Article Cite This: Macromolecules XXXX, XXX, XXX−XXX
pubs.acs.org/Macromolecules
Atypical Multiple Site Behavior of Hafnocene Catalysts in Ethylene/ 1-Hexene Copolymerization Using Trioctylaluminum and Borate Saeid Mehdiabadi,† João B. P. Soares,*,† and Jeffrey Brinen‡ †
Department of Chemical and Materials Engineering, University of Alberta, Edmonton, AB T6G 2R3, Canada ExxonMobil Chemical Company, Baytown, Texas 77522, United States
Macromolecules Downloaded from pubs.acs.org by UNIV OF SOUTH DAKOTA on 09/07/18. For personal use only.
‡
ABSTRACT: Solution copolymerizations of ethylene and 1hexene were performed using the dichloride (HfCl2) and dimethyl (HfMe 2 ) analogues of the bis(n-propylcyclopentadienyl)hafnium precatalyst. Tetrakis(pentafluorophenyl) borate dimethylanilinium salt ([B(C6F5)4]−[Me2NHPh]+) (B) or methylaluminoxane (MAO) was employed as activator, and tri-n-octylaluminum (TOA) was used as scavenger. Crystallization analysis fractionation (CRYSTAF) or crystallization elution fractionation (CEF) was used to measure the chemical composition distributions (CCD). Above a minimum 1-hexene threshold concentration, the CCD profiles of all copolymers were bimodal, and the areas under the peaks depended on the B/HfCl2 and B/Al ratios. Decreasing the TOA concentration reduced the weight fraction of the higher crystallinity copolymer. Copolymerizations using MAO also produced copolymers with bimodal CCDs. Copolymerizations using HfMe2/B without TOA produced copolymers with unimodal CCD. These results suggest that TOA reacts with either the catalyst or the activator, generating a new active site with a lower propensity for incorporating comonomer than the original catalyst system.
■
INTRODUCTION
Alternatively, perfluoroarylboron-based activators, such as tris(pentafluorophenyl)borane, B(C6H5)3, or anilinium salts, such as [PhNR 2 H] + [B(C 6 F 5 ) 4 ] − (R = Me, Et) and triphenylcarbenium tetrakis(pentafluoropheny1)borate, [CPh3]+[(BC6F5)4]−, can also activate metallocenes.12−19 Their main function is to abstract an alkyl group from dialkylmetallocenes and stabilize the resulting olefin polymerization catalyst site.9,17,20−32 In contrast with MAO, these cocatalysts can be used in near-stoichiometric proportions with the precatalyst to yield catalysts that have similar activities than those activated with MAO. When borates are used as cocatalysts, alkylaluminum or aluminoxanes may also be added to the reactor for two main reasons: (1) if the precatalyst is supplied as a chlorinated complex, the alkylaluminum is required to alkylate the precatalyst before it can be reduced by the borate cocatalyst, and (2) the alkylaluminum is also needed to scavenge polar impurities that may deactivate the catalyst sites formed by the reaction of the precatalyst and the cocatalyst. The addition of this third component, however, may lead to surprising catalyst behavior, which is at the center of the present investigation. In previous publications from our group,33,34 the effects of activating rac-dimethylsilylbis(indenyl)hafnium dimethyl with MAO or tetrakis(pentafluorophenyl) borate dimethylanilinium salt were compared using statistical techniques to estimate the
Polyolefins are the synthetic polymers with the farthest impact in the commodity market, accounting for more than 60% of the total production of the plastic industry today. Except for low-density polyethylene, LDPE, all other polyolefins are made with transition metal coordination catalysts. These catalysts (excluding some chromium oxide catalysts) are formed when two components are combined: the precatalyst and the cocatalyst. Precatalysts (often called “catalysts” for simplicity) are organometallic transition metal complexes, and cocatalysts are typically alkylaluminum compounds, aluminoxanes, or borates. Even though metallocenes are as old as Ziegler−Natta catalysts,1 they remained mostly an academic curiosity until the mid-1970s, when it was shown that they could make polyethylene with high productivities if alkylaluminums were replaced with methylaluminoxane (MAO) as the cocatalyst.2,3 This serendipitous discovery led to a true revolution in polyolefin production technology.4,5 Methylaluminoxane remains the most common cocatalyst for the production of polyolefins with metallocenes, but its structure is still controversial.6−8 MAO alkylates the transition metal−chloride bonds, abstracts the second chloride atom to yield a metallocenium cation with a vacant coordination site that is active for olefin polymerization, and scavenges polar impurities that may poison the catalyst.9−11 Unfortunately, especially in solution polymerization, a large MAO/metallocene ratio is needed to achieve high catalyst activities. © XXXX American Chemical Society
Received: May 15, 2018 Revised: August 23, 2018
A
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
polymer solution was heated to 160 °C and held for 2 h to ensure complete dissolution, followed by decreasing the temperature to 105 °C and stabilizing it for 55 min. A constant cooling rate of 0.1 °C/min was applied during analysis until the final temperature reached 30 °C. The polymer concentration in solution was monitored using an in-line infrared detector. Crystallization elution fractionation (CEF) was performed using a CEF instrument from Polymer Char. The samples were prepared in TCB stabilized with 300 ppm of 2,6-di-tert-butyl-4-methylphenol, at a concentration of 1 mg/mL. The volume of the injected sample was 200 μL. The polymer solution was stabilized at 95 °C for 5 min before being injected in the CEF column. The cooling cycle started by decreasing the column temperature from 95 to 35 °C under a constant cooling rate of 2 °C/min and a constant crystallization flow rate of 0.03 mL/min. At the end of the cooling cycle, the column temperature was held at 35 °C for 3 min under a constant elution flow rate of 1 mL/min. The elution cycle started with a heating rate of 3 °C/min and temperature changing from 35 to 140 °C. The concentration of polymer in the eluent was monitored as a function of the elution temperature by a dual wavelength infrared detector placed at the exit of the CEF column. The dimensions of the CEF column was 15 cm length and 9.5 mm id. The CEF column was filled with inert stainless-steel shots. CRYSTAF and CEF generate essentially the same microstructural information, and were used interchangeably in the current investigation. Cross fractionation chromatography (CFC) was performed in a CFC instrument from Polymer Char. In CFC, two fractionation techniques, temperature rising elution fractionation (TREF) and GPC, are combined to measure the bivariate distribution of chemical composition and molecular weight of polyolefins. A 32 mg sample was dissolved in 8 mL of TCB at 160 °C for 1 h. A sample volume of 400 μL was injected into the TREF column, followed by the crystallization step at a cooling rate of 0.5 °C/min, during which the polymer is fractionated by the deposition of polymer layers of decreasing crystallinity onto the TREF column support. The dimensions of the TREF column were 15 cm length and 9.5 mm id. After the crystallization step, the discontinuous elution process was started, at increasing temperature isothermal steps of 3 °C. For each 3 °C temperature interval, and after a given dissolution time, an injection valve was switched to allow the solvent to elute the polymer that dissolved at that temperature interval from the TREF column. The elution step started at 35 °C and finished at 120 °C. Size exclusion chromatography of each fraction eluted from the TREF column was performed using three GPC columns similar to those discussed above for molecular weight determination and under the analysis same conditions.
main polymerization kinetic parameters for these systems. This line of research was extended in a subsequent publication to study the kinetics of ethylene polymerization using the bis(npropylcyclopentadienyl)hafnium dichloride/borate/TOA system.35 The present investigation continues our efforts to elucidate these intriguing systems by by providing additional details on the copolymerization of ethylene and 1-hexene using bis(n-propylcyclopentadienyl)hafnium dichloride (HfCl2) and bis(n-propylcyclopentadienyl)hafnium dimethyl (HfMe2) activated with tetrakis(pentafluorophenyl) borate dimethylanilinium salt (B), in the presence of trioctylaluminum (TOA). In this paper, we show for the first time how CCD bimodality can be tuned by varying the ratio of catalyst/TOA and catalyst/B for either one of the molecular catalysts studied in this investigation.
■
EXPERIMENTAL SECTION
Materials. Ethylene and high purity nitrogen (Praxair) were flown through one bed packed with a mixture of 4A and 13X sieves and another with copper(II) oxide. HPLC grade toluene (Aldrich) was refluxed over metallic sodium for 40 h and distillated under a nitrogen atmosphere. Triisobutylaluminum (TIBA) and trioctylaluminum (TOA) were purchased from Sigma-Aldrich. The catalysts bis(npropylcyclopentadienyl)hafnium dichloride (HfCl2) and bis(npropylcyclopentadienyl)hafnium dimethyl (HfMe2) and cocatalyst tetrakis(pentafluorophenyl) borate dimethylanilinium salt (B) were provided by ExxonMobil. All air-sensitive compounds were handled under inert atmosphere in a glovebox. Polymer Synthesis. All polymerizations were conducted in a 500 mL Parr autoclave reactor operated in semibatch mode. The polymerization temperature was controlled using an electrical band heater and an internal cooling coil. The reaction medium was mixed using a pitched-blade impeller connected to a magneto-driver stirrer, rotating at 1500 rpm. Prior to use, the reactor was heated to 125 °C, evacuated and refilled with nitrogen six times, and charged with 250 mL of toluene and 0.5 g of TIBA as a scavenger. The reactor temperature was then set to 120 °C and kept constant for 20 min. Finally, the reactor contents were blown out under nitrogen pressure. In a typical polymerization run, 200 mL of toluene was charged into the reactor, followed by the desired amount of TOA, introduced via a 10 mL vial at room temperature. The comonomer 1-hexene was then introduced into the reactor using a 50 mL vial and a needle under ethylene pressure. The reactor temperature was then increased to 120 °C, and ethylene was supplied to saturate the toluene to the desired pressure. The catalyst and borate solutions were transferred into the reactor via a 5 mL tube and a 20 mL sampling cylinder connected in series using an ethylene pressure differential of 40 psig. 5 mL of toluene was placed in the sampling cylinder before injection to wash the tube walls from any remaining catalyst solution. Ethylene was supplied on demand to maintain a constant reactor pressure and monitored with a mass flow meter. With the exception of a 1−3 °C fluctuation upon catalyst injection, the temperature was kept within 120 ± 0.2 °C throughout the polymerizations. After the required time, the polymerization was stopped by closing the ethylene feed valve and immediately blowing out the reactor contents into a 2.0 L beaker filled with 400 mL of ethanol. The polymer suspension was kept overnight, filtered, washed with ethanol, dried in air, and further dried under vacuum. Polymer Characterization. Molecular weight distributions (MWD) were determined with a Polymer Char high temperature gel permeation chromatograph (GPC), run at 145 °C under a 1 mL/ min flow rate of 1,2,4-trichlorobenzene (TCB). The GPC was equipped with three detectors in series (infrared, light scattering, and differential viscometer) and calibrated with polystyrene narrow standards. Crystallization analysis fractionation (CRYSTAF) was performed using a Polymer Char CRYSTAF model 200. Polymer samples were dissolved in 47 mL of TCB at a concentration of 0.6 mg/mL. The
■
RESULTS AND DISCUSSION Ethylene and 1-Hexene Copolymerization with HfCl2/ B/TOA. Twelve copolymerizations were carried out, in random order, at eight different equally spaced 1-hexene concentrations with HfCl 2 /B/TOA, composed of bis(n-propylcyclopentadienyl)hafnium dichloride, tetrakis(pentafluorophenyl) borate dimethylanilinium salt (B), and trioctylaluminum (TOA). Table 1 summarizes the polymerization conditions, yields, and polymer molecular weight averages for these experiments. Polymer samples were named according to the convention: H-16B means an ethylene/1-Hexene copolymer made with 16 g of 1-hexene in replicate polymerization B. Figure 1 illustrates the effect of 1-hexene feed concentration on polymer yield. As the concentration of 1-hexene increases, the insertion of ethylene into the growing polymer chains with bulkier 1-hexene ends is slower than into chains terminated with ethylene units. Consequently, the polymerization rate slows down, and the polymer yield drops as more 1-hexene is added to the reactor. B
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Table 1. Polymerization Conditions, Yields, and Polymer Molecular Weight Averages for HfCl2/B/TOAa sample
1-hexene (g)
1-hexene in feed (mol/L)
yield (g)
Mn
Mw
PDI
H-0A H-0B H-0C H-8 H-12 H-16A H-20 H-24 H-4 H-28 H-16B H-16C
0 0 0 8 12 16 20 24 4 28 16 16
0.000 0.000 0.000 0.421 0.631 0.841 1.052 1.262 0.210 1.472 0.841 0.841
4.86 4.90 4.42 4.35 3.38 3.39 2.65 2.11 4.91 1.68 2.92 2.95
133900 126400 131600 63800 49100 40600 35100 32400 87800 26600 38400 41600
267600 263000 265900 140000 110100 89400 77700 68800 180300 56900 84000 89500
2 2.08 2.02 2.19 2.24 2.2 2.21 2.12 2.05 2.14 2.19 2.15
Figure 2. Effect of 1-hexene concentration on ethylene uptake profiles (legends are 1-hexene reactor concentrations) with HfCl2/B/TOA.
[TOA] = 932 μmol/L, [Hf] = 635 nmol/L, [B] = 1262 nmol/L, B/ Hf = 2, TOA/Hf = 1470, T = 120 °C, P = 120 psig, tp = 10 min, and total liquid feed volume (solvent + 1-hexene) = 226 mL. The sample labels indicate the amount of 1-hexene loading and replicate number. For instance, H-16A indicates ethylene copolymerization (H stands for 1-hexene) using 16 g of 1-hexene. Letter A indicates that this run is the first run, and further replicates are shown using letters B and C. a
Figure 3. Effect of 1-hexene concentration on ln(FM,in/VR) profiles (legends are 1-hexene reactor concentrations) with HfCl2/B/TOA.
merizations at low 1-hexene concentrations, the ethylene uptake profiles obey first-order behavior, but not for higher 1hexene concentrations. These copolymerizations cannot be explained with the polymerization kinetics model described by eqs 1 and 2. This deviation may be attributed to the presence of more than one active site type, as will be demonstrated later. Figure 4 compares the CRYSTAF profiles of selected samples from Table 1. Unimodal chemical composition distributions (CCD) are observed for homopolymers and copolymers made with 1-hexene concentrations less than or equal to 0.21 mol/L. Increasing the 1-hexene concentration in the feed causes the composition profiles to broaden and eventually become bimodal for higher 1-hexene concentrations, suggesting that at least two types of sites were active during the copolymerizations. For the bimodal profiles, the crystallization peak temperatures (Tc) of both peaks decreased with increasing 1-hexene concentration (Figure 5); both site types, therefore, behaved as expected: when more 1-hexene was added to the reactor, more 1-hexene was incorporated in the copolymer. Because CRYSTAF peak temperatures can be correlated to the fraction of 1-hexene in the copolymer using a linear calibration curve that relates Tc to 1-hexene content, Figure 4 shows that ethylene/1-hexene copolymers made with HfCl2/
Figure 1. Effect of 1-hexene concentration on polymer yield for HfCl2/B/TOA.
When catalyst deactivation and ethylene propagation rates follow first-order kinetics and also assuming that the site activation is instantaneous, the following equation describes ethylene uptake curves35,36 FM,in = k p[Ci]0 [M]VR e−kdt
(1)
where FM,in, is the molar flow rate of ethylene to the reactor, [M] is ethylene concentration in the liquid phase, [Ci]0 is the initial catalyst concentration, and VR is the volume of the reaction medium. Equation 1 can be linearized through the simple transformation
ji FM,in zyz lnjjj z = ln(k p[Ci]0 [M]) − kdt j VR zz (2) k { Figure 2 compares all ethylene uptake curves, and Figure 3 compares their corresponding ln(FM,in/VR) × t plots. For ethylene homopolymerizations and ethylene/1-hexene copoly-
C
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 6. Ratio of the areas under high (AH) and low (AL) CRYSTAF peak temperatures for copolymers made with HfCl2/B/TOA.
Figure 4. CRYSTAF profiles of selected polymers from Table 1. The z-axis represents the 1-hexene concentration in the reactor feed in mol/L.
Figure 7. Effect of 1-hexene concentration on molecular weight averages and polydispersity index of copolymers made with HfCl2/B/ TOA.
Figure 5. CRYSTAF peak temperatures versus 1-hexene concentration in the feed for the polymerizations listed in Table 1.
B/TOA have bimodal CCD beyond a certain 1-hexene concentration for the polymerization conditions investigated herein. Interestingly, the ratio between the areas under the low (AL) and high (AH) CRYSTAF temperature peaks was almost constant (AL/AH ≅ 3.0) for all copolymers, decreasing slightly only at the highest 1-hexene concentrations, as shown in Figure 6. This implies that the fraction of polymer made on the two site types does not vary substantially for different 1-hexene concentrations in the reactor. Figure 7 shows that Mn and Mw decrease when more 1hexene is added to the reactor, likely due to decrease in apparent propagation rate constant, as expected since 1-hexene is less reactive than ethylene. The polydispersity index increases marginally when the CRYSTAF profiles start to become bimodal but remains nearly constant after this threshold value. The slight increase in PDI is another indication that a second site type may become active in this catalyst system at higher 1-hexene concentrations. The presence of two polymer populations, likely made on two distinct active site types, comes to life in the CFC profiles for samples H-12, H-16B, H-24, and H-28, depicted in Figure 8. Note that the elution temperature (Te) in the CFC plots
Figure 8. CFC plots for sample H-12, H-16B, H-24, and H-28.
correlates linearly with 1-hexene content in the copolymer: the lower the Te, the higher the 1-hexene incorporation in the copolymer. Surprisingly, all of these samples have bimodal CFC profiles, despite being produced in solution polymerization with a discrete molecular catalyst. The high temperature peak curiously “stands in front” of the low temperature peak when viewed perpendicularly to the molecular weight plane. This means that both polymer populations have approximately the same D
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
relationship between the mass of 1-hexene added to the reactor and the resulting 1-hexene concentration in the liquid phase is almost linear under the conditions investigated herein. Table 2 summarizes CRYSTAF peak temperatures and their corresponding 1-hexene contents, calculated using the
MWD, which explains why their PDIs are close to 2.0, as measured by GPC. Reactivity Ratios for Each Site Type of HfCl2/B/TOA. The CRYSTAF peak temperature data (Figure 5) were used with the CRYSTAF calibration curve relating Tc to 1-hexene fraction in the copolymer, the Mayo−Lewis equation, and a thermodynamic equation of state to calculate 1-hexene and ethylene concentrations in the liquid phase at the polymerization conditions and to estimate the reactivity ratios for ethylene/1-hexene using HfCl2/B/TOA. The CRYSTAF calibration curve is given by the expression F2 = −1.317 × 10−3Tc + 1.12467 × 10−1
Table 2. CRYSTAF and Copolymer Composition Data for Copolymers Made with HfCl2/B/TOAa
(3)
where F2 is the molar fraction of 1-hexene in the copolymer and Tc is the CRYSTAF crystallization temperature. The Mayo−Lewis equation, relating the mole fraction of ethylene in the copolymer, F1, to the ethylene mole fraction in the feed, f1, is expressed as F1 =
r1f12 + f1 f2 r1f12 + 2f1 f2 + r2f2 2
(4)
where r1 and r2 are the reactivity ratios, defined as k r1 = 11 , k12
k r2 = 22 k 21
M1 M1 + M 2
1-hexene (mol/L)
Tc,max (°C)
Tc,min (°C)
H-0A H-0B H-0C H-8 H-12 H-16A H-20 H-24 H-4 H-28 H-16B H-16C
0.000 0.000 0.000 0.341 0.513 0.687 0.861 1.037 0.170 1.214 0.687 0.687
86.9 86.6 86.7 78.6 76.7 74.4 72.5 70.3
86.9 86.6 86.7 74.9 67.8 62.1 56.2 50.3 79.8 46 62.1 62
67.9 75.3 74.5
A L/ AH
2.98 3.05 2.96 2.85 3.13 3.11
f1
F1,max
F1,min
1.000 1.000 1.000 0.564 0.460 0.388 0.334 0.292 0.723 0.259 0.388 0.388
1.000 1.000 1.000 0.991 0.989 0.986 0.983 0.980
1.000 1.000 1.000 0.986 0.977 0.969 0.962 0.954 0.993 0.948 0.969 0.969
0.977 0.987 0.986
T = 120 °C, P = 120 psig, [Hf] = 617 nmol/L, B/Hf = 2, [TOA] = 905 μmol/L, and TOA/Hf = 1470.
a
(5)
CRYSTAF calibration curve, eq 3, and their F1 and f1 values. Because the 1-hexene conversion for all runs was low, changes in its concentration (composition drift) were assumed to be negligible during the polymerizations. Figure 10 plots the 1-hexene content for the copolymer populations corresponding to the two CRYSTAF peaks versus the 1-hexene concentration in the liquid phase in the reactor.
and k11 and k21 are propagation rate constants for ethylene reacting with polymer chains ending with ethylene or 1-hexene, respectively, and k12 and k22 are the corresponding propagation rate constants for 1-hexene reacting with polymer chains ending with ethylene or 1-hexene, respectively. The variable f1 in eq 4 can be calculated using ethylene and 1-hexene molar concentrations in the reactor liquid phase, M1 and M2, respectively. f1 = 1 − f2 =
sample
(6)
Aspen Plus was used to estimate ethylene, 1-hexene, and toluene concentrations in the liquid phase at the polymerization conditions. The Peng−Robison equation of state was selected to estimate the fugacity coefficient of each component in the vapor and liquid phases. Figure 9 shows that the
Figure 10. 1-Hexene mole percent in copolymer versus 1-hexene concentration in the liquid phase (the red diamond symbols refer to the high Tc peakslower 1-hexene incorporation in the copolymer) for copolymers made with HfCl2/B/TOA.
Reactivity ratios estimates for the two site types in HfCl2/B/ TOA were obtained by fitting eq 4 to the experimental values, minimizing the sum of the squares of the residuals using Newton’s method. Table 3 summarizes these estimates and their confidence intervals. The confidence intervals for r2 estimated for both site types include zero, but this does not alter the conclusion that r2
Figure 9. Relationship between 1-hexene concentration in the liquid phase and the mass of 1-hexene added to the reactor. E
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Table 3. Reactivity Ratios for Ethylene/1-Hexene Copolymerization with HfCl2/B/TOA parameter low-Tc peaka
high-Tc peakb
value
lower confidence limit
higher confidence limit
r1
50.15
46.94
53.36
r2
0.014 94 113.8
−0.0139
0.044
96.6
131.1
0.000 55
−0.066
0.067
r1 r2
Table 4. Polymerization Conditions to Study the Effect of Borate Concentration on Ethylene/1-Hexene Copolymerization with HfCl2/B/MAOa sample
[B] (nmol/L)
B/Hf
yield (g)
activity kg polymer/ (mol catalyst·h)
B-8a B-6 B-2a B-4a B-16a B-12 B-4b B-20 B-24 B-8b B-16b B-134 B-4c B-2b
5040 3780 1260 2520 10080 7560 2520 12599 15119 5040 10080 83995 2520 1260
8 6 2 4 16 12 4 20 24 8 16 133 4 2
5.19 5.21 3.02 3.76 5.50 5.91 4.43 5.93 5.66 4.94 5.30 7.13 3.45 2.8
225600 226700 131400 163500 239200 256700 192500 257800 245900 214600 230600 310100 150000 121700
Sum of the squares of the residuals = 6.12 × 10−6, R2 = 1. bSum of squares of residuals = 6.21 × 10−6, R2 = 1.
a
values are much smaller than their corresponding r1 values for each site type. The large differences between r1 and r2 for each site type show that ethylene is much more reactive than 1hexene on both site types. Therefore, copolymers containing appreciable fractions of 1-hexene could not be made at the conditions investigated in this study, which explains the wide confidence intervals for the r2 estimates. Figure 11 shows graphical representations of the Mayo− Lewis equation for the instantaneous copolymer composition
[Hf] = 630 nmol/L, [TOA] = 930 μmol/L, TOA/Hf = 1476, T = 120 °C, P = 120 psig, Vs = 219.5 mL, tp = 10 min, and 1-hexene in feed = 0.841 (mol/L). a
(1-hexene in feed = 0.841 mol/L for all samples) made with a B/Hf molar ratio of 16, and the letter “a” indicates it is the first polymerization replicate for this copolymer. Molecular weight averages for the polymerizations in Table 4 are given in Table 5. Table 5. Molecular Weight Averages and CRYSTAF Peak Temperatures for the Copolymers Listed in Table 4 Made with HfCl2/B/TOA
Figure 11. Copolymer composition versus liquid phase composition for polymer populations made under low and high temperature CRYSTAF peaks with HfCl2/B/TOA. The solid curves are the calculated F1 values using eq 4 and the estimated reactivity ratios listed in Table 3.
as a function of the molar fraction of ethylene in the reactor for both catalyst sites. Despite the poor accuracy of the r2 estimates, these plots show that nevertheless the Mayo− Lewis equation can track copolymer composition well under our experimental conditions. Effect of Borate Concentration on CRYSTAF Profiles for HfCl2/B/TOA. Interestingly, Figure 6 shows that the AL/ AH ratio remained relatively constant for all copolymerizations (AL/AH ≅ 3.0). To elucidate the effect of borate on the magnitude of AL/AH ratio−that is, on the proportion of polymer made on each site type−14 additional copolymerizations were performed at varying borate concentrations, while keeping all other conditions constant. Table 4 summarizes polymerization conditions, yields and catalyst activities for this set of experiments. These new samples were named according to the convention: B-16a is an ethylene/1-hexene copolymer
run
B/Hf
Mn
Mw
PDI
Tc,min (°C)
Tc.max (°C)
AL/AH
B-8a B-6 B-2a B-4a B-16a B-12 B-4b B-20 B-24 B-8b B-16b B-134 B-4c B-2b
8 6 2 4 16 12 4 20 24 8 16 133 4 2
35600 40600 41500 38000 39800 39600 40000 38000 38900 37000 38100 37600 36000 36900
79600 87000 89500 86000 87700 86400 82100 82800 85300 85000 82000 82500 78000 82500
2.24 2.14 2.15 2.26 2.20 2.18 2.05 2.18 2.19 2.30 2.15 2.19 2.17 2.24
59.4 62 61.8 60.7 61.2 61.9 61.2 61.1 61.2 61.6 60.4 60.7 60.2 62.1
73.8 74.5 74.3 74.2 73.9 74 75 73.9 74 74.3 74 74.1 74.1 74.4
5.8 4.7 3.1 3.6 7.2 5.5 4.3 5.3 5.5 4.7 5.0 6.7 3.8 3
Figure 12 shows that polymer yield increases as the B/Hf ratio increases, until a plateau is reached. Figure 13 shows that molecular weight averages and polydispersity indices are not affected by borate concentration, as expected, since borate does not act as a chain transfer agent. More interestingly, Figure 14 shows how the B/Hf ratio affected the CRYSTAF profiles. All profiles were bimodal, but AL/AH depended strongly on the B/Hf ratio. Increasing B/Hf from 2 to about 10−12 increased AL/AH from 3 to about 6, but above this value the AL/AH ratio remained constant up to B/Hf = 24 (Figure 15). To confirm that this plateau was true for higher B/Hf values, another copolymerization was conducted at B/Hf = 133. Even at this extreme value, the AL/AH ratio was F
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 12. Effect of B/Hf ratio on yield of polymer made with HfCl2/ B/TOA.
Figure 15. Effect of B/Hf ratio on AL/AH of copolymers made with HfCl2/B/TOA.
The low and high CRYSTAF peak temperatures are plotted as a function of B/Hf ratio in Figure 16 to verify whether the
Figure 13. Effect of B/Hf ratio on molecular weight averages and polydispersity of polymers made with HfCl2/B/TOA. Figure 16. Effect of borate-to-catalyst ratio on CRYSTAF peak temperatures for copolymers made with HfCl2/B/TOA.
B/Hf ratio affected their positions. No correlation was observed, which indicates that excess borate does not influence 1-hexene incorporation (that is, it has no effect on r1 and r2) in this system. The ethylene uptake curves for polymerizations with varying B/Hf ratios are given in Figure 17. The curves were then linearized according to eq 2, as shown in Figure 18. Except for B/Hf = 2, all polymerizations followed first-order kinetics. Following the approach outlined above, we estimated kd and kp[M] from the slopes and y-axis intercepts of the linearized curves. Figure 19 shows how kp[M] and kd depend on the B/ Hf ratio. The values for kd decreased with increasing B/Hf until a limiting value was reached, while the apparent kp[M] increased, but eventually also reached a plateau value. Considering that a high B/Hf ratio increases the fraction of polymer made under the low CRYSTAF temperature peak, we may conclude that the active site type responsible for making copolymers with higher 1-hexene content has higher propagation frequencies and is less prone to deactivate than
Figure 14. CRYSTAF profiles for the copolymers listed in Table 5.
found to be 6.7, which is within the range of the plateau value of 6. G
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
The fact that a change in borate concentration affects the relative proportion of the two copolymer populations without altering their combined molecular weight is intriguing. To confirm the trend shown in Figure 15, another copolymerization was performed at the same TOA concentration of the experiments in Table 4, but with B/Hf = 1 and a higher Hf concentration to compensate for the decrease in activity due to the low B/Hf ratio. Figure 20 shows the CRYSTAF profile for
Figure 17. Ethylene uptake curves for the polymerizations described in Table 4.
Figure 20. CRYSTAF profile for a copolymer made at B/Hf = 1 with HfCl2/B/TOA.
this sample. The AL/AH ratio dropped to 1.02, confirming the trend that lower B/Hf ratios decrease the weight percent of copolymer made on the active site that makes copolymer with low crystallization temperature (high 1-hexene content). The number- and weight-average molecular weights for this sample were 43000 and 91000, respectively, which are within the range of the values reported in Table 5 for higher B/Hf ratios, confirming that the B/Hf ratio does not affect polymer molecular weight, even when B and Hf are present in a stoichiometric ratio. It could be argued that if the B/Hf ratio is indeed regulating the proportions of copolymer populations made in the two different active site typesas strongly supported by the CRYSTAF profiles in Figures 14 and 20then these active sites make copolymers with very similar molecular weight averages. It is only by doing so that varying their fractions by altering the B/Hf ratio would still lead to copolymers with essentially identical Mn and Mw and PDI of ≈2.0. This hypothesis was already confirmed by the CFC profiles shown in Figure 8. To further support this proposition, we also analyzed the copolymer made when B/Hf = 1 by CFC (Figure 21), confirming that the MWDs of polymer made under both CRYSTAF peaks are nearly the same. CFC analysis for sample B-24, made at the high B//Hf ratio of 24, further confirms this hypothesis (Figure 22). Taking these results together, it seems reasonable to assume that the B/Hf ratio regulates the proportion of the two types of active sites in this system: high B/Hf ratios favor the active site type with higher 1-hexene reactivity ratio (low CRYSTAF peak temperatures, high 1-hexene incorporation). Figure 19 also implies that these sites have a higher propagation rate constant and a lower deactivation rate constant. The next sections of this article will continue exploring additional features of this intriguing catalyst system. Ethylene and 1-Hexene Copolymerization with HfCl2/ MAO. The previous sets of experiments reveal the presence of
Figure 18. Plot of ln(FM,in/VR) versus time for the polymerizations shown in Figure 17. All curves can be linearized, with the exception of B/Hf = 2.
Figure 19. Effect of B/Hf ratio on kd and kp[M] estimates for polymerizations with HfCl2/B/TOA.
the active site type that makes polymer with lower 1-hexene fraction. H
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 21. CFC plot for a copolymer made at B/Hf = 1 with HfCl2/ B/TOA. Figure 23. CRYSTAF profiles of polymers in Table 6. Labels on the curves show 1-hexene concentration in mol/L.
Figure 22. CFC plot for sample B-24.
two site types for HfCl2/B/TOA, but a question remains to be answered: Is this a consequence of the B/TOA activator system or an intrinsic property of the HfCl2 catalyst? To answer this question, a set of seven copolymerization runs using MAO instead of B/TOA were conducted under similar temperature, pressure and 1-hexene concentrations of the runs reported in Table 1. These polymerization experiments are summarized in Table 6. Samples were named according to the convention: HM-8 means an ethylene/1-Hexene copolymer made with HfCl2/MAO system with 8 g of 1-hexene. Figure 23 depicts the CRYSTAF profiles of all copolymers in Table 6. Similarly to the CRYSTAF profiles in Figure 4, when the 1-hexene concentration in the feed exceeded 0.2 mol/L, all profiles became bimodal. Both CRYSTAF peak temperatures depended linearly on 1-hexene concentration (Figure 24), indicating that two active site types were also present when MAO replaced B/TOA as activator. Intriguingly, the AL/AH ratio for these copolymers is also ≈6, as observed for high B/Hf ratio. Keep in mind that the Al/Hf ratio for the polymerizations in Table 6 is also high (Al/Hf = 1400). At this point, it may be concluded that a bimodal CCD is either (1) an inherent property of the HfCl2 catalyst or (2) a
Figure 24. CRYSTAF peak crystallization temperatures versus 1hexene concentration in the reactor feed for copolymers made with HfCl2/MAO.
result of a new site type created through interaction with alkylaluminums like TOA (in the B/TOA polymerizations described above) or with TMA present in equilibrium with MAO. Following the same procedure used for estimating 1-hexene content and reactivity ratios for the B/TOA activator, the corresponding values for copolymerization of ethylene/1hexene using Hf/MAO were obtained. Table 7 summarizes CRYSTAF peak temperatures and their corresponding 1hexene contents, and Table 8 lists the reactivity ratio estimates and their confidence limits. Because of the fewer number of data points at higher 1-hexene incorporation, the confidence intervals for reactivity ratios were relatively high in this case. Figure 25 shows the graphical representation of the Mayo− Lewis equation for the instantaneous copolymer composition as a function of ethylene fraction in the reactor for both active sites when MAO was used as a cocatalyst. Despite the wide
Table 6. Copolymerization Conditions for Ethylene/1-Hexene Copolymerizations Using HfCl2/MAOa run
HM-0
HM-4
HM-8
HM-12
HM-16
HM-20
HM-24
1-hexene (g) 1-hexene (mol/L)
0.00 0.00
4 0.206
8 0.402
12 0.588
16 0.843
20 1.054
24 1.265
[Hf] = 1640 nmol/L, [MAO] = 2300 μmol Al/L, Al/Hf = 1400, T = 120 °C, P = 120 psig, Vs = 226 mL, and tp = 15 min.
a
I
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Table 7. Summary of Copolymerizations with HfCl2/MAO run
[1-hexene] (mol/L)
Tc,max (C)
Tc,min (C)
HM-0 HM-4 HM-8 HM-12 HM-16 HM-20 HM-24
0.000 0.169 0.329 0.482 0.691 0.864 1.037
86.8
86.8 80.4 73.5 67 56.1 48.2 42.5
77.5 70.6 67.3 64.3
AL/AH
1-hexene (mol %) in copolymer (Tc,min)
1-hexene (mol %) in copolymer (Tc,max) 0.000
6.1 6.65 6.56
0.000 0.658 1.567 2.423 3.858 4.899 5.649
low-Tc peaka high-Tc peakb
F1,min 1.000 0.993 0.984 0.976 0.961 0.951 0.944
0.990 0.981 0.976 0.972
higher confidence limit
parameter
value
r1
46.04
36.05
56.03
r2 r1
0.095 101.4
−0.0291 38.04
0.218 164.8
r2
0.11
−0.254 −5
F1,max 1.000
molecular weight decreases as 1-hexene concentration increases (Figure 26).
Table 8. Reactivity Ratios for Ethylene/1-Hexene Using HfCl2/MAO lower confidence limit
1.040 1.949 2.383 2.778
f1 1.000 0.725 0.572 0.477 0.386 0.332 0.293
0.476
Sum of squares of residuals = 1.27 × 10 , R = 0.994. bSum of squares of residuals = 5.73 × 10−6, R2 = 0.989.
a
2
Figure 26. Effect of 1-hexene concentration on molecular weight averages and polydispersity index of copolymers made with HfCl2/ MAO.
Figure 27 compares the CRYSTAF peak temperatures for copolymers made with HfCl2/MAO and HfCl2/B/TOA under
Figure 25. Copolymer composition versus liquid phase composition for both low and high temperature peaks for HfCl2/MAO. Solid curves are the F1 values calculate with eq 2 using the estimated reactivity ratios listed in Table 8.
confidence intervals, the point estimates for r 1 and r 2 adequately describe the copolymer composition curve within the range investigated in the study. Molecular weight averages for these copolymers are listed in Table 9. The polydispersity indices of all polymers are larger than 2.0 (in contrast with the B/TOA case, where PDI was only slightly higher than 2.0), a fact that further supports the hypothesis that two catalyst sites are present in the HfCl2/ MAO system. The PDI also increases as 1-hexene concentration increases. Similar to what we observed in Figure 7,
Figure 27. Effect of cocatalyst type on CRYSTAF peak temperatures.
Table 9. Molecular Weight Results for the Copolymers in Figure 34 run
HM-0
HM-4
HM-8
HM-12
HM-16
HM-20
HM-24
Mn Mw PDI
113000 260000 2.3
68000 169000 2.49
60000 153000 2.55
51000 137000 2.69
50000 141000 2.82
495000 143000 2.88
38300 118300 3.09
J
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
reaction with borate, making alkylation the rate-limiting step. When present in high excess, TOA also deactivates the active sites. Because its concentration is low for polymerization TH160, the rate of active site generation is greater than their deactivation rate by TOA. Thus, the interplay of these two competing reactions would result in the acceleration-type profile at the beginning of the polymerization, eventually followed by a decay-type profile for rest of the polymerization. Increasing TOA concentration favors the site deactivation step, which manifests itself as a decay-type ethylene uptake profile. The trends observed in ethylene uptakes curves for polymerizations TH-294 and TH-900 confirm this tentative explanation. Figure 29 compares the CEF profiles for the polymerizations reported in Table 10. These copolymers were analyzed by CEF
different 1-hexene concentrations. The CRYSTAF peak temperatures for copolymers made with MAO were always lower than those made using B/TOA, indicating that the sites activated with MAO were better 1-hexene incorporators. Nevertheless, both catalyst systems clearly made copolymers with bimodal CCDs and reflect the presence of two types of active sites. Thus, one may conclude that the presence of two active sites is a property of the HfCl2 catalysts and not induced by the B/TOA or MAO activator. Effect of TOA Concentration on HfCl2/B/TOA. Four different ethylene/1-hexene copolymers were made at different TOA concentrations to study the effect of TOA on copolymer CCD bimodality (Table 10). Samples were named according to the convention: TH-900 means an ethylene/1-Hexene copolymer made with a TOA/Hf ratio of 900. Table 10. Copolymerization Conditions for Ethylene/1Hexene Copolymerizations Using HfCl2/TOA/Ba run
[TOA] (μmol/L)
TOA/Hf
yield (g)
activity kg polymer/ (mol catalyst·h)
TH-88 TH-900 TH-294 TH-160
88 905 302 160
88 900 294 160
0 4.23 6.37 9.96
0 111700 131400 163500
a [Hf] = 1000 nmol/L, B/Hf = 1, T = 120 °C, P = 120 psig, tp = 10 min, total liquid feed volume (solvent + 1-hexene) = 226 mL, and 1hexene in feed = 0.841 mol/L.
Figure 28 illustrates how the ethylene uptake curves are affected by changing TOA/Hf ratio. At the lowest TOA Figure 29. Effect of TOA concentration on CEF profile of copolymers made with HfCl2/TOA/B.
instead of CRYSTAF because CEF analysis times are shorter, but both techniques provide essentially the same results and can be compared, at least qualitatively, without any difficulties. Interestingly, as the TOA concentration decreases, the weight fraction of the polymer population under the hightemperature peak decreases. It is not unreasonable to infer, therefore, that TOA reacts with the Hf catalyst in the reactor, promoting the formation of a second active site (Hf·TOA) responsible for the production of the polymer population with low 1-hexene content (high-temperature peak). Therefore, this set of experiments revealed that not only the B/Hf but also the TOA/Hf regulates CCD bimodality of the copolymers made with Hf/B/TOA. Ethylene/1-Hexene Copolymerization Using HfMe2/ B/TOA. To explore whether the methylated version of the Hf catalyst would also produce copolymers with bimodal CCD, ethylene and 1-hexene were copolymerized using the dimethyl analogue of the catalyst: HfMe2/B/TOA. Experimental conditions for the copolymerizations and polymer yields are listed in Table 11. Samples were named according to the convention: Hme-8 means an ethylene/1-Hexene copolymer made with HfMe2/B/TOA system with 8 g of 1-hexene. Figure 30 depicts the ethylene uptake curves for these four runs. All the polymerizations showed an abrupt rise in ethylene uptake curve as soon as the HfMe2 catalyst was injected, likely because the precatalyst molecules were already methylated prior to injection, and their subsequent reaction with borate to generate active sites was faster than for the non-methylated
Figure 28. Effect of TOA concentration on ethylene uptake profiles for polymerizations with HfCl2/TOA/B.
concentration, all TOA seems to be consumed to scavenge impurities in the liquid phase. Hence, not enough TOA remains to alkylate the catalyst precursor and start the polymerization. As more TOA was added to the reactor, the TOA not consumed in scavenging reactions was available to alkylate the catalyst precursor molecules, which could then react with the borate activator to produce metallocenium cations, which are the active sites for polymerization. The acceleration-type ethylene uptake profile observed for polymerization TH-160 (TOA/Cat. = 160) may be explained by proposing that at this reduced TOA concentration the alkylation rate is low compared to the subsequent rate of K
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Table 11. Copolymerization Conditions for Ethylene/1Hexene Copolymerizations Using HfMe2/B/TOAa run
1-hexene (g)
1-hexene in liquid feed (mol/L)
yield (g)
Hme-0 Hme-8 Hme-16 Hme-24
0 8 16 24
0.000 0.42 0.841 1.262
8.85 9.35 7.65 6.57
[TOA] = 930 μmol/L, [Hf] = 1550 nmol/L, [B] = 3083 nmol/L, B/ Hf = 2,TOA/Cat. =600, T = 120 °C, P = 120 psig, tp = 10 min, and total liquid feed volume (solvent + 1-hexene) = 226 mL. a
Figure 32. CEF profiles of the copolymers reported in Table 11. Numbers on the curve show 1-hexene concentration in the reactor feed. The high temperature should for the homopolymer is an artifact of the CEF analysis.
interpreted as the presence of an additional polymer population made on a separate active site type. Effect of TOA Concentration on HfMe2/B/TOA. Similarly to the study performed to investigate the effect of TOA on CCD bimodality of copolymers made with HfCl2/B/ TOA, four copolymerizations using HfMe2/B/TOA were also conducted. Table 12 summarizes the conditions, polymer yield, and catalyst activity data for these polymerizations.
Figure 30. Ehylene uptake curves for the polymerizations reported in Table 11.
Table 12. Experimental Conditions for Ethylene/1-Hexene Copolymerization Using HfMe2/B/TOAa
analogue. Adding 1-hexene to the polymerization lowered the rate of catalyst deactivation, as apparent by the shallower slopes of the ln(FM,in/VR) versus time plots shown in Figure 31.
run
[TOA] (μmol/L)
TOA/Hf
yield (g)
activity kg polymer/ (mol catalyst·h)
THme-170 THme-300 THme-580 THme-865
170 290 580 865
170 300 580 865
31.9 9.08 5.93 2.58
826000 240000 157000 75000
a [Hf] = 1000 nmol/L, B/Hf = 2, T = 120 °C, P = 120 psig, tp = 10 min, total liquid feed volume (solvent+1-hexene) = 226 mL, and 1hexene in feed = 0.841 mol/L.
Figure 33 compares the ethylene uptake curves for these polymerizations. As the TOA concentration decreases, the ethylene flow rate rises and, consequently, the polymer yield increases. Previous studies from our group have shown that excess TOA deactivates catalyst sites, and this trend is also observed for these polymerization experiments.33,34 Figure 34 compares the CEF profiles of these four polymers. When the TOA concentration decreased below a TOA/Hf ratio of about 300, the CEF profiles became unimodal. Therefore, it may be concluded that TOA participates in the reaction that creates the site type (lets called it Hf·TOA) that makes the polymer population under the high-temperature CEF peak (low 1-hexene incorporation). Because the Hf·TOA site is less active than the original Hf site (see Figure 19), the ethylene uptake rate decreases when more TOA is added to the reactor, as seen in Figure 33. Elimination of TOA: Using the Catalyst as a Scavenger. All previous results support the hypothesis that TOA is related to the production of copolymers with bimodal CCD, either when HfCl2 or its pre-methylated version, HfMe2,
Figure 31. Effect of 1-hexene concentration on ln(FM,in/VR) profiles for the polymerizations reported in Table 11.
Figure 32 compares the CEF profiles of these copolymers. Similarly to the CRYSTAF profiles of copolymers made with HfCl2/B/TOA, the CEF profiles also became bimodal after a certain 1-hexene concentration when HfMe2/B/TOA was used to make the copolymers. The bimodal profile for the homopolymer is an artifact of the CEF analysis. This phenomenon is commonly observed when ethylene homopolymers are analyzed by CEF and TREF and should not be L
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
temperature for 35 min. Toluene was then saturated with ethylene to 120 psi, before the borate/catalyst mixture was injected into the reactor. In another copolymerization run (THme-900, Table 13) TOA replaced the catalyst as the impurity scavenger, but all other conditions remained the same. Figure 35 compares ethylene uptake curves for these polymerizations. Interestingly, for the case where TOA was
Figure 33. Effect of TOA on ethylene uptake curves with HfMe2/B/ TOA.
Figure 35. Ethylene uptake curves for the runs THme-0 and THme900 with HfMe2/B.
not used, the ethylene consumption rate was much higher, which may result from a combination of two factors. First, excess TOA decreases catalyst activity, and TOA was not added to run THme-0. Second, it is conceivable that some of the catalyst injected as scavenger did not react with the impurities present in toluene and became activated when the HfMe2/B mixture was added later to the reactor, which would also increase the overall polymerization rate. As suspected, the CEF profile is unimodal (see Figure 36) for run THme-0, when no TOA was used, but it was bimodal for run THme-900, where TOA was added as impurity scavenger. This seems to be the definite proof that TOA promotes the generation of the second type of active site that produces copolymers with low 1-hexene content.
Figure 34. CEF profiles of the samples reported in Table 12.
is used to make the copolymers. The next set of experiments was designed to determine whether the HfMe2/B catalyst could produce a bimodal polymer without adding any TOA to the reactor. The main experimental difficulty in this case arises from trace amounts of impurities present in toluene (used as the polymerization media) which are scavenged with TOA. In the absence of TOA, these impurities may completely poison the catalyst, unless the catalyst itself is added in enough excess, so that part of it replaces TOA as an impurity scavenger. In run THme-0 (Table 13), 200 mL of toluene and the desired amount of 1-hexene were added to the reactor, followed by 70 μmol of HfMe2 as “impurity scavenger”. The reactor temperature was increased to 120 °C and kept at that Table 13. Summary of Conditions for Polymerizations to Study Elimination of TOA in the HfMe2/B/TOA Systema run
[TOA] (μmol/L)
TOA/ HfMe2
yield (g)
activity kg polymer/ (mol catalyst·h)
THme-0 THme-900
0.0 867
0.0 900
6.81 2.85
240000 75000
a [Hf] = 963 nmol/L, B/Hf = 2, T = 120 °C, P = 120 psig, tp = 10 min, total liquid feed volume (solvent+1-hexene) = 236 mL, and 1hexene in feed = 0.841 mol/L.
Figure 36. CEF profiles for copolymers made in runs THme-0 and THme-900 with HfMe2/B. M
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules Effect of Polymerization Time for Copolymerizations with HfCl2/TOA/B. Thus far, it can be concluded that TOA was responsible for the formation of the second active site type (Hf·TOA) with either HfCl2/B/TOA or HfMe2/B/TOA systems. Another set of polymerization experiments were designed to test the kinetics of the formation of the conjectural Hf·TOA site, which could be reflected by a change of AL/AH ratio of CEF and/or CRYSTAF profiles. This effect can only be elucidated by investigating how the AL/AH ratio varies as a function of polymerization time. To study the effect of polymerization time on the microstructure of the copolymers made with HfCl2/TOA/B, four polymerizations with different durations were performed. In one of the polymerizations, samples were collected from the reactor following polymerization times of 1, 3, 5, 7, and 10 min. In the other three experiments, the polymerization times were 10, 14, and 30 s. The last three runs were conducted separately because the mass of polymer formed at these short polymerization times was very small. The polymerization conditions of all these runs were the same except polymerization time (Table 14).
Figure 38. CEF profiles for the runs reported in Table 14. Figure also shows the ratio of AL/AH versus time (see the inset shown on the right side).
temperature peak decreases, eventually reaching a steadystate value at about 2 min of polymerization (see the inset on the right-hand side of Figure 38). To calculate the area for each peak, the minimum point between the peaks was taken as the separation point. It seems, therefore, that the reaction between TOA and Hf is fast, leading to the formation of the Hf·TOA site (high elution temperature/low 1-hexene content) at the very beginning of the polymerization, but the proportion of polymer made by this site decreases and reaches a steady-state value soon after the polymerization begins, either because part of it is transformed back into regular Hf sites (low elution temperature, high 1-hexene incorporation) or because Hf sites are activated more slowly, causing the fraction of polymer made on these sites to increase with polymerization time. Figure 39 shows that how the molecular weight averages and PDI of the samples vary with time. Their numerical values are reported in Table 15. At the beginning of the polymerization, the molecular weight is slightly higher because the weight fraction of the polymer population made on the Hf·TOA site is higher than that made on the Hf site. We may speculate that the Hf·TOA site makes polymer with higher molecular weights
Table 14. Effect of Polymerization Time with HfMe2/B/ TOAa run
polymerization time (s)
yield (g)
t-600 t-30 t-14 t-10
600 30 14 10
6.81 0.03 0.05 0.04
[Hf] = 963 nmol/L, [TOA] = 867 μmol/L, B/Hf = 1, T = 120 °C, P = 120 psig, TOA/Hf = 900, total liquid feed volume (solvent + 1hexene) = 236 mL, and 1-hexene in feed = 0.841 mol/L. a
Figure 37 shows the ethylene uptake curve for run t-600. Samples were taken at 1, 3, 5, and 7 min during the
Figure 37. Ethylene uptake curve for run t-600.
polymerization. This is reflected by the sudden increase in ethylene flow rate due to the drop in the volume of the polymerization medium. After 10 min, the whole reactor content was forced out into a beaker containing ethanol. Samples were filtered and then dried in an oven for further analysis. Figure 38 compares the CEF profiles of all copolymers. As polymerization time increases, the area under the high
Figure 39. Variation of molecular weight averages and PDI with polymerization time for HfMe2/B/TOA. N
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 40 is a graphical representation of eq 9 for arbitrary values of k1/k2 and [TOA]/[B]. It is apparent that this
Table 15. Effect of Polymerization Time on Molecular Weight Averages of Polymer Made with HfMe2/B/TOA run t-600 t-600 t-600 t-600 t-600 t-10 t-14 t-30
(1 min sample) (3 min sample) (5 min sample) (7 min sample) (10 min sample)
time (s)
Mn
Mw
PDI
60 180 300 420 600 10 14 30
NA 42100 43800 42100 41600 52700 49400 NA
NA 82200 91500 86700 87800 108400 101400 NA
NA 2.1 2.09 2.06 2.11 2.06 2.05 NA
due to its lower 1-hexene affinity, a common feature of other multiple sites catalysts for olefin polymerization. As the polymerization time increases, the polymer molecular weight averages become constant, at about the same time the ratio AH/AL reaches a stable plateau, and the proportions of polymer populations made by both site types reach a steady-state value.
Figure 40. Graphical representation of eq 9.
■
simplistic model can describe how the proportions of the two proposed site types are affected by the relative concentrations of TOA and B vary in the reactor. From the copolymerization kinetic studies, it may also be concluded that the Hf site is more stable and has a higher propagation rate constant than the Hf·TOA site. Even though a much wider set of experiments would be required to test this mechanism, it is clear that both catalyst forms, HfCl2 and HfMe2, make copolymers having two populations with distinct 1-hexene fractions and that the relative amounts of these population can be “tuned-in” by varying the [TOA]/[B] ratio during the polymerization.
CONCLUSIONS The copolymerization of ethylene and 1-hexene in a solution reactor using bis(n-propylcyclopentadienyl)hafnium dichloride (HfCl2), its methylated version (HfMe2), activated with TOA/ B or MAO was investigated. Every catalyst/cocatalyst combination tested in this study can produce copolymers with bimodal CRYSTAF and/or CEF profiles, indicating that two different site types were present, each with a different reactivity ratio toward 1-hexene incorporation. Raising the B/Hf ratio increased the fraction of copolymer with high 1-hexene content (low CRYSTAF crystallization temperature, Tc, or low CEF elution temperature, Te), up to a limiting value (B/Hf ≅ 10−12), when the ratio of the area under the low-Tc peak (AL) and high-Tc peak (AH) reached a plateau value of AL/AH ≅ 6. Interestingly, the AL/AH ratio was also regulated by the TOA/Hf ratio used during polymerization. In fact, using HfMe2/B in the absence of TOA, the CEF profiles of a bimodal ethylene/1-hexene copolymer (made in the presence of TOA) became unimodal, suggesting that TOA is required to form the second site type. These two observations suggest that TOA participates in the creation of a second type of active site but that the borate cocatalyst also plays a role shifting the equilibrium between the two site types. It is likely that the same effect may be observed with other alkylaluminum compounds. A rudimentary kinetic model that would explain such behavior is suggested in the equations below: (7) Hf + TOA → Hf·TOA Hf·TOA + B → Hf + TOA·B
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected] (J.B.P.S.). ORCID
João B. P. Soares: 0000-0001-8017-143X Notes
The authors declare no competing financial interest.
■
REFERENCES
(1) Breslow, D. S.; Newburg, N. R. Bis-(Cyclopentadienyl)Titanium Dichloride - Alkylaluminum Complexes as Catalysts for the Copolymerization of Ethylene. J. Am. Chem. Soc. 1957, 79, 5072− 5073. (2) Andresen, A.; Kaminsky, W.; Pein, J.; et al. Halogen-Free Soluble Ziegler Catalysts for the Polymerization of Ethylene. Control of Molecular Weight by Choice of Temperature. Angew. Chem., Int. Ed. Engl. 1976, 15, 630−632. (3) Kaminsky, W. The Discovery of Metallocene Catalysts and their Present State of the Art. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 3911−3921. (4) Kaminsky, W. Discovery of Methylaluminoxane as Cocatalyst for Olefin Polymerization. Macromolecules 2012, 45, 3289−3297. (5) Busico, V. Metal-Catalysed Olefin Polymerization into the New Millenium: A Perpective Outlook. Dalton Transactions 2009, 41, 8794−8802. (6) Sinn, H. Proposals for Structure and Effects of Methylalumoxane based on Mass Balances and Phase Separation Experiments. Macromol. Symp. 1995, 97, 27−52. (7) Koide, Y.; Bott, S. G.; Barron, A. R. Alumoxanes as Cocatalysts in the Palladium-Catalyzed Copolymerization of Carbon Monoxide and Ethylene: Genesis of a Structure-Activity Relationship. Organometallics 1996, 15, 2213−2226.
(8)
where the Hf site is a better 1-hexene incorporator, making copolymers with lower Tc, and the Hf·TOA site is a worse 1hexene incorporator, making copolymers with higher Tc. Assuming that the rate constants for the reactions described in eqs 7 and 8 are k1 and k2, respectively, and assuming that the equilibrium concentrations between the two site types is reached quickly (as indicated in Figure 38), it is easy to show that the molar fraction of Hf sites in the reactor is given by 1 x Hf = k1 [TOA] 1 + k [B] (9) 2
O
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules (8) Resconi, L.; Bossi, S.; Abis, L. Study on the Role of Methylaluminoxane in Homogeneous Olefin Polymerization. Macromolecules 1990, 23, 4489−4491. (9) Bryliakov, K. P.; Semikolenova, N. V.; Yudaev, D. V.; Ystenes, M.; Rytter, E.; Zakharov, V. A.; Talsi, E. P. 1H and 13C NMR Study of the Intermediates Formed by (Cp-R)2ZrCl2 Activation with MAO and AlMe3/[CPh3][B(C6F5)4]. Correlation of Spectroscopic and Ethene Polymerization Data. Macromol. Chem. Phys. 2003, 204, 1110−1117. (10) Bryliakov, K. P.; Semikolenova, N. V.; Yudaev, D. V.; Zakharov, V. A.; Brintzinger, H. H.; Ystenes, M.; Rytter, E.; Talsi, E. P. 1H, 13C NMR and Ethylene Polymerization Studies of Zirconocene/MAO Catalysts: Effect of the Ligand Structure on the Formation of Active Intermediates and Polymerization Kinetics. J. Organomet. Chem. 2003, 683, 92−102. (11) Bryliakov, K. P.; Semikolenova, N. V.; Zakharov, V. A.; Talsi, E. P. 13C NMR Study of Ti(IV) Species Formed by Cp*TiMe3 and Cp*TiCl3 Activation with Methylauminoxane (MAO). J. Organomet. Chem. 2003, 683, 23−28. (12) Brintzinger, H. H.; Fischer, D.; Mülhaupt, R.; Rieger, B.; Waymouth, R. M. Stereospecific Olefin Polymerization with Chiral Metallocene Catalysts. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143− 1170. (13) Jordan, R. F. Chemistry of Cationic Dicyclopentadienyl Group 4 Metal-Alkyl I Complexes. Adv. Organomet. Chem. 1991, 32, 325− 387. (14) Bochmann, M. Cationic Group 4 Metallocene Complexes and their Role in Polymerisation Catalysis: The Chemistry of Well Defined Ziegler Catalysts. J. Chem. Soc., Dalton Trans. 1996, 3, 255− 270. (15) Diesner, T.; Troll, C.; Rieger, B. Hafnocene-Based Olefin Polymerizations. In Metal Catalysts in Olefin Polymerization, Guan, Z., Ed.; Springer-Verlag: Berlin, 2009; Vol. 26, pp 47−63. (16) Jia, L.; Yang, X.; Stern, C. L.; Marks, T. J. Cationic Metallocene Polymerization Catalysts based on Tetrakis(pentafluorophenyl)borate and its Derivatives. Probing the Limits of Anion “Noncoordination” via a Synthetic, Solution Dynamic, Structural, and Catalytic Olefin Polymerization Study. Organometallics 1997, 16, 842−857. (17) Chien, J. C. W.; Tsai, W. M.; Rausch, M. D. Isospecific Polymerization of Propylene Catalyzed by rac-Ethylenebis(indenyl)methylzirconium Cation. J. Am. Chem. Soc. 1991, 113, 8570−8571. (18) McInnis, J. P.; Delferro, M.; Marks, T. J. Multinuclear Group 4 Catalysis: Olefin Polymerization Pathways Modified by Strong MetalMetal Cooperative Effects. Acc. Chem. Res. 2014, 47, 2545−2557. (19) Bochmann, M. The Chemistry of Catalyst Activation: The Case of Group 4 Polymerization Catalysts. Organometallics 2010, 29, 4711−4740. (20) Chen, M. C.; Roberts, J. A. S.; Marks, T. J. Marked Counterion Effects on Single-Site Olefin Polymerization Processes. Correlations of Ion Par Structure and Dynamics with Polymerization Activity, Chain Transfer, and Syndioselectivity. J. Am. Chem. Soc. 2004, 126, 4605−4625. (21) Sarsfield, M. J.; Ewart, S. W.; Tremblay, T. L.; Roszak, A. W.; Baird, M. C. Synthesis of a Series of New, Highly Electrophilic, Monocyclopentadienyltitanium Olefin Polymerization Initiators. J. Chem. Soc., Dalton Trans. 1997, 18, 3097−3104. (22) Yang, X. M.; Stern, C. L.; Marks, T. J. Cation-like Homogeneous Olefin Polymerization Catalysts based upon Zirconocene Alkyls and Tris(pentafluorophenyl) Borate. J. Am. Chem. Soc. 1991, 113, 3623−3625. (23) Yang, X. M.; Stern, C. L.; Marks, T. J. Cationic Zirconocene Olefin Polymerization Catalysts based on the Organo-Lewis Acid Tris(pentafluorophenyl)borane. A Synthetic, Structural, Solution Dynamic, and Polymerization catalytic study. J. Am. Chem. Soc. 1994, 116, 10015−10031. (24) Yang, X. M.; Stern, C. L.; Marks, T. J. Models for Organometallic Molecule-support Complexes. Very Large Counterion Modulation of Cationic Actinide Alkyl Reactivity. Organometallics 1991, 10, 840−842.
(25) Siedle, A. R.; Newmark, R. A. Stereochemical Nonrigidity in Metallocenium Ions. J. Organomet. Chem. 1995, 497, 119−125. (26) Chen, Y. X.; Stern, C. L.; Yang, S.; Marks, T. J. Organo-Lewis Acids as Cocatlysts in Cationic Metallocene Polymerization Catalysis. Unusal Characteristics of Sterically Encumbered Tris(perfluorobiphenyl)borane. J. Am. Chem. Soc. 1996, 118, 12451− 12452. (27) Beck, S.; Prosenc, M. H.; Brintzinger, H. H.; Goretzki, R.; Herfert, N.; Fink, G. Binuclear Zirconocene Cations with μ-CH3Bridges in Homogenous Ziegler-Natta Catalyst Systems. J. Mol. Catal. A: Chem. 1996, 111, 67−79. (28) Bochmann, M.; Lancaster, S. J. Monomer-Dimer Equilibria in Homo- and Heterodinuclear Cationic Alkylzirconium Complexes and their Role in Polymerization Catalysis. Angew. Chem., Int. Ed. Engl. 1994, 33, 1634−1637. (29) Song, F. Q.; Cannon, R. D.; Bochmann, M. The Kinetics of Propene and Hexene Polymerisation with [(SBI)ZrR]+X−: Evidenc for Monomer-Dependent Early or Late Transition States. Chem. Commun. 2004, 5, 542−543. (30) Moscato, B. M.; Zhu, B. L.; Landis, C. R. Mechanistic Investigations into the Behavior of a Labeled Zirconocene Polymerization Catalyst. Organometallics 2012, 31, 2097−2107. (31) Theurkauff, G.; Bader, M.; Marquet, N.; Bondon, A.; Roisnel, T.; Guegan, J. P.; Amar, A.; Boucekkine, A.; Carpentier, J. F.; Kirillov, E. Discrete Ionic Complexes of Highly Isoselective Zirconocenes. Solution Dynamics, Trimethylaluminum Adducts, and Implications in Propylene Polymerization. Organometallics 2016, 35, 258−276. (32) Bryliakov, K. P.; Babushkin, D. E.; Talsi, E. P.; Voskoboynikov, A. Z.; Gritzo, H.; Schröder, L.; Damrau, H. R. H.; Wieser, U.; Schaper, F.; Brintzinger, H. H. ansa-Titanocene Catalysts for α-Olefin Polymerization. Synthesis, Structures, and Reactions with Methylaluminoxane and Boron-Based Activators. Organometallics 2005, 24, 894−904. (33) Mehdiabadi, S.; Soares, J. B. P.; Bilbao, D.; Brinen, J. A polymerization Kinetics Comparision between a Metallocene Catalyst Activated by Tetrakis(pentafluorophenyl) Borate and MAO for the Polymerization of Ethylene in a Semi-Batch Solution Reactor. Macromol. React. Eng. 2011, 5, 418−430. (34) Mehdiabadi, S.; Soares, J. B. P.; Bilbao, D.; Brinen, J. Ethylene Polymerization and Ethylene/1-Octene Copolymerization with racDimethylsilylbis(indenyl)hafnium Dimethyl using Trioctyl Aluminum and Borate: A Polymerization Kinetics Investigation. Macromolecules 2013, 46, 1312−1324. (35) Mehdiabadi, S.; Soares, J. B. P.; Brinen, J. Ethylene Polymerization with a Hafnocene Dichloride Catalyst using Trioctyl Aluminum and Borate: Polymerization Kinetics and Polymer Characterization. Macromol. React. Eng. 2017, 11, 1600044. (36) Mehdiabadi, S.; Soares, J. B. P. In-Depth Investigation of Ethylene Solution Polymerization Kinetics with rac-Et(Ind)2ZrCl2/ MAO. Macromol. Chem. Phys. 2013, 214, 246−262.
P
DOI: 10.1021/acs.macromol.8b01028 Macromolecules XXXX, XXX, XXX−XXX
