| 摘要: | 金屬有機框架(metal-organic framework, MOFs)可透過調節劑 (modulator synthesis)以改善結晶結果。然而，當調節劑添加量增加而影響到反應酸鹼值(pH)時，則無法達成預期結果，其配位基與調節劑的交換機制仍不明確。
在本研究中，我們以Zr6O4(OH)4(C4H3O4)11(CHO2)為模型，並模擬其在pH 4.6 及 pH 5.0下能量級距圖(energy landscapes)，利用密度泛函理論(density function theory, DFT)進行計算。我們的結果顯示，在質子濃度充足環境下，配體交換會遵循質子催化的結合取代(associative-substitution)機構，與先前的文獻計算結果相符；而在質子相對不足的條件下，則進行配位基與鋯氧團簇中既有配體的交換，形成無效的反應。
基於對調節劑協助MOF合成機制的研究基礎，我們希望建立一個專門的調節劑資料庫，以促進對其在MOFs結晶與結構控制中作用的更全面理解。然而目前既有的MOF資料庫多以提供孔洞特性、氣體吸附性或電子性質為導向，作為後續理論計算與模擬的基礎。少部分合成相關資料庫則結合Python程式及其他模型輔助自動化資料擷取，但使用者常面臨程式能力不足、版本衝突與除錯困難等挑戰，且目前語言模型自動化抽取文獻參數的準確率亦有限。
本研究中以人工方式擷取實驗參數，統一單位與配體名稱，將結果輸入MySQL資料庫，並結合網頁介面設計，以供使用者以簡單直觀的方式進行資料的搜尋。
;Metal organic frameworks (MOFs) can be synthesized with the aid of modulators (modulator synthesis) to improve crystallization outcomes. However, when the addition of modulators alters the pH of the reaction, the desired results may not be achieved. The exchange mechanism between linkers and modulators remains unclear.
In this study, we use Zr6O4(OH)4(C4H3O4)11(CHO2) as a model and simulated its energy landscapes under pH 4.6 and pH 5.0 using density functional theory (DFT). Our results show that under proton-rich conditions, linker exchange proceeds via s proton-catalyzed associative-substitution mechanism, which aligns with previous computational studies. In contrast, under proton-deficient conditions, linker exchange occurs between the incoming linker and existing linkers on the oxo-zirconium cluster, resulting in unproductive reactions.
Building upon our investigation on the modulator-assisted MOF synthesis mechanism, we aim to establish a dedicated database of modulators to support a more comprehensive understanding of their role in MOF crystallization and structural control. However, most existing MOF databases focus on pore properties, gas adsorption, or electronic characteristics, which serve as foundations for subsequent theoretical calculations and simulations. A few synthesis-related databases incorporate Python scripts and other models to facilitate automated data extraction, but users often face challenges such as limited programming skills, version conflicts, and debugging difficulties. Moreover, the accuracy of current language models in automatically extracting parameters from literature remains limited.
In this study, we manually extracted experimental parameters, standardized units and linker names, and imported the results into the MySQL database. A web-based interface was also developed, allowing users to search the data in a simple and intuitive manner. |
