Software

DFT study on the sensitivity of open edge graphene toward CO2 gas

Description
We investigated the sensitivity and reactivity of open edge zigzag and armchair graphene slabs toward CO2 gas, using density functional theory calculations. Different levels of theory (B3LYP, ωB97xD, and Minnesota 06 functionals) and basis sets were
Categories
Published
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  DFT study on the sensitivity of open edge graphene toward CO 2  gas Maziar Noei Department of Chemistry, Mahshahr Branch, Islamic Azad University, Mahshahr, Iran a r t i c l e i n f o  Article history: Received 13 May 2016Received in revised form20 June 2016Accepted 21 June 2016Available online 22 June 2016 Keywords: GrapheneComputational studySensorElectronic properties a b s t r a c t We investigated the sensitivity and reactivity of open edge zigzag and armchair graphene slabs towardCO 2  gas, using density functional theory calculations. Different levels of theory (B3LYP,  u B97xD, andMinnesota 06 functionals) and basis sets were employed. The results reproduce and explain the electricalbehavior which has been observed experimentally from the graphene upon the exposure to the CO 2  gas.We found that, unlike the pristine graphene, the open edge graphene layers may be promising chemicalsensor for the CO 2  detection in agreement with the experimental results. Zigzag edges are more sensitiveand reactive than the armchair types. ©  2016 Elsevier Ltd. All rights reserved. 1. Introduction Carbon dioxide (CO 2 ) is a colorless, odorless and non 󿬂 ammablegas which is used in large scales every years in the industry forcarbonated beverage, and for production of carboxylic acids car-bonates, urea, and carbon monoxide [1]. It is also a product of fossilfuel burning, ammonia production, grain fermentation, cellularrespiration, and petroleum operations [2,3]. Developing a CO 2 chemical sensor with high sensitivityand fast response is of a greatinterest, and is necessary for the atmospheric control, medicalapplications, monitoring global warming and indoor climate,  etc. [4 e 8]. The CO 2  detection is a challenging problem for researchersbecause of its inherent stability [9]. It is a non-polar molecule withzero electric dipole moment which makes it stable with smallreactivity. However, researchers work to develop a selective, highlysensitive, portable, cost effective, and low power consuming CO 2 sensor with short recovery times and fast response [10,11].Some of commonly employed materials for gas detection arepolymers and metal oxides [12 e 18]. Recently, by advent of nano-materials with hollow structures and particularly high surface/volume ratio the sensor industry has been revolted [19 e 24].Application of different nanostructured materials such as nano-sheets, nanotubes, nanocones, nanoclusters,  etc.  as gas sensors hasbeen widely investigated [25 e 31]. In particular, graphene which isa sheet of carbon with sp 2 hybridization has unique thermal,electrical,mechanical,optical,andchemicalpropertieswhichmakeit promising candidate for chemical sensor industry [32 e 35]. Atroom temperature, graphene has very large electron mobility, andtherefore, its electronic sensitivity is very high. However, it hasbeen shown that the pristine graphene is not proper for detectionof some gases, and numerous studies have focused on the sensi-tivity improvement by structural engineering such as chemicalfunctionalization, doping impurity atoms, decoration by transitionmetals, generating defects,  etc   [36 e 41].Recently an experiment has shown that the CO 2  gas can beef  󿬁 ciently detected by the graphene which is manipulated viamechanical cleavage [42]. This gas sensor has been fabricated bymechanical cleavage and unlike other solid-state gas sensors, thissensor can be operated under ambient conditions and at roomtemperature [42]. After the infusion of CO 2  gas, the electricalconductance of the graphene increases sharply at the beginning,then gradually levels off until it reaches a steady-state value. It hasbeen revealed that this type of graphene has short recovery time,fastresponsetime,highsensitivity,andlowpowerconsumption.Inthis work, following the experimental work [42], we will cut theedges of graphene and investigate the reactivity and sensitivity of these open sites toward CO 2  gas based on the density functionaltheory (DFT) calculations. Our main purpose is to  󿬁 nd and explainthe srcin of the electronic behavior of graphene which has beenobserved in the experimental work in the presence of CO 2  gas. Wehopeto 󿬁 ndarelation between thecalculatedelectronic propertiesand the experimentally observed electrical conductance changeupontheCO 2 adsorption.DFTmethodshavebeenpreviouslyworkspredicting increased reactivity of similar graphene-like cluster E-mail address:  noeimaziar@gmail.com. Contents lists available at ScienceDirect Vacuum journal homepage: www.elsevier.com/locate/vacuum http://dx.doi.org/10.1016/j.vacuum.2016.06.0180042-207X/ ©  2016 Elsevier Ltd. All rights reserved. Vacuum 131 (2016) 194 e 200  systems such as coronene, benzene and circumcoronene [43,44]. 2. Computational details A graphene slab constructing from 116 carbon atoms wasselected which its ends were saturated with 28 H atoms to reducethe boundary effects. The structural optimization, energy calcula-tions, natural bond (NBO) and frontier molecular orbitals analyseswere done at B3LYP level of theory, employing 6-31G* basis set asexecuted in the GAMESS suite of program [45]. The B3LYP has beenrevealed to be a consistent and commonly used level of theory inthe study of different nanostructures [46 e 53]. As there is no uni-versal exchange-correlation density functional for all propertycalculations, and functionals may give different results. Therefore,the effect of density functional was investigated. To this aim, fourMinnesota functionals of Truhlar group [54 e 56] including M06-L,M06, M06-2X, and M06-HF with 0, 27, 54, and 100% Hartree Fock(HF) exchange, respectively, were tested. To modeling the effect of graphene cleavage on the adsorption of CO 2  and the electronicsensitivityof graphene, weremoved some H atoms fromthe zigzagand armchair edges and studied the CO 2  adsorption behavior inthese sites.  2.1. The HOMO-LUMO gap is de  󿬁 ned as E g ¼ E LUMO  E HOMO  (1)where E LUMO  and E HOMO  are the HOMO and LUMO energies. Theelectronic sensitivity of the graphene is calculated by the ratiobetween the numerical difference of the initial HOMO-LUMO gapmeasured in the clean cage (E g 1: reference value) and the HOMO-LUMO gap calculated in the presence of CO (E g 2), with respect tothe reference value, (s  ¼  [(E g 2  E g 1)/E g 1]*100). The adsorptionenergy is de 󿬁 ned in the usual way as:E ad ¼ E (CO 2 /graphene)  E (graphene)  E (CO 2 ) (2)whereE(CO 2 /graphene)correspondstotheenergyof thegraphene(pristine of manipulated) in which the CO 2  was adsorbed on thesurface,E(graphene)istheenergyof theisolatedgraphene,E(CO 2 )is the energy of an isolated CO 2  molecule. 3. Results and discussion  3.1. The CO  2  adsorption on the pristine graphene The optimized graphene and its complex with CO 2  molecule areshown in Fig. 1. The calculated adsorption energy based on the B3LYP/6-31G* method is about   1.1 kca1/mol which is in goodagreement with the results of Lee et al. [57]. We also repeated thecalculations with  u B97xD functional. The adsorption energy isabout  4.1 kcal/mol at u B97xD level of theory with the same basisset. The more negative adsorption energy at the dispersion cor-rected u B97xD level shows that the main nature of this interactionis weak dispersion which B3LYP cannot predict it well. As a com-parison, it has been shown that nonpolar H 2  molecule is adsorbedon a coronene surface with an adsorption energy of    5.0 kJ/mol(~  1.2 kcal/mol) at dispersion corrected B3LYP level of theory [58].Wanno et al. have demonstrated that the adsorption energy of COmoleculeonthepristinegrapheneisabout  1.28kcal/molatB3LYPlevel of theory. They showed that doping the graphene with tran-sition metal atoms signi 󿬁 cantly increases the reactivity and sensi-tivity [59].The NBO charge transfer is about 0.002 e from the graphene tothe CO 2 . Also, the results of  Table 1 show that the electronicproperties HOMO, LUMO, and HOMO-LUMO are not sensitive to-ward CO 2  adsorption. For example, the HOMO-LUMO gap ischanged slightly by about less than 1% after the CO 2  adsorption atthe all density functionals. It can be concluded that the presence of CO 2  cannot be detected by the pristine graphene due to the weakinteraction and small charge transfer which is in good agreementwith experimental results. The molecular electrostatic potentialsurface (MEP) plot for the CO 2 /graphene in Fig. 1 indicates a cleardiscontinuityalong CO 2 -graphene surface. This discontinuous zoneis associated with less directional bonds that are usually weakconsistentwithlessnegativeadsorptionenergy.Also,itcanbeseenthat the CO 2  molecule has no effect on the electrostatic potential of the adsorbing site which is in agreement with the small NBOcharge-transfer.  3.2. The CO  2  adsorption on the edges of pristine graphene 3.2.1. Armchair edges For considering the reactivity and sensitivity of the armchairedges of graphene toward the CO 2  gas, we removed the hydrogenatoms (six H atoms) from one armchair edge of the graphene andoptimized the geometry (Fig. 2). There exist two possible spinstates for this structure including singlet or septet. In the singletstate, it has been assumed that all electrons are paired but in theseptet one, six electrons are unpaired on six carbon atoms whosehydrogen atoms are removed. Our calculations indicate that thesinglet state is more stable than the septet one by about 69.4 kcal/mol. It seems that two electrons on each of C e C bond tend to bepaired. Calculations show that after the hydrogen removing, theC e C bonds are shortened from 1.39 to 1.25 Å, con 󿬁 rming theelectron pairing.For CO 2  adsorption on the armchair edge, we explored severalinitial adsorption con 󿬁 gurations, and found four local minima asshown in Fig. 3. In the con 󿬁 guration  A  , a CO 2  molecule attachesfromits carbon atom to a carbon atomof the edge, with adsorptionenergy about 19.7 kcal/mol (Table 2). The formed C e C bond iscalculatedtobeabout 1.48ÅandanNBOchargeof0.156 e transfersfrom the graphene to the CO 2  molecule. We checked the spin statefor this con 󿬁 guration and found that the triplet state with twounpaired electrons is more stable than the singlet one by about31.4 kcal/mol. Our spin density calculations (Fig. 4) indicates thatthese electrons more localized on the adsorbing C carbon and itsneighboring carbon atom. This reveals that the CO 2  adsorptiondisturbs the C e C bond and weakens it, separating two pairedelectrons. Bond length calculations shows that after the CO 2 adsorption the C e C bond is enlarged by about 0.11 Å (Fig. 3).In the con 󿬁 guration  B , the CO 2  molecule attacks to a carbonatom of the edge from one O head and forms an O e C bond withlength of 1.39 Å, with adsorption energy about 36.6 kcal/mol(Table 2). An NBO charge of 0.109  e  transfers from the graphene tothe CO 2  molecule. Similar to the con 󿬁 guration  A  , the triplet state ismore stable than the singlet one by about 29.1 kcal/mol. Also, theC e C bond which adsorbs the CO 2  molecule is weakened and itslength is increased from 1.25 to 1.35 Å. In the con 󿬁 guration  C , theCO 2  attaches from its oxygen atoms to two carbon atoms of theedge, forming a heptagonal ether ring, with adsorption energy of 31.0 kcal/mol. Our calculations show that the stable spin state forthis con 󿬁 guration is quintet. This indicates that the adsorptionprocess weakens the adsorbing C e C bonds generating four un-paired electrons. The spin density plot in Fig. 4 indicates that thesefour electrons is distributed on the newly formed heptagonal ringand two adjacent carbon atoms.In the con 󿬁 guration  D , a CO 2  molecule attaches from its carbon M. Noei / Vacuum 131 (2016) 194 e  200  195  and an oxygen atom to the armchair edge of graphene, forming alactone group as shown in Fig. 3. The adsorption energy for thisprocess is about   8.5 kcal/mol, and a charge of 0.201  e  transfersfrom graphene to CO 2  molecule. The formed C e C and O e C bondsare calculated to be about 1.46, and 1.37 Å, respectively. Like thecon 󿬁 guration  C , the spin state for this con 󿬁 guration is triplet statewith two unpaired electrons. From the adsorption energy analysis(Table 2), it can be concluded that only lactone formation is ener-geticallyfavorableinaccordancewithpreviousreports[60]andtheothers are rejected because of positive adsorption energies.  3.2.2. Zigzag edges Similar to the case of armchair edge, we removed the hydrogenatoms (six H atoms) from one zigzag edge of graphene and opti-mized the geometry (Fig. 2). Unlike the armchair edge, our calcu-lations shows that six electrons on the edge carbon are unpairedandthespinstateisseptet.Singletstateislessstablethattheseptetone by about 91.2 kcal/mol. For the CO 2  adsorption on the zigzagedge, several initial adsorption con 󿬁 gurations were investigated,and like the case of armchair, four local minima are predicted asshown in Fig. 5. In the con 󿬁 guration  P , a CO 2  molecule attachesfromits carbon atom to a carbon atom of the edge, with adsorptionenergy of about   4.3 kcal/mol (Table 3). The formed C e C bond iscalculatedtobeabout1.48ÅandanNBOchargeof0.213 e transfersfrom graphene to CO 2  molecule. Unlick the case of armchair edge,the adsorption process of CO 2  form its C carbon on the zigzag edgeis energetically favorable based on the negative adsorption energy. Fig. 1.  Structural geometries of the pristine graphene and its complex with CO 2 . Also, molecular electrostatic potential surface of the complex. Color ranges, in a.u.: blue, morepositive than 0.015; green, between 0.015 and 0; yellow, between 0 and  0.015; red, more negative than  0.015. (For interpretation of the references to colour in this  󿬁 gure legend,the reader is referred to the web version of this article.)  Table 1 Adsorption energy (E ad , kcal/mol) for CO2 adsorption on the pristine graphene(Fig. 1) at different level of theory with 6-31G* basis set. Energy of HOMO, andLUMO, and HOMO-LUMO energy gap (E g ) in eV. The  D E g  indicates the change of E g after the adsorption process.Functional System E ad  E HOMO  E LUMO  E g  D E g B3LYP Graphene  e   3.83   3.38 0.45  e Complex   1.1   3.83   3.39 0.45   0.49 u B97xD Graphene  e   4.63   2.83 1.80  e Complex   4.1   4.64   2.82 1.81   0.48 Fig. 2.  Optimized structure of open ended zigzag and armchair graphenes. Distances in Å. M. Noei / Vacuum 131 (2016) 194 e  200 196  Itseemsthatsincealltheelectronsarepairedonthearmchairedge,this isomer is more stable and, thus, less reactive than zigzag edgewhich has radical character with unpaired electrons. Our calcula-tions con 󿬁 rm that the singlet armchair construction is more stablethan the septet zigzag one by about 66.1 kcal/mol.Similar to the con 󿬁 gurations  B ,  C , and  D  of the armchair con-structions (Fig. 3), we have predicted the corresponding  Q  ,  R  ,  S con 󿬁 gurations (Fig. 5),for the zigzagone, respectively. However, alltheadsorption energiesforzigzagcasesaremorenegativethan thecorresponding armchair ones. In the both cases, the lactone form-ing is the most favorable process. The positive adsorption energyof the con 󿬁 guration  Q   indicates that the adsorption of CO 2  from anoxygenheadisenergeticallyunfavorable.Calculationsshowthatallof the complexes for zigzag structure are septet and the adsorptionprocess does not help to electron pairing.  3.3. Electronic properties and sensing characteristic  One method to explore the sensitivity of an adsorbent toward achemical is calculating the electrical change after the adsorptionprocess [61 e 63]. In the computational studies, the electricalconductance has been frequently simulated by the change of HOMO-LUMO gap of semiconductor based on the following equa-tion [64]: s ¼ A T 3/2 exp(  E g /2kT) (3) Fig. 3.  Different con 󿬁 gurations of CO 2  adsorption at the armchair open end of graphene. Distances in Å.  Table 2 Adsorption energy (E ad , kcal/mol) for different states of CO 2  adsorption at the openarmchair edge of graphene (Fig. 3). Energy of HOMO, and LUMO, and HOMO-LUMOenergy gap (E g ) in eV. The  D E g  indicates the change of E g  after the adsorption pro-cess. The result obtained using B3LYP/6-31G*.System E ad  E HOMO  E LUMO  E g  D E g Armchair   3.97   3.46 0.51  e  A   19.7   4.31   3.96 0.35   16.4 B  36.6   3.99   3.62 0.37   13.9 C  31.0   4.20   3.77 0.43   7.9 D   8.5   4.34   3.59 0.76 24.6 Fig. 4.  Spin density plots for CO 2 /armchair graphene complexes (see Fig. 3). M. Noei / Vacuum 131 (2016) 194 e  200  197  where k is the Boltzmann ’ s constant and A (electrons/m 3 K 3/2 ) is aconstant. Herein, we compare the results of this equation withthoseofanexperimentalwork[42].Experimentally,Yoonetal.[42] have shown that the mechanically cleaved graphene sheets showsigni 󿬁 cant conductance increase when they were exposed to CO 2 gas. Also, they demonstrated that the response time of the gra-pheneislessthan10s,showingthatthedevicehasfastresponsetoCO 2  gas. They concluded that these types of graphene may bepromising sensor for CO 2  gas can operate at room temperature andunder ambient conditions.Our results indicates that the zigzag case in more appropriateenergetically for CO 2  adsorption compared to the armchair one.Here,wewillfocusontheeffectofCO 2 adsorptionontheelectronicproperties of the zigzag case and compare the results with theexperimental  󿬁 ndings. The results of  Table 3 show that the HOMOand LUMO levels of the zigzag-case graphene lie at   4.45 and  2.82 eV, respectively, producing a HOMO-LUMO gap of 1.63 eV.After the CO 2  adsorption, in the all complexes, the HOMO andLUMO are stabilized compared to the bare graphene. As this sta-bilization is more occurred for the LUMO level, the HOMO-LUMOgap is decreased in comparison to the bare graphene. Forexample,in thecon 󿬁 guration  S  which is the moststable among theall, after the CO 2  adsorption a new state is appeared within theHOMO-LUMO gap at  3.44 eV, narrowing the gap by about 36.3%.BasedontheEq.(3),theelectricalconductivityofthegrapheneslabwill exponentially increase by decreasing its HOMO-LUMO gapafter the CO 2  adsorption. It can be concluded that zigzag edgegraphene can show the presence of CO 2  gas by generating anelectrical signal. This  󿬁 nding is in good agreement with theexperimental results and con 󿬁 rms that the open edge graphenemay be promising candidate for CO 2  gas sensing.As shown in Table 4, for the con 󿬁 guration  S , the adsorptionenergy is slightly increased (became more negative) by increasingthe HF exchange percentage of density functional. The HOMO,LUMO,andHOMO-LUMOgapofthegrapheneanditscomplexwithCO 2  signi 󿬁 cantly depend on the kind of density functional. TheM06-L as a generalized gradient approximation (GGA) functional(with zero percentage of HF exchange) gives lower LUMO andhigher HOMO compared to the hybrid M06, M06-2X, and M06-HF.Therefore, itgivesasmallHOMO-LUMOgapwhich isinaccordancewith the fact that GGA functionals underestimate the gap [65]. Byadding and increasing the %HF exchange, the HOMO and LUMO arestabilized and destabilized, respectively, thereby enlarging theHOMO-LUMO gap. Although the absolute values of HOMO, LUMO,and HOMO-LUMO gap are depends on the density functional, allfunctionals show that the HOMO-LUMO gap is sensitive to thepresent of CO gas. The effect of the larger basis sets including 6-31 þ G* and 6-311 þþ G** on the adsorption energies and electronicproperties of the most stable  S  complex was investigated, at theB3LYP level of theory. Table 4 displays that enlarging the basis setslightly affects the, HOMO, LUMO levels, HOMO-LUMO gap, andadsorption energies. Also it has no signi 󿬁 cant effect on the HOMO-LUMO gap change upon the adsorption process. Fig. 5.  Different con 󿬁 gurations of CO 2  adsorption at the zigzag open end of graphene. Distances in Å.  Table 3 Adsorption energy (E ad , kcal/mol) for different states of CO 2  adsorption at the openzigzag edge of graphene (Fig. 5). Energy of HOMO, and LUMO, and HOMO-LUMOenergy gap (E g ) in eV. The  D E g  designates the change of E g  after the adsorptionprocess. The result obtained using B3LYP/6-31G*.System E ad  E HOMO  E LUMO  E g  D E g Zigzag  e   4.45   2.82 1.63  e P   4.3   4.62   3.33 1.29   20.7 Q   14.6   4.46   3.13 1.31   19.6 R    33.2   4.54   2.95 1.59   2.3 S   89.5   4.48   3.44 1.04   36.3  Table 4 Results of different methods for adsorption energy (E ad , kcal/mol) of CO 2  on thezigzag edge of graphene ( S , Fig. 5). Energy of HOMO, and LUMO, and HOMO-LUMOenergy gap (E g ) in eV. The  D E g  indicates the change of E g  after the adsorptionprocess.Method System E ad  E HOMO  E LUMO  E g  D E g M06-L/6-3G* Graphene  e   4.11   3.42 0.69  e Complex   83.3   4.18   3.82 0.36   48.6M06/6-3G* Graphene  e   4.74   2.84 1.90  e Complex   84.2   4.73   3.36 1.37   27.9M06-2X/6-3G* Graphene  e   5.34   2.56 2.78  e Complex   92.8   5.32   3.05 2.27   18.4M06-HF/6-3G* Graphene  e   6.47   1.98 4.49  e Complex   98.9   6.42   2.53 3.89   13.4B3LYP/6-31 þ G* Graphene  e   4.70   3.23 1.47  e Complex   86.2   4.73   3.73 1.01   31.6B3LYP/6-311 þþ G** Graphene  e   4.75   3.29 1.46  e Complex   84.3   4.79   3.79 1.00   31.7 M. Noei / Vacuum 131 (2016) 194 e  200 198
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks