Synthesis, crystal structures and magnetic characterization of four β-diketonate-alkoxide iron(III) dimers. Dependence of the magnetic properties on geometrical and electronic parameters

Article abstract of DOI:10.1016/S0020-1693(97)05482-0

Synthesis, crystallographic characterization and magnetic properties of the dinuclear iron(III) complexes [Fe(OMe)(dbm)₂]₂ (1), [Fe(OMe)(dpm)₂]₂ (2), [Fe(OEt)(bpm)₂]₂ (3) and [Fe(OProp)(npm)₂]₂ (4) are reported. Complexes 1 and 2 have symmetric β-diketonate ligands dbm (dibenzoylmethanate) and dpm (dipivaloylmethanate), respectively, whereas complexes 3 and 4 contain asymmetric β-diketonate ligands bpm (benzoylpivaloylmethanate) and npm (naphthoylpivaloylmethanate), respectively. Complex 1 crystallizes in the triclinic system, space group P1 (No. 2), a=9.634(1), b=10.946(2), c=13.079(1) A, α=79.95(1), β=88.01(1), γ=82.57(1)°, Z=1. Complex 2 crystallizes in the triclinic system, space group P1 (No. 2), a=10.980(2), b=14.255(2), c=17.979(1) A, α=85.70(1), β=89.63(1), γ=71.65(1)°, Z=2. Compound 3 crystallizes in the monoclinic system, space group P2₁/c (No. 14), a=11.546(2), b=18.539(1), c=13.595(2) A, β=113.18(1)°, Z=2. Compound 4 crystallizes in the monoclinic system, P2₁/c (No. 14) space group, a=13.746(2), b=18.933(2), c=14.158(2) A, β=117.37(1)°, Z=2. Each complex consists of two iron(III) ions that are symmetrically bridged by two alkoxide groups. The antiferromagnetic couplings between the S=5/2 iron centers have been fitted by using the Hamiltonian H=JS₁·S₂ to give values of J=15.4(1) cm^-1 with g=1.96(1) for 1, J=19.0(6) cm^-1 with g=1.98(1) for 2, J=14.8(5) cm^-1 with g=1.98(1) for 3 and J=18.0(5) with g=2 for 4. The structural and magnetic parameters of 1-4 and other iron(III)-alkoxo clusters synthesized in our laboratory are compared. No simple correlation exists between the isotropic exchange-coupling constant J and the average Fe-O(bridge) bond distance, whereas an approximately linear correlation is found between the J value and the Fe-O-Fe angle. The observed trend and the predictions of Extended Hueckel calculations on the model [Fe(OH)H₄]₂^4- are compared and discussed.

Full text of DOI:10.1016/S0020-1693(97)05482-0

Inorganica Chimica Acta 262 (1997) 123–132
Synthesis, crystal structures and magnetic characterization of four
b-diketonate-alkoxide iron(III) dimers. Dependence of the magnetic
properties on geometrical and electronic parameters
Florence Le Gall a, Fabrizia Fabrizi de Biani b, Andrea Caneschi c, Patrizia Cinelli c,
U
Andrea Cornia d, Antonio C. Fabretti d, Dante Gatteschi c,
a Department of Chemistry, Universite´ de Rennes, Rennes, France
b Dipartimento di Chimica, Universita` degli Studi di Siena, Siena, Italy
c Dipartimento di Chimica, Universita` degli Studi di Firenze, Via Maragliano 75, 50144 Florence, Italy
d Dipartimento di Chimica, Universita` degli Studi di Modena, Modena, Italy
Received 20 August 1996; revised 5 November 1996; accepted 2 December 1996
Abstract
Synthesis, crystallographic characterization and magnetic properties of the dinuclear iron(III) complexes [Fe(OMe)(dbm)2]2 (1),
[Fe(OMe)(dpm)2]2 (2), [Fe(OEt)(bpm)2]2 (3) and [Fe(OProp)(npm)2]2 (4) are reported. Complexes 1 and 2 have symmetric b-
diketonate ligands dbm (dibenzoylmethanate) and dpm (dipivaloylmethanate), respectively, whereas complexes 3 and 4 contain asymmetric
b-diketonate ligands bpm (benzoylpivaloylmethanate) and npm (naphthoylpivaloylmethanate), respectively. Complex 1 crystallizes in the
#
˚
triclinic system, space group P1 (No. 2), as9.634(1), bs10.946(2), cs13.079(1) A, as79.95(1), bs88.01(1), gs82.57(1)8, Zs1.
#
˚
Complex 2 crystallizes in the triclinic system, space group P1 (No. 2), as10.980(2), bs14.255(2), cs17.979(1) A, as85.70(1),
bs89.63(1), gs71.65(1)8, Zs2. Compound 3 crystallizes in the monoclinic system, space group P21/c (No. 14), as11.546(2),
˚
bs18.539(1), cs13.595(2) A, bs113.18(1)8, Zs2. Compound 4 crystallizes in the monoclinic system, P21/c (No. 14) space group,
˚
as13.746(2), bs18.933(2), cs14.158(2) A, bs117.37(1)8, Zs2. Each complex consists of two iron(III) ions that are symmetrically
bridged by two alkoxide groups. The antiferromagnetic couplings between the Ss5/2 iron centers have been fitted by using the Hamiltonian
HsJS1ØS2 to give values of Js15.4(1) cmy1 with gs1.96(1) for 1, Js19.0(6) cmy1 with gs1.98(1) for 2, Js14.8(5) cmy1 with
gs1.98(1) for 3 and Js18.0(5) with gs2 for 4. The structural and magnetic parameters of 1–4 and other iron(III)–alkoxo clusters
synthesized in our laboratory are compared. No simple correlation exists between the isotropic exchange-coupling constant J and the average
Fe–O(bridge) bond distance, whereas an approximately linear correlation is found between the J value and the Fe–O–Fe angle. The observed
trend and the predictions of Extended Hu¨ckel calculations on the model [Fe(OH)H4]24y are compared and discussed.
Keywords: Crystal structures; Magnetism; Iron complexes; Alkoxide complexes; Dimeric complexes
1. Introduction
tion is therefore an essential step for a correct approach to
larger systems.
We herein present the synthesis and the structural and
magnetic characterization of diiron(III) di(alkoxo)-bridged
complexes in which b-diketonates act as ancillary ligands.
The reported complexes have been isolated in the course of
an extensive investigation of iron(III) solvolytic routes con-
trolled by b-diketonates [3–5] and have general formula
[Fe(OR)(L)2]2 where RsMe, Et, n-Pr and HLsHdbm,
Hdpm, Hbpm, Hnpm 1. This class of compounds was studied
Alkoxide ligands can be successfully used to build up
polynuclear metal complexes displaying unusual structures
[1,2]. In particular, iron(III)–alkoxo clusters [3–9] are of
special interest to material and bioinorganic chemistry [10]
because they are naturally related to soluble hydroxo-bridged
complexes which act as precursors in the formation of iron
oxides [11,12]. Low nuclearity complexes, such as Fe2 and
Fe3 [4], are likely to represent the ‘molecular bricks’ for the
formation of high-nuclearity clusters, and their characteriza-
1 Abbreviations used in the text: Hdbmsdibenzoylmethane; Hdpms
dipivaloylmethane; Hbpmsbenzoylpivaloylmethane; Hnpmsnaphthoyl-
pivaloylmethane; H3salmps2-bis(salicylideneamino)methylphenol; Hpm-
dbms1,3-di(p-methoxyphenyl)-1,3-propandione.
U Corresponding author. Tel.: q39 55-354 841/2/3; fax: q39 55-354
845.
0020-1693/97/$17.00 q 1997 Elsevier Science S.A. All rights reserved
PII S0020-1693(97)05482-0

124
F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
by Gray and co-workers more than twenty years ago, but no
X-ray structure determination was performed and only rough
estimatesoftheexchange-couplingconstantsJwereprovided
[13]. Accurate magnetic measurements down to 3 K have
now evidenced a possible correlation between the J values
and the Fe–O–Fe angle in the reported dimers and in related
Fe4 and Fe6 alkoxo-bridged complexes.
3, C56H70O10Fe2: C, 66.27; H, 6.95; Fe, 11.01. Found C,
65.89; H, 6.82; Fe, 10.85%.
2.1.5. Step B [Fe(OPr)(npm)2]2 (4)
The deep-red solid obtained in Step A with HLsHnpm
was dissolved in a 1:1 (vol./vol.) mixture of CHCl3 and n-
PrOH. Orange crystals of 4 were obtained by slow diffusion
of n-PrOH vapors. Anal. Calc. for 4, C74H82O10Fe2: C, 71.5;
H, 6.65; Fe, 8.99. Found C, 70.44; H, 6.35; Fe, 8.78%.
2. Experimental
2.2. Physical measurements
2.1. Synthetic procedures
The magnetic susceptibility of polycrystallinesampleswas
measured by using a Me´tronique Ingenie´rie MSO3 SQUID
magnetometer in the temperature range 3–300 K with an
applied field of 1 T. Diamagnetic corrections were estimated
from Pascal’s constants. The magnetic susceptibilities were
corrected for the contribution of the sample holder, which
was measured separately in the same temperature range and
field. An isotropic spin-Hamiltonian defined as HsJS1ØS2
(S1sS2s5/2) was used to fit the experimental data. J and
g were treated as adjustable parameters. Corrections for tem-
perature independent paramagnetism were not applied.
The starting materials FeCl3, Hdbm and Hdpm (Aldrich)
were used as received. Methanol (Fluka) was dried and dis-
tilled over Mg/I2 under a dinitrogen atmosphere shortly
before use. Ethanol and n-propanol (Carlo Erba) were used
without further purification. Kopecki’s reaction between the
appropriate methylester and methylketone in the presence of
sodium hydride was used to prepare the ligands Hbpm and
Hnpm [14,15]. A two-step procedure was adopted which
follows with minor variations that described by Gray and co-
workers for the synthesis of [Fe(OMe)(dpm)2]2 [13]. Step
A is identical for all the reported compounds.
2.3. X-ray crystallography
2.1.1. Step A
X-ray diffraction studies were performed with an Enraf-
Nonius CAD4 diffractometer and graphite monochromatized
radiation. The selected crystal of compound 1 was mounted
on the top of a glass fiber, whereas a large individual of 2
was sealed in a quartz capillary; both crystals were collected
at room temperature; crystals of 3 and 4 were bathed in a
stream of cold dinitrogen as soon as they were removed from
their mother solution, were mounted on a quartz fiber with
silicone grease and collected at 210 and 160 K, respectively.
Collected intensities for complexes 1 and 2 were corrected
for absorption following the measurements of C scans. Initial
heavy atom positions were determined by using the SIR92
program [16] and structures were refined by using the
SHELX76 program [17]. Final models were obtained by
least-squares refinement in combination with differenceFou-
rier syntheses. Hydrogen atoms were fixed at a C–H distance
A 2.37 N solution of MeONa in methanol was prepared by
careful addition of metallic Na (2.18g)to40ml ofanhydrous
methanol. Anhydrous FeCl3 (0.162 g, 1 mmol) was dis-
solved in 10 ml of MeOH. The resulting solution was added
dropwise to a solution obtained by dissolving HL (1 mmol)
in 10 ml of MeOH and 2 ml of the MeONa solution (4.74
mmol MeONa). The yellow-to-red precipitate was isolated
by filtration, washed with methanol and dried under vacuum.
2.1.2. Step B for [Fe(OMe)(dbm)2]2 (1)
The orange–yellow solid obtained in Step A with
HLsHdbm wasdissolvedinchloroform. Thenmethanolwas
carefully layered over the filtered solution and allowed to
stand at room temperature, yielding red crystals of 1 in 2–3
days. Anal. Calc. for 1, C62H50O10Fe2: C, 69.80; H, 4.72; Fe,
10.47. Found: C, 69.42; H, 4.65; Fe, 10.23%.
˚
of 0.95 A and were assigned isotropic thermal parameters
Bs1.2Beq(C). All non-hydrogen atoms except for the car-
bon atoms of the alkyl chains in 3 and 4 were refined with
anisotropic thermal parameters. The alkyl chainsof thebridg-
ing propoxide ligands in 4 were subject to major disorder
effects and were modeled over two positions: Cx1, Cx2, Cx3,
and Cxa, Cxb, Cxc with occupancy factors 0.48 and 0.52,
respectively. O–C distances of the propoxide groups were
2.1.3. Step B for [Fe(OMe)(dpm)2]2 (2)
The bright yellow solid obtained in Step A with
HLsHdpm was dissolved in chloroform. Large orange crys-
tals of 2 were obtained by slow diffusion of methanol vapors.
Anal. Calc. for 2, C46H82O10Fe2: C, 60.94; H, 9.37; Fe, 12.32.
Found C, 60.51; H, 9.12; Fe, 12.21%.
˚
fixed at 1.42 A. Though a somewhat high thermal parameter
2.1.4. Step B for [Fe(OEt)(bpm)2]2 (3)
resulted for the carbon atom Cx2, the model converged rea-
sonably well. The highest residual peak (0.8 e Ay3) was
˚
The deep-red solid obtained in Step A with HLsHbpm
was dissolved in a 1:1 (vol./vol.) mixture of CHCl3 and
EtOH. Large orange crystals of 3 were obtained either by
EtOH vapor diffusion or by slow evaporation. Anal. Calc. for
located near Cx2. Crystal data and experimental parameters
are reported in Table 1 while atomic coordinates are reported
in Tables 2–5.

F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
125
Table 1
Crystal data and experimental parameters for [Fe(OMe)(dbm)2]2 (1), [Fe(OMe)(dpm)2]2 (2), [Fe(OEt)(bpm)2]2 (3) and [Fe(OPr)(npm)2]2 (4)
1
2
3
4
Formula
Fe2C62H50O10
1066.77
Fe2C46H82O10
906.86
Fe2C56H70O10
1014.86
Fe2C74H82O10
1243.15
FW (g moly1
)
Crystal system
triclinic
P1 (No. 2)
triclinic
P1 (No. 2)
monoclinic
P21/c (No. 14)
11.546(2)
18.539(1)
13.595(2)
monoclinic
P21/c (No. 14)
13.746(2)
18.933(2)
14.158(2)
Space group a
#
#
˚
a (A)
9.634(1)
10.946(2)
13.079(1)
79.95(1)
88.01(1)
82.57(1)
1347(3)
1
10.980(2)
14.255(2)
17.979(1)
85.70(1)
89.63(1)
71.65(1)
2663.2(6)
2
˚
b (A)
˚
c (A)
a (8)
b (8)
113.18(1)
117.37(1)
g (8)
3
˚
V (A )
2675.2(2)
2
1.260
3272.2(2)
2
1.262
Z
D
calc (g cmy3
)
1.315
1.131
T (K)
l (A)
Scan type
298
0.71069
u–2u
298
1.5418
u–2u
210
0.71069
u–2u
160
0.71069
u–2u
˚
Data collection range (8)
Data limits
No. data collected
Rav
4F2uF52
"h, "k, ql
5509
4F2uF120
"h, "k, ql
6193
5F2uF50
5F2uF44
"h, qk, ql
4369
"h, qk, ql,
3949
0.066
0.020
0.029
0.061
No. unique data
4296
4970
2557
3437
No. obs. unique data
No. parameters
2250 with Fo)9s(Fo)
335
6.7
5.5
0.95/1.00
0.048
0.047
0.001
0.29
2927 with Fo)9s(Fo)
454
6.4
4.7
0.72/1.00
0.058
0.063
0.001
0.50
2143 with Fo)16s(Fo)
2727 with Fo)4s(Fo)
309
6.9
5.5
388
7.0
4.6
Data/parameter ratio
Absorption coefficient (cmy1
)
Transmission coefficient, min./max. b
R c
Rw
0.071
0.072
0.001
0.54
0.096
0.097
0.001
0.61
Largest shift/e.s.d., final
Largest residual peak (e A
y3
˚
)
a Ref. [44].
b An empirical absorption correction based on c scan was applied.
2
c Rs8NNFoNyNFcNN/8NFoN and Rws[8w(NFoNyNFcN)2/8wNFoN ]1/2 where ws1/s2(F).
Table 2
Table 2 (continued)
Final atomic positional parameters a of [Fe(OMe)(dbm)2]2 (1)
Atom
x/a
y/b
z/c
Atom
x/a
y/b
z/c
C(13)
C(14)
C(15)
C(16)
C(17)
C(18)
C(19)
C(20)
C(21)
C(22)
C(23)
C(24)
C(25)
C(26)
C(27)
C(28)
C(29)
C(30)
C(31)
0.5392(8)
0.6537(8)
0.6368(7)
0.5056(7)
0.2556(6)
0.2131(6)
0.0967(6)
0.3757(6)
0.4046(7)
0.5150(8)
0.5991(8)
0.5714(7)
0.4598(7)
0.0510(6)
y0.0234(7)
y0.0720(9)
y0.0499(8)
0.0211(8)
0.0728(7)
0.1626(7)
0.1783(7)
0.4430(6)
0.3804(6)
0.2796(6)
0.2422(5)
y0.0319(4)
0.0340(4)
0.0995(5)
y0.1061(4)
y0.1501(5)
y0.2203(6)
y0.2451(5)
y0.2025(5)
y0.1323(5)
0.1631(5)
0.2545(5)
0.3133(7)
0.2822(7)
0.1910(6)
0.1312(5)
Fe(1)
O(1)
O(2)
O(3)
O(4)
O(5)
C(1)
C(2)
C(3)
C(4)
C(5)
C(6)
C(7)
C(8)
C(9)
C(10)
C(11)
C(12)
0.0719(9)
0.3837(7)
0.0718(6)
0.1015(4)
y0.0333(4)
0.2417(4)
0.2000(4)
0.0219(4)
0.2338(6)
y0.0001(6)
0.1332(6)
0.5570(3)
0.4256(3)
0.3490(4)
0.3166(3)
0.2112(3)
0.5965(5)
0.3936(5)
0.3358(5)
0.3180(5)
0.4265(5)
0.5131(5)
0.5490(6)
0.4971(7)
0.4082(7)
0.3744(6)
0.2675(5)
0.2077(6)
0.0106(3)
0.1989(3)
0.1621(3)
y0.0353(3)
0.1119(3)
y0.0138(5)
0.2937(4)
0.3258(4)
0.2608(4)
0.3698(4)
0.3355(5)
0.4067(5)
0.5100(5)
0.5441(5)
0.4745(5)
0.3038(4)
0.4056(5)
0.2374(7)
0.2797(6)
0.2031(5)
0.1009(5)
0.1079(5)
0.1876(5)
0.0799(6)
0.0700(7)
0.1607(8)
0.2666(8)
0.2806(6)
y0.0052(5)
0.0050(6)
y0.0976(7)
y0.2124(7)
y0.2239(6)
y0.1227(5)
0.2477(6)
y0.1107(6)
y0.2243(6)
y0.3259(7)
y0.3158(7)
y0.2058(7)
y0.1028(6)
0.3883(6)
0.4058(7)
(continued)
a Numbers in parentheses are errors in the last significant digit.

126
F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
Table 3
Table 4
Final atomic positional parameters a for [Fe(OEt)(bpm)2]2 (3)
Final atomic positional parameters a for [Fe(OMe)(dpm)2]2 (2)
Atom
x/a
0.4500(1)
y/b
0.0147(8)
0.0533(4)
y0.0302(3)
y0.0383(4)
0.0969(4)
z/c
0.3796(1)
Atom
x/a
y/b
z/c
Fe(1)
O(1)
O(2)
O(3)
O(4)
O(5)
C(1)
C(2)
C(3)
C(4)
C(5)
C(6)
C(7)
C(8)
C(9)
C(10)
C(11)
C(12)
C(13)
C(14)
C(15)
C(16)
C(17)
C(18)
C(19)
C(20)
C(21)
C(22)
C(23)
C(24)
C(25)
C(26)
C(27)
C(28)
0.5703(5)
0.5939(5)
0.3393(6)
0.4694(5)
0.2981(5)
0.5938(9)
0.4810(8)
0.3619(1)
0.4004(1)
0.2891(9)
0.2482(9)
0.7189(8)
0.8181(9)
0.9378(1)
0.9533(1)
0.8552(1)
0.7372(1)
0.1334(1)
0.1282(1)
0.0253(1)
y0.0720(1)
y0.0715(1)
0.0344(1)
0.2469(1)
0.1724(1)
0.1622(1)
0.2760(1)
0.4410(9)
0.5028(2)
0.3559(1)
0.5535(1)
0.6898(9)
0.6961(1)
0.5181(6)
0.3529(7)
0.2496(7)
0.2912(6)
0.3775(6)
0.2618(9)
0.1700(9)
0.1677(1)
0.2497(1)
0.2643(1)
0.3314(1)
0.2599(1)
0.3187(1)
0.3160(1)
0.2521(1)
0.1920(1)
0.1922(1)
0.3440(1)
0.4423(1)
0.4677(1)
0.3930(1)
0.2931(1)
0.2706(1)
0.0629(1)
0.0458(1)
0.0754(1)
y0.0338(1)
0.1834(1)
0.1225(2)
0.1143(1)
0.2653(1)
0.5303(1)
0.5691(1)
Fe(1)
Fe(2)
O(1)
O(2)
O(3)
O(4)
O(5)
O(6)
O(7)
O(8)
O(9)
O(10)
C(1)
C(2)
C(3)
C(4)
C(5)
C(6)
0.5963(1)
y0.09478(9)
0.09244(9)
0.0286(4)
y0.1501(4)
y0.1891(4)
y0.2027(4)
y0.0327(4)
y0.0294(4)
0.1982(4)
0.0291(4)
0.1445(4)
0.1898(4)
0.0781(7)
y0.2328(7)
y0.2953(7)
y0.2713(7)
y0.2567(8)
y0.256(1)
y0.178(1)
y0.3572(9)
y0.3449(7)
y0.4509(8)
y0.337(1)
y0.315(1)
y0.2274(7)
y0.1697(7)
y0.0759(7)
y0.3282(7)
y0.4058(9)
y0.343(1)
y0.333(1)
y0.0157(7)
y0.0767(9)
0.0553(9)
0.049(1)
y0.02862(8)
0.53101(8)
0.0187(3)
y0.0197(3)
0.0603(3)
y0.0880(3)
y0.1278(3)
0.4817(3)
0.5915(4)
0.6295(3)
0.5220(3)
0.4422(3)
0.0093(5)
0.0110(5)
0.0566(5)
0.0789(5)
y0.0076(6)
y0.0906(7)
0.021(1)
0.0216(1)
0.5887(5)
0.7878(5)
0.5965(5)
0.5741(6)
0.6275(5)
0.1116(5)
y0.0886(6)
0.1005(5)
0.1700(5)
y0.0520(6)
0.6816(8)
0.8575(8)
0.8104(9)
0.6822(9)
0.9985(9)
1.014(1)
1.047(1)
1.083(1)
0.640(1)
0.718(1)
0.507(1)
0.640(2)
0.6152(9)
0.6682(8)
0.6707(8)
0.609(1)
0.0680(4)
y0.0487(5)
y0.0573(5)
y0.0504(6)
0.1511(5)
0.1654(5)
0.1255(6)
y0.0610(5)
y0.0185(6)
y0.0279(8)
y0.0802(8)
y0.1254(7)
y0.1144(6)
0.1485(6)
0.1331(6)
0.1529(7)
0.1900(8)
0.2037(7)
0.1843(6)
y0.0614(6)
0.0129(7)
y0.1189(7)
y0.0746(8)
0.2015(5)
C(7)
C(8)
C(9)
0.0287(8)
0.1324(6)
0.1279(9)
0.114(1)
C(10)
C(11)
C(12)
C(13)
C(14)
C(15)
C(16)
C(17)
C(18)
C(19)
C(20)
C(21)
C(22)
C(23)
C(24)
C(25)
C(26)
C(27)
C(28)
C(29)
C(30)
C(31)
C(32)
C(33)
C(34)
C(35)
C(36)
C(37)
C(38)
C(39)
C(40)
C(41)
C(42)
C(43)
C(44)
C(45)
C(46)
0.2076(8)
y0.1513(6)
y0.1989(5)
y0.1871(5)
y0.1715(6)
y0.1268(9)
y0.2538(7)
y0.1477(9)
y0.2448(6)
y0.2939(6)
y0.2885(7)
y0.2028(7)
0.4898(5)
0.6532(6)
0.7002(6)
0.6867(5)
0.6759(7)
0.648(1)
0.1586(8)
0.2537(8)
0.2468(9)
0.0889(6)
0.704(2)
0.618(2)
0.476(2)
0.7225(9)
0.823(1)
0.1613(7)
0.614(1)
0.785(1)
a Numbers in parentheses are errors in the last significant digit.
0.2455(8)
y0.0656(9)
0.0356(9)
0.1129(9)
y0.151(1)
y0.286(2)
y0.162(2)
y0.106(2)
0.2164(9)
0.158(1)
y0.0796(7)
0.2245(7)
0.1673(7)
0.0703(7)
0.3249(7)
0.341(1)
2.4. Calculations
Extended Hu¨ckel (EH) [18] calculations were performed
on the model molecule [Fe(OH)H4]24y in an idealized D2h
symmetry (see inset in Fig. 8). All the bond distances and
angles were fixed to the average of the experimental values
apart from the Fe–O–Fe bond angle (u) which was varied
from 115 to 908 along the calculation. The orbital parameters
are reported in Table 6. A more complex model like
[Fe(OH)(H2O)4]24q was used in trial calculations explor-
ing a few values of u. The trend in the orbital energies was
found consistent with that obtained for [Fe(OH)H4]24y, so
justifying the use of the simpler model.
0.334(1)
0.7561(9)
0.637(1)
0.4003(9)
0.0092(7)
y0.0574(8)
y0.0562(9)
0.0726(8)
0.2683(7)
0.2871(7)
0.2267(7)
0.3391(7)
0.4415(9)
0.311(1)
0.353(2)
0.2464(7)
0.246(1)
0.343(1)
0.165(1)
0.7443(5)
0.7926(6)
0.7025(6)
0.7934(6)
0.4214(5)
0.4426(5)
0.4898(5)
0.3657(6)
0.388(1)
0.3277(9)
0.265(1)
y0.0248(9)
0.0845(9)
0.1789(9)
y0.120(1)
y0.165(2)
y0.236(2)
y0.064(2)
0.302(1)
0.355(2)
0.2923(8)
0.5058(6)
0.5897(8)
0.4722(8)
0.479(1)
3. Results and discussion
0.317(1)
0.311(1)
0.411(1)
3.1. Synthesis
The reaction of an iron(III) salt with Hdpm and a base in
methanol was used many years ago by Gray and co-workers
a Numbers in parentheses are errors in the last significant digit.

F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
127
Table 5
b-diketone (HL) and 4–5 equiv. of sodium methoxide in
anhydrous methanol, a yellow-to-red precipitate, which ana-
lyzes approximately as Fe(OMe)2(L), is formed. The solid
is quite soluble in moderately polar organic solvents such as
CHCl3 and CH2Cl2, though insoluble materials are quickly
formed when it is exposed to wet air. Slow diffusion of ROH
vapors or spontaneous evaporation of a CHCl3 or CHCl3/
ROH solution lead to dimer complexes 2–4 for RsMe, Et,
n-Pr and HLsHdpm, Hbpm, Hnpm, respectively. It is to be
noted that the formation of 3 and 4 involves ligand substitu-
tion: methoxide ligands which presumably are present in the
starting material are replaced by ethoxide and n-propoxide
species, respectively. Complex1isformedbyliquiddiffusion
of methanol in a CHCl3 solution of the appropriate solid
obtained in Step A. On the contrary spontaneous evaporation
of CHCl3/MeOH solutions of the same solid leads to the
hexairon(III) complex [NaFe6(OMe)12(dbm)6]q [3a].By
simply playing with crystallization conditions in Step B we
have in fact uncovered a completely new pool of molecular
iron(III)–alkoxo clusters [3–5]. Interestingly dbm ligands
have allowed the isolation of Fe2, Fe3 [4], Fe4 [19], Fe6 [3]
and Fe10 [5] species from CHCl3/MeOH mixtures whereas
dpm, bpm and npm have so far afforded Fe2 complexes only.
The combined nature of the solvent and of the R9 and R0
groups of the b-diketonates (defined as R9–C(O)–CH–
C(O)–R0) may therefore play a fundamental role in the
aggregation processes.
Final atomic positional parameters a of [Fe(OPr)(npm)2]2 (4)
Atom
x/a
0.4575(1)
y/b
0.01731(9)
z/c
0.3808(1)
Fe(1)
O(1)
O(2)
O(3)
O(4)
O(5)
C(1)
C(2)
C(3)
C(6)
C(5)
C(4)
C(7)
C(15)
C(25)
C(10)
C(22)
C(20)
C(21)
C(18)
C(19)
C(23)
C(24)
C(31)
C(32)
C(27)
C(17)
C(30)
C(28)
C(29)
C(33)
C(34)
C(26)
C(11)
C(16)
C(9)
0.3556(6)
0.4745(6)
0.3243(6)
0.5802(5)
0.5574(6)
0.410(1)
0.3086(10)
0.2754(9)
0.374(1)
0.4804(10)
0.5757(10)
0.2676(10)
0.1705(9)
0.6839(10)
0.294(1)
0.0200(9)
0.117(1)
y0.115(1)
0.0714(10)
y0.022(1)
y0.016(1)
0.077(1)
0.996(1)
0.883(2)
0.815(1)
0.0742(9)
1.014(2)
0.840(2)
0.938(2)
0.863(2)
0.763(1)
0.704(1)
0.4479(9)
0.1675(9)
0.190(1)
0.554(1)
0.460(1)
0.2222(10)
0.363(1)
0.570(3)
0.680(4)
0.731(3)
0.664(1)
0.698(3)
0.619(3)
y0.0335(4)
0.0992(4)
0.0693(4)
y0.0262(4)
0.0530(4)
0.1452(7)
0.1576(6)
0.1205(7)
y0.0501(6)
y0.0541(6)
y0.0431(6)
y0.0627(6)
0.1412(6)
y0.0459(7)
y0.0730(8)
0.1796(6)
0.1958(7)
0.1984(6)
0.1554(6)
0.1743(7)
0.1844(6)
0.1655(7)
y0.0369(9)
y0.0436(9)
y0.0979(9)
0.1565(6)
y0.086(1)
y0.1435(8)
y0.139(1)
0.0040(8)
y0.0015(9)
y0.0973(8)
0.1914(6)
0.1342(6)
y0.0048(7)
0.2270(9)
0.1460(9)
y0.1314(6)
0.2459(8)
0.1272(7)
0.133(3)
0.2477(6)
0.2979(6)
0.3686(6)
0.3605(6)
0.5222(5)
0.244(1)
0.2416(10)
0.3088(9)
0.169(1)
0.1790(9)
0.271(1)
0.0658(10)
0.3122(10)
0.271(1)
y0.029(1)
0.3228(10)
0.427(1)
0.334(1)
0.513(1)
0.518(1)
0.224(1)
0.219(1)
0.279(2)
0.271(1)
0.214(1)
0.4170(10)
0.218(2)
0.157(1)
0.157(2)
0.334(1)
0.332(1)
0.211(1)
0.1771(9)
0.4060(9)
0.046(1)
0.253(1)
0.097(1)
0.082(1)
0.111(1)
0.546(3)
0.547(6)
0.591(3)
0.545(2)
0.600(2)
0.524(3)
3.2. Structural studies
Figs. 1–4 display ORTEP [20] views of the four com-
plexes, which have similar structures. The asymmetric unit
of 1, 3 and 4 comprises one half-dimer molecule, whereasthe
crystal lattice of 2 is generated by two crystallographically
independent half-dimers. The two iron(III) ions are related
by an inversion center and are symmetrically bridged by two
alkoxide groups. The geometry around the metal atoms is
roughly octahedral, the six coordination sites being occupied
by the oxygen donors of two cis-alkoxide groups and two b-
diketonate ligands. No evidence was found in either structure
for intermolecular contacts between dimer units. The average
Fe–O(bridge) bond distances are equal within the experi-
mental errors (see Table 7) and agree well those reported in
the literature for similar diiron(III) complexes [21–23]. The
C(13)
C(12)
C(8)
C(14)
Cx(1)
Cx(2)
Cx(3)
Cxa
0.204(2)
0.080(1)
0.151(1)
0.203(2)
Cxb
Cxc
a Numbers in parentheses are errors in the last significant digit.
˚
˚
Fe–Fe separations vary from 3.049(3) A for 3 to 3.105 A for
2, while the Fe–O–Fe angles range from 101.8(3)8 for 3 to
103.78 for 2. Selected bond distances and angles for com-
pounds 1–4 are reported in Tables 8–11, respectively.
for the synthesis of [Fe(OMe)(dpm)2]2 [13] 2. However,
a number of dimer complexes with general formula
[Fe(OR)(L)2]2 (RsMe, Et, n-Pr; HLsHdbm, Hdpm,
Hbpm, Hnpm) can be easily prepared by similar procedures.
When iron(III) chloride is allowed to react with 1 equiv. of
Complexes 1–4 are structurally related to a few iron(III)
clusters that contain Fe2(OR)2 units, namely [Fe4(OMe)6-
(dpm)6] (5) [19], [NaFe6(OMe)12(dbm)6]ClPCHCl3P
12MeOH (6) [3a] and [NaFe6(OMe)12(pmdbm)6]ClO4-
xCHCl3yMeOH (7) [3b]. Their structures are sketched in
Fig. 5. In the tetrairon(III) complex 5 double methoxide
bridges connect the central iron(III) ion to the peripheral
ones, which are arranged at the vertices of an approximately
2 The same authors described alternative synthetic procedures which
involve the reaction of sodium alkoxides with tris(b-diketonate) complexes
or oxidation of iron(II) chloride in the appropriate alcohol in the presence
of HL. It has recently been demonstrated that under strictly anaerobic con-
ditions the latter reaction leads to the cubane tetrairon(II) clusters
[Fe4(OMe)4(L)4(HOMe)4] (HLsHdbm, Hdpm) [2a].

128
F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
Table 6
Valence ionization potential, coefficient and exponents of the atomic Slater functions used in the EH calculations
Atom
Fe
Orbital
Hii
c1
z1
c2
z2
4s
4p
3d
y9.17
1.0000
1.0000
0.5366
1.900
1.900
5.350
y5.37
y12.70
0.6678
1.800
O
H
2s
2p
y32.30
y14.80
1.0000
1.0000
2.275 a
2.275
1s
y13.60
1.0000
1.300
a The z coefficient was altered to 3.050 in the calculation (see text).
Fig. 1. ORTEP view of the complex [Fe(OMe)(dbm)2]2 (1).
Fig. 2. ORTEP view of the complex [Fe(OMe)(dpm)2]2 (2).
Fig. 3. ORTEP view of the complex [Fe(OEt)(bpm)2]2 (3).
Fig. 4. ORTEP view of the complex [Fe(OPr)(npm)2]2 (4).
equilateral triangle and are further coordinated by two b-
diketonate ligands (Fig. 5(b)). Hexairon(III) clusters 6 and
7 are isostructural: the metal atoms form a six-memberedring
and are connected by double methoxide bridges. In this case
only one b-diketonate ligand is found on each metal ion,
Fig. 5(c). Interestingly, both clusters encapsulate a sodium
ion, which is coordinated by the oxygen atoms of six meth-
oxide ligands.
3.3. Magnetic properties
The temperature dependence of the molar magnetic sus-
ceptibility of complexes 1–4 is typical of antiferromagneti-
cally-coupled systems with Ss0 in the ground state. Fig. 6
shows the x versus T plot for compound 3, but a very similar
behavior was observed for compounds 1 and 2.

F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
129
Table 7
Structural and magnetic parameters for 1–7
Compound
Fe–O(bridge) c
(A)
Fe∆Fe
(A)
Fe–O–Fe
(8)
J
g
(cmy1
)
˚
˚
[Fe(OMe)(dbm)2]2 (1)
[Fe(OMe)(dpm)2]2 (2)
[Fe(OEt)(bpm)2]2 (3)
1.987
1.974
1.965
1.974
1.960
2.015
2.020
3.087(1)
3.105 c
102.0(2)
103.7 c
15.4(1)
19.0(6)
14.8(5)
18.0(5) a
21.5(3)
20.0(3)
19.9(1)
1.96(1)
1.98(1)
1.98(1)
2.00 a,b
1.98(1)
2.00 b
3.049(3)
3.093(3)
3.133 c
101.8(3)
103.1(5)
105.8 c
[Fe(OPr)(npm)2]2 (4)
[Fe4(OMe)6(dpm)6] (5)
[NaFe6(OMe)12(dbm)6]q (6)
[NaFe6(OMe)12(pmdbm)6]q (7)
3.195 c
104.8 c
3.215(5)
105.6 c
2.00 b
a Values obtained by fitting only the high temperature data.
b The value was kept fixed.
c Averaged value.
Table 8
Table 10
Selected bond lengths (A) and angles (8) for [Fe(OEt)(bpm)2]2 (3)
a
a
˚
˚
Selected bond lengths (A) and angles (8) for [Fe(OMe)(dbm)2]2 (1)
O1–Fe1
O2–Fe1
O3–Fe1
Fe19∆Fe1
1.978(3)
2.007(4)
2.007(4)
3.087(1)
O4–Fe1
O5–Fe1
O19–Fe1
2.007(3)
1.986(4)
1.995(3)
O1–Fe1
O2–Fe1
O3–Fe1
Fe1∆Fe19
1.971(9)
2.015(1)
1.986(9)
3.049(3)
O4–Fe1
O5–Fe1
O19–Fe1
2.007(9)
2.005(1)
1.959(8)
Fe1–O1–Fe19
O1–Fe1–O19
O1–Fe1–O2
O1–Fe1–O3
O1–Fe1–O4
O1–Fe1–O5
O2–Fe1–O3
O2–Fe1–O4
102.0(2)
80.0(2)
96.5(1)
95.8(2)
90.8(1)
170.1(2)
85.5(1)
168.7(1)
O2–Fe1–O5
O2–Fe1–O19
O3–Fe1–O4
O3–Fe1–O5
O3–Fe1–O19
O4–Fe1–O5
O4–Fe1–O19
O5–Fe1–O19
87.7(2)
89.2(1)
85.2(2)
93.5(2)
171.4(2)
86.5(1)
100.8(1)
93.1(1)
Fe1–O1–Fe19
O1–Fe1–O19
O1–Fe1–O3
O2–Fe1–O4
O1–Fe1–O4
O2–Fe1–O3
101.8(3)
78.8(9)
171.5(2)
84.9(6)
95.4(5)
85.6(6)
O2–Fe1–O5
O1–Fe1–O2
O3–Fe1–O4
O3–Fe1–O5
O1–Fe1–O5
O4–Fe1–O5
168.7(5)
90.0(1)
91.6(1)
88.7(3)
96.8(5)
85.4(5)
a Numbers in parentheses are errors in the last significant digit.
a Numbers in parentheses are errors in the last significant digit.
Table 11
a
˚
Selected bond lengths (A) and angles (8) for [Fe(OPr)(npm)2]2 (4)
Table 9
Selected bond lengths (A) and angles (8) for [Fe(OMe)(dpm)2]2 (2)
a
˚
Fe1–O1
Fe1–O2
Fe1–O3
Fe1∆Fe19
2.001(1)
2.025(1)
2.020(1)
3.093(3)
Fe1–O4
Fe1–O5
Fe1–O59
2.011(1)
1.958(1)
1.989(1)
O1–Fe1
O2–Fe1
O3–Fe1
Fe1∆Fe19
1.988(5)
2.002(6)
2.010(5)
3.103(2)
O4–Fe1
O5–Fe1
O19–Fe1
2.013(6)
2.007(5)
1.965(5)
Fe1–O5–Fe19
O1–Fe1–O2
O1–Fe1–O3
O1–Fe1–O4
O1–Fe1–O5
O2–Fe1–O4
103.1(5)
96.6(5)
95.7(6)
92.9(6)
170.8(5)
84.3(2)
O2–Fe1–O5
O3–Fe1–O4
O3–Fe1–O5
O2–Fe1–O3
O4–Fe1–O5
92.4(4)
167.3(5)
86.7(1)
85.4(5)
86.2(6)
Fe1–O1–Fe19
O1–Fe1–O1
O1–Fe1–O3
O2–Fe1–O4
O1–Fe1–O4
O2–Fe1–O3
O19–Fe1–O2
O19–Fe1–O1
O19–Fe1–O5
103.7(5)
76.6(3)
102.2(2)
95.7(3)
168.8(2)
85.1(2)
168.5(3)
76.6(3)
100.7(2)
O2–Fe1–O5
O1–Fe1–O2
O3–Fe1–O4
O3–Fe1–O5
O1–Fe1–O5
O4–Fe1–O5
O19–Fe1–O4
O19–Fe1–O3
85.3(2)
93.8(2)
84.6(2)
165.0(2)
90.0(2)
84.9(2)
94.6(2)
90.8(2)
a Numbers in parentheses are errors in the last significant digit.
through a narrow maximum at 30 K. The presence of mag-
netically dilute ferric impurities was detected at the lowest
temperatures. The value Js18.0(5) cmy1 reported in
Table 7 was obtained by a fit on the high-temperature data
with gs2.00. Though we have at present no definitive expla-
nation for the magnetic behavior of 4, we suggest that a
structural phase transition may occur.
Experimental and theoretical work [24–28] on alkoxo-
and hydroxo-bridged copper(II) and chromium(III) dimers
has demonstrated that the substituents R on the bridging
ligands OR have a major influence on magnetic coupling, the
greater the electron withdrawing power of R the smaller the
value of J. However, J may be influenced by non-bridging
a Numbers in parentheses are errors in the last significant digit. The geo-
metrical parameters of the molecule containing Fe2 are equal within the
experimental error.
The results of the least-squares fitting of the experimental
susceptibility data and selected geometrical parametersofthe
Fe2(OR)2 units are reported in Table 7 for complexes 1–4.
Complex 4 shows a rather unusual x versus T curve for an
iron(III) dimer. The susceptibilityincreasesoncoolingdown
from room temperature, seems to approach a broadmaximum
around 100 K but then increases rather rapidly and passes

130
F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
in the past [29–43], the trend of the J values in weakly
coupled systems is still largely unexplained. If one defines
the‘couplingdistance’Pashalfoftheshortestsuperexchange
pathway connecting two metal centers A and B, JAB shows
an approximate inverse-exponential dependence upon P
[29]. Though it is widely accepted that P has a major influ-
ence on the JAB value, plots of JAB against P show a greater
scattering of the experimental points when the coupling is
small. It has therefore been recognized that when P is large
more subtle electronic effects may become important [29].
The classical work on copper(II) and chromium(III) oxo-
bridged dimers has inspired a number of investigations on
the possible dependence of JAB upon the Fe–O–Fe angle
[34,35]. In m2-O bridged non-heme diferric complexes, for
which strong couplings are in general observed(Js160–240
cmy1), a decrease in the Fe–O–Fe angle leads to a small but
significant decrease of J, as pointed out by Kurtz [33]. Some
experimental results indicate that the Fe–O–Fe angle may
play a role in weakly-coupled systems as well, though a
general correlation has not been found. For instance, the fer-
romagnetic exchange interaction detected in the compound
[Fe2(salmp)2]P2DMF has been attributed to an unusually
acute angle at the bridging OPh ligands (f978) [32]. More
recently surprisingly large antiferromagnetic interactions
mediated by m6-O bridges have been detected in a few
hexairon(III) clusters (Js19.0–24.5 cmy1 with Ps
2.25(1) and Fe–O–Fes1808) [7,8].
Fig. 5. Sketched structures of dimeric complexes 1–4 (a), and of tetrameric
5 (b) and hexameric 6, 7 (c) clusters. The Fe2(OR)2 units are highlighted.
Solid arcs represent the b-diketonate ligands.
The stronger couplings found in 5–7 with respect to 1–4
cannot be due to different Fe–O(bridge) distances, since
these compounds exhibit larger Fe–O(bridge) separations
than 1–4 (see Table 7). Indeed a plot of J versus Fe–O-
(bridge) distances does not reveal any particular trend. On
the contrary a nice linear correlation was found between J
and the Fe–O–Fe angle u (Fig. 7), an increase of the latter
leading to a larger J value. The best-fit straight line through
the experimental points is
Fig. 6. x vs. T plot for [Fe(OEt)(bpm)2]2 (3). The solid line represents
the best fit curve.
ligands as well. Inspection of Table 7 shows that the anti-
ferromagnetic interactions in complexes 1 (RsMe;
R9sR0sPh) and 2 (RsMe; R9sR0st-Bu) are very sim-
ilar to each other. Therefore, the influence of R9 and R0 on
the value of J is probably small. However, J for complex 3
is significantly larger in spite of the fact that R9st-Bu and
R0sPh. This is probably due to the different bridgingspecies
in 3 (RsEt) and is in agreement with the stronger electron
donating power of Et with respect to Me [24]. A somewhat
similar trend is apparent from the rough estimates of J which
were reported by Gray and co-workers for relatedcompounds
[13].
Js1.48uy135
Complexes 5–7 are structurally similar to 1–4 in that they
contain alkoxide and b-diketonate ligands only, and the
exchange interactions are invariably mediated by
bis(alkoxide) bridges. Inspection of Table 7 shows that they
display larger Fe–Fe couplings than 1–4 with J values in the
range 19.9–21.5 cmy1 (Table 7). In order to rationalize this
observation, we tried to correlate the experimental J values
with the geometrical parameters of the Fe2(OR)2 unit as
described in Section 3.4.
3.4. Magneto-structural correlations
Though several attempts to find magneto-structural corre-
lations in iron(III) polynuclear compounds have been made
Fig. 7. Plot of the experimental J values vs. the Fe–O–Fe angle u. The best
fit line Js1.48uy135 is also drawn.

F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
131
According to the above equation, the change of sign of J
(i.e. the crossover from antiferro- to ferromagnetic interac-
tion) occurs for us918, in good agreement with the predic-
tions of a simple angular-overlap model.
not support the experimental dependence of J upon u. How-
ever, it is well known that the predictions of EH calculations
are dramatically influenced by the orbital parameters used. In
order to check this point, we repeated the calculation by
contracting the 2s(O) orbital, so reducing its contribution to
the MOs. For this purpose the z coefficient of the 2s(O) Slater
orbital was fixed at 3.05. As previously reported for cop-
per(II) dimers [27], the different angular dependence of the
d(Fe)–s(O) and d(Fe)–p(O) overlaps may in fact not have a
negligible influence on the trend of JAF. Along with a general
lowering of the 8iDEi2 quantity, we found a minimum in the
8iDEi2 versus u curve at uf1048 (see Fig. 8(b)). It can
therefore be anticipated that by contracting further or even
eliminating the 2s(O) orbital we can shift the minimum to
smaller angles. The model would therefore correctly predict
larger JAF values for increasing u. As already mentioned the
same behavior was found in similar Cu(II) dimers [27]; the
authors also needed to contract the 2s(O) orbital to better
reproduce the experimental trend. The origin of such a behav-
ior was explained in terms of the different possibilities of
interaction between symmetric and antisymmetric combina-
tions of the metal orbitals and the s and p orbitals of the
bridging oxygens, but as a matter of fact no theoretical jus-
tification for the necessity to contract the 2s(O) orbital was
given. It seems that this has to be interpreted as a way to
compensate for other weaknesses of the EH approach. It is
evident that the results of the EH calculation are strongly
influenced by the orbital parameter used; this, together with
the presence of ten magnetic orbitals and therefore a large
number of magnetic exchange pathways and with a scarce-
ness of homogeneus experimental data, makes any analysis
done with such a simple theoretical model difficult .
3.5. Extended Hu¨ckel calculations
As shown in Section 3.4, larger Fe–O–Fe angles are asso-
ciated with stronger antiferromagnetic interactions in our
series of compounds. Since for small angles (and short
Fe∆Fe separations) direct interactions between metal orbi-
tals are expected to enhance antiferromagnetic coupling, the
observed trend suggests that indirect interactions predomi-
nate. In order to theoretically support these observations
we performed EH calculations on the model complex
4y
[Fe(OH)H4]2 for various values of the Fe–O–Fe angle
(u).
According to the theoretical model developed by Hay et
al. [27] the antiferromagnetic contribution JAF to the cou-
pling constant can be studied in terms of pairwiseinteractions
of MOs, holding the relationship JAFA8iDEi2, where DEi is
the energy separation between the symmetric and antisym-
metric combinations of couples of magnetic orbitals. There-
fore the influence of geometrical distortions on JAF in a series
of closely related compounds can be studied by determining
the corresponding variation of the quantities DEi. Surpris-
ingly, our calculations predict an increase of 8iDEi2, and
hence of JAF, with decreasing Fe–O–Fe angle (Fig. 8(a)).
This trend is opposite to the experimental one and suggests
prevalent direct interactions. In order to support this conclu-
sion, we performed
a calculation on the fragment
[FeH4]22y, for which we assumed the same geometry as in
the hydroxo-bridged complex. As shown in Fig. 8(c), the
quantity 8iDEi2 increases quite rapidly as the two iron atoms
approach each other. Hence, at this stage our calculations do
4. Conclusions
Four diiron(III) complexes 1–4 have been prepared by
controlled alcoholysis of iron(III) chloride in the presence
of b-diketonates and characterized by single-crystal X-ray
diffraction methods. Attention has been focussed on the
geometry of the Fe2(OR)2 moiety (RsMe, Et, n-Pr) in the
solid state and on its influence on the strength of antiferro-
magnetic coupling. A linear dependence has been found
between the J value and the Fe–O–Fe angle in 1–4 and in
three other iron(III) polynuclear compounds containing
alkoxide and b-diketonate ligands. On the other hand, EH
calculations applied to the model developed by Hay et al.
[27] do not permit the observed trend to be fully supported.
The results of the EH calculation are in fact strongly influ-
enced by the orbital parameters used. This aspect, together
with the presence of ten magnetic orbitals and of a large
number of magnetic-exchange pathways, makes any analysis
done with such a simple theoretical model difficult. Finally,
it has to be noted that our EH calculations explored quite a
large range of u values, while the experimental data fall in a
small range. The variation in the estimated strength of the
2
4y
Fig. 8. Plot of 8iDEi vs. the Fe–O–Fe angle u for the [Fe(OH)H4]2
model dimer calculated at the EH level by using the orbital parameters
reported in Table 6 (a) or with a higher exponent zs3.05 for the 2s(O)
orbital (b). The curve (c) shows the results obtained for the model
[FeH4]22y, without the bridging groups. The inset shows a picture of the
di(hydroxo)-bridged model with the bond distances (A) and reference
system used in the calculations.
˚

132
F. Le Gall et al. / Inorganica Chimica Acta 262 (1997) 123–132
[10] D. Gatteschi, A. Caneschi, R. Sessoli and A. Cornia, Chem. Soc. Rev.,
(1996) 101.
[11] W. Schneider, Chimia, 42 (1988) 9.
[12] W. Schneider, Comments Inorg. Chem., 3 (1984) 205.
[13] C.-H.S. Wu, G.R. Rossman, H.B. Gray, G.S. Hammond and H.J.
Schugar, Inorg. Chem., 11 (1972) 990.
[14] S.N. Poelsma, A.H. Servante, F.P. Fanizzi and P.M. Maitlis, Liq.
Cryst., 16 (1994) 675.
coupling is small as well and could eventually be influenced
by other factors. However the homogeneity of the samples,
both with respect to the nature of the ligands and to their
structural features, supports the thesis of the dependence of
the J value on the angle at the bridge in this class of weakly-
coupled oxo-bridged polyiron(III) complexes. To our
knowledge this represents the first study where such a cor-
relation is well documented by experimental results.
[15] K.R. Kopecki, D. Nonhebel, G. Morris and G.S. Hammond, J. Org.
Chem., 27 (1962) 1036.
[16] A. Altomare, M.C. Burla, G. Camalli, G. Cascarano, C. Giacovazzo,
A. Guagliardi and G. Polidori, J. Appl. Crystallogr., 27 (1994) 435.
[17] G.M. Sheldrick, SHELX76, program for crystal structure
determination, University of Cambridge, Cambridge, UK, 1976.
[18] CACAO Package, C. Mealli and D. M Proserpio, J. Chem. Educ., 67
(1990) 399.
[19] A. Caneschi, A. Cornia, A.C. Fabretti, D. Gatteschi, F. Fabrizi de Biani
and R. Sessoli, unpublished results.
[20] C.K. Johnson, ORTEP, Rep. ORNL-3794, Oak Ridge National
Laboratory, Oak Ridge, TN, 1965.
[21] J.A. Bertrand, J.L. Breece and P.G. Eller, Inorg. Chem., 13 (1974)
125.
[22] J.A. Thich, C. Chih Ou, D. Powers, B. Vasiliou, D. Mastropaolo, J.A.
Potenza and H.J. Schugar, J. Am. Chem. Soc., 98 (1976) 1425.
[23] B. Chiari, O. Piovesana, T. Tarantelli and P.F. Zanazzi, Inorg. Chem.,
23 (1984) 3398.
5. Supplementary material
Tables of anisotropic thermal parameters for 1, 2, 3, 4,
SIa–d (6 pages); coordinates of hydrogen atoms for 1, 2, 3,
4, SIIa–d (4 pages); bond distances and angles for 1, 2, 3, 4,
SIIIa–d (8 pages); structure factors for 1, 2, 3, 4, SIVa–d
(28 pages); Fig. SI, plot of x versus T for compound 4
(1 page) are available from the authors on request.
Acknowledgements
[24] E. Dixon Estes, R.P. Scaringe, W.E. Hatfield and D.J. Hodgson, Inorg.
Chem., 16 (1977) 1179, 1605.
[25] H.E. LeMay, Jr., D.J. Hodgson, P. Pruettiangkura and L.J. Theriot, J.
Chem. Soc., Dalton Trans., (1979) 781.
[26] V.H. Crawford, H.W. Richardson, J.R. Wasson, D.J. Hodgson and
W.E. Hatfield, Inorg. Chem., 15 (1976) 2107.
This work was supported by Ministero dellaRicercaScien-
tifica e Tecnologica (MURST), by ConsiglioNazionaledelle
Ricerche (CNR) and by the Human Capital and Mobility
Program through Contract ERBCHRX-CT20080.
J.P. Hay, J.C. Thibeault and R. Hoffmann, J. Am. Chem. Soc., 97
(1975) 4884.
[27]
R.D. Willett, in R.D. Willet, D. Gatteschi and O. Kahn (eds.),
Magneto-Structural Correlations in Exchange Coupled Systems,
NATO ASI Series C, Vol. 140, 1985, p. 389.
[28]
References
S.M. Gorun and S.J. Lippard, Inorg. Chem., 30 (1991) 1625.
J.R. Hart, A.K. Rappe´, S.M. Gorun and T.H. Upton, Inorg. Chem., 31
(1992) 5254.
R.N. Mukherjee, T.D.P. Stack and R.H. Holm, J. Am. Chem. Soc., 110
(1988) 1850.
B.S. Snyder, G.S. Patterson, A.J. Abrahamson and R.H. Holm, J. Am.
Chem. Soc., 111 (1989) 5214.
D.M. Kurtz, Jr., Chem. Rev., 90 (1990) 585.
K. Takahashi, Y. Nishida, Y. Maeda and S. Kida, J. Chem. Soc., Dalton
Trans., (1985) 2375.
C.C. Ou, R.G. Wollmann, D.N. Hendrickson, J.A. Potenza and H.J.
Schugar, J. Am. Chem. Soc., 100 (1978) 4717.
C.C. Ou, R.A. Lalancette, J.A. Potenza and H.J. Schugar, J. Am. Chem.
Soc., 100 (1978) 2053.
J.A. Bertrand, J.L. Breece, A.R. Kalyanaraman, G.J. Longand and
W.A. Baker, Jr., J. Am. Chem. Soc., 91 (1970) 5233.
R.G. Wollmann and D.N. Hendrickson, Inorg. Chem., 17 (1978) 926.
B. Chiari, O. Piovesana, T. Tarantelli and P.F. Zanazzi, Inorg. Chem.,
21 (1982) 1396.
S. Menage and L. Que, Jr., Inorg. Chem., 29 (1990) 4293.
S.J. Barclay, P.E. Riley and K.N. Raymond, Inorg. Chem., 23 (1984)
2005.
[29]
[30]
[1] (a) M.I. Khan, Q. Chen, H. Ho¨pe, S. Parkin, C.J. O’Connor and J.
Zubieta, Inorg. Chem., 32 (1993) 2929; (b) M. Cavaluzzo, Q. Chen
and J. Zubieta, J. Chem. Soc., Chem. Commun., (1993) 131.
[2] (a) K.L. Taft, A. Caneschi, L.E. Pence, C.D. Delfs, G.C.
Papaefthymiou and S.J. Lippard, J. Am. Chem. Soc., 115 (1993)
11753; (b) A. Caneschi, A. Cornia, S.J. Lippard, G.C. Papaefthymiou
and R. Sessoli, Inorg. Chim. Acta, 243 (1996) 295; (c) K.L. Taft,
G.C. Papaefthymiou and S.J. Lippard, Inorg. Chem., 33 (1994) 1510.
[3] (a) A. Caneschi, A. Cornia and S.J. Lippard, Angew. Chem., Int. Ed.
Engl., 34 (1995) 467; (b) A. Caneschi, A. Cornia, A.C. Fabretti, S.
Foner, D. Gatteschi, R. Grandi and L. Schenetti, Chem. Eur. J., 2
(1996) 1379.
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[4] A. Caneschi, A. Cornia, A.C. Fabretti, D. Gatteschi and W. Malavasi,
Inorg. Chem., 34 (1995) 4660.
[5] A. Caneschi, A. Cornia, A.C. Fabretti and D. Gatteschi, Angew. Chem.,
Int. Ed. Engl., 34 (1995) 2716.
[38]
[39]
[6] V.S. Nair and K.S. Hagen, Inorg. Chem., 31 (1992) 4048.
[7] (a) K. Hegetschweiler, H.W. Schmalle, H.M. Streit and W. Schneider,
Inorg. Chem., 29 (1990) 3625; (b) K. Hegetschweiler, H.W.
Schmalle, H.M. Streit, V. Gramlich, H.-U. Hund and I. Erni, Inorg.
Chem., 31 (1992) 1299.
[8] (a) A. Cornia, D. Gatteschi and K. Hegetschweiler, Inorg. Chem., 33
(1994) 1559; (b) A. Cornia, D. Gatteschi, K. Hegetschweiler, L.
Hausherr-Primo and V. Gramlich, Inorg. Chem., 35 (1996) 4414.
[9] K.L. Taft, C.D. Delfs, G.C. Papaefthymiou, S. Foner, D. Gatteschi and
S.J. Lippard, J. Am. Chem. Soc., 116 (1994) 821.
[40]
[41]
B. Chiari, O. Piovesana, T. Tarantelli and P.F. Zanazzi, Inorg. Chem.,
21 (1982) 2444.
G.D. Fallon, A. Markiewicz, K.S. Murray and T. Quach, J. Chem. Soc.,
Chem. Commun., (1991) 198.
T. Hahn (ed.), International Tables for X-ray Crystallography,Reidel,
Dordrecht, Netherlands, 1983.
[42]
[43]
[44]