Effect of water on the solubility and diffusivity of hydrocarbons in polyethylene.
This article is devoted to analyzing the effect of water on the removal of the residual monomer and solvent used during the polymerization of ethylene in order to define a more effective devolatilization process. There are a number of well-known polyethylene polymerization licenses which are applied in about 80% of the world's polyethylene production. UNIPOL, HIMONT, and BOSTAR are among the most common. In these processes, the ethylene polymerization takes place in a fluidized bed reactor or in a combination of this with a loop reactor depending on which stage the catalyst is fed into the process. In another commercial process, polymerization occurs in solution using a mixture of hydrocarbons as the polymerization medium.
If the polymer is produced by suspension polymerization, there is a stripping stage to lower the concentration of residual solvents and unconsumed monomers to safe levels in the final product, which is typically pellets. In gas phase polymerization granular polymer from the fluidized reactor go to a product purge bin to remove the residual hydrocarbons and then is fed into an extruder to be pelletized. In either case failure to remove the volatile organics can cause a building-up of dangerous gas mixtures after long storages periods or result in not meeting the specifications for the final product. This removal is usually accomplished by passing hot gas and/or steam through the material. As the removal of solvent involves the diffusion of the volatiles through the solid into the gas phase, higher stripping rates are obtained at the maximum temperature at which the solids can be handled without excessive softening [l]. Steam, being an excellent heat transfer medium and possessing a high latent heat of condensation, is preferred to other hot gases in order to minimize the flow rates of the stripping agent.
The design of this equipment is based on the correct estimation of diffusivities of the organic volatiles in the polymer phase, since most of the mass transfer resistance is located in the solid phase rather than in the gas phase. In addition, the solubility of the residual components has a key role in the stripping process because their initial concentrations will dictate the amount of stripping agent to be used as well as the retention time of the pellets in the devolatilization unit. Thus, the experimental measurement of solubility and diffusion coefficients is of profound importance. A great number of experimental studies dealing with both gases and vapors in polyethylene have been reported in the literature. Solubility of gases, generally light olefins such as ethylene or propylene, is usually studied at high pressures, similar to the ones encountered during the gas-phase polymerization reaction [2-4], On the other hand, studies involving organic vapors are conducted at atmospheric pressures. Rogers et al.  used a quartz helix microbalance to measure solubilities and diffusivities of n-hexane, n-heptane, and benzene in polyethylene. Castro et al.  measured the solubility of other linear paraffins using a permeation cell. Doong and Ho , measured the diffusivity of aromatics in semicrystalline polyethylene by means of a gravimetric sorption technique.
All these studies have been limited to the case where the polymer is completely dry. When steam is used as the stripping medium, it is commonly found that the polymer pellets have surfaces covered with a film of water produced from the condensation of steam. Therefore, knowledge of thermodynamic and mass transfer effects that water may have on the system is vital to properly designing stripping units. Two possible effects of water on the system are (a) lower the solubility of solvents and monomers in the polyethylene-water mixture or (b) plasticize the polyethylene resulting in increased diffusion rates.
A few studies were found in the literature where the effect of water was analyzed with some contradictory outcomes. Mathews et al.  studied a commercial stripping unit comprised of a precipitator and a steam stripper. Experiments conducted at 105[degrees]C and 115[degrees]C using this unit showed that the presence of water substantially reduced the diffusion coefficient for n-hexane while testing sheets of the material. When the actual crumbs or pellets were used, the stripping rate was unaffected by the liquid phase surrounding the particles. They conclude that the rate of removal was controlled by particle structure and surface morphology. On the other hand, Humkey et al.  patented an apparatus for stripping residual solvent out of polymer pellets using steam as the stripping medium. They found a considerable increase in the removal rate of cyclohexane at 95[degrees]C from polyethylene when they used a mixture of steam and nitrogen instead of pure nitrogen. They claimed that the wet system shifted the equilibrium conditions compared with the dry one with just nitrogen.
Data for the sorption of water and the effect of water on the solubility and diffusivity of solvents into polyethylene were measured by three different methods: pressure decay, finite inverse gas chromatography (IGC), and the static sorption method. Each of these methods has been described previously in the literature so only a brief summary of the methods and their characteristics is presented here.
The pressure decay technique measures the uptake of a solvent by monitoring the pressure change in a constant volume-constant temperature sample chamber as the solvent sorbes into the polymer , Using an accurate equation of state the pressure drop is converted into a mass uptake. Using the numerical analyses provided by Crank  the diffusivity is determined primarily from the initial slope of the curve and the solubility by the long-time equilibrium uptake. In the current work this method was used to examine the solubility and diffusivity of ethylene in polyethylene at high pressure.
Finite IGC was used by Tihminlioglu et al.  in 1997 to obtain data for toluene in polystyrene and polyfvinyl acetate) at finite concentrations. A capillary column is brought to equilibrium with a plateau concentration of the solvent and then an additional pulse of the solvent is introduced and the resulting peak is numerically analyzed for the solubility and diffusivity. The results were shown to be in agreement with independent measurements made by gravimetric sorption and piezoelectric sorption techniques. In the current work this method was used to measure the solubility and diffusivity in polyethylene of water and of 1-octene and cyclohexane at different water activities and low pressure.
The static sorption technique was used by Palamara et al.  to measure the diffusivity and solubility of gases in polymers at elevated pressures. In this method the solvent was introduced into a set of four capsules all at the same temperature. The pressure of the solvent was held constant and the capsules were sealed at increasing times. By measuring the uptake in each capsule by weight the slope of the uptake curve was determined. The solubility was found after allowing one of the capsules to come to equilibrium. In the current work this method was used to examine the uptake of ethylene at elevated pressure in polyethylene in the presence of water.
Solubility of Water
The actual solubility of water in polyethylenes depends of course on the characteristics of the polyethylene (density, crystallinity, void volume, etc.). One would expect, however, that in general the solubility of water would be relatively the same. A number of commercial polyethylenes are characterized by their distributors as having a water solubility of less than 5 x [10.sup.-4] weight fraction. There are some estimates in the literature. McCall et al.  studied the effect of oxygen on water sorption in a low-density polyethylene and reported for the unoxidized sample a weight fraction of about 1 x [10.sup.-4] at 60[degrees]C and low pressure. Suherman et al.  studied a high-density polyethylene at 25[degrees]C and low pressure and also reported a water solubility of about 1 X [10.sup.-4] weight fraction.
In the current study the solubility of water on a low-density polyethylene was measured using finite inverse gas chromatography. The polyethylene used to coat the capillary column with a film thickness of 3.0 pm was obtained from NOVA Chemical, Corp. It had a molecular weight of approximately 145,000, a polydispersity of 5.5, a crystallinity of 49 wt%, and a melt temperature of approximately 110[degrees]C. The detector for the IGC was a mass spectrometer which proved to be very sensitive to water. It was difficult to use the regression algorithm previously describe by Tihminlioglu et al.  to determine the partition coefficients because of the very low solubility of the water. (The partition coefficient is the concentration in the polymer divided by the concentration in the vapor when both concentrations are in the same units.) Therefore, the partition coefficients, K, were calculated using the traditional equation .
K = [273.2[[rho].sub.2]/T[m.sub.2]]([t.sub.r] - [t.sub.c])Fj
Here [[rho].sub.2] is the density of the polymer, [m.sub.2] is the mass of polymer, [t.sub.r] and [t.sub.c] are the retention times of solute and marker gas, T is the column temperature, F is the flow rate of gas, and j is a pressure drop correction which is negligible for capillary columns.
The results are shown in Fig. 1. For a water activity up to about 0.4 the weight fraction of water in low density polyethylene was about 7 x [10.sup.-4]. The lack of temperature dependency when solubility is plotted as a function of activity is the expected behavior. At 125[degrees]C the pressure at an activity of 0.4 is about 169 kPa.
Effect of Water at Low Pressures
A form of the finite IGC method was used to evaluate the effect of water on the behavior of cyclohexane and 1-octene in low density polyethylene. The experiments consisted of two parts. In the first instance pure helium was used as the carrier gas: in the second case the helium was bubbled through a saturator containing distilled water and the column was brought into equilibrium with the wet vapor. In both cases the column was pulsed with the solvent of interest. Thus, the results are for infinitely dilute solvent in the dry polymer or at a fixed concentration of water in the polymer.
Data for both the solubility and diffusivity of cyclohexane and 1-octene were collected for five temperatures between 100[degrees]C and 125[degrees]C. The results for the solubility at the two extreme temperatures are shown in Fig. 2 in tenus of the partition coefficient. The vapor pressure leaving the saturator ranged from 4 to 586 kPa. Clearly the partioning of the two solvents is not affected by the water concentration in the polyethylene. Figure 3 shows the diffusion coefficients for the same cases. Within the error estimates there is no effect of the water concentration on the diffusivities.
The final measurements were on the behavior of ethylene in low-density polyethylene pellets at elevated pressures using both the static sorption and the pressure decay methods. The pellets obtained from NOVA Chemical, Corp. had a spherical geometry with an average diameter of 3.33 [+ or -] 0.04 cm and a crystallinity of 49 wt%. The pressure decay method was used at 80[degrees]C and 100[degrees]C over a pressure range of about 500-6000 kPa only for the dry pellets. The pressure was measured as a function of time using the data acquisition method described by Perez-Bianco et al.. For the static capsule method the pellets were loaded into the capsules and data were first collected on the dry polymer.
Over a period of time the ethylene was shut off and the capsule was weighed to get a measure of the uptake as a function of time in order to determine the diffusivity. Long time exposure was used to determine the equilibrium solubility. For the wet experiments water was added to the capsules. The capsules were then very briefly exposed to a vacuum to remove the air before the measurements with the ethylene. Figures 4 and 5 show the results for the solubility of the ethylene with both the dry and wet pellets at 80[degrees]C and 100[degrees]C. Figure 6 depicts the diffusion coefficients for both cases at both temperatures. Excellent agreement was found between the pressure decay method and the static capsule method for the dry pellets. There is some scatter in the diffusivity data, but the conclusion is clear--there is essentially no effect of the water on either the solubility or the diffusivity.
Three different experimental methods have been used to investigate the effects of water on the sorption and diffusion of liquid solvents and gases in polyethylene. The data clearly indicate that there is no measureable effect of having water in a devolatilization process, at least in terms of the solubility and rate of mass transfer.
[1.] F.J. Matthews, J.R. Fair, J.W. Barlow, D.R. Paul, and C. Cozewith, Ind. Eng. Client. Prod. Res. Dev., 25, 65 (1986).
[2.] C. Kiparissides, V. Dimos, T. Boultouka, A. Anastasiadis, and A. Chasiotis, J. Appl. Polym. Sci., 87, 953 (2003).
[3.] S.S. Kulkarni, and S.A. Stern, J. Polym. Sci.: Parr B Polym. Phys., 21, 441 (1983).
[4.] N.N. Li, and R.B. Long, AlChE J., 15, 73 (1969).
[5.] C.E. Rogers, V. Stannett, and M. Szwarc, J. Polym. Sci., 45, 61 (1960).
[6.] E.F. Castro, E.E. Gonzo, and J.C. Gottifredi, J. Membr. Sci., 31, 235 (1987).
[7.] H.K. Dong, B.W. Liang, M.C. Ho, S. Hung, and C.W. Tu, J. Cryst. Growth, 124, 181 (1992).
[8.] R.G. Humkey, and D.J. Ryan, U.S. Patent 3,958,585 (1976).
[9.] M. Perez-Bianco, J.R. Hammons, and R.P. Danner, J. Appl. Polym. Sci., 116, 2359 (2010).
[10.] J. Crank, The Mathematics of Diffusion, Oxford University Press, London, England (1975).
[11.] F. Tihminlioglu, R.K. Surana, R.P. Danner, and J.L. Duda, J. Polym. Sci.: Part B Polym. Phys., 35, 1279 (1997).
[12.] J.E. Palamara, K.A. Mulcahy, A.T. Jones, R.P. Danner, and J.L. Duda, Ind. Eng. Client. Res., 44, 9943 (2005).
[13.] D.W. McCall, D.C. Douglass, L.L. Blyler, G.E. Johnson, L.W. Jelinski, and H.E. Bair, Macromolecules, 17, 1644 (1984).
[14.] M.P. Suherman, and E. Tsotsas, Reaktorr, 13, 89 (2010).
[15.] J.R. Conder, and C.L. Young, Physicochemical Measurements by Gas Chromatography, Wiley, New York (1979).
Jose Roman Galdamez, Adam T. Jones, Ronald P. Danner
Department of Chemical Engineering, The Pennsylvania State University, University Park, Pennsylvania 16802
Correspondence to: Ronald P. Danner; e-mail: email@example.com
Contract grant sponsor: NOVA Chemicals Corporation, Calgary, Alberta, Canada.
Published online in Wiley Online Library (wileyonlinelibrary.com).
|Printer friendly Cite/link Email Feedback|
|Author:||Galdamez, Jose Roman; Jones, Adam T.; Danner, Ronald P.|
|Publication:||Polymer Engineering and Science|
|Date:||Jun 1, 2015|
|Previous Article:||Determination of compounding formulation of cured rubber by reverse engineering.|
|Next Article:||Long chain branched impact copolymer of polypropylene: microstructure and rheology.|