Molecular properties of artificial sweeteners in water

Abstract

Currently, artificial sweeteners – saccharin, acesulfame, cyclamic acid, and aspartame – are used as food additives. These compounds in their molecular and ionised forms have been studied by quantum chemical methods: the B3LYP hybrid variant of density functional theory and the post-Hartree–Fock DLPNO-CCSD(T) method, which includes a major part of the electron correlation energy. Full geometry optimisation and vibrational analysis were performed using the B3LYP method for neutral molecules, their cations and their anions. For neutral molecules, calculated molecular properties include electronic, electrical, and thermodynamic properties. Calculated ionisation energies and electron affinities were used to predict redox properties – the absolute oxidation and reduction potentials. All data refer to water as a solvent. Three electron–proton transfer mechanisms were investigated for studied systems: (i) the SPLET (Sequential Proton Loss Electron Transfer) mechanism starts with proton transfer (deprotonation) and then continues with electron transfer (oxidation); (ii) the SET-PT (Single Electron Transfer followed by Proton Transfer) mechanism has interchanged steps, so that electron transfer is followed by proton transfer; and (iii) the HAT (Hydrogen Atom Transfer) mechanism assumes simultaneous oxidation and deprotonation. According to the reaction Gibbs energy (B3LYP) and/or total electronic energy (DLPNO-CCSD(T)), the SPLET mechanism is favoured. High ionisation energies prevent the functionality of studied artificial sweeteners as antioxidants.

---

Introduction

The consumption of sucrose (natural sugar) is still growing. This substance has both positive and negative effects on the human body. The positives are the high energy content, fast energy source and fast absorbability.^1–10 The negative effects are, for example, diabetes mellitus, irreversible damage to teeth and obesity. Artificial sweeteners reduce the negative effects, and their sweetness significantly exceeds the sweetness of sucrose. However, their daily use can also cause prediabetes and can trigger other diseases.^11–21 In one sentence, sugar substitutes belong to food additives that provide a sweetness effect; unlike sugar, they contain much less food energy (low-calorie or zero-calorie sweeteners). Depending on the source, they can be classified as plant extracts or products of chemical synthesis. A specific class consists of sugar alcohols, some with simple aliphatic structures, such as arabitol, erythritol, glycerol, mannitol, sorbitol and xylitol. The most used artificial sweeteners are (Table 1) 1H-1λ6,2-benzothiazole-1,1,3(2H)-trione – saccharin (E954),^22,27 potassium 6-methyl-2,2-dioxo-2H-1,2λ6,3-oxathiazin-4-olate – acesulfame potassium (E950),^23,28 sodium cyclohexylsulfamate – cyclamate (E952)^24,29 and methyl L-α-aspartyl-L-phenylalaninate – aspartame (E951).^25,26,30

Table 1 Structures of artificial sweeteners^a

a For selected physicochemical properties, see Table S1.

---

Saccharin is a cyclisation product that can be prepared by the following route: toluene → o-chlorosulfonic acid → o-sulfonamide → [oxidation] carboxylic acid → [cyclisation] saccharin. It is a white, thermally stable solid that is not very soluble in water, has an acidic N–H bond (K_a = 1.3), and is hydrophilic (log P_ow = 0.9). It is odourless with negligible nutritional value. It is about 500 times sweeter than sucrose and is therefore used as a safe artificial sweetener for patients suffering from diabetes or prediabetes. It is applied in the form of sodium or calcium salts in drinks, pastries, and candies and as an additive to some medicines to cover their bitter taste. These salts are highly soluble in water: 0.67 g cm⁻³. It is often used in combination with other sweeteners such as cyclamate and aspartame.

Acesulfame is structurally closely related to saccharin but has a six-membered ring instead of a five-membered one. It has similar properties to saccharin: a white solid, thermally stable, soluble in water, with an acidic N–H bond (K_a = 2.0); it is hydrophobic (log P_ow = −1.3). It is about 200 times sweeter than sucrose, which implies its usage as an artificial sweetener. At higher concentrations, it has a slightly bitter aftertaste. It is sold in the form of potassium salts and is often mixed with other artificial sweeteners, e.g., aspartame.

Cyclamic acid (cyclamate, cyclohexanesulfamic acid) also belongs to the family of compounds with the [double bond splayed left]SO₂ group. This white, odourless solid is very water-soluble (1000 mg cm⁻³), stable when heated, and 30–50 times sweeter than sucrose. It is cheaper than most artificial sweeteners and is used in the form of the sodium or potassium salts.

Aspartame is a methyl ester of a dipeptide formed from L-asparatic acid and L-phenylalanine. Aspartame can undergo hydrolysis that is strongly dependent on pH: it is most stable at pH = 4.3 with a half-life of about 300 days. At pH = 7 its half-life is only a few days. Sweetened drinks have pH = 3–5. Under strongly acidic or basic conditions and upon digestion, aspartame is hydrolysed to aspartic acid, phenylalanine and methanol. High levels of phenylalanine can be hazardous for individuals suffering from phenylketonuria. Aspartame is about 180–200 times sweeter than sucrose. As a low-calorie sweetener, it is used to substitute sugar in beverages and foods, especially for patients suffering from diabetes. Aspartame is metabolised and absorbed very quickly and is not accumulated in the body. Aspartame crystallises in the zwitterion form; it forms aspartame hydrochloride, having an acidic carboxyl group (aspartamium).

Recent research on aspartame has focused on investigating the effects of its metabolites on the human brain.³¹ Also, long-term studies of people with diabetes who consumed artificial sweeteners have demonstrated measurable decline in cognitive functions.^32,33 Moreover, animal studies administering aspartame have shown increased insulin release, contributing to atherosclerosis, memory deficits, and other complications.^34,35

The aim of the quantum-chemical study is to characterise the redox properties of artificial sweeteners in water, such as absolute oxidation and reduction potentials, molecular electronegativity, chemical hardness, electrophilicity index, and mechanisms of electron–proton transfer –problems not studied so far.

Methods

The ORCA package was used for quantum-chemical calculations of the electronic structure and molecular properties of the studied molecules.^36–38 Density functional theory with a hybrid variant B3LYP/UKS was used for full geometry optimisation followed by a complete vibrational analysis. The basis set of valence-triple-zeta-polarisation and diffuse functions quality (abbr. def2-TZVPD) was applied.³⁹

The post-Hartree–Fock method, namely Domain Localised Pair Natural Orbitals – Coupled Cluster Singles + Doubles + perturbative Triples method, abbr. DLPNO-CCSD(T), has also been applied with the aug-cc-pVTZ (Dunning correlation consistent with Polarisation Valence Triple-Zeta, extended by diffusion functions) basis set;^40,41 the UHF variant was applied for open shell systems. This method involves the main part of the electron correlation energy. A conductor-like polarisable continuum model (CPCM) was used to account for the solvent effect in water with relative permittivity ε_r = 80.1.⁴² Details of the method of calculations are presented in the SI.

Results and discussion

Molecular geometry

Geometry optimisation was performed by the B3LYP method for three classes of species: (i) canonical forms (C) of saccharin, acesulfame, cyclamic acid and aspartame; (ii) ionic forms of saccharinate, acesulfamate, cyclamate and aspartamium; and (iii) zwitterionic forms of cyclamic acid and aspartame. Each species was considered in three oxidation states: as a closed shell system L⁰, an oxidized form L⁺, and a reduced form L⁻. A total of 30 molecules per molecular ions were processed. The final geometries are shown in Fig. S1–S3.

Saccharin and acesulfame do not have rotatable C–C bonds, so they do not have rotamers. Aspartame has 3⁸ = 6561 rotamers. The calculated geometries for neutral molecules in water closely copy the solid-state structures determined by X-ray diffraction (Fig. 1 and Fig. S4).^22–30

Fig. 1 Overlaid structures of saccharine, acesulfame, cyclamic acid, and aspartame taken from solid-state X-ray diffraction (green) and B3LYP calculations in water (purple).

For neutral amino acids, the zwitterionic form is more stable in water compared to the canonical form by ΔE⁰ = 12.2 and ΔG^ø = 20.5 kcal mol⁻¹ for cyclamic acid and ΔE⁰ = 4.2 and ΔG^ø = 3.2 kcal mol⁻¹ for aspartame (B3LYP data). This is caused by the charge polarisation in the zwitterionic forms, which results in higher hydration energy: ΔE_CPCM = −48.1 (aspartame Z) vs. −23.4 kcal mol⁻¹ (aspartame C) using B3LYP. The DLPNO-CCSD(T) method gave electron-correlated values of ΔE⁰ = 10.6 and 1.8 kcal mol⁻¹ for cyclamic acid and aspartame, respectively. The stabilisation of the zwitterionic forms in the aqueous solution is a typical feature of amino acids.^43,44

Vibrational spectra

Calculated unscaled vibrational transitions (harmonic approximation, B3LYP method) are presented in Fig. S5 for saccharine, sacchatinate (1−), acesulfame, acesulfamate (1−), cyclamic acid in canonical and zwitterionic forms, cyclamate (1−), aspartame in canonical and zwitterionic forms, and aspartamium (1+). All spectra show two domains: high-frequency referring to the valence vibrations (C–H, N–H, and O–H) in the region 2800–3700 cm⁻¹ and low-frequency referring to the deformation vibrations below 1800 cm⁻¹. The experimental data are also included for comparison.^45–48 The calculated unscaled spectra are broader when compared to the highest stretching frequencies (O–H or N–H). The absence of the valence N–H vibrations in saccharinate (1−) and acefulfamate (1−) species is evident in the high-frequency domain. The effect of the scaling of the B3LYP calculated frequencies by a factor of 0.962 is demonstrated for aspartame.

Electron–proton transfer

The antioxidant activity of the species is usually accompanied by electron transfer coupled with proton transfer (removal). Three mechanisms have been proposed depending on the reaction steps: (i) The SPLET (sequential proton loss electron transfer) mechanism starts with proton transfer (deprotonation) and then continues with electron transfer (oxidation); (ii) the SET-PT (single electron transfer followed by proton transfer) mechanism has interchanged steps, so that electron transfer is followed by proton transfer; (iii) the HAT (hydrogen atom transfer) mechanism assumes the removal of a hydrogen atom, so the transfer of electrons and protons takes place in one step.⁴⁹

Energy profiles for electron–proton coupled transfer are shown in Table 2 for saccharin, acesulfame, cyclamic acid and aspartamium. The reaction energy for proton transfer includes the energy of the hydrated hydrogen radical (H⁺)_aq released; several values have been proposed, and “the best estimate” is ΔG^ø(H⁺)_aq = −262 kcal mol⁻¹.⁵⁰ However, the calculated values are sensitive to the method used, and the value suitable for the B3LYP method is ΔG^ø(H⁺)_aq = −274 kcal mol⁻¹.⁵¹

Table 2 Energy profile for electron–proton coupled transfer^a

a Gibbs energies G^ø refer to B3LYP calculations, and total electronic energies E⁰ to DLPNO-CCSD(T). For amino acids the canonical form is selected as a reference.

---

Taking into account the release of (H⁺)_aq, deprotonation of saccharin within the SPLET mechanism is associated with the Gibbs energy change Δ_d1G(corr) = 277 − 274 = 3 kcal mol⁻¹. Compared to the second step – oxidation, for which Δ_o1G = 142 kcal mol⁻¹, deprotonation is a much less energetically demanding process. The concurrent reaction through the SET-PT mechanism starts with an energy demand Δ_o2G = 170 kcal mol⁻¹ and then with a release Δ_d2G(corr) = 248 − 274 = −26 kcal mol⁻¹. Finally, the HAT mechanism requires the addition Δ_HATG(corr) = 3 + 142 = 170 − 26 = 145 kcal mol⁻¹. All these values refer to B3LYP calculations.

According to DLPNO-CCSD(T) calculations, the first deprotonation step within the SPLET mechanism of saccharin has Δ_d1E(corr) = 285 − 274 = 11 kcal mol⁻¹ while the second step has Δ_d2E = 174 kcal mol⁻¹. This is still more favourable compared to the SET-PT mechanism with steps Δ_o2E = 182 and Δ_d2E(corr) = 276 − 274 = 2 kcal mol⁻¹ or Δ_HATE(corr) = 184 kcal mol⁻¹.

For acesulfame, Δ_d1G(corr) = 273 − 274 = −1, Δ_d2G = 146, Δ_HATG(corr) = 145 kcal mol⁻¹ by B3LYP, and Δ_d1E(corr) = 282 − 274 = 8, Δ_d2E = 156 kcal mol⁻¹, Δ_HATE(corr) = 164 kcal mol⁻¹ by DLPNO-CCSD(T).

For cyclamic acid, Δ_d1G(corr) = 265 − 274 = −9 kcal mol⁻¹, Δ_d2G = 131, Δ_HATG(corr) = 122 kcal mol⁻¹ by B3LYP, and Δ_d1E(corr) = 274 − 274 = 0, Δ_d2E = 143, Δ_HATE(corr) = 143 kcal mol⁻¹ by DLPNO-CCSD(T).

For the aspartame pathway, the starting point is aspartamium (1+) present in the hydrochloride salt. Then Δ_d1G(corr) = 274 − 274 = 0, Δ_d2G = 138, Δ_HATG(corr) = 138 kcal mol⁻¹ by B3LYP, and Δ_d1E(corr) = 285 − 274 = 11, Δ_d2E = 151, Δ_HATE(corr) = 162 kcal mol⁻¹ by DLPNO-CCSD(T).

As the steps in the energy profile of electron–proton transfer decrease, the hypothetical antioxidant capacity increases in the series of saccharin, acesulfame, aspartamium, and cyclamic acid. The above data refer to the thermodynamic predispositions for the oxidation of sweeteners (Fig. 2). (Kinetic studies will require identification of the transition state and evaluation of the activation Gibbs energy.)

Fig. 2 Steps for oxidation (hydrogen atom removal) of studied sweeteners in water.

Molecular descriptors

The calculated molecular descriptors for electroneutral (reference) molecules are shown in Tables 3 and 4 for B3LYP calculations and in Tables 5 and 6 for DLPNO-CCSD(T) calculations.

Table 3 Molecular properties calculated by the B3LYP/def2-TZVPD method in water^a

	Molecule	p	−Q	α	S	V	E zpe	ST^ø
a All energy quantities in units of kcal mol^−1; conversion: 1 hartree = 627.5095 kcal mol^−1; 1 eV = 23.06054; 1 kcal mol^−1 = 4.184 kJ mol^−1. Standard temperature T^ø = 298.15 K. Dipole moment p/debye; isotropic quadrupole moment Q/ea0^2, isotropic dipole polarisability α/a0^3, solvated surface area S/a0^2, solvated volume V/a0^3, debye, D = 3.336 × 10^−30 A m s; angstrom, Å = 10^−10 m; bohr, a0 = 5.292 × 10^−11 m; zero-point energy Ezpe, total entropic term ST^ø.
Canonical neutral forms
1C	Saccharin	5.38	56.4	164.9	682	1248	72.5	28.2
2C	Acesulfame	4.70	49.4	120.6	619	1085	63.1	27.6
3C	Cyclamic acid	7.71	58.3	148.3	739	1336	126.3	30.2
4C	Aspartame	3.02	87.2	277.8	1252	2347	197.4	44.4
Ionic forms
1I	Saccharin-ate(−) (Na)	12.7	67.4	177.6	674	1247	64.8	27.7
2I	Acesulfam-ate(−) (K)	9.81	58.2	132.0	613	1071	55.4	27.3
3I	Cyclam-ate(−) (Na)	14.6	69.1	158.0	749	1331	118.9	29.8
4I	Aspartamium(+) (Cl)	15.7	66.4	268.5	1246	2309	206.8	44.9
Zwitterionic forms
3Z	Cyclamic acid	10.9	59.8	150.0	746	1354	127.9	29.9
4Z	Aspartame	15.4	108.2	279.5	1234	2334	198.2	43.7

---

Table 4 Adiabatic redox properties calculated by the B3LYP/def2-TZVPD method in water^a

	Molecule	E i	E eg	χ	η	ω	E ^øox	E ^øred
a Redox properties in kcal mol^−1, ionisation energy Ei = E^+ − E^0, electron affinity Eeg = E^− − E^0, molecular electronegativity, chemical hardness, electrophilicity index, absolute oxidation potential E^øox = −ΔoxG^ø/F in V, absolute reduction potential E^øred = −ΔredG^ø/F in V, Faraday constant F = 96 485 A s mol^−1. n.a. – not available because of dissociation of the H atom for negatively charged cyclamic acid.
Canonical neutral forms
1C	Saccharin	173.0	−62.3	117.7	55.4	125.0	−7.39	2.81
2C	Acesulfame	166.5	−56.7	111.6	54.9	113.4	−7.14	2.57
3C	Cyclamic acid	156.3	n.a.				−6.68
4C	Aspartame	139.2	−31.0	85.0	54.1	66.8	−5.99	1.44
Ionic forms
1I	Saccharin-ate(−) (Na)	143.3	−44.7	94.0	49.3	89.6	−6.16	2.07
2I	Acesulfam-ate(−) (K)	148.1	−36.6	92.4	55.8	76.5	−6.34	1.70
3I	Cyclamate(−) (Na)	132.4	−10.8	71.6	60.8	42.2	−5.70	0.50
4I	Aspartam-ium(+) (Cl)	152.1	−37.6	94.9	57.3	78.6	−6.54	1.73
Zwitterionic forms
3Z	Cyclamic acid	175.2	−17.9	96.6	78.6	59.3	−7.40	0.94
4Z	Aspartame	150.8	−31.7	91.3	59.6	69.9	−6.47	1.50

---

Table 5 Molecular properties calculated by the DLPNO-CCSD(T)/aug-cc-pVTZ method in water^a

	Molecule	p	−Q	S	V	HOMO	LUMO
a Units as in Table 3. HOMO – highest occupied molecular orbital, LUMO – lowest unoccupied molecular orbital at the Hartree–Fock level.
Canonical neutral forms
1C	Saccharin	6.163	57.0	682	1248	−233.6	19.1
2C	Acesulfame	5.103	49.9	619	1085	−248.4	19.6
3C	Cyclamic acid	8.108	58.5	739	1336	−258.6	20.4
4C	Aspartame	2.728	86.9	1252	2347	−208.8	18.1
Ionic forms
1I	Saccharin-ate(−) (Na)	13.76	67.8	674	1247	−221.1	20.4
2I	Acesulfam-ate(−) (K)	10.52	58.4	613	1071	−229.2	21.1
3I	Cyclamate(−) (Na)	15.43	68.1	710	1290	−239.8	21.6
4I	Aspartam-ium(+) (Cl)	16.38	66.3	1246	2309	−216.2	16.1
Zwitterionic forms
3Z	Cyclamic acid	11.23	60.1	746	1354	−273.8	20.7
4Z	Aspartame	15.66	109.1	1234	2334	−209.4	17.9

---

Table 6 Adiabatic redox properties calculated by the DLPNO-CCSD(T)/aug-cc-pVTZ method in water^a

	Molecule	E i	E eg	χ	η	ω
a Units as in Table 4.
Canonical neutral forms
1C	Saccharin	182.4	−59.8	121.1	61.3	119.6	−7.91	2.59
2C	Acesulfame	174.9	−54.7	114.8	60.1	109.6	−7.58	2.37
3C	Cyclamic acid	168.6	n.a.				−7.31
4C	Aspartame	151.0	−28.5	89.7	61.2	65.8	−6.55	1.24
Ionic forms
1I	Saccharin-ate(−) (Na)	173.6	−43.0	108.3	65.3	89.8	−7.53	1.86
2I	Acesulfam-ate(−) (K)	156.4	−35.3	95.9	60.6	75.9	−6.78	1.53
3I	Cyclamate(−) (Na)	143.4	−8.7	76.1	67.4	42.9	−6.22	0.38
4I	Aspartam-ium(+) (Cl)	160.8	−35.6	98.2	62.6	77.0	−6.97	1.54
Zwitterionic forms
3Z	Cyclamic acid	188.1	−17.4	102.7	85.3	61.9	−8.16	0.75
4Z	Aspartame	160.4	−29.2	94.8	65.6	68.5	−6.96	1.27

---

These ground-state properties can be classified into two groups. Bulk properties increase with increasing molar mass of the molecule; in the series of acesulfame (2C), saccharin (1C), cyclamic acid (3C) and aspartame (4C), the solvated surface (and volume) increases as S = 619, 682, 739, and 1252a₀² (and V = 1085, 1248, 1336, and 2347a₀³); the absolute value of the quadrupole moment (and dipole polarisability) varies as |Q| = 49, 56, 58, and 87ea₀² (and α = 121, 165, 148, and 278a₀³); the zero-point vibration energy (and the total entropic term) increases as E_zpe = 63, 72, 126, and 197 (and ST^ø = 27.6, 28.2, 30.1, and 44.4) kcal mol⁻¹. The dipole moment shows no systematic trend (p = 4.70, 5.38, 7.71, and 3.02 debye) because it depends on the spatial separation of barycentres of negative and positive charges.

The second class of molecular descriptors refer to global properties such as energies of frontier orbitals (HOMO and LUMO) and the adiabatic redox properties (ionisation energy E_i, electron affinity E_eg, chemical hardness η, molecular electronegativity χ, electrophilicity index ω, absolute oxidation E^ø_ox and reduction E^ø_red potentials).

While the chemical hardness of the studied molecules is approximately the same (η ∼ 55 kcal mol⁻¹), aspartame shows the lowest electronegativity (χ = 85 kcal mol⁻¹) and the lowest electrophilicity (ω = 67 kcal mol⁻¹).

The zwitterionic forms of cyclamic acid (3Z) and aspartame (4Z) exhibit an increased dipole moment (p = 10.89 and 15.37 debye); bulk molecular properties remain close to their canonical forms. The ionisation energy is slightly higher for 4Z compared to 4C: 151 vs. 139 kcal mol⁻¹ by B3LYP and 160 vs. 151 kcal mol⁻¹ by DLPNO-CCSD(T).

Ionic forms carrying negative charge (1I, 2I, 3I) have lower ionisation energies compared to neutral prototypes; in contrast, aspartamium has the lowest ionisation energy.

Redox properties

The properties associated with redox processes were evaluated in the adiabatic way – after the complete geometry optimisation of the neutral and ionised species. These are also collected in Tables 3 and 4. The adiabatic ionisation energy based on the difference in electronic energies decreases in the order of the canonical forms saccharin (C1), acesulfame (C2), cyclamic acid (C3) and aspartame (C4) between E_i = 173–139 kcal mol⁻¹ (B3LYP values). The ionization energy predetermines the reaction Gibbs energy for oxidation, Δ_oxG^ø ∼ E_i, which in turn relates to the absolute oxidation potential, E^ø_ox = –Δ_oxG^ø/F (F – Faraday constant). Therefore, the absolute oxidation potential correlates with the adiabatic ionisation energy along a straight line – see Fig. 3.

Fig. 3 Correlation of the absolute oxidation potential with the adiabatic ionisation energy (left) and the absolute reduction potential with the electrophilicity index; linear equation y = b₀ + b₁x.

Data on ionisation energies can be classified into three groups. (i) 17 amino acids (glycine, α-alanine, GABA, DAVA, β-alanine, valine, leucine, isoleucine, AABA, BABA, AAIBA, BAIBA, hydroxyglycine, hydroxyalanine, serine, threonine, and homoserine) in their 34 forms (canonical C and zwitterionic Z) cover the interval of ionisation energies between 133 (DAVA, Z) and 155 (hydroxyglycine, Z) kcal mol⁻¹. (ii) Five dopaminergic molecules (dopamine, noradrenaline, adrenaline, DOPA, and tyrosine) in their 10 forms (canonical C and zwitterionic Z) have ionisation energies between 98 (dopamine Z) and 137 (tyrosine Z) kcal mol⁻¹. Low E_i values predispose the functionality of dopaminergic molecules as effective antioxidant agents.

(iii) Four artificial sweeteners (saccharin, acesulfame, cyclamic acid, and aspartame) in their 10 forms (canonical C, 2 zwitterionic Z, and 4 ionic I) have ionisation energies between 139 (aspartame, C) and 175 (cyclamic acid, Z) kcal mol⁻¹. High E_i values prevent the functionality of artificial sweeteners as antioxidants. On the other hand, they can be considered as substances stable against oxidation.

Molecular electrostatic potential for the neutral molecules is plotted in Fig. 4. Molecules 1C, 2C and 3C show two charged regions: a negative one, which is spread over the O₂S–NH–C(O) backbone, and a positive one involving the carbon and hydrogen atoms of a ring. In contrast, aspartame has three domains of negative charge.

Fig. 4 Molecular electrostatic potential drawn at the isosurface electron density for 1C, 2C, 3C, 4C, 3Z and 4Z (in the reading direction). Red – negative, blue – positive.

Conclusions

Four commonly used artificial sweeteners – saccharin, acesulfame, cyclamic acid and aspartame – were investigated by quantum-chemical computational methods. The B3LYP method was used to optimise the geometry of the neutral and ionised species (cations and anions). The optimised geometries in water copy the solid-state structure obtained by X-ray diffraction.

Some molecular properties of neutral molecules increase with increasing molar mass: solvated surface and volume, electric quadrupole moment, dipole polarisability, zero-point vibration energy and total entropic term.

The absolute oxidation potentials of the investigated species are too negative; therefore, they do not act as antioxidants and are redox stable.

According to the reaction Gibbs energy, hydrogen transfer occurs preferentially in the two-step SPLET mechanism (deprotonation, followed by oxidation), compared to the alternative SET-PT mechanism (oxidation and then deprotonation).

Within the series of electroneutral sweeteners, aspartame has the lowest ionisation energy compared to three sulphur-containing molecules.

The zwitterionic forms of cyclamic acid and aspartame are more stable in water compared to the canonical forms due to the increased hydration energy.

The ionised (hydrogen-deficient) forms of saccharinate, acesulfamate and cyclamate exhibit lower ionisation energies, while the aspartamium possesses increased ionisation energy compared to electroneutral molecules.

The calculated absolute oxidation potentials correlate with the ionisation energies along a straight line (r² = 0.997). Analogously, the correlation of absolute reduction potentials with the electrophilicity indices is proven.

The results obtained by the highly correlated DLPNO-CCSD(T) method are qualitatively the same as those obtained by B3LYP; however, they are quantitatively much more reliable.

Author contributions

C. R. wrote the core of the manuscript and discussion. J. S. wrote a part of the manuscript. including the literature search. R. B. did quantum chemical calculations by DLPNO-CCSD(T) methods and made figures. All authors reviewed the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

Computer protocols are available from the corresponding author upon request.

Supplementary information (SI): details about the methods, calculated total energies, figures of optimised structures, and calculated vibrational transitions. See DOI: https://doi.org/10.1039/d5nj03504j.

Acknowledgements

The Slovak grant agency VEGA (project 1/0191/22) is acknowledged for the financial support.

Notes and references

1. T. A. Khan, J. J. Lee, S. Ayoub-Charette, J. C. Noronha, N. McGlynn, L. Chiavaroli and J. L. Sievenpiper, Eur. J. Clin. Nutr., 2023, 77, 1009.
2. Artificial sweeteners. Meyler's Side Effects of Drugs, in The International Encyclopedia of Adverse Drug Reactions and Interactions, ed. J. K. Aronson, Elsevier, 16th edn, 2016, p. 713.
3. M. C. Yebra-Biurrun, Sweeteners, in Encyclopedia of Analytical Science, ed. P. Worsfold, A. Townshend and C. Poole, Elsevier, 2nd edn, 2005, p. 562.
4. J. F. Lawrence, Cyclamates, in Encyclopedia of Food Sciences and Nutrition, ed. B. Caballero, Elsevier, 2nd edn, 2003, p. 1712.
5. M. J. Prival, Cancer-Carcinogens in the Food Chain, in Encyclopedia of Food Sciences and Nutrition, ed. B. Caballero, Elsevier, 2nd edn, 2003, p. 799.
6. S. Pavanello, A. Moretto, C. La Vecchia and G. Alicandro, Regul. Toxicol. Pharmacol., 2023, 139, 105369.
7. K. Jorge, Soft drinks-Chemical Composition, in Encyclopedia of Food Sciences and Nutrition, ed. B. Caballero, Elsevier, 2nd edn, 2003, p. 5346.
8. K. Walton, R. Walker, J. J. M. van de Sandt, J. V. Castell, A. G. A. A. Knapp, G. Kozianowski, M. Roberfroid and B. Schilter, Food Chem. Toxicol., 1999, 37, 1175.
9. R. A. Lawrence and R. M. Lawrence, Maternal Nutrition and Supplements for Mother and Infant, Breastfeeding, Elsevier, 7th edn, 2011, p. 283.
10. A. R. Lobach, A. Roberts and I. R. Rowland, Food Chem. Toxicol., 2019, 124, 385.
11. A. Otabe, F. Ohta, A. Takumi and B. Lynch, Regul. Toxicol. Pharmacol., 2019, 103, 345.
12. S. A. Elmore, J. E. Rehg, T. R. Schoeb, J. I. Everitt and B. Bolon, Food Chem. Toxicol., 2023, 171, 113504.
13. D. Kirkland and D. Gatehouse, Food Chem. Toxicol., 2015, 84, 161.
14. M. Marinovich, C. L. Galli, C. Bosetti, S. Gallus and C. La Vecchia, Food Chem. Toxicol., 2013, 60, 109.
15. C. Wöber and Ç. Wöber-Bingöl, Handb. Clin. Neurol., 2010, 97, 161.
16. R. C. Guy, Aspartame, in Encyclopedia of Toxicology, ed. P. Wexler, Elsevier, 4th edn, 2024, 1, p. 835.
17. S. J. Borghoff, S. S. Cohen, X. Jiang, I. A. Lea, W. D. Klaren, G. A. Chappell, J. K. Britt, B. N. Rivera, N. Y. Choski and D. S. Wikoff, Food Chem. Toxicol., 2023, 172, 113549.
18. S. Yalamanchi, R. Srinath, A. Dobs and K. Acesulfame, in Encyclopedia of Food and Health, ed. B. Caballero, P. M. Finglas and F. Toldrá, 2016, p. 1.
19. M. Carocho, P. Morales and I. C. F. R. Ferreira, Food Chem. Toxicol., 2017, 107, 302.
20. J. R. Araújo, F. Martel and E. Keating, Reprod. Toxicol., 2014, 49, 196.
21. C. Fitch, Saccharin – How Sweet It Is, in Encyclopedia of Food and Health, ed. B. Caballero, P. M. Finglas, F. Toldrá, 2016, p. 659.
22. CCDC 274592 (SCCHRN02): J. L. Wardell, J. N. Low and G. Glidewell, Acta Crystallogr., Sect. E: Struct. Rep. Online, 2005, 61, o1944.
23. CCDC 1211737 (MGSACD10), G. Jovanovski and B. Kamenar, Cryst. Struct. Commun., 1982,(11), 247.
24. CCDC 766997 (WURMOM): S. P. Velaga, V. R. Vangala, S. Basavoju and D. Bostrom, Chem. Commun., 2010, 46, 3562.
25. CCDC 1198198 (KMOTZD), E. F. Paulus, Acta Crystallogr., Sect. B: Struct. Sci., 1975, 31, 1191.
26. CCDC 650627 (RIFVIM): I. Leban, D. Rudan-Tasic, N. Lah and C. Klofutar, Acta Crystallogr., Sect. B: Struct. Sci., 2007, 63, 418.
27. CCDC 650628 (RIFVOS), I. Leban, D. Rudan-Tasic, N. Lah and C. Klofutar, Acta Crystallogr., Sect. B: Struct. Sci., 2007, 63, 418.
28. CCDC 1136816 (DAWGOX), M. Hatada, J. Jancarik, B. Graves and S.-H. Kim, J. Am. Chem. Soc., 1985, 107, 4279.
29. CCDC 608804 (KETXIR) C. Guguta, H. Meekes and R. de Gelder, Cryst. Growth Des., 2006, 6, 2686.
30. CCDC 1161343 (FUPFAX), C. H. Gorbitz, Acta Chem. Scand., 1987, 41, 87.
31. W. Dar, Int. Immunopharmacol., 2024, 135, 112295.
32. N. G. Gonçalves, E. Martinez-Steele, P. A. Lotufo, I. Bensenor, A. C. Goulart, S. M. Barreto, L. Giatti, C. Perim de Faria, M. del Carmen Bisi Molina, P. Caramelli, D. M. Marchioni and C. K. Suemoto, Neurology, 2025, 105, e214023.
33. S. A. A. Shaher, D. F. Mihailescu and B. Amuzescu, Nutrients, 2023, 15, 3627.
34. W. Wu, W. Sui, S. Chen, Z. Guo, X. Jing, X. Wang, Q. Wang, X. Yu, W. Xiong, J. Ji, L. Yang, Y. Zhang, W. Jiang, G. Yu, S. Liu, W. Tao, C. Zhao, Y. Zhang, Y. Chen, C. Zhang and Y. Cao, Cell Metab., 2025, 37, 1075.
35. H. Bai, X. Zuo, C. Zhao, S. Zhang and X. Feng, J. Agric. Food Chem., 2024, 72, 23478.
36. F. Neese, The ORCA program system, Wiley Interdiscip. Rev.: Comput. Mol. Sci., 2012, 2, 73.
37. F. Neese, F. Wennmohs, U. Becker and C. Riplinger, The ORCA quantum chemistry program package, J. Chem. Phys., 2020, 152, 224108.
38. F. Neese, ORCA – An ab initio, density functional and semi-empirical program package, version 5.0.4., 2022.
39. F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297.
40. R. A. Kendall, T. H. Dunning Jr. and R. J. Harrison, J. Chem. Phys., 1992, 96, 6796.
41. F. Weigend, A. Kohn and C. Hattig, J. Chem. Phys., 2002, 116, 3175.
42. Y. Takano and K. N. Houk, J. Chem. Theory Comput., 2005, 1, 70.
43. R. Boča, R. Imrich, J. Štofko, B. Vranovičová and C. Rajnák, Amino Acids, 2024, 56, 5.
44. R. Boča, R. Imrich, J. Štofko, B. Vranovičová and C. Rajnák, J. Phys. Chem. A, 2024, 128, 8088.
45. I. G. Binev, B. A. Stamboliyska and E. A. Velcheva, Spectrochim. Acta, Part A, 1996, A52, 1135.
46. A. D. Popova, E. A. Velcheva and B. A. Stamboliyska, J. Mol. Struct., 2012, 1009, 23.
47. A. B. Brizuela, A. B. Raschi, M. V. Castillo, L. Davies, E. Romano and S. A. Brandán, J. Mol. Struct., 2014, 1074, 144.
48. S. Gunasekaran, P. Rajesh, T. Gnanasambandan, A. Manikandan and S. Seshadri, Int. J. Sci. Technol. Human., 2014, 1, 18.
49. D. Dimič, D. Milenkovič, J. Dimitrič Markovič and Z. Markovič, Phys. Chem. Chem. Phys., 2017, 19, 12970.
50. C.-G. Zhan and D. A. Dixon, J. Phys. Chem. A, 2001, 105, 11534.
51. F. R. Dutra, C. de Souza Silva and R. Custodio, J. Phys. Chem. A, 2021, 125, 65.

---