Article

pubs.acs.org/IC

Titanium(IV) and Vitamin C: Aqueous Complexes of a Bioactive Form

of Ti(IV)
Katherine M. Buettner,† Joseph M. Collins,† and Ann M. Valentine*,‡
†
Department of Chemistry, Yale University, P.O. Box 208107, New Haven, Connecticut 06520-8107, United States
‡
Department of Chemistry, Temple University, 1901 North 13th Street, Philadelphia, Pennsylvania 19122, United States
*
S Supporting Information

ABSTRACT: Ascorbic acid is among the biorelevant ligands that render Ti(IV)
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

stable in aqueous solution. A series of pH-dependent titanium(IV) coordination

complexes of L-ascorbic acid is described. Directed by spectropotentiometric
methods, important aspects of the aqueous interactions in this system are
investigated, including ligand binding mode, pH-dependent metal−ligand
Downloaded via HANYANG UNIV on April 14, 2022 at 17:16:00 (UTC).

stoichiometry, and the importance of metal ion-promoted hydrolysis and the

binding of hydroxide. Stability constants are determined for all metal ion−ligand−
proton complexes by a process of model optimization and nonlinear least-squares
fitting of the combined spectropotentiometric titration data to the log βMLH values in the model. A speciation diagram is
generated from the set of stability constants described in the model. In the range pH 3−10, the aqueous speciation is
characterized by the sequential appearance of the following complexes as a function of added base: Ti(asc)20 → Ti(asc)32‑ →
Ti(asc)2(OH)22‑ → Ti(asc)(OH)42‑. These species dominate the speciation at pH < 3, pH 4−5, pH ∼ 8, and pH > 10,
respectively, with minimum log stability constants (β values) of 25.70, 36.91, 16.43, and −6.91. Results from electrospray mass
spectrometry, metal−ligand binding experiments, and kinetics measurements support the speciation, which is characterized by
bidentate chelation of the ascorbate dianion to the titanium(IV) ion via proton displacement, and a pH-dependent metal−ligand
binding motif of ligand addition followed by metal ion-promoted hydrolysis, stepwise ligand dissociation, and the concomitant
binding of hydroxide ion. Additionally, the kinetics of ligand exchange of titanium ascorbate with citrate are reported to
understand better the possible fate of titanium ascorbate under biologically relevant conditions.

■ INTRODUCTION
L-Ascorbic acid is an important molecule in both chemistry and
is invoked, characterized by monodentate or bidentate
coordination via the oxygen atoms on C2 or C3, with the
ligand singly or doubly deprotonated. Crystal structures are
biology, and its complexes with metals are of particular interest
available for metal ascorbate complexes of the soft ions
in both of these areas.1−3 Their study is complicated because thallium(I)2 and platinum(II),3,4 which employ various
the ligand has multiple protonation and oxidation states (Figure coordination modes.5,6 A multinuclear mixed-valent Fe(II)Fe-
1), because the metal−ligand complexes are in general not as (III)2 complex with the related tetramethyl reductic acid ligand
tight as one might expect for an ene-diol ligand, and because was characterized by X-ray crystallography and in solution.7
the complexes have mostly defied crystallographic character- One study reported bidentate binding on titanium dioxide
ization.1 In most cases, coordination via one of the modes 4−6 surfaces through binding mode 6, supported by IR, EXAFS, and
XANES data.8
Complexes of ascorbic acid with titanium(IV) should be
relatively strong,10 and not prone to the metal-catalyzed ligand
oxidation that renders many metal ascorbate complexes so
reactive. Accordingly, complexes of ascorbic acid with titanium,
first reported in 1936,11 have been used for detection and
quantitation of Ti(IV).12 Early work was reviewed by
Sommer,13 who combined spectroelectrochemistry and electro-
phoresis to address the pH-dependent speciation of titanium
ascorbate. The data supported the predominant species
Ti(Hasc)3+ (pH < 1), TiO(Hasc)+ (1 < pH < 1.6), TiO(Hasc)2
Figure 1. L-Ascorbic acid (H2asc, 1), doubly deprotonated ascorbic (1.6 < pH < 2.2), and Ti(asc)32‑ (2.5 < pH < 4.8). The latter
acid (asc2‑, 2) (pKa of C3-OH = 4.04 and pKa of C2-OH = 11.34),9 species, the one used for metal quantitation,12 was characterized
dehydroascorbic acid (3), and selected metal−ligand binding modes,
including monodentate (mono)protonated (4), bidentate (mono)-
protonated (5), and bidentate deprotonated (6). Received: July 16, 2012
Published: September 27, 2012

© 2012 American Chemical Society 11030 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039
Inorganic Chemistry Article

by a broad absorption at λmax = 360 nm, ε = 6800 M−1cm−1. terminal water ligands on the order of 3400 s−1 at 25 °C.19
The extinction coefficient was determined at a ligand/metal Titanium ascorbate injected into human blood serum at pH 7.4,
ratio of nearly 2000:1, and the ligand absorption tails into this for example, for an imaging application, where the ascorbate
region; it represents an upper limit.13 Minority species concentration is on the order of 50 μM,47,48 should undergo
TiO(Hasc)22‑ and Ti(asc)2 were also detected. Above pH 4.8, ligand exchange reactions with citrate, which is abundant in
the color disappeared, and hydroxo complexes were blood serum at 100 μM48 and makes quite stable titanium(IV)
suspected.13 This quantitative work agreed well with the complexes.49 Likewise, titanium ascorbate formulated near pH
original qualitative titrations.11 Another study invoked TiO- 5 and sprayed onto plants or fed to animals as a growth
(Hasc)+ (λmax = 355 nm, ε = 7800 M−1cm−1) and a second, promoter might encounter citrate or similar α-hydroxy acid
anionic complex with λmax = 375−385 nm that formed at lower competitor ligands. It remains to be evaluated how the rates of
ascorbate concentration and higher pH.14 Complexes formu- these exchanges, involving presumably bidentate ligands, might
lated as TiO(Hasc)2 (λmax = 364.4 nm, ε = 1202 M−1 cm−1) compare with the benchmark value for water exchange.
and TiO(OH)(Hasc) (λmax = 344.8 nm, ε = 692 M−1 cm−1) In light of the demonstrated bioactivity of titanium ascorbate,
were prepared and characterized by elemental analysis and IR the current work revisits these complexes, to further explore
and 1H NMR spectroscopies.15,16 The former complex is one this system. This work enumerates the important complexes
predicted by Sommer to occur at low pH; the latter, a and equilibria that define the aqueous speciation, particularly in
presumed hydrolysis product prepared by methanol extraction the most bioactive formulations. The ascorbate ligand binding
of the former, was not one of the species thought to mode is addressed. Finally, ligand exchange reactions are
predominate in aqueous solution.13 The IR spectra were not studied to understand the kinetics of such exchange under
reported below 1500 cm−1,16 so the presence or absence of a biorelevant conditions.

■
TiO stretch could not be ascertained. A complex formulated
as K2Ti(asc)3 was isolated from EtOH and characterized by EXPERIMENTAL SECTION
elemental analysis, but the ligand dianion was thought to be too
Reagents and Chemicals. All aqueous solutions were prepared in
basic to form in aqueous solution.17 Nanopure-quality water (18.2 MΩ cm resistivity; Barnstead), and all
The titanyl ion, TiO2+(aq), is often invoked in the aqueous chemicals were ACS grade or better. Stock solutions of 1.0 M
coordination chemistry of Ti(IV), though the importance of potassium chloride (KCl, J.T. Baker) and 0.2000 M potassium
discrete, monomeric TiO2+ units in solution has been hydrogen phthalate (KHP, ACS primary standard, Mallinckrodt) were
challenged.18 One study used 17O NMR spectroscopy to prepared in volumetric flasks by weighing out an appropriate amount
reinvestigate the issue of the aqueous titanyl ion.19 The “yl” of the salt, dissolving in water, and diluting to volume. Solutions of 0.1
oxygen of TiO2+(aq) exchanged 9 orders of magnitude faster M hydrochloric acid (HCl, 1 N volumetric solution, J.T. Baker) and
than the “yl” oxygen of VO2+(aq). This finding led to the 0.2 M potassium hydroxide (KOH, 1 N carbonate-free Dilut-It
conclusion that, at low pH and in the absence of any stabilizing Analytical Concentrate, J.T. Baker) were diluted from their respective
stock solutions volumetrically, and their concentrations were
ligand, a rapid equilibrium exists among TiO2+(aq), Ti- determined accurately (4 significant figures) by titrating the base
(OH)22+(aq), and Ti4+(aq). against KHP, and the acid against the base. Phenolphthalein (1% in
The bioactivity of titanium(IV) ascorbate is well established, 95% alcohol, Mallinckrodt) was used as an end point indicator, and the
but in general the complex is prepared in situ and not analyses were done in triplicate. All solutions were prepared fresh, and
characterized. In one early use of titanium ascorbate in were stored under an inert gas (nitrogen or argon) in parafilm-
medicine, a solution was used to treat tumors in humans.20 wrapped Nalgene bottles.
45
Ti ascorbate has been investigated as a radiopharmaceutical.21 For the spectropotentiometric titrations, the ascorbate concen-
Metal delivered as the complex by injection into rats bound to tration was constant and the Ti(IV) concentration was varied. An
accurately weighed amount (70.45 mg, 0.400 mmol) of L-ascorbic acid
albumin in plasma and accumulated in the spleen, liver, and crystals (99+ %, Sigma-Aldrich) was added to an argon-purged and
blood. 45Ti concentrated in and was used to image a glioma in foil-wrapped 100 mL volumetric flask containing an aqueous solution
the brain of a rat.21 More recently, ascorbic acid has been of 0.1 M KCl (aq). The titanium(IV) source for all experiments was a
shown to increase the uptake of titanium in rat everted gut neat solution of TiCl4 (l) (99.9%, with SureSeal cap, Sigma-Aldrich)
sacs.22 István Pais described the use of titanium(IV) ascorbate which is highly reactive with water and water vapor and was added to
as a water-soluble and relatively stable compound capable of the titration solutions by using a gastight glass syringe (Hamilton Co.).
stimulating the growth of plants.23 Several reviews have covered Following addition of metal ion (2−15 μL, 30−170 μmol), a 50.0-mL
the research in this area.24−27 The generally beneficial effects of aliquot of this solution was transferred to the titration vessel by using a
titanium solutions on various plants, including peppers,28−30 volumetric pipet. For other applications, a 0.1000 M stock solution of
L-ascorbic acid was prepared volumetrically from the acid crystals and
wheat,31,32 oats,33 apples,34,35 and plums,36,37 have been an stored under inert gas; additionally, this solution was stored in a foil-
important focus. Titanium ascorbate was patented as a growth wrapped bottle to prevent photo-oxidation. The Ti(IV) concentration
promoter for plants,38 animals,39 or both.40 Other patents cover was determined from the remaining 50-mL titration solution by flame
its preparation as a solid41 or for aquaculture applications42 or atomic absorption analysis (SpectrAA-20 flame atomic absorption
as a component of a complex fertilizer mixture.43−46 The spectrometer, Varian) by using the method of internal standards.
formulation applied to organisms is not always described in For kinetic experiments, stock solutions of 10 mM titanium
great detail, but it often comprises titanium(IV) with at least a ascorbate were made by dissolving ascorbic acid crystals (0.35 g, 2.0
20-fold excess of ascorbate, with the pH adjusted to between 5 mmol) in 0.5 M NaCl. To this solution, neat TiCl4 (25 μL, 250 μmol)
and 7. The earlier work on titanium ascorbate speciation does was added, and allowed to mix for 1 h. The pH of the solution was
adjusted to either 7.4 or 4.5 by using NaOH. Individual samples were
not address the complex speciation under these conditions, but diluted immediately before kinetics measurements to 125 μM by using
instead focuses on lower pH values. 0.5 M NaCl. A stock citric acid solution was made by dissolving citric
Once administered into a biological system, titanium acid monohydrate (2.6 g, 12 mmol) in 0.5 M NaCl, and the pH was
ascorbate complexes may undergo ligand exchange reactions. adjusted to either 7.4 or 4.5 using NaOH. Further dilutions were made
Titanium(IV) is a labile metal ion, with exchange rates for from this solution to give the range of concentrations used.

11031 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039

Inorganic Chemistry Article

Spectropotentiometric Titrations. All titrations were conducted

as described previously,49 according to the general methods
described,50,51 with minor modifications. The entire assembly was
foil-wrapped to prevent any photo-oxidation of the ligand, and kept
under positive pressure from argon to prevent oxygen contamination.
The total number of equilibrated titration points for which data were
collected was between 70 and 90. The fitting software was Specfit/
32.52−56 Full details of the experimental procedure are given in the
Supporting Information.
Metal−Ligand Binding Stoichiometry. Ligand-to-metal stoi-
chiometry at pH 4.5 was determined independently by using an
aqueous mole ratio method, probed by UV−vis spectroscopy.
Repeated volumes of TiCl4 (l) were added to a solution of 4.00
mM H2asc in 0.1 M KCl under the same conditions discussed above
for the titrations. The pH was readjusted to 4.5 for each successive
solution by adding acid or base, and total solution volume was
monitored. Following each equilibration, a 1.0 mL aliquot was
removed, and its absorbance was measured. After data collection, the
remaining solution was used for titanium concentration analysis. For
each point, an L/M molar ratio value was determined, and the
corresponding absorbance at 366 nm, corrected for dilution and
reported as an observed molar (Ti(IV)) absorptivity, was calculated.
Mass Spectrometry. Electrospray mass spectra were collected in
positive ion mode on a Waters/Micromass ZQ spectrometer. The
pertinent voltages were as follows: capillary, 3 kV; cone, 20 V;
extractor, 3 V; RF lens, 0.3 V. The aqueous samples were injected
directly, and molecular ions were detected in a Waters 2996
Photodiode Array Detector.
Kinetics. Citric acid solutions were prepared in 0.5 M NaCl at
concentrations ranging from 3.75 to 250 mM to ensure pseudo-first- Figure 2. (A) Potentiometric titration curves, as a function of molar
order kinetics. The final solutions contained 62.5 μM titanium equivalents of base added, for 4.00 mM L-ascorbic acid in the presence
ascorbate and between 1.88 and 125 mM citrate. of various concentrations of Ti(IV). (B) Ligand deprotonation (ZL)
Kinetics measurements were made by using a Cary 50 Bio UV−vis curves, as a function of pH, for 4.00 mM L-ascorbic acid in the
spectrophotometer. The absorbance was measured every 6 s between presence of various concentrations of Ti(IV). In both panels, data are
300 and 600 nm, or every 0.6 s at 370 nm for single wavelength as follows: no symbol, ascorbic acid alone with no titanium; open
measurements. Longer time scale experiments were performed over 20 circles (○), [Ti(IV)] = 0.297 mM (L/M = 13.5); filled circles (●),
h to confirm complete conversion to the citrate species. These [Ti(IV)] = 0.530 mM (L/M = 7.5); open triangles (Δ), [Ti(IV)] =
measurements were made every 6 s for the first 2 min, every 30 s for 1.058 mM (L/M = 3.8); filled triangles (▲), [Ti(IV)] = 1.665 mM
the next 8 min, and then every 10 min for the remainder of the 20 h. (L/M = 2.4). For clarity, in each case, only every fifth data point is
The pH was unbuffered except by the solution components, but was symbolized. Dashed lines represent equivalents of acidic ligand
constant to within ±0.1 pH unit over the course of the reaction. protons. All titrations were conducted at 25 °C, in 0.1 M KCl, and
Data were fit by using Specfit/32.53−56 Data at pH 7.4 were best fit under an argon atmosphere.
to a sequential two step model. Data at pH 4.5 were best fit to a
sequential three step model. All data were collected at least in
triplicate; error bars represent the standard deviations of all trials. the basis of both its position relative to the titration of ligand

■
alone and the steepness of the inflection region. For titrations
involving a diprotic acid such as L-ascorbic acid, having pKa
RESULTS AND DISCUSSION values that widely straddle the isopotential pH of water (pH
L-Ascorbic Acid: Protonation States and Oxidation. 6.90 in the presence of 0.1 M supporting electrolyte and at 25
The spectropotentiometry of ascorbic acid in the absence of °C) are particularly useful, because the region of inflection is
metal was performed, in some cases with deliberate ligand isolated from the two ligand deprotonations. Furthermore, in
oxidation (Figures S1−S3), so that the latter could be the absence of metal-ion promoted hydrolysis and subsequent
recognized and avoided. In the absence of oxidation, the hydroxide binding by the metal ion, the inflection region will be
measured pKa values and species spectroscopic properties as steep as that seen for the titration of ligand alone.
agreed with those reported in the literature.9 Exactly 1 molar equivalent of base was needed to reach the
Potentiometric Data Analysis. By examining the unfitted inflection point in the titration of ligand alone (Figure 1A, no
potentiometric data for a series of spectropotentiometric symbols). As metal ion was introduced in subsequent titrations
titrations varying only in ligand-to-metal (L/M) molar ratio, at constant ligand concentration, the position of the inflection
several important points relating to the aqueous interactions of was shifted along the x-axis, reflecting the extra base needed to
Ti(IV) with L-ascorbic acid were established. The raw consume the additional protons in solution that resulted solely
potentiometric data can be presented as plots of pH as a from the presence of metal ion. A comparison was made
function of equivalents of base added to a given amount of between experimental and predicted inflection point values,
ligand in the presence of varying amounts of metal ion, or as with the former determined from the graph and the latter
ligand deprotonation curves showing the pH-dependent calculated on the assumption that the metal ion was
removal of acidic ligand protons for the same data (Figure coordinatively saturated by anionic oxygen donors originating
2). In the former case the most important observations from water or ascorbate. Because the calculated values
concerned the inflection point for the various titrations. An accounted for the 1 equiv of ligand proton consumed by
inflection point in a metal ion−ligand titration is informative on added base, but did not differentiate between a proton
11032 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039
Inorganic Chemistry Article

(presumably the O(2) hydroxyl proton) of ascorbate versus [H+]bound to L = [H+]total − [H+]free (4)
water as the source of additional protons, the analysis was
limited by an inability to distinguish between species related to and
two limiting cases: Ti(Hasc)3(OH)32‑ and Ti(asc)32‑. This [H+]total = n[H nL]total − [OH‐]added (5)
comparison can be described by the following equations:
In cases where the L/M ratio was greater than 3, the curves
equiv H+ titrated to midpoint, experimental thus generated (Figure 2B), when compared to the curve for
mol base added to midpoint the ligand alone, were uniformly shifted to lower ZL values as a
= function of added Ti(IV), and were essentially parallel to the
mol ligand (1) curve for the ligand alone. This finding is an indication
primarily that their speciations are similar. This observation is
equiv H+ titrated to midpoint, calculated reinforced by comparison with the curve for the L/M 2.4
3 titration, which has a different overall shape from the others,
=1+ and which crosses into negative ZL territory at pH ∼ 4. Both of
L/M molar ratio (2)
these differences point to a situation where the lack of sufficient
The experimental and calculated midpoint values for the three ligand for tris-coordination of the metal ion becomes a crucial
curves (Figure 1A) describing titrations involving a minimum factor in understanding the speciation of the system.
of a 3-fold excess of ligand were in close agreement (exp/calc To address the inability of the above plots to discriminate
for L/M 13.5, 7.5, 3.8 were 1.22/1.22, 1.44/1.40, 1.86/1.90, conclusively between bidentate metal−ligand binding with
respectively). While this finding supports the contention (see proton displacement and metal-promoted hydrolysis and
below) that, in the region before inflection, and in the presence binding of hydroxide in mixed metal−ligand−hydroxo
of sufficient ligand for tris-coordination, the prominent metal− complexes in the coordination sphere of Ti(IV), a further
ligand complex is the aqueous Ti(asc)32‑ species, it confirms graphical tool was employed. This tool, referred to as a Bjerrum
only that, in acidic aqueous solution, Ti(IV) complexation plot or n-plot, defines a function n as the average number of
involves octahedral coordination of anionic oxy-donors. The ligands bound per metal ion.51 The analysis utilizes total ligand
titration at L/M ratio of 2.4, where insufficient ligand precluded and metal concentrations (Ltotal and Mtotal, respectively), base
the complete tris-coordination of ascorbate by metal ion, gave concentration ([base]), initial solution volume (Vinitial), and the
values not in agreement, with the experimental value conditional pKw and pKa values as known quantities, and can be
significantly higher than the calculated (exp/calc = 2.48/ expressed in terms of a diprotic acid ligand:
2.25). This finding indicates that, in the presence of insufficient
ascorbate for tris-coordination, coordinative saturation is n(pH, v base , mf H2asc, mf Hasc−)
maintained, and metal-ion-promoted hydrolysis and binding v base·[base]
2L total − − 10−pH + 10 pH − pK w
of hydroxide has occurred, with the loss of at least one bound L total −
Vinitial + v base
2·mf H2asc + mf Hasc−
ligand. =
The argument for metal-ion-promoted hydrolysis and M total (6)
binding of hydroxide in titrations lacking sufficient ligand for
tris-coordination is supported by several further observations. The experimental quantities determined for each titration point
In the case of the L/M 2.4 titration, from the fact that the include the pH-dependent mole fractions for the protonation
inflection fell outside of the limit of titratable ligand protons states of the ligand (mfH2asc, mfHasc‑), the pH, and the volume of
(>2); it follows that because the sole source of protons in this titrant added (vbase). A derivation of the equation used to
region is water, these titratable protons were necessarily generate these plots can be found in Supporting Information,
produced by hydrolytic binding of hydroxide by the metal Derivation S1. The n-curves for a particular metal−ligand
ion. Also, as mentioned above, for an aqueous titration in the system can be shown mathematically to coincide independent
presence of supporting electrolyte and an acid ligand such as of total metal and ligand concentration (and L/M ratio)
ascorbate the inflection region should be both sharp and steep. provided that the system of interest is characterized by a series
The presence of metal ion, however, alters this behavior. The of ML−MLx complexes. If, on the other hand, the aqueous
trace for the L/M 2.4 titration, for instance, and to a lesser speciation is more complicated, including ternary metal−
extent for the L/M 3.8 titration, exhibited some buffering in the hydroxo−ligand species and protonated metal−ligand species,
inflection region, which can be attributed to the hydrolytic the assumptions contained in the n-plot algorithm are
activity of the metal ion. invalidated, and the curves will be unique. Finally, because n
The ZL plot, adapted from the work of Ö hman,51 is another will likely have an upper limit of 3 for the Ti(IV)−ascorbate
convenient tool for visualizing the average number of protons coordination system, values exceeding this limit reflect, as noted
bound per ligand molecule as a function of pH, and, by above, portions of the pH range for which simple ML−MLx
extension, for identifying regions where mixed ligand−hydroxo complexes are of minor importance.
complexes are formed, pure Ti(IV) hydroxo species having For each experimental titration data pair (volume of base
been shown, in previous work,49 to be insoluble above pH ∼ 3 added, pH) comprising a given titration, the n-value was
at the millimolar concentrations used here. The function ZL was generated and plotted against log[L2‑], which was independ-
derived from the potentiometric data by using the following ently derived from the speciation of the ligand alone at the
equation: particular pH. The results are shown in Figure 3. Three of the
curves, all representing titrations with >3-fold ligand excess,
[H+]bound to L were very nearly identical, flattening at n ∼ 3 before curving
Z L(pH, added base) = upward again above pH 7. The fourth curve, representing the
[HnL]total (3)
L/M 2.4 titration, was distinct from the other three, with
where incomplete flattening at n ∼ 2.5. These observations are
11033 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039
Inorganic Chemistry Article

Figure 3. Bjerrum plots, also referred to as n-plots, describing

equivalents of metal-bound asc2‑ as a function of log[asc2‑] (bottom
axis), with corresponding pH values (top axis), for titrations of 4.00
mM L-ascorbic acid in the presence of varying amounts of Ti(IV).
Symbols as follows: open circles (○), [Ti(IV)] = 0.297 mM (L/M =
13.5); filled circles (●), [Ti(IV)] = 0.530 mM (L/M = 7.5); open
triangles (Δ), [Ti(IV)] = 1.058 mM (L/M = 3.8); filled triangles (▲),
[Ti(IV)] = 1.665 mM (L/M = 2.4). For clarity, only every fifth data
point is symbolized.

evidence that (1) between pH 3 and 6, and in the presence of

at least a 3-fold excess of ligand, the aqueous speciation of
Ti(IV) ascorbate is characterized by the binding of a third
equivalent of asc2‑; (2) above pH ∼ 6 ternary metal−ligand−
hydroxo species begin to dominate, regardless of L/M ratio;
and (3) displacement of the O(2) proton leading to bidentate
binding of asc2‑ is the characteristic aqueous binding mode of
the ligand.
Spectral Data Analysis. The UV−vis absorbance changes
associated with the titrations were examined independently. For
Figure 4. Selected UV−vis absorbance spectra for the spectropo-
titrations spanning L/M 2−15, the pH-dependent trends were tentiometric titration of 4.00 mM L-ascorbic acid in the presence of
similar, and reflected in several key ways the conclusions of the 0.297 mM Ti(IV) (L/M = 13.5). Ranges as follows: (A) pH 2.9−4.1,
potentiometric analysis. (B) pH 4.1−4.8, (C) pH 4.8−5.8, (D) pH 5.8−8.8, (E) pH 8.8−10.0.
The pH-dependent UV−vis behavior for one titration (L/M All traces were normalized against a baseline of 0.1 M KCl (25 °C).
13.5) is shown in Figure 4. Overall, the data agree well with the For clarity, only six traces are shown per panel out of a total of 76
qualitative trends described by Ettori.11 From pH 3−4 (Figure traces collected. Dashed line is at 366 nm. Dotted line is at 348 nm.
4A), a steady increase in the absorbance at 366 nm was
observed. This trend supports the conclusions of the Ti(asc)32 ‐ + 2H 2O → Ti(asc)2 (OH)2 2 ‐ + Hasc‐ + H+
potentiometric data, specifically the n-plots, which supported (8)
for this region the following binding event:
2‐
Ti(asc)2 (OH)2 + 2H 2O
Ti(asc)2 0 + Hasc ‐ → Ti(asc)32 ‐ + H+ (7)
→ Ti(asc)(OH)4 2 ‐ + Hasc ‐ + H+ (9)
The absorbance then remained unchanged from pH 4−5
(Figure 4B), reflecting the stability of the putative tris-ligand This qualitative description of the aqueous speciation of
complex in this region. In Figure 4C, encompassing the range Ti(IV) with L-ascorbic acid was further investigated by plotting
pH 5−6, the λ = 366 nm absorbance began to decrease slightly, the observed 366 nm molar absorptivity profiles as a function of
an indication of the onset of metal-promoted hydrolysis and pH for each of the L/M ratios considered (Figure S4). The
ligand loss. Finally, in the region pH 6−10 (Figure 4D,E), the most striking observation was that, for the three titrations with
decrease was more dramatic, and was accompanied by a shift in L/M > 3, a clear absorbance maximum was seen at pH 4.5. The
the LMCT absorbance to 348 nm. This result is significant L/M 2.4 titration peaked slightly before pH 4.5 due to the
because it agrees with the λmax proposed for a ternary Ti(IV)− inability of the metal ion to coordinate fully a third ascorbate
ascorbate−hydroxo species,16 and it supports the findings of the ligand. Additionally, the apparent molar absorptivity at a given
n-plots regarding metal-promoted hydrolysis and replacement pH increased as L/M increased. This correlation can be
of bound ligand by hydroxide. The binding events proposed for explained by considering the effect of increasing L/M ratio on
this region can be summarized by the following equations: any aqueous speciation. It can be seen that, as L/M increases,
11034 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039
Inorganic Chemistry Article

both of the following equilibria will be forced to the right in of L/M ratio. Looking at the values generated for L/M > 3, this
response to the conditions (charges omitted for simplicity): trend was linear for each particular species (Figure S9),
allowing for the reporting of stability constants with reference
M + large excess L → ML 2 (10)
to a singular L/M ratio. Because a minimum 10-fold ligand
M + large excess L → ML 3 (11) excess is a common condition for studies of this type, and
because Figure S9 showed that asymptotic linearity began at L/
However, if the two equilibria are in direct competition over a M ∼ 10, this ratio was chosen for reporting the log β values
certain pH range, as they are in the speciation model being characteristic of this system (Table 1). When applied to the
constructed, the second equilibrium will be favored. By this spectropotentiometric data, these results produced a speciation
reasoning, the Ti(IV) tris-ascorbate complex will dominate the that was consistent in all respects with the benchmark findings
region around pH 4.5 as L/M ratio increases by excluding the used to direct the fitting process.
neighboring species; this phenomenon is seen in the The UV−vis absorbance characteristics were consistent with
experimental data as an increase in the observed absorptivity those predicted from the various analyses of the spectral data,
of the species whose wavelength is being monitored. Finally, and with the UV−vis characteristics of the species predicted by
the pH 4.5 absorbance profiles were investigated as a function Jabs et al.,16 with binary metal−ligand and ternary metal−
of L/M ratio (Figure S5). The results concurred with the above ligand−hydroxo species exhibiting different λmax values and
discussion in that the effect of increasing L/M ratio from 1 to extinction coefficients based on the numbers of bound ligand
12 changed the speciation, seen in the shift of the λmax from 348 and hydroxide (Table 2). Viewed graphically (Figure 5), these
to 366 nm, and in the increasing εobserved, without changing the
equilibrium model. Table 2. Spectral Characteristics of Aqueous Complexes in
pH-Dependent Speciation of Aqueous Titanium(IV)− the Ti(IV) Ascorbate System Predicted for L/M = 10
Ascorbate Complexes. The data were analyzed first by ([H2asc] = 4.00 mM) in 0.1 M KCl (25 °C, Argon
model-free evolving factor analysis (EFA) (Figures S6 and S7). Atmosphere)
Specific speciation models were proposed, on the basis of the
considerations outlined above (Figure S8) and complemented species λmax (nm) εmax (M−1 cm−1)
by modeling of the data. Details of proposed models are Ti(asc)20 370 1260
outlined in the Supporting Information. The optimized fit Ti(asc)32‑ 365 1660
returned the following species and stability constants (Table 1). Ti(asc)2(OH)22‑ 348 1060
Ti(asc)(OH)42‑ 362 700
Table 1. Minimum Stability Constants Determined for H2asc 243 9800
Aqueous Ti(IV)−Ascorbate−Hydroxo Species Using L/M = Hasc− 266 15 100
10 ([H2asc] = 4.00 mM) in 0.1 M KCl (25 °C, Argon asc2‑ 299 9700
Atmosphere)
equilibrium log βmin
Ti4 + + 2asc 2 ‐ ⇆ Ti(asc)2 0 25.70
Ti4+
+ 3asc 2‐
⇆ Ti(asc)3 2‐ 36.91
Ti4+
+ 2asc + 2H 2O ⇆ Ti(asc)2 (OH)2
2‐ 2‐
+ 2H + 16.43
Ti4 + + asc 2 ‐ + 4H 2O ⇆ Ti(asc)(OH)4 2 ‐ + 4H+ −6.91

Because free Ti4+(aq) was not detected even at the lowest

pH values, these β values are best considered minimum values.
Their relative magnitudes are firm, however. An important
aspect of a series of stability constants generated for a particular
system is the confluence of these constants over a range of L/M
values. Assuming that the model is unchanged in the range of
conditions employed, the values optimized for each independ-
ent data set should produce the same speciation based on the
same log β values. This condition depends on the magnitude of
the stability constant(s), and the hydrolytic propensity of the
metal ion. For a strong complexation with a low propensity for
metal ion-promoted hydrolysis, the log β values will be virtually
identical regardless of L/M ratio. The opposite is true for values
optimized in this system, incorporating relatively weak binding
with a strong propensity toward hydrolysis (Table S1 and
Figure S9). The reasoning is a reflection of the relative
concentrations of metal ion and ligand with respect to pH, Figure 5. (A) Speciation diagram and (B) corresponding absorbance
wherein solutions with L/M < ∼50 will be differentially prone spectra for spectropotentiometric titration of 4.00 mM L-ascorbic acid
to ligand binding and hydrolysis.57 This effect was seen in the in the presence of 0.297 mM Ti(IV) fitted to optimized L/M = 10 log
log β values optimized for the Ti(IV) ascorbate system at β values. Data as follows: open circles (○), Ti(asc)20 (β120); filled
different L/M ratios (Table S1), for which slightly different circles (●), Ti(asc)32‑ (β130); open triangles (Δ), Ti(asc)2(OH)22‑
values were determined for each particular species as a function (β12−2); filled triangles (▲), Ti(asc)(OH)42‑ (β11−4).

11035 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039

Inorganic Chemistry Article

results, applied to the L/M 13.5 titration data, produced a Extrapolation of the linear portions of the data showed an
speciation and absorbance paradigm that could be generalized intersection at almost exactly L/M 3, with an extinction
in form to all L/M ratios. Furthermore, the log β values coefficient in close agreement with that predicted from an
generated very similar speciation diagrams for all L/M > 3 data examination of the raw spectral data (1630 M−1 cm−1).
sets, changing in a predictable way below this ratio for the Therefore, in addition to predicting the relatively weak binding
reason of insufficient ligand for tris-coordination (Figure S10). interaction between metal ion and ligand, this technique
These findings correspond well with the conclusions of the supported a Ti(IV) tris-ascorbate species in the aqueous
present study, in that, for the pH range 2.5−10, and in the speciation of the system. Additional experiments were
presence of excess ligand (ascorbate), no evidence was found conducted at pH = ∼2, ∼6, and ∼9 (data not shown). The
for the titanyl ion. This result can be explained by considering asymptote at high L/M was not sufficiently horizontal to allow
the lability of the “yl” oxygen, and the resulting equilibria quantitation, but qualitatively, the L/M ratio for binding in each
between the titanyl ion, the bis-hydroxo species, and the of these regions was clearly less than 3.
aqueous metal ion. Specifically, as discussed below, the most Electrospray mass spectrometry (ESI-MS) offered an addi-
acidic species described in this investigation, the Ti(asc)20 tional means of probing the formation of the tris-coordinated
complex, was not anionic according to the potentiometric species. Several processes, including ligand oxidation, degrada-
and fitting results; a putative Ti(O)(asc)22‑ species was tion, and protonation, and metal precipitation (as TiO2), were
therefore ruled out. Furthermore, the binding of a third difficult to avoid in the instrument despite the “soft” ionization
ascorbate ligand was not found to be accompanied by proton that is the hallmark of electrospray; these constraints limited
consumption, which would be the case if either Ti(O)(asc)22‑ the pH range and L/M ratios under which results could be
(aq) or Ti(OH)2(asc)22‑ (aq) were in significant concentration obtained. At pH = ∼3 and L/M = ∼7.5, with low cone and
at this pH. On the contrary, the binding event in this pH region capillary voltages, however, a spectrum was detected in positive
was found to entail proton liberation from the ascorbate ion mode depicting, based on m/z+ and isotope distribution, a
monoanion. The combination of these characteristics led to the Ti(IV) tris-ascorbate species (Figure 7).
conclusion that titanyl species are not prevalent in the pH
values surveyed.
Metal−Ligand Stoichiometry. Several further analytical
procedures were undertaken to substantiate the existence of the
tris-coordinated titanium(IV) ascorbate species. In one experi-
ment a solution of ascorbic acid was maintained at pH 4.5 as
aliquots of TiCl4 (l) were added. After each addition, the pH of
the solution was readjusted to 4.5, and the solution was allowed
to re-equilibrate. UV−vis scans were then taken by using a
cuvette, and adjusted ligand and metal concentrations were
determined. The resulting L/M ratios were plotted against
observed molar absorbance at 366 nm, giving the plot shown in
Figure 6. The shape of the curve is indicative of a relatively
weak binding interaction, and points to the potential for
hydroxide to compete with ascorbate for metal binding.

Figure 7. Portion of an electrospray mass spectrum (positive ion

mode) for an aqueous solution of Ti(IV) ascorbate (L/M = ∼7.5, pH
= ∼3). Peaks centered on m/z = 573 are consistent with the
formulation H3[Ti(asc)3]+ (C18H21O18Ti) in terms of mass and
isotope distributions (dotted trace). Peak at m/z = 567 is similarly
consistent with the formulation K(H2asc)3+ (C18H24O18K).

Rates of Ligand Exchange. The binding of Ti(IV) by

ascorbate is sufficiently strong to prevent hydrolytic precip-
itation, but weaker than binding by other common biological
ligands such as citrate.10,49 Speciation models suggest that
Ti(IV) ascorbate complexes introduced to blood serum, for
example,58 would undergo ligand exchange with citrate. The
rapid exchange rates reported for Ti(IV) using water as a
benchmark ligand (>1000 s−1)19 suggest that this exchange
among biomolecules might be very fast. But other data, such as
Figure 6. Mole ratio method for ligand binding stoichiometry in the the kinetics of delivery of Ti(IV) from citrate to the metal
aqueous Ti(IV) ascorbate system at pH 4.5, monitored by the
absorbance at 366 nm. Experiments were conducted at 25 °C and
transport protein transferrin,59 suggest that exchange kinetics
under an argon atmosphere by adding aliquots of neat TiCl4 (l) to a might be much slower, on the order of minutes or hours.
solution containing 4.00 mM L-ascorbic acid and 0.1 M KCl. The pH The competition for Ti(IV) binding between ascorbate and
4.5 ligand-alone absorbance trace was subtracted from each, and all citrate was studied under pseudo-first-order conditions of
traces were corrected for dilution. The two solid lines intersect at L/M excess citrate. Ligand exchange was probed by monitoring the
= 3.009 and ε(366 nm) = 1632 M−1 cm−1. Dashed line is at L/M = 3. disappearance of absorbance due to the Ti(IV) ascorbate
11036 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039
Inorganic Chemistry Article

complexes at 355 nm after citrate was introduced. Experiments match well with titanium species having two, one, and no
were carried out at the physiological serum pH of 7.4 and also ascorbate ligands. At pH 4.5, a tris-ascorbate species dominates,
at pH 4.5, near the pH of formulation of the patented growth- and the UV−vis spectra of the kinetic species match well with
promoting complexes, where the tris-ascorbate species is titanium species having three, two, one, and no ascorbate
predicted to predominate. The dominant species predicted ligands. The rate constants at the two pH values are remarkably
under these two conditions are49 similar: the first ascorbate ligand in the tris-ascorbate species
exchanges at ∼0.2 s−1 (t1/2 ∼ 3 s), the penultimate ascorbate at
Ti(asc)2 (OH)2 2 ‐ + 3Hcit 3 ‐ ∼0.06−0.08 s−1 (t1/2 ∼ 10 s), and the last ascorbate at ∼0.006
→ Ti(cit)38 ‐ + 3Hasc ‐ (at pH 7.4) (12) s−1 (t1/2 ∼ 2 min). The last kinetic step does not exhibit a
concentration dependence on citrate, which is consistent with a
Ti(asc)32 ‐ + 3H 2cit2 ‐/H3cit ‐ dissociative mechanism. At pH 4.5, for example, eq 14 is
consistent with the observed data.
→ Ti(Hcit)2 (H 2cit)4 ‐ + 3Hasc‐/H 2asc (at pH 4.5)
(13)
The data at pH 7.4 fit best to two sequential rate constants
(Table S2). The first rate constant was dependent on citrate
concentration (Figure 8A) and saturable, yielding a dissociation

At pH 4.5, citrate exists as a mixture of predominantly the

H2cit2‑ (57%) and H3cit− (39%) forms, whereas ascorbate exists
as Hasc‑ (73%) and H2asc (27%). These species provide
buffering for protonation changes during ligand exhanges. The
starting and ending species are those that predominate at pH
4.5 under the solution conditions. The intermediate species are
highly speculative, but consistent with the data.
These data serve as a model of processes that may be
occurring in blood serum where the ligands studied in the
exchange are in relatively high concentrations. The Kd values
for the citrate-dependent steps are far higher than the
physiological serum citrate concentration of 0.1 mM,48 and
so this model predicts that in serum the conversion of titanium
ascorbate to titanium citrate complexes would occur on the
order of minutes or longer, and that the rates would depend
strongly on the citrate concentration. Physiological citrate
concentration varies in certain disease states such as epilepsy
Figure 8. Observed rate constants for the disappearance of the and renal failure.60,61
Ti(Hasc)2(OH)22‑ or Ti(asc)32‑ absorbance as a function of citrate The time scale of this exchange is far slower than the rates
concentration. Error bars represent standard deviations among at least reported for monodentate ligands, the exchange of which
three trials. (A) Data at pH 7.4. (●) k1: kmax = 0.081 ± 0.007 s−1, Kd = showed no concentration dependence on the incoming ligand,
10 ± 4 mM (■) k2 = 0.0060 ± 0.009 s−1. (B) Data at pH 4.5. (●) k1: and had forward rate constants ranging from kobs = 4 s−1 to kobs
kmax = 0.22 ± 0.02 s−1, Kd = 18 ± 5 mM (■) k2: kmax = 0.055 ± 0.002 > 100 s−1.19,62 The faster rates are likely due to the difference in
s−1, Kd = 5.4 ± 1.7 mM (⧫) k3 = 0.0056 ± 0.0008 s−1.
the nature of the ligands binding to the Ti center. In the current
case, ascorbate having lost one or both M−L bonds might
constant of approximately 11 mM for the first detectable step of remain associated with the complex, rebinding rather than
exchanging for citrate.

■
the exchange. The second rate constant was not dependent on
citrate concentration. For experiments carried out at pH 4.5,
three steps were necessary to fit the data. The first two were CONCLUSIONS
saturable, with apparent Kd values of 18 mM and 5.4 mM for k1 The aqueous speciation model determined for the Ti(IV)−
and k2 (Figure 8B). The third did not depend on citrate ascorbate complexation system under the reported conditions
concentration. gave considerable insight into several important facets of the
The observation that, at 7.4, two distinct rate constants were binding. The binding mode was suggested to feature bidentate
detectable by UV−vis, whereas at pH 4.5, three steps were chelation, with the deprotonated oxy-groups at C2 and C3 as
detectable may be related to the titanium ascorbate speciation the donors. This finding supports the stipulation that a hard
at those pH values. For the former, a bis-ascorbate species metal ion like Ti(IV) is capable of out-competing protons for
predominates, and the removal of each ascorbate may be the binding of these hard donor atoms. Additionally, direct and
detectable in distinct steps. Consistent with this interpretation, indirect evidence for a Ti(IV) tris-ascorbate species at pH =
the absorbance change for each step is comparable, and the ∼4.5 was described. The question of metal-promoted
UV−vis spectra of the kinetic species modeled by Specfit/32 hydrolysis was investigated. The results indicated that, even
11037 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039
Inorganic Chemistry Article

while coordinative saturation was maintained, binding of (11) Ettori, J. C. R. Hebd. Seances Acad. Sci. 1936, 202, 852−854.
hydroxide was indicated below physiological pH for this (12) Hines, E.; Boltz, D. F. Anal. Chem. 1952, 24, 947−948.
metal ion. Titanium ascorbate complexes convert smoothly to (13) Sommer, L. Collect. Czech. Chem. Commun. 1963, 28, 449−462.
titanium citrate complexes under physiological concentrations (14) Susic, M. V. Bull. Boris Kidric Inst. Nucl. Sci. 1963, 14, 125−133.
(15) Jabs, W.; Gaube, W. Z. Anorg. Allg. Chem. 1984, 514, 179−184.
of the ligands in multistep processes, with exchange times on (16) Jabs, W.; Gaube, W. Z. Anorg. Allg. Chem. 1984, 514, 185−195.
the order of minutes. Taken together, the results reported here (17) Jabs, W.; Gaube, W.; Fehl, C.; Lukowski, R. Inorg. Chim. Acta
delineate many thermodynamic and kinetic features of one of 1990, 175, 273−276.
the longest-known and most bioactive forms of Ti(IV).

■
(18) Clark, R. J. H. The Chemistry of Titanium and Vanadium;
Elsevier: Amsterdam, 1968.
ASSOCIATED CONTENT (19) Comba, P.; Merbach, A. Inorg. Chem. 1987, 26, 1315−1323.
*
S Supporting Information (20) Arloing, F.; Morel, A.; Josserand, A. C. R. Hebd. Seances Acad.
UV−vis spectra of ascorbic acid protonation states; potentio- Sci. 1937, 204, 824−825.
metric plots and UV−vis spectra depicting oxidation of ascorbic (21) Kawamura, M.; Ido, T.; Ishiwata, K.; Inoue, K.; Kimura, S.;
acid; mathematical derivation of n-plot; graphical displays Matsuda, K.; Kawashima, K.; Kameyama, M. J. Labelled Compd.
showing effect of L/M ratio on λ(366 nm) absorbance as a Radiopharm. 1986, 23, 1360−1362.
function of pH, and on pH 4.5 absorbance as a function of (22) Aarabi, M. H.; Moshtaghie, A. A.; Mirhashemi, M. Pak. J. Biol.
wavelength; graphical representation of EFA smoothing; Sci. 2011, 14, 945−949.
(23) Pais, I.; Feher, M.; Farkas, E.; Szabo, Z.; Cornides, I. Commun.
comparison of EFA profiles at different L/M ratios; speciation
Soil Sci. Plant Anal. 1977, 8, 407−410.
diagram comparison of 3- vs 4-species model; table and graph (24) Pais, I. J. Plant Nutr. 1983, 6, 3−131.
comparing fitted stability constants as a function of L/M ratio; (25) Dumon, J. C.; Ernst, W. H. O. J. Plant Physiol. 1988, 133, 203−
comparison of speciation diagrams for different L/M ratios 209.
using singular log β values. Full spectra kinetics and single (26) Pais, I.; Fehér, M.; Bokori, J.; Nagy, B. Water, Air, Soil Pollut.
wavelength cuts across citrate concentrations for 4 min 1991, 57−58, 675−680.
exchange reactions. UV−vis absorbance of the full spectra for (27) Carvajal, M.; Alcaraz, C. F. J. Plant Nutr. 1998, 21, 655−664.
overnight kinetics of exchange at pH 4.5. Specfit/32 (28) Carvajal, M.; Gimenez, J. L.; Riquelme, F.; Alcaraz, C. F. Acta
determined species and speciation for both pH 7.4 and 4.5 Aliment. 1998, 27, 365−375.
kinetics data. This material is available free of charge via the (29) Reverte, S.; Carbonell-Barrachina, A. A.; Gimenez, J. L.;
Internet at http://pubs.acs.org. Carvajal, M. Acta Aliment. 2000, 29, 9−23.

■
(30) Abellan-Palazon, M.; Carbonell-Barrachina, A. A.; Gimenez-
AUTHOR INFORMATION Sanchez, J. L.; Lopez-Segura, M.; Martinez-Sanchez, F. Acta Aliment.
2001, 30, 159−171.
Corresponding Author
(31) Lesko, K.; Stefanovits-Banyai, E.; Pais, I.; Simon-Sarkadi, L.
*E-mail: ann.valentine@temple.edu. Phone: 215-204-7836. Novenytermeles 2001, 50, 71−81.
Notes (32) Lesko, K.; Stefanovits-Banyai, E.; Pais, I.; Simon-Sarkadi, L. J.
The authors declare no competing financial interest. Plant Nutr. 2002, 25, 2571−2581.

■ ACKNOWLEDGMENTS
We thank the New England Division of the American Cancer
(33) Hruby, M.; Cígler, P.; Kuzel, S. J. Plant Nutr. 2002, 25, 577−
598.
(34) Wojcik, P. J. Plant Nutr. 2002, 25, 1129−1138.
(35) Wojcik, P.; Gubbuk, H.; Akgul, H.; Gunes, E.; Ucgun, K.; Kocal,
Society for a Research Scholar Award (Research Scholar Grant
H.; Kucukyumuk, C. J. Plant Nutr. 2010, 33, 1914−1925.
RSG-06-246-01-CDD) to A.M.V., the donors of the Petroleum (36) Alcaraz-Lopez, C.; Botia, M.; Alcaraz, C. F.; Riquelme, F. J. Plant
Research Fund administered by the American Chemical Physiol. 2003, 160, 1441−1446.
Society, and the Research Corporation’s Research Innovation (37) Alcaraz-Lopez, C.; Botia, M.; Alcaraz, C. F.; Riquelme, F. J. Plant
Award RI0961, for partial support of this research.

■
Nutr. 2004, 27, 713−729.
(38) Pais, I.; Feher, M.; Soptei, C.; Tomordi, E. Titanium Plant
REFERENCES Conditioning Compositions. Great British PatentGB2090585, 1982.
(1) Martell, A. E. Chelates of Ascorbic Acid: Formation and Catalytic (39) Pais, I.; Feher, M.; Nagy, B.; Bokori, J.; Szabo, Z. Fodder and
Properties. In Ascorbic Acid: Chemistry, Metabolism, and Uses; Seib, P. Fodder Additives Promoting the Weight Increase of Domestic Animals
A., Tolbert, B. M., Eds.; American Chemical Society: Washington, DC, and a Process for the Use Thereof. United States Patent US4482550,
1982; Vol. 200, pp 153−178. 1984.
(2) Hughes, D. L. Dalton Trans. 1973, 2209−2215. (40) Pais, I.; Nagy, B.; Feher, M.; Szabo, Z.; Palfalvi, A.; Reibling, J.;
(3) Hollis, L. S.; Amundsen, A. R.; Stern, E. W. J. Am. Chem. Soc. Laszlo, M.; Kimura, S. Method for Increasing the Multiplication
1985, 107, 274−276. Biological Properties Using Titanium (IV) Ascorbinate Chelate.
(4) Yuge, H.; Miyamoto, T. K. Chem. Lett. 1996, 375−376. Australian Patent AU592211, 1990.
(5) Kriss, E. E.; Kurbatova, G. T.; Kuts, V. S.; Prokopenko, V. P. Russ.
(41) Yu, Y.; Ming, L.; Deng, W. Synthesis of Solid Titanium-Ascorbic
J. Inorg. Chem. 1976, 21, 1644−1645.
Acid Complex as Plant Growth Regulating Agent. Chinese Patent
(6) Kriss, E. E.; Kurbatova, G. T.; Kuts, V. S.; Prokopenko, V. P. Zh.
Neorg. Khim. 1976, 21, 2978−2981. CN85107690, 1987.
(7) Kim, Y.; Xudong, F.; Lippard, S. J. Inorg. Chem. 2007, 46, 6099− (42) Kato, T.; Kawano, T.; Poisu, I. Method for cultivating plant with
6107. nutrient solution. Japanese Patent JP62249902, 1987.
(8) Rajh, T.; Nedeljkovic, J. M.; Chen, L. X.; Poluektov, O.; (43) Dankiewicz, M.; Sas, J.; Malczewski, Z.; Zernik, Z.; Borowik, M.;
Thurnauer, M. C. J. Phys. Chem. B 1999, 103, 3515−3519. Turczyn, A. Method of Obtaining a Nitrogenous Fertilizing
(9) Martell, A. E.; Smith, R. M. Critical Stability Constants; Plenum Concentrate Containing Microelements. Polish Patent PL172272,
Press: New York, 1974−1977. 1997.
(10) Buettner, K. M.; Valentine, A. M. Chem. Rev. 2012, 112, 1863− (44) Ye, X.; Zhang, L.; Zhao, J. Preparation of Fertilizer Containing
1881. Soluble Silicon. Chinese Patent CN11525631997.

11038 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039

Inorganic Chemistry Article

(45) Ma, Y. Specific Fertilizer for Preventing Decreasing Production

due to Continuous Cropping and Alternate Cropping. Chinese Patent
CN1850737, 2006.
(46) Ma, Y. Method for Manufacturing Specific Fertilizer for
Preventing Decreasing Production due to Continuous and Alternate
Cropping and Therefore Increasing Crop Production. Chinese Patent
CN1850738, 2006.
(47) Schleicher, R. L.; Carroll, M. D.; Ford, E. S.; Lacher, D. A. Am. J.
Clin. Nutr. 2009, 90, 1252−1263.
(48) Lentner, C. Geigy Scientific Tables; Ciba Geigy: West Caldwell,
NJ, 1984; Vol. 3.
(49) Collins, J. M.; Uppal, R.; Incarvito, C. D.; Valentine, A. M. Inorg.
Chem. 2005, 44, 3431−3440.
(50) Martell, A. E.; Motekaitis, R. J. Determination and Use of Stability
Constants, 2nd ed.; Wiley-VCH: New York, 1992.
(51) Ö hman, L. O. Chem. Geol. 1998, 151, 41−50.
(52) Binstead, R. A.; Jung, B.; Zuberbühler, A. D. Specfit/32;
Spectrum Software Associates: Marlborough, MA, 2001.
(53) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbühler, A. D.
Talanta 1985, 32, 95−101.
(54) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbühler, A. D.
Talanta 1985, 32, 257−264.
(55) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbühler, A. D.
Talanta 1985, 32, 1133−1139.
(56) Gampp, H.; Maeder, M.; Meyer, C. J.; Zuberbühler, A. D.
Talanta 1986, 33, 943−951.
(57) Ö hman, L. O. Inorg. Chem. 1988, 27, 2565−2570.
(58) Ishiwata, K.; Ido, T.; Monma, M.; Murakami, M.; Fukuda, H.;
Kameyama, M.; Yamada, K.; Endo, S.; Yoshioka, S.; Sato, T.;
Matsuzawa, T. Int. J. Radiat. Appl. Instrum. Part A 1991, 42, 707−712.
(59) Tinoco, A. D.; Valentine, A. M. J. Am. Chem. Soc. 2005, 127,
11218−11219.
(60) Natelson, S.; Miletich, D. J.; Seals, C. F.; Visintine, D. J.;
Albrecht, R. F. Clin. Chem. 1979, 25, 889−897.
(61) Marangella, M.; Vitale, C.; Manganaro, M.; Cosseddu, D.;
Martini, C.; Petrarulo, M.; Linari, F. Nephron 1991, 57, 439−443.
(62) Thompson, G. A. K.; Taylor, R. S.; Sykes, A. G. Inorg. Chem.
1977, 16, 2880−2884.

11039 dx.doi.org/10.1021/ic301545m | Inorg. Chem. 2012, 51, 11030−11039