*Article* **Bimetallic PdCo Nanoparticles Loaded in Amine Modified Polyacrylonitrile Hollow Spheres as Efficient Catalysts for Formic Acid Dehydrogenation**

**Yulin Li 1,†, Ping She 1,†, Rundong Ding <sup>1</sup> , Da Li <sup>2</sup> , Hongtan Cai <sup>1</sup> , Xiufeng Hao 1,\* and Mingjun Jia <sup>1</sup>**


**Abstract:** Polyacrylonitrile hollow nanospheres (HPAN), derived from the polymerization of acrylonitrile in the presence of polystyrene emulsion (as template), were modified by surface amination with ethylenediamine (EDA), and then used as support for loading Pd or PdCo nanoparticles (NPs). The resultant bimetallic catalyst (named PdCo0.2/EDA-HPAN) can efficiently catalyze the additivefree dehydrogenation of formic acid with very high activity, selectivity and recyclability, showing turnover frequencies (TOF) of 4990 h−<sup>1</sup> at 333 K and 915 h−<sup>1</sup> at 303 K, respectively. The abundant surface amino groups and cyano group as well as the hollow structure of the support offer a suitable environment for achieving high dispersion of the Pd-based NPs on the surface of EDA-HPAN, thus generating ultra-small bimetallic NPs (bellow 1.0 nm) with high stability. The addition of a small portion of Co may adjust the electronic state of Pd species to a certain extent, which can further improve their capability for the dehydrogenation of formic acid. In addition, the surface amino groups may also play an important role in synergistically activating formic acid to generate formate, thus leading to efficient conversion of formic acid to hydrogen at mild conditions.

**Keywords:** formic acid decomposition; polyacrylonitrile hollow nanospheres; ethylenediamine modification; palladium cobalt nanoparticles

## **1. Introduction**

Hydrogen production and storage are very topical issues in the field of green energy, which is a novel field in engineering with a target to develop idealistic energy systems that have no negative environmental, economic and societal impacts [1,2]. Formic acid (FA) is the simplest and most easily available carboxylic acid, which has the characteristics of relatively high hydrogen content (53 g H2/L), low toxicity and easy storage [3,4]. Since FA can readily release hydrogen (H2) and carbon dioxide (CO2) on demand under mild conditions in presence of suitable catalysts, it has been considered as a safe and convenient hydrogen storage compound [3–9]. Moreover, the concomitant CO<sup>2</sup> can also be reutilized to regenerate FA by combining another catalytic process of CO<sup>2</sup> hydrogenation, thus being of great significance to achieve the hydrogen storage cycle and CO<sup>2</sup> utilization [10–12]. During the past decade, considerable efforts have been devoted to developing highly efficient and stable heterogeneous catalysts for FA dehydrogenation, owing to its great potential in sustainable energy storage and utilization [13–20].

Among the numerous reported catalysts, supported Pd nanoparticles (NPs) catalysts have received increased attention because of their relatively high catalytic activity and dehydrogenation selectivity for FA decomposition [19,20]. In order to enhance the catalytic efficiency and the structure stability of Pd NPs (against aggregation), more recent work has

**Citation:** Li, Y.; She, P.; Ding, R.; Li, D.; Cai, H.; Hao, X.; Jia, M. Bimetallic PdCo Nanoparticles Loaded in Amine Modified Polyacrylonitrile Hollow Spheres as Efficient Catalysts for Formic Acid Dehydrogenation. *Catalysts* **2022**, *12*, 33. https:// doi.org/10.3390/catal12010033

Academic Editors: Florica Papa, Anca Vasile and Gianina Dobrescu

Received: 13 December 2021 Accepted: 27 December 2021 Published: 29 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

been focused on optimizing the dispersity, electronic state, and coordination environments of Pd species. One of the most effective ways is to addition of a second or even third metal component, which can usually have a positive effect on the physicochemical properties, including catalytic performance of the supported Pd-based NPs [21–26]. For instance, Jiang et al. reported that a carbon supported Co0.30Au0.35Pd0.35 nanoalloy could be used as an active and stable catalyst for additive-free FA dehydrogenation with an initial TOF of 80 h−<sup>1</sup> at room temperature [21]. Wang et al. prepared carbon black-supported PdCo-based nanocatalysts, which can efficiently catalyze FA dehydrogenation at room temperature with sodium formate as a promoter [22]. They proposed that the electron status of the Pd surface could be modified by Co, leading to the decrease in Pd 3d binding energy, which can enhance the CO anti-toxicity ability of Pd [22].

In addition, it was also well known that changing the types of supports or adjusting their surface properties may considerably influence the activity and stability of the supported Pd based NPs catalysts [27–31]. Previous works in the literature demonstrated that introducing some basic groups (such as amino groups) into porous supports could be beneficial for achieving high dispersion of Pd-based NPs and enhancing the structure stability of the metal NPs by building suitable metal-support interaction and space-confined effect [32–40]. For instance, Lu et al. reported that amine-functionalized mesoporous SBA-15 supported bimetallic PdIr NPs exhibited very high catalytic activity and 100% H<sup>2</sup> selectivity for FA dehydrogenation. The amine functional groups of SBA-15-NH<sup>2</sup> can interact strongly with the metal ions, thus leading to the formation of ultrafine and stable PdIr NPs distributed inside the mesopores of the support [32]. Moreover, the surface amino groups may also serve as basic sites to facilitate the O-H bond dissociation in the FA molecule, thus acting as a co-catalyst for achieving FA dehydrogenation at mild condition [32–37]. Currently, it is still a highly attractive subject to modify the surface properties and the morphologies of some easily available supports for constructing more efficient and stable supported Pd-based catalysts.

Previously, the authors and our coworkers also carried out some studies on the preparation of supported Pd-based NPs catalysts for FA dehydrogenation [41–44]. Interestingly, by using low-cost polymer materials such as polyacrylonitrile (PAN) and modified PAN as supports, relatively active supported Pd, Pd-Me (Me = Fe, Co, or Ni) NPs catalysts were obtained [41–43]. It was found that modification of PAN with a suitable amount of ethylenediamine could further improve the dispersity and stability of Pd NPs, thus resulting in formation of a highly active supported Pd catalyst with ultra-small particle size (abound 1.2 nm), which can work well for FA dehydrogenation with a TOF 688 h−<sup>1</sup> at 303 K [41]. As a continuation of the above work, we here tried to use ethylenediamine modified PAN hollow nanospheres (HPAN) as support to prepare supported Pd and PdCo NPs catalysts. A variety of characterization results demonstrated that the as-synthesized EDA-HPAN could show great advantages in generating and stabilizing the ultrasmall Pd-based NPs with average particle size below 1 nm, thus leading to the formation of more active and stably supported Pd-based NPs catalysts for FA dehydrogenation at ambient temperature.

#### **2. Results and Discussion**

## *2.1. SEM Studies*

The HPAN support with hollow sphere structure was synthesized on the basis of a literature work reported for synthesis of polystyrene/polyacrylonitrile composite nanospheres (PS@PAN) [45]. The preparation process of Pd/EDA-HPAN and PdCox/EDA-HPAN catalysts can be divided into three steps: the preparation of HPAN, the synthesis of aminefunctionalized HPAN, and the synthesis of Pd/EDA-HPAN and PdCox/EDA-HPAN (Scheme 1). As shown in Figure 1, the SEM images of PS, PS@PAN, HPAN and EDA-HPAN reveal that all the samples possess microsphere morphology. The PS microspheres have a smooth surface with particle size around 300–400 nm (Figure 1a). Core-shell PS@PAN formed by polymerization of AN on the surface of PS core shows a larger particle size

(500–600 nm) and a coarse surface (Figure 1b). The support of HPAN hollow sphere, which was obtained by dissolving the PS core of the PS@PAN with THF, shows a much rougher surface with sphere morphology. The hollow structure of HPAN could be confirmed by the appearance of some semi-spherical shell in the SEM image of HPAN (Figure 1c). After modification with EDA, both the particle size and the morphology of EDA-HPAN (Figure 1d) have no obvious change in comparison with the parent of HPAN, indicating that amino modification has little effect on the structure and the morphology of the polymer support. hollow sphere, which was obtained by dissolving the PS core of the PS@PAN with THF, shows a much rougher surface with sphere morphology. The hollow structure of HPAN could be confirmed by the appearance of some semi-spherical shell in the SEM image of HPAN (Figure 1c). After modification with EDA, both the particle size and the morphology of EDA-HPAN (Figure 1d) have no obvious change in comparison with the parent of HPAN, indicating that amino modification has little effect on the structure and the morphology of the polymer support. hollow sphere, which was obtained by dissolving the PS core of the PS@PAN with THF, shows a much rougher surface with sphere morphology. The hollow structure of HPAN could be confirmed by the appearance of some semi-spherical shell in the SEM image of HPAN (Figure 1c). After modification with EDA, both the particle size and the morphology of EDA-HPAN (Figure 1d) have no obvious change in comparison with the parent of HPAN, indicating that amino modification has little effect on the structure and the morphology of the polymer support.

microspheres have a smooth surface with particle size around 300–400 nm (Figure 1a). Core-shell PS@PAN formed by polymerization of AN on the surface of PS core shows a larger particle size (500–600 nm) and a coarse surface (Figure 1b). The support of HPAN

microspheres have a smooth surface with particle size around 300–400 nm (Figure 1a). Core-shell PS@PAN formed by polymerization of AN on the surface of PS core shows a larger particle size (500–600 nm) and a coarse surface (Figure 1b). The support of HPAN

*Catalysts* **2022**, *12*, x FOR PEER REVIEW 3 of 13

*Catalysts* **2022**, *12*, x FOR PEER REVIEW 3 of 13

**Scheme 1.** Preparation route of PdCox/EDA-HPAN. Step 1: removal of PS template by THF to obtain HPAN. Step 2: EDA-HPAN obtained by EDA modification. Step 3: PdCo/EDA-HPAN obtained by impregnation method. **Scheme 1.** Preparation route of PdCox/EDA-HPAN. Step 1: removal of PS template by THF to obtain HPAN. Step 2: EDA-HPAN obtained by EDA modification. Step 3: PdCo/EDA-HPAN obtained by impregnation method. **Scheme 1.** Preparation route of PdCox/EDA-HPAN. Step 1: removal of PS template by THF to obtain HPAN. Step 2: EDA-HPAN obtained by EDA modification. Step 3: PdCo/EDA-HPAN obtained by impregnation method.

**Figure 1.** SEM images of the fresh samples. (**a**) PS, (**b**) PS@PAN, (**c**) HPAN, (**d**) EDA-HPAN. **Figure 1. Figure 1.**  SEM images of the fresh samples. ( SEM images of the fresh samples. (**aa** ) PS, ( ) PS, ( **b b** ) PS@PAN, ( ) PS@PAN, (**cc**) HPAN, ( ) HPAN, ( **d d** ) EDA-HPAN. ) EDA-HPAN.

#### *2.2. TEM Studies*

The TEM images of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN show that a large number of ultra-small Pd-based NPs are uniformly dispersed on the surface of the EDA- HPAN (Figure 2). The Pd-based NPs observed in the region of the two images have an average diameter of about 0.88 nm and 0.81 nm, respectively, smaller than the previous reported EDA-PAN supported Pd NPs (1.2 nm) [41]. This might be mainly related to the fact that the HPAN support with hollow sphere structure has a rougher spherical surface and a higher specific surface area, as revealed by the following N<sup>2</sup> adsorption-desorption measurements, which can provide more space for the uniform distribution of the amino groups as well as the introduced metal NPs. It is generally believed that the well-distributed amino groups on the solid support have a strong ability to coordinate with the cations (such as Pd2+ and Co2+), thus playing an important role in the nucleation and growth of Pd-based NPs, and, meanwhile, inhibiting the aggregation of the generated small NPs during the NaBH<sup>4</sup> reduction process. This results in a reduction in the size of Pd NPs. Compared with Pd/EDA-HPAN, the size of metal NPs in the image of PdCo0.2/EDA-HPAN is slightly smaller, indicating that the introduction of a small amount of the second metal component may effectively decrease the particle size of Pd-based NPs. This might be mainly caused by the alloy effect, which can adjust the electronic state of Pd-based NPs to a certain extent, and then bring about a positive effect on achieving high dispersion of the bimetallic NPs [42]. diameter of about 0.88 nm and 0.81 nm, respectively, smaller than the previous reported EDA-PAN supported Pd NPs (1.2 nm) [41]. This might be mainly related to the fact that the HPAN support with hollow sphere structure has a rougher spherical surface and a higher specific surface area, as revealed by the following N2 adsorption-desorption measurements, which can provide more space for the uniform distribution of the amino groups as well as the introduced metal NPs. It is generally believed that the well-distributed amino groups on the solid support have a strong ability to coordinate with the cations (such as Pd2+ and Co2+), thus playing an important role in the nucleation and growth of Pd-based NPs, and, meanwhile, inhibiting the aggregation of the generated small NPs during the NaBH4 reduction process. This results in a reduction in the size of Pd NPs. Compared with Pd/EDA-HPAN, the size of metal NPs in the image of PdCo0.2/EDA-HPAN is slightly smaller, indicating that the introduction of a small amount of the second metal component may effectively decrease the particle size of Pd-based NPs. This might be mainly caused by the alloy effect, which can adjust the electronic state of Pd-based NPs to a certain extent, and then bring about a positive effect on achieving high dispersion of the bimetallic NPs [42].

The TEM images of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN show that a large number of ultra-small Pd-based NPs are uniformly dispersed on the surface of the EDA-HPAN (Figure 2). The Pd-based NPs observed in the region of the two images have an average

*Catalysts* **2022**, *12*, x FOR PEER REVIEW 4 of 13

*2.2. TEM Studies* 

**Figure 2.** TEM images and histograms of the PdCo nanoparticles size distribution for the fresh catalyst (**a**) Pd/EDA-HPAN and (**b**) PdCo0.2/EDA-HPAN. **Figure 2.** TEM images and histograms of the PdCo nanoparticles size distribution for the fresh catalyst (**a**) Pd/EDA-HPAN and (**b**) PdCo0.2/EDA-HPAN.

#### *2.3. IR Spectra Analysis of Samples 2.3. IR Spectra Analysis of Samples*

Figure 3 shows the infrared spectra of PS, PS@PAN, HPAN, EDA-HPAN and Pd2+CO2+/EDA-HPAN, respectively. For the sample of PS sphere (Figure 3a), the stretching vibration peak of C-H on the benzene ring appears at 3060 cm−1, the absorption peak at 2922 cm−1 proves the existence of methylene, while the signals appeared at 1452 and 727 cm−1 are related to the skeleton vibration peak of the benzene ring and the out of plane bending vibration absorption peak of C-H, respectively [46,47]. As for PAN (Figure 3b), a new absorption peak of 2242 cm−1 appears, which belongs to the C≡N functional group in PAN. In the spectrum of HPAN (Figure 3c), the disappearance of the characteristic peak of PS suggests that the PS core has been completely removed after treating the precursor of PS@PAN with THF. Combined with the above characterization results of SEM and FTIR, it could be deduced that the resultant HPAN should possess hollow sphere structure after dissolving the inside core (PS sphere) from PS@PAN. For EDA-HPAN (Figure 3d) and Pd2+Co2+/EDA-HPAN (Figure 3e), the signals appear at 1578, 1633 and 1669 cm−<sup>1</sup> should originate from the stretching vibration peaks of N-H, NH2 and N-C=N, respectively [48]. Moreover, as shown in Figure S1, the TG curves of HPAN and EDA-HPAN are well consistent with that of PAN reported in the literature [49,50], indicating that the Figure 3 shows the infrared spectra of PS, PS@PAN, HPAN, EDA-HPAN and Pd2+CO2+/ EDA-HPAN, respectively. For the sample of PS sphere (Figure 3a), the stretching vibration peak of C-H on the benzene ring appears at 3060 cm−<sup>1</sup> , the absorption peak at 2922 cm−<sup>1</sup> proves the existence of methylene, while the signals appeared at 1452 and 727 cm−<sup>1</sup> are related to the skeleton vibration peak of the benzene ring and the out of plane bending vibration absorption peak of C-H, respectively [46,47]. As for PAN (Figure 3b), a new absorption peak of 2242 cm−<sup>1</sup> appears, which belongs to the C≡N functional group in PAN. In the spectrum of HPAN (Figure 3c), the disappearance of the characteristic peak of PS suggests that the PS core has been completely removed after treating the precursor of PS@PAN with THF. Combined with the above characterization results of SEM and FTIR, it could be deduced that the resultant HPAN should possess hollow sphere structure after dissolving the inside core (PS sphere) from PS@PAN. For EDA-HPAN (Figure 3d) and Pd2+Co2+/EDA-HPAN (Figure 3e), the signals appear at 1578, 1633 and 1669 cm−<sup>1</sup> should originate from the stretching vibration peaks of N-H, NH<sup>2</sup> and N-C=N, respectively [48]. Moreover, as shown in Figure S1, the TG curves of HPAN and EDA-HPAN are well consistent with that of PAN reported in the literature [49,50], indicating that the modification of the PAN with EDA did not have an obvious effect on the thermal stability of the polymer support. This might be attributed to the fact that only a small amount of EDA molecules was anchored on the surface of the HPAN support, through a reaction between the -C≡N group in PAN and the amino group in EDA. This point could be further confirmed by the fact that the nitrogen content in EDA-HPAN reached 25.34 wt %, higher than that in HPAN

(23.83 wt %), as given in Table 1. After the introduction of Pd2+ and Co2+ species, the peak position of N-H bond in the spectrum of Pd2+Co2+/EDA-HPAN shifts slightly from 1578 cm−<sup>1</sup> to 1560 cm−<sup>1</sup> , which may be due to the coordination action between the metal cations (Pd2+ and Co2+) and the amino groups distributed on the surface of EDA-HPAN (Figure 3e). than that in HPAN (23.83 wt %), as given in Table 1. After the introduction of Pd2+ and Co2+ species, the peak position of N-H bond in the spectrum of Pd2+Co2+/EDA-HPAN shifts slightly from 1578 cm−1 to 1560 cm−1, which may be due to the coordination action between the metal cations (Pd2+ and Co2+) and the amino groups distributed on the surface of EDA-HPAN (Figure 3e).

modification of the PAN with EDA did not have an obvious effect on the thermal stability of the polymer support. This might be attributed to the fact that only a small amount of EDA molecules was anchored on the surface of the HPAN support, through a reaction between the -C≡N group in PAN and the amino group in EDA. This point could be further confirmed by the fact that the nitrogen content in EDA-HPAN reached 25.34 wt %, higher

*Catalysts* **2022**, *12*, x FOR PEER REVIEW 5 of 13

**Figure 3.** The infrared spectra of (**a**) PS (**b**) PS@PAN (**c**) HPAN (**d**) EDA-HPAN (**e**) Pd2+Co2+/EDA-HPAN. **Figure 3.** The infrared spectra of (**a**) PS (**b**) PS@PAN (**c**) HPAN (**d**) EDA-HPAN (**e**) Pd2+Co2+/EDA-HPAN.

**Table 1.** The N contents, Pd loading amount and the specific surface areas of various samples. **Table 1.** The N contents, Pd loading amount and the specific surface areas of various samples.


#### *2.4. Elemental Analyses and Structure Properties of Samples 2.4. Elemental Analyses and Structure Properties of Samples*

The ICP-AES results show that the loading amounts of Pd species in Pd/EDA-HPAN and PdCo0.2/EDA-HPAN samples are 3.12 wt % and 3.09 wt %, respectively (Table 1), which are very close to the theoretical value of 3.19 wt %. It can be inferred from the TEM image in Figure 2 that the amine groups introduced by amination should be uniformly distributed on the surface of EDA-HPAN. Combined with results of the N2 adsorption isotherms of HPAN and EDA-HPAN, the specific surface areas of the HPAN and EDA-HPAN supports calculated by the Brunauer-Emmett-Teller method are 45 and 48 m2/g, respectively (Table 1), which are higher than those of coral-like PAN and EDA-PAN reported in our previous work [41]. The ICP-AES results show that the loading amounts of Pd species in Pd/EDA-HPAN and PdCo0.2/EDA-HPAN samples are 3.12 wt % and 3.09 wt %, respectively (Table 1), which are very close to the theoretical value of 3.19 wt %. It can be inferred from the TEM image in Figure 2 that the amine groups introduced by amination should be uniformly distributed on the surface of EDA-HPAN. Combined with results of the N<sup>2</sup> adsorption isotherms of HPAN and EDA-HPAN, the specific surface areas of the HPAN and EDA-HPAN supports calculated by the Brunauer-Emmett-Teller method are 45 and 48 m2/g, respectively (Table 1), which are higher than those of coral-like PAN and EDA-PAN reported in our previous work [41].

The XRD patterns of HPAN, EDA-HPAN, Pd/EDA-HPAN and PdCo0.2/EDA-HPAN samples are shown in Figure 4. All the samples present a strong signal at 17°, which is the The XRD patterns of HPAN, EDA-HPAN, Pd/EDA-HPAN and PdCo0.2/EDA-HPAN samples are shown in Figure 4. All the samples present a strong signal at 17◦ , which is the characteristic peak of PAN appearing at 17◦ [41,51]. Compared with the support of HPAN and EDA-HPAN, the characteristic peaks in the supported Pd-based catalysts became weaker due to the incorporation of a certain amount of metal NPs. It is worth noting that the characteristic peaks of the metal NPs are not detected in the patterns, confirming further the high dispersion of the small metal NPs on the surface of the polymer support.

**Figure 4.** XRD patterns of (**a**) HPAN, (**b**) EDA-HPAN, (**c**) Pd/EDA-HPAN and (**d**) PdCo0.2/EDA-HPAN. **Figure 4.** XRD patterns of (**a**) HPAN, (**b**) EDA-HPAN, (**c**) Pd/EDA-HPAN and (**d**) PdCo0.2/EDA-HPAN.

characteristic peak of PAN appearing at 17° [41,51]. Compared with the support of HPAN and EDA-HPAN, the characteristic peaks in the supported Pd-based catalysts became weaker due to the incorporation of a certain amount of metal NPs. It is worth noting that the characteristic peaks of the metal NPs are not detected in the patterns, confirming further the high dispersion of the small metal NPs on the surface of the polymer support.

#### *2.5. X-Ray Photoelectron Spectroscopy Analyses 2.5. X-ray Photoelectron Spectroscopy Analyses*

The Pd 3d XPS spectra of the Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts were shown in Figure 5. There are two groups of symmetrical characteristic peaks in the energy spectra, corresponding to the existence of two kind of Pd species in both catalysts. For Pd/EDA-HPAN (Figure 5a), the peaks at 342.8 and 337.8 eV are related to the Pd 3d3/2 and Pd 3d5/2 characteristic peaks of Pd2+, while the peaks at 341.1 and 335.9 eV are associated to the Pd 3d3/2 and Pd 3d5/2 characteristic peaks of Pd0. For PdCo0.2/EDA-HPAN (Figure 5b), both kinds of Pd species are still present, while the signals belonging to Pd0 shift slightly toward lower binding energies compared with the single metal catalyst, possibly related to the alloy effect of Pd and Co, i.e., some electrons transferred from Co to Pd [21]. Combined with the results of the above TEM images (Figure 2), it can be deduced that the Pd NPs with ultra-small particle size should be easily oxidized, thus resulting in the appearance of a number of oxidized Pd2+ species in the supported Pd-based catalyst. It can be seen from Figure 5c that the characteristic signals of Co 2p in Co/EDA-HPAN appear at 780.9 and 796.4 eV, while the signals of Co 2p in PdCo0.2/EDA-HPAN move upward, appearing at 781.1 and 797.3 eV, respectively. Combined with the above results of Pd 3d spectra, it can be proposed that there is indeed a strong electron interaction between Pd and Co in PdCo0.2/EDA-HPAN. Such strong interaction could lead to the change of the electronic state of palladium, which may considerably influence the catalytic activity and stability of the supported Pd-based NPs. The Pd 3d XPS spectra of the Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts were shown in Figure 5. There are two groups of symmetrical characteristic peaks in the energy spectra, corresponding to the existence of two kind of Pd species in both catalysts. For Pd/EDA-HPAN (Figure 5a), the peaks at 342.8 and 337.8 eV are related to the Pd 3d3/2 and Pd 3d5/2 characteristic peaks of Pd2+, while the peaks at 341.1 and 335.9 eV are associated to the Pd 3d3/2 and Pd 3d5/2 characteristic peaks of Pd<sup>0</sup> . For PdCo0.2/EDA-HPAN (Figure 5b), both kinds of Pd species are still present, while the signals belonging to Pd<sup>0</sup> shift slightly toward lower binding energies compared with the single metal catalyst, possibly related to the alloy effect of Pd and Co, i.e., some electrons transferred from Co to Pd [21]. Combined with the results of the above TEM images (Figure 2), it can be deduced that the Pd NPs with ultra-small particle size should be easily oxidized, thus resulting in the appearance of a number of oxidized Pd2+ species in the supported Pd-based catalyst. It can be seen from Figure 5c that the characteristic signals of Co 2p in Co/EDA-HPAN appear at 780.9 and 796.4 eV, while the signals of Co 2p in PdCo0.2/EDA-HPAN move upward, appearing at 781.1 and 797.3 eV, respectively. Combined with the above results of Pd 3d spectra, it can be proposed that there is indeed a strong electron interaction between Pd and Co in PdCo0.2/EDA-HPAN. Such strong interaction could lead to the change of the electronic state of palladium, which may considerably influence the catalytic activity and stability of the supported Pd-based NPs.

**Figure 5.** XPS spectra of (**a**) Pd/ EDA-HPAN, (**b**) PdCo0.2/EDA-HPAN, (**c**) Co/EDA-HPAN, (**d**) PdCo0.2/EDA-HPAN. **Figure 5.** XPS spectra of (**a**) Pd/ EDA-HPAN, (**b**) PdCo0.2/EDA-HPAN, (**c**) Co/EDA-HPAN, (**d**) PdCo0.2/EDA-HPAN.

#### *2.6. Evaluation of Catalytic Activity*

*2.6. Evaluation of Catalytic Activity*  The catalytic properties of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts were measured for the dehydrogenation of FA. As shown in Figure 6a,b, both Pd/EDA-HPAN and PdCo0.2/EDA-HPAN exhibit good catalytic activity in the temperature range of 303– 353 K, and the dehydrogenation rate of PdCo0.2/EDA-HPAN is higher than that of Pd/EDA-HPAN. No CO signal was detected by GC analysis during the reaction term (Figure S2). In addition, after CO2 was absorbed by NaOH solution trap (10 mol/L), the total volume of gas product was reduced by half (Figure S3), indicating that the catalyst had excellent dehydrogenation selectivity. It should be pointed out that both of HPAN and EDA-HPAN are nearly inactive for FA dehydrogenation (Figure S4). The dehydrogenation rate increases with the increase of reaction temperature. The activation energies of FA dehydrogenation on Pd/EDA-HPAN and PdCo0.2/EDA-HPAN are 39.30 and 37.25 kJ/mol, respectively, which are comparable with the most active supported Pd-based catalysts reported in studies in the literature [33–37]. At 333 K, the TOF value of PdCo0.2/EDA-HPAN catalyst is 4990 h−1, which is higher than that of Pd/EDA-HPAN catalyst (4330 h−1). It is worth noting that the aminated PdCo0.2/EDA-HPAN catalyst could also efficiently convert FA to H2 and CO2 at ambient temperature (303 K) with a TOF of 915 h−1, higher than Pd/EDA-HPAN and the previously reported PAN supported Pd-based catalysts [41– 43]. The catalytic activity of PdCo0.2/EDA-HPAN is comparable or even better than that of the recently reported MOF- or carbon-supported Pd-based catalysts (Table S1), suggesting that HPAN with hollow sphere structure could be used as a suitable support to prepare supported Pd-based NPs catalysts with excellent catalytic activity for FA dehydrogena-The catalytic properties of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts were measured for the dehydrogenation of FA. As shown in Figure 6a,b, both Pd/EDA-HPAN and PdCo0.2/EDA-HPAN exhibit good catalytic activity in the temperature range of 303–353 K, and the dehydrogenation rate of PdCo0.2/EDA-HPAN is higher than that of Pd/EDA-HPAN. No CO signal was detected by GC analysis during the reaction term (Figure S2). In addition, after CO<sup>2</sup> was absorbed by NaOH solution trap (10 mol/L), the total volume of gas product was reduced by half (Figure S3), indicating that the catalyst had excellent dehydrogenation selectivity. It should be pointed out that both of HPAN and EDA-HPAN are nearly inactive for FA dehydrogenation (Figure S4). The dehydrogenation rate increases with the increase of reaction temperature. The activation energies of FA dehydrogenation on Pd/EDA-HPAN and PdCo0.2/EDA-HPAN are 39.30 and 37.25 kJ/mol, respectively, which are comparable with the most active supported Pd-based catalysts reported in studies in the literature [33–37]. At 333 K, the TOF value of PdCo0.2/EDA-HPAN catalyst is 4990 h−<sup>1</sup> , which is higher than that of Pd/EDA-HPAN catalyst (4330 h−<sup>1</sup> ). It is worth noting that the aminated PdCo0.2/EDA-HPAN catalyst could also efficiently convert FA to H<sup>2</sup> and CO<sup>2</sup> at ambient temperature (303 K) with a TOF of 915 h−<sup>1</sup> , higher than Pd/EDA-HPAN and the previously reported PAN supported Pd-based catalysts [41–43]. The catalytic activity of PdCo0.2/EDA-HPAN is comparable or even better than that of the recently reported MOF- or carbon-supported Pd-based catalysts(Table S1), suggesting that HPAN with hollow sphere structure could be used as a suitablesupport to prepare supported Pd-based NPs catalysts with excellent catalytic activity for FA dehydrogenation.

tion. In addition, the stability and recyclability of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts were also studied for FA dehydrogenation at 333 K. As shown in Figure 7, both Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts still show very high activity after five cycles, suggesting the excellent stability of these catalysts. SEM images shown in Figure S5 demonstrate that the morphologies of the spent catalysts of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN are kept well in comparison with the fresh catalysts. The TEM results revealed that the average particle sizes of the Pd-based NPs in the spent catalysts of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN are still quite small (around 1.0 nm and 0.9 nm) after five catalytic cycles (Figure S6), suggesting that these two supported Pd-based In addition, the stability and recyclability of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts were also studied for FA dehydrogenation at 333 K. As shown in Figure 7, both Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts still show very high activity after five cycles, suggesting the excellent stability of these catalysts. SEM images shown in Figure S5 demonstrate that the morphologies of the spent catalysts of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN are kept well in comparison with the fresh catalysts. The TEM results revealed that the average particle sizes of the Pd-based NPs in the spent catalysts of Pd/EDA-HPAN and PdCo0.2/EDA-HPAN are still quite small (around 1.0 nm and 0.9 nm) after five catalytic cycles (Figure S6), suggesting that these two supported Pd-based catalysts have strong capability against aggregation.

catalysts have strong capability against aggregation.

*Catalysts* **2022**, *12*, x FOR PEER REVIEW 8 of 13

**Figure 6.** (**a**,**b**) Gas evolution volume (CO2 + H2) versus time for the dehydrogenation of FA over Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts and (inset) the corresponding Arrhenius plot; (**c**,**d**) TOF values at different temperatures. **Figure 6.** (**a**,**b**) Gas evolution volume (CO<sup>2</sup> + H<sup>2</sup> ) versus time for the dehydrogenation of FA over Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts and (inset) the corresponding Arrhenius plot; (**c**,**d**) TOF values at different temperatures. **Figure 6.** (**a**,**b**) Gas evolution volume (CO2 + H2) versus time for the dehydrogenation of FA over Pd/EDA-HPAN and PdCo0.2/EDA-HPAN catalysts and (inset) the corresponding Arrhenius plot;

(**c**,**d**) TOF values at different temperatures.

PdCo0.2/EDA-HPAN catalysts at 333 K. The content of Co in the supported Pd-based NPs should be also a key factor in in-**Figure 7.** Durability tests for the dehydrogenation of FA solution over (**a**) Pd/EDA-HPAN and (**b**) PdCo0.2/EDA-HPAN catalysts at 333 K. **Figure 7.** Durability tests for the dehydrogenation of FA solution over (**a**) Pd/EDA-HPAN and (**b**) PdCo0.2/EDA-HPAN catalysts at 333 K.

fluencing the catalytic performance of the catalysts. To check this point, a series of supported PdCox/EDA-HPAN catalysts with different Co/Pd molar ratios were also prepared by using the same procedure as described in the experimental section. As shown in Figure S7, the catalytic activities of the supported bimetallic PdCox catalysts are all higher than that of the single metal Pd based catalyst. Among them, the catalyst with Co/Pd ratio of 0.2 exhibits the highest TOF, suggesting that optimizing the Pd/Co ratio can adjust the catalytic performance to a certain extent, which might be due to the fact that the addition of different Co contents could generate different effects on the electron density of Pd as well as the dispersion state of Pd-based NPs, thus changing the activation ability for FA molecules. Further work is still required to clarify this point. The content of Co in the supported Pd-based NPs should be also a key factor in influencing the catalytic performance of the catalysts. To check this point, a series of supported PdCox/EDA-HPAN catalysts with different Co/Pd molar ratios were also prepared by using the same procedure as described in the experimental section. As shown in Figure S7, the catalytic activities of the supported bimetallic PdCox catalysts are all higher than that of the single metal Pd based catalyst. Among them, the catalyst with Co/Pd ratio of 0.2 exhibits the highest TOF, suggesting that optimizing the Pd/Co ratio can adjust the catalytic performance to a certain extent, which might be due to the fact that the addition of different Co contents could generate different effects on the electron density of Pd as well as the dispersion state of Pd-based NPs, thus changing the activation ability for FA molecules. Further work is still required to clarify this point. The content of Co in the supported Pd-based NPs should be also a key factor in influencing the catalytic performance of the catalysts. To check this point, a series of supported PdCox/EDA-HPAN catalysts with different Co/Pd molar ratios were also prepared by using the same procedure as described in the experimental section. As shown in Figure S7, the catalytic activities of the supported bimetallic PdCo<sup>x</sup> catalysts are all higher than that of the single metal Pd based catalyst. Among them, the catalyst with Co/Pd ratio of 0.2 exhibits the highest TOF, suggesting that optimizing the Pd/Co ratio can adjust the catalytic performance to a certain extent, which might be due to the fact that the addition of different Co contents could generate different effects on the electron density of Pd as well as the dispersion state of Pd-based NPs, thus changing the activation ability for FA molecules. Further work is still required to clarify this point.

On this basis of the above results and the related literature reports [15,37], a possible mechanism of FA dehydrogenation on PdCox/EDA-HPAN catalyst can be proposed. As

On this basis of the above results and the related literature reports [15,37], a possible mechanism of FA dehydrogenation on PdCox/EDA-HPAN catalyst can be proposed. As

*2.7. Catalytic Mechanism* 

*2.7. Catalytic Mechanism* 

#### *2.7. Catalytic Mechanism Catalysts* **2022**, *12*, x FOR PEER REVIEW 9 of 13

On this basis of the above results and the related literature reports [15,37], a possible mechanism of FA dehydrogenation on PdCox/EDA-HPAN catalyst can be proposed. As shown in Scheme 2, the surface functional groups such as cyanide and amino can act as basic sites to promote deprotonation of FA to generate [HCOO]− (Step 1). Then the adjacent PdCo species interacted with [HCOO]− to form an intermediate PdCo-[HCOO]− (Step 2). The C–H bond of the intermediate was then dissociated to produce CO<sup>2</sup> and [H]− (Step 3). Finally, [H]<sup>−</sup> combined with the initial H<sup>+</sup> to produce H<sup>2</sup> to complete the whole cycle. shown in Scheme 2, the surface functional groups such as cyanide and amino can act as basic sites to promote deprotonation of FA to generate [HCOO]<sup>−</sup> (Step 1). Then the adjacent PdCo species interacted with [HCOO]− to form an intermediate PdCo-[HCOO]<sup>−</sup> (Step 2). The C–H bond of the intermediate was then dissociated to produce CO2 and [H]− (Step 3). Finally, [H]− combined with the initial H+ to produce H2 to complete the whole cycle.

**Scheme 2.** Possible mechanism of dehydrogenation of FA. **Scheme 2.** Possible mechanism of dehydrogenation of FA.

#### **3. Materials and Methods 3. Materials and Methods**

#### *3.1. Chemicals and Materials*

*3.1. Chemicals and Materials*  Acrylonitrile (AN), styrene (St) and polyvinyl pyrrolidone (PVP) were supplied by Sinopharm Chemical Reagent Co., Ltd. (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China). Azodiisobutyronitrile (AIBN) was purchased from Tianjin Institute of Fine Chemicals retrocession (Tianjin Institute of Fine Chemicals retrocession, Tianjin, China). FA and EDA were procured from Aladdin Chemistry Co., Ltd. (Aladdin Chemistry Co., Ltd., Los Angeles, SC, USA). Ethanol and tetrahydrofuran (THF) were bought from Beijing Chemical Factory (Beijing Chemical Factory, Beijing, China). PdCl2 was supplied by Sinopharm Chemical Reagent Co., Ltd. (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China). NaBH4 and Co(NO3)2 were purchased from Shandong Xiya Chemical Industry Co., Ltd. Acrylonitrile (AN), styrene (St) and polyvinyl pyrrolidone (PVP) were supplied by Sinopharm Chemical Reagent Co., Ltd. (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China). Azodiisobutyronitrile (AIBN) was purchased from Tianjin Institute of Fine Chemicals retrocession (Tianjin Institute of Fine Chemicals retrocession, Tianjin, China). FA and EDA were procured from Aladdin Chemistry Co., Ltd. (Aladdin Chemistry Co., Ltd., Los Angeles, SC, USA). Ethanol and tetrahydrofuran (THF) were bought from Beijing Chemical Factory (Beijing Chemical Factory, Beijing, China). PdCl<sup>2</sup> was supplied by Sinopharm Chemical Reagent Co., Ltd. (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China). NaBH<sup>4</sup> and Co(NO3)<sup>2</sup> were purchased from Shandong Xiya Chemical Industry Co., Ltd. (Xiya Chemical Industry Co., Ltd., Chengdu, China).

#### *3.2. Preparation of Polyacrylonitrile Hollow Nanospheres (HPAN)*

(Xiya Chemical Industry Co., Ltd., Chengdu, China).

*3.2. Preparation of Polyacrylonitrile Hollow Nanospheres (HPAN)*  HPAN was produced by emulsifier-free emulsion polymerization and dissolution method [45]. To begin, 200 mg of PVP and 10 mL of St were added to a three necked flask containing 50 mL of deionized water and 50 mL of ethanol. After stirring for 5 min, 50 mg of AIBN was added. The resulting mixture was then heated at 358 K for 4 h. Subsequently, heating was stopped and the polystyrene (PS) emulsion was cooled to room temperature. Next, 55 mL of PS emulsion was mixed with 5 mL of AN for 2 h. Under the magnetic stirring, 20 mg of AIBN, 10 mL of deionized water and ethanol were added. Subsequently, the mixture was heated for 5 h at 360 K, then heating was stopped, and the emulsion was cooled to room temperature. The centrifuged powder product of PS@PAN was obtained by grinding. Finally, 700 mg of the powder and 10 mL of THF were mixed and dissolved by magnetic stirring for 12 h, then centrifuged and washed. The above operation was re-HPAN was produced by emulsifier-free emulsion polymerization and dissolution method [45]. To begin, 200 mg of PVP and 10 mL of St were added to a three necked flask containing 50 mL of deionized water and 50 mL of ethanol. After stirring for 5 min, 50 mg of AIBN was added. The resulting mixture was then heated at 358 K for 4 h. Subsequently, heating was stopped and the polystyrene (PS) emulsion was cooled to room temperature. Next, 55 mL of PS emulsion was mixed with 5 mL of AN for 2 h. Under the magnetic stirring, 20 mg of AIBN, 10 mL of deionized water and ethanol were added. Subsequently, the mixture was heated for 5 h at 360 K, then heating was stopped, and the emulsion was cooled to room temperature. The centrifuged powder product of PS@PAN was obtained by grinding. Finally, 700 mg of the powder and 10 mL of THF were mixed and dissolved by magnetic stirring for 12 h, then centrifuged and washed. The above operation was repeated twice and the product was lyophilized to obtain a white powder product of HPAN.

peated twice and the product was lyophilized to obtain a white powder product of HPAN.

First, 400 mg of HPAN powder was mixed with 32 mL of H2O and treated with ul-

*3.3. Synthesis of Amine-Functionalized HPAN* 

#### *3.3. Synthesis of Amine-Functionalized HPAN*

First, 400 mg of HPAN powder was mixed with 32 mL of H2O and treated with ultrasound for 3 min. Then, 4 mL of EDA was added by stirring and the mixture was heated to 368 K for 3 h. The sample was collected by centrifugation and washed three times with water. Finally, the product was dried overnight at 333 K to obtain EDA-HPAN.

## *3.4. Synthesis of Pd/EDA-HPAN and PdCox/EDA-HPAN*

To begin, 100 mg of EDA-HPAN powder was dispersed in 20 mL water by ultrasonic treatment, and then 1.5 mL H2PdCl<sup>4</sup> (0.02 mol/L) and 0.3 mL Co(NO3)<sup>2</sup> (0.02 mol/L) were added after ultrasonic treatment. After stirring for 8 h, the mixture was centrifuged, and the separated solid samples of Pd2+Co2+/EDA-HPAN were washed several times with ethanol and deionized water and dried in an oven at 333 K. Subsequently, the solid samples of Pd2+Co2+/EDA-HPAN were ultrasonically dispersed in 20 mL deionized water, and the newly prepared NaBH<sup>4</sup> solution (30 mg NaBH<sup>4</sup> dissolved in 2 mL H2O) was rapidly injected under strong stirring for 6 h. The reduced samples of PdCo0.2/EDA-HPAN were collected by centrifugation and washed with water three times. Finally, the product was dried overnight in an oven at 333 K. For comparison, PdCox/EDA-HPAN samples (where x represents the ratio of Co to Pd, which is 0, 0.4 and 0.6, respectively) were prepared according to the same procedure as above.

#### *3.5. Catalyst Characterisation*

FTIR: The samples were characterized by a Nicolet iS5 Spectrometer in the region of from 500 to 4000 cm−<sup>1</sup> .

Elemental analysis: The content of N was analyzed by using a Vario EL cube.

SEM: Patterns of the catalysts powder was collected by a SU8020 microscope (HI-TACHI Company, Tokyo, Japan) scanning electron microscopy (SEM) with an acceleration voltage of 10.0 kV and 100 kV.

TEM: Transmission electron microcopy (TEM) measurement was performed on FEI Tecnai F20 instrument operated at 200 kV.

ICP-AES: The contents of Pd and Co were analyzed by Thermo iCAP Qc inductively coupled plasma atomic emission spectrometry (ICP-AES).

TG: The pyrolysis experiments were performed on Bruker V70, the sample was heated with a higher heating rate of 10 ◦C/min from room temperature to 600 ◦C.

BET: The surface area of samples was calculated using the Brunauer-Emmett-Teller (BET) method, which performed on ASAP2020PlusHD88 at 77 K. All samples were degassed at 110 ◦C for 12 h prior to analyses.

XRD: The patterns of the catalysts powder were performed on a PANalytical B.V. Empyrean diffractometer with a CuKα radiation (λ = 1.5406 Å) at a voltage of 40 kV and a current of 40 mA.

XPS: X-ray photoelectron spectroscopy (XPS) measurements were performed on a Thermo ESCA LAB 250 spectrometer equipped with an MgKα source (1254.6 eV).

Gas Analysis: The gas composition analysis by a Shimadzu GC-8A gas chromatograph (GC) equipped with a TCD (detection limit: 10 ppm for CO).

#### *3.6. Catalytic Test*

An FA dehydrogenation experiment at different reaction temperatures using FA aqueous solution as raw material was conducted. In the test, 0.5 mL of FA (4 mol/L) aqueous solution was injected into a double-necked flask containing 9.5 mL of water and 50 mg of the catalyst powder without any additives under agitation. After each reaction, the catalyst was centrifuged and washed. Finally, it was dried in an oven at 333 K. The gas released from FA aqueous solution was collected and measured by the gasometric method, and real-time mapping.

In order to check the recoverability of the catalysts, Pd/EDA-HPAN and PdCox/EDA-HPAN samples were collected after the catalytic test by centrifugation. The separated solid

catalyst was washed several times and finally dried in an oven at 333 K before circulation experiment.

The catalytic activity of the catalysts is calculated as follows:

$$\text{TOF} = P\_0 V / \text{ (2RT} \eta\_{Pd} t) \tag{1}$$

In which *P*<sup>0</sup> represents the atmospheric pressure (101,325 Pa), *V* is the volume of generated H<sup>2</sup> when the conversion rate of FA reaches 20% (m<sup>3</sup> ), and *R* is the universal gas constant. *T* is the ambient temperature (298 K), *nPd* is the total mole number of the catalyst (mol), and *t* is the reaction time (h) when the conversion rate of FA reaches 20%.

#### **4. Conclusions**

In summary, PdCo NPs supported on surface aminated polyacrylonitrile hollow nanospheres (EDA-HPAN) were prepared and could be used as a highly active and stable catalyst for dehydrogenation of FA. PdCo NPs with ultra-small particle size (below 1 nm) could be uniformly dispersed on the surface of EDA-HPAN support due to the special hollow sphere structure, rough surface and the abundant basic groups of the polymer support. The existence of a relatively strong interaction between Pd and Co, as well as the metal-support interaction could modify the electronic structure of Pd to a certain extent, and lead to the formation of a stable structure and catalytically active Pd-based NPs for FA dehydrogenation. The abundant surface basic groups such as cyanide and amino groups could enhance the dispersity of the Pd-based NPs, and may also participate in the activation of FA molecules, thus resulting in the formation of a highly efficient FA dehydrogenation catalyst that can work well under ambient conditions. We believe that more efficient Pd-based catalysts will be developed for the dehydrogenation of FA under mild conditions by further tuning the morphologies and surface properties of the polymer support of polyacrylonitrile, which can certainly provide novel opportunities for promoting the development of green-energy engineering in the near future.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/catal12010033/s1, Figure S1: The pyrolysis TG curves of HPAN and EDA-HPAN, Figure S2: Gas chromatography analysis results of pure H<sup>2</sup> (**b**), pure CO (**d**) and generated gas for PdCo/EDA-HPAN catalyzed additive-free dehydrogenation of FA at 333 K (a) and (**c**), Figure S3: The volume of generated gas versus time for PdCo/EDA-HPAN catalyzed additive-free dehydrogenation of FA at 298 K in the absence (**a**) or presence (**b**) of NaOH trap, Figure S4: The volume of generated gas versus time for HPAN and EDA-HPAN catalyzed additive-free dehydrogenation of FA at 333 K. Figure S5: SEM images of the spent catalysts of (**a**) Pd/EDA-HPAN and (**b**) PdCo/EDA-HPAN after five times reaction, Figure S6: The TEM images and the corresponding Pd particle size distributions of the spent catalysts of (**a**) Pd/EDA-HPAN and (**b**) PdCo/EDA-HPAN after five times reaction, Figure S7: The TOF values of PdCox/EDA-HPAN catalysts with different Co/Pd molar ratios. Table S1: Comparison of catalytic activities with literature reported heterogeneous catalysts.

**Author Contributions:** Conceptualization, X.H. and M.J.; Methodology and writing, Y.L. and P.S.; Investigation, H.C.; Data curation, D.L. and R.D. The manuscript was amended and supplemented by all authors. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors greatly appreciate the financial support of the Jilin Institute of Chemical Technology School-level Major Project (No. 2018006) and General Project (No. 2017022), the National Natural Science Foundation of China (No. 22172058), and the Shanghai Engineering Research Center of Green Energy Chemical Engineering (No. 18DZ2254200).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

