Theoretical Study of CO Adsorption Interactions with Cr-Doped Tungsten Oxide/Graphene Composites for Gas Sensor Application.

The present study explores the CO adsorption properties with graphene, tungsten oxide/graphene composite, and Cr-doped tungsten oxide/graphene composite using density functional theory (DFT) calculations. The results of the study reveal the Cr-doped tungsten oxide/graphene composites, g-CrW n-1O3n (n = 2 to 4), to have a lowered highest occupied molecular orbital-lowest unoccupied molecular orbital (HOMO-LUMO) energy gap, high surface reactivity, and a strong cluster-graphene binding energy, hence exhibiting a strong adsorption interaction with CO. The CO adsorption interaction shows physisorption properties by having a greater tendency for Mulliken and natural bond orbital (NBO) charge transfer supported by a strong physisorption interaction toward the g-CrW n-1O3n (n = 2 to 4) composite with HOMO-LUMO energy gaps of -0.638, -0.486, and -0.327 eV, respectively. The calculated photoelectron spectroscopy (PES) and infrared spectra combined with the visualized electrostatic potential and contour line confirm the population density of the physisorption interaction. The calculated results show that the g-CrW n-1O3n composite achieves a greater sensing ability by possessing the highest sensitivity, adsorption, and desorption characteristics for n = 2 (g-CrWO6 composite). In conclusion, Cr-doped tungsten oxide/graphene has high sensitivity toward CO gas.

Introduction

Rapid development of industry and technology has become a concern for the increasing growth of air pollution and toxic gas. Detection of toxic gas is extremely important to maintain and ensure a safe and healthy environment.[1,2] Therefore, the development of gas sensors with high sensitivity and selectivity and fast response and recovery time is in great demand. Over past decades, graphene-based composites have attracted interest because they have been promoted and shown to have enhanced properties, and they are applicable in many fields, including the field of gas sensors.[3−5] Graphene is regarded as an important support layer for conjugated material because of its outstanding physical and chemical properties.[6,7] Since graphene-based sensors are restricted by low response signals, long recovery time, and poor selectivity for sensing application, a number of studies have explored conjugating graphene with metal, conducting polymer, and especially with metal oxide nanoparticles because of the synergistic effect of their structural and chemical compositions.[8−12] In the context of gas sensing applications, the ease in modifying the metal oxide structure helps to achieve a high surface-to-volume ratio, a parameter required for improving the chemical reactivity.[13−15]

The use of tungsten oxide (WO3) in the graphene–metal oxide composite is popular because WO3 has unique physical and chemical properties attributed to its various morphologies.[16,17] In recent studies, the hydrothermal method has been shown to be successful in synthesizing the graphene/WO3 nanocomposite.[18−23] Gui et al.[20] have shown the sensor properties of hemispherical WO3/graphene toward triethylamine. They concluded that the hollow structure of WO3/graphene shows a significant sensing performance for amine gases at room temperature, especially for triethylamine. Srivastava et al.[22] have shown graphene/WO3 nanocomposites to have a fast sensing response toward NO2 gas. Furthermore, Adhyapak et al.[23] have observed a linear increase in sensitivity in NO sensing because of NO affinity toward the nanostructured WO3/graphene composites.

However, to the best of our knowledge, the gas sensing properties of WO3/graphene and Cr-doped WO3/graphene toward CO gas molecules have not yet been studied. In our previous study, we reported the chemical, electronic, and energy properties and physisorption effect of CO, H2S, and H2 gas toward pure tungsten oxide (WO3) (n = 2 to 4) and its chromium-doped (CrWO3) (n = 2 to 4) nanoclusters by density functional theory (DFT) analysis.[24] It is reported that the sensing ability decreases as the cluster and dopant density increases, and these nanoclusters have high selectivity toward CO molecules. Martínez et al.[25] investigated different graphene model approaches for utilization of aromatic molecules (finite system) and periodic systems (supercell) for sensor applications. For instance, Saidi[26] investigated the stability and the interaction between MX2 (M = Mo or W and X = S or Se) layered on a graphene supercell. They found that there is van der Waals epitaxy of MX2 on undoped graphene. In modeling van der Waals (dispersion) interaction, using the vdW-TS scheme, Grimme’s DFT-D2, and DFT-D3 are the common approaches.[27,28] Since in the finite system of graphene, coronene molecules were considered as prototypes for the graphene model,[29−34] Hughes et al.[35] assessed the dispersion interaction with and without Grimme’s D3 corrected DFT for the interaction between the graphene surface and H2, NO2, H2O, and Ar. However, they concluded that there is an undistinguishable result between the functionals and demonstrate the problem with empirically corrected DFT as the large size of the graphene system. Jadoon et al.[36] have demonstrated that the interaction energies of Ag6 composited on coronene and circum-coronene for nitroaniline are quite similar.

Hence, in this present work, the coronene molecule is used as our graphene model. We have investigated the physisorption interaction of CO with graphene, graphene/tungsten oxide (g-WO3 (n = 2 to 4)), and graphene/Cr-doped tungsten oxide (g-CrWO3 (n = 2 to 4)) composites based on DFT analysis using our previous DFT configuration. The study consists of a systematic theoretical investigation of optimized ground-state structures; the highest occupied molecular orbital–lowest unoccupied molecular orbital (HOMO–LUMO) energy gap; chemical hardness; chemical potential; electrophilicity index; and binding energy of graphene, graphene/WO3, and graphene/CrWO3 composites. Furthermore, in this study, we have calculated the adsorption energy, electron transfer, photoelectron spectroscopy (PES), and infrared spectra as evidence to support the sensor performance properties of the composites. We believe that our investigation provides insight into the fundamental characteristics of sensor material by the DFT simulation method.

Results and Discussion

Geometric Structure and Electronic Behavior of Graphene, g-WO3, and g-CrWO3 Composites

Graphene was initially optimized before it was used to form the composite with the WO3 and CrWO3 (n = 2 to 4) clusters. The composite formation was carried out by attaching graphene C atom with either W of WO3 or Cr of CrWO3 clusters. This procedure is iterated for n = 2 to 4 and optimized at the ground-state energy level. Figure represents the ground-state optimization of pristine graphene and its composite structures.

Figure 1

(a) Top and (b) side views of the ground-state optimized structure of graphene, g-WO3, and g-CrWO3 composites.

The geometry-optimized graphene exhibits three bonding types: C–C, C=C, and C–H. The relevant average bond lengths of graphene range from 1.384 to 1.435 for C=C and C–C and 1.088 Å for C–H, and these values agree with those in the literature.[34,37] When graphene is decorated with the WO3 and CrWO3 clusters, there are no significant changes in the bond lengths, as shown in Figure a. However, a slight increase in the bond lengths has been observed for the isolated WO3 and CrWO3 clusters when compared with our previous work.[24] WT and CrT represent the terminal W–O and Cr–O bonds, respectively, whereas WB and CrB are the corresponding bridging bonds, as shown in Figure b. Furthermore, the binding distance between the cluster and the graphene surface is increased for the composites of g-W2O6 and g-W3O9. However, there is a slight decrease for g-W4O12. The relevant binding distances are 2.785, 3.883, and 3.849 Å for WO3 (n = 2 to 4), respectively. The same pattern is noticed for the CrWO3 cluster (n = 2 to 4), and the binding distances are 2.760, 3.916, and 3.981 Å for n = 2 to 4, respectively.

The binding energy (Eb) of the metal oxide cluster on the graphene surface is defined as follows[38]where Ecluster/graphene, Egraphene, and Ecluster are the total energy of the cluster on graphene, the total energy of graphene, and the total energy of the cluster, respectively. The calculated binding energies for the interaction of clusters on the graphene surface are presented in Table .

Table 1

Calculated Binding Energy (Eb), HOMO Energy (EHOMO), LUMO Energy (ELUMO), HOMO–LUMO Energy Gap (Eg), and Charge Transfer of Mulliken (ΔQMulliken) and NBO (ΔQNBO), and the Charge Transfer Is between C (Graphene) and W (WO3) or Cr (CrWO3)

structure	Eb (eV)	ΔEgraphene (eV)	ΔEcluster (eV)	ΔEGC (eV)	EHOMO (eV)	ELUMO (eV)	Eg (eV)	ΔQMulliken (e)	ΔQNBO (e)
graphene					–5.653	–1.628	4.025
g-W2O6	–3.83 × 10–1	–1.60 × 104	–2.51 × 104	4.11 × 104	–6.219	–4.211	2.008	–0.280	–0.113
g-W3O9	–1.83 × 10–1	–2.40 × 104	–2.51 × 104	4.91 × 104	–5.869	–4.739	1.130	–0.014	–0.037
g-W4O12	–2.51 × 10–1	–3.20 × 104	–2.51 × 104	5.70 × 104	–6.041	–4.638	1.403	–0.001	–0.010
g-CrWO6	–5.43 × 10–1	–1.65 × 104	–2.51 × 104	4.16 × 104	–6.387	–5.067	1.321	–0.007	–0.016
g-CrW2O9	–3.08 × 103	–2.45 × 104	–2.51 × 104	4.65 × 104	–6.249	–5.477	0.772	0.040	0.043
g-CrW3O12	–4.61 × 10–1	–3.25 × 104	–2.51 × 104	5.75 × 104	–6.343	–5.562	0.782	0.045	0.056

The decreasing order of binding energy of the composites is g-CrW2O9 > g-W3O9 > g-W4O12 > g-W2O6 > g-CrW3O12 > g-CrWO6. The results indicate that the CrW2O9 cluster strongly binds to the graphene surface. To obtain further insights into the stabilizing forces for graphene/cluster, the binding energy is partitioned into[39,40]where ΔEgraphene is the graphene deformation energy defined as the energy difference between the graphene in bonding and isolated state, ΔEcluster measures the energy penalty corresponding to the deformation of the cluster from the composites and is defined as the energy difference between the cluster in bonding and isolated state, and ΔEGC is defined as the energy difference between the total binding energy of the composites and the sum of ΔEgraphene and ΔEcluster. The calculated partitioning of binding energies for the interaction of the clusters on the graphene surface into ΔEgraphene, ΔEcluster, and ΔEGC is shown in Table . From the three partition components of the binding energy, both ΔEgraphene and ΔEGC show contribution to Eb. ΔEgraphene becomes more negative as the size of the cluster composited on the graphene increases. On the other hand, ΔEGC becomes more negative as the cluster size increases. The more negative ΔEgraphene and more positive ΔEGC are from g-CrWO3 than g-WO3 for n = 2 and 4.

Further investigation of the binding energy of graphene composite, HOMO–LUMO energy gap distribution, and density of state (DOS) plot is elucidated. The energy gap (Eg) between the HOMO and LUMO energy level was calculated as follows[29]where ELUMO and EHOMO are the energy values of LUMO and HOMO, respectively.

As shown in Figure a, the results indicated that no shift occurred in the DOS peak of the graphene composite clusters of WO3 and CrWO3 (n = 2 to 4). This denotes that the clusters had undergone physisorption upon binding to the graphene surface. However, the HOMO–LUMO energy gap was observed to be reduced significantly for both the g-WO3 and g-CrWO3 composites. When the cluster size (n) is increased from 2 to 3, the energy gap decreases and then slightly increases when n is increased to 4, as summarized in Table . Therefore, less energy is required for the excitation electrons to be polarized and magnetized due to the smaller energy gap.[24,41]

Figure 2

DOS of the ground-state optimized structure of graphene, g-WO3, and g-CrWO3 composites including (a) HOMO–LUMO energy profiles and (b) s, p, and d angular moment orbitals. The dashed vertical lines represent the zero-point Fermi level.

The Mulliken and natural bond orbital (NBO) charge population analyses were applied to elucidate the charge transfer on the system and were calculated using the equation[42]where X represents Mulliken or NBO charge transfer of graphene, metal oxide cluster, or gas molecule. Qafter and Qbefore are the charges of X after and before adsorption, respectively. The negative value of ΔQX indicates the electron loss of X.

As seen in Table , both the ΔQMulliken and ΔQNBO values increased with the cluster size. In the g-WO3 composite, the increase in ΔQMulliken for n = 2 to 4 is from −0.280, −0.014, and −0.010 e, whereas the ΔQNBO values are −0.113, −0.037, and −0.010 e, respectively. For g-CrWO3, the increased value of −0.007, 0.40, and 0.045 e (ΔQMulliken) and −0.016, 0.043, and 0.056 e (ΔQNBO) correspond to n = 2 to 4, respectively. Moreover, greater charge transfer occurred as the CrWO3 cluster binds to the surface of graphene, and this increases the electron density.

To further understand the nature of charge transfer within graphene composites, the DOSs of s, p, and d angular moment orbitals were obtained and are presented in Figure b. Near the Fermi level, only the peak of the p-orbital is observed and this is due to the sp2 hybridization π electron cloud as indicated by Kumar et al.[34] As the composite is formed, the peaks are observed to be shifted slightly toward the negative energy level, and more peaks emerge within the p-orbital with greater peak spikes of the following order: g-CrWO3 > g-WO3 > graphene. The result indicates that more electrons within the hybridization of sp2 π clouds of g-CrWO3 are involved in the charge transfer.

With Koopman’s principle,[43,44] the surface activity of graphene and its composites is generally investigated through chemical hardness (η), chemical potential (μ), and electrophilicity index (ω), and the calculated results are shown in Table . The surface activity of graphene and its composites was investigated by calculating the chemical hardness, chemical potential, and electrophilicity index based on Koopman’s approach, which can be expressed as[43−45]where I and A are the ionization potential (≅–EHOMO) and electron affinity (≅–ELUMO), respectively.

Table 2

Calculated Chemical Hardness (η), Chemical Potential (μ), and Electrophilicity Index (ω) for Graphene and Its Composites

structure	η (eV)	μ (eV)	ω (eV)
graphene	4.02	–3.64	1.65
g-W2O6	2.01	–5.21	6.77
g-W3O9	1.13	–5.30	12.45
g-W4O12	1.40	–5.34	10.16
g-CrWO6	1.32	–5.73	12.42
g-CrW2O9	0.77	–5.86	22.26
g-CrW3O12	0.78	–5.95	22.67

Chemical hardness is directly associated with the HOMO–LUMO energy gap, explaining the structural stability and reactivity. The decreasing order of structural stability is as follows: graphene > g-W2O6 > g-W4O12 > g-W3O9 > g-CrWO6 > g-CrW3O12 > g-CrW2O9. The result indicates that graphene is the most stable but least reactive, while the g-CrW2O9 composite exhibits the highest reactivity but with the lowest structural stability.

The tendency of electrons escaping from the composite is described by the chemical potential (μ). The chemical potential of graphene composites decreases as the cluster size is increased from n = 2 to 4 corresponding to WO3 and CrWO3 clusters being stuck onto the graphene surface. The order of the chemical potential (μ) is as follows: g-CrW3O12 < g-CrW2O9 < g-CrW2O9 < g-W4O12 < g-W3O9 < g-W2O6 < graphene. This also denotes that the Cr-doped clusters on graphene (g-CrWO3) show a lower chemical potential when compared to the g-WO3 composites. Thus, we can conclude that g-CrWO3 composites can more readily donate their electrons to the nearby molecules.

The electrophilicity index (ω) is a measure of the capacity to accept electrons from the environment to form a stable energy state. Electrophilicity has the following order: g-CrW3O12 > g-CrW2O9 > g-CrWO6 > g-W3O9 > g-W4O12 > g-W2O6 > graphene. The results elucidate that g-CrWO3 composites are more stable in accepting electrons. Based on the above results, both g-WO3 and g-CrWO3 composites are potential materials for a gas sensor with high stability and reactive surface activity. Moreover, the decoration of the graphene surface with CrWO3 composites demonstrates a highly reactive and stabilized structure for accepting/donating electrons when compared to g-WO3 composites.

Geometric Structure and CO Adsorption Property on Graphene, g-WO3, and g-CrWO3 Composites

The interaction of WO3 and CrWO3 clusters on graphene with CO is evaluated, and the results are presented in Figure . The ground-state optimized structure reveals that the interaction exhibits only physisorption behavior as there is no formation of a bond between CO and graphene and its composite. Furthermore, weak physisorption does not cause significant changes to the structural geometry of graphene and its composites.

Figure 3

(a) Top and (b) side views of the ground-state optimized structure of graphene, g-WO3, and g-CrWO3 composites after CO adsorption.

The adsorption energy (Eads) of CO molecule on the surface (graphene and its composites) was calculated using the equation[46]where ECO–surface, Esurface, and ECO are the total energy of the surface with CO, the total energy of the surface without CO, and the total energy of CO gas molecule, respectively. The negative value of Eads indicates an exothermic adsorption reaction. The calculated adsorption energy for CO interaction with graphene and its composites are tabulated in Table . The DFT calculations show that the adsorption distance increases as the absorption of CO occurs with the larger clusters. In g-WO3 composites, the adsorption distance increases from 2.261, 2.283, and 2.341 Å for n = 2 to 4, while the adsorption energy increases from −0.823, −0.658, and −0.390 eV, respectively. The same trend of adsorption distance is observed for the g-CrWO3 composites, and their values are 2.260, 2.290, and 2.382 Å for n = 2 to 4 with the adsorption energies of −0.638, −0.486, and −0.327 eV, respectively, although graphene exhibits a significantly higher adsorption distance, 3.635 Å, and has the lowest adsorption energy, −0.0472 eV. If the adsorption energy is near zero or low, the adsorption becomes improbable.[47] It also reveals that among the composites, the adsorption energy is slightly higher in g-CrWO3 when compared to g-WO3 composites, and this also indicates that composites are more favorable for the absorption of CO gas by having a smaller adsorption distance and higher adsorption energies.

Table 3

Calculated CO Adsorption Energy (Eads) at a Distance (dCO), HOMO energy (EHOMO), LUMO Energy (ELUMO), HOMO–LUMO Energy Gap (Eg), and Charge Transfer of Mulliken (ΔQMulliken) and NBO (ΔQNBO)a

structure	Eads (eV)	dCO (Å)	EHOMO (eV)	ELUMO (eV)	Eg (eV)	ΔQMulliken (e)	ΔQNBO (e)
graphene-CO	–0.0472	3.635	–5.647	–1.629	4.018	0.005	0.003
g-W2O6-CO	–0.823	2.261	–5.893	–3.813	2.080	–0.080	–0.376
g-W3O9-CO	–0.658	2.283	–5.924	–4.770	1.155	–0.121	–0.356
g-W4O12-CO	–0.390	2.341	–5.832	–4.763	1.069	–0.145	–0.323
g-CrWO6-CO	–0.638	2.260	–5.831	–4.918	0.913	–0.043	–0.369
g-CrW2O9-CO	–0.486	2.296	–6.017	–5.330	0.687	–0.104	–0.334
g-CrW3O12-CO	–0.327	2.382	–6.169	–5.396	0.773	–0.156	–0.307

The charge transfer is between C of CO and C/W of graphene/composites.

To obtain further insight into the influence of CO adsorption on the structures, the DOS with HOMO–LUMO energy distributions and its s, p, and d electron orbitals are presented in Figure a,b. The figures reveal that there is no shift in peaks after and before the adsorption of CO on the structures. This indicates a significant decrease in the HOMO–LUMO energy gap as the cluster size is increased from n = 2 to 4. In addition, the HOMO–LUMO energy gap of g-CrWO3 composites with CO adsorption is lower in comparison with those of the respective g-WO3. The energy gaps for g-CrWO3 (n = 2 to 4) composites are 0.913, 0.687, and 0.773 eV, respectively, while the energy gaps of g-WO3 (n = 2 to 4) composites are 2.080, 1.155, and 1.069 eV, respectively. The calculated ΔQMulliken and ΔQNBO are shown in Table , indicating that g-CrWO3 gained more electrons than its corresponding g-WO3 composites. Moreover, the physisorption between CO and g-CrWO3 composites exhibits strong hybridization due to the high overlapping of the electron cloud near the Fermi level, as shown in Figure b.

Figure 4

DOS of the ground-state optimized structure of graphene, g-WO3, and g-CrWO3 composites after CO adsorption including (a) HOMO–LUMO energy profiles and (b) s, p, and d angular moment orbitals. The dashed vertical lines represent the zero-point Fermi level.

Vibrational Analysis of the CO Adsorption Effect on Graphene, g-WO3, and g-CrWO3 Composites

The nature of the molecular interaction is investigated using the vibrational frequency and the intensity of infrared (IR) spectra. The computed IR spectra of graphene and its composites before and after the adsorption of CO are shown in Figure . The experimental studies for the g-WO3 and g-CrWO3 composites with CO gas are not available in the literature. However, the experimental IR vibrational spectra of the g-WO3 composite are available in the literature.[18,20] There is a significant difference in the spectra between graphene and its composites g-WO3 and g-CrWO3 with CO. The characteristic peak of CO appears at 2027 cm–1 for graphene when the before and after CO adsorption spectra are compared. As the CO is adsorbed by the graphene composites, this peak is shifted to the left from 2027 cm–1 to a range from 2135 to 2140 cm–1. The same trend in the peak shift is observed for the O–W–O peak in g-WO3 composites, in which the shifts are from 608 to 702 cm–1 and 930 to 934 cm–1 for g-W2O6 and g-W4O12, respectively. However, g-W3O9 shifts to the opposite direction, from 887 to 882 cm–1. The same peak for the g-W3O9 composite at 887 cm–1 was reported by Chu et al.[18] However, many different values were reported for the O–W–O peak of g-W2O6 and g-W4O12 composites at different temperature variations in the literature.[18,20,48] The O–W–O peak of the g-CrWO3 composite is observed and shifts to the right from 750 to 668, 807 to 790, and 846 to 842 cm–1 for n = 2 to 4, respectively. The relevant peak of O–Cr–O for CrWO3 composites is at 1650 cm–1, and there is no observed shift for this peak. Moreover, a slight shift is also seen for the stretching of the C–H bond peak, that is, from 3213 to 3216 cm–1 for both g-WO3 and g-CrWO3 composites after the adsorption of CO. Hence, the observed small shift indicates the physisorption characteristic between the CO and graphene and its composites as mentioned previously.

Figure 5

Infrared spectra of graphene, g-WO3, and g-CrWO3 (a) before and (b) after adsorption of CO molecule on the surface.

Photoelectron Spectroscopy Analysis of the CO Adsorption Effect on Graphene, g-WO3, and g-CrWO3 Composites

The photoelectron spectroscopy (PES) data are presented in Figure . PES provides information on the adsorption interaction of CO on the graphene and the graphene composites, g-WO3 and g-CrWO3. PES involves the ejection of the nonlocalized electron onto the surface via excitation by an electromagnetic wave.[49] The first peak is associated with the HOMO energy, and it is referred to as the vertical detachment energy (VDE), whereas the other peaks are related to the deeper orbitals that excite the higher binding energy.[50]Figure shows changes caused by CO adsorption in all of the composites except graphene itself and the g-W3O9 composite. After the adsorption of CO, the peaks are generally shifted to the left. The three major peaks from the g-W2O6 spectra that experienced shifts are 6.247, 7.868, and 9.403 eV, and they were shifted to 5.921, 7.598, and 9.088 eV after CO was adsorbed. There are five major peaks for both g-W4O12 and g-CrWO6 composites. The shift is from 6.057, 7.332, 7.772, 9.128, and 9.718 eV to 5.842, 7.152, 7.568, 8.923, and 9.448 eV, respectively, for the g-W4O12 composite. The shift is from 6.432, 7.668, 8.117, 8.718, and 9.033 eV to 5.837, 7.132, 7.553, 8.990, and 9.803 eV for the g-CrWO6 composite, respectively. The g-CrW2O9 and g-CrW3O12 composites show four major peaks, and the peaks shift from 6.367, 7.658, 8.088, and 9.608 eV to 6.197, 7.488, 7.968, and 9.523 eV for the g-CrW2O9 composite, while the shift for the g-CrW3O12 composite is from 6.382, 7.663, 8.098, and 9.623 eV to 6.202, 7.482, 7.918, and 9.528 eV, respectively. From the shift in the peak, it is evident that there is physisorption between CO and the structure of the composites, and this leads to the dipole moment polarization of the upper part of the structure.[51] Thus, it can be deduced that there is physisorption interaction in the deeper orbitals. The decreasing order of the physisorption interaction is g-W4O12 > CrWO6 > g-CrW2O9 and g-CrW3O12 > g-W2O6 > g-W3O9 > graphene.

Figure 6

Photoelectron spectroscopy (PES) intensity–binding energy of graphene, g-WO3, and g-CrWO3 composites before and after CO adsorption represented in black and red solid curves, respectively.

Electrostatic Potential (ESP) Analysis of CO Adsorption on Graphene, g-WO3, and g-CrWO3 Composites

The electrostatic potential (ESP) analysis is performed to elucidate the position and strength of the physisorption interaction for CO adsorption on graphene and its composite.[52,53] To aid the visualization of the charge density of ESP, a contour line is included. The ESP and the contour line for CO adsorption on graphene and its composite are shown in Figure . The ESP distribution for all structures is configured with 0.02 and −0.02 for positively and negatively charged regions, respectively, with the isovalue set as 0.002. Blue and red refer to the most positively and negatively charged regions, respectively, for its physisorption surface extrema.

Figure 7

Electrostatic potential (ESP) distribution maps including the contour line for the optimized structure of graphene, g-WO3, and g-CrWO3 composites (isovalue is set to 0.002).

From Figure , it is obvious that the charged region between the CO and the absorbent surface that has the least positive ESP belongs to graphene/CO (in green) than in its composite/CO structure (in blue), which indicates that the physisorption interaction of CO is stronger with the composite. Among the composites, the ESP distributions of both g-WO3 and g-CrWO3 (n = 2 to 4) are identical to each other. The major difference is that the ESP distribution belonging to the Cr atom of the g-CrWO3 composite decreases the negative charge distribution, leading to a more saturated positive charge region after CO is absorbed for n = 2 to 4. Thus, this indicates that the adsorption of CO is more prominent in g-CrWO3 composites when compared to the others because of the stronger physisorption interactions.

Sensing Characteristics of Graphene, g-WO3, and g-CrWO3 Composites for CO Gas Detection

Based on the above discussion, the interaction of CO gas is relatively more reactive with g-CrWO3 as compared to g-WO3 composites and has stronger physisorption interaction. Herein, the effectiveness of the structures for use in the gas sensor application is discussed. The sensing performance of gas sensors is dependent on parameters such as sensitivity, response time, selectivity, recovery time, etc. To improve the sensing performance, considerable adsorption energy with the targeted gas molecule is needed, and sufficient charge transfer between the gas and the structure is also required. These parameters ensure spontaneous adsorption of the target gas molecule, and they influence the electrical conductivity of the sensor. The sensor electrical conductivity (σ) is[54]where Eg represents the HOMO–LUMO energy gap, A is a constant (electrons/(m3 K3/2)), KB is Boltzmann constant (8.62 × 10–5 eV K–1), and T is the operating temperature in Kelvin.

A decrease in the HOMO–LUMO energy gap, Eg, after CO gas is adsorbed produces an increase in electrical conductivity. According to the results in Tables and 3, the increase trend in the electrical conductivity is deduced while CO adsorbs on the graphene and g-WO3 (n = 2 and 3) composites, whereas an opposite trend is observed upon adsorption interaction with the g-WO3 (n = 4) composite.[24]

Based on the empirical equation, the derived sensitivity of the sensor is expressed as followswhere R1 and R2 are the electrical resistances of the composite before and after the CO adsorption process, respectively.

The sensitivity of the sensor is expressed as below after using the inverse relationship between electrical resistance and conductivitywhere σ1 and σ2 are the electrical conductivities of the composite before and after CO adsorption on the composite, respectively. ΔEg is the difference between the energy gap before and after CO adsorption on the composite. t is calculated as

The strength of the gas adsorption is critically essential for measuring the desorption process. Higher adsorption strength hinders the process of desorption and leads to an increase in the sensor recovery time. Based on the transition states, the recovery time (τ) is given as follows[55]where ν, KB, and T are the attempt frequency, Boltzmann constant, and operating temperature, respectively. The attempt frequency and temperature are assumed to be kept constant with an order of ν = 1 THz at 300 K.

The calculated sensing parameters of sensitivity and recovery time at a constant temperature of 300 K are plotted in Figure . As seen in Figure a, the least sensitive for CO is graphene as it exhibits weak reactivity and physisorption interaction as compared to its composite structures. Generally, the sensitivity decreases in the g-CrWO3 composites as the cluster size increases when compared to their corresponding g-WO3 composites. This is due to the lesser chemical hardness and chemical potential with higher electrophilicity except for the g-W4O12 composite. The greater sensitivity of g-CrWO6 and g-W4O12 composites is probably attributed to the PES peak shift allowing for more interaction of physical adsorption to occur within their deeper orbitals. Hence, it is deduced that there is a fast reaction in accepting/donating electrons from/to occurring between CO gas and g-CrWO3 than between CO and the g-WO3 composites. Furthermore, there is a dramatic change in the energy gap after adsorption on the g-CrWO6 composite allowing for more electrons to be transferred within its deeper orbitals. Therefore, we conclude that the g-CrWO6 composite is highly sensitive in detecting CO gas.

Figure 8

Calculated (a) sensitivity and (b) recovery time of graphene, g-WO3, and g-CrWO3 composites for CO adsorption.

As shown in Figure b, the highest value of recovery time is obtained when CO is adsorbed on the graphene surface. This reveals the strong adsorption stability of CO with graphene and prevents the desorption process from occurring for the gas. As CO is adsorbed with the increasing cluster size composite, the recovery time also increases due to the higher electrophilicity nature after accepting electrons from the gas. Interestingly, the recovery time is more elevated in g-CrWO3 than g-WO3 composites, which indicates slightly high adsorption energy of CO gas with the structure. This is due to the greater physisorption strength as CO is adsorbed on g-CrWO3 than is observed in the other composite. The increasing order of recovery time is g-W2O6 < g-W3O9 < g-CrWO6 < g-CrW2O9 < g-W4O12 < g-CrW3O12 < graphene. This indicates that the recovery time and desorption process are slightly longer for g-CrWO3 than for the g-WO3 composite, which is considered acceptable. Hence, we deduce that the g-CrWO6 composite has more desirable adsorption and desorption characteristics with CO gas.

Conclusions

In summary, we have presented a systematic theoretical DFT investigation, at the ground-state B3LYP/LanL2DZ level with GenECP labeled basis sets, on the electronic structure and physisorption interaction of CO gas molecule with graphene and graphene composites g-WO3 and g-CrWO3 (n = 2 to 4). The results indicate that ta strong binding exists between graphene and the WO3 and CrWO3 clusters. In addition, there is no evidence of change in both graphene and cluster structural morphologies. Both g-WO3 and g-CrWO3 composites show a significant decrease in energy gap, Eg, with greater adsorption energy, charge transfer, and reactivity toward CO with an increase of the cluster size. The calculated photoelectron spectroscopy (PES) and infrared spectra when aligned with the respective electrostatic potential and contour line reveal that the composite structures have stronger physisorption interaction in g-CrWO3 composites when interacted with CO. The sensing properties for both composites exhibit a short adsorption time, and they are significantly more sensitive toward CO gas molecules. The results also show that the g-CrWO6 composite achieves the highest sensitivity, adsorption, and desorption characteristics. These findings conclude that Cr-doped tungsten oxide/graphene composite is a prominent material for CO sensing and CO adsorption.

Computational Details

The Gaussian 16 Software package was utilized for all calculations, including GaussView 6 software for molecular visualization.[56,57] All atoms were treated as LanL2DZ (Los Alamos National Laboratory 2 Double-Zeta) under GenECP labeled basis sets.[58−60] Additionally, LanL2DZ was used to provide a more consistent output with pseudopotential approximation in the periodic system.[61,62] The modeled graphene surface consists of 24 carbon atoms with 12 terminal hydrogen atoms used to prevent the dangling of graphene structural bonds. The frequency calculations along with geometry optimizations were based on the standard hybrid exchange–correlation function of Becke’s three parameters of Lee, Yang, and Parr (B3LYP).[63,64] The optimization was considered at the ground-state energy level with a net neutral charge and single spin multiplicity. The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) are modeled with respect to the vacuum level. The Mulliken and natural bond orbital (NBO) charge population analyses were carried out using the NBO 3.1 program embedded within the Gaussian 16 package.[56,65] The electrostatic potential (ESP) and contour line surfaces of CO adsorption on graphene and its composites were obtained using the Cubegen utility of the Gaussian 16 package.[56] Density of states (DOS), infrared (IR) spectra, and photoelectron spectroscopy (PES) were analyzed and plotted using Multiwfn v3.8 software.[66]