Long Chain Branching Detection and Quantification in LDPE with

Oct 17, 2018 - The Dow Chemical Company, Lake Jackson , Texas 77566 , United States. Macromolecules , Article ASAP. DOI: 10.1021/acs.macromol...
0 downloads 0 Views 3MB Size
Download PDF
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Long Chain Branching Detection and Quantification in LDPE with Special Solvents, Polarization Transfer Techniques, and Inverse Gated 13C NMR Spectroscopy Zhe Zhou,* Stacy Pesek, Jerzy Klosin, Mari S. Rosen, Sukrit Mukhopadhyay, Rongjuan Cong, Dan Baugh, Bill Winniford, Hayley Brown, and Kamin Xu Macromolecules Downloaded from pubs.acs.org by AUCKLAND UNIV OF TECHNOLOGY on 10/24/18. For personal use only.

The Dow Chemical Company, Lake Jackson, Texas 77566, United States ABSTRACT: Recent global demand of polyethylene was 187 billion pounds with 23% being low density polyethylene (LDPE). LDPE is used in a variety of applications including liners, consumer bags, heavy duty sacks, caps, closures, toys, lamination, and agricultural films. A key feature of LDPE is the presence of long chain branching (LCB) which has a significant impact on physical and rheological properties, including melt strength. Previously it was reported that there are no C6 branches in LDPE, and therefore the common practice has been to use the resonance around 32.2 ppm as a measure of LCB content. Herein, the existence of C6 branches in LDPE is reported for the first time, enabled by the use of 1-chloronaphthalene as the NMR solvent. It is shown that the C6 branches have a contribution to the resonance around 32.2 ppm. This finding suggests that C6 branches should be measured in LDPE samples and be excluded from LCB quantification because they do not affect LDPE rheological properties as LCB does. The data obtained strongly suggest that C7 branches are present in much lower concentrations in the LDPE samples studied. The quantification of these resonances traditionally requires long acquisition time, even with the use of a cryoprobe. In this work, different polarization transfer techniques, refocused insensitive nuclei enhanced by polarization transfer (RINEPT), and distortionless enhancement by polarization transfer (DEPT) were compared in terms of their ability to enhance CH2 sensitivity, thus leading to a greater ability to observe the CH2 resonances belonging to the C6 branches. It was found that RINEPT with hard 180° 13C pulses is most suitable for this purpose. A much faster method is proposed for measuring LCB in LDPE that employs a combination of a conventional quantitative 13C NMR technique and a polarization transfer technique.

■

INTRODUCTION Low density polyethylene (LDPE), commercially produced since 1939, is produced via free radical polymerization at high temperature and high pressure autoclave, autoclave/tube, or tubular reactors.1 A majority of manufacturers use tubular reactors now because they can produce material with both narrow and broad molecular weight distributions. Recent global consumption of polyethylene was 187 billion pounds with 23% being LDPE.2 LDPE is used in liners, consumer bags, heavy duty sacks, caps, closures, toys, lamination, and agricultural films, among other uses. It has excellent optical properties and processability, good shrink properties, and no catalyst residuals. LDPE has a density in the range 0.918− 0.928 g/cm.3 Typically LDPE has about 10−20 short chain branches (SCB)/1000 carbons, the level of which impacts the crystallinity of the material. A key feature of LDPE is long chain branching (LCB) which has a significant impact on physical and rheological properties, including melt strength.3−5 LCB in LDPE has been characterized via a variety of techniques, including 13C NMR spectroscopy, triple-detector gel permeation chromatography (TD-GPC), and rheology.1 For substantially linear ethylene polymer (SLEP), TD-GPC is a reliable approach to measure LCB, but for LDPE certain corrections to TD-GPC measurements are needed. Rheological properties such as rheological ratio, shear thinning, and © XXXX American Chemical Society

melt strength can provide some information about LCB content, but such measurements do not enable direct quantification of LCB. 13 C NMR spectroscopy, when performed correctly,6 provides an intrinsically quantitative method for LCB quantification, and it has been widely used to measure LCB level in SLEP.7−9 13C NMR spectroscopy has also been used extensively to measure LCB content in LDPE by quantifying the resonance from the CH2 group located at the third carbon from the chain end (CE3) resonating at approximately δ 32.1−32.3 ppm (see Scheme 1; the drawing is based on previous NMR reports1,7). Prior papers report values for LCB content in LDPE where any contribution from C6 branches to longer branches is included in the LCB value because it was not possible to resolve the C6 resonance from the resonances corresponding to longer chains.1,3,7,10−15 In addition, it has been reported previously that C6 branches are not present in LDPE.16 Recently, we reported a new 13C NMR method in which halogenated naphthalenes are used as NMR solvents in experiments to detect and quantify LCB content in SLEP. In that example, LCB signals (branch methine, α- and βmethylenes) were well separated from resonances correspondReceived: August 21, 2018 Revised: October 1, 2018

A

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Macromolecules

■

Scheme 1. Illustration of Chemical Structure of LDPE with Labeling of Some of the Branch Types Observed with 13C NMR Spectroscopy1,7 a

Article

EXPERIMENTAL SECTION

NMR. Polyolefins are provided by The Dow Chemical Company. All solvents were obtained from Sigma-Aldrich without further purification. The NMR experiments were done on a Bruker 600 MHz (Avance III HD console and Ascend magnet) with CH dual 5 mm cryoprobe. Tune and match were adjusted before every experiment. 1 H and 13C pulse widths were measured with a few drops of TCE in DMSO-d2 with 0.025 M Cr3+ at 120 °C. 1H and 13C pulse widths were 10.7 and 15.5 μs, respectively. The acquisition time was 2.62 s. For LDPE analysis, 50 mg of polymer was added to 0.5−0.7 g of solvent in a 5 mm NMR tube. The solvent with polymer was purged for 5 min with N2, and then the polymer was dissolved at 135 °C. The carrier frequencies for 13C and 1H were set at 30.0 and 1.50 ppm, respectively. For the experiments in which 1-chloronaphthalene (ClNA) was used, 10 wt % of p-dichlorobenzene-d4 (PDCB-d4) was added to the sample for locking and shimming. Detailed NMR parameters are listed in figure captions. Integrals were obtained with Mnova software version 12.02. Computational Methodology. The geometry of the reactant (straight chain primary alkyl radical, containing 12 carbon atoms) and products (straight chain secondary radicals, with different branch lengths; C4 to C10) were computed using unrestricted open shell density functional theory with hybrid functional (B3LYP)18,19 and 631g* basis set.20,21 The ground states of reactants and products were confirmed by lack of imaginary frequency. The transition state geometries (where the primary radical is converted to secondary radicals) were optimized at the same level as that of the ground state; the optimized geometry of each transition states was confirmed by the presence of one imaginary frequency. Furthermore, the reaction coordinate for the hydrogen transfer was confirmed by viewing the normal mode, having an imaginary frequency. The barrier height for the radical transfer (termed backbiting in the subsequent text) was calculated as the energy difference between the transition state for forming the secondary radicals and the ground state of the primary radical, as shown in Scheme 2. To understand the effect of basis set on the computed energies, single point energy calculations were performed on optimized geometries of the ground state primary radical and the transition states using the B3LYP/6311+g** method (termed high level method in subsequent text). Similar methodologies were used to compute backbiting processes in polyethylene, polystyrene, polyacrylate, and poly(vinyl chloride)s.22,23 Because our

a

Dashed lines represent polymer chains. CE1−3 represent carbons that are at various positions relative to the chain end (CE). The red C6 branch is a branch observed in LDPE with 13C NMR spectroscopy in this work.

ing to C6 side chains.17 Unlike C6 and intermediate length side chains, LCB in LDPE significantly impacts rheological properties, and thus it is important to be able to quantify LCB content accurately. The data in this paper, obtained using 1-chloronaphthalene as the NMR solvent, answer the question of whether C6 branches can be detected in LDPE. In addition, a new approach is reported for the quantification of LCB in LDPE by using a combination of a conventional quantitative 13 C NMR technique and a polarization transfer technique.

Scheme 2. Molecular Structures of Reactants, Transitions States, and Productsa

a

The hydrogen atom, involved in the backbiting process, was explicitly indicated in the transition state structures; the bond making and bond breaking processes were also indicated by dashed lines. Backbiting leads to formation of different branches (C4 to C10), depending on the position of the hydrogen transfer, leading to different types of secondary radicals. B

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules focus is primarily directed toward understanding the difference in concentration of the short chain branches (C4 to C10) in LDPE, the quantum tunneling effect for hydrogen transfer is assumed to be constant and is not expected to affect the trends in relative concentration of these branches.22 These calculations were performed using G16 suite of programs.24

level is a good indicator of LCB content in LDPE. Thus, method C is more reliable compared with methods A and B. As C1−C5 branches have no signals in region C,7 this method calculates LCB6+, i.e., the chains longer than a C5 branch, which includes any C6 branches that may be present. The presence of C4 and C5 branches in LDPE is wellestablished.26 Given the complex structure of LDPE (see Scheme 1),7 we wanted to explore the possibility of C6 branches even though prior art suggested that C6 branches were not present in LDPE.16 Establishing the presence of C6 branches in LDPE would contribute to a greater understanding of the measured value that is commonly attributed to LCB content. Detection of C6 Branches and Other Short Chain Branches in LDPE. To investigate the question of C6 branches in LDPE, a 13C NMR spectrum of a typical LDPE sample was compared with the 13C NMR spectrum of an ethylene−1-octene (EO) linear low density polyethylene (LLDPE) sample. As can be seen in the top two spectra in Figure 2 in which the samples are dissolved in tetrachloro-

■

RESULTS AND DISCUSSION There are predominately three methods which may be considered for calculating LCB content in LDPE: LCBmethod A = A area integral − C4 branch integral − C5 branch integral LCBmethod B = B area integral − 2 × C4 branch integral − 3 × C5 branch integral LCBmethod C = C area integral

Method A uses the area of the branch CH resonance, as branches from C4 up to LCB all have CH peaks in the A region (Figure 1). Consequently, the level of branches longer than a

Figure 1. A portion of a 13C NMR spectrum of a typical LDPE sample in tetrachloroethane-d2 with 0.025 M Cr(acac)3 at 120 °C. The green boxes highlighting regions A, B, and C indicate the integral range used for these three areas of the spectrum. The resonances corresponding to the short C4 and C5 chains are also labeled here.

Figure 2. Top two spectra: 13C NMR spectra of a typical LDPE polymer and an ethylene−1-octene (EO) linear low density polyethylene (LLDPE) sample in tetrachloroethane-d2 (TCE-d2) with 0.025 M Cr(acac)3. Bottom two spectra: the same samples in TCE-d2/o-dichlorobenzene (ODCB) (50:50, w:w) with 0.025 M Cr(acac)3. All NMR spectra were acquired at 120 °C. The C6 branch signal labeled here is from the third carbon counting from methyl from the C6 side chain in EO.

C5 branch was evaluated by subtracting the areas of the C4 and C5 branches from the value of the area of region A. However, LDPE is highly branched, and the CH chemical shifts of C4 and longer branches are affected by nearby branches, generally moving upfield, so negative values were often obtained with method A. Method B uses the area of the CH2 resonances α to the CH branch. Because C4 branches have two α carbon signals and C5 branches have two α carbon signals and one CH2 carbon signal from the side chain for a total of three signals in the B area, the content of branches longer than a C5 branch was evaluated by subtracting 2 × C4 branch area and 3 × C5 branch area from the area of the B region. However, as with method A, the chemical shifts of CH2 α carbons to C4 -LCB branches are affected by nearby branches; therefore, method B is also not reliable. In contrast, method C uses the area of the third CH2 carbon from the chain end (CE3, see Scheme 1). This method has been widely used to measure LCB content in LDPE with 13C NMR. Specifically, method C measures the amount of saturated, nonbranched long chain ends (LCE)25 rather than directly measuring LCB. The levels of LCE and LCB are often used interchangeably as the LCE

ethane-d2 (TCE-d2) with 0.025 M Cr(acac)3, a shoulder on the left side of the resonance corresponding to the CE3 carbon (Scheme 1) is visible. A comparison with the EO polymer sample in the same solvent suggests that this shoulder may result at least in part from C6 branches. Further separation of the shoulder from the main portion of the LDPE CE3 resonance was desired, so the combination of TCE-d2 and odichlorobenzene (ODCB) (50:50, w:w) with 0.025 M Cr(acac)3 was employed as the solvent. This solvent system provided greater separation of the shoulder from the main CE3 resonance as can be seen in the bottom two spectra in Figure 2, and a resonance corresponding to C6 branches was clearly visible. Further inspection of the CE3 resonance in the TCEd2/ODCB solvent system revealed a subtle shoulder on the left side of the peak. C

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules In 2017, Zhou et al. reported a measurement of LCB in SLEP by 13C NMR spectroscopy using halogenated naphthalenes as the solvent.17 Based on those results, which demonstrated a difference in chemical shift between the methines of SCB and LCB, 1-chloronaphthalene was explored here to evaluate whether it would lead to a better separation of the resonance corresponding to the C6 branches and the shoulder of the CE3 resonance of LCB in LDPE. 13C NMR spectra were acquired (Figure 3) for a sample of LDPE and a

Figure 4. 13C NMR spectra of other two LDPE samples, referred to here as LDPE-Y and LDPE-Z, and an EO in 1-chloronaphthalene (ClNA)/p-dichlorobenzene (PDCB-d4) (9:1, w:w) with 0.025 M Cr(acac)3 at 120 °C.

Figure 3. 13C NMR spectra of a typical LDPE and an EO polymer in 1-chloronaphthalene (Cl-NA)/p-dichlorobenzene (PDCB-d4) (9:1, w:w) with 0.025 M Cr(acac)3 at 120 °C.

sample of EO copolymer prepared in 1-chloronaphthalene (ClNA)/p-dichlorobenzene-d4 (PDCB-d4) (9:1, w:w; PDCB-d4 is used for locking and shimming) and with 0.025 M Cr(acac)3. In contrast to previous 13C NMR studies of LDPE,13,16 a welldefined C6 branch peak is clearly observed along with additional peaks that correspond to other branches which are observed between the C6 branch peak and the CE3 peak. To determine whether this observation is general across multiple LDPE samples, 13C NMR spectra of two more LDPE samples, referred to here as LDPE-Y and LDPE-Z, and an EO copolymer in 1-chloronaphthalene (Cl-NA)/p-dichlorobenzene (PDCB-d4) (9:1, w:w) with 0.025 M Cr(acac)3 were acquired (Figure 4). Clearly, LDPE-Y and LDPE-Z also have a well-separated resonance arising from C6 branches along with other resonances between the C6 branch peak and the resonance of the CE3. A polyethylene−1-nonene (EN) and a polyethylene−1decene (ED) samples were also studied to elucidate the identity of other visible peaks between C6 branch and the CE3 of the main LCB. Figure 5 shows the 13C NMR spectra of a typical LDPE sample along with an EO, an EN, and an ED copolymers. From a comparison between the EN copolymer and the LDPE sample, a C7 branch should be found at 32.160 ppm, but no strong resonance is observed at this location for the LDPE sample, indicating that the level of C7 branches is much lower than the C6 branch level if C7 branches exists at all, which corroborates with computational results (see section below). The 13C NMR spectrum of the ED sample demonstrates that the resonance corresponding to C8 branches should be at 32.133 ppm, but it is unclear whether there are C8 branches in LDPE given the location of the CE3 resonance.

Figure 5. 13C NMR spectra of a typical LDPE sample, along with ethylene−1-octene (EO), ethylene−1-nonene (EN), and ethylene−1decene (ED) samples in 1-chloronaphthalene (Cl-NA)/p-dichlorobenzene (PDCB-d4) (9:1, w:w) at 120 °C.

However, the computational results (see below) indicate that the frequency of C8 branches is expected to be comparable to that of C7 branches, and thus C8 branches are likely to be very infrequent or nonexistent altogether. Because the resonance labeled with a question mark in Figure 5 has similar intensity as that of the C6 branches and the chemical shift is consistent with 2-ethyloctyl branch, it is plausible it originates from 2ethyloctyl branches. This will be the subject of future studies. If the resonances observed between the C6 branch resonance and the CE3 resonance do in fact correspond to short chain branches, they should also not be included in the peak integration for measuring LCB levels in LDPE. Computational Results. Figure 6A depicts the computed energies of forming different types of secondary radicals via unimolecular back biting. Because the primary radical is converted to secondary radicals, the backbiting process is expected to be exothermic in nature, as validated by computational results. Furthermore, the computed energies D

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 6. (A) Computed energies of secondary radicals with different branch lengths with respect to ground state of primary radicals, using lower level and higher level methodologies. (B) Computed barrier heights of forming secondary radicals with different branch lengths with respect to the ground state of primary radicals, using lower level and higher level methodologies.

(the integral considered in prior arts to be LCB in LDPE) in the LDPE sample in Figure 5, and they are spread across about 0.07 ppm. A very long NMR acquisition time is needed to obtain a good signal-to-noise ratio for quantification of these peaks with conventional inverse gated quantitative 13C NMR spectroscopy and the use of a cryoprobe.28,29 To shorten the NMR analysis time we explored polarization transfer techniques RINEPT, shown at the top in Figure 7 where the enhancement factor depends on time T, and DEPT, shown at the bottom in Figure 7, where the enhancement

of different types of secondary radicals are comparable (irrespective of the methodology used), which indicates that the relative concentration of different branches in LDPE is guided by the kinetics of backbiting rather than the thermodynamics of the secondary radical formation. This led us to compute the barrier heights of backbiting process, which results in different branch lengths in LDPE. Figure 6B depicts the barrier heights for backbiting. The computed barrier height for forming the C4 branch is between 16.5 and 17.1 kcal/mol, depending on the chosen methodology, which is comparable to computed barrier heights at high level post-Hartree−Fock method (17.4 kcal/mol at the QCISD(T) level).27 Because the barrier height for formation of C4 branches is the lowest, it is expected that C4 branches will dominate in LDPE if backbiting is the major mechanism for short chain branching formation. The barrier height for the formation of C5 branches is marginally higher than that of C4 branches, which indicates that C5 branches should be less frequent than C4 branches (within an order of magnitude) in LDPE. The computational trend matches well with the experimental observation of the relative concentration of C4 and C5 branches in LDPE (see Table 3). On the other hand, the barrier for the formation of C6 branches is ∼3.7 kcal/mol higher than that of the C4 branches. Thus, the computational results indicate that the formation of C6 branches is less likely than those of C4 and C5 branches. This is also in accordance with the experimental results (see Table 3), which indicates that the concentration of the C6 branches is lower than C4/C5 branches by an order of magnitude. However, the formation of C6 branches is more likely than that of C7 branches by the backbiting mechanism. This finding is also consistent with experimental observations (Figure 5), which indicate very low level of C7 branches, if it is present at all. Moreover, the barrier heights for forming C7 and longer branches (C8 and higher) are comparable to each other and are higher than that of C6 branch by 4.8−5.8 kcal/mol (depending on the chosen computational methodology). Thus, a low level of C7 branches in LDPE is likely to indicate that the concentration of longer short chain branches (C8 to C10) is also low in LDPE. REINEPT and DEPT. When method C is employed for measuring LCB in LDPE, the area corresponding to the C6 branch along with the areas from the resonances between that and that of the CE3 should be subtracted from the LCB value assuming these resonances are from SCB. These areas account for about 20% of the peak integral from 31.99 to 32.25 ppm

Figure 7. Theoretical signal enhancement factors of RINEPT and DEPT. ECH, ECH2, and ECH3 are enhancement factor for CH, CH2, and CH3, respectively. E

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 8. RINEPT pulse sequences: (top) this sequence has an adiabatic 180° 13C pulse; (bottom) this sequence has a hard 180° 13C pulse. The green bars represent adiabatic 180° 13C pulses.

factor depends on the last 1H channel pulse width in the DEPT pulse sequence. The goal is to increase CH2 sensitivity of all the signals from 31.99 to 32.25 ppm. RINEPT30−32 and DEPT33−35 were developed many years ago and had been improved since then, and DEPT-135 has been routinely used to aid in peak assignments. There are two RINEPT pulse sequences in Bruker pulse sequence library: Figure 8 (top) shows one with an adiabatic 180° 13C pulse sequence (ineptrdsp), while Figure 8 (bottom) shows the other one with a hard 180° 13C pulse sequence (ineptrd). The parameter CNST11 is used to adjust τ2 for signal enhancement in Bruker RINEPT pulse sequences. The integral increase factor is defined as the ratio of the integral obtained from the RINEPT sequence divided by the integral from the conventional quantitative inverse gated pulse (Bruker pulse sequence is ZGIG). Figure 9 (blue color) shows the integral increase factor for CH2 resonances (29.4−30.8 ppm of an EO sample) for the RINEPT sequence with an adiabatic 180° 13C pulse. Theoretically, the integral increase factor should be equal to the enhancement factor shown in Figure 7. It is influenced by pulse imperfection, signal relaxation, and the integral scale factor of different experiments. The last factor dominates here, so the integral increase factor obtained in Figure 9 is higher than the theoretical enhancement factor shown in Figure 7. However, this is not an issue in this case as only a relative sensitivity increase is being sought. A fourthorder polynomial trend line is added as a visual aid. The highest enhancement is obtained when CNST11 is 9. The integral increase factor for the CH2 resonances (29.4−30.8 ppm of an EO sample) for the RINEPT sequence with a hard 180° 13C pulse is shown in Figure 9 (orange color). A fourthorder polynomial trend line is also added here as a visual guide. With this sequence, the highest enhancement occurs when

Figure 9. Integral increase factor for two RINEPT sequences. (blue) An adiabatic 180° 13C pulse. (orange) A hard 180° 13C pulse. The integral increase factor is defined as the integral from RINEPT divided by the integral from inverse gated pulse (ZGIG). The sample used here is an EO copolymer in TCE-d2 with 0.025 M Cr(acac)3 at 120 °C. Number of scans (NS) = 32.

CNST11 is 8. Therefore, we can conclude that the best CNST 11 for CH2 signal enhancement for the RINEPT sequence with an adiabatic 180° 13C pulse and for the RINEPT sequence with a hard 180° 13C pulse are different. In general, T = 2τ2 is used to describe signal integral increase factor.36 The blue color in Figure 10 shows this relationship for the RINEPT sequence with an adiabatic 180° 13C pulse, while the orange color in Figure 10 is from the RINEPT sequence with a hard 180° 13C pulse. While the blue color adiabatic 180° 13 C pulse reaches a maximum when T is 2.8 ms, the orange color hard 180° 13C pulse reaches a maximum when T is 2.0 ms. It takes less time to reach the maximum integral increase factor with the RINEPT sequence with the hard 180° 13C NMR pulse. F

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 10. Integral increase factor as a function of time (T). Blue color is from RINEPT sequence with an adiabatic 180° 13C pulse and the orange color is from RINEPT with hard 180° 13C pulse. The sample here is an EO copolymer in TCE-d2 with 0.025 M Cr(acac)3 at 120°C. Number of scans (NS) = 32.

Figure 12. CH2 resonances obtained with the four polarization options described in this section along with a ZGIG pulse (the conventional sequence used for quantification) are shown here. DEPT-45 HP: θ = 45°, hard 180° 13C pulse; DEPT-45 SP: θ = 45°, adiabatic 180° 13C pulse. RINEPT HP: hard 180° 13C pulse; RINEPT SP: adiabatic 180° 13C pulse. The sample is an EO copolymer in TCE-d2 with 0.025 M Cr(acac)3 at 120 °C.

There are two Bruker DEPT pulse sequences. Figure 11 (top) shows one with an adiabatic 180° 13C pulse (deptsp), while Figure 11 (bottom) shows the other one with a hard 180° 13C pulse (dept). As can be seen in Figure 7 (green line at the bottom), the highest signal enhancement for CH2 resonances is achieved when θ equals 45° or 135°. CH2 resonances obtained with the four polarization options described above, along with a ZGIG pulse, are shown in Figure 12. It can be seen that the RINEPT sequence with hard 180° 13 C pulse (abbreviated as RINEPT HP) has the least phase distortion under the experimental conditions used. To evaluate the polarization efficiency of different polarization pulse sequences, a DEPT-135 experiment was used to identify CH2 regions. Figure 13 shows the results of a DEPT-

135 experiment on an EO copolymer. The shaded CH2 regions A−C are used for comparison. Table 1 shows the integral increase factor for CH2 regions A−C (Figure 13) obtained via the four different NMR methods: (1) DEPT-45 HP: θ = 45°, hard 180° 13C pulse; (2) DEPT-45 SP: θ = 45°, adiabatic 180° 13C pulse; (3) RINEPT HP: hard 180° 13C pulse; (4) RINEPT SP: adiabatic 180° 13C pulse. From these data in Table 1, it can be seen that the two DEPT experiments lead to smaller signal integral increases than the two RINEPT experiments. One of the reasons for this may be the longer DEPT pulse duration than that in the

Figure 11. DEPT pulse sequences: (top) DEPT sequence with an adiabatic 180° 13C pulse; (bottom) DEPT sequence with a hard 180° 13C pulse. The green bar represents an adiabatic 180° 13C pulse. G

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

the smaller integral increase factor with the phase distortion shown in Figure 12 that was observed with the DEPT experiments, these two DEPT techniques were not further investigated for CH2 signal enhancement. With regards to the RINEPT experiments, the data suggest that the RINEPT sequence with the hard 180° 13C pulse performs better than the RINEPT sequence with the adiabatic 180° 13C pulse in terms of both phase distortion (Figure 12) and integral increase factor (Table 1). Consequently, these two RINEPT experiments were examined further. Table 2 shows the integral increase factor for regions A−C (Figure 13) obtained from four experiments each of the RINEPT sequence with an adiabatic 180° 13C pulse (RINEPT SP) and the RINEPT sequence with a hard 180° 13C pulse (RINEPT HP). For each of the experiments, 1600 scans were performed. From these data, the RINEPT experiment with a hard 180° 13C pulse reproducibly gives a slightly higher integral increase factor, averaging 6.61 ± 0.02, than does the RINEPT experiment with the adiabatic 180° 13C pulse, averaging 5.86 ± 0.03. One possible reason for this difference may be that the RINEPT experiment with a hard 180° 13C pulse is about 2 times shorter than the RINEPT sequence with an adiabatic 180° 13C pulse (see Figure 8). In practice, while the integral increase factor for the RINEPT with a hard 180° 13 C pulse is highest of the evaluated techniques, the RINEPT with an adiabatic 180° 13C pulse may also be used as the difference between the two RINEPT experiments is not dramatic. The RINEPT experiment with an adiabatic 180° 13C pulse might offer some advantages when a large 13C sweep width is needed.36 In RINEPT experiments, the pulse delay (D1 + acquisition time (AQ)) needed for a quantitative measurement is determined by 1H relaxation instead of 13C relaxation, and the pulse delay D1 may be reduced from 6.0 s in a ZGIG pulse to a shorter time. Figure 14 shows 13C NMR spectra resulting from RINEPT experiments with hard 180° 13C pulses where D1 = 6.0 and 1.7 s (AQ was set to 2.62 s). Clearly all the peaks of interest are fully relaxed when D1 = 1.7 s. In other words, the NMR time needed can be reduced by about 2 times (2.26 + 1.7) ∼ (2.26 + 6.0)/2. The signal enhancement factor is 3.32 with the RINEPT sequence with a hard 180° 13C pulse compared with the ZGIG pulse (see S/N in Table 1). Therefore, the total NMR time saved with RINEPT is about 2 × 3.322 = 22 times. Assuming a 5.5 times S/N increase with a cryoprobe over a conventional broad band observe (BBO) probe at 120 °C,28 another S/N increase of 3.3 with

Figure 13. DEPT-135 pulse sequence used on an EO copolymer in TCE-d2 with 0.025 M Cr(acac)3 at 120 °C. The shaded CH2 regions A−C are used for evaluation of the integral increase factor with different NMR methods. 13C transmitter frequency is set to 30 ppm.

Table 1. Integral Increase Factor for CH2 Regions A−C (Figure 13) Obtained via Different NMR Methods integral increase factor for regions A−C

region NMR method ZGIG RINEPT SP-CH2 RINEPT HP-CH2 DEPT-45 SP DEPT-45 HP

A

B

C

average increase: all regions

standard deviation: all regions

S/Na

1 5.8

1 5.9

1 5.8

1 5.9

0 0.1

12538 40931

6.6

6.6

6.5

6.6

0.1

41688

4.3

3.8

3.7

3.9

0.3

24777

5.1

4.8

4.6

4.8

0.2

34653

a

S/N was measured with a Bruker Topspin 3.5 Pl7; highest peak in region C, noise 0 to −20 ppm, and noise range 2 ppm were used.

RINEPT sequence (see Figures 8 and 11). It is also noted that DEPT-45 HP performed better than DEPT-45 SP. Combining

Table 2. Integral Increase Factor for Regions A−C (Figure 13) Obtained from RINEPT Experiments with an Adiabatic 180° 13 C Pulse (RINEPT SP) and with a Hard 180° 13C Pulse (RINEPT HP) region NMR method and run number ZGIG RINEPT RINEPT RINEPT RINEPT RINEPT RINEPT RINEPT RINEPT

SP-CH2-1 SP-CH2-2 SP-CH2-3 SP-CH2-4 HP-CH2-1 HP-CH2-2 HP-CH2-3 HP-CH2-4

integral increase factor for regions A−C

A

B

C

average increase: all regions

standard deviation: all regions

1 5.83 5.85 5.88 5.79 6.65 6.66 6.70 6.63

1 5.92 5.95 5.96 5.90 6.65 6.67 6.70 6.67

1 5.82 5.83 5.84 5.79 6.50 6.51 6.52 6.50

1 5.86 5.87 5.89 5.83 6.60 6.62 6.64 6.60

0 0.06 0.07 0.06 0.06 0.09 0.09 0.10 0.09 H

average increase: all runs

standard deviation: all runs

5.86

0.03

6.61

0.02

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

acquisition of a quantitative 13C NMR spectrum using a pulse sequence such as ZGIG with a small number of scans (quick NMR acquisition) and, second, the acquisition of a RINEPT spectrum to enhance the sensitivity of the desired type(s) of carbon (CH or CH2 or CH3), such as CH2 in the case discussed above. With the ZGIG spectrum (Figure 15), we can set the integral of all peaks to 1000 and, then, for example measure the integral from a CH2 peak that has a S/N greater than the limit of quantification (LOQ; S/N = 10), assumed to be 98 in this example. With the ZGIG pulse alone, it is not possible to obtain good quantification for the CH2 peak at left in Figure 15 because the S/N is less than LOQ for the CH2 peak. With RINEPT focusing on CH2, it is possible to boost the CH2 sensitivity by a factor of 3.3 and reduce the NMR time by 22 times as discussed above. Now that the S/N of the CH2 peak at left in Figure 15 is well above LOQ, the integral of the CH2 peak at right in Figure 15 can be set to 98 based on the value obtained from the ZGIG pulse, and the corresponding integral Y for the CH2 peak at left in Figure 15 can now be obtained. The microstructure represented by this CH2 signal can be expressed as Y/1000C. A similar idea of using a combination of a quick quantitative 1H NMR and a sensitivity enhancement 1H NMR has been extensively used in industry in the past to measure unsaturation in polyolefins.37−42 This approach can also be used to obtain information about CH and CH3 in the material of interest. It is likely to work for other nuclei such as 15N and 29Si. The key advantages are (1) it is applicable to samples that have carbons without attached protons (although these carbons’ signals cannot be enhanced), such as LDPE, poly(ethylene-co-acrylic acid), and EPDM; (2) it should also work for functional groups which have drastically different C−H J coupling constants, such as vinyl groups; and (3) the tolerance of error of T (CNST 11 in Bruker pulse sequence) in RINEPT and the error of pulse width in DEPT are much higher, as shown in Figure 7. In other words, time T in RINEPT and θ pulse width in DEPT do not need to be precise to get very good signal enhancement. An example of this technique applied to a real material is shown in Figure 16 in which a typical LDPE sample is analyzed. The peak at 32.669 ppm arising from a CH2 in a C5 branch is used as the reference peak. With a ZGIG pulse (Figure 16, top and middle), the S/N of this peak is 20 with a 58 min NMR acquisition time, and the integral of this peak is 2.67/1000C (the LCB measured with method C discussed earlier is 3.41/1000C). The resonance arising from C6 branches is at 32.177 ppm, and the CE3 resonance (Scheme 1) in this LDPE sample is at 32.122 ppm. It can be seen from the expanded portion of the spectrum from the ZGIG pulse (Figure 16, middle) that the S/N of the peak of the C6 branch and that of the shoulder peak on the left side of the CE3 peak is only about 2.5, which is far below a LOQ value of 10. However, when RINEPT is used to maximize CH2 enhancement (Figure 16, bottom), the peak at 32.669 ppm has a S/N of 260 with 6.3 h acquisition, and now the S/N of the peaks from the C6 branch and shoulder peak on the left side of the CE3 peak is well above a LOQ value of 10. After setting the integral of the peak at 32.669 ppm equal to 2.67 in the RINEPT spectrum, we can obtain LCB = 3.41/1000C with method C discussed earlier. However, as discussed earlier, the resonance arising from the C6 branch and those peaks between the C6 branch and the CE3 peak should be subtracted from 3.41, assuming these peaks are from SCB. The contribution of

Figure 14. Comparison of the effect of varying delay times D1 with a RINEPT sequence with a hard 180° 13C pulse (acquisition time was set to 2.62 s). The sample is an EO copolymer in TCE-d2 with 0.025 M Cr(acac)3 at 120 °C.

polarization transfer, and a further 50% reduction of NMR time needed through pulse delay (see discussion above), the NMR time saved could be equal to 2 × 5.52 × 3.32 = 659 times through this combination. Quantification of LCB in LDPE with Quick Conventional NMR and RINEPT. Although RINEPT can boost sensitivity and reduce NMR time substantially, it can be seen from Figure 7 that the sensitivity enhancements for CH (blue line), CH2 (orange line), and CH3 (gray line) resonances are different at any time. Hou et al. proposed to find the “good point”,36 a specific time when the enhancements for CH2 and CH3 resonances are equal (T = 2.92 ms was reported with RINEPT with an adiabatic 180° 13C pulse by the authors). A theoretical enhancement factor can then be applied to CH resonances to make CH, CH2, and CH3 resonances have the same enhancement factor to enable quantification. However, because LDPE has quaternary carbons where signals will disappear with RINEPT, it is therefore difficult to analyze LCB content in LDPE with RINEPT directly. Here we propose a different approach for quantification of microstructures in polymers: the use of a combination of a conventional quantitative pulse such as ZGIG with RINEPT (DEPT may also work for certain situations). This approach, as shown in Figure 15, has two experimental steps: first, the

Figure 15. A combination of a conventional quantitative pulse such as ZGIG and a polarization transfer technique (PT) such as RINEPT or DEPT for quantification of microstructures in polymers. LOQ: limit of quantification. I

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 16. 13C NMR spectra of a typical LDPE sample in Cl-NA/PDCB-d4 (9:1, w:w) with 0.025 M Cr(acac)3 at 120 °C. The top and middle are spectra from a ZGIG pulse (AQ = 2.7 s, D1 = 6 s, NS = 400). The bottom spectrum is from RINEPT focusing on CH2 (AQ = 2.7 s, D1 = 1.7 s, NS = 5120).

the C6 branch is 0.31/1000C, and the contribution of those peaks between the C6 branch and the CE3 peak is 0.34/1000C, which is about 20% of the 3.41 value. So the LCB content in this typical LDPE is actually 2.76/1000C ((3.41−0.31−0.34)/ 1000C). Branches of this typical LDPE are summarized in Table 3. In conclusion, C6 branches were identified for the first time in LDPE via 13C NMR spectroscopy using 1-chloronaphthalene as the NMR solvent. In addition, on the basis of experimental and computational results, it is concluded that

the number of longer short chain branches (C7 and above) is much lower compared to the amount of C6 branches present in LDPE. Furthermore, the presence of ethyl−octyl branches is hypothesized based on the data presented here, and this will be the subject of future studies. As was discussed, all peaks corresponding to SCB should be excluded from calculations of LCB content in LDPE, as SCB do not have the same impact on polymer properties as do LCB. Different polarization transfer techniques were evaluated, and it was found that a RINEPT sequence with hard 180° 13C pulses is the best method for signal enhancement used in the quantification of LCB in LDPE under the experimental conditions used here. A much faster method for quantifying LCB in LDPE is demonstrated to be significantly faster than the conventional quantitative ZGIG method. This method can also be used to measure other functional groups in polymers, which have drastically different C−H J coupling constants, such as vinyl and aromatic groups. It is likely to work for other nuclei such as 15N and 29Si, as well.

Table 3. Branches/1000C of a Typical LDPEa C4

C5

C6

C*

LCB6+

LCB7+

LCB**

8.02

2.67

0.31

0.34

3.41

3.10

2.76

a C*: peaks between the C6 branch and the CE3 peak; LCB6+: LCB value obtained with method C; LCB7+: LCB value obtained with method C minus contribution from C6 branches; LCB**: LCB value obtained with method C minus contributions from C6 branch and from peaks between the C6 branches and the CE3 peak.

J

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

■

(16) Usami, T.; Takayama, S. Fine-branching structure in highpressure, low-density polyethylenes by 50.10-MHz 13C NMR analysis. Macromolecules 1984, 17, 1756−1761. (17) Zhou, Z.; Baugh, D.; Fontaine, P.; He, Y.; Shi, Z.; Mukhopadhyay, S.; Cong, R.; Winniford, B.; Miller, M. Long-chain branch measurement in substantially linear ethylene polymers by 13C NMR with halogenated-naphthalenes as solvents. Macromolecules 2017, 50, 7959−7966. (18) Becke, A. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652. (19) Lee, C.; Yang, W.; Parr, R. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (20) Ditchfield, R.; Hehre, W.; Pople, J. Self-consistent molecular orbital methods. IX. An extended Gaussian-type basis for molecularorbital studies of organic molecules. J. Chem. Phys. 1971, 54, 724− 728. (21) Hehre, W.; Ditchfield, R.; Pople, A. Self-consistent molecular orbital methods. XII. Further extensions of GaussianType bsis sets for use in molecular orbital studies of organic molecules. J. Chem. Phys. 1972, 56, 2257−2261. (22) Cuccato, D.; Dossi, M.; Polino, D.; Cavallotti, C.; Moscatelli, D. Is quantum tunneling relevant in free-radical polymerization? Macromol. React. Eng. 2012, 6, 496−506. (23) Liu, S.; Srinivasan, S.; Grady, M.; Soroush, M.; Rappe, A. Backbiting and β-scission reactions in free-radical polymerization of methyl acrylate. Int. J. Quantum Chem. 2014, 114, 345−360. (24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian, Inc.: Wallingford, CT, 2009. (25) This method measures saturated, straight aliphatic chains/ branches and does not count chain ends with unsaturation, i.e., vinyls, nor does it count chain ends terminated by certain chain transfer agents, such as propionaldehyde. (26) Mattice, W.; Stehling, F. Branch formation in low-density polyethylene. Macromolecules 1981, 14, 1479−1484. (27) Toh, J.; Huang, D.; Lovell, P.; Gilbert, R. Ab initio calculation of the rate coefficient for short chain branching in free-radical polymerizations. Polymer 2001, 42, 1915−1920. (28) Zhou, Z.; Kuemmerle, R.; Stevens, J. C.; Redwine, D.; He, Y.; Qiu, X.; Cong, R.; Klosin, J.; Montanez, N.; Roof, G. 13C NMR of polyolefins with a new high temperature 10 mm cryoprobe. J. Magn. Reson. 2009, 200, 328−333. (29) Zhou, Z.; Stevens, J. C.; Klosin, J.; Kummerle, R.; Qiu, X.; Redwine, D.; Cong, R.; Taha, A.; Mason, J.; Winniford, B.; Chauvel, P.; Montanez, N. NMR Study of Isolated 2,1-Inverse Insertion in Isotactic Polypropylene. Macromolecules 2009, 42, 2291−22294. (30) Morris, G.; Freeman, R. Enhancement of nuclear magnetic resonance signals by polarization transfer. J. Am. Chem. Soc. 1979, 101, 760−762. (31) Burum, D.; Ernst, R. Net polarization transfer via a J-ordered state for signal enhancement of low-sensitivity nuclei. J. Magn. Reson. 1980, 39, 163−168. (32) Sorensen, O.; Ernst, R. Elimination of spectral distortion in polarization transfer experiments. Improvements and comparison of techniques. J. Magn. Reson. 1983, 51, 477−489.

AUTHOR INFORMATION

Corresponding Author

*(Z.Z.) Tel 1-979-238-1387; Fax 1-979-238-0752; e-mail [email protected]. ORCID

Zhe Zhou: 0000-0003-0869-2112 Jerzy Klosin: 0000-0002-9045-7308 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

The authors thank Linh Le, Kebede Beshah, Leslie Fan, Dave Meunier, Wei Gao, Tony Gies, and Arkady Krasovskiy of The Dow Chemical Company for very helpful discussions.

(1) Rungswang, W.; Narkchamnan, K.; Petcharat, N.; Thitisak, B.; Pathaweeisariyakul, T. Primitive structure and its morphology for describing highly branched structure of low-density polyethylene. Polym. Bull. 2017, 74, 3229−3242. (2) Spalding, M. A.; Chatterjee, A. M. Handbook of Industrial Polyethylene and Technology; Wiley Scrivener Publishing: 2017. (3) Kouda, S. Prediction of processability at extrusion coating for low-density polyethylene. Polym. Eng. Sci. 2008, 48, 1094−1102. (4) Vega, J.; Aguilar, M.; Peon, J.; Pastor, D.; Martinez-Salazar, J. Effect of long chain branching on linear-viscoelastic melt properties of polyolefins. e-Polym. 2002, 046. (5) Graessley, W. W. Effect of long branches on the temperature dependence of viscoelastic properties in polymer melts. Macromolecules 1982, 15, 1164−1167. (6) Zhou, Z.; Kuemmerle, R.; Qiu, X.; Redwine, D.; Cong, R.; Taha, A.; Baugh, D.; Winniford, B. A new decoupling method for accurate quantification of polyethylene copolymer composition and triad sequence distribution with 13C NMR. J. Magn. Reson. 2007, 187, 225−233. (7) Randall, J. C. A review of high resolution liquid 13carbon nuclear magnetic resonance characterizations of ethylene-based polymers. J. Macromol. Chem. Phys. 1989, C9, 201−317. (8) Wood-Adams, P.; Dealy, J.; deGroot, W.; Redwine, D. Effect of molecular structure on the linear viscoelastic behavior of polyethylene. Macromolecules 2000, 33, 7489−7499. (9) Baugh, D.; Redwine, D.; Taha, A.; Reichek, K.; Potter, J. Using solvents to improve the chemical shift differences between short-chain branch methines and long-chain branch methines in polyethylene copolymers. Macromol. Symp. 2007, 257, 158−161. (10) Stapleton, R.; Chai, J.; Nuanthanom, A.; Flisak, Z.; Nele, M.; Ziegler, T.; Rinaldi, P.; Soares, J.; Collins, S. Synthesis of LDPE using nickel iminophosphonamide complexes. Macromolecules 2007, 40, 2993−3004. (11) Bugada, D.; Rudin, A. Long chain branching indices of LDPE from GPC and 13C NMR. Eur. Polym. J. 1987, 23, 847−850. (12) Axelson, E.; Levy, G.; Mandelkern, L. A quantitative analysis of LDPE by carbon-13 FT-NMR at 67.9 MHz. Macromolecules 1979, 12, 41−52. (13) Usami, T.; Gotoh, Y.; Takayama, S. Sizes of LCB in a LDPE as a function of molecular weight. J. Appl. Polym. Sci. 1991, 43, 1859− 1863. (14) Nordmeier, E.; Lanver, U.; Lechner, M. The molecular structure of LDPE. 1. LCB and solution properties. Macromolecules 1990, 23, 1072−1076. (15) Li, P.; Xue, Y.; Wu, X.; Sun, G.; Ji, X.; Bo, S. Microstructure characterization of one high-speed extrusion coating polyethylene resin fractionated by solvent gradient fractionation. J. Polym. Res. 2018, 25, 113−122. K

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (33) Bendall, M.; Doddrell, D.; Pegg, D. Editing of carbon-13 NMR spectra. 1. A pulse sequence for the generation of subspectra. J. Am. Chem. Soc. 1981, 103, 4603−4605. (34) Doddrell, D.; Pegg, D.; Bendall, M. Distortionless enhancement of NMR signals by polarization transfer. J. Magn. Reson. 1982, 48, 323−327. (35) Schenker, K.; Von Philipsborn, W. Off-resonance effects and their compensation in the multiple-pulse sequences INEPT, DEPT, and INADEQUATE. J. Magn. Reson. 1986, 66, 219−229. (36) Hou, J.; He, Y.; Qiu, X. Speedy, Robust and quantitative analysis of polyolefins using sensitivity-enhanced 13C NMR spectroscopy. Macromolecules 2017, 50, 2407−2414. (37) Zhou, Z.; Cong, R.; He, Y.; Paradkar, M.; Demirors, M.; Cheatham, M.; deGroot, W. Unsaturation characterization of polyolefins by NMR and thermal gradient NMR with a high temperature cryoprobe. Macromol. Symp. 2012, 312, 88−96. (38) Hermel, T.; Demirors, M.; Hayne, S.; Cong, R. Ethylene-based polymer compositions. US 8,729,200B2, 2014. (39) Karjala, T.; Kardos, L.; Yau, W.; Ortega, J.; Vigil, A. Ethylenebased polymers and process to make the same. US 9,303,107B2, 2016. (40) Ewart, S.; Karjala, T.; Zogg, M.; Munjal, S. Free-radical processes to make ethylene-based polymers using alkylated phenols. US 9,328,180B2, 2016. (41) Demirors, M.; Kapur, M.; Jian, P.; Fontaine, P.; Ginger, D.; Gillespie, D.; Bilgen, M. Polyolefin composition. US 9,631,059B2, 2016. (42) Effler, J.; Karjala, T.; Demirors, M.; Savargaonkar, R.; Bensason, S.; Zhou, Z. Process to produce enhanced melt strength ethylene/ alpha-olefin copolymers and articles thereof. US 9,908,956, 2018.

L

DOI: 10.1021/acs.macromol.8b01806 Macromolecules XXXX, XXX, XXX−XXX