Abstract

   Leveraging functional materials to develop advanced personal protective
   equipment is of significant importance for cutting off the propagation
   of infectious diseases, yet faces ongoing challenges owing to the
   unsatisfied antimicrobial efficiency. Herein a hierarchically porous
   cerium metal–organic framework (Ce‐MOF) boosting the antimicrobial
   performance by intensifying catalytic reactive oxygen species (ROS)
   generation and bacterial entrapment simultaneously is reported. This
   Ce‐MOF presents dendritic surface topography and hierarchical pore
   channels where the Lewis acid Ce sites are dispersedly anchored.
   Attributing to this sophisticated nanoarchitecture rendering the
   catalytic Ce sites highly accessible, it shows a ca. 1800‐fold activity
   enhancement for the catalytic conversion of atmospheric oxygen to
   highly toxic ROS compared to traditional CeO[2]. Additionally, the
   dendritic and negative‐charged surface engineered in this Ce‐MOF
   substantially enhances the binding affinity toward positive‐charged
   bacteria, enabling the spatial proximity between the bacteria and the
   short‐lived ROS and therefore maximizing the utilization of highly
   toxic ROS to inactivate bacteria. It is demonstrated that this
   Ce‐MOF‐integrated face mask displays almost 100% antimicrobial efficacy
   even in insufficient light and dark scenarios. This work provides
   important insights into the design of antibacterial MOF materials by a
   pore‐ and surface‐engineering strategy and sheds new light on the
   development of advanced self‐antimicrobial devices.

   Keywords: antibacterial material, catalysis, metal–organic framework,
   nanozyme, reactive oxygen species
     __________________________________________________________________

   Herein, the development of a protective face mask based on a
   hierarchically porous cerium metal–organic framework intensifying the
   bacterial entrapment ability and catalytic reactive oxygen species
   (ROS) generation at ambient condition is presented. It presents almost
   100% antimicrobial efficacies for different bacteria even though in
   insufficient light (e.g. on cloudy and rainy days) and dark scenarios
   (e.g. at night).

   graphic file with name ADVS-12-2410703-g006.jpg

1. Introduction

   Globalization leads to the transmission of infectious diseases more
   irregular and faster. This issue was well presented in the microbial
   infection events such as coronavirus disease 2019 (COVID‐19),^[ [46]^1
   ^] influenza virus,^[ [47]^2 ^] Middle East Respiratory syndrome
   coronavirus (MERS‐CoV)^[ [48]^3 ^] and Ebola virus,^[ [49]^4 ^] which
   cost millions of lives and disturbed global commerce and pose a serious
   global safety concern. Infectious diseases are mainly transmitted from
   person to person through direct contact and/or respiration. To dispute
   the spread of infectious diseases, wearing protective equipment such as
   face mask, bioprotective clothing, and medical glove, is highly advised
   by the standard infection prevention and control guidance.^[ [50]^5 ,
   [51]^6 ^] Despite wearing relevant protective equipment is capable of
   minimizing the biological threats, the infection risk still cannot be
   totally eradicated, as most of the available protective equipment can
   only intercept rather than kill the microbes.^[ [52]^7 , [53]^8 ^] Even
   worse, the accumulated microbes on/in protective equipment are prone to
   trigger cross‐contamination and aggravate infectious disease
   transmission.^[ [54]^9 , [55]^10 ^] Hence, it is highly desired to
   exploit next‐generation antibacterial protective devices capable of
   physical intercept and chemical neutralization of threatening germs.

   A feasible approach to confer protective equipment antimicrobial is
   incorporating them with biocides.^[ [56]^11 , [57]^12 , [58]^13 ,
   [59]^14 , [60]^15 , [61]^16 ^] To this point, photocatalytic
   antimicrobial materials have recently been extensively engineered to
   meet the basic demands for bioprotective applications, because of their
   ability to photocatalytically produce reactive oxygen species (ROS),
   which are capable of serving as highly active oxidants to inactivate
   harmful microorganisms.^[ [62]^11 , [63]^12 , [64]^13 , [65]^15 ,
   [66]^16 , [67]^17 , [68]^18 ^] The photocatalytic equipment, however,
   are highly relied on the input of photo energy. Although sunlight is
   sustainably accessible, their immediate biocidal functions suffer from
   the fast decline or quench under dim light (e.g., on cloudy and rainy
   days) and/or dark conditions (e.g., at night). This limits the
   antimicrobial efficacy and practicability of the photocatalytic
   equipment in diverse occasions.

   In nature, some oxidase enables the catalytic generation of ROS using
   well‐sourced oxygen as the electron acceptor in ambient condition.^[
   [69]^19 , [70]^20 , [71]^21 ^] Limited by the structural vulnerability
   and expensive cost of natural oxidase,^[ [72]^22 ^] the stable and
   cost‐effective Ce nanoparticle, e.g. ceria (CeO[2]), has garnered great
   interest in mimicking the oxidase‐like catalytic reaction, attributing
   to its strong Lewis acidity and facile electron cycle between Ce^3+ and
   Ce^4+.^[ [73]^23 , [74]^24 ^] This offers new opportunities to design
   the next‐generation self‐antimicrobial protective equipment in an
   energy input‐ and chemical‐addition‐free principle. However, the ROS
   generation efficiency of Ce nanomaterials, to date, is far from
   satisfactory, because of their dense structure that significantly
   reduces the accessibility of active Ce sites. At the same time, the
   lifespan of catalytic ROS is extremely short with an effective radius
   of action less than 200 nm,^[ [75]^19 , [76]^25 , [77]^26 ^]
   complicating the utilizing of ROS for targeting and inactivating
   bacteria.

   Herein, we present the development of a protective face mask based on a
   hierarchically porous cerium metal–organic framework (Ce‐MOF), which
   can boost self‐antimicrobial activity by virtue of intensifying ROS
   catalytic generation and bacterial entrapment simultaneously (Scheme
   [78]1 ). The hierarchically porous Ce‐MOF is synthesized by a polymer
   microemulsion template‐directed coordination assembly method (Figure
   [79]1a). In this method, the amphipathic block copolymers of
   Pluronic®P‐123 (PEO[20]PPO[70]PEO[20], termed P123) and Pluronic®F‐127
   (PEO[106]PPO[70]PEO[106], termed F127) are aggregated and then form
   linear micelles in aqueous solution. The addition of hydrophobic
   toluene transforms these linear micelles into columnar microemulsion
   templates, which co‐stabilized by P123 and F127. Finally, the abundant
   PEO segments on the columnar microemulsion surface form
   crown‐ether‐type complexes with Ce^4+, facilitating the in‐place growth
   of Ce‐UiO‐66 around the columnar microemulsion templates.^[ [80]^27 ,
   [81]^28 ^] After removing the columnar microemulsion templates, the
   Ce‐MOF featuring dendritic surface topography and hierarchical pore
   channels is formed. This hierarchically porous Ce‐MOF renders the
   catalytic Ce sites highly accessible, facilitating the catalytic
   conversion of atmospheric oxygen to highly toxic ROS, which is ca.
   1800‐fold higher than traditional CeO[2]. Importantly, the dendritic
   surface shows a much higher binding affinity toward bacteria,
   significantly shortening the migration distance of short‐life ROS to
   bacteria and thus maximally harnessing the ROS to inactivate bacteria.
   This Ce‐MOF is functionally and structurally stable, evidenced by the
   fact that 70% activity of ROS catalytic generation is preserved after
   storing at room temperature up to 300 days. The feasibility of
   utilizing the Ce‐MOF for fabricating antimicrobial protective mask is
   also demonstrated, presenting almost 100% antimicrobial efficacies for
   different bacteria even under insufficient light (e.g., on cloudy and
   rainy days) and dark scenarios (e.g., at night).

Scheme 1.

   Scheme 1
   [82]Open in a new tab

   Schematic illustration of the hierarchically porous Ce‐MOF intensifying
   catalytic ROS generation and bacterial entrapment simultaneously for
   engineering antimicrobial mask.

Figure 1.

   Figure 1
   [83]Open in a new tab

   a) Schematic representation of the polymer microemulsion
   template‐mediated coordination assembly for synthesizing MD‐Ce‐UiO‐66.
   The SEM b) and TEM images c,d) of MD‐Ce‐UiO‐66 nanoagent. e) Left: The
   Cryo‐EM image showing the (111) lattice planes throughout the
   MD‐Ce‐UiO‐66 nanoagent. Right: The crystallographic model of UiO‐66
   highlights the (111) lattice plane.

2. Results and Discussion

2.1. Synthesis of Ce‐MOF With Hierarchically Porous Channels and Dendritic
Surface

   The nanoarchitecture of a material dictates both its catalytic activity
   and interface chemistry, impacting the subsequent antimicrobial
   performance.^[ [84]^29 , [85]^30 , [86]^31 ^] With this in mind, we
   herein engineered a Ce‐MOF with hierarchically porous channels and
   dendritic surface, aiming at maximally harnessing the catalytic ROS to
   inactivate bacteria. The synthesis principle was based on a
   microemulsion polymer‐templated coordination assembly between cerium
   (IV) ammonium nitrate and 1,4‐benzenedicarboxylic acid (BDC)^[ [87]^27
   , [88]^28 ^] (Figure [89]1a). After removing the microemulsion
   template, the Ce‐MOF with mesopore‐macropore channels and dendritic
   surface were obtained, named MD‐Ce‐UiO‐66 (M and D mean
   mesopore‐macropore channels and dendritic shape surface, respectively).
   Powder X‐ray diffraction (PXRD) revealed the high crystallinity of this
   MD‐Ce‐UiO‐66 nanoagent, and the Bragg diffraction pattern was
   consistent well with standard UiO‐66 MOF (Figure [90]S1, Supporting
   Information), the structure of which is formed by the coordination
   linkage of 12‐connected Ce[6] cluster and 1,4‐benzenedicarboxylic acid
   in a fcu topology.^[ [91]^32 ^] In addition, the elemental analysis by
   the inductively coupled plasma mass spectrometry (ICP‐MS) revealed the
   Ce ratio of 39.33 wt% in MD‐Ce‐UiO‐66. This was in good agreement with
   the theoretical value of 41.51%, estimated based on the chemical
   formula of Ce[6]O[4](OH)[4](BDC)[6] of UiO‐66 (Figure [92]S2 and Table
   [93]S1, Supporting Information). The slightly decreased Ce content was
   plausible to be caused by the defective formation in 12‐connected fcu
   MOF^[ [94]^33 ^] (vide infra). Further insight into the
   crystallographic structure of MD‐Ce‐UiO‐66 was profiled by cryoelectron
   microscopy (Cryo‐EM) imaging. The lattice fringes of 1.2 nm, assigned
   to the (111) plane of fcu topology, were observed throughout the
   nanoagent, which was well in line with the simulated fcu UiO‐66 model
   (Figure [95]1e).

   The scanning electron microscopy (SEM) image showed that this
   as‐prepared MD‐Ce‐UiO‐66 nanoagent possessed a well‐defined spherical
   structure with a diameter ≈1 µm (Figure [96]S3, Supporting
   Information). A closer observation disclosed the existing of numerous
   mesopore and macropore channels (ca. 40–80 nm, Figure [97]1b). The
   dendritic surface and highly open porous channels of MD‐Ce‐UiO‐66 were
   further demonstrated by the transmission electron microscopy (TEM)
   images (Figure [98]S4, Supporting Information). The TEM image at higher
   amplification (Figure [99]1c) revealed that the opening pore channels
   deviated out from the core of the sphere and distributed radially in
   all directions. Besides, the pore channels were closely adjacent
   throughout the whole nanoparticle with a pore wall of ∼10 nm
   (Figure [100]1d), which can maximize the mass transfer rate and the
   accessibility of the interior Ce active sites. Additionally, the N[2]
   adsorption/desorption isotherms (Figure [101]S5, Supporting
   Information) described the type‐IV isotherm (Figure [102]S5a,
   Supporting Information) with a large hysteresis loop at high relative
   pressure (P/P[0]) higher than 0.48, and the simulated pore‐size
   distributions (Figure [103]S5b–d, Supporting Information) indicated the
   structural integration of micropore, mesopore and macropore in
   MD‐Ce‐UiO‐66. Collectively, these results demonstrated that the Ce‐MOF
   with a hierarchically porous structure and dendritic surface was
   formulated.

2.2. Catalytic Generation of ROS at Ambient Condition

   We first inspected the Ce valence of MD‐Ce‐UiO‐66 by X‐ray
   photoelectron spectroscopy (XPS). At the same time, the standard
   Ce‐UiO‐66 was also synthesized as a comparison (Figure [104]S6,
   Supporting Information). The C, O, and Ce elements were recorded in
   both samples (Figure [105]S7, Supporting Information), and the
   high‐resolution Ce 3d spectra disclosed the presence of Ce^3+ and Ce^4+
   in both MD‐Ce‐UiO‐66 and Ce‐UiO‐66 samples (Figure [106] 2a). The
   Ce^4+/Ce^3+ ratio (Table [107]S2, Supporting Information) of
   MD‐Ce‐UiO‐66 was calculated to be 1.68, which was close to that of
   standard Ce‐UiO‐66 (1.69), indicating that the pore and surface
   engineering in MD‐Ce‐UiO‐66 could not alter the chemical compositions
   and the valences of Ce[6] clusters. The closer structural insights by
   thermogravimetric analysis (TGA) revealed that the ligand‐missing
   defects were formed in both MD‐Ce‐UiO‐66 (0.52 linker deficiency per
   Ce[6] cluster) and Ce‐UiO‐66 (1.04 linker deficiency per Ce[6] cluster)
   (detailed calculation method was described in Quantitative analysis of
   TGA and Figure [108]S8, Supporting Information). Such defect formation
   is also observed in Zr‐UiO‐66 reported in the previous work.^[ [109]^33
   , [110]^34 , [111]^35 ^] This can be interpreted by the fact that the
   highly ligand‐connected number and dense topology of UiO‐66 trends to
   form ligand‐ or cluster‐missing defects. Of specific note, these
   defects favor the exposure of Ce active sites, which enables the
   molecular oxygen binding and consequently the catalytic generation of
   ROS (vide infra).

Figure 2.

   Figure 2
   [112]Open in a new tab

   a) The Ce 3d XPS spectra of Ce‐UiO‐66 and MD‐Ce‐UiO‐66. b) The
   catalytic kinetics curves of bulky CeO[2], microporous Ce‐UiO‐66, and
   MD‐Ce‐UiO‐66 under the same Ce amount (15 µg mL^−1). c) The EPR spectra
   of DMPO‐O[2] ^−• for CeO[2], Ce‐UiO‐66, and MD‐Ce‐UiO‐66. d) The
   simulated structure of O[2] molecule adsorbed on the defective Ce[6]
   cluster, in which the ligated aqua was pre‐removed. The corresponding
   adsorption energy and O─O bond length of the adsorbed O[2] molecule
   were highlighted. e) The charge density difference of O[2] molecule
   adsorbed on defective Ce[6] cluster. f) The magnetic moment of O[2]
   molecule adsorbed on defective Ce[6] cluster. Red ball represented O
   atom; purple ball represented adsorbed O atom; yellow ball represented
   Ce atom; gray ball represented C atom; pink ball represented H atom.

   The oxidase‐like activity was estimated using
   3,3′,5,5′‐tetramethylbenzidine (TMB) as the oxidase substrate. TMB can
   be oxidized by O[2] under the catalytic function of oxidase, and the
   catalytic product has a blue color, which is easy to be monitored by
   UV–vis spectrometer. Both Ce‐UiO‐66 and MD‐Ce‐UiO‐66 presented the
   catalytic activity for TMB oxidization at ambient condition, without
   the extra addition of chemical reagents or light irradiation
   (Figure [113]2b). In addition, the catalytic activity was not affected
   under either light irradiation or dark conditions (Figure [114]S9,
   Supporting Information), suggesting a spontaneous catalytic process.
   The initial catalytic rate and conversion (at 90 s) of MD‐Ce‐UiO‐66
   were 1839‐fold and 96.71‐fold as the bulky CeO[2] under the same Ce
   amount (Figure [115]S10, Supporting Information), indicating that the
   discrete Ce active sites in MD‐Ce‐UiO‐66 were favorable for the
   catalytic reaction compared to the dense Ce sites in CeO[2]
   (Figure [116]2b). In addition, the MD‐Ce‐UiO‐66 presented 3.01‐fold
   initial catalytic rate and 2.77‐fold conversion (at 90 s) compared to
   Ce‐UiO‐66 (Figure [117]2b). Of note, as it was difficult to control the
   microemulsion templates, the resulting MD‐Ce‐UiO‐66 had a non‐uniform
   particle size. We concurrently prepared three batch samples of
   MD‐Ce‐UiO‐66 (Figure [118]S11a, Supporting Information), all of which
   presented uneven particle size but formed rich mesopore and macropore
   channels in each particle (Figures [119]S11b–d, Supporting
   Information). Upon assessing the oxidase activity of the three batch
   samples, we found a minimal variation of activity with a low RSD value
   of 3.2% (Figure [120]S12a, Supporting Information). Additionally, we
   also conducted triplicate activity tests on the same batch sample and
   found that they presented negligible differences of activity (RSD =
   0.49%, Figure [121]S12b, Supporting Information). These results
   indicated that the heterogeneity in particle size of MD‐Ce‐UiO‐66 had a
   negligible impact on its oxidase activity.

   Considering that both the Ce[6] cluster formulation and valence state
   were identical in these two samples, the catalytic activity enhancement
   could be ascribed to the consecutive meso‐macropore channels in
   MD‐Ce‐UiO‐66 that facilitated the mass transfer and rendered the
   interior Ce active sites more accessible. This view was further
   supported by the observation of differential activity in MD‐Ce‐UiO‐66
   before and after polymer template removal (Figure [122]S13, Supporting
   Information). MD‐Ce‐UiO‐66 showed 2.88‐fold activity enhancement after
   polymer template removal, validating the key role of opening
   meso‐macropore channels for catalytic efficiency.

   The mechanism of oxidase‐like catalytic activity was then investigated.
   To verify that the Ce[6] cluster was indeed the catalytic center, we
   first synthesized an isostructural MOF, the Ce[6] cluster of which was
   replaced by Zr[6] cluster (denoted as Zr‐UiO‐66, Figure [123]S14,
   Supporting Information). No oxidase‐like activity was recorded in
   Zr‐UiO‐66 (Figure [124]S15, Supporting Information), manifesting that
   the catalytic center was originated from the Ce[6] cluster in
   MD‐Ce‐UiO‐66. Under the N[2] atmosphere, we found that the catalytic
   reaction of MD‐Ce‐UiO‐66 and Ce‐UiO‐66 completely stopped (Figure
   [125]S16, Supporting Information), indicating that O[2] was the
   indispensable electron acceptor in the oxidase‐like reaction. Further
   insight into the catalytic mechanism was surveyed by density functional
   theory (DFT) calculation. The anterior TGA analysis implied that the
   linker‐missing defect was formed in the Ce[6] cluster of MD‐Ce‐UiO‐66
   (Figure [126]S8, Supporting Information), and the defect sites would be
   ligated by one hydroxo and one aqua ligand, respectively.^[ [127]^36 ^]
   Based on this consideration, we modeled the optimal Ce[6] cluster with
   a linker‐missing site, onto which one hydroxo and one aqua ligand was
   ligated (Figure [128]S17, Supporting Information). The calculated O[2]
   adsorption energy (−2.579 eV, Figure [129]2d) onto this defective Ce
   site was much lower than the original aqua ligand adsorption energy
   (−1.183 eV, Figure [130]S17, Supporting Information), indicating that
   the adsorption of O[2] molecule onto this defective Ce site was
   energetically favorable. The charge density difference image suggested
   that the electron transferred from Ce[6] cluster to the bonded O[2]
   molecule, as well evidenced by the electron density increased in the
   O[2] molecule while reduced in the Ce atoms (Figure [131]2e).
   Furthermore, the O‐O bond length of the bonded O[2] molecule was
   determined to be 1.33 Å (Figure [132]2d) and the total magnetic moment
   of bonded O[2] was ≈1.26 μB (Figure [133]2f), both of which were well
   in line with the characteristic parameters of a superoxide radical.^[
   [134]^14 ^] These theoretical calculations revealed that the highly
   accessible Ce site in MD‐Ce‐UiO‐66 facilitated the catalytic conversion
   of O[2] to superoxide radical, attributing to the distinct valence
   electronic structure of Ce (4f^15d^16s^2) favoring the catalytic cycle
   of Ce^3+/Ce^4+.^[ [135]^24 ^] A closer examination of ROS species was
   carried out by the electron paramagnetic resonance (EPR) experiment,
   which indeed identified the strong signal of superoxide radical (O[2]
   ^−•) using (5,5‐diemthyl‐1‐pyrroline N‐oxide (DMPO) as a spin‐trap
   agent in both the catalytic reaction of MD‐Ce‐UiO‐66 and Ce‐UiO‐66
   (Figure [136]2c). In contrast, almost no O[2] ^−• signal was detected
   in CeO[2] under the same Ce amount, well in line with its poor
   oxidase‐like catalytic activity. At the same time, the signal of
   hydroxyl radical (•OH) using DMPO as a spin‐trap agent was also
   identified (Figure [137]S18a, Supporting Information), whereas no
   singlet oxygen (^1O[2]) was found using 2,2,6,6‐tetramethylpiperidine
   (TEMP) as the spin‐trap agent during the catalytic reaction (Figure
   [138]S18b, Supporting Information). Based on these observations, we
   infer the following catalytic process occurring at the Ce sites: After
   the adsorption of O₂ onto the Ce sites, an electron is transferred from
   the Ce site to the O₂ molecule, leading to the formation of O₂⁻^•.
   Subsequently, the H₂O molecule is oxidized to •OH at the Ce site, which
   simultaneously regenerates the Ce sites for further catalysis.

   The durability of this MD‐Ce‐UiO‐66 for ROS catalytic generation was
   then examined. As seen in Figure [139]S19a,b (Supporting Information),
   MD‐Ce‐UiO‐66 could reserve more than 85% of its original oxidase‐like
   activity after soaking it in water for 12 d, and the catalytic
   abilities for O[2] ^−• and •OH generation were also well retained
   (Figure [140]S20, Supporting Information), attributing to the stable
   crystallographic structure (Figure [141]S21, Supporting Information)
   and the negligible change of Ce constituents (Figure [142]S22 and Table
   [143]S3, Supporting Information). The long‐lasting activity of this
   MD‐Ce‐UiO‐66 was further estimated by storing it at room temperature up
   to 300 d. It still could maintain 70% catalytic activity (Figures
   [144]S19c,d, Supporting Information), with a preserved catalytic
   function for O[2] ^−• and •OH generation (Figure [145]S23, Supporting
   Information).

2.3. The Interfacial Interaction Between MD‐Ce‐UiO‐66 and Bacteria

   Apart from the catalytic activity, it entails the interfacial
   interactions between the materials and bacteria to make full use of the
   short‐life ROS. In nature, the macrophages in the human immune system
   can effectively capture bacteria, benefiting by its uneven surface
   which enhances the contacting area and binding affinity toward
   bacteria.^[ [146]^37 , [147]^38 ^] Previously reports have found that
   the electrophilic Ce can readily bind with electron‐rich phosphate
   groups,^[ [148]^39 , [149]^40 ^] which are abundantly expressed on the
   cytomembrane. Therefore, we speculated that the dendritic‐shaped
   surface engineered in MD‐Ce‐UiO‐66 might promote interfacial
   interactions with bacteria. To verify this, a batch of Ce‐MOFs with
   various morphologies were synthesized, including the 2D Ce‐BTB‐MOL,
   cuboidal Ce‐PCN‐224, clavate Ce‐NU‐1000 and octahedral Ce‐MOF‐808
   (Figures [150]S24 and [151]S26, Supporting Information). All these
   Ce‐MOFs presented comparable Ce^4+/Ce^3+ constituents (Figures [152]S27
   and [153]S28, and Table [154]S2, Supporting Information), and also held
   the oxidase‐like activity at ambient condition (Figure [155]S29,
   Supporting Information), owing to the generation of highly active O[2]
   ^−• and •OH (Figure [156]S30, Supporting Information). In addition, a
   similar catalytic mechanism like MD‐Ce‐UiO‐66 was also verified in
   these Ce‐MOFs (Figures [157]S31–S34, Supporting Information). Of
   specific note, our MD‐Ce‐UiO‐66 still presented the highest catalytic
   activity (Figure [158]S35, Supporting Information), ascribed to the
   widespread meso‐macropore channels favoring the mass transfer and
   rendering the interior Ce sites more accessible.

   The interfacial interaction between Escherichia coli (E. coli) bacteria
   and these Ce‐MOFs was then examined. We found that the surface charges
   of the Ce‐MOFs were different. Ce‐PCN‐224, Ce‐NU‐1000, and Ce‐BTB‐MOL
   possessed negative surface charges while MD‐Ce‐UiO‐66, Ce‐UiO‐66, and
   Ce‐MOF‐808 had positive surface charges (Figure [159]S36, Supporting
   Information). This differential surface charge was plausible caused by
   the different surface termination patterns of Ce[6] cluster termination
   and carboxylate linker termination in these MOFs. In this regard,
   MD‐Ce‐UiO‐66, Ce‐UiO‐66, and Ce‐MOF‐808 were easy to adhere negatively
   charged bacteria through electrostatic attraction, as evidenced by the
   SEM images (Figure [160]3a). While Ce‐PCN‐224, Ce‐NU‐1000, and
   Ce‐BTB‐MOL exhibited very limited ability for bacteria capture because
   of the electrostatic repulsion (Figure [161]3a). A further inspection
   of the binding affinity between these Ce‐MOFs and bacteria was
   conducted by atomic force microscopy (AFM) (Figure [162]3b; Figure
   [163]S37, Supporting Information). Both positively charged Ce‐MOFs of
   MD‐Ce‐UiO‐66, Ce‐UiO‐66, and Ce‐MOF‐808 presented higher binding forces
   toward bacteria compared to the negatively charged counterparts of
   Ce‐PCN‐224, Ce‐NU‐1000, and Ce‐BTB‐MOL. Of note, although the surface
   potentials were similarly positive among MD‐Ce‐UiO‐66, Ce‐UiO‐66, and
   Ce‐MOF‐808, the MD‐Ce‐UiO‐66 dominated the strongest interaction with
   bacteria, with an average binding force of 5.22 N, which was 3.55‐fold
   and 3.92‐fold higher than that of Ce‐UiO‐66 and Ce‐MOF‐808,
   respectively (Figure [164]3c; Figure [165]S38, Supporting Information).
   Additionally, the confocal laser scanning microscope (CLSM) experiments
   were carried out to validate the strong interaction between
   MD‐Ce‐UiO‐66 and bacteria. The overlap of green and red fluorescence
   was observed after immediately mixing SYTO9‐labeled E. coli (green
   fluorescence) with rhodamine‐loaded MD‐Ce‐UiO‐66 (RhB‐MD‐Ce‐UiO‐66, red
   fluorescence, Figure [166]S39, Supporting Information), suggesting the
   high bacteria‐capturing ability of the MD‐Ce‐UiO‐66. Furthermore, the
   Cryo‐EM imaging showcased that the rod‐shaped E. coli was tightly
   adhered to the dendritic‐shaped surface of MD‐Ce‐UiO‐66 (Figure
   [167]S40, Supporting Information). At the same time, in the
   high‐resolution Ce 3d XPS spectra, we found that the binding energy of
   Ce 3d was shifted to lower energy after the interfacial interaction
   formation between MD‐Ce‐UiO‐66 and bacteria (Figure [168]3d),
   indicating that the molecular interactions between the surface
   electrophilic Ce site on MD‐Ce‐UiO‐66 and the electron‐rich phosphate
   moieties of bacteria cytomembrane (Figure [169]3e).^[ [170]^38 ,
   [171]^39 ^]

Figure 3.

   Figure 3
   [172]Open in a new tab

   a) SEM images showing the interaction between the Ce‐MOFs and bacteria.
   b) The AFM set up for the interaction force estimations between various
   E. coli and different nanoagnets. c) The estimated interaction forces
   by AFM force‐distance curves. d) The Ce 3d XPS spectra of MD‐Ce‐UiO‐66
   before and after incubation with E. coli. e) The schematic illustration
   of molecular interactions between the Ce[6] cluster of MD‐Ce‐UiO‐66 and
   the phosphate moieties of bacteria cytomembrane.

2.4. In Vitro Antibacterial Performances

   After validating the dual functions of MD‐Ce‐UiO‐66 for ROS catalytic
   generation and bacteria capture, we then examined the antibacterial
   performances against the representative gram‐negative bacteria of E.
   coli at ambient condition using a classical plate counting method. As
   shown in Figure [173]4a, incubating the E. coli with different Ce‐MOFs
   under the same Ce amount, the number of bacterial colonies in
   Ce‐BTB‐MOL (colony‐forming units reduction, CFU reduction = 25.2%,
   Figure [174]4b), Ce‐PCN‐224 (CFU reduction = 21.7%, Figure [175]4b) and
   Ce‐NU‐1000 (CFU reduction = 15.1%, Figure [176]4b) were only slightly
   lower than that of the blank group, implying that the antibacterial
   efficacies were feeble attributing to the inferior catalytic activities
   and poor bacteria‐capturing capacities (Figure [177]3a). Whereas the
   Ce‐MOF‐808 (CFU reduction = 91.5%, Figure [178]4b) and Ce‐UiO‐66 (CFU
   reduction = 62.3%, Figure [179]4b) presented an enhanced performance
   for combating E. coli bacteria, owing to their enhanced interactions
   with bacteria (Figure [180]3a). Of specific note, our MD‐Ce‐UiO‐66
   showed nearly 100% antibacterial efficacy (CFU reduction = 99.9%,
   Figure [181]4b). At the same time, the two‐color living/dead staining
   experiment confirmed that when incubating E. coli bacteria with
   Ce‐MOF‐808 and Ce‐UiO‐66, respectively, partial bacteria were still
   survive (Figure [182]S41, Supporting Information), in line well with
   classical plate counting results (Figure [183]4b). While in the case of
   using MD‐Ce‐UiO‐66 under the same Ce amount, no survival E. coli
   bacteria was found, suggesting the best antibacterial ability.

Figure 4.

   Figure 4
   [184]Open in a new tab

   a) Digital images of bacterial colonies of E. coli after treatments
   with different Ce‐MOFs and b) the corresponding bacterial viabilities.
   c) Principal components analysis in blank group and
   MD‐Ce‐UiO‐66‐treated group. d) Volcano plot of upregulated and
   downregulated genes in blank group and MD‐Ce‐UiO‐66‐treated group. e)
   Heatmap showing the differentially expressed genes between
   MD‐Ce‐UiO‐66‐treated group and the blank group. f) Bar plot showing the
   biological processes differentially regulated by MD‐Ce‐UiO‐66 compared
   with the blank group by gene ontology (GO) enrichment analysis. g) Bar
   plot showing the pathways differentially regulated by MD‐Ce‐UiO‐66
   compared with the blank group through Kyoto Encyclopedia of genes and
   genomes (KEGG) pathway enrichment analysis. h) Expression changes
   between MD‐Ce‐UiO‐66‐treated group and the blank group in the
   Alpha−amino acid catabolic process. i) The expression of cell wall
   organization‐related genes in MD‐Ce‐UiO‐66‐treated group and blank
   group.

   To deeply understand the molecular mechanism for the antibacterial
   activity of MD‐Ce‐UiO‐66, RNA sequencing (RNAseq) technology was
   applied to unveil the relevant pathways involved in the treatment of E.
   coli with MD‐Ce‐UiO‐66. The distinct differences in transcriptome
   landscapes between the MD‐Ce‐UiO‐66‐treated group and the blank group
   were found in the principal component analysis (PCA), indicating the
   obvious changes in gene expression patterns of E. coli after
   MD‐Ce‐UiO‐66 treatment (Figure [185]4c). Further analysis revealed that
   after MD‐Ce‐UiO‐66 treatment, there were a total of 1451 diversely
   expressed genes (DEGs) (flod change > = 2, adj. p‐value < 0.01),
   containing 716 upregulated and 735 downregulated genes
   (Figure [186]4d). Figure [187]4e presented a heat map of the
   hierarchical clustering of DEGs, showing the expression levels of each
   gene in the MD‐Ce‐UiO‐66‐treated group and the blank group. Enrichment
   analysis indicated that several signaling pathways, such as
   transcriptional regulation, nucleotide metabolism, protein homeostasis,
   and cell wall maintenance, were enriched by treating with MD‐Ce‐UiO‐66
   (Figure [188]4f,g). Notably, a large part of genes associated with
   amino acid metabolism were suppressed, suggesting that MD‐Ce‐UiO‐66
   affected the protein utilization in bacteria and triggered the
   depletion of biological processes including energy metabolism and
   oxidative phosphorylation. In addition, we noticed the significant
   inhibition of genes associated with bacterial cell wall maintenance
   (Figure [189]4h,i), which might be caused by the strong adhesion of
   MD‐Ce‐UiO‐66 on E. coli.

2.5. Self‐Antimicrobial Mask Workable in Different Occasions

   We next investigated the antibacterial activities of MD‐Ce‐UiO‐66
   against other bacteria including Staphylococcus aureus (S. aureus),
   Klebsiella pneumoniae (K. pneumoniae) and Pseudomonas aeruginosa (P.
   aeruginosa), the most common antibiotic‐resistant pathogens causing
   hospital‐associated infections. At ambient condition, it was found that
   the colony counts in the experimental groups of S. aureus (CFU
   reduction = 97.6%), K. pneumoniae (CFU reduction = 100%) and P.
   aeruginosa (CFU reduction = 94.7%) treated with MD‐Ce‐UiO‐66 were much
   lower than the blank groups without MD‐Ce‐UiO‐66 (Figures [190]5a and
   [191]S42a, Supporting Information). Furthermore, the two‐color
   dead/live staining analysis validated the nearly entire dead bacteria
   with red fluorescence (labeled by propidium iodide, PI, λ[ex] = 488 nm,
   λ[em] = 635 nm) in all experimental groups treated by MD‐Ce‐UiO‐66. As
   a striking contrast, numerous alive bacteria with bright green
   fluorescence (labeled by SYTO‐9, λ[ex] = 488 nm, λ[em] = 500 nm) were
   identified in the blank groups (Figure [192]5b). These results verified
   the broad‐spectrum antibacterial performance of our MD‐Ce‐UiO‐66 at
   ambient condition.

Figure 5.

   Figure 5
   [193]Open in a new tab

   Self‐antimicrobial mask workable in different occasions. a) Digital
   images of bacterial colonies of S. aureus, K. pneumoniae and P.
   aeruginosa in the blank groups and the MD‐Ce‐UiO‐66 groups, and b) the
   corresponding SYTO9/PI two‐color fluorescent images for the live (green
   fluorescence) and dead (red fluorescence) bacterial staining assay of
   different bacteria. Scale bar: 25 µm. c) The antibacterial plate assays
   (for E. coli) showing the antibacterial performances of MD‐Ce‐UiO‐66 in
   the dark and on rainy and sunny days.

   We next carried out antibacterial plate‐killing assays to verify the
   bacteriostatic efficiencies of MD‐Ce‐UiO‐66 in the dark, as well as on
   rainy and sunny days (Figure [194]5c). As indicated in agar plate
   images, the bacterial survival rates in all experimental groups treated
   by MD‐Ce‐UiO‐66 were close to zero and the CFU reduction in the dark
   and on rainy and sunny days were calculated to be 99.9%, 99.9% and
   99.8%, respectively (Figure [195]S42b, Supporting Information),
   demonstrating the spontaneous antimicrobial behavior of this
   MD‐Ce‐UiO‐66 without the dependency on light energy.

   The feasibility of using MD‐Ce‐UiO‐66 as a biocidal coating for the
   fabrication of personal protective equipment was then examined. As a
   proof of concept, we integrated MD‐Ce‐UiO‐66 into commercially
   available KN95 face masks through a simple spray deposition method. The
   formed MD‐Ce‐UiO‐66 mask consisted of three layers, of which the first
   and third layers were the woven fabric similar to KN95 masks and the
   middle layer was the MD‐Ce‐UiO‐66‐coated woven fabric, with a loading
   of 5 mg MD‐Ce‐UiO‐66 per 3×3 cm^2 square woven fabric (Figure [196]S43,
   Supporting Information). The sterilization efficiency of this
   conceptual MD‐Ce‐UiO‐66 face mask was evaluated. In this sterilization
   experiment, the outermost surfaces of the MD‐Ce‐UiO‐66‐integrated masks
   and original KN95 face masks were sprayed with the E. coli aerosols for
   5 min using an aerosol generator (flow rate: 0.3 mL min^−1, Figure
   [197]6a,b) for simulating the propagation process of the respiratory
   droplet. After incubating at 37 °C for 4 h, each layer of the
   as‐treated masks was then soaked into sterile water and vortexed. The
   resulting supernatants were further diluted and plated onto
   Luria‐Bertani agar plates, followed by incubation at 37 °C overnight
   for residual analysis of adhered visible colonies. As seen in
   Figure [198]6c,d, we noticed that many living bacteria were visible on
   the first layers of both the MD‐Ce‐UiO‐66 mask and KN95 mask. Whereas
   in our MD‐Ce‐UiO‐66 mask, no alive bacterium was identified in the
   middle or third layers that close to or in direct contact with the
   human face. In stark contrast, numerous living bacteria are still
   located on the middle and third layers of the KN95 mask
   (Figure [199]6c,d). Furthermore, we conducted a series of antimicrobial
   experiments on the MD‐Ce‐UiO‐66 face mask under varying humidity and
   temperature conditions. Specifically, we tested the mask in the
   following environments: 53% humidity at 23.5 °C; 36% humidity at
   36.9 °C; 86% humidity at 10.2 °C. As illustrated in Figure [200]S44
   (Supporting Information), in our MD‐Ce‐UiO‐66 masks, no alive bacterium
   was identified in the middle or third layers that close to or in direct
   contact with the human face, regardless of varying humidity levels and
   temperatures, indicating that our MD‐Ce‐UiO‐66 mask showed considerable
   antibacterial activity under different humidity and temperature
   conditions.

Figure 6.

   Figure 6
   [201]Open in a new tab

   a) Photographs of bioaerosol generation apparatus as well as the
   trilaminar KN95 mask and MD‐Ce‐UiO‐66 mask. b) Schematic representation
   of the bioaerosol passing across the face masks. (c, d) E. coli
   residual levels on the first, second, and third layers of our
   MD‐Ce‐UiO‐66 mask and the KN95 mask counterpart. # in c) indicates an
   undetectable level of CFU mL^−1 count.

   It was worth noting that the preparation of this MD‐Ce‐UiO‐66 was not
   associated with any energy‐intensive steps and the whole synthesis
   procedure was completed within 40 min in an aqueous solution at 40 °C.
   The total price of MD‐Ce‐UiO‐66 for one face mask is ca. $0.4, and the
   cost of this antimicrobial MD‐Ce‐UiO‐66 mask is estimated to be ca.
   $0.43 per mask (detailed analysis was described in The cost analysis of
   MD‐Ce‐UiO‐66 mask and Table [202]S4, Supporting Information).
   Considering the functional stability for ROS catalytic generation over
   300 d, this versatile and cost‐efficient MD‐Ce‐UiO‐66 would hold huge
   potential for scale‐up fabrication of multifunctional protective
   equipment. We also assessed the biosafety of the MD‐Ce‐UiO‐66 mask by
   evaluating both the water stability of the loaded particles and their
   migration potential. First, we studied the structural integrity of
   MD‐Ce‐UiO‐66 in an aqueous solution. In this experiment, 5 mg of
   MD‐Ce‐UiO‐66 powder was immersed in 1 mL of water for three days. ICP
   analysis showed that the Ce ratio in the supernatant was only 0.2%,
   indicating that the MD‐Ce‐UiO‐66 particles were highly water‐stable and
   exhibited excellent resistance to leaching in aqueous conditions. In
   addition, after the bioaerosol simulation experiment, the third layer
   of MD‐Ce‐UiO‐66 mask, which was in direct contact with the human face,
   was removed and examined by SEM. It showed the absence of MD‐Ce‐UiO‐66
   powder (Figure [203]S45, Supporting Information), indicating that the
   MD‐Ce‐UiO‐66 coated in the middle layer could not migrate to the
   innermost layer of the mask. Based on these results, we concluded that
   the MD‐Ce‐UiO‐66‐loaded mask was biosafe.

3. Conclusion

   In summary, by virtue of pore and surface engineering, we develop an
   antimicrobial Ce‐MOF intensifying ROS catalytic activity and bacteria
   entrapment. This synergistic effect can maximally harness the
   short‐life ROS to inactivate bacteria, achieving a dynamical promotion
   of antibacterial efficacy against different kinds of pathogens at
   ambient condition. Note that this antimicrobial Ce‐MOF is functionally
   stable for up to 300 days and its synthesis is procedurally simple and
   low‐cost, with the cost of the antimicrobial face mask estimated to be
   as low as $0.43 per face mask, holding great potential for the scale‐up
   industrial manufacture. This work provides new insights into the access
   to next‐generation antibacterial materials by MOF pore‐ and
   surface‐engineering approach, advancing the development of
   self‐antibacterial protective devices.

Conflict of Interest

   The authors declare no conflict of interest.

Supporting information

   Supporting Information
   [204]ADVS-12-2410703-s001.docx^ (17.1MB, docx)

Acknowledgements