---
title: 'Ethaline: Deep Eutectic Solvent Insights'
url: https://www.emergentmind.com/topics/ethaline
type: topic
---

# Ethaline: Deep Eutectic Solvent Insights

Ethaline is a prototypical deep eutectic solvent (DES) formed by mixing choline chloride (ChCl) and ethylene glycol (EG) in a 1:2 molar ratio, often written equivalently as a 2:1 mixture of EG and ChCl. It is commonly classified as a type-III DES and is among the most widely studied choline-chloride-based eutectics. Its defining feature is a pronounced melting-point depression driven by strong hydrogen bonding between chloride and hydroxyl-bearing species, producing a homogeneous, viscous liquid with an extended hydrogen-bond network near room temperature. That combination has made ethaline central in studies of electrochemistry, biomass processing, transport in glass-forming liquids, hydration effects, and confinement. At the same time, recent work has qualified its “green” status by reporting partial decomposition into chloromethane and dimethylaminoethanol under mild conditions [2510.05882] [2410.05498].

## 1. Definition, composition, and eutectic character

Ethaline consists of choline chloride, \([\mathrm{HO\!-\!CH_2\!-\!CH_2\!-\!N(CH_3)_3}]^+\mathrm{Cl^-}\), and ethylene glycol, \(\mathrm{HO\!-\!CH_2\!-\!CH_2\!-\!OH}\), in a strict 1:2 molar ratio of ChCl:EG or, equivalently, 2:1 EG:ChCl. In shorthand form,
\[
\mathrm{Ethaline} = 2\,\mathrm{EtGl} + 1\,\mathrm{ChCl}
\quad\longrightarrow\quad
\{\,\mathrm{EtGl},\,\mathrm{Ch^+},\,\mathrm{Cl^-}\,\}.
\]
It is routinely prepared by mixing the components until homogeneous; reported procedures include stirring ChCl and EG at \(350\,\mathrm{K}\) for \(24\,\mathrm{h}\) to obtain a liquid with \(<1\,\mathrm{wt}\,\%\) water, and heating at \(60\,^\circ\mathrm{C}\) for \(30\,\mathrm{min}\) until homogenization [1902.02207] [2103.14312].

Its eutectic identity is tied to strong, directional hydrogen bonding between chloride and hydroxyl groups. One study states that the eutectic mixture melts near \(288\,\mathrm{K}\), more than \(100\,\mathrm{K}\) below the individual components, because hydrogen bonding between the quaternary ammonium cation and EG hydroxyls disorders the lattice and depresses the melting point [2510.05882]. A force-field study likewise describes the mixture as behaving like a single “liquid salt,” with the melting point dropping to near room temperature because the neutral hydrogen-bond donor interacts strongly and directionally with the hydrogen-bond acceptor chloride [2109.14007].

Ethaline has often been presented as attractive because of low toxicity, biodegradability, and low cost, and because it is useful in electrochemistry, biomass processing, metal leaching, and carbon capture [2510.05882] [2410.05498]. That application profile, however, must be read alongside later evidence for decomposition chemistry, discussed below.

## 2. Hydrogen-bond network and microscopic structure

At the molecular level, ethaline is characterized by an extensive, percolating hydrogen-bond network in which chloride is the dominant acceptor. X-ray and molecular-dynamics studies cited in the decomposition work report that each \(\mathrm{Cl^-}\) typically accepts \(3\)–\(4\) hydrogen bonds from nearby hydroxyl donors, stabilizing the liquid [2410.05498]. A hydrated-state NMR study similarly describes EG–\(\mathrm{Cl^-}\), EG–EG, and \(\mathrm{Ch^+}\)–\(\mathrm{Cl^-}\) contacts as the dominant interactions, yielding a percolating H-bond matrix [2410.11447].

More explicit local-structure metrics were reported in the fine-tuned polarizable CL&Pol force-field study. In that model, the chloride–hydroxyl radial distribution functions exhibit first-shell peaks at \(g_{(\mathrm{Cl-HO})}(r)_{\max}\simeq 9.9\) at \(r\simeq 2.05\,\text{\AA}\) for choline hydroxyl protons and \(g_{(\mathrm{Cl-HE})}(r)_{\max}\simeq 7.7\) at \(r\simeq 2.00\,\text{\AA}\) for ethylene-glycol hydroxyl protons. Integration to the first minimum near \(3.2\,\text{\AA}\) gives coordination numbers \(N_{\mathrm{Cl-HO}}\simeq 2.0\) and \(N_{\mathrm{Cl-HE}}\simeq 4.0\), implying that each chloride accepts on average about two OH\(\cdots\)\(\mathrm{Cl^-}\) bonds from choline and about four from EG [2109.14007]. In the same study, the total X-ray structure factor \(S(q)\) shows the characteristic charge-alternation peak around \(q\simeq 14\,\mathrm{nm^{-1}}\), corresponding to \(\sim 0.45\,\mathrm{nm}\) real-space correlations.

The structural description is not independent of the interaction model. The original CL&Pol formulation produced unphysically strong pre-peaks and anti-peaks at very low \(q\) (\(\sim 2.4\,\mathrm{nm^{-1}}\)), interpreted as spurious nano-segregation on \(\sim 3\,\mathrm{nm}\) length scales. After re-optimizing \(\sigma(\mathrm{Cl-OHG})=0.345\,\mathrm{nm}\) and \(\sigma(\mathrm{Cl-HO})=0.356\,\mathrm{nm}\) against ab initio RDFs of Alizadeh et al., and applying Tang–Toennies damping with \(b=4.5\,\text{\AA}^{-1}\) and \(n=4\), the low-\(q\) features shift to \(\sim 4.3\,\mathrm{nm^{-1}}\), corresponding to a more realistic \(\sim 1.5\,\mathrm{nm}\) heterogeneity [2109.14007].

This structural picture is consistent across several studies: ethaline is not a simple binary liquid but a strongly associated ionic-hydrogen-bonded network in which chloride-centered coordination plays the organizing role. A plausible implication is that small changes in hydrogen-bond donor identity, water content, or confinement can alter dynamics without necessarily destroying the underlying network.

## 3. Bulk transport, relaxation, and multiscale dynamics

Bulk dynamical studies place ethaline among strongly glass-forming ionic liquids with marked non-Arrhenius transport. Broadband dielectric spectroscopy from \(0.1\,\mathrm{Hz}\) to \(\approx 3\,\mathrm{GHz}\) and from \(350\,\mathrm{K}\) down to \(\approx 163\,\mathrm{K}\) showed that the dc conductivity decreases by \(\approx 5\)–\(6\) decades on cooling and is well described by a Vogel–Fulcher–Tammann law,
\[
\sigma_{dc}(T)=\sigma_0 \exp\!\left[-\frac{D_\sigma T_0^\sigma}{T-T_0^\sigma}\right].
\]
For one study of neat ethaline, the fitted parameters were \(T_0^\sigma=111\,\mathrm{K}\), \(D_\sigma=13.7\), and \(\sigma_0=81\,\Omega^{-1}\mathrm{cm}^{-1}\); the same work reported \(T_0^\tau=113\,\mathrm{K}\), \(D_\tau=13.2\), \(\tau_0=2.8\times 10^{-14}\,\mathrm{s}\), fragility index \(m\approx 60\), \(T_g\approx 155\,\mathrm{K}\), and room-temperature conductivity \(\sigma_{dc}\approx 1\times 10^{-2}\,\Omega^{-1}\mathrm{cm}^{-1}\) [1902.02207].

The relation between conductivity, viscosity, and dipolar relaxation has been interpreted in two closely related but not identical ways. One dielectric study found \(\rho_{dc}\propto \langle\tau\rangle\) over the entire temperature range and described this proportionality in terms of a Debye–Stokes–Einstein-type relation, with a “revolving-door” or “paddle-wheel” mechanism in which dipolar rotations open transient pathways for ion hopping [1902.02207]. A later rheology-plus-dielectric analysis concluded that, for ethaline, \(\sigma_{dc}\propto \eta^{-1}\) with Walden exponent \(\alpha=1\) and \(\sigma_{dc}\eta=\mathrm{const}\simeq 2\times 10^{-3}\,\mathrm{Pa\cdot s\cdot S\;cm^{-1}}\), and argued that the data can be understood without invoking a revolving-door mechanism, instead as viscosity-controlled translational–rotational coupling [2101.11042]. The common empirical point is that ionic transport, dielectric \(\alpha\)-relaxation, and viscosity display essentially identical VFT temperature dependences in ethaline.

Neutron scattering and NMR extend that description across shorter length and time scales. A 2025 multiscale study combined pulsed-field-gradient NMR with time-of-flight and backscattering QENS on isotopically labelled samples. On the micrometer scale, the self-diffusion coefficients obey classical hydrodynamics: \(D_{\mathrm{EG}}(\text{pure}) > D_{\mathrm{EG}}(\text{in ethaline}) > D_{\mathrm{ChCl}}(\text{in ethaline})\), with \(D_{\mathrm{EG}}/D_{\mathrm{ChCl}}\approx 1.5\)–\(1.7\). Using the Stokes–Einstein relation,
\[
D=\frac{k_B T}{6\pi \eta R_h},
\]
with \(R_h(\mathrm{EG})\simeq 1.5\)–\(2\,\text{\AA}\) and \(R_h(\mathrm{ChCl})\simeq 3.3\,\text{\AA}\), the expected ratio \(D_{\mathrm{EG}}/D_{\mathrm{ChCl}}\simeq 1.6\)–\(2.2\) agrees with experiment [2509.18896].

At nanometer scales that hydrodynamic size dependence disappears. The same study found that at \(T\approx 300\,\mathrm{K}\),
\[
D_{\mathrm{QENS}}(\mathrm{EG\ in\ ethaline}) \simeq
D_{\mathrm{QENS}}(\mathrm{ChCl\ in\ ethaline}) \simeq
D_{\mathrm{QENS}}(\mathrm{pure\ EG})
\simeq (0.8\text{–}1.2)\times 10^{-10}\,\mathrm{m^2\,s^{-1}},
\]
indicating dynamically correlated supramolecular units held together by \(\mathrm{Cl^-}\cdots\mathrm{H-O}\) hydrogen bonds [2509.18896]. The sub-nanometer motions preceding Fickian diffusion were described as jumps of length \(\ell\simeq 0.7\)–\(1.0\,\text{\AA}\) separated by residence times \(\tau_j\simeq 8\)–\(30\,\mathrm{ps}\), with localized motions of amplitude \(\sqrt{\langle r^2\rangle}\lesssim 1\,\text{\AA}\) and correlation times \(\tau_L\simeq 2\)–\(3\,\mathrm{ps}\) [2509.18896].

A cholinium-selective QENS study on IRIS at ISIS resolved these same ideas with an explicit cage-jump model on \(\sim 1\)–\(10\,\text{\AA}\) and \(\sim\mathrm{ps}\)–\(10\,\mathrm{ps}\) scales:
\[
S(Q,E) = [A(Q)L_{\mathrm{loc}}(Q,E,\Gamma_{\mathrm{loc}})
+ (1-A(Q))L_j(Q,E,\Gamma_j)] \otimes R(E).
\]
The jump component was fitted by the Singwi–Sjölander expression,
\[
\Gamma_j(Q)=D_j Q^2 [1+(D_j Q^2 \tau)]^{-1},
\]
and the mean jump length followed from the Chudley–Elliott relation,
\[
D_j=\frac{L^2}{6\tau}.
\]
For ethaline, \(D_j\) increased from \(1.16\pm 0.09\) to \(2.43\pm 0.19\) in units of \(10^{-6}\,\mathrm{cm^2\,s^{-1}}\) between \(300\) and \(345\,\mathrm{K}\), while \(\tau\) decreased from \(26.4\) to \(5.32\,\mathrm{ps}\); \(L\) remained essentially temperature invariant at \(\simeq 1.05\,\text{\AA}\). Compared with glyceline and reline, ethaline showed the shortest residence times at all temperatures and the highest cholinium self-diffusion coefficients, whereas reline exhibited larger jump lengths and crossed over to higher \(D_j\) than glyceline above \(\simeq 330\,\mathrm{K}\) [2510.05882].

## 4. Hydration, glassy freezing, and aqueous regimes

Hydration reorganizes ethaline’s phase behavior without immediately destroying its native network. Dielectric studies on mixtures with water mass fraction \(W=0\)–\(90\,\mathrm{wt}\,\%\) distinguished a “water-in-DES” regime for \(W<40\,\mathrm{wt}\,\%\), where the liquid remains macroscopically homogeneous, from a “DES-in-water” regime for \(W>40\,\mathrm{wt}\,\%\), where cooling induces phase separation into ice plus a maximally freeze-concentrated DES solution of composition \(W'\approx 30\,\mathrm{wt}\,\%\) [2103.14312]. A calorimetric phase-diagram study reached the same threshold, reporting \(W_g'\approx 30\%\) as the boundary above which crystallization occurs and below which neat and moderately hydrated mixtures remain glass-forming [2201.01067].

In the homogeneous regime, conductivity and dipolar reorientation remain inversely proportional, \(\sigma\propto \tau^{-1}\), over the timescale covered by the dielectric study. For \(W\le 30\,\mathrm{wt}\,\%\), the coupling exponent is \(n\approx 1.00\), corresponding to classical Debye–Stokes–Einstein behavior; for \(W>40\,\mathrm{wt}\,\%\), \(n\approx 0.90\), indicating fractional decoupling once freeze-concentration and ice domains emerge [2103.14312]. The same study reported that the fragility index decreases from \(m\simeq 48\) for neat ethaline to \(\simeq 40\) at \(W\approx 30\,\mathrm{wt}\,\%\), while the stretching parameter inferred from \(\beta_{\mathrm{KWW}}\simeq (\alpha\beta)^{1/1.2}\) rises from \(\sim 0.60\) at \(W=0\) to \(\sim 0.70\) for \(W\ge 40\,\mathrm{wt}\,\%\).

Component-selective magnetic resonance studies refine that picture. In a sample containing \(17\,\mathrm{wt}\,\%\) water, stimulated-echo experiments yielded \(\beta\approx 0.46\) for both \(^{2}\mathrm{H}\) and \(^{17}\mathrm{O}\) probes, showing no increase in dynamic heterogeneity upon hydration [2410.11447]. The same work found that water and ethylene glycol display very similar mobilities over \(150\)–\(300\,\mathrm{K}\), and that adding \(\approx 17\,\mathrm{wt}\,\%\) \(\mathrm{H_2O}\) reduces \(\eta_0\) and \(\tau_J\) by factors of \(2\)–\(3\) near \(160\)–\(180\,\mathrm{K}\). It also reported that water acts as an antifreeze, shifting \(T_g\) down by \(\approx 9\,\mathrm{K}\) at \(17\,\mathrm{wt}\,\%\) \(\mathrm{H_2O}\) [2410.11447].

These results support a consistent description of moderate hydration as plasticization rather than demixing. The aqueous ethaline studies explicitly state that, in the water-in-DES regime, water “blends in” with the EG hydrogen-bond network and does not detectably increase heterogeneity [2410.11447]. A plausible implication is that the principal crossover with increasing water is not immediate local disruption but the onset of phase-separated freeze concentration near the \(30\)–\(40\,\mathrm{wt}\,\%\) threshold.

## 5. Nanoconfinement, phase behavior, and interfacial thermodynamics

When ethaline is confined in mesoporous silica, neutron scattering indicates substantial structural robustness. In cylindrical mesopores of SBA-15 with \(D_p\approx 8.1\,\mathrm{nm}\) and MCM-41 with \(D_p\approx 3.5\,\mathrm{nm}\), neutron diffraction found no evidence of core-shell segregation within the pore cross-section. SBA-15 filled with ethaline showed Bragg intensities reduced in proportion to contrast without \(q\)-shift, consistent with uniform filling across the \(4.05\,\mathrm{nm}\) pore radius. In MCM-41, the smaller intensity reduction was reproduced by the homogeneous-filling model if only \(\sim 60\%\) of the pore volume was filled, again without evidence for core-shell segregation or microphase separation [2607.07090].

The confined dynamics preserve the bulk jump-diffusion phenomenology but with longer waiting times between jumps. At \(298\,\mathrm{K}\), the translational linewidths were fitted with
\[
\Gamma(Q)=\frac{D_T Q^2}{1+D_T Q^2 \tau_0},
\qquad
D_T=\frac{\langle r^2\rangle}{6\tau_0}.
\]
Representative values were \(D_T\approx 1.5\times 10^{-11}\,\mathrm{m^2\,s^{-1}}\) and \(\tau_0\approx 1.0\,\mathrm{ps}\) in bulk ethaline, \(D_T\approx 1.3\times 10^{-11}\,\mathrm{m^2\,s^{-1}}\) and \(\tau_0\approx 4\)–\(8\,\mathrm{ps}\) in SBA-15, and \(D_T\approx 1.1\times 10^{-11}\,\mathrm{m^2\,s^{-1}}\) and \(\tau_0\approx 10\,\mathrm{ps}\) in MCM-41 [2607.07090]. Localized in-cage motion was described through the elastic incoherent structure factor,
\[
A_0(Q)=\left[\frac{3j_1(Qa)}{Qa}\right]^2,
\]
with \(a\approx 1.0\,\text{\AA}\) and \(t_L\approx 3\,\mathrm{ps}\) in bulk, \(a\approx 0.9\,\text{\AA}\) and \(t_L\approx 4\,\mathrm{ps}\) in SBA-15, and \(a\approx 0.8\,\text{\AA}\) and \(t_L\approx 5\)–\(6\,\mathrm{ps}\) in MCM-41 [2607.07090]. The principal confinement effect is therefore a substantial increase in residence and relaxation times, not a collapse of translational mobility.

Calorimetric studies of hydrated ethaline under confinement add a thermodynamic dimension. For bulk and confined systems alike, \(W_g'\approx 30\%\) marks the threshold above which ice crystallizes and below which glass-forming solutions are observed [2201.01067]. The melting-point depression in confinement was analyzed using an extended Gibbs–Thomson–Raoult relation,
\[
T_m(r_p,a_w)=T_{\mathrm{bulk}}
-\frac{2\gamma_{sl}V_{\mathrm{ice}}}{r_p\Delta H_m}
+\frac{R T_m}{\Delta H_m}\ln a_w.
\]
The study reported melting depressions of up to \(50\,\mathrm{K}\), good agreement of the model for bulk and SBA-15 at high water content, and systematic deviation in MCM-41, where measured \(T_m\) values lie above the predictions [2201.01067]. Back-calculated water activities \(a_w(W)\) are slightly elevated in SBA-15 relative to bulk and exceed bulk values in MCM-41, even exceeding unity in extreme confinement. Behboudi et al. vapor-pressure data for bulk were reported to agree with calorimetry-derived \(a_w\) values [2201.01067].

Taken together, the confinement studies indicate that ethaline can remain structurally homogeneous and dynamically recognizable even inside nanometer-scale pores, while its freezing, water activity, and cage lifetimes become strongly geometry dependent.

## 6. Chemical stability, decomposition pathways, and computational description

The main qualification to ethaline’s benign reputation is chemical instability associated with the choline chloride component. A 2024 study reported partial decomposition of ethaline at room temperature into toxic chloromethane and dimethylaminoethanol, and concluded that choline chloride is susceptible to decomposition in strongly hydrogen-bound mixtures [2410.05498]. Experimentally, isothermal TGA at \(60\,^\circ\mathrm{C}\) for \(4\,\mathrm{h}\) showed \(17\,\mathrm{wt}\,\%\) mass loss. GC–TCD and GC–MS identified chloromethane in the headspace at retention \(\approx 12.6\,\mathrm{min}\), dimethylaminoethanol in the condensed phase at retention \(\approx 10.9\,\mathrm{min}\), with \([\mathrm{DMAE}]\approx 26\,\mathrm{mM}\) after \(30\,\mathrm{min}\) at \(60\,^\circ\mathrm{C}\), together with minor trimethylamine and 2-methoxyethanol [2410.05498].

The principal pathway was written as an \(\mathrm{S_N2}\) process,
\[
\ce{[HO-CH2-CH2-N(CH3)3]^+ Cl^- -> CH3Cl + HO-CH2-CH2-N(CH3)2},
\]
initiated by hydrogen-bond fluctuations that bind chloride near reaction sites [2410.05498]. The same work reported a vacuum barrier \(\Delta G^\ddagger \approx 1.68\,\mathrm{eV}\), a “rigid” minimum-energy-pathway barrier of \(\approx 3.0\,\mathrm{eV}\), and an average barrier of \(\approx 2.07\,\mathrm{eV}\) after \(200\,\mathrm{ps}\) solvent relaxation plus \(500\,\mathrm{fs}\) NVE AIMD at \(298\,\mathrm{K}\), with fluctuations of \(\pm 0.3\)–\(0.7\,\mathrm{eV}\) at the critical coordinate. At the transition state, \(\mathrm{Cl^-}\) is held by two strong hydrogen bonds of \(\sim 0.7\,\mathrm{eV}\) each [2410.05498].

That decomposition study also introduced a quantum-chemically accurate workflow based on PBE0\((68)\)-D3 with \(\alpha=0.6851\), TZVP-MOLOPT, and CP2K, benchmarked against CCSD(T) for ethylene-glycol ionization potentials. Active learning through “FLARE + OPLS” generated \(\approx 800\) training snapshots spanning intramolecular distortions, intermolecular hydrogen-bond configurations, and reactive \(\mathrm{S_N2}\) geometries; the final Allegro-v2 model reproduced bulk RDFs and thermodynamics at \(298\,\mathrm{K}\) within \(2\,\mathrm{meV/atom}\), gave force RMSE \(\approx 0.08\,\mathrm{eV/\AA}\), and reproduced the \(\mathrm{S_N2}\) barrier within \(0.1\,\mathrm{eV}\) of PBE0\((68)\)-D3 [2410.05498].

Independent simulation work has focused on equilibrium structure and transport rather than reaction chemistry. The corrected CL&Pol model reproduced density, viscosity, and surface tension at \(323\,\mathrm{K}\) and \(1\,\mathrm{atm}\) with close agreement to experiment: \(\rho=1.114\pm 0.002\,\mathrm{g\,cm^{-3}}\) versus \(1.10\,\mathrm{g\,cm^{-3}}\), \(\eta=18.5\pm 1.2\,\mathrm{cP}\) versus \(18.8\,\mathrm{cP}\), and \(\gamma=48.1\,\mathrm{mN\,m^{-1}}\) versus \(47.6\,\mathrm{mN\,m^{-1}}\). In the same comparison, CL&Pol gave \(D^+=4.6\), \(D^0=13.3\), and \(D^-=11.7\) in units of \(10^{-11}\,\mathrm{m^2\,s^{-1}}\), while experimental values were \(D^+=7.5\) and \(D^0=13.2\) [2109.14007]. These results show that accurate modeling of ethaline requires both chemically realistic short-range hydrogen-bond structure and controlled polarization response.

Ethaline therefore occupies a technically important but nontrivial position within the DES literature. It is prototypical in stoichiometry, hydrogen-bond organization, glassy transport, and confinement behavior, yet it also exposes a central limitation of choline-chloride-based design: the same hydrogen-bond environment that produces strong eutectic behavior can stabilize reactive configurations that compromise chemical stability.

Source: https://www.emergentmind.com/topics/ethaline