Microwave spectroscopy of MgO in the a3Π and X1Σ states

Article abstract of DOI:10.1016/j.molstruc.2006.02.026

Pure rotational transitions of MgO in the a³Π_i state and the ground X¹Σ⁺ state (v=1-3) were observed in the 210-400 GHz region. The MgO molecule was produced by a DC discharge in a mixture of Mg vapor and N₂O. A simultaneous deperturbation analysis for the X¹Σ⁺, a³Π, and A¹Π states was carried out for the observed transition frequencies including previous high-resolution data, where the energy levels were calculated by diagonalizing the energy matrix including the spin-orbit and orbit-rotation interactions among three electronic states with vibrational levels up to v=9. A centrifugal term of the spin-orbit interaction between the a³Π_i and X¹Σ⁺ states was newly introduced to improve quality of the fitting, to explain the observed Λ-type doubling of the a³Π_i state.

Full text of DOI:10.1016/j.molstruc.2006.02.026

Journal of Molecular Structure 795 (2006) 179–184
www.elsevier.com/locate/molstruc
Microwave spectroscopy of MgO in the a3P and X1S states
E. Kagi a, K. Kawaguchi b,
*
a Information Science, Hiroshima City University, Asaminami-ku, Hiroshima 731-3194, Japan
b Department of Chemistry, Faculty of Science, Okayama University, Tsushima-naka 3-1-1, Okayama 700-8530, Japan
Received 6 December 2005; received in revised form 1 February 2006; accepted 16 February 2006
Available online 29 March 2006
Dedicated to Gisbert Winnewisser on the occasion of his 70th birthday.
Abstract
Pure rotational transitions of MgO in the a3Pi state and the ground X1SC state (vZ1–3) were observed in the 210–400 GHz region. The MgO
molecule was produced by a DC discharge in a mixture of Mg vapor and N2O. A simultaneous deperturbation analysis for the X1SC, a3P, and
A1P states was carried out for the observed transition frequencies including previous high-resolution data, where the energy levels were
calculated by diagonalizing the energy matrix including the spin–orbit and orbit–rotation interactions among three electronic states with
vibrational levels up to vZ9. A centrifugal term of the spin–orbit interaction between the a3Pi and X1SC states was newly introduced to improve
quality of the fitting, to explain the observed L-type doubling of the a3Pi state.
q 2006 Elsevier B.V. All rights reserved.
Keywords: Microwave spectroscopy; Magnesium oxide; Rotational spectrum; Perturbation; Excited electronic state
1. Introduction
already published data and calculated the term values of
rotational levels [4].
The pure rotational transitions of MgO were first observed
by Steimle et al. [5] with a microwave-optical double
Magnesium oxide, MgO, possesses two very low-lying
excited electronic states, namely the a3P (2620 cmK1) and
A1P (3560 cmK1) states, where numbers in parentheses denote
the energy differences from the ground X1SC state. The
investigation of these two and the ground states is of interest
because of mixing by spin–orbit and orbit–rotation inter-
actions. Many theoretical and experimental studies have been
performed on these low-lying electronic states, as summarized
in a recent report [1]. In our previous study [2], the A1P–X1SC
transition was observed by Fourier transform absorption
spectroscopy, where a perturbation analysis was carried out
to determine the interaction constant between the A and X
states. Mu¨rtz et al. [3,4] reported Faraday laser magnetic
resonance (LMR) spectroscopy with CO lasers. The method is
a very sensitive and makes it possible to observe much weaker
a3P0,1–X1SC intercombination bands as well as A1P–X1SC
bands. They carried out a systematic analysis considering
interactions among three low-lying electronic states, including
¨
resonance (MODR) technique. Later, Torring and Hoeft [6]
observed 13 rotational transitions of 24MgO, 25MgO, and
26MgO in the vZ0–2 states with a conventional millimeter-
wave spectrometer, and determined molecular constants
precisely. However, the pure rotational transitions in a3Pi
have not been reported so far. Although the accuracy of the
LMR data [4] is better than 1!10K3 cmK1 (30 MHz),
millimeter-wave spectroscopy provides highly precise tran-
sition frequencies, which are useful for perturbation analysis.
In the present study, we observed the pure rotational
transitions in the a3Pi state and carried out extended
measurements in the vibrationally excited states (vZ1–3) of
X1SC. From a least-squares analysis, the molecular constants
for the three low-lying electronic states were derived, and
interactions among three states were analyzed.
2. Experimental details
The microwave spectrometer used in the present study is
shown in Fig. 1 schematically, which was constructed at
Nobeyama radio observatory and moved to Okayama
University. Millimeter wave radiation is obtained by two
Gunn oscillators [7] in the frequency range of 65–115 GHz
*
Corresponding author. Tel.: C81 86 251 7848; fax: C81 86 251 7853.
E-mail address: okakent@cc.okayama-u.ac.jp (K. Kawaguchi).
0022-2860/$ - see front matter q 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molstruc.2006.02.026

180
E. Kagi, K. Kawaguchi / Journal of Molecular Structure 795 (2006) 179–184
QMC
Liq.-He cooled
InSb Detector
Teflon lens
Multiplier (× 2 or 3 or 4)
Absorption cell
Gunn oscillator
Polarizer
Roof top mirror
Amp.
Multiplier ( × 4)
Harmonic Mixer
νLO
3.2-3.6 GHz
Cs time standard
10 MHz
PLL
IF
Frequency Synthesizer
HP8656B
~1GHz
Lock-in Amplifier
EG&G 5205
Control module
Reference
FM
5 kHz
Reference
IF Monitor
A-D Converter
Computer
Multifunction Synthesizer
70 MHz
Spectrum Analyzer
Synthesizer
NF WF 1956
Printer
GP-IB
Fig. 1. A schematic diagram of the millimeter and submillimeter wave spectrometer. The system consists of four basic parts: radiation source, absorption cell,
detector, and control/data acquisition system.
with output power of 30–100 mW. Source radiation above
115 GHz is generated by multiplying the fundamental
frequencies of Gunn oscillators using several multipliers. A
frequency from the Gunn oscillator is phase-locked to
the 88th–120th order harmonics of the reference signal
(0.8–0.9 GHz) generated by a frequency synthesizer (HP
8656B). Since the phase-lock system (XL) is operated with a
reference signal above 2.8 GHz, the synthesizer frequency is
four times multiplied to give 3.2–3.6 GHz and amplified. The
oscillator frequency of the synthesizer was calibrated with
the hydrogen maser or the cesium frequency standard at
Nobeyama radio observatory. About 10% of the Gunn
oscillator output is coupled into a Schottky diode harmonic
mixer, where the multiplied frequency of the synthesizer is
mixed to generate an intermediate frequency (IF) of 70 MHz.
Thus, the Gunn oscillator frequency is phase-locked by
comparing the IF signal with a 70 MHz reference signal. A
spectrum analyzer also monitors the 70 MHz IF signal.
amplifier (EG&G 5205) is operated for phase sensitive
detection (PSD) with 10 kHz, and spectral lines are observed
as second derivative line shapes of absorption profiles. The
output from the lock-in amplifier is fed into a 12 bit A/D
converter, and the data are stored in a hard disk of a personal
computer. The error of the frequency measurement was
estimated to be 30–50 kHz depending on the signal-to-noise
ratio.
The sensitivity of the spectrometer was checked
by observing the JZ31–30 transition of 18O13C34S at
343–654 MHz [8] in natural abundance (0.93 ppm). The
observation was carried out at the OCS pressure of 8.6 Pa
(65 mTorr) and at room temperature (293 K) using a 1-m long
cell, and the signal-to-noise ratio (S/N) of the spectrum was
estimated to be 20. The peak absorption coefficient of the
transition is calculated to be 1.91!10K8 cmK1, and hence the
minimum detectable peak absorption coefficient (S/NZ2) is
estimated to be 1.9!10K9 cmK1, which corresponds to the
minimum detectable numbers of 2!108 molecules cmK3 in
volume density, or 4!1010 cmK2 in column density.
The source radiation is fed into the absorption cell through a
feedhorn, a polarizing grid, and a Teflon lens, which also acts
as a vacuum seal of the cell. The polarizing grid is set to pass a
linear polarized light of the source radiation. At the end of the
cell, a rooftop mirror is mounted to reflect the radiation back
through the cell, rotating its polarization by 908.
The detector we employed is a liquid helium cooled InSb
hot electron bolometer (QMC QFI/2BI). The detector element
is set inside a magnet to have sensitivity in the frequency
region more than 600 GHz. The noise equivalent power is
0.77!10K12 W HzK1/2. The frequency of the source radiation
is modulated about 800 kHz at 330 GHz in full width applying
a bipolar square wave with a 5 kHz frequency. A lock-in
The arrangement of the high-temperature cell is shown in
Fig. 2. The central part of the cell was made by a stainless steel
pipe with 120 cm length and 66 mm inner diameter. The both
ends were connected with Pyrex glass tubes, which have side-
arm electrodes. The cell was evacuated with a mechanical
booster pump (SINKO SMB-300D) followed by a rotary pump
(EDWARDS E2M40). The pumping speed was about 10 l sK1
.
The MgO molecule was generated by a discharge in a
mixture of N2O and Mg vapor. Solid metal was vaporized in
the cell wrapped with two ceramic fiber heaters (WATLOW),
which can heat up to 1200 8C. The cell temperature is

E. Kagi, K. Kawaguchi / Journal of Molecular Structure 795 (2006) 179–184
181
Fig. 2. A schematic diagram of the discharge-coupled high temperature cell.
monitored with a thermocouple on the outside surface of the
stainless steel tube, and a feedback loop (CHINO DB500)
is used for stabilizing the temperature within 1–2 8C. It was
necessary to heat the tube to 630 8C for producing magnesium
vapor. For the production of the ground state MgO, the
discharge was not necessary, but the discharge was
indispensable to observe the spectrum in the a3Pi state.
The optimum partial pressures of Ar and N2O were 5.3 Pa
(40 mTorr) and 2.7 Pa (20 mTorr), respectively, and the
discharge current was 150 mA, where a Baratron pressure
gauge was used. A solenoid coil was set around the central part
of the cell to observe the Zeeman effect of a spectral line. The
current of 20 A was applied for the solenoid coil to generate
magnetic field of 40 G at the center of the cell.
eight transitions in vZ0–2 in the 68–270 GHz region. In the
present study, we observed newly 13 rotational transitions in
vZ1–3. Tables 1 and 2 list the observed rotational transition
frequencies of MgO in the a3Pi and the ground states from
210 to 400 GHz.
We first tried to analyze the observed spectrum in the a3Pi
state by a least-squares method using a conventional 3P
Hamiltonian [12]. The standard deviation of the least-squares
fitting was 2 MHz, and it was found that the rotational levels of
the a3Pi state were strongly perturbed by spin–orbit and orbit–
rotation interactions from nearby states. This indicated
necessity of a simultaneous deperturbation analysis for the
X1SC, a3Pi, and A1P states, as carried out by Mu¨rtz et al. [3,4]
This spectrometer was also used for observations of pure
rotational spectra of FeCO [9], NaSH [10], and vibrationally
excited MgNC [11].
3. Observed spectrum and analysis
According to the results of Mu¨rtz et al. [4], we started
a search for the JZ12–11 transition in the lowest spin
component UZ2 of the a3P state around 354.6 GHz.
Paramagnetic lines were observed with a doublet structure
separated by about 19 MHz because of the L-type doubling.
Their observed frequencies 353,929 and 353,910 MHz are
226–228 MHz lower than the predicted values. Transitions in
other spin components, UZ1 and 0, were observed with about
200 MHz lower frequencies than expected from term values of
the LMR study. Fig. 3 shows an example of the observed
spectrum of MgO in the a3Pi state.
Rotational transition frequencies in the vibrationally excited
states of the ground electronic state (X1SC) were predicted
¨
using the parameters of Torring and Hoeft [6] who observed
Fig. 3. Observed spectrum of the JZ13–12 rotational transition of MgO in the
a3P state measured near 383 GHz taken in one 15 s scan.

182
E. Kagi, K. Kawaguchi / Journal of Molecular Structure 795 (2006) 179–184
Table 1
Observed rotational transition frequencies of MgO in the a3Pi state (MHz unit)
J0
J00
UZ2
UZ1
UZ0
Obs.
Calc.–obs.
Obs.
Calc.–obs.
Obs.
Calc.–obs.
7
6
e
f
210 588.907
K0.028
7
6
7
8
e
f
235,991.407
235,985.771
324,450.566
324,435.806
353,929.035
353,910.062
383,402.806
383,378.877
0.086
K0.027
K0.029
0.045
240,651.232
240,116.132
330,782.128
330,063.344
360,803.689
360,026.812
390,812.534
389,979.278
K0.043
K0.087
K0.013
K0.006
0.006
243,383.371
243,635.316
334,536.801
334,868.712
364,898.636
365,254.347
0.082
0.056
8
7
11
11
12
12
13
13
10
10
11
11
12
12
e
f
0.016
K0.028
K0.065
K0.020
e
f
K0.038
0.030
0.007
e
f
K0.083
0.037
0.058
0.047
who treated optical and ultraviolet data, infrared Fourier
transform data, in addition to their LMR data.
constant. Similarly, the interaction between A1P and a3P is
considered as !vAjvaOAeAla. The orbit–rotation interaction
(HOR) between A1P and X1SC is expressed as
In the present study, we constructed an energy matrix
including the A1P, X1SC, and a3Pi states. Since the energy
matrix has been given by Ip et al. [13] in the case of one
vibrational state, we simply extended the matrix to include
vibrational levels up to vZ9 for each electronic state. The
rotational energy levels are expressed by standard formulae for
hA1P;vAjKBðJCLKCJKLCÞjX1SC;vXi
1=2
el
AX
1=2
ZKð2Þ hvAjBðrÞjvXiL ½JðJ C1Þꢀ ð1G1Þ=2
(5)
el
where L ZhA1PjLCjX1SCi corresponds to HeRlE in the
1
1
the P and SC states, as follows
AX
previous paper [2] and B(r)Zh/8p2cmr2. The overlap integrals
and hvAjB(r)jvXi were calculated among vibrational levels up
to vZ9 by using Le Roy’s RKR (Rydberg–Klein–Rees) and
LEVEL programs [14]. In the calculation, we used the
vibrational and rotational parameters reported by Mu¨rtz et al.
[4] as input data initially, and later new data determined by the
analysis were used, and the process was repeated until the
discrepancy between the input and determined parameters was
removed.
2
EðA1PÞ Z Eev CBv½JðJ C1ÞK1ꢀKDv½JðJ C1ÞK1ꢀ
GqvJðJ C1Þ=2
EðX1SCÞ Z Eev CBvJðJ C1ÞKDvJ2ðJ C1Þ
(1)
(2)
2
where Eev denotes the electronic and vibrational energy, and
Bv, Dv, qv are the rotational, centrifugal distorsion, the L-type
doubling constants, respectively. These parameters are
expressed by a general formula
We analyzed the pure rotational transitions observed in the
present work, together with the ro-vibrational transitions [15]
observed by infrared diode laser spectroscopy, the A1P –
X1SC, a3Pi–X1SC bands by LMR spectroscopy [3,4], the
A1P–X1SC band by FT-IR spectroscopy [2] and the rotational
transitions in X1SC[6]. The energy levels were calculated by
diagonalizing the energy matrix with the maximum dimension
of 50!50. Compared with the previous analysis of Mu¨rtz et al.
[4], we did not included optical and ultraviolet data, so the total
number of fitted lines were 594. The weight was set to be 1.0
for LMR zero-field data (65 lines) and diode laser data
(23 lines), 0.5 for FT data (460 lines), and (0.5K1)!106
for the microwave data. Three lines (a–X, vZ0–0, UZ0–0,
JZ8–9, 15–14, and A–X, vZ0–2, UZ1–0, JZ15–14) of the
LMR data were not included in the least squares analysis,
Pv Z Pe Cað1Þðv C1=2Þ Cað2Þðv C1=2Þ C/
(3)
2
Pe
Pe
with dependence of the vibrational quantum numbðe1rÞ v. In the
case of the electronic and vibrational energy, Pe; aPe , and aPð2eÞ
correspond to Te, ue, and Kuexe, respectively. In the
v-dependence of Bv, relations BeZPe, KaB ZaðP1eÞ, etc. hold
for each electronic state. For the rotational energy levels of the
3Pi state, Brown and Merer’s expression [12] is used.
The interaction matrix element between the a3P and X1SC
states is expressed as
el
aX
ha3P;vajHSOjX1SC;vXi Z hvajvXiA ð1G1Þ=2
where C and K signs refer to e and f levels, respectively, and
hvajvXi is an overlap integral and AeaXl the spin–orbit interaction
Table 2
(4)
Observed rotational transition frequencies of MgO in the vibrational excited states of X1SC (MHz unit)
J0
J00
vZ1
vZ2
vZ3
Obs.
Calc.–obs.
Obs.
Calc.–obs.
Obs.
Calc.–obs.
7
6
237,803.212
271,757.273
339,643.260
373,573.314
0.046
0.076
0.016
0.018
235,563.057
269,197.039
336,442.787
370,052.682
403,652.783
K0.012
0.015
233,223.138
K0.089
8
7
9
10
11
12
K0.033
K0.024
K0.040
333,096.403
366,369.977
399,633.248
K0.030
0.030
10
11
0.081

E. Kagi, K. Kawaguchi / Journal of Molecular Structure 795 (2006) 179–184
183
because of relatively large deviation compared with other data.
By using 33 parameters including the L-type doubling
parameter o in the a3P state, the fitting gave a standard
deviation of 0.0018 cmK1, but the L-type doubling of a3P
(UZ2) was not explained within the experimental measure-
ment accuracy of 30–50 kHz, where the deviations of the fitting
microwave lines, differences between calculated and observed
frequencies are listed in Tables 1 and 2.
Mu¨rtz et al. [4] considered influence of perturbing levels in
the calculations of each energy level, and the matrix was
truncated below vZ7 or 8 in most cases, but in the calculation
of the vZ4 level of X1SC, the perturbing levels were included
up to vZ11 of a3P. In the present study, we included all
vibrational levels up to vZ9 for each electronic state. When we
truncated the matrix below the level of vZ8, the standard
deviation of the fitting became 1.5 times larger than the case of
vZ9. Inclusion of the vZ10 level did not improve the quality
of the fitting. This may be due to uncertainty in energies of the
vibrational levels, especially in the a3P state, where only the
ue and uexe parameters have been used for describing higher
vibrational energy levels.
were about 200 kHz. So we introduced the following
el
aX
centrifugal term for A [16]:
el
aX
el
aX
A
/A CAJaXJðJ C1Þ:
(6)
By this term, the deviation in the L-type doubling of the
a3P state was less than 90 kHz, and the standard deviation of
the fitting became 0.00098 cmK1, where the o parameter was
not determinable and set to be zero in the final analysis. The
AJaX term may have a similarity with the Ad term for the spin–
orbit diagonal term A. The term is also included in the analysis
of a perturbation between a4SK and X2P of NHC [16]. The
4. Discussion
el
Aa
same kind of term may be present for A , and the inclusion
decreased the standard deviation of the fitting slightly. So the
term AJAa was not adjusted and set to zero in the final analysis.
The spin–rotation coupling term g in a3P was not determinable
in the fitting and set to be zero, because the AD term was
adjusted, and indeterminacy is hold between AD and g.
The present study reports the pure rotational spectrum of
MgO in the excited electronic state for the first time. The data
in the vibrational excited state of the ground electronic state
were also obtained, and those precise data are utilized for
understanding of perturbation between low-lying electronic
states. In our previous study of the A1P–X1SC band,
we recognized an effect of perturbation in the vZ1–3
levels of A1P, and the analysis needed higher-order L-type
doubling terms qJ[J(JC1)]2 and qJJ[J(JC1)]3, in addition to
the qvJ(JC1) term of Eq. (1). Therefore, an energy matrix
including the orbit–rotation interaction term was constructed
between a vibrational level of the A1P state and vibrational
levels of the ground electronic state. The interaction between
A1P and a3P was not considered in the analysis. Direct
diagonalization of the energy matrix explained the perturbation
to give ‘unperturbed’ molecular constants of the A1P state
The fitting is satisfactory in the point of the measurement
error, which has been reported to be 0.001 cmK1 for the LMR
data [4]. The standard deviation of the present analysis is
compared with the value of 0.0014 cmK1 obtained by the
previous analysis of FT data [2], although we used fewer
parameters than the previous analysis. Table 3 lists the
molecular constants determined for the three low-lying a3Pi,
A1P, and X1SC states. The parameter uexe of the a3Pi state
was fixed to the values reported by Ip et al. [13], because we
only used infrared data, which were not enough for
determining the parameter. For the presently observed
Table 3
Molecular constants for the MgO A1P, a3P, and X1SC states
ð1Þ
ð2Þ
ð3Þ
a
Pe
Pe
a
a
Pe
Pe
A1P
aP
T
3558.50124(86)
0.50544923(94)
0.117097(33)
K0.8536(39)
2618.9457(30)
0.502681(11)
0.120026(14)
K62.8431(20)
K0.10847(19)
0.015138(35)
K0.15667(62)
[0.0]
664.4360(11)
K3.92853(62)
0.00001786(16)
0.01374(11)
B
K0.00466439(74)
K0.000410(11)
K0.0468(37)
105D
104q
T
650.18028(77)
K0.004670(23)
[K4.2]a
B
105D
A
0.1088(36)
103AD
P
K0.000000280(14)
103q
X1SC
T
785.26262(89)
K5.12379(45)
0.000010458(48)
0.018831(97)
B
105D
0.57482854(42)
0.123371(20)
73.380(35)
K0.00532718(18)
el
aX
A
AJaX
K0.0002961(76)
60.597(13)
el
A
Aa
el
AX
0.8327(13)
L
ð1Þ
ð2Þ
The unit cmK1 except for L (dimensionless). The expansion Pv ZPe Ca ðvC1=2ÞCa ðvC1=2Þ C/is used in the fit. Numbers in parentheses denote one
el
AX
2
Pe
Pe
standard deviation.
a
Fixed to the value of Ip et al. [13].

184
E. Kagi, K. Kawaguchi / Journal of Molecular Structure 795 (2006) 179–184
without the qJ and qJJ terms. However, we needed different LAelX
5. Conclusion
RE
terms (H in Ref. [2]) in the vAZ1–3 states. The L-type
el
Pure rotational transitions in the excited electronic a3Pi vZ0
state of MgO were observed in the 210–400 GHz region.
Measurements of the ground state was extended up to vZ3.
These observed transitions (23 lines in a3P and 13 lines in
X1SC) were subjected to a simultaneous analysis with previous
high-resolution infrared and microwave data. Highly accurate
microwave data make it possible to determine molecular
constants describing interactions among three low-lying
electronic states.
doubling constant q was also changed 30% from vAZ1 to 3. In
the present study, only one parameter of L was enough for
el
AX
explaining the perturbation, and the v-dependence term of the
L-type doubling constant was only 5% of the q value.
Mu¨rtz et al. [4] considered the r-dependence of the
electronic interaction between three electronic states, that is,
they used hvajHSO(r)jvXi, hvAjHOR(r)jvXi and hvAjLC(r)jvai
instead of hvajvXi, hvAjB(r)jvXi, and hvAjvai, respectively. On
the other hand in the present analysis, instead of the
r-dependence, the centrifugal term for the off-diagonal spin–
orbit interaction was considered to obtain a satisfactory fitting.
By this inclusion, the determined parameters listed in Table 3
Acknowledgements
We thank Professor R.J. Le Roy for computer programs to
calculate the RKR potential and Franck-Condon factors. The
present study was partially supported by a Grant-in-Aid from
the Ministry of Education, Culture, Sports, Science and
Technology of Japan (Grant No. 13440181).
differ only slightly from those of Mu¨rtz et al. [4]. However,
el
relatively large differences are recognized in parameters A
aX
and aðA1Þ which is the vibrational dependence term of A in a3P.
Mu¨rtz et al. [4] reported aðA1ÞZK0:224ð14Þ cmK1, and the value
well explained the data included in the present least squares
analysis except for two LMR lines of a3P–X1SC, vZ1–0, UZ
0–0, JZ8–9 and 11–12. By inclusði1oÞn of these two lines, the
References
parameter was determined to be aA Z0:1495ð35Þ cmK1 with
[1] C.B. Dreyer, J.W. Daily, A. Abbud-Madrid, M.C. Branch, Appl. Opt. 40
(2001) 2561.
el
aX
the opposite sign. The A value determined in the present
study was 73.683 (34) cmK1 which is larger than the value
[2] E. Kagi, T. Hirano, S. Takano, K. Kawaguchi, J. Mol. Spectrosc. 168
(1994) 109.
el
A
A
Z68:73ð4Þ cmK1 of Mu¨rtz et al. [4], but close to the value
aX
el
Z72:4ð3Þ cmK1 of Ip et al. [13].
[3] P. Mu¨rtz, S. Richter, C. Pfelzer, H. Thu¨mmel, W. Urban, Mol. Phys. 82
(1994) 989.
aX
The majority of the microwave spectroscopic studies of
[4] P. Mu¨rtz, H. Thu¨mmel, C. Pfelzer, W. Urban, Mol. Phys. 86 (1995) 513.
[5] T.C. Steimle, Y Azuma, P.G. Carrick, Astrophys. J. 277 (1984) L21.
¨
[6] T. Torring, J. Hoeft, Chem. Phys. Lett. 126 (1986) 477.
MgO have been motivated by the search for magnesium-
containing interstellar species. The cosmic abundance of Mg is
relatively large: almost same as Si and 10 silicon containing
species have been detected so far. Among these Si compounds,
SiO is known to be a ubiquitous molecule in relatively high
temperature regions in various interstellar clouds and circum-
stellar envelopes. This may imply existence of MgO with
detectable abundance in interstellar space or circumstellar
envelope. Many searches for MgO have been carried out, and
most recent one is reported by Sakamoto et al. [17], where the
[MgO]/[SiO] abundance ratio is reported to be !10K1.9 in the
foreground molecular clouds toward the galactic center region.
Lack of effective formation mechanism of MgO in interstellar
space may explain the low MgO/SiO abundance ratio in the gas
phase. It is interesting to note that the first Mg-bearing
molecule detected by radio observations was MgNC in the
circumstellar envelope of a late type star IRCC10216 [18]. In
the case of Si, the corresponding CN compounds SiCN and
SiNC, where SiCN has a lower energy than SiNC, are less
abundant than MgNC by a factor of 20 in IRCC10216 [19,20].
The present ground state data may be useful for further
sensitive search for MgO in higher frequency region.
[7] J.E. Carlstrom, R.L. Plambeck, D.D Thornton, IEEE Trans. Microwave
Theory Technol. (1985) 610 MTT-33.
[8] Z. Kisiel, D.J. Millen, J. Phys. Chem. Ref. Data 11 (1982) 101.
[9] E. Kagi, Y. Kasai, H. Ungerechts, K. Kawaguchi, Astrophys. J. 488
(1997) 776.
[10] E. Kagi, K. Kawaguchi, Astrophys. J. 491 (1997) L129.
[11] E. Kagi, K. Kawaguchi, J. Mol. Spectrosc. 199 (2000) 309.
[12] J.M. Brown, A.J. Merer, J. Mol. Spectrosc. 74 (1979) 488.
[13] P.C.F. Ip, K.J. Cross, R.W. Field, J. Rostas, B. Bourguignon,
J. McCombie, J. Mol. Spectrosc. 146 (1991) 409.
[14] R.J. Le Roy, LEVEL 7.7: A Computer Program for Solving the Radial
¨
Schrodinger Equation for Bound and Quasibound Levels, University of
Waterloo Chemical Physics Research Report CP-661, 2005. see the
‘Computer Programs’ link at http://leroy.uwaterloo.ca
[15] S. Civis, H.G. Hedderich, C.E. Blom, Chem. Phys. Lett. 176 (1991) 489.
[16] K. Kawaguchi, T. Amano, J. Chem. Phys. 88 (1988) 4584.
[17] S. Sakamoto, G.J. White, K. Kawaguchi, M. Ohishi, K.S. Usuda,
T. Hasegawa, Mon. Not. R. Astron. Soc. 301 (1998) 872.
[18] K. Kawaguchi, E. Kagi, T. Hirano, S. Takano, S. Saito, Astrophys. J. 406
(1993) L39.
´
[19] M. Guelin, S. Muller, J. Cernicharo, A.J. Apponi, M.C. McCarthy,
C.A. Gottlieb, P. Thaddeus, Astron. Astrophys. 363 (2000) L12.
´
[20] M. Guelin, S. Muller, J. Cernicharo, M.C. McCarthy, P. Thaddeus,
Astron. Astrophys. 426 (2004) 49.