Professional Documents
Culture Documents
James L. Gole
Schools of Physics and Mechanical Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0430
ReceiVed: March 5, 2010; ReVised Manuscript ReceiVed: May 26, 2010
The hydrolysis of titanium tetrachloride (TiCl4) to produce titanium dioxide (TiO2) nanoparticles has been
studied to provide insight into the mechanism for forming these nanoparticles. We provide calculations of
the potential energy surfaces, the thermochemistry of the intermediates, and the reaction paths for the initial
steps in the hydrolysis of TiCl4. We assess the role of the titanium oxychlorides (TixOyClz; x ) 2-4, y ) 1,
3-6, and z ) 2, 4, 6) and their viable reaction paths. Using transition-state theory and RRKM theory, we
predicted rate constants including the effect of tunneling. Heats of formation at 0 and 298 K are predicted for
TiCl4, TiCl3OH, TiOCl2, TiOClOH, TiCl2(OH)2, TiCl(OH)3, Ti(OH)4, and TiO2 using the CCSD(T) method
with correlation consistent basis sets extrapolated to the complete basis set limit and compared with the
available experimental data. Clustering energies and heats of formation are calculated for neutral clusters.
The calculated heats of formation were used to study condensation reactions that eliminate HCl or H2O. The
reaction energy is substantially endothermic if more than two HCl molecules are eliminated. The results
show that the mechanisms leading to formation of TiO2 nanoparticles and larger ones are complicated and
will have a strong dependence on the experimental conditions.
TABLE 1: Reaction Enthalpy (H, kcal/mol) at 298 K for step in the combustion system and is relevant to the hydrolysis
Reactions 1a and 1ba mechanism as well. The formation of the species described
H298K H1500K G298K G1500K herein can be coupled with those based on the thermal
decomposition of TiCl4 in further studies of the formation of
reaction 1b, n ) 1 114.2 110.8 103.7 64.9
reaction 1b, n ) 2 105.4 103.0 96.6 64.3
TiO2 nanoparticles. For a number of species, the thermodynam-
reaction 1b, n ) 3 86.6 85.5 77.8 44.4 ics have been obtained at the coupled cluster CCSD(T) level at
reaction 1b, n ) 4 66.7 66.7 65.1 59.8 the complete basis set (CBS) limit using approaches developed
reaction 1a -43.4 -41.5 -39.0 -23.0 in our group in collaboration with Washington State Univer-
a sity.32 We have previously used such an approach to study the
Heats of formation for (TiO2)n from ref 19 and for TiCl4
(-182.4 kcal/mol) and TiO2(s,rutile) from ref 21. thermodynamic properties of TiO2 and other transition-metal
oxide clusters.19 The focus of the current study is on the species
TABLE 2: Reaction Enthalpy (H, kcal/mol) at 298 K for formed by the reaction of H2O with TiCl4 and their subsequent
Reactions 2a and 2ba clustering reactions.
H298K H700K H1500K G298K G700K G1500K
Computational Methods
reaction 2c, n ) 1 141.4 140.8 139.4 121.8 95.7 44.8
reaction 2c, n ) 2 159.8 159.9 160.3 132.9 96.6 24.1 The potential energy surfaces were initially calculated at the
reaction 2c, n ) 3 171.7 172.8 174.8 135.8 86.8 -12.4 density functional theory level with the B3LYP exchange-
reaction 2c, n ) 4 175.5 177.4 181.2 137.7 85.8 -20.6 correlation functional33,34 and the aug-cc-pVDz/aug-cc-pVDZ-
reaction 2b -16.2 -15.2 -12.8 -20.9 -27.8 -43.1
PP basis set described below. Equilibrium geometries and
reaction 2ab -66.6 -51.0
harmonic vibrational frequencies were calculated at the second-
a
Heats of formation for (TiO2)n from ref 19 and for TiCl4 order Mller-Plesset perturbation theory (MP2) level for
(-182.4 kcal/mol) and TiO2(s,rutile) from ref 21 unless noted various structures on the potential energy surface using the
otherwise. b Heats of formation from ref 22. If aqueous TiCl4 is Gaussian 03 program.35 These calculations were done with the
used, then H ) 1.3 kcal/mol and the reaction is essentially
aug-cc-pVnZ basis sets36 for H and O, the aug-cc-pV(n+d)Z
thermoneutral.
basis sets for Cl,37 and the small core effective core potential
(ECP) based aug-cc-pVnZ-PP basis sets19,38 for Ti with n ) D
reagent.18 An important issue is that both reactions 1b and 2c
and T. (We denote this combination of basis sets as aVnZ.)
are endothermic to produce small (TiO2)n nanoparticles in the
For the hydrates and transition states, the geometries calculated
gas phase. Therefore, one must partially oxidize or hydrolyze
at the MP2/aVTZ level were subsequently used in single-point
the TiCl4 to build up an oxychloride/hydroxyoxychloride particle
CCSD(T) calculations with the aVDZ, aVTZ, and aVQZ basis
large enough so that the reaction thermodynamics begins to
sets. The geometries of TiCl4, TiCl3OH, TiOCl2, TiOClOH, and
resemble that of the solid as evidenced in Table 1 for TiO2
TiO2 were also optimized at the CCSD(T) level with the aVDZ
clusters starting from Cl2 (reactions 1a and 1b) and in Table 2
and aVTZ basis sets. The geometry calculated at the CCSD(T)/
starting from H2O.19-22 The enthalpy and free energy in Table
aVTZ level was then used in single-point CCSD(T)/aVQZ
1 at 298 K show that, as the particle gets larger, the energy to
calculation. All of the CCSD(T) calculations were performed
form it decreases. At 1500 K, the structure of the individual
with the MOLPRO 2006.1 program.39 The CCSD(T) energies
particle begins to play a more important role as the presence of
were then extrapolated to the CBS limit using a mixed Gaussian/
low-lying vibrational modes in the cluster can significantly
exponential formula (eq 3).40 n ) 2, 3, and 4 for aVDZ, aVTZ,
contribute to the entropy contribution to the free energy. In Table
and aVQZ, respectively.
2, the addition of H2O leads to more exothermic (in terms of
the free energy) reactions as the temperature increases, especially
E(n) ) ECBS + Be-(n-1) + Ce-(n-1)
2
as the nanoparticle size gets larger, due to the release of more (3)
HCl particles in the gas phase. The hydrolysis reaction to form
rutile is overall exothermic, independent of whether the process
takes place with gas-phase reactants or in the aqueous phase. To predict thermochemical properties to high accuracy, it is
The gas-phase processes for the generation of TiO2 have been necessary to include additional corrections. For the complexes
modeled in terms of the oxidation of TiClx radicals which then with one Ti atom, the core-valence correlation corrections
cluster to form (TiO2)n nanoparticles for large n.23 In another (ECV) were obtained at the CCSD(T)-DK/aug-cc-pwCVTZ-
case, TiOCl2 has been suggested to be a key intermediate.24,25 DK level.38 For the calculation of transition-metal compound
West et al.26 have predicted the thermochemical parameters for atomization energies, the core-valence calculations should be
a number of TiOxCly intermediates using density functional calculated at the CCSD(T)-DK level to achieve the best
theory (DFT)27 with different functionals and with coupled accuracy.19 In addition, we also account for relativistic effects
cluster28 CCSD(T)29 calculations using modest basis sets. These in atoms and molecules. The first is the spin-orbit correction
data were then used in kinetic models to predict the formation (ESO), which lowers the sum of the atomic energies (decreasing
of TiO2 nanoparticles.14,30,31 The major reactions they studied D0) by replacing energies that correspond to an average over
were based on the thermal decomposition of TiCl4 to from the the available spin multiplets with energies for the lowest
TiCl3 and Cl radicals. The compounds that we studied are similar multiplets. This correction is required as most electronic
to those given by Kraft and co-workers14,30,31 with the addition structure codes produce only spin multiplet averaged wave
of the presence of hydroxyl groups, and their compounds could functions. The experimental ground-state atomic spin-orbit
also play a role in the hydrolysis process. corrections are ESO(Ti) ) 0.64 kcal/mol, ESO(Cl) ) 0.84
Although the combustion approach to the synthesis of TiO2 kcal/mol, and ESO(O) ) 0.22 kcal/mol.41 The second correction
particles has been used in industry for many years, the overall is the scalar relativistic correction (ESR). As we already have
mechanism is still poorly understood. The goal of this work is such a correction for Ti through the use of the ECP, we
to study the potential energy surface for the initial steps in the evaluated ESR for O and Cl as expectation values of the two
gas-phase hydrolysis of TiCl4, which may serve as an initiation dominant terms in the Breit-Pauli Hamiltonian, the mass-
Hydrolysis of TiCl4 J. Phys. Chem. A, Vol. 114, No. 28, 2010 7563
velocity and one-electron Darwin (MVD) corrections, at the TABLE 3: Calculated and Experimental Ti-X (X ) Cl, O,
configuration interaction singles and doubles (CISD) level with OH) Bond Lengths () at the CCSD(T)/aVTZ or MP2/aVTZ
the aVTZ basis set at the MP2/aVTZ geometry. Our computed Levels
total atomization energy (D0) values were obtained with eq 4. molecule state method Ti-Cl Ti-O Ti-OH
1
TiCl4 A1/Td CCSD(T)/aVTZ 2.186
D0 ) ECBS + EZPE + ECV + ESR + ESO exptla 2.170
1
TiOCl2 A1/C2V CCSD(T)/aVTZ 2.250 1.619
(4) B97-1/6-311+G(d,p)b 2.234 1.587
1
TiOClOH A/Cs CCSD(T)/aVTZ 2.270 1.627 1.843
1
TiO2 A1/C2V CCSD(T)/aVTZc 1.666
Given the known heats of formation at 0 K for the elements, exptld 1.651
Hf,0K(Ti) ) 112.4 ( 0.7 kcal/mol,42 Hf,0K(Cl) ) 28.59 kcal/ TiCl3OH 1
A/Cs MP2/aVTZ 2.192 1.767
mol,21 Hf,0K(O) ) 58.98 kcal/mol,21 and Hf,0K(H) ) 51.63 2.191
1
kcal/mol,21 we can derive Hf,0K values for the molecules under TiCl2(OH)2 A/C1 MP2/aVTZ 2.213 1.784
study. The heats of formation at 298 K can be obtained by 1.775
1
TiCl(OH)3 A/C3 MP2/aVTZ 2.238 1.797
following the procedures outlined by Curtiss et al.43 The Ti(OH)4 1
A/S4 MP2/aVTZ 1.813
temperature dependence of the enthalpy and the entropy
a
calculations were done in the rigid-rotor, harmonic oscillator Reference 48. b Reference 26. c Reference 19. d Reference 50.
approximation with hindered rotors treated as a vibration.44 At
higher temperatures, this may lead to modest errors in the the gas-phase electron diffraction48 value of 2.170 . Lower
thermodynamic properties due to the transition of a hindered level calculations are also in agreement.49 The Ti-Cl bond
rotor to a free rotor. distances in TiCl3OH, TiOCl2, TiOClOH, TiCl2(OH)2, and
The CCSD(T) method scales approximately as N7 with N TiCl(OH)3 are predicted to lengthen from that in TiCl4 with an
basis functions, and large basis sets are required to reach the increase in the number of O atoms and with changes in the
CBS limit. The approach described above for the CCSD(T)/ number of ligands bonded to the Ti atom. The Ti-Cl bond
CBS calculations has been used to predict a wide range of distance in TiOCl2 of 2.234 calculated26 at the B97-1/6-
thermodynamic properties to chemical accuracy.19,20,32,45 DFT 311+G(d,p) level is shorter than our CCSD(T)/aVTZ value by
has a much better scaling, scaling as N3 to N4 depending on the 0.02 . The calculated Ti-O bond distance in the series from
exchange-correlation functional, but is not as accurate as the TiO2 to TiOClOH to TiOCl2 decreases as the number of ligands
CCSD(T)/CBS approach for the calculation of a broad range increases from two to three and Cl is substituted for OH. The
of thermodynamic properties including atomization energies. Ti-O bond distance in TiOCl2 calculated at the B97-1/6-
For small clusters with one Ti atom, accurate thermodynamic 311+G(d,p) level26 is shorter than our CCSD(T)/aVTZ value
properties can be calculated using the CCSD(T)/CBS approach, by 0.03 . The Ti-O bond distance in TiO2 has been
which is more reliable than similar DFT-based approaches.19,20,46 measured by Fourier transform microwave spectroscopy to be
In addition, the CCSD(T)/CBS values can be used to benchmark 1.651 ;50 the CCSD(T)/aVTZ calculations overestimate this
different DFT functionals in the future. We avoided the use of distance by 0.015 .19 The Ti-OH bond distance increases by
isodesmic and similar types of reaction energies as used 0.067 from TiCl3OH to TiOClOH.
previously because of issues with the experimental heats of The role of multireference character in the wave function can
formation of similar Ti compounds and because there are so be estimated from the T1 diagnostic51 for the CCSD calculation.
few experimental values.47 For the larger clusters, we used the The values for the T1 diagnostics are modest (<0.03), showing
normalized clustering energy method that we have previously that the wave function is dominated by a single electron
developed19,20 for the prediction of the heats of formation of configuration. TiO2 has a slightly larger T1 diagnostic but is
transition-metal oxide clusters. well-behaved on the basis of our recent work.19
All calculations were performed on a Xeon- and Opteron- Atomization Energies and Heats of Formation. The dif-
based Penguin Computing Linux cluster in our group, the ferent energetic components for the total atomization energies
Itanium 2-based SGI Altix and the Opteron-based dense memory (TAEs) are given in Table 4. The core-valence corrections for
cluster Linux cluster at the Alabama Supercomputer Center, the the Ti compounds are found to increase the TAEs by 1.3-4.8
Xeon-based Dell Linux cluster at The University of Alabama, kcal/mol. The scalar relativistic corrections for the Ti compounds
and the Opteron-based Linux cluster at the Pacific Northwest are usually smaller than the core-valence corrections and are
National Laboratory. found to decrease the TAEs by 0.7-2.0 kcal/mol.
The calculated heats of formation are also given in Table 4,
Results and Discussion
and we use the values at 298 K in our discussion below unless
Geometries and Frequencies. The calculated Ti-X (X ) specified otherwise. The estimated error bars for the calculated
Cl, O, OH) bond lengths at the CCSD(T)/aVTZ level for heats of formation are (1.5 kcal/mol. Here we include our
selected reactants, intermediates, and products are given in Table estimates of the errors in the energy extrapolation, vibrational
3. They are compared with available experimental values. The frequencies, and other electronic energy components, as well
electronic states and symmetry labels for these molecules are as the error associated with the heat of formation of the Ti atom
given in Table 3 and will not be repeated hereafter. The based on our recent work.20 The calculated heat of formation
optimized molecular structures for these molecules together with of TiCl4 is in good agreement with the experimental value20
those of the transition states are given in the potential energy considering the experimental error limits of (0.7 kcal/mol for
surface plots (Figure 1). The total CCSD(T) energies, calculated both Hf(TiCl4) and Hf(Ti). The calculated heat of formation
harmonic frequencies compared with the available experimental of TiOCl2 is 11.4 kcal/mol more negative than the estimated
values, and T1 diagnostics are given as Supporting Information. value of -130.4 kcal/mol.21 For TiOCl2, the calculated heat of
The calculated Ti-Cl bond distance of 2.186 at the formation by Green and co-workers26 at the B97-1/6-311+G(d,p)
CCSD(T)/aVTZ level in TiCl4 is slightly longer (<0.02 ) than level of -142.9 ( 4.8 kcal/mol is in good agreement with our
7564 J. Phys. Chem. A, Vol. 114, No. 28, 2010 Wang et al.
TABLE 4: Total Atomization Energies (D0,0K, kcal/mol) at 0 K and Heats of Formation at 0 and 298 K (Hf,0K and Hf,298K,
kcal/mol) Calculated at the CCSD(T) Level
molecule ECBSa EZPEb ECVc ESRd ESOe D0,0Kf Hf,0Kg Hf,298Kg Hf,298K (exptl)
TiCl4 413.56 -3.65 3.32 -1.46 -4.00 407.78 -181.0 -181.5 -182.4 ( 0.7h
TiCl3OH 533.91 -10.27 3.88 -1.67 -3.38 522.47 -213.7 -214.5
TiOCl2 375.57 -3.23 1.35 -1.17 -2.54 369.98 -141.4 -141.8 -130.4h
TiOClOH 490.09 -10.41 1.87 -1.32 -1.92 478.31 -167.7 -168.8
TiO2i 299.77 -3.26 2.92 -0.78 -1.08 297.57 -67.2 -67.8 -73.0 ( 3i
TiCl2(OH)2 653.35 -17.32 4.04 -1.86 -2.76 635.45 -244.7 -246.1
TiCl(OH)3 770.77 -24.69 4.74 -2.00 -2.14 746.69 -273.9 -276.2
Ti(OH)4 885.14 -31.81 4.34 -2.14 -1.52 855.01 -300.2 -303.2
H2 O 232.95 -13.44 0.34 -0.29 -0.22 219.34 -57.1 -57.8 -57.7978 ( 0.0096h
TiCl4 H2O 655.90 -19.22 3.89 -1.87 -4.22 634.48 -245.5 -247.2
TS1 632.94 -15.24 4.05 -1.88 -4.22 611.60 -226.6 -228.2
TiCl3OH HCl 644.14 -15.93 4.20 -1.90 -4.22 626.30 -237.3 -238.3
HCl 107.40 -4.36 0.19 -0.23 -0.84 102.17 -21.9 -22.0 -22.06 ( 0.024i
TS2 495.63 -7.80 3.80 -1.50 -3.38 486.75 -178.0 -179.5
TiOCl2 HCl 499.92 -8.95 2.98 -1.48 -3.38 489.09 -180.3 -181.0
TiOCl2 H2O 640.95 -19.01 2.99 -1.64 -2.76 620.54 -229.7 -231.4
a
CCSD(T)/CBS energies obtained using eq 3. b MP2/aVTZ. c CCSD(T)/aug-cc-pwCVTZ-DK. d Expectation value of the MVD operators on
the CISD/aVTZ wave function. e Experimental atomic spin orbit corrections from ref 41. f Equation 4. g See the text. h Reference 17. i Reference
15.
TABLE 5: Calculated Reaction Energies (kcal/mol) and Rate Constants k (cm3/(molecule s)) from TST at 298, 700, and 1500 K
for Bimolecular Reactions
reaction H298K H700K H1500K G298K G700K G1500K k(298 K)a k(700 K) k(1500 K)
-3
TiCl4 + H2O f TS1 f TiCl3OH + HCl (5a) b
2.7 3.5 4.9 0.7 -2.2 -9.7 2.0 10 3.0 102 5.0 104
TiCl3OH + H2O f TS3 f TiCl2(OH)2 + HClc (7) 4.1 4.8 6.3 4.4 4.3 3.3 5.1 10-4 3.4 101 4.3 103
TiCl2(OH)2 + H2O f TS4 f TiCl(OH)3 + HCld (8) 5.6 6.3 7.7 6.2 6.7 6.3 4.8 10-3 8.4 101 6.1 103
TiCl(OH)3 + H2O f TS5 f Ti(OH)4 + HCle (9) 8.7 9.4 10.8 9.4 10.0 10.1 3.5 10-3 1.3 102 1.3 104
TiOCl2 + H2O f TiOClOH + HCl (6a) 8.7 9.3 10.8 8.8 7.6 7.5 f f f
a
The tunneling factors are not included in the rate constants. b Qtunnel(298 K) ) 2.38. c Qtunnel(298 K) ) 1.94. d Qtunnel(298 K) ) 1.65.
e
Qtunnel(298 K) ) 1.53. f There is no reaction barrier for the bimolecular reaction other than the reaction endothermicity.
TABLE 6: Calculated Rate Constants from TST (k, s-1) and the RRKM Rate Constant (s-1) at 298, 700, and 1500 K for the
Unimolecular Reactions from the Precomplex to the Postcomplexa
reaction k(298 K) k(700 K) k(1500 K)
-2
TiCl4 H2O f TS1 f TiCl3OH HClb
5.1 10 4.7 10 6
8.9 108
TiCl3OH H2O f TS3 f TiCl2(OH)2 HClc 3.0 10-2 2.0 106 6.2 108
TiCl2(OH)2 H2O f TS4 f TiCl(OH)3 HCld 1.9 10-2 9.9 105 4.3 108
TiCl(OH)3 H2O f TS5 f Ti(OH)4 HCle 3.5 10-1 2.1 107 1.8 109
TiOCl2 H2O f TS6 f TiOClOH HCl 3.1 10-6 4.0 104 1.4 108
the TST rate constants assuming a bimolecular reaction (reac- temperatures. Figure 3 shows the calculated molecular structures
tions 5a, 7, 8, and 9) at 298, 700, and 1500 K in Table 5 because with key geometry parameters for Ti2O3Cl2, Ti2O2Cl4, Ti2OCl6,
the precomplexes may not have long lifetimes at the higher Ti3O5Cl2, Ti3O4Cl4, Ti3O3Cl6, Ti4O6Cl4 and for a number of
temperatures. Both the unimolecular and bimolecular reaction TixOyClz(OH)w clusters. These molecules were optimized at the
rates are substantial at T g 700 K. DFT level with the B3LYP, BP86,56,57 PBE,58,59 and PW9160,61
TiOCl2 has been observed experimentally in some of the exchange-correlation functionals and the aVDZ basis set. The
systems,24,25 so we also studied the reaction of TiOCl2 with water geometries obtained with the aVDZ basis set were then used in
(reaction 6). The first step as shown in the following reaction single-point DFT energy calculations with the aVTZ basis set.
begins with the formation of an initial Lewis acid-base complex The geometries obtained at the B3LYP/aVDZ level were also
with a much larger complexation energy of 37.8 kcal/mol used in single-point energy calculations at the CCSD(T)/aVDZ
(Figure 2): level.
In our previous work on TiO2 clusters,19 we studied the
clustering energies of (TiO2)n (n ) 2-4) clusters, and we found
TiOCl2 + H2O f TiClOH + HCl (6a)
that the normalized clustering energies at 0 K range from 60 to
100 kcal/mol, increasing from the dimer to the tetramer. The
The hydrogen transfer reaction from the complex has a higher normalized clustering energy is the average binding energy of
barrier of 24.2 kcal/mol, and the resulting complex is a Lewis the monomers in a cluster and for a (TiO2)n cluster with n Ti
acid-base adduct of HCl with TiO(OH)Cl with a Lewis atoms, Enorm,n is defined by the following equation:
acid-base adduct bond energy of 16.1 kcal/mol. The overall
reaction to form HCl + TiO(OH)Cl is endothermic by 9.2 kcal/ Enorm,n ) [nE(monomer) - E(cluster)]/n (10)
mol at 0 K. The value of Keq at 700 K for reaction 6a is 4
10-3, which is large enough for some product to be formed
which can be involved in further clustering reactions. There is For the TiO2 clusters previously studied, the CBS extrapolation
no reaction barrier on the bimolecular energy surface other than effect from the aVDZ basis set is small in contrast to the basis
the endothermicity of the reaction as the transition state for set extrapolation in the total atomization energies. We can then
hydrogen transfer is below the energy of the reactants. At higher use the following equation to obtain the heat of formation of a
temperatures, the reaction will occur at the collision frequency cluster where we are summing over the heats of formation of
reduced by the appropriate exponential expression involving the the monomers:
endothermicity (a factor of 10-20 at 1500 K). The loss of the
second HCl is highly endothermic as expected. Hf,0K(cluster) ) Hf,0K(monomer) - nEnorm,n
Titanium Oxychloride Hydroxide Species. We have noted
(11)
above that the formation of TiO2 nanoclusters must proceed by
way of TixOyClz (or TixOyClz(OH)w) clusters as discussed in the
Introduction and have shown that it is possible to generate the The advantage of this approach is that the normalized clustering
intermediates needed for clustering reactions at the experimental energies (Enorm,n) are not strongly dependent on the basis set.19
Hydrolysis of TiCl4 J. Phys. Chem. A, Vol. 114, No. 28, 2010 7567
TABLE 7: Calculated Heats of Formation (kcal/mol) of the Clusters and Reaction Energies (kcal/mol) at the CCSD(T)/aVDZ//
B3LYP/aVDZ Level
reaction H0K H298K H700K H1500K Hf,0K Hf,298K
2TiOCl2 f Ti2O2Cl4 -86.3 -87.0 -85.7 -82.5 -361.9 -370.6
3TiOCl2 f Ti3O3Cl6 -155.9 -156.4 -153.5 -147.1 -580.1 -581.8
TiOCl2 + TiO2 f Ti2O3Cl2 -105.7 -106.4 -105.1 -102.0 -314.3 -316.0
TiOCl2 + 2TiO2 f Ti3O5Cl2 -203.7 -204.2 -201.4 -195.1 -479.5 -481.6
2TiOCl2 + TiO2 f Ti3O4Cl4 -200.9 -201.9 -199.2 -192.9 -550.9 -553.3
TiOCl2 + TiCl4 f Ti2OCl6 -54.9 -55.2 -54.5 -52.9 -377.3 -378.5
2TiOCl2 + 2TiO2 f Ti4O6Cl4 -296.9 -298.4 -294.1 -284.5 -714.1 -717.6
TiCl4 + TiOClOH f Ti2OCl5OH -58.6 -59.1 -57.0 -53.8 -407.8 -409.4
2TiOClOH f Ti2O2Cl2(OH)2 (cis) -92.2 -92.5 -90.9 -87.7 -427.6 -430.1
2TiOClOH f Ti2O2Cl2(OH)2 (trans) -93.2 -93.5 -92.1 -88.9 -428.6 -431.1
2TiOClOH f Ti2O2Cl2(OH)2 (iso) -91.1 -92.4 -90.9 -87.8 -426.5 -430.0
Ti(OH)4 + TiO(OH)2 f Ti2O(OH)6 -55.1 -54.4 -52.6 -49.3 -552.7 -557.8
2TiO(OH)2 f Ti2O2(OH)4 -88.8 -88.7 -86.9 -83.5 -483.6 -487.8
TiOClOH + TiO(OH)2 f Ti2O2Cl(OH)3 -91.5 -91.7 -93.1 -120.2 -456.6 -460.0
dative bond energies.45,62 These interactions are not dominated ature rate constant is independent of pressure, and the effect
by hydrogen bonding. The DFT functionals used may of tunneling on the rate constant is an increase of about a
underestimate the reaction barriers as found in some of the factor of 2 at 298 K. At higher temperatures, the predicted
hydrogen transfer reactions. The relative energies of TiCl4 rate constants are large enough for the reactions to occur
H2O, TiCl3OH H2O, TiCl2(OH)2 H2O, TiCl3OH H2O, and readily and show that key oxygenated intermediates can be
TiOCl2 H2O calculated at the B3LYP/aVDZ level are greater formed. The calculations show, at the temperatures relevant
than those at the CCSD(T) level for ECBS + ECV by 2.8, to the experimental conditions for the gas-phase hydrolysis
5.8, 4.8, 4.9, and 3.1 kcal/mol, respectively. Thus, DFT/ of TiCl4 or the combustion of TiCl4 leading to TiO2 particles,
B3LYP provides a qualitatively correct potential energy that partially oxygenated Ti molecules can be generated.
surface but not a quantitative one. These molecules can condense to form partially oxygenated
Ti clusters which are precursors to TiO2 nanoparticles. Our
Conclusions calculated data will enable the development of more detailed
Coupled cluster [CCSD(T)] theory, Mller-Plesset per- gas-phase mechanisms for the formation of TiO2 nanopar-
turbation theory (MP2), and density functional theory (DFT) ticles and larger particles which incorporate the hydrolysis
have been used to study the structural and energetic properties of TiCl4. Within this framework, our calculations also show
of the intermediates in the hydrolysis of TiCl4 and the that the reported values for the heat of formation of TiOCl2
formation of titanium oxychloride species (TixOyClz; x ) 2-4, should be remeasured.
y ) 1, 3-6, and z ) 2, 4, 6). The normalized clustering
energies at 0 K for the ground state of (TiOCl2)n (n ) 2, 3) Acknowledgment. This work was supported by the Na-
clusters range from 40 to 55 kcal/mol, increasing from the tional Science Foundation (Grant CTS-0608896), through the
dimer to the trimer. The reaction energy of the condensation NIRT program, and by the Chemical Sciences, Geosciences
reactions is substantially endothermic if more than two HCl and Biosciences Division, Office of Basic Energy Sciences,
molecules are eliminated. In addition, we also predict the U.S. Department of Energy (DOE), under Grant DE-FG02-
reaction paths for the initial steps of hydrolysis of TiCl4. We 03ER15481 (Catalysis Center Program). D.A.D. also thanks
predict that the hydrolysis of TiCl4 will not form TiOCl2 and the Robert Ramsay Chair Fund of The University of Alabama
HCl or TiO2 and HCl due to the substantial endothermicities for support. Part of this work was performed in the Molecular
associated with the formation of gas-phase TiO2. Using Sciences Computing Facility at the W. R. Wiley Environ-
transition-state theory and RRKM theory, we predicted rate mental Molecular Sciences Laboratory, a national scientific
constants including the effect of tunneling. The low-temper- user facility sponsored by the DOEs Office of Biological
Hydrolysis of TiCl4 J. Phys. Chem. A, Vol. 114, No. 28, 2010 7569
and Environmental Research and located at Pacific Northwest (27) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and
National Laboratory, operated for the DOE by Battelle. We Molecules; Oxford University Press: New York, 1989.
(28) Bartlett, R. J.; Musial, M. ReV. Mod. Phys. 2007, 79, 291.
thank Prof. K. A. Peterson of Washington State University (29) (a) Purvis, G. D., III; Bartlett, R. J. J. Chem. Phys. 1982, 76, 1910.
for providing the Ti basis set. (b) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. Chem.
Phys. Lett. 1989, 157, 479. (c) Watts, J. D.; Gauss, J.; Bartlett, R. J. J. Chem.
Supporting Information Available: Calculated geometry Phys. 1993, 98, 8718.
(30) West, R. H.; Celnik, M. S.; Inderwildi, R.; Kraft, M.; Beran,
parameters for the molecules, CCSD(T)/aVnZ total energies G. J. O.; Green, W. H. Ind. Eng. Chem. 2007, 46, 6147.
(Eh) as a function of the basis set, calculated harmonic (31) West, R. H.; Shirley, R. A.; Kraft, M.; Goldsmith, C. F.; Green,
frequencies for molecules at the MP2/aVTZ level, calculated W. H. Combust. Flame 2009, 156, 1764.
harmonic frequencies for Ti2O3Cl2, Ti2O2Cl4, Ti2OCl6, (32) (a) Feller, D.; Dixon, D. A. J. Phys. Chem. A 2000, 104, 3048. (b)
Feller, D.; Dixon, D. A. J. Chem. Phys. 2001, 115, 3484. (c) Dixon, D. A.;
Ti3O4Cl4, Ti3O5Cl2, Ti3O3Cl6, and Ti4O6Cl4 at the B3LYP/ Feller, D.; Peterson, K. A. J. Chem. Phys. 2001, 115, 2576. (d) Ruscic, B.;
aVDZ, BP86/aVDZ, PBE/aVDZ, and PW91/aVDZ levels, T1 Wagner, A. F.; Harding, L. B.; Asher, R. L.; Feller, D.; Dixon, D. A.;
diagnostics calculated at the CCSD(T)/aVQZ level, contribu- Peterson, K. A.; Song, Y.; Qian, X.; Ng, C.; Liu, J.; Chen, W.; Schwenke,
tions to the relative energies on the different potential energy D. W. J. Phys. Chem. A 2002, 106, 2727. (e) Feller, D.; Dixon, D. A. J.
Phys. Chem. A 2003, 107, 9641. (f) Pollack, L.; Windus, T. L.; de Jong,
surfaces, CCSD(T) relative energies for the potential energy W. A.; Dixon, D. A. J. Phys. Chem. A 2005, 109, 6934. (g) Feller, D.;
surfaces as a function of the basis set, x, y, and z coordinates Peterson, K. A.; Dixon, D. A. J. Chem. Phys. 2008, 129, 204015.
of molecules at the MP2/aVTZ level (), complete potential (33) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(34) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
energy surface for the hydrolysis of TiCl4 to produce TiO2
(35) Frisch, M. J. Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
with different calculation methods (kcal/mol), and plots of M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
the rate constant (k) vs temperature. This material is available Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
free of charge via the Internet at http://pubs.acs.org. Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
References and Notes X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
(1) Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements, 2nd Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
ed.; Pergamon Press: Oxford, U.K., 1984. Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
(2) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
AdVanced Inorganic Chemistry, 6th ed.; Wiley: New York, 1999. Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
(3) Thompson, T. L.; Yates, J. T. Chem. ReV. 2006, 106, 4428. Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
(4) Anpo, M.; Dohshi, S.; Kitano, M.; Hu, Y. In Metal Oxides: P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Chemistry and Applications; Fierro, J. L. G., Ed.; CRC Press: Boca Raton, Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
FL, 2006; pp 595-622. B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03,
(5) Gratzel, M. Nature 2001, 414, 338. revision E.01; Gaussian, Inc.: Wallingford, CT, 2004.
(6) Fujishima, A.; Hashimoto, K.; Watanabe, H. TiO2 Photocatalysis: (36) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. J. Chem. Phys.
Fundamentals and Applications; BKC, Inc.: Tokyo, 1997. 1992, 96, 6796.
(7) Hattori, A.; Yamamoto, M.; Tada, H.; Ito, S. Chem. Phys. Lett. (37) Dunning, T. H., Jr.; Peterson, K. A.; Wilson, A. K. J. Chem. Phys.
1998, 8, 707. 2001, 114, 9244.
(8) Sopyan, I.; Murasawa, S.; Hasimoto, K.; Fujishima, A. Chem. Phys. (38) Peterson, K. A.; Figgen, D.; Dolg, M; Stoll, H. J. Chem. Phys.
Lett. 1998, 10, 723. 2007, 126, 124101. Balabanov, N.; Peterson, K. A. J. Chem. Phys. 2005,
(9) Dayte, A. K.; Riegel, G.; Bolton, J. R.; Huang, M.; Prairie, M. R. 123, 064107.
J. Solid State Chem. 1995, 115, 236. (39) Amos, R. D.; Bernhardsson, A.; Berning, A.; Celani, P.; Cooper,
(10) Cervais, C.; Simth, M. E.; Pottier, A.; Jolivet, J. P.; Babonneau, F. D. L.; Deegan, M. J. O.; Dobbyn, A. J.; Eckert, F.; Hampel, C.; Hetzer,
Chem. Mater. 2001, 13, 462. G.; Knowles, P. J.; Korona, T.; Lindh, R.; Lloyd, A. W.; McNicholas, S. J.;
(11) Gonzalez, R. A.; Musick, C. D.; Tilton, J. N. Process for controlling Manby, F. R.; Meyer, W.; Mura, M. E.; Nicklass, A.; Palmieri, P.; Pitzer,
agglomeration in the manufacture of TiO2. U.S. Patent 5,508,015, April R.; Rauhut, G., Schutz, M.; Schumann, U.; Stoll, H., Stone, A. J.; Tarroni,
16, 1996. R.; Thorsteinsson, T.; Werner, H.-J. MOLPRO, a package of ab initio
(12) Deberry, J. C.; Robinson, M.; Pomponi, M. D.; Beach, A. J.; Xiong, programs designed by H.-J. Werner and P. J. Knowles, version 2006.1,
Y.; Akhtar, K. Controlled vapor phase oxidation of titanium tetrachloride Universitat Stuttgart, Stuttgart, Germany, and University of Birmingham,
to manufacture titanium dioxide. U.S. Patent 6,387,347, May 14, 2002. Birmingham, U.K.
(13) Akhtar, M. K.; Pratsinis, S. E.; Mastrangelo, S. V. R. J. Mater. (40) Peterson, K. A.; Woon, D. E.; Dunning, T. H., Jr. J. Chem. Phys.
Res. 1994, 9, 1241. 1994, 100, 7410.
(14) Shirley, R.; Liu, Y.; Totton, T. S.; West, R. H.; Kraft, M. J. Phys. (41) Moore, C. E. Atomic Energy LeVels As DeriVed from the Analysis
Chem. A 2009, 113, 13790. of Optical Spectra, Vol. 1, H to V; U.S. National Bureau of Standards
(15) Addamo, M.; Augugliaro, V.; Di Paola, A.; Garca-Lopez, E.; Circular 467; National Technical Information Service, COM-72-50282; U.S.
Loddo, V.; Marc, G.; Palmisano, L. Colloids Surf., A 2005, 265, 23. Department of Commerce: Washington, DC, 1949.
(16) Xia, B.; Huang, H; Xie, Y. Mater. Sci. Eng., B 1999, 57, 150. (42) Cox, J. D.; Wagman, D. D.; Medvedev, V. A. CODATA Key Values
(17) ORegan, B.; Moser, J.; Anderson, M.; Gratzel, M. J. Phys. Chem. for Thermodynamics; Hemisphere Publishing Corp.: New York, 1989. The
1990, 94, 8720. Liu, D.; Kamat, P. J. Phys. Chem. 1993, 97, 10769. heat of formation for Ti is given in this reference at 298 K. The heat of
(18) Stathatos, E.; Lianos, P.; del Monte, F.; Levy, D.; Tsiourvas, D. J. formation of Ti at 0 K is obtained from this value with the correction from
Sol-Gel Sci. Technol. 1997, 10, 83. 0 to 298 K given in ref 21.
(19) Li, S.; Dixon, D. A. J. Phys. Chem. A 2008, 112, 6646. (43) Curtiss, L. A.; Raghavacchari, K.; Redfern, P. C.; Pople, J. A.
(20) Li, S.; Hennigan, J. M.; Dixon, D. A. J. Phys. Chem. A 2009, 113, J. Chem. Phys. 1997, 106, 1063.
7861. (44) McQuarrie, D. A. Statistical Mechanics; University Science Books:
(21) Chase, M. W. NIST-JANAF Themochemical Tables, 4th ed.; J. Phys. Sausalito, CA, 2000.
Chem. Ref. Data, Monogr. 1998, No. 9, 1-1951. (45) (a) Dixon, D. A.; Gutowski, M. J. Phys. Chem. A 2005, 109, 5129.
(22) Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; (b) Matus, M. H.; Anderson, K. A.; Camaioni, D. M.; Autrey, S. T.; Dixon,
Halow, I.; Bailey, S. M.; Churney, K. L.; Nuttall, R. L. J. Phys. Chem. D. A. J. Phys. Chem. A 2007, 111, 4411.
Ref. Data 1982, 11, Suppl. 2. (46) Li, S.; Dixon, D. A. J. Phys. Chem. A 2010, 114, 2665.
(23) Pratsinis, S. E.; Bai, H.; Biswas, P.; Frenklach, M.; Mastrangelo, (47) Hildenbrand, D. L. J. Phys. Chem. A 2009, 113, 1472; High Temp.
S. V. R. J. Am. Ceram. Soc. 1990, 73, 2158. Mater. Sci. 1996, 35, 151.
(24) Karlemo, B.; Koukkari, P.; Paloniemi, J. Plasma Chem. Plasma (48) Mrino, Y.; Uehara, U. J. Chem. Phys. 1966, 45, 4543.
Process. 1996, 16, 59. (49) Webb, S. P.; Gordon, M. S. J. Am. Chem. Soc. 1999, 121, 2552.
(25) Hildenbrand, D. L.; Lau, K. H.; Mastrangelo, S. V. R. J. Phys. (50) Brunken, S.; Muller, H. S. P.; Menten, K. M.; McCarthy, M. C.;
Chem. 1991, 95, 3435. Thaddeus, P. Astrophys. J. 2008, 676, 1367.
(26) West, R. H.; Beran, G. J. O.; Green, W. H.; Kraft, M. J. Phys. (51) Lee, T. J.; Taylor, P. R. Int. J. Quantum Chem. Symp. 1989, 23,
Chem. A 2007, 111, 3560. 199.
7570 J. Phys. Chem. A, Vol. 114, No. 28, 2010 Wang et al.
(52) Steinfeld, J. I.; Francisco, J. S.; Hase, W. L. Chemical Kinetics (59) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1997, 78,
and Dynamics, 2nd ed.; Prentice Hall: Englewood Cliffs, NJ, 1999. 1396.
(53) Holbrook, K. A.; Pilling, M. J.; Robertson, S. H. Unimolecular (60) Burke, K.; Perdew, J. P. Wang, Y. In Electronic Density Functional
Reaction, 2nd ed.; Wiley: Chichester, U.K., 1996. Theory: Recent Progress and New Directions; Dobson, J. F., Vignale, G.,
(54) Wigner, E. Z. Z. Phys. Chem. B 1932, 19, 203. Das, M. P., Eds.; Plenum: New York, 1998.
(55) Skodje, R. T.; Truhlar, D. J. J. Chem. Phys. 1981, 85, 624. (61) Perdew, J. P.; Wang, Y. Phys. ReV. B 1991, 45, 13244.
(56) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (62) Gilbert, T. M. J. Phys. Chem. A 2004, 108, 2550.
(57) Perdew, J. P. Phys. ReV. B 1986, 33, 8822.
(58) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. ReV. Lett. 1996, 77,
3865. JP102020H