Elsevier


膜科学杂志


第712卷，2024年12月，123239
Journal of Membrane Science


用于H 2纯化的大面积α-氧化铝管上含有苯并咪唑键的分子级杂化膜

https://doi.org/10.1016/j.memsci.2024.123239 获取权利和内容

 强调

  • •

    在大面积粗糙 α-氧化铝管上制备分子级杂化膜。
  • •

    使用简便的界面聚合方案。
  • •

    该膜表现出良好的H 2 /CO 2分离性能。
  • •

    该膜具有耐高温、耐压的性能。
  • •

    设计了两级膜工艺。

 抽象的


H 2 /CO 2分离在工业中很重要。聚合物、无机和杂化膜已有报道，但仍缺乏具有良好可扩展性和大面积的高性能膜。在这里，我们在粗糙的 α-氧化铝管上制造了一系列含有苯并咪唑键的分子级杂化膜 (MHM)。每个面积为37.7 cm 2的膜在423 K下表现出14.2的H 2 /CO 2选择性（相应的H 2渗透率为29.5 GPU）和310 GPU的高H 2渗透性（相应的H 523 K 时( 2 /CO 2选择性为9.37)。膜具有良好的耐压和耐温性。此外，还设计了两级膜系统来研究压力和膜面积对产氢的影响。

 关键词


分子级杂化膜
管状膜
H 2纯化
苯并咪唑键
金属氧簇

 一、简介


能源是现代社会的支柱。清洁、廉价、丰富的能源对于经济社会可持续发展具有重要意义。氢是一种良好的能源载体，人们相信氢将在能源转型中发挥重要作用[ 1 , 2 ]。尽管最近已经尝试基于可再生能源生产绿色氢，但目前氢生产的主要来源是化石燃料[ 3 ]。化石燃料经过气化/重整和水煤气变换后可以转化为以H 2和CO 2为主的变换气。该过程每年排放二氧化碳500吨[ 4 ]。发展绿色可持续的氢气提纯技术对于氢能的普及具有重要意义。

工业上使用的氢气分离技术包括变压吸附、吸收和低温分离。气体分离膜技术具有无相变、占地面积小、能耗低等优点，已在多个领域得到应用[ 5 ]。膜技术的使用有望降低氢气纯化的能耗和成本[ 6 , 7 ]。

用于氢气纯化的膜有无机膜、有机膜和有机-无机杂化膜。钯膜、二氧化硅膜、沸石膜等无机膜表现出高性能。然而，在生产和使用过程中存在一些限制[ 8 , 9 ]。有机膜具有生产成本低、加工性能高等优点[ 10 ]。人们已经设计和合成了多种材料来增强膜的性能，包括聚苯并咪唑（PBI）[ 11 ]、聚酰亚胺（PI）[ 12 ]、多孔芳香骨架（PAF）[ 13 ]、多孔有机骨架（POF）如共价键有机框架（COF）[ 14 ]、苯并咪唑连接聚合物（BILP）[ 15 ]以及苯并咪唑和亚胺连接聚合物（BIILP）[ 16 ]。基于尺寸依赖性扩散选择性，BILP和BIIIP被证明具有良好的H 2 /CO 2分离能力。此外，BILP 和 BIILP 膜可以通过界面聚合（IP）轻松制造[ 16 , 17 ]。 有机-无机杂化膜，如普通混合基质膜（MMM），结合了无机填料和有机基质的优点，有望提高氢气净化膜的综合性能[ 18 ]。此外，刘等人。用无机单体在分子水平上调节聚合物膜的结构，并提出了分子级杂化膜（MHM）的概念[ 8 ]。分子级杂化膜避免了MMM的界面相容性问题，有效提高了H 2 /CO 2分离性能。 [Zr 6 (O 4 ) (OH) 4 (H 2 O) 8 (Gly) 8 ]•12Cl•8H 2 O (CP-2) [ 19 ]是一种Zr 6簇，其丰富的氨基官能团为能与醛基反应形成网络结构。据报道，引入 CP-2 单体形成 MHM 可以提高聚合物膜的性能[ 8 ]。此外，使用大分子量单体有利于在粗糙基材上合成无缺陷层，因为它们可以使粗糙基材变得光滑[ 20 , 21 ]。
Although many membranes show excellent H2/CO2 separation performance according to reports, related research (e.g., BILP membranes and MHMs) is in the primary stage, and the membrane area is limited especially. It is worth noting that synthesis of membranes with large area is essential for the application.
Besides the selective layers, membrane substrates are also important. Organic materials have the advantages of low cost and have been widely used in various membrane processes [22,23]. Inorganic materials such as ceramics are recommended as substrates for hydrogen purification membranes due to their good thermal stability, chemical stability and high mechanical strength [24,25]. According to reports, inorganic substrates have been used not only in hydrogen purification [24,26], but also in other harsh conditions such as high temperature nanofiltration [24,27] and high pressure organic separation [22].
Commercial ceramic tubes such as α-alumina tubes are easy to scale up and have been widely used in water treatment and pervaporation [28,29]. Besides, α-alumina is more stable than γ-alumina, the latter was used as substrates for MHMs and BILP membranes [8,15]. However, direct synthesis of continuous selective layers (e.g., BILP membranes and MHMs) on commercial α-alumina tubes is a challenging work, because normally their surfaces are rough and pore size are large [27,30,31].
In this work, we directly fabricated defect-free molecular-scale hybrid membranes (MHMs) on α-Al2O3 tubes by interfacial polymerization (IP). Each of the membrane area reached 37.7 cm2, compared with 1–2 cm2 of the reported MHMs or BILP membranes. The introduction of CP-2 monomers during IP significantly increased the separation performance of the bare polymer layers. Besides, the thermal stability of membranes was improved by post cross-linking with α,α′-dibromo-p-xylene (DBX). Finally, we designed a two-stage membrane process based on the synthesized membranes. The parameters of membrane process were studied.

2. Experimental section

2.1. Materials

1,2,4,5-benzenetetramine tetrahydrochloride (BTA, Heowns, 98 %), terephthal aldehyde (TPA, Heowns, 98 %), α,α′-dibromo-p-xylene (DBX, Heowns, 98 %), trimesoyl chloride (TMC, TCI, >98 %), zirconyl chloride octahydrate (ZrOCl2·8H2O, Sigma-Aldrich, 98 %), glycine (Gly, Meryer, 99.5 %), hydrochloric acid (HCl, Dongguan Liyuan Chemical Co. Ltd, AR), toluene (Tianjin Jiangtian Chemical Industry Co., Ltd, AR), n-hexane (Tianjin Yuanli Chemical Industry Co., Ltd, AR) were used without further purification. Deionized water was made in our laboratory. Porous α-alumina tubes (100 mm in length, 12 mm in outer diameter and 2 mm in thickness, membrane area (outer surface) of 37.7 cm2, average pore size of the outer surface of 50 nm) were supplied by Guangdong Lishun technology Co., Ltd. and used as substrates. The H2 permeance and H2/CO2 selectivity of the bare α-alumina tubes are 5.19 × 105 GPU and 1.22, respectively, at 1.1 bar and room temperature. CP-2 was synthesized referring to our previous work [8].

2.2. Fabrication of tubular membranes

The membranes were fabricated on the outer surfaces of α-Al2O3 tubes by IP (see Fig. 1 and Table 1). Monomers in aqueous phase were BTA and CP-2 (or only with BTA). The monomer in toluene phase was TPA. Before IP, the substrates were heated under 700 °C for 6 h to remove stains. A photo of a α-Al2O3 tube is shown in Fig. 2 (A). In a standard IP procedure, the tubes were immersed in the aqueous phase for 20 min under reduced pressure (0.80 bar) in a vacuum oven at room temperature. All the pressures mentioned in this work are absolute pressures. Then, the tubes were taken out from the aqueous phase, sealed on both ends with polytetrafluoroethylene tapes, and dried under ambient conditions. Afterwards, the tubes were put into toluene phase carefully and immediately when the outer surfaces were dry so that the IP took place on the outer surfaces rather than the inner sides. After 30 min, the tubes were consequently taken out from the toluene phase, and dried under ambient conditions for two days to remove the solvents. A photo of an as-synthesized membrane (a M5) is shown in Fig. 2B.
Fig. 1
  1. Download: Download high-res image (509KB)
  2. Download: Download full-size image

Fig. 1. (A) Fabrication procedures of tubular membranes; selective layer structures of M5 (B), M7 (C) and M8 (D).

Table 1. Preparation parameters of the membranes studied in this work.

Membrane codeAqueous phaseToluene phaseIP durationPost cross-linking reagent
M11.2 wt% BTA0.5 wt% TPA30 min
M21.2 wt% BTA+0.5 wt% CP-20.5 wt% TPA30 min
M31.2 wt% BTA+0.3 wt% CP-20.5 wt% TPA30 min
M41.2 wt% BTA+0.2 wt% CP-20.5 wt% TPA30 min
M51.5 wt% BTA+0.3 wt% CP-20.5 wt% TPA30 min
M6
2.0重量%BTA+0.3重量%CP-2		 0.5 重量% TPA 30分钟
M7
1.5% BTA+0.3% CP-2		 0.5 重量% TPA 30分钟TMC
M8
1.5% BTA+0.3% CP-2		 0.5 重量% TPA 30分钟DBX
Fig. 2

  1. 下载：下载高分辨率图像 (268KB)

  2. 下载：下载全尺寸图像


图2 . (A) α-Al 2 O 3管（长度和外径分别为 10 cm、1.2 cm）、(B) 膜 (M5) 和 (C) 用于性能评估的膜电池的数码照片。


2.3.管状膜的后交联


IP 后管状膜进一步交联。在这项工作中我们采用TMC和DBX作为交联剂。将TMC溶解在正己烷中，浓度为0.2wt%。将DBX溶解在甲苯中，浓度为0.5wt%。在后交联之前将管干燥至少两天以除去溶剂。将膜在室温下交联24小时并在环境条件下干燥两天。


2.4.表征方法


通过扫描电子显微镜（SEM，Regulus 8100）观察膜的形貌。通过SEM（日立，S-4800）的能量色散光谱仪（EDS）分析膜内元素的分布。通过原子力显微镜（AFM，Bruker Dimension 图标）测量观察表面形貌和粗糙度。采用非接触方式，扫描尺寸为5μm×5μm。 X 射线光电子能谱 (XPS) 在 K-Alpha + 能谱仪 (Thermo) 上进行。结合能用受污染的碳 C1s (284.8 eV) 进行校准。结果是通过测试膜的表面获得的。使用iS50 FT-IR光谱仪（Thermo）在400-4000 cm -1范围内以1 cm -1的分辨率测量衰减全反射傅里叶变换红外光谱（ATR-FTIR）光谱。


2.5.膜性能评估


使用Wicke-Kallenbach方法用自制膜池（如图2C所示）测试气体渗透性。膜的内表面和外表面被O形环分开（图S1 ）。密封后有效膜面积为26.4cm 2 。通过质量流量控制器调节进料气体的流速和组成。原料气为等摩尔比的H 2和CO 2 ，​​体积流量各300mL·min -1 。进料温度和压力分别由对流烘箱和背压阀控制。使用氩气稀释并携带渗透气体，流量为300 mL min -1 。透过气相色谱仪(Shimadzu, GC-2014)热导检测器分析渗透液的组成和摩尔浓度，柱箱温度120℃，检测器电流55mA，载气流量25mL分钟-1 )。每个数据在稳定10小时左右后记录。气体渗透率（ Pi ) 根据以下等式(1)计算： (1)Pi=NiAΔpi N为组分i的透过量(mol·s -1 )， A为有效膜面积(m 2 )， Δp为组分i的跨膜压差(Pa)。本文中气体渗透率采用气体渗透单元（GPU）作为单位，1 GPU = 3.35 × 10 -10 mol m -2 s -1 ·Pa -1 。气体选择性（ αi/j ）通过以下等式（2）定义为它们的磁导率之比： (2)αi/j=PiPj


2.6。膜工艺设计


先前的工作表明单级膜系统很难实现高H 2纯度和高H 2产率[ 32 ]。在这项工作中，设计了模拟两级膜系统来讨论管式膜的技术可行性。过程中使用的性能选自M5-1。操作温度为150℃。两级的进料压力相同。氢气纯度和氢气产率根据以下方程(3) 、 (4) 、 (5)计算： (3)Ni=Pi×A×(pi,retentate−pi,permeate) (4)hydrogenpurity=NH2,permeateNH2,permeate+NCO2,permeate (5)hydrogenyield=NH2,permeateNH2,feed

N为组分 i 透过膜的透过率（mol∙s −1 ）， P为组分 i 的透过量（mol∙m −2 ∙s −1 ∙Pa −1 ）， A为有效膜面积（m 2 )、 pi ,渗余物是组分i在渗余物侧的分压(bar)， pi ,permeate是组分i在渗透物侧的分压(bar)。


3。结果与讨论


3.1.膜的表征


裸基板和膜的SEM图像如图3所示。 α-Al 2 O 3管的外表面不光滑，表面孔径达数百纳米（图3A ）。如图3 B和C所示，膜层（M5）是连续的，厚度为40-50 nm，并随着基底的表面形貌而波动。后交联后，膜表面稍微光滑，厚度略有增加（图3D和E）。值得注意的是，孔隙渗透不明显，有利于H 2渗透。从 EDS 图可以看出，Zr、C 和 N 元素均匀分布在膜层中（图 3 （G-I））。 C、N和Zr的归一化质量百分比分别为59.7%、39.0%和1.3%。经过523 K的测试后，M5的表面部分受损（图3 J），而M8则完好保留（图3 K），表明用DBX进行后交联可以增强热稳定性。
Fig. 3

  1. 下载：下载高分辨率图像 (2MB)

  2. 下载：下载全尺寸图像


图3 . (A) α-Al 2 O 3管的表面。合成膜的 SEM 图像：(B) M5 的表面和 (C) 横截面； (D) M8 的表面和 (E) 横截面； (F) 用于 EDS 映射分析的 M5 表面。图像 (F) 上扫描的 EDS 映射：(G) Zr 信号； (H) C 信号； (I) N 个信号。 (G)、(H)和(I)的比例尺与(F)相同。 523 K 测试后膜的表面 SEM 图：(J) M5； （K）M8。


图4显示了裸露的α-Al 2 O 3基板和膜(M5)的3D图像和高度测量结果。 α-Al 2 O 3基材和膜的均方根粗糙度(Rq)分别为21.2nm和11.2nm。另外，α-Al 2 O 3基材和膜的算术平均粗糙度(Ra)分别为16.8nm和8.68nm。这表明基材具有粗糙的表面，这也通过SEM观察到（图3A ）。这将增加其表面形成无缺陷膜的难度。选择性层形成后，表面变得更加光滑。这一变化与之前对聚酰胺膜的研究结果一致[ 27 , 33 ]。
Fig. 4

  1. 下载：下载高分辨率图像 (704KB)

  2. 下载：下载全尺寸图像


图4 .裸露的α-Al 2 O 3管表面的3D图像（A）和高度测量（B）；膜 (M5) 表面的 3D 图像 (C) 和高度测量 (D)。


图5显示了膜的XPS数据。 M1和M2的N1s光谱的窄扫描分为两个峰成分（-NH-和-N=），证明了亚胺和苯并咪唑环的形成（图1B ）[ 16 ]。值得注意的是，M5中CP-2的使用降低了–N=的比例（从44.5%降至21.0%），表明CP-2中的部分–NH 2基团不与TPA反应。 CP-2的引入可以产生新的孔隙，调节聚合物链的结构并提高气体渗透性[ 8 ]。后交联后，分配给 –N< 的新峰出现，表明 TMC 或 DBX 成功交联（图 1 C 和 D）[ 34 ]。
Fig. 5

  1. 下载：下载高分辨率图像 (517KB)

  2. 下载：下载全尺寸图像


图5 。 N1s 光谱的 XPS 窄扫描 (A) 不含 CP-2 的膜（即 M1）； (B)膜：用CP-2(即M5)； (C)具有CP-2并通过TMC后交联的膜（即M7）； (D) 具有 CP-2 并通过 DBX 后交联的膜（即 M8）。


此外，对合成的膜进行了 ATR-FTIR 分析。如图6所示，膜中存在1290 cm -1 、1566 cm -1和1616 cm -1处的特征峰，表明苯并咪唑和亚胺键的形成（图1 B）[ 8 , 16 , 17 ]。可以发现，CP-2的添加降低了苯并咪唑环吸收峰的相对强度，因为XPS证明CP-2中的一些NH 2没有与TPA反应（图5B ）。 M7 1646 cm -1处的峰对应于C double bond 酰胺的O伸缩振动[ 35 ]，证明TMC和聚合物主链成功交联（图1C ）。交联膜的光谱中2920cm -1和2850cm -1处的特征峰的相对强度增强。它们归因于交联剂（即TMC和DBX）的C-H和新形成的C-N键（图1C和1D）[ 36 ]。这些进一步证实了后交联的成功。
Fig. 6
  1. Download: Download high-res image (342KB)
  2. Download: Download full-size image

Fig. 6. FT-IR spectra of the as-synthesized membranes.

3.2. Effect of preparation parameters on membrane performance

We had investigated different preparation parameters in order to enhance the membrane performance. Table 2 lists the performance of the membranes fabricated in this work. The monomer of organic phase was TPA, and its concentration was fixed at 0.5 wt%. The IP duration was 30 min. We had found that the H2 permeance of M1 was very low, only 12.7 GPU. The introduction of CP-2 could regulate the structure of polymer chains (As Fig. 5 shown), create more H2 channels, and consequently the H2 permeance was improved in M2, M3 and M4. In details, as shown in Fig. 5A, after adding 0.2 wt% CP-2, the H2 permeance increased from 12.7 GPU to 941 GPU. However, the H2/CO2 selectivity decreased from 4.03 to 3.48, demonstrating the formation of non-selective channels possibly caused by loose arrangement of polymer chains. The H2/CO2 selectivity increased to 4.85 with a 0.3 wt% CP-2 content, since more imine linkages could be formed to enhance the crosslinking degree and to improve the sieving capacity. However, further increase of CP-2 content to 0.5 wt% resulted in the decrease of both H2 permeance and H2/CO2 selectivity. It is largely due to the overpacking of the hybrid chains caused by excess CP-2.

Table 2. Performance of the membranes studied in this work. Feed: equal molar H2 and CO2, 2 bar and 423 K.

Membrane codeH2 Permeance (GPU)CO2 Permeance (GPU)H2/CO2 Selectivity
M112.73.154.03
M270.316.44.28
M386.1 ± 6.117.8 ± 2.34.87 ± 0.29
M49412703.48
M5203 ± 6.525.7 ± 1.87.91 ± 0.28
M6168 ± 9.030.8 ± 2.65.47 ± 0.17
M729.52.0814.2
M890.210.88.35
Apart from CP-2, the effect of BTA content was studied (Fig. 7B). Both H2 permeance and H2/CO2 selectivity were the highest when the BTA content was 1.5 wt%, reached 196 GPU and 8.18, respectively. Proper BTA content could well regulate the structure of the hybrid chains and form more H2 selective channels.
Fig. 7
  1. Download: Download high-res image (279KB)
  2. Download: Download full-size image

Fig. 7. Performance of the tubular membranes prepared with different CP-2 content (A) and BTA content (B). Feed conditions: equimolar H2 and CO2, 423 K and 2 bar. Error bars are standard deviations of performance of independent membranes fabricated with the same recipe.

In addition, we had tried to enhance the performance by post cross-linking. M7 and M8 had higher H2/CO2 selectivity and lower H2 permeance than M5, indicating denser structures were formed after post cross-linking. The selectivity of membrane crosslinked by TMC (M7) (n-hexane was used as the solvent) was higher than the membrane crosslinked by DBX (M8) (toluene was used as the solvent) while the change of H2 permeance was opposite. On the one hand, this phenomenon could be attributed to the cross-linking reagents since TMC has more reaction sites. On the other hand, it may be due to the influence of the solvent. The polarity of toluene (2.4) was higher than n-hexane (0.0). It means that toluene could promote the movement of polymer chains and facilitate the cross-linking.

3.3. Effect of feed conditions on membrane performance

It is important for H2/CO2 separation membranes to withstand high pressure. As shown in Fig. 8A, the H2 permeance of M5 slightly decreased from 196 GPU to 187 GPU when feed pressure rose from 2 bar to 6 bar. The H2/CO2 selectivity decreased from 8.18 to 6.86. These can be interpretated by the compaction of the hybrid chains under higher pressure, which resulted in a decrease of H2 selective channels. However, CO2 did not drop with the pressure possibly due to the slight swelling effect by virtue of the affinity between CO2 and the membrane layer.
Fig. 8
  1. Download: Download high-res image (530KB)
  2. Download: Download full-size image

Fig. 8. Effect of feed pressure (A) and temperature (B) on M5, and effect of feed temperature on M7 (C) and M8 (D). Feed conditions: equimolar H2 and CO2, 423 K and 2 bar. When one feed condition was studied, the others were fixed.

As shown in Fig. 8B, the H2 permeance of M5 increased from 196 GPU to 216 GPU as test temperature rose from 423 K to 473 K, indicating an activated diffusion, and the H2/CO2 selectivity dropped from 8.18 to 7.48. When the feed temperature was increased to 523 K, the H2 permeance and H2/CO2 selectivity sharply changed to 452 and 5.33, respectively. The decrease of selectivity may be caused by the violent movement of the hybrid chains, resulting in an increase of effective pore size of the membrane. Moreover, the selectivity didn't recover with the decrease of temperature. Large defects were caused by high temperature, which was demonstrated by Fig. 3J. In this case, viscous flow was an additional contributor to gas permeation besides the size-dependent selectivity of diffusion in the membrane, resulting in higher permeance and lower selectivity.
Since the high temperature of shift-gas, we tried to enhance the temperature resistance of membranes by post cross-linking. Fig. 8C shows the effect of temperature on the performance of the membrane crosslinked by TMC (M7). Compared with M5 (Fig. 8B), M7 exhibited the similar trend of performance change. This phenomenon was consistent with the acyl chloride crosslinked polyimides [37,38]. In contrast to M7 and M5, the membrane crosslinked by DBX (M8) delivered a simultaneous increase of H2 permeance and H2/CO2 selectivity with the rise of temperature (Fig. 8D). The H2 permeance and H2/CO2 selectivity reached 310 GPU and 9.37 at 523 K, respectively. The pore size of the M8 possibly expanded with the temperature, thereby created more H2 selective channels. The performance was well recovered as the temperature went back to 423 K. No visible morphology change was recognized after this loop test (Fig. 3K). These phenomena highlight that the membrane crosslinked by DBX (M8) possesses high thermal tolerance and can be used in the conditions with temperature variation.
A M5 was assessed for 74 h to evaluate the stability (Fig. 9). The membrane showed a steady H2 permeance around 210 GPU and H2/CO2 selectivity around 7.60 at 423 K and 2 bar. The results reveal that the molecular-scale hybrid tubular membranes have a great potential to be applied in industrial H2 purification.
Fig. 9
  1. Download: Download high-res image (253KB)
  2. Download: Download full-size image

Fig. 9. Membrane (a M5) stability for H2/CO2 separation. Feed conditions: equimolar H2 and CO2, 423 K and 2 bar. The membrane was placed on stream from 0 to 74th h while data were not collected during 13th to 64th h.

3.4. Membrane process

The flow diagram is shown in Fig. S2, the flow rate of feed gas was 4 mol s−1 in total (equal molar H2 and CO2). The calculation of the membrane process was based on the performance of M5-1. As shown in Fig. 10A, the H2 purity decreased from 98.3 % to 90.9 % with the increase of membrane area of second stage. In the case of limited selectivity, the CO2 concentration in permeate increased with the membrane area, resulting in a low H2 purity. However, it is beneficial to the H2 yield, the H2 yield increased from 26.4 % to 87.8 %. Although high H2 purity is not mandatory for gas turbines and industrial refining, certain H2 purity is wanted for these processes [4]. As shown in Fig. 10B, the H2 purity increased from 89. 3 %–94.2 % with the feed pressure increased from 2 bar to 4 bar. When the feed pressure was 6 bar, the H2 purity decreased to 90.9 % due to the degradation of membrane performance. It demonstrates that the pressure resistance of the membrane is of great importance to membrane process. Meanwhile, the increase of feed pressure is beneficial to the H2 yield, indicating the necessity of high-pressure operation.
Fig. 10
  1. Download: Download high-res image (283KB)
  2. Download: Download full-size image

Fig. 10. (A) Effect of membrane area of the second stage on H2 purity and H2 yield (the membrane area of the first stage was 500 m2, Feed conditions: 423 K, 6 bar); (B) Effect of feed pressure on H2 purity and H2 yield (the membrane area of the first stage and second stage were 500 m2 and 100 m2, respectively, Feed temperature: 423 K).

4. Conclusions

In summary, we successfully fabricated molecular-scale hybrid membranes (MHMs) containing benzimidazole linkages on large-area rough α-Al2O3 tubes for H2/CO2 separation. CP-2 was important for the creation of H2 selective channels. The membrane performance was enhanced by post cross-linking. The tubular membranes could effectively separate H2 and CO2 at high temperature (e.g., 523 K) and high pressure (e.g., 6 bar). Specially, the resulting membranes have a membrane area of 37.7 cm2 and easy to scale up with interfacial polymerization. Furthermore, we evaluated the feasibility of the application of resulting membranes in H2 production by a two-stage membrane system. The system could produce H2 with a purity of more than 95 %.

CRediT authorship contribution statement

Puxin Shi: Writing – original draft, Methodology, Investigation. Liping Luan: Writing – review & editing, Methodology. Bo Zhang: Methodology. Shenzhen Cong: Methodology. Zhi Wang: Writing – review & editing. Xinlei Liu: Writing – review & editing, Resources, Investigation, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was financially supported by the National Key Research and Development Program (2021YFB3801200), the Inner Mongolia Autonomous Region Unveiling Project (2022JBGS0027), and the Seed Foundation of Tianjin University (No. 2024XJD-0058).

Appendix A. Supplementary data

The following is the Supplementary data to this article:Download: Download Word document (2MB)

Multimedia component 1.

Data availability

Data will be made available on request.

References

Cited by (0)

View Abstract
© 2024 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.