Negative Mixing Enthalpy

• Negative mixing enthalpy is a thermodynamic measure that indicates exothermic mixing where specific hetero-atomic bonds lower the overall mixture energy compared to unmixed components.
• Experimental techniques like calorimetry and computational methods including DFT and machine learning quantify ΔHₘᵢₓ by elucidating the formation of stable cross-species bonds and favorable electronic interactions.
• The phenomenon underpins improved miscibility, stable alloy formation, and tailored functional properties, influencing materials design in systems from molecular liquids to high-entropy alloys.

Negative mixing enthalpy (ΔHₘᵢₓ < 0) characterizes mixtures and alloys in which hetero-molecular or hetero-atomic interactions yield a more stable, lower-enthalpy state than the weighted average of the unmixed constituents. Negative ΔHₘᵢₓ arises in a wide range of systems—molecular liquids, multicomponent alloys, melt mixtures, high-entropy solids—whenever the formation of specific cross-species bonds or favorable electronic, structural, or packing effects dominates over the energy cost of disrupting self-associations. This property is a major thermodynamic driver for miscibility, compound formation, liquid stability, and design of functional multicomponent materials.

1. Thermodynamic Definition and Quantitative Framework

The enthalpy of mixing, ΔHₘᵢₓ, quantifies the deviation in enthalpy between a real mixture and the corresponding ideal (unmixed) reference state. For a multicomponent system,

$$\Delta H_{\mathrm{mix}} \equiv H_{\mathrm{mixture}} - \sum_{i} x_{i} H_{i}^0$$

where $H_{\mathrm{mixture}}$ is the molar enthalpy of the mixture, $x_{i}$ is the mole fraction, and $H_{i}^0$ the molar enthalpy of pure component $i$ (Chaban et al., 2021). This is equivalent to the excess (non-ideal) enthalpy $H^\mathrm{E}$. Negative values (ΔHₘᵢₓ < 0) indicate that average hetero-interactions are more stabilizing than the sum of homo-interactions, leading to exothermic mixing.

The sign and magnitude of ΔHₘᵢₓ serve as a direct thermodynamic measure of the strength and specificity of unlike (A–B) contacts relative to their like (A–A, B–B) counterparts.

2. Molecular Origin and Physical Mechanisms

Negative mixing enthalpy is rooted in the formation of highly stable hetero-contacts or bonds upon mixing:

• Directional hydrogen bonding: In molecular liquids, strong specific H-bonds such as the O(DESO)…H(MeOH) network drive significant exothermicity, with bond distances (e.g., r_O…H = 0.18 nm) and energetics far surpassing those of the pure liquids (Chaban et al., 2021). Similarly, in 1-alkanol + amine mixtures, OH–NH₂ cross-H-bonds (ΔH_{OH–NH₂} ≈ –41 kJ·mol⁻¹ for methanol–benzylamine) are substantially more exothermic than disruption of homo-H-bonded networks (Sanz et al., 2024).
• Electrostatic and quadrupole effects: Mixtures of π-rich aromatics and quadrupolar molecules (e.g., hexafluorobenzene + benzene) exhibit negative ΔHₘᵢₓ due to strong π–π quadrupolar hetero-association (González et al., 2024).
• Ionic and solvation structure: In molten salt mixtures, negative ΔHₘᵢₓ is correlated with the formation of coordinated oligomeric complexes (e.g., LaCl₆³⁻, La₂Cl_n) whose network energetics are lower than those of the parent melts (Goncharov et al., 2024). Optimized first-shell hetero-coordination enhances exothermicity.
• Electronic and lattice effects in alloys: Charge transfer, electron-density mismatch, and s–s cross-gap interactions can render heteroatomic bonds in metallic or semiconducting alloys more stable than the average of the pure phases. For instance, cross-band-gap s–s repulsion in IVB–IIB halide perovskites yields both upward band-gap bowing and negative ΔHₘᵢₓ at the same time (Zhang et al., 31 Jan 2026). In metallic liquids, work-function and electron-density differences (Miedema P- and Q-terms) are dominant (Deffrennes et al., 2024).

3. Empirical, Computational, and Machine-Learning Approaches

A variety of methods have been deployed for quantifying and predicting negative mixing enthalpy:

• Direct calorimetry: High-resolution differential scanning calorimetry or drop calorimetry establishes ΔHₘᵢₓ directly (e.g., LaCl₃–LiCl–KCl at 873 K, with minima near x₁ ≈ 0.42, ΔHₘᵢₓ ≈ –5.4 kJ·mol⁻¹ (Goncharov et al., 2024); benzylamine + MeOH, ΔHₘᵢₓ = –2.69 kJ·mol⁻¹ at x₁ = 0.49 (Sanz et al., 2024)).
• Molecular simulation and electronic structure: Cluster-based DFT (e.g., hybrid DFT with B3LYP-D3) is used for analyzing PES minima and enthalpic trends in small molecular clusters (Chaban et al., 2021). Large-scale ab initio molecular dynamics and molecular dynamics with explicit enthalpy calculations (NPT ensemble, TIP4P/ε water, UAM-I-EW methanol) provide excess enthalpies for liquid mixtures (Sanchez et al., 24 Mar 2025).
• Thermodynamic models:
  • DISQUAC & ERAS: These frameworks decompose ΔHₘᵢₓ into chemical (H-bond), dispersive, quasi-chemical, and combinatorial terms. They capture S-shaped enthalpy curves and minima positions; model parameters are fit to experiment and display systematic variations with chain length and heteronuclear contact strength (Sanz et al., 2024, Sanz et al., 2024).
  • MIVM: The molecular interaction volume model links ΔHₘᵢₓ to coordination numbers, pair potential parameters, and molar volumes in molten salts (Goncharov et al., 2024).
  • Redlich-Kister (RK) polynomial: Widespread in alloy thermodynamics (CALPHAD), where ΔHₘᵢₓ(x) is fitted as $x_A x_B \sum_{i} L_i (x_A-x_B)^i$ (Vincely et al., 25 Apr 2025, Deffrennes et al., 2024).
• Machine learning: LightGBM and neural networks achieve sub-kJ·mol⁻¹ accuracy for ΔHₘᵢₓ prediction using composition-based and atomic descriptors, outperforming classical empirical models for large data sets of binary and multicomponent alloys (Vincely et al., 25 Apr 2025, Deffrennes et al., 2024). These models can be directly integrated into CALPHAD methodology and improve database completeness, especially for exothermic (ΔHₘᵢₓ < 0) systems with sparse empirical data.

4. Systematic Experimental and Theoretical Trends

Negative mixing enthalpy shows robust, quantitatively reproducible trends across chemical classes and thermodynamic conditions:

System	Composition (min)	ΔHₘᵢₓ,min	Mechanism	Reference
Water–methanol	x_m ≈ 0.42–0.6	–3.8 kJ/mol	Strong H-bonding	(Sanchez et al., 24 Mar 2025)
1-alkanol–cyclohexylamine	x₁ ≈ 0.5	–3.8 to –2.3 kJ/mol	OH–NH₂ H-bonds, dispersion	(Sanz et al., 2024)
Benzylamine–1-alkanol	x₁ ≈ 0.5–0.7	–2.7 to –0.7 kJ/mol	OH–NH₂ cross-association	(Sanz et al., 2024)
Hexafluorobenzene–aromatic hydrocarbon	x₁ = 0.5	–0.5 to –1.7 kJ/mol	Quadrupole–quadrupole π–π	(González et al., 2024)
Molten LaCl₃–(LiCl–KCl)	x₁ ≈ 0.42	–5.4 kJ/mol	Oligomeric/strong La–Cl complexes	(Goncharov et al., 2024)
Liquid Fe–Al, Fe–Ti	x₁ = 0.5	L₁ ≈ –75,000 J/mol	Charge/e-structure, short-range order	(Vincely et al., 25 Apr 2025)
Cu–Zn–Sn–S (CZTS) film	bulk composition	–493 kJ/mol	Compound semiconducting alloy formation	(Baryshev et al., 2014)

The magnitude of exothermicity decreases with increased steric hindrance, loss of H-bonding acceptor ability, or addition of longer aliphatic chains (Sanz et al., 2024, Sanz et al., 2024). For metallic and semiconducting alloys, negative ΔHₘᵢₓ correlates with large differences in work function, electron density, or optimized s–s electronic interactions across the band gap (Zhang et al., 31 Jan 2026, Deffrennes et al., 2024).

5. Methodological Advances and Modelling Limitations

Advancements in both experimental and computational protocols have improved the fidelity of ΔHₘᵢₓ measurement and prediction:

• Experimental precision: Modern Tian–Calvet calorimetry achieves ΔHₘᵢₓ uncertainties down to 1 % for multicomponent molecular systems (Sanz et al., 2024).
• Model limitations: Cluster-based QM calculations may neglect bulk many-body and cooperative effects; classical force fields can underpredict enthalpy magnitudes due to limitations in cross-species parameterization (Chaban et al., 2021, Sanchez et al., 24 Mar 2025).
• Machine-learning uncertainty: LightGBM uncertainty < 1 kJ/mol; for neural networks trained on adequately diverse binaries, errors are below typical calorimetric error bars (Vincely et al., 25 Apr 2025, Deffrennes et al., 2024).

For accurate macroscopic ΔHₘᵢₓ, periodic boundary techniques (DFT supercells, SQS, high-entropy sampling) are critical in large, disordered, or compositionally complex solid solutions (Novick et al., 2022, Zhang et al., 31 Jan 2026).

6. Implications for Materials Design and Thermodynamic Modelling

Negative mixing enthalpy is a central parameter in the rational design of stable multicomponent systems:

• Phase stability and compound formation: Strongly negative ΔHₘᵢₓ prevents phase separation in liquid alloys, stabilizes multinary compounds over binary precursors (e.g., CZTS formation from Cu₂S, ZnS, SnS₂: ΔHₘᵢₓ = –493 kJ/mol (Baryshev et al., 2014)).
• Functional property tuning: In halide perovskites, the coexistence of negative ΔHₘᵢₓ and upward band gap bowing enables direct band gap engineering (Zhang et al., 31 Jan 2026).
• CALPHAD and database augmentation: Machine-learning–driven predictions of negative ΔHₘᵢₓ permit rapid filling of data “gaps,” especially for alloy systems lacking experimental enthalpy data (Vincely et al., 25 Apr 2025, Deffrennes et al., 2024).
• Melt processing and separation: For molten salt nuclear technologies, negative ΔHₘᵢₓ underpins optimized electrolyte design and element activity control (Goncharov et al., 2024).

Guidelines emphasize maximizing hetero-associative bonding, leveraging electronic or structural mismatch, and employing compositional “high-entropy” strategies to further enhance the thermodynamic driving force for mixing.

7. Representative Equations and Modelling Protocols

Key equations underpinning quantitative analysis are used across the literature:

• Generic definition:

$$\Delta H_{\mathrm{mix}} = H_{\mathrm{mixture}} - \sum_i x_i\,H_i^0$$

(Chaban et al., 2021)

• Redlich–Kister expansion (binary alloy):

$$\Delta H_{\mathrm{mix}}(x) = x_A x_B \sum_{n=0}^N L_n (x_A - x_B)^n$$

(Vincely et al., 25 Apr 2025, Deffrennes et al., 2024)

• Empirical (Miedema) model:

$$\Delta H_{\mathrm{mix}}^{\mathrm{Miedema}} = c_A H_{\mathrm{inter}}(B\text{ in }A) + c_B H_{\mathrm{inter}}(A\text{ in }B)$$

(Deffrennes et al., 2024)

• MIVM for molten salts:

$$\Delta H_{\mathrm{mix}} = \sum_{i,j} x_i x_j Z_i V_{m,j} B_{ji}(1+\ln B_{ji}) - \sum_{i,j} x_i x_j V_{m,j} B_{ji} \ln B_{ji}$$

(Goncharov et al., 2024)

Each method requires careful parametrization (e.g., exchange-correlation functionals in DFT, group parameters in DISQUAC, coordination statistics in MIVM) for the system of interest.

---

Negative mixing enthalpy constitutes a rigorous, experimentally accessible, and theoretically tractable signature of strong hetero-interaction in multicomponent systems, underpinning both practical materials processing and fundamental thermodynamic theory. Its determination and modelling remain at the core of solution chemistry, alloy theory, and computational materials design (Chaban et al., 2021, Vincely et al., 25 Apr 2025, Sanz et al., 2024, Baryshev et al., 2014, Zhang et al., 31 Jan 2026, Deffrennes et al., 2024, Goncharov et al., 2024, Sanz et al., 2024, González et al., 2024, Sanchez et al., 24 Mar 2025, Novick et al., 2022).