Fabrication of Pd NPs-supported porous carbon by integrating the reducing reactivity and carbon-rich network of lignin

The renewable resource as a major feedstock to prepare porous carbon has showed many advantages compared to fossil-based materials. This study proposes a new strategy to synthesize palladium nanoparticles (Pd NPs)-supported porous carbon, utilizing both the chemical reactivity and the carbon-rich 3D network of lignin. The Pd NPs-supported porous carbons were prepared in one-pot synthesis, with Pd(NH3)2Cl2 as precursor, lignin as reducing and stabilizing agents of Pd NPs, nano SiO2 as hard-template, followed by carbonization and removal of the template. The results reveal a positive effect of Pd precursor dosage on the development and excellent texture of the Pd NPs-supported porous carbon. Accordingly, the synthesized porous carbon was proved to have large micropore volume and good micro-mesopore porous structure, revealing it a promising hydrogen adsorbent.

The structural change of lignin during the reducing process was illustrated by FT-IR spectra in Fig. 2(b). Compared to original lignin (spectrum III), the band at 3393 cm −1 ascribed to O−H stretching of lignin after reduction (spectrum I and II) was weakened, as well as the absorption band of -CH 2 − and -CH 3 stretching assigned to lignin structure at 2800 cm −1 to 3000 cm −1 , indicating the reduction of hydroxyl within lignin, and some cut offs of lignin structure. In the spectrum of raw lignin, there was a 1710 cm −1 band belonged to the unconjugated C=O stretching of ester. While after the reaction, the unconjugated C=O units of ester were probably transformed to carboxylic anion (COO − ), resulting in the disappearance of 1710 cm −1 band, and the occurrence of 1580 cm −1 band. The band around 1221 cm −1 corresponded to the phenolic hydroxyl stretching of lignin was diminished after the reduction, implying that some phenolic hydroxyl units of lignin may be reacted during Pd NPs formation. Furthermore, compare spectrum I with II, it can be seen with increasing reaction time, the hydroxyl stretching band decreased and shifted to lower wavenumber, which may be a result of interaction between hydroxyl of lignin and the synthesized Pd NPs. structure characterization of pd nanoparticles-supported porous carbon (ppC). The XRD patterns in Fig. 3 show that the characteristic reflection peaks of face-centered cubic structure of metallic Pd (JCPDS, No. 05-0681) of PPC samples are stronger than those of non-carbonized lignin-Pd NPs composites. Besides four peaks at 2θ = 40.1°, 46.6°, 68.1° and 82.0° corresponding to the (111), (200), (220) and (311) facets of PdNPs 15,16 , there was one more peak at 2θ = 87.2° ascribed to (222) facet for the PPC-0.3 sample, as shown in Fig. 3(a). The residual Pd content detected by ICPMS is shown in Table 1, it can find that the Pd contents in Pd NPs-supported lignin-based porous carbon are higher than those of lignin-Pd composite, indicating that the reducing effects of lignin still remained in the one-pot synthesis, and the carbonization process also induced the crystal of Pd. Contrast to the Pd contents of samples prepared with SiO 2 , the samples prepared without SiO 2 have relatively small amount of real Pd content. It indicates that the SiO 2 as a pore-making agent could produce more pores and adsorb more Pd nanoparticles. As we all know, the shapes of peaks are basically dependent on the particle size of samples. Therefore, sharpness of the peaks in Fig. 3(a) means the size distributions of PPC samples are more uniform and concordant than those of non-carbonized lignin-Pd NPs composites in Fig. 3(b).
The XPS spectra in Fig. 4 show the elemental analysis of the PPC with different Pd contents and the samples prepared without SiO 2 with different Pd contents, the binding energies were calibrated by C1s at 284.8 eV. It can be seen there were not only C and O signals ascribed to lignin in the samples, but also the signal of Pd due to the generation of Pd NPs. Compared with lignin-Pd NP composite, PPC samples after carbonization showed much stronger intensity of both C and Pd signals, which are agreement with XRD and ICPMS results. The high-resolution spectrums of Pd3d in Fig. 4c,d reveal the state of Pd element in detail. The peaks at 335.80 eV and 341.10 eV from  www.nature.com/scientificreports www.nature.com/scientificreports/ Pd 0 correspond to the Pd3d 5/2 and Pd3d 3/2 , respectively 17 . In addition, a small proportion of Pd Ox/Pd mixtures at 336.40 eV and 341.70 eV are observed in Fig. 4c, which may be due to the interaction between Pd NPs and oxygen-contained groups within lignin during the carbonization process 18 . It is noted that the Pd3d spectrum of Pd-0.3 sample showed no peaks corresponding to Pd 2+ at 338.00 eV and 343.30 eV, respectively, and indicated the negligible presence of Pd 2+ species. The results confirmed the successful reduction of Pd NPs in one-pot synthesis, even though a small proportion of Pd Ox/Pd mixtures were introduced during carbonization process.
Scanning electron micrographs are utilized to investigate the influence of Pd NPs on the texture properties of the resultant porous carbons. From the SEM images in Fig. 5(a-c), all the PPC samples exhibited excellent 3D porous structure, which are similar to the LPC sample shown in Fig. 5d. It is noteworthy that with the increasing Pd content, the porous structure of PPC sample seemed to be looser. It seems that when there were more Pd NPs dispersed in the network of lignin, there would be less interaction sites between SiO 2 template and lignin molecules, so that looser structure would be obtained during carbonization process 13 .
Transmission electron microscope images in Fig. 6a show the porous structure of PPC-0.3 sample in large area, while the TEM images of LPC in Fig. 6f exhibit mesopores of 10 nm to dozens of nanometers in diameter, which are agreement with SEM results. In Fig. 6b, it can be observed that large amounts of small Pd NPs were homogeneously embedded in the carbon network and had good dispersity. And in the high-resolution TEM image (Fig. 6c), the spherical Pd NPs can be clearly seen, with similar shape and size to Pd NPs prepared with lignin as the reducing agent as shown in Fig. 6d,e, implying that carbonization process hardly influenced the shape and size of Pd NPs. The crystal detail of one Pd nanoparticle was further analyzed by high-resolution transmission electron microscopy (HRTEM). Figure 6e shows the grain boundaries, faults and dislocations on the surfaces of Pd nanoparticle. The measured d-spacing of 0.22 nm and 0.19 nm as shown in Fig. 6e corresponded to the (111) and (200) planes of the Pd crystals 19 , respectively, which is agreement with the XRD results.
properties of pd nanoparticles-supported porous carbon (ppC) for hydrogen storage. Ligninbased porous carbon samples without Pd NPs were prepared under different lignin/SiO 2 ratios and carbonization temperatures, as given in Table S1. As shown in Figs S1, S2 and Table S2, it was found that the LPC samples with lignin/SiO 2 ratios of 1:2 at 750 °C showed BET specific surface area of 893 m 2 /g, which is relatively high among lignin-based porous carbon materials without activation process 4,13,20 . Therefore, the Pd NPs-supported porous carbons were also prepared with lignin/SiO 2 ratios of 1:2 at 750 °C. The structural properties of Pd NPs-supported porous carbons and LPC are evaluated by nitrogen adsorption-desorption isotherms, and the results are shown in Table 2 and Fig. 7. Nitrogen adsorption-desorption isotherms at 77 K (−196 °C) for PPC samples with different Pd contents are presented in Fig. 7a. As shown in Fig. 7a, the PPC samples had IV Type isotherms, according to the IUPAC nomenclature 21 , and had clearer hysteresis loops of H3 shapes than LPC sample, confirming the presence of mesopores. Moreover, the rapid adsorption in low relative pressure range (smaller than P/P 0 = 0.1) also represented the presence of micropores 22 . Thus, the as-prepared PPC samples can be identified as micro-mesopore porous materials 23 . And with increasing residual Pd contents in final samples, the nitrogen adsorption of PPC samples increased, which is consistent with the results of BET specific surface area and total pore volume, as listed www.nature.com/scientificreports www.nature.com/scientificreports/ in Table 2. It is noteworthy that the BET specific surface area of PPC-0.3 was 1467 m 2 /g, which is much higher than that of LPC sample without supporting Pd NPs. This is because that high hydrogen spillover effects produced by residual Pd contents in samples greatly increased their specific surface area. And this indicated the improvement of texture properties by high Pd loading, which is agreement with the observed pore results of SEM and TEM. The mesopores size of PPC samples can be seen from the size distribution curves (Fig. 7b) by BJH method, as listed in Table 2. From the size distribution curves, two peaks were observed at around 15 and 30 nm, implying that there may be two types of mesopores structures within PPC samples. Moreover, with more Pd contents in lignin-Pd NPs composites, the mesopores size of resultant PPC sample was larger. The pore size distribution curves in the micropore range by the DFT method (Fig. 7c) indicate that under the same template dosage and carbonization temperature, higher Pd loading contents had positive impact on micropore volume (see Table 2).
The hydrogen storage capacities of the Pd nanoparticle-supported porous carbons were conducted and the results are shown in Table 2. It can be seen the H 2 uptake capacity increases in the order of PPC-0.1 < PPC-0.2 < PPC-0.3, which means that with increasing residual Pd contents in final samples, the hydrogen storage capacity increased. These results can be ascribed to the process called spillover effect, which generally happens in  www.nature.com/scientificreports www.nature.com/scientificreports/ such Pd doped samples 11,24 . When Pd NPs were dispersed in lignin-derived porous carbon, apart from physisorption of hydrogen in porous carbon, there was also chemisorption of hydrogen in Pd NPs, result in the formation of H atoms. The H atoms spill over towards the pores in PPC samples. Thus, higher Pd contents lead to stronger spillover effect, and enhanced hydrogen storage capacity 5,25 . Furthermore, it can be seen from Table 1 and Table 2 that PPC-0.3 sample with the highest micropore volume (0.18 cm 3 /g) exhibited much higher H 2 uptake than the samples with lower micropore volume (0.06 cm 3 /g and 0.08 cm 3 /g). These results indicate that micropore volume of the porous carbons also has remarkable effect on the hydrogen storage capacity. When compared with some carbonaceous materials under similar condition (Table 2), the H 2 uptake of prepared PPC samples was lower than Pd/3D porous carbon 9 , and activated carbon aerogels (with ultrahigh surface area) 26 . However, the PPC samples performed better than some metal-doped carbon materials such as Co-modified carbon aerogels 26 , Pd-carbon nanotubes 27 , nanostructured carbon-Pd 28 , Pd-decorated carbon nanotubes 29 , which was 3.52 wt. % under 77 K and 60 bar. And the maximum H 2 uptake can reach to 3.98 wt. % under 77 K and 180 bar. As shown in Table 2, with the increase of H 2 uptake temperature from 77 K to 298 K, H 2 uptake capacity of all prepared samples decreased. In addition, the N 2 adsorption isosteric technique was also applied to determine the isosteric heat with the calorimeter at 298 K for each sample as shown in Table 3. For N 2 , the isosteric heats decreased continuously with the increase of coverage. The parameter which may influence the H 2 uptake is the surface chemistry. It can be explained that when the temperature is higher, higher internal energy and repulsive energy between hydrogen molecule make the adsorption reaction more difficult, and influence the spillover of hydrogen on prepared materials 28 . In this study, PPC samples were prepared without activation process after carbonization. It is believed that under optimized carbonization and activation process in upcoming study, the hydrogen storage capacity of PPC will be much better than now.  Table 2. Hydrogen uptake of the PPC samples.  Table 3. Isosteric heats of adsorption of N 2 on the PPC samples.

Conclusions
A green strategy was applied for the synthesis of Pd NPs-supported porous carbon, utilizing lignin as not only reducing and stabilizing agents, but also the carbon source. The Pd NPs-supported porous carbon synthesized via hard-template method exhibited 3D interconnected mesoporous/macroporous carbon networks over a large area. The loading of Pd NPs was proved to increase the hydrogen storage capacity due to the spillover effects. Thus, the present study not only provides a promising hydrogen adsorbent based on lignin, more importantly, proposes a greener and more available strategy for efficient utilization of renewable materials.  Table 4 exhibits the preparation conditions and residual Pd contents of all carbon materials. For the preparation of Palladium nanoparticle-supported porous carbon, lignin/SiO 2 mixture (1.5 g) with mass ratio of 1:2 was firstly dispersed in 1 wt. % NaOH solution. Pd(NH 3 ) 2 Cl 2 powder (0.1 mmol, 0.2 mmol, 0.3 mmol) was also dissolved in 20 mL 1 wt. % NaOH solution to obtain the Pd solution. The lignin/SiO 2 dispersion was then added to the Pd solution, and the whole reactant was reacted under microwave irradiation (800 W) at 90 °C for 4 h. After lyophilization, the resultant samples were carbonized under 750 °C for 2 h under nitrogen flow (heating rate of 5 °C/min, flow rate = 60 cm 3 /min). After dissolving the SiO 2 template in 10% HF at 80 °C for 6 h, the products were washed by deionized water and dried in the oven at 60 °C for 24 h. The resulting carbon materials were designated as PPC-p, where the p stands for different Pd dosages (mmol).

synthesis of pd nanoparticles-supported porous carbon (ppC).
For comparison, Pd NPs prepared with lignin as the reducing agent was conducted without SiO 2 and prepared under the same microwave irradiation procedure, and the resultant black solution was dialyzed by ultrapure water and the solid product was obtained after lyophilization at −40 °C. The obtained products were denoted as Pd-0.1, Pd-0.2, Pd-0.3 respectively, corresponding to different Pd dosages. In addition, the lignin/SiO 2 mixture was also treated without Pd NPs under the same recipe and procedure for comparison, and the as-prepared carbon was labeled as LPC.
Characterization of pd nanoparticles-supported porous carbon (ppC). The UV-Vis spectrometry analysis was performed using a UV-1800 spectrophotometer (Shimadzu, Japan) with a scan range of 200~500 nm at a scan interval of 0.5 nm. Fourier transform infrared (FT-IR) analysis was performed using a Nicolet 5700 spectrophotometer (Madison, USA) under a dry air at room temperature by the KBr pellet method. The spectra were collected over the range of 4000 to 400 cm −1 . X-ray diffractometer (XRD) analysis was performed on a D8 Advance XRD system (Cu Kα radiation, 40 kV, 50 Ma) (Bruker, Germany). X-ray photoelectron spectroscopy (XPS) analysis was performed using an AXIS Ultra DLD spectrometer (Kratos, UK) with Mg Kα radiation (hγ = 1253.6 eV) in step size of 0.1 eV. Transmission electron microscopy (TEM) images were obtained using a JEM-2100 (JEOL, Japan) instrument at an accelerating voltage of 200 kV. The residual Pd contents in final samples were detected using an Inductively Coupled Plasma Mass Spectrometer (ICPMS) (5300DV, Perkin Elmer, USA). Particle size analysis of Pd NPs was performed using a Malvern 3000HSA analyzer (Malvern, England). Pore textural analysis of lignin-based porous carbon was performed using an ASAP 2020 volumetric analyzer (Micromeritics, USA). The specific surface area was calculated using the Brunauer-Emmett-Teller (BET) method applied to adsorption data within the relative pressure (P/Po) range of 0.05-0.2. The total pore volume was calculated from the amount of nitrogen adsorbed at a relative pressure of 0.99, and the micropore volume was evaluated by t-plot method. The mesopore volume was obtained by subtraction of the micropore volume from the total pore volume. The pore size distribution (PSD) analysis was obtained via the Barrett-Joyner-Halenda (BJH) method from the desorption branch. In addition, the micro porosity of the porous carbon was estimated by the classical density functional theory (DFT) method using the nitrogen adsorption data.  Table 4. Preparation conditions and residual Pd contents of all carbon materials.