Abstract

Self-assembly of chromophores nanoconfined in porous materials such as zeolite L has led to technologically relevant host-guest systems exploited in solar energy harvesting, photonics, nanodiagnostics and information technology. The response of these hybrid materials to compression, which would be crucial to enhance their application range, has never been explored to date. By a joint high-pressure in situ synchrotron X-ray powder diffraction and ab initio molecular dynamics approach, herein we unravel the high-pressure behaviour of hybrid composites of zeolite L with fluorenone dye. High-pressure experiments were performed up to 6 GPa using non-penetrating pressure transmitting media to study the effect of dye loading on the structural properties of the materials under compression. Computational modelling provided molecular-level insight on the response to compression of the confined dye assemblies, evidencing a pressure-induced strengthening of the interaction between the fluorenone carbonyl group and zeolite L potassium cations. Our results reveal an impressive stability of the fluorenone-zeolite L composites at GPa pressures. The remarkable resilience of the supramolecular organization of dye molecules hyperconfined in zeolite L channels may open the way to the realization of optical devices able to maintain their functionality under extreme conditions.

1. Introduction

Zeolites are crystalline natural or synthetic porous materials consisting of corner-sharing tetrahedral units, characterized by a regular arrangement of cages and channels of molecular size [1,2]. The nanometric-scale geometry of zeolite pore systems allows the intensive use of zeolites in several fields such as in molecular separation processes and in heterogeneous catalysis [3,4].

Besides being instrumental in traditional applications, the ordered arrangements of zeolitic cages has long been recognized as a route to create advanced materials based on confined, organized nanostructures, such as luminescent metal clusters [5,6,7,8], quantum dots/wires [9] or lanthanides [10]. Turning to more complex guests, the incorporation of fluorescent molecules in zeolite cages generally enhances their emission properties [11,12,13] because it allows to obtain high concentrations of chromophores while limiting the formation of aggregates (which negatively affect the emission intensity) [14]. Incorporated dyes exhibit a preferential orientation of the electronic transition dipole moments, which endows -dye-zeolite host–guest compounds with exceptional energy transfer properties, similar to antenna systems in natural photosynthetic organisms [11,15,16,17,18,19,20,21,22,23,24,25].

Zeolite L (Figure 1) features a unique one-dimensional 12-membered (12 MR) channel system running along the [001] direction. One of the greatest advantages of zeolite L is that high-quality crystals with a desired aspect ratio can be easily synthesized [26,27]. Also importantly, their surfaces can be modified selectively by the attachment of molecules carrying a specific chemical functionality [28,29], thus allowing zeolite L crystals to be sealed (to avoid leakage of dye molecules) [30,31,32], or interfaced to molecules [33], nanoparticles [34,35], living cells [36,37] and rigid supports [38]. Based on these features, zeolite L has been fully exploited as matrix for dye encapsulation and the molecules organization has been hierarchically enhanced up to the macroscopic scale, resulting into functional materials with remarkable electro-optical and energy-transfer performances, already exploited in devices [39]. Composites based on zeolite L as host matrix have reached an advanced stage of development [40] and are presently used as effect pigments [41], in solar energy harvesting [11], or in biomedical technology [42,43].

Such ongoing progress in applications has been accompanied, in the latest years, by a deeper molecular-level understanding of the confined photoactive assemblies, which has been achieved through computational modelling [44,45,46,47,48,49,50], often combined with multi-technique experimental analyses [45,50,51,52,53,54,55,56,57]. The key role of water in tuning the organization of the confined chromophores has been revealed [45,46,58], as well as the stabilizing effect of potassium cations in composites with carbonyl dyes [44,57,59].

The high-pressure behaviour of zeolites has been intensively studied in the last 15 years, as documented in various reviews [50,60,61,62,63,64,65,66], both with pore penetrating and non-penetrating pressure transmitting media (PTM). One reason for this interest is that high pressures, combined with the confining environments of nanometric pores, may reveal unexpected chemical phenomena (e.g., pressure-induced hydration [67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82], ionic conductivity [83], guest exchange [84,85,86,87,88], or realization of new materials otherwise unattainable [89,90,91,92,93,94,95,96,97]). With this paper we would like to answer the question: how do confined supramolecular assemblies of water and chromophores respond to compression?

In this paper, it is reported the study of the high-pressure evolution of zeolite–dye hybrids. Specifically, we selected the zeolite L (ZL)–fluorenone (FL) hybrid (ZL/FL from now on), subject of previous investigations performed by our group [56]. We are interested in exploring whether the compression could favour a more ordered distribution of the dye molecules in the zeolite L channels and if this process leads to an improvement of the optical properties of the hybrid material. In fact, it has been shown that the properties of dye-zeolite systems depend on the molecular orientation, arrangement and packing inside the channel [14,15,20,44,45,46,56,58], which control the guest-guest and host-guest interactions. To unravel and understand at molecular level the compression behaviour of the inclusion composites we adopted an integrated experimental-theoretical approach, based on the use of high pressure (HP) in situ synchrotron X-ray powder diffraction (XRPD) and ab initio molecular dynamics [98] simulations, which already captured a nice example of pressure-induced supramolecular organization in zeolites [97]. Herein, high-pressure experiments were performed up to 6 GPa using non-penetrating pressure transmitting media on samples characterized by three different contents of fluorenone molecules per unit cell, in order to verify how the dye loading influences the structural properties of the hybrid material under compression.

2. Materials and Methods

2.1. Zeolite Dye Hybrids: Structural Details at Ambient Conditions

The LTL framework [1] is built from columns of cancrinite cages stacked with double six membered rings (D6R) along the c axis. These columns are connected to form larger circular 12-ring (12 MR) channels and smaller elliptical 8-ring (8 MR) channels both running along the c axis (8 MR ⊥ [001], channel ǁ [001]). The main channels—with dimensions ranging from of 7.1 to 12 Å—are connected to the parallel 8 MR channels by a non-planar boat shaped 8 MR (8 MR ǁ [001]). In the as-synthesized K-LTL [56,99] [K_8.46(Al_8.35·Si_27.53)O₇₂·17.91H₂O, s.g. P6/mmm, a = 18.3795, c = 7.5281], three positions of K cations were located: site KB in cancrinite cage, site KC in the 8 MR and site KD in the main 12 MR channel.

The ZL/FL composites here investigated are those synthesized and previously characterized at ambient pressure (P_amb) by Gigli et al. [56]. Specifically, three hybrids at different fluorenone loadings were investigated, containing 0.5, 1 and 1.5 molecules of colorant per unit cell (from now on ZL/0.5FL, ZL/1FL, ZL/1.5FL).

In the structures of ZL/0.5FL and ZL/1FL at P_amb [56] the fluorenone molecules are sited in the 12 MR on the mirror planes parallel to the c axis, statistically occupying only one of the six equivalent positions (Figure 2a,b). The oxygen of the carbonyl group of the FL molecule is strongly interacting with potassium atoms KD located along the walls of the main channel. Along with the dye, water molecules were also located in the 12 MR channel: specifically, 14.7 molecules/u.c. in the ZL/0.5FL composite and 9.7 molecules/u.c. in the ZL/1FL one. The structure of ZL/1.5FL resulted to be too complex to be unravelled by structural refinement (i.e., due to high symmetry constrains imposed and to the low electronic density of the atoms), thus a theoretical approach was used. Density functional calculations were indeed able to provide a reliable structural model of the packing of fluorenone inside the composite at room pressure conditions: more specifically, the supramolecular organization of the confined dyes consisted of pairs of molecules—positioned roughly on top of each other, with their long axes nearly parallel to the channel axis—alternated by a dye molecule oriented at about 45° with respect to the zeolite channel axis (Figure 3) [56]. Moreover, calculations evidenced that such a peculiar arrangement—originally called “dye-nanoladder”—was due not only to the structural constraints imposed by the ZL-nanochannel geometry but also to a complex network of intermolecular interactions, among which the dominant one was the coordination of the carbonyl oxygen to the K⁺ extra framework cations (KD)).

2.2. Synchrotron X-ray Powder Diffraction Experiments

All the samples were studied with silicon oil (s.o.) as non-penetrating PTM. The in-situ HP XRPD experiments were performed at the SNBL1 (BM01a) beamline at ESRF (European Synchrotron Radiation Facility) with fixed wavelength of 0.72 Å, using a modified Merril-Basset Diamond Anvil Cell (DAC) [100,101]. The pressure was measured using the ruby fluorescence method on the non-linear hydrostatic pressure scale [102]. The estimated error in the pressure values is 0.05 GPa. Bidimensional diffraction patterns were recorded on a PILATUS2M-Series detector (commercialised by Dectris-Switzerland, Baden-Dättwil, Switzerland) (pixel dimension 172 μm) at a fixed distance of 195 mm from the sample; the exposure times were 300 s for each collected pattern. One dimensional diffraction patterns were obtained in the 2θ range 0°–43° by integrating the two dimensional images with the program Fit2D [103]. Some patterns were collected upon pressure release (labelled (rev) in Tables and Figures). Unit cell parameters determination through Rietveld refinements was possible for all the ramps (Table S1). The Rietveld structural refinements were performed on the ZL/0.5FL, ZL/1FL composites at 2.09 and 2.01 GPa, respectively (from now on will be both labelled 2 GPa (i.e., the highest pressures at which the quality of the diffraction patterns allowed the structural refinement) and upon complete decompression (P_amb (rev)). All the structural refinements were carried out in the space group P6/mmm starting from the atomic coordinates reported by Gigli et al. [56], using the GSAS package [104] with EXPGUI interface [105]. The background curves were fitted by a Chebyschev polynomial function with 24 coefficients for all the samples. The pseudo-Voight profile function proposed by Thomson et al. [106] was used with refined Gaussian (GW) e Lorentzian (LX) terms and a 0.1% cut-off was applied to the peak intensities. The scale factor and 2θ zero shift were accurately refined in all patterns of the data set. Soft constraints were imposed on tetrahedral bond lengths (Si–O = 1.63 Å) as well as on the C−C (in the range 1.39–1.48 Å) and C−O (1.19 Å) distances, with tolerance values of 0.03 Å. The isotropic displacement parameters were constrained in the following way: the same value for all tetrahedral cations, a second value for all framework oxygen atoms and a third value for water molecules. For both the ZL/FL systems, isotropic displacement parameters for FL atoms were kept equal to P_amb values. Occupancy factor for FL were not varied and kept equal to P_amb values for all the refinement [56] and a fourth for the oxygen atoms of the water molecules. Details of the structural refinements are reported in

Table 1. Experimental and structural parameters for selected XRPD refinements performed on the ZL/0.5FL and ZL/1FL composites compressed in silicon oil (s.o.) at 2 GPa and upon decompression (P_amb (rev)). ¹ From Ref. [56].

ZL/0.5FL
P (GPa)	Pamb 1	2 GPa	Pamb (rev)
Space group	P6/mmm	P6/mmm	P6/mmm
a (Å)	18.3860 (4)	18.1788 (9)	18.4349 (8)
c (Å)	7.5228 (2)	7.3866 (4)	7.5498 (6)
V (Å)	2202.4 (1)	2114.0 (2)	2222.0 (2)
R F**2 (%)	7.3	12.3	11.2
No. variables	73	82	82
No. obs.	1319	2474	2474
No. refl.	944	618	618
ZL/1FL
P (GPa)	Pamb 1	2 GPa	Pamb (rev)
Space group	P6/mmm	P6/mmm	P6/mmm
a (Å)	18.3940 (6)	18.250 (1)	18.3962 (8)
c (Å)	7.5203 (3)	7.4063 (5)	7.5189 (4)
V (Å)	2203.5 (1)	2136.3 (2)	2203.6 (2)
R F**2 (%)	7.8	11.9	13.7
No. variables	81	83	83
No. obs.	1319	2422	2404
No. refl.	946	679	639

. Atomic coordinates, site occupancies and isotropic displacement parameters are reported in Table S2 while interatomic distances in Table S3.

2.3. Theoretical Modelling

The behaviour of the ZL/1.5FL sample was simulated both at ambient conditions and at compression corresponding to 1.95 GPa in s.o. adopting the experimental cell parameters. The simulation cell consisted of two crystallographic cells along the 12 MR channel (c axis). The corresponding stoichiometry of the simulated system is K₁₈[Al₁₈Si₅₄O₁₄₄]13·(H₂O) 3·FL, as determined in [56]. The systems were simulated, at room temperature, via first-principles molecular dynamics [98]. The PBE approximation to Density Functional Theory was adopted for the electron-electron interactions and augmented with dispersion corrections via the Grimme D2 approach [107] for the FL-FL interactions. Pseudopotentials and basis set expansion were the same adopted in [56]. Such a computational approach has demonstrated to provide a satisfactory description of diverse systems involving adsorbate-surfaces [108,109,110,111,112,113,114,115,116,117,118,119,120,121,122] or host-guest interactions [123,124,125,126,127,128,129,130,131,132,133,134] including also processes at zeolite interfaces [32,59] and high pressure conditions [97,135,136,137,138,139,140].

Equations of motion were integrated with the Car Parrinello Lagrangean [98,141] adopting a time step of 0.121 fs. The starting configuration for the P_amb simulation corresponded to the minimum energy structure determined in [56]. An equilibration trajectory of 5 ps was followed by 12 ps of production run. The final configuration obtained at P_amb conditions was used as starting configuration for the 1.95 GPa compression simulation. Also in this case, data were gathered from a 12 ps trajectory after a 5 ps equilibration run. In all cases, temperature was controlled via Nose-Hoover thermostats set at 300 K. In all the simulations, performed with the CPMD code [142], no constraints was imposed to the systems except for the experimentally determined cell parameters.

3. Results

3.1. Structure of the ZL-FL Composites at Different Pressure Conditions

Figure 4 shows selected powder patterns of the three ZL-FL composites compressed in s.o. as a function of pressure. With increasing pressure the peak intensities decrease and the peak profiles become broader. Notwithstanding this, complete X-ray amorphization is not achieved up to the highest investigated pressure (about 6 GPa). All the observed peaks are consistent with the P/6mmm s.g., thus ruling out any P-induced phase transition. The patterns collected upon decompression demonstrate that the P-induced effects are almost completely reversible. In fact, the features of the ambient-pressure pattern (P_amb) and the unit cell parameters are rather well recovered upon P release. The decrease of the cell parameters for ZL/0.5FL, ZL/1FL, ZL/1.5FL samples up to 6 GPa are: ∆a = −3%, −2.7%, −2.3%; ∆c = −4.6%, −4.4%, −3.7% accounting for a ∆V = −10.1%, −9.5% and −8.0%, respectively (See Figure 5 and Table S1). These values indicate that the compressibility changes with the loading and that the FL molecules hosted in the channels stiffen the structure.

The hexagonal lattice undergoes a slight anisotropic compression (Figure 5), with c as the most compressible axis. This is probably due to the presence of bonds among the FL carbonyl groups and KD potassium cations, lying along the a direction, that stiffen the structure.

Framework Modifications

The framework P-induced deformations can be summarized as follows (see

Table 2. Experimental framework distances obtained from the refinements performed on the ZL/0.5FL and ZL/1.0FL composites at P_amb (from Ref. [56]), at 2 GPa and upon decompression (P_amb (rev)). ¹ From ref. [56].

	ZL/0.5FL			ZL/1FL
	Pamb 1	2 GPa	Pamb (rev)	Pamb 1	2 GPa	Pamb (rev)
	12 MR			12 MR
O1–O1	10.10	9.76	9.95	10.14	10.30	10.00
O2–O2	10.52	10.43	10.56	10.46	10.12	10.45
	8 MR ⊥ [001]			8 MR ⊥ [001]
O1–O1	8.29	8.41	8.48	8.26	7.95	8.39
O5–O5	4.63	4.54	4.75	4.60	4.70	5.22
* E	1.79	1.85	1.78	1.79	1.69	1.61
	8 MR ‖ [001]			8 MR ‖ [001]
O1–O1	7.52	7.39	7.54	7.52	7.41	7.51
O6–O6	4.66	4.45	4.63	4.68	4.42	4.60
	D6R			D6R
O5–O3–O5	147.67	151.82	145.63	149.18	152.6	143.00
O3–O5–O3	91.68	86.51	93.23	90.09	87.7	95.97
12 MR maximum diameter
O6–O6	15.60	15.44	15.69	15.67	15.67	15.91

* E = (Ellipticity) is the ratio between the largest and the smallest O–O diameters.

and Figure 6):

(i)

In the ZL/0.5FL at 2 GPa, both the diameters (O1–O1 and O2–O2) of the 12 MR shorten. The shortening of the 12 MR O1–O1 diameter is reflected in the lengthening of the O1–O1 diameter of 8 MR channel (ǁ [001]), which becomes more elliptical. Upon pressure release, the original values of the 12 MR diameters are almost regained (remaining slightly smaller than those observed at P_amb), while the 8 MR ones are strongly lengthened with respect to the original values, in accordance with the increase of a parameter (see Table 1 and Table S1).

(ii)

In the ZL/1FL sample, O2–O2 (12 MR) diameter decreases with pressure while O1–O1 increases—probably due to the presence of a larger numbers of FL molecules in the channels with respect to the ZL/0.5FL system. The O1–O1 (12 MR) increase is balanced by the decrease of O1–O1 diameter of the 8 MR channel running alongside (ǁ [001]): as a consequence, 8 MR becomes less elliptical. Upon pressure release, the 12 MR opening remains slightly smaller than that observed at P_amb. The significant enlargement of the 8 MR leads to an overall value for a parameter comparable with that at P_amb.

(iii)

The 8 MR window (8 MR ‖ [001]) (O1–O1, O6–O6), parallel to the c axis, becomes more circular at 2 GPa in both the samples. Once the pressure is released the starting values are regained.

(iv)

The O3–O5–O3 and O5–O3–O5 angle variations indicate that the D6R slightly increases its ditrigonal distortion in both samples. Upon pressure release the P_amb features are almost recovered.

Extra Framework Species

The distribution of the K cations in both composites at P_amb can be described as follows (see Table S2 and S3):

(i)

site KB—in the centre of the cancrinite cage—is fully occupied and coordinated to six framework oxygen atoms O3;

(ii)

site KC—in the centre of the 8 MR channel—is fully occupied and coordinated to four oxygen atoms O5;

(iii)

site KD—near the wall of the main 12 MR channel—is partially occupied and coordinated to six oxygen atoms (O4, O6), two water molecules (WH and WI) and to the oxygen atom of FL molecule (OFL).

At P_amb, in ZL/0.5FL and ZL/1.0FL composites the water molecules (14.7 and 9.7 per unit cell, respectively) are distributed over three extra framework sites (WH, WI, WJ). All of them are located in the main channel. WH site is present only in ZL/0.5FL composite, the other two sites have the same positions of the oxygen atom and of the C3 carbon atom of FL molecule (labelled OFL/WI and C3/WJ, respectively, in Ref. [56]).

Upon compression at 2 GPa the following structural features are observed:

(i)

the distances between KB, KC and the coordinating framework oxygen atoms (Table S3 and Ref. [56]) decrease as a consequence of the shape modifications of both 8 MR channel aperture and D6R. The distances between the cation in the main channel (KD) and O4 and O6 decrease as well. All these effects are more marked in the ZL/1FL system.

(ii)

OFL–KD distance decreases in ZL/0.5FL and remains almost constant in ZL/1.0FL.

(iii)

Compression induces the splitting of OFL/WI and C3/WJ sites, which in the P_amb structures of ZL/0.5FL and ZL/1FL samples [56], occupy single sites. After the splitting, WJ increases its distance from the framework O2 atom, approaching WI site (Table S3). After pressure release, the original positions are recovered in the ZL/0.5FL composite, while this does not happen in the ZL/1FL sample.

(iv)

At 2 GPa, the shape, orientation and arrangement of the fluorenone molecules in the main channel do not change with respect to ambient pressure.

All these P-induced deformations are reversible and, once the pressure is released, the original features of the zeolite and the distances among the FL molecules are almost recovered (Table 2, Table S2 and Table S3).

3.2. Structure of the ZL/1.5FL Composite from First-Principles Molecular Dynamics

As evidenced by the above-discussed data, XRPD refinements provided a satisfactory description of the P-induced structural modifications of the composites characterized by low and moderate dye content. On the other hand, the great number of low-occupancy sites found for water and fluorenone molecules and the high symmetry hindered the structural refinement of the sample containing the maximum amount of dye. Such a difficulty was previously encountered in the room pressure refinement of the ZL/1.5FL composite and it was overcome by integrating the experimental data on cell parameters with theoretical modelling for the atomic coordinates [56].

Hence, encouraged by this result, we exploited again theoretical modelling for achieving an atomistic structural description of the ZL/1.5FL composite at high-pressure conditions. Practically, we used the room-pressure coordinates as an initial guess to determine the composite structure at cell parameters corresponding to 1.95 GPa and we run first-principles molecular dynamics simulations for both P = P_amb and P = 1.95 GPa in order to study the pressure-induced changes of the supramolecular organization inside the zeolite nanochannels at molecular-level detail.

The first remarkable observation is that the unique dye-architecture found at room pressure remains stable at high pressure conditions (Figure 7), with minimal alterations of its intermolecular distances and essentially without significant perturbation of the FL molecular geometry, apart from slight instantaneous distortions from the ideal gas-phase structure, mainly ascribable to thermal motion.

Significantly, the leading interaction stabilizing the confined fluorenone superstructure, i.e., the coordination to potassium cations, not only is maintained but it appears also to be, on average, strengthened upon compression. Indeed, by comparing the pair distribution functions relative to the K-OFL interaction, we deduce a significant shortening of the minimum coordination distance, which passes from 3.1 to 2.9 Å in going from room pressure to 1.95 GPa (Figure 8). Moreover, the splitting of the first maximum into two peaks becomes much more evident upon compression, indicating that the interaction of the carbonyl group and potassium may have different degrees of strength among the hyperconfined fluorenone molecules. Taken as a whole, these data indicate that the supramolecular architecture of dyes responds to pressure essentially by approaching with the carbonyl groups the potassium cations in the 12 MR channel, without undergoing appreciable modifications of both intra- and inter-molecular distances.

By considering now water molecules, which share the channel space with fluorenone, a striking similarity emerges by comparing their pair distribution functions at room pressure and 1.95 GPa (Figure 9). Actually, for both water protons (Figure 9a) and water oxygens (Figure 9b) the two curves are almost superposable, showing also very close positions for first maximum peaks of water-water hydrogen bonding (corresponding, at room pressure, to H_w–O_w and O_w–O_w distances of 1.85 and 2.80 Å, respectively—see upper panels of Figure 9a,b). In particular, only the shoulder of the H_w–O_w peak—found at 1.78 Å at P_amb—appears to be very slightly displaced towards greater distances. Also the interaction of water with framework oxygens—which, as normally found for zeolitic water, is weaker than water–water hydrogen bonding [143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164,165,166,167,168]—appears to be nearly unperturbed upon compression and characterized by H_w–O_f and O_w–O_f distances of 1.95 and 2.90 Å, respectively (see centre panels of Figure 9a,b). Furthermore, the interaction of water with potassium cations is nearly unaffected by pressure at short range distances (with first maximum position at 2.90 Å, see Figure 9b, bottom panel), indicating that the coordination environment of the zeolite L extra framework cations is, on average, spectacularly insensitive to compression: no variation of the K coordination number occurs and distances from water oxygens remain basically unaltered. Finally, very minor changes are detected for the water-carbonyl interaction: the first peak position passes from 1.80 to 1.75 Å upon compression, indicating a slightly stronger interaction of water molecules with the carbonyl group of the dye (Figure 9a, bottom panel).

Interestingly, whereas in the ZL/0.5FL and ZL/1FL composites the O6–O6 distance (i.e., the maximum diameter of the 12 MR channel) either remains constant (ZL/1FL) or slightly shortens (ZL/0.5FL) with pressure (Table 2), such a distance increases in the ZL/1.5FL composite. Accordingly, the calculations also predict a decrease of the 12 MR window (i.e., the channel opening) of ZL/1.5FL with compression (see O1–O1 and O2–O2 distances in

Table 3. Calculated average diameters of 12 MR channel in Å from room-temperature FPMD simulations at different compression for the hydrated ZL/1.5FL system.

Distance	Pamb	1.95 GPa
O1–O1	9.934	9.878
O2–O2	10.431	10.409
O6–O6	15.520	15.614

). This behaviour is due to the peculiar close-packing arrangement of the extra framework content at maximum dye loading (ZL/1.5FL). Specifically, as compression occurs mainly along the channel axis, the guest species—water molecules and dye nanoladder—can only respond by further clustering in the maximum-diameter region of the channel (Figure 7), thus explaining the O6–O6 lengthening and the 12 MR window narrowing. The different pressure response of the composites according to the dye loading is a consequence of the stiffening/template effects of the extra framework content, already observed in several high-pressure studies on zeolites and zeolite-based materials (see, e.g., Refs. [62,63,66,77,81,139,169,170,171]).

Taken together, besides the impressive stability of the fluorenone nanoladder, these results underline a quite surprising and important feature of the confined supramolecular system: even though water is smaller than the dye and hence in principle more easily displaceable upon compression, the arrangement of the water molecules in ZL/0.5FL and ZL/1FL is only slightly modified by the application of hydrostatic pressure—in particular, WJ increases its distance from the framework O2 atoms—and remains essentially unperturbed in ZL/1.5FL. This is ascribable to the fact that the P_amb structure of the ZL/1.5FL composite has already a close-packing arrangement of the extra framework species. In particular, all water molecules are fully stabilized by coordination to potassium cations and by the network of hydrogen-bond interactions with fluorenone carbonyl groups, water and framework oxygens [56].

4. Discussion

The main motivation of our study was to elucidate the behaviour of luminescent hybrid materials, obtained via the inclusion of the fluorenone dye in zeolite L microcrystals, under compression. To answer this question, a thorough study has been performed using non-penetrating pressure transmitting fluid (silicon oil) in the high pressure X-ray diffraction experiments and complementing those results by ab initio molecular dynamics simulations, which allow to explore the finite temperature behaviour of the system at atomistic level detail.

If we consider that the technological applications of these composites rely on the presence of confined regular arrays of photoactive species, it is important to understand how this supramolecular organization of dye molecules responds to non-ambient conditions, specifically an applied pressure. Remarkably, all of the experiments and simulations evidenced the approaching of the fluorenone carbonyl group to potassium—i.e., the strengthening of the main stabilizing interaction in zeolite L composites with carbonyl dyes. Also notably, the applied pressure did not cause appreciable distortions of the molecular geometry of the chromophore, because of the zeolite framework ability to withstand compression effects, as already demonstrated in other high pressure studies of zeolites with non-penetrating fluids [135,138,139,140,169,170,171].

Experiments performed on the samples with low and medium fluorenone contents showed that the arrangement and the distances of the dye molecules, upon increasing pressure, do not change and their behaviour is consistent with the observed shortening of the c-parameter of the zeolite host. This pressure-induced shortening of intermolecular distances between dye molecules is however reversible upon returning to P_amb, because it depends solely on non-covalent interactions. Also, such an effect is much less pronounced in the case of the high-concentrated composite (1.5FL/ZL), where the dye molecules are organized in a very close-packed arrangement already at room pressure conditions.

All these observations suggest that also the optical properties of the composites should not be significantly perturbed by external pressures of GPa-scale, an extremely important feature in the perspective of potential applications.

5. Conclusions

Overall, this study highlights the stability upon compression of the zeolite L/fluorenone adducts. This very important fact is of relevance in view of possible applications of such a kind of dye-zeolite composites at conditions different from the standard ones. We have also monitored an intriguing fluorenone-loading dependence of the system response to the applied pressure. However, some common features have been evidenced as well. In particular, compression brings about a strengthening of the fluorenone–zeolite L interactions, as shown by the shortening of the distance between the carbonyl oxygen atoms of the dye and the K⁺ cations. Also, the molecular structure of the dye guests is barely influenced by compression. This is essentially due to the zeolite framework, which bears most of the compression effects—as evidenced by the (moderate) changes in the cell parameters and by the (moderate) changes in the 12 MR channel—that are relevant for the zeolite pores volume. Such a framework stability upon compression induces only a modest variation of the pores volumes, which affects only to a minor extent the molecular structure of the organic guests. In conclusion, we have shown for the first time the stability towards pressure of dye-zeolite L adducts, which suggests that also the technologically relevant photophysical properties of these appealing materials could be exploited at high pressure conditions.