Process intensification for the enhancement of growth and chlorophyll molecules of isolated Chlorella thermophila: A systematic experimental and optimization approach

Abstract In our current work, we have optimized six physicochemical parameters (light intensity, light period, pH, inoculum size, culture period, and salt concentration) toward growth and chlorophyll synthesis using isolated fresh water microalgae Chlorella thermophila [contains ∼6% (w/w on dry biomass basis) chlorophyll]. Here, both experimental and computational [Taguchi orthogonal array (TOA), artificial neural network (ANN), and genetic algorithm (GA)] approaches were employed for the process intensification. Results revealed that the content of biomass and chlorophyll were enhanced by 118% and 95%, respectively, with productivity enhancement of 30% for biomass and 61% for chlorophyll from the optimization of physicochemical parameters. Further, optimum light intensity was found to be 128 µmol m−2 s−1 after conducting experiments in optimized chemical and physicochemical conditions, contributing to the enhancement of productivity of 46% for biomass and 106% for chlorophyll. Urea was found to be the most effective nitrogen source with an increase of 70% and 160% biomass and chlorophyll productivity, respectively. Moreover, sucrose as a carbon source contributed to an increase of 97% and 264% biomass and chlorophyll productivity.


Introduction
Microalgae have remained an area of scientific research for eco-friendly renewable sources of biodiesel because of their high content of cellular lipids. [1,2] However, recent reports revealed that commercial biodiesel production as a single product from microalgae is economically infeasible. [3] Therefore, efforts are made to cultivate high cell density microalgal biomass which is used in the hydrothermal liquefaction process to produce bio-oil crudes using the biorefinery concept to produce multiple products. [3] Besides bio-fuels, the attention of microalgae research has switched to producing high-value pigment molecules (e.g., chlorophyll), which have high demand in pharmaceuticals, nutraceuticals, cosmeceuticals, and biotechnological industries. [3] Nowadays, chlorophylls are extracted on a commercial scale from alfalfa, corn, and spinach. [4] It may be noted that Chlorella sp. of microalgae contains substantially high chlorophyll ($7% of dry biomass) than commercial sources of terrestrial plant sources such as alfalfa ($0.2% of dry biomass). [5] Therefore, Chlorella sp. can be used as a potential source of chlorophyll (chlorophyll a and chlorophyll b) that can be used as natural coloring agents in various applications. For example, the use of chlorophyll (chlorophyll a and chlorophyll b) as a food coloring agent has been approved by the Food Safety and Standards (Food Products Standards and Food Additives) Regulations (2011) in various countries like the USA, India, and European countries. [6] Chlorophyll pigment is widely used in resins, coloring inks, cosmetics, liniments, waxes, soap, edible fats, mouth washes, perfumes, lotions, etc. as a cosmetic ingredient. [4] Derivatives of chlorophyll pheophorbide a 17(3)-dimethyl ester, chlorin e6 13(1)-N-methylamide-15(2)-diethylene glycol-17(3)-methyl ester are also used in photodynamic therapy as photo-sensitizer. [7] Further, research has shown the application of chlorophyll in the biomedical field as it has wound healing, antioxidant as well as anti-mutagenic properties. [8] Therefore, industries have shown interest in developing a technology for extracting chlorophyll from microalgae on a commercial scale. [8] To make chlorophyll production at a commercial scale, optimization of cultivation parameters of algal biomass is an important factor to achieve high cell density leading to high chlorophyll productivity. In this direction, chemical parameters (media compositions) and physicochemical parameters influence the microalgae cultivation conditions. Different physicochemical factors include light intensity, light period, pH, inoculum size, and culture period for the cultivation of green microalgae. The rate of photosynthesis is enhanced with increasing light intensity up to intensity saturation leading to a higher growth rate [9] and thereafter, photosynthesis decreases due to photo inhibition. [10] The requirement of optimal light intensity varies within the microalgal species and needs to evaluate for the specific microalgal strain of interest. [11] The light period or dark/light cycle also acts as an influential factor and the optimal light period for each microalgal strain must be determined before large-scale cultivation. According to Khoeyi et al., an increase in light duration by changing the ratio of light phase and dark phase from 8:16 to 16:8 h had a positive impact on biomass productivity. [12] They reported that at 62.5 mmol m À2 s À1 light intensity and 16:8 h photoperiod, the maximum biomass yield was 2.05 g L À1 , while at 37.5 mmol m À2 s À1 light intensity and 8:16 h photoperiod, the minimum biomass yield was 0.6 g L À1 .
In photosynthesis, one phase is a light-dependent photochemical phase and another is a light-independent biochemical dark phase. The molecules that are synthesized in the light phase (NADPH, ATP) are utilized in the dark phase to produce metabolic molecules that are important for microalgal growth. [13] A particular pH is another important factor in achieving the highest microalgal growth rate since pH affects growth by modulating the metabolic activity of the organism. Further, culture periods for microalgal cultivation ought to be accurate for every algal species. Depending on the environment and algal strain, the culture period ranged from 5 to 25 days. However, culture time should be low with high cell density to achieve higher productivity. Inoculum size also has a positive impact on microalgae cultivation. Research had shown that biomass growth and productivity are increased with the increment in inoculum size. [14] However, it increases processing costs leading to an effect on the economics of the productivity and therefore, optimum inoculum size is important. Another factor, salt concentration in the media has unfavorable effects on the microalgae growth and synthesis of chlorophyll. High salt concentrations in culture media impair the dynamic equilibrium between the production and consumption of reactive oxygen species. [15] Oxidative stress influences cell metabolism, limiting cell development and potentially causing cell death. [16] A high level of salt concentration in waste and freshwater algal cultivation media induces the efflux of potassium ion (K þ ) leakage from the cell. The K þ leakage causes a change in membrane ionization and significantly decreased the osmotic tolerance power of the cell which is triggered by K þ present inside the cell. [17] A high salt concentration resulted in the destruction of chlorophyll molecules. [18] It had been reported that chlorophyll content decreased by 33.45% than the control when treated with 0.01 M NaCl and was decreased by 93.42% when treated with 0.20 M NaCl. [18] Therefore, chlorophyll content became substantially low while the salt concentration was high in the cultivation media during the growth of microalgae.
The optimization of microalgae culture process parameters is critical, and different computational techniques were used to optimize the process parameters. The commonly used conventional optimization method ignores interaction among the different experimental parameters (e.g., chemical parameters, physicochemical parameters) even after carrying out many experiments. [19] To resolve these problems, many statistical methodologies such as response surface methodology (RSM) and Taguchi methodology were employed. [20] Himabindu et al. used the RSM method to optimize fermentation medium components for a maximum 110 times greater amount of gentamicin production in Micromono sporaechinospora than in un-optimized conditions. [21] Recently Taguchi orthogonal array (TOA) method is used to design factorial experiments by many researchers because it possesses some advantages over RSM. Only a few experimental trials are required to extract quantitative information, and it also gives a systematic and efficient methodology for conducting tests while taking into account the interacting effects of the control factors. [22] Compare to the RSM method, the Taguchi method is an effective optimization tool and requires half as long as the RSM method. The signal-to-noise ratio (S/N) is a Taguchi tool for determining the influence of each parameter and also the level of each parameter on the process responses. The S/N ratio is used to calculate the ratio of the mean to the undesired value. As a result, this ratio is utilized to assess quality attributes that deviate from the target value. Moreover, the S/N ratio is used to identify the interaction between the parameters. [23] Rao et al. applied Taguchi DOE to optimize the factors for the maximum yield of xylitol from Candida sp. [24] In another study, Prakasham et al. showed that Taguchi DOE is an efficient process for the maximum yield of I-asparagine from isolated Staphylococcus sp. [20] Artificial neural networks (ANN) coupling with genetic algorithms (GAs) are another mathematical tool for optimization. In biotechnology and pharmaceutical technology, the ANN is utilized to solve nonlinear issues such as pattern recognition, optimization, and fermentation, as well as online optimization processes and bioreactor management. [25] Many publications have compared the ANN to statistical designs and found that the ANN outcomes are far superior to the statistical results. [25] GA is focused on unconventional search and optimization algorithms that aid in the search for a solution to a problem by simulating some of the natural processes of evolution. GA conducts a direct random searching process across a set of possibilities to select the best options based on provided goodness of fit criteria expressed as a fitness function. In a number of reports, it is found that 12-14 variables are optimized at a time using GA. [26] Rao et al. reported that a hybrid neural network and GA (ANN-GA) is the superior optimization methodology. According to the results of this study, the use of the ANN-GA method resulted in a 2.5-fold increase in the yield of alkaline protease by isolated Bacillus circulans. [27] It has been reported that mixotrophic cultivation of microalgae with a suitable nitrogen source and carbon source enhance the growth and productivity of metabolites in higher amount. [28] Microalgae cells require a higher amount of nitrogen because nitrogen is a crucial component of chlorophylls, proteins, peptides, enzymes, DNA, RNA, ATP, and other cellular components. [29] Further, light acts as an energy source for photosynthesis and has a direct impact on biomass production and chlorophyll productivity. [30,31] Interestingly, different microalgal species prefer a different type of nitrogen source and carbon source and different light intensity for higher productivity. Therefore, identification of microalgae-specific nitrogen sources, carbon sources, and specified light intensity is required to enhance chlorophyll productivity.
In this current work, six physicochemical parameters (light intensity, light period, culture period, pH, inoculum size, and salt concentration) are optimized to maximize the biomass and chlorophyll content and productivity using isolated Chlorella thermophila. Experiments were designed using Taguchi orthogonal array (TOA) design experiment for the optimization of physicochemical factors and experiments were performed. Thereafter, TOA was employed using experimental data for the optimization of physicochemical parameters for the enhancement of the content and productivity of biomass and chlorophyll. Thereafter, the ANN-GA hybrid was performed for further optimization process for the enhancement of the content and productivity of biomass and chlorophyll. After optimizing the physicochemical factors, we have studied the effects of key influence factors such as light intensity, different carbon sources (e.g., dextrose, fructose, lactose, sucrose, cellulose, sodium acetate, rice powder, rice straw, and banana stem) and nitrogen sources (e.g., ammonium sulfate, ammonium chloride, ammonium acetate, sodium nitrate, glycine, and urea) individually on biomass and chlorophyll productivity.

Materials and methods
Materials and methods are divided into five subheadings (supplementary Figure S1) as per the need of the experimental and computational methodologies which are discussed below in detail.

Microalgae strain preservation, chemicals, and instruments
An isolated C. thermophila microalgae strain was employed in this study. This microalgal strain was rich in high chlorophyll content ($6%). [13] After isolation, this isolated axenic culture of microalgae was submitted to the National Repository for Cyanobacteria and Microalgae (Fresh water), Department of Biotechnology, Institute of Bio-resources and sustainable development (IBSD), Imphal, Manipur, India with BTA 9035 accession number. [13] This isolated microalga was stored at 19 C, 29 mmol m À2 s À1 of light intensity with an 8:16 day-night ratio and used for all experiments.
Extraction of chlorophyll was performed using methanol (Hi-media laboratories, India). Other chemicals were procured from Merck, India. Estimation of biomass and extracted chlorophyll was determined by UV-visible spectrophotometer (Evolution 201, Thermo Fisher Scientific, USA). Centrifuge (2-16KL, Sigma, Germany) was used for separation. The statistical software package Qualitek-4 software (Nutek Inc., Grand Rapids, MI, USA) was used to create experimental design and analyze experimental results. The ANN and GA toolbox of MATLAB (version 8.5, Mathworks, Natick, MA) was employed to build the modeling and for the further optimization of the factors.

Morphological analysis by optical microscopy and scanning electron microscopy (SEM)
A sample of cultured microalgae was taken after 15 days of culturing. An optical microscope (Olympus, model: BX 53) was used to analyze morphological parameters at a magnification of 100Â in differential interference contrast (DIC) mode.
An ultra-structure morphological investigation was carried out with the help of a scanning electron microscope (SEM) (Nova NanoSEM 450, FEI, USA). The cultivated microalgae sample was first rinsed with a 0.1 M phosphate buffer at pH 7.0 for this examination. After that, the sample was fixed with 2.5% formaldehyde at 4 C for 24 h. Thereafter, the sample was dehydrated by washing it with 50%, 70%, 80%, 90%, 95%, and 100% ethanol for 20 min each. [32] The sample was dehydrated at room temperature after being accumulated on an aluminum metal slab with a conductive double-sided adhesive tape made up of carbon. Following that, a sample preparation system (Q150R, Quorum Technologies, UK) was used to coat the sample with a thin layer of gold.

Biomass growth
Seed culture (100 mL) with BG 11 media was inoculated from 1% of the stock culture of the isolated microalgal strain. Thereafter, seed culture was cultivated in a 250 mL conical flask maintaining 28 C, 74 mmol m À2 s À1 light intensity, and 14:10 day-night ratio. [33] Then, applying L-18 TOA, 18 experimental sets were performed to optimize six physicochemical parameters at three levels. For 18 sets of experiments, 2% of seed culture was transferred into 50 mL growth media in a 100 mL conical flask and cultivated for 15 days with the same conditions mentioned for seed culture. These cultivated cultures were used for measuring growth and chlorophyll content. For growth measurement, a known amount of (2 mL) of the sample was drawn from the cultivated culture and a biomass pellet was obtained by discarding the supernatant after centrifugation at 10,000 rpm for 10 min. Thereafter, biomass was dissolved in an equal amount of water and optical density (OD) was measured at 680 nm using a UV visible spectrometer. The biomass content (g-dry biomass L À1 ) of the culture was determined using a previously developed relationship between dry cell weight and optical density. [34] Biomass productivity (g L À1 day À1 ) was determined using the differences in biomass concentration over a specific cultivation period by using the following equation: Where x 1 and x 0 are the biomass content (g-dry biomass L À1 ) on sampling day (t 1 ) and initial day (t 0 ) respectively.

Total chlorophyll content
For the quantification of chlorophyll content, 10 mL of cultivated algal sample was taken and centrifuged at 10,000 rpm for 5 min at 4 C. [35] The supernatant was removed, and the collected pellet was treated with 10 mL of 99.9% pure methanol. The mixture was stored at 45 C for 24 h in a closed container shaking incubator at 150 rpm. [36] The absorbance of the supernatant was determined using a UV visible spectrometer at 652.4 and 665.2 nm after the incubation period. The actual content of the pigment molecule was measured using the following formulas: Chlorophyll a lg mL À1 À Á

Design of experiment, modeling, and optimization
In the present study, TOA was used for the optimization of physicochemical factors for the improvement of the content and productivity of biomass and chlorophyll. Then ANN modeling was performed using experimental data and the output weights and bias of ANN were used for further optimization using the GA (supplementary Figure S2). The first step of the TOA was the formulation of the problem where the objective function, factors, and levels were defined. This methodology consisted of four phases such as planning of the experiment, conduction, analysis, and validation of the experiments. Each phase had a different goal and was connected in a sequential order to complete the entire optimization process. The results obtained after the successful execution of the experiments were analyzed by determining the S/N ratio analysis and ANOVA analysis. After analyzing the experimental data, the optimum level of the factors was identified. Validations of the experiments were conducted for confirmation (supplementary Figure S2a). Beside TOA, ANN was also used to construct a non-statistical model to reduce experimental error with back-propagation. After getting a satisfactory result from ANN, GA was used to predict the optimum experimental parameters for the maximum result (supplementary Figure S2b).

Taguchi orthogonal array (TOA) methodology
Taguchi's approach entailed orthogonal arrays to create several experimental conditions to minimize experimental mistakes and improve the reproducibility and efficiency of the experiment. The robust design was used in this study because it reduces the impact of noise factors during the optimization process, resulting in a dynamic or robust experimental design. TOA was used to optimize various factors that had a noteworthy impact on the growth of biomass and productivity of chlorophyll from microalgae. All of the parameters were investigated within their feasible ranges, such that the intrinsic fluctuation in the process did not obscure the effects of the parameters. In this work, six physicochemical factors namely light intensity, light period, pH, inoculum size, culture period, and salt concentration were considered important factors for the growth and chlorophyll productivity of microalgae. All these factors were investigated at three levels shown in Table 1. The matrix was then developed with appropriate OAs to select the parameters and their levels for the next stage. Three levels of six factors were investigated in this investigation, and the size of the experiment was represented by a symbolic array of L-18, indicating 18 experimental trials ( Table 2). The data obtained from the experiments were analyzed and performance was predicted by using Qualitek-4 software (Trial version, Nutek Inc., MI, USA). The influence of individual parameters, optimum conditions of the parameters, and the process performance were determined by using this software. In this study, quality characteristics (QC) or S/N analysis was used to estimate performance at the optimum conditions for all experimental cases using the bigger-is-better principle.
In the Taguchi technique, the term "signal" refers to the desired value while "noise" refers to the undesirable value for the output values. The S/N ratio represents the ratio of the desired value with the undesirable value. To estimate the resilience of the overall process, Taguchi employed the S/N ratio as a performance evaluation of a dynamic system. The following equation gives the mathematical equation for the S/N ratio in the "bigger is better" condition for performance statistics that quantify variation from the target, called mean square deviation (MSD). A smaller value of MSD was desired for bigger values of results (Y). The use of multiplier 10 in Eq. (5) was arbitrary, and the negative sign assured higher S/N for reduced MSD/variability: The optimized cultivation conditions were then assessed by completing experiments in order to validate the optimized methods.

Modeling by artificial neural network (ANN)
Here, a multilayer neural network was used consisting of input layers, output layers, and hidden layers. In this analysis, one input layer of six neurons, one hidden layer of five neurons, and one output layer of one neuron (6-5-1) were used to determine the performance parameters as illustrated in supplementary Figure S3. It is worth noting that this topology was repeated for four outputs (biomass content, biomass productivity, chlorophyll content, and chlorophyll productivity). The input layer of six neurons was comprised of six input variables such as light intensity, light period, culture period, inoculum size, and salt concentration. In the input and output layer, this network also contained additional neurons, known as a bias for the input layer (B I ) and bias for the hidden layer (B H ). All the nodes present in each layer were linked to subsequent layer neurons. Synaptic weights, such as weight on the link between input neurons and hidden neurons (W H ) and weight on the connection between hidden neurons and output neurons (W O ), were used to connect the input and secret layers, as well as the hidden and output layers (weight on the connection between hidden and output neurons). When defining the most relevant variable in the process to be modeled, the efficacy of any ANN model relies on a high-quality view of the problem being solved. The number of hidden neurons was determined by stepwise increase till the greatest correlation was achieved. In this study, 5 hidden neurons were used. If a network was allowed to train for an extended period of time, it will over-train and would lose its capability because of increasing process noise. A total of 54 experiments were carried out in this study; 38 runs ($70%) were chosen for training and the remaining 16 runs ($30%) were used for testing. As per Zhu et al. and Coleman et al., all the data in the input and output layer was normalized within a uniform range (À1, 1) before training to avoid any type of numerical overflow. [37] In this investigation, the feed-forward neural network was utilized which did not have a connection back from the output to the input neurons and hence did not retain track of its prior output values. Information from the input was feed-forwarded through the network for the optimization of the weights between neurons. In this study, one hidden layer feed-forward back-propagation along with the Levenberg-Marquardt training algorithm (TRAINLM) was utilized to train the network which offered faster solutions because of the inclusion of an additional second derivative of error information and automated internal design changes to the learning parameters. In this specified feed-forward network, the Trans-sig activation function was selected for the hidden and Purelin for output neurons since it improved performance over other functions throughout the development of the perfect model because it was a nonlinear and continuous differentiable function. [38] Nonlinearities were incorporated into neural networks for the selection of the network parameters, by utilizing activation functions, in order to make them more effectual than linear transitions. [38] This nonlinear function mapping device determined the N-dimensional nonlinear function vector (f) f : X ! Y, X was a set of "n" number of input vectors, x ¼ ½x n ; n ¼ 1, 2, … .N and x ¼ ½x 1 , x 2 , ::::x n T and Y was a set of corresponding output vectors (biomass content, productivity, and chlorophyll content and productivity). Y ¼ ½y n ; n ¼ 1, 2, :::::::N and y ¼ y 1 , y 2 , ::::, y N ½ Initially, weight and bias values were taken randomly. ANN read the input and output values in the training data set and changed the values of the weighted links to minimize the difference between the predicted and target values. The error was reduced by many training cycles until the network reached a required level of accuracy to achieve the similarity in training and testing experimental titer values. The neural network training resulted in the following equation, which relates the input to an output variable in terms of weights and biases: The percent participation of each specified factor was determined using factor fitness scores acquired during network training, and the contribution of each factor was calculated using the following formulae.

¼
Factor fitness score Total factors score Â 100 (7) The accuracy of the model was determined by the regression coefficient or Pearson product-moment correlation coefficient (R), the absolute function of variance (R 2 ). The inaccuracy was determined to assess the modeling performance of ANN based on the difference between experimental and predicted values such as mean square relative error (MSRE), root mean square error (RMSE), mean absolute percentage error (MAPE), and mean square error (MSE). [34] Optimization by genetic algorithm (GA) ANN output values were used to optimize GA using Eqs. (6) and (7). Then the subsequent optimization of biomass cultivation and productivity and also chlorophyll content and productivity was achieved by using GA. GA was applied to investigate and optimize the maximum of a function over a certain domain space. Chromosomes were bit strings that encode points in the search domain space (generally real numbers). Each bit position in the string was called a gene and each starting solution (here six physicochemical parameters) was known as population. Every population was assessed for their fitness during the problem-solving process. Three genetic processes were used to establish a new population of chromosomes from the previous population: reproduction (creation of new population), crossover (exchange of bits in a pair of chromosomes), and mutation (flipping of bits in offspring) Each GA output simulation was utilized to investigate in different subspaces and to positioned the global maximum on the objective function surface. By establishing the lower and upper boundaries, the underlying optimization objective was employed to discover the Ldimensional optimal decision variables ( Table 1) (x n L < x n < x n U , l ¼ 1, 2, 3, … , L). x denoted the optimizing parameters (L ¼ 6) and x n L and x n U represented the lower and upper bounds on x n . Based on literature studies, different GA optimization parameters such as chromosomal length of 200, a population size of 200, crossover probability of 0.8, and mutation probability of 0.01 were assumed in this work. [23] After evaluating GA for 500 generations, the optimal parameters for achieving maximum chlorophyll and biomass content and productivity were determined in the given range of input variables. Neural networks and GA toolboxes of MATLAB (version 8.5, Mathworks, Natick, MA) were used in this study.
Cultivation of optimized microalgae under different light intensity, carbon and nitrogen sources for the maximization of biomass and chlorophyll productivity Here, we were interested to determine the influence of light intensity, different carbon sources, and nitrogen sources since these parameters may play a crucial role in the growth of biomass and chlorophyll productivity. To study the influence of these parameters, we have used optimized media parameters (evaluated in our previous work [39] ) and optimized physicochemical parameters from this current work. For the investigation of the effect of light intensity on growth and chlorophyll productivity, seven different light intensities (37,78,128,168,292,392, and 478 mmol m À2 s À1 ) were applied using optimized media and physicochemical conditions. Another important factor was the nitrogen sources which would require a high amount for the anabolism of nitrogenous compounds (e.g., proteins, genetic materials) in growing microalgal cells. [29] Thus, in another set of experiments, different nitrogen sources such as ammonium sulfate, ammonium chloride, ammonium acetate, sodium nitrate, glycine, and urea were employed to investigate the influential effect of these nitrogen sources on biomass and chlorophyll productivity. Further, carbon was an essential component of cell division, and physiological activities promote biomass growth. In the photoautotrophic culture of microalgae, there are two ways to provide carbon to the media. Carbon can be supplied in two forms: gaseous CO 2 or organic/inorganic sources. [40] In this study, different carbon sources were used (e.g., dextrose, fructose, lactose, sucrose, cellulose, sodium acetate, and other organic sources from biomass such as rice powder, rice straw, and banana stem extract) to study the influence of different types of carbon sources on biomass and chlorophyll productivity.
The influential effect of different nitrogen sources was observed by applying different nitrogen sources in the optimized BG11 medium with equimolar concentration (0.018 M nitrogen) of ammonium sulfate, ammonium chloride, ammonium acetate, sodium nitrate, glycine, and urea. After that, the significant effect of different carbon sources was observed by performing the experiment with different carbon sources such as dextrose, fructose, lactose, sucrose, cellulose, sodium acetate with equimolar concentration of 6 M carbon and some other organic carbon sources derived from organic biomass such as rice powder, rice straw and banana stem with 60 g L À1 . To study the effect of different nitrogen and carbon sources, the experimental process was carried out under the light intensity of 44 mmol m À2 s À1 with day-night ratio of 14:10 h at pH 9. Then after 15 days of the culture period, biomass and chlorophyll productivity was measured for every set of the experiments.

Results and discussion
Morphological characteristics by optical microscopy and scanning electron microscopy (SEM) Isolated C. thermophila (BTA 9035) was observed under optical microscopy (Olympus, BX-53) to specify the morphological properties of the cells (Figure 1a) which show the same morphological properties as Chlorella sp. The optical observation revealed that these cells are 3 $ 4 mm in size with spherical to oval shapes under 100Â magnification in differential interference contrast (DIC) mode. These nonflagellated microalgae cells with a single parietal chloroplast also formed clusters or existed in a solitary state.
An image of an isolated microalgae strain (BTA 9035) was captured with a scanning electron microscope at 15 KV with 12,000 times magnification (Figure 1b). SEM image had shown the microalgae cells were about 4 mm long and ellipsoidal. It was also found that cells were trapped together on the surface and their cell surface is smooth. The appearance of the cells was found non-flagellated and nonfilamentous from the SEM image.

Influence of physicochemical factors
To investigate the significant effects of physicochemical factors on microalgal growth and chlorophyll synthesis, experiments were performed using Chlorella thermophilus as per the experimental design using TOA (Section "Taguchi orthogonal array (TOA) methodology") and the results were represented in Table 2. Experimental values revealed a significant variation in the content and productivity of biomass and chlorophyll. The effect of individual physicochemical parameters and their interactions play a significant role in  growth (Figures 2 and 3; Table 3) and chlorophyll synthesis (Figures 4 and 5; Table 3) and details are provided next.

Influence of physicochemical factors on biomass content and productivity
After experimental data was obtained from DOE (Table 2), the S/N ratio was calculated using TOA simulation. The calculated S/N ratio is represented against three levels to investigate the influential effect of six physicochemical factors on biomass content and productivity (Figures 2 and 3). Table 3 also provides quantitative information on the effect of six physicochemical factors on biomass content and productivity. The magnitude of difference between level 2 and level 1 (L2-L1) and level 3 and level 2 (L3-L2) represents the relative influence of the factors at different levels ( Table 3). The larger the difference between L2-L1 and L3-L2 indicates the stronger influence. [41] The positive influence on biomass content in the order as LI > pH > CP > LP > IS with the negative influence of salt concentration. Similarly, positive influence on biomass productivity in the order as LI > pH > LP > CP > IS with the negative influence of salt concentration. Thus, simulation results revealed that light intensity had the highest impact on biomass growth ( Figure  2a) and productivity (Figure 3a) at level 2 (Table 3). Thereafter, the effect of light intensity started to decrease at level 3 on both biomass content and productivity. The less biomass content and productivity can be explained due to an inhibitory effect of higher light intensity. [42] Similarly, the light period has an optimum effect at level 2 on both biomass content (Figure 2b) and biomass productivity (Figure 3b; Table 3). Then there is no significant increase in biomass content and productivity on level 3. Interestingly, pH had a positive influential effect on both biomass content ( Figure 2c) and biomass productivity (Figure 3c) because level 3 has the highest S/N ratio (Table 3). Therefore, simulation results confirmed that alkaline pH was necessary for this microalgae strain for higher growth. A similar trend was noticed in the case of inoculum size where the S/N ratio at level 3 of inoculum size produced the highest biomass content (Figure 2d) and biomass productivity (Figure 3d; Table 3). This study also revealed that increasing the culture period leads to the enhancement of biomass content ( Figure  2e). But, in the case of biomass productivity, the culture period showed a linear pattern of effect at their respective levels indicating the culture period was less significant for biomass productivity (Figure 3e; Table 3). However, an increased amount of salt concentration causes a negative effect on both biomass content ( Figure 2f) and productivity (Figure 3f; Table 3).
This simulation study confirms that specific light intensity is essential for optimum growth and productivity of microalgae because light intensity is one of the most significant factors for growth. [43] Microalgae growth enhances up to the saturation point of light intensity and thereafter, growth is inhibited due to photo-inhibition. [43] Besides light intensity, the light period or dark/light cycle also acts as an influential factor in the growth of biomass and productivity. A specific light period is required for a photochemical phase to synthesize sugar, NADPH, and ATP. On the other hand, dark respiration is the light-independent process by which sugar and oxygen are utilized to provide energy for growth and maintenance. [12] Experimental investigation revealed that an increase in light duration was directly proportional to an increase in microalgal growth. [12] Wahidin et al. applied three different photoperiods (24:0, 18:6, and 12:12 h day-night ratio) and 3 different light intensities (50, 100, and 200 mmol m À2 s À1 ) to show the growth of biomass and lipid contents on microalgae. Continuous illumination in a photo bioreactor system was widely applied to improve biomass production. On the other hand, excessive light energy cannot be transformed into chemical energy, resulting in photo inhibition, which inhibits algal cell growth. [44] In our simulation study, significant growth (L2-L1 ¼ 1.59, Table 3) was observed while the light period increased from 10 to 14 h. Thereafter, less significant growth (L3-L2 ¼ 0.27, Table  3) was found by increasing the light period (increased from 14 to 18 h). This result can be explained by the photo inhibition in microalgae at higher light periods. An optimum pH is another key factor to obtain the maximum microalgal growth. Favorable growth of microalgae can be observed at pH (7.0-9.0). [45] In an alkaline environment, the availability of OH À ions would be more which would react with CO 2 to form HCO 3 À and CO 3 2À , resulting in higher availability of carbon sources in the form of bicarbonate and carbonate in microalgal growth media. [46] Further, the bioavailability of nitrogen sources and phosphorus sources along with other nutrients would be in a favorable condition for the microalgae species in alkaline pH. [45] Therefore, the ease of transport of these nutrients into the microalgal cells would facilitate the growth of microalgae cells. It was also reported that alkaline pH would require for the cultivation of the microalgae otherwise cells can collapse and unexpected cell death may rise. [45] In our study, alkaline pH (pH ¼ 9) showed an optimum influential effect on the content and productivity of biomass (Figures 2c and 3c). In our simulation study, we found that an increased amount of salt concentration caused a negative effect on biomass content and biomass productivity (Figures 2f and 3f). Our observation was supported by other investigations where the growth of microalgae cells was hindered by high concentrations of NaCl. Salinity can induce cell death because high salt concentration disrupts the dynamic equilibrium between the generation and consumption of reactive oxygen species (ROS), resulting in oxidative stress. Oxidative stress in the cells damage different cellular organelles (nucleic acids, proteins) that would affect the metabolism of the cell, resulting in cell growth inhibition and even may cause cell death. [16] Influence of physicochemical factors on chlorophyll content and productivity Similar to biomass content and productivity, the S/N ratio was calculated for the determination of the quantitative effect of six physicochemical factors at their three respective levels for chlorophyll content and productivity (Figures 4 and 5; Table 3). In our simulation study, the influence of light intensity was found highest at level 2 (Table 3) for both chlorophyll content and productivity (Figures 4a and  5a). In the case of the light period, the linear pattern of the effect was found for the three respective levels of chlorophyll content (Figure 4b; Table 3) while chlorophyll productivity increased with the increasing levels of the light period ( Figure 5b; Table 3). It is worth to mention here mentioning that, the effect of pH (5-9) had the highest impact at level 3 Table 3.  ).
on chlorophyll content (Figure 4c; Table 3) and productivity ( Figure 5c; Table 3). On the other hand, the effect of inoculum size (2-8% v/v) was highest at level 2 then it showed a slightly declined pattern of effect at level 3 for both chlorophyll content (Figure 4d; Table 3) and productivity ( Figure  5d; Table 3). The culture period (14-28 days) had the highest effect at level 2 for chlorophyll content (Figure 4e; Table 3) but in the case of chlorophyll productivity, level 3 showed the highest influence (Figure 5e; Table 3). Results revealed that increasing salt concentration caused a negative impact on both chlorophyll content (Figure 4f; Table 3) and productivity (Figure 5f; Table 3). This finding indicates that  chlorophyll content and productivity decreased with the increasing salt concentration.
Microalgae require a specific amount of light energy for photosynthesis to produce ATP and NADPH. [47] The rate of the light-dependent reaction increases with increasing light intensities and hence photosynthesis increases proportionally. More photons would cause ionization of more chlorophyll molecules attributing to more generation of ATP and NADPH. If the light intensity would increase further, however, the rate of photosynthesis eventually would decrease due to damage to chlorophyll. In our study, 44 mmol m À2 s À1 of light intensity was optimum for the synthesis of chlorophyll by cultivating the C. thermophila. An optimum light period is essential for photosynthesis which plays an important role to achieve a higher amount of chlorophyll content and productivity. It has been reported that a continuous light period causes photo-inhibition of algal photosynthetic apparatus. [44] In our simulation study, chlorophyll synthesis was highest at 18 h light exposure. The protein and pigment content of microalgae is substantially influenced by the media pH. Scientific reports revealed that acidic pH would cause the degradation of chlorophyll molecules. [45] Our simulation study supported the previous experimental observations and chlorophyll synthesis was observed at alkaline pH (pH ¼ 9) for our isolated C. thermophila. Earlier studies had suggested that high salt concentration had an inhibitory effect on chlorophyll content. With a high concentration of NaCl (30-60 g L À1 ), the activity of chlorophyllase enzyme (chlorophyllase a and chlorophyllase b enzyme) is enhanced. [48] This active chlorophyllase a and chlorophyllase b enzyme would destroy the chlorophyll molecule resulting in a decreased amount of chlorophyll. Our simulation results follow a similar trend reported in the literature. [48] Influences for the interactions of the six physicochemical factors The information about the interaction of many factors would be important due to their important role to enhance productivity. A factor might interact with any or all of the other factors by introducing the probability of many interactions. [49] The influences of interactions of two factors were determined by means of severity indexes (SI) which were evaluated by TOA and values of SI were presented in Table 4. Higher the SI percentage was attributed to the large influence of the interaction parameters. In Table 4, "Levels" represented the desired levels of factors for optimized conditions. In the case of biomass content, the highest SI interaction (65.13%) was found between light intensity and inoculum size (at levels 2 and 1) followed by light intensity and light period (SI 61.39%), light period and culture period (SI 33.22%), and culture period and salt concentration (SI 27.55%) respectively in descending order (Table 4). It may be noted that the individual contribution of light intensity on biomass content was highest while the individual contribution of inoculum size had the lowest. However, the combination of light intensity with level 2 and inoculum size having level 1 resulted in the highest SI (SI 65.13%). Similarly, in the case of biomass productivity, the highest SI interaction (72.1%) was observed between light intensity (high impact factor individually) and inoculums size (lowest impact factor individually) eventually followed by light intensity and light period (SI 67.95%), light period and culture period (SI 38.06%). For the chlorophyll content, the highest interaction (SI 71.51%) was observed between light intensity (high impact factor individually) and pH (lowest impact factor individually). However, in the case of chlorophyll productivity, two relatively high impact factors at their individual level were light intensity and inoculum size (highest interaction (SI 65.06%)). It was evident from the simulation result that, the content and productivity of biomass had the same trend in the influence of the interaction of parameters. However, the interaction of parameters of chlorophyll content and productivity showed a different trend. This might be due to the chlorophyll content per unit of biomass, which was the inherent property of chlorophyll synthesis depending on the specified physicochemical factors. However, chlorophyll productivity was dependent on both chlorophyll content and biomass productivity and therefore, Abbreviations: LI: light intensity (mmol m À2 s À1 ); LP: light period (hr), pH; IS: inoculum size (%, v/v); CP: culture period (days); SC: salt concentration (g L À1 ).
the order of interaction of parameters of chlorophyll productivity was the coupling of biomass productivity and chlorophyll content.

Analysis of variance
The TOA-designed experiments were statistically analyzed using analysis of variance (ANOVA) to determine the partial contribution of each parameter to the responses ( Table 5). The F-ratio was used to distinguish the significant factors from those less significant factors statistically significant within a 95% confidence level. [50] Results showed that the salt concentration contributed the maximum impact on biomass content (77.6%). The next levels of significant factors for biomass contents are light intensity (8.4%), culture period (6.1%), pH (2.0%), and inoculum size (0.6%) in descending order. For the biomass productivity, salt concentration has maximum contribution (82.4%) followed by light intensity (9.0%), pH (2.2%), inoculums size (0.6%) and culture period (0.1%). Like biomass content and productivity, salt concentration also had a negative impact on chlorophyll content (17.4%) and productivity (77.0%). The next significant factors for the chlorophyll content in order of importance are inoculum size (13.0%), pH (5.4%), and culture period (5.2%). In the case of chlorophyll productivity, the next level of significant factors is light intensity (8.0%), pH (3.9%), and inoculum size (1.4%). In summary, apart from salt concentration, light intensity (LI) had the highest contribution to growth (both biomass content and productivity) and chlorophyll productivity while the contribution of inoculum size (IS) was highest for chlorophyll content.

Optimum process parameters and experimental validation of Taguchi OA DOE
The optimum conditions in terms of contribution to achieving higher biomass and chlorophyll content and biomass and chlorophyll productivity have been shown in Table 6. Level description suggested that the higher level of biomass content and productivity can be achieved with light intensity (44 mmol m À2 s À1 ), light period (18 h), pH (9), inoculum size (8 mL), and culture period (28 days) in the absence salt concentration in the culture media. The contribution of the S/N ratio of light intensity was found to be 2.7 after salt concentration in both cases with a total contribution from all the factors of 22.0 for biomass content and 19.1 for biomass productivity. By transforming the S/N ratio in terms of quality characteristics unit (QC), biomass content and productivity were 1.37 g L À1 and 48.69 mg L À1 respectively. Table 6 also represented the expected result of chlorophyll content and productivity at an optimum condition in S/N ratio is 38.5 and 10.6 respectively. The total contribution from all the factors for chlorophyll content and productivity was 6.9 and 24.4, respectively. By the conversion of the QC factor, chlorophyll content was achieved at 84.49 mg g À1 biomass, and chlorophyll productivity was 3.37 mg L À1 day À1 at optimum conditions. After performing validation experiments of predicted values from TOA, it was observed that the biomass content and productivity were 1.2 g L À1 , 46.49 mg L À1 day À1 while the chlorophyll content and productivity were recorded as 82.83 mg g À1 biomass and 2.58 mg L À1 day À1 , respectively (Table 7). Experimental results were within the acceptable limit of TOA predicted values indicating the efficient capability of TOA for the optimization problem. It might be noted that approximately 118.2% increase in biomass content, 23.1% increase in biomass productivity, 94.5% increase in chlorophyll content, and 61.3% increase in chlorophyll productivity were achieved compared to the standard physicochemical conditions in BG-11 media.

Prediction and experimental validation of ANN modeling and GA optimization
The maximum R-value and minimum MSE, and MAPE values were achieved from a network with a hidden layer of 5 neurons during training of the ANN network with six input variables for biomass and chlorophyll. Therefore, a topology of (6-5-1) was selected for both biomass and chlorophyll (supplementary Figure S3). It may be noted that input variables were light intensity, light period, pH, inoculum size, culture period, and salt concentration. The overall R-value for biomass content and biomass productivity was observed to be 0.976 and 0.996, wherein 0.977 and 0.995 for training, 0.990 and 0.991 for testing, and 1 and 0.999 for validation of the proposed model, respectively (supplementary Figures  S4 and S5). Similar satisfactory overall R-values were found in the case of chlorophyll content and productivity ( Figures  S6 and S7). The reliability of the proposed model was further evaluated with other statistical parameters [mean squared error (MSE), mean absolute percentage error Abbreviations: LI: light intensity (mmol m À2 s À1 ); LP: light period (hr), pH; IS: inoculum size (%, v/v), CP: culture period (days); SC: salt concentration (g L À1 ). Table 6. Optimum culture conditions and their contribution to achieving higher content and productivity of biomass and chlorophyll.
(MAPE), mean squared relative error (MSRE), and root mean squared error (RMSE)] and found satisfactory (supplementary Table S1) supplementary Figure S8 showed the correlation between experimentally measured and model-predicted values for the biomass content (supplementary Figure S8a), biomass productivity (supplementary Figure S8b), chlorophyll content (supplementary Figure S8c) and chlorophyll productivity (supplementary Figure S8d). The R 2 was found to be 0.98 for biomass content, 0.99 for biomass productivity, 0.95 for chlorophyll content, and 0.97 for chlorophyll productivity. All these statistical analyses revealed the developed ANN model had excellent predictive capabilities.
The ANN output data (weight and bias) was used to evaluate the optimized physicochemical conditions for the maximum biomass and chlorophyll synthesis using GA. Table 6 depicts the most excellent possible conditions obtained after performing various ANN-GA trails. From Table 6, it was observed that maximum predicted biomass content (0.89 g L À1 ) from the ANN-GA model could be achieved with light intensity (74 mmol m À2 s À1 ), light period (10 h), pH (9), inoculums size (8 mL), culture period (28 days) and without any salt concentration in the culture media. The ANN-GA predicted optimized culture conditions for biomass productivity was 52.25 mg L À1 day À1 ) with light intensity (68 mmol m À2 s À1 , light period (10 h), pH (9), inoculums size (8 mL), culture period (28 days) with no salt concentration in the media. Table 6 also depicted the maximum of 77.95 mg g À1 biomass chlorophyll content and 2.2 mg L À1 day À1 chlorophyll productivity that could be achieved by ANN-GA predicted optimized parameters. For the chlorophyll content, ANN-GA predicted optimized parameters were as light intensity (50 mmol m À2 s À1 ), light period (10 h), pH (5), inoculum size (2 mL), culture period (14 days), and salt concentration (58.7 g L À1 ) and maximum amount of chlorophyll productivity could be achieved with light intensity (39 mmol m À2 s À1 ), light period (18 h), pH (8.9), inoculums size (7.9 mL), culture period (14 days) without salt concentration. After the validation experiments, 0.88 g L À1 biomass content and 49.55 mg L À1 day À1 biomass productivity were achieved (Table 7), which was close to ANN-GA optimized biomass content and productivity. The experimental values of model validation were recorded for chlorophyll content and productivity and found as 74.63 mg g À1 biomass and 1.98 mg L À1 day À1 , respectively (Table 7). These experimental values were close to the ANN-GA optimized values (77.95 mg g À1 biomass chlorophyll content and 2.2 mg L À1 day À1 chlorophyll productivity). Using the ANN-GA hybrid optimization technique an increase of 60% in biomass content, 31.17% increase in biomass productivity, 75.27% increase in chlorophyll content, and 23.75% increase in chlorophyll productivity were achieved compared to the standard physicochemical conditions in BG-11 media (Table 7).
In the TOA optimization technique, it was found the optimized physicochemical conation's values were within the values of the three levels of each physicochemical parameter. It indicated that the TOA optimization technique worked on the discrete values of experimental design. On the other hand, the ANN-GA optimization technique reached the optimized conditions by both interpolation and extrapolation of the experimental values rather than discrete experimental values. In this case, more biomass and chlorophyll synthesis was achieved using TOA than ANN-GA. It might be the case where ANN-GA would have better capability than TOA. Therefore, both optimization approaches should be employed for a given optimization problem and decisions should be taken based on predictive capabilities for modeling and optimization of other bioprocesses.
Effect of different light intensity, nitrogen, and carbon sources on biomass and chlorophyll productivity in optimized physicochemical and BG-11 media conditions

Effect of different light intensity
After evaluation of optimized physicochemical parameters from our current study, it was observed that light intensity was the most positive influential factor for the synthesis of biomass and chlorophyll from ANOVA analysis (Table 5). Therefore, for further evaluation of the effect of light intensity on biomass and chlorophyll productivity, experiments were carried out under photoautotrophic culture conditions by applying different light intensities ranging from 37 to 478 mmol m À2 s À1 (Figure 6a). In this experiment, optimized BG-11 media compositions were used from our previous work [39] with optimized physicochemical conditions with two reasonable modifications. Here, we employed a day-night ratio of 14:10 h instead of 18: 6 h (optimized condition) since an insignificant difference was observed from 14:10 h to 18: 6 h ( Figure 2a) and it reduced the energy consumption significantly ($30%). Further, inoculum size was kept at 4% instead of 8% which was also a reasonable constraint for bioprocess engineering. The experimental results showed that the biomass and chlorophyll productivity enhanced with increasing light intensity up to 128 mmol m À2 s À1 and thereafter decreasing trend was observed. At higher light intensity the color of microalgae cells became yellowish indicating the degradation of chlorophyll. At 128 mmol m À2 s À1 light intensity, optimum biomass productivity (55.1 mg L À1 day À1 ) and chlorophyll productivity (3.30 mg L À1 day À1 ) were observed. In the case of phototrophic microalgae, light acts as an energy source for photosynthesis and has a direct impact on biomass production and chlorophyll productivity. [30,31] However, algal growth is increasing up to a certain threshold, beyond this level a decline in algal productivity was observed due to photo-inhibition. [9] Our results supported this phenomenon and 128 mmol m À2 s À1 was the optimum light intensity for the isolated C. thermophila. [10] Effect of different nitrogen sources To investigate the significant effect of nitrogen sources on the productivity of biomass and chlorophyll, the culture was grown under the conditions mentioned earlier. Here, we used the light intensity of 44 mmol m À2 s À1 obtained from TOA optimization (Table 6), and sodium nitrate in BG-11 media was replaced with other nitrogen sources (e.g., Figure 6. Effect of (a) different light intensity, (b) nitrogen sources, and (c) carbon sources on the productivity of biomass and chlorophyll. Effect of light intensity ranging from 37 to 478 mmol m À2 s À1 for a photoperiod of 14:10 day-night ratio (a ammonium sulfate, ammonium chloride, ammonium acetate, sodium nitrate, glycine, and urea) with the same concentration. The experimental result showed that urea had the highest positive influence to support maximum biomass productivity (69.40 mg L À1 day À1 ) and chlorophyll productivity (4.16 mg L À1 day À1 ) (Figure 6b). Microalgae cells require a higher amount of nitrogen because nitrogen is a crucial atom present in chlorophylls, proteins, peptides, enzymes, DNA, RNA, ATP and other cellular molecules. [29] The utilization of urea as a nitrogen source gave an energy benefit due to its spontaneous hydrolysis to ammonia and CO 2 in the alkaline culture media, which is readily assimilated in the microalgae biomass. [51] Urea can be used directly after being converted to bicarbonate and ammonium by the urease enzyme. [52] It is important to note that urea release CO 2 and acts as an additional source of carbon CO 2 in the growth media attributing to the enhanced growth. [53] Results also showed that glycine and sodium nitrate were better nitrogen sources next to urea. Glycine dissolved freely in the aqueous media and the microalgae could easily absorb glycine for faster growth. [54] Nitrate is more thermodynamically stable than other forms of nitrogen that are oxidized in aqueous conditions. To be incorporated into the cell, nitrate must be chemically converted to ammonium after being translocated across the plasma membrane. Two enzymes, nitrate reductase, and nitrite reductase control the chemical reduction of nitrate. [55] With the help of NADPH, nitrate reductase catalyzes the bi-electrontransfer in the cytosol. In microalgae, the enzyme nitrate reductase is linked to pyridine nucleotide oxidation. [56] Nitrite reductase lowers the nitrite in a six-election transfer reaction. Nitrite reductase, which is found in the chloroplast, makes use of ferredoxin, which comes from photosynthetic electron flow in microalgae. The experimental result showed that this microalgae species were unable to grow in presence of ammonium sulfate and ammonium chloride and less amount of growth was found in presence of ammonium acetate. This might be due to the presence of ammonium ions on media that had an inhibitory impact on biomass and chlorophyll productivity. It had previously been shown that ammonium decreases the growth of microalgae primarily through the poisoning of the photosynthetic system. [57] The primary role of microalgal pigments is to harvest and transform light energy in the photosystems to synthesize ATP and NADPH, both of which are required for growth. In autotrophic cultivations, the microalgae obtained all of their energy from light. Therefore, inadequate energy leads to the toxicity of ammonium, which inhibits growth and chlorophyll synthesis. Furthermore, ammonia uptake by cells results in the release of H þ ions into the medium, causing the medium to become more acidic and the commencement of the early stationary phase. [58] Effect of different carbon sources The effect of carbon sources on biomass and chlorophyll productivity was obtained by growing the culture in dextrose, fructose, lactose, sucrose, cellulose, sodium acetate and other organic sources from biomass such as rice powder, rice straw in optimized physicochemical and BG-11 media culture condition mentioned earlier. The experimental results showed that the organism was not able to grow in presence of dextrose, fructose, and sodium acetate (Figure 6c). This might be due to the absence of transport enzymes in the cell membrane or inability to oxidize these carbon compounds or the lack of appropriate permease enzyme in the cell membrane. [59] Among the remaining carbon sources, sucrose supported maximum biomass and chlorophyll productivity and cellulose-supplemented media resulted in a significant reduction in biomass and chlorophyll productivity. The presence of carbon source obtained rice powder caused a significant increase in biomass and chlorophyll productivity than optimum culture conditions. However, rice straw and banana stem extract have no significant effects on biomass and chlorophyll productivity.
Comparative analysis of media concentration parameters, physicochemical factors, light intensities, nitrogen sources, and carbon sources A comparative study has been represented in Table 8 to evaluate the effect of different media concentration parameters, physicochemical factors, light intensities, nitrogen sources, and carbon sources on biomass and chlorophyll productivity. In standard physicochemical conditions and standard BG11 media, biomass productivity and chlorophyll productivity were observed to be 37.78 and 1.60 mg L À1 day À1 , respectively. Compared to the standard physicochemical and standard BG-11 media, biomass productivity increased by 51.36%, and chlorophyll productivity increased by 53.12% in optimized media concentrations of BG-11. [39] In standard BG-11 media with optimum physicochemical circumstances, the increase in biomass productivity and chlorophyll productivity was 23.07% and 61.25% compared to the standard condition. In optimum light intensity (128 mol m À2 s À1 ) and optimized media and physicochemical conditions, biomass production increased by 45.86% and chlorophyll productivity increased by 106.25% as compared to the standard condition. In the presence of urea, biomass productivity increased by 83.73%, and chlorophyll productivity increased by 160%. However, in the presence of sucrose as a carbon source in optimum media and physicochemical conditions, a significant increase in biomass and chlorophyll productivity was observed. The experimental results showed that biomass productivity increased by 157.46% and chlorophyll productivity increased by 264.38%. In summary, significant enhancement of growth and chlorophyll synthesis was found by varying different regulating factors. Therefore, the assimilation of these factors with careful process engineering would lead to high cell density microalgae biomass.

Conclusion
The isolated C. thermophila, containing a high amount of pigment molecules ($6% of chlorophyll) can be used as a potent microalgae strain for commercial scale-up operations. In this work, we performed experiments, modeling, and optimization of the cultivation process by varying physicochemical parameters to enhance the content and productivity of biomass and chlorophyll. Initially, optimization of the physicochemical process parameters resulted in a significant improvement of the content and productivity of biomass (118.18% and 23.07%, respectively) and chlorophyll (94.52% and 61.25%, respectively) than the standard physicochemical condition in BG-11 media. Further, using optimum media and physicochemical conditions, light intensity was optimized (128 mol m À2 s À1 ) with the enhancement of biomass and chlorophyll productivity of 45.86% and 106.25%, respectively. In the presence of urea as a nitrogen source in optimized physicochemical and media conditions, biomass productivity increased by 83.73% and chlorophyll productivity increased by 160%. Thus, urea can be used as an effective nitrogen source due to its minimum production cost and low price per kilogram. [60] A considerable increase in productivity of biomass (157.46%) and chlorophyll productivity (264.37%) was observed in the presence of sucrose as a carbon source in physicochemical and media optimized culture conditions. Current work provided valuable information that can be used for the scale-up of the microalgae cultivations. A careful synchronization by integrating the investigated information will lead to high cell density biomass of isolated C. thermophila for a sustainable amount of microalgae-based bio-products.

Authors' contributions
Sreya Sarkar, Sambit Sarkar, and KG conceptualized the work and developed the experimental methodologies. Sreya Sarkar and Sambit Sarkar performed the experimental data curation. Sreya Sarkar and KG analyzed the data and drafted the manuscript. TKB reviewed and edited the manuscript.
Disclosure statement