Autoxidation of Ni ( II ) and Co ( II ) Tetra , Penta and Hexaglycine Complexes Accelerated by Oxy Sulfur Radicals

A autoxidação dos complexos de Ni(II) e Co(II) com tetra, penta e hexaglicina, em meio de tampão borato, é acelerada por espécies de enxofre(IV) (H 2 SO 3 , HSO 3 – and SO 3 ). A formação dos complexos de Ni(III) e Co(III) foi acompanhada espectrofotometricamente em 325 e 265 nm, respectivamente. Técnicas eletroquímicas também foram empregadas para caracterizar a geração destes complexos. A velocidade da reação de autoxidação aumenta com a concentração de S(IV) e é máxima em pH ≅ 8,5. O processo é autocatalítico com Ni(III) ou Co(III) atuando como iniciadores, formados pela oxidação espontânea de Ni(II) ou Co(II) pelo oxigênio molecular. A dependência da constante de velocidade de pseudo-primeira ordem com a concentração de sulfito evidenciou possíveis reações paralelas com formação de um complexo com ligantes mistos antes da etapa da oxidação.


Experimental
Reagents All reagents were of analytical grade (Merck or Sigma) and were used as received.Solutions were prepared by using deionised water purified with a Milli-Q Plus Water system (Millipore).
Stock solutions of sulfite (1.00×10 -2 mol L -1 ) were daily prepared by dissolving the Na 2 S 2 O 5 salt in water previously purged with nitrogen.Water was flushed with nitrogen for at least half an hour to remove dissolved oxygen.To prepare diluted solutions of sulfite, small volumes of the stock solutions were properly added to air saturated water.The sulfite content of the stock solution was determined by iodimetry. 21i(II) (0.200 mol L -1 ) stock solution was prepared from the direct reaction of Ni (powder) (99.99%) with double distilled nitric acid.
Co(II) (0.965 mol L -1 ) stock solution was prepared from the direct reaction of Co(II) carbonate with perchloric acid followed by standardization with EDTA. 22
The kinetics was followed at the wavelength of maximum absorption of the Ni III G n or Co III G n complexes by using an HP8453A diode array spectrophotometer.
The data were obtained by mixing an equal volume of sulfite solution (2.0 -14) × 10 -5 mol L -1 and the metal ion complex solution (in borate buffer) in a double compartment cell (0.875 cm optical path length quartz cell) for slow reactions.Pro-K.2000Stopped-Flow Mixing Accessory was used for the experiments at short time intervals.The final concentrations after the mixture are indicated in all figures.
Air saturated solutions were employed in all experiments and the oxygen concentration was considered to be 2.5×10 -4 mol L -1 . 23Water was used as reference solution (blank).
A pHmeter Metrohm model 713 with a glass electrode (filled with sat.NaCl) was used in the pH measurements.The temperature was kept at 25.0 °C.

Electrochemical measurements
An Autolab PGSTAT 30 (Eco Chemie) bipotentiostat with the data acquisition software (GPES 4.8 version) was used.Experiments were done in an electrochemical cell with a Ag/AgCl (saturated KCl) and a platinum wire as reference and counter electrodes, respectively.Voltammetry with a rotating ring-disc electrode (RRDE) was carried out using an analytical rotator (AFMSRX) connected to the bipotentiostat.The glassy carbon/glassy carbon ring-disk electrodes (AFMT29) had the following dimensions: disk radius = 0.5613 cm, inner radius of the ring = 0.6248 cm and outer radius of the ring = 0.7925 cm.During rotating ring-disk electrode voltammetric experiments, the disk electrode potential was scanned between the limits 0.2 and 1.0 V, the ring being maintained at 0.0 V to collect the material generated at the disk.0.05 mol L -1 NaClO 4 in borate buffer was used as supporting electrolyte.
The voltammograms were obtained from air saturated solutions containing the metal ion complexes in borate buffer.In order to evaluate the effect of S(IV) on the voltammetric profile, very small volumes of S(IV) solutions were added to the electrochemical cell.The S(IV) concentration in the final solutions is indicated in the figures.

The sulfite induced autoxidation of Ni II G n
Ni II G n complexes have maximum absorbance peaks at 250 and 410 nm.In the case of Ni II G 4 complex, literature 24 reports an equilibrium between the blue octahedral [Ni II G 4 ] + and the yellow square planar [Ni II (H -3 G 4 )] 2-complexes.The amount of [Ni II (H -1 G 4 )] and [Ni II (H -2 G 4 )] -in solution is not appreciable at pH = 9.0 (H -1 G n 2-, H -2 G n 3-and H -3 G n 4-refer to the peptide ligand with 1, 2 and 3 deprotonated peptide nitrogen coordinated to nickel ion).In the present work, the transition from the octahedral to the square-planar geometry could be followed at 250 nm by the absorbance increasing, a constant value being reached after 10 minutes.
Ni II G 5 (2.0×10 -4 mol L -1 ) reacts slowly with oxygen in an autocatalytic process with induction period of 10000 s (at pH=8.50, borate buffer).However, in the presence of S(IV) the formation of Ni III G 5 complexes becomes much faster and reaches an absorbance limit after 1 s.
The S(IV) induced autoxidation of Ni II G n complexes was studied following the absorbance changes at 325 nm. Figure 2 shows the effect of the medium acidity on the formation and decomposition of Ni III G 5 .The maximum absorbance is reached at approximately 0.8 s as result of the fast S(IV) induced autoxidation of Ni II G 5 complex; a maximum value was observed at pH=8.50.The further decomposition of Ni III G 5 complex increases with the increase of the pH, a similar behavior being observed with Ni III G 4 and Ni III G 6 complexes.The Ni III G 4 complex formed in absence of sulfite decomposes with oxidation of the ligand. 20he Olis Kinfit 25 and Pro-K.2000 26set of programs were employed to fit the absorbance-time traces.However, no good approach was achieved.Therefore, the first order formation of Ni III G n (k obs ) was obtained by the initial slope (ln (Absorbance t ) -ln (Absorbance i ) vs. time plots, neglecting the induction period (less than 0.1 s).Unfortunately, the software from Pro-K.2000Stopped-Flow Mixing Accessory (on line with HP8453A diode array spectrophotometer) allows data acquisition with time interval of 0.1 s, such as after the induction period only three or four points were available to obtain the initial slope.This part of the kinetic trace exhibits the maximum rate for Ni III G n formation. 2,27he k obs vs. pH, represented in Figure 2(B), can be subject of some error due to the interference of the induction period (an evidence of autocatalytic process) and further decomposition of Ni III G 5 complex (especially at higher pH).k obs values increase with pH in the range of 7.5 -9.0 likely because of the different degree of protonation of the nickel complexes, which is dependent on the medium acidity.The [Ni II (H -3 G 5 )] 2- is probably the main species in solution at pH 8.5 -9.0.
For comparative studies, a kinetic investigation on the S(IV) induced autoxidation of Ni II G n complexes was carried out at pH=8.50, in air saturated solution.At this pH, the Ni(III) formation was more efficient and its decomposition is slower (Figure 2(A)).
The dependence of Ni III G 5 formation on the S(IV) concentration (Figure 3   such lower S(IV) concentrations were not significant.Due to the lack of sufficiently detailed experimental information, the proposed kinetic data treatment yields a semi-quantitative description.Therefore, in the present work only the main pathways in the intrinsic mechanism are described.
Figure 3(A) shows a small induction period, which depends on the initial [S(IV)].Absorbance values after 1 s are linear with initial [S(IV)], hence they may be used for analytical purposes.At longer scale (t > 3s) the absorbance decreases as a consequence of the decomposition of Ni III G n complexes.

Mechanism
As already shown in our previous work, 7,17 the sulfite induced autoxidation of Ni II G 4 is an autocatalytic process involving a chain of redox reactions with free radical formation.
The Scheme 1 describes the main reactions involving a redox cycling for the oxidation of Ni II G n complexes promoted by the strong oxidizing agents (SO 4 •-, SO 5 •-and HSO 5 -) generated in the reaction.
The autocatalytic process needs the initiator Ni III G n to start the reaction (equation 6).Small amounts of this species are produced by direct oxidation of Ni II G n by dissolved oxygen (equation 1). 17,20This step initiates the reaction between Ni III G n and SO and S(IV) oxidation.9][30][31][32][33][34][35][36] In this process type the interaction of the metal ion (such as Fe 3+ , Mn 3+ , Co 3+ and Ni 3+ ) with HSO 3 -or SO 3 2-, at low concentrations, results in the oxidation of SO 3 2-to the radical SO 3 •-and the reduction of the metal ion to the 2+ state.In the presence of oxygen, the metal ion is oxidized back to higher oxidation state.This redox cycling was clearly demonstrated for Co(II)/(III)/ N 3 -, 33 Mn(II)/(III)/N 3 -, 33 Fe(II)/(III)/H 2 O, 32 Ni(II)/(III)/ OH -, 6 Ni(II)/(III)/cyclam. 8 Initiation Scheme 1. Mechanism of the sulfite induced autoxidation of Ni II G n . 17, where H -1 G n 2- , H -2 G n 3-and H -3 G n 4-refer to the peptide ligand with 1, 2 and 3 deprotonated peptide nitrogen coordinated to the nickel ion.J. Braz.Chem.Soc.Some studies in the literature 9,31 clearly show the importance of Fe(III) as one possible initiator, present as unavoidable impurity in chemicals or even in highly purified water.If Fe(III) is considered as the initiator, SO 3 •could be generated by the reduction of Fe(III) by SO 32-(similar to equation 6), with further oxidation of Ni II G n (equations 8, 10-12).As the present studies were carried out at pH = 8.50, Fe(III) impurities may be present as hydrolyzed species and equation 6 must be important in the initiation process.
The intercept and constant value of k obs at higher S(IV) concentration (in the case of Ni II G 5 and Ni II G 6 autoxidation) in Figure 3 •-) (L=lysylglycylhistidine carboxiamide), where the SO 4 •-radical may coordinate to Ni III L complex.Green et al. 37 considered the formation of reactive dimmer species of Ni(II) and Ni(III)Gly 2 HisGly complexes in the oxidative self-decomposition of Ni(III) complex.

The sulfite induced autoxidation of Co II G n
Figure 1B shows the spectrum of a Co II G 5 solution (pH = 9.05) recently prepared.The Co III G 5 formation in air saturated solution, in absence or presence of S(IV), can be followed by the absorbance changes at 265 nm (Figure 1B').
A solution of 1.0×10 -4 mol L -1 of Co II G n complexes was chosen for the kinetic studies since Co(OH) 3 may be formed at higher concentrations of Co(II) after 120 minutes.The precipitation of Co(OH) 3 could also be prevented by keeping G n ligand in excess (10%).
In the absence of sulfite the autoxidation of Co II G 5 is relatively slow and becomes faster by increasing the pH (Figure 4(A)).At pH 8.48 -9.05 (Figure 4(B) a-c) the effect of S(IV) by accelerating Co III G 5 formation can be better evaluated (compare Figure 4(A) and 4(B)), since the spontaneous oxidation by dissolved oxygen is slower.
Similarly to Ni II G n complexes, the reaction rate and effectiveness of Co(III) formation may be related to the different reactive species of Co II G 5 due to the different protonation degree of the coordinated ligand, which is dependent on the solution acidity.As pK values for [Co II (H -x G n )] (1-x) or [Co III (H -x G n )] (2-x) are unknown, it is not possible to define the distribution diagram of the different species in the solution.Besides, a shift in the HSO 3 _ /SO 3 2equilibrium (pK 2 = 7.2) 38 will lead to an increase in the redox rate constant at pH>7.The spontaneous oxidation of Co II G n by dissolved oxygen was faster by increasing the pH, Co II G n and ligand concentrations.For instance, the Co III G n formation in the presence or absence of sulfite exhibits a remarkable dependence on the free ligand concentration (G 5 was kept in excess 0 -200% over Co II G 5 , data not shown), the reaction becoming faster with higher Co III G 5 formation.
In order to better evaluate the accelerating effect of S(IV) on the Co III G n formation, experimental conditions must be such that the spontaneous oxidation by O 2 is minimized (Figure 5, [S(IV)]=zero) and the S(IV) effect is more pronounced.
Comparing Ni II G 5 and Co II G 5 ([M II G 5 ]=1.0×10 -4 mol L -1 , 100% G 5 in excess, at pH = 9.05, data not shown), the later reacts faster with O 2 (in absence of S(IV)).However, Co(III) complexes formation is much slower in the presence of sulfite (compare Figures 3(A   The Olis Kinfit 25 set of program was employed to fit the absorbance-time traces and two exponential functions could be fitted (Table 1), resulting k 1obs and k 2obs .It means that there are at least two consecutive steps, with rate constant values with different order of magnitude.However, it was not possible to fit the complete absorbance vs time trace with three or four exponential functions.
For Co II G 4 complex, k 1obs showed a dependence with S(IV), while k 2obs = 2.0x10 -3 s -1 was found to be constant.These results are in agreement with our previous work. 17k 1obs and k 2obs do not depend on S(IV) concentration for Co III G 5 and Co III G 6 complexes formation.
The fact that the sulfite induced autoxidation of Co II G 4 is faster than Co II G 5 and Co II G 6 complex could be associated to a steric effect of the ligands.
The Co III G n formation, especially in the case of Co III G 5 (Figure 5(B)) and Co III G 6 (Figure 5(C)), is the sum of spontaneous oxidation by dissolved oxygen and S(IV) accelerating effect.Initially, the sulfite induced autoxidation of Co II G n occurs with simultaneous oxidation of S(IV).After the S(IV) consumption, the autocatalytic Co III G n formation is due to the oxygen still remaining in solution (the absorbance did not reach a limit value).
The sulfite induced autoxidation of Co II G n complexes also shows evidence for a mechanism involving formation of radicals.The initiation may also involve a parallel reaction with formation of a mixed ligand complex.Accordingly, literature shows that Co II SO 3 2-is fairly stable. 39n the Scheme 1 (proposed for the sulfite induced autoxidation of Ni II G n ), additional steps must be considered in the initiation process for Co III G n generation, with formation of dimeric complexes with oxygen adduct and peroxo bridges.
Studies 40 with Co(II) peptides (gly-gly, gly-ala, alagly and ala-ala) showed the formation of dimeric complexes with μ-superoxo bridges.In the case of Co II aspargine, the formation of dioxygen complex occurs prior to the oxidation of the metal center. 41he pH dependence of Co II G n oxidation by oxygen (in the presence or absence of sulfite), Figure 4, may be explained by the possible formation of oxygen adduct or μ-superoxo bridge.This property of Co(II)/(III) complexes may explain the different behavior of Co II G n compared to Ni II G n in the presence of oxygen and sulfite.nature of the Co(II) electroactive species still remaining in the solution.

Conclusions
In the present work the formation of Ni(III) and Co(III) complexes could be followed by spectrophotometry and cyclic voltammetry.The exact nature of these complexes and of the products (structures, protonation degree of the ligands and mixed complexes) is unknown.
The complexity of the system does not allow a definitive assignment of the involved species.The redox process is more effective when oxygen is kept in large excess over sulfite concentration and higher pH.An important aspect to consider is the decomposition of Ni III G n complexes at pH higher than 9.0, so it was not possible the accurate data treatment.
The mechanism elucidation for the autoxidation of Co II G n complex is more complicated since species as mixed ligand complex and dimeric complexes with μ-superoxo bridges may be involved.
The redox cycling represented in the Scheme 1, involving changes in the oxidation state of the metal ion complex, is active as long as sulfite and oxygen are present in the solution to generate the SO 5 •-, HSO 5 -and SO 4 •species.
The understanding of the redox cycling process of Ni III G n /Ni II G n and Co III G n /Co II G n is of interest not only in atmospheric processes but also in the treatment of gaseous effluents to assist pollution control systems development since these metal ion complexes can be efficient catalysts of sulfur compounds.][50][51]

Figure
Figure 1A and 1A' show the spectra obtained before and after 0.8 s of addition of 5.0x10 -5 mol L -1 sulfite to a Ni II G 5 solution.An absorption band at 325 nm is observed due to the formation of Ni III G 5 .
(A)) could only be studied over a limited concentration range where the initial [S(IV)] = (0.05 -0.35) × [Ni(II)].At [S(IV)] > 7×10 -5 mol L -1 the reduction of Ni III G 5 formed may occur in an air saturated solution ([O 2 ] ≅ 2.5×10 -4 mol L -1 ). 23Experiments following Ni III G 4 and Ni III G 6 formation showed similar Absorbance × time profiles (data not shown).Experiments at [S(IV)] lower than 1×10 -5 mol L -1 would better define the intercept observed in the Figure 3(B), which is due to either a back or parallel reaction.However, the absorbance changes at
with formation of SO 3•-, which reacts with O 2 producing SO 5 •-(equation 7).The oxidation of Ni II G n occurs with simultaneous consumption of O 2 6,8,9
(B) can be interpreted as an evidence of back or parallel reaction, as represented by equation 5.The formation of mixed ligand complex (equations 2 and 3), prior to the oxidation step (equation 4), may involve a slow dechelation process of the G n ligand with coordination of sulfite and O 2 to Ni II G n .An internal electron transfer could produce [Ni III (H -2 G n )SO 3 ] 2-(equation 4).In a posterior step Ni(III) is reduced by SO 3 2-to give [Ni II (H -3 G n )] 2-and SO 3 •-.Two studies in the literature about Ni(II)/(III) peptide complex are relevant in the present discussion.Lepentsiotis et al. 36 proposed the formation of Ni III L (SO 4 )

Figure 5
Figure 5 shows no clear induction period, which may indicate the formation of a significant amount of Co(III) as a result of spontaneous oxidation by oxygen.Also in this case, the dependence of Co(III) formation with S(IV) concentration was studied over a limited concentration range [S(IV)] = (0.1-0.5) × [Co(II)].The experiments were not carried out under pseudo-first order conditions (Co(II) and O 2 in large excess over S(IV)).The Olis Kinfit 25 set of program was employed to fit the absorbance-time traces and two exponential functions could be fitted (Table1), resulting k 1obs and k 2obs .It means that there are at least two consecutive steps, with rate constant values with different order of magnitude.However, it was not possible to fit the