- Email: [email protected]
Thermorheological behavior of peroxide-induced long chain branches linear low density polyethylene
Accepted Manuscript Title: Thermorheological behavior of peroxide-induced long chain branches linear low density polyethylene Author: M. Golriz H.A. Khonakdar J. Morshedian PII: DOI: Reference:
S0040-6031(14)00332-3 http://dx.doi.org/doi:10.1016/j.tca.2014.07.010 TCA 76948
To appear in:
Thermochimica Acta
Received date: Revised date: Accepted date:
16-4-2014 23-6-2014 10-7-2014
Please cite this article as: M.Golriz, H.A.Khonakdar, J.Morshedian, Thermorheological behavior of peroxide-induced long chain branches linear low density polyethylene, Thermochimica Acta http://dx.doi.org/10.1016/j.tca.2014.07.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Thermorheological behavior of peroxide-induced long chain branches linear low density polyethylene M. Golriza, H.A. Khonakdarb*, J. Morshediana Iran Polymer and Petrochemical Institute, 14965/115, Tehran, Iran
b
Leibniz Institute of Polymer Research, D-01067 Dresden, Germany
ip t
a
cr
*Corresponding Author: H.A. Khonakdar Email: [email protected] Tel.: +49-351-
us
4658647
Highlights
an
• Performed a correlations between thermorheological behaviour and induced branching
M
in LLDPE
• Determined the changes in gel content, Mw and MWD in terms of processing
d
conditions.
Ac ce pt e
• Found good agreement between induced microstructural changes with SEC and thermorheology data
• Potential of thermorheological approach for analyzing branching in modified PE was ed.
Abstract
In this article, the correlation between thermorheological behaviour and molecular structure of linear low density polyethylene upon peroxide modification was explored. For this purpose, a commercial grade LLDPE (Exxon MobileTM LL4004EL) was reacted with different amounts of dicumyl peroxide (DCP) and their viscoelastic behavior were studied. Moreover, the samples were analyzed by size-exclusion chromatography coupled with a light scattering detector. Increasing the DCP content at roughly constant molar mass led to broadening of Page 1 of 18
molecular weight distribution as well as increasing the number of long-chain branches. The latter consequently resulted in enhanced activation energy and delayed relaxation times of the LLDPE. The thermorheological behaviour of peroxide modified samples was investigated and the results showed that the induced long chain branching changes the thermorheological behaviour from simple to complex. The plotted results (activation energy versus phase angle) showed constant activation energy at low peroxide level (i.e. increasing the long chain
ip t
branching) (up to 125 ppm) and a distinct variation from low to high phase angle with increasing the peroxide level. The theromorheological complexity threshold could be
cr
determined using these plots. The potential of thermorheological approach as an alternative powerful tool for analyzing LCB issue in peroxide modified LLDPE could be highlighted.
us
Keywords: Thermorheological behavior, LLDPE; Peroxide modification; Long-chain branching.
an
Introduction
A glance at the current literature witnesses that modification of polyolefins (PE) has been the
M
subject of many investigations due to its importance from economical and processing viewpoints. In principle, modification of PE with organic peroxide is associated with an
d
insoluble molecular network or gel, which affects the level of chain-linking. There exist only
Ac ce pt e
limited publications on the modification of PE using peroxide concentrations below the critical gel formation concentration. In such situation, the modified PE still behaves like a thermoplastic. However, the molecular structure of the PE changes in such a manner that the average molecular weight follows an ascending trends, at the same time its distribution becomes slightly broader, altering its properties in the molten and solid states [1–8]. Rheological investigation is a key in achieving useful information about events at molecular level. Accordingly, determination of degree of long chain branching (LCB) of polymers is essential for understanding their rheological response and optimizing their processing behaviour [9-16].
Typically, a very low level of LCB has a significant impact on the processability of polymers, especially their melt strength. As a result, different authors have tried to exploit this rheological effect to characterize LCB [17]. The correlations established between degree of LCB and rheological response can be expressed as a fast and robust method for monitoring the microstructural changes during reactive processing of linear low density polyethylene (LLDPE).
Page 2 of 18
In general, the rheological response of PEs is influenced by the type, content, and structure of branches. So far, diverse rheological methods have been examined and revealed that LCBmPEs feature a higher temperature dependence than linear ones, but the introduction of LCBs also leads to the failure of the time temperature superposition principle, i.e., to a thermorheological complexity. For LCB-PEs, a drop in the activation energy (Ea) together with an increase in the storage modulus is characteristic of thermorheological complexity [18-
ip t
19].
One way to distinguish the complex thermorheological behavior from the simple behavior
cr
would be determination of Ea. In the case that Ea shows no dependency on frequency, the modulus and phase angle of master curve can be explained in view of shifting in viscoelastic
us
properties at different temperatures. This is a sign of simple thermorheological behavior, which is the case seen for linear PE. Increase of long-chain branching in linear PEs caused by
an
dependency of Ea on frequency, modulus, and phase angle establishment of master curve seems impossible.
The activation energy of linear PEs takes the minimum value and will be affected by the long
M
or short chain branches (SCBs). The activation energy of high density PE (HDPE) in the melt is estimated to be around 26-28 kJ/mol, while it is slightly higher (30–34 kJ/mol) for linear
d
PEs containing SCBs. For LLDPE, the Ea takes the value of 34 kJ/mol. An increase in the
Ac ce pt e
LCB content possesses stronger effects on the activation energy, rather than the increase in the SCB content. The maximum activation energy of about 65 kJ/mol has been reported for LDPE, which has a great amount of LCBs. In general, the activation energy is independent of molecular weight (Mw) and molecular weight distribution (MWD), therefore the higher activation energy values of LDPE can be attributed to the presence of more LCBs that slow down the segmental motions[20-22].
For linear polyethylene, a simple thermorheological behavior was found, which means Ea is constant. Accordingly, Ea is independent of frequency or modulus, shear rate and relaxation time and therefore the determined viscoelastic properties in different temperatures can be shifted toward each other to obtain a master curve [23-25]. In the other words, the accuracy of time-temperature superposition represents the simple thermorheological behavior [25]. In some cases, the effect of temperature cannot be described by a single time-shift factor and different shift factor is required corresponding to times. As a result, the time-temperature superposition is no longer applicable. This kind of materials is called thermorheologically complex. The materials containing LCBs often shows the simple/complexity theromorheological behaviour. [26]
Page 3 of 18
In recent years, lots of attentions were focused on the influence of branching on the rheological and thermorheological properties of PE and its blends. Investigations on the PEs reveal that LCBs have stronger effects on the rheological properties and more rheological temperature dependency as compared to SCBs. Rheological response of PEs depends on the type, content, and structure of branches. For LCB-PEs, a decrease in activation energy (Ea) with an increase in storage modulus was reported which is considered as a sign for
ip t
thermorheological complexity. [27-35]
Dordinejad et al. [36] performed different analytical approaches based on rheological
cr
measurements on two grades of neat m-LLDPE samples to define thermorheological behavior for the assigned system. In the other work [37], they focused on some model blend systems
us
based on PE in order to examine the applicability and ability of diverse analytical techniques in defining thermorheological behavior of the blends and their relation with branching
an
structure. Elsewhere [38], they qualitatively assessed long chain branching content in the LLDPE, LDPE and their blends via thermorheological analysis.
Golriz et al. [39], proposed a correlation between reactive modification conditions and degree
M
of LCB in chemically-modified LLDPE. Moreover, by the use of a response surface method, correlations between the processing parameters and degree of LCB was established. In
d
continuation [40], the Monte Carlo (MC) simulation could successfully monitor the
Ac ce pt e
microstructural changes due to the induced LCB, decreased terminal double bonds, and increased trans vinyl bonds in LLDPE during peroxide modification. Elsewhere [41], they studied the rtheological assessment of variable molecular chain structures of LLDPE under reactive modification. By the use of a simple rheological method that uses melt rheology, and is reasonable and consistent with estimations on the degree of LCB (x) and the volume fraction of the various molecular species produced during peroxide modification of LLDPE. The main objective of the current work is to investigate and define thermorheological behavior of peroxide modified LLDPE and degree of imposed LCB, which is one of the most important characteristics of polyethylene reflecting changes in molecular chain structure upon peroxide modification. In particular, we report on the potential of thermorheological approach as an alternative powerful tool for analyzing LCB in the peroxide modified LLDPE.
2. EXPERIMENTAL 2.1. Materials A commercial LLDPE (LL4004EL), with a melt flow index of 3.6 g/10 min (190 °C,2.16 kg) and a density of 0.924 g/cm3 (20°C), was purchased from ExxonMobil Chemical. The
Page 4 of 18
average molecular weight (Mw) of this grade of LLDPE was about 95,000 g/mol. DCP (molecular weight 270 g/mol) and xylene (a mixture of o-, m-, and p-xylene with boiling point of 140°C and density of 0.87 g/cm3 at 20°C) were chemically pure and were obtained from Aldrich Chemical Co. 2.2. Modification procedure LLDPE was melt compounded with various amounts of peroxide i.e. 0, 15, 125, 500, 700,
ip t
1000, (ppm) using a DACA twin-screw Microcompounder (DACA Instruments, Goleta, CA, USA) with a screw speed of 100 rpm and temperature of 185 °C within various residence
cr
times. The preparation conditions are shown in Table 1.
Mixing Time (min)
an
0 15 125 500 700 1000 700
3 3 3 3 3 3 7.5
Ac ce pt e
d
LL/0 LL/15 LL/125 LL/500 LL/700 LL/1000 LL/700/7.5
Peroxide Concentration (ppm)
M
Sample Code
us
Table 1- Preparation conditions of peroxide modified LLDPE at various residence times and peroxide contents
2.3. Gel analysis
The most highly modified LLDPE samples i.e. the samples treated with high concentarations of DCP were examined for content of gel according to ASTM 2765. Based on this standard the extraction in boiling xylene was the basis of all the analyses.
2.4. Size exclusion chromatography (SEC) Average molecualr weights and molecular weight distributions (MWDs) were determined by high-temperature SEC coupled with a multi angle laser light scattering (MALLS) detector and a refractive index (RI) detector. The polymer molecules are fractionated in the SEC by their hydrodynamic volume, which depends on the density in the dissolved state, molar mass and degree of LCD. Therefore, the conventional SEC using linear polymer standards for the calibration is not suitable for investigations of the molar mass of branched polymer structures due to the fact that the calculated molar mass averages would be lower than the true values. [42] By coupling SEC with MALLS, the absolute molar mass MLS of every fraction can be determined directly. The SEC experiments were carried on a PL-GPC 220 (Polymer
Page 5 of 18
Laboratories) at 150 °C coupled with a MALLS (Heleos II, Wyatt Technology Corp., USA) and an RI detector. The column set consisted of two columns PL Mixed-B-LS (Polymer Laboratories). The eluent was TCB (1, 2, 4- trichlorobenzene, Merck) stabilized with BHT. The software used for data processing and for calculation of the number of LCB was Astra 5 (Wyatt Technology Corp., USA).
ip t
2.5. Rheometry
Rheological Measurements Parallel-plate rheometry was performed to determine the linear
cr
viscoelastic properties of peroxide modified LLDPE. The measurements were performed using a Paar-Physica Rheometer (MCR300, Ostfildern, Germany) in oscillatory shear mode
us
with parallel plates (25 mm in diameter with a gap of 1 mm) at a frequency range from 0.03 to 100 rad/s. The measurements were performed at four different temperatures, i.e. 150, 170,
an
190, and 210 °C with accuracy of ±0.5°C under N2 atmosphere. All the rheological
3. Results and Discussion 3.1. Gel content analysis
M
measurements were performed within linear viscoelastic region.
d
The gel content of peroxide modified LLDPE was measured by a solvent extraction
Ac ce pt e
technique. The gel content measurement results for all peroxide modified LLDPE revealed that the insoluble network structure fraction was negligible. This implies that incorporation of DCP to LLDPE predominately leads to branching rather than crosslinking. In other words, long-chain branches can be introduced into linear PE by peroxide modification, without gel formation.
3.2. Changes in Molecular Mass Distributions The MWDs of the LLDPE’s modified with different amounts of peroxide i.e., 0-, 125-, 500-, 700-, and 1000-ppm prepared at 185 °C for 3 min. are plotted in Figure 1. A reduction in distribution frequency together with the deviation of the distribution mode toward higher values implies that the dominant termination reaction mechanism is combination. It is seen that the changes in average molecular weight and molecular weight distribution (MWD) induced by peroxide modification under various conditions are small and there are no indications of formation of low-molecular-weight fractions due to chain scission. This fact leads to the conclusion that the degradation process is not predominant in the high molecular mass area of the polymer and long-chain branches can be introduced into linear PE by
Page 6 of 18
peroxide modification, without degradation and crosslinking. A slight increase of MWD is generally observed which can attriburtd to the occurrence of modification reactions, chain branching, and extension. These findings are in good agreement with the previous reports from Gloor et al. [43] and Pedernera et al. [44] on the modification of PE with low amounts of peroxide. When the aim is the formation of a three dimensional, insoluble network, the extent of peroxide is selected to be beyond 2000 ppm [43-45]. However, in this work, the highest
ip t
peroxide concentration used is 1000 ppm, which is well below the critical level for insoluble
Ac ce pt e
d
M
an
us
cr
network formation.
Figure 1. MWDs (dW/d log M) for the initial LLDPE (LL/0) and DCP modified LLDPEs processed at 185 °C for 3 min.
3.3. Rheological Studies
Effect of Peroxide Modification on Viscoelastic Behaviour of LLDPE As it was discussed earlier, since MWD is not affected significantly after modification, therefore the changes observed in the rheological behavior can be attributed to the changes in molecular architectures from linear to branched structure. It is worth to mention that the extent of degradation of the samples and the respective reporoducibility have been examined by means of dynamic tests. Figure 2 reveal the variation of G’(ω) and G”( ω) as a function of frequency for the DCP modified samples. As it is clearly seen, on increasing the DCP concentration, the magnitude of both moduli increases, in particular that of G’(ω) is more pronounced. Concerning the
Page 7 of 18
0
plateau of G’(ω), one can realize that G N
is almost independent of the extent of branching,
whereas the terminal region is considerably affected by branching. The similar investigation was done by Park et al [46]. They used a metallocene-based polyethylene which contained various extents of the long chain branches. The trend observed in his works are very similar to the one reported in the present work.
ip t
An enhancement of the moduli due to the branching indicates that the contribution of the end groups motions to the stress relaxation is not dominant and therefore they are low in number. It is then concluded that the branches should be of branching configuration. Moreover,
cr
increasing the peroxide content in LLDPE, results in increasing the activation energy of the
us
peroxide-modified LLDPE. As it is known the activation energy is independent of molecular weight (Mw) and molecular weight distribution (MWD), therefore the higher activation energy values of modified LLDPE with higher peroxide content can be due to the presence of
an
more branches. This increase in activation energy at higher branch content can be related to the slowed segmental dynamics [23]. The activation energy is representative of the potential
M
energy from the flow of a molten polymer, therefore introducing branches retards the overall
Ac ce pt e
d
dynamics, and consequently higher activation energy is required for the segmental motion.
Figure 2. a) Effect of DCP modification on storage modulus and b) loss modulus of initial LLDPE and the samples modified with different amounts of peroxide processed at 185 °C for 3 min.
In our previous work [39] we studied the correlations between rheological behaviour and degree of long chain branching (LCB) of LLDPE upon a peroxide (DCP) modification process under various conditions. It was reported that at low frequencies, the viscosity of the modified samples is greatly affected by peroxide modification and the zero-shear rate viscosity (η0) values of modified LLDPE is higher than that of the neat LLDPE. Therefore, peroxide modification of the LLDPE increases the dynamic viscosity of the samples. In the
Page 8 of 18
other words, the η0 values increase with increasing of long chain branching. As was discussed earlier, the molecular weight and MWD of modified LLDPE did not significantly change, therefore this deviation could be attributed to structural changes in LLDPE induced by long chain branching. The degree of long-chain branching (LCB) was evaluated based on the zero shear rate viscosity data and the results in the form of van Gurp-Palmen plots are presented in Figure 3.
ip t
A plot of the phase angle δ versus the complex shear modulus |G*(w)|, the so-called van Gurp-Palmen plot, is frequently used to get an insight on long-chain branching. This method
cr
is rather sensitive in molecular structural change. The influence of peroxide concentration and time of mixing on phase angle δ of samples is shown in Figure 3. As it is seen from the figure
us
with increasing peroxide level from 200ppm (L1/185/3) to 1200ppm (L3/185/12) and time of mixing from 3(L1/185/3) to 12 min. (L1/185/3) the phase angle δ decreases. In the medium
an
range of the complex modulus between rubbery plateau region and the Newtonian region a deviation of the phase angle δ beneath the curve of linear LLDPE is observed for the samples modified with 200ppm peroxide and higher. A shift of the phase angle towards small values at
M
fixed |G*(w)| can be caused by a broader molar mass distribution or an introduction of longchain branching. In this work for the peroxide modified LLDPE the decrease is only attributed
d
to the introduction of long-chain branching since the polydispersity remains approximately
Ac ce pt e
constant according to the SEC analysis.
Figure 3. δ(|G*|) -plot of the peroxide modified LLDPE Thermorheological Investigation Page 9 of 18
As mentioned earlier, materials obeying time-temperature superposition principle have "simple thermorheological" behavior. The shift factor, aT, facilitates superposition of dynamic moduli measured at different temperatures. However, the effect of temperature cannot be simply depicted via aT for long chain branched (LCB) polymers. Due to existence of molecules of different length and number of branches LCB polymers are thermorheologically complex and a spectrum of activation energy, Ea(λ), is necessary for description of their
ip t
behavior. Ea increases with increasing LCB. Vega et al. suggested the following equation to establish a relation between the long chain branching and the activation energy for the linear
us
E a ,b − E a , L Ea ,L
(5)
an
I LCB =
cr
and branch polymers [13].
It is to be noted that in the literature, activation energy values of 24–26 kJ per mole for linear polyethylene have been reported [20]. Therefore, for linear polyethylene Ea,L value assumed
M
with Ea,L of HDPE (Ea=24kJ/mol.).
As it is clear, the virgin sample in this work has linear structure and thus expected to have
d
simple thermorheological behavior. But our structure modification brings about complexity to the sample behavior. In this section, attempt is made to obtain the threshold amount of
Ac ce pt e
peroxide to induce thermorheological complexity. Thus temperatures of 150, 170 and 190o C for the behavior study, and 190oC as the reference temperature for determining the horizontal and the vertical shift factors were selected.
To compare the relaxation behavior of modified LLDPE, the weighted relaxation spectra evaluated from the linear viscoelastic data using the US200 software and the shift factors have been definitely optimized.
The curves of the horizontally and vertically shifted storage and loss moduli versus frequency for the modified samples with 15 to 1000 ppm peroxide at 180oC for 3 minutes are depicted in figures 4 to 5. The more observed non-superimposibility of the curves at difference temperatures, the more complexity indication of the thermorheological behavior. It is observed that increasing amount of peroxide from 15 to 1000 ppm intensifies this behavior. Figure 4-5 shows the phase angle of the modified samples with 15 to 1000 ppm peroxide at 180oC for 3 minutes and the neat LLDPE as a function of reduced frequency. The curves were shifted along the x-axis by multiplying the values with the coefficients shown on the diagram for better visualization. Figure 4 and 5 indicates that there is a good superposition to obtain a
Page 10 of 18
master curve resulting in simple thermorheological behavior for LLDPE. Nevertheless, there is obvious non-superposition for other samples that makes it impossible to obtain a master curve. The more observed non-superimposibility of the curves at difference temperatures, the more complexity indication of the thermorheological behavior. It is observed that increasing amount of peroxide from 15 to 1000 ppm intensifies this behavior. In other words, the reason behind this fact that the curves corresponding to the different temperatures can not be
ip t
superimposed on each other (after applying shift factors) is indeed due to the complicated thermorheological behavior. As the concentration of peroxide is raised and long chain
cr
branching thereby increases, the thermorheological behavior becomes more and more sophisticated. This results in further lack of superposition of the curves which is clearly
Ac ce pt e
d
M
an
us
visible in Figures 4 and 5.
Figure 4. Master curves of the shifted storage modulus (bT G’) and shifted loss modulus (bT G’’) as a function of the reduced frequency (aTω) for modified LLDPE with 15 ppm peroxide at temperature of 180 °C for 3 min.
Page 11 of 18
ip t cr us
an
Figure 5. Master curves of the shifted storage modulus (bT G’) and shifted loss modulus (bT G’’) as a function of the reduced frequency (aTω) for modified LLDPE with 1000 ppm peroxide at temperature of 180 °C for 3 min.
Furthermore, curves of phase angle δ against reduced frequency for peroxide modified samples (accroding to Table 1) are depicted in figures 6 and 7.
M
Figure 6 and 7 shows the phase angle of peroxid modified LLDPE and the neat LLDPE as a function of reduced frequency. It is observed that with increasing peroxide (increasing LCB)
d
the curves shift to smaller values of δ. Figure 7 compares δ-frequency curves for samples P700/185/3 and P700/185/7.5 distinguished only in mixing time. For enhanced unambiguity,
Ac ce pt e
curves of different sample are shifted by some factors along the frequency axis.
Figure 6. Phase angle δ as a function of the reduced frequency (aT.ω) at different temperatures (150, 170 and 190 °C) for modified LLDPE mixed for 3 min (the curves were shifted along the ω-axis by the factors indicated, for the matter of a better visualization) (T0 =150°C).
Page 12 of 18
ip t cr us
an
Figure 7. Phase angle δ as a function of the reduced frequency (aT.ω) at different temperatures for two modified LLDPE (P700/185/3 and P700/185/7.5) (the curves were shifted along the ω-axis by the factors indicated, for the matter of a better visualization) (T0 =150°C). A close look at the samples behavior in figure 6, one reveals that a concavity change occurs
M
from the third sample, i.e. above 125 ppm in peroxide content, illustrating simple-to-complex threshold in thermorheological behavior of the samples. This behavior has been observed in
d
the graphs for moduli and complex viscosities of the samples. Figure 6 well illustrates the
Ac ce pt e
simple-complex change in behavior of the samples. Similar works by Keßner et al. in 2010 [32], and Stadler et al. in 2007 [15] have been made on thermorheology of metallocene LLDPE with sparse long branches as well as LDPE. In figure 6, samples 4 and 5 exhibit a behavior similar to metallocene samples with sparse long branches and sample 6 shows a behavior similar to LDPE. To evaluate the thermorheological behavior in greater detail, the activation energy curves as functions of phase angle (δ) can be used. To establish the activation energy curve, first determine frequency at constant phase angles and then obtain shift factor for each value of δ with respect to reference temperature of 190oC. The shift factor curve against reciprocal temperature gives activation energy values for each δ. Thus activation energy curve as a function of δ is established. According to Keßner and Münstedt method [32] logarithms of the time-scale shift factors at a constant δ are plotted against the reciprocal absolute temperature 1/T. By using Arrhenius relationship for shift factors, the activation energies of peroxide modified LLDPE (15 to 1000 ppm peroxide) and the neat LLDPE are plotted as a function of δ in Figure 8. It is interesting to see the difference between activation energies of samples containing more and less than 125 ppm peroxide. As demonstrated in figure 6, the curves of
Page 13 of 18
these samples are different in concavity, although curves of activation energy reveal this difference more clearly. Indeed in case of simple thermorheological behavior, curve of activation energy against any change of phase angle remains unchanged (linear polyethylenes without any long branches show this type of behavior) [15,35,25, 21]. However, for complex thermorheological behavior activation energy change is observable. LLDPE exhibits constant activation energy. Therefore, based on this method a simple
ip t
thermorheological behavior is concluded for LLDPE. This is similar to the result reported by Keßner et al. [32, 34] LLDPE and LLDPE/LDPE blends show phase angle dependent
cr
activation energies that indicates complex thermorheological behaviour which is attributed to the presence of long branches. The reason for the observed simple thermorheological behavior
us
of LLDPE containing high amount of SCB can be the presence of short branches in each LLDPE molecule and similarity of the structures of these branches. [33]
an
At high values of phase angle ,we observe an activation energy of about 60 pertaining to LCB where as at low values of phase angle, an activation energy of about 25-30 due to linear chains of PE is observed. With regard to figure 8, for the linear LL sample, LL/15 and LL/125
M
the activation energy is almost constant but with increasing peroxide (hence increasing LCB), for samples LL/500 to LL/1000, the activation energy modifies from high values (at large
d
phase angle) to low values (at small phase angle). Somehow one may assume such samples to
Ac ce pt e
be blends of linear chains and chains containing long branches.
Figure 8. Activation energies as a function of the phase angle for peroxide modified LLDPE
Page 14 of 18
In a similar manner one can deduce curve of activation energy against elastic modulus. Figure 9 shows the activation energy of modified LLDPE and neat LLDPE as a function of storage modulus according to Wood-Adams and Costeux method [21]. In these curves, also, if the thermorheological behavior is simple, then the activation energy is nearly constant, and one has the complex behavior, the activation energy alters (decreases) with increasing of elastic modulus. For all samples, dependency of activation energy on the storage modulus was
ip t
evident. This indicates thermorheological complexity of the samples based on this method. This finding is in agreement with the results reported by Keßner and Münstedt [32]. For all
cr
samples, activation energy decreases with the increase in storage modulus. Furthermore, the decreasing rate of the activation energy increases with increase in the long branch content
Ac ce pt e
d
M
an
us
(higher peroxide content).
Figure 9. Activation energies as a function of the storage modulus for peroxide modified LLDPE at 150, 170, 190 °C Conclusion The correlations between microstructural changes induced in a linear low density polyethylene (LLDPE) upon peroxide modification with corresponding thermorheological responses were explored. Chromatography results showed that with increasing peroxide content molecular weight distribution broadened a little and long chain branches increased.
Page 15 of 18
The gel content analysis revealed that the insoluble crosslinked fraction was negligible in all the modified samples. This implies that incorporation of DCP to LLDPE predominately leads to branching rather than crosslinking. Results showed that both G'(ω) and G''(ω) increased with peroxide content, although variation of G'(ω) was more pronounced. With regard to plateau section of G'(ω) curve, one can deduce that the plateau modulus, GoN, is nearly independent of branches presence, while effect of
ip t
branches are marked in the terminal zone. Results of shifted storage and loss moduli against frequency for 15 to 1000 ppm peroxide modified samples showed non-superimpossiblity of
cr
the curves with increasing the peroxide content, indicating pronounciation of complex thermorheological behavior. Results of activation energy as a function of phase angle showed
us
that at low levels of peroxide, the polymer enjoys of simple thermorheological behavior (constant activation energy) while at higher amounts from 125 to 1000 ppm, the behavior
an
converted to complex thermorheology (change in activation energy). Results of activation energy curves versus elastic modulus indicated change in activation energy with increasing the elastic modulus for peroxide modified samples. The amount of change is little at low level
M
of peroxide but the activation energy decreases considerably with increasing of peroxide. The threshold of sample-to-complex thermorheological behavior can be detected from these
d
curves.
Ac ce pt e
Acknowledgments: The author expresses his gratitude to the Alexander von Humboldt foundation for financial support. Also, technical assistance from A. Lederer, B. Kretzschmar and A. Ivanov is deeply appreciated.
References
[1] C. J. Perez, G. A. Cassano, E. M. Vallés, M. D. Failla, L. M. Quinzani, Polymer 2002, 43, 2711. [2] M. Pedernera, C. Sarmoria, E. M. Vallés, A. Brandolin, Polym. Eng. Sci. 1999, 39, 2085. [3] W. S. Lambert, P. J. Phillips, J. S. Lin, Polymer 1994, 35, 1809. [4] P. Sajkiewicz, P. J. Phillips, J. Polym. Sci., Part A: Polym. Chem.1995, 33, 949. [5] T. Bremner, A. Rudin, J. Appl. Polym. Sci. 1995, 57, 271. [6] T. K. Kang, C. S. Ha, Polym. Testing 2000, 19, 773. [7] H. Fang, Y. Zhang, J. Bai, Zh.Wang, Zh.Wang, RSC Advances, 2013, 3(23), 8783. [8] Y. Wang, L. Yang, Y. Niu, Z. Wang, J. Zhang, F. Yu, H. Zhang, J. Appl. Polym. Sci., 2011,122, 1857. [9] C. Gabriel, H. Münstedt, Rheol. Acta ,38,393-403,1999. [10] H. Xu, H. Fang, J. Bai, Y. Zhang, Zh. Wang, Ind. Eng. Chem. Res. 2014, 53, 1150-1159.
Page 16 of 18
[11] P. Lehmus, E. Kokko, O. Harkki, R. Leino, HJG. Luttikhedde, J.H. Nasman, J.V. Seppala, ,Macromolecules , 32,3547-3552,1999. [12] A. Malmberg, E. Kokko, P. Lehmus, B. Lofgren, J. Seppala, ,Macromolecules, 31, 84488454,1998. [13] J.F. Vega, A. Santamaria, A. Munoz-Escalona, P. Lafuente, Macromolecules , 31, 36393647,1998.
ip t
[14] Niu Y-H., Wang Z-G, Duan X-L, Shao W., Wang D-J., Qiu J., Journal of Applied Polymer Science, 2011, 119, 530.
[15] F.J. Stadler, A. Nishioka, J. Stange, K. Koyama, H. Münstedt, Rheol. Acta, 46,1003-1012, 2007.
Polyethylenes”, Dissertation, 2010.
us
[17] X. Chen, C. Costeux, R.G. Larson, J. Rheol. 2010, 54,1185.
cr
[16] U. Keßner, “Thermorheology as a Method to Investigate the Branching Structures of
[18] B.H. Zimm, W.H. Stockmayer, J. Chem. Phys. , 17, 1301-1314,1949.
an
[19] K. Klimke, M. Parkinson, C. Piel, W. Kaminsky, H.W. Spiess, M. Wilhelm, Macromolecular Chemistry and Physics, 207, 382-395, 2006.
[20] W.W. Graessley, Account. Chem. Res., 10, 332-339,1977.
M
[21] Wood-Adams P.M., Costeux S., Macromolecules, 34,6281-6290, 2001. [22] P.J. Rouse, J. Chem. Phys., 21, 1272-1279, 1953.
[23] F.J. Stadler, C. Gabriel, H. Münstedt, Macromol. Chem. Phys., 208, 2449-2454, 2007
d
[24] P.M. Wood-Adams, J.M. Dealy, Macromolecules, 33, 7481-7488, 2000.
Ac ce pt e
[25] U. Keßner, J. Kaschta, F.J. Stadler, S. Le DuffC, X. Drooghaag, H. Münstedt, Macromolecules, 43, 7341–7350, 2010.
[26] J. A. Resch, U. Keßner, F. J. Stadler, Rheol. Acta, 50,559-575, 2011. [27] P. M. Wood-Adams, S. Costeux, Macromolecules 2001, 34, 6281. [28] H. M. Laun, Prog. Colloid Polym. Sci. 1987, 75, 111. [29] D. Bonchev, A.H. Dekmezian, E. Markel, A. Faldi, J. Appl. Polym. Sci. 2003, 90, 2648. [30] D. J. Lohse, S.T. Milner, L.J. Fetters, M. Xenidou, N. Hadjichristidis, J. Roovers, R.A. Mendelson, C. A. Garcia-Franco, M.K. Lyon, Macromolecules 2002, 35, 3066. [31] E. Kokko, A. Malmberg, P. Lehmus, B. Lofgren, J.V. Seppala, J. Polym. Sci. Part A: Polym. Chem. 2000, 38, 376.
[32] U. Keßner, H. Munstedt, Polymer 2010, 51, 507. [33] F.J. Stadler, J. Kaschta, H. Munstedt, Macromolecules 2008, 41, 1328. [34] U. Keßner, J. Kaschta, F.J. Stadler, C.S. Le Duff, X. Drooghaag, H. Munstedt, Macromolecules 2010, 43, 7341. [35] F.J. Stadler, C. Gabriel, H. Munstedt, Macromol. Chem. Phys 2007, 208, 2449. [36] A. K. Dordinejad, S. H. Jafari, H. A. Khonakdar, U. Wagenknecht, G. Heinrich, J. Appl. Polym. Sci. 2013, 129, 458.
Page 17 of 18
[37] A. K. Dordinejad, S. H. Jafari, Polym. Eng. Sci. 2014,54:5,1081–1088. [38] A. K. Dordinejad, S. H. Jafari, J. Appl. Polym. Sci. 130: 3240–3250, 2013. [39] M. Golriz, H.A. Khonakdar, J. Morshedian, H. Abedini, S.H. Jafari, A. Lederer, U. Wagenknecht, Macromol. Mater. Eng. 2014, 299, 154–164. [40] M. Golriz, H.A. Khonakdar, J. Morshedian, S.H. Jafari, Y. Mohammadi, U. Wagenknecht, Macromol. Theory Simul. 2013, 22, 426–438.
ip t
[41] M. Golriz, H.A. Khonakdar, J. Morshedian, J. Appl. Polym. Sci. 2014, 131, 39617. [42] H.A. Khonakdar, Polymers for Advanced Technologies,2014, DOI: 10.1002/pat.3314. [43] P. E. Gloor , Y. Tang, A.E. Kostanska, A.E. Hamielec Polymer, 35 ,1012-1019, 1994.
cr
[44] M. N. Pedernera, M. Sarmoria Valles, A. Brandolin Polym. Eng. Sci., 39, 2085-2095., 1999.
Ac ce pt e
d
M
an
[46] S. J. Park, R.G. Larson, J. Rheol., 49, 523-536, 2005.
us
[45] A. E. Hamielec, P. E. Gloor, S. Zhu Can J. Chem Eng., 69, 611-617, 1991.
Page 18 of 18
S0040-6031(14)00332-3 http://dx.doi.org/doi:10.1016/j.tca.2014.07.010 TCA 76948
To appear in:
Thermochimica Acta
Received date: Revised date: Accepted date:
16-4-2014 23-6-2014 10-7-2014
Please cite this article as: M.Golriz, H.A.Khonakdar, J.Morshedian, Thermorheological behavior of peroxide-induced long chain branches linear low density polyethylene, Thermochimica Acta http://dx.doi.org/10.1016/j.tca.2014.07.010 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Thermorheological behavior of peroxide-induced long chain branches linear low density polyethylene M. Golriza, H.A. Khonakdarb*, J. Morshediana Iran Polymer and Petrochemical Institute, 14965/115, Tehran, Iran
b
Leibniz Institute of Polymer Research, D-01067 Dresden, Germany
ip t
a
cr
*Corresponding Author: H.A. Khonakdar Email: [email protected] Tel.: +49-351-
us
4658647
Highlights
an
• Performed a correlations between thermorheological behaviour and induced branching
M
in LLDPE
• Determined the changes in gel content, Mw and MWD in terms of processing
d
conditions.
Ac ce pt e
• Found good agreement between induced microstructural changes with SEC and thermorheology data
• Potential of thermorheological approach for analyzing branching in modified PE was ed.
Abstract
In this article, the correlation between thermorheological behaviour and molecular structure of linear low density polyethylene upon peroxide modification was explored. For this purpose, a commercial grade LLDPE (Exxon MobileTM LL4004EL) was reacted with different amounts of dicumyl peroxide (DCP) and their viscoelastic behavior were studied. Moreover, the samples were analyzed by size-exclusion chromatography coupled with a light scattering detector. Increasing the DCP content at roughly constant molar mass led to broadening of Page 1 of 18
molecular weight distribution as well as increasing the number of long-chain branches. The latter consequently resulted in enhanced activation energy and delayed relaxation times of the LLDPE. The thermorheological behaviour of peroxide modified samples was investigated and the results showed that the induced long chain branching changes the thermorheological behaviour from simple to complex. The plotted results (activation energy versus phase angle) showed constant activation energy at low peroxide level (i.e. increasing the long chain
ip t
branching) (up to 125 ppm) and a distinct variation from low to high phase angle with increasing the peroxide level. The theromorheological complexity threshold could be
cr
determined using these plots. The potential of thermorheological approach as an alternative powerful tool for analyzing LCB issue in peroxide modified LLDPE could be highlighted.
us
Keywords: Thermorheological behavior, LLDPE; Peroxide modification; Long-chain branching.
an
Introduction
A glance at the current literature witnesses that modification of polyolefins (PE) has been the
M
subject of many investigations due to its importance from economical and processing viewpoints. In principle, modification of PE with organic peroxide is associated with an
d
insoluble molecular network or gel, which affects the level of chain-linking. There exist only
Ac ce pt e
limited publications on the modification of PE using peroxide concentrations below the critical gel formation concentration. In such situation, the modified PE still behaves like a thermoplastic. However, the molecular structure of the PE changes in such a manner that the average molecular weight follows an ascending trends, at the same time its distribution becomes slightly broader, altering its properties in the molten and solid states [1–8]. Rheological investigation is a key in achieving useful information about events at molecular level. Accordingly, determination of degree of long chain branching (LCB) of polymers is essential for understanding their rheological response and optimizing their processing behaviour [9-16].
Typically, a very low level of LCB has a significant impact on the processability of polymers, especially their melt strength. As a result, different authors have tried to exploit this rheological effect to characterize LCB [17]. The correlations established between degree of LCB and rheological response can be expressed as a fast and robust method for monitoring the microstructural changes during reactive processing of linear low density polyethylene (LLDPE).
Page 2 of 18
In general, the rheological response of PEs is influenced by the type, content, and structure of branches. So far, diverse rheological methods have been examined and revealed that LCBmPEs feature a higher temperature dependence than linear ones, but the introduction of LCBs also leads to the failure of the time temperature superposition principle, i.e., to a thermorheological complexity. For LCB-PEs, a drop in the activation energy (Ea) together with an increase in the storage modulus is characteristic of thermorheological complexity [18-
ip t
19].
One way to distinguish the complex thermorheological behavior from the simple behavior
cr
would be determination of Ea. In the case that Ea shows no dependency on frequency, the modulus and phase angle of master curve can be explained in view of shifting in viscoelastic
us
properties at different temperatures. This is a sign of simple thermorheological behavior, which is the case seen for linear PE. Increase of long-chain branching in linear PEs caused by
an
dependency of Ea on frequency, modulus, and phase angle establishment of master curve seems impossible.
The activation energy of linear PEs takes the minimum value and will be affected by the long
M
or short chain branches (SCBs). The activation energy of high density PE (HDPE) in the melt is estimated to be around 26-28 kJ/mol, while it is slightly higher (30–34 kJ/mol) for linear
d
PEs containing SCBs. For LLDPE, the Ea takes the value of 34 kJ/mol. An increase in the
Ac ce pt e
LCB content possesses stronger effects on the activation energy, rather than the increase in the SCB content. The maximum activation energy of about 65 kJ/mol has been reported for LDPE, which has a great amount of LCBs. In general, the activation energy is independent of molecular weight (Mw) and molecular weight distribution (MWD), therefore the higher activation energy values of LDPE can be attributed to the presence of more LCBs that slow down the segmental motions[20-22].
For linear polyethylene, a simple thermorheological behavior was found, which means Ea is constant. Accordingly, Ea is independent of frequency or modulus, shear rate and relaxation time and therefore the determined viscoelastic properties in different temperatures can be shifted toward each other to obtain a master curve [23-25]. In the other words, the accuracy of time-temperature superposition represents the simple thermorheological behavior [25]. In some cases, the effect of temperature cannot be described by a single time-shift factor and different shift factor is required corresponding to times. As a result, the time-temperature superposition is no longer applicable. This kind of materials is called thermorheologically complex. The materials containing LCBs often shows the simple/complexity theromorheological behaviour. [26]
Page 3 of 18
In recent years, lots of attentions were focused on the influence of branching on the rheological and thermorheological properties of PE and its blends. Investigations on the PEs reveal that LCBs have stronger effects on the rheological properties and more rheological temperature dependency as compared to SCBs. Rheological response of PEs depends on the type, content, and structure of branches. For LCB-PEs, a decrease in activation energy (Ea) with an increase in storage modulus was reported which is considered as a sign for
ip t
thermorheological complexity. [27-35]
Dordinejad et al. [36] performed different analytical approaches based on rheological
cr
measurements on two grades of neat m-LLDPE samples to define thermorheological behavior for the assigned system. In the other work [37], they focused on some model blend systems
us
based on PE in order to examine the applicability and ability of diverse analytical techniques in defining thermorheological behavior of the blends and their relation with branching
an
structure. Elsewhere [38], they qualitatively assessed long chain branching content in the LLDPE, LDPE and their blends via thermorheological analysis.
Golriz et al. [39], proposed a correlation between reactive modification conditions and degree
M
of LCB in chemically-modified LLDPE. Moreover, by the use of a response surface method, correlations between the processing parameters and degree of LCB was established. In
d
continuation [40], the Monte Carlo (MC) simulation could successfully monitor the
Ac ce pt e
microstructural changes due to the induced LCB, decreased terminal double bonds, and increased trans vinyl bonds in LLDPE during peroxide modification. Elsewhere [41], they studied the rtheological assessment of variable molecular chain structures of LLDPE under reactive modification. By the use of a simple rheological method that uses melt rheology, and is reasonable and consistent with estimations on the degree of LCB (x) and the volume fraction of the various molecular species produced during peroxide modification of LLDPE. The main objective of the current work is to investigate and define thermorheological behavior of peroxide modified LLDPE and degree of imposed LCB, which is one of the most important characteristics of polyethylene reflecting changes in molecular chain structure upon peroxide modification. In particular, we report on the potential of thermorheological approach as an alternative powerful tool for analyzing LCB in the peroxide modified LLDPE.
2. EXPERIMENTAL 2.1. Materials A commercial LLDPE (LL4004EL), with a melt flow index of 3.6 g/10 min (190 °C,2.16 kg) and a density of 0.924 g/cm3 (20°C), was purchased from ExxonMobil Chemical. The
Page 4 of 18
average molecular weight (Mw) of this grade of LLDPE was about 95,000 g/mol. DCP (molecular weight 270 g/mol) and xylene (a mixture of o-, m-, and p-xylene with boiling point of 140°C and density of 0.87 g/cm3 at 20°C) were chemically pure and were obtained from Aldrich Chemical Co. 2.2. Modification procedure LLDPE was melt compounded with various amounts of peroxide i.e. 0, 15, 125, 500, 700,
ip t
1000, (ppm) using a DACA twin-screw Microcompounder (DACA Instruments, Goleta, CA, USA) with a screw speed of 100 rpm and temperature of 185 °C within various residence
cr
times. The preparation conditions are shown in Table 1.
Mixing Time (min)
an
0 15 125 500 700 1000 700
3 3 3 3 3 3 7.5
Ac ce pt e
d
LL/0 LL/15 LL/125 LL/500 LL/700 LL/1000 LL/700/7.5
Peroxide Concentration (ppm)
M
Sample Code
us
Table 1- Preparation conditions of peroxide modified LLDPE at various residence times and peroxide contents
2.3. Gel analysis
The most highly modified LLDPE samples i.e. the samples treated with high concentarations of DCP were examined for content of gel according to ASTM 2765. Based on this standard the extraction in boiling xylene was the basis of all the analyses.
2.4. Size exclusion chromatography (SEC) Average molecualr weights and molecular weight distributions (MWDs) were determined by high-temperature SEC coupled with a multi angle laser light scattering (MALLS) detector and a refractive index (RI) detector. The polymer molecules are fractionated in the SEC by their hydrodynamic volume, which depends on the density in the dissolved state, molar mass and degree of LCD. Therefore, the conventional SEC using linear polymer standards for the calibration is not suitable for investigations of the molar mass of branched polymer structures due to the fact that the calculated molar mass averages would be lower than the true values. [42] By coupling SEC with MALLS, the absolute molar mass MLS of every fraction can be determined directly. The SEC experiments were carried on a PL-GPC 220 (Polymer
Page 5 of 18
Laboratories) at 150 °C coupled with a MALLS (Heleos II, Wyatt Technology Corp., USA) and an RI detector. The column set consisted of two columns PL Mixed-B-LS (Polymer Laboratories). The eluent was TCB (1, 2, 4- trichlorobenzene, Merck) stabilized with BHT. The software used for data processing and for calculation of the number of LCB was Astra 5 (Wyatt Technology Corp., USA).
ip t
2.5. Rheometry
Rheological Measurements Parallel-plate rheometry was performed to determine the linear
cr
viscoelastic properties of peroxide modified LLDPE. The measurements were performed using a Paar-Physica Rheometer (MCR300, Ostfildern, Germany) in oscillatory shear mode
us
with parallel plates (25 mm in diameter with a gap of 1 mm) at a frequency range from 0.03 to 100 rad/s. The measurements were performed at four different temperatures, i.e. 150, 170,
an
190, and 210 °C with accuracy of ±0.5°C under N2 atmosphere. All the rheological
3. Results and Discussion 3.1. Gel content analysis
M
measurements were performed within linear viscoelastic region.
d
The gel content of peroxide modified LLDPE was measured by a solvent extraction
Ac ce pt e
technique. The gel content measurement results for all peroxide modified LLDPE revealed that the insoluble network structure fraction was negligible. This implies that incorporation of DCP to LLDPE predominately leads to branching rather than crosslinking. In other words, long-chain branches can be introduced into linear PE by peroxide modification, without gel formation.
3.2. Changes in Molecular Mass Distributions The MWDs of the LLDPE’s modified with different amounts of peroxide i.e., 0-, 125-, 500-, 700-, and 1000-ppm prepared at 185 °C for 3 min. are plotted in Figure 1. A reduction in distribution frequency together with the deviation of the distribution mode toward higher values implies that the dominant termination reaction mechanism is combination. It is seen that the changes in average molecular weight and molecular weight distribution (MWD) induced by peroxide modification under various conditions are small and there are no indications of formation of low-molecular-weight fractions due to chain scission. This fact leads to the conclusion that the degradation process is not predominant in the high molecular mass area of the polymer and long-chain branches can be introduced into linear PE by
Page 6 of 18
peroxide modification, without degradation and crosslinking. A slight increase of MWD is generally observed which can attriburtd to the occurrence of modification reactions, chain branching, and extension. These findings are in good agreement with the previous reports from Gloor et al. [43] and Pedernera et al. [44] on the modification of PE with low amounts of peroxide. When the aim is the formation of a three dimensional, insoluble network, the extent of peroxide is selected to be beyond 2000 ppm [43-45]. However, in this work, the highest
ip t
peroxide concentration used is 1000 ppm, which is well below the critical level for insoluble
Ac ce pt e
d
M
an
us
cr
network formation.
Figure 1. MWDs (dW/d log M) for the initial LLDPE (LL/0) and DCP modified LLDPEs processed at 185 °C for 3 min.
3.3. Rheological Studies
Effect of Peroxide Modification on Viscoelastic Behaviour of LLDPE As it was discussed earlier, since MWD is not affected significantly after modification, therefore the changes observed in the rheological behavior can be attributed to the changes in molecular architectures from linear to branched structure. It is worth to mention that the extent of degradation of the samples and the respective reporoducibility have been examined by means of dynamic tests. Figure 2 reveal the variation of G’(ω) and G”( ω) as a function of frequency for the DCP modified samples. As it is clearly seen, on increasing the DCP concentration, the magnitude of both moduli increases, in particular that of G’(ω) is more pronounced. Concerning the
Page 7 of 18
0
plateau of G’(ω), one can realize that G N
is almost independent of the extent of branching,
whereas the terminal region is considerably affected by branching. The similar investigation was done by Park et al [46]. They used a metallocene-based polyethylene which contained various extents of the long chain branches. The trend observed in his works are very similar to the one reported in the present work.
ip t
An enhancement of the moduli due to the branching indicates that the contribution of the end groups motions to the stress relaxation is not dominant and therefore they are low in number. It is then concluded that the branches should be of branching configuration. Moreover,
cr
increasing the peroxide content in LLDPE, results in increasing the activation energy of the
us
peroxide-modified LLDPE. As it is known the activation energy is independent of molecular weight (Mw) and molecular weight distribution (MWD), therefore the higher activation energy values of modified LLDPE with higher peroxide content can be due to the presence of
an
more branches. This increase in activation energy at higher branch content can be related to the slowed segmental dynamics [23]. The activation energy is representative of the potential
M
energy from the flow of a molten polymer, therefore introducing branches retards the overall
Ac ce pt e
d
dynamics, and consequently higher activation energy is required for the segmental motion.
Figure 2. a) Effect of DCP modification on storage modulus and b) loss modulus of initial LLDPE and the samples modified with different amounts of peroxide processed at 185 °C for 3 min.
In our previous work [39] we studied the correlations between rheological behaviour and degree of long chain branching (LCB) of LLDPE upon a peroxide (DCP) modification process under various conditions. It was reported that at low frequencies, the viscosity of the modified samples is greatly affected by peroxide modification and the zero-shear rate viscosity (η0) values of modified LLDPE is higher than that of the neat LLDPE. Therefore, peroxide modification of the LLDPE increases the dynamic viscosity of the samples. In the
Page 8 of 18
other words, the η0 values increase with increasing of long chain branching. As was discussed earlier, the molecular weight and MWD of modified LLDPE did not significantly change, therefore this deviation could be attributed to structural changes in LLDPE induced by long chain branching. The degree of long-chain branching (LCB) was evaluated based on the zero shear rate viscosity data and the results in the form of van Gurp-Palmen plots are presented in Figure 3.
ip t
A plot of the phase angle δ versus the complex shear modulus |G*(w)|, the so-called van Gurp-Palmen plot, is frequently used to get an insight on long-chain branching. This method
cr
is rather sensitive in molecular structural change. The influence of peroxide concentration and time of mixing on phase angle δ of samples is shown in Figure 3. As it is seen from the figure
us
with increasing peroxide level from 200ppm (L1/185/3) to 1200ppm (L3/185/12) and time of mixing from 3(L1/185/3) to 12 min. (L1/185/3) the phase angle δ decreases. In the medium
an
range of the complex modulus between rubbery plateau region and the Newtonian region a deviation of the phase angle δ beneath the curve of linear LLDPE is observed for the samples modified with 200ppm peroxide and higher. A shift of the phase angle towards small values at
M
fixed |G*(w)| can be caused by a broader molar mass distribution or an introduction of longchain branching. In this work for the peroxide modified LLDPE the decrease is only attributed
d
to the introduction of long-chain branching since the polydispersity remains approximately
Ac ce pt e
constant according to the SEC analysis.
Figure 3. δ(|G*|) -plot of the peroxide modified LLDPE Thermorheological Investigation Page 9 of 18
As mentioned earlier, materials obeying time-temperature superposition principle have "simple thermorheological" behavior. The shift factor, aT, facilitates superposition of dynamic moduli measured at different temperatures. However, the effect of temperature cannot be simply depicted via aT for long chain branched (LCB) polymers. Due to existence of molecules of different length and number of branches LCB polymers are thermorheologically complex and a spectrum of activation energy, Ea(λ), is necessary for description of their
ip t
behavior. Ea increases with increasing LCB. Vega et al. suggested the following equation to establish a relation between the long chain branching and the activation energy for the linear
us
E a ,b − E a , L Ea ,L
(5)
an
I LCB =
cr
and branch polymers [13].
It is to be noted that in the literature, activation energy values of 24–26 kJ per mole for linear polyethylene have been reported [20]. Therefore, for linear polyethylene Ea,L value assumed
M
with Ea,L of HDPE (Ea=24kJ/mol.).
As it is clear, the virgin sample in this work has linear structure and thus expected to have
d
simple thermorheological behavior. But our structure modification brings about complexity to the sample behavior. In this section, attempt is made to obtain the threshold amount of
Ac ce pt e
peroxide to induce thermorheological complexity. Thus temperatures of 150, 170 and 190o C for the behavior study, and 190oC as the reference temperature for determining the horizontal and the vertical shift factors were selected.
To compare the relaxation behavior of modified LLDPE, the weighted relaxation spectra evaluated from the linear viscoelastic data using the US200 software and the shift factors have been definitely optimized.
The curves of the horizontally and vertically shifted storage and loss moduli versus frequency for the modified samples with 15 to 1000 ppm peroxide at 180oC for 3 minutes are depicted in figures 4 to 5. The more observed non-superimposibility of the curves at difference temperatures, the more complexity indication of the thermorheological behavior. It is observed that increasing amount of peroxide from 15 to 1000 ppm intensifies this behavior. Figure 4-5 shows the phase angle of the modified samples with 15 to 1000 ppm peroxide at 180oC for 3 minutes and the neat LLDPE as a function of reduced frequency. The curves were shifted along the x-axis by multiplying the values with the coefficients shown on the diagram for better visualization. Figure 4 and 5 indicates that there is a good superposition to obtain a
Page 10 of 18
master curve resulting in simple thermorheological behavior for LLDPE. Nevertheless, there is obvious non-superposition for other samples that makes it impossible to obtain a master curve. The more observed non-superimposibility of the curves at difference temperatures, the more complexity indication of the thermorheological behavior. It is observed that increasing amount of peroxide from 15 to 1000 ppm intensifies this behavior. In other words, the reason behind this fact that the curves corresponding to the different temperatures can not be
ip t
superimposed on each other (after applying shift factors) is indeed due to the complicated thermorheological behavior. As the concentration of peroxide is raised and long chain
cr
branching thereby increases, the thermorheological behavior becomes more and more sophisticated. This results in further lack of superposition of the curves which is clearly
Ac ce pt e
d
M
an
us
visible in Figures 4 and 5.
Figure 4. Master curves of the shifted storage modulus (bT G’) and shifted loss modulus (bT G’’) as a function of the reduced frequency (aTω) for modified LLDPE with 15 ppm peroxide at temperature of 180 °C for 3 min.
Page 11 of 18
ip t cr us
an
Figure 5. Master curves of the shifted storage modulus (bT G’) and shifted loss modulus (bT G’’) as a function of the reduced frequency (aTω) for modified LLDPE with 1000 ppm peroxide at temperature of 180 °C for 3 min.
Furthermore, curves of phase angle δ against reduced frequency for peroxide modified samples (accroding to Table 1) are depicted in figures 6 and 7.
M
Figure 6 and 7 shows the phase angle of peroxid modified LLDPE and the neat LLDPE as a function of reduced frequency. It is observed that with increasing peroxide (increasing LCB)
d
the curves shift to smaller values of δ. Figure 7 compares δ-frequency curves for samples P700/185/3 and P700/185/7.5 distinguished only in mixing time. For enhanced unambiguity,
Ac ce pt e
curves of different sample are shifted by some factors along the frequency axis.
Figure 6. Phase angle δ as a function of the reduced frequency (aT.ω) at different temperatures (150, 170 and 190 °C) for modified LLDPE mixed for 3 min (the curves were shifted along the ω-axis by the factors indicated, for the matter of a better visualization) (T0 =150°C).
Page 12 of 18
ip t cr us
an
Figure 7. Phase angle δ as a function of the reduced frequency (aT.ω) at different temperatures for two modified LLDPE (P700/185/3 and P700/185/7.5) (the curves were shifted along the ω-axis by the factors indicated, for the matter of a better visualization) (T0 =150°C). A close look at the samples behavior in figure 6, one reveals that a concavity change occurs
M
from the third sample, i.e. above 125 ppm in peroxide content, illustrating simple-to-complex threshold in thermorheological behavior of the samples. This behavior has been observed in
d
the graphs for moduli and complex viscosities of the samples. Figure 6 well illustrates the
Ac ce pt e
simple-complex change in behavior of the samples. Similar works by Keßner et al. in 2010 [32], and Stadler et al. in 2007 [15] have been made on thermorheology of metallocene LLDPE with sparse long branches as well as LDPE. In figure 6, samples 4 and 5 exhibit a behavior similar to metallocene samples with sparse long branches and sample 6 shows a behavior similar to LDPE. To evaluate the thermorheological behavior in greater detail, the activation energy curves as functions of phase angle (δ) can be used. To establish the activation energy curve, first determine frequency at constant phase angles and then obtain shift factor for each value of δ with respect to reference temperature of 190oC. The shift factor curve against reciprocal temperature gives activation energy values for each δ. Thus activation energy curve as a function of δ is established. According to Keßner and Münstedt method [32] logarithms of the time-scale shift factors at a constant δ are plotted against the reciprocal absolute temperature 1/T. By using Arrhenius relationship for shift factors, the activation energies of peroxide modified LLDPE (15 to 1000 ppm peroxide) and the neat LLDPE are plotted as a function of δ in Figure 8. It is interesting to see the difference between activation energies of samples containing more and less than 125 ppm peroxide. As demonstrated in figure 6, the curves of
Page 13 of 18
these samples are different in concavity, although curves of activation energy reveal this difference more clearly. Indeed in case of simple thermorheological behavior, curve of activation energy against any change of phase angle remains unchanged (linear polyethylenes without any long branches show this type of behavior) [15,35,25, 21]. However, for complex thermorheological behavior activation energy change is observable. LLDPE exhibits constant activation energy. Therefore, based on this method a simple
ip t
thermorheological behavior is concluded for LLDPE. This is similar to the result reported by Keßner et al. [32, 34] LLDPE and LLDPE/LDPE blends show phase angle dependent
cr
activation energies that indicates complex thermorheological behaviour which is attributed to the presence of long branches. The reason for the observed simple thermorheological behavior
us
of LLDPE containing high amount of SCB can be the presence of short branches in each LLDPE molecule and similarity of the structures of these branches. [33]
an
At high values of phase angle ,we observe an activation energy of about 60 pertaining to LCB where as at low values of phase angle, an activation energy of about 25-30 due to linear chains of PE is observed. With regard to figure 8, for the linear LL sample, LL/15 and LL/125
M
the activation energy is almost constant but with increasing peroxide (hence increasing LCB), for samples LL/500 to LL/1000, the activation energy modifies from high values (at large
d
phase angle) to low values (at small phase angle). Somehow one may assume such samples to
Ac ce pt e
be blends of linear chains and chains containing long branches.
Figure 8. Activation energies as a function of the phase angle for peroxide modified LLDPE
Page 14 of 18
In a similar manner one can deduce curve of activation energy against elastic modulus. Figure 9 shows the activation energy of modified LLDPE and neat LLDPE as a function of storage modulus according to Wood-Adams and Costeux method [21]. In these curves, also, if the thermorheological behavior is simple, then the activation energy is nearly constant, and one has the complex behavior, the activation energy alters (decreases) with increasing of elastic modulus. For all samples, dependency of activation energy on the storage modulus was
ip t
evident. This indicates thermorheological complexity of the samples based on this method. This finding is in agreement with the results reported by Keßner and Münstedt [32]. For all
cr
samples, activation energy decreases with the increase in storage modulus. Furthermore, the decreasing rate of the activation energy increases with increase in the long branch content
Ac ce pt e
d
M
an
us
(higher peroxide content).
Figure 9. Activation energies as a function of the storage modulus for peroxide modified LLDPE at 150, 170, 190 °C Conclusion The correlations between microstructural changes induced in a linear low density polyethylene (LLDPE) upon peroxide modification with corresponding thermorheological responses were explored. Chromatography results showed that with increasing peroxide content molecular weight distribution broadened a little and long chain branches increased.
Page 15 of 18
The gel content analysis revealed that the insoluble crosslinked fraction was negligible in all the modified samples. This implies that incorporation of DCP to LLDPE predominately leads to branching rather than crosslinking. Results showed that both G'(ω) and G''(ω) increased with peroxide content, although variation of G'(ω) was more pronounced. With regard to plateau section of G'(ω) curve, one can deduce that the plateau modulus, GoN, is nearly independent of branches presence, while effect of
ip t
branches are marked in the terminal zone. Results of shifted storage and loss moduli against frequency for 15 to 1000 ppm peroxide modified samples showed non-superimpossiblity of
cr
the curves with increasing the peroxide content, indicating pronounciation of complex thermorheological behavior. Results of activation energy as a function of phase angle showed
us
that at low levels of peroxide, the polymer enjoys of simple thermorheological behavior (constant activation energy) while at higher amounts from 125 to 1000 ppm, the behavior
an
converted to complex thermorheology (change in activation energy). Results of activation energy curves versus elastic modulus indicated change in activation energy with increasing the elastic modulus for peroxide modified samples. The amount of change is little at low level
M
of peroxide but the activation energy decreases considerably with increasing of peroxide. The threshold of sample-to-complex thermorheological behavior can be detected from these
d
curves.
Ac ce pt e
Acknowledgments: The author expresses his gratitude to the Alexander von Humboldt foundation for financial support. Also, technical assistance from A. Lederer, B. Kretzschmar and A. Ivanov is deeply appreciated.
References
[1] C. J. Perez, G. A. Cassano, E. M. Vallés, M. D. Failla, L. M. Quinzani, Polymer 2002, 43, 2711. [2] M. Pedernera, C. Sarmoria, E. M. Vallés, A. Brandolin, Polym. Eng. Sci. 1999, 39, 2085. [3] W. S. Lambert, P. J. Phillips, J. S. Lin, Polymer 1994, 35, 1809. [4] P. Sajkiewicz, P. J. Phillips, J. Polym. Sci., Part A: Polym. Chem.1995, 33, 949. [5] T. Bremner, A. Rudin, J. Appl. Polym. Sci. 1995, 57, 271. [6] T. K. Kang, C. S. Ha, Polym. Testing 2000, 19, 773. [7] H. Fang, Y. Zhang, J. Bai, Zh.Wang, Zh.Wang, RSC Advances, 2013, 3(23), 8783. [8] Y. Wang, L. Yang, Y. Niu, Z. Wang, J. Zhang, F. Yu, H. Zhang, J. Appl. Polym. Sci., 2011,122, 1857. [9] C. Gabriel, H. Münstedt, Rheol. Acta ,38,393-403,1999. [10] H. Xu, H. Fang, J. Bai, Y. Zhang, Zh. Wang, Ind. Eng. Chem. Res. 2014, 53, 1150-1159.
Page 16 of 18
[11] P. Lehmus, E. Kokko, O. Harkki, R. Leino, HJG. Luttikhedde, J.H. Nasman, J.V. Seppala, ,Macromolecules , 32,3547-3552,1999. [12] A. Malmberg, E. Kokko, P. Lehmus, B. Lofgren, J. Seppala, ,Macromolecules, 31, 84488454,1998. [13] J.F. Vega, A. Santamaria, A. Munoz-Escalona, P. Lafuente, Macromolecules , 31, 36393647,1998.
ip t
[14] Niu Y-H., Wang Z-G, Duan X-L, Shao W., Wang D-J., Qiu J., Journal of Applied Polymer Science, 2011, 119, 530.
[15] F.J. Stadler, A. Nishioka, J. Stange, K. Koyama, H. Münstedt, Rheol. Acta, 46,1003-1012, 2007.
Polyethylenes”, Dissertation, 2010.
us
[17] X. Chen, C. Costeux, R.G. Larson, J. Rheol. 2010, 54,1185.
cr
[16] U. Keßner, “Thermorheology as a Method to Investigate the Branching Structures of
[18] B.H. Zimm, W.H. Stockmayer, J. Chem. Phys. , 17, 1301-1314,1949.
an
[19] K. Klimke, M. Parkinson, C. Piel, W. Kaminsky, H.W. Spiess, M. Wilhelm, Macromolecular Chemistry and Physics, 207, 382-395, 2006.
[20] W.W. Graessley, Account. Chem. Res., 10, 332-339,1977.
M
[21] Wood-Adams P.M., Costeux S., Macromolecules, 34,6281-6290, 2001. [22] P.J. Rouse, J. Chem. Phys., 21, 1272-1279, 1953.
[23] F.J. Stadler, C. Gabriel, H. Münstedt, Macromol. Chem. Phys., 208, 2449-2454, 2007
d
[24] P.M. Wood-Adams, J.M. Dealy, Macromolecules, 33, 7481-7488, 2000.
Ac ce pt e
[25] U. Keßner, J. Kaschta, F.J. Stadler, S. Le DuffC, X. Drooghaag, H. Münstedt, Macromolecules, 43, 7341–7350, 2010.
[26] J. A. Resch, U. Keßner, F. J. Stadler, Rheol. Acta, 50,559-575, 2011. [27] P. M. Wood-Adams, S. Costeux, Macromolecules 2001, 34, 6281. [28] H. M. Laun, Prog. Colloid Polym. Sci. 1987, 75, 111. [29] D. Bonchev, A.H. Dekmezian, E. Markel, A. Faldi, J. Appl. Polym. Sci. 2003, 90, 2648. [30] D. J. Lohse, S.T. Milner, L.J. Fetters, M. Xenidou, N. Hadjichristidis, J. Roovers, R.A. Mendelson, C. A. Garcia-Franco, M.K. Lyon, Macromolecules 2002, 35, 3066. [31] E. Kokko, A. Malmberg, P. Lehmus, B. Lofgren, J.V. Seppala, J. Polym. Sci. Part A: Polym. Chem. 2000, 38, 376.
[32] U. Keßner, H. Munstedt, Polymer 2010, 51, 507. [33] F.J. Stadler, J. Kaschta, H. Munstedt, Macromolecules 2008, 41, 1328. [34] U. Keßner, J. Kaschta, F.J. Stadler, C.S. Le Duff, X. Drooghaag, H. Munstedt, Macromolecules 2010, 43, 7341. [35] F.J. Stadler, C. Gabriel, H. Munstedt, Macromol. Chem. Phys 2007, 208, 2449. [36] A. K. Dordinejad, S. H. Jafari, H. A. Khonakdar, U. Wagenknecht, G. Heinrich, J. Appl. Polym. Sci. 2013, 129, 458.
Page 17 of 18
[37] A. K. Dordinejad, S. H. Jafari, Polym. Eng. Sci. 2014,54:5,1081–1088. [38] A. K. Dordinejad, S. H. Jafari, J. Appl. Polym. Sci. 130: 3240–3250, 2013. [39] M. Golriz, H.A. Khonakdar, J. Morshedian, H. Abedini, S.H. Jafari, A. Lederer, U. Wagenknecht, Macromol. Mater. Eng. 2014, 299, 154–164. [40] M. Golriz, H.A. Khonakdar, J. Morshedian, S.H. Jafari, Y. Mohammadi, U. Wagenknecht, Macromol. Theory Simul. 2013, 22, 426–438.
ip t
[41] M. Golriz, H.A. Khonakdar, J. Morshedian, J. Appl. Polym. Sci. 2014, 131, 39617. [42] H.A. Khonakdar, Polymers for Advanced Technologies,2014, DOI: 10.1002/pat.3314. [43] P. E. Gloor , Y. Tang, A.E. Kostanska, A.E. Hamielec Polymer, 35 ,1012-1019, 1994.
cr
[44] M. N. Pedernera, M. Sarmoria Valles, A. Brandolin Polym. Eng. Sci., 39, 2085-2095., 1999.
Ac ce pt e
d
M
an
[46] S. J. Park, R.G. Larson, J. Rheol., 49, 523-536, 2005.
us
[45] A. E. Hamielec, P. E. Gloor, S. Zhu Can J. Chem Eng., 69, 611-617, 1991.
Page 18 of 18