Platform · Demonstration 2
No extraction route wins on every objective
Take one documented separation: carnosic acid and rosmarinic acid from dried rosemary leaves. In the published studies, pressurized ethanol gave the highest extract yield and supercritical CO₂ the highest carnosic acid content. A deep eutectic solvent needs the least heat unless it must be recovered. This page computes energy and solvent loss for five routes from reference property data, and quotes yield and purity only from the studies that measured them.
Proposed capabilityDemonstration only. EUTECNIQ has run no extractions and owns no equipment. Energy and solvent figures are educational model outputs from CoolProp 8.0.0 property equations and the inputs stated on this page; yields and selectivity are quoted from the cited studies.
Five routes, per kg of dried rosemary leaf Heat plus pump work (MJ) and solvent not recovered (kg), at this page's default inputs
Values in the figure
| Route at default inputs | Heating, MJ | Evaporation, MJ | Work, MJ | Solvent lost, kg |
|---|---|---|---|---|
| Hot water, 90 °C | 3.6 | 24.8 | 0.00 | 1.2 |
| Ethanol/water, 70 °C | 1.8 | 12.7 | 0.00 | 1.2 |
| Pressurized ethanol/water, 150 °C | 5.2 | 9.7 | 0.21 | 1.2 |
| Supercritical CO₂ + 5% ethanol | 4.6 | 0.9 | 0.34 | 1.0 |
| DES, extract left in solvent | 1.8 | 0 | 0.00 | 20.0 |
| DES, 90% recovered with water | 1.8 | 46.6 | 0.00 | 2.0 |
In brief · Sensitivity
Solvent recovery reorders the routes; scale-up evidence and solvent status then narrow them
At this page's default inputs a deep eutectic solvent needs the least heat, because none of it is recovered. Recover 90% of it with water as the antisolvent and it needs the most: 48.5 MJ per kg of leaf, against 28.4 MJ for hot water. In the liquid routes every kilogram of solvent recovered is a kilogram evaporated, so higher recovery trades solvent loss for evaporator heat.
| Route | Model at defaults, per kg leaf heat and work; solvent lost | What solvent recovery changes | Scale-up evidence; capital intensity judgment | Regulatory status and hazard of the solvent |
|---|---|---|---|---|
| Hot water | 28.4 MJ; 1.2 kg at 90% recovery | Evaporation is 87% of the heat and rises in proportion to the fraction recovered | Simplest equipment, but fails the carnosic acid target (Hirondart); lower | Not flammable |
| Ethanol/water | 14.5 MJ; 1.2 kg at 90% recovery | Evaporation is 87% of the heat and rises in proportion to the fraction recovered | Conventional solid–liquid extraction; moderate | Not in 21 CFR Part 173 Subpart C; § 184.1293 affirms GRAS for a named use (antimicrobial on pizza crusts) and was issued before a general evaluation of ethanol. Flammable |
| Pressurized liquid | 15.1 MJ; 1.2 kg at 90% recovery | Evaporation is 65% of the heat, because part of the solvent flashes on letdown; it rises in proportion to the fraction recovered | Sánchez-Camargo pooled laboratory runs because pilot-scale PLE equipment was unavailable to them; higher | As ethanol/water, plus hot liquid under pressure; browning at 150 °C and above |
| Supercritical CO₂ | 5.8 MJ, plus 4.8 MJ of chilled condensing; 1.0 kg CO₂ at 95% recycled | Cycle duty scales with the CO₂-to-leaf ratio (19.6 kg/kg); 5% ethanol cosolvent brings back 0.94 MJ of evaporation | More than 150 industrial SFE plants by 2014 (cited by Sánchez-Camargo); highest | Carbon dioxide listed in 21 CFR 184.1240; high-pressure gas |
| Deep eutectic solvent | 1.8 MJ and 20 kg if kept in the extract; 48.5 MJ and 2.0 kg if 90% recovered | Recovery needs a second separation; with 1 kg of water per kg recovered, 18 kg of water must be evaporated | Laboratory studies only among the sources here; lower vessels, recovery train unknown | No DES in 21 CFR 173 Subpart C; component status depends on intended use |
Educational model Published evidence Energy and solvent figures are this page's model outputs at its default inputs, not measurements. The model is per kilogram of leaf: it has no economy of scale and omits heat integration, multiple-effect evaporation and vapor recompression, any of which would reduce the steam needed. Capital intensity is EUTECNIQ's qualitative reading. To see a route move, change the recovery in the model (section 03); detail in sections 07 and 08.
01 · The decision
One documented separation puts five routes on the same leaf
Rosemary leaf carries two kinds of target molecule with opposite polarity. Carnosic acid and carnosol are phenolic diterpenes, fat-soluble antioxidant molecules; rosmarinic acid is a water-soluble ester of caffeic acid. A solvent that dissolves one tends to leave the other behind, which makes rosemary a demanding test. It is also a commercial separation: Vieira et al. note that the European Commission authorized extracts of rosemary, standardized on carnosic acid and carnosol, as food additive E 392 in 2010.
Hot water
Water at up to 100 °C and atmospheric pressure. The cheapest solvent to buy, with the highest heat of vaporization: 2,257 kJ/kg at its boiling point (CoolProp).
Ethanol/water
80% ethanol by weight, below its boiling point. It dissolves both polarities. In Hirondart's pressurized screen (100 °C, 80 bar), 80% ethanol extracted the most rosmarinic and carnosic acid together. Ethanol is flammable.
Pressurized liquid
Pressurized liquid extraction (PLE) keeps ethanol/water or water liquid above its boiling point: 150 °C at 100 bar and 183 °C at 130 bar in the cited studies. With water alone it is called pressurized hot water extraction.
Supercritical CO₂
CO₂ above its critical point, 30.98 °C and 73.8 bar (CoolProp), where it is dense enough to dissolve lipophilic compounds. The cited pilot runs used 150 to 300 bar at 40 °C, with 5 to 7% ethanol as cosolvent for the more polar compounds.
Deep eutectic solvent
A deep eutectic solvent (DES) is a mixture of hydrogen-bond donors and acceptors that stays liquid far below its components' melting points, such as menthol with lauric acid or lactic acid with glucose. Their vapor pressure is very low. For hydrophobic, menthol-based mixtures it has been measured and remains far below that of common organic solvents such as toluene (Dietz et al., ACS Sustain. Chem. Eng. 2019;7:4047–4057).
02 · Published yield and selectivity
Pressurized ethanol recovered more carnosic acid; supercritical CO₂ made the purer extract
Sánchez-Camargo et al. ran both routes on one lot of leaves. Pressurized 76% ethanol at 150 °C turned 38.5% of the leaf into extract at 105 mg carnosic acid per gram. Two-step supercritical CO₂ gave 4.7% at 443 mg/g. Hirondart et al. state the same trade for their own pressurized extraction: higher yield, lower purity of the active compounds.
Extract yield and carnosic acid content, as published Percent of dried leaf mass; milligrams of carnosic acid per gram of extract
What else the studies found
- Water extracts only the polar half. In Hirondart's solvent screen (100 °C, 80 bar), water and 20% ethanol extracted rosmarinic acid but no carnosic acid; carnosic acid appeared from 40% ethanol.
- High temperatures brown the extract. Hirondart observed Maillard reaction products, a brown color and burnt smell, at 150 °C and above, and notes that such temperatures are usually not recommended for products; their purpose was analytical.
- Long heating converts carnosic acid to carnosol. Their 8-hour Soxhlet reference held more carnosol than pressurized extraction (4.4 against 1.9 mg/g leaf), which the authors read as carnosic acid degrading to carnosol.
- A pair of DES phases splits the polarities. Vieira et al. used two immiscible natural DES at once. The polar phase took the rosmarinic acid; the nonpolar phase took 17.54 mg carnosic acid per gram of leaf, 91% of what methanol Soxhlet extraction recovered from the same leaves (computed from their values).
Every published value used on this page
| Route and study | Conditions | Reported result | Same-study comparison |
|---|---|---|---|
| Water, pressurized (Hirondart et al. 2020) | PLE screen, 0% and 20% ethanol in water, 100 °C, 80 bar, 3 cycles of 5 min | Rosmarinic acid extracted; carnosic acid not extracted | 80% ethanol: 10.13 mg rosmarinic acid and 20.6 mg carnosic acid per g leaf |
| Soxhlet reference, solvent not stated (Hirondart et al. 2020) | 78 °C, 8 h, 1 g leaf per 12 mL; the paper does not name the Soxhlet solvent and lists only ethanol and water as extraction solvents | Yield 27.91%; per g leaf: rosmarinic acid 9.86 mg, carnosic acid 17.68 mg, carnosol 4.39 mg | Carnosic acid mean 84% of the pressurized mean (computed) |
| Pressurized 80% ethanol (Hirondart et al. 2020) | 183 °C, 130 bar, 3 × 1 min; about 50 mL solvent for 3.1 g leaf | Yield 46.20%; per g leaf: rosmarinic acid 10.01 mg, carnosic acid 21.13 mg, carnosol 1.86 mg | Means of 8 runs each; the authors report no significant difference from Soxhlet for rosmarinic acid, carnosic acid or carnosol |
| Pressurized 76% ethanol (Sánchez-Camargo et al. 2016) | 150 °C, 100 bar, 20 min, laboratory scale | Yield 38.46%; per g extract: carnosic acid 105.0 mg, carnosol 10.7 mg | 40.4 mg carnosic acid per g leaf (computed: yield × content) |
| Supercritical CO₂, two steps (Sánchez-Camargo et al. 2016) | Pilot, 2 L vessel, 60 g CO₂/min, 40 °C: 300 bar for 60 min, then 150 bar with 7% ethanol for 120 min | Yield 4.68%; per g extract: carnosic acid 443.3 mg, carnosol 88.9 mg | 20.7 mg carnosic acid per g leaf, 51% of the pressurized-ethanol value; 4.2 times its purity (computed). Single step at 150 bar with 7% ethanol: yield 6.74% |
| Supercritical CO₂ with 5% ethanol (Vicente et al. 2013) | Pilot, 2 L vessel, 0.55 kg leaf, 60 g CO₂/min for 180 min, 150 bar, 40 °C; separators at 40 °C, recirculation at 50 bar | Yield 7.26%; carnosic acid 25.66% of the extract | 18.6 mg carnosic acid per g leaf; 19.6 kg CO₂ per kg leaf (both computed) |
| Biphasic natural DES (Vieira et al. 2022) | Lactic acid:glucose (5:1) with menthol:lauric acid (2:1), ultrasound, 60 °C, 60 min, 1 g leaf per 20 g solvent | Polar phase: rosmarinic acid 1.00 mg, caffeic acid 0.26 mg per g leaf. Nonpolar phase: carnosol 2.30 mg, carnosic acid 17.54 mg per g leaf | Methanol Soxhlet on the same leaves: rosmarinic acid 1.36, carnosol 7.28, carnosic acid 19.28 mg/g |
Why numbers from different studies cannot be ranked. The two Soxhlet references disagree by a factor of 7.3 on rosmarinic acid: 9.86 mg/g in Hirondart's leaves and 1.36 mg/g in Vieira's (computed ratio). Plant lot, drying and analysis moved the result more than most route changes do. EUTECNIQ's first proposed experiment therefore fixes one lot and one analytical method for every route.
03 · Energy and solvent model
Evaporating the solvent takes 65 to 87% of the heat in the liquid routes
At 12 kg of solvent per kg of leaf and 90% recovery, evaporating the recovered solvent takes 87% of the heat for hot water and for ethanol/water, and 65% for pressurized ethanol/water, whose hot extract flashes part of its solvent when the pressure is let down. Change the solvent ratio, the recovery and the temperature to see which term moves.
Mass and energy flows per kg of dried leaf Band width proportional to mass (kg), arrow width to energy (MJ), widths are capped for legibility; labels give the computed values
Liquid routes (shown: ethanol/water at defaults)
Supercritical CO₂ at defaults
Deep eutectic solvent at defaults
All five routes at their current inputs
Per kg of dried leaf. Each route keeps its own inputs when you switch routes above.
| Route | Inputs | Heat, MJ | Work, MJ | Cooling, MJ | Solvent lost, kg |
|---|---|---|---|---|---|
| Hot water | 12 kg/kg, 90% recovered, 90 °C | 28.4 | 0.00 | 24.4 | 1.2 |
| Ethanol/water | 12 kg/kg, 90% recovered, 70 °C | 14.5 | 0.00 | 12.2 | 1.2 |
| Pressurized liquid | 12 kg/kg, 90% recovered, 150 °C, 100 bar | 14.9 | 0.21 | 12.2 | 1.2 |
| Supercritical CO₂ | 19.6 kg/kg, 95% recovered, 40 °C, 150 bar, ethanol 5% of CO₂ | 5.5 | 0.34 | 5.6 | 1.0 |
| Deep eutectic solvent | 20 kg/kg, 0% recovered, 60 °C, water 0%, antisolvent 1 kg/kg | 1.8 | 0.00 | 0 | 20.0 |
Method, assumptions and limits
What it calculates. Heat, pump work and cooling per kilogram of dried leaf to bring the solvent to the extraction conditions and to recover it, and the solvent that is not recovered. It does not calculate yield, purity or cost.
Liquid routes (1 to 3). Heating: Qheat = S Σi wi [hi(T, P) − hi(T0, P)] + cp,leaf(T − T0). Recovery: Qevap = S r Σi wi [hivap(1 atm) − hi(T, P)], the enthalpy from the extract's state to saturated vapor at 1 atm; above the boiling point this credits the flash on letdown. Cooling: S r ΣwiΔhvap,i(1 atm). Pump work for route 3: isentropic compression of the liquid at T0 to P, divided by an efficiency of 0.70.
Supercritical CO₂ (route 4). Pump cycle: 4 liquid at 50 bar, 5 K below saturation (9.3 °C); 5 after the pump (isentropic, efficiency 0.70); 1 heated to the extraction pressure and temperature; 2 after an isenthalpic valve to 50 bar; 3 heated in the separator to 40 °C, the separator and recirculation conditions reported by Vicente et al. Per kg CO₂: work h5 − h4, preheat h1 − h5, separator heat h3 − h1, condenser h3 − h4. If the pump outlet is hotter than the extraction temperature (high pressure, low temperature), the preheater becomes a cooler and that heat is counted as cooling duty. Cosolvent ethanol, given as a fraction of the CO₂ mass, is heated and pumped with the CO₂ and then fully evaporated from the extract at 1 atm, from 40 °C.
Deep eutectic solvent (route 5). Heat capacity of choline chloride:glycerol 1:2 and its water mixtures, measured by Leron and Li (2012) at 30 to 80 °C (NIST TRC ThermoML), converted to a mass basis and fitted linearly in temperature for each composition (largest residual below 0.01 kJ/kg·K), interpolated in water content. This DES stands in for the menthol- and lactic acid-based DES of the selectivity data, whose heat capacities were not found. Recovery is modeled as antisolvent precipitation with water at the stated kilograms per kilogram of DES recovered, the water then evaporated from 20 °C at 1 atm. No rosemary study reports that ratio; it is an input.
Variables and units. S kg solvent per kg leaf; r fraction recovered; T °C; P bar; wi mass fraction; h kJ/kg; results in MJ per kg leaf. T0 = 20 °C. Leaf heat capacity 1.5 kJ/kg·K is assumed; it adds at most 0.3 MJ.
Assumptions. Ethanol and water mix ideally (heat of mixing ignored), and each component evaporates at its own normal boiling point. There is no heat integration, multiple-effect evaporation or vapor recompression, any of which would reduce the steam needed, and no heat loss. Solvent not recovered leaves the process. CO₂ losses are a fraction of circulated CO₂ (in practice most of the loss is the vessel inventory vented at each depressurization). Ethanol cosolvent is fully recovered.
Valid range. Inputs are limited to the slider ranges: atmospheric liquids below their boiling points, pressurized liquids at 50 to 150 bar and 100 to 200 °C, CO₂ at 100 to 400 bar and 35 to 80 °C, DES at 30 to 80 °C (the measured range).
What it does not establish. Yield, purity, extraction kinetics, product quality, equipment size, operating cost or environmental impact. Heat and electricity are reported in the same unit but differ in cost and primary energy.
Properties. CoolProp 8.0.0 (Bell et al. 2014): water, ethanol and carbon dioxide reference equations of state; CO₂ enthalpy on the IIR reference, 200 kJ/kg for saturated liquid at 0 °C. Tabulated for the page every 2 °C and at six pressures (liquids) or every 10 bar and 2 °C (CO₂) and interpolated linearly.
Verification. The same equations were evaluated independently in CoolProp for 19 input sets, and the page's model agrees with every result within 1%.
04 · Supercritical CO₂
The separator state and the CO₂-to-leaf ratio fix the heat duty of a CO₂ pump cycle
Per kilogram of CO₂ in a closed cycle, pump work and the two heaters add up to what the condenser removes. With separation at 50 bar and 40 °C that is 244.5 kJ/kg, whatever the extraction pressure.
- 1 Extractor CO₂ at the extraction pressure and temperature, 780 kg/m³ at 150 bar and 40 °C.
- 2 After the valve At 50 bar the stream is a cold mixture, 26.5% vapor at 14.3 °C, and the extract drops out.
- 3 Separator Heated to 40 °C: 181.6 kJ/kg at 150 bar extraction.
- 4 Condenser Liquefied 5 K below its 14.3 °C saturation temperature, so the coolant must be colder still.
- 5 Pump Back to the extraction pressure: 16.0 kJ/kg at 150 bar, 39.0 kJ/kg at 300 bar.
Raising the pressure from 150 to 300 bar moves heat from the heaters to the pump without changing their sum. The duty per kilogram of leaf therefore scales with the CO₂-to-leaf ratio, 19.6 kg/kg in Vicente's pilot runs. Ethanol cosolvent brings evaporation back: 5% of that CO₂ is 0.98 kg ethanol per kg leaf, 0.94 MJ to evaporate. Vicente et al. describe cosolvent evaporation as an expensive task.
The modeled CO₂ cycle on a pressure–enthalpy diagram Pressure (bar, log scale) against enthalpy (kJ/kg); updates with the CO₂ inputs above
05 · Deep eutectic solvents
A deep eutectic solvent stays in the extract unless a second separation removes it
A DES cannot be boiled off like ethanol. That spares the evaporator and creates three engineering problems instead: viscosity, recovery and residue.
Thick solvents extract slowly or not at all
Vieira et al. measured 14 to 1,654 mPa·s at 40 °C for ten natural DES, against 0.65 mPa·s for water (CoolProp): 21 to 2,533 times as viscous (computed). The most viscous, citric acid:glycerol, extracted none of the four target compounds, which the authors attribute to compromised mass transfer. Viscosity falls with temperature and added water, and water also shifts polarity.
Taking the solvent back costs a second separation
Rosa et al. (2026) identify low volatility as the reason extract purification and solvent recovery remain difficult, and list resin adsorption, antisolvent precipitation and direct formulation as the approaches explored. In this page's model, recovering 90% of the DES with 1 kg of water per kg leaves 18 kg of water to evaporate: 48.5 MJ per kg leaf, more than three times the ethanol/water route.
Left in place, the solvent is the product
Vieira et al. propose using the extract in the DES, which kept the compounds stable for 90 days. At 1 g leaf per 20 g solvent the nonpolar phase holds about 1.8 mg carnosic acid per gram, against 443 mg/g in the two-step CO₂ extract (computed, assuming the solvent splits evenly by mass between the two phases). The DES components then need a regulatory status that fits the product: 21 CFR Part 173 Subpart C, the U.S. list of solvents in food, names no DES (eCFR, as of November 2026).
Heat capacity is the DES advantage: 2.20 to 2.36 kJ/kg·K for choline chloride:glycerol at 30 to 80 °C (Leron and Li 2012, converted to a mass basis), about half that of water.
Viscosity at 40 °C Millipascal-seconds, log scale; DES abbreviations as in Vieira et al. 2022
06 · Life-cycle view
A route is greener only after a full-process comparison per unit of product
Calling a solvent natural or green describes the solvent, not the process. Life-cycle assessment (LCA) compares whole systems against one functional unit, from raw material to waste, following ISO 14040 and ISO 14044. The energy model above is one input to such a comparison, and an incomplete one.
Per gram of carnosic acid at specification
Per kilogram of leaf, pressurized ethanol looks efficient. Per gram of carnosic acid at a purity specification, the four-times-purer CO₂ extract may need less purification downstream, while pressurized ethanol needs about half as much leaf (Sánchez-Camargo values). The ranking can reverse with the unit.
What this model leaves out
Steam, electricity and refrigeration differ in primary energy and emissions. The model also omits solvent production (including DES components, ethanol and CO₂), solvent losses and waste treatment, drying of the extract, cleaning and the equipment itself.
An eco-footprint radar is not an LCA
Hirondart et al. scored pressurized extraction at 2.96 against 100 for Soxhlet on six green-extraction criteria for an analytical method. That is a useful screen; it is not an ISO life-cycle assessment and was not scaled.
Comparative LCAs exist for other materials
Bruna et al. (2025) compared subcritical water and supercritical CO₂ for apple pomace, and Gaalich et al. (2026) reviewed LCA studies of supercritical processes. The sustainability and LCA page summarizes the first result.
07 · Capital intensity and scale-up
Pressure, flammability and the recovery train set capital intensity before any vendor quote
No costs are estimated here. The table ranks routes qualitatively from the operating conditions in the cited studies and the equipment each would need.
| Route | Pressure in the cited studies | Solvent status and hazard | Recovery equipment | Scale-up evidence | Relative capital intensity |
|---|---|---|---|---|---|
| Hot water | 1 atm | Not flammable | Evaporator with the largest duty per kg solvent | Simplest equipment; fails the carnosic acid target (Hirondart) | Lower |
| Ethanol/water | 1 atm | Not in 21 CFR Part 173 Subpart C; § 184.1293 affirms GRAS for a named use (antimicrobial on pizza crusts) and was issued before a general evaluation of ethanol. Flammable | Evaporator or distillation, condenser, solvent storage | Conventional solid–liquid extraction | Moderate |
| Pressurized liquid | 100–130 bar at 150–183 °C | As ethanol/water, plus hot liquid under pressure; browning at 150 °C and above | Pressure-rated vessels and pumps, letdown, evaporator | Sánchez-Camargo pooled laboratory runs because pilot-scale PLE equipment was unavailable to them | Higher |
| Supercritical CO₂ | 150–300 bar at 40 °C | Carbon dioxide listed in 21 CFR 184.1240; high-pressure gas | High-pressure extractor, separators, chilled condenser, pump; cosolvent evaporator | More than 150 industrial SFE plants by 2014 (cited by Sánchez-Camargo) | Highest |
| Deep eutectic solvent | 1 atm | No DES in 21 CFR 173 Subpart C; component status depends on intended use | Undefined: antisolvent or resin step plus water evaporation, or none if the extract stays in the DES | Laboratory studies only among the sources here | Lower vessels; recovery train unknown |
Qualitative judgment The last column is EUTECNIQ's reading of operating pressure and the number of unit operations, not a cost estimate. Hazard analysis for flammable solvent and combustible dust is covered on the process-engineering page.
08 · How EUTECNIQ would decide
EUTECNIQ would keep every route that no other route beats on all five objectives
Recovery, purity, energy, solvent loss and scale-up risk compete. A weighted score hides the trade. EUTECNIQ proposes to keep the non-dominated set, the routes that no other route matches or beats on every objective (a Pareto set), and to spend experiments on the numbers that could change its membership.
| Route | Recovery of carnosic acid published | Purity of the product published | Heat and work, MJ/kg leaf model, defaults | Solvent lost, kg/kg leaf model, defaults | Scale-up risk judgment |
|---|---|---|---|---|---|
| Hot water | None detected | Not reported | 28.4 | 1.2 | Low; fails the carnosic acid target |
| Ethanol/water | 17.7 mg/g in the Soxhlet reference (Hirondart et al.; solvent not stated), 84% of the PLE mean in the same study (difference not significant per the authors) | 63 mg/g extract in the same Soxhlet reference (computed) | 14.5 | 1.2 | Low to moderate; flammable solvent |
| Pressurized liquid | 21.1 to 40.4 mg/g, the highest mean in both studies | 46 to 105 mg/g extract | 15.1 | 1.2 | Moderate to high; pressure, browning above 150 °C |
| Supercritical CO₂ | 18.6 to 20.7 mg/g, 51% of PLE in the same study | 257 to 443 mg/g extract | 5.8, plus 4.8 MJ of chilled condensing | 1.0 CO₂ | Moderate; established at industrial scale, capital-intensive |
| Deep eutectic solvent | 17.5 mg/g, 91% of methanol Soxhlet in the same study | About 1.8 mg/g solvent phase (computed) | 1.8 if kept; 48.5 if 90% recovered | 20 if kept; 2.0 if recovered | High; recovery and residue unresolved |
On these numbers no route is matched or beaten on all five columns by another route. Hot water leaves the set on a hard limit: it does not extract carnosic acid. Cross-study recovery figures are indicative only (section 02).
Confirmatory experiments, in the order EUTECNIQ would run them
- Set one reference basis. One leaf lot, one particle size, one validated HPLC method for rosmarinic acid, carnosic acid and carnosol, and an exhaustive extraction (pressurized 80% ethanol, as Hirondart used for raw-material assay) as the 100% basis. This defines "recovery" for every route.
- Run every route on that lot. Two solvent ratios each, mass balance closed on solids and solvent. This puts recovery and purity on one basis, replacing the cross-study table.
- Measure extraction curves. Recovery against solvent-to-leaf ratio for the two leading routes. This sets the ratio, which the model shows drives energy almost linearly.
- Measure solvent retention and loss. Solvent held by the spent leaf, vent losses and CO₂ vented per batch. This sets the recovery fraction, an assumption on this page.
- Test thermal limits. Carnosic acid to carnosol conversion and browning against temperature and time. This sets the temperature ceiling for pressurized extraction.
- Resolve the DES questions. Measure viscosity against water content and temperature, recovery by resin or antisolvent and residual DES in the final extract, and confirm the regulatory status of each component. This decides whether the DES route stays in the set.
- Meter energy at pilot scale. Steam, electricity and refrigeration for the remaining routes. Measurements then replace this page's model.
- Build the life-cycle inventory. ISO 14044 inventory from the measured balances, per gram of carnosic acid at specification. This decides whether any route can be called greener.
Rule for dropping a route. A route leaves the set as soon as another route matches or beats it on every objective, or when it fails a hard limit such as the purity specification or the regulatory status of a solvent. If no route meets the specification at an acceptable energy and solvent loss, EUTECNIQ would say so and stop.
09 · Sources
References cited in this demonstration
- Sánchez-Camargo AP, García-Cañas V, Herrero M, Cifuentes A, Ibáñez E. Comparative study of green sub- and supercritical processes to obtain carnosic acid and carnosol-enriched rosemary extracts with in vitro anti-proliferative activity on colon cancer cells. Int J Mol Sci. 2016;17(12):2046. doi.org/10.3390/ijms17122046 (extraction conditions and yields used; no biological activity is discussed here)
- Hirondart M, Rombaut N, Fabiano-Tixier AS, Bily A, Chemat F. Comparison between pressurized liquid extraction and conventional Soxhlet extraction for rosemary antioxidants, yield, composition, and environmental footprint. Foods. 2020;9(5):584. doi.org/10.3390/foods9050584
- Vicente G, Molina S, González-Vallinas M, García-Risco MR, Fornari T, Reglero G, Ramírez de Molina A. Supercritical rosemary extracts, their antioxidant activity and effect on hepatic tumor progression. J Supercrit Fluids. 2013;79:101–108. doi.org/10.1016/j.supflu.2012.07.006 (extraction conditions and yields used)
- Vieira C, Rebocho S, Craveiro R, Paiva A, Duarte ARC. Selective extraction and stabilization of bioactive compounds from rosemary leaves using a biphasic NADES. Front Chem. 2022;10:954835. doi.org/10.3389/fchem.2022.954835
- Rosa GP, Barreto MC, Seca AML, Pinto DCGA. Ionic liquids and deep eutectic solvents for polyphenol extraction: opportunities and limitations. Int J Mol Sci. 2026;27(8):3538. doi.org/10.3390/ijms27083538
- Leron RB, Li M-H. Molar heat capacities of choline chloride-based deep eutectic solvents and their binary mixtures with water. Thermochim Acta. 2012;530:52–57. doi.org/10.1016/j.tca.2011.11.036; data from NIST TRC ThermoML
- Bell IH, Wronski J, Quoilin S, Lemort V. Pure and pseudo-pure fluid thermophysical property evaluation and the open-source thermophysical property library CoolProp. Ind Eng Chem Res. 2014;53(6):2498–2508. doi.org/10.1021/ie4033999 (CoolProp 8.0.0 used)
- ISO 14044:2006. Environmental management — Life cycle assessment — Requirements and guidelines (Amd 1:2017, Amd 2:2020; confirmed 2022). iso.org/standard/38498.html
- Bruna L, Marty C, Draye M, Cravotto G, Chatel G. Subcritical or supercritical? A comparative life cycle assessment of bioactive compound extraction from apple pomace. Clean Environ Syst. 2025;18:100311. doi.org/10.1016/j.cesys.2025.100311
- Gaalich I, Lee D, Aymonier C, Sonnemann G, Olchowka J, Philippot G, Loubet P. Environmental impacts of supercritical fluids processes: a critical review of life cycle assessment studies. Environ Impact Assess Rev. 2026;116:108106. doi.org/10.1016/j.eiar.2025.108106
- Electronic Code of Federal Regulations. 21 CFR Part 173, Subpart C (solvents), 21 CFR 184.1240 (carbon dioxide) and 21 CFR 184.1293 (ethyl alcohol). Section list retrieved November 1, 2026. ecfr.gov
Health information notice. These statements have not been evaluated by the Food and Drug Administration. Nothing on this site is intended to diagnose, treat, cure or prevent any disease, and nothing here is medical advice.
Read the full notice
The information is for general education only. It does not replace advice, diagnosis or treatment from a qualified health professional, or any medication a clinician has prescribed. Do not start, stop or change any medicine, supplement or herbal product because of something you read here. Talk to your doctor or pharmacist first, especially if you are pregnant or breastfeeding, have a medical condition, take any medicines, are preparing for surgery, or are caring for a child. Research findings, traditional uses and safety data are reported as their sources state them; they are not recommendations or claims about any product. If you think you may have a medical emergency, call your doctor or 911 at once.
EUTECNIQ Sciences sells no supplements or health products today and receives nothing from any product or ingredient mentioned on this site; any commercial interest it later has in an ingredient will be stated on that ingredient’s page.