Astrohemistry: Reent Suesses and Current Challenges
Proeedings IAU Symposium No. 231, 2005
D.C. Lis, G.A. Blake & E. Herbst, eds.
2005 International Astronomial Union
DOI: 00.0000/X000000000000000X
What Constitutes Spetrosopi Proof for
the Detetion of Large "Hot Core"
Moleules?
Luy M. Ziurys and Aldo J. Apponi
Departments of Chemistry and Astronomy, University of Arizona, 933 N. Cherry Avenue,
Tuson, AZ 85721 USA
The presene of large organi speies in interstellar gas has important impliations
for the origin of life and pre-bioti hemistry. Aurate identiations of suh moleules, however, is problemati in moleular louds. There are many reasons for suh diÆulties. One is
that the spetral density is very high in objets where the hemistry is suÆiently omplex
to produe suh speies - at least 10 lines per 100 MHz in Sgr B2(N), for example, at 3 mm,
at a sensitivity of 10 mK peak-to-peak. Hene, the possibility of hane oinidenes is large.
Another reason is the presene of many large organi moleules at relatively high temperatures;
these large asymmetri tops have vast numbers of favorable transitions under these onditions,
inluding those originating from low-lying vibrational states. Confusion and blending of transitions of one large moleule with those of another add to the risks of an inaurate identiation.
A ase in point is that of glyolaldehyde, CH2 OHCHO. In order to onrm the identiation
of this moleule in Sgr B2(N), we searhed for its most favorable transitions in the 2 and 3 mm
windows, spanning the energy range of 10-100 K - a total of 43 individual transitions. Of all
these lines, only seven were not heavily blended or ontaminated by other moleules, i.e. so-alled
"lean" features. Emission, however, was visibly deteted at 34 of the other transitions, but one
transition was learly absent. The "missing" line orresponds to a weak transition originating in
the Ka = 3 ladder, and its absene is onsistent with the other deteted features. Based on the
lean features only, glyolaldehyde has a VLSR = 61.7 1.5 km/s and V1=2 = 7.8 1.8 km/s
with intensities onsistently in the range TR 20-70 mK. Given these data, the identiation of
glyolaldehyde is13 99.9%2 seure. A rotational diagram from this data set yields a olumn density
of Ntot 6 10 m for CH2 OHCHO - roughly a fator of 27 less than that of H2 CO. These
data illustrate the problems and subtleties in identifying large organi moleules in spae.
astrobiology - astrohemistry - ISM: abundanes - ISM: moleules - moleular
proesses - radio lines: ISM
Abstrat.
Keywords.
1. Introdution
The large moleules found in dense interstellar louds are typially simple organi
speies ontaining ommon funtional groups. Ketones, aldehydes, aids, esters, and
amides, for example, are well represented in interstellar moleules, as demonstrated by
the presene of ompounds suh as dimethyl ether (CH3 OCH3 ), formi aid (HCOOH),
methylformate (HCOOCH3 ), aetaldehyde (CH3 CHO), and formamide (NH2 CHO). In
fat, interstellar spae is the one other ommon plae that organi moleules are routinely
found other than in living organisms on Earth.
The presene of organi moleules in these two vastly dierent environments suggests
that there may be a onnetion. In fat, it is possible that interstellar hemistry atually
supplied the initial ompounds that led to terrestrial life. It is now thought that the
119
120
Ziurys & Apponi
evolution of life on Earth did not always result in the best biohemial solutions (e.g.
Benner et al. 2004). There are many other hemial possibilities for living systems. For
example, it is not lear why ribose and deoxyribose are used in DNA and RNA. Why not
glyol, a tetrose or even a hexose? Work by Albert Eshenmoser (1999) has suggested
that other sugars are viable substitutes for ribose in the DNA struture. Why are only
twenty standard amino aids used in living proteins? Studies have shown that there are
no biohemial reasons to exlude other unnatural amino aids in proteins (e.g. Hohsaka
& Masahiko 2002). However, life may have hosen one set of hemial possibilities over
others not beause they were the best solution to the problem, but beause of syntheti
ontingeny, namely, that what evolved is a result of what was available as starting
materials. Could interstellar hemistry have provided these ontingenies via moleule
formation in dense louds? Suh ompounds may have been transported to planet surfaes via omets and meteorites, and hene beame the substanes that led to living
systems.
To examine suh a hypothesis, it is learly neessary to establish what hemial ompounds are present in dense moleular louds and in what quantities. Only then an
the piees be put together to examine the possible ontribution of interstellar moleules
to early biohemistry. However, aurate identiation of suh large organi speies is
not trivial. In this paper, we disuss suh diÆulties and formulate riteria for seure
detetions of large moleules in dense interstellar louds.
2. The Millimeter-Wave Spetrum of SgrB2(N)
The spetrum of Sgr B2(N) shown in Figure 1 is a good illustration of a onfusion
limited spetrum. This 6 hour integration on Sgr B2(N) using the ARO SMT 10 meter
telesope has a theoretial 3- noise level of about 10 mK. No signiant baseline regions
are found in this spetrum at this level, indiating that the entire bandpass is made up
of partially overlapping emission features. Only the strongest of these features are easily
identiable, i.e. vinyl yanide, arbon monosulde, ethyl yanide, and arbonyl sulde.
The intensity of these lines typially are greater than 500 mK. Lines below that level
start to beome ontaminated with other features and show up as blends of two or more
transitions. These data illustrate the diÆulty in unambiguously assigning the lines near
the onfusion limit.
The lines that make up the onfusion-limited spetra of moleular louds in general
an be attributed to relatively large asymmetri top organi moleules. These types of
speies possess intrinsially omplex spetra owing to their three unique moments of
inertia and their system of omplex energy levels that ontribute to large rotational partition funtions. The result is a dense population of rotational transitions that peak in
the millimeter region of the spetrum.
Figure 2 and 3 illustrate the three levels of omplexity that an be expeted while
observing astronomial soures ontaining large organi moleules. Figure 2 is a model
of the 2mm spetrum of Sgr B2(N) only onsidering the 11 moleules listed in the legend. At rst glane, the spetrum is dominated by strong emission lines of the most
abundant speies present (in this ase methanol). These lines are very strong (several
Kelvin) and are sparsely distributed throughout the spetrum. It is noteworthy that the
density of lines this strong is small (only a few per 5000 km/s); hene, any ontamina-
U + C2H 5OH
C 2H3CN + HCOOCH3
C2H3CN
C2H 5CN
?
3
13
0.2
121
U
0.4
C2H5OH + (CH2OH)2
CH3CHO + U
HCOOCH 3 ?
C2H3CN
HCOOCH 3 ?
U U (CH3)2O
CH3CHO
C2H5OH
C H CN?
C 2H3CN2 5
U
C2H5OH
NH2CHO
CH3NH2
C2H5CN
HNCO
(CH3)2O
0.6
OCS
*
TR (K)
0.8
CS
C 2H3CN
1.0
C 2H3CN
C2H5CN
HCOOCH 3
1.2
D2CO? + HCOOCH3 ?
C2H5OH
CH3CHO
U
HCOOCH
HCOOH
C2H3CN
CH3CHO + U
CH3CHO + C2H 5CN
CH3CHO + C2H 3CN
U? C2H 3CN + C2H5CN
Spetrosopi Proof for Hot Core Moleules
0.0
231200
231400
231600
Frequency (MHz)
231800
232000
Single sideband spetrum of Sgr B2(N) taken with the ARO SMT 10 meter telesope
near 231.6 GHz. With an integration time of 6 hours, the spetrum is near the onfusion limit
with no baseline regions at the theoretial 3 noise level ( 10 mK).
Figure 1.
glycol_i
ethylene_oxide_i
vinylalcohol_i
glycolaldehyde_i
methanol_i
vinylcyanide_i
methylformate_A_i
methylformate_E_i
ethylcyanide_i
ethanol_i
5
3
*
TR (K)
4
2
1
0
130000
140000
150000
160000
Frequency (MHz)
170000
180000
Simulated 2mm spetrum of Sgr B2(N) using the olumn density and rotational
temperatures given in the literature for 11 interstellar moleules of varying abundane. The
ombination of only these few moleules quite eetively shows that soure onfusion starts to
beome very signiant in Sgr B2(N) at about the 500 mK level. See the on-line presentation
for the olor version of this gure.
Figure 2.
tion of weak features on these lines will be insigniant. These lines are easily identiable.
The next level of omplexity found in Sgr B2(N) is at about the 300 to 500 mK level.
Here a pattern of lines of inreasing density is found, resulting from moleules that are of
lesser abundane, as well as weaker features from moleules that are of high abundane.
The emission features of methylformate, ethyl yanide and vinyl yanide were seleted
for this illustration. These lines are about the same size as many of the weak methanol
lines, whih adds a level of omplexity. The transitions are more densely spaed beause
these moleules are generally heavier and more asymmetri. The line density at this level
of sensitivity is about an order of magnitude higher.
The last level of omplexity is at about the 40 mK level, where there is nearly a
ontinuum of losely spae lines (see Figure 3). The weak features of the moleules of high
and intermediate abundane plus the strongest features of yet lesser abundant speies
all ontribute at this level. The moleules seleted to illustrate this ase are ethanol,
ethylene glyol, ethylene oxide, vinyl alohol and glyolaldehyde. Even with these few
122
Ziurys & Apponi
0.10
glycol_i
ethylene_oxide_i
vinylalcohol_i
glycolaldehyde_i
methanol_i
vinylcyanide_i
methylformate_A_i
methylformate_E_i
ethylcyanide_i
ethanol_i
0.06
*
TR (K)
0.08
0.04
0.02
0.00
130000
Figure 3.
140000
150000
160000
Frequency (MHz)
170000
180000
Same as gure 2, but the intensity sale is magnied to illustrate the inreased
omplexity that ours at about the 40 mK level in Sgr B2(N).
additional speies, the problem in assigning spetra beomes very apparent. The line
density inreases yet again by about an order of magnitude to about 5 lines per 100
km/s in Sgr B2(N). Contamination at this level is inevitable and lean transitions will
be few in number.
3. Approahing This Complex Problem
The statistial probability of nding a partiular spetral feature of a moleule at a
partiular frequeny is a funtion of: (i) the physis assoiated with the transition under
investigation (Ntot , Trot , Sij , Eu , et.) and (ii) the number of abundant ontaminating
speies with favorable transitions that signiantly overlap the same spetral region. Ideally, we would like the rst of these two fators to dominate, and we have all beome
austom to assuming that this is true. However, as the sensitivity of an observation is
inreased, the seond fator beomes signiant. Hene, only through the observation of a
suÆient number of lean lines (lines free from signiant ontamination) an a moleule
be unambiguously identied near the onfusion limit.
To date, spetral surveys at millimeter wavelengths have not been sensitive enough
to identify new moleules or even some weak known moleules. For instane, Nummelin
et al. 1998 surveyed the 1.3 mm atmospheri window down to about 100 mK peak to
peak and Friedel et al. (2005) sparsely overed the 3mm window at modest sensitivity
ompared to that required for identifying new moleules. Neither of these surveys of
Sgr B2(N) was able to detet glyolaldehyde or ethylene glyol nor were they able to
help onrm or dismiss other reently laimed moleular identiations. To identify new
asymmetri organi moleules with low abundane, it is lear that higher sensitivity is
needed. Moreover, owing to high ontamination levels, a vast number of transitions need
to be overed to ensure that lean lines are observed over the most probable energy ranges.
The approah that we have adopted to help solve this problem is to ondut deep,
onfusion-limited, spetral line surveys that over a large frequeny range (65 - 270 GHz).
This large dataset will provide the means to loate lean transitions of new moleules
over a large energy range. The overall method an be summarized as follows: (i) obtain
an aurate set of rest frequenies, preferably by diret laboratory measurement, (ii) lo-
Spetrosopi Proof for Hot Core Moleules
123
ate lean transitions that math in line shape and width, (iii) analyze the intensity of
those lines to ensure that they represent a physially onneted system, (iv) loate a sufient number of lines that sample an appropriate energy range, (v) identify and model
intensities of all know speies, partiularly weak lines that may overlap to make strong
lines, (vi) attempt to separate out ontaminating features to strengthen the ase, and
(vii) aount for beam-dilution where appropriate (Snyder et al. 2005; Halfen et al. 2005).
4. Case Studies of Glyolaldhyde and Dihydroxyaetone
Two ase studies illustrate the power of this tehnique for identiation of new interstellar moleules. These studies involve the moleules glyolaldehyde and dihydroxyaetone.
Glyolaldehyde was rst reported by Hollis et al. (2000) to be present in SgrB2(N) using
the Kitt Peak 12 meter telesope at millimeter wavelengths. These observations were
followed by measurements at the NRAO GBT telesope in the K-band region (Hollis et
al. 2004). Dihydroxyaetone was rst reported by Widius Weaver & Blake (2005) to be
present in Sgr B2(N) using the CSO 10 meter telesope.
4.1. Conrmation of Glyolaldehyde
In Hollis et al. (2000), several lines were reported in the 3mm window; however, most
of these are apparently ontaminated with other lines beause they exhibit unusually
broad line widths of 20 to 30 km/s. Reently, Halfen et al. (2005) have searhed for an
additional 43 transitions of this moleule. In this work, we disovered that only about
20% of the transitions were free from ontamination, but we were able to observed 7 lean
features that agree in line width, line enter (VLSR ), and their intensities are onsistent
for a gas with a rotational temperature of 34 K.
Figure 4 is an example of some of the 40 spetra that we have measured for glyolaldehyde. These spetra represent the omplete dataset. The degree of ontamination
ranges from total obsuration (102.5 GHz) to partial blending (103.4 GHz). The spetrum at 103.7 GHz shows an example of a lean feature, whih exhibits the anonial
line width and line enter of glyolaldehyde in Sgr B2(N). Still, some ambiguities remain.
For example, in the spetrum at 95.1 GHz, vibrationally exited lines of vinyl yanide
are seen throughout the spetrum (Nummelin & Bergman 1999; Halfen et al. 2005). The
relative intensities of these lines losely math the observed data. However, there does
seem to be some exess emission near the enter frequeny, whih an be attributed to
glyolaldehyde although it is diÆult to quantify how muh. Contamination suh as this
is seen throughout the dataset and highlights the importane of a omplete aounting
of all the emission from known onstituents.
To help prove that the lines observed for glyolaldehyde are all physially onneted, a
rotational diagram was used (see Figure 5). As shown in the gure, both the lean lines
and the partially blended lines all t quite losely to the same physial model.
4.2. A Rigorous Attempt to Conrm Interstellar Dihydroxyaetone
Widius Weaver & Blake (2005) reported the detetion of 9 transitions of dihydroxyaetone (DHA) in Sgr B2(N) at 1.3 mm that losely mathed rest frequenies measured
by the same group (Widius et al. 2004). From their observations, they derived a olumn density of 5 1015 m 2 and a rotational temperature of 220 K. Figure 6 shows
124
Ziurys & Apponi
Figure 4.
Single sideband spetra of Sgr B2(N) taken with the ARO KP 12 meter telesope
(Halfen et al. 2005).
Rotational Diagram for Glycolaldehyde
12
Trot = 35 K
13
Ntot = 5.9 x 10
"Clean" transitions
-2
cm
3
2
log (3kW/8S QSP )
11
10
"Clean" and partially
blended transitions
9
8
0
10
20
30
40
50
60
70
80
Eu / k (K)
Figure 5.
Rotational energy diagram for lean and partially blended lines of glyolaldehyde in
Sgr B2(N).
a single sideband spetrum of DHA taken at the ARO SMT 10 meter telesope that
reprodues Figure 1 shown in Widius Weaver & Blake (2005). The 220 K rotational
temperature does not appear to reprodue the relative intensities of the observed lines as
well as had been reported. The gure also shows that if the temperature is lowered, the
Spetrosopi Proof for Hot Core Moleules
125
Sgr B2(N-LMH)
Simulated @ 220 K
Simulated @ 50 K
0.4
TR (K)
0.2
222820
222840
Frequency (MHz)
605,56 - 594,55
-0.4
1511,4 - 1410,5
1511,5 - 1410,4
-0.2
614,58 - 603,57
613,58 - 604,57
0.0
222860
222880
Single sideband spetrum taken at the ARO SMT 10 meter telesope overing the
same spetral features as shown in Figure 1 of Widius Weaver & Blake (2005). Overlaid on
the spetrum are simulated spetra at 220 K and 50 K showing that these data are not well
reprodued at either temperature even though they do show a lose math in frequeny.
Figure 6.
0.5
0.3
*
TMB (K)
0.4
0.2
0.1
0.0
120000
150000
180000
Frequency (MHz)
210000
240000
Simulated spetrum of DHA in Sgr B2(N) using the physial parameters listed in
(Widuus Weaver & Blake 2005). For a olor version of this gure see the on-line presentation.
Figure 7.
predited relative intensities hange dramatially. Also, the reported olumn density is
about a fator of 6 higher than that of the ommon interstellar moleule H2 CO towards
the same region (Apponi et al. 2005), learly raising some questions about the validity of
their detetion. The intense lines observed at 1.3 mm (400 mK) and the high rotational
temperature proposed indiate that DHA should possess a very intense spetrum from
100 GHz through 300 GHz. Figure 7 shows the simulated spetrum of DHA from 100
GHz to 245 GHz. The blue trae (see on-line presentation for olor version) indiates
the positions of the transitions, while the red trae is the modeled spetrum assuming 10
km/s line width and 10 mK rms noise. It is easy to see that there should be no problem
deteting this moleule at lower frequenies.
In addition, a loser examination of the rotational diagram of the original 9 lines
reorded at 1.3 mm further shows that their intensities are not reprodued well (see
Figure 8). On the left is the rotational diagram reprodued from the results of Widius
Weaver & Blake (2005). At rst glane, the data all seem reasonable, exept that not all
of the data points fall on the tted line. The right panel of the gure shows the optial
depth for eah transition that would be required to bring the respetive point onto the
tted line. These data show that the optial depths vary from -1 to 1.5 and suggest that
these lines are not all physially onneted.
126
Ziurys & Apponi
Rotational diagram analysis of the 9 lines taken from Widius Weaver & Blake (2005)
(left) and the respetive opaities that would be required to orret eah data point bak onto
the tted line (right).
Figure 8.
Spetra overing 4 favorable regions that should have signiant DHA emission. The
spetra were take with the ARO 12 meter telesope in single sideband mode (Apponi et al.
2005). The dashed lines indiate the predited intensity of DHA emission based on Trot = 220
K and Trot = 50 K as labeled.
Figure 9.
The Ziurys group, in ollaboration with Hollis and oworkers, also onduted searhes
for DHA in the same soure in the K-band at the NRAO GBT and at 2 and 3 mm at
the ARO 12 meter telesope at nearly the same time as the CSO measurements. Of 40
separate frequenies overing 61 favorable rotational transitions of DHA, only 2 of them
show reasonable intensity orrelation with the original work reported by Widius Weaver
& Blake (2005). Fourteen of the regions were ontaminated with other moleules and 20
of them showed no signiant emission. Figure 9 shows a few of these spetra. With
these new results, we have onstrained the temperature of DHA to below 50 K and have
redued the upper limit of the olumn density to 5 1013 m 2 . It is doubtful from these
new data that DHA is abundant enough in Sgr B2(N) for a denitive detetion.
Spetrosopi Proof for Hot Core Moleules
127
5. Conlusion
To obtain reliable spetrosopi evidene for the presene of large moleules, a areful and onsistent dataset needs to be assembled. Although suh data olletion is time
onsuming, it is learly neessary for a reliable moleular identiation. Previous identiations of large speies based on a limited number of transitions need to be losely
re-examined. An aurate piture of interstellar organi hemistry will never be obtained
otherwise.
Aknowledgements
This material is based upon the work supported by the National Aeronautis and Spae
Administration through the NASA Astrobiology Institute under Cooperative Agreement
No. CAN-02-055-02 issued through the oÆe of Spae Siene.
Referenes
Apponi, A. J., Halfen, D. T., Ziurys, L. M., Hollis, J. M., Lovas, F. J, & Remijan, A., in
preparation.
Benner, S. A., Riardo, A., & Carriga, M. A., 2004, Curr. Opin. Chem. Biol., 8, 672
Eshenmoser, A., 1999, Siene, 284, 2118
Friedel, D. N., Snyder, L. E., Turner, B. E., & Remijan, A., 2004, ApJ, 600, 234
Halfen, D. T., Apponi, A. J., Woolf, N., Polt, R., & Ziurys, L. M. 2005, ApJ, aepted
Hohsaka, T. & Masahiko, S. M., 2002, Curr. Opin. Chem. Biol., 6, 809
Hollis, J. M., Lovas, F. J., & Jewell, P. R. 2000, ApJ, 540, L107
Hollis, J. M., Jewell, P. R., Lovas, F. J., & Remijan, A. 2004, ApJ, 613, L45
Nummelin, A. & Bergman, P., 1999, A&A, 341, L59
Nummelin, A. et al., 2000, ApJS, 128, 213
Snyder, L. E. et al. 2005, ApJ, 619, 914
Widius, S. L., Braakman, R., Kent, D. R., & Blake, G. A., 2004, J. Mol. Spetros., 224, 101
Widius Weaver, S. L., & Blake, G. A., 2005, ApJ, 624, 33