Complexation of uranium(VI) with glyphosate in acidic to basic aqueous solutions

Abstract

The complexation of uranium(VI) with glyphosate (N-(phosphonomethyl)glycine) has been investigated across acidic to basic aqueous solutions using multiple analytical techniques. Eleven uranium(VI)–glyphosate complexes, including 1 : 1, 1 : 2, and 2 : 2 species, were identified under experimental conditions. Thermodynamic parameters—formation constants, as well as enthalpies and entropies of complexation—were determined through potentiometry and calorimetry. The results indicate that the complexation between uranium(VI) and glyphosate (in both L and HL forms) is endothermic and driven by a favorable entropy change. These thermodynamic data, in conjunction with structural information from ³¹P NMR spectroscopy and theoretical validation from DFT calculations, help to identify the coordination modes in the uranium(VI)–glyphosate complexes. In addition, the speciation of glyphosate under environmental conditions, both in the absence and presence of uranium(VI), is discussed.

---

Introduction

Glyphosate (N-(phosphonomethyl)glycine), the most widely used herbicide worldwide, typically exists in a zwitterionic form. Its introduction has been regarded as a revolutionary advancement in agriculture. The C–P bond in glyphosate is highly stable, rendering the compound resistant to hydrolysis and photolysis in soil environments.¹ As a result, glyphosate is considered “relatively persistent” in certain soils, with a half-life ranging from less than a week to several months, depending on soil adsorption capacity and microbial activity.² Persistence increases in sandy soils and cooler climates; for example, residues were detected in Sweden up to three years after application.³

The glyphosate molecule contains three functional groups capable of metal coordination: the amino, carboxylate, and phosphonate groups. From a coordination chemistry perspective, glyphosate acts as a highly effective ligand for various metal ions and can inhibit metalloenzymes by blocking their active metal centers. Extensive studies have been conducted on the coordination behavior of glyphosate with di- and trivalent metal ions,^4–8 with most research focused on determining stability constants and elucidating complex structures.^9–12 In contrast, few studies have addressed the complexation between glyphosate and uranium(VI). When uranium enters soil systems, it can interact with glyphosate, altering uranium speciation and consequently influencing the transport behavior of both uranium and glyphosate in the environment.

Only one study has investigated the complexation of glyphosate with uranium(VI) in aqueous solution. Using multinuclear NMR spectroscopy, Zoltán Szabó et al.² identified three uranium(VI)–glyphosate complexes (UO₂L⁻, UO₂HL₂³⁻, and UO₂H₂L₃⁵⁻) within the pH range of 7–10 at T = –5 °C. The stoichiometries and stability constants of these complexes were derived from integral values of coordinated and free ligands in ¹H, ³¹P, and ¹⁷O NMR spectra, with reported values of log β = 13.65 for UO₂L⁻, log β = 25.62 for UO₂HL₂³⁻, and log β = 36.62 for UO₂H₂L₃⁵⁻. However, complex formation in acidic to neutral solutions was not examined in that study. Glyphosate is a strongly basic ligand and protonated species are unlikely under alkaline conditions. The behavior across a broader pH range remains unexplored.

To date, no thermodynamic data have been reported in the literature on the complexation between uranium(VI) and glyphosate from acidic to basic solutions at 25 °C. This lack of data hinders the interpretation and prediction of UO₂²⁺ and glyphosate speciation, as well as its transport behavior in environmental systems. Studying U(VI)–glyphosate complexation across the full pH range not only aids in understanding its environmental fate but also provides fundamental insights into the coordination chemistry of U(VI) with ligands containing amide, carboxylate, and phosphonate functional groups simultaneously.

Therefore, the aim of this study is to identify the uranium(VI)–glyphosate complexes formed across acidic to basic aqueous solutions. The stability constants and enthalpies of complexation are determined by potentiometry and calorimetry. Additionally, NMR spectroscopy is employed to assist in structural characterization. The speciation of glyphosate both in the absence and presence of uranium(VI) under environmental conditions is also discussed.

Experimental section

Chemicals

All chemicals were of reagent grade or higher. Distilled and deionized water was used in preparations of all the solutions. The stock solutions of uranyl chlorate were prepared by dissolving uranium trioxide (UO₃, >98%, Aldrich) in concentrated hydrochloric acid. The concentration of uranium in the stock solution was determined by the gravimetric method.¹³ Gran's potentiometric method¹⁴ was used to determine the concentration of hydrochloric acid in the stock solutions. The stock solutions of glyphosate were prepared by dissolving glyphosate solid (>98%, Aldrich) in solution. The concentration of glyphosate in the stock solution was determined by potentiometric titration. The ionic strength of all the solutions used in potentiometry and calorimetry was adjusted to 0.5 mol dm⁻³ at 25 °C by adding appropriate amounts of sodium chloride as the background electrolyte.

Potentiometry

Potentiometric titrations were performed to determine the protonation constants of glyphosate and the stability constants of uranium(VI)–glyphosate complexes. Details of the titration setup and procedure have been given elsewhere.¹⁵ In this work, the original inner solution of the glass electrode (3 M KCl) was replaced with 1 M sodium chloride to reduce the electrode junction potential. Electromotive force (EMF, in millivolts) was measured using a potentiometric titrator (888 Titrando, Metrohm) equipped with a combination pH electrode (6.0259.100 Unitrode, Metrohm). In acidic and basic regions, E can be obtained as in eqn (1) and (2), respectively,

E = E° + RT/F ln[H⁺] + γ_H[H⁺] (1)

E = E° + RT/F ln k_w − RT/F ln[OH⁻] + γ_OH[OH⁻] (2)

where R is the gas constant, F is the Faraday constant, and T is the temperature. k_w is the ionic product of water (= [H⁺][OH⁻]). The terms γ_H[H⁺] and γ_OH[OH⁻] are the electrode junction potentials for the hydrogen and hydroxide ions, respectively. Prior to each protonation or complexation titration, an acid/base titration with standard hydrochloric acid and sodium hydroxide solutions was performed to obtain the parameters E°, γ_H, and γ_OH. These parameters allowed the calculation of hydrogen ion concentrations from the electrode potential in the subsequent titration.

Glyphosate molecules possess multiple coordination sites, exhibit strong alkalinity and high molecular flexibility, and are capable of forming various complexes with U(VI). To detect all the uranium(VI)–glyphosate complexes that can be formed, multiple titrations need to be performed by titrating a solution containing U(VI) and glyphosate (C_L/C_U = 2.2 and 1) with NaOH. Uranium complexes and/or uranium hydrolyzed products have low solubility. When C⁰_U > 1 mmol L⁻¹, the solution system remains clear only under neutral (pH = 6.7–7.5) conditions, and when C⁰_U ≤ 1 mmol L⁻¹, the solution system remains clear throughout the pH range (pH = 3–10). In this work, C⁰_U < 1 mmol L⁻¹. Multiple titrations were conducted with solutions of different concentrations (C_L, C_H, and C_U). Approximately 60 data points were collected in each titration. The protonation constants of glyphosate and the stability constants of uranium(VI)–glyphosate complexes were calculated using the program HyperQuad 2000.¹⁶

Calorimetry

Calorimetry was used to determine the enthalpy of glyphosate protonation and complexation with uranium(VI). It was carried out using an isoperibol solution calorimeter (TAM IV, Calorimetry Sciences Corp). Due to the poor solubility of the complexation reaction of glyphosate with uranium(VI) under the acidic and alkaline conditions, only the enthalpy data of some species (UO₂HL(aq), UO₂L⁻, UO₂H₂L₂²⁻, and UO₂L₂⁴⁻) under the neutral conditions could be measured in this experiment. The initial cup solutions (0.75 cm³ at 25 °C) usually contained 0.005–0.007 mol dm⁻³ UO₂Cl₂ and 0.01–0.014 mol dm⁻³ glyphosate solution, C_L/C_U = 2. The pH of the solution is controlled between 6.7 and 7.5 with HCl or NaOH. The titrant was 0.021 mol dm⁻³ NaOH, and V = 0.25 mL. At least three titrations with different concentrations of UO₂Cl₂ and glyphosate were conducted. For each titration, n experimental values of the total heat produced in the reaction vessel (Q_ex,j, where j = 1 to n, usually n = 50–70) were calculated as a function of the volume of the titrant added. These values were corrected for the heat of dilution of the titrant (Q_dil,j), which was determined in separate runs. The net reaction heat at the jth point (Q_r,j) was obtained from the difference: Q_r,j = Q_ex,j − Q_dil,j. These data, in conjunction with the protonation and complexation constants obtained by potentiometry, were used to calculate the enthalpy of protonation and complexation using the computer program HypDeltaH.¹⁷

NMR

³¹P NMR experiments were performed by using a 400 MHz Bruker Avance spectrometer. A D₂O solution, containing 40 mM U(VI) and 80 mM glyphosate, was prepared for the NMR experiment. The pH of the solution was adjusted by adding NaOH or HCl. All chemical shifts were measured with reference to a solution of phosphoric acid in D₂O.

DFT calculations

To theoretically verify the possible structures of the uranium(VI)–glyphosate complexes in solution, we performed density functional theory (DFT) calculations. All quantum chemical calculations were carried out using the Gaussian 16 software package. The hybrid PBE0 functional was employed, incorporating Grimme's D3(BJ) dispersion correction to accurately describe non-covalent interactions. For the heavy element uranium, the Stuttgart–Dresden (SDD) effective core potential and its associated basis set were used to account for scalar relativistic effects. For light elements (H, C, N, O, P, etc.), the 6-31+G(d,p) basis set was adopted. Solvation effects in an aqueous environment were simulated using the solvation model based on density (SMD). Full geometry optimizations were first performed for the five proposed complex structures (UO₂L⁻, UO₂L₂⁴⁻, UO₂HL⁰, UO₂H₂L₂²⁻, and UO₂H₂L⁺), followed by frequency analysis at the same theoretical level to ensure that the obtained structures corresponded to true minima on the potential energy surface (no imaginary frequencies). Based on the optimized structures, single-point energy calculations were conducted using the higher-level 6-311+G(d,p) basis set (with SDD retained for uranium) to obtain high-quality wavefunctions. Finally, wavefunction analysis was performed using the Multiwfn program to calculate Mayer bond orders (MBO) to assess the strength of key coordination bonds.

Results and discussion

Protonation of glyphosate

The thermodynamic parameters of glyphosate protonation are given in Table 1. The protonation constants from this work are in good agreement with previous values obtained in the 1.0 mol L⁻¹ NaCl ionic media.¹⁸ The enthalpies of protonation indicate that, stepwise, the three steps of protonation are all exothermic, while the second step is slightly endothermic. The first two steps are favored by both enthalpy and entropy, while the third step is only enthalpy-driven.

Table 1 The thermodynamic parameters of glyphosate protonation, T = 25 °C and I = 0.5 mol L⁻¹ NaCl

Reaction	log Ka	ΔH	ΔS	ΔG
		kJ mol^−1	J (mol K)^−1	kJ mol^−1
a Ref. 18, T = 25 °C and I = 1.0 mol L^−1 NaCl.
H^+ + L^3− = HL^2−	9.85 ± 0.06	−32.20 ± 0.30	81.2 ± 3.1	−56.2 ± 1.1
	9.76^a
2H^+ + L^3− = H2L^−	15.16 ± 0.12	−34.30 ± 0.30	175.8 ± 3.1	−86.5 ± 1.6
	14.84^a
3H^+ + L^3− = H3L	17.32 ± 0.15	−59.2 ± 2.00	133.6 ± 7.5	−98.8 ± 2.1
	16.61^a

---

The stability constants of uranium(VI)–glyphosate complexes

Glyphosate could form a variety of complexes with U(VI), including the protonated complexes MH_jL_k in acidic solutions and the mixed hydroxyl complexes, M(OH)_jL_k or MH_−jL_k, in basic solutions. Besides, polynuclear species could form when the ligand-to-metal ratio is low. To obtain stability constants for such a large number of complexes, it is necessary to perform multiple titrations under different conditions. In this work, a total of 11 potentiometric titrations with varying concentrations of U(VI), glyphosate, and H⁺ have been conducted. Analysis of all of the titrations was performed using HyperQuad 2008 to obtain the stability constants of the complexes. Fig. 1 shows the fitting of a representative titration (experimental and calculated pC_H), as well as the percentages of U(VI) species in the titration. The calculated stability constants of the uranium(VI)–glyphosate complexes are summarized in Table 2. As shown in Table 2, 11 uranium(VI)–glyphosate complexes are detected in the experiment. It was observed that U(VI) forms a number of 1 : 1, 1 : 2, and 2 : 2 complexes with glyphosate in solution. The complexes range from highly protonated (e.g., UO₂H₂L, and UO₂H₂L₂) in strongly acidic solutions to unprotonated (e.g., UO₂L and UO₂L₂) in neutral solutions, mixed hydroxyl glyphosate complexes (e.g., UO₂(OH)₂L and UO₂(OH)₂L) in basic solutions, and polynuclear complexes (e.g., (UO₂)₂HL₂ and (UO₂)₂L₂). Using multinuclear NMR spectroscopy, Zoltán Szabó detected three uranium(VI)–glyphosate complexes (UO₂L, UO₂HL₂, and UO₂H₂L₃) in the pH range of 7–10 at T = −5 °C. The last two complexes have not been observed in our titrations, even if the ligand-to-metal ratio increased to 4. The stability constants for uranium(VI)–glyphosate complexes in Table 2 formed at pH = 3–10 and T = 25 °C are the first such data that have been experimentally determined.

Fig. 1 Potentiometric titration of uranium(VI)–glyphosate complexation. I = 0.5 M NaCl, T = 25 °C, V₀ = 18 mL, and titrant 6.924 mM NaOH. C⁰_U = 0.1 mM, C⁰_L = 0.22 mM, and C⁰_H = 0.604 mM. (◇) Experimental pC_H, left y axis. (dashed line) Calculated pC_H, left y axis. (solid lines) Percentages of U(VI) species, right y axis. 1: UO₂H₂L⁺, 2: UO₂HL⁰, 3: UO₂L⁻, 4: UO₂H₂L₂²⁻, 5: UO₂L₂⁴⁻, 6: UO₂(OH)₂L₂⁶⁻, and 7: (UO₂)₃(OH)⁷⁻.

Table 2 The species and stability constants of uranium(VI)–glyphosate complexation (I = 0.5 M NaCl and T = 25 °C)

MiHjLk			Reaction	log β	Ref.
i	j	k
1	2	1	UO2^2+ + 2H^+ + L^3− = UO2H2L^+	23.53 ± 0.07
1	1	1	UO2^2+ + H^+ + L^3− = UO2HL(aq)	19.90 ± 0.07
1	0	1	UO2^2+ + L^3− = UO2L^−	13.71 ± 0.06
				13.65 ± 0.2	2
1	2	2	UO2^2+ + 2H^+ + 2L^3− = UO2H2L2^2−	32.95 ± 0.20
1	0	2	UO2^2+ + 2L^3− = UO2L2^4−	19.41 ± 0.09
1	−2	2	UO2^2+ + 2L^3− + 2H2O = UO2(OH)2L2^6− + 2H^+	−0.49 ± 0.11
1	−2	1	UO2^2+ + L^3− + 2H2O = UO2(OH)2L^3− + 2H^+	−3.1 ± 0.02
2	1	2	2UO2^2+ + H^+ + 2L^3− = (UO2)2HL2^−	38.79 ± 0.03
2	0	2	2UO2^2+ + 2L^3− = (UO2)2L2^2−	31.27 ± 0.10
2	−1	2	2UO2^2+ + 2L^3− + H2O = (UO2)2(OH)L2^3− + H^+	25.26 ± 0.02
2	−2	2	2UO2^2+ + 2L^3− + 2H2O = (UO2)2(OH)2L2^4− + 2H^+	17.22 ± 0.03

---

Fig. 2 illustrates the correlation between the stability constants (log β) of the ML-type complexes formed by glyphosate with UO₂²⁺ and other metal ions (Fe³⁺, Fe²⁺, Ca²⁺, Ni²⁺, Co²⁺, Mn²⁺, and La³⁺) and their ionic potential (Z/r). It should be noted that the log β data for the metal ions other than UO₂²⁺ were taken from the literature (experimental conditions: I = 0.1 M KNO₃ and T = 25 °C),¹⁰ while the data for UO₂²⁺ in this study were obtained under conditions of I = 0.5 M NaCl and T = 25 °C. Although differences in ionic strength and background electrolyte may have some influence on the absolute values, the figure still clearly reveals important qualitative trends. As shown, the log β values of the metal ions other than those of UO₂²⁺ exhibit a good linear positive correlation with Z/r (the fitting equation based on all eight data points is log β = 4.52 × (Z/r) − 5.57 and R² = 0.978). This trend indicates that for these ions, the complexation with glyphosate is predominantly governed by classical electrostatic interactions.^19–21 For the UO₂²⁺ data point (13.71), the residual in this model is +0.996, and its 95% prediction interval is [11.00, 14.43]. The observed value falls within this interval (t-test, p = 0.206), which statistically confirms the reliability of the stability constant for the UO₂²⁺–glyphosate complex measured in this work and shows that its variation trend is essentially consistent with the classical electrostatic model. The minor and statistically non-significant deviation observed for the UO₂²⁺ point may be attributed to the unique properties of UO₂²⁺ as a linear actinyl hard acid ion.

Fig. 2 The correlation between the stability constant (log β) of ML (glyphosate) complexes and their ionic potential (Z/r).

Enthalpy of complexation

As mentioned earlier, when C⁰_U > 1 mmol L⁻¹, the solution system remains clear only under neutral conditions (pH = 6.7–7.5). Based on the stability constants measured by potentiometric titration, there exist four U(VI) species (UO₂HL, UO₂L, UO₂H₂L₂, and UO₂L₂) within this pH range. Experimental calorimetric titrations were performed at pH = 6.7–7.5 to measure the enthalpy of the four U(VI) species. Experimental data of a calorimetric titration are shown in Fig. 3: (a) the heat flux density in the titration system as a function of time and (b) the total reaction heat Q_r and U(VI) species in the titration system as a function of the titrant volume added. These data were used, in conjunction with the protonation constants and enthalpy of glyphosate and the stability constants of the four uranium(VI)–glyphosate complexes, to calculate the enthalpies of complexation. The values of enthalpy and entropy are summarized in Table 3. Using the enthalpies and formation constants of the four uranium(VI)–glyphosate complexes in Tables 3 and 2, a curve simulating the calorimetric titration was calculated and is shown in Fig. 3(b) (the line). The good agreement between the curve and the experimental points confirms the mutual consistency of the calorimetric and potentiometric data on the complexation (Tables 3 and 2) as well as the reliability of the data on the protonation of glyphosate (Table 1).

Fig. 3 Experimental data of a calorimetric titration of uranium(VI)–glyphosate complexation (I = 0.5 M NaCl and T = 25 °C): (a) the heat flux density in the titration system as a function of time; (b) the total reaction heat Q_r and U(VI) species in the titration system as a function of the titrant volume added (left y axis: solid square, experimental heat of reaction (Q_r); solid line, calculated heat of reaction. Right y axis: black, UO₂HL(aq); red, UO₂L; green, UO₂H₂L₂; blue, UO₂L₂). Titrant: 21 mM NaOH (50 additions of 0.005 mL each). Initial solution: V₀ = 0.75 mL, C⁰_U = 5.25 mM, C⁰_L = 10.5 mM, and C⁰_H = 7 mM.

Table 3 The enthalpies and entropies of part of uranium(VI)–glyphosate complexes

Reaction	log β	ΔH	ΔS	ΔG
		kJ mol^−1	J (mol K)^−1	kJ mol^−1
UO2^2+ + H^+ + L^3− = UO2HL^0	19.90 ± 0.07	1.0 ± 0.4	384.3 ± 2.7	−113.5 ± 0.4
UO2^2+ + L^3− = UO2L^−	13.71 ± 0.06	37.0 ± 0.3	386.6 ± 2.2	−78.2 ± 0.3
UO2^2+ + 2H^+ + 2L^3− = UO2H2L2^2−	32.95 ± 0.20	−21.0 ± 0.8	560.3 ± 6.5	−188.0 ± 1.1
UO2^2+ + 2L^3− = UO2L2^4−	19.41 ± 0.09	34.0 ± 0.6	485.7 ± 3.7	−110.7 ± 0.5

---

Comparison of the enthalpy and entropy data could provide insight into the energetics of complexation and the nature of the complexes. The enthalpy and entropy of complexation for the complexes with L³⁻ (UO₂²⁺ + L³⁻ = UO₂L⁻ and UO₂²⁺ + 2L³⁻ = UO₂L₂⁴⁻) are both positive: UO₂L (ΔH = 37 kJ mol⁻¹ and ΔS = 386 J (mol K)⁻¹) and UO₂L₂ (ΔH = 34 kJ mol⁻¹ and ΔS = 485 J (mol K)⁻¹), consistent with the formation of inner-sphere complexes between hard acid and hard base.²² The positive enthalpy and entropy values for uranium(VI)–glyphosate (L³⁻) complexation can be interpreted to be due to the strong contribution of dehydration of the interacting cation and anion.²³ The energy required to dehydrate UO₂²⁺ and glyphosate (L³⁻) exceeds the energy released when UO₂²⁺ and L³⁻ combine, resulting in a positive enthalpy. The large positive entropy essentially reflects the increase in disorder when a number of water molecules are released from UO₂²⁺ and L³⁻. Similarly, the enthalpy and entropy of complexation for the complexes with HL²⁻ (UO₂²⁺ + HL²⁻ = UO₂HL, UO₂²⁺ + 2HL²⁻ = UO₂H₂L₂²⁻) are both positive: UO₂HL (ΔH = 33 kJ mol⁻¹ and ΔS = 300 J (mol K)⁻¹) and UO₂H₂L₂ (ΔH = 43 kJ mol⁻¹ and ΔS = 399 J (mol K)⁻¹), calculated from the parameters for the reaction (UO₂²⁺ + H⁺ + L³⁻ = UO₂HL⁰ and UO₂²⁺ + 2H⁺ + 2L³⁻ = UO₂H₂L₂²⁻) in Table 3 and the parameters for the reaction (H⁺ + L³⁻ = HL²⁻) in Table 1. So, we assume that UO₂HL and UO₂H₂L₂ are also inner-sphere complexes.

Coordination modes in the uranium(VI)–glyphosate complexes

Since there are few structural data in the literature on the uranium(VI)–glyphosate complexes in aqueous solutions, the coordination modes in these complexes in solution remain undefined. We have made efforts to obtain structural information on the uranium(VI)–glyphosate complexes by X-ray crystallography. However, attempts to prepare single crystals of the uranium(VI)–glyphosate complexes were unsuccessful. In this work, we attempt to use the thermodynamic parameters from potentiometry, in conjunction with the ³¹P NMR data, to shed light on the coordination modes in the uranium(VI)–glyphosate complexes.

UO₂L

Carboxylate/phosphonate-containing ligands oxydiacetic acid (ODA), iminodiacetic acid (IDA), and glyphosate share similar molecular structures. When coordinating with U(VI), potential donor atoms (carboxylate oxygen, phosphonate oxygen, amino nitrogen, and ether oxygen) can all participate in coordination. For UO₂L-type complexes, if their coordination modes are identical, their stability constants should correlate with the overall basicity of the ligand, typically measured by the sum of its protonation constants (ΣpK_a).^24,25 To verify this correlation and deduce the coordination mode of glyphosate, we compared the data for the UO₂²⁺–glyphosate complex obtained in this study (experimental conditions: I = 0.5 M NaCl and T = 25 °C) with the literature data for the UO₂²⁺–ODA²⁶ and UO₂²⁺–IDA²⁷ systems (experimental conditions in the literature: I = 1.0 M NaClO₄ and T = 25 °C). As shown in Fig. 4, despite slight differences in the ionic strength (I) and background electrolyte (NaCl vs. NaClO₄) used across the systems, which may have some influence on the absolute values of the stability constants, a strong linear relationship is still observed between the log β values of the three types of complexes and the ΣpK_a of their corresponding ligands. This robust qualitative trend suggests that the three UO₂L complexes most likely share the same coordination mode. It is known from previous studies that both ODA²⁶ and IDA^27,28 adopt a tridentate coordination mode with UO₂²⁺, involving two carboxylate oxygen atoms plus either an ether oxygen atom (for ODA) or an amino nitrogen atom (for IDA) bound to the uranium center. Based on the consistency revealed by the aforementioned linear correlation, we reasonably infer that in the UO₂L (glyphosate) complex, glyphosate also coordinates in a tridentate mode, specifically through one carboxylate oxygen, one phosphonate oxygen, and the amino nitrogen atom to the U(VI) center. This proposed structure is depicted in Fig. 7(a).

Fig. 4 The stability constant log β of UO₂L complexes for ODA, IDA, and glyphosate ligands vs. ΣpK_a of the protonation constants for the ligands.

UO₂L₂

It is usually more difficult to reveal the coordination mode in the consecutive second and/or third complexes merely by evaluation of the thermodynamic parameters. However, a comparison of the stepwise formation constants among U(VI)/ODA (log K₁ = 5.01 and log K₂ = 2.63),²⁶ U(VI)/IDA (log K₁ = 8.75 and log K₂ = 7.67),²⁷ and U(VI)/glyphosate complexes (log K₁ = 13.71 and log K₂ = 5.7) may provide insight into the coordination modes in the complexes. For the U(VI)/ODA system, the second complex is more than 2 orders of magnitude weaker than the first complex, and for the U(VI)/IDA system, it is 1 order of magnitude. However, for the U(VI)/glyphosate system, the second complex is 8 orders of magnitude weaker than the first complex. This rather large decrease in the stability of the second U(VI)/glyphosate complex may imply that the second glyphosate ligand is probably not tridentate as the first glyphosate, but bidentate exhibits coordination: one phosphonate oxygen and one carboxylate oxygen coordinate with U, while N is not involved in coordination. The greater electrostatic repulsion (glyphosate: UO₂L⁻ + L³⁻ = UO₂L₂⁴⁻ and ODA/IDA: UO₂L⁰ + L²⁻ = UO₂L₂²⁻) and steric hindrance of PO₃²⁻ resulting from the crowdedness of two glyphosate ligands in the equatorial plane might be responsible for the weakening of the second glyphosate complexation, compared to the ODA/IDA.

This inference, derived from thermodynamic data, is directly corroborated by ³¹P NMR spectroscopy. As shown in Fig. 5, the ³¹P NMR spectrum of the U(VI)–glyphosate system at pH 7 displays three major signals. Based on the stability constants obtained from potentiometric titration, the speciation at this pH, calculated using Hyss2009 software, is as follows: HL, 7%; UO₂L, 10%; UO₂L₂, 60%; UO₂HL, 3%; and UO₂H₂L₂, 20%. The signal at ∼24.5 ppm, assigned to the phosphorus nuclei in the UO₂L and UO₂L₂ complexes, exhibits clear splitting. This observation is consistent with the coexistence of multiple species (UO₂L₂⁴⁻ ∼ 60% and UO₂L⁻ ∼ 10%) predicted using the thermodynamic model. More critically, even within the single UO₂L₂⁴⁻ complex, the proposed nonequivalent coordination modes of the two glyphosate ligands—one tridentate and the other bidentate—would place their respective phosphonate groups in distinct micro-chemical environments. Consequently, the signal splitting in the ∼24.5 ppm region can be reasonably attributed to the superimposed contributions from UO₂L⁻ and the two inequivalent phosphonate moieties within UO₂L₂⁴⁻. This spectroscopic evidence strongly supports the conclusion that the second glyphosate ligand adopts a different (bidentate) coordination mode. The proposed structure of the UO₂L₂ (glyphosate) complex is depicted in Fig. 7(b).

Fig. 5 ³¹P-NMR spectrum of the uranium(VI)–glyphosate system: C_U = 40 mM, C_L = 80 mM, and pH = 7.0.

UO₂HL and UO₂H₂L₂

The protonated complexes UO₂HL and UO₂H₂L₂ predominantly form under weakly acidic to neutral conditions. Given the high proton affinity of the nitrogen atom (pK_a = 9.85) in this pH range, the N atom is typically protonated. We therefore propose that the nitrogen atoms in both UO₂HL and UO₂H₂L₂ are protonated. It is usually more difficult to reveal the coordination mode in the protonated complexes merely by evaluation of the accumulated thermodynamic parameters. However, the analysis of the thermodynamic constants for stepwise complexation reactions may provide insight into the coordination modes in the complexes. For the UO₂HL⁰ complex, the stepwise complexation reaction is UO₂²⁺ + H₂L²⁻ ⇌ UO₂HL⁰, with log β = 19.90 − 9.85 = 10.05 (N is protonated). The stability constant of UO₂L (log β = 13.71) confirms tridentate coordination. This indicates that in UO₂HL⁰, coordination involves one phosphonate oxygen and one carboxylate oxygen bound to U(VI), while the nitrogen remains protonated and non-coordinating. These donor atoms form an eight-membered chelate ring with uranium. The neutral character of UO₂HL⁰ enhances its thermodynamic stability. For UO₂H₂L₂²⁻, the stepwise reaction is UO₂HL⁰ + H₂L²⁻ ⇌ UO₂H₂L₂²⁻, with log β = 32.95 − 19.90 − 9.85 = 3.2 (N is protonated). The second ligand likewise exhibits bidentate coordination through one phosphonate oxygen and one carboxylate oxygen, forming a second eight-membered chelate ring. The significantly weaker binding of the second ligand compared to the first arises from electrostatic repulsion between anionic complexes and the instability of the eight-membered ring structure. The conclusion is further supported by ³¹P NMR data. The ³¹P NMR spectra of the uranium(VI)–glyphosate system at pH = 7 are shown in Fig. 5. Based on the stability constants of the uranium(VI)–glyphosate complexes obtained by potentiometric titration, the composition of the solution under these conditions was calculated using Hyss2009 software as follows: HL: 7%, UO₂L: 10%, UO₂L₂: 60%, UO₂HL: 3%, and UO₂H₂L₂: 20%. Three major peaks around 7.8 ppm, 15.0 ppm, and 24.5 ppm were observed. According to the description of the literature,²⁹ the peak at 7.8 ppm can be assigned to the uncomplexed ligand species HL²⁻. The peak at 24.5 ppm corresponds to the P nucleus in the UO₂L and UO₂L₂ complexes. In these species, phosphonate oxygen coordinates with U, and N is not protonated. This configuration reduces electron density around phosphorus, resulting in deshielding and consequently a larger chemical shift. Conversely, the peak at 15 ppm is assigned to phosphorus in UO₂HL⁰ and UO₂H₂L₂²⁻. Here, phosphonate oxygen coordinates to U(VI), whereas nitrogen is protonated. The positive charge on protonated nitrogen distributes through the N–H bonds, enabling effective hydrogen bonding interactions with phosphoryl (P=O) and carboxylate (C=O) oxygen atoms.³⁰ This causes a reduction in the C–P–O bond angle, increased electron density around phosphorus, and enhanced shielding effects. Consequently, the ³¹P chemical shift moves upfield from 24.5 ppm (UO₂L/UO₂L₂) to 15 ppm (UO₂HL⁰/UO₂H₂L₂²⁻). The proposed structures of the UO₂HL and UO₂H₂L₂ complexes are depicted in Fig. 7(c) and (d).

UO₂H₂L

Under strongly acidic conditions (pH < 1, pH = 0.4 in this work), the uranium(VI)–glyphosate species in solution is completely dominated by the UO₂H₂L⁺ complex (100% abundance). At this pH, the glyphosate ligand exists in its diprotonated form (H₂L⁻), and the amino nitrogen has been confirmed to be protonated. The ³¹P NMR spectrum (Fig. 6) exhibits a broad peak at 12.5 ppm, indicating that the oxygen atoms of the phosphonate group are involved in coordination with the uranyl center (UO₂²⁺). The calculated stability constant (log β = 11.52) for the stepwise complexation reaction UO₂²⁺ + H₂L⁻ = UO₂H₂L⁺ is consistent with this coordination mode. The results of DFT geometry optimizations (Table 4) in this study show that the uranium center forms two nearly equivalent coordination bonds with the two oxygen atoms (O6 and O7) of the phosphonate group, with bond lengths of 2.3149 Å and 2.3145 Å, respectively. Further Mayer bond order analysis reveals that the bond orders of these two U–O bonds are 0.4176 and 0.4038, respectively, quantitatively confirming that both are effective chemical bonds. These structural parameters are highly consistent with the characteristics of a terminal bidentate coordination mode. Therefore, we propose that in the UO₂H₂L⁺ complex, glyphosate coordinates with the uranyl ion through its phosphonate group in a terminal bidentate fashion, while the protonated amino nitrogen and carboxylic oxygen do not participate in coordination. The proposed structure is illustrated in Fig. 7(e).

Fig. 6 ³¹P-NMR spectrum of the uranium(VI)–glyphosate system: C_U = 40 mM, C_L = 80 mM, and pH = 0.4.

Table 4 Bond lengths and Mayer bond order data of key coordination bonds in the five complexes

Uranium(VI)–glyphosate complexes	Bond	Bond length (Å)	Bond order
UO2L	U–O5	2.168	0.625
	U–O6	2.348	0.364
	U–N	2.568	0.336
UO2L2	U–O3	2.296	0.475
	U–O4	2.423	0.331
	U–N1	2.676	0.338
	U–O5	2.226	0.537
	U–O6	2.368	0.325
	U–N2	—	0.053
UO2HL	U–O5	2.28	0.404
	U–O6	2.385	0.277
UO2H2L2	U–O3	2.224	0.501
	U–O4	2.334	0.301
	U–O5	2.224	0.501
	U–O6	2.335	0.301
UO2H2L	U–O6	2.315	0.418
	U–O7	2.342	0.404

---

Fig. 7 Proposed structures of uranium(VI)–glyphosate complexes in solution: (a) UO₂L, (b) UO₂L₂, (c) UO₂HL, (d) UO₂H₂L₂, and (e) UO₂H₂L.

We attempted to compare this coordination mode with those reported in the literature for simple U–phosphonate systems. However, most of the existing studies on simple U–phosphonate systems have focused on solid-state structures and lack thermodynamic data and solution ³¹P NMR data under strictly comparable acidic conditions with high ionic strength. The comprehensive characterization of UO₂H₂L⁺ in this work, including NMR spectroscopy, thermodynamic calculations, and DFT analysis, provides important insights into the specific coordination behavior of the multifunctional ligand glyphosate in extremely acidic environments. Future systematic studies on simple model phosphonic acids (such as methylphosphonic acid) under similar conditions will help establish a more solid experimental benchmark for such comparisons.

DFT calculations

DFT calculations provided strong theoretical support for the coordination modes of the five uranium(VI)–glyphosate complexes proposed in this work. The optimized structures of the complexes are shown in Fig. 8, and the bond lengths and Mayer bond orders for key coordination bonds are summarized in Table 4.

Fig. 8 Optimized complex structure models from DFT calculations: (a) UO₂L, (b) UO₂L₂, (c) UO₂HL, (d) UO₂H₂L₂, and (e) UO₂H₂L.

For the UO₂L⁻ complex, the calculations revealed three distinct coordination bonds between U and the ligand: U–O5 (phosphonate oxygen, bond length 2.168 Å and bond order 0.625), U–O6 (carboxylate oxygen, bond length 2.348 Å and bond order 0.364), and U–N (amino nitrogen, bond length 2.568 Å and bond order 0.336). All three bond orders are significantly greater than zero, confirming the tridentate coordination mode of glyphosate via carboxylate oxygen, phosphonate oxygen, and amino nitrogen. This is consistent with the conclusion inferred from the correlation between thermodynamic data and ΣpK_a.

In the UO₂L₂⁴⁻ complex, the calculated structure shows different coordination modes for the two glyphosate ligands. The first ligand (L¹) coordinates in a tridentate mode (U–O3, U–O4, U–N1, with bond orders of 0.475, 0.331, and 0.338, respectively), while the second ligand (L²) binds to the uranium center only through a phosphonate oxygen (U–O5, bond order 0.537) and a carboxylate oxygen (U–O6, bond order 0.325). The U–N2 bond order is only 0.053, indicating that the amino nitrogen is essentially non-coordinating. This directly confirms our earlier inference of “the first ligand being tridentate and the second bidentate” and provides a reasonable explanation for the significant decrease in the stepwise stability constant (log K₂).

For the protonated complexes UO₂HL⁰ and UO₂H₂L₂²⁻, the calculations show that the amino nitrogen in the ligands is protonated and does not participate in coordination (U–N bond orders are very low or absent). In UO₂HL⁰, coordination is achieved through a phosphonate oxygen (U–O5, bond length 2.280 Å and bond order 0.404) and a carboxylate oxygen (U–O6, bond length 2.385 Å and bond order 0.277), forming an eight-membered chelate ring. In UO₂H₂L₂²⁻, both ligands coordinate in a similar bidentate mode symmetrically (U–O3/O5 and U–O4/O6, with comparable bond lengths and bond orders), supporting the structural model involving two eight-membered rings bound to uranium.

For the UO₂H₂L⁺ complex, which exists under strongly acidic conditions, the DFT-optimized structure shows that the two oxygen atoms of the phosphonate group (O6 and O7) coordinate with the uranium center in a nearly equivalent manner (bond lengths of 2.315 Å and 2.342 Å and bond orders of 0.418 and 0.404, respectively), while the protonated amino and carboxyl groups do not participate in coordination. This confirms that under these conditions, glyphosate coordinates solely via the phosphonate group in a terminal bidentate mode.

In summary, the bond lengths and bond orders obtained from DFT calculations are in excellent agreement with the coordination modes inferred for each complex from potentiometric titrations, ³¹P NMR spectroscopy, and thermodynamic trend analysis. This theoretically strengthens the solution coordination chemistry model for uranium(VI)–glyphosate proposed in this study.

The species of glyphosate and its uranium(VI) complexes in the environment

Fig. 9 illustrates the distribution of glyphosate species as a function of pH. As shown, the HL²⁻ species predominates within the pH range of 5.31 to 9.85. Consequently, glyphosate primarily exists as the HL²⁻ anion in environmental solutions ranging from weakly acidic to weakly alkaline conditions. In this HL²⁻ form, the amino nitrogen atom is protonated. The two hydrogen atoms attached to this nitrogen form intramolecular hydrogen bonds with the phosphonate oxygen and carboxylate oxygen atoms, respectively, stabilizing the molecule in a spirocyclic configuration. Carrying two negative charges, the HL²⁻ species readily undergoes complexation reactions with metal cations.

Fig. 9 The distribution of glyphosate species as a function of pH.

Fig. 10 shows the distribution of glyphosate complex species as a function of pH in the presence of uranium. As indicated, the species UO₂HL and UO₂L predominate within the pH range of 3.5 to 8.3. Therefore, in environmental waters—from weakly acidic to weakly alkaline conditions and in the presence of uranium—glyphosate complexes primarily occur as UO₂HL and UO₂L. Experimental results demonstrate that these U(VI)–glyphosate complexes exhibit high aqueous solubility under neutral pH conditions. This high solubility enhances the mobility of uranium in environmental systems, thereby complicating pollution control and remediation efforts.

Fig. 10 The distribution of glyphosate complex species as a function of pH in the presence of uranium.

Conclusions

Across acidic to basic aqueous solutions, a total of eleven uranium(VI)–glyphosate complexes were identified through potentiometric titration. These include highly protonated species (UO₂HL, UO₂H₂L, and UO₂H₂L₂) in strongly acidic media, unprotonated complexes (UO₂L and UO₂L₂) under neutral conditions, mixed hydroxyl–glyphosate species (UO₂(OH)₂L and UO₂(OH)₂L₂) in basic environments, and polynuclear complexes such as (UO₂)₂HL₂, (UO₂)₂H₋₁L₂, (UO₂)₂H₋₂L₂, and (UO₂)₂L₂. The complexation reactions of UO₂L, UO₂L₂, UO₂HL, and UO₂H₂L₂ are all endothermic, with both positive enthalpy and entropy changes, and are classified as inner-sphere complexes. Thermodynamic data, supported by structural evidence from ³¹P NMR spectroscopy and further validated by DFT calculations, indicate that glyphosate adopts a tridentate coordination mode in the 1 : 1 UO₂L complex. In contrast, in the 1 : 2 UO₂L₂ complex, the second glyphosate ligand coordinates in a bidentate manner, without involvement of the nitrogen atom, as directly confirmed from the calculated bond orders. In the protonated complexes UO₂HL and UO₂H₂L₂, each ligand binds to U(VI) through one oxygen atom from the phosphonate group and one from the carboxylate group, while the nitrogen remains protonated and does not participate in coordination—a structural assignment corroborated by DFT-optimized geometries. Furthermore, in the strongly acidic UO₂H₂L complex, DFT calculations confirm a terminal bidentate coordination solely through the phosphonate group, with both the amino and carboxylate groups remaining protonated and non-coordinating. Finally, the speciation of glyphosate in environmental contexts, both in the absence and presence of uranium, is also discussed.

Author contributions

Qiang Li: conceptualization, methodology, investigation, writing – original draft, supervision, and project administration. Xingliang Li: conceptualization, funding acquisition, formal analysis, and validation. Yanqiu Yang: investigation, data curation, and visualization. Huining Sun: resources and validation. Lujun Jia: resources and writing – review and editing. Haijiao Xie: calculation and formal analysis.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article are available within the manuscript. Additional raw data (potentiometric and calorimetric titrations) are available from the corresponding author upon reasonable request.

Acknowledgements

The authors gratefully acknowledge Mianyang Polytechnic for providing financial support and resources, and the Institute of Nuclear Physics and Chemistry, China Academy of Engineering Physics for providing experimental facilities. Special thanks are extended to Professor Xingliang Li and Professor Yanqiu Yang for their valuable discussions and suggestions, and to Haijiao Xie of Hangzhou Yanqu Information Technology for performing the DFT calculations.

References

1. S. V. Kononova and M. A. Nesmeyanova, Biochemistry, 2002, 67, 184–195.
2. Z. Szabó, J. Chem. Soc., Dalton Trans., 2002, 21, 4242–4247.
3. P. Eberbach and L. Douglas, Soil Biol. Biochem., 1983, 15, 485–487.
4. R. J. Motekaitis and A. E. Martell, J. Coord. Chem., 1985, 14, 139–149.
5. L. H. E. Madsen, H. H. Christensen and C. Gottlieb-Petersen, Acta. Chir. Scand., 1978, A32, 79.
6. P. H. Smith and K. N. Raymond, Inorg. Chem., 1988, 27, 1056–1061.
7. M. Dhansay and P. Linder, J. Coord. Chem., 1993, 28, 133–145.
8. P. Buglyó, T. Kiss, M. Dyba, M. Jezowska-Bojczuk, H. Kozlowski and S. Bouhsina, Polyhedron, 1997, 16, 3447–3451.
9. P. G. Daniele, C. De Stefano, E. Prenesti and S. Sammartano, Talanta, 1997, 45, 425–431.
10. R. J. Motekaitis and A. E. Martell, J. Coord. Chem., 1985, 14, 139–149.
11. M. B. McBride, Soil Sci. Soc. Am. J., 1991, 55, 979–985.
12. J. Sheals, P. Persson and B. Hedman, Inorg. Chem., 2001, 40, 4302–4309.
13. C. W. Sill and H. E. Peterson, Anal. Chem., 1947, 19, 646–651.
14. G. Gran, Analyst, 1952, 77, 661–670.
15. J. Jiang, L. Rao, P. Di Bernardo, P. Zanonato and A. Bismondo, J. Chem. Soc., Dalton Trans., 2002, 1832–1838.
16. P. Gans, A. Sabatini and A. Vacca, Talanta, 1996, 43, 1739–1753.
17. R. Arnek, Ark. Kemi, 1970, 32, 81.
18. B. C. Barja and M. Dos Santos Afonso, Environ. Sci. Technol., 1998, 32, 3331–3335.
19. Y. Yang, L. Lv, Y. Liu, B. Chen, J. Liu, X. Li and S. Luo, Dalton Trans., 2021, 50, 705–714.
20. Y. Yang, Z. Zhang, L. Yang, J. Liu, C. Xu, Sh Luo and L. Rao, Inorg. Chem., 2019, 58, 6064–6074.
21. S. A. Ansari, Y. Yang, Z. Zhang, K. J. Gagnon, S. J. Teat, S. Luo and L. Rao, Inorg. Chem., 2016, 55, 1315–1323.
22. E. N. Rizkalla and G. R. Choppin, Handbook on the Physics and Chemistry of Rare Earths, in Lanthanides/Actinides: Chemistry, ed. K. A. Gschneider, L. Eyring Jr., G. R. Choppin and G. H. Lander, Elsevier Science BV, New York, 1994, vol. 18, p. 572.
23. G. H. Nancollas, Interactions in Electrolyte Solutions, Elsevier, Amsterdam, The Netherlands, 1966, pp. 133–140.
24. L. Lv, B. Chen, J. Liu, J. Chen, C. Xu and Y. Yang, Dalton Trans., 2019, 48, 566–577.
25. Ed. A. E. Martell, R. M. Smith and R. J. Motekaitis, Critically Selected Stability Constants of Metal Complexes, Database Version 8, NIST, Gaithersburg, MD, 2004.
26. L. Rao, A. Y. Garnov, J. Jiang, P. D. Bernardo, P. Zanonato and A. Bismondo, Inorg. Chem., 2003, 42, 3685–3692.
27. P. Di Bernardo, G. Tomat, A. Bismondo, O. Traverso and L. Magon, J. Chem. Res., Miniprint, 1980, 3144–3171.
28. J. Jiang, J. C. Renshaw, M. J. Sarsfield and F. R. Livens, Inorg. Chem., 2003, 42, 1233–1240.
29. B. Liu, L. Dong, Q. Yu, X. Li, F. Wu, Z. Tan and S. Luo, J. Phys. Chem. B, 2016, 120, 2132–2137.
30. S. Castellino, G. C. Leo and R. D. Sammons, Biochemistry, 1989, 28, 3856–3868.

---