The need for better sample preparation in this field is genuine. Drugs such as antidepressants, sedative-hypnotics, and antipsychotics typically have narrow therapeutic windows, meaning the difference between an effective dose and a toxic one can be small. Accurate measurement of blood concentrations is therefore central to clinical toxicology and therapeutic drug management. Yet plasma is a difficult matrix for chromatographic analysis, loaded with proteins and phospholipids that can interfere with detection and damage instruments. Some form of extraction and cleanup is essential before gas or liquid chromatography can deliver trustworthy numbers, and the extraction step often determines how fast, cheap, and environmentally friendly the entire workflow can be.

Dispersive liquid-liquid microextraction, or DLLME, has long been a favorite miniaturized technique because it is simple in principle and delivers strong enrichment. A few dozen microliters of extraction solvent are dispersed into the aqueous sample, analytes partition into the tiny droplets, and the solvent phase is then recovered for analysis. The trouble has always been the solvent itself. Traditional extractants such as chloroform or trifluoroacetic acid are toxic and volatile, and even newer green alternatives have struggled with two practical bottlenecks: many weakly basic drugs require careful pH adjustment to extract efficiently, and recovering a few microliters of floating solvent from a centrifuge tube is fiddly and error-prone.

The team behind the new study, led by researchers at the Seventh People’s Hospital of Zhengzhou and Zhengzhou University, attacked both problems at once. Their key invention is a type V alkaline natural deep eutectic solvent, abbreviated ANADES. Deep eutectic solvents form when a hydrogen bond donor and a hydrogen bond acceptor are mixed and interact through hydrogen bonding to produce a liquid with a lower melting point than either component. Type V solvents are a recent category in which both components are neutral, non-ionic molecules. That neutrality turned out to be the crucial design feature: because menthol, carvone, and eucalyptol carry no acidic groups, they do not neutralize added base, allowing ammonia to dissolve into the solvent matrix through hydrogen bonding and remain there as a built-in alkaline reservoir.

The choice of base was not arbitrary. When the researchers tested saturated solutions of sodium hydroxide, potassium hydroxide, and ammonia, only ammonia produced a functional alkaline solvent. Sodium and potassium hydroxides dissociate completely into ions that, by virtue of their charge and high hydration energy, cannot partition into the hydrophobic solvent phase and are simply removed during centrifugation. Ammonia, by contrast, can hydrogen-bond with the menthol hydroxyl and the carbonyl groups of carvone or eucalyptol, integrating directly into the solvent. Measurements confirmed the effect: pure solvent gave near-neutral pH of about 6.8 to 7.0 in water and plasma, while the alkaline version shifted samples to roughly pH 9.3 to 9.5, reproducibly across plasma from six different donors.

That alkalinity matters because the seven target drugs, including lidocaine, amitriptyline, imipramine, promethazine, diazepam, chlorpromazine, and clozapine, are weakly basic compounds with conjugate-acid pKa values between roughly 8.0 and 10.5. At an extraction pH of 9.49, above their pKa values, the drugs exist predominantly as neutral, uncharged molecules that partition readily into the hydrophobic solvent. Under acidic conditions they become protonated and water-soluble, and extraction collapses. Comparative experiments confirmed the pattern: alkaline conditions pushed extraction recoveries above 80 percent for all seven analytes, while acidic conditions performed worst. The built-in pH regulation thus eliminates an entire operational step while simultaneously improving the chemistry.

The second innovation addresses solvent recovery. Rather than centrifuging to collect floating droplets, the team added iron oxide magnetic nanoparticles, which associate with the solvent droplets, likely through hydrogen bonding between surface hydroxyl groups on the particles and the polar functional groups of the solvent components. A magnet then pulls the solvent-coated particles to the bottom of the tube in seconds. Optimization showed that 50-nanometer particles gave the best combination of recovery and reproducibility, and that about 12 milligrams sufficed to capture the entire solvent phase. Desorption was equally streamlined: one minute of vortexing in 200 microliters of ethyl acetate released the analytes for injection into a gas chromatograph coupled to triple quadrupole mass spectrometry.

Getting to those numbers required systematic optimization. Single-factor experiments identified the menthol-to-carvone molar ratio of 1:1 as optimal, along with 23 microliters of solvent, 400 microliters of ammonia, and a two-minute extraction, after which longer mixing brought no benefit because the tiny droplets reach partitioning equilibrium almost immediately. A three-factor Box-Behnken response surface design then refined the settings while accounting for interactions among variables. The resulting regression model showed an R-squared of 0.9897 and a non-significant lack-of-fit statistic, and experimental recoveries under the predicted optimum, 97.9 percent with a standard deviation of 10.5 percent, matched the model’s predictions closely.

Validation performance was solid across the board. Calibration curves were linear from 5 to 1000 nanograms per milliliter for lidocaine, amitriptyline, and promethazine, and from 10 to 1000 for the other four drugs, with determination coefficients between 0.990 and 0.999. Limits of detection ranged from 0.61 to 3.36 nanograms per milliliter, limits of quantification from 2.05 to 9.89, extraction recoveries from 85.01 to 99.89 percent, and enrichment factors from 6.65 to 8.79. Precision, expressed as relative standard deviation, stayed below 13.3 percent, and spiked recoveries fell between 95.1 and 114.9 percent. Matrix effects, measured by comparing plasma-spiked samples with neat standards at three concentrations, ranged from 84.9 to 108.0 percent, indicating negligible interference from the plasma itself. In single-blind tests on fortified plasma at 25, 40, and 80 nanograms per milliliter, all seven drugs were detected with recoveries of 95.1 to 112.7 percent.

The method’s green credentials were assessed with three complementary metrics. The AGREEprep score of 0.63 clears the 0.6 threshold for classification as green, the ComplexMoGAPI score of 86 reflects strong environmental performance, and the BAGI score of 67.5 confirms practical applicability alongside sustainability. The authors are candid that the ethanol used for protein precipitation and the ethyl acetate used for desorption are still volatile organic compounds, but per-sample consumption is minimal. Compared with previously reported DLLME methods for CNS drugs, which relied on chloroform, ionic liquids, or multi-solvent mixtures, required manual pH adjustment, and often took more than ten to thirty minutes, the new approach uses just 23 microliters of non-toxic extractant and finishes extraction in about five minutes overall. The authors acknowledge limitations: validation relied on spiked plasma rather than authentic patient specimens, and the dependence on GC-MS/MS instrumentation limits accessibility in some laboratories. Still, with linear ranges covering the therapeutic windows of five of the seven drugs and detection limits at or below the lower clinical bounds, the technique stands as a credible candidate for routine therapeutic drug monitoring, pending confirmation in larger real-world sample sets.

Subject of Research: Development of an alkaline natural deep eutectic solvent-based magnetic dispersive liquid-liquid microextraction method for quantifying CNS drugs in plasma

Article Title: A novel type V natural deep eutectic solvent with built-in pH-regulation for magnetic dispersive liquid-liquid microextraction of central nervous system drugs in plasma samples

Article References: Han, W.-C., Yang, Y.-L., Yang, X.-B., Zhang, H.-J., Chen, Y.-Y., Pei, X.-R., Yuan, D.-D., & Chen, D. (2026). A novel type V natural deep eutectic solvent with built-in pH-regulation for magnetic dispersive liquid-liquid microextraction of central nervous system drugs in plasma samples. Frontiers in Chemistry, 14, 1986305. https://doi.org/10.3389/fchem.2026.1986305

Image Credits: AI Generated

DOI: 10.3389/fchem.2026.1986305

Keywords: natural deep eutectic solvent, magnetic dispersive liquid-liquid microextraction, central nervous system drugs, plasma, GC-MS/MS, green analytical chemistry, pH regulation, therapeutic drug monitoring, sample preparation, menthol, carvone, response surface methodology