- Email: [email protected]
Protonation and silver(I) complex-formation equilibria of some amino-alcohols
Talanta Talanta 44 (1997) 2059 2067
ELSEVIER
Protonation and silver(I) complex-formation equilibria of some amino-alcohols Silvia Canepari b Vincenzo Carunchio ~, Paola Castellano a, Antonella Messina
~'*
" Department q/' Chemistry, Universi O, of Rome, La Sapienza, P.le Aldo Moro, 5-00185 Rome. ltah' h ~TMS-Bioanalitieal Chemist O, Section, Universi O, q/' Rome, La Sapienza, P.le Aldo Moro, 5-00185 Rome, lta@
Received 7 August 1996; receivedin revised lbrm 21 January 1997: accepted 4 February 1997
Abstract
Formation constants of the silver(I) complexes with some amino-alcohols have been determined at 25°C in 0.5 M KNO3 by means of two independent potentiometric measurements employing glass and silver electrode. The ligands considered are: sec-butylamine, 2-amino-l-propanol, 2-amino-l-methoxy-propane, 2-amino-2-methyl-l-propanol, 2amino-l-butanol, 2-amino-l-pentanol, 2-amino-l-hexanol, 2-amino-l,3-propanediol, 2-amino-l,3-hexandiol, 2-amino2-methyl-l,3-propanediol and Tris(hydroxymethyl)-aminomethane. Protonation constants of the selected ligands have also been determined. Calculations were made using HYPERQUAD computer program. The influence exerted by the introduction of hydroxy groups and by the presence of alkyl residuals in the ligand structure on the formation equilibria, is discussed. © 1997 Elsevier Science B.V. Keywords: Amino-alcohols; Complex formation equilibria: Silver(I)
1. Introduction
The amino-alcohols represent a largely diffused class of biologically active substances. They characterise the moiety of several drugs and natural substances like ephedrine, adrenaline and noradrenaline and of some naturally occurring sphingoid like sphingosine and sphinganine. Due to the coordinating properties of amino and hydroxy groups, the study of formation equilibria of complex species with metal ions can help to clarify the biological role of these substances. The knowledge of complexing properties of amino-alcohols is not yet exhaustive and only a few data related to their * Corresponding author.
complex equilibria are known [1 14]. In particular sphingosine and sphinganine in the last few years have received great attention because of their activity as inhibitors of tumoral cells growth [15-18]. Notwithstanding this, the mechanism of their biological activity has not yet been clarified and their pharmaceutics application is nowadays hampered by the cytotoxic effect shown at high concentration [19]. However, it is known that the antitumoral effect of these sphingoids depends on the structure of their polar head and in particular on the presence of an amino-group, positively charged, in position 2 [15,20], so that when the -NH2 is chemically bonded, as an example by acylation, the inhibiting effect does not occur.
0039-9140/97;$17.00 ¢7 1997 Elsevier Science B.V. All rights reserved. Pll S0039-9 140(97)00044-1
2060
S. Canepari et al. / Talanta 44 (1997) 2059-2067
Then, is presumable that a coordination bond involving the amino group also influences the pharmacological activity of these compounds. Furthermore, as the interaction with metal ions is usually reversible and rapid, the study of complex formation equilibria may allow an efficient regulation of active species concentration. In this context, the present paper is concerned with the study of complexation equilibria of some amino-alcohols with Ag(I). The considered ligands are: sec-butylamine (L0, 2-amino-1-propanol (Ln) , 2-amino-l-methoxy-propane (LIll) , 2amino-2-methyl-l-propanol (Lw), 2-amino-l-butanol (Lv), 2-amino-l-pentanol (Lw), 2-amino-1hexanol (Lw0, 2-amino-l,3-propanediol (Lvni), 2-amino-l,3-hexandiol (L~x), 2-amino-2-methyl1,3-propanediol (Lx) and Tris(hydroxymethyl)aminomethane (Lx0. The ligands, which have a structure similar to the polar head of sphingosine and sphinganine, have been chosen in order to evidence the effect of the number of substituent hydroxy-groups and of the addition of alkyl residual on the coordinating properties of 2-amino-alcohol moiety. The study has been extended to Tris(hydroxymethyl)-aminomethane (TRIS, THAM), whose complexing properties have been not sufficiently documented even if it is often added as a buffer to solutions prepared to investigate biochemical reaction involving transition metal ions [12-14,21-
24]. Among the elements of the same periodic group, silver has been the least studied one, probably due to its extremely low presence in biological systems. However, our attention has been focused on silver(I) compounds, firstly because of the lack of thermodynamic data and secondly because, as already observed for other metal ions whose presence is not relevant in biological systems [25-27], their applicative importance can not be excluded.
2. Experimental 2.1. Materials
All the commercial products employed for
the preparation of L n I and L i v , w e r e by Fluka. Twice distilled and deionized water was used for preparing all solutions. All the reagents were obtained in high purity. Ultra pure KNO 3 (Suprapur grade, Merck) was used as ionic strength buffer. Silver(I) nitrate (Carlo Erba) stock solutions were standardized by titration with standard reagent NaCI (Merck) and using K2CrO 4 (Merck) as indicator following the Mohr's method [28]. Ligand stock solutions. As all the considered ligands are highly hygroscopic and uptake easily CO2, the direct weighing was not possible and all the solutions had to be handled in strictly air-free conditions. Stock solutions were then prepared by taking a volumetric sample in a glove-box under an argon atmosphere and diluting with appropriate volumes of water. The titre was successively controlled from series of crossed acid-base titrations with a potentiometric end point detection. 2-Amino-l,3-propanediol has been supplied as oxalate salt. The quantitative precipitation of oxalate ion with calcium(II) was then performed before preparing stock solutions. The presence of a 0.5% excess of calcium(II) nitrate, needed in order to assure the quantitative precipitation of oxalate, has been checked by EDTA titration. Potassium hydroxide 0.1 M (COz-free solution, Merck) was standardized by potentiometric titrations against a solution of potassium hydrogen phthalate (Merck pro analysis, dried at 120°C) using the Gran method [29] for the evaluation of the end point. Nitric acid solution (0.1 M) was prepared by diluting the pure concentrated product (Merck) and standardized by potentiometric titrations against tris(hydroxymethyl)aminomethane (Merck), dried at 100°C for 24 h. An independent check of the analytical procedure was also carried out by a direct titration of potassium hydroxide solution against the nitric acid.
S. Canepari et al. / Talanta 44 (1997) 2059-2067
2. I. 1. Synthesis of 2-amino- 1-hexanol and 2-amino- 1,3-hexandiol As 2-amino-l-hexanol and 2-amino-l,3-hexandiol are not commercially available, these products were prepared following the synthetic path summarized in Scheme 1 [30].
2.1.2. E M F measurements Glass electrode. The alkalimetric titrations were carried out adding KOH to the test solution placed in a thermostatic vessel, under agitation and under a stream of purified argon presatured with a 0.5 M KNO3 solution. The temperature control (25_+0.1°C) was achieved by means of water circulation in the jacketed vessel from a water thermocryostat (thermostat HAAKE DC3 and chryostat HAAKE K15). The initial volumes of the titrated solutions were 25 ml. The KOH solution was delivered in the titration vessel by a Metrohm Dosimat 655 digital burette with a total volume of 5 ml. In order to have CO2-free potassium hydroxide, the solution was preserved in a bottle stopped by a soda lime plug and kept under argon atmosphere. In these conditions the solution was stable up to 1 month. Potentiometric titrations were carried out using a Metrohm mod. 665 pH/mVmeter, with a glass electrode (INGOLD) and a double junction Ag/ AgCI electrode (INGOLD) as a reference. The outer solution of the reference electrode was 0.5 M KNO 3. The potentiometric apparatus was made completely automatic by applying an IBM mod.30 personal computer with a parallel interface and using an appropriate software program. [31] OBn Br~
osn m
oe~
b,c,d,e •
o,1
NH2
A 4O%
s
~
/ g,h.i
OB.
Br
2061
The electrode standard potentials E ° were determined every day from the acid region of the titration of a 0.005 M HNO 3 solution (0.5 M ionic strength buffer) against 0.1 M KOH. Similarly the response of the glass electrode was checked every week evaluating the value of pKw from the basic portion of the calibration curve. The linearity of E versus pH function was also checked; the slope of the straight line obtained was, within the experimental errors, coincident with the theoretical value. For the evaluation of the protonation constants, solutions 0.5 M in KNO 3 containing different quantities of ligand and acidified with a known excess of HNO3 were titrated with potassium hydroxide solution. For the formation constant determination of Ag(I) complexes, solutions at various concentrations of Ag(I) nitrate and ligands were titrated with KOH. Before the titrations all the solutions were acidified with a known excess of nitric acid. The ionic strength was always adjusted to 0.5 M by adding solid potassium nitrate. Ag electrode. All the measurements regarding AgO) complexation equilibria were repeated also by employing an Ag electrode (Metrohm) in order to have an independent measurement. The calibration of the Ag electrode was performed by measuring the potential of five solutions at different Ag(I) concentration containing KNO 3 0.5 M as ionic strength buffer. E ° and slope were determined by linear regression of data (r2= 0.99) and the obtained values were in good agreement with the theoretical ones (error < 1%). The operative conditions employed for the determination of stability constants were the same as those related to glass electrode measurements. All the calculations for the calibration of the glass electrode and for the determination of ligand protonation and complex formation constants were carried out by employing HYPERQUAD least squares computer program [32].
3. Results and discussion NH2
2-aminQ-1,3-hexandiol
Scheme 1.
Particular attention had then to be paid to the choice of concentration ranges, considering the
S. Canepari et al. /Talanta 44 (1997) 2059 2067
2062
Table 1 Initial analytical concentrations (raM) of ligand (CL) and acid (CH) and - l o g [ H +] ranges for the determination of the protonation constant of 2-amino-l-hexanol at l = 0.5 M KNO3 and T = 25°C Ligand
Expt. no
1, 2, 3 2-Amino-l-hex- 4 anol 5
CL 1.8 3.7 6.7
CH
- l o g [ H 7] ranges
3.0 3.08-10.61 5.0 3.07 10.34 10
3.08-10.43
method sensitivity and the best operative conditions for the calculation of the constants. Good results were obtained in all the examined systems when the ligand concentration was maintained lower than 7 raM. In fact, by a preliminary examination of the experimental data, turns out that the ligand concentration influences the results. In particular, there is a progressive worsening of the fit on increasing the ligand concentration in the titration vessel.
3.1. Protonation equilibria The calculation of the protonation constants has been performed by elaborating the data obtained by five titration runs for each system. In Table 1 the operative conditions are reported for 2-amino-l-hexanol as an example. Similar conditions have been employed for all the other ligands. As can be noted, one of the titrations has been repeated three times in order to check the reproducibility of the experimental measurements. Repeated runs also allowed a better control of the ligand concentrations, whose determination requires particular care (Section 2). In fact the amounts of ligands in the titration vessel (COL) have been refined by HYPERQUAD program and the resulting values were in a good agreement with the calculated ones (2% error). In Table 2 the pK,~ values obtained by the simultaneous elaboration of all the titration curves are reported. The error between brackets is the standard deviation provided by HYPERQUAD calculations. The statistical parameters are satisfactory also considering that, due to the
basicity of ligands, E.F.M. measurements had to be performed in the basic region. The pK, values related to sec-butylamine [33], 2-amino-l-butanol [34] and THAM [12 14,2124] are in a good agreement with literature data. From the examination of the reported data, two main effects on the acidity of protonated ligands can be evidenced. Firstly, the insertion of hydroxy group causes a relevant lowering of pK~ values. This effect could be caused either by the intramolecular hydrogen bond formation [35], which is stronger in the deprotonated form than in the protonated one, or by an electron-withdrawing polar effect due to the hydroxy group. Comparing the pK~, values related to L , and L m it is evident that the second effect is prevalent. Furthermore, the methoxy group seems to be more effective than hydroxy group, in spite of the lower electronic-withdrawing capabilities of this group with respect to hydroxy one. This reversal, already observed for similar compounds [36] was ascribed to the modification of the true inductive effect of the hydroxy group by solvation. Secondly, the addition of alkyl residuals in the ligand structure causes an increase of the pK~ values. The effect is mainly ascribable to a field effect of the alkyl group, which results in an increase in the electron density on the nitrogen atom [36]. Furthermore, from the results it can be Table 2 Protonation constant values and related statistical parameters for the considered ligands at I = 0.5 M KNO 3 and T = 25°C Ligands
pK,,
Z2
o"
Sec-butylamine 2-Amino-l-propanol 2-Amino- 1-methoxy-propane 2-Amino-2-methyl- 1-propanol 2-Amino-l-butanol 2-Amino-l-pentanol 2-Amino-l-hexanol 2-Amino- 1,3-propanediol 2-Amino-1,3-hexandiol 2-Amino-2-methyl- 1,3propanediol Tris(hydroxymethyl) aminomethan
10.599(3) 9.450(6) 9.424(2) 9.612(6) 9.554(5) 9.766(3) 9.90(6) 8.786(I) 8.901(2) 8.780(6)
19.23 13.54 10.29 14.37 15.00 17.50 18.41 10.32 11.73 15.68
3.72 2.01 1.34 2.98 2.80 3.01 4.00 1.09 3.21 2.90
8.156(1 )
5.23
0.56
S. Canepari et al. ' Tahmta 44 (1997) 2059- 2067
Table 3 Initial analytical concentrations (mM) of ligand (C0, metal (CM) and acid (C,), CL/CM ratios and p[H] ranges employed in calculations of complexation constant of 2-amino-l-hexanol -Ag(l) at I = 0 . 5 M KNO3 and T = 25°C Expt. no
C~
CM
CL/CM
Cii
p[H] ranges
I 2 3 4 5 6 7 8
4.10 1.49 2.84 3.70 4.24 3.20 2.01) 4.115
4.10 0.75 1.42 1.85 1.41 1.09 0.69 0.81
1 2 2 2 3 3 3 5
6.10 3.50 4.85 5.70 6.10 3.52 2.21) 6.10
7.42 8.26 7.78- 9.25 7.61) 8.93 7.50 8.85 7.92 9.48 7.78 9.56 7.65 -9.75 7.54 10.20
noted that the insertion of methyl group in position 2 in L~v and Lx causes an increase of pK~ lower than the one expected on the basis of the field effect. This is in agreement both with the steric hindrance on the quaternary carbon, which could stabilize the intramolecular hydrogen bond with the consequent lowering in basicity and with the differential hydration [37] which is less effective when a -CH~ group is added in position 2.
3.2. Complexation equilibria In order to evidence both mononuclear and polynuclear species, the determination of complexation constants has been carried out for each considered system on a stock of eight solutions containing various CL/CM ratios and at different ligand and metal initial concentrations. In Table 3 are reported, as an example, the operative conditions for 2-amino-l-hexanol-Ag(I) system. As can be noted, the - l o g [ H +] ranges are variable depending on experimental conditions and have been identified considering all the equilibria involved, in order to avoid precipitation equilibria. All the experimental runs were firstly elaborated by H Y P E R Q U A D one by one, afterwards an elaboration of each set obtained with glass or Ag electrode was performed. Finally all the titration curves were elaborated simultaneously. Different equilibrium models were examined. The species ML, MLe M2L, M2L2, MLOH, ML2OH
2063
and ML 3, were systematically considered in different combinations. These species were either rejected by H Y P E R Q U A D or caused a worsening of the statistical parameters a and Z 2 [38] or lead to not acceptable chemical models. The absence of systematic errors was tested refining also the dangerous parameters as COM, COL, Co. (total mmoles of metal, ligand and proton, respectively, in the reaction vessel). The deviation of calculated values from input ones can be considered within the experimental error (about 2%). Hydrolysis constants [39] were considered in the calculation. However they do not influence the log fl values, apart from a slight ( ~ 1%) effect on the results related to the titration curves with C L / C M = 1,
The calculations related to Ag electrode measurements give statistical parameters better than those obtained with glass electrode, probably due to the working pH range in which the alkaline error is not negligible. Anyway, the stability constants obtained with the two independent measurements are in good agreement as evident in Table 4 in which the results related to Ag and glass electrode measurements for Ag(I)-2-amino1,3-propanediol system are reported. In Table 5 the results of the overall elaboration obtained for all the considered systems, are summarized. Two different models, i.e., M L - M L 2 (model 1) and ML-ML2-MLOH (model 2), are reported as the most significative ones. Is worth noting that is not possible from the potentiomettic measurements, to deduce whether the species indicated as M L O H is a hydroxy complex or a complex where ligand acts as a chelate, with the deprotonation of the hydroxy groups. Due to steric reasons and considering logfl values the first alternative seems to be more probable [21]. On the other hand, the enhancement of the affinity for OH-compared with the hydrated silver ion has been already observed from other authors in the presence of aromatic [40] or aliphatic [21,41] N-ligands. From a critical view of the Table 5 it is evident that the model 2 is the most representative at least for the ligands L w . , Lix, Lx and Lx~, in which two or three hydroxy groups are present. The
S. Canepari et al. ,/Talanta 44 (1997) 2059-2067
2064
Table 4 Formation constants of complexes of 2-amino-l,3-propanediol with AGO), and related statistical parameters obtained at I = 0.5M KNO3 and T = 25°C Ag electrode
log fl1.1,o (log K 0 log f12,1,0 log K 2 log fl~ j _ ~ log K~(AgL +) 2' 2 o"
Glass electrode
Model 1
Model 2
3.39 (1) 6.72 (1) 3.33 --28.00 3.85
3.407 (6) 6.64 ( 1) 3.23 - 6.24 (4) -9.6 21.00 1.58
Total
Model 1 3.40 (6) 6.79 (2) 3.39 -62.13 5.18
presence of the MLOH species is evident in Fig. 1 in which the comparison between the two models for Ag(I)-THAM system is reported as a plot of experimental and calculated data. For all the other systems, the presence of ternary species MLOH is uncertain and its addition to the speciation model does not improve significantly the fitting. This may be due both to the reduction in acidity of the bonded water (see log Ka(AgL +) values in Table 5) and to the formation constant values, so that the presence of ML2 species is predominant in the examined experimental conditions (Fig. 2 and Fig. 3). As far as the stability order of the complexes is concerned, the same trend as observed for pK~
Model 2
Model 1
Model 2
3.39 (5) 6.69 (3) 3.30 - 6.40 (7) -9.8 30.20 2.60
3.38 (1) 6.76 (1) 3.38 --50.25 4.31
3.39 (1) 6.65 (2) 3.26 - 6.34 (5) -9.7 22.15 2.19
values has been found. It is then evident that the coordination is due to the aminic nitrogen and that no chelate structures are formed. As found
&O
6,0
7,0
0,0
9,O
I0,0
Fig. 2. Plots of formation percentages vs p[H] related to ML, ML 2 and MLOH species for Ag(I)-THAM. CL = 3.2 m M , C M = 1.1 m M , C H = 3 . 2 raM. Formation constants as reported in Table 5 (model 2).
,.,:::L'.I" 3,60
3,40
i
0. 20.
300 0,10
0,15
0,20
0,25 0,30 tltrant (mL)
0,35
040
0,45
Fig. I. Comparison between experimental and calculated curves related to either model l (ML, ML2) or model 2 (ML, ML2, M L O H ) for the Ag(I)-THAM system with Ag electrode. C L = 3.2 m M , CM = l.l m M , C H = 3.5 m M , 1 = 0 . 5 M KNO~ and T = 25°C.
0. - 5.5
6.5
7,5
&5
9.5
10,5
Fig. 3. Plots of formation percentages vs p[H] related to ML, M L : and MLOH species for Ag(I)-2-amino-l-propanol. C L = 3.2 raM, C M = 1.1 m M , CH = 3.2 m M . Formation constants as reported in Table 5 (model 2).
S. Canepari et al.
Talanta 44 (1997) 2059-2067
,.o o
o II
keo o_
~
Z
oo i
II ~G ~EP. -o
~'~ o
-
I
a2~ =
=
~
~'~
._= .~
~
I ~
~.~
..Y
~
~
'~
~~
.-~
7"
~'
2065
2066
S. Canepari et al. /Talanta 44 (1997) 2059-2067
for other systems [6,42], (log KO/pK, ratio is constant for all the considered ligands, the average value being 0.37 _+ 0.02. It is worth noting that in the systems containing Liv and L x as ligands, the stability of the ML species is lower than the one expected considering the pK~ values. These results can be ascribed to the steric hindrance which is more influent on the tetrahedral structure of ML species with respect to the linear one characteristic of ML2 species in silver(I)-aliphatic amines systems. Furthermore, it can be noted that log K 2 values are higher than log K~ ones, with the exclusion of 2-amino-l,3-hexandiol and 2-amino-l,3propane- diol. This different behaviour can be explained assuming that the K values, as suggested by Orgel [43] and interpreted in detail by Martin [44] in terms of cooperative effect, are made up of two parts which are in competition each other. The first is mainly due to a statistical-electronic effect and causes a fall off of stability from K~ to K,,. The second part can be attributed to an energetic effect associated to the change of the complex structure from tetracoordinate to linear and determinates an increasing of stability. Depending on which is the prevalent effect, K2 value will be lower or higher than KI.
Acknowledgements This work was supported financially by a National Project of Italian M.U.R.S.T., Rome, Italy.
References [l] A. Braibanti, G. Mori, F. Dallavalle, L. Leporati, J. Chem. Soc. Dalton (1975) 1319. [2] A. Braibanti, G. Mori, F. Dallavalle, J. Chem. Soc. Dalton (1976) 826. [3] A. Braibanti, G. Mori, F. Dallavalle, J. Chem. Soc. Dalton (1979) 1050. [4] C. Van Poucke, Z. Eeckhaut, Bull. Soc. Chim. Beiges 81 (1972) 363. [5] B.A. Timini, D.H. Everett, J. Chem. Soc (B) (1968) 1380. [6] D.J. Alner, R.C. Lansbury, A.G. Smeeth, J. Chem. Soc (A) (1968) 417.
[7] J.J. Christensen, R.M. lzatt, D.P. Wrathall, L.D. Hansen, J. Chem. Soc (A) (1969) 1212. [8] A.Y. Sychev, A.P. Gerbeleu, P.K. Migal, Russ. J. Inorg. Chem., 8 (1963) 1081. [9] P.K. Migal, A.N. Pushnyak, Russ. J. lnorg. Chem. 4 (1959) 601. [10] P.K. Migal, A.N. Pushnyak, Russ. J. Inorg. Chem. 5 (1960) 293. [11] J.M. Antelo, F. Arce, J. Casado, M.C. Regal, A. Varela, J. Chem. Research(s) (1982) 144. [12] L. Bologni, A. Sabatini, A. Vacca, lnorg. Chim. Acta 69 (1983) 71. [13] K.S. Bai, E. Martell. J. lnorg. Nucl. Chem. 3l (1969) 1697. [14] W. Forsling, Acta Chem. Scand. A 32 (1978) 57. [15] Y.A. Hannun, R.M. Bell, Science 243 (1989) 500. [16] A.H. Merrill Jr., M. Sereni, V.L. Stevens, Y.A. Hannun, R.M. Bell, J.M. Kinkade Jr., J. Biol. Chem. 261 (1986) 2610. [17] M. Faucher, N. Girones, Y.A. Hannun, R.M. Bell, R.J. Davis, J. Biol. Chem. 263 (1988) 5319. [18] L.A. Speizer, M.J. Watson, U Brunton, Fed. Proc. 46 (1987) 2065. [19] Y.A. Hannun, R. Bell, Science 235 (1987) 678. [20] E. Wilson, M.C. Olcott, R.M. Bell, A.H. Merrill Jr., J.D. Lambeth, J. Biol. Chem. 261 (1986) 12616. [21] 1. Granberg, W. Forsling, S. Sjoberg, Acta Chem. Scand. 36 (1982) 819. [22] K. Tanaka, R.G. Bates, Anal. Chem. 53 (1981) 1021. [23] D. Masi, C. Mealli, M. Sabat, A. Sabatini, A. Vacca, F. Zanobini, Helv. Chim. Acta 67 (1984) 1818. [24] J.L. Hall, J.A. Swisher, D.G. Brannon, T.M. Liden, lnorg. Chem. 1 (1962) 409. [25] P. Kopf-Maier, H. Kopf, Chem Rev. 87 (1987) 1137. [26] 1. Haiduc, C. Silvestru, Coord. Chem. Rev. 99 (1990) 253. [27] M. Zimmer, Chem Rev. 95 (1995) 2629. [28] A.I. Vogel, A Text Book of Quantitative Inorganic Analysis, 3rd ed., Chap. 3, Longmans, London, 1961. [29] G. Gran, Analyst 77 (1952) 661. [30] P. Passacantilli, private communication. [31] G. Arena, G. Maccarrone, private communication. [32] P. Gans, A. Sabatini, A. Vacca, Coord. Chem. Rev. 120 (1992) 389. [33] R.D. Hancock, J. Chem. Soc. Dalton Trans. 3 (1980) 416. [34] L.C. Van Poucke, Z. Eekhaut, Bull. Soc. Chim. Beiges 81 (1972) 363. [35] M. Eckert-Maksic, in S. Patai (Ed.), The chemistry of functional groups-Supplement E2: The chemistry of hydroxyl, ether and peroxide groups, Wiley, New York, 1993, pp. 299. [36] H.C. Brown, D.H. Mc Daniel, O. Hofliger, in E.A. Braude and F.C. Nachod (Eds.), Determination of organic structures and physical methods, Academic Press, New York, 1955, pp. 567. [37] J. March, Advanced Organic Chemistry, 2nd ed., Mc Graw-Hill Book Company, New York, 1977, pp. 243.
S. Canepari et al./ Talanta 44 (1997) 2059-2067
[38] P. Gans, A. Sabatini, A. Vacca, J. Chem. Soc. Dalton Trans., 1985, pp. 1195 [39] R.M. Smith, A.E. Martell, Critical Stability Constants, vol. 4, Plenum Press, New York, 1976, pp. 8. [40] I. Granberg, S. Sjoberg, Acta Chem Scand. 33 (1979) 531. [41] H. Ohtaki, Y. lto, J. Coord. Chem. 3 (1973) 131.
2067
[42] R.J. Breuehlman, F.H. Verhoek, J. Am. Chem. Soc. 70 (1948) 1401. [43] I_.E. Orgel, An Introduction to Transition-Metal Chemistry-Ligand-Field Theory, 2nd ed., Chap. 5, Methuen and Co, London, 1966. [44] R.B. Martin. Comments Inorg. Chem. 18 (1996) 249.
ELSEVIER
Protonation and silver(I) complex-formation equilibria of some amino-alcohols Silvia Canepari b Vincenzo Carunchio ~, Paola Castellano a, Antonella Messina
~'*
" Department q/' Chemistry, Universi O, of Rome, La Sapienza, P.le Aldo Moro, 5-00185 Rome. ltah' h ~TMS-Bioanalitieal Chemist O, Section, Universi O, q/' Rome, La Sapienza, P.le Aldo Moro, 5-00185 Rome, lta@
Received 7 August 1996; receivedin revised lbrm 21 January 1997: accepted 4 February 1997
Abstract
Formation constants of the silver(I) complexes with some amino-alcohols have been determined at 25°C in 0.5 M KNO3 by means of two independent potentiometric measurements employing glass and silver electrode. The ligands considered are: sec-butylamine, 2-amino-l-propanol, 2-amino-l-methoxy-propane, 2-amino-2-methyl-l-propanol, 2amino-l-butanol, 2-amino-l-pentanol, 2-amino-l-hexanol, 2-amino-l,3-propanediol, 2-amino-l,3-hexandiol, 2-amino2-methyl-l,3-propanediol and Tris(hydroxymethyl)-aminomethane. Protonation constants of the selected ligands have also been determined. Calculations were made using HYPERQUAD computer program. The influence exerted by the introduction of hydroxy groups and by the presence of alkyl residuals in the ligand structure on the formation equilibria, is discussed. © 1997 Elsevier Science B.V. Keywords: Amino-alcohols; Complex formation equilibria: Silver(I)
1. Introduction
The amino-alcohols represent a largely diffused class of biologically active substances. They characterise the moiety of several drugs and natural substances like ephedrine, adrenaline and noradrenaline and of some naturally occurring sphingoid like sphingosine and sphinganine. Due to the coordinating properties of amino and hydroxy groups, the study of formation equilibria of complex species with metal ions can help to clarify the biological role of these substances. The knowledge of complexing properties of amino-alcohols is not yet exhaustive and only a few data related to their * Corresponding author.
complex equilibria are known [1 14]. In particular sphingosine and sphinganine in the last few years have received great attention because of their activity as inhibitors of tumoral cells growth [15-18]. Notwithstanding this, the mechanism of their biological activity has not yet been clarified and their pharmaceutics application is nowadays hampered by the cytotoxic effect shown at high concentration [19]. However, it is known that the antitumoral effect of these sphingoids depends on the structure of their polar head and in particular on the presence of an amino-group, positively charged, in position 2 [15,20], so that when the -NH2 is chemically bonded, as an example by acylation, the inhibiting effect does not occur.
0039-9140/97;$17.00 ¢7 1997 Elsevier Science B.V. All rights reserved. Pll S0039-9 140(97)00044-1
2060
S. Canepari et al. / Talanta 44 (1997) 2059-2067
Then, is presumable that a coordination bond involving the amino group also influences the pharmacological activity of these compounds. Furthermore, as the interaction with metal ions is usually reversible and rapid, the study of complex formation equilibria may allow an efficient regulation of active species concentration. In this context, the present paper is concerned with the study of complexation equilibria of some amino-alcohols with Ag(I). The considered ligands are: sec-butylamine (L0, 2-amino-1-propanol (Ln) , 2-amino-l-methoxy-propane (LIll) , 2amino-2-methyl-l-propanol (Lw), 2-amino-l-butanol (Lv), 2-amino-l-pentanol (Lw), 2-amino-1hexanol (Lw0, 2-amino-l,3-propanediol (Lvni), 2-amino-l,3-hexandiol (L~x), 2-amino-2-methyl1,3-propanediol (Lx) and Tris(hydroxymethyl)aminomethane (Lx0. The ligands, which have a structure similar to the polar head of sphingosine and sphinganine, have been chosen in order to evidence the effect of the number of substituent hydroxy-groups and of the addition of alkyl residual on the coordinating properties of 2-amino-alcohol moiety. The study has been extended to Tris(hydroxymethyl)-aminomethane (TRIS, THAM), whose complexing properties have been not sufficiently documented even if it is often added as a buffer to solutions prepared to investigate biochemical reaction involving transition metal ions [12-14,21-
24]. Among the elements of the same periodic group, silver has been the least studied one, probably due to its extremely low presence in biological systems. However, our attention has been focused on silver(I) compounds, firstly because of the lack of thermodynamic data and secondly because, as already observed for other metal ions whose presence is not relevant in biological systems [25-27], their applicative importance can not be excluded.
2. Experimental 2.1. Materials
All the commercial products employed for
the preparation of L n I and L i v , w e r e by Fluka. Twice distilled and deionized water was used for preparing all solutions. All the reagents were obtained in high purity. Ultra pure KNO 3 (Suprapur grade, Merck) was used as ionic strength buffer. Silver(I) nitrate (Carlo Erba) stock solutions were standardized by titration with standard reagent NaCI (Merck) and using K2CrO 4 (Merck) as indicator following the Mohr's method [28]. Ligand stock solutions. As all the considered ligands are highly hygroscopic and uptake easily CO2, the direct weighing was not possible and all the solutions had to be handled in strictly air-free conditions. Stock solutions were then prepared by taking a volumetric sample in a glove-box under an argon atmosphere and diluting with appropriate volumes of water. The titre was successively controlled from series of crossed acid-base titrations with a potentiometric end point detection. 2-Amino-l,3-propanediol has been supplied as oxalate salt. The quantitative precipitation of oxalate ion with calcium(II) was then performed before preparing stock solutions. The presence of a 0.5% excess of calcium(II) nitrate, needed in order to assure the quantitative precipitation of oxalate, has been checked by EDTA titration. Potassium hydroxide 0.1 M (COz-free solution, Merck) was standardized by potentiometric titrations against a solution of potassium hydrogen phthalate (Merck pro analysis, dried at 120°C) using the Gran method [29] for the evaluation of the end point. Nitric acid solution (0.1 M) was prepared by diluting the pure concentrated product (Merck) and standardized by potentiometric titrations against tris(hydroxymethyl)aminomethane (Merck), dried at 100°C for 24 h. An independent check of the analytical procedure was also carried out by a direct titration of potassium hydroxide solution against the nitric acid.
S. Canepari et al. / Talanta 44 (1997) 2059-2067
2. I. 1. Synthesis of 2-amino- 1-hexanol and 2-amino- 1,3-hexandiol As 2-amino-l-hexanol and 2-amino-l,3-hexandiol are not commercially available, these products were prepared following the synthetic path summarized in Scheme 1 [30].
2.1.2. E M F measurements Glass electrode. The alkalimetric titrations were carried out adding KOH to the test solution placed in a thermostatic vessel, under agitation and under a stream of purified argon presatured with a 0.5 M KNO3 solution. The temperature control (25_+0.1°C) was achieved by means of water circulation in the jacketed vessel from a water thermocryostat (thermostat HAAKE DC3 and chryostat HAAKE K15). The initial volumes of the titrated solutions were 25 ml. The KOH solution was delivered in the titration vessel by a Metrohm Dosimat 655 digital burette with a total volume of 5 ml. In order to have CO2-free potassium hydroxide, the solution was preserved in a bottle stopped by a soda lime plug and kept under argon atmosphere. In these conditions the solution was stable up to 1 month. Potentiometric titrations were carried out using a Metrohm mod. 665 pH/mVmeter, with a glass electrode (INGOLD) and a double junction Ag/ AgCI electrode (INGOLD) as a reference. The outer solution of the reference electrode was 0.5 M KNO 3. The potentiometric apparatus was made completely automatic by applying an IBM mod.30 personal computer with a parallel interface and using an appropriate software program. [31] OBn Br~
osn m
oe~
b,c,d,e •
o,1
NH2
A 4O%
s
~
/ g,h.i
OB.
Br
2061
The electrode standard potentials E ° were determined every day from the acid region of the titration of a 0.005 M HNO 3 solution (0.5 M ionic strength buffer) against 0.1 M KOH. Similarly the response of the glass electrode was checked every week evaluating the value of pKw from the basic portion of the calibration curve. The linearity of E versus pH function was also checked; the slope of the straight line obtained was, within the experimental errors, coincident with the theoretical value. For the evaluation of the protonation constants, solutions 0.5 M in KNO 3 containing different quantities of ligand and acidified with a known excess of HNO3 were titrated with potassium hydroxide solution. For the formation constant determination of Ag(I) complexes, solutions at various concentrations of Ag(I) nitrate and ligands were titrated with KOH. Before the titrations all the solutions were acidified with a known excess of nitric acid. The ionic strength was always adjusted to 0.5 M by adding solid potassium nitrate. Ag electrode. All the measurements regarding AgO) complexation equilibria were repeated also by employing an Ag electrode (Metrohm) in order to have an independent measurement. The calibration of the Ag electrode was performed by measuring the potential of five solutions at different Ag(I) concentration containing KNO 3 0.5 M as ionic strength buffer. E ° and slope were determined by linear regression of data (r2= 0.99) and the obtained values were in good agreement with the theoretical ones (error < 1%). The operative conditions employed for the determination of stability constants were the same as those related to glass electrode measurements. All the calculations for the calibration of the glass electrode and for the determination of ligand protonation and complex formation constants were carried out by employing HYPERQUAD least squares computer program [32].
3. Results and discussion NH2
2-aminQ-1,3-hexandiol
Scheme 1.
Particular attention had then to be paid to the choice of concentration ranges, considering the
S. Canepari et al. /Talanta 44 (1997) 2059 2067
2062
Table 1 Initial analytical concentrations (raM) of ligand (CL) and acid (CH) and - l o g [ H +] ranges for the determination of the protonation constant of 2-amino-l-hexanol at l = 0.5 M KNO3 and T = 25°C Ligand
Expt. no
1, 2, 3 2-Amino-l-hex- 4 anol 5
CL 1.8 3.7 6.7
CH
- l o g [ H 7] ranges
3.0 3.08-10.61 5.0 3.07 10.34 10
3.08-10.43
method sensitivity and the best operative conditions for the calculation of the constants. Good results were obtained in all the examined systems when the ligand concentration was maintained lower than 7 raM. In fact, by a preliminary examination of the experimental data, turns out that the ligand concentration influences the results. In particular, there is a progressive worsening of the fit on increasing the ligand concentration in the titration vessel.
3.1. Protonation equilibria The calculation of the protonation constants has been performed by elaborating the data obtained by five titration runs for each system. In Table 1 the operative conditions are reported for 2-amino-l-hexanol as an example. Similar conditions have been employed for all the other ligands. As can be noted, one of the titrations has been repeated three times in order to check the reproducibility of the experimental measurements. Repeated runs also allowed a better control of the ligand concentrations, whose determination requires particular care (Section 2). In fact the amounts of ligands in the titration vessel (COL) have been refined by HYPERQUAD program and the resulting values were in a good agreement with the calculated ones (2% error). In Table 2 the pK,~ values obtained by the simultaneous elaboration of all the titration curves are reported. The error between brackets is the standard deviation provided by HYPERQUAD calculations. The statistical parameters are satisfactory also considering that, due to the
basicity of ligands, E.F.M. measurements had to be performed in the basic region. The pK, values related to sec-butylamine [33], 2-amino-l-butanol [34] and THAM [12 14,2124] are in a good agreement with literature data. From the examination of the reported data, two main effects on the acidity of protonated ligands can be evidenced. Firstly, the insertion of hydroxy group causes a relevant lowering of pK~ values. This effect could be caused either by the intramolecular hydrogen bond formation [35], which is stronger in the deprotonated form than in the protonated one, or by an electron-withdrawing polar effect due to the hydroxy group. Comparing the pK~, values related to L , and L m it is evident that the second effect is prevalent. Furthermore, the methoxy group seems to be more effective than hydroxy group, in spite of the lower electronic-withdrawing capabilities of this group with respect to hydroxy one. This reversal, already observed for similar compounds [36] was ascribed to the modification of the true inductive effect of the hydroxy group by solvation. Secondly, the addition of alkyl residuals in the ligand structure causes an increase of the pK~ values. The effect is mainly ascribable to a field effect of the alkyl group, which results in an increase in the electron density on the nitrogen atom [36]. Furthermore, from the results it can be Table 2 Protonation constant values and related statistical parameters for the considered ligands at I = 0.5 M KNO 3 and T = 25°C Ligands
pK,,
Z2
o"
Sec-butylamine 2-Amino-l-propanol 2-Amino- 1-methoxy-propane 2-Amino-2-methyl- 1-propanol 2-Amino-l-butanol 2-Amino-l-pentanol 2-Amino-l-hexanol 2-Amino- 1,3-propanediol 2-Amino-1,3-hexandiol 2-Amino-2-methyl- 1,3propanediol Tris(hydroxymethyl) aminomethan
10.599(3) 9.450(6) 9.424(2) 9.612(6) 9.554(5) 9.766(3) 9.90(6) 8.786(I) 8.901(2) 8.780(6)
19.23 13.54 10.29 14.37 15.00 17.50 18.41 10.32 11.73 15.68
3.72 2.01 1.34 2.98 2.80 3.01 4.00 1.09 3.21 2.90
8.156(1 )
5.23
0.56
S. Canepari et al. ' Tahmta 44 (1997) 2059- 2067
Table 3 Initial analytical concentrations (mM) of ligand (C0, metal (CM) and acid (C,), CL/CM ratios and p[H] ranges employed in calculations of complexation constant of 2-amino-l-hexanol -Ag(l) at I = 0 . 5 M KNO3 and T = 25°C Expt. no
C~
CM
CL/CM
Cii
p[H] ranges
I 2 3 4 5 6 7 8
4.10 1.49 2.84 3.70 4.24 3.20 2.01) 4.115
4.10 0.75 1.42 1.85 1.41 1.09 0.69 0.81
1 2 2 2 3 3 3 5
6.10 3.50 4.85 5.70 6.10 3.52 2.21) 6.10
7.42 8.26 7.78- 9.25 7.61) 8.93 7.50 8.85 7.92 9.48 7.78 9.56 7.65 -9.75 7.54 10.20
noted that the insertion of methyl group in position 2 in L~v and Lx causes an increase of pK~ lower than the one expected on the basis of the field effect. This is in agreement both with the steric hindrance on the quaternary carbon, which could stabilize the intramolecular hydrogen bond with the consequent lowering in basicity and with the differential hydration [37] which is less effective when a -CH~ group is added in position 2.
3.2. Complexation equilibria In order to evidence both mononuclear and polynuclear species, the determination of complexation constants has been carried out for each considered system on a stock of eight solutions containing various CL/CM ratios and at different ligand and metal initial concentrations. In Table 3 are reported, as an example, the operative conditions for 2-amino-l-hexanol-Ag(I) system. As can be noted, the - l o g [ H +] ranges are variable depending on experimental conditions and have been identified considering all the equilibria involved, in order to avoid precipitation equilibria. All the experimental runs were firstly elaborated by H Y P E R Q U A D one by one, afterwards an elaboration of each set obtained with glass or Ag electrode was performed. Finally all the titration curves were elaborated simultaneously. Different equilibrium models were examined. The species ML, MLe M2L, M2L2, MLOH, ML2OH
2063
and ML 3, were systematically considered in different combinations. These species were either rejected by H Y P E R Q U A D or caused a worsening of the statistical parameters a and Z 2 [38] or lead to not acceptable chemical models. The absence of systematic errors was tested refining also the dangerous parameters as COM, COL, Co. (total mmoles of metal, ligand and proton, respectively, in the reaction vessel). The deviation of calculated values from input ones can be considered within the experimental error (about 2%). Hydrolysis constants [39] were considered in the calculation. However they do not influence the log fl values, apart from a slight ( ~ 1%) effect on the results related to the titration curves with C L / C M = 1,
The calculations related to Ag electrode measurements give statistical parameters better than those obtained with glass electrode, probably due to the working pH range in which the alkaline error is not negligible. Anyway, the stability constants obtained with the two independent measurements are in good agreement as evident in Table 4 in which the results related to Ag and glass electrode measurements for Ag(I)-2-amino1,3-propanediol system are reported. In Table 5 the results of the overall elaboration obtained for all the considered systems, are summarized. Two different models, i.e., M L - M L 2 (model 1) and ML-ML2-MLOH (model 2), are reported as the most significative ones. Is worth noting that is not possible from the potentiomettic measurements, to deduce whether the species indicated as M L O H is a hydroxy complex or a complex where ligand acts as a chelate, with the deprotonation of the hydroxy groups. Due to steric reasons and considering logfl values the first alternative seems to be more probable [21]. On the other hand, the enhancement of the affinity for OH-compared with the hydrated silver ion has been already observed from other authors in the presence of aromatic [40] or aliphatic [21,41] N-ligands. From a critical view of the Table 5 it is evident that the model 2 is the most representative at least for the ligands L w . , Lix, Lx and Lx~, in which two or three hydroxy groups are present. The
S. Canepari et al. ,/Talanta 44 (1997) 2059-2067
2064
Table 4 Formation constants of complexes of 2-amino-l,3-propanediol with AGO), and related statistical parameters obtained at I = 0.5M KNO3 and T = 25°C Ag electrode
log fl1.1,o (log K 0 log f12,1,0 log K 2 log fl~ j _ ~ log K~(AgL +) 2' 2 o"
Glass electrode
Model 1
Model 2
3.39 (1) 6.72 (1) 3.33 --28.00 3.85
3.407 (6) 6.64 ( 1) 3.23 - 6.24 (4) -9.6 21.00 1.58
Total
Model 1 3.40 (6) 6.79 (2) 3.39 -62.13 5.18
presence of the MLOH species is evident in Fig. 1 in which the comparison between the two models for Ag(I)-THAM system is reported as a plot of experimental and calculated data. For all the other systems, the presence of ternary species MLOH is uncertain and its addition to the speciation model does not improve significantly the fitting. This may be due both to the reduction in acidity of the bonded water (see log Ka(AgL +) values in Table 5) and to the formation constant values, so that the presence of ML2 species is predominant in the examined experimental conditions (Fig. 2 and Fig. 3). As far as the stability order of the complexes is concerned, the same trend as observed for pK~
Model 2
Model 1
Model 2
3.39 (5) 6.69 (3) 3.30 - 6.40 (7) -9.8 30.20 2.60
3.38 (1) 6.76 (1) 3.38 --50.25 4.31
3.39 (1) 6.65 (2) 3.26 - 6.34 (5) -9.7 22.15 2.19
values has been found. It is then evident that the coordination is due to the aminic nitrogen and that no chelate structures are formed. As found
&O
6,0
7,0
0,0
9,O
I0,0
Fig. 2. Plots of formation percentages vs p[H] related to ML, ML 2 and MLOH species for Ag(I)-THAM. CL = 3.2 m M , C M = 1.1 m M , C H = 3 . 2 raM. Formation constants as reported in Table 5 (model 2).
,.,:::L'.I" 3,60
3,40
i
0. 20.
300 0,10
0,15
0,20
0,25 0,30 tltrant (mL)
0,35
040
0,45
Fig. I. Comparison between experimental and calculated curves related to either model l (ML, ML2) or model 2 (ML, ML2, M L O H ) for the Ag(I)-THAM system with Ag electrode. C L = 3.2 m M , CM = l.l m M , C H = 3.5 m M , 1 = 0 . 5 M KNO~ and T = 25°C.
0. - 5.5
6.5
7,5
&5
9.5
10,5
Fig. 3. Plots of formation percentages vs p[H] related to ML, M L : and MLOH species for Ag(I)-2-amino-l-propanol. C L = 3.2 raM, C M = 1.1 m M , CH = 3.2 m M . Formation constants as reported in Table 5 (model 2).
S. Canepari et al.
Talanta 44 (1997) 2059-2067
,.o o
o II
keo o_
~
Z
oo i
II ~G ~EP. -o
~'~ o
-
I
a2~ =
=
~
~'~
._= .~
~
I ~
~.~
..Y
~
~
'~
~~
.-~
7"
~'
2065
2066
S. Canepari et al. /Talanta 44 (1997) 2059-2067
for other systems [6,42], (log KO/pK, ratio is constant for all the considered ligands, the average value being 0.37 _+ 0.02. It is worth noting that in the systems containing Liv and L x as ligands, the stability of the ML species is lower than the one expected considering the pK~ values. These results can be ascribed to the steric hindrance which is more influent on the tetrahedral structure of ML species with respect to the linear one characteristic of ML2 species in silver(I)-aliphatic amines systems. Furthermore, it can be noted that log K 2 values are higher than log K~ ones, with the exclusion of 2-amino-l,3-hexandiol and 2-amino-l,3propane- diol. This different behaviour can be explained assuming that the K values, as suggested by Orgel [43] and interpreted in detail by Martin [44] in terms of cooperative effect, are made up of two parts which are in competition each other. The first is mainly due to a statistical-electronic effect and causes a fall off of stability from K~ to K,,. The second part can be attributed to an energetic effect associated to the change of the complex structure from tetracoordinate to linear and determinates an increasing of stability. Depending on which is the prevalent effect, K2 value will be lower or higher than KI.
Acknowledgements This work was supported financially by a National Project of Italian M.U.R.S.T., Rome, Italy.
References [l] A. Braibanti, G. Mori, F. Dallavalle, L. Leporati, J. Chem. Soc. Dalton (1975) 1319. [2] A. Braibanti, G. Mori, F. Dallavalle, J. Chem. Soc. Dalton (1976) 826. [3] A. Braibanti, G. Mori, F. Dallavalle, J. Chem. Soc. Dalton (1979) 1050. [4] C. Van Poucke, Z. Eeckhaut, Bull. Soc. Chim. Beiges 81 (1972) 363. [5] B.A. Timini, D.H. Everett, J. Chem. Soc (B) (1968) 1380. [6] D.J. Alner, R.C. Lansbury, A.G. Smeeth, J. Chem. Soc (A) (1968) 417.
[7] J.J. Christensen, R.M. lzatt, D.P. Wrathall, L.D. Hansen, J. Chem. Soc (A) (1969) 1212. [8] A.Y. Sychev, A.P. Gerbeleu, P.K. Migal, Russ. J. Inorg. Chem., 8 (1963) 1081. [9] P.K. Migal, A.N. Pushnyak, Russ. J. lnorg. Chem. 4 (1959) 601. [10] P.K. Migal, A.N. Pushnyak, Russ. J. Inorg. Chem. 5 (1960) 293. [11] J.M. Antelo, F. Arce, J. Casado, M.C. Regal, A. Varela, J. Chem. Research(s) (1982) 144. [12] L. Bologni, A. Sabatini, A. Vacca, lnorg. Chim. Acta 69 (1983) 71. [13] K.S. Bai, E. Martell. J. lnorg. Nucl. Chem. 3l (1969) 1697. [14] W. Forsling, Acta Chem. Scand. A 32 (1978) 57. [15] Y.A. Hannun, R.M. Bell, Science 243 (1989) 500. [16] A.H. Merrill Jr., M. Sereni, V.L. Stevens, Y.A. Hannun, R.M. Bell, J.M. Kinkade Jr., J. Biol. Chem. 261 (1986) 2610. [17] M. Faucher, N. Girones, Y.A. Hannun, R.M. Bell, R.J. Davis, J. Biol. Chem. 263 (1988) 5319. [18] L.A. Speizer, M.J. Watson, U Brunton, Fed. Proc. 46 (1987) 2065. [19] Y.A. Hannun, R. Bell, Science 235 (1987) 678. [20] E. Wilson, M.C. Olcott, R.M. Bell, A.H. Merrill Jr., J.D. Lambeth, J. Biol. Chem. 261 (1986) 12616. [21] 1. Granberg, W. Forsling, S. Sjoberg, Acta Chem. Scand. 36 (1982) 819. [22] K. Tanaka, R.G. Bates, Anal. Chem. 53 (1981) 1021. [23] D. Masi, C. Mealli, M. Sabat, A. Sabatini, A. Vacca, F. Zanobini, Helv. Chim. Acta 67 (1984) 1818. [24] J.L. Hall, J.A. Swisher, D.G. Brannon, T.M. Liden, lnorg. Chem. 1 (1962) 409. [25] P. Kopf-Maier, H. Kopf, Chem Rev. 87 (1987) 1137. [26] 1. Haiduc, C. Silvestru, Coord. Chem. Rev. 99 (1990) 253. [27] M. Zimmer, Chem Rev. 95 (1995) 2629. [28] A.I. Vogel, A Text Book of Quantitative Inorganic Analysis, 3rd ed., Chap. 3, Longmans, London, 1961. [29] G. Gran, Analyst 77 (1952) 661. [30] P. Passacantilli, private communication. [31] G. Arena, G. Maccarrone, private communication. [32] P. Gans, A. Sabatini, A. Vacca, Coord. Chem. Rev. 120 (1992) 389. [33] R.D. Hancock, J. Chem. Soc. Dalton Trans. 3 (1980) 416. [34] L.C. Van Poucke, Z. Eekhaut, Bull. Soc. Chim. Beiges 81 (1972) 363. [35] M. Eckert-Maksic, in S. Patai (Ed.), The chemistry of functional groups-Supplement E2: The chemistry of hydroxyl, ether and peroxide groups, Wiley, New York, 1993, pp. 299. [36] H.C. Brown, D.H. Mc Daniel, O. Hofliger, in E.A. Braude and F.C. Nachod (Eds.), Determination of organic structures and physical methods, Academic Press, New York, 1955, pp. 567. [37] J. March, Advanced Organic Chemistry, 2nd ed., Mc Graw-Hill Book Company, New York, 1977, pp. 243.
S. Canepari et al./ Talanta 44 (1997) 2059-2067
[38] P. Gans, A. Sabatini, A. Vacca, J. Chem. Soc. Dalton Trans., 1985, pp. 1195 [39] R.M. Smith, A.E. Martell, Critical Stability Constants, vol. 4, Plenum Press, New York, 1976, pp. 8. [40] I. Granberg, S. Sjoberg, Acta Chem Scand. 33 (1979) 531. [41] H. Ohtaki, Y. lto, J. Coord. Chem. 3 (1973) 131.
2067
[42] R.J. Breuehlman, F.H. Verhoek, J. Am. Chem. Soc. 70 (1948) 1401. [43] I_.E. Orgel, An Introduction to Transition-Metal Chemistry-Ligand-Field Theory, 2nd ed., Chap. 5, Methuen and Co, London, 1966. [44] R.B. Martin. Comments Inorg. Chem. 18 (1996) 249.