열역학 이론과 분자 동역학 전산모사를 결합한 고분자 용액의 물성 연구

Abstract

We performed the thermodynamic calculation and molecular simulation to understand the phase behavior of polymer solution. The thermodynamic potentials were derived accurately based on Gibbs rule, and were calculated adequately. The molecular simulation was adopted which was applicable to obtain the model parameters of mathematical equation or the structural behavior of polymer solution at the given temperature and composition. In chapter I, Poly N-isopropylacrylamide-co-ethylacrylate [P(NIPAM-co-EA)], Poly N-isopropylacrylamide-co-2-hydroxyethylmethacrylate [P(NIPAM-co-HEMA)], and Poly N-isopropylacrylamide-co-2-hydroxyethylacrylate [P(NIPAM-co-HEA)] nano-sized particle copolymer hydrogels were synthesized to investigate their volume phase transition behavior. Increasing the hydrophobic or hydrophilic monomer content of hydrogels led to uniform changes in transition temperature and swelling ratio. The classical interaction energy parameter in the thermodynamic model, which is not suitable for accurately representing the equilibrium swelling of a copolymer hydrogel, was modified semi-empirically. This step was emphasized in order to reflect the individual contribution of each hydrogel constituent. In addition, we investigated the volume phase transition of PNIPAM gels as a function of crosslinker chain length. The crosslinkers used in this study were N,N'-methylenebisacrylamide (BIS) and two different molecular weights of Poly(ethyleneglycol) diacrylate (PEGDA, 575 and 700). Interestingly, the use of a long chain crosslinker resulted in a decreased volume change compared to that of a short chain crosslinker without any significant changes in transition temperature. To understand this unexpected result, we used a molecular dynamics simulation for BIS and PEGDA 575 to demonstrate the micro-scale conformation change of crosslinker chain mixed with water or ethanol. The electrochemical properties of polymer electrolyte solution were predicted by combining the thermodynamic model and molecular simulation in chapter 2. We obtained all parameters in an ionic conductivity model from an atomistic simulation and removed all adjusted model parameters. From a microscopic point of view, the hydrated perfluorosulfonic acid (PFSA) membrane shows micro-phase segregation which separated into hydrophilic and hydrophobic phases. This chapter originated with this phenomenon and we treated this phase segregation as if it was a continuous phase for each of which the proton (H+) was transported inside the PFSA membrane/solvent (water and alcohols) mixture. The chemical potential for a given system was estimated using a molecular simulation technique to predict the van der Waals interaction energy between the polymer and solvent. In addition, the self diffusion coefficients were calculated from the molecular dynamics simulation. We studied various polymer/solvent compositions to understand the concentration dependence of self diffusion coefficient. In chapter III, the morphological characteristics of proton exchange membranes were estimated using coarse-grained beads into a dissipation particle dynamics simulation. The thermodynamic potentials were not directly involved in this work, however, the particle dynamics simulation required the thermodynamic model parameters to calculate non-bonded interaction. We enlarged the time and sizes of the simulation above the micro-scale, resulting in stable particle diffusivity during simulation. Two membranes, a well known material PFSA and a recently developed material disulfonated poly(arylene ether sulfone), were used in this study. The simulated cells reached an equilibrium state after time scales of 500 (DPD units), then showed a microphase segregation suggested by many researchers during over the last decade. We compared the structures of the as-hydrated membrane with the cluster network model by Gierke, the cluster channel network model by Kreuer and the modified cluster network model by Newman. The simulation results correspond well with experimental AFM data as well as various suggested physical models at different simulation conditions. |고분자 용액의 상 거동을 이해하기 위해 열역학적 계산과 분자 시뮬레이션을 수행하였다. 모든 열역학 포텐셜 식들은 깁스 룰을 기반으로 하여 정확히 유도 되었고 알맞게 계산 되였다. 분자 시뮬레이션은 수학식의 모델 변수를 구하거나 혹은 특정 온도와 조성 조건에서의 고분자 용액의 구조 변이 특성을 예측하는 것에 적용하기 위하여 도입되었다. 챕터 I 에서는 Poly N-isopropylacrylamide-co-ethylacrylate [P(NIPAM-co-EA)]와 Poly N-isopropylacrylamide-co-2-hydroxyethylmethacrylate [P(NIPAM-co-HEMA)], Poly N-isopropylacrylamide-co-2-hydroxyethylacrylate [P(NIPAM-co-HEA)], 총 세가지의 공중합체 하이드로겔을 부피 변이 거동을 알아보기 위하여 나노 사이즈 입자 젤로 합성하였다. 하이드로겔의 부피변이 온도와 팽윤 비율의 변화를 주기 위하여 공단량체의 친수성과 소수성의 차이를 두었다. 기존에 사용하던 열역학 이론의 상호 에너지 변수는 이러한 변화를 제대로 표현할 수 없는 한계를 가지고 있어 식의 형태를 반경험적 형태로 변형시켰다. 이 과정은 하이드로겔의 공단량체와 가교제 각각의 상호 에너지 기여도를 모두 고려할 수 있기 때문에 매우 중요한 과정이라 할 수 있다. 추가적으로 가교제의 사슬 길이에 따라 PNIPAM 하이드로젤의 부피 상 전이에 미치는 영향을 알아보기 위한 내용도 포함되어 있다. 이 논문에 쓰인 가교제는 N,N'-methylenebisacrylamide (BIS)와 두개의 다른 분자량을 가진 Poly(ethyleneglycol) diacrylate (PEGDA, 575 and 700)를 사용하였다. 흥미롭게도 긴 사슬을 가진 가교제를 쓴 하이드로겔의 부피 팽창비는 가장 짧은 사슬길이를 쓴 하이드로겔보다 더 작았다. 이런 예측하지 못한 결과를 이해하기 위하여 BIS와 PEGDA575 두 가지 가교제를 물과 에탄올에 혼합 시키는 분자 동역학 시뮬레이션이 도입되었다. 챕터 II 에서는 열역학 모델과 분자 시뮬레이션을 결합하여 고분자 전해질의 전기화학적인 물성을 얻어내는 것에 주력하였다. 모든 이온 전도도 식의 변수들은 분자 수준의 시뮬레이션을 이용하여 얻어졌다. 미시적인 관점으로 보자면, perfluorosulfonic acid (PFSA) 분리막은 친수성 영역과 소수성 영역이 나뉘는 미시적 상 분리가 일어난다. 이 챕터에서 다루는 기본적인 요소는 이러한 미시적 상 분리가 연속적인 상으로 일어난다고 가정하고 수소이온이 이 연속 채널을 이용하여 이동한다고 생각하였다. 화학 퍼텐셜 값은 분자 시뮬레이션을 이용하여 고분자와 용매간 반데르발스 상호 에너지를 얻어냄으로써 구해졌다. 추가적으로, 자가 확산 계수 값은 여러 농도 에서의 분자 동역학 시뮬레이션으로 구할 수 있었다. 챕터 III 에서는 앞에서 알아보았던 PFSA분리막의 미시상 구조를 coarse-grained bead로 연성하여 입자 동역학 시뮬레이션으로 그 형태적 특성을 파악하였다. 열역학적 퍼텐셜들은 이 챕터에 직접적인 연관을 주지는 않았지만 입자 동역학 시뮬레이션을 수행하는데 핵심적인 변수 값을 얻어내는 과정에 꼭 필요한 요소로 존재했다. 우리는 분자 동역학과는 다르게 시간 조건과 시뮬레이션을 수행할 수 있는 가상 공간의 사이즈를 넓히고 입자들의 확산이 안정적으로 이동할 수 있도록 하는 조정을 거치게 하였다. 이번 챕터에 다룬 이온 전도성 분리막은 PFSA 와 최근 개발된 disulfonated poly(arylene ehter sulfone) 분리막이었다. 시뮬레이션 공간은 DPD unit으로 500 시간 단위에서 안정화가 일어났고 그 이후 미시적 구조 변형이 일어났다. 이 챕터에서는 이온전도성 고분자의 여러 물리적 모형을 비교하고 어떤 조건에서 모형과 흡사한 형태를 보일 것인지를 보여줄 것이며 실제 원자현미경(AFM) 데이터와 의 비교를 보여줄 것이다.; We performed the thermodynamic calculation and molecular simulation to understand the phase behavior of polymer solution. The thermodynamic potentials were derived accurately based on Gibbs rule, and were calculated adequately. The molecular simulation was adopted which was applicable to obtain the model parameters of mathematical equation or the structural behavior of polymer solution at the given temperature and composition. In chapter I, Poly N-isopropylacrylamide-co-ethylacrylate [P(NIPAM-co-EA)], Poly N-isopropylacrylamide-co-2-hydroxyethylmethacrylate [P(NIPAM-co-HEMA)], and Poly N-isopropylacrylamide-co-2-hydroxyethylacrylate [P(NIPAM-co-HEA)] nano-sized particle copolymer hydrogels were synthesized to investigate their volume phase transition behavior. Increasing the hydrophobic or hydrophilic monomer content of hydrogels led to uniform changes in transition temperature and swelling ratio. The classical interaction energy parameter in the thermodynamic model, which is not suitable for accurately representing the equilibrium swelling of a copolymer hydrogel, was modified semi-empirically. This step was emphasized in order to reflect the individual contribution of each hydrogel constituent. In addition, we investigated the volume phase transition of PNIPAM gels as a function of crosslinker chain length. The crosslinkers used in this study were N,N'-methylenebisacrylamide (BIS) and two different molecular weights of Poly(ethyleneglycol) diacrylate (PEGDA, 575 and 700). Interestingly, the use of a long chain crosslinker resulted in a decreased volume change compared to that of a short chain crosslinker without any significant changes in transition temperature. To understand this unexpected result, we used a molecular dynamics simulation for BIS and PEGDA 575 to demonstrate the micro-scale conformation change of crosslinker chain mixed with water or ethanol. The electrochemical properties of polymer electrolyte solution were predicted by combining the thermodynamic model and molecular simulation in chapter 2. We obtained all parameters in an ionic conductivity model from an atomistic simulation and removed all adjusted model parameters. From a microscopic point of view, the hydrated perfluorosulfonic acid (PFSA) membrane shows micro-phase segregation which separated into hydrophilic and hydrophobic phases. This chapter originated with this phenomenon and we treated this phase segregation as if it was a continuous phase for each of which the proton (H+) was transported inside the PFSA membrane/solvent (water and alcohols) mixture. The chemical potential for a given system was estimated using a molecular simulation technique to predict the van der Waals interaction energy between the polymer and solvent. In addition, the self diffusion coefficients were calculated from the molecular dynamics simulation. We studied various polymer/solvent compositions to understand the concentration dependence of self diffusion coefficient. In chapter III, the morphological characteristics of proton exchange membranes were estimated using coarse-grained beads into a dissipation particle dynamics simulation. The thermodynamic potentials were not directly involved in this work, however, the particle dynamics simulation required the thermodynamic model parameters to calculate non-bonded interaction. We enlarged the time and sizes of the simulation above the micro-scale, resulting in stable particle diffusivity during simulation. Two membranes, a well known material PFSA and a recently developed material disulfonated poly(arylene ether sulfone), were used in this study. The simulated cells reached an equilibrium state after time scales of 500 (DPD units), then showed a microphase segregation suggested by many researchers during over the last decade. We compared the structures of the as-hydrated membrane with the cluster network model by Gierke, the cluster channel network model by Kreuer and the modified cluster network model by Newman. The simulation results correspond well with experimental AFM data as well as various suggested physical models at different simulation conditions.