Origin of ultraweak fluorescence of 8-hydroxyquinoline in water: photoinduced ultrafast proton transfer

Abstract

8-Hydroxyquinoline (8-HQ), commonly known as oxine, is the most popular among the hydroxyquinolines due to its excellence in complex formation with metal ions and with a wide spectrum of pharmacological applications. 8-HQ and many of its derivatives are fluorogenic ligands when complexed with metal ions. Thus they are regularly used for detection, separation, and quantitative analysis of metal ions as well. For example when chelated with aluminum, the coordinated complex exhibits strong visible emission, which is applicable for the fabrication of organic light-emitting diodes. Although metal complexes with 8-HQ and its derivatives are fluorescent promising wide applications, the ligand itself, 8-HQ, is surprisingly a very weak fluorophore in most of the media, because of its short lifetime in the excited state. To address the nature of ultrashort-lived 8-HQ in the excited sate, we study its photochemical and photophysical processes in acidic, basic, and neutral aqueous media. Our results show that 8-HQ as well as its protonated and deprotonated forms, undergoes ultrafast excited-state proton transfer within femto-to-picoseconds to produce a tautomer as a product, the lifetime of which is also ultrashort. This can give a clue as to why metal complexes of 8-HQ generally show strong fluorescence whereas 8-HQ itself is rarely fluorescent in aqueous solutions; when 8-HQ binds to metal cations, the ligand is in the deprotonated (anionic) form, which cannot undergo the ultrafast proton transfer without generating the short-lived dark product, but survives much longer to tens of nanoseconds to “turn on” fluorescence.

---

1. Introduction

8-Hydroxyquinoline (8-HQ) and its derivatives have been extensively used in various fields of materials science and pharmaceutics. The wide applicability of 8-HQ and its derivatives is mainly due to metal chelation properties that has been known for more than a century.¹ 8-HQ behaves as a bidentate ligand through N and O donor atoms, which is considered as the second chelating agent in significance after EDTA.² 8-HQ is not relatively well soluble in water thus being used as a chelating ligand to separate metal ions in liquid–liquid extraction.³ In this regard, 8-HQ and its derivatives have been used analytically⁴ and industrially⁵ for the detection and the separation of metal ions. Besides, the derivatives of 8-HQ have a broad spectrum of biological activities^6,7 as neuro-protection agents,⁸ anti-cancer agents,⁹ and anti-HIV agents.¹⁰

8-HQ and its derivatives are also famous as fluorogenic ligands for different type of metal ions,^11,12 which include rare earth metal ions,¹³ transition metal ions,¹⁴ and lanthanides.¹⁵ The controllable fluorogenic property of 8-HQ and its family, when complexed with distinct metal ions, is characterized as posing the strong potential for optoelectronic materials.¹⁶ For example, the aluminium complex of 8-HQ, known as Alq3,^17,18 has been applied to solar cells and optoelectronics.¹⁹ Ruthenium complexes of 8-HQ are promising for catalysis in artificial photosynthetic systems for fuel production.²⁰

Although some metal complexes of 8-HQ are strongly fluorescent, e.g., fluorescence quantum yield (Φ) = 15% for Alq3, it is intriguing that the ligand itself is very weekly emissive in water (Φ = 0.01–0.2%).^21–24 The quantum yield significantly increases when 8-HQ is dissolved in acidic or alkaline aqueous solutions and also at low temperatures.^25,26 For instance, the emission intensity of 8-HQ is enhanced by 100 times in solutions of NaOH upon increasing basicity from pH ≈ 13 to H₀ ≈ 17.5 (H₀ is the Hammet acidity function for strongly acidic or alkaline solutions).²⁵ Similarly, there is a large increase of fluorescence when H₀ changes from −6 to −10 in a concentrated sulfuric acid solution.²⁶ Substantial amounts of ambiguities are still present on the reason behind the fluorescence enhancement of 8-HQ in acidic and alkaline solutions.²⁷ Interestingly, because of its well-known low quantum yield in mild conditions in terms of pH, there are only limited reports on the photophysics and photochemistry of 8-HQ in contrast to its various application to light emitting devices and fluorogenic sensing of metal ions.

In general the fluorescence quantum yield of a dye molecule is inversely related to its fluorescence lifetime, typically on the timescale of 0.2–10 nanoseconds (ns) depending on its oscillator strength with no radiationless transition.^28–30 Although the time-resolved studies on 8-HQ fluorescence are rare, other hydroxyquinolines (HQs) showing higher quantum yields have been reported intensively to reveal excited-state prototropy, which is completed within hundreds of picoseconds.^31–34 In general, the bifunctional groups of the acidic enol and the basic imine in HQs allow for the existence of four prototropic species in the ground state depending on the pHs of the solutions; normal (N), tautomeric (T), cationic (C), and anionic (A) species (Scheme 1).^34–38 This is also the case for 8-HQ. As each prototropic reaction, proton transfer, is reversible, the position of equilibrium depends on the pH of the solutions. Upon photoexcitation to the first excited singlet state, the enol and the imine groups of HQs become more acidic and basic, respectively, relative to those in the ground state, shifting the equilibriums with new pK_a values in the excited states (pK^*_a) and driving excited-state proton transfer (ESPT) to occur. Each prototropic species has characteristic lifetime and emission wavelength.

Scheme 1 Ground-state equilibria among different prototropic species of HQs in water.

The mechanism of ESPT is diverse for each HQ of different molecular structure regarding the distance between the prototropic bifunctional groups (Scheme 1) and specific solvation environment because adjacent protic solvent molecules play an important role in bifurcating the mechanism of ESPT and changing rates related.^31–34 Aqueous 3-HQ, as well as 6-HQ, with the enol and the imine group being distal to each other is reported to undergo ESPT on both the functional groups with adjacent independent water clusters to complete prototropic tautomerisation via forming anionic intermediate species (Scheme 2). On the other hand, the ESPT fashion of 7-HQ becomes more complicated in protic solvents due to the decreased distance between the functional groups compared to those in 3-HQ and 6-HQ. trans-7-HQ, depending on the orientation of the enol group with respect to the imine group, follows the same ESPT mechanism as that for 3-HQ and 6-HQ. However, cis-7-HQ in both polar aprotic and nonpolar solvents can form a cyclically hydrogen (H)-bonded complex with protic guest molecules, with which proton relay from the acidic to the basic site of the molecule takes place, resulting in the production of a ketonic tautomer (Scheme 3).

Scheme 2 Excited-state tautomerisation of 3-HQ and 6-HQ in water via acid–base reactions.

Scheme 3 Conformation-dependent ESPT pathways of 7-HQ in water.

The notion attained from the previous studies on the ESPT of 3, 6, and 7-HQ in different media and the spectroscopic studies on 8-HQ leads us to envision that ultrafast ESPT operative possibly on the timescale of femto (fs)-to-picoseconds (ps) may be the reason for the very low quantum yield of 8-HQ fluorescence.^27,39 From literature survey the formation of intramolecular H-bonding between the –OH and the ≥N group is claimed mainly to be responsible for the low fluorescence quantum yield.²⁴ According to successive studies, 8-HQ is reported to form a very stable homo-dimer in nonpolar solvents at the ground state. Upon photoexcitation the dimer can undergo ultrafast ESPT as 7-azaindole.^40–43 Bardez and co-workers have reported the formation of the monohydrated complex of 8-HQ with residual water present in chlorinated solvents and the possibility of intramolecular proton transfer between the two functional groups via the formation of a five membered ring.^27,44 Some other spectroscopic and theoretical studies also support this idea.^39,45 However, the very weak fluorescence and the ultrashort lifetime of the excited species have hampered experimental studies to reveal detailed mechanism in the proton transfer of 8-HQ in water.

The extensive H-bond forming nature of water may allow 8-HQ for posing a variety of H-bonded configurations because of the proximity between of the two functional groups to each other: intramolecularly H-bonded, H-bonded homo-dimeric, and cyclically H-bonded with a protic solvent molecule. The possibility of the existence of the multiple H-bonded configurations can complicate the elucidation of photophysics and photochemistry of 8-HQ as well. In the present contribution, with the aid of combined ultrafast fluorescence spectroscopy and theoretical methods, we address the ESPT mechanism of aqueous 8-HQ solutions that turned out to be the main reason of the deactivation of the excited-state 8-HQ.

2. Experimental

2.1 Materials

8-HQ (purity > 98%) was purchased from Acros Organics and used as received. NaOH, HCl, D₂O, DCl, and NaOD were purchased from Sigma-Aldrich. Isotopic purity of all the deuterated samples was at least 99.9%. Aqueous solutions of 8-HQ were prepared by dissolving it in distilled water to achieve [8-HQ] = 5 × 10⁻⁴ M. Due to the low solubility of 8-HQ prolonged sonication (for 30 min) was performed with the use of a bath sonicator (Powersonic-603, Hwashin Technology). The pHs and pDs of 8-HQ solutions of [8-HQ] = 5 × 10⁻⁴ M were adjusted with the addition of HCl/NaOH or DCl/NaOD. The pD values were corrected from pH-meter reading.⁴⁶

2.2 Spectroscopic methods

UV-vis absorption spectra were collected using a UV-vis spectrophotometer (V-730, Jasco). Emission spectra were obtained with a fluorometer (QM-400, Photon Technology International). Fluorescence up-conversion technique was used to measure fs-resolved fluorescence kinetic profiles. We used the output of an amplified ytterbium-based laser system (Pharos, Light Conversion), which produces ∼170 fs pulses centered at 1030 nm at a 50 kHz repetition rate and a output power of 300 mW. The output beam was split into two parts to generate the pump and gate pulse trains. For the pump beam generation, the fundamental beam was directed to pump an optical parametric amplifier (Orpheus, Light Conversion). The output idler beam was sum-frequency mixed with the residual fundamental in a β-barium borate (BBO) crystal (type II) to generate the pump pulse at 325 nm or 365 nm, which was attenuated to ∼20 nJ at the sample in a fluorescence up-conversion spectrometer (Chimera, Light Conversion), equipped with a photomultiplier (PMC-100, Becker & Hickl) and a monochromator (MSA-130, Solar Laser System). The pump-beam polarisation was set at the magic angle (54.7°) with respect to the up-conversion crystal axis to eliminate the influence of anisotropy in the signal. The Igor program (WaveMetrics) was used to convolute the IRF and the exponential functions and fit the measured fluorescence kinetic profiles.

2.3 Computational methods

Density functional theory (DFT) and classical molecular dynamics (MD) were used to investigate the dynamics of the 8-HQ in water system. For the DFT calculations, we used the Dmol³ program package^47,48 with B3LYP exchange correlation functional^49–51 and Grimme dispersion correction with DNP 4.4 basis set. The SCF convergence criteria was set to 1.0 × 10⁻⁶. The partial charges of the geometrically optimized structures of 8-HQ, i.e., cis-8-HQ, trans-8-HQ, and a 8-HQ dimer, which are shown in Fig. S1 of ESI,† were estimated under the implicit water condition imposed by the COSMO (conductor-like screening model) method.⁵² For the MD simulations, we used COMPASS force field.^53,54 The flexible SPC water model was used with the partial charges of hydrogen and oxygen as +0.41e and −0.82e, respectively. After assigning the estimated partial charges from DFT to 8-HQ molecules, we constructed the systems containing cis-8-HQ, trans-8-HQ, or the 8-HQ dimer with 3658 water molecules, which made the system concentrations as 0.015 M, 0.015 M, and 0.03 M, respectively. First, MD with the isothermal-isobaric (NPT) ensemble was performed for the equilibration of the system for 500 ps. The densities were obtained as 0.96 g cm⁻³ for the cis- and trans-systems and 0.97 g cm⁻³ for the dimer system, respectively. Then, MD with the canonical (NVT) ensemble was performed for the production of data for 1 ns. Pressure and temperature were set to be 1 atm and 298 K, respectively, which were kept constant with Berendsen barostat and thermostat, respectively. The time step for all MD simulations was 1 fs.

3. Results and discussion

3.1 Steady-state absorption and emission spectra

We have recorded absorption and emission spectra of aqueous 8-HQ solutions at three representative pHs; acidic (pH = 1), neutral (pH = 7.5), and alkaline (pH = 13). At the neutral pH, 8-HQ shows strong absorption maximum at 305 nm and a very weak broad band around 440 nm (Fig. 1a). The strong band around 305 nm results from the lowest electronic transition of N because N is the major prototropic species in the neutral solution. From the reported prototropic equilibria of 8-HQ as presented in Table 1 and the pH condition, the composition of the N species is estimated to be 97.2% at pH = 7.5; those of C, A, and T are 0.4%, 0.4%, and 2.0%, respectively, at the same pH.⁵⁵ In the acidic solution, where the C species is exclusively present, a new band arises with a lowest-energy peak at 358 nm, which corresponds to the lowest electronic transition of C. In the basic solution, A is found to have the lowest peak at 354 nm. Accordingly, the small band appeared around 440 nm at pH = 7.5 can be attributed to arise from the electronic absorption of the T form, the fraction of which is 2% as calculated above.

Fig. 1 Absorption (a) and emission (b) spectra of 8-HQ at different pHs. The excitation wavelength is 325 nm for the pH = 7.5 solution and 365 nm for the acidic and alkaline solutions. The dashed line in the panel (a) is the magnified spectrum for the pH 7.5 solution to show T absorption (∼440 nm), while dotted spikes in the panel (b) are Raman scatter from water.

Table 1 pK_a^a and pK^*_a^b values for the prototropic equilibrium of 8-HQ in aqueous solutions

pK1 (pK^*1)	pK2 (pK^*2)	pK3 (pK^*3)	pK4 (pK^*4)
a Prototropic equilibriums are defined in Scheme 1.b pK^*a are estimated from the Förster cycle calculation;^36 pK^*a = pKa − (hν1 − hν2)/2.3RT, where hν1 and hν2 are the electronic transition energies of the conjugate acid and base, respectively.
5.1 (13.4)	9.9 (0.4)	6.8 (−5.2)	8.2 (20.8)

---

Emission spectra of 8-HQ were collected by exciting major species at the three representative pHs (Fig. 1b). It is noteworthy that emission intensity is very weak at each experimental pH that is evidenced from much weaker intensity of all the emission bands compared to Raman scatter at 366 nm and 417 nm when excited at 325 nm and 365 nm, respectively, by solvent water. Fluorescence maxima for the acidic and basic solutions are located at 495 nm and 546 nm, respectively. In the case of pH = 7.5, N* is found to emit around 431 nm. Long fluorescence tail extending above 600 nm is noticed without resolvable band feature. We could not detect any noticeable emission band from T* by exciting T at 450 nm that may be due to its low absorbance and ultrashort lifetime (see below). Steady-state emission spectra reveal that N*, C*, and A* emit at the different wavelengths with peaks at 431 nm, 495 nm, and 546 nm, respectively. From the previous studies with other HQs and the relative electronic transition energy of T to those of other prototropic species, T* is expected to emit at a longer wavelength than C* and A*, i.e., ≥550 nm.^31–34

3.2 Time-resolved fluorescence of the basic solution

Fig. 2 presents fs-resolved fluorescence kinetic profiles measured with a basic solution (pH = 12). Kinetic constants obtained by fitting the kinetic profiles are shown in Table 2. At 500 nm, a blue wavelength with respect to the peak of A* fluorescence at 546 nm, the fluorescence triexponentially decays in 140 ± 30 fs (73%), 940 ± 90 fs (22%), and 20 ± 2.9 ps (5%). At 550 nm, A* fluoresces with only two time constants of 660 ± 50 fs (61%) and 7.0 ± 0.9 ps without the 100 fs component, which is dominant at 500 nm and thus originates from hydration. Solvation dynamics in water (hydration) is generally known to occur on multiple timescales: the ultrafast component (∼100 fs) from the inertial solvation by librational motions and a slow picosecond component from the diffusive-type motions such as rotations.²⁸ For the solvation dynamics in deuterated water (D₂O), the kinetic isotope effect (KIE) has been reported to be as small as ∼1.4.^57–59 The time constant of ∼700 fs observed at 550 nm is coincidently similar to the hydration timescales. However, from the fact that the transient was collected at the maximum of A* fluorescence, solvation components has to be absent at this wavelength. Upon photoexcitation of A, the pK_a value immediately increases from 8.2 to 20.8 promoting A* to a very strong photobase, readily undergoing protonation at the imine site to generate T*. It follows that the ∼700 fs component is the decay of A* undergoing ESPT. At the red side of the A* band, 600 nm, it is found that the fluorescence decays in 1.1 ± 0.2 ps (37%) and 9.3 ± 1.0 ps (60%). The compositions of the ∼1 ps and the ∼10 ps component change with increase of monitored wavelength and the 10 ps component becomes dominant.

Fig. 2 Fluorescence kinetic profiles of 8-HQ in the basic solution ([NaOH] = 0.01 M) at various monitoring wavelengths. The excitation wavelength is 365 nm, and solid lines are the best-fitted curves to extract kinetic constants, which are given in Table 2. Isotopes and monitoring wavelengths are indicated in the panels.

Table 2 pH-dependent kinetic constants of 8-HQ in aqueous solutions observed from femtosecond-resolved fluorescence measurements

Solution	λexc^a/nm	λmon^a/nm	H/D	Time constant (ps)
a Excitation (λexc) and monitoring (λmon) wavelengths to obtain femtosecond fluorescence transients.b Negative and positive values of initial fractional amplitudes indicate rise and decay time constants, respectively.
Acidic ([HCl] = 0.01 M)	365	500	H	0.08 ± 0.04 (79%)^b + 0.47 ± 0.05 (20%)
			D	0.14 ± 0.06 (72%) + 0.78 ± 0.08 (27%)
		600	H	0.7 ± 0.02 (92.1%) + 6.97 ± 0.99 (7.9%)
Basic ([NaOH] = 0.01 M)	365	500	H	0.14 ± 0.03 (73%) + 0.94 ± 0.09 (22.3%) + 20 ± 2.9 (4.7%)
		550	H	0.66 ± 0.05 (60.7%) + 7.0 ± 0.86 (36.6%) + ∞ (2.7%)
			D	1.1 ± 0.07 (62.4%) + 19.8 ± 3.5 (32.7%) + ∞ (4.9%)
		600	H	1.14 ± 0.24 (36.5%) + 9.3 ± 1.04 (59.8%) + ∞ (3.7%)
		630	H	0.1 ± 0.09 (−82%)^b + 7.6 ± 0.27 (100%)
Neutral (pH 7.5)	325	430	H	0.7 ± 0.1 (52.7%)^b + 2.6 ± 1.3 (32%) + 8.8 ± 3.1 (15%)
			D	1.1 ± 0.1 (53.1%) + 5.6 ± 2.3 (22.1%) + 21 ± 3.2 (24.8%)
		500	H	0.29 ± 0.1 (31%) + 4.23 ± 0.5 (47%) + 16.6 ± 5.8 (20%)
			D	3.8 ± 0.4 (34%) + 21.9 ± 1.1 (64%)
		600	H	0.75 ± 0.1 (−47.1%) + 4.49 ± 5 (−19%) + 11.6 ± 1.4 (100%)
			D	0.8 ± 0.12 (−34.2%) + 13 ± 6.2 (−44.9%) + 27.8 ± 2.2 (100%)

---

From the finding that there are the two global lifetimes, i.e., ∼1 ps (700 fs to 1.1 ps) and ∼10 ps (7–10 ps) except solvation-related one, and the presence of the 10 ps component at the blue wavelength of A*, the simplest mechanism that accommodates the results is proposed to be a two-state reversible intermolecular proton transfer involving A* and T* as depicted in Scheme 4. In this model, k_p and k_−p refer to the rate constants for the protonation of the imine group of A* (≥N) and the deprotonation of the protonated group (≥NH⁺), respectively. k_A and k_T are the relaxation rates of A* and T*, respectively, to their ground states. The time evolution of the fluorescence of the two states follows the well-known equations as:

I_A*(t) = A₁ exp(−λ₁t) + A₂ exp(−λ₂t) (1)

I_T*(t) = A₃{exp(−λ₁t) − exp(−λ₂t)} (2)

where λ₁ and λ₂ are the reciprocals of the longest (τ₁) and the shortest (τ₂) decay times, respectively, given by:⁵⁶with

Y = k_−p + k_T = λ₁ + λ₂ − X (5)

Scheme 4 ESPT of 8-HQ in the basic solution.

The ratio of the pre-exponential factors is given by R = A₂/A₁. The four rate constants in Scheme 4 can be deduced from the three parameters obtainable from the above decays (λ₂, λ₁ and R) if the fluorescence lifetime of A* is available. The lifetime of A* was obtained to be ∼1 ns with a very basic solution. Then, from the eqn (3)–(5), k_p, k_−p, and k_T are calculated to be (1.0 ps)⁻¹, (2.3 ps)⁻¹, and (4.6 ps)⁻¹, respectively. The k_p and k_−p can be further expressed as:

k_p = k′_p + k′′_p[H₃O⁺] (6)

k_−p = k′_−p + k′′_−p[OH⁻] (7)

where k′_p and k′_−p are water-catalyzed rates and k′′_p and k′′_−p are the diffusion-limited rate constants with the hydronium and the hydroxide ion, respectively. Because the concentration of the hydronium ion is negligible at pH = 12, A* undergoes the protonation with adjacent water molecules so that k_p = k′_p. In addition, k′′_−p[OH⁻] can be estimated to be 1 to 10 × 10⁸ s⁻¹ at pH = 12 using typical diffusion-limited rate constants of the hydronium ion in water ranging 1 to 10 × 10¹⁰ M⁻¹ s⁻¹, which translate into the time constants of 1 to 10 ns. These values are much smaller than the deduced k_−p value that the deprotonation of T* to form the parent A* back occurs effectively with water molecules as well. This argument also applies to the deprotonation of C* in the acidic solution discussed in the next section.

In D₂O at pD = 12, k_p, k_−p, and k_T are obtained to be (1.6 ps)⁻¹, (3.7 ps)⁻¹, and (13.3 ps)⁻¹, respectively. The kinetic isotope effects (KIEs) of ∼1.6 for the ultrafast imine protonation of A* (k_p) and the deprotonation of T* (k_−p) by water molecules nearby are in line with the free-energy relationship; the more the reaction is exothermic, the smaller the KIE because the reaction rate approaches an asymptotic value for isotope-insensitive solvent motions (KIE ∼ 1.4) on the timescale of several ps or faster; at room temperature the Debye relaxation time for water is 7 ps.⁶⁰ Our data shows small KIE of 1.6, indicating that the ESPT of 8HQ in water is barrierless and solvent-motion controlled.

3.3 Time-resolved fluorescence of the acidic solution

From the Förster-cycle analysis,³⁶ ΔpK_a upon photoexcitation of C is estimated to be −12. It follows that the C* species is highly susceptible for undergoing deprotonation to form T*. In the acidic solution at pH = 2 where the C species exclusively exists, fluorescence monitored at the maximum (500 nm) shows a biexponential decay profile comprised of 80 ± 40 fs (79%) and 470 ± 50 fs (20%), while that probed at 600 nm decays on the time scale of 700 ± 20 fs (92%) and 7.0 ± 1.0 ps (8%) (Fig. 3). In the deuterated acidic solution, the decay times of C* are 140 ± 60 fs (72%) and 780 ± 80 fs (27%), showing small KIEs. From the fact that the C* fluorescence was measured at the wavelength of its maximum both the components of 80 fs and 470 fs are attributed to the lifetime of C* undergoing ESPT. As similarly observed in the basic solution, the emergence of the 7 ps component at a red wavelength compared to the C* fluorescence maximum is ascribed to the lifetime of the T* species.

Fig. 3 Fluorescence kinetic profiles of 8-HQ in the acidic solution ([HCl] = 0.01 M) at various monitoring wavelengths. The excitation wavelength is 365 nm, and solid lines are the best-fitted curves to extract kinetic constants, which are given in Table 2. Isotopes and monitoring wavelengths are indicated in the panels.

To understand the ESPT dynamics of 8-HQ in the acidic aqueous solution we have analysed the observed biexponential kinetic constants of C* and T* by employing the irreversible two-state model considering the absence of the T* lifetime in the fluorescence wavelength of C*. The k_C of Scheme 5 denotes the overall relaxation rate constant of C* except ESPT. Then, the temporal behaviours of C* and T* can be expressed as follows:

I_C*(t) = A₄ exp{−(k_dp + k_C)t} (8)

I_T*(t) = A₅[exp(−k_Tt) − exp{−(k_dp + k_C)t}] (9)

Scheme 5 ESPT of 8-HQ in the acidic solution.

k_C is taken by measuring the C* fluorescence lifetime in a very acidic solution ([HCl] = 11.3 M), which is in the order of ns. Therefore, the weighted average of the ESPT-related time constants (80 fs and 470 fs) at 500 nm effectively corresponds to the reciprocal of k_dp. Finally, k_dp and k_T are deduced to be (160 fs)⁻¹ and (7 ps)⁻¹, respectively. From similar analysis with the deuterated acidic solution KIE of k_dp is found to be ∼2, indicating that the deprotonation of C* is also a barrierless process and solvent-motion (hydration) controlled according to the free-energy relation. It is remarkable that the k_dp value for C* of 8-HQ is among the largest values reported for the intermolecular ESPT of photoacids.^28,61

3.4 Hydrogen-bonding configuration of 8-HQ in water

The MD study was implemented to search for the possible and the favourable structures of 8-HQ in solvent water. This information is essential to decipher the ESPT mechanism in neutral water, which bifurcates depending on the H-bonding environment of 8-HQ. We considered several models with different charge states because 8-HQ can exist in the form of cis-, trans-, and dimer configurations as shown in Fig. 4. Based on the torsional angle analysis (Fig. S2a of ESI†), the ratio of cis-8-HQ to trans-8-HQ was found to be 2.57 from the averaged fraction of 8-HQ geometries in the MD simulations over 1 ns time frame. This result agrees with the previous finding⁶² on the cis-8-HQ as a major conformation. The formation of the 8-HQ dimer was checked by tracking the time-dependent distance between the centroids of the nitrogen atom and the hydroxyl group of two 8-HQ molecules in a simulation box. Note that criterion distance for the ESPT-reactive dimer formation was set to be less than 3.17 Å, which was estimated from DFT calculation for the optimized 8-HQ dimer model (Fig. S2b of ESI†). The fractional time was found to be 0.016 ns ns⁻¹, which indicates the negligible formation of the dimer in water.

Fig. 4 Possible 8-HQ H-bonded configurations in water : 8-HQ monomer (a); 1 : 1 cyclic cis-8HQ (b); 1 : 2 cyclic cis-8HQ (c); 1 : 2 non-cyclic cis-8HQ (d); 1 : 2 non-cyclic trans-8HQ (e); and homo-dimer (f).

To check the abundance of various H-bonded configurations of 8-HQ with water, we considered following candidate models, which are expected to participate in ESPT, as an 8-HQ monomer, a 1 : 1 cyclic cis-8-HQ complex with water, a 1 : 2 cyclic cis-complex, a 1 : 2 non-cyclic cis-complex, and a 1 : 2 non-cyclic trans-complex (Fig. 4). They were classified according to the distance and the donor–acceptor information of H-bonds around an 8-HQ molecule. Note that the specifically classified models were monitored during ∼0.18 ns ns⁻¹, i.e., in fractional time, because other water molecules around 8-HQ interrupted the formation of the classified H-bonding configurations. Within the ∼0.18 ns time frame, the fraction for the 8-HQ monomer, the 1 : 1 cyclic cis-complex, 1 : 2 cyclic cis-complex, 1 : 2 non-cyclic cis-complex, and 1 : 2 non-cyclic trans-complex were found to be 5.6, 17.3, 0.8, 26.3, and 50.0%, respectively. This indicates that 1 : 2 non-cyclically H-bonded complexes of 8-HQ with water is the major configuration in neutral water. Also in water, the presence of the 1 : 2 cyclic cis-complex and the intramolecularly H-bonded 8-HQ monomer are insignificant.

Among the classified models above, the non-cyclically H-bonded configurations, whether 1 : 1- or 1 : 2-complexed with water, and the cyclically H-bonded ones may follow different trajectory for ESPT. When 8-HQ is cyclically H-bonded with water, it can undergo ESPT via a catalytic water bridge, whereas for the non-cyclically H-bonded one the tautomerisation can occur by acid–base reactions with independent water molecules at the two functional groups of 8-HQ. From our MD results, the probability ratio of cyclic vs. non-cyclic complexes was deduced to be 1 : 4.22, indicating that the ESPT of 8-HQ in neutral water mainly takes place by the latter mechanism. To check the inter-conversion between the two distinct classes, we ran MD simulations repeating 11 times with the 1 : 2 noncyclic cis-complex as a starting model and examined the configuration of 8-HQ at every 20 fs. We found that averaged transformation time is ∼18.3 ps, which turned out to be about 5 times slower than the experimentally observed lifetime of N* in this study (see the next section). Therefore, we can consider that the transformation of the H-bonding configuration in the 1 : 2 non-cyclic cis-complex is effectively frozen during ESPT.

3.5 Time-resolved fluorescence of the neutral solution

In Fig. 5, N* fluorescence at 430 nm shows a tri-exponential decay composed of 700 ± 100 fs (53%), 2.6 ± 1.3 ps (32%), and 8.8 ± 3.1 ps (15%). The 700 fs decay is the major component observed at the maximum wavelength of the N* fluorescence, and this component cannot be that of ultrafast solvation, which appears as a decay at the blue wavelengths with respect to the fluorescence maximum. The fluorescence at 600 nm, presumably to selectively monitor the time dependence of T* rises in 750 ± 100 fs and 4.5 ± 5 ps, and relaxes in 11.6 ± 1.4 ps. At 500 nm, which is an in-between wavelength of N* and T* and may reveal the fluorescence of C* or A*, a possible intermediate species if any, during tautomerisation by ESPT, the fluorescence transient was fit with three kinetic constants of 290 ± 100 fs (31%), 4.2 ± 0.5 ps (47%), and 16.6 ± 5.8 ps (20%).

Fig. 5 Fluorescence kinetic profiles of 8-HQ in the neutral solution (pH = 7.5) at various monitoring wavelengths. The excitation wavelength is 325 nm, and solid lines are the best-fitted curves to extract kinetic constants, which are given in Table 2. Isotopes and monitoring wavelengths are indicated in the panels.

The observation made in the neutral solution seems complicated in that there are several different time constants revealed and the possible interplay of the intermediate species in the ESPT. However, two unique features are noticed in the fluorescence kinetics compared to those measured in the acidic and basic solutions. First, there appears the new time constant of 2.6–4.5 ps, which is absent in the solutions at the other two pHs. Second, in the fluorescence transient measured at 600 nm, where we anticipate the unknown T* band resides, rise components are clearly resolved that is in contrast to the fluorescence transients obtained with the other solutions at different pHs.

From the fact that the several-ps component is present only in the neutral solution, in which the precursor for ESPT is N*, it is plausible that the new component originates from its fluorescence lifetime. Likewise, the ∼10 ps component can be attributed to the lifetime of T* because the T* is the only common prototropic species and the final product at the excited state at all the pHs as a consequence of ESPT from A*, C*, and N* (see previous sections). The emergence of the rise components at 600 nm indicates that the fluorescence of T* is well separated from the fluorescence of parent N*. Also it is explained in a way that the accumulation of a possible intermediate species, which can more strongly interfere the fluorescence at 600 nm, can be temporally negligible because of its short-lived nature, during the tautomerisation from N* to T*.

As we have seen in the previous section, the H-bonding configuration of N can be classified mainly into two: one is non-cyclically H-bonded to water molecules as a major (up to 76%), where the enol and the imine group are independently hydrated, and the other minor, cyclically H-bonded with a water molecule. When non-cyclically H-bonded, N* is expected to undergo ESPT to form T* stepwise via forming an intermediate such as A* or C*. For other HQs (3-HQ, 6-HQ, and 7-HQ) in neutral water it has been reported that the consecutive ESPT occurs through forming transient A* as the intermediate.^34,35,63,64 For 8-HQ in the neutral solution, we find that the time constant of 700 fs, which also coincides with the lifetime of A* in the basic solution, indicative for the identification of the intermediate species. Therefore, it is persuasive to assign the 700 fs component to the A* species as the intermediate during the tautomerisation. This assignment is in line with the negligible accumulation of A* fluorescing around 550 nm, of which the tail to longer wavelengths can cause phenomenological, spectral congestion at 600 nm for monitoring T*, allowing us for resolving the formation of T*. In addition, as a control experiment we investigated the ESPT of 8-methoxyquinoline, in which the acidic, hydroxyl group of 8-HQ is replaced by a methoxy group so that only the possible prototropic species are N* and C*. In this case, we found that the protonation of N* to form C* with ΔpK_a = −9.7, which is comparable to the ΔpK_a value for N* (8-HQ), −9.2, to form C*, takes several hundred ps, which is not the any observed time constants for the ESPT of 8-HQ (data not shown). This also supports the idea that the identity of the intermediate species during the ESPT of 8-HQ is A*.

Bearing in mind that there exist three main prototropic species of N*, A*, and T*, we now categorize all the observed time constants into three: 700 fs, 3.4 ps, and 11.6 ps. The choice of the time constants of 700 fs and 11.6 ps is made from the fit at 430 nm and 600 nm, respectively, based on the fact that their fractions are dominant and thus the time constants are the most reliable at the corresponding wavelengths. 3.4 ps was determined by taking the averaged value of 2.6 ps and 4.2 ps extracted from the fits for the transients measured at 430 nm and 500 nm, respectively, because the wavelength of 430 nm is well defined for N* fluorescence and the fraction of the corresponding component is pronounced at 500 nm. Using the three lifetimes and their fractions at 430 nm and 600 nm, we performed global analysis to deduce the rate constants in Scheme 6 for a reversible three-state model. The mechanism translates into the following differential equations:

where X = k_dp + k_N, Y = k_−dp + k_p + k_A, and Z = k_−p + k_T. k_N, the relaxation rate of N* other than deprotonation, and k_A are assumed to be 1 ns, which is at least two orders-of-magnitude larger than any time constants reported in this study. The selection of the relaxation rates on the order of (1 to 10 ns)⁻¹ does not significantly affect the much larger ESPT rate constants. The set of rate-law equations was numerically solved. The aforementioned procedure and analysis was conducted for the fluorescence transients measured for the deuterated aqueous solution as well. Finally, Table 3 summarizes the evaluated rate constants for the ESPT of non-cyclically H-bonded 8-HQ in the neutral solutions. The first and the second step are ultrafast on the timescales of 1–3 ps (in H₂O) and take place barrierless controlled by solvent motions, which is evident from negligible KIEs of 1–2 for both the processes.

Scheme 6 ESPT of 8-HQ in the neutral solution.

Table 3 ESPT and relaxation kinetic constants of 8-HQ in the neutral solution

H/D	(kdp)^−1/ps	(k−dp)^−1/ps	(kp)^−1/ps	(k−p)^−1/ps	(kT)^−1/ps
H	1.35	2.06	2.53	16.7	8.13
D	2.51	2.61	4.95	11.4	11.8

---

We note that about 34% of T* fluorescence rises instantly, faster than our instrument response function (IRF) having FWHM of 240 fs; 21% for the deuterated solution. This may result from the possible small spectral congestion of A* fluorescence at 600 nm. It is also possible explanation that the ESPT of the cyclically H-bonded 8-HQ with one water molecule, which was found to exist as much as 17% among all H-bonded configurations from the MD simulations in the previous section, takes place within the IRF, reflected as an immediate rise component. The fluorescence of the trace amount of intramolecularly H-bonded 8-HQ can also contribute to the instant rise signal at 600 nm if the timescale of intramolecular ESPT is a few hundred fs or shorter, typical in other systems.⁶⁵

4. Summary

Although some metal complexes of 8-HQ, such as Alq3, are reported to be strongly fluorescent, 8-HQ itself is very weakly fluorescent in water. In this report, to provide the photochemical mechanism for the ultraweak fluorescence of aqueous 8-HQ, we have investigated the excited-state proton transfer of 8-HQ in aqueous solutions at various pHs. As a result, we have found that the cationic form and the anionic form of 8-HQ in the excited states are extraordinarily reactive undergoing deprotonation in 100 fs to 500 fs and protonation in 700 fs, respectively. Our MD simulation shows that 8-HQ in water mainly exists as non-cyclically H-bonded to adjacent water molecules with the two acidic enol and the basic imine functional groups being independently hydrated. This indicates that relatively small fraction of 8-HQ form cyclically H-bonded complex with a water molecule. The non-cyclically H-bonded 8-HQ is proposed to undergo photoinduced prototropic tautomerisation stepwise via forming the anionic species as a reaction intermediate to yield a tautomer as a product in the excited state. The proton-transfer dynamics is found to be more facile compared to that of other hydroxyquinolines, resulting in the ultrashort lifetimes of all the involved excited species: ∼3 ps for the parent 8-HQ; 700 fs for the anionic intermediate; and ∼10 ps for the tautomer as a product. This is found to be the main reason for very weak fluorescence of 8-HQ in aqueous solutions in contrast to well-known strong fluorescence of 8-HQ when chelated with metal ions without undergoing proton transfer. The remaining questions after this study include spectral identification of the fluorescence of the tautomeric form of 8-HQ and its short-lived nature in the excited state. Accordingly, further investigations for quantum mechanical calculations and fs-resolved fluorescence spectral measurements are in progress.

Acknowledgements

This work was supported by NRF Korea funded by the Ministry of Science, ICT and Future Planning (MSIP) (2014R1A1A1008289). O. H. K. also received support from the 2013 Research Fund of UNIST and the Institute for Basic Science (IBS-R020-D1). S. K. K. acknowledges the financial supports from CRC-14-1-KRICT and the computational resources from UNIST-HPC and KISTI-HPC/PLSI. We thank Prof. Mi Hee Lim for careful reading of the manuscript.

Notes and references

1. J. J. Fox, J. Chem. Soc., 1910, 1119.
2. S. F. Mason, J. Philp and B. E. Smith, J. Chem. Soc. A, 1968, 3051.
3. J. Stary, Anal. Chim. Acta, 1963, 28, 132.
4. L. L. Merritt Jr and J. K. Walker, Ind. Eng. Chem., 1944, 16, 387.
5. G. P. Demopoulos and P. A. Distin, Hydrometallurgy, 1983, 11, 389.
6. A. M. Mancino, S. S. Hindo, A. Kochi and M. H. Lim, Inorg. Chem., 2009, 48, 9596.
7. J.-S. Choi, J. J. Braymer, R. P. R. Nanga, A. Ramamoorthy and M. H. Lim, Proc. Natl. Acad. Sci., 2010, 107, 21990.
8. P. J. Crouch and K. J. Barnham, Acc. Chem. Res., 2012, 45, 1604.
9. S. H. Chan, C. H. Chui, S. W. Chan, S. H. Kok, D. Chan, M. Y. Tsoi, P. H. Leung, A. K. Lam, A. S. Chan, K. H. Lam and J. C. Tang, ACS Med. Chem. Lett., 2012, 4, 170.
10. K. Mekouar, J.-F. Mouscadet, D. Desmaele, F. Subra, H. Leh, D. Savoure, C. Auclair and J. d'Angelo, J. Med. Chem., 1998, 41, 2846.
11. M. Albrecht, M. Fiege and O. Osetska, Coord. Chem. Rev., 2008, 252, 812.
12. S. Cho, M. W. Mara, X. Wang, J. V. Lockard, A. A. Rachford, F. N. Castellano and L. X. Chen, J. Phys. Chem. A, 2011, 115, 3990.
13. M. A. Palacios, Z. Wang, V. A. Montes, G. V. Zyryanov and P. Anzenbacher Jr, J. Am. Chem. Soc., 2008, 130, 10307.
14. H. Zhang, L.-F. Han, K. A. Zachariasse and Y.-B. Jiang, Org. Lett., 2005, 7, 4217.
15. N. M. Shavaleev, R. Scopelliti, F. Gumy and J.-C. G. Bunzli, Inorg. Chem., 2009, 48, 2908.
16. Z. Yin, B. Wang, G. Chen and M. Zhan, J. Mater. Sci., 2011, 46, 2397.
17. C. W. Tang and S. A. VanSlyke, Appl. Phys. Lett., 1987, 51, 913.
18. M. Colle, R. E. Dinnebier and W. Brutting, Chem. Commun., 2002, 2908.
19. L. G. C. Rego, R. da Silva, J. A. Freire, R. C. Snoeberger and V. S. Batista, J. Phys. Chem. C, 2010, 114, 1317.
20. L. Duan, L. Wang, F. Li, F. Li and L. Sun, Acc. Chem. Res., 2015, 48, 2084.
21. C. Perez-Bolivar, S. Takizawa, G. Nishimura, V. A. Montes and P. Anzenbacher Jr, Chem.–Eur. J., 2011, 17, 9076.
22. E. M. Filip, I. V. Humelnicu and C. I. Ghirvu, Acta Chem. Iasi, 2009, 17, 85.
23. Y. Huo, J. Lu, S. Hu, L. Zhang, F. Zhao, H. Huang, B. Huang and L. Zhang, J. Mol. Struct., 2015, 1083, 144.
24. M. Goldmanl and E. L. Wehry, Anal. Chem., 1970, 42, 1178.
25. S. G. Schulman, Anal. Chem., 1971, 43, 285.
26. R. E. Ballard and J. W. Edwards, J. Chem. Soc., 1964, 4868.
27. E. Bardez, I. Devol, B. Larrey and B. Valeur, J. Phys. Chem. B, 1997, 101, 7786.
28. O. F. Mohammed, D. Pines, E. T. J. Nibbering and E. Pines, Angew. Chem., Int. Ed., 2007, 46, 1458.
29. R. Hu, E. Lager, A. Aguilar-Aguilar, J. Liu, J. W. Y. Lam, H. H. Y. Sung, I. D. Williams, Y. Zhong, K. S. Wong, E. Pena-Cabrera and B. Z. Tang, J. Phys. Chem. C, 2009, 113, 15845.
30. K. M. Solntsev, S. P. Laptenok and P. Naumov, J. Am. Chem. Soc., 2012, 134, 16452.
31. T. G. Kim, Y. Kim and D.-J. Jang, J. Phys. Chem. A, 2001, 105, 4328.
32. S.-Y. Park, H.-B. Kim, B. K. Yoo and D.-J. Jang, J. Phys. Chem. B, 2012, 116, 14153.
33. O.-H. Kwon, Y.-S. Lee, B. K. Yoo and D.-J. Jang, Angew. Chem., Int. Ed., 2006, 45, 415.
34. H. Yu, O.-H. Kwon and D.-J. Jang, J. Phys. Chem. A, 2004, 108, 5932.
35. E. Bardez, A. Chatelain, B. Larrey and B. Valeur, J. Phys. Chem., 1994, 98, 2357.
36. L. M. Tolbert and K. M. Solntsev, Acc. Chem. Res., 2002, 35, 19.
37. S.-Y. Park, H. Jeong, H. Yu, S. Y. Park and D.-J. Jang, Photochem. Photobiol., 2010, 86, 1197.
38. B. Kang, K. C. Ko, S.-Y. Park, D.-J. Jang and J. Y. Lee, Phys. Chem. Chem. Phys., 2011, 13, 6332.
39. M. Amati, S. Belviso, P. L. Cristinziano, C. Minichino, F. Lelj, I. Aiello, M. L. Deda and M. Ghedini, J. Phys. Chem. A, 2007, 111, 13403.
40. C. M. Cheatum, M. M. Heckscher and F. F. Crim, Chem. Phys. Lett., 2001, 349, 37.
41. L. R. Naik and N. N. Math, Indian J. Pure Appl. Phys., 2005, 43, 743.
42. O.-H. Kwon and A. H. Zewail, Proc. Natl. Acad. Sci., 2007, 104, 8703.
43. S. Takeuchi and T. Tahara, Proc. Natl. Acad. Sci., 2007, 104, 5285.
44. E. Bardez, Isr. J. Chem., 1999, 39, 319.
45. S.-C. Lan and Y.-H. Liu, Spectrochim. Acta, Part A, 2015, 139, 49.
46. P. K. Glasoe and F. A. Long, J. Phys. Chem., 1960, 64, 188.
47. B. J. Delley, Chem. Phys., 1990, 92, 508.
48. B. J. Delley, Chem. Phys., 2000, 113, 7756.
49. A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
50. B. Miehlich, A. Savin, H. Stoll and H. Preuss, Chem. Phys. Lett., 1989, 157, 200.
51. C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 37, 785.
52. B. Delley, Mol. Simul., 2006, 32, 117.
53. H. Sun, J. Phys. Chem. B, 1998, 102, 7338.
54. H. Sun, Comput. Theor. Polym. Sci., 1998, 8, 229.
55. O. K. Abou-Zied, J. Husband, N. Al-Lawatiaa and T. B. Steinbrecher, Phys. Chem. Chem. Phys., 2014, 16, 61.
56. L. Giestas, C. Yihwa, J. C. Lima, C. Vautier-Giongo, A. Lopes, A. L. Maçanita and F. H. Quina, J. Phys. Chem. A, 2003, 107, 3263.
57. N. Nandi, K. Bhattacharyya and B. Bagchi, Chem. Rev., 2000, 100, 2013.
58. N. Nandi, S. Roy and B. Bagchi, J. Chem. Phys., 1995, 102, 1390.
59. B. Zolotov, A. Gan, B. D. Fainberg and D. Huppert, Chem. Phys. Lett., 1997, 265, 418.
60. H. Sona, D.-H. Choia, S. Jung, J. Park and G.-S. Park, Chem. Phys. Lett., 2015, 627, 134.
61. R. Simkovitch, K. Akulov, S. Shomer, M. E. Roth, D. Shabat, T. Schwartz and D. Huppert, J. Phys. Chem. A, 2014, 118, 4425.
62. M. Amati, S. Belviso, P. L. Cristinziano, C. Minichino and F. Lelj, J. Phys. Chem. A, 2007, 111, 13403.
63. H. J. Park, O.-H. Kwon, C. S. Ah and D.-J. Jang, J. Phys. Chem. B, 2005, 109, 3938.
64. O.-H. Kwon and O. F. Mohammed, Phys. Chem. Chem. Phys., 2012, 14, 8974.
65. E. Driscoll, S. Sorenson and J. M. Dawlaty, J. Phys. Chem. A, 2015, 119, 5618.

---

Footnote

† Electronic supplementary information (ESI) available: Charge distribution of optimized 8-HQ monomer and 8-HQ dimer and probability distribution of torsional angle analysis of 8-HQ monomer. See DOI: 10.1039/c5ra23802a

---