LED spectrum optimisation using steady-state fluorescence gains

The use of light emitting diodes (LEDs) in greenhouses entails the possibility to control the light in a better way, since both spectrum and light intensity can be adjusted. We aim at developing a method to automatically find the optimal spectrum in terms of energy consumption and plant growth. Previous work shows that chlorophyll fluorescence (ChlF) at 740 nm strongly correlates with the photosynthetic rate (carbon dioxide uptake rate) and that the net efficiency of a LED group therefore is coupled to the fluorescence gain w.r.t. energy consumption, i.e., the slope of a curve depicting steady-state ChlF versus applied power to the LED group. In the present work we compare the fluorescence gains for six different LED types (wavelength peaks from 400 to 660 nm) and six different species: tomato, cucumber, basil, lettuce (two species) and dill. We also compare two different kinds of experiments: steadystate experiments, waiting for the fluorescence to reach a steady state at a few incident light intensities, and ramp experiments, where the light intensity is increased slowly. The ramp experiment gives essentially the same information as the steady-state experiment, but was found to slightly overestimate the gains of the blue LEDs. Being aware of this, it should be possible to initially use the faster (ramp) method in order to find the right light composition, possibly using steady-state experiments for a few LED colours to fine tune the lamp. The relative order of the fluorescence gains among the tested LED groups is similar, but not identical, for all species tested. LED660 has the highest fluorescence gain w.r.t. incident photon flux density, and LED400 and/or LED530 have the lowest. However, the important quantity is in fact the fluorescence gain w.r.t. applied electrical power. If the individual electrical efficiencies of the LEDs change the most efficient power split on the different LEDs might change