Solid state fluorescence properties

In the following the solid state fluorescence properties of 21-28 will be presented.

Previous research has shown that in general all selenium oxidized phosphoryl anthracenes show distinctly weaker fluorescence emission than their sulfur oxidized counterparts.[44b, 44c]

This can be assigned to the presence of a heavy atom in the molecule which may generally produce a red-shift of the emission maximum and also favour non-radiative decay pathways over fluorescence emission.[7a, 7b, 7d, 70]

Hence

comparison is only reasonable among the four sulfur oxidized compounds 21-24 and among the four selenium oxidized compounds 25-28.

The solid state excitation spectra of 21-24 are nearly identical in shape. They show a broad range of circa 100 nm of possible excitation wavelengths, reaching a maximum at 450 nm. Similar behavior has also been observed for SPAnPS (c.f. 3.2). The excitation band is located in the spectral range of visible light, which is particularly attractive for fluorescent materials, leaving the employment of UV light for excitation obsolete.

Furthermore, due to the broadness of the band, monochromatic light is not required for excitation, which again is beneficial for fluorescent materials.

Figure 3-44: Normalized solid state excitation (green) and emission (red) spectra of MeAnPS(NMe2)2 (21).

The corresponding emission spectra show each one fairly broad maximum, lacking the characteristic vibrational band structure of many anthracene derivatives.^[49-50]

Additionally, the emission maximum is red-shifted by circa 70-80 nm compared to unsubstituted anthracene. This phenomenon was also observed for the strongly emitting SPAnPS structures 15 and 20 (c.f. 3.2) The lack of vibrational bands has often been assigned to exciplex formation in solid state for other anthracene derivatives,^[52,

55] but does not appear to be present in this case in view of the molecular arrangement. The normalized emission spectra of 21-24 show a notable bathochromic shift of the emission maxima of the gold complexes (~15 nm) compared to free phosphoryl anthracenes. This is in accordance with theory that predicts a possible red-shift of emission in the presence of heavy atoms in the structure. Also the emission maximum of MeAnPS(NEt2)2 (23) is slightly red-shifted compared to MeAnPS(NMe2)2

(21), though there is no heavy atom present in either of the structures. Similar phenomena have repeatedly been explained by π-π interaction, with increasing overlap and shorter π-π distance leading to a stronger red-shift of emission.[42c, 53-54]

This does not apply to 21 and 23. While 21 exhibits a π-π overlap of circa 35% with a comparatively short π-π distance of 3.51 Å, there is virtually no π-π interaction in the solid state structure of 23. Though the bathochromic shift of emission between 21 and 23 is fairly small, the expected effect of the π-π overlap is not observed.

Table 3-10: Summarized structural properties of 21, 22, 23, and 24.

21 22 23 24

[MeAnP(NMe2)2(S)AuCl] (22) and [MeAnP(NEt2)2(S)AuCl] (24) also differ significantly in terms of π-π overlap, with π systems in 24 achieving an overlap of 25% at 3.21 Å and 22 showing virtually none. Again the expected red shift of emission for 24 is not found, as 22 and 24 have nearly identical maximum emission wavelengths.

Emission intensities of aromatic molecules can also be influenced by the degree of π-π interaction. A stronger interaction has often been named as a quenching factor of solid state fluorescence.[7d, 42c, 48b, 50]

Furthermore, the deformation of the anthracene moiety caused by steric strain on the fluorophore has been shown to have a quenching effect on fluorescence of anthracene derivatives, which was confirmed by the results obtained in 3.2.^{[49, 52]} As mentioned before, heavy atoms in the molecular structure can also exert similar effects on emission properties.

Figure 3-45: Left: Normalized solid state emission spectra of MeAnPS(NMe2)2 (21, blue), [MeAnP(NMe2)2(S)AuCl] (22, cyan), MeAnPS(NEt2)2 (23, red), and [MeAnP(NEt2)2(S)AuCl] (24, green);

right: solid state emission spectra of MeAnPS(NMe2)2 (21, blue), [MeAnP(NMe2)2(S)AuCl] (22, cyan), MeAnPS(NEt2)2 (23, red), and [MeAnP(NEt2)2(S)AuCl] (24, green).

When comparing the emission intensities of 21-24 it becomes clear at first sight that the presence of the heavy gold ion in the molecular structures of [MeAnP(NMe2)2(S)AuCl] (22) and [MeAnP(NEt2)2(S)AuCl] (24) does not lead to appreciable fluorescence quenching, as both gold complexes show stronger emission than the corresponding free ligands 21 and 23, respectively. Considering the fluorescence quenching factors named above, excluding the effect of heavy atoms, the strongest emission of 22 seems plausible, as the solid state structure of 22 displays no π-π overlap and the weakest steric deformation of the fluorophore.

Table 3-11: Solid state fluorescence properties of 21-24.

21 22 23 24

λEx (max) [nm] 449 449 449 449

λEm (max) [nm] 482 499 488 502

Irel 0.67 1 0.07 0.39

Consequently the similar emission of MeAnPS(NMe2)2 (21) and 24 is due to their comparable π-π interactions (slightly larger overlap in 21, slightly shorter π-π distance in 24). 24 exhibits a stronger deformation of the fluorophore than 21. The significantly weaker emission of MeAnPS(NEt2)2 (23) surprisingly does not fit into this series. The solid state structure of 23 shows no π-π overlap (as does the strongest fluorescent [MeAnP(NMe2)2(S)AuCl] (22)) and only moderate deformation of the anthracene

moiety, which suggests that other factors must be taken into account to explain the observed emission properties.

C-H^…π hydrogen bonds

The results derived from the investigation of SPAnPS (c.f. 3.2) have indicated the sensitivity of fluorescence emission of phosphoryl anthracenes towards C-H interactions with the aromatic π system. Two strong and nearly orthogonal sp² CH^…π bonds to each fluorophore dictate the presence of fluorescence emission in the case of SPAnPS@tol (15).[44b, 44c] As stated in the description of the crystal structures, C-H^…π type interactions are also found in 21-28. When these are taken into account in addition to the factors considered above, the emission properties of 21-24 become conclusive. Employing the assumptions made earlier regarding the influence of length, polarity and angle of CH^…π bonds, the strength of the interactions found in the packing plots can be classified. MeAnPS(NEt2)2 (23) displays the weakest and least directed C-H^…π interaction of all four compounds, as the methyl group is not rotated to the optimum position, producing a low angle of only 31.2°. Furthermore the anthracene moieties of the “head-to-tail” arrangement are shifted to a degree that the methyl C-H can barely interact with the π system. This also prevents the aromatic C-H-bonds from adopting a position above the ring centres of the adjacent anthracene moiety like in the structures of 21, 22, and 24. The strongly fluorescent [MeAnP(NMe2)2(S)AuCl] (22) shows one fairly short sp³ C-H^…π bond (2.776 Å) and one sp² C-H^…π bond (2.980 Å), both enclosing larger angles with the anthracene plane. These can both be considered clearly stronger than the single C-H^…π interaction found in 23. MeAnPS(NMe2)2 (21)

Figure 3-46: Normalized solid state excitation (green) and emission (red) spectra of [MeAnP(NMe2)2(Se)AuCl] (26).

Though the emission of 25-28 is distinctly weaker than of the sulfur oxidized compounds (approximately by the factor of 15) the fluorescence phenomena are directly comparable. The excitation spectra of 25-28 again are nearly identical, showing a broad maximum (~110 nm). The emission band is also fairly broad and lacks a vibrational band structure as previously observed for 21-24. Compared to the sulfur oxidized species, the emission maxima of the selenium derivatives are red-shifted by approximately 40 nm.

Table 3-12: Summarized structural properties of 25, 26, 27, and 28.

25 26 27 28

Especially MeAnPSe(NMe2)2 (25) and MeAnPSe(NEt2)2 (27) show excitation and emission maxima which lie very close together, separated only by 15-20 nm. The normalized emission spectra of 25-28 reveal similar tendencies as deduced for 21-24.

The emission maxima of the gold complexes [MeAnP(NMe2)2(Se)AuCl] (26) and 28 show a bathochromic shift of nearly 35 nm compared to the uncoordinated phosphoryl anthracenes 25 and 27. Again, no significant shift of emission caused by π-π overlap is found. 25 and 27 have nearly identical emission maxima despite the fact that MeAnPSe(NMe2)2 (25) shows π-π overlap of 35% at a distance of 3.60 Å while 27 shows no overlap at all. This also applies to [MeAnP(NMe2)2(Se)AuCl] (26) and [MeAnP(NEt2)2(Se)AuCl] (28) with identical emission maxima but clear differences in π-π overlap. Thus, the observed red-shift of the gold complexes can solely be assigned to the presence of a heavy gold atom in the structures.

The emission intensities of 25-28 indicate that the gold atoms in 26 and 28 do not lead to significant fluorescence quenching, as both gold complexes show the strongest emissions of the four compounds. The weakest emission is observed for MeAnPSe(NEt2)2 (27), which is isostructural to 23, which exhibited the weakest emission of all sulfur oxidized compounds. Hence, the structural constitution of both MeAnPS(NEt2)2 (23) and 27 appears to be particularly obstructive for fluorescence emission. The comparatively strong deformation of the anthracene moiety (13.8°

folding angle) and a single low-angle C-H^…π interaction account for these findings like in 23.

Figure 3-47: Left: normalized solid state emission spectra of MeAnPSe(NMe₂)₂ (25) (blue), [MeAnP(NMe₂)₂(Se)AuCl] (26) (cyan), MeAnPSe(NEt₂)₂ (27) (red), and [MeAnP(NEt₂)₂(Se)AuCl] (28) (green); right: solid state emission spectra of MeAnPSe(NMe₂)₂ (25) (blue), [MeAnP(NMe₂)₂(Se)AuCl]

(26) (cyan), MeAnPSe(NEt₂)₂ (27) (red), and [MeAnP(NEt₂)₂(Se)AuCl] (28) (green).

MeAnPSe(NMe2)2 (25), which shows the second weakest emission, has the strongest π-π overlap (~35% at 3.60 Å) and the strongest distortion of the fluorophore (16.7° folding angle) of all four compounds. These factors would generally lead to an expected emission even weaker than of MeAnPSe(NEt2)2 (27). But the solid state structure of 25 also contains a sp³ C-H^…π and a sp² C-H^…π interaction, both of medium strength. Taking these into account, the emission of 25 becomes comprehensible.

Slightly stronger emission is observed for [MeAnP(NEt2)2(Se)AuCl] (28), which has a smaller π-π overlap (~15%) than 25 in combination with a moderate folding angle of 11.1°.

Table 3-13: Solid state fluorescence properties of 25-28.

25 26 27 28 folding angle of 12.7° which is close to the angle observed in the structure of 28. While no π-π overlap is found, anthracene bound hydrogen atoms are located exactly above the π systems of the parallel oriented anthracene moiety at a distance of only 3.41 Å.

Additionally the shortest C-H^…π bond of all eight compounds is found, measuring only 2.617 Å with a fairly acute angle of 65.5° to the anthracene plane. As for 21-24, the observed emission properties follow the trend of number and strength of C-H^…π interactions present.

The analysis of the solid state fluorescence properties of 21-28 suggests that the influence of C-H^…π interactions on emission is an important factor in the process of understanding the basic requirements for solid state fluorescence phenomena. When excluding these from argumentation, by just considering quenching factors as deformation of the fluorophore and π-π overlap, the observed properties cannot be explained conclusively. Especially the effects of π-π overlap, which are widely acknowledged in the fluorescence literature, do not occur in the expected manner.

Although the observed π-π distances are fairly short, no bathochromic shift of

emission was detected for any of the compounds showing π-π overlap. The proposed quenching from π-π overlap also does not appear to be the predominant factor in the quenching of emission, as compounds MeAnPS(NMe2)2 (21), [MeAnP(NEt2)2(S)AuCl]

(24) and MeAnPSe(NMe2)2 (25) exhibit considerable emission in spite of π-π overlap.

Similar phenomena have been documented before, e.g. in the work of Dreuw et al.

where π stacked naphthalene derivatives show stronger emission than the corresponding monomer arrangement.^[42b] The deformation of the fluorophore, represented by folding of the anthracene moiety ranging from 6.0° to 20.7°, seems to have a stronger influence on the emission properties of 21-28. The twist deformation does not appear to have a strong impact on the emission properties. While for the different structures of SPAnPS, even a fluorescence enhancing effect of twist deformation was presumed, the solid state fluorescence of 21-28 cannot confirm this principle. As stated earlier, the concept of solid state fluorescence quenching by distortion of the anthracene moiety was introduced by Mizobe et al. in the course of their work on 2,6-anthracenedisulfonic acid (2,6-ADS).[49, 51-52] The distortions found in their compounds are, however, minimal compared to the downright deformed fluorophores of 21-28. All anthracene moieties are almost perfectly planar in the work of Mizobe et al. and IR-spectroscopy was necessary to detect small deviations in steric distortion. Therefore folding angles of over 20° and additional twisting of the ring system generating ring angles clearly deviating from the ideal 120° sp²-angle are expected to give nearly complete fluorescence quenching, which again is not observed.

On the other hand T-shaped or “herringbone”-like arrangements described in literature are often strongly fluorescent. Though the C-H^…π interactions resulting from this arrangement are sometimes recognized (as in Mizobe’s work)^[49], they often are not identified as direct contributors to the fluorescence properties of the given compound. Especially the T-shaped arrangement of anthracene moieties in the solid state often provides several short aromatic C-H^…π interactions per molecule which can be considered strong due to the polarity of the aromatic C-H bonds and the steep angle to the π system of frequently close to 90° (Figure 3-48). Especially in the case of 2,6-ADS, large differences between virtually non-fluorescent one-dimensional arrangements and strongly fluorescent two-dimensional “zig-zag” arrangements of the molecules are found. While by the geometrical operations applied for quantification of deformation in this thesis no difference between both forms is determinable, the two dimensional form exhibits strong C-H^…π bonding (Figure 3-48), which is absent in the

one-dimensional structure. The This arrangement comes close to that found in SPAnPS@tol (15) which has shown the possible impact of such interactions and the results from 21-28 also suggest that C-H^…π bonds cause a decisive influence on the solid state fluorescence properties of anthracene derivatives.

Figure 3-48: C-H^…π interactions in a strongly solid state fluorescent compound by Mizobe et al.^[49]

Summarizing the solid state fluorescence properties of 21-28 from a more remote perspective, it was found, despite the chalkogen or complexed gold atom, that in general every dimethylamino substituted compound showed stronger emission than the diethylamino substituted equivalent (21 > 23; 22 > 24; 25 > 27; 26 > 28). Broken down to a simple resumé, bulkier substituents produce a stronger deformation and stronger steric repulsion in the solid state which again leads to larger distances and weaker C-H^…π interactions between the fluorophores, which in the case of 21-28 results in suppression of fluorescence emission.