Article Preview
TopIntroduction
In recent years there has been an increasing demand for acrylic fibres due to their extensive application as the main precursor for high strength carbon fibres (Paiva, 2003). It has been shown that the properties of carbon fibres strongly depend on the structure of the initial precursor (Mital, 1998; Thunemann, 2000). Hence, understanding the mechanisms of structure development of acrylic fibre during spinning is a key to design better precursors and improve mechanical properties of the resulting carbon fibre (Mital, 1998; Paiva, 2003; Thunemann, 2000). The acrylic fibres used in carbon fibre industry are mainly produced by wet-spinning process (Salem, 2001). Wet-spinning is an immersion precipitation process in which the extruded polymer solution (spinning dope) precipitates as a porous fibre in the coagulation bath (Chen, 2006; Masson, 1995; Wang, 2007). It is a complex process, including phase separation, rheological and diffusion phenomena (Hou, 2005, 2006; Ziabicky, 1976). Phase separation occurs through nucleation and growth of the polymer lean phase leading to void formation with controlled morphology. Solvent/nonsolvent exchange rate between fibre precursor and coagulation medium has been mentioned as the main parameter affecting void size, size distribution and morphology (Knaul, 1997; Lai, 1999; Li, 2004; Lin, 2002; Liu, 1990; Oh, 1996; Um, 2004). Nonsolvent inflow leads to nucleation of the polymer lean phase often surrounded by polymer rich phase forming core-shell nuclei (Mckelvey, 1996). Osmotic pressure difference between the polymer lean phase and coagulation medium drives an intrusive mass into the fibre and thus core of nuclei grows against the polymer rich shells (Lin, 2002). Therefore, the mixing affinity between fibre precursor and nonsolvent on one hand and the mechanical resistance of the polymer reach phase against deformation on the other hand, determines the kinetics of phase separation (exchange rate and demixing time) and the corresponding void structure. Smolders et al. (1992) conclude that instantaneous and delayed demixing often yield finger-like and sponge-like structures, respectively. Phase separation kinetics can be manipulated in many ways such as variation of spinning dope composition, coagulation bath composition and spinning temperature. For example, increasing polymer concentration in the spinning dope, reduces solvent concentration gradient between fibre and coagulation bath and consequently solvent / nonsolvent diffusion coefficient decreases, solution viscosity increases and growth of the nucleated voids is inhibited (Hou, 2005, 2006). Won et al. (1999) investigate the effect of dope viscosity on void formation. They notice that low solution viscosity results usually to macrovoids formation. But, increasing solution viscosity via increasing polymer concentration leads to a significant reduction in size and number of macrovoids. They attribute this effect to gelation of the polymer phase. Gelation onset causes a reduction in the rate of phase separation and prevents voids growth. Van de Witte et al. (1996) attribute macrovoids reduction to the increase of phase separation delay time as a result of increasing solution concentration or viscosity. Precipitant addition to dope or spinning solution also affects demixing time via reducing the rate of solvent outflow and nonsolvent inflow both rheologically and thermodynamically. In other words, assuming constant exchange rate of solvent outflow and nonsolvent inflow, dope containing less precipitant would show the least viscosity, fastest phase separation and formation of finger-like voids. Precipitant addition, however, strongly affects pair-wise interactions among polymer/nonsolvent, polymer/solvent and solvent/nonsolvent and often leads to the reduction of mass transfer rate between the spinning dope and the coagulation medium. According to the calculations and modelling by Yilamz and Machugh (1986) and Tsay and Machugh (1990), an increase in solvent/nonsolvent interaction parameter (
) enhances the miscibility gap in the phase diagram which corresponds to delayed demixing. Furthermore, their results verify the decrease of the miscibility gap via increase of the polymer/solvent (
) or polymer/nonsolvent (
) interaction parameter leading to instantaneous demixing (Yilmaz, 1986; Tsay, 1990). Based on aforementioned qualitative data evaluations, Ruaan et al. (1999) propose 
criteria for selecting solvent and coagulation medium in view of macrovoid formation in wet phase-inversion processes:
where
,
and
are solubility differences among polymer/solvent, polymer/nonsolvent and solvent/nonsolvent, respectively and
is polymer solubility parameter. By addition of nonsolvent to dope or solvent to coagulation medium, however,
and
are substituted with
and
, respectively where m stands for mixed solvents or nonsolvents. Using
criterion, Ruaan et al. (1999) proceed to predict correctly void morphologies in some wet phase inversing systems while they also report some discrepancies. In addition, van de Witte et al. (1996) obtain membranes with void structures on the contrary with what is expected based on phase separation time scale. In other words, nonsolvent addition to the dope in their systems led to shorter delayed time contrary to what is anticipated. On the other hand, the role of polymer reach phase rheological resistance against polymer lean phase (nuclei) growth during morphological evolution in wet phase inversion process is often ignored by many research groups or considered just qualitatively.