Abstract
Photoredox catalysis is a rapidly evolving platform for synthetic methods development. The prominent use of acridinium salts as a sustainable option for photoredox catalysts has driven the development of more robust and synthetically useful versions based on this scaffold. However, more complicated syntheses, increased cost, and limited commercial availability have hindered the adoption of these catalysts by the greater synthetic community. By utilizing the direct conversion of a xanthylium salt into the corresponding acridinium as the key transformation, we present an efficient and scalable preparation of the most synthetically useful acridinium reported to date. This divergent strategy also enabled the preparation of a suite of novel acridinium dyes, allowing for a systematic investigation of substitution effects on their photophysical properties.
Keywords: photoredox, catalysis, synthesis, acridinium, organocatalyst
Graphical Abstract
Over the past decade, photoredox catalysis has rapidly emerged as a platform for synthetic methods development. Through photoinduced electron transfer (PET), this chemistry allows for the generation of highly reactive intermediates under operationally mild conditions, and typically utilizes low-energy visible light.1 Owing to their long excited-state lifetimes and versatile redox potentials, the polypyridyl complexes of ruthenium and iridium currently serve as the flagship catalysts for PET-based catalysis.1a–c Despite their utility, high cost and the requirement for precious metals engender doubt about their sustainability. Organic dyes, such as the acridinium salt first introduced by Fukuzumi (1, Scheme 1A)2 and popularized as a photoredox catalyst by our group,3 provide an attractive alternative. As acridinium-based photoredox catalysis matured, structural modifications to the catalyst’s core have led to improvements in its stability and photophysical properties. In particular, N-arylation and the introduction of tert-butyl substituents were implemented to prevent catalyst bleaching by dealkylation and nucleophilic addition, respectively.1e,4 However, a more complicated synthesis, concomitant increase in cost,4 and limited commercial availability have obstructed the widespread adoption of catalyst 2 in the greater synthetic community.
Scheme 1.
Synthesis of acridinium photoredox catalysts
Acridinium salts have traditionally been prepared by the addition of an organometallic nucleophile to an acridone.5 Protonation of the resulting tertiary alcohol results in rapid dehydration, furnishing the aromatic acridinium core. In the case of 2, construction of the di-tert-butyl acridone precursor requires several challenging steps involving long reaction times, difficult purifications, and expensive starting materials. Yields are also inconsistent in practice.
Our group previously developed several new electron-rich acridinium photocatalysts with attenuated excited-state reduction potentials.3 The synthetic approach to these compounds exploited symmetrical triarylamine precursors that can engage in a Friedel-Crafts reaction with a benzoyl chloride derivative, delivering the acridinium core (Scheme 1B). Inspired by the efficiency of this strategy, we sought to develop an improved synthesis of our more highly oxidizing catalyst 2 by using a similar disconnection. In the midst of our efforts, Sparr and co-workers disclosed an efficient preparation of electron-rich acridinium salts via the addition of 1,5-bifunctional nucleophiles to aromatic esters (Scheme 1B).6 However, this method cannot be directly applied to the synthesis of 2 because the preparation of an organomagnesium precursor would not provide an advantage over the original route, and the required selectivity for a double directed ortho-metalation of a triarylamine is not feasible. Instead, we elected to simplify the starting material to a biaryl ether to access the corresponding xanthylium salt 3 (Scheme 1C). We envisioned then converting the latter directly into 2 by condensation with aniline (Scheme 1C), in a reaction analogous to the well-developed transformation of pyrylium into pyridinium salts.7 These transformations are scarce in the literature owing to the multiple electrophilic sites on the xanthylium core.8 However, the requirement for sterically demanding tert-butyl and mesityl substituents on 3 prompted us to evaluate this chemistry.
We sought to prepare xanthylium 3 by the double directed ortho-lithiation of biaryl ether 4 and addition of the resulting 1,5-bifunctional nucleophile to a 2,4,6-trimethyl benzoate ester (Equation 1). We postulated that the desired lithiation regioselectivity would be sterically dictated by the presence of the tert-butyl groups. Adapting conditions originally reported by Ogle and co-workers,9 double ortho-lithiation using n-butyllithium and TMEDA resulted in only moderate yields of 3, in addition to various byproducts likely arising from incomplete and/or nonselective lithiation. The use of sec-butyllithium drastically improved the yields of 3 and delivered much cleaner material. In parallel, we investigated the conversion of 3 to the desired acridinium photocatalyst 2.
We began by evaluating conditions that are commonly used for the transformation of pyrylium into pyridinium salts, which involves treatment of the pyrylium with a primary amine in refluxing ethanol.7a With the inclusion of catalytic amounts of acetic acid, we were able to achieve conversions of up to 79%; however, the reaction could not be driven to completion. This was problematic because the starting material and product are difficult to separate. A mechanistic study of the pyrylium into pyridinium transformation by Katritzky et al. suggests the initial addition of the amine and ring opening of the pyrylium are facilitated by base, and the subsequent ring closure is facilitated by acid.7b As a result, sequential additions of triethylamine followed by acetic acid allowed this transformation to proceed rapidly at room temperature when dichloromethane was employed as the solvent. By using this procedure, we attained rapid and near quantitative conversion of 3 into acridinium 2. Optimization revealed that a slight excess of aniline resulted in the complete consumption of 3, allowing for the isolation of higly pure 2 in excellent yield (Equation 2). Furthermore, the order of addition was found to be inconsequential in terms of yield and reaction time, and that the simultaneous addition of triethylamine and acetic acid resulted in a cleaner reaction mixture.
Having optimized the synthetic sequence for the preparation of 2, we next evaluated its performance on a gram-scale. To this end, we were able to produce more than 10.5 g of material in a single pass with a minimal loss in efficiency.10 Furthermore, the entire sequence could be completed in roughly three days by using readily available and affordable starting materials, representing a substantial improvement over the original synthetic route.
Recognizing the modularity of this divergent synthetic strategy, we then set out to prepare a suite of novel acridinium salts and evaluate their photophysical properties. To perform a systematic evaluation, we elected to study three structural variables individually: the aryl substitution in the 9-position, the acridinium core structure, and substitution on the acridinium nitrogen (Figure 1). Accordingly, several xanthylium salts (5-11) with differing cores and 9-substitution were prepared for transformation into their acridinium counterparts (Scheme 2).
Figure 1.
Structural design elements of new photocatalysts
Scheme 2.
Preparation of xanthylium salts
The mesityl acridinium salt 1 was originally designed to be an electron donor-acceptor molecule that, upon excitation, forms a charge-transfer (CT) state in which the positive charge is localized on the mesityl ring.2a This structural feature was carried over when the dye was co-opted as a photoredox catalyst. However, being less oxidizing than the locally excited (LE) state11 and more prone to decomposition,12 this CT phenomenon can be counterproductive in photoredox reactions. Our group has demonstrated that the incorporation of a 2,6-dimethylphenyl (xylyl) group, being slightly less electron-rich, inhibits formation of the CT state.11 As such, the xylyl group was selected for catalysts in which the 9-substitution was held constant. Indeed, conversion of mesityl acridinium (2) into its xylyl analogue (12) resulted in an increased excited-state lifetime (Scheme 3). The extended lifetime may be partially attributed to attenuation of a nonradiative decay pathway associated with C-C bond rotation of the additional methyl group in 2. However, acridinium salts 13 and 14, which possess substitution in that position, exhibit fluorescence lifetimes that are comparable to that of 12, suggesting that inhibition of a CT state is likely the predominating contributor to the extended lifetime.
Scheme 3.
Scope and photophysical properties of acridinium photocatalysts. a Excited-state reduction potentials were estimated from the ground-state reduction potentials and excited-state energies (E0,0). b Ground-state reduction potentials were determined by cyclic voltammetry (reported vs. SCE). c Excited-state energies were determined from the point of intersection of the normalized absorption and emission spectra. d Fluorescence lifetimes were determined from time-correlated single-photon counting (see the Supporting Information).
In addition to studying substitution effects at the 9-position, we investigated the effects of modulating the core structure of the acridinium. Unsurprisingly, extending the π-system, as in benzannulated acridiniums 17 and 18, produced redshifted absorbance profiles and slightly diminished excited-state reduction potentials. The 2,7-di-tert-butyl acridinium 16, however, possesses similar redox properties to that of its isomeric counterpart 12, but exhibits a significantly longer excited-state lifetime. This scaffold may prove useful in photoredox catalysis as it can be synthesized from cheaper starting materials than 12, and preparation of the requisite xanthylium salt would not suffer from regioselectivity issues.
We selected a suite of electronically and sterically differentiated anilines to explore the effects of various N-substitutions on the 3,6-di-tert-butyl acridinium scaffold. Reaction yields were diminished when sterically encumbered or electron-deficient anilines were used, but still delivered sufficient material to evaluate their properties. Although varying the N-aryl substitution does not significantly influence excited-state reduction potentials, fluorescence lifetimes are increased with the incorporation of electron-withdrawing groups. Notably, the trifluoromethyl-substituted acridiniums 22 and 23 both exhibit lifetime measurements greater than 20 ns. N-Xylyl acridinium 25 also possesses a lifetime in this regime, although it is likely a result of more hindered rotation about the C-N bond, tempering nonradiative decay. Of the lifetimes reported in Scheme 3, N-benzyl acridinium 26 exhibits the longest at nearly 24 ns, but its utility in certain photoredox reactions is questionable because N-alkyl acridinium salts are readily dealkylated.1e
Finally, having identified some trends for structural effects on photophysical properties, we sought to design an acridinium that will possess a further enhanced fluorescence lifetime. We elected to synthesize 27 (Figure 2), which encompasses the best-in-class structural modifications from each category highlighted in Figure 1: a 2,6-dichlorophenyl group at the 9-position, a 2,7-di-tert-butyl substituted acridinium core, and benzyl-substituted nitrogen atom. The substitution effects appear to be additive, eliciting an excited-state lifetime exceeding 25 ns, which is longer than any of the acridinium salts studied in Scheme 3.
Figure 2.
Custom acridinium salt designed to possess an enhanced excited-state lifetime
In conclusion, an efficient reaction manifold has been developed to access our most stable and potent acridinium photocatalyst (2). This new approach is a marked improvement over the previous generation synthesis in terms of step count, time, overall yield, scalability, and cost of starting materials.13 We have also applied this versatile strategy to the synthesis of a library of novel acridinium salts and measured their photophysical properties. The insights attained from this study are guiding ongoing efforts in catalyst design aiming to strike an optimal balance between cost, ease of preparation, stability, and photophysical properties.
Supplementary Material
Funding Information
This project was supported by Award No. R01 GM098340 from the National Institute of General Medical Sciences and a Camille Dreyfus Teacher-Scholar Award (D.A.N.). L.W. was supported by the International Postdoctoral Exchange Fellowship Program. Photophysical measurements were performed in the UNC-ERFC Instrumentation Facility established by the UNC-EFRC (Center for Solar Fuels, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Award Number DE-SC0001011).
Footnotes
Supporting Information
Supporting information for this article is available online at https://doi.org/10.1055/s-0037-1611744.
References and Notes
- (1). For recent reviews, see:; a) Prier CK; Rankic DA; MacMillan DWC Chem. Rev 2013, 113, 5322. [DOI] [PMC free article] [PubMed] [Google Scholar]; (b) Narayanam JMR; Stephenson CR J. Chem. Soc. Rev 2011, 40, 102. [DOI] [PubMed] [Google Scholar]; (c) Yoon TP; Ischay MA; Du J Nat. Chem 2010, 2, 527. [DOI] [PubMed] [Google Scholar]; (d) Nicewicz DA; Nguyen TM ACS Catal. 2014, 4, 355. [Google Scholar]; (e) Romero NA; Nicewicz DA Chem. Rev 2016, 116, 10075. [DOI] [PubMed] [Google Scholar]; (f) Shaw MH; Twilton J; MacMillan DW C. J. Org. Chem 2016, 81, 6898. [DOI] [PMC free article] [PubMed] [Google Scholar]; (g) Wang C-S; Dixneuf PH; Soulé J-F Chem. Rev 2018, 118, 7532. [DOI] [PubMed] [Google Scholar]
- (2).(a) Fukuzumi S; Kotani H; Ohkubo K; Ogo S; Tkachenko NV; Lemmetylnen HJ Am. Chem. Soc 2004, 126, 1600. [DOI] [PubMed] [Google Scholar]; (b) Kotani H; Ohkubo K; Fukuzumi SJ Am. Chem. Soc 2004, 126, 15999. [DOI] [PubMed] [Google Scholar]
- (3).Joshi-Pangu A; Lévesque F; Roth HG; Oliver SF; Campeau L-C; Nicewicz D; DiRocco DA J. Org. Chem 2016, 81, 7244. [DOI] [PubMed] [Google Scholar]
- (4). Acridinium 2 can be purchased for US $871/g from Sigma-Aldrich at the time of this publication.
- (5).Romero NA; Margrey KA; Tay NE; Nicewicz DA Science 2015, 349, 1326. [DOI] [PubMed] [Google Scholar]
- (6).(a) Fischer C; Sparr C Angew. Chem. Int. Ed 2018, 57, 2436. [DOI] [PubMed] [Google Scholar]; (b) Fischer C; Sparr C Tetrahedron 2018, 74, 5486. [Google Scholar]
- (7).(a) Katritzky AR; Brownlee RTC; Musumarra G Tetrahedron 1980, 36, 1643. [Google Scholar]; (b) Katritzky AR; Manzo RH; Lloyd JM; Patel RC Angew. Chem. Int. Ed. Engl 1980, 19, 306. [Google Scholar]
- (8).Wu D; Feng X; Takase M; Haberecht MC; Müllen K Tetrahedron 2008, 64, 11379. [Google Scholar]
- (9).Birman VB; Chopra A; Ogle CA Tetrahedron Lett. 1996, 37, 5073. [Google Scholar]
- (10). Gram-Scale Preparation of Acridinium Salts; Typical Procedure for Acridinium 2. 3,6-Di-tert-butyl-9-mesitylxanthylium Tetrafluoroborate (3) To a flame-dried 250 mL round-bottom flask under argon were added 4 (8.02 g, 28.4 mmol, 1 equiv), TMEDA (8.71 mL, 58.2 mmol, 2.05 equiv), and anhydrous n-hexane (28 mL). The resulting solution was cooled in an ice bath and sec-butyllithium (1.4 M solution in cyclohexane, 42.0 mL, 58.2 mmol, 2.05 equiv) was added dropwise. The ice bath was removed and the reaction mixture was stirred at room temperature for 4 h. The reaction was cooled to −78 °C and a solution of methyl 2,4,6-trimethylbenzoate (5.11 g, 28.7 mmol, 1.01 equiv) in anhydrous n-hexane (28 mL) was added slowly via cannula. After the addition, the mixture was allowed to slowly warm to room temperature and stirred for 12 h. The reaction was quenched with water (25 mL) and the biphasic mixture was stirred vigorously for 30 min. The mixture was diluted with Et2O (100 mL) and the layers were separated. The organic layer was washed with water (2 × 150 mL) and brine (1 × 150 mL). The organic layer was transferred to a 250 mL round-bottom flask equipped with a stir bar. To the vigorously stirred solution was added conc. HCl (12 mL), resulting in a bright-yellow precipitate. The suspension was stirred vigorously for 30 min then diluted with water (150 mL). The layers were separated and the organic layer was extracted with water (3 × 150 mL or until the washings become colorless). To the combined aqueous layers was added solid NaBF4 (9.35 g, 85.2 mmol, 3 equiv), resulting in a bright-yellow precipitate. The resulting suspension was extracted with dichloromethane (3 × 150 mL or until the washings become colorless). To the combined organic layers was added HBF4·Et2O complex (3.46 mL, 28.4 mmol, 1 equiv). The solution was swirled to achieve homogeneity then washed with water (1 × 100 mL) and aq. NaBF4 (1 M, 1 × 100 mL). The organic layer was dried over solid NaBF4, filtered, and concentrated to dryness. The residue was purified by trituration with hexanes and filtered. The solid was rinsed with n-pentane and dried in vacuo to give xanthylium 3 (10.6 g, 21.3 mmol, 75% yield) as a yellow-orange solid. 1H NMR (600 MHz, CDCl3): δ = 8.50 (s, 2 H), 7.97–7.88 (m, 2 H), 7.76 (d, J = 8.9 Hz, 2 H), 7.17 (s, 2 H), 2.49 (s, 3 H), 1.87 (s, 6 H), 1.54 (s, 18 H). 13C NMR (151 MHz, CDCl3): δ = 174.2, 171.0, 158.4, 141.3, 135.3, 129.2, 129.1, 128.5, 127.4, 122.0, 116.7, 37.5, 30.4, 21.3, 20.1. 19F NMR (376 MHz, CDCl3): δ = −153.90, −153.96. HRMS (ESI+): m/z [M]+ calcd for C30H35O: 411.2688; found: 411.2687. 3,6-Di-tert-butyl-9-mesityl-10-phenylacridin-10-ium Tetrafluoroborate (2): To an oven-dried 250 mL round-bottom flask under argon were added 3,6-di-tert-butyl-9-mesitylxanthylium tetrafluoroborate 3 (10.0 g, 20.1 mmol, 1 equiv) and dry, degassed dichloromethane (40 mL). To the resulting solution were added acetic acid (3.40 mL, 60.2 mmol, 3 equiv) followed by NEt3 (4.20 mL, 30.1 mmol, 1.5 equiv). Aniline (2.20 mL, 24.1 mmol, 1.2 equiv) was then added dropwise. The flask was covered with aluminum foil and stirred at room temperature for 12 h. The reaction was transferred to a separatory funnel and washed with water (1 × 50 mL) followed by sat. aq. NaHCO3 (1 × 50 mL). To the organic layer was added HBF4·Et2O complex (2.44 mL, 20.1 mmol, 1 equiv). The solution was swirled to achieve homogeneity then washed with water (1 × 100 mL) and aq. NaBF4 (1 M, 1 × 100 mL). The organic layer was dried over solid NaBF4, filtered, and concentrated to dryness. The residue was purified by trituration with 1:2 Et2O/hexanes and filtered. The solid was rinsed with n-pentane and dried in vacuo to give acridinium 2 (10.5 g, 18.3 mmol, 91% yield) as a bright-yellow solid. 1H NMR (500 MHz, CDCl3): δ = 7.96 (t, J = 7.6 Hz, 2 H), 7.90 (t, J = 7.5 Hz, 1 H), 7.83–7.75 (m, 4 H), 7.72 (d, J = 7.7 Hz, 2 H), 7.40 (s, 2 H), 7.16 (s, 2 H), 2.48 (s, 3 H), 1.85 (s, 6 H), 1.28 (s,18 H). 13C NMR (151 MHz, CDCl3): δ = 163.6, 162.3, 142.1, 140.2, 136.8, 136.1, 131.8, 131.6, 129.2, 128.9, 128.3, 128.0, 127.5, 124.0, 115.0, 36.6, 30.2, 21.3, 20.2. 19F NMR (376 MHz, CDCl3): δ = −154.39, −154.44. HRMS (ESI+): m/z [M]+ calcd for C36H40N: 486.3161; found: 486.3157.
- (11).Romero N; Nicewicz DA J. Am. Chem. Soc 2014, 136, 17024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (12).Benniston AC; Elliott KJ; Harrington RW; Clegg W Eur.J. Org. Chem 2009, 253. [Google Scholar]
- (13). As of February 2019, the 3-tert-butyl phenol required for the synthesis of biaryl ether 4 can be purchased for approximately US $160/mol (available in 500 g quantities from Frontier Scientific). For comparison, the 3-tert-butyl aniline required for the previous synthesis of 2 can be purchased for approximately US $707/mol (available in 100 g quantities from Combi-Blocks). Furthermore, the 2-bromo-4-(tert-butyl)benzoic acid coupling partner is prohibitively expensive (upwards of US $100,000/mol) and must be synthesized.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.