Non-photochemical spectral hole-burning mediated by water molecules in interligand pockets of [Cr(terpy)2]3+

Low temperature fluorescence line-narrowing (FLN) and spectral hole-burning experiments (SHB) were performed in the 2E←4A2 spin–flip transition of [Cr(2,2′:6′,2″-terpyridine)2]3+ in frozen ethylene glycol/water (2 ∶ 1) and DMSO/water (2 ∶ 1) glasses.

In the FLN experiments an average 2E splitting of 23 cm−1 is observed.

It is concluded that the interaction with water molecules in pockets provided by the ligands is most likely to be responsible for the relatively efficient non-photochemical hole-burning.

Fast spectral diffusion and spontaneous hole-filling prevent the observation of holes above 20 K. The FLN and SHB experiments were performed by using a diode laser.


Electronic transitions are inhomogeneously broadened in condensed phases due to the variation of local fields.

This broadening obscures valuable spectroscopic information at low temperatures.

Often it can be overcome by the application of laser techniques such as fluorescence line narrowing (FLN) and spectral hole-burning (SHB).1–3

Laser techniques facilitate investigations of subtle details of the electronic structure of molecules or molecular ions.3,4

Non-photochemical spectral hole-burning is a general phenomenon in amorphous systems.

It is based on some rearrangement of host–guest interactions upon photoexcitation, resulting in a slight shift of the transition frequency of the laser-selected subset of chromophores.

Recently, metal complexes of the 2,2′:6′,2″-terpyridine ligand (abbreviated as terpy in the following) and its derivatives have obtained a great deal of attention.5

The [Cr(terpy)2]3+ complex displays a very strong nephelauxetic effect in comparison with very similar systems such as [Cr(bpy)3]3+ (bpy = 2,2′-bipyridine).

This follows from the 760 cm−1 shift of the R lines from ≈729 nm for [Cr(bpy)3]3+ to 772 nm for [Cr(terpy)2]3+.6

The latter complex assumes a distorted meridional geometry, resulting in interligand pockets which facilitate a solvation of the chromium(iii) ion.7

Efficient non-photochemical hole-burning processes can be expected due to the interaction of the solvent molecules with the terpy ligands.

Terpy, like bpy, can stabilise the +2 oxidation state of chromium and hence its complexes have some potential as optical recording materials.

Here we present preliminary results of fluorescence line-narrowing (FLN) and spectral hole-burning (SHB) studies of the 2E←4A2 spin–flip transition in [Cr(terpy)2]3+.


2,2′:6′,2″-terpyridine (terpy) was used as obtained from Aldrich.

[Cr(terpy)2](PF6)3 was synthesised from CrCl3·6H2O and terpy following modified literature procedures.8,9

CrCl3·6H2O was treated with zinc amalgam to form a solution of Cr(ii),8 and a portion of this solution was added to a deoxygenated solution of terpy in methanol.

The resulting red–brown [Cr(terpy)2]2+ solution was bubbled through with oxygen to produce yellow [Cr(terpy)2]3+, which was then isolated as the hexafluorophosphate salt by treatment with NaPF6(aq).

The chromium(iii) concentration in the ethylene glycol/water (2 ∶ 1) and DMSO/water (2 ∶ 1) glasses was ≈ 0.005 M. The samples were cooled to 2.5 K by a closed-cycle refrigerator (Janis/Sumitomo SHI-4.5).

Non-selective luminescence spectra were excited by the 488-nm line of an Ar+ laser (Spectra Physics Stabilite 2017).

The luminescence was dispersed by a Spex 1704 monochromator (1200 grooves mm−1 holographic grating) and detected by a cooled RCA31034 photomultiplier.

After current–voltage conversion by a broadband preamplifier (Products-For-Research) the signal was processed by a lock-in-amplifier (EG&G 5210).

The FLN spectra were excited by a current- and temperature-stabilised (Thorlabs LDC500 and TEC2000 controllers with TCLDM9 thermoelectric mounts) single frequency laser diode (Sharp LT025MD0).

Two frequency and phase locked optical choppers (Thorlabs MC1000) with custom-made blades) were used in tandem with a phase shift of 180° for the resonant FLN experiments.

The first chopper employed a blade with a duty cycle of about 20% to modulate the laser and the second chopper was used at the entrance slit of the monochromator.

This experimental arrangement prevents laser light from entering the monochromator.

In the hole-burning spectra the wavelength of the laser was kept constant for the burn period and subsequently scanned by modulating the current (triangle ramp).

The hole-burning spectra were measured in excitation and the monochromator was used in zeroth order with a RG715 filter.

To reject laser light, the tandem chopper configuration (see above) was employed.

In the hole-burning spectra the output of the lock-in amplifier was averaged by a digital storage oscilloscope (Tektronix TDS210).

The laser scans were calibrated by a spectrum analyser (Coherent model 240) with a free spectral range of 1.5 GHz.

Results and discussion

Fig. 1 shows the non-selectively excited luminescence spectra of [Cr(terpy)2]3+ in ethylene glycol/water (2 ∶ 1) and DMSO/water (2 ∶ 1) at 2.5 K in comparison with the resonantly narrowed luminescence spectrum in the DMSO/water (2 ∶ 1) glass at 20 K.

The FLN spectrum was measured at 20 K to avoid hole-burning.

The 4A22E emission is strongly red-shifted in [Cr(terpy)2]3+ due to a pronounced nephelauxetic effect (reduction of the Racah B parameter caused by electron delocalisation onto the ligands) and peaks at ≈772 nm.

The inhomogeneous linewidth of the electronic origin (R1) is ≈90 cm−1 in both systems.

This is comparable to the value reported for [Cr(bpy)3]3+.10

In the FLN spectrum well defined vibrational sidelines are observed at –121, −206, −327, −209 and −455 cm−1.

The laser predominantly excites chromophores via the R1 line.

However, some molecular cations get excited via the R2 line.

This leads to relatively broad R1-emission at longer wavelength (see inset of Fig. 1) since molecules excited into the R2 line will relax to R1 and energy levels of transition metal complexes in amorphous hosts are poorly correlated (the 2E splitting varies within the inhomogeneous distribution).10

Note that the resonant line in the inset of Fig. 1 is truncated.

The average 2E splitting is ≈ 23 cm−1 with a standard deviation of ≈ 15 cm−1.

Thus the 2E splitting is better defined than in [Cr(bpy)3]3+.10

We partly assign this difference to the nominal D2d symmetry of the [Cr(terpy)2]3+ complex but particularly to the higher rigidity of terpy compared to bpy.

We have found that the 2E splitting in trigonal complexes may vary drastically in amorphous systems due to low symmetry perturbations.10

At 2.5 K the excited state lifetime is 584 ± 1 µs in the ethylene glycol/water glass and 694 ± 1 µs in nafion.

This indicates that solvent molecules are present in the interligand pockets providing an efficient pathway for the non-radiative deactivation of the 2E state by direct energy transfer to high frequency acceptor modes.11

It is important to realize that this electronic to vibrational energy transfer is facilitated by the coincidence of the luminescence of [Cr(terpy)2]3+ with the absorption spectrum of the fourth overtones of O–H and C–H stretch frequencies.

Fig. 2 presents a ≈30% deep spectral hole burnt into the 2E←4A2 transition of [Cr(terpy)2]3+ in ethylene glycol/water (2 ∶ 1).

The initial hole-burning efficiency is estimated to be ≈0.3%.

We tentatively assign this relatively efficient hole-burning process to rearrangements of solvent molecules in the interligand pockets.

The holewidth is strongly dependent on the hole depth.12

For example, a 10% deep hole has a width of 400 MHz compared to the 780-MHz width of the 30% hole.

Taking into account the 70-MHz laser linewidth an upper limit of an effective homogeneous linewidth Γh of 130 MHz results by using eqn. (1).Γhole = 2Γh + 2ΓlaserThis is orders of magnitude larger than the lifetime-limited linewidth of Γh = 1/2πT1 (T1 is the excited state lifetime).3

Fast spectral diffusion caused by rearrangements of host–guest interactions, and electronic and nuclear spin fluctuations must be responsible for the observed holewidth.

For example, the indirect electronic spin–spin interactions may contribute significantly to the holewidth for the concentration used in the present experiment.

Furthermore, superhyperfine interactions with the proton spins of the water molecules in the interligand pockets provides another mechanism for the broadening.

Sample heating can be excluded since the optical density at the laser frequency is very low (<0.001) and relatively low laser powers were used.

Relatively fast spontaneous hole-filling occurs.

For example, the hole area of an initially 20% deep spectral hole decreases by 30% in 500 s.

Spectral holes cannot be burnt at temperatures above 20 K. These observations are indicative of a non-photochemical hole-burning mechanism and two-level systems with relatively low activation barriers enabling fast spontaneous hole-filling at higher temperatures.1–3

If deep spectral holes in the [Cr(terpy)2]3+/ethylene glycol/water system are cycled from 2.5 to 20 to 2.5 K over 10 minutes a significant increase of the holewidth is observed (0.8 to 2.55 GHz).

This is a direct indication of thermally activated spectral diffusion by the dynamic rearrangement of host–guest interactions.

It has been suggested that in frozen solutions the parameters D and E of the spin Hamiltonian of the 4A2 ground state are subject to a statistical distribution caused by the solvent molecules in the interligand pockets.13

This idea is supported by the present SHB experiments: no distinct sideholes are measurable in the range of the reported zero field splitting, 0.28–0.34 cm−1, of [Co(terpy)2](ClO4)2·2.5H2O ∶ Cr(iii).


We have applied an inexpensive laser diode in FLN and SHB studies of a coordination compound in an amorphous host.

The [Cr(terpy)2]3+ system shows a relatively high efficiency of the non-photochemical hole-burning process.

We tentatively assign this to the interaction of solvent molecules in interligand pockets. Further investigations of the solvent–ligand interactions by laser techniques and deuteration effects are currently undertaken.