A hydrogen adsorbing PUR/Pd nanocomposite nanofibrous membrane prepared by electrospinning technology

Jakub Hoskovec*^a, Pavla Čapková^a, Petr Ryšánek^a, Dániel Gardenö^c, Karel Friess^c, Jaroslava Jarolímková^a, Viktor Greguš^b, Pavel Kaule^b, Tereza Dušková^b, Magda Škvorová^b, Václav Šícha^b and Oldřich Benada^de
aCentre of Nanomaterials and Biotechnology, Faculty of Science, J.E. Purkyne University, Pasteurova 3632/15, 400 96 Usti nad Labem, Czech Republic. E-mail: Jakub.Hoskovec@ujep.cz; Tel: +420 475 286 647
^bDepartment of Chemistry, Faculty of Science, J. E. Purkyne University, Pasteurova 3632/15, 400 96 Usti nad Labem, Czech Republic
^cUniversity of Chemistry and Technology Prague, Department of Physical Chemistry, Technicka 5, Prague 6 – Dejvice, 16628, Czech Republic
^dDepartment of Biology, Faculty of Science, Jan Evangelista Purkyne University, Za Valcovnou 1000/8, 400 96 14 Usti nad Labem, Czech Republic
^eInstitute of Microbiology of the Czech Academy of Sciences, Videnska 1083, 142 00 Prague 4, Czech Republic

First published on 28th August 2024

---

---

1 Introduction

Hydrogen storage materials and methods represent a significant challenge in development of the hydrogen energy industry.^1–6 Among the various methods of hydrogen storage, those based on adsorption prevail, using various materials.^4–7 Requirements for H₂-storage materials are primarily high storage capacity and low thermodynamic stability for easy sorption and desorption at low temperatures. Various hydrogen-absorbing materials have been considered for hydrogen storage, particularly membranes based on palladium, platinum, and their alloys that form hydrides.^8–13 Other metals such as Zr, Ti, Nb, Al, and Mg^14–17 have also been successfully tested in hydrogen capture membranes, but the energy and temperatures required for H₂ sorption and desorption are significantly higher for these metals and their alloys than for Pd and Pt. Other gas separation media aimed at hydrogen capture reported in the literature are modified zeolites,^18–22 MOFs,^23–28 and a wide range of carbon-based materials like carbon nanotubes (SWCNTs), activated carbon, alkali-doped graphite, and pure and metal-doped graphite nanofibers.^29–33

The advantage of palladium and platinum compared to other hydrogen sorbents is their ability to form hydrides even under mild conditions.^34,35 Other advantages of these metals include the simplicity of the technology for the preparation of functional units whose design uses thin metal layers. The disadvantage of these metal parts is the well-known fact that the hydration of metals makes them unbearably brittle. This is why palladium nanoparticles and other transition metals have attracted interest as promising hydrogen sorbents.^36–40 Nanoparticles will not only eliminate the problem of embrittlement of metal membranes, but their unique properties will also multiply the hydrogen sorption capacity. However, nanoparticle use requires anchoring on a suitable scaffold that meets conditions like high specific surface area, high porosity, and sufficient thermal stability. From this point of view, an electrospun polymeric 2D nanofibrous sheet appears to be a suitable scaffold for hydrogen-sorbing nanomaterials. In addition, the air-permeable membrane in the form of a nanofibrous layer will contribute to the acceleration of sorption–desorption cycles.

An electrospun nanofiber PAN membrane as the carrier of Zr-MOF and Cr-MOF nanocrystals was reported in ref. 27. Although MOF nanocrystals anchored on fibers achieved a lower sorption capacity than pure MOF powders, hydrogen sorption was more efficient in the case of the nanofiber/MOF composite. With a 20 wt% loading of MOF nanocrystals, the composites were able to achieve over 50% of the H₂ uptake capacity of individual MOF nanocrystals. Therefore, the anchoring of absorbent substances to nanofibers is advantageous, especially in the case of expensive sorbents. Xia et al. achieved a good hydrogen sorption capacity of electrospun composite nanofibers PVA/Li₃N, which were calcined at 550 °C,⁴¹ which makes the design of functional devices difficult for practical applications. Calcination of electrospun PVP and PAN/PFSA (PFSA – perfluorinated sulfonic acid) nanofiber membranes at 900 °C, as reported in ref. 42, also led to excellent hydrogen storage capacity, and calcination of electrospun materials as a final technological step is also recommended in the review paper.⁴³ Although calcination leads to a dramatic increase in hydrogen sorption capacity, the disadvantage of this technology is the change in the macroscopic structure of the membrane, which loses its self-supporting properties. In fact, most of the articles describing the technology for the preparation of hydrogen-sorbing nanofibrous materials with final calcination present the resulting sorbing material in powder form, which complicates the design of functional devices for practical application. An exception is the work of Alicia Vergara-Rubio, who reported a self-supporting membrane created from an electrospun PVA membrane with subsequent acidic activation treatment followed by pyrolysis.⁴⁴

There are various approaches for the preparation of selective sorbent materials, and a general problem in practical applications is the design of a functional device. If the active sorbents are in powder form, the design of the functional unit requires additional technological steps to fix the active powder substances on a solid substrate.

In this work, we tried to create a self-supporting, breathable hydrogen-sorbing nanocomposite membrane based on PUR polymer electrospun nanofibers decorated with Pd nanoparticles with a minimum number of technological steps. The primary motivation of this work is to achieve optimal functionality of the membrane, which will also be easily usable in working devices.

2 Materials and methods

2.1 Materials and spinning solution

Polyurethane (PUR)-Larithane AL 286 was purchased from CoimGroup; dimethyl-formamide 99% (DMF) (MW = 73.10 g mol⁻¹) and tetrahydrofuran (THF) (MW = 72.11 g mol⁻¹) were purchased from AvantorTM; hydrazine (MW = 32.05 g mol⁻¹ [anhydrous basis]) and palladium(II) chloride (PdCl₂) (MW = 177.33 g mol⁻¹) were purchased from Sigma-Aldrich Company.

Electrospinning technology generally offers great variability in the selection and use of spinning and collection electrodes. These specific modifications of the technology can then be advantageously combined with given types of experimental methods and desired requirements, which are also used in this work, where two electrospinning technologies within the InoSPIN device (InoCure s. r. o., Czech Republic) and the corresponding initial optimized PUR solution were used.

In the case of the one-step modification procedure of nanofiber membranes, which was described in more detail in subsection 2.3 Modification methods, denoted as approach (2.3.1), needle electrospinning was used, which in previous experiments and comparisons proved to be more advantageous for the spinning of mixtures using a needle as the spinning electrode and a rotating cylinder as the collecting electrode. Within this technological setup, an optimised PUR/DMF stock solution of 45 wt% was prepared. To this pre-prepared, perfectly homogenised working solution of the base material, a quantity of 2 wt% palladium chloride-modifying salt was subsequently added under constant stirring (600 rpm). The solution was stirred for 24 hours at 30 °C.

For the method of postspinning modification of a pure nanofibrous PUR membrane (2.3 Modification methods (2.3.2)), the initial needle electrode was replaced by an edge electrode consisting of a conductive prism with a thickness of 3 mm and a length of 30 cm for higher yield and better homogeneity over the entire collection area within the same device (hereinafter referred to as edge electrospinning). The rewinding collector was used as a collection electrode. Here again, a starting solution of PUR 45 wt% optimised for the technology was prepared using a DMF/THF two-component solvent system in a 1:1 ratio. Homogenization was carried out under constant stirring (300 rpm) at room temperature.

Finally, a stock-reducing solution of hydrazine and distilled water was prepared in a 1:200 ratio for both of these approaches to modify the PUR nanofiber membranes.

2.2 Spinning conditions

Needle electrospinning: the modified solution with the dissolved chloride salt Pd was successively subjected to needle electrospinning with one needle as the spinning electrode and a rotating cylinder as the collecting electrode. The following working parameters were used for the PdCl₂ modified solution during the process: applied voltage (potential difference): 50 kV; working distance: 123 mm; solution dispensing rate: 0.1 ml min⁻¹; needle diameter: 0.7 mm; and rotating collector speed: 500 rpm. The controlled ambient conditions corresponded to 23 °C and 30% humidity. A standard type of non-woven antistatic polypropylene (PP) fabric (hereafter referred to as spunbond) was used as the collection material for the final product. After the electrospinning process, polyurethane fiber-based membranes were obtained, in which PdCl₂ chlorides were retained in their structure and on their surface, mainly due to electrostatic forces.

Edge electrospinning: the initial PUR/DMF/THF solution was transported by an injection pump at a dosing rate of 0.5 ml min⁻¹ into a special slider, which, at a set speed of 30 mm s⁻¹, performs a repetitive motion on the edge, thus wetting its surface. The whole process was carried out at an applied voltage (potential difference) of 80 kV between the electrodes, whose working distance was 150 mm. The solvent-free nanofibrous membrane was collected on the planar surface of the rewinding collector via a spunbond with a defined collection rate for the desired yield.

2.3 Modification methods

2.4 Sample characterization and the sorption test

Nanofibrous samples of the initial unmodified PUR (hereinafter referred to as PUR pristine) and PUR containing 2 wt% modifying chloride salt of palladium (Pd) prepared using the needle electrospinning process and treated in a post-reduction step (hereinafter referred to as PUR/PdCl₂) were characterised by a series of analytical methods: Scanning Electron Microscopy (SEM), High Resolution Electron Microscopy (HRSEM), Energy Dispersive Spectroscopy (EDS), X-Ray Diffraction (XRD), X-Ray Fluorescence (XRF), air permeability, zeta potential (ζ), Pressure Decay (PD), and dynamic Gas Chromatography (dGC) were used to determine morphology, crystallinity, nanoparticle size, elemental percentage composition of the material, surface chemistry, and resulting sorption capacity with respect to H₂.

Standard SEM analysis was performed using a VEGA3 instrument (Tescan, Brno, Czech Republic). For the analysis itself, the samples were sputtered on a Q 150T ES sputtering station (Quorum Technologies Ltd, Ringmer, UK) with 20 nm gold for optimal SE yield. All images were taken in resolution mode with the detector set to secondary electrons (SE) at a magnification of 10000× and a constant accelerating voltage of 20 kV. The average values of nanofiber diameters were directly evaluated using the dedicated TESCAN software. High-resolution electron microscopy (HRSEM) and X-ray microanalysis were performed using an FEI Nova NanoSEM (FEI, Brno, Czech Republic) equipped with an Ametek ®EDAX Octane plus SDD detector and TEAM EDS analysis system (AMETEK B. V.; Tilburg, The Netherlands). The nanotissue samples were prepared in the following way: six circular cutouts with a diameter of 3 mm were mounted onto standard SEM stubs in two rows. The first was sputter-coated with 3 nm of platinum for high-resolution imaging at 3 kV. For X-ray microanalysis performed at 15 kV, five nanometers of carbon were evaporated onto the second row. Final images from carbon-coated samples were acquired in backscattered electron mode (BSE) at 3 kV using a concentric CBS detector at magnifications of 25000× and 100000×. The EDS spectra were preprocessed using a TEAM EDAX analysis system, with final post-processing in NIST DTSA-II software Microscopium, 2022-0606 revision (Newbury and Ritchie, 2015;⁴⁵ Ritchie et al., 2012⁴⁶).

The structure and phase composition of nanofiber membranes were analysed by X-ray diffraction (XRD). A PANalytical X'Pert PRO X-ray diffractometer in the symmetrical Bragg–Brentano configuration under CuKα radiation (λ = 1.5418 Å) was used. A parallel X-ray mirror was located in the path of the primary beam, and a parallel plate collimator was located in the path of the diffracted beam. This configuration was used due to the low intensity of the diffracted beam of nanofiber textiles. For the detection of diffracted X-ray beams, a 1D detector X'Celerator was used.

X-ray fluorescence spectroscopy (XRF) was performed on a wave dispersive XRF spectrometer Rigaku Primus IV using standard-less SQX analysis based on the fundamental parameter method. The concentration of the polymer was calculated by normalising the result to 100%, which is typical of standard methods.

The zeta potential of fiber surfaces was evaluated on a SurPASS instrument (Anton Paar, Austria). At room temperature, the samples were studied inside the adjustable gap cell in contact with the electrolyte (0.001 mol per dm³ KCl). All the samples were measured four times at constant pH (pH = 6.9) with a relative error lower than 5%, and the current streaming method and the Helmholtz–Smoluchowski equation have been applied to calculate the zeta potential.

The air permeability parameter indicating the material's degree of porosity/breathability was analysed for individual samples on the equipment of Polymertest (Otrokovice, Czech Republic). The measurement was carried out according to the standard procedure defined by ČSN EN ISO 9237 for determining the air permeability of fabrics. Parameters such as a test area of 20 cm² and a pressure gradient of 100 Pa were constant during the measurement. Tests for individual membranes were repeated under the same conditions ten times, constantly at a place where no measurements had been made before.⁴⁷

2.5 Hydrogen sorption test

The sorption of the H₂ Pressure Decay (PD) method up to 4 bar was determined using a self-developed dual-chamber PD apparatus.⁴⁸ The device contains a sample chamber fabricated from titanium grade 5 (Ti–6Al–4V) with a volume of 10 cm³ and stainless-steel capillary tubing (0.55 cm³) between inlet and outlet valves serving as a gas chamber. Prior to all experiments, the system's volume was calibrated for each tested gas and tested as leakage-proof up to 20 bar. The entire apparatus was immersed in a water bath and thermostated by using an immersion circulator (Julabo Corio C). Sorption experiments were performed at 30 ± 0.1 °C. The results were further processed with a MATLAB routine using the Peng–Robinson EOS to obtain the sorption isotherms.

The method of dynamic Gas Chromatography (dGC), implemented with the help of a gas chromatograph Agilent USA Inc., 7890B and 7697A, programme OpenLab CDS 2.3 (2.3.0) with an external stainless steel fixed bed cell (Galaktech China; internal volume of 20 cm³) placed in an air chamber with an automatically controlled temperature with an accuracy of ±0.5 °C, enables faster and more repeatable determination of sorption and desorption capacities for hydrogen in bulky, low-mass flexible solid samples in a flow system at atmospheric pressure. The development of the dGC breakthrough method for H₂ sorption/desorption capacity determination and device was inspired mainly by the work of Bhatta,⁴⁹ Yang,⁵⁰ and García⁵¹ based on the analysis of sorption or desorption capacities for CO₂. During the measurement, the input gas mixture stably contains 1.0 ± 0.1% H₂ in N₂ (Messer Technogas ČR) or N₂ of purity 99.999% and flows through the entire sample at a constant flow rate of 30 cm³ min⁻¹. Before the start of the measurement, each sample is desorbed from most of the adsorbed H₂ by vacuum desorption at least 10 minutes after the stabilisation of the limiting vacuum (0.2 Torr) at an automatically controlled temperature with an uncertainty of ±0.5 °C. The resulting adsorption/desorption capacities of the materials were evaluated using the so-called breakthrough curves, which express the dependence of gas concentration on the time of sorption or desorption. The points of these curves were calculated using a formula used, for example, in the papers^49–51 to calculate the sorption capacities for CO₂, usually from triplicates of experimentally obtained H₂ values using a numerical integration method, the trapezoidal method, in the Excel programme of Microsoft Office Standard 2019.

3 Results and discussion

3.1 Nanofiber fabric morphology and structure

The method of preparation and labels of all prepared samples are summarised in Table 1. The use of different approaches to modify the nanofibrous materials was, as expected, also reflected in the resulting morphology of the materials, as can be seen from the overview images (Fig. 1A–D), which compare PUR in the unmodified initial state (Fig. 1A), in the state modified using the one-step method (Fig. 1B), and modified using the postspinning method (Fig. 1C and D). Considering the practice and the reproducibility of the fiber production process in relation to PUR modification with PdCl₂, the production of the initial PUR fibers followed by subsequent postspinning modification proved to be a more advantageous approach. This approach balances the undesirable effects that were observed in the modification using mixed spinning (one-step approach). The addition of the palladium salt directly to the spinning solution completely changes its chemistry, which, while increasing the conductivity of the solution, is also reflected in the electrospinning itself, which is highly unstable despite the optimized technological setup of the process and further increases with increasing amounts of modifying salts, as demonstrated previously.⁵²

All these effects lead to fibers with a smaller diameter than that of the postspinning-modified fibers, whose value within the measurement error corresponds to the initial unmodified PUR (see Table 2). Furthermore, there are significant defects in the form of beads and fused layers related to the local accumulation of Pd in areas of the solution exposed to a strong electric field, which was also ultimately manifested in the case of air permeability measurements, when due to this high defectivity, its value was slightly increased against the expected decrease despite the increased surface mass generated by the aforementioned conductivity of the solution. Based on this fact, the so-called maximum threshold value of the concentration of the modifying substance was determined experimentally, at which it is still possible to spun the mixed material in a stable process. This concentration corresponds to 2 wt% PdCl₂, which was used in the case of the sample shown in Fig. 1B. Thus, in summary, the postspinning modification method offers a number of advantages compared to the mixed approach, which are (1) preserving the non-defectivity of the initial spun polymer, whose overall morphological nature is easily controlled and, above all, tunable thanks to the mere presence of the components (polymer/solvent). (2) The fact of using a chemical or physical approach to activate the surface before the modification itself. (3) Possible use of a higher concentration of the modifying substance above the established maximum threshold value.

---

High-resolution scanning electron microscopy (HRSEM) images (Fig. 2A–D) capture a more detailed view of the surface of the modified fibers with reduced Pd nanoparticles, the size of which (Table 2) was determined from the peak profiles using X-ray diffraction analysis based on the calculation of the Scherer formula.

The imaging of treated samples in backscattered electrons clearly shows the presence of metal nanoparticles as bright spots not appearing on pure PUR (Fig. 2A). Thus, it is an efficient tool for comparison of the modified and unmodified surfaces. Thanks to this visualisation, it is also possible to optically compare their distribution and the tendency to cluster when, as can be seen, very good coverage of the polymer surface with nanoparticles applies to the samples treated by the postspinning approach (Fig. 2C and D). Here, the effect of the applied physical approach for surface activation through UV radiation is also very visible, which caused a more homogeneous coverage with a simultaneous reduction in the formation of aggregates (Fig. 2D).

In the case of Fig. 2B, representing a sample of one-step approach modification, a preferential orientation of nanoparticles was revealed in the resulting defect formations in the form of beads, the surface of which is captured in the relevant image. The size and distribution of nanoparticles in space are much more random in this case, but on the other hand, no tendency to cluster into larger objects, e.g., coatings, was observed, which can most likely be attributed to the steric effect of PUR macromolecules on PdCl₂ before and during the spinning process itself.

XRD structure analysis revealed the crystallinity of the PUR nanofibers and supported the conclusive presence of Pd nanoparticles on the membrane surface. Fig. 3 shows the diffraction patterns for the original PUR membrane and the PUR/Pd prepared by various methods. The black profile corresponding to the pristine PUR membrane shows a single dominant broadened peak for ∼20° 2θ (hereafter referred to as the PUR peak), which indicates two distinctive features of the fiber structure: (1) the small size of PUR crystallites manifested by the broaden PUR peak; and (2) the preferred orientation of PUR crystallites, where the polymer chains are ordered parallel to the fiber axis, which means the absence of reflections belonging to hkl planes perpendicular to the fiber axis.

XRD profile broadening for palladium reflections indicates the small size of Pd crystallites. The lowest halfwidth of Pd reflections corresponds to the largest Pd nanocrystallites in the sample PUR_Pd_SOL. A comparison of XRD profiles for all PUR/Pd samples is shown in Fig. 4.

For the peak occurrence areas characteristic of Pd falling within the range of values ∼40–82 2θ (Fig. 4A and B), the manifestation of the amount of Pd nanoparticles is quite noticeable. An obvious phenomenon can be stated: as the percentage of Pd in the PUR matrix increases, there is a rise in the number of reflective positions and, consequently, the intensity of individual profiles.

In addition, the effect of UV activation of the fiber surface is also visible, which led to the binding of a larger amount of Pd (see also Table 3 and Subsection 3.2) than in the case of the non-UV-exposed sample (green vs. blue profile).

---

3.2 Surface chemistry and sorption capacity of membranes

Zeta potential is a reliable indicator of changes in the chemical composition of the membrane surface and confirms the presence of additives on the fiber surface. Table 3 shows the zeta potential changes of the PUR/Pd membranes compared to the pure PUR membrane.

Here, too, the zeta potential values confirm the fact of the positive benefit of UV surface activation, together with the performed EDS analysis as shown ub Fig. 5. (Green profile) and at the same time indicate that, within the framework of the one-step modification (PUR_Pd_SOL), Pd is partially absorbed by the volume of PUR nanofibers. This is evident when comparing the peaks belonging to C Kα and Pd Lβ and Pd Lα. The surface of pure PUR, knowing its chemical structure, has a dominant representation of C and O₂ in the range of energy used for EDS analysis, which confirms the course of the red profile for the sample with one-step modification, when due to the dominant presence of Pd in the fiber volume, the Pd intensities for Pd Lβ and Pd Lα, were suppressed, with a simultaneous increase in intensities for C Kα and O Kα, which distorts Pd Mγ in the overlap. Finally, the peak belonging to Cl Kα indicates the residual HCl product that remained in the samples despite thorough washing.

Fig. 6 and Table 4 show the pressure dependence of H₂ sorption capacity obtained by the PD method. Surprisingly, the PUR nanofibrous substrate itself shows a significant sorption capacity. Since the sample PUR_Pd_PS(UV) showed the highest sorption capacity, further sorption tests using the dGC method were performed only for this sample.

---

Compared to other methods used in the field, the dGC breakthrough method makes it possible to selectively determine the content of sorbed and desorbed hydrogen under mild conditions in the presence of nitrogen, which competes with sorbed hydrogen on the surface of the material under given conditions. The yields of the repetitive cycles of sorption and hydrogen desorption in a stream of nitrogen (Fig. 7) can thus approach the real values of hydrogen capture, for example, from pyrolysis gas mixtures. Therefore, the values of hydrogen sorption capacities for all three materials obtained by the pressure decay method (Fig. 6) are approximately 4.8–8.3 times higher than the results of the dynamic GC method, as described in Tables 4 and 5.

---

Desorption capacities obtained exclusively by the dGC breakthrough method also provide information about H₂ desorption at 25 °C by changing the sorption equilibrium due to the flow of pure nitrogen. Amazingly, PUR pristine material excels with the best yields of 66% H₂, and 60% H₂ at 50 °C after three adsorption–desorption cycles measured at ambient pressure. Fig. 7 (left diagram) shows the H₂ sorption capacity measured by this method for the PUR_Pd_PS(UV) sample. The right diagram in Fig. 7 illustrates the desorption test measured by the dGC method at atmospheric pressure in a stream of nitrogen. It can be seen that the maximum hydrogen loss of about 11 ml g⁻¹ is reached after about 15 minutes, where the majority of the hydrogen loss comes from the PUR substrate. After 15 minutes, the membrane is stable, i.e., it retains a constant hydrogen content at room temperature and normal atmospheric pressure.

The low values of H₂ desorption yields in the material PUR_Pd(UV) after three repetitions can be explained mainly by the formation of a thermodynamically more stable interstitial hydride made of palladium nanoparticles after the material comes into contact with hydrogen. This hypothesis can be supported by the presence of changes in the Bragg angles measured by XRD in the material exposed to H₂ in comparison with the starting material stored in air (Fig. 8).

The advantage of nanoPd modified PUR after surface pretreatment by UV radiation is the selectivity of Pd in chemisorption of H₂ with the formation of interstitial PdxHy hydride. Unfortunately, the PUR used as a matrix with a large surface area for nanoPd deposition is not a suitable material for the study of H₂ desorption at higher temperatures because, above 60 °C, the molten fibers stick together.

Therefore, for the use of Pd nanoparticles deposited only on the surfaces of flexible nanostructured fibers obtained by the method of electrostatic spinning of polymers, it will be necessary to further search for a material that will effectively cover its surface with Pd nanoparticles and at the same time selectively and efficiently desorb chemisorbed H₂ at higher temperatures or under other experimental conditions.

4 Conclusions

The membranes prepared in this work have several advantages compared to previously published membranes, designed in the form of powdered materials or thin Pd layers.

The maximum achieved sorption capacity under normal pressure at 1 bar is 207.7 ml g⁻¹ (1.9 wt%), and at 4 bar, it is 1173.3 ml g⁻¹ (10.6 wt%). The membrane shows the stability needed for practical use, which means the hydrogen content captured in the membrane remains stable under ambient conditions.

Comparing the achieved results with those of other authors is actually difficult due to the different experimental conditions of sorption—high pressures and temperatures—used by various authors, for example.^35,39,40 Nevertheless the sorption capacity of our PUR/Pd membrane measured at 1 bar and room temperature of ∼207.7 ml g⁻¹ corresponds to 1.9 wt%, which can be compared with the value of ∼1 wt% achieved at 1 bar by J. Ren in ref. 27 with a PAN nanofiber membrane decorated with MOF nanocrystals.

The results of this work yielded a number of useful partial findings significant for both technology and membrane characterization, including the methodology of sorption and desorption tests, which can be summarised as follows:

• First, functional additives mixed directly into the spinning solutions reduce the number of process steps; however, additives localised within the fibers lose functionality, which is a problem especially for metal nanoparticles and MOFs.

• Second, UV activation of nanofiber surfaces leads to a higher content of modifying agents on the surface of the membranes and also brings a higher homogeneity of distribution of the modifying agent on the fiber surface itself.

• Third, although the choice of polymeric substrate can enhance the sorption capacity through chemisorption, however, sorption on the substrate is inefficient for hydrogen retention in the membrane.

This work explores advanced methods for preparing nanofibrous composite materials based on PUR/Pd. It introduces a novel dGC characterization method, unpublished until now, that enables simultaneous study of the sorption and desorption of H₂. The method simulates real-life process conditions using a gas stream.

Author contributions

Jakub Hoskovec: investigation, methodology, formal analysis, visualization, validation, writing – original draft; Pavla Čapková: conceptualization, funding acquisition, project administration, resources, supervision, writing – review & editing; Petr Ryšánek: investigation, formal analysis, project administration; Dániel Gardenö: investigation, formal analysis; Karel Friess: supervision, resources, writing – review & editing; Jaroslava Jarolímková: investigation; Viktor Greguš: investigation; Pavel Kaule: investigation; Tereza Dušková: investigation; Magda Škvorová: formal analysis; Václav Šícha: supervision, resources, writing – review & editing; Oldřich Benada: investigation, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge the assistance provided by the Research Infrastructure NanoEnviCz, supported by the Ministry of Education, Youth and Sports of the Czech Republic under project NanoEnviCz (No. LM2023066). This work was also supported by the Technology Agency of the Czech Republic [TAČR, No. TK05020080], by the SGS project (No. UJEP-SGS-2022-53-006-3), and project ERDF/ESF “UniQSurf – Centre of biointerfaces and hybrid functional materials” (No. CZ.02.1.01/0.0/0.0/17_048/0007411). The authors gratefully acknowledge the access to the electron microscopy facility of IMIC, supported by the Czech Academy of Sciences (RVO CZ61388971). The financial support from the Ministry of Education, Youth and Sports of the Czech Republic for the specific university research (MSMT No. 21-SVV/2022 and 2023) is gratefully acknowledged.

References

1. M. R. Usman, Renewable Sustainable Energy Rev., 2022, 167, 112743 CrossRef CAS .
2. N. S. Muhammed, B. Haq, D. Al Shehri, A. Al-Ahmed, M. M. Rahman and E. Zaman, Energy Rep., 2022, 8, 461–499 CrossRef .
3. R. R. Shahi, A. K. Gupta and P. Kumari, Int. J. Hydrogen Energy, 2023, 48, 21412–21428 CrossRef CAS .
4. M. Hirscher, V. A. Yartys, M. Baricco, J. Bellosta von Colbe, D. Blanchard, R. C. Bowman, D. P. Broom, C. E. Buckley, F. Chang and P. Chen, et al., J. Alloys Compd., 2020, 827, 153548 CrossRef CAS .
5. R. P. Prachi, M. M. Wagh and C. A. Gangal, Adv. Energy Power, 2016, 4, 11–22 Search PubMed .
6. A. Züttel, Mater. Today, 2003, 6(9), 24–33 CrossRef .
7. Y. Xu, Y. Deng, W. Liu, X. Zhao, J. Xu and Z. Yuan, Sustain. Energy Technol. Assessments, 2023, 55, 102974 CrossRef .
8. S. Dekura, H. Kobayashi, K. Kusada and H. Kitagawa, ChemPhysChem, 2019, 20, 1158–1176 CrossRef CAS PubMed .
9. S. Wagner and A. Pundt, Int. J. Hydrogen Energy, 2016, 41, 2727–2738 CrossRef CAS .
10. Ø. Hatlevik, S. K. Gade, M. K. Keeling, P. M. Thoen, A. P. Davidson and J. D. Way, Sep. Purif. Technol., 2010, 73, 59–64 CrossRef .
11. C. Goyhenex and L. Piccolo, Phys. Chem. Chem. Phys., 2017, 19, 32451–32458 RSC .
12. M. Baldauf and D. M. Kolb, Electrochim. Acta, 1993, 38, 2145–2153 CrossRef CAS .
13. I. A. Pašti, N. M. Gavrilov and S. V. Mentus, DFT Study, Adv. Phys. Chem., 2011, 2011, 305634 CrossRef .
14. J. Cermak and L. Kral, J. Alloys Compd., 2018, 744, 252–259 CrossRef CAS .
15. C. Zlotea, et al., Hydrogen sorption in TiZrNbHfTa high entropy alloy, J. Alloys Compd., 2019, 775, 667–674 CrossRef CAS .
16. H. R. Z. Sandim, B. V. Morante and P. A. Suzuki, Mater. Res., 2005, 8(3), 293–297 CrossRef CAS .
17. L. Kral and J. Cermak, Int. J. Hydrogen Energy, 2019, 44, 13561–13568 CrossRef CAS .
18. F. Cacho-Bailo, B. Seoane, C. Téllez and J. Coronas, J. Membr. Sci., 2014, 464, 119–126 CrossRef CAS .
19. Y. Li, F. Liang, H. Bux, W. Yang and J. Caro, J. Membr. Sci., 2010, 354, 48–54 CrossRef CAS .
20. X. Gu, Z. Tang and J. Dong, Microporous Mesoporous Mater., 2008, 111, 441–448 CrossRef CAS .
21. A. L. Khan, A. Cano-Odena, B. Gutiérrez, C. Minguillón and I. F. J. Vankelecom, J. Membr. Sci., 2010, 350, 340–346 CrossRef CAS .
22. M. Rezakazemi, K. Shahidi and T. Mohammadi, Int. J. Hydrogen Energy, 2012, 37, 14576–14589 CrossRef CAS .
23. K. Suresh, D. Aulakh, J. Purewal, D. J. Siegel, M. Veenstra and A. J. Matzger, J. Am. Chem. Soc., 2021, 143, 10727–10734 CrossRef CAS PubMed .
24. S. S. Kaye, A. Dailly, O. M. Yaghi and J. R. Long, J. Am. Chem. Soc., 2007, 129, 14176–14177 CrossRef CAS PubMed .
25. W. Bao, et al., Int. J. Hydrogen Energy, 2023, 48, 36010–36034 CrossRef CAS .
26. S. Zhou, et al., Int. J. Hydrogen Energy, 2013, 38, 5338–5347 CrossRef CAS .
27. J. Ren, N. M. Musyoka, P. Annamalai, H. W. Langmi, B. C. North and M. Mathe, Int. J. Hydrogen Energy, 2015, 40, 9382–9387 CrossRef CAS .
28. Y.-G. Lee, H. Moon, Y. Cheon and M. Suh, Angew. Chem., Int. Ed., 2008, 47, 7741–7745 CrossRef CAS PubMed .
29. Y. Yürüm, A. Taralp and T. N. Veziroglu, Int. J. Hydrogen Energy, 2009, 34, 3784–3798 CrossRef .
30. P. Ramirez-Vidal, et al., ACS Appl. Mater. Interfaces, 2021, 13, 12562–12574 CrossRef CAS PubMed .
31. Y. Kojima, Y. Kawai, A. Koiwai, N. Suzuki, T. Haga, T. Hioki and K. Tange, J. Alloys Compd., 2006, 421, 204–208 CrossRef CAS .
32. F. Song, L. Huang, H. Ding, S. Zhang and J. Yu, Sustainability, 2023, 15, 8722 CrossRef CAS .
33. B. Panella, M. Hirscher and S. Roth, Carbon, 2005, 43, 2209–2214 CrossRef CAS .
34. P. D. Cobden, B. E. Nieuwenhuys, V. V. Gorodetskii and V. N. Parmon, Platin. Met. Rev., 1998, 42, 141–144 CrossRef CAS .
35. B. D. Adams and A. Chen, Mater. Today, 2011, 14, 282–289 CrossRef CAS .
36. M. Yamauchi, H. Kobayashi and H. Kitagawa, ChemPhysChem, 2009, 10, 2566–2576 CrossRef CAS PubMed .
37. C. Zlotea and M. Latroche, Colloids Surf., A, 2013, 439, 117–130 CrossRef CAS .
38. H. Kobayashi, M. Yamauchi, H. Kitagawa, Y. Kubota, K. Kato and M. Takata, J. Am. Chem. Soc., 2008, 130, 1828–1829 CrossRef CAS PubMed .
39. H. Kobayashi, M. Yamauchi, H. Kitagawa, Y. Kubota, K. Kato and M. Takata, J. Am. Chem. Soc., 2008, 130, 1818–1819 CrossRef CAS PubMed .
40. G. Li, H. Kobayashi, S. Dekura, R. Ikeda, Y. Kubota, K. Kato, M. Takata, T. Yamamoto, S. Matsumura and H. Kitagawa, J. Am. Chem. Soc., 2014, 136, 10222–10225 CrossRef CAS PubMed .
41. G. Xia, D. Li, X. Chen, Y. Tan, Z. Tang, Z. Guo, H. Liu, Z. Liu and X. Yu, Adv. Mater., 2013, 25, 6238–6244 CrossRef CAS PubMed .
42. X. Chen, et al., RSC Adv., 2021, 11, 4053–4061 RSC .
43. J. Seong Mu, in Hydrogen Storage, ed. L. Jianjun, IntechOpen, Rijeka, 2012, p. 8 Search PubMed .
44. A. Vergara-Rubio, L. Ribba, D. E. P. Borregales, K. Sapag, R. Candal and S. Goyanes, ACS Appl. Nano Mater., 2022, 5, 15353–15361 CrossRef CAS .
45. D. E. Newbury and N. W. M. Ritchie, J. Mater. Sci., 2015, 50(2), 493–518 CrossRef CAS PubMed .
46. N. W. M. Ritchie, D. E. Newbury and J. M. Davis, Microsc. Microanal., 2012, 18(4), 892–904 CrossRef CAS PubMed .
47. K. Zdeňka, et al., in Polymer Science, ed. Y. Faris, IntechOpen, Rijeka, 2013, p. 8 Search PubMed .
48. M. Klepić, Gas permeation and sorption in polymer-ionic liquid membranes, PhD thesis, UCT Prague, 2020 .
49. L. K. G. Bhatta, S. Subramanyam, M. D. Chengala, U. M. Bhatta, K. Venkatesh and V. Raghavendra, Curr. Sci., 2017, 112, 835–838 CrossRef CAS .
50. Y. Yang, H. Li, S. Chen, Y. Zhao and Q. Li, Langmuir, 2010, 26, 13897–13902 CrossRef CAS PubMed .
51. S. García, M. V. Gil, C. F. Martín, J. J. Pis, F. Rubiera and C. Pevida, Chem. Eng. J., 2011, 171, 549–556 CrossRef .
52. J. Hoskovec, P. Capkova, Diploma thesis, University of J.E. Purkyne, Usti nad Labem, 2021 .

This journal is © The Royal Society of Chemistry 2024