纳米通道控制合成超高氮掺杂效率的介孔 Fe/N/C 催化剂用于氧还原反应

X射线光电子能谱(XPS)分析用于研究m表面的电子结构和化学成分 -Fe/N/C-800, m -Fe/N/C-900 和 m -Fe/N/C-950。根据调查的 XPS 光谱（图 4 和附加文件 1：图 S2），我们发现所有 Fe/N/C 型催化剂都含有 C、N、O 和 Fe 元素，证明 Fe 和 N原子被成功地掺杂到碳结构中。 XPS 表面分析中 Fe、C 和 N 的总含量以及所制备的介孔 Fe/N/C 型 ORR 电催化剂中的所有 N/C 比率总结在表 1 中。 米 -Fe/N/C-900高达10.3，明显大于m -Fe/N/C-800 (~ 10.0) 或 m -Fe/N/C-950 (~ 4.4)，进一步表明 m 中最高的氮掺杂效率和总氮含量 -Fe/N/C-900。 C1s m 的 XPS 分析 -Fe/N/C-900（图 4b）由于与 sp ² 对应的特征峰的出现，也证实了氮原子的掺杂 C=N 键的结合能 (B.E.) 约为 286.5 eV。此外，所制备的 Fe/N/C 型电催化剂的 XPS 光谱中 Fe 元素的信号相对较弱，因为在酸性溶液中浸出后 Fe 含量非常低。为了检查 Fe 原子的电子态，我们分析了 Fe 2p m 的 XPS 谱 图 2b 中的 -Fe/N/C-900，用 B.E. 拟合成四个峰。分别为 710.3、712.9、716.8 和 725.3 eV。位于~710.3 eV的峰可以归因于Fe-N键的特征峰，表明Fe ²⁺ 的出现 m 中的状态 -Fe/N/C-900 [29, 30]。该会。位于约 712.9 eV 和 725.3 eV 对应于 Fe ³⁺ 2p 3/2 和 Fe ³⁺ 2p 分别为 1/2（图 4c），这表明 Fe-TPTZ 复合物中的 Fe 原子可以被部分氧化；然而，位于 716.8 eV 的峰可以代表 Fe-N 键的卫星峰。为了进一步确定 Fe 氧化物的化学状态，我们分析了 O1s m 的频谱 -Fe/N/C-900（图 4d）。它用 B.E. 分为三个峰。 531.0、532.3 和 533.2 eV，分别对应于 C–O–C、C=O 和 Fe–O [31]。上述结果可能表明，Fe原子的存在状态主要由m中的Fe-N和Fe-O键组成。 -Fe/N/C-900电催化剂。

调查 (a ), Fe 2p (b ), O 1s (c ), 和 C 1s (d ) m 的 XPS 光谱 -Fe/N/C-900

拟合的 N1s m 的 XPS 光谱 -Fe/N/C-800 和 m -Fe/N/C-900（参见图 5a、b）显示了四个 B.E. 398.3、399.5、401.1 和 406.1 eV，这分别归因于吡啶-N、Fe-N、石墨-N 和氧化-N (-NO2) [10, 16, 25, 26, 32, 33] .然而，拟合的 N1s m 的 XPS 谱 -Fe/N/C-950（图 5c）只显示了三个 B.E. 399.2、401.4 和 406.4 eV，分别对应于 Fe-N、石墨-N 和氧化-N (-NO2)。注意到由于氧原子的电负性更强，氮原子的电子云密度发生了移动，导致氧化峰出现在高能阶段。石墨-N基团的相对比例在总氮含量中占主导地位，但随着热解温度的增加，惰性氧化-N（-NO2）的相对比例从17.8增加到38.7 at.%，如图5d所示.此外，较高的热解温度会导致m中吡啶-N基团的消失。 -Fe/N/C-950。上述结果表明，Fe-N 基团可以在所制备的 Fe/N/C-型催化剂中形成，但活性吡啶-和石墨-N 基团对确定 Fe/N/C- 的 ORR 活性具有重要意义。型电催化剂。

N 1s m 的 XPS 光谱 -Fe/N/C-800 (a ), m -Fe/N/C-900 (b ), m -Fe/N/C-950 (c ), 以及它们在总 N 含量中的相对含量 (d )

m的形态分析 -Fe/N/C-900 催化剂如图 6 和附加文件 1：图 S3 所示。 m 可以存在大量海绵状的形状 -Fe/N/C-900（图 6a、b）通过金属有机物 (Fe-TPTZ) 复合物的高温热解合成，受限地固定在分子筛的纳米通道中。高分辨率 TEM 图像（图 6c、d）显着显示 m 内部有许多高度有序的介孔结构 -Fe/N/C-900，这在很大程度上可以源自作为纳米通道限制反应器的分子筛的去除。此外，在图 6e 中可以清楚地看到边缘上的无序碳结构和几个介孔，这是由于氮原子的掺杂。因此，m 的多孔特性和布鲁诺-埃米特-特勒 (BET) 表面积 -Fe/N/C-900 还通过氮吸附/解吸等温线进行了研究（图 6f 的插图）。可以观察到 Langmuir IV 型等温线，表明 m 具有高度介孔特性 -Fe/N/C-900 电催化剂。 m的BJH孔径分布可以进一步证实 -Fe/N/C-900（图 6f），显示出高 BET 表面积（A BET ~ 1035 m ² g ^-1 ) 和总孔体积 (V 总计 ~ 1.22 cm ³ g ^-1 ) 平均孔径 (D p) 约为 4.7 nm。所展示的两个最大孔径（4.3 和 9.3 nm）归因于 m 中介孔的最大位置 -Fe/N/C-900。这些介孔可以为电解质、反应物和产物的快速运输提供更方便的纳米通道；降低氧分子向掺氮活性位点的传输阻力；并提高m的ORR催化性能 -Fe/N/C-900。此外，使用分子筛的纳米通道作为纳米限域反应器有利于生产具有超高表面积的海绵状三维介孔碳，有利于增强催化 ORR 活性位点的暴露。元素映射图像（图 7）显示了四种主要元素（Fe、N、C 和 O）在 m 表面的均匀分布 -Fe/N/C-900，这在一定程度上可能是分子筛的纳米通道限域效应和Fe-TPTZ配合物的稳定分子配位效应的协同作用所致。

SEM (a ) 和高分辨率 TEM 图像 (b –e ) m -Fe/N/C-900； f m的孔径分布 -Fe/N/C-900；插图为氮气吸附-解吸等温线

TEM 图像 (a ) m -Fe/N/C-900 和相应的 C、O、Fe 和 N 元素映射图像 (b –f )

通过循环伏安图 (CV) 或线性扫描伏安图 (LSV) 测试所有制备的催化剂对 ORR 的电催化活性。图 8a 显示了 m 的 CV -Fe/N/C-900 in N2 vs O2-saturated 0.1 mol l ^-1 KOH 电解液。在 O2 饱和电解质中，ORR 峰值出现在 0.86 V vs. RHE 处；然而，在 N2 饱和电解质中观察到无特征的 CV 曲线，表明 m 的 ORR 催化活性 -Fe/N/C-900 具有起始电位 (E 起始）1.0 V。此外，热处理温度（800-950 °C）对ORR性能的影响在图8b中进行了研究。根据最大峰值电流密度(j p)，以及最正的峰值电位 (E p) 可以在 m 处获得 -Fe/N/C-900。 m 的 ORR 活性 -Fe/N/C-900 可以与其他报道的掺杂碳催化剂相媲美（参见附加文件 1：表 S2）。为了深入了解 Fe/N/C 型电催化剂的 ORR 动力学行为，我们通过 LSV 方法结合 RRDE 进一步测试了 ORR 极化曲线，如图 8c 所示。根据 RRDE 数据，转移的电子数 (n ) 和 ORR 期间的 H2O2 产率 (H2O2%) 是通过使用以下方程估计的。分别为 (1) 和 (2)。计算公式如下[34]：

一 m的CV曲线 -Fe/N/C-900 在 N2 与 O2 饱和的 0.1 M KOH 溶液中。 b m的CV曲线 -Fe/N/C-800, m -Fe/N/C-900 和 m -Fe/N/C-950 in O2-saturated 0.1 mol l ^-1 氢氧化钾溶液。 c 在 m 的 RRDE 上使用 LSV 获得的盘和环电流 -Fe/N/C-800, m -Fe/N/C-900 和 m -Fe/N/C-950 in O2-saturated 0.1 mol l ^-1 氢氧化钾溶液。 d m对应的电子转移数和H2O2产率 -Fe/N/C-800, m -Fe/N/C-900, and m -Fe/N/C-950 derived from c

其中我 d is the Faradaic disk-current, I r is the Faradaic ring-current, and N is the collection efficiency of ring electrode (0.38). The Pt-ring potential was set at 1.5 V (vs. RHE) as reported elsewhere. Figure 8d displays the relatively calculated results. The H2O2 yield (<6.5%) and n value (3.88–3.94) are obtained on m -Fe/N/C-900, dominating a four-electron ORR pathway. It suggests that this catalyst is a valuable substitute for the traditional 20 wt.% Pt/C catalyst (purchased from Aladdin Industrial Co. Ltd.), although the H2O2 yield on m -Fe/N/C-900 is slightly higher. In addition, the half-wave potential (E 1/2) for ORR of m -Fe/N/C-900 is about 0.841 V approaching that of 20 wt.% Pt/C (~ 0.848 V), and the limited current density (j d) of m -Fe/N/C-900 is almost identical to that of 20 wt.% Pt/C. Compared to the m -Fe/N/C-900 catalyst, higher H2O2 yield and smaller n value can be gained on both m -Fe/N/C-800 and m -Fe/N/C-950, but the electron transfer number on both m -Fe/N/C-800 and m -Fe/N/C-950 still belongs to the range (3.4–4.0), showing that the ORR on two Fe/N/C-type happens with a mixed process of two and four-electron transfer pathways. The above results further approve that the m -Fe/N/C-900 synthesized by nanochannel-confined control of the pyrolysis process to improve the nitrogen-doping efficiency and increase the nitrogen-doped active-site density has exhibited the optimal ORR catalytic performance in alkaline electrolyte.

We also discuss the effect of different transition metals on the ORR catalytic activity of Fe/N/C-type electrocatalysts. The obtained LSV curves on the RRDE are indicated in Fig. 9a, and the corresponding n value and H2O2 yield are demonstrated in Fig. 9b. E 1/2 values are about 0.785 V for m -Cu/N/C-900 and 0.780 V for m -Ni/N/C-900, respectively, which are lower compared to the m -Fe/N/C-900. The j d follows the order of m -Fe/N/C-900> m -Cu/N/C-900> m -Ni/N/C-900, further suggesting the best ORR catalytic activity of m -Fe/N/C-900 in 0.1 mol l ^-1 KOH solution. Compared to the m -Fe/N/C-900 catalyst, higher H2O2 yield and smaller n value are obtained on both m -Cu/N/C-900 and m -Ni/N/C-900. What is noteworthy is that the H2O2 yield on m -Cu/N/C-900 and m -Ni/N/C-900 is over twice as large as that on m -Fe/N/C-900. However, the n value on both m -Cu/N/C-900 and m -Ni/N/C-900 is 3.5–4.0, indicating that the ORR process on two Fe/N/C-type electrocatalysts follows a two- and four-electron mixed transfer pathway but is dominant in a four-electron reaction pathway. Besides, the electrochemical long-term stability for ORR catalysis of m -Fe/N/C-900 is of great significance in the practical applications. An accelerated aging test (AAT) was carried out by successive CV scanning tests from 0.2 to 1.2 V vs RHE for 5000 cycles at 200 mV s ^-1 in oxygen-saturated 0.1 M KOH electrolyte. The ORR electrocatalytic behavior of m -Fe/N/C-900 has been further evaluated under the same conditions as above experiments. CV curves for ORR activity of m -Fe/N/C-900 before and after doing the AAT are almost unchanged on the E p (~ 0.86 V), but the j p is slightly reduced (Fig. 9c). LSV curves of m -Fe/N/C-900 (Fig. 9d) also reveal an only ~ 12 mV negative shift in the E 1/2 and a negligible decrease in the j d. However, the commercial Pt/C (20 wt.% Pt) catalyst after doing the AAT has exhibited about 55 mV of the negative shift in E 1/2 and an obvious reduction in the j d (Fig. 9d). Results show that the m -Fe/N/C-900 has more excellent electrocatalytic stability compared to the Pt/C catalyst, further suggesting that it is a valuable and promising substitute for the conventional Pt-based materials in alkaline electrolytes.

一 Disk and ring currents obtained with LSVs on RRDE for m -Ni/N/C-900, m -Cu/N/C-900, and m -Fe/N/C-900 in O2-saturated 0.1 mol l ^-1 KOH solution. b The corresponding electron transfer number and H2O2 yield of m -Ni/N/C-900, m -Cu/N/C-900, and m -Fe/N/C-900 derived from a . c CV curves and d LSV curves of m -Fe/N/C-900 before and after continuous scanning for 5000 cycles in O2-saturated 0.1 M KOH solution

In order to discuss the catalytically active sites of Fe/N/C-type catalysts and study the role of the molecular coordination and nanochannel-confined effects, we have further examined the N/C-900, Fe/N/C-900, and 20 wt.% Pt/C catalysts for comparison of the ORR catalytic behavior. The tested results of the ORR activity are indicated in Fig. 10a. The onset potentials of ORR are about 0.683 V for N/C-900 and 0.740 V for Fe/N/C-900, being largely lower than those of m -Fe/N/C-900 (0.841 V) and 20 wt.% Pt/C catalysts (0.848 V). Figure 10b shows the corresponding H2O2 yields and transferred electron numbers in the ORR process. Given other Fe/N/C-type catalysts and the Pt/C catalyst, the H2O2 yield on N/C-900 is the highest and the transferred electron number on N/C-900 is the smallest, suggesting the worst ORR catalytic activity of N/C-900. In addition, the H2O2 yield on Fe/N/C-900 is mainly higher than that on m -Fe/N/C-900 and the transferred electron number on Fe/N/C-900 is far lower than that on m -Fe/N/C-900 in the same range (0.2–0.8 V vs RHE), indicating a relatively inferior ORR activity. Thus, it can be concluded that the ORR performance complies with the sequence of Pt/C> m -Fe/N/C-900> Fe/N/C-900> N/C-900. These results show that the formation of Fe-TPTZ compounds with the strong molecular-level coordination effect is beneficial to produce the Fe/N/C-type catalysts with high ORR activity, and the utilization of the nanochannel-confined effect can reduce the decomposition speed of Fe-TPTZ compounds and protect the nitrogen-rich active sites (e.g., Fe–N, graphitic-N, or pyridinic-N) from the thermal loss during the pyrolysis process (see the TG analysis, Fig. 2a), which can enhance the ORR performance of Fe/N/C-type catalysts in alkaline medium.

一 Disk and ring currents obtained with LSVs on RRDE for N/C-900, Fe/N/C-900, m -Fe/N/C-900, and Pt/C (20 wt.%) in O2-saturated 0.1 mol l ^-1 KOH solution. b The corresponding electron transfer number and H2O2 yield of N/C-900, Fe/N/C-900 m -Fe/N/C-900, and Pt/C (20 wt.%) derived from a

To deepen the understanding of active sites and their ORR catalysis mechanism, we also characterized the prepared Fe/N/C-type catalysts by the spectra of X-ray diffraction (XRD) and X-ray photoelectric spectroscopy (see Additional file 1:Figures S4 and S5). The XRD data display that the density of carbon (002) peak follows the sequence of N/C-900> Fe/N/C-900> m -Fe/N/C-900, and the positions for carbon (002) peak in three catalysts are negatively shifted because of the production of more sp ² C–N groups into the graphitic layers and the decrease of graphitization. It also implies that the different N content may be doped into the carbon skeleton in the catalyst, and both N-doping efficiency and N content can follow the similar order with their ORR activity. More significantly, we further compare the structural and porous differences between Fe/N/C-900 and m -Fe/N/C-900 to better study the nanochannel-confined effect of molecular sieves. The tested XPS survey spectrum of Fe/N/C-900 was indicated in Additional file 1:Figure S5, and its surface contents of Fe, C, and N and the N/C ratio were summarized in Table 1. The Fe content and N content are ~ 0.41 at.% and ~ 4.48 at.%, but the N/C ratio is only 5.1 in the Fe/N/C-900, being lower compared to the m -Fe/N/C-900. It suggests the N-doping efficiency was improved by introducing the nanochannel-confined protection strategy into the carbonization process, which can prove our key views of this work. Besides, the fitted N1s XPS spectrum of Fe/N/C-900 is indicated in Fig. 10c. It displays the existence of four peaks with B.E. of 398.6, 399.6, 401.2, and 406.4 eV, which still correspond to the pyridinic-N, Fe–N, graphitic-N, and oxidized-N (–NO2) with a relative percentage of 14.0, 10.2, 35.4, and 40.4 at.%, respectively. Compared with the m -Fe/N/C-900, the total ratio of active N-rich groups such as pyridinic-N, Fe–N, and graphitic-N obviously decreases about 19.8 at.%; however, the relative ratio of Fe–N groups is reduced about 15.2 at.%. Thus, associating with fore-mentioned XPS data and the catalytic activity data, we conclude that the electrocatalytically active sites may be pyridinic- and graphitic-N groups for holding the ORR performance, but the enhancement of the ORR activity may be related to the relative ratio of Fe–N groups for our system. The role of the nanochannel-confined effect cannot only reduce the loss of total N content, but also can largely increase the N-doping efficiency and improve the effective ORR active-site density in the catalyst. Besides, we further analyzed porous characteristic and BET specific surface area of Fe/N/C-900 without the usage of the KIT-MS nanoreactor. The nitrogen adsorption/desorption isotherms with a similar Langmuir IV-type isotherm curve are seen in the inset of Fig. 10d. It suggests that highly mesoporous characteristic is still existed in Fe/N/C-900, supported by the analysis of BJH pore-size distribution (Fig. 10d). A BET (~ 875 m ²  g ^-1 ) and V total (~ 0.76 cm ³  g ^-1 ) with an average D P of only ~ 3.5 nm are obtained on the Fe/N/C-900, which are obviously lower than those on the m -Fe/N/C-900. The large difference on pore structures between Fe/N/C-900 and m -Fe/N/C-900 can be derived from the stable coordination effect and the nanochannel-confined role of a KIT-MS reactor in the preparation of m -Fe/N/C-900. It will also influence their inherent ORR performance because higher A BET and V total can help to supply abundant catalytic sites and increase the exposed surface active-site density, being beneficial to the adsorption and electro-reduction process of O2 molecule [35]. Notably, we should pay much attention to the effect of the conductivity characteristic. Generally, a higher conductivity characteristic of m -Fe/N/C-900 corresponds to a relatively faster ORR electron transportation process. Therefore, facile design and control of active nitrogen-rich groups (pyridinic-N, graphitic-N, and Fe–N, etc.) is of great importance to fabricate mesopore-structured Fe/N/C electrocatalysts for the ORR, but further improving the conductance, N-doped active-site density, and mesoporous characteristic is another key issue of concern to obtain the high performance.