X‐ray Excited Optical Fluorescence and Diffraction Imaging of Reactivity and Crystallinity in a Zeolite Crystal: Crystallography and Molecular Spectroscopy in One

Abstract Structure–activity relationships in heterogeneous catalysis are challenging to be measured on a single‐particle level. For the first time, one X‐ray beam is used to determine the crystallographic structure and reactivity of a single zeolite crystal. The method generates μm‐resolved X‐ray diffraction (μ‐XRD) and X‐ray excited optical fluorescence (μ‐XEOF) maps of the crystallinity and Brønsted reactivity of a zeolite crystal previously reacted with a styrene probe molecule. The local gradients in chemical reactivity (derived from μ‐XEOF) were correlated with local crystallinity and framework Al content, determined by μ‐XRD. Two distinctly different types of fluorescent species formed selectively, depending on the local zeolite crystallinity. The results illustrate the potential of this approach to resolve the crystallographic structure of a porous material and its reactivity in one experiment via X‐ray induced fluorescence of organic molecules formed at the reactive centers.

Abstract: Structure-activity relationships in heterogeneous catalysis are challenging to be measured on as ingle-particle level. Forthe first time,one X-ray beam is used to determine the crystallographic structure and reactivity of as ingle zeolite crystal. The method generates mm-resolved X-ray diffraction (m-XRD) and X-raye xcited optical fluorescence (m-XEOF) maps of the crystallinity and Brønsted reactivity of az eolite crystal previously reacted with as tyrene probe molecule.T he local gradients in chemical reactivity (derived from m-XEOF) were correlated with local crystallinity and framework Al content, determined by m-XRD.T wo distinctly different types of fluorescent species formed selectively,d epending on the local zeolite crystallinity.T he results illustrate the potential of this approacht or esolve the crystallographic structure of aporous material and its reactivity in one experiment via X-ray induced fluorescence of organic molecules formed at the reactive centers.
Zeolites are microporous aluminosilicates that play amajor role as solid acid catalysts in industries. [1][2][3] Zeolite framework aluminum is commonly related to the catalytically active Brønsted acid sites. [4,5] Thes ingle crystal architecture and distribution of Al sites over short-and long-range distances [6] influence the overall catalytic activity and success of various post-treatment methods aiming to improve mass transport by controlled dealumination and desilication. [7][8][9] Ar emarkable example of the compositional and structural complexity of zeolites is ZSM-5 with the MFI topology, often found with pronounced Al zoning [10][11][12][13] and complex internal intergrowth structures. [14,15] Both Al zoning and architecture of the crystals may strongly affect the outcome of post-synthesis modifications and lead to remarkable differences in mesoporosity [8,16] and reactivity. [17,18] Whereas various micro-spectroscopy methods previously introduced provided aw ealth of information about inter-a nd intra-particle heterogeneities in structure and reactivity, [15,[19][20][21] direct structure-reactivity relationships remain difficult to establish.
Forthis study we used large zeolite ZSM-5 crystals [18,[30][31][32] and aB r ønsted acid-catalyzed probe reaction based on the oligomerization of 4-methoxystyrene.U pon the protonation of 4-methoxystyrene on zeolite ZSM-5, oligomeric carbocations are formed, revealing the location of accessible Brønsted acid sites. [20,32] If excited by X-rays,these molecules undergo photoemission in the optical region (UV/Vis), ap henomenon that is generally known as X-ray excited optical luminescence, [33][34][35] here referred to as XEOF. Recently,s everal strategies were developed at synchrotrons to utilize the XEOF emission of visible light for studies of functional materials. [36][37][38][39][40][41][42][43][44] Our method makes use of al ess common method to simultaneously excite electronic transitions in organic molecules and resolve the crystallographic structure of asingle crystal. Figure 1illustrates the approach for measuring m-XRD/m-XEOF maps of asingle steamed ZSM-5 crystal stained with 4methoxystyrene in one experiment. Details of the setup can be found in the Supporting Information. Hard X-rays (8.5 keV) focused to as pot size of 500 nm were used for the successive m-XRD and m-XEOF imaging of as ingle ZSM-5 crystal. Figure 1a shows the response of an X-ray detector for specific (16 00 )a nd (0 16 0) Bragg reflections,w hich were previously used to study the intergrowth structure of zeolite ZSM-5. [13] An optical fiber for the collection of the XEOF signal was placed in the close proximity to the sample stage at ca. 200-300 mmdistance (Supporting Information, Figure S1). TheX -ray excitation of the formed cyclic and linear dimeric styrene species takes place along the beam trajectory,w hich results in aX EOF spectrum (Figure 1b). Ther esulting fluorescence is related to the accessible and reactive Brønsted acid sites,where the formation of the fluorescent species takes place.
Thee valuation of the XEOF signal with fluorescence microscopy and the inherent photobleaching processes of the fluorescent carbocations in the presence of X-rays are described in the Supporting Information. Ther ecorded fluorescence intensity decayed with the time constant of 4.5 AE 0.3 s( Supporting Information, Figure S2). Thep hotobleaching showed clear dose-dependent behavior but did not cause the formation of new fluorescent bands.Prior to the 2D XEOF scans,t he beam damage to organic molecules was minimized by using neutral density filters.Acquisition time of 1.95 sp er point was chosen to collect good quality XEOF spectra and avoid potential artefacts that are due to photobleaching.Asacompromise between the sampling frequency and scanning time,t he spatially correlated maps were acquired in steps of 4 mmf or m-XEOF (1.95 se xposure time, ET) and 2 mmf or diffraction (20-50 ms ET) with the typical X-Y scanning pattern presented in Figure 1c.T he diffraction rocking maps were collected after the m-XEOF intensity maps (a single X-Y scan), by changing the incident angle of the beam and repeating X-Y scans for 13 rocking angles,insteps of Dq = 0.18 8.The X-Y scans were carried out in afast scanning mode (K-map), as described by Chahine et al. [45] To evaluate the impact of steaming on reactivity,w e investigated as teamed ZSM-5 crystal with am ore complex intergrowth structure (Figure 2a). The908 8 intergrowth of the selected ZSM-5 crystal seems to be interconnected in an anomalous manner, when compared to previous reports. [14,46] Thes patial distribution of the crystallographic phases was resolved by integrating the contributions of the higher-order (16 00)and (0 16 0) Bragg reflections for agiven range of X,Y positions ( Figure 2b). Ther esulting spatially resolved diffraction maps,obtained in X-ray strain orientation calculation software (XSOCS), [45] reveal the anomalous and asymmetrical crystal growth (Figure 2c). Thec ontribution of each phase in the diffraction signal will depend on the orientation of the phase with respect to the optical path of the X-ray beam ( Figure 2d).
TheZ SM-5 crystal was tested for XEOF response in the visible region by collecting X-ray excited fluorescence light during ar aster scan of the crystal. An averaged XEOF spectrum summed over all collected data points is shown in Figure 3a.A ni ntense emission band with the highest intensity in all recorded XEOF spectra appeared at about 530 nm, followed by two less intense emission bands at circa 615 nm and circa 670 nm. Thelatter two bands appeared to be red-shifted (up to 20 nm) as compared to the fluorescence microscopy spectra (600 and 650 nm) of the same species (Supporting Information, Figure S3). These two emission bands have been previously attributed to linear dimeric and trimeric species that are confined along the straight pores of ZSM-5. [20,32,47] Theh igher-energy XEOF band at 530 nm is assigned to cyclic dimeric species.U nlike the lower-energy bands,t he band at 530 nm was not detected in the m-XEOF experiments with parent zeolite crystals (Supporting Information, Figure S3). Furthermore,Fornes et al. [48] and Stavitski et al. [47] have reported the UV/Vis absorption band at 490 nm originating from cyclic dimeric carbocations.T he same   [45] The diffraction intensities were summed over all 13 rocking angles for the regions of interest defined in (b). The yellow lines denote the vertical (X-Z) cross-sections shown in (d). d) Exposure of the different crystallographic subunits along the optical path of an X-ray beam;the dotted lines illustrate the propagationo fthe X-ray beam throughout the crystal resulting in the diffraction information from different crystalline domains.
species are found to be formed at the near-surface acid sites and crystalline defects induced by steaming. [49,50] Figure 3a presents two XEOF spectra taken from different positions along the crystal (1 and 2inFigure 3b). We note that the position and intensity of the emission maximum between 610-615 nm change depending on the extent of reactivity.T he emission maximum was about 615 nm for the highly reactive domains (spectrum 2, Figure 3a), and shifted towards higher energies (600 nm) for the domains with lower XEOF intensity (spectrum 1, Figure 3a). We attribute the observed shift to intermolecular interactions of the closely packed oligomeric carbocations. [32,33] The2 Dm-XEOF map in Figure 3b shows an otable gradient in XEOF intensity towards the bottom side of the crystal. Clearly,s teaming has unevenly affected different parts of the crystal. To resolve the differences in the positions and amplitudes of the emission bands we have applied aGaussian deconvolution of the XEOF spectra, by fitting the XEOF spectra with three Gaussians centered at the emission maxima at 530, 610, and 670 nm. Thespatially resolved maps of the XEOF intensities are shown in Figure 3c for the cyclic species and Figure 3d for the linear dimeric species.
Thevertical positions of the studied reflections on the Xray detector can be translated into the corresponding 2q values (Figure 1a). In this way,X -ray diffractograms were constructed as 1D representations of the 2D detector response.T oi llustrate the complexity of the m-XRD/m-XEOF data set, seven different points were chosen along the crystal (Figure 4a)t os how both the recorded X-ray diffractograms (Figure 4b)and corresponding XEOF spectra (Figure 4c). Principal component analysis (PCA)a nd subsequent clustering turned out to be very powerful to classify the recorded data sets according to their spectral feature. Analysis of the m-XRD data set divided the 2D diffraction intensity map into five clusters that have distinct diffraction features,w hich are represented by different colors in the PCA-XRD map (Figure 4d). Similar classification was made with the m-XEOF data set (Figure 4e). Thec olor-coded diffractograms and XEOF spectra in Figure 4b,c highlight spectral differences between the individual clusters in Figure 4d,e,respectively.A veraged cluster spectra of both PCA-XRD and PCA-XEOF clusters are shown in the Supporting Information, Figure S4.
PCA divided the 2D map of crystallinity based on the intensities and positions of the (16 00)and (0 16 0) reflections, which translate directly into the strain in the crystal lattice that is imposed by Al enrichment/depletion. Theo uter regions of the crystal (blue clusters,p oints 4a nd 5i n Figure 4b)s how low XRD peak intensities;t he (0 16 0) peak is notably shifted towards higher d-spacings (lower 2q values) as observed previously for parent zeolite crystals, [13] whereas amaximum of the (16 00)peak seems to be shifted towards lower d-spacings.Inaparent ZSM-5 crystal, the outer region is Al-rich [13] and subsequent steaming leads to dealumination and contraction of the unit cell along the a lattice vector. Theo verlay of the PCA-XRD clustered regions and

Angewandte Chemie
Zuschriften XEOF intensity map in Figure 4findicates the lowest XEOF intensity,m eaning the lowest reactivity,i nt hese clusters. m-XRD intensity maps in Figure 2c suggest the highest diffraction intensity originating from the middle of the crystal. Thed iffractograms of the inner clusters,d epicted in orange (points 2, 3a,b) and red (point 1) in Figure 4b,d, confirm the higher content of framework Al and al ower degree of dealumination. Consequently,t heir reactivity is higher, as visible from the most intense XEOF emission from the inner regions in Figure 4f.Avery distinct feature is the green cluster (point 6i nF igure 4b,d) that represents ah ighly crystalline domain with the lowest d-spacing for the (16 00 ) reflection, which is an indication of the Al-poor phase that is more resistant to steaming and less reactive due to lower accessibility of the microcrystalline domains.
TheXEOF intensity ratio of the cyclic and linear dimeric species can be used as an indication of the extent of reactivity that is determined by crystallinity and accessibility of the zeolite domains,a ss hown in the intensity map of this ratio (Figure 4g). This map resembles the PCA-m-XEOF map from Figure 4e,h ighlighting the differences in the XEOF spectra compared in Figure 3a.T he higher amount of cyclic dimeric species with respect to linear dimeric species correlates well with the total XEOF intensity and the loss of crystallinity in the ZSM-5 crystal.
Our experimental approach in combination with PCA shows that a2 Dm-XRD mapping can provide useful crystallographic information about the 3D structure of asingle zeolite crystal. This is possible owing to the presence of 908 8 intergrowths and pronounced Al zoning that divide the analyzed volume into distinct crystallographic phases,visible also in PCA cluster maps (Figure 4d). Although precise 3D information cannot be extracted from our measurements,the positions and orientations of different crystallographic phases can still be identified based on the previous knowledge of the crystallographic and compositional anisotropy within the parent crystals,a sillustrated with the intergrowth model in Figure 2d. [13] Theobserved 2D zones of different crystallinity would not be present if the crystals would consist of as ingle homogeneous phase.
As ar esult of the described crystal anisotropy,c rystallographic phases within one zeolite crystal may be unevenly affected by steaming and result in distinctly different reactivity.During steaming, the outer Al-rich phase is more prone to dealumination than the inner Al-poor crystalline domains, which is the direct consequence of the local Al concentration that affects the dealumination rate. [51][52][53] In recent work, we have measured higher catalytic turnover rates in the inner regions of steamed ZSM-5 crystals and detected severe dealumination and clustering of Al atoms at the surface of the crystals. [54] It is important to note that the crystal lattice of ap arent ZSM-5 expands at the outer rim in both aa nd b directions due to Al zoning, [13] with the lattice parameters of a = 20.10 AE 0.02 a nd b = 19.92 AE 0.02 . Upon steaming these parameters change to a = 20.03 AE 0.02 a nd b = 19.93 AE 0.02 . A0 .07 c ontraction in the a lattice parameter implies ac rystallographic change and dealumination along the sinusoidal pores.HR-SEM and FIB-SEM studies by Karwacki, Aramburo et al. noticed higher susceptibility of sinusoidal pores towards steaming and the unidirectional nature of mesopores along the sinusoidal channels. [30,31] In summary,w eh ave demonstrated that hard X-rays can be used to acquire information from both X-ray and visible spectral regions when studying the impact of the crystalline structure and mesoporous defects on Brønsted reactivity of azeolite crystal in asingle X-ray shot. Thestudy demonstrates that the intra-particle differences in zeolite reactivity are determined by the underlying local crystalline structure and composition. Such important structure-reactivity relationships are difficult to derive from other characterization approaches;h ence the developed method has the potential to substantiate,s ynchronously,i ns pace and time,t he structural and reactivity properties of many other important functional materials.