Optimization to Hydrothermal Liquefaction of Low Lipid Content Microalgae Spirulina sp. Using Response Surface Methodology

+e production and nature of the biocrude obtained from Spirulina sp. by hydrothermal liquefaction (HTL) technology is focused in this investigation. Our aim is to evaluate the interaction of different factors on the bio-oil production through HTL using microalgae that contains relatively low lipid content and high protein. Optimization of three key parameters—concentration (mass of algae per mass of solvent), reaction temperature, and holding time—was carried out by response surface methodology (RSM). In this work, we used central composite design to conduct the experiment process. Graphical response surface and contour plots were used to locate the optimum point. +e final results showed that the optimum concentration, temperature, and holding time were 10.5%, 357°C, and 37min, respectively. Under the optimum conditions established, yield of the biocrude (41.6± 2.2%) was experimentally obtained using the fresh microalgae. +is study showed the potential of bio-oil production of Spirulina sp. by HTL technology, but it still needs more improvement of the biocrude for utilization.


Introduction
Microalgae, which could fast convert CO 2 into biomass, have got an increasing interesting role in biofuel production [1].
ey are considered more photosynthetically efficient than any other energy plants [2].
Usually, microalgae contain the lipid in the range of 20-50% [10].e lipid content is dependent upon strain and growth conditions [11].After solvent extraction or physical extraction, these lipids of microalgae can then be further transesterified to biodiesels.One of the problems of this approach is that the wet aquatic biomass requires drying before it can be processed [12].Hydrothermal liquefaction is one of the alternatives being increasingly considered, especially at low temperatures and pressures near the water critical pressure [13].Wet microalgae with high water content could be converted into crude bio-oil by thermally and hydrolytically decomposing the biomacromolecules such as protein and lipid into smaller compounds.e biocrude is an energy dense product that can potentially be used as a substitute for petroleum crudes [14].Some reports showed that hydrothermal liquefaction could be widely applied to various microalgae as the oil yield usually exceeds the crude fat content of microalgae [15].Some of the most productive microalgae in terms of biomass production are lower in lipid and contain larger amounts of protein and carbohydrate [16].Growing these algae for biodiesel is unlikely to be economical, and the alternative-processing routes would be advantageous such as Spirulina.
e Spirulina industry in China is developing rapidly as a national strategic programme [17].By the mid-1990s, China has become the biggest country in Spirulina production in the world.Just in 2009, 3,500 t (dw) of Spirulina have been produced [18].e supply of Spirulina as the functional food has much exceeded the demand.Some studies showed the possibility of producing bio-oil using Spirulina by liquefaction technology [19][20][21][22].If biooil is to be obtained efficiently from mass-cultivated Spirulina by liquefaction, this will be one of the both promising methods for energy production and one of the Spirulina consumption.
To date, there is still lack of the studies about the interaction of multifactors on bio-oil production using Spirulina.
is work focuses on the optimization of hydrothermal liquefaction condition of Spirulina sp. using response surface methodology (RSM).
e influences of process variables containing feedstock concentration, temperature, and holding time have been studied.We hope to evaluate the maximum production rate and further analyze the characteristic of biocrude by using the wet microalgae as the feedstock under the optimal condition.

Strains and Culture Media.
e Spirulina sp.strain was bought from Freshwater Algae Culture Collection at the Institute of Hydrobiology (FACHB-collection).Strains were cultured for three weeks at 25 °C with a continuous illumination of 120 μ•mol•m −2 s −1 in Spirulina medium.Per 1 liter, Spirulina medium contained 13.61 g of NaHCO 3 , 4.03 g of Na 2 CO 3 , 0.50 g of K 2 HPO 4 , 2.50 g of NaNO 3 , 1.00 g of K 2 SO 4 , 1.00 g of NaCl, 0.20 g of MgSO    We measured the moisture of the fresh Spirulina sp. after reducing most of the water by centrifugation, and then, we diluted the sludge into the specific concentration after the optimization calculation.Meanwhile, the lipid and protein content of Spirulina sp.powder and fresh (Figure 1) were measured by the Soxhlet extraction method and Dumas combustion method, respectively.

Elements Analyses and Higher Heating Value (HHV)
Estimate.
e basic elements of the biomass are listed in Table 1.C, H, N, and S contents of the biomass were measured using an elemental analyzer (Flash EA 1112 series, CE Instruments, Italy).All measurements were repeated in triplicate, and a mean value was reported.
Estimation of HHV from the elemental composition of fuel is one of the basic steps in performance modeling and calculations on thermal systems [23].As HHV is an important fuel property which defines the energy content of the fuel, we calculated the biomass and biocrude by the following equations, respectively [24,25]: where C, H, O, and S are the weight percentages of carbon, hydrogen, oxygen, and sulfur, respectively.

Apparatus and Experimental
Procedure.We applied central composite design of three key factors and five levels to simulate our experiment (Design-expert, V8.0, Stat-ease, Inc., USA).374 °C is the critical temperature, a dramatic increase of biomass degradation rate could appear near this critical point owing to the hydrolyze capability of water [26].
And after some single factor trials, the parameters were set like in Table 2. Finally, we obtained sixty biocrude samples from each of the conversion process that were triple duplicated.e hydrothermal liquefaction was performed in a reactor (2 L, Parr, USA) at heating rate of the reactor approximately 2 °C•min −1 .In each case, different weight of microalgae was mixed in deionized water.Microalgae were 2 Journal of Chemistry added into the reactor premixed as slurry.en, the reactor vessel was sealed and nitrogen was introduced to purge the residual air.e microalgae slurry was stirred during the whole process.e speed of magnetic stir bar was set at 100 rpm.Meanwhile, the stir was cooled down by the condense water.e reaction started by heating the autoclave with an electric furnace.After heating the autoclave up to the required temperature, the temperature was maintained constant for the desired holding time, and then, the autoclave was allowed to cool to the room temperature.

Yield.
After the conversion finished, we opened the reactor and dumped the reaction mixture into a beaker.e reactor and stir bar were washed with trichloromethane, and then, they were poured into the beaker too (Figure 2).And then, the solid residue was separated by the glass microfiber filter.e trichloromethane together with the reaction solvent was separated from the water-insoluble substance, and then, the trichloromethane in the mixture was evaporated using the rotary evaporator (RV 10 digital, IKA, Staufen, German) at 60 °C under a vacuum condition.e material remaining in the flask was the biocrude.e weight of biocrude was calculated by using the overall weight of the remaining materials after evaporation subtracting the initial trichloromethane-soluble substrate.e yield of biocrude is determined on a dry basis using the following equation: Yield of biocrude (wt.%) � weight of biocrude weight of algae powder × 100%.
(3) 2.5.FT-IR Analysis.Infrared (IR) spectra for biocrude samples were acquired using a FT-IR spectrometer (4100, JASCO Inc., Tokyo, Japan) to determine the main organic components based on the peaks of the functional groups present.e measurement wavenumber range is 7,800 to 350 cm −1 and resolution is 4 cm −1 , controlled by JASCO's exclusive Spectra Manager ™ cross-platform software.

Constituent Analysis of the Fresh Biomass-Based Bio-Oil.
e biocrude is analyzed by GC/MS on an Agilent 6890N GC/5975B MSD.A volume of 0.5 mL was injected for each sample, and the inlet temperature and split ratio are 300 °C and 3 : 1, respectively.Two minutes solvent delay was set to protect the filament.e column was initially held at 40 °C for 4 min.
e temperature was ramped to 300 °C at 4 °C•min −1 and held isothermally for 4 min, giving a total runtime of about 60 min.Helium flowing at 3 mL•min −1 served as the carrier gas.NIST Mass Spectra Database was used for compound identification.

Sample Workup and Analysis.
e optimization process was carried out to determine the optimum value of bio-oil yield using the Design Expert 8.0 software.e biocrude is a dark viscous liquid.Table 3 shows the yield of each experiment.e yield of biocrude was in the range of 37.2-44.4% and the average was 40.9%.Table 4 shows that the HHV was in the range of 26.7-36.0%and the average was 32.0 MJ•kg −1 .e HHV are much higher than the feedstock.
e oxygen content of microalgal biocrude in our study was a little higher than in some other microalgal liquefaction studies [20,27].e sulfur content of microalgal bio-oil was less than 1% in all cases.
We used the Design Expert software to analyze the experimental results by multiple regressions fitting analysis.Following is the quadric multiple regression equation of yield:  After that, we carried on a significance test of the regression equation.From Table 5, we can see that first degree terms of temperature and concentration are very significant (p < 0.01), the holding time is significant (p < 0.05), and quadratic terms of the three variables are very significant (p < 0.01).e Model F value of 23.02 implies the model is significant.ere is only a 0.01% chance that a "Model F value" this large could occur due to noise.e "Lack of Fit F value" of 2.70 implies the Lack of Fit is not significantly relative to the pure error.ere is a 15.02% chance that a "Lack of Fit F value" this large could occur due to noise.Nonsignificant lack of fit is good, and we want the model to fit.

Graphical Interpretation of the Response Surface Models.
In order to determine the effect of the independent variables on the yield of biocrude, a three-dimensional diagram and contour plot for each response were generated as a function of two variables, while the other one variable was held constant.Figure 3 shows the response surface and contour plots for biocrude yield as a function of concentration (x 1 ) and temperature (x 2 ) with a holding time of 40 min.As can be seen from Figure 3, with the increase of temperature, the yield increases and finally tends towards stability.e yield  rstly increases and then decreases with the increase of the concentration.e highest yield (44.4%) occurs when the concentration and temperature are kept at about 10.5% and 355 °C, respectively.
Figure 4 shows the response surface and contour plots for biocrude yield as a function of temperature (x 2 ) and holding time (x 3 ) with the concentration of 12.5%.Under this condition, the yield rstly increases and then decreases with the increase of temperature or holding time.e yield reaches to the highest (44.4%) when the concentration and holding time are kept at about 354 °C and 38.5 min, respectively.
Figure 5 illustrates the response surface and contour plots for bio-oil yield as a function of concentration (x 1 ) and holding time (x 3 ) with speci c temperature.e yield also shows increase at the beginning and then decrease with the increase of concentration or holding time with the temperature of 345 °C.
e highest yield is 44.4% when the concentration and holding time are kept at about 11% and 38 min, respectively.

Determination of HTL Process of Fresh Spirulina sp. with
Optimal Variables.According to the software optimization step, the desired goal for each operational condition (temperature, holding time, and concentration) was chosen "within the range," while the response (the yield of biocrude) was de ned as "maximum" to achieve the highest performance.
e program combines the individual desirability into a single number and then searches to maximize this function.Accordingly, optimum working conditions are concentration of 10.5%, temperature of 357 °C, and holding time of 37 min.e largest yield is 44.4%.In order to verify the optimal condition, we used cultured microalgae for the further experiment.e moisture of fresh microalgae was 82.1 ± 1.6% after centrifugation.We added some water to the microalgae slurry for the HTL process.e result of three replicated experiments was 41.6 ± 2.2%.e result indicated that when each of the parameters was set as the optimum value, the bio-oil yield was in agreement with the value predicted by the model.It implies that the strategy to optimize the HTL conditions using Spirulina sp. and to obtain the maximal biocrude yield by RSM in this study is feasible.Dimitriadis and Bezergianni showed that the yields from algae by HTL obtained from the literature are more dispersed, but 40% of them render a 45% oil yield [26].Our results indicated that the biocrude yield of Spirulina sp. is a little less than that of some microalgae by HTL, but the performance is not bad, and it is still worth for further analysis.

FT-IR Analysis.
e FT-IR analysis of bio-oil provided results consistent with the GC-MS characterization and the elemental analysis.In the FT-IR analysis, the curves (Figure 6) of the bio-oil obtained under di erent condition showed some di erence.But most of them are similar, suggesting that the same types of functional groups exist in these samples.e composition of the bio-oil is complex as there are many absorbance band and some main bands are shown in Table 6.e bio-oil showed a strong absorbance between 2850 and 3000 cm −1 , indicating a high content of methyl and methylene groups.
ere was also a strong absorbance around between 1500 and 1700 cm −1 , and C O stretching indicated the presence of ketones, aldehydes, esters, or acids.

GC-MS Analysis
. Speci c compositions of the biocrude products were not easily identi able by FT-IR, so the biocrude obtained from fresh Spirulina sp. by the HTL was characterized by GC-MS for identi cation of their chemical compositions too.For a better understanding of the properties of the biocrude, a mass spectral library and computer matching were used to facilitate compound identi cation.Hundreds of peaks were detected by GC-MS in this investigation.e results reveal that the HTL-based microalgal biocrude is an extremely complex mixture of numerous compounds; some major chemical categories are listed in Table 7. e compositions are far di erent from the bio-oil obtained from the Chlorella pyrenoidosa or Dunaliella tertiolecta cake by HTL [27][28][29].Some compounds are similar  Journal of Chemistry with the compounds in bio-oil obtained by hydrothermal liquefaction of Chlorella vulgaris and Spirulina using alkali and organic acids as catalyst [20,30].e nal hydrocarbon product has about 19 wt.% in naphtha range (mainly C 5 -C 10 ) and 84 wt.% in diesel range (mainly C 10 -C 20 ). e hydrocarbons have about 7 wt.%larger than C 20 .
at is probably why the biocrude was a dark viscous liquid.

Conclusion
In this study, bio-oil crude was produced by hydrothermal liquefaction using Spirulina sp.By RSM optimization, we got the optimal reactor condition that was 10.5% of concentration, 37 min of holding time, and 357 °C of temperature and nally got the yield of 41.6 ± 2.2% using the fresh Spirulina sp. as the feedstock.Among the three variables examined in this study, temperature and concentration were the most in uential factors on the product oil yield.
e characteristic of the bio-crude displayed the potential ability of Spirulina sp. to be a valuable and environmentally friendly feedstock candidate for biofuels and biochemicals.As some of the compounds in biocrude, such as fatty acids and aldehydes, are directly related to the undesirable properties of bio-oil, the biocrude still needs more improvement to be used as the liquid fuel.

Figure 3 :
Figure 3: Contour plot and response surface plot of biocrude yield as function of temperature and concentration.

Figure 4 :
Figure 4: Contour plot and response surface plot of biocrude yield as function of temperature and holding time.

Table 1 :
Elemental analysis and HHV estimate.

Table 2 :
Experimental factors and the levels.

Table 3 :
Experimental design and the results.

Table 4 :
e element analysis of each sample.

Table 5 :
Test of significance of the quadratic equation coefficient and ANOVA for the response surface quadratic model.

Table 6 :
e main FT-IR function group composition of biocrude from HTL of Spirulina sp.

Table 7 :
Major chemical compositions of bio-oil from hydrothermal liquefaction of fresh Spirulina sp.biomass at the concentration of 10.5%, temperature of 357 °C, and holding time of 37 min.