Al-Doped MoSe2 Monolayer as a Promising Biosensor for Exhaled Breath Analysis: A DFT Study.

Exhaled breath analysis by nanosensors is a workable and rapid manner to diagnose lung cancer in the early stage. In this paper, we proposed Al-doped MoSe2 (Al-MoSe2) as a promising biosensor for sensing three typically exhaled volatile organic compounds (VOCs) of lung cancer, namely, C3H4O, C3H6O, and C5H8, using the density functional theory (DFT) method. Single Al atom is doped on the Se-vacancy site of the MoSe2 surface, which behaves as an electron-donor and enhances the electrical conductivity of the nanosystem. The adsorption and desorption performances, electronic behavior, and the thermostability of the Al-MoSe2 monolayer are conducted to fully understand its physicochemical properties as a sensing material. The results indicate that the Al-MoSe2 monolayer shows admirable sensing performances with C3H4O, C3H6O, and C5H8 with responses of -85.7, -95.6, and -96.3%, respectively. Also, the desirable adsorption performance and the thermostability endow with the Al-MoSe2 monolayer with good sensing and desorbing behaviors for the recycle detection of three VOCs. We are hopeful that the results in this paper could provide some guidance to the experimentalists fulfilling their exploration in the practical application, which can also broaden the exploration of transition-metal dichalcogenides (TMDs) in more fields as well.

Introduction

Lung cancer is ren class="Chemical">cognized around the world as the second most prevalent cancers in adult men and women, and its morbidity and mortality rates are the highest among all cancers.[1] As reported, the tumors growing in the body will generate some specific substance, releasing volatile organic compounds (VOCs) into the blood and exchanging in the lung.[2] Besides, the exhaled VOCs of lung cancer patients are significantly more than those of the healthy people,[3] which provides the possibility to clarify the possible patients and evaluate the severity of diagnosed ones. Therefore, exhaled breath analysis becomes an attractive, rapid, and workable method to diagnose lung cancer without trauma issue, and the characterized VOCs are accepted as the biomarkers to reflect the potential dysfunction of human lungs. These typical VOCs include hydrocarbons such as isoprene (C5H8) and methyl cyclopentane (C6H12), hydrocarbon derivatives such as acetone (C3H6O) and 2-propenal (C3H4O), and aromatic hydrocarbons such as benzene (C6H6) and ethylbenzene (C8H9).[4−7]

In terms of VOC detection, a chemical resistance-type n class="Chemical">sensor is full of potential with advantages of rapid response, high sensitivity, and low cost.[8,9] Recently, many two-dimensional (2D) materials with favorable chemical reactivity and high electron mobility are demonstrated with strong interaction with the gas molecules,[10,11] including transition-metal dichalcogenides (TMDs), III–IV compounds, and V group monolayer. Besides, metal-doping could significantly enhance the adsorption and sensing behaviors of the materials upon gas species due to the strong catalytic property of metal atom(s).[12−14] This would be beneficial in guaranteeing their application in some harsh environment with good sensitivity. Specifically, the semiconducting MoSe2 monolayer with a direct band gap of 1.55 eV has received great attention as a sensing material for small gases,[15−17] which stimulates us to theoretically study its performance upon VOC sensing. Besides, aluminum (Al) is a common and inexpensive metal with a superior catalytic behavior upon gas interactions, making it a frequently used dopant for the surface to promote the sensing behavior of the material.[18,19]

In this work, we propose the Al-doped MoSe2 (Al–MoSe2) monolayer as a possible biosensor for sensing the VOCs of lung cancer based on density functional theory (DFT). Since the Se-vacancy inevitably exists in the MoSe2 monolayer in the engineering synthesis, the Al–MoSe2 monolayer is determined as the Al-doping on the Se-defected MoSe2 monolayer to better meet the real condition, and we select C3H4O, C3H6O, and C5H8 as the typical VOCs to perform their adsorption behavior onto the Al–MoSe2 surface. The sensing mechanism, desorption behavior, and thermostability of the Al–MoSe2 monolayer are conducted as well to fully understand its property as a chemical sensor. Our results manifest the feasibility of the Al–MoSe2 monolayer as a reusable sensor for the detection of VOCs, which provides the possibility for its further exploration in the diagnosis of lung cancer in daily life. From this aspect, our work can offer some guidance to the experimentalists and is important to broaden the application of TMDs in more fields.

Results and Discussion

Al-Doping Behavior on the Se-Defected MoSe2 Monolayer

Chaln class="Chemical">cogen vacancies in TMDs play a crucial role in their geometric and electronic behaviors.[20] Herein, we first analyze the Se-vacancy behavior on the MoSe2 surface and then the Al-doping effect on the geometric and electronic property on the Se-defected monolayer. Figure exhibits the process of establishing the Al–MoSe2 monolayer based on the pristine MoSe2 surface. The Se-defected MoSe2 monolayer is established by removing a Se atom from the upper layer of the pure MoSe2 supercell. After full optimization, somewhat deformations could be identified compared with the pristine counterpart. Then, an Al dopant is adsorbed on the Se-vacancy of the Se-defected MoSe2 monolayer to form the Al–MoSe2 monolayer. Upon Al-doping on the Se-defected MoSe2 surface, the binding force (Eb) is calculated to be −4.27 eV, which suggests the strong interaction between the Al dopant and the Se-defected MoSe2 surface. Besides, the Al–Mo bonds are measured as 2.52 Å, slightly shorter than the original Se–Mo bond of 2.55 Å and the sum of the covalent radii of Se and Mo atoms (2.64 Å[21]), confirming the strong binding force between Al dopant and Mo atoms that leads to the formation of the chemical bonds of Al–Mo.[22] In addition, the obtained Eb is much larger than the cohesive energy of the Al atom (3.39 eV), indicating the stable doping of Al dopant on the Se-vacancy without the clustering problem.[23,24]

Figure 1

Formation process of the Al–MoSe2 monolayer from (a) pure, (b) Se-defected, and (c) Al-doped MoSe2 monolayers. In CDD, the green (rosy) area indicates electron accumulation (depletion). The isosurface is 0.01 eV/Å3.

After doping, the Al adatom is positively charged by 0.13e acn class="Chemical">cording to the Hirshfeld method, meaning its electron-losing property on the Se-defected MoSe2 surface. From the charge density difference (CDD) distribution, one can see that the Al adatom is surrounded by the electron depletion, whereas the neighboring Mo atoms are surrounded by electron accumulation. The evident electron overlaps on the Al–Mo bonds verify their ionic nature and a strong orbital interaction during Al-doping.

Figure depicts the band structure (BS) of various systems and orbitn class="Chemical">al density of state (DOS) of Al and Mo atoms to illustrate the deformation of the electronic behavior in the formation of the Al–MoSe2 monolayer. First, in the pure MoSe2 system, the BS implies that it shows direct semiconducting property with a band gap of 1.55 eV, in accordance with the previous work[25] manifesting the good accuracy of our calculations. After the removal of a Se atom, there exist several novel states within the band gap of the pure MoS2 system, narrowing the band gap to 1.02 eV accordingly. At the same time, the BS states become much denser, suggesting the enhanced electron mobility in the Se-defected system. Both perfect and Se-defected MoSe2 monolayers are nonmagnetic given the symmetric distribution of the BS states. However, the Al–MoSe2 system has a magnetic moment of 1 μB according to our result. The spin up (black line) is not symmetric with a spin down (red line).

Figure 2

BS of (a) pure, (b) Se-defected, and (c) Al-doped MoSe2 monolayers and (d) orbital DOS of Al and Mo. The Fermi level is set to 0. In BS, the black line is spin up and the red line is spin down.

To explicitly expound the magnetic property of the Al–n class="Chemical">MoSe2 monolayer, its spin density is plotted in Figure . It is found that the dipole moment is mainly localized on the Al dopant and the bonding Mo atoms, accounting for 0.148 and 0.509 μB, respectively. Apart from that, the spin up and spin down both shift to lower regions by about 0.12 and 0.46 eV, respectively. This finding not only supports the electron-losing property of the Al dopant causing n-doping in the system[26] but also evidences a decline in the band gap of 0.68 eV for the Al–MoSe2 monolayer. In the atomic DOS, the Al 3p orbital is highly overlapped with the Mo 3d orbital ranging at −0.5 to 1.2 eV, which confirms the orbital hybridizations and further verifies the strong binding force of the Al–Mo bonds. Moreover, the top of the valence band is occupied by the Mo atom, while the bottom of the conduction band is occupied by the Al dopant, agreeing with the charge-transfer path from the Al dopant to the Se-defected MoSe2 surface according to the Hirshfeld analysis.[27]

Figure 3

Spin density of the Al–MoSe2 monolayer.

Adsorption Performance of the Al–MoSe2 Monolayer

Figure shows the most stable configurations (MSC) for (a) C3H4O, (b) C3H6O, and (c) C5H8 adsorption on the Al–MoSe2 monolayer as well as related CDD distributions.

Figure 4

MSC and CDD for C3H4O (a), C3H6O (b), and C5H8 (c) adsorption on the Al–MoSe2 monolayer. The characteristic of CDD are the same as in Figure .

In the C3H4O system, one can n class="Chemical">see that the C3H4O molecule prefers to be adsorbed on the Al–MoSe2 surface through the molecule-parallel position with a small slope to the plane, in which the C=C bond approaches the Al dopant, while the −CHO group does not. One C atom of the C=C bond is captured by the Al dopant with the Al–C bond length of 2.13 Å, which suggests the stronger chemical reactivity of the C=C bond than the −CHO group when interacting with the Al dopant. The Ead in this system is −1.45 eV, indicating chemisorption for C3H4O adsorption on the Al–MoSe2 monolayer.[28] According to the Hirshfeld analysis, the C3H4O molecule accepts 0.16e from the Al-doped surface, as verified by the CDD wherein the electron accumulation is mainly localized on the adsorbed gas molecule, while the electron depletion is mainly localized on the Al dopant. Besides, the electron hybridization could be observed on the Al–C bond identifying the formation of a chemical bond.

For the C3H6O system, we find that the preferred n class="Chemical">configuration for the C3H6O molecule adsorption on the Al–MoSe2 surface is through the molecule-vertical position with the O atom oriented to the Al dopant. That is, the O atom in the ketone group is chemically active to interact with the Al dopant. The newly formed Al–O bond is measured as 1.80 Å, shorter than the sum covalent radii of Al and O atoms (1.89 Å[21]), confirming the strong binding force between them. The large Ead of −1.80 eV also manifests the strong chemical interaction between the Al–MoSe2 monolayer and the C3H6O molecule. Different from that in the C3H4O system, the C3H6O molecule is positively charged by 0.06e after adsorption indicating its weak electron-donating property. From the CDD, one can see that for the trapped C3H6O molecule, the electron depletion is mainly on the ketone group, which accounts for the donated charge to the Al–MoSe2 monolayer, while the overlaps between electron accumulation and electron depletion on the Al–O bond expound its chemical nature as well.

When it comes to the n class="Chemical">C5H8 system, the preferred configuration for the C5H8 adsorption is similar to that in the C3H4O system, in which the C5H8 molecule is almost parallel to the Al–MoSe2 plane and one C atom in the C=C bond forms a new bond with the Al dopant with the equivalent bond length of 2.13 Å. These findings suggest the strong chemical reactivity of Al dopant upon C=C bond. The Ead (−2.00 eV) obtained in this system is the largest among the three adsorption configurations, which means the best adsorption performance of the Al–MoSe2 monolayer upon C5H8 molecule compared with those of the C3H4O and C3H6O molecules. The Hirshfeld analysis indicates that 0.23e transfers from the C5H8 molecule to the Al–MoSe2 surface. The largest value of QT for the C5H8 molecule also suggests the strongest electron redistribution for the whole system, which is supposed to change the electronic behavior of the Al–MoSe2 monolayer to the maximum extent. From the CDD, one can see that the charge source mainly results from the trapped C=C bond of the C5H8 molecule, which is transferred by the Al–C bond mainly to the Al dopant. These findings explicate the charge-transfer path, the formation of the chemical Al–C bond, and the charge localization.

In short, the adsorption performance of the n class="Chemical">Al–MoSe2 monolayer upon three VOCs are in the order of C5H8 > C3H6O > C3H4O. Based on their calculated Ead, chemisorption could be identified in three systems. Furthermore, the morphologies of the adsorbed gas molecules also undergo different levels of deformations compared with their isolated structures, which indicates the geometric activation by the Al–MoSe2 monolayer in adsorption. Besides, the magnetic behavior of the Al–MoSe2 monolayer disappears after gas adsorption, which we assume attributes to the charge transfer that eliminates the effect of lone pair electron on the magnetic property of the whole system. At the same time, the electron redistribution caused by charge transfer could deform the electronic behavior of the Al–MoSe2 system, which will be analyzed in Section .

Electronic Behavior of the Al–MoSe2 Monolayer upon Gas Adsorptions

To comprehend the electronic behavior of the n class="Chemical">Al–MoSe2 monolayer upon VOC adsorption, BS and DOS are conducted. Furthermore, the frontier molecular theory is also employed to obtain the distributions of frontier molecular orbitals (FMOs), including highest occupied molecular orbital (HOMO), lowest unoccupied molecular orbital (LUMO), and related energies, which is also an important method to analyze the sensing mechanism of the chemical resistance-type sensor.[29,30] To perform the FMO calculations, the smearing is set to 10–4 Ha to guarantee the accuracy of their energies. The BS, FMO with related energies, and DOS of various systems are exhibited in Figure .

Figure 5

BS, FMO with related energies, and DOS of (a) Al–MoSe2 system, (b) C3H4O system, (c) C3H6O, and (d) C5H8 system. The Fermi level is set to 0. In BS, the black line is spin up and the red line is spin down.

Initially, we fon class="Chemical">cus on the BS and FMO analysis for the isolated Al–MoSe2 system. As mentioned above, the Al–MoSe2 monolayer shows magnetic property with a band gap of 0.68 eV. From the FMO, the HOMO and LUMO both distribute on the Al dopant, highlighting its strong chemical reactivity in the surroundings. The energy gap between HOMO and LUMO is calculated as 0.66 eV, close to the band gap in the BS implying the high accuracy of FMO calculations. In the C3H4O system, the band gap gets narrowed to 0.58 eV after gas adsorption, the energies of HOMO and LUMO are both down-shifted, with the energy gap calculated as 0.56 eV. The HOMO is mainly localized on the Al dopant, while the LUMO is mainly localized on the C3H4O molecule, which verifies the charge-transfer path from the C3H4O molecule to the Al–MoSe2 monolayer. After adsorption, one can see that molecular DOS of isolated C3H4O is left-shifted and split into several small states in the lower region, which means the electronic activation behavior of Al dopant in adsorption. Besides, the Al 3p orbital is highly hybrid with the C 2p orbital at −5.2 to 0 eV, which indicates the formation of the Al–C chemical bond.

Whereas in the n class="Chemical">C3H6O and C5H8 systems, the band gaps are narrowed to 0.52 and 0.51 eV, respectively. The HOMO and LUMO distributions are afflicted with different deformations in contrast to the isolated Al–MoSe2 system. Specifically, the HOMO is mainly localized on the bonding atoms (Al and O atoms for the C3H6O system, while Al and C atoms for the C5H8 system), and the LUMO is mainly localized on the Al dopant. These findings not only manifest the strong reactivity on the Al–O and Al–C bonds but also support the Hirshfeld analysis that C3H6O and C5H8 molecules show an electron-donating property. At the same time, the energies of HOMO and LOMO in such two systems experience remarkable up-shift, which is different from that in the C3H4O system. However, the energy gaps are similarly reduced to 0.51 eV for the C3H6O system and to 0.49 eV for the C5H8 system. Based on the molecular and orbital DOS of the C3H6O system, it is seen that there is little state deformation in the DOS of the adsorbed C3H6O compared with that of the isolated phase, and really weak orbital mixing between Al 3p and O 2p orbital is determined. These results may be attributed to the small charge transfer in the C3H6O adsorption that contributes not strong electron redistribution to the whole system. On the contrary, the state split could be identified in the DOS of the adsorbed C5H8 molecule, and there is obvious electron hybridization between the Al 3p and O 2p orbitals at −5.1 to 0.2 eV. These results illustrate the stronger interaction of Al dopant with the C5H8 molecule compared with that with the C3H6O molecule, in agreement with the largest Ead and QT in the C5H8 system.

In short, the band gap and energy gap of the Al–n class="Chemical">MoSe2 monolayer have identically declining trends after the adsorption of three VOCs, which confirms its increasing electrical conductivity with three VOCs adsorbed. We assume that the decline of the band gap and energy gap results from the DOS state contributions of the adsorbed molecules around the Fermi level within the band gap of the isolated Al–MoSe2 system. The orbital overlap of the bonding atoms indicates their intensity of hybridization, reflecting the strength of the biding force. The charge-transfer path could be identified from the FMO analysis, which provides a workable manner to evaluate the chemical reactivity of typical species and judge the position where the reaction occurs.[31]

Sensing Explanation of the Al–MoSe2 Monolayer

Based on the obtained results in n class="Chemical">Section , the sensing mechanism of the resistance-type gas sensor could be identified. As a chemical resistance-type sensor, the sensing response (S), determined by the change in the electrical resistance after and before gas adsorption, is important to evaluate its usability for detecting typical gas species, which could be assessed using the following formula[32,33]In formula , σ is the electrical conductivity, λ is a constant, Bg is the band gap of a certain system, k is the Boltzmann constant, and T is the working temperature; in formula , σgas and σpure, respectively, are the conductivity of the Al–MoSe2 monolayer after and before gas adsorption, respectively. On the basis of these two equations, the responses for the sensing C3H4O, C3H6O, and C5H8 molecules are calculated to be −85.7, −95.6, and −96.3%, respectively. In other words, the Al–MoSe2 monolayer has desirable negative responses to such three VOCs, and the decreased electrical resistance of the Al–MoSe2 monolayer upon exhaling gas is the basic sensing mechanism and evidence of the possible lung cancer in real clinic diagnosis.

The recovery property, mn class="Chemical">eanwhile, is another important parameter to evaluate the reusability of a chemical gas sensor. To this end, the recovery time (τ), the minimum time for the adsorbed gases desorption from the Al–MoSe2 surface, is defined based on the van’t-Hoff–Arrhenius[34]where A is the attempted frequency (1012 s–1 [35]), T is the temperature, and kB is the Boltzmann constant (8.318 × 10–3 kJ/(mol·K)). Ea is the potential barrier of desorption which in this work is determined to be equal to Ead. From this formula, it could be inferred that the recovery time is related to the working temperature of the sensor. Based on this, we plot the recovery time of the Al–MoSe2 monolayer at three typical temperatures in Figure . One can see from this figure that C5H8 is the hardest to desorb from the Al–MoSe2 surface, while C3H4O is the easiest to adsorb. Even so, the desorption of three VOCs at room temperature is somewhat unrealistic. On the other hand, the desorption of C3H4O become feasible as the temperature increases to 498 K; furthermore, as the temperature increases to 598 K, the recovery time of the Al–MoSe2 monolayer for the desorption of all three VOCs becomes acceptable. In that case, the Al–MoSe2 monolayer becomes reusable as a gas sensor for sensing VOCs from the exhaled breath, and the related devices based on the Al–MoSe2 monolayer will have a longer lifespan in real applications.

Figure 6

Recovery property of the Al–MoSe2 monolayer upon VOC desorption.

When it comes to the enhanced desorption performance through heating at high temperatures, the stability of the Al–MoSe2 monolayer becomes another issue. Therefore, we conduct the molecular dynamic simulation in a period of 1 ps (1000 fs) at 500 and 800 K for the Al–MoSe2 monolayer to verify its thermostability. The obtained configurations of the Al–MoSe2 monolayer after simulation are displayed in Figure . From this figure, one can see that the Al–MoSe2 surface suffers somewhat deformation in the high temperatures. However, these slight distortions have not impacted the morphology of the whole system, and the Al dopant experiencing tiny displacement remains its original doping site on the Se-vacancy MoSe2 monolayer. These findings imply the desirable thermostability of the Al–MoSe2 monolayer at 500 and 800 K. Meanwhile, the vibrational analysis suggests that the calculated frequency of the Al–MoSe2 monolayer ranges from 80.54 to 840.39 cm–1, in which the nonincluded virtual frequency further confirms the chemical stability of the Al–MoSe2 monolayer. From all these aspects, it could be inferred that the Al–MoSe2 monolayer possesses desirable thermostability and desorption of gas species from its surface is completely feasible without impairing its morphology.

Figure 7

Configurations of the Al–MoSe2 monolayer at (a) 500 K and (b) 800 K.

Conclusions

In this paper, we theoretically investigate the adsorption and sensing performance of the Al–MoSe2 monolayer upon three characteristic VOCs of exhaled breath to explore its potential as a resistance-type chemical sensor for an early diagnosis of lung cancer. The desorption behavior and the thermostability of Al–MoSe2 are also performed to fully understand its property as a gas sensor. The main conclusions are as follows:

Al dopant behaves as an electron-donor, leading to 1 μB of magnetic moment and narrowing the band gap for the whole system.

The adsorption performance of the Al–MoSe2 monolayer upon three VOCs are in the order of C5H8 > C3H6O > C3H4O, with sensing responses calculated as −85.7, −95.6, and −96.3%, respectively.

The Al–MoSe2 monolayer with desirable adsorption performance and thermostability offers good sensing and desorbing behaviors for the recycle VOC detection.

Our work can provide guidance for experimentalists to explore its practical use in the diagnosis of lung cancer and is important to broaden the further application of TMDs in more fields.

Computational Details

The spin-polarized calculations were implemented using DMol3 package[36] to obtain the results below. The Perdew–Burke–Ernzerhof (PBE) function within the generalized gradient approximation (GGA) was employed to deal with the electron exchange–correlation terms.[37] In terms of van der Waals force and long-range interactions, the Grimme method based on DFT-D2 was adopted.[38] Double numerical plus polarization (DNP) was selected as the atomic orbital basis set,[39] and DFT semicore pseudopotential (DSSP) method was selected to resolve the relativistic effect.[40] We adopted the Monkhorst–Pack k-point mesh of 6 × 6 × 1 for the supercell geometry optimizations and of 9 × 9 × 1 for electronic structure calculations.[41] The energy tolerance accuracy, maximum force, and displacement were set as 10–5 Ha, 2 × 10–3 Ha/Å, and 5 × 10–3 Å,[42] respectively. Self-consistent loop energy of 10–6 Ha, global orbital cutoff radius of 5.0 Å, and smearing of 0.005 Ha were applied to ensure the accuracy of the total energy.[43]

We established a 4 × 4 × 1 pure MoSe2 supercell with a van class="Chemical">cuum region of 15 Å to perform the calculations throughout this work, which contains 16 Mo and 32 Se atoms. It has been proved that a 4 × 4 supercell is large enough to conduct the gas adsorption process,[44] while a 15 Å slab is proper to prevent the interaction between adjacent units.[45] The lattice constant of the relaxed MoSe2 configuration here was 3.30 Å, which is in agreement with other theoretical work (3.31 Å[46]). The binding force (Eb) for Al-doping on the Se-vacancy MoSe2 monolayer is calculated aswhere EAl–MoSe, Evac-MoSe, and EAl are the energies of the Al–MoSe2 monolayer, the isolated Se-vacancy MoSe2 monolayer, and Al dopant, respectively. Besides, the adsorption energy (Ead) was calculated by the following formula[47]where EAl–MoSe, EAl–MoSe, and Egas are the energies of the adsorbed system, pure Al–MoSe2 monolayer, and gas molecule, respectively. The Hirshfeld method was used to consider the atomic charge of Al (QAl) and molecular charge of the adsorbed molecule (QT) in every system. The positive charge meant the electron-donating property of the analytes and vice versa. Only the most favorable adsorption configurations were plotted and analyzed in this work.