Cyanoketene: The microwave spectrum and structure of an unstable molecule

Article abstract of DOI:10.1016/j.jms.2003.10.010

The unstable molecule cyanoketene has been prepared by pyrolysis in a flow system, and the microwave spectra of five isotopic species, a- and b-type, have been measured in the frequency range from 8 to 40 GHz. Precise rotational constants and centrifugal distortion parameters up to the sixth order were obtained. The molecule has been shown to be planar, and a reliable structure was derived, Also the N-quadrupole coupling constants and the dipole moment components have been determined. The results could be a basis for interstellar spectroscopy of this molecule.

Full text of DOI:10.1016/j.jms.2003.10.010

Journal of Molecular Spectroscopy 223 (2004) 138–147
www.elsevier.com/locate/jms
Cyanoketene: the microwave spectrum and structure
of an unstable molecule
M. Hahn,1 H.-K. Bodenseh,* and M. Ferner
Universit€at Ulm, Fakult€at f€ur Naturwissenschaften, Abteilung Chemische Physik, D-89069 Ulm, Germany
Received 30 July 2003; in revised form 21 October 2003
In Memoriam Silke Balla
Abstract
The unstable molecule cyanoketene has been prepared by pyrolysis in a flow system, and the microwave spectra of five isotopic
species, a- and b-type, have been measured in the frequency range from 8 to 40 GHz. Precise rotational constants and centrifugal
distortion parameters up to the sixth order were obtained. The molecule has been shown to be planar, and a reliable structure was
derived. Also the N-quadrupole coupling constants and the dipole moment components have been determined. The results could be
a basis for interstellar spectroscopy of this molecule.
Ó 2003 Elsevier Inc. All rights reserved.
Keywords: Cyanoketene; Microwave spectroscopy; Dipole moment; Nuclear quadrupole coupling constants; Molecular structure
1. Introduction
diffraction [9]. Microwave spectra, however, which allow
the determination of much more important quantities of
a molecule have so far been obtained only from 3 of
such ketenes: chloroketene [10], fluoroketene [11], and
bromoketene [12]. Here we report now the microwave
spectra of 5 isotopomers and the structure of another
unstable ketene: cyanoketene. This molecule has found
increasing interest during the recent years: a photo
electron spectrum, published in 1981 by Bock et al. [13],
can be considered as the starting point and in 1985 we
presented for the first time the microwave spectrum of
the parent molecule 14NB12C–(12CH)@12C@16O [14].
Extensive theoretical calculations about the structure,
reactivity, and energetic quantities have been carried out
[15–17]. Production of the molecule under high vacuum
conditions by flash thermolysis of different precursors
allowed the measurement of its heat of formation and an
estimate of the ionization energy by mass spectroscopic
techniques [18], the determination of the IR-bands in an
Ar matrix at 12 K [19], and to find the pathways of its
formation also by trapping the different species in an
Ar matrix at 14 K [16]. The latter paper contradicts
strongly a finding of Efremov et al. [20] who claim that
cyanoketene can be isolated as a liquid, boiling at
)34 °C, and who published an NMR spectrum. In [16]
In 1905 Staudinger [1] synthesized the first ketene:
diphenylketene. Only two years later Wilsmore [2] suc-
ceeded in preparing the unsubstituted compound. Since
then a great number of alkyl- and aryl-substituted ket-
enes have been described [3]. Also derivatives with one
or even both hydrogen atoms replaced by such electro-
negative groups as halides or pseudo-halides have long
been claimed by organic chemists to exist as intermedi-
ates in in situ reactions [4]. But only much later some of
them could be identified as transient species in the gas
phase: dicyanoketene by UV photoelectron spectros-
copy [5] and by Ar matrix IR spectroscopy after flash
pyrolysis of a precursor [6], mono- and di-bromo- and -
chloroketene by HeI photoelectron spectroscopy [7], and
dichloroketene also by IR spectroscopy [8] and electron
* Corresponding author. Present address: Universit€at Ulm, c/o
Abteilung
Germany. Fax: +49-731-50-22839.
E-mail address: hans-karl.bodenseh@chemie.uni-ulm.de (H.-K.
Bodenseh).
Quanteninformationsverarbeitung,
D-89069
Ulm,
1
Present address: COR AG, Karlsruher Str. 3, D-70771
Leinfelden-Echterdingen, Germany.
0022-2852/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved.
doi:10.1016/j.jms.2003.10.010

M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
139
however, the authors point out that the compound does
not survive 80 K in condensed phase. Another important
aspect: because this molecule is only composed of the
atoms C, H, N, and O there is some evidence that it may
be detected in interstellar space by radio astronomy just
like the unsubstituted ketene [21], and laboratory sim-
ulations have already been carried out [22] how it could
be formed in outer space. Looked at from this point of
view the microwave transitions published herein could
be useful.
with a yield of 95%. The resulting cyanoacetic acid was
then processed as above.
2.1.4. N13CCHCO
The same procedure was used as described under
Section 2.1.3 but with 13C-labeled KCN.
2.1.5. NCCHC18O
To obtain this isotopomer, the simplified procedure
of not isolating the cyanoacetyl chloride could not be
used. Instead, the reaction mixture from the chlorina-
tion of the cyanoacetic acid was, again after removal of
ether and POCl3, carefully distilled under a pressure of
20 Pa, yielding the pure chloride at 33 °C. This was
hydrolyzed with H218O according to
2. Experimental procedures
2.1. Chemical preparation of cyanoketene
NBC–CH2–COCl þ H218O ! NBC–CH2–CO18OH
þ HCl
producing a cyanoacetic acid containing one 18O-atom.
Chlorinating this acid as described under Section 2.1.1
resulted in a 1:1-mixture of ‘‘normal’’ and 18O-cyano-
acetyl cloride. The final pyrolysis therefore also produced
a 1:1-mixture of cyanoketene and 18O-cyanoketene.
2.1.1. Parent molecule NCCHCO
We intended to prepare cyanoketene by dehydro-
chlorination of cyanoacetic acid chloride, NC–CH–
COCl, under pyrolytic conditions in the gas phase.
Since, however, this precursor is known to be unstable
[23], we at first did not attempt to isolate this compound.
It was prepared, instead, in the usual way by chlori-
nating the free acid in dry diethyl ether with PCl5 at
room temperature. Then the reaction mixture was
cooled down to )40 °C, and the ether was distilled off.
After removal of the ether the remaining mixture was
further cooled to )60 °C and connected to the flow
system. This consisted of an oven of 10 cm in length
surrounding a quartz tube of 1 cm i.d. and loosely
packed with rock-wool for better heat transfer, spec-
trometer cell, LN2-traps, and vacuum line. The oven still
kept cool, the by-product POCl3 was withdrawn
through the absorption cell, its final disappearance being
monitored by some microwave transitions. Then the
oven was heated to 750 °C; this temperature giving
the strongest signals for almost all isotopomers. Finally
the sample was allowed to warm up until the pressure
was sufficient for optimum spectroscopic conditions; ice
water was convenient. After use the sample had always
to be kept below )20 °C.
2.2. Spectrometer Set-up
The microwave spectrograph was of conventional
Stark-type operating at 30 kHz and equipped with a
self-configurated and -programmed ELTEC computer—
based on a MOTOROLA 68K processor—for time
averaging because the pyrolytic reaction is not really
stationary. 1000 points for spectrum and also for fre-
quency markers could be taken and shown on the screen
with a repetition rate of 9 Hz. For survey spectra 50MHz
ranges were used with real-time frequency markers, the
MARCONI-sweeper swept via its FM-input. The fre-
quency accuracy of 300–500 kHz was sufficient for as-
signments. By means of the frequency markers these
spectra could be concatenated with arbitrary frequency
scaling for viewing larger spectral ranges. For precise
frequency measurements the sweeper was phase-stabi-
lized against the frequency standard and swept via the
reference frequency of 30 MHz by means of a voltage-
frequency converter. Ranges of 1–3 MHz were used, also
with real-time frequency markers and resulting in an
accuracy better than 10 kHz. Finally each line was fitted
to Lorentzian shape to obtain the center frequency. The
frequency standard was monitored against the standard
signals of DCF77-station, Mainflingen, Germany, the
relative deviations kept <10ꢀ8. The cell temperature
was )15 °C and the pressure 4 Pa for the survey spectra
and down to 0.2 Pa for frequency measurements, de-
pending upon the intensity of the transition. Figs. 1 and
2 show examples of a survey spectrum and of a fre-
quency determination. The small structures in Fig. 1 are
2.1.2. NCCDCO
The acidity of the hydrogen atoms in cyanoacetic acid
is sufficiently high to exchange them against deuterium
by shaking the acid with D2O. The degree of exchange
was monitored by means of a quadrupole mass spec-
trometer. After drying, the acid was processed as de-
scribed above. A remarkable isotope effect was observed
for the pyrolysis of just this isotopomer: cyanoketene
formation had its optimum already at 600 °C, and above
660 °C no more signals were found.
2.1.3. 15NCCHCO
The sodium salt of chloroacetic acid was reacted with
15N-labeled KCN according to a Kolbe-synthesis and

140
M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
Table 1
Observed transitions of cyanoketene
#
Transition
JK0 a0Kc0 –JK00a00Kc00
Frequency
(MHz)
Resid.
(kHz)
NCHCCO
1
11;1–00;0
11;0–10;1
32162.344
27034.909
10750.015
37290.208
16120.638
27285.529
15754.483
16499.684
27664.606
21485.914
21003.834
21997.329
21522.654
21505.486
21514.823
26844.235
26251.418
27493.009
26900.768
28824.761
32193.889
31496.783
32311.700
29617.225
36739.477
37533.040
38476.559
37622.067
37717.977
37658.173
37663.241
31663.803
26843.942
25446.206
20275.666
34387.124
13597.647
37870.406
24140.243
39924.394
14151.480
16152.670
14095.979
37656.786
21798.906
21618.421
24995.938
38446.179
25867.213
21944.294
27035.797
14119.014
38423.737
21996.716
32416.505
38449.166
26325.130
13562.343
38344.14
)155
19
37
0
2
3
20;2–10;1
21;2–10;1
4
5
3
0;3–20;2
1;1–20;2
63
44
12
17
29
6
6
2
7
31;3–21;2
31;2–21;1
31;2–30;3
8
Fig. 1. Example of survey spectrum (lines not included in the fit
because later identified).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
4
0;4–30;3
41;4–31;3
1;3–31;2
42;2–32;1
2;3–32;2
)6
12
27
48
45
1
4
4
43;2–33;1
50;5–40;4
51;5–41;4
4
5
1;4–41;3
4;1–44;0
)17
20
)17
52
10
)41
0
5
51;4–50;5
0;6–50;5
6
61;6–51;5
62;4–52;3
61;5–60;6
7
1;7–61;6
70;7–60;6
1;6–61;5
72;6–62;5
2;5–62;4
73;4–63;3
44–643
81;7–80;8
2;7–91;8
90;9–81;8
2;8–101;9
)40
3
7
)2
10
)19
)24
38
)30
38
33
)82
)20
)12
8
Fig. 2. Example of frequency determination (transition 515 414 of
NBC–CH@C@18O).
7
7
not noise but rather weak transitions lying very dense
and making it impossible to find rare isotopomers in
natural abundance.
8
9
101;9–100;10
102;9–111;10
121;11–120;12
122;10–131;13
131;12–130;13
151;14–142;13
152;13–161;16
162;14–171;17
171;16–171;17
213;18–222;21
242;22–242;23
262;24–253;23
264;23–273;24
284;25–293;26
304;26–313;29
312;29–321;32
342;33–333;30
365;32–374;33
373;34–373;35
375;33–384;34
383;35–374;34
385;34–394;35
415;36–424;39
466;41–475;42
474;43–465;42
486;43–495;44
3. Results
24
7
3.1. Rotational spectra and spectroscopic constants
4
9
7
About 60 rotational transitions between 8 and
40 GHz, a- and b-type, were measured and assigned for
4 of the 5 isotopomers with standard deviations of about
23 kHz, being in good agreement with the estimated
uncertainty of the frequency measurements (10–30 kHz).
All calculations were done with WatsonÕs [24] A-reduc-
tion in Ir-representation, the transitions weighted ac-
cording to their accuracy. All five quartic centrifugal
distortion constants, five of the seven sextic, and one
octic could be determined with sufficient accuracy and
yielded the best fits. Only for the 18O-species we had to
restrict the measurements from 18 to 26 GHz and to 13
a-type transitions with a standard deviation of 54 kHz.
Besides the rotational constants only two quartic dis-
tortion constants could be fitted; the rest was fixed to the
corresponding values of the parent molecule. Table 1
35
1
)11
)30
)21
1
3
)8
10
)14
14
7
)16
19
6
1
26582.94
26615.99
37
21

M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
141
Table 1 (continued)
Table 1 (continued)
#
Transition
JK0 a0Kc0 –JK00a00Kc00
Frequency
(MHz)
Resid.
(kHz)
#
Transition
JK0 a0Kc0 –JK00a00Kc00
Frequency
(MHz)
Resid.
(kHz)
62
63
64
65
66
67
68
69
70
503;47–512;50
504;47–495;44
514;48–505;45
513;48–522;51
524;49–515;46
567;50–576;51
615;57–606;54
676;61–667;60
686;62–677;61
21031.55
27684.26
32454.84
26338.68
37100.815
38402.896
37622.995
20600.11
26815.36
47
5
Standard deviation of the fit: 19 kHz
N13CHCCO
19
1
2
11;1–00;0
1;2–10;1
32130.980
37217.903
21306.630
21809.500
20832.180
26620.560
26036.960
27258.360
26654.640
26687.900
31239.590
37221.480
38148.470
37400.540
37307.750
37347.970
37359.855
37375.910
27044.420
27290.920
27663.780
29583.260
30510.290
31593.680
27370.820
31174.550
34267.050
31165.780
37426.180
27719.900
37684.180
27519.040
26620.150
27044.090
30978.720
31092.040
27510.880
26151.060
28095.990
37157.470
31734.020
34454.720
28443.350
30822.470
28766.280
34854.120
30786.120
35010.330
30825.870
29192.570
30302.000
30740.550
29984.960
29426.220
34738.000
)7
20
3
)23
)13
)6
1
2
3
40;4–30;3
41;3–31;2
41;4–31;3
4
3
5
)2
0
25
)33
6
50;5–40;4
7
51;5–41;4
51;4–41;3
52;4–42;3
11
)8
)13
12
18
5
8
Standard deviation of the fit: 24 kHz
NCDCCO
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
52;3–42;2
61;6–51;5
70;7–60;6
71;6–61;5
1
2
11;1–00;0
40;4–30;3
41;3–31;2
27118.620
21309.420
21919.700
20747.450
21368.598
21339.218
21350.690
26614.540
25928.720
26728.680
26670.000
26693.146
22823.510
23433.810
24212.930
25171.470
26321.850
24399.500
24311.110
21661.170
18487.350
24791.440
26822.230
18223.270
22224.830
25164.680
22137.880
26008.530
21877.820
21387.360
24063.980
18602.420
21134.360
24725.830
24880.550
23769.900
22492.440
18161.780
18410.930
23380.000
24125.570
24067.220
21125.840
20882.140
26351.100
23792.060
18464.010
24614.820
24369.050
24749.340
)30
)4
)18
11
)13
)16
7
3
72;5–62;4
4
4
1;4–31;3
2;2–32;1
72;6–62;5
74;4–64;3
5
4
61
42
)18
25
30
)18
)11
16
1
6
42;3–32;2
3;1–33;0
7
5;3–65;2
76;1–66;0
1;0–10;1
21;1–20;2
1;2–30;3
61;5–60;6
1;6–70;7
81;7–80;8
2;7–91;8
7
4
)42
)11
6
8
50;5–40;4
51;5–41;4
52;3–42;2
1
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
)3
)26
54
3
52;4–42;3
54;1–44;0
31;2–30;3
41;3–40;4
7
23
)5
6
7
12
8
51;4–50;5
10
100;10–91;9
8
61;5–60;6
71;6–70;7
62;5–71;6
)3
)16
13
101;9–100;10
102;8–111;11
110;11–101;10
112;9–121;12
121;11–120;12
171;16–162;15
183;16–192;17
203;17–212;20
272;25–263;24
272;25–272;26
284;25–293;26
312;29–321;32
373;34–364;33
375;32–384;35
385;33–394;36
375;33–384;34
385;34–394;35
403;37–403;38
433;41–424;38
453;43–444;40
484;44–475;43
476;42–485;43
486;42–495;45
486;43–495;44
514;48–505;45
587;51–596;54
587;52–596;53
605;56–596;53
615;57–606;54
5
)3
6
80;8–71;7
18
11
50
)28
5
92;7–101;10
102;8–111;11
143;12–152;13
153;12–162;15
153;13–162;14
163;13–172;16
222;20–213;19
262;24–271;27
272;25–281;28
303;27–294;26
315;27–324;28
315;26–324;29
325;27–334;30
343;32–334;29
353;33–344;30
394;35–385;34
396;33–405;36
396;34–405;35
406;34–415;37
404;37–395;34
414;38–405;35
433;40–442;43
444;40–444;41
475;42–466;41
485;44–476;41
495;45–486;42
523;50–514;47
556;49–547;48
558;47–567;50
558;48–567;49
566;50–557;49
1
)28
)8
17
)1
)2
)1
6
9
)12
17
)4
)1
)10
)4
2
)13
5
)21
)17
14
)16
7
12
)8
)15
)5
4
)19
)17
)10
10
)5
22
0
18
15
)1
)2
6
32
)6
0
6
16
7
)15
)6
0
Standard deviation of the fit:15 kHz
15NCHCCO
2
10
)11
)11
1
2
40;4–30;3
1;3–31;2
20903.976
21393.448
)8
)2
4

142
M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
Table 1 (continued)
shows the transition frequencies; the spectroscopic
constants are given in Table 2. The inertia defects
#
Transition
JK0 a0Kc0 –JK00a00Kc00
Frequency
(MHz)
Resid.
(kHz)
2
ꢀ
(0.34. . . amu A , Table 5) confirm the expected planarity
of the molecule, the 18O-species again off by ꢁ25%. As
indicated above, all spectra were very dense. Obviously
cyanoketene was not the only molecule that was formed
by the pyrolysis and was giving rise to rotational tran-
sitions: this could be concluded from changing relative
intensities of adjacent assigned and unassigned lines
during the course of the reaction. But we could not
identify this one or even more other transient species
and we did also not succeed in assigning vibrational
satellites.
3
4
41;4–31;3
20441.799
20938.173
20922.283
20931.139
36519.517
37421.124
36602.359
36691.305
36635.387
36636.157
36641.342
36652.966
36668.796
31742.827
26750.750
36735.370
28463.783
29221.496
30123.142
33813.008
31176.540
32390.533
30341.312
31151.544
39049.053
36089.205
36537.326
31974.292
34829.041
36633.691
34418.825
29375.342
32380.224
35941.948
30070.294
36859.807
30433.657
33382.834
36068.457
36927.970
34744.797
37075.343
38571.780
32457.828
36087.809
26433.862
)4
)16
32
)11
)2
)7
8
42;2–32;1
42;3–32;2
43;1–33;0
5
6
7
70;7–60;6
71;6–61;5
8
9
7
2;6–62;5
2;5–62;4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
7
)15
36
)34
9
73;5–63;4
73;4–63;3
74;3–64;2
7
5;2–65;1
76;1–66;0
1;1–00;0
11;0–10;1
1;2–10;1
)21
50
2
1
3.2. Nuclear quadrupole coupling
31
19
)29
)11
23
)23
7
2
The small electric quadrupole moment of the 14N-
nucleus gives rise to an observable splitting of only the
low-J transitions. Three of them have been resolved as
shown in Table 3. A least squares analysis based on first
order quadrupole perturbation yielded the values
51;4–50;5
61;5–60;6
71;6–70;7
7
2;6–81;7
1;7–80;8
8
91;8–90;9
100;10–91;9
13
6
vaa ¼ ꢀ2:461ð13ÞMHz; vbb ¼ 1:220ð12ÞMHz;
vcc ¼ 1:241ð18ÞMHz:
102;8–111;11
131;12–130;13
171;16–171;17
183;15–192;18
193;16–202;19
274;24–283;25
282;26–273;25
284;24–293;27
294;25–303;28
332;31–341;34
375;33–384;34
385;34–394;35
423;39–423;40
443;42–434;39
453;43–444;40
463;44–454;41
476;42–485;43
494;45–485;44
534;50–525;47
577;50–586;53
587;52–596;53
605;55–596;54
605;56–596;53
18
0
20
)21
)6
)14
)19
)5
0
The 14 transitions of Table 3 allowed a simultaneous fit
of two independent coupling constants and the three
unsplit frequencies. Their values agree well with those of
Table 1 which were obtained at higher pressure and
therefore unsplit. Fig. 3 shows as an example the hy-
perfine structure of the transition 212 101: measured
and calculated with the above constants. Using the angle
of 28.22° between a-axis and NBC-bond from the
structure calculations and assuming that in such a nitrile
the z-axis of the coupling tensor coincides with the bond
axis, we calculated the coupling constants in the bond
axis system as
)5
10
8
)2
9
)13
)6
2
14
)1
)8
12
)7
3
vxx ¼ 2:709ð44ÞMHz; vzz ¼ ꢀ3:950ð82ÞMHz;
vyy ¼ vcc ðdue to the molecular symmetry planeÞ:
The asymmetry parameter therefore works out to
g ¼ ꢀ0:372ð15Þ.
Standard deviation of the fit:17 kHz
NCHCC18O
3.3. Dipole moment
1
2
4
1;41–31;31
40;4–30;3
2;3–32;2
43;2–33;1
2;2–32;1
20054.663
20503.253
20520.727
20528.882
20535.620
20977.816
25065.117
25617.576
25648.655
25662.694
25668.427
25678.748
26219.088
147
)23
41
The cell—cross section 8 ꢂ 2 cm2 for better homoge-
neity of the electric field—was calibrated with OCS, and
MuenterÕs [25] value was used for its dipole moment.
The DC-voltage was taken from a digitally controlled
high voltage and high precision power supply and the
modulating square-wave was superimposed. The tran-
sitions 313 212 and 404 303 were found to be isolated
enough that the Stark shifts could be followed up to
10 MHz. Because of the small quadrupole moment of
the 14N-nucleus, the strong field case was expected even
at low field strengths leading to doublets for each
3
4
4
)381
)102
)51
)7
5
4
6
41;3–31;2
51;5–41;4
50;5–40;4
7
8
50
9
52;4–42;3
)13
58
10
11
12
13
53;3–43;2
54;2–44;1
52;3–42;2
)155
26
51;4–41;3
)9
Standard deviation of the fit: 54 kHz

M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
143
Table 2
Rotational and centrifugal distortion constants of cyanoketene isotopomers
NCCHCO
NCCDCO
N13CCHCO
15NCCHCO
NCCHC18
O
Rotational constants (MHz)
A
B
C
29601.197 (5)
2812.1669 (6)
2563.7072 (5)
24599.240 (5)
2813.9840 (7)
2520.8095 (7)
29590.292 (4)
2787.6639 (4)
2543.2470 (3)
29249.319 (5)
2734.0364 (5)
2496.0392 (5)
29142. (58)
2680.308 (15)
2449.381 (17)
Centrifugal distortion constants (quartic in kHz, sextic and octic in Hz)
DJ
1.5596 (20)
)115.021 (63)
1.5192 (32)
)83.652 (76)
1.5210 (13)
)113.941 (42)
1.4855 (17)
)111.612 (51)
1.73 (31)
)110.79 (92)
2812.93a
0.36895a
8.079a
0.006862a
DJK
DK
dj
2812.93 (81)
0.36895 (9)
1574.94 (59)
0.39490 (9)
2818.46 (60)
0.35893 (8)
2768.22 (65)
0.34708 (6)
dk
8.097 (17)
0.00686 (15)
6.674 (24)
0.00766 (26)
7.915 (14)
0.00661 (18)
7.748 (23)
0.00648 (13)
UJ
UJK
b
0.0
)25.54 (75)
0.0
)10.56 (68)
0.0
)26.89 (68)
0.0
)25.1 (11)
0.0
)25.54a
836.a
0.00242a
0.0
27.17a
UKJ
UK
836. (24)
0.00242 (1)
0.0
275. (15)
0.00265 (3)
0.0
886. (24)
0.00243 (3)
0.0
835. (33)
0.00227 (3)
0.0
/
j
b
/
/
jk
27.17 (88)
)0.0163 (38)
16.04 (76)
)0.0073 (23)
23.1 (12)
)0.0153 (22)
24.3 (21)
)0.0140 (39)
k
kk
)0.0163a
a Values fixed to parent molecule.
b Fixed to 0.0.
Table 3
Resolved hyperfine transitions of cyanoketene
J0Ka0 Kc0
J00Ka00Kc00
Observed (MHz)
obs. ) calc.
F 0
F 00
Rel. Int.
munperturbed
1 1 1
0 0 0
32161.726
32162.283
32162.629
0.002
0.009
0
2
1
1
1
1
0.11
0.55
0.33
32162.355 (6)
)0.011
2 1 2
1 0 1
37289.265
37289.782
37290.159
37290.506
37291.133
)0.016
0.013
1
2
3
2
1
0
2
2
1
1
0.11
0.08
0.47
0.25
0.08
37290.200 (5)
27285.513 (4)
)0.007
0.000
0.009
2 1 1
2 0 2
27284.596
27285.196
27285.438
27285.824
27286.210
27286.429
0.001
)0.007
0.011
2
1
3
2
3
1
1
1
3
2
2
2
0.05
0.15
0.41
0.23
0.05
0.05
0.000
)0.007
)0.006
MJ -component according to MI2 ¼ 0; 1. This was true for
some of them from which the dipole moment compo-
nents could be determined straightforwardly as
could be measured without quadrupole perturbation.
The results:
la ¼ 2:844ð12ÞD; lb ¼ 2:112ð9ÞD; jltotj ¼ 3:542ð15ÞD:
la ¼ 2:816ð12ÞD; lb ¼ 2:110ð9ÞD; jltotj ¼ 3:519ð17ÞD:
This increase in la beyond the error margins by the
substitution shows that the total dipole moment is al-
most parallel to the NBC-bond with an angle of 36.8° to
the a-axis, compared to the 28.2° between a-axis and
that bond.
However, some components did not decouple even at
high field strengths and showed a multiplet behavour.
To confirm the results, these transitions were calculated
as intermediate field cases by solving the secular equa-
tion of Mizushima [26] and using the above compo-
nents. The agreement was excellent; an example is
shown on Fig. 4 with the values given in Table 4. After
the 15N-species had been prepared, the field dependence
of all four M-components of the transition 404 303
3.4. Structure calculations
As a planar molecule with 6 atoms, cyanoketene has 9
structural parameters: 5 bond distances and 4 bond

144
M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
ertheless, we can give a reliable geometry for the mole-
cule by using two different methods leading to similar
results. They are based on the moments of inertia as
calculated from WatsonÕs determinable constants [24]
and given in Table 5.
3.4.1. Method I
In a rigid rotor approximation the five isotopomers
yield 10 linear independent moments of inertia, one more
than parameters. Therefore they can not only be solved
for the parameters, but also a fit is possible with the one
in excess. Because the very small moments Ia are ex-
tremely sensitive to the effects of zero point vibrations the
Ib and Ic moments are used. One more approximation
has to be made: because the a-coordinate of the H-atom
is very close to zero, the moment Ib of the heavier D-
species is smaller than that of the parent molecule due to
non-rigidity effects. This is in contradiction to the physics
of a rigid body and leads to imaginary results. To avoid
this difficulty, IbðDÞ was set equal to IbðHÞ which is
tantamount to setting the coordinate equal to zero. With
this modification these ten moments of inertia can be
Fig. 3. Hyperfine structure of the transition 212 101: measured and
calculated.
2
ꢀ
reproduced with a mean deviation of 0.01 amu A , cor-
responding to 0.005%, by the structure shown in Table 6
under ÔMethod I.’’ Because the errors are determined by
the non-rigidity effects, those given by the fitting proce-
dure are not meaningful. Instead, they are estimated
from the degree of agreement or disagreement between
the two structures resulting from the two different
methods and as outlined under Section 3.4.3.
Fig. 4. Stark multiplet of the transition 404 303 at 500 V: measured
and calculated.
3.4.2. Method II
angles. For a complete rs-structure [27] substitution of
the two C-atoms in the ketenyl-chain would be neces-
sary in addition to the actual measurements. Therefore
only a mixed rs ꢀ r0-structure can be calculated. Nev-
KraitchmanÕs equations [28] allow to calculate the
coordinates of the four substituted atoms in the parent
molecule, yielding at least one rs-distance: NBC–. For
Table 4
Stark multiplet MJ00 ¼ 3 of the transition 404 303 at 500 V
Field str. (V/cm)
Meas. (MHz)
Calc. (MHz)
Rel. intens.
MJ0
MI0
MJ00
MI00
587.3
21486.690 (45)
21487.190 (39)
21487.310 (45)
21487.710 (45)
21486.698
21487.148
21487.318
21487.702
21488.054
21488.098
0.12
0.11
0.24
0.15
0.08
0.10
ꢃ3
ꢃ3
ꢃ3
ꢃ2
ꢃ4
ꢃ2
ꢄ1
ꢃ0
ꢃ1
ꢃ0
ꢄ1
ꢃ1
ꢃ3
ꢃ3
ꢃ3
ꢃ3
ꢃ3
ꢃ3
ꢄ1
ꢃ0
ꢃ1
ꢄ1
ꢃ0
ꢃ0
21488.080 (51)
Table 5
Moments of inertia of cyanoketene from WatsonÕs determinable constantsa
Isotopomer
Ia
Ib
Ic
D
Ia0 ¼ Ic ꢀ Ib
Parent
D
17.07293
20.54450
17.07922
17.27832
17.34191
179.71985
179.60162
181.29954
184.85568
188.56138
197.13554
200.48814
198.72148
202.47991
206.33691
0.3428
0.3420
0.3427
0.3459
0.4336
17.41568
20.88652
17.42195
17.62423
17.77552
13C
15N
18O
a Ref. [24].

M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
145
Table 6
Structure of cyanoketene (distances in pm, angles in degrees)
Parameter
Method I
Method II
Proposed structure
1
r(NBC–)
r(BC–C@)
r(–C@C@)
r(@C@O)
115.86
143.02
132.64
115.33
108.15
119.76
119.29
178.47
177.49
115.71
143.79
134.19
114.33
103.67
120.07
117.65
177.76
175.43
115.8 (1)
143.4 (6)
133.4 (12)
114.8 (8)
105.9 (34)
119.9 (2)
118.5 (12)
178.1 (5)
176.5 (15)
2
3
4
5
r(@C–H)
6
°(H–C@C@)
°(BC–C@C@)
°(NBC–C@)
°(–C@C@O)
7
8a
9a
a N- resp. O-atom bent towards H-atom.
the complete structure four more coordinates are nee-
ded: those of the two ketenyl-C-atoms. They have to
meet the conditions for the center of mass in the parent
P
molecule
migi ¼ 0; i ¼ 1 . . . 6, g ¼ a; b, as well as to
reproduce the moments of inertia of the parent mole-
cule. In addition, we can make use of the condition for
the moment of deviation in the principal axis system of
P
the parent molecule:
miaibi ¼ 0. So we have five
equations for the four unknown coordinates, and a fit-
ting procedure can be applied. From these 12 coordi-
nates of the six atoms in the a; b-plane the structure can
be evaluated. For these calculations the moments Ia and
Ib are needed. If the values of Ia are used as they are
shown in Table 5, the results are unrealistic because the
vibrationally induced uncertainties represent a relatively
large percentage of these small moments. If, instead, the
moments Ia0 ¼ Ic ꢀ Ib as also given in Table 5 are used, a
reasonable structure results, listed in Table 6 under
‘‘Method II.’’
Fig. 5. Geometry of cyanoketene: parameters in Table 6.
the dying out could be very well described by a first
order rate law. The time constant varied from 2.4 to
6.5 min, almost linearly with the dimensions of the cell
cross section. Broad (circa 5 MHz) bumps appeared then
at some other frequencies, presumably a reaction
product of cyanoketene with itself and perhaps HCl. We
also trapped the products of the pyrolysis behind the
absorption cell by LN2 during 4 h, monitoring the pro-
duction of cyanoketene by observing a rotational tran-
sition. Then the trap was allowed to warm up very
slowly, the escaping gases drawn back through the cell
and still monitoring a strong microwave transition. Af-
ter some time the gauges indicated a buildup of pressure,
and for half an hour gas was flowing through the cell,
but no microwave transition could be observed. This gas
was obviously the HCl from the pyrolysis reaction
which was also trapped at the LN2-temperature. The
remaining white solid in the trap first turned yellowish
and then brown when further warmed up very carefully.
No significant amounts of gas were given off during this
process. So it was not possible to recover gaseous cya-
noketene from the solid phase. To exclude the possibility
that cyanoketene had reacted with part of the HCl, we
made another trapping test with dry ice cooling. At this
temperature, )80 °C, no HCl at all is trapped, but once
3.4.3. Proposed structure
Table 6 shows that the two methods result in
structures which are not significantly different; only the
C–H-distance exhibits a remarkable discrepancy. The
essentials, however, are in good agreement, especially
the behavior of the NBC–C– and C@C@O-chains: it is
indeed impossible to reproduce the measured moments
of inertia if these chains are fixed to a linear arrange-
ment. Therefore, it seems to be justified to propose a
structure whose parameters are the mean of the values
obtained by the two methods and to assume error limits
which exceed the range between the two values by 50%.
This structure is also given in Table 6 and shown on
Fig. 5.
3.5. Stability of cyanoketene
As mentioned in Section 1, severe contradictions exist
concerning the stability of the compound. We therefore
carried out some tests to elucidate these discrepancies.
We observed the decay of strong absorption lines after
the cell had been closed at both ends, and we found that

146
M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
more a white deposit was collected during 2 12 h. While
warming up again very cautiously this turned yellow,
but only at room temperature pressure was built up.
This gas, in contrast to cyanoketene coming from the
oven, condensed into the cell at )15 °C and did not show
any rotational transitions. So with this method too,
cyanoketene could not be re-sublimated from the solid
state. These results corroborate strongly the findings of
[16], and it seems very unlikely that a liquid phase of
cyanoketene could be obtained as claimed in [20].
a factor of 20–30 greater than in the other cyano-
compounds and by a factor of 10 or 5 than in the
haloketenes, bearing in mind that the bonding is dif-
ferent because of the missing C-atom in between. But it
shows clearly that there is considerable interaction
between the triple bond and the ketene system, leading
to a contribution of a ionic structure to the cumulene
system:
N
ðꢀÞ@C@ðCHÞ–CðþÞ@O $ NBC–ðCHÞ@C@O:
There are not many possibilities to compare the struc-
ture of cyanoketene with that of similar compounds. In
the case of chloroketene and bromoketene assumptions
had to be made which influence strongly the essential
parameters, and we are left with ketene itself and flu-
oroketene. This comparison is done in Table 8; two
theoretical calculations for cyanoketene are also in-
cluded. We see that the C@C-double bond is longer than
in ketene and fluoroketene whereas the C@O-bond is
shorter, in accordance with the above resonance struc-
ture. The two ab initio-calculations point into the same
direction. They also predict the bend of the two ‘‘linear’’
chains towards the H-atom, now definitely confirmed by
experiment.
4. Discussion
With cyanoketene the series of haloketenes which
have been investigated by microwave spectroscopy has
been extended to a member of the pseudo halogen
derivatives. In many aspects it fits into the series, but
the bond system shows a remarkable difference, not
only compared to chloro- and bromoketene but also to
other cyano-groups attached to an sp2-carbon atom.
This is easily recognized from the quantities in Table 7
which are deduced from the quadrupole coupling ten-
sor: asymmetry parameter and p-character. Both are by
Table 7
Comparison of coupling constants
Cyanoketenea
Cyanoalleneb
Acrylnitrilec
Chloroketened
Bromoketenee
vxx (MHz)
vyy (MHz)
vzz (MHz)
g
2.695 (44)
1.241 (18)
)3.936 (82)
)0.369
2.13 (23)
2.08 (15)
)4.21 (25)
)0.011
2.169 (24)
2.100 (5)
)4.269 (51)
)0.016
40.2 (15)
43.19 (4)
)83.4 (10)
0.036
)316.0 (13)
)341.0 (9)
656.9 (13)
0.038
p-char. (%)
10
0.3
0.5
2
2
a This work.
b Ref. [30].
c Ref. [31].
d Ref. [10].
e Ref. [12].
Table 8
Experimental and calculated structures of ketenes (distances in pm, angles in degrees)
Parameter
Cyanoketene exp.a
Ketene exp.b
Fluoroketene exp.c
Cyanoketene calc.d
Cyanoketene calc.e
1
r(NBC–)
r(BC–C@)
r(–C@C@)
r(@C@O)
115.8
143.4
133.4
114.8
105.9
119.9
118.5
178.1
176.5
—
—
113.7
142.4
132.1
113.3
107.2
118.1
120.6
179.4
178.7
116.4
142.9
133.3
115.5
108.3
118.2
121.3
179.0
179.0
2
—
—
3
131.4
116.2
108.3
118.5
—
131.7
116.7
110.0
122.3
119.5f
4
5
r(@C–H)
6
°(H–C@C@)
°(BC–C@C@)
°(NBC–C@)
°(–C@C@O)
7
8g
9g
—
180.0
178.0
a This work.
b Ref. [29].
c Ref. [11].
d Ref. [15]a.
e Ref. [16].
f °(F-C@C@).
g N-resp. O-atom bent towards H-atom.

M. Hahn et al. / Journal of Molecular Spectroscopy 223 (2004) 138–147
147
[14] M. Hahn, H. Bodenseh, Ninth Colloquium on High Resolution
Molecular Spectroscopy, Riccione 1985, C26 (Postdeadline).
[15] (a) L. Gong, M.A. McAllister, T.T. Tidwell, J. Am. Chem. Soc.
113 (1991) 6021–6028;
Acknowledgments
We appreciate gratefully support from the Deutsche
Forschungsgemeinschaft and the Fonds der Chemischen
€
Industrie and we thank Dr. Hans Hubner for use of his
program CAESAR.
(b) M.A. McAllister, T.T. Tidwell, Can. J. Chem. 72 (1994) 882–
887;
(c) M.A. McAllister, T.T. Tidwell, J. Org. Chem. 59 (1994) 4506–
4515;
(d) N.H. Werstiuk, J. Ma, M.A. McAllister, T.T. Tidwell, D.
Zhao, J. Chem. Soc. Faraday Trans. 90 (1994) 3383–3390.
[16] D.W.J. Moloney, M.W. Wong, R. Flammang, C. Wentrup, J.
Org. Chem. 62 (1997) 4240–4247.
References
[1] H. Staudinger, Ber. Dtsch. Chem. Ges. 38 (1905) 1735–1739.
[2] N.T.M. Wilsmore, J. Chem. Soc. 91 (1907) 1938–1941.
[3] W. Weyler Jr., W.G. Duncan, H.W. Moore, J. Am. Chem. Soc. 97
(1975) 6187–6192.
[17] K. Sung, Canad. J. Chem. 78 (2000) 562–567.
[18] J.L. Holmes, P.M. Mayer, M. Vasseur, P.C. Burgers, J. Phys.
Chem. 97 (1993) 4865–4870.
[19] G. Maier, H.P. Reisenauer, K. Rademacher, Chem. Eur. J. 4
(1998) 1957–1963.
[4] (a) W.T. Brady, T.C. Cheng, J. Org. Chem. 42 (1977) 732–734,
and previous papers;
[20] D.A. Efremov, P.M. Zavlin, N.S. Essentseva, J.C. Tebby, J.
Chem. Soc. Perkin Trans. 1 (1994) 3163–3168.
[21] S.B. Charnley, P. Ehrenfreund, Y.-J. Juan, Spectrochim. Acta
Part A 57 (2001) 685–704.
(b) D.J. Pollart, H.W. Moore, J. Org. Chem. 54 (1989) 5444–5448.
[5] A. Hotzel, R. Neidlein, R. Schulz, A. Schweig, Angew. Chem. Int.
Ed. 19 (1980) 739–740.
[6] J.E. Gano, R.H. Jacobson, R.H. Wettach, Angew. Chem. Int. Ed.
22 (1983) 165.
[22] F. Borget, T. Chiavassa, J.-P. Aycard, Chem. Phys. Lett. 348
(2001) 425–432.
[7] D. Colbourne, N.P.C. Westwood, J. Chem. Soc. Perkin Trans. II
(1985) 2049–2054.
[23] G. Schroeter, Ch. Seidler, J. Pract. Chem. 105 (1922) 165–
176.
[8] M.C.L. Gerry, W. Lewis-Bevan, N.P.C. Westwood, Can. J.
Chem. 63 (1985) 676–677.
[24] J.K.G. Watson, in: J.R. Durig (Ed.), Vibrational Spectra and
Structure: A Series of Advances, vol. 6, Elsevier, Amsterdam,
1977, pp. 1–89.
[9] B. Rozsondai, J. Tremmel, V.N. Khabashesku, N.D. Kagrama-
nov, O.M. Nefedov, J. Am. Chem. Soc. 111 (1989) 2845–2849.
[10] M.C.L. Gerry, W. Lewis-Bevan, N.P.C. Westwood, J. Chem.
Phys. 79 (1983) 4655–4663.
[25] J.S. Muenter, J. Chem. Phys. 48 (1968) 4544–4547.
[26] M. Mizushima, J. Chem. Phys. 21 (1953) 539–541.
[27] C.C. Costain, J. Chem. Phys. 29 (1958) 864–874.
[28] J. Kraitchman, Am. J. Phys. 21 (1953) 17–24.
[29] R.D. Brown, P.D. Godfrey, D. McNaughton, A.P. Pierlot, W.H.
Taylor, J. Mol. Spectrosc. 140 (1990) 340–352.
[30] A. Bouchy, J. Demaison, G. Roussy, J. Barriol, J. Mol. Struct. 18
(1973) 211–217.
[11] R.D. Brown, P.D. Godfrey, K.H. Wiedenmann, J. Mol. Spec-
trosc. 136 (1989) 241–249.
[12] N.P.C. Westwood, W. Lewis-Bevan, M.C.L. Gerry, J. Mol.
Spectrosc. 136 (1989) 93–104.
[13] H. Bock, T. Hirabayashi, S. Mohmand, Chem. Ber. 114 (1981)
2595–2608.
[31] M. Stolze, D.H. Sutter, Z. Naturforsch. 40a (1985) 998–1010.