The impact of ionic liquids on the coordination of anions with solvatochromic copper complexes

View Article Online

View Journal

Dalton

Transactions

Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: O. Kuzmina, H.

Hassan, L. Patel, C. Ashworth, E. Bakis, A. J. P. White, P. Hunt and T. Welton, Dalton Trans., 2017, DOI:

10.1039/C7DT02372C.

This is an Accepted Manuscript, which has been through the

Volume 45 Number

1 7 January 2016 Pages 1–398

Royal Society of Chemistry peer review process and has been

accepted for publication.

Dalton

Transactions

Accepted Manuscripts are published online shortly after

An international journal of inorganic chemistry

www.rsc.org/dalton

acceptance, before technical editing, formatting and proof reading.

Using this free service, authors can make their results available

to the community, in citable form, before we publish the edited

article. We will replace this Accepted Manuscript with the edited

and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the

author guidelines .

Please note that technical editing may introduce minor changes

to the text and/or graphics, which may alter content. The journal’s

standard T erms & Conditions and the ethical guidelines, outlined

in our author and reviewer resource centre, still apply. In no

event shall the Royal Society of Chemistry be held responsible

for any errors or omissions in this Accepted Manuscript or any

consequences arising from the use of any information it contains.

ISSN 1477-9226

PAPER

Joseph T. Hupp, Omar K. Farha et al.

Eﬃcient extraction of sulfate from water using

framework

a Zr-metal–organic

rsc.li/dalton

Please do not adjust margins

Dalton Transactions

Page 1 of 16

View Article Online

DOI: 10.1039/C7DT02372C

Journal Name

ARTICLE

The impact of ionic liquids on the coordination of anions with

solvatochromic copper complexes

O. Kuzmina^a, N. H. Hassan,^{a, b}L. Patel,^aC. Ashworth,^aE. Bakis,^aA. J. P. White,^aP. A. Hunt^aand T.

Welton^a

Received 00th January 20xx,

Accepted 00th January 20xx

DOI: 10.1039/x0xx00000x

Solvatochromic transition metal (TM)-complexes with weakly associating counter anions are often used to evaluate

traditional neutral solvent and anion coordination ability. However, when employed in ionic liquids (IL) many of the

common assumptions made are no longer reliable. This study investigates the coordinating ability of weakly coordinating

IL anions in tradtional solvents and within IL solvents employing a range of solvatochromic copper complexes. Complexes

of the form [Cu(acac)(tmen)][X] (acac = acetylacetonate, tmen = tetramethylethylenediamine) where [X]^–= [ClO₄]^-, Cl^–,

[NO₃]^–, [SCN]^–, [OTf]^–, [NTf₂]^–and [PF₆]^–have been synthesised and characterised both experimentally and computationally.

ILs based on these anions and imidazolium and pyrrolidinium cations, some of which are functionalised with hydroxyl and

nitrile groups, have been examined. IL-anion coordination has been investigated and compared to typical weakly

coordinating anions. We have found there is potential for competition at the Cu-centre and cases of anions traditionally

assigned as weakly associating that demonstrate a stronger than expected level of coordinating ability within ILs.

[Cu(acac)(tmen)][PF₆] is shown to contain the least coordinating anion and is established as the most sensitive probe

studied here. Using this probe, the donor numbers (DNs) of ILs have been determined. Relative donor ability is further

confirmed based on the UV-Vis of a neutral complex, [Cu(sacsac)₂] (sacsac = dithioacetylacetone), and DNs evaluated via

²³Na NMR spectroscopy. We demonstrate that ILs can span a wide donor range, similar in breadth to conventional

solvents.

www.rsc.org/

loosely associated ions forming and breaking over time. Each

ion is interacting with multiple other ions, however at any

instant in time most of the ions are not strongly associated and

Introduction

Ionic liquids (ILs) are versatile ionic systems used in wide range

of applications¹including synthesis,²electrochemistry,^3,4metal

extraction processes,⁵homogeneous and heterogeneous

catalysis.^2,6ILs have significant advantages over conventional

molecular solvents, for example; a very low vapour pressure

under ambient conditions and therefore low-volatility, in

addition ILs exhibit polarities similar to short chain alcohols

combined with a soft electronic character. ILs are good

solvents for a broad spectrum of inorganic, organic and

polymeric materials and are immiscible with numerous organic

solvents.²There is a wide range of possible cation and anion

combinations, meaning that the properties of ILs can be tuned

to fit a specific purpose.⁷

therefore have been considered to be “non-associating” or

"non-coordinating”. The reader should not be misled by these

terms as the coordination ability of “non-coordinating” anions

was presented as early as 1973.⁸Given an ion pair, an anion

can be non-coordinating dependant on the electrophilicity of

the associated cation⁹However a range of weak interactions is

usually possible and therefore we use terms “weakly

coordinating” or “effectively non-associating” anions.

ILs have been studied as solvents for biphasic catalysis with

transition metal complexes since the 1990’s.¹⁰ILs can affect

TM-based catalysis through solvation of the starting materials,

catalyst and/or products, as well as relevant intermediates and

transition states, particularly those of the TM-catalyst. Solute-

solvent interactions can influence reactivity, reaction rates,

reaction mechanisms, side reactions and the distribution of

final products.^11,12ILs can play a key role in TM-based catalysis

and it is important that we develop a robust understanding of

how IL solvents interact with TM complexes.

IL solvents are typically composed of large organic cations and

smaller inorganic or organic anions. The IL solvent can be

represented as a dynamic system, an equilibrium exists

between individual ions, ion-pairs, and larger networks of

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 1

Please do not adjust margins

Dalton Transactions

Page 2 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

In biphasic systems the catalyst is often charged to achieve

optimum retention in the IL phase, thus typically the catalyst is

a cation present with an associated counter-anion. Recent

studies have indicated that some metal salts do not remain

associated in ILs; the solute cation and counter-anion are

individually solvated by the IL ions.¹³This means that the

"weakly interacting" IL ions are capable of disrupting the

attractive Coulombic interactions within the metal salt ion

pairs. Other studies have shown that ILs do not usually form

ideal solutions with metal salts.⁵Ideal solutions form when

there are no preferential interactions between the solute and

solvent and dissolution is driven by the entropy of mixing.

Non-ideality provides evidence that preferential interactions

are occurring, i.e. a positive enthalpy of association between

the solute and solvent, in this case the IL ions and the TM

cation and its counter-anion. Thus, it is uncertain whether IL

solvents interact with the catalyst TM-complex, behaving

similarly to a traditional neutral non-interacting solvent or

whether ILs associate in a manner similar to the counter-anion,

i.e. forming an ion-pair with the solute.

neutral solvent is poorly coordinating, and the counter-anion

remains associated (Cu---S << Cu---[X]^–), as the donor strength

of the complexing anion increases coordination is stronger and

a neutral ion-paired square based pyramid geometry evolves

(Figure 2, left). If the counter-ion is poorly coordinating (Cu---

[X]^–<< Cu---S), then as the donor strength of the solvent

increases, solvent coordination is stronger and the complex

will move from a "tetragonally" distorted structure to the limit

of an "octahedrally" coordinated cationic complex, (Figure 2

right). An octahedral environment associated with the

coordination of two counter-anions is unlikely due to the

stoichiometry moreover, Cu(II) complexes have a d⁹electronic

configuration which leads to a strong Jahn-Teller distortion.²⁰

In addition an overall negatively charged complex will be

destabilised in a neutral solvent. It is generally assumed that

distortion away from the square planar complex (by either

anion or solvent coordination) leads to a reduction in the

energy of the primary electronic transition leading to the

observed absorbance peak in spectra shifting: the wavelength

Solvatochromic transition metal complexes are ideal probes

for understanding how TM catalysts are solvated.

Solvatochromism measures the influence of

a solvent

environment on, for example, the electronic absorption

spectra of molecules.¹⁴Ni, Fe and Cu complexes are common

solvatochromic indicators used to describe solute-solvent

interactions and the properties of the solvent.¹⁵

(Acetylacetonato)(N,N,N’,N’-tetramethyl-

ethylenediamine)copper(II) salts (Figure 1), abbreviated to

[Cu(acac)(tmen)][X],

are

known

to

be

strongly

solvatochromic.^16–18These Cu complexes are favourable in that

coordination of counter ions or solvation molecules is believed

to lead to significant structural change, which alters the energy

of the d–d transitions, leading to significant colour changes.¹⁵

Simple empirical models have been developed to structurally

and electronically classify the observed colour changes of Cu-

complexes.¹⁹Figure 2, summarises the typical interaction

modes and the resultant crystal field theory (CFT) splitting

patterns that evolve when [Cu(acac)(tmen)][X] is dissolved in a

traditional neutral solvent. Simple models like these provide a

good starting point for discussion, however they should not be

over-emphasised when examining non-neutral systems with a

low symmetry or when considering transition metal complexes

in detail; in these cases modern MO theory should be

employed.

λ

_maxincreases while the wavenumber ν_maxdecreases.

Experimentally, if the donor strength of the solvent is higher

than that of the anion (Cu---S >> Cu---[X]^–), the solvent will be

preferably coordinated to the Cu complex (pseudo-octahedral

coordination) and ν_maxwill be responsive to the nature of the

solvent. The donor strength of a solvent is evaluated relative

to a reference, typically 1,2-dichloroethane (DCE), which is

known to be a very poorly coordinating solvent and defined to

have a donor number (DN) of zero. Meinershagen and Bein

have examined the shift of λ_maxin solvents with different

donor

number

introduced

as

a

vapour

into

The

zeolites/[Cu(acac)(tmen)]⁺

complexes.²¹

[Cu(acac)(tmen)][ClO₄] and [Cu(acac)(tmen)][BPh₄] complexes

are commonly used for determining the basicity of a solvent;

these anions are assumed to be "effectively non-associating".¹⁷

If neither the solvent (S) or the counter-anion ([X]^–) are

associating, [Cu(acac)(tmen)]⁺remains uncoordinated and

takes on a square planar geometry (Figure 2, centre). If the

The

degree

of

coordination

indicated

by

2 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Dalton Transactions

Page 3 of 16

Journal Name

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

[Cu(acac)(tmen)][BPh₄] in different solvents has been

correlated to the turnover frequency for catalytic reactions.²²

In a weakly coordinating solvent ν_maxwill remain at constant

value until a counter anion with a larger donor strength than

the solvent is used, subsequently ν_maxwill be responsive to the

increasing donor ability of the anion (Cu---S << Cu---[X]^–).¹⁶

Thus, within DCE (a poorly coordinating solvent) ion-pair

formation is favoured and the strength of the counter-anion

coordination can be evaluated.¹⁹Soukup and Sone reported

λ_maxin DCE for the series of salts [Cu(acac)(tmen)][X] where

Materials and methods

Chemicals were purchased from Sigma-Aldrich, Apollo

Scientific, Acros Organic and Alfa Aesar. All solvents were

purified employing standard drying agents prior to use.

All ionic liquids were synthesised following established

procedures which are presented in the ESI, Section 1.

[Cu(acac)(tmen)][X] complexes were synthesised according to

the literature.^16,18The [Cu(sacsac)₂] complex was synthesised

according to the literature procedure for [Cu(acac)₂] using

dithioacetylacetone.^30,31Further synthesis details for both the

[Cu(acac)(tmen)][X] and [Cu(sacsac)₂] complexes can be found

in the ESI, Section 2.

[X]^-= [NO₃]^–, [ClO₄]^–, [BPh₄]^–.²³

Linert et al found a

relationship between increasing donor ability and lower ν_max

for a range of counter-anions in DCE.^15,19Fukuda et al

analysed ν_maxfor the [NO₃]^–and [ClO₄]^–Cu salts in various

organic solvents, concluding that a degree of ion-pairing

occurs in DCE and that closer association occurs with the

[NO₃]^–anion.¹⁸

Elemental analyses were carried out using a Carlo Erba CE

1108 Elemental Analyser at the London Metropolitan

University. Melting points were measured with

a Sanjo

Gallencamp apparatus. Mass spectroscopic analyses were

recorded under FAB conditions on a VG Autospec-Q Mass

Spectrometer. Infra-Red spectra were measured on a Perkin

Elmer Spectrum 100 FT-IR spectrometer with a diamond

Complications can potentially arise when the solvent is an IL;

as the IL anions are considered "effectively non-associating".

Now both the solvent-anion and counter-anion are "effectively

non-associating" and may have a comparable coordinating

ability. In this case, the weakly coordinating counter-ion could

have an impact on the experimental results. Classical “non-

coordinating” anions include [ClO₄]^-, [OTf]^-, [BF₄]^-, [PF₆]^-all of

which can coordinate to metal ions.²⁴Nevertheless, there are

numerous examples of spectra taken within ILs using

solvatochromic salts with the counter-anions [BPh₄]^-or [ClO₄]^-

where it has been assumed that the counter-anion is non-

coordinating^22,25,26and thus not interfering with the

coordination of the IL ions. In contrast, the presence of a

counter-ion to [Cu(acac)(tmen)]⁺has also been shown to

influence the ability of the IL ions to coordinate to the metal

centre.²²

crystal universal ATR sampling accessory.

The crystal

structures were solved using an OD XCalibur PX Ultra.

UV-Vis spectra were recorded on a Perkin Elmer Lambda 25

UV/Vis spectrometer, with the thermostabilised chamber set

at 298 K. For solid state measurements, dyes were dissolved in

DCM and applied to a microscope slide. Measurements were

carried out after DCM evaporation and sample drying in a

vacuum chamber.

²³Na NMR measurements of 0.2M NaClO₄solutions in ILs were

performed at 293K using a Bruker DRX-400 spectrometer.

NMR tubes with IL solutions contained a sealed DMSO-d₆

capillary as the lock solvent and an aqueous NaCl reference

capillary (3 mol/L in H₂O, δ

(²³Na) = 0 ppm). All other NMR

measurements were carried out using DMSO-d₆or CDCl₃on a

This study aims to investigate the coordinating ability of

“effectively non-coordinating” IL anions with solvatochromic

Cu-complexes and to study the effects of the surrounding IL

solvent on anion donor ability. In Section 1 of the results and

discussion we set out some of the potential differences

between traditional neutral solvents and ILs and we describe

the MO theory underpinning the observed UV-vis spectra of

the solvatochromic Cu-centre. In Section 2 we examine the

donor nature of the anion through [Cu(acac)(tmen)][X] where

[X]^–is a range of conventional weakly associating anions. We

study "isolated" [Cu(acac)(tmen)][X] in the gas-phase, in the

solid state and in the non-coordinating solvent DCE. Once a

reference state has been established we explore the impact of

the IL solvation environment in Section 3. The same anion is

employed for both the IL and the counter-anion, thus

eliminating any potential influence from the counter-anion.

This section presents our most novel and unanticipated

results. In Section 4 donor numbers (DN) of ILs are predicted

using [Cu(acac)(tmen)][X], employing a range of neutral Cu-

Bruker Avance 400 instrument.

In the computational studies each structure was fully

optimised, confirmed as a minimum via frequency analysis,

further detail can be found in the ESI Section 3. The NBO and

CHelpG partial charge distributions of the isolated IL anions

were determined at the B3LYP-D3/6-311+G(d,p) level in the

gas phase. A detailed MO analysis was carried out for the

[Cu(acac)(tmen)]⁺cation in the gas phase using the B3LYP

functional in conjunction with aug-cc-pVTZ basis sets. The Cu-

atom was modelled using a scalar relativistic Stuttgart PP

(ECP10MWB) and the associated basis set. The

[Cu(acac)(tmen)]⁺cation was further examined employing the

M06, M06-L and M06-HF density functionals in conjunction

with aug-cc-pVDZ basis sets. TD-DFT determined electronic

excitations (UV-Vis) of the copper complexes in the gas and

solution phase (IL [C₄C₁im][PF₆]) were computed and model

UV-Vis spectra produced. Four selected copper complexes

with alternative ligands were further examined using DFT(M06

and M06-L) in the gas-phase and partial charges were

determined using NBO and CHelpG methods.

complexes and via

27–29

spectroscopy.

a

method based on ²³Na NMR

Each sub-section contains brief

a

concluding paragraph outlining the main results.

Results and Discussion

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 3

Please do not adjust margins

Dalton Transactions

Page 4 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

1. Factors to consider in this study

Increased complexity introduced by the IL solvent

In a traditional neutral solvent there is only one constituent -

the solvent molecule, however an IL is composed of two

species, the cation [C]⁺and anion [A]^–, and a complex set of

equilibria can occur, Figure 3. The cation and anion are

generally assumed to be non-associating, however ion-pairs

[C]⁺[A]^–can briefly form and rapidly dissociate. Within the IL

solvent the TM-complex [ML_n]⁺can remain associated with its

counter anion [X]^–forming the salt ion-pair [ML_n]⁺[X]^–(Figure

3a). The metal complex and counter-anion can be solvated as

an ion pair (Figure 3b) or they can dissociate into individually

solvated ions (Figure 3c). Moreover, interactions with the IL

ions can be sufficiently strong that new ion-pairings are

formed; [ML_n]⁺[A]^–and [C]⁺[X]^–(Figure 3d). Because the IL ions

are charged, interactions between the IL cation and counter-

anion ([C]⁺[X]^–) are stronger, and are likely to make more of an

impact in terms of the thermodynamic free energies driving

association/dissociation. In addition, an interaction between

the IL cation and the Cu-complex counter-anion may impact on

[ML_n]⁺[X]^–association, this may be through solvation [ML_n]⁺[X]^–

---[C]⁺or if strong, through the formation of a charged cluster,

([ML_n][X][C])⁺(Figure 3e). Within a traditional neutral solvent

it has been assumed that a neutral association complex with a

single counter-anion [ML_n]⁺[X]^–would be favoured, however,

when the solvent is an IL then stabilisation of an overall

anionic or overall cationic complex is more likely; primarily

([ML_n][A]₂])^–or ([ML_n][A][X])^–(Figure 3e). Therefore, in

probing individual anion interactions with the TM-centre (in a

non-coordinating environment such as the gas phase, solid

state or DCE) we explore the Cu-[X] interactions. In probing

the solvent effects on a probe Cu-complex we explore the IL

behaving as a solvent.

complex) a large energy gap separating the fully occupied and

partially occupied highest occupied MOs, leading to a large

ν_maxfor the primary peak in the absorbance spectra. In a real

system ligand orbital effects are prominent and the MO

diagram is significantly more complex, see Figure 4. To

expedite interpretation, symmetry labels have been given but

these are only approximate and relate back to the original CFT

labels (and not the actual symmetry or axial orientation of the

molecule). While the overall structure is square planar, this

complex is not D_4hsymmetric due to the presence of two

different ligands. The ligands contribute to the highest

occupied MOs substantially more than is normally assumed, in

addition ligand-only MOs interleave within the standard

(assumed) dAO manifold.

Moreover, the different

electronegativity of the interacting N and O atom orbitals

results in a miss-alignment in the energy of the ligand orbitals

and unsymmetrical ligand contributions to significant dAOs;

both of these effects change the dAO splitting pattern.

Disentangling all of these effects and attempting to relate the

actual MOs back to the simplified picture is difficult, further

details on constructing and interpreting the MO diagram of

Figure 4 are provided in the ESI Section 4.1.

IL anions as counter-anions to [Cu(acac)(tmen)]⁺

In the presence of a non-coordinating counter-anion and in the

absence of an interacting solvent, the [Cu(acac)(tmen)]⁺cation

has a square planar coordination geometry. In this section we

establish the electronic structure of the isolated TM-complex

and confirm the electronic transitions associated with the UV-

Vis spectra. In the presence of a coordinating counter-anion

the structure of the complex will change, depending principally

on the donor ability, charge density and shape of the anion,

Figure 2. Subsequently, we examine the experimental UV-Vis

spectra for [Cu(acac)(tmen)]⁺evaluated with various counter-

anions in a non-interacting solvent DCE and in the solid state.

We examine the X-ray structures of complexes formed with

different IL anion counter-anions, and consider the donor

ability of these ions based on a computed partial charge

analysis.

TD-DFT calculations can provide some insight into the principal

Isolated [Cu(acac)(tmen)]⁺cation electronic structure

Schemes like those shown in Figure 2 are very useful in

building up a general impression of the origin of UV-Vis

spectral features. Based on the highly simplified diagram

shown in Figure 2 it is typical to assume (for a square planar

4 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Dalton Transactions

Page 5 of 16

Journal Name

transitions, details of the calculations can be found in the ESI

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Table 1 UV-Vis absorption maxima of the [Cu(acac)(tmen)][X] complexes in the

Section 4.2. We tested the functional employed (B3LYP and solid state and in DCE solution, literature values in brackets.^14,19

M06), the amount of exchange included (M06-HF 100%, M06

Anion

Solid state

_max, /10³cm^-1

DCE solution

_max, /10³cm^-1

27% and M06L 0%) ESI Figure S4.2.1a, and the local

environment (gas-phase or an ionic liquid [C₄C₁im][PF₆]), ESI

Figure S4.2.1b.

ν

Cl^–

13.8 (14.0)

14.9 (14.8)

16.3 (16.5)

17.3 (17.2)

18.6 (18.6)

13.8 (14.0)

14.8 (14.7)

16.3 (16.5)

17.1 (17.2)

18.6 (18.6)

The lowest energy excitation (S1) is a HOMO to LUMO

transition Lπ

*-> dx²-y²and is formally forbidden. The formally

[SCN]^–

[NO₃]^–

[OTf]^–

[ClO₄]^–

allowed principal transitions are identified as those with a

large Frank-Condon factor (ESI Tables S4.2.1-2) the most

significant (allowed) transition is from a nominal dz²type MO

to the dx²-y²type MO. Two other transitions had appreciable

Frank-Condon factors, one is a lower energy ligand to metal

charge transfer and the other is a higher energy ligand to

ligand excitation. However, all of the allowed transitions occur

outside of the visible region of the spectrum. Thus, the

allowed transitions are not giving rise to the spectroscopic

features examined through experimental UV-Vis spectroscopy.

d-d transitions such as those giving rise to the observed colour

of this complex are formally forbidden, and are driven by

vibronic coupling. We can examine the DFT computed

(formally forbidden) transitions to find those experimentally

[NTf₂]^–

[PF₆]^–

18.9 (n.a.)

19.9 (20.0)

18.6 (n.a.)

19.6 (19.6)

Our values for ν_maxin DCE are consistent, within experimental

error (±0.1x10³cm^-1), with those of Linert et. al., except for

[NO₃]^–where our value is 16.3x10³cm^-1and Linert et. al.

obtained 16.5x10³cm^-1.¹⁷The ν_maxvalues provide insight into

the interaction strength of the counter-anion.

On anion association [Cu(acac)(tmen)]⁺forms a pseudo-square

based pyramidal geometry, the stronger the donor ability of

the anion the stronger the geometric distortion and the

smaller the energy gap for the lowest energy d-d transition

and thus the lower ν_max. Based on the ν_maxreported in Table 1

relevant. The key visible region dz²

→

dx²-y²transition for the

the coordination ability of these counter-anions is consistent

with the Kamlet-Taft β values for the corresponding ILs with

imidazolium cations.^32–36(ESI Table 5.1) and decreases as

follows:

different functionals (M06, B3LYP and M06L) in the gas-phase,

and in the least interacting IL [C₄C₁im][PF₆] are listed in the ESI,

Table S4.2.3. This is a transition from MO 65 to MO 69 in

Figure 4, larger images of these MOs are also depicted in ESI

Figure S4.2.2.

Cl^–> [SCN]^–> [NO₃]^–> [OTf]^–> [ClO₄]^–> [NTf₂]^–> [PF₆]^–

β values: 0.87 > 0.71 ~ 0.74 > 0.46 > 0.41 > 0.24 > 0.21

The experimental wavelength of the peak maxima for the

[Cu(acac)(tmen)]⁺cation in [C₄C₁im][PF₆] is at 514nm and

coincides well with the B3LYP gas-phase computed wavelength

The slightly different solvatochromic responses observed for

[Cu(acac)(tmen)][X] in the solid state and solvated in DCE

indicate that weakly coordinating IL-anions possess some

associating ability in a non-associating solvent. The Cl^–, [NO₃]^–

and [ClO₄]^–complexes show the same absorbance maxima

both in the solid state and in DCE solution. For [SCN]^–the

difference between the solid state and DCE values is within

experimental error and for [OTf]^–there is very little difference

for the S2 transition (dz²

→

dx²-y²) which occurs at 506nm.

2. Copper(II) complexes in 1,2-dichloroethane (DCE) and the solid

state

The Cu complexes [Cu(acac)(tmen)][X] undergo solvatochromic

changes in UV-Vis spectra as a result of counter ion or

solvation molecule association, i.e. we are exploring the left

hand side of Figure 2. An estimate of the "pure" anion-Cu-

complex interaction is difficult to obtain, for example, solid

state crystal structures may show some variation due to crystal

packing effects, and in solution the ability of the counter-anion

[X]^–to coordinate to the copper centre is in competition with

the solvent. Nevertheless, some solvents such as DCE are

considered to be very poorly associating (DCE has a donor

number DN=0) and thus measurements are assumed to result

from the association of [X]^–only. This may not be an accurate

assumption, for example Fukuda et al has concluded that a

degree of ion-pairing occurs in DCE.¹⁸

between ν_maxin the solid state and DCE spectra (ꢀ

ν_max=0.2 10³

cm^-1). This similarity implies that similar geometries are

present in the solid state and in DCE.

Cl^–is the smallest anion with the largest Kamlet-Taft β value of

those studied here. The chloride anion has all the charge

localised on a single atom and thus Cl^–can bind strongly to the

copper centre.³⁷The crystal structure of [Cu(acac)(tmen)]Cl is

known and demonstrates a square-based pyramidal geometry

with no intermolecular interactions.³⁷[SCN]^–is a larger anion

with the charge distributed over a larger volume. [SCN]^–can

bind to the metal centre through either of the S or N atoms,

Cu(II) is an intermediate acidic metal and it tends to bind

strongly via the nitrogen atom of the [SCN]^–which is an

intermediate ligand. Consequently one could predict that

[SCN]^-, coordinates to the copper centre with a weaker

coordination than observed for Cl^–, which has been confirmed

here.

The values of maximum absorption wavenumber (ν_max) for

[Cu(acac)(tmen)]⁺with the counter-anions [X]^–=Cl^-, [NO₃]^-,

[SCN]^-, [OTf]^-, [ClO₄]^–, [NTf₂]^-and [PF₆]^-in the solid state and in

DCE solution are presented in Table 2.

Corresponding literature data for DCE have also been noted

where available.^14,19

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 5

Please do not adjust margins

Dalton Transactions

Page 6 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

[NO₃]^–has a Kamlet-Taft β value very close to that of [SCN]^–

but greater than [OTf]^–(primarily due to a stronger

based

interaction).³⁸However, [NO₃]^–has been reported as non-

coordinating in the crystal structure of [Cu(2,4-diacetyl-5-

hydroxy-5-methyl-3-phenylcyclohexane)(tmen)][NO₃],

and more diffuse charge indicate a reduced interaction

Cl^–.

π

compared

to

The

crystal

structure

of

[Cu(acac)(tmen)][OTf] (Figure 6) contains two independent

cation:anion pairs, the dominant feature is chains of

alternating cations and anions with distinctly asymmetric

Cu···O linkages one shorter and one longer; for Cu1A

resulting in a square planar Cu-complex.³⁹Our attempts to

produce [Cu(acac)(tmen)][NO₃] crystals under anhydrous

conditions yielded crystals of poor quality. Diffraction quality

crystals could only be formed when one molecule of water was

present. Two different crystalline polymorphs were obtained.

One crystal structure contained two independent cation:anion

pairs forming an extended 1-dimensional ribbon, Figure 5. The

ribbon is comprised of a main thread of Cu1B-based cations

interacting with two oxygen atoms from different nitrate

anions, Cu1B–O distances are similar, being 2.7960(12) and

2.8385(14) Å. The Cu1A-based cation forms a terminal side-

chain and interacts only with one oxygen atom of the nitrate

anion, but with a much shorter Cu1A–O distance of 2.4495(11)

Å. Thus the Cu-cations interact primarily with oxygen atoms

and not with the anion as a whole, moreover there are two

types of Cu-anion complex present; Cu1A is 5-coordinate with

a square based pyramid geometry whilst Cu1B is octahedral

and 6-coordinate. In this crystal structure a single [NO₃]^–anion

is able to "neutralize" two Cu-cations. The spectroscopic

response observed from the solid state must be an average of

the response from two distinct Cu- centres, the square

pyramidal and octahedral structures.

2.4395(16) and 3.226(2)

Å respectively, and for Cu1B

2.3510(15) and 4.152(2) Å respectively. The short atomic

distances are considerably shorter than the sum of the van der

Waals radii of the copper and oxygen atoms (2.92 Å), strongly

indicating "coordination" of one oxygen atom to each Cu-

centre. These bond distances are also similar to the Cu-Cl

bond distance in [Cu(acac)(tmen)Cl] reported by Jalilehvand et.

al.³⁷(Cu-Cl = 2.50 Å). The other Cu···O linkage is considerably

longer and indicates there is a second "associated" [OTf]^–.

Overall, we conclude [Cu(acac)(tmen)][OTf] has a square based

pyramidal geometry, and a pseudo-octahedral coordination

when the longer distances are considered. While [OTf]^–is a

larger more diffuse anion than Cl^–it appears that interactions,

rather than being with the [OTf]^–anion as a whole are

primarily between the Cu-centre and the oxygen atoms.

A

single [OTf]^–anion both "coordinates" one Cu-centre, and

"solvates" another Cu-centre. The spectroscopic response

observed from the solid state must be an average of these

strong and weak interactions at the same Cu-centre.

[ClO₄]^–is difficult to place within the anion series, this anion

can be thought of as a structurally expanded Cl^–and known

β-

values of [C₂C₁im][ClO₄] and [C₄C₁im][ClO₄] ILs (0.46 and 0.41

respectively)^36,40indicate a weaker or similar donor ability

compared to [OTf]^–. However, despite of intermediate

β-value

The other nitrate anion is involved in

a self-contained

hydrogen bonding network with water molecules, (ESI Figure

S5.1). It is surprising that the second nitrate prefers to H-bond

with water, indicating that stabilisation from H-bonding

exceeds that from Coulomb interactions. Should we be

[ClO₄]^–is often treated as a weakly coordinating anion

producing a rather weak square based pyramidal structure.⁴¹

If [ClO₄]^–is weakly coordinating then, based on our

spectroscopic data, both [NTf₂]^–and [PF₆]^–should be

considered as weakly coordinating. It has been assumed in the

literature that the charge distribution in [NTf₂]^–is delocalised

over S-N-S and, combined with the large size of [NTf₂]^–, this

anion is expected to exhibit only a very weak interaction.^22,40

[PF₆]^–is highly symmetrical and is assumed to exhibit a high

degree of charge delocalisation over all of the fluoride atoms,

generating a diffuse sphere of charge; as a result, [PF₆]^–is

expected to exhibit the weakest coordinating ability of all the

considering this more as

a

[CuL₂-NO₃-CuL₂]⁺complex

interacting with [NO₃-H₂O]^–? The other crystal structure

obtained exhibited water coordination at the Cu-centre with

the nitrate anions forming a H-bond cluster with additional

water. Given the difficulty in obtaining a water free structure,

spectra from of Cu-complexes with [NO₃]^–are, unless water

has been rigorously excluded, likely to represent a water

rather than nitrate based response.

Compared to Cl^–, [OTf]^–has the negative charge primarily

(70%) localised over the SO₃group,³⁸the larger molecular size

6 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Dalton Transactions

Page 7 of 16

View Article Online

DOI: 10.1039/C7DT02372C

Journal Name

ARTICLE

interaction ability, however, the order of [NTf₂]^–and [ClO₄]^–is

reversed, indicating that based purely on the partial charge of

the oxygen atoms, [ClO₄]^–is less associating than [NTf₂]^–.

[ClO₄]^–is substantially less associating than Cl^–; [ClO₄]^–is larger,

the charge is more diffuse, and while the partial charge on Cl^–

is necessarily 1e the partial charge on the individual oxygen

atoms of [ClO₄]^–is much less (0.59e). If [ClO₄]^–is to be

considered as "effectively non-coordinating" then any ion with

a charge less than 0.59e on the primary interacting atoms can

also be considered as “effectively non-coordinating”.

The [NTf₂]^–anion can coordinate to a metal centre via its

nitrogen atom acting as a Lewis base, via the oxygen atoms

donating electron pairs to electrophilic metal centres or as a

bidentate ligand via either an oxygen–oxygen or oxygen–

nitrogen bonding mode.⁴²Based on the solid state structure

we could expect [NTf₂]^–to coordinate primarily through the O-

atoms, however based on the partial charges we would expect

[NTf₂]^–to coordinate through the N-atom. Computational

studies have indicated that the N-atom is the primary

interaction site in [NTf₂]^–,⁴³but this interaction is not evident

in the present crystal structure. However, weak coordination

of Pd(II) complexes with the N-atom of [NTf₂]^–has been

observed via XRD⁴²and mostly N-atom coordination has been

reported for TM complexes in [C₂C₁im][NTf₂].⁴⁴In solution

more conformational variation is possible; thus, the enhanced

coordination response for [NTf₂]^–could be due to the

existence of a N-atom coordination mode in the liquid state.

Thus, the ChelpG charges support the stronger association

implicated by the DCE solution studies, where coordination via

the N-atom is possible. The real situation is likely to be a

combination of effects, a reduction in secondary interactions

which weaken the solid state association and an additional

coordination mode which strengthens the liquid state

association, both leading to an enhanced coordination

response for [NTf₂]^–in DCE.

studied anions. The ꢀν_maxfor [ClO₄]^–and [NTf₂]^–are very

similar, while that of [PF₆]^–is greater, indicating [PF₆]^–is less

associating than [NTf₂]^–.

Figure

7

shows part of the crystal structure of

[Cu(acac)(tmen)][NTf₂]. This crystal structure is very similar to

the [OTf]^–analogue, and a dominant feature is the formation

of a chain of alternating cations and anions. The Cu···O

linkages are also distinctly asymmetric, being 2.5354(18) and

3.614(2) Å. The short Cu-O linkages in [Cu(acac)(tmen)][NTf₂]

complex are only slightly longer (≈0.1Å) than the axial Cu-O

interaction encountered in the [OTf]^-complexes (2.35-2.44Å).

This is in contrast to the weaker coordination expected from

[NTf₂]^–relative to [OTf]^–. The longer interaction for [NTf₂]^–lies

between those found for the crystalographically distinct [OTf]^–

anions; compare 3.61Å for [NTf₂]^–with 3.23 and 4.15 Å for

[OTf]^–. The [NTf₂]^–also forms secondary interactions; CF₃

…

F₃C interactions and O-atoms and the central N-atom all

interact with H-atoms of the Cu-complex ligands. We conclude

that the Cu1 centre has a square based pyramid geometry and

weak axial interactions which indicate a pseudo-octahedral

coordination. Similar to [OTf]^–a single [NTf₂]^–anion

"coordinates" one Cu-centre, while "solvating" another.

For the [NTf₂]^–and [PF₆]^–complexes ν_maxdecreases in DCE

solution compared to the solid state, this implies a closer

association of the anion and the Cu-centre in DCE, or repulsive

crystal packing effects reducing the association in the solid

state. It was noted above that secondary H-bonding

interactions and fluorous domains formed for [NTf₂]^–; these

could be weakening the primary Cu-O [NTf₂]^–interaction in the

solid state. Alternatively, the association in DCE may be

stronger.

The NBO charges represent electron density localisation

around atoms, and do not represent the electrostatic

potential. Thus, although the results of a ChelpG and NBO

analysis are both called "charges" they measure different

quantities. NBO charges better reflect a build-up of electron

density around specific atoms, MO contributions and the

effects of relative electronegativity and lone pair localisation.

Based on NBO charges the following order is obtained:

Cl^–> [OTf]^–> [NTf₂]^–> [ClO₄]^–> [PF₆]^–> [SCN]^–> [NO₃]^–

Thus, the NBO charges do not match with the UV-Vis spectra

β

-value based ordering of the anions. Both [OTf]^–and

Thus, it was found that these weakly coordinating IL-anions do

possess some associating ability in the non-associating solvent

DCE.

and

[NTf₂]^–appear as much stronger donors than expected based

on electrostatic arguments alone. [NO₃]^–appears low in the

NBO series, based on their electron density the O-atoms in

[NO₃]^–have the lowest localised charge, likely due to the lack

of polarity in the N-O bond (Pauling electronegativity of

N=3.04, O=3.44), meaning that electron density in [NO₃]^–

cannot easily accumulate on the O-atoms. In contrast, [OTf]^–

and [NTf₂]^–appear high in the NBO series, S-O bonds are highly

polar (Pauling electronegativity of S=2.58) and the O-atoms

easily accumulate electron density pulling it away from the S-

atoms.

Computational analysis of the counter-anions and copper(II)

complex

Population analysis was carried out for the isolated anions (see

ESI Section 3 for computational details). Key interacting atoms

within the anion are identified based on the solid-state

structures and the partial charges for these atoms are given in

Table 2. The Coulomb based (ChelpG) charges are mostly

consistent with the spectroscopically determined order of

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 7

Please do not adjust margins

Dalton Transactions

Page 8 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

The NBO charges represent the local electron density and

Journal Name

Table 3 NBO and CHelpG partial charges and gas phase basicity values of the

"soft" interactions, in contrast electrostatic potential derived anions.

charges represent "hard" interactions and a more global

picture taking into account the full distribution of charges

surrounding each atom. We can expect the hard Coulomb

characteristics to dominate here because the associating

anions are not covalently bind to the Cu centre and because

[Cu(acac)(tmen)]⁺is a charged cation. This conclusion is

supported by the much better match between the CHelpG

charge ordering and the spectroscopic results.

Anion

principal

atom

Cl*

NBO

CHelpG Anion gas phase

basicity, kJ/mol

Cl^–

-1.000

-0.466

-0.587

-0.560

-0.971

-0.372

-0.839

-1.199

-1.195

-0.912

-0.911

-0.352

-0.350

-0.592

-1.000

-0.733

-0.793

-0.720

-0.696

-0.218

-0.586

-0.777

-0.761

-0.611

-0.594

-0.174

-0.217

-0.452

1391⁴⁶

[SCN]^–

[NO₃]^–

[OTf]^–

[ClO₄]^–

[NTf₂]^–

S

-

N*

O*

1330⁴⁵

1225⁴⁵

1233⁴⁵

O*

F

Gas phase basicity is the change in Gibbs free energy upon

deprotonation of the [anion]H molecule in the gas phase, and

provides a measure of the stability of the anion (in the gas

phase) relative to the stability of the neutral acid. The results

reported in Table 2 are a mixture of experimental and scaled

G2(MP2) and scaled B3LYP/6-311+G(d,p) data.^45–47

O*

N cis

N trans

O* cis

O* trans

F cis

F trans

F*

[NTf₂]^–

[PF₆]^–

1218⁴⁷

Table 2 NBO and CHelpG partial charges and gas phase basicity

values of the anions.

1182⁴⁵

The order of gas phase basicities can provide an approximate

order for the associating ability of the anions, and in this case

the [ClO₄]^–appears higher in the series (compared to that

Spectra of a positive TM-complex with a "non-associating"

counter-anion within an IL will therefore not reliably reflect the

donor ability of the IL; the counter-anion will be influential.

Nevertheless we can avoid this issue; matching the counter-

anion [X]^–of the probe complex with the anion of the ionic

liquid [A]^–will ensure there is no competition for coordination

at the Cu centre ([X]^–=[A]^–), and will result in a more accurate

based on the IL β-values), however the remaining anions keep

the same relative ordering as for the spectroscopic results.

Thus, we have found that hard Coulomb characteristics

dominate the association of anions to the Cu centre of the

cation, therefore CHelpG and anion gas phase basicity data can

provide insight into the associating abilities of the anions.

measurement of IL donor properties.

Solutions of

[Cu(acac)(tmen)][X] complexes in selected ILs [C₄C₁im][A], [A]^–

= [SCN]^–, [NO₃]^–, [OTf]^–, [NTf₂]^–and [PF₆]^–in addition to ILs

with different cations [C][NTf₂] where [C]⁺= [C₄C₁C₁im]⁺and

3. Copper(II) complexes in ILs: Solvent interactions

We have considered impact of the anion on the [C₄C₁pyrr]⁺have been investigated, Table

3 shows the

[Cu(acac)(tmen)][X] complexes in a "non-associating" solvent experimental ν_maxvalues. DCE values from Table 1 are

DCE and isolated in the solid state. Now we consider the reproduced as the "DCE" row in Table 3.

ability of [Cu(acac)(tmen)][X] complexes to report on the The higher values of

associating ability of these same anions as part of an IL solvent, general suggest that [PF₆]^–is the best "effectively non-

i.e. we are now exploring the right hand side of Figure 2.

associating" (the weakest coordinating) anion (among those

ν_maxfor the probe [Cu(acac)(tmen)][PF₆] in

The ILs studied here, have a relative dielectric constant (ε) studied here) for both traditional solvents and ILs. In each

which varies over a very limited range; between 12.9 and 15.3 case, when the [Cu(acac)(tmen)][PF₆] probe is employed, the IL

(ESI Table 5.1). Moreover, ε=10.36 for DCE at 25 ^oC ,⁴⁸which is anions have displaced [PF₆]^–and interacted with the Cu-centre.

very similar to the ε of the ILs examined. The similarity of ε for These results are consistent with [PF₆]^–being successfully

all the ILs and DCE means that general solvent polarity effects employed as a “non-coordinating” ion.¹⁴Moreover, these

should be very similar for all of the liquids studied here. Thus, results are consistent with the coordinating abilities of the

any differences observed are likely to derive from specific anions falling in the order: [SCN]^-> [OTf]^-> [NTf₂]^-> [PF₆]^-It has

interactions, such as H-bonding, anion-

π

and localised surface been reported in the literature that the ν_maxvalue is not

charge interactions between ions.

dependant on the cation of ionic liquid,^35,40which has been

Typically "non-associating" counter-anions [X]^–such as [BPh₄]^–confirmed here through the similarity in

ν_maxin [C₄C₁im][NTf₂],

and [ClO₄]^–are employed in evaluating the response of [C₄C₁C₁im][NTf₂] and [C₄C₁pyrr][NTf₂]. However, our results

different solvents, and it is expected that these counter anions have also shown that replacing an alkyl chain methyl group

will be replaced by the IL anion [A]^–and thus spectra will with OH^-forming [(OH)³C₃C₁im][NTf₂]

has significantly

ν

_maxvalue from 18.2 x10³to 16.8 x10³cm^-1using

reflect the coordinating ability of the IL anion in an IL solvent decreased the

environment. However, when studying ILs, both the counter- the anion-matched complex, and from 18.1 x10³to 17.7 x10³

anion [X]^–and the IL anion [A]^–are considered to be weakly cm^-1using [Cu(acac)(tmen)][ClO₄]. Similar but less dramatic

interacting, and thus there may be competition for (weak) changes in absorption wave number occurred when replacing

coordination at the Cu-centre.

the methyl group with

[(CN)³C₃C₁im][NTf₂].

.

a

CN^-functionality forming

8 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Dalton Transactions

Page 9 of 16

View Article Online

DOI: 10.1039/C7DT02372C

Journal Name

ARTICLE

Table 4 Experimental ν

max values for IL anion-matched [Cu(acac)(tmen)]⁺. A dash indicates a value not measured or reported. Literature values are reported for

νmax employing [BPh₄]^-.²²

ν

_max/10³cm^-1

[SCN]^-

[NO₃]^-

[OTf]^-

[NTF₂]^-

[ClO₄]^-

13.2

15.9

17.7

16.6

17.9

18.1

18.0

18.2

19.4

18.6

[PF₆]^-

14.3

-

[BPh₄]^-

14.3

-

[C₄C₁im][SCN]

[C₄C₁im][NO₃]

[(OH)³C₃C₁im][NTf₂]

-

16.3

-

16.8

-

17.2

17.0

-

16.6

-

[C₄C₁im][OTf]

[(CN)³C₃C₁im][NTf₂]

[C₄C₁im][NTf₂]

[C₄C₁C₁im][NTf₂]

[C₄C₁pyrr][NTf₂]

[C₄C₁im][PF₆]

DCE

-

17.4

18.2

18.3

18.6

-

18.1

-

18.3

-

18.6

19.4

19.6

-

19.4

20.0

14.8

16.3

17.1

18.6

anions [SCN]^–, [OTf]^–and [NO₃]^–have higher values of ν_max

measured with anion-matched complexes, compared to those

measured with [ClO₄]^-and [BPh₄]^-counter-anions. For

example, the value of ν_maxfor [Cu(acac)(tmen)][OTf] is 17.2

while that of [Cu(acac)(tmen)][BPh₄] is 16.6 (both complexes

are in [C₄C₁im][OTf]). This can be rationalised by noting that in

weakly associating ILs (and DCE) the larger more diffuse anion

is better solvated (Figure 8a) and thus only weakly interacts

with [Cu(acac)(tmen)]⁺. While in a higher

β-value IL (composed

of smaller charge dense anions) the more diffuse "effectively

non-associating" anion is not well solvated (Figure 8b), leading

to a stronger association with [Cu(acac)(tmen)]⁺.

In contrast, the more charge dense anion is better solvated,

particularly so if the anion is the anion of the IL, leading to a

weaker association with the Cu-complex. This rationalisation

is consistent with studies that have shown that many ILs form

ideal mixtures, but those that produce slight but important

deviations from ideality have very different anions.⁴⁹A general

conclusion therefore is that the anions [ClO₄]^-and [BPh₄]^-

cannot be considered as "effectively non-coordinating" in high

Thus, the IL cation does have an effect on [Cu(acac)(tmen)]⁺

and [X]^–association when functional groups on the cation have

a stronger coordinating ability than the anions present.

Nevertheless, there are a number of disconcerting results that

ran counter to our initial expectations; these are outlined in

the following. For example, the association with

[Cu(acac)(tmen)]⁺is weaker for the more strongly coordinating

IL-anion (as counter-anion) than for a traditional weakly

coordinating counter-anion. The more strongly coordinating

β

-value ILs, they are only “non-coordinating” in conventional

neutral solvents or low -value ILs. In a high -value IL it is

important to use an anion-matched probe complex.

β

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 9

Please do not adjust margins

Dalton Transactions

Page 10 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

There is a close similarity between the results obtained for

anion-matched probe Cu-complexes in [C₄C₁im][PF₆] and

[C₄C₁im][NTf₂] and literature ν_maxvalues obtained with “non-

coordinating” [ClO₄]^-and [BPh₄]^-complexes. For example, the

value of ν_maxfor [Cu(acac)(tmen)][PF₆], [Cu(acac)(tmen)][BPh₄]

and [Cu(acac)(tmen)][ClO₄] in [C₄C₁im][PF₆] is 19.4 in all cases.

These results suggest that matching of IL anions and counter-

anions has no effect on the coordination of copper centre for

anions and ILs of very low associating ability. While the

association ability for the counter-anion complexes ([ClO₄]^–,

[BPh₄]^–and [PF₆]^–) in [C₄C₁im][PF₆] is similar, there are

differences in other ILs, for example, a significant difference is

observed for the ν_maxof [Cu(acac)(tmen)][X] in [C₄C₁im][SCN]

between [X]^–=[PF₆]^–(14.3 10³cm^-1) or [ClO₄]^–(13.2 10³cm^-1).

These results indicate that there are anion effects when the

anion of the IL is more associating, thus the typical "non-

associating" anions should be treated with caution. We

conclude that weakly associating anions have the same

association ability in ILs of very low associating ability but vary

when the IL anion is more associating.

We expected to find stronger [Cu(acac)(tmen)][X] ion-pairing

in a neutral solvent DCE, than in the ILs. However, the ν_max

values obtained for [Cu(acac)(tmen)][X] in the anion-matched

[C₄C₁im][X] ILs for [NO₃]^–and [OTf]^–are identical (within

experimental error) to those in DCE. Moreover, ν_maxvalues for

[Cu(acac)(tmen)][X] are higher in DCE compared to the anion

matched IL for [SCN]^–, [NTf₂]^–and [PF₆]^–, indicating a weaker

Cu-coordination interaction in DCE. Thus, we have an

unexpected set of contra-indicators; (a) the more strongly

associating anions [NO₃]^–and [OTf]^–do not show a stronger

association in DCE, and (b) the interaction of [Cu(acac)(tmen)]⁺

and [X]^–is weaker in DCE for the remaining ions [SCN]^–, [NTf₂]^–

and [PF₆]^–. The apparent loss of systematic variations with

respect to the anion matched probe molecules in different ILs

can be rationalised as outlined below.

Here we need to remember that the anion-association with

the Cu-complex represents, in part, the competition provided

by the solvation of the counter-anion, Figure 9. Because the IL

ions are charged, interactions between the IL cation and

counter-anion are stronger, and likely to make more of an

impact, Figure 9b-d. While the competition [Cu]⁺---[X]^–vs

[Cu]⁺---[A]^–is eliminated when [X]^–=[A]^–, competition of the Cu

complex cation or the solvent cation for the IL anion [X]^–(or

[A]^–) is not eliminated, the competition between [Cu]⁺---[X]^–vs

[C]⁺---[X]^–is still active. Thus, if the interaction strength [Cu]⁺--

-[X]^–<< [C]⁺---[X]^–(Figure 9b) then spectra will reflect a non-

coordinated [Cu(acac)(tmen)]⁺complex. Alternatively if the

interaction strength [Cu]⁺---[X]^–>> [C]⁺---[X]^–(Figure 9a) then

spectra will reflect an association of the anion with the Cu

complex, [Cu(acac)(tmen)][X].

We have made an implicit assumption that, similarly to

traditional neutral solvents, the anion will associate with the

Cu-centre, taking on a five coordinate geometry and forming a

neutral salt (as established for the solid state and DCE

solution). However, in the IL there is a very large excess of the

anions in comparison to the concentration of the copper

complex; stoichiometry no longer constrains the coordination

to only one anion. Moreover, when the solvent is an IL,

stabilisation of an overall anionic or overall cationic complex is

favoured. Hence, it is possible for the copper to take on a

higher coordination number, Figure 9. The formation of an

anionic complex via the coordination of two (as opposed to

one) anion is now possible leading to a pseudo-octahedral

geometry ([ML_n][X]₂)^–, ([ML_n][A]₂)^–or ([ML_n][A][X])^–(Figure 9a

and 9e) analogous to the cationic complex [ML_nS₂]⁺expected

to form within a traditional neutral solvent.

This also connects to the unexpected result outlined above,

that the association with [Cu(acac)(tmen)]⁺is weaker for the

(stronger coordinating) IL-anions than for the weakly

coordinating counter-anion. Stronger solvation of the IL

counter-anion

by

the

IL

cation

weakens

the

[Cu(acac)(tmen)][X] interaction. Using the above example,

stronger solvation of [OTf]^–in [C₄C₁im][OTf] than [BPh₄]^–in

[C₄C₁im][OTf], results in [OTf]^–staying associated with the IL

cation, while [BPh₄]^–prefers to associate with the Cu-complex

cation.

We suspect a fine balance between competing solvation

effects is occurring and could be a rationalisation for the

observed variation in ν_maxvalues, i.e. that they do not

correlate with anion association ability between ILs and DCE.

Subtle influences are significant and appear to have essentially

scrambled any systematic correlation.

10 | J. Name., 2012, 00, 1-3

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Dalton Transactions

Page 11 of 16

Journal Name

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

.

Table 5 IL donor numbers predicted by linear best fit from the experimental ν_maxvs DN correlation plots, * is derived from the [PF₆] complex. DN for the solvent

19

.

calculated using Eqn. 4 (probe [Cu(acac)(tmen)][PF₆]) and Eq. 1 (probe [Cu(acac)(tmen)][BPh₄]), literature values taken for the anion DN in DCE

Linear fit DN equation

DN of solvent

from [PF₆]^–

DN of solvent

using matched [X]^–

DN of solvent DN of anion

from [BPh₄]^-

Ionic liquid

from

matched

anion

in DCE¹⁹

probe

[C₄C₁im][SCN]

[C₄C₁im][NO₃]

*

-

44.4

26.2

63.1

31.9

-3

(10 ν

max

0.08DN)

=

18.48

18.73

–

27.3

35.8

21.1

-3

(10 ν

max

0.09DN)

–

[C₄C₁im][OTf]

[C₄C₁im][NTf₂]

17.0

7.3

19.8

9.8

32.8

16.4

16.9

-

-3

(10 ν

max

0.10DN)

18.91

19.18

-3

[C₄C₁im][PF₆]

-2.0

0.7

6.2

8.9

-

(10

ν

max

0.11DN)

[C₄C₁C₁im][PF₆]

8.4

For example, differences of the absorption values for different (alternative) DNs for conventional molecular solvents.^19,23

complexes in the same solvent (comparing columns in Table 3) When different probes are employed different DNs are

vary more strongly than differences in ν_maxfor different determined, nevertheless, the two types of DN (Gutmann or

solvents with the complex (comparing rows in Table 3). This [Cu(acac)(tmen)]⁺based) correlate; ν_maxbased DN correlations

means that changing the probe creates a more substantial for different anions over a range of traditional solvents have

change in ν

_maxthan switching the IL. Overall, our conclusion is been studied.^18,23However, although [Cu(acac)(tmen)]⁺has

that results for a specific probe ([Cu(acac)(tmen)][X]) in been used with ILs, reports tend to be of absorption maxima

different solvents can be compared, and that different probes values without conversion to DNs.^3,11The apparent DNs of ILs

in the same solvent can be compared, but making systematic can be predicted using a standard linear best fit procedure, for

comparison across both domains (solvent and probe) is example, Eqn. 1 was found by measuring the ν_maxof

problematic. [Cu(acac)(tmen)][BPh₄] in a range of molecular solvents with

Thus, we have studied the coordination of copper complexes known DN (DCE, MeNO₂, MeCN, Me₂CO, H₂O, MeOH, EtOH,

in IL solvents, to avoid the influence of the counter-anion we DMF and DMSO, DNs range 0-29.8).¹⁹Subsequently measuring

have matched the anion of the TM complex with that of the IL. ν_maxof a new solvent with [Cu(acac)(tmen)][BPh₄] and entering

We have used the solvation abilities of DCE and IL to explain this value into Eqn. 1 will deliver a DN for that solvent

the unexpected coordination of [SCN]^-to the Cu centre. [PF₆]^–10^-3ν_max= 20 - 0.103DN_solv

was found to be the weakest coordinating anion among those

(for [BPh₄]^–)

Eqn. 1

studied. We concluded that [ClO₄]^-and [BPh₄]^-should not to be We have measured

ν

_maxof [Cu(acac)(tmen)][X] where X=[SCN]^-

considered as "effectively non-coordinating" in a high

β

-value , [NO₃]^–, [OTf]^–and [NTf₂]^–in a range of molecular solvents

IL. Finally, we have made a statement on applicability of with known DN (nitromethane, benzonitrile, acetonitrile, 1,4-

solvatochromic probes for comparison of coordination abilities dioxane, methanol, DMF , DMSO and pyridine, DNs range 2.7-

of anions in ILs.

33.1), ESI Section 6, Table S6.1.

4. Donor Numbers for IL solvents and for IL anions

A linear fit can be established for each probe complex

[Cu(acac)(tmen)][X], two representative correlation plots for

X=[SCN]^–and X=[PF₆]^–are shown in Figure 10, the full set are

presented in the ESI, Figure S6.1, equations are reported in

4.1 Donor Number of ILs Evaluated [Cu(acac)(tmen)]⁺

In this section we study the donor characteristics of each IL

using donor numbers, DN. According to Gutmann’s definition

the donor number, also known as “donicity”, is a quantitative

measure for the tendency of a solvent to donate electron pairs

to an acceptor.⁵⁰The original Gutmann's DNs were determined

from an association enthalpy with SbCl₅. Solvents can have a

DN, for example the DNs of H₂O, EtOH, MeCN and DMSO are

19.5, 19.0, 14.1 and 19.8 respectively.¹⁹However, individual

anions can also have DNs, for example the DNs of [Cl]^–, [SCN]^–,

[NO₃]^–, [OTf]^–, [ClO₄]^–and [BPh₄]^–are 36.2, 31.9, 21.1, 16.9, 8.4

and 0.0 in DCE respectively.¹⁹This raises an interesting

question which cannot arise when considering traditional

solvents; how well do the DN of IL solvents match the DNs of

the individual anions making up the IL?

Table 4. Subsequently ν_maxfor each [Cu(acac)(tmen)][X] probe

has been measured in the anion-matched [C₄C₁im][X] IL, and

the DN for the IL solvent determined, Table 4. The [SCN]^-salt

posed some problems, as the absorption maximum vs DN gave

an almost horizontal line in the linear fit with a gradient of

approximately zero (Figure 10). This may be due to [SCN]^-

coordinating strongly to the copper centre irrespective of the

solvent. Therefore, the linear fit established for the [PF₆]^-

complex (our reference non-associating anion) was used to

determine the DN of [C₄C₁im][SCN]. Comparing the equations

in Table 4, anions with a larger DN (i.e. a stronger association

with the Cu-complex) have a shallower (lower) slope and lower

intercept, however adjacent anions exhibit differences only

just outside of the error boundaries (intercept ±0.2*10³and

slope ±0.01*10³). Eqn 1 for [X]^–= [BPh₄]^–has a slope

DNs can be determined using other probe molecules, for

example [Cu(acac)(tmen)]⁺has been used to determine

This journal is © The Royal Society of Chemistry 20xx

J. Name., 2013, 00, 1-3 | 11

Please do not adjust margins

Dalton Transactions

Page 12 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

(0.103*10³) lying between that of [PF₆]^–and [NTf₂]^–, and a anion to the metal centre can be affected by the nature of the

similar intercept (20*10³) to [PF₆]^–. The key point is that each IL cation.⁵²Improvements in a Stille reaction have been

probe [Cu(acac)(tmen)][X] must be calibrated individually, and reported

where

the

nitrile

pendant

group

of

that a probe with a larger slope is more sensitive and hence [(CN)³C₃C₁im][NTf₂] or [(CN)³C₃C₁im][BF₄] interact with the

more desirable. Based on the slope of the ν_maxvs DN palladium centre.⁵¹More recently, Hardacre and co-workers

correlation plots [Cu(acac)(tmen)][PF₆] is the best probe reported that the basicity of several amine functionalised ILs

examined here.

was dependent on the amine tether to the cation and on the

As [Cu(acac)(tmen)][PF₆] is the best probe (of those examined choice of linker between the two nitrogen centres.⁵³Thus, the

here) we have evaluated ν_maxfor the following ILs: [C₄C₁im][X] difference in DN values established here, strengthens the

where [X]^–

=

[SCN]^–, [NO₃]^–, [OTf]^–, [NTf₂]^–, [PF₆]^–and proposition that hydroxyl and nitrile groups (on the

[C₄C₁C₁im][PF₆]. Subsequently Eqn. 2 has been employed to imidazolium cations) compete with the [NTf₂]^-anion to

determine the DN for each IL, Table 4. The donor numbers for coordinate to the copper centre.

these solvents exhibit a wide range, from -2.0 (or assumed 0) We calculated DNs of ILs using a Cu(II) complex with matching

for [C₄C₁im][PF₆] through to 44.4 for [C₄C₁im][SCN]. Negative anions, using a Cu(II) complex with the weakest coordination

DN have been obtained previously for [NTf₂]^–and [PF₆]^–anion defined in this study [PF₆]^-and a Cu(II) complex with

compared to a DCE reference and simply indicate that this IL classical “non-coordinating” [BPh₄]^-and concluded that the Cu-

(solvent) is a weaker donor than DCE.²⁵

complex counter-ion influenced the obtained result and must

be considered when taking the measurements.

10^-3ν_max= 19.18(±0.22) - 0.11(±0.01)DN_solv(for [PF₆]^–) Eqn. 2

4.2 Donor Number of ILs Evaluated with a Neutral Cu Complex

The [Cu(acac)(tmen)]⁺complexes used in this study are

positively charged and thus may coordinate IL anions more

strongly due to enhanced Columbic interactions. To reduce

the Coulombic interactions, three neutral complexes, based on

The variation in DN of the ILs reflects the coordinating abilities

of the anions, consistent with the donor number of ILs being

determined by the anionic component:

[SCN]^–> [NO₃]^–> [OTf]^–> [NTf₂]^–> [PF₆]^–.

DNs for a given IL predicted using [Cu(acac)(tmen)][PF₆] or a

matching anion complex are smaller than those obtained from

[Cu(acac)(tmen)][BPh₄], emphasising that the Cu-complex

counter-ion must be taken into account when attempting to

predict donor numbers. Because correlations have been

established based on conventional molecular solvents, the

predicted DNs now allow for a direct comparison between ILs

and more traditional solvents. On this basis, the weakest

(donor anion) IL solvent [C₄C₁im][PF₆] has a low DN similar

(within experimental error) to that of DCE (DN=0), while the

strongest (donor anion) IL solvent [C₄C₁im][SCN] has a high DN

(44.4) above that of pyridine (33.1) and approaches that of

water (54.8). Thus, like conventional solvents ILs can span a

full range of solvent donor environments. These results

illustrate the important point that IL anions should not always

be considered as "non-coordinating".

a

[Cu(acac)₂] core structure have been computationally

assessed. In addition to [Cu(acac)₂], the methyl R groups of

acac were replaced with CF₃forming [Cu(hfac)₂] and the

oxygen atoms of acac were replaced with sulphur forming a

dithioacetylacetone complex [Cu(sacsac)₂], Figure 11.

Structures were optimised employing the M06 and M06-L

functionals with an aug-cc-pVDZ basis set and charges were

evaluated using both orbital (NBO) and electrostatic potential

(CHelpG) based methods (ESI Section 7). The complex with the

smallest charge on the Cu centre [Cu(sacsac)₂] was synthesised

and used to evaluate donor numbers to compare with

[Cu(acac)(tmen)][X].

From an electron book-keeping

perspective, while these complexes are neutral, the formal

oxidation state of the Cu centre is still +2, and the Cu is still a

Lewis acid. Nevertheless, with a smaller partial charge on the

Cu centre, the Cu is expected to be less acidic, and the

complex (as a whole) less polarised.

Similar to previously reported [Cu(acac)₂] spectra,⁵⁴

[Cu(sacsac)₂] shows a complex absorption spectrum, spectra

were deconvoluted using a Gaussian fit procedure described in

It has been thought that the donor ability of an IL does not

depend on the cation, at least with respect to non-

functionalised ILs.

ν_maxhas been determined for

[Cu(acac)(tmen)][NTf₂] in [NTf₂] based ILs with non-

functionalised and with hydroxyl and nitrile functionalised

cations. DNs were computed using the equation reported for

[Cu(acac)(tmen)][NTf₂] in Table 4. DNs for [C₄C₁im][NTf₂],

[C₄C₁C₁im][NTf₂] and [C₄C₁pyrr][NTf₂] are close and equal to

7.3, 6.1 and 3.1, correspondingly. However, the DNs for

functionalised [(CN)³C₃C₁im][NTf₂] and [(HO)³C₃C₁im][NTf₂] are

15.1 and 21.1 respectively, which is very different from the DN

7.1 for their non-functionalised analogue. The Kamlet-Taft

hydrogen bond acceptor ability β value of these ILs is very

the ESI Section 8. The absorption maxima

(ν_max) for

[Cu(sacsac)₂] have been obtained for a range of molecular

solvents (DMSO, methanol, acetonitrile, chloroform and DCE),

ESI Table S8.1.

,51

similar spanning 0.22-0.29.⁴⁰

Thus, in contrast to the

Kamlet-Taft β value there are significant differences in DN for

hydroxyl and nitrile functionalised ILs compared to the related

non-functionalised ILs. Thus, the degree of coordination of the

12 | J. Name., 2012, 00, 1-3

Please do not adjust margins

Dalton Transactions

Page 13 of 16

View Article Online

DOI: 10.1039/C7DT02372C

Journal Name

ARTICLE

a

Table 6 Calculated DNs of molecular solvents and ILs employing [Cu(acac)(tmen)][BPh₄], anion matched [Cu(acac)(tmen)]X , [Cu(sacsac)₂] and ²³Na NMR. Gutmann's

19.

b

DN for the neutral solvents and [Cu(acac)(tmen)][BPh₄] values from ref

linear fit for molecular solvents²⁷

Literature spectroscopic DN for selected ILs from ref ²⁵. * value was calculated using the

[Cu(acac)(tmen)][BPh₄] [Cu(acac)(tmen)] [X]

Solvent

Other DN

29.8 ^a

19.1 ^a

14.1 ^a

0.0 ^a

DN

[Cu(sacsac)₂] DN ²³Na NMR DN

DMSO

34.0

28.2

24.3

0.0

-

27.1

23.7

17.4

0.2

32.7²⁷

25.3*

18.2²⁷

-

Methanol

Acetonitrile

DCE

-

[C₄C₁im][SCN]

[C₄C₁im][OTf]

[(HO)³C₃C₁im][NTf₂]

[(CN)³C₃C₁im][NTf₂]

7.86 ^b

1.57 ^b

-

63.1

32.8

22.0

18.7

16.4

6.2

44.4

17.0

21.1

15.1

7.3

-2.0

28.3

24.2

15.4

10.1

6.1

46.3

19.8

14.1

8.9

-

[C₄C₁im][NTf₂]

[C₄C₁im][PF₆]

-3.44 ^b

-4.21 ^b

8.4

-5.8

-7.6

Where possible a linear best fit correlation of ν_maxvs DN was our ²³Na NMR based DNs with those obtained employing

evaluated, Eqn 3 (ESI Table S8.2). This fit exhibits a very solvatochromic dyes, Table 6. Our ²³Na NMR based results

shallow slope, and thus its

ν_maxis not very sensitive to DN. The report slightly lower DN than those obtained using

indications are therefore that this probe is not sufficiently [Cu(acac)(tmen)][BPh₄] however, the same trend is obtained.

sensitive.

Nevertheless, the DNs obtained from ²³Na NMR data are

predominantly in-line with those obtained using the

Eqn. 3 [Cu(sacsac)₂] complex and matching anion complexes.

Alternatively, the low DN obtained from ²³Na NMR may have

10^-3ν_max= 22.4(±0.5) - 0.05(±0.03)DN_solv

Absorption maxima for [Cu(sacsac)₂] in the ILs ([C₄C₁im][X] [X]^–an experimental origin. Chemical shift values are established

= [SCN]^–, [OTf]^–, [NTf₂]^–and [PF₆]^–, and [(HO)³C₃C₁im][NTf₂]) at 293K while the UV-Vis values are evaluated at 298K. It is

were determined, and the DN for each IL evaluated, Table 5. A possible that the temperature decrease of 5K could lead to

negative DN is obtained for [C₄C₁im][PF₆] and simply indicates slightly lower DNs, however, Schmeisser and co-workers

that this IL (solvent) is a weaker donor than DCE.²⁵

suggest the change in DN will not exceed ≈1 unit for a

As expected with the reduced Coulombic attraction, the DNs difference of 5 K, which is insufficient to account for the full

obtained with [Cu(sacsac)₂] are less than those obtained shift.²⁷The concentration of a probe substance in UV-Vis

employing [Cu(acac)(tmen)][X]. Similar trends are obtained for spectroscopy is negligible compared to that required to obtain

both [Cu(sacsac)₂] and [Cu(acac)(tmen)][X] complexes.

a reasonable ²³Na NMR measurement. ²³Na⁺is introduced in

the system as a [ClO₄]^–salt (c=0.2 mol/L), [ClO₄]^-has a low

DN=8.4 which may reduce the overall DN of the system.

However, if adding the Na[ClO₄] salt had a significant effect,

then the DN for [C₄C₁im][PF₆] should increase, as the DN of

[ClO₄]^–is greater than that of [PF₆]^–, which is not the case.

Nevertheless, a shift dependent on the quantity of added

Na[ClO₄] salt cannot be excluded.

4.3 Donor Number of ILs evaluated using ²³Na NMR

DNs for ILs have been previously been evaluated using ²³Na

NMR spectroscopy, and a linear fit established by Schmeisser

et al.²⁷and reassessed by Gal et al., Eqn 4.²⁸We have assessed

²³Na chemical shifts for 0.2M NaClO in a range of ILs.

4

Chemical shifts were converted into DNs employing Eqn 4, and

Our ²³Na NMR based results have shown the same trend as

those obtained from UV-Vis solvatochromic probes, confirming

the reliability of both approaches.

the resultant DNs are reported in Table 6.

23

DN[kcal/mol]=33(±1) + 2.1(±0.2)*

δ(

Na)

Eqn. 4

The quantitative reliability of DN determination by ²³Na NMR

has been questioned previously.^27–29Thus, we have compared

4.4 Donor Number of ILs Compared

J. Name., 2013, 00, 1-3 | 13

Please do not adjust margins

Dalton Transactions

Page 14 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

Figure 12. presents the collected DN data for a range of non-

functionalised ILs, [C₄C₁im][X] where [X]^–= [SCN]^–, [OTf]^–,

[NTf₂]^–and [PF₆]^–. All of the methods for evaluating DN

preserve the same trend. These results confirm the

applicability of a range of spectroscopic methods to determine

the donating properties of the ILs relative to molecular

solvents. Employing the neutral [Cu(sacsac)₂] complex as a

solvatochromic probe provided viable results, consistent with

those obtained via ²³Na NMR. The DN values obtained here are

smaller

than

Gutmanns's

DNs

obtained

with

[Cu(acac)(tmen)][BPh₄] (black squares Figure. 12), but are

larger than the DNs obtained via other spectroscopic methods,

for example those of Holzweber et al,²⁵(green diamonds

Figure. 12). The distribution of the DN values narrows with a

weakening of the anion coordination ability. This has

consequences in the relative comparison of the donor ability

of ILs, for example using [Cu(sacsac)₂], [C₄C₁im][SCN] has ≈5

times the donor ability of [C₄C₁im][NTf₂], however using

[Cu(acac)(tmen)][BPh₄], [C₄C₁im][SCN] has ≈4 times the donor

ability of [C₄C₁im][NTf₂].

interacted with the Cu-centre. Structures that exhibited both

long and short Cu---O interactions for the same Cu centre were

found, or two distinct structures (square pyramidal and

octahedral) exhibiting different anion interaction strength

could exist. Thus, the observed spectroscopic response must

represent an average of different interaction modes, or

different structures. Structurally, [Cu(acac)(tmen)][NO₃] was

unusual in that H-bonding between [NO₃]^–and water was

preferred over association with the Cu-centre indicating a

Therefore the ILs with stronger coordinating anions have wide

spread of DN results, while IL with weakly coordinating anions

have less of the deviation in DN obtained with different

probes.

Conclusions

The computed B3LYP/6-311+G(d,p) electronic structure for the

isolated square planar [Cu(acac)(tmen)]⁺cation has been

evaluated. Analysis of the MOs of [Cu(acac)(tmen)]⁺cation

showed that the familiar scheme associated with tetragonal

distortion is substantially oversimplified. The ligands

contribute to the highest occupied MOs substantially more

than is normally assumed, in addition ligand-only MOs can

interleave within the standard (assumed) dAO manifold,

leading to the identification of L->M and L->L transitions.

Nevertheless, a primary transition from a nominal dz²MO to

the dx²-y²LUMO was identified using TD-DFT and confirmed by

comparison with experiment.

possible [CuL₂-NO₃-CuL₂]⁺

+

[NO₃-H₂O]^–conformation is

favoured over 2[CuL₂][NO₃]+H₂O, a water free crystal structure

could not be obtained. For the [NTf₂]^–and [PF₆]^–complexes

the anion and the Cu-centre exhibit a closer association in DCE

than in the solid state. [NTf₂]^–in the solid state does not

interact via the more strongly donating N-atom indicating

crystal packing effects may be active for this anion. We

rationalised that the enhanced coordination response for

[NTf₂]^–in DCE could be due to the existence of a N-atom

coordination mode in the liquid state. Anion ordering based

on the "hard" CHelpG charges, representing a picture of

distributed partial charges was consistent with that obtained

spectroscopically. The NBO charges, which reflect the softer

characteristics of bonding, through orbital effects and a local

build-up of electron density provided a very different anion

ordering.

Anion coordination interactions with [Cu(acac)(tmen)]⁺

complexes have been evaluated.

UV-Vis spectra for

[Cu(acac)(tmen)][X] where [X]^–= Cl^–, [NO₃]^–, [SCN]^–, [OTf]^–,

[NTf₂]^–and [PF₆]^–in the solid state and in a DCE solution have

been collected and ν_maxreported. X-ray crystallographic

analyses of [Cu(acac)(tmen)][OTf], [Cu(acac)(tmen)][NTf₂] and

After evaluating the associating ability of the individual anions

we examined the associating ability of ILs composed of the

same anions. Typically one adds [Cu(acac)(tmen)][X] to a

solvent to measure the extent of association, however when

both the counter-anion [X]^–and IL anion [A]^–are weakly

associating, they will complete for coordination at the Cu-

centre of [Cu(acac)(tmen)]⁺. ILs are unique, in that matching

the counter-anion [X]^–of the probe complex with the anion of

the ionic liquid [A]^–ensures there is no competition for

coordination at the Cu centre resulting in a more accurate

measurement of IL donor properties.

[Cu(acac)(tmen)][NO₃]•(H₂O)

have

been

reported.

Population analysis of the individual gas-phase anions, both

electrostatic (CHelpG) and electron density (NBO) have been

discussed.

The coordination ability of the ILs was found to be consistent

with the Kamlet-Taft β values for the corresponding ILs. The

solvatochromic response observed for [Cu(acac)(tmen)][X] in

the solid state and solvated in DCE were very similar, from

which, it was concluded that even classical "effectively non-

coordinating" IL-anions possess some associating ability in the

non-associating solvent DCE. For all the O-containing anions, it

was found that the O-atoms, rather than the anion as a whole

14 | J. Name., 2012, 00, 1-3

Please do not adjust margins

Dalton Transactions

Page 15 of 16

Journal Name

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

The solvent interactions of a set of [C₄C₁im][X] ILs where [X]^–the donating ability of [C₄C₁im][X] ILs in comparison to other

= Cl^–, [NO₃]^–, [SCN]^–, [OTf]^–, [NTf₂]^–and [PF₆]^–have been conventional solvents.

investigated using probes with matching anions: consistent with

DNs ranged from 44.4 to -2.0

a

decreasing coordination ability and

[Cu(acac)(tmen)][X]. We established that [PF₆]^–is the least demonstrating that ILs can span a wide donor range, similar in

associating anion (among those studied here) for both breadth to conventional solvents (water=54.8, DCE=0).

traditional solvents and ILs. In addition [C][NTf₂] ILs with [C]⁺= We designed a probe with reduced Coulombic interactions;

[C₄C₁C₁im]⁺, [C₄C₁pyrr]⁺, [(CN)³C₃C₁im]⁺and [(HO)³C₃C₁im]⁺calculations were employed to evaluate a range of neutral

have been compared to the base IL [C₄C₁im][NTf₂]. Evidence of complexes based on [Cu(acac)₂]. The complex with the

competition to coordinate to the Cu centre between -OH or - smallest charge on the Cu centre, [Cu(sacsac)₂], was

CN groups on the cation and a [NTf₂]^-anion for selected synthesised and used to evaluate DNs to compare with

functionalised ILs was found.

On closer examination

[Cu(acac)(tmen)][X]. DN values were very close to those

number of counter-intuitive obtained via ²³Na NMR, strengthening the viability of both

a

observations arose which have been rationalised. Comparing methods, and suggesting some of the sensitivity of

counter-anions for the same ILs, recorded ν_maxvalues [Cu(acac)(tmen)]⁺is due to the overall positive charge of the

suggested that anion association with [Cu(acac)(tmen)]⁺is complex.

weaker for the higher

the supposedly non-coordinating anions, e.g. [BPh₄]^–. In traditional "non-associating" anions are not non-coordinating.

higher -value ILs, the larger more diffuse anions are poorly The use of anion-matched solvatochromic probes in ILs is

β

-value IL-anions e.g. [SCN]^–, than for Overall this study has presented evidence indicating that

β

solvated leading to a stronger association with the Cu- pioneered. The applicability of spectroscopic methods to

complex, while the smaller charge dense ions are more determine the donor properties of ILs in comparison to more

effectively solvated leading to a weaker association with the conventional molecular solvents is established.

Cu-complex, thus giving rise to the counter intuitive results

obtained. A general conclusion is that the anions [ClO₄]^-and

[BPh₄]^-cannot be considered as "effectively non-coordinating"

Acknowledgements

in a high

β-value IL, in this case an anion-matched probe

We would like to acknowledge the ERC for an Advanced

Investigator Award, ‘Mixing Ionic Liquids’

should be employed.

Weakly associating anions have the same association ability in

ILs of very low associating ability but vary when the IL anion is

more associating. This was rationalised as an IL cation effect,

because ILs ions are charged there can be stronger counter-

anion solvent interactions. Where the ILs are more associating

a stronger competition occurs between the probe cation and IL

cation for the counter anion [Cu]⁺---[X]^–---[C]⁺. Results

obtained for typical "non-associating" anions should be

treated with caution, as they will also reflect cation association

ability. Further consideration of the potential equilibria set up

in an IL suggests it is possible for the Cu-centre to interact with

two anions, forming a number of singly negatively charged

species ([ML_n][X]₂)^–, ([ML_n][A]₂)^–or ([ML_n][A][X])^–. Comparing

the ν_maxvalues obtained for [Cu(acac)(tmen)][X] anion

matched probes with classical "non-coordinating" anion

probes in both the anion-matched ILs [C₄C₁im][X] and in DCE,

showed no systematic trends and we surmised that a delicate

balance between competing solvation effects have impacted

on the results. Overall, our conclusion is that results for a

specific probe ([Cu(acac)(tmen)][X]) in different solvents can

be compared, and that different probes in the same solvent

can be compared, but making systematic comparison across

both domains (solvent and probe) is problematic.

References

1

O. Kuzmina and J. P. Hallett, Application, Purification and

Recovery of Ionic Liquids, Elsevier, Amsterdam, 1st edn.,

2016.

2

3

T. Welton, Chem. Rev., 1999, 99, 2071.

X. Lu, G. Burrell, F. Separovic and C. Zhao, J. Phys. Chem. B,

2012, 116, 9160.

4

Y.-S. Liu, G.-B. Pan, K. a Venkatesan, T. G. Srinivasan and P.

R. V. Rao, in Ionic Liquids: Applications and perspectives,

ed. A. Kokorin, InTech Europe, Rijeka, 2009, vol. 10, pp. 1–

6.

5

6

7

O. Kuzmina, E. Bordes, J. Schmauck, P. A. Hunt, J. P. Hallett

and T. Welton, Phys. Chem. Chem. Phys., 2016, 18, 16161.

P. Wasserscheid, Transition Metal Catalysis in Ionic Liquids.

Handbook of Green Chemistry. Part 6. Ionic Liquids, 2010.

P. Wasserscheid and T. Welton, Ionic Liquids in Synthesis,

WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim, 2008.

M. R. Rosenthal, J. Chem. Educ., 1973, 50, 331.

I. Krossing and I. Raabe, Angew. Chem. Int. Ed. Engl., 2004,

43, 2066.

8

9

10

11

12

13

Y. Chauvin, L. Mussmann and H. Olivier, Angew. Chem. Int.

Ed. Engl., 1996, 34, 2698.

Correlation plots of ν_maxvs DN for [Cu(acac)(tmen)][X] where

X=[SCN]^–, [NO₃]^–, [OTf]^–and [NTf₂]^–in a range of molecular

solvents with known DN have been established. Based on the

highest value of ν_max[PF₆]^–can be considered the least

coordinating anion, and based on the largest slope for the ν_max

vs DN correlation plots, [Cu(acac)(tmen)][PF₆] was established

as most sensitive probe (out of the anions studied here).

Subsequently, we employed [Cu(acac)(tmen)][PF₆] to evaluate

O. A. El Seoud, A. Koschella, L. C. Fidale, S. Dorn and T.

Heinze, Biomacromolecules, 2007, 8, 2629.

A. Abbott, G. Frisch, H. Garrett and J. Hartley, Chem.

Commun., 2011, 47, 11876.

M. Y. Lui, L. Crowhurst, J. P. Hallett, P. A. Hunt, H.

Niedermeyer and T. Welton, Chem. Sci., 2011, 2, 1491.

J. Name., 2013, 00, 1-3 | 15

Please do not adjust margins

Dalton Transactions

Page 16 of 16

View Article Online

DOI: 10.1039/C7DT02372C

ARTICLE

Journal Name

14

15

16

17

18

19

S. I. Noro, N. Yanai, S. Kitagawa, T. Akutagawa and T.

Phys. Chem. B, 2005, 109, 16474.

D. B. Williams, M. E. Stoll, B. L. Scott, A. Costa and W. J.

Nakamura, Inorg. Chem., 2008, 47, 7360.

W. Linert, Y. Fukuda and A. Camard, Coord. Chem. Rev.,

2001, 218, 113.

44

45

46

47

Oldham, Chem. Commun., 2005, 613, 1438.

I. A. Koppel, P. Burk, I. Koppel, I. Leito, T. Sonoda and M.

Mishima, J. Am. Chem. Soc., 2000, 122, 5114.

J. D. D. Martin and J. W. Hepburn, J. Chem. Phys., 1998,

109, 8139.

Y. Fukuda and K. Sone, Bull. Chem. Soc. Jpn., 1972, 45, 465–

469.

Y. Fukuda and K. Sone, J. Inorg. Nucl. Chem., 1972, 34,

2315.

I. Koppel, J. Koppel, P.-C. Maria, J.-F. Gal, R. Notario, V. M.

Vlasov and R. W. Taft, Int. J. Mass Spectrom. Ion Process.,

1998, 175, 61.

Y. Fukuda, A. Shimura, M. Mukaida, E. Fujita and K. Sone, J.

Inorg. Nucl. Chem., 1974, 36, 1265.

W. Linert, R. F. Jameson and A. Taha, J. Chem. Soc. Dalt.

Trans., 1993, 3181.

48

49

A. A. Maryott and E. R. Smith, Natl. Bur. Stand. Circ. 514,

1951, 1–40.

20

21

22

M. A. Halcrow, Chem. Soc. Rev., 2013, 42, 1784.

J. Meinershagen and T. Bein, Adv. Mater., 2001, 13, 208.

M. J. Muldoon, C. M. Gordon and I. R. Dunkin, J. Chem. Soc.

Perkin Trans. 2, 2001, 433.

M. T. Clough, C. R. Crick, J. Gräsvik, P. a. Hunt, H.

Niedermeyer, T. Welton and O. P. Whitaker, Chem. Sci.,

2015, 6, 1101.

50

51

V. Gutmann, Coord. Chem. Rev., 1976, 18, 225.

C. Chiappe, D. Pieraccini, D. Zhao, Z. Fei and P. J. Dyson,

Adv. Synth. Catal., 2006, 348, 68.

23

R. W. Soukup and K. Sone, Bull. Chem. Soc. Jpn., 1987,

2286.

24

25

S. H. Strauss, Chem. Rev., 1993, 93, 927.

M. Holzweber, R. Lungwitz, D. Doerfler, S. Spange, M. Koel,

H. Hutter and W. Linert, Chem. - A Eur. J., 2013, 19, 288.

J. Bartosik and A.-V. Mudring, Phys. Chem. Chem. Phys.,

2010, 12, 4005.

52

53

C. Chiappe, G. Imperato, E. Napolitano and D. Pieraccini,

Green Chem., 2004, 6, 33.

S. A. Forsyth, U. Fröhlich, P. Goodrich, H. Q. N. Gunaratne,

C. Hardacre, A. McKeown and K. R. Seddon, New J. Chem.,

2010, 34, 723.

26

27

M. Schmeisser, P. Illner, R. Puchta, A. Zahl and R. Van Eldik, 54

Chem. - A Eur. J., 2012, 18, 10969.

R. L. Belford, M. Calvin and G. Belford, 1957, 26, 1165.

28

29

J.-F. Gal and C. Laurence, Chem. - A Eur. J., 2013, 19, 16832.

M. Schmeisser, P. Illner, R. Puchta, A. Zahl and R. Van Eldik,

Chem. - A Eur. J., 2013, 19, 16835.

30

31

M. M. Jones, J. Am. Chem. Soc., 1958, 81, 3188.

V. Polshettiwar and M. P. Kaushik, Tetrahedron Lett., 2004,

45, 6255.

32

33

H. Ohno and Y. Fukaya, Chem. Lett., 2009, 38, 2.

R. Lungwitz, M. Friedrich, W. Linert and S. Spange, New J.

Chem., 2008, 32, 1493.

34

35

L. Crowhurst, P. R. Mawdsley, J. M. Perez-Arlandis, P. A.

Salter and T. Welton, Phys. Chem. Chem. Phys., 2003, 5,

2790.

M. A. Ab Rani, A. Brant, L. Crowhurst, A. Dolan, M. Lui, N.

H. Hassan, J. P. Hallett, P. A. Hunt, H. Niedermeyer, J. M.

Perez-Arlandis, M. Schrems, T. Welton and R. Wilding,

Phys. Chem. Chem. Phys., 2011, 13, 16831.

S. Zhang, X. Qi, X. Ma, L. Lu and Y. Deng, J. Phys. Chem. B,

2010, 114, 3912.

36

37

38

39

40

F. Jalilehvand, Y. Ishii, M. Hidai and Y. Fukuda, Dalt. Trans.,

1996, 3251.

R. P. Matthews, C. Ashworth, T. Welton and P. A. Hunt, J.

Phys. Condens. Matter, 2014, 26, 284112.

Y. Y. Lim, W. Chen, L. L. Tan and K. H. Lim, Polyhedron,

1994, 13, 2379.

L. Crowhurst, P. R. Mawdsley, J. M. Perez-Arlandis, P. A.

Salter and T. Welton, Phys. Chem. Chem. Phys., 2003, 5,

2790.

41

42

43

C. Su, S. Wu, C. Wu and T. Chang, Polyhedron, 1995, 14,

267.

P. Illner, R. Puchta, W. Heinemann and R. Van Eldik, Dalt.

Trans., 2009, 2795.

S. Tsuzuki, H. Tokuda, K. Hayamizu and M. Watanabe, J.

16 | J. Name., 2012, 00, 1-3

Please do not adjust margins

Article Doi

The impact of ionic liquids on the coordination of anions with solvatochromic copper complexes

DOI: 10.1039/c7dt02372c

Source and publish data:

Authors:

Article abstract of DOI:10.1039/c7dt02372c

Full text of DOI:10.1039/c7dt02372c

Products guided by the article

R&D Labs maybe for 7223-96-3

Relevant to this article

Hot Product