on the basis of our analysis of the previous single-dish survey of Sgr B2(N) (5). In this way, the risk
of misassigning a line to a new species is reduced.
About 50 and 120 transitions of i-PrCN and n-PrCN,
respectively, are detected toward the northern
hot core (Fig. 2, A and B, and figs. S1 and S2).
On the basis of this model, we selected the least-contaminated transitions and produced contour
maps of their intensity integrated over their line
profile (Fig. 2, C and E, and figs. S3 and S4). From
these maps, we derived a deconvolved angular size
of 1.0″ T 0.3″ (FWHM) for the region where both
species emit. We used the population diagram
method (11) to estimate a rotation temperature of
153 T 12 K (SEM), which characterizes the emission of both molecules (9).
With the size and temperature derived above
and a linewidth measurement of 5 km s−1, we
obtained a good fit to all transitions of i-PrCN
and n-PrCN detected toward the northern hot
core of Sgr B2(N). After correction for the contribution of vibrationally excited states (9), we
derived column densities of 7.2 × 1016 T 1.4 ×
1016 cm−2 and 1.8 × 1017 T 0.4 × 1017 cm−2 (SEM),
respectively, which yielded an abundance ratio
[i-PrCN]/[n-PrCN] of 0.40 T 0.06 (SEM). The latter uncertainty assumes the same source size and
rotation temperature for both isomers. With the
H2 column density derived from the continuum
emission (9) (Fig. 2D), we deduced average abundances relative to H2 of 1.3 × 10–8 T 0.2 × 10–8 for
i-PrCN and 3.2 × 10–8 T 0.5 × 10–8 for n-PrCN. The
latter uncertainties assume the same rotation
and dust temperatures and take neither possible
contamination of the continuum emission by
free-free emission nor uncertainties on the dust
properties into account.
The recently developed chemical kinetics mod-
el MAGICKAL [Model for Astrophysical Gas and
Ice Chemical Kinetics And Layering (12)] was
used to simulate the time-dependent chemistry
of the source. The model begins with a cold col-
lapse phase, during which abundant ice mantles,
composed of simple H-, O-, C-, and N-bearing
molecules, are formed on dust-grain surfaces.
The cold stage is followed by a warm-up stage,
during which the dust and gas temperature
rises from ~8 to 400 K. The majority of complex
organic molecule formation occurs during this
stage, through the addition of simple and com-
plex radicals within and upon the ice mantles.
The model produces time-dependent chemi-
cal abundances (with respect to H2) for various
cyanide molecules (Fig. 3). The model temper-
atures at which each molecule’s peak abundance
is attained (table S2) may be considered repre-
sentative of the excitation temperature at which
most of the emission from each molecule would
occur, assuming local thermodynamic equilib-
rium. The desorption temperature of each PrCN
isomer is ~150 K, with peak gas-phase abundan-
ces achieved at 160 K. This agrees well with the
rotation temperatures that we determined from
observational data for these molecules.
The majority of each of the two PrCN isomers
forms in or on the dust-grain ices at around 55
to 75 K in the model. However, the model also
indicates that, whereas many similar chemical
pathways are open to both i-PrCN and n-PrCN, the
dominant formation route in each case is different.
The greatest contribution to i-PrCN production
comes from the reaction of CN radicals (which
are accreted from the gas) with the CH3CHCH3
radical. The latter derives from the earlier gas-
phase formation of C3, which is hydrogenated and
stored on the grains as C3H2, C3H4, and C3H6. The
addition of atomic hydrogen to propylene (C3H6)
strongly favors the production of CH3CHCH3, whose
radical site lies at the secondary carbon atom (13).
Because the production, either by hydrogen
addition or abstraction processes, of radicals such
as CH2CH2CH3 and CH2CH2CN (whose radical site
is at the primary carbon atom) is strongly dis-
favored versus their equivalent iso forms, we
find that the dominant formation mechanism for
n-PrCN is the addition of C2H5 and CH2CN, a pro-
cess that has no equivalent for the production of
i-PrCN. The radicals form through the abstrac-
tion of hydrogen by OH, from C2H6 and CH3CN,
respectively. i-PrCN production dominates all
reaction mechanisms for which parallel processes
are available to both isomers.
Although the overall peak gas-phase values of
i-PrCN and n-PrCN produced by the models are
similar, they show a slight bias toward i-PrCN
production (2.2:1), rather than the observed bias
toward n-PrCN (0.4:1). This may be caused by the
poorly defined rates for barrier-mediated surface
reactions, such as H + C3H6 → CH3CHCH3 and
OH + CH3CN → CH2CN + H2O, for which only gas-
phase rates have been measured and whose behav-
ior may be somewhat different on an ice surface.
Amid the growing understanding that com-
plex organic molecules could form on the surface
of dust grains, the formation of branched alkyl
molecules in the ISM was suggested theoretically
in the 1980s (14), but no such molecules were de-
tected until now. The detection of a branched alkyl
molecule in Sgr B2, with an abundance similar
to that of its straight-chain isomer, indicates a
further divergence between the chemistry of star-
forming regions like Sgr B2 and quiescent regions.
These less-active regions seem to produce only
linear molecules—the largest one known to date
being HC11N (15). The detection of a branched
alkyl molecule also suggests a further link between
interstellar chemistry and the molecular compo-
sition of meteorites for which branched amino
acids are even found to dominate over their straight-
chain isomers (16). The inherent bias toward the
production of secondary rather than primary
radical sites on precursor radicals suggests that
branched molecules may be prevalent, and in-
deed dominant, in star-forming regions where
chemistry of sufficient complexity is reached. The
detection of the next member of the alkyl cyanide
series, n-butyl cyanide (n-C4H9CN), and its three
branched isomers would allow the testing of this
REFERENCES AND NOTES
1. See, e.g., www.astro.uni-koeln.de/cdms/molecules.
2. J. E. Elsila, J. P. Dworkin, M. P. Bernstein, M. P. Martin,
S. A. Sandford, Astrophys. J. 660, 911–918 (2007).
3. O. Botta, J. L. Bada, Surv. Geophys. 23, 411–467 (2002).
4. A. Belloche et al., Astron. Astrophys. 499, 215–232 (2009).
5. A. Belloche, H. S. P. Müller, K. M. Menten, P. Schilke,
C. Comito, Astron. Astrophys. 559, A47 (2013).
6. A. Belloche et al., Astron. Astrophys. 482, 179–196 (2008).
7. M. J. Reid et al., Astrophys. J. 783, 130 (2014).
8. H. S. P. Müller, A. Coutens, A. Walters, J.-U. Grabow,
S. Schlemmer, J. Mol. Spectrosc. 267, 100–107 (2011).
9. Additional information is available in the supplementary
materials on Science Online.
10. H. S. P. Müller, F. Schlöder, J. Stutzki, G. Winnewisser,
J. Mol. Struct. 742, 215–227 (2005).
11. P. F. Goldsmith, W. D. Langer, Astrophys. J. 517, 209–225
12. R. T. Garrod, Astrophys. J. 765, 60 (2013).
13. H. J. Curran, Int. J. Chem. Kinet. 38, 250–275 (2006).
14. T. J. Millar, P. D. Brown, H. Olofsson, A. Hjalmarson,
Astron. Astrophys. 205, L5–L7 (1988).
15. M. B. Bell et al., Astrophys. J. 483, L61–L64 (1997).
16. J. R. Cronin, S. Pizzarello, Adv. Space Res. 3, 5–18 (1983).
We thank D. Petry and E. Humphreys at the European ALMA
Regional Center (ARC), F. Gueth at the IRAM ARC node, and
D. Muders at the MPIfR for their help with the reduction of the
ALMA data; S. Bardeau, S. Maret, J. Pety, M. Lonjaret, and
P. Hily-Blant for the new developments implemented in the
Weeds software; and M. Koerber for help with preparing Fig. 1.
This work made use of the National Institute of Standards and
Technology Chemical Kinetics Database. This work has been
supported in part by the Deutsche Forschungsgemeinschaft
1586 26 SEPTEMBER 2014 • VOL 345 ISSUE 6204
Fig. 3. Simulated
abundances of alkyl cyanides with respect to
molecular hydrogen, H2.
These abundances represent the warm-up phase of
hot-core evolution. Solid
lines indicate gas-phase
species; dotted lines of the
same color indicate the
same species in the solid
phase. The main phase
change from solid to gas
for each molecule is caused
by thermal desorption from
the grain surfaces,
according to species-specific binding energies.