A Sustainable Synthetic Approach to the Indaceno[1,2-b:5,6-b′]dithiophene (IDT) Core through Cascade Cyclization–Deprotection Reactions

Abstract

Bulk heterojunction organic solar cells (BHJs) are competitive within the emerging photovoltaic technologies for solar energy conversion because of their unique advantages. Their development has been boosted recently by the introduction of nonfullerene electron acceptors (NFAs), to be used in combination with a polymeric electron donor in the active layer composition. Many of the recent advances in NFAs are attributable to the class of fused-ring electron acceptors (FREAs), which is now predominant, with one of the most notable examples being formed with a fused five-member-ring indaceno[1,2-b:5,6-b′]dithiophene (IDT) core. Here, we propose a novel and more sustainable synthesis for the IDT core. Our approach bypasses tin derivatives needed in the Stille condensation, whose byproducts are toxic and difficult to dispose of, and it makes use of cascade reactions, effectively reducing the number of synthetic steps.

1. Introduction

Among the organic photovoltaic device family, bulk heterojunction solar cells (BHJs) have developed amongst the most promising and efficient emerging technologies for solar energy conversion, because of their flexibility, stretchability, and the possibility to be integrated into many diverse materials [1,2,3]. BHJs usually possess a “sandwich” structure, in which the photoactive layer, made by a blended mixture of an electron donor material, usually a conjugated organic polymer, as the p-type semiconductor, and an electron acceptor small molecule (SMA) as the n-type semiconductor, is encapsulated between two interlayers—an electron-transporting layer (ETL) and a hole-transporting layer (HTL) [4,5]. Recently, the focus on the development of organic solar cells has shifted to nonfullerene electron acceptors (NFAs) as the SMA in combination with the polymeric electron donor in BHJs. In fact, the photoconversion efficiencies (PCEs) of such cells have increased dramatically since 2015, now reaching a high value of over 20% [6]. Ideal NFAs exhibit (a) strong absorption coefficients in regions of the visible and NIR spectrum that are complementary to those in which the available electron-donor polymers absorb, (b) suitably matched energy levels for achieving photoinduced charge separation with donors, (c) the ability to form appropriate morphologies for charge separation and suitable percolation pathways for charge transport, and (d) good molecular and morphological thermal and photostability. A wide variety of material classes have been examined as NFAs [7,8,9,10,11,12,13,14,15]. Many of the recent advances in NFA-containing organic solar cells are attributable to the class of fused-ring electron acceptors (FREAs) [16,17,18,19,20,21,22,23,24,25,26,27,28,29], which is now predominant. FREAs are well-defined organic compounds characterized by a push–pull architecture in which π-extended donating cores having four or more aromatic fused rings are flanked by two electron-withdrawing units (A–D–A or A–D–A–D–A molecular structures, where A and D are electron-withdrawing or electron-donating structures, respectively).

Amongst FREAs, the indacenodithiophene capped with 3-oxo-2,3-dihydro-1H-indene-2,1-diylidene))dimalononitrile (IDIC) family of NFAs stood out in the last five years in terms of performances. The prototypical IDIC structure and its precursors are shown in Figure 1. It is composed of an electron-donating, fused five-member-ring indaceno[1,2-b:5,6-b′]dithiophene (IDT), terminated by two electron-withdrawing units. Several IDIC NFAs have been synthesized by modification of the central electron-rich core, side chains, and end-cap groups. Despite that, the synthesis of the IDT core and its functionalization remain challenging in terms of sustainable chemistry.

We have recently reported innovative methodologies for the synthesis of fused-ring π-extended heterocycles through cascade reactions and sustainable approaches [30,31,32,33,34,35,36]. Here, we propose a novel synthesis for the IDT core, which starts from relatively cheap starting materials and allow us to assemble a fully functionalized IDT core with a reduction in synthetic steps, compared with the synthetic routes reported in the literature. Our approach is innovative and more sustainable, especially considering two main aspects: (a) it avoids the use of tin derivatives, whose byproducts are well known to be toxic, and (b) it makes use of cascade reactions to reduce the number of synthetic steps.

2. Results and Discussion

The two traditional main routes for the synthesis of the indaceno[1,2-b:5,6-b′]dithiophene (IDT) scaffold present in the literature are described in Scheme 1, top. They both require several steps, harsh conditions, and toxic reagents, so that large-scale production, despite the high performance shown in devices, remains unforeseeable. They both involve an initial Stille coupling of a monostannylated thiophene residue. The route developed by Luscombe et al. [37] for the IDT scaffold is followed by a series of steps involving ester hydrolysis, an intramolecular acid-catalyzed cyclization, and a reduction, followed by the alkylation of the newly formed CH₂ positions with long alkyl chains for imparting overall solubility and processability to the final product. The final step, the formylation of the terminal thiophene residues in the α-position, has been reported by other groups [38] with the use of POCl₃/DMF and affords the final IDT core, ready to be fully functionalized via a Knoevenagel reaction with the electron-deficient indanedione-based derivatized scaffold. The route developed by Jen et al. [39] for the IDT scaffold, instead, proceeds through the formation of tertiary alcohol, obtained by the double addition of an organometallic Grignard nucleophile to the ester, which is followed by acid-catalyzed cyclization. Such cascade sequence allows aryl substituents to be directly placed onto the IDT scaffold, able to impart solubility and avoid intermolecular aggregation. Additionally, in this case, the final step involves the formylation of the terminal thiophene residues in the α-position and has been reported by other groups [40] with the use of n-BuLi and DMF to afford the IDT core ready for the final Knoevenagel step. Other approaches have appeared in the literature [41,42,43], which include the use of boron chemistry. The approach presented here inverts the introduction of the formyl groups. Instead of introducing them in a final step, they feature in the thiophene starting materials in the protected form, and the advantage is given by the fact that the thiophene synthon we used is commercially available.

We aimed to achieve a more sustainable and expeditious approach to the IDT core, to further elaborate on the cascade reaction protocol elaborated by Jen et al., and to focus on alternatives to the Stille reaction used in the first step. In fact, we discovered that 5-(1,3-dioxolan-2-yl)-2-thienylzinc bromide is a commercially available starting material, sold at a reasonable price (ca. 35 EUR/g from commercial sources). Such zinc derivative as starting material, bearing an acetal into its molecular structure, allowed us to bypass the conventional route based on Stille reaction in favor of an eco-friendlier Negishi reaction for the synthesis of both diester and dibromo aryl derivatives 1a and 1b (Scheme 1, bottom). Compound 1a is prone to react with an aryl Grignard reagent, and the resulting tertiary alcohol could undergo acid-catalyzed intramolecular cyclization to yield the target IDT core. Moreover, the presence of the acetal ensures, in the final acid-catalyzed cyclization step, its deprotection, to restore the aldehyde functionality without the need for a further, subsequent synthetic step, the final formylation. Compound 1b could react, in an inverted fashion, by generating the nucleophile on the linear scaffold. Attack on a secondary ketone would generate the tertiary alcohol functionality, prone to undergo the acid-catalyzed intramolecular cyclization and deprotection mentioned above.

In the case of derivative 1a, the Negishi protocol afforded the title compound in essentially quantitative yield, after purification of the reaction mixture by column chromatography. In the case of 1b, the reaction was selective toward the functionalization of the iodide aryl substituents, with an overall yield of 70% after purification by column chromatography (Scheme 2, top). Column chromatography had to be carried out on deactivated silica to avoid deprotection of the acetal moieties directly into the column during the separation. The corresponding acetal 1a was tested in the next step, the nucleophilic attack by the Grignard reagent, followed by a cascade intramolecular cyclization–deprotection step to yield the corresponding IDIC core 2a (Scheme 2, bottom).

The cyclization–deprotection step has been extensively studied to find the most suitable conditions, as reported in

Table 1. Cyclization conditions for the synthesis of 2a.

Entry	H+ Source	Conditions	Yield 1
1	HCl/AcOH	Reflux, 3 h	7%
2	H2SO4/AcOH	Reflux, 3 h	7%
3	H2SO4/AcOH	r.t., 16 h	28%
4	SiO2/H2O	r.t., 16 h	-
5	Amberlyst	r.t., 16 h	-
6	BF3·Et2O	0 °C to r.t., 2 h	-

¹ Isolated yield after column chromatography.

. Noteworthy, in the classical conditions reported in the literature (entries 1 and 2), the reaction proceeded with a low yield; lowering the temperature (entry 3) allowed us to obtain derivative 2a in a somewhat improved yet modest yield. We investigated the possibility to promote this reaction using different sources of acids: We tried SiO₂ and amberlyst as proton sources (entries 4 and 5), but after 16 h, we did not observe the formation of derivative 2a. In the case of BF₃·OEt₂, after 2 h, we observed only degradation products.

We ascribe the low yields to the consequence of competitive reaction pathways: deprotection and cyclization (Scheme 3). We speculate that, if the first reaction starting from intermediate 1a* that occurs is deprotection (Path A), the unprotected aldehyde reduces the electron density over the thiophene ring and does not allow the intramolecular cyclization, while when the first reaction is the cyclization (Path B), the cascade reaction is ensured.

Since the deprotection of the acetal functionality requires the stoichiometric addition of water, we performed the reaction in our best conditions (entry 3 in Table 1) in the presence of molecular sieves, with the aim of activating the acetal deprotection in the aqueous workup procedures, only after the successful cyclization had occurred. Unfortunately, yields were not improved in this case either.

In the light of the overall improved sustainability of the OPV devices, solubility and processability of the components in green solvents such as alcohols are desirable. In our approach to developing an alcoholic soluble IDT core, we deprotected molecule 2a to restore the phenolic moiety; the demethylation reaction was carried out in presence of BBr3 to provide the corresponding molecule 3a (Scheme 2) in good yields. The free phenol functionalities could be functional as a solubilizing anchor for the overall core, because of their enhanced polarity and the possibility of the formation of organic salts of tunable solubility in different solvents, by using suitable cationic counterparts.

For acetal 1b, as detailed in Scheme 1, bottom, we sought to generate the corresponding lithium derivative, followed by a reaction with several diaryl ketones. However, despite several attempts, we were never able to isolate the correct cyclized IDT core. As reported for the derivative 2a, competition between deprotection and cyclization could be the reason for such failure, with the former pathway which, when occurring before, switches off the reactivity of the electron-rich thiophene moiety for the latter cyclization. In the case of acetal 1b, since the correct product could not be obtained, pathway A seems to be the only one operating.

We then considered established sustainability indices such as the E factor (kg of organic waste/kg of product) [44]. E factor values for organic semiconductors are often in the excess of 10⁴, in some cases largely surpassing those for organic small molecules, which are active components of pharmaceutical formulations. Although the yield of the final step was far from optimal, E factor calculations (see Supplementary Materials) for our synthesis of molecule 2a gave a value of 201.8. The value calculated on the conventional synthesis reported in the literature, considering the shortest route elaborated by Jen et al. (Scheme 1) on a slightly different IDT core (bearing 4-hexylphenyl aromatic spiro residues), and the formylation carried out on another slightly different IDT core (bearing 4-hexylspiro residues) [40], gave a value of 562.5. We did not include a chromatography solvent in our calculations because it can be recovered by vacuum distillation using a Rotovapor, as reported recently in the literature [45]. Our approach, therefore, allows a reduction by a factor of two of the E factor, and the synthesis is carried out in only three steps, avoiding the further formylation step required in the published approach.

3. Experimental Section

3.1. General Experimental

All commercially available reagents and solvents were purchased from Sigma-Aldrich, Fluorochem, Alfa Aesar. Flash chromatography was carried out using Merck silica gel 60 (pore size 60 Å, 270–400 Mesh). ¹H and ¹³C NMR spectra were recorded from solutions in deuterated solvents on 300 Bruker or 400 Jeol spectrometers with the residual solvent as the internal standard. Mass spectra of pure compounds were recorded using an Electron Spray Ionization Agilent Technologies mass spectrometer, a Direct Exposure Probe mass spectrometer, or a GC–MS ThermoScientific spectrometer.

3.2. Synthesis of New Compounds

2,5-bis(5-(1,3-dioxolan-2-yl)thiophen-2-yl)terephthalate (1a). In a flame dried dry Schlenk tube, under Argon, diethyl 2,5-dibromoterephthalate (1 g, 2.63 mmol), (5-(1,3-dioxolan-2-yl)thiophen-2-yl)zinc bromide (11.5 mL, 5.8 mmol), and Pd(PPh₃)₄ (122 mg, 0.105 mmol) were added in previously degassed dry THF (20 mL, 0.15 M). The resulting black solution was refluxed under stirring. After 16 h, the reaction was stopped, and the resulting solution was extracted three times with DCM/NaHCO₃ solution. The combined organic phases were dried (Na₂SO₄); then, the solvent was removed under reduced pressure to yield a dark solid, which was purified by flash chromatography (basic SiO₂, 7:3 v/v hexanes/EtOAc) to give a yellow solid in quantitative yield. ¹H NMR (300 MHz, CDCl₃): δ = 7.84 (s, 2H), 7.13−7.11 (d, 2H, J = 3.6 Hz), 6.98−6.96 (d, 2H, J = 3.5 Hz), 6.31 (s, 2H), 4.25−3.97 (m, 12H), 1.21−1.14 (t, 6H, J = 14 Hz). ¹³C NMR (101 MHz, CDCl₃): δ = 167.31, 142.91, 141.23, 133.81, 133.26, 131.82, 126.33, 126.22, 107.5, 65.15, 61.17, 13.68. ESI-MS: 553 [M + Na]⁺ (100). HRMS calculated for C₂₆H₂₆O₈S₂: 530.1069; found: 530.1078.

2,2′-((2,5-dibromo-1,4-phenylene)bis(thiophene-5,2-diyl))bis(1,3-dioxolane) (1b). In a flame dried dry Schlenk tube, under Argon, 1,4-dibromo-2,5-diiodobenzene (100 mg, 0.20 mmol), 5-(1,3-dioxolan-2-yl)thiophen-2-yl)zinc bromide (0.9 mL, 0.45 mmol), and Pd(PPh₃)₄ (12 mg, 0.010 mmol) were added in previously degassed dry THF (2 mL, 0.1 M). The resulting black solution was stirred at 60 °C. After 1 day, the reaction was stopped, and the resulting reaction mixture was extracted with DCM/NaHCO₃ solution. The combined organic phases were dried (Na₂SO₄); then, the solvent was removed under reduced pressure.to yield a black solid. The desired compound was purified by flash chromatography (basic SiO₂, 9:1 to 8:2 hexanes/EtOAc) to give a light-yellow solid (110 mg, 73%). ¹H NMR (400 MHz, CDCl₃): δ = 7.77 (s, 1H), 7.25−7.24 (d, 1H, J = 4Hz), 7.17−7.16 (d, 1H, J = 4Hz), 6.13 (s, 1H), 4.21−4.00 (m, 4H). ¹³C NMR (101 MHz, CDCl₃): δ = 143.33, 140.54, 136.01, 135.74, 128.11, 126.25, 121.16, 100.21, 65.35, 65.28. ESI-MS: 1111 [2M + Na]⁺ (100). HRMS calculated for C₂₀H₁₆Br₂O₄S₂: 541.8857; found: 541.8853.

4,4,9,9-tetrakis(4-methoxyphenyl)-4,9-dihydro-s-indaceno[1,2-b:5,6-b′]dithiophene-2,7-dicarbaldehyde (2a). In a flame dried dry Schlenk tube, under Argon, a light-yellow solution of 1a (246 mg, 0.46 mmol) in dry THF (3 mL, 0.15 M), the resulting solution was cooled to 0 °C; then 4-methoxyphenylmagnesium bromide (4.10 mL, 2.05 mmol) was added dropwise under stirring. The solution immediately changed its color from light yellow to red. The resulting red solution was allowed to reach r.t., under stirring. After 16 h, the reaction was stopped, and the resulting red solution was quenched with a NaHCO₃ solution and then extracted with DCM/H₂O to yield an orange solid. The resulting orange solid was dissolved at r.t. in AcOH and a catalytic amount of conc. H₂SO₄ was added, and the resulting solution immediately changed its color from orange to violet. After 16 h under stirring at r.t., the resulting violet solution was quenched with an aqueous solution of NaHCO₃ and extracted three times with DCM/H₂O. The combined organic phases were dried (Na₂SO₄), and the solvent was removed under reduced pressure to yield an orange solid. The resulting orange solid was purified by flash chromatography (7:3 v/v hexanes/EtOAc) to give compound 2a as a yellow solid (95 mg, 28% yield). ¹H NMR (300 MHz, CDCl₃): δ = 9.87 (s, 2H), 7.63 (s, 2H), 7.58 (s, 2H), 7.18−7.14 (d, 8H, J = 8.8 Hz), 6.84−6.80 (d, 8H, J = 8.8 Hz), 3.78 (s, 12H) ppm. ¹³C NMR (75 MHz, CDCl₃) δ = 182.69, 158.77, 157.14, 155.38, 149.98, 146.44, 135.70, 135.23, 131.57, 128.71, 118.83, 113.92, 62.14, 55.17. ESI-MS: 747.33 [M + 1]⁺ (80). HRMS calculated for C₄₆H₃₆O₆S₂: 748.1953; found: 748.1955.

Synthesis of 4,4,9,9-tetrakis(4-hydroxyphenyl)-4,9-dihydro-s-indaceno[1,2-b:5,6-b′]dithiophene-2,7-dicarbaldehyde (3a). In a flame-dried, dry Schlenk tube, under Argon, a solution of 2a (95 mg, 0.127 mmol) in dry DCM (1.2 mL, 0.1 M) was cooled to 0 °C; then, BBr₃ (0.980 mmol, 170 µL) was added dropwise under stirring. The solution immediately changed its color from light yellow to dark red. The resulting dark red solution was allowed to reach r.t. under stirring. After 2 h, the reaction was stopped, and the resulting dark blue solution was extracted three times with EtOAc/H₂O. The combined organic phases were dried (Na₂SO₄), and the solvent was removed under reduced pressure.to yield a dark red solid. The resulting dark red solid was purified by precipitation in hexanes from THF to give 3a as a red solid compound (44 mg, 50%). ¹H NMR (300 MHz, acetone-d₆): δ = 9.91 (s, 2H), 8.35 (s, 4H), 7.96 (s, 2H), 7.90 (s, 2H), 7.15−7.11 (d, 8H, J = 8.5 Hz), 6.77−6.73 (d, 8H, J = 8.6 Hz). ¹³C NMR (75 MHz, acetone-d₆): δ = 184.41, 157.91, 148.18, 135.78, 133.48, 130.17, 119.79, 118.69, 116.55, 55.32. HRMS calculated for C₄₂H₂₈O₆S₂: 692.1171; found: 692.1165.

4. Conclusions

We successfully developed a novel synthesis for the IDT core, which can be carried out from relatively cheap starting materials and allow us to assemble a fully functionalized indaceno[1,2-b:5,6-b′]dithiophene scaffold in only three synthetic steps. Our approach is more sustainable than current approaches reported in the literature, as testified by the reduction in a commonly used green matric parameter, the E factor, by a factor of two. Two other aspects must be stressed as evidence pointing to the decisive improvement of our procedure over existing ones: (a) our synthesis avoids the use of Stille condensation and related tin derivatives, whose byproducts are toxic and difficult to dispose of, and (b) it makes use of cascade reactions to reduce the number of synthetic steps. Further studies will focus on the possible improvement of the key cyclization–deprotection strategy, and on the possibility to use the phenolic scaffold 3a to produce novel NFAs, which could be soluble and processable in green polar solvents such as alcohols.