In this study, 4-week-old rice plants (Oryza sativa cv. Koshihikari) were used because laser irradiation has been reported to alleviate salt stress in rice (Cheng et al., 2025), implying a potentially high physiological sensitivity to laser light. The seeds were first sterilized completely in 70% ethanol for 30 s, followed by an additional sterilization step with 50% sodium hypochlorite for 30 min. After sterilization, the seeds were rinsed eight times with distilled water to remove residual sterilizing agents. The sterilized seeds were then placed on filter paper moistened with distilled water in 9-cm Petri dishes. The dishes were kept in an incubator (FCI-280, AS ONE Co., Tokyo, Japan) in complete darkness at 30°C for three days to promote germination. On the fourth day, LED illumination was initiated at a light intensity of 100 μmol m⁻² s⁻¹ and maintained for two days. Subsequently, individual seedlings were transplanted into urethane blocks (2.2 cm × 2.2 cm × 2.5 cm) at a density of 20 seedlings per 100 cm². The urethane blocks were soaked in hydroponic nutrient solution (Otsuka-A & B, Inochio Holdings Inc., Aichi, Japan) to prevent desiccation, and the seedlings were cultivated in a growth chamber (NK Systems, Nippon Medical & Chemical Instruments Co., Ltd., Osaka, Japan) until the fourth week after sowing. The growth chamber, equipped with cool-white fluorescent lamps providing broad-spectrum white light, was maintained at 70% relative humidity (RH) with a photosynthetically active photon flux density (PPFD) of 300 µmol m⁻² s⁻¹, a CO₂ concentration of 400 µmol mol⁻¹ and 26°C in a 14 h light period. The nutrient solution (EC 1.2 dS m⁻¹) was replaced once a week.

To investigate the effects of LDs and LEDs on photosynthesis, we introduced a novel LD-based light source (ALT-9H59, ALT Inc., Tokyo, Japan). This unit is designed to be interchangeable with the standard LED light module supplied with the LI-6800P gas-exchange system (LI-COR, Lincoln, NE, USA) and can be mounted onto the chamber head, enabling uniform illumination over a 1 cm × 3 cm leaf area (Figure 1). The light source consists of two LD modules: a red LD with a peak wavelength at 660 nm and a blue LD with a peak at 452 nm. These two light sources can be combined at every ratio to achieve desired spectral compositions. While LDs inherently emit highly directional light, the light emitted from this device is diffused by an optical expander to ensure even illumination across the leaf surface; however, the angle between the optical axis and the outermost edge of the irradiated leaf area is limited to approximately 8°.

The spectral characteristics of each light source were measured using a spectrometer (Ocean SR4, Ocean Optics, CA, USA). Spectral data were collected at 1 nm intervals across the 350–800 nm range while each light source was turned on. Subsequently, background noise was measured with the light source turned off. Spectral intensities were expressed as relative values by normalizing to the peak intensity (set to 1). For each light source, the full width at half maximum (FWHM) was calculated as a measure of peak sharpness, defined as the wavelength interval between the two points on either side of the peak at which the spectral intensity equals half of the maximum.

Rice plants were dark adapted overnight prior to measurement. Fully expanded third or fourth leaves from 4-week-old seedlings were enclosed in the gas exchange chamber of the measurement system. Only the enclosed leaf area inside the gas-exchange chamber was illuminated during measurements, while the rest of the plant remained under dark conditions. As the actinic light source, either the newly developed LD light described in the previous section or a commercially available LED light (LI-6800-02P, LI-COR, Lincoln, NE, USA) was used under two light colors, 100% red light (“R₁₀₀”) or 80% red and 20% blue light (“R₈₀B₂₀”).

To investigate the effects of different light sources on photosynthetic induction, time-course responses of photosynthetic parameters were measured. The chamber was maintained at a temperature of 25°C, a RH of 70%, and a PPFD of 950 µmol m⁻² s⁻¹, respectively, and data were collected at 60 s intervals for 60 min under R100 or R80B20 conditions. The measured photosynthetic parameters included CO₂ assimilation rate (A, µmol m⁻² s⁻¹), transpiration rate (E, mmol m⁻² s⁻¹), stomatal conductance (g_s, mmol m⁻² s⁻¹). The intrinsic water-use efficiency (iWUE) was calculated using A and g_s as shown in Equation 1:

For the photosynthetic induction curves, the final three data points were taken as representative of the steady-state photosynthetic parameters. Each treatment was replicated at least 4 times using independently grown plants, with one fully expanded leaf measured per plant.

To assess the periodicity of the photosynthetic induction curve, its frequency components were extracted using fast Fourier transform (FFT). The time series of photosynthetic parameters during the induction phase were multiplied by a Hanning window and then subjected directly to FFT in R software (http://www.r-project.org/) using the built−in FFT function. Power spectral density was computed over the frequency band from 0.0003 Hz (the lower limit set by the total observation duration) to 0.006 Hz, and dominant frequencies were identified as peaks.

To obtain the A-C_i (intercellular CO₂ concentration) curve and g_s-C_i curve, the C_a (ambient CO₂ concentration) in the gas chamber was sequentially changed to 0, 50, 100, 200, 400, 600, 800, 1000, and 1500 µmol mol⁻¹, and A, g_s and C_i were measured at each C_a level. The chamber was maintained at a temperature of 25°C, a RH of 70%, and a PPFD of 300 µmol m⁻² s⁻¹, respectively. Different PPFD levels were used depending on the experimental purpose: high irradiance (950 μmol m⁻² s⁻¹) was applied for photosynthetic induction measurements to detect dynamic stomatal responses, whereas moderate irradiance (300 μmol m⁻² s⁻¹) was used for A–C_i and g_s–C_i analyses to ensure stable parameter estimation under non-light-limiting conditions. Vc_max, the maximum rate of Rubisco carboxylation, and J_max, the maximum rate of electron transport supporting RuBP (ribulose-1,5-bisphosphate) regeneration, were estimated by fitting the A-C_i response curves to the Farquhar’s model (Farquhar et al., 1980), using nonlinear least squares (nlsLM) in R software. The Farquhar model of photosynthesis can be formulated as shown in Equation 2:

Where A_c is Rubisco-limited CO₂ assimilation rate, A_j is electron transport-limited CO₂ assimilation rate, and R_d is day respiration rate (μmol CO₂ m⁻² s⁻¹). To ensure stable parameter estimation, R_d was fixed at 1 following earlier studies for C₃ plants (e.g., Von Caemmerer, 2000; Yamori et al., 2005). A is limited by the smaller of A_c or A_j, and these two limiting rates can be described in detail by Equations 3 and 4:

Where C_i is intercellular [CO₂] (μmol mol⁻¹), Γ^* is CO₂ compensation point (μmol mol⁻¹), K_c and K_o are Michaelis-Menten constants for CO₂ and O₂, and O is ambient O₂ concentration (mmol mol⁻¹). Temperature-dependent parameters in the Farquhar model, including Γ^*, K_c, and K_o, were adjusted according to exponential temperature response functions based on activation energies (Bernacchi et al., 2001). O was assumed to be 210 mmol mol⁻¹. In the g_s–C_i curve, the intercellular CO₂ concentration at which stomatal conductance reached its maximum (g_s^*) was defined as C_i^*. The regions preceding and following C_i^* were designated as the low-C_i and high-C_i regions, respectively. Within each region, linear regressions were performed using up to four data points, and the slopes of the regression lines were used as indices of changes in stomatal conductance.

During the photosynthetic induction experiment described in the previous section, a rice leaf was removed from the gas-exchange cuvette during one of the following time points after the onset of the lighting: 0, 900, 1200, 1800, or 2400 s. Immediately after removal, cyanoacrylate adhesive (Aron Alpha, Toagosei Co., Ltd., Tokyo, Japan) was applied to the abaxial side of the leaf at the irradiated region and left to dry. Once dried, the adhesive film was peeled off to obtain stomatal footprints representing the shape and aperture of stomata. The footprints were observed under a light microscope (CH-40, Olympus Corp., Tokyo, Japan), and at least 200 stomata per sample were visually classified into three categories based on the degree of stomatal opening: closed, partially open, or open (Toda et al., 2018). The relative frequency of each category was then calculated.

In this study, statistics were analyzed in R software. Data were analyzed by two-way ANOVA followed by the Tukey’s HSD test (P< 0.05) for comparing between light sources and light color.

The wavelengths of red and blue LD and LED light sources were measured (Figure 2). The emission spectrum of each light source was characterized by its peak wavelength and full width at half maximum (FWHM), which indicates the spectral bandwidth corresponding to half of the maximum intensity. The results showed that the peak wavelength of blue light was 452 nm for the LD and 454 nm for the LED. Similarly, the red LD and LED exhibited peak wavelengths of 660 nm and 661 nm, respectively. Thus, for both blue and red light, the peak wavelengths were nearly identical regardless of the light source. When the spectral width was defined as FWHM, the blue LD had a FWHM of 2.17 nm and the red LD 1.80 nm, whereas the blue and red LEDs exhibited much broader FWHM of 17.1 nm and 13.6 nm, respectively (Table 1).

In rice under R₁₀₀ LD, steady‐state net CO₂ assimilation rate (A) was significantly higher than under R₁₀₀ LED (Figure 3A), and stomatal conductance (g_s) and transpiration (E) rate were elevated (Figures 3B, C). Because the relative increase in stomatal conductance exceeded that of net CO₂ assimilation rate, intrinsic water‐use efficiency (iWUE) was reduced under R₁₀₀ LD (Figure 3D). By contrast, when rice was illuminated with R₈₀B₂₀ LD versus LED, there were no significant differences in A, g_s, E, or iWUE (Figures 3A–D). Comparison of absolute values across all treatments revealed that A under R₈₀B₂₀ LD and LED, and R₁₀₀ LD did not differ and were uniformly higher than under R₁₀₀ LED, which exhibited the lowest A (Figure 3A). Likewise, g_s and E were highest and equivalent under R₈₀B₂₀ LD and LED, intermediate under R₁₀₀ LD, and lowest under R₁₀₀ LED (Figures 3B, C).

Next, we compared the induction curves of each photosynthetic parameter. In rice under R₁₀₀ LD, periodic oscillations over time were observed most prominently in g_s and E compared with R₁₀₀ LED (Figures 4B, C). A and iWUE, on the other hand, exhibited weaker or less distinct oscillatory tendencies compared with the stomatal parameters (Figures 4A, B). To quantify these periodicities, we applied fast Fourier transform (FFT) to the induction curves and found that A and iWUE showed only minor differences in their frequency spectra between R₁₀₀ LD and R₁₀₀ LED (Figures 4E, H), whereas g_s and E under R₁₀₀ LD exhibited a distinct frequency component at 1.12 × 10⁻³ Hz (≈ 900 s) that was not observed under R₁₀₀ LED (Figures 4F, G).

Under R₈₀B₂₀ illumination, as in the R₁₀₀ condition, periodic oscillations in the induction curves of g_s and E were observed under LDs but not under LEDs (Figures 5B, C). However, these oscillations appeared to have longer periods than those under R₁₀₀ LDs (Figures 5B, C). In contrast, oscillatory behavior in A and iWUE was much less apparent than in g_s and E (Figures 5A, D). To compare these responses with those under LEDs, we again performed FFT analysis. The FFT results showed no substantial differences in oscillation amplitude of A, g_s, or E between R₈₀B₂₀ LD and LED treatments (Figures 5E–G), whereas the amplitude of iWUE was greater under R₈₀B₂₀ LDs (Figure 5H). Moreover, the frequency component around 900 s that was evident in g_s and E under R₁₀₀ LDs was absent under R₈₀B₂₀ LD (Figures 5F, G).

Because g_s and E showed dynamic temporal fluctuations under R₁₀₀ LD (Figures 4B, C), we used microscopy to visually examine stomatal morphology. Following a classification of Toda et al. (2018), stomata were divided into three categories, “Closed”, “Partially Open”, and “Open” (Figure 6A). We then tracked the relative proportion of each category over time in rice under R₁₀₀ LDs versus LEDs. Under R₁₀₀ LDs, the percentage of open stomata (“Partially Open” + “Open”) increased to 94.5% at 900 s and to 72.5% at 1800 s (Figure 6B). By contrast, under R₁₀₀ LEDs the proportion of open stomata peaked at 76.5% at 900 s and then gradually declined to 50.0% by 3600 s (Figure 6B). These shifts in open‐stomata ratio reflected the time courses of g_s and E observed under R₁₀₀ LDs and LEDs (Figures 4B, C).

By varying ambient CO₂ (C_a) and measuring A, g_s and intercellular CO₂ concentration (C_i), we generated both A-C_i and g_s-C_i curves to probe stomatal dynamics under LD illumination. Under R₁₀₀ LDs versus LEDs, A-C_i responses did not differ significantly (Figure 7A), whereas under R₈₀B₂₀ LDs, A remained relatively lower as C_i increased compared with R₈₀B₂₀ LEDs (Figure 7B). To discover why CO₂ assimilation was depressed under R₈₀B₂₀ LDs, we estimated photosynthetic capacity parameters, maximum carboxylation rate (Vcmax) and maximum electron transport rate (Jmax) from the A-C_i curves by using Farquhar’s model (Farquhar et al., 1980). Vcmax values tended to cluster into higher (R₁₀₀ LDs and LEDs) and lower (R₈₀B₂₀ LDs) groups, with statistical overlap leaving R₈₀B₂₀ LEDs intermediate (Figure 7C). In contrast, Jmax in rice grown under R₈₀B₂₀ LD was comparable to that under the other light treatments (Figure 7D).

By varying C_a, we obtained g_s corresponding to C_i. Under the R₁₀₀ condition, the maximum g_s (g_s*) did not differ between LD and LED treatments, but the C_i corresponding to g_s* (C_i*) was lower under LDs than LEDs (Figure 8A). Under the R₈₀B₂₀ condition, C_i* was also lower under LDs, and g_s* was greater under LDs than LEDs (Figure 8B). To quantify stomatal responsiveness to CO₂, linear regressions were performed before and after C_i*, and the regression slopes were extracted. In both the before C_i* and after C_i* regions, LD treatments exhibited significantly steeper slopes than LEDs treatments, with a particularly pronounced difference under the R₈₀B₂₀ condition (Figures 8C, D).

In the present study, g_s and E of rice grown under R₁₀₀ LDs were higher than those under R₁₀₀ LEDs, but lower than those under R₈₀B₂₀ in both LD and LED conditions (Figures 3B, C). By contrast, A under R₁₀₀ LDs remained high and did not differ significantly from that under R₈₀B₂₀ (Figure 3A). These findings suggest that stomatal opening in rice under R₁₀₀ LDs may be driven predominantly by a red light–dependent mechanism.

Extensive LED-based studies have shown that supplementing red light with even a small proportion of blue light markedly enhances photosynthetic parameters such as A and g_s compared with monochromatic red light (Hogewoning et al., 2010; Savvides et al., 2012; Wang et al., 2016). For instance, Savvides et al. (2012) demonstrated that providing cucumber leaves with a mixture of red LED and 30% blue LED increased stomatal conductance to more than three times that under red light alone, accompanied by a rise in photosynthetic rate. Similarly, Hogewoning et al. (2010) reported that adding as little as 7% blue light to red illumination more than doubled the maximal photosynthetic rate relative to pure red light. These studies underpin the prevailing view that stomatal opening is primarily driven by blue light.

However, in our study, g_s under R₁₀₀ LDs exceeded that under R₁₀₀ LEDs (Figure 3B), and A under R₁₀₀ LDs was comparable to that under R₈₀B₂₀ (Figure 3A). These observations indicate that stomatal opening in rice under R₁₀₀ LDs is likely mediated mainly by red light. Red light–dependent photosynthesis in mesophyll or guard cells has been shown to trigger stomatal responses independently of blue-light signaling (Baroli et al., 2008; Ando and Kinoshita, 2018; Bernardo et al., 2023). Such mechanisms may have been activated more strongly under LD than LED illumination. The possibility that red LDs promotes stomatal opening has also been proposed previously (Li et al., 2025). Nevertheless, LDs differ from LEDs in several optical properties—monochromaticity, coherence, and beam collimation among them—and the physiological effects of such extremely narrowband illumination remain poorly understood and have not been systematically investigated, making it unclear which of these characteristics contribute to enhanced stomatal opening.

In this study, rice under R₁₀₀ LDs exhibited pronounced periodicity in g_s and E (Figures 4B, C, F, G), although this periodicity was diminished in rice under R₈₀B₂₀ (Figures 5B, C, F, G). In contrast, under LED illumination, g_s and E increased once after the onset of measurement but subsequently declined, irrespective of whether blue light was present (Figures 4B, C; Figure 5B, C). Furthermore, the proportion of “Partially Open” and “Open” stomata in rice under R₁₀₀ LD and LED conditions (Figure 6B) closely mirrored the rise-and-fall patterns observed in the induction curves of g_s and E (Figures 4B, C; 5B, C). Collectively, these observations indicate that red LDs promote rapid and periodic stomatal opening and closing in rice, whereas blue LDs appear to suppress stomatal dynamics.

Many physiological processes display oscillations on time scales ranging from 10 min to 24 h or longer, referred to as “biological rhythms” (McClung, 2006; Damineli et al., 2022). Stomatal movements are also strongly influenced by such rhythms: oscillations in stomatal conductance and transpiration rate with periods of tens of minutes, classically described as oscillatory stomatal behavior (Farquhar and Cowan, 1974) or transpiration oscillation (Johnsson et al., 1979), are sometimes referred to as “ultradian rhythms” in more recent studies (Zait et al., 2017; Peak et al., 2023). Our study indicated that stomatal oscillations during photosynthetic induction in rice were most pronounced under red LDs, whereas these oscillations were suppressed under blue LDs. Periodic Ca²⁺ oscillations in guard cells are considered a primary driver of stomatal ultradian rhythms, as they rhythmically regulate ion transporters such as H⁺-ATPases and K⁺ channels (Yang et al., 2005; Damineli et al., 2022), giving rise to membrane-potential oscillations. These electrical oscillations interact with hydraulic cycles produced by transpiration-driven declines and subsequent recovery of leaf water potential (Steppe et al., 2006), together generating the characteristic periodic opening and closing of stomata. We confirmed that applying Ca²⁺ to the leaf surface of lettuce enhanced stomatal conductance and transpiration rate under red LD illumination compared with plants without Ca²⁺ application (data not shown). Collectively, these results suggest that red LDs may modulate the amplitude and persistence of stomatal oscillations during photosynthetic induction through changes in intracellular Ca²⁺ concentrations in leaves. Although a LED-based study has suggested that red light–dependent stomatal opening requires the activation of ion transporters such as H⁺-ATPase and K⁺ channels (Ando and Kinoshita, 2018), it remains unclear why red LDs is able to activate these mechanisms more strongly.

In rice under R₈₀B₂₀ LDs, stomatal movements were far more pronounced at lower C_i, and the amplitude of those movements exceeded that of the other three treatment groups (Figures 8A, B). In other words, despite low C_i, plants under R₈₀B₂₀ LDs opened their stomata more widely and took up CO₂ more effectively. This behavior follows the basic blue‐light-dependent stomatal opening mechanism: blue light absorbed by the phototropin receptors phot1 and phot2 in guard cells activates plasma membrane H⁺-ATPase and K⁺ channels, driving stomatal aperture (Inoue and Kinoshita, 2017). Moreover, because g_s^* under R₈₀B₂₀ LDs exceeded that under R₁₀₀ LDs (Figure 3B), the blue LDs appear to act additively with red LDs to enhance stomatal opening.

A theoretical model predicts that when stomatal conductance responds strongly even to small changes in C_i, CO₂ assimilation capacity should increase accordingly (Medlyn et al., 2011). However, contrary to this model, rice under R₈₀B₂₀ LDs exhibited lower CO₂ assimilation rates across rising C_i than the other three treatments (Figures 7A, B). In addition, Vcmax under R₈₀B₂₀ LDs tended to be lower than that of the other treatments, whereas Jmax remained unchanged (Figures 7C, D). These results imply that blue LDs may impose a non-stomatal limitation on photosynthesis, despite promoting blue‐light‐dependent stomatal opening. This negative impact may arise from the interaction of LD’s inherent coherence and the heterogeneous absorption of blue light in the leaf. When coherent LD light illuminates a leaf, speckle contrast creates a patchwork of high‐ and low‐intensity zones determined by surface microtopography (Zhong et al., 2013). Thus, transmitted LD photons may interfere to form localized light‐intensity “hot” and “cold” spots within the tissue. Furthermore, blue light is absorbed almost entirely (≈98%) in the palisade mesophyll, whereas red light penetrates both palisade and spongy mesophyll layers (Vogelmann and Evans, 2002). Concentrated absorption of blue light in the palisade has been shown to slow electron transport and reduce photosynthetic rate by about 11% (Earles et al., 2017). It should be emphasized that this proposed mechanism remains speculative and requires direct experimental validation, for example through imaging of intra-leaf light absorption and distribution under LD versus LED illumination. Taken together, under blue LDs, its coherent speckle‐induced heterogeneity and localized overabsorption in the palisade are likely to suppress overall photosynthetic efficiency.

In this study, we evaluated photosynthetic parameters of rice under blue LDs with a peak wavelength of 452 nm and red LDs with a peak wavelength of 660 nm, both of which exhibit much narrower spectral bandwidths than LEDs (Figure 2; Table 1). In contrast, lighting systems currently used in PFALs predominantly rely on white or purple LEDs, which provide mixed spectra composed of multiple monochromatic components rather than narrowband light sources (Rihan et al., 2022; Yano et al., 2023). Therefore, when considering LDs as alternative light sources for PFALs, it is necessary to systematically evaluate which wavelengths and spectral characteristics confer physiological advantages, including scenarios in which LDs of different colors are combined.

Under both LD and LED conditions, rice under R₁₀₀ lighting exhibited significantly higher iWUE than rice under R₈₀B₂₀ lighting (Figure 3D). Blue light is known to accelerate stomatal and photosynthetic induction; however, under steady-state conditions, increases in stomatal conductance can exceed gains in CO₂ assimilation, potentially resulting in reduced water-use efficiency (Matthews et al., 2020). Such mechanisms likely underlie the lower iWUE observed in rice under R₈₀B₂₀ conditions compared with R₁₀₀ LDs or LEDs. In addition, rice under R₁₀₀ LD, but not under R₁₀₀ LED, maintained CO₂ assimilation rates comparable to those observed under R₈₀B₂₀ LDs or LEDs (Figure 3A). These results indicate that rice leaves illuminated with red LDs peaking at 660 nm can achieve high photosynthetic rates while minimizing water loss, suggesting that 660 nm red LDs may represent an effective cultivation light source for PFALs.

Nevertheless, because the red LDs used in this study had an extremely narrow spectral width (Table 1), it remains uncertain whether similar physiological responses would be observed under red LDs with different peak wavelengths or broader spectral widths. Consistent with this concern, tobacco grown under red LDs peaking at 660, 664, and 673 nm exhibited maximal net photosynthetic rates at 660 nm, with a sharp decline at longer wavelengths (Li et al., 2025). Moreover, A. thaliana cultivated under a combination of a 671 nm red LDs and a 473 nm blue LDs showed reduced chlorophyll content and dry mass compared with plants grown under white LED illumination (Ooi et al., 2016). Taken together, these findings indicate that while red LDs peaking at 660 nm can simultaneously enhance water-use efficiency and photosynthetic performance, their effectiveness is strongly dependent on spectral properties, underscoring the need for precise spectral control when applying LDs to PFAL systems.

Beyond spectral optimization, practical implementation of LD lighting in cultivation systems also requires consideration of engineering constraints such as energy efficiency and water availability. It should be noted that energy consumption and total operational costs of LD- and LED-based lighting systems were not directly evaluated in this study and therefore remain topics for future investigation. Future studies should therefore integrate direct measurements of electrical power input with photosynthetic performance to assess energy-use efficiency under practical PFALs. Advances in manufacturing processes and yield improvements associated with large-scale production have substantially reduced the cost of LED lighting (Tsao et al., 2010), suggesting that LEDs generally require lower initial investment than LDs as cultivation light sources. In contrast, LDs can deliver narrowband light closely matching chlorophyll absorption peaks, a property that has been suggested to improve the efficiency with which light energy is utilized for photosynthesis. For example, red LD illumination at 660 nm has been shown to significantly enhance photosynthetic capacity compared with red LEDs of the same peak wavelength (Li et al., 2025).

High water-use efficiency under LD illumination may be particularly advantageous in cultivation environments where water availability is severely constrained, such as space. On the International Space Station (ISS), more than 90% of consumed water is recycled (Ichimura and Yamashiki, 2025), whereas the remaining fraction must be resupplied from Earth via cargo missions every one to two months (Lynch et al., 2024). In such environments, cultivation systems that maximize water-use efficiency are essential. As demonstrated in this study, rice under R₁₀₀ LDs exhibited higher water-use efficiency while maintaining photosynthetic rates comparable to those under LED illumination (Figures 3A, D), suggesting that LD lighting may serve as a promising light source candidate for extreme cultivation environments, including space agriculture, which lies along the conceptual extension of PFALs.