【作者】 陈文斌；

【导师】 谭明秋；

【作者基本信息】 浙江大学 ， 凝聚态物理， 2008， 博士

【摘要】 第一性原理研究在许多领域已经取得了巨大的成功,使我们从微观的角度更深入的理解固体和表面特性的物理本质,并预测体系许多性质及变化的趋势。随着计算理论和方法的不断改进,第一性原理计算已经在适用性和准确度上都获得了极大地提高。但是,对于复杂固体和表面体系的第一性原理研究,尤其是对包含第一行元素或者d／f电子的系统,尚处于开始阶段或者尚未得到理想的结果。关于上述系统计算的理论和方法及其应用,仍处于不断发展和改进的过程中,并得到越来越多地关注。本人在攻博期间从事的第一性原理计算研究,以下是主要的研究内容以及重要结果：一、首先我们采用第一性原理的密度泛函理论方法计算了清洁Cu(110)表面和吸附O原子的(2×1)表面的原子结构,驰豫和电子结构,得到了表面结构参数的各种数据。分别计算了O原子在Cu(110)表面三个可能吸附位置吸附后的能量,并在其中能量最低的吸附位置上给出了各层原子的弛豫特性和态密度。结果表明O吸附后的Cu(110)表面有附加列(added-row)再构的特性,O原子吸附在最表层铜原子上方,与衬底Cu原子的垂直距离为0.016nm,以氧分子为能量基准的吸附能为-1.94 eV；同时由于Cu 3d-O 2p态的杂化作用使得在费米能级以下5.5-6eV的范围内出现了局域的表面态。计算得到清洁的和氧吸附的Cu(110)表面的功函数分别为4.51 eV和4.68 eV。电子态密度的结果表明：.在Cu(110)(2×1)表面O吸附的结构下,吸附O原子和衬底之间的结合主要是由于最表层Cu原子3d态和O原子2p态的相互作用。二、其次用密度泛函理论研究了氢原子的污染对于Ti(0001)表面结构的影响。通过PAW总能计算研究了p(1×1)、p(1×2)、(3(1/2)×3(1/2))R30°和p(2×2)等几种氢原子覆盖度下的吸附结构,以及在上述结构下Ti(0001)面fcc格点和hcp格点的氢原子吸附。本章所得结果表明：在p(1×1)-H、p(1×2)-H、(3(1/2)×3(1/2))R30。-H和p(2×2)-H几种H原子覆盖度下,以p(1×1)-H结构的单个氢原子吸附能为最大。在p(1×1)-H吸附结构下,由于氢原子吸附导致的Ti(0001)表面Ti原子层收缩的理论计算数值分别为-2.85%(hcp吸附)和-4.31%(fec吸附),因此实际上最有可能的情况是两种吸附方式都有一定的几率。而实验中观察到的所谓“清洁”Ti(0001)表面实际上是有少量氢原子污染的表面。本章的结果还说明,不同覆盖度和氢分压下氢原子吸附的污染对Ti(0001)表面结构有极大的影响,其表面的各种特性都会随覆盖度的不同而产生相应的变化。三、接下来用密度泛函理论研究了氧原子的吸附对于Ag(100)表面结构和电子态的影响。通过PAW总能计算研究了p(1×1)、c(2×2)和(2(1/2)×22(1/2))R45°等几种原子氧覆盖度下的吸附结构,以及在上述结构下Ag(100)表面的弛豫特性、吸附能量、功函数等一系列物理量。研究表明：在(2(1/2)×22(1/2)R45°-20吸附Ag(100)表面的情况下,每格两列就会缺失一列银原子,即产生了缺列再构,这导致了银原子层间的垂直距离的差异和水平位置的偏移。进一步的局域原子态密度计算表明：在Ag(100)表面(2(1/2)×22(1/2)R45°-20吸附的结构中,吸附氧原子和衬底银原子层之间的结合主要来源于表层银原子的4d态和吸附氧原子的2p态的强烈的轨道杂化。文章的后一部分,我们还模拟计算了在不同偏压和针尖高度的STM图像,为实验工作者研究该表面的STM图像提供了丰富的数据和理论支持。四、然后我们角基于PAW方法的密度泛函理论总能计算得到了Ni(100)／H20表面的原子结构和吸附位型。在p(3×3)结构的Ni(100)表面研究了水分子可能吸附的四种不同位置。研究表明：水分子最可能的吸附位置是T1位,并且其轴线向偏离表面法线方向呈现一定的弯折。由于其吸附能非常小(33meV),很显然就可以得到：Ni(100)表面H2O吸附是典型的物理吸附。五、论文的第五章用第一性原理的密度泛函理论研究了W(100)c(2×2)再构表面的表面弛豫以及STM图像和衬底偏压的关系。计算所得到的表面原子的畸变位移δ0.27A,畸变能△E为80.6 meV/atom,表面原子的弛豫分别为-7.6%(△d12／d0)和+0.8%(△d23/d0),功函数φ4.55 eV。STM图像模拟表明：由于表面原子沿[-110]方向的位移,会寻致出现平行于[110]方向的亮暗带状条纹。STM针尖突起所对应的并不是表面或次表面的钨原子,而是zig-zag型W原子链中线位置；而STM暗区对应于原子位置畸变形成的相邻zig-zag型W原子链中间区域。当衬底负偏压时,STM针尖典型起伏高度大约在0.08-0.13A之间；而当衬底正偏压时,针尖起伏高度在0.19-0.24A之间变化六、最后我们用第一性原理的密度泛函理论研究了氧吸附的Mo(110)p(2×2)表面的表面弛豫以及STM图像和衬底偏压的关系。氧吸附的Mo(110)p(2×2)表面的原子弛豫Ad12/d0为-2.10%,。清洁的Mo(110)表面的功函数φ约为4.65eV,表面能σ为0.019eV/atom,而氧吸附的Mo(110)p(2×2)表面Mo-O间距dMo-O为1.20 A,而氧原子的吸附能Eads约为-4.52eV。计算了在不同的偏压v下,针尖相对于表面钼原子分别为4.5A、5.0A和5.5 A时的STM图像。STM图像呈现出突起和凹陷交错排列的状态。突起就是表面氧原子所占据的位置,而凹陷所代表的就是最表层相邻Mo原子的中间地带。从STM针尖起伏高度来看,当衬底偏压为负时,针尖起伏高度随着偏压的降低而缓慢减小,.当偏压接近O时,针尖起伏高度迅速减小；而在正偏压下,STM的针尖起伏随着偏压的增加而减小,到0.7V左右时趋于相对平稳。比较在三种不同针尖高度下的起伏关系,可以看出针尖高度越高,其最大起伏也越大。更多还原

【Abstract】 First-principles calculations have made significant contributions to our understanding of condensed-matter systems and solid-state properties. We can do profound investigation to the physical nature about the properties of solid and surface, and predict many material qualities and their tendency of variations from microscopic view by calculations. With the rapid development of theories and methods, first principle methods are greatly improved in applicability and accuracy. But the first principles investigations for the complicated solids and surface systems, especially for the first-row elements or systems with d or f electrons, are still in initial stage or have not obtain ideal result. Recently, more and more interests have been attracted for study those systems. The main contents being studied and important results in this research presented as following:1.. First, the surface atomic geometry, structural relaxations, and electronic states of clean Cu(110) surface and oxygen-adsorbed Cu(110) (2×1)-O surface have been studied by using ab initio total energy calculations. The density functional calculations have been carried out for three possible adsorbed positions of oxygen and the most favorable one has been determined by total energy comparison. It is reveled that the added-row reconstruction is the most stable one with maximum adsorption energy in Cu(110) (2×1)-O surface and the adsorbed 0 atom is beyond the outmost surface Cu layer slightly. The adsorbates lie approximately 0.016nm above the outermost Cu layer and the hybridized band derived from Cu 3d-O 2p hybridization locates in the range of-5.5 --6.0 eV below EF. The adsorption energy of oxygen in this configuration is measured to be-1.94 eV with respect to oxygen molecule. The work functions of clean Cu(110) and oxygen-adsorbed Cu(110) (2×1) surface are calculated to be 4.51 eV and 4.68 eV, respectively. The surface electronic structures show that the cohesive effect between adsorbates and the substrate Cu atoms is essentially due to the Cu 3d-O 2p interaction.2. Then, we have studied the influence of hydrogen contamination on the atomic geometry of Ti(0001) surface by using the density-functional theory and the projector-augmented wave method. Based on the optimized structural parameters of hcp Ti from the PAW total energy calculation, the surface relaxations, surface energy and work function of clean Ti(0001) surface were calculated in the same way. The adsorption geometries and total energies of several coverages of hydrogen on Ti(0001) surface including p(1×1)、p(1×2)、((?))R30°, and p(2×2), were studied for the hcp and fcc site absorptions combined with the both sites occupation in p(1×1) structure. These results suggest that the Ti(0001) p(lxl)/H geometry has the largest energy gain among each conformation, so under the condition of low coverage and low H2 pressure, the most possible conformation is p(1×1)-H adsorption. The shrink of Ti(0001) surface with H contamination was-3.7% from available experiments and this work yields-2.85% for hcp and-4.31% for fcc adsorption geometries, respectively. It is deduced that the most possible adsorption configuration for a hydrogen contaminated Ti(0001) surface is a mixture of hcp and fcc adsorptions. For a clean Ti(0001) surface the surface contraction is calculated to be near-7.0% while the experimental measurement predicted-4.9%. This observation implies that even for a "clean" Ti(0001) surface there is still about 13.6% surface area covered with hydrogen adsorption. These results reflect that the hydrogen contamination could affect the Ti(0001) surface structure dramatically. Furthermore the present study could yield a conclusion naturally that the shrink of the Ti(0001) surface will be reduced with the increase of H atom adsorption below 1.0 ML.3. We also have investigated the influence of oxygen adsorption on the surface geometry and electronic properties of Ag(100) surface by using the density-functional theory calculations. The total energy calculations based on projector-augmented wave (PAW) method have been preformed to describe the adsorption geometry at several coverages of oxygen adsorption including p(1×1), c(2×2), and ((?))R45°, and a series of essential physical properties at these coverages on Ag(100) surface such as the surface relaxation, adsorption energy, work function and so on. The results presented in this work show that for an Ag(100) ((?))R45°-2O geometry, the most stable atomic structure is a type of missing-row reconstruction. Eventually, this structural change causes the various displacements of surface atoms which have been calculated in this work. The calculations on the local density of states reveal that in the Ag(100) ((?))R45°-2O geometry the cohesive effect between the adsorbed oxygen atoms and the substrate Ag layer is essentially due to the sufficient Ag4d-O2p orbital hybridization. Finally we have.simulated the STM images for several bias-voltages and tip heights, providing experiments abundant data and theoretical supports. 4. Our next work is that total energy calculations on the atomic geometry and adsorption properties of Ni(100)/H2O surface are performed by using the state-of-art density-functional theory in the implementation of the projector-augmented wave (PAW) method. The adsorptions of molecular water have been investigated in the surface periodicities of p(3×3) combined with 4 different sites. It is concluded that water adsorbed on the Ni(100) surface prefers the T1 site with its axis tilted away from the surface normal. In considering the small adsorption energy-33 meV, it is easily to conclude that adsorption behind in the Ni(100)/H2O surface has the type of physical adsorption.5. In chapter 5, the STM images and surface relaxations of reconstructed W(100) c(2×2) surface have been investigated by using the density-functional theory calculations. The distorted displacementδof tungsten atom along [110] presents as 0.27 A combined with a distorted energy of 80.6 meV/atom, and the relaxation of W(100) c(2×2) surface is calculated to be-7.6% for△d12/d0, and+0.8% for△d12/d0 respectively while the surface work functionΦis 4.55 eV. The calculated STM images of the W(100) c(2×2) surface display unusual features as follows:The protrusions in STM image along [110] axis lie in the middle of zig-zag chain of tungsten atoms while the dark regions in STM images correspond to the valley between neighbor zig-zag chains due to surface reconstruction. The typical corrugation of STM tip has been calculated between 0.08-0.13 A for the negative bias voltages while in the positive bias voltage regions the corrugation varies from 0.19 A to 0.24 A.6. At the end, we have investigated the STM images and surface relaxations of oxygen adsorption on Mo(110) p(2×2) surface by using the density-functional theory calculations. The relaxation of oxygen adsorption on Mo(110)p(2×2) surface is calculated to be-2.10% for△d12/d0 which is much smaller than that of clean one. The surface work functionΦof clean Mo(110) surface is 4.65 eV while the surface energyσrepresents 0.019eV/atom with the distance of Mo-O dMo-O 1.20 A and the adsorption energy Eads-4.52eV. The calculated STM images of the Mo(110)p(2×2) surface on different bias voltages display unusual features as follows:The protrusions in STM image along [111] axis is the oxygen atoms while the valley in STM images correspond to the region between neighbor Mo atoms. The typical corrugation of STM tip has been calculated to show that:higher the STM tip is, sharper the corrugation will be.更多还原