Methyl tert-butyl ether is a branched aliphatic ether with molecular formula C₅H₁₂O and molar mass 88.15 g/mol. In fuel oxygenate applications it was blended into reformulated gasoline at volume fractions up to 15% in the United States before widespread groundwater detections led to state-level bans and subsequent use restrictions. The compound is a colourless volatile liquid with a boiling point of 55.2 °C and a density near 0.7405 g/mL at 20 °C. Its aqueous solubility is central to environmental transport modelling, yet two numbers dominate secondary literature and safety data sheets: 4.8 g per 100 g and 26 g per Litre. These values cannot be reconciled as unit variants. The first corresponds to approximately 48 g/L in dilute aqueous solution; the second is a separate lower value that appears without consistent temperature or methodological specification. The discrepancy carries regulatory and engineering consequences because a factor of nearly 1.8 changes groundwater loading estimates, air-stripper sizing, and phase-separation thresholds in fuel storage.

The mass-fraction expression 4.8 g per 100 g derives from saturation solubility measurements expressed as grams of MTBE per 100 g of water or solution. Because the aqueous solution density remains close to that of pure water, the accepted conversion is 4.8 g per 100 g = 48 g/kg, which is equivalent to 48 g/L when solution density is assumed to be 1.0 kg/L. At 25 °C, the density of water is 0.997 kg/L; therefore, 48 g/kg corresponds to 47.9 g/L. The lower value of 26 g/L cannot be obtained by decimal misplacement, density correction, or mass-fraction versus mass-per-volume distinction. In saturation mass terms, 26 g/L is approximately 2.6 wt%, which is only about 54% of the 4.8 wt% value.

When 4.8 g per 100 g Is Recalculated as Mass Per Litre

The conversion path is unambiguous. A saturated water solution containing 4.8 g MTBE per 100 g solution contains 48 g MTBE per 1000 g solution. If the solution density is 0.997 kg/L at 25 °C, the volume occupied by 1000 g solution is 1003 mL; the resulting concentration is 47.9 g/L. If the original determination was based on 4.8 g MTBE per 100 mL solvent, the concentration is exactly 48 g/L. Either form rounds to 48 g/L, not 26 g/L. This indicates that the two circulating values are not equivalent representations of one measurement but reflect either different experimental conditions or a citation error propagated among safety data sheets and compilation databases.

The saturated concentration of a partially miscible organic compound in water is a thermodynamic endpoint governed by temperature, pressure, solute purity, and ionic strength of the aqueous phase. MTBE is fully miscible with many organic solvents but only partially miscible with water. In binary MTBE-water equilibrium, two liquid phases can coexist up to the mutual solubility boundary. The organic-rich phase may contain dissolved water, and the aqueous phase contains dissolved MTBE. If the aqueous solubility is expressed on a solvent-only basis, a measurement of 4.8 g MTBE per 100 g water yields a saturated mass fraction of 4.58 g per 100 g solution, which corresponds to 45.8 g/L after density correction. If expressed on a total-solution basis, 4.8 g per 100 g solution yields 48 g/L. The gap between 45.8 g/L and 48 g/L is only 2.2 g/L, far smaller than the gap to 26 g/L. The lower value therefore cannot be explained by the solvent-versus-solution mass basis alone.

Does Shake-Flask Methodology Explain the 26 g/L Value?

Standard water solubility determinations for liquid organic compounds are carried out according to OECD Test No. 105 or equivalent shake-flask procedures. A known excess of MTBE is added to reagent water in a temperature-controlled vessel, agitated for a minimum equilibration period, and then allowed to separate before analysis of the aqueous phase. The most critical operator-dependent variable is the separation step. MTBE has a boiling point of 55.2 °C and a high vapour pressure relative to heavier ethers. Filtration, syringe sampling, or decanting can volatilise dissolved MTBE from the aqueous phase; even a 2 °C temperature rise during sampling can reduce the apparent concentration. A low recovery would artificially depress the apparent solubility. In contrast, direct closed-vessel UV absorption or headspace gas chromatography with internal standard calibration may report higher values because the aqueous phase is not exposed to atmosphere. A low-biased result such as 26 g/L could arise from an open transfer step if the compound was quantified by purge-and-trap without a closed-loop sampling train. However, published data for the specific configuration that generated 26 g/L is limited; the number often appears in secondary compilations without a cited temperature or method.

Analytical methods for MTBE in water rely on volatility. EPA Method 5030C couples purge-and-trap concentration with gas chromatography and mass spectrometry under EPA Method 8260D. The method quantifies MTBE at part-per-trillion to part-per-billion concentrations after spiking with isotopically labelled internal standards such as d₃-methyl tert-butyl ether. For solubility-scale measurements, direct split injection of water samples is not practical for capillary columns without polar stationary phases; backflush techniques and thick-film cyanopropyl columns are used to retain the ether. In comparison, the determination of MTBE in gasoline is standardised in ASTM D4815-22, in which two-dimensional gas chromatography separates oxygenates from hydrocarbon interferences. Those fuel-range methods do not address saturated aqueous solutions because the concentration range differs by several orders of magnitude and the sample matrix is water rather than hydrocarbon. Method-to-method variability between direct injection, static headspace, and purge-and-trap can exceed 15% for volatile ethers in unfiltered water, but it does not produce a 48 g/L versus 26 g/L divergence unless sample aeration or temperature control is lost.

Groundwater Partitioning and Fuel Storage Phase Equilibrium

In a phase-separated gasoline release, the concentration of MTBE in water is not governed solely by the solubility of pure MTBE but by its mole fraction in the gasoline and Raoult’s-law partitioning. Nevertheless, the aqueous solubility of the pure ether is used as a bounding condition for source-term models and saturated groundwater concentration estimates. If 4.8 g per 100 g is used, the upper aqueous concentration is about 48 g/L; if 26 g/L is used, the bound is 26 g/L, a reduction of 46%. For a gasoline containing 10% MTBE by volume, equilibrium partitioning can still produce water concentrations far below the pure-phase solubility limit, but the selection of solubility value controls the scaling in fugacity calculations and air-stripper mass balances.

Air stripping and granular activated carbon are standard ex situ groundwater treatment technologies. Air stripper design uses the Henry’s law constant and the influent aqueous concentration. With a lower apparent solubility of 26 g/L, the mass transfer driving force is underestimated in the high-concentration region, and pack height or air-to-water ratio may be undersized. Packed-tower air strippers treating MTBE-impacted groundwater typically require an air-to-water ratio that is 2 to 3 times higher than for benzene under the same removal target because MTBE’s dimensionless Henry’s law constant is comparatively small. If the solubility value is incorrect by a factor of 1.8, the equilibrium concentration at the air-water interface in the tower sump is incorrectly assigned, and the expected effluent breakthrough curve shifts. This is a process-level consequence of the duplicated solubility figures.

Separating MTBE from Aqueous Process Streams

During liquid-phase synthesis of MTBE from isobutylene and methanol over acidic ion-exchange resin, the crude product stream is contacted with water to extract methanol and trace unreacted reagents. The water solubility of MTBE is sufficiently high that aqueous extraction streams cannot be discharged without ether recovery. In a countercurrent water wash column using structured packing, the aqueous phase extracts methanol and dissolves MTBE at a concentration that approaches the solubility limit only when the water-to-hydrocarbon ratio is high. The choice between 48 g/L and 26 g/L as the design saturation concentration alters the liquid-liquid extraction mass balance and the reflux requirement in the downstream distillation column. If the lower figure is accepted, the organic loading in the water draw is underestimated and the recovery column may be specified with insufficient stripping steam. Published data for this specific configuration is limited because plant-specific water compositions and column temperatures vary; however, industrial practice in MTBE production units treats the aqueous phase as containing significant dissolved ether rather than a trace-level contaminant.

Temperature exerts a measurable but insufficient shift to bridge the gap. Solubility data for ethers in water generally show decreasing mass fraction with increasing temperature when the system approaches the lower critical solution temperature. For MTBE, published values at ambient conditions cluster near 4.5 g per 100 g to 4.8 g per 100 g, with lower values sometimes reported at higher temperatures. A shift to 26 g/L would require a temperature excursion well beyond normal groundwater or standard laboratory conditions, or a saline aqueous phase. No widely cited experimental data set at 20–40 °C supports 26 g/L as the saturation limit for pure MTBE in reagent water. The value is more plausibly an analytical loss or a transcription from a different compound such as ethyl tert-butyl ether or tert-amyl methyl ether, but the original source is not consistently identified.

Safety data sheet preparation requires the basis to be stated clearly as mass fraction of solution, mass fraction of solvent, or mass of solute per volume of solution. Regulatory exposure models under REACH use water solubility to derive predicted environmental concentrations. If 26 g/L is entered into a Level III fugacity model, the distribution shifts away from the water compartment and toward air. The difference changes the predicted no-effect concentration dilution, but not the classification threshold because MTBE is not classified as a chronic aquatic toxicant at the concentrations likely to be modelled. However, the groundwater source term in a spill scenario is directly proportional to the solubility boundary. The safer engineering convention is to use the higher value for conservative source-term estimates and to state the test temperature and method alongside any cited figure.