Matrix isolation infrared spectroscopic and density functional theoretical study of the aluminum, gallium and indium borates

Article abstract of DOI:10.1016/j.chemphys.2005.01.024

The aluminum, gallium and indium borates have been produced by co-deposition of metal and boron atoms with oxygen molecules in excess argon at low temperature. The metal borate molecules were identified on the basis of the experimentally observed vibrations with isotopic effects (¹⁶O/ ¹⁸O substitutions, and ¹⁰B/¹¹B and ⁶⁹Ga/⁷¹Ga in natural abundance) as well as density functional calculations. The aluminum, gallium and indium borates were predicted to have linear geometries with a highly ionic electronic structure.

Full text of DOI:10.1016/j.chemphys.2005.01.024

Chemical Physics 313 (2005) 325–329
www.elsevier.com/locate/chemphys
Matrix isolation infrared spectroscopic
and density functional theoretical study of the aluminum,
gallium and indium borates
*
Guangjun Wang, Mohua Chen, Gongyu Jiang, Han Zhou, Mingfei Zhou
Department of Chemistry, Laser Chemistry Institute, Shanghai Key Laboratory of Molecular Catalysts and Innovative Materials,
Fudan University, No. 220, Handan Road, Shanghai 200433, PR China
Received 18 November 2004; accepted 24 January 2005
Available online 13 February 2005
Abstract
The aluminum, gallium and indium borates have been produced by co-deposition of metal and boron atoms with oxygen mol-
ecules in excess argon at low temperature. The metal borate molecules were identified on the basis of the experimentally observed
vibrations with isotopic effects (16O/18O substitutions, and 10B/11B and 69Ga/71Ga in natural abundance) as well as density func-
tional calculations. The aluminum, gallium and indium borates were predicted to have linear geometries with a highly ionic elec-
tronic structure.
Ó 2005 Elsevier B.V. All rights reserved.
1. Introduction
The reactions of group 13 metal atoms with oxygen
and the geometrical, electronic and bonding properties
of small metal oxide clusters provide useful information
in understanding the structure and bonding between
group 13 metal and oxygen, and have been extensively
studied both experimentally and theoretically [6–18]. It
was found that the boron atoms reacted with oxygen to
form the linear OBO insertion molecule in solid argon
[6], while the Al, Ga and In atoms reacted with O2 to form
the cyclic M(O2), which rearranged to the more stable lin-
ear OMO isomer upon excitation [7]. The previous stud-
ies were focused on the binary metal oxide species, and
the ternary oxides of group 13 metals have not been re-
ported. In this paper we present a matrix isolation infra-
red spectroscopic and density functional theory study of
the MOBO (M = Al, Ga, and In) molecules.
The oxides of group 13 (boron, aluminum, gallium
and indium) possess versatile structures and functions,
and are widely used in catalytic and optical material
fields. Recent investigations have shown that hetero-
atom incorporation or doping can improve the proper-
ties of group 13 metal oxide materials [1–5]. For
example, Al2O3 is a widely used catalyst and catalyst
supporter for petroleum refining, coal liquefaction and
organic synthesis. It is found that incorporation of bor-
on into Al2O3 can improve the activity or selectivity of
the catalyst [1–3]. As a kind of glass, some borates of al-
kali, alkaline earth or rare earth metals are of interest
from a view point of nonlinear optical property. It is re-
ported that some properties of the borates were signifi-
cantly improved when part of B was substituted by Al,
Ga or In [4,5].
2. Experimental and computational methods
*
Corresponding author. Tel.: +86 2165643532; fax: +86
The experimental setup for pulsed laser ablation and
matrix isolation infrared spectroscopic investigation has
2165643532.
E-mail address: mfzhou@fudan.edu.cn (M. Zhou).
0301-0104/$ - see front matter Ó 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.chemphys.2005.01.024

326
G. Wang et al. / Chemical Physics 313 (2005) 325–329
been described previously [19]. Briefly, the 1064 nm
Nd:YAG laser fundamental (Spectra Physics, DCR
150, 20 Hz repetition rate and 8 ns pulse width) was fo-
cused onto the rotating metal targets through a hole in a
CsI window, which was mounted on a cold tip of a
closed-cycle helium refrigerator (Air Products, Model
CSW202). Typically, 5–10 mJ/pulse laser energy was
used. The ablated boron and metal atoms were co-
deposited with molecular oxygen in excess argon onto
the 12 K CsI window for one to two hours at a rate of
approximately 5 mmol/h. The mixed metal targets were
prepared by pressing the mixtures of boron and metal
powders. Natural abundance boron powders (10B:
19.8%, 11B: 80.2%, Merck), aluminum, gallium (69Ga:
60%, 71Ga: 40%) and indium particles (Shanghai Chem-
ical Reagent Corporation, >99.99%) were used in the
experiments. 16O2 and 18O2 (99%, Isotec) were used to
prepare the O2/Ar mixtures. After sample deposition,
infrared spectra were recorded on a Bruker IFS113v
spectrometer at 0.5 cmꢀ1 resolution with a DTGS detec-
tor in the spectral range of 4000–400 cmꢀ1. Matrix sam-
ples were annealed at different temperatures, and
selected samples were subjected to broadband irradia-
tion using a 250-W high-pressure mercury arc lamp.
Density functional theory calculations (DFT) were
performed using the Gaussian 03 program [20]. The
most popular BeckeÕs three-parameter hybrid func-
tional, with additional correlation corrections due to
Lee, Yang, and Parr was utilized (B3LYP) [21,22]. The
6-311+G* basis set for B, O, Al, and Ga, and Lanl2dz
pseudopotential and basis for In were used [23–25].
Geometries were fully optimized and vibrational fre-
quencies calculated with analytical second derivatives,
and zero point vibrational energies were derived.
product absorptions were produced. The band positions
are listed in Table 1. Representative infrared spectra in
the 2080–1960 cmꢀ1 region from co-deposition of
laser-ablated B/Ga atom mixtures (using a B:Ga molar
ratio = 2:1 target) with 0.2% O2 in argon are shown in
Fig. 1. The experiments were repeated with the 18O2,
18O2 + 16O2 and 18O2 + 18O16O + 16O2 samples, and
the isotopic counterparts of the new product absorp-
tions are also listed in Table 1. The spectra in selected
regions with different isotopic O2 samples after 30 K
annealing are shown in Figs. 2–4, respectively.
Reactions of B/Al, B/Ga, and B/In atoms with oxy-
gen molecules revealed new absorptions in the 2100–
1900 cmꢀ1 region (Table 1). Taking the B/Ga system
as an example, two weak doublet absorptions at
2063.4/2061.8 and 1993.6/1992.3 cmꢀ1 were observed
on sample deposition when a target mixed with natural
abundance boron and gallium was used (Fig. 1). These
absorptions increased together on 25 and 33 K anneal-
ing, and kept almost unchanged upon broadband irradi-
ation. The 2063.4/2061.8 cmꢀ1 doublet is approximately
25% in intensity of the 1993.6/1992.3 cmꢀ1 doublet,
which clearly indicates that only one boron atom is in-
volved in the vibrational mode. The relative intensities
of the 1993.6 and 1992.3 cmꢀ1 bands matched natural
isotopic abundance gallium (69Ga: 60%, 71Ga: 40%)
[26] and confirms that only one gallium atom is involved
in the mode. These absorptions shifted to 2038.4/2037.1
and 1963.5/1962.1 cmꢀ1 with 18O2 (Fig. 3, trace c). The
isotopic frequency ratios (10B/11B (69Ga): 1.0350,
16O/18O (11B69Ga): 1.0153) imply that these absorptions
are mainly due to terminal B–O stretching vibrations. In
the spectrum when a scrambled 18O2 + 18O16O + 16O2
sample was used (Fig. 3, trace b), each band split into
a quartet with approximately 1:1:1:1: relative intensities,
indicating that two inequivalent O atoms are involved in
the vibration. These observations demonstrated that the
new product molecule has the BGaO2 stoichiometry
with two inequivalent O atoms. Accordingly, we assign
the 2063.4/2061.8 and 1993.6/1992.3 cmꢀ1 absorptions
to the B–O stretching vibrations of the GaOBO
isotopomers.
3. Result and discussion
The reactions of laser-ablated boron, aluminum, gal-
lium and indium atoms with oxygen in solid argon have
been investigated previously [6,7]. We first repeated
these experiments, and the resulting infrared spectra
are about the same as those reported previously. When
the mixed B/Al, B/Ga or B/In targets were used, new
Similar absorptions at 2086.5 and 2016.1 cmꢀ1 in the
B/Al + O2 experiments (B:Al molar ratio: 1:2), and
Table 1
Infrared absorptions (cmꢀ1) of the MOBO (M = Al, Ga and In) molecules in solid argon
16O2
18O2
16O2 + 18O2
16O2 + 16O18O + 18O2
Assignment
2086.5
2016.1
2063.4
2061.8
1993.6
1992.3
2049.4
1979.9
2057.5
1985.5
2038.4
2037.1
1963.5
1962.1
2015.2
1950.4
2086.5, 2057.5
2016.1, 1985.5
2063.4, 2038.4
2061.8, 2037.1
1993.6, 1963.5
1992.3, 1962.1
2049.4, 2015.2
1979.9, 1950.4
2086.5, 2079.8, 2064.3, 2057.5
2016.1, 2009.1, 1993.2, 1985.5
2063.4, 2051.6, 2042.2, 2038.4
2061.8, 2049.5, 2040.5, 2037.1
1993.6, 1986.6, 1971.7, 1963.5
1992.3, 1985.2, 1970.1, 1962.1
2049.4, 2039.8, 2029.3, 2015.2
1979.9, 1972.1, 1959.0, 1950.4
AlO10BO
AlO11BO
69GaO10BO
71GaO10BO
69GaO11BO
71GaO11BO
InO10BO
InO11BO

G. Wang et al. / Chemical Physics 313 (2005) 325–329
327
0.05
0.04
0.03
0.02
0.01
0.00
0.06
(d)
Ga18O11B18
O
Ga18O10B18
O
(c)
0.04
0.02
0.00
GaO11BO
(c)
GaO10BO
(b)
(a)
(b)
(a)
Ga16O11B16
O
Ga16O10B16
O
2100
2050
2000
1950
2080
2060
2040
2020
2000
1980
1960
Wavenumber (cm-1)
Wavenumber (cm-1)
Fig. 3. Infrared spectra in the 2100–1950 cmꢀ1 region from co-
deposition of laser-ablated B and Ga atoms with isotopic substituted
O2 samples in argon. Spectra were taken after sample deposition
followed by 30 K annealing. (a) 0.2% 16O2, (b) 0.125% 16O2 + 0.25%
16O18O + 0.125% 18O2, and (c) 0.2% 18O2.
Fig. 1. Infrared spectra in the 2080–1960 cmꢀ1 region from co-
deposition of laser-ablated B and Ga atoms with 0.2% O2 in argon. (a)
1 h of sample deposition at 12 K, (b) after 25 K annealing, (c) after 33
K annealing, and (d) after 20 min of broad band irradiation.
Al18O11B18O
0.08
0.08
In18O11B18
O
Al18O10B18O
10B18O2-
(d)
(c)
0.06
0.04
0.02
0.00
In18O10B18
O
0.06
(d)
0.04 (c)
(b)
(a)
Al16O11B16O
(b)
Al16O10B16O
0.02
10BO2-
2000
In16O11B16
O
In16O10B16
(a)
O
O11BOO
2010
2080
2040
1960
0.00
Wavenumber (cm-1)
2040
1980
-1
1950
Wavenumber (cm )
Fig. 2. Infrared spectra in the 2100–1940 cmꢀ1 region from co-
deposition of laser-ablated B and Al atoms with isotopic O2 samples in
argon. Spectra were taken after 2 h of sample deposition followed by
30 K annealing. (a) 0.2% 16O2, (b) 0.2% 16O2 + 0.2% 18O2, (c) 0.125%
16O2 + 0.25% 16O18O + 0.125% 18O2, and (d) 0.2% 18O2.
Fig. 4. Infrared spectra in the 2060–1940 cmꢀ1 region from co-
deposition of laser-ablated B and In atoms with isotopic substituted O2
samples in argon. Spectra were taken after 2 h of sample deposition
followed by 30 K annealing. (a) 0.2% 16O2, (b) 0.2% 16O2 + 0.2% 18O2,
(c) 0.15% 16O2 + 0.3% 16O18O + 0.15% 18O2, and (d) 0.2% 18O2.
2049.4 and 1979.9 cmꢀ1 in the B/In + O2 experiments
(B:In molar ratio:1:1) are assigned to the AlO10BO/
AlO11BO and InO10BO/InO11BO molecules, respec-
tively, following the example of GaOBO.
Table 2
˚
Calculated geometrical (bond length in A) and electronic properties of
MOBO (M = Al, Ga and In)
AlOBO
GaOBO
InOBO
Density functional theory calculations support the
experimental assignment and provide insight into the
electronic structure and bonding in MOBO (M = Al,
Ga and In). The B3LYP calculations predicted the
MOBO molecules to have linear structures. The geomet-
rical parameters and some electronic properties for
the computed equilibrium structure are presented in
Table 2. The calculations indicated that the linear MOBO
structure is the global minimum on the potential energy
surfaces of BMO2. We also computed all the other
possible structural isomers with BMO2 formula, such as
RM–O
RO–B
RB–O
E (a.u.)a
qM
1.763
1.307
1.218
1.880
1.303
1.221
2.013
1.299
1.224
ꢀ417.942799
ꢀ2100.357107
ꢀ177.381222
+0.884
+0.875
+0.918
a
After zero point energy corrections.
OBMO, BOMO, BMO2, MBO2, and cyclic-B(l-O2)M.
All of these isomers are higher in energy than MOBO
by more than 60 kcal/mol.

328
G. Wang et al. / Chemical Physics 313 (2005) 325–329
Table 3
Calculated vibrational frequencies (cmꢀ1) and intensities (km/mol) of the MO11BO molecules (M = Al, Ga and In)
Frequency (intensity, mode)
OBO (2Pg)þ
AlOBO (1R)
GaOBO (1R)
InOBO (1R)
422.2 (83, pu) 509.7 (65, pu) 1089.0 (0, rg) 1472.2 (474, ru)
597.1 (47, pu) 597.1 (47, pu) 1095.2 (0, rg) 1954.8 (740, ru)
80.7(0, p) 80.7(0, p) 512.0 (104, r) 539.0 (61, p) 539.0 (61, p) 1161.1 (231, r) 2057.0 (858, r)
8.1(0, p) 8.1 (0, p) 359.1 (102, r) 538.2 (66, p) 538.2 (66, p) 1129.9 (215, ) 2040.7 (924, r)
51.2 (0, p) 51.2 (0, p) 303.3 (81, r) 553.5 (63, p) 553.5 (63, p) 1124.2 (163, r) 2036.3 (942, r)
OBOꢀ ð Rg Þ
1
The calculated vibrational frequencies and intensities
evaporated boron and metal atoms with O2 in excess ar-
gon. Sample annealing allowed the BO2 molecules to
diffuse and react with metal atoms to form MOBO:
of the linear MOBO molecules are listed in Table 3. As
can be seen, the most intensive IR mode is the terminal
B–O stretching mode, which was computed at 2057.0,
2040.7 and 2036.3 cmꢀ1 for AlO11BO, 69GaO11BO and
InO11BO, respectively. The calculated frequencies
should be scaled by 0.980, 0.977 and 0.972 to reproduce
the experimentally observed frequencies. This is in ac-
cord with the expected accuracy of the B3LYP calcula-
tions [27]. The calculated isotopic frequency ratios are
also in excellent agreement with the experimental values.
Taking the GaOBO molecule for example, the calcula-
tions gave ratios of 69GaO10BO/69GaO11BO = 1.0356
and 69Ga16O11B16O/69Ga18O11B18O = 1.0154, while the
experimental values are 69GaO10BO/69GaO11BO =
1.0350 and 69Ga16O11B16O/69Ga18O11B18O = 1.0153.
Beside the terminal B–O stretching mode, the MO–
BO stretching mode was predicted to absorb around
1100–1200 cmꢀ1 region with appreciable IR intensities.
The calculated IR intensity of this mode is about 20%
of the intensity of the terminal B–O stretching mode.
This mode should be observed if the predicted relative
intensities are correct. However, the spectra in this fre-
quency region are quite clean, and no such absorption
was observed in this frequency region. We suspect that
the IR intensity of the MO–BO stretching mode is over-
estimated by the DFT/B3LYP calculations.
The MOBO molecules can be regarded as the borates
of aluminum (I), gallium (I) and indium (I). The natural
atomic charges calculated for MOBO indicate a large
amount charge transfer, indicating a mostly ionic bond-
ing. The natural atomic charges of the Al, Ga and In
atoms in MOBO are +0.884, +0.875, and +0.918,
respectively. Therefore, the MOBO molecules can be
denoted as M+(OBO)ꢀ. Consistent with that notion, the
B–O stretching frequencies of MOBO are significantly
higher than the antisymmetric BO stretching mode of
the linear neutral OBO molecule (1299 cmꢀ1 for 11BO2),
and are very close to that of the linear OBOꢀ anion
(1920 cmꢀ1 for 11BO2ꢀ) observed in solid argon [6].
Our experimental observations indicated that the
MOBO molecules are most likely formed by the associ-
ation reactions between metal atoms and BO2 molecules
on annealing, reactions (1)–(3). DFT calculations pre-
dicted that these charge transfer reactions are highly
exothermic and require no activation energy. BO2 is
the main reaction product from codeposition of laser-
Al þ OBO ! AlOBO
Ga þ OBO ! GaOBO
In þ OBO ! InOBO
DE ¼ ꢀ127:2 kcal=mol
DE ¼ ꢀ112:3 kcal=mol ð2Þ
DE ¼ ꢀ112:6 kcal=mol ð3Þ
ð1Þ
Although weak absorptions due to BO, BOB and
other metal oxides were observed on sample deposition,
the absence of intermediate absorptions in the mixed
16O2 + 18O2 experiments excludes the possibility for
the formation of MOBO via the BO + MO reactions.
Previous investigations have shown that silicon monox-
ide could form complexes with metal atoms such as Li,
Na and Ag [28,29]. However, no obvious absorptions
due to species such as MOB were observed in the present
experiments.
4. Conclusions
The vibrational spectra and structures of the alumi-
num, gallium and indium borates MOBO (M = Al,
Ga, and In) have been studied by matrix isolation infra-
red spectroscopy and density functional theory calcula-
tions. The MOBO molecules were produced by
co-deposition of metal and boron atoms with oxygen
molecules at low temperatures in solid argon. The metal
borate molecules were identified on the basis of the
experimentally observed vibrations with isotopic effects
(16O/18O substitutions, and 10B/11B and69Ga/71Ga in
natural abundance). These metal borate molecules were
predicted to have linear geometries with a highly ionic
electronic structure.
Acknowledgements
We greatly acknowledge financial support from
NNSFC (20125311) and the NKBRSF of China.
References
[1] D. Li, T. Sato, M. Imamura, H. Shimada, A. Nishijima, Appl.
Catal. B: Environ. 16 (1998) 255.

G. Wang et al. / Chemical Physics 313 (2005) 325–329
329
[2] G. Murali Dhar, B.N. Srinivas, M.S. Rana, K. Manoj, S.K.
Maity, Catal. Today 86 (2003) 45.
[3] O.V. Buyevskaya, M. Baerns, Catal. Today 42 (1998) 315.
[4] H. Kimura, T. Numazawa, M. Sato, Cryst. Growth 174 (1997)
308.
[5] R.W. Smith, M.A. Kennard, M. Dudik, J. Mater. Res. Bull. 32
(1997) 649.
[6] T.R. Burkholder, L.J. Andrews, Chem. Phys. 95 (1991) 8697.
[7] L. Andrews, T.R. Burkholder, J.T. Yustein, J. Phys. Chem. 96
(1992) 10182.
[16] M.V. Pak, M.S. Gordon, Chem. Phys. Lett. 344 (2001) 236.
[17] C. Chang, A.B.C. Patzer, E. Sedlmayr, T. Steinke, D. Sulzle,
Chem. Phys. Lett. 324 (2000) 108.
[18] Y.X. Bu, X.Y. Song, J. Chem. Phys. 113 (2000) 4216.
[19] M.H. Chen, X.F. Wang, L.N. Zhang, M. Yu, Q.Z. Qin, Chem.
Phys. 242 (1999) 81.
[20] M.J. Frisch et al., GAUSSIAN 03, Revision B.05, Gaussian, Inc.,
Pittsburgh, PA, 2003.
[21] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
[22] C. Lee, E. Yang, R.G. Parr, Phys. Rev. B 37 (1988) 785.
[23] A.D. McLean, G.S. Chandler, J. Chem. Phys. 72 (1980)
5639.
[8] L.B. Knight, T.J. Kirk, J. Herlong, J.G. Kaup, E.R. Davidson, J.
Chem. Phys. 107 (1997) 7011.
[9] S.R. Desai, H.B. Wu, L.S. Wang, Int. J. Mass Spectrom. 159
(1996) 75.
[24] R. Krishnan, J.S. Binkley, R. Seeger, J.A. Pople, J. Chem. Phys.
72 (1980) 650.
[10] S.R. Desai, H.B. Wu, C.M. Rohlfing, L.S. Wang, J. Chem. Phys.
106 (1997) 1309.
[25] P.J. Hay, W.R. Wadt, J. Chem. Phys. 82 (1985) 299.
[26] CRC Handbook, CRC Press, Boca Raton, FL, 1985.
[27] See for example: I. Bytheway, M.W. Wong, Chem. Phys. Lett.
282 (1998) 219.
[28] R. Koppe, H. Schnockel, Heteroatom Chem. 3 (1992) 329.
[29] B. Tremblay, M.E. Alikhani, L. Manceron, Chem. Phys. 218
(1997) 37.
[11] E.F. Archibong, R. Sullivan, J. Phys. Chem. 99 (1995) 15830.
[12] E.F. Archibong, R. Sullivan, J. Phys. Chem. 100 (1996) 18078.
[13] E.F. Archibong, A. St-Amant, J. Phys. Chem. A 102 (1998) 6877.
[14] E.F. Archibong, A. St-Amant, J. Phys. Chem. A 103 (1999) 1109.
[15] M.V. Pak, M.S. Gordon, J. Chem. Phys. 118 (2001) 4471.