Introduction

Energy issue has raised increasing attention over the past decades. Since fossil fuels are expected to be used up within 21^th century, several types of alternative energy, including solar energy, biofuel, wind power, tidal power, and geothermal energy, are currently under intensive investigation. It is beautiful to extract electricity directly from the sun, which is clean, safe, and crucial for our lives. Thus, different kinds of photovoltaic devices (PVs) have emerged and become a prominent subject in the past few decades.¹ Silicon-based solar cells, because of their well-studied properties, long stability, and high power conversion efficiency (PCE), have been widely used for generating electricity. Despite several advantages of silicon-based solar cells, there are some drawbacks, like high production cost and related environment impact, which have impelled us to develop new types of PVs continuously.¹ Among all, organic solar cells (OSCs) are one of the possible solutions.

The fascinating features of OSCs are the potential of large-area and roll-to-roll fabrication, light-weight and flexible devices, lower cost, and low impact to environment. Nowadays, different kinds of OSCs have been investigated. The best-known designs of device are dye-sensitized solar cells (DSSCs) and organic thin-film solar cells. DSSCs have been well studied and the maxima PCEs are steadily over 10%.² Organic thin-film solar cells comprise an electron acceptor, which is usually fullerene derivatives, and an organic electron donor. When an electron donor absorbs a photon, an electron in the highest occupied molecular orbital (HOMO) is excited to the lowest unoccupied molecular orbital (LUMO). The photogenerated exciton diffuses within the electron donor towards the interface to the electron acceptor. If the energy level difference between the electron donor and the electron

acceptor is larger than the exciton binding energy, usually 0.3–1.0 eV,³ charge transfer from the donor molecule to the acceptor molecule occurs. The electron is transferred to the acceptor molecule via the energetically favorable driving force, resulting in a polaron pair. After dissociation, the hole stays and transports on the electron donor, which is thus called p-type material, whereas the electron travels on the electron acceptor, thus called n-type material. The polarons can transport to electrodes and deliver photocurrent. The donor-acceptor heterojunction plays an important role in device performance. Since the diffusion length of excitons in organic electronic materials is typically within 20 nm,⁴ large interface area is essential for the excitons to reach the heterojunction interface. In bulk heterojunction cells (BHJ), an electron donor and an electron acceptor are cast simultaneously. Hence, two components are allowed to interpenetrate, creating huge extension of interface area. Consequently, BHJ is the most widely used architecture for organic thin-film solar cells.^3-4

An electron donor should capture photons from solar radiation efficiently to provide high photocurrent. The most intense irradiance region of solar spans 400–700 nm at Earth's surface (Figure 1-1).⁵ To gain power conversion efficiency, the optical absorption of an electron donor should span visible range and extend to near-infrared region. Luckily, powerful synthetic methods have been developed to invent electron donors with appropriate properties.

Figure 1-1. The spectrum of solar radiation at Earth's surface.

Judicious design of electron donors is crucial for controlling their bandgaps and thus utilizing solar radiation. As shown in Figure 1-2, four polyaromatic conjugated polymers, namely polyphenylene, poly(phenylenevinylene), polythiophene, and polyisothianaphthene, are chose to illustrate the relationship between aromaticity and bandgap.⁶ In ground state, all polymers exhibit two forms with nondegenerate energy:

aromatic and quinoid form. In aromatic form, the π-electrons are confined in each ring to gain resonance stabilized energy, resulting in extra stability. In quinoid form, however, the π-electrons are delocalized and the aromaticity is destroyed, which is thus energetically less stable than aromatic form. Consequently, there is more population in aromatic form for the first three polymers. The large bandgap (3.2 eV) of polyphenylene results from the high resonance stabilized energy of benzene (36 kcal mol^-1), whereas the relatively low aromaticity of thiophene (29 kcal mol^-1)⁷ gives a small bandgap (2.0 eV) for polythiophene. For poly(phenylenevinylene), insertion of double bonds can dilute the overall aromaticity and thus reduce the bandgap (2.5 eV). To increase the quinoid population effectively, fused rings are adopted to maintain the aromaticity in quinoid form. The higher resonance stabilized energy of benzene than thiophene lets polyisothianaphthene tend to favor quinoid form to keep the aromaticity of benzene. To sum up, higher aromaticity components result in larger bandgap; however, the aromaticity can be utilized in fused rings to increase the quinoid population. In this thesis, an electron-withdrawing [2,1,3]benzothiadiazole moiety with high quinoidal character was utilized in conjugated backbone to facilitate π-electron delocalization.⁸

Figure 1-2. Aromatic and quinoid resonance forms for some conjugated polymers and their relative population in ground state.

After fabrication into photovoltaic devices, there are three key parameters to determine the power conversion efficiency of a PV: the open-circuit voltage (Voc), short-circuit current density (Jsc), and fill factor (FF). The fill factor is defined as the following equation: FF =^{VMPP ∗ JMPP}

V_oc ∗ J_sc , where MPP is the abbreviation of the maximum power point. As a result, power conversion efficiency can be written as follows: PCE =^{Voc ∗ Jsc ∗ FF}

P_in , where Pin denotes the input power. In order to obtain high Jsc, PV devices should possess large shunt resistance (Rsh) and small series resistance (Rs).⁹ On the other hand, since Voc is mainly determined by the energy difference between the HOMO of an electron donor and the LUMO of an electron acceptor,¹⁰ lower HOMO of an electron donor may realize higher Voc. FF implies morphology in the active layer, and the possibility the photogenerated carriers can be extracted.

In bulk heterojunction cells, electron donors can be roughly classified into two categories: small molecule and polymer. Small molecule organic solar cells

(SMOSCs), because of their well-defined structure and batch-to-batch reproducibility compared to polymer solar cells (PSCs), are attracting more attention. Several small molecules were employed as electron donors and achieved PCE over 8% (Scheme 1-1). DR3TSBDT, featuring benzodithiophene (BDT) as the electron-donating unit for solution-processed BHJ, showed a certified PCE of 9.9%.¹¹ From our previous research, vacuum-deposited DTDCPB solar cell performed a maximum PCE of 8.2%, which is one of the highest PCE reported for asymmetric dipolar electron donors.¹² A cascade architecture non-fullerene OSC comprising an electron donor α-6T and two electron acceptors, SubPc and SubNc, reached a remarkable PCE of 8.4%.¹³ Furthermore, a tandem OSC comprising DTTz for visible absorption and DTDCTB for near-infrared absorption could achieve a high PCE of 9.2%.¹⁴ Such results indicate that SMOSCs can yield high PCE and are worth further investigation.

Scheme 1-1. Small molecule electron donors for OSCs achieving PCE over 8%.

Recently, there are extensive studies of the introduction of heteroatoms on electron donors, one of which is a fluorine atom. Some representative chemical structures and device performance of fluorinated electron donors are shown in Scheme 1-2. Dithienosilole (DTS) centered p-DTS(FBTTh2)2,¹⁵ connecting with fluorobenzothiodiazole (FBT) as electron withdrawing group (EWG), performed a maximum PCE of 9.0%,¹⁶ which is prominently higher than the PCE 0.2% of the non-fluorinated counterpart 4.¹⁷ BIT-4F-T, comprising difluorobenzothiadiazole (dFBT) as EWG, delivered a maximum PCE of 8.1%.¹⁸ Polymer BHJ utilizing PTB7 as an electron donor, which features a fluorinated-thienothiophene moiety, could obtain an impressive PCE of 9.2%.¹⁹ All-polymer BHJ, utilizing both fluorinated electron donor PBDTT-TT-F and electron acceptor P(NDIDT-FT2), yielded a maximum PCE of 6.7%,²⁰ which is among the highest PCE reported all-polymer solar cells. Based on these investigations, fluorinated EWGs are promising components to access high power conversion efficiency for OSCs.

Scheme 1-2. Fluorinated electron donors for OSCs achieving PCE over 6%.

Many studies revealed the relationship between fluorination and energy level.

Some representative series of fluorinated electron donors are illustrated in Scheme 1-3 and summarized in Table 1-1. As the introduction of fluorine atom(s), both HOMO and LUMO energy levels were lowered for all electron donors in three conditions: (1) the decrement of HOMO is larger than that of LUMO, resulting in larger bandgap, like X1 series²¹ and PBT-0F series.²² (2) The decrement of HOMO is equal to that of LUMO, resulting in equal bandgap, like HH series²³ and p-DTS(FBTTh2)2 series.^15,

17 (3) The decrement of HOMO is smaller than that of LUMO, resulting in smaller bandgap, like PBnDT-HTAZ series.²⁴ In all conditions, the lowered HOMO can contribute to high Voc. Furthermore, in the third condition, smaller bandgap may lead to bathochromic absorption, which is beneficial for pursuing higher Jsc. However, the third condition is generally less reported.

Scheme 1-3. Chemical structures of several fluorinated electron donors.

Table 1-1. Electrochemical parameters and device performance of several fluorinated both HOMO and LUMO, such conclusion is typically based on polymers. There are fewer literatures about small molecules. Most of them are based on centrosymmetric configurations, e.g. D2-A1-D1-A1-D2 or D2-A2-D1-A1-D1-A2-D2 architecture. Research based on asymmetric configuration, such as D-A-A type, is relatively rare. This thesis provided examples for the second and the third condition based on D-A-A small molecules.

From our previous study of electron donors for SMOSCs, DTCTB and DTCPB, in which two electron-donating ditolylaminophenyl and ditolylaminothienyl moieties were respectively connected to an electron-withdrawing cyano group through another

electron-withdrawing [2,1,3]benzothiadiazole moiety, performed a maximum PCE of 4.4% and 6.4% on C70-based devices (Scheme 1-4). It is noteworthy that the fill factor is very high, especially for DTCPB (FF = 0.65). Organic photovoltaic (OPV) devices with high FF have a high Rsh and a low Rs, suggesting the morphology after blended with fullerenes is very beneficial for reducing the internal loss of current produced by the devices. Encouraged by the promising results, we decide to introduce a fluorine atom onto BT moieties of DTCTB and DTCPB as a pendant group to alter energy levels. Since the atomic weight is 19 Da for a fluorine atom, the change of molecular weight (MW) is small for a D-A-A molecule with MW 400–500 Da. We expect that replacing a hydrogen atom with a fluorine atom will make little change to the benign morphology while providing additional interactions or steric hindrance.

Scheme 1-4. Chemical structures and performance of DTCTB and DTCPB.

In this thesis, four donor-acceptor-acceptor (D-A-A) structured electron donors were synthesized and characterized, namely DTCTiFBT, DTCToFBT, DTCPiFBT, and DTCPoFBT (Scheme 1-5). The difference in aromaticity of benzene and thiophene along with the different position of the inserted fluorine atom can lead to different performances and meaningful discussions. Their corresponding BHJ devices and solid film properties were fabricated and measured by Prof. Jiun-Haw Lee and Mr.

Yi-Ze Hsiao at the Graduate Institute of Photonics and Optoelectronics, National Taiwan University. For comparison, this thesis also included the non-fluorinated

counterparts, DTCTB and DTCPB, synthesized and characterized by Dr. Hao-Chun Ting and Mr. Chia-Hsun Chen. The density functional theory (DFT) and time-dependent density functional theory (TD-DFT) theoretical calculations of six molecules were performed at CAM-B3LYP/6-311G(d,p) level by Dr. Shu-Hua Chou.

Scheme 1-5. Four D-A-A electron donors synthesized and characterized in this thesis.