How does excited-state antiaromaticity affect the acidity strengths of photoacids?

Abstract

Photoacids like substituted naphthalenes (X = OH, NH₃⁺, COOH) are aromatic in the S₀ state and antiaromatic in the S₁ state. Nucleus independent chemical shifts analyses reveal that deprotonation relieves antiaromaticity in the excited conjugate base, and that the degree of “antiaromaticity relief” explains why some photoacids are stronger than others.

Baird first proposed a set of rules suggesting that the electron-counting rules of aromaticity and antiaromaticity reverse in the lowest triplet states of π-conjugated cycles.¹ Based on this set of rules, [4n + 2] π-rings are antiaromatic and [4n] π-rings are aromatic in the first ππ* state. These predictions were later extended to the first singlet ππ* states (S₁), and explained the reactivities of many Baird-type antiaromatic [4n + 2] π-systems,^2–7 Benzene, in the S₁ state, is reactive and readily isomerizes to fulvene.^7,8 Among other [4n + 2] π-ring systems, salicylic acid, in the S₁ state, undergoes intramolecular proton transfer.^9,10 Here we show that, in the S₁ state, differences in the acidity strengths of photoacids might be rationalized by the effects of “antiaromaticity relief" upon deprotonation, followed by a redistribution of electrons in the excited conjugate base.

Following the early independent works of Förster^11,12 and Weller,^13–15 it was recognized that some aromatic acids (e.g., with hydroxyl or ammonium groups) can turn into stronger Brønsted acids in their first excited (S₁) states.^11–16 2-Naphthol, a prototypical organic “photoacid,” is a weak acid in the ground state, but shows enhanced acidity in the state (pK_a = 9.5, pK_a* = 2.8, ΔpK_a = −6.7), and can deprotonate to the solvent producing an electronically excited conjugate base.¹⁵ A Stokes’ shift in the fluorescence spectrum of 2-naphthol in water was interpreted by Forster as radiative decay emitting from the excited conjugate base (see depiction of “Förster cycle” in Scheme 1).¹¹ Yet, despite a large body of theoretical and experimental efforts towards understanding excited-state proton transfer reactions in aromatic acids,^17–27 reasons underlying the occurrence of photoacidity remain unclear. The disparate effects of substituents on photoacidity are even more puzzling. 2-Naphthylammonium displays increased acidity in the S₁ state (pK_a = 4.1, pK_a* = −0.8, ΔpK_a = −4.9), deprotonating from an NH₃⁺ group, but the change in acidity, ΔpK_a, is two-folds less than that of 2-naphthol.²⁸ Aromatic acids, like the 2-naphthoic acid, show the opposite effect and exhibit decreased acidity in the S₁ state (pK_a = 4.2, pK_a* = 11.5, ΔpK_a = +7.3)(Table 1).²⁹

Scheme 1.

The Förster cycle of 2-naphthol. As indicated by the color scheme above, 2-naphthol is [4n + 2] π-aromatic in the ground states (red) but becomes [4n + 2] π-antiaromatic in the S₁ state (green). Deprotonation relieves excited-state antiaromaticity, stabilizing the excited conjugate base (light green). In the S₀ state, the conjugate base (light red) is only moderately less aromatic than the acid.

Table 1.

Experimental pK_a pK_a*, and ΔpK_a values

Compound	pKa	pKa*	ΔpKa	Ref.
2-Naphthol	9.5	2.8	−6.7	15
2-Naphthylammonium	4.1	−0.8	−4.9	28
2-Naphthoic acid	4.2	11.5	+7.3	29
8-Cyano-2-naphthol	8.4	−0.8	−9.2	19
5,8-Dicyano-2-naphthol	7.8	−4.5	−12.3	19
Salicylamide	−2.6	−5.3	−2.7	35

Why do substituents have such disparate effects on the photoacidities of aromatic acids? Here, we relate the effects of photoacidity to a switch in the ground and excited-state (anti)aromatic character of aromatic acids. According to the Hiickel rule, cyclic π-conjugated rings with [4n + 2] π-electrons are aromatic, and those with [4n] π-electrons are antiaromatic.³⁰ But this electron-counting rule reverses in the first ππ* state following Baird’s rule.^1–7 2-Naphthol is [4n + 2] Hückel aromatic (ten π-electrons in naphthalene) in the ground state, but becomes [4n + 2] antiaromatic in the S₁ state. Upon deprotonation, an excited conjugate base forms, and negative charge on the O⁻ delocalizes into the ring, giving rise to a resonance structure with breached cyclic [4n + 2] π-electron delocalization—this alleviates antiaromaticity in the S₁ state of the acid (see Scheme 1 and resonance structure for the excited conjugate base in Fig. 1a, note delocalization of the negative charge into the ring). In this way, photoacidity might be considered as a consequence of antiaromaticity relief in the S₁ states of aromatic acids.⁶ Based on a more bond equalized S₁ vs. S₀ state of 2-naphtholate, Agrnon et al. pointed out similarly that the excited conjugate base of 2-naphthol might be stabilized by increased aromatic character.^19,20 The effects of ground and excited-state (anti)aromaticity also have been recognized in other excited-state proton transfer processes.^31,32

Fig. 1.

Computed NICS(1)_ZZ and ΔANICS(1)_ZZ (sum of NICS(1)_ZZ values of the conjugate base minus that of the acid) (in ppm) values for the S₁ states of the acid and conjugate base at CASSCF(12,12)/6–311+G(d,p), for (a) 2-naphthol, (b) 2-naphthylammonium, and (c) 2-naphthoicacid. Negative ΔNICS(1)_ZZ values indicate antiaromaticity relief, and positive values indicate antiaromaticity gain, upon formation of the excited conjugate base. Experimental ΔpK_a values are included for reference.

Even though not realized at the time, the possible effects of excited-state antiaromaticity relief were implied in Weller’s original explanation (1950’s) of photoacidity—argued based on a redistribution of ring π-electrons in the S₁ states of the aromatic acids;^13–15 notably, these ideas were published roughly ten years prior to Dewar³³ and Zimmerman’s³⁴ independent works and Baird’s¹ proposal (60’s-70’s) of a reversed Hückel π-electron-counting rule for aromaticity and antiaromaticity in the first ππ* states of transition states^33,34 and of π-conjugated rings.¹ Weller reasoned that when 2-naphthol is electronically excited to the S₁ state, intramolecular charge transfer from the hydroxyl oxygen to the aromatic ring increases acidity of the OH group. Later, it was suggested that even more pronounced charge redistribution happens upon deprotonation (as indicated by shortened C-O bond lengths and changes in dipole moments),²¹ stabilizing the excited conjugate base. We now relate the effects of “charge redistribution” in the excited conjugate base to “relief of excited-state antiaromaticity.” This rationale also may explain why photoacidity only is observed for aromatic acids [i.e., with [4n + 2] ring π-electrons), but not for other hydroxyl, amine, or ammonium compounds.

Compared to 2-naphthol, the effect of charge redistribution for alleviating antiaromaticity in the excited conjugate base of 2-naphthylammonium is much weaker, since delocalization of a neutral nitrogen lone pair into the naphthalene ring is less effective (see Fig. 1b, note charge separated resonance form). In the S₁ state of 2-naphthylammonium, deprotonation produces a neutral amine (NH₂); proton transfer alleviates antiaromaticity of the excited naphthalene ring, but to a lesser degree compared to that of 2-naphthol. Notably, compounds with competing deprotonation sites like salicylamide^35,36 and 3-ammonium-2-naphthol³⁷ undergo proton transfer from NH₃⁺ in the ground state [i.e.,, to retain aromaticity of the π-ring), but deprotonate from the OH group in the S₁ state [i.e.,, to alleviate excited-state antiaromaticity of the π-ring) when solvated in water. In the S₁ atate of 2-naphthoic acid, deprotonation of the carboxylic group gives a carboxylate (COO⁻). But negative charge is mostly delocalized between the two oxygen atoms, and does not help lessen excited-state antiaromaticity in the naphthalene ring (Fig. 1c).

We performed dissected nucleus independent chemical shifts,^38,39 NICS(1)_zz, to quantify excited-state antiaromaticity in the S₁ states⁴⁰ of the acids and conjugate bases of 2-naphthol, 2-naphthylammonium, and 2-naphthoic acid (Fig. 1). The computed ring NICS(1)_zz values of excited 2-naphthol are large and positive (+71.8, +72.3 ppm, strongly antiaromatic) but become much less so in the excited conjugate base (+24.1, +34.0 ppm, weakly antiaromatic), suggesting decreased antiaromaticity upon deprotonation of the excited acid (ΔNICS(1)_zz = −86.1 ppm, Fig. 1a). 2-Naphthylammonium reveals a lesser degree of antiaromaticity relief upon deprotonation (ΔNICS(1)_zz = −26.6 ppm, Fig. 1b). Accordingly, the computed exocyclic CO and CN bond distances of 2-naphthol (1.350 Å) and 2-naphthylamionium (1.476 Å), are longer in the S₁ state acid, and shorter in the excited conjugate base (1.247 Å and 1.373 Å, respectively), indicative of electron delocalization from the deprotonated site into the excited naphthalene ring (see optimized geometries in the ESI^†).

In contrast, computed ring NICS(1)_zz value for 2-naphthoic acid in the S₁ state are positive for the acid (+66.1, +69.2 ppm, strongly antiaromatic) but become even more so in the excited conjugate base (+78.6, +78.1 ppm, strongly antiaromatic), suggesting increased antiaromaticity upon deprotonation of the excited acid (ΔNICS(1)_zz= +21.4 ppm, Fig. 1c). The exocyclic C-C bond of the excited acid is 1.471 Å (cf. 1.40 Å CC length of benzene), indicating modest π-conjugation between the carboxylic group and the naphthalene ring. In the excited conjugate base, the exocyclic C-C bond lengthens to 1.534 Å(cf.1.53 Å CC length of ethane), suggesting little resonance between the exocyclic carbonate group and the excited (antiaromatic) naphthalene (see geometries in the ESI^†). In the ground state, deprotonation has less effect on the 10 π-electron aromatic character of the naphthalene ring, in 2-naphthol (ΔNICS(1)_zz = +8.5 ppm), 2-naphthylammonium (ΔNICS(1)_zz = +2.2 ppm), and 2-naphthoic acid (ΔNICS(1)_zz = −1.2 ppm); positive/negative values indicate aromaticity loss/gain (see full data in the ESI^†).

Naphthols with cyano (CN) substituents at the C5 and C8 positions are very strong photoacids: 8-cyano-2-naphthol (pK_a = 8.4, pK_a* = −0.8, ΔpK_a = −9.2) and 5,8-dicyano-2-naphthol (pK_a = 7.8, pK_a* = −4.5, ΔpK_a =−−12.3) show increased acidites of up to 12 units in the state.^17,18,41,42 These strong photoacids can undergo excited-state proton transfer reactions in methanol, methylsulfonyl, and other organic solvents in the absence of water, first expanding the possibility of studying proton transfer kinetics in non-aqueous solvents.⁴¹ Computed ring NICS(1)_zz values for the S₁ state of 8-cyano-2-naphthol and its excited conjugate base (ΔNICS(1)_zz = −13.72 ppm, Fig. 2a) show significant excited-state antiaromaticity relief upon deprotonation, and the effect in 5,8-dicyano-2-naphthol (ΔNICS(1)_zz = −167.3 ppm, Fig. 2b) is even greater (cf. ΔNICS(1)_zz = −86.1 ppm, for 2-naphthol). As suggested by the resonance forms in Fig. 2, the electron-withdrawing CN groups help increase charge redistribution in the excited conjugate base, by inductive effects, but also by resonance stabilization (see resonance contributors with negative charges delocalized to the nitrogen atoms). We note that other known strong photoacids with electron-withdrawing groups, e.g., sulfonyls, also have π-systems on the substituents capable of delocalizing negative charge of the excited conjugate base.

Fig. 2.

Computed NICS(1)_ZZ and ΔNICS(1)_zz (sum of NICS(1)_zz values of the conjugate base minus that of the acid) (in ppm) values for the S₁ states of the acid and conjugate base at CASSCF(12,12)/6–311+G(d,p), for (a) 8cyano-2-naphthol and (b) 5,8-dicyano-2-naphthol. Negative ΔNICS(1)_zz values indicate antiaromaticity relief upon formation of the excited conjugate base. Experimental ΔpK_a values are included for reference. Note resonance form showing delocalization of the negative charge into the CN groups.

Computed deprotonation reaction energies (ΔE) based on the equation: ArOH + H₂O → ArO⁻ + H₃O⁺, document the energetic effects of (anti)aromaticity loss in the S₀ and S₁ states upon deprotonation, and agree with the conclusions based on NICS analyses. Compared to computed ΔE values in the S₀ state, deprotonation is less endothermic for the S₁ states of 2-naphthol (ΔΔE = −17.99 kcal mol⁻¹, ΔΔE = ΔE(S₁) − ΔE(S₀)), 2-naphthylammonium (−2.77 kcal mol⁻¹), 8-cyano-2-naphthol (−28.85 kcal mol⁻¹), and 5,8-dicyano-2-naphthol (−29.28 kcal mol⁻¹), but more endothermic for the S₁ states of 2-naphthoic acid (+1.39 kcal mol⁻¹) (see full data in the ESI^†).

Aromatic acids with competing deprotonation sites like salicylamide undergo proton transfer from NH₃⁺ in the S₀ state, but deprotonate from the OH group in the S₁ state.^33,36 In the S₀ state, computed NICS(1)_zz for the acid (−24.4 ppm) and conjugate base (−24.5 ppm, deprotonated at NH₃⁺) give nearly the same values (ΔNICS(1)_zz = +0.1 ppm)(Fig. 3, bottom). But in the S₁ state, the ring NICS(1)_zz value for protonated salicylamide is large and positive (+40.2 ppm) while that for the excited conjugate base is modestly negative (−4.2 ppm, deprotonated at OH), documenting the effects of excited-state antiaromaticity relief (ΔNICS(1)_zz = −44.4 ppm) (Fig. 3, top). Notably, when a proton is removed from the NH₃⁺ site of the electronically excited acid, computed NICS(1)_zz for the excited conjugate base (+65.8 ppm) show increased excited-state antiaromaticity in the benzene ring.

Fig. 3.

Computed NICS(1)_zz and ΔNICS(1)_zz (sum of NICS(1)_zz values of the conjugate base minus that of the acid) (in ppm) values for the S₀ and S₁ states of protonated salicylamide an its conjugate base at CASSCF(12,12)/6–311+G(d,p). Negative ΔNICS(1)_zz values indicate antiaromaticity relief upon formation of the excited conjugate base. Positive ΔNICS(1)_zz values in the ground state indicate aromaticity loss upon formation of the conjugate base. An experimental ΔpK_a value is included for reference.

A long-standing and much debated anomaly is the stronger photoacidity of 1-naphthol (pK_a = 9.2, pK_a* = −0.5, ΔpK_a = −9.7) compared to its structurally similar 2-naphthol isomer (pK_a = 9.5, pK_a* = 2.8, ΔpK_a = −6.7); note 3-fold ΔpK_a difference!^43–45 Based on time-resolved emission spectroscopy and steady state spectrofluorometry experiments, 1-naphthol displayed a rate of deprotonation (k₁*) greater than 2-naphthol by 280 times.⁴⁴ It was proposed that the stronger photoacidity of 1-naphthol was a consequence of populating and emitting from the ¹L_a state, while 2-napththol shows near degenerate ¹L_a and ¹L_b states, emitting from the ¹L_b state.^17,46,47 We speculate that naphthols promoted to the ¹L_a state deprotonate more readily because of pronounced antiaromatic character in the ¹L_a state of naphthalene. Computed geometric indices of aromaticity for excited naphthalene show a more bond alternated ¹L_a state and a more bond equalized ¹L_b state (see data in the ESI^†). See also resonance structures of naphthalene in the ¹L_a (B_2u) state (“diradical form,” note “antiaromatic” Clar sextet structure in the ¹L_a) and ¹L_b (B_3u) state (“allyl radical form”) (Scheme 2). We note that Baird’s original paper on the effects of triplet (anti)aromaticity also recognized a more antiaromatic “diradical form” for triplet benzene.¹

Scheme 2.

(a) Resonance structures for the ¹L_a and ¹L_b states of excited state naphthalene, and (b) the "diradical" (B_1u)and "allyl radical" (B_2u) forms of triplet benzene.

Although substituents are typically considered to have negligible effects on aromaticity (unless charged or in the presence of other push/pull substituents),^48,49 they can easily perturb the excited-state antiaromaticity of organic compounds,⁵⁰ having tremendous effects on reactions such as the excited-state proton transfer of photoacids. These findings are another manifestation of the increasingly recognized effects of excited-state (anti)aromaticity on the photochemical reactivity of organic compounds.

Computational methods

Geometry optimizations for the ground (S₀) and excited (S₁, ¹ππ*) states of all acid and conjugate base structures were performed at CASSCF(12,12)/6–311+G(d,p) with C_s symmetry, employing Molpro2012.1.⁵¹ The S₀ and S₁ geometries of 5,8-dicyano-2-naphthol and its conjugate base were computed at CASSCF(10,10)/6–311+G(d,p) with C_s symmetry. Computed dissected nucleus-independent chemical shifts, NICS(1)_zz,^38,39 were performed at CASSCF(12,12)/6–31G(d,p), employing the Dalton2016 program,⁵² to quantify the magnetic effects of aromaticity and antiaromaticify in the S₀ and S₁ states⁴⁰ of the acids and conjugate bases. NICS(1)_zz values were computed at 1 Å above each of the six membered ring centers and include only contributions from the “out-of-plane” (zz) tensor component perpendicular to the ring plane. ΔNICS(1)_zz values were calculated based on the sum of ring NICS(1)_zz values of the conjugate base minus that of the acid.