染料敏化太陽能電池(Dye-Sensitized Solar Cells， DSC)因具備透光性、可撓曲性及色彩多樣化等優點，極具應用潛力，但DSC元件在組裝完成後顏色無法再度調整，限制了其在建築上的應用。因此本研究的目標是合成具光致變色的有機染料並將其應用於DSC。目前以光致變色染料NPI作為敏化劑之DSC元件具有最高光電轉換效率，NPI結構為典型的D-π-A (Donor–π-bridge–Accepter)結構，以二苯胺為推電子基經由Naphthopyran變色單元(作為π-bridge)連接氰乙酸拉電子基而成。然而，其吸收範圍窄(半高寬約130 nm)，為擴展其吸收範圍以提高DSC之效率，因此本研究提出兩種方案。第一個方案是於NPI結構中引入Benzothiadiazole (BT) 單元作為輔助拉電子基(Auxiliary acceptor， Aa)形成D-Aa-π-A (Donor–Auxiliary acceptor–π-bridge–Accepter)結構，合成出有機染料DBT-NP；第二個方案是將NPI錨定基團前的苯環替換為Benzothiadiazole (BT) ，使染料的LUMO分布集中於拉電子區，合成出有機染料DNP-BT。兩染料於甲苯中在未照光時的吸收光譜，皆表現π→π*與分子內電荷轉移(ICT)兩個主要吸收峰，DBT-NP 溶液呈亮橘色(兩吸收峰的λmax分別位於377 nm與487 nm)；DNP-BT 溶液呈淺黃色(兩吸收峰的λmax則位於312 nm與395 nm)。在AM 1.5G模擬太陽光(強度100 mW/cm2)照射 20 分鐘後，DBT-NP 轉為暗橘色，原始吸收峰強度隨照光時間遞增且半高寬增加(144 nm→154 nm)，並於602 nm產生微弱吸收峰；DNP-BT 則轉為深綠色，且原始吸收峰強度隨照光時間遞減，並於 617 nm 產生明顯新峰，半高寬約148 nm。將兩染料停止照光再置於暗處20~25分鐘後皆可回復初始狀態，再經過5次循環測試，確認了DBT-NP、DNP-BT於溶液中的光致變色具有不錯的可逆性。由光致變色動力學與能隙分析證實，因DNP-BT相較於DBT-NP能產生更強 ICT 效應，使其褪色速度較慢，表示其開環結構較穩定；此外，前置軌域分佈圖顯示，DBT-NP的LUMO軌域被侷限於輔助拉電子基(BT)上，限制了電子傳遞，而DNP-BT的LUMO更多貢獻在拉電子基上，有效促進電子傳遞至拉電子基再注入至TiO2電極。最後，DNP-BT敏化之DSC元件的最高光電轉換效率為1.33%，優於DBT-NP所敏化元件的1.21%。;Dye-Sensitized Solar Cells (DSCs) have great application potential due to advantages such as transparency， flexibility， and color diversity. However， the color of DSC devices cannot be adjusted once they are assembled， which limits their applications in architecture. Therefore， the goal of this study is to synthesize photochromic organic dyes and apply them to DSCs. Currently， DSC devices using NPI as a sensitizer exhibit the highest power conversion efficiency amongst the photochromic dyes used. NPI features a typical D-π-A (Donor–π-bridge–Acceptor) structure， consisting of a diphenylamine electron donor connected to a cyanoacetic acid electron acceptor via a naphthopyran photochromic unit (acting as the π-bridge). However， its absorption band is narrow (with a full width at half maximum， FWHM， of ca. 130 nm). To broaden its absorption range and improve DSC efficiency， this study proposes two approaches. The first approach introduces a benzothiadiazole (BT) unit into the NPI structure as an auxiliary acceptor (Aa) to form a D-Aa-π-A (Donor–Auxiliary acceptor–π-bridge–Acceptor) configuration， resulting in the synthesis of DBT-NP. The second approach replaces the benzene ring in front of the anchoring group of NPI with a BT unit， allowing the dye′s LUMO to mainly distribute on the anchoring group， resulting in the synthesis of DNP-BT. In toluene solutions， the absorption spectra of both as-prepared dyes exhibit two main absorption peaks corresponding to π→π* and intramolecular charge transfer (ICT) transitions. The DBT-NP solution appears bright orange (the λmax of the two peaks are located at 377 nm and 487 nm， respectively); the DNP-BT solution is light yellow (the λmax of the two peaks are at 312 nm and 395 nm). After 20 minutes of irradiation under AM 1.5G simulated sunlight (intensity of 100 mW/cm²)， DBT-NP turns dark orange; the intensity of its original absorption peak increases with irradiation time， its FWHM broadens (144 nm to 154 nm)， and a weak absorption peak emerges at 602 nm. Conversely， DNP-BT turns dark green; the intensity of its original absorption peak decreases with irradiation time， and a distinct new peak appears at 617 nm with an FWHM of about 148 nm. When the irradiation is stopped and both dyes are placed in the dark for 20 to 25 minutes， they can both return to their initial states. After 5 cycle tests， it was confirmed that both DBT-NP and DNP-BT exhibit good photochromic reversibility in solution. DNP-BT exhibits a slower fading speed than DBT-NP， indicating that its ring-opened structure is more stable. Frontier orbital distribution maps show that the LUMO of DBT-NP is localized on the auxiliary acceptor (BT)， restricting electron transfer. In contrast， the LUMO of DNP-BT exhibits more contribution from the electron acceptor， which effectively promotes electron transfer to the acceptor and subsequent injection into the TiO₂ electrode. Finally， the maximum power conversion efficiency of the DSC device sensitized by DNP-BT is 1.33%， outperforming the 1.21% efficiency of the device sensitized by DBT-NP.
