Vanadium(V) complexes of polydentate amino alcohols: Fine-tuning complex properties

Article abstract of DOI:10.1021/ja9808200

In aqueous solution, the stability, the structure, and the lability of 12 vanadium(V) complexes were determined and related to the amino alcohol ligand properties. The amino alcohols were chosen to maintain a five- coordinate vanadium geometry in the vanadium complex and were based on diethanolamine as the parent ligand. Apparent and composite stability constants were determined for all 12 complexes. Selected complexes were further examined to determine solution structure, lability, and thermodynamic properties. We empirically show that the formation constants for the series of 12 complexes, when plotted as a function of the pK(a) value for the protonated amino alcohol, generate a curve that initially increases as the pK(a) value increases and then decreases as the pK(a) value continues to increase. The observed associations suggest that the electron-donating capacity of the amino alcohol plays a major role in complex stability; however, metal coordination number, sterics in the amino alcohol, and solvation also are found to affect the stability. Determination of the thermodynamic parameters of selected complexes showed that the enthalpic component was the major contributor to the stability of the complex, but that the entropic component opposed this term. A series of studies was conducted to examine whether the lability of these complexes varied from the parent complexes. Unfortunately, the variation in the lability of these complexes was much less than the variation in complex stability. In summary, these studies describe quantitatively the variation in complex stability and lability as the structure of amino alcohol is modified and explain why some complexes show less stability than predicted on the basis of the pK(a) value. In short, this study provides valuable information for the design of additional vanadium complexes with specific properties.

Full text of DOI:10.1021/ja9808200

J. Am. Chem. Soc. 1998, 120, 8069-8078
8069
Vanadium(V) Complexes of Polydentate Amino Alcohols:
Fine-Tuning Complex Properties
Debbie C. Crans* and Iman Boukhobza†
Contribution from the Department of Chemistry, Colorado State UniVersity,
Fort Collins, Colorado 80523-1872
ReceiVed March 11, 1998
Abstract: In aqueous solution, the stability, the structure, and the lability of 12 vanadium(V) complexes were
determined and related to the amino alcohol ligand properties. The amino alcohols were chosen to maintain
a five-coordinate vanadium geometry in the vanadium complex and were based on diethanolamine as the
parent ligand. Apparent and composite stability constants were determined for all 12 complexes. Selected
complexes were further examined to determine solution structure, lability, and thermodynamic properties. We
empirically show that the formation constants for the series of 12 complexes, when plotted as a function of the
pKa value for the protonated amino alcohol, generate a curve that initially increases as the pKa value increases
and then decreases as the pKa value continues to increase. The observed associations suggest that the electron-
donating capacity of the amino alcohol plays a major role in complex stability; however, metal coordination
number, sterics in the amino alcohol, and solvation also are found to affect the stability. Determination of the
thermodynamic parameters of selected complexes showed that the enthalpic component was the major contributor
to the stability of the complex, but that the entropic component opposed this term. A series of studies was
conducted to examine whether the lability of these complexes varied from the parent complexes. Unfortunately,
the variation in the lability of these complexes was much less than the variation in complex stability. In
summary, these studies describe quantitatively the variation in complex stability and lability as the structure
of amino alcohol is modified and explain why some complexes show less stability than predicted on the basis
of the pKa value. In short, this study provides valuable information for the design of additional vanadium
complexes with specific properties.
Introduction
exploring the development of vanadium compounds that are
more stable and less labile under physiological conditions.1,7
The mechanism of the insulin mimetic properties of vanadate,
and other vanadium complexes, remains elusive, in part because
of the rich aqueous chemistry of vanadium under physiological
conditions as well as the complexity of the cell-signaling cascade
in which vanadate1 and other vanadium compounds induce their
observed insulin-like action.2-4 We are investigating the
fundamental properties of selected vanadium compounds in
order to be able to design compounds with a given desired
structure, stability, and lability. These types of studies are
important because it has recently been recognized that most of
the compounds currently being investigated for their insulin
action properties have limited lifetimes under physiological
conditions.1 In view of the complex redox and hydrolytic
chemistry these compounds can undergo, it is very difficult to
determine what vanadium species are actually responsible for
the insulin-like action. Furthermore, the situation is complicated
by the fact that many vanadium compounds induce biological
responses.5,6 We, and others, have therefore recently been
Modifying ligand structure to generate a complex with more
favorable properties with respect to complex stability is well-
known.8-13 In the case of the stability of metal complexes with
amine ligands, the inductive effects of the amine have been
correlated with the pKa value of the protonated amine.8-11 In
contrast, the reaction of bis(2,4-pentanediono)copper(II) with
pyridine bases showed no correlation between the stability
constant and the substituent on the pyridine ring consistent with
the greater contribution of solvation and steric effects than that
of the electron-donating effects of the amine.12 Sterics were
also found to influence the stability of Ni(II) complexes with
primary aliphatic amines11 and trialkylamines,13 and Mo(V)
complex formation with triphenyl-substituted diethanolamine
(6) Crans, D. C.; Keramidas, A. D.; Hoover-Litty, H.; Anderson, O. P.;
Miller, S. M.; Lemoine, L. M.; Pleasic-Williams, S.; Rossomando, A. J.;
Sweet, L. J. Am. Chem. Soc. 1997, 119, 5447.
(7) Gibney, B. R.; Stemmler, A. J.; Pilotek, S.; Kampf, J. W.; Pecoraro,
V. L. Inorg. Chem. 1993, 32, 6008.
(8) Carlin, R. L.; Walker, F. A. J. Am. Chem. Soc. 1965, 87, 2128.
(9) Walker, F. A.; Hui, E.; Walker, J. M. J. Am. Chem. Soc. 1975, 97,
2390.
(10) Massould, S. S.; Labib, L.; Iskander, M. F. Polyhedron 1994, 13,
517.
(11) Sacconi, L.; Lombardo, G.; Ciofalo, R. J. Am. Chem. Soc. 1960,
82, 4182.
† 1994-1997, Ph.D. Candidate at Abd El Malik Essaadi University, BP
2121, Tetouan, Morocco. Present address: Department of Chemistry,
Colorado State University, Ft. Collins, CO 80523-1872.
(1) Crans, D. C.; Mahroof-Tahir, M.; Keramidas, A. D. Mol. Cell.
Biochem. 1995, 153, 17.
(2) Shechter, Y. Diabetes 1990, 39, 1.
(3) Stern, A.; Yin, X.; Tsang, X. Y.; Davison, A.; Moon, J. Biochem.
Cell Biol. 1993, 71, 103.
(4) Goldfine, A. B.; Simonson, D. C.; Folli, F.; Patti, M.-E.; Kahn, C.
R. Mol. Cell. Biochem. 1995, 153, 217.
(12) May, W. R.; Jones, M. M. Inorg. Nucl. Chem. 1963, 25, 507.
(13) Sacconi, L.; Lombardo, G.; Paoletti, P. J. Am. Chem. Soc. 1960,
82, 4185.
(14) Barbaro, P.; Belderrain, T. R.; Bianchini, C.; Scapacci, G.; Masi,
D. Inorg. Chem. 1996, 35, 3351.
(5) Crans, D. C. Comments Inorg. Chem. 1994, 16, 1.
S0002-7863(98)00820-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/01/1998

8070 J. Am. Chem. Soc., Vol. 120, No. 32, 1998
Crans and Boukhobza
([HOCH2-PhCH]NH[PhCH-PhCHOH]).14 Sterics also af-
fected the heat of the reaction between vanadyl acetylacetonate
and a number of nitrogen and oxygen donors.8
these studies provide information to assist the ongoing quest to
design desired properties within this and other families of
compounds.
Modification of ligand structure has also been used to change
the rate and kinetic parameters of metal complexes.15-19 In
general, one expects the ligand pKa to affect the rate of complex
formation if the reaction occurs by an associative mechanism
or by a dissociative mechanism where the lability is determined
by the leaving group. A variety of vanadium(V) complexes
and their complex formation rates have been examined in
detail.20-25 All show similar rates of complex formation. Thus,
on the basis of these studies, one would not expect lability
changes but, on the basis of other complexes, one might expect
sterics to influence the lability. Increasing the steric bulk in
aliphatic amines has been found to decrease the rate of Cr-
amine complex formation.16 The increase in the enthalpy of
activation term has been determined to be the major contributor
to this rate decrease since in the nonpolar solvents studied, no
evidence for solvation effects were observed. On the other hand,
the electrostatic attraction between V(III) complexes and ami-
nopolycarboxylic acids was found to be important in determining
the rate of ligand substitution.18
A series of vanadium(V) complexes with diethanolamine
derivatives has been found to form 1:1 complexes with a charge
of -1.26 In this study we are systematically examining the
effects on complex formation by small ligand perturbations. We
have previously shown that the vanadium(V) complex of
triethanolamine is labile since the complex undergoes exchange
with free ligand in aqueous solution.23,27 In addition, the
aqueous solution structure of this complex is different from the
crystal structure of the compound.28 In the current study, we
have selected a series of amino alcohol-vanadium(V) com-
plexes which previous studies have suggested would favor
formation of five-coordinate vanadium species in aqueous
solution.29,30 We have modified the parent amino alcohol ligand,
diethanolamine, to determine the effect of each modification
on the corresponding stability and lability of the corresponding
vanadium(V) complexes. To interpret the observed stability
pattern, we have determined thermodynamic parameters for
several complexes. As described in this paper these studies
required that we determine the temperature dependence of the
chemical shift for the external reference VOCl3. Collectively,
Experimental Section
A. Materials. The reagents used in this work were all of reagent
grade. The water was distilled and deionized (DDI) on an ion-exchange
column. Chemicals were purchased from Aldrich or Sigma and checked
for purity by 1H NMR spectroscopy. Several ligands, including meso-
2,2′-diphenyliminodiethanol and C2-2,2′-diphenyliminodiethanol,31 2-[2-
(hydroxyethyl)amino-2-hydroxymethyl]-1,3-propanediol32 and 2-[2-
(hydroxyethyl)amino]-2-methyl-1-propanol, were prepared following
the procedure previously described.32a
B. Synthesis. 2-Hydroxy-5-nitrobenzyl Chloride. p-Nitrophenol
(5.00 g, 35.9 mmol) was dissolved in a mixture of 65 cm3 of
concentrated HCl, 0.5 cm3 of concentrated H2SO4, and 6.6 cm3 of
dimethoxymethane. The stirred solution was heated at 343 K for 5 h
under the continuous addition of HCl gas (20 cm3 of concentrated HCl
was added dropwise into 30 cm3 of concentrated H2SO4, generating
HCl gas which was slowly bubbled into the reaction mixture). Once
the reaction was complete, the mixture was kept at 253 K for 3 h,
filtered, and dried under vacuum to give the crude product. After
recrystallization from benzene (4.80 g, 25.6 mmol, 72%), a white solid
resulted. 1H NMR 300 MHz (D2O): 4.60 (s, 2H, CH2Cl), 7.00 (d,
1H, ArH), 8.20 (dd, 1H, ArH), 8.30 (d, 1H, ArH) ppm. 13C{1H} 75
MHz (D2O): 61.5, 118.0, 118.3, 128.3, 129.1, 130.3, 163.4 ppm.
N-(2-Hydroxy-5-nitrobenzyl)iminodiethanol (HNBIDE). First,
2,5-hydroxynitrobenzyl chloride was prepared from p-nitrophenol and
then the addition of diethanolamine completed the synthesis. Dietha-
nolamine (1.45 g, 13.8 mmol) was dissolved in 20 cm3 of 2-propanol,
and after the addition of 2-hydroxy-5-nitrobenzyl chloride (1.20 g, 6.40
mmol), the colorless solution immediately turned yellow. This reaction
mixture was refluxed overnight. The solution was cooled to room
temperature and filtered to remove decomposition products, and a
yellow solid was obtained after evaporation of the solvent. Water (10
cm3) was then added, and the solution was poured into 150 cm3 of
boiling acetone and 5 cm3 of ethyl acetate. The solution was
concentrated to 70 cm3 and then kept at 253 K overnight to give a
yellow solid which was recrystallized from water. The crystals (1.32
g, 5.15 mmol, 80%) were isolated by filtration. (Anal. Found: C,
51.20; H, 6.3; N, 10.77. Calcd for C11H16N2O5: C, 51.55; H, 6.3; N,
10.93%). 1H NMR 300 MHz (D2O): 3.20 (t, 4H, CH2CH2OH), 3.90
(t, 4H, CH2CH2OH), 4.15 (s, 2H, ArCH2N), 6.55 (d, 1H, ArH), 8.10
(dd, 1H, ArH), 8.20 (d, 1H, ArH) ppm. 13C{1H} 75 MHz (D2O): 57.3,
58.2, 59.9, 121.2, 121.6, 131.1, 136.2, 178.6 ppm.
2-[2-(Hydroxyethyl)amino-2-hydroxymethyl]-1,3-propanediol
(TDEA). Tris(hydroxy methyl)aminomethane (2.00 g, 16.5 mmol) was
dissolved in 50 cm3 of water. To the stirred solution, at 277 K, liquid
ethylene oxide (2.47 cm3, 49.5 mmol) was added, and stirring was
continued for 5 h while warming to ambient temperature. The solvent
was then removed to give a white solid which was recrystallized twice
from 2-propanol and three times from ethanol to give white crystals
(2.10 g, 12.7 mmol, 80%). (Anal. Found: C, 43.20; H, 9.13; N, 8.61.
Calcd for C6H15NO4: C, 43.62; H, 9.15; N, 8.47%). 1H NMR 300
MHz (D2O): 2.75 (t, 2H, CH2CH2OH), 3.60 (s, 2H, C(CH2OH)3), 3.70
(t, 2H, CH2CH2OH) ppm. 13C{1H} 75 MHz (D2O): 45.1, 63.3, 63.9,
65.8 ppm.
(15) Jones, T. E.; Cole, J. R.; Nusser, B. J. Inorg. Chem. 1978, 17, 3680.
(16) Burkey, T. J. Polyhedron 1989, 8, 2681.
(17) Cooper, T. H.; Mayaer, M. J.; Leung, K.-H. Ochrymowycz, L. A.;
Rorabacher, D. B. Inorg. Chem. 1992, 31, 3796.
(18) Ikeda, Y.; Hassan, R. M.; Abd El-Fattah, M. M.; Park, Y. Y.;
Tomiyasu, H. Collect. Czech. Chem. Commun. 1994, 59, 1077.
(19) Argonne, L.; Yu, Q.; Ochrymowycz, L. A.; Rarabacher, D. B. Inorg.
Chem. 1995, 34, 1844.
(20) Whittaker, M. P.; Assay, J.; Eyring, E. M. J. Phys. Chem. 1966,
70, 1005.
(21) Crans, D. C.; Rithner, C. D.; Theisen, L. A. J. Am. Chem. Soc.
1990, 112, 2901.
2-[2-(Hydroxyethyl)amino]-2-methyl-1-propanol (DMeDEA).
2-Amino-2-methyl-1-propanol (1.00 g, 11.2 mmol) and ethylene oxide
(0.839 cm3, 16.8 mmol) were dissolved in 50 cm3 of absolute ethanol
at 277 K. After the reaction was stirred for 2 h at 277 K, the solution
was allowed to warm to ambient temperature. After stirring the solution
overnight, the solvent was evaporated, 50 cm3 of water was then added,
and the solution was extracted with chloroform (5 × 30 cm3). After
drying the solution over anhydrous Na2SO4, the solvent was removed
in vacuo to yield the desired compound (1.04 g, 7.80 mmol, 70%) as
(22) Kustin, K.; Toppen, D. L. J. Am. Chem. Soc. 1973, 95, 3564.
(23) Crans, D. C.; Ehde, P. M.; Shin, P. K.; Pettersson, L. J. Am. Chem.
Soc. 1991, 113, 3728.
(24) Kustin, K.; Liu, S.-T.; Nicolini, C.; Toppen, D. L. J. Am. Chem.
Soc. 1974, 96, 7410.
(25) Kustin, K.; Nicolini, C.; Toppen, D. L. J. Am. Chem. Soc. 1974,
96, 7416.
(26) Crans, D. C.; Shin, P. K. Inorg. Chem. 1988, 27, 1797.
(27) Crans, D. C.; Shin, P. K.; Armstrong, K. B. In Mechanistic
Bioinorganic Chemistry; Thorp, H., Pecoraro, V., Eds.; American Chemical
Society: Washington, DC 1995; p 303.
(28) Crans, D. C.; Chen, H.; Anderson, O. P.; Miller, S. M. J. Am. Chem.
Soc. 1993, 115, 6769.
(29) Crans, D. C.; Shin, P. K. J. Am. Chem. Soc. 1994, 116, 1305.
(30) Crans, D. C.; Keramidas, A. D.; Amin, S.; Anderson, O. P.; Miller,
S. M. J. Chem. Soc., Dalton Trans. 1997, 2799.
(31) Scialdone, M. A.; Meyers, A. I. Tetrahedron Lett. 1994, 35, 7533.
(32) (a) VanGreat, L. Anal. Chem. 1970, 42, 679. (b) Bonningue, A.-C.;
Brazier, J. F.; Houalla, D.; Osman, F. H. Phosphorus, Sulfur Relat. Elem.
1979, 5, 291. (c) Hoffman, R. E. Magn. Reson. Chem. 1988, 26, 523. (c)
Hindman, J. C. J. Phys. Chem. 1966, 44, 4582.

V(V) Complexes of Polydentate Amino Alcohols
J. Am. Chem. Soc., Vol. 120, No. 32, 1998 8071
a colorless oil which crystallizes upon standing at room temperature.
(Anal. Found: C, 54.06; H, 10.99; N, 10.35. Calcd for C6H15NO4:
C, 54.10; H, 11.35; N, 10.51%). 1H NMR 300 MHz (D2O): 1.00 (s,
6H, C(CH3)2), 2.60 (t, 2H, CH2CH2OH), 3.35 (t, 2H, CH2CH2OH), 3.60
(s, 2H, C(CH3)2CH2OH) ppm. 13C{1H} 75 MHz (D2O): 25.2, 45.4,
55.8, 63.7, 70.2 ppm.
Because it is not possible to distinguish different protonated forms
by NMR spectroscopy, only composite formation constants are directly
determined. Let the total concentration of free V(V) monomer be V1,
where [V1] ) [VO43-] + [HVO42-] + [H2VO4-] + [H3VO4]. Thus,
the reaction as defined in eq 6 leads to the definition of Kcom in eq 7.
N-(2,5-Dimethylbenzyl)iminodiethanol (DMeBIDE). Diethanol-
amine (14.0 g, 133 mmol) and 2,5-dimethylbenzyl chloride (9.12 cm3,
62.0 mmol) were dissolved in 100 cm3 of 2-propanol, and the solution
was refluxed for 48 h. After cooling to ambient temperature, 20 cm3
of water was added and the solution was extracted with ethyl acetate
(3 × 50 cm3). The combined ethyl acetate solutions were dried over
Na2SO4. The solvent was removed in vacuo to yield N-(2,5-dimeth-
ylbenzyl)iminodiethanol (12.4 g, 55.6 mmol, 90%) as a colorless oil.
(Anal. Found: C, 69.22; H, 9.58; N, 6.20. Calcd for C13H21NO2 C,
69.92; H, 9.47; N, 6.27%. 1H NMR 300 MHz (D2O): 2.50 (2s, 6H,
CH3ArCH3), 2.65 (t, 4H, CH2CH2OH), 3.50 (t, 4H, CH2CH2OH), 3.60
(s, 2H, ArCH2N), 7.05 (2d, 2H, ArH), 7.10 (s, 1H, ArH) ppm. 13C-
{1H} 75 MHz (D2O): 20.7, 22.6, 57.5, 57.8, 58.6, 130.3, 133.8, 134.1,
135.0, 138.1,139.7 ppm.
V1 + Lfree a complex + H2O
(6)
(7)
[complex]
Kcom
)
[V1][Lfree
]
1
Structural and Lability Studies. The H and 13C NMR spectra
were recorded on a Bruker ACP spectrometer (7.0 T) operating at 300
MHz for 1H and 75 MHz for 13C. A 2 Hz exponential line broadening
was typically applied to the 13C FIDs prior Fourier transformation. The
13C NMR spectra were acquired with 200 ppm spectral window, a 90°
pulse width, and a relaxation delay of 700 ms.
NMR Experiments Leading To Determination of Thermody-
namic Parameters. The thermodynamic parameters were measured
for vanadium complex (51V and 1H), amino alcohol (1H), and vanadate
(51V) using variable-temperature NMR spectroscopy. The temperature
range examined was from 298 to 338 K for most experiments, although
all vanadium complexes were also studied down to 278 K in one series
of experiments. The temperature for the variable-temperature experi-
ments was calibrated using an 80% ethylene glycol sample in DMSO-
d6 to an accuracy of (1°.32a It is well-known that the chemical shift
of H2O is strongly dependent on temperature32b,c and potentially will
cause erroneous results if measuring the chemical shift against the H2O
signal. For convenience, some of the repeat experiments were carried
out in this manner. As a result, we have measured the temperature
dependence of the H2O signal under our conditions using the internal
standards tetramethylammonium chloride (TMAC)32b and DSS, as well
as the instrumental frequency for calibration. The chemical shifts were
also checked against external DSS at ambient temperature. We find
that the change in chemical shift of H2O signal (0.016 ppm/deg) under
our conditions is within the range of values reported in the literature.32a
The concentration of vanadium complexes were measured in samples
containing a ratio of 1:2 vanadium to ligand at the optimum pH (the
formation constant was determined using 51V NMR spectroscopy at
the optimum pH). The concentrations of vanadium complex and V1
were used to determine the Keq from which the ∆H° was calculated.
Four experiments using a minimum of five different temperatures were
performed for each complex, and two experiments using five different
temperatures were performed for the vanadate parent solution. The
thermodynamic parameters for the amino alcohol protonation reaction
were determined by measuring the 1H NMR chemical shift at the
appropriate series of pH values at various temperatures. These
experiments were performed once for the five different temperatures,
and additional studies were carried out to ensure no changes were
observed when increasing the temperature range studied. The typical
experiment was initiated at ambient temperature, and the temperature
was then raised. The temperature dependence of the chemical shift
for H2O is well documented,32b,c but similar information is not available
in the literature for vanadium compounds. It is therefore first necessary
to determine the effect of temperature on the chemical shift of the
reference compound VOCl3 (0 ppm, 298 K) by recording 51V NMR
spectra in the unlocked mode at various temperatures. The chemical
shift of VOCl3 changes 0.054/deg, and the data is shown in Figure 4.
Furthermore, such information is complicated by the known exchange
processes that vanadium(V) complexes undergo.5,21 Equipped with this
information we were able to measure accurately the chemical shift of
the species under study in this work at various temperatures.
C. Potentiometric Measurements of pKa Values. Using standard-
ized stock solutions of HCl and NaOH, the pKa values of the new
amines, as well as several known amines, were measured by titration
at ambient temperature. Amine concentrations ranged from 10 to 70
mM. The titrant volumes (at 0.20 cm3) were added, and the pH change
was measured. The number of experimental points used to define the
titrations ranged from 100 to 150 points. The pKa values were thus
calculated from the titration curves generated using the spreadsheet
program Wingz. If appropriate an iterative process was used to
minimize the deviation in the pKa value. In addition, the pKa values
1
for selected amino alcohols were measured by the H NMR chemical
shift as described under thermodynamic parameters.
D. NMR Spectroscopy. Formation Constant Measurements and
Calculations. 51V NMR spectra were recorded on a ACP-300 Bruker
NMR spectrometer (7.0 T) at 78.9 MHz. A spectral window of 150
ppm, a 90° pulse angle, and an acquisition time of 0.25 s with no
relaxation delay were used. A 15 Hz exponential line broadening was
applied before Fourier transformation. 51V NMR chemical shifts are
reported with respect to the external reference standard VOCl3 at 0
ppm. Complex and oligomeric vanadate mole fractions were measured
using the Bruker integration software. We found no changes in mole
fractions of species by integrating spectra obtained with different
parameters. Assuming that all vanadium present in solution is in the
form of vanadium(V), the total added vanadium concentration combined
with the mole fractions of each vanadium complex, concentrations of
complex and oligomeric vanadates can be calculated. The reactions
describing the system and its pH dependence are shown below in eqs
1-7. The Kcom and the Keq values (eqs 4 and 7) were derived by linear
least squares from plots of formation constants as a function of the
multiple of concentrations of amino alcohol and vanadate in the
appropriate protonated forms (see text), respectively.
-
Both H2VO4 and H2L undergo protonation and deprotonation
reactions, and those described in eqs 1 and 2 are important to the
complex formation examined here. Complex formation is shown in
eq 3 and the equilibrium constant in eq 4.
H2VO4- h HVO42- + H+
H3L+ + H2O h H2L + H3O+
H2L + H2VO4- h [VO2L-] + 2H2O
[VO2L-]
(1)
(2)
(3)
Keq
)
(4)
-
Sample Preparation for NMR Studies. Vanadate stock solutions
were prepared by dissolving sodium metavanadate (NaVO3) in doubly
distilled and deionized water. All NMR samples were prepared at room
temperature immediately before NMR spectroscopic determinations.
The samples all contained 20% (v/v) D2O. Vanadate and ligand
concentrations were varied as needed. The ionic strength was
maintained at 0.40 with 4.00 M KCl. The pH measurements were
carried out at room temperature using a Corning pH-meter-140 before
[H2L][H2VO4
]
In general, the Kcom for each complex was measured near the pH
(pHexp) where optimum amount of complex was observed (pHopt
defined as shown in eq 5.
)
pHopt ) 1/2(pKa(H2VO4-) + pKa(HL+))
(5)

8072 J. Am. Chem. Soc., Vol. 120, No. 32, 1998
Crans and Boukhobza
Table 1. List of 12 Amino Alcohol Ligands, Their Trivial Names, Abbreviations, and pKa Values of the Protonated Aminesa
ligand
abbreviation
R1
R2
R3
pKa
diethanolamine
N-methyldiethanolamine
triethanolamine
triisopropanolamine (tri-2-propanolamine)
N-ethyldiethanolamine
N-butyldiethanolamine
N-(2-hydroxy-5-nitrobenzyl)iminodiethanol
N-2,5-(dimethylbenzyl)iminodiethanol
C2-2,2′-diphenyliminodiethanol
meso-2,2′-diphenyliminodiethanol
2-[2-(hydroxyethyl)amino-2-hydroxymethyl]-1,3-propandiol
2-[2-(hydroxyethyl)amino]-2-methyl-1-propanol
DEA
MeDEA
TEA
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
H
CH3
CH2CH2OH
8.88
8.52
7.76
7.86
8.66
8.99
TPA
CH2CH(CH3)OH CH2CH(CH3)OH CH2CH(CH3)OH
EtDEA
BuDEA
HNBIDE
DMeBIDE
cDPDEA
tDPDEA
TDEA
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH(Ph)CH2OH
CH(Ph)CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH2CH2OH
CH(Ph)CH2OH
CH(Ph)CH2OH
C(CH2OH)3
CH2CH3
CH2CH2CH2CH3
CH2(p-OHPhNO2) 5.60
CH2(p-CH3PhCH3) 7.58
H
H
H
H
5.90
6.30
8.00
9.63
DMeDEA
C(CH3)2CH2OH
a N-Phenyldiethanolamine (pKa ) 4.13) was also examined in this work, but since no stable five-coordinate complex formed it is not included
in this table.
and after each spectroscopic measurement to ensure that no significant
changes ((0.05) occurred during that time. The majority of the samples
used in the NMR measurements were prepared and analyzed in
duplicate. Selected samples were analyzed in triplicate.
Results and Discussion
Synthesis of Ligands. The new ligands were prepared by
conventional procedures involving the nucleophilic attack of
amines on ethylene oxide or benzyl chloride derivatives. 2-[2-
(Hydroxyethyl)amino-2-hydroxymethyl]-1,3-propanediol was
prepared in good yield by the reaction of tris(hydroxymethyl)-
aminomethane and ethylene oxide. The hydrochloride salt of
this ligand was previously prepared from 1-bromo-2-trityloxy-
ethane and tris(hydroxymethyl)aminomethane in the presence
of acetyl chloride.33 We used an excess of ethylene oxide due
to its high volatility and isolated pure compound after recrys-
Figure 1. The ratio of the concentration of vanadium complex to the
concentration of monomeric vanadate plotted as a function of pH. These
reactions were examined in solutions of 10 mM vanadate and 20 mM
ligand (TDEA (9), BuDEA (0), and EtDEA (b)). The ionic strength
tallizations from ethanol and 2-propanol. The preparation of
DMeDEA from 1 equiv of 2-amino-2-methyl-1-propanol with
1 equiv of ethylene oxide lead to a yield of 35% in a mixture
of products.32 We increased the amount of ethylene oxide to
was adjusted to 0.40 using 4.00 M KCl. The solid line connects data
1.5 equiv and thus increased the yield by 35%. We prepared
N-(2-hydroxy-5-nitrobenzyl)iminodiethanol (HNBIDE) by the
condensation of 2-hydroxy-5-nitrobenzyl chloride with dietha-
nolamine in 2-propanol. The synthesis of 2-hydroxy-5-ni-
trobenzyl chloride was carried out in the presence of excess
hydrochloric acid to reduce the formation of 2-hydroxymethyl-
4-nitrophenol. The preparation of an isomer (N-(4-hydroxy-3-
nitrobenzyl)iminodiethanol) was previously reported;34 however,
our isolated yield of 80% is a significant improvement of the
reported 64%.
Characterization of Vanadium(V) Complexes in Aqueous
Solution. In aqueous solution vanadate and diethanolamine
derivatives form complexes with a 1:1 stoichiometry. Table 1
lists names, abbreviations, and structures of the 12 amino
alcohols used in this work. The vanadium(V) complexes were
characterized using 51V, 1H, and 13C NMR as described
previously.26 Here we describe the pH dependence of the
complexes, complex stoichiometry, the apparent composite
formation constants (pH dependent Kcom and pH independent
Keq), the solution structure of the complexes, and the complex
points and does not represent a fit.
lability and thermodynamic properties of selected complexes,
vanadate, and amino alcohols.
Complex Stoichiometry and pH Dependence. To deter-
mine the experimental conditions appropriate for measuring the
formation constants for the series of vanadium(V) complexes
generated from the amino alcohols listed in Table 1, the pH
dependence (see for example Figure 1) and complex stoichi-
ometry were first determined. Complex stoichiometry was
determined by measuring the amount of complex in a series of
solutions containing varying vanadate and ligand concentrations.
The preliminary data were analyzed, and it was found that one
-
molecule each of H2VO4 and H2L was needed to form
complex. Since these studies also involve measurements at
different pH values, the number of protons involved in the
reaction is also determined. On the basis of the previously
reported vanadium(V) complexes of this type (V-DEA and
-
V-TEA), we expected that the stoichiometries each of H2VO4
and H2L were 1:1. Furthermore, if the complex is analogous
to the V-TEA complex, no protons will form (or be used) in
this reaction and accordingly these complexes will have a charge
of -1 (and a composition of VO2L- as the V-TEA complex).26
(33) Fowler, P. A.; Haines, A. H.; Taylor, J. K.; Chrystal, E. J. T.;
Gravestock, M. B. J. Chem. Soc., Perkin Trans. 1994, 2229.
(34) Stephanyan, G. M.; Margaryan, S. S.; Iradyan; M. A.; Garibdzh-
anyan, B. T. Khim.-Farm. Zh. 1981, 15, 24.

V(V) Complexes of Polydentate Amino Alcohols
J. Am. Chem. Soc., Vol. 120, No. 32, 1998 8073
Table 2. Summary of the Apparent Equilibrium Constants of
We confirmed these stoichiometries by observing the relation-
ships implied in eq 4.
Vanadium(V) Complexes Studied in This Worka
DEA
derivatives
complex 51
δ(ppm)
V
Kcom
Keq
ligand
Although the stability of these complexes is due to the
protonation states of each vanadate and amino alcohol, the fact
that many of these amino alcohols are used as biological buffers
makes it relevant for the life scientist to consider the concentra-
tion of total complex in solution as a function of pH. The pH
at which the greatest concentration of complex is present can
be calculated from the pKa of the protonated amino alcohol and
b
c
(M-1
)
pHexp
(M-1
)
pKa
DEA
-485
-477
-476
-474
-478
-478
-487
-487
-479
-480
-484
-520
58
240
44
40
15
28
130
55
210
430
770
392
8.54
510 (pH 7-11)
8.88d
8.52d
8.66
8.99
MeDEA
EtDEA
BuDEA
DMeBIDE
DMeDEA
tDPDEA
cDPDEA
TEA
8.40 1100 (pH 7-11)
8.47
8.60
7.90
8.90
7.25
7.00
8.00
8.00
300 (pH 7-10)
440 (pH 7-10)
22 (pH 7.5-9.5) 7.58
2 (pH 7-9)
9.63
6.30
-
the pKa of H2VO4 (eq 5).26 We measured the pKa values for
150 (pH 6-10)
61 (pH 6.5-10) 5.90
new and known amino alcohols by titration at ambient temper-
ature and for selected amino alcohols using the chemical shift
dependence as a function of pH at various temperatures. The
pKa of H2VO4- was measured by 51V NMR spectroscopy for a
series of solutions at different pH, at ambient temperature and
0.40 M KCl. By plotting the chemical shift as a function of
pH, the pKa of H2VO4- could be calculated to 8.2 which agreed
closely with that reported previously.26 Plots of [VO2L-]:[V1]
against pH display distinct maxima at pH 8.4 for EtDEA, at
pH 8.6 for BuDEA and at pH 8.0 for TDEA.
520 (pH 7-11)
950 (pH 7-11)
7.76d
7.86d
8.00
5.60
TPA
TDEA
HNBIDE
8.16 1600 (pH 7-10)
6.90
360 (pH 6-10)
a The reactions were studied at 5-50 mM concentrations of amino
alcohols and 5-20 mM concentrations of vanadate. The ionic strength
was adjusted to a value of 0.40 using a 4.00 M KCl solution. The
indicated formation constants have been calculated on the basis of
concentrations and have not been corrected for activity coefficients since
the assumption that the activity coefficients are 1.0 is not appropriate
at this ionic strength. All samples were prepared at ambient temperature
b
in duplicate. Kcom (M-1) is a pH-dependent measurable quantity (eq
Formation Constants. It is the intention in this work to
document the effects of alkyl substitution on the amino alcohol
ligands on the stability of the resulting vanadium(V)-amino
alcohol complex. The formation constants were determined by
51V NMR spectroscopy using the studies described above to
guide pH range and concentration ranges. As defined in eq 3
this equilibrium constant takes into account the protonation state
of both vanadate and amino alcohol which are defined in eqs 1
and 2 for the pH range 6 to 10.36 (Although it is known that
both HVO42- and H2VO4- react with L and/or HL+ when L is
alizarin and EDTA,22 eq 3 describes the fundamental reaction
relevant for derivation of thermodynamic properties for this
reaction.) Each Keq value was initially calculated from data
obtained from solutions containing varying vanadate and ligand
concentrations at an appropriate pH value and confirmed by
measurement at several other pH values. The Keq values are
given in Table 2 for the 12 different amino alcohols.
3), and Keq is a pH-independent value (eq 6) obtained from experimental
studies. c The value pHexp was chosen near pHopt and based on the
preliminary plot generated from [complex]:[V1] versus pH. d These pKa
values were reported in ref 34. The other pKa values were determined
as described in the Experimental Section.
One series of representative data obtained for the V-EtDEA
and V-BuDEA complexes is shown in a plot of the complex
concentration as a function of [H2VO4-] [H2L] in Figure 2. Each
set of data was determined in two or three different experiments.
These data were used to calculate both the pH-dependent and
the pH-independent (see below) equilibrium constants listed in
Table 2. All equilibrium constants calculated in this work were
determined at an ionic strength at which activity coefficients
deviate significantly from 1.36c Thus, in Table 2 we report the
apparent formation constants with the units M-1 to indicate
clearly that these constants have not been corrected by their
respective activity coefficient. The Kcom values for the 12 amino
alcohols were calculated at pHopt from the measurements
described above for determination of the Keq values listed in
Table 2. The resulting Kcom values are given in Table 2.
Effects of Ligand Perturbation on the Formation Con-
stants of H2VO4--H2L Complexes. First, we consider
substitution on the central nitrogen atom of DEA by simple alkyl
groups and in this and other comparisons below we focus on
the comparison of the Keq values unless otherwise noted. It
Figure 2. Representative example of the data for determination of
the equilibrium constant Keq of the complexes V-EtDEA (9) and
V-BuDEA (0) formed in aqueous solution at pH of 8.40 and 8.60,
respectively.
has previously been shown that the amine nitrogen atom is
essential for formation of these V-DEA-type complexes,26 and
indeed we confirm that N-phenyldiethanolamine does not form
a strong complex with vanadate (see footnote Table 1). On
the other hand, MeDEA forms a more stable complex with
vanadate than does DEA. Presumably, the MeDEA complex
is more stable due to greater electron donation by the amine
nitrogen. However, if electron donation were the major factor
determining complex stability, the EtDEA and BuDEA com-
plexes should be less stable than the complex MeDEA, but more
stable than the DEA complex. As seen in Table 2 the
V-MeDEA complex is 4-fold more stable than the V-EtDEA
and V-BuDEA complexes and the V-DEA complex is slightly
more stable than either of these complexes suggesting that other
factors are contributing to the stabilities of the vanadium(V)
complexes. The stability pattern was confirmed by examining
the complex formed between vanadate and DMeBIDE; its
complex is 14- to 20-fold less stable than the V-EtDEA and
V-BuDEA complexes.
(35) (a) Gresser, M. J.; Tracey, A. S. J. Am. Chem. Soc. 1986, 108, 1934.
(b) Tracey, A. S.; Jaswal, J. S. Inorg. Chem. 1993, 32, 4235.
(36) (a) Perrin, D. D.; Dempsey, B. In Buffers for pH and Metal Ion
Control. Perrin, D. D., Dempsey, B., Eds.; Chapman and Hall: New York,
1974. (b) Perrin, D. D.; Dempsey, B.; Serjeant, E. P. In pKa Prediction for
Organic Acids and Bases; Perrin, D. D., Dempsey, B., Serjeant, E. P. Eds.;
Chapman and Hall: New York, 1981. (c) Timmermans, J. The Physico-
chemical Constants of Binary Systems in Concentrated Solutions; Inter-
science: New York, 1959.
Second, we consider alkyl substitution on the carbons adjacent
to the central nitrogen atom of DEA. Substituents on the carbon
atoms adjacent to the nitrogen should increase the electron

8074 J. Am. Chem. Soc., Vol. 120, No. 32, 1998
Crans and Boukhobza
when plotted against the pKa of the protonated amino alcohol.
The bell-curve dependence demonstrates that the stability of
the vanadium(V) complex is not solely dictated by the basicity
of the amine nitrogen; the better the base does not imply the
more stable complex. The presence of the appropriate proto-
nated form of vanadate is clearly also crucial to the stability of
the complex, and it seems reasonable that amino alcohols
forming the most stable five-coordinate vanadium(V)-amino
alcohol complexes would have a pKa value of the protonated
amino alcohol in the neighborhood of 8-8.2.
Given the above considerations, how far could one modify
the structure of the amino alcohol before the resulting complex
no longer will follow the empirically observed bell-shaped
correlation of Kcom (or Keq) values shown in Figure 3? It is
well-known that alkylating a secondary amine to generate a
tertiary amine lowers its basicity in solution even though one
would have expected the basicity to increase.36 In addition,
modification of the amine on the carbon framework will
significantly affect the pKa of its protonated form. Thus, alkyl
group substitution on the carbon framework increases the pKa
value for the protonated amine as evidenced by the pKa value
for protonated DMeDEA compared to DEA, whereas aryl group
substitution decreases the pKa value for the protonated amine
as observed for cDPDEA and tDPDEA.
Figure 3. The composite pH-dependent and -independent formation
constants Kcom (0) and Keq (9) plotted as a function of the pKa of the
protonated amine.
donation of the nitrogen atom, but to a much smaller extent.
Accordingly, we may anticipate small changes in the formation
constant. As seen from Table 2 the V-DMeDEA complex is
300-fold less stable than the V-DEA complex. On the other
hand, the complexes V-cDPDEA and V-tDPDEA are 3- to
8-fold less stable than the V-DEA. This series of alkyl
substitutions clearly demonstrates that the electron donation of
the nitrogen atom is not the only important contributor to the
vanadium(V) complex stability. As we show the match between
the pKa of the protonated amine and pKa(H2VO4-) is also
important, as well as the alkyl substitution on the ligand.
Third, we consider substitution on the central nitrogen atom
of DEA by alkyl groups containing functionalities capable of
coordinating to the vanadium atom. As described previously,
the V-TEA complex is as stable as the V-DEA complex (4-
fold more stable if considering Kcom).26 It is interesting that
the V-TPA complex is 2-fold more stable than the V-DEA
complex and the V-TEA complex. These results suggest that
the extra chelating functionality, even though it is pendant in
aqueous solution,29 may (the V-TPA and V-TDEA complexes)
or may not (the V-TEA complex) contribute to stabilizing the
vanadium(V) complex.
The ability to provide estimates for complex stabilities would
be useful for future complex design. Thus it would be useful
to identify complexes that fit the empirical bell-curve relation-
ship between Kcom and pKa of the protonated amino alcohol
and compare their properties to selected complexes that fall off
the bell-curve. Such considerations should provide insight into
the other factors determining the stability and other properties
of aqueous vanadium(V) complexes. Given the fact that 12
complexes have been examined in this work and the importance
of amino alcohol pKa value for complex stability, we examined
the possibility that pKa value for ligand could be a first level
predictor for complex stability. Since this kind of crude analysis
would be most useful for the life-scientist, we carried out this
analysis focusing on the comparisons of Kcom values. The pKa
values for the protonated forms of DEA, TEA, TPA, TDEA,
and DMeDEA follow the generally anticipated trend for aliphatic
amines in aqueous solution; the tertiary amine is a poorer base
than the secondary amines. On the basis of this information,
we examined the relationship between the Kcom values for the
vanadium(V) complexes versus the pKa values of the protonated
form of the amino alcohol. As seen from Figure 3 a bell-curve
results with a maximum in the vicinity of pKa(H2VO4-).
Similarly, we find that the Keq values also follow a bell-curve
The amines cDPDEA and tDPDEA are particularly interesting
since they have identical formulas, functionalities, and substit-
uents; the only difference between these amines is the stereo-
chemistry of the two phenyl groups. Despite the structural
similarities, the difference in the pKa values for the protonated
amines is only 0.4 pH units. The observation that the vanadium-
(V) complexes formed from these ligands do not adhere to the
trend in Kcom values shown in Figure 3 is perhaps not surprising
and suggestive that other factors are also important to the
stability of these complexes. Since the pKa value for the
protonated DMeDEA is larger than the pKa value for DEA one
would expect that the Kcom for the former would be smaller
than that observed for DEA. As seen from Table 2 this is indeed
observed. However, the Kcom (or the Keq) value for the two
secondary amines, tDPDEA and cDPDEA do not follow the
trend shown in Figure 3; both complexes are significantly more
stable than anticipated given the low pKa value of their
respective protonated amino alcohols. The complex formed
from tDPDEA has a Keq value significantly larger than that for
the cDPDEA complex so that even a comparison of these
complexes with each other underlines the importance of other
factors. Alternative effects that may be important involve
phenyl groups stacking which may affect solvent organization
around the complex. The effects of differences in water
organization should be observable by comparing the thermo-
dynamic parameters of selected complexes and will be examined
below.
Solution Structure. The five-coordinate geometry of the
V-TEA and V-TPA complexes with a -1 charge in aqueous
solution has been well-documented.28,29 All 1:1 complexes
formed from vanadate with DEA, EtDEA, BuDEA, HNBIDE,
and tDPDEA are likely to have the five-coordinate geometry
of the V-TEA and V-TPA complexes with a -1 charge
observed for complexes formed from vanadate and TEA or
TPA.28 Detailed multinuclear NMR studies of the V-TEA and
V-TPA systems in aqueous and organic solution compared with
the solid-state characterization of these complexes showed the
aqueous solution complex to contain a VO2 unit chelated to a
tridentate TEA or TPA ligand.28 The suggestion of structural
similarity between the new complexes described here and the

V(V) Complexes of Polydentate Amino Alcohols
J. Am. Chem. Soc., Vol. 120, No. 32, 1998 8075
Table 3. 13C CIS Valuesa for Vanadium(V) Complexes Formed in
Aqueous Solutionb from EtDEA, BuDEA, and tDPDEA and
Compared to Carbon Atoms in DEAc
1
using H and 13C NMR spectroscopy.23,27,28 Presumably, all
complexes examined here also undergo these dynamic processes
in aqueous solution similar to those characterized in detail
previously. We will briefly summarize the dynamics of these
complexes inferred from previous detailed studies and the
experiments carried out to examine these expectations using 51
V
NMR line width and T1 measurements as well as variable-
1
temperature H NMR spectroscopic experiments on the series
of complexes presented in this manuscript.
The line widths of the eight complexes (V-DEA, V-MeDEA,
V-TPA, V-BuDEA, V-EtDEA, V-TDEA, V-DMeDEA,
V-DMeBIDE) fall within a range of 90-180 Hz, which is close
to that of V-TEA (110 Hz). The V-HNBIDE complex (D.
C. Crans and A. D. Keramidas, unpublished results), as well as
related derivatives30,37,38 form complexes with vanadium(V)
which contain a six-coordinate vanadium atom. Within this
structurally limited group of vanadium complexes (i.e., com-
plexes derived from DEA- or EDTA-based ligands), six-
coordinate complexes have line widths that fall outside the range
of those previously observed for five-coordinate complexes (∆ν
> 180 Hz).29,30,37,38 The vanadium(V) complexes of cDPDEA
and tDPDEA also have ∆ν values outside this range (290 and
310 Hz). The possibility that the line broadening in the 51V
NMR signals caused by increased lability of these complexes,
was therefore of interest and addressed by two methods: by
measuring the T1 values for the 51V NMR signals and by
complex
V-DEAd
V-EtDEA
V-BuDEA
V-tDPDEA
C1
C2
C1′
C2′
pH
14.3
14.5
14.5
12.6
5.4
4.5
4.6
2.2
14.3
14.5
14.5
12.9
5.4
4.5
4.6
2.6
9.01
8.53
8.53
7.20
a CIS ) δ(complex) - δ(free ligand). b The 13C NMR spectra were
obtained at ambient temperature in solutions added 600-1000 mM
ligand and 400 mM vanadate. c Key to ligand carbon atoms is
illustrated, and the chemical shifts of part of free ligands are indicated
in footnote e. d Previously described in refs 23, 26, 28, and 29. e DEA/
EtDEA/BuDEA/tDPDEA: C1 (61.0/58.6/58.6/67.6); C2 (52.4/57.1/57.7/
64.8).
1
other well-known complexes was investigated using H NMR
and 13C NMR spectroscopy and coordination-induced shift (CIS)
values. These results are listed in Table 3 for complexes formed
from DEA, EtDEA, BuDEA, and tDPDEA. In the V-TEA/
V-TPA complexes a CIS value of ∼15 ppm was observed for
coordinated C1, -1.2 ppm for pendant C1, and 4.6 ppm for C2
adjacent to the coordinated amine in a pendant hydroxyethyl
arm. The close similarity of for example, the C1 value of 14.5
ppm for V-EtDEA and V-BuDEA to the CIS values for the
V-DEA complex,29 is indicative that the complexes have very
similar solution structures. In contrast, HNBIDE does not have
a pendant hydroxyethyl arm: all four functionalities in this
ligand are coordinated to the metal ion. Thus, the data suggest
that all but the V-HNBIDE complex contain five-coordinate
vanadium(V) with a tridentate coordinated ligand with an overall
charge of -1.
1
variable-temperature H NMR studies. Assuming that these
vanadium compounds are within the motional limit and that T1
= T2, the relationship ∆ν ) (πT1)-1 should hold (where ∆ν is
the line width of the signal) and can be used as an indication of
contributions of dynamic processes in this series of complexes.
Since the measured T1 values in the 51V NMR spectra were all
within 30% of that predicted from the measured ∆ν, we
conclude that all of these complexes, including those from
tDPDEA and cDPDEA, show the same lability as the V-DEA
and V-TEA complexes on the 51V NMR time scale. The
increased line widths of V-cDPDEA and V-tDPDEA are thus
not likely to be attributable to a higher degree of chemical
exchange on the 51V NMR time scale. This result was
Recent studies with pyridyl and benzimidazole DEA deriva-
tives show that these ligands form extremely stable complexes
with vanadate in aqueous solution.30 However, structural
characterization of these materials shows that they contain six-
coordinate vanadium in contrast to most complexes examined
in this work.30 Thus, it is possible that a ligand forming a six-
coordinate complex will no longer form complexes that adhered
to the relationship shown in Figure 3. Indeed, Kcom for the
HNBIDE complex is 6-fold larger (although Keq is slightly less)
than that for the DEA complex. However, in the study with a
limited series of six-coordinate vanadium complexes of pyridyl
and benzimidazole containing DEA derivatives, we point out
that a bell-curve dependence is followed when comparing these
six-coordinate complexes among themselves.
Lability Studies. The V-TEA complex is known to undergo
both intra- and intermolecular exchange processes in aqueous
solution.27 Multinuclear NMR studies of the V-DEA, V-TEA,
and V-Tricine systems23,27,28 have previously shown that the
complex undergoes dynamic processes of several types. Vana-
date oligomers undergo intermolecular exchange in solution in
the neutral pH range containing above 1 mM uncomplexed
vanadate. These processes can be characterized using 51V
EXSY spectroscopy and/or 51V NMR line width analysis.21 Free
ligand amine exchanges with chelated amine (intermolecular
exchange). The pendent hydroxyethyl arm undergoes intramo-
lecular exchange with the chelated hydroxyethyl arms. These
two types of dynamic processes can be characterized in detail
1
confirmed by the variable-temperature H NMR experiments
for the V-DEA, V-MeDEA, V-TEA, V-cDPDEA, and
V-tDPDEA complexes which showed similar levels of dynamic
processes in the temperature range of 278-338 K. The spectra
for the V-cDPDEA and V-tDPDEA complexes deviated only
slightly from the general pattern observed above with respect
to the sharpness of the signals attributed to the methylene
protons and the methine proton adjacent to the phenyl groups.
We conclude that higher lability of the DPDEA complexes
cannot account for the observed differences in the 51V line width
between the V-TEA, V-cDPDEA, and V-tDPDEA com-
plexes.
Thus, the vanadium complexes of cDPDEA and tDPDEA
are exceptions to the empirical correlation between the line width
and the 51V NMR chemical shift previously observed for five-
and six-coordinate vanadium complexes.29 Since factors other
than dynamic exchange processes can affect the 51V NMR line
widths, we determined the thermodynamic parameters (∆H°,
∆S°, and ∆G°) to obtain further information on the inherent
stability of these complexes.
Thermodynamic Parameters for the Formation of Selected
Vanadium(V) Complexes, Amino Alcohols, and Vanadate
in Aqueous Solution. Before using the Keq for vanadium
(37) Plass, W. Inorg. Chim. Acta 1996, 244, 221.
(38) Mahroof-Tahir, M.; Keramidas, A. D.; Goldfarb, R. B.; Anderson,
O. P.; Miller, S. M.; Crans, D. C. Inorg. Chem. 1997, 36, 1657.

8076 J. Am. Chem. Soc., Vol. 120, No. 32, 1998
Crans and Boukhobza
complex formation to extract the thermodynamic parameters, a
few comments are in order. First, the concentration of H2O
has an activity of 1 given the fact that H2O is solvent. Second,
activity coefficients of charged species deviate significantly from
1 under high ionic strength conditions. However, the fact that
the complex and H2VO4- both have a charge of -1, combined
with the fact that L is neutral, the overall ratio of the activity
coefficients is likely to approach 1. Thus, we will not further
consider adjustments of the measured formation constants at
0.40 M KCl for calculation of the thermodynamic parameters.
Third, for interpretation of the thermodynamic parameters for
the complexes it is important that the corresponding parameters
-
for H2VO4 be available under the conditions of the study. In
Figure 4. The chemical shift for the reference compound VOCl3 (0.00
ppm, 298 K) is shown at variable temperatures.
the case of H2VO4-, we expect that the T∆S° term will dominate
over the enthalpic component on the basis of the study reported
at lower ionic strength.39a However, since some differences are
likely based on the differences in conditions of the studies, it is
necessary to redetermine these parameters in about 0.40 M KCl.
Fourth, as described previously in studies of molybdenum and
tungsten complexes of citrate and malate,40-42 the protonation
equilibria of respective ligands and their dependence on
temperature must be considered.
before eq 4 could be used to calculate the Keq. Deprotonation
of the amino alcohol affects the chemical shift for the adjacent
methylene groups in the amino alcohol ligand, and measuring
the chemical shift dependence as a function of pH is a sensitive
indicator of the pKa value for the amino alcohol. Although,
1
the H NMR chemical shifts changed as a function of pH for
DEA, MeDEA, EtDEA, TEA, tDPDEA, and cDPDEA, we
found no observable difference in their respective inflection
points (pKa values). The data for these studies are provided in
the Supporting Information. The inflection points were mea-
sured from the plots as well as calculated by an iteration using
Excel. Plotting the pKa as a function of 1/T (eq 9) led to the
determination of the enthalpic and entropic contributions. For
all of the ligands, the enthalpic contribution was experimentally
indistinguishable from 0 (a very small positive slope ranging
from 0.01 to 0.005), since no changes in pKa values were
measured in the temperature range examined (from 298 to 338
K). The entropic contributions were calculated from the
intercept of the plots of pKa against 1/T and the resulting
parameters are given in Table 4. As expected for a deproto-
nation reaction, the entropic component of the reaction is
responsible for the changes in free energy of the system. The
parameters for protonation of the amino alcohol ligands showed
in general that an enthalpic contribution to the protonation
reaction was experimentally indistinguishable from 0. Correla-
tion of the entropic component with observed ∆G° values has
been observed for other, structurally distinct ligands.39d
For the vanadate salt system, the situation is furthermore
complicated by the exchange processes that occur between the
monomeric and oligomeric ions which also affect chemical shifts
of these species.26 The protonation reaction of vanadate was
measured using 51V NMR chemical shift, as a function of pH
and temperature and the pKa,temp value could be derived for each
-
temperature (eq 8). Using eq 8 the concentrations of H2VO4
and HVO42- (and Ka,temp) could be calculated for each temper-
ature and the thermodynamic properties for the protonation
reaction of HVO4 can be calculated using eq 9.39a We
2-
pH ) pKa,temp + (log δ(low pH) - δtemp)/
(δtemp - δ(high pH)) (8)
pKa ) -∆H/2.303RT + ∆S/2.303R
(9)43
measured the temperature dependence of the chemical shift for
the reference VOCl3 in order to make appropriate adjustment
of chemical shift at various temperatures (Figure 4). Although
vanadate and vanadium complexes are labile, under the condi-
tions of this study, such effect contributes little to the observed
chemical shift changes. The parameters ∆H°, ∆G°, and ∆S°
-
Using the changes in the deprotonation constant for H2VO4
the Ka,temp was calculated for each of the 51V NMR experiments
at each temperature. Using these temperature-dependent Ka,temp
values, the temperature-dependent Keq,temp values for each
complex were calculated and used to determine the thermody-
namic parameters for complexes. The Keq,temp of different
complexes were plotted as a function of the reciprocal temper-
ature, and the thermodynamic parameters, given in Table 4, were
obtained. An example of the plots used for deriving the
parameters given in Table 4 is provided in Figure 5 for three
complexes: V-DEA, V-EtDEA, and V-tDPDEA. All va-
nadium(V) complexes have negative entropy terms and negative
enthalpy terms reflecting the opposing effects of these terms in
determination of the overall stability of the complexes. The
negative entropic contributions presumably reflect a combination
of the chelate formation with loss in the degree of freedom and
charges with associated solvation of products and reactants.
The enthalpic component is the major contributor to the
stability of V-DEA and V-MeDEA complexes at ambient
temperatures, and thus the overall free energy is favorable for
complex formation. The small increase in the pKa of the
protonated amine changed slightly the enthalpic term when
comparing the parameters for V-DEA and V-MeDEA. The
obtained in 0.40 M KCl (no buffer) were -5.6 (( 1.0) kJ mol-1
,
-47.8 (( 2.1) kJ mol-1 and 141.0 (( 3.9) J mol-1 K-1. These
values are all in reasonable agreement if compared to those
reported previously39a-c although these studies were all carried
out under different conditions (ranging from differences in
buffers and ionic strength).
Equation 8 was used to calculate the concentrations of
-
H2VO4 at each temperature. Furthermore, the effect of
temperature on the amino alcohol protonation reaction under
the conditions of the present study needed to be determined
(39) (a) Tracey, A. S.; Jaswal, J. S.; Angus-Dunne, S. J. Inorg. Chem.
1995, 34, 5680. (b) Larson, J. W. J. Chem. Eng. Data 1995, 40, 1276. (c)
Cruywagen, J. J.; Heyns, J. B.; Westra, A. N. Inorg. Chem. 1996, 35, 1556.
(d) Cruywagen, J. J.; Rohwer Miller, E. A. J. Chem. Soc., Dalton Trans.
1995, 3433.
(40) Cruywagen, J. J.; van de Water, R. F. Polyhedron 1986, 5, 521.
(41) Cruywagen, J. J.; Rohwer, E. A.; Wessels, G. F. S. Polyhedron 1995,
14, 23.
(42) Cruywagen, J. J.; Rohwer, E. A.; van de Water, R. F. Polyhedron
1997, 16, 243.
(43) Although deprotonation reactions are often considered, we, as well
as previous workers, on the simple vanadate anion (see ref 39), examined
the protonation reaction. This changes the sign change in eq 9.

V(V) Complexes of Polydentate Amino Alcohols
J. Am. Chem. Soc., Vol. 120, No. 32, 1998 8077
Table 4. Thermodynamic Parametersa,b for V-DEA, V-MeDEA,
V-EtDEA, V-TEA, V-tDPDEA, V-cDPDEA, and
stability. Although the enthalpic contribution for the V-EtDEA
is greater than that for V-TEA, the entropic contribution for
V-EtDEA is significantly larger resulting in complexes of
similar stability. In comparing the parameters for the V-cD-
PDEA and V-tDPDEA complexes, a difference of 4 kJ is
observed in the enthalpic component, whereas a much larger
difference is found in the entropic component. The fact that
V-cDPDEA is less stable than the V-tDPDEA complex can
mainly be attributed to the large unfavorable entropic compo-
nent.
Corresponding Free Ligands in Aqueous Solution at 298 Kb,c
V complexesa/
ligandsb
∆H°/kJ
∆S°/J mol-1
∆G°/kJ
mol-1 d
K-1 d
mol-1 d
pH
V-DEA
-21.2 ( 3.7 -13.5 ( 1.7 -17.2 ( 3.2 8.7
-22.2 ( 2.1 -13.9 ( 1.3 -18.1 ( 1.7 8.5
-27.5 ( 4.7 -42.0 ( 2.2 -15.0 ( 4.0 8.7
-22.2 ( 1.3 -22.1 ( 2.3 -15.6 ( 2.0 8.1
-26.8 ( 2.9 -50.4 ( 1.7 -11.8 ( 2.4 7.3
V-MeDEA
V-EtDEA
V-TEA
V-tDPDEA
V-cDPDEA
-30.6 ( 3.1 -83.6 ( 2.0
-6.0 ( 2.6 7.6
The large entropic component found for the vanadium(V)
complex V-cDPDEA (-83.6 J mol-1 K-1) may reflect solvent
organization around the complex. The two cis phenyl groups
on the R-carbons adjacent to the chelating nitrogen center are
too far apart for π-π stacking. The water molecules must
therefore interact with both phenyl groups in the V-cDPDEA
complex, whereas the trans phenyl groups in the corresponding
V-tDPDEA complex are close enough for some π-π stacking
interactions. We suggest that ligands with large hydrophobic
groups such as cDPDEA and tDPDEA are particularly sensitive
probes and that the entropic contribution to complex stability
can be modified by ligand design. The entropic term clearly is
crucial for the stability of this type of complex since the
protonated amine pKa value would have incorrectly predicted
both the quantitative and the relative stabilities of these two
complexes.
Previously, several observations made for related systems
documented the importance of both sterics and the pKa values
of the protonated amines on complex stability. For example,
in nitrobenzene solution, thermodynamic quantities for the
association reactions between vanadyl acetylacetonate and
substituted pyridines or substituted aliphatic amines are sensitive
to both steric effects and inductive effects.8 Similar corre-
sponding observations have been made with other metal
systems.13 In this work we show that thermodynamic properties
for the vanadium(V)-amino alcohol complexes are also being
governed by both the pKa of protonated ligand and sterics in
the ligand. Corresponding changes were not observed in the
lability of these complexes upon ligand changes.
DEAb
e
e
e
e
e
e
180.7( 0.8
-53.8 ( 0.2 8.7
MeDEAb
EtDEAb
TEAb
173.2 ( 0.5 -51.6 ( 0.2 8.5
175.3( 1.2
158.5( 0.7
132.6 ( 1.1 -39.5 ( 0.3 7.3
123.2 ( 0.2 -36.7 ( 0.1 7.6
-52.2 ( 0.4 8.7
-47.2 ( 0.2 8.1
tDPDEAb
cDPDEAb
a For the vanadium complexes the enthalpy (∆H°) and the entropy
(∆S°) were determined from the graph ln(Keq) versus 1/T containing
6-7 experimental points determined in triplicate using a concentration
of 10 mM vanadate and 20 mM ligand. The temperature range studied
was from 278 to 338 K. In all measurements, the ionic strength was
adjusted to 0.40 M using a solution of 4.00 M KCl at a pH near the
optimum pH ((0.05). ∆G° was obtained from the relationship ∆G° )
∆H° - T∆S°. b The thermodynamic parameters associated to the
protonation equilibrium of the ligands were determined from separate
experiments using the appropriate concentration of the amino alcohol
and otherwise the same condition as in the determinatiom of those
parameters for the complexes (pH and ionic strength). c The parameters
determined for deprotonation of dissolved 10 mM NaVO3 at 0.40
(adjusted with 4.00 M KCl) were for the enthalpy -5.6 ((1.0) kJ mol-1
the entropy 141 ((4) J mol-1 K-1, and free energy -47.8 ((2.1) kJ
mol-1 d The indicated errors in the thermodynamic parameters are
,
.
obtained as standard deviation on the resulting data and rounded up
from a 95% confidence limit (2σ). e The enthalpic contribution was
experimentally indistinguishable from 0 ranging from 0.01 to 0.005 kJ
mol-1 for all amino alcohol ligands.
To summarize, changes in the pKa values of the protonated
ligand are accompanied by corresponding changes in complex
stability. In addition increased hydrophobicity (i.e., steric bulk)
of the amine increases the entropic component which leads to
an overall decrease in complex stability and deviation from the
empirically found bell-curve-like dependence of the formation
constant of this type of complex.
Figure 5. Plots of ln(Keq) as a function of 1/T: (9) V-DEA, (0)
-
Conclusions
V-EtDEA, and (b) V-tDPDEA. The KaH VO of the deprotonation
2
4
-
reaction for H2VO4 used in the calculation of Keq of different V
complexes at 278 K was extrapolated from the experimental data
determined in the temperature range from 298 to 338 K.
The apparent formation constants for 12 vanadium(V)-amino
alcohol complexes were determined. The pH-dependent com-
posite formation constants as well as the pH-independent
equilibrium constants of the different vanadium(V) complexes,
when plotted against the pKa values of various protonated amino
alcohols display distinct maxima. This bell-curve is not to be
mistaken for the bell-curve that describes the stability of each
of these complexes individually and has previously been
characterized. The electron-donating ability as described by the
pKa values of the amine is not the only factor that influences
the stability of the vanadium(V) complexes; complex coordina-
tion number, sterics in the amine, and solvation were also found
to affect the vanadium complex stability. Determination of the
thermodynamic parameters of selected complexes showed that
the enthalpic component was the dominant term defining the
stability of these complexes and that the entropic component
change in the enthalpic term in these complexes is directly
reflected in the changes in complex stabilities. In a comparison
of the parameters for V-DEA with the parameters for V-TEA,
a similar trend is observed for the enthalpic component;
however, the stability of the V-TEA complex is lower since
the pKa value for this amino alcohol is significantly lower (we
note that a comparison of Kcom shows that more complex is
generated in such solutions). It is interesting that the enthalpic
(and entropic) component for V-EtDEA is significantly larger
than those of the other complexes (with the exception of
V-cDPDEA). Comparing the parameters for the V-MeDEA,
V-EtDEA, and V-TEA gives the opportunity to examine the
effect of the additional functionality in TEA on complex

8078 J. Am. Chem. Soc., Vol. 120, No. 32, 1998
Crans and Boukhobza
opposed this term. Modification of the amino alcohol increased
the entropic contribution and generally resulted in a decrease
in the overall stability of the complex from that observed for
the parent systems. Modification of the amino alcohol to
generate a ligand that forms a six-coordinate complex in general
generates vanadium(V) complexes that are more stable than the
corresponding five-coordinate complexes. On the basis of our
observations, we suggest that complex stability increases by
addition of alkyl substituents, if water soluble functionalities
are added in conjunction with the alkyl substituent and the pKa
value for the protonated amino alcohol remains high and above
8.
We conclude that structural modifications of amino alcohol
ligands do not significantly affect the lability of vanadium(V)
complexes as they change the vanadium(V) complex stability.
Acknowledgment. We thank the National Institute for
General Medical Sciences, the National Institutes of Health, for
funding this research (D.C.C.). We also thank Dr. A. I. Meyers
for providing us with samples of cDPDEA and tDPDEA. In
addition we thank Dr. A. D. Keramidas for access to his results
regarding the ligand HNBIDE and the corresponding vanadium
complex prior to publication. We also thank Dr. Sean S. Amin
for technical assistance, Drs. Alan S. Tracey, and Kenneth
Kustin for stimulating discussions.
Variable-temperature 1H and 13C NMR spectroscopy provided
little evidence for significant changes in complex lability despite
the range of differences in pKa values of protonated amine, in
sterics, and in solvation. The lack of sensitivity of the kinetics
of aqueous vanadium(V) complexes has previously been at-
tributed to a dissociative mechanism for complex formation.
Supporting Information Available: Plots of ligand chemi-
cal shift versus pH and T (5 pages, print/PDF). See any current
masthead page for ordering information and Web access
instructions.
JA9808200