The study was carried out in a walk-in growth chamber (3.36 × 2.86 × 2.78 m) designed to ensure precise environmental control. The chamber features sandwich panel walls with 80 mm polyurethane insulation, covered with pre-painted galvanized steel to guarantee thermal stability [K = 0.23 kcal/(h·m²·°C)]. The environmental parameters [temperature, relative humidity (RH), CO₂ concentration, and lighting] were managed via an integrated Programmable Logic Controller -based electronic control unit with touchscreen interface and remote web access. The system allowed the creation of up to eight programmable daily time slots for each parameter, with real-time and historical data monitoring. Lighting was provided by dimmable LED lamps (3.2–3.4 µmol·J⁻¹), with customized spectra for each shelf level and a total photon flux of 900–1100 µmol·s⁻¹ for the ceiling-mounted units (Aquila series, Ambralight, Italy). The growth chamber hosted three movable aluminum racks (each 60 × 189 × 200 cm) equipped with hydroponic aluminum trays and independent irrigation systems, including stainless-steel centrifugal pumps, mesh filters, and nutrient solution tanks (200 L each). Temperature and RH were regulated by an air-cooled condensing unit and a dual-air evaporator with stainless-steel heating resistors. Air recirculation, de-stratification, and automated humidification ensured homogeneity throughout the growing space. The CO₂ injection system included an electrovalve, pressure reducer, and monitoring sensors. All climatic and technical parameters were continuously monitored and recorded, ensuring standardized and reproducible growth conditions. Two cultivars (cv) of crisphead lettuce (Lactuca sativa L. var. crispa) - ‘Falstaff’ (green) and ‘Copacabana’ (red) (Isi Sementi) - were grown in two crop cycles.

The seeds were sown in rockwool plugs (Grodan plantop plug NG2.0, Roermond, The Netherlands) in polystyrene trays (600 × 400 × 50 mm) with 240 holes (1,000 seedlings·m^-2). At the stage of two true leaves, the seedlings were transplanted into 0.5 L pots filled with a 1:1 (v/v) mixture of fine and medium-grade perlite (Perlite Italiana, Italy). The fine-grade perlite had a particle size of 0.5–1.5 mm, while the medium-grade perlite ranged from 1.5–3.0 mm. On each shelf, 16 pots (eight for each of the two lettuce cultivars) were placed.

The pots were placed at 15 cm between the rows and 20 cm within the row, with a final shelf density of 33.3 plants·m^-² (Figure 1). Fertigation was managed in an ebb and flow system (closed cycle), with a total of four irrigations per day, with a timing of 10 minutes (pump flow rate 20 L·min⁻¹).

The nutrient solution (NS) had the following concentrations of macro-nutrients (expressed in mmol·L⁻¹): 5.42 N-NO₃, 0.29 N-NH₄, 2.71 K, 0.65 P, 0.74 Mg, 1.75 Ca, 1.00 S. While the micronutrients composition, expressed as µmol·L^-1, was: 20 Fe, 5 Mn, 2.5 Zn, 25 B, 0.75 Cu, and 0.5 Mo. These concentrations represent calculated target values, derived from the stoichiometric formulation of the nutrient solution based on the fertilizers used. During the crop cycles, the electrical conductivity (EC) and pH of the NS were maintained within a range of 1.2-2.0 dS·m⁻¹ and 5.5-6.5, respectively. During the germination phase, the growth chamber was kept with a temperature of 20 °C, RH of 80%. The CO₂ concentration inside the growth chamber was at ambient atmospheric levels (≈400 µmol·mol⁻¹). Nevertheless, after transplanting, air temperature and relative humidity were set to 24 °C and 65%, respectively, as these conditions are widely reported as optimal for lettuce cultivation under controlled-environment agriculture. This temperature range supports high photosynthetic activity and biomass accumulation while avoiding heat stress, whereas a relative humidity around 60–70% ensures adequate transpiration and stomatal conductance without increasing the risk of physiological disorders or microbial development. These settings are commonly adopted in growth chamber and vertical farming studies to standardize environmental conditions and minimize confounding effects unrelated to lighting treatments (Pennisi et al., 2020).

LED lamps were used, with a “full light spectrum” (Figure 2), a photosynthetic photon flux (PPF) of 80–200 µmol s⁻¹ and efficiency ranging from 3.0 to 3.4 µmol J⁻¹. Each unit (length:120 cm, weight:680 g) operates at 25–60 W and is rated IP65, with integrated protections against short circuit, overvoltage, and overheating. The emission spectrum of the LED lamps was measured in situ using a LI-180 Spectroradiometer (LI-COR Biosciences, Lincoln, NE, USA). PPFD and spectral distribution were measured at a fixed reference height corresponding to the average canopy level of lettuce plants during the main vegetative phase, rather than directly at the tray surface, using a portable spectroradiometer (LI-180, LI-COR Biosciences, Lincoln, NE, USA), operating in the 380–780 nm range, with a spectral resolution of 1 nm and a cosine-corrected sensor. PPFD measurements were collected at multiple points across each shelf and are reported as mean ± standard deviation. The instrument provides PPFD measurements with an accuracy of ± 5% and was factory calibrated prior to use. PPFD measurements were performed at leaf level to characterize the actual light quality received by the plants under each lighting treatment, following a grid-based approach, with multiple points evenly distributed across each shelf to account for spatial variability. Reported PPFD values represent the mean ± standard deviation of these measurements. Spectral data are expressed as PPFD (μmol·m⁻²·s⁻¹; Figure 2). Each shelf (or light treatment) was isolated using blackout curtains to prevent light contamination between treatments. Photoperiod and PPFD treatments were set as follows:

All treatments were set to achieve the same DLI value of 14.4 ± 0.4 mol·m⁻²·day⁻¹.

At harvest time (26 days after transplant; DAT) Destructive measurements were performed on four plants per experimental unit to evaluate key morphological traits. Specifically, the length and width of the largest fully expanded leaf, leaf area (LA), shoot fresh weight, and shoot dry weight were recorded. Shoot fresh weight was measured immediately after harvest using a precision electronic balance, while shoot dry weight was obtained by drying leaves in a ventilated oven at 80 °C until a constant weight was reached. LA was measured using a LI-3100C Area Meter (LI-COR Biosciences, Lincoln, NE, USA). Leaves were manually arranged flat on the illuminated reading surface and scanned with a high-resolution optical system capable of accurately detecting the two-dimensional profile of the leaf blade. This method enabled rapid and reproducible LA estimation, minimizing operator-related bias even in leaves with irregular margins. Colorimetric measurements were conducted with a CM-700d handheld spectrophotometer (Konica Minolta, Tokyo, Japan), featuring a wavelength range of 400–700 nm, spectral interval of 10 nm, and repeatability of ΔE*ab ≤ 0.04 under standard conditions and operating in diffuse reflectance mode (D65 illumination, 10° observer angle), following the CIE guidelines. The instrument was calibrated with a white reference tile, and the aD65*, bD65*, L*D65, and Chroma values of each leaf were recorded based on the CIE Lab* system.

The day before harvest operation, 25 DAT, chlorophyll fluorescence was measured using a pulse-amplitude-modulated fluorometer (PAM-2500, Heinz Walz GmbH, Germany), equipped with a blue measuring light (λ ≈ 470 nm) and saturating pulses exceeding 6,000 µmol photons·m⁻²·s⁻¹. The device allows detection of fluorescence signals with high temporal resolution (up to 10 µs), ensuring accurate determination of PSII photochemical parameters. The measurements were made at 25 DAT, including: i) Fv/Fm: the maximum quantum efficiency of Photosystem II (PSII), calculated after dark adaptation (it represents the potential efficiency of PSII photochemistry when all reaction centers are open; healthy, non-stressed plants typically show values around 0.83; a lower Fv/Fm indicates stress or photoinhibition; ii) ETR (Electron Transport Rate): the rate of electron flow through PSII during photosynthesis (it is an indicator of photosynthetic); iii) Y(II): the effective quantum yield of PSII in light-adapted conditions (it reflects the proportion of light absorbed by chlorophyll that is used in photochemistry under real conditions); iv) NPQ (Non-Photochemical Quenching): a measure of non-photochemical energy dissipation (it reflects how much excess light energy is being dissipated as heat to protect the photosynthetic apparatus, especially under high light stress).

Before measurements, leaves were kept in the dark for 30 minutes using a dark clip, allowing a full re-open of all PSII reaction centers. This dark adaptation ensures accurate detection of i) Fo: the minimal fluorescence when all PSII reaction centers are open (no photochemical activity); ii) Fm: the maximal fluorescence when all PSII centers are closed after a saturating light pulse.

In addition to these steady-state parameters, at 25 DAT, fast chlorophyll fluorescence kinetics were recorded to analyze the rapid changes of fluorescence occurring in milliseconds after illumination. Rapid light curves were performed with the PAM-2500 with gradually increasing irradiance in seven steps. For each step, the irradiance was 40, 140, 270, 500, 870, 1400, and 2000 μmol·m^-2·s^-1, and the fluorescence signal was recorded. Data were recorded using the software by PamWin-4 V4.02v. In addition, the maximum quantum yield for whole chain electron transport, at low light intensities (Alpha), the maximum electron transport capacity, at light saturation (ETRmax) and the light saturation coefficient (Ik), was taken. Fluorescence was simultaneously detected at wavelengths above 700 nm and below 710 nm using dual detectors to differentiate between PSII and PSI signals. This dual-emission analysis enables a more complete understanding of the interplay between the two photosystems and how plants respond to light stress or physiological changes.

Furthermore, gas exchange parameters, including net assimilation rate (An), stomatal conductance (gs), and transpiration rate (E), were performed at 25 DAT, using a portable photosynthesis system (LI-6400XT, LI-COR Biosciences, Lincoln, NE, USA), equipped with a 2 × 3 cm clear-top leaf chamber and an infrared gas analyzer with a CO₂ measurement accuracy of ±1 µmol·mol⁻¹ and H₂O accuracy of ±0.1 mmol·mol⁻¹. The system was configured with a CO₂ concentration of 400 ppm, a leaf chamber temperature of 25 °C, and an airflow rate of 500 μmol·s⁻¹, ensuring stable and controlled environmental conditions. For the plant physiology parameters, two plants per experimental unit were measured, and one representative, fully expanded leaf was used per plant.

LED-related electricity consumption (considering different lighting conditions as described in Figure 1) was monitored using energy meters. In the same way, the consumption of other system components (HVAC system and cultivation facilities, e.g., water pumps and fertigation units) was monitored for the entire growth period. The Energy use efficiency (EUE), expressed as g·kWh^-1, was calculated as follow (Equation 1); Plant fresh weight (PFW) was expressed as g·plant⁻¹ and measured at harvest; plant density is 33.3 plants·m^-², and LED power consumption is the electricity consumption per square meter of surface of artificial light treatment, obtained by multiplying the power consumption of the LEDs over their operating time (kWh).):

The calculation of LEDs electrical energy cost, expressed as €·m^-2, was made by multiplying the hourly energy consumption per square meter of surface of LEDs system for each treatment by the corresponding hourly tariff rates, as obtained from the Italian electricity market data (https://www.mercatoelettrico.org/It/Statistiche/ME/DatiSintesi.aspx - accessed on 07/07/2025) Equation 2:

Finally, light use efficiency (LUE) was evaluated (Equation 2). It describes how effectively the plants converted light into biomass and indicates how much plant growth results from the light absorbed. It is expressed as g · mol^-1, and it was calculated with the following formula (2):

where the DLI is obtained by multiplying the PPFD of the LEDs by the duration of their operation and converting the result into moles of photons per square meter per day. These parameters provided quantitative benchmarks for optimizing energy input, guiding the adoption of energy-saving technologies and light management strategies to enhance sustainability and cost-effectiveness in indoor cultivation.

To investigate the effects of light treatments on the biochemical content of lettuce, biochemical analyses were performed on freeze-dried leaf samples. Total chlorophyll (a + b) and carotenoids were quantified by spectrophotometric analysis in acetone extracts, according to Lichtenthaler and Buschmann (2001). Absorbance was recorded at 661.6, 644.8, and 470 nm, and pigment concentrations were calculated using established equations. Total phenolic content (TPC) and antioxidant activity (ABTS assay) were determined following Admane et al. (2023) and Tarantino et al. (2020), using ethanolic extracts of freeze-dried leaves. TPC was expressed as mg gallic acid equivalents (GAE) per g of dry weight, while antioxidant capacity was evaluated by ABTS (2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) assay and expressed as µmol Trolox equivalents (TE) per g of dry weight. Total anthocyanin content (TAC) was measured only in GCII by UV–vis spectrophotometry, using extracts obtained from the entire aboveground biomass of each plant, following extraction with methanol–water–formic acid (50:44:6, v/v/v). Absorbance at 535 nm was used to quantify TAC, expressed as mg malvidin-3-O-glucoside per g of dry weight. Absorbance measurements were performed using a UV–Vis spectrophotometer with a wavelength accuracy of ±1 nm and photometric accuracy of ±0.005 absorbance units, ensuring reliable quantification of pigments and antioxidant-related compounds.

Experimental design was a split-plot scheme, with light treatments representing the main plots and cultivars the sub-plots (Figure 1). Statistical analysis was conducted using R software (version 4.4.2. - R Core Team (2024)) within the RStudio environment (RStudio for Windows, Version 2025.05.0 + 496, Posit Software, PBC, Boston, MA, USA) (Posit Team 2025). To assess differences between experimental treatments, a two-way analysis of variance (ANOVA) was applied, after checking the assumptions of normality and homoscedasticity by the Shapiro-Wilk and Levene test, respectively. In the presence of significant effects (p< 0.05), multiple comparisons between the means, Tukey HSD post-hoc test was implemented.

Despite having conducted two experimental trials, the results presented here referred only to the second trial, which was performed in the growth chamber under strictly controlled and consistent environmental and cultivation conditions. This approach ensures data reliability and minimizes variability not caused by experimental factors.

The growth cycle of the two lettuces cultivated in the growth chamber with three light treatments (12 L:12 D, 16 L:8 D, and 24 L:0 D) was 26 DAT. Light treatments did not influence plant morphology: on average, number, length and width of leaves per plant and plant fresh weight (PFW) were 14.6, 161.4 mm, 106.4 mm and 39.5 g·plant^-1, respectively (Table 1). Independently to light treatments, plants of green lettuce had 2.6 leaves more than red ones, and PFW was 18% more in green than red lettuce (Table 1). Differently, the leaves of red lettuce were 43.2% longer compared to those of green one (Table 1).

Furthermore, the treatments did not influence leaf area, specific leaf area and yield (p = 0.08) with the average values of 805 cm², 115.2 cm²·g^-1 and 1.3 kg·m^-2 respectively (Table 2). Between the cultivars, red lettuce showed 9% more dry matter content than green one but green lettuce produced 222 g·m^-2 more than red one (Table 2).

Chlorophyll fluorescence parameters were not affected by light treatments or cultivars, with overall mean values of 0.45 for Y(II), 0.95 for NPQ, 50.0 µmol m⁻² s⁻¹ for ETR, and 0.78 for Fv/Fm (Supplementary Table 1).

Consistently with chlorophyll fluorescence parameters, no effects of photoperiod or cultivar were observed for the parameters derived from the light response curves. Overall mean values were 62.3 µmol·m⁻²·s⁻¹ for ETRmax, 0.38 for α, and 170.5 µmol photons·m⁻²·s⁻¹ for Ik (Supplementary Table 2).

Net assimilation rate (An) was not affected by light treatments and cultivars, and its average value was 37.9 µmol CO₂·m^-2·s^-1 (Table 3). Differently, gs was 0.20 mol H₂O·m^-2·s ^-1 more for green than for red lettuce (Table 3). Similarly, E was not affected by light treatments, but it was 1.5 times more for green lettuce than for red ones (Table 3).

During the growth cycle, the average LEDs electrical consumption was 130 kWh·m^-2, independently to light treatments (Table 4). But the cost of LEDs application was 16.5% lower for the continuous light (24 L:0 D) compared with the other treatments (Table 4). Consequently, the best performances in terms of EUE and LUE were found for the lettuces grown under continuous light treatment, with EUE and LUE respectively 21% and 12% better than for the lettuces grown under 16 L:8 D and 12 L:12 D (Table 4). Furthermore, green lettuce showed 1.7 g·kWh^-1 (EUE) and 0.6 g·mol^-1 (LUE) more than red lettuce (Table 4).

Light treatments did not influence the colorimetric parameters of green (Supplementary Table 3) and red lettuce (Supplementary Table 4). On average, among light treatment: aD65* was -13.3 and 2.23, respectively for green and red lettuce (Supplementary Tables S3, S4); bD65* was 19.33 and 0.97, respectively for green and red lettuce (Supplementary Tables S3, S4); L*D65* was 42.07 and 28.47, respectively for green and red lettuce (Supplementary Tables S3, S4); Chroma was 23.47 and 2.97, respectively for green and red lettuce (Supplementary Tables S3, S4).

Similarly to color parameters, light treatments did not affect chlorophyll content or antioxidant capacity (ABTS) of lettuce (Table 5). However, differences between cultivar were observed: green lettuce exhibited a 22% higher chlorophyll content, whereas red lettuce showed ABTS values approximately 2.38-fold higher (Table 5). Carotenoid content was influenced by cultivar x photoperiod interaction. The highest carotenoid concentration was observed in red lettuce grown under 12 L:12 D (Figure 3a). Red lettuce grown under 16 L:8 D and 24 L:0 D still exhibited carotenoid contents 21% higher than those observed in green lettuce grown under 12 L:12 D (Figure 3a). On average, total phenolic content (TPC) was 11% higher under the 16 L:8 D photoperiod compared with other light treatments and was 2.65-fold higher in red lettuce compared with green lettuce (Table 5). Anthocyanins were not detected in green lettuce, except for a negligible amount in plants grown under 12 L:12 D (Figure 3b). In contrast, substantially elevated anthocyanin levels were measured in red lettuce. Red lettuce grown under 12 L:12 D showed anthocyanin contents 13% and 22% higher than those observed under 16 L:8 D and 24 L:0 D, respectively (Figure 3b).