Off-line and on-line optical monitoring of microalgal growth

Microalgae cultures

Chlamydomonas reinhardtii (CC-124) microalgae were purchased from the Chlamydomonas Resource Center (USA) and were grown photoautotrophically in a Sueoka medium (Sueoka, 1960). Growth conditions included continuous air bubbling (1 VVM = 1 L − air/min/L − medium) under controlled room temperature conditions (298 ± 2) K. Experiments were performed in two series. The first series was conducted for the off-line OD and fluorimetry measurements (Experiments A and B). The second series was conducted for the on-line techniques that used digital images (Experiments 1–5). A portable spectrometer (EPP2000; StellarNet, Tampa, FL, USA) was used to measure the OD, fluorescence and color of the microalgae cultures, with a detection range between 200 and 850 nm. Spectrawiz software (OS V5.0©2011; StellarNet) was used for spectrometer measurements. In order to monitor cell growth, OD measurements at λ = 640 nm were taken. Each measurement was repeated three times for every sample.

For the off-line experiments, three 1 − L Roux culture bottles were used, each with 0.9 L of algae culture and continuous air bubbling (one L/min), under controlled room temperature conditions (Fig. 1A). The cultures were illuminated continuously with two fluorescent lamps (Philips, F20T12/D 20W). To assess the microalgae growth kinetics, three samples of three mL each were taken from every bottle every 12 h for 5 days. Additionally, a final control measurement was taken on the 7th day. Triplicate OD and fluorescence measurements were taken for every sample. Two experiments were performed with the only difference between them being the initial microalgae concentration (x₀). These were labeled Experiment A (x₀ = 34 ± 2 mg/L), and Experiment B (x₀ = 42 ± 2 mg/L). In summary, a total of 270 measurements were made for every off-line experiment and the number of measurements provided statistical significance to this study.

For the on-line experiments, a 3 − L column PBR was used (Fig. 1B). A transparent acrylic tube with a thickness of 25 mm and an inner diameter of 95 mm was used to build the 95 − cm length PBR. Four fluorescent lamps (T4 6500K20W; Tecno Lite, China) were used to illuminate the center of the column with around 100 μmol photons m⁻² s⁻¹. To avoid external light, the PBR was placed inside a dark cabin. Five experiments were performed, each with 5 days of sampling. These included three experiments with continuous illumination (Experiments 1, 2, and 3), and two experiments with 12-h light/12-h dark cycles (Experiments 4 and 5). Both off-line samples and digital images were taken every 6 or 12 h. A webcam (Carl Zeiss Tessar HD 1080 p; Logitech, China) connected to a PC was used to capture the PBR digital images. For the color analysis, the images were captured with two fluorescent lamps located behind the PBR. These lamps were turned on while the other two lamps remained turned off. The fluorescent images, were captured in darkness and only the culture was illuminated with a blue beam from a super luminescent diode (maximal wavelength of 440 nm) and a light filter (LSR-GARD ARGON, model 2204) with a protection range within 190–520 nm. In order to measure fluorescence after a dark-adapted period, at least 15 min of darkness was provided before fluorescent stimulation. The images were analyzed in a LCD screen with an integrating Cube (IC2; StellarNet) and portable spectrometer. The International Commission on Illumination (CIELAB) scale was used for color measurements. To achieve a homogeneous representation of the color of every image, five measurements were taken from different regions of the PBR.

Fluorescence cabin

The experimental system to measure fluorescence was composed of two main parts: a fluorescence cabin and the portable spectrometer. Based on the spectrometer, the fluorescence cabin was designed, manufactured, and coupled to the system using an optical fiber (F400; StellarNet). Through this fiber, light was guided from the sample cuvette to the spectrometer detector. The fluorescence cabin configuration is shown in Fig. 2A. The main components are a dark cabin, a cylindrical (14 mm i.d.) glass sample cuvette (four mL), six light emission diodes (LED), feed and switching electronic circuits, an AC/DC electric current converter (output 5.4 V), and a multi-modal optical fiber connector. The dark cabin is a space where light does not come in from external sources. Inside the dark cabin, a base was used to fix the cuvette in a normal position (at 90°) relative to the floor. As exciting radiation, six LEDs (Ultra Blue; Steren, Shangai, China) were placed parallel to the floor, three to the right side of the sample cuvette and three to the left. This configuration ensured homogeneous illumination conditions. In Fig. 2B, the spectrum of the six LEDs is shown, with a maximum wavelength emission at around λ = 464 nm, luminosity of 7 cd and 400 mW as maximum power.

To improve the quality of fluorescence measurement, the optical fiber was positioned on top of the glass sample cuvette, at 90° relative to the LEDs’ radiation. Fluorescence traditionally is measured through the sample cuvette wall, but our fluorescence measurements were taken at the uncovered top of the cuvette, thus reducing losses from reflection and refraction in the interface of the cuvette. The cuvette used for the samples was a four-mL glass cylinder, but with this fluorescence system, the cuvette’s geometry is not important. Light information was processed and digitized using the spectrometer SpectraWiz software. Before measurements, the reference blank was defined by setting the spectrometer with the Sueoka medium, then the fluorescence measurement was performed on the microalgae sample. The direct fluorescence of C. reinhardtii culture samples at room temperature was measured successfully for 7 days with this experimental setup. The ratio of fluorescence intensity between λ = 685 and λ = 740 nm was calculated.

The F₆₈₅/F₇₄₀ fluorescence ratio

At low Chl concentrations, fluorescence emissions increase with increasing amounts of Chl. At higher concentrations, the increase of fluorescence with the increment of Chl is mainly detected around 740 nm. For in vivo cultures, fluorescence emission at 740 nm is favored and fluorescence emission at 685 nm is not favored. This is due to the following factors: (i) the re-absorption of photons from the fluorescence emitted by neighboring molecules, (ii) light interference between the short (685 nm) and long wavelengths (740 nm), and (iii) the increment of Chl (the new Chl molecules preferentially absorb energy at 685 nm) (Gouveia-Neto et al., 2011).

There is a good inverse correlation between photochemistry and Chl fluorescence. The ratio of fluorescence intensity between maximal wavelengths (F₆₈₅/F₇₄₀) is influenced by photosynthetic activity. In mature microalgae cultures, the chloroplast structure, CO₂ uptake rate, carbon metabolism, etc., are better than in younger cells. Higher F₆₈₅/F₇₄₀ values signal young cultures or cultures with a photosynthetic apparatus that isn’t developing. Low values of this rate indicate mature cultures with a fully developed photosynthetic apparatus. In other words, a decrement in F₆₈₅/F₇₄₀ values is indicative of increased photosynthetic activity. Measured by induction fluorescence, F₆₈₅/F₇₄₀ exhibits a curvilinear relationship with cell concentration (x). This relationship is successfully expressed in Eq. (1), where c and d are constants (Hák, Lichtenthaler & Rinderle, 1990): (1) $$\frac{F_{685}}{F_{740}}=c{x}^{-d}$$

This technique has been applied to all kinds of leaves, chloroplast suspensions, and acetone extracts of photosynthetic pigments. Our study demonstrated that this technique is also applicable to microalgae cultures.

Microbial growth

It has been established that the Gompertz model represents C. reinhardtii growth better than the classical Monod model (Del Campo et al., 2014). Actually, the Monod and Gompertz models can be seen as particular cases of a more universal growth model (Castorina, Delsanto & Guiot, 2006). For the experiments in the Roux bottles (A and B), as well as in the PBR (1–5), the Gompertz specific growth rate is reported in Table 1. For the PBR experiments, it was possible to observe that the Gompertz model is best fitted when the light regime is continuous (see R² for Experiments 1–3) than when light/dark cycles are performed (Experiments 4–5). Moreover, the specific growth rate is favored when the PBR is used. This stands in contrast to the cultures in the Roux bottles.

Fluorescence measurements

Variations in fluorescence intensity were successfully measured according to the increment in C. reinhardtii concentration. Some selected spectra are shown in Fig. 3 for experiment A. The fluorescence dataset for experiments A and B are shown in File S1. In all cases, Chl fluorescence exhibits a peak around λ = 685 nm and a broad shoulder around λ = 740 nm. This is a general observation at room temperature for Chl a (Gouveia-Neto et al., 2011; Krause & Weis, 1984). The fluorescence around λ = 685 nm is attributed to the PSII antenna, and the fluorescence around λ = 740 nm is due to the PSI antenna (Gouveia-Neto et al., 2011). The fluorescence signal/noise ratio in the measurements was around 7 at λ = 740 nm and 14 at λ = 685 nm.

In Fig. 4, fluorescence evolution at λ = 685 nm and at λ = 740 nm is shown for all bottles of Experiments A. The plots include trend lines. In both Experiments A and B, the maximum fluorescence intensity at λ = 685 nm occurred between 48 and 60 h. After that time, fluorescence decreased. Regarding fluorescence at λ = 740 nm, in Experiment A, the maximum value occurred at 72 h in all cultures. In Experiment B, the maximum value happened between 96 and 108 h. For both experiments, the maximum fluorescence at λ = 685 nm occurred before that of λ = 740 nm. Maximum fluorescence at λ = 740 nm occurred at the highest algae concentrations. In general, the results were consistent with those described in the literature for fluorescence in plants (Gouveia-Neto et al., 2011). Namely, at low Chl concentrations, fluorescence emissions increased with increasing Chl concentration. At higher concentrations, the increase of fluorescence with the increment of Chl was detected only around λ = 740 nm.

Regarding the F₆₈₅/F₇₄₀ ratio, we observed a similar behavior in every culture of both experiments, regardless of the initial concentration. For that reason, Fig. 5 shows the average of the three cultures for each of the two experiments over 5 days. Over time, the F₆₈₅/F₇₄₀ fluorescence ratio decreased, meaning that the photosynthetic processes had improved. The same trend has been reported for green leaves in plants, as stated in Eq. (1) (Hák, Lichtenthaler & Rinderle, 1990).

Based on the information in Fig. 5, and considering the data of 168 h as the minimum possible value (aged cultures), we figured that cultures in both experiments reached around 70% maturity at 96 h. Therefore, every culture evolved successfully and the conditions were appropriate to grow the microalgae and keep them in a good state of health. In addition, the cultures that reached the lowest F₆₈₅/F₇₄₀ values, that is, the highest photosynthetic activity, were those of Experiment A, which started with the lowest initial concentration. Moreover, after 72 h, these values did not change significantly. This is the moment when illumination may not be enough for the culture because cell concentration reduces the passage of light. Equation (2) expresses a very useful linear correlation between the logarithmic concentration of microalgae and the F₆₈₅/F₇₄₀ ratio through time (Fig. 6): (2) $$\text{ln}\left(\frac{x}{x_0}\right)=3.27-0.7084(F_{685}/F_{740})$$

Digital images

When the PBR was used, digital images of the cultures were taken to monitor the change in color and to measure the penetration of a fluorescent beam during microalgae growth. Figure 7 shows a selection of images illustrating a typical experiment. We observed that cultures got darker with time due to the increase of biomass concentration in the PBR preventing the passage of light throughout the reactor. For the fluorescence measurements, the flashes due to the blue super-luminescent diode were filtered in order to measure only fluorescence light contribution. This contribution diminishes with time due to a shadow effect produced by cells as the microalgae concentration gets denser.

International Commission on Illumination measurements include three values to characterize the color of a sample: L is the luminosity, the parameter “a” represents colors from green to red, and the parameter “b” represents colors from blue to yellow. Both parameters “a” and “b” remained almost constant throughout the experiment during microalgae growth. This means that, technically, color does not change. This is expected since the photosynthetic pigments are always the same and luminosity is what more importantly diminishes during cell growth since cells deflect or shadow light sources. Figure 8 shows the logarithmic correlation between microalgae concentration x, and luminosity L for Experiments 1–3. For these experiments, Eq. (3) is proposed to get x from on-line measurements of L from digital images: (3) $$\text{ln}\left(\frac{x}{x_0}\right)=(1.6\pm 0.2)-(0.44\pm 0.04)(L/W)$$

In this equation, the intercept and the slope values are presented as the mean and the corresponding standard deviation of the individual correlations in the three experiments (see Fig. 8). However, this same correlation was not observed in Experiments 4–5, where light/dark cycles were performed (Fig. 9). The correlations between the values of the three experiments confirm that the experiments are reproducible, which gives confidence to the study. Table 2 shows the L(W) and Table 3 shows the correlation between ln(x/x₀) values.

Finally, the fluorescence beam penetration was characterized as an image changing with time. Both the beam surface area (data not shown) and penetration distance were used to monitor these changes. Similar results were obtained when comparing the cell concentration with the changes in the fluorescent images; therefore, only the distance beam penetration was used since its measurement was much simpler than that of the surface area. Figure 10 shows the correlation of this distance measured for the fluorescent beam penetration with the inverse of the OD. As stated before, the OD is related to the biomass concentration (Del Campo et al., 2014). It is important to note that the linear correlations were obtained for all experiments, with Eq. (4) proposed to calculate the OD of the culture directly from the on-line measure of the beam penetration: (4) $$\text{OD}=\frac{1}{\text{Beam\hspace{0.17em}penetration}/\text{cm}}$$

This simple equation is proposed since the values for the mean and the corresponding standard deviation for the intercept and the slope in the individual linear correlations for the five experiments are (0.0 ± 0.5) and (1.0 ± 0.1) cm⁻¹ respectively.