Producing Energy-Rich Microalgae Biomass for Liquid Biofuels: Inﬂuence of Strain Selection and Culture Conditions

: Energy-storage metabolites such as neutral lipids and carbohydrates are valuable compounds for liquid biofuel production. The aim of this work is to elucidate the main biological responses of two algae species known for their effective energy-rich compound accumulation in nitrogen limitation and day–night cycles: Nannochloropsis gaditana , a seawater species, and Parachlorella kessleri , a freshwater species. Lipid and carbohydrate production are investigated, as well as cell resistance to mechanical disruption for energy-rich compound release. Nitrogen-depleted N. gaditana showed only a low consumption of energy-storage molecules with a non-signiﬁcant preference for neutral lipids (TAG) and carbohydrates in day–night cycles. However, it did accumulate signiﬁcantly fewer carbohydrates than P. kessleri . Following this, the highest levels of productivity for N. gaditana in chemostat cultures at four levels of nitrogen limitation were found to be 3.4 and 2.2 × 10 − 3 kg/m 2 · d for carbohydrates and TAG, respectively, at 56% NO 3 limitation. The cell disruption rate of N. gaditana decreased along with nitrogen limitation, from 75% (at 200% NO 3 ) to 17% (at 13% NO 3 ). In the context of potentially recoverable energy for biofuels, P. kessleri showed good potential for biodiesel and high potential for bioethanol; by contrast, N. gaditana was found to be more efﬁcient for biodiesel production only.

As well as the process by which bottlenecks are associated with biomass harvesting and TAG/carbohydrate recovery [17][18][19], there are several other challenges to overcome in order for these energy-rich metabolites to be produced efficiently, especially when they are produced in variable outdoor conditions. These include the preferred ways by which microalgae store energy in the molecules concerned [15,20], their resilience to environmental changes [21] and their specific response to light and day-night cycles (DNc) [14,22]. Nonetheless, microalgae biodiversity offers many species capable of adapting to these outdoor conditions with different sets of metabolite profiles [23][24][25][26][27][28][29].
Taleb et al. [30] screened 14 microalgae strains for TAG productivity for the purpose of biodiesel production. The most efficient strains found for biodiesel production were Nannochloropsis gaditana CCMP527 and Parachlorella kessleri UTEX2229, which accumulated up to 56% X and 40% X TAG content, respectively. The corresponding maximum areal TAG

Strains
The marine microalgae Nannochloropsis gaditana was first described by Lubian [44] as part of the Eustigmatophyceae class. It has been reported as reproducing asexually by fission. It can reach 3.5-4 × 2.5-3 µm in size, describing an ellipsoidal shape. N. gaditana CCMP527 (National Center for Marine Algae and Microbiota , USA) was the strain used for the experiments in this work.
Parachlorella kessleri is a fresh water microalgae belonging to the Chlorophyceae class, with a spherical or ellipsoidal shape. This species has also been reported as reproducing asexually by auto-spores (2, 4, 6 or 8) of around of 2.5-3.5 × 3-4.5 µm. In the mature stage, it can reach 5-10 µm in size [31,33,45]. In this work, the experiments on P. kessleri UTEX2229 (Culture Collection of Algae, University of Texas, USA) by Taleb et al. [30] and Kandilian et al. [35] have been used for comparison with N. gaditana.

Culture Medium
Bold Basal Medium (8.02 mM) was used for the cultivation of P. kessleri (the composition of the medium can be found in Kandilian et al. [35] For this work, four levels of nitrogen (N) limitation and one full depletion were used to trigger lipid accumulation. These levels were established as a percentage of the original NaNO 3 in the CONWAY formulation (100% NO 3 for 10.6 mM). The limitation levels were 200%, 56%, 46%, 29% and 13% NO 3 (21.2, 5.94, 4.88, 3.07 and 1.38 mM, respectively) and 0% NO 3 (0 mM) for nitrogen depletion. These values were chosen for the purposes of comparison with the work of Taleb et al. [30] and Kandilian et al. [35] on P. kessleri. The 200% NO 3 value was used to set the non-limited (replete) physiological state.

Culture Systems
Two culture systems were used in this study. The nitrogen-depleted culture was in a 170 L flat-panel airlift photobioreactor (HECtor technology). A detailed description of the photobioreactor (PBR) can be found in Pruvost et al. [47]. This system was operated in batch mode and supplied with artificial LED light to simulate day-night (DNc) cycles, in relation to the average annual irradiation (photon flux density, PFD 270 µmol/m 2 ·s) of Saint-Nazaire, France. Meteorological information on the region was obtained from the Meteonorm database, Switzerland. For the inoculation, a continuous culture of N. gaditana in non-limited conditions was established prior to initiating the batch. Once the culture was found to be in an exponential growth phase, enough biomass was harvested to inoculate the batch culture at 0.2 kg/m 3 in a fresh nitrogen-depleted (0% NO 3 ) medium. The temperature of the culture was regulated at 23 • C and the pH was set at 8 by automatic 98% CO 2 (gas) injection.
For the nitrogen limitation studies, a 1 L flat panel airlift PBR was used. This PBR has been described by Pruvost et al. [20]. The system was operated in continuous mode (chemostat) and under continuous artificial LED light, supplied at 250 µmol/m 2 ·s. The nitrogen-full (200 % NO 3 ) and nitrogen-limited (56%, 46%, 29% and 13% NO 3 ) culture media were supplied continuously using a peristaltic pump (Reglo ICC, Ismatec, Wertheim, Germany) to maintain the dilution rate at 0.01 h −1 . The volume in the PBR was maintained and continuously harvested with an overflow into a sterile harvest bottle. The bottle was replaced every time a sample was taken from it. As with the batch experiment, the same non-limited continuous culture was also used to provide the biomass for inoculating the corresponding nitrogen-limited experiments. The temperature was convectively controlled by the room temperature at 23 • C and the pH was set at 8 with the automatic injection of 98% CO 2 (gas).
For each nitrogen limitation level, a steady state was identified in the culture when the daily measurements of biomass, carbohydrate, pigment and total fatty acid concentrations showed no more than 5% variation with respect to the previous day for five consecutive days. Once the steady state was obtained in the culture, cell counting commenced and nitrate concentration was measured. Another level of nitrogen-limited medium was then supplied. The reported measurements at each steady state represent the average over 5 days at steady state. All the previous sampling set-ups were repeated successively using the corresponding medium for each limitation level, from 200% to 13% NO 3 .

Cell Disruption Rate and Cell Counting
Cell counting was done by analyzing (in triplicate) a 10 µL aliquot from the nitrogenlimited cultures at steady state. A malassez chamber and an optical microscope equipped with camera (Axio MRC Cam at Axio Scope A1 microscope, Carl Zeiss, Oberkochen, Germany) were used to take 40 pictures of each sample, as described by Zinkoné et al. [48]. Then, using an image treatment software (ImageJ v.1.52o, NIH, Bethesda, Maryland, USA) and a self-developed algorithm in MATLAB (MathWorks, Natick, Massachusetts, USA), the corresponding image data were processed to obtain statistical information on the number and size of the cells.
To determinate the cell disruption rate, another sample was adjusted to obtain 30 mL at a biomass concentration of 1 kg/m 3 using a phosphate buffer saline (PBS) solution. This was done to standardize the measurements from different nitrogen limitations and avoid the effect of concentration in the disruption system. The preparation was then passed through a high-pressure homogenization device (TS5, Constant Systems Limited, Daventry, UK) at 1750 bar.
The cell disruption rate (τ D ), which represents the mechanical resistance of cells, is reported below as the complementary proportion of the ratio of cells counted after bead milling, compared to those counted before the process.

Dry Weight Measurement
For biomass concentration analysis, 10 mL samples were collected in triplicate and passed through a pre-weighed 0.45 µm glass-fiber filter (Whatman GF/F, Cytiva, Marlborough, Massachusetts, USA). The filters were then washed with 1.19 M NH 4 HCO 2 and MiliQ water to remove residue salts from the culture medium. The filters were left at 103 • C for about 1 h or until the weight was stable. Biomass concentration X is reported as the difference in weight before filtration and after, in terms of the volume used. Similarly, pigment, carbohydrate, TFA and TAG content are reported as a percentage of the biomass (% X ). Biomass concentration was used to determine volumetric productivity P X and areal productivity S X . The corresponding definitions can be found in Pruvost et al. [49].

Pigment Content
Pigment content analysis was carried out in accordance with Ritchie [50] and Strickland and Parsons [51]. Samples of 0.5 mL were centrifuged at 6000 RFC, 4 • C for 10 min. The pellet was resuspended in 1.5 mL of methanol anhydrous 99.8%. After 1 min in a sonic bath, samples were incubated at 45 • C for 50 min. Then, the supernatant was recovered after centrifugation, and light absorption at 750, 665, 652 and 480 nm was measured (V-730, Jasco, Easton, Maryland, USA). The pigment concentrations were obtained using the equations from the literature previously cited. Pigment content is expressed as a percentage of the overall dry biomass (% X Pig).
Additionally, the ratio 480/662 nm (stress index) was calculated as a representation of the carbon/nitrogen content and level of stress in the cell. The higher the value, the greater the carotenoid content compared to chlorophyll. This indicates that nitrogen limitation has a direct effect, triggering the decrease in chlorophyll compared to carotenoids [52].

Total Fatty Acid and Triacylglycerol Content
The whole-cell extraction method described by Van Vooren et al. [53] was applied to quantify the total fatty acid (TFA) and triacylglycerol (TAG). Following the protocol, a 30 mL sample was centrifuged for 10 min at 6000 RFC and 4 • C. The pellet was recovered and freeze-dried for 24 h. A total of 6 mL of chloroform/methanol 2:1 v/v (Fisher Sci, Hampton, New Hampshire, USA) and 0.01% w/w of butylated hydroxytoluene was then added to the samples immediately and left for 6 h in incubation at 25 • C. Following sample sedimentation, the solvent was recovered and analyzed for TFA concentration by gas chromatography with a flame ionization detector, GC-FID (Thermo Fisher Sci, Waltham, Massachusetts, USA), and for TAG concentration by high-performance thin-layer chromatography, HPTLC (CAMAG, Muttenz, Switzerland). TFA/TAG content is expressed as a percentage of the overall dry biomass (% X TFA/TAG).
After determining the TAG content, the corresponding volumetric and areal productivities (P TAG , S TAG ) were also determined, in accordance with Pruvost et al. [49].

Total Carbohydrate Content
Total carbohydrates were measured following the protocol developed by DuBois et al. [54]. A 10 mL sample was centrifuged for 10 min at 6000 RFC and 4 • C. The pellet was recovered and resuspended in 1 mL of MiliQ water. A 0.5 mL aliquot was mixed with 0.5 mL of phenol solution (0.53 M) and three consecutive 2.5 mL doses of 97% H 2 SO 4 . Between each dose, the solution was vigorously agitated. The reaction was then left for 30 min at room temperature and wet-bath-incubated at 35 • C for 30 min, after which absorbance was measured at 483 nm (V-730, Jasco, Easton, Maryland, USA). A calibration curve from 0.025 to 0.1 g/L of glucose was prepared to calibrate the method and subsequently obtain the total carbohydrate concentration of the sample. The total carbohydrate content is expressed as a percentage of the overall dry biomass (% X Sg).
The areal productivity was also calculated based on carbohydrate content (S Sg ), using the same method as for TAG values.

Nitrate Concentration Measurement
Once a steady state was reached in the continuous cultures, the outflow culture medium was recovered for nitrate measurement. A volume of 3-5 mL was filtered using a 0.2 µm acetate cellulose filter (Minisart, Sartorius, Gottingen, Germany). The sample was injected into an ion chromatography system (ICS 900, Thermo Fisher Sci, Waltham, Massachusetts, USA) to find the NO 3 − ion concentration. If needed, concentration was adjusted by dilution with MiliQ water to make it readable under the ion chromatography method. Using the NO 3 − concentration, the Nitrogen uptake yield (Y N/X )was calculated with the equation where ∆N t is the difference between the NO 3 − concentration in the supplied and worn-out culture media.

Potentially Recoverable Energy Calculation
A theoretical estimate of the amount of energy that can be recovered from microalgae was calculated as the sum of the energy derived from the biofuels potentially obtained from the TAG (converted into biodiesel) and carbohydrates (converted into bioethanol) produced. This was expressed per unit of time and illuminated surface for simplicity of comparison with other applications.
First, the energy from biodiesel via TAG transesterification (E BioD in J/m 2 ·d) was obtained where MW algallipids is the molecular weight of the lipids (and TAG molecules) from the harvested biomass, MW FAME is the molecular weight of fatty acid methyl esters (FAME, biodiesel) (g/mol), η E,TAG is the efficiency of TAG recovery, Y Trans is the efficiency of the transesterification stage, and ∆H • comb,BioD is the heat of combustion of biodiesel (26.7 MJ/kg BioD ). The value of 3 is the stoichiometric coefficient of the FAME product in the transesterification step.
To analyze the biodiesel production, it was proposed that about 80% of the TAG produced was actually recovered. This value may be accomplished using the wet-extraction process [55,56]. Transesterification was then proposed based on a conversion yield of 98%. Note, however, that such values require six moles of alcohol per mole of triglyceride [57]. For the molecular weight of the fatty acid methyl esters (FAME, biodiesel), the calculation was based on the proposed fatty acid profile of Botryococcus braunii, given in Ashokkumar et al. [58]. The corresponding heat of combustion of biodiesel is derived from this value.
Second, the energy from bioethanol via carbohydrate fermentation (E BioE in J/m 2 ·d) is the product of where η E,Sg is the efficiency of fermentable carbohydrate recovery, Y Fer is the bioethanol conversion yield at the fermentation step and ∆H • comb,BioE is the heat of combustion of bioethanol (40.4 MJ/kg BioE ).
The term η E,Sg represents the recovery and acid hydrolysis of the total carbohydrate, according to Karemore and Sen [9]. The fermentation yield from the same work was 0.23 g of bioethanol per gram of fermentable sugar. Bioethanol was produced by converting 90% of the fermentable sugars. The value was reported to be 40% of the theoretical value for the Saccharomyces cerevisiae fermentation.
Finally, the total value of the recoverable fuel energy (J/m 2 ·d) was estimated as

Response of N. gaditana to Nitrogen Depletion in Day-Night Cycles
Investigating the effects of nitrogen depletion on N. gaditana cultivated in day-night cycles (DNc) and batch mode was based on the approach described by Taleb et al. [36] for P. kessleri in a 1 L PBR. For the present study, a plateau was obtained after 168 h. Since the light source remained unchanged (same DNc's) during the whole cultivation time, the plateau may indicate nitrogen depletion (i.e., no growth). The culture was then maintained for 11 days, as in the above-mentioned reference, to allow lipid accumulation during this time.
The final composition of the culture is shown in Table 1. N. gaditana was found to be considerably less productive than P. kessleri under the same culture conditions, with biomass concentrations of 0.9 and 1.3 kg/m 3 , respectively. The energy-storage molecules in N. gaditana attained 43cX: 17% X carbohydrates and 26% X TFA. About 70% of the TFA consisted of TAG molecules at this stage (18% X ). It is interesting to compare the difference in the accumulation preference between energy-storage molecules. Carbohydrate concentration was 0.65 times less than that of TFA in N. gaditana, whereas the carbohydrates in P. kessleri were about 1.3 times more than the concentration of TFA. This difference could be explained by the preference of P. kessleri to accumulate carbohydrates as an energy-storage molecule, to tolerate the dark periods during DNc. P. kessleri was indeed reported to consume up to 10 and 5% X of carbohydrates and TAG storage, respectively, in DNc (reported as non-significant) [36]. In the present work N. gaditana was found to consume only 1.4 and 1% X of carbohydrates and TAG, respectively, for the dark periods between 137 and 167 h (Figure 1). This could also explain the difference in biomass development between the two strains in DNc. Therefore, the fact that the energy-rich compounds of N. gaditana do not vary greatly during the night could be beneficial for biofuel production, because the productivity would not be greatly affected by outdoor cultivation in actual day-night cycles. More detailed studies will be required, since other parameters, such as temperature and night duration, could also influence the consumption dynamics of energy-rich compounds. The areal TAG productivity under DNc obtained in the present work was compared with the results obtained by Taleb et al. [36] for both N. gaditana and P. kessleri, under continuous light and DNc regimes. In the present work, N. gaditana achieved 1.3 × 10 −3 kg/m 2 ·d in TAG at the end of 7 days under batch nitrogen depletion in simulated DNc. This value is around 30% less than the final areal productivity reported for P. kessleri in the same culture conditions (DNc). It can also be observed that the difference between strains is only 15% under continuous light, which could also be related to the night consumption of energy-storage molecules. Figure 2 shows a complementary study of the mechanical resistance of the cells. In batch-cultivated, nitrogen-depleted medium and continuous light (culture data not shown) N. gaditana shows a more marked difference in the cell disruption rate compared to nitrogen-replete conditions. In nitrogen-depleted conditions, cell disruption was 11% (SE 5%), which is 85% less than the disruption rate found in replete conditions [30]. This trend, which indicates an increase in the mechanical resistance of the cells in nitrogen-depleted conditions, has also been observed by Angles et al. [55].

N. gaditana Culture under Nitrogen Limitation and Continuous Mode
Nitrogen limitation applied to continuous culture proved useful in obtaining a biomass rich in the relevant storage molecules, in a single production step [22,59]. The response of N. gaditana to four consecutive levels of nitrogen limitation in continuous mode (56%, 46%, 29% and 13% NO 3 ) was investigated. The results were compared to those obtained for P. kessleri under the same conditions as used by Kandilian et al. [35]. The PBR was operated in continuous mode until steady-state was achieved, to allow the microalgae to adapt to each NO 3 − concentration (around 20 days between each steady state). Continuous light (photon flux density 250 µmol/m 2 ·s) and CO 2 supply were regulated so as not to limit cell growth. The results are given in Table 2; Figure 3 shows the main results.

Nitrogen Uptake and Stress Index
The nitrogen uptake yield (Y NO 3 − /X ) as a value for indicating the ability of the cell to assimilate the surrounding nitrogen sources, and the stress index as an indirect correlation of the C:N ratio [52], were calculated at steady state for each limitation level. Note that all the limited cultures left an undetectable amount of NO 3 − in the medium (<5 mg NO 3 /L, below the detection level), except for the 200% NO 3 limitation. This corroborates the level of nitrogen depletion expected to stress the cultures by limiting the nitrogen source.
The non-limited culture was characterized for the highest nitrogen uptake yield, 411 g NO 3 /g X . The value was reduced to 189 and 201 g NO 3 /g X for the two subsequent limitation values, 56% and 46% NO 3 , respectively. Then, values were reduced again by half, for stronger nitrogen limitations: 105 and 98 g NO 3 /g X at 29% and 13% NO 3 , respectively. Rafay et al. [60] recently calculated the maximum specific nitrate uptake of N. gaditana as 210 ± 11.2 mg NO 3 /g X in a batch culture (F/2 medium for initial NO 3 − 100 mg/L), which falls within the range of values obtained in the present study.
As mentioned above, a decrease in pigment content is another physiological change that takes place in nitrogen-limited cultures [22,61]. Stress levels, along with nitrogen limitation, can be monitored by measuring the cell pigment content [52]. Specifically, the stress index behaved in a similar but inverted trend compared to the previously reported nitrogen uptake values. The representative stress index for the non-limited culture was 0.55. For higher limitations, the value increased progressively to 1.70 at 13% NO 3 . Dortch [62] showed that as soon as external nitrogen starts to limit growth, phytoplankton may accumulate pools of un-assimilated nitrogen, allowing the cell to rapidly react to changes in medium composition (which may be read as the small stress index or C:N ratio). Thereafter, while the surrounding nitrogen concentration remains low, the cells consume this nitrogen pool first, before any other organic nitrogen stock (proteins or amino acids) and then re-adapt to the new environment. This leads to a reduction in the total nitrogen in the cell, increasing the stress index.

Biomass, TAG and Total Carbohydrates
The total biomass concentration of N. gaditana decreased from 2.13 kg/m 3 in nonlimited conditions (200% NO 3 ) to 0.89 kg/m 3 at 13% NO 3 . The total biomass did not change at the same rate as the nitrogen limitation. Reducing the nitrogen concentration from 200 to 56% NO 3 only reduced the biomass concentration by 3%. However, a higher limitation, from 29 to 13% NO 3 , reduced the biomass concentration by 36%. The biomass trend is very similar to that noted with P. kessleri, as reported by Kandilian et al. [35]. The most notable difference between the two strains is related to the level of stress needed to accumulate energy-storage molecules.
When N. gaditana is in favorable growth conditions, a small fraction of the total biomass (2% X TAG) appears as TAG molecules (which are 16% of TFA). In comparison, as can be noted in Table 2, when limitations were introduced, both strains gained up to 24 % X TAG at the highest limitation. However, strains showed differences in nitrogen limitation levels: P. kessleri needed between 29 and 46 % NO 3 limitation to obtain an average 17% X TAG; N. gaditana immediately achieved 15% X TAG at the first limitation value (56 % NO 3 ). This might be a useful optimization indicator for N. gaditana, since low limitation values do not compromise biomass production as much as higher limitations do.
Regarding carbohydrates, N. gaditana does not accumulate as much as P. kessleri (Table 2). In optimal growth conditions, it accumulates 19% X in carbohydrates, which, at this point, is higher than the TAG content. A limitation of 56% X therefore triggers maximum carbohydrate content up to 23% X . At the lowest limitation, only 18% X carbohydrates are accumulated, and at this point there are already more TAG molecules accumulated in the cell. On the other hand, the values reported for P. kessleri indicate that no less than 45% X carbohydrates are achieved for all limitations (maximum 64% X carbohydrates at 46% NO 3 ). It therefore seems that P. kessleri has a marked storage preference for carbohydrates over TAG molecules in all the nitrogen-limited conditions tested here. However, although N. gaditana has a slight preference for carbohydrates at medium limitations (56% and 46% NO 3 ) and more for TAG at stronger limitations (29% and 13% NO 3 ), the difference is balanced between the two energy-storage molecules compared to P. kessleri. Simionato et al. [38] mention the possibility of some sugars being recycled into TAG molecules, which may explain the low carbohydrate content of N. gaditana. Figure 3 shows the areal productivity for biomass, TAG and carbohydrates. The combined effect of the growth and composition of biomass in N. gaditana revealed the maximum levels of carbohydrate and TAG productivity to be equal to 3.41 and 2.19 × 10 −3 kg/m 2 ·d (SE 0. 17 and 0.19), respectively, at 56% NO 3 limitation.
It is also interesting to note that, in large-scale outdoor production, it would be harder to apply the given limitation levels precisely. It is nonetheless important to know the limitation levels that trigger the greatest TAG and carbohydrate productivity

Cell Resistance
The fact that N. gaditana is more resistant to mechanical stress in nitrogen-depleted conditions has been reported previously in the literature [40,43,55] and is confirmed in the present batch experiment. A more detailed evolution of this change at different levels of nitrogen limitation is given in Table 2.
A non-limited culture of N. gaditana can be disrupted at 75% in a single stage through high-pressure homogenization at 1750 bar. This resistance is maintained with a low limitation of 56% NO 3 , but the disruption rate τ D was reduced to 63% at 46% NO 3 limitation, and fell to 17% at the lowest limitation. Cell concentration was rejected as an explanation for this phenomenon since all the samples were prepared at 1 kg/m 3 . An evolution in the cell wall composition and thickness could be the explanation for this [39,40,43,63,64].
The cell resistance of microalgae cultivated in nitrogen-depleted conditions has proven to be an important factor for recovering TAG and carbohydrate molecules in wet-biomass treatment. The values shown here imply that even if productivity is high at some limitation levels, not all the compounds can be recovered if the cell is resistant to the disruption (or more energy is required to obtain sufficient cell disruption). In other words, the final metabolite productivity and overall efficiency of the whole process depends on how easy it is to recover these compounds in this physiological state. The results affect the entire energy investment in the biofuel production process.

Potentially Recoverable Energy
Previous results have shown that the nitrogen limitation levels applied may impact both TAG and carbohydrate productivity in the microalgae cells in different ways. At some levels of limitation, more carbohydrates are produced than TAG, and the reverse is true for others levels of limitation. This could impact the amount of biofuel energy that can be obtained. For this reason, a theoretical analysis of the potentially recoverable energy (E P ) from the two strains shown above, cultivated in nitrogen-limited conditions, is discussed here. This analysis is based on Equation (4), using the values given in Table 3. The results in Figure 4 show the impact that the obtained TAG and carbohydrate productivity have on a theoretical biofuel process, in terms of global energy valorization. Nannochloropsis gaditana achieved maximum energy values of 67.8 and 18.7 J/m 2 ·d for biodiesel and bioethanol, respectively, at the 56% NO 3 limitation level. The corresponding total amount of recoverable energy E P was, therefore, 86.6 J/m 2 ·d. Note that the E P values for N. gaditana did not take into account the changing disruption rates presented in Table 2.
In comparison, the maximum amount of potentially recoverable energy E P from P. kessleri [35] was 128 J/m 2 ·d at the 46% NO 3 limitation level, with 47 J/m 2 ·d accounted for by the energy obtained from bioethanol and 81 J/m 2 · by biodiesel. Note that although the carbohydrate productivity of P. kessleri at 56% NO 3 was almost twice that of N. gaditana at the same limitation level, the total amount of recoverable energy from N. gaditana at the same level was higher due to the greater proportion of biodiesel energy (heat of combustion 40.4 MJ/kg) and also the greater TAG content. In summary, carbohydrate productivity with Nannochloropsis gaditana is not as great as with Parachlorella kessleri. The latter has more potential for dual-biofuel valorization (biodiesel and bioethanol), where more energy can be obtained from the same amount of biomass produced due to the large accumulation of both carbohydrates and TAG. However, N. gaditana may be of interest in terms of biodiesel production in particular, due to its preference for TAG accumulation.

Conclusions
This study has demonstrated the effect that nitrogen limitation has on the physiology of Nannochloropsis gaditana. During day-night cycles, N. gaditana did not show any preference or significant TAG or carbohydrate consumption during the night period. This stability could be beneficial in outdoor cultivation. Under continuous light and nitrogenlimited conditions, the highest carbohydrate and TAG productivity of N. gaditana was 3.41 and 2.19 × 10 −3 kg/m 2 ·d, respectively, for the 56% NO 3 limitation in continuous mode (PFD 250 µmol/m 2 ·s). The disruption rate τ D of N. gaditana decreased significantly by around 77% from non-limited conditions to the lowest limitation level. This trend may compromise the eventual recovery of metabolites. The same analysis was applied to data available in the literature for Parachlorella kessleri. The results showed that under moderate limitation levels, P. kessleri produces larger amounts of energy-storage molecules, which should be easier to recover due to the high cell disruption rate. This, in addition to a high rate of carbohydrate production, is encouraging for the development of both biodiesel and bioethanol. However, N. gaditana remains a promising producer of energy-storage metabolites for biodiesel production, using seawater in particular.