B R 2 BodPR 2 : highly ﬂ uorescent alternatives to PPh 3 and PhPCy 2 †

The syntheses of highly ﬂ uorescent analogues of PPh 3 and PhPCy 2 based on the Bodipy chromophore are described. The ligands have been incorporated into two-to four-coordinate group 11 metal complexes. The synthesis, characterisation and photophysical properties of the novel ligands and their metal complexes are reported; many of these compounds have also been characterised by single-crystal X-ray di ﬀ raction. Incorporation of the phosphino group and complexation to the group 11 metal centre has little e ﬀ ect on the absorption and emission pro ﬁ les; high molar extinction coe ﬃ cients and ﬂ uorescence quantum yields were still obtained. In particular, incorporation of the dicyclohexylphosphino substituent signi ﬁ cantly increases the quantum yields relative to the parent dyes.


Introduction
Metal complexes of fluorescent phosphines have potential applications in diagnostic cell imaging 1 and catalytic reaction monitoring, 2 by virtue of the sensitivity of the fluorescence technique.As an imaging tool used in vitro, fluorescence microscopy produces high spatial resolution of the nanometre order, giving accurate images of processes at the subcellular level; 1,3 there is current interest in incorporating fluorescent tags onto radiopharmaceuticals, because it is otherwise difficult to image the localisation of such probes in detail.A fluorescent radiopharmaceutical would instead facilitate both in vivo and in vitro imaging, 3 allowing one to gain a better understanding of the probe's mechanism and localisation within cells.The sensitivity of fluorescence spectroscopy compared to its NMR counterpart (10 −6 to 10 −7 M), also signifies that the detection of low concentrations of catalytically active speciesundetectable by other meansought to also be possible. 2However, phosphines conjugated to organic fluorophores often suffer from fluorescence quenching 4 due to reductive photoinduced electron transfer, therefore synthetic routes to this class of compound have not been extensively reported and fluorescent phosphines remain somewhat rare. 5-Bodipy has desirable photophysical properties which include a high fluorescence quantum yield, strong UV-absorp-tion, chemical robustness and good solubility (Fig. 1).6 A common site of modification is at the meso position due to its easy synthetic incorporation.6 Over the last few years new synthetic procedures have been developed for the substitution of the fluorides with aryl/alkyl (C-Bodipy) or ethynyl (E-Bodipy) groups; this development has allowed more sophisticated functions to be introduced on the Bodipy backbone.6 Given the desirable properties of Bodipy and the extensive synthetic routes to derivatives, 6 we aimed to create fluorescent tertiary phosphines based on this versatile fluorophore, coordinate the new ligands to transition metals and study their behaviour.
Group 11 metal phosphine complexes are sought after, in part, due to their known medicinal applications. 7Gold(I), 8 silver(I) 9 and copper(I) 10 phosphine complexes have all shown cytotoxic activity, with significant anti-tumour propertiescurrent complexes are based on monodentate and bidentate phosphines.One breakthrough was the discovery of auranofin (Fig. 1) in the early 1980s by Sutton, 11 a gold(I) phosphine complex, that was approved for clinical use in 1985 to treat rheumatoid arthritis, but which has also been shown to exhibit anticancer properties, 12 and led to the development of several two-coordinate gold(I) phosphine analogues. 8o further develop the possibility of using gold-based drugs, a greater knowledge of their subcellular distribution and mechanism of action is desirablea fluorescent gold phosphine complex could help to understand the biodistribution of such compounds at high resolution and precision. 1,3Both gold and silver also have potential in therapeutic nuclear medicine due to the beta-emitting radioisotopes 199 Au and 111 Ag, which have long half-lives of 3.15 and 7.5 days respectively. 13opper has a range of radionuclides, but the most commonly investigated is 64 Cua positron emitterand thus can be used for diagnostic nuclear medicine purposes in Positron Emission Tomography (PET) imaging; its relatively long halflife of 762 minutes is considered attractive. 14For the aforementioned reasons it would therefore be interesting to investigate the coordination chemistry of novel fluorescent monodentate phosphines with the group 11 metals and measure the photophysical properties of any complexes so synthesised, to ascertain if they are suitable for study in the applications already discussed above.

Synthesis of Bodipy monodentate phosphines
Scheme 1 details our synthesis of the four novel Bodipy monodentate tertiary phosphines 2a/2b and 3a/3b, substituted at the meso position; aryl bromides 1a and 1b 15 were lithiated by reacting with n-butyllithium in diethyl ether at −78 °C to room temperature, followed by the addition of chlorodiphenylphosphine or chlorodicyclohexylphosphine.Both aryl-and alkylchlorophosphines reacted in a similar manner and we also found that the substituent at the boron atom had a limited effect on the overall reactivity of the aryl bromide; all four ligands were produced in good yields ranging from 60 to 84%.The route thus depicts a mild synthetic method for preparing phosphines containing a Bodipy fluorophore.
The 31 P{ 1 H} NMR spectra of the triarylphosphines in d-chloroform showed 2a/2b at δ −5.5 ppm; aryldialkylphosphines 3a and 3b are shifted downfield to δ 2.7 and 2.8 ppm respectively, due to the electron-donating cyclohexyl rings.
Crystals of 2a were analysed by X-ray crystallography and its molecular structure is depicted in Fig. 2; similar atom numbering schemes are used for all the crystal structures.The P-C bond lengths of 1.8302 (19), 1.834(2) and 1.827(2) Å and C-P-C bond angles of 104.03 (9), 101.66 (9) and 102.40(9)°are typical for tertiary phosphines and compare well to triphenylphosphine. 16

Coordination chemistry
Copper.Tetrakis(acetonitrile)copper(I) hexafluorophosphate was treated with two equivalents of 2a/2b or 3a/3b in dichloromethane which led to the formation of [Cu(2a/2b) 2 (CH 3 CN)]-[PF 6 ] (4a/4b) and [Cu(3a/3b) 2 (CH 3 CN)][PF 6 ] (5a/5b), as depicted in Scheme 2; two phosphines coordinate to the copper(I) centre, replacing the labile acetonitrile ligands.Coordination of the phosphines causes a broadening and downfield shift of the 31 P{ 1 H} NMR signal from δ −5.5, and 2.7/2.8ppm for the free phosphines 2a/2b and 3a/3b to δ 0.0/0.1, and 13.4/12.7 ppm for the complexes 4a/4b and 5a/5b respectivelythese values compare well to other copper(I) acetonitrile phosphine complexes. 17,18Crystals of 4a and 4b suitable for X-ray crystallography were obtained by slow diffusion from ethanol/pentane; the molecular structure of 4b is depicted in Fig. 3, whilst the structure of a compound obtained from the attempted recrystallization of 4a is given in the ESI.† Complex 4b contains a three-coordinate copper(I) centre with a non-coordinating PF 6 − anion and one ethanol molecule, which has exchanged for the labile acetonitrile ligand, coordinated to the copper centre.The Cu-P bond lengths of 2.2384(11) and 2.2273(11) Å are typical for copper(I) complexes. 18,19The complex has a dis-Scheme 1 Synthesis of the C-Bodipy substituted tertiary phosphine derivatives. . 17,20Ligand 2b is sterically more bulky than triphenylphosphine, and therefore the four-coordinate structure may be too crowded for the first-row d 10 transition metal in this case.Complex 5a was crystallised by slow solvent diffusion (chloroform-diethyl ether); two types of crystals were produced and their molecular structures are shown in Fig. 4. Structure A is a three-coordinate copper(I) complex with two phosphines and one acetonitrile ligand coordinated to the copper centre, and a non-coordinating anion, PF 6 − .The Cu-P bond lengths of 2.2273(11) and 2.2564(10) Å are typical of copper(I) phosphine complexes. 18,19The complex has a distorted trigonal planar geometry as shown by the P1-Cu-P2, P1-Cu-N5 and P2-Cu-N5 bond angles of 136.11(4)°, 120.68(10)°and 102.36(10)°respectively; the P1-Cu-P2 angle is larger than in complex 4b due to the steric bulk of the cyclohexyl groups.The anion⋯Cu(I) interaction is again weak, as signified by the closest Cu⋯F distance of 3.228 Å. Structure B is a distorted trigonal-planar copper(I) complex with two phosphines coordinated and no bound acetonitrile ligands, but the PF 6 − anion (modelled with disorder) is weakly coordinated to the copper centre as shown by a stronger anion⋯Cu(I) interaction than in structure A; the alternative closest Cu⋯F distances are 2.741(7) Å to F6A as shown in Fig. 4, and 2.964(11) Å to F6B (not shown); F6A is disordered over two symmetry-related positions in the trigonal plane with the copper centre (on a twofold rotation axis) and two P atoms.All other Cu⋯F distances are >4 Å and are thus non-bonded.A search of the Cambridge Structural Database 21 reveals 11 structures recorded with weakly coordinated PF 6 − and BF 4 − anions having Cu⋯F > 2.7 Å. 22 The P-Cu-P bond angle of 149.98(9)°is large for a three-coordinate structure but is consistent with the steric bulk of the cyclohexyl groups and is similar to that in the three-coordinate complex [Cu(PCy 3 ) 2 (ClO 4 )], which has a P-Cu-P angle of 144.46(6)°. 23The symmetry-equivalent Cu-P bonds lengths of 2.2280(15) Å are typical and compare well with [Cu(PCy 3 ) 2 ][PF 6 ], which was prepared by Che et al. 18 In that complex the crystal structure reveals a two-coordinate linear copper(I) structure, with Cu-P bond lengths of 2.213(1) and 2.313(1) Å, and a much larger P-Cu-P angle of 179.47(3)°, consistent with a linear geometry and ionic PF 6 − . 18However, when the counter-ions tetrafluoroborate and perchlorate were employed, three-coordinate structures were obtained (similar to 5a, structure B) with copper-fluorine and copper-oxygen distances of 2.420(6) and 2.220(7) Å and P-Cu-P angles of 159.98(8)°and 144.46(6)°respectively. 23,24 Willet et al. showed  that dicyclohexylphenylphosphine (which is the closest analogue to 3a and 3b) produced [Cu(PCy 2 Ph) 2 ][ClO 4 ] with the perchlorate anion, but with tetrafluoroborate a different copper(I) complex formed, [Cu(PCy 2 Ph) 2 (F-BF 3 )].Currently the crystal structures are not known to confirm whether the counter-ions have any significant coordinating interaction. 25ilver.Treatment of a dichloromethane solution of (1,5cyclooctadiene)hexafluoroacetylacetonatosilver(I) ([Ag(η 2 :η 2cod)(hfa)]) with two equivalents of 2a/2b or 3a/3b led to the formation of the neutral silver(I) complexes [Ag(2a/2b) 2 (hfa)] (6a and 6b) and [Ag(3a/3b) 2 (hfa)] (7a and 7b), as depicted in Scheme 3. Two phosphines coordinate to the silver, displacing the labile 1,5-cyclooctadiene, to give the products in nearly quantitative yields.Silver has two NMR-active isotopes, 107 Ag (I = 1/2, natural abundance 52%) and 109 Ag (I = 1/2, natural abundance 48%), therefore the expected 31 P{ 1 H} NMR spectra for 6a/6b and 7a/7b would consist of two doublets arising from 107 Ag-P and 109 Ag-P spin-spin coupling. 26osphine coordination resulted in a downfield shift of the 31 P{ 1 H} NMR signal at room temperature from δ −5.5 ppm for the free phosphines 2a/2b to broad peaks at δ 10.2 and 6.6 ppm for the complexes 6a and 6b respectively (Fig. 5 and ESI †); the signal was a broad singlet at elevated temperatures in both cases, which likely results from rapid phosphine exchange. 26his lability is reduced at low temperatures and the Ag-P coupling can be observed. 26t low temperature (−40 °C) two doublets appeared; 1 J 107AgP = 444 Hz and 1 J 109AgP = 511 Hz, and 1 J 107AgP = 446 Hz and 1 J 109AgP = 507 Hz coupling constants were observed for 6a and 6b respectively; these values are typical for silver(I) complexes with two phosphines bound 26,27 and compare well to [Ag-(PEt 3 ) 2 (hfa)] ( 1 J 107AgP = 468 Hz). 28For both 6a and 6b a second set of two low-intensity doublets was observed at −40 °C and indicates that a second silver species was present in solution.For these signals the following Ag-P couplings were measured: 1 J 107AgP = 696 Hz and 1 J 109AgP = 803 Hz, and 1 J 107AgP = 590 Hz and 1 J 109AgP = 677 Hz for 6a and 6b respectively.The increase in Ag-P coupling constants indicates a change in the hybridisation state of the silver(I) complex 26,27 and is attributed to [Ag-(2a/2b)(hfa)] with only one phosphine ligand bound.The coupling constants are similar to the reported values for [Ag(PR 3 )-(hfa)] (R = Ph, Me, Et): 1 J 107AgP = 700-760 Hz. 28,29 Puddephatt et al. reported that an equilibrium is established on the addition of excess phosphine to [Ag(PR 3 )(hfa)] complexes and showed that the Ag-hfa bonding is more ionic when extra phosphines are present. 28,30Mass spectrometry confirmed the [Ag(2a/2b) 2 ] + product, giving peaks at m/z of 1469.6056(6a) and 1217.5443(6b), with loss of the hfa ligand.Low-intensity peaks at 786.2561 and 663.2251 were also observed, which correspond to [Ag(2a/2b)] + , in addition to peaks at 2148.9477 and 1777.8560 for the tris-cations [Ag(113a/ 113b) 3 ] + , but these latter species were not detected by NMR spectroscopy.
On coordination of the dicyclohexyl phosphines 3a and 3b the 31 P{ 1 H} NMR signals at room temperature were also shifted downfield for the complexes 7a and 7b; however, instead of a broad singlet, two broadened doublets were observed at δ 23.7 ppm and δ 23.9 ppm for 7a and 7b respectively (Fig. 6 and ESI †).The following Ag-P coupling constants were observed for 7a and 7b respectively: 1 J 107AgP = 453 Hz and 1 J 109AgP = 517 Hz, and 1 J 107AgP = 453 Hz and 1 J 109AgP = 519 Hz.At higher temperatures (120 °C) 7a and 7b gave a broad singlet.The observation of Ag-P coupling at room temperature was unusual since the rapid exchange of the phosphine ligands usually causes this coupling to be averaged to zero; this ligand exchange is normally slowed down by cooling the solution to low temperatures (as is the case for 6a and 6b). 27However, the observation of Ag-P coupling at room temperature has been reported previously, where the phosphine is sterically hindered or chelating. 27,31Phosphines 3a and 3b are also bulky which may explain the lower rate of phosphine exchange and the observation of Ag-P coupling.The electrospray mass spectra gave peaks for [M − (hfa)] + at m/z of 1489.7998 for 7a, and 1242.7354 for 7b.Complexes 6a and 7b were also characterised by X-ray crystallography (Fig. 7).The complexes have neutral silver(I) tetrahedral four-coordinate geometries with two phosphines and one hfa ligand bound to the metal.The Ag-P bond lengths of 2.4255(10) Å and 2.4028(13) Å are typical for silver(I) phosphine complexes. 25,30,32The complexes are somewhat distorted, shown by the P-Ag-P, P-Ag-O and O-Ag-O bond angles (Fig. 7 caption).The synthesis of a related four-coordi-nate silver(I) complex [Ag(PPh 3 ) 2 (hfa)] has been reported in a similar fashion, but currently no crystal structure is known. 33owever, the three-coordinate complex [Ag(PPh 3 )(hfa)] has had its solid state structure determinedshorter bond lengths are observed: 2.346(3) for Ag-P, and 2.341(5) and 2.218(5) Å for Ag-O. 29For 7b the Ag-O bond distances of 2.337( 6) Å (Ag-O1) and 2.718( 8) Å (Ag-O2) are significantly different, and the latter is longer than the normal covalent silver(I)-oxygen bond length of ca.2.3 Å, which indicates that the second oxygen is only weakly bonded to the silver; 28 the hfa ligand is disordered over two positions related by the twofold rotation axis passing through Ag.The bite angle of the phosphines is larger than in 6a, due to the increased steric bulk of the phosphine 3b.No β-diketonate silver complexes with dicyclohexylphenylphosphine or tricyclohexylphosphine have been reported, however cationic two-and three-coordinate silver(I) complexes do form with dicyclohexylphenylphosphine and the non-coordinating perchlorate, tetrafluoroborate, hexafluorophosphate and hexafluoroantimonate anions. 25,32,33old.Treatment of chloro(tetrahydrothiophene)gold(I) [AuCl-(tht)] with one equivalent of 2a/2b or 3a/3b in dichloromethane led to the formation of the neutral gold(I) complexes [AuCl(2a/ 2b)] (8a and 8b) and [AuCl(3a/3b)] (9a and 9b), as depicted in Scheme 4. In each case one phosphine coordinates to the gold, replacing the labile tht ligand; purification was achieved by column chromatography, resulting in high yields of 75-91%.
Coordination of the phosphines resulted in a downfield shift of the 31 P{ 1 H} NMR signal of the free phosphines 2a/2b and 3a/3b to δ 33.2/33.3and 51.4/51.6 ppm for the complexes 8a/8b and 9a/9b respectively.Complexes 8a, 9a and 9b were also characterised by X-ray crystallography (Fig. 8 and ESI †).The solid-state structures show them to be gold(I) two-coordinate linear complexes, as expected.

Optical properties
After the synthesis of several group 11 metal complexes of 2a/2b and 3a/3b, it was important to determine and understand their photophysical properties.Our initial concern was whether the phosphorus donor 4 or the heavy metals 35 themselves would quench the fluorescence of the Bodipy fluorophore, which has been shown to occur for other fluorophores in phosphorus systems.Photophysical data were collected for all ligands and complexes in dry degassed tetrahydrofuran to minimise photobleaching and phosphine oxidation in solution (Table 1).
The four phosphine ligands 2a/2b and 3a/3b all show a typical Bodipy profile 6 with absorption maxima at either 517/513 nm or 518/512 nm depending on the groups at the boron centre; changing from diphenyl to dimethyl causes a hypsochromic shift of the absorption maxima (4-6 nm).The lowest energy maximum is assigned to the S 0 -S 1 (π-π*) tran-   sition for the Bodipy core.All the ligands have typically high molar absorption coefficients, ranging from 77 000 to 92 000 M −1 cm −1 . 6Secondly, a lower intensity, broader absorption band can be seen between 370 and 385 nm (ε = 3000-10 000 M −1 cm −1 ), which is attributed to the S 0 -S 2 (π-π*) transition of the Bodipy core.There are also low-intensity, high-energy peaks centred between 250-300 nm for the dicyclohexyl derivatives 3a/3b.The absorption spectra for phosphines 2b/3b and their complexes are given in Fig. 9; the corresponding spectra for 2a/3a and their complexes is given in the ESI.† Room-temperature fluorescence of the phosphine ligands was readily detected, with maxima of 534/527 nm or 533/526 nm, again depending on the boron substituents (Fig. 10).The Stokes shifts of 14-17 nm are small, which is common for the Bodipy fluorophore, and suggests that only small structural changes occur on excitation. 6The fluorescence quantum yields of 2b (0.29) and 3b (0.44) are typically high for Bodipy molecules and compare well to the parent Bodipy 10b (0.35, ESI †), which shows the phosphorus donor does not quench the fluorescence.This is unusual and, as with our previously reported Bodipy primary phosphines, 15 contradicts several phosphine examples. 4owever, it is perhaps to be expected, given that the DFT calculations (see Fig. 11 and ESI †) show there is no phosphorus character in the HOMO, HOMO−1 or HOMO−2 for 2a and 3a, nor in the HOMO or HOMO−1 for 2b and 3b, and that the energy difference between the HOMO and the first phosphorus-containing orbital is 0.9 eV for all the phosphine ligands (Fig. 11 and ESI †).The quantum yield for 3b is, perhaps surprisingly, also higher than that of the parent Bodipy which has no substituents on the meso phenyl ring (3b = 0.44 and 10b = 0.35).More detailed investigations into a Bodipy phosphine series could reveal how the quantum yield is affected by both the electronic and steric nature of the substituents on the phosphorus.Changing the methyl groups at the boron atom for phenyl groups severely quenches the fluorescence (compare 2b: Φ F = 0.29 to 2a: Φ F = 0.042), which is consistent with our previous findings. 15The absorption spectra of the complexes are very similar to those of the un-  coordinated ligands (Fig. 9); wavelengths of the low energy transition fall in the range 519-512 nm.The corresponding molar extinction coefficients are large for the copper (4-5) and silver (6-7) complexes (ε = 133 000-171 000 M −1 cm −1 ), due to the presence of two Bodipy ligands; the monodentate gold complexes (8-9) retain similar values to the free ligands (ε = 74 000-85 000 M −1 cm −1 ).The fluorescence spectra of the complexes 4b-9b are displayed in Fig. 10.On complexation, the fluorescence quantum yields are generally retained, with only a slight decrease observed on descending the group (for instance 2b: Φ F = 0.29, 8b: Φ F = 0.20).The gold species (8-9) have the lowest relative quantum yields, likely due to the heavy atom effect.Quenching is less pronounced for the aryldialkylphosphine complexes of 3b compared to those of the triarylphosphine 2b (Table 1, Fig. 10) and all three metal complexes of 3b have higher quantum yields than the parent Bodipy 10b (Φ F = 0.35).It is noteworthy that Gray et al. reported group 11 complexes which contained separated azadipyrromethene and triphenylphosphine ligands; however, the molar extinction coefficient and fluorescence quantum yields are significantly lower in those cases, 30 000-65 000 M −1 cm −1 and 0.0024-0.0039respectively. 37

Conclusions
The synthesis of fluorescent monodentate triaryl and aryldialkyl tertiary phosphines has been achieved in excellent yields via the lithiation of the Bodipy aryl bromides 1a/1b, and subsequent addition of the relevant chlorophosphine.This route may be transferable to a range of aryl and alkyl chlorophosphines and thus could be an excellent route to a library of new Bodipy tertiary phosphines.The photophysical properties of the phosphines are intriguing; the fluorescence is not quenched significantly compared to their precursor and, in fact, in the case of aryldialkylphosphines 3a and 3b, the emission is enhanced in comparison to their parent Bodipys 10a and 10b (see ESI †).The novel phosphines 2a/2b and 3a/3b readily coordinate to low-oxidation-state coinage metals at room temperature in nearly quantitative yield; upon coordination, the fluorescence of the phosphines is not significantly quenched.These novel ligands and their complexes have potential applications in medicinal imaging, as the fluorescent Bodipy functionality will facilitate cell imaging.Several group 11 metal phosphine complexes have shown cytotoxic properties against several cancer cell lines, [8][9][10] and our novel derivatives may well display similar attributes.Future work will focus on their use in diagnostic imaging and therapy.
Flash chromatography was performed on silica gel from Fluorochem (silica gel, 40-63 μ, 60 Å, LC301).Thin-layer chromatography was performed on Fisher aluminium-based plates with silica gel and fluorescent indicator (254 nm).Infrared spectra were recorded on a Varian 800 FT-IR spectrometer and mass spectrometry was carried out by the EPSRC National Mass Spectrometry Service Centre, Swansea. 1 H, 13 C{ 1 H}, 31 P{ 1 H}, 19 F{ 1 H} and 11 B{ 1 H} NMR spectra were recorded on a JEOL Lambda 500 ( 1 H 500.16 MHz) or JEOL ECS-400 ( 1 H 399.78 MHz) spectrometer at room temperature (21 °C) using the indicated solvent as internal reference, unless specified otherwise; 1 H and 13 C shifts were relative to tetramethylsilane, 31 P relative to 80% H 3 PO 4 , 11 B relative to BF 3 •Et 2 O and 19 F relative to CFCl 3 .It is worth noting that, as for dicyclohexylphenylphosphine, 39 six unique cyclohexyl carbon signals are seen in the 13 C{ 1 H} NMR spectra for phosphines 3a and 3b and their respective complexes.The six signals arise as the two cyclohexyl rings are equivalent but the carbon atoms that are symmetrically equivalent are diastereotopic; these diastereotopic carbon atoms are also non-equivalent with respect to P-C coupling constants.
All calculations were carried out using Spartan 10 software. 40Full geometry optimizations of the studied compounds were performed using density functional theory with a B3LYP/ 6-31G* basis set.A vibrational analysis was performed at the same level to characterize calculated structures as minima.

X-ray crystallography
Data were collected at 150 K (120 K for 2a) on an Agilent Technologies Gemini A Ultra diffractometer with MoKα (λ = 0.71073 Å, for 4a, 5a structures A and B, and 9b) or CuKα (λ = 1.54178Å, for 4b, 6a, and 8a) radiation, 41 and on a Crystal Logics diffractometer equipped with a Rigaku Saturn 724+ detector at beamline I19 of Diamond Light Source using a synchrotron X-ray wavelength of 0.6889 Å (for 2a, 7b, and 9a). 42elected crystallographic information is given in Table 2. Absorption corrections were based on multiple and symmetryequivalent data; the structures were solved by direct and heavyatom methods, and refined on all unique F 2 values with appropriate constraints and/or restraints in each case, particularly for the treatment of disordered structural components. 43Unidentified solvent and/or counter-ion in the structure of the complex obtained from the attempted recrystallization of 4a was treated by the Squeeze procedure of PLATON; 44 H atoms could not be observed in difference maps for this relatively low-precision structure, so it is not clear whether the complex is a neutral Cu(II) complex with two ethoxide ligands or a cationic Cu(I) complex with two ethanol ligands and a highly disordered small anion, though the latter is more likely from the observed tetrahedral geometry.Further discussion and evidence from spectroscopic data is provided in the ESI.† CCDC references: 987389 (2a), 987388 (4a), 987387 (4b), 987386 (5a A), 987385 (5a B), 987390 (6a), 987384 (7b), 987391 (8a), 987383 (9a) and 987382 (9b).