蒸氣誘導式相分離成膜過程之研究

Investigation on the Membrane Formation Mechanism of Vapor-Induced Phase Separation

蘇酉申(Yu-Shen Su)

指導教授 ： 王大銘

國立臺灣大學/工學院/化學工程學研究所/博士(2009年)

https://doi.org/10.6342/NTU.2009.02452

摘要

本研究探討以polysulfone/N-methyl pyrrolidinone, polyetherimide)/NMP及polymethylmethacrylate)/NMP進行水蒸氣誘導式相分離(VIPS, vapor-induced phase separation)之成膜過程。對於PSf/NMP及PEI/NMP系統，可觀察到將高分子溶液暴露於潮濕空氣中，可以引發離相分解(spinodal decomposition)並生成暫時性的雙連續(bi-continuous)結構，再進一步發展成為胞狀(cellular)孔洞。關於PMMA/NMP，其截面結構初期形成雙連續結構，之後進一步由表層至底部發展成為薄層、顆粒狀、雙連續及近似胞狀結構。

為了進一步了解蒸氣誘導式相分離過程中的相分離機制及結構演變過程，研究中以傅立葉紅外光顯微鏡(FT-IR microscope)來分析成膜過程中鑄膜液不同位置處各成分的濃度變化。因為所用溶劑NMP對水的高親和性，在鑄膜液與濕空氣接觸時會將水吸入溶液中，當水的濃度越來越高，鑄膜液會由均相狀態進入介穩態(meta-stable sate)，甚至進入非穩態(unstable state)而產生相分離。由傅立葉紅外光顯微鏡所量得的各成分濃度變化，配合高分子/NMP/水的熱力學相圖，可以定出不同位置處的鑄膜液是否進入介穩區或非穩區，並將溶液中達到介穩區的最深位置稱為介穩前緣(metastability front)，達到非穩區的最深位置稱為非穩前緣(unstability front)。在成膜過程中，由於水不斷進入鑄膜液中，介穩前緣及非穩前緣會隨成膜時間朝較深的地方前進。另外溶液相分離後，所生成的高分子富相及貧相區域會成長合併，當富相或貧相區長大到足以散射可見光時，溶液就會變混濁。可以利用光學顯微鏡觀察鑄膜液變混濁的情形，將變混濁的最深位置稱為混濁前緣(turbidity front)。

比較介穩前緣、非穩前緣及混濁前緣的移動情形，再配合成膜過程中膜結構的演變過程，可以對蒸氣誘導相分離過程有更深入的了解。成膜過程中，因為進入介穩區所需的水量較進入非穩區少，所以溶液會先進入介穩區。在成膜初期時，進入介穩區的溶液並不會相分離，要等到非穩區的前緣抵達後，溶液進入非穩區，才會進行spinodal decomposition而產生雙連續結構。這些雙連續結構會因要降低高分子貧、富相間的界面張力而成長，進而轉變成胞狀或顆粒狀的結構使得溶液變混濁。所以在成膜初期可觀察到介穩前緣領先非穩前緣，而混濁前緣落於最後。但混濁前緣一旦形成後，前進的速度有機會較非穩前緣的速度更快。若混濁前緣趕不上非穩前緣時，鑄膜液各處的相分離機制均是spinodal decomposition，初期的結構均為雙連續。但當混濁前緣可以趕上非穩前緣後，在介穩區(在介穩前緣與非穩前緣之間)的鑄膜液會受到干擾，失去穩定度而進行成核成長(nucleation and growth)的相分離機制，形成胞狀結構；此時混濁前緣會領先非穩前緣，並會趕上介穩前緣。在混濁前緣趕上非穩前緣之前，溶液的相分離機制為spinodal decomposition，但當混濁前緣趕上之後，相分離機制就由nucleation and growth取代。

上述機制可以解釋為何10 wt% PSf/NMP鑄膜液及20 wt% PMMA/NMP鑄膜液的離相分解得以發生於整層薄膜，而20 wt% PSf/NMP鑄膜液及20 wt% PEI/NMP鑄膜液的離相分解只能發生在薄膜表面附近。

關鍵字

蒸氣誘導式相分離 ； 質量傳遞 ； 傅立葉轉換紅外光顯微鏡 ； 成核成長 ； 離相分解

並列摘要

Investigated in the present work is the preparation of membranes by the vapor-induced phase separation (VIPS) of polysulfone (PSf) /N-methyl-pyrrolidinone (NMP), polyetherimide (PEI)/NMP and poly methyl methacrylate (PMMA)/NMP solutions. For the PSf/NMP and PEI/NMP systems, it was observed that the exposure of the polymer solution to humid air was able to bring about spinodal decomposition to the solution, resulting in transient bi-continuous structure that later developed into cellular pores. For the PMMA/NMP system, the initial structure was bi-continuous throughout the whole cross-section, but it became non-uniform, forming a thin dense skin, followed by nodular, bi-continuous, and cellular-like structures from top to bottom.

To get insight into the mechanism of phase separation and morphology evolution associated with VIPS, we employed FT-IR microscopy to measure the composition change during membrane formation at different positions of the cast film. Because of the high affinity of NMP for water, water vapor was imbibed into the cast film when the film was exposed to humid air. As the water concentration at a certain position was high enough, the polymer solution moved from the homogeneous state to the meta-stable state or even the unstable state, resulting in phase separation. On the basis of the time dependence of the solution compositions at a certain position in a cast film, measured with FT-IR microscope, we can determine if the solution was in the meta-stable or unstable region, with the help of the polymer/NMP/water ternary phase diagram. We then defined the deepest position where the solution was in the meta-stable region as the metastability front, and the deepest position where the solution was in the unstable region as the unstability front. During membrane formation, due to the intake of water from the humid air, the metastability and unstability fronts moved toward the deeper region of the cast film. In addition, after phase separation, the formed polymer-rich and polymer-poor domains coalesced and grew. When the domains grew large enough to scatter visible light, the solution became turbid. We employed the optical microscope to observe the propagation of the turbidity in the cast film, and defined the deepest position where the solution was turbid as the turbidity front.

With comparison of the movement of the metastability, unstability, and turbidity fronts, together with the observation of the morphology evolution during membrane formation, new understanding of the mechanism of VIPS was obtained. Because it took less water to reach the meta-stable region than to reach the unstable region, the solution entered first into the meta-stable region. During the initial period of membrane formation, the solution in the meta-stable region did not phase separate. It performed spinodal decomposition and resulted in bi-continuous structure as the solution entered into the unstable region, when the unstability front reached. To reduce the interfacial energy between the polymer rich and poor phases, the bi-continuous structure coarsened and turned into nodular or cellular structure, making the solution turbid. Therefore, we observed that, during the initial period, the unstability front lagged the metastability front, but led the turbidity front. But, once the turbidity front occurred, it might travel faster than the unstability front. When the turbid front cannot catch the unstability front, the cast film underwent spinodal decomposition everywhere. On the other hand, when the turbidity front caught the unstability front, the solution in the meta-stable region between the metastability and unstability fronts, was disturbed by the arrival of the turbidity front and demixed via the nucleation and growth mechanism, forming cellular structure. Under this circumstance, the turbidity front led the unstability front and might catch the metastability front. To summarize, the phase separation mechanism was spinodal decomposition before the turbidity front caught the unstability front, but after that the mechanism turned to nucleation and growth.

As a result, the competition between the travel of the unstability front driven by the water mass-transfer, and that of the turbidity front, initiated by the interfacial tension between the phase rich and poor domains, would determine the thickness of the membrane region containing bi-continuous structures. The above mechanism can explain why spinodal decomposition occurred throughout the whole cast film for the cases with 10 wt% PSf/NMP and 20 wt% PMMA/NMP as the casting solution, while spinodal decomposition was confined in a small region near the film surface for other cases with 20 wt% PSf/NMP and 20 wt% PEI/NMP.

並列關鍵字

vapor-induced phase separation ； mass transfer ； FT-IR microscopy ； nucleation and growth ； spinodal decomposition

參考文獻

1. Mulder, M., Basic Principles of Membrane Technology. 2nd ed.1996, Dordrecht, Netherlands, Boston: Kluwer Academic.

2. Matsuyama, H., M. Teramoto, and T. Uesaka, Membrane formation and structure development by dry-cast process. Journal of Membrane Science, 1997, 135(2), 271-288.

3. Wang, D.M., T.T. Wu, F.C. Lin, J.Y. Hou, and J.Y. Lai, A novel method for controlling the surface morphology of polymeric membranes. Journal of Membrane Science, 2000, 169(1), 39-51.

4. Srinivasarao, M., D. Collings, A. Philips and S. Patel, Three-dimensionally ordered array of air bubbles in a polymer film. Science, 2001, 292(5514), 79-83.

5. Brown, G.L., Formation of films from polymer dispersions. Journal of Polymer Science, 1956, 22(102), 423-434.