5 Discussion
5.6 The effects of bromide (and other dissolved molecules)
It is obvious from the results that bromide inhibits coordination of Cu2+ to adamanzanes (figure 4.21), and the question is how. Does it block the required UV light, chemically inhibit the effect of the light or inhibit chelation regardless of UV light? Perhaps all three!
5.6.1 UV filter effect (physical inhibition of light induction)
The bromide filter absorbs UV light with a wavelength of less than 220 nm (figure 3.4). This means that the bromide absorbs photons with a higher energy than 5.6 eV.
If the spectral increase from pH 5 to pH 6 in figure 4.1 is the LMCT band of the B intermediate, then the majority of the band is at λ > 220 nm, which means that the bromide filter should only cause a small reduction of chelation speed. The larger effect on chelation rate, which is observed when using the bromide filter (data not shown), can be explained by a secondary reaction, which is rare at energies under 5.6 eV (figure 5.3) [66]. If the F intermediate is a result of a high energy photon, then the nitrogen radical can remove an electron from a C-C bond and form a N-C double bond leaving the other C atom as a radical (J1).
Alternatively, the F intermediate can be irradiated and then form J1. Cu(I) can then deliver the electron back to the ligand resulting in K. K can also be formed by irradiation of E, which in most cases will result in L. L like F can break a C-C bond forming J2. In both these cases, the formation of J does not require much additional energy if the N=C bond is included. However, the C-C bond has to be broken first, so the activation energy makes this process unlikely unless additional energy is added.
There is very little chelation when UV light with λ < 280 nm is removed by a filter, so the minimum energy required for oxidation of an adamanzane amine must be 4.4 eV. Thus, a simultaneous adamanzane oxidation and breakage of an N-C bond, which is required for the E → M reactions, requires 4.4 eV + 3.2 eV (N-C bond) = 7.6 eV, which corresponds to 163 nm. Photons with a wavelength of 163 nm are absorbed by water, so breakage of N-C bond by a single photon is rare unless as a thermal reaction [181]. Nonetheless,
molecules at the side of the cuvette where the UV light enters could be hit by a 7.6 eV photon. A 2 mm water filter removes that possibility and this is probably the explanation for the effect of a water filter on chelation rate. The chance for a two photon excitation increases with the photon flux and can be avoided by flash irradiation. Multiple flashes separated by a second or more should also prevent breakage of the adamanzane, even with UV irradiation at λ < 220 nm. Moreover, it is likely that the effect of the bromide filter in
preventing a secondary reaction is to reduce the photon flux enough to make irradiation of radical intermediates rare. In that case, increasing the flux sufficiently would bring back secondary reactions.
Other dissolved molecules able to absorb UV light have a similar effect, depending on their cut-off.
E 2+
NH N
HN
N (II) Cu
F 2+
N N
N N
H (I)
Cu
H J1 2+
N N
N H2C N
H (I)
Cu
H K 2+
N N
N H3C N
(II) Cu
H
L + N
N H N N
(I)
Cu
hν
M2 +
NH N
HN
N (I)
Cu
M1 + NH
N H N N
(I)
Cu
hν
hν
N2 2+
NH N
HN
HN (II) Cu
N1 2+
NH N
H N NH
(II) H+ Cu
H+
H+ H+
H+
H+ (hν)
J2 +
N N
H N H2C N
(I) (hν) Cu
Figure 5.3: Proposed reaction scheme for high energy side reaction of F intermediate in coordination of Cu2+ to [24.31]adz and of the chelate E. E → L does not require high energy photons. Only coordination to the adamanzane is shown.
5.6.2 Competitive coordination
Other molecules able to coordinate Cu(II), such as acido ligands, can compete with adamanzanes for the copper ions. Since Kf for Cu(II) coordination to adamanzanes are more than 1010 fold higher than for coordination to acido ligands and buffer anions, this will only affect the rate of chelation unless another strong chelator is added. The kinetic effect can be considerable though, as long as the competing molecules have a much faster kon than the adamanzane. Therefore, it is important to consider the competitive inhibitory effect of buffer molecules and ion strength modifiers. Unfortunately, most buffers in the biological range are Cu(II) coordinating. Bromide only coordinates Cu(II) weakly, so it is unlikely that competitive coordination contributes significantly to the inhibitory effect of bromide at low concentrations.
Acido ligands can also coordinate to Cu(II) when it is partly or fully coordinated by an adamanzane. The coordination to chelated Cu(II) stabilizes the chelate, which should increase the rate if dissociation would be significant otherwise. It is also possible though, that an acido ligand coordinating Cu(II) in the B
intermediate will decrease the chelation rate by reducing the positive charge and thereby the acidity of the intermediate, which is rate determining.
5.6.3 Redox interference by acido ligands (chemical inhibition of light induction) Upon UV irradiation, another effect is added to the pallet. If Cu(II) oxidizes the adamanzane, its acidity increases enabling a fast removal of the last proton bound in the cavity as shown in figure 5.2. However, Cu(II) can also oxidize acido ligands upon UV irradiation resulting in O (figure 5.4 left). This has been shown for UV irradiation of Cu(II) macrocycles [66,181]. O has a lower acidity than B and since the adamanzane is not a radical, it depends on X• for further reaction. Cu(I) chelates are much more labile than Cu(II) chelates, and Cu(I) have different donor group preferences [20]. Therefore, it is likely that Cu(I) will dissociate from the adamanzane unless reoxidized by X• before that. If X• reoxidizes Cu(I) to Cu(II), then the result is A or B, depending on whether the copper ion is still coordinated by the adamanzane. It is also possible, especially at high concentrations, for two acido radicals to combine. The result is production of Cu(I) and X2. This can also be a result of UV irradiation of the stable chelate E. Even without X2 formation, UV light can cause dissociation of copper from E when an acido ligand in coordinated (figure 5.4 right). If Cu(II) is reduced by an acido ligand the result is L, but L is much less robust than E, making reductive transchelation likely. This is the case for most Cu(II) chelates, when used in vivo [222]. An adamanzane capable of coordinating in tetrahedral conformation is more likely to keep Cu(I) coordinated [20].
B
hν N
N N
N H
Cu (II)
F +
N N
N N
H
Cu(I)
H H
H O
N N
N N
H (I)
Cu
H H X
X +
hν X
E
N N
N N
Cu(II) H H
I
N N
N N
Cu(I) H
H+
X
2+
+
X +
X
L
N N
N N
(I)
Cu
H H
+ hν
hν H+ 2+
Figure 5.4: Proposed reaction scheme for side reactions involving acido ligands. A LMCT can either cause an oxidation of the adamanzane or the acido ligand.
The effect of bromide in solution under irradiation has not been compared directly to the effect as UV filter only; however, chloride is far more inhibitory when added to the solution than as UV filter only, and bromide is easier to oxidize than chloride, so the redox interference by bromide can be expected to be even larger.
The rate increasing effect by adding nitrate at 25 °C and 40 °C is inversed at 55 °C (figure 4.22). A possible explanation is that Cu(I) is oxidized by nitric acid but only at a significant rate at 55 °C. If Cu(I) is oxidized by nitric acid, it cannot afterwards reduce the adamanzane to form the final chelate.
5.6.4 Redox interference by oxygen
It is evident from the results (figure 4.25) that dissolved gases inhibit UV light induced chelation of Cu2+. Not only that, but they also decrease the initial spectrum (of the reactants). One obvious possibility is that bobbling argon through the Cu2+ and [24.31]adz solutions causes some evaporation. However, if evaporation alone should have caused the more than 50 % increase in absorption around 270 nm, more than a third of the volume should have evaporated and that would have been clearly visible. Consequently, evaporation might be a part of the effect but only a small part.
Carbonates can coordinate to Cu(II) but was not removed from the [24.31]adz solution due to its high pH (>
10) before mixing with the Cu2+ solution, and the Cu2+ solution is too acidic (pH < 4) to contain carbonates in significant amounts (at atmospheric pressure) before mixing with the [24.31]adz solution. That leaves O2, and removal of oxygen by argon increases the rate of chelation similar to the increase in Cu(I) formation observed by Moore [152].
O2 is known to form Cu(II) superoxo complexes with Cu(II) by oxidation of Cu(I) [84]. These can be quite stable and are likely to increase to life-time of the C, F and G intermediates (figure 5.2) but with Cu(II) instead of Cu(I). Oxidation of Cu(I) in C and G prevents simple reversion to B and H respectively.
Redelivery of the electron would require Cu(III) formation, which is possible but less likely. Therefore, O2
presence can be expected to increase formation and stabilization of the F intermediate but with Cu(II). The superoxide formed by oxidation of Cu(I) is basic and is likely to form HO2• with the H+ released from the C intermediate [167]. Redelivery of the electron to N• from HO2•, results in B. Thus the overall result of O2 is a higher rate of H→F and F→B, and consequently a lower rate of B→E.
O2 presence can also result in photodegradation of the adamanzane [167]. This is probably a result of oxygen radical formation, such as HO•. It is possible that the effect of the bromide filter in preventing a secondary reaction is preventing formation of oxygen radicals. In that case, UV filtering is not necessary when chelation is performed under argon.
5.6.5 pH stabilization
Very high pH lowers chelation rate due to Cu(OH)2 formation and perhaps because HO- can bind to the B intermediate and form HO· by redox interference. At pH 5 or lower the B intermediate does not form, but between pH 6 and 8 the rate does not seem to change. If a buffer is considered necessary, it is important to choose one that does not cause any of the three problems that bromide does.