1、固体表面化学的理论研究方法、模型和应用,吕鑫 2005.5.19,State Key Laboratory for Physical Chemistry of Solid Surfaces,厦门大学固体表面物理化学国家重点实验室,物理体系,物理、化学性质 （实验研究）,理论模型,理论方法,表面吸附是固体表面化学研究的一个中心问题，是一切表面化学现象的根源,固体表面化学的理论研究 方法、模型与应用,分类(理论方法、模型方法、物理体系） 层板模型方法与应用（Slab Model and Its Applications) 3. 簇模型方法及其应用（Cluster Model and Its App
2、lications),分类,理论方法分类：经典力学方法（MM, MD, MC) 、量子力学方法 (DFT, HF, CI) 、杂交方法（QM/MM, AIMD) 、其他半经验方法（AM1，PM3等） 模型分类：局域模型（簇模型方法）、周期性模型 3. 应用体系分类：共价体系、离子体系、金属体系、HB体系、VDW体系,2 Slab Model 2.1 First-Principle Method,量子化学问题均在于求解Schrdinger方程，对于大块固体，其Schrdinger方程表示为：H(Rm,rn) = E (Rm,rn) (1.1)Problem: H将是无限维的，上式很难求解。Sol
3、utions: Introducing some approximations.,A. Born-Oppenheimer Approximation: H(Rm) (rn) = E(Rm) (rn) (1.2)(核运动和电子运动分离) B. Single-Particle Approximation for Solving The Wavefunctions of Electrons(电子波函数的单粒子近似) C. Energy Band Theory (DFT) and Crystal Orbital Theory (HF) (see A. Gross, Surf. Sci. Rep. 19
4、98, 32, 291),Eec(n) exchange-correlation functional ec(n) exchange-correlation energy per particle,2.2 Density functional Theory, Kohn-Sham Equation,2.3 PSPW and Super Cell,Pseudopotentials for inner shells Plane-wave functions for valence shells Periodic Boundary Conditions and Super Cell Method fo
5、r Solid Slab Model for Solid Surface Car-Parrilleno Molecular Dynamics Method,Slab model: 4 atomic layers QM Method: DFT-GGA, PSPW,(see J. A. Rodriguez et al, J. Phys. Chem. B 2000, 104, 7439.),Example 1: SO2 on MgO(100) and CuMgO(100),Eads (kcal/mol) Cu-Free 2-O,O on Mg 8 1-S on O 11 3-S,O,O 21 Cu-
6、dopping 2-O,O 28 1-S on O 25,Bonding Modes of SO2 on MgO(100) Surface,其基本思想在于用一小簇原子组成的簇来类比表面, 其首要问题就是如何消除簇模型的“边界效应” 。,定域化,3. Cluster Model 3.1 Concept,Localization: Adams-Gilbert Equation,1) FC: 小簇C的Fock算符，包括簇C内的动能与各种相互作用能； 2) VSlr : 环境S对簇C的长程作用势，包括簇C与环境S间的电子电子、电子核、核电子、核核等四种库仑势； 3) VSsr : 环境S对簇C的短程作
7、用势，包括簇C与环境S间的电子交换势，反映出簇C与环境S间的轨道相互作用。 4) VSsr : 定域化势（亦称屏蔽势）,3.2 How to reach a successful cluster modeling?,关键问题： 怎样选择簇模型，使之与环境的短程作用尽可能小（必须注意，这并不意味着簇与环境的相互作用能很小）?怎样合理地考虑环境对簇的长程作用?,3.3 Schemes of Cluster Modeling,Simple Cluster Model Embedded Cluster Model (for ionic solids) Saturated Cluster Model (
8、for covalent solids) ONIOM Model (hybrid QM/QM or QM/MM method, readily for covalent solids),3.4 Simple Cluster Model,Simple cut-out ! Capacity? (may give qualitatively reasonable simulation results for VDW, HB, metal and ionic solids) How to make a reasonable cut-out? How to determine the electroni
9、c state of the cluster?,3.4.1 Cluster Model for Metal Surface,Dilemma: The larger, the more reasonable, but more expensive; the smaller, the more economical with higher accuracy, but less reasonable. Whats the way out?,“Surface molecule”,Convergence problem,*H/Ni(111):Ni19 (2.75kcal/mol) Ni22(15 kca
10、l/mol)Ni40(46.5kcal/mol) Ni4(55kcal/mol) Expt(63kcal/mol),Examples: 1) P.S. Bagus, et al , J. Chem. Phys., 78 (1983) 1390; 2) C.W. Bauschlicher Jr., Chem. Phys. Lett., 129(1986) 586; 3) P.E.M. Siegbahn et al., Chem. Phys. Lett., 149(1988) 265.*,Concept of “Metallic Atom”,Two kind of motions of elect
11、rons in bulk metal: 1) Localized ; 2) Delocalized-Free electrons. The atom in a bulk metal should be quite different from a simple atom, e.g. a) R(Cr-Cr):1.68 ( Cr2 ), 2.49 (bulk Cr) b) Pd atom: 4d10/ bulk Pd:(4d9.635sp0.37),Metallic Basis Functions,The attractive potential of a metallic atom is: m(
12、r) = -(Z*/r)exp(-kSr) vs a(r) = -Z*/r 1/kS - Thomas-Fermi Screening Length. Slater exponents: m = a + (1) With the help of Free Electron Theory, we have: = - (a(n)/n (inner shell) (2) = (a(n-1)/n (outermost valence-shell) (3),( N. Wang et al., J. Mol. Struct. (Theochem), 262(1992) 105.),Metallic m a
13、nd Atomic a of Co Atom.,UHF/STO-3G Calculations M-CO cluster,X. Xu et al., Surf Sci., 274 (1992) 378,Choice of Multiplicity,Metallic Cr: 3d5.244s0.76 (3d64s0 - 3d54s1) 3d64s0: 5, 3, 1; 3d54s1: 7, 5, 3,1 Note: UHF wavefunctions of a quintet are mixtures of wavefunctions from quintet and septet, rathe
14、r than a pure quintet.,Multiplicity Dependency in the UHF Calculations of Cr-CO,*Fe: 3d7.344s0.61/ (3d84s0 - 3d74s1)/(3),1 - 5,3,1 *Co: 3d8.374s0.63/(3d94s0 - 3d84s1)/(2) - 4,2,Metallic State Principle,M Mn M Mn M Ground State Bulk Metallic State Composition process Adiabatic decomposition proce,Som
15、e relative methods,Bond-Prepared State Principle (P.E. M. Siegbahn et al., Stockholm, 1988) DAM (Dipped Adcluster Model)(H. Nakatsuji, Kyoto, 1991)Many-Electron Embedding Theory (J.L. Whitten, 1980; 1987),Example 2: NO2/Au(111),X. Lu, J.Phys.Chem. A, 103 (1999) 10969.,NO2/Au2,Properties of Au2 clust
16、er and bulk Au (in eV),B3LYP calculations of NO2Au2,NO2 (2A1) + Au2 (1g) NO2Au2 (2A1) NO2 (2A1) + Au2 (3u) NO2Au2 (2B2),More Cluster Models: Au7 and Au12,Results omitted from here,3.4.2 Simple cluster model for ionic solids,How to cut out a cluster? Three Principles: Neutrality, Stoichiometry and Co
17、ordination Principles. Coordination number principle: 1) fewest dangling bonds at the edge of a cut-out; 2) maintain the stronger dative bonds within the cluster.,X. Lu et al., 1) Chem. Phys. Lett. 291(1998) 457; 2) Int. J. Quant. Chem. 73 (1999) 377; 3) Theor. Chem. Acc. 102(1999) 179.,CO/MgO,X. Lu
18、 et al., J. Phys. Chem. B, 105(2001) 10024.,C2O32- Surface Species,C3O42- Surface Species,3.5 Embedded Cluster Model for Ionic Solid,For ionic solid, VSsr can be replaced by VIsr :,For ideally ionic solid, VIsr would be negligible:,i.e. Simple embedded cluster model,3.5.1 Simple embedded cluster mod
19、el,A cut-out cluster is embedded into an array of point charges (always in formal charge) to represent the Madelung Potential of the ionic surroundings.,Example 4: CO/MgO(100) and NiO(100),See in G. Pacchioni et al. Surf. Sci. 255 (1991) 344.,Simple embedded cluster model for MgO(100) and NiO(100) (
20、 Mg(Ni) +2; O: -2 ),Demerits of simple embedded cluster model,Most of the ionic solids are not ideally ionic. Hence, the ionic charges are always fractional; the short range interaction between the cut-out cluster and its surrounding is seldom negligible. Way-out: Charge consistency Minimize the sho
21、rt range interaction.,Charge Consistence between the Embedded cluster and its PCC surrounding,Different embedding charge Q gives different C with different charges at the in-cluster atoms. Hence charge consistence between the embedding charges and the equivalent in-cluster atoms is essential and can
22、 be readily reached.,自洽条件 探讨,电荷自洽 偶极矩自洽,电荷密度自洽 偶极矩自洽,势自洽,SPC Embedded Cluster Model,X. Lu et al, J. Phys. Chem. B 103(1999) 2689.,Spherical Point Charges,Self-consistency of Charge Density,Cutout Cluster,SPC Embedding,Coordination Principle,Stoichiometry Principle,Nuetrality Principle,Example: SPC C
23、luster Models for MgO,X. Lu et al., J. Phys. Chem. B, 103(1999) 3373.,X. Lu et al., J. Phys. Chem. B, 103(1999) 5657.,NxOx+12- (X=1,2) Species Formed on MgO,3.6 Saturated Cluster Model for Covalent Solids,Saturating the radical-like dangling bonds at the edge of the cut-outs by using suitable satura
24、tors (e.g. H or other pseudoatoms). Widely employed in the study of covalent solid surfaces, e.g., Silicon, Diamond, Zeolite and so on. Examples shown below include Chemical Reactions on Silicon Surfaces.,Atomic arrangements of a) X(100)-21 (X= Si, Ge) and b) Si(111)-77 reconstructed surfaces.,buckl
25、ing,Three models describing the bonding within a buckled X=X dimer,Reconstruction of X(100) X= C, Si, Ge,In the solid state, each atom adopts sp3 hybridization and tetrahedral coordination.,Two widely used cluster models for X(100)-2x1 surface,X9H12 X15H16,2+2 addition of Alkene on Si(100),Possible
26、pathways,p-complex mechanism: FTIR spectra of dideuterioethylene/Si(100) suggested that the adsorption is stereospecific and stereoselective. (Liu et al., J. Am. Chem. Soc., 1997, 119, 7593.) Radical mechanism: STM images of 2-butene/Si(100) indicates the adsorption is not stereospecific, thought wi
27、th a high stereoselectivity of 98%. (Lopinski et al., J. Am. Chem. Soc., 2000, 122, 3548.),Controversy on the Mechanism,X. Lu, J. Am. Chem. Soc. 2003, 125, 6384,C4H4X(X=S,O) on Si(100)-2x1 surface,X. Lu et al, J. Phys. Chem. B, 105(2001) 10069.,Example: HN3 reaction with C(100)-2x1,X. Lu et al., Che
28、m. Phys. Lett. 343(2001) 212.,1,3-Dipolar Cycloadditions on C(100)-2x1,X. Lu et al., 1) J. Org. Chem. 67(2002) 515; 2) J. Phys. Chem. B, 106(2002) in press.,Example: NH3 on Si(111)-7x7,X. Lu et al, Chem. Phys. Lett. 355(2002) 365.,Profile of Energy Surface,Organic functionalization of Si(111),(X. Lu
29、 et al, J. Am. Chem. Soc. 2003, 125, 7923),Benzene/Si(111),Prediction: C4H2 on X(100),Possible pathways,Is the direct 4+2 pathway realistic? No!,The key point P4 on this pathway is indeed diradicaloid! Its UB3LYP wavefunction is 3.4 kcal/mol more stable than the RB3LYP one!,C4H2/Ge(100),PES,Is the d
30、irect 4+2 pathway realistic? No!,The key point P4b on this pathway is indeed diradicaloid! Its UB3LYP wavefunction is more stable than the RB3LYP one!,C4H2/Si(111): Prediction,PES,EONIOM= Ehow(A+H) Elow(A+H) + Elow(A+B),(K. Morokuma et al., J. Mol. Struct. (Theochem) 461-462(1999) 1.),3.7 ONIOM Mode
31、l,Adsorption of Methanol, Formaldehyde and Formic Acid on Si(100)-21 Surface,( see X. Lu et al., Phys. Chem. Chem. Phys. 3(2001) 2156.),Methanol,CCSD(T):B3LYP,Si2H4Si9H12,formaldehyde,Formic acid,b) ONIOM中最内层的C24簇,a) SWNT(10,0)片断,Sidewall functionalization by F and H (Bauschilicher, Chem. Phys. Lett
32、. 322(2000) 237.),ONIOM(B3LYP:UFF),F atoms appear to favor bonding next to existing F atoms. Hydrogenation of the sidewall of SWNT is probably endothermic.,Results:,Sidewall Functionalization of SWNT by1,3-Dipolar Cycloadditions,SWNT(5,5) 片断,ONIOM(B3LYP/6-31G*:AM1),Predicted Reaction Energies (kcal/
33、mol),1,2,3,1,3-DC of nitrile ylide with an olefin,X. Lu et al., 1) J. Phys. Chem. B, 106(2002), 2136; 2) J. Am. Chem. Soc., 2003, 125, 10459-10464.,3,3.8 Cluster modeling of electrodes,Charged cluster: Cluster Cluster in electric field. More realistic models are required.,Liao, M. et al. Int. J. Qua
34、nt. Chem., 67(1998), 175.,Concluding Remarks,Methods of simulation vary with and depend largely on the solids to be concerned. A simulation process is meaningless itself, unless certain physical criteria have been introduced to guarantee the consistence between the physical model and the real physical system. More significant is the scientific problem to be concerned. Simulation can be found everywhere nowadays.,
