Enhancing the stability of trinickel molecular wires and switches: Ni36+/Ni37+

Article abstract of DOI:10.1021/ic0341486

This paper describes in detail four new compounds that contain extended metal atom chains (EMACs) of three nickel atoms wrapped by either di(2-pyridyl)amide (dpa) or the new homologous ligand with an ethyl group at the para position of each pyridyl group, depa, and compares them to the precursor Ni₃(dpa)₄Cl₂ (1) and the oxidized and rather unstable Ni₃(dpa)₄(PF₆)₃ (2). The new molecules are Ni₃(depa)₄Cl₂ (3), Ni₃(depa)₄(PF₆)₃ (4), [Ni₃(dpa)₄(CH₃CN)₂] (PF₆)₂ (5), and [i₃(depa)₄(CH₃CN)₂] (PF₆)₂ (6). These compounds are fully described as to preparation, elemental composition, structure, infrared spectra, ¹H NMR spectra (where possible), electrochemistry, magnetic susceptibility, and an EPR spectrum for 4. The effects of (a) introducing the ethyl substituents on the ligands, (b) replacing axial anions by neutral axial ligands, and (c) oxidizing the Ni₃ chains are reported and discussed. The point of major interest is how oxidation profoundly alters the electronic structure of the EMAC.

Full text of DOI:10.1021/ic0341486

Inorg. Chem. 2003, 42, 3595−3601
Enhancing the Stability of Trinickel Molecular Wires and Switches:
6+
7+
Ni3 /Ni3
,†
,†
John F. Berry,† F. Albert Cotton,* Tongbu Lu,†,‡ Carlos A. Murillo,* and Xiaoping Wang†
Laboratory for Molecular Structure and Bonding, Department of Chemistry, P.O. Box 30012,
Texas A&M UniVersity, College Station, Texas 77842-3012, and School of Chemistry & Chemical
Engineering, Sun Yat-Sen UniVersity, Guangzhou 510275, P. R. China
Received February 12, 2003
This paper describes in detail four new compounds that contain extended metal atom chains (EMACs) of three
nickel atoms wrapped by either di(2-pyridyl)amide (dpa) or the new homologous ligand with an ethyl group at the
para position of each pyridyl group, depa, and compares them to the precursor Ni3(dpa)4Cl2 (1) and the oxidized
and rather unstable Ni3(dpa)4(PF ) (2). The new molecules are Ni3(depa)4Cl2 (3), Ni3(depa)4(PF ) (4), [Ni3(dpa)4(CH -
6 3
6 3
3
CN)2](PF ) (5), and [Ni3(depa)4(CH CN)2](PF ) (6). These compounds are fully described as to preparation, elemental
6 2
3
6 2
composition, structure, infrared spectra, 1H NMR spectra (where possible), electrochemistry, magnetic susceptibility,
and an EPR spectrum for 4. The effects of (a) introducing the ethyl substituents on the ligands, (b) replacing axial
anions by neutral axial ligands, and (c) oxidizing the Ni3 chains are reported and discussed. The point of major
interest is how oxidation profoundly alters the electronic structure of the EMAC.
1. Introduction
and in the ensuing years an immense growth has occurred,
primarily in this laboratory4 and that of Prof. S.-M. Peng in
Taipei.5
We have recently reported1 on the effects of introducing
ethyl substituents into the pentanickel compound Ni5(tpda)4-
Cl2, where tpda is the dianion a, to obtain Ni5(etpda)4Cl2,
where etpda is b. We have now completed a similar but more
extensive study of analogous trinickel compounds. While it
might seem more logical to publish the work on the trinickel
compounds first, the actual course of research did not allow
us to do that.
Quite apart from the historical importance of 1, it and its
derivatives and homologues continue to be of prime impor-
tance in a fundamental understanding of the entire class of
EMACs. For this reason we have continued to investigate
very actively 1 and its relatives that also contain nickel. In
1999 we published a definitive study of Ni3(dpa)4Cl2 itself.6
We have now proceeded well beyond 1 itself to determine
the influence of three factors: (1) oxidation to produce Ni3-
(dpa)4(PF6)3,7 2; (2) change of axial ligands; and (3) the effect
of substituents on the dpa ligand on the characteristics of
the Ni3 chain. In our report of the oxidation of 1 to produce
(3) Aduldecha, S.; Hathaway, B. J. Chem. Soc., Dalton Trans. 1991, 993.
(4) See for example: (a) Cotton, F. A.; Daniels, L. M.; Murillo, C. A.;
Pascual, I. J. Am. Chem. Soc. 1997, 119, 10223. (b) Cle´rac, R.; Cotton,
F. A.; Daniels, L. M.; Dunbar, K. R.; Kirschbaum, K.; Murillo, C.
A.; Pinkerton, A. A.; Schultz, A. J.; Wang, X. J. Am. Chem. Soc.
2000, 122, 6226. (c) Cotton, F. A.; Daniels, L. M.; Lu, T.; Murillo,
C. A.; Wang, X. J. Chem. Soc., Dalton Trans. 1999, 517.
(5) See for example: (a) Yang, E.-C.; Cheng, M.-C.; Tsai, M.-S.; Peng,
S.-M. J. Chem. Soc., Chem. Commun. 1994, 2377. (b) Sheu, J.-T.;
Lin, C.-C.; Chao, I.; Wang, C.-C.; Peng, S.-M. Chem. Commun. 1996,
315. (c) Zhu, L.-G.; Peng, S. M. Wuji Huaxue Xuebao 2002, 18, 117.
(6) Cle´rac, R.; Cotton, F. A.; Dunbar, K. R.; Murillo, C. A.; Pascual, I.;
Wang, X. Inorg. Chem. 1999, 38, 2655.
The entire field of extended metal atom chains (EMACs)
can be traced back to 1968 when purple Ni3(dpa)4Cl2, 1 (dpa
) anion of di(2-pyridyl)amine (shown as c)), was reported
but not correctly formulated.2 With the structure correctly
determined in 1991,3 the field can truly be said to have begun,
* To whom correspondence should be addressed: cotton@tamu.edu
(F.A.C.); murillo@tamu.edu (C.A.M.).
† Texas A&M University.
‡ Sun Yat-Sen University.
(1) Berry, J. F.; Cotton, F. A.; Lei, P.; Lu, T.; Murillo, C. A. Inorg. Chem.,
in press.
(2) Hurley, T. J.; Robinson, M. A. Inorg. Chem. 1968, 7, 33.
(7) (a) Berry, J. F.; Cotton, F. A.; Daniels, L. M.; Murillo, C. A. J. Am.
Chem. Soc. 2002, 124, 3212. (b) Berry, J. F.; Cotton, F. A.; Daniels,
L. M.; Murillo, C. A.; Wang, X. Inorg. Chem., 2003, 42, 2418.
10.1021/ic0341486 CCC: $25.00 © 2003 American Chemical Society
Published on Web 05/02/2003
Inorganic Chemistry, Vol. 42, No. 11, 2003 3595

Berry et al.
2-amino-4-ethylpyridine hydrochloride (12.9 g, 81.0 mmol) was
heated at 290 °C under a nitrogen atmosphere for 8 h. After cooling
to room temperature, 40 mL of water was added; the resulting
solution was made basic by addition of solid NaOH and extracted
with chloroform (3 × 50 mL). After the chloroform extracts were
dried over K2CO3, the chloroform was evaporated, and the residue
was distilled under vacuum (0.2 mm). This gave two fractions. The
first contained about 2 g of unreacted 2-amino-4-ethylpyridine. The
second fraction contained the product, which was distilled at 135-
140 °C as a yellow oil (7.0 g, 38%). After standing overnight, the
oil became a light yellow solid, which was purified by sublimation.
1H NMR (CDCl3, 300 MHz, δ): 8.12 (d, J ) 4.8 Hz, 2H), 7.82
(br s, 1H), 7.36 (s, 2H), 6.68-6.66 (dd, 2H), 2.66-2.58 (q, J )
7.8 Hz, 4H), 1.22 (t, J ) 7.2 Hz, 6H).
Ni3(depa)4Cl2, 3. A 100 mL round-bottom flask was charged
with anhydrous NiCl2 (0.26 g, 2.0 mmol), Hdepa (0.45 g, 2.0
mmol), and naphthalene (10 g). The mixture was heated at 185-
190 °C for 5 min, and then n-butanol (3 mL) was carefully added;
heating was continued while the system was purged with nitrogen
until the n-butanol completely evaporated. Then, a solution of
t-BuOK (0.23 g, 2.0 mmol) in 5 mL of n-butanol was added
dropwise. Heating was continued until the remaining n-butanol was
evaporated completely. After the mixture was cooled to about 80
°C, hexanes (3 × 60 mL) were used to remove naphthalene. The
remaining deep purple solid was extracted with toluene (2 × 10
mL) and this solution was then layered with hexanes. After 1 week,
a large crop of deep purple crystals of 3‚0.5C6H14 formed. Yield:
0.38 g, 66%. Anal. Calcd for C59Cl2H71N12Ni3 (Ni3(depa)4Cl2‚
0.5C6H14): C, 59.29; H, 5.99; N, 14.06. Found: C, 59.37; H, 5.98;
N, 13.91. IR (KBr, cm-1): 3448 (br, w), 3056 (w), 2965 (s), 2928
(m), 2869 (m), 1608 (s), 1536 (s), 1473 (s), 1414 (s), 1346 (s),
1290 (s), 1227 (m), 1179 (s), 1126 (w), 1061 (m), 1014 (s), 937
(s), 813 (s), 739 (m), 656 (w), 544 (m), 444 (s).
2 we discussed how, upon oxidation, the Ni-Ni separations
of about 2.43 Å decrease to 2.28 Å. Such a diminution by
0.15 Å is contrary to what would be expected for an increase
by one positive charge on proximate but nonbonded,
positively charged Ni ions. Thus it was concluded that
metal-metal bond formation is responsible for the large
contraction in Ni-Ni separations. It can be envisioned that
a molecular switch8 could be produced as a consequence of
a change from a non-interacting (presumably nonconducting)
chain of metal atoms to an interacting (presumably conduct-
ing) chain by an appropriate change in the applied potential.
We have now studied some of the factors that affect the
stability of these trinickel species, especially the oxidized
ones, and we have produced a considerably more thermally
stable switch. The importance of having products that are
easily handled at room temperature (and above) cannot be
overestimated. A full report on that work, which deals with
four new compounds, is presented here. The new compounds
are (3) Ni3(depa)4Cl2, (4) Ni3(depa)4(PF6)3, (5) [Ni3(dpa)4-
(CH3CN)2](PF6)2, and (6) [Ni3(depa)4(CH3CN)2](PF6)2 where
depa is the anion of 4,4′-diethyl-2,2′-dipyridylamine, the
diethyl-substituted ligand, d.
2. Experimental Section
Unless otherwise stated, reactions were carried out under nitrogen
with the use of standard Schlenk techniques. Anhydrous NiCl2 and
AgPF6 were purchased from Strem Chemicals, and 4-ethylpyridine
was purchased from Aldrich. All the solvents used were dried and
distilled under N2 by following standard procedures. Ni3(dpa)4Cl2
was prepared according to our previous report,7b and 2-amino-4-
ethylpyridine was prepared according to a published method.9
Physical Measurements. Elemental analyses were performed
by Canadian Microanalytical Services in British Columbia, Canada.
The IR spectra were recorded on a Perkin-Elmer 16PC FT-IR
spectrophotometer using KBr pellets. 1H NMR spectra were
obtained on a VXR-300 NMR spectrometer. Cyclic voltammograms
(CVs) were taken on a BAS 100 electrochemical analyzer with
Ni3(depa)4(PF6)3, 4. A Schlenk flask containing Ni3(depa)4Cl2
(0.120 g, 0.104 mmol) and AgPF6 (0.105 g, 0.416 mmol) was
charged with 15 mL of CH2Cl2 at -78 °C. The resulting solution
was stirred at -78 °C in the dark for 1 h to give a deep blue mixture,
which was filtered while cold through Celite to remove AgCl and
Ag. The filtrate was layered with hexanes and kept in a freezer.
After about 1 week, large blue crystals of 4‚3CH2Cl2 formed.
Yield: 0.14 g, 76%. Anal. Calcd for C57.5Cl3F18H67N12Ni3P3 (Ni3-
(depa)4(PF6)3‚1.5CH2Cl2): C, 42.02; H, 4.11; N, 10.23. Found: C,
41.98; H, 4.24; N, 10.45. IR (KBr, cm-1): 3435 (br, m), 2970 (w),
1616 (s), 1535 (w), 1476 (m), 1426 (vs), 1230 (w), 1181 (w), 1093
(br, m), 843 (vs), 734 (w), 558 (m), 472 (w).
[Ni3(dpa)4(MeCN)2](PF6)2, 5. A dark purple solution of Ni3(dpa)4-
Cl2 (317 mg, 0.342 mmol) in 15 mL of acetonitrile was stirred as
a solution of AgPF6 (173 mg, 0.684 mmol) in 15 mL of acetonitrile
was added dropwise. The mixture became more intensely purple,
and a white precipitate was observed. After stirring for several
hours, the mixture was filtered with the aid of Celite to give a clear,
dark purple solution. This solution was concentrated by means of
vacuum and layered with diethyl ether. Within 1 week, large block-
shaped purple crystals of 5‚3.90 MeCN grew along the sides of
the flask. Yield: 0.325 g, 77%. IR, (KBr, cm-1): 3448 (w, br),
3026 (w), 2965 (w), 2500 (vw), 2291 (w, CN), 1604 (s), 1595 (s),
1550 (m), 1469 (vs), 1460 (vs), 1425 (vs), 1358 (s), 1315 (m),
1283 (m), 1263 (w), 1243 (w), 1150 (m), 1111 (m), 1053 (w), 1017
(m), 894 (w), 839 (vs), 763 (s), 740 (m), 641 (w), 557 (m), 517
(w), 497 (w), 426 (m). Anal. Calcd for C40H34N12OP2F12Ni3: C,
41.25; H, 2.94; N, 14.43. Found: C, 41.54; H, 3.41; N, 14.88.
Ni3(depa)4(CH3CN)2(PF6)2, 6. A Schlenk flask containing
Ni3(depa)4Cl2 (0.20 g, 0.17 mmol) and AgPF6 (0.090 g, 0.36 mmol)
Bun NPF6 (0.1 M) electrolyte, Pt working and auxiliary electrodes,
4
a Ag/AgCl reference electrode, and a scan rate of 100 mV/s. The
magnetic susceptibility data were collected on a Quantum Design
SQUID magnetometer MPMS-XL from 2 to 300 K at a field of
1000 G on finely ground, vacuum-dried samples. The data were
corrected for the sample holder and for the diamagnetic contribu-
tions calculated from Pascal’s constants. The X-band EPR spectrum
of a THF glass of 4 was recorded at 6 K on a Bruker ESP 300
spectrometer.
2,2′-Di(4-ethyl)pyridylamine (Hdepa). This was made by a
modification of a literature reaction for a related compound.10
A
mixture of 2-amino-4-ethylpyridine (10.0 g, 81.0 mmol) and
(8) For a recent review on molecular electronics, see: Carroll, R. L.;
Gorman, C. B. Angew. Chem., Int. Ed. 2002, 41, 4378.
(9) Hansch, C.; Carpenter, W.; Todd, J. J. Org. Chem. 1958, 23, 1924.
(10) (a) Baxter, C. E.; Rodig, O. R.; Schlatzer, R. K.; Sinn, E. Inorg. Chem.
1979, 18, 1918. (b) Dyadyusha, G. G.; Verbovskaya, T. M.; Kiprianov,
A. I. Ukr. Khim. Zh. (Russ. Ed.) 1966, 32, 357.
3596 Inorganic Chemistry, Vol. 42, No. 11, 2003

The Stability of Trinickel Molecular Wires and Switches
Table 1. Crystal Data
3‚0.5hexane
4‚3CH2Cl2
5‚3.14CH3CN
6‚0.33H2O
formula
fw
cryst syst
space group
a, Å
b, Å
c, Å
C59H71Cl2N12Ni3
1195.31
tetragonal
P4hn2
19.0825(7)
19.0825(7)
15.771(1)
90
90
90
5743.0(5)
4
1.382
C59H70Cl6F18N12Ni3P3
1771.01
monoclinic
P21/n
13.920(1)
31.703(3)
16.727(1)
90
95.023(2)
90
7353(1)
4
1.600
0.0756, 0.1757
0.0993, 0.1884
C50.27H47.41F12N17.14Ni3P2
1357.70
triclinic
P1h
12.56(1)
12.802(9)
20.01(2)
107.26(2)
98.83(3)
101.54(5)
2932(4)
2
C60H70.67F12N14Ni3O0.33P2
1459.38
cubic
Pn3hn
21.5984(5)
21.5984(5)
21.5984(5)
90
R, deg
â, deg
90
90
γ, deg
V, Å3
10075.5(4)
6
1.443
0.0508, 0.1319
0.0729, 0.1639
Z
d(calc), g cm-3
R1,a wR2b [I > 2σ(I)]
R1,a wR2b (all data)
1.538
0.0908, 0.2365
0.1185, 0.2653
0.0323, 0.0888
0.0383, 0.0955
a R1 ) ∑||Fo| - |Fc||/∑|Fo|. b wR2 ) [∑[w(Fo - Fc )2]/∑[w(Fo )2]]1/2, w ) 1/σ2(Fo ) + (aP)2 + bP, where P ) [max(0 or Fo ) + 2(Fc )]/3.
2
2
2
2
2
2
was charged with 15 mL of acetonitrile. The resulting solution was
stirred at room temperature in the dark for 3 h to give a deep purple
mixture, which was filtered through Celite to remove AgCl. To
the filtrate was added 15 mL of ether. The clear solution was
allowed to evaporate slowly in air. After about 2 weeks, deep purple
crystals of 6‚0.33H2O formed from the solution. Yield: 0.12 g,
48%. IR (KBr, cm-1): 3427 (s, br), 2969 (s), 2368 (w), 2281 (w,
CN), 1612 (vs), 1535 (s), 1474 (s), 1417 (vs), 1345 (m), 1229 (w),
1181 (m), 1060 (w), 1016 (w), 938 (w), 844 (vs), 738 (w), 557
(m), 447 (m).
Crystallography. In each case, a suitable crystal was mounted
on the end of a quartz fiber with a small amount of stopcock grease
and was transferred to a goniometer head. X-ray diffraction data
for 3‚0.5C6H14, 4‚3CH2Cl2, and 6‚0.33H2O were collected on a
Bruker Smart 1000 CCD area detector system at 213 K. Data were
corrected for Lorentz and polarization effects using the program
SAINTPLUS.11 Absorption corrections were applied using SAD-
ABS.12 These three structures were solved by direct methods.
For 5‚3.14MeCN X-ray data were gathered on a Nonius FAST
diffractometer at 213 K using the software program MADNES.13
Reflections for indexing were found via a routine which collects a
series of images from several different sections of reciprocal space.
The reflections showed signs of high mosaicity by the broadening
in ω. A triclinic unit cell was refined using 247 strong reflections
with 18.1° < 2θ < 41.8°. The cell dimensions were confirmed by
examination of axial images. The program PROCOR14 was used
to process the integrated data into SHELX format. The data were
corrected for Lorentz and polarization effects, and absorption
corrections were applied using the program SORTAV.15 The
structure was solved using the Patterson method.
methane in 3‚0.5C6H14 and 4‚3CH2Cl2, and PF6 anions in 5‚
3.90MeCN and 6‚0.33H2O were found to be disordered. All
refinement cycles were performed using the SHELXL-97 program.16
The crystallographic data are listed in Table 1.
3. Results and Discussion
Preparative Chemistry. The precursor of ligand d,
Hdepa, was prepared in moderate yield from commercially
available starting materials. Ni3(depa)4Cl2 (3) was prepared
according to eq 1 using naphthalene as solvent, which allows
the reaction temperature to be raised to 185-190 °C.
Conversion of 3 to Ni3(depa)4(PF6)3 (4) was accomplished
by reaction with an excess of AgPF6 at -78 °C as shown in
eq 2. By treating 3 with exactly 2 equiv of AgPF6 in
acetonitrile at 20-25 °C [Ni3(depa)4(CH3CN)2](PF6)2 (6) was
obtained as shown in eq 3. The dpa analogue of 6, namely,
[Ni3(dpa)4(CH3CN)2](PF6)2 (5), was obtained in an analogous
way from Ni3(dpa)4Cl2 (1). Reaction 3 succeeds, in part,
because AgPF6 is a much milder oxidant in acetonitrile than
it is in CH2Cl2.17 This is no doubt attributable to the ability
of CH3CN to stabilize the Ag+ ion.
BuOH, C10H8
3NiCl2 + 4Hdepa + 4 t-BuOK
8
Ni3(depa)4Cl2 + 4 t-BuOH + 4KCl (1)
CH2Cl2
Ni3(depa)4Cl2 + 3AgPF6
8
In all structures, hydrogen atoms were placed at calculated
positions based on a riding model. Some of the ethyl groups in
3‚0.5C6H14, 4‚3CH2Cl2, and 6‚0.33H2O, hexane and dichloro-
Ni3(depa)4(PF6)3 + Ag + 2AgCl (2)
Ni3(depa)4Cl2 + 2AgPF6 + 2CH3CN CH CN8
3
(11) SAINTPLUS, V 6.28A Software for the CCD Detector System; Bruker
Analytical X-ray System, Inc.: Madison, WI, 2001.
(12) SADABS. Program for absorption correction using SMART CCD data
based on the method of Blessing: Blessing, R. H. Acta Crystallogr.
1995, A51, 33.
(13) Pflugrath, J.; Messerschmidt, A. MADNES, Munich Area Detector
(New EEC) System, Version EEC 11/1/89, with enhancements by
Enraf-Nonius Corp., Delft, The Netherlands. A description of MADNES
appears in the following: Messerschmitt, A.; Pflugrath, J. J. Appl.
Crystallogr. 1987, 20, 306.
[Ni3(depa)4(CH3CN)2](PF6)2 + 2AgCl (3)
Structural Results. Crystal structures of 3-6 were solved
and refined satisfactorily. Thermal ellipsoid drawings of the
structures are shown in Figures 1-4, and some important
dimensions are listed for compounds 1-6 in Table 2. More
detailed comparison of these structures will be made below,
(14) (a) Kabsch, W. J. Appl. Crystallogr. 1988, 21, 67. (b) Kabsch, W. J.
Appl. Crystallogr. 1988, 21, 916.
(15) Program for absorption correction for Enraf-Nonius FAST diffracto-
meter using the method of Blessing: Blessing, R. H. Acta Crystallogr.
1995, A51, 33.
(16) Sheldrick, G. M. SHELXTL97; University of Go¨ttingen: Go¨ttingen,
Germany, 1997.
(17) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877.
Inorganic Chemistry, Vol. 42, No. 11, 2003 3597

Berry et al.
Figure 3. Thermal ellipsoid plot of the dication [Ni3(dpa)4(NCMe)2]2+
from 5‚3.14CH3CN. Ellipsoids are drawn at the 50% probability level;
pyridyl hydrogen atoms have been removed for clarity.
Figure 1. Thermal ellipsoid plot of 3 with ellipsoids drawn at the 50%
probability level. Only one orientation of the disordered ethyl groups is
shown; hydrogen atoms and solvent molecules have been omitted for clarity.
Figure 4. Thermal ellipsoid plot of 6 with ellipsoids drawn at the 40%
probability level. Only the cation is shown; hydrogen atoms and solvent
molecules have been omitted for clarity.
Figure 2. Thermal ellipsoid plot of the cation [Ni3(depa)4(PF6)2]+ from
4‚3CH2Cl2. Ellipsoids are drawn at the 30% probability level. Only one
orientation of the disordered ethyl groups is shown; hydrogen atoms have
been omitted for clarity.
rings. The resonances for the ethyl groups are found as a
peak at 1.301 ppm for the methyl groups and as two broad
peaks at -0.430 and -0.816 ppm for the methylene protons.
The reason for the splitting of the latter is not obvious.
The 1H NMR spectrum for [Ni3(dpa)4(CH3CN)2](PF6)2 (5)
is shown in Figure 5. While the chemical shifts are
somewhat, but not greatly, different from those for 1, the
general similarity, especially for the line widths, is notable.
There are also signals at 1.940 and 2.124 ppm, which we
assign to CH3CN and H2O (present in the solvent), respec-
tively. Thus, it appears that the coordinated CH3CN mol-
ecules are exchanging with those in the solvent.
but it may be observed here that in every case the Ni-N
bonds for the center nickel atom are in the range 1.87-1.89
Å.
Nuclear Magnetic Resonance Spectra. Often, NMR
spectra of paramagnetic compounds are difficult to detect
or unobservable because of the short relaxation times
exhibited by such compounds. Occasionally, a spectrum can
be obtained, and sometimes interpretation is possible. We
previously6 reported the 1H NMR spectrum of Ni3(dpa)4Cl2
(1) in CD2Cl2 which showed four signals displaying large
shifts and line-broadening at 48.79, 32.32, 15.69, and 9.08
ppm. This was consistent with a symmetrical molecular
structure. The spectrum of Ni3(depa)4Cl2 (3) recorded at room
temperature in CDCl3 is very similar except that the peak
for 1 at 9.08 ppm is missing (see Table 3). This clearly
identifies that peak as being due to the H(4) of the pyridyl
For Ni3(depa)4(PF6)3 (4) in CD3CN, which has a com-
pletely different electronic structure from 3 and 5, only three
signals were observed: peaks at 2.002 and 1.585 ppm, in
an approximate intensity ratio of 3:2, can be assigned to the
CH3 and CH2 components of the ethyl groups; there is also
a very broad signal (approximately 0.4 ppm width at half-
height) centered at -3.44 ppm. This might be due to some
3598 Inorganic Chemistry, Vol. 42, No. 11, 2003

The Stability of Trinickel Molecular Wires and Switches
Table 2. Selected Internuclear Distances (Å) for Compounds 1-6a
1b,c
2c
3‚0.5C6H14
4‚3CH2Cl2
5‚3.14CH3CN
6‚0.33H2O
Ni-Ni
2.430[9]
2.10[8]
1.89[6]
2.33[4]
2.284[1]
1.927[5]
1.881[5]
2.4325(3)
2.094[2]
1.893[1]
2.336(1)
2.293[4]
1.930[6]
1.871[6]
2.374[2]
2.079[7]
1.882[8]
2.415(1)
2.071(3)
1.891(5)
Ni-N1 outer
Ni-N1 inner
Ni-Cl
Ni‚‚‚F
2.424, 2.429
2.458, 2.439
Ni-NCCH3
2.049[7]
50.3
2.037(7)
45.4
mean ød
51.7
46.1
49.8
46.5
a Esds in parentheses when there is only one crystallographically distinct value. Esds in square brackets when several chemically equivalent but
crystallographically distinct values have been averaged. b Averages from three independent structures discussed in ref 6. c Data from ref 7b. d ø is the average
of the four torsion angles.
1
Table 3. Chemical Shifts (δ, ppm) from H NMR Spectra
a
1 in CD2Cl2
3 in CDCl3
48.79
48.20
42.50
32.32
30.50
27.63
15.69
14.97
15.25
9.080
1.301
-0.430
-0.816
8.800
{
5 in CD3CN
a From ref 6.
of the pyridyl protons, but the absence of other expected
signals make an assignment difficult. Thus, the difference
in the electronic structure of the oxidized species is also
evident in the NMR spectrum. The numerical data for 1, 3,
and 5 are presented in Table 3.
Electrochemistry. A solution of compound 3, Ni3(depa)4-
Cl2, in CH2Cl2 has a CV in the range 0-1.2 V that consists
of a single reversible wave at an E1/2 of 0.778 V, which may
be compared with the behavior of 1, Ni3(dpa)4Cl2, in the same
solvent, which has a wave at 0.908 V. Thus, the effect of
the electron-donating ethyl groups is significant; they make
the oxidation easier by 0.130 V. A similar effect has already
been seen for the two Ni5 species, Ni5(tpda)4Cl2 and
Ni5(etpda)4Cl2.1
Figure 5. 1H NMR spectrum of 5 taken at 300 MHz in CD3CN at 23 °C.
This species is predicted to have a three-center four-electron
σ bond along the trinickel chain.7b
The effect of replacing the axial chloride ligands in
Ni3(dpa)4Cl2 (1) by acetonitrile molecules in [Ni3(dpa)4(CH3-
Ni3(dpa)4Cl2 ) [Ni3(dpa)4Cl2]+ + e-
(4)
2+
CN)2]2+ is very striking. There is a shift of the Ni3(dpa)4
/
[Ni3(dpa)4(NCMe)2]2+ ) [Ni3(dpa)4(NCMe)2]3+ + e- (5)
[Ni3(dpa)4(NCMe)2]3+ ) [Ni3(dpa)4(NCMe)2]4+ + e- (6)
3+
Ni3(dpa)4 couple to 1.205 V. While the direction is as
expected, the magnitude of this shift, 0.297 V, is large
enough to tempt one to think that other properties of the
2+
Ni3(dpa)4 core might also be seriously altered, but, as
Magnetic Susceptibility and EPR. Although the magnetic
properties of 1 have been reported,6,18 data were remeasured
for a standardized comparison with the new Ni3 com-
pounds. These data are shown in Figure 6 and are consistent
with two antiferromagnetically coupled high-spin Ni2+ ions
in the terminal positions of the Ni3 chains and a diamagnetic
square planar Ni2+ species in the center. Therefore, eq 7 can
be used to fit the data for 1, 3, 5, and 6:
shown later in this paper, they are not. In CH2Cl2 the
oxidation of 1 is probably represented reasonably well by
eq 4, while for the oxidation of 5 in CH3CN, eq 5 is
applicable. Removing an electron from a +2 ion to make
the corresponding +3 ion as in eq 5 is bound to be more
difficult than removing an electron from a neutral molecule
to make the corresponding +1 ion in eq 4, even though the
solvent for eq 5, CH3CN, has a higher dielectric constant,
namely, 38, than that for eq 4, CH2Cl2, where it is 9. In
other words, the large shift is of mainly electrostatic origin
and does not necessarily imply that the internal electronic
structures (and hence bond lengths, bond angles, or magnetic
properties) are very different. There is, however, a second
redox wave for 5 centered at E1/2 ) 1.523 V. This we assign
to the process given in eq 6, which produces a [Ni3(dpa)4]4+
ion. It should be noted that no Nin(2n+2)+ species are known
within the framework of four bridging ligands, and attempts
6+
6Ng2â2
3kT
exp(J/kT) + 5 exp(3J/kT)
1 + 3 exp(J/kT) + 5 exp(3J/kT)
ø )
(7)
This equation is based on the isotropic interaction between
two S ) 1 centers with the following Hamiltonian: H )
-JS1‚S3 - gâS‚H. Here, N is Avogradro’s number, g is the
(18) Note that there is a typographical error in ref 6. Since that work was
done using the Hamiltonian H ) -2JS1‚S3, the value of J must
therefore be half the value obtained here (using H ) -JS1‚S3). Thus,
the expression “2J ) -108 cm-1” should read “J ) -108 cm-1.”
8+
are being made to isolate the Ni3 unit from this system.
Inorganic Chemistry, Vol. 42, No. 11, 2003 3599

Berry et al.
show axial symmetry with g| > g and hyperfine splitting.
This is likely due to the interaction of the unpaired electron
with the two I ) /2 19F nuclei in the axial positions of the
1
molecules. It is interesting that this coupling is observed,
despite the long Ni‚‚‚F distances of >2.4 Å in the solid state.
Furthermore, the existence of this splitting supports the
7+
electron delocalization along the Ni3 unit as the unpaired
electron sees equally well each of the terminal PF6 anions.
4. Additional Remarks
Comparison of 1 and 3. The introduction of the two
4-ethyl substituents, that is, changing from Ni3(dpa)4Cl2 to
Ni3(depa)4Cl2, has essentially no structural effect. Both
molecules have essentially symmetrical structures with D4
symmetry in the Ni3N12Cl2 cores. To within 0.01 Å the
corresponding bond lengths are the same, including the
Ni-N distances despite ca. 0.8 difference in the pK values
of the pyridyl and 4-ethylpyridyl groups.21 In five separate
crystal structures of 1 the overall torsion angle ranged from
49.6° to 53.2°, while in 3 the torsion angle is in the range
49.1-51.2°.
Figure 6. Magnetic susceptibility data for 1 and 3-6. The solid lines
represent the theoretical fits of the data (see text).
Table 4. Magnetic Susceptibility Results
formula
Ni3(dpa)4Cl2
Ni3(depa)4Cl2
[Ni3(dpa)4(CH3CN)2](PF6)2
[Ni3(depa)4(CH3CN)2](PF6)2
Ja
g
Rb
1
3
5
6
-218.2(7)
-217.5(7)
-267.6(5)
-236.6(3)
2.067(5)
2.245(7)
2.132(3)
2.197(2)
0.99998
0.99997
0.99997
0.99998
a J (in cm-1) is defined by the term -JS1‚S3 for the spin-spin coupling.
Electrochemically, there is a significant and helpful
difference. For both 1 and 3 there is a reversible, one-electron
oxidation, but for 3 oxidation is more than 100 mV easier
b R is the goodness of fit.
Lande´ factor, â is the Bohr magneton, k is Boltzmann’s
constant, J is the exchange parameter, and T is temperature.
Also included were parameters accounting for paramagnetic
S ) 1 impurities (corresponding to less than 1% of the
sample in every case) and TIP (in all cases <300 × 10-6
emu/mol).
(E1/2 ) +0.778 V vs Ag/AgCl in CH2Cl2) than in 1 (E1/2
)
0.908 V vs Ag/AgCl in CH2Cl2), in keeping with the greater
basicity of the ethyl-substituted ligand.
Oxidations To Produce 2 and 4. Reaction of purple 1 or
3 with 3 equiv of AgPF6 in CH2Cl2 gives the corresponding
7
blue compounds Ni3(dpa)4(PF6)3 or Ni3(depa)4(PF6)3 in
The g values, shown in Table 4, are similar to each other
and to those of typical paramagnetic nickel compounds.19
Because the room-temperature øT values of 1, 3, 5, and 6
are not the same, different values of J for each compound
are to be expected. Indeed, the values of J for the complexes
with axial acetonitrile ligands are more negative than those
of the corresponding chloro complexes (see Table 4),
indicative of stronger antiferromagnetic coupling in the
former. Since the Ni‚‚‚Ni distances in the acetonitrile
complexes are shorter by 0.06 Å, it is possible that an
important contribution to the exchange pathway between the
terminal Ni2+ ions is through direct Ni‚‚‚Ni interaction. This
has also been seen in the Cu3(dpa)4X2 system, where the
Cu‚‚‚Cu distances of 2.47 and 2.40 Å for Cu3(dpa)4Cl2 and
Cu3(dpa)4(BF4)2, respectively, led to stronger antiferromag-
netic coupling in the latter.20
In contrast, there is no temperature dependence of øT for
the oxidized compound 4 (see Figure 6). The øT value of
0.406 emu‚K/mol corresponds to one unpaired electron with
g ) 2.08.
The EPR spectrum of 4 shows an axial signal with g )
2.165 and g| ) 2.117. The signal for g| was split into a triplet
with A| ) 150 G. This is similar to that of 2.7b Both are
consistent with the presence of one unpaired electron and
isolated yields of better than 75%. Structurally 2 and 4 are
very similar. In both compounds there is a highly significant
decrease in the Ni-Ni separations upon oxidation. For 2
there are two crystallographically independent Ni-Ni sepa-
rations of average value 2.284[1] Å which are ca. 0.15 Å
shorter than those of 1.7 For 4, there are also two crystal-
lographically independent but chemically insignificant sepa-
rations of 2.296(1) and 2.289(1) Å. These are ca. 0.14 Å
shorter than those in the precursor 3. In both precursors there
are Ni-Cl distances of ∼2.33 Å, but in 2 and 4 the axially
situated PF6 anions are relatively far away (Ni‚‚‚F separations
of over 2.4 Å). There is also a significant shortening of the
Ni-N distances for 2 and 4 as compared to those in the
precursors, clearly attributable to the increase in overall
positive charge which would be expected to increase attrac-
tion for the nitrogen donor atoms. The greater effect is
observed on the outer Ni-N distances. For 2, the central
Ni-N distances are 1.881[5] Å and the outer ones are 1.927-
[5] Å.7 These are significantly shorter than the corresponding
distances (1.89 and ∼2.10 Å, respectively) in 1.7b A similar
situation is found in the 3/4 couple where the outer Ni-N
distances are in the ranges 2.086(2)-2.101(2) and 1.918-
(6)-1.940(6) Å.
It is clear that the large shortening in Ni-Ni separation
upon oxidation would not be expected for nonbonded metal
(19) Drago, R. S. Physical Methods in Chemistry; W. B. Saunders
Company: Philadelphia, 1977; p 486.
(20) Berry, J. F.; Cotton, F. A.; Lei, P.; Murillo, C. A. Inorg. Chem. 2003,
42, 377.
(21) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum
Press: New York, 1982.
3600 Inorganic Chemistry, Vol. 42, No. 11, 2003

The Stability of Trinickel Molecular Wires and Switches
atoms in 1 and 3 if they remained nonbonded. An increase
from -216 to -268 cm-1 in the magnitude of the coupling
constant (in the term -JS1‚S3), on going from 1 to 5. This
probably indicates that at least some (perhaps all) of the
coupling is transmitted through the central nickel atom since
it accords well with the decrease in Ni‚‚‚Ni distances whereas
the entire indirect path from Ni(1) to Ni(3) via the Ni-N-
C-N-C-N-Ni chains does not appear to be significantly
changed as has been demonstrated for Cu3 compounds.21
Similar results are found in the cases of 3 and 6, although
the difference in J is not as pronounced.
6+
7+
in the total chain charge from Ni3 to Ni3 would be
expected to increase repulsion between the positively charged
nickel ions. Such an increase has been shown by oxidizing
Cu3(dpa)4Cl2 to Cu3(dpa)4Cl2+ where the Cu‚‚‚Cu separation
increases by 0.04 Å from 2.47 to 2.51 Å.7b Neither one of
these compounds has metal-metal bonds. Therefore, the
shortening by ca. 0.15 Å of the metal-metal distance in the
trinickel compounds must be attributed to metal-metal bond
formation. This is consistent with transformation to a
delocalized system that is likely to be a conductor. We
believe that these molecules could therefore serve as on-
off switches, which could be controlled by an applied
oxidation potential. They could also be expected to serve as
physical switches as the length of the Ni3 unit decreases by
0.30 Å upon oxidation.
An especially interesting observation is the great increase
in thermal stability of the ethyl-substituted 4 versus that of
the unsubstituted 2. The latter is stable as a solid and in
solution but only at low temperatures. However, 4 appears
to be stable indefinitely at room temperature. This improved
stability is presumably associated with the significant
decrease in the oxidation potential of the ethyl-substituted
precursor, as mentioned above.
5. Concluding Remarks
The effect of introducing the ethyl substituents to 1, 2,
and 5 so as to obtain 3, 4, and 6 is 2-fold. First, as expected
and intended, it increases the solubilities. We are thus
reporting the first examples of what will be an extensive,
planned program of making EMACs of many lengths and
with many metals that will have better solubilities than do
those of dpa and its longer homologues, loosely called
polypyridyl amides.
The second important effect is how the introduction of
alkyl substituents changes the redox chemistry. The electron-
donating influence of the ethyl groups makes oxidized
species significantly easier to obtain and improves their
stability. We have an excellent illustration of this not only
in comparing the redox potentials of 1 and 3 but in a very
practical sense by the greater thermal stability of 4 compared
to 2. There was an expectation of such an effect from the
fact that 4-ethylpyridine is considerably more basic than
pyridine by about 0.8 pK units. It has been a pleasant surprise
to see how well this has translated into the measured and
observed properties of the trinickel EMACs.
In parallel work1 we have shown that this effect is
observable in longer EMACs and the corresponding five-
metal chains are more readily oxidized than the unsubstituted
ones and thermally stable also. Another important advantage
of the ethyl-substituted compounds is a significant increase
in solubility relative to the corresponding unsubstituted
compounds. This factor should be of great value as the length
of the EMAC increases.
Comparisons of 1 and 5 and of 3 and 6. The replacement
of axial Cl- ligands by axial CH3CN ligands has several
major effects on the Ni3N12 core. There is considerable
shortening of the Ni‚‚‚Ni distances, from values in the range
of 2.417(1)-2.443[1] Å in 1 to 2.374[2] Å in 5. At the same
time, the corresponding (i.e., inner and outer) equatorial
Ni-N distance scarcely change and the difference between
them remains 0.20 Å. Even though the Ni‚‚‚Ni distances
shorten by about 0.06 Å, it would be expected that, as in 1,
there would be no Ni-Ni bonding in 5, and this is borne
out by the magnetic data. Both 1 and 5 are to be viewed as
having inner S ) 0 Ni(II) atoms and outer Ni(II) atoms with
S ) 1, with magnetic spin coupling between the two outer
magnetic moments. However, there is a significant increase,
Acknowledgment. This work was supported through a
Nanoscale Science and Engineering/NIRT Grant (DMR-
0103455) and the Telecommunications and Information Task
Force at Texas A&M University. J.F.B. thanks the National
Science Foundation for support in the form of a predoctoral
fellowship. We also thank Dr. A. Murali and Mr. C. Fewox
for assistance in the EPR measurements and Peng Lei for
assistance in obtaining NMR spectra.
Supporting Information Available: X-ray crystallographic data
in CIF format for 3‚0.5hexane, 4‚3CH2Cl2, 5‚3.14CH3CN, and
6‚0.33H2O and EPR spectrum for 4 as a PDF file. This material is
available free of charge via the Internet at http://pubs.acs.org.
IC0341486
Inorganic Chemistry, Vol. 42, No. 11, 2003 3601