Development of high-performance nickel-based catalysts for production of hydrogen and carbon nanotubes from biogas

Selecting a suitable catalyst for implementing the simultaneous production of hydrogen-rich syngas and multi-walled carbon nanotubes through the integration of dry reforming and methane decomposition reactions has recently gained great interests. In this study, a series of bimetallic (NiMo/MgO) and trimetallic (CoNiMo/MgO, FeNiMo/MgO, CoFeMo/MgO) catalysts was prepared and evaluated for a catalytic activity of CH4 and CO2 conversions of biogas in a fixed bed reactor at 800 °C and atmospheric pressure. Among the investigated catalysts, the bimetallic NiMo/MgO catalyst showed the outstanding catalytic performance with 86.4% CH4 conversion and 95.6% CO2 conversion as well as producing the highest syngas purity of 90.0% with H2/CO ratio = 1.1. Moreover, the characterization of the synthesized solid products proved that the well-aligned structured morphology, high purity, and excellent textural properties of CNTs were obtained by using NiMo/MgO catalyst. On the other hand, using trimetallic catalysts which have the composition of Co and Fe leads to the severe deactivation. This could be attributed the catalyst oxidation with CO2 in biogas, resulting in the transformation of metals into large metal oxides. The integrative process with NiMo/MgO catalyst is regarded as a promising pathway, which has a high potential for directly converting biogas into the high value-added products and providing a green approach for managing the enormous amounts of wastes.


Scientific Reports
| (2022) 12:15195 | https://doi.org/10.1038/s41598-022-19638-y www.nature.com/scientificreports/ ogy and particle-sized distribution of synthesized CNTs were identified by a Field-emission scanning electron microscope (FE-SEM, JEOL, JSM-7610F) at the operating voltage of 10 kV. The internal structure of nano carbon deposited on the spent catalyst was characterized by a Field-emission transmission electron microscope (HR-TEM, JEOL, JEM-3100, Japan) at the operating voltage of 300 kV. The purity of the synthesized carbon nanotube was determined by thermogravimetric analysis (TGA, Metteler Toledo, TGA/DSC1) with an oxygen gas flow rate of 50 ml/min and a heating rate of 10 °C/min from room temperature to 800 °C. The hydrogen programmed reduction (H 2 -TPR, BEL JAPAN, BELCAT-B) method was used to assess the reducibility and metal-supported interaction of calcined catalysts. Prior to measurements, 100 mg of calcined catalysts was pretreated at 200 °C for 1 h under a He flowrate of 30 ml/min. Afterward, the H 2 -TPR profile was obtained using a mixture of 5% H 2 / Ar flowing at a rate of 30 ml/min with the sample heated from 100 to 900 °C with a fixed heating rate of 10 °C/ min. The chemical composition of the reduced catalysts was analyzed by X-ray fluorescence (XRF, S8 TIGER, Series 2). The graphitic and disorder carbon of synthesized CNTs was examined by Raman spectrophotometer (PerkinElmerÒ Spectrum™ GX). The textural properties of spent catalysts were identified by the N 2 sorption measurement (Micromeritics, TriStar II 3020).
Production process. The evaluation of the catalytic activity was performed under atmospheric pressure.
The 0.5 g catalyst held on a quartz boat (15 mm width, 100 mm length, 8 mm height) was inserted at the center of a fixed-bed horizontal quartz reactor (26 mm I.D. and 1100 mm length) with PID temperature controller (KP1000 series, Chino) and Type-K thermocouple inserted at the center of the reactor. The experimental apparatus is schematically shown in Fig. 1. Prior to the reaction, the catalyst was reduced by H 2 (75 ml/min) with a ramp rate of 10 °C/min until the temperature reached 1000 °C. Once the reactor temperature reached 1000 °C, biogas with a composition of 60% CH 4 and 40% CO 2 was continually fed into the reactor at the flow rate of 400 ml/min, corresponding to GHSV = 48,000 ml/g-h. During the reaction, the gas composition of the outlet stream was examined every 10 min using Gas chromatograph (TCD, Shimadzu, GC-2014). The volumetric flow rate of the inlet and outlet streams was controlled and calibrated by a mass flow controller (HORRIBA METRON, S48-32HMMT) and a soap bubble meter, respectively. After 3 h of the reaction time, the biogas was switched to the He flow at 50 ml/min to cool down the reactor to room temperature. The spent catalyst with deposited carbon was then weighed and collected in a desiccator for further characterization.
Product analysis. The catalytic performance of the bimetallic and trimetallic catalysts in the integration of dry reforming.
For the gas analysis, the effluent gas of the outlet stream, H 2 purity, and H 2 /CO ratio were calculated using Eqs. (5,6), where F i is the flow rate of component i, [i] represents the concentration of CH 4 , CO 2 , H 2 or CO. F in and F out are the total gas volumetric flow rate of inlet and outlet from the reactor, in term of (ml/min). The production yield, purity of synthesized CNTs and gas hourly space velocity were calculated according to Eqs. (7)(8)(9)(10).

Results and discussion
Characterization of catalyst. All catalysts were prepared by wetness impregnation method. The total loading of metals (Ni, Co, Fe, and/or Mo) on MgO was fixed at 30 wt.%. The equal mass ratio of metals was used. Each catalyst was reduced by H 2 (75 ml/min) before the characterization. The actual elemental composition of the reduced catalyst was analyzed by X-ray diffractometer (XRF) as shown in Table 1. The total metal loading of bimetallic catalyst was observed to be 31.6 wt.%. Likewise, the total metal loading of trimetallic catalyst is in the range of 30.0-30.8 wt.%. All catalyst results have a total metal loading that is quite close to the prescribed values. Furthermore, a trace of Ca impurity in all of the metal catalysts was detected with 0.2 wt.%. X-ray diffraction (XRD) measurement was principally carried out to identify the crystalline phases arising during the process. The results of the contemporary catalysts analyzed by XRD after calcination at 500 °C for 3 h are depicted in Fig. 2a. For all samples, many diffraction peaks that mainly ascribed to magnesium oxide support (JCPDS 45-0946) were observed at 2θ = 36.96°, 42.92°, 62.29°, 74.59°, and 78.58°, which is perfectly in accordance with (111), (200), (220), (311), and (222) planes. However, the crystal phase of NiO (JCPDS 47-1049) was identified only in the pattern of NiMo/MgO catalyst, showing a higher number of particles consisting of Ni species on the surface of the support. Moreover, the Co 3 O 4 phases (JCPDS 42-1467) are almost absent as identified in the XRD results profile of CoNiMo and CoFeMo/MgO catalysts. The possible reason is that it can be well dispersed in the support and/or too small to be detected. The appearance of MoO 3 species (JCPDS 05-0508) in these catalysts are known to form due to the oxidation of Mo at temperatures above 500 °C 37 . That can be attributed to a high composition of molybdenum oxide phases on the catalysts surface. The reflection peaks of MgMoO 4 (JCPDS 72-2153) are also detected with very low or less intensity in the diffraction pattern of all catalysts, which formed via interaction between MgO and MoO 3 at a temperature higher than 400 °C 38 . Figure 2b illustrates XRD diffractograms of fresh catalysts obtained after carrying out the activation process by reducing under certain conditions, along with the structural parameters analyzed by the XRD technique are presented in Table 2. It was noticed that the characteristic peak of MgO phases appears extensively in all catalysts. Also, it can be clearly seen that the MgO crystal domain size significantly diminished in all catalysts compared to the reduced MgO. In addition, the reflections attributed to metallic forms such as Ni (JCPDS 04-0850), Co (ICPDS 15-0806), Fe (JCPDS 65-4899), and Mo (ICSD 64-3962) were detected. In comparison, the crystallinity of metal Mo peaks for using bi-metallic NiMo/MgO catalyst presents a higher than other tri-metallic catalysts, revealing a high percentage of Mo in the catalyst. Meanwhile, no other diffraction patterns reflexed to their oxides were observed, expecting that metallic phases can be completely reduced into the activated catalyst by the usage of hydrogen reducibility at a high temperature.
The reducibility of all the fresh calcined catalysts used in this study was determined by the H 2 -TPR technique and the results are shown in Fig. 3. For the bimetallic NiMo/MgO catalyst, the TPR profile illustrates two hydrogen consumption regions: in the first low-temperature region, between 300 and 600 °C, a broad peak with  www.nature.com/scientificreports/ a slightly low intensity observed at 399 °C is assigned to the reduction of loosely attached NiO bound on the support surface, and the second, shoulder peak centered at 523 °C could be attributed to the reduction of NiO species strongly interacted with the support 37,39 . At the higher temperature region between ranges of 600-1000 °C, the profile shows the presence of two distinct peaks that appear centered at 630 and 720 °C, indicating two differentiated reduction processes. According to previous studies [40][41][42] , the former represents the reduction of octahedral Mo-oxo species (Mo 6+ → Mo 4+ ) together with the first reduction step of MgMoO 4 , and the following peak consumed in the high reduction temperature is associated with the second reduction step of MgMoO 4 as well as the further reduction from tetrahedral Mo 6+ to metallic Mo 0 passing through the Mo 4+ phase which has been reported that is more difficult to reduce than that of octahedral ones. On the other hand, the H 2 consumption peaks of the trimetallic with Co-containing catalysts appeared centered at temperatures of 372 °C are related to the reduction from Co 3 O 4 to metallic Co 0 phase. It is reported that if the size of these peaks is small, the relative amount of these crystals is very low. Thus, it can be said that  . Also, the peaks at a temperature range of 400-500 °C were detected in the catalysts composed of Fe content, which reflected the formation of metallic iron from the two-step reduction of magnetite as a sequence of Fe 3 O 4 to FeO and Fe 21,41,45 . In addition, the TPR profiles of all trimetallic catalysts display a single broad peak with a maximum at around 650 °C which is deconvoluted into mainly two peaks corresponding to the reduction of MgMoO 4 and Mo species as mentioned before. These peaks begin with temperatures higher than 500 °C and seem to continue even up to 900 °C, especially in the profiles of FeNiMo/MgO and CoFeMo/MgO catalysts. The broad tail at high temperature is probably due to the result of a larger amount of Fe oxides were not completely reduced to Fe 0 , indicating a strong interaction between iron species and the support 46 . This suggests that both catalysts needed temperatures higher than 900 °C for the full reduction into the metallic form.
Evaluation of catalytic performance and durability. The performance of bimetallic and trimetallic catalysts was evaluated under the integration of dry reforming catalytic decomposition of methane. All catalytic has been carried out in a fixed bed reactor at 800 °C, 1 atm under biogas (CH 4 :CO 2 = 1.5:1) feeding, with a corresponding space velocity of GHSV = 48,000 ml/gCat-h. The feed gas conversion and the amount of H 2 and CO product were calculated based on the measurement of the effluent gas composition of the outlet stream during the reaction for 3 h. In this process, the decomposition of CH 4 on the catalyst surface leads to the production of H 2 and carbon nanotubes (CNTs) 47,48 . Meanwhile, CO 2 interacts with CH 4 through a dry reforming reaction, which is accompanied by a parallel reaction that produces syngas (H 2 + CO) 49 . As a result, the outlet stream effluent gas contains H 2 , CO, and a small amount of unconverted CO 2 and CH 4 . Figure 4a,b shows the CH 4 and CO 2 conversion as a function of time on stream. It is clearly observed that the coexistence of Ni and Mo reveals exceptional performance with achieving higher CH 4 conversion of 86.4% and CO 2 conversion of 95.6% in NiMo catalyst as well as showing excellent stability over 3 h. Besides, H 2 -rich syngas was produced over NiMo/MgO catalyst having H 2 /CO ratio of 1.1. In addition, the syngas purity of 90.0% was obtained as listed in Table 3.
Meanwhile, the addition of Co and Fe in the catalyst is found to be that the conversion of CH 4 and CO 2 decreases gradually. This occurrence may imply that the catalyst surface of Co and Fe would have undergone oxidation while exposure to CO 2 in biogas led to the catalyst deactivation 26,[50][51][52] . Table 4 compares the performance of NiMo/MgO catalyst, with respect to Ni-based catalysts from the previous reports for converting biogas in various circumstances. In the case of the CH 4 /CO 2 ratio equal to 1.5/1, NiMo/MgO catalyst in the present study outperforms the competition in CH 4 and CO 2 conversions. Our catalyst possesses outstanding performance for syngas production having the H 2 /CO ratio > 1, a resource for producing high-value products such as methanol and liquid hydrocarbons by Fischer-Tropsch synthesis 53,54 . Furthermore, NiMo/MgO catalyst can perform superior for the formation of valuable carbon nanomaterials without encountering the deactivation of the catalyst during the tested time. These findings assure that our catalyst is regarded as a promising catalyst for biogas transformation into syngas and carbon nanomaterials through the integrative process of DRM and CDM. However, as indicated in Table 4, the concentrations of CH 4 and CO 2 in biogas are Characterization of CNTs products. CNTs generated from the decomposition of CH 4 were examined for yield, purity, crystallinity, graphitization, morphology, and textural properties. Table 3 presents the production yield of CNTs in terms of gProduct/gCat-h was synthesized over bimetallic (NiMo/MgO) and trimetallic (CoNiMo/MgO, FeNiMo/MgO, CoFeMo/MgO). The bimetallic NiMo/MgO catalyst exhibited high performance for the CNTs production of 2.60 gProduct/gCat-h than all trimetallic catalysts. The production yield of CNTs obviously that the cohabitation of Ni and Mo particles drastically encouraged the creation of CNTs through CDM, Eq. (2) 60 as compared to that obtained without Ni catalyst 60 as compared to that obtained without Ni catalyst. On the contrary, the addition of Co and Fe to the catalyst induced a significant abatement yield of synthesized carbon nanotubes. This could be owing to the metals Co and Fe in biogas being easily oxidized by CO 2 and forming metal oxides 26,[50][51][52] . Figure 5 provides more insight into the morphological differences of the deposited carbon over various catalysts used, as well as the analyzing parameters of synthesized CNTs which were calculated from FE-SEM and HR-TEM images are presented in Table 5. From the SEM images, it was clearly observed that the presence of condensed nanocarbon was observed on the catalyst surface, especially on catalysts that contain Ni metal particles while a small amount of CNTs detected on the surface of the catalyst with the absence of Ni (Fig. 5d). This indicates the high activity and capability of producing CNTs by using Ni-based catalysts. In addition, TEM images as shown in Fig. 5, confirm the formation of multi-walled carbon nanotubes (MWCNTs) on all catalysts used in this study. Also, it is obvious that the addition of Fe in the catalyst forming the bamboo type in carbon nanostructures 21,25 which may be useful in some practical applications, such as hydrogen storage, electrochemical capacitors, and lithium-ion batteries 61 .
From Fig. 5 and Table 5, it was found that the narrowest and smallest carbon nanotubes can be achieved over NiMo catalyst without any further combination with other transition metals. Meanwhile, the addition of Fe and Co apparently gives the higher range of the wall numbers and diameters of CNTs. This might be attributed to the agglomeration of oxide species (Fe 3 O 4 and Co 3 O 4 ) on the catalyst surface which leads to an increase in metal particle size, consequently, large CNTs are produced 44,62 .
The graphitization of all synthesized CNTs produced onto bimetallic and trimetallic catalysts at 800 °C was evaluated by Raman spectroscopy. All of these spectra show three distinct peaks, including D-band, G-band, and G′-band. The D-band at around 1300-1400 cm −1 is associated with the structural defect and impurity which represents the disordered carbon or amorphous carbon deposited on the outer surface of carbon nanotubes 63 . The G-band at around 1500-1600 cm -1 is related to the tangential stretching mode of all pairs of sp 2 atoms in both rings and chains which represents the graphitic carbon structure 61 , while G'-band observed at around 2600-2700 cm −1 is associated with the process of two-photon elastic scattering 64 . The appearance of both G-band and G'-band can be used as an indicator for assuring the formation of graphite nanotube. This data is in accordance with the morphology observation in Fig. 5. The relation between D-band and G-band in the aspects of intensity ratio Ig/Id is used to evaluate the quality of MWCNTs on the used catalyst, which the low value of the Ig/ Id ratio (> 1) indicates a good degree of graphitization. The results from Raman analyze show that all synthesized CNTs (Ig/Id = 1.14-1.60) presented the Ig/Id ratio higher than commercial CNTs (Ig/Id = 0.74). As seen in Fig. 6,    Figure 7 displays XRD diffraction data for the CNTs generated after conducting in the isothermal test at 800 °C for 3 h. All the pattern results, besides the CoFeMo/MgO catalyst, the intensity of the 002 plane of the graphitic peaks at 2 θ = 26.1° are more prominent compared to the relevant peaks that are assigned to metallic and other phases. This peak is known to quantify the crystallinity of carbon materials. However, in our study, there seems to be more associated with the formation of CNTs, where the sharp diffraction peak shows the high activity of the catalysts for achieving the carbon products, which confirms the higher yield of the carbon nanofilaments grown over these catalysts. The interlayer distance (d 002 ) or d-spacing value derived from the XRD data can be used to the structural degree of synthesized CNTs as well, which was calculated by applying the Bragg's equation ( d = nl/2sinθ) 65,66 and the results are presented in Table 6. It was found that the used catalysts have produced the CNTs with the interplanar spacing value in the range between 0.344 and 0.348 nm, which is very nearly the value    (Figs. 3b, 7). This could be implied that the portion of Ni in the catalyst makes the metal particles to be small, resulting in the increase of the active surface area thus improving the dispersion and catalytic activity of active nanoparticles. After the reaction, The Mo 2 C peaks slightly appeared for all cases, which have been known to be generated during the formation of CNTs via the carbide cycle mechanism 28 . In addition, the presence of MoO 3 in the consumed CoFeMo/MgO catalyst can be ascribed to the oxidation of Co as mentioned above. It might be corresponded to the conversion of Mo into their oxides 26 . The several reflection peaks of Co 3 O 4 and Fe 3 O 4 are detected in the XRD pattern of the Co-and Fe-containing catalysts. This certifies that Co and Fe metallic phases could be oxidized with CO 2 to form metal oxides. Besides, some spot of Fe 2 O 3 , and Fe 3 C peaks were also identified with very low intensities especially on the surface of the CoFeMo/MgO catalyst. This suggested that some Fe particles were possibly converted into Fe 2 O 3 , and Fe 3 C through the oxidation of Fe with carbon dioxide (CO 2 + Fe → Fe 2 O 3 + Fe 3 C). In order to better understand what exactly happens, another two approaches are performed, and latter being thoroughly discussed in "The influence of CO 2 oxidation" section. Moreover, the bigger diameter of CNTs generated with the CoNiMo/MgO, FeNiMo/MgO, and CoFeMo/MgO catalysts can be regarded as the huge proportion of their metal oxide forms hold on the surface of catalysts, which are responsible for the growing of large diameter CNTs. Since the crystal size of oxide species is normally higher than its usual metallic form as obviously seen in Table 6 and TEM images in Fig. 7.
Thermal gravimetric analysis (TGA) was used to determine the quality and the purity of CNTs formed over these catalysts, and the results are depicted in Fig. 8. The weight loss is ascribed to the burning of deposited carbon in oxygen and then corresponds to the yield of solid carbon in the catalysts 67 . Thus, as can be seen in Fig. 8a, the largest weight loss of about 84% was obtained by using the NiMo/MgO catalyst, indicating the presence of high carbon content in the sample, and hence high purity CNTs was observed. The impurities in the CNTs sample were possibly attributed to the residual catalysts that remained after the reaction and other species including oxygen from CO 2 which would possibly be contained in the samples. In general, the classic chemical technique for separating CNTs from other entities such as residual catalysts has been done by treating the CNTs sample with acid leaching 68 . Moreover, the purity of CNTs can be further increased by extending the reaction time as well, as reported in our previous work 26 . The TGA profiles of the carbonaceous materials deposited on all catalysts show similar oxidation behaviors with single-step degradation, regarding the absence of amorphous carbon which is closely related with their textural and structural properties. Additionally, a thorough analysis Table 6. Structural parameters analyzed by XRD for carbon products (interplanar spacing (d 002 )), and other components (crystal size).  67 . Particularly again in the case of the NiMo catalyst, the CNTs was produced with remarkably outstanding thermal stability better than the commercial as demonstrated in Fig. 8b. However, the purity of synthesized CNTs has lower than the commercial CNTs ones. The textural properties of CNTs were assessed by N 2 sorption measurements, the data are listed in Table 7 and shown in Fig. 9 The highest surface area was obtained by synthesized CNTs over NiMo/MgO catalysts. In contrast, the addition of Co into NiMo/MgO catalysts is reduced the BET surface area of synthesized CNTs owing to the oxidation of Co leading to the deactivation of the catalyst. Additionally, the effect of the addition of Fe into NiMo/MgO catalysts results in a decrease in the BET surface area of synthesized CNTs due to Fe could be readily oxidized in exposure CO 2 in Biogas, which could have caused the catalyst deactivation [50][51][52] . For these reasons, the synthesized CNTs over CoFeMo/MgO exhibit the lowest surface area. According to Fig. 9a, N 2 absorption-desorption isotherms of all synthesized CNTs, present type IV isotherms according to IUPAC classification, indicating that the majority of porosity in the CNTs is mesoporous 69 . These results consistent with the mean pore size of synthesized CNTs is in 10-13 nm which is in the range of mesopore material (2-50 nm), as shown in Table 7. The hysteresis loops formed by the capillary condensation effect can be classified as H3 hysteretic loops, usually found in solids consisting of aggregates or agglomerates of particles forming slit-shaped pores, with a non-uniform size and/or shape 70 . The pore diameter distributions based on the BJH method as shown in Fig. 9b present the curves, which can be ascribed that all the synthesized CNTs exhibit a bimodal feature that has a pore size with a board distribution ranging from 2 to 150 nm, including a small pore-size fraction (2-5 nm) and a large pore-size fraction (5-150 nm). Table 7. Summary of surface area and porosity of CNTs synthesized with different catalysts.

Sample
Isotherms S BET (m 2 /g) V meso (cm 3 /g) V total (cm 3 /g) D avg (nm) www.nature.com/scientificreports/ The influence of CO 2 oxidation. One aspect that the influence of CO 2 oxidation was revealed in the former discussion. Herein, an effort is exerted to comprehend how CO 2 oxidation influences Fe-based metal catalysts through performing under two different conditions including CH 4 /He (3/2), CH 4 /CO 2 (3/2), and He/ CO 2 (3/2) feeding into the reactor. All this condition was operated over FeNiMo/MgO catalyst at 800 °C with the flow rate of 400 ml/min, corresponding to GHSV = 48,000 ml/gCat-h. The reaction time was set to 3 h. Figure 10 shows the effluent gas of the outlet stream for each condition. Considering the CH 4 decomposition reaction as seen in Fig. 10a www.nature.com/scientificreports/ were observed as to CH 4 /He condition. These findings assure that operating under the existence of Fe on the catalyst suffered from CO 2 oxidation resulting in deactivation. The graphitization of CNTs produced by the mean of CH 4 decomposition are shown in Fig. 11b. The results show that the synthesized CNTs on CH 4 /He had a higher graphitization degree of carbon structure (Ig/Id = 1.98) in comparison to that obtained by operating at normal CH 4 /CO 2 condition (Ig/Id = 1.60). This may be explained by the decrease in the proportion of Fe metallic phase due to it can be oxidized into Fe oxides with CO 2 , corresponding to the lower level of graphitization in CNTs 47 .
To further observe the catalyst oxidation by CO 2 in biogas (CO 2 + CH 4 ), the two experiments of sole CO 2 and sole CH 4 balanced with He were carried out as shown in Fig. 12. Figure 12c reveals that the CO 2 causes higher degree of oxidation, resulting in the transformation of metal into larger metal oxide particles. The formation of metal oxide particles regards as another key factor that is correlated with CNTs diameter. Thus, the synthesized CNTs using FeNiMo/MgO catalysts had enlarged sizes of CNTs. This is evident by TEM analysis as shown in Fig. 5.

Conclusion
In this study, the bimetallic NiMo, trimetallic CoNiMo, FeNiMo, and CoFeMo supported on MgO were used as catalysts for CNTs growth and syngas production. The experimental results show that the introduction of the third composition (Co or Fe) to NiMo/MgO catalyst does not give any further increase in the catalytic performance. The NiMo/MgO catalyst can perform the highest activity for both reforming and methane cracking. In addition, only slight deactivation of NiMo/MgO catalyst was observed over 3 h. The high purity of syngas was obtained, while the narrowest distribution and the smallest diameter of CNTs with graphitization degree are comparable to those of commercial CNTs. This makes NiMo/MgO a remarkably outstanding catalyst to directly www.nature.com/scientificreports/ convert biogas into syngas and multi-walled carbon nanotubes. In the case of the Fe-containing catalysts, they are known to catalyze the carbon oxidation, causing the agglomeration of oxide phases, and the formation of large diameter CNTs. However, the existence of Fe in the catalyst greatly enhances the bamboo structure with high grades which may be useful in some practical applications.

Data availability
All data related to the finding of this study are accessible upon request from the corresponding author Sakhon Ratchahat.