Adduct formation between alkali metal ions and divalent metal salicylaldimine complexes having methoxy substituents. A structural investigation

Article abstract of DOI:10.1021/ic990496p

Sodium, potassium, and cesium salts (iodides, nitrates, acetates, and tetraphenylborates) form 1/1, 1/2 and 2/3 adducts with ML(n) [M = Co, Ni, Cu, and Zn; n = 1-4; H₂L₁ = N,N'-(3-methoxysalicylidene)ethane-1,2-diamine; H₂L₂, H₂L₃, and H₂L₄ are the -propane-1,2-diamine, -o-phenylenediamine, and -propane-1,3-diamine analogues of H₂L₁). Metal salicylaldimine, alkali metal, and anion all exert influence on stoichiometry and reactivity. Sodium ions tend to reside within the planes of the salicylaldimine oxygens, as in Na(NO₃)(MeOH)·NiL₄ (1), Na(NO₃)-(MeOH)·CuL₁ (2; both with unusual seven-coordinated sodium), and Na·(NiL₄)₂I·EtOH·H2O (3; with dodecahedral sodium coordination geometry). Potassium and cesium tend to locate between salicylaldimine ligands as in KI· NiL₄ (4) and [Cs(NO₃)·NiL₄]₃·MeOH (5; structures with infinite sandwich assemblies), CsI·(NiL₂)₂·H₂O (6), CsI₃·(NiL₄)₂ (7; simple sandwich structures), and [K(MeCN)]₂·(NiL₄)₃ (8; a triple-decker sandwich structure). Crystal data for I are the following: triclinic, P1, a = 7.3554(6) A, b = 11.2778(10) A, c = 13.562(2) A, α = 96.364(10)°, β = 101.924(9)°, γ = 96.809(10)°, Z = 2. For 2, triclinic, P1, a = 7.2247(7) A, b = 11.0427(6) A, c = 13.5610(12) A, α = 94.804(5)°, β = 98.669(7)°, γ = 99.26(6)°, Z = 2. For 3, orthorhombic, Pbca, a = 14.4648(19) A, b = 20.968(3) A, c = 28.404(3) A, Z = 8. For 4, triclinic, P1, a = 12.4904(17) A, b = 13.9363-(13) A, c = 14.1060(12) A α = 61.033(7)°, β= 89.567(9)°, γ = 71.579(10)°, Z = 2. For 5, monoclinic. P2₁/n, a = 12.5910(2) A, b = 23.4880(2) A, c = 22.6660(2) A, β = 99.3500(1)°, Z = 4. For 6, orthorhombic, Pbca, a = 15.752(3) A, b = 23.276(8) A, c = 25.206(6) A Z = 8. For 7, triclinic, P1, a = 9.6809(11) A, b = 10.0015-(13) A, c = 11.2686(13) A, α = 101.03°, β = 90.97°, γ = 100.55°, Z = 2. For 8, monoclinic, C2/c, a = 29.573(5) A b = 18.047(3) A, c = 23.184(3) A, β = 122.860(10)°, Z = 8.

Full text of DOI:10.1021/ic990496p

Inorg. Chem. 2000, 39, 1639-1649
1639
Adduct Formation between Alkali Metal Ions and Divalent Metal Salicylaldimine
Complexes Having Methoxy Substituents. A Structural Investigation
D. Cunningham,* P. McArdle, M. Mitchell, N. N´ı Chonchubhair, and M. O’Gara
Department of Chemistry, National University of Ireland, Galway, Ireland
Federico Franceschi and Carlo Floriani
Institut de Chimie Mine´rale et Analytique, Universite´ de Lausanne, CH-1015 Lausanne, Switzerland
ReceiVed May 7, 1999
Sodium, potassium, and cesium salts (iodides, nitrates, acetates, and tetraphenylborates) form 1/1, 1/2 and 2/3
adducts with MLn [M ) Co, Ni, Cu, and Zn; n ) 1-4; H2L1 ) N,N′-(3-methoxysalicylidene)ethane-1,2-diamine;
H2L2, H2L3, and H2L4 are the -propane-1,2-diamine, -o-phenylenediamine, and -propane-1,3-diamine analogues
of H2L1). Metal salicylaldimine, alkali metal, and anion all exert influence on stoichiometry and reactivity. Sodium
ions tend to reside within the planes of the salicylaldimine oxygens, as in Na(NO3)(MeOH)‚NiL4 (1), Na(NO3)-
(MeOH)‚CuL1 (2; both with unusual seven-coordinated sodium), and Na‚(NiL4)2I‚EtOH‚H2O (3; with dodecahedral
sodium coordination geometry). Potassium and cesium tend to locate between salicylaldimine ligands as in KI‚
NiL4 (4) and [Cs(NO3)‚NiL4]3‚MeOH (5; structures with infinite sandwich assemblies), CsI‚(NiL2)2‚H2O (6),
CsI3‚(NiL4)2 (7; simple sandwich structures), and [K(MeCN)]2‚(NiL4)3 (8; a triple-decker sandwich structure).
Crystal data for 1 are the following: triclinic, P1, a ) 7.3554(6) Å, b ) 11.2778(10) Å, c ) 13.562(2) Å, R )
96.364(10)°, â ) 101.924(9)°, γ ) 96.809(10)°, Z ) 2. For 2, triclinic, P1, a ) 7.2247(7) Å, b ) 11.0427(6) Å,
c ) 13.5610(12) Å, R ) 94.804(5)°, â ) 98.669(7)°, γ ) 99.26(6)°, Z ) 2. For 3, orthorhombic, Pbca, a )
14.4648(19) Å, b ) 20.968(3) Å, c ) 28.404(3) Å, Z ) 8. For 4, triclinic, P1, a ) 12.4904(17) Å, b ) 13.9363-
(13) Å, c ) 14.1060(12) Å, R ) 61.033(7)°, â ) 89.567(9)°, γ ) 71.579(10)°, Z ) 2. For 5, monoclinic. P21/n,
a ) 12.5910(2) Å, b ) 23.4880(2) Å, c ) 22.6660(2) Å, â ) 99.3500(1)°, Z ) 4. For 6, orthorhombic, Pbca,
a ) 15.752(3) Å, b ) 23.276(8) Å, c ) 25.206(6) Å, Z ) 8. For 7, triclinic, P1, a ) 9.6809(11) Å, b ) 10.0015-
(13) Å, c ) 11.2686(13) Å, R ) 101.03°, â ) 90.97°, γ ) 100.55°, Z ) 2. For 8, monoclinic, C2/c, a )
29.573(5) Å, b ) 18.047(3) Å, c ) 23.184(3) Å, â ) 122.860(10)°, Z ) 8.
Introduction
has been indicated as the key intermediate compound in
electron-transfer reactions between two metal salicylaldimine
molecules.5b
Over these past 2 decades there has been enormous interest
in the coordination chemistry of alkali metal ions, much of this
interest originating from a desire to develop molecular systems
that can mimic naturally occurring molecules responsible for
the selective transport of ions. Structure/selectivity studies have
focused strongly on crown ethers, cryptands, and related ligands.
Although these are particularly good classes of complexing
agents for alkali metal ions, there are many other simpler and
certainly more affordable ligands that are also extremely effec-
tive in this respect. Metal salicylaldimines represent a fascinating
group of ligands that are not only effective complexing agents
for p- and d-block elements but also for alkali metal ions.
Furthermore, some of the bimetallic systems resulting from
alkali metal complexation have particularly important properties.
The structures of a number of lithium, sodium, and potassium
adducts of metal salicylaldimines have been determined,1 as well
as structures of sodium salts of the free ligands.2 In some
instances these adducts are precursors for other interesting
molecular species concerning small molecule activation,3 elec-
tron storage,4-7 and the carrying of polar organometallics8 or
have remarkable magnetic properties (the alkali cation being
crucial in determining the three-dimensional network in the solid
state9). In fact, an analogous adduct of Na‚(NiL4)2I (see Figure
1 for nomenclature), which is described in the present paper,
When the metal salicylaldimine complexes have methoxy
groups in the 3,3′ positions (see, for example, Figure 1), they
are potentially tetradentate donor ligands. Under such circum-
(1) (a) Floriani, C.; Calderazzo, F.; Randaccio, L. J. Chem. Soc., Chem.
Commun. 1973, 384-385. (b) Milburn, H.; Truter, M. R.; Vickery,
B. J. Chem. Soc., Dalton Trans. 1974, 841-846. (c) Bresciani-Pahor,
N.; Calligaris, M.; Delise, P.; Nardin, G.; Randaccio, L.; Zotti, E.;
Fachinetti, G.; Floriani, C. J. Chem. Soc., Dalton Trans. 1976, 2310-
2316. (d) Pasquali, M.; Marchetti, F.; Floriani, C.; Cesari, M. Inorg.
Chem. 1980, 19, 1198-1202. (e) Gambarotta, S.; Fiallo, M. L.;
Floriani, C.; Chiesi-Villa, A.; Guastini, C. J. Chem. Soc., Chem.
Commun. 1982, 503-505. (f) Gambarotta, S.; Arena, F.; Floriani, C.;
Gaetani-Manfredotti, A. J. Chem. Soc., Chem. Commun. 1982, 835-
837. (g) Floriani, C.; Fiallo, M. L.; Chiesi-Villa, A.; Guastini, C. J.
Chem. Soc., Dalton Trans. 1987, 1367-1376. (h) van Veggel, F. C.
J. M.; Harkema, S.; Bos, M.; Verboom, W.; van Staveren, C. J.;
Gerritsma, G. J.; Reinhoudt, D. N. Inorg. Chem. 1989, 28, 1133-
1148. (i) Pessoa, J. C.; Silva, J. A. L.; Vieria, A. L.; Vilas-Boas, L.;
O’Brien, P.; Thornton, P. J. Chem. Soc., Dalton Trans. 1992, 1745-
1749. (j) Ball, S. C.; Cragg-Hine, I.; Davidson, M. G.; Davies, R. P.;
Lopez-Solera, M. I.; Raithby, P. R.; Reed, D.; Snaith, R.; Vogl, E.
M. J. Chem. Soc., Chem. Commun. 1995, 214-2149. (k) Atwood, D.
A.; Rutherford, D. Inorg. Chem. 1995, 34, 4008-4010.
(2) Solari, E.; De Angelis, S.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C.
J. Chem. Soc., Dalton Trans. 1991, 2471-2476.
(3) (a) Fachinetti, G.; Floriani, C.; Zanazzi, P. F. J. Am. Chem. Soc. 1978,
100, 7405-7407. (b) Gambarotta, S.; Arena, F.; Floriani, C.; Zanazzi,
P. F. J. Am. Chem. Soc. 1982, 104, 5082-5092.
10.1021/ic990496p CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/29/2000

1640 Inorganic Chemistry, Vol. 39, No. 8, 2000
Cunningham et al.
of the metal salicylaldimines of Figure 1 and reactions of NaBPh4 with
ML4 were carried out using a common procedure as follows. Ap-
proximately 0.8 nmol of the solid sodium salt was added to an equimolar
quantity of the metal salicylaldimine complex in 30 cm3 methanol
(generally, the metal salicylaldimine was incompletely dissolved) at
room temperature. In most instances, solutions were obtained and
precipitation from these solutions were generally obtained only after a
number of hours. After 48 h solids were isolated by filtration and dried
under vacuum. Exceptions were as follows: [Na(NO3)‚NiL3]‚MeOH‚
H2O, [NaI(NiL3)2]‚MeOH‚H2O, Na(ML4)2]BPh4 (M ) Ni, Cu, and Co),
[Na(ZnL4)2]BPh4‚MeOH, and [NaI(NiL2)2]‚2H2O precipitated almost
immediately from solution. [Na(NO3)‚NiL2]‚MeOH, [Na(NO3)‚NiL4]‚
MeOH, and [NaI(NiL3)2]‚MeOH‚H2O were obtained as solids only after
the methanol was reduced to approximately one-third of the original
volume and after several days had elapsed. Adduct formation did not
take place between either sodium nitrate or iodide and ZnL4.
Figure 1. Divalent metal salicylaldimines employed as donor ligands.
Preparation of Potassium and Cesium Adducts. Reactions of
potassium and cesium nitrates and iodides with all of the metal
salicylaldimines in Figure 1 were carried out as follows. Approximately
0.8 nmol of the solid salt was added to an equimolar quantity of the
metal salicylaldimine in methanol (30 cm3) at room temperature. In
the case of the reactions involving copper, nickel, and cobalt salicyl-
aldimines the addition of the salt resulted in instant visible color
changes. After the mixture was stirred for 12 h, solid products were
isolated by filtration and dried under vacuum. Unlike organotin adducts
of cobalt salicylaldimines, the alkali metal adducts were not prone to
cobalt oxidation. Adducts of copper and zinc were least soluble, while
iodides tended to be more soluble than nitrates. Yields of adducts were
always greater than 85%. In most cases reactions were also carried out
with a 2-fold excess of metal salicylaldimine complex, but in no case
did the excess metal salicylaldimine alter the stoichiometry of the
adduct.
Reactions of M′A (M′ ) K and A ) thiocyanate or acetate; M′ )
K or Cs and A ) tetraphenylborate) with ML4 were carried out as
described for the nitrates and iodides. Thiocyanates and tetraphenylbo-
rates precipitated immediately as did [K(acetate)‚ZnL4]‚MeOH. To
obtain other acetates in high yield, it was necessary to reduce the volume
of methanol to approximately one-quarter of the original. Potassium
and cesium tetraphenylborates did not form adducts with ZnL4.
Analytical Data for Key Complexes of the Discussion. Anal. Calcd
for Na(NO3)/NiL4‚MeOH (C20H24N3O8NaNi): C, 46.53; H, 4.45; N,
8.14. Found: C, 46.41; H, 4.31; N, 8.69. Anal. Calcd for Na(NO3)‚
CuL1‚MeOH (C19H22N3O8CuNa): C, 45.62; H, 4.40; N, 8.40. Found:
C, 44.98; H, 4.15; N, 8.81. Anal. Calcd for [Cs(NO3)‚NiL4]‚MeOH
(C58H64N9O22Cs3Ni3): C, 38.42; H, 3.55; N, 6.95. Found: C, 37.99;
H, 3.21; N, 6.83. Anal. Calcd for [Na‚(NiL4)2]I‚EtOH‚H2O (C40H48N4O10-
INaNi2): C, 48.32; H, 4.63; N, 5.64; I, 12.77. Found: C, 48.22; H,
4.55; N, 5.77; I, 13.00. Anal. Calcd for CsI‚(NiL2)2‚H2O (C38H44N4O9-
ICsNi2): C, 42.41; H, 3.90; N, 5.21; I, 11.80. Found: C, 42.64; H,
3.79; N, 4.98; I, 11.20. Anal. Calcd for CsI3‚(NiL4)2 (C38H40N4O8I3-
CsNi2): C, 34.79; H, 3.07; N, 4.27; I, 29.02. Found: C, 35.10; H,
2.98; N, 4.42; I, 28.67. Anal. Calcd for KiNiL4 (C19H20N2O4IKNi): C,
40.37; H, 3.54; N, 4.96; I, 22.47. Found: C, 40.24; H, 3.53; N, 4.94;
I, 21.60. Anal. Calcd for (K‚MeCN)2‚(NiL4)3(BPh4)2 (C109H106N7O12B2K2-
Ni3): C, 65.59; H, 5.35; N, 5.61. Found: C, 65.21; H, 5.37; N, 6.14.
Anal. Calcd for (CsI)2‚(CuL3)3‚2H2O (C66H58N6O14I2Cu3Cs2): C, 42.40;
H, 3.13; N, 4.50; I, 13.58. Found: C, 41.70; H, 3.06; N, 4.46; I, 15.59.
Anal. Calcd for (NaI)2‚(CuL3)3‚2H2O (C66H58N6O14I2Cu3Na2): C, 47.88;
H, 3.50; N, 5.08; I, 15.34. Found: C, 48.13; H, 3.55; N, 5.07; I, 16.04.
Crystallization of Adducts. Attempts were made to grow crystals
from the solvents methanol, ethanol, tetrahydrofuran, acetone, and
acetonitrile using the following procedure. Saturated solutions of the
adducts, close to the refluxing temperature of the solvent, were placed
in preheated narrow-bore tubes that were immersed in hot water in a
thermos flask. In this fashion suitable crystals of Na(NO3)‚NiL4‚MeOH,
Na(NO3), CuL1‚MeOH, [Cs(NO3)‚NiL4]3‚MeOH, CsI‚(NiL2)2‚H2O, and
KI‚NiL4 were obtained from methanol, while crystals of [Na‚(NiL4)2]‚
EtOH‚H2O and [(K‚MeCN)2‚(NiL4)3](BPh4)2 were obtained from
ethanol and acetonitrile, respectively. It is noteworthy that CsI‚(NiL2)2‚
H2O was obtained from the attempted crystallization of the 1/1 adduct.
Poor quality crystals of CsI‚NiL4 were obtained from methanol.
stances, the coordinated metal ion may be either retained in the
plane of the donor phenolic and methoxy oxygens (as has been
demonstrated in the case of diorganotin(IV)10 and tin(II)11
complexation) or sandwiched between two sets of Schiff base
oxygens (as has been demonstrated by the complexation of lead-
(II) ions).12 It would thus be anticipated that the size of an alkali
metal ion would influence its mode of complexation by a metal
methoxysalicylaldimine complex and that this, in turn, might
have a critical influence on selectivity patterns. The present
paper focuses primarily on solid-state structural aspects of adduct
formation between methoxy-substituted metal salicylaldimine
complexes and alkali metal ions. The study reveals novel
coordination behavior of the alkali metal ions and provides some
understanding of the structural factors that contribute to the
formation and stoichiometry of the adducts. Furthermore, the
manner in which infinite stacking assemblies are achieved in
two of the adducts may serve as a useful model for the design
of functional molecular assemblies.
Experimental Section
All of the salicylaldimine complexes ML1, ML2, ML3, and ML4 (see
Figure 1) were prepared by standard literature methods.13
Preparation of Sodium Nitrate and Iodide and Tetraphenyl-
borate Adducts. Reactions of NaA (A ) nitrate or iodide) with each
(4) (a) Gambarotta, S.; Floriani, C.; Chiesi-Villa, A.; Guastini, C. J. Chem.
Soc., Chem. Commun. 1982, 756-758. (b) Gambarotta, S.; Urso, F.;
Floriani, C.; Chiesi-Villa, A.; Guastini, C. Inorg. Chem. 1983, 22,
3966-3972.
(5) (a) Gambarotta, S.; Mazzanti, M.; Floriani, C.; Zehnder, M. J. Chem.
Soc., Chem. Commun. 1984, 1116-1118. (b) Franceschi, F.; Solari,
E.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C.; Rosi, M. Chem. Eur. J.,
in press.
(6) Gallo, E.; Solari, E.; Re, N.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C.
J. Am. Chem. Soc. 1997, 119, 5144-5154.
(7) (a) Fachinetti, G.; Floriani, C.; Zanazzi, P. F.; Zanzari, A. R. Inorg.
Chem. 1979, 18, 3469-3475. (b) Arena, F.; Floriani, C.; Zanazzi, P.
F. J. Chem. Soc., Chem. Commun. 1987, 183-184. (c) De Angelis,
S.; Solari, E.; Gallo, E.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C. Inorg.
Chem. 1996, 35, 5995-6003.
(8) Gallo, E.; Solari, E.; Floriani, C.; Chiesi-Villa, A.; Rizzoli, C. Inorg.
Chem. 1997, 36, 2178-2186.
(9) (a) Miyasaka, H.; Matsumoto, N.; Okawa, H.; Re, N.; Gallo, E.;
Floriani, C. J. Am. Chem. Soc. 1996, 118, 981-994. (b) Re, N.; Gallo,
E.; Floriani, C.; Miyasaka, H.; Matsumoto, N. Inorg. Chem. 1996,
35, 5964-5965.
(10) Clarke, B.; Cunningham, D.; Gallagher, J. F.; Higgins, T.; McArdle,
P.; McGinley, J.; N´ı Cholchu´in, M.; Sheerin, D. J. Chem. Soc., Dalton
Trans. 1994, 2473-2482.
(11) Cunningham, D.; Gallagher, J. F.; Higgins, T.; McArdle, P.; Sheerin,
D. J. Chem. Soc., Chem. Commun. 1991, 432-433.
(12) Clarke, B. Ph.D. Thesis, National University of Ireland, Galway,
Ireland, 1997.
(13) McCarthy, P. J.; Hovey, R. J.; Ueno, K.; Martell, A. E. J. Am. Chem.
Soc. 1955, 77, 5820-5824.

Adduct Formation
Inorganic Chemistry, Vol. 39, No. 8, 2000 1641
However, when iodine was added to a hot methanolic solution of the
adduct, the solubility of the adduct was greatly increased, and this led
to the isolation of good quality crystals of the triiodide CsI3‚(NiL4)2.
The tetraphenylborates, on crystallization from acetonitrile, retained
the 2/3 stoichiometry but contained coordinated acetonitrile, as sug-
gested by analytical and infrared data and confirmed by the crystal
structure determination of the adduct [(K‚MeCN)2‚(NiL4)3](BPh4)2.
X-ray Powder Diffractometry. Powder diffraction data were
collected on a JEOL model DX-Go-S diffractometer, which was
modified by the incorporation of a stepping motor and a digital data
collection and display system. Cu KR (λ ) 1.5405 Å) was used; a
nickel filter was employed to remove Kâ. Scans using 0.1° steps and
3 s count times were employed.
Single-Crystal X-ray Diffractometry. With the exception of CsI‚
(NiL2)2‚H2O, crystals of other adducts were mounted on glass fibers.
Because of instability of crystals of CsI‚(NiL2)2‚H2O, the crystal selected
for data collection was coated in liquid paraffin and mounted in a sealed
capillary. Crystallographic details are in Table 1. The structures were
solved by direct methods, SHELXS-86,14 and refined by full matrix
least squares using SHELXL-97.15 SHELX operations were rendered
paperless using ORTEX, which was also used to obtain the drawings.16
Data were corrected for Lorentz and polarization effects but not for
absorption. Hydrogen atoms were included in calculated positions with
thermal parameters 30% larger than the atom to which they were
attached. The non-hydrogen atoms were refined anisotropically. All
efforts to refine the structure of [Na(NO3)‚NiL4]‚MeOH to a final value
of R1 less than approximately 10% failed, though at this point the esd
values were not unreasonable. The failure to refine the structure to a
more satisfactory level may be the result of the very poor diffracting
power of the crystal, as a result of which there was an abnormally
large number of reflections of low intensity. All calculations were
performed on a Pentium PC.
Results and Discussion
The metal salicylaldimine ligands investigated in the present
study are shown in Figure 1. These were chosen to investigate
not only the relative effects of the metal M but also the effects
of nitrogen bridging group on mode of complexation of alkali
metal ions. The emphasis on the bridging group stems from
the fact that studies in this laboratory have shown, as expected,
that the number of carbon atoms linking the imine nitrogen
atoms has a crucial bearing on the behavior of these metal
salicylaldimines as ligands. For example, when the bridging
group is the 1,3-propylene group, metal atoms (which are being
complexed) tend to sit in the plane of the four Schiff base
oxygens where they are held by strong donor bonds from the
phenolic oxygen atoms and weaker donor bonds from the
methoxy oxygen atoms (see, for example, refs 9 and 10). This
mode of complexation was not observed thus far when the
nitrogen bridging group was a two-carbon-atom bridge as, for
example, in ML1, ML2, and ML3. Reactions of the last metal
salicylaldimines with Lewis acids generally resulted in the
formation of aqua adducts with the donor water linked to the
metal salicylaldimne oxygen atoms through hydrogen bonding.17
Exceptions were found in the case of [K‚(MnL1)2Fe(CN)6]n and
the mixed valent lead complex PbCl‚[NiL1]2(PbEt2Cl3) in
which K+ and PbCl+ are sandwiched between two molecules
of ML1, being retained by eight donor bonds from Schiff base
oxygen atoms.9,11 In the former case, different substitution on
the Schiff base skeleton led to adducts with different stoichi-
ometry. Stereoelectronic modulation of the ligand is thus
confirmed to be crucial for the final adducts’ composition.
(14) Sheldrick, G. M. Acta Crystallogr. 1990, A46, 467-473.
(15) Sheldrick, G. M. SHELXL-97 (a computer program for crystal structure
determination); University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(16) McArdle, P. J. Appl. Crystallogr. 1995, 28, 65-65.
(17) Cunningham, D.; Gallagher, J. F.; Higgins, T.; McArdle, P.; McGinley,
J.; O’Gara, M. J. Chem. Soc., Dalton Trans. 1993, 2183-2190.

1642 Inorganic Chemistry, Vol. 39, No. 8, 2000
Cunningham et al.
Adduct formation between all of the nickel and copper
salicylaldimines of Figure 1 and M′A (M′ ) Na, K, and Cs; A
) iodide and nitrate) was investigated. The cobalt salicylaldi-
mine complexes were less thoroughly investigated (having
discovered that these were behaving similarly to the copper and
nickel salicylaldimines), and in the case of the zinc salicyla-
ldimines, investigations were limited to the salicylaldimines
ZnL4. Since the investigation revealed that the anion was having
an influence on adduct stoichiometry, the study was broadened
somewhat to include adduct formation between ML4 and both
KA (A ) thiocyanate or acetate) and M′B(Ph)4 (M′ ) Na, K,
and Cs). Practically every acid/base combination studied led to
the isolation of an adduct with a stoichiometry that was
positively identified by a combination of powder diffractometry
and elemental analytical data.
Adduct Formation by Alkali Metal Nitrates. Both potas-
sium and cesium nitrates reacted with all of the metal Schiff
base complexes to give exclusively adducts of 1/1 stoichiometry.
These precipitated rapidly from methanol in greater than 90%
yield. Sodium reacted with the copper, nickel, and cobalt Schiff
base complexes but not with the zinc analogues; methanolic
solutions containing the last complexes in combination with
sodium nitrate ultimately yielded crystalline sodium nitrate.
Where sodium adducts formed, they were of 1/1 stoichiometry
and had much greater solubility than their potassium and cesium
analogues, the one exception being Na(NO3)‚NiL3, which
precipitated immediately from solution in greater than 90%
yield. The adduct Na(NO3)‚NiL4 was the most soluble, remain-
ing in methanol solution for several days following reaction.
Some significant differences were noted in the behavior of
the divalent metal Schiff base complexes as donor ligands in
reactions where alkali metal ions were in competition. Despite
the fact that the adducts of Cs(NO3) with NiL4, NiL2, and CuL3
had very low solubilities in methanol, the presence of sodium
nitrate along with each of these Schiff base complexes in
equimolar quantities in methanol completely prevented the
precipitation of the cesium nitrate adducts. After several days,
the sodium nitrate adduct crystallized from solution (as it would
do in the absence of cesium nitrate), and this was followed by
crystallization of cesium nitrate. When a methanol solution of
sodium and cesium nitrates in equimolar quantities was added
to a solution/suspension of NiL3, the sodium adduct was
preferentially precipitated (the adduct Na(NO3)‚NiL3, as men-
tioned earlier, has low solubility in methanol). However, in the
case of other Schiff base complexes of this study, the presence
of sodium nitrate did not prevent the precipitation of the cesium
nitrate adduct nor did its presence prevent the precipitation of
any potassium nitrate adducts.
Figure 2. Asymmetric units of (a) [Na(NO3)(MeOH)]‚NiL4 and (b)
[Na(NO3)(MeOH)]‚CuL1 (30% ellipsoids).
along with NiL4, the adduct that precipitated (rapidly) contained
both cesium and potassium, and X-ray fluorescence spectroscopy
suggested a considerable cesium content. This latter observation
was further suggested by carbon, hydrogen, and nitrogen
analytical data, which were approximately midway between
those expected for the potassium and cesium adducts. Despite
the fact that the diffraction patterns of potassium and caesium
adducts of NiL4 are strikingly similar and indeed very similar
to the diffraction pattern of the mixed ion precipitate, there was
sufficient evidence from weaker lines of the diffraction patterns
that the mixed ion precipitate represented a new solid-state phase
rather than a mixture of the two adducts. The observations
suggest that the new phase may also have an infinite sandwich
structure broadly similar to that of the cesium nitrate adduct
(the infinite sandwich structure of this latter adduct is described
later). In the case of the mixed metal adduct, both potassium
and cesium ions are being sandwiched (either in random or in
ordered fashion) between nickel salicylaldimine molecules.
It proved to be possible to grow crystals of sodium nitrate
adducts with CuL1 and NiL4, and their structures were deter-
mined. The asymmetric units of the adducts are in Figure 2.
The geometry about sodium is extremely similar in these
adducts. The sodium sits in the plane of the Schiff base oxygen
atoms and is involved in electrostatic interactions with all four.
Further significant interactions with two nitrate oxygen atoms
and a methanol oxygen, trans to the nitrate, complete the
coordination sphere, thus resulting in an unusual seven-
coordination geometry about sodium. A particular point of note
regarding the copper/sodium adduct is the fact that it represents
In view of the fact that sodium ions are not complexed by
ZnL4 whereas potassium and cesium ions form adducts that
precipitate in high yield, selective precipitation of either of these
ions in the presence of sodium ions is clearly possible using
the zinc salicylaldimine as the complexing ligand. However,
since both potassium and cesium adducts precipitate immediately
from methanolic solutions in high yield, it is clear that the metal
salicylaldimines do not offer the possibility of selective
precipitation of either of these ions, at least not under the
conditions employed. On the other hand, from the fact that the
presence of sodium ions along with NiL4 totally suppressed the
precipitation of the cesium adduct but not the potassium adduct,
it would appear that this would offer a means of effecting their
separation.
Rather surprisingly, however, when potassium nitrate was
added to a solution containing both cesium and sodium ions

Adduct Formation
Inorganic Chemistry, Vol. 39, No. 8, 2000 1643
Table 2. Selected Bond Lengths (Å) for [Na(NO3)MeOH]‚NiL4 and
[Na(NO3)MeOH]‚CuL1
[Na(NO3)MeOH]‚NiL4
[Na(NO3)MeOH]‚CuL1
Ni(1)-O(1)
Ni(1)-O(2)
Ni(1)-N(1)
Ni(1)-N(2)
Ni(1)-Na(1)
Na(1)-O(1)
Na(1)-O(2)
Na(1)-O(3)
Na(1)-O(4)
Na(1)-O(5)
Na(1)-O(6)
Na(1)-O(8)
1.864(3)
Cu(1)-O(1)
1.8799(14)
1.8855(15)
1.9184(17)
1.9278(17)
3.3468(8)
2.3197(17)
2.3743(16)
2.5482(16)
2.6388(17)
2.549(2)
1.864(3)
1.894(4)
1.921(3)
3.466(2)
2.342(4)
2.375(3)
2.367(4)
2.397(4)
2.615(6)
2.535(5)
2.445(5)
Cu(1)-O(2)
Cu(1)-N(1)
Cu(1)-N(2)
Cu(1)-Na(1)
Na(1)-O(1)
Na(1)-O(2)
Na(1)-O(3)
Na(1)-O(4)
Na(1)-O(5)
Na(1)-O(7)
Na(1)-O(8)
2.456(2)
2.368(2)
the first confirmed case where a metal methoxysalicylaldimine
complex containing a two-carbon-atom bridge retains the Lewis
acid metal in the plane of the four ligand oxygen atoms.
The most significant difference in Na-O bond lengths
between the two adducts is seen in the bonds involving the
methoxy oxygens where they are much longer in the case of
the copper adduct (see Table 2). The difference can be traced
to the shorter imine nitrogen bridge in the copper complex,
resulting in a N-Cu-N bond angle of 86.07(7)° compared to
a N-Ni-N bond angle of 98.4(4)° in the nickel complex. The
smaller angle results in a methoxy oxygen separation of 5.15
Å in the copper complex compared to a separation of 4.75 Å in
the case of the nickel complex. The lengthening of the sodium-
methoxy oxygen contacts in the sodium/copper adduct results
in a strengthening of the sodium water and sodium nitrate
interactions.
Of particular interest is the precise location of the sodium
ion in the sodium/nickel complex compared to the location of
p-block metals in their adducts with the same nickel salicyl-
aldimine. In the case of the latter metals, and indeed silver in
Ag(NO3)NiL4,18 it is invariably found that, as a direct conse-
quence of the trapezium arrangement of the bonding oxygen
atoms, the metal is located significantly closer to the phenolic
than to the methoxy oxygens to minimize repulsive interactions
between bonding electrons (as predicted from Kepart’s calcula-
tions19). That sodium does not do so is a direct consequence of
the purely electrostatic nature of the Na-O interactions.
It has recently been shown that Ni(MeOsalNH)2 [HMeOsal-
NH ) 3-methoxy salicyideneamine) forms an adduct with
sodium perchlorate.20 The location of the sodium ion in this
complex is very similar to that in the present nickel-sodium
adduct. However, the perchlorate provides only one significant
Ni-O interaction, and this, coupled with an interaction with a
water molecule, resulted in a coordination number of 6 for
sodium (in an unusual pentagonal pyramidal geometry) com-
pared with 7 in the present structures.
Figure 3. Section of the infinite sandwich structure of [Cs(NO3)‚
NiL4]3‚MeOH (atoms portrayed as spheres for clarity). Dotted lines
correspond to contacts greater than 3.5 Å.
Table 3. Selected Bond Lengths (Å) for [Cs(NO3)‚NiL4]3‚MeOH
Cs(1)-O(1)
Cs(1)-O(3)
Cs(1)-O(5)
Cs(1)-O(7)
Cs(1)-O(13)
Cs(1)-O(18)
Cs(2)-O(2)
Cs(2)-O(4)
Cs(2)-O(10)
Cs(2)-O(12)
Cs(2)-O(18)
Cs(2)-O(21)
Cs(3)-O(6)′
Cs(3)-O(8)′
Cs(3)-O(10)
Cs(3)-O(12)
Cs(3)-O(21)
Ni(1)-O(1)
Ni(1)-N(1)
Ni(1)-O(13)
Ni(2)-O(5)
Ni(2)-N(3)
Ni(3)-O(9)
Ni(3)-N(5)
Ni(3)-O(17)
3.142(6)
3.218(7)
3.253(6)
3.391(7)
3.297(8)
3.138(8)
3.136(6)
3.208(7)
3.232(6)
3.232(7)
3.589(10)
3.934(8)
3.215(6)
3.131(7)
3.033(6)
3.161(7)
3.269(9)
1.996(5)
2.015(7)
2.268(8)
1.866(5)
1.881(8)
1.991(6)
2.038(8)
2.209(8)
Cs(1)-O(2)
Cs(1)-O(4)
Cs(1)-O(6)
Cs(1)-O(8)
Cs(1)-O(14)
Cs(2)-O(1)
Cs(2)-O(3)
Cs(2)-O(9)
Cs(2)-O(11)
Cs(2)-O(17)
Cs(2)-O(19)
Cs(3)-O(5)′
Cs(3)-O(7)′
Cs(3)-O(9)
Cs(3)-O(11)
Cs(3)-O(14)′
Cs(3)-O(22)
Ni(1)-O(2)
Ni(1)-N(2)
Ni(1)-O(19)
Ni(2)-O(6)
Ni(2)-N(4)
Ni(3)-O(10)
Ni(3)-N(6)
Ni(3)-O(22)
3.007(6)
3.188(6)
3.047(6)
3.361(8)
3.268(8)
2.989(6)
3.286(7)
3.037(6)
3.313(7)
3.110(8)
3.193(8)
3.105(6)
3.237(6)
3.204(7)
3.164(7)
3.191(8)
4.006(8)
1.991(5)
2.042(8)
2.133(8)
1.852(6)
1.868(7)
2.007(6)
2.042(8)
2.153(7)
Although it proved to be possible to grow crystals of several
other alkali metal nitrate adducts, it was only in the case of the
cesium nitrate adduct of NiL4 that crystals suitable for a
crystallographic study were obtained. The asymmetric unit of
this fascinating adduct (see Figure 3 and Table 3 for selected
bond lengths) corresponding to the formula {[Cs(NO3)‚NiL4}3‚
MeOH is part of an infinite sandwich structure generated by
the 21 axis lying close to the cesium ions; the cesium ions are
in an oval-shaped helical array about the 21 axis. This is the
first example of a metal salicylaldimine ligand being involved
in an infinite sandwich structure, though not the first such
structure involving cesium ions. The first infinite sandwich
structure containing cesium ions was recently reported employ-
ing a crown ether as ligand.21 The three crystallographically
independent cesium ions are in similar locations insofar as they
are each sandwiched between two sets of five oxygen atoms.
Four of the five oxygens in each set belong to a nickel
salicylaldimine molecule while the fifth belongs to a nitrate
group, the three nitrate oxygens in question being O(14), O(18),
(18) O’Gara, M. Ph.D. Thesis, National University of Ireland, Galway,
Ireland, 1993.
(19) Kepert, D. L. J. Organomet. Chem. 1976, 107, 49-54.
(20) Costes, J.-P.; Dahan, F.; Laurent, J.-P. Inorg. Chem. 1994, 33, 2738-
2742.
(21) Baldas, J.; Colmanet, S. F.; Williams, G. A. J. Chem. Soc., Chem.
Commun. 1991, 954-956.

1644 Inorganic Chemistry, Vol. 39, No. 8, 2000
Cunningham et al.
and O(21). The oxygens of each set are approximately coplanar
and lie at the corners of an elongated pentagon, the elongation
in each case originating from the displacement of the nitrate
oxygen from the more ideal pentagonal geometry location. As
a result of the sandwich arrangement so described, each cesium
has a coordination geometry that is most aptly described as
pentagonal antiprismatic. Cs-O interactions within these co-
ordination environments are generally strong, apart from those
involving O(18) and O(21). There are, however, additional
Cs-O contacts resulting from nitrate and methanol (see Table
3) such that the overall coordination geometries of Cs(1) and
Cs(3) can be described as monocapped pentagonal antiprismatic,
while that of Cs(2) can be described as bicapped pentagonal
antiprismatic.
The salicylaldimine oxygen atoms provide, on average, much
stronger contacts with the cesium ions than do the crown ether
oxygens in the infinite sandwich structure of [Cs(18-crown-
6)]+[tcNCl4]- 21. In the case of the structure of the latter, three
of the oxygens provide Cs-O contacts of 3.34(1), 3.34(1), and
3.35(1) Å (most of the Cs-O distances involving salicylaldimine
oxygens are less than these values), while the other three provide
significantly longer contacts of 3.66(1), 3.67(1), and 3.68(1)
Å. All of the Cs+-Cs+ separations in the present structure
(3.872, 3.909, and 4.050 Å) are smaller than the Cs+-Cs+
separation of 4.275 Å in the crown ether structure.
Apart from forming close contacts with cesium ions, O(13)
and O(19) and O(17) also form bonds to nickel ions. The former
two oxygens complete the octahedral geometry about Ni(1),
while the latter oxygen, along with the methanol oxygen O(22),
complete the octahedral geometry about Ni(3). There are no
bonds to Ni(2) other than those provided by the Schiff base
ligand. Thus, the infinite sandwich structure consists of pairs
of nickel salicylaldimines [nickel salicylaldimines (1) and (3)
with octahedral nickel separated by single square-planar nickel
salicylladimines [nickel salicylaldimine (2)].
Figure 4. Asymmetric unit of [Na‚(NiL4)2]I‚EtOH (30% ellipsoids).
The iodide and lattice ethanol are omitted.
Table 4. Selected Bond Lengths (Å) for [Na‚(NiL4)2]I‚EtOH
Ni(1)-O(1)
Ni(1)-O(2)
Ni(1)-N(1)
Ni(1)-N(2)
Ni(1)-Na(1)
Na(1)-O(1)
Na(1)-O(2)
Na(1)-O(3)
Na(1)-O(4)
1.847(4)
1.864(4)
1.868(6)
1.886(5)
3.530(2)
2.424(4)
2.452(4)
2.486(5)
2.493(5)
Ni(2)-O(5)
Ni(2)-O(6)
Ni(2)-N(3)
Ni(2)-N(4)
Na(1)-Ni(2)
Na(1)-O(5)
Na(1)-O(6)
Na(1)-O(7)
Na(1)-O(8)
1.846(4)
1.839(4)
1.863(5)
1.866(4)
3.472(2)
2.412(4)
2.413(4)
2.519(5)
2.544(4)
metal atom in a salicylaldimine complex having a two-carbon-
atom (imine) nitrogen-bridging group is very much less likely
than the metal of an analogous complex with a three-carbon-
atom bridge to expand its coordination beyond 4 when the
salicylaldimine complex adopts a donor ligand role.
Adduct Formation by Alkali Metal Iodides. Both potassium
and cesium iodides, like their nitrate analogues, reacted with
the metal salicylaldimines to yield adducts that precipitated
rapidly from methanol but in somewhat reduced yields (ranging
from approximately 80% to 85%). Most of the adducts were of
1/1 stoichiometry, but there were notable exceptions. Both
potassium and cesium iodides formed 1/2 adducts with NiL3,
and potassium iodide formed an adduct with the same stoichi-
ometry with CuL3. Cesium iodide formed an adduct of 2/3
stoichiometry with CuL3.
Adducts of sodium iodide precipitated from solution very
slowly except for those involving NiL2 and NiL3, which
precipitated very rapidly. All of the sodium adducts with the
nickel salicylaldimines were of 1/2 stoichiometry, whereas those
with the copper salicylaldimines were of 1/1 stoichiometry
except for the adduct with CuL3, which appears to be of 2/3
stoichiometry. As had been the case with sodium nitrate, sodium
iodide did not form adducts with ZnL4.
In several instances, crystals of reasonable quality were
obtained, thus allowing the structures of the sodium and
potassium iodide adducts of NiL4, the cesium iodide adduct with
NiL2, and the cesium triiodide adduct of NiL4 to be determined
crystallographically.
The asymmetric unit of the 1/2 adduct of sodium iodide with
NiL4 is shown in Figure 4, and selected bond lengths are in
Table 4. Iodide has no meaningful contacts with either sodium
or nickel but is involved in weak hydrogen-bonding interactions
with water and ethanol molecules of crystallization (these are
omitted from Figure 4). Two nickel salicylaldimines approach
In view of the 1/1 stoichiometry of all of the adducts of both
cesium and potassium nitrate with the metal salicylaldimines
of Figure 1 and the apparent inability of these large ions to be
located within the plane of the Schiff base oxygen atoms, the
coordination requirements of the ions are most easily achieved
in infinite sandwich structures broadly similar to that already
described for the cesium nitrate adduct of NiL4 or that of KI‚
NiL4 (to be described later). The marked similarity of the powder
diffraction patterns of all of the 1/1 adducts, including that of
[Cs(NO3)‚NiL4]‚MeOH, is consistent with, though not total
proof of, a common infinite sandwich type structure. It is clear,
however, from magnetic data that while the adducts may share
a common infinite sandwich type structure they do not feature
the diversity of nitrate roles that led to three nickel environments
in the cesium nitrate adduct of NiL4.
The adduct K(NO3)‚NiL4 has a magnetic moment of 3.31 B
M per nickel, while the adducts Cs(NO3)‚CoL4 and Cs(NO3)‚
CoL4 have magnetic moments of 4.97 and 4.84 B M per cobalt,
respectively, indicating that these do not have the complication
of square-planar geometries (about the transition metal ions)
coexisting with other geometries within their structures. How-
ever, in view of the magnetic moments and the apparent
anhydrous nature of the adducts, it is clear that the nitrates do
coordinate to the nickel and cobalt centers.
By contrast, nitrate coordination to the transition metal ions
does not occur in adducts containing ML1, ML2, and ML3 (M
) Ni or Co), since the nickel adducts are diamagnetic and the
cobalt analogues are low-spin complexes. These observations
are in accord with our earlier observations that indicate that a

Adduct Formation
Inorganic Chemistry, Vol. 39, No. 8, 2000 1645
Figure 6. Section of the nominal chain polymeric structure of CsI3‚
((NiL4)2‚ (atoms portrayed as spheres for clarity).
Table 6. Selected Bond Lengths (Å) for CsI3‚(NiL4)2
Figure 5. Asymmetric unit of CsI‚(NiL2)2‚H2O (30% ellipsoids) with
Cs(1)-O(1)
Cs(1)-O(3)
Cs(1)-I(1)
Ni(1)-O(1)
Ni(1)-N(1)
3.027(3)
3.262(4)
3.9666(6)
1.861(3)
1.883(4)
Cs(1)-O(2)
Cs(1)-O(4)
I(1)-I(2)
Ni(1)-O(2)
Ni(1)-N(2)
3.149(4)
3.127(4)
2.9047(5)
1.862(3)
1.881(4)
the water molecule omitted.
Table 5. Selected Bond Lengths (Å) for CsI‚(NiL2)2‚H2O
Cs(1)-O(1)
Cs(1)-O(2)
Cs(1)-O(3)
Cs(1)-O(4)
Cs(1)-O(5)
Cs(1)-O(6)
Cs(1)-O(7)
Cs(1)-O(8)
Cs(1)-I(1)
3.072(7)
3.063(7)
3.139(9)
3.142(8)
3.111(7)
3.081(7)
3.175(9)
3.120(9)
4.090(1)
Ni(1)-O(1)
Ni(1)-O(2)
Ni(1)-N(1)
Ni(1)-N(2)
Ni(2)-O(5)
Ni(2)-O(6)
Ni(2)-N(3)
Ni(2)-N(4)
1.851(7)
1.867(7)
1.821(10)
1.846(9)
1.869(7)
1.843(8)
1.841(12)
1.845(13)
from the planes defined by O(1)-O(4) and O(5)-O(8),
respectively, the shorter distance probably resulting from the
extra attraction to the iodide lying close to the plane defined
by O(1)-O(4).
The 1/1 addition complex formed between cesium iodide and
NiL4 had limited solubility in methanol, and all attempts to grow
single crystals from this solvent resulted in the formation of
small clumps of twinned crystals. Addition of I2 to the solution
greatly enhanced the solubility of the cesium iodide adduct, and
from the more concentrated solution, crystals of the adduct CsI3‚
(NiL4)2 were obtained. This is an interesting adduct having a
nominal chain polymeric structure as a result of bridging
triiodide groups. A section of the polymeric chain is shown in
Figure 6, while selected bond lengths are in Table 6. The chain
packing arrangement coupled with short aromatic ring contacts
suggests the presence of attractive π-aromatic interactions
linking all neighboring chains. Both cesium and the central
iodine of the triiodide groups are on centers of inversion. Thus,
the structure contains true triiodide ions with I-I bond lengths
of 2.904(1) Å.
As had been the case in the cesium iodide adduct structure
described above, the cesium ions are sandwiched between two
sets of Schiff base oxygens giving eight Cs-O contacts with
contact distances in the range 3.027(3)-3.262(4) Å. Each set
of Schiff base oxygens is essentially coplanar, and these planes
are close to being parallel. An iodine of a triiodide ion lies close
to each of these planes, and these iodine atoms give rise to Cs-I
contacts at 3.967(1) Å, compared with a Cs-I contact at 4.090-
(2) Å in the iodide structure discussed above. Cesium is 1.81
Å from the planes defined by the Schiff base oxygens. It is
10-coordinate, and the coordination geometry can be described
as distorted pentagonal antiprismatic, a geometry that is very
similar to that in the 1/2 adduct of potassium chloride with O,O′-
catecholdiacetic acid.22 However, the resemblance between these
two structures runs deeper than this. Potassium is sandwiched
between two sets of oxygens disposed toward the corners of a
trapezium, quite similar to that defined by the Schiff base
oxygens, and just as the iodine atoms complete the pentagonal
faces of the coordination polyhedron in the present structure,
oxygen atoms from neighboring molecules complete the pen-
tagonal faces in the structure of the potassium adduct.
the sodium ion such that the planes defined by their oxygen
atoms are inclined at an angle of approximately 81.5° to each
other. Sodium is within each of these planes and is located such
that there is reasonable uniformity in the sodium-oxygen
electrostatic interactions (Na-O contacts range from 2.405(7)
to 2.537(8) Å). To achieve this, sodium is located close to the
outer edge of each of the trapeziums of oxygen atoms such that
the methoxy oxygens of each ligand subtend angles of 173.3-
(1)° and 175.9(3)° at sodium. As a result of the interlacing of
the two trapeziums, sodium achieves fairly regular dodecahedral
geometry, an unusual coordination geometry for sodium. It
seems highly likely that all of the sodium iodide adducts of 1/2
stoichiometry have the same dodecahedral coordination geom-
etry about sodium. A more usual six-coordinate geometry could
easily have been achieved by sodium by forming 1/1 addition
complexes with the Schiff base oxygens and two water oxygen
atoms involved in bonding interactions with the cation. This
probably is the situation in the 1/1 adducts (Na‚CuL2)I‚2H2O
and (Na‚CuL4)I‚2H2O.
Although the adduct that immediately precipitated from
solution following reaction of cesium iodide with NiL2 was of
1/1 stoichiometry, the adduct that was obtained from slow
crystallization was the 1/2 adduct CsI‚(NiL2)2‚H2O. The struc-
ture of this adduct is strikingly different from that of the sodium
iodide adduct structure described above, the difference being
dictated by the much larger size of cesium relative to that of
sodium. In this case (see Figure 5) the cesium ion is sandwiched
between two sets of Schiff base oxygen atoms with Cs-O
contacts ranging from 3.06(1) to 3.21(1) Å (see Table 5); these
are comparable with the shorter Cs-O contacts in the cesium
nitrate structure described earlier. Each set of Schiff base
oxygens is essentially coplanar, and these planes are essentially
parallel. The iodide is 0.24 Å from the plane defined by the set
of oxygens O(1)-O(4) and situated at distances of 4.51 and
4.52 Å from the methoxy oxygens O(3) and O(4), respectively.
There is a Cs-I contact at 4.090(2) Å, which brings the overall
cesium coordination number to 9. Cesium is 1.67 and 1.78 Å
(22) Green, E. A.; Duax, W. L.; Smith, G. M.; Wudl, F. J. Am. Chem.
Soc. 1975, 97, 6689-6692.

1646 Inorganic Chemistry, Vol. 39, No. 8, 2000
Cunningham et al.
a
Table 7. Selected Bond Lengths (Å) for KI‚NiL4
K(1)-O(1)
K(1)-O(3)
K(1)-O(5)
K(1)-O(7)
K(1)-I(1)
K(2)-O(1)
K(2)-O(3)
K(2)-I(1)
K(3)-O(6)
K(3)-O(8)
Ni(1)-O(1)
Ni(1)-N(1)
Ni(1)-I(1a)′
Ni(2)-O(5)
Ni(2)-N(3)
Ni(2)-I(1)
2.886(2)
3.056(2)
2.819(2)
3.171(3)
3.7111(10)
2.812(2)
3.263(3)
4.196(1)
2.818(2)
3.218(2)
1.898(2)
1.911(3)
3.1760(9)
1.975(2)
2.016(3)
3.0655(8)
K(1)-O(2)
K(1)-O(4)
K(1)-O(6)
K(1)-O(8)
K(1)-I(2)
K(2)-O(2)
K(2)-O(4)
K(3)-O(5)
K(3)-O(7)
K(3)-I(2)
Ni(1)-O(2)
Ni(1)-N(2)
Ni(1)-I(2a)
Ni(2)-O(6)
Ni(2)-N(4)
Ni(2)-I(2)′
2.856(2)
3.244(3)
2.821(2)
3.060(2)
4.180(1)
2.826(2)
2.989(2)
2.806(2)
3.067(2)
3.7892(6)
1.897(2)
1.920(3)
3.0367(18)
1.970(2)
1.994(3)
3.0135(7)
a O(1) and O(2) are the phenolic oxygens and O(3) and O(4) the
methoxy oxygens of NiL4(1). O(5) and O(6) are the phenolic oxygens
and O(7) and O(8) the methoxy oxygens of NiL4(2) (see Figure 7).
and 3.0135(7) Å, respectively. By contrast, there are no
significant interactions between Ni(1) and iodide ions, and thus,
this nickel remains square-planar. As a result of the centers of
inversion on K(2) and K(3), pairs of nickel salicylaldimine
molecules exist with octahedral nickel and these alternate with
pairs containing square-planar nickel. When both iodide ions
occupy their 30% occupation sites, Ni(2) becomes diamagnetic
and Ni(1) becomes octahedrally coordinated as a result of
the formation of Ni(1)-I(1) and Ni(1)-I(2) bond lengths of
3.1760(9) and 3.0367(18) Å, respectively [note that these bond
lengths are slightly different from those that had been to Ni-
(2)]. Thus, the infinite sandwich still consists of pairs of square-
planar nickel salicylaldimines alternating with pairs of octahedral
nickel salicylaldimines. Since the iodides are crystallographically
independent, the possibility must also be considered where one
iodide is in its 70% occupation site while the other is in its
30% occupation site. Under such circumstances, irrespective
of which iodide is in the 70% occupation site, all nickels have
square pyramidal geometry.
None of the 1/1 adducts of CsI and KI contain solvent or
water molecules that could contribute to the coordination spheres
of the large potassium and cesium ions. Thus, the coordination
demands of the cations are most easily achievable in an infinite
sandwich structure, broadly similar to the two infinite sandwich
structures that were confirmed by the crystallographic data. The
pronounced similarity of the powder diffraction patterns of
potassium and caesium nitrate adducts extends to those of the
iodide analogues, and this is consistent with (though not
necessarily proof of) a common structural type. Magnetic data
do however point to differences in the role of iodide in the
adduct structures.
Figure 7. Section of the infinite sandwich structure of KI‚NiL4 (10%
ellipsoid outlines).
The most intriguing and crystallographically the most com-
plicated structure of this study is that of the 1/1 adduct of
potassium iodide with NiL4. In broad terms, it is similar to the
structure of the cesium nitrate adduct with NiL4 insofar as it is
an infinite sandwich structure (see Figure 7) in which potassium
ions are sandwiched between sets of salicylaldimine oxygen
atoms. All of the oxygen are involved in significant electrostatic
interactions with the cations. Furthermore, as had been the case
with the cesium nitrate structure, ion pairing is evident.
The asymmetric unit consists of two nickel salicylaldimine
complexes, ML4(1) and ML4(2) in Figure 7, two iodide ions,
I(1) and I(2), and three potassium ions. Two of the latter, K(2)
and K(3), are on centers of inversion and the third, K(1), is
close to a center of inversion. Potassium ions alternate through
the sandwich structure in the order K(1), K(2), K(1), K(3) etc.
Considerable complication is introduced by the fact that both
iodides are disordered over two positions with site occupancies
of 70% and 30%. In Figure 7 the iodides are shown in their
70% positions.
With the iodides located in their 70% occupancy sites, all
three potassium ions achieve 10 coordination as a result of eight
contacts with the Schiff base oxygens, those to the phenolic
oxygens being on average approximately 0.2 Å shorter than
those to the methoxy oxygens, and two contacts with iodide
ions. In the case of K(2) both K-I distances are long at 4.196
Å, while for K(1) one of the K-I distances is long at 4.178 Å.
With the iodide ions in their 30% occupation positions, the
coordination numbers of all potassium ions remains unaltered,
but it is now K(3) rather than K(2), which has the two long
K-I contacts (see Table 7 for selected bond lengths).
In contrast to the paramagnetic nature of KI‚NiL4, the cesium
iodide analogue is diamagnetic, as are all other 1/1 adducts of
cesium iodide and potassium iodide with NiL1, NiL2, and NiL3.
Thus, only in the case of the potassium iodide structure do the
iodide ions coordinate to nickel. If the iodide ions in the
diamagnetic adducts are located close to the planes of the Schiff
base oxygen atoms, as they do in the 1/2 adduct of cesium iodide
with NiL2 (the structure of which was described earlier), they
can be involved in electrostatic interactions with the alkali metal
ions without forming bonds to nickel. Presumably this is the
situation that prevails in the diamagnetic adducts.
With the iodides located in their 70% occupancy sites, Ni(2)
achieves six coordination as a result of the Ni-O and Ni-N
bonds within the metal salicylaldimine molecule and trans Ni-
(2)-I(1) and Ni(2)-I(2) bonds with bond lengths of 3.0655(8)
The 1/1 adducts of cesium and potassium iodides with CoL4
contain high-spin cobalt with magnetic moments of 4.97 and
4.80 B M per cobalt, respectively, whereas all of the other 1/1
adducts with CoL1, CoL2, and CoL3 contain low-spin cobalt.

Adduct Formation
Inorganic Chemistry, Vol. 39, No. 8, 2000 1647
sandwiches generate. This shape is dictated by the alkali metal
ions, demanding that successive nickel salicylaldimine molecules
be approximately related by 180° rotations. Since the columns
in each of the structures have essentially the same shape, they
pack very similarly in each of the lattices. Figure 10 shows how
neighboring columns, which provide the total content of a row
of unit cells stacked parallel to the b axis, are mutually arranged
in the cesium nitrate adduct structure. In the extended structure
this gives a packing arrangement resembling the intermeshing
of cogwheels. Since the nitrate and iodide ions are within the
columnar structures, they clearly have little influence on the
lattice packing. If the thiocyanates and acetates adducts adopt
the infinite sandwich structures, these anions can likewise remain
within the columnar structures.
Reactions of Alkali Metal Tetraphenylborates with ML4.
A necessary condition for efficient lattice packing of infinite
sandwich columns, such as shown in Figure 10, is that the anions
are relatively small and contained within the boundaries of the
columnar sandwich assemblies. This is achieved by ion-pairing
and, in some cases, by the additional complexation of the anion
to the transition metal. The question arises as to what type of
stoichiometry and structures would be preferred if the anions
were too bulky to be accommodated comfortably within the
boundaries of the columns, and with this in mind attention was
focused on tetraphenylborate salts.
NaBPh4 reacted with all of the complexes ML4 to give 1/2
adducts, whereas cesium and potassium tetraphenylborates
reacted when M was Cu, Co, or Ni to give adducts of 2/3
stoichiometry. Reactions with ZnL4 yielded surprising results.
Sodium tetraphenylborate yielded an adduct with the zinc
salicylaldimine, whereas the nitrate and iodide salts failed to
do so. The reverse situation was observed in the case of the
potassium and cesium adducts. Potassium and cesium tetraphen-
ylborates failed to give adducts with this zinc salicylaldimine,
whereas the nitrate and iodide salts readily yielded 1/1 adducts.
The fact that ZnL4 was the only complex that in any case failed
to yield adducts suggests that of all the salicylaldimine
complexes investigated, it is the weakest donor ligand toward
alkali metals.
Figure 8. Powder diffraction patterns of potassium salt adducts of
NiL4.
The latter complexes, in all probability, have iodide ions in
locations similar to those in the nickel analogues. In view of
the absence of water and methanol in the lattices of the high-
spin complexes, it is clear that the iodide ions are located such
that they form Co-I bonds. In view of the normal magnetic
moments of these adducts, the iodide ions clearly do not exhibit
the disorder featured in the structure of KI‚NiL4, thus contribut-
ing to its abnormally low magnetic moment of 2.57 B M per
nickel ion.
Reactions of Potassium Thiocyanate and Acetate with
ML4. Potassium thiocyanate and acetate reacted with ML4 to
give exclusively 1/1 adducts. The potassium acetate adducts with
ML4, where M ) Cu, Co, and Ni, were extremely soluble,
whereas all other adducts precipitated rapidly from solution.
The potassium thiocyanate and acetate adducts of NiL4 each
had magnetic moments of 3.2 B M, whereas the cobalt analogues
were high-spin with magnetic moments of 4.6 and 4.7 B M,
respectively. Since these adducts apparently contain neither
donor solvent molecules nor water, the implication is that the
anions are coordinated to the transition metal. Presumably they
are also involved in electrostatic interactions with the potassium
ions.
The powder diffraction patterns of the potassium thicyanate
and acetate adducts bear a remarkable resemblance, and indeed,
this resemblance extends to diffraction patterns of all of the
nitrate and iodide adducts. The diffraction patterns of the
potassium adducts of NiL4 shown in Figure 8 clearly illustrate
this point where it is seen that the resemblance is particularly
pronounced in the 2θ range 5-25°. A possible reason for the
resemblance of the diffraction patterns can be appreciated by
reference to Figure 9, which shows sections of the infinite
sandwich arrays for the cesium nitrate and potassium iodide
adducts of NiL4 when viewed along the direction of propagation
of the sandwiches. The most striking observation is the similarity
of the cross-sectional shapes of the columns that the infinite
Adduct formation in the present study implies that the adduct
precipitated in preference to starting materials, and thus, all of
the factors that contribute to the lattice energies of starting
materials and adducts have an important bearing on the outcome.
While the strength of electrostatic interaction between the metal
salicylaldimine complex and cation is one of these factors, other
factors such as ion-pairing, donor bond formation between the
anion and the transition metal, and lattice packing are also
important factors. The latter three factors assume increasing
significance as the interaction between the salicylaldimine and
cation weakens. Thus, the inability of either sodium iodide or
nitrate to yield an adduct with ZnL4 when sodium tetraphen-
ylborate did so may well reflect a higher lattice energy asso-
ciated with the packing of Na(ZnL4)2 and tetraphenylborate
groups than the packing of the same complex cation with either
iodide or nitrate. The ability to isolate adducts of cesium and
potassium iodides and nitrates with ZnL4, contrasting with the
inability to do so with either sodium nitrate or sodium iodide,
may reflect a very favorable packing energy associated with
the packing of the infinite sandwich assemblies as described
earlier. Furthermore, the inability of potassium and cesium
tetraphenylborates to yield adducts with ZnL4 may then be
attributed to the fact that this favorable packing is now no longer
possible following the introduction of the bulky tetraphenyl-
borate groups.

1648 Inorganic Chemistry, Vol. 39, No. 8, 2000
Cunningham et al.
Figure 9. Comparison of molecular stacking in the cesium nitrate and potassium iodide adducts [parts a and b, respectively] of NiL4 when viewed
along stacking directions.
Figure 10. Adjacent columns, providing the contents of a row of unit cells stacked along the b axis in the structure of [Cs(NO3)‚NiL4]3‚MeOH.
It seems highly likely that all of the sodium tetraphenylborate
adducts have sodium in the same type of dodecahedral environ-
ment as that described for Na(NiL4)2I‚EtOH‚H2O (in no case
did it prove possible to obtain suitable crystals of a sodium
tetraphenylborate adduct for a crystallographic study). The nickel
adduct is diamagnetic while cobalt in the cobalt adducts is low-
spin, thus indicating that the metals are in square-planar
environments in these adducts, as was nickel in the sodium
iodide adduct.
It did not prove to be possible to obtain crystals of any of
the potassium or cesium tetraphenylborates adducts from meth-
anol, but when acetonitrile was employed as a preparative sol-
vent, adducts precipitation was rapid and, following removal
of these solids, crystals were observed, in some instances, to
grow slowly from the filtrates. In this fashion reasonable quality
crystals of the potassium tetraphenylborate adduct with NiL4
were obtained, one of which was selected for a crystallographic
study.
aldimine molecule, the other half being generated by a crystal-
lographic 2-fold axis passing through nickel. In actual fact, this
2-fold axis generates the full cation (see Figure 11) consisting
of three nickel salicylaldimine molecules and two potassium
ions (with associated acetonitrile molecules) in a club sandwich
type structure, thus accounting for the 2/3 stoichiometry.
Each of the symmetry-related potassium ions is involved in
electrostatic interactions with all four Schiff base oxygens of a
terminal nickel salicylaldimine molecule, these interactions, on
average, being somewhat stronger with the phenolic oxygens
(see Table 8), while it is also involved in interactions with three
oxygens of the central nickel salicylaldimine. On average, the
K-O distances to these central oxygens are greater than those
to the terminal oxygens, an observation that can be rationalized
on the basis of the central oxygens being involved in interactions
with two potassium ions. Furthermore, there may be a degree
of repulsion between the two potassium ions. The coordination
of each potassium is increased to 8 as a result of an interaction
with an acetonitrile molecule.
The asymmetric unit of this adduct consists of a full nickel
salicylaldimine molecule, a potassium ion with an acetonitrile
molecule coordinated to it, and half of a second nickel salicyl-
The three metal salicylaldimine molecules are arranged such
that on each wing of the cation there are three phenyl groups

Adduct Formation
Inorganic Chemistry, Vol. 39, No. 8, 2000 1649
All of the potassium and cesium tetraphenylborate adducts
isolated from acetonitrile, being of the same stoichiometry and
containing acetonitrile (as suggested by both infrared spectros-
copy and analytical data) presumably share the same type of
club sandwich type structure as that described above. The ad-
ducts isolated from methanol show no evidence, from analytical
or infrared data, for either methanol or water molecules that
could adopt the role of acetonitrile in the structure described.
The loss of the coordinating donor molecule probably results
in changes in salicylaldimine conformation and relative orienta-
tions such that the alkali metals are coordinated by eight salicyl-
aldimine oxygens as they are in the infinite sandwich structures.
Apart from the tetraphenylborate adducts, there are two other
adducts that appear to be of 2/3 stoichiometry, these being
adducts formulated as (CsI)2‚(CuL3)3‚2H2O and (NaI)2‚(NiL3)3‚
2H2O (see Experimental Section). The sodium adduct is of
particular interest, since the stoichiometry will almost inevitably
demand the club sandwich structure. Thus, in this instance the
sodium ions are not in the plane of the salicylaldimine oxygens
as they were in the sodium nitrate adduct structures described
earlier. The fact that a metal ion is of sufficient size to be located
in the plane of the Schiff base oxygens does not guarantee that
it will do so. For example, whereas magnesium ion is located
in the plane of the Schiff base oxygens in the 1/1 adduct of
magnesium nitrate with NiL4, it is sandwiched between two sets
of salicylaldimine oxygens in the structure of the 1/2 adduct of
magnesium bromide with NiL4. In this latter structure, electro-
static interactions are with six of the oxygens (three from each
salicylaldimine ligand), and presumably this will be the case in
the sodium adduct of 2/3 stoichiometry.
Figure 11. Triple decker structure of [K(MeCN)]2‚(NiL4)3 (40%
ellipsoids).
Table 8. Selected Bond Lengths (Å) for
[K(MeCN)]2‚(NiL4)3(BPh4)2
K(1)-O(1)
K(1)-O(3)
K(1)-O(5)
K(1)-O(5)′
K(1)-N(4)
Ni(1)-O(2)
Ni(1)-N(2)
Ni(2)-N(3)
2.736(3)
2.746(3)
2.859(3)
2.901(3)
2.981(7)
1.849(3)
1.904(4)
1.872(4)
K(1)-O(2)
K(1)-O(4)
K(1)-O(6)
K(1)-O(6)′
Ni(1)-O(1)
Ni(1)-N(1)
Ni(2)-O(5)
2.663(3)
2.901(4)
2.894(3)
3.750(4)
1.862(3)
1.860(4)
1.851(3)
Acknowledgment. Two of us, N. N´ı C. and M. O’G., are
thankful to Enterprise Ireland for financial support.
Supporting Information Available: Eight crystallographic files,
in CIF format, are available. This material is available free of charge
via the Internet at http://pubs.acs.org.
arranged with their planes being close to parallel while there
are a number of C-C and C-N contacts between neighboring
aromatic and azomethine carbon and nitrogen atoms lying in
the range 3.417-3.0.9 Å. These features suggests the possibility
of net attractive π interactions.24
IC990496P
(24) Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112, 5525-
5534.
(23) McArdle, P.; Cunningham, D. J. Appl. Crystallogr., in press.
(25) Mitchell, M. Ph.D. Thesis, National University of Ireland, 1998.