E L S E V I E R Inorganica Chimica Acta 240 (1995) 159-168

1,4,8,11 -Tetrakis { ( 2,2'-bipyridyl-5 '-ylmethyl) -bis ( 2,2'- bipyridyl) ruthenium (II) } - 1,4,8,11-tetraazacyclotetradecane ( L 1 ), a

macrocyclic pH and transition metal ion fluorescence sepsor. Equilibrium, and stopped-flow kinetic, fluorimetric studies of the reactions of L 1 with

nickel(II) and copper(II) ions in aqueous solution at neutral pH"

Ana M. Josceanu, Peter Moore *, Simon C. Rawle, Philippa Sheldon, Stephen M. Smith Department of Chemistry, University of Warwick, Coventry, CV4 7AL, UK

Received 7 January 1995; revised 5 April 1995

Abstract

A new pendent-arm derivative of 1,4,8,11-tetraazacyclotetradecane (cyclam), 1,4,8,11-tetrakis(2,2'-bipyridyl-5'-ylmethyl)-l,4,8,11- tetraazacyclotetradecane (1) (L) has been synthesised and characteri sed. L reacts with four mole equiv, of cis-[ Ru (bipy)2C12 ] (bipy = 2,2'- bipyridine) to give the highly fluorescent macrocyclic ligand 2 (L I) (Ll=[{(bipy)2Ru}4L]S+), which has been isolated as the monoprotonated derivative, [ { (bipy)2Ru}4(LH) ] [CIO419. Molecular mechanics and dynamics calculations indicate that the cyclam group of 2 adopts the R,S,S,R (trans-IV) rather than the more common R,R,S,S (trans-III) set of N-configurations. This modified cyclam has four covalently attached [Ru(bipy)3] 2+ units, and is highly fluorescent at pH 7 (Aex =450 nm, Aem=600 nm). The fluorescence of 2 is much reduced upon protonation, or as metal ions enter the macrocyclic cavity. The compound is a pH and transition metal ion sensor, with a first photo excited state pKa of 5.92 + 0.11 (from fluorescence measurements) and a ground state pKa of 7.84 + 0.04 (from absorbance measure- ments). The results of kinetic and thermodynamic studies are reported for the fluorescence quenching which occurs when aqueous copper(II) or nickel(II) ions coordinate at the cyclam cavity of L t.

Keywords: Kinetics and mechanism; Fluorimetric studies; Nickel complexes; Copper complexes; Macrocyclic ligand complexes

1. Introduction

Studies of pendent arm macrocycles and their metal com- plexes continue to attract significant attention. Such ligands combine the advantages of a macrocyclic framework with the added flexibility of the attached pendent arm(s). Poly- azamacrocyclic ligands carrying up to six N-pendent 2,2'- bipyridyl (bipy) arms have been reported [ 1-10], including two derivatives of 1,4,8,11-tetraazacyclotetradecane (cyclam), each with a single 2,2'-bipyridyl (bipy) arm linked via a methylene group to one of the macrocyclic N- atoms [6-8] . In 3 (L 2) the methylene link is at the 6-position of bipy, and this leads to the formation of mononuclear six-coordinate metal complexes in which the N-atoms of the

* This paper is dedicated to Professor F. Basolo. * Corresponding author.

0020-1693/95/$09.50 © 1995 Elsevier Science S.A. All fights reserved SSD10020-1693(95)0453 I -D

cyclam and bipy groups are bound to a single metal ion [6,7]. In 4 (L3) , the methylene link is positioned at the 5-position of the bipy, and this leads to the formation of polynuclear metal complexes since it is stereochemically impossible for the bipy and cyclam groups to encompass a single metal ion. For example, reaction of L 3 with an equimolar amount of Cu 2÷ gives [CuL3] 2÷ in which the bipy is uncoordinated, and the Cu 2÷ resides in the macrocyclic cavity. In contrast, reaction of 3 equiv, of L 3 with 1 equiv, of Fe 2+ gives [Fe(L3)3] 2+, which has an [Fe(bipy)3] 2+ core and three peripheral and uncomplexed cyclam groups. [ C u E 3 ] 2 + and [Fe(L3)3] 2+ may be reacted further with divalent metal ions ( M 2+ ) to give the dinuclear [ C u ( L 3 M ) ]4+ and tetranuclear [ Fe (LaM) 3 ] s + complexes respectively. A pH and transition metal ion fluorescence sensor 5 (L 4) was also obtained from the reaction of L 3 with an equimolar amount of cis- [Ru(bipy)2C12] [ 8]. L 4 was shown to be highly fluorescent at high pH (A,x=450 nm, A~m= 600 nm), and to sense the presence of cations in the cyclam cavity by showing much

160 A. M. Josceanu et aL /Inorganica Chimica Acta 240 (1995) 159-168

N L N N ~ N

1 (L)

H, ~ /H

3 (L 21 4 (L 3)

N N N N

N N__

N N

%"i"-.

2 (L 1)

8+

reduced fluorescence either by protonation, or by complex- ation with transition metal ions such as Cu 2 ÷ and Ni 2 + [ 8]. An analogous pH sensor 6 (R1 = Me, R2 = Ph), in which the fluorescence decreases at low pH has also been reported, although the opposite type of behaviour is found for L 5 when R 1 = Me, and R 2 is a phenyl group carrying electron releasing methoxy substituents [ 11 ]. For the latter systems, the fluo- rescence increases upon protonation of the two NR1R 2 groups, and this is postulated to be the result of R u N bond fission in the photo excited state [ 11 ]. A [ Ru (bipy) 3]: ÷ appended derivative of calix[4]arene has recently been reported to show increased fluorescence as the pH is reduced and the calixarene phenolate groups become protonated [ 12 ].

In this study a cyclam derivative carrying four N-pendent bipy groups 1 (L) has been isolated and characterised. 1 has been used to obtain the new highly fluorescent sensor [{(bipy)zRu}4L] 8÷ (2) (LI), which has been isolated as the perchlorate salt of the monoprotonated macrocycle, [ { (bipy) 2Ru } 4 ( L ] H ) ] [ C10 4 ] 9" The conformation of L1 has been investigated by molecular mechanics and molecular dynamics calculations, and the four pKa values of L ~ in the ground- and photoexcited-states have been determined by pH titration using absorption spectrophotometry and fluorimetry respectively. Finally, stability constants and rate data for the macrocyclic complexation of Cu 2+ and Ni 2+ ions by L ~ have been determined by fluorimetric titration and stopped-flow fluorimetry.

R

~ - N N - -

H

5 (L 4)

2+

~ N ~ N ~ | N ~ ~ N R 1 R 2

6 (L 5)

2. Experimental

2.1. Materials and methods

1,4,8,11-tetraazacyclotetradecane (cyclam) was prepared by the published method [ 13]. All other chemicals were the best commercially available, and were used without further purification. Standard solutions of the metal ions were ana- lysed either by [edtaH2] 2- titration, or by using a Leeman Labs PS ICP/Echelle spectrometer. C, H, N analyses were carried out with a Leeman Labs CE440 instrument. Fluori- metric measurements were made using aerated solutions in quartz cells with a Perkin Elmer LS-3 or an Applied Photo- physics DX. 17MV stopped-flow spectrofluorimeter, using an excitation wavelength of 450 nm, and a detection (emission) wavelength of 600 nm. Absorption spectra were recorded with a Philips-Unicam PU 8700 spectrophotometer coupled to a 486 PC for data analysis. ~H and 13C NMR spectra were recorded either at 250 and 62.89 MHz respectively using a Bruker ACF 250 spectrometer, or at 400 and 100.62 MHz respectively using a Bruker WH-400 instrument. Mass spec- tra were obtained with a Kratos MS80 spectrometer.

2.2. Syntheses

2.2.1. Synthesis o f l,4,8,11-tetrakis(2,2'-bipyridyl-5'- ylmethy l )- l , 4,8,11-tetraazacyclotetrade cane (1) ( L )

5-(Bromomethyl)-2,2'-bipyridine [8] (1.01 g, 4.1 mmol), 1,4,8,11-tetraazacyclotetradecane (0.20 g, 1.0 mmol) and triethylamine (0.50 g, 5 mmol) were stirred at 100 °C in chlorobenzene (200 cm 3) for 1 h. The solution was then filtered and the solvent evaporated. The residue was recrystallised from the minimum quantity of acetonitrile and the product collected as a white crystalline material. Yield: 570 mg, 65%. IH NMR (250 MHz, CDC13): 6 8.64 (d, 4H), 8.56 (s, 4H), 8.33 (d, 4H), 8.27 (d, 4H), 7.77 (m, 8H), 7.26 (m, 4H), 3.46 (s, 8H), 2.63 (s, 8H), 2.52 (t, 8H), 1.79 m, 4H) ppm. ]3C NMR (62.89 MHz, CDCI3): 6 156.1 (4C), 154.9 (4C), 149.6 (4C), 149.0 (4C), 137.4 (4C), 136.7

A. M. Josceanu et al. I lnorganica Chimica Acta 240 (1995) 159-168 161

\ \ ' \ /-,"/Y / / / 'V /

s

Fig. 1. Energy minimised structure of L 1 as determined by molecular mechanics and molecular dyna~cs calculations using Hyperchem Version IV.

(4C), 135.4 (4C), 123.4 (4C), 120.9 (4C), 120.5 (4C), 56.2 (4C), 51.3 (4C), 50.5 (4C), 24.3 (2C) ppm. Mass spectrum (CI/NH3): m/z 873 (calc. for LH ÷, 873).

2.2.2. Synthesis of [{ Ru(bipy)2 }4(LH)I(CI04)9 (2. HCI04) 1 ( 100 mg, 0.11 mmol) and cis-[ (bipy)2RuCl2] • 2H20

(235 mg, 0.45 mmol) were refluxed in H20 (100 cm 3) for 14 h. The solution was then filtered (to remove any unreacted cis-[ (bipy)2RuC12] ) and the volume reduced to approxi- mately 25 cm 3 with a rotary evaporator. Dropwise addition of excess saturated aqueous NaC104 gave an orange precip- itate which was collected by suction filtration and dried in vacuo over silica gel. The product was obtained as an orange powder. Yield: 260 mg, 69%. ~H NMR (400 MHz, CDaNO2): 8 8.43 (m, 24H), 7.95 (m, 20H), 7.85-7.71 (m, 24H), 7.64 (s, 4H), 7.29 (m, 20H), 5.29 (s, 1H), 3.25 (s, br, 8H), 2.15 (s, br, 8H), 2.05 (s, br, 8H), 1.21 (s, br, 4H) ppm. 13C NMR (100.62 MHz, (CD3)2SO): 6 158.6 (8), 158.5 (8), 157.1 (4), 153.1 (4), 153.0 (4), 152.9 (8), 152.8 (4), 152.4 (4), 142.3 (4), 139.1 (20), 139.0 (4), 128.8 (16), 128.6 (4), 125.5 (20), 125.3 (4), 125.1 (4), 57.0 (4), 52.2 (4), 51.8(4), 24.0 (2) ppm. UV-Vis spectrum in H20 (A~x/nm (~/dm 3 mol - l cm-1)) : 453 (46000), 285 (283 000), 243 ( 105 000). Anal. Found: C, 47.12; H, 3.96; N, 11.41. Calc. for CI34H121C19N28036Ru4: C, 47.02; H, 3.56; N, 11.46%.

2.2.3. Synthesis of [ { Ru(bipy)z } ~(LNi)]( ClO~)w ([NiLl ](ClO4) w)

2- HCIO4 (50 mg, 0.015 mmol) and Ni (CIO4) 2" 6H20 (27 mg, 0.075 mmol, a five-fold equivalence) were refluxed in CH3OH/H20 (1:1, 50 cm 3) for 1 h. Dropwise addition of

excess sat. aq. NaCIO4 resulted in the precipitation of [ { Ru (bipy) 2 } 4 L3Ni ] (CIO4) x o as an orange crystalline solid. Yield: 32 mg, 62%. Anal. Found: C, 44.4; H, 3.54; N, 10.5. Calc. for Ci34H12oClloN2sNiO4oRu4.2H20: C, 44.51; H, 3.46; N, 10.85%.

The Cu 2÷ complex was made in an analogous way. Yield: 90%. Anal. Found: C, 44.4; H 3.28; N, 10.4. Calc. for CI34H12oCIloCuN2sO4oRu4"2H20: C, 44.4; H, 3.45; N 10.85%.

2.3. Molecular modelling

Molecular mechanics and molecular dynamics calcula- tions were undertaken with Hyperchem Version 4 including the ChemPlus modification. Molecular mechanics was car- fled out using MM + with the Polak-Ribiere algorithm of Hyperchem. Molecular dynamics was also used (simulated heating to 3 000 K) to ensure that the true energy minimum had been reached. A conformational search (random walk) about the macrocyclic ring was also undertaken to find all five isomers expected for a cyclam ring. Computing was carried out with a PC containing a Pentium P90 processor. Initial bond lengths and force constants were taken from literature values [ 14,15] 1. The lowest energy minimised structure is shown in Fig. 1. The relative energies of the five possible [ 16] cyclam N-conformations of L ~ were found to be R,S,S,R (trans-IV, 104.0), R,R,S,S (trans-III, 109.5), R,S,S,S (110.1), R,R,R,R (trans-I, 115.7) and R,S,R,S ( 121.5). For 1,4,8,11-tetraazacyclotetradecane (cyclam) the

The C o N force constant was used.

162 A. M. Josceanu et al. / Ino rganica Chimica Acta 240 (1995) 159-168

relative energies were found to be R,R,S,S (8.64), R,S,R,S (11.47), R,R,R,R (11.50), R,S,S,R (11.78) and R,S,S,S (12.92).

calculated final fluorescence (end-point) for each stopped- flow trace was used to evaluate the stability constants.

2.4. pKa determinations 3. Results and discussion

pKa values for 2 (L 1) in both the ground- and photo excited-states were determined using absorbance and fluo- rescence measurements respectively. Increasing amounts of a solution of L ~ ( 2 × 10 -6 mol dm -3) in 0.2 mol dm -3 NaOH were added to a solution o f L ~ (2 × 10 -6 mol dm -3) in 0.2 mol dm -3 HCI or HBF4, and the absorption and emis- sion spectra of each solution was recorded over the pH range 0.8 to 11.0. pH values were measured with a combined glass/ calomel electrode, using a Radiometer PM64 pH meter, whose electrode was calibrated before each study using two standard buffer solutions of pH 4.00 + 0.02 and 9.20___ 0.02. The temperature was maintained at 25.0 + 0.1 °C in all exper- iments.

Absorption measurements were recorded with a Philips Analytical PU 8700 instrument in the range 225-535 nm, using a 4 cm pathlength cell. A long pathlength cell was used since the absorbance differences between the totally proton- ated and the fully deprotonated ligand are rather small. Exper- imental data were acquired in transmission mode, then transferred to a 486 PC for analysis with a non-linear least- squares routine which was programmed to simultaneously analyse several data sets taken at different pH values, and at 4 wavelengths.

Fluorescence measurements were recorded with a Perkin Elmer LS-5 spectrometer. The excitation wavelength used was 450 n m ( 10 nm bandpass), and emission was recorded at 600 nm.

All reagents were analytical grade, the stock acid and base solutions being standardised before use by titration using phenolphthalein indicator. Deionised distilled water was used for all solutions.

2.5. Stability constants and rate measurements

The stabilities of complexes formed by L 1 with C u 2+ and Ni 2 + in an aqueous solution buffered at pH 6.4 were esti- mated fluorimetrically using an Applied Photophysics kinetic workstation. A solution of ligand ( 2 × 10- 6 tool dm - 3) was mixed 1:1 in the stopped-flow with a set of solutions contain- ing increasing concentrations of each metal ion. In the case of Cu 2+, metal concentrations were varied from 6 × 10 -7 mol dm -3 to 8 × 10 -4 mol dm -3 (after mixing), whereas for Ni 2+, because the stability constant of the complex is smaller than that of the Cu 2 + complex, the range used was 1×10 -6 to 5 × 1 0 -3 tool dm -3. 2,6-Lutidine buffer (5 × 10 -3 mol dm -3 to which HCIO4 was added to adjust the pH) was added to both the metal and the ligand solutions to control pH.

Kinetic traces were recorded and analysed as either pseudo first-order or second-order processes as necessary, and the

L l (2) contains four [Ru(bipy)3] 2+ groups, all of which can exist as either A or A isomers. In addition there are 5 possible isomers arising from the various N-configurations of the cyclam moiety. The NMR data indicate the presence of a single symmetric N-configurational isomer, and the molecular modelling suggests this is most likely to be the one with an R,S,S,R (trans-IV) arrangement (Fig. 1 ). The phys- ical data refer to the racemic mix of all possible isomers arising from the four [ Ru (bipy) 3 ] 2 + groups.

3.1. Acidity constants

The reactivity of azamacrocycles with metal ions in aque- ous solutions is known to be very pH dependent due to sequential protonation of the nitrogen atoms which markedly lowers the rate of the reactions [ 17-21].

For cyclam the following equilibria apply:

[L] [H + ] L + H + ~ H L + Ka, = [HL+ ] (1)

HL + + H + ~ H z L2+ Ka2 = [HL+] [H+] (2) [ H2L 2 + ]

[H2L 2+ ] [H + ] H2L z + + H + ~ H3L 3 + K~ o = [ H3L3 + ] (3)

[H3L 3+ ] [H + ] Ha L3+ + H + # H 4 L 4+ K~a = [H4L4+ ] (4)

If we represent L, and [L] as the total and unprotonated ligand concentrations respectively, then:

L,= [L] + [HL +] + [H2L 2+]

+ [H3L 3+ ] + [H4L 4+ ] (5)

Combining Eqs. ( 1 ) - (5 ) gives Eq. (6):

L, = 1 + 10 pral -pH + 10PK, i +PK,2- 2pH [L]

+ 10PKal +pKa2 +pXa3 -3pH _~_ 10pgal +pXa2+pra3+pKa4-4pH

(6)

The observed property (e.g. fluorescence or absorbance), Pob,, is given by Eq. (7).

Pobs =PL[L] +PnL[HL + ] +Png[H2L 2+ ]

+PH3L[H3L 3+ ] +PH4L[H4 L4+ ] (7)

PL, PHL, PHi, Ptc3L, PH, L are the intrinsic molar properties (fluorescence or absorbance) of the respective species. Com- bining Eqs. ( 1 ) - ( 7 ) gives Eq. (8):

A. M. Josceanu et al. / Inorganica Chimica Acta 240 (1995) 159-168 163

//\\/,

I ! !

wavelength/nm Fig. 2. Comparison of the fluorescence spectra of (a) [Ru(bipy)3] ~÷ (SX 10 .6 tool din-3), (b) [Lt] ~÷ at high pH, (c) [Ni(L])] ~°÷, (d) lCu(L ~) ] ~o+ (e) H4L ~+, ((b)-(e) 2 x 10 -6 tool (Ira-~; &~ = 450 nm)).

t :

o

E.,

70.0

60.0

50.0

40.0

30.0

::'0.0

~.0.0

i i i i i : i i i

t ~ . . . . . . . . . . 4 - - ' ~ ' M - - " t ' . . _ _ _ , . . . . . . L..----'.L ....... . . . . . " ] " - " - ] ' - " ~ " " ~ t t : t : :

i i i ii ~.. i i i i

--~: . . . . . ~ - - ! - - + - ~ - - - - + ....... ! ....... t .......

. . . . . . . L . - - £ - - . " . . . . ~ ~ - .~ . . . . ; - - . - k - - - ' - - - ~ - - . : ....... ~ . . . . . i i i [ i [ ~ i

: I : ~ ~ t : t : !

E i 7 i I 1 I i i

i i [ i ! i i * [ i i i i i i i i

5, .0 3 . 0 5 . 0 7 . 0 9.0 ~. t .0

p H

F ig . 3. T i t r a t i on c u r v e f o r the c h a n g e in f l u o r e s c e n c e o f L t ( 2 X 10 - 6 m o l

d m - 3 ) w i t h p H at 25 °C a n d I = 0 . 2 m o l d m - 3 [ N a C I O 4 ] . A¢x = 4 5 0 n m .

The relatively low solubility of L 1 prevents a potentiomet- ric determination of the ground state pK~ values, and although UV-Vis spectra of the protonated and unprotonated com- plexes in aqueous solution are not markedly different, a care- fully conducted experiment using 4 cm pathlength cells provides enough information for an approximate absorption versus pH titration curve (Fig. 4) . Data from such titration curves obtained at four wavelengths were fitted simultane- ously to four equations of type (8) , where in this case values of PL, PnL, PH2L, PH3L, P n ~ are the products of the cell path- length and the molar extinction coefficients of the various species at each of the four wavelengths chosen. Data were recorded at each wavelength as percentage transmission (T), and values of Twere converted to absorbances (A) using Eq. (9). The absorbances (A) were analysed as a function o f p H with an equation analogous to Eq. (8) (A =Pobs).

A = 2 - l o g j o ( T ) ( 9 )

The calculated ground- and photo excited-state pK~ values of L 1 are compared with the ground state values of cyclam in Table 1.

It can be seen that in its photo excited state, L ~ is less basic than in its ground state at all levels of protonation. In the ground state, the four pendent [ Ru (bipy) 3 ] z + groups, which give the overall 8 + charge to unprotonated L ~, result in a lowering of the first pK~ of the cyclam core by ~ 3.7 pK a units, and for first protonation the photo excited state of L ~ is more acidic than the ground state by ~ 1.9 pK~ units. Similar differences between the ground- and photo excited states are seen for the subsequent protonations. Interestingly, pKa2- pKa3 is 9.6 for cyclam, but only 1.6 (ground state) or 1.9 (photo excited state) for L I. The large difference between the values of pKa2 and PKa3 for unfunctionalised cyclam is explained by the need to break a hydrogen bond to enable the third proton to become attached, since in the diprotonated state of cyclam the two protons on the N-atoms in the 1 and 8 positions form H-bonds to the N-atoms in the 4 and 11

P o b s

L, = (PL + PHL 10 p K a 1 - - p H 3r. P n ~ 10 pKal + pK~Z - 2pH

+... + P n ~ 10 pg~t + pK~z + PKa3 + pKa4 - - 4 p H )

X (1 + I 0 P K a l - P H + 1 0 pKal + p K a 2 - 2 p H

+ . . . + 1 0 P K a l +pKa2+pKa3+PKa4--4pH) - - 1 ( 8 )

In Fig. 2 is shown the marked lowering of the fluorescence of L t as the pH is reduced, or as metal ions enter the macro- cyclic cavity. For comparison, also shown in Fig. 2 is the fluorescence of [ Ru (bipy) 3 ] 2 + at four times the molar con- centration of L ~. The fluorescence pH titration curve for L ] is presented in Fig. 3. Data from this titration curve were fitted to Eq. (8) by non-linear least-squares analysis to give four photo excited-state pKa values, together with values of PL, PILL, PH2L, PHaL, Pnd... The r e s u l t a n t p K a v a l u e s t o g e t h e r

with their standard deviations are collected in Table 1.

s 4 . o | .,'J ~ . o ~ - - - ~ . . . . . . . . ~ - . . . . . . -~ . . . . . . . . . . ... . . . . . . . . ~ . . . . . . . . . . . ..-- . . . . . . . . . .

• I ' ' . . . . . . . . t . . . . . . . . . . . . . . . . . . . . . ! . . . . . . . . . . . . i . . . . . . . . . . . . . . . . . . . . . . .

, , . o . . . . . . . i . . . . . . . . . . ............ i . . . . . . . . . . . + ............

7 ° . o ~ - - , - . ~ - - - i - - - i . . " ' ~ - " i . . : . . . . . . . . i . . . . . . . . . t . . . . . . . . . .

. . . . . . . . , . . . . . . . . . . . . . . . . . . . .

6 6 . ° . . . . . . . . . . . . . . . . . . . . . . . . i . . . . . . . . . . i . . . . . . . . . . .

6 4 . o - - - ~ . . . . ~ - - ~ . : . ~ , . . . . . . . ?. . . . . . . . .

62.0 I J I I I ; 2.0 6.0 so.o s4.o

pH

Fig. 4. Titration curve for the change in absorbance of L t (2X 10 -¢ mol dm- 3) with pH at 396 nm, 25 °C and 1 = 0.2 mol dm- 3 using 4 cm pathlength cells.

164 A. M. Josceanu et aL I lnorganica Chimica Acta 240 (1995) 159-168

Table 1 Comparison of the pK, values of cyclam with those of L ~ determined in the ground- and photo excited-states (1 = 0.2 mol dm-3, 25 °C)

Ligand pKa, pK~ pK~, pK~, Method

Cyclam 11.5 10.2 1.6 0.9 Potentiometry a L l 5.92 + 0.11 4.78 + 0.06 2.87 5:0.18 0.50 + 0.18 Fluorimetry b L l 7.84 + 0.04 5.27 + 0.08 3.67 + 0.06 0.86 + 0.13 UV-Vis ~

aRef. [21]. b pt = 62.9 + 0.9 ' ptcL = 51.8 + 2.1, pn2L= 21. 3 +0.6, PmL = 17.8 5:0.2, PHi= 3.1 5:4.1.

At 396, 400, 404 and 408 nm respectively, 102 PL (:1:0.08)= 19.49, 19.70, 19.94, 20.52; 102 PHL (+0.16)= 14.98, 15.34, 15.87, 16.72; 102 PH2L ( "4-0.31 ) = 10.98, 11.49, 12. I0, 13.07; 102 PmL ( 5: 0.09) = 10.53, 11.00, 11.64, 12.56; 102 PH, L ( + 0.05) = 7.49, 8.08, 8.74, 9.76.

0 . 1 2 ~ : , , : ,

t: o. 1o / i ,," ~ ,:' ,," ~ i o . . . . . . . .

o . . . . . . . . . .

0 . 0 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

o . . . . . . . . . .

o . o o r r - - - - - - ~ - - - - , ~ ~ _ . . . . . . . T . . . . . . . .

- 0 . 0 2 1 - t . . . . . . . ~ . . . . + . . . . . . . .

- o . . . . . . . . . . . . . . . .

-olo:: i i i i . . . . . . i . . . . . . . . . 4 - . . . . . . . . ~ . . . . . . . . i . . . . . . . . . 4 - . . . . . . . ~. . . . . . . . . . ~ . . . . . . . . . + . . . . . .

-0.0BI i i i i i i i 0 .5 1 .5 2 .5 3 .5 4 .

105 [C'u(II)] /mol din-3

Fig. 5. Titration curve for the change in fluorescence of L ~ ( 10- 6 mol dm- 3) with [Cu 2+ ] at pH 6.4 (5 × 10 .3 mol dm -3 2,6-1utidine buffer).

0 .12 : : : : : : t ; : _

_8 o.os i i ~ o . t o .... -.: . . . . i - - - - f . . . . . i . . . . f - - -+----~ : -------r ........

o. . . . . . . . .

o o.of i . . . . . . . .

0 . . . . . . . .

. . . . . .

- o . o . ~ ~ . . . . . . . . :+,. . . . . . . .

- - i . . . . i - - - i . . . . . . b - - - f , ' - - - , i - - - t - - i . . . . . . . +,,' . . . .

-0 i ~ " - i . . . . . ~----4i . . . . g . ___ ; . . . . . 4 - - - - - - 4 - - - - - 4 . . . . . . . 4----- 0 . 4 1.2 2.0 2 . 8 3 .6

104 [Ni(II)] /mol din-3

Fig. 6. Titration curve for the change in fluorescence of L' ( 10-6 mol dm - 3) with [Ni 2÷ ] at pH 6.4 (5 × 10 -3 mol dm -3 2,6-1utidine buffer).

positions respectively [21 ]. The trend in the pKa values of L a is more typical of a tetrabasic amine in which hydrogen bonding of the type observed for cyclam is not possible. This would suggest that the lowest energy conformation of the cyclam core in L 1 is not that found in cyclam, where the favoured set of N-atom configurations is R,R,S,S (known as trans-III) [ 16]. Molecular modell ing studies were under-

taken to shed further light on the likely conformation of L ~, and the results of these calculations are shown in Fig. 1. The lowest energy conformation was found to be that with the R,S,S,R set of N-atom configurations (known as trans-IV) [ 16]. Both the trans-III and trans-IV conformations have

been found to coexist in the crystal lattice of the analogous macrocycle 1,4,8,11-tetramethyl- 1,4,8,11-tetraazacyclote- tradecane ( tmc) [22]. H-bonding of the type found in [cyclamH2] z÷ (i.e. from an NH ÷ proton spanning the 4,8 and the 1,11-positions) is not possible for the trans-IV isomer shown in Fig. I, although such bonding would appear to be possible across the 1,4 and 8,11 positions. The titration results are not consistent with this conclusion, and indicate further conformational changes as successive protonation proceeds.

3.2. Stability constants

L ~ reacts readily with Cu( I I ) and Ni ( I I ) ions in neutral buffered aqueous solution. The reactions are accompanied by fluorescence quenching (Fig. 2) , which in turn may be used to measure the degree of complex formation, and hence the stability constants.

At pH 6.4, the following reactions take place (L = L 1) :

kL

M 2+ + L . "ML 2+ (10)

kHL

M E + + H L + - " M L 2 + + H + (11)

The observed fluorescence, Fobs, is given by Eq. (12) (Fr ,

Fro., Fn~, FML are the intrinsic molar fluorescences of L, HL +, H2L 2 ÷ and ML 2 ÷ respectively):

Fob s = FL[L] + FnL[HL + ] +FH2L[H2 L2+ ]

+ FML[ML z+ ] (12)

The effect on Fobs of adding increasing amounts of either Cu ( II ) or Ni ( I I ) to L 1 is shown in Figs. 5 and 6, respectively. Defining M,, [M] and [ML] as the total, uncomplexed and complexed metal ion concentrations, and [L] , [HL] and [H/L] as the unprotonated, monoprotonated and diproton- ated ligand concentrations, respectively, Eqs. ( 1 3 ) - ( 1 5 ) apply (KML = [ ML ] / [ M ] [ L ], is the stability constant of the complex formed):

A. M. Josceanu et al. / Inorganica Chimica Acta 240 (1995) 159-168 165

Table 2 Stability constants (1ogmKML) for the interaction of azamacrocycles with divalent transition metal ions in aqueous solution (25 °C)

Ligand Ni (II) Cu ( II )

14N4(cyclam) " 22.2 ~ 27.7 b L l 4.43 5-0.01 6.01 +0.01

a Ref. [23]. bRef. [19].

Mt = [M] + [ML] = [M] ( 1 + KuL[L] ) (13)

L,= [L] + [HL] + [H2L] + [ML] (14)

Hence,

[ML] = L , - [L] a (15)

where

Q= (1 + [H + ]/Kal + [H+]2/KR,K,2) (16)

Eqs. (13) and (15) combine to give Eq. ( 17):

L , - [L] Q KML = ( 17 )

Mr[L]/( 1 + KML[L] )

Eq. (17) rearranges to the quadratic Eq. (18), which may be solved for [ L ].

[L] 2KMLQ - [L] {KML(L , -M,) -- Q} - L, = 0 (18)

Alternatively, from Eqs. (13) and (14),

Lt=Q[L] +Mr- [M] (19)

and hence,

[L] = (L , -M,+ [M] ) / a (20)

Substitution of Eq. (20) into Eq. (13) gives Eq. (21), which rearranges to Eq. (22), a quadratic equation in [M].

M,= [M]{ 1 +KML(L,--Mt+ [M] )/Q} (21)

KML[M]2--[M]{KML(M,--L,)--Q}--QM,=O (22)

Eq. (18) was used when L,>M,, and Eq. (22) when Mt > L,. Knowing the pH and the values of K~, and K~2 deter- mined previously, [L] may be calculated from Eq. (18), then substitution of [L] into Eq. (15) gives [ML], and substitution of [ML] into Eq. (13) gives [M]. Hence KML may be calculated from the equation KML = [ML] / [M] [L]. Alternatively, [M] may be estimated from Eq. (22), which gives [ML] from Eq. (13), and [L] from Eq. (15). Calcu- lated K~L values for the interaction of the fluorescent complex with Cu(II) and Ni(II) are presented in Table 2. The curves shown in Figs. 5 and 6 are the best fits to these equations using non-linear least-squares analysis.

As expected, and in line with the much reduced pK, values, the stability constants are significantly less than the values reported for unfunctionalised cyclam. The Cu 2 ÷ complex of L ~ is more stable than the Ni 2+ complex, as expected from

the Irving-Williams stability sequence, and in line with values reported for cyclam.

Whereas [pK,~(cyclam)-pK,,(L1)] = 3.7 in the ground- state, and 5.6 in the photo excited-state, the values of [logloK~L(cyclam)--logloKML(L 1) ] =21.8 (for Cu e+) and 17.9 (for Ni 2÷ ). This can be partly understood because of the higher charge on the metal ions compared with a proton, but more importantly because the metal ions interact with all four N-atoms. Coordination of all four N-atoms of L ~ results in the Ni 2 ÷ and Cu 2 ÷ cations being brought within range of the electrostatic repulsion from all four of the pendent [Ru (bipy)3] 24 groups, whereas each proton is only close to a single [Ru(bipy)3] 2÷ arm (or at most two arms if H- bonding to the adjacent N-atom occurs). It is not surprising, therefore, that in comparing L t with cyclam, the stability constants are lowered much more than the pK, values.

3.3. Kinetic and mechanistic studies

3.3.1. Reaction of L 1 with copper(ll) The reactions were studied by stopped-flow fluorimetry

under pseudo-first-order conditions, using a large excess of copper(II) (7-25 × 10 -6 mol dm -3) over L 1 (5 × 10 -~ to 10 - 6 mol dm- 3), at five pH values in the range 5.81 to 7.45. At each pH, plots of the pseudo-first-order rate constants (kobs/s -1) versus [Cu(II)] /mol dm -3 are linear (Fig. 7), and a weighted linear least-squares analysis was used to obtain the rate constants for formation (kf/dm 3 mol-1 s - l ) and dissociation (kd/S- i) and their standard deviations using Eq. (23). The results are collected in Table 3.

kob,= + k:[ Cu(II) ] (23)

Values of k/were analysed on the basis of reactions (10) and ( 11 ), assuming kL >> knL which is normal for the reac- tions of azamacrocycles [ 17-20]. With this assumption Eq. (24) applies:

kobs/S -1

60

s0 j

J

30 / / < J / Y

20 / f

10 .~._X~.7~Z]~ ~ _ , m._.____~____-, w

o 0.7 1 1.3 1.6 1.9 2.2 2,5

lOs[Cu(ll)] / tool dm -3

[ ~pH5.81 -~pH6.33 -~pH6.53] ~-pH 7.01 ~-pH 7.45

Fig. 7. Variation of the pseudo-first-order rate constants (kobs) with [ Cu 2 + ] at 5 buffered pH values (2,6-1utidine and 2,4,6-collidine buffers were used).

166 A. M. Josceanu et al. / Inorganica Chimica Acta 240 (1995) 159-168

k_k~ ./-~ (24)

Q is defined by Eq. (16), and since fluorimetric detection was used, the photoexcited state pKa values are applicable. Values of k/in Table 3 as a function of pH were fitted to Eq. (24), allowing pKa, to vary but holding pKa2 at a value of 4.78 determined previously (Table 1). Although pKa2 was determined at an ionic strength/z = 0.2, and values of k/were obtained at an average/z= 1.8 x 10 -3, it was found that for the pH range investigated, varying pK~2 between 3.8 and 5.3 had negligible effect on the determined values of kL and pK~,, and so a correction to pKa2 for the variation in ionic strength was not applied. However, if pK~ was held at the value of 5.92 obtained at/x = 0.2 (from the fluorimetric titra- tion), the fit of the observed k/values to Eq. (24) was rather poor ( Eiwi( k~ b~ - k~ ~1~) 2 = 356), whereas the fit obtained was improved by allowing pKa, to vary (whereby ~']iWi ( k~fbs - - k~falc) 2 = 4 5 ). T h e r e s u l t s o f t h e l a t t e r analysis a r e

shown graphically in Fig. 8, where the best-fit curve is that calculated with pK~, = 6.90 + 0.13, and I 0 - 6 kL = 2.31 + 0.39 dm 3 mo1-1 s -I (if the value of pK~, is fixed at 5.92, then 1 0 - 6 kL = 0.90 + 0.18 dm 3 mol- 1 s - ', the % error in kL again indicating a poorer data fit). The kinetic, fluorimetrically determined photoexcited state value of pK~, at /z ~ 1 . 8 × 10 -3 compares with the value of 5.92 +_0.11 from

"7

O

2"5/ i i i i i i i i i i i i

. . . . . ...... i . . . .

5..0 - - ~ ----4-----4.---4: : : ....... 4------i: : . . . . . . . . i ~----~----~--.-.4-z ' I : . . . . i i i i i i i ~ i i i I : I i I I I [ [ I [ I i i ~ i / - ! ! ! I I !

o.~ ..... i ...... 4--~7~--,.."---#--~---~-......" .... ~- .....

0 . 0 I I I I s . e s . 2 6 . 6 7 . 0 7 . 4 7 . e

pH

Fig. 8. Variation of the second-order rate constants for the formation of [CuL l ] m+ (k:) with pH fitted to Eq. (24).

the fluorimetric pH titration a t /x=0 .2 (Table 1). The esti- mated value of k~. when compared with the rate of water exchange rate for [ C u ( H 2 0 ) 6 ] 2+ (kex ,,~ 1010 s - 1 ) [24] is as expected based on the Eigen-Wilkins mechanism (ky=k,~V,o), and indicates a very small outer-sphere pre- equilibrium constant, Ko of ~ 10 -4 dm 3 mol-1. This is as expected for a 2 + cation reacting with an 8 + charged ligand,

Table 3 Pseudo-first-order rate constants at 25.0 °C (kobs) for the reaction of L ~ with Cu 2+ at five pH values, and the derived formation (k:) and dissociation (k4) rate constants from Eq. (23)

pH 10S[Cu(II) ] (mol dm -3) kobs (s - I ) a 10-s k / (dm 3 mol - I s -1) ka (s -1 )

5.81 0.958 4.02 _+ 0.23 3.13 + 0.27 1.43 + 0.55 1.14 5.01 5=0.28 1.46 6.62 + 0.30 1.96 7.76 _+ 0.13 2.46 9.00 + 0.12

6.33 0.869 5.16 5:0.33 4.49 + 0.26 1.17 5:0.41 0.960 5.74 + 0.25 1.46 7.41 + 0.25 ! .96 10.12 -t- 0.49 2.47 12.44 + 0.30

6.53 0.910 7.77 + 0.14 6.68 -I- 0.22 1.58 + 0.33 1.10 8.56 + 0.25 1 . 4 7 1 1 . 4 2 + 1 . 0 0

1.98 14.50 + 0.24 2.47 18.37 + 0.26

7.01 0.861 12.58 5:0.21 12.17 + 0.56 2.91 + 0.93 0.958 15.55 +0.23 1.46 20.59 + 0.37 1.96 27.20 + 0.33 2.46 32.74 + 0.21

7.45 0.857 20.83 +0.16 18.63 +0.64 4 .74+0.80 0.963 22.79 + 0.25 1.57 33.22 -t- 0.24 2.46 51.43 + 0.42

"Average of at least 5 kinetic runs.

A. M. Josceanu et al. I lnorganica Chimica Acta 240 (1995) 159-168 167

~ ° ! ! i z i i i

0 i i i i ! I i ~ i 5 , _ . . . . . . ~ . . . . . ~ . . . . . . . ~ . . . . . ~ . . . . . . . ~ - - - - d . - . _ _ " . . . . J . . . . . ~ . - . I : : t : : : l t

i i i i i i i i i 3 . o ....... ~ ....... ~-"-'~i ~ .... . . ~L ........ ~i . . . . . ~ ' - - ' ~ I " ~ . . . . . ~ - - - "

i i i i i i i i i e. o ....... ~ ..... ~ ..... ~ ...... ~ ....... ~

i i : i i i

I . o ...... ! ..... t ................... t - - - t . . . . . . . . . . . . . . . . . . . . ! i i i

0.c I I I I 5 . 8 6.2 6.6 7 . 0 7 . 4

pH

Fig. 9. Variation of the rate constants for the dissociation of [ CuL 1 ] io + with pH fitted to Eq. (25).

Tab le 4

Pseudo- f i r s t -o rder ra te cons tan t s at 25 .0 *C (ko~) for the reac t ion o f L 1 wi th Ni 2+ at p H 6 .40 ( 10 - 3 m o l d m -3 , 2 ,6-1ut idine buf fe r )

103[Ni(II) ] 0.501 1.00 2.00 3.01 4.01 (mol dm -3)

103ko~ 5.88+0.04 15.4+0.1 46.0+0.2 104+1 182+1 ( s - ' ) '

a A v e r a g e o f th ree k ine t i c runs.

0 . 2 0

.# 0.15 ........ i ......... i .......... i ......... i ......... T ......... i ......... ............... i i i i i i i i i i i i i i i J i i / I I

. . . . . g i ....... L i I ! 1 1 1 i O . 1 0 . . . . . . . . ! . . . . . . . . . t " . . . . . . . . . ~ . . . . . . . . . ! . . . . . . . . . t - . . . . . . . . . .

I i i i I / t ! i : I : I I I I l : l : : 1 I i

! i I / ! I i i I : I I l I I I

° . ° 5 . . . . . . . . i . . . . . . . . . i .......... ~ ~ . . . . . i ........

0 . 0 0 0.5 1.5 a.5 3.5 4.5

103 [Ni(IT)] / t o o l d in -3

Fig. 10. Var i a t ion o f the pseudo- f i r s t -o rde r ra te cons tan t s (ko~) wi th [ Ni 2 + ]

at p H 6.4 (2 ,6-1ut id ine bu f f e r ) , f i r e d to Eqs . ( 2 3 ) and ( 2 6 ) (ka=O).

and bearing in mind the likelihood of extensive ion-pairing with [LI] 8+ (q.v.) .

Values of ka also increase with pH as shown graphically in Fig. 9. The data were fitted to Eq. (25) using non-linear least-squares, to give ~ = 1 . 1 1 + 0 . 1 3 s - t , and 1 0 7 k/if :

1.31 +0 .16 tool dm -3 s-1

[ H + ] (25)

The kinetically determined stability constant 1Ogto(KML) = logl0(kL/~) = 6.32 + 0.37; this compares rea- sonably with the value from the equilibrium measurements (Table 2) [ IogI0(KML) = 6.01 ]. The presence of the k~d term in Eq. (25) indicates a base catalysed dissociation pathway, probably due to the formation of an hydroxy complex by axial ligation of an O H - ion by the macrocyclic bound Cu 2 +.

3.3.2. Reaction of L 1 with nickel( ll) The lower stability of the nickel(II) complex (Table 2)

requires the use of a higher range of metal ion concentrations to ensure complete complex formation. The slower rate of reaction with nickel(II) compared with Cu( I I ) makes this experimentally possible. The reaction was studied by stopped-flow fluorimetry at pH 6.4 (2,6-1utidine buffer) with [Ni 2+ ] ( 1 - 4 X 10 -3 mol dm -3) in large excess over [L 1] (2 X 10 -6 mol d m - 3 ) . The pseudo-first-order rate constants (Table 4) increase non-linearly with [ Ni 2 + ] as shown graph- ically in Fig. 10. This is attributed to variations in ionic strength (/~) due to the much higher range of metal ion concentrations used. Values of kobs were fitted to Eq. (23) assuming ka is negligibly small (q.v.) , and a l lowing/9 to vary with/z according to Eq. (26) [25]. This gave the rate constant at zero ionic strength (k~), and also gave an indi- cation of the product of the charges on the two reactants (ZAZB).

20tZAZB 3/--~ log(k1) - - l o g ( ~ ) -~ (26)

1+ ~/-~

At 25 °C, in water 2c~= 1.04 [26], and the fit shown in Fig. 10 was obtained with ZAZB=+IO.3+0.2, and k~ = 3.84 + 0.07 dm 3 m o l - 1 s - i. The value ofzAzB is less than the maximum possible value of + 16 for a 2 + cation reacting with an 8 + charged ligand, and suggests significant ion pair- ing of the ligand, which is reasonable for such a highly charged species. Application of Eq. (24) assuming Q has the value from the Cu(I I ) kinetic study gives a value for kL of 15.7 dm 3 mol 1 s -1 for the reaction with Ni 2÷ at 25 °C and zero ionic strength. The rate of water exchange with [Ni(H20)6] 2+ is 3.2 x 104 s-1 at 25 °C [27] , and the value of k~ is again in line with the Eigen-Wilkins mechanism, if the value of Ko is ~ 10 -4 dm 3 m o l - 1 as found for the reaction of L 1 with Cu 2+. Using the stability constant in Table 2, an approximate calculated value of ka is therefore 5.8 x 10 -4 s - 1, which justifies the omission of ka in fitting the values of /Cobs to Eq. (23).

A c k n o w l e d g e m e n t s

We thank the SERC for financial support and for provision of mass spectral facilities at Swansea, and the University of Warwick for a grant to AMJ.

168 A. M. Josceanu et al. / lnorganica Chimica Acta 240 (1995) 159-168

References

[ 1 ] R. Ziessel and J.-M. Lehn, Helv. Chim. Acta., 73 (1990) 1149. [2] L. Prodi, M. Maestri, R. Ziessel and V. Balzani, Inorg. Chem., 30

( 1991 ) 3798. [3] R. Ziessel, M. Maestri, L. Prodi, V. Balzani and A. van Dorsselaer,

Inorg. Chem., 32 (1993) 1237. [4] J.-M. Lehn and R. Ziessel, J. Chem. Soc., Chem. Commun., (1987)

1292. [5] V. Balzani, E. Berghmans, J.-M. Lehn, N. Sabbatini, R. Ter6rde and

R. Ziessel, Helv. Chim. Acta, 73(1990) 2083. [6] N.W. Alcock, F. McLaren, P. Moore, G.A. Pike and SM. Roe, J.

Chem. Soc., Chem. Commun., (1989) 629. [7] F. McLaren, P. Moore and A.M. Wynn, J. Chem. Soc., Chem.

Commun., (1989) 798. [8] P. Moore, S.C. Rawle and N.W. Alcock, J. Chem. Soc., Chem.

Commun., (1992) 684. [9] E. Kimura, S. Wada, M. Shionoya, T. Takahashi and Y. Iitak, J. Chem.

Soc., Chem. Commun., (1990) 397. [10] R.K. Beeston, S.L. Larson and M.C. Fitzgerald, lnorg. Chem., 28

(1989) 4187. [11] R. Grigg and W.D.J.A. Norbert, J. Chem. Soc., Chem. Commun.,

(1992) 1300.

[12] R. Grigg, J.M. Holmes, S.K. Jones and W.D.J.A. Norbert, J. Chem. Soc., Chem. Commun., (1994) 185-187.

[13] E.K. Barefield, F. Wagner, A.W. Herlinger and A.R. Dahl, lnorg. Synth., 16 (1976) 220.

[14] A.G. Orpen, L. Brammer, F.H. Allen, O. Kennard, D.G. Watson and R. Taylor, J. Chem. Soc., Dalton Trans., (1984) 532.

[ 15] G. Brubaker and D.W. Johnson, Coord. Chem. Rev., 53 (1984) 6. [ 16] B. Bosnich, C.K. Poon and M.L. Tobe, Inorg. Chem., 4 (1965) 1102. [ 17] A.P. Leugger, L. Hertli and T.A. Kaden, Helv. Chim. Acta., 61 (1978)

2296 [ 18 ] M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., (1976) 116. [ 19 ] M Kodama and E. Kimura, J. Chem Soc., Dalton Trans., (1976) 1720. [20] M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., (1977)

1473. [21] M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., (1976)

2269. [22] G.R. Willey, M.T. Lakin, N.W. Alcock and C.J. Samuel, J. Inclusion

Phenom., 15 (1993) 293. [23] F. P. Hinz and D. W. Margerum, lnorg. Chem., 13 (1974) 2941; J.

Am. Chem. Soc., 96 (1974) 4993. [24] R.G. Wilkins, Kinetics and Mechanisms of Reactions of Transition

Metal Complexes, VCH, Weinheim, Germany, 2nd edn., 1991, p.415. [25 ] R.G. Wilkins, Kinetics and Mechanisms of Reactions of Transition

Metal Complexes, VCH, Weinheim, Germany, 2nd edn., 1991, p.111.