Don’t gamble with physical properties of polymers
Advancing the chemical engineering fundamentals
Thermodynamics: General (T2-1e)
Keywords: LDPE, property models, polymers, EOS
Authors: Anton A. Kiss, Abdelghani Belkou, Alexandre C. Dimian, Piet Iedema
Affiliation: University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, NL
The physical properties models are the base of any process modeling and simulation hence the appropriate selection of the model influences all future steps of the simulation. In any polymerization system there is a separation of the polymer from monomer, which is recycled in case of incomplete conversion. This is not a trivial task because mixtures of small molecules and long-chain polymers are involved. Thus it is essential to choose the suitable properties model to get reliable results from simulation. The problem is how to select the correct properties model for which reliable parameters can be calculated, when insufficient experimental data is available.
To solve this problem we explore the most used properties models (e.g. Sanchez-Lacombe, SAFT, Flory¬-Huggins, SRK) available in polymerization systems, perform regression and parameter fitting, and show that different results are predicted with respect to monomer-polymer separation and polymer properties. The properties models used in this work describe with good accuracy the behavior of both conventional molecules and polymer chains.
As both conventional molecules and polymers are present in the process, the key physical properties of all types of molecules must be predicted with acceptable accuracy by the models used. A major advantage of using equation-of-state (EOS) is that pure-component parameters of many conventional molecules encountered in polymer processes are already available in literature. Nevertheless there is a need to estimate somehow the binary interaction parameters for these models. If the binary parameters are not fitted to experimental data the models tested reveal unsatisfactory performance.
In this work the high-pressure low-density poly-ethylene (HP-LDPE) system was selected as case study.[1] Experimental data was used to calculate the model parameters.[2] The parameters should be segment-based or scalable for the size of the polymer. This condition is necessary because the size of the polymer changes in the process, and the model used must be able to accommodate this fact. [3-4]
Different results are predicted by the EOS models with respect to monomer-polymer separation and polymer properties. This has a major impact on the results of the simulation of a certain design. Since the analysis was limited to a single process (HP-LDPE) we can not generalize the conclusions, but these results are an effective starting point in approaching other polymerization systems. Note that all calculations, including an industrial reactor simulations, were rigorously performed using the state-of-the-art software Aspen Polymers Plus that is a layered product built on top of AspenTech Aspen Plus.
Each EOS properties model has exclusive characteristics that affect the results of modeling the pure monomer and polymer mixture behavior. Out of the four EOS models used, Sanchez-Lacombe and Polymer-SRK EOS produced the best acceptable fit of the data. Both models are capable of predicting properties at both high and low-pressure, and give accurate results especially for high temperatures.
References
1. A. A. Kiss, C. S. Bildea, A. C. Dimian, P. D. Iedema, Chem. Eng. Sci., 2003, 58, 2973.
2. W. Hao, H.S. Elbro, P. Alessi, Chemistry Data Series, XIV, DECHEMA, 1992.
3. N. Koak, R.M. Visser, TH.W. de Loos, Fluid Phase Equilibria, 1990, 158-160, 835.
4. Stanley H. Huang and Maciej Radosz, Ind. Eng. Chem. Res., 1990, 29, 2284.
Presented Tuesday 18, 15:00 to 15:20, in session Thermodynamics: General (T2-1e).
