Editor’s introduction: In 2018 Dr. John MacKay wrote a short series of posts on supercritical CO2 extraction. A chemist who spent 34 years at Waters Corporation, he is known in the botanical industry as an expert in sub- and supercritical fluid extraction and chromatography. This page combines three of those posts in his own words: what a supercritical fluid is and why CO2, how the extraction works, and the process parameters that decide the result. A final section looks at what research since 2018 adds, with numbers where he was qualitative. For how CO2 compares with the other methods, see our complete guide to cannabis extraction methods.
Extraction can be simply defined as the removal of soluble material from an insoluble residue, either liquid or solid, by treatment with a liquid solvent. The extraction process itself is well defined, documented and relatively straightforward. In fact, plant extraction has been around for centuries and methods have been developed to “remove” and separate certain key elements from the plant. However, with cannabis, this “simple” process can become quite complex indeed.
In my role as a cannabis processing consultant, I focus on using CO2 as the solvent in a supercritical fluid extraction process.
What Is a Supercritical Fluid?
A supercritical fluid is any substance at a temperature and pressure above its critical point, where distinct liquid and gas phases do not exist. It can diffuse through solids like a gas, and dissolve materials like a liquid. In addition, close to the critical point, small changes in pressure or temperature result in large changes in density, allowing many properties of a supercritical fluid to be “fine-tuned”. Carbon dioxide and water are the most commonly used supercritical fluids, while ethane, propane, ethylene, propylene, methanol, ethanol and acetone possess supercritical properties as well.
Why CO2?
Ideally, an extraction method should be rapid, simple and inexpensive; should give quantitative recovery of analytes without loss or degradation; should yield a solution of the analyte that is sufficiently concentrated to permit the final measurement to be made without the need for concentration; and should generate little or no laboratory wastes that have to be disposed of. Carbon dioxide in supercritical mode fits the requirements nicely. CO2 is easy to remove simply by reducing the pressure, leaving almost no trace. It is environmentally benign — non-toxic, non-flammable, odorless, tasteless, inert, and inexpensive. CO2 has been commonly used in decaffeination processes and more recently shown to be an effective supercritical solvent for extracting botanicals.
Finding the sweet spot (aka the optimum extraction method) that will deliver the highest yield of pure compound with the most efficient extraction process, whichever compound you are focusing on, is the holy grail. Most importantly, the extraction process must be performed correctly to produce a safe product for human use, and the results should be confirmed independently with proper analytical testing.
How a Supercritical Extraction Works
Supercritical Fluid Extraction (SFE) is the process of separating one component (the extractant) from another (the matrix) using supercritical fluids as the extracting solvent. It is considered by many to be the most efficient mechanism to extract botanical components. SFE can be used as a sample preparation step prior to further analysis by other techniques, such as High Performance Liquid Chromatography, or on a larger scale to either strip unwanted material from a product or collect a desired product. With cannabis, for example, the goal is to collect CBD, THC, terpenes and other compounds while stripping away unwanted residues, like pesticides, chlorophyll and waxes, and leaving behind the biomass.
In an earlier blog post, I discussed the characteristics of a supercritical fluid and how they can diffuse through solids like a gas and dissolve materials like a liquid. I also described how supercritical fluids are capable of breaking down structures where they can be separated or fractionated. And, I also noted that CO2 is the ideal solvent to be used in a supercritical fluid extraction. Now let’s take a closer look at the equipment required to perform a supercritical fluid extraction.
A typical SFE system includes a pump for the CO2, a pressure cell to contain the sample, a means of maintaining pressure in the system, and a collection vessel(s). The first step in the SFE process is to take gaseous CO2 and run it through a chamber where it will be subjected to low temperatures and enough pressure to cause the gas to change into a fluid. Within this state, CO2 exhibits special properties that when reheated and pressurized become supercritical. In this state, the supercritical fluid passes through a chamber containing the raw cannabis material. Because of its unique properties, this supercritical fluid can pass through the raw cannabis very easily while gently dissolving the membrane of the trichomes to capture their many active compounds.
Next, the extraction process occurs when the compound-enriched solvent passes into another pressurized separation vessel, only this time the pressures and temperatures will fluctuate in order to fractionate the compounds such as cannabinoids and terpenes away. The dissolved material is swept from the extraction cell into a separator (or a series of in-line separators) at lower pressure, and the extracted material settles out. The CO2 extraction provides the ability to separate individual compounds as the extraction is taking place. Different molecular weight compounds drop out at different temperatures and pressures which enables the fractionation to take place in different separators. The CO2 can then be cooled, recompressed and recycled, or discharged to atmosphere.
SFE Advantages
A research review from the International Journal of Chemical Sciences written by G.N. Sapkale et al.1, provides an excellent overview of the Supercritical Extraction Process. The review outlines in detail the many advantages of SFE:
Environmental improvement and reduced product contamination: SFE is an alternative to liquid extraction using solvents such as hexane or dichloromethane. In conventional liquid-liquid extractions, there will always be some residual solvent left in the extract and matrix, and there can be some level of environmental contamination from their use. In contrast, carbon dioxide is easy to remove simply by reducing the pressure, leaving almost no trace, and it is also environmentally benign.
Selectivity: The solvent strength of a supercritical fluid can be varied by changes in the pressure and temperature which enables the process to be fine-tuned to collect compounds of interest. We’ll talk more about this in future posts.
Speed: Extractions can be completed in just 10 to 60 minutes.
Purity: A supercritical fluid can separate from analytes by simply releasing pressure. Supercritical fluids are cheap, inert and nontoxic. Thus, they are readily disposed of after an extraction is completed by allowing them to evaporate into the atmosphere.
SFE Equipment Considerations
There are many considerations regarding the configuration of SFE equipment itself. What are the most desirable mechanical properties of the extractor — vessel configuration and size, tubing size, configuration of piping, valves, heat exchangers and other components of the extractor. For example, the extraction system can incorporate one, two, three or more extraction vessels (what is the optimal size of the vessel?), multiple collection vessels, and other mechanical options that most efficiently accomplish the extraction based on the composition of the materials.
The advantages of SFE are clear and well documented over the years. However, the equipment is expensive and there is some level of expertise required to perform an extraction to achieve the highest efficiency. An ROI analysis will help determine whether it is more cost effective to contract an outside lab to perform the extractions or for you to make the investment in equipment, disposables and operator training.
Process Parameters
Tunable selectivity
Supercritical fluid extraction mechanics are derived from the pressurization of a fluid under subcritical or supercritical conditions, applying heat, then moving the fluid into an extraction vessel where the desired cannabis constituents are separated, isolated and collected. The separation mechanics are controlled by changing temperature and pressure, as each compound reacts differently and has a different degree of solubility to the operating conditions. Solubility, density and processing time are essential factors in the recovery of a specific compound.
One of the key aspects of using pure supercritical CO2 or CF3CH2F as the extraction solvents is their unique properties in the supercritical phase, providing a level of tunability that can be achieved more easily than other solvent choices. Supercritical CO2 and CF3CH2F have been proven to be inexpensive, recyclable and environmentally safe industrial solvents that have been successful in the extraction processes used for coffee, vanilla, tea, omega-3 oils, fragrances, and others. By altering the pressure and temperature conditions during the extraction, the operator can change the solvent characteristics of the CO2 and/or CF3CH2F providing the ability to extract only the compounds of interest.
When I talk about tunability, I like to use an analogy of an orchestra. The conductor (method developer) can organize the musicians (cannabis components) and ask them to leave the room (extraction) in sections upon request. So, for example, if ethanol, which has less tunability at room temperature, is used as the extraction solvent, the conductor could only ask the musicians to leave the room altogether. With CO2, however, the conductor could ask the string section to leave the room, then the percussion, then the brass section and so on. Upon leaving, the conductor could even ask them to overlap a bit providing the ability to test different interactions between the musicians (components), with the ultimate goal of getting the entire orchestra to play in sync.
The conductor has to also assess the unique attributes of each group. For cannabis extraction, those important attributes are:
Moisture
Impact of moisture – too high a moisture level (higher than 8%) will result in CO2 acting as a co-solvent and removing polar compounds, or if too little, will result in the compounds being less amenable to extraction. Water plus CO2 will result in a lower pH due to carbonic acid. CF3CH2F will not have this effect.
Particle size
Particle size – the grinding process should result in controlled, reproducible particle sizes to ensure repeatable methodology. Too small particles will often result in channeling and clumping of the material and too large will make the interstitial distance too far apart, and the CO2 would have to travel too far and reduce the total amount of extracted material and at too slow a rate.
Mass balance
Mass balancing – need an accurate determination of the cannabinoid and terpene content of the starting raw feedstock material, the CO2 concentrate and what was left behind, as this material could be used in other products, like cosmetics.
With all of this knowledge on hand, the conductor can keep his orchestra in sync and deliver the most pleasing sound. And, the extraction method developer can provide the highest quality end product.
What Has Changed Since 2018
This section was added by the editors in 2026.
The conditions
CO2 becomes supercritical above its critical temperature of 31.1 °C and critical pressure of 7.4 MPa (about 74 bar).[2] At a constant temperature, raising the pressure increases the fluid’s density, which changes what it dissolves. That is the tunability MacKay describes.[2] A 2023 review of cannabis studies put the operating conditions used for cannabinoids and terpenoids in the range of 10 to 38 MPa (about 100 to 380 bar) and 35 to 80 °C.[2]
A note on CF3CH2F
The second solvent MacKay names, CF3CH2F, is the refrigerant HFC-134a. Calling it environmentally safe needs a qualifier: the US Environmental Protection Agency describes HFC-134a as a potent greenhouse gas with a global warming potential 1,430 times that of CO2.[6]
Terpene runs and cannabinoid runs
Studies since 2018 put numbers on MacKay’s orchestra. In a 2022 optimization study on industrial hemp flowers, lower pressure favored monoterpene-rich extracts, while higher pressure favored sesquiterpenes and CBD. Temperatures close to 60 °C emerged as optimal across the response combinations modeled by the researchers, although they cautioned that optimum conditions are specific to the starting material and process.[3] The modeled optima, which also weighed total yield and the extracts’ antioxidant and antibacterial activity, were about 131 bar for a monoterpene-rich extract, about 285 bar for CBD and about 320 bar for sesquiterpenes.[3] Other work follows the same pattern: a terpene study cited in the 2023 review extracted hemp volatiles at 10 and 14 MPa (100 and 140 bar) and 40 °C,[2] while a 2017 cannabinoid study ran at 17 to 34 MPa (170 to 340 bar) and 55 °C.[5]
One run cannot capture everything at its best, and a CO2 extract’s terpene profile can end up quite different from the flower’s. Our report on a flower-versus-CO2-extract comparison describes monoterpenes lost in extraction.
Separators, waxes, chlorophyll and pesticides
Fractionation can drop some unwanted material early. In one hemp study reviewed in 2023, the first separator ran at 7 MPa and 25 °C and the second at 5 MPa and 15 °C; cuticular waxes precipitated in the first separator because they were less soluble in supercritical CO2 than the volatile compounds, which were recovered almost entirely in the second. Adding a co-solvent such as ethanol, however, gives extracts with more waxes and chlorophyll, which then need more clean-up.[2] In practice, CO2 extraction commonly produces a crude oil that still needs refinement, including winterization to remove waxes and lipids, as our extraction guide explains.
Pesticides deserve a firmer caution. Extraction can concentrate pesticide residues from the flower rather than strip them away, to an extent that depends on the pesticide and the process; see our article on how pesticides are removed from cannabis extracts. In regulated markets, a CO2 extract must still meet the applicable pesticide-testing requirements; extraction is not a substitute for contaminant testing. Requirements differ by jurisdiction.
Moisture and particle size
MacKay’s 8% moisture ceiling is a practitioner’s rule of thumb. The 2023 review states the direction rather than a threshold: extraction yield increases as the raw material’s moisture falls, because water reduces contact between solvent and solute, so material with a high moisture content may need drying before extraction.[2] The review also supports his warning on grinding: smaller particles increase surface area and mass transfer, but very small particles can compact the bed and hinder it.[2]
Speed at larger scale
The 10-to-60-minute figure comes from the Sapkale review’s general description of SFE and should not be treated as a typical cannabis-production cycle.[1] At larger scale, runs are longer. An Australian process-development study that extracted medicinal cannabis at one kilogram per batch tested runs of 240 to 600 minutes; the highest yield came from the 600-minute run at the highest flow rate and pressure, and CO2 flow rate had the largest effect on CBD recovery.[4]
Cost and residues
Two of MacKay’s points have held up. A 2023 review still lists the absence of solvent residue as a defining feature of supercritical CO2 extraction, and still names the significant capital cost of SFE plants as a drawback.[2] The residue advantage applies to the primary extraction; refinement steps afterwards may use ethanol. For a direct comparison with butane for edibles, including cost, see BHO or CO2 for edibles.
References
1. SAPKALE, G.N., PATIL, S.M., SURWASE U.S., P. K. BHATBHAGE, P.K., SUPERCRITICAL FLUID EXTRACTION, Department of Pharmacognosy, ASPM,S K. T. Patil College of Pharmacy, Siddharth Nagar, Barshi Road, OSMANABAD – 413501 (M.S.) INDIA Int. J. Chem. Sci.: 8(2), 2010, 729-743
Added by the editors:
2. de Aguiar AC, Vardanega R, Viganó J, Silva EK. Supercritical carbon dioxide technology for recovering valuable phytochemicals from Cannabis sativa L. and valorization of its biomass for food applications. Molecules. 2023;28(9):3849.
3. Jokić S, Jerković I, Pavić V, Aladić K, Molnar M, Kovač MJ, Vladimir-Knežević S. Terpenes and cannabinoids in supercritical CO2 extracts of industrial hemp inflorescences: optimization of extraction, antiradical and antibacterial activity. Pharmaceuticals. 2022;15(9):1117.
4. Rochfort S, Isbel A, Ezernieks V, Elkins A, Vincent D, Deseo M, Spangenberg G. Utilisation of design of experiments approach to optimise supercritical fluid extraction of medicinal cannabis. Scientific Reports. 2020;10:9124.
5. Rovetto LJ, Aieta NV. Supercritical carbon dioxide extraction of cannabinoids from Cannabis sativa L. Journal of Supercritical Fluids. 2017;129:16–27.
6. US Environmental Protection Agency. Refrigerant transition & environmental impacts. Last updated November 18, 2020 (archived).
Hero image: AIDEPCUL, “Osma tarozi” / Wikimedia Commons, CC BY 4.0, resized.
Editor’s note: This page combines three posts by John MacKay, Ph.D., first published on the blog of his extraction consulting firm, Synergistic Technologies & Associates, in September and October 2018, and on Terpenes and Testing on October 5, 2018: “Cannabis Extractions Can Be a Bumpy Ride”, “Supercritical Fluid Extraction – More Pros Than Cons” and “Tuning the Orchestra in Cannabis Extraction”. They were combined here in October 2026. References to earlier and future posts are his and date from when the pieces ran separately.
Last updated October 2026 · Originally published October 5, 2018.
