close

Вход

Забыли?

вход по аккаунту

?

Calixarene-Based Macrocyclic Nonet (S = 4) Octaradical and its Acyclic Sextet (S = 52) Pentaradical Analogue.

код для вставкиСкачать
COMMUNICATIONS
and [Cp(dppe)FeCN] in THF and recrystallized from CH,Cl,/'
petroleum ether in 62 "/o yield. Both isomers of 2. just like those
of I , cannot be interconverted thermally up to their decomposition points around 150 "C. 2 a (red) and 2 b (yellow) differ clearly
in their colors and in their optical absorption maxima (472 nm
for 2 a and 401 nm for 2 b ) . This time the fi(Cp) and v(CN)
v ~ ~ u K sare
~ "quite
'
similar, and the electronic difference is mostly
reflected in the CO vibrations.['I Specifically the A t band of the
Cr(CO), group which corresponds to the CO vibration /runs to
C N is about 30 cm-' higher for 2 a (1898 cm-') than for 2b.
This might be expected because N-bonded C N is a weaker
acceptor than C-bonded CN. A comparison of 2 b with
[(CO),CrNCMe] ( A : bands at 1869 and 1930 cm-', respectively) reveals clearly that the "ligand" [Cp(dppe)FeCN] in 2 b is
extremely electron-rich. Irrespective of this, the HOMOS of 2 a
and 2 b seem to be lower in energy than those of 1 a and 1 b. since
both complexes 2 are about 0.2 V more difficult to oxidize than
their counterparts
reveal that
The structure determinations of 2 a and 2b'".
it is almost impossible to distinguish the compounds in the solid
state. The crystals are isomorphous (but clearly distinguished by
their colors), and both structures can be presented by one drawing (Fig. l ) . The bonding parameters are all normal. The differ-
($
@=-a-
Fig. 1. Molecular geometry of 2 a and 2b. Pertinenl distances [pm] and angles [ ]
(values for 2 b in ?quare brackets) Cr-CCr-N 206.4(5) [208.6(3)].C-N 115.8(7)
[ I IS.l(S)]. Fe-N'Fe-C 193.5(4) [18Y.7(4)], Cr-C(CO t r ( i m ) 183.8(6) [184.4(5)], CrC ( C 0 ( I \ . iiverage) 188.6(6)[189.5(5)].Fe-P (average) ?20.2(1) [218.5(1)]: Cr-C(N)N(C) 1 70.7(4) [I 65. 1( 3 )I. Fe-N( C)-C(N) 169 .8(4) [I 74.3(4)]
ences between 2 a and 2 b are almost insignificant, but the trends
in the Cr-CN-Fe backbones are as expected. Thus, the M - N
distances are longer than the corresponding M -C distances
because ofweaker M-N vs. M-C backbonding, and the Fe-C
bond in 2 b is the shortest of the M-(CN) bonds because the
electron-rich iron unit has the most backbonding to offer, while
the C-N bond length shows very little variation. The largest
difference between 2 a and 2 b lies in the deviations from h e a r ity at the bridging C and N atoms. We consider this bending of
the moiecules to result from packing forces. Isomorphism, a
similar bending. and the near-to-identity of all bonding parameters were also observed for the pair of complexes [(NH3)sCo(p-L)Co(CN),] ( L = CN, NC).'*]
The spectroscopic and structural similarity within the pair of
isomers 2 a and 2 b, which may be explained by their delocalized
nature. extends itself to their cationic species. The molecular
structure of the cation in [2b] [BF,]["] is again most similar to
that of 2 a and 2b. and a detailed investigation has revealed that
the radical ion 2 b + possesses a high degree of charge delocalization.""
The physical properties of the isomers and their tolerance of
a change in the electron count underline the ability of the
cyanide ligand to act as a mediator and transducer of electronic
effects. The position of the v(CN) vibration in the IR spectrumLY1
and preliminary extended-Huckel MO calculations['31
indicate that the bridging cyanide ligand can respond to the
donor/acceptor properties of both organometallic units by variation of its electronic situation without altering its bond length.
just as has been observed for many solid-state metal-cyanide
compounds."41 Thus. it does not seem to be utterly unrealistic
to search for "organometallic Prussian Blue".
Received: May 9. lY94 [Z 6912 IE]
German version: Anpc,ii.. Chc,ni. 1994. 106. 2166
[I] C. Creutz. Prop. lnorg. C%rm. 1983. 30. 1-73.
[2] W. P. Fehlhammer, M . Friti. Chrm. Krv. 1993. Y3. 1243--1280.
[3] F. L. Atkinson, A. Christofides. N.G. Connelly. H. J. Lawson. A . C. Loyns.
D u l t o ~ il k u n \ . 1993.
A. G Orpen. G . M. Rosair. G . H. Worth. J. Cheni..%I(
1441-1450, and references therein: F. Scandola. R. Argazri, C. A. Bignozri. C .
Chiorboli. M. T. Indelli, N. A. Rdmpi. c'oorif. c'him. R e v 1993. 125. 283-292.
and references therein.
[4] First communication: B. Oawald. A. K. Powell. F. Rashwan. J. Heinie. 11.
Vdhrenkamp. Cheni. B w . 1990. 123. 243- 250.
[S] For a case of terminal M -NC + M -CN isomerization cl: S. Alviarez. C'.
Lopez. /iior~q. Chim. A c f o 1982, 64. LYV-LIOO
[6] Among the inert classical complexes two such pairs have been reported:
[(H,O),Cr(/i-CN)Co(CN),1 (spectroscopic characterrzition) [7] and
[ (NH,),Co(/i-CN)Co(CN),1 (structure determinations) [8].
[7] D. Gaswick, A. H a m . J. h i o , ~ N. u d . C'hetit. 1978. 40, 437- 439.
[8] F. R. Fronczek. W. P. Schaefer. /tior,q Client 1974. 13, 727 732.
[9] I R (CH,CI,): r(CO): l a : F[cm-l] = 2062s. 2017s. 1920s, 1843s: I b 2063\.
2018s. 1922s. 1854s; 2 a 2 0 5 6 ~ .1 9 3 0 ~ s 1898m:
.
2b: 2064%. 192Xvs. 1869ni.
i>(CN):la:2147vw: l b : 2 0 9 9 r n : 2 a : 2 1 1 5 ~ ~ : 2 b : 2 1 0 3 w' H
. NMR(h(Cp)TMS
mt., CDCI,): l a ' 5.07. 4.41; Ib: 5.08. 4 40: 2 a . 4.15: 2 b - 4.22. Cyclovoltammetric data (first oxidation. Evs. Ag;AgCl. CH,CN): I a : 0.1 1 V (quasi-rev.).
1 b: 0.35 V (rev.): 2 a : 0.31 V (rev.): 2 b : 0.53 V ( m e \ . ) .
[lo] N. Zhu. H. Vdhrcnkamp, J. Orjiimonicr. Chcm.1994. 472. CS C7.
[ I l l Crystal data: 2 a : space group P 2 , : c . 0=1501.4(1). /J =1299.5(1).
c =1718.2(l)pm. /l
= 95.49(1) . 5034rcflcctions. 424 parametcrs. R = 0.068.
2 b . space group P2,,c. u =1500.7(1). h =1297.6(1). ( =1716.4(2) pm. =
Y5.60(1) , 5314 reflectionc. 424parameters, K = 0.045 [I21
[I21 Further details of the crystal structure investigation muy be obtained upon
request from the Fachinforinalionsrentrum Karlsruhe. D-76344 EggensteinLeopoldshafen (FRO). on qnotinz the depository numbers CSD-380055 (for
2 a ) and CSD-380056 (for 2b).
1131 C Marchand. H Grutrmacher. private communication
1141 A . M . Golub, H. Kohler, V. V. Skopenko. C ~ I z m i i i s t r i(.I / P\e~uddiulirL.s( 7 i p . . \
in Iiiwquiiic fiiiil Gcnwiil Chiwti.vtrj, I id. 21 (Ed : R J. H. C l x k ) ) . Elsevier.
Amsterdam. 1986. pp. 77 185.
Calixarene-Based Macrocyclic Nonet (S = 4)
Octaradical and its Acyclic Sextet (S = 5/2)
Pentaradical Analogue**
A n d r z e j Rajca,* Suchada Rajca, a n d Raghavakaimal
Padmakumar
Very high-spin organic molecules are of current interest in
relation to fundamental aspects of bonding and novel approaches to materials.['] Synthesis of mesoscopic-size and well-defined
high-spin organic molecules is one of the ultimate goals in the
area of organic magnetism.12.31 The organic molecule with the
[*I
[**I
Prof A. Rajca. Dr. S. Rajca. Dr. R. Padinakumar
Department of Chemistry. University of Nebraska
Lincoln. N E 68588-0304 (USA)
Telekx: Int. code +(402) 472-9402
This research was supported by the National Science Foundation (CHE920391 8). Mass spectral determinations were performed by the Midwest Center for Mass Spectrometry with partial support by the National Science Foundation (DIR-9017262). Wc thank Professor S. H. Liou for access to a SQUID
ma_metometer.
COMMUNICATIONS
highest spin quantum number S is the nonacarbene ( S = 9)
prepared by Iwamura et al.I4]So far, strong ferromagnetic coupling could be maintained in molecules with up to around ten
carbene or radical sites.[4' Dendritic polyradicals with up to 31
radical sites were reported to possess lower than expected spin
values. It was suggested that one of the possible reasons for the
interruption of the spin coupling was the presence of defects,
which were defined as failure to generate a radical site.['] In
acyclic spin-coupled structures, such as typical organic highspin molecules, even a single defect may interrupt or weaken the
only available pathway for spin coupling.16* This is in contrast
to inorganic clusters or extended networks,['] where an alternative pathway (to that containing the defect) is possible, as long
as the number of defects is moderate.
We now report on two macrocyclic high-spin organic molecules, the calix[4]arene-based nonet ( S = 4) octaradical 1 and its
acyclic analogue 2.19]Both polyradicals are designed to mini-
+
2
mize the potential impact of defects on spin coupling (conjugation): in l cyclic connectivity and in 2 the number of terminal
sites was maximized. Spin coupling in the above polyradicals is
interrupted when a defect divides the spin system into at least
t h o uncoupled spin systems with lower spin. Because the minimum requirements for such interruptions (namely, more than
me defect at the four-site calix[4]arene ring in 1 and one defect
alt the solitary central site in 2) are quite stringent, we expect to
dhtain high-spin polyradicals with negligible amounts of low
spin ( S =1/2, 1, etc.) contaminants.
First, polyethers 9 and 12 were synthesized (Scheme 1). In the
key step, the assembly of calix[4]arene ring, a single isomer of 8
was isolated in approximately 20 % yield."'] The characteristic
features of the 'H NMR spectrum of 8 are: one signal for the
hydroxyl groups, four for the methoxy groups (2/2/1/1), and
seven for the tBu groups (2/2/2/2/2/1/1); the symmetry in 8 is
corroborated by the 13CNMR spectrum. NMR spectroscopic
investigations confirmed the fact that the two equivalent hydroxyl groups were replaced with two equivalent methoxy
gfloups in the conversion of 8 to 9.[lo1
Treatment of octaether 9 with Na/K alloy in T H F or diethyl
ether and of pentaether 12 with Li metal in T H F for 2-4 days
gave solutions of the corresponding carbopolyanions. This was
confirmed by the quenching products 10 and 13 obtained with
dsuterated methanol. Addition of one equivalent (or a small
excess) of iodine to the carbopolyanions in T H F at 180 K gave
polyradicals 1 and 2, respectively.[51When a solution of 1 in
TMF/2-MeTHF was allowed to react with an excess of Na/K
alloy at low temperature and then quenched with methanol,
nandeuterated isotopomers of 10 were obtained; this suggests
that the macrocyclic ring is intact in 1.
The Am, = 1 regions of the electron spin resonance (ESR)
spectra for 1 and 2 in T H F or THF/2-MeTHF (10-4-10-3 M)
0 VCH Verlugsgeseiischuft mbH. 0-69451 Wernheim, 1994
t,
I E
lF
Scheme 1 . Synthesis of the polyethers: A: 1) nBuLi/ether, 2) 4,4'-di-rerf-butylbenzophenone; B: NaH/THF, MeI; C: 1 ) fBuLi/ether, 2) methyl 4-rerl-butylbenzoate;
D : 1) tBuLilTHF, 2) 1-(4-tert-butylbenzoyl)pyrrolidine;E: 1 ) 6/tBuLi/THF, 2) 7; F:
1) nBuLi/THF, 2) 4,4'-di-tert-butylbenzophenone;
G : 1) 9/Na/K in THF or ether,
2) MeOD; H: 1) lZ/Li in THF, 2) MeOD.
%
% ' %
2092
+ +
's3*
at 77 K are not sufficiently resolved to permit a meaningful
spectral fit for these sterically congested polyradicals.[' 2a1 A
weak Ams = 2 signal is obseived for 1but not for 2 at 77 K; this
is what one would qualitatively expect for such polyradicals
with integral (S = 4) and half-integral (S = 5/2) spin quantum
numbers.['2b1 The product of the intensity I of the Ams = 2
signal for 1 and temperature T is constant in the range T = 2080 K.
The bulk magnetization ( M ) of solutions of 1 and 2 in THF
(8 x 10-4-1 x
M) were measured as a function of magnetic
field H (0-5.5 Tesla) at T = 2, 3, 5, and 10 K and as a function
of temperature ( T = 2-35 K) at H = 1.0 and 0.5 Tesla. The
sdmple consisted of a band of a frozen polyradical solution (ca.
4 mm high) wedged between two bands of pure frozen solvent
(ca. 60 mm high) in a 4 mm OD flame-sealed quartz tube. Such
a sample does not allow for measurement of the amount of
thermally unstable polyradical, but gives an adequate signal in
the Quantum Design SQUID magnetgmeter and guarantees
almost complete cancellation of the diamagnetic contribution to
M.l2,51 An adequate correction for diamagnetism was made for
1. This was verified by comparing the plot of MTvs T ( T = 235 K) with that of IT vs T ( T = 20-80 K), where Z is the
Ams = 2 ESR-signal intensity for 1. For the most dilute solutions, the data for M vs HIT were fitted to a Brillouin function
with variable S and magnetization at saturation M,,,
.[I3] The
coincidence of the plots for T = 2, 3, 5, and 10 K is consistent
with the absence of intermolecular magnetic interactions in the
dilute solution. In the concentrated solutions, the curves for M
vs HIT at different temperatures do not overlap, and intermolecular antiferromagnetic interactions are accounted for with
a mean-field parameter 6' (6' < 0, 16'1 O.l), that is, HIT is replaced with H/(T - 0).1141The standard errors of the parame-
0570-0833/94/2020-2092$10.00+ .25/0
Angew. Chem. Int. Ed. EngI. 1994. 33, No. 20
COMMUNICATIONS
ters for the fits at each temperature are < 1 O h ; discrepancies
between the parameters for different temperatures and samples
are < 5 0'0, and the parameter dependence for S a n d M,,, ranges
from around 0.3 at T = 2 K to around 0.8 at T = 5 K."51 The
values for S obtained from fitting are 3.8 and 2.3 for 1 and 2,
respectively (Fig. 1 ) . In the absence of defects, high-spin
polyradicals I and 2 with all unpaired electrons coupled ferromagnetically would be expected to have S = 4.0 and 2.5. respec-
1.0
[2] A. Rajca, S. Utamapanya. J. ,4117.Chrrii Sot.. 1993. 113. 106x8.
[3] D. D. Aushalom. D. P. DiVincenzo. J. F. Sinyth. Scicnr.r 1992. 2SX. 414
[4] N . Nakamura. K . Inoue. H. Iwamura. An,qw C/io,r. 1993. 105. 900: Angcbi.
Chcm. I n ! . Ed. Engl. 1993. 32. 872
[5] A. Rajca. S. Utamapanya, S . Thayumana\wn. .I .4m. Chi~ni.Sw. 1992. 114.
1884.
[6] N. M a t a p , T / i ~ wChin?.
.
,4<./a1968, 10. 372: for acyclic tiiradicals urltll
S = 3/2 that are based upon 1.3.5-connected benzene. see J. Vcciana, C. Rovir;i.
N. Ventosa. M.I. Crespo. F. Palacio. J. A m Clicwr. So<. 1993. 11.7, 57, and
references therein: a tetracarbene based upon the calix[4]arene ring ia inentioned in ref. [l a]
[7] S. Sasdki. H. Iwamura. C'hrni Lt'rt. 1992. 1759.
[8] D. P. Goldberg. A. Caneschi, S. I . Lippard. J Am. Ch~wi.So( 1993. II3.9299.
[9] Culirurcrres A Evsutilc C l u s ~of tMowo<:v<./,c Compound\ (Ed
Bohmer). Kluwer. Dordrecht. 1991
[lo] a) Compounds 3-9. 11 and 12, have adequate 'H NMR. "C NMR. and
FABMS data. The IR spectra for compounds 3-9 confirm the presence o1'thc
O H and C - 0 groups. C:H analyses (3, 5. 6. and 9) or high resolution FABMS
(4) are satisfactory. h) Calix[4]arene 8. M.p. 132-133 C . FAB-US (ofdionitrophenyl octyl ether matrix), cluster: n : : (peak height) for (.W O H ) ' :
2196.3(3.5).2197.3(5),2198.3(4.5),2199.3(2.S);
f o r ( M O C H , ) ' : 2182.3(7).
2183..3(10). 2184.3(8.5), 2185.3(5): calcd for ( M O C H , ) ' , C , 5 q H , 8 5 0 7 :
2182.41(5.5), 2183.42(10), 2184.42(9). 2185.42(5). ' H N M R (300 M H L CDCI,. 20 C): d =7.45-6.50 (m. 60H), 2.821 (s, 2 H . exchange with D,O). 2.654
(s. 6 H ) . 2.629 (s. 6 H ) . 2.519 (s. 3H). 2.225 (s, 3 H ) . 1.258 (s. 18H). l . 3 Y (\.
9 H j . 1.203 (s. YH), 1 185 (s. 18H). 1.175 (5, 1 8 H ) . 1 163 (s, 18H). 1.127 (s,
1 8 H ) . "CC('H) N M R (75 MHL, CDCI,. 20 Cj: ii = 150.0-13X.O ( 2 0 \ignals).
130.0-124.0 (14 overlapping signals). 87.39. 86.97. 86.77. 86.62. X2.43. 52.40.
52.11. 51.87. 51.78, 35.0-31.0 (8 overlapping signals) I R (ZnSc):
i. = 3500cm-' (OH). - Calix(4larene 9: M.p. 166-167 C . calcd foiC,,,H,,,O,: C 85.67. H 8.639'0: found: C 86.05. H 8 88% FAB-MS ( O N POEj. cluster: in:: (peak height) for ( M - OCH,)': 2209.6(4). 2210.5(7).
2213.4(10), 2212.4(8). 2213.4(5). calcd for ( M O C H , ) ' . C,5,,H,8c,0,:
2210.44(5.5).2211.45(10). 2212.45(9), 2213.45(5). ' H N M R : A =7.30 6.XO
(m. 60H). 2.654 (s. h H ) , 2.641 (s. 6 H ) . 2.601 (5. 6 H ) . 7.560 (s. 3 H ) . 2.445 (s.
3 H ) . 3.208. 1.181. 1.165 (s. 108Hj " C [ ' H ; N M R . 5 =150.0 140O(l9overlapping signals). 129.0-124.0 (16overlapping signals). X7.24. 87.15, X7.13.
86.84. 86.78. 52.18. 52.03, 52.00. 51.98. 51 93, 35.0 31.0 (5overlapping
signals). IR. OH not detected.
[I I] Compounds LO and 13 have adequate ' H NMR and FABMS spectra: oiie of
the isomers of LO was separated. Reaction o f 9 with Li in THF gave complex
reaction mixtures. the FAB mass spectrum wggests incorporation of THF-like
fragments.
[12] a) With the exception ofone heptaiadical ( S =7:2. ref. [ j ] ) . all polyarylmethyl
polyradicals with S > 2 have unresolved dipolar couplings in their ESR spectra. b) One of the differences between the integral and half-integral spin system5
is the presence in the former of an intense center peak in the Am, = 2 region:
A Rajca, S. Utamapanya. J Am. Chenr. Sac 1993, 115. 2396.
[13] R. L. Carlin. M ~ ~ ~ f i " o r . h e ~ ~ i Springer,
irlrv.
Berlin. 1986. p. 14 18.
[14] A. Bino. D. C. Johnston. D. P. Goshorn. T. R. Halbert. E. I . Stiefel. S r r r n < c
1988. 241. 1479.
[lS] The reliability of the M vs II:Tfits is indicated by low parameter dependence
[dependence = 1 (variance of the parameter. other parameters constant):
(variance of the parameter. other parameters changing)]. For example. application of a standard technique (ESR signal intensity vs temperature in the 15
YO K range) to a favorable case of dirddical with J:h = 20 K gives a twoparameter fit with dependence of 0.6.
[16] One of the referees suggested fitting the magnetization data for I to a inixture
o f t h e polyradicals with S = 4 (defect-free 1) and S = 3.5 ( 1 with one defect).
which would correspond to a simplified version of the procedure in ref. [?I. For
a sample with S = 3.83 ? 0.03. such fitting gives 66% f 10% octaradical with
s = 4.
~
0.6
~
~
1
0.6
2
~
3
0.2
4-
5 - -
P
0.0
0.0
1 .o
H/T [TK-']
2.0
-
3.0
Fig. 1. Magnetization studies of I (8.5 x
M in T H F ) performed with a SQUID
magnetometer. The plots of M : M,,, vs H:Tare shown. The tit to a Brillouin function
with variable S gives S = 3.8. The solid line corresponds to the Brillouin function
with S = 4.
The magnetization data establishes that 1 and 2 have nonet
(S = 4) and sextet ( S = 5 / 2 ) ground states, respectively. The
strength of the ferromagnetic coupling in 1 may be judged by the
lack of a detectable thermal population of the low-spin excited
states up to T = 35 K (or 80 K ) , as determined by either SQUID
magnetometry or ESR spectroscopy. The magnetization data
for I are drastically different from those for polyradicals with a
comparable or larger number of sites, such as star-branched
hepta- and decaradicals and their dendritic homologues.[21For
the latter, magnetization data could not be fitted to a single
Brillouin function and the average values for S (between the low
and high field values) were well below the theoretical values
expected for high-spin systems without defects.
In summary, macrocyclic polyradical I with nonet (S = 4)
ground state was prepared. This compound may provide a first
step towards addressing the problem of defects in molecules
with large numbers of interacting radical sites.
Received. May 11. 1994 [Z 6921 IE]
German version. Angen. Chrm. 1994. 106, 2193
[ l ] a ) H. luamura. A d i . P h n Org. Clicm. 1990.26, 179; h) D. A. Dougherty. Acc.
Cllt'f7i. Rc.5. 1991. 24. 88; C) H. IWdmura, N. Koga. h i d . 1993, 26, 346: d ) W. T.
Borden. H . Iwamura. J. A. Berson. h i d . 1994. 27. 109: e) A. Rajca, Ch~r77.R e i ,
1994. Y4. 871.
~
~
Документ
Категория
Без категории
Просмотров
0
Размер файла
385 Кб
Теги
base, calixarenep, sextet, acyclic, pentaradical, macrocyclic, none, octaradical, analogues
1/--страниц
Пожаловаться на содержимое документа