Tri-s-triazine derivatives. Part I. From trichloro-tri-s-triazine to graphitic C3N4 structures

Article abstract of DOI:10.1039/b111062b

The first detailed structural characterisation of a functionalised tri-s-triazine derivative, trichloro-tri-s-triazine, is reported, which is a promising starting material for numerous compounds including graphitic C₃N₄ phases. DFT calculations show that a C₃N₄ structure based on tri-s-triazine should exist and that it is ～30 kJ mol^-1 more stable than the previously reported C₃N₄ phase of lowest energy.

Full text of DOI:10.1039/b111062b

View Online / Journal Homepage / Table of Contents for this issue
Tri-s-triazine derivatives. Part I. From trichloro-tri-s-triazine
to graphitic C N structuresy
3
4
a
a
a
b
c
Edwin Kroke,* z Marcus Schwarz, Elisabeth Horath-Bordon, Peter Kroll, Bruce Noll
c
and Arlan D. Norman
a
Darmstadt University of Technology, FB 11, FG Disperse Feststoffe, Petersenstrasse 23, 64287
Darmstadt, Germany. E-mail: kroke@hrzpub.tu-darmstadt.de; Fax: þ49 6151 166346;
Tel: þ49 6151 166347
b
c
Institut fu¨r Anorganische Chemie, Rheinisch-Westfa¨lische Technische Hochschule (RWTH)
Aachen, Professor-Pirlet-Strasse 1, 52056, Aachen, Germany
Chemistry & Biochemistry Department, University of Colorado, Boulder, CO 80309, USA
Received (in Montpellier, France) 11th December 2001, Accepted 19th February 2002
First published as an Advance Article on the web 15th April 2002
The first detailed structural characterisation of a functionalised
tri-s-triazine derivative, trichloro-tri-s-triazine, is reported,
which is a promising starting material for numerous compounds
including graphitic C N phases. DFT calculations show that
3
4
a C
3
N
4
structure based on tri-s-triazine should exist and that
À1
it is ꢀ30 kJ mol more stable than the previously reported
C N phase of lowest energy.
3
4
1
Acid chlorides are important starting materials. For example,
cyanuric chloride (trichloro-s-triazine) 1 in Scheme 1 the
trichloride of cyanuric acid 5ab, is used to produce many
2
triazine containing compounds.
3
The sym-heptazine nucleus C N consists of three fused s-
6
7
triazine rings. A compound structurally analogous to cyanuric
4
acid, cyameluric acid 6ab, has been reported in the literature.
Its chemical properties have been examined several decades
ago, but no detailed spectroscopic and physical data except IR
5
spectra are known. The acid chloride of cyameluric acid 6ab,
trichloro-tri-s-triazine 2, is a useful starting material to intro-
duce the sym-heptazine unit, an electron-poor and thermally
extremely stable fragment. In fact, tri-s-triazine is the most
stable member of the azacycl[3.3.3]azine series of compounds,
6
which are isoelectronic to the phenalene anion. The sym-
heptazine unit is most likely contained in melon, melem and
similar C=N=H materials, which also posses high thermal
ꢁ
stability, sometimes even to >500 C. Investigations on these
compounds date back to the 1830s. Not only then were
7
investigators challenged by the insolubility and chemical
inertness—possibly one reason why their structures remained
unclear until today. Pauling and Sturdivant were the first to
suggest a coplanar tri-s-triazine unit as the basic structural
8
motif of these compounds. Pauling apparently maintained an
interest in tri-s-triazine, for reasons unknown, as the molecular
formula of compound 3 was preserved on his office chalkboard
Scheme 1 Comparison of the triazine and tri-s-triazine derivatives
cyanuric chloride 1, trichloro-tri-s-triazine 2, as well as cyanuric acid
5ab and cyameluric acid 6ab. The parent tri-s-triazine molecule 4 and
an unknown azido derivative 3, which was found on the chalkboard of
Linus Pauling after his death, are also depicted.
9
at the time of his death in 1994.
To our knowledge, there is only one tri-s-triazine compound
described in the literature that has been comprehensively
characterised, namely the molecule tri-s-triazine, C N H 4. It
6
7
3
was synthesised by a three-step route starting from 2,4-di-
6
amino-1,3,5-triazine and methyl N-cyanomethanimidate. Its
y Part II: Alkalicyamelurates M
3
[C
6
N
7
O
3
], M ¼ Li, Na, K, Rb, Cs,
structural and spectroscopic properties including an X-ray
crystal structure analysis were discussed. UV photoelectron
spectra were taken and ab initio calculations performed in
manuscript in preparation.
z Current address: Materials Department, University of California,
Santa Barbara, CA 93106, USA. E-mail: kroke@hrzpub.tu-darm-
stadt.de; Fax: þ01 805 893 8486; Tel: þ01 805 893 4341.
6
order to explain the low basicity and the high stability.
5
08
New J. Chem., 2002, 26, 508–512
DOI: 10.1039/b111062b
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2002

A search in the CAS Registry database gave 129 substances
View Online
1
containing the tri-s-triazine unit. However, many of these
0
compounds have been only briefly mentioned in communica-
5
tions or patents and were not characterised at all. Some have
been examined by the determination of melting points, C=H=N
analysis and IR spectroscopy. This is also true for trichloro-tri-
s-triazine 2, which was first prepared by Redemann and Lucas
in 1940 by chlorination of cyameluric acid 6ab or its tri-
5d
potassium salt with PCl
5
.
The product was not characterised
except for its N and Cl content. It was reported to be impure,
containing small amounts of phosphorus. According to the
authors compound 2 could not be purified by sublimation or
5
d
re-crystallisation from organic solvents.
Schroeder and
PCl instead
of PCl , which was unsuccessful. Therefore, they treated an
Kober tried to improve this synthesis using C
H
6 5
4
5
b
5
5
excess of PCl with the tripotassium salt of cyameluric acid and
purified the product by washing it very rapidly with ice water.
The dried product was characterised without further purifica-
tion by determination of its melting=decomposition point and
the chlorine content. It was claimed in a patent that the
chlorination of potassium cyamelurate can be performed in
Fig. 1 Ortep plot of 2 (50% probability) indicating the distortion of
the six-membered triazine rings. However, the molecule is completely
planer as shown by the sum of bond angles of 360 for all C and N
atoms.
ꢁ
5e
an organic solvent like o-dichlorobenzene. Another patent
describes a different route to 2 that is based on a high tem-
perature transformation of 1 or cyanogen chloride ClCN using
5f
charcoal as a solid phase catalyst.
Table 1 Average bond lengths and bond angles for molecule 2 (see
Fig. 1)
We prepared 2 by a modified method similar to those
described in ref. 5(b–d), followed by purification through
sublimation and recrystallisation (see Experimental). The
bright yellow crystalline product was characterised using MS,
a
+
+
b
a
ꢁ
ꢁ
b
Bond length
d =A (s=A)
Bond angle
ff = (s= )
0
1
3
N(1)–C(1)
C(1)–N(2)
C(2)–N(2)
C(2)–Cl
1.397 (0.004)
1.334 (0.004)
1.328 (0.008)
1.716 (0.004)
C(1)–N(1)–C(1)
120.0 (0.6)
119.7 (0.5)
115.1 (0.4)
130.2 (0.4)
114.9 (0.4)
FTIR, C-NMR, UV=Vis absorption and photoluminescence
PL) spectroscopy as well as elemental analysis and single
N(1)–C(1)–N(2)
C(1)–N(2)–C(2)
N(2)–C(2)–N(2)
Cl–C(2)–N(2)
(
crystal X-ray diffraction. The C=N element content was in
excellent agreement with the theoretical values, and IR data
correspond well with the reported spectra, including three
characteristic vibrations for the tri-s-triazine unit at 1610, 1505
0
a
Avarage values for all equivalent bond lengths and angles of
molecules 1 through 3 (see Experimental). The sigma values refer to
b
À1
13
and 825 cm . T he C NMR signal at 159 ppm is due to the C
atoms bonded to three N neighbours, which is almost identical
to the value reported for tri-s-triazine. A resonance for the Cl-
bonded C atoms was detected at 176 ppm, which is at lower
field compared to the unsubstituted molecule as expected. The
yellow colour of compound 2 is caused by an absorption band
with a maximum at 310 nm. Photoluminescence was detected
at 468 nm.
the calculated mean values (the experimental sigma values for each
bond length and angle are available together with the crystallographic
data as ESI.
Table 2 Intermolecular ClÁ Á ÁN distances found in the crystals of 2
+
with d < 3.5 A. The first twelve lines refer to molecules within layer
B, while the last two lines are related to NÁ Á ÁCl contacts between the
For X-ray structure analysis, suitable crystals were grown
by slow diffusion of diethyl ether vapour into a solution of 2
in acetone (see Experimetal). The structure solution in the
monoclinic space group C2=c revealed the expected molecule
layers A and B (see Fig. 2)
a
+
a
ff=ꢁ
ClÁ Á ÁN contact
d=A
Bond angle
1
2
with an asymmetric unit comprised of 3 molecules of 2, half a
molecule of diethyl ether and a molecule of acetone. Since
Z ¼ 8 there are thus 28 molecules of 2, four of diethyl ether and
eight of acetone in the unit cell. All trichloro-tri-s-triazine
molecules are planar as indicated by the sums of bond angles
at each C and N atom bonded to three neighbours being
Cl(3)Á Á ÁN(9)
Cl(4)Á Á ÁN(18)
Cl(8)Á Á ÁN(7)
Cl(4)Á Á ÁN(5)
Cl(8)Á Á ÁN(11)
Cl(9)Á Á ÁN(12)
Cl(9)Á Á ÁN(14)
Cl(7)Á Á ÁN(7)
Cl(2)Á Á ÁN(17)
Cl(1)Á Á ÁN(9)
Cl(6)Á Á ÁN(19)
Cl(4)Á Á ÁN(24)
Cl(2)Á Á ÁN(23)
3.070
3.091
3.111
3.257
3.344
3.369
3.373
3.387
3.456
3.458
3.492
3.411
3.429
C(6)–Cl(3)Á Á ÁN(9)
C(8)–Cl(4)Á Á ÁN(18)
C(16)–Cl(8)Á Á ÁN(7)
C(8)–Cl(4)Á Á ÁN(5)
C(16)–Cl(8)Á Á ÁN(11)
C(18)–Cl(9)Á Á ÁN(12)
C(18)–Cl(9)Á Á ÁN(14)
C(14)–Cl(7)Á Á ÁN(7)
C(4)–Cl(2)Á Á ÁN(17)
C(2)–Cl(1)Á Á ÁN(9)
C(12)–Cl(6)Á Á ÁN(19)
C(8)–Cl(4)Á Á ÁN(24)
C(4)–Cl(2)Á Á ÁN(23)
150.4
164.0
148.6
98.6
93.7
67.6
93.3
120.4
95.0
121.0
105.3
118.9
138.0
ꢁ
ꢁ
3
60 (± 0.2 ). This is due to the aromaticity of 2 based on the
1
considerably distorted (Fig. 1 and Table 1). The average C–N
4 p electrons. However, the three triazine units of 2 are
+
bond lengths to the central N atom of 1.397 A are longer than
the (Cl)C–N bonds (1.328 A) as well as the remaining C–N
+
+
bonds (1.334 A). Besides, the endocyclic N–C–N angles of the
ꢁ
C atoms bonded to the chlorine atoms are larger (ꢂ130 ) than
ꢁ
a
the endocyclic C–N–C angles of ꢂ115 at the peripheral N
The atom numbering refers to the crystallographic data available as
2
atoms, which is to be expected due to the lone pair sp orbitals
ESI.
located on these nitrogen atoms.
Azaaromatic chlorides, which are aromatic compounds
containing many N=C–Cl subunits, are a remarkable class of
1
compounds. Like cyanuric chloride 1 they frequently show
1
arrangement of the molecules in the unit cell can be described
by ABAB layers that are parallel to the ab-plane (Fig. 2).
Within layer A the molecules of 2 are almost coplanar forming
thin sheets. Since solvent molecules separate these molecules
there are no ClÁ Á ÁN interactions within these layers. Twelve
+
short intermolecular ClÁ Á ÁN contacts of 3.0 to 3.5 A displaying
1
1
extensive donor-acceptor networks in the solid state. We
therefore examined the intermolecular distances in the crystals
of 2 and found indeed several such contacts (Table 2). The
New J. Chem., 2002, 26, 508–512
509

For the integration over the Brillouin zone we used the k -Vp i oe wi n Ot nline
1
scheme of Monkhorst and Pack. Forces were relaxed to
9
À2
À1
values below 1 ꢃ 10 eV A and stresses below 1 kbar. We
confirmed the achieved structural properties at this level of
accuracy by using finer grids for the k-point sampling and
larger plane-wave basis sets. The total energy differences are
À1
found to be converged to values better than 1 kJ mol . T he
approach we used has been proven to successfully calculate
structure, relative energies, and properties of solid-state
2
0
materials, defect structures, and surfaces. Our own experi-
ence in this field stems from several nitride systems, such as
1
6
21
22
carbon nitride, silicon nitride, and phosphorous nitride.
Previously, it was shown that 7 is the C structure with
While we do reproduce all the relative
energies of diamond-like and graphitic-C N structures pre-
3 4
N
1
4,15a
lowest energy.
3
4
viously suggested, we now find that a tri-s-triazine based
layered structure 8 is energetically more favorable, by ꢀ30 kJ
À1
Fig. 2 View along the b-axis of the unit cell. The layers A and B can
clearly be distinguished (further layers have been omitted for clarity
and space limitations). Due to the solvent molecules in layer A there
are no ClÁ Á ÁN interactions found within this layer. A close packing of
mol in comparison to 7, and thus should exist. We tested for
several different stacking variants of the layered structure 8,
and found the structure with lowest energy to have space
group Cmc21, for which an ortho-hexagonal setting is due to
the stacking sequence. Similar to hexagonal boron nitride,
nitrogen atoms avoid being stacked on top of each other so as
to reduce repulsive interactions along the stacking direction.
Due to the size of the structural building block, however, there
exists at least three different stacking variants based on the
motif 8, all separated by an energy difference of less than 10 kJ
molecules of 2 in layer B corresponds to several weak ClÁ Á ÁN contacts
+
of 3.0 to 3.5 A.
+
different ClÁ Á ÁN contacts of 3.070 to 3.492 A with C–ClÁ Á ÁN
ꢁ
ꢁ
angles of 67.6 to 164.0 are present within the layer B. This is
due to the absence of solvent molecules. Interestingly, there are
only two rather long ClÁ Á ÁN contacts between the two layers.
À1
mol . We, therefore, suggest that in many cases, because of
the high temperatures usually used to prepare carbon nitrides
by CVD, PVD or bulk synthesis, tri-s-triazine based structures
are formed. The experimental data should be re-examined
carefully, comparing them to the data of tri-s-triazine and
tricholoro-tri-s-triazine 2.
Based on ab initio calculations it was suggested that the
À1 11
strength of each of the weak contacts is about 5 kJ mol
Several compounds may be prepared from 2, for example
.
1
graphitic C N structures. These materials are expected to be
2
3
4
suitable precursors for the high pressure synthesis of superhard
C N phases in analogy with the conversion of graphite and
Currently, we are trying to use compound 2 for reactions
with nitrogen sources like Li
Varying and very interesting results have been reported for
N or N[Si(CH ) ] to form 8.
3 3 3 3
3
4
h-BN to diamond and c-BN, respectively. Theoreticians have
proposed several structural models of C N to have proper-
1
2
similar attempts to obtain 7 from 1. Similarly, 11 (or 12) may
be obtained by reacting 2 with 1.5 equiv of bis(trimethyl-
3
4
1
3,14
ties comparable to diamond.
This motivated numerous
1
2
23
silyl)carbodiimide for example, as was tried for 9. Besides, we
have prepared a single source precursor for 8 by reacting 2
experimentalists to try to synthesise these phases. Typical
products of most attempts are ‘‘graphite-like’’ materials,
which are believed to be based on the s-triazine network 7
with 1 equiv of LiN[Si(CH
triazine compounds. The C N
3 4
3
)
3
]
2
, again similar to analogous
-structures 15 (or 16), con-
taining hepatzin and triazin units, may be synthesised from
2
4
(
derived from triazine rings connected by N-atoms, has been
Scheme 2). Another carbon(IV) nitride structure 13, also
1
5
suggested recently. Alternatively, the C
25
3
N
3
rings may be
interconnected with carbodiimide (N=C=N) or cyanamide (N–
3 6 9
tricyanmelamiates like Na [C N ] and the title compound 2.
A detailed report about these studies will be published else-
where. Very recently similar approaches to carbon(IV) nitrides
C:N) groups to form the (soft) 2D or 3D networks 9 and 10
with C N stoichiometry and an alternating C–N arrange-
2
6
have been reported by Komatsu. However, none of the
attempts was successful so far, since elemental analyses
indicated that all products contained significant amounts of
residual hydrogen.
3
6
4
1
3 4
ment. Analogous 2D or 3D C N materials are formed by
connecting the tri-s-triazine units with N atoms (8, 14),
carbodiimide (11) or cyanamide (12) groups (Scheme 2).
Furthermore, C N structures based on both triazine and tri-
3
4
s-triazine units may be postulated, as indicated by the net-
works 15 and 16. To our knowledge, none of the authors
Experimental
3 4
discussing graphitic C N structures have ever tried to explain
the experimental results in terms of these tri-s-triazine con-
taining structures.
Synthesis of 2
To access the likeliness of such structures we performed total
energy calculations of various structures using density func-
Potassium cyamelurate was synthesised by refluxing a sus-
pension of 50 g melon powder (obtained from Degussa AG,
Germany) in 500 ml of a 2.5 molar aqueous KOH solution
for 45 min. The hot reaction mixture was filtered and upon
1
7
tional theory. We used a standard ab initio total energy and
molecular dynamics program (VASP, Vienna Ab-initio Simu-
lation Package) for this purpose, which combines ultra-soft
ꢁ
slow cooling to 0 C fine needles precipitated from the filtrate.
1
8
pseudo-potentials with a plane-wave basis set. The exchange
correlation energy of the electrons is treated within the local
density approximation (LDA). The generalised-gradient appro-
ximation (GGA) yields almost identical results for energy
differences, since no significant changes in co-ordination (e.g.,
tetrahedral vs. octahedral co-ordination) are present among
the investigated structures. All structures we considered were
completely optimised using a cut-off of 400 eV for the
expansion of the wavefunction into the plane-wave basis set.
The potassium cyamelurate was separated, washed with
EtOH, dried and identified by IR spectroscopy and C=H=N
analysis. The potassium cyamelurate (18.7 g) was intimately
mixed with 42.27 g of PCl
walled Schlenk tube that was subsequently evacuated to 10
5
. The mixture was filled in a thick
À1
ꢁ
mbar. The closed tube was heated to 130 C (caution!). T he
pressure generated inside the tube was carefully released
several times followed by evacuation. After 10 h the solid
reaction product was quickly washed with ice water. Drying
5
10
New J. Chem., 2002, 26, 508–512

View Online
3 4
Scheme 2 Triazine and tri-s-triazine based C N structures, which can be considered as ‘‘graphitic’’ carbon(IV) nitrides containing an alternating
arrangement of C and N atoms. None of these phases has been unambiguously proven to exist. Most authors tried to explain experimental results
in terms of 7. DFTcalculations indicate that the tri- s-triazine based structure 8 seems to be more stable than 7.
New J. Chem., 2002, 26, 508–512
511

over P
2
O
5
afforded 14 g (91%) of 2. The crude product was
Notes and references
View Online
À2
À3
ꢁ
purified by sublimation at 10 –10 mbar and 240–290 C.
We also found that Soxhlet extraction with toluene may be
used to purify 2. EA (Labor Pascher, Remagen, Germany):
C: 26.22 (calcd: 26.07), N: 35.5 (calcd: 35.46); MS (EI) m=z:
1
2
N. O. V. Sonntag, Chem. Rev., 1953, 52, 251.
(a) L. Person, Ind. Chim., 1966, 53, 235; (b) K. Huthmacher and
D. Most, in Ullmann’s Encylopedia of Industrial Chemistry, Wiley-
VCH, Weinheim, Germany, 2002, online-version, chapter
þ
þ
þ
13
2
77.5 (100%, MH ), 276.5 (M ), 240 (MH À Cl); C NMR
‘‘Cyanuric acid and cyanuric chloride’’.
(
(
D
8
-THF, 300 MHz): d 158.23 (CN
3
), 175.00 (CClN
2
); FTIR
KBr pellet, cm ): 1610(vs), 1505(vs), 1310(s), 1205(s),
3
Synonyms: tri-s-triazine, 1,3,4,6,7,9,9b-heptaazaphenalene, 1,3,4,
6,7,9-hexaazacycl[3.3.3]azine or cyamelurine.
W. Henneberg, Annalen, 1850, 73, 288.
À1
4
5
9
41(s), 825(m); PLS (T ¼ 290 K) lmax,abs 310, lmax,emission 466
(a) A. I. Finkelshtein and N. V. Spiridonova, Russ. Chem. Rev.,
1
nm (t < 700 ns).
964, 33, 400; (b) H. Schroeder and E. Kober, J. Org. Chem., 1962,
7, 4262; (c) H. Schroeder (Olin Mathieson Chem. Corp., Virgi-
2
nia, USA), US Pat., 3089875, 1963; (d) C. E. Redemann and H. J.
Lucas, J. Am. Chem. Soc., 1940, 62, 842; (e) R. Neef (Bayer AG,
Leverkusen, Germany), Ger. Pat., DE 1102321, 1961; ( f ) C.
Gremmelmaier and J. Riethmann (Ciba-Geigy, New York, USA),
US Pat., US 4205167 A, 1980.
(a) M. A. Rossman, N. J. Leonard, S. Urano and P. R. LeBreton,
J. Am. Chem. Soc., 1985, 107, 3884; (b) R. S. Hosmane, M. A.
Rossman and N. J. Leonard, J. Am. Chem. Soc., 1982, 104, 5497;
(c) M. Shahbaz, S. Urano, P. R. LeBreton, M. A. Rossman, R. S.
Hosmane and N. J. Leonard, J. Am. Chem. Soc., 1984, 106, 2805;
Crystal structure determination
For X-ray diffraction a plate-shaped crystal was placed under
Exxon Paratone N oil, mounted onto a glass fiber, and moved
to the goniometer head of a Siemens SMARTCCD dif-
6
fractometer. The structure was solved using direct methods
2
and refined by full-matrix least-squares on
F
with
2
7
SHELXTL. A detailed examination of the refined data
revealed that molecules 1 through 3 all have well-behaved
displacement parameters, while the half molecule that is
located on the twofold rotational axis shows somewhat larger
ellipsoids elongated in the direction perpendicular to the
symmetry axis. Therefore, we carefully re-examined the
experimental data for possible supercells. Besides, the system
was re-refined in space group Cc, the only sensible second
choice. Geometry was distorted, a common result when
refining a centrosymmetric system in a non-centrosymmetric
space group. The 2-fold symmetry was readily apparent.
PLATON found 100% fit in the original space group C2=c.
Therefore, we suppose that static disorder of the molecule on
the rotational axis is the reason for its slightly elongated
ellipsoids. The structural parameters presented in Table 1 are
confined to molecules 1 through 3 and exclude the half mole-
cule 4, although the deviations of bond lengths and bond angle
(
Chem., 1984, 88, 4324.
(a) J. Liebig, Ann. Pharm., 1834, 10, 10; (b) L. Gmelin, Ann.
Pharm., 1835, 15, 252.
L. Pauling and J. H. Sturdivant, Proc. Natl. Acad. Sci. U.S., 1937,
d) M. A. Rossman, R. S. Hosmane and N. J. Leonard, J. Phys.
7
8
2
3, 615.
9
0
Chem. Eng. News, Aug. 7th, 2000, p. 62 and Oct. 2nd, 2000, pp. 8–9.
Chemical Abstract Service (CAS), Columbus, OH, registry data-
base.
1
11 K. Xu, D. M. Ho and R. A. Pascal, Jr., J. Am. Chem. Soc., 1994,
16, 105.
Recent reviews: (a) T. Malkow, Mater. Sci. Eng., A, 2001, 302,
11; (b) S. Matsumoto, E.-Q. Xie and F. Izumi, Diamond Relat.
1
1
2
3
Mater., 1999, 8, 1175; (c) S. Matsumoto, E.-Q. Xie and F. Izumi,
Diamond Relat. Mater., 2000, 9, 94; (d) L. R. Shaginyan, Powder
Metall. Met. Ceram., 1999, 37, 648; (e) P. F. McMillan, Curr. Opin.
Solid State Mater. Sci., 1999, 4, 171; ( f ) V. J. Badding, Adv.
Mater., 1997, 9, 877.
1
3
A. Y. Liu and M. L. Cohen, Science, 1989, 245, 841.
+
14 D. M. Teter and R. J. Hemley, Science, 1996, 271, 53 [graphite-
like C structure based on 7 (Scheme 2), space group P6m2].
ꢁ
values are smaller than 0.02 A and 1 , respectively. Crystal
1
3 4
N
1
2
data for 3 2Á acetoneÁ diethyl ether: C26
24.50 1.50
H11Cl10.50N O ,
2
15 (a) L. Alves, G. Demazeau, B. Tanguy and F. Weill, Solid State
Commun., 1999, 109, 697; (b) M. Mattesini, S. F. Matar and
J. Etourneau, J. Mater. Chem., 2000, 10, 709.
1 P. Kroll and R. Hoffmann, J. Am. Chem. Soc., 1999, 121, 4698.
17 P. Hohenberg and W. Kohn, Phys. Rev. A, 1964, 136, 864.
M ¼ 1062.82, monoclinic, space group C12=c1, a ¼ 23.2550(3),
+
ꢁ
b ¼ 13.4512(2), c ¼ 26.8143(4) A, b ¼ 111,4770(10) , U ¼
+
3
À3
6
7
805.31(19) A , T ¼ 142 K, Z ¼ 8, D ¼ 1.809
g
cm ,
c
À1
m(Mo-Ka) ¼0.815 mm , 40 278 reflections measured, 8520
unique (R ¼ 0.0997). The final R values were R ¼ 0.1148,
1
8
(a) G. Kresse and J. Hafner, Phys. Rev. B, 1993, 47, 558; (b) G.
Kresse and J. Hafner, Phys. Rev. B, 1994, 49, 14 251; (c) G. Kresse
int
1
wR
6
2
¼ 0.1511 [for all data] and R
1
¼ 0.0798, wR
2
¼ 0.1360 [for
¨
and J. Furthmuller, Comput. Mater Sci., 1996, 6, 15; (d) G. Kresse
¨
and J. Furthmuller, Phys. Rev. B, 1996, 55, 11 169.
354 data with I > 2s(I)].
CCDC reference number 175532. See http:==www.rsc.org=
19 H. J. Monkhorst and J. D. Pack, Phys. Rev. B, 1976, 13, 5188.
0
2
M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias and J. D.
Joannopoulos, Rev. Mod. Phys., 1992, 64, 1045.
suppdata=nj=b1=b111062b/ for crystallographic data in CIF or
other electronic format.
2
1
A. Zerr, G. Miehe, G. Serghiou, M. Schwarz, E. Kroke, R. Riedel,
H. Fueß, P. Kroll and R. Boehler, Nature, 1999, 400, 340.
P. Kroll and W. Schnick, manuscript in preparation.
(a) R. Riedel, E. Kroke, A. Greiner, A. O. Gabriel, L. M.
Ruwisch, J. Nicolich and P. Kroll, Chem. Mater., 1998, 10, 2964;
2
2
2
3
Acknowledgements
(
b) E. Kroke, A. Greiner and R. Riedel, US Pat. 6.127.510, 2000.
2
4
(a) J. Kouvetakis, A. Bandari, M. Todd and B. Wilkens, Chem.
Mater., 1994, 6, 811; (b) M. Todd, J. Kouvetakis and T. L. Groy,
Chem. Mater., 1995, 7, 1422.
The authors gratefully acknowledge financial support by
the Deutsche Forschungsgemeinschaft (Kr 1739=2–3) and
Fonds der Chemischen Industrie. E. K. is indebted to the
A. v. Humboldt Foundation for granting a Lynen Fellowship
as well as to R. Riedel (Darmstadt) and F. F. Lange (Santa
Barbara) for their generous support. Dr R. Schmechel
25 B. Ju
2
¨
rgens, E. Irran, J. Schneider and W. Schnick, Inorg. Chem.,
000, 39, 665.
(a) T. Komatsu, J. Mater. Chem., 2001, 11, 799; (b) T. Komatsu,
J. Chem. Mater., 2001, 11, 802.
G. M. Sheldrick, SHELXTL, A Program for Crystal Structure
Determination, v. 5.03, Siemens Analytical X-ray Instruments,
Madison, WI, 1995.
2
6
7
2
(Darmstadt) is acknowledged for performing PL spectro-
scopy.
5
12
New J. Chem., 2002, 26, 508–512