Nature of metal binding sites in Cu(II) complexes with histidine and related N-coordinating ligands, as studied by EXAFS

Article abstract of DOI:10.1021/ic049699q

Knowledge of the complexes formed by N-coordinating ligands and Cu(II) ions is of relevance in understanding the interactions of this ion with biomolecules. Within this framework, we investigated Cu(II) complexation with mono-and polydentate ligands, such as ammonia, ethylenediamine (en), and phthalocyanine (Pc). The obtained Cu-N coordination distances were 2.02 A for [Cu(NH₃)₄]²⁺, 2.01 A for [Cu(en) ₂]²⁺, and 1.95 A for CuPc. The shorter bond distance found for CuPc is attributed to the macrocyclic effect. In addition to the structure of the first shell, information on higher coordination shells of the chelate ligands could be extracted by EXAFS, thus allowing discrimination among the different coordination modes. This was possible due to the geometry of the complexes, where the absorbing Cu atoms are coplanar with the four N atoms forming the first coordination shell of the complex. For this reason multiple scattering contributions become relevant, thus allowing determination of higher shells. This knowledge has been used to gain information about the structure of the 1:2 complexes formed by Cu(II) ions with the amino acids histidine and glycine, both showing a high affinity for Cu(II) ions. The in-solution structure of these complexes, particularly that with histidine, is not clear yet, probably due to the various possible coordination modes. In this case the square-planar arrangements glycine-histamine and histamine-histamine as well as tetrahedral coordination modes have been considered. The obtained first-shell Cu-N coordination distance for this complex is 1.99 A. The results of the higher shells EXAFS analysis point to the fact that the predominant coordination mode is the so-called histamine-histamine one in which both histidine molecules coordinate Cu(II) cations through N atoms from the amino group and from the imidazole ring.

Full text of DOI:10.1021/ic049699q

Inorg. Chem. 2004, 43, 6674−6683
Nature of Metal Binding Sites in Cu(II) Complexes with Histidine and
Related N-Coordinating Ligands, As Studied by EXAFS
Flora Carrera,† Enrique Sa´nchez Marcos,‡ Patrick J. Merkling,‡ Jesu´s Chaboy,§ and
Adela Mun˜oz-Pa´ez*,†
Departamentos de Quimica Inorganica y Quimica Fisica, ICMSE, UniVersidad de SeVilla,
CSIC, 41012 SeVilla, Spain and Instituto de Ciencia de Materiales de Aragon, CSIC,
UniVersidad de Zaragoza, 50009 Zaragoza, Spain
Received March 8, 2004
Knowledge of the complexes formed by N-coordinating ligands and Cu(II) ions is of relevance in understanding the
interactions of this ion with biomolecules. Within this framework, we investigated Cu(II) complexation with mono-
and polydentate ligands, such as ammonia, ethylenediamine (en), and phthalocyanine (Pc). The obtained Cu−N
coordination distances were 2.02 Å for [Cu(NH3)4]2+, 2.01 Å for [Cu(en)2]2+, and 1.95 Å for CuPc. The shorter bond
distance found for CuPc is attributed to the macrocyclic effect. In addition to the structure of the first shell, information
on higher coordination shells of the chelate ligands could be extracted by EXAFS, thus allowing discrimination
among the different coordination modes. This was possible due to the geometry of the complexes, where the
absorbing Cu atoms are coplanar with the four N atoms forming the first coordination shell of the complex. For this
reason multiple scattering contributions become relevant, thus allowing determination of higher shells. This knowledge
has been used to gain information about the structure of the 1:2 complexes formed by Cu(II) ions with the amino
acids histidine and glycine, both showing a high affinity for Cu(II) ions. The in-solution structure of these complexes,
particularly that with histidine, is not clear yet, probably due to the various possible coordination modes. In this
case the square-planar arrangements glycine-histamine and histamine-histamine as well as tetrahedral coordination
modes have been considered. The obtained first-shell Cu−N coordination distance for this complex is 1.99 Å. The
results of the higher shells EXAFS analysis point to the fact that the predominant coordination mode is the so-
called histamine-histamine one in which both histidine molecules coordinate Cu(II) cations through N atoms from
the amino group and from the imidazole ring.
Introduction
of the naturally occurring amino acids forming the proteins
may form stable five-membered chelate rings with metal ions,
where the donor groups are amino Nam and carboxylate OCOO
groups. Lately, several works have been devoted to the study
of normal and abnormal interactions that might take place
between transition metals and proteins, which eventually
could be at the heart of some neurodegenerative diseases,
such as Parkinson’s, Alzheimer’s, or Creutzfeldt-Jacob’s
disease.3,4 Copper is the third most abundant transition-metal
element in the human body, following iron and zinc, mainly
occurring in the divalent oxidation state. It has a relevant
role in redox processes and is present in brain and liver
tissues. In the interaction with proteins, in addition to the
Metal ions play a vital role in a vast number of widely
differing biological processes.1 The biochemical function of
the metal ion in all processes is a matter of fundamental
importance. Most metal cations in living organisms interact
with proteins; thus, knowledge of metal-protein binding sites
has attracted the attention of many research groups.2 Each
* Corresponding author. E-mail: adela@us.es.
† Departamento de Quimica Fisica Inorganica, Universidad de Sevilla.
‡ Departamento de Quimica Fisica, Universidad de Sevilla.
§ Universidad de Zaragoza.
(1) (a) Hughes, M. N. In The Inorganic Chemistry of Biological Processes;
J. Wiley: New York, 1972; Chapter 1, p 3. (b) Trautwein, A. X.
Bioinorganic Chemistry: Transition Metals in Biology and their
Coordination Chemistry; Deutsche Forschungsgemeinschaft Weinheim
Wiley-VCH: Weinheim, 1997.
(2) Wilkins, P. C.; Wilkins, R. G. Inorganic Chemistry in Biology; Oxford
Scientific Publishing: Oxford, 1997.
(3) Bush, A. I. Curr. Opin. Chem. Biol. 2000, 4, 184-191.
(4) Stockel, J.; Safar, J.; Wallace, A. C.; Cohen, F. E.; Prusiner, S.
Biochemistry 1998, 37, 7185-7193.
6674 Inorganic Chemistry, Vol. 43, No. 21, 2004
10.1021/ic049699q CCC: $27.50
© 2004 American Chemical Society
Published on Web 09/24/2004

Nature of Metal Binding Sites in Cu(II) Complexes
Scheme 1. Equilibrium Structures Proposed for Cu(His)2 Complexes at Neutral pH9a
coordination with Nam and OCOO, Cu(II) shows a remarkable
affinity for nitrogen atoms of imidazole rings, Nim, probably
related with the high stability of the complexes this metal
forms with N-coordinating ligands.1,5 Due to the presence
of the imidazole ring, histidine is the amino acid showing
the highest affinity for Cu(II) ions (log â2 ) 18.3), and
consequently, knowledge of the structure of the complexes
that it forms with Cu(II) ions is of relevance in understanding
Cu(II)-protein interactions. However, although these com-
plexes have been extensively studied, they are far from being
fully understood.
modes of coordination being proposed, and in all of them
Cu(II) is bound to histidine residues.14
The first fully solved structure of a Cu-His complex,
reported by Evertsson in 1969 from an XRD study of a
crystal of the nitrate of bis-histidinecopper(II) dihydrate
prepared at pH ) 3.7,7 shows that each histidine molecule
behaves as a bidentate ligand coordinating Cu(II) ions
through the Nam atom (at 1.98 and 1.99 Å for each molecule)
and through the OCOO atom at (1.93 and 1.99 Å). This
coordination mode is called “glycine-type”, since it is the
most stable one in Cu(II)-glycine complexes. Further studies
of 1:2 Cu(II):histidine complexes8 showed that at neutral/
basic pH each histidine molecule coordinated through the
Nim and Nam atoms in the so-called ‘histamine-histamine’
mode, and at pH ) 8.1 there was an equilibrium between
this and the mixed “glycine-histamine” mode, the last
coordination mode being predominant (see Scheme 1).9a This
coordination mode is shown by the [Cu(His)(Asn)] complex,
as solved by XRD by Ono et al.9b Further studies of the Cu-
(His)2 complex using CD,6,10,11a UV-vis, EPR and ENDOR
spectroscopies11b reached similar conclusions. In contrast,
Ozutsumi et al.,12 who studied a series of complexes formed
between Cu(II) and one or two amino acid molecules with
EXAFS, including histidine, propose the formation of
aminocarboxylate complexes, i.e., “glycine-glycine” coor-
dination type, although the complex was prepared at neutral
pH, with coordination distances equal to 1.98 Å for Cu-N
and 2.02 Å for Cu-O.
Through these three coordination sites, Nam, Nim, and OCOO
,
histidine might act either as a monodentate or bidentate
ligand, and some authors have proposed simultaneous
coordination to the same cation through the three binding
sites.6 The type of coordination is determined by several
factors, the most important being the ligand-to-metal ratio,
which determines the stoichiometry of the complex, and the
pH of the solution, because it changes the protonation degree
1,5
of the coordinating atoms Nam, Nim, and OCOO
.
The
relevance of the knowledge of the metal binding sites in these
complexes cannot be overemphasized, and thus, this subject
has attracted the attention of many research groups who have
used a wealth of techniques to determine their structures.6-12
To name but one example, indirect evidence exists attributing
the native cellular prion protein, PrPC, a role in copper
metabolism. This protein is proposed to play a role in bovine
spongiform encephalopathy and Creuzelft-Jacob disease in
humans.13 Cu(II) is said to bind the N-terminus region
particularly rich in histidine and glycine residues with several
Another key point to understand Cu(II)-metalloprotein
interactions is knowledge of the complexes it forms with
the amino acid glycine, which has a high affinity for Cu(II)
ions as well, forming stable complexes with metal:glycine
ratios of 1:1, 1:2, or 1:3. Each glycine molecule can act as
a monodentate or bidentate ligand, coordinating through the
Nam atom and/or OCOO atoms. In the 1:2 complexes, glycine
usually acts as a bidentate ligand, coordinating through both
atoms in the so-called “glycine-type” mode, referred to
above.15 The most stable isomer is the one in which the two
(5) Gazo, J.; Bersuker, I. B.; Garaj, J.; Kabesova, M.; Kohot, J.;
Langfelderova, H.; Melnik, M.; Serator, M.; Valach, F. Coord. Chem.
ReV. 1976, 19, 253-297.
(6) Casella, L.; Gulloti, M. J. Inorg. Biochem. 1983, 18, 19-31.
(7) Evertsson, B. Acta Crystallogr. 1969, B25, 30-41.
(8) Valensin, G.; Basosi, R.; Antholine, W. E.; Gagelli. E. J. Inorg.
Biochem. 1985, 23, 125-130.
(9) (a) Sigel, H.; McCormick, B. D. J. Am. Chem. Soc. 1971, 89, 2041-
2044. (b) Ono, T.; Shimanouchi, H.; Sasada, Y.; Sakurai, T.;
Yamaguchi, O.; Nakahara, A. Bull Chem. Soc. Jpn. 1979, 52, 2229-
2234.
(10) Ramonelli, M.; Basosi, R.; Chem. Phys. Lett. 1988, 143, 404-408.
(11) (a) Grommen, R.; Manikandan, P.; Gao, Y.; Shane, T.; Shane, J. J.;
Schoonheydt, R. A.; Weckhuysen, B. M.; Goldfarb, D. J. Am. Chem.
Soc. 2000, 122, 11488-11496. (b) Manikandan, P.; Epel, B.; Goldfarb,
D. Inorg. Chem. 2001, 40, 781-787.
(13) Aronoff-Spencer, E.; Burns, C. S.; Avdievich, N. J.; Gerfen, G. J.;
Peisach, J.; Antholine, W. E.; Ball, H. L.; Cohen, F. E.; Prusiner, S.
B.; Milhauser, G. L. Biochemistry 2000, 40, 13760-13771.
(14) Hasnain, S. S.; Murphy, L. M.; Strange, R. W.; Grossmann, J. G.;
Clarke, A. R.; Jackson, G. S.; Collinge J. J. Mol. Biol., 2001, 311,
467-473.
(12) Ozutsumi, K.; Miyata, Y.; Kawashima, T. J. Inorg. Biochem. 1991,
44, 97.
(15) Condrate, R. A.; Nakamoto, K. J. Chem. Phys. 1965, 42, 2590-2598.
Inorganic Chemistry, Vol. 43, No. 21, 2004 6675

Carrera et al.
carboxylic groups occupy contiguous positions (cis coordina-
tion), so that the first coordination shell is formed by two
oxygen and two nitrogen atoms. The crystal structure of this
complex has been solved,16 yielding 1.98 and 2.02 Å for
Cu-N coordination distances and 1.95 and 1.96 for Cu-O
coordination distances. Very similar values have been
obtained from an EXAFS study of this complex in aqueous
solution.17 From the point of view of the technique, this
complex might serve to check if EXAFS spectroscopy is able
to solve two contributions at similar distances where the
backscatters are consecutive atoms in the periodic table, i.e.,
N and O atoms.
Although all these studies provide relevant information
concerning Cu(II)-histidine/glycine interactions, only two
provide values for the metal-ligand coordination distances
in solution. The other works are for crystalline compounds,
where compact packing is said to be governed by minimum
van der Waals plus hydrogen-bond interaction energy, which
might not be the same in solution than in the solid state.
Moreover, it is of relevance to obtain structural information
in a medium as similar as possible to that of the physiological
fluids, i.e., low concentrations and neutral pH since, as
explained above, these factors determine metal-ligand
binding sites. To fill this gap we have undertaken a study
aimed at obtaining direct structural information about the
complexes formed by Cu(II) cations and amino acids
histidine and glycine for 1:2 metal to amino acid ratio, bis-
histidinecopper(II), [Cu(His)2], and bis-glycinatocopper(II),
[Cu(Gly)2], in aqueous solutions at physiological pH ) 7.3
and low concentrations (0.01-0.05 M). As a first step we
studied Cu(II) model complexes with mono-, bis-, and
tetradentate N-coordinating ligands of known structures
which are closely related with those proposed for [Cu(His)2]
and [Cu(Gly)2]. The selected complexes are tetrammine-
copper(II), [Cu(NH3)4]2+, bis-ethylenediaminecopper(II),
[Cu(C2N2H8)2]2+, hereafter [Cu(en)2]2+, and phthalocyanine-
copper (II), [Cu(C32N8H16)]), hereafter CuPc.
proposed for CuPc, from electron diffraction and molecular
dynamic studies,26,27 similar values, 1.97-2.02 Å, were
obtained for complexes formed with imidazole and histi-
dine.14,28-30
The requirements concerning concentration and media
drastically reduce the number of experimental techniques that
might be of application to the study of these systems.
Extended X-ray absorption fine structure (EXAFS) is one
of the most appropriate ones since it is element sensitive
and can be applied in a wide range of concentrations to
species showing either short- or long-range order.31 For this
reason it has been applied successfully to the study of
biomolecules containing a metal center, such as metallopro-
teins.32 It has been successfully applied as well to the
structural determination of several other Cu(I) and Cu(II)
complexes in aqueous and nonaqueous solutions.33 Given
the available structural information, the general aim of this
study is determination of the in-solution structure of the
selected complexes, paying special attention to shells beyond
the first one. Moreover, it is of general interest to study the
effect of chelation on coordination distances and Debye-
Waller (DW) factors of these species in media similar to
those formed by the physiological fluids. The detection of
labile species such as coordinated water molecules is of
relevance as well since they might play important roles in
the conformational behavior of the protein.1,2
Experimental and Computational Methods
Sample Preparation. [Cu(NH3)4]2+ was prepared by dissolving
the required amount of Cu(II) perchlorate hexahydrate in 30%
commercial ammonia. Under these conditions ammonia molecules,
as tested by UV-vis spectroscopy, substitute water molecules.18
To prepare the [Cu(en)2]2+ complex in aqueous solution, the
required amount of Cu(ClO4)2 was dissolved in distilled water, and
a 25% ethylenediamine solution was added with shaking. The
complex is stable and highly soluble.34 The CuPc in solution was
obtained by dissolving the corresponding amount of copper(II)
In the model complexes the local coordination around
copper can be considered to be square planar, the first
coordination shell being formed by four N atoms, at ca. 2
Å.18 For instance, Cu-N coordination distances for these
complexes from XRD studies of crystalline salts with several
anions, such as chloride, bromide, sulfate, and selenate,19-22
range from 1.97 to 2.15 Å. These parameters range from
1.93 to 2.03 Å in the data obtained from EXAFS for the
dissolved species.23-25 While a shorter distance, 1.935 Å, is
(25) Valli, M.; Matsuo, S.; Wakita, H.; Yamaguchi, T.; Nomura, M. Inorg.
Chem. 1996, 35, 5642-5645.
(26) Brown C. J. J. Chem. Soc. A 1968, 2488.
(27) (a) Ogawa, M. Y.; Martinsen, J.; Palmer, S. M.; Stanton, J. L.; Tanaka,
J.; Greenee, R. L.; Hoffman, B. M.; Ibers, J. A. J. Am. Chem. Soc.
1987, 109, 1115-1121.(b) Mastryukov, V.; Ruan, C.-Y.; Fink, M.;
Wang, Z.; Pachter, R. J. Mol. Struct. 2000, 556, 225-237.
(28) Strange, R. W.; Blackburn, N. J.; Knowles, P. F.; Hasnain, S. S. J.
Am. Chem. Soc. 1987, 109, 7157-7162.
(29) Boswell, J. S.; Reedy, B. J.; Kulathila, R.; Merkler, D.; Blackburn,
N. J. Biochemistry 1996, 35, 12241-12250.
(30) Kurzdiel, K.; Glowiak, T.; Jezierska, J. Polyhedron 2002, 21, 1857-
1863.
(31) (a) X-ray Absorption: Principles, Applications, Techniques of EXAFS,
SEXAFS and XANES; Koningsberger, D. C., Prins, R., Eds.; Wiley:
New York, 1988. (b) Penner-Hahn J. E. Coord. Chem. ReV. 1999,
190-192, 1101-1123.
(32) Scott, R. A. In Physical Methods in Bioinorganic Chemistry; Que,
L., Jr., Ed.; University Science Books: Sausalito, CA, 2000. (b)
Ascone, I.; Meyer-Klaucke, W.; Murphy, L. J. Synchrotron Rad. 2003,
10, 16-22. (c) Hasnain, S. S.; Strange, R. W. J. Synchrotron Rad.
2003, 10, 9-15.
(16) Freeman, H. C.; Snow, M. R.; Nitta, I.; Tanita, K. Acta Crystallogr.
1964, 17, 1463-1470.
(17) D’Angelo, P.; Botani, E.; Festa, M. R.; Nolting, H.-F.; Pavel, N. V.
J. Phys. Chem. B 1998, 102, 3114-3122.
(18) Cotton, F. A.; Wilkinson, G.; Murillo, C. A.; Bochmann, M. AdVanced
Inorganic Chemistry, 6th ed.; John Wiley & Sons: New York, 1999;
Chapter 1, p 3.
(19) Mazzi, F. Acta Crystallogr. 1955, 8, 137.
(20) Koniyama, Y.; Lingafelter, E. C. Acta Crystallogr. 1964, 17, 1145-
1148.
(21) Distler, T.; Vaughan, P. A. Inorg. Chem. 1967, 6, 126-129.
(22) Morosin, B. Acta Crystallogr. 1969, B25, 19-30.
(23) Fujita, T.; Ohtaki, H. Bull Chem. Soc. Jpn. 1983, 56, 3276-3283.
(24) Sano, M.; Matsuo, T.; Yamatera, H. Bull. Chem. Soc. Jpn. 1983, 56,
3287-3290.
(33) (a) Kau, L.-S.; Spira-Solomon, D. J.; Penner-Hahn, J. E.; Hodgson,
K. O.; Solomon, E. I. J. Am. Chem. Soc. 1987, 109, 6433-6442. (b)
Huang, H.; Liang, C. H.; Penner-Hahn, J. E. Angew. Chem., Int. Ed.
1998, 37, 1564-1566.
(34) Grossman, H.; Schuck, B. Z. Anorg. Chem. 1906, 50, 1.
6676 Inorganic Chemistry, Vol. 43, No. 21, 2004

Nature of Metal Binding Sites in Cu(II) Complexes
Scheme 2. Single Scattering Paths and Most Significant Multiple
Scattering Paths Included in the Fits of the EXAFS Spectra of the Cu(II)
Complexesa
3,4′,4′′,4′′′′-tetrasodiumtetrasulfonate phthalocyanine in distilled
water.27 This derivative of phthalocyanine was used due to its higher
solubility.
To prepare the cis-bis(glycinate)copper(II) monohydrate, a 3.0
g sample of CuSO4‚5H2O was dissolved in 17 mL of HCl, then
1.5 g of glycine was added, and the mixture washeated at 40 °C.
NaHCO3 was added to increase the pH and induce complex
precipitation. The solid was suction filtrated for 5 min to dry it.15
To obtain [Cu(His)2] complex, the required amounts of Cu(II)
perchlorate hexahydrate and histidine were dissolved in water in a
molar ratio of 1:5. The solution was shaken for 1 h, and afterward
the pH was adjusted to 7.3 by adding sodium hydroxide.8
With the exception of the bis(glycinate)copper(II), the complexes
are moderately soluble in water, so medium (0.1 M in [Cu(NH3)4]2+
,
[Cu(en)2]2+, and CuPc, 0.05 M in [Cu(His)2]) and low (0.01 M)
concentration solutions were measured to search for possible effects
of concentration on the complex structure. The reported solubility
of the [Cu(Gly)2] is 0.025 M.15 Thus, only a saturated solution was
measured. In addition to the spectra of the complexes in aqueous
solution, the spectrum of the solid compound CuPc was measured
to check for changes induced by the solvation process. The amount
of solid complex required to obtain the optimum absorption was
mixed with BN to obtain a self-supported wafer.
a The number and type of legs are indicated as well as the amplitudes
compared to that of the first shell for zero Debye-Waller factors, σ2
)
0.0.
X-ray Absorption Measurements and Data Analysis. XAS
spectra of the copper K-edge (8979 eV) were measured at the
synchrotron radiation source ESRF (Grenoble, France) at beamline
BM29. Ring energy was 6 GeV, and ring current was 200 mA.
Energy calibration was carried out with a copper foil. A double-
crystal Si (311) monochromator was used. For this energy and using
0.3 mm vertical slits its estimated resolution is better than 0.5 eV,
which is confirmed by the resolution of a sharp preedge feature in
the XANES spectrum of CuPc. Higher harmonic rejection was
carried out by detuning both crystals 50%. The solid and liquid
sample spectra were recorded at room temperature in transmission
mode. Ion chambers, filled with a He/Ar gas mixture and optimized
to absorb 20% of the beam intensity in I0 and 80% in It, were used
as detectors. Counting time was evenly increased through the
recording of each scan, and several scans were averaged to improve
the signal-to-noise ratio.
The EXAFS functions ø(k) were obtained from the X-ray
absorption spectra by subtracting a Victoreen curve followed by a
cubic spline background removal using the program AUTOBKX35
and normalizing to the height of the measured absorption edge. E0
was defined as the energy where the absorption coefficient is one-
half of the height of the atomic absorption jump. The EXAFS
spectra thus obtained were analyzed using FEFF 8.1036 and FEFFIT
2.54,37 as detailed below.
the fitting range, were taken into account. Only those paths whose
maximum effective distance is equal to or less than 5.0 Å were
considered.
The first coordination shell, formed in all complexes by four N
atoms in a square-planar arrangement (except in [Cu(Gly)2], where
this shell is formed by two N atoms and two O atoms), gives rise
to the most intense contribution to the EXAFS spectra, which
corresponds to the SS path labeled A in Scheme 2 In all complexes
except in [Cu(NH3)4]2+ there is a second coordination shell formed
by carbon atoms (path B in Scheme 2). Due to the high symmetry
and stability of the CuPc chelate complex, higher shell contributions
cannot be neglected. The SS paths corresponding to these shells
are paths C and D in Scheme 2. Similar paths had to be taken into
account in [Cu(His)2] and [Cu(Gly)2]. It has been assumed that all
atoms are coplanar. This is strictly true for the CuPc complex in
the gas phase and solid state, as observed from ED and XRD.26,27
In [Cu(Gly)2] some deviations from planarity have been observed
by XRD in the crystalline compound, from 0.02 to 0.04 Å,16 which
would induce changes in Cu(II)-ligand coordination distances far
beyond the detection limits of the technique (0.015-0.05 Å for
first and higher shells, respectively).
Since octahedral Jahn-Teller-distorted symmetry is the one most
frequently found in Cu(II) ions and the existence of axial water
molecules is normal in this type of complex,5,18 the spectrum of
each complex was analyzed with and without an additional single-
scattering contribution from two axial water molecules (path AX
in Scheme 2).
MS paths within these arrangements have to be taken into account
as well. Common contributions to all complexes are the ones within
the first coordination shell, paths including aligned atoms labeled
I-III as well as triangular paths labeled IV. Inclusion of these MS
paths in the analysis procedure is crucial since they are relevant
only in a square-planar geometry, where the absorbing atom and
those forming the first shell are coplanar. In contrast, in a tetrahedral
arrangement paths I-IV are negligible. Thus, their inclusion might
serve to discriminate between both coordination environments. Paths
of type V have a high amplitude in complexes with a second
coordination shell. Paths of type VI, derived from SS path D, have
to be considered in the CuPc, Cu(Gly)2, and Cu(His)2 complexes.
In a first step, ab initio calculations (with FEFF code) of the
contributions to the EXAFS signal of each scattering path were
performed, either single scattering (SS) or multiple scattering (MS),
which includes relevant backscatterers, i.e., C and N and excludes
hydrogen atoms. A curved wave amplitude filter was used,
according to which only paths with amplitude equal to or higher
than 5% of the most intense path, obtained by integrating through
(35) (a) Cook, J. W.; Sayers, D. E. J. Appl. Phys. 1981, 52, 5024-5029.
(b) Sakane, H.; Miyanaga, T.; Watanabe, I.; Matsubayashi, N.; Ikeda,
S.; Yokoyama, Y. Jpn. J. Appl. Phys. 1993, 32, 4641-4645.
(36) (a) Mustre de Leo´n, J.; Rehr, J. J.; Zabinsky, S. I.; Albers, R. C. Phys.
ReV. B 1991, 44, 4146-4156. (b) Zabinsky, S. I.; Rehr, J. J.;
Ankudinov, A.; Albers, R. C.; Eller, M. J. Phys. ReV. B 1995, 52,
2995-3009. (c) Ankudinov, A.; Ravel, B.; Rehr, J. J.; Conradson, S.
D. Phys. ReV. B 1998, 58, 7565-7576.
(37) Stern, E. A.; Newville, M.; Ravel, B.; Yacoby, Y.; Haskel, D. Physica
B 1995, 208/209, 117-120.
Inorganic Chemistry, Vol. 43, No. 21, 2004 6677

Carrera et al.
MS paths involving second and higher shells (type V and VI) are
crucial to discriminate between the different types of ligands. Note
that except glycine, the first coordination shell is formed by four
N atoms; thus, higher shells are the key feature to distinguish
between them.
a single shell either. The Cu-O third shell is formed by two O
atoms at a well-defined distance, but it is not easy to fit either
because the adjacent MS paths of type VI in Scheme 2 are not
well defined due to the poor resolution of the first and second
coordination shells. Thus, the number of SS and MS contributions
is roughly multiplied by two and concomitantly the number of free
parameters. A fit of the EXAFS function cannot be performed, but
rather a calculation using the structural parameters from XRD
studies has to be done.
Between 6 and 13 paths (depending on the complex) have to be
taken into account. The number of free parameters could become
quite high in some fits, thus exceeding the number of independent
points according to the Nyquist theorem.38 Therefore, some
strategies were undertaken to reduce the number of fitting param-
eters. First, assuming the initial structures, coordination numbers
are fixed for each shell. Moreover, in complexes with higher
coordination shells, which have chelate ligands, bond distances
within the ligand have to be coherent. Concerning Debye-Waller
factors of MS contributions, those corresponding to paths I-IV
were not considered free parameters, rather they were defined as
functions of the Debye-Waller factors of SS path A; then according
to the independent vibration approximation model39
2
The amplitude reduction factor, S0 , was set equal to 0.81 in all
complexes. This was the value heuristically obtained in the analysis
of the EXAFS spectra of another first transition series cation,
Cr(III), in a similar environment and measured in similar condi-
tions.42 This value is similar to that obtained in the fit of the spectra
2
of a Cu metal foil, S0 ) 0.85.37
Results and Discussion
The raw EXAFS spectra of the five complexes dissolved
in aqueous solutions have been included as Supporting
Information in Figure 1aSI, and the magnitude of the Cu-N
phase-corrected Fourier transform (FT) of the spectra
included in Figure 1aSI is given in Figure 1bSI. The EXAFS
function becomes increasingly complex when the number
of atoms in the ligand becomes larger: it is a quasi simple
wave in the tetrammine, it shows a beating at around 5.5
Å-1 in the [Cu(en)2]2+ spectra, and in the complex formed
with the tetradentate ligand phthalocyanine the function is
rather complex. A peak centered at around 2.00 Å, which
should be ascribed to the Cu-N first coordination shell, is
the main contribution in all spectra. Additional contributions
appear beyond 2.8 Å, thus allowing determination of the
structural properties of higher coordination shells. The plotted
spectra are those corresponding to the higher concentration
in each case: 0.1 M in [Cu(NH3)4]2+, [Cu(en)2]2+, and CuPc,
0.05 M in [Cu(His)2], 0.025 M in [Cu(Gly)2]. Nevertheless,
the fit of the 0.01 M solutions was carried out as well for
the first three complexes, finding exactly the same results,
since the only difference between both series of spectra was
the signal-to-noise ratio, which was better in the more
concentrated solutions.
All these spectra were analyzed following the procedure
described above. Fit results have been included in Table 1
and fitting ranges and goodness of fit as Supporting Informa-
tion in Table 1SI. The latter table also includes the number
and type of free parameters used during the fitting procedure
in each case. Fit weight was 2 in all cases.
[Cu(NH3)4]2+. [Cu(NH3)4]2+ is the most simple Cu(II)
complex with monodentate N-coordinating ligands. Consid-
ering the restrictions described above for path lengths and
amplitude filters, the computation with FEFF yielded the
single-scattering path corresponding to the Cu-N coordina-
tion shell (path A in Scheme 2) and the associated four
multiple scattering ones (paths I-IV in Scheme 2).
2
σI2 ) σII2 ) σIV2 ) 2σA
2
σIII ) 4 σA
This model has shown to be useful in systems with similar
structure40 as well as in other systems with octahedral coordination
geometry.41
The EXAFS signals calculated for the paths thus obtained were
used to fit the experimental spectra using FEFFIT program. The
free parameters in each fit were coordination distances and Debye-
Waller (DW) factors for each SS path and DW factors for MS paths
type V and VI. Additionally, a unique value of the inner potential
correction, ∆E0, was taken as a free parameter in the analysis of
each complex. Thus, the number of free parameters was kept
between 5 and 11, below the number of independent points in
experimental spectra. According to the Nyquist theorem, this
number was between 22 and 24, depending on the fitting range.
Moreover, among the possible results of the fit, we took into account
only those having physical meaning, both concerning coordination
distances and DW factors. First, in the chelate complexes the only
distance left completely free for the fit is the first-shell one, since
bond distances within the ligand are fixed, which constrains the
values of the distances between Cu(II) and second and higher shells.
Second, concerning DW factors of SS paths, they have to increase
with coordination distances.
The spectra of the [Cu(Gly)2] complex could not be fitted with
the above-described procedure. The most probable reason is the
split that, according to XRD16 data, appears in the first shell, formed
by two O and two N atoms. In fact, Cu-N and Cu-O contributions
are too close in distance to be fitted like independent shells, but
since they have different backscatterer atoms they cannot be fitted
as a single shell. In addition, there are two pairs of C atoms in the
second coordination shell 0.2 Å apart, so they cannot be fitted as
(38) Sayers D. E. Report on Standard and Criteria Committee 2000
(http://ixs.csrri. it.edu).
(39) (a) Yokoyama, T.; Kobayashi, K.; Ohta, T.; Ugawa, A. Phys. ReV. B
1996, 53, 6111.(b) Haskel, D. Ph.D. Thesis, University of Washington,
Seattle, WA, 1998.
(40) (a) Ayala, R.; Sa´nchez Marcos, E.; D´ıaz-Moreno, S.; Sole´, V. A.;
Mun˜oz-Pa´ez, A. J. Phys. Chem. B 2001, 105, 7588-7593. (b) Mun˜oz-
Pa´ez, A.; D´ıaz-Moreno, S.; Sa´nchez Marcos, E.; Rehr, J. J. Inorg.
Chem. 2000, 39, 3784-3790.
(41) (a) D´ıaz-Moreno, S.; Mart´ınez, J. M.; Mun˜oz-Pa´ez, A.; Sakane, H.;
Watanabe, I. J. Phys. Chem. A 1998, 102, 7435-7441. (b) D´ıaz-
Moreno, S.; Mun˜oz-Pa´ez, A.; Sa´nchez Marcos, E. J. Phys. Chem. B
2000, 104, 11794-11800.
Inclusion of one additional SS contribution, formed by two
axial water molecules, improves the fit slightly. For each
SS contribution there are two free parameters: interatomic
(42) Merkling, P. J.; Mun˜oz-Pa´ez, A.; Sa´nchez Marcos, E. J. Am. Chem.
Soc. 2002, 124, 10911-10920.
6678 Inorganic Chemistry, Vol. 43, No. 21, 2004

Nature of Metal Binding Sites in Cu(II) Complexes
Table 1. Structural Parameters Obtained from the EXAFS Fitsa (see
Scheme 2 for path definition)
shell
R (Å)
CN
σ2 (Å2)
SS paths
MS paths
[Cu(NH3)4]2+
0.004
Cu-N
Cu-Oax
2.02
3.3
4
2
A
AX
I, II, III, IV
0.011
[Cu(en)2]2+
0.0036
0.006
Cu-N
Cu-C
Cu-Oax
2.01
2.84
3.4
4
4
2
A
B
AX
I, II, III
V
0.01
CuPc
0.0008
0.002
0.004
0.007
Cu-N
Cu-C
Cu-N
Cu-C
1.95
2.97
3.37
4.1
4
8
4
8
A
B
C
D
I, II, III
V,V′
Figure 1. (a) Raw EXAFS data, k2-weighted, of [Cu(NH3)4]2+: experi-
mental function (solid line) and fit obtained by using the parameters included
in Table 1 (dashed line). (b) Cu-N phase-corrected FT of the EXAFS
spectra included in Figure 1a. Magnitude and imaginary part: experimental
(solid line); best fit (dashed line).
VI,VI′
[Cu(His)2]
0.0034
0.007
0.007
0.008
Cu-N
Cu-C
Cu-C
Cu-N
1.99
2.97
3.32
4.1
4
2
4
2
A
B
B
D
I, II, III
V, V′
V, V′
[Cu(en)2]2+. Bis(ethylenediamine)copper(II) is a complex
closely related with tetrammine copper(II) from the point of
view of both reactivity and structure. In [Cu(en)2]2+ each
ethylenediamine molecule acts as a bidentate ligand coor-
dinating through the N atom of each -NH2 group. Its
chelating character makes the resulting complex more stable
(log K1 ) 9.7) than the tetrammine complex (log K1 ) 4.1).18
It forms five-membered chelate rings similar to the ones
found in the glycine-type coordination mode described above.
The beat appearing at around 5.5 Å-1 and a new peak in the
FT at 3 Å (see Figure 1SI) are clear indications of a second
coordination shell, formed by the two C atoms of each
ethylenediamine ligand, corresponding to path B in Scheme
2. As expected, the values obtained for coordination distances
and DW factors for the first shell, formed by four N atoms
at 2.01 Å, are similar to those found for tetrammine copper-
(II), Table 1. The chelate effect does not induce significant
changes in Cu-N coordination distances compared to those
of tetrammine complex, as already observed by Koniyama20
and Mazzi19 in crystalline compounds. These result do not
confirm those reported by Fujita et al.,23 who from an XRD
study of an aqueous solution report a Cu-N coordination
distance of 1.93 Å. The second shell is formed by four carbon
atoms at 2.84 Å from Cu(II) ions, a value consistent with
bond distances within the ligand and similar to that obtained
in the crystalline complex20 and in solution.23 The low values
of the DW of this shell, 0.006 Å2, can be attributed to the
chelate effect. As observed in the previous complex, the
inclusion of two axial water molecules improves the fit,
goodness-of-fit decreases from 0.03 to 0.01, and although
VI, VI′
[Cu(Gly)2]
0.003
0.004/0.008
0.015
Cu-N/O
Cu-C/C
Cu-O
1.95
2.74/2.8
4.0
2+2
2+2
2
A × 2
B × 2
D
(I, II) × 2
V × 2
VI × 2
a
R: Coordination distances; CN: coordination numbers, σ2: Debye-
Waller factors or mean square displacements, SS: Single Scattering,
MS: Multiple Scattering. Uncertainties (except for Cu(Gly)2, see text): R
in first coordination shell: 0.01 Å, R second and higher coordination
shells: 0.015-0.05 Å Debye-Waller factors 0.002 Å2.
distances R(Cu-X) and Debye-Waller factors, σ2 . The best
X
fit was obtained for a square-planar structure with four
molecules of ammonia at 2.02 Å. These values compare well
with those previously obtained by XRD or EXAFS.19,21-25
DW factors are similar to those found in stable square-
planar40 complexes and smaller than those of hexahy-
drates.42,43 Concerning axial molecular water, the optimized
values for coordination distance R(Cu-Oax) and DW factor
σ2 were 3.3 and 0.011 Å2. This distance is much longer
AX
than that found in tetrammine and ethylenediamine in
crystalline square-planar complexes, 2.5-2.8 Å.5 Likewise,
axial water molecules in [Cu(H2O)6]2+ complexes appear
around 2.3 Å.5,44 Nevertheless, a similar value has been found
in crystalline [Cu(NH3)4]SO4‚H2O.19,23
Comparative plots for experimental spectra and best fit,
both in k and R spaces, taking into account all these
contributions, are included in Figure 1. As can be seen in
the figure, the fit quality, both in R and k spaces, is good
and no well-defined contribution that would improve the fit
appears to be missing.
two new free parameters are included, the value of ø2
red
decreases from 26 to 14. The values obtained for the
coordination distance, 3.4 Å, and DW factor, 0.01 Å2, are
very close to those of [Cu(NH3)4]2+ as well. As seen in Figure
2, including comparative plots in k and R spaces and in the
parameters included in Table 1SI, fit quality is quite good.
To show the relative amplitudes of the contributions taken
into account to fit this spectrum, comparative plots of the
amplitudes of their FT’s have been included in Figure 3SI.
As seen there, the dominant contributions are those corre-
sponding to the first-shell SS Cu-N and to the second-shell
SS Cu-C. The two following contributions correspond to
MS path type V involving Cu-N-C atoms and to SS Cu-
Since the amplitude of the path from axial water molecules
is rather small, an additional comparative plot without this
shell has been included in the Supporting Information (Figure
2SI). Fit quality is slightly worse in the EXAFS spectra for
low k values, 2.5-6 Å-1, and around 3 Å in the magnitude
of the Fourier transform. The parameters of the Cu-N
coordination shell are not affected by this suppression.
(43) Mart´ınez, J. M.; Pappalardo, R. R.; Sa´nchez Marcos, E.; Refson, K.;
D´ıaz-Moreno, S.; Mun˜oz-Pa´ez, A. J. Phys. Chem. B 1998, 102, 3272-
3282.
(44) Persson, I.; Persson, P.; Sandstro¨m, M.; Ullstrom, A. S. J. Chem. Soc.,
Dalton Trans. 2002, 1256-1265.
Inorganic Chemistry, Vol. 43, No. 21, 2004 6679

Carrera et al.
Figure 2. (a) Raw EXAFS data, k2-weighted, of [Cu(en)2]2+: experimental
function (solid line) and fit obtained by using the parameters included in
Table 1 (dashed line). (b) Cu-N phase-corrected Fourier transform of the
EXAFS spectra included in Figure 2a. Magnitude and imaginary part:
experimental (solid line); best fit (dashed line).
Figure 3. (a) Raw EXAFS data, k2-weighted, of CuPc: experimental
function (solid line) and fit obtained by using the parameters included in
Table 1 (dashed line). (b) Cu-N phase-corrected Fourier transform of the
EXAFS spectra included in Figure 3a. Magnitude and imaginary part:
experimental (solid line); best fit (dashed line).
OAX. The EXAFS functions in k space corresponding to these
contributions as well as the sum of three main MS paths
within the first coordination shell have been included in
Figure 4SI. As shown there some of these contributions are
opposite in phase, so they partially cancel each other.
An additional figure has been made including the fit
without axially coordinated water molecules (Figure 5SI) to
show the relevance of this contribution. Reproduction of the
EXAFS spectra at low k values is slightly worse, see Figure
5SIa, at 3, 5, and 7 Å-1. The imaginary part of the FT
worsens between 2 and 4 Å, and the magnitude is worse at
ca. 3.5 Å, see Figure 5SIb.
CuPc. Phthalocyanine is a macrocyclic conjugated planar
molecule formed by four isoindole units, joined by four
nitrogen atoms, thus rendering it a unique fully symmetric
tetradentate ligand. More than 70 different ions can be
incorporated into the central cavity,26,27 making the resulting
complexes very stable. Although CuPc is a rather large
complex, formed by 41 atoms excluding H, this tetradentate
ligand is very rigid, probably due to the so-called macrocyclic
effect, the greater thermodynamic stability of a complex with
a cyclic polydentate ligand compared to a complex formed
by a comparable noncyclic ligand.18 This higher stability
together with its high symmetry makes the analysis of distant
shells (R > 4.0 Å) feasible. Phthalocyanine forms six-
membered chelate rings similar to the ones found in the
“histamine” coordination mode.
As shown in Figure 1SI, the EXAFS function of this
compound is rather complex, and three clear new peaks
appear in the amplitude of its FT. A good fit of this spectrum
was attained including the first four shells (four N, eight C,
four N, eight C) with the corresponding SS (A-D in Scheme
2) and MS paths. MS contributions are particularly relevant
in this case due to the high symmetry of this complex as
well as its rigidity. The relevant MS paths are those involving
quasi aligned atoms: the ones within the four N atoms of
the first shell, paths I, II, and III in Scheme 2; those involving
atoms from the first and second shell, similar to that
appearing in [Cu(en)2], path V in Scheme 2; and overall,
those involving atoms from the first, second, and fourth
shells, paths V, V′,VI, VI′. Note that N atoms of the third
shell are not aligned with any other and thus do not contribute
with significant MS paths. The obtained Cu-N first-shell
coordination distance, 1.95 Å, is close to that obtained from
XRD of the crystal structure,26 from ED of gas-phase
molecules, and from DFT calculations.27b Inclusion of axial
water molecules did not improve the fit, probably due to
the existence of intense single- and multiple-scattering
contributions at similar distances. To show the relative
amplitude of each individual single- and multiple-scattering
path taken into account in this fit, the magnitude of the
Fourier transforms of the SS paths and of some of the MS
ones has been plotted in Figure 6SI. The main contributions
are those corresponding to SS first, second and third shells,
while that of the fourth shell is negligible. In contrast, MS
associated to the fourth shell are particularly intense, which
is due to the focusing effect of the quasi-aligned Cu-C-N
atoms. Figure 7SI includes comparative plots in k space of
MS and SS contributions. Note again the high amplitude of
MS paths associated with the fourth shell (paths V,V′,VI,
VI′) It is remarkable as well that MS within the first shell
keep a high amplitude for high k values, rather unusual in
MS contributions, which can be attributed to the extremely
low DW factor within the first Cu-N shell.
Figure 3 shows comparative plots of the experimental and
best fit of the EXAFS spectra of CuPc in k and R spaces
including the SS and MS paths described above, the quality
of the fit being rather good. Given the relevance of MS paths
shown in Figures 6SI and 7SI, their exclusion yields a poor
reproduction of the spectrum both in k and R spaces, as seen
in the comparative plots included as Supporting Information
(Figure 8SI).
Although fit quality is good, it can be seen in Figure 3
that the intensity of the signal beyond 4 Å is still high enough
to allow for analysis of the fifth and sixth shells, which are
formed by eight C atoms at 5.5 and 6.5 Å, respectively.26
However, inclusion of these additional two coordination
shells adds a high number of relevant MS paths, 42 paths
with intensity higher than 10%, for which there are not simple
ways of computing DW factors. Consequently, a too high
number of parameters has to be adjusted, and thus, analysis
of the EXAFS spectra could not resolve these shells. Note
that exclusion of these shells from the fit affects the quality
of it around 4 Å (see Figure 3b).
Analysis of the solid compound yielded the same results
with a significant difference: the DW factor for the first shell
6680 Inorganic Chemistry, Vol. 43, No. 21, 2004

Nature of Metal Binding Sites in Cu(II) Complexes
was 0.002 Å2 instead of 0.0008 Å2. The difference is already
visible in the magnitude of the FT (not plotted), smaller in
the spectrum of the solid than in that of the dissolved species.
This value indicates that smaller mean-square deviations
appear in the dissolved species as compared to the solid-
state compound. In principle, the opposite effect was
expected, since the disorder is usually higher in liquids than
in solid systems. Nevertheless, the lack of packing restrictions
in solution might produce a symmetrically averaged structure.
Thus, although up to now the structure of the crystalline
compounds has been considered as a good approximation
to the structure of the dissolved species, there might be
significant differences between the structures of the same
species in different aggregation states, as we have already
observed in [Ni(CN)4 ]2-.40b
[Cu(Gly)2]. The previous analyses are a test of the
accuracy of phase shift and backscattering amplitude func-
tions calculated theoretically for this type of complex. They
also show the capability of the technique to reproduce
previously reported values of first-shell metal-ligand coor-
dination distances within 0.015 Å. For chelate complexes,
such as CuPc, with well-defined symmetric structure, up to
four coordination shells can be analyzed using a number of
free parameters small enough to make the obtained param-
eters statistically significant.
Analysis of the spectra of [Cu(Gly)2] was carried out with
a similar strategy. Nevertheless, although this complex looks
simple, especially when compared with CuPc, the splitting
of the first shell, formed by two N atoms plus two O atoms,
renders the fit unfeasible as explained before. As it happens
with ED,27 EXAFS cannot solve adjacent contributions very
close in distance, such as those formed by the first (two N
plus two O) or second (two C plus two C) coordination
shells, although some authors do not take into account the
obvious limitations of the technique and have overanalyzed
the EXAFS data of this complex.17 In an alternative approach,
we introduced as fixed parameters coordination distances,
averaged from those obtained from XRD,16 and DW factors,
similar to those obtained for the complexes already analyzed.
They were changed step by step until the EXAFS spectrum
was reproduced. It required consideration of the 13 paths
indicated in Table 1 (five SS +eight MS) with the parameters
included in this table as well. Although, according to XRD,
the average value of the Cu-N coordination distance is 2.00
Å and that of Cu-O is 1.955 Å, the best reproduction of
the spectrum was obtained with a common value of
coordination distance, 1.95 Å, and the same DW factor, 0.003
Å2. Since these values were not yielded by the fitting
procedure, goodness of fit values are not provided nor
indeterminations in the parameters. Figure 4 includes the
comparative plots of the calculated and experimental spectra.
Analysis of this spectrum confirms that the technique is
sensitive to small changes in the structure: note that although
[Cu(en)2]2+ and [Cu(Gly)2] have rather similar structures, the
spectrum of the first was fitted straightforwardly and that of
the second was not. The difficulties undergone to analyze
this spectrum suggest that a different strategy is required to
gain insight into its structural elucidation, for instance, one
Figure 4. (a) Raw EXAFS data, k2-weighted, of Cu(Gly)2: experimental
function (solid line) and spectrum obtained by using the parameters included
in Table 1 (dashed line). (b) Cu-N phase-corrected Fourier transform of
the EXAFS spectra included in Figure 4a. Magnitude and imaginary part:
experimental (solid line); best fit (dashed line).
procedure could be based on the optimization of a few key
parameters belonging to a structure of the complex by XAS-
independent techniques, such as other spectroscopies or
theoretical computations.42,45
[Cu(His)2]. As explained above, proposed coordination
modes for [Cu(His)2] include histamine-histamine, histamine-
glycine, and glycine-glycine. In all of them histidine acts as
a bidentate ligand coordinating through Nam and Nim or OCOO
,
whereas there is only one set of values of M-L coordination
distances in solution obtained from EXAFS,12 which, on the
other hand, correspond to a structure in conflict with that
proposed from NMR studies.11 Under these premises, in our
first approach we carried out a fit of the spectrum of the
Cu(His)2 complex considering the symmetric structure
histamine-histamine like. This is sound taking into account
the studies by Ramonelli,10 Grommen,11a and Manikandan,11b
who point out that for a neutral pH the most stable
coordination is histamine-histamine like, since the spectrum
was measured at physiological pH. For this type of coordina-
tion both the square-planar and tetrahedral arrangements, a
coordination found in some copper(II) complexes,5 were
considered. Since both structures were rather symmetric, they
have the minimum number of shells and free parameters, as
compared to the other structures already proposed.
No good fit was attained with the tetrahedral arrangement.
As outlined before, a distinction between those arrangements
was possible due to MS contributions within the first shell
because paths I-III are not relevant for tetrahedral geometry.
In contrast, a rather good fit was obtained for the square-
planar arrangement, taking into account only four coordina-
tion shells, with the corresponding four SS paths and nine
MS paths. Only nine free parameters (four coordination
distances, four DW factors, and ∆E0) were used, obtaining
the values included in Table 1. The first shell is formed by
four N atoms at 1.99 Å. This value is between that obtained
for CuPc and [Cu(NH3)4]2+.
The coordination distance is the same as that obtained by
Otzusumi for related Cu(II) complexes,12 similar to that found
by Boswell et al. in the amidating enzyme for Cu-Nim,29
1.97 Å, where the imidazole N atom comes from histidine,
and to that obtained by Strange et al.28 for Cu-Nim, 2.00 Å,
(45) Merkling, P. J.; Mun˜oz-Pa´ez, A.; Mart´ınez, J. M.; Pappalardo, R. R.;
Sa´nchez Marcos, E. Phys. ReV. B 2001, 64, 012201-4.
Inorganic Chemistry, Vol. 43, No. 21, 2004 6681

Carrera et al.
Figure 5. (a) Raw EXAFS data, k2-weighted, of Cu(His)2: experimental
function (solid line) and fit obtained by using the parameters included in
Table 1 (dashed line) in the histamine-histamine coordination mode. (b)
Cu-N phase-corrected Fourier transform of the EXAFS spectra included
in Figure 5a. Magnitude and imaginary part: experimental (solid line); best
fit (dashed line).
Figure 6. (a) Raw EXAFS data, k2-weighted, of Cu(His)2: experimental
function (solid line) and fit obtained for histamine-glycine (dashed line).
(b) Cu-N phase-corrected Fourier transform of the EXAFS spectra included
in Figure 6a. Magnitude and imaginary part: experimental (solid line); best
fit (dashed line).
in the complex Cu(im)4. The last paper is the first EXAFS
study taking into account MS contributions in this type of
system. In fact, due to the presence of four imidazole rings,
MS contributions are very relevant in these systems, and their
EXAFS spectra are more complex than the spectrum here
presented, where there are only two imidazole rings, as can
be seen by comparing Figure 5 in this paper with Figures
2-6 in ref 28. Although the coordination proposed here,
bidentate with two histidine molecules coordinating through
Nim and Nam, is different from that proposed by Strange et
al., monodentate with four imidazole rings coordinating Cu-
(II), in both cases the existence of imidazole rings coplanar
with Cu(II) ions yields the amplitude of MS paths very high.
Note the fit presented here, both in k and R spaces,
imaginary part and magnitude, is as good as that obtained
in the previous complexes studied for which the already
known structures were reproduced. Moreover, all the reported
values for coordination distances and Debye-Waller factors
are coherent within the complex and consistent with the
values found for the other complexes.
Taking into account that other structures were proposed
for this complex, e.g., the glycine-glycine found by Evertsson
using XRD7 at pH ) 3.7 and by Ozutsumi12 by EXAFS at
neutral pH, an additional calculation of the spectra was
carried out using as a starting structure that of the Cu(Gly)2
complex, with the difficulties already described. No good
fit could be obtained with this approach. Similarly, we tried
the other structure appearing in Scheme 1, i.e., histamine-
glycine. Although it implies a higher number of paths and
thus of free parameters, reproduction of the spectrum is
slightly worse than that attained for the histamine-histamine
coordination mode (cf. plots in Figure 6a, showing some
discrepancies between 4 and 7 Å-1 and Figure 6b around 4
Å with the similar plots in Figure 5). It is quite clear that
there are only minor differences between the two coordina-
tion modes.
analysis of higher shells which eventually could allow the
discrimination among the different coordination modes.
By comparing Figures 5 and 6, we propose that the
predominant structure is the so-called histamine-histamine,
with the coordination parameters included in Table 1.
Nevertheless, since an X-ray absorption spectrum provides
average values of all the coordination sites in which the
absorbing atom is involved, coming back to Scheme 1, we
cannot discard the existence of small amounts of histamine-
glycine coordination mode. This model is consistent with
the recently published studies by Grommen and Manikandan
using EPR and ENDOR spectroscopies11 and does not
confirm those of the early work by Ozutsumi obtained by
EXAFS.12 The fact that the latter authors did not take into
account MS contributions might be the reason for the
disagreement. EPR and ENDOR studies do not provide
M-ligand coordination distances, although it can be inferred
from the values of H-H coordination distances provided by
the ENDOR technique. We complete these results by
providing coordination distances for the first four coordina-
tion shells, which are formed by four N atoms at 1.99 Å,
two C atoms at 2.97 Å, four C atoms at 3.32 Å, and two N
atoms at 4.1 Å, values very similar to those found for the
CuPc complex.
Conclusions
The structures of [Cu(NH3)4]2+, [Cu(en)2]2+ and CuPc have
been determined by fitting their EXAFS spectra using a
reduced number of SS and MS contributions (6-13) and
free parameters (5-10), and results compare well with the
available information. The short coordination distance found
for the Cu-N first shell in CuPc, 1.95 Å, might be attributed
to the macrocycle effect. Higher values are found in the other
two complexes, 2.01 and 2.02 Å. Both sets of distances can
be solved with EXAFS in these symmetric complexes. The
existence of chelating ligands with strong bonds allows the
rather precise determination of distant shells, Cu-X (X )
C, N, O) with bond distances longer than 4.0 Å. Concerning
the labile axial water molecules, the weakness of the
interaction renders the amplitude of the contribution rather
small, so their existence can be assessed only in simple
complexes where no other significant contribution appears
As outlined before, the key to discriminate between the
proposed structures is the number of imidazole rings within
the plane of the complex, from which there are two in the
first coordination mode and only one in the second. This
geometry enhances the amplitude of MS paths V and VI, as
already pointed out by Strange et al.,28 thus allowing the
6682 Inorganic Chemistry, Vol. 43, No. 21, 2004

Nature of Metal Binding Sites in Cu(II) Complexes
at similar distances, as happens for [Cu(NH3)4]2+ or
[Cu(en)2]2+.
parent structure glycine-glycine can be fully discarded as well
as the tetrahedral arrangement.
In the case of the [Cu(His)2] complex (pH ) 7.3), several
coordination models have been tested, obtaining the best
results for the symmetric histamine-histamine coordination
mode in which both histidine molecules coordinate Cu(II)
cations through N atoms from the amino group and from
the imidazole ring, forming two six-membered chelate rings.
The obtained coordination distance for the first Cu-N shell
is 1.99 Å. In this coordination mode two imidazole rings
are coplanar with the atoms forming the first coordination
shell of the complex. For this reason multiple scattering
contributions become relevant, thus allowing determination
of higher shells. Two C atoms at 2.97 Å, four C atoms at
3.37 Å, and two N atoms at 4.1 Å form the successive shells.
Only a slightly worse fit was obtained for the related
coordination mode histamine-glycine, where only one imi-
dazole ring is coplanar with Cu(II) ions. Nevertheless, the
Through these analyses we observed that the inclusion of
MS paths provides indirect information about coordination
angles (i.e., discriminating between tetrahedral and square-
planar arrangements) since three- and four-body correlations
have a significant effect on the spectra.
Acknowledgment. ESRF is acknowledged for beamtime
allocation at the BM29 line (Experiment number CH-1192
and CH-1397). We thank Drs. S. Ansell, G. Subias Peruga,
and S. D´ıaz-Moreno for their cooperation during the experi-
ments. Spanish DGICYT is acknowledged for financial
support (BQU2002-02217) and BQU2002-04364-C02-01).
Supporting Information Available: Additional figures and
tables. This material is available free of charge via the Internet at
http://pubs.acs.org.
IC049699Q
Inorganic Chemistry, Vol. 43, No. 21, 2004 6683