Unexpectedly High Acidity of Water-Soluble Phosphacoumarins

Abstract

In this work, we report the one-pot synthesis and characterization of four water-soluble 2-hydroxybenzo[e][1,2]oxaphosphinine 2-oxides. The compounds were obtained by cascade reactions of (2-ethoxyvinyl)phosphonic dichloride with phenol or naphthol derivatives, and their acid–base, structural, and photophysical properties were investigated using a combination of experimental and computational methods. These compounds exhibit UV–vis absorption maxima at 209–341 nm and fluorescence maxima at 300–394 nm. Notably, these cyclic phosphonic acids exhibit unusually strong acidity with pK_a values from −1.3 to 0, comparable to mineral acids; complete protonation is not achieved even in concentrated HCl. The acidity trends and spectra were further analyzed by DFT using both explicit and implicit solvation models.

1. Introduction

Organophosphorus compounds are a large family of phosphorus-containing compounds that exhibit properties of catalysts, optical materials, chelating and extraction agents, and biologically active chemicals [1,2]. The phosphonic acid RP(O)(OH)₂ has two acidic OH groups. When R is an aromatic group, the first pKa ranges from ca 1.1 to 1.7, rendering these compounds moderately acidic [3]. Because the organic group can be changed in a modular and systematic way, such compounds have been widely used to design Brönsted acid catalysts. Besides the well-known Akiyama-Terada catalysts, novel classes of phosphorus-acid-based catalysts continue to emerge, highlighting ongoing interest in the application of phosphonic acid derivatives in this area [4,5,6].

We have previously developed a one-pot synthesis of some phosphocoumarins [7,8], a promising class of condensed cyclic phosphonic acids characterized by unique reactivity and promising biological properties: inhibition of cholesterol esterase [9], of protein tyrosine phosphatase SHP-1 [10], which is of interest from the point of view of creating antidiabetic drugs, their antibacterial activity [11] and cytotoxicity towards cancer cell lines [12,13] are also known. In our study [14] of a new reaction of phosphoralkylation of various phenols, aromatic hydrocarbons 2-hydroxy-5,7,8-trimethylbenzo[e][1,2]oxaphosphinine 2-oxide, which occurs according to the Friedel-Crafts reaction type, we found that this reaction is realized only in the presence of trifluoroacetic acid, which was used as a catalyst and solvent. It is known that one of the important conditions of the Friedel-Crafts reaction is the use of very strong acids of various types and their compositions. In our case, trifluoroacetic acid is an acid of medium strength, and since the reaction proceeds successfully, the question arose: what is the strength of this cyclic phosphonic acid that it creates such a synergistic effect of enhancing the acidity of the system? Given the presence of an acidic proton [14], we sought to investigate the acid-base behavior of cyclic phosphonic acid. Understanding these properties not only explains their reactivity in acidic media but also serves as a basis for the development of new Brønsted acid catalysts. Recently, we have demonstrated the effectiveness of the Cox-Yates method [15] for the determination of acidic parameters of phosphoric acid in strongly acidic media [16]. Thus, the aim of this work was a synthesis of derivatives of phosphacoumarins and study their acid-base properties in aqueous solution. For this study, we selected a series of phosphocoumarins synthesized by a new method, presented in Scheme 1.

2. Materials and Methods

2.1. Chemicals Used

All chemicals for the study of equilibrium processes in solution were purchased from commercial suppliers and used without further purification: aminoacetic acid (≥99%), hydrochloric acid (≥99.9%, ChimReact, Nizhniy Novgorod, Russia), acetic acid (≥99.9%, ReaChim, Moscow, Russia), sodium hydroxide (≥98%, ReaChim, Moscow, Russia), and sodium chloride (≥99%, ReaChim, Moscow, Russia). Buffer solutions were prepared as follows: glycine (98%, ChRS, Ufa, Russia)–HCl for pH 1.80–3.60, CH₃COOH–CH₃COONa(98%, ChRS, Ufa, Russia) for pH 3.60–5.60, and Tris(98%, ChRS, Ufa, Russia)–HCl for pH 7.00–8.00. For solutions with pH < 2, acidity was adjusted using hydrochloric acid. The concentration of HCl was determined by titration with a standardized Na₂CO₃ (99%, ChRS, Ufa, Russia) solution. For FTIR spectroscopy FTIR Grade Potassium Bromide (KBr, ICL, Garfield, NJ, USA) was used.

2.2. Synthesis of Compounds

2-Hydroxybenzo[e][1,2]oxaphosphinine 2-oxides P1–P4 were obtained by cascade reactions of (2-ethoxyvinyl)phosphonic dichloride (97.5%, IOPC, Kazan, Russia) with phenol (98%, IOPC, Kazan, Russia) or naphthol (98%, DAlChIM, Nizhniy Novgorod, Russia) derivatives (Scheme 2).

General procedure for the synthesis of compound P1–P4

(2-Ethoxyvinyl)phosphonic dichloride (1.00 g, 5.3 mmol, 97.5%, IOPC, Kazan, Russia) was dissolved in benzene (25 mL, 99%, ECOS-1, Moscow, Russia) and cooled in a water bath (t = 5–8 °C) under argon atmosphere (99.99%, TechnoRemStroy, Kazan, Russia) with vigorous stirring. A solution of the appropriate phenol (5.3 mmol, 98%, IOPC, Kazan, Russia) and triethylamine (0.53 g, 5.3 mmol, 99.5%, Prime Chemical Group, Moscow, Russia) in benzene (5 mL, 99%, ECOS-1, Moscow, Russia) was added dropwise. The reaction mixture was then allowed to warm to room temperature (t = 23 °C) and stirred until a precipitate of triethylammonium chloride formed, which was removed by filtration. Trifluoroacetic acid (1 mL, 99.5%, SRL, New Mumbai, India) was added to the filtrate. The mixture was stirred for 15 h at room temperature (t = 23 °C), after which the formed crystalline precipitate was filtered, washed with diethyl ether (99%, Chimprom-M, Noscow, Russia), and dried to constant weight.

2-Hydroxy-5,7,8-trimethylbenzo[e][1,2]oxaphosphinine 2-oxide P1 [7]

Yield 0.99 g (83%), white powder. Melting point 208 °C. ¹H NMR (DMSO-d₆, 600 MHz): 2.14 (s, 3H, CH₃), 2.24 (s, 3H, CH₃), 2.34 (s, 3H, CH₃), 6.24 (dd, J = 13.0 Hz, J = 21.2 Hz, 1H), 6.85 (s, 1H_arom), 7.55 (dd, J = 12.8 Hz, J = 42.8 Hz, 1H).

³¹P NMR (DMSO-d₆, 250 MHz): 5.17;

¹³C NMR (DMSO-d₆, 150 MHz): 97.53 (s, CH₃), 97.69 (s, CH₃), 97.85 (s, CH₃), 112.31 (s), 113.45 (s), 117.20 (d, J_CP = 17.0 Hz, C-4), 123.43 (d, J_CP = 6.0 Hz), 126.22 (s), 133.52 (s), 139.40 (s), 139.52 (s), 150.20 (d, J_CP = 9.0 Hz, C-6).

FTIR (KBr): 3450 (OH stretch), 3040 (C-H^aromatic stretch), 2560 (broad, weak, O-H overtone), 2270 (P=O combination bands), 1630–1560 (C=C^aromatic stretch), 1460 (C-H bending of ring deform), 1400 (C-H bending), 1320 (C-O stretch or ring def), 1232 (C-O stretch), 1225 (P=O stretch), 1070 (P-O-C stretch), 1000 (C-H^arom out-of-plane).

3-Hydroxynaphtho[1,2-e][1,2]oxaphosphinine 3-oxide P2

Yield 0.60 g (50%), white powder. Melting point 230 °C. Find (%): С 62.30; Н 3.68; Р 13.02. С₁₂H₉O₃P. Calc. (%): С 62.07; Н 3.88; Р 13.66. GS-MS (MALDI-TOF, m/z): 232.58 [M]⁺, 272.61 [М + К]⁺.

¹H NMR (DMSO-d₆, 600 MHz): 6.46 (dd, J = 12.5 Hz, J = 20.0 Hz, 1H), 7.35 (d, J = 9.0 Hz, 1H_arom), 7.50 (t, J = 7.5 Hz, 1H_arom), 7.61 (t, J = 7.5 Hz, 1H_arom), 7.94 (d, J = 8.2 Hz, 1H_arom), 7.98 (d, J = 8.2, 1H_arom), 8.27 (dd, J = 12.5 Hz, J = 42.0 Hz, 1H_arom), 8.34 (d, J = 8.2 Hz, 1H), 10.47 (br.s, 1H, OH);

³¹P NMR (DMSO-d₆, 250 MHz): 4.83;

¹³C NMR (DMSO-d₆, 150 MHz): 114.51 (d, J_CP = 18.7 Hz, C-3), 116.56 (d, J_CP = 168.0 Hz, C2-P), 119.64 (d, J_CP = 6.7 Hz, C-4), 122.61 (C-6), 125.63 (C-7), 126.60 (C-8), 128.29 (C-9), 130.14 (C-5), 130.64 (C-10), 132.15 (C-11), 137.48 (C-12), 150.75 (d, J_CP = 9.0 Hz, C-13).

FTIR (KBr): 3446 (OH stretch), 3055 (C-H^aromatic stretch), 2580 (broad, weak, O-H overtone), 2272 (P=O combination bands), 1630–1560 (C=C^aromatic stretch), 1460 (C-H bending of ring deform), 1400 (C-H bending), 1329 (C-O stretch or ring def), 1228 (C-O stretch), 1213 (P=O stretch), 1068 (P-O-C stretch), 1003 (C-H^arom out-of-plane).

2-Hydroxy-[1,3]dioxolo[4′,5′:4,5]benzo[1,2-e][1,2]oxaphosphinine 2-oxide P3

Yield 0.78 g (65%), brown powder. Melting point 242 °C. Find (%): C 47.87; H 3.07; P 13.71. С₉H₇O₅P. Calc. (%): С 47.81; Н 3.12; Р 13.7. GS-MS (MALDI-TOF, m/z): 226 [M]⁺; 227 [M + H]⁺; 249 [M + Na]⁺; 265 [M + K]⁺.

¹H NMR (DMSO-d₆, 600 MHz): 6.03 (s, 2H, O-CH₂-O), 6.13 (dd, J = 12.5 Hz, J = 20.0 Hz, 1H), 6.83 (s, 1H_arom), 7.00 (s, 1H_arom), 7.24 (dd, J = 12.5 Hz, J = 42.0 Hz, 1H_arom), 9.81 (br. s, 1H, OH);

³¹P NMR (DMSO-d₆, 250 MHz): 6.21;

¹³C NMR (DMSO-d₆, 150 MHz): 100.65 (d, J_CP = 7.5 Hz, C-4), 102.40 (s, OCO), 108.10 (C-8), 113.53 (d, J_CP = 168.0 Hz, C2-P), 114.54 (d, J_CP = 20.0 Hz, C-3), 141.96 (C-5), 143.60 (C-6), 147.76 (C-7), 149.40 (C-9).

FTIR (KBr): 3420 (OH stretch), 3049 (C-H^arom stretch), 2967 (C-H^aliphatic asymmetric stretch), 2845 (C-H^aliphatic sym stretch), 1623 (C=C^aromatic stretch), 1380 (C-H), 1319 (C-O stretch), 1251, 1230, 1192 (P=O stretch + C-O^aromatic stretch), 1168 (P=O stretch), 1140–1100 (P-O-C stretch), 1026, 1002 (C-O stretch), 850–800 (C-H^aromatic out-of-plane).

2-Hydroxy-7-methoxybenzo[e][1,2]oxaphosphinine 2-oxide P4 [17]

Yield 0.84 g (75%), white powder. Melting point 159 °C. ¹H NMR (DMSO-d₆, 600 MHz): 3.76 (s, 3H, OCH₃), 6.1 (dd, J = 12.5 Hz, J = 20.0 Hz, 1H), 6.68–6.76 (m, 2H_arom), 7.25–7.39 (m, 2H_arom), 9.53 (br.s, 1H, OH).

³¹P NMR (DMSO-d₆, 250 MHz): 5.91;

¹³C NMR (DMSO-d₆, 150 MHz): 56.14 (OCH₃), 104.04 (d, J_CP = 6.8 Hz, C-4), 110.45 (C-8), 113.21 (d, J_CP = 168.0 Hz, C-2), 114.80 (С-3), 131.20 (C-6), 141.80 (C-5), 153.25 (С-9), 161.67 (C-7).

FTIR (KBr): 3441 (OH stretch), 3059 (C-H_arom stretch), 2915 (C-H_aliph stretch), 2582 (overtone or P-H), 2291 (overtone P-O stretch), 1630–1500 (C=C^arom stretch), 1482 (C-H bending), 1260–1200 (C-O-C stretch), 1157 (P=O stretch), 1130–1060 (P-O-C stretch), 1034 (C-O methoxy stretch), 1000 (C-H^arom out-of-plane).

2.3. FTIR Spectroscopy

A Tensor 27 Fourier-transform infrared spectrometer (Bruker, Billerica, MA, USA) was used for recording FTIR spectra in the range of 4000–600 cm⁻¹ with a resolution of 4 cm⁻¹ and 32 scans in potassium bromide (FTIR grade) at room temperature. Spectra are available in the Supplementary Materials (Figure S1).

2.4. NMR Measurements

The ¹H, ¹³C{¹H} and ³¹P{¹H} NMR spectra were acquired on an AVANCE III 600 NMR spectrometer (Bruker, Billerica, MA, USA) equipped with a 5 mm probe. For analysis, each compound (15 mg) was dissolved in 0.7 mL of DMSO-d₆. The two-dimensional ¹H-¹³C-HSQC and ¹H-¹³C-HMBC correlation maps were obtained using a pulse program package by Bruker library. All spectra were measured at room temperature in DMSO-d₆ solution. Chemical shifts are given relative to the d-solvent signal. For ³¹P, chemical shifts are shown referenced to the H₃PO₄ signal as an external standard. All spectra were processed using TopSpin 3.2 software (Bruker, Billerica, MA, USA). Spectra are available in the Supplementary Materials (Figures S2–S19).

The electrophoretic NMR experiments (eNMR) were performed by incrementing the electric field in 8 steps between −200 V/cm and +200 V/cm. The samples were blown with a gas flow to minimize convection by Joule heating of the sample. The diffusion time Δ was 0.1 s, the gradient pulse duration δ was 2 ms, and the applied magnetic field gradient was 0.24 T/m. The phase shift values were corrected for the solvent shift to eliminate the influence of convection within the sample. A double stimulated echo pulse field gradient sequence with a diffusion time of 0.1 s and a gradient pulse duration of 2 ms was applied for 32 gradient values.

2.5. Steady-State and Time-Resolved Spectral Measurements

To measure absorption spectra in the range of 200–450 nm with a slit width of 1 nm, a 3600 Plus scanning spectrophotometer (Shimadzu, Kyoto, Japan) equipped with a double monochromator, and a one-centimeter quartz cell was used for the detection of electronic absorption spectra. The cell temperature (298 ± 0.1 K) was maintained by a TCC-100 thermoelectrically temperature-controlled cell holder (Horiba Scientific, Edison, NJ, USA). The pH meter “F20” (Mettler Toledo, Columbus, OH, USA) was used for the determination of the acidity level. The concentration of the studied solutions was 1 × 10⁻⁵ M.

The steady-state excitation and fluorescence spectra, as well as fluorescence time-resolved decays, were recorded using a Fluorolog 3–22 spectrofluorometer (Horiba Scientific, Edison, NJ, USA) with a DeltaHub module (Horiba Scientific, Edison, NJ, USA), operating in the time-correlated single photon counting, to obtain the lifetimes. In addition, spectral measurements were performed using an impulse spectrofluorometer “Fluo-rat-02-Panorama” (Lumex, Saint-Petersburg, Russia) with an external double monochromator.

The fluorescence and excitation spectra were corrected for reabsorption, detection system sensitivity, and solvent background. Standard 10 × 10 mm quartz cells were used to study aqueous solutions in L-excitation geometry. All measurements were performed at room temperature.

For the excitation of time-resolved fluorescence, we used a NanoLED N-295 pulse diode (Horiba Scientific, Edison, NJ, USA) with a maximum at a wavelength of 296 nm, and the pulse duration was <1.2 ns. The fluorescence lifetimes were measured from the decays using the deconvolution procedure in the DAS6 program (Horiba Scientific, Edison, NJ, USA).

2.6. Calculation of Dissociation Constant

The process of mono-deprotonation of the neutral form of phosphacoumarins to an anion (in a strongly acidic solution) was studied:

$$\mathrm{HL}\leftrightarrow {\mathrm{H}}^{+}+{\mathrm{L}}^{-}.$$

(1)

The dissociation constant (К_a) was determined based on a set of absorption spectra for each compound at varying pH values. A F20 pH meter (Mettler Toledo) was used to obtain acidity. The value of Кa of the reaction (1) was calculated using the non-linear Cox–Yates method [15] with the excess acidity function χ [18] in strongly acidic solutions (m* is the solvation coefficient):

$$A_i={\displaystyle \frac{A_{L^{-}}-A_{HL}}{1+(\frac{C_{H^{+}}}{K_a}){10}^{(m*\chi )}}}+A_{HL},$$

(2)

where A_i, A_HL, and A_L⁻ are the absorbance of the current solution, the neutral compounds, and their conjugate base, respectively. The optimized values of dissociation constants were determined from the least squares analysis [19] (K_i = K_a):

$${\displaystyle \sum _{i=1}^n{(A_i^{\exp }-A_i^{calc})}^2\stackrel{K_i,{\epsilon }_i}{\to }m}inimum$$

(3)

The coefficient of determination (R²) was at least 0.97 for all spectrophotometric measurements. The “±” values represent confidence intervals (P = 0.95) throughout the article. Absorption matrices and the absorbance values for single wavelength are presented in the Supporting Information. Calculations of all equilibrium constants and molar extinction coefficients were performed using the GNU Octave software package (8.4.0, Free Software Foundation, Boston, MA, USA) [20].

2.7. Electronic Structure Calculations

Ab initio simulations were carried out using the GAMESS US [21] program package on the MVS-1000M cluster of the Institute of Computational Modeling SB RAS. Quantum-chemical estimations were computed with a temperature of 298 K. The optimization of geometry was performed by density functional theory (DFT-D3) with the hybrid Perdew–Burke–Ernzerhof (PBE0) density functional [22] under Grimme’s empirical correction [23]. The full-electron cc-pVDZ [24] basis set functions were used for C, O, and H atoms. For the P atom, the CRENBL effective core potential [25] was applied for calculations. The choice of the method was according to data about the efficiency of hybrid functionals and correlation-consistent double-zeta basis set functions to conjugated systems [26,27]. The Gibbs free energy of dissociation (∆G^aq) was calculated taking into account three parts: energy in the gas phase (∆G^gas), the free energy of the solvation (δ∆G^solv), and zero-point energy with thermal correction (∆E^ZPE) [28]:

logK^calc = −∆G^aq/2.303RT,

(4)

∆G^aq = ∆G^gas + δ∆G^solv + ∆E^zpe + (ΔP)ΔG^0→* + E^corr,

(5)

E^corr = −(ΔP)RTln([H₂O]).

(6)

Here, E^corr is the change in free energy associated with the movement of the solvent from of the standard-state solution phase concentration of 1 M to the standard state of the pure liquid at 55.34 M [29]; ΔG^0→* is the correction relates with condensation of one mole from the gas phase at the standard condition of one atmosphere (24.46 L/mol) to solution at the standard state of one mole per liter [30]; ΔG^0→* = RTln(24.46 L/mol) = 1.89 kcal/mol at 298 K; ΔP is the difference between the number of species during the dissociation reaction. The zero-point energy corrections were estimated from the results of normal mode analysis of the complexes in the gas phase. The solvation long-range effects were evaluated using the SMD solvation model [31].

The maximum wavelengths of ultraviolet absorption spectra of compounds were reproduced from the vertical excitation energies for the first 15 singlet excited states by Time-Dependent Density Functional Theory (TD-DFT) [32].

3. Results and Discussion

3.1. Spectral and Photophysical Properties

The synthesized compounds demonstrated solubility in aqueous solutions; thus, the spectral parameters described in this section were measured in aqueous media. All compounds exhibit a set of π-π* absorption transitions and rather weak emission in the ultraviolet and visible spectral regions. The emission and excitation spectra are shown in Figure 1a. The maxima of the most intense absorption bands lie in the range of 231–319 nm (Figure 1b), and the maxima of the emission spectra lie in the range of 300–394 nm. From Figure 1 and

Table 1. Spectral and photophysical parameters of P1–P4 in water.

Compound	${\mathbf{\epsilon }}_{\mathbf{abs}}\left(\mathbf{\lambda }\right)$, a ${10}^3\mathbf{L}{\mathbf{mol}}^{-1}{\mathbf{cm}}^{-1}$	${\mathbf{\lambda }}_{\mathbf{abs}}^{\mathbf{max}}$, b nm	${\mathbf{\lambda }}_{\mathbf{fl}}^{\mathbf{max}}$, c nm	Stokes Shift, nm	τfl, d ns	ϕfl, e %
P1	14.1 ± 0.7 (273)	273, 216	385	112	0.73 ± 0.05	12.6 ± 1.7
P2	9.2 ± 0.1 (310)	341, 310, 231	365	55	0.28 ± 0.03	4.6 ± 0.5
P3	8.4 ± 0.1 (319)	319, 280, 234, 216	394	75	0.24 ± 0.02	3.9 ± 0.4
P4	12.6 ± 0.6 (276)	276, 209	300	24	0.06 ± 0.02	1.8 ± 0.3

^a ${\mathsf{\epsilon }}_{\mathrm{abs}}^{\max }\left(\lambda \right)$ is the molar extinction coefficient at a given wavelength (nm) indicated in parentheses. ^b ${\mathsf{\lambda }}_{\mathrm{abs}}^{\max }$ are the visible maxima of the main absorption bands. The longest wavelength and most intense absorption band used to calculate the Stokes shift are shown in bold. More detailed information can be found in Figure S20. ^c ${\mathsf{\lambda }}_{\mathrm{fl}}^{\max }$ is the maximum of the fluorescence spectrum. ^d τ_fl is the fluorescence lifetime measured at the emission maximum when excited at 296 nm. ^e ϕ_fl is the fluorescence quantum yield calculated as the ratio of the measured fluorescence lifetime τ_fl and the radiative lifetime τ_rad; τ_rad was determined from the absorption and fluorescence spectra using the Strickler-Berg equation [33].

, it can be seen that the spectral (absorption, excitation, and emission maxima) and photophysical (emission lifetime, quantum yield, Stokes shift) properties of the synthesized compounds P1–P4 differ significantly and depend on the number of methyl groups. The greatest value of Stokes shift is observed for P1, which contains the largest number of methyl groups. The emission lifetimes of P1–P4 were obtained from the time-resolved decays after pulsed excitation at 296 nm (Figure 2a). The emission lifetimes of P1–P4 (0.06–0.73 ns, see Table 1 for details) correspond to the fluorescence, the radiative transition between singlet states. The fluorescence quantum yield for all compounds is low, ranging from 2 to 13% (Figure 2b, Table 1). Compound P1 exhibits the highest emission efficiency (12.6%) and the longest lifetime (0.73 ns) among the synthesized phosphacoumarins, while compound P4, on the contrary, exhibits the shortest emission lifetime (0.06 ns) and a very low fluorescence quantum yield (1.8%).

3.2. Acidic Properties

Since the synthesized compounds are derivatives of orthophosphoric acid, they should exhibit acidic capacity. Typical values of the dissociation constants for phosphoric acids and their organic derivatives are in the range of 2–3 logarithmic units (for orthophosphoric acid pK_a1 = 2.15 [34]). Initially, we expected the acidity of new compounds to be close to the typical value. However, the absorption spectra of the studied compounds do not change in the pH range of 2–14 and above in alkali media. The significant spectral changes were detected only in strongly acidic media (Figure 1b,c).

Because compounds P1–P4 were unstable over time in the sulfuric acid, all measurements were performed in hydrochloric media. The acidic character is so pronounced that even in concentrated HCl, the dissociation cannot be completely suppressed. The inability to obtain the molar excitation coefficients for the neutral form imposes limits on the accuracy of measurements. Statistical analysis of the obtained data shows that all compounds are strong acids with negative values of pK_a. The determined acidic parameters are presented in

Table 2. Acidic parameters.

Compound	logεanion	logεneutral	pKa	m* (Solvation Coefficient)
P1 (292 nm)	3.78 ± 0.01	4.02 ± 0.01	0.19 ± 0.10	-
P2 (251 nm)	3.51 ± 0.01	3.78 ± 0.01	−1.30 ± 0.15	0.35 ± 0.10
P3 (325 nm)	3.40 ± 0.01	3.73 ± 0.01	−1.26 ± 0.12	0.45 ± 0.12
P4 (298 nm)	4.13 ± 0.01	4.33 ± 0.01	−0.70 ± 0.06	0.12 ± 0.02

.

In addition, the dissociation of the studied compounds in water was obtained by electrophoretic NMR analysis. Using P2 as an example, the electrophoretic behavior was examined in heavy water media (H₂O/D₂O 1:4 vol.) for greater signal stability and accuracy in determining chemical shifts. The appearance of a phase shift in signals in an applied electric field (Figure 3) indicates the presence of charge carriers (ions). The electrophoretic mobility of the aqueous solution was 2.2 × 10⁻⁸ m²/Vs with an effective anion charge of −0.93. The diffusion coefficient by DOSY measurements was 6 × 10⁻¹⁰ m²/s.

The obtained parameters correspond to the strong acids such as trifluoroacetic acid (pK_a = 0.23 [35]) or chloric acid HClO₃ (pK_a ≈ −2.7 [36]). In fact, we can see a decrease in the dissociation constants from 2.15 (pure H₃PO₄) to −1.3 (P2) logarithmic units. This unexpected result emphasizes the role of aromatic substituted groups, which are highly delocalized π-conjugated systems. The problem of explaining acidity by π-delocalization lies in the nature of the O-H bond, which has a σ-electron structure. Electronic effects should have a significant impact on the central phosphorus atom, which in turn affects the O-H group. A quantitative assessment of the state of the phosphorus atom can be obtained using ³¹P-NMR spectra (Figures S3, S6, S11 and S16), but the chemical shifts have typical values (around zero ppm). Aside from electronic effects, changes in acidic properties may be related to solvation effects. Indeed, as donor atoms, phosphorus and oxygen can form a solvation shell of explicit water molecules around their places. This conclusion is difficult to confirm by experimental methods but can be estimated by DFT modeling.

3.3. DFT Modeling

Modeling of spectral parameters by the TD-DFT approach allows us to identify the nature of the observed electronic transitions in absorption spectra. Herein, we estimated the spectra that were recorded in water media, i.e., the anionic form of compounds P1–P4.

Table 3. Results of TD-DFT estimations.

Compound	λmaxexp	λmaxcalc	Oscillator Strength
P1	273	253	0.188
P2	231	251	0.319
P3	234	216	0.178
P4	276	281	0.077

presents a comparison of the theoretically calculated and experimental parameters. All transitions correspond to intracycle π-π* electronic transitions.

To describe the acidic properties of the studied compounds, theoretical values of the dissociation constants were calculated. Since the solvation effects are one of the most possible factors leading to anomalous acidity, different solvation forms of phosphacoumarins and protons were involved. Theoretical modeling of the dissociation process was performed using explicit solvation of the hydrogen ion in the form H·(H₂O)_n⁺, where n was set to 4. Explicit solvation of the compounds was approximated by one water molecule, which connected with the free oxygen atom of the compound by a hydrogen bond (Figure 4, top). The thermodynamic cycle for the DFT calculation is shown in Scheme 3. The values of the theoretically calculated equilibrium constants and their energy contributions are summarized in

Table 4. Energy contribution (kJ/mol) and values of the theoretically calculated acidity constants (log units).

Compound	δΔGsolv	ΔEZPE	ΔGaq.	logKcalc	logKexp
P1	−41.5	40.3	−1.2	0.22	−0.19
P2	−57.0	46.0	−11.1	1.94	1.3
P3	−53.2	40.7	−12.5	2.19	1.26
P4	−53.4	40.3	−14.1	2.48	0.7

.

As we can see, the inclusion of explicit solvation in the computational protocol leads to the prediction of acidic properties within one logarithmic unit. At the same time, the solvated form of the proton was necessary to achieve agreement between theoretical computations and the experimental data obtained. The formation of solvated species of P1–P4 explains their unexpected acidity, but only within the first-order approximation. An important feature of the optimized structures is the difference in the P–O bond distances between the anionic and neutral forms. In the anionic form, the P–O bond is significantly weaker than in the neutral form; the bond order varies from 0.67 to 0.56. This fact may be related to the anomalous acidity value, but further research is required for a more specific explanation (Figure 4, bottom).

The cyclic fragment containing phosphorus atoms is larger than that of the neutral forms. In addition, the P-O bond order in the neutral form is significantly smaller than the corresponding order in H₃PO₄ (all three P-O bonds have an order of 1.16, calculated at the same level of theory cc-pVDZ/CRENBL ECP/PBE0/SMD). Thus, two factors reducing the stability of the neutral form are the weak P-O bonds and the larger size of the phosphorus cycle compared to the anionic form.

4. Conclusions

Four water-soluble 2-hydroxybenzo[e][1,2]oxaphosphinine 2-oxides, P1–P4, were prepared by a novel method via cascade reactions of (2-ethoxyvinyl)phosphonic dichloride with phenol or naphthol derivatives and characterized using a combination of physicochemical methods and DFT modeling. The structures of two compounds were confirmed by NMR, FTIR, GC-MS, and elemental analysis. These compounds exhibit UV-violet emission with a fluorescence quantum yield of 2–13% and possess diverse spectral and acid-base properties.

The acid-base parameters were determined by electron absorption spectroscopy using the non-linear Cox–Yates method in strongly acidic media. Unusually high acidities were observed: in aqueous solution, the compounds predominantly exist as singly deprotonated anionic forms, and dissociation occurs even under strongly acidic conditions. The measured pKₐ values were 0.2, −1.3, −1.26, and −0.7 for P1, P2, P3, and P4, respectively. Electrophoretic NMR confirmed that these anionic forms represent the monodeprotonated forms of neutral compounds and remain dominant throughout the studied pH range.

DFT calculations of the acid–base properties using an explicit solvation model for both the compounds and the proton achieved the best agreement with experiment when the proton was represented as an H₉O₄⁺ cluster and the compounds were modeled in their monosolvated neutral form. Within this model, the predicted acid–base parameters agreed with experimental values within ±0.5–1.5 log units.

The unique result obtained in this study: the unexpectedly high acidity of phosphate coumarins P1–P4 is a new scientific fact, allowing us to confirm that they are among the most active organophosphorus acids currently known. This study will serve as an impetus for the further development of methods for synthesizing a wide range of phosphocoumaric acids containing electron-withdrawing substituents in the aromatic ring, the identification of patterns in the influence of these substituents on acidity, and the creation of unique new types of photoactive phosphorus-containing acids.