In solvent-replacement programmes for heat-sensitive botanical extract purification, the substitution of dichloromethane by methyl tert-butyl ether introduces a set of correlated physical property changes that alter evaporation rate, liquid-liquid extraction behaviour, flammability classification, peroxide management, and residual solvent compliance. Dichloromethane has a normal boiling point of 39.6 °C at 101.3 kPa, a vapour pressure of approximately 47.0 kPa at 20 °C, a density of 1.326 g/cm³, and a dielectric constant of 8.93 at 25 °C. Methyl tert-butyl ether has a normal boiling point of 55.2 °C, a vapour pressure of approximately 26.7 kPa at 20 °C, a density of 0.740 g/cm³, and a dielectric constant of 2.6. The increase in normal boiling point of 15.6 °C requires deeper vacuum or longer thermal exposure to achieve comparable solvent removal, while the lower dielectric constant reduces solvation of moderately polar polyphenol glycosides and organic acids but can improve selectivity toward neutral terpenoids, waxes, and lipid-soluble oleoresins. The reduction in solvent density from 1.326 g/cm³ to 0.740 g/cm³ shifts phase orientation in liquid-liquid extraction; MTBE forms the upper organic layer against water, whereas dichloromethane forms the lower layer. The shift changes decanter draw-off configuration, interface-level sensing, and dense-phase pumping requirements in existing plant installations. Methyl tert-butyl ether is flammable with a closed-cup flash point of approximately -28 °C, whereas dichloromethane is non-flammable under conventional process conditions. Residual solvent monographs differ materially: ICH Q3C(R8) assigns dichloromethane to Class 2 with a permitted daily exposure of 6 mg/day, whereas methyl tert-butyl ether is listed as Class 3 with a permitted daily exposure of 50 mg/day. These differences make MTBE favourable from a pharmacopoeial residual solvent perspective for certain botanical extracts, provided that extraction selectivity, peroxide stability, and fire safety controls are re-engineered rather than assumed.

Production decanters processing crude MTBE-containing extract show inverted phase behaviour relative to dichloromethane-based campaigns. With MTBE as the upper layer, the main interface moves to the lower portion of the decanter, and level sensors calibrated for dense-phase dichloromethane service must be repositioned or replaced. The higher water solubility of MTBE, 4.8 g/100 g at 20 °C, increases the aqueous carryover into the extract and raises the drying duty on subsequent anhydrous sodium sulfate or molecular sieve beds. Demisting equipment such as coalescers and electrostatic coalescers may require reconfiguration because the lower interfacial tension and phase-density difference against water, 0.260 g/cm³ for MTBE versus 0.326 g/cm³ for dichloromethane, reduce phase-separation driving force. Batch-to-batch variance has been observed in production decanters when feed moisture content varies by more than 2 %, changing the rag-layer thickness and requiring operator intervention. Pre-drying of the botanical feedstock to below 5 % residual moisture is recommended when MTBE-based extraction is used at production scale; otherwise, the additional water transferred into the extract can exceed the capacity of the drying train and prolong cycle time.

Thermal Degradation Pathways in Cannabinoid- and Terpene-Rich Oleoresins During Solvent Evaporation

Solvent removal from heat-sensitive botanical extracts is constrained by the thermal lability of monoterpenes, sesquiterpenes, and acidic cannabinoid constituents. Acidic cannabinoids undergo decarboxylation at appreciable rates above 105 °C, but terpene oxidation and acid-catalysed rearrangements can occur at lower temperatures when dissolved oxygen and trace acids are present. The higher normal boiling point of MTBE, 55.2 °C versus 39.6 °C for dichloromethane, shifts solvent-removal operations to higher jacket temperatures unless vacuum is increased. For a condenser temperature of -5 °C, the vacuum required to maintain the evaporator sump below 45 °C is deeper for MTBE than for dichloromethane; plant vacuum pumps, condenser surface area, and vapour ducts may require re-rating. Published data for specific botanical matrices of this type are limited, but a process control window of ±5 °C around a target sump temperature is commonly used in commercial wiped-film evaporation to avoid thermal bleaching of chlorophyll and degradation of heat-sensitive esters. The substitution therefore demands evaluation of evaporator vacuum capability, heating medium turndown, and residence time distribution rather than a simple replacement of the solvent drum.

What Limits Liquid Hourly Space Velocity in Packed-Bed Botanical Extraction Columns?

Liquid hourly space velocity in packed-bed botanical extraction is limited by bed compaction, solute diffusion, pressure-drop constraints, and the phase-density difference between the extraction solvent and the aqueous or ethanol-water phase. MTBE has a viscosity of approximately 0.35 mPa·s at 20 °C, slightly lower than the 0.44 mPa·s typical of dichloromethane, and a density of 0.740 g/cm³ that reduces static head in the column. The lower viscosity and density reduce pressure drop per unit bed length, but the lower density also reduces the gravitational term in liquid-holdup correlations, which can lead to premature flooding in countercurrent operation if the phase ratio is not adjusted. In a 316L column packed with corrugated structured packing having a specific surface area of 250 m²/m³, the pressure drop is dominated by vapour-phase entrainment rather than liquid-phase density; published hydraulic data for this specific configuration are limited. The Ergun equation is commonly applied for packed beds, and the lower liquid density of MTBE reduces the gravity-driving force for phase disengagement. Liquid hourly space velocity must therefore be reduced when the extractor is operated at temperatures above 30 °C because elevated vapour pressure can generate gas pockets in the bed. For botanical feedstocks with high fines content, the wetting behaviour of the cellulosic matrix also changes; MTBE does not swell lignocellulose to the same extent as dichloromethane, and this may reduce bed permeability if the feedstock is not pre-conditioned with a compatible co-solvent.

Where dichloromethane has been used to extract free-base alkaloids from basified aqueous phases, the change to MTBE lowers the distribution coefficient for moderately polar alkaloids because the solvent polarity is lower and hydrogen-bond basicity differs. Morphinan and isoquinoline alkaloids with phenolic hydroxyl groups may require the addition of 5 vol% to 15 vol% ethyl acetate or 2-propanol to maintain recovery above 90 %; published data for specific alkaloids are limited. The phase inversion also requires that the lower aqueous raffinate be drawn from the bottom and the upper MTBE product from the top. In continuous centrifugal extractors, the weir heights must be adjusted to account for the lower density and lower interfacial tension of MTBE, which can increase emulsion tendency when saponins are present. Centrifugal extractor manufacturers specify maximum density-difference and flow-ratio envelopes; operation outside these envelopes causes the interface to migrate and alkylaloid losses to the raffinate. A separate alkaloid precipitation step may be required because MTBE has a weaker dipole and does not promote the same crystal habit as dichloromethane in hydrochloride salt formation.

Analytical-scale purification of botanical extracts with MTBE in liquid-liquid extraction and solid-phase extraction offers a lower-chlorinated background for gas chromatography–mass spectrometry because MTBE is less likely to form hydrochloric acid or phosgene under thermal desorption conditions. The extract must be dried with anhydrous sodium sulfate or magnesium sulfate before concentration; sodium sulfate is preferred because it has less surface acidity than magnesium sulfate and minimises acid-catalysed rearrangements of labile terpenes. Concentration is performed in a centrifugal vacuum evaporator equipped with PTFE seals; the vacuum is ramped to avoid bumping due to the lower surface tension of MTBE. Published data for specific configurations are limited, but laboratories using autosamplers have noted that residual MTBE can swell certain perfluoroelastomer septa if the injection solvent is not evaporated completely. This compatibility boundary is addressed by using butyl/PTFE laminated septa and replacing them at the interval specified by the autosampler manufacturer. Headspace gas chromatographic determination of residual solvents in the finished extract is performed under USP <467> or the corresponding Ph Eur method, with system suitability mixtures containing MTBE at the Class 3 limit and dichloromethane at the Class 2 limit.

Peroxide Formation and Stabiliser Depletion in Recovered MTBE

Peroxide accumulation in MTBE occurs via free-radical autoxidation at the tertiary carbon adjacent to the ether oxygen. Commercial MTBE is inhibited with 2,6-di-tert-butyl-4-methylphenol at concentrations typically between 5 mg/kg and 100 mg/kg; during extraction and distillation, the stabiliser can partition into the extract and become depleted. Peroxide levels in recovered solvent should be tested by ASTM E298 or an equivalent iodometric titration before reuse; peroxide concentrations above 50 mg/kg are generally considered unacceptable for pharmaceutical botanical extraction because of oxidative degradation risk. The higher boiling point of MTBE relative to dichloromethane increases distillation residence time in the reboiler, which can accelerate peroxide formation if oxygen ingress occurs. Recovery systems for MTBE should therefore be blanketed with nitrogen and fitted with low-oxygen shutdown interlocks at 1.0 vol% oxygen. Dichloromethane does not form peroxides, so the shift to MTBE requires new operational controls and batch-record fields for stabiliser concentration and peroxide index. The stabiliser content in recovered solvent may be supplemented with fresh inhibited solvent to maintain the peroxide index below the rejection threshold; the addition rate is determined by the mass balance across the distillation unit and the antioxidant partition coefficient in the specific extract matrix.

When the Boiling Point Differential Alters Wiped-Film Evaporator Turndown

A wiped-film evaporator with a heated surface area of 0.25 m² and rotor speed of 300 rpm processes dichloromethane-containing extract at a feed rate of 20 kg/h under 20 kPa; switching to MTBE while maintaining the sump below 45 °C may require reducing the feed rate by 15 % to 30 % because the boiling point at 20 kPa for MTBE is higher than for dichloromethane. The exact turndown depends on the vacuum system, condenser temperature, and the heat-transfer coefficient of the liquid film. Published data for this specific configuration are limited, but manufacturers of agitated thin-film evaporators provide sizing correlations based on solvent vapour load and heat-transfer coefficient. The vapour loading per unit mass of solvent is influenced by the lower density of MTBE, which increases volumetric vapour flow per unit mass and changes entrainment behaviour in the vapour disengagement zone. Rotary vacuum pumps handling MTBE require flame-proof motors when located in classified areas; dichloromethane degradation products may also corrode aluminium pump parts if moisture and acidic metabolites are present. The replacement therefore requires a joint review of evaporator heat transfer, vacuum pump classification, and condenser fouling behaviour rather than an isolated boiling-point comparison.

In pilot-scale botanical extraction campaigns, the lower density of MTBE changes the liquid head and pump net positive suction head margin. Centrifugal pumps with magnetic drives sized for dichloromethane may not generate adequate suction with MTBE due to its higher vapour pressure and lower net positive suction head margin. A gear pump or diaphragm pump with PTFE wetted parts is generally specified. The packed-column pressure drop is lower for MTBE, but the phase inversion in the decanter requires new interface-level control programming and recalibration of capacitance probes. The lower density also alters the hydraulic load on the falling-film evaporator distributor; a distributor designed for dense dichloromethane may produce uneven wetting with MTBE, reducing heat-transfer efficiency and increasing localised thermal degradation. Published data for specific distributor configurations are limited, but the operational boundary is typically evaluated by measuring the heat-transfer coefficient and the colour change of the concentrated extract during solvent-removal performance qualification.

Residual Solvent Compliance Matrix for Botanical Extracts

The residual solvent specification for a finished botanical extract intended for pharmaceutical use is commonly set at 600 ppm for dichloromethane and 5000 ppm for methyl tert-butyl ether under ICH Q3C(R8), but the actual limit may be tightened by customer monograph or inhalable dose calculations. USP <467> Option 1 uses a headspace gas chromatographic method with a water-insoluble sample preparation where dichloromethane and MTBE are detected against system suitability standards. For botanical extracts that are not pharmaceutical grade, these limits may not apply, and regulatory status under food contact or cosmetic frameworks must be assessed separately. The matrix below is used as a gate check for solvent substitution in pharmaceutical-grade botanical extraction campaigns.

An operational incompatibility exists with amine-based additives in MTBE-containing extracts because trace amines can promote base-catalysed decomposition of adventitious peroxide species and generate odorous low-molecular-weight carbonyls. This incompatibility is not present with dichloromethane, which is comparatively inert toward amines under cold extraction conditions. Equipment elastomers must also be re-specified: MTBE swells nitrile and ethylene-propylene-diene monomer gaskets to a greater extent than dichloromethane, whereas polytetrafluoroethylene and perfluoroelastomer wetted parts remain acceptable. Published data for specific botanical extract purification campaigns are limited, but the documented physical property differences and the compliance boundaries described above establish the minimum re-qualification scope for a dichloromethane-to-MTBE solvent replacement programme.