By-products: not what we intended. Secondary products. We made something, and in the process of making, bonus products were born—maybe they were tossed in the dumpster. Hemp and cannabis cultivation generates by-products. In Cannabis sativa (hemp or intoxicating cannabis), this may include flower, leaf, seed, root, dust, shives, and seed meal. Recently, researchers  reviewed the phenolic compounds and terpenes that can be extracted from these by-products.
Conventional solvent extraction is typical when profiling the phenolics in cannabis by-products. This usually means maceration with ethanol, methanol, or acetone. Increasing the temperature and time may boost yield when going the conventional route. One study  employed 2-hydroxypropyl-cyclodextrin, an “eco-friendly solvent,” with results that trumped conventional solvents. The authors note that where a given cultivar is grown can drastically change the phenolic contents, and thus “results from different studies cannot be compared with each
other.” That comes down to terroir.
Ultrasound-assisted extraction (UEA) has also been used for this purpose. UAE may reduce extraction time and boost yield, but “excessive sonication” can degrade compounds. Microwave-assisted extraction, pressurized liquid extraction, and rapid solid-liquid dynamic extraction are less-explored possibilities.
With these techniques in mind, the extractor may wonder: what is it I can get out of these by-products?
In terms of phenolic compounds, the review notes 26 identified options. Flavonoids unique to cannabis include cannflavin A, B, and C. The researchers point out that cannflavins “can relieve pain up to thirty times more than aspirin.” Flavonoids in general are known for potent antioxidant activity. Cannabis houses these chemicals in varying ways. The flavonoid orientin, for example, is higher in leaves than seeds with similar amounts across male and female plants. Quercetin boasts greater presence in male flowers compared to female flowers.
The review proceeds toward cannabinoids and terpenes, more familiar molecules to cannabis extractors.
Here, hydrodistillation (HD) and steam distillation (SD) are most typical. The researchers note ideal temperatures of 110ºC and 130ºC for HD and SD, respectively. They consider HD superior as it extracts more caryophyllene and cannabidiol (due to more decarboxylation). They note that for SD, “steam does not penetrate uniformly the plant material.” Not surprisingly, monoterpenes are extracted early (0-10 minutes) and sesquiterpenes later (80 minutes), noted from a study we covered recently on Extraction Magazine.
Conventional solvents have been used for triterpenes from roots and stem bark. Supercritical carbon dioxide (CO2) may be superior to HD when extracting volatile terpenes given lower temperature. This technique can also be used to sequentially extract cannabinoids and terpenes. One study  found that low temperature (35ºC) and no co-solvent were optimal for terpene extraction with supercritical CO2.
Monoterpenes and sesquiterpenes tend to vary among flowers even on the same plant. The authors note triterpenes are available from roots, stems, and leaves. These include friedelin, the most abundant triterpene in roots, with “antiulcer, antidiabetic and cytotoxic effects.”
The takeaway message is that there are many powerful compounds throughout the entire cannabis plant. Extraction—particularly “intensification and fractionation techniques”—allows producers to harness these molecules rather than dumping them. 
1- Isidore E, et al. Extraction of phenolic compounds and terpenes from Cannabis sativa by-products: from conventional to intensified processes. Antioxidants. 2021;10:942. [Impact Factor: 6.312; Times Cited: n/a]
2- Mourtzinos I, et al. A green extraction process to recover polyphenols from byproducts of hemp oil processing. Recycling. 2018;3(15). [Impact Factor: n/a; Times Cited: 8 (Semantic Scholar)]
3- Omar J, et al. Optimisation and characterisation of marihuana extracts obtained by supercritical fluid extraction and focused ultrasound extraction and retention time locking GC-MS. J Sep Sci. 2013;36(8):1397–1404. [Impact Factor: 3.645; Times Cited: 54 (Semantic Scholar)]