Synthesis of Carbon Nano Materials from Carbohydrate Rich Biogenic Precursors using Chemical Vapor Deposition Method and Optimization of Parameters by Taguchi Optimization Method

Carbon Nano Materials (CNM) from biogenic carbohydrate rich non-edible precursors such as Syzygium cumini (Jamun), Tamarindus indica (Imli) and Litchi chinensis (Imli) seeds were synthesized by chemical vapour deposition method (CVD). Parameters were optimized by Taguchi optimization method, four parameters such as precursor, temperature, carrier gas and duration and three levels S. cumini, T. indica and L. chinensis as a precursor, Argon (Ar), Nitrogen (N2) and Hydrogen (H2) as a carrier gas, 1 hour (hr), 2 hours and 3 hours (hrs) for duration of pyrolysis, 800C, 900C and 1000C for temperature were selected. In the present work, impact of precursor is 35.84% which is the most effective factor than temperature (29.59%) and other parameters such as duration (19.50%) and the least effective factor is carrier gas (15.07%). Impact of parameter on the yield which described by the deviation of signal to noise (S/N) ratio. The result of deviation of S/N ratio shows T. indica seed (precursor), 900C (temperature), 2 hr (duration) and Ar (carrier gas) are the best parameters. The morphology of CNM is studied by SEM characterization and the nature of synthesized CNM by RAMAN spectroscopy.


Introduction
Using petroleum derived materials as precursor, Carbon nano materials (CNMs) are usually synthesized. As petroleum derived materials are comparatively expensive and also they are at the verge of depletion in the near future. Hence, there is the need to look for other sources of precursors to synthesize CNMs. While following the principle of Green Chemistry, it is desirable that the raw material for any industrial process must be renewable rather than the depleting natural resources. It is one of the prime requirements to explore regenerative materials for synthesis of CNM, however, with high efficiency. Sharon and co-workers [1]- [4] firstly synthesized carbon nano tubes (CNT) from plant-derived precursors and since then there have been efforts in this direction in different parts of the world. Scientists have been exploring environment-friendly sources of CNMs. Efforts are being made to establish the conditions for growing multi walled carbon nanotube (MWCNTs), singlewalled carbon nano tubes (SWCNTs) and vertically aligned MWCNTs using suitable catalytic support by a simple and inexpensive chemical vapour deposition (CVD) techniques. There are different plant derivatives as well as plant parts which have successfully been used for the synthesis of carbon nano materials under pyrolytic conditions (5)(6)(7)(8)(9). They also synthesize CNM by a modified tradition sources from vegetable sources [10]- [12]. Almost all plant parts (roots, stem, leaves, seeds etc.) and plant derived products (pinene, latex, oil, juice, camphor) have been used as precursors of CNMs. The choice of precursor is primarily based on wide availability of the plant or plant parts. The next criterion is the percentage of hydrocarbon content (oil or fatty acid, cellulose, polysaccharide, lignin etc.) of the chosen plant precursors. Researchers have also used sugarcane juice, an algae Euglena, cotton, fibre, rice straw, corn cob, latex of Calotropis, bamboo, and pretreated rice straw as raw materials, however, with promising results. The use of palm oil has yielded aligned CNTs arrays. Spray pyrolysis of biodiesel has yielded well defined multi walled carbon nano tubes [13]. The Carbon Nano Material (CNM) is determined by the nature of precursor. There are several reports with regard to the carbon nano materials being derived from the seeds including Grapes seeds, T. indica, Soybean, Anacardium, Ritha and Castor [14]- [17].
The aims of the present work are to synthesize carbon nano materials from precursors derived from carbohydrate rich nonedible seeds of plants and parameters were optimized by Taguchi optimization method using CVD process. It has been decided to use various types of carbohydrate rich seeds as Jamun (Syzygium cumini (L.) Skeels), Imli (Tamarindus indica L.) and Litchi (Litchi chinensis Sonn.) to synthesize carbon 432 nano materials. These seeds contain different amount of carbohydrate, oil and protein constituents, these are rich source of carbon.

Materials and Methods
For the synthesis of CNM, the seeds of S. cumini, T. indica and L. chinensis were purchased from local market, CVD furnace was used for pyrolysis.CVD method was used for Pyrolysis of these three type of seeds. The parameters considered were 1) three different gases i.e. Argon (Ar), Nitrogen (N2), and Hydrogen (H2) 2) three different temperatures i.e. 800 0 C, 900 0 C and 1000 0 C and 3) duration of pyrolysis 1hr, 2 hrs, and 3 hrs. The seeds were washed with D/W and dried in the muffle furnace at 60 0 C for 6 hrs and the crushed into powder form with the help of mortar and pestle. After crushing, took known amount of precursor into a quartz boat for pyrolysis. CVD apparatus was used for the synthesis of carbon nano materials from carbohydrate rich non-edible seeds. Known amount of crushed seeds were taken in a quartz boat and placed in the centre of 1 m long quartz tube (inside furnace) that was inserted in the furnace. The carrier gas was flushed into the quartz tube at a higher flow rate (150 ml/min) so as to remove oxygen and the quartz tube was closed. Then the purging of gas was continued at the flow rate (25 ml/min) during pyrolysis. Desired temperatures were set when the temperature reaches 450 0 C -500 0 C, the biomass starts vaporizing inside the tube and flows along with the carrier gas toward outlet of the tube. By the time the temperature reached 700 0 C the flow rate was gradually decreased and stopped. When it reached at set temperature, it was maintained at that temperature for desired set duration. At high temperature, the hydrocarbons are decomposed and converted into simpler form of carbon. The non-carbonaceous material gets removed along with the carrier gas from the furnace and remnant carbonized carbon material are left in the quartz boat.

A. Standardization of synthesis parameters (Taguchi Optimization of Parameters)
For the present CVD process four parameters, each with three variables were to be tried. Therefore, for the present work, experiment was designed by Taguchi Optimization method [18]. This method is a statistical method developed by Genichi Taghuchi (1950) of Nippon Telephones and Telegraph Company, Japan for improving the manufactured goods, marketing & advertising, more recently, even applicable for the research field. In this method statistically designed orthogonal arrays (OA) [19]- [20] are being of used to evaluate comparatively smaller number of experiments. It is a simple efficient and systematic approach for the optimization of experimental parameters [21]- [22]. Using Taguchi method one can get information about the most effective parameters controlling the formation of a required product by carrying out only 9 experiments. The Taguchi philosophy of variability reduction is based on the fact that there is a 'best' value for the product. The Taguchi strategy attempts to find the combination of the values of the controllable design variables that minimizes the expected loss over the uncontrollable noise space.

B. Application of Taguchi Optimization for Present work
Four parameters considered for the synthesis of CNM from plant seeds are Carrier Gases, Precursors, Temperatures, and Duration of Pyrolysis. In total, three levels of each parameter were selected (Table 1). Using above mentioned parameters and its level, L9 orthogonal matrix was designed ( Table 2).
Taguchi optimization method is a statistical method [23] in which simple mean of analysis and optimization of complex systems based on the statistical analysis of data are optimized. A special design of orthogonal arrays (OA) is used in Taguchi optimization method to study the entire parameter space with only a small number of experiments. This approach of the statistical analysis is comparatively simple for optimization of experimental designs in order to evaluate performance and quality of the experiments. Statistical analysis of signal to noise (S/N) ratio and an analysis of variance (ANOVA) [24] has to be employed for determining relative importance of various factors. Orthogonal array (OA) analysis is completed by ANOVA and S/N ratio. ANOVA is used to analyze the results of the OA (Orthogonal Arrays) experiment and to determine how much variation each quality (influencing factor) has contributed. S/N ratio is log function of desired output for optimization, which thereby, helps in data analysis and as well as in prediction of optimal results. An S/N ratio is a key part of the Taguchi optimization method and is often evaluated using a parameter design. Taguchi's S/N ratios are the performance statistics that recommends for selecting the best combination of the design variables. Larger the S/N ratio more robust the design. For the best design, the parameter which gives S/N ratio highest value is considered as the favoured parameter. Thus, to minimize the expected loss, one always finds the combination of the design variables, which maximizes the SN ratio performance statistic.
The S/N is the ratio of the average response of the root mean square variation about the average response. Larger the S/N ratio, smaller the measurement of error because the S/N is the reciprocal of the variance of the measurement error. S/N ratio is used to analyze the test run results because the S/N ratio signifies the mean mean and variation (scatter) of the experimental results.
Calculations for 'Larger the Better': There are three categories of S/N ratios (i) smaller the better, (ii) larger the better and (iii) nominal the best. In the present work target was to get larger quantity of carbon, therefore, "larger the better" is used for calculation using following equation: Where "yi" is the mean response calculated as y =1/n Ʃyi and n is the number of experiments carried out under similar conditions.
Calculations for Effect of each parameter: To determine the effect of each parameter level (mi) average value of S/N ratios was calculated for each parameter, using analysis of mean (ANOM). For this calculation, the S/N ratios of each experiment with corresponding parameter levels are calculated using following equation: Where, ni is the number of experiments repeated with the same parameter levels.
There are two types of average value of S/N ratio is calculated. One is the overall mean S/N ratio calculated from the entire experiments for example from L9 experiment of L9 orthogonal array (i.e. from S/N values given in table). Secondly, average S/N ratio is calculated for each parameter from equation-1. The advantage of the average value of S/N ratio, are considered to be the least effective parameters as compared to those parameters whose S/N ratios are larger than mean S/N ratio.
The parameters effects, i.e. the contribution of each experimental parameter to the quality characteristic are calculated by the analysis of variance (ANOVA). ANOVA is done by summing of squares (SOS) of variance of all levels of a given parameters. The relations that are used to determine the sum of the squares and the factor effect are given by the following equation: Where mi gives the average of the levels contributions for each parameters levels to S/N ratio, <mi> is the average of mi for a given parameter and the coefficient and Ni represents the no. of repeats the experiment is conducted with the same factor level.
SOS is normalized with respect to the degree of freedom (DOF) of the corresponding process parameters. DOF = number of parameter level -1 If there are three levels for each parameters then DOF = 3-1 = 2 Some of square (SOS) of variances for all levels for a given parameter are divided by degree of freedom (DOF) of corresponding process parameter to obtain factor of effects (FOE) of various experimental parameters. The equation of calculation of FOE is following: Finally, percentage parameter effect is calculated as following: Parameter effect % = 100 x FOE (6) Using L9 orthogonal array nine sets of experiments were carried out. Results were calculated based on equations 1 to 3 mentioned above, nano and micro carbon materials were synthesized by CVD method from different non-edible seeds.

C. Purification of CNM derived from Seeds
Acid-refluxing the sample is an effective measure in reducing the amount of metal particles, amorphous carbon, and raw material based impurities and also for residual functional groups. Various acids such as HCl, HNO3, H2SO4 are commonly used for purification of CNMs. For SEM treatment of our experimental seeds samples, the synthesized CNM was purified utilizing acid reflux with HCl and HNO3. In the present work, 1.01 gm of crushed CNM produced was weighed and taken in chemical synthetic reactor/ three necked round bottomed flask. In the reaction vessel 100mL Conc. HCl (36.5%) and 100mL Conc. HNO3 (63%) were added along with the sample and those were refluxed at 110°C for 12 hrs. Then International Journal of Research in Engineering, Science and Management Volume-3, Issue-7, July-2020 https://www.ijresm.com | ISSN (Online): 2581-5792 | RESAIM Publishing 434 after, the flask was allowed to cool for 6-8 hrs. The resulting black residual of the reflux process was centrifuged and subsequently the dark liquid was discarded. The powder thus obtained was washed with luke warm double distilled water for several times till the traces of acid was completely removed. The obtained powder was then dried in vacuum oven (24 hours). The sample was now ready for SEM analysis. Obtained CNM was analysed by SEM, and Raman Spectroscopy.

A. Carbon Nano Materials from carbohydrate rich non-edible seeds
Taguchi optimization method is a statistical method by which experimental parameters are optimized in a small number of experiments. On that basis the orthogonal table was constructed considering different combinations (four possible parameters) with a minimum of three variables of each parameter. We had selected the variable parameters such as carrier gas, temperature, duration and precursors (different type of carbohydrate rich non-edible seeds) and they all are orthogonal to each other. By using Taguchi Optimization technique, an orthogonal table was constructed in which three levels of parameters were used. Therefore, L9 orthogonal table was constructed by using the parameters and their levels as mentioned (Table 2). Pyrolysis was carried out for each set of conditions. Percentage (%) of yield and S/N ratio was calculated for each experimental result by using equation 1. The yields of CNM as well as S/N ratio calculated from each set are also given in table 3. Deviation of S/N ratio was also calculated and factor of effect (FoE) was evaluated.

B. Taguchi optimization on the yield of carbon
So far as yield of carbon from pyrolysis of different seeds are concerned, Syzygium cumini and Litchi chinensis yielded 22 -24% carbon, whereas Tamarindus indica produced upto 23% carbon. Earlier, Chaudhary et al. (2014) derived Carbon nano particles (21% wt./wt.) from Litchi seeds, however, as compared to the amount of precursor used. Mopung et al.
(2015) observed comparatively higher 28.89% of carbon yield the seed of Tamarindus indica at a temperature range of 500ºC -700ºC. Various parameters of pyrolysis do not seem to have much impact on the yield. However, it was observed that out of three seeds taken Syzygium cumini and Tamarindus indica yielded higher levels of carbon at lower temperature compared to Litchi chinensis. The decrease in CNM yield with increasing temperature could either be due to greater primary decomposition of biomass at higher temperature or through secondary decomposition of carbonaceous materials residues (Ming et. al., 2016). Signal to noise (S/N) ratio of CNM produced under different conditions is plotted in figure 2. The purpose of the proposed work was to find out the best set of parameters for pyrolysis which could give the maximum amount of CNM. For calculation of the S/N ratio "Larger the Better "condition was used. Among three types of carbohydrate rich non-edible seeds Tamarindus indica seeds appeared to give the best yield. Following Taguchi optimization, the deviation S/N ratio showed that 900 0 C temperature (as applied) and two hours' time of pyrolysis gave the higher yield of CNM. Amongst the three carrier gases applied, Argon was better suited than nitrogen and hydrogen for the yield of CNM. Since hydrogen is a reducing gas, it might have reduced the carbon during secondary decomposition of carbonaceous materials (Li et al., 2017).  (Table 4) suggests that precursor (35.84%) has maximum impact on the experimental designing followed by temperature (29.59%) then duration (19.50%) and the least impact is that of carrier gas (15.07%) which was shown in the pie diagram ( figure 3). Sharon et al., 2011 [25] working at 800ºC using Argon as carrier gas and pyrolyzing duration of 2 hours showed 15% impact compared to other parameters.  (Fig. 4) showed that precursor had maximum level of impact followed by temperature and duration, however the impact of carrier gas had least impact.

C. Scanning Electron Microscopy (SEM) micrographs of pyrolised Carbon materials from carbohydrate rich non-edible seeds
Morphological studies of SEM micrograph and observation of pyrolised different carbohydrate rich non-edible seeds (Syzygium cumini, Tamarindus indica and Litchi chinensis.) are represented in Table 5.
The SEM micrographs of CNMs (synthesized through CVD of the three precursors: Carbohydrate rich non-edible seeds: Syzygium cumini, Tamarindus indica and Litchi chinensis) are shown in this table.SEM micrographs are the experiments of different L levels (L1 -L9) as optimized by 'Taguchi method'. The SEM micrograph were obtained from the powder of precursors which revealed plate like structures.
The SEM micrographs of Syzygium cumini (L1, L5 and L9 (Table 5)) it was observed that the CNM morphology was multilayered smooth surface plates with porous side and the size range was approx. 30-750 nm. In L1 the size of larger structures were about 150-750 nm and the size of smaller fragments was in the range of 30-100 nm in range. In L5, the size of larger structure was 100-2000 nm and the size of smaller fragments was approx. 50-100 nm. In L9, the length of the larger structure was about 2200 nm and the size of smaller fragments was about 20-100 nm which was impregnated on the larger structures.
On observation and respective comparison of the SEM micrographs of CNM of Tamarindus indica (L2, L6 and L7 (Table 5)) it was found that the CNM morphology of L2 was multilayered plate with rough surface and porous sides, the size range was 40-2000 nm. The morphology of L6 showed that the structure with pits (bowl like) contained small fragments of CNM. The pit diameter was about 100 nm, and larger fragments of 2200 nm in size while smaller fragments were of 30-60 nm size. The SEM micrograph of L7 showed that smooth surface plate with multilayered plates. The size of larger fragments was 1500 nm and smaller fragments of 30-100 nm. While working on T. indica seed parts, Munusamy et al. (2011) [15] found that the decoated T. indica seeds were non-porous, whereas the T. indica seeds char (TSC) and T-char were of porous structures.  Multi-layer smooth surface plates with porous sides also small fragments.

L2
Ar -900 0 C -2 hrs -Tamarindus indica seed Multilayer plate with rough surface and porous sides. The observation of SEM micrograph of pyrolysed CNM of Litchi chinensis seeds (L3, L4 and L8 (Table 5)) had plate like morphology and the length of CNM was about 40-2500 nm in size. The morphology of all the three levels (L3, L4 and L8) of experiment were almost same but with variable of CNM. The larger structure of L3 was about 1500 nm and the smaller fragments of were of 40-200 nm in size. In L4, the larger structures was 2000 nm and the smaller fragments were of 50-250 nm in size. The size of L8 varied from 100-2500 nm. Chaudhary and Bhowmick (2014) [26] observed 40-70nm range of CNPs in aggregated form in Litchi peel.

D. Result and Inference of Raman spectra of Carbohydrate rich seeds E. Raman spectra of Jamun seeds (Syzgyium cumini)
In the Raman spectra of Taguchi optimized Syzygium cumini seeds, there was sharp peak of D-band L1 at 1357, L5 at 1366 & L9 at 1367 and G-band of L1 at 1604, L5 at 1607 & L9 at 1608 which revealed the presence of sp 3 and sp 2 carbon with defect ( fig 5). In pure diamond this peak remains at 1332 and in pure graphite peak at 1580. The ratio of it it's intensities show the defect in those materials. ID/IG ratio of L1 is 0.874, L5 is 0.924 & L9 is 0. 945.An additional blunt 2D peak revealed at 2915 in L1, at 2915 in L5 & at 2919 in L9 which indicated the graphitic nature of these Carbon nano materials. In above three Raman spectra it was studied that with respect to temperature, as the temperature increased, the ratio of defect in turn increased, it may be due to aggregation of molecules. Similarly, in the case of Hydrogen gas produced more defects than Nitrogen and Argon and with respect to duration of pyrolysis, 2 hrs revealed more defect than 3 hrs and 1 hr.   (c) T. indica seed (H2, 800 0 C, 3 hrs) Fig. 6. a) Raman spectra of L2 (Precursor-Tamarindus indica seed, Temp-900 0 C, Gas-Argon, Duration-2 hrs), b) Raman spectra of L6 (Precursor-Tamarindus indica seed, Temp-1000 0 C, Gas-Nitrogen, Duration-1 hr), c) Raman spectra of L7 (Precursor-Tamarindus indica seed, Temp-800 0 C, Gas-Hydrogen, Duration-3 hrs).

G. Raman Spectra of Litchi chinensis Seed
In the Raman spectra of Taguchi optimized Litchi chinensis seeds ( fig. 7)  In L3, L4 & L8, the presence of broad peak of 2D peak at 2911, 2900 & 2920 showed samples being graphitic in nature. Bhardwaj et al. (2007) [27] studied on carbon materials of different seeds and found that IG/ID ratio in soap nut seeds and bamboo fibre had highest value suggesting that it contained more graphitic carbon than disorder carbon atoms.
In the present studies, the analysis of CNMs were done on the basis of D banding, G banding and 2D peaks obtained with this technique and structural information was interpreted (table 6). It was found that the CNMs from carbohydrate rich seeds (Tamarindus indica, Syzygium cumini and Litchi chinensis) produced: D band -1354 -1367 cm-1 G band -1600 -1619 cm-1 2D peak -2900 -2920 cm-1 On account of the above observation it is sumrised that the CNM synthesized from the aforesaid non-edible seed precursors possessed prevailing graphitic constitution along with some graphene like structure.