Complexation of a four-strand tetraalcohol with labile metal ions probed by electrospray mass spectrometry

www.elsevier.com/locate/ica Inorganica Chimica Acta 328 (2002) 159– 168

Complexation of a four-strand tetraalcohol with labile metal ions probed by electrospray mass spectrometry Geoffrey A. Lawrance *, Mark J. Robertson, Sutrisno, Ellak I. von Nagy-Felsobuki Discipline of Chemistry, School of Biological and Chemical Sciences, The Uni6ersity of Newcastle, Callaghan 2308, Australia Received 11 July 2001; accepted 4 October 2001

Abstract The saturated, stereodefined tetraalcohol 2,3,5,6-endo,endo,endo,endo-tetrakis(hydroxymethyl)bicyclo[2.2.1]heptane (tetol, L1) and the simple alcohol butane-1,3-diol (L2) form complexes with alkali metal ions (lithium, sodium, potassium, rubidium and caesium), alkali earth cations (magnesium, calcium, strontium and barium) and Ga(III) and Ce(IV) in aqueous solution, characterised by electrospray ionisation mass spectrometry (ESMS). Metal ion exchange between the Li+ complex of L1 and the other metal ions is rapid, with a range of M(L1)nm + species detected, in addition to solvated species. With the alkal metal ions, M(L1)+ and M(L1)2+ are dominant, although speciation varies with metal ion size. For the alkaline earth ions, a range of complex ions up to n =8 are observed, although n=1–3 dominate. A preference for M(L1)22 + with Mg2 + versus M(L1)32 + with Ca2 + may again relate to a larger ion size. For the higher-charged Ga(III) and Ce(IV), hydroxo species M(OH)(L1)n(m − 1) + are dominant reflecting bulk solution behaviour, which the ESMS studies appear to map generally. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Polyalcohol; Alkali metal ions; Alkaline earth metal ions; Metal complexes; Electrospray ionisation mass spectrometry

1. Introduction Polyalcohols (polyols), as potential ligands, offer a set of hydroxy groups for binding metal ions. The interaction with metal ions is generally weak, and the complex formation by a neutral polyalcohol may not be significant in aqueous solution, although deprotonated polyalcohols represent strong and efficient metal binding agents [1], as, since polyols are weak acids (pKa \ 12), the deprotonated polyols are even stronger bases than polyamines. Due to the high tendency of the deprotonated hydroxy group to bind to more than one metal ion, polyalcohols are particularly suitable for the stabilisation of polynuclear complexes, with bridging by m-alkoxo groups a basic structural element in such complexes [2]. For metal ions in alkaline media, complex formation with polyols competes with hydrolytic polymerisation reactions. Moreover, partially hy* Corresponding author. Tel.: + 61-2-492 15471; fax: + 61-2-492 15472. E-mail address: [email protected] (G.A. Lawrance).

drolysed species (polynuclear polyolato complexes with additional oxo or hydroxo bridges) are also known. Such aggregates can be regarded to be composed of an inorganic metal-oxide or metal-hydroxide core that is protected by the peripherally coordinated polyolato ligand, thereby preventing complete hydrolysis and yielding species that are intermediate between welldefined low molecular weight complexes and solid phases. This intermediate state is currently of particular interest with respect to applications in materials science (e.g. the sol –gel technique) [3]. The wide structural diversity found in the naturally occurring polyols gives rise to a variety of different geometries for donor sets, each with characteristic coordination properties. This diversity could possibly be used to control complex stability in terms of metal ion size [4]. Further, selective metal complex formation could be of interest in terms of a controlled protection or activation of individual hydroxy groups providing opportunities for devising new routes in synthetic organic chemistry [5]. The ability of a variety of natural occurring polyol ligands, such as glycerol or sorbitol, to act as sequester-

0020-1693/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 0 - 1 6 9 3 ( 0 1 ) 0 0 7 2 8 - 9

160

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

ing agents— preventing formation of solid metal oxides or hydroxides in alkaline aqueous solution— has been well known for many years [6]. However, the correct structures of the complexes that were formed were not elucidated until a number of recent X-ray crystal structures [7,8]. The use of polyols as selective receptors for metal ions has been reviewed [9]. However, the detailed synthetic and structural studies of sugar– metal complexes have been limited due to their complicated stereochemistry, multi-functionality and the hygroscopic nature of species. Synthetic polyols, particularly those with pre-defined isomeric character, present themselves as more practical molecules for study of metal ion– polyol interactions. We have sought ways of examining the relatively weak complexation between metal ions and neutral polyalcohols in aqueous solution, and have identified electrospray mass spectrometry (ESMS) as a potentially revealing method [10]. Spectroscopic methods have been used to investigate polyol and related polyaminol ligand complexation previously [11], but were not able to provide definitive answers regarding solution structures. Examination of polyol– metal ion interactions by mass spectrometry (MS) does not appear to have been pursued previously, although complexation of weak interactions of the polyethers 18-crown-6 and cryptand[2.2.2] with alkali metal cations using ESMS has been probed [12]. In the present study, we have applied ESMS to characterise the tetraalcohol 2,3,5,6endo,endo,endo,endo-tetrakis(hydroxymethyl)bicyclo[2.2.1]heptane (tetol, L1), a saturated, stereodefined molecule, interacting with alkali metal ions (lithium, sodium, potassium, rubidium and caesium), alkali earth cations (magnesium, calcium, strontium and barium) and higher-charged ions, gallium(III) and cerium(IV). The interaction of a simple chelating model ligand with alkali metal cations is also examined. Tetol is the first example of a four-strand ligand where four pendant groups extend from a rigid core in one direction. This study involves the direct observation of metal ion exchange in solution by means of ESMS influenced by the relative stability of complexes formed. The ‘soft’ ionisation technique of ESMS, which has now been applied to a range of both inert and labile complexes, appears to reflect speciation in solution [13,14]. Thus this ESMS study offers the potential to provide information on complexation in bulk media.

2. Experimental

2.1. Materials All reagents were commercially certified ACS grade or better. Chloride salts of metal ions were usually

employed, except for cerium(IV), where the sulfate was used. Butane-1,3-diol (L2) was obtained from Sigma (98%). Doubly deionised Milli-Q water (Millipore) was used as solvent. All chemicals were used as purchased without further purification. The tetraalcohol 2,3,5,6endo,endo,endo,endo-tetrakis(hydroxymethyl) bicyclo[2.2.1] heptane (tetol) was synthesised by a variation [10] of a reported method [15] isolated as the lithium salt LiL1Cl. Some was converted to KL1Cl salt by ion exchange on a column of Dowex 50W cation exchange resin. The LiL1Cl has been characterised by an X-ray crystal structures analysis [10].

2.2. Electrospray ionisation mass spectrometry For LiL1Cl and KL1Cl, sample solutions were prepared by dissolving the solid samples in deionised water. For butane-1,3-diol (L2), the solution was prepared by dissolution of a weighed amount of the alcohol in water. The concentration of all sample solutions was 10 − 4 M. For investigation of cation exchange of lithium complexes with other cations, aqueous solutions of LiL1Cl with the cations in equimolar concentrations were prepared. Similarly, for the simple model ligand butane-1,3-diol (L2), 1:1 solutions of metal:ligand were prepared. All analyte samples (in 1.8-ml vials) were introduced readily using an auto injector, which was thoroughly rinsed with solvent between injections. All the ESMS experiments were performed on a VG Platform II single quadrupole mass spectrometer (Micromass Ltd, Altrincham, UK) coupled to a high performance liquid chromatography (HPLC) binary pump system. For all spectral acquisitions, the tip of the capillary was at a potential of 9 3.5 kV relative to ground. The source temperature was maintained at 80 °C, and nitrogen was used as the bath and nebulising gas. The mobile phase solution was the same as the solvent of the analyte solution. An optimum flow rate of 10 ml min − 1 was employed. The cone voltage (CV) was varied between +20 to + 60 for positive-ion and − 20 to − 50 V for negative-ion mode. A Reodyne injector fitted with a 10-ml loop was used to inject the sample solution into the flow of the mobile phase. Mass spectra were acquired by scanning the quadrupole mass filter from m/z 2000 to 2, and ions were detected by means of a scintillator detector. Approximately 40 scans were summed to give a mass spectrum. All data were acquired and processed using the Micromass MassLynx system. Experimental peak values throughout this study are identified by the m/z ratio of the most abundant peak in the parent group. Calculated m/z values tabulated are those based on the most abundant isotopes. Peak intensities are cited as percentages of the base (major) peak intensity (%BPI).

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

3. Results and discussion The molecule L1 is a stereochemically defined molecule with four CH2OH arms pendant in one direction to a stereorigid but saturated bicyclo[2.2.1]heptane frame. The arms are sufficiently flexible to either wrap up a metal ion in a 1:1 mode, or else to share binding of ions with other molecules in dimeric or other polymeric motifs. The X-ray structure of the LiL1Cl complex displays a polymeric structure in the solid state [10]. However, this is not anticipated to persist in solution. To pursue the solution structure of the labile and relatively spectroscopically silent alkali and alkaline earth ion complexes, we have turned to the technique of ESMS, which has been used previously to

161

provide information about other labile systems, including inorganic clusters. Although not necessarily defining the bulk solution structure strictly, speciation in this ‘soft’ MS technique does appear to identify thermodynamically significant species.

3.1. Positi6e-ion ESMS of LiL 1Cl, KL 1Cl and H+L 2 The positive-ion mode ESMS for the LiL1Cl complex reveals a series of lithium complexes and their solvated ions (Fig. 1(a)). The dominant peaks at a CV of 30 V arise from complexes with observed m/z of 223.0, (100% BPI) and 439.3 (78% BPI) assigned to (LiL1)+ (calc. m/z 223) and (Li(L1)2)+ (calc. m/z 439). Further, smaller peaks assigned to higher ratio Li:ligand species and solvated ions are observed. For KL1Cl (Fig. 1(b)), a series of potassium complexes with increasing numbers of ligands (up to four) was observed [namely (KL1)+, (K(L1)2)+, (K(L1)3)+and (K(L1)4)+] as well as a series of solvated ions. For both lithium and potassium complexes, minor peaks at m/z 216.2, 198.2, 181.2 and 164.2 were identified at a CV of 30 V assigned to mono-protonated ligands that have lost successive hydroxy groups through dehydration reactions. Alcohol dehydration reactions to form alkenes can occur sequentially for L1 (Scheme 1), with loss of a water molecule in each step leading to formation of a double bond. The final step to a tetraene leads to a species that may not undergo ready protonation or complexation and, therefore, is less likely to be detected. The positive-ion ESMS spectra of the simple model ligand butane-1,3-diol exhibits similar behaviour. The spectrum is simple, with a major peak assigned to the monoprotonated diol (L2 + H)+ (m/z obsd. 91.1 (100% BPI), calc. 91). Dehydrated charged species are assigned to two smaller peaks (m/z obsd. 73.1 (25% BPI), calc. 73 and m/z obsd. 55.1 (20% BPI), calc. 55). The behaviour is consistent with the scheme:

Fig. 1. Positive-ion ESMS of (a) LiL1Cl and (b) KL1Cl complexes in water at neutral pH (cone voltage 30 V).

Scheme 1.

The behaviour of L2 in the absence of any metal ion confirms the chemistry observed in the more complex molecule L1. However, these reactions are minor at low CV, although they increase as the CV is increased. At the cone voltages typically employed in this study, this chemistry does not occur at a significant level for L1 at least. Addition of a mixture of Li+, Na+, K+, Rb+ and Cs+ produces a range of (ML)+ species and their solvated forms indicating that metal exchange is rapid as anticipated. This occurs with both L1 and the simple chelate L2; Fig. 2 illustrates the behaviour with L2. The

162

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

Fig. 2. Positive-ion ESMS of butane-1,3-diol (L2) after addition of equimolar amounts of Li+, Na+, K+, Rb+ and Cs+ at neutral pH (dehydrated ligand species are identified by asterisks; refer also to Scheme 1).

Fig. 3. A plot of %BPI of lithium complexes vs. species at different cone voltages.

relative peak intensities in the ESMS may reflect rela+ tive abundances, but not directly, since the Maq ions alone in equimolar concentrations do not produce peaks of equal intensity. In fact, they are significantly different in relative peak size. Variation of CV from 20 to 60 V provides additional information. Fig. 3 shows a plot of %BPI against species of lithium complexes. At CV 20 V, the most abundant peak is that of the (LiL2)+ ion. When the CV is increased from 20 to 60 V, the (LiL)+ ion becomes the dominant species and the other lithium complexes decrease. This would be anticipated due to an increase in the collision cross-section for these larger complexes. The trend towards the lower molecular weight species (LiL1)+ with increasing CV reflects the stabilisation of the smaller species, which have a much smaller collision cross-section and so under high CV conditions are more able to survive collisions with the neutral carrier gas in the cone-to-skimmer region. These trends are machine-dependent phenomena and do not reflect any bulk solution properties; in general, the lower the cone voltage the more likely is the technique to reflect bulk

solution equilibria. Consequently, the appearance of (LiL2)+ as a major species at low CV suggests that a species that binds lithium ion in a ‘sandwich’ structure between two sets of four alcohol donor groups may be particularly stable. The detailed assignments of the positive-ion ESMS of LiL1Cl, KL1Cl and butane-1,3diol (L2) in aqueous solution appear in Table 1. ESMS were recorded for samples with additional lithium ion added to a sample of LiL1Cl complex to yield molar ratios of L1:Li ranging from 1:10 to 1:1 for a 10 − 4 molar ligand concentration in water. Table 2 gives detail of %BPI of lithium complexes with different molar ratio from positive-ion ESMS at CV 20 V. The peak at m/z 223.2 assigned as (LiL)+ is the dominant species, and becomes progressively dominant as the concentration of Li+ grows. An excess of Li+ in solution does not contribute to any polymer complex formation. In contrast, however, Wang et al. [16] reported that upon addition of tenfold excess Na+ to aryl methyl ether, the peak intensity of polymer species in the ESMS spectra increased. The behaviour of (LiL1n )+ species in the ESMS reflects what is expected to take place in the bulk solution, with higher Li+ concentration favouring (LiL1)+ over other species. The negative-ion ESMS of LiL1Cl complex in water at a CV of −30 V shows two series of peaks. A series of chloride anion species (LiLClH)− m/z 260.9 (100% BPI), (L2Cl)−, 467.3 (42% BPI), (L3Cl)−, 683.3 (10% BPI) and a small peak due to (L4Cl)− 899.0 (1% BPI) are observed. The chloride anion compounds can be rationalised as a series of weak complexes formed by hydrogen bonding interactions. The other series observed is the deprotonated free-ligand, with both monomer (L1H)− and dimer (2L1H)− species at m/z 215.3 (28%) and 432.2 (12%), respectively. All observed species appear in Table 3.

3.2. Alkali cation exchange in the Li(L 1)+ complex From FAB MS studies of ligand-cation interactions, two types of binding experiments are in common usage. Competitive binding studies [17] are those in which a ligand is mixed with a molar equivalent of a series of metal ions. Since the metals are competing for a limited number of binding sites, the distribution of ionic complexes, revealed in the mass spectrum, should reflect the selectivity of binding. Relative peak intensity measurement [18] is where ratios of intensities of the signals resulting from ‘free’ and complexed ligands are taken as direct indicators of the extent of complexation. In this work, competitive binding studies were employed. ESMS were measured for an aqueous solution containing LiL1Cl (10 − 4 M) and a mixture of four other alkali metal chlorides (each 10 − 4 M). Major mass peaks attributable to the 1:1 ion:ligand and 1:2 ion:ligand complexes are observed, along with those of

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

163

Table 1 Positive-ion ESMS of LiL1Cl and KL1Cl in water at CV 20 V Species

Major ions observed a

m/z (Theoretical)

m/z (Experimental), (%BPI)

LiL1Cl

(L1***+H)+ (L1**+H)+ (L1*+H)+ (L1+H)+ (LiL1)+ (LiL1(H2O))+ (LiL1(H2O)11)+ (LiL12 )+ (LiL12 (H2O))+ (LiL13 )+ (LiL14 )+ (L1***+H)+ (L1**+H)+ (L1*+H)+ (L1+H)+ (KL1)+ (KL1(H2O)4)+ (KL12 )+ (KL12 (H2O)4)+ (KL13 )+ (KL13 (H2O)4)+ (KL14 )+

164 181 198 216 223 241 421 439 457 655 871 164 181 198 216 255 327 471 543 687 759 904

163.9, 181.1, 198.2, 216.0, 223.0, 239.5, 421.3, 439.3, 455.4, 655.7, 871.5, 164.3, 181.5, 198.3, 216.2, 255.3, 327.3, 471.2, 544.0, 687.2, 760.1, 905.2,

KL1Cl

a

5 25 8 11 100 30 11 88 6 18 8 4 7 9 9 100 10 60 10 28 10 4

Refer to Scheme 1 for ligand species identification.

Table 2 Variation of peak intensities of lithium complexes in water with excess Li+ at CV 20 V Molar ratio Li:LiL1Cl

(LiL)+ (%BPI)

(LiL2)+ (%BPI)

(LiL3)+ (%BPI)

(LiL4)+ (%BPI)

1 2 4 6 8 10

92.0 100 100 100 100 100

100 88.0 73.4 67.4 55.6 34.5

18.8 21.0 12.6 11.1 4.7 0.74

9.8 6.5 5.1 4.2 1.1 0.5

the 1:3 complexes (Table 4). Fig. 4(a) shows the spectrum of the mixture of equimolar ratios of LiL1Cl with Na+, K+, Rb+ and Cs+, whereas Fig. 4(b) shows the spectrum of an equimolar mixture of free alkali ions. The total %BPI for all (ML1n )+ complexes are 172 (Li), 162 (Na), 35.3 (K), 27.2 (Rb) and 20.2 (Cs), which equates with peak ratios of 1:0.94:0.21:0.16:0.12. It is suggested that the ligand prefers lithium ion (ionic radius 0.76 A, ) for complexation rather than others cations, which have larger ionic radii. This behaviour is consistent with a similar preference for lithium complexes in the mixture of alkali ions with the simple model ligand (L2), where %BPI of the lithium complex is the largest relative to the others (see Fig. 5). To survey the selective cation binding, relative peak areas were obtained for each metal complex by dividing the peak area of the ESMS of 1:1 metal complexes by those of the corresponding metal ions alone. The need to normalise against the free metal ions is of paramount

importance due to the ESI process yielding different conversion efficiencies for the alkali metal ions (i.e. conversion of bulk ions into the gas-phase are different for each alkali ion). This is clearly evident in Fig. 4(b),

Table 3 Negative-ion ESMS of LiL1Cl in water at CV −30 V Major ions

m/z (Theoretical)

m/z (Experimental), (%BPI)

(L1H)− (L1Cl)− (LiL1Cl-H)− (LiL1Cl2(H2O))− (L12 H)− (L12 Cl)− (L13 Cl)− (L14 Cl)−

215 251 260 313 432 467 683 899

214.8, 251.1, 260.9, 313.0, 432.2, 467.3, 683.3, 899.0,

28 50 100 38 12 42 10 1

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

164

Table 4 Positive-ion ESMS for equimolar amounts of alkali cations Na+/K+/ Rb+/Cs+ with LiL1Cl or L2 in water at CV 20 V Ligand

Major ions observed

m/z (Theoretical)

m/z (Experimental), (%BPI)

L1

(LiL1)+ (NaL1)+ (KL1)+ (RbL1)+ (CsL1)+ (LiL12 )+ (NaL12 )+ (KL12 )+ (RbL12 )+ (CsL12 )+ (LiL13 )+ (NaL13 )+ (KL13 )+ (RbL13 )+ (CsL13 )+ (L2+H)+ (LiL2)+ (NaL2)+ (KL2)+ (RbL2)+ (CsL2)+ (LiL22 )+

223 239 255 301 349 439 455 471 517 565 656 672 687 733 780 91 97 113 128 176 222 187

223.1, 69 239.2, 100 255.0, 27 301.5, 25 450.2, 19 439.2, 7 456.3, 4 471.2, 8 517.5, 2 565.4, 1 656.3, 16 672.2, 8 687.3, 0.3 733.5, 0.2 780.5, 0.2 91.4, 100 97.3, 78 113.1, 30 129.2, 3 175.5, 1 223.0, 0.5 187.2, 22

L2

Table 5 reports the relative peak areas of complexes involving the alkali metal ions. Here, any straightforward quantification of the ESMS data in the gas-phase presents some difficulty even for the complexes since it is uncertain whether these peak area values reflect exactly the solution phenomena under study [19]. Possible misinterpretation of the ESMS data for the alkali metal ion binding of oxaamide ionospheres [20] and for the cyclodextrin inclusion of amino acids and small peptides have been noted [21]. In contrast, several workers have discovered some correlation of solution and gas-phase complexation of 18-crown-6 [11] and cryptand[2.2.2] [22] with alkali metal cations when comparing cation-binding selectivities by ESMS and by well established methods [23]. The selective cation binding caused by the preferential ion–dipole interaction can more explicitly be visualised graphically. This is done via a plot of the relative peak areas of metal ion complexes divided by peak areas of free metal ions (data of Table 5) versus metal ionic radius after normalising the major peak in each set to one (Fig. 5). The preferential formation of (LiLn )+ could be explained by its high surface charge density being ‘felt’ by the ligand. The ligand may also be too small to envelop the alkali metal cations K+, Rb+ and Cs+ as effectively. Notably, preference for ML2 species falls from Li+ to Na+ to K+ to Rb+, but rises again for the largest Cs+. A consistent fall is expected on the basis of electrostatic effects alone. The rise for Cs+ suggests that this large ion may be stabilised somewhat in a ‘sandwich’ eight-coordinate structure between two L1 molecules.

3.3. Cation exchange in [Li(L 1)]+ with alkali earth, Ga(III) and Ce(IV) ions Addition of a molar equivalent of the higher charged alkaline earth ions Mg2 + , Ca2 + , Sr2 + and Ba2 + to a

Fig. 4. Positive-ion ESMS of (a) an equimolar mixture of L1 and Li+, Na+, K+, Rb+, Cs+ and (b) an equimolar mixture of alkali cations alone in water at neutral pH (cone voltage 20 V).

where equimolar mixtures of the alkali cations yield vastly different ion abundances.

Fig. 5. Plot of relative peak intensities of alkali ion complexes ML+, ML2+ and ML3+ vs. ionic radii of metal (Li+, Na+, K+, Rb+, Cs+ from left to right in the sets above).

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

165

Table 5 Relative amounts of alkali metal complexes observed in the ESMS of a solution containing equimolar concentrations of all M(I) ions and L1 M

Ionic radii (A, )

Total (ML1n)+ (normalised to Li =100)

(ML)+

(ML2)+

(ML3)+

Li+ Na+ K+ Rb+ Cs+

0.76 1.02 1.38 1.52 1.67

100 77.3 2.67 0.40 0.17

23.84 28.19 1.15 0.18 0.08

29.51 14.9 0.32 0.035 0.011

4.65 1.76 0.079 0.012 0.007

Calculated by dividing ESMS peak areas of alkali complexes by peak areas of the free alkali ions.

solution of LiL1Cl displays exchange and M2 + complexation in the positive-ion ESMS. Fig. 6 gives selected positive-ion ESMS of LiL1Cl interacting with Ca2 + and Sr2 + in water at CV of 20 V. Major species observed with calcium are (CaL)2 + , (CaL2)2 + , (CaL3)2 + and (CaL4)2 + while minor higher ratio and solvated hydroxo species such as (CaLn (OH)(H2O)n )+ also appear. In addition, the free hydrated lithium ion is observed (necessarily released upon metal exchange) as well as only small amounts of lithium complexes (LiLn )+. Similarly, these metal ion exchange trends are observed with others alkali earth cations, such as Sr2 + (Fig. 6). In general, the presence of alkali earth cations (Mg2 + , Ca2 + , Sr2 + , Ba2 + ) in the LiL1Cl solution at a 1:1 molar ratio produces species from [M(L1)]2 + to [M(L1)8]2 + (where M= alkali earth cations). This behaviour is consistent with expectations for speciation in the bulk solution, where the higher charged alkali earth ions are expected to compete more effectively than lithium ions, and infers that the ESMS provides, in this instance, information that reflects in a qualitative sense solution behaviour. Fig. 7 plots %BPI against alkali earth complexes, which shows that (ML3)2 + is the most stable metal complex (except for magnesium, where it is ML22 + ). The variation from Mg2 + to Ba2 + (which identifies the 1:2 and 1:3 species as dominant, respectively) is consistent with the ability of the larger ion to ‘pack’ more ligands around it and achieve a higher coordination number [24]. Of course, the ESMS method may be detecting, for example, (LiL2)nn + assemblies rather than single molecular units, since successive fragmentation of drying droplets need not yield purely species with n =1 [25]. However, we find no evidence for mixed-metal species such as (LiCaL2)3 + in the mass spectrum (which should be expected if large clusters are present and prevalent). All the speciations are tabulated in Table 6. The ratio of total %BPI of all species of alkali earth complexes in the mixture from magnesium to barium is 1:0.221:0.205:0.189. This behaviour of the divalent metal ion (M2 + ) is consistent with dominant electrostatic influences on the formation constants, where the trend Ba2 + B Sr2 + BCa2 + BMg2 + is expected.

Fig. 6. Positive-ion ESMS of an equimolar mixture of LiL1Cl with (a) Ca2 + and (b) Sr2 + in water at neutral pH (cone voltage 20 V).

Fig. 7. Plot of %BPI for various MLn species of alkali earth complexes (M =alkali earth cation; X =ligand).

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

166

Table 6 ESMS for equimolar concentration of LiL1Cl with individual alkali earth cations in water at CV 20 V Added M2+ 2+

Mg

Major ion observed 1 2+

(MgL ) (MgL1(H2O))2+ (MgL1(H2O)2)2+ (MgL1(H2O)3)2+ (L1H(H2O)3)+ (L1H(H2O)2)+ (L1H(H2O))+

(LiL1)+

Ca2+

Sr2+

Ba2+

(MgL12 )2+ (Mg(OH)L1(H2O))+ (MgL13 )2+ (LiL12 )+ (MgL14 )2+ (Mg(OH)L12 (H2O))+ (MgL15 )2+ (MgL16 )2+ (Mg(OH)L13 (H2O))+ (MgL17 )2+ (MgL18 )2+ (CaL1)2+ (CaL1(H2O))2+ (CaL1(H2O)2)2+ (CaL1(H2O)3)2+ (L1H(H2O)2)+ (LiL1)+ (CaL12 )2+ (CaL12 (H2O))2+ (Ca(OH)L1(H2O))+ (CaL13 )2+ (LiL12 )+ (CaL14 )2+ (Ca(OH)L12 (H2O))2+ (CaL15 )2+ (CaL16 )2+ (Ca(OH)L13 (H2O))+ (CaL17 )2+ (CaL18 )2+ (SrL1)2+ (SrL1(H2O))2+ (SrL1(H2O)2)2+ (SrL1(H2O)3)2+ (L1H(H2O)3)+ (L1H(H2O)2)+ (L1H(H2O))+ (LiL1)+ (SrL12 )2+ (Sr(OH)L1(H2O))+ (SrL13 )2+ (LiL12 )+ (SrL14 )2+ (Sr(OH)L12 (H2O))+ (SrL15 )2+ (SrL16 )2+ (Sr(OH)L13 (H2O))+ (SrL17 )2+ (SrL18 )2+ (BaL1)2+ (BaL1(H2O))2+ (BaL1(H2O)2)2+ (BaL1(H2O)3)2+ (L1H(H2O)2)+ (LiL1)+ (BaL12 )2+ (BaL12 (H2O))2+ (Ba(OH)L1(H2O))+ (BaL13 )2+ (LiL12 )+ (BaL14 )2+ (Ba(OH)L12 (H2O))+ (BaL15 )2+ (BaL16 )2+ (Ba(OH)L13 (H2O))+ (BaL17 )2+ (BaL18 )2+

m/z (Theoretical)

m/z (Experimental), %BPI

120 129 138 147 164 181 198 223 228 275 386 439 444 491 552 660 707 768 876 128 137 146 155 181 223 236 245 291 344 440 452 507 560 668 723 776 884 152 161 170 180 164 182 198 223 259 338 368 440 476 555 584 692 772 800 908 177 186 195 204 181 223 285 294 388 393 440 501 406 609 717 622 825 833

119, 20 128.9, 20 137.9, 6 147.2, 100 164.1,38 180.8, 25 198.2, 5 223.0, 19 226.8, 52 275.7, 13 386.4, 30 438.0, 3 439.4, 10 491.0, 20 552.2, 30 660.0, 3 706.7, 60 768.1, 6 876.0, 0.4 128.0, 19 137.1, 25 145.9, 18 155.2, 3 181.0, 18 223.0, 54 236.0, 52 245.1, 3 291.2, 5 343.2, 100 439.2, 39 452.1, 22 507.1, 11 561.5, 12 668.0, 2 723.1, 0.8 775.1, 0.5 884.2, 0.2 152.2, 28 161.2, 30 170.0, 20 179.1, 12 164.0, 5 181.2, 25 198.0, 8 223.2, 62 260.1, 58 338.0, 4 368.0, 100 440.3, 30 476.2, 53 555.2, 5 584.2, 23 692.2, 18 772.1,7 800.2, 2 908.3, 0.4 176.2, 9 186.1, 12 195.1, 11 204.1, 5 181.1, 18 223.2, 40 285.0, 40 294.2, 10 388.1, 2 393.2, 68 440.2, 18 501.2, 33 406.2, 2 607.2, 11 717.2, 5 622, 0.2 825.2, 3 833.2, 0.2

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

167

Table 7 ESMS for equimolar concentration of LiL1Cl with Ga(III) and Ce(IV) cations in water at CV 20 V Added M2+

Major ion observed

m/z (Theoretical)

m/z (Experimental), (%BPI)

Ga3+

(Ga(OH)L1(H2O))2+ (Ga(OH)L1(H2O)3)2+ (Ga(OH)L1(H2O)5)2+ (Ga(OH)L1(H2O)7)2+ (LiL1)+ (Ga(OH)L12 (H2O)5)2+ (Ga(OH)2L1)+ (Ga(OH)2L1(H2O))+ (Ga(OH)2L1(H2O)2)+ (CeL1(H2O)2)4+ (CeL1(H2O)4)4+ (Ce(OH)L1)3+ (Ce(OH)L1(H2O)2)3+ (Ce(OH)L13 )3+ (Ce(OH)L13 (H2O))3+ (Ce(OH)L13 (H2O)2)3+ (Ce(OH)2L12 )2+ (Ce(OH)2L12 (H2O))2+ (Ce(OH)2L1)2+ (Ce(OH)3L1(H2O))2+

160 178 196 214 223 304 320 338 356 98 108 124 136 268 274 280 303 312 407 425

160.5, 179.0, 196.1, 214.2, 223.2, 304.6, 320.3, 338.4, 356.2, 98.2, 108.2, 124.4, 136.3, 268.4, 274.5, 280.3, 304.3, 312.3, 407.2, 425.3,

Ce4+

100 88 49 48 15 9 40 9 20 61 87 100 68 40 28 38 16 24 28 13

were observed, namely, (CeL(H2O)n )4 + , (Ce(OH)L(H2O)n )3 + , (Ce(OH)L3(H2O)n )3 + , (Ce(OH)2L22+ (H2O)n ) and (Ce(OH)3L(H2O)n )+ (Fig. 8(b)). Both gallium and cerium complex species are assigned in Table 7. Both a tendency of the Ga3 + and Ce4 + ions to hydrolyse readily, and a general trend in ESMS towards favouring lower-charged species support the observation of hydroxo complexes. With the higher charged ions, exchange equilibria lie strongly towards the multiply-charged ions and little if any Li+ complexes are observed.

4. Conclusion

Figure 8. Numerical solutions of the extended Boussinesq equations model for the propagation of a deep water wave (kh= 3.14).

Positive-ion mode ESMS of an equimolar mixture of LiL1Cl with Ga3 + in aqueous solution was also investigated. Two series of charge reduction gallium complexes, (Ga(OH)L(H2O)n )2 + and (Ga(OH)2L(H2O)n )+ were present in the spectra. Only a small residual (LiL)+ peak was observed (Fig. 8(a)). For Ce4 + , under comparable conditions, a series of cerium complexes

The polyalcohol (L1) is capable of forming complexes with mainly alkali and alkaline earth metal ions. Rapid metal ion exchange is implied from ESMS experiments, consistent with expectations for these metal ions. Preferences for higher charged and smaller ions support a dominant electrostatic character to the bonding. However, observations such as higher than anticipated amounts of (SrL12 )2 + in the ESMS infers some shapebased effect on complex stability. Given the tetrapodal shape of the ligand, this is not unreasonable. Formation of species as high in ligand number as (ML18 )n + infers that chelation does not operate in all species, although the dominant species have fewer L1 molecules around and are more likely to involve multidentate chelators. Some further structural studies in the solidstate could infer more fully some of the species that could also exist in solution. Elucidation of stereochemistries is not obtained from ESMS, although it is

168

G.A. Lawrance et al. / Inorganica Chimica Acta 328 (2002) 159–168

likely that coordination numbers can vary appreciably and range up to at least 12, as found in the solid-state and in solution under equilibrium conditions used for metal ions in this study. The value of ESMS to probe labile metal– ligand interactions evident in earlier studies has been further defined in this study, the first of structurally defined synthetic polyalcohols.

Acknowledgements Support of this research by the Australian Research Council is gratefully acknowledged.

References [1] [2] [3] [4]

S. Yono, Coord. Chem. Rev. 92 (1988) 113. D. Klaassan, P. Klufers, Inorg. Chem. 619 (1993) 661. L.G. Hubert-Pfalzgraf, New J. Chem. 11 (1997) 663. A. Appelt, A.C. Willis, S.B. Wild, J. Chem. Soc., Chem. Commun. (1986) 1001. [5] T. Tanase, F. Shimizu, S. Yono, S. Yoshikawa, J. Chem. Soc., Chem. Commun. (1986) 1009. [6] P. Goldschmidt, E. Weinggardt, W. Bachmann, Koloid Z. 47 (1929) 49. [7] J. Burger, C. Gack, P. Klufers, Angew. Chem. 107 (1995) 2950.

[8] K. Hegetschweiler, L. Prino, H. Koppenol, V. Gramlich, W. Meyer, Angew. Chem. 107 (1995) 2421. [9] K. Hegetschweiler, Chem. Soc. Rev. 28 (1999) 239. [10] P.V. Bernhardt, G.A. Lawrance, S.M. Luther, M. Maeder, M.J. Robertson, Sutrisno, Inorg. Chem. Commun. 3 (2000) 410. [11] K. Hegetschweiler, Bol. Soc. Chilena Quim. 42 (1997) 257. [12] E. Leize, J. Jaffrezic, A. van Dorsselaer, J. Mass Spectrom. 31 (1996) 537. [13] D.K. Walanda, R.C. Burns, G.A. Lawrance, E.I. von Nagy-Felsobuki, J. Chem. Soc., Dalton Trans. (1999) 311. [14] Sutrisno, Y. Baran, G.A. Lawrance, E.I. von Nagy-Felsobuki, D.C. Richens, H. Xiao, Inorg. Chem. Commun. 2 (1999) 107. [15] D.N. Butler, T.J. Munshaw, Can. J. Chem. 59 (1981) 3365. [16] K. Wang, X. Han, R.W. Gross, G.W. Gokel, J. Am. Chem. Soc. 117 (1995) 7680. [17] R.A. Johnstone, M.E. Rose, J. Chem. Soc., Chem. Commun. (1983) 1268. [18] A.E. Lewis, M.C. Grossel, Tetrahedron 39 (1993) 1597. [19] G.J. Angley, D.G. Hamilton, M.C. Grossel, J. Chem. Soc., Perkin Trans. 2 (1995) 929. [20] M. Goodall, P.M. Kelly, D. Parker, K. Gloe, H. Stephan, J. Chem. Soc., Perkin Trans. 59 (1997) 442. [21] J.B. Cunniff, P. Vouras, J. Am. Soc. Mass Spectrom. 6 (1996) 4693. [22] K. Wang, G.W. Gokel, J. Org. Chem. 61 (1996) 4693. [23] R.M. Izatt, J.S. Bradshaw, S.A. Nielsen, J.D. Lamb, Chem. Rev. 85 (1995) 271. [24] F.A. Cotton, G. Wilkinson, A.C. Murillo, M. Bochmann, Advanced Inorganic Chemistry, sixth ed., Wiley, New York, 1999. [25] G. Horlick, G.R. Agnes, Appl. Spectrosc. 46 (1992) 401.