Simultaneous Observations of H2O and SiO Masers toward OH/IR Stars

We performed simultaneous observations of SiO v = 1, 2, 29SiO v = 0, J = 1–0, and H2O maser lines toward 252 OH/IR stars using the individual 21 m telescopes of the Korean VLBI Network (KVN). The observations for studying SiO and H2O maser properties associated with the different evolutionary stages of OH/IR stars were carried out from 2011 November to 2012 July. Both H2O and SiO masers were detected from 50 sources with a detection rate of 20% in one epoch of observation. One-sided SiO maser emissions without H2O were detected from 108 sources and H2O maser emission was detected from 11 sources, of which the detection rates were 43% and 4%, respectively. The overall detection rate of the SiO maser was 63%, and that of the H2O maser was 24%. There were 65 new detections in the SiO maser lines, 22 new detections in the H2O maser line, and 4 new detections in the 29SiO maser line. For the H2O and/or SiO maser- detected sources, mutual relations between SiO and H2O maser properties (including peak and integrated antenna temperatures, and full widths at zero power, etc.) are investigated based on a statistical analysis. We also investigate these maser properties on an IRAS two-color diagram related to stellar evolutionary sequences. In particular, a large number of SiO v = 2-only detected sources appear among the SiO-only detected sources compared to those of both H2O and SiO maser detected sources and also appear in the later evolutionary stages of asymptotic giant branch AGB stars in the IRAS two-color diagram. These results may be associated with the development of a hot and thick dust envelope at later stages of AGB evolution and the different excitation conditions of SiO v = 1 and v = 2 masers. Our observational results will be useful for statistical studies of circumstellar envelopes of OH/IR stars related to their late evolution and future very long baseline interferometry (VLBI) observations.


Introduction
OH/IR stars are surrounded by thick dust layers and represent asymptotic giant branch (AGB) stars with tremendous mass-loss rates of approximately 10 −5 to 10 −4  M yr −1 (Baud & Habing 1983). These stars are generally optically opaque and bright in the infrared. OH/IR stars are classified into two types, type I stars, which are strong in the 1665 and 1667 MHz OH masers, and type II stars, which are strong in the double-peak 1612 MHz OH maser. In initial OH/IR star studies (Wilson & Barrett 1972), Bowers & Hagen (1984) and Baud & Habing (1983) performed 1612 MHz OH maser observations for studying the Galactic dynamics because the centers of OH maser double-peaks provide the stellar velocity. After that, based on the IRAS catalog, extensive observations were performed by many researchers, including Chengalur et al. (1993), David et al. (1993), Le Squeren et al. (1992, te Lintel Hekkert et al. (1991), Lewis (1990), and Eder et al. (1988), who all found a large number of new OH/IR stars. Additionally, OH/IR star catalogs were created by te Lintel Hekkert et al. (1989) and Benson et al. (1990), and a database was created by Engels (see http://www.hs.uni-hamburg.de/ maserdb).
OH/IR stars have both H 2 O and SiO masers together with OH masers. Therefore, these masers are very useful probes for investigating the stratified structure of OH/IR stars because OH, H 2 O, and SiO masers trace different distances from the central star according to their different chemical and physical conditions (Reid & Moran 1981). Engels & Lewis (1996) performed a survey of 22 GHz H 2 O masers using the Arecibo set of 382 OH/IR stars and obtained 181 new H 2 O maser detections. Sjouwerman et al. (2002) detected H 2 O masers from 25 sources and 43 GHz SiO masers from 18 sources in the Galactic center OH/IR stars using the Very Large Array (VLA). Nyman et al. (1986Nyman et al. ( , 1993Nyman et al. ( , 1998 examined the characteristics of SiO masers in OH/IR stars through observations of SiO v=1, 2 J=1-0, and v=1, J=2-1 maser lines. However, these observations of H 2 O and SiO masers were carried out at different times and using different telescopes and receivers. Therefore, the observational data obtained from different epochs and different systems have significant limits for statistical studies on rapidly varying maser phenomena. On the other hand, simultaneous observations of H 2 O 22 GHz and SiO 43,86,and 129 GHz bands with the Korean Very Long Baseline Interferometry Network (KVN) can overcome these limitations and provide homogeneous data for the combined statistical study of H 2 O and SiO masers because KVN has adopted a quasi-optical system for simultaneous observations at the above four bands (Han et al. 2008). In this study, we aim to investigate the correlations and differences among the H 2 O and SiO masers in OH/IR stars based on simultaneous observational results of 22 GHz H 2 O and 43 GHz SiO masers at the initial stage of KVN operation. This kind of work with the         We present our source selection and observations in Section 2. In Section 3 we present our observational results. Section 4 presents a discussion including a comparison between H 2 O and SiO masers in terms of the relative intensity and line width, and a comparison of the observed source and IRAS two-color diagram, as well as an explanation of the characteristics of these properties. Section 5 presents a summary.

Source Selection
In order to obtain as many samples of statistical analyses as possible for H 2 O and/or SiO maser properties from OH/IR stars, we selected our observing sources following two steps. As the first step, we selected 170 sources that included many H 2 O maser detected sources. Namely, 74 sources with more than 0.6 Jy and 30 sources with 0.2 ∼0.6 Jy in H 2 O maser emission were selected among 197 H 2 O maser detected sources (Engels & Lewis 1996). We also included 55 sources that were not detected in H 2 O maser emission, considering their variability and detection upper limits. In addition, 11 H 2 O and/or SiO maser detected sources from Engels et al. (1986) and Engels & Lewis (1996) were added.
As the second step, we added 65 sources from Chen et al. (2001), who conducted a comparative analysis of 1065 OH/IR stars, by comparing the sources that were detected to emit the 1612 MHz OH maser and the IRAS LRS (low resolution spectra) from Kwok et al. (1997). In addition, we included four sources from Lewis et al. (2004), who established the corresponding relationship of Arecibo OH/IR stars with 2 MASS (2 Micron All Sky Survey) and MSX (Midcourse Space Experiment) lists, and 13 sources from te Lintel Hekkert et al.  Table 1 presents the information related to the 252 observational objects in ascending order of right ascension and declination. The first and second columns list the identification number and the name of the sources, and the third and fourth columns list their right ascension and declination. The fifth and sixth columns relay the stellar velocity (line-of-sight velocity) of each central star and pulsation period. The seventh and eighth columns contain the peak flux densities of H 2 O and SiO masers obtained by past observations from the reference papers in column 9.

Observations
Note. Position reference. NASA/IPAC Infrared Science Archive. observations were conducted from 2011 November to 2012 July using the 21 m KVN single-dish radio telescopes located at Yonsei University in Seoul, Ulsan University in Ulsan, and Tamna University on Jeju Island. The quasi-optic system of KVN enables us to perform simultaneous observations of H 2 O 22 GHZ and SiO 43, 86, 129 GHz through three low-pass filters (Han et al. 2008). However, only 22 GHz and 43 GHz band receivers were installed at the first stage of the KVN operation.
The half-power beam widths and aperture efficiencies were 122″/65% (22 GHz) and 64″/67% (43 GHz) for the KVN Yonsei telescope, 123″/62% (22 GHz) and 66″/59% (43 GHz) for the KVN Ulsan telescope, and 124″/66% (22 GHz) and 64″/60% (43 GHz) for the KVN Tamna telescope (Lee et al. 2011). We used 22 GHz and 43 GHz band HEMT (high electron mobility transistor) receivers with both left and right circular polarized feeds. In this observation, only the left circular polarized feed was used. The system noise temperature in each band corresponded to 80 K-290 K at 22 GHz and 140 K-350 K at 43 GHz, which was affected by the weather and altitude. The four 64 MHz bandwidths in the digital spectrometer were used for four maser lines of H 2 O  6 5 12 23 , SiO v=1, 2, J=1-0, and 29 SiO v=0, J=1-0 transitions. The velocity ranges of these bandwidths corresponded to 860 kms −1 at 22 GHz and 440 kms −1 at 43 GHz and velocity resolutions are 0.21 kms −1 (4096 channels at 22 GHz) and 0.22 kms −1 (2048 channels at 43 GHz). The pointing of the telescope was checked every two hours using nearby strong SiO maser sources.
The observational data were calibrated by the chopper wheel method, which corrects the changes in antenna gain and atmospheric attenuation for yielding an antenna temperature ( * T A ). Integration time was set to 40 minutes at least to 90 minutes at most in order to achieve sensitivities of ∼0.05 K. The conversion factors from the antenna temperature to the flux density are 12.27 JyK −1 at 22 GHz and 11.90 JyK −1 at 43 GHz, respectively. Figure 1 shows the spectra of 169 detected sources in the H 2 O and/or SiO v=1, 2, J=1-0 maser emission. The overall observational results are summarized in Table 2 Table 3, the first and second columns present the identification number and name of each source. The third through sixth columns show the peak antenna temperature for each maser. The seventh through tenth columns show the LSR velocity of the peak antenna temperature. The eleventh and twelfth columns list the peak antenna temperature ratios of the SiO v=2 maser to v=1 and those of the SiO v=1 maser to H 2 O. The thirteenth and fourteenth columns list the observation dates and telescopes used for observation (YS: KVN Yonsei telescope, US: KVN Ulsan telescope, TN: KVN Tamna telescope).

Observational Results
In Table 4, the third through sixth columns present the overall integrated intensities for both H 2 O and SiO maser detected sources. The seventh through tenth columns show the mean velocities for each maser. The eleventh and twelfth columns list the ratios of SiO v=2, J=1-0 and H 2 O integrated intensities with respect to those of the SiO v=1, J=1-0 maser. The thirteenth and and fourteenth columns list the observation date and telescopes used for observation. Table 5 lists the peak antenna temperatures and peak velocities for SiO-only maser detected sources and Table 6 lists the integrated antenna temperatures and mean velocities for SiOonly maser detected sources. Table 7 presents the peak antenna temperatures, peak velocities, integrated antenna temperatures, and mean velocities for H 2 O-only maser detected sources. Table 8 presents the upper limit of the antenna temperatures for detected sources with neither H 2 O nor SiO masers.
The SiO v=2-only maser was detected from 28 sources among 158 SiO maser detected sources in Table 2, while the SiO v=1-only maser was detected from 13 sources. As shown in Figure 1 and Table 3, the H 2 O maser shows various shapes of line profiles compared to SiO masers. In addition, the peak of the H 2 O maser was blueshifted or redshifted with respect to that of the SiO masers. Regarding the spectrum shapes of the H 2 O maser, in Table 9 we classified the shapes as one-way single-peak, double-peaks, and multiple-peaks, similar to Engels et al. (1986) and Kim et al. (2010). The characteristics of several individual sources can be described as follows.
V1018 Sco (OH354.88-0.54, IRAS 17317-3331, No. 41): this source is classified as IRAS LRS (low resolution spectra) 79 and has a long period of 1,433 days (Nyman et al. 1993 David et al. (1993). In the study by Cohen et al. (2005), it is classified as a recently formed planetary nebula (PN) because an Hα ring was found around a long-period variable star (V1018 Sco) with a 1612 MHz OH maser. Cohen et al. (2006)   8 the ratio of integrated intensity of SiO v=1, J=1-0 to v=1, J=2-1 maser is higher than 18 and the intensity ratio of SiO v=2, J=1-0 to v=1, J=1-0 maser is 4.4. The intensity ratio of SiO v=2, J=1-0 to the v=1, J=1-0 maser showed 1.96 in our observations, which is similar to the value of 2 in found for warmer OH/IR stars by Nyman et al. (1993).
OH20.7+0.2(IRAS18251-1048, No. 88): the H 2 O maser showed a noticeable double-peak line profile with intensities of 6.7 Jy and 2.4 Jy at 73.6 kms −1 and 107.7 kms −1 (Engels et al. 1986), respectively. However, our observation indicated that there was only a blueshifted single-peak at 73.5 kms −1 with an intensity of 0.4 K. It may be attributable to the variability of the H 2 O maser. The H 2 O maser intensity was comparable to that of the SiO v=2, J=1-0 maser. The SiO v=2, J=1-0 maser showed stronger intensity compared to v=1 maser (Tables 3 and 4). The intensity ratio of SiO v=2, J=1-0 to v=1, J=1-0 maser is 4.4. IRAS18282-0943 (No. 96): toward this OH/IR star, the SiO v=1, J=1-0 maser was detected as double-peaks at 83 kms −1 and 134.8 kms −1 , with a velocity separation of 51.8 kms −1 (Figure 1 and Table 3). The SiO v=2, J=1-0 maser was also detected as double-peaks at 81.7 kms −1 and 130.9 kms −1 , with a velocity separation of 49.2 kms −1 . However, the H 2 O maser was detected as a single-peak at 78.5 kms −1 . The right component of spectra No. 96 (Figure 1) was associated with the OH/IR star IRAS18282-0943 because the stellar velocity of this star is 130.3 kms −1 (Table 1). Therefore, the left component of the SiO maser accompanied by the H 2 O maser seems to originate from another evolved star within the KVN single-dish beam. The detection rate of this kind of double-SiO maser source was reported and discussed by Deguchi et al. (1999) and Shiki & Deguchi (1997).
V1365 Aql (OH32.8-0.3, No. 129): Cato et al. (1976) detected the H 2 O maser emission at 60 kms −1 and 77 kms −1 for the first time. The SiO v=1, J=1-0 maser was detected at 61.8 kms −1 for the first time by Jewell et al. (1984). However, we detected the SiO v=2-only maser without the v=1 detection at 62.1 kms −1 and the 29 SiO v=0. J=1-0 maser at 61.5 kms −1 . The H 2 O maser from this OH/IR star was detected at 73 kms −1 and 78.5 kms −1 as a redshifted emission with respect to the stellar velocity 60.7 kms −1 in our observations. Engels et al. (1986) detected the H 2 O maser as double-peaks at 47.8 kms −1 (4.9 Jy) and 71.7 kms −1 (0.6 Jy). The H 2 O maser showed an active variation in both intensity and peak velocity, and this needs long-term monitoring observations. IRAS18525+0210 (No. 135): the double-peaks of the 1612 MHz OH maser were detected at 51.4 kms −1 and 88.7 kms −1 by Chengalur et al. (1993) and the stellar velocity was determined at 70.2 kms −1 ( Table 1). Engels & Lewis (1996) detected the double-peaks of the H 2 O maser with the intensity of 1.0 Jy at 53.8 kms −1 and 7.8 Jy at 84.3 kms −1 . However, the H 2 O maser was not detected in our observations. The peak intensity ratios of the SiO v=2 to v=1, J=1-0 masers PT(SiO v=2)/PT(SiO v=1) were 4.87 and the integrated intensity ratios IT(SiO v=2/IT(SiO v=1) were 5.46, which were significantly higher than those of other sources.

Relative Intensities and Full Width at Zero Power between H 2 O and SiO Masers
In Figures 2 and 3, we examined some characteristic distributions for the sources in which both H 2 O and SiO masers were detected. Figures 2(a) and (b) presents the histograms for the numbers of sources with respect to the peak antenna temperature ratios between H 2 O and SiO (v=1, J=1-0) masers PT(H 2 O)/PT(SiO v=1) and the integrated antenna temperature ratios IT(H 2 O)/(IT(SiO v=1), respectively. In Figure 2(a), the number of sources with ratios between 0 and 1 is 23 (49%) and the number for those larger than 1 is 24 (51%). Namely, the number of sources in which the peak antenna temperatures of the H 2 O maser are weaker than those of the SiO maser are comparable to the number of sources in which the peak antenna temperatures of the H 2 O maser were stronger than those of the SiO maser. In Figure 2(b), there was a small number of sources (11 sources, 23%) in which integrated antenna temperatures of the H 2 O maser were stronger than those of the SiO maser. In particular, the number of sources in which the peak antenna temperature of the H 2 O maser is less than half of that where the SiO maser 16 sources in Figure 2( Figure 2 and Tables 3 and 4). The SiO maser intensities in OH/IR stars are stronger than those of the H 2 O maser in most sources, similar to the case of Mira variables. The average photon luminosities of the H 2 O and SiO masers in OH/IR stars are higher than those in Mira variables. These intensity ratios and photon luminosities of H 2 O and SiO masers between Mira variables and OH/IR stars were discussed by Kim et al. (2013Kim et al. ( , 2014. In Table 10, for both H 2 O and SiO maser detected sources (47 samples), the average peak and integrated antenna temperature ratios of the H 2 O maser with respect to the SiO (v=1) maser are given. We compare our results of the OH/IR sample stars with those of Kim et al. (2014) for both H 2 O and SiO detected sources (32 samples). The average values of our results for 47 sources are 1.38 for peak temperature ratios and 0.99 for integrated temperature ratios. These average values of Kim et al. (2014) for 32 sources are 1.56 and 1.09. Our values are slightly smaller than those of Kim et al. (2014), but do not represent a significant difference.  Figure 3 and Table 11 present the distribution of FWZP (full width at zero power) of H 2 O and SiO masers for investigating the development of expansion velocity from the SiO maser region to the H 2 O maser region in OH/IR stars. The FWZP is defined as the velocity extent of maser lines between outer blueshifted and redshifted emission boundaries at the 3σ level SiO with respect to stellar velocity, which reaches the baseline of the spectrum. In the case of one-way and double-peak spectra of the H 2 O maser, we measured the FWZP of the one-way spectrum as two times the velocity extent of a one-way spectrum (blueshifted or redshifted) with respect to stellar velocity, and measured the FWZP of the double-peak spectrum as the velocity extent between the outer blueshifted emission boundary of the blueshifted component and the outer redshifted emission boundary of the redshifted component. The FWZP of the H 2 O maser shows a broad feature compared to that of the   Kim et al. (2013). The differences of the FWZP distributions between OH/IR stars and Mira variables may be caused by the differences of the mass-loss rates between them. The OH/IR stars with a very extended envelope show higher mass-loss rates than Mira variables, as indicated by Kim et al. (2013).
The H 2 O maser emission can be divided into type A and type B depending on the shape of the line profile (Engels & Lewis 1996). Type A has a single-peak spectrum close to the   The Astrophysical Journal Supplement Series, 232:13 (36pp), 2017 September stellar velocity, while type B has a double-peak spectrum with respect to the stellar velocity. The number of type B H 2 O maser spectra in OH/IR stars is larger than that in Mira variables (Engels & Lewis 1996). Our observational results show a similar tendency. However, we can add type C, which shows one-way H 2 O maser emission (blueshifted or redshifted) with respect to stellar velocity (Kim et al. 2010). The number of type C H 2 O maser spectra in OH/IR stars is also larger than that in Mira variables. The sources with type B and C spectra ( Table 9) will be discussed in Section 4.3. IRAS two-color diagram of SiO and H 2 O observed sources.   Figure 4, the peak and integrated antenna temperatures of the SiO v=2 maser are stronger than those of the SiO v=1 maser in 20 sources (49%) for peak temperatures and in 11 sources (27%) for integrated temperatures among 41 of both SiO and H 2 O maser detected sources. Meanwhile, the peak and integrated antenna temperatures of the SiO v=2 maser are stronger than those of the SiO v=1 maser in 42 sources (56%) for peak temperatures and in 38 sources (51%) for integrated temperatures among 75 SiO-only maser detected sources, as shown in Figure 5. Namely, in SiO-only maser detected sources, there are larger numbers of sources that have a stronger intensity for the SiO v=2 maser compared to that of the v=1 maser.

Relative Intensities between SiO v=1 and 2, J=1-0 Masers
In Figure 4, V1016 Sco (No. 41) and OH20.7+0.2 (No. 88) had a peak temperature ratios of 1.96 and 3.25 and integrated temperature ratios of 2.71 and 2.18, respectively, which were significantly larger values than those of other sources. They were positioned in a more evolved region compared to other sources in the IRAS two-color diagram ( Figure 6).
In Figure 5, 75 SiO-only maser detected sources show a slightly different tendency from those in Figure 4. There are more sources that have stronger intensities for the SiO v=2 maser compared to the intensity of the v=1 maser. In relation to the peak temperature, there are 42 sources (about 56%) with a stronger SiO v=2 (1) maser, among 75 sources. Regarding the integrated temperature ratios, there are 38 sources (51%) with a stronger SiO v=2 maser. Among them, V669 Cas (No. 3), IRAS18525+0210 (No. 135), and IRAS19085+0755 (No. 163) had peak temperature ratios of 3.00, 4.87, and 3.55, and integrated temperature ratios of 4.45, 5.46, and 5.56, respectively, that were significantly larger values than those from other sources.
In Table 10 intensity with respect to the v=1 intensity are higher in SiO-only maser detected sources than those in both H 2 O and SiO maser detected sources. The average values of those peak temperature and integrated temperature ratios were estimated to be 1.10 and 1.13, respectively, for 116 OH/IR stars that included 41 both H 2 O and SiO maser detected sources and 75 SiO-only maser detected sources. Lindqvist et al. (1991) showed that the average value of IT(SiO v=2)/ IT(SiO v=1) is ∼1.5 for 11 OH/IR stars. Nyman et al. (1986Nyman et al. ( , 1993 also showed that these average values are ∼4 for 5 OH/IR stars, and ∼1.8 for 19 OH/IR stars, respectively. Unlike Mira variable stars, in OH/IR stars there were more sources with stronger v=2 maser emission compared to v=1 maser emission. Such a tendency supports the OH/IR stars with thicker dust formation layer tending to have a larger value of IT(SiO v=2)/IT(SiO v=1). Our results are in good agreement with these tendencies.
All results need to be discussed from the viewpoint of a different evolutionary stage associated with a different excitation condition and pumping mechanism in the following two-color diagram ( Figure 6). Figure 6 shows the IRAS two-color diagram for our sample 252 OH/IR stars observed in SiO and H 2 O masers. The two-    Habing (1988), which is also associated with the extent of mass-loss. The blue dashed line represents two post-AGB star groups of LI (left of IRAS: left of the evolutionary sequence) and RI (right of IRAS: right of the evolutionary sequence), and the LI sources have higher masses and higher outflow velocities than the RI sources (Sevenster 2002a(Sevenster , 2002b V439 Sct (No. 116). These results are consistent with those in the case of Mira variables and semi-regular variables by Kim et al. (2010Kim et al. ( , 2013. The detection or non-detection of the H 2 O maser is clearly divided, even within the same kinds of OH/IR stars, according to the border line between Regions IIIa and IIIb in Figure 6. These trends in the H 2 O maser detectability from OH/IR stars were also indicated and discussed by Gómez et al. (1990). The different detectability of H 2 O and SiO masers from OH/IR stars between Regions IIIa and IIIb may be associated with a different pumping mechanism and excitation conditions between H 2 O and SiO masers, according to the evolutionary sequence of OH/IR stars from Region IIIa to Region IIIb. Gómez et al. (1990)    Note. "Type" indicates the line profile shapes of H 2 O maser spectra. "D", "OW", and "M" indicate double, one-way (blueshifted or redshifted emission with respect to stellar velocity), and multiple-peaks of the H 2 O maser spectra. OW (B) and OW (R) indicate a one-way spectrum with blueshifted emission with respect to stellar velocity and a one-way spectrum with redshifted emission with respect to stellar velocity, respectively.

IRAS Two-color Diagram of Observed OH/IR Stars in SiO and H 2 O Masers
of the optical central source from Mira variables to OH/IR stars (van der Veen & Habing 1988). Gómez et al. (1990) suggested that the cause of the H 2 O maser detectability drop is the H 2 O maser region becoming too cool for excitation due to the increase in optical depth in the envelope, rather than being due to the thermalization of the H 2 O maser. Therefore, we can also adopt a similar explanation as Gómez et al. (1990) in that OH/IR stars from Region IIIb become redder and thicker than in Region IIIa according to the development of the dust layer and steep increase of the mass-loss rate, consequently inducing the increase in optical depth. However, H 2 O-only detected sources appear in the post-AGB RI area of Region VIII together with water fountain sources, as shown in Figure 3 of Yoon et al. (2014). We need to investigate the appearance and disappearance of H 2 O and SiO masers in OH/IR stars associated with their late evolution. We investigate the distribution of the SiO v=1-only and v=2-only maser detected sources including both SiO v=1 and v=2 maser detected sources for both H 2 O and SiO maser detected sources and SiO-only maser detected sources in the IRAS two-color diagram in Figures 7(a) and (b), respectively. In Figure 7(a), most of both of the H 2 O and SiO maser detected sources are present in Region IIIa. The SiO v=1-only detected sources are present only in Regions II, IIIa, and VII, while the SiO-only v=2 maser detected sources are present in Regions VIb and VIII, although the number of detected sources is one each. Both SiO v=1 and v=2 masers were detected in sources of IRAS18034-2441 (No. 65) and V439 Sct (No. 116) in Region IIIb, and their peak temperature ratios of PT(SiO v=2)/PT(SiO v=1) and intensity ratios of IT(SiO v=2)/ IT(SiO v=1) had a value equal to or exceeding 1. In addition, V1018 Sco (No. 41) and OH20.7+0.2 (No. 88) in Regions IV and VIII had higher ratios. Figure 7(b) shows the distribution for the SiO v=1-only, v=2-only, and both v=1 and v=2 maser detected sources for the SiO-only maser detected sources. The SiO v=1-only maser detected sources are present only in Regions IIIa and VII, except for OH49.8-0.8 (No. 180), while the SiO v=2-only maser detected sources are present mostly in Regions IIIb and IV. Moreover, many sources have a value equal to or exceeding 1 in relation to the intensity ratio of those sources, compared to the sources in Region IIIa. There are clear differences in SiO v=1 and v=2 maser distributions between both H 2 O and SiO maser detected sources and SiO-only maser detected sources, as shown in Figures 7(a) and (b). These trends also appeared in the results of the SiO and H 2 O maser survey toward 83 known SiO maser sources (58 Mira variables, 11 OH/IR stars, 7 semi-regular variables etc., Cho & Kim (2012)), and 164 post-AGB stars (Yoon et al. 2014). The SiO v=2-only maser detected sources are more common in a more evolutionary stage, i.e., in Regions IIIb and IV. The development of a hot dust shell at a later stage of AGB evolution and the thermalization of the SiO v=1 maser were suggested by Cho & Kim (2012) and Ramstedt et al. (2012). These trends appeared more clearly in our observed OH/IR stars, which contained a large number of samples compared to previous samples. We also need to investigate the possibility of infrared line overlap, suggested by Olofsson et al. (1985), even in the SiO v=2, J=1-0 masers. This is because the SiO v=2 (J=1-0)-only maser and the higher intensity of the v=2 maser compared to the v=1 maser appear much more from SiO-only maser detected sources rather than from both H 2 O and SiO maser detected sources, as shown in Tables 10 and 11. Such a line overlap may be possible through the inversion of the rotational levels and population transfer as the infrared radiation of H 2 O and SiO (v=1, J=0  v=2, J=2) overlap one another (Olofsson et al. 1985). A radiative or collisional pump leads to an inversion of the rotational levels and a moderate population  transfer in radiating levels of SiO v=2 and J=1, and results in the generation of an emission curve corresponding to SiO v=2, J=1-0, v=2, J=0  v=1, J=1, v=1, J=1-0 (Olofsson et al. 1985). In particular, the significant mass-loss in OH/IR stars leads to the intensive overlapping of an infrared line and makes the decay rate of the v=2, J=1-0 transition easier. The SiO v=2, J=1-0 maser is highly sensitive to the infrared flux, and therefore undergoes greater effects (Nyman et al. 1986). This may make the IT(SiO v=2)/IT(SiO v=1) ration of OH/IR stars larger than 1. However, intensity ratios of SiO v=2 to SiO v=1 masers for SiO-only maser detected sources are larger than those for both H 2 O and SiO maser detected sources, as shown in Tables 10 and 11. In addition, the large number of SiO-only maser detected sources, including the SiO v=2-only maser detected sources, is distributed at a later stage in the IRAS two-color diagram compared to both H 2 O and SiO maser detected sources, as shown in Figures 7(a) and (b). Therefore, the larger value of intensity ratios of SiO v=2 maser with respect to SiO v=1 for SiO-only maser detected sources compared to those for both H 2 O and SiO maser detected sources may originate from the differences in their evolutionary stage rather than the line overlap between H 2 O and SiO maser lines. To investigate the line overlap effect in detail, we need to perform simultaneous observations of H 2 O and SiO v=1, 2, J=1-0 and J=2-1 masers, including the SiO v=1, J=3-2 maser, because our observations are limited to simultaneous observations of H 2 O and SiO v=1, 2, J=1-0 masers.
We also examined the distribution of sources with one-way peaks or double-peaks of H 2 O maser emission with respect to the stellar velocity on the IRAS two-color diagram in Figure 6. This is because these one-way peak or double-peaks of the H 2 O maser can be associated with an asymmetric or bipolar outflow, which often appears at the late-stage of AGB evolution and in protoplanetary nebulae (Lewis 1989;Zijlstra et al. 2001;Engels 2002). The sources marked with identification numbers are distributed in Regions IIIa, IIIb, IV, VIb, and VIII, as  Kim et al. (2014). b Kim et al. (2010Kim et al. ( , 2013, Cho & Kim (2012). shown in Figure 6. Among them, all sources in a relatively latestage of AGB evolution of Regions IIIb, IV, VIb, andVIII (No. 41. V1018 Sco, No. 65. IRAS18034-2441, No. 88. OH20.7 +0.2, No. 103. OH26.3+0.1, No. 116. V439 Sct, No. 144. IRAS18572+0618, andNo. 129. V1365 Aql) show one-way peak and double-peaks of H 2 O maser emission, except for AI CMi (No. 16) in Region V. These trends are also found in the results of Kim et al. (2010Kim et al. ( , 2013 and Yoon et al. (2014). In addition, these sources are both H 2 O and SiO maser detected sources. Among them, the sources OH20.7+0.2 (No. 88) and V1365 Aql (No. 129) in the post-AGB area LI show a one-way peak of H 2 O maser, which is much blueshifted (−17.5 km s −1 in OH20.7+0.2) and redshifted (17.8 km s −1 in V1365 Aql) with respect to the stellar velocities. The FWZP ratios between H 2 O and SiO masers show a large value 4.22 for OH20.7+0.2 (No. 88) and 4.06 for V1365 Aql (No. 129) compared to other sources ( Table 12). The asymmetric structure and bipolar outflow from these sources need to be confirmed by VLBI observations as the candidates of precursors of planetary nebulae.