Water Molecule-Triggered Anisotropic Deformation of Carbon Nitride Nanoribbons Enabling Contactless Respiratory Inspection

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 Jun 2021Water Molecule-Triggered Anisotropic Deformation of Carbon Nitride Nanoribbons Enabling Contactless Respiratory Inspection Yuye Zhang†, Yongxiu Song†, Yanfei Shen, Kaiyang Chen, Qing Zhou, Yanqin Lv, Hong Yang, Ensheng Xu, Songqin Liu, Lei Liu and Yuanjian Zhang Yuye Zhang† Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 †Y. Zhang and Y. Song contributed equally to this work.Google Scholar More articles by this author , Yongxiu Song† Institute for Advanced Materials, School of Material Science and Technology, Jiangsu University, Zhenjiang 212013 †Y. Zhang and Y. Song contributed equally to this work.Google Scholar More articles by this author , Yanfei Shen Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Kaiyang Chen Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Qing Zhou Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Yanqin Lv Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Hong Yang Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Ensheng Xu Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Songqin Liu Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author , Lei Liu *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Institute for Advanced Materials, School of Material Science and Technology, Jiangsu University, Zhenjiang 212013 Google Scholar More articles by this author and Yuanjian Zhang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Jiangsu Engineering Laboratory of Smart Carbon-Rich Materials and Device, Jiangsu Provincial Hi-Tech Key Laboratory for Biomedical Research, School of Chemistry and Chemical Engineering, Medical School, Southeast University, Nanjing 211189 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.020.202000361 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesTrack Citations ShareFacebookTwitterLinked InEmail The exploitation of the interaction between nanostructured matter and small molecules, such as H2O at interfaces via dynamic hydrogen bonding, is essentially the key for smart, responsive nanodevices but remains challenging. Herein, the authors report that the carbon nitride nanoribbons (CNNRs) with an anisotropic intraplanar and interplanar molecular arrangement underwent a deformation by H2O triggering. Both experiments of bulk samples and single nanoribbons disclosed that the reversible formation of a hydrogen-bonded H2O adsorption layer was the source of the CNNRs deformation, reminiscent of the hydration-triggered twist of natural bean pods in seeding. Nonetheless, CNNRs had a more balanced H2O affinity, enabling a superior response and recovery time. By coupling with carbon nanotubes, the authors also converted the deformation of CNNRs into more straightforward electrical readouts with record-fast response time. Further applied to capture fluctuations in humidity in real-time respiration, a higher detection sensitivity was obtained in a contactless mode, which compared favorably with the clinical breath-testing station. Given the carbon nitride family with various C/N ratios, surface properties, and topography, this finding that CNNRs are an outstanding H2O transducer would significantly pave the way for the H2O-triggered smart devices in broad prospective applications. Download figure Download PowerPoint Introduction The exploration of nanostructured and low-dimensional materials with exceptional electronic, optical, and mechanical properties has increasingly attracted interest in many fields.1,2 For instance, such external stimuli as heat, light, electricity, pH, and solvents would significantly perturb the microscopic shape of graphene, carbon nanotubes (CNTs), Mxene, and molybdenum disulfide that feature high surface-to-volue ratios, abundant interfacial active sites, and/or unique crumpled topography.3–6 As often strongly associated to their electronic, magnetic, and/or chemical properties, the exploiting of microscopic change of these low-dimensional matters in shape is envisioned to open up a new way for their prospective applications, for example, as smart, responsive nanodevices.7,8 As such, a comprehensive understanding of the interaction between the nanostructured matters and small molecules, such as H2O at surfaces and interfaces via a dynamic hydrogen bonding, is essentially the key and at present the subject of great interest, but remains elusive. As one of the few semiconducting two-dimensional (2D) matters, polymeric carbon nitride (p-CN, also often denoted as g-C3N4) has unusual optoelectronic properties, thus already gaining great success in artificial photosynthesis.9–13 Along with the mainstream use of p-CN as a photocatalyst, nonetheless, an unusual but much less explored feature of p-CN is that its stacked structures are made of repetitive triangular heptazine units with hydrophilic edge groups. When tailored into nanoribbons,14 such molecular arrangements along the intraplanar and interplanar directions are anisotropic and very similar to that of natural bean pods (a biopolymer with gradient structure).15 Bean pods are well known for the hydration-triggered twist in seeding due to the inner-stress generated after losing water molecules (Figures 1a and 1b). These intriguing facts inspire us to hypothesize that p-CN holds great potential as a biomimetic mechanical transducer for water molecules, but this has scarcely been investigated.16 Figure 1. | (a) Scheme and photographs of anisotropic deformation of natural bean pod with gradient structure upon dehydration. (b) Proposed reversible anisotropic deformation of CNNRs with similar gradient hydrophilic/hydrophobic domains. (c) TEM image of CNNRs in a dry state [see the larger version and simulated structure in (d)]. AFM height images of CNNRs in (e) 40% RH and (f) 70% RH. Insets: scheme of coiled and partially recovered CNNR. Line profiles of single CNNR indicated by arrows 1 (g), 2 (h), and 3 (i) in (e) and (f). Download figure Download PowerPoint Here, we reported the first discovery of anisotropic deformation of p-CN nanoribbons (CNNRs) triggered by interacting with water molecules and forming a hydrogen-bonded adsorption layer. Moreover, after coupling with CNTs via π–π interaction, such delicate deformation could be simply and efficiently converted into an easy-to-measure resistance change. As a result, the CNNRs-based devices exhibited an ultrafast response of ca. 50 ms, high reproducibility, and selectivity and linearity almost in the full humidity range, outperforming the clinical pulmonary function testing station in a demonstration of real-time respiratory monitoring. This study would initiate a conceptually new application of p-CN in the form of anisotropic nanoribbons as a low-cost water molecule transducer for a diverse range of emerging fields with outstanding performances. Experimental Methods Chemicals and materials Dicyandiamide (DCDA, 99%) was purchased from Sigma-Aldrich (St. Louis, USA). Sodium hydroxide (NaOH, 96+%) was obtained from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Multiwalled CNTs (product no. S-MWNT-1020) were purchased from Shenzhen Nanotech Port Co., Ltd. (Shenzhen, China). The interdigitated Au and Pt electrodes (product nos. ED-IDE3-Au and ED-IDE3-Pt) were obtained from Micrux Technologies (Asturias, Spain). Ultrapure water (18.2 MΩ cm), obtained from a Smart2Pure 3 purification system (Thermo, Waltham, USA), was used in all experiments. Preparation of CNNRs and CNNRs–CNTs CNNRs were prepared according to our previous report. Briefly, the bulk p-CN was initially prepared by thermal condensation of DCDA at 550 °C and ground into fine powders by agate mortar before use. Then, the p-CN powder (500 mg) was added into NaOH solution (3 M, 20 mL), and the mixture was stirred at 60 °C for 12 h and treated under ultrasonication for 1 h every 4 h. After that, the CNNRs were obtained by removing excessive NaOH via dialysis in a membrane of molecular weight cutoff of 3500 Da (D306-50; Biodesign Inc., Shanghai, China) against the water until neutral. The CNNRs–CNTs were synthesized by adding CNTs powder (1 mg) to 0.5 mL CNNRs dispersion (8.9 mg/mL), and then the volume of the dispersion was adjusted to be the same (1 mL) using water. The mixtures were ultrasonicated for 15 min to obtain the homogeneous CNNRs–CNTs dispersion. Characterization Scanning electron microscopy (SEM) images were collected on a Zeiss Ultra Plus SEM (Potsdam, Germany). Transmission electron microscopy (TEM) images were obtained on a JEM-2100 microscope (JEOL, Tokyo, Japan). The circular dichroism experiment was measured by a Chirascan spectrometer (Applied Photophysics, Leatherhead, United Kingdom) on quartz. Raman spectra were measured with a DXR™-laser micro Raman spectrophotometer (Ex = 532 nm; Thermo). Elemental analysis (EA) was carried out on a Vario EL Cube analyzer (Elementar, Langenselbold, Germany). The X-ray diffraction (XRD) patterns were recorded on SmartLab and Ultima IV diffractometers (Rigaku, Tokyo, Japan) in sealed capillaries. The Brunauer–Emmett–Teller (BET) surface area was calculated from 77 K N2 adsorption–desorption isotherms acquired on a Nova 1000e instrument (Quantachrome, Boynton Beach, FL). All electrical measurements were carried out on a potentiostat (Gamry Reference Gamry Reference 600, Philadelphia, USA). Atomic force microscopy measurements CNNRs solution (10 μL) was deposited onto the freshly cleaved mica surface and air-dried for 10 min. Then the residues were removed, and the sample was dried before measurement. The experiments were performed under ambient conditions (Multimode SPM and Nanoscope V controller; Veeco Instruments, Bruker, Billerica, USA). The cantilever (ultrasharp silicon probe, OMCL-AC160TS-R3; Olympus, Tokyo, Japan) had a spring constant of 26 N m−1. For nanoscale material property mapping, the cantilever was calibrated by ramp and thermal tuning beforehand. We imaged samples in Veeco 8 quantitative nanomechanical mapping mode atomic force microscopy (AFM) under ambient conditions, which is named PeakForce quantitative nanomechanical mapping (PF-QNM, Bruker, USA) mode with precise force control. PeakForce Tapping® mode oscillates, but far below the cantilever resonant frequency, the vertical motion of the cantilever using the (main) Z piezo element and relies on peak force for feedback. Peak interaction force and nanoscale material properties were collected for each individual tap. AFM images were performed in tapping mode or PeakForce Tapping mode at a scan frequency of 1 Hz with optimized feedback parameters and 512 × 512-pixel resolution. The air humidity of 40% and 70% used for the hygroscopic properties test of CNNRs was controlled by a Parkoo dehumidifier (no. YDA-870EB; Parkoo, Guangzhou, China). The AFM images were collected from up to down at eight random places from three samples to confirm the consistency of the height. The average height of CNNRs was obtained by calculating all the nanoribbons in one image after subtracting the substrate height. The data were analyzed with NanoScope Analysis software (version 1.7, Bruker, Billerica, USA), and Gwyddion software (version 2.51, Czech Metrology Institute, Brno, Czech Republic) was also used for image processing. Fabrication of CNNRs–CNTs sensor- and humidity-sensing measurements Briefly, 5 μL of CNNRs–CNTs dispersion (CNNRs: 0.89 mg/mL, CNTs: 0.2 mg/mL) was deposited on clean Au or Pt interdigitated electrode by drop casting and natural drying to obtain a CNNRs–CNTs humidity sensor. The static humidity sensing measurement was carried out in a homemade sealed container (25 cm×18 cm×15 cm). To simulate an environment of different humidity, 200 μL of water was dropped on a 5 cm × 5 cm cellulose paper in the container. An electric fan was used to force air convection and accelerate volatilization of water to form a moist atmosphere. High purity dry air (99.999%, 79% N2 and 21% O2) was used to replace part of the moist air in the container to obtain different humidity. The real-time humidity of the atmosphere was detected by a commercial humidometer (Benetech, China). The current–voltage (I–V) plots were performed on the potentiostat. For response-/recovery-time measurement, the CNNRs–CNTs sensor was placed in a silicone pipeline of 8-mm diameter, with a dry airflow passing through (0.1 MPa). To form a humid air pulse, 1 mL of 100% relative humidity (RH) air was injected into the up-pipeline (ca. 5 cm) with a syringe. The CNNRs–CNTs sensor responded to the humidity, and the signal change was recorded by the potentiostat. For respiratory monitoring, the CNNRs–CNTs sensor was placed in front of the mouths or noses (ca. 5–7 cm away) of volunteers in a noncontact manner and open environment ( Supporting Information Movie S1). The CNNRs–CNTs sensor responded to the breathing, and the signal change was recorded by the potentiostat. Results and Discussion CNNRs were prepared by chemical tailoring of pristine bulk p-CN under an alkaline hydrolysis condition.14 As shown in the TEM image (Figure 1c), CNNRs had a uniform width of tens of nanometers and length of several micrometers, distinct to bulk p-CN that mostly consisted of particles several micrometers in size ( Supporting Information Figure S1). The morphology of CNNRs was further confirmed by AFM measurement ( Supporting Information Figure S2a), and the average thickness of CNNRs was measured to be 3.26 ± 0.30 nm according to the height profile of the AFM image ( Supporting Information Figure S2b), which corresponded to a few layers. More interestingly, the CNNRs had a twisted structure from the TEM observation (Figure 1d), which differed from the traditional two-dimensional planar structure. A similar phenomenon was also clearly observed in AFM image, shown as the height periodic fluctuations of the CNNRs, for example, axial periodicity of a single CNNR indicated by arrows in Supporting Information Figure S2a and depicted in Supporting Information Figure S2c. Besides, the circular dichroism spectrum of CNNRs exhibited stronger absorption intensity than bulk p-CN ( Supporting Information Figure S3), indicating higher chirality of CNNRs. The chemical structures and surface properties of CNNRs were further explored. Supporting Information Figure S4 demonstrated that the Fourier transform infrared (FT-IR) spectrum of CNNRs had an enhanced vibration peak nearby at 3200 cm−1, ascribed to −OH/NHx groups. It indicated that these new groups were decorated on the basic heptazine skeleton of p-CN, consistent with the combustion EA. Along this line, the surface hydrophilicity of CNNRs would alter in comparison with p-CN, which was further studied by contact angle measurement. In general, the existing –NHx and –OH terminals/defects made bulk p-CN hydrophilic. Nevertheless, a slight increase in the initial contact angle of CNNRs was also observed, which could be explained by the wetted contact owing to the nanorough surface of CNNRs ( Supporting Information Figure S5a). More interestingly, CNNRs exhibited complete absorption of a water droplet in <4 s, much faster than that of bulk p-CN (at least >30 s), indicating that water absorbability was substantially improved. The remarkably wrinkled surface of CNNRs after absorbing water was also observed, providing evidence of the strong deformation of CNNRs in the presence of water molecules ( Supporting Information Figure S5b). It was worth noting that the twisted structure of CNNRs was obtained in a dry atmosphere because the aforementioned TEM and AFM images were carried out under low humidity. Owing to the dedicated interaction with water molecules from terminal hydrophilic groups, the twisted CNNRs would be remitted after combining with water molecules in a high humidity environment. To demonstrate the deformation of twisted CNNRs upon adsorption and desorption of water molecules, the AFM measurements were carried out at different humidity. The AFM images in Figures 1e–1g and the height profiles in 1g show that the height of a single CNNR decreased from 3.4 to 1.5 nm, and the width of CNNR increased from 62.0 to 101.8 nm, when the environmental RH increased from 40% to 70%. According to the loss of periodic fluctuation in height from low humidity to high humidity (Figures 1h and 1i), such significant anisotropic swelling could be ascribed to the uncoiling effect after absorbing water molecules,17,18 similar to the structural transformation of bean pods between wet and dry conditions. To understand the interactions between CNNRs and water molecules, FT-IR spectra of CNNRs exposed to different environmental humidity were collected in the first set of experiments. Figure 2a revealed that the peaks between 3080 and 3460 cm−1 increased along with humidity rising from 6% RH to 86% RH, which can be attributed to the formation of a hydrogen-bonded H2O adsorption layer. This absorbance restored to the initial strength when the humidity was decreased to the initial dry state of 7% RH, indicating the reversibility of the adsorption/desorption of water molecules on the surface of CNNRs. From the enlarged FT-IR spectra between 1200 and 1700 cm−1 ( Supporting Information Figure S6), the influence of humidity on the heterocycles stretching mode of CNNRs was also proved reversible. The interaction of water molecules with CNNRs could also be proven by capillary XRD (Figures 2b and 2c). The interlayer (002) diffraction peak of bulk p-CN and CNNRs after absorbing water molecules was found to slightly shift to a smaller angle, indicative of expanded interlayer distance, but the latter had a more evident change. It suggested that after tailoring, the absorbing interaction between carbon nitride and water was strengthened and led to more massive swelling, presumably ascribable to the higher density of exposed edges along with more –NHx and –OH terminals/defects of CNNRs. Figure 2. | (a) FT-IR spectra of CNNRs under different environmental humidity. XRD 5 patterns of (b) pristine bulk p-CN and (c) CNNRs before and after humidification. The dashed lines show the (002) diffraction peaks. Insets: scheme (not in scale) of the expanded interlayer of p-CN and CNNRs after absorbing water molecules. AFM adhesion map of CNNRs in (d) 40% RH and (e) 70% RH. Representative adhesion force curve (up panel) and adhesion line profiles (bottom panel) of (f) CNNR indicated by arrow 1 in (d, e) at a different RH of 40% and 70%. Download figure Download PowerPoint To further verify the active interaction between water molecules and CNNRs, it is highly desired to characterize a single CNNR before and after absorbing water molecules. As a fundamental quantity, the adhesion force between the AFM tip and a single CNNR was measured. As shown in the AFM adhesion map (Figures 2d and 2e) and adhesion force curve (Figure 2f, upper panel), the maximum adhesion force of CNNR was up to approximately 20.30 nN in 70% RH, while that at the same position was only 8.53 nN when decreased to 40% RH. Line statistical data for the CNNR (Figure 2f, bottom panel) further confirmed this general trend. It was supposed that the AFM probe was identical in different humidity and had relatively good hydrophilicity, but the Van der Waals interaction between the AFM tip and the sample would be enlarged when the CNNRs absorbed water in high humidity. As a control, the adhesion force of the mica surface with excellent hydrophilicity was maintained almost the same (ca. 20.80 nN) in both 40% RH and 70% RH. Exploring by quantitative nanomechanical mapping, the change in the Young’s modulus provided quantitative evidence of the softening of CNNRs after adsorption of water.19,20 As shown in Supporting Information Figure S7, the modulus of CNNRs decreased from 1578 to 48 MPa when the RH increased from 40% to 70%. Note that the change of adhesion force and Young’s modulus of CNNRs upon humidity variation was compatible with some polymers, further indicative of their macromolecular nature.21 Therefore, the CNNRs efficiently and reversibly interacted with water molecules by forming of hydrogen-bonded adsorption layer and underwent an internal deformation under the different concentration of water molecules in the environment. As a polymer, such significant anisotropic swelling could be ascribed to the uncoiling effect via the release of residual stress after bonding to water molecules. Different from the mainstream of photocatalysis study, it should be noted that such polymeric nature of carbon nitride has been rarely reported so far,17,18 but has prospective applications in responsive smart devices, as discussed in the following context. Nevertheless, the readout of this deformation on the nanoscale using AFM is slow and inconvenient for practical water molecules-transducing applications. In general, electrical signals are widely used in sensing systems because they can be quickly recorded and processed in integrated circuits for application in most electronic devices. Since the conductivity of p-CN and its derivatives is inadequate ( Supporting Information Figure S8), CNTs were further used as a reinforcement agent for CNNRs. CNTs sensed humidity poorly when used alone, but CNTs were supposed to boost signal readout, improve water mass transfer, and optimize the sensing interface in devices due to their high conductivity, nanostructure (approximately 20 nm in diameter and micrometers in length; Supporting Information Figures S9a and S9b), and high surface area (144.3 m2/g; Supporting Information Figure S10a). Moreover, it should be noted that the nanoribbon structure also led to a much higher slit-type porosity (194.7 m2/g) than pristine bulk one (8.6 m2/g; Supporting Information Figure S10a). The higher porosity would be helpful for efficient transport of water molecules to adsorption sites (–NHx and –OH terminals) throughout the whole p-CN network and quickly reaching equilibrium, and vice versa for the desorption processes. Moreover, due to the similarity in conjugated framework, CNTs were ready to be coupled with CNNRs by effective noncovalent π–π interaction ( Supporting Information Figures S9c, S9d, a10b). As shown, CNTs could be well dispersed in water by cooperation with the hydrophilic CNNRs by a simple one-step ultrasonic agitation process ( Supporting Information Figure S11). Because pristine CNTs were difficult to disperse in aqueous solution, and water is envisioned as the most environmentally friendly solvent, the successful synthesis of homogeneous CNNRs–CNTs in aqueous solution was highly expected for device fabrication and maximization of the unique properties of CNNRs in humidity sensing. Because CNNRs deformed in the presence of water molecules, we envisioned that the interconnection of CNTs in the as-prepared CNNRs–CNTs network would be influenced, and subsequently conductivity would be altered (Figure 3a). Based on this principle, to investigate sensing performance, the CNNRs–CNTs were deposited on an interdigital Au electrode for conductivity measurements via a two-electrode system (see more details in Supporting Information). I–V curves (Figure 3b) were first measured under different RH atmospheres. The linear I–V curves passed through the coordinate origin, following Ohm’s law and the typical conductor behavior of CNNRs–CNTs. With a gradual increase in the RH, the decrease in the slope of the I-V curves was observed. It was supposed that upon high humidity the anisotropic deformation of CNNRs occurred, the CNTs were more covered by the low conductive CNNRs, and the interconnection among CNTs would decrease, thus increasing the resistance of the composite (Figure 3a). The sensing signals showed high linearity in a wide range of humidity (Figure 3c), which shows the attractive possibility for realistic humidity sensing. To eliminate the influence of contact resistance between the CNNRs–CNTs network and the substrate, a Pt interdigital electrode was used as a control. Compared with those on the Au electrode, the CNNRs–CNTs on the Pt electrode performed similarly over the entire humidity range (Figure 3c), indicating that the conductivity change originated exclusively from the CNNRs–CNTs rather than from the interface between the CNNRs–CNTs and the metal electrodes. Figure 3. | (a) Scheme of CNNRs–CNTs deformation by the absorption/desorption of water molecules. (b) I–V curves of CNNRs–CNTs in different RH atmospheres. (c) Relative resistance of CNTs and CNNRs–CNTs on the different interdigital electrodes in RH ranging from 5% to 97%. (d) Response and recovery time under a pulsed flow of humid air. (e) Time-resolved dynamic tests of the sensing signal under different humidity changing between 5% RH and 55% RH for reproducibility test. (f) The normalized current response of the CNNRs–CNTs sensor (inset: I–T curves) in different gases and solvent vapors showing high selectivity. Download figure Download PowerPoint The development of a humidity sensor with a fast response is critical for real-time monitoring of a rapidly fluctuating environment. The response time in this study is defined as the time for the output signal to reach 90% of the final amplitude.22 The proposed humidity sensor exhibited a response time as low as approximately 50 ms (see the high current to low current, according to the low to high humidity in Figure 3d inset) when humid air was pulsed through dry air. Such a superfast response time was competitive with that of the state-of-the-art humidity sensor ( Supporting Information Table S1).23–34 Nevertheless, it should be noted that the exact value of the response time here does not fully represent the real sensing performances, as different evaluation methods were used in different studies in previous reports. To this end, as discussed in the following context, we compared the sensitivity of CNNRs-based devices with the commercial testing station. Notably, the full recovery time (Figure 3d) reached 1 s, which was supposed to be longer than the real recovery time, as the humid air pulse passing through the sensor in the testing pipeline also took a specific time (see Supporting Information for more details). The fast response and recovery times for the CNNRs–CNTs sensor could be ascribed to the following factors: (1) its rapid absorption of water by the nanoribbon structure, (2) the balanced affini