Multinuclear NMR studies and Ab initio structure calculations of hindered phencyclone Diels-Alder adducts from symmetrical cyclic dienophiles: Cyclohexene, vinylene carbonate, vinylene trithiocarbonate and two N-aryl maleimides

Article abstract of DOI:10.1002/jhet.5570420522

A series of hindered Diels-Alder adducts have been prepared from phencyclone, 1, with various unusual symmetrical cyclic dienophiles, including cyclohexene, 2a; vinylene carbonate, 2b; vinylene trithiocarbonate, 2c; and the N-aryl maleimides: N-(4-dimethy

Full text of DOI:10.1002/jhet.5570420522

Multinuclear NMR Studies and Ab Initio Structure Calculations of
Hindered Phencyclone Diels-Alder Adducts from Symmetrical Cyclic
Dienophiles: Cyclohexene, Vinylene Carbonate, Vinylene
Trithiocarbonate and Two N-Aryl Maleimides
Jul-Aug 2005
889
Ronald Callahan [a], Orlando Ramirez, III [b], Kerstin Rosmarion [b],
Robert Rothchild [b,c]*, Kevin C. Bynum [d]
[a] New York University, Chemistry Department, 29 Washington Place (569 Brown), New York, NY 10003
[b] The City University of New York, John Jay College of Criminal Justice, Science Department,
th
445 West 59 Street, New York, NY 10019-1128
[c] Doctoral Faculty, Graduate School and University Center, City University of New York
[d] Novartis Pharmaceuticals Corp., Bldg. 401, East Hanover, NJ 07936
Received February 8, 2005
A series of hindered Diels-Alder adducts have been prepared from phencyclone, 1, with various unusual
symmetrical cyclic dienophiles, including cyclohexene, 2a; vinylene carbonate, 2b; vinylene trithiocarbon-
ate, 2c; and the N-aryl maleimides: N-(4-dimethylamino-3,5-dinitrophenyl)maleimide ("Tuppy's
maleimide"), 2d; and N-[3,5-bis(trifluoromethyl)phenyl]maleimide, 2e. The highly hindered adducts, 3a-e,
respectively, were extensively characterized by one- and two-dimensional NMR methods, observing proton,
carbon-13 and fluorine-19. High resolution COSY45 spectra permitted rigorous proton NMR assignments.
The 2D heteronuclear C-H chemical shift correlation spectra (HETCOR, XHCORR) were obtained for
adducts 3a-d, allowing specific assignments for protonated carbons. Corrections to earlier proton NMR
13
assignments for the vinylene carbonate adduct are given; results of the gated decoupling C NMR experi-
ment for this adduct supported endo adduct stereochemistry. Relative proton chemical shifts for bridgehead
phenyls of adduct 3c appeared anomalous relative to other adducts, suggesting possible special anisotropic
interactions (with endocyclic sulfur or other anisotropic groups in the product) due to the unusual calculated
orientation of the phenyls. The unsubstituted bridgehead phenyls in all adducts were shown to exhibit slow
1
13
exchange limit (SEL) H and C spectra on the NMR timescales at ambient temperatures (7 tesla) showing
3
2
slow rotations about the C(sp )-C(aryl sp ) bonds. The rapid rotation of the N-aryl rings of the maleimide
adducts was indicated by fast exchange limit spectra, suggesting that ortho substitution of the N-aryl ring
may be necessary to slow this rotation to the SEL regime. Ab initio geometry optimizations at the Hartree-
Fock level were carried out for each adduct, with the 6-31G* basis sets. Appreciable geometry differences
were seen in calculated structures, and significant NMR chemical shift differences were experimentally
observed, depending on the nature of the groups attached to the (Z)-HC=CH moiety of the dienophiles.
J. Heterocyclic Chem., 42, 889 (2005).
Introduction.
The hindered adducts from the potent Diels-Alder diene,
phencyclone, 1 [1], have been of considerable interest for
NMR studies of hindered rotations and magnetic
anisotropic shielding effects, e.g., for the N-n-hexyl-
maleimide adduct [2]. For example, the unsubstituted
bridgehead phenyls in many Diels-Alder adducts of 1 have
3
been shown to exhibit slow rotations about the C(sp )-
2
C(aryl sp ) bonds on the NMR timescales based on slow
exchange limit (SEL) proton or carbon-13 spectra at ambi-
ent temperatures (at 7 and 4.7 tesla, i.e., 300 or 200 MHz
1
13
for H, 75 or 50 MHz for C). Selected references include
studies of the phencyclone adducts from N-n-propyl-
maleimide [3]; N-n-butylmaleimide [4]; N-carbamoyl-
maleimide [5]; N-pentafluorophenyl maleimide [6]; N-
phenylmaleimide and N-(2-trifluoromethylphenyl)-
maleimide [7]. Anisotropic effects on the bridgehead
phenyls of the adducts are suggested by dispersion in pro-
ton chemical shifts of 1 ppm or more, a rather wide range
for phenyl rings bearing no substituents or conjugated
Figure 1. Structures and atom numbering for key reagents, precursors and
possible side-products.
3
groups and connected only to a quaternary sp carbon.

890
R. Callahan, O. Ramirez, III, K. Rosmarion, R. Rothchild, K. C. Bynum
Vol. 42
Depending on the nature of the dienophiles, 2, used to form
the adducts, 3, rather dramatic anisotropic shielding effects
may be seen for some nuclei derived from the dienophiles,
positioned towards the phenanthrenoid moiety of the endo
adducts, and experiencing shielding attributable to the
phenanthrenoid rings, as with phencyclone adducts of N-
(2,6-dialkylphenyl]maleimides [8] and a series of N-n-
alkylmaleimides [9]. Structures and compound numbering
are shown in Figures 1 (for 1 and 2) and 2 (for the adducts,
3). Ab initio calculations at the Hartree-Fock level have
diverse assortment of symmetrical cyclic dienophiles in
conjunction with high level (ab initio) structure calcula-
tions for the adducts.
EXPERIMENTAL
General synthetic procedures and spectral methods followed
techniques described earlier [2-9]. Reagents were obtained from
Aldrich Chemical (Sigma-Aldrich, Milwaukee, WI) and used
directly without further purification. NMR spectra were routinely
acquired with a Bruker AC300F spectrometer with Aspect 3000
data system, with observe frequencies of ca. 300 MHz for proton
and 75 MHz for carbon-13. [Note that the proton NMR spectrum
of N-[3,5-bis(trifluoromethyl)phenyl]maleamic acid was
acquired at 400.13 MHz with a Bruker 400 AVANCE instru-
ment.] NMR chemical shifts were referenced to internal tetra-
1
methylsilane (TMS) at 0.0 ppm for H, to the central line of
13
CDCl at 77.0 ppm for C, and to internal CFCl at 0.0 ppm for
3
3
19
13
19
F. Standard 1D C and F NMR were obtained with compos-
ite pulse decoupling (WALTZ16) of protons. Normal relaxation
delays (RD) were 1 sec for proton, 3 sec for carbon-13, and 5 sec
for fluorine-19, unless noted. For the high resolution COSY45
spectra, the aryl proton region (typically ca. 700 Hz spectral
width from about 6.6 to 8.9 ppm) was examined with 256 incre-
ments in the t dimension, with 2 dummy scans and 16 transients
1
per increment. For the HETCOR (XHCORR) spectra, typically
256 scans were acquired for each of 128 increments in t . For
1
DEPT45 spectra, between 128 and 1024 scans were acquired.
See Table 1 for proton shifts of adducts 3a-e. Infrared spectra
were acquired on KBr pellets with a Perkin Elmer 1640 or Midac
-1
M2000 spectrophotometer with DTGS detector at 4 cm resolu-
tion; only major peaks are listed. Molecular modeling calcula-
tions were performed using Spartan '04 ESSENTIAL (v. 2.0.0,
Wavefunction, Inc. 1991-2003; Oct. 8, 2003) or Spartan '04 for
Windows (full version, v. 1.0.0, Wavefunction, Inc. 1991-2003;
Sept. 17, 2003) on Dell Pentium 4 platforms with 2.4 or 3.06
GHz processor speeds and 512 or 1024 MB memory. Replicate
calculations for several compounds were run using one or more
of the available software packages to confirm reproducibility or
locate the lowest energy optimized structures. Data presented
here reflect the lowest energy calculations, regardless of the soft-
ware or platform used. Calculation parameters included turning
symmetry off and convergence on.
Figure 2. Structures and atom numbering for phencyclone adducts.
Phencyclone Adduct, 3a, from Cyclohexene.
been carried out on the crowded adducts to gain insight into
expected preferred conformations. This present report
compares the phencyclone adducts from a diverse range of
symmetrical cyclic dienophiles, with varied substituents on
the (Z)-HC=CH bond of the dienophile. Here we examine
adducts from cyclohexene, 2a; vinylene carbonate, 2b;
vinylene trithiocarbonate, 2c; and the N-aryl maleimides:
N-(4-dimethylamino-3,5-dinitrophenyl)maleimide
("Tuppy's maleimide"), 2d; and N-[3,5-bis(trifluoro-
methyl)phenyl]maleimide, 2e. Geometry minimizations for
the corresponding adducts, 3a-e, were performed at the
Hartree-Fock level with the 6-31G* basis sets. We believe
that this is the first report describing the hindered bridge-
head phenyl rotations of phencyclone adducts from a
Cyclohexene was washed with aq. NaHSO (3%) and dried
3
(anh. Na SO ). Phencyclone (1.0705 g, 2.799 mmol), 2,6-di-
2
4
tert-butyl-4-methylphenol (BHT, ca. 3 mg) [2] and a magnetic
stirbar were combined in a round-bottom flask with the puri-
fied cyclohexene (16.2675 g, 198 mmol), capped and stirred at
ambient temperature for 3 weeks. Discharge of the intense
green-black phencyclone color had not completed, so the cap
was replaced by a water-cooled condenser, topped by a drying
tube, and the system was refluxed for 2 days, resulting in
decolorization of the mixture to a light yellow. Partial solvent
removal (rotary evaporator) produced abundant crystals, col-
lected and dried to give the crude adduct (693.3 mg, 53.1%
yield). This material contained considerable cyclohexene (by
proton NMR) and was recrystallized from CH Cl – ethanol to
give white crystalline flakes, used for NMR studies, with mpr
2
2

Jul-Aug 2005
Multinuclear NMR Studies and Ab Initio Structure Calculations
891
3a
3b
3d
3f
3c
3e
Figure 3. Calculated structures from Hartree-Fock/6-31G* geometry optimizations of the phencyclone adducts. (Selected hydrogens may be hidden for
clarity.) (a) Cyclohexene adduct 3a anti; (b) cyclohexene adduct 3a syn; (c) vinylene carbonate adduct 3b; (d) vinylene trithiocarbonate adduct 3c; (e)
Tuppy maleimide adduct 3d; (f) N-[3,5-bis(trifluoromethyl)phenyl]maleimide adduct 3e.
13
270-271° (dec., with darkening). C (CDCl ): 203.79 (ketone
126.08 C3,6; 126.03 C1,8; 123.23 C4,5; 64.18 C H -C; 38.63
3
6
5
3
C=O); 136.39 Q; 135.98 Q; 131.92 C6'; 130.60 Q; 128.88 C3';
128.56 C5'; 128.31 Q; 127.51 C4'; 126.62 C2'; 126.10 C2,7;
bridgehead sp (alpha); 20.03 CH (beta); 18.95 CH
2
2
(gamma).

892
R. Callahan, O. Ramirez, III, K. Rosmarion, R. Rothchild, K. C. Bynum
Vol. 42
Table 1
Proton NMR Shifts for Phencyclone Adducts of Cyclic Dienophiles, in ppm (with observed coupling constants in Hz) in CDCl . See Notes.
3
3a
3b
3c
3d
3e
Nucleus
H-1,8
H-2,7
H-3,6
H-4,5
H-2'
H-3'
H-4'
H-5'
7.08 (8.26,0.88)
7.21 (7.59)
7.54
8.74 (8.32)
7.68 (7.75)
7.58 (7.38)
7.42 (7.20)
7.35 (7.51,1.38)
7.13 (7.72)
3.31 m
7.00 (8.34)
7.20 (7.55)
7.56 (7.52)
8.70 (8.37)
7.84 (7.73)
7.70 (7.52,0.95)
7.56 (7.52)
7.47 (7.54,1.16)
7.29 (7.76)
6.08
6.91 (7.93)
7.21 (7.01)
7.57 (7.73,1.16)
8.73 (8.28)
7.37 (7.70)
7.65 (7.58,1.16)
7.53
7.45 (7.56,1.19)
7.22 (7.01)
6.01
7.15 (8.39)
7.24 (7.65)
7.57 (7.05,1.1)
8.67 (8.45)
8.27 (7.85)
7.73 (7.57)
7.55 (7.34)
7.47 (7.56)
7.30 (7.81)
4.61
7.16 (8.30)
7.24 (7.30)
7.56 (7.34)
8.66 (8.42)
8.29 (7.84)
7.73 (7.58)
7.56 (7.34)
7.47 (7.57)
7.31 (7.82)
4.66
H-6'
Bridgehead
H(a) beta
H(b), H(c)
H(d) beta
H2",6"
H4"
1.86 d (14.25)
1.52 br s
0.73 m
6.30
6.33
7.54
N(Me)
2.62
2
3
Notes: Approximate observed coupling constants are shown in parentheses, and refer usually to the gross vicinal ( J) couplings. Fine splitting
from long range couplings have been shown when seen for aryl proton signals. Gross multiplicities for the aryl protons (d or t) were consistent
with the numbers of vicinal proton neighbors (one or two, respectively). The large coupling for H(a) of 3a is a geminal coupling.
Table 2
Calculated energies (atomic units) and geometric parameters for Diels-Alder adducts,
including selected distances (angstroms) and dihedral angles (degrees). See Notes (below) and Discussion.
Energies
Dihedral angles
O=C-C-C1'-C6'
Interatomic distances
H2'….
Phenanthrene
pucker
Adduct
H6'…ketone C=O
H6'…H1
H2'…H1
bridgehead CH
3a anti
∆E, kcal/mol
-1416.5813400
6.39
59.41, -59.36
2.915, 2.914
3.218, 3.219
3.218, 3.219
3.333, 3.332
0.281
3a syn
3b
3c
3d
3e
-1416.5711500
-1522.9257163
-2490.8310070
-2310.5204718
-2441.7687126
58.92, -58.99
48.99, -49.01
55.88, -55.84
51.50, -53.07
52.38, -52.70
2.896, 2.898
2.669, 2.669
2.812, 2.811
2.707, 2.750
2.730, 2.740
3.210, 3.209
2.257, 2.257
2.251, 2.252
2.277, 2.274
2.272, 2.269
3.210, 3.209
3.444, 3.444
3.289, 3.291
3.406, 3.423
3.237, 3.395
3.351, 3.353
2.998, 2.998
3.235, 3.234
3.065, 3.069
3.079, 3.080
0.287
0.248
0.257
0.286
0.268
Notes: Where pairs of values of a parameter are given, replicate measurements were obtained from corresponding nuclei on each "half" of the molecule.
See Figures for atom numbering. Signs for dihedral angles depend on local chirality. Phenanthrenoid pucker is the distance from the midpoint of the line
joining H2 and H7, to the midpoint of the C4a-C4b bond. (The midpoint of the line joining H2 and H7 was closer to the dienophile-derived group.).
Phencyclone Adduct, 3b, from Vinylene Carbonate.
per Harrison [10] with ca. 15 hr reflux in xylene [with substantial
decolorization of the phencyclone in ca. 2 hr] and had mpr 275-
277° (dec., gas evoln.). Lit. 264-266° (dec.) [10] (See
Discussion.) IR (KBr): 1816 vs, carbonate; 1795 vs, strained
In a 50 mL round bottom flask was placed vinylene carbonate
(877.6 mg, 10.199 mmol), 25 ml of toluene, phencyclone (3.607
g, 9.432 mmol), and a magnetic stirbar. The flask was fitted with
a water-cooled condenser topped with a drying tube (anhyd.
ketone C=O; 1498 m; 1448 m; 1360 m; 1148 s; 1089 s; 962 w;
13
936 w; 756 s; 724 m; 698 s. C (CDCl ): 194.39 (ketone C=O);
3
CaCl ). The reaction mixture was initially dark green-black,
2
153.52 (carbonate); 132.87 Q; 132.37 Q; 131.50 Q; 130.92 C6';
129.54 C3'; 129.28 C5'; 128.87 C4' or 3,6; 127.51 C2'; 127.26
C4' or 3,6; 126.96 Q; 126.71 C2,7; 125.17 C1,8; 123.42 C4,5;
becoming light orange-red after ca. 20 hr reflux with stirring.
Partial solvent removal and cooling to room temperature gave
white crystals, collected by vacuum filtration and dried (4.646 g,
105% crude yield). Preliminary proton NMR showed large
amounts of toluene present. Toluene removal was achieved by
treating an aliquot of the crude adduct (280.1 mg) with 10 ml por-
77.90 (bridgehead CH); 66.27 C H -C. (See Discussion for
6
5
13
results of the gated decoupling C NMR spectrum.).
Phencyclone Adduct, 3c, from Vinylene Trithiocarbonate.
tions of CH Cl and removing solvent (rotary evaporator).
2
2
As in the preparation of adduct 3b, above, vinylene trithiocar-
bonate (647.9 mg, 4.826 mmol) in 25 mL toluene with phency-
clone (1.7253 g, 4.51 mmol) was refluxed with stirring for 36 hr.
Essentially toluene-free adduct (117.9 mg) was thus obtained and
used for NMR studies. Mpr 243-245° (vigorous gas evolution
and darkening). A separate batch of this adduct was prepared as

Jul-Aug 2005
Multinuclear NMR Studies and Ab Initio Structure Calculations
893
Table 3
Additional calculated geometric parameters for adducts. Distances are in angstroms; angles and dihedral
angles are in degrees. See Discussion and Notes from preceding Table 2 and below.
Distances from Pt. 1 (center of middle ring of phenanthrenoid) to:
Distances from H2' to:
Adduct
H(a) out
H(d) in
H(c) out
H(b) in
H(c) out
H(d) in
H(a) out
3a anti
4.477, 4.475
5.107, 5.108
3.193, 3.191
5.906, 5.906
5.528, 5.525
3.654,3.653
H(b) in
4.408,4.408
3.677,3.677
2.280,2.280
3a syn
4.413, 4.414
4.591, 4.591
2.965, 2.965
2.271,2.272
3.152,3.152
Angles from Point 2 to Point 1 to specified nucleus:
Dihedral angles:
C(Q)CHCH CH (γ)
O=CC(Q)CHCH (β)
2
173.33, -173.34
78.47, -78.45
2
2
3a anti
3a syn
54.96, 54.99
46.14, 46.14
63.83, 63.84
53.67, 53.66
47.17, 47.21
73.15, 73.14
64.14, 64.18
77.83, 77.81
-158.63, 158.57
-162.13,162.16
H2' to -X-
2.767, 2.768
2.890, 2.890
H1 to -X-
3.606, 3.606
3.616, 3.614
CX to Pt. 1
3.694
4.153
Pt.2-Pt.1-CX
57.90
62.26
CX Planarity
0.006
0.004
Ring Planarity
0.005, 0.005
0.000, 0.000
3
3
3
3b
3c
Dihedral angle
O=CNC1"C6"(in)
128.09, -50.58
119.42, -59.48
Pt.1 - H6" (in)
3.809
3.646
Pt.2-Pt.1-H6"
95.40
97.32
H2' - NC=O
2.562, 2.573
2.575, 2.581
H1 - NC=O
3.305, 3.214
3.294, 3.255
N planarity
0.009 N out
0.010 N out
3d
3e
Notes: Point 1 is the midpoint of the line joining C8a and C10a, and is the approximate center of the central ring of the phenanthrenoid moiety.
3
Point 2 is the midpoint of the C9-C10 bond. C(Q) is the quaternary sp carbon, C H -C. CX Planarity for adducts 3b and 3c refers to the -X-
6
5
3
C(=X)-X- groups, expressed as the distance of the CX carbon from the plane defined by the three attached heteroatoms. Ring Planarity for
3
adducts 3b and 3c refers to the distances of the endocyclic heteroatoms to the plane defined by the three carbons in the carbonate or trithiocar-
bonate ring. N planarity for the adducts 3d and 3e refers to the distance of the nitrogen from the plane defined by the three attached atoms. For
the latter two adducts, the N atom was on the side of the plane away from the adduct cavity and phenanthrenoid nucleus.
The reaction mixture, although still somewhat green, was cooled
to room temperature. Pale yellow-green precipitate was collected
bridgehead; 41.97 NMe . Traces of unreacted Tuppy maleimide
were suggested by weak peaks at 134.55 and 42.24 ppm, consis-
tent with reported signals [11] at 134.50 (HC=CH or aryl C2,6)
2
by vacuum filtration, washed with chilled CH Cl and dried to
2
2
give 1.126 g of light yellow adduct (48.7% yield) for spectral
studies, mpr 269-279° (vigorous gas evolution and darkening).
IR (KBr): 1784 vs, strained ketone C=O; 1497 m; 1446 s; 1437
m; 1243 br. m; 1071 vs, trithiocarbonate; 902 w; 886 m; 853 w;
and 42.19 ppm (NMe ).
2
N-[3,5-Bis(trifluoromethyl)phenyl]maleamic Acid.
This maleamic acid was prepared in the normal way [2,19] by
reaction of maleic anhydride with equimolar 3,5-bis(trifluo-
13
784 m; 751 s; 735 m; 723 m; 693 s. C (CDCl ) (RD=60 sec,
3
NS=5441): 219.95 (trithiocarbonate); 196.03 (ketone C=O);
135.23 Q; 133.42 Q; 131.39 Q; 131.26 C2,7 or 6'; 129.51 C3';
129.30 C5'; 128.77 C4'; 127.43 Q; 127.03 C3,6; 126.92 C2';
126.54 C2,7 or 6'; 125.83 C1,8; 123.43 C4,5; 66.85 C H -C;
romethyl)aniline in warm CH Cl , ca. 5-10 mmol scale, in 40-
2
2
60% yields. Crude solid was obtained on trituration and standing,
collected by filtration and washed with cold CH Cl . The dried
2
2
1
white crystals had mpr 144-150°, lit. 163-165° [12]. H (CDCl ):
6
5
3
3
63.92 (bridgehead sp CH).
9.29 (1H, br s, NH); 8.12 (2H, sl br s, H2,6); 7.73 (1H, sl br s,
H4); 6.52 (1H, d, J=12.80, half of AB q, cis HC=CH); 6.46 (1H,
d, J=12.80, half of AB q, cis HC=CH). This sample was directly
converted to the corresponding maleimide.
Phencyclone Adduct, 3d, from N-(4-Dimethylamino-3,5-dinitro-
phenyl)maleimide ["Tuppy's maleimide"].
In a screw-cap vial (ca. 20 mL capacity) was placed a mag-
netic stirbar, phencyclone (383.8 mg, 1.001 mmol), Tuppy's
maleimide (321.8 mg, 1.0508 mmol), BHT (ca. 3 mg) [2] and
sufficient CH Cl to bring the mixture level to within 3-4 mm of
N-[3,5-Bis(trifluoromethyl)phenyl]maleimide.
Prepared by standard methods [13,14], by cyclodehydration of
the precursor N-[3,5-bis(trifluoromethyl)phenyl]maleamic acid
in acetic anhydride with anhydrous sodium acetate at 90-100° for
ca. 1.5 hr. The reaction mixture had become deep reddish-black.
After cooling and quenching with ice-water, the mixture was par-
2
2
the vial lip. After capping with a Teflon©-lined cap, the mixture
was stirred at ambient temperature, with the initial intense green-
black color becoming brown-orange in 5 min and yellow-orange
after 15 min. After a total 60 min stirring, the light yellow-orange
mixture gave 383 mg of orange solid upon solvent removal,
washing with cold CH Cl and drying (55.6% yield), mpr (with
titioned into CH Cl , extensively washed with 5% aqueous
2
2
NaHCO and water. The organic layer was separated and dried
3
(anh. Na SO ), and solvent removal gave the crude solid
2
2
2
4
slow heating): early darkening ca. 270°, then dec. with gas evolu-
maleimide, 2e, in ca. 20-30% yield. Recrystallization from EtOH
13
1
tion 280-286°, sufficiently pure for NMR studies. C (CDCl ):
gave white crystals, mpr 77-81°. H (CDCl ): 7.95 (2H, sl. br s,
3
3
19
196.64 ketone C=O; 172.64 imide NC=O; 144.05 Q; 139.29 Q;
133.28 Q; 133.17 Q; 131.14 Q; 130.86 C6'; 129.49 C3'; 128.87
C5'; 128.78 C2'; 128.71 C4'; 127.68 C3,6; 127.12 C2,7; 127.03
C2",6"; 126.04 Q; 125.50 C1,8; 123.38 C4,5; 121.63 Q; 109.35
H2,6); 7.87 (1H, sl br s, H4); 6.95 (2H, s, HC=CH). F (CDCl ):
3
-63.42 s. Impurity ca. 28% acetoxypyrrolidinedione: 5.55 (1H,
3
2
dd, J=9.06 and 5.06 Hz, AcOCH, H ); 3.39 (1H, dd, J=18.64,
A
3
2
3
J=9.06, H ); 2.96 (1H, dd, J=18.64, J=5.06, H ); 2.22 (3H, s,
M
X
3
weak, unassigned impurity?; 63.59 C H -C; 44.86 sp CH
CH CO). See Discussion.
6
5
3

894
R. Callahan, O. Ramirez, III, K. Rosmarion, R. Rothchild, K. C. Bynum
Vol. 42
Phencyclone Adduct, 3e, from N-[3,5-Bis(trifluoromethyl)-
phencyclone analog, 3,6-dibromophencyclone and its adduct
phenyl]maleimide.
from Tuppy's maleimide. Literature values for chemical shifts of
1
cyclohexene (in CDCl ) were: H: 5.67 (HC=CH); 1.99 (allylic
3
This adduct has mpr 267-269° (dec. with vigorous gas evolu-
tion and darkening). C (CDCl , 66,126 scans): 196.59 (ketone
C=O); 172.64 (NC=O); 133.31 Q; 133.20 Q; 133.24 (q, J=34.2,
C3",5"); 132.14 Q (sl. br., NC1"); 131.17 Q; 130.88 (2 CH);
129.48 (2 CH); 128.85 (2 CH); 128.78 (2 CH); 128.69 (2 CH);
13
CH -C=C); 1.61 ppm (CH -CH C=C). C: 127.21 (HC=CH);
2
2
2
13
3
25.21 and 22.70 ppm [11]. For the adduct, 3a, from phencyclone
with cyclohexene, the bridgehead methine (alpha) protons res-
onated at 3.31 ppm as a symmetrical multiplet (six lines), the
2
diastereotopic (beta) CH protons at 1.86 (H in Figure 2) and
2
a
3
127.56 (2 CH); 127.02 (2 CH); 126.16 (br. q, J=3.15, C2",6");
0.73 ppm (H in Figure 2), and the diastereotopic (gamma) pro-
d
126.04 Q; 125.57 (2 CH); 123.23 (2 CH); 122.26 (app. quintet,
tons (H and H in Figure 2) appeared as a 4H broad singlet at
b
c
3
1
J=3.73, C4"); 122.19 (q, J=273.2, CF ); 63.62 C H -C; 44.90
3
6 5
1.52 ppm (overlapped with the HOD signal). Adduct assignments
were based on the COSY45 spectrum, in which the bridgehead
methine showed strong crosspeaks to both the 1.86 and 0.73 ppm
signals, but not to the 1.52 ppm signal. The 1.52 ppm signal cor-
related only with the 1.86 and 0.73 ppm signals, and the 1.86 and
0.73 ppm signals were strongly correlated to each other, consis-
3
19
(2 sp CH bridgehead). F (CDCl ): -63.75 s.
3
Results and Discussion.
The dienophiles 2a-d were directly available commercially;
phencyclone was prepared as described earlier [2]. Dienophile 2e
was prepared with standard methods (Figure 1) by reaction of
maleic anhydride with 3,5-bis(trifluoromethyl)aniline, 4e, to
form the N-[3,5-bis(trifluoromethyl)phenyl]maleamic acid, 5e.
The intermediate 5e underwent cyclodehydration in modest yield
(see below) to give the corresponding maleimide, 2e.
2
tent with their large J(geminal) coupling. Based on these assign-
ments, we believe that the 0.73 ppm signal may be attributed to
the syn protons, H (Figure 2) directed "into" the adduct cavity
d
and toward the phenanthrenoid moiety, where it could experience
an anisotropic shielding magnitude of ca. 1.1 ppm relative to the
For each of the phencyclone adducts described here, the basic
question of bridgehead phenyl rotation rates was addressed based
13
anti geminal neighbors, H . In the C spectrum of this adduct,
a
no less than three partly overlapped signals for aryl CH pairs
were seen between 126.03-126.10 ppm, not baseline resolved,
but assignments could be made from the HETCOR spectrum.
HETCOR CH crosspeaks were seen for the bridgehead (alpha)
methines and for the (gamma) methylenes [with the 1.52 ppm
proton signal]. We could not observe crosspeaks from the (beta)
proton signals at 1.86 or 0.73 ppm to the carbon signal at 20.03
ppm, which we ascribe to signal-to-noise consequences. We note
1
13
on the H and C NMR spectra. Slow exchange limit (SEL)
spectra, implying slow rotations of these unsubstituted bridge-
head phenyls on the NMR timescales, should result in observa-
tion of nine signals (in the absence of coincidental overlaps) from
×
pairs of aryl methines, 2 CH. Four of the pairs would derive
from the phenanthrenoid nucleus, 1,8; 2,7; 3,6; and 4,5, and five
pairs would result from the bridgehead phenyls, positions 2', 3',
4', 5', and 6'. (See Figure 2.) Fast rotation of the bridgehead
phenyls would give fast exchange limit (FEL) spectra due to
exchanging and averaging of the 2',6' pairs and the 3',5' pairs,
resulting in just three expected CH signals from these phenyls,
with 4:4:2 relative intensities.
1
13
that the above H and C NMR assignments for adduct 3a
reflect the major spectral peaks; even after recrystallizations from
CH Cl /ethanol, there appeared to be numerous smaller peaks,
2
2
perhaps suggesting ca. 10% of a minor component. The other
component may have been the stereoisomeric exo Diels-Alder
adduct, but we have not attempted to rigorously define this.
Reversibility of the Diels-Alder reaction with 1 is known for
some dienophiles, and this could account for some formation of
With adduct 3a, a simple tetramethylene chain, (CH ) , links
2 4
the two bridgehead methines (labeled "alpha" in Figure 2). For
adducts 3b and 3c, the two bridgehead methines of the adducts
are directly attached to heteroatoms, i.e., -X-C(=X)-X-, where X
= O or S. In the final set of adducts, 3d and 3e, the "alpha"
bridgehead methines are bonded to carbonyls of the imide group,
-C(=O)-N(Aryl)-C(=O)-. Thus, these selected adducts provide
examples of a wide range of phencyclone adducts.
The initially obtained crude adduct, 3a, from phencyclone
with cyclohexene, contained considerable cyclohexene (by H
NMR, ca. equimolar with adduct), consistent with the reported
propensity for some phencyclone adducts to form solvates [15-
17]. We have sometimes removed complexed solvent from phen-
1
the exo isomer of adduct [1]. Cyclohexene adduct 3a exhibits H
13
and C NMR spectra supporting slow exchange limit (SEL)
spectra, implying slow rotation of the bridgehead phenyls on the
NMR timescales. The suggested preferred conformation of the
cyclohexyl ring in the endo adduct is further discussed below.
Preparation and characterization of the phencyclone adduct of
vinylene carbonate was of particular interest to us. In 1975,
Harrison had reported that the reactions of various substituted
cyclopentadienones with 2b in refluxing bromobenzene for 24 hr
yielded substituted phenols containing no carbonyl, rather than
the normal Diels-Alder adducts. Phencyclone itself, however,
was said to yield a carbonyl-containing compound rather than a
phenol, but thermolysis of the normal adduct, 3b, (Figure 2) gave
a moderate yield of phenolic product [20]. [No data were given
for adduct 3b.] The intermediacy of 3b (from reaction with 1)
1
cyclone adducts simply by repeated co-distillations with CH Cl
2
2
or CHCl , with a final co-distillation using CDCl if desired to
3
3
avoid [proton] NMR peaks of the proton-bearing solvents. It
should be noted that Harano and co-workers [15-17] have pub-
lished X-ray crystal structures for several of the phencyclone
adducts, and their results provide strong support for the structures
of the analogs we present here. Addition of a trace of BHT, a
free-radical trap, in these preparations is expected to potentially
suppress undesired free-radical reactions, such as free-radical
additions to the maleimides, free-radical polymerizations of the
maleimides, or a free-radical process of oxidative decarbonyla-
tion from phencyclone (leading to 9,10-dibenzoylphenanthrene
byproduct) [2,18,19]. Reference [19] describes work with the
-1
was postulated, based on IR peaks at ca. 1770 and 1810 cm
[phase not stated] with subsequent decarbonylation and decar-
boxylation from the adduct to give the observed substituted phe-
nol product. The high reaction temperature (bromobenzene boils
at ca. 156°) and long reaction time were possible factors in the
formation of the reported products. In a later report, Harrison and
Ammon [10] elegantly clarified and extended the earlier findings.
Refluxing 1 and 2b in bromobenzene for 48 hrs produced the

Jul-Aug 2005
Multinuclear NMR Studies and Ab Initio Structure Calculations
895
isolated rearranged cyclic carbonate, 3b', (see Figure 1) as a
result of decarbonylation from the intermediate adduct, 3b, fol-
lowed by a concerted [1,5] sigmatropic shift of carbonate. Under
milder reaction conditions of reflux in xylene (bp ca. 140°) for 15
hrs, the authentic normal Diels-Alder adduct, 3b, was obtained in
good yield. The authors' published data [10] showed IR bands
With our vinylene carbonate adduct, 3b, we employed the
method of Warrener, et al. [22], to establish endo stereochemistry
13
in the Diels-Alder adduct. Use of the "gated decoupling"
C
NMR experiment, with broadband proton decoupling gated off
during the FID acquisition, allows development of nuclear
Overhauser enhancement in a fully proton-coupled carbon spec-
trum. The key signal is that of the strained bridging ketone car-
bonyl, C=O. In the normally expected endo Diels-Alder adduct,
the bridgehead methine hydrogens are directed "upwards," away
from the adduct cavity and the phenanthrenoid moiety, and
towards the ketone C=O. The dihedral angle magnitudes of the
bridgehead methine hydrogens relative to the ketone carbonyl, H-
C-C-C=O, would be ca. 88.6° based on the ab initio HF/6-31G*
calculations for 3b (described below). For dihedral angles near
-1
(paraffin oil) at 1815 and 1795 cm and mpr 264-266° (dec).
These IR bands are close to the values that we observed (in a dif-
ferent phase), and could be consistent with cyclic carbonate and
strained bridging carbonyl group of the Diels-Alder adduct.
Published IR spectra for vinylene carbonate (neat) showed bands
at 1831.2 (very strong) and 1777.0 (shoulder, med.-strong); eth-
ylene carbonate (as a melt) showed bands at 1861.2 (shoulder,
med.), 1802.1 (very strong) and ca. 1775 (strong) [21]. The dif-
ference in our observed mpr (decomposition ranges) versus that
reported earlier [10] may be attributed, in part, to differences in
heating rates, because many phencyclone Diels-Alder adducts
characteristically undergo decomposition at elevated tempera-
tures, as by decarbonylation of the bridging ketone carbonyl. We
also note that when we prepared a separate batch of the vinylene
carbonate adduct following the Harrison procedure [10] with ca.
15 hr reflux in xylene, this sample displayed proton and carbon-
13 NMR spectra essentially identical with the material obtained
in refluxing toluene. But our product from xylene exhibited a
decomposition range of 275-277°, closer to Harrison's value. It is
possible that some difference in observed decomposition range
may reflect the specific solvent used for the synthesis (and poten-
tially complexed with the adduct sample), e.g., toluene versus
xylene.
3
90°, the vicinal J CCCH coupling should be very small, accord-
ing to the Karplus relationship. For the hypothetical exo Diels-
Alder adduct stereochemistry, the corresponding dihedral angle
magnitudes from the bridgehead methine hydrogens to the ketone
carbonyl were ca. 159.1° (semi-empirical AM-1 calculation).
Warrener has reported that in such exo adducts, the observed vic-
3
inal J CCCH coupling was ca. 5-9 Hz, with the ketone carbonyl
signal appearing as a distinct triplet, in clear contrast to the endo
adducts, which exhibited an unsplit singlet for the ketone C=O
signal. Our sample of 3b gave a clean, unsplit singlet, w ca. 2.3
1/2
Hz or less, at 194.4 ppm for the ketone C=O in the gated decou-
13
pling C NMR spectrum, fully consistent with the assigned endo
stereochemistry.
Adduct 3c from vinylene trithiocarbonate exhibited anomalous
proton chemical shifts for the bridgehead phenyls, with the H2'
resonance at surprisingly high field. For the other adducts, the
shifts of the bridgehead phenyl protons appear at progressively
higher field (toward TMS) in the sequence H2', to H3', to H4', to
H5', to H6' (most shielded). In contrast, 3c exhibited these shifts
in the sequence H3', to H4', to H5', to H2', to H6' (most shielded).
Our initial interpretation was that this might reflect the direct
effect of the trithiocarbonate sulfur atoms with a proximal C2',
perhaps as a consequence of the large sulfur atoms and their
unshared electron pairs. We return to this in our Discussion
below regarding calculated structures. The reported carbon-13
1
Our greatest concern was the reported H NMR data from
Harrison and Ammon, which indicated an 18H aromatic multi-
plet from 7.05-7.83 ppm (plus a 2H singlet at 6.05 ppm for the
bridgehead methines). We have found in our own earlier studies
that the phencyclone adducts consistently display a very distinc-
tive low-field 2H doublet at, e.g., ca. 8.7 ppm, assigned to the
phenanthrenoid H4,5 protons. (This signal may be considered as
diagnostic for the phencyclone adducts that we have examined.)
Thus, the NMR data for 3b published earlier seemed question-
1
able. But, perhaps most important to us, the earlier H NMR
work was carried out on a low field spectrometer with 60 MHz
observe frequency, under which conditions there is inadequate
dispersion to observe and assign the different adduct aryl proton
signals, so that the question of potential SEL spectra from slow
rotations of the bridgehead phenyls simply cannot be investi-
gated. Together with the other aspects of the earlier work regard-
ing 3b, we were especially interested in rigorous NMR examina-
tion of the compound.
NMR shifts for reagent vinylene trithiocarbonate (CDCl )
3
showed the C=S at 213.13 ppm and the HC=CH at 129.06 ppm.
For ethylene trithiocarbonate, the C=S resonated at 228.52 ppm
[11]. For the adduct, the C=S peak was extremely weak and was
convincingly observed only with the long 60 sec relaxation
delay; it appeared at 219.95 ppm, almost 7 ppm further downfield
than in the reagent 2c, but not as deshielded as in ethylene trithio-
carbonate. Apparently, the shift of the C=S group is very sensi-
tive to surrounding structure, much more so than the C=O of the
analogous carbonate systems. The strong observed IR band
Our results for the proton spectra of 3b confirmed our earlier
findings for the highly deshielded H4,5 resonance. With the
available dispersion of the 300 MHz spectrometer, the aryl pro-
ton region is quite simplified and is consistent with the highly
symmetrical [near-mirror symmetry] Diels-Alder adduct struc-
ture of 3b, but not consistent with the less symmetrical structure
of the decarbonylated and rearranged cyclic carbonate, 3b', iso-
lated after bromobenzene reflux by Harrison and Ammon. For
-1
(KBr) for our adduct 3c at 1071 cm is consistent with the strong
-1
band at 1063.5 cm for ethylene trithiocarbonate (melt) [21].
The phencyclone adduct, 3d, from N-(4-dimethylamino-3,5-
1
dinitrophenyl)maleimide ["Tuppy's maleimide"] gave a H NMR
spectrum with a sharp 2H singlet at 6.30 ppm for the ortho
H2",6" protons of the N-aryl ring, indicating fast rotation of the
N-aryl. Relative to the reported chemical shift of 7.97 ppm for
these protons in the maleimide itself [11], this suggests a mag-
netic anisotropic shielding magnitude in the adduct of ca. 1.67
ppm as a result of H2",6" adduct protons spending 50% of their
time in the adduct cavity, towards the phenanthrenoid moiety.
This "inside," or syn, orientation would correspond to the edge
13
the vinylene carbonate adduct, 3b, C NMR showed that the car-
bonate C=O chemical shift of 153.52 ppm was virtually identical
to the reported value of 153.42 ppm for vinylene carbonate and
close to the value of 155.58 ppm for ethylene carbonate [11].
1
13
Both H and C NMR of the adduct implied SEL spectra, con-
sistent with slow rotation of the bridgehead phenyls.

896
R. Callahan, O. Ramirez, III, K. Rosmarion, R. Rothchild, K. C. Bynum
Vol. 42
(of the N-aryl ring) labeled (b) in the structure of 3d shown in
reflects predominant vicinal couplings to the CF fluorines. The
3
Figure 2. (Since the N-aryl rotation of a nominal 180° about the
sharper approximate quintet assigned to C4" is consistent with
2
2
N(sp )-C(aryl sp ) bond would constitute a twofold degenerate
interconversion, each conformer would equally contribute a 50%
share to the equilibrium system.) Presumably, if one of these
ortho protons spent 100% of its time in the adduct cavity, syn
with respect to the phenanthrenoid system, the proton might
exhibit an anisotropic shielding magnitude of ca. 3.5 ppm, and
would resonate near 4.5 ppm. This might be investigated experi-
mentally by low-temperature NMR spectra, if the N-aryl rotation
could be "frozen out" to provide SEL N-aryl spectra. The postu-
lated 4.5 ppm shift would be in excellent agreement with the
splitting by an even number of vicinal fluorines (i.e., the six fluo-
19
rines of the 3,5-bis-(CF ) groups). The F NMR with proton
3
decoupling shows one sharp singlet (w
ca. 1.4 Hz). Together
1/2
with the proton NMR results, we conclude that N-aryl rotation is
fast on the proton, carbon-13 and fluorine-19 NMR timescales.
Bridgehead phenyl rotation is slow. Relative to the H2,6 proton
resonances of the precursor maleimide, which resonate at 7.95
ppm, the H2",6" signal of the adduct appears at 6.33 ppm, imply-
ing an anisotropic shielding magnitude of ca. 1.6 ppm for the pro-
tons as a consequence of their spending half their time directed
"into" the adduct cavity, toward the phenanthrenoid moiety. This
is consistent with results described above for adduct 3d and ear-
1
observed H NMR spectrum (ambient temperature, 300 MHz)
for the H-6" proton of the N-aryl ring in the phencyclone adduct
from N-(2-trifluoromethylphenylmaleimide [7]. The C-13 spec-
trum of the adduct, 3d, exhibited the ten predicted intense signals
for the aryl CH pairs, as expected for a slow exchange limit sys-
tem (of the two bridgehead phenyls) with fast exchange limit for
the N-aryl ring. The HETCOR spectrum allowed assignment of
all protonated aryl carbon signals, despite having three closely
spaced signals between 128.71 – 128.87 ppm and a close pair at
127.03 and 127.12 ppm. The highest-field carbon-13 signal for
protonated aryl (C4,5) correlates with the lowest field proton sig-
19
lier data [7]. The change in chemical shifts from the F NMR for
the maleimide, 2e to adduct, 3e, were minimal, ca. 0.3 ppm.
1
For all five of the adducts described here, the relative H NMR
shifts of Table 1 seem quite striking with respect to the consis-
tency of the shifts from adduct to adduct for some of the protons
of the phenanthrenoid and bridgehead phenyls, and the much
larger variations for other protons. The absorption for H2' of the
phenyls is most unusual. This signal appears as far upfield as 7.37
ppm for 3c (considered anomalously high field relative to all the
other adducts), and as far downfield as 8.29 ppm for 3e, a range
of 0.92 ppm. The next largest range was 0.25 ppm for H1,8. The
ranges for the other protons were as follows (in ppm): H2,7
(0.04); H3,6 (0.03); H4,5 (0.08); H3' (0.15); H4' (0.14); H5'
(0.12); H6' (0.18). It is tempting to consider that the protons H2'
and H1,8 with the largest ranges might be [spatially] closest to
13
nal (H4,5). Of the seven C signals attributed to adduct non-pro-
tonated aryl carbons (Q), five were of relatively greater area (19.5
to 28.1 area units) and two were much weaker (5.9 and 9.3 area
units) [relative to average areas for each of the aryl methine pairs
of 59.6 area units, with RD=3 sec]. Of the seven expected Q sig-
nals, five would be attributed to pairs of carbons and two to indi-
vidual carbons (ipso C1" and para C4"). Since the N-aryl ring is
itself rotating rapidly, we cannot draw conclusions regarding the
3
the varying substituents on the (alpha) sp bridgehead methines,
and therefore most susceptible to possible anisotropic effects.
Alternatively, the generally larger shift ranges for the bridgehead
phenyl protons (relative to, e.g., the phenanthrenoid H2,7; H3,6;
H4,5), may reflect varying conformations of the phenyls with
respect to their dihedral angles relative to the ketone carbonyl. To
further explore possible adduct structures, we carried out geome-
try optimizations for 3a-e.
rotation rate of the dimethylamino group about the bond to the N-
2
aryl ring, i.e., the Me N-C(aryl sp ) bond.
2
Published proton and carbon-13 NMR data for 2e have
appeared, for d -DMSO solutions [14], so our present data in
6
CDCl are complementary. Laschewsky and co-workers had
3
19
reported the F NMR absorption for a CDCl solution at -60.55
3
ppm, somewhat different from our observed value of -63.42 ppm,
but fluorine chemical shifts can be temperature- and concentra-
tion-dependent. Our observed proton NMR of N-[3,5-bis(trifluo-
romethyl)phenyl]maleimide, 2e, exhibited slightly broadened
singlet signals for the aryl protons, consistent with small unre-
solved long-range H-H or H-F couplings. This crude maleimide
showed proton absorptions consistent with substantial amounts
of the (undesired) acetoxypyrrolidinedione side-product, 6
(Figure 1, Nu = OAc) [23], with the three 1H dd patterns of the
AMX system and the 3H singlet of the acetate group. The aryl
protons of the acetoxypyrrolidinedione partly overlapped with
the resonances of the target maleimide. In our hands, the tradi-
tional method for maleimide formation by cyclodehydration of
the precursor maleamic acid using anhydrous sodium acetate in
acetic anhydride (at 90-100°) gave cleaner (i.e., solid) crude
product than the alternative method of Wang [23], which initially
gave an oil.
Ab Initio Calculations.
For each of the phencyclone adducts, ab initio structure calcu-
lations were performed at the Hartree-Fock level with the 6-31G*
basis set. This method is considered to provide reasonable accu-
racy for energies and conformer structures with respect to
required calculation times [24-27]. Selected geometric parame-
ters from these optimized structures are presented in Tables 2 and
3. Several calculated geometric parameters are defined here and
evaluated for characterization of the adduct structures. The center
of the middle ring of the phenanthrenoid moiety, Point 1, is
defined as the midpoint of the line joining C8a and C10a. Point 2
is the midpoint of the C9-C10 bond. Distances (in Angstroms)
from selected nuclei to Point 1 are provided, as well as some
angles from Point 2 to Point 1 to specified nuclei. "Puckering" of
the phenanthrenoid moiety refers to the butterfly-like folding of
this aryl system, with the outer rings ("wings") folded towards
the portion of the adduct derived from the dienophile. Pucker
magnitude is expressed as the distance between the midpoint of
the line joining H2 and H7, and the midpoint of the C4a-C4b
bond. Another key parameter expresses the dihedral angles of the
bridgehead phenyls relative to the strained bridging ketone car-
bonyl. The tabulated dihedrals express the smaller magnitude
values measured from the carbon of the ketone carbonyl to the
For the phencyclone adduct, 3e, from N-[3,5-bis(trifluo-
romethyl)phenyl]maleimide, we have tentatively assigned the
weak signal in the carbon-13 NMR at 132.14 ppm to NC1" (the
ipso carbon) of the N-aryl ring, because of its slightly broadened
appearance (w
= 2.3 Hz), which could be attributed to small
1/2
4
unresolved long-range J coupling to fluorine. The slightly
broadened approximate quartet assigned to the C2",6" carbons

Jul-Aug 2005
Multinuclear NMR Studies and Ab Initio Structure Calculations
897
bridgehead phenyl carbon (designated C6') directed "upwards"
and proximal to the ketone, i.e., O=C-C(sp )-C1'(ipso)-C6'. Note
tances from H2' to the endocyclic heteroatoms for 3c, 2.89 versus
2.77 Å, which might also influence direct anisotropic effects.
Calculated structures for adducts 3b and 3c are shown in Figures
3c and 3d, with the bridgehead methines, phenanthrenoid hydro-
gens H1,8 and bridgehead phenyl hydrogens H2' and H6'
included to highlight potential interactions.
For the adduct 3d, from Tuppy's maleimide, three separate
optimizations were performed, with two showing relatively good
agreement to within 0.000062 au (0.039 Kcal/mol), and one
appreciably higher (by 0.0008814 au, 0.55 Kcal/mol). The opti-
mized structures appeared to differ mainly in the conformations
3
that the observed pairs of dihedral angles reported have opposite
signs reflecting the local chiralities of each "half" of the adduct.
For adducts 3b and 3c, non-planarity in the –X-C(=X)-X- group
is expressed as the distance from the central carbon to the plane
defined by the three bonded heteroatoms. Both of these adducts
show 0.006 Å or less of non-planarity by this measure. The Ring
Planarity of these adducts measures "envelope folding" of these
five-membered heterocycles; they are highly planar. For the
maleimide adducts, 3d and 3e, non-planarity at the nitrogen is
expressed as the distance from the nitrogen to the plane defined
by the three directly bonded atoms. This N-planarity for each of
the latter adducts was ca. 0.01 Å, with the nitrogen directed "out-
wards," and the N-aryl rings actually bent "inwards" toward the
phenanthrenoid system.
of the –NO and –NMe groups on the N-aryl ring. Some notable
2
2
features of the N-aryl ring substituents follow. The plane of the
N-aryl ring was defined by carbons C1", C3", C5", and the planes
of each nitro group were defined by the NO atoms. For the
2
"inner" nitro (on the (b) edge of the N-aryl ring at the 5" posi-
tion), the angle between the N-aryl and the NO planes was
For the cyclohexene adduct, 3a, two different minima were
located, referred to as the anti (lower energy) and the syn (higher
energy) conformers, differing in the orientation of the "boat-like"
conformations of the tetramethylene chain. For the more stable
2
41.44°; the corresponding angle was 49.62° for the "outer" nitro
at the 3" position (on the (a) edge of the N-aryl ring). The
observed dihedral angle from the innermost oxygen of the "inner"
5"-nitro to C4", i.e., O=N(O)-C5"-C4", was -140.72°. For the
"outer" 3"-nitro, the dihedral angle from the oxygen syn to the
ketone C=O to C4", i.e., O=N(O)-C3"-C4", was 130.20°. The
anti structure, the (gamma) (CH ) bridge is further from the
2 2
phenanthrenoid central ring; in particular, the (beta "in") hydro-
gens, H , are relatively close to the center of that ring (Point 1,
d
defined above), within 3.19 Å, and proximal (63.8°) to the axis
perpendicular to the plane of the phenanthrenoid central ring and
passing through the ring center. Anisotropic shielding effects
nitrogen of the –NMe group was quite pyramidalized with the
nitrogen located 0.252 Å from the plane of its three directly
2
bonded neighbors. The –NMe group was highly skewed with
2
should be favored for the H protons in this orientation [28]. In
respect to the N-aryl ring. The tabulated data correspond to those
for the lowest energy of the three calculations.
d
contrast, the 3a syn conformer would have the H (gamma "in")
b
protons calculated as closest to the middle of the phenanthrenoid
central ring. Tabulated data for the 3a conformers includes dis-
The most striking difference between adducts 3d and 3e
(from N-[3,5-bis(trifluoromethyl)phenyl]maleimide) appeared
to be in the relative angles between the N-aryl rings and the
pyrrolidinedione rings. These angles were estimated from the
calculated dihedral angles from the imide C=O to the ("inner")
C6" of the N-aryl, O=C-N-C1"-C6". The N-aryl and imide rings
were closer to coplanarity for 3d by about 9°, i.e., ca. 50°
between the rings for the Tuppy's maleimide adduct and 59° for
3e. This would account for the "inner" N-aryl H6" in 3e being
closer to the center of the middle ring of the phenanthrenoid
moiety (Point 1), 3.65 versus 3.81 Å, and presumably should
make it somewhat more shielded. This does not appear to be
experimentally verifiable, however, since the apparent
anisotropic shielding magnitudes determined by comparing N-
aryl ortho proton chemical shifts in the adducts and the precur-
sor maleimides may indicate the contrary. Thus, the published
tances from H2' to each of the (beta) protons, H and H , and to
a
d
the closer of the (gamma) protons. The closer (gamma) proton is
H "out" for 3a anti and H "in" for 3a syn. The calculated energy
c
b
difference, ∆E, for the 3a anti/syn pair was 0.01019 au or 6.39
Kcal/mol (1 au = 627.5 Kcal/mol). This relatively large energy
difference implies that the anti conformer would be overwhelm-
ingly favored at equilibrium, which would be consistent with the
appreciable shielding for the H (beta "in") proton signal
d
assigned at 0.73 ppm versus the H (beta "out") signal assigned to
a
the lower field 1.86 ppm absorption. Calculated structures for the
3a anti/syn pair are shown in Figures 3a and 3b, with aryl hydro-
gens hidden for clarity.
The vinylene carbonate adduct, 3b, appeared to be strikingly
different from the anti cyclohexene adduct in several respects.
The dihedral angles, O=C-C-C1'-C6', for the bridgehead phenyls
are more than 10° smaller for the vinyl carbonate adduct, bring-
ing H6' of the phenyls closer to the bridging ketone oxygen (2.67
Å in 3b versus 2.91 Å in 3a anti) and bringing H2' of the phenyls
much closer to the phenanthrenoid H1 (3.00 versus 3.33 Å).
Increased anisotropic contributions might be expected from the
closer proximities.
Duplicate optimizations for the adduct, 3c, from vinylene
trithiocarbonate, showed close agreement in the optimized ener-
gies, to within ca. 0.0000033 au. Comparing 3b to 3c, nearly 7°
greater dihedral angles, O=C-C-C1'-C6', for the bridgehead
phenyls are seen for the trithiocarbonate adduct, bringing H6'
closer to the phenanthrenoid H1 (3.29 versus 3.44 Å) and H2' fur-
ther from H1 (3.23 versus 3.00 Å); the greater H2' to H1 distance
might decrease deshielding by the phenanthrenoid and partly
account for the higher field chemical shift observed for H2' in 3c.
Also, longer bond lengths to sulfur would account for greater dis-
1
H NMR spectrum for Tuppy's maleimide [11] shows the aryl
proton resonance at ca. 7.97 ppm while the corresponding
adduct signal for 3d is at 6.30 ppm implying a shielding magni-
tude of about 1.67 ppm. For the bis-trifluoromethyl adduct 3e,
the N-aryl ortho proton signal appears at 6.33 versus 7.95 ppm
for the maleimide, 2e, indicating a slightly smaller shielding
magnitude of 1.62 ppm. It may be that attempting to reconcile
the apparent shielding magnitudes in the adducts with the dis-
tances of H6" to Point 1 is an oversimplification, since the angle
Point 2 to Point 1 to H6" is greater in 3e (97.3°) than in 3d
(95.4°), and the latter angle, closer to 90°, might tend to
enhance anisotropic shielding. Lastly, we note that in 3e, for the
"inner" 5"-trifluoromethyl group, the fluorine closest to Point 1
is 5.232 Å away. Dihedral angles from each fluorine of this CF
3
group to C4" were 148.22, 27.76, and -91.96°. Calculated struc-
tures for adducts 3d and 3e are shown in Figures 3e and 3f, with
all hydrogens hidden for clarity.

898
R. Callahan, O. Ramirez, III, K. Rosmarion, R. Rothchild, K. C. Bynum
Vol. 42
REFERENCES AND NOTES
Correspondence should be sent to RR at John Jay College (e-mail:
rrothchild@jjay.cuny.edu).
[1] T. Sasaki, K. Kanematsu and K. Iizuka, J. Org. Chem., 41,
1105 (1976).
[2] R. Callahan, K. Bynum, R. Prip and R. Rothchild, The Chem.
Educator, 3 S1430 (1998) [02192-4 Avail. URL: http://journals.springer-
ny.com/chedr].
[3] K. Bynum and R. Rothchild, Spectrosc. Lett., 29, 1599(1996).
[4] K. Bynum and R. Rothchild, Spectrosc. Lett., 29, 1621 (1996).
[5] K. Bynum and R. Rothchild Spectrosc. Lett., 30, 727 (1997).
[6] M. Amin, K. Bynum, R. Callahan, R. Prip and R. Rothchild,
Spectrosc. Lett., 31, 673 (1998).
[7] K. Marshall and R. Rothchild, Spectrosc. Lett., 37, 469 (2004).
[8] O. Ramirez, III, R. Tokhunts and R. Rothchild, Spectrosc.
Lett., 37 421 (2004).
[9] O. Sklyut, R. Prip, N. Azar, R. Callahan and R. Rothchild,
Spectrosc. Lett., 37, 493 (2004).
Conclusions.
A series of five phencyclone adducts, 3a-e, have been prepared
from a diverse group of symmetrical cyclic dienophiles, including
cyclohexene, vinylene carbonate, vinylene trithiocarbonate,
Tuppy's maleimide and N-[3,5-bis(trifluoromethyl)phenyl]-
maleimide, with detailed proton NMR assignments for the
adducts obtained by high resolution COSY45 spectra for each
adduct, and assignments for protonated carbons obtained by HET-
COR spectra for adducts 3a-d. Spectra were obtained at ambient
temperatures at 300 MHz for proton, 75 MHz for carbon-13, and
282 MHz for fluorine-19 (for adduct 3e). For all five adducts,
bridgehead phenyl rotation was slow on the proton and carbon-13
NMR timescales, with slow exchange limit spectra observed. In
contrast, N-aryl rotations for both adducts, 3d and 3e, from the N-
*
aryl maleimides, were fast based on sharply averaged fast
exchange limit spectra for proton and carbon-13 NMR (and
19
F
NMR, for 3e), suggesting that meta disubstitution at the N-aryl
C3",5" positions is not sufficient to retard the N-aryl ring rotation.
Substantial anisotropic shielding of the N-aryl ortho H2",6" pro-
tons in 3d and 3e is consistent with the expected effect of the
phenanthrenoid moiety, with each of the H2",6" protons spending
half their time in a conformation placing them proximal to the
central aromatic ring of the phenanthrenoid. Geometry optimiza-
tions for all five adducts were performed at the Hartree-Fock level
with the 6-31G* basis set. Two distinct conformers were found for
the cyclohexene adduct, with the more stable 3a anti more consis-
tent with proton assignments. A substantial range of adduct
geometries for the calculated structures was found, especially
with respect to the conformations of the bridgehead phenyls rela-
tive to the bridging ketone carbonyl. Some anomalous chemical
shifts in the vinylene trithiocarbonate adduct were found (relative
to the other adducts). The calculated structures, for the cyclic car-
bonate and trithiocarbonate rings in adducts 3b and 3c were
highly planar. The N-aryl rings of the maleimide adducts were
twisted by ca. 50-59° away from coplanarity with the planes of
the imide systems. Proton NMR assignments for the vinylene car-
bonate adduct presented here differ somewhat from an earlier
report. The endo stereochemistry in this adduct was confirmed by
the sharp, unsplit singlet signal for the ketone C=O carbon in the
[10] E. A. Harrison, Jr. and H. L. Ammon. J. Org. Chem., 45, 943
(1980).
13
[11] C. J. Pouchert and J. Behnke, The Aldrich Library of C and
H FT NMR Spectra, 1st ed., Aldrich Chemical, Milwaukee, WI, 1993.
1
cyclohexene: 1 (1) 57C; vinylene carbonate: 1 (1) 933A; ethylene carbon-
ate: 1 (1) 932A; vinylene trithiocarbonate: 1 (1) 933B; ethylene trithiocar-
bonate: 1 (1) 932A; Tuppy's maleimide: 1 (2) 1456B.
[12] M. Roth, Helv. Chim. Acta, 62, 1966 (1979).
[13] M. Matsumoto, H. Hiratani, K. Nakada, Y. Yamamoto and Y.
Tarumi. Jpn. Kokai Tokkyo Koho JP 06,279,403 [94,279,403] (Cl.
C07D207/452), 04 Oct. 1994, Appl. 93/69,886, 29 Mar. 1993; Chem.
Abstr., 122, 134130p (1995).
[14] M. El-Guweri, P. Hendlinger and A. Laschewsky, Macromol.
Chem. Phys., 198. 401 (1997).
[15] M. Eto, K. Setoguchi, A. Harada, E. Sugiyama and K. Harano,
Tetrahedron Lett., 39, 9751 (1998).
[16] Y. Yoshitake, J. Misaka, K. Setoguchi, M. Abe, T. Kawaji, M.
Eto and K. Harano, J. Chem. Soc., Perkin Trans. 2, 1611 (2002).
[17] Y. Yoshitake, J. Misaka, M. Abe, M. Yamasaki, M. Eto and K.
Harano, Org. Biomol. Chem., 1 1240 (2003).
[18] A. P. Marchand, G. Srinivas and W. H. Watson; ARKIVOC
(Online ed.) ISSN 1424-6369, (2003) Part (iii) (SD-571C) 8-15. URL:
http://www.arkat-usa.org/ark/journal/2003/Sukh%20Dev/SD-
571C/571C.htm.
[19] R. Callahan, K. Marshall, R. Rothchild and K. Rosmarion, The
Chem. Educator, 6, 227234 (2001) [DOI 10.1007/s00897010488a].
[20] E. A. Harrison, Jr., Organic Preparations and Procedures Intl.,
7, 71 (1975).
[21] C. J. Pouchert, The Aldrich Library of FT-IR Spectra, Edition
I, Aldrich Chem. Co., Inc. (Milwaukee, WI) 1985. Vinylene carbonate: 1
(1) 644A; ethylene carbonate: 1 (1) 612A; ethylene trithiocarbonate: 1 (1)
612D.
13
gated decoupling C NMR spectrum. Lastly, we would like to
mention some suggestions from the Reviewers that might be very
fruitful for further studies, including variable temperature (VT)
NMR studies and GIAO (Gauge Independent Atomic Orbital) cal-
culations for chemical shifts; our available NMR hardware did not
allow us to do the VT experiments and our current computational
software does not provide GIAO calculations. A second Reviewer
suggested the use of the powerful NMR techniques of HMQC,
HSQC, EXSIDE, HSQMBC, etc., which could not be performed
on our present instrument. We are grateful for these suggestions.
[22] R. Y. S. Tan, R. A. Russell and R. N. Warrener, Tetrahedron
Lett., 20, 5031 (1979).
[23] Z. Y. Wang, Synthetic. Commun., 20, 1607 (1990).
[24] F. Jensen, Introduction to Computational Chemistry, John
Wiley, New York, 1999.
Acknowledgements.
[25] Spartan '04 Windows Tutorial and User's Guide,
Wavefunction, Inc., Irvine, CA, 2001-2003.
[26] W. J. Hehre, A. J. Shusterman and J. E. Nelson, The Molecular
Modeling Workbook for Organic Chemistry, Wavefunction, Inc., Irvine,
CA, 1998.
[27] W. J. Hehre, A Guide to Molecular Mechanics and Quantum
Chemical Calculations, Wavefunction, Inc. , Irvine, CA, 2003.
[28] H. Günther, NMR Spectroscopy: Basic Principles, Concepts,
and Applications in Chemistry, 2nd ed. John Wiley, New York, 1995.
Partial support has been provided to R. R. by the U.S.
Education Department (Hispanic-Serving Institutions-Title V,
and the Minority Science and Engineering Improvement
Program), National Science Foundation and the Professional
Staff Congress-City University of New York Research Award
Program. Further support has been provided by the Project
ASCEND/McNair Program to O. R. and K. R. Technical assis-
tance was provided by Kimberly Marshall.