Supporting Information for
First-principles study of trimethylamine adsorption on anatase TiO2nanorod surfaces
Leonardo Triggiani,a,b,c Ana Belén Muñoz-García,a Angela Agostiano,b,c and Michele Pavonea,*
a. Dept. of Chemical Sciences, University of Naples Federico II, via Cintia 26, 80126 Naples, Italy
b. Dept. of Chemistry, University of Bari Aldo Moro, via E. Orabona 4, 70125 Bari, Italy
c. Institute of Physico-Chemical Processes of the National Research Council (CNR-IPCF), Bari Division, c/o Dept. of Chemistry, via E. Orabona 4, 70125 Bari, Italy
* E-mail:
- Comparison of PAW potentials
Different PAW potentials were tested in order to determine the combination that best describes the bulk properties. Ti and Tipv potentials differ for the number of electrons explicitly considered for the self-consistent cycle (3p3d4s and 3d4s, respectively); O and Osoft for the length of the cut-off radius (1.520 and 1.850 Å, respectively).
Table S1: Anatase structural parameters calculated with VASP using different PAW potentials to reproduce titanium and oxygen nuclei and core electrons.
(Å3) / (GPa) /Experimental / 136.27(a) / 179 ± 2(b) / 4.5 ± 1.0(b)
Ti + O / 141.542 / 192.5 / 4.69
Ti + Osoft / 141.248 / 190.9 / 4.76
Tipv + Osoft / 140.682 / 184.0 / 4.55
a (Å) / c (Å) / dax (Å) / deq(Å) / δ (°) / θ (°)
Experimental(a) / 3.7845 / 9.5143 / 1.9796 / 1.9339 / 101.9 / 156.2
Ti + O / 3.8344 / 9.6398 / 1.9997 / 1.9606 / 102.08 / 155.84
Ti + Osoft / 3.8301 / 9.6288 / 1.9985 / 1.9581 / 102.05 / 155.91
Tipv + Osoft / 3.8031 / 9.7510 / 2.0089 / 1.9493 / 102.71 / 154.58
V0 is the unit cell volume, B0 the bulk modulus and B0’ its derivative with respect to pressure at constant temperature; a and c are the cell dimensions; dax and deq the Ti-O distances in the direction of c (axial) and in the plane orthogonal to it (equatorial); greek letters indicate angles formed by two consecutive Ti-O bonds: δ between an axial and an equatorial bond and θ between two equatorial bonds.
(a) From ref. 1
(b) From ref. 2
- Validation of the plane-wave code for isolated molecules
In order to validate the use of the plane-wave code for the representation of the molecular structures of the two chosen adsorbates, we run analogous simulations on VASP and GAUSSIAN. The PBEXC density functional was used in both calculations, and in the latter case the Aug-cc-pVTZ Gaussian basis set was adopted. Optimized structures and corresponding geometrical parameters are reported in Figure S1 and Table S2. Evidently, isolated molecule structures and geometries are reliably reproduced with VASP’s plane-wave code, as well as with the nucleocentric basis set of the GAUSSIAN suite.
Table S2: Structural parameters of the two molecules chosen as adsorbates. Calculated values are compared to the experimental ones reported in literature for water and TMA (Ref. 3 and 4, respectively)
H2O / TMAd(O-H) / d(C-N) / d(C-Hs) / d(C-Ha) / θ(CNC)
VASP / 0.986 / 1.455 / 1.115 / 1.099 / 111.3
GAUSSIAN / 0.970 / 1.455 / 1.114 / 1.099 / 111.3
Experimental / 0.958 / 1.451±0.003 / 1.109±0.008 / 1.088±0.008 / 110.9±0.6
θ / θ(NCHs) / θ(NCHa) / θ(HaCHs) / θ(HaCHa)
VASP / 104.9 / 112.8 / 109.8 / 108.0 / 108.3
GAUSSIAN / 103.5 / 112.8 / 109.8 / 108.0 / 108.2
Experimental / 104.5 / 111.7±0.4 / 110.1±0.5 / 108.1±0.7 / 108.6±0.8
All distances (d) are expressed in angströms and all angles (θ) in degrees.
- Further details about water adsorption
- Contributions to molecular adsorption energies
Table S3 reports the three contributions to water molecular adsorption energies (similarly to Table 3 for TMA molecular adsorption in the main text). Also in this case, the driving force for the adsorption process is the electronic stabilization, since the negative EELEC terms largely prevail on the other two. In particular, the (positive) cohesive terms ECOH are always nearly negligible, while the structural relaxation terms (EREL), also positive, are very similar for the three surfaces investigated.
Table S3: Contributions to water adsorption energies, calculated with and without the dispersion corrections. All values are expressed in eV. Values in square brackets refer to calculations including D2 correction directly in the structural relaxation algorithm.
(001) / (100) / (101)PBE / EELEC / -0.508 / -0.635 / -0.597
EREL / +0.129 / +0.091 / +0.095
ECOH / +0.002 / +0.001 / +0.013
EADS / -0.377 / -0.544 / -0.489
PBE-D2 / EELEC / -0.638 [-0.641] / -0.781 [-0.783] / -0.760 [-0.766]
EREL / +0.127 [+0.124] / +0.085 [+0.085] / +0.100 [+0.095]
ECOH / +0.002 [+0.001] / +0.000 [+0.000] / +0.012 [+0.011]
EADS / -0.509 [-0.515] / -0.695 [-0.698] / -0.647 [-0.660]
PBE-D3 / EELEC / -0.649 / -0.789 / -0.766
EREL / +0.135 / +0.089 / +0.104
ECOH / +0.001 / +0.000 / +0.012
EADS / -0.512 / -0.699 / -0.649
3.2.Charge density difference plots
As for the case of TMA, also for water adsorption we found evidence of coordinative bond formation between the molecule and the Ti5c site. The charge density difference plots for the three investigated surfaces are reported in Figure S2.
3.3.PDOS
In Figure S3 are shown the densities of states of the three surfaces before and after water adsorption (both molecular and dissociative). It comes out that the nature of valence and conduction bands is almost unmodified by the adsorption process.
- Comparison of TMA adsorption energies with different degrees of coverage
Very small energy differences are found comparing the two degrees of coverage (θ = 1/4 and 1/12 ML): a systematic increase of less than 0.1 eV is observed on EADS, strictly related with a corresponding enhancement of the electronic stabilization.
Table S4. Comparison of the contributions to the TMA adsorption energies (in eV) with different degrees of coverage. Values in square brackets refer to calculations including D2 correction directly in the structural relaxation algorithm.
θ = 1/4 ML / θ = 1/12 ML(001) / (100) / (101) / (001) / (100) / (101)
PBE / EELEC / -0.937 / -0.802 / -0.844 / ‒ / -0.928 / -0.921
EREL / +0.300 / +0.191 / +0.256 / ‒ / +0.222 / +0.277
ECOH / +0.046 / +0.032 / +0.043 / ‒ / +0.043 / +0.046
EADS / -0.591 / -0.579 / -0.545 / ‒ / -0.664 / -0.598
PBE-D2 / EELEC / -1.446
[-1.477] / -1.581
[-1.346] / -1.369
[-1.440] / ‒ / -1.434
[-1.461] / -1.471
[-1.520]
EREL / +0.283
[+0.310] / +0.461
[+0.190] / +0.266
[+0.280] / ‒ / +0.215
[+0.206] / +0.281
[+0.295]
ECOH / +0.025
[+0.031] / +0.017
[+0.023] / +0.025
[+0.027] / ‒ / +0.035
[+0.038] / +0.036
[+0.043]
EADS / -1.138
[-1.136] / -1.103
[-1.133] / -1.078
[-1.132] / ‒ / -1.184
[-1.217] / -1.154
[-1.181]
PBE-D3 / EELEC / -1.494 / -1.615 / -1.419 / ‒ / -1.470 / -1.515
EREL / +0.278 / +0.471 / +0.275 / ‒ / +0.226 / +0.291
ECOH / +0.030 / +0.023 / +0.034 / ‒ / +0.044 / +0.047
EADS / -1.186 / -1.120 / -1.110 / ‒ / -1.200 / -1.177
- Comparison TMA vs. NH3 molecular adsorption energies
TMA adsorption is systematically less favorable than ammonia on all three surfaces of anatase: EADS for TMAare almost a half of the corresponding values for NH3, except for the (100) surface, for which the method used in literature is quite different from ours, so the results should not be directly compared. Moreover, the lack of selectivity toward any specific surface is a feature of the adsorption processes of both molecules.
Table S5. Comparison of calculated TMA adsorption energies with literature data of ammonia.
Surface / Molecule / Method / EADS (eV) / Reference(0 0 1) / TMA / PW:DFT-GGA(PBE) / -0.590 / This work
NH3 / Cluster, DFT/B3LYP/6-31G** / -0.997 / 5
NH3 / PW:DFT-GGA(PW91) / -1.084 / 6
NH3 / Cluster, B3LYP/6-311G**//B3LYP/6-31G* / -1.136 / 7
NH3 / Cluster, DFT/B3LYP/6-31G** / -1.171 / 8
(1 0 0) / TMA / PW:DFT-GGA(PBE) / -0.581 / This work
NH3 / PP-MSINDO / -1.648 / 9
(1 0 1) / TMA / PW:DFT-GGA(PBE) / -0.546 / This work
NH3 / PBC-DFT(B3LYP)/6-31G* / -1.045 / 10
NH3 / Cluster, B3LYP/6-311G**//B3LYP/6-31G* / -1.432 / 7
NH3 / Cluster, DFT/B3LYP/6-31G** / -1.113 / 8
References to Supporting Information
1. Howard, C.J.; Sabine, T.M.; Dickson, F., Structural and Thermal Parameters for Rutile and Anatase. Acta Crystallogr., Sect. B: Struct. Sci. 1991, 47 (4), 462-468.
2. Arlt, T.; Bermejo, M.; Blanco, M.A.; Gerward, L.; Jiang, J.Z.; Staun Olsen, J.; Recio, J.M., High-Pressure Polymorphs of Anatase TiO2. Phys. Rev. B 2000, 61 (21), 14414-14419.
3. Csaszar, A.G.; Czako, G.; Furtenbacher, T.; Tennyson, J.; Szalay, V.; Shirin, S.V.; Zobov, N.F.; Polyansky, O.L., On Equilibrium Structures of the Water Molecule. J. Chem. Phys. 2005, 122 (21), 214305-214310.
4. Wollrab, J.E.; Laurie, V.W., Structure and Conformation of Trimethylamine. J. Chem. Phys. 1969, 51 (4), 1580-1583.
5. Erdogan, R.; Onal, I., AnONIOM and DFT Study of Water and Ammonia Adsorption on Anatase TiO2 (001) Cluster. Int. J. Quantum Chem. 2011, 111 (9), 2149-2159.
6. Erdogan, R.; Ozbek, O.; Onal, I., A Periodic DFT Study of Water and Ammonia Adsorption on Anatase TiO2 (001) Slab. Surf. Sci. 2010, 604 (11–12), 1029-1033.
7. Wanbayor, R.; Ruangpornvisuti, V., Adsorption of Di-, Tri- and Polyatomic Gases on the Anatase TiO2 (001) and (101) Surfaces and Their Adsorption Abilities. J. Mol. Struc. THEOCHEM2010, 952 (1–3), 103-108.
8. Onal, I.; Soyer, S.; Senkan, S., Adsorption of Water and Ammonia on TiO2-Anatase Cluster Models. Surf. Sci. 2006, 600 (12), 2457-2469.
9. Homann, T.; Bredow, T.; Jug, K., Adsorption of Small Molecules on the Anatase (100) Surface. Surf. Sci. 2004, 555 (1–3), 135-144.
10. Wanbayor, R.; Ruangpornvisuti, V., Adsorption of CO, H2, N2O, NH3 and CH4 on the Anatase TiO2 (001) and (101) Surfaces and Their Competitive Adsorption Predicted by Periodic DFT Calculations. Mater. Chem. Phys. 2010, 124 (1), 720-725.