Adsorption and Thermal Reactions of H2O and H2S on Ge(100) - The

Dec 28, 2009 - Ki Yeol Byun , Pete Fleming , Nick Bennett , Farzan Gity , Patrick McNally , Michael Morris , Isabelle Ferain , Cindy Colinge. Journal ...
0 downloads 0 Views 4MB Size
Download PDF
J. Phys. Chem. C 2010, 114, 1019–1027

1019

Adsorption and Thermal Reactions of H2O and H2S on Ge(100) Tsung-Fan Teng,† Wei-Lin Lee,*,† Yi-Fu Chang,† Jyh-Chiang Jiang,‡ Jeng-Han Wang,*,† and Wei-Hsiu Hung*,† Department of Chemistry, National Taiwan Normal UniVersity, Taipei 116, Taiwan, and Department of Chemical Engineering, National Taiwan UniVersity of Science and Technology, Taipei 106, Taiwan ReceiVed: August 12, 2009; ReVised Manuscript ReceiVed: NoVember 23, 2009

The adsorption and thermal decomposition of H2O and H2S on Ge(100) were studied with temperatureprogrammed desorption (TPD) and X-ray photoelectron spectra (XPS), using synchrotron radiation. At 105 K, H2O molecules either adsorb molecularly or dissociate to form surface OH and H for exposures of all durations. Chemisorbed H2O dissociates to form surface OH on annealing to 270 K, whereas H2S dissociates to form surface SH and H on an initial exposure and further adsorbs molecularly on protracted exposure to the surface at 105 K. The calculation with density-functional theory (DFT) shows that H2S undergoes dissociative adsorption with a negligible barrier (2.1 kJ/mol) but H2O dissociates with a larger barrier (26.8 kJ/mol). On annealing to 550 K, surface OH mainly recombines with surface H to evolve H2O, but a small proportion of surface OH decomposes to form surface O and H. Most surface SH undergoes decomposition to form surface S and H at 520 K. According to a DFT calculation, surface OH has an activation energy for decomposition greater than that for recombination with surface H, whereas activation energies for decomposition and recombination of surface SH show the reverse order. Surface H resulting from the dissociation of H2O and H2S is thermally activated to combine and to desorb as H2 at 620 K. For H2S, a small proportion of surface H recombines with surface S to desorb as H2S. Finally, surface O and S are removable from the surface with desorption of GeO at 710 K and GeS at 695 K, respectively. Introduction Germanium (Ge) is a prospective semiconductor material for high-performance integrated circuits because electrons and holes both possess great mobility therein. This Ge semiconductor also has a direct transition of which the energy is only slightly greater than that of the indirect band gap. As a consequence, Ge material exhibits a greater absorption coefficient than Si, making Ge desirable for many optoelectronic or photovoltaic applications. The corresponding oxides are commonly considered to act as insulating or protective layers in electrical devices. H2O serves as a source of oxygen for the deposition of an oxide layer on a semiconductor surface via direct oxidation. The interaction of H2O with semiconductors has thus been an important issue because of fundamental and technological interests.1,2 Most research on the adsorption and thermal reactions of H2O on semiconductors has been performed on Si because this material predominates in the fabrication of electronic devices. As a Ge(100) surface resembles a Si surface both structurally and electronically, the difference of adsorption of H2O between Ge and Si warrants further investigation.3,4 The chemical reactions and functionalization on a Ge surface were reviewed by Buriak and Loscutoff/Bent.5,6 According to previous measurements of photoelectron and vibrational spectra, all H2O molecules adsorb dissociatively on Ge(100) at 300 K,7–9 whereas the results of scanning tunneling microscopy (STM) indicated that H2O adsorbs on Ge(100) at room temperature in two mannerssmolecular and dissociative.10 Some disagreement between measurements thus exists as to whether H2O adsorbs * Towhomcorrespondenceshouldbeaddressed.E-mail:[email protected]. Phone: +886-2-29350749, ext 307. Fax: +886-2-29324249. † National Taiwan Normal University. ‡ National Taiwan University of Science and Technology.

molecularly on Ge(100) at room temperature. H2 and GeO were observed as desorption products from the thermal decomposition of H2O on a Ge surface.3,11,12 A lack of stable Ge oxides of satisfactory quality is a drawback in the application of Ge for metal-oxide-semiconductor electronic devices. As widely reported, the deposition of a passivating layer can eliminate interface states, thereby improving the electrical performance of a Ge/dielectric stack in Gebased devices.13,14 Considerable effort has thus been devoted to investigate the formation of a passivating layer, as an alternative to Ge oxides, on Ge substrates. A surface S adlayer is commonly considered as a passivating and protective layer on the semiconductors. It has been demonstrated that a S adlayer becomes deposited on a Ge surface on immersion in a solution of (NH4)2S15,16 or on exposure to gaseous S2 or H2S.17,18 According to quantum-chemical calculations, the passivating layer formed from adsorption of H2S exhibits properties better than those from adsorption of elemental S.19 The treatment of a Ge surface with H2S was expected to remove effectively the surface states and to produce an electrically passivated surface that is a critical requirement for the use of Ge in electronic devices.20 Previous work with ultraviolet photoelectron spectra showed that H2S adsorbs dissociatively on a Ge surface at 300 K, and decomposes completely to form surface S at 550 K.21 An ordered (2×1)S/Ge(100) overlayer was produced on adsorption of H2S at 573 K, in which a surface S adatom was proposed to occupy a bridge site between two Ge atoms.22 H2 and GeS were the only desorption products obtained from this decomposition,3,18 but a detailed mechanism for the thermal reaction of H2S on a Ge surface is lacking. Both H2O and H2S are members of chalcogen hydrides and likely exhibit varied reactivities and reaction paths on a Ge

10.1021/jp907791f  2010 American Chemical Society Published on Web 12/28/2009

1020

J. Phys. Chem. C, Vol. 114, No. 2, 2010

surface.1 An improved understanding of the adsorption and thermal reactions of H2O and H2S on Ge might provide information about the limitations or merits of H2O and H2S as a source of O and S atoms, respectively, for the deposition of passivating layers. Mechanistic studies of the thermal reactions of H2O and H2S on a Ge substrate are thus fundamentally important for the fabrication and application of Ge-based devices. We report here an investigation of the adsorption and thermal reactions of H2O and H2S on a Ge(100) surface, utilizing temperature-programmed desorption (TPD) and X-ray photoelectron spectra (XPS); such examination elucidates the reaction paths of H2O and H2S on a Ge surface. The adsorption of H2O and H2S on Ge(100) has been studied with calculations based on density-functional theory (DFT).10,23,24 Previous calculations focused mainly on the initial adsorption near 295 K without further investigating the mechanisms of thermal reactions. In this work, we employed DFT calculations to study the potential-energy surfaces (PES) for the adsorption and thermal reactions of H2O and H2S on Ge(100). A doubledimer model was used to determine the activation energies of thermal reactions on Ge(100). We accordingly compare and correlate the experimental and theoretical results for the adsorption and thermal reactions of H2O and H2S on Ge(100).

Teng et al.

Figure 1. (a) XPS spectra of O 1s for a Ge(100) surface exposed to H2O at 105 K for various durations. Dots represent data collected after background subtraction; solid lines are fitted curves, and various components are shown with dashed lines. The photon energy used to collect these spectra is 610 eV.

Experiments and Calculations The experiments were conducted in an ultrahigh vacuum (UHV) chamber with a base pressure of 2 × 10-10 Torr. The UHV system was equipped with a quadrupole mass filter (EPIC, Hiden), a low-energy electron diffraction (LEED) analyzer, and an electron-energy analyzer (HA100, VSW). The thickness of our Ge(100) samples (Sb-doped n-type, 1-10 Ω cm) was 0.3 mm. The Ge sample was mounted on a Si sample of the same dimensions. A Ta strip (thickness 0.025 mm) was uniformly pressed between the Ge and Si samples with Ta foils at both ends, which were in turn mounted on a copper block. The sample was cooled to 105 K with liquid N2 via conduction through the copper block, and the Ta strip was resistively heated. The sample temperature was monitored with a K-type thermocouple spot-welded onto a thin Ta foil inserted between the Ge and Si samples. The surface was cleaned through Ar+ ion sputtering and annealed to 870 K; according to LEED, at 105 K the surface exhibited a c(4×2) pattern, so consists of surface dimers.25–27 The cleanliness of the Ge surface was verified with XPS measurements. Deionized water was purified with several freeze-pump-thaw cycles; H2S (99.5%, Matheson) was not further purified before being introduced to the Ge surface. During dosing, the partial pressures of H2O and H2S were controlled at 1 × 10-9 Torr. The sample was placed ∼3 cm before the doser to minimize contamination of the UHV system. XPS were measured at the HSGM beamline of the National Synchrotron Radiation Research Center, Taiwan; the angle of incidence of photons was 45° from the surface normal. Emitted photoelectrons were collected with an electron analyzer at an angle 10° from the surface normal in an angle-integrated mode. Collected spectra were numerically fitted with Voigt functions after Shirley background subtraction with a third-order polynomial to each side of the feature. The onset of photoemission from an Au foil attached to the sample holder indicated the Fermi level, corresponding to zero binding energy. A quadrupole mass filter served for analysis of desorption products in the TPD measurement. The mass spectrometer was enclosed in a differentially pumped cylinder, at the end of which was a skimmer with an entrance aperture of diameter 2.8 mm. For TPD measurement, the sample surface was placed about 2

mm before the aperture and in line of sight of the ionizer of the mass spectrometer; TPD scans were recorded on ramping the sample at a linear rate ∼1.5 deg/s. The calculations on the cluster model were performed with Gaussian03.28 A Ge15H16 cluster that contains two surface dimers served as a model of the Ge(100) surface. The hybrid Hartree-Fock/DFT method, B3LYP, includes Becke’s threeparameter nonlocal-exchange function with the correlation function of Lee, Yang, and Parr.29–32 This method was employed in the energetic calculation and geometrical optimization of intermediates (local minima, LM) and transition states (TS) without constraining the degree of freedom. The basis set applied in this calculation was the standard all-electron split-valence basis set designated 6-31G(d), involving a polarization dfunction on each non-hydrogen atom.31 All potential energies were calculated with unscaled zero-point-energy (ZPE) corrections at a B3LYP/6-31G(d) level. Details of the cluster calculations with a hybrid DFT using the model Ge(100) surface and two dimers are found elsewhere.34,35 Results and Discussion The chemical identity of a species on the Ge surface was characterized with XPS measurements. Figure 1 shows the spectra of O 1s recorded for a Ge(100) surface at 105 K with varied exposure to H2O. Two O 1s features are observed at 532.1 and 533.8 eV for all exposures, corresponding to two adsorption features. The latter feature is attributed to adsorbed H2O molecules. The former feature of smaller binding energy is assigned to surface OH, produced from dissociation of H2O according to H2O(g) f OH(ad) + H(ad). Previous measurements with a STM showed that OH and H resulting from dissociation of H2O bind covalently to a Ge-Ge dimer on Ge(100) in a configuration H-Ge-Ge-OH.10 According to the relative intensities of O 1s at 532.1 and 533.8 eV, adsorbed H2O dissociates only partially to form OH at 105 K. The intensity of the O 1s feature assigned to surface OH decreased with duration of exposure increasing beyond ∼35 s, but the intensity of O 1s due to H2O increased with exposure without saturation

J. Phys. Chem. C, Vol. 114, No. 2, 2010 1021

Figure 2. XPS spectra of Ge 3d for a clean Ge(100) surface before and after exposure to H2O and H2S at 105 K for 10 and 35 s, respectively. The photon energy used to collect these spectra is 250 eV.

extended to 100 s. The surface sites for chemisorption hence became saturated, and further H2O molecules physisorbed on the surface. Figure 2 shows a comparison of Ge 3d spectra for a Ge(100) surface at 105 K before and after exposure to H2O. The clean Ge(100)-c(4×2) surface consists of asymmetric Ge-Ge dimers; XPS spectra of 3d core-electrons were deconvoluted into several components corresponding to bulk, surface, and subsurface Ge atoms,17,36 but could be simplified into two components that pertain to bulk and surface Ge, so as to illustrate the adsorption of adsorbates on the Ge surface.12,37 The resolution of our spectral measurement is estimated to be about 200 meV, which may not be significant enough to differentiate the components of surface and subsurface Ge atoms in the resulting spectra. The spectra of 3d for the clean Ge surface were thus deconvoluted into two d5/2 features at 29.4 and 29.0 eV, attributed to bulk and surface Ge, respectively. Upon adsorption of H2O, the intensity of Ge 3d5/2 assigned to surface Ge decreased, and two new features appeared at 29.8 and 30.2 eV. Previous STM measurements and DFT calculations showed that molecular H2O and dissociative OH interact with the Ge surface via the O atom.10 The Ge 3d5/2 signals at 29.8 and 30.2 eV are attributed to surface Ge bonded to chemisorbed H2O and OH, respectively.11,37 On addition of OH, surface H formed through dissociative adsorption of H2O was bound to surface Ge. In fitting the Ge 3d region, we assumed the signal due to surface Ge bound to H to overlap that of bulk Ge because their binding energies are similar.37 Our XPS data indicate that H2O adsorbed dissociatively and molecularly on Ge(100) at 105 K, consistent with an interpretation of electron-energy-loss spectra (EELS).9,38 At a saturation coverage, H2O underwent mainly molecular adsorption on Ge(100) at 105 K, with the dissociative adsorption to a smaller extent. In contrast, previous workers deduced that H2O dissociated preferentially to form OH on Si(100) at a similar temperature, ∼100 K.23,24,39 Previous DFT calculations showed that the activation barrier for dissociation of H2O on Ge(100) is larger

Figure 3. TPD spectra of H2 (m/e ) 2), H2O (m/e ) 18), and GeO (m/e ) 90) collected from Ge(100) exposed to H2O at 105 K with various durations: (a) 10, (b) 20, (c) 35, and (d) 50 s.

than that on Si(100).23,24 The smaller extent of H2O dissociation on Ge(100) is attributed to a weak Ge-O bond that draws the adsorption system toward a dissociated state less effectively than the strong Si-O bond on Si(100).42 The dissociation of H2O on Ge(100) thus exhibits an activation barrier, whereas the analogous reaction on Si(100) is unactivated. In detecting possible products evolved during thermal decomposition of H2O on Ge(100), we observed several possible fragments. The desorption products are H2 (m/e ) 2), H2O (m/e ) 18), and GeO (m/e ) 90). Figure 3 depicts the TPD scans as a function of duration of exposure of H2O. A desorption state of H2 is observed at 625 K and attains a maximum at an exposure duration of ∼35 s. This desorption temperature is similar to that for adsorption of H2 on a Ge(100) surface;43,44 this desorption is hence attributed to a combination of surface H resulting from dissociation of H2O. Desorption of H2O is observed with signal maxima at 175, 220, and 550 K. The former two features are observed only for prolonged exposures and are thus attributed to desorption of physisorbed and chemisorbed H2O molecules, respectively. The third feature is attributed to recombination of surface OH and H, as OH(ad) + H(ad) f H2O(g). In contrast, surface OH preferentially underwent decomposition to form surface O and H on Si(100), rather than recombination to evolve molecular H2O.45 On Ge(100), the surface O produced from the decomposition of H2O was removed from that surface through desorption of GeO. The desorption temperature of GeO is less than that of SiO on Si(100) by ∼200 K.45 The GeO and SiO desorptions require breaking their bonds to the underlying Ge and Si atoms, i.e., Ge-GeO and Si-SiO. The difference between the desorption temperatures of GeO and SiO is explained simply by the fact that the energy of the Ge-Ge bond (263 kJ/mol) is less than that of Si-Si (326 kJ/mol).46 The thermal evolution of XPS spectra was used to characterize the variation of surface composition during the thermal decomposition of H2O on Ge(100), and was correlated with TPD results to elucidate the reaction intermediates. Figure 4 shows O 1s spectra for a Ge surface at 105 K exposed to H2O for 50 s, and subsequently annealed to various temperatures. All XPS spectra were recorded for samples at 105 K after a sample

1022

J. Phys. Chem. C, Vol. 114, No. 2, 2010

Teng et al.

H2O(ad) f OH(ad) f H(ad)

(170 K)

OH(ad) + H(ad) + H2O(g) (470-600 K) (major) OH(ad) f O(ad) + H(ad) (470-600 K) (minor) 2H(ad) f H2(ad)

(600-650 K)

O(ad) + Ge(s) f GeO(g)

Figure 4. XPS spectra of O 1s for a Ge(100) surface at 105 K exposed to H2O for 50 s and subsequently heated to the indicated temperatures.

was heated to a desired temperature and cooled abruptly. Upon annealing the sample to 170 K, the O 1s component at 533.8 eV due to physiorbed H2O disappeared because of molecular desorption as shown in the TPD data. The intensity of the O 1s signal at 532.1 eV due to surface OH increased, indicating the decomposition of chemisorbed H2O molecules to form more surface OH species. At 270 K, all chemisorbed H2O desorbed intact or dissociated to form OH and H. On further annealing to 500 K, a new O 1s feature appeared at 530.5 eV, attributed to surface O formed on dehydrogenation of surface OH. According to the previous DFT calculations, the O atom was proposed to reside on a bridge site between two Ge atoms.24 The temperature of OH decomposition on Ge(100) is similar to that observed on Si(100);1 this observation disagrees with an expectation from a calculation that the Si-H bond does not dissociate at temperatures below 670 K.20 Upon annealing to 600 K, the signal of O 1s at 532.1 eV due to OH disappeared and only O was left on the surface. The total integrated intensity of the O 1s signal decreased significantly because most OH recombined with surface H to desorb H2O in this temperature range as shown by the TPD data. If the integrated intensity is proportional to the coverage of O-containing species on the surface, about 70% of surface OH underwent recombinative reaction. When the sample was annealed to 700 K, the intensity of O 1s further decreased because the O adatom evolved from the surface with desorption of GeO. The surface O was removed completely from the surface at 770 K and a clean Ge surface was regained.

(650-750 K)

(2)

(3) (4) (5)

Both H2O and H2S are hydrogen chalcogenides (group 16 hydrides). For comparison with H2O, the adsorption and reaction of H2S on Ge(100) was investigated. Figure 5 shows XPS data of S 2p recorded for a Ge(100) surface at 105 K exposed to H2S with various durations. At small exposures, only a S 2p3/2 signal is observed at 162.6 eV; an additional S 2p3/2 feature appears at 164.0 eV for a duration of exposure greater than ∼25 s. The former feature is attributed to surface SH, whereas the latter feature is due to adsorbed H2S molecules. The surface SH is produced from the decomposition of H2S as observed for H2O. As shown in Figure 2, two features of Ge 3d5/2 develop at 29.8 and 30.0 eV upon the adsorption of H2S at 105 K, which indicate bonding of surface Ge atoms with S atoms of chemisorbed H2S and SH, respectively. As shown in Figure 5, no significant change in the S 2p spectrum is observed on exposure prolonged beyond ∼50 s; the chemisorption of H2S is hence nearly saturated at an exposure for ∼50 s and the physisorption coefficient of H2S is small, unlike for H2O that effectively physisorbs on the surface. The interaction between physisorbed H2S and a Ge surface saturated with chemisorbed H2S is much weaker than the analogous interaction observed for the adsorption of H2O that exhibits a strong hydrogen bond to the chemisorbed H2O covered surface. Figure 6 shows TPD spectra for the Ge surface exposed to H2S at 105 K with varied durations. The desorption products are H2 (m/e ) 2), H2S (m/e ) 34), and GeS (m/e ) 106). A desorption state of H2 is observed at 620 K, similar to that for adsorbed H2O, and is due to the recombinative reaction of surface H. Desorption of H2S is observed with signal maxima at 120, 160, 530, and 620 K. The first two

On the basis of the TPD and XPS data, the adsorption and decomposition of H2O are summarized according to the following reactions.

H2O(g) f H2O(ad) or OH(ad)+H(ad)

(105 K)

(1)

Figure 5. XPS spectra of S 2p for a Ge(100) surface exposed to H2S at 105 K for various durations.

J. Phys. Chem. C, Vol. 114, No. 2, 2010 1023

Figure 6. TPD spectra of H2 (m/e ) 2), H2S (m/e ) 34), and GeS (m/e ) 106) collected from Ge(100) exposed to H2S at 105 K with various durations: (a) 10, (b) 20, (c) 50, and (d) 100 s.

features observed only for prolonged exposures are attributed to desorption of physisorbed and chemisorbed H2S molecules, respectively. The third feature is attributed to the recombination of surface SH and H (SH(ad) + H(ad) f H2S(g)), analogous to the major channel of reaction 3 in the case of H2O. The desorption temperature of H2O due to recombination is greater than that of H2S by ∼20 K, reflecting the energy of bond O-Ge, 659 kJ/mol, being greater than that of S-Ge, 551 kJ/mol. The latter desorption state occurs at a temperature at which desorption of H2 is observed. We propose that surface H is thermally activated at ∼620 K and reacts subsequently either with another surface H to desorb molecular H2 or with surface S to desorb H2S (S(ad) + 2H(ad) f H2S(g)). The latter reaction involves diffusion and combination of absorbed surface atoms, which is not observed between surface H and O in the case of H2O. The desorption of GeS is observed with an intensity maximum at 695 K. Analogous to the reaction on the Ge surface, desorption of SiS was also generally observed when a Si surface was exposed to sulfurcontaining compounds, e.g., elemental sulfur and alkanethiols.47–50 The desorption temperature of GeS is less than that of SiS by ∼100 K. These desorption temperatures correlate with the strengths of their bonds to the underlying substrate atoms. Conversely, their desorption temperatures differ for the same reason as for the desorption temperatures of GeO and SiO described above. The desorption temperature of GeS is only slightly less than that of GeO; desorption of GeS and GeO is hence determined mainly by their bonds to the underlying Ge atoms; that is, Ge-GeS and Ge-GeO. GeS and GeO thus exhibit similar thermal stabilities because both desorptions involve a breakage of the Ge-Ge bond. Figure 7 shows S 2p3/2 spectra for the Ge surface saturated with H2S at 105 K for 60 s and subsequently annealed to various temperatures. The S 2p3/2 component at 164.0 eV disappeared on annealing the sample to 220 K because the adsorbed H2S desorbed intact or dissociated to form SH and H. On further annealing to 500 K, a new S 2p3/2 feature appeared at 161.8 eV, attributed to surface S formed on further dehydrogenation of surface SH. Upon annealing to 570 K, all SH species disappeared and only surface S was left on the surface, consistent

Figure 7. XPS spectra of S 2p for a Ge(100) surface at 105 K exposed to H2S for 60 s and subsequently heated to the indicated temperatures.

with previous EELS data.22 The total integrated intensity of S 2p gradually decreased in the temperature range 500-570 K, consistent with the desorption feature of H2S at 530 K shown in Figure 6, which occurred from the recombination of surface SH and H. On the basis of the integrated S 2p intensities, about 30% of SH recombined with surface H to desorb as H2S. The intensity of S 2p decreased further on annealing the sample to 670 K, corresponding to another desorption feature of H2S at 620 K. As proposed above, a small fraction of surface S reacts with two surface H atoms to desorb as H2S. The S 2p intensity completely disappeared at 770 K. All surface S was removed from the surface on desorption of GeS. Our XPS data show that H2O mainly adsorbed molecularly on Ge(100) at 105 K, whereas H2S preferentially dissociated to form surface SH on the initial exposure, and H2S molecularly adsorbed to only a small extent at large exposures. H2S is hence more active than H2O on the Ge surface, consistent with their relative acidities. Surface OH preferentially underwent recombination to desorb as H2O, whereas SH predominantly dissociated to form surface S and H. A desorption feature of H2S observed at 620 K is attributed to the recombination of surface S and H adatoms. This reaction is proposed to be initiated on thermal activation of surface H because this reaction temperature is identical with the desorption temperature of surface H. This observation might apply to the removal of surface S from the Ge surface in the cleaning procedure on dosing H atoms at ∼620 K. In contrast, the analogous reaction does not occur for the case of surface O adatoms. We performed DFT calculations to elucidate the various thermal reactions of H2O and H2S on Ge(100) as described above. Figure 8 shows the PES for the adsorption and dissociation of H2X (X ) O or S) on Ge(100); the potential energy is referred to the reactants, H2X + Ge(100). Figure 9 depicts the structures of possible intermediates and the corresponding transition states formed during the dissociation of H2O on Ge(100). The structures of intermediates and the transition states for the adsorption and decomposition of H2S are similar to those

1024

J. Phys. Chem. C, Vol. 114, No. 2, 2010

Teng et al.

Figure 8. Potential-energy surfaces for adsorption and reactions of (a) H2O and (b) H2S on Ge(100). XLM and XTS denote reaction intermediates and transition structures of H2X (X ) O and S) on Ge(100), respectively. The numbers in parentheses specify the potential energies (kJ/mol) of X LM and XTS. All potential energies of XLM and XTS are referred to the H2X molecule separated from the cluster surface.

obtained for the case of H2O (Supporting Information, Figure S1). The geometric parameters and relative energies of the intermediates and transition states are listed in Table 1. As the bond lengths between the Ge atoms on the surface and at the subsurface vary negligibly during surface reactions, these values are omitted from this table. As shown in Figure 8, adsorbed H2X can undergo the first dissociation of hydrogen via three routes: 1,2, 1,3, and 1,4 dissociation, corresponding to the dissociated H occupying Ge(2), Ge(3), and Ge(4) sites, respectively. The first route results in a surface configuration of HX(a)1 + H(a)2 (XLM2) in which HX and H each adsorb on Ge of a dimer. In the second route, HX and H adsorb on Ge atoms on two neighboring dimers, HX(a)1 + H(a)3 (XLM3). The smaller dissociation barrier for the (1,3)-dissociation is attributed to the fact that the fivemember ring structure of XTS2 is more stable than the fourmember ring of XTS1 in the 1,2-dissociation. The 1,4dissociation routes for both H2O and H2S systems are energetically less feasible because of the larger separations between the Ge(1) and Ge(4) atoms.

The adsorbed HX (XLM2 and XLM3) can further dissociate to liberate a H atom to neighboring Ge sites via two possible routes: HX(a)1 + H(a)2 (XLM2) f XTS3 f X(a)1 + H(a)2 + H(a)3 (XLM4) and HX(a)1 + H(a)3 (XLM3) f XTS4 f X(a)1 + H(a)2 + H(a)3 (XLM4). The transition states XTS3 and XTS4 exhibit barriers larger than the transition states XTS1 and XTS2 for the first dissociation of hydrogen in H2X, because the adsorption energy of HX is greater than that for H2X. In contrast, the dissociation of HX via XTS4 in the 1,3-dissociation route has a smaller activation barrier than that via XTS3 in the 1,2dissociation route. In addition, the atomic intermediate state X LM4 can undergo rearrangement via XTS5, X(a)1 + H(a)2 + H(a)3 (XLM4) f XTS5 f X(a)12 + H(a)3 + H(a)4 (XLM5). This rearrangement consists of transfers of surface X from the ontop site to the bridge site and surface H from the Ge(2) atom to the Ge(4) atom. The corresponding activation barrier is similar to those of dissociation of HX via XTS3 and XTS4. The bridgebonded X adatom seems to be formed readily upon the dissociation of HX.

J. Phys. Chem. C, Vol. 114, No. 2, 2010 1025

Figure 9. Optimized geometric structures of intermediates (OLM) and transition structures (OTS) as indicated in Figure 8a, obtained from the calculation at the B3LYP/6-31G(d) level.

An additional reaction path for dissociation of HX (XLM2) is considered via XTS6, HX(a)1 + H(a)2 (XLM2) f XTS6 f X(a)12 + H(a)1 + H(a)2 (XLM6). This reaction results in the formation of an X adatom bridged between the Ge(1) and Ge(2) sites. However, the corresponding activation barrier is greater than that for two other dissociation routes via XTS3 and XTS4. The greater barrier is attributed to the unstable structure of XTS6 in which one X and one H are bonded to Ge(1) with a strongly repulsive electron interaction as shown in Figure 9. The greatest reaction barrier of XTS6 on the entire PES implies that X-H dissociation is least likely to occur via this route. As shown in Figure 8, the reaction is less exothermic (or more endothermic) and the activation energy is greater for each elementary step in system H2O/Ge(100) than for system H2S/ Ge(100). The distinction is associated mainly with the greater dissociation energy of the O-H bond, ∼428 kJ/mol, than that of S-H, ∼344 kJ/mol.46 In addition, the computed PES for the adsorption and dissociation of H2O and H2S is consistent with the experimental results. H2S undergoes dissociative adsorption

with a negligible activation barrier via STS2, 2.1 kJ/mol, but H2O proceeds to dissociation with a higher activation energy of 26.8 kJ/mol. This calculation is consistent with XPS data that show H2S to dissociate preferentially to form surface SH and H, whereas H2O mainly undergoes molecular adsorption on Ge(100) at 105 K. The TPD data show also that OH tends to undergo recombination to desorb as H2O, whereas SH predominantly dissociates to form surface S and H. This distinction between OH and SH is explained with the calculated results showing the energy barriers of recombination and dissociation. The activation barriers for the recombination of OH and H via routes OLM2 f OTS1 f OLM1 and OLM3 f OTS2 f OLM1 are 151.8 and 101.5 kJ/mol, respectively, which are significantly smaller than those for dissociation of OH via routes OLM2 f OTS3 f OLM4 (183.8 kJ/mol) and OLM3 f OTS4 f OLM4 (227.3 kJ/mol). The recombination of OH with surface H thus predominates in the thermal reaction over the dissociation. In contrast, the activation barriers for SH dissociation via routes SLM2 f STS3

1026

J. Phys. Chem. C, Vol. 114, No. 2, 2010

Teng et al.

TABLE 1: Optimized Bond Distances (Å) and Relative Energies (kJ/mol) of Intermediates and Transition Structures for Reactions of (a) H2O and (b) H2S on Ge(100)a H2O/Ge(100) part a Ge(1)-Ge(2) Ge(1)-O Ge(2)-O Ge(1)-H Ge(2)-H Ge(3)-H Ge(4)-H S-H(1) S-H(2) ∆E

2.348

O

LM1

2.431 2.144

O

LM2

2.421 1.808

O

LM3

O

LM4

2.449 1.812

2.480 1.681

1.519

1.533 1.519

1.523

O

LM5

2.374 1.860 1.861

O

LM6

1.812 1.797 1.525 1.524

O

2.403 1.943

0.987 0.976 -32.9

0.967 -229.5

0.967 -207.7

-201.9

-188.9

-62.7

H2S/Ge(100) 2.348

S LM1 2.454 2.502

S LM2 2.421 2.282

S LM3 2.467 2.293

S LM4 2.442 2.133

S LM5 2.362 2.306 2.306

S

1.518

1.527 1.517

TS2

O

TS3

O

TS4

O

TS5

2.449 1.927

2.451 1.706

2.399 1.736

2.473

1.733

1.529 1.543

1.637 1.517

1.690 1.517 2.178

1.357 0.976 -77.7

1.356 0.972 -106.2

2.306 -46.0

1.603 19.6

8.4

S TS1 2.906 2.286

S TS2 2.459 2.438

S TS3 2.428 2.172

S TS4 2.414 2.218

S TS5 2.458 2.118

1.957

1.526 1.577

1.686 1.516

1.691 1.516 1.998

1.572 1.351 -111.2

2.210 -154.2

1.955 -98.2

1.807

1.526 1.526

0.942 0.942 0.0

O

TS1

O

TS6

2.496 1.698 2.980 1.590 1.530

2.810 110.4

part b Ge(1)-Ge(2) Ge(1)-S Ge(2)-S Ge(1)-H Ge(2)-H Ge(3)-H Ge(4)-H S-H(1) S-H(2) ∆E a

1.527 1.347 1.347 0.0

1.424 1.352 -113.3

1.352 -244.5

1.351 -221.5

-224.9

LM6

2.272 2.254 1.541 1.540

1.585

1.529 1.529 -318.9

-160.5

2.796 1.355 -94.9

-94.1

S TS6 2.585 2.013 3.040 1.485 1.519

3.030 -12.5

The relative energies, ∆E, are set to zero for a H2X molecule separate from the cluster surface.

f SLM4 (90.3 kJ/mol) and SLM3 f STS4 f SLM4 (123.3 kJ/ mol) are smaller than those for the recombination between SH and H via routes SLM2 f STS1 f SLM1 (149.6 kJ/mol) and S LM3 f STS2 f SLM1 (110.3 kJ/mol). This difference of the activation energies between OH and SH on Ge(100) can be attributed to the relative strengths of O-H and S-H bonds as well. Conclusion Our TPD and XPS data elucidate the thermal reactions of H2O and H2S on Ge(100). The adsorption and decomposition of H2O and H2S proceed with essentially similar mechanisms although their corresponding reactions occur at separate temperatures. At 105 K, H2O adsorbs molecularly and dissociatively on the Ge(100) surface, whereas H2S preferentially dissociates to form surface SH and H. The DFT calculation shows that H2O has a barrier to dissociation greater than that of H2S. On annealing the surface to 270 K the chemisorbed H2O molecules readily dissociate to form OH, which mainly combines with surface H to desorb as H2O at 550 K. A small proportion of surface OH further decomposes to form surface O adatoms. In a similar temperature range, surface SH mostly dissociates to form surface S and H whereas a small fraction of SH undergoes recombination to desorb as H2S. The PES calculation also shows that surface OH has a greater barrier to dissociation than for recombination, whereas surface SH has the opposite trend. Different from surface O, on annealing the sample to 620 K, a small portion of surface S reacts with surface H to desorb as H2S. Surface O and S adatoms are completely removed at 770 K through desorption of GeO and GeS, respectively. Our data indicate that a passivating S adlayer can be prepared on Ge(100) through adsorption of H2S at temperatures above 570 K. The surface S is thermally stable at 600 K in the presence of surface H and is sustainable to 670 K in the absence of surface H. Acknowledgment. The National Science Council (grant nos. NSC95-2113-M-003-012-MY3 and NSC97-2113-M-009-003)

supported this work. We are grateful also for the support of the National Synchrotron Radiation Research Center and the National Center for High-performance Computing. Supporting Information Available: Optimized geometric structures of intermediates (SLM) and transition structures (STS) as indicated in Figure 8b, obtained from the calculation at the B3LYP/6-31G(d) level. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Henserson, M. A. Surf. Sci. Rep. 2002, 46, 1. (2) Thiel, P. A.; Madey, T. E. Surf. Sci. Rep. 1987, 7, 211. (3) Cohen, S. M.; Yang, Y. L.; Rouchouze, E.; Jin, T.; D’Evelyn, M. P. J. Vac. Sci. Technol. A 1992, 10, 2166. (4) Duke, C. B. Chem. ReV. 1996, 96, 1237. (5) Buriak, J. M. Chem. ReV. 2002, 102, 1271. (6) Loscutoff, P. W.; Bent, S. F. Annu. ReV. Phys. Chem. 2006, 57, 467. (7) Kuhr, H. J.; Ranke, W. Surf. Sci. 1987, 187, 98. (8) Larsson, C. U. S.; Flodstro¨m, A. S. Phys. ReV. B 1991, 43, 9281. (9) Papagno, L.; Caputi, L. S.; Frankel, D.; Chen, Y.; Lapeyre, G. J. Surf. Sci. 1987, 189/190, 199. (10) Jung, S. J.; Lee, J. Y.; Hong, S.; Kim, S. J. Phys. Chem. B 2005, 109, 24445. (11) Larsson, C. U. S.; Flodstro¨m, A. S. Vacuum 1991, 42, 297. (12) Larsson, C. U. S.; Flodstro¨m, A. S.; Karlsson, U. O.; Yang, Y. J. Vac. Sci. Technol. A 1987, 7, 2044. (13) Dimoulas, A.; Mavrou, G.; Vellianitis, G.; Evangelou, E.; Boukos, N.; Houssa, M.; Caymax, M. Appl. Phys. Lett. 2005, 86, 32908. (14) Spiga, S.; Wiemer, C.; Tallarida, G.; Scarel, G.; Ferrari, S.; Seguini, G.; Fanciulli, M. Appl. Phys. Lett. 2005, 87, 112904. (15) Anderson, G. W.; Hanf, M. C.; Norton, P. R.; Lu, Z. H.; Graham, M. J. Appl. Phys. Lett. 1995, 66, 1123. (16) Lyman, P. F.; Sakata, O.; Marasco, D. L.; Lee, T.-L.; Breneman, K. D.; Keane, D. T.; Bedzyk, M. J. Surf. Sci. 2000, 462, 594. (17) Go¨thelid, M.; Lelay, G.; Wigren, C.; Bjo¨rkqvist, M.; Rad, M.; Karlsson, U. O. Appl. Surf. Sci. 1997, 115, 87. (18) Nelen, L. M.; Fuller, K.; Greenlief, C. M. Appl. Surf. Sci. 1999, 150, 65. (19) Cakmak, M.; Srivastava, G. P. Phys. ReV. B 1999, 60, 5497. (20) Houssa, M.; Nelis, D.; Hellin, D.; Pourtois, G.; Conard, T.; Paredis, K.; Vanormelingen, K.; Vantomme, A.; Bael, M. K. V.; Mullens, J.; Caymax, M.; Meuris, M.; Heyns, M. M. Appl. Phys. Lett. 2007, 90, 222105.

J. Phys. Chem. C, Vol. 114, No. 2, 2010 1027 (21) Kuhr, H. J.; Ranke, W. Surf. Sci. 1987, 189/190, 420. (22) Leung, K. T.; Terminello, L. J.; Hussain, Z.; Zhang, X. S.; Hayashi, T.; Shirley, D. A. Phys. ReV. B 1988, 38, 8241. (23) Cho, J. H.; Kleinman, L.; Kim, K. S. Phys. ReV. B 2002, 66, 112206. (24) Mui, C.; Senosiain, J. P.; Musgrave, C. B. Langmiur 2004, 20, 7604. (25) Culbertson, R. J.; Kuk, Y.; Feldman, L. C. Surf. Sci. 1986, 167, 127. (26) Kevan, S. D. Phys. ReV. B 1985, 32, 2344. (27) Xi, M.; Bent, B. E. J. Vac. Sci. Technol. B 1992, 97, 2440. (28) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; HeadGordon, M.; Replogle, E. S.; Pople, J. A. Gaussian03, Gaussian, Inc., Pittsburgh, PA, 2003. (29) Becke, A. D. J. Chem. Phys. 1992, 96, 2155. (30) Becke, A. D. J. Chem. Phys. 1992, 97, 9173. (31) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (32) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (33) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry; Plenum: New York, 1977.

(34) Wang, J. H.; Lin, M. C. J. Phys. Chem. B 2004, 108, 9189. (35) Wang, J. H.; Lin, M. C. Surf. Sci. 2005, 579, 197. (36) Landemark, E.; Karlsson, C. J.; Johansson, L. S. O.; Uhrberg, R. I. G. Phys. ReV. B 1994, 49, 16523. (37) Ranke, W.; Wasserfall, J. Surf. Sci. 1993, 292, 10. (38) Papagno, L.; Frankel, D.; Chen, Y.; Caputi, L. S.; Anderson, J.; Lapeyre, G. J. Surf. Sci. 1991, 248, 343. (39) Chabal, Y. J.; Christman, S. B. Phys. ReV. B 1984, 29, 6974. (40) Konecny, R.; Doren, D. J. J. Chem. Phys. 1997, 106, 2426. (41) Waltenburg, H. N.; Yates, J. T. Chem. ReV. 1995, 95, 1589. (42) Fo¨raker, A.; Doren, D. J. J. Chem. Phys. 2003, 107, 8507. (43) Lewis, L. B.; Segall, J.; Janda, K. C. J. Chem. Phys. 1995, 102, 7222. (44) Surnev, L.; Tikhov, M. Surf. Sci. 1984, 138, 40. (45) Zhou, X.-L.; Flores, C. R.; White, J. M. Appl. Surf. Sci. 1992, 62, 223. (46) Lide, D. R. CRC Handbook of Chemistry and Physics: a readyreference book of chemical and physical data; CRC Press: Boca Raton, FL, 2005-2006. (47) Han, M.; Luo, Y.; Camillone, N., III; Osgood, R. M., Jr. J. Phys. Chem. B 2000, 104, 6576. (48) Lai, Y. H.; Yeh, C. T.; Lin, Y. H.; Hung, W. H. Surf. Sci. 2002, 519, 150. (49) Lai, Y. H.; Yeh, C. T.; Yeh, C. C.; Hung, W. H. J. Phys. Chem. B 2003, 107, 9351. (50) Papageorgopoulos, A.; Kamaratos, M. Surf. Sci. 1996, 352-354, 364.

JP907791F