1H, 13C, and 19F NMR studies of phencyclone adducts of N-(polyhalophenyl)maleimides: Evidence for dynamic NMR in maleamic acids - Ab initio calculations for optimized structures

Article abstract of DOI:10.1016/j.jfluchem.2004.07.002

NMR methods, including one- and two-dimensional techniques (at 7.05 T) for ¹H, ¹³C and ¹⁹F, have been applied to studies of hindered rotations and magnetic anisotropy in some crowded Diels-Alder adducts of phencyclone (1).

Full text of DOI:10.1016/j.jfluchem.2004.07.002

Journal of Fluorine Chemistry 125 (2004) 1893–1907
www.elsevier.com/locate/fluor
13
19
1H, C, and F NMR studies of phencyclone adducts of
N-(polyhalophenyl)maleimides: evidence for dynamic NMR
in maleamic acids
Ab initio calculations for optimized structures
Kimberly Marshalla, Kerstin Rosmarionb, Olga Sklyutb, Nikolay Azarb,
Ronald Callahanc, Robert Rothchildb,d,
*
aChemistry and Chemical Engineering Department, California Institute of Technology, Pasadena, CA 91125, USA
bScience Department, John Jay College of Criminal Justice, The City University of New York,
445 West 59th Street, New York, NY 10019-1128, USA
cChemistry Department, New York University, 29 Washington Place (569 Brown), New York, NY 10003, USA
dDoctoral Faculty, Graduate School and University Center, City University of New York, New York, NY, USA
Received 1 April 2004; received in revised form 9 July 2004; accepted 13 July 2004
Available online 3 October 2004
Abstract
NMR methods, including one- and two-dimensional techniques (at 7.05 T) for 1H, 13C and 19F, have been applied to studies of hindered
rotations and magnetic anisotropy in some crowded Diels–Alder adducts of phencyclone (1). Symmetrically substituted N-aryl maleimides
(2) bearing numerous halogens on the N-aryl ring, were employed as dienophiles to form the target adducts (3). The maleimides included: N-
(4-bromo-2,6-difluorophenyl)maleimide (2a); N-(2,3,5,6-tetrafluorophenyl)maleimide (2b); N-(4-bromo-2,3,5,6-tetrafluorophenyl)malei-
mide (2c); N-(2,3,4,5,6-pentachlorophenyl)maleimide (2d); and N-(2,4,6-tribromophenyl)maleimide (2e). Maleimides (2a–2c) were pre-
pared from the precursor N-aryl maleamic acids (5a–5c). Ambient temperature fluorine-19 NMR of these maleamic acids in d6-acetone
showed substantial unusual peak broadening consistent with intermediate exchange rate processes, which may correspond to the N-aryl
rotation process. Maleimides (2d) and (2e) were produced in one step from pentachloroaniline or 2,4,6-tribromoaniline, respectively, and
maleic anhydride with anhydrous ZnCl2 at ca. 200 8C. For the adducts (3), we observed slow exchange limit spectra on the 1H, 13C, [and 19F,
for (3a–3c)] NMR timescales for the rotation of the unsubstituted bridgehead phenyls about the C(sp3)–C(sp2) bonds, and for the rotations of
the N-aryl rings about the N(sp2)–C(aryl sp2) bonds. Ab initio calculations for geometry optimizations at the Hartree–Fock level with 6-31G*
(or LACVP*) basis sets were performed for the adducts. We believe that this is the first report of detailed 1H, 13C, and 19F NMR data for a
substantial collection of N-aryl maleamic acids, maleimides and their phencyclone adducts bearing multiple fluorines or other halogens
directly on the N-aryl ring, together with complementary quantitative geometric parameters from high-level HF/6-31G* (or LACVP*)
calculations.
# 2004 Published by Elsevier B.V.
Keywords: Hindered rotations; Magnetic anisotropy; Dynamic NMR; One- and two-dimensional (1D and 2D) NMR; COSY; HETCOR; Hartree–Fock;
Molecular modeling
1. Introduction
multinuclear NMR studies of hindered systems derived from
phencyclone (1), by Diels–Alder additions with symmet-
rically substituted halogen-containing N-aryl maleimides,
bearing multiple fluorine, chlorine and/or bromine atoms
directly on the N-aryl ring. Phencyclone is well known as an
effective diene component for Diels–Alder cycloaddition
reactions [1] and N-substituted maleimides (2), are excellent
This report presents the first compilation of the results of
ab initio calculations (Hartree–Fock level) together with
* Corresponding author. Tel.: +1 212 237 8886.
E-mail address: rrothchild@jjay.cuny.edu (R. Rothchild).
0022-1139/$ – see front matter # 2004 Published by Elsevier B.V.
doi:10.1016/j.jfluchem.2004.07.002

1894
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
dienophiles with (1) [2]. The normally expected endo
stereochemistry for the Diels–Alder adducts (3), results in
highly hindered systems. (Fig. 1 shows the relevant
structures and synthetic pathways.) Slow exchange limit
routine one-dimensional (1D) spectra, we used relaxation
1
delays (RD) of 1 s for H, 3 s for 13C, and 5 s for 19F.
Because of extensive splitting of some carbon signals by
fluorine, not all non-protonated carbons produced positively
observable absorptions, even though we routinely used
15,000–17,000 FID acquisitions (i.e., number of scans, NS)
for the 1D 13C NMR. For two of the adducts (3d) and (3e),
1D 13C NMR spectra were obtained, using RD values of 1,
5 and 60 s, and the number of scans was as high as 27,000.
[Note: Further details regarding interpretations and assign-
ments of the NMR spectra are presented in the Section 3 for
most compounds, including maleamic acids (5a–5c);
maleimides (2d) and (2e); and adducts (3a–3e), and should
be consulted in conjunction with the brief summary details
that are included below in this Section 2.] Solvents and
reagents were obtained from Aldrich Chemical (Milwau-
kee, WI) and were used without further purification.
Melting points were obtained on Mel-Temp apparatus
(Laboratory Devices, Cambridge, MA) and are uncorrected.
Infrared (IR) spectra for two of the adducts, were obtained
on Type 61 (polyethylene) 3M Disposable IR Cards; only
stronger peaks are reported, and peaks in the 2800–
3000 cmꢀ1 region in which the cards absorb strongly have
been omitted. Spectra were acquired on a Perkin-Elmer
1640 instrument with DTGS detector, using 4 cmꢀ1
resolution. For computational studies, Spartan ’04 for
Windows (full version, v.1.0.0 #1991–2003; 17 September
2003) or Spartan ’04 Essential (v.2.0.0 #1991–2003; 8
October 2003) was obtained from Wavefunction Inc.,
Irvine, CA and installed on Dell Pentium 4 platforms with
processor speeds of 2.26 or 3.06 GHz, using 512 or 1024
MB memory. Calculation parameters included turning
symmetry OFF and convergence ON.
1
(SEL) H and 13C NMR spectra at ambient temperatures
were observed for hindered rotations about the C(sp2)–
C(sp3) bonds to the unsubstituted bridgehead phenyls in the
adducts, using medium field strength NMR spectrometers,
e.g., ca. 200–300 MHz for proton and 50–75 MHz for
carbon-13 [3–7]. When the adducts are formed from N-aryl
maleimides, potential hindered rotations about the N(sp2)–
C(aryl sp2) bonds in the N-aryl moieties could be examined
[8]. With the endo adduct stereochemistry, striking examples
of magnetic anisotropic shielding effects may be observed as
a result of proximity of the N-aryl substituent to the
phenanthrenoid group of the adduct [9,10]. Modern
molecular modeling software for ab initio calculations of
optimized adduct structures has been used to gain insight
into adduct geometries [10,11]. When fluorine atoms are
incorporated into the N-aryl substituent, NMR studies of (3)
can be extended to a third nucleus, 19F [12]. These present
and ongoing studies are broadly aimed at obtaining a better
understanding of molecular motions (especially hindered
rotations) and the geometric aspects of magnetic anisotropic
effects.
2. Experimental
General NMR and other techniques were described
earlier [2,13]. Spectra were obtained on a Bruker ACF300
1
NMR spectrometer (7.05 T magnet) at 300 MHz for H,
75 MHz for 13C, and 282 MHz for 19F, using the QNP (quad
nuclear probe) and Aspect 3000 data system. Acquisitions
were performed at ambient temperatures in CDCl3 unless
otherwise noted, e.g., d6-acetone for maleamic acids.
Proton shifts were referenced to internal tetramethylsilane,
TMS, at 0.0 ppm. Carbon-13 shifts were referenced to the
center line of the CDCl3 triplet at 77.0 ppm (or to internal
TMS at 0.0 ppm with other solvents). Fluorine spectra were
referenced to internal CFCl3 at 0.0 ppm. Standard Bruker
microprograms were routinely used for two-dimensional
(2D) 1H–1H and 19F–19F homonuclear chemical shift
correlation NMR by COSY45, and for the 2D 1H–13C
heteronuclear HETCOR. NMR signal multiplicities are
abbreviated as: s (singlet), d (doublet), t (triplet), q
(quartet), m (multiplet). Q refers to quaternary non-
protonated aryl carbon signals. High resolution COSY
spectra for the aryl proton regions typically employed 256
increments in the t1 dimension with spectral widths of ca.
600 Hz, e.g., the regions from about 6.8–8.8 ppm. In listing
the 1H NMR data, we generally show gross multiplicities, d
or t, for signals of the bridgehead phenyls and phenan-
threnoid moieties, without explicitly listing the observed
vicinal splittings, which were in the normal range, ca. 8 Hz
(or the smaller long-range splittings that could be seen). For
2.1. Syntheses of maleamic acids
The maleamic acids (5a–5c) were prepared by similar
procedures, from maleic anhydride and the corresponding
substituted anilines (4a–4c), as shown in Fig. 1. Details of
the synthesis of (5a) are given as an illustrative procedure
(see Section 3 for further details regarding NMR spectra of
compounds (5a–5c)).
2.2. N-(4-Bromo-2,6-difluorophenyl)maleamic
acid (5a)
A solution of 4-bromo-2,6-difluoroaniline (1.024 g,
4.924 mmol) in ca. 30 mL CH2Cl2 was added dropwise
with magnetic stirring to a suspension of maleic anhydride
(0.490 g, 4.997 mmol) in 30 mL CH2Cl2. After 30 min at
room temperature, solvent removal gave the crude maleamic
acid as an off-white crystalline solid (1.340 g, 89% yield,
mpr 170–172 8C), used without further purification (see
Section 3 for further details of NMR). 1H NMR (d6-acetone,
ppm): 10.14 (1H, br s, NH); 7.46 (2H, approx. d, apparent
3J(HCCF) = 7.14, aryl H); 6.78 (1H, d, 3J = 12.71, half of AB

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1895
Fig. 1. Summary of synthetic routes with general compound structures and atom numbering. Syntheses of the maleamic acids and maleimides are shown at the
top. Formation of the adducts (3a–3e), from phencyclone with the respective maleimide, together with atom numbering for the adducts, is shown across the
center of the Figure. The halogen substitution patterns of the N-aryl rings in each series of compound (a–e) is shown at the lower left of the Figure. The general
structure of the acetoxypyrrolidinone impurities (6), observed by NMR in some samples of the maleimides (2a–2c), appears at the bottom right.
3
q, cis HC=CH); 6.47 (1H, d, J = 12.69, half of AB q, cis
HC=CH). 19F NMR (d6-acetone): ꢀ114.87 (slightly
2.4. N-(4-bromo-2,3,5,6-tetrafluorophenyl)maleamic
acid (5c)
3
broadened apparent d, J = 8.29).
Prepared analogously to (5a), above, after standing at
ambient temperature (1 week) with slow evaporation of
solvent, the crude maleamic acid crystals were obtained
(97% yield) mpr 129–135 8C, used directly for conversion to
the maleimide (see Section 3 for further details of NMR). 1H
NMR (d6-acetone): 10.27 (1H, very broad s, NH); 6.75 (1H,
d, 3J = 12.52, cis, HC=CH, half of AB q); 6.47 (1H, d, 3J =
12.53, cis, HC=CH, half of AB q). 19F NMR (d6-acetone):
ꢀ135.01 (2F, slightly broadened symmetrical complex m,
ꢀ143.18 (2F, very broad d, apparent J = 18.3).
2.3. N-(2,3,5,6-Tetrafluorophenyl)maleamic acid (5b)
Prepared analogously to (5a), above, the crude maleamic
acid was washed with hexane and dried to give pure white
crystals of the maleamic acid (mpr 117–119.5 8C, lit. mp
121 8C [14]). On standing, a second crop was obtained,
collected by vacuum filtration, washed with cold hexane,
and dried, to give a tan solid (combined total 61% yield; see
Section 3 for further details of NMR). 1H NMR (d6-
acetone): 10.25 (1H, br s, NH); 7.55 (1H, tt, J = 10.46 and
3
7.33, aryl H); 6.76 (1H, d, cis J = 12.58, half of AB q,
2.5. Syntheses of maleimides
HC=CH); 6.46 (1H, d, cis 3J = 12.59, half of AB q, HC=CH).
19F NMR (d6-acetone): ꢀ140.25 (2F, q, J = 10.5); ꢀ144.99
(2F, br s).
The maleimides (2a–2c) were prepared by similar
procedures, from the respective maleamic acids (5a–5c),

1896
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
by reaction in acetic anhydride (boiling water bath), promoted
with anhydrous NaOAc. The illustrative procedure for (2a) is
given in detail. For maleimides (2d) and (2e), direct one-step
high-temperature reactions (ca. 200 8C) were used, with the
corresponding anilines (4d) or (4e), in maleic anhydride as
solvent, promoted by anh. ZnCl2. The detailed procedure for
preparation of (2d) is given below (see Section 3 for further
details regarding NMR spectra of compounds (2d) and (2e)).
combined with anh. sodium acetate (90 mg, 1.10 mmol) and
ca. 4 mL acetic anhydride. After heating in a boiling water
bath for 30 min, the dark reddish-brown mixture was cooled
and transferred to a separatory funnel with 50 mL H2O and
50 mL CH2Cl2. After extensive washings as above, the
yellowish solution was treated with decolorizing carbon,
and filtered through a Celite pad. Solvent removal gave the
crude beige maleimide (557 mg, 1.72 mmol, 44% yield)
1
mpr 80–90 8C, used for adduct formation. H NMR: 6.99
(2H, s, HC=CH). In addition to the maleimide signal, three
dd resonances were observed, consistent with ca. 8%
2.6. N-(4-Bromo-2,6-difluorophenyl)maleimide (2a)
3
impurity acetoxypyrrolidinedione: 5.65 (1H, dd, J = 8.95
The crude maleamic acid (5a) (565.3 mg, 1.847 mmol),
acetic anhydride (20 g), anhydrous sodium acetate (20 mg)
and 2,6-di-t-butyl-4-methylphenol (BHT, ca. 3 mg) were
placed in a flask fitted with reflux condenser topped by a
drying tube and heated in a boiling water bath for 1.2 h. The
dark reddish-brown mixture was cooled and quenched with
ice water, extracted with CH2Cl2 and washed repeatedly with
water, 5% NaHCO3, and brine. The organic layer was separa-
ted and dried, and solvent was removed (rotary evaporator).
The resulting solid was recrystallized from toluene to give
light tan needles of the maleimide (210 mg, 39.5% yield, mpr
107.5–112 8C), which was used for subsequent reaction with
and 5.20, AcOCH, HX); 3.44 (1H, dd, 2J = 18.58, 3J = 8.97,
2
3
HB); 3.01 (1H, dd, J = 18.60, J = 5.22, HA). 19F NMR:
ꢀ132.04 (2F, complex symmetr. m [eight lines]); ꢀ142.50
(2F, complex symmetr. m [eight lines]). Each eight-line
multiplet was sharp, and the two multiplets were essentially
identical.
2.9. N-(2,3,4,5,6-Pentachlorophenyl)maleimide (2d)
2,3,4,5,6-Pentachloroaniline (1.3254 g, 5.00 mmol),
maleic anhydride (2.5196 g, 25.69 mmol) and anhydrous
ZnCl2 (350.2 mg, 5.270 mmol) were combined in a test tube
and maintained at ca. 200 8C for 65 min, using a
thermometer to monitor the temperature and provide
mixing. After cooling to room temperature, the material
was transferred with ca. 100 mL boiling water to an
erlenmeyer flask, heated further (steam bath) and allowed to
cool and crystallize. The resulting crude light beige
maleimide had mpr 138–153 8C, lit. 1528 [15], 1.546 g,
4.48 mmol, 89.6% yield (see Section 3 for further details of
NMR). 1H NMR: 6.96 (H–C=C–H). 13C NMR: 167.00
(imide C=O); 134.88 (H–C=C–H); 135.85 very weak (Q);
133.82 (Q); 132.52 (Q); 127.95 very weak (Q).
1
phencyclone. H NMR (CDCl3, ppm): 8.05 (1H, very br s,
NH); 7.23–7.29 (2H, complex symmet. m, aryl H); 6.94 (2H,
s, HC=CH). Note trace of acetoxypyrrolidinedione with
three equal area dd signals: 5.66 (dd, 3J = 5.3 and 3J = 9.0,
HX); 3.40 (dd, 2J = 18.75, 3J = 8.82, HB); 2.95 (dd, 2J = 18.75,
3J = 5.15, HA). 19F NMR: ꢀ115.16 (s, aryl F).
2.7. N-(2,3,5,6-Tetrafluorophenyl)maleimide (2b)
Prepared analogously to (2a), above. The crude maleamic
acid (5b) (1.910 g, 7.26 mmol), anh. sodium acetate
(243 mg, 2.96 mmol) and acetic anhydride (ca. 2.5 mL)
were heated in a boiling water bath for 30 min, becoming
dark reddish-black, and allowed to cool to room temperature
and stand for 3 days. The mixture was transferred to a
separatory funnel with water (25 mL) and CH2Cl2 (25 mL),
and the organic layer was repeatedly washed with water,
aqueous NaHCO3 and water. After drying (anh. MgSO4),
solvent removal from the organic layer gave the crude
maleimide as an off-white solid (1.206 g, 67% crude yield)
mpr 124–132 8C, lit. mp 1378 [14], used directly for adduct
formation. 1H NMR: 7.23 (1H, tt, J = 9.69 and 7.25, aryl H);
6.98 (2H, s. HC=CH). Also present were signals consistent
with ca. 35% of the acetoxypyrrolidinedione: 5.67 (1H, dd,
3J = 8.92 and 5.20, HX); 3.44 (1H, dd, 2J = 18.55, 3J = 8.95,
2.10. N-(2,4,6-Tribromophenyl)maleimide (2e)
Prepared analogously to (2d), above. 2,4,6-Tribromoani-
line (2.6528 g, 8.04 mmol), maleic anhydride (5.0085 g,
51.08 mmol) and anhydrous ZnCl2 (686.4 mg, 5.037 mmol)
were maintained at ca. 190–200 8C for 75 min, using a
thermometer to monitor the temperature and provide
occasional mixing. (A bath of boiling ethylene glycol, bp
196–1988, was used to provide a stable heat source.)
Following workup as for (2d), the resulting crude tan
maleimide had mpr 129–135 8C, lit. 1428 [15], 2.9562 g,
7.21 mmol, 89.7% yield (see Section 3 for further details of
NMR). 1H NMR: 7.82 (2H, s, aryl H); 6.91 (2H, s, HC=CH).
Ca. 3% or less of pyrrolidinedione impurities were seen
2
3
HB); 3.00 (1H, dd, J = 18.59, J = 5.21, HA); 2.22 (3H, s,
OAc). 19F NMR: ꢀ138.11 (2F, complex mult.); ꢀ143.98
(2F, complex mult.); plus impurity peaks.
3
based on small dd resonances at: 4.85 (1H, dd, J = 8.47,
3.83, HX); 3.56 (1H, dd, 2J = 18.89, 3J = 8.66, HB); 3.17 (1H,
dd, 2J = 18.92, 3J = 3.77, HA). 13C NMR: 167.34 (approx. t, J
= 9.49, imide); 135.05 (2 ꢁ CH, dd, 1J = 175.57, 3J = 6.21,
2.8. N-(4-Bromo-2,3,5,6-tetrafluorophenyl)maleimide (2c)
1
2
Prepared analogously to (2a), above. The crude maleamic
acid (5c) (1340 mg, 3.918 mmol) prepared above, was
aryl H3,5); 134.66 (2 ꢁ CH, dd, J = 185.13, J = 2.34,
HC=CH); 129.84 (very weak, t, 3J = 7.73, ipso NC1); 125.73

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1897
(weak, t, 2J = 2.27, 2 ꢁ CBr, C2,6); 124.73 (very weak, t, 2J
= 4.09, 1 ꢁ CBr, C4).
3,6); 7.43 (2H, approx. t, H50); 7.11–7.24 (6H, m, including
H60 at 7.23, H2,7 at 7.20, H1,8 at 7.13); 6.88 (1H, tt, J = ca.
9.7 and 7.1, C6F4H); 4.69 (2H, 2 ꢁ sp3). 13C NMR: 195.74
(ketone C=O); 171.30 (imide NC=O); weak multiplets ca.
147.19, 143.84 and 140.38; 135.24 very weak (assigned to
trace impurity of maleimide HC=CH); 133.39 (Q); 133.07
(Q); 131.73 and 131.71 (poss. split peak? Q); 130.96 (2 ꢁ
CH); 129.40 (2 ꢁ CH); 128.82 (2 ꢁ CH); 128.68 (2 ꢁ CH);
128.54 (2 ꢁ CH); 127.14 (2 ꢁ CH); 126.53 (2 ꢁ CH);
126.08 and 126.07 (poss. split peak? Q); 125.61 (2 ꢁ CH);
123.12 (2 ꢁ CH); very weak multiplet ca. 111.1; 107.27 (1 ꢁ
CH, t, J = 22.6, C400); 63.51 (C6H5–C); 45.67 (2 ꢁ CH, sp3).
19F NMR: ꢀ137.77 (1F, dd, J = 21.85 and 11.52); ꢀ138.69
(1F, dd, J = 21.87 and 11.75); ꢀ144.15 (1F, ddd, J = 21.83,
11.59, 5.34); ꢀ144.57 (1F, ddd, J = 21.83, 11.59, 5.34).
2.11. Syntheses of phencyclone adducts
Similar procedures were used for the five adducts (3a–
3e), and illustrative details are provided for (3a) and (3b)
with noteworthy differences cited for the other adducts (see
Section 3 for further details regarding NMR spectra of
compounds (3a–3e)).
2.12. N-(4-Bromo-2,6-difluorophenyl)maleimide
adduct (3a)
An excess of the crude 4-bromo-2,6-difluorophenylma-
leimide (114.3 mg, 0.397 mmol) and phencyclone (76.4 mg,
0.200 mmol) were refluxed in 25 mL toluene. The dark
green–black phencyclone color was discharged in 20 min to
give a clear yellow solution. Solvent removal by rotary
evaporator gave the crude adduct which was recrystallized
from ethanol to give white solid (45 mg, 34% yield), mpr
273–280 8C (dec. with darkening and gas evoln., see Section
3 for further details of NMR). 1H NMR (CDCl3, ppm): 8.67
(2H, d, H4,5); 8.31 (2H, d, H20); 7.71 (2H, t, H30); 7.53 (4H,
t, H40 and H3,6); 7.43 (2H, t, H50); 7.12–7.23 (6H, m,
including H60 at 7.22, H2,7 at 7.19, H1,8 at 7.14); 7.01 (1H,
2.14. N-(4-Bromo-2,3,5,6-tetrafluorophenyl)maleimide
adduct (3c)
Similar to the procedure for (3b), phencyclone (114 mg.
0.298 mmol), excess crude maleimide (207 mg, nominal
0.639 mmol), and CH2Cl2 (ca. 15 mL) were combined in a
screw-cap vial and stirred overnight. The intense dark
green–black phencyclone color was discharged to give a
pale yellow supernatant and some solid. Partial solvent
removal, filtration, washings with ice-cold hexanes and
drying gave the crude adduct as a white solid (90.6 mg,
30.4%) mpr 260–262 8C (gas evoln. and dec.) IR (cmꢀ1):
1793.6 (strained ketone C=O); 1731.3; 1497.3; 1448.2;
1351.5; 1268.6; 1219.7; 1176.4; 978.2; 777.1; 756.0; 724.7;
698.6 (see Section 3 for further details of NMR). 1H NMR:
8.69 (2H, d, H4,5); 8.28 (2H, d, H20); 7.72 (2H, t, H30); 7.54
(4H, t, overlap of H3,6 and H40); 7.44 (2H, t, H50); 7.11–7.23
(6H, complex m, including H60 at 7.23, H2,7 at 7.20 and H1,
8 at 7.14); 4.73 (2H,s, 2 ꢁ sp3). 13C NMR: 195.65 (ketone
C=O); 171.12 (imide NC=O); four very weak multiplets at
ca. 146.3, 144.1, 143.0 and 140.6; 133.33 (Q); 133.04 (Q);
131.72 and 131.71 (tentative split peak, apparent J = 0.86,
Q); 130.96 (2 ꢁ CH); 129.43 (2 ꢁ CH); 128.77 (2 ꢁ CH);
128.71 (2 ꢁ CH); 128.58 (2 ꢁ CH); 127.21 (2 ꢁ CH);
126.56 (2 ꢁ CH); 126.04 (Q); 125.62 (2 ꢁ CH); 123.13 (2 ꢁ
CH); 109.90 (tentative, very weak possible t, apparent J ca.
16, possible C100); 102.02 (t, J ca. 22, C400); 63.53 (C6H5–C);
45.73 (2 ꢁ CH). 19F NMR: ꢀ131.68 (1F, ddd, J = 22.58,
9.11 and 4.20); ꢀ132.61 (1F, ddd, J = 22.67, 9.34 and 4.20);
ꢀ142.69 (1F, ddd, J = 22.67, 9.10 and 5.77); ꢀ143.03 (1F,
ddd, J = 22.53, 9.30 and 5.74).
3
3
d, J(HCCF) = 8.57, H300); 6.59 (1H, d, J(HCCF) = 8.08,
H500); 4.69 (2H, s, sp3). 13C NMR: 195.92 (ketone); 171.62
(imide); 157.50 (dd, C200); 157.23 (dd, C600); 133.56 (Q);
133.18 (Q); 131.72 (Q); 131.00 (C60); 129.40 (C30); 128.94
(C20); 128.66 (C50); 128.50 (C40 or 3,6); 127.03 (C3,6 or 40);
126.48 (C2,7); 126.21 (Q); 125.77 (C1,8); 123.37 (t, C100);
123.00 C4,5); 116.10 (C300); 115.69 (C500); 103.58 CBr);
63.55 (C6H5–C); 45.54 (2 ꢁ sp3). Two weak apparent dd
patterns appear at low field, ca. 159.1 and 155.6 ppm. 19F
NMR: ꢀ115.705 (1F, s); ꢀ115.816 (1F, s).
2.13. N-(2,3,5,6-Tetrafluorophenyl)maleimide adduct (3b)
The crude maleimide (2b), produced as above, was used
in excess. Thus, 904 mg (3.69 mmol) of the maleimide and
318 mg (0.832 mmol) of phencyclone were combined with a
magnetic stirbar in a small screw-cap vial with Teflon1-
lined cap, with sufficient CH2Cl2 (ca. 15 mL) to bring the
level of the reaction mixture to within 2–3 mm of the vial lip.
After tightly capping the vial and stirring at ambient
temperature for 3 days, the intense green–black color of the
phencyclone was discharged. The resulting light-colored
solid was collected by vacuum filtration, washed with
hexane and dried to give the crude adduct (471 mg, 90%
yield); a portion recrystallized from ethanol had mpr 262–
268 8C (dec., gas evoln.). IR (cmꢀ1): 1793.8; 1731.5;
1521.8; 1498.5; 1448.0; 1347.6; 1262.0; 1179.4; 940.2;
775.7; 753.4; 724.0; 698.3; 633.3 (see Section 3 for further
details of NMR). 1H NMR: 8.68 (2H, d, H4,5); 8.28 (2H, d,
H20); 7.71 (2H, approx. t, H30); 7.53 (4H, approx. t, H40 and
2.15. N-(2,3,4,5,6-Pentachlorophenyl)maleimide
adduct (3d)
As for (3b) above, in a small screw-cap vial with
Teflon1-lined cap was placed phencyclone (382.1 mg,
0.999 mmol), CH2Cl2 (ca. 15 mL) and a slight excess of the
pentachlorophenylmaleimide (prepared as above, 362.7 mg,
1.051 mmol). The vial was capped and stirred at ambient

1898
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
temperature for 3 days. The intense dark green–black
phencyclone color had decolorized to give a light yellow
solution. After partial solvent removal, the resulting solid
was collected, washed with cold CH2Cl2 and dried to give
the crude tan adduct (680.2 mg, 93.4% yield), mpr 267–
270 8C (early shrinkage, gas evoln. and dec, see Section 3
most of the compounds prepared here are presented below in
this Section 3. The discussions of the selected compounds
are arranged in the same order as for their syntheses in the
Section 2, with the maleamic acids (5), first; followed by the
maleimides (2); and the phencyclone adducts (3), last.)
The desired maleimides (2), are generally accessible
according to Fig. 1. The top of Fig. 1 shows the reaction of
maleic anhydride with the appropriately substituted anilines
(4a–4e). (Specific substitution patterns of the aniline N-aryl
rings are shown at the lower left of the Figure.) In the more
standard two-step approach, the substituted aniline (4), is
condensed with maleic anhydride at mild temperatures to
give the corresponding maleamic acids, e.g. (5a–5c), as
isolated intermediates. A separate cyclodehydration step, as
with anh. sodium acetate in acetic anhydride at ca. 90–
100 8C, converts the maleamic acids (5a–5c), to the
maleimides (2a–2c). With anilines of low basicity or
reduced reactivity, direct reaction at high temperature (ca.
180–200 8C) of, e.g. (4d) or (4e), with maleic anhydride
(without solvent, in the presence of an acid catalyst, e.g.,
anh. ZnCl2) can give (2d) or (2e) directly, in one step,
without isolation of the maleamic acids (5). In the two-step
method, using NaOAc/Ac2O for the cyclodehydration, we
frequently found a common impurity in the crude product
maleimides. This impurity (6) (shown at the bottom right of
Fig. 1), characterized by NMR, would formally correspond
to an acetic acid addition product of (2), an acetoxy-
pyrrolidinedione [16]. In the absence of acetic anhydride/
sodium acetate, other nucleophiles could potentially add to
maleimides to give corresponding pyrrolidinediones.
Infrared (IR) spectra of two representative phencyclone
adducts were obtained to confirm the presence of the sharp
band ca. 1794 cmꢀ1, attributed to the strained ketone
carbonyl stretch. This band would imply that undesired
decarbonylation of the C=O has not occurred during the
reactions or workup procedures. Phencyclone adducts at
elevated temperatures can potentially undergo decarbonyla-
tion, retro-Diels–Alder reactions, etc. We did not see NMR
spectral evidence for the potential stereoisomer, i.e., the exo
adducts, in our isolated products. Recent X-ray studies of
some phencyclone adducts have confirmed the endo
configuration normally preferred for Diels–Alder adducts
[17–19].
1
for further details of NMR). H NMR: 8.69 (2H, d, H4,5);
8.32 (2H, d, H20); 7.72 (2H, approx. td, H30); 7.56 (H3,6)
overlapped with 7.54 (H40) to give a 4H approx. t; 7.43 (2H,
approx. td, H50); 7.14–7.25 (6H, m, including H2,7 at 7.21,
H60 at 7.20, and H1,8 at 7.16); 4.74 (2H, s, 2 ꢁ sp3). 13C
NMR: 195.31 (ketone C=O); 171.41 (imide NC=O); 135.52
(very weak, Q, CX); 134.86 (very weak, trace impurity of
maleimide HC=CH); 133.425 (2Q); 133.402 (2Q); 132.84
(very weak, Q, CX); 132.37 (very weak, Q, CX); 131.91
(very weak, Q, CX); 131.73 (2Q); 131.40 (very weak, Q,
CX); 130.94 (2 ꢁ CH); 129.42 (2 ꢁ CH); 128.93 (2 ꢁ CH);
128.65 (2 ꢁ CH); 128.55 (2 ꢁ CH); 128.27 (very weak, Q,
CX); 127.33 (2 ꢁ CH); 126.74 (2 ꢁ CH); 126.57 (2Q);
125.91 (2 ꢁ CH); 122.99 (2 ꢁ CH); 63.59 (C6H5–C); 45.61
(2 ꢁ sp3 CH).
2.16. N-(2,4,6-Tribromophenyl)maleimide adduct (3e)
As for the preparation of (3d), phencyclone (382.0 mg,
0.999 mmol), a slight excess of the tribromophenylmalei-
mide (prepared as above, 430.7 mg, 1.051 mmol), and
CH2Cl2 were stirred at ambient temperature for 7 days in a
small capped vial. The intense dark green–black phency-
clone color was largely discharged to give a light beige
solution. After partial solvent removal, the resulting solid
was collected, washed with cold CH2Cl2 and dried to give
the crude pale tan adduct (706.5 mg, 89.0% yield), mpr 224–
234 8C (early shrinkage, gas evoln. and dec, see Section 3
1
for further details of NMR). H NMR: 8.63 (2H, d, H4,5);
8.36 (2H, d, H20); 7.72 (2H, td, H30); 7.64 (1H, d, 4J = 1.98,
‘‘outer’’ H of N-aryl); 7.53 (4H, approx. t, overlap of H3,6
and H40); 7.42 (2H, approx. td, H5); 7.16–7.24 (7H, complex
m, including ‘‘inner’’ H of N-aryl at 7.23; H2,7 at 7.21; H60
at 7.20; H1,8 at 7.19); 4.71 (2H, s, 2 ꢁ CH, sp3). Traces of
maleimide and pyrrolidinedione impurities appeared to be
present. 13C NMR: 195.36 (ketone); 171.65 (imide); 134.90
(1 ꢁ CH); 134.43 (1 ꢁ CH); 133.64 (2Q); 133.55 (2Q);
131.85 (2Q); 130.93 (2 ꢁ CH); 130.45 (1Q); 129.39 (2 ꢁ
CH); 129.09 (2 ꢁ CH); 128.59 (2 ꢁ CH); 128.48 (2 ꢁ CH);
127.21 (2 ꢁ CH); 126.86 (2Q); 126.69 (2 ꢁ CH); 126.14 (2
ꢁ CH); 124.43 (1Q); 124.40 (1Q); 123.48 (1Q); 122.74 (2 ꢁ
CH); 63.58 (C6H5–C); 45.53 (2 ꢁ CH, sp3).
The presence of fluorine in a molecule potentially
1
complicates the H and 13C NMR spectra because of the
spin–spin couplings, J(HF) or J(CF). For 13C NMR, this
splitting by 19F can be especially problematic because the
splitting leads to lower signal-to-noise ratios for each
component of the resulting carbon multiplet signal;
particularly for [normally weak] signals of non-protonated
carbons, signal detection is more difficult. (Our NMR
spectrometer system does not allow for fluorine decoupling.)
Adding to this problem was the rather low CDCl3 solubility
of most of the adducts. (Note that precursor maleamic acids
also exhibited low CDCl3 solubility; d6-acetone was often a
better solvent.)
3. Results and discussion
3.1. Syntheses and NMR interpretation
(Note that specific discussions, where appropriate, of
NMR interpretation (and rationales for assignments) for

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1899
The key NMR experiments for evaluating bridgehead
phenyl rotations were the high resolution two-dimensional
(2D) COSY45 spectra for homonuclear (1H–1H) chemical
shift correlation. Under our conditions, routine detection of
crosspeaks for 3J, 4J and even 5J couplings in the
phenanthrenoid and bridgehead phenyl systems was gen-
erally possible. It was possible to map out the (CH)4 and
(CH)5 spin systems for the phenanthrenoid and phenyl
systems, respectively, allowing full proton assignments even
when numerous signal overlaps occurred in the one-
timescale), attributed to hindered rotation about the
N(sp2)–C(aryl sp2) bond due to repulsions between the
imide carbonyl oxygens and the ortho substituents (i.e.,
200,600 attachments) on the N-aryl. (Adduct structures and
atom numbering are shown in Fig. 1.) These repulsions force
the N-aryl to twist away from coplanarity with the imide
system of the pyrrolidinedione ring. The equilibrium
dihedral angle between the pyrrolidinedione and N-aryl
ring can approach 908 for bigger 200,600-substituent groups on
the N-aryl. This conformation results in one edge (the b
edge) of the N-aryl ring being directed ‘‘into’’ the cavity of
the adduct (3), and therefore ‘‘towards’’ the phenanthrenoid
moiety. Substituents on these inner positions might
experience buttressing effects and magnetic anisotropic
shielding from the proximal phenanthrenoid. Substituents
on the opposite, or outer edge (the a edge) of the N-aryl ring
are directed ‘‘away’’ from the phenanthrenoid, and would
not experience these effects; the chemical shifts of
substituents aimed ‘‘out of’’ the adduct cavity would show
relatively normal NMR chemical shifts.
1
dimensional (1D) H spectra. With slow phenyl rotation
on the NMR timescale, five 2H intensity signals are expected
for the two bridgehead phenyls, and four 2H intensity signals
for the phenanthrenoid moiety (in the absence of accidental
overlaps) in the aryl proton regions of adducts of (1). (If
protons are present on the N-aryl rings of the adducts,
additional signals or overlaps would be expected.) If the
bridgehead phenyls rotated rapidly on the NMR timescale,
the resulting fast exchange limit (FEL) spectra would show
one 4H signal for the ortho phenyl protons, one 4H signal for
the meta protons, and one 2H signal for the para protons.
The above analyses assume that the adducts (3), exhibit an
effective mirror symmetry plane on the NMR timescale. For
the 1D 13C NMR spectra with proton decoupling, the SEL
regime for bridgehead phenyl rotation would predict signals
for nine pairs of aryl methines, including five from the
phenyls and four from the phenanthrenoid, plus signals of
four pairs of non-protonated (quaternary, Q) aryl carbons
from these groups. Potential complications and additional
aryl carbon signals result from the N-aryl groups and
possible splitting by 19F. Using a standard 3 s relaxation
delay (RD) for the 1D 13C spectra resulted in peak areas
from signals of protonated carbons being roughly propor-
tional to the numbers of these carbons. Carbon-13 NMR
spectra can directly support slow bridgehead phenyl
rotations as well as slow rotations for the N-aryl ring.
Assignments for the protonated carbons can be straightfor-
ward by use of the 2D heteronuclear chemical shift
correlation spectra (HETCOR) in the absence of accidental
proton signal overlaps, but the non-protonated aryl
(quaternary, Q) carbons may not be readily assigned. If
severely split by fluorines, these non-protonated aryl carbons
may not always be observable above baseline noise. In some
cases, we have made tentative assignments based on
chemical shift arguments or expected C–F coupling
constants. In compounds with fluorine, 19F NMR spectra
were acquired, first with coarse digital resolution with CFCl3
as internal reference, and then with fine digital resolution
(narrow spectral width) to resolve complex fluorine signal
multiplicities.
3.2. N-(4-Bromo-2,6-difluorophenyl)maleamic acid (5a)
Note that the fluorine NMR for the N-(4-bromo-2,6-
difluorophenyl)maleamic acid showed an apparent broad-
ened doublet when acquired with proton decoupling (see
Fig. 2a). This doublet was incompletely resolved, with
valley height about 64% of the average height of the two
peaks. The full width of the doublet measured at half-height
(i.e., somewhat below the valley) was ca. 14.9 Hz. If this
maleamic acid exhibited fast rotation about the N-aryl bond,
the two fluorines should be rendered chemically equivalent
(isochronous). With decoupling of protons, a singlet should
be observed for the 19F signal. With N-aryl amides, two
possible important bond rotation processes may be envi-
sioned [20–22]. One process involves rotation about the N–
C=O bond, which would interconvert syn and anti amide
conformers. The second process would be the N-aryl
rotation. The former process, N–C=O rotation, is generally
found to be slow at ambient temperatures on typical proton
or carbon-13 NMR timescales. In the 1H NMR for the crude
maleamic acid, only a single set of sharp signals is observed
for the AB q of the HC=CH portion, implying just one
predominant amide conformer. (If both syn and anti amide
conformers were present in appreciable amounts under a
slow exchange limit (SEL) regime, two sets of AB q signals
would be expected.) This implies that N–C=O rotation is not
a factor in the observed pair of broad lines in the fluorine
NMR. If N-aryl rotation was slow and the aryl group was
effectively near-coplanar with the amide group, the two
fluorines are non-equivalent. They could produce two
singlets in the 19F NMR if the 4J(FCCCF) magnitude is very
small, or an AB q pattern (approaching a pair of doublets,
depending on the difference in chemical shifts of the
fluorines) if the 4J coupling is appreciable. With inter-
mediate N-aryl rotation rates, chemical exchange broad-
With highly fluorinated N-aryl rings in the adducts (3),
19F–19F COSY spectra have been employed to rigorously
identify vicinal pairs of fluorines. With symmetrical 2,3,5,6-
tetrafluoro substitution on the N-aryl ring, observation of
four separate fluorine signals in the 19F NMR shows slow
rotation of the N-aryl group (on the fluorine NMR

1900
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
Fig. 2. Fluorine-19 NMR spectra, 282 MHz, at ambient temperatures. Chemical shifts are in ppm relative to CFCl3 at 0.0 ppm. Spectra were acquired with
proton decoupling. For the maleamic acids, spectra were recorded in d6-acetone. Trace (a) (top left) is N-(4-bromo-2,6-difluorophenyl)maleamic acid
(compound (5a)). Traces (b and c) (upper middle and right) show the spectrum for N-(2,3,5,6-tetrafluorophenyl)maleamic acid (compound (5b)). The spectrum
of compound (5c), N-(4-bromo-2,3,5,6-tetrafluorophenyl)maleamic acid is seen in traces (d and e) (bottom left and middle). Spectral traces (b and c) are
displayed with identical horizontal and vertical scales and represent equal area peaks; this applies also to traces (d and e). The symmetrized 19F–19F COSY45
spectrum, f (bottom right), corresponds to the phencyclone adduct (3b), of N-(2,3,5,6-tetrafluorophenyl)maleimide (compound (5b)), in CDCl3. The spectral
window ran from ꢀ137 to ꢀ145 ppm, using two dummy scans and four transients at each of 64 t1 increments.
ening can be expected, which might obliterate F–F splittings
or the 19F chemical shift differences (Dn), depending on the
kinetic rate and the magnitudes of couplings and Dn. We
shows that: (a) full proton decoupling was effective, and (b)
the downfield doublet assigned to the maleamic acid was
abnormally broadened.] We also point out that the aryl
proton NMR spectrum shows a sharp resonance for the
believe that this is the explanation for the broad doublet 19
NMR signal.
F
1
meta aryl protons, H3,5. We believe that the H NMR for
[A trivial alternative explanation is possible; if proton
decoupling were incomplete, then even with fast N-aryl
rotation, the fluorine resonance could be a gross doublet due
these protons is effectively FEL on the proton timescale,
with the chemical shifts of these protons essentially sharply
averaged, and the observed near-doublet signal showing
spacing consistent with an approximate vicinal 3J H–F
coupling. This may be deceptively simple, since the
maleamic acid’s bromodifluorophenyl ring is an AA0XX0
spin system. The aryl proton spectrum can be in the FEL
3
to vicinal J(HCCF) coupling. This is not the case, as is
indicated by the following. The proton NMR of the crude
maleamic acid showed the presence of about 11% of the
starting material, 4-bromo-2,6-difluoroaniline, based on the
presence of the sharp complex multiplet, ca. 7.05–
7.15 ppm, for the absorption signal of the aryl protons of
this material. (This aryl proton multiplet was found to be
identical to a reference 1H NMR spectrum in d6-acetone for
the pure starting aniline.) In the fluorine NMR of the
crude maleamic acid, a sharp singlet at ꢀ130.66 ppm is
seen, ca. 12% of the area of the lower field doublet. This
singlet may be assigned to the bromodifluoroaniline; the
width at half-height was only 2.8 Hz. For the fluorine
resonance of the aniline to be such a sharp singlet clearly
regime, showing a sharply averaged multiplet, while the 19
F
NMR may only be at an intermediate exchange regime, if
there is a small Dn for the protons and a larger Dn
magnitude for the fluorines.
3.3. N-(2,3,5,6-Tetrafluorophenyl)maleamic acid (5b)
The fluorine-19 NMR spectrum appears to be a striking
example of a dynamic NMR spectrum with an intermediate
exchange rate. (Throughout this paper, we use the term

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1901
‘‘dynamic NMR’’ in its broadest sense for describing our
studies, that is, use of NMR spectra to gain insight regarding
possible kinetic rate processes, to judge whether a particular
system might be characterized as exhibiting an SEL or FEL
spectrum, or [of special interest] whether the spectrum
suggests a system at some intermediate exchange rate, as
suggested by significant signal broadening. Thus, simple
observation of the numbers of spectral absorption signals,
and their integrated areas, may directly suggest, e.g., aryl
ring rotation rates as being slow, fast or intermediate [on the
respective NMR time scales] even without explicit use of
variable-temperature studies or line-fitting analyses aimed at
explicit determination of process rates or barriers to
interconversions.) For the specific case of (5b), the
fluorine-19 NMR is shown in Fig. 2b and c. Two equal-
area signals are seen, with one being a sharp apparent
quartet, and the other being a broad, featureless signal. Our
explanation is that the favored conformation can be
visualized as having the N-aryl roughly coplanar with the
amide and alkene moieties, so that the environments of F2
and F6 are not the same, the F3 and F5 pair also being non-
equivalent. Fast N-aryl rotation would render the ortho F2
and F6 chemically equivalent (isochronous), and the meta F3
and F5 pair would also be a chemically equivalent
(isochronous) pair; two sharply exchange-averaged multi-
plets would be expected (one for each pair of fluorines).
With N-aryl rotation at the slow exchange limit, four discrete
sharp multiplets might be seen. An intermediate rate for the
N-aryl rotation can cause partial averaging and broadening
of the splittings and chemical shift differences. Note that the
1H NMR spectrum for the para proton of the tetrafluor-
ophenyl ring shows a sharp tt pattern, and we believe that
there is only one significant amide conformer present. This
proton lies on the rotation axis for the N-aryl rotation, so its
chemical shift would not be affected by the aryl ring
rotation. Splitting of the proton by the four fluorines is
potentially complex, since the system would be considered
AA0MM0X, but moderate N-aryl rotation rates may be
sufficient to sharply average the H–F couplings, to give a
deceptively simple tt pattern expected from an A2M2X
system.
(polyhalophenyl) maleamic acids, encompassing three
different fluorine substitution patterns (2,6; 2,3,5,6; and
2–6), all of which exhibit fluorine-19 NMR signals
broadened to a greater or lesser extent. That these are all
N-aryl amides, for which hindered N-aryl rotations are
known to be potentially significant on various NMR
timescales, would seem to be very suggestive that we are,
indeed, observing the effects of intermediate exchange rate
19F NMR spectra for these maleamic acids. All four of these
maleamic acids possess fluorines at both the 2- and 6-
positions [i.e., ortho to nitrogen] of the N-aryl rings, where
the effects on the N-aryl rotation rates would most largely be
determined, and we would therefore suspect that the four
compounds should have similar energy barriers to the N-aryl
rotations. This would be fully consistent with the fact that
NMR line broadenings for all four maleamic acids were
observable at the same temperature, ca. 298 K, on our
NMR.)
3.4. N-(4-Bromo-2,3,5,6-tetrafluorophenyl)maleamic
acid (5c)
The fluorine NMR spectrum for the maleamic acid is
consistent with an intermediate exchange rate for N-aryl
rotation, causing some exchange-broadening, much more
severe in the higher field absorption (see Section 3 above for
(5b)) The fluorine-19 NMR is shown in Fig. 2d and e.
3.5. N-(2,3,4,5,6-Pentachlorophenyl)maleimide (2d)
This compound’s carbon-13 NMR was obtained using a
relaxation delay (RD) of 60 s with 1230 transients acquired.
The long RD provided enhanced response for the weak
signals of the non-protonated aryl carbons (relative to the
protonated vinyl carbons). The two signals at 135.85 and
127.95 ppm were each about half the area of the signals at
133.82 and 132.52 ppm, and we attribute the weaker pair of
signals to the single carbons C1 (N–C, ipso) and C4 (para),
with the higher field signal tentatively assigned to the N–C
carbon. Note that the ipso and para carbons of an aryl ring
are often found to exhibit shorter T1 relaxation times than the
ortho or meta carbons (if similarly substituted), because the
ipso and para carbons, lying on the rotation axis of the aryl
ring, have slower effective motions, longer correlation
times, and (for relatively small molecules) shorter T1. The
ortho or meta carbons, lying off the rotation axis of the aryl
ring, would have faster effective motions and shorter
correlation times [23]. However, this would only be
applicable for aryl rings that are rotating rapidly. For the
N-aryl rings in the maleimides and phencyclone adducts
described in this paper, the point is that these rings are
essentially prevented from fast rotation about the N-aryl
bonds because of the ortho substituents. If this is the case, we
would not expect appreciable T1 differences for on-axis
versus off-axis N-aryl ring carbons (due to this effect), and
carbon-13 NMR peak areas might more closely reflect the
(We note that a referee has suggested that NMR line
broadening may not necessarily be an indication of an
intermediate exchange-rate process, such as a hindered
rotation, but may reflect other factors affecting relaxation
rates or resulting from extensive unresolved coupling.
However, we strongly favor intermediate exchange rates of
the N-aryl rotations for the maleamic acids (5a), (5b) and
(5c) (discussed below) as the most reasonable explanation in
the present case, since line-broadening in the 19F NMR of
the maleamic acids is seen here for three different N-aryl
ring substitution patterns (2,4,6; 2,3,5,6; and 2,3,4,5,6). If
the earlier case of N-(pentafluorophenyl)maleamic acid is
also considered [8,12], for which varying line broadening of
signals in the 19F NMR spectrum was observed when
temperature was varied, we would have four different N-

1902
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
numbers of each type of carbon [23]. We make the
presumption of slow N-aryl rotation in the maleimides, as
well as in the adducts, because the expected interatomic
repulsions of ortho N-aryl ring substituents with the imide
carbonyls should be grossly similar for the maleimides and
the adducts. In the maleimides, however, since the two faces
of the maleimide ring are identical, we cannot observe
different NMR chemical shifts for the two ortho (C2,6)
positions of the N-aryl ring (and likewise for the two meta
(C3,5) positions). In the phencyclone adducts, slow rotations
of the N-aryl rings are readily demonstrable because the two
faces of the pyrrolidinedione ring are quite different from
one another.
as eight equal lines, with a central double doublet and a pair
of flanking doublets. We cannot be certain whether this
represents a pair of adjacent dd patterns or whether the dd
patterns overlap. These are assigned to the C300 and C500
carbons on the N-aryl ring, with long range 4J(CF) coupling
ca. 3.9 Hz, and 2J(CCF) couplings ca. 22.8 Hz (or 31.1 Hz).
Using the smaller values for the geminal couplings, and
assigning higher field position to the ‘‘inner’’ carbon, C500,
we would have C500 at 115.69 ppm and the ‘‘outer’’ C300 at
116.10 ppm. Protonated carbons on the bridgehead phenyls
and the phenanthrenoid groups are rigorously known from
the HETCOR spectrum, which shows nine crosspeaks for
the aryl methine pairs. We could not observe the crosspeaks
for the C300 and C500 methines, which absence we attribute to
inadequate signal-to-noise ratios as a result of the splitting of
the signals of these carbons into dd patterns by the fluorines.
The proton decoupled fluorine spectrum showed two
sharp singlets, separated by 0.11 ppm, with peak widths at
half-height ca. 3.1 Hz, implying negligible 4J(FF) coupling
magnitude.
3.6. N-(2,4,6-Tribromophenyl)maleimide (2e)
Some clarification for carbon assignments was obtained
by acquiring the proton coupled spectrum with RD = 60 s in
addition to the usual proton decoupled spectrum. Carbon
signal multiplicities and observed splittings are shown in the
Section 2. Assuming that the magnitudes for the aryl ring
3
vicinal J(CH) couplings with ca. 1808 dihedral angles are
greater than the geminal J couplings in these systems, we
3.8. N-(2,3,5,6-Tetrafluorophenyl)malemide adduct (3b)
2
have made the indicated tentative assignments [21]. Signals
attributed to the individual non-protonated aryl carbons C1
and C4 had areas about half or less than those assigned to the
pair of ortho CBr carbons.
Several uncertainties in the 13C NMR spectrum assign-
ments may be attributed to substantial splitting by fluorine,
with the resulting high multiplicity for the carbon signals
causing poor signal-to-noise ratios, despite acquisition of
more than 15,000 accumulated transients. These uncertain
features include the following: (a) The weak complex
multiplets ca. 147.2, 143.8 and 140.4 ppm appear roughly as
a 1:2:1 triplet, with spacings between adjacent pairs of these
multiplets of ca. 253 and 261 Hz. These spacings are
consistent with the direct 1J(CF) couplings for aryl fluorides
[23], and we believe that the apparent triplet pattern is the
result of overlap of two gross doublets for fluorine-bearing
carbons; the centers of these doublets would be 145.5 and
142.1 ppm, assigned to either the ortho or the meta pair of
CF carbons of the tetrafluorophenyl group. This would
account for two of the four aryl CF groups; we are not
confident of the location for the other pair. (b) The signal at
135.24 ppm is exceedingly weak, and we have assigned it to
the HC=CH absorption from a trace impurity of the
precursor maleimide. A small sharp singlet is seen in the
proton NMR of the adduct at ca. 6.94 ppm, attributable to the
HC=CH of the maleimide. [The absence of a corresponding
peak in the carbon spectrum of the adduct of the brominated
analog described below would suggest that this weak peak
is, in fact, not part of the adduct.] Only four of these aryl Q
carbon signals are expected for the bridgehead phenyl and
phenanthrenoid moieties, and related phencyclone adducts
[from N-phenylmaleimide and N-n-alkylmaleimides] have
generally shown three signals at relatively low field, i.e.,
131–134 ppm, with a higher field Q signal at ca. 126 ppm,
and peaks of seven of the nine intense aryl methine pairs
appearing within the 126–131 ppm window [10,24]. (c) Our
sample showed two absorptions consistent with these aryl Q
3.7. 4-(Bromo-2,6-difluorophenyl)maleimide adduct (3a)
In the proton NMR, triplets for H40 and for H3,6 were
essentially isochronous, giving a 4H t signal. A complex 6H
multiplet from 7.12 to 7.23 ppm included H60, H2,7 and
H1,8. Approximate shifts were estimated from the COSY
spectrum. Note that higher field 1H doublets were seen at
7.01 and 6.59 ppm, assigned to the ‘‘outer’’ H300 and the
‘‘inner’’ H500, respectively.
In the 13C NMR, two weak apparent dd patterns appear at
low field, ca. 159.1 and 155.6 ppm. Fluorine is strongly
deshielding for directly bonded aryl carbon. We tentatively
assign these signals to the CF carbons at the 200 and 600
positions of the N-aryl ring, with each carbon showing the
large 1J(CF) direct coupling and a smaller 3J(CCCF)
coupling. The most consistent interpretation of the observed
couplings suggests the following assignments, with the
higher field carbon being designated as the ‘‘inner’’ carbon,
C600, on the ‘‘b’’ edge of the N-aryl. C200 resonates at
1
3
157.50 ppm, J = 258 Hz, J = 4.8 Hz. C600 resonates at
1
3
157.23 ppm, J = 262.8 Hz, J = 5.1 Hz. The very weak
singlet at 103.58 ppm is assigned to C400, the CBr, shielded
by the heavy-atom effect of the directly bonded bromine. An
extremely weak approximate triplet is centered at
123.37 ppm, assigned to the ipso carbon, CN, C100, showing
2
geminal J splitting by the two fluorines of ca. 11.2 Hz. A
strong set of peaks is centered at ca. 115.9 ppm, with total
area consistent with two methines. The absorption appears

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1903
signals ca. 131.7 and 126.1 ppm, each of which appeared as
a doublet with slightly split peaks; the apparent splittings
were 1.35 Hz for the 131.7 ppm signal and 0.86 Hz for the
126.1 ppm signal. If these observed splittings were real, and
not artifactual, we could not account for them as being the
result of through-bond coupling (J coupling, spin–spin
coupling) with a fluorine on the N-aryl group; the fluorines
are just too many bonds removed. We have speculated that
this might represent a through-space dipolar coupling effect
resulting from the unusual proximity of the ‘‘inner’’ ortho
fluorine to quaternary carbons of the phenanthrenoid system,
but this is not certain. (d) An extremely weak multiplet was
seen at 111.1 ppm as a possible triplet of multiplets; the
spacing between the three branches was about 15 Hz, which
could be consistent with 3J(FCCC) or 2J(FCC) for the
tetrafluorophenyl ring; we might tentatively assign this as
the N-aryl C100. The one carbon assignment for the C6F4H
ring which is firm is for the protonated carbon, C400,
appearing as a clean triplet with J = 22.6 Hz, exhibiting peak
area consistent with a single methine. Note that the simple
triplet appearance means that the carbon is split by only one
pair of fluorines.
Four separate sharp multiplets for the 19F absorptions are
consistent with the SEL N-aryl rotation causing different
chemical shifts for the ‘‘inner’’ versus the ‘‘outer’’ positions
(relative to the cavity of the adduct). Two dd signals
appeared at lower field, ca. ꢀ138.2 ppm, and two ddd
multiplets (each consisting of eight near-equal intensity
lines) appeared at higher field, ca. ꢀ144.4 ppm. If the
nitrogen is shielding for ortho fluorines (as it is for ortho
protons), the high field ddd multiplets would be F2 and F6.
The 19F–19F COSY45 spectrum, seen in Fig. 2f, showed
stronger crosspeaks, suggesting vicinal orientation, for the
lower field dd signal with the higher field ddd absorption,
and a comparable crosspeak correlated the higher field dd
resonance with the lower field ddd signal. Thus, any effect
from the phenanthrenoid moiety on the fluorines in the
cavity of the adduct appears to have opposite influence on
the vicinal pair of fluorines, moving their resonances in
opposite directions. See below for discussion of the spectra
for the brominated analog with the 4-bromo-2,3,5,6-
tetrafluorophenyl ring.
interpreted as a pair of doublets with 1J(CF) ca. 250–260 Hz.
If the four multiplets in the bromotetrafluoro analog are
considered to be overlapping doublets, approximate direct
one-bond C–F couplings of ca. 250 Hz (for a doublet
centered at ca. 144.8 ppm) and ca. 260 Hz (for a doublet
centered at ca. 142.4 ppm) can be estimated. Interpreting the
four multiplets as two adjacent doublets would imply
approximate direct one-bond C–F couplings of ca. 166 and
176 Hz, which are less likely [21]; we favor the former
rationale. These two gross doublets, then, could be assigned
to a pair of non-equivalent carbons on the C6BrF4 ring, either
the ortho C200 and C600, or the meta C300 and C600. If these two
doublets represent one of these carbon pairs, the signals for
the other pair of fluorine-bearing carbons are not obvious, as
was the case for the non-brominated C6F4H ring in the analog
described earlier. (b) In the case of the C6BrF4 ring, as was
seen for the non-brominated C6F4H ring in the preceding
analog, one of the weak Q signals appeared as a slightly split
peak, ca. 131.7 ppm. We had tentatively postulated that this
might result from a through-space dipolar C–F coupling. In
the less hindered 4-bromo-2,6-difluorophenyl system, no
such splittings of any quaternary aryl carbon signals were
seen, suggesting that such splitting may require very
crowded, ‘‘buttressed’’ systems, i.e., the two tetrafluorophe-
nyl systems. (c) At higher field for aryl carbon absorptions, a
few exceedingly weak peaks were observed ca. 109.9 ppm,
which we tentatively ascribe to the aryl CN, C100, appearing
as a possible triplet. (d) A more secure assignment is for the
triplet at 102.02 ppm (J ca. 22 Hz). We believe that this is the
signal for CBr, C400, based (in part) on the expected shielding
by bromine for directly bonded carbon; the observed splitting
2
could be either J or 3J.
In the 19F NMR spectrum, the four sharp, highly resolved
fluorine multiplets confirm that the N-aryl rotation is at the
SEL, and that each fluorine is in a distinct environment.
Because of similarities of coupling constants, we assume
that the lower field pair of signals arise from one pair of
fluorines, either the ortho 200, 600 or the meta 300, 500 pair, with
the higher field signals assigned to the other pair. The
homonuclear fluorine–fluorine COSY45 experiment was
performed on this compound. Strong crosspeaks, signifying
vicinal couplings, were seen between the lowest field and the
highest field absorptions, i.e., ꢀ131.7 and ꢀ143.0 ppm (and
also between ꢀ132.6 and ꢀ142.7 ppm signals), as was the
case for the non-brominated tetrafluorophenyl analog
described above. For both of these compounds (with similar
fluorine substitution patterns), we tentatively suggest that
the lower field pair of fluorine-19 NMR signals may be
assigned to the ortho pair of fluorines at the 200 and 600
positions because they exhibit a larger Dd, ca. 0.9 ppm, than
the higher field pair of signals which exhibit a magnitude of
Dd ca. 0.3–0.4 ppm. It would be reasonable that a larger
difference in environments for ‘‘inner’’ versus ‘‘outer’’
positions on the N-aryl ring would be experienced by the
ortho than by the meta (300, 500) fluorines, and this could
express itself in the chemical shift differences.
3.9. N-(4-Bromo-2,3,5,6-tetrafluorophenyl)maleimide
adduct (3c)
Over 17,000 transients were acquired for the carbon-13
NMR spectrum of this adduct. Nevertheless, extensive
splitting of carbon signals by fluorine caused difficulty in
assignments for the halogenated aryl ring. Several salient
points: (a) In the 140–147 ppm region, four roughly equal
intensity complex multiplets were recognizable, potentially
being either two adjacent gross doublets or two overlapped
gross doublets. For the N-(2,3,5,6-tetrafluorophenyl)malei-
mide adduct described above, an approximate 1:2:1 triplet
was seen in a similar chemical shift region, and was

1904
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
3.10. N-(2,3,4,5,6-Pentachlorophenyl)maleimide
adduct (3d)
each of the protonated aryl methines had an average peak
area of 90.25 integral units (per carbon). Using RD = 60 s
and NS = 3954, with the protonated aryl methines exhibiting
an average peak area of 90.25 integral units per carbon for
their signal peaks, the quaternary pairs now showed average
peak areas of 54.89 integral units (per carbon) but the six
single carbon signals of the NC6Cl5 ring were only 36.18
integral units (range 28.34 for the peak at 128.26 ppm, to
48.27). This implies exceedingly long relaxation times for
the NC6Cl5 carbons, even when compared to the other
quaternary pairs. (Note that these samples were not degassed
in any way.) Thus, slow rotation of the NC6Cl5 ring on the
carbon-13 NMR timescale is demonstrated.
With the 4-bromo-2,6-difluorophenyl, 2,3,5,6-tetrafluor-
ophenyl, and 4-bromo-2,3,5,6-tetrafluorophenyl as the N-
aryl moieties in the phencyclone adducts, a qualitative sense
of the N-aryl rotation rates can potentially be obtained by
NMR of a sensitive nucleus, either proton (for adduct (3a))
or fluorine-19 for adducts (3a), (3b) or (3c). However, for the
pentachlorophenyl system of adduct (3d), the problem is
more difficult, since the N-aryl ring has no sensitive nucleus
attached. It is still possible to answer the question if carbon-
13 NMR is used. This requires a rather pure sample of (3d)
to avoid impurities with interfering carbon signals, and a
very high number of scans (NS) to obtain adequate signal
intensity to detect the expected exceedingly weak responses
from the non-protonated carbons. We used up to 27,000
scans with a 1 s relaxation delay (RD) and compared the
resulting spectrum to that acquired with RD = 60 s (NS =
3954). Except for a trace peak of the precursor maleimide
(HC=CH signal at 134.86 ppm), the 13C spectra were very
clean, and have allowed, we believe, firm identification of all
six carbon signals for the NC6Cl5 ring. For each of the six
single non-protonated carbon signals expected for the SEL
spectrum of the adduct, weak (but similar intensity) signals
were seen. Five of the six signals were at even lower field
than the lowest field peak (at 130.94 ppm) of the methine
pairs, consistent with direct attachment of an electronegative
atom, i.e., chlorine. The sixth of these weak peaks was seen
at 128.27 ppm, possibly the ipso N–C, C100. With RD = 1 s
and NS = 27,157, the average intensity of each of these six
weak peaks was 7.81 integral units (range from 5.96 to 9.49)
compared to an average of 22.67 integral units per carbon
(range from 13.45 to 32.80) for the quaternary pairs of
carbons from the bridgehead phenyls and phenanthrenoid
moieties, i.e., C10, C4a,4b; C8a,10a; and C9,10. This
suggests extremely long relaxation times for the NC6Cl5
carbons, consistent with their considerable distance from
protons, which might assist in relaxation. For comparison,
3.11. N-(2,4,6-Tribromophenyl)maleimide
adduct (3e)
The carbon spectra were acquired with RD = 1 s (NS =
20,328) and 60 s (NS = 2670) to aid in distinction and
assignments of the single methine and the non-protonated
quaternary carbons. Thus, with the short 1 s relaxation delay,
the areas of protonated carbon signals are quite closely
proportional to their numbers, while non-protonated carbon
signals are much weaker; even the pairs of aryl Q carbons
have areas that are barely half that for a single CH. With RD =
5 s, the strongest signal for a quaternary pair, 2Q, at
133.55 ppm, actually has area greater than either single
methine peak. But the four 1Q signals of the N-aryl ring each
still have substantially less area than the peaks of the 2Q
signals. With the long RD = 60 s (NS = 2670), enhanced
response for quaternary aryl carbons of the adduct provided
very favorable signal-to-noise ratio, so that the required 19
aryl carbon signals are unambiguously defined: nine methine
pairs and four quaternary pairs from the bridgehead phenyls
and the phenanthrenoid moieties, plus two unique methines
and four single quaternary carbons from the brominated
N-aryl. We have clearly demonstrated slow rotations and
SEL spectra on both proton and 13C NMR timescales for both
the bridgehead phenyls and the N-aryl ring.
Table 1
Calculated parameters for phencyclone adducts of polyhalophenylmaleimides
Adduct
Optimized
energies (au)a
Dihedral angle
O=C–C–C–C20
Dihedral angle
O=C–N–C–C600 (inner)c
Distance of N-Ar C600
substituent to phenanthrenoidd
N pyramidalizatione
Phenanthrenoid
puckerf
b
(3a)
(3b)
(3c)
(3d)
(3e)
a
ꢀ1980.574906
ꢀ2165.904533
ꢀ2178.247712
ꢀ4064.980568
ꢀ1807.572845
51.77
52.56
51.97
52.26
51.53
76.04
90.67
73.69
92.49
93.07
3.473
3.431
3.490
3.539
3.720
0.004
0.010
0.005
0.034
0.042
0.289
0.277
0.284
0.294
0.331
Energies in atomic units (1 au = 627.5 kcal/mol).
Dihedral angles in degrees. The magnitude of the average dihedral angles for the two bridgehead phenyls is given. C20 is arbitrarily designated as the carbon
b
proximal to the ketone carbonyl.
c
Dihedral angles in degrees. The magnitude of the dihedral angle is given as measured to the closer imide carbonyl.
Distance in angstroms, measured from the ‘‘inner’’ substituent at C600 of the N-aryl ring to the midpoint of the line joining C8a and C10a, the ‘‘center’’ of the
d
middle ring of the phenanthrenoid moiety.
e
˚
Pyramidalization expressed as distance (A) of nitrogen from the plane defined by the three directly bonded atoms.
˚
Pucker defined as the distance (A) between the midpoints of the lines joining C4a–C4b and H2–H7.
f

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1905
3.12. Ab Initio geometry optimizations
dihedral angles of the N-aryl rings relative to the imide;
closest approach distances (angstroms) of the inner ortho
substituent atom of the N-aryl ring to the center of the middle
ring of the phenanthrenoid moiety; pyramidalization at
nitrogen expressed as the distance (angstroms) of the
nitrogen from the plane defined by the three directly-bonded
atoms; and puckering of the phenanthrenoid system. The
latter parameter is defined as the distance between the
midpoints of: (a) the line joining H2 and H7, and (b)
the C4a–C4b bond. We have defined the center of the
Ab initio geometry optimizations for the adducts have
been performed at the Hartree–Fock level with 6-31G* basis
set (or LACVP* for bromine-containing compounds), using
Spartan ‘04 for Windows (full version) or Spartan ‘04
Essential (Wavefunction Inc., Irvine, CA) [25–28]. To
conserve space, our summary data (shown in Table 1) for
these computational results is limited to: optimized energies;
dihedral angles of bridgehead phenyls relative to the ketone;
Fig. 3. Representative views of calculated structures for phencyclone adducts. In some cases, the hydrogens are ‘‘hidden’’ for clarity. (a) Compound (3a), from
4-bromo-2,6-difluorophenylmaleimide; (b) compound (3b), from 2,3,5,6-tetrafluorophenylmaleimide; (c) compound (3c), from 4-bromo-2,3,5,6-tetrafluor-
ophenylmaleimide; (d) compound (3d), from pentachlorophenylmaleimide; and (e) compound (3e), from 2,4,6-tribromophenylmaleimide.

1906
K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
phenanthrenoid middle ring as the midpoint of the line from
C8a to C10a. All five adducts had similar dihedral angles for
the bridgehead phenyls, ca. 528. For adducts (3a) and (3c),
these dihedrals differed by about 0.58 between the phenyl on
one side of the adduct and the phenyl on the other side,
although for the other adducts, the dihedrals differed by less
than 0.158. It is also striking that for the same adducts (3a)
and (3c), the N-aryl ring was found to form a dihedral angle
of ca. 758 relative to the imide, much less symmetrical than
for the other adducts, for which the N-aryl ring was nearly
perpendicular to the imide. The adducts (3a) and (3c) both
possess a pair of ortho 200,600 fluorines and a para bromine
on the N-aryl ring, but it is not obvious why these
compounds should deviate as they do in their calculated
values. Nitrogen pyramidalization is notably greater for
adducts (3d) and (3e), in which larger, lower row atoms, i.e.,
Cl and Br, occupy the crowded ortho ‘‘inner’’ positions on
C600 of the N-aryl ring, such that the N-aryl tips away from
the phenanthrenoid system. This is also reflected for (3e)
in which there is appreciably more puckering of the
phenanthrenoid moiety. If this system is regarded as
‘‘butterfly-like’’ in shape, the outer rings (i.e., the butterfly
wings) are folded towards the N-aryl, while the central ring
of the phenanthrenoid is relatively more remote, to reduce
hindrance. Fig. 3 shows representative calculated structures
for some of the adducts.
sets, resulted in calculated structures that reflected consider-
able hindrance, and were consistent with observed slow
rotation rates (on the NMR timescales) for the unsubstituted
bridgehead phenyl and N-aryl rings.
Acknowledgments
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 K.M. and K.R.
References
[1] T. Sasaki, K. Kanematsu, K. Iizuka, Molecular design by cycloaddi-
tion reactions. XXV. High peri- and regiospecificity of phencyclone, J.
Org. Chem. 41 (7) (1976) 1105–1112.
[2] R. Callahan, K. Bynum, R. Prip, R. Rothchild, Synthesis of Diels–
Alder adducts of phencyclone and NMR studies of hindered rotations
of unsubstituted bridgehead phenyls: microscale experiments in
organic chemistry, The Chemical Educator 1998 3 (2): S1430-4171
(98) 02192-4, http://journals.springer-ny.com/chedr.
[3] K. Bynum, R. Rothchild, NMR studies of hindered rotation. The
Diels–Alder adduct of phencyclone with N-n-propylmaleimide:
restricted motion of bridgehead phenyls, Spectrosc. Lett. 29 (8)
(1996) 1599–1619.
4. Conclusions
[4] K. Bynum, R. Rothchild, NMR studies of hindered rotation. The
Diels–Alder adduct of N-n-butylmaleimide with phencyclone:
restricted motion of bridgehead phenyls, Spectrosc. Lett. 29 (8)
(1996) 1621–1634.
We have prepared, and presented NMR data for,
phencyclone adducts from five N-arylmaleimides (2),
including: (2a), N-(4-bromo-2,6-difluorophenyl)maleimide;
(2b), N-(2,3,5,6-tetrafluorophenyl)maleimide; (2c), N-(4-
bromo-2,3,5,6-tetrafluorophenyl)maleimide; (2d), N-(2,3,4,
5,6-pentachlorophenyl)maleimide; and (2e), N-(2,4,6-tribro-
mophenyl)maleimide. Maleimides (2a–2c) were prepared in
two steps, with isolation of intermediate N-aryl maleamic
acids (5). The corresponding maleamic acids (5a–5c), each
showed signal broadening in their ambient temperature
fluorine-19 NMR spectra (at 282 MHz) consistent with an
intermediate exchange regime, which we have ascribed to the
process involving N-aryl rotation. In the phencyclone adducts
(3a–3e), proton and carbon-13 NMR (at 300 and 75 MHz,
respectively) showed SEL spectra for rotation of the
bridgehead phenyls about the C(sp2)–C(sp3) bonds. Adducts
(3a–3c) displayed SEL spectra for the N-aryl rotations based
on, e.g., 1H, 13C and fluorine-19 NMR. For adducts (3d) and
[5] L.A. LaPlanche, Y. Xu, R. Benshafrut, R. Rothchild, E.A. Harrison Jr.,
NMR studies of hindered rotation of unsubstituted bridgehead phenyl
rings in the phencyclone-norbornadiene adduct, Spectrosc. Lett. 26 (1)
(1993) 79–101.
[6] Y. Xu, L.A. LaPlanche, R. Rothchild, Assignment of 13C resonances
in the phencyclone-norbornadiene adduct via 2D NMR. 13C evidence
for hindered rotation of unsubstituted bridgehead phenyl rings, Spec-
trosc. Lett. 26 (1) (1993) 179–196.
[7] K. Bynum, R. Rothchild, Analysis of hindered rotation and magnetic
anisotropy by NMR. Models for drugs and agricultural compounds.
The Diels–Alder adduct of phencyclone with N-carbamoylmaleimide,
Spectrosc. Lett. 30 (4) (1997) 727–749.
[8] M. Amin, K. Bynum, R. Callahan, R. Prip, R. Rothchild, Studies of
hindered rotation and magnetic anisotropy by 1H, 13C and 19F NMR.
The Diels–Alder adduct of N-pentafluorophenyl maleimide and phen-
cyclone: a model for drugs, Spectrosc. Lett. 31 (3) (1998) 673–692.
[9] O. Ramirez III, R. Tokhunts, R. Rothchild, NMR studies of crowded
Diels–Alder adducts of phencyclone with N-(2,6-dialkylphenyl)ma-
leimides. Hindered rotations and magnetic anisotropy, Spectrosc. Lett.
37 (4) (2004) 421–436.
1
(3e), slow N-aryl rotation was shown by H or 13C NMR.
Significant magnetic anisotropic effects were seen for protons
on the N-aryl rings of adducts (3a) and (3e), depending on
their orientation‘‘into’’ or‘‘outof’’ the adductcavity;withthe
‘‘inner’’ orientation, the nuclei are directed towards, and
shielded by, the phenanthrenoid moiety. Ab initio geometry
optimizations for the five phencyclone adducts, performed at
the Hartree–Fock level, using the 6-31G* or LACVP* basis
[10] K. Marshall, R. Rothchild, NMR studies of hindered rotation and
magnetic anisotropy: The Diels–Alder adducts of phencyclone with N-
phenylmaleimide and N-(2-trifluoromethylphenyl)maleimide. Ab
initio calculations for optimized structures, Spectrosc. Lett. 37 (5)
(2004).
[11] R. Tokhunts, O. Ramirez III, R. Rothchild, NMR studies of crowded
Diels–Alder adducts of 3,6-dibromophencyclone with N-(2,6-dialkyl-
phenyl)maleimides. Hindered rotations and magnetic anisotropy. Ab

K. Marshall et al. / Journal of Fluorine Chemistry 125 (2004) 1893–1907
1907
initio calculations for optimized structures of the precursor malei-
mides and their adducts with phencyclone and 3,6-dibromophency-
clone, Spectrosc. Lett. 37 (5) (2004).
[19] Y. Yoshitake, J. Misaka, M. Abe, M. Yamasaki, M. Eto, K. Harano,
Carboxylic acid clathrate hosts of Diels–Alder adducts of phencyclone
and 2-alkenoic acids. Role of bidentate C–HꢂꢂꢂꢂO hydrogen bonds
between the phenanthrene and carbonyl groups in host-host networks,
Org. Biomol. Chem. 1 (7) (2003) 1240–1249.
[12] K. Bynum, R. Prip, R. Callahan, R. Rothchild, 19F NMR studies of the
Diels–Alder adduct of N-pentafluorophenylmaleimide with phency-
clone. Hindered rotation about a N–C6F5 bond, J. Fluorine Chem. 90
(1) (1998) 39–46.
[20] T.H. Siddall III, W.E. Stewart, Slow rotations in some substituted
anilides, J. Org. Chem. 34 (10) (1969) 2927–2933.
[13] R. Callahan, K. Marshall, R. Rothchild, K. Rosmarion, Syntheses of
phencyclone analogs. Applications for NMR studies of hindered
rotations and magnetic anisotropy in crowded Diels–Alder adducts.
The Chemical Educator [online] 2001, 6, 227234 DOI 10.1007/
s00897010488a.
[21] W.E. Stewart, T.H. Siddall III, Nuclear magnetic resonance studies of
amides, Chem. Rev. 70 (5) (1970) 517–551.
¨
[22] J. Sandstrom, Dynamic NMR Spectroscopy, Academic Press, New
York, 1982.
[23] R.J. Abraham, J. Fisher, P. Loftus, Introduction to NMR Spectroscopy,
Wiley, New York, 1988.
[14] J.M. Barrales-Rienda, J. Gonzalez Ramos, M. Sanchez Chaves, Synth-
esis of N-(fluorophenyl)maleamic acids and N-(fluorophenyl)malei-
mides, J. Fluorine Chem. 9 (4) (1977) 293–308.
[24] O. Sklyut, R. Prip, N. Azar, R. Callahan, R. Rothchild, Proton and
carbon-13 NMR studies of the phencyclone adducts of medium alkyl
chain-length N-n-alkylmaleimides. Hindered rotations and magnetic
anisotropic effects. Ab initio calculations for optimized structures,
Spectrosc. Lett. 37 (5) (2004).
´
[15] K. Bouchal, M. Ryska, F. Hrabak, Homopolymerization of N-(penta-
chlorophenyl)maleimide and N-(2,4,6-tribromophenyl)maleimide,
Makromol. Chem. 178 (1) (1977) 55–68 (Chem. Abstr. 86: 90323).
[16] Z.Y. Wang, Syntheses of some N-alkylmaleimides, Synth. Commun.
20 (11) (1990) 1607–1610.
[25] F. Jensen, Introduction to Computational Chemistry, Wiley, New
York, 1999.
[17] M. Eto, K. Setoguchi, A. Harada, E. Sugiyama, K. Harano, Phencyclone
Diels–Alder adducts as a new crystalline host. Role of C–Hꢂꢂꢂꢂp and
C–HꢂꢂꢂꢂO interactions, Tetrahedron Lett. 39 (52) (1998) 9751–9754.
[18] Y. Yoshitake, J. Misaka, K. Setoguchi, M. Abe, T. Kawaji, M. Eto, K.
Harano, Non-hydroxylic clathrate hosts of [4 + 2]p cycloadducts of
phencyclone and N-arylmaleimides: recognition of aromatic guests, J.
Chem. Soc., Perkin Trans. 2 (9) (2002) 1611–1619.
[26] TITAN Tutorial and User’s Guide, Wavefunction Inc. (Irvine, CA) and
¨
Schrodinger Inc. (Portland, OR), 1999.
[27] W.J. Hehre, A.J. Shusterman, J.E. Nelson, The Molecular Modeling
Workbook for Organic Chemistry, Wavefunction Inc., Irvine, CA,
1998.
[28] W.J. Hehre, A Guide to Molecular Mechanics and Quantum Chemical
Calculations, Wavefunction Inc., Irvine, CA, 2003.