Light Quality Impacts Vertical Growth Rate, Phytochemical Yield and Cannabinoid Production Efficiency in Cannabis sativa

2.1. Morphology

2.2. Inflorescence Yield

Significant differences (p < 0.05) were observed between replicates for average fresh inflorescence yield with 116.2 g plant⁻¹ ± 16.6 and 104.5 g plant⁻¹ ±18.3 for replicate 1 and 2 respectively. Similarly, mean dry inflorescence was significantly different (p < 0.0005) between replicate 1 and 2, with average values of 23.3 g plant⁻¹ ± 3.7 and 19.4 g plant⁻¹ ± 3.7 (p = 0.0004) for replicate 1 and 2, respectively (Supplementary Table S1). The mean inflorescence fresh and dry mass for each light treatment per replicate is summarized in Table 2. HPS resulted in the highest average fresh and dry inflorescence mass for both replicates. Light intensity was significantly different between replicates (p < 0.0001), with replicate 1 having ~7% more PPFD per plant. Light treatments had a slightly different effect when ranking inflorescence yield between replicates for light treatments rose, purple, amber, and red. Blue LED light resulted in the lowest average fresh and dry inflorescence mass in both replicates. Drying reduced inflorescence mass by approximately 80% for each light treatment, and the moisture concentration was between 12.5 ± 1.5%.

2.3. Phytochemical Concentration

Cannabis plant inflorescence was analyzed to determine the concentration of the major cannabinoids (THC, CBD, and CBG) and terpenes. Significant differences (p < 0.005) were observed for THC and CBD concentrations between replicates, but not for CBG and terpenes (Supplementary Table S1). Table 3 summarizes the effect of light treatments on total THC, CBD, and CBG concentration per replicate.

2.4. Total Phytochemical Yield

3.1. Morphology

Light treatment with a high blue fraction caused smaller plant height and vertical growth rate, as has been observed in other studies [26,27]. A high fraction of blue light, or blue-dominant light, resulted in the lowest plant height. This was expected, as dwarfing of plants grown under blue-dominant light has been observed for other greenhouse crops [44,45]. Blue light induces cryptochrome photoreceptor activation, which slows vertical plants, although not all plants experience shortening with blue light [46]. This typical blue-light-induced dwarfism in C. sativa plants under blue is consistent with the observation in other plants, where cryptochrome seems to be involved.

Amber light resulted in the greatest plant height while monochromatic red light resulted in smaller plants and a significantly lower growth rate during the flowering phase. In other plants, the stimulation of internodal stretching and subsequent plant elongation by red-light induced shade avoidance response through phytochrome activation [47]. Red-light-induced vertical growth increases have been previously reported in C. sativa [18] and plant elongation of other greenhouse crops [48,49,50]. LEDs with a higher proportion of red wavelength increased plant height in other studies [26,33], but no study investigated the effect of amber light on cannabis plants.

The molecular mechanism of amber light-induced vertical growth in C. sativa during flowering is unknown. It could be hypothesized that this occurs through different molecular pathways other than the characterized shade avoidance response/photoreceptor activation, as amber light is not believed to activate phytochrome [41]. The higher plant height observed with the amber light treatment could be caused by specific protein over-expression. Proteomic analysis on Arabidopsis thaliana showed that amber light results in a 2- to 6-fold increase in the expression of key proteins implicated in carbon and amino acid metabolism (GAD2, RBCS-1A), protein synthesis, folding, and degradation (HSC70–1), as well as cytoskeleton and cell wall (PME3) [51]. Increased expression of these proteins could cause the observed internodal stretching.

Regardless of the mechanism implied in amber-light-induced vertical growth, growers want better control of plant height. Whether to achieve better air circulation in the canopy or homogenize canopy height, amber light could be used instead of red light. Integrated plant management programs focusing on promoting canopy air flow to decrease fungal infection in the phyllosphere might consider incorporating additional amber lighting to the grow room. Known greenhouse pathogens such as Botrytis, Fusarium, and Phytium are affected by humidity management via airflow increase [52,53]. Conversely, growers wanting to increase productivity via vertical farming could benefit from blue light-induced smaller plants. Vertical farming increases facility productivity for some crops but studies on cannabis are lacking [54].

3.2. Inflorescence Mass

Inflorescence mass is a key factor in determining cannabis yield for growers. HPS had the greatest inflorescence mass, followed closely by Rose and Red light. A study conducted by Magagnini et al. [18] supports these results, reporting that C. sativa plants grown under HPS yielded more inflorescence than plants grown under LEDs. Although some similarities in the spectral compositions of HPS and amber light can be noted, the latter yielded 30% less fresh inflorescence mass. Differences between these two spectra had HPS light with a lower percentage of 595-nm light, higher peaks from 420 nm to 460 nm, and 500 nm, when compared to the Amber LED spectrum. These differences in spectral composition, particularly for 500-nm light, may be why there is a considerable difference in inflorescence mass between HPS and Amber. However, when comparing Red and Rose light treatments, supplementing with a small fraction of blue light does not seem to impact fresh inflorescence mass.

Danziger and Bernstein [26] tested different light spectra of three different chemotypes. Their study used LED with a B:R ratio of 1:1 and 1:4 versus the B:R ratios used in the present study (2:1 and 1:10). Inflorescence yield was almost doubled compared to HPS when the intermediate chemotype ‘CS12′ was grown under a LED light having a 1:4 B:R ratio. It should be mentioned that only one of the three chemotypes tested in their study had a significantly higher yield under certain LEDs when compared to HPS. Of note, the light intensity used in this study is half of what was used in their study (950 versus 400 μmol m⁻² s⁻¹). Other studies, such as the one performed by Wei et al. [33] did not normalize PPFD between treatments. The question of a potential interaction between light intensity and spectrum composition should be part of future studies. If growers want to achieve higher yield with LEDs, they should consider the need for fine-tuning the B:R ratio, light intensity, and chemotype selection.

3.3. Cannabinoids Concentration, Yield, and Profitability

In this work, plants grown under blue-dominant light (blue or purple) yielded the highest THC and CBG concentration but it did result in lower inflorescence mass when compared to the control. This study aligns with others indicating that THC concentration increases with increasing blue light [18,26,27]. It should be highlighted that whereas the Rose spectrum (B:R ratio 1:10) and HPS had similar inflorescence mass, the difference in THC percentage did cause a significant decrease in g THC plant⁻¹. At a glance, the difference between the two spectra is a 6-fold increase in light intensity in the 700–799 nm range for HPS, as well as an absence of any wavelength >800 nm in the Rose LED spectrum. Influence of the Far-Red spectrum (700–780 nm) on cannabinoid yield should be further investigated.

Unlike THC, CBD concentration did not seem to respond as much to increasing fractions of blue light. Comparable CBD percentages were found in the inflorescence of plants cultivated among blue, purple, rose, and HPS light treatments, while lower CBD percentages were observed in amber and red treatment.

Conflicting results have been reported regarding the effect of blue light on CBD biosynthesis depending on chemotypes. THC-dominant chemotype (drug-type) C. sativa seems to increase CBD biosynthesis with increasing blue light fraction [17,18,27]). Increasing the blue-light fraction has conflicting results in CBD-dominant (fiber-type) chemotypes. Some studies report no significant impact of light spectrum on CBD biosynthesis [17,34], whereas others did report a significant change in CBD concentration for plants grown under a higher blue-light fraction [18,25,33]. Studies reporting increased CBD concentration with an increase in blue-light fractions did compare treatment with different PPFD levels. The increase in CBD concentration could be solely caused by an increase in PPFD. Interestingly, light treatment in other studies impacted cannabinoid biosynthesis so much that it blurred previously described legal thresholds. For example, Westmoreland et al. [34] showed how legally defined hemp (THC < 0.03%) is transformed into legally defined drug-type cannabis (THC > 0.03%) [55,56] if grown under higher blue-light fractions. Further studies should consider chemotype-specificity when measuring light quality-driven cannabinoid biosynthesis, as CBD-dominant chemotypes seem to be less receptive to blue-light-driven increases in cannabinoid biosynthesis.

There was a consistent increase in CBG concentration as the fraction of blue light increased between the Rose, Purple, and Blue treatments. This finding agrees with other reports where LED light containing blue light resulted in higher CBG concentration than HPS [18]. As CBG is still being studied for new medical indications [57], strategies to increase its concentration are worth investigating further. Again, the interaction between light quality and chemotypes should be studied further, especially for growers working with CBG-dominant (type IV) chemotypes [58].

Cannabinoid concentration in C. sativa should always be analyzed in conjunction with biomass accumulation. In this study, HPS treatment produced the most cannabinoids (THC and CBD) on a per-plant basis. The light treatments containing blue light (blue and purple) resulted in 22 to 43 % less total THC yield per plant. It is important to note that although blue light may induce higher THC and CBD percentages in inflorescence biomass, it appears to suppress inflorescence growth, resulting in less total secondary metabolite produced per plant. This could be of concern to growers as blue light can lead to a greater reduction in overall cannabinoid yield.

Of further industrial relevance, we suggest reporting cannabinoid yield as a per-plant metric. This metric is relevant from a grower’s perspective, since it shows how much THC was produced per plant per growing cycle, rather than presenting THC concentration percentage-wise (% mass or % m/m) [2,18]. These units should facilitate the translation of scientific findings to growers using different plant densities in their facilities.

Using a novel CPE_light metric, it was possible to show how LEDs can help produce cannabinoids more efficiently. The same amount of energy was necessary to produce 1 g of THC under HPS than Rose or Purple, even when plants had significantly lower inflorescence mass or THC concentration. Because lighting is one of the main operational costs for growers, increasing light intensity in an already operational grow room could be rendered more economical if using LED lamps instead of adding extra HPS lamps.

3.4. Terpenes

Terpenes are responsible for inflorescence odor and flavor profiles [59]. Light treatments with blue-dominant spectra led to higher total terpene production, monoterpenes, and sesquiterpenes. Purple light treatment resulted in the highest total terpene concentrations (mg g⁻¹), followed by the blue light treatment. Higher concentrations of monoterpenes and sesquiterpenes were observed under these two light treatments, with a higher fraction of blue light, and this data agree with a previous study [17], in which higher concentrations of monoterpenes such as α-pinene and limonene were reported with supplemental blue, green and red light. Although there was a relatively low fraction of blue light in the HPS treatment, it induced comparable total terpene concentrations to the Blue light treatment. Although different cultivation approaches were applied, this finding regarding HPS treatment agrees with that of Namdar et al. [27]. The authors reported that a higher total terpene concentration was observed when plants were grown under LEDs with a high B:R ratio in the vegetative phase and flowered under HPS light. In this work, the same spectral treatment was applied throughout the vegetative and flowering stages, and higher total terpene concentrations were observed under both HPS and purple treatments.

3.5. Limitations (Differences between Replicates)

Significant differences in inflorescence mass (fresh and dry), total THC%, and total CBD% were observed between two temporal replicates, and plants with different origins may have affected how the plants grew and developed. The plants used to conduct replicate 1 were transported over a two-day period and arriving in a closed box, suggesting there was little air exchange for two days. Seedlings from replicate 1 were on average double the height as the ones used for replicate 2. Plants used for replicate 2 were cloned from the original lot of plants received, as the original licensed producer stopped shipping seedlings mid-way through replicate 1. Plants were given a week of standardized lighting before being placed in a treatment to get their growth rate as similar as possible. Seedling plant height might be a cause for the observed significant difference in final harvested inflorescence and phytochemical concentration.

Other environmental parameters that exhibited differences between replicates were PPFD levels (5–10% difference) and night-time temperatures (4 °C). The smaller seedlings from replicate 2 were exposed to colder nights and lower PPFD. They produced less inflorescence but were more concentrated in phytochemicals. Interestingly, higher PPFD in the first replicate should have resulted in higher cannabinoid concentration, as observed in other studies [23,60]. Future studies could look at the effect of seedling height and night-time temperature on phytochemical biosynthesis.

4.2. Light Treatments

The six light treatments used in this experiment were as follows: (1) HPS light (control; Gavita Pro 6/750e DE FLEX, Aalsmeer, Netherlands), (2) ‘Amber’ light; 595 nm monochromatic wide spectrum (VQ-GLIB600W-595, Vanq Technology, Shenzhen, China); (3) ‘Red’ light; 630 nm monochromatic narrow spectrum (VQ-GLIB600W-630, Vanq Technology, Shenzhen, China); (4) ‘Rose’ light; 430 nm and 630 nm polychromatic narrow spectrum with a 1:10 B:R ratio (VQ-GLIB600W-630/430n, Vanq Technology, Shenzhen, China); (5) ‘Purple’ light; 430 nm and 630 nm polychromatic narrow spectrum with a 2:1 B:R ratio (VQ-GLIB600W-630/430m, Vanq Technology, Shenzhen, China) and (6) ‘Blue’ light; 430 nm monochromatic narrow spectrum (VQ-GLIB600W-430, Vanq Technology, Shenzhen, China). All LED spectra emitted 600 W from four 150 W LED chips equipped with glass circular optic lenses (90° viewing angle). The HPS lamp was a larger single-cylinder bulb with an aluminum refection hood, which was powered at 750 W. All spectral composition data were confirmed with a spectroradiometer (ALP00051300010731, Asensetek, Gatineau, QC, Canada). Electrical parameters were measured with a current monitor (KW47-US, Kuman Tech, Shenzhen, China). The lighting fixture and spectra are listed in Table 5.

The average (photosynthetic photon flux density) PPFD level and photoperiod were adjusted based on plant growth stages. During the first two weeks (days 0–13, vegetation stage), plants were grown with an 18 h d⁻¹ photoperiod at 250–270 μmol m⁻² s⁻¹. At the end of the vegetative stage (day 13), the remaining nine plants were switched to inductive photoperiod (12 h d⁻¹) for 8 weeks (day 14–70) until harvest. The light uniformity in each light treatment was between 0.91 to 0.94; this was determined by the ratio of the minimum to average PPFD level as described by Balasus et al. [61]. During cultivation, plants were randomly reorganized every 3 days, which avoided inconsistent PPFD levels caused by light uniformity. In the meantime, PPFDs under each light treatment were confirmed and adjusted to the set point by adjusting the height of the lights. Once being switched to inductive photoperiod, PPFD levels were increased by ~20 μmol m⁻² s⁻¹ every week until plants stopped growing vertically (day 35, third week of flowering), then maintained at 400 μmol m⁻² s⁻¹. Stray light testing was performed with all lights on, except the treatment zone being tested. Measurements were made at the center base of the table and stray light bleeding was negligible (<2%).

4.3. Plant Materials and Cultivation

One hundred cuttings (100) of C. sativa from the ‘Babbas Erkle Cookies’ accession (WeedMD Inc., Aylmer, ON, Canada) were obtained. This accession is an intermediate chemotype (Type II; THC concentration> 0.3% and CBD concentration >0.5%) [62]. Out of the initial 100 cuttings, 72 plants uniform in size, 12 per experimental group, were transplanted and placed in a growth chamber for the first crop cycle (Replicate 1). The remaining plants were used as mother plants. Cuttings for the second crop cycle (Replicate 2) came from these mother plants. Mother plants were maintained under fluorescent light (RAZR2, Fluence, Austin, TX, USA). Cuttings for Replicate 2 were rooted using indole-3-butyric acid gel (Technaflora, Mission, BC, Canada) and rapid rooter plugs (General Hydroponics, Santa Rosa, CA, USA), with an 18 h d-1 photoperiod and a PPFD level of 125 μmol m⁻² s⁻¹ under high humidity (>90%), using a propagating tray with transparent dome cover. PPFD was determined with an LI-250A Light Meter and an LI-193 Spherical Underwater Quantum Sensor (LI-COR, Lincoln, NE, USA). Successfully propagated cuttings showing adventitious roots were transplanted into 750 mL square pots with Canna-coco coconut husk mixture (Canna, Toronto, ON, Canada), and 10 mL Myke Tree and Shrub Mycorrhizae were added to the pots (Premier Tech, Rivière-du-Loup, QC, Canada).

Plants were irrigated with the commercial nutrient solution every 2–4 days. Nutrient solutions comprised tap water (Montreal, QC, Canada), Coco A&B nutrient solutions (Canna, Toronto, ON, Canada), KH₂PO_4, and K₂SO₄. Electrical conductivity (EC) and pH were monitored with a handheld EC and pH meter (HI98130, Hanna Instruments, Woonsocket, RI, USA) and pH was corrected to 6 with an H₃PO₄ solution (HGDi Technologies, Montreal, QC, Canada) before irrigation. Representative samples were analyzed by an independent laboratory (A&L Canada, London, ON, Canada). The average electrical conductivity (mean ± SE, n = 25) was 1.9 ± 0.3 mS cm⁻¹ with nutrient element concentration of N 137.1 ± 6.4 mg L⁻¹, P 88.9 ± 15.0 mg L⁻¹, K 149.9 ± 20.2 mg L⁻¹, Ca 141.7 ± 5.9 mg L⁻¹, Mg 44.7 ± 1.8 mg L⁻¹, S 86.4 ± 16.7 mg L⁻¹, Cl 10.5 ± 0.3 mg L⁻¹, Fe 0.6 ± 0.02 mg L⁻¹, Zn 0.2 ± 0.01 mg L⁻¹, Mn 0.3 ± 0.01 mg L⁻¹, B 0.2 ± 0.01 mg L⁻¹, and Cu 0.047 ± 0.001 mg L⁻¹.

Total nitrogen was analyzed by combustion/thermal conductivity (Dumas method), and chloride by K₂SO₄ extraction with standard method 4500-Cl- G Mercuric Thiocyanate Flow Injection Analysis. Other elements were analyzed by acid digestion (ICP-OES Ref. EPA3050B/EPA6010B).

The first watering was performed by hand to compact the coco grow medium and sequential watering was provided by two submersible pumps (728305, EcoPlus, Austin, TX, USA) every 2–4 days. The amount of water provided increased during the study and tripled once plants were transplanted into their final pots. No nutrients were added to the irrigation water from week 8 until harvest at week 10. Plants were trimmed on day 35 (third week of flowering) using the following guidelines: (1) All fan leaves on the bottom two-thirds of the primary stem were removed; (2) branches that grew from the bottom one-third of the main stem were stripped of all fan leaves except for the top two fan leaves while all inflorescence leaves were kept; (3) Leaves in contact with inflorescences were removed. This technique is similar to the bottom branches and leaves removal (BBLR) method described by Danziger et al. [26].

4.4. Vertical Growth Rate

Plant vertical growth rate was recorded under each light treatment. Height was measured weekly from the base of the plant to its highest point (excluding leaves). The vertical growth rate for the flowering period was measured only for the first three weeks which is when most (>90%) of the vertical growth happens [26]. The growth rate was calculated with Equation (1):

where Δ is vertical growth rate,  is plant height at the start of a given week (cm) and +1 is plant height at the start of the following week (cm).

4.5. Plant Images

Plant images were taken one week before harvest. For this purpose, the lighting from each treatment was switched off and a flash from a digital camera was used. The camera was placed above the plant canopy, and the distance from the camera to the plant canopy was consistent (46 cm). Images were analyzed with a color histogram using ImageJ 1.48v software (Bethesda, MD, USA), to determine the effect of each light treatment on inflorescence and leaf coloration.

4.6. Plant Harvest and Phytochemical Measurement

Plants were harvested eight weeks after initiating the flowering photoperiod. Inflorescence was weighed separately and placed on a screen for drying. The inflorescence was placed in a dark room with a dehumidifier at 18 °C, to a moisture concentration of 12.5 ± 1.5%. Dried inflorescence was placed into a plastic bag for curing. Over a one-week curing period, bags were opened once a day for 10 min.

Cannabinoid and terpene analyses were performed on dried plant material from three plants in each treatment. Three separate 4-g inflorescence samples per plant were subjected to cannabinoid and terpene analyses, and each 4-g sample comprised a portion of the biggest inflorescence as well as equal parts of medium and small-sized inflorescences. Cannabinoid and terpene analyses were performed by Laboratoire PhytoChemia (Saguenay, QC, Canada).

Measured cannabinoids included THC, tetrahydrocannabinolic acid (THCA), CBD, cannabidiolic acid (CBDA), cannabigerol (CBG) and cannabigerolic acid (CBGA). Cannabinoids were analyzed based on the procedures described by De Backer [43]. Briefly, 0.25 g samples were mixed with extraction solvent (methanol:dichloromethan, 9:1). The sample was filtrated before injection in a high-performance liquid chromatograph (HPLC). The column used was Kinetex 2.5 μM C18, 150 × 4.6 mm, and the detector was a diode-array (DAD). The primary mobile phase was deionized water with formic acid (0.1%) and the secondary mobile phase was methanol:acetonitrile (75:25) with formic acid (0.1%). References used for THCA, CBDA, CBGA, THC, CBD and CBG have CAS numbers 23987-85-0; 1244-58-2; 25555-57-1;1972-08-3; 13956-29-1; 25654-31-3 respectively. Methyl octanoate at ≥99% purity has a CAS of 111-11-5.

Terpenes were analyzed by mixing 1 g of dried biomass with 5 mL of extraction solvent: pentane with octanoate methyl (0.16 mg L⁻¹). Samples were injected in Agilent 127-5012 DB-5 and 127-7012 DB-WAX columns for gas chromatography and mass spectrometry (GC-MS). Identification was made by spectral matching with available databases [63,64,65,66,67]. Instrument parameter and quantification of unique terpenes was based on guidelines by the International Organization of the Flavor Industry (IOFI) [68].

Total THC was defined as Equation (2). This equation was applied to total CBD. For CBG the value used was acidic form × 0.878

where total cannabinoid concentration, acidic form and neutral form is in percent (%). Phytochemical yield per plant was calculated following Equation (3).

where total phytochemical yield is in g plant⁻¹, total phytochemical is in % or mg g⁻¹ and dry inflorescence mass is in g. No significant difference in average phytochemical yield per plant between replicates was observed (Supplementary Table S1).

4.7. Light Associated Cannabinoid Production Efficacy

Each light treatment had its cannabinoid production efficacy (CPE_light) measured for THC and CBD. Energy used by the lamp was divided per quantity of cannabinoid produced per plant and was calculated following Equation (4).

where CPE_light is the cannabinoid production efficacy (kWh g⁻¹), P_light is lamp power (kW) with values of 0.6 kW for LED lamps and 0.75 kW for HPS, operation time is in hours (h) with a value of 924 and total cannabinoid yield in g cannabinoid.

The authors would like to acknowledge Laboratoire Phytochemia and A&L Canada Laboratories for their help in sharing their analytical methods. The authors would also like to thank Daniel Ducharme for his role as a consultant for data analysis.