Extraction & purification Peer-reviewed methods NIST, NIOSH and ICH data Educational models Proposed capability

Extraction decides what leaves the plant, and the method decides how fast, how hot and what solvent stays behind

Extraction methods, supercritical CO₂ physics, solvent recovery and solvent safety, for readers who must judge a process, not only a yield.

Status. EUTECNIQ owns no extraction equipment and has run no extractions. Laboratory extraction is planned for Phase I, pilot extraction and solvent recovery for Phase II (platform plan). Calculators are educational models.

629 → 910kg/m³: CO₂ density at 40 °C as pressure rises from 10 to 30 MPa

30.98 °Cand 7.38 MPa: the critical point of CO₂, above which it is a supercritical solvent

25 ppmhexane residue allowed in spice oleoresins (21 CFR 173.270)

13 °Cflash point of ethanol: its vapor can ignite at room temperature

Span and Wagner, J Phys Chem Ref Data 1996 (via CoolProp; matches NIST WebBook); eCFR, 21 CFR 173.270, October 2026; NIOSH Pocket Guide, ethanol.

Grinding, solvent and pressure set the yield, the purity and the residue

Grinding sets the pace. Broken cells release their contents fast. Intact cells leak slowly, by diffusion.
CO₂ is a tunable solvent. At 40 °C, raising the pressure from 10 to 30 MPa makes CO₂ 45% denser, and its dissolving power rises steeply with density.
A laboratory yield is not a plant cost. Solvent-to-feed ratio, the energy to recover solvent, cycle time, downstream purification and the fire-safety class of the solvent decide the cost per kilogram.

How it works

Solvent gets in, actives dissolve, then they get out: fast from broken cells, slowly from intact ones

Milling opens the cells on each particle's surface. Their contents wash out quickly. Cells deeper inside stay closed, so their contents must cross cell walls first.

One ground particle, and the yield curve it producesIllustrative schematic and model

Broken-and-intact-cells concept: Sovová H. Chem Eng Sci. 1994;49(3):409–414. Curve: the simplified two-period form described under the extraction-curve explorer below, with invented parameters.

Finer grinding opens more cells, so more of the yield arrives in the fast first phase.

Explanation, details and sources

The five steps a solvent goes through, in order:

GrindMilling breaks cells open on each particle's surface.
WetSolvent soaks into the particle's pores.
DissolveActives dissolve until the solvent is saturated or they run out.
Wash outFree solute from broken cells leaves quickly.
Diffuse outSolute in intact cells crosses cell walls slowly.

Why it mattersParticle size and grinding are cheap levers. The Science lab diffusion calculator shows the slow tail: halve the particle size and it runs four times faster.

Six methods

Six ways to get actives out of a plant, each trading time, heat and solvent

Schematics, not to scale Hover a part for detail.

Maceration and percolation

How it works, conditions, strengths and limits

How it works: In maceration, ground plant soaks in solvent. In percolation, solvent flows through a packed bed, so saturated solvent is constantly replaced by fresh solvent.
Typical conditions: Room temperature, atmospheric pressure, long contact times, large solvent volumes. Published example: soak 24 h in 70% ethanol, then percolate with 20 times the herb's amount of 70% ethanol.
Strengths: Simple equipment; no heat, so suited to heat-sensitive components.
Limits: Long extraction time and low efficiency; high solvent use.

Zhang et al. 2018 (Table 1 and text); Nayak et al. 2026 (Table 1)

Soxhlet extraction

How it works, conditions, strengths and limits

How it works: Solvent boils, condenses and drips onto the plant in a porous thimble. When the chamber fills, a siphon drains it back to the flask. The plant always meets freshly distilled solvent.
Typical conditions: Organic solvents boiled in the flask, atmospheric pressure, long runs, moderate solvent volume.
Strengths: Automatic and continuous; less time and solvent than maceration or percolation.
Limits: Heat and long runs raise the risk of degradation. With tea, Soxhlet at 70 °C recovered less total polyphenol and alkaloid than maceration at 40 °C.

Zhang et al. 2018 (Table 1 and text)

Ultrasound-assisted extraction

How it works, conditions, strengths and limits

How it works: Ultrasound (above 20 kHz) makes tiny bubbles form and collapse in the solvent. The collapse breaks and erodes plant tissue, which speeds up dissolving and diffusion.
Typical conditions: 20–120 kHz and 20–700 W in published fruit and vegetable studies. Published example: 50% ethanol, 1:30 solid-to-solvent, 0.3 mm particles, 15 min.
Strengths: Short times at lower temperature; less solvent and energy; suits heat-sensitive compounds.
Limits: Too much intensity can degrade actives. Cavitation becomes uneven at large scale.

Kumar et al. 2021; Zhang et al. 2018; Nayak et al. 2026 (Table 1)

Microwave-assisted extraction

How it works, conditions, strengths and limits

How it works: Microwaves heat polar molecules such as water and ethanol (ionic conduction and dipole rotation). Moisture inside the plant heats up, so heat and mass move outward together.
Typical conditions: 2.45 GHz microwaves; the solvent must absorb them (polar, such as water, ethanol or methanol). Published example: 80% ethanol, 25:1 liquid-to-solid, 1.5 kW, 7 min.
Strengths: Fast; less organic solvent; heats the plant material selectively.
Limits: Uneven heating creates hot spots that can degrade actives. Limited penetration at 2.45 GHz makes large volumes heat unevenly.

Zhang et al. 2018; Nayak et al. 2026 (Table 1 and text)

Pressurized liquid extraction

How it works, conditions, strengths and limits

How it works: Pressure keeps the solvent liquid above its normal boiling point. Hot liquid solvent dissolves more, diffuses faster and penetrates the plant better.
Typical conditions: The U.S. EPA method (written for soils, sediments, sludges and waste solids) runs 100–180 °C at 1,500–2,000 psi (10.3–13.8 MPa) with 5-minute static steps.
Strengths: Much less time and solvent than older methods; better repeatability.
Limits: High temperature can harm heat-sensitive actives: Maillard reactions appeared at 200 °C with grape pomace. Specialized, costly equipment.

Zhang et al. 2018; U.S. EPA Method 3545A (2007); Nayak et al. 2026 (Table 1)

Supercritical CO₂ extraction

How it works, conditions, strengths and limits

How it works: Above 31 °C and 7.38 MPa, CO₂ diffuses like a gas but is dense like a liquid. Pressure tunes what it dissolves. Lowering the pressure drops the extract out, and the CO₂ can be recycled.
Typical conditions: Published example: CO₂ with 2% ethanol at 30 MPa (300 bar) and 40 °C, to extract vinblastine from Catharanthus roseus.
Strengths: Low critical temperature suits heat-sensitive compounds; CO₂ is non-toxic, inert and low-cost; little or no organic solvent; CO₂ leaves the extract as a gas.
Limits: CO₂ is non-polar, so polar actives need a co-solvent such as ethanol. High equipment cost, mainly from the high pressure.

Zhang et al. 2018; Tzima et al. 2023; Nayak et al. 2026 (Table 1); critical point: Span and Wagner 1996

Faster methods use energy (ultrasound, microwaves, heat, pressure) to speed up the same two steps: dissolving and diffusing.

Compare

No method wins on every axis: the molecule, the residue the product can tolerate and the scale decide

The matrix follows a published relative comparison. The table below it adds what sets selectivity, what drives cost and what limits scale.

Seven methods on four criteria: the fast, low-solvent methods pay in pressure or in heatPublished relative ratings

Time, organic solvent use and pressure: relative ratings in Zhang Q-W, Lin L-G, Ye W-C. Chin Med. 2018;13:20, Table 1. Heat-sensitive actives and limitations: the review text of Zhang et al. 2018 and Nayak A, et al. Ultrason Sonochem. 2026;124:107712, Table 1; Tzima S, et al. Molecules. 2023;28(3):1410.

Text version: the published relative comparison

Relative comparison of seven extraction methods
Method	Pressure	Time	Organic solvent use	Compounds reached	Reported limitation
Maceration	Atmospheric	Long	Large	Depends on the solvent	Long time, low efficiency (Zhang)
Percolation	Atmospheric	Long	Large	Depends on the solvent	Slow; high solvent use (Zhang; Nayak)
Soxhlet	Atmospheric	Long	Moderate	Depends on the solvent	Heat and long runs can degrade actives (Zhang)
Pressurized liquid	High	Short	Small	Depends on the solvent	High temperature can degrade heat-sensitive actives; specialized equipment (Nayak)
Supercritical CO₂	High	Short	None or small	Non-polar to moderately polar	Poor for polar actives without a co-solvent; high equipment cost (Nayak; Tzima)
Ultrasound-assisted	Atmospheric	Short	Moderate	Depends on the solvent	High intensity can degrade actives; uneven cavitation at scale (Nayak)
Microwave-assisted	Atmospheric	Short	None or moderate	Depends on the solvent	Hot spots; uneven heating in large volumes (Nayak)

Zhang Q-W, Lin L-G, Ye W-C. Chin Med. 2018;13:20, Table 1 and text. Nayak A, et al. Ultrason Sonochem. 2026;124:107712, Table 1. Tzima S, et al. Molecules. 2023;28(3):1410. Pills: relative ratings as given by Zhang et al.

[1] Zhang et al. 2018, Table 1 and text. [2] Kumar K, et al. Ultrason Sonochem. 2021;70:105325. [3] Nayak et al. 2026. [4] U.S. EPA SW-846 Method 3545A (2007). [5] Tzima et al. 2023. [6] Cheng Y, Xue F, Yu S, Du S, Yang Y. Subcritical water extraction of natural products. Molecules. 2021;26(13):4004; water's critical point: IAPWS-95 via CoolProp 8.0.0. [7] Puri M, Sharma D, Barrow CJ. Trends Biotechnol. 2012;30(1):37–44. [8] Cassano A, et al. Int J Mol Sci. 2018;19(2):351; Zhang et al. 2018. [9] McCabe WL, Smith JC, Harriott P. Unit Operations of Chemical Engineering. 7th ed. McGraw-Hill; 2005 (leaching). Unbracketed cells are general engineering judgments, not data.

A method that wins in a published laboratory comparison can lose in a plant, where solvent recovery, pressure rating, cycle time and fire-safety class set the cost.

Explanation, details and sources

Engineering comparison of ten extraction approaches: principle, operating window, selectivity levers, scale-up, cost drivers and limits
Approach	Principle	Operating window (published)	Selectivity levers	Scale-up	Main cost drivers	Main limits and hazards
Maceration, percolation	Soaking, or solvent flowing through a packed bed; diffusion-limited	Room temperature, atmospheric pressure [1]	Solvent and its water content, particle size, contact time	Simple vessels; straightforward to enlarge	Solvent volume to recover; long cycle times tie up vessels	Slow; large flammable-solvent inventory when ethanol is used
Soxhlet	Repeated reflux of freshly distilled solvent	Solvent boiling point, atmospheric [1]	Solvent choice	Mainly a laboratory and analytical reference method	Energy for continuous boiling	Prolonged heating degrades heat-sensitive actives [1]
Ultrasound-assisted (UAE)	Cavitation disrupts tissue and speeds mass transfer	20–120 kHz, 20–700 W in reviewed studies [2]	Solvent, power, time, temperature	Cavitation becomes uneven in large volumes [3]	Transducer power and probe or flow-cell design	Excess intensity can degrade actives [3]
Microwave-assisted (MAE)	Dielectric heating of the solvent and plant moisture	2.45 GHz; solvent must absorb microwaves [3]	Solvent polarity, power, time	Limited penetration depth heats large volumes unevenly [3]	Microwave equipment and power	Hot spots; closed vessels build pressure
Pressurized liquid (PLE, ASE)	Liquid solvent held above its boiling point by pressure	100–180 °C, 10.3–13.8 MPa in EPA Method 3545A (written for soils and wastes) [4]	Temperature, solvent	Standard in analytical laboratories	Pressure-rated cells and pumps	Thermal degradation; Maillard products at 200 °C in grape pomace [3]
Supercritical CO₂ (SFE)	Dense CO₂ above 30.98 °C and 7.38 MPa; density sets dissolving power	Example: 30 MPa, 40 °C, 2% ethanol [1]	Pressure, temperature, co-solvent, staged separators [5]	Vessel volume and pressure rating set the cost of each step up	Capital for high-pressure equipment [3]; compression energy; CO₂ make-up	Polar actives need a co-solvent [3]; high-pressure and CO₂ asphyxiation hazards (see process safety)
Subcritical water (pressurized hot water)	Liquid water heated below its critical point (373.95 °C, 22.06 MPa); its polarity falls as it heats	Water at 200, 250 and 300 °C has a dielectric constant similar to acetonitrile, methanol and ethanol [6]	Temperature, time	Pressure vessel and heat recovery needed; mostly laboratory and pilot reports [6]	Heating water; drying an aqueous extract (water's latent heat is the highest of common solvents)	Hydrolysis of glycosides and esters; Maillard and caramelization products; line plugging [6]
Enzyme-assisted (EAE)	Enzymes hydrolyze cell-wall polysaccharides to release contents	Cellulase, α-amylase and pectinase; room temperature or heated after treatment [1]	Enzyme choice and dose, pH, time	Fits existing aqueous vessels [7]	Enzyme cost; incubation time; enzyme inactivation step	Can change composition versus solvent-only extracts; the enzyme preparation needs a lawful basis for the product category
Membrane-assisted processing	Pressure-driven filtration separates by molecular size (MF, UF, NF)	Selection by molecular-weight cut-off, pressure, temperature and feed flow [8]	Membrane material and cut-off [8]	Modular; area scales with throughput	Membrane area and replacement; cleaning; pumping	Fouling; similar-sized molecules are not cleanly separated [8]
Continuous countercurrent extraction	Solids and solvent move in opposite directions, so the leanest solvent meets the most-depleted solids	Set by the solvent and the equipment design	Number of stages, solvent-to-feed ratio	Industrial leaching equipment; needs steady feed [9]	Capital for continuous equipment; lower solvent per kilogram for a given recovery	More complex control; staging is modeled on Process engineering

Supercritical CO₂

Pressure turns CO₂ from a thin gas into a dense, liquid-like solvent

CO₂ density over temperature and pressure: steep near the critical point, flatter at extraction pressuresReference equation of state

Span R, Wagner W. J Phys Chem Ref Data. 1996;25(6):1509–1596, evaluated with CoolProp (computed October 24, 2026). Marked conditions: Zhang et al. 2018 (30 MPa, 40 °C, with 2% ethanol; density shown for pure CO₂) and Prado et al. 2012 (35 MPa, 313 K). Drag to rotate.

Data table: CO₂ density, kg/m³

Pressure	32 °C	40 °C	60 °C	80 °C
7.5 MPa	366	232	173	147
10 MPa	750	629	290	222
15 MPa	835	780	604	427
20 MPa	881	840	724	594
30 MPa	940	910	830	746
40 MPa	982	956	890	823

30.98 °Ccritical temperature of CO₂

7.38 MPacritical pressure (about 73 atmospheres)

629 → 910kg/m³ at 40 °C, from 10 to 30 MPa (NIST)

−10%Peng–Robinson density error at 40 °C and 10 MPa

CO₂ density versus pressure at four temperaturesReference equation of state

Span R, Wagner W. A new equation of state for carbon dioxide. J Phys Chem Ref Data. 1996;25(6):1509–1596, evaluated with CoolProp 8.0.0 (Bell IH, et al. Ind Eng Chem Res. 2014;53(6):2498–2508). Values match the NIST Chemistry WebBook at 40 °C (5–30 MPa), retrieved October 22, 2026.

Show data

CO2 density in kilograms per cubic meter by pressure and temperature
Pressure (MPa)	32 °C	40 °C	50 °C	60 °C
5	121.5	113.1	104.8	98.3
7.5	365.9	231.5	193.9	172.8
10	750.0	628.6	384.3	290.0
15	834.5	780.2	699.8	604.1
20	880.7	839.8	784.3	723.7
25	914.0	879.5	834.2	786.6
30	940.5	909.9	870.4	829.7
40	981.9	956.1	923.3	890.1

Density in kg/m³ from the Span–Wagner equation of state (CoolProp 8.0.0).

Near the critical point, a small pressure change moves density a lot. That is the control knob for what CO₂ dissolves.

Peng–Robinson equation of stateP = RTV − b − a(T)V(V + b) + b(V − b)

Temperature Pressure

Reference density (NIST)

Peng–Robinson density

Peng–Robinson error

Assumptions

Pure CO₂; T_c = 304.13 K, P_c = 7.3773 MPa, ω = 0.22394 (Span–Wagner values in CoolProp)
Standard Peng–Robinson (1976) with no volume translation
Across 32–80 °C and 1–40 MPa, its error ranges from −15% (32 °C, 7.6 MPa) to +4%

Simple model versus reference data at the chosen temperatureModel comparison

Peng D-Y, Robinson DB. A new two-constant equation of state. Ind Eng Chem Fundam. 1976;15(1):59–64. Reference: Span & Wagner (1996) via CoolProp.

Method, assumptions and limits: Peng–Robinson versus reference density

What it calculates: The density of pure CO₂ at the chosen temperature and pressure from the Peng–Robinson cubic equation, and its error against the Span–Wagner reference equation of state.
Equations: a = 0.45724 R²T_c²/P_c · α(T); b = 0.07780 RT_c/P_c; α = [1 + κ(1 − √(T/T_c))]², κ = 0.37464 + 1.54226ω − 0.26992ω². The cubic in the compressibility factor Z is solved analytically; where three roots exist, the root with the lowest fugacity coefficient is kept. Density ρ = PM/(ZRT).
Variables and units: T in K (slider in °C), P in Pa (slider in MPa), R = 8.314462618 J/(mol·K), M = 0.0440098 kg/mol, ρ in kg/m³. Reference densities come from a CoolProp grid (0.1 MPa steps from 6 to 12 MPa, 0.5 MPa elsewhere, every 1 °C) with linear interpolation in pressure.
Valid range: 32–80 °C and 1–40 MPa; inputs are clamped to that range. Every state shown is above the critical temperature, so no phase boundary is crossed.
What it does not establish: It does not predict solubility, extraction yield or the behavior of CO₂ with a co-solvent or dissolved extract. It shows only why process design uses a reference equation of state near the critical point.
Verification: Independent Python cases (CoolProp 8.0.0 Span–Wagner and CoolProp's own Peng–Robinson backend) in tests/numerics/ref/technologies-extraction.py; at 40 °C and 10 MPa the reference is 628.61 kg/m³ (NIST WebBook) and Peng–Robinson gives 563.5 kg/m³ (−10.4%).

A one-line equation of state misses CO₂ density by up to 15% near the critical point; the error shrinks far from it.

Chrastil (1982) density relationln S = k ln ρ + aT + b

Exponent k (association number)

Density gain, 10 → 30 MPa

Solubility gain, same step

Assumptions

Fixed temperature (40 °C), so the a/T term cancels in the ratio
k is fitted from solubility data for each compound; the slider value is illustrative
Real solubility also depends on the solute's vapor pressure

Relative solubility in CO₂ at 40 °C (1 = solubility at 10 MPa)Illustrative exponent

Chrastil J. Solubility of solids and liquids in supercritical gases. J Phys Chem. 1982;86(15):3016–3021; association-number interpretation as described in Bawazeer S. Sci Rep. 2025;15:42267. Densities: Span & Wagner (1996) via CoolProp.

Method, assumptions and limits: Chrastil solubility ratio

What it calculates: How many times more of a solute CO₂ can dissolve at 30 MPa than at 10 MPa, both at 40 °C, for an assumed exponent k.
Equation: At fixed temperature, S₂/S₁ = (ρ₂/ρ₁)^k, because the a/T + b terms cancel. ρ from the Span–Wagner reference grid.
Variables and units: S in g/L of CO₂ (only the ratio is shown, so units cancel); ρ in kg/m³; k dimensionless, 2–12 (clamped).
Assumptions: The Chrastil form holds for the solute over this density range; k is a fitted constant for each solute and is illustrative here.
What it does not establish: It does not give the solubility of any named botanical compound. Fitted Chrastil parameters from measured solubility data are needed for that.
Verification: Python reference cases at k = 2, 6 and 12 using CoolProp densities (628.61 and 909.89 kg/m³).

Because solubility scales with a power of density, a 45% density gain can mean several times more dissolving power.

Pressure against temperature for carbon dioxide: solid, liquid, gas and supercritical regions, the triple point at −56.56 °C and 0.518 MPa, the critical point at 30.98 °C and 7.377 MPa, and lines of constant density from 100 to 1,100 kg per cubic meter. Above the critical point a small change in pressure or temperature changes the density of the fluid strongly. Computed from the Span and Wagner (1996) reference equation of state. — Span R, Wagner W. J Phys Chem Ref Data 1996;25:1509 (vapor, melting and sublimation pressure equations); isochores from the same equation of state in CoolProp 8.0. Generated illustration, EUTECNIQ. Full-size SVG

Extraction curve

A supercritical extraction curve has a fast straight start and a slow tail

Two-period broken-and-intact-cells curve (simplified from Sovová, type A)e = y_s q for q ≤ q₁ = r x_u/y_s
e = x_u[1 − (1 − r) e^{−C₂(q − q₁)}] for q > q₁

Broken cells r (grinding efficiency) Solubility in CO₂, y_s Intact-cell release rate, C₂

End of fast period, q₁

Yield after 50 kg CO₂/kg

CO₂ for 90% of the oil

Assumptions

Extractable oil x_u = 0.20 kg per kg of oil-free plant solids; q and e use the same basis. All parameter values are illustrative
Type A curve: one straight section, then an exponential tail. Some materials show a second straight section (type B)
Real parameters are fitted from pilot runs at fixed pressure, temperature and flow

Share of the oil recovered versus CO₂ passedModel output

Simplified from Sovová H. J Supercrit Fluids. 2005;33(1):35–52 (type A); published equations as given by Mouahid A, Claeys-Bruno M, Clercq S. Processes. 2024;12(9):1865, Eqs. 1–4. Model origin: Sovová H. Chem Eng Sci. 1994;49(3):409–414.

Isometric drawing of a generic supercritical carbon dioxide extraction loop: make-up CO2 from a supply cylinder; a condenser that liquefies and subcools the CO2; a high-pressure pump; a heater that takes it above 31 degrees C; an extraction vessel holding a packed bed of ground plant; a heated back-pressure regulator that lets the pressure down; a separator where the CO2 turns to gas and the extract drops out to a collection vessel; and the CO2 gas returning to the condenser. Instrument bubbles use ISA-5.1 tags. — Critical point: Span R, Wagner W. J Phys Chem Ref Data 1996;25:1509. Equipment order and functions are standard practice; no operating condition or vendor design implied. Generated illustration, EUTECNIQ. Full-size SVG

Method, assumptions and limits: extraction curve

What it calculates: The share of the extractable oil recovered after a given mass of CO₂ per kilogram of oil-free plant solids has passed through a packed bed.
Equations: First period: e = y_sq until the broken-cell oil r x_u is exhausted at q₁ = r x_u/y_s. Second period: first-order release from intact cells, e = x_u[1 − (1 − r)exp(−C₂(q − q₁))]. CO₂ for a fraction f > r: q = q₁ + ln[(1 − r)/(1 − f)]/C₂.
Variables and units: q: kg CO₂ per kg oil-free solids; e: kg oil per kg oil-free solids; y_s: kg oil per kg CO₂ (slider in g/kg); r: fraction of oil in broken cells; C₂: per (kg CO₂/kg solids); x_u = 0.20.
Simplification: This form is continuous at q₁ and ignores release from intact cells during the first period. Sovová's published type A approximation (Mouahid et al. 2024, Eqs. 1–3) writes the tail as x_u[1 − C₁exp(−C₂q)] with r = 1 − C₁exp(−C₂q_c/2), which counts that release. The simplified curve therefore shows a sharper transition than a fitted curve.
Valid range: r 10–90%, y_s 1–15 g/kg, C₂ 0.005–0.08 (clamped). Plug flow, constant solubility, no co-solvent.
What it does not establish: It does not predict the yield of any real material. Pilot runs at fixed pressure, temperature and flow would be needed to fit y_s, r and C₂.
Verification: Independent Python cases (closed-form and numerical root-finding) for the three readouts at the default, low and high settings.

Better grinding stretches the fast straight start; higher solubility (more pressure) makes it steeper; neither speeds up the slow tail much.

Explanation, details and sources

The straight start is free oil from broken cells, limited by how much CO₂ can carry. The tail is oil diffusing out of intact cells. Move the sliders to see which lever changes what.

Why it mattersThe CO₂ needed per kilogram of extract sets run time and energy; pilot runs would fit this model to price each lever. Staging is on Process engineering.

Yield, purity, recovery

Yield, purity and recovery are three different numbers, tied together by one mass balance

Mass balance on the target constituentY · P = R · x_f so P = R x_f / Y

DER and the "higher ratio is better" misconception: Monagas M, et al. Understanding plant to extract ratios in botanical extracts. Front Pharmacol. 2022;13:981978.

One kilogram of plant through extraction and solvent recovery, at the calculator's default settingsEducational model

Calculator defaults: ethanol, 10 L per kg of dry plant, Y = 12.5%, x_f = 2%, R = 80%; ethanol density 0.785 kg/L and latent heat 0.850 MJ/kg (CoolProp 8.0.0; the measured 38.56 kJ/mol in the NIST WebBook gives 0.837 MJ/kg, the value used on the process-engineering page, 1.5% lower). Same assumptions as the calculator: single pass, all solvent evaporated once and none retained in the spent plant, latent heat only. Letters A–F mark where the cost per kilogram is set; the schematic is not to scale and implies no equipment design.

Solvent Solvent-to-feed ratio (liters per kg of dry plant) Extraction yield, Y (dry extract, % of dry plant) Target content of the plant, x_f (% w/w) Target recovery, R

Target content of the extract, P

Enrichment over the plant

Native DER (plant : extract)

Solvent per kg of extract

Latent heat per kg of extract

Same, in kilowatt-hours

Heat to evaporate the solvent, per kg of dry extract, at the chosen ratio and yieldEducational model with reference data

Show data

Solvent mass and latent heat per kilogram of dry extract at the chosen settings
Solvent	Solvent per kg extract (kg)	Latent heat per kg extract (MJ)

Solvent property data used by the model
Solvent	Density at 25 °C (kg/L)	Normal boiling point (°C)	Latent heat at boiling point (MJ/kg)	Flash point (°C)
Water	0.997	100.0	2.257	not flammable
Ethanol	0.785	78.4	0.850	13
Methanol	0.786	64.5	1.101	11
Acetone	0.785	56.1	0.501	−18
Hexane	0.655	68.7	0.335	−22
Heptane	0.680	98.4	0.317	−4

Density, boiling point and latent heat: CoolProp 8.0.0 reference equations of state. Flash points: NIOSH Pocket Guide, converted from °F.

Thermophysical properties: Bell IH, et al. Ind Eng Chem Res. 2014;53(6):2498–2508 (CoolProp 8.0.0). Flash points: NIOSH Pocket Guide to Chemical Hazards, retrieved October 22, 2026.

Method, assumptions and limits: mass balance and solvent load

What it calculates: From the yield, the plant's target content and the recovery, the target content of the extract; and from the solvent-to-feed ratio and yield, the mass of solvent and the latent heat needed to evaporate it, per kilogram of dry extract.
Equations: P = R x_f/Y; enrichment = P/x_f; native DER = 1/Y; solvent per kg extract m = (S/F)·ρ/Y; heat Q = m·Δh_vap(T_b); 1 kWh = 3.6 MJ.
Variables and units: S/F in L of solvent per kg of dry plant; ρ in kg/L at 25 °C; Y, x_f, R as fractions (sliders in %); Δh_vap in MJ/kg at the normal boiling point.
Assumptions: Single pass: all solvent fed is evaporated once and none is retained in the spent plant. Pure solvents (ethanol–water blends, which are common, would fall between the two pure values). Latent heat only: no sensible heat, no losses, no plant moisture.
Valid range and guards: Inputs are clamped to the slider ranges. If R x_f > Y, the purity would exceed 100%, which is impossible; the tool says so instead of showing a number as valid.
What it does not establish: It is not an energy or cost estimate for any plant. Multiple-effect evaporation and vapor recompression supply far less steam per kilogram evaporated than a single-effect evaporator, solvent retained in the spent plant must also be recovered, and solvent losses, waste treatment and labor are not included.
Verification: Independent Python cases that evaluate the properties directly in CoolProp 8.0.0 for each solvent, including an inconsistent-input case.

Text version of the cost drivers

The cost per kilogram at scale depends on the solvent-to-feed ratio and how much of that solvent must be recovered, the energy for evaporation, cycle time and vessel occupancy, the purification needed to reach the specification, the pressure rating of the equipment, the fire-code class of the solvent and the operators' exposure limits.

Explanation, details and sources

A quoted "15% yield" says how much extract came out, not how much active it contains or how much of the plant's active was captured. A buyer needs all three, and the solvent each kilogram cost.

Extraction yield, Y

Dry extract per dry plant. Its inverse is the native drug-to-extract ratio (DER); a higher ratio is not a stronger or better extract.

Purity (content), P

Target constituent per mass of extract, by a validated assay. Purification raises it, usually at some cost in recovery.

Recovery, R

Share of the plant's target that ends up in the extract; it needs an assay of the plant too.

At the same ratio and yield, water needs about ten times the evaporation heat of hexane and over three times that of ethanol, while hexane's flash point is −22 °C. Energy and fire risk pull in opposite directions.

A better laboratory yield does not guarantee better industrial economics. EUTECNIQ would compare candidate processes on modeled cost of goods and safety class at pilot scale before committing capital (proposal).

Fractionation & purification

Purification trades recovery for purity, so each step must earn its place in the specification

Six purification families, each separating on a different propertySchematic, not to scale

Zhang et al. 2018; Ito Y. J Chromatogr A. 2005;1065(2):145–168; Friesen JB, Pauli GF. J Liq Chromatogr Relat Technol. 2005;28(17):2777–2806; Soto ML, et al. J Food Eng. 2011;105(1):1–27; Cassano A, et al. Int J Mol Sci. 2018;19(2):351; Tzima et al. 2023. Simulated moving bed chromatography (continuous, less solvent) is described in the text version.

Text version of the six purification families

Partition

Liquid–liquid partition and solvent-system optimization

The extract is split between two immiscible liquids; each compound distributes according to its partition coefficient. Changing the solvent pair or the ratio within a solvent family moves the cut.

Ito 2005; Friesen & Pauli 2005

Support-free chromatography

Countercurrent and centrifugal partition chromatography (CCC, CPC)

Partition chromatography in which the stationary phase is itself a liquid held by gravity or centrifugal force, so there is no solid support to adsorb the sample irreversibly. The two-phase solvent system is chosen so the target's partition coefficient falls in a workable range.

Zhang et al. 2018; Ito 2005; Friesen & Pauli 2005

Adsorption

Macroporous adsorbent resins

Non-ionic polymer adsorbents take up compounds from an aqueous extract and release them when eluted with a stronger solvent, typically aqueous ethanol. Reviewers cite high capacity, relatively low cost, easy regeneration and easy scale-up; surface area, pore diameter and polarity set capacity.

Zhang et al. 2018; Soto et al. 2011

Adsorption

Preparative chromatography

Flash and preparative HPLC on silica or reversed-phase media resolve closely related compounds, but solvent and media consumption make them costly per kilogram. Simulated moving bed chromatography runs continuously and uses less solvent.

Zhang et al. 2018

Size

Membranes

Microfiltration clarifies; tight ultrafiltration and nanofiltration concentrate and fractionate by molecular-weight cut-off at low temperature. Separation depends on membrane material, cut-off, pressure, temperature and flow, and fouling limits throughput.

Cassano et al. 2018; Zhang et al. 2018

Pressure

Fractional separation of supercritical extracts

Two or more separators at falling pressure drop out different fractions of a CO₂ extract, because dissolving power falls with density.

Tzima et al. 2023

Explanation, details and sources

A crude extract holds hundreds of compounds. Enrichment uses differences in partition, adsorption, molecular size or, for CO₂ extracts, density. Equations, a purity–recovery calculator and lot blending are on Purification and standardization.

Every purification step also changes what a "standardized" extract means: an enriched fraction is a different material from the native extract, and comparability, labeling and, for supplements, whether a new dietary ingredient notification is needed must be assessed for the new material (see NDIN).

Solvents

Solvent choice is a safety and quality decision, not only a yield decision

Residual-solvent limits in medicines (ppm)ICH Q3C

U.S. FDA. Q3C—Tables and List: Guidance for Industry, Revision 3 (June 2017). Class 2 values are the concentration limits listed in the guidance; for Class 3 solvents, 50 mg/day or less (5,000 ppm, 0.5%) is acceptable without justification. Water and CO₂ do not appear in the Q3C tables. Food limits for solvent residues in spice oleoresins (21 CFR 173) are charted in the CFR atlas.

Show data

ICH Q3C class, permitted daily exposure and concentration limit by solvent
Solvent	Class	PDE (mg/day)	Limit (ppm)
Benzene	1	not set (avoid)	2
Chloroform	2	0.6	60
Hexane	2	2.9	290
Acetonitrile	2	4.1	410
Dichloromethane	2	6.0	600
Toluene	2	8.9	890
Methanol	2	30.0	3,000
Ethanol	3	50 or less without justification	5,000 (default)
Acetone	3	50 or less without justification	5,000 (default)
Ethyl acetate	3	50 or less without justification	5,000 (default)
Heptane	3	50 or less without justification	5,000 (default)
2-Propanol	3	50 or less without justification	5,000 (default)

Hansen solubility parameters: polar versus hydrogen-bonding (MPa½)Published values

Ethanol in water (by volume)

Blend δP (MPa½)

Blend δH (MPa½)

Distance to soybean oil, R_a (hexane: 4.6)

Show data

Hansen solubility parameters by liquid
Liquid	δD	δP	δH	Source
Hexane	14.9	0	0	Carré et al. 2024
Heptane	15.3	0	0	Hansen 2007
Ethyl acetate	15.8	5.3	7.2	Carré et al. 2024
Acetone	15.5	10.4	7.0	Carré et al. 2024
2-Propanol	15.8	6.1	16.4	Carré et al. 2024
Ethanol	15.8	8.8	19.4	Carré et al. 2024
Methanol	15.1	12.3	22.3	Carré et al. 2024
Water	15.6	16.0	42.3	Hansen 2007
Soybean oil	16.5	2.0	2.7	Hansen 2004, via Carré et al. 2024

All values in MPa½.

Carré P, et al. OCL. 2024;31:31, Table 2 (hexane, alcohols, acetone, ethyl acetate; soybean oil from Hansen 2004). Hansen CM. Hansen Solubility Parameters: A User's Handbook, 2nd ed. CRC Press; 2007, as tabulated by Accu Dyne Test (water, heptane; accessed October 22, 2026). Blend rule and distance R_a² = 4ΔδD² + ΔδP² + ΔδH²: hansen-solubility.com, HSP Basics. Hansen parameters split a liquid's cohesion into dispersion (δD), polar (δP) and hydrogen-bonding (δH) parts; the closer two sets are, the more alike the liquids. δD varies little across these solvents (14.9–15.8 MPa½; soybean oil 16.5), so the chart shows the two that separate them; R_a uses all three.

Screening seven solvents on safety, residue limit and oil solvency at once: brush an axis to keep only what passesValues cited on this page

Residual-solvent limits from FDA, Q3C—Tables and List, Revision 3 (June 2017; Class 3 shown at the 5,000 ppm default; food limits for spice oleoresins: 21 CFR 173 in the CFR atlas); NIOSH Pocket Guide flash points and lower flammable limits; Hansen parameters from OCL 2024 Table 2 and Hansen 2007; distance to soybean oil R_a² = 4ΔδD² + ΔδP² + ΔδH², lower for a better oil solvent. Boiling points and δP, δH are in the table.

Data table

Solvent	ICH limit, ppm	Flash point, °C	LEL, %	Boiling point, °C	δP	δH	R_a to soybean oil
Hexane	290	-22	1.1	69	0	0	4.6
Heptane	5,000	-4	1.05	98	0	0	4.1
Ethyl acetate	5,000	-4	2	77	5.3	7.2	5.8
Acetone	5,000	-18	2.5	56	10.4	7	9.6
2-Propanol	5,000	12	2	83	6.1	16.4	14.4
Ethanol	5,000	13	3.3	78	8.8	19.4	18.1
Methanol	3,000	11	6	64	12.3	22.3	22.3

Method, assumptions and limits: Hansen blend and distance

What it calculates: The Hansen parameters of an ethanol–water blend and its Hansen distance to soybean oil, as a guide to how alike the two liquids are.
Equations: Blend: δ_i = φ_Eδ_i,E + (1 − φ_E)δ_i,W for i = D, P, H, with φ_E the ethanol volume fraction. Distance: R_a = [4(ΔδD)² + (ΔδP)² + (ΔδH)²]^½.
Assumptions: Ideal volume-fraction mixing (no volume change on mixing); 25 °C parameter values.
What it does not establish: A smaller distance indicates likely miscibility; it does not give the solubility of any botanical compound or the selectivity of an extraction. Strongly hydrogen-bonding mixtures such as ethanol–water deviate from ideal blending.
Verification: Independent Python cases at 0%, 70% and 100% ethanol.

Deep eutectic solvents and ionic liquids are extraction research until four questions are answered for a useProposed screening sequence

eCFR, 21 CFR Part 173 Subpart C, up to date as of October 15, 2026. Dai Y, et al. Anal Chim Acta. 2013;766:61–68 (more than 100 natural deep eutectic mixtures). Passos H, Freire MG, Coutinho JAP. Green Chem. 2014;16:4786–4815 (limits of toxicity, cost and melting point). The sequence is an EUTECNIQ proposal.

Text version: research solvents and what they mean for a product

Research solvents

Deep eutectic solvents (DES, NADES) and ionic liquids are extraction research, not ready formulation components

Natural deep eutectic solvents are liquids formed when certain plant primary metabolites, solid on their own, are mixed in the right ratio; one group reported more than 100 such mixtures and found many biomolecules far more soluble in them than in water. Ionic liquids are salts that are liquid at low temperature. The same authors note that ionic liquids and deep eutectic solvents still face limits of toxicity, cost and, for many DES, being solid at room temperature.

Dai Y, et al. Anal Chim Acta. 2013;766:61–68. Passos H, Freire MG, Coutinho JAP. Green Chem. 2014;16:4786–4815.

What that means for a product

A good extraction solvent is not automatically an acceptable residue, excipient or food component

U.S. food regulations list specific extraction solvents and their permitted uses and residues (21 CFR Part 173, Subpart C); neither DES nor ionic liquids appear there as a class.
A component that is used in foods on its own does not establish that the mixture, at the residue left in the product, is acceptable.
Any residue would need its own analytical method, safety assessment and regulatory basis for the product category; see Regulatory & safety.
EUTECNIQ would treat these solvents as research tools until those questions are answered for a specific use (proposal).

eCFR, 21 CFR Part 173 Subpart C, up to date as of October 15, 2026.

Explanation, details and sources

The residual-solvent limits below come from the ICH Q3C guideline for medicines. EUTECNIQ would use them as its internal benchmark for all products (proposal). U.S. food regulations name specific extraction solvents for specific uses: for example, 21 CFR 173.270 permits hexane residue in spice oleoresins up to 25 ppm.

Hexane sits closest to soybean oil, yet its limit is 17 times tighter than ethanol's default. Ethanol–water blends slide along a straight line, so the blend ratio is a continuous design knob.

Process safety

The common extraction solvents can ignite at room temperature, so the plant is designed around their vapors

All seven organic solvents fall in OSHA Category 2, and each burns over its own range of vapor in airNIOSH data; OSHA definitions

NIOSH Pocket Guide to Chemical Hazards (n-hexane, acetone, ethyl acetate, n-heptane, methyl alcohol, isopropyl alcohol, ethyl alcohol), retrieved October 22, 2026; °F converted to °C and rounded. Categories: 29 CFR 1910.106(a)(19), up to date as of October 21, 2026 (Category 1: flash point below 23 °C, boiling point at or below 35 °C; Category 2: flash point below 23 °C, boiling point above 35 °C; Category 3: flash point 23–60 °C; Category 4: above 60 and at or below 93 °C). Carbon dioxide is not flammable.

Text version: flammability data by solvent

Flammability data for common extraction solvents and CO2
Solvent	Flash point (°C)	Flammable range in air (% by volume)	Boiling point (°C)	OSHA 1910.106 category
Hexane	−22	1.1–7.5	69	Category 2
Acetone	−18	2.5–12.8	56	Category 2
Ethyl acetate	−4	2.0–11.5	77	Category 2
Heptane	−4	1.05–6.7	98	Category 2
Methanol	11	6.0–36	64	Category 2
2-Propanol	12	2.0–12.7	83	Category 2
Ethanol	13	3.3–19	78	Category 2
Carbon dioxide	not flammable	not flammable	sublimes	not a flammable liquid; OSHA exposure limit 5,000 ppm (8-hour average); 1.53 times as dense as air

How a proposed ethanol extraction room would be zoned and controlledSchematic, not to scale

Class I, Division 1 and 2 definitions and the note that Division 1 usually includes locations with fat and oil extraction equipment using volatile flammable solvents: 29 CFR 1910.399. Equipment in classified locations must be intrinsically safe, approved for the location, or safe for it; classified areas must be documented: 29 CFR 1910.307(b)–(c). UL 1389 listing or technical report: 2025 California Fire Code § 3904.2. eCFR up to date as of October 21, 2026. Controls 3, 4 and 7 restate the fire code and OSHA; 1, 2, 5 and 6 are EUTECNIQ design intentions, and a process hazard analysis would precede each first run and each scale-up.

Text version of the rules and the proposed design

Rules that apply Regulatory requirement

Codes set thresholds and design reviews before any extraction runs

Process safety management. OSHA's PSM standard covers a process with 10,000 pounds (4,536 kg) or more on site in one location of a flammable liquid with a flash point below 37.8 °C, with exceptions that include atmospheric storage kept below the normal boiling point (29 CFR 1910.119(a)(1)(ii)).
Fire code. Chapter 39 of the 2025 California Fire Code, Processing and Extraction Facilities, requires plant-oil extraction equipment to be listed to UL 1389 or approved through a technical report by a registered design professional, reviewed by the fire code official before installation (§ 3904.2).
Exposure. OSHA's 8-hour limit for CO₂ is 5,000 ppm. CO₂ is heavier than air and collects in pits and low rooms, so CO₂ systems need monitoring and ventilation.

eCFR, 29 CFR 1910.106 and 1910.119, up to date as of October 21, 2026. 2025 California Fire Code (Title 24, Part 9), Chapter 39, § 3904, as published by GoCodeBook; confirm with the authority having jurisdiction. NIOSH Pocket Guide, carbon dioxide.

How EUTECNIQ would design for it Proposal

Small inventories, closed transfers and engineered controls

Keep flammable-solvent inventory at the point of use small, and recover solvent in closed equipment.
Remove ignition sources: electrical equipment rated for the classified area, bonding and grounding against static, and hot-work permits.
Ventilation and gas or CO₂ detection interlocked with shutdown where the fire code calls for it.
Pressure equipment (supercritical CO₂, pressurized hot water) designed with relief devices, interlocks and documented inspection.
A process hazard analysis for each new process before first run, and again before scale-up.

Explanation, details and sources

Every organic solvent below has a flash point under 23 °C and boils above 35 °C, which places it in OSHA's Category 2 flammable liquids. Its vapor can form an ignitable mixture with air at ordinary room temperature.

Process safety management. OSHA's PSM standard covers a process with 10,000 pounds (4,536 kg) or more on site in one location of a flammable liquid with a flash point below 37.8 °C, with exceptions that include atmospheric storage kept below the normal boiling point (29 CFR 1910.119(a)(1)(ii)). Regulation
Carbon dioxide. OSHA's 8-hour limit is 5,000 ppm. CO₂ is 1.53 times as dense as air and collects in pits and low rooms, so CO₂ systems need monitoring and ventilation. Regulation

No site, building or equipment has been selected; these are design intentions to be reviewed by qualified fire-protection and process-safety engineers.

EUTECNIQ's role

What EUTECNIQ proposes to do in extraction, and what it would place with partners

Screening and purification in Phase I, a pilot process in Phase II, and partners for commercial scale and fire-protection engineeringProposed division of work

Facility plan, October 2026. The same four rows are in the table below. EUTECNIQ owns no extraction equipment.

Show the table (4 rows): Proposed in-house extraction work by phase and work placed with partners

Proposed in-house extraction work by phase and work placed with partners
Work	Planned in house	Qualified external partner
Method screening	Phase I Laboratory percolation, ultrasound-assisted and supercritical CO₂ extraction; yield, purity and recovery measured against a written specification	Partner Methods outside the planned equipment list
Pilot process	Phase II Pilot extraction, evaporation and solvent recovery; fitted kinetic models and a cost-of-goods model	Partner Toll extraction at commercial scale until Phase III
Purification	Phase I Laboratory resin, partition and membrane trials	Partner Preparative chromatography at scale
Safety and compliance	Planned Process hazard analysis and solvent-residue testing	Partner Fire-protection engineering and code review

Explanation, details and sources

Founder-reported prior experience The founder reports four years (2007–2011) as a senior process engineer in industrial extraction and separation, including process design, debottlenecking and energy optimization, at a prior employer. These are not EUTECNIQ results; see Founder.

Have a botanical that needs a process, not only a yield?

Describe the material, the target constituents and the product category. EUTECNIQ is in formation and would respond with a proposed development plan, not a quotation for work it cannot yet perform.

Discuss a development program