DNA‐Incorporated Biomimetic Olfactory Neuroepithelium That Facilitates Artificial Intelligence

Designing biomimetic olfactory neuroepithelium (BONe) with subnanosized active domains as artificial olfactory receptors (ORs) is highly desirable to sense various colorless and odorless hazardous odorants which find no appropriate ORs in the human olfactory neuroepithelium (ONe), yet challenging because of the unsuitability of biomolecules for a design that requires effective electronic features and stability. Herein, a DNA‐incorporated 3D BONe is introduced, where DNA facilitates optimal tuning of d‐band center, and in situ anchoring of PdO2 subnanoscale clusters (PdO2‐sNCs) on the exfoliated single‐layer reduced graphene oxide (SL rGO), to mimic wrinkled morphology of natural ONe. Unprecedentedly, BONe demonstrates benchmarked H2‐sensing performance (small recovery time of ≈30 s with a limit of detection of 50 ppb) at room temperature with yearlong durability, satisfying prerequisites of safe adoption of H2 clean energy. The great recovery is innovatively illustrated by the downshift of d‐band center of PdO2‐sNCs and strong electron transport of SL‐rGO network. An adsorption/desorption model is proposed to clarify the sensing mechanism. BONe design may eventually be integrated with artificial intelligent electronics for ppb‐level sensing of harmful gases to ensure accident prevention in modern public and military environments.

DOI: 10.1002/aisy.202200396 Designing biomimetic olfactory neuroepithelium (BONe) with subnanosized active domains as artificial olfactory receptors (ORs) is highly desirable to sense various colorless and odorless hazardous odorants which find no appropriate ORs in the human olfactory neuroepithelium (ONe), yet challenging because of the unsuitability of biomolecules for a design that requires effective electronic features and stability. Herein, a DNA-incorporated 3D BONe is introduced, where DNA facilitates optimal tuning of d-band center, and in situ anchoring of PdO 2 subnanoscale clusters (PdO 2 -sNCs) on the exfoliated single-layer reduced graphene oxide (SL rGO), to mimic wrinkled morphology of natural ONe. Unprecedentedly, BONe demonstrates benchmarked H 2 -sensing performance (small recovery time of %30 s with a limit of detection of 50 ppb) at room temperature with yearlong durability, satisfying prerequisites of safe adoption of H 2 clean energy. The great recovery is innovatively illustrated by the downshift of d-band center of PdO 2 -sNCs and strong electron transport of SL-rGO network. An adsorption/desorption model is proposed to clarify the sensing mechanism. BONe design may eventually be integrated with artificial intelligent electronics for ppb-level sensing of harmful gases to ensure accident prevention in modern public and military environments.
tailoring a biomolecule-incorporated biomimetic sensor-design via tuning the d-band center has not been reported. Therefore, realizing a DNA-incorporated biomimetic design of optimally tuned d-band center to modulate the adsorption/desorption features via regulating the reaction barriers and kinetics is of paramount interest.
Herein, a novel approach to design DNA-incorporated BONe via a scalable route without relying on sophisticated technologies is introduced. In this design, salmon sperm DNA is applied to optimally tune the d-band center of Pd oxides and in situ homogeneous dispersion of PdO 2 subnanoscale clusters (PdO 2 -sNCs) on the exfoliated ultrathin single-layer reduced graphene oxide (SL-rGO). Practically, this unprecedented BONe design not only outperforms natural counterparts but also all existing synthetic systems for room-temperature H 2 sensing, exhibiting complete recovery in short time (%30 s) with a small limit of detection (50 ppb). In this BONe design, DNA incorporation is killing two birds with one stone, providing robust design and benchmark performance with multipronged benefits such as roomtemperature ppb-level H 2 sensing, quick and complete recovery, and excellent stability.
The schematic diagram ( Figure 1) reveals the strategy to design 3D active BONe. In this 3D BONe design, single-stranded DNA (ssDNA) incorporation achieves in situ homogenous anchoring of PdO 2 -sNCs on highly exfoliated ultrathin SL-rGO owing to the unique arrangement of DNA base pairs and -PO 4 À groups in its backbone, [9a,11] and then lyophilization makes wrinkles in the design. Subsequently, large surface area with abundant wrinkles and plenteous subnanosized active domains, sNCs as artificial ORs, make the 3D BONe (Figure 1c,d) a perfect mimic of the natural counterpart (Figure 1b), to demonstrate benchmark sensing characteristics.

Results and Discussion
Analytical investigations were performed on as prepared BONe-240 and various control samples (BONe-DNA, BONe-AAS, BONe-120, BONe-360, BONe-480, DG-240, and Pd-rGO) to unveil surface, structural, compositional, geometrical, and vibrational features of BONe-240. The scanning electron microscopic (SEM) images of BONe-240 exhibit exfoliated ultrathin SL-rGO (Figure 2a,b) with wrinkled textures (Figure 2c), mimicking the surface roughness of human ONe (Figure 1b), which preserves heterogeneity in the surface morphology owing to the patchy distribution of the characteristic dense mat of olfactory cilia. [12] Evidently, SEM confirms the absence of aggregation of sNCs and agglomeration of SL-rGO at different scales of magnification (Figure 2a-c), thus, showing homogenous integration of components in the biomimetic design. Visibly, SEM analysis collected no evidence for sNCs owing to their size in subnanoscale range. However, SEM coupled with energydispersive X-ray spectroscopy (SEM-EDX) identifies homogeneous dispersion of Pd and other elements ( Figure S1, Supporting Information). Furthermore, transmission electron microscopic (TEM) images ( Figure 2d) exhibit wrinkle-induced surface roughness of ultrathin nanosheets.
Alike SEM examination, TEM exhibits no proof of agglomeration in the design of BONe-240. For further insights, spherical aberration-corrected scanning TEM (Cs-STEM) examination (Figure 2e,f ) was performed at atomic-scale resolution. Therefore, an edge of BONe-240 was purposely selected for high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) investigations (Figure 2e), where Z-contrast mapping undoubtedly verified the extremely homogeneous distribution of Pd atoms of PdO 2 -sNCs ( Figure 2f ). Moreover, to assess surface cracks or deterioration, crest of a wrinkle was chosen knowingly ( Figure 2g); fortuitously, the chosen part exhibited no damage in the surface. To further evaluate the surface homogeneity, STEM-EDX linescan was performed to assess elemental distributions at a nanometer area, marked with red in Figure 2g. The corresponding linescan spectra (Figure 2h) represent uniform distribution of constituent elements. Likewise, to authenticate the incorporation of ssDNA in BONe design, HAADF-STEM-EDX mapping test was conducted at a relatively larger section of BONe-240 to achieve some patterns corresponding to ssDNA and sNCs (Figure 2i,j). As expected, HAADF-STEM-EDS mapping shows all constituent elements in a typical pattern, where every element follows the same typical pattern, thus  verifying the presence of ssDNA and subsequent homogenous anchoring of sNCs. To determine the significance of DNA, different samples were prepared for surface analyses. SEM, TEM, and Cs-STEM examinations ( Figure S2i-t, Supporting Information) revealed that increasing concentration of DNA leads to agglomerated particles due to enhanced number of nucleobases and subsequently, low electrostatic repulsion between the clusters, [13] which severely hamper surface homogeneity of sNCs. Structural features of the prepared composite samples were studied by X-ray diffraction (XRD). As displayed in Figure 2d, the XRD peak of GO at 2θ = 10.4°is attributed to oxygencontaining functional groups on the surface of graphene sheets, whereas DNA reveals no sharp peaks. [14] Compared to BONe-DNA, BONe-AAS has high-intensity peaks at 2θ values of 40.1°, 46.9°, 68.2°, and 82.2°corresponding to the facecentered-cubic Pd (JCPD No. 46-1043). [15] Conspicuously, there is a positive shift in the diffraction pattern of DG-240 along with an obvious decrease in the intensity of a broad peak at 2θ = %21.25°, which can be attributed to the reduction of GO and the appropriate π-π interactions between exfoliated SL-rGO and DNA, respectively. BONe-240 follows the similar tend with an additional positive shift and decline in its intensity due to the anchoring of sNCs and subsequently more surface modifications. As evidence, a similar trend can be observed in Unlike both BONe-DNA and BONe-AAS, BONe-240 reveals no observable diffraction peak, including that of Pd 0 , confirming that DNA incorporation prohibits the reduction of metal cations and facilitates the formations of PdO 2 -sNCs, where electrostatic repulsions between PdO 2 -sNCs and -PO 4À groups of DNA backbone prevent the aggregation of sNCs. [9a,11a,16] Therefore, they offer no diffraction to the incident X-Ray beam. [14] Generally, a positive shift in the binding energy of X-ray photoelectron spectroscopy (XPS) is attributed to d-band downshift away from the E F of metals. The downshift of the d band center brings more antibonding states below the E F , which makes adsorbate-metal interaction Pauli repulsive, thereby weakening the bond strength of adsorbates. [17] XPS demonstrates the oxidation states which are usually correlated with the changes in the d-band center of metal. Figure 3b shows that the Pd 3d XPS signal of the Pd-rGO is fit to two pairs of doubles: Pd 3d 5/2 (335.7 eV), Pd 3d 3/2 (337.1 eV) and Pd 3d 5/2 (340.9 eV), Pd 3d 3/2 (342.2 eV), which can be related to Pd 0 and surface Pd 2þ linked to oxygen atoms, respectively. [18] Interestingly, XPS spectrum of BONe-240 reveals a positive shift in the binding energy and shows two prominent peaks, which on deconvolution represent four peaks. Based on the binding energy values, two peaks Pd 3d 5/2 (337.4 eV) and Pd 3d 3/2 (342.5 eV) can be related to the Pd 2þ of PdO, whereas other two peaks Pd 3d 5/2 (338.4 eV) and Pd 3d 3/2 (343.7 eV) can be unambiguously attributed to Pd 4þ of PdO 2 . [19] XPS study ( Figure 3b; and S3d, Supporting Information) gives information that the BONe design excludes Pd 0 and keeps Pd 4þ as a dominant species in sNCs. Thus, XPS analysis shows that BONe design provides effective downshift of d-band center which can, in turn, weaken the interaction for adsorbates. Moreover, Raman and Fourier transform infrared (FTIR) spectroscopic techniques were applied to further elucidate the structural and vibrational properties. Figure 3c presents comparative analysis of Raman spectra, where GO indicates small I D /I G ratio of 1.00 corresponding to D (1350 cm À1 ) and G bands (1598 cm À1 ). Raman spectrum of BONe-AAS locates D band at 1347 cm À1 while G band (1574 cm À1 ) exhibits a negative shift in the wavelength with the emergence of two additional bands D* (%1190) and D** (%1500 cm À1 ), which can be ascribed to sp 3 carbons and C-H vibrations of hydrogenated carbons, respectively. [20] BONe-AAS shows low I D /I G ratio of 1.06, which confirms inappropriate reduction owing to the attachment of bulky particles on the surface ( Figure S2a-d, Supporting Information) and subsequently, insufficient π-conjugation of the graphene-like layers. [21] In contrast, D (1347.75 cm À1 ) and G (1590.01 cm À1 ) bands of BONe-DNA attain high I D /I G ratio (1.20) corresponding to unfavorably high reduction of GO. [22] However, unlike BONe-AAS, BONe-DNA maintains relatively small particles ( Figure S2e-h, Supporting Information) and ascribes exceptional features of DNA to create electrostatic repulsion among the particles to prohibit their aggregation. [9a] Raman spectrum of DG-240 exhibits D band www.advancedsciencenews.com www.advintellsyst.com at 1346 and G band at 1584 cm À1 to provide an I D /I G ratio of 1.13, which is higher than that of GO. Interestingly, the D band shift from 1350 to 1346 cm À1 confirms πÀπ interaction of DNA with graphene and appropriate reduction of GO. Both BONe-AAS and BONe-DNA remained unsuccessful to achieve homogeneous anchoring of sNCs on the exfoliated ultrathin SL-rGO, which is realized in BONe-240 owing to the effective linkage of -PO 4À groups of DNA backbone with -NH 2 of sodium 2-((2-aminoethyl)amino)ethanesulfonate (AAS). [23] Subsequently DNA bases prohibit aggregation and develop fine coordination with sNCs. [9a] Therefore, BONe-240 exhibits I D /I G value of 1.16, corresponding to D (1342 cm À1 ) and G (1583 cm À1 ) bands, higher than that of DG-240. FTIR analysis offers further information on different types of binding sites, especially for sNCs-ssDNA interaction (Figure 3d). The peak at 1696 cm À1 which is associated with nucleobases of ssDNA undergoes redshift and appears at 1566 cm À1 in DG-240 and 1578 cm À1 in BONe-240, can be ascribed to the ssDNA-rGO interactions, and consequently decreased distance between ssDNA and rGO. [24] Moreover, the asymmetric stretch and symmetric stretch of PO 4 À group of DNA backbone which appeared at 1232 and 1066 cm À1 , respectively, are shifted to 1217 and 1078 cm À1 in DG-240 and 1209 and 1056 cm À1 in BONe-240. This redshift can be attributed to an effective linkage between PO 4À of ssDNA and NH 2 group of AAS. [23] Importantly, the peak at 1578 cm À1 for BONe-240, unlike DG-240, appeared after a blueshift of 12 cm À1 , confirming the coordination of sNCs with the N-atoms of DNA, rather than to phosphate. [24] Thermogravimetric analysis (TGA) was performed under N 2 atmosphere to obtain decomposition curves at elevated temperatures ( Figure 3e) and to estimate the thermal stability and the composition of BONe design. GO shows lowest thermal stability with 98.4% weight loss owing to the removal of oxygen functionalities and the burning of ring carbon. [25] Unlike GO, DNA shows much higher thermal stability (total weight loss of 74.7 wt%). Compared to DG-240, BONe-240 has enhanced thermal stability and therefore, the weight remaining of DG-240 (34%) was increased to 56% BONe-240, thus indicating fine anchoring of PdO 2 -sNCs. [26] The higher thermal stability and lower weight loss of 18.17 wt% of Pd-rGO sample is associated with the bulky aggregates of Pd with rGO. To assess the number of Pd atoms in the PdO 2 -sNCs, matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS) analysis was performed for both BONe-240 and DG-240 in positive ionization mode. BONe-240 shows strongest signal at 665.62 m/z which corresponds to sNCs containing total number of Pd atoms of %6.2 (Figure 3f ). In addition, BONe-240 also shows a relatively weak signal at 603.98 m/z ( Figure S4a, Supporting Information) corresponding to sNCs containing Pd atoms equal to %5.6. Hence, it is evident that BONe-240 contains a variety of sNCs where the number of Pd atoms vary from %5.6 to %6.2.
The gas-sensing characteristics of the resultant chemiresistive sensor devices were evaluated by monitoring the variation of current at a fixed bias of 5 V upon exposure to H 2 , NH 3 , NO 2 , CO, or ethanol. The response (%) is defined as the relative current response ðResponse ð%Þ ¼ jI 0 ÀI g j I 0 Â100%Þ, where I 0 and I g are the balance currents of the sensor before and after exposure to analyte gases, respectively. The I-V characteristics of all sensor devices (BONe-120, 240, 360, and 480, and DG-240) indicate ohmic contact between the sensing layer and electrodes ( Figure S7a-e, Supporting Information). BONe-240 exhibits the best response (%15% to 5 ppm H 2 ) at room temperature, with the response/recovery time of 20/30 s (Figure 4a). Meanwhile, BONe-360 shows a relatively weak H 2 response (%13% to 5 ppm H 2 ) with slightly longer response/recovery time (%25/35 s) compared to BNOe-240, indicating a wide window for the preparation of BONe systems with great H 2 sensing. However, the H 2 response is dramatically lowered (%5% to 5 ppm H 2 ) with further increase in DNA content (Figure 4a: and S8a-d, Supporting Information) and leads to aggregation of PdO 2 -sNCs (evident from Figure S2i-t, Supporting Information) which undoubtedly reduces the number of active sites. [13] Besides, BONe-120 exhibits very small H 2 responses of %2.2% to 5 ppm H 2 , due to less PdO 2 -sNCs resulted by less DNA. Notably, all the BONe design (BONe-120, 240, 360, and 480) exhibit a prominent decrease in the current during exposure to H 2 gas, suggesting p-type behavior of BONe design ( Figure 4a: and S8a-c, Supporting Information). Figure 4b shows positive response values of BONe-240 versus H 2 concentration in the range of 0.05-5 ppm, which is fit with Langmuir isotherm for molecule adsorption on a surface using equatio: Response (%) = 11.435C 1þ0.58C (where C is the concentration in ppm). Similar Langmuir isotherm fitting is performed for BONe-360 ( Figure S8f, Supporting Information). The fitting further confirms the charge transfer when the BONe systems respond to H 2 gas. [27] What's more, slight responses to 50 ppm NH 3 , NO 2, CO ethanol confirm the desired superior H 2 selectivity of BONe-240 and BONe-360 (Figure 4c and S8d, Supporting Information). In contrast, for the DG-240 sample excluding PdO 2 -sNCs, no obvious H 2 response was observed ( Figure S8d, Supporting Information). Importantly, the selective H 2 response of BONe is associated with the inherent ability of PdO 2 , which is explained in the following section. Negligible change in the repeated cyclic response to 5 ppm H 2 after 1 month and 1 year aging for BONe-240 ( Figure 4d) and BONe-360 ( Figure S8e, Supporting Information) confirms excellent repeatability and unprecedent long-term stability of the BONe design. Unlike complete recovery in short time of BONe with PdO 2 -NCs, the Pd-rGO failed to demonstrate any level of recovery ( Figure S9, Supporting Information) mainly owing to big particles of Pd 0 as it is clearly shown in the SEM and TEM images ( Figure S2u-x, Supporting Information) and, subsequently, their unfavorable behavior to adsorb H 2 very strongly. [17] Compared with the state-of-the-art H 2 sensors operating at room temperature (Table S1, Supporting Information), the prepared BONe exhibits superior H 2 -sensing performances in terms of complete recovery in short time (%30 s) and experimental limit of detection (50 ppb) which is much lower than that of other systems (hundreds of ppb).
In order to clarify the H 2 -sensing performance of BONe, we fit the adsorption and desorption processes of H 2 molecules on BONe. Taking 1 ppm response as an example, the adsorption curves are well fit with both firstand second-order kinetics with same fitting error (Figure 4e; and S8f, Supporting Information), based on Equation (1) and (2) as given. [28] First-order exponential Second-order exponential where I is the instantaneous current, I f is the final current value at the equilibrium state after adsorption and desorption process, t is the time, A is the amplitude, and τ is the time constant; subscripts of f and s represent the parameters for fast and slow components, respectively. As shown in Figure 4f and S8f, Supporting Information, the desorption curves are better fit with second-order kinetics with a smaller fitting error than that of fitting with first-order kinetics. This suggests that there are two mechanisms associated with H 2 molecules adsorption/ desorption in BONe. [28] Density functional theory (DFT) calculations were performed to investigate the origin of high performance, quick response, and complete recovery of BONe-design. DFT study shows a downshift of d-band center of PdO 2 -sNCs, attributed weak adsorption of H 2 on PdO 2 -sNCs, and it is innovatively proposed to clarify H 2 sensing. [29] It is a fact that increasing concentration of a more electronegative atom (e.g., O) in a system can predominantly capture electrons from the neighboring atom (e.g., Pd) with relatively lower electronegativities. [10] Thus, the lower electronegative atoms become electron deficient and reveal a downshift of d-band centers. [29] Here, DFT calculations demonstrate the probable H 2 -sensing mechanism, where PdO 2 shows a d-band downshift of À2.58 eV (Figure 5c), which is much better than that of PdO (À2.16 eV) ( Figure 5b) and Pd (À0.179 eV) (Figure 5a). The high electronegativity difference between O (3.44) and Pd (2.20) regulates electron transfer from Pd atoms to nearest O atoms [10] and thus, electron-deficient Pd causes significant downshift of d-band center of PdO 2 . [29b] It is well known that the adsorption energy of adsorbates depends on the occupancy of the bonding and antibonding states formed between adsorbates and metals. When a downshift occurs in the position of d-band center, then more antibonding states move below the E F , turns adsorbate-metal linkage highly repulsive, and thereby makes chemisorption very weak. [17] To calculate the adsorption strength of H intermediates, dominant phases of PdO 2 , PdO, and Pd were selected for calculations (shown in Figure 5d-f ), where relatively less negative adsorption energy of PdO 2 (E ads = À0.2772 V) compared to PdO (E ads = À0.690) and Pd (E ads = À0.896 eV) shows weak interactions with adsorbates, thus optimally tuning the reaction kinetics to achieve high sensing performance, especially complete and fast recovery.
In order to further explicate the specific adsorption and desorption processes of H 2 on BONe, an adsorption/desorption model is proposed to clarify the sensing mechanism. [3b,30] Unlike natural olfactory system, which lacks appropriate ORs to sense H 2 gas, the artificial ORs of the BONe system achieve effective interactions with H 2 (Figure 6) to demonstrate benchmarked ppb-level sensing performance with superior selectivity and yearlong durability (Figure 4c,d; and S8d-e, Supporting Information), where PdO 2 -sNCs anchored along DNA strands   (Equation (3)), and then, physisorbed O 2 with high electron affinity of 0.43 eV extracts electrons from PdO 2 and transforms to chemisorbed oxygen ions (O 2 À, Equation (4)). [31] Additionally, PdO 2 is well known to be a strong electron acceptor.
[4a] In the BONe system, electrons transfer from DNA and p-type rGO to PdO 2 due to the heterojunctions between their interfaces. Therefore, more oxygen ions (O 2 À ) chemisorb on the BONe surface. Decreasing concentration of electrons in p-type PdO 2 leads to an increased electrical conductivity owing to the development of hole accumulation layer (HAL) [32] (Figure S10, Supporting Information). As a result, the conductivity of BONe system is enhanced, which is demonstrated by the I-V results in Figure S7, Supporting Information.
When H 2 is exposed to BONe, H 2 molecule will be dissociated into H atom on the surface of PdO 2 catalyst (Equation (5) and (6)). Then, the H atom will spill over onto PdO 2 and diffuse along the graphene surface. The dissociation of H 2 molecule into H atom is an exothermic chemisorption process, which facilitates the subsequent spillover and surface diffusion. [33] The diffused H atoms will react with chemisorbed O 2 À and form desorbed H 2 O (Equation (7)), releasing electrons back to the transport network of rGO via PdO 2 , which narrows the HAL of PdO 2 (Figure S10, Supporting Information). As a result, the electrical conductivity of BONe system is reduced. [4a] On the other hand, the H atom could weaken the bond between O ion and Pd ion in PdO 2 , generating localized H 2 O molecule along with oxygen vacancy forming. Similar performance has been demonstrated. [30,34] The oxygen vacancy will behave as a donor and decrease the electrical conductivity of the material. Possibly the rate-determining steps 1 (Equation (7)) and 2 (Equation (8)) almost go halves on the adsorption process, so, the adsorption curves show are fit with both firstand second-order kinetics with same fitting error. When H 2 exposure is stopped, air is let in, density of the surface oxygen increases. Thus, chemical adsorbed oxygen is formed with trapped electrons (Equation (4)), increasing carrier density of the p-type PdO 2 . Meanwhile, oxygen will occupy the vacancy and capture electrons, reforming the bond with Pd ions. Consequently, the electrical conductivity of the material is completely and quickly increased. For desorption, the chemisorption process of oxygen is much stronger than that of replacement of oxygen vacancy. So, the desorption curves are fit better with second-order kinetics.
In BONe-design, DNA facilitates homogeneous anchoring of PdO 2 -sNCs on exfoliated ultrathin SL-rGO and prohibits the aggregation of PdO 2 -sNCs and agglomerations of SL-rGO that remained a serious problem in the existing systems. [11a] The robustness of H 2 -sensitive BONe-design ascribes synergy that emerged from the variety of DNA's interactions including πÀπ interaction with graphene, and electrostatic interactions with AAS and PdO 2 -sNCs, which strictly hinder the reduction of sNCs and degradation of DNA. [11a,35]

Conclusion
In conclusion, artificial ORs-like H 2 sensors are prepared based on DNA-incorporated 3D BONe design. The BONe design mimics human ONe, wherein PdO 2 -sNCs perform as artificial ORs and SL-rGO as carrier-transporting networks to transfer electrical signals to test terminals. Interestingly, DNA incorporation enables separation of graphene layers, and subnano restriction of PdO 2 -sNCs to optimally tune its d-band center, and thus the obtained BONe design demonstrates benchmarked ppb-level H 2 sensing with excellent selectivity (especially a short recovery time of %30 s, low limit of detection of 50 ppb, and unprecedented yearlong durability). The modulation of d-band center of PdO 2 -sNCs is innovatively proposed to understand the improvement of H 2 -sensing performance. The ultrathin graphene layers network domains the transmission of electrical signals, owing to the ultrafast electronic transport. Further advancement will be dedicated to realize a BONe design with a variety of artificial ORs to achieve pattern recognition with improved sensing performance, selectivity, and durability. Combining the versatility of sensing performances of the robust DNA-incorporated BONe design with a variety of artificial ORs, it is predicted that the integration of BONe design with AI electronics will be readily adaptable for the ppb-level sensing of various biochemicals, harmful, and explosive gases in many industrial processes such as H 2 production, petroleum transformation, and most prominently, lethal chemical warfare agents of the modern era.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.