J Mol Model (2015) 21: 116 DOI 10.1007/s00894-015-2648-7
ORIGINAL PAPER
Adsorption of carbon monoxide on the pristine, B- and Al-doped C3N nanosheets Mansoureh Pashangpour 1 & Ali Ahmadi Peyghan 2
Received: 27 November 2014 / Accepted: 9 March 2015 / Published online: 15 April 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract The potential application of the intrinsic and extrinsic (8,0) zigzag single-walled C3N nansheets as chemical sensor for CO molecules has been investigated using density functional theory calculations. The calculation shows that the pristine sheet is a semiconductor with a HOMO-LUMO gap (Eg) of about 2.19 eV. The pristine and B-doped sheets can weakly adsorb a CO molecule with the adsorption energies of -4.8 and −4.6 kcal mol-1, and their electronic properties are not sensitive to this molecule. By replacing a C atom with an Al atom, localized impurity states are induced under the conduction level of the sheet. The binding interaction between the CO molecule and the Al-doped sheet becomes much stronger (Ead =−17.8 kcal mol-1). After the adsorption of CO on the Al-dope sheet, the Eg of the sheet is significantly decreased from 1.07 to 0.73 eV. This leads to a sizable increase in the resistance of the tube. Thus, the Al-doped sheet can show the presence of CO molecules by an electronic signal because of the change in its resistance and conductivity.
Keywords Sensor . B3LYP . DFT . Graphene
Introduction Carbon monoxide (CO) is one of the most dangerous gases in air pollution. It is produced by the incomplete combustion of * Ali Ahmadi Peyghan [email protected] 1
Department of Physics, College of Science, Islamshahr Branch, Islamic Azad University, Islamshahr, Iran
2
Young Researchers and Elite club, Central Tehran Branch, Islamic Azad University, Tehran, Iran
fuels and is commonly found in automobile exhaust, the burning of domestic fuels, and so on. It is highly toxic and extremely dangerous because it is colorless and odorless. The health effects of CO depend on the CO concentration and length of exposure. At CO levels above 70 ppm, symptoms can include headache, fatigue, and nausea. At sustained CO concentrations above 150–200 ppm, disorientation, unconsciousness, and death are possible [1]. Hence, detecting and measuring acetone concentrations in the workplace or human body are necessary for our safety and health. Since carbon nanotube (CNT) was discovered by Iijima [2], the properties and applications of this novel material have been investigated extensively. CNTs have recently emerged as a promising substitute for materials of different properties and various applications in hydrogen storage, gas sensors, textiles, and many more [3–7]. In addition to CNT, there are also some other nanotubes found experimentally such as carbon nitride (CNx). The CNx nanotubes are synthesized by pyrolyzing ferrocence/melamine mixture at 1050 Cinargonatmosphere [8]. Hales et al. [9] have recently studied structural and thermodynamic stability of small CxN nanotubes with x=1, 2, 3, 5, and 7. They have shown that the C3N nanotube (C3NNT) exhibits a distorted structure that is strongly independent of the chiral index [10]. Some predictions of the possible crystal structures for C3N can be found in the literature [11, 12]. By the analogy with CNTs, C3NNT can be considered to be formed by rolling C3N sheet. Graphene-like C3N sheet (hC3N) with sp2 bond may be implied in electronic devices in future. In contrast to the half-metal behavior of graphene, hC3N possesses polar N–C bond and wider band gap, which promises the potential application of h-C3N in semiconductorbased sensors. The fundamental sensing mechanism in these materials relies upon the change in electrical conductivity as a result of charge transfer with adsorbates [13]. However, not being able
116 Page 2 of 7
J Mol Model (2015) 21: 116
to change the electrical conductivity during the adsorption process a group of toxic molecules cannot be detected by pristine nanostructured materials. In order to conquer this drawback, considerable experimental and theoretical studies have been focused on improving the sensing performance of several nanostructures by manipulating the structure [14–20]. Among many possible strategies to tailor the sensitivity of nanostructures, generating new active sites on their surface is classically the premier method to overcome the drawback of the intrinsically inert character. Embedding of foreign atoms or doping in these structures may enable them to detect a wide range of gases [21–26]. Here, we are interested in finding out if there is a potential possibility for pristine hC3N serving as a chemical sensor to CO molecule, and if not, can we find a method for improving the sensitivity of sheet to CO? To these ends, we have studied the reactivity of CO to the pristine, B-doped and Al-doped C3N sheet, using density functional theory (DFT) calculations.
Computational methods We selected a C3N sheet (h-C3N), consisted of 102 C and 34 N atoms, in which the end atoms have been saturated with hydrogen atoms to reduce the boundary effects. Geometry optimizations and energy calculations have been performed on hC3N and different CO/h-C3N complexes using B3LYP [27, 28] functional augmented with an empirical dispersion term (B3LYP-D) [29] with 6-31G(d) basis set as implemented in the GAMESS suite of program [30]. GaussSum program [31] was used to obtain the density of states (DOS) results. The B3LYP density functional has been previously shown to reproduce experimental properties and has been commonly used in nanostructure studies [32, 33]. We have defined adsorption energy (Ead) in the usual way as: Ead ¼ EðCO=h−C3 NÞ–Eðh−C3 NÞ–EðCOÞ þ EBSSE ;
ð1Þ
where E (CO/h-C3N) corresponds to the energy of the h-C3N in which the CO has been adsorbed on the surface, E (h-C3N) is the energy of the isolated sheet, E (CO) is the energy of a single CO molecule, and EBSSE is the energy of the basis set Fig. 1 Geometrical parameters of the optimized h-C3N. Bonds are in angstrom
superposition error [34]. The Boys and Bernardi counterpoise correction was used to calculate BSSE [35].
Results and discussion In Fig. 1, we have shown partial structure of the optimized C3N sheet, where two types of bonds namely N-C, and C-C can be identified, with corresponding lengths of 1.40 and 1.41 Å, respectively. The h-C3N has two different kinds of hollow sites, involving the hollow center of the C6 hexagon (H1) and C4N2 hexagon (H2) as shown in Fig. 1. As shown by the calculated DOS and the energy gaps (Eg) between the HOMO and LUMO in Fig. 1 and Table 1, the pure sheet is found to be a semiconductor with Eg of 2.19 eV. At first, we investigated the most stable configuration of CO adsorbed on the pristine h-C3N. Several initial configurations have been considered in order to study the adsorption of CO on the sheet. A CO molecule was initially placed above a carbon and nitrogen atoms or the center of a six-membered ring (H1 or H2), with the CO molecule oriented perpendicular (with the C or O atom pointing toward the sheet) to the h-C3N. Several other configurations with the CO molecule placed parallel to the h-C3N plane were also tested. For simplification, only the three stable complexes are shown in Fig. 2. After the CO adsorption on the pristine sheet, the optimization dose not causes noticeable geometry change for both CO and the sheet as compared with the isolated molecule and sheet. This means that the interaction between molecule and the pristine sheet is very weak, as demonstrated by the Ead ranged between −4.5 and −4.8 kcal mol-1 and large interaction distance of 3.07–3.10 Å. A configuration with the adsorbed CO axis aligned parallel to the h-C3N plane atop of the H2 ring (configuration P.1) was found to be the most stable one for the pristine h-C3N. The large interaction distance and less negative value of Ead indicate that the nature is the interaction physisorption. Similarly, using overlap population DOS, Armaković et al. have shown that the adsorption of CO on a sumanene is dictated mainly by the physisorption mechanism [36]. The charge transfer between CO and pristine sheet was obtained from natural bond orbital (NBO) analysis. In the
J Mol Model (2015) 21: 116
Page 3 of 7 116
Table 1 Calculated adsorption energies (Ead, kcal mol-1), HOMO, and LUMO energies, and HOMO-LUMO energy gap (Eg) of the systems ΔEg (%)
System
Ead
a
Q (e)
EHOMO (eV)
ELUMO (eV)
Eg (eV)
b
h–C3N P.1 P.2 P.3 B-C3N B.1 B.2 Al-C3N Al.1
– −4.8 −4.7 −4.5 – −4.6 −4.2 – −17.8
– −0.024 −0.006 −0.008 – −0.006 0.000 – +0.127
−3.20 −3.28 −3.27 −3.28 −3.25 3.26 −3.26 −3.25 −3.22
−1.01 −1.04 −1.04 −1.06 −1.10 −1.15 −1.12 −2.18 −2.49
2.19 2.24 2.23 2.22 2.15 2.11 2.14 1.07 0.73
– 2.3 1.8 1.3 – 1.8 0.4 – 31.7
−6.2
+0.108
−3.19
−2.45
0.74
30.8
Al.2 a
Q is defined as the average of total NBO charge on the molecule
b
The change of HOMO-LUMO gap of sheet after CO adsorption
configuration P.1, the calculated charges on the C and O atoms of the CO are about 0.170 and −0.194 e, respectively. Meanwhile, a small charge (0.024 e) is transferred from the Fig. 2 Model for three stable adsorption of CO on the h-C3N and their density of states (DOS)
h-C3N to CO. To verify the effects of the adsorption of CO on the h-C3N electronic properties, the DOS plots of CO/h-C3N adsorption systems were calculated (Fig. 2). Such weak interactions are also evident in their Eg (Table 1), which show little change after the adsorption process. For example, the DOS of the h-C3N (Fig. 1) is similar to that of the most stable CO/hC3N (configuration P.1). The contribution of the CO electronic levels to the DOS for all systems is far away from the Fermi level. This change in the electronic properties is negligible indicating that C3N sheet is still a semi-conductor after CO adsorption. Thus, we conjecture that the electronic properties of pure h-C3N are insensitive to the CO molecule. To overcome the insensitivity of the h-C3N, the N atom was replaced by a B atom. As shown in Fig. 3a, substituting the N atom by an impurity of B atom, the geometric structures of the doped hC3N are slightly distorted, but the planarity of the sheet is not affected. The calculated bond lengths are 1.47 Å for the neighbor B-C bond in the doped sheet, being longer than the corresponding N-C bonds in the pure sheet. Also, the C–B–C angle in the doped sheet is 120° which are similar to C–N–C in the pure one.
116 Page 4 of 7
J Mol Model (2015) 21: 116
Fig. 3 Structure of optimized (a) B-C3N, (b) Al-C3N and their density of states (DOS). Bonds are in Å
Subsequently, we have explored CO adsorption on the B-C3N by locating the molecule above the B atom with different initial orientations. We identified two distinct adsorptive configuration of CO in doped sheet, as shown in Fig. 4. Configuration B.1 gives rise to an E ad of −4.6 kcal mol-1, which is less negative than the Ead value of configuration B.2 (Ead =−4.2 kcal mol-1). On comparison, CO adsorption on B-C3N energetically is almost similar to the pristine sheet. In configuration B.1 and B.2, carbon and oxygen atoms of CO was located on the top of B atom with the corresponding bond length of 3.24 and 3.37 Å, respectively. The small Ead of NH3 on sheet in these structures and large adsorption distance reveals the weak van der Waals interaction. In Table 1, we have Fig. 4 Model for two stable adsorption of CO on the B-C3N and their density of states (DOS)
summarized the results for Ead, and Eg for the CO adsorption on the B-C3N. Calculated DOS of B-doped C3N sheet (B-C3N) is shown in Fig. 3a, indicating that its Eg value is reduced to 2.15 eV compared to the pristine sheet. The DOS plot clearly shows that the B-C3N is still a semiconductor. Figure 4 shows that the CO adsorption through configurations B.1 and B.2 have no sensible effects on the electronic properties of the sheet so that the E g of the B-C 3 N has slightly decreased to 2.11 and 2.14 eV, respectively. The calculated DOS show that the CO adsorption on B-C3N can be generally classified as certain type of Belectronically harmless modification^. In other words, the electronic properties of the B-C3N are negligibly changed by the CO adsorption.
J Mol Model (2015) 21: 116
Page 5 of 7 116
Fig. 5 Model for two stable adsorption of CO on the Al-C3N and their density of states (DOS)
Thus, in the next step, the effects of substituting one of the nitrogen atoms by an Al atom on the geometrical structure and electronic properties of the C3N sheet and also on the adsorption behaviors were investigated. We have observed that the aluminum atom also projects out of the sheet and creates local deformation due to larger covalent radius of Al atom (1.18 Å) as compared with nitrogen atom (0.75 Å). The calculated bond lengths are 1.83 Å for the neighbor Al-C bond in the doped sheet, being longer than the corresponding N-C bonds in the pure sheet (Fig. 3b). The bond angles between C–Al–C Fig. 6 Calculated LUMO profile of Al-C3N
bonds are 103° and 104°, which indicate tetrahedral structure of bonding and prefer sp3 hybridization. This structural deformation results into a significant change in properties such as Ead, Eg and charge transfer. NBO analysis shows that Al atom acquires positive charge of magnitude 0.200 e. This clears that the charge is transferred from the Al atom to the vicinity carbon atoms. Thus the dopant site (Al atom) acts as an affinity center for strong adsorption of the CO molecule. We have performed calculations to predict the interaction of CO with Al-doped C3N sheet (Al-
116 Page 6 of 7
J Mol Model (2015) 21: 116
C3N) and find the most stable configurations. Similar to the BC3N, we identified two distinct adsorptive configuration of CO in Al-C3N, as shown in Fig. 5. Configuration Al.1 gives rise to an Ead of −17.8 kcal mol-1, which is more negative than the Ead value of configuration Al.2 (Ead =−6.2 kcal mol-1). In the most stable configuration, the carbon atom of molecule binds to the Al atom by forming a sigma (σ) Al–C bond with length of 2.07 Ǻ (Fig. 5b) with corresponding charge transfer of 0.123 e from CO to the Al-C3N. The more negative value of Ead indicates the interaction of CO molecule with Al-C3N is enhanced as compared to the pure and B-doped sheets. We have investigated frontier molecular orbital (FMO) analysis of Al-C3N. It can be seen in Fig. 6, LUMO profile of Al-C3N located mainly on the Al-doped area. As a result, the HOMO of CO, locating on the C atom, donates electrons preferably to the LUMO centered on the exposed Al site of the sheet. Also, Al has two fewer electrons than nitrogen, thus the Al-C3N is an electron-deficient system. When the CO atoms are adsorbed on the Al-C3N, there will be large electron transferring from the CO molecule to the Al-C3N, which makes the CO molecule more strongly bound than Al-C3N. Calculated DOS of Al-C3N is shown in Fig. 3b, indicating that its Eg value is reduced to 1.07 eV compared to the pristine sheet. After CO adsorption in configuration Al.1 (as the most stable structure), valence level remains constant but conduction level shifted to lower energies (−2.49 eV) compared to the bare Al-C3N (−2.18 eV) so that the Eg of sheet decreased from 1.07 to 0.73 eV (Fig. 5a). It is well known that the Eg (or band gap in bulk materials) is a major factor determining the electrical conductivity of a material and there is a classic relation between them as follows [17]: 3
σ ¼ CT 2 qe ðμe þ μh Þ
−Eg ; ekT
ð2Þ
where σ is the electric conductivity, qe, μe, μh, and k are the charge of electron, mobility of electrons and holes, and the Boltzmann constant, respectively. It should be noted that this equation works only for semiconductors in which by increasing the temperature the electrical conductivity will be increased. According to the equation, smaller values of Eg at a given temperature lead to larger electric conductivity. The considerable change of about 0.34 eV (Table 1) in the Eg value, demonstrates the high sensitivity of the electronic properties of Al-C3N toward the CO adsorption on its surface. We think that the proposed sensor may benefit from some advantages including high sensitivity: the Eg of Al-doped C3N sheet is appreciably sensitive to the presence of acetone so that it increases by about 31.7 % upon the adsorption of CO molecule, short recovery time: the adsorption energy of CO molecule is not so large as to hinder the recovery of Al-C3N and therefore the sensor will possess short recovery times since
based on the conventional transition state theory, the recovery time can be expressed as [16]: τ ¼ ν 0 −1 expð−Ead =kTÞ;
ð3Þ
where T is the temperature, k is the Boltzmann’s constant, and ν0 is the attempt frequency. According to this equation, more negative Ead values will prolong the recovery time in an exponential manner.
Conclusion The adsorption of CO molecule on the pristine, B− and Al− doped C3N sheets are investigated using DFT calculations. It is found that CO molecule is only weakly adsorbed onto the pristine and B-doped C3N sheet with low negative Ead and large separation. The electronic properties of the pure Bdoped C3N sheet have very limited change upon the adsorption of CO molecule. In contrast, the CO molecule shows strong interactions with the Al-C3N sheets. The more negative Ead and charge transfers of CO on the Al-C3N sheet are expected to induce significant changes in the electrical conductivity of sheet. These results may help to seek appropriate chemical modification methods to widen the application fields of the C3N nanomaterials.
Ethical statement We guarantee that authorship is limited to those who have made a significant contribution to the attached manuscript. All coauthors have seen and approved the final version of this manuscript and have agreed to its submission for publication. We verify that the manuscript is entirely original, and all words of others has been appropriately cited and quoted. We verify that this manuscript has not been submitted to any other journal or primary publication. We agree to allow the journal to review the data if requested. The authors confirm having read Bethical responsibilities of authors^ and affirm that this paper is consistent with those guidelines. The authors disclose no conflicts of interest. This project has been financed by Iran Nanotechnology Initiative Council.
References 1. 2. 3. 4. 5. 6. 7.
Fukui K, Nakane M (1995) Sens Actuators B Chem 24/25:486–490 Iijima S (1991) Nature 354:56–58 Chelmecka E, Pasterny K, Kupka T, Stobiński L (2012) J Mol Model 18:1463–1472 Dinadayalane TC, Kaczmarek A, Łukaszewicz J, Leszczynski J (2007) J Phys Chem C 111:7376–7383 Beheshtian J, Ahmadi Peyghan A, Bagheri Z (2012) J Mol Model 19:255–261 Beheshtian J, Peyghan AA, Bagheri Z (2012) Monatsh Chem 143: 1623–1626 Turabekova MA, Dinadayalane TC, Leszczynska D, Leszczynski J (2012) J Phys Chem C 116:6012–6021
J Mol Model (2015) 21: 116 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Czerw R, Terrones M, Charlier JC, Blase X, Foley B, Kamalakaran R, Grobert N, Terrones H, Tekleab D, Ajayan P (2001) Nano Lett 1: 457–460 Hales J, Barnard AS (2009) J Phys Condens Matt 21:144203– 144209 Kim E, Chen C, Köhler T, Elstner M, Frauenheim T (2001) Phys Rev B 64:094107–094112 Sandré É, Pickard CJ, Colliex C (2000) Chem Phys Lett 325:53–60 Azevedo S, De Paiva R (2006) Europhys Lett 75:126–131 Peng N, Zhang Q, Chow CL, Tan OK, Marzari N (2009) Nano Lett 9:1626–1630 Sasaki I, Minami N, Karthigeyan A, Iakoubovski K (2009) Analyst 134:325–330 Ahmadi Peyghan A, Omidvar A, Hadipour NL, Bagheri Z, Kamfiroozi M (2012) Physica E 44:1357–1360 Beheshtian J, Peyghan AA, Bagheri Z (2012) Sens Actuators B Chem 171–172:846–852 Peyghan AA, Hadipour NL, Bagheri Z (2013) J Phys Chem C 117: 2427–2432 Peyghan AA, Soltani A, Pahlevani AA, Kanani Y, Khajeh S (2013) Appl Surf Sci 270:25–32 Rastegar SF, Peyghan AA, Hadipour NL (2012) Appl Surf Sci 265: 412–417 Beheshtian J, Peyghan AA, Bagheri Z (2013) Struct Chem 24:165– 170 Beheshtian J, Peyghan AA, Bagheri Z (2013) Struct Chem 24: 1565–1570
Page 7 of 7 116 22.
Baei MT, Peyghan AA, Bagheri Z, Tabar MB (2012) Phys Lett A 377:107–111 23. Beheshtian J, Ahmadi Peyghan A, Bagheri Z (2013) Struct Chem 24:165–170 24. Beheshtian J, Peyghan AA, Noei M (2013) Sens Actuators B Chem 181:829–834 25. Peyghan AA, Noei M, Yourdkhani S (2013) Superlattices Microstruct 59:115–122 26. Saha S, Dinadayalane TC, Murray JC, Leszczynska D, Leszczynski J (2012) J Phys Chem C 116:22399–22410 27. Becke AD (1993) J Chem Phys 98:5648–5652 28. Lee C, Yang W, Parr RG (1988) Phys Rev B 37:785–789 29. Grimme S, Antony J, Ehrlich S, Krieg H (2010) J Chem Phys 132: 154104–154119 30. Schmidt MW, Baldridge KK, Boatz JA, Elbert ST, Gordon MS, Jensen JH, Koseki S, Matsunaga N, Nguyen KA, Su S, Windus TL, Dupuis M, Montgomery JA (1993) J Comput Chem 14: 1347–1363 31. O’Boyle N, Tenderholt A, Langner K (2008) J Comput Chem 29: 839–845 32. Beheshtian J, Soleymanabadi H, Kamfiroozi M, Ahmadi A (2012) J Mol Model 18:2343–2348 33. Ahmadi A, Beheshtian J, Kamfiroozi M (2012) J Mol Model 18: 1729–1734 34. Liu B, McLean AD (1973) J Chem Phys 59:4557–4558 35. Boys SF, Bernardi F (1970) Mol Phys 19:553–566 36. Armaković S, Armaković S, Šetrajčić J, Jaćimovski S, Holodkov V (2014) J Mol Model 20:1–14
ORIGINAL PAPER
Adsorption of carbon monoxide on the pristine, B- and Al-doped C3N nanosheets Mansoureh Pashangpour 1 & Ali Ahmadi Peyghan 2
Received: 27 November 2014 / Accepted: 9 March 2015 / Published online: 15 April 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract The potential application of the intrinsic and extrinsic (8,0) zigzag single-walled C3N nansheets as chemical sensor for CO molecules has been investigated using density functional theory calculations. The calculation shows that the pristine sheet is a semiconductor with a HOMO-LUMO gap (Eg) of about 2.19 eV. The pristine and B-doped sheets can weakly adsorb a CO molecule with the adsorption energies of -4.8 and −4.6 kcal mol-1, and their electronic properties are not sensitive to this molecule. By replacing a C atom with an Al atom, localized impurity states are induced under the conduction level of the sheet. The binding interaction between the CO molecule and the Al-doped sheet becomes much stronger (Ead =−17.8 kcal mol-1). After the adsorption of CO on the Al-dope sheet, the Eg of the sheet is significantly decreased from 1.07 to 0.73 eV. This leads to a sizable increase in the resistance of the tube. Thus, the Al-doped sheet can show the presence of CO molecules by an electronic signal because of the change in its resistance and conductivity.
Keywords Sensor . B3LYP . DFT . Graphene
Introduction Carbon monoxide (CO) is one of the most dangerous gases in air pollution. It is produced by the incomplete combustion of * Ali Ahmadi Peyghan [email protected] 1
Department of Physics, College of Science, Islamshahr Branch, Islamic Azad University, Islamshahr, Iran
2
Young Researchers and Elite club, Central Tehran Branch, Islamic Azad University, Tehran, Iran
fuels and is commonly found in automobile exhaust, the burning of domestic fuels, and so on. It is highly toxic and extremely dangerous because it is colorless and odorless. The health effects of CO depend on the CO concentration and length of exposure. At CO levels above 70 ppm, symptoms can include headache, fatigue, and nausea. At sustained CO concentrations above 150–200 ppm, disorientation, unconsciousness, and death are possible [1]. Hence, detecting and measuring acetone concentrations in the workplace or human body are necessary for our safety and health. Since carbon nanotube (CNT) was discovered by Iijima [2], the properties and applications of this novel material have been investigated extensively. CNTs have recently emerged as a promising substitute for materials of different properties and various applications in hydrogen storage, gas sensors, textiles, and many more [3–7]. In addition to CNT, there are also some other nanotubes found experimentally such as carbon nitride (CNx). The CNx nanotubes are synthesized by pyrolyzing ferrocence/melamine mixture at 1050 Cinargonatmosphere [8]. Hales et al. [9] have recently studied structural and thermodynamic stability of small CxN nanotubes with x=1, 2, 3, 5, and 7. They have shown that the C3N nanotube (C3NNT) exhibits a distorted structure that is strongly independent of the chiral index [10]. Some predictions of the possible crystal structures for C3N can be found in the literature [11, 12]. By the analogy with CNTs, C3NNT can be considered to be formed by rolling C3N sheet. Graphene-like C3N sheet (hC3N) with sp2 bond may be implied in electronic devices in future. In contrast to the half-metal behavior of graphene, hC3N possesses polar N–C bond and wider band gap, which promises the potential application of h-C3N in semiconductorbased sensors. The fundamental sensing mechanism in these materials relies upon the change in electrical conductivity as a result of charge transfer with adsorbates [13]. However, not being able
116 Page 2 of 7
J Mol Model (2015) 21: 116
to change the electrical conductivity during the adsorption process a group of toxic molecules cannot be detected by pristine nanostructured materials. In order to conquer this drawback, considerable experimental and theoretical studies have been focused on improving the sensing performance of several nanostructures by manipulating the structure [14–20]. Among many possible strategies to tailor the sensitivity of nanostructures, generating new active sites on their surface is classically the premier method to overcome the drawback of the intrinsically inert character. Embedding of foreign atoms or doping in these structures may enable them to detect a wide range of gases [21–26]. Here, we are interested in finding out if there is a potential possibility for pristine hC3N serving as a chemical sensor to CO molecule, and if not, can we find a method for improving the sensitivity of sheet to CO? To these ends, we have studied the reactivity of CO to the pristine, B-doped and Al-doped C3N sheet, using density functional theory (DFT) calculations.
Computational methods We selected a C3N sheet (h-C3N), consisted of 102 C and 34 N atoms, in which the end atoms have been saturated with hydrogen atoms to reduce the boundary effects. Geometry optimizations and energy calculations have been performed on hC3N and different CO/h-C3N complexes using B3LYP [27, 28] functional augmented with an empirical dispersion term (B3LYP-D) [29] with 6-31G(d) basis set as implemented in the GAMESS suite of program [30]. GaussSum program [31] was used to obtain the density of states (DOS) results. The B3LYP density functional has been previously shown to reproduce experimental properties and has been commonly used in nanostructure studies [32, 33]. We have defined adsorption energy (Ead) in the usual way as: Ead ¼ EðCO=h−C3 NÞ–Eðh−C3 NÞ–EðCOÞ þ EBSSE ;
ð1Þ
where E (CO/h-C3N) corresponds to the energy of the h-C3N in which the CO has been adsorbed on the surface, E (h-C3N) is the energy of the isolated sheet, E (CO) is the energy of a single CO molecule, and EBSSE is the energy of the basis set Fig. 1 Geometrical parameters of the optimized h-C3N. Bonds are in angstrom
superposition error [34]. The Boys and Bernardi counterpoise correction was used to calculate BSSE [35].
Results and discussion In Fig. 1, we have shown partial structure of the optimized C3N sheet, where two types of bonds namely N-C, and C-C can be identified, with corresponding lengths of 1.40 and 1.41 Å, respectively. The h-C3N has two different kinds of hollow sites, involving the hollow center of the C6 hexagon (H1) and C4N2 hexagon (H2) as shown in Fig. 1. As shown by the calculated DOS and the energy gaps (Eg) between the HOMO and LUMO in Fig. 1 and Table 1, the pure sheet is found to be a semiconductor with Eg of 2.19 eV. At first, we investigated the most stable configuration of CO adsorbed on the pristine h-C3N. Several initial configurations have been considered in order to study the adsorption of CO on the sheet. A CO molecule was initially placed above a carbon and nitrogen atoms or the center of a six-membered ring (H1 or H2), with the CO molecule oriented perpendicular (with the C or O atom pointing toward the sheet) to the h-C3N. Several other configurations with the CO molecule placed parallel to the h-C3N plane were also tested. For simplification, only the three stable complexes are shown in Fig. 2. After the CO adsorption on the pristine sheet, the optimization dose not causes noticeable geometry change for both CO and the sheet as compared with the isolated molecule and sheet. This means that the interaction between molecule and the pristine sheet is very weak, as demonstrated by the Ead ranged between −4.5 and −4.8 kcal mol-1 and large interaction distance of 3.07–3.10 Å. A configuration with the adsorbed CO axis aligned parallel to the h-C3N plane atop of the H2 ring (configuration P.1) was found to be the most stable one for the pristine h-C3N. The large interaction distance and less negative value of Ead indicate that the nature is the interaction physisorption. Similarly, using overlap population DOS, Armaković et al. have shown that the adsorption of CO on a sumanene is dictated mainly by the physisorption mechanism [36]. The charge transfer between CO and pristine sheet was obtained from natural bond orbital (NBO) analysis. In the
J Mol Model (2015) 21: 116
Page 3 of 7 116
Table 1 Calculated adsorption energies (Ead, kcal mol-1), HOMO, and LUMO energies, and HOMO-LUMO energy gap (Eg) of the systems ΔEg (%)
System
Ead
a
Q (e)
EHOMO (eV)
ELUMO (eV)
Eg (eV)
b
h–C3N P.1 P.2 P.3 B-C3N B.1 B.2 Al-C3N Al.1
– −4.8 −4.7 −4.5 – −4.6 −4.2 – −17.8
– −0.024 −0.006 −0.008 – −0.006 0.000 – +0.127
−3.20 −3.28 −3.27 −3.28 −3.25 3.26 −3.26 −3.25 −3.22
−1.01 −1.04 −1.04 −1.06 −1.10 −1.15 −1.12 −2.18 −2.49
2.19 2.24 2.23 2.22 2.15 2.11 2.14 1.07 0.73
– 2.3 1.8 1.3 – 1.8 0.4 – 31.7
−6.2
+0.108
−3.19
−2.45
0.74
30.8
Al.2 a
Q is defined as the average of total NBO charge on the molecule
b
The change of HOMO-LUMO gap of sheet after CO adsorption
configuration P.1, the calculated charges on the C and O atoms of the CO are about 0.170 and −0.194 e, respectively. Meanwhile, a small charge (0.024 e) is transferred from the Fig. 2 Model for three stable adsorption of CO on the h-C3N and their density of states (DOS)
h-C3N to CO. To verify the effects of the adsorption of CO on the h-C3N electronic properties, the DOS plots of CO/h-C3N adsorption systems were calculated (Fig. 2). Such weak interactions are also evident in their Eg (Table 1), which show little change after the adsorption process. For example, the DOS of the h-C3N (Fig. 1) is similar to that of the most stable CO/hC3N (configuration P.1). The contribution of the CO electronic levels to the DOS for all systems is far away from the Fermi level. This change in the electronic properties is negligible indicating that C3N sheet is still a semi-conductor after CO adsorption. Thus, we conjecture that the electronic properties of pure h-C3N are insensitive to the CO molecule. To overcome the insensitivity of the h-C3N, the N atom was replaced by a B atom. As shown in Fig. 3a, substituting the N atom by an impurity of B atom, the geometric structures of the doped hC3N are slightly distorted, but the planarity of the sheet is not affected. The calculated bond lengths are 1.47 Å for the neighbor B-C bond in the doped sheet, being longer than the corresponding N-C bonds in the pure sheet. Also, the C–B–C angle in the doped sheet is 120° which are similar to C–N–C in the pure one.
116 Page 4 of 7
J Mol Model (2015) 21: 116
Fig. 3 Structure of optimized (a) B-C3N, (b) Al-C3N and their density of states (DOS). Bonds are in Å
Subsequently, we have explored CO adsorption on the B-C3N by locating the molecule above the B atom with different initial orientations. We identified two distinct adsorptive configuration of CO in doped sheet, as shown in Fig. 4. Configuration B.1 gives rise to an E ad of −4.6 kcal mol-1, which is less negative than the Ead value of configuration B.2 (Ead =−4.2 kcal mol-1). On comparison, CO adsorption on B-C3N energetically is almost similar to the pristine sheet. In configuration B.1 and B.2, carbon and oxygen atoms of CO was located on the top of B atom with the corresponding bond length of 3.24 and 3.37 Å, respectively. The small Ead of NH3 on sheet in these structures and large adsorption distance reveals the weak van der Waals interaction. In Table 1, we have Fig. 4 Model for two stable adsorption of CO on the B-C3N and their density of states (DOS)
summarized the results for Ead, and Eg for the CO adsorption on the B-C3N. Calculated DOS of B-doped C3N sheet (B-C3N) is shown in Fig. 3a, indicating that its Eg value is reduced to 2.15 eV compared to the pristine sheet. The DOS plot clearly shows that the B-C3N is still a semiconductor. Figure 4 shows that the CO adsorption through configurations B.1 and B.2 have no sensible effects on the electronic properties of the sheet so that the E g of the B-C 3 N has slightly decreased to 2.11 and 2.14 eV, respectively. The calculated DOS show that the CO adsorption on B-C3N can be generally classified as certain type of Belectronically harmless modification^. In other words, the electronic properties of the B-C3N are negligibly changed by the CO adsorption.
J Mol Model (2015) 21: 116
Page 5 of 7 116
Fig. 5 Model for two stable adsorption of CO on the Al-C3N and their density of states (DOS)
Thus, in the next step, the effects of substituting one of the nitrogen atoms by an Al atom on the geometrical structure and electronic properties of the C3N sheet and also on the adsorption behaviors were investigated. We have observed that the aluminum atom also projects out of the sheet and creates local deformation due to larger covalent radius of Al atom (1.18 Å) as compared with nitrogen atom (0.75 Å). The calculated bond lengths are 1.83 Å for the neighbor Al-C bond in the doped sheet, being longer than the corresponding N-C bonds in the pure sheet (Fig. 3b). The bond angles between C–Al–C Fig. 6 Calculated LUMO profile of Al-C3N
bonds are 103° and 104°, which indicate tetrahedral structure of bonding and prefer sp3 hybridization. This structural deformation results into a significant change in properties such as Ead, Eg and charge transfer. NBO analysis shows that Al atom acquires positive charge of magnitude 0.200 e. This clears that the charge is transferred from the Al atom to the vicinity carbon atoms. Thus the dopant site (Al atom) acts as an affinity center for strong adsorption of the CO molecule. We have performed calculations to predict the interaction of CO with Al-doped C3N sheet (Al-
116 Page 6 of 7
J Mol Model (2015) 21: 116
C3N) and find the most stable configurations. Similar to the BC3N, we identified two distinct adsorptive configuration of CO in Al-C3N, as shown in Fig. 5. Configuration Al.1 gives rise to an Ead of −17.8 kcal mol-1, which is more negative than the Ead value of configuration Al.2 (Ead =−6.2 kcal mol-1). In the most stable configuration, the carbon atom of molecule binds to the Al atom by forming a sigma (σ) Al–C bond with length of 2.07 Ǻ (Fig. 5b) with corresponding charge transfer of 0.123 e from CO to the Al-C3N. The more negative value of Ead indicates the interaction of CO molecule with Al-C3N is enhanced as compared to the pure and B-doped sheets. We have investigated frontier molecular orbital (FMO) analysis of Al-C3N. It can be seen in Fig. 6, LUMO profile of Al-C3N located mainly on the Al-doped area. As a result, the HOMO of CO, locating on the C atom, donates electrons preferably to the LUMO centered on the exposed Al site of the sheet. Also, Al has two fewer electrons than nitrogen, thus the Al-C3N is an electron-deficient system. When the CO atoms are adsorbed on the Al-C3N, there will be large electron transferring from the CO molecule to the Al-C3N, which makes the CO molecule more strongly bound than Al-C3N. Calculated DOS of Al-C3N is shown in Fig. 3b, indicating that its Eg value is reduced to 1.07 eV compared to the pristine sheet. After CO adsorption in configuration Al.1 (as the most stable structure), valence level remains constant but conduction level shifted to lower energies (−2.49 eV) compared to the bare Al-C3N (−2.18 eV) so that the Eg of sheet decreased from 1.07 to 0.73 eV (Fig. 5a). It is well known that the Eg (or band gap in bulk materials) is a major factor determining the electrical conductivity of a material and there is a classic relation between them as follows [17]: 3
σ ¼ CT 2 qe ðμe þ μh Þ
−Eg ; ekT
ð2Þ
where σ is the electric conductivity, qe, μe, μh, and k are the charge of electron, mobility of electrons and holes, and the Boltzmann constant, respectively. It should be noted that this equation works only for semiconductors in which by increasing the temperature the electrical conductivity will be increased. According to the equation, smaller values of Eg at a given temperature lead to larger electric conductivity. The considerable change of about 0.34 eV (Table 1) in the Eg value, demonstrates the high sensitivity of the electronic properties of Al-C3N toward the CO adsorption on its surface. We think that the proposed sensor may benefit from some advantages including high sensitivity: the Eg of Al-doped C3N sheet is appreciably sensitive to the presence of acetone so that it increases by about 31.7 % upon the adsorption of CO molecule, short recovery time: the adsorption energy of CO molecule is not so large as to hinder the recovery of Al-C3N and therefore the sensor will possess short recovery times since
based on the conventional transition state theory, the recovery time can be expressed as [16]: τ ¼ ν 0 −1 expð−Ead =kTÞ;
ð3Þ
where T is the temperature, k is the Boltzmann’s constant, and ν0 is the attempt frequency. According to this equation, more negative Ead values will prolong the recovery time in an exponential manner.
Conclusion The adsorption of CO molecule on the pristine, B− and Al− doped C3N sheets are investigated using DFT calculations. It is found that CO molecule is only weakly adsorbed onto the pristine and B-doped C3N sheet with low negative Ead and large separation. The electronic properties of the pure Bdoped C3N sheet have very limited change upon the adsorption of CO molecule. In contrast, the CO molecule shows strong interactions with the Al-C3N sheets. The more negative Ead and charge transfers of CO on the Al-C3N sheet are expected to induce significant changes in the electrical conductivity of sheet. These results may help to seek appropriate chemical modification methods to widen the application fields of the C3N nanomaterials.
Ethical statement We guarantee that authorship is limited to those who have made a significant contribution to the attached manuscript. All coauthors have seen and approved the final version of this manuscript and have agreed to its submission for publication. We verify that the manuscript is entirely original, and all words of others has been appropriately cited and quoted. We verify that this manuscript has not been submitted to any other journal or primary publication. We agree to allow the journal to review the data if requested. The authors confirm having read Bethical responsibilities of authors^ and affirm that this paper is consistent with those guidelines. The authors disclose no conflicts of interest. This project has been financed by Iran Nanotechnology Initiative Council.
References 1. 2. 3. 4. 5. 6. 7.
Fukui K, Nakane M (1995) Sens Actuators B Chem 24/25:486–490 Iijima S (1991) Nature 354:56–58 Chelmecka E, Pasterny K, Kupka T, Stobiński L (2012) J Mol Model 18:1463–1472 Dinadayalane TC, Kaczmarek A, Łukaszewicz J, Leszczynski J (2007) J Phys Chem C 111:7376–7383 Beheshtian J, Ahmadi Peyghan A, Bagheri Z (2012) J Mol Model 19:255–261 Beheshtian J, Peyghan AA, Bagheri Z (2012) Monatsh Chem 143: 1623–1626 Turabekova MA, Dinadayalane TC, Leszczynska D, Leszczynski J (2012) J Phys Chem C 116:6012–6021
J Mol Model (2015) 21: 116 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Czerw R, Terrones M, Charlier JC, Blase X, Foley B, Kamalakaran R, Grobert N, Terrones H, Tekleab D, Ajayan P (2001) Nano Lett 1: 457–460 Hales J, Barnard AS (2009) J Phys Condens Matt 21:144203– 144209 Kim E, Chen C, Köhler T, Elstner M, Frauenheim T (2001) Phys Rev B 64:094107–094112 Sandré É, Pickard CJ, Colliex C (2000) Chem Phys Lett 325:53–60 Azevedo S, De Paiva R (2006) Europhys Lett 75:126–131 Peng N, Zhang Q, Chow CL, Tan OK, Marzari N (2009) Nano Lett 9:1626–1630 Sasaki I, Minami N, Karthigeyan A, Iakoubovski K (2009) Analyst 134:325–330 Ahmadi Peyghan A, Omidvar A, Hadipour NL, Bagheri Z, Kamfiroozi M (2012) Physica E 44:1357–1360 Beheshtian J, Peyghan AA, Bagheri Z (2012) Sens Actuators B Chem 171–172:846–852 Peyghan AA, Hadipour NL, Bagheri Z (2013) J Phys Chem C 117: 2427–2432 Peyghan AA, Soltani A, Pahlevani AA, Kanani Y, Khajeh S (2013) Appl Surf Sci 270:25–32 Rastegar SF, Peyghan AA, Hadipour NL (2012) Appl Surf Sci 265: 412–417 Beheshtian J, Peyghan AA, Bagheri Z (2013) Struct Chem 24:165– 170 Beheshtian J, Peyghan AA, Bagheri Z (2013) Struct Chem 24: 1565–1570
Page 7 of 7 116 22.
Baei MT, Peyghan AA, Bagheri Z, Tabar MB (2012) Phys Lett A 377:107–111 23. Beheshtian J, Ahmadi Peyghan A, Bagheri Z (2013) Struct Chem 24:165–170 24. Beheshtian J, Peyghan AA, Noei M (2013) Sens Actuators B Chem 181:829–834 25. Peyghan AA, Noei M, Yourdkhani S (2013) Superlattices Microstruct 59:115–122 26. Saha S, Dinadayalane TC, Murray JC, Leszczynska D, Leszczynski J (2012) J Phys Chem C 116:22399–22410 27. Becke AD (1993) J Chem Phys 98:5648–5652 28. Lee C, Yang W, Parr RG (1988) Phys Rev B 37:785–789 29. Grimme S, Antony J, Ehrlich S, Krieg H (2010) J Chem Phys 132: 154104–154119 30. Schmidt MW, Baldridge KK, Boatz JA, Elbert ST, Gordon MS, Jensen JH, Koseki S, Matsunaga N, Nguyen KA, Su S, Windus TL, Dupuis M, Montgomery JA (1993) J Comput Chem 14: 1347–1363 31. O’Boyle N, Tenderholt A, Langner K (2008) J Comput Chem 29: 839–845 32. Beheshtian J, Soleymanabadi H, Kamfiroozi M, Ahmadi A (2012) J Mol Model 18:2343–2348 33. Ahmadi A, Beheshtian J, Kamfiroozi M (2012) J Mol Model 18: 1729–1734 34. Liu B, McLean AD (1973) J Chem Phys 59:4557–4558 35. Boys SF, Bernardi F (1970) Mol Phys 19:553–566 36. Armaković S, Armaković S, Šetrajčić J, Jaćimovski S, Holodkov V (2014) J Mol Model 20:1–14
