Elsevier

Fluid Phase Equilibria

Volume 428, 25 November 2016, Pages 38-47
Fluid Phase Equilibria

Cloud point pressure in the system polyethylene + ethylene – Impact of branching

This paper is dedicated to Prof. T.W. de Loos in honor to his scientific impact and to his activities encouraging young scientists.
https://doi.org/10.1016/j.fluid.2016.06.032Get rights and content

Highlights

  • A simple theory for arbitrarily branched polydisperse polymer solutions is proposed.

  • Pure compound parameters are estimated for ethylene and polyethylene.

  • Only critical properties of ethylene and simple arguments for polyethylene are used.

  • A binary, temperature-dependent parameter is fit to cloud points of one polyethylene fraction in solution.

  • Cloud point data for nine other fractions are predicted and compared to experiment.

Abstract

We propose here a simple theory for polydisperse, branched polyethylene dissolved in ethylene based on a cubic equation, namely the Sako-Wu-Prausnitz-equation of state, in combination with the Lattice Cluster Theory in the incompressible version, which is predictive concerning the branching, while using continuous thermodynamics to account for polydispersity with respect to the molecular mass. The new theory considers polymers to follow the Schulz-Flory distribution and to adopt an average degree of branching. It does not require different parameters for different molar masses, polydispersities or degrees of branching and hence, the number and mass average molar mass as well as average branch content are sufficient. We parameterize this theory for ethylene based on critical data and for polyethylene based on simple arguments and some experimental data of n-alkanes. The binary parameter describing the dispersion energy and its temperature dependence stemming from Lattice Cluster Theory is fit to experimental cloud point data from literature of one specific sample of polyethylene in ethylene. The theory is then used to predict the cloud point curves and critical properties of nine other polyethylene samples in ethylene. The fit and predicted cloud point curves show quantitatively correct temperature dependence. For high polydispersities, the concentration dependence is only qualitatively correct, which is most probably due to the oversimplified Schulz-Flory distribution. The influence of branching on cloud point pressures and critical pressures at constant temperature is predicted quantitatively correct.

Introduction

Polyethylene (PE) is one of the most widespread plastic materials, produced commercially on a large scale in a continuous way at relatively high temperatures and pressures by free-radical bulk polymerization in supercritical ethylene [1]. Folie and Radosz [1] explained in detail the significance of the involved phase behavior for the industrial production. These authors [1] emphasize the complexity of the phase behavior of this system, which is caused by a) the materials behaving highly nonideal at high pressure, b) the polymer and the solvent greatly differing in size, and c) commercial polymers being composed of many molecules differing in molar mass and chemical composition. In the case of polyethylene, the last one can be characterized as branching. The phase behavior of this system drastically changes with the polymers molar mass distribution and with branching. Both of these properties are usually not well known, especially in combination, i.e. the distribution of polymers with respect to both molar mass and branching. If anything, number and mass average molar mass are known, and sometimes the average branch content of the whole polymer sample. The latter number can make up a difference of up to 40 MPa in the high pressure cloud point curve in the mixture of ethylene + polyethylene [1]. In addition to the afore mentioned difficulties concerning the molar mass and chemical distribution function, often no pure compound properties are known and therefore the equation of state necessary for the description of high pressure phase equilibria can only be parameterized on mixture data.

Caused by the high industrial importance of this mixture, a large amount of experimental e.g. [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14] as well as theoretical work e.g. [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30] has been performed on these systems in the past. For modelling, different approaches, including different equations of state (Perturbed Hard Chain Theory [8]; Sanchez-Lacombe-EOS or their modifications [10], [11], [12], [13], [14], [16], [17], [19], [26]; Statistical Association Fluid Theory [7], [16], [17], [19], [22]; Statistical Association Fluid Theory – Variable Range [19]; Perturbed Chain- Statistical Association Fluid Theory [25], [26]; Group Contribution - Statistical Association Fluid Theory –Variable Range [30]; Sako-Wu-Prausnitz-EOS [26]) or models for the excess Gibbs energy of mixing (GE-model) [6], [15], [26] were applied. In the theoretical framework, the polydispersity was taken into account mostly by the pseudo-component method [9], [11], [15], [16], [17], [22], [26]. Usually, the pure-component parameters of the branched and the linear polymer were fit to experimental data [19]. However, this approach requires a large experimental effort. Only, few papers consider the branching directly in the model using different ideas. Kleintjens et al. [31] suggest the use of topological arguments in their GE-model. Dominik and Chapman [25] developed a thermodynamic model for branched polyolefins within the PC-SAFT approach by hetero-segmentation. Marshall and Chapman [32] suggested a new EOS for branched molecules based on Wertheim’s perturbation theory. Yang et al. [33] developed a theoretical framework, incorporating the degree of branching, by fitting the long-range correlation to simulation data. Using molecular simulation [34], [35], [36] or renormalization group theory [37] also allows studying the impact of branching on the phase behavior.

We focus our attention here to the impact of branching on the cloud-point pressure of the system PE and ethylene. The importance of branching can be recognized by the fact that linear and branched polymers of the same type and approximately the same molecular weight are not even completely miscible [38], [39], [40], [41], [42], [43], [44], [45]. Next to the phase behavior e.g. [46], [47], the branching also has a large impact on other physical properties, for instance crystallization [48], osmotic pressure [49], rheological properties [48] and theta temperature [50]. From experiments it is known that the cloud point pressure decreases with increasing degree of branching [2], [5]. The goal of this contribution is the calculation of this equilibrium using a minimum number of adjustable parameters. To this end, the pure compound parameters are fit only to pure compound properties and derived from arguments where this is not possible. Furthermore, the binary parameters are fit only to the phase equilibrium of one specific polymer sample in solution and transferred to all other polymer samples discussed.

The Lattice Cluster Theory (LCT), developed by Freed and Coworkers [51], [52], [53], can reach this goal in principle. This theory allows the calculation of the Gibbs energy for molecules having an arbitrary architecture. In the last years, the LCT was successfully applied for calculation of phase equilibria of systems containing branched [54], [55], [56] or hyperbranched polymers [57], [58], [59], [60], [61], [62], [63], [64], [65], [66], [67]. Arya and Panagiotopoulos [34] compare results obtained by grand canonical Monte Carlo simulations with results obtained by a simplified version of LCT. It was observed that the LCT significantly underestimates the impact of polymer branching on the critical behavior of polymers [34]. However, the authors use the coordination number of 26 for a cubic lattice, which should be 6 in the notation of lattice cluster theory for that lattice [34]. Since in LCT, architecture is developed in the inverse coordination number, the underestimation using this too high coordination number is not surprising. By introducing empty lattice sites, an equation of state based on LCT can be derived [68], [69], [70]. However, the theoretical foundation of LCT is based on a lattice theory, which reaches its limitation at very high pressures. Therefore, we suggest in this contribution an alternative approach, which is a combination of a cubic EOS, namely the Sako-Wu-Prausnitz- EOS (SWP-EOS) [71] and the LCT, where the lattice is taken as incompressible. This approach is similar to the idea of Vidal [72], who combined a cubic EOS with a model for the Gibbs energy of mixing. The phase behavior of n-alkane mixtures [73] and polystyrene in different solvents [74], [75] could be modelled using SWP-EOS close to experimental data. Additionally, the equations for the critical point were already derived by Browarzik and Kowalewski [76].

Section snippets

Theory

The SWP-EOS, originally suggested by Sako et al. [71], was used in the theoretical framework of continuous thermodynamics [77], [78], where the polydispersity can be taken into account by a continuous distribution function, by Browarzik et al. [73], [74], [75]. The stability of a semi-continuous mixture, made from a discrete component (solvent, A) and a continuous ensemble (polymer, B), was treated by continuous thermodynamics based on an EOS [73]. Its parameters were assumed to depend on the

Parameter estimation procedure

Since ethylene is rigid, it is considered as the base unit here, hence rA = 1, which also implies cs,A = 1. It can be shown that the segment-molar critical volume Vs,cr,A of ethylene divided by the co-volume bs per segment has to fulfill the equationV˜s,cr,A33V˜s,cr,A23V˜s,cr,A1=0where V˜s,cr,A=Vs,cr,A/bs,A. The only real root of this equation is approximately:Vs,cr=3.84732bs

We choose bs,A = 34.05 cm3/mol such that the calculated critical volume is equal to the experimental critical volume

Model predictions

Since all model parameters are determined, the remaining systems’ cloud point curves can be predicted. This is a quite stringent test of the theory, because molar mass, non-uniformity and branching degree of the respective polymers vary for these systems. For the polymer with the highest non-uniformity (BPE1) the p-T-behavior of a 3 Ma.-% solution in ethylene is depicted in Fig. 3. The agreement is quantitatively correct for almost all temperatures. Only at low temperatures there are some minor

Conclusion

In this contribution we use a combination of a simple cubic equation of state, namely the Sako-Wu-Prausnitz-EOS [71], with the Lattice Cluster Theory of Freed and co-workers [51], [52], [53], where polydispersity is explicitly accounted for by continuous thermodynamics [77], [78], to model the phase behavior of 10 different polyethylene samples in solutions of ethylene. The EOS has three pure compound parameters that are fit to the critical point for ethylene and constructed for polyethylene

Acknowledgment

The authors are grateful for the financial support by the Deutsche Forschungsgemeinschaft (German Science Foundation) in Germany (EN 291/7-2) to accomplish this work.

References (84)

  • T.W. de Loos et al.

    J. Supercrit. Fluids

    (1995)
  • T.W. De Loos et al.

    Fluid Phase Equilibria

    (1996)
  • S.J. Moore et al.

    Chem. Eng. Sci.

    (2001)
  • H. Trumpi et al.

    J. Supercrit. Fluids

    (2003)
  • X. Chen et al.

    Fluid Phase Equilibria

    (2004)
  • I. Nagy et al.

    J. Supercrit. Fluids

    (2006)
  • M. Haruki et al.

    J. Supercrit. Fluids

    (2009)
  • M. Haruki et al.

    Fluid Phase Equilibria

    (2011)
  • E. Kiran et al.

    J. Supercrit. Fluids

    (1993)
  • N. Koak et al.

    Fluid Phase Equilibria

    (1999)
  • K. Gauter et al.

    Fluid Phase Equilibria

    (2001)
  • R.A. Krenz et al.

    Fluid Phase Equilibria

    (2005)
  • J. Yang et al.

    Fluid Phase Equilibria

    (2006)
  • D.N. Theodorou

    Chem. Eng. Sci.

    (2007)
  • M.J. Hill et al.

    Polymer

    (1991)
  • S.I. Mumby et al.

    Polymer

    (1995)
  • C. Schipp et al.

    Polymer

    (1996)
  • F. Samadi et al.

    Eur. Polym. J.

    (2007)
  • D. Papanagopoulos et al.

    Eur. Polym. J.

    (2004)
  • D. Browarzik et al.

    J. Chem. Thermodyn.

    (2013)
  • T. Zeiner et al.

    Fluid Phase Equilibria

    (2009)
  • T. Zeiner et al.

    Fluid Phase Equilibria

    (2011)
  • T. Zeiner et al.

    J. Chem. Thermodynamics

    (2011)
  • T. Zeiner et al.

    Chem. Eng. Sci.

    (2011)
  • C. Browarzik et al.

    Fluid Phase Equilibria

    (2012)
  • A. Kulaguin-Chicaroux et al.

    Fluid Phase Equilibria

    (2014)
  • S. Enders et al.

    Thermodynamics

    (2014)
  • T. Goetsch et al.

    Chem. Eng. Proc. Process Intensif.

    (2016)
  • K. Langenbach et al.

    Fluid Phase Equilibria

    (2012)
  • K. Langenbach et al.

    J. Chem. Thermodyn.

    (2013)
  • K. Langenbach et al.

    Fluid Phase Equilibria

    (2014)
  • J. Vidal

    Chem. Eng. Sci.

    (1978)
  • D. Browarzik et al.

    Fluid Phase Equilibria

    (1998)
  • D. Browarzik et al.

    Fluid Phase Equilibria

    (2002)
  • D. Browarzik et al.

    Fluid Phase Equilibria

    (1999)
  • M.T. Rätzsch et al.

    Prog. Polym. Sci.

    (1989)
  • S. Enders et al.

    Fluid Phase Equilibria

    (1997)
  • B. Folie et al.

    Ind. Eng. Chem. Res.

    (1995)
  • P.P. Maloney et al.

    AIChE J.

    (1976)
  • D.P. Maloney et al.

    Ind. Eng. Chem. Process Des. Dev.

    (1976)
  • R. Spahl et al.

    Ber. Bunsenges. Phys. Chem.

    (1982)
  • T.W. de Loos et al.

    Macromolecules

    (1983)
  • View full text