Experimental investigation of the effect of residual stresses and interfacial adhesion on rapid crack propagation in bilayered polypropylene/polyethylene (PP/PE100) pipes.
Nowadays pipe for water and gas distribution systems are routinely manufactured from polyethylene (PE). Among the key attractions of PE pipe systems are the ability to use non-dig methods for pipe installation, and the ability to weld the pipes and fittings together giving leak-free, fully restrained joints. However, trenchless installation methods such as pipe bursting and directional drilling may damage the pipe , and this could have an impact on the pipe performance as cracks may grow from flaws and scratches introduced during pipe laying. In addition, although laboratory tests on fusion joints have proven that for well-made contamination-free joints the joint strength can be equal or greater than the pipe itself, it is also known that the joint quality can be reduced by poor pipe preparation--leading to earlier failure [2, 3].
Almost two decades ago, the technology for manufacturing a bilayered or "skinned" pipe composed of a pressure rated core PE pipe, and a thin sacrificial and peelable polypropylene (PP) skin was developed. The core PE pipe and the skin are co extruded, meeting in the liquid phase in the die. The skinned pipe is then cooled conventionally, with the skin shielding the PE core from contact with the cooling water . The external surface of the PE core pipe is not in contact with the atmosphere until the PP skin is removed. The core PE pipe meets the outside diameter and wall thickness requirements for solid wall pipe, that is, the PP skin constitutes extra material.
The skin serves to protect the core pipe during transportation, handling on site, and installation. It absorbs any damage and thus preserves the long-term strength of the core pipe , When the skin is removed it exposes a clean, virgin surface ready for electrofusion (EF) joining, thus increasing on-site productivity as scraping is not necessary. Studies have shown that skinned pipes offer other advantages such as an increase in falling weight impact strength when compared to solid pipes, and a good long-term stress crack performance , Full Scale (FS) Rapid Crack Propagation (RCP) tests on skinned pipes (630 mm SDR 21) confirmed that they met the RCP requirements .
However, previous studies by Leevers et al. [7, 8] and Munch  revealed another possible effect of the skin layer: surface embrittlement. S4 and three point bending impact tests showed that a PP skin bonded to a PE100 pipe increased its RCP critical temperature and its transition temperature. They concluded that the effect of the skin is essentially similar to that of thickness-induced constraint [10. 11], and that modulus inequality between skin and core material, skin thickness, and interfacial adhesion has an important role in determining the fracture toughness of the bilayered structure.
In this article, the effect of residual stresses and interfacial adhesion on the RCP performance of bilayered PP/PE100 pipes is investigated.
Surface embrittlement occurs when a thin, brittle layer bonded to the surface of a normally ductile material causes it to fail in a brittle manner. This phenomenon may arise from surface weathering and degradation  or from common procedures such as the application of a thin layer of paint or a protective film layer onto a tough material .
In order to investigate whether the PP skin has a significant embrittling effect on PE100 pipe. S4 critical temperature tests, at 5 bar, were performed on 110/11 pipes (110-mm outside diameter, and a "standard dimensional ratio" of outside diameter to nominal wall thickness SDR 11) with different nominal skin thickness (0.4, 0.8. and 1.5 mm). Table 1 shows that skinned pipes have a higher S4 critical temperature [T.sub.c-S4]. Note that solid PE pipe of this strength class will normally have [T.sub.c-S4] well below its expected service temperature, making sustained RCP impossible, and thus the slight increase in [T.sub.c-S4] in bilayered pipes still met the RCP requirements.
EFFECT OF SKIN STIFFNESS AND THICKNESS
Leevers et al.  developed a linear elasticity model to explain and predict the effect on brittle-tough transition temperature of a polymer when a skin is attached to it. The model uses elastic plate theory and linear elastic fracture mechanics (LEFM) to investigate the constraining effect of the skin on plastic deformation. It is assumed that the skin is perfectly attached to the core material. It was concluded that the effect of the skin is similar to that of thickness-induced constraint.
The model, as illustrated in Fig. 1a and b, assumes a relatively thick core with a stiff skin layer of a different material bonded to each of its free surfaces. First, it calculates the dimpling or core surface deflection that would occur around the crack under plane stress conditions and in the absence of the skin (Fig. lc). Then, it uses plate theory to calculate the external pressure distribution needed to conform the skin to the deformed stress-free surface (Fig. 1d). It then assumes that this pressure distribution is actually applied to the skin by tensile tractions acting through the skin--core interface. Finally, it compares these tensile tractions [[sigma].sub.i] to the through-thickness tensile stresses [[sigma].sub.zz] that would be generated by plane strain constraint. This comparison provides a criterion for skin-induced fracture mode based on a dimension less parameter C which is strongly dependent on skin thickness h:
C [equivalent to] [[[sigma].sub.i]/[[sigma].sub.zz]] = [25/384] [1/ (1 - [v.sup.2.sub.2]][[E.sub.2]/[E.sub.1]][B.sub.l][h.sup.3][r.sup.-4] (1)
where [E.sub.1] and [B.sub.1] are the elastic modulus and thickness of the core, [E.sub.2] and [v.sub.2] the modulus and Poisson's ratio of the skin, and r the radius of the dimpling zone at the crack tip. According to this model, when r = [r.sub.p] (the plastic zone size) the brittle to tough transition temperature (TBT) is defined as the point at which C [equivalent to] 1. Thus when C > 1 plane strain conditions are induced by the skin into a surface previously under plane stress, removing the toughness usually associated with it.
Previous work on layered structures [8, 9J. where the skin and core elastic moduli are different, has shown that the predicted transition temperature using this method agrees well with the measured one. It was concluded that the principal source of surface embrittlement is the skin-core modulus inequality, and that the skin thickness plays a predominant role in the fracture behavior.
For the present work, tensile properties of the three skins were measured using an instrumented low rate tensile test machine. Dumb-bell specimens were stamped out from sheets of peeled PP skin , An elastic modulus of 640, 625, and 621 MPa was found for the 0.4, 0.8, and 1.5 mm skin, respectively (strain rate of 0.02 [s.sup.-1]). Similar values were reported by Munch , who worked with pipes from the same manufacturer. Although all the skinned pipes tested during this investigation were co-extruded from the same base materials, the amount of additives in the skin is modified in each pipe to achieve the desired adhesion, and that explains the variations in modulus. However, this variation is relatively small, and because C varies proportionally with the third power of the skin thickness h (Eq. 1), the constraint and thus the embrittlement effect was expected to be more significant for thicker skins.
As shown in Table 1, pipes with 0.4 and 0.8 mm skin have the same critical temperature ([T.sub.c-S4] = -10[degrees]C) while the thickest skin (1.5 mm) has a lower [T.sub.c-S4] = -11[degrees]C. Although the difference is minimal, this slightly better RCP performance for 1.5 mm pipes can be explained with another variable: the adhesion between the pipe and the skin, which varies from one pipe to another.
EFFECT OF ADHESION
Munch  reported that a change in the level of adhesion between the skin and the core pipe affects the RCP performance of the whole composite, and presented a model in which the skin de-bonds from the core within a certain radius from the crack tip. If the plane stress plastic zone is smaller than this debond radius then the constraining effect of the skin is completely lost. If the plastic zone is larger, the skin will delay yield and thus constrain the core. The higher the interfacial adhesion, the smaller the de-bond radius and the more the effect of the skin approaches the assumptions from the previous constraint model.
PEEL TESTS RESULTS
The following analysis is based on the work by Williams and coworkers [15, 16], Figure 2a shows an illustration of the fixed arm peel test. In order to peel the skin from the substrate or core material, it is necessary to provide energy in the form of external work to the laminate. An energy balance is used to calculate the adhesive fracture toughness [G.sub.A] as
[G.sub.A] = [d[U.sub.ext]/bda] - [d[U.sub.s]/bda] - [d[U.sub.dt]/bda] - [d[U.sub.db]/bda] (2)
where [U.sub.ext], [U.sub.s], [U.sub.dt], and [U.sub.db] refer to external work, stored strain, tensile deformation, and bending deformation energies. respectively, and da is the peel fracture length. Considering a peeling arm of thickness h and width b which is peeling in a steady-state under a constant force P at and applied peel angle [theta], then
[U.sub.ext] = Pda(1 + [epsilon] - cos [theta]) (3)
d([U.sub.s] + [U.sub.dt]) = bhda [[integral].sup.[epsilon].sub.0][sigma]d[epsilon] (4)
where [epsilon] is the tensile strain in the peeling arm. Therefore, if any tensile deformation of the peeling arm is considered, but the bending of the peeling arm is assumed to be only elastic, the adhesive fracture energy is given by
[G.sub.A] = [G.sup.eb.sub.A] = [P/b](1 + [epsilon] - cos [theta]) - h [[integral].sup.[epsilon].sub.0][sigma]d[epsilon] (5)
where the superscript eb indicates elastic bending.
Now, if plastic or viscoelastic bending of the peeling arm occurs near the crack front, the plastic strain energy release rate [G.sup.db] has to be subtracted from [G.sub.A]
[G.sub.A] = [G.sup.eb.sub.A] - [G.sup.db] = [P/b](1 + [epsilon] - cos [theta]) - h [[integral].sup.[epsilon].sub.0][sigma]d[epsilon] - [G.sup.db] (6)
where [G.sub.db] = d[U.sub.ab]/bda. Different expressions for calculating [G.sup.db] have been derived for different cases [15. 16], where the values of elastic modulus, plastic yield strain, and a work-hardening parameter are needed. In this study, a linear-elastic stiffness approach is considered, and thus [G.sub.A] was calculated by knowing the average peel force and peel arm deformation (Eq. 5).
Peel tests were performed on a low rate instrumented tensile machine using a jig with a linear bearing system. An angle [theta] = 90[degrees] was used. The specimens were cut axially from the pipe and were 150 mm long and 10 mm wide (arm width). They were secured by two screws in a sample holder, as shown in Fig. 2b.
Results are included in Table 1. Results show that adhesion increased with skin thickness. Similar results were reported by Munch . This means that thicker skins, which have higher interfacial adhesion, should exert a greater constraint in the core and induce plane strain conditions, decreasing the RCP resistance of the pipe. However, as highlighted previously, the RCP critical temperature of pipes with thicker skins was lower.
It is possible that the RCP performance of 1.5 mm skinned pipe is better because, as shown in Table 1, it has residual stresses 5.7 and 7.5% lower than the 0.8 and 0.4 mm skinned pipes, respectively. The effect of residual stresses on the RCP performance of bilayered pipes will be analyzed in the following sections.
To investigate the effect of the adhesion independently from the effect of residual stresses, "zero-adhesion" skinned pipe specimens were prepared and tested. Normally, there should be a minimum level of adhesion between the skin and the core (peel forces of at least < 0.2 N/mm) to avoid the skin becoming detached during pipe coiling and storage and when using some no-dig pipe installation procedures , Here 1.5 mm skinned pipe specimens were carefully peeled and the skin was then used to wrap the same pipe and welded in place (Fig. 3). These specimens were then tested using a crack path at 180[degrees] from the welded line.
As shown in Table 1, zero-adhesion specimens performed better than as-received 1.5 mm skinned pipes ([T.sub.c-S4] = -11[degrees]C) and also better than the solid specimens ([T.sub.c-S4] = -14[degrees]C) as RCP was not sustained above -17[degrees]C ([T.sub.c-S4] = -16[degrees]C). This observation is in agreement with Munch's suggestion that as the adhesion falls below a threshold at which the skin no longer constrains the core pipe, bilayered pipes may even present better RCP resistance than monolayered ones. In the present tests, adhesion was completely removed rather than decreased to a threshold value, and the only effect of the skin was to add its own unconstrained RCP resistance to that of the pipe wall.
Annealed Bilayered Pipes
Approximately 0.8 mm skinned pipes were annealed in a fan oven at 80[degrees]C for different time intervals, and cooled at room temperature. Residual stresses were then measured using the ring slitting method , in which rings sliced from the pipe are slit axially, releasing residual stresses causing local bending. Having measured the pipe diameter before and after slitting, the bending moment is calculated and thus the residual strains and stresses. Peel tests were also performed to determine the adhesion between skin and core. Because material was by then in short supply, critical temperatures were not determined by the multi-specimen procedure specified in ISO 13477. Instead, tests were performed at -18[degrees]C (compared to [T.sub.c-S4] = -10[degrees]C for "as received" 0.8 mm skinned pipes) and crack lengths were measured. Results are shown in Table 2.
With increasing annealing time, measured residual stresses decrease while adhesion increases. After 120 hr, residual stresses were reduced by 40% while adhesion was doubled. A relatively small increase in elastic modulus (~2%) was found for pipes annealed for 5 days. This increase in adhesion with annealing time has been reported previously for PE-PP interfaces  and blends  and is attributed to the thermodynamical motion of the polymer chains, which allows them to interdiffuse through the interface and entangle with each other. The longer the annealing time, the higher will be the entanglement density and the greater the adhesion. Table 2 also shows that crack length decreased with annealing time (i.e., RCP resistance improved), which means that the positive effect of residual stress relaxation on RCP performance was more significant than that of increasing adhesion and constraint. This improvement in RCP performance by relieving residual stresses during annealing has been reported previously [20-22] for solid pipes (no skin). It has been suggested that during annealing, residual stresses relax, reducing the stored strain energy in the pipe wall that could help to drive RCP.
S4 critical temperatures were measured for 0.4 mm skinned pipes which were annealed at 80[degrees]C for 4 and 24 hr. Results are shown in Table 3. Again, residual stresses decrease while adhesion increases with annealing time. [T.sub.c-S4] decreased significantly (from -10 to -19[degrees]C) for the 24 hr annealed pipe, reinforcing (he conclusion that residual stress relaxation has a more significant effect than that of increasing constraint due to stronger adhesion.
Surface Heal Treatment to Modify Interfacial Adhesion
To investigate the effects of residual stresses and interfacial adhesion on RCP performance separately, a surface heat treatment rig , in which the pipe is mounted on a spit rod between two rows of infrared heaters, was used to modify the adhesion between the skin and the core. The distance between the heaters and the pipe surface was set to 100 mm. and the pipes were treated for 5 and 10 min. After the heat treatment, the pipes were left to cool down at room temperature. Residual stresses in the core pipe remained almost unchanged (Table 4) while the interfacial adhesion was modified by ~ 12-25%. For 0.4 mm skinned pipes the adhesion increased but for 0.8 and 1.5 mm skinned pipes it decreased.
A decrease in adhesion with a similar heat treatment was reported by Munch , who treated molded bilayered laminates with infrared heaters approximately 100 mm above their surface. As mentioned previously, higher temperatures promote the mobility of polymer chains, allowing them to interdiffuse more rapidly and entangle to increase the interfacial adhesion. This is the case of annealed specimens, where the pipes are left at high temperatures for long periods of time. However, during this surface heat treatment the 0.8 and 1.5 mm skins are preventing the core pipe to reach a temperature high enough for interdiffusion to take place, only the PP skin molecules are moving and disentangling and thus there is a decrease in interfacial adhesion. In the case of a 0.4-mm skin the heat is able to reach the core pipe surface and interdiffusion is promoted increasing the adhesion.
S4 tests were performed on the heat treated pipes to find the critical temperature [T.sub.c-S4]. Results are shown in Table 4. It is observed that in general, the lower the adhesion, the lower [T.sub.c-S4], except for the 0.4-mm skinned pipe treated for 10 min in which [T.sub.c-S4] slightly decreased with a small increase in adhesion. This can be an effect of the relatively small residual stress relaxation, which as suggested in the previous section, might be outweighing the effect of interfacial adhesion.
The effect of adhesion and residual stresses on the RCP performance of skinned pipes was investigated through a series of S4 tests on pipes with different thermal history: "as received," annealed, and heat treated to modify the interfacial adhesion between the skin and the core. It was found that the addition of a thin skin (0.4-1.5 mm) increased the critical temperature of the pipe by 3-4[degrees]C. Tensile tests revealed that the three skins have similar elastic modulus before and after the heat treatments. Therefore, modulus inequality between skin and core was not a significant variable in the analysis as all the pipes have the same level of inequality.
Adhesion was investigated as it was expected that higher adhesion levels enhance the constraint effect of the skin over the core, promoting plane strain conditions and increasing the critical temperature. Peel tests showed that adhesion of the skin to the core increased with its thickness. However, for the 1.5-mm skinned pipe, with higher adhesion levels, a better RCP performance was found, which was attributed to its lower residual stresses.
Annealing decreased residual stresses and increased the interfacial adhesion as high temperatures promote chain motion which promotes interdiffusion across the interface. The RCP performance of these pipes improved (9[degrees]C for 0.4 mm skinned pipes), an indication that the effect of relaxing residual stresses by approximately 40% had a greater impact than that of doubling the adhesion.
However, tests on surface heat treated pipes showed that when only the adhesion is modified, it indeed plays an important role in the RCP pipe performance, with pipes with lower adhesion having lower critical temperature.
[1.] W. Elzink, "Trenchless Replacement with Dedicated PE pipes," in Trenchless Middle East 2007 Conference, Dubai (2007)
[2.] P. Vanspeybroeck, "Evaluation of Electrofusion Welders." in Plastics Pipes XIII, Washington, DC (2006).
[3.] M. Troughton and C. Brown, "Comparison of Long-Term and Short-Term Tests for Electrofusion Joints in PE Pipes," in Plastics Pipes XIII, Washington, DC (2006).
[4.] S. Wood, L. Richards, R. Street, D. Muckle, and J. Bowman. "Second Generation Skinned Pipes with Enhanced Fracture Resistance," in Plastic Pipes XIII, Washington, DC (2006).
[5.] D. Harget and J. Bowman, Pipeline Gas J., 228, 12 (2001).
[6.] G. Palermo and J. Bowman, Considerations and Thermochemical Kinetic Simulation of Initial Product-Forming Pathways, Vol. 3, American Gas Association, Washington, DC (2010).
[7.] P.S. Leevers and L. Moreno. Eng. Fract. Mech., 72, 947 (2005).
[8.] L. Moreno and P.S. Leevers, Polym. Eng. Sci., 44, 1627 (2004).
[9.] S. Munch. Influence of Inteifacial Adhesion on the Fracture Resistance of Polymer Multilayers, Imperial College London. London, UK (2007).
[10.] A.G. Atkins and Y.M. Mai, Int. J. Fract., 12, 923 (1976).
[11.] R.W. Truss, R.A. Duckett, and I.M. Ward. Polym. Eng. Sci., 23, 708 (1983).
[12.] P. So and L.J. Broutman, Polym. Eng. Sci., 22, 888 (1982).
[13.] C. Verpy, J.L. Gacougnolle, A. Dragon. A. Vanlerberghe. A. Chesneau. and F. Cozette, Prog. Org. Coat., 24, 1 (1994).
[14.] British Standard, Plastics: Determination of Tensile Properties. Part 2: Test Conditions for Moulding and Extrusion Plastics. International Organization for Standardization, Geneva, Switzerland. ISO 527-2.
[15.] D R. Moore and J.G. Williams, ESIS Protocol: A Protocol for Determination of the Adhesive Fracture Toughness of Flexible Laminates by Peel Testing: Fixed Arm and T-Peel Methods, Imperial College London, London, UK (2006).
[16.] A.J. Kinloch, C.C. Lau, and J.G. Williams, Int. J. Fract., 66, 45 (1994).
[17.] J.G. Williams, J.M. Hodgkinson, and A. Gray, Polym. Eng. Sci., 21, 822 (1981).
[18.] G.X. Lin, W. Wenig, and J. Petermann, Die Angew. Makromol. Chem., 255, 33 (1998).
[19.] J. Li, R.A. Shanks, and Y. Long,./. Appl. Polym. Sci., 76, 1151 (1999).
[20.] R.K. Krishnaswamy and M.J. Lamborn, Adv. Polym. Technol., 24, 3 (2005).
[21.] L.B. Morgan, Pliilos. Trans. R. Soc. London Ser. A, 247, 13 (1954).
[22.] A. Guevara-Morales and P.S. Leevers, Polym. Eng. Sci., 53, 6 (2013).
Andrea Guevara-Morales, (1) Patrick Leevers (2)
(1) Escuela de Ingenieria y Cienclas, Tecnologico de Monterrey, Atizapan de Zaragoza 52926, Mexico
(2) Mechanical Engineering Department, Imperial College London, London SIV7 2AZ, UK
Correspondence to: A. Guevara-Morales; e-mail: firstname.lastname@example.org
Published online in Wiley Online Library (wileyonlinelibrary.com).
Caption: FIG. 1. (a) Unloaded through-thickness crack in a multilayer plate, (b) cross-section of crack plane near front, (c) plane stress core deformation on loading of crack front, and (d) axisymmetric bending to restore skin-core contact .
Caption: FIG. 2. (a) Schematic of a fixed arm peel test and (b) sample holder for peel tests.
Caption: FIG. 3. "Zero-adhesion" bilayered pipe.
TABLE 1. S4 critical temperature ([T.sub.c-S4]), residual stresses ([[sigma].sub.[theta]]), and adhesive fracture energy for 110/11 skinned PE100 pipes. Pipe Skin thickness [T.sub.c-S4] (mm) ([degrees]C) Solid -- --14 Skinned 0.4 --10 0.8 --10 1.5 --11 Skinned 1.5 --16 (zero-adhesion) Pipe [[sigma].sub.[theta]] Adhesive fracture (MPa) energy (J/[m.sup.2]) Solid 2.21 -- Skinned 1.59 161 1.56 216 1.47 245 Skinned 1.47 0 TABLE 2. Crack length, residual stresses and adhesive fracture energy for annealed 0.8 mm skinned pipes at -18[degrees]C. Hours Crack length/ [[sigma].sub.[theta]] Adhesive fracture annealed diameter (MPa) energy (J/[m.sup.2]) 0 6.13 (P) 1.56 216 4 2.59 (A) 1.48 231 7 1.45 (A) 1.36 280 24 1.36 (A) 1.21 320 48 1.00 (A) 1.19 322 72 2.36 (A) 1.12 343 96 0.95 (A) 1.01 407 120 1.04 (A) 0.95 448 P: crack propagation; A: crack arrest. TABLE 3. S4 critical temperature, residual stresses, and adhesive Fracture energy for 0.4 mm skinned pipes annealed al 80[degrees]C for various limes. Hours [T.sub.c-S4] [[sigma].sub.[theta]] Adhesive fracture annealed ([degrees]C) (MPa) energy (J/[m.sup.2]) 0 -10 1.59 161 4 -13 1.47 219 24 -19 1.29 278 TABLE 4. S4 critical temperature, residual stresses, and adhesive Fracture energy for 110/11 skinned pipes after a surface heat treatment to modify the adhesion between the skin and the pipe. Heat [T.sub.c-S4] treatment ([degrees]C) Pipe time (min) 0.4 0 --10 5 --8 10 --9 0.8 0 --10 5 --10 10 --11 1.5 0 --11 5 --13 10 --13 [[sigma].sub.[theta]] (MPa) Pipe 0.4 1.59 1.56 1.51 0.8 1.56 1.51 1.44 1.5 1.47 1.42 1.35 Adhesive fracture energy Pipe (J/[m.sup.2]) 0.4 161 198 205 0.8 216 198 183 1.5 245 212 187
|Printer friendly Cite/link Email Feedback|
|Author:||Guevara-Morales, Andrea; Leevers, Patrick|
|Publication:||Polymer Engineering and Science|
|Date:||Apr 1, 2017|
|Previous Article:||Crystallization and component properties of polyamide 12 at processing-relevant cooling conditions.|
|Next Article:||Development and properties evaluation of bio-based PLA/PLGA blend films reinforced with microcrystalline cellulose and organophilic silica.|