Quantitative reaction cascades of ninhydrin in the solid state

Article abstract of DOI:10.1002/1521-3765(20020201)8:3<594::AID-CHEM594>3.0.CO;2-5

Crystalline ninhydrin (1) undergoes waste-free solid-state cascade reactions with dimedone, L-proline, three o-phenylenediamines, o-mercaptoaniline, two ureas, three thioureas, and methyl 3-aminocrotonate. The yields are quantitative and give pure crystal

Full text of DOI:10.1002/1521-3765(20020201)8:3<594::AID-CHEM594>3.0.CO;2-5

FULL PAPER
Quantitative Reaction Cascades of Ninhydrin in the Solid State
[a]
Gerd Kaupp,* M. Reza Naimi-Jamal, and Jens Schmeyers
Abstract: Crystalline ninhydrin (1) un-
dergoes waste-free solid-state cascade
reactions with dimedone, l-proline,
three o-phenylenediamines, o-mercap-
toaniline, two ureas, three thioureas, and
methyl 3-aminocrotonate. The yields are
quantitative and give pure crystalline
products without workup just by milling
stoichiometric mixtures of the crystal-
line reagents. The structures of the new
and the previously obtained products
with lower yields from solutions are
established or confirmed by spectro-
scopic data and density functional cal-
culations at the B3LYP/6-31G* level.
The success of 3- and 4-cascade reac-
tions in the crystal without melting is
unusual and of unmatched atom econo-
my. They are mechanistically investigat-
ed with atomic force microscopy techni-
ques (AFM) on six different faces of 1
when o-phenylenediamine was the re-
agent (substitution, elimination, cycliza-
tion, elimination) and interpreted on the
basis of known crystal structure data.
Strict correlations to the crystal packings
are observed. The characteristic surface
features grow to mm heights in some
cases at distances of 0.5 mm from the
contact edge of the reacting crystals. The
waste-free and easy syntheses of highly
ˆ
ˆ
functionalized (C O; O-H; C N) het-
erocycles or of a tetraketone are also of
interest for synthetic use.
Keywords: environmentally benign
¥ heterocycles ¥ ninhydrin ¥ scan-
ning probe microscopy ¥ solid-state
reactions
Introduction
Gas-solid and solid-solid reactions profit from the crystal
packing and have been mechanistically investigated.[1] They
require long-range molecular movements (phase rebuilding),
phase transformation into the product lattice and creation of
fresh surface by crystal disintegration. Frequently they run to
completion with highest selectivity and give 100% yield under
moderate reaction conditions without any necessity for work-
up. Even multistage cascade reactions were quantitative in the
solid state, a fact that provides unsurpassed efficiency in terms
of environmental protection and savings of both labor and
resources. We now extended those principles to versatile
reactions of ninhydrin 1 in 2-, 3- and 4-cascades of various
reaction types.
Scheme 1. Quantitative condensation of ninhydrin with dimedone.
solution with 82% yield.[2] No dehydratisation was observed
in the solid state to yield 4 even with gaseous trimethylamine
or hydrogen chloride or p-toluenesulfonic acid. Anyhow, the
liquid-state synthesis of 4 in anhydrous dioxane with trifluoro-
acetic anhydride and pyridine[3] gains importance, as 3 is now
more easily accessible without producing any waste. Such
reluctance for solid-state elimination of water was not typical
for all condensation reactions of ninhydrin in the solid state.
Results
Quantitative one-step condensation: The solid state reaction
of ninhydrin 1 with dimedone 2 in a ball-mill proceeds only as
a
one-step condensation to give 3 with 100% yield
(Scheme 1). Compound 3 has previously been synthesized in
Quantitative 3-cascade with proline: The presumed wealth of
amino acids in solid-state reactions has not yet been exploited
so far. It is thus of high interest that l-proline 5 reacts
quantitatively in a 3-cascade (substitution, elimination, de-
carboxylation) with 1 to give the versatile azomethine ylide 6
(Scheme 2) waste-free and in quantitative yield. Again, 6 can
[a] Prof. Dr. G. Kaupp, Dr. M. R. Naimi-Jamal, Dr. J. Schmeyers
University of Oldenburg, FB Chemie, Organische Chemie I
Postfach 2503, 26111 Oldenburg (Germany)
Fax : (‡49) 441-7983-409
E-mail: kaupp@kaupp.chemie.uni-oldenburg.de
594
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0594 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

594 600
be synthesized in solution but
only with 82% yield.[4] It has
been used for 1,3-dipolar cyclo-
additions.[5] The zwitterionic
structure of 6 is secured by the
symmetry of the aromatic pro-
ton resonances, the presence of
an iminium H (d ˆ 9.24) and of
an iminium C (d ˆ 213.78) in
the NMR spectra. In support of
this structure, density function-
al theory (DFT) calculations at
the B3LYP/6-31G* level find
the cyclized valence tautomers
by 15.25 (spiroaziridine) or
Scheme 2. Quantitative 3-cascade in the solid-state reaction of ninhydrin with l-proline.
17.49 kcalmolÀ1
(oxazoline)
higher in energy content.
Quantitative 4-cascades with
o-phenylenediamines: Various
solid o-phenylenediamines
7
react quantitatively with solid
ninhydrin 1 in 4-cascades (sub-
stitution, elimination, cycliza-
tion, elimination) to give the
indenoquinoxaline ketones 8
(Scheme 3). Similar reactions
proceed in solution with yields
of 61% to almost quantita-
tive.[6] The presumed sequence
Scheme 3. Quantitative solid-state 4-cascades of ninhydrin with o-phenylenediamines.
of events (substitution, elimination, addition, and elimina-
tion) allows for a rationalization of the orientation selectivity
found. While 7a and 7c have no orientational choice (see
products 8a and 8c), 7b yields the 6-Me isomer 8b exclusively
(75% in solution[6a]) obviously for steric reasons in the
primary substitution step. The structures of the compounds 8
have been derived from the NMR data and by comparison
with the known isomers from the solution reactions. The
possibility of a symmetric diketo spiro-aminal structure is
excluded by the asymmetry in the NMR spectra and the lack
Abstract in German: Kristallines Ninhydrin (1) geht mit
Dimedon, l-Prolin, drei o-Phenylendiaminen, o-Mercapto-
anilin, zwei Harnstoffen, drei Thioharnstoffen und 3-Amino-
crotons‰uremethylester abfallfreie Kaskadenreaktionen im
festen Zustand ein. Die Ausbeuten sind quantitativ und geben
reine kristalline Produkte beim Vermahlen stˆchiometrischer
Mischungen der kristallinen Reaktanden, ohne dass reinigende
Aufarbeitung erforderlich w‰re. Die Konstitution der neuen
und fr¸her aus Lˆsungen mit schlechterer Ausbeute erhaltenen
Produkte werden mittels spektroskopischer Daten und Dichte-
funktional Theorie Rechnungen auf dem B3LYP/6-31G*
Niveau begr¸ndet und best‰tigt. Der gute Erfolg der 3- und
4-Kaskaden-Reaktionen im Kristall ohne fl¸ssige Phase ist
ungewˆhnlich und von ¸berragender Atom-÷konomie. Sie
werden daher mittels Kraftmikroskopie (AFM) auch mecha-
nistisch untersucht. Dies gelingt mit o-Phenylendiamin auf
sechs verschiedenen kristallographischen Fl‰chen von 1 (Sub-
stitution, Eliminierung, Cyclisierung, Eliminierung). Die In-
terpretation gelingt auf Grundlage der bekannten Rˆntgen-
strukturanalyse: Es wird die strikte Korrelation zur Kristall-
packung best‰tigt. Die charakteristischen Oberfl‰chen-
strukturen wachsen auf 1 in einigen F‰llen bis zur Mikro-
meter-Hˆhe obwohl der Reagenzkristall 0.5 mm entfernt auf-
liegt. Die abfallfreien und besonders einfachen Synthesen hoch
À
of N H vibrations in the IR spectrum.
Quantitative 3-cascade with o-mercaptoaniline hydrochlor-
ide: Stoichiometric reactions of solid 1 with solid o-mercap-
toaniline hydrochloride (9) gives the salt 10 ¥ HCl (Scheme 4)
in quantitative yield by substitution, cyclization, and elimi-
nation. The free base 10 has a very low solubility in water and
can thus be obtained in quantitative yield by a NaHCO3
treatment. Similar reactions in solution were reported to
provide only 60% yield.[7] Compound 10 is remarkably stable:
It does not form a spiro-N/S-acetal upon heating up to its
melting point at 2278C.
Quantitative 2-cascades with ureas: Ureas (11) and more
easily thioureas (11) are well suited for substitution and
addition to 1 (Scheme 5). The interesting heterocyclic bis-
ˆ
ˆ
funktionalisierter (C O; O-H; C N) Heterocyclen oder eines
Tetraketons sind auch von synthetischem Interesse.
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0595 $ 17.50+.50/0
595

G. Kaupp et al.
FULL PAPER
Table 1. Syntheses of 12 in mortar, ball-mill (bm), and solution.
12
Reaction conditions
Yield
[%]
Yield in
solution [%]
Ref.
[8]
a
mortar, 3 d at 808C
bm, 1 h at 808C
bm, 45 min, RT
mortar, 3 d
100
100
100
100
100
100
100
100
78
[9]
b
c
91
92
[10, 11]
bm, 2 h, RT
d
e
bm, 2 h,(water bath, 208C)
mortar, 3 d, RT
bm, 1 h, 608C
N/O-semiaminal 15 could not be obtained in a 4-cascade
sequence in the solid state. Compound 15 had been synthesized
in solution, but only in 82% yield.[12] The structure of 15 is
Scheme 4. Quantitative solid-state 3-cascade of ninhydrin and o-mercap-
toaniline.
1
derived from the H NMR data which exclude a symmetric
diketo spiro[5,4]azaoctadien structure. DFT calculations at the
B3LYP/6-31G* level give the imine tautomer (in brackets) 5.86
kcalmolÀ1 higher in energy than the enamine tautomer 15.
AFM results: It appears interesting to study the influence of
the molecular packing on the molecular movements in
multistage cascade reactions in the solid state and to exclude
(nano)liquid intermediate phases. Thus, six different crystal
faces of 1 were studied with the AFM next to a reactant crystal
of o-phenylenediamine (7a) on it. Ninhydrin (1) crystallizes in
monoclinic prisms in the chiral space group P21.[13] Hydrogen
bonded double layers appear as principle motif. The
O-H ¥¥¥ O-H hydrogen bonds have O ¥¥¥ O distance of
Scheme 5. Quantitative 2-cascades of ninhydrin and (thio)ureas in the
solid state.
[13]
ˆ
2.7981 ä, O-H ¥¥¥ O C of 2.805 ä.
One of the carbonyl
groups of 1 remains unbridged. There are no O-H hydrogen
bonds between the double layers. The orientation of double
layers determines the solid-state reactivity at the various
faces. The molecules of 7a migrate into the crystal of 1 as can
be judged from a light yellow coloration. The (100)-face is not
strongly present on the prismatic crystals of 1, however, these
cleave easily at that face and thus freshly cleaved (100)-
surfaces were used in the measurements of Figure 1. A flat
crystal of 7a contacted well and was placed in [001]-direction
from the AFM tip. The features of Figure 1 indicate only
minor reactions on that face. The depth of the craters after
210 min does not exceed 20 nm and even the deepest crater
after 380 min is only 65 nm deep. Evidently the reaction does
not extend easily over the cleavage plane.
N/O-semiacetals 12 are stable compounds that form with
100% yield in all cases studied (a e). Some of these have
been previously prepared in solution with inferior yields
(Table 1). It is remarkable that neither 12a nor 12b eliminate
water during heating in the DSC/TGA experiments up to the
melting points. Even HCl (gas) or p-toluenesulfonic acid
(solid) do not produce the anticipated elimination products 13
in the solid state. However, the now waste-free access to the
compounds 12 opens certainly numerous versatile possibilities
for further syntheses using their a-hydroxyketone, a-amino-
ketone and N/O-semiacetal functionalities. The structures 12
are secured by the NMR spectra which exclude symmetrical
diketo spiro[5,4]aminal structures.
Quantitative 3-cascade with
methyl 3-aminocrotonate: The
solid enaminoester 14 reacts
quantitatively with 1 in stoi-
chiometric ball-mill runs with-
out intermediate liquid phase
to give the heterocyclic building
block 15 with a-hydroxyketone
and N/O-semiaminal function-
ality (Scheme 6). The 3-cascade
consists of vinylogous substitu-
tion, cyclization and 1,3-hydro-
gen shift. Again, the presumed
Scheme 6. Quantitative solid-state 3-cascade of ninhydrin and methyl 3-aminocrotonate.
elimination of water from the
596
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0596 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

Solid-State Reactions
594 600
densely
Figure 3d.
These
packed hills grow slowly further
to 1.35 mm after 3 h. Unlike the
behavior in Figure 1 we have a
strong transport of material
above the (10-1)-surface. This
behavior is derived from the
crystal packing in Figure 4. The
functional groups are available
to 7a for entering deeply into
the crystal while reacting and
spreading to the inclined cleav-
age planes from where appa-
rently 8a exits above the hori-
zontal surface.
The major (1-10)- and (-1-10)-
faces of 1 give the same type
of features upon reaction
with 7a. Figure 5 shows the
results on (1-10). Irrespective
Figure 1. 9.5 mm AFM images of 1 on its (100)-cleavage plane at 0.1 mm distance from a crystal of 7a on it after
the times given, showing the formation of small craters that grow gradually. The z scale is 50 nm, the direction of
preference runs along the [001]-direction (the scan direction was changed after 210 min).
of the roughness in Figure 5a, isolated island hills formed at
the remarkably large distance of 0.5 mm by long-range
molecular movements. These increased to edged mountains
of considerable height without direction of preference (Fig-
ure 5c, d) and grew to twice that height after 20 h.
The packing under the (100) double layers is shown in
Figure 2. The molecules 7a have to penetrate through the
The crystal packing in Figure 6 clearly indicates that the
functional groups are available for reaction on the (1-10)
surface. The product molecule 8a can exit along the steep
cleavage plane (100) and thus build the island mountains
around nucleation sites while reacting molecules of 7a enter
the lattice. The molecular packing under the (-1-10) surface is
the same as in Figure 6. It is thus not surprising that the same
type of features ensue, as expected (this was shown at 0.3 mm
distance). Furthermore, Figure 6 indicates that the packing
under (-110) is different from the one on (1-10). It is the same
as the packing under (110) and has the functional groups
somewhat shielded by the skew benzo groups. Thus, at least a
Figure 2. Stereoscopic view of the crystal packing of 1 on (010), but turned
around y b y 58 for a better view, showing the structure of the double layers
(height ca. 11 nm) and the horizontal (100) cleavage plane; all hydrogen
bonds are given.
double layers from the top,
react and spread along the
cleavage plane where the func-
tional groups are shielded. It
can thus be assumed that no
rapid reaction occurs. The flat
molecules 8a appear to com-
plex the water of the reaction
while forming small craters
around nucleation sites.
The behavior on the (10-1)-
face of 1 that could be mounted
after cutting the crystal across
the major prism faces, is com-
pletely
different.
Figure 3
shows how the reaction with
7a produces volcano type fea-
tures all over the surface which
grow together rapidly and as-
sume heights of up to 1 mm in
Figure 3. 9.5 mm AFM topographies on the (10-1)-face of 1 at 0.1 mm distance from a flat crystal of 7a; a) fresh;
b) after 2 min; c) after 15 min; d) after 60 min.
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0597 $ 17.50+.50/0
597

G. Kaupp et al.
FULL PAPER
Again the behavior on (110) is closely comparable, due to
the identical packing. After difficulties with its start, the rate
approaches the one of the more favorable (-1-10)-face.
It appears, that the molecules 7a have initially difficulties to
enter the crystal on (-110) or (110) due to the impeding benzo
groups except at very few nucleation centers. However, after
considerable molecular movements the reaction gains in rate
and the transport along the (100) cleavage plane is the same as
on the opposite side of the crystal.
Figure 4. Stereoscopic view of the crystal packing of 1 with the (10-1)-face
horizontal on top showing steep double layers; all hydrogen bonds are
given.
Discussion
Ninhydrin (1) is a rich source
for waste-less quantitative sol-
id-state cascades. Its melting
point is quite high. Various
reaction types were already
shown with solution reactions
which produce waste due to
lower yield and workup re-
quirements.
The unsurpassed perform-
ance of the presented solid-
state cascades in stoichiometric
runs is unusual and highly prof-
itable in terms of sustainable
synthesis with utmost efficiency
in all respects. Their mechanis-
tic basis was developed in nu-
merous investigations of vari-
ous types of solid state reac-
tions using supermicroscopy
(AFM and SNOM),[1] and graz-
Figure 5. 9.5 mm AFM topographies on the (1-10)-face of 1 at 0.5 mm distance from a flat crystal of 7a; a) fresh;
b) after 15 min; c) after 2 h; d) after 4 h.
ing
incidence
diffraction
(GID),[14] and correlations to
crystal packings.[1] Reaction
cascades in the solid state are
very complicated, however,
they also profit from the gen-
eral three-step solid state mech-
anism of 1) phase rebuilding,
2) phase transformation, and
3) crystal disintegration at any
local sites and development.
Unlike the gas/solid reactions
step (2) does not occur sudden-
ly in the solid/solid reactions
Figure 6. Stereoscopic packing diagram (crystal surfaces model) of 1 with (1-10) horizontal on top and (-110)
horizontal at the bottom, showing the steep double layers. All hydrogen bonds are given.
and it is hard to obtain the crystal disintegration with the
studies at single crystals for technical reasons due to the very
long distances. Nevertheless, the heights of the features at
0.5 mm distance show the high efficiency of the molecular
migrations. Generally, step 3) and the generation of new
contacts between the reacting crystals is attained by mechan-
ical grinding, milling and other techniques.
The products of the condensation reactions of ninhydrin are
preparatively useful building blocks which are now available
without producing waste. The reactions are too slow to be
completed in a mortar at moderate times, but proceed very
different reactivity is expected and indeed, we assigned the
more efficient feature formations to (1-10) and (-1-10) and the
less efficient ones to the opposing (-110) and (110)-faces of the
same crystal, respectively.
Figure 7 depicts the AFM results on the (-110) surface of 1.
The reaction of 7a is indeed impeded, even though the initial
natural surface is rather rough. Only few island features
develop in the first 20 min (Figure 7b), but the reaction
becomes fast at 90 min and similar features develop of
comparable height than those in Figure 5, but their width
remains smaller.
598
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0598 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3

Solid-State Reactions
594 600
100%) was obtained. M.p. 193 1958C (de-
comp) (ref. [7]: 196 1988C, decomp);
1H NMR (CDCl3/[D6]DMSO): d ˆ 1.03 (s,
6H), 2.23 (s, 4H), 2.5 3.2 (brs, 2H), 7.60
7.80 (AA'BB', 2H), 7.80 8.00 (BB'AA', 2H).
1-Oxo-2-(1-pyrrolin-1-ylium)-1H-indene-3-
olat (6): Ninhydrin (1; 356 mg, 2.00 mmol)
and l-proline (5) (230 mg, 2.00 mmol) were
ball-milled for 1 h. The teflon seal was
tightened at 15 Nm to avoid dangerously
high pressure. The water of the reaction was
removed in a vacuum at 808C for 1 h to yield
the title compound (421 mg, 99%). M.p.
2398C (ref. [15]: 2408C); IR (KBr): nÄ ˆ
1663, 1625, 1577, 1408 cmÀ1
;
1H NMR
(CDCl3/[D6]DMSO): d ˆ 2.20 2.34 (m,
2H), 3.00 3.12 (m, 2H), 4.83 4.95 (m,
2H), 7.46 (s, 4H), 9.24 (s, 1H); 13C NMR
(CDCl3/[D6]DMSO): d ˆ 17.61, 31.72, 57.31,
118.21 (2C), 130.19 (2C), 135.92, 150.82 (2C),
181.46 (2C), 213.78.
11H-Indeno[1,2-b]quinoxaline-11-one(8a):
Ninhydrin (1) (356 mg, 2.00 mmol) and o-
phenylenediamine (7a) (216 mg, 2.00 mmol)
were ball-milled for 15 min at À58C. Com-
Figure 7. 9 mm AFM topographies on the (-110)-face of 1 at a 0.5 mm distance from a flat crystal of 7a that
was placed on it; a) fresh; b) after 20 min; c) after 2 h; d) after 4 h.
pound 8a (461 mg, 99%) was obtained after drying in a vacuum at 808C.
M.p. 217 2188C (ref. [6a]: 217 2198C); IR (KBr): nÄ ˆ 1732, 1605,
well in a ball-mill at milling times of less than one hour. We
reported here on 12 quantitative and waste-free reaction
cascades. Many more reaction cascades with ninhydrin 1 and
diamines or diamides can be envisaged, as the crystal lattice of
1 appears very permissive to solid reaction partners.
1571 cmÀ1
;
1H NMR (CDCl3): d ˆ 7.56 (yt, 1H), 7.77 (m, 3H), 7.90 (yd,
1H), 8.09 (m, 2H), 8.20 (yd, 1H); 13C NMR (CDCl3/[D6]DMSO): d ˆ
122.2, 124.3, 129.3, 129.9, 131.1, 132.1, 132.2, 136.2, 136.5, 141.1, 142.2, 142.7,
148.9, 156.2, 189.5.
11H-6-Methylindeno[1,2-b]quinoxaline-11-one (8b): Ninhydrin (1;
178 mg, 1.00 mmol) and 7b (122 mg, 1.00 mmol) were ball-milled for 1 h.
After drying in a vacuum at 808C pure 8b (264 mg, 100%) was obtained.
M.p. 2278C (ref. [6a]: 222 2248C); IR (KBr): nÄ ˆ 1726, 1568, 1193,
Experimental Section
770 cmÀ1
;
1H NMR (CDCl3): d ˆ 2.83 (s, 3H), 7.52 7.68 (m, 3H), 7.69
7.78 (yt, 1H), 7.82 7.94 (yd, 1H), 7.95 8.12 (yt, 2H); 13C NMR (CDCl3):
d ˆ 17.2, 122.3, 124.5, 129.4, 129.8, 132.1, 132.6, 136.5, 138.2 (2C), 141.8,
142.0, 142.7, 148.7, 155.4, 190.2.
Melting points were determined with
a Gallenkamp melting point
apparatus and are uncorrected. Infrared (IR) spectra were recorded with
a Perkin Elmer 1720-X FT-IR spectrometer using potassium bromide
pellets. All NMR spectra were taken at a Bruker WP 300 at 300 MHz (1H)
or 75 MHz (13C). CDCl3/[D6]DMSO mixtures contained up to 25%
[D6]DMSO. Mass spectra were obtained on a Finnigan MAT 212 System.
DSC/TGA measurements were performed with a Perkin Elmer DSC
model 7 and TGA 7. The ball-mill was a Retsch MM 2000 swing mill with a
10 mL stainless steal double-walled beaker with fittings for circulating
coolants. Two stainless steal balls with 12 mm diameter were used. Ball-
milling was performed at 20 25 Hz frequency usually at room temperature
(without circulating liquid the temperature did not rise above 308C). Water
or methanol of the appropriate temperature was circulated for heating or
cooling. Completeness of the solid-state reactions was checked by IR
spectroscopy in KBr, product purity by m.p. and 1H NMR spectroscopy.
11H-7,8-Dimethylindeno[1,2-b]quinoxaline-11-one (8c): Ninhydrin (1;
356 mg, 2.00 mmol) and 7c (272 mg, 2.00 mmol) were ball-milled for 1 h.
After drying in a vacuum at 808C pure 8c (517 mg, 100%) was obtained.
1
M.p. 2488C; IR (KBr): nÄ ˆ 1720, 1606, 1566 cmÀ1; H NMR ([D6]DMSO):
d ˆ 2.49 (2s, 6H), 7.65 (yt, 1H), 7.82 (m, 2H), 7.91 (s, 2H), 8.01 (yd, 1H);
13C NMR (CDCl3/[D6]DMSO): d ˆ 19.6, 19.9, 121.5, 123.7, 128.4, 129.9,
131.5, 135.7, 136.0, 140.2, 140.7, 141.0, 141.2, 142.9, 147.6, 155.4, 189.5; HR-
MS (CI, isobutane): calcd for C17H12N2O‡H: 261.1040; found: 261.1041.
10a-Hydroxyindeno[2,1-b]benz[1,4]thiazin-11(10aH)-one (10): Ninhydrin
(1; 356 mg, 2.00 mmol) and o-aminothiophenol hydrochloride (9) (323 mg,
2.00 mmol) were ball-milled for 1 h. The solid product was washed with
saturated NaHCO3 solution and dried to obtain pure 10 (529 mg, 99%).
M.p. 2278C (ref. [7]: 2288C); IR (KBr): nÄ ˆ 1733, 1639, 1594 cmÀ1; 1H NMR
(CDCl3/[D6]DMSO): d ˆ 7.24 (yt, 1H), 7.33 (yt, 1H), 7.46 (yd, 1H), 7.58 (s,
OH), 7.65 (yd, 1H), 7.79 (yt, 1H), 7.93 (m, 2H), 8.20 (yd, 2H); 13C NMR
(CDCl3/[D6]DMSO): d ˆ 119.5, 122.2, 122.9, 125.2 (2C), 126.3, 127.5, 131.9,
133.6, 135.6, 140.8 (2C), 141.6, 154.0, 193.6.
B3LYP (basis set 6-31G*) calculations with full geometry optimization
were performed with the program TITAN, version 1.01 of Wavefunction,
Inc., Irvine, USA. Single crystals of 1 were obtained by slow evaporation of
solutions in water. The prisms selected had lengths of 3 4 mm and
maximal widths of 2 mm. The AFM technique and the imaging have been
described elsewhere.[1]
3a,8a-Dihydroxy-1,3,3a,8a-tetrahydroindeno[1,2-d]imidazole-2,8-dione
(12a): Ninhydrin (1; 534 mg, 3.00 mmol) and urea (11a) (180 mg,
3.00 mmol) were ball-milled at 808C for 1 h. After drying at 1408C for
5 min in a vacuum pure 12a (696 mg, 100%) was obtained. M.p. 2188C
(decomp) (ref. [8]: 218 2208C); IR (KBr): nÄ ˆ 3347, 1728, 1692, 1604, 1413,
Cleavage along the (100)-face of 1 is achieved by touching the (110)-face
with a needle. Cutting across on the (110)-face required considerable
pressure with a razor blade applied in the suitable direction. These cuts for
(10-1)-faces were not always straight, but flat regions were chosen for the
AFM measurements. Distances were judged by comparison with the crystal
width that was measured with a slide gauge. Opposing (-1-10)/(110) and
(1-10)/(-110)-faces were mounted from the same crystal species that was
cut across with a razor blade so that the opposing surfaces could be
mounted separately to the AFM. The faster reacting ones were assigned to
(-1-10) and (1-10) for crystallographic reasons.
1
1116 cmÀ1; H NMR ([D6]DMSO): d ˆ 6.39 (s, 1H), 6.50 (s, 1H), 7.57 (yt,
1H), 7.78 (yt, 2H), 7.85 (s, 2H), 8.00 (s, 1H); 13C NMR ([D6]DMSO): d ˆ
86.6 (2C), 123.4, 125.1, 130.0, 132.1, 136.6, 152.0, 156.7, 197.9.
3a,8a-Dihydroxy-2-thioxo-2,3,3a,8a-tetrahydro-1H-indeno[1,2-d]imida-
zole-8-one (12b): Ninhydrin (1; 178 mg, 1.00 mmol) and thiourea (11b)
(76 mg, 1.00 mmol) were ball-milled for 45 min. Pure 12b (236 mg, 100%)
was obtained. M.p. 2208C (decomp) (ref. [9]: 240 2428C); IR (KBr): nÄ ˆ
2-Hydroxy-2-(4,4-dimethyl-2,6-dioxocyclohexane-1-yl)-1H-indene-1,3(2H)-
dione (3): Ninhydrin (1; 178 mg, 1.00 mmol) and dimedone (2; 140 mg,
1.00 mmol) were ball-milled for 1 h at room temperature. Pure 3 (300 mg,
3310, 1728, 1510, 1107, 615 cmÀ1
(brs, 1H), 6.60 (brs, 1H), 7.50 7.60 (yt, 1H), 7.70 7.95 (m, 3H), 9.13 (brs,
;
1H NMR (CDCl3/[D6]DMSO): d ˆ 6.40
Chem. Eur. J. 2002, 8, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0599 $ 17.50+.50/0
599

G. Kaupp et al.
FULL PAPER
1H), 9.62 (brs, 1H); 13C NMR (CDCl3/[D6]DMSO): d ˆ 88.6, 89.5, 122.7,
Acknowledgement
124.2, 129.1, 131.8, 135.4, 149.5, 177.7, 194.8.
3a,8a-Dihydroxy-1,3-dimethyl-1,3,3a,8a-tetrahydroindeno[1,2-d]imida-
zole-2,8-dione (12c): Ninhydrin (1; 178 mg, 1.00 mmol) and N,N'-dimethyl-
urea (11c) (88 mg, 1.00 mmol) were ball-milled for 2 h. Pure 12c (248 mg,
100%) was obtained. M.p. 2488C (ref. [11]: 247 2518C); IR (KBr): nÄ ˆ
This work was supported by the Deutsche Forschungsgemeinschaft and the
Fonds der Chemischen Industrie. We thank Dr. W. Saak for the Miller
indices of ninhydrin and Mrs. Ludmila Hermann for her skillful help in the
experimental work.
3308, 3153, 1732, 1686, 1604, 1490, 1418, 1400 cmÀ1 1H NMR (CDCl3/
;
[D6]DMSO): d ˆ 2.82 (s, 3H), 2.88 (s, 3H), 6.00 6.30 (brs, 1H), 6.35 6.60
(brs, 1H), 7.40 7.55 (m, 1H), 7.65 7.78 (m, 3H); 13C NMR (CDCl3/
[D6]DMSO), d ˆ 23.5, 23.8, 85.7, 86.7, 123.2, 123.7, 129.1, 131.8, 135.1, 148.0,
154.2, 194.6.
[1] G. Kaupp, in Comprehensive Supramolecular Chemistry, Vol. 8 (Ed.:
J. E. D. Davies), Elsevier, Oxford, 1996, pp. 381 423.
[2] N. P. Peet, E. W. Huber, J. C. Huffman, J. Hetercycl. Chem. 1995, 32,
33.
[3] F. Campagna, A. Carotti, G. Casini, M. Ferappi, Gazz. Chim. Ital.
1983, 113, 507.
[4] A. W. Johnson, D. J. McCaldin, J. Chem. Soc. 1958, 817.
[5] M. F. Aly, H. Ardill, R. Grigg, S. Leong-Ling, S. Rajviroongit, S.
Surendrakumar, Tetrahedron Lett. 1987, 28, 6077.
[6] a) L. W. Deady, J. Desneves, A. C. Ross, Tetrahedron 1993, 49, 9823;
b) S. Ruhemann, J. Chem. Soc. 1910, 97, 1438.
[7] A. Schˆnberg, E. Singer, G.-A. Hoyer, D. Rosenberg, Chem. Ber. 1977,
110, 3954.
[8] R. Caputo, C. Ferreri, G. Palumbo, V. Adovasio, M. Nardelli, Gazz.
Chim. Ital. 1987, 117, 731.
[9] N. Chatterjie, R. A. Stephani, C. H. Strom, J. Pharm. Sci. 1980, 69,
1431.
[10] R. Shapiro, N. Chatterjie, J. Org. Chem. 1970, 35, 447.
[11] P. A. Crooks, T. Deeks, Chem. Ind. 1975, 793.
[12] N. Chatterjie, R. Shapiro, S.-G. Quo, R. A. Stephani, Tetrahedron Lett.
1975, 30, 2535.
3a,8a-Dihydroxy-1,3-dimethyl-2-thioxo-2,3,3a,8a-tetrahydro-1H-indeno-
[1,2-d]imidazole-8-one (12d): Ninhydrin (1; 178 mg, 1.00 mmol) and N,N'-
dimethylthiourea (11d) (104 mg, 1.00 mmol) were ball-milled at 208C for
2 h. Pure 12d (264 mg, 100%) was obtained. M.p. 2198C (decomp); IR
(KBr): nÄ ˆ 3327, 3215, 1700, 1600, 1467, 1381, 1303, 1224 cmÀ1
;
1H NMR
(CDCl3/[D6]DMSO): d ˆ 3.25 (s, 3H), 3.32 (s, 3H), 6.40 6.70 (brs, 1H),
6.70 6.90 (brs, 1H), 7.50 7.62 (m, 1H), 7.72 7.88 (m, 3H); 13C NMR
(CDCl3/[D6]DMSO): d ˆ 28.0, 28.4, 88.5, 90.1, 123.8, 124.1, 129.8, 132.4,
135.8, 147.5, 177.9, 193.6; HR-MS (CI, isobutane): calcd for
C12H12N2O3S‡H: 265.0674; found: 265.0681.
3a,8a-Dihydroxy-1,3-diphenyl-2-thioxo-2,3,3a,8a-tetrahydro-1H-indeno-
[1,2-d]imidazole-8-one (12e): Ninhydrin (1; 178 mg, 1.00 mmol) and N,N'-
diphenylthiourea (11e) (228 mg, 1.00 mmol) were ball-milled at 608C for
1 h. Pure 12e (388 mg, 100%) was obtained. M.p. 2128C (decomp);
1H NMR (CDCl3/[D6]DMSO): d ˆ 6.85 6.95 (m, 1H), 7.10 7.20 (brs, 1H),
7.30 7.53 (m, 11H), 7.55 7.65 (m, 2H), 7.80 7.90 (m, 1H); 13C NMR
(CDCl3/[D6]DMSO): d ˆ 90.4, 92.0, 123.6, 125.0, 127.6, 127.7, 127.9, 130.1,
132.4, 135.4, 135.8, 136.0, 147.6, 179.8, 193.4; HR-MS (EI): calcd for
C22H16N2O3S: 388.0882; found: 388.0885.
[13] R. C. Medrud, Acta Crystallogr. Sect. B 1969, 25, 213.
[14] A. Herrmann, G. Kaupp, T. Geue, U. Pietsch, Mol. Cryst. Liq. Cryst.
1997, 293, 261.
3a,8a-Dihydroxy-2-methyl-8-oxo-3,3a,8,8a-tetrahydro-3-aza-cyclopenta-
[a]indene-1-carboxylic methyl ester (15): Ninhydrin (1; 356 mg, 2.00 mmol)
and methyl 3-aminocrotonate (14) (230 mg, 2.00 mmol) were ball-milled
for 1 h. The solid was dried in a vacuum at 808C to yield 15 (545 mg, 99%).
M.p. 2018C (ref. [12]: 201 2028C); IR (KBr): nÄ ˆ 3518, 3437, 3313, 1706,
[15] A. Schwarz, G. Uray, H. Junek, Liebigs Ann. Chem. 1980, 1919.
1644 cmÀ1
;
1H NMR (CDCl3/[D6]DMSO): d ˆ 2.05 (s, 3H), 3.50 (s, 3H),
5.38 (s, OH), 6.35 (s, OH), 7.48 7.60 (yt, 1H), 7.65 (yd, 2H), 7.78 (yd, 1H),
8.80 (s, 1H); 13C NMR (CDCl3/[D6]DMSO): d ˆ 14.4, 49.5, 85.2, 91.5, 94.7,
122.5, 124.7, 129.8, 134.2, 134.7, 149.7, 160.0, 165.5, 198.9.
Received: July 16, 2001 [F3417]
600
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
0947-6539/02/0803-0600 $ 17.50+.50/0
Chem. Eur. J. 2002, 8, No. 3