A systematic investigation of long-range couplings in the 1H NMR spectra of dimeric indanones

Article abstract of DOI:10.1007/s007060170136

During the synthesis of methyl substituted indanones by intramolecular cyclization of 3-phenyl-propionic acids, dimerization led to by-products which can be considered as dimeric indanones. The proton NMR spectra of these compounds exhibit pronounced scalar couplings over up to seven bonds. A series of structures of the above type were investigated, and their NMR spectroscopic behaviour is discussed.

Full text of DOI:10.1007/s007060170136

Monatshefte fuÈr Chemie 132, 267±278 (2001)
A Systematic Investigation of Long-Range
1
Couplings in the H NMR Spectra of Dimeric
Indanones
Ã
Susanne Felsinger, Hermann Kalchhauser , and Horst Neudeck
Institute of Organic Chemistry, University of Vienna, A-1090 Vienna, Austria
Summary. During the synthesis of methyl substituted indanones by intramolecular cyclization of 3-
phenyl-propionic acids, dimerization led to by-products which can be considered as dimeric
indanones. The proton NMR spectra of these compounds exhibit pronounced scalar couplings over
up to seven bonds. A series of structures of the above type were investigated, and their NMR
spectroscopic behaviour is discussed.
Keywords. NMR spectroscopy; Indanones; Long-range couplings.
Introduction
In the context of an ongoing project concerning the synthesis of optically active
spiro compounds [1±4], a series of substituted indanones were prepared as key
intermediates [5]. Upon intramolecular Friedel-Crafts acylation of phenylpropa-
noic acid under PPA-catalysis (PPA ˆ polyphosphoric acid), by-products were
obtained which could be spectroscopically (NMR, MS) identi®ed as dimerized
indanones resulting from an intermolecular aldol condensation of two molecules of
product. The process can be well understood bearing in mind the reaction
conditions employed: attack of a molecule of 1-indanone in its enol form on a
protonated species of the same kind results in the formation of a dimeric structure
by an acid-catalyzed aldol condensation [6]. Subsequent acidic dehydratization
yields the title compounds (Scheme 1). Only isomers of the (E) series with respect
to the newly formed double bond are obtained, obviously because of pronounced
steric hindrance due to the close proximity of the carbonyl oxygen and H-C(30) in
the corresponding (Z) products.
During the routinely performed NMR spectroscopic characterization of some
of the above compounds, a priori unexpected cross-peaks were observed in the
ge-DQF-COSY spectra which could be traced back to scalar couplings over up to
seven bonds. A detailed investigation of this behaviour is the subject of the present
paper.
Ã
Corresponding author

268
S. Felsinger et al.
Scheme 1. Formation of compounds 1±5 as by-products in the synthesis of substituted indanones by
intramolecular Friedel-Crafts acylation of phenylpropanoic acid under PPA-catalysis
Results and Discussion
A summary of the structures dealt with in this article together with the numbering
system applied is presented in Scheme 2.
Two major structural features characterize compounds 1±5 as can immediately
be seen by inspection of a Dreiding molecular model or a rough minimization using
a PC-based program [7]: planarity and rigidity. These are rather favourable
properties with respect to NMR spectroscopy, especially for experiments making
Ê
use of0NOE effects: the interproton distances are small (e.g. about 2 A for H-C(8) ±
giving rise to strong and easily detectable cross-peaks in NOESY spectra.
0
0
Ê
H-C(3 ) and ca. 3 A for H-C(8) ± H-C(6) or H-C(8 ) ± H-C(6 )) and invariant, thus

Long-Range Couplings in Dimeric Indanones
269
Scheme 2. Dimeric indanones discussed in this paper together with the numbering scheme chosen
for convenience; formally, the compounds have to be viewed as 2-(1-indylidene)-1-indanones
Consequently, a complete and ± with few exceptions ± unambiguous NMR
spectroscopic assignment (1H, 13C) could easily be established employing standard
two-dimensional techniques (ge-DQF-COSY, ge-NOESY, ge-HSQC, ge-HMBC).
The corresponding shift values are collected in Tables 1 and 2.
1
Table 1. H chemical shifts of compounds 1±5 (ꢀ/ppm)
1
2
3
4
5
H-C(3)
7.83
7.39
7.55
7.52
4.02
7.80
7.34
7.33
7.40
3.12
3.56
±
7.66a
7.28
7.35
±
7.63
±
7.57
7.16
±
7.46
±
H-C(4)
H-C(5)
7.38
7.43
3.98
7.59
±
7.18
±
H-C(6)
±
H-C(8)
3.81
7.65a
7.27
7.18
±
3.76
7.53
7.14
±
3.78
7.43
±
H-C(30)
H-C(40)
H-C(50)
H-C(60)
H-C(80)
H-C(90)
CH3-C(4)
CH3-C(5)
CH3-C(6)
CH3-C(40)
CH3-C(50)
CH3-C(60)
7.19
7.30
3.07
3.55
2.41
±
7.02
±
±
2.98
3.54
±
2.98
3.53
±
2.94
3.54
2.37
±
±
±
2.35
2.29
±
±
2.41
±
±
2.40
2.42
±
±
2.44
±
±
±
2.31
2.19
±
2.29
±
2.26
a
Assignment interchangeable together with the corresponding 13C shifts (cf. Table 2)

270
S. Felsinger et al.
Table 2. 13C chemical shifts of compounds 1±5 (ꢀ/ppm)
1
2
3
4
5
C(1)
195.16
139.51
123.56
127.26
133.56
125.90
148.50
32.97
126.25
154.95
140.85
125.93
130.38
126.81
125.73
151.67
30.89
31.50
±
195.58
139.34
120.96a
127.45
134.14
135.06c
147.49
31.69
126.26
155.44
140.70
123.40a
127.09
131.07
134.98c
150.69
29.80
31.43
±
195.37
139.79
123.68
137.15
134.78
125.62
145.93
32.74
126.66
154.92
141.20
126.34
136.40
131.50
125.42
148.95
30.52
31.90
21.22
±
195.43
137.64
120.76
129.32
142.51
133.14
147.75
32.07
125.99
155.29
138.65
123.10
128.92
139.15
133.40
150.82
30.20
31.68
±
195.71
139.60
121.05
137.28b
135.40
134.62
144.91
31.37
126.58
155.42
140.99
123.70
136.63b
132.28
134.62
147.98
29.39
31.79
21.13
±
C(2)
C(3)
C(4)
C(5)
C(6)
C(7)
C(8)
C(9)
C(10)
C(20)
C(30)
C(40)
C(5)0
C(60)
C(70)
C(80)
C(90)
CH3-C(4)
CH3-C(5)
CH3-C(6)
CH3-C(40)
CH3-C(50)
CH3-C(60)
±
±
20.28
14.67
±
±
17.89
±
±
17.90
21.57
±
±
21.29
±
±
±
19.93
15.49
±
18.86
±
18.78
a
Assignment interchangeable together with the corresponding 1H shifts (cf. Table 1); b;c assignment
interchangeable
The assignment procedure was based on the easily detectable proton NMR signal of H-C(8) as an
anchor point (broad singulet at ca. 4 ppm) in all cases. A strong NOE connectivity allowed to identify
H-C(30), whereas a weaker effect was observed with H-C(6) (1, 3) or CH3-C(6) (2, 4, 5). A suitable
combination of dipolar and scalar interactions depending on the compound under investigation then
afforded the chemical shifts of all protons. In some instances, especially in the case of 1, the shift
values of the aromatic protons had to be extracted from the 2D data because of severe signal overlap
1
in the 1D H NMR spectra. The protonated carbon atoms were assigned straightforwardly from a
HSQC experiment; the somewhat trickier task of identifying all quarternary carbons was successfully
tackled via the HMBC technique optimized for different 1H,13C long-range couplings. As for
compound 5, the 13C resonances of C(6) and C(60) are almost perfectly isochronous; however,
adequate data processing allowed to prove the existence of two signals of almost equal frequency
(Áꢀ ˆ 0:96 Hz). The 13C NMR spectra show no noticeable features; in particular, the methylation
increments of the aromatic carbons are in good accordance with literature values [8], ranging from
about 6 to 12 ppm with a mean value of 8.7 ppm (cf. Table 2).
When doing routine NMR spectroscopic characterization of 3, cross-peaks were
detected in the ge-DQF-COSY spectrum which at a ®rst glance seemed to contradict
the structures proposed for the title compounds (Fig. 1). In particular, scalar
couplings were found between H-C(8) and H-C(90) (5J) and, still more surprising,

Long-Range Couplings in Dimeric Indanones
271
Fig. 1. ge-DQF-COSY spectrum of 3 (600 MHz; detail); the cross-peaks discussed in the text
(H-C(8)!H-C(90), H-C(8)!CH3-C(4), H-C(80)!CH3-C(40)) are marked by arrows
between H-C(8) and CH3-C(4) on the one hand and H-C(80) and CH3-C(40) on the
other hand (7J in both cases).
To exclude phenomena different from J-coupling as the source of the unex-
pected cross-peaks (e.g. intermolecular association, stacking, cross correlation
effects between CSA and dipolar interactions [9, 10], etc.), measurements under
varying conditions (dilution series, temperature variation, solvent change) were
performed; in any case, the spectroscopic pattern remained unaffected. When,
however, a COSY experiment optimized for long-range couplings (lr-COSY) was
carried out with 3, the J-origin of the above cross-peaks became immediately
evident: whereas the signals due to vicinal couplings decreased in intensity,
increasing peak heights were found for resonances indicating long-range inter-
actions (Fig. 2).

272
S. Felsinger et al.
Fig. 2. COSY spectrum of 3 optimized for long-range couplings (600 MHz; detail); the cross-peaks
discussed in the text (H-C(8)!H-C(90), H-C(8)!CH3-C(4), H-C(80)!CH3-C(40)) are marked by
arrows
An additional feature of the lr-COSY experiment performed on 3 coming in handy concerns the
aromatic part of the spectrum. All protons of a speci®c aromatic moiety exhibit long-range cross-
peaks to the methyl substituent of the respective nucleus, thus allowing for a cross-check of the NMR
spectroscopic assignment or even to establish this assignment in cases where the information derived
from NOE spectroscopy is insecure or incomplete. The corresponding region of the lr-COSY
spectrum of 3 is shown in Fig. 3 from which the TOCSY-like character can easily be perceived.
As an encore, homodecoupling experiments were performed with 3. As
expected, the removal of unresolved long-range couplings led in the ®rst place to
line narrowing and therefore to an increased intensity of the corresponding signals
(Fig. 4). As discussed above for the aromatic part of the lr-COSY spectrum, this
feature can also be exploited for assignment purposes: the required experiment is
simple and time-saving, and the effect is pronounced (ca. 30%). In some cases,
however, especially in the aromatic region, besides intensity enhancement ad-

Long-Range Couplings in Dimeric Indanones
273
Fig. 3. COSY spectrum of 3 optimized for long-range couplings (600 MHz); horizontal cross-
sections at the frequencies of the methyl groups contain the complete proton spin system of the
corresponding aromatic moiety
ditional line splittings were observed in the decoupled spectra which were obscured
in the fully coupled spectra by signal overlap (Fig. 5).
Long-range couplings in NMR spectroscopy have been a matter of interest from
the very beginning of its application to structure elucidation, and various early
compilations deal with this subject both from practical and theoretical points of
view (see e.g. Refs. [11, 12]). The advent of more sophisticated and powerful
techniques for structural analysis, especially the development of multidimensional
methods, however, has led to a loss of interest in the ®eld as can easily be seen from
1
the citation data in a recent review article on this topic [13], and text books on H
NMR spectroscopy usually fall back upon statements like e.g. `couplings over ®ve
and more bonds can only very rarely be detected' [14].
Among the prerequisites required for the occurrence of long-range couplings,
W-arrangement, coplanarity, conjugation, and multiple coupling pathways are
among the most important. All these features are present in compounds 1±5 to an
extremely high extent, and scalar interactions over many bonds are accordingly
strongly favoured. In the following, the observed connectivities will be discussed
systematically.
A summary of proton-proton long-range couplings in compounds 1±5 (with the
exception of the trivial ones occuring within the spin system of the aromatic protons
of one benzene ring) is presented in Table 3. In an attempt to tidy up the vast amount
of information available, the data were classi®ed according to four different groups.
a) Inter-residue couplings. Long-range couplings which connect protons located in
different monomeric units of the dimers are termed inter-residue couplings in
this context. Two interactions of this type can be found in any of the compounds
1±5 connecting H-C(8) with H-C(90) and H-C(80). The former (®ve bonds),
which is rather pronounced, belongs to the group of homoallylic couplings
which have been discussed extensively in the older literature [15], mentioned in
all review articles on the topic as well as in basic and advanced text books on
NMR spectroscopy, and found applications in stereochemistry [16]. The

274
S. Felsinger et al.
Fig. 4. Methyl 1H resonances of 3; left: undisturbed spin system, middle: irradiation at the frequency
of H-C(80), right: irradiation at the frequency of H-C(8)
coupling between H-C(8) and H-C(80) (6 bonds) is considerably weaker and less
readily explained; maybe additional information transfer via the pathway
H-C(8) ± C(8) ± C(9) ± C(10) ± C(20) ± C(70) ± C(80) ± H-C(80) contributes to
render the overall effect observable. Actually, taking into account conjugation,
this route might even be the more important one, although one additional bond is
involved.
b) Couplings between aromatic hydrogens and protons of the adjacent ®ve-
membered rings. Whenever possible from the point of view of the substitution
pattern, scalar couplings between H-C(4) and H-C(8) and, analogously, H-C(40)
and H-C(80) are observed (1, 2, 4). Although we here deal with a 6JH;H like in the
H-C(8) ± H-C(80) case discussed above, this interaction is detected more readily
due to the two identical possible coupling pathways. Further couplings via 4
respectively 5 bonds belonging to this subgroup can be detected in 3 and 4
(cf. Table 3).

Long-Range Couplings in Dimeric Indanones
275
1
Fig. 5. Aromatic region of the H NMR spectrum of 3 (400 MHz); lower trace: undisturbed spin
system, middle trace: irradiation at the frequency of H-C(80), upper trace: irradiation at the frequency
of H-C(8); the spectra are normalized to equal intensity of the residual solvent signal (7.24 ppm)
c) Couplings between methyl groups and protons of the corresponding aromatic
rings. As already discussed extensively for compound 3 (vide supra, cf. Fig. 3),
in all cases the methyl groups are correlated to all aromatic protons of the same
ring via long-range couplings. Again, degenerate coupling pathways play an
important role within this group, especially for the cross-talk between methyl
groups and protons in para-position (6JH;H). With respect to practical appli-
cations, this is perhaps the most useful feature, being by no means restricted to
rather peculiar overall structures as in the present case. It allows the extraction

276
S. Felsinger et al.
Table 3. Proton-proton long-range couplings observed in compounds 1±5; a) inter-residue interac-
tions, b) couplings between aromatic hydrogens and protons of the ®ve-membered rings,
c) connectivities correlating methyl groups and the protons of the corresponding aromatic moiety,
d) others
1
2
a)
b)
a)
b)
c)
d)
a)
b)
c)
d)
a)
b)
c)
H-C(8)$H-C(90); H-C(8)$H-C(80)
H-C(4)$H-C(8); H-C(40)$H-C(80)
H-C(8)$H-C(90); H-C(8)$H-C(80)
H-C(4)$H-C(8); H-C(40)$H-C(80)
CH3-C(6)$H-C(3), H-(C4), H-(C5); CH3-C(60)$H-C(30), H-(C40), H-(C50)
CH3-C(6)$H-C(8); CH3-C(60)$H-C(80)
3
4
H-C(8)$H-C(90); H-C(8)$H-C(80)
H-C(6)$H-C(8); H-C(60)$H-C(80)
CH3-C(4)$H-C(3), H-C(5), H-C(6); CH3-C(40)$H-C(30), H-C(50), H-C(60)
CH3-C(4)$H-C(8); CH3-C(40)$H-C(80)
H-C(8)$H-C(90); H-C(8)$H-C(80)
H-C(4)$H-C(8); H-C(30)$H-C(80); H-C(40)$H-C(80)
CH3-C(5)$H-C(3), H-C(4); CH3-C(6)$H-C(3), H-C(4);
CH3-C(50)$H-C(30), H-C(40); CH3-C(60)$H-C(30), H-C(40)
CH3-C(6)$CH3-C(5), H-C(8); CH3-C(60)$CH3-C(50), H-C(80)
H-C(8)$H-C(90); H-C(8)$H-C(80)
d)
a)
b)
c)
5
±
CH3-C(4)$H-C(3), H-C(5); CH3-C(6)$H-C(3), H-C(5);
CH3-C(40)$H-C(30), H-C(50); CH3-C(60)$H-C(30), H-C(50)
CH3-C(4)$CH3-C(6), H-C(8); CH3-C(40)$CH3-C(60), H-C(80)
CH3-C(6)$H-C(8); CH3-C(60)$H-C(80)
d)
and, most likely, the assignment of the complete spin system of methyl
substituted benzene moieties in a compound from the TOCSY-like traces at the
frequencies of the corresponding methyl resonances provided these are known.
In a more general sense, the applicability of the method to alkyl substituted
aromatics, starting from the methylene group adjacent to the ring in this case,
might also be taken into consideration.
d) Others. These couplings include interactions between methyl groups and H-C(8)
or H-C(80), respectively, and between different methyl groups in disubstituted
representatives of the title compounds (4, 5). The most extended connectivities
observed in the course of this investigation (e.g. CH3-C(4)$H-C(8) and CH3-
7
C(40)$H-C(80) in 3; JH;H) can be found in this group, again enhanced and
promoted by the existence of two equivalent pathways available for information
transmission. Similar observations have been described for p-xylene (scalar
7
coupling between the protons of the two methyl groups; j Jj ˆ 0:6 Hz) [13]);
however, due to molecular symmetry (Áꢀ ˆ 0), this effect does not manifest
1
itself in H NMR spectra.
A parameter which has not been addressed so far is the size of the observed
long-range coupling constants. Although the phenomena can easily be detected
employing the appropriate two-dimensional techniques, their manifestation in the
1
1D H spectra is mostly restricted to line broadening. This is, however, not only
caused by the low values of nJH;H, but to a considerable extent also by the wealth of

Long-Range Couplings in Dimeric Indanones
277
information available, leading to numerous splittings and thus providing mainly
unresolved multiplets. In homodecoupled spectra, where part of the couplings are
removed, ®ne structure can be observed within some of the signals (cf. e.g. Fig. 5),
and at least approximate values can be extracted. All far-distance coupling
constants checked range considerably below 1 Hz, not exceeding 0.5 Hz in most
cases. No attempts have been made to analyze the complex high-order spin systems
of H-C(8), H-C(80), and H-C(90).
Experimental
Details on the synthesis and isolation of compounds 1±5 will be published elsewhere [5]. NMR
spectra: Bruker DRX 400 WB (1H: 400 MHz, 13C: 100 MHz) and DRX 600 (1H: 600 MHz, 13C:
150 MHz) spectrometers, 5 mm tubes, T ˆ 27ꢀC (300 K), sample concentration: ca. 30 mM, solvent:
CDCl3, referencing: relative to internal TMS using the solvent resonances as secondary standards
(7.24/77 ppm for 1H and 13C, respectively). Oxygen was removed by Ar bubbling for 5 min prior to
the experiments. 13C spectra were recorded in the J-modulated mode (APT, [17]). The 2D methods
used comprise the following experiments: ge-DQF-COSY [18, 19], long-range-COSY [20] (delay
for evolution of long-range couplings: 100 to 300 ms depending on the size of the long-range
interaction to be detected), ge-HSQC [21±24], ge-HMBC with low-pass J-®lter to suppress direct
lr
C;H
connectivities [25, 26] (optimized for
J
ˆ 7:5 Hz corresponding to an evolution delay of
66.7 ms), ge-NOESY [27, 28] (mixing time: 800 ms). In experiments employing gradient coherence
selection, gradient lengths of 1 ms and recovery delays of 100 ms were used throughout. Except for
the COSY experiments optimized for long-range couplings and the HMBC experiments (magnitude
mode), all spectra were recorded in the phase-sensitive mode. For all experiments, pulse programs
supplied by the spectrometer manufacturer were used; the ge-DQF-COSY sequence was modi®ed by
introducing a steady-state sandwich (gradient ± ꢁꢁ=2† ± gradient) prior to each scan to destroy
x
unwanted magnetization and suppress artifacts [29]. 1 k complex data points per FID were acquired,
and 128 experiments were performed in !1. Before Fourier transformation, zeroes were added in
both dimensions to afford 1 k  512 real data points in the frequency domain. Squared sine bells
shifted by ꢁ=2 were applied as ®lter functions in both dimensions for the phase-sensitive spectra,
whereas unshifted sine bells gave the best results in the case of magnitude-mode representations.
Acknowledgements
The advice of Prof. Dr. O. Hofer in questions of nomenclature is gratefully acknowledged.
References
[1] Neudeck H (1996) Monatsh Chem 127: 185
[2] Neudeck H (1996) Monatsh Chem 127: 201
[3] Neudeck H, Melmer M (1996) Monatsh Chem 127: 275
[4] Neudeck H (1996) Monatsh Chem 127: 417
[5] Neudeck H (in preparation)
[6] Baigrie LM, Cox RA, Slebocka-Tilk H, Tencer M, Tidwell TT (1985) J Am Chem Soc 107: 3640
[7] Alchemy II (Tripos Associates)
[8] Kalinowski H-O, Berger S, Braun S (1984) 13C-NMR-Spektroskopie, 1st edn. Thieme, Stuttgart
[9] Wimperis S, Bodenhausen G (1987) Chem Phys Lett 140: 41
[10] Wimperis S, Bodenhausen G (1989) Mol Phys 66: 897
[11] Jackman LM, Sternhell S (eds) (1969) Application of Nuclear Magnetic Resonance Spectros-
copy in Organic Chemistry, 2nd edn. Pergamon Press, Oxford, p 312

278
S. Felsinger et al.: Long-Range Couplings in Dimeric Indanones
[12] Bar®eld M, Chakrabarti B (1969) Chem Rev 69: 757
[13] Schaefer T (1996) In: Grant DM, Harris RK (eds) Encyclopedia of Nuclear Magnetic Resonance,
vol 7. Wiley, New York, p 4571
[14] Friebolin H (1992) Ein- und zweidimensionale NMR-Spektroskopie, 2nd edn. VCH, Weinheim
[15] Sternhell S (1964) Pure Appl Chem 14: 15
[16] Garbisch jr EW, Grif®th MG (1968) J Am Chem Soc 90: 3590
[17] Brown DW, Nakashima TT, Rabenstein DL (1981) J Magn Reson 45: 302
[18] Davies AL, Laue ED, Keeler J, Moskau D, Lohman J (1991) J Magn Reson 94: 637
[19] Shaw AA, Salaun C, Dauphin J-F, Ancian B (1996) J Magn Reson Ser A 120: 110
[20] Bax A, Freeman R (1981) J Magn Reson 44: 542
[21] Palmer III AG, Cavanagh J, Wright PE, Rance M (1991) J Magn Reson 93: 151
[22] Kay LE, Keifer P, Saarinen T (1992) J Am Chem Soc 114: 10663
[23] Schleucher J, Schwendinger M, Sattler M, Schmidt P, Schedletzky O, Glaser SJ, Sùrensen OW,
Griesinger C (1994) J Biomol NMR 4: 301
[24] Kontaxis G, Stonehouse J, Laue ED, Keeler J (1994) J Magn Reson Ser A 111: 70
[25] Ruiz-Cabello J, Vuister GW, Moonen CTW, van Gelderen P, Cohen JS, van Zijl PCM (1992) J
Magn Reson 100: 282
[26] Willker W, Leibfritz D, Kerssebaum R, Bermel W (1993) Magn Reson Chem 31: 287
[27] Wagner R, Berger S (1996) J Magn Reson Ser A 123: 119
Â
[28] Parella T, Sanchez-Ferrando F, Virgili A (1997) J Magn Reson 125: 145
È
[29] Kahlig H (private communication)
Received September 5, 2000. Accepted September 14, 2000