrix parallel to the layer will occur, generating tensile stresses, extension, and finally fracture of the layer [11,25]. The shape of the boudin is determined by the properties of the layer, where strong brittle layers would produce rectangular boudins with sharp corners as those shown in Fig. 2a [26]. More ductile layers would allow plastic flowage and lateral elongation producing structures termed pinch-and-swell which result from necking. These structures may eventually fracture producing more oblong or lensshape called lenticular boudins, or barrel-shape boudins [11,14]. Fig. 2b shows a barrel-shaped boudin with concave end that is believed to be the result of high plastic deformation imposed on the
originally rectangular boudin. This kind of rock structure is also known as ‘‘fish-mouth boudins’’ [23]. Other rock structures known as folds occur when competent viscous layers in a less stiff matrix buckle when subjected to layer-parallel compression [15]. Structures such as the pinch-and-swell, and folding were explored by Smith’s theory on formations of folds, boudinage, and mullions in Newtonian and non-Newtonian materials [21,22]. Using a general approach, Smith hypothesized that folding and boudinage are related structures that occur when a single layer is squeezed between two thick layers of different viscosity, establishing a state of motion with unstable small disturbances. Fig. 3 portrays a summary of the layer instabilities, where the stronger layer is depicted as the darker region, and the more ductile as the lighter region. The disturbances could grow into flow, fracture boudinage, folding, or mullions. The parallels between the boudin and fold structures and the development of fracture and folding in a metal-alloy/polymer
Fig. 2. Images of typical boudin fracture pattern in rock structures. (a) Rectangular boudin: Tourmaline grains in a quartz matrix from Western Australia [26]. (b) Shapedbarrel boudin with ‘fish-mouth’ shape ends: Boudin in calc-silica layer in marble, Khan Gorge, Namibia [23]. Coin shown for scale.
1346
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 3. Four classifications of dynamic instability cases in a single layer. The more competent layer is depicted in darker gray. Figure adapted from [21].
composite during cold rolling motivated a more detailed investigation. We sought to identify the roles that rolling process, material properties, interfacial adhesion, and load transfer play in the deformation and fracture of an inclusion material sandwiched between layers of dissimilar material. We discuss here the results of two of the experimental techniques used in our work, cold rolling and pullout tests.
3. Experimental methods The process investigated is called cold rolling (or calendering) of a sandwich composite. Fig. 4a shows a schematic of the rolling process. The sandwich composite is introduced to the rollers which run at a constant speed. The nip-gap (distance between the rollers) is fixed at a dimension smaller than the sandwich thickness. The sandwich composites were made using polypropylene (PP), polycarbonate (PC), or high-density polyethylene (HDPE) (McMaster, #8742K131, #8574K24, #8619K421, size 1200 1200 1/1600 ) as the matrix phase, and strips of pseudoelastic nickel titanium alloy (Nitinol Devices & Components, NiTi foil, Alloy type SE 508, pseudoelastic at room temperature, 55.6 at.% Ni, 0.0127 mm thick) or titanium (Alfa Aesar, Ti foil, 99%, 0.127 mm thick) as the inclusion phase [1,27]. The surface of the NiTi and Ti strips were etched using a solution of 1HF + 4HNO3 + 5H2O. The composite sandwich samples were made by placing strips of metal or alloy between two layers of polymers and heating them while under 5 lbs of compressive load. The molding temperature for the PP composites was 197 °C, 230 °C for the HDPE, and 290 °C for the PC.
Two types of cold rolling experiments were performed. In the repeated rolling experiments, a metal/polymer or alloy/polymer sandwich composite containing several strips of metal/alloy was subjected to cold rolling several times. With each rolling pass the nip-gap (the spacing between the rollers) was reduced in order to decrease the sandwich composite thickness and induce plastic deformation. All of the repeated rolling experiments reported here were performed at the same roller speed with a Stanat rolling mill (model TA-315). The sandwich composites were made with 8 cm 8 cm polymer squares and 2.5 cm 0.5 cm metal/alloy strips [2]. A second set of experiments to produce one fracture in the inclusion was performed to investigate the process parameters that were relevant during the cold rolling. Samples with only one metal/alloy strip between polymer layers were prepared and subjected to the minimum number of rolling passes to fracture the inclusion during their passage through the rollers [1,27]. These tests were performed with a 5471 150 New Durston rolling mill (model D2-120). The strip width was 5 mm and the length varied from 5 mm to 20 mm [2]. In order to quantify the strength of the bonding between the metal and the polymeric matrix, pullout tests on the composites were performed where a tensile force was applied to the inclusion to produce debonding between the two constituents of the sandwich composite [27,28]. A double overlapping configuration following that of Al-Sheyyab et al. [29] was used with a 5 mm wide embedded strip. Testing was conducted with an Instron 5566 using a constant crosshead speed of 5 mm/min. The embedded length of the strips ranged between 10 and 12 mm [27]. The surface of all strips was subjected to the same etching treatment applied to the cold rolling strips. Fig. 4b shows a schematic of the pullout test performed. All the experiments described were done at room temperature.
4. Results The repeated rolling experiments produced a range of behaviors depending on the material combination and rolling parameters used (e.g. roller speed, nip-gap, and sample dimensions). Under the same rolling conditions, NiTi in PP fractured producing particles with rectangular corners (see Figs. 1a and 5), similar to the behavior exhibited by Ti in PP. In contrast, NiTi in HDPE did not fracture (Fig. 1b), and Ti in a HDPE produced particles with distorted fracture surfaces (Fig. 1d). Under other conditions, fracture was not observed, and instead delamination of the composite constituents occurred and the strip acquired out-of-plane
Fig. 4. (a) Schematic of the cold rolling test for one pass. (b) Lateral and frontal view of a pullout sample.
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 5. Image of NiTi strip in PP matrix. The NiTi fractured producing two particles with rectangular and sharp corners after four rolling passes at a roller speed of 11.2 rpm. The original sample thickness was 1.55 mm and it was rolled at the following nip gaps: 1.5, 1.0, 0.8, and 0.6 mm. The strip was 20.17 mm long, and 5.65 mm wide; the thickness varied between 0.11 and 0.12 mm. Ruler in millimeter scale.
deformations. For example, for Ti in PC after only one rolling pass the Ti strip exhibited periodically repeating folds (Fig. 1c) and loss of adhesion as shown in Fig. 6. The three polymers were specifically chosen for the difference in their materials properties [30]. Under rolling conditions in the sandwich composite, the PC displayed brittle behavior, while the PP matrix deformed plastically and the HDPE matrix deformed extensively. Although the NiTi and Ti strips were prepared identically and expected to have the same surface oxide, differences in composite behavior were consistently observed between the pseudoelastic NiTi and ductile Ti. Under conditions that produce fracture for a particular composite combination, multiple rolling passes can be made to produce additional fractures either by folding the composite to double its thickness or decreasing the nip-gap. In general, fracture of the strip initiated at the strip edge and propagated perpendicular to the direction of rolling. Fractures in multiple directions can be achieved by rotating the sample as shown in Fig. 7 for NiTi in PC. Fig. 8 shows two samples subjected to ten cycles of cold rolling, folding and heating. This repeated rolling and folding produces a composite with a particle distribution. The size of particles was smaller for Ti in PP compared with the NiTi in PP for the same number of passes. The set of rolling experiments designed to initiate one to two fractures in the metal or alloy strip within the minimum number
1347
Fig. 7. Particles of NiTi in PC matrix after 1 rolling cycle, 90° rotation, and an additional rolling pass; particles measured approximately 0.2 cm.
of rolling passes provided additional information about the relationship between rolling parameters and fracture. Parameters such as rolling speed, nip-gap, sample thickness, strip length, and stretch of the composite were studied. Prior to deformation, lines were drawn on the exterior of the sample 1 cm from each end and outside the location of the metal strip ends. The distance between these lines were again measured after cold rolling in order to calculate the composite stretch, (Dllf/Dll0). The nip-gap roller separation, 2h0, was also compared to the original composite thickness, T0. The key relationship found was between the overall composite stretch (Dllf/Dll0) and the dimensionless nip-gap (2h0/T0) [1,2]. Fig. 9a shows results for one roller speed and Fig. 9b shows for all roller speeds and all composite material combinations. The results for the pull out tests are summarized in Fig. 10 [28]. The use of embedded strips (rather than wires) in the pull out testing was chosen to better emulate the sample configuration in cold rolling and also avoided issues of having the metal/alloy undergo plastic deformation or transformation prior to debonding. The tests applied load to the strips until the maximum force needed to induce failure of the polymer/strip interface was attained. After this point, only friction opposes the sliding at the interface. Adhesion strength is defined as the maximum force needed to produce debonding over the embedded strip surface area. Results were discarded for samples where torsion of the strip occurred due to incorrect placing of the sample in the grips and for samples containing voids in the matrix material that were produced during manufacture of the samples. 5. Discussion
Fig. 6. Images of a Ti/PC composite showing loss of adhesion during cold rolling. (a) Top view of entire sample. (b) Side view of central buckled regions. (c) Side view of strip end. Ruler in millimeter scale.
The cold rolling of sandwich composites with a metallic strip embedded in a polymeric matrix produces different fracture patterns for different combinations of constituents for the same processing parameters as portrayed in Fig. 1. The polymers used as matrices were two semicrystalline polymers, PP and HDPE, and one amorphous, PC. PP exhibits high dynamic loading capacity, and a higher strength and stiffness than HDPE. HDPE is a tough polymer with low strength and stiffness, and high elongation at break; also less dense than other polymers. PC is a transparent polymer with a high strength and stiffness, a great resistance to impact, and susceptible to crack under stress [31,32]. Titanium and pseudoelastic nickel titanium strips were used as inclusions. Titanium is a transition metal, of high strength
1348
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 8. Images of particle dispersed composites obtained by cold rolling and folding. (a) NiTi particles in PP matrix, (b) Ti particles in PP matrix. The original samples measured 8 cm 8 cm and contained the same number of metal strips 2.5 cm 0.5 cm, with thicknesses between 0.28 and 0.29 mm. They were subjected to 10 rolling cycles aligned with the length of the inclusion. Ruler in millimeter scale.
Fig. 9. Composite stretch (Dllf/Dllo) plotted against dimensionless nip gap (2ho/To) for samples whose strip inclusion fractured in one or two locations within three or fewer rolling passes. (a) Data for samples that fractured at a roller speed of 3.7 rpm. The line given is a logarithmic fit curve (R = 0.97). (b) Data for the six composite combinations tested and the roller speeds: 3.7, 11.2, 18.3, 22, 25, and 29 rpm [1].
and toughness, low density, nonmagnetic, and highly resistant to corrosion [33]. Pseudoelastic nickel titanium is a shape memory alloy (SMA), which can be deformed to strains up to 10% when a load is applied and recover its original shape upon unloading. It is nonmagnetic, dense, of high strength, and moderate impact resistance [34]. Both materials used as inclusions, titanium and nickel titanium alloy, oxidize to form a resistant titanium oxide layer on the surface [35,36]. Table 1 contains relevant properties of the composite constituents used in this work. When comparing the rolled composites with the rock structures, the metallic inclusions represented the competent surrounded by less competent polymeric matrices. During the cold rolling process the NiTi in PP fractured producing particles with rectangular corners as shown in Fig. 1a and Fig. 5 similar to the boudins shown in Fig. 2a; Ti in PP (not shown in this paper) exhibited the same behavior. However, a range of behaviors was observed with cold rolling of the polymer/metal combinations described here. Under the same processing parameters (roller speed, nip-gap, and sample dimensions), NiTi embedded in HDPE did not fracture (Fig. 1b), while the cold rolling of Ti in a HDPE matrix (Fig. 1d) produced particles with oblong shapes similar to the shaped-barrel
boudin of Fig. 2b. The Ti strip exhibited elongation of approximately 10%. In both of the HDPE composites the matrix exhibited extensive plastic deformation. In the case of NiTi in PC the metal strip fractured resembling rectangular boudins with minimal plastic deformation of the matrix (see Fig. 7). That was not the case of the Ti in PC, which exhibited folds that repeated periodically along the metal strip after only one rolling pass (Fig. 1c). These folds are very much like the rock fold structures that occur due to buckling of the more competent layer when subjected to layer-parallel compression. However in the experimental case the loading applied was the same for all the samples. Fig. 1c also shows loss of adhesion between the Ti and the PC matrix at both sides of the metal strips. The debonding between the two constituents starts from the strip end and increases as the folds grow as displayed in detail in Fig. 6 where the side view of the strip is portrayed. In order to identify the role of the parameters involved in the cold rolling process, a second cold rolling test was performed. The goal of this test was to get one or two fractures of the inclusion after the minimum number of passes of the composite sample through the rollers. Varying the nip-gap and the speed of the rollers it was possible to fracture all the metallic inclusions in all
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
Fig. 10. Adhesion strength for pullout samples of nickel titanium (NiTi) and titanium (Ti) strips embedded within polypropylene (PP), polycarbonate (PC), or high-density polyethylene (HDPE) polymers [28].
Table 1 Properties for composite constituents showing: density, Young modulus, ultimate strength, strain at fracture, and yield strength.
a b
Constituent
q (g/cm3)
E (GPa)
rU (MPa)
ef (%)
ry (MPa)
NiTia Tia PPb PCb HDPEb
6.5 4.5 0.3 1.2 0.94
16 ± 2 23 ± 5 1.1–1.3 2.1–2.4 0.7–1.4
743 ± 11 305 ± 5 21–37 56–67 18–35
22 ± 3 35 ± 3 20–800 100–130 150
324 ± 9 270 ± 12
Mechanical properties from reference [27]. Properties from reference [31].
polymeric matrices even in cases such as the NiTi in HDPE (Fig. 1b) and Ti in PC (Fig. 1c) samples. The nip-gap needed to induce fracture depended on the sample initial thickness and the speed; and was different for all inclusion/matrix combinations. For the same initial sample thickness and roller speed, the NiTi embedded in PC matrix fractured at larger nip-gaps than in PP. In order to fracture the Ti in PP, under the same parameters, a smaller nip-gap than the NiTi in PP was needed. Fracture was also obtained at the same roller speed in the case of the Ti in PC but with extremely small nip-gap. For NiTi in HDPE, the speed of the rollers had to be increased more than six times to fracture the inclusion. In those cases where fracture did not occur, the inclusion folded. Dimensional analysis was used to find the relationship between the variables involved in the cold rolling process that induced fracture. As illustrated by Fig. 9a, the composite stretch was shown to have a logarithmic relationship with the dimensionless nip-gap, which was defined as the spacing between the rollers divided by the initial sample thickness. This relation is given by Dllf/Dll0 = 0.97–1.28 log (2h0/T0) [1,27]. It is notable that each inclusion/matrix combination is clustered in a different region of the curve. This trend is maintained when the data for all roller speeds for the six composites is plotted as shown in Fig. 9b. The transfer of the load applied by the rollers to the composite and thus to the metal strip depends on the interfacial adhesion between the two constituents. The strength of this interfacial bonding was measured using pullout tests. The results of these tests, summarized in Fig. 10, show that the adhesion strength of PC and NiTi is similar to that of PC and Ti [27,28]. However, although the preparation of these composites was the same, the cold rolling
1349
tests gave a different behavior for these two material combinations. Figs. 1c and 6, show images of a Ti/PC sample where loss of adhesion occurred between the Ti strip and the PC after an attempt to produce a particle disperse composite using cold rolling. For the same cold rolling parameters, the NiTi in PC fractured. To obtain fracture of the metal strip in Ti/PC samples with only one rolling pass, an extremely small spacing between rollers was needed as proved with the one-rolling-pass tests, producing also cracking of the matrix of PC. For the PP composites, the interfacial bonding between the polymer and the Ti strips was higher than the value obtained for the NiTi and PP. After subjecting these composites to ten cycles of cold rolling, folding and heating, a composite with smaller and more uniform particle size distribution was obtained for the Ti/PP compared with the NiTi/PP composites as shown in Fig. 8 [2]. It would seem that a stronger interfacial adhesion could provide a more efficient load transfer between the Ti strip and the PP matrix. However, the comparison of the adhesion strength values for the HDPE composites show a nearly equivalent value for the NiTi and Ti, while under the same rolling parameters the Ti in HDPE fractured but the NiTi in HDPE did not. Loss of adhesion was also observed in Ti/PP and NiTi/PP composites where fracture was easily obtained. Fig. 11 displays images of a Ti/PP sample where folding occurred and the interfacial bonding between matrix and inclusion was lost during cold rolling. When the process parameters were analyzed, it was found that the nip-gap imposed during the cold rolling was larger than the necessary nip-gap predicted by the relation obtained with onerolling-pass-test [1,27]. For a smaller nip-gap, fracture after one pass through the rollers was obtained in Ti in PP matrix as shown in Fig. 12. The 4.82 mm long and 0.09 mm thick Ti strip fractured in two locations, in one of them the segments appear overlapped while in the second they are completely separated from each other. Although NiTi has pseudoelastic capability, the main differences relevant to fracture behavior between the two inclusions types (NiTi and Ti) are the ultimate strength and elongation at break. In comparison, the ultimate strength of NiTi is significantly higher while the elongation at break for Ti is significantly greater. For the polymers, the ultimate strength of PC is higher while the elongation at break for PP is significantly greater. No obvious explanation for the behavior of these material combinations under cold rolling can be provided by a simple comparison of the mechanical
Fig. 11. Images of a Ti/PP composite (49.80 mm 36.72 mm 2.41 mm) showing loss of adhesion during cold rolling at a speed of 3.7 rpm and a nip gap of 1.91 mm. (a) Top view of entire sample. (b) Side view of strip end. (c) Side view of central buckled region.
1350
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351
The phenomena of boudinage and folding encountered in the cold rolling of metal/polymer composites is observed to be similar to geology formations, although the time and size scale of these events are several orders of magnitude apart. These observations warrant further study due to their applicability to both geological problems and composite material processing. Acknowledgments
Fig. 12. Images of a Ti/PP sample showing fracture that occurred during cold rolling at a speed of 3.7 rpm and a nip gap of 0.25 mm. (a) Top view of sample. (b) Side view of overlapped fractured segments. (c) Side view of fractured segments separated by a distance.
properties of the constituents. Clearly there is a more complex interplay between the materials properties of the constituents, their adhesion behavior in the combinations explored, and the rolling parameters. The experimental results obtained in this work show that cold rolling can be used to study the phenomena of boudinage and folding using a metal strip as the competent layer sandwiched between layers of a polymer as the less competent layer. While the state of stress to which the composite is subjected when going through the rollers is complex and may differ from the layer-parallel extension or layer-normal compression cited in the geology references, it produces similar structures. Furthermore cold rolling provides a convenient and simple method to obtain different particle shapes just varying the composite constituents and the processing parameters at room temperature. 6. Conclusions In this research we have shown that the cold rolling process of a composite material consisting of a metallic inclusion sandwiched between layers of a polymeric matrix induces deformation and fracture behaviors that depend not only on the mechanical properties of the constituents, deformation behavior of the polymeric matrices, and interfacial adhesion, but also on processing parameters such as roller speed and nip-gap. In the experimental work described, composites with polypropylene (PP), polycarbonate (PC), or high-density polyethylene (HDPE) as the matrix phase, and strips of pseudoelastic nickel titanium alloy or titanium as the inclusion phase were subjected to cold rolling. Additionally, pullout tests on the composites were performed to quantify the strength of the bonding between the metal and the polymeric matrix. The cold rolling experiments aimed to investigate which process parameters and constituent properties were relevant in the inclusion fracture event. Results show that under the same rolling conditions, NiTi and Ti in PP fractured producing particles with rectangular corners, similar to boudins. In contrast, NiTi in HDPE did not fracture, and Ti in a HDPE produced particles with distorted fracture surfaces that resembled fish mouth boudins. In the PC composites, the Ti strip exhibited periodically repeating folds without fracturing while NiTi fractured producing rectangular particles. The adhesion strength results showed similar values for the interfacial bonding between PC and NiTi and PC and Ti, while a greater strength was obtained for PP and Ti compared to PP and NiTi. Neither adhesion test results or the cold rolling results obtained with the one rolling pass explain the difference in fracture patterns obtained with different combinations of constituents. A more complex analysis which also includes materials properties must be undertaken. However, a logarithmic relationship was observed between the composite stretch and the dimensionless nip-gap in those cases where fracture occurred with the minimum number of rolling passes, regardless of the material combination.
The authors are grateful to Prof. Basil Tikoff for sharing his expertise on geological structures and geophysics. The authors acknowledge funding support from the AFOSR Grant (FA 955004-1-0109), the National Defense Science and Engineering Graduate fellowship (NDSEG), and the University of Puerto Rico at Mayaguez, from where B. Calcagno was on leave of absence. References [1] Calcagno BO, Rinaldi C, Osswald T, Crone WC. Fracture induced by cold rolling in metal/polymer composites. In: Society for experimental mechanics – SEM annual conference and exposition on experimental and applied mechanics, vol. 4; 2009. p. 2357. [2] Calcagno BO. Fracture of a shape memory alloy in polymer composites induced by cold rolling. Master thesis. University of Wisconsin-Madison; 2008. [3] Menon M, Chen I-W. Bimaterial composites via colloidal rolling techniques: I, microstructure evolution during rolling. J Am Ceram Soc 1999;82:3413. [4] Liang Z, Blackburn S. Analysis of crack development during processing of laminated ceramic tubes. J Mater Sci 2002;37:4227. [5] Crone WC, Yahya AN, Perepezko JH. Bulk shape memory NiTi with refined grain size synthesized by mechanical alloying. Metastable, mechanically alloyed and nanocrystalline materials, vol. 386. Zurich-Uetikon: Trans Tech Publications Ltd.; 2002. p. 597. [6] Yazar Ö, Ediz T, Öztürk T. Control of macrostructure in deformation processing of metal/metal laminates. Acta Mater 2005;53:375. [7] Crone WC, Wu D, Perepezko JH. Pseudoelastic behavior of nickel–titanium melt-spun ribbon. Mater Sci Eng A-Struct Mater Prop Microstruct Process 2004;375–77:1177. [8] Antoun GG. Improving to adhesion between a NiTi wire and polymer matrix for composite applications and mechanical fabrication of nanometer-sized shape memory alloy NiTi particles from bulk by rolling and folding. Master science thesis. Madison: Wisconsin; 2003. [9] Perepezko JH, Hebert RJ, Wilde G. Synthesis of nanostructures from amorphous and crystalline phases. Mater Sci Eng A-Struct Mater Prop Microstruct Process 2004;375–77:171. [10] Hebert RJ, Perepezko JH. Deformation-induced synthesis and structural transformations of metallic multilayers. Scr Mater 2004;50:807. [11] Ramberg H. Natural and experimental boudinage and pinch-and-swell structures. J Geol 1955;63:512. [12] Ramberg H. Fluid dynamics of viscous buckling applicable to folding of layered rocks. Bull Am Assoc Pet Geol 1963;47:484. [13] Lloyd GE, Ferguson CC, Reading K. A stress-transfer model for the development of extension fracture boudinage. J Struct Geol 1982;4:355. [14] Schmalholz SM, Schmid DW, Fletcher RC. Evolution of pinch-and-swell structures in a power-law layer. J Struct Geol 2008;30:649. [15] Hudleston PJ, Treagus SH. Information from folds: a review. J Struct Geol 2010;32:2042. [16] Zhao PL, Ji SC. Refinements of shear-lag model and its applications. Tectonophysics 1997;279:37. [17] Ji S, Zhu Z, Wang Z. Relationship between joint spacing and bed thickness in sedimentary rocks: effects of interbed slip. Geol Mag 1998;135:637. [18] Mandal N, Deb K, Khan D. Evidence for a non-linear relationship between fracture spacing and layer. J Struct Geol 1994;16:1275. [19] Price NJ, Cosgrove JW. Analysis of geological structures. Cambridge (New York): University Press; 1990. p. 405. [20] Pollard DD, Fletcher RC. Fundamentals of structural geology. Cambridge (UK, New York): Cambridge University Press; 2005. p. 169. [21] Smith RB. Unified theory of onset of folding, boudinage, and mullion structure. Geol Soc Am Bull 1975;86:1601. [22] Smith RB. Formation of folds, boudinage, and mullions in non-newtonian materials. Geol Soc Am Bull 1977;88:312. [23] Treagus SH, Lan LB. Deformation of square objects and boudins. J Struct Geol 2004;26:1361. [24] Mengong ME, Zulauf G. Coeval folding and boudinage under plane strain with the axis of no change perpendicular to the layer. Int J Earth Sci 2006;95:178. [25] Kidan TW, Cosgrove JW. The deformation of multilayers by layer-normal compression; an experimental investigation. J Struct Geol 1996;18:461. [26] Masuda T, Kimura N, Hara Y. Progress in microboudin method for palaeostress analysis of metamorphic tectonites: application of mathematically refined expression. Tectonophysics 2003;364:1.
B.O. Calcagno et al. / Composites Science and Technology 72 (2012) 1344–1351 [27] Calcagno BO. Deformation and fracture in multilayer systems during calendering processes. University of Wisconsin-Madison. PhD. Thesis; 2010. [28] Calcagno BO, Hart KR, Crone WC. Adhesion strength in metal/polymer composites. Proceedings of the 2010 annual conference on experimental and applied mechanics 2011;vol. 6:p. 49. [29] Al-Sheyyab A, Kuehnert I, Schmachtenberg E. Insert coating as a preprocessing approach for improvement of adhesive bonding in plastic–metal hybrid structures. In: Society of plastics engineers annual technical conference. plastics encounter at ANTEC 2007, conference proceedings, vol. 2; 2007. p. 996. [30] Calcagno BO, Roderic L, Tim O, Crone WC. Comparison of viscoplastic properties for polycarbonate, polypropylene, and high density polyethylene. J Plast Technol 2010;6:229.
1351
[31] Osswald T, Baur E, Brinkmann S, Oberbach K, Schmachtenberg E. International plastics handbook. Hanser Publishers; 2006. p. 522. [32] Osswald T, Hernández-Ortíz JP. Polymer processing: modeling and simulation. Hanser Publishers; 2006. p. 30. [33] Callister WD. Materials science and engineering, an introduction, 6th edition. John Wiley and Sons, Inc.; 2003. [34] Neuking K, Abu-Zarifa A, Youcheu-Kemtchou S, Eggeler G. Polymer/NiTicomposites: fundamental aspects, processing and properties. Adv Eng Mater 2005;7:1014. [35] Chan CM, Trigwell S, Duerig T. Oxidation of an NiTi alloy. Surf Interface Anal 1990;15:349. [36] Pelton AR, DiCello J, Miyazaki S. Optimisation of processing and properties of medical grade nitinol wire. Minim Invasive Ther Allied Technol 2000;9:107.