Non-Periodic Phenomena in Variable Stars
IAU Colloquium, Budapest, 1968
THE LIGHT VARIATIONS OF THE NUCLEI IN HERBIG-HARO
OBJECT NO. 2, 1946-1968*
G. H. HERBIG
Lick Observatory
University of California, Santa Cruz, California, USA
The first three of the peculiar semi-stellar emission nebulae now
known as Herbig-Haro objects were detected by Herbig (1948, 1951) and by
Haro (1950, 1952). The initial discoveries were made in the region south
of the Orion Nebula, but a total of about 40 such Objects have now been
found in Orion, Taurus, Perseus, and elsewhere. All known examples occur
in heavily-obscured regions that are also rich in T Tauri stars. A number
of other very small nebulous spots have been observed by Haro (1953, Table 2),
Méndez (1967) and others, particularly near the Orion Nebula, but these are
spectroscopically distinct from the H-H Objects in having either a strong
continuum especially in the near infrared, or a rather conventional emission
spectrum. The H-H Objects have quite characteristic emission spectra:
the H emission lines are strong, and [O I] and [S II] are unusually intense.
The [N II] lines are also strong, and in those Objects not too heavily reddened,
[O II] lambda lambda3726-29 as well. These properties can be explained
(Böhm 1956, Osterbrock 1958) if H and O are only partially ionized, and if
T_epsilon ~ ~ 7500 deg and n_epsilon ~ ~ 3 X 10^3 cm^-3. Only the brightest
of the H-H Objects, Herbig No. 1 (= Haro 11a, near NGC 1999) has been observed
in any detail. It shows a number of weaker emission lines that are not
ordinarily found in appreciable strength in gaseous nebulae: H and K of Ca II,
the infrared [Ca II] lines, Mg I lambda4571, and lines of [Fe II] and [Fe III].
It is significant that many T Tauri stars show, in integrated light, nebular
lines very similar to those of the H-H Objects.
With only a few exceptions, most of the known H-H Objects have fairly simple
structure: a small, bright, diffuse nucleus that may be elongated, often with
a short fainter tail or extension attached. Herbig No. 2 (= Haro 10a) is one
of the exceptions: at the present time it consists of a number (at least 8) of
bright nuclei together with considerable fainter structure enclosed in an
elliptical area about 25" x 40". This Object was first photographed at adequate
scale in 1946-47, and when the plates were repeated in 1954-55 it was found
that two new nuclei had appeared within this complex Object in the interim
(Herbig 1957). Thereafter the region was photographed annually with the same
telescope and emulsions (Crossley reflector, no filter, and Kodak 103aO or
IIaO) through 1959. A number of plates were obtained with the 120-inch
reflector in 1959-63+, but the Crossley series was not resumed until 1968.
Some further but rather minor changes were noted in the structure of
this Object after 1954 (Herbig 1966a). However, it was not until after the
discovery by Magnan (1967) that by 1966 one of the nuclei had brightened still
more, that all the Lick material was systematically re-examined and an effort
made to fill gaps in the early record by inspection of other plate series.
* Contributions from the Lick Observatory, No. 282.
+ A 120-inch photograph of this area appears in Sky and Telescope, 20, 338, 1960.
Determination of the magnitudes of the individual nuclei in Object No. 2 is
beset by problems of overlap and the background, the non-stellar nature of the
condensations even at Crossley scale (39"/mm), the effects of differing seeing
from plate to plate, and complications in the comparison of emission-line
sources with stars having continuous spectra. A sequence of comparison stars
was set up in the vicinity of Object No. 2 by direct comparison and extension,
through objective-grating images on a plate taken with the 20-inch astrograph,
with Special Selected Area 20 (Van Rhijn 1952). No high accuracy is claimed for
these m_pg's (listed in Table I) but they should be sufficient for this
investigation, considering the other sources of uncertainty. Many of the
fainter stars in this heavily-obscured region around NGC 1999 are variable,
and several comparison stars had to be discarded for that reason, but the stars
in Table 1 seemed to be constant over this time interval.
Table 1
Approximate m_pg's and Relative Coordinates* of Comparison Stars
m_pg Delta x cos delta Delta delta
16.0 +11'.3 +1'.3
16.5 +0.7 +2.7
17.0 -5.8 +0.2
17.6 +6.2 -1.8
18.0 -1.3 -3.4
18.5 +4.8 +2.7
19.5: -3.4 +1.5
* With respect to Nucleus A in Object No. 2.
The principal nuclei in Object No. 2 are identified by letters in Fig. 1.
The m_pg's of the 7 brightest of these are listed in Table 2, estimated from
all adequate plates available to me. It would have been desirable to use m_pg's
obtained only from the Crossley material in order to minimize the effect of the
systematic errors already mentioned, but for the critical years 1947-1954 all
available photographs were utilized. The 20-inch astrograph plates especially
have to be employed with caution because of their small scale (55"/mm) and
color-curve and low ultraviolet-transmission effects. Plates of the red region
were used only to pronounce upon the presence or absence of a condensation.
The magnitudes given for nuclei C, D and particularly E are of lower accuracy
than for the others, because of their clearly non-stellar appearance.
None of the 120-inch plates were used for magnitude estimates, because at that
scale (13"/mm) all the nuclei in Object No. 2 appear as complex, non-stellar
structures. The best of the series of Crossley and 120-inch blue-ultraviolet
plates of Object No. 2 are shown in Fig. 2.
Fig. 1. Object Herbig No. 2, photographed with the 120-inch reflector in
red light on 1959 Dec. 6. The letter designations of the individual
nuclei are shown. The scale can be judged by the fact that the distance
between Nuclei A and B is 8.2".
Fig. 2. Object No. 2, photographed with the Crossley and 120-inch
reflectors in blue-ultraviolet between 1946 and 1968.
Fig. 3. Photographic light curves of nuclei G, H, A and B in Object No. 2.
Vertical bars indicate uncertain points, and carets show plate limits.
The light-curves of nuclei A, B, G and H are plotted in Fig. 3. The rapid
rise of nuclei G and H from fainter than about m_pg = 20.0 (only a rough
estimate because of the lack of faint standards in SSA 20) to about m_pg = 17.5
is apparent. The data in Table 2 suggest that the brightening of G was certainly
complete by late 1953, and if the evidence of the single 20-inch plate of 1951.9
can be accepted, may have been completed some time between 1950.2 and
1951.9. The rise of H was completed between 1952.8 and 1953.9. No
further change took place in H until after 1959.9. Magnan's (1967)
published photograph shows H in 1966.0 to have been the brightest
nucleus in the Object (in red light); by 1968.0, H was fully 1 mag.
brighter than it had been in 1954-1960.
Table 2
Photographic Magnitudes of Nuclei, Object No. 2 between 1946 and 1968
Date (UT) Telescope* Nucleus
(Number of
plates) A B C D E G H
1946 Jan. 24 Cr (2) 17.3 16.5 17.6: - 18.8+- <20.0 <19.5
1946 Feb. 23 Cr 17.5: 17.0 17.3 17.7: - <18.5 <18.5
1947 Jan. 20 Cr (3) 16.5 16.7 17.6: - 20.0+- <20.0 <19.5
1947 Jan. 19 20 (2) - - - - - <18.5: <18.5:
1949 Aug. 28 48, red - - - - - Invis Invis
1950 Mar. 21 48 - - - - - <18.5 ? <19.5
1951 Nov. 30 20 - - - - - 17 ?^1 <18.5:
1952 Oct. 26 Cr, red - - - - - Invis ?^2 Invis ?
1953 Nov. 17 48 16.5: 17.5+- - - - 17.5 17.5:
1954 Jan. 2 Cr (2) 16.4+- 16.9+- - - - - Present^3
1954 Dec. 20 Cr 16.8: 17.3 18.0: 17.8 20.0+- 17.4 17.6:
1955 Feb. 25 Cr 16.4: 17.0: 17.7 17.8 19.5+- 17.0: 17.3:
1955 Sept 20 Cr 16.6: 17.4 17.9: 18.5: 19.5+- 17.4 17.5
1955 Dec. 13 Cr 16.4 17.4 17.8: - 20.0+- 17.3 17.3
1956 Sept. 3 Cr (2) 17.0 17.2 17.7 18.0: 20.0+- 17.3 17.2
1956 Nov. 13 Cr 16.7 17.3 17.8 17.9 20.0+- 17.2 17.1
1957 Sept.23 Cr 17.2 17.4 17.8 17.8 - 17.1 17.3
1958 Nov. 9 Cr 17.0: 17.4 17.9 17.9 20.0+- 17.1 16.9
1959 Nov. 10 Cr 17.1 17.7 18.0: 18.2: 20.0+- 17.4
1968 Jan. 5 Cr (2) 17.7^4 17.5 17.6: <19.0+- 20.0+- 17.0: 16.2
* Cr = Crossley reflector (Lick);
48 = 48-inch Schmidt (Palomar);
20 = 20-inch astrograph, blue lens (Lick).
Notes: 1 G seems to be present but at the scale and definition of the 20-inch,
this result is marginal.
2 Unwidened slitless spectrogram. The overlapping of many monochromatic
images makes a decision difficult, but probably neither G nor H are
present within a magnitude of their final brightness in H alpha.
3 Object No. 2 occurs near the corner of these plates, and the definition
is very poor. The magnitudes given are only rough estimates.
4 The nucleus A on the 3 plates of 1968 seems to be slightly north of
its former position.
Nucleus A is also variable. It was about 1 mag. fainter on the first
Crossley plates of the area taken in early 1946 than it was a year
later. It remained constant until about 1955. Beginning about 1956, A
fatted again until by 1968.0 it was fainter than at any other time
during the 22 years of observation. Unfortunately no 120-inch plates
have been taken of Object No. 2 since 1963.0. A good, modern large-scale
photograph is required to check the suspicion, based on the 3 Crossley
plates of 1968.0, that the condensation now identified as A lies
slightly northward of the position A had when it was bright. It will be
noted that in those earlier years (Fig. 1), A had a very short tail
extending toward the northeast; possibly this extension now dominates
the total light of A as observed at lower resolution.
None of the other nuclei in Object No. 2 show any convincing evidence of
variability, although there is a possibility that D is now fainter than
formely.
Prior to 1946, the only large-scale, good quality plate of the area of
Object No. 2 known to me was one centered on NGC 1999 that was taken with the
30-inch Helwan reflector in 1915. Unfortunately the exposure was rather short,
and H-H Object No. 1 is just above threshold while No. 2 does not appear
at all.* The earliest, ordinary wide-angle camera photographs that I have
examined were exposed in 1901 (Wolf 1903, Roberts 1927), and show No. 2
as a faint spot. Of course the variety of lenses, focal lengths, and emulsions
used for photographs in the literature make a trustworthy determination of the
total magnitude very difficult. However it can be said firmly that since 1901
there has been an image at the position of Object No. 2 of roughly the same
character and brightness as is observed there at the present time. In view
of the known history of the individual condensations in the Object since 1946,
it would not be surprising if a roughly constant magnitude for the Object as
a whole could be maintained by the net effect of the random brightening and
fading of individual nuclei. One would predict on this basis, however, that
immediately prior to 1946 the total brightness must have been substantially
lower than after 1954, since G and H had not yet appeared and A was faint.
* I am indebted to Dr. A. Samaha, Director of the Helwan Observatory,
for very kindly sending me a print from this negative.
Several explanations of the changes observed in Object No. 2 have, or
could be put forward.
(1) The correlation of T Tauri stars with Herbig--Haro-like nebulae,
observable both spectroscopically as well as directly in the cases of T Tauri
and HL Tauri (Herbig 1968), strongly suggests that there may be some real
organic connection between the two. Possibly the existence of an H-H Object
without an associated star marks the site of some electromagnetic activity
preliminary to star formation, wherein is produced the flux of protons, of
about 100 KeV energy that Magnan and Schatzman (1965) calculate can explain
the observed level of ionization. But the observed fading of nucleus A shows
that an increase in brightness of an individual nucleus does not
necessarily represent a progressive or a permanent change. Thus the
spectacular brightening of G and H in the early 1950's must not be
associated or identified with the "birth" of two "new" stars. At best,
one might expect that at one of these sites an event like that of FU
Orionis might eventually take place (Herbig 1966).
(2) The appearance of G and H so close (about 3") to A and B,
respectively, gave rise to one suggestion that this might be no more
than the resolution, due to rapid orbital motion, of two binary pairs
that had previously been too close to be separated. This explanation can
be dismissed quite convincingly for the following reasons.
(a) Prior to 1954, A and B, interpreted as A + H and B + G, would have
had to be about 0.75 mag. brighter than afterwards. This was not the
case.
(b) The photocenters of A + H and B + G would have had to lie on the
line connecting the present positions of A and B, and of B and G.
Careful measurements were made of the position of the nuclei on Crossley
plates, with respect to a reference frame defined by nearby field stars.
These show convincingly that A and B in 1947 were, within the errors of
measurement (about 0".2), at the same positions as they were in 1954
after G and H had appeared.
(c) The fact that A and H have subsequently been observed to change in
brightness demonstrates that the individual nuclei in Object No. 2 are
quite capable of intrinsic variation.
(3) The variations might be interpreted as due to the fluctuations of a
variable star within each nucleus. A strong objection to this is that,
especially in the red, essentially all the energy of the nuclei is
contained in the emission lines, yet the contribution of a variable star
within - presumably a T Tauri star - would largely be in the form of a
continuous spectrum. Furthermore, the best 120-inch direct plates exposed in
the continuum between strong emission lines (pass bands lambda lambda5200-5800
and lambda lambda6800-8800) show the nuclei to be clearly non-stellar
in these spectral regions. Thus there is no direct or spectroscopic
evidence of a star image within any of these condensations, variable or
not. This point has also been stressed by Haro. A further argument to
this same conclusion is based on the fact that T Tauri is known to be
central in a small complex emission-line nebula that appears itself to
be a H-H Object. There is strong evidence, although unfortunately most
of it based on early visual work, that this nebula varies in brightness
by a substantial amount (see Herbig 1950 for modern observations and
references to the early work), and that this variation is not obviously
correlated with that of the star. These facts support the belief that
the presence of a faint T Tauri-like variable could not explain the
light variations of a H-H nucleus.
(4) It might be argued that the light variations of the nuclei are not
intrinsic, but rather are due to variable extinction by dense dust
concentrations moving across the line of sight in the foreground. This
hypothesis meets with the following difficulty. None of the nuclei in
Object No. 2 are truly stellar; diameters of 1"- 2" are measured on the
120-inch plates. Now, if the motion of dust in the vicinity is typical
of gas motion in H I regions, a velocity of 1 km/sec for an element of
the cloud would be reasonable. But at 500 pc, the distance of Object No. 2,
this corresponds to an angular cross-motion of 1" in 2000 years.
Clearly, an opaque screen having such a slow motion would not be able to
uncover nuclei G and H quickly enough to explain their observed
brightening in an interval of 2-3 years or less.
(5) It has been suggested that the sudden appearance of new nuclei
within Object No. 2 may have been the result of the sudden ionization of
a small volume of a dense, neutral interstellar cloud by a blast of radiation
or particles from some invisible source within. If this were the correct
explanation, then the fading of nucleus A subsequent to 1959 has to be
interpreted as recombination due to the withdrawal of excitation.
At T_epsilon = 7500 deg and n_epsilon = 3 X 10^3 cm^-3, about 5 years are
required to reduce n_epsilon to 0.8 n_epsilon (Alley 1956) and thus to lower
the surface brightness in the Balmer lines by the observed 0.5 mag. This is
in fact approximately the observed time scale of the fading of A. On the other
hand, if the H recombines on such a short time scale, it means that the
hypothetical exciting sources must remain active in most H-H Objects. Otherwise,
they all would quickly fade away. Furthermore, if this decay were going on in
all Objects, one would expect a considerable range in their degrees of
ionization, yet the "new" nuclei in Object No. 2 seemed to be spectroscopically
indistinguishable from all the others, presumably much older.
The fact that H and O are partially ionized, apparently throughout the
whole volume of the Object, seems to rule out radiative ionization from
a single source within, as pointed out by Osterbrock (1958). Ionization
by high-energy particles is a more acceptable explanation for this
effect. Magnan and Schatzman (1965) have calculated the amount and
energy (~~ 100 KeV) of a proton flux that would be required to maintain
the ionization in Object No. 1 by collisional processes. But there is no
explanation of why all H-H Objects are fed by particles of just these
special properties.
(6) A very speculative interpretation is that the bright nuclei in
Object No. 2 are only transient phenomena on the surface of a very
dense, dark cloud, and are thus distantly analogous to surface phenomena
on the sun. This presumably requires that the cloud be a coherent,
gravitationally stable unit. The elliptical area of about 25" X 40"
which contains all the bright structure of Object No. 2 would, in 3
dimensions and at a H density of n~~ 10^4 cm^-3, enclose a mass of about
0.05 m_sun. This mass would not be gravitationally bound because the surface
escape velocity is only about 0.1 km/sec. If the mean density of the cloud
were however raised to n > 10^6 H cm^-3, then v_escape >v_internal and
the cloud could remain intact. Such a mean density seems reasonable
since the value of n = 10^4 cm^-3 was inferred from the brightline spectra
of the nuclei, which here are interpreted as surface phenomena on a body
which could have strong central condensation. This hypothesis is appealing
in that it provides at least intuitively, through analogy with the sun,
an explanation for the light variations of the nuclei in Object No. 2.
It also might explain why three nuclei brightened in the years 1946-1953:
possibly the degree of surface activity on such an object would vary
cyclically, as in the sun.
The foregoing outline has demonstrated that although some possibilities can
be eliminated, there is still no completely straightforward interpretation
of the light variations of the nuclei in Object No. 2. The most acute need in
clarifying some of the questions raised here is a continuing series of
high-resolution direct photographs, taken over a period of 5-10 years with a
very large reflector: Such a series could demonstrate, for example, whether
the position and the fine structure of a nucleus changes as its total magnitude
varies. Possibly it would also help to answer also the question whether the same
nucleus is able to reappear again, or whether each brightens and fades
only once. It is not now clear whether variability is the exception or
the rule among H-H Objects. The variability of the Object surrounding T Tauri
has already been mentioned. Small changes are believed to have been observed
in several other Objects as well, but to date, either the plate material is
too limited or the Crossley scale too small to confirm this suspicion.
It is a pleasure to express my thanks to Dr. G. O. Abell for allowing me
to examine some early 48-inch Schmidt negatives in his collection; to
Mr. E. Harlan for taking the Crossley plates of Object No. 2 in January, 1968;
to Dr. G. Haro for helpful and stimulating correspondence; to Dr. W. J. Luyten
for sending me copies of some 48-inch Schmidt plates in his possession;
to M. C. Magnan for interesting discussions as well as for communication of his
results in advance of publication; and to Dr. S. Vasilevskis for taking several
plates for me in 1959 with the 120-inch reflector.
REFERENCES
Aller, L. H., 1956, Gaseous Nebulae (New York: John Wiley and Sons), p. 66.
Böhm, K. H., 1956, Astrophys. J., 123, 379.
Haro, G., 1950, Astr. J., 55, 72.
Haro, G., 1952, Astrophys. J., 115, 572.
Haro, G., 1953, Astrophys. J., 117, 73.
Herbig, G. H., 1948, Thesis. University of California.
Herbig, G. H., 1950, Astrophys. J., 111, 11.
Herbig, G. H., 1951, Astrophys. J., 113, 697.
Herbig, G. H., 1957, Non-Stable Stars, ed. G. H. Herbig (I. A. U. Symposium No. 3.
London and New York: Cambridge University Press), p. 3.
Herbig, G. H., 1962, in The Universe, by R. Bergamini (New York: Time, Inc.), p. 142.
Herbig, G. H., 1966, Vistas in Astronomy, 8, 109.
Herbig, G. H., 1968, June: paper presented at Liége Astrophysical Symposium.
Magnan, C., 1967, l'Astronomie, 81, 49.
Magnan, C. and Schatzman, E., 1965, C. r. hebd. Seanc. Acad. Sci., Paris, 260, 6289.
Méndez, M. E., 1967, Bol. Tonantzintla y Tacubaya, 4, 104.
Osterbrock, D. E., 1958, Publ. astr. Soc. Pacific, 70, 399.
Roberts, Mrs. I., 1927, Isaac Roberts' Atlas of 52 Regions, Chart 22.
Van Rhijn, P. J., 1952, Durchmusterung of Selected Areas of the Special Plan, 1.
(Groningen: Kapteyn Astronomical Laboratory).
DISCUSSION
Seitter: Could the hypothesis of the object being a protostar with spot-like
activity on its surface be checked from proper motions of the group
of nuclei as a whole? If it is indeed a protostar one would expect it
to be in rotation. The observation of a common proper motion of the
expected order of magnitude could thus strengthen the hypothesis.
Herbig: Yes, in principle. But the cross-motion due to rotation at 1 km/sec
would amount only to about 0.0005"/year.
Rosino: 1) I should like to know whether any radial velocity determination on
H-H Object No. 2 has been made and, in the affirmative case, whether
this radial velocity corresponds to that of the Orion Nebula.
2) Three H-H objects 1, 2, 3 are found in a peculiar region, near the
small nebula NGC 1999, which is abnormal for the spectrum and the
aspect. I am wondering whether there may be any connection between
these objects and the three H-H objects.
Herbig: Observations of the radial velocities of the individual nuclei
in No. 2 have been made at the prime focus of the 120-inch reflector.
These results are beset by the well-known difficulties of velocity work
with a spectrograph having a thick-mirror Schmidt and un-flattened
field. Probably one should say only that there is no evidence for large
internal motions greater than about 50 km/sec.
Anderson: Are there any infrared or radio observations?
Herbig: My own I-red photography extends only to 0.8 micron - these show the same
structure in Objects No. 1 and 2 as in the emission lines in the red
and blue. I understand that No. 1 has also been observed by one of the
Arizona - Tonantzintla observers at 3 micron, but no detectable infrared
radiation was found. Thus these H-H objects are not "infrared stars" in
sense of the recent use of that word.
Feast: 1) Did I understand that there is no measurable relative proper motion
of the various nuclei?
2) Is there any evidence for changes in size of the individual
condensations?
Herbig: to 1)There was no relative motion between 1947 and 1954 greater
than 0.2". This is based on Crossley material; the 120-inch plates have
not been measured.
to 2) The scale of the Crossley plates is certainly too small to answer
this question. The 120-inch material is perhaps adequate but has not
been examined from this point of view.