an analysis of the line shape for h2o maser emission peaks in star-forming regions

Astronomy Letters, Vol. 28, No. 2, 2002, pp. 89–99. Translated from Pis’ma v Astronomicheskiı Zhurnal, Vol. 28, No. 2, 2002, pp. 106–117.Original Russian Text Copyright c© 2002 by Lekht, Silant’ev, Mendoza-Torres, Tolmachev.

An Analysis of the Line Shape for H2O Maser Emission Peaksin Star-Forming Regions

E. E. Lekht1, 2, *, N. A. Silant’ev1, 3, J. E. Mendoza-Torres1, and A. M. Tolmachev4

1Instituto Nacional de Astrofısica, Optica y Electronica,Luis Enrique Erro No. 1, Apdo Postal 51 y 216, 72840 Tonantzintla, Puebla, Mexico2Sternberg Astronomical Institute, Universitetskii pr. 13, Moscow, 119899 Russia

3Pulkovo Astronomical Observatory, Russian Academy of Sciences,Pulkovskoe sh. 65, St. Petersburg, 196140 Russia

4Pushchino Radio Astronomy Observatory, Astrospace Center, Lebedev Physical Institute,Russian Academy of Sciences, Pushchino, Moscow oblast, 142292 Russia

Received September 17, 2001

Abstract—We analyze the line shape for emission peaks of H2O maser sources associated with star-forming regions by using the spectra obtained with the RT-22 radio telescope at the Pushchino RadioAstronomy Observatory. For five sources, we found the line profile of emission peaks to be asymmetric.In all cases, the left (high-frequency) line wing is higher than the right wing. Our analysis of the lineshape yielded additional information on the structure and evolution of the maser sources under study. InG43.8–0.1, the emission feature was found to split up into two components. To explain the evolution of the16.8 km s−1 line in NGC 2071, we propose a model in which the line-of-sight velocity gradient changesunder the effect of a (non-shock) wave. The observed short-duration flares of individual emission featuresin W75N can emerge due to a chance projection of the numerous clumps of matter involved in Keplerianmotion onto each other. c© 2002 MAIK “Nauka/Interperiodica”.

Key words: radio sources, star formation

INTRODUCTION

The spectra of H2Omasers have a complex struc-ture and are highly variable in time. Such spec-tra are attributable to the presence of both regularand turbulent motions with various scales and life-times in the source. A long-term monitoring revealslong-lived emission features (spots) and yields thecharacteristic parameters of the relatively small-scale(<1 AU) turbulence (Lekht et al. 1999). In addition,studies of such features provide information on thecomparatively regular physical processes both in thespot itself and in the possible amplification corridoralong the propagation path of the maser emission(Matveenko et al. 2000).

At present, there are several studies aimed atsearching for flux and line-FWHM variability ofsingle features and for a correlation between thevariations of these parameters. The line shape wasfirst analyzed for an intense flare of the feature at+8 km s−1 of the H2O maser source in Orion in1979 (Matveenko 1981; Strel’nitskii 1982). Theexistence of asymmetry was pointed out: the left,

*E-mail: [email protected];[email protected]

1063-7737/02/2802-0089$22.00 c©

high-frequency wing was higher. The asymmetryin the profile was explained by the possible super-position of two components (Matveenko et al. 1980;Matveenko, 1981).

Apart from the natural quantum broadening, theline FWHM is determined by the thermal scatter ofvelocities of the emitting molecules. In addition, theline can be broadened by small-scale chaotic tur-bulent motions. In a medium with an inverse levelpopulation, the line significantly narrows by a factorof ∼(τ + 1)1/2 (for an unsaturated maser). Here, τis the optical depth of the amplification path at theline center. In this case, estimates of the mean ther-mal velocity and the velocity of small-scale turbulentmotions depend on τ . Goldreich and Kwan (1974)showed that under certain conditions, the line canalso narrow with the same dependence on τ for asaturated maser. The line FWHM and the flux willthen be related by

∆V ∝ (lnF0)−1/2 (1)

for an unsaturated maser and

∆V ∝ F−1/20 (2)

2002 MAIK “Nauka/Interperiodica”

90 LEKHT et al.

600

200

0

12 14 16

800

400

0

8 12 16

800

0

–4 0 4

Flux

den

sity

, Jy

Radial velocity, km/s

February 8, 1995

March 26, 1998

March 10, 1999

(‡)

(b)

(c)

S

A

0.67

0.73

0.58

0.52

0.47

W75N

S

S

S

Fig. 1. Three spectra of W75N during flares of individual spectral features. The number denotes the line FWHM, in km s−1.

for a saturated maser (Mattila et al. 1985). Here, F0

is the flux at the line center. The formation of the lineprofile for various masing modes was considered byStrel’nitskii (1974).

The passage and amplification of the backgroundemission or the spontaneous emission (at τ � 1) ina medium at rest and in a rigidly rotating medium(e.g., a rotating homogeneous maser spot) gives asymmetric line at the output of an unsaturated maser.In our case, a Gaussian fit satisfactorily describes theshape of the observed line. For a more complex spotstructure, the line can be asymmetric (Matveenkoet al. 1980, 2000).

Here, we investigate the shapes of the emissionlines in some H2O maser sources. By the line shape,we mean the following: the line FWHM, whether itcoincides with or differs from a Gaussian, and the lineasymmetry (the difference between the right and leftline wings).

DATA PRESENTATION

We analyzed the spectra of theH2Omaser sourcesNGC 2071, S269, G43.8−0.1, ON1, W75N, andS128, which are associated with star-forming re-gions. All spectra were obtained with the RT-22

ASTRONOMY LETTERS Vol. 28 No. 2 2002

AN ANALYSIS OF THE LINE SHAPE 91

radio telescope at the Pushchino Radio AstronomyObservatory. The spectral resolution at λ = 1.35 cmwas 0.101 km s−1. The radio telescope, equipment,and observing techniques were described in detail, forexample, by Sorochenko et al. (1985).

We selected individual emission components dur-ing flares and long-lived components. The emissionpeaks during flares are most frequently single. Theirblending with other features can be insignificant andcan take place mainly in the wings. In addition, itis highly probable that the main characteristics ofthe single line (flux, radial velocity, line FWHM, andshape) will vary. There may also be a functionalrelationship between the flux and line-FWHM vari-ations. The H2O spectra of the long-lived featureswere taken for the periods when these features weresingle or weakly blended. We selected lines that wereno broader than 0.7 km s−1 and lines for which therewas a correlation between the flux and FWHM vari-ations. Of course, this did not completely guaranteethat the lines were definitely single.

Our selection presents sources of three differenttypes: a Keplerian disk (W75N and S269), an ex-panding envelope (G43.8−0.1, ON1, and S128), anda jet (NGC 2071). The maser S128 has two groupsof spots located at the boundary between interacting(probably colliding) clouds. The maser W75N is alsoa double source. One of the sources is associatedwith an ultracompact H II region (VLA 2), whilethe other is associated with a jet (VLA 1) (Torrelleset al. 1997). The RT-22 radio telescope receives theemission from both sources. NGC 2071 has a similarstructure (Torrelles et al. 1998).

Figure 1 shows the H2O spectra of the maserW75N when strong emission flares took place. Thespectra of this source are rather complex and containa large number of emission features, which overlapin radial velocity. Nevertheless, the flaring featuresare fairly strong and may be considered as singlefeatures. The letters S and A denote the symmetricand asymmetric lines, respectively.

Figure 2 shows the H2O spectra of some sourcesat the epochs of their study here. The arrow inthe spectrum of G43.8−0.1 indicates the possiblevelocity of the central star taken from Lekht andSorochenko (1999).

Some emission peaks were very narrow. Forexample, during the 1982 flare in G43.8−0.1, theline FWHM at the flare maximum was a mere0.35 km s−1. In such cases, the bandwidth of thefilter-bank analyzer (0.1 km s−1) was not consider-ably narrower than the line FWHM. For this reason,the observed lines proved to be slightly distorted; i.e.,the fluxes in the line peak were underestimated. Suchnarrow lines were corrected.


400

200

0

10 20

4000

2000

036 40 44

400

0

–78 –74 –70

V

★


S269August 8, 2000

May 8, 1982February 2, 1989

G43.8-0.1

S128December 16, 1999

Flux

Den

sity

, Jy

Fig. 2.H2O spectra.

Since the maser emission is highly variable intime, to compare the line profiles at different epochs,they were normalized in amplitude and centered in ve-locity. The centering stems from the fact that a radial-velocity drift of the emission features was observedin some cases. In addition, the random errors andthe errors in the radial velocity were eliminated. Thisprocedure was carried out for all line profiles.

For the source G43.8−0.1, which is identified withan expanding envelope, we chose two strong flares.One of them occurred in 1982 at VLSR = 38.6 km s−1.The flare lasted for about 8 months. The mean lineFWHM was 0.4 km s−1. Figure 3 shows five lineprofiles at epochs when the line intensities were highenough to study their shapes. The superposition of all

92 LEKHT et al.

1.0

0.5

0

1.0

0.8

0.6

0.4

0.2

0

–0.8 –0.4 0 0.4 0.8Velocity, km/s

(g)

V

LRS

= 38.6 km/s

–0.8 –0.4 0 0.4 0.8–0.8 –0.4 0 0.4 0.8

Nor

mal

izer

flu

x

(‡)

(c)

(e)

(b)

(d)

(f)

G43.8–0.1

October 22, 1981

February 4, 1982

May 8, 1982

December 23,

March 18, 1982

October, 1981–May, 1982

1.0

0.5

0

1.0

0.5

0

1981

Fig. 3. The 1981 flare at a radial velocity of 38.6 km s−1 in G43.8−0.1. The line profiles for five consecutive epochs with fittedGaussians (a–e), the superposition of five lines (f), and the average line with a fitted Gaussian (g) are shown. The Gaussiansare indicated by the dotted lines.

five line profiles is presented in Fig. 3f. The averageline and the fitted Gaussian are shown in Fig. 3g.

During a different flare (1988–1990), the line at a

radial velocity of 37.6 km s−1 was broader (Fig. 4a),of the order of 0.7 km s−1. The line splitting into twocomponents (Fig. 4b) was observed. The spectrum



0

37 38 39

200

400

Flux

den

sity

, Jy

(b)

April, 1989–April, 1990


1.0

0.5

0

–1 0 1

G43.8–0.1

V

LSR

= 37.6 km/sSeptember, 1988–March, 1989

(‡)

Nor

mal

izer

flu

x

Velocity, km/s

Fig. 4. The 1988–1990 flare in G43.8−0.1 at VLSR = 37.6 km s−1: (a) the average profile of 12 lines for the period September1988–April 1989 and a fitted Gaussian (dotted line), (b) the line superposition when the line became asymmetric and thendouble (May 1989–April 1990).

of G43.8−0.1 exhibits a long-lived component at aradial velocity of 42.2 km s−1 (see Fig. 2), whosevelocity was fairly stable. To study the radial-velocitydrift of the 37.6-km s−1 feature, its velocity was mea-sured relative to the 42.2-km s−1 feature. The resultsare presented in Fig. 5a. The flux and line-FWHMvariations in the ln F − ∆V −2 coordinates are shownin Fig. 5b.

In the H2O spectrum of NGC 2071, we chosethe component at VLSR ≈ 17 km s−1 and monitoredits variability from October through December 1988.Figure 6a shows five spectra obtained on October


6; November 11, 12, 24; and December 21, 1988.The spectra are numbered 1 through 5, respectively.During this entire period, the line was asymmetric.We observed a small regular drift of the emission peakin radial velocity within 0.2 km s−1 and a rise influx from 1740 to 2100 Jy with a simultaneous linenarrowing from 0.82 to 0.61 km s−1.

To compare the line shapes, all five profiles werereduced to the same minimum FWHM, which wasobserved on December 24, 1988. The lines were thencentered to a zero velocity (Fig. 6b). The normalizedlines closely coincide. Figure 6c shows the averageline (1), which was fitted by two Gaussians (2). Their

94 LEKHT et al.

0.9

8.07.2 8.8ln

F

, Jy

1.0

1.1

1.2

∆

V

–2

, (km

/s)

–2

(b)

19901989Year

–4.0

Rel

ativ

e ve

loci

ty, k

m/s

(‡)

–4.2

–4.4

–4.6

–3.8

G43.8–0.1

V

LSR

= 37.6 km/s

Fig. 5. (a) Radial-velocity variations of the emission feature at 37.6 km s−1 (of two components, A and B) in G43.8−0.1relative to the 42.2-km s−1 feature and (b) the relationship between the flux and line FWHM for the same feature. The pointswere joined by straight lines to show their temporal sequence. The arrow indicates the direction of evolution. The large circlemarks the flare maximum. The dashed line represents an extrapolation of the drift curve for componentB.

sum is indicated in Fig. 6c by line 3, which virtuallycoincides with the average line.

The results of our study of the asymmetry in theobserved lines are presented in Fig. 7. For the leftplot, distances from the line center and deviation ofthe observed line from the average line are along thehorizontal and vertical axes, respectively. The averageline was obtained as the ordinary mean of the left andright wings in the observed line, F (U) = (FL(U) +FR(U))/2, where F (U) is the emission flux. In thiscase, the deviation from the mean (along the horizon-

tal axis) can be written as U − |UL| for the left wingof the line and as U − |UR| for its right wing. In fivecases, the line was asymmetric: S269, G43.8−0.1(flare 1982), ON1 (the 15-km s−1 emission featurefor two time intervals), and NGC 2071.

The difference in the patterns of the line FWHMvariability for the symmetric and asymmetric lines isshown in the right-hand part of Fig. 7. Here, ∆U isthe range within which the line FWHM varies as afunction of the distance from its center. The letters L



0

16Radial velocity, km/s

17 18

1000

2000NGC 2071

12

3

4 5

(‡)

1.0

0.5

0

0

123

(b)

(c)

Nor

mal

izer

flu

xFl

ux d

ensi

ty, J

y

1–1

1.0

0.5

0

Velocity, km/s

Fig. 6. (a) H2O line profiles of the 16.9-km s−1 emission feature in NGC 2071 obtained during October–December 1998;(b) the superposition of lines normalized in the flux and line FWHM; (c) the average line and fitted Gaussians.

and R denote the curves for the left and right linewings, respectively.

DISCUSSION

The Source W75N

According to Torrelles et al. (1997), there are twomain groups of maser spots in this region. One ofthem, W75N (Ba), that coincides with the ultracom-pact continuum source VLA 1 is elongated roughlyin the direction of the bipolar molecular outflow. Thegroup of spots W75N (Bb) that coincides with thethe ultracompact continuum source VLA 2, forms arotating protoplanetary disk. The separation between


the groups of maser spots does not exceed 1′′. Insingle-dish observations, we receive the emission si-multaneously from both H2Omasers.

The H2O spectrum of the maser W75N (Ba) hasa smaller velocity dispersion than W75N (Bb) (Tor-relles et al. 1997). Therefore, the strong flare of 1999at −2.6 km s−1 (Fig. 1c) may be assumed to haveoccurred in the maser W75N (Bb) identified with aKeplerian disk. The flare lasted no longer than 3–4 months. The line had a symmetric shape. Despitethe significant flux variations, the line FWHM did notvary (0.67 km s−1). Given that the line was broadand that there was no correlation between the flux andline FWHM variations, the maser may be assumed to

96 LEKHT et al.

–0.08–0.4 0 0.4

U, km/s

0

0.08–0.08

0

0.08

–0.08

0

0.08

–0.08

0

0.08S26920.1 km/s

U

–

–

|

U

|

, km

/s

G43.8–0.1

38.6 km/s

37.6 km/s

–71.9 km/sS128

ON115 km/s

a

b

a

—1994–1996

b

—1997

0.02

0.20 0.4 0.6

|

U

|

, km/s

0.06

0.10

0.02

0.10

0.06

0.02

0.06

0.10

∆

U, k

m/s

S26920.1 km/s

L

R

G43.8–0.1

L

R

R

L

38.6 km/s

–71.9 km/sS128

Fig. 7. Line asymmetry of individual H2Omaser emission components.

be saturated in this case. The flare was apparentlylocal, because there were no significant variations inthe H2O spectrum as a whole during this period.

The maser W75N (Bb) has a complex spatialstructure and contains a large number of maser spots(clumps of matter), which individually do not producenoticeable maser emission because of the insufficientoptical depth. Since all these structures are involved

in Keplerianmotion, we cannot rule out a chance line-of-sight projection of two maser spots with similarvelocities onto one another. In this case, a strongshort flare can be observed, as in 1999.

Such short flares in W75N were observed in1995–1997 (see Figs. 1a and 1b). The line shapeof these emission peaks at the flare maximum wassymmetric, and the line FWHMwas 0.5–0.7 km s−1.



Outside the peaks, there was no significant excess ofemission of these short-lived components over theothers. Therefore, for these periods, we cannot obtainthe actual estimate of their line shape.

The Source G43.8–0.1

During the 1982 flare, the line was narrowest:the minimum FWHM was 0.35, and the mean was0.4 km s−1. The line had an asymmetry of ∼11%(Figs. 3 and 7). The flare duration was estimated tobe 6–7 months. Despite the significant flux and lineFWHM variations, the asymmetry was preserved.There was a correlation between the flux and lineFWHM variations, which is typical of an unsaturatedmaser (Mattila et al. 1985). No radial-velocity driftof the line was observed, which argues for the absenceof any appreciable radial-velocity gradient in the linegeneration region.

If the Doppler line width is taken to be 2 km s−1,then the observed line at maximum will be narrowerthan the Doppler one by almost a factor of 6, whichcorresponds to an amplification increment at the linecenter of τ ∼35. This is enough to account for thehigh brightness temperature in maser spots (TB ≈1015 K) that requires an amplification of at least∼1010 (Strel’nitskii 1982). The observed line wasso narrow that even the assumption of simultaneousemission from the two closest hyperfine-structurecomponents of the 616–523 transition cannot explainits considerable (11%) asymmetry. Indeed, at aseparation between the two closest components of33 kHz (0.44 km s−1), the observed line cannotbe fitted by the superposition of two Gaussians.Therefore, we may assume that in this source (in amaser spot), only one hyperfine-structure componentis amplified.

In contrast to the left wing, the right line wing(Fig. 3g) is satisfactorily fitted by a Gaussian. At adistance larger than 1.5 FWHM from the line center,the left line wing is greatly raised with respect tothe Gaussian. The deviations from the Gaussian aremuch larger than those that can be for an unsaturatedmaser (Strel’nitskii 1982). The observed line asym-metry may be related to the spot geometry, i.e., thespot may have a more complex spatial structure thansimply a homogeneous clump of matter.

The 1988–1990 flare differed significantly from the1982 flare. The line was broader and initially symmet-ric (Fig. 4a). As the flare developed until its maximum(September 1988–March 1999), the velocity of theemission feature decreased. The change in velocitywas 0.05 km s−1 (Fig. 5a). Such an accuracy wasachieved because the velocity was measured relativeto the fairly stable component at 42.2 km s−1. The


line FWHMvaried within a narrow range of radial ve-locities (0.67–0.80 km s−1), and no correlation withthe flux variations was observed. The observed lineFWHM variations were most likely erratic (Fig. 5b).The absence of a correlation between the line FWHMand flux variations, its relatively large FWHM, andthe symmetry are a possible argument for a saturatedmaser.

After the flare maximum, the drift direction re-versed. The drift in this direction continued for 1 year.In addition, the line broadened immediately after theflare maximum and half a year after the maximum,the second component B appeared in the spectrum(Fig. 4b), which receded from the main component A.The velocity of the latter (A) increased in this time by0.1 km s−1; the velocity of component B increasedby 0.7 km s−1, if it is assumed to have appeared im-mediately after the maximum. Indeed, extrapolationindicates that component B emerged at the flare max-imum (Fig. 5a), as confirmed by Fig. 5b. In addition,an analysis of the line shape showed that the line wasslightly asymmetric at the emission maximum (theMarch 31, 1989 spectrum): the right line wing washigher than the left one. The asymmetry was ≈4%.Thus, our study of the line shape allowed us to refinethe occurrence time of the second component.

The radial velocity of the central star in G43.8−0.1is estimated to be≈40 km s−1 (Lekht and Sorochen-ko 1999). Thus, we can explain the observed vari-ations in the line radial velocity and shape as fol-lows. The motion of the maser condensation atVLSR = 37.6 km s−1 was initially accelerated relativeto the central star. At the time corresponding tothe emission maximum, the deceleration in a densemedium (or collisions) resulted in a breakup of thecondensation into two parts. That is why there wasno further rise in flux and its decline began. Duringthe interaction with a dense medium, the motion ofboth components slowed down, which accounts forthe reversal of the drift direction. The deceleration ofcomponent B was larger than that of component A.

The variety of the flares that occurred in G43.8−0.1 is not limited to the cases considered. As anexample, we can note the 1997 flare (Lekht 2000),which occurred at a radial velocity of 38.2 km s−1.The flare lasted ∼1 year. The line was narrow andsymmetric; its minimum and maximum FWHMswere 0.43 and 0.6 km s−1, respectively. We observedthe flux-correlated line FWHN variations that arewell described by relation (1). In addition, there was aradial-velocity drift of the feature during the flux rise.

The Source NGC 2071According to Torrelles et al. (1998), there are two

groups of maser spots in this region, which are asso-ciated with compact H II regions and the IR sources

98 LEKHT et al.

IRS 1 and IRS 3. One of them located near IRS 1is extended along a jet. The other group of spotsassociated with IRS 3 forms a rotating protoplanetarydisk whose plane is perpendicular to the jet axis.

For the component at VLSR ≈ 17 km s−1, we ob-served a small, regular velocity drift of the emissionpeak from 16.7 to 16.9 km s−1. The high rate ofchange in the radial velocity of the maser in a pro-toplanetary disk seems unlikely. In addition, thecomponent at ≈17 km s−1 must be on the peripheryof the Keplerian disk in the red amplification corridor(Torrelles et al. 1998). As was noted previously,the rate of change in the radial velocity here is at aminimum. This maser spot is most likely located nearIRS 3 and is, thus, associated with the jet.

We were able to closely fit the normalized profileby two Gaussians of the same width (0.46 km s−1)and with an amplitude ratio of 3 : 1. The equalityof the line FWHMs is apparently not accidental andcan be evidence that the emission originates froma single spot with a nonuniform density. The fluxdifference between the two components by a factor of3 (unsaturated maser) requires a difference betweenτ no more than 5% (τ ≈ 20–25 at the line center),which essentially does not result in a line FWHMdifference.

The coincidence of the line shapes for the normal-ized profiles also rejects the model of the approachof two maser spots in radial velocity, i.e., when theymove toward each other along the line of sight. Itis unlikely that for such an approach of the featuresduring a short period (2.5 months), both featuressimultaneously narrowed. In addition, to preservethe shape of the normalized lines, only small fluxvariations are admitted.

The line asymmetry may be assumed to be causedby a radial-velocity gradient at a nonuniform densityof the matter. Previously, we assumed this emissionfeature to be associated with a molecular jet. Theradial-velocity gradient may have changed, whichcaused a change in the profile of the amplificationfactor of the medium. A wave (not a shock) prob-ably passed through the jet, which changed theradial-velocity gradient of the medium whose H2Omolecules were involved in the maser line formation.Initially, the gradient decreased, resulting in a linenarrowing and in a displacement of the emission peakin the H2O spectrum. Subsequently, the gradientagain increased, resulting in a line broadening and ina reverse radial-velocity displacement of the line peak.

Other Sources

Our monitoring of the source ON1 allowed usto trace the evolution of the emission at VLSR =

15.5 km s−1 over more than 15 years. We chose twotime intervals (1994–1996 and the entire 1997), whenthis emission feature was not blended with others.Within each interval, the line FWHM changed onlyslightly and the line itself was asymmetric (its profilesare not given here). Between these intervals, the linebroadened from 0.58 to 0.69 km s−1, while the lineasymmetry decreased from 12 to 7%. These facts in-dicate that the maser spot must have a more complexstructure than a homogeneous clump of matter. Inaddition, there may be small-scale turbulent motionsboth in the maser spot and in its surroundings.

The line shape of the 20-km s−1 component inthe source S269 was studied previously (Lekht et al.2001). It was shown that the line had an asymmetricshape and its radial velocity varied according to asinusoidal lawwith a period of 26 years. As themodel,we proposed a rotating turbulent vortex located at thecenter of the red amplification corridor of a Kepleriandisk. Here, we numerically estimated the line asym-metry. The mean asymmetry for the entire monitoringof S269 is ∼11%.

In S128, the H2O emission in the radial-velocityrange from −73 to −70 km s−1 was continuouslyobserved during our entire monitoring. The emissionfrom several features alternately appeared in this partof the spectrum. Since 1998, the emission has origi-nated from a single feature with a mean radial velocityof −71.9 km s−1. The line shape was symmetric.There was a correlation between the flux and lineFWHM variations (Lekht et al. 2002), characteristicof an unsaturated maser. The absence of line asym-metry indicates that the model of a homogeneousmaser spot is most plausible.

Asymmetric lines are common in the H2O maserspectra. The causes of the asymmetry can be densitynonuniformities and a radial-velocity gradient (in-cluding a gradient nonlinear along the line of sight)in the masing region (maser spot).

In addition, we explored the possibility of absorp-tion of the emission in the medium near the maserspot. The molecules in this medium and in the spothave similar radial velocities; the difference is thatthey have no inverse level populations. The resultsof such studies will be presented in a forthcomingpublication.

The line asymmetry, together with the flux andradial velocity, is an important parameter for choosinga maser spot model. Collectively, these characteris-tics of maser emission make it possible to study thestructure of maser spots and their interaction with themedium where they are localized.



CONCLUSIONS

Below, we present our main results.(1) The shapes of the emission components of

H2O maser sources in star-forming regions can beboth symmetric and asymmetric. The asymmetry,i.e., the deviation from a symmetric line of the sameFWHM, ranges from 5 to 15%.

(2) Despite the variations in the flux and FWHMof the single asymmetric line, its asymmetry waspreserved (S269, G43.8−0.1, and ON1).

(3) The strong short flares of H2O emission inW75N (Bb), which is identified with a protoplanetarydisk, could result from a chance projection of numer-ous clumps of matter onto each other.

(4) As the 1988–1990 flare developed in G43.8−0.1, the emission feature at 37.6 km s−1 broke up intotwo components. We found a small drift of both com-ponents with different radial velocities. The breakupof the condensation responsible for the emission at37.6 km s−1 into two parts could be caused by itsinteraction with the dense envelope matter.

(5) In NGC 2071, we observed a narrowing ofthe asymmetric line (VLSR ≈ 16.8 km s−1) with thesimultaneous displacement of the emission peak inradial velocity. This phenomenon can be explained bythe passage of some wave (not a shock) that changedthe radial-velocity gradient of the medium where themaser emission originated.

ACKNOWLEDGMENTS

This study was carried out with the RT-22 radiotelescope (registration number 01-10). It was sup-ported by the Ministry of Information of Russia andthe Russian Foundation for Basic Research (projectno. 99-02-16293). We are grateful to the staff ofthe Pushchino Radio Astronomy observatory for helpwith the observations.


REFERENCES

1. P. Goldreich and J. Kwan, Astrophys. J. 190, 27(1974).

2. E. E. Lekht, Astron. Astrophys., Suppl. Ser. 141, 185(2000).

3. E. E. Lekht and R. L. Sorochenko, Astron. Zh. 758,76 (1999) [Astron. Rep. 43, 663 (1999)].

4. E. E. Lekht, S. B. Likhachev, R. L. Sorochenko,and V. S. Strel’nitskii, Astron. Zh. 70, 731 (1993)[Astron. Rep. 37, 367 (1993)].

5. E. E. Lekht, J. E. Mendoza-Torres, and N. A. Silan-t’ev, Astron. Zh. 76, 248 (1999).

6. E. E. Lekht, N. A. Silant’ev, J. E. Mendoza-Torres,et al., Astron. Astrophys. 377, 999 (2001).

7. E. E. Lekht, J. E. Mendoza-Torres, and I. I. Berulis,Astron. Zh. 79, 63 (2002) [Astron. Rep. 46, 57(2002)].

8. K. Mattila, N. Holsti, M. Toriseva, et al., Astron.Astrophys. 145, 192 (1985).

9. L. I. Matveenko, Pis’ma Astron. Zh. 7, 100 (1981)[Sov. Astron. Lett. 7, 54 (1981)].

10. L. I. Matveenko, L. R. Kogan, and V. I. Kostenko,Pis’ma Astron. Zh. 6, 505 (1980) [Sov. Astron. Lett.6, 279 (1980)].

11. L. I. Matveenko, P. J. Diamond, and D. A. Graham,Astron. Zh. 77, 669 (2000) [Astron. Rep. 44, 592(2000)].

12. R. L. Sorochenko, M. M. Berulis, V. A. Gusev, et al.,Tr. Fiz. Inst. Akad. Nauk SSSR 159, 50 (1985).

13. V. S. Strel’nitskiı, Usp. Fiz. Nauk 113, 463 (1974)[Sov. Phys. Usp. 17, 507 (1974)].

14. V. S. Strel’nitskiı, Pis’ma Astron. Zh. 8, 165 (1982)[Sov. Astron. Lett. 8, 86 (1982)].

15. J. M. Torrelles, J. F. Gomes, L. F. Rodrıguez, et al.,Astrophys. J. 489, 744 (1997).

16. J. M. Torrelles, J. F. Gomes, L. F. Rodrıgues, et al.,Astrophys. J. 505, 756 (1998).

Translated by G. Rudnitskiı

an analysis of the line shape for h2o maser emission peaks in star-forming regions

Documents