Consequences of secondary zeolite growth on catalytic performance in DMTO studied over DDR and CHA

Abstract

Zeolites with DDR (Sigma-1 and ZSM-58) and CHA (SSZ-13) topology were synthesized by seed assisted and direct hydrothermal synthesis in order to investigate the effects of fast crystal growth on catalytic performance. Application of small amount of seeds (0.1% wt) significantly reduced synthesis time of all the studied zeolites. XRD and NH₃-TPD analyses did not reveal any difference in crystallinity and acidity. On the other hand, IR spectroscopy clearly demonstrates the presence of multiple defects, internal silanols (3729 cm⁻¹) and silanol nests (3400 cm⁻¹), as a result of accelerated crystal growth kinetics. Comparison of catalytic properties in the dimethyl ether to olefins (DMTO) reaction at 400 °C and 450 °C revealed that, despite smaller crystal sizes, zeolites prepared by secondary growth display shorter lifetimes due to faster coking rates, the latter being a result of silanols promoting hydrogen transfer reactions. Fluoride treatment of CHA removed silanols and prolonged its lifetime. This work highlights the importance of zeolite quality for catalytic application and the necessity to optimize current synthetic protocols based on secondary growth.

---

Introduction

Zeolites are microporous aluminosilicates widely employed in chemical industry as molecular sieves and catalysts.¹ Due to their unique properties such as ability to ion-exchange, thermal and chemical stability and shape selectivity; they are the working horses in many processes ranging from waste water treatment² to gas separation^3–5 and catalysis.^6,7 Fluid catalytic cracking (FCC), hydrocracking and methanol to hydrocarbons (MTH) are among the most important industrial processes utilizing zeolites as catalysts. In all of them, the active site is a proton – Brønsted acid – counterbalancing negative charge of zeolite framework due to the insertion of Al. However, catalysis by zeolites is not limited to Brønsted acidity. For instance, similar isomorphous substitutions with tetravalent metals such as Sn- or Zr-may act as Lewis acid sites to catalyse Baeyer–Villiger reaction,⁸ aldol condensation⁹ and many other reactions¹⁰ of interest.¹¹ Besides, zeolites can host different metals in the form of nanoparticles or as counter ions, such as in case of Cu-CHA for the selective catalytic reduction (SCR) of NOx.¹²

Given the relatively large number of known zeolite frameworks and their outstanding properties, one could easily argue that zeolites should be even more extensively used in large scale processes. However, issues associated with the high price of certain structure directing agents, large synthesis times and, in some cases, reproducibility issues, still hamper zeolite scale up.^8,13–16

In order to address most of these issues, the application of secondary (or seeded) growth has been proposed.¹⁷ The idea behind this method is rather simple – certain amount of crystals of the desired topology is added to the mixture prior to hydrothermal synthesis. If the amount of crystals is big enough (10–25 wt% relative to silica source) crystallization proceeds without structure-directing agent (SDA).¹⁸ On the other hand, even negligible amount of crystals (≈ 0.1%) significantly decreases the crystallization time in the presence of an SDA. For example, Yu et al.¹⁹ reported ultrafast synthesis of SAPO-34 in the presence of seeds in 10 min. The faster formation of products is attributed to the fact that addition of seeds suppresses the necessity of a nucleation step either by partial dissolution of seeds to provide growth surface for new crystals²⁰ or via a core–shell growth mechanism.²¹

Naturally, fast crystal growth might lead to the formation of some framework imperfections, which in turn can affect physical and chemical (read as catalytic) properties. Thus, the aim of this work is to compare structural and catalytic properties of zeolites obtained via seeded growth approach with the ones synthesized via direct hydrothermal synthesis. In this spirit, we synthesized a series of 8MR zeolites with two different topologies (DDR and CHA) using both methods and targeting the same morphology and acidity and assessed their catalytic properties in the dimethyl ether-to-olefins reaction. DMTO (and the methanol to olefins process, MTO) is very sensitive to parameters related to quality of crystals, such as Si and Al distribution,²² presence of defects,²³ and thus serves as an excellent probe reaction to assess even small differences in zeolite quality.

Experimental

Synthesis of zeolites

N,N,N-Trimethyl-1-adamantammonium hydroxide (TMAdaOH) was provided by SACHEM, Cab-O-Sil M5 was provided by Cabot. The rest of the chemicals were purchased from Sigma-Aldrich. All chemicals were used as received.

The detailed synthetic procedures for the preparation of zeolites with DDR topology (ZSM-58 and Sigma-1) was reported previously by our group.^24,25

SSZ-13 (CHA) was synthesized following the work by Zhu et al.²⁶ with some modifications. In a typical procedure, 25% solution of TMAdaOH was mixed with a solution of 0.077 g sodium aluminate and 0.32 g sodium hydroxide in 8.51 g of deionized water. When the solution became clear, 2.78 g of fumed silica Cab-O-Sil M5 and 21.43 g of deionized water were added and kept for aging for 4 h under the stirring. The obtained mixture was transferred to autoclaves and subjected to hydrothermal synthesis at 160 °C for 24 or 120 hours.

After crystallization, zeolites were thoroughly washed with deionized water and dried overnight. These solids were used as starting material for the preparation of seeds via ball-milling. The exact procedure can be found elsewhere.²⁴

Subsequent syntheses of zeolites with DDR and CHA topology were performed by seed assisted growth using the ball-milled crystals (0.1 wt% relative to silica source).

The following nomenclature based on synthesis type and framework type code (FTC) was used in the course of the current work: X – FTC, where X stands for type of synthesis (D – direct hydrothermal, S – seeded), while FTC is either DDR₁ (Sigma-1), DDR₂ (ZSM-58) or CHA (SSZ-13). Details on gel composition, synthesis time and nomenclature can be found in Table 1. The as-synthesized crystals were calcined for 10 h at 650 °C and converted to their protonic forms by triple ion-exchange in an aqueous NH₄NO₃ solution (1 M, 80 °C, 2 h, 100 mL per gram of zeolite) followed by calcination at 550 °C.

Table 1 Experimental details on zeolite synthesis procedure

Sample	Zeolite	Type of growth	Seeds	Gel composition	Time, h
				SDA : Na2O : SiO2 : Al2O3 : H2O
a Ball-milled.
D-DDR1	Sigma-1	Direct	—	20 : 3 : 60 : 1 : 2400	144
S1-DDR1	Sigma-1	Seeded	BM^a D-DDR1	20 : 3 : 60 : 1 : 2400	16
S2-DDR1	Sigma-1	Seeded	S1-DDR1	20 : 3 : 60 : 1 : 2400	16
D-DDR2	ZSM-58	Direct	—	15 : 10 : 60 : 1 : 2400	120
S-DDR2	ZSM-58	Seeded	BM^a D-DDR2	15 : 10 : 60 : 1 : 2400	16
D1-CHA	SSZ-13	Direct	—	20 : 10 : 100 : 1 : 4400	120
D2-CHA	SSZ-13	Direct	—	20 : 10 : 100 : 1 : 4400	24
S1-CHA	SSZ-13	Seeded	BM^a D1-CHA	20 : 10 : 100 : 1 : 4400	16

---

Fluoride modified S₁-CHA-F was obtained from S₁-CHA following the method described by Xu et al.²⁷ For a typical procedure, 1.2 g of S₁-CHA was dispersed in 20 ml of 0.013 M NH₄F solution and left stirring at room temperature until complete evaporation of water (approximately 48 h). The obtained solid was further dried 80 °C for 16 h followed by calcination at 600 °C for 6 h.

Characterization of zeolites

The XRD patterns of the powders were recorded in Bragg–Brentano geometry with a Bruker D8 Advance X-ray diffractometer equipped with a LynxEye position-sensitive detector. Measurements were performed at RT by using monochromatic CoKα (λ = 1.788970 Å) radiation between 2θ = 5° and 50°.

Images were recorded using a JEOL JSM-6010LA with a standard beam potential of 10 kV and an Everhart–Thornley detector. X-ray microanalysis (SEM/EDX) confirmed the elemental composition in the sample by scanning microscopy (SEM) coupled with a dispersive X-ray microanalysis system (EDX) with a silicon-drift detector.

A Malvern Nano ZS zetasizer was used for the analysis of particle size of the wet ball-milled crystals with dynamic light scattering (DLS).

N₂ adsorption was carried out using Tristar II 3020 Micromeritics sorptometer at 77 K. Prior to the experiment samples were outgassed at 350 °C for 16 h.

Temperature-programmed NH₃ desorption (NH₃-TPD) was measured by AutoChem II chemisorption analyzer (Micromeritics). Approximately 0.2 g of the material was first degassed under He flow at 400 °C and then saturated with NH₃ at 200 °C during 1 h using a flow of 1.65% NH₃ in He. The gas mixture was then switched back to He and the sample was purged at 200 °C for about 1 h to remove weakly adsorbed NH₃ molecules. TPD was subsequently recorded under He flow, from 200 °C to 800 °C. All flow rates were adjusted to 25 mL min⁻¹, and the heating rates were 10 °C min⁻¹ during different stages of experiment.

Transmission FT-IR spectroscopy was performed using a Nicolet Nexus spectrometer at 4 cm⁻¹ resolution equipped with an extended KBr beam splitting and a mercury cadmium telluride (MCT) cryo-detector. The pellets were placed in an IR quartz cell equipped with CaF₂ windows. A movable sample holder allows the sample to be placed in the infrared beam for the measurements or into the furnace for thermal treatments. The cell is connected to a vacuum line for pretreatment. The specimen is activated in vacuum at 400 °C for 16 h to remove adsorbed species.

Catalytic experiments

Dimethyl ether to olefins (DMTO) experiments were performed in a Microactivity Reference unit (PID Eng&Tech) at 400 °C and 450 °C and ambient pressure. For each experiment, 0.5 g of catalyst (pressed, crushed and sieved to particle sizes 250–420 μm) was placed in a fixed-bed with internal diameter of 9 mm. Conversion, selectivities and yields were calculated on a molar carbon basis. Detailed description of the set-up and calculations can be found elsewhere.^28,29

Dimethyl ether (DME) was chosen instead of methanol due to several reasons. Application of DME allowed to achieve very high dilution with nitrogen (DME : N₂ = 1 : 19) at the same time avoiding exothermic dehydration of methanol to DME, thus getting rid of heat effects intervention.²⁹ For a typical experiment, a mixture of 5 mol% of dimethyl ether (DME) in nitrogen was used at a weight-hourly space velocity (WHSV) of 1.23 g_DME g_cat⁻¹ h⁻¹. The product mixture was analyzed online with an Interscience CompactGC equipped with a 15 m capillary RTX-1 (1% diphenyl-, 99% dimethylpolysiloxane) column and a flame ionization detector.

Catalytic results are presented as a function of the DME throughput per amount of catalyst used (g_DME g_cat⁻¹) and defined as the total amount of DME fed to the catalytic bed before the conversion of DME dropped below 80%. Selectivities reported in Table 4 are integral values calculated during the active period of the catalyst. The typical evolution of ethylene and propylene selectivities as a function of time is shown in the ESI† (Fig. S7–S9).

Results and discussion

Results

Characterization of DDR and CHA zeolites. The cage structures of DDR (7.1 × 9.4 Å) and CHA (6.7 × 10.9 Å) are shown in Fig. 1. Both topologies belong to the 8-membered ring family (8MR) characterized by large cavities with small window openings. DDR has a 2D porosity, while CHA possesses a 3D pore structure.

Fig. 1 Cage structures of DDR (A) and CHA (B) frameworks. Arrangement of 8MR-joined 2D pore substructure layers of DDR in (C) spherical and (D) disk-like crystals.

Comparison of powder XRD patterns of the synthesized materials with their reference patterns³⁰ reveals that in all cases the desired topology was obtained and no crystalline impurities could be observed (Fig. S1†).

Notably, the experimental diffraction pattern of the sample D-DDR₁ shows around 13.3° higher diffraction intensity than expected from the theoretical XRD pattern revealing a preferred orientation of the particles. Consequently, the XRD pattern was subjected to a whole pattern refinement in order to figure out the habit of the crystallites and the direction of their preferred orientation (Fig. S2†). The refinement results reveal that the D-DDR₁ particles tend to orient preferably along the (011) plane. Moreover, the G₁-factor of 0.023 suggests a platy habit of the crystals, which has been confirmed by scanning electron microscopy. The D-DDR₁ crystals have a disk-like shape and grow preferentially along directions perpendicular to the (0kl) crystallographic planes developing a (011) crystal surface. Therefore, the top and bottom surfaces of the D-DDR₁ crystals preferably expose (011) lattice planes along with equivalent (0kl) ones, as represented in Fig. 1D. These 2D substructures lay in the (001) plane and intersect the top and bottom surfaces of the D-DDR₁ round slabs with inclination angle of 73°, as shown in Fig. 1D. Considering this structural arrangement within the crystallite, the diffusion through 8MR channel system is expected to be faster in the disk-like crystals than in spherical particles (Fig. 1C and D).

Analysis of SEM pictures shows that direct hydrothermal synthesis led to the formation of crystals with disk (Sigma-1), spherical (ZSM-58) and cubic (SSZ-13) shape with relatively broad size distribution (Fig. S3,†Table 2). On the other hand, seeded growth resulted in the formation of crystals with a narrow size distribution, all particles having similar crystal size (400–600 nm), independently of the topology (Fig. 2). DLS analyses of seeding solutions employed for the seeded growth synthesis show that in all cases they contained nanoparticles with an average size of 100 nm. These results suggest that the final crystal size obtained via seeded synthesis is more a function of the initial size of seeds than of the zeolite topology. Such interdependence is an indisputable advantage of the seeded growth technique as it allows tuning the particle size.

Table 2 Catalytic properties of the zeolites with the DDR topology under the study

Sample	Particle size^a (μm)	S meso (m^2 g^−1)	S BET (m^2 g^−1)	V micro (cm^3 g^−1)	SiO2/Al2O3^d (mol mol^−1)	NH3 capacity^d (μmol NH3 g^−1)
a From SEM images. b t-Plot method applied to the N2 isotherm. c BET method applied to the N2 isotherm. d From NH3-TPD measurements.
D-DDR1	12–15	20	250	0.10	111	295
S1-DDR1	0.4	30	320	0.13	107	306
S2-DDR1	5	9	300	0.13	103	318
D-DDR2	0.5–1.6	22	370	0.14	111	296
S-DDR2	0.5	17	310	0.13	113	290
D1-CHA	5–8	7	780	0.30	159	207
D2-CHA	0.3–1.4	20	770	0.30	168	196
S1-CHA	0.6–0.7	8	750	0.29	158	208
S1-CHA-F	0.6–0.7	18	540	0.23	—	—

---

Fig. 2 NH₃-TPD profiles of zeolites with DDR (A) and CHA (B) topology.

NH₃-TPD shows that all catalysts possess a similar amount of acid sites independently of the topology and synthesis protocol. For all zeolites, TPD curves are characterized by the presence of one desorption maximum at 410 °C (ZSM-58), 450 °C (Sigma-1) or 460 °C (SSZ-13). The desorption curves of ZSM-58 (D-DDR₂ and S₁-DDR₂) are also characterized by the presence of a shoulder at 280 °C that can be indicative for the presence of weaker acid sites (Fig. 3).

Fig. 3 Particle size distribution of seeds (A₁, B₁, C₁) obtained by DLS, and SEM images of zeolites (A₂, B₂, C₂) obtained from these seeds. A₂ – S₁-DDR₁, B₂ – S-DDR₂, C₂ – S₁-CHA.

Structural differences in the hydroxyl region were investigated by means of IR (Fig. 4). The whole spectra can be found in the ESI† (Fig. S4). For all samples, the two expected absorbances are always present at 3747 cm⁻¹, characteristic from the ν(OH) of terminal silanols from external surfaces and at 3616 cm⁻¹, associated to the ν(OH) of bridging silanols (Brønsted acid sites).³¹

Fig. 4 FTIR spectra in the OH-stretching region of activated Sigma-1 (A), ZSM-58 (B), SSZ-13 (C) and modified SSZ-13 (D) zeolites.

For D-DDR₁, the band at 3747 cm⁻¹ is less prominent, in good agreement with the relatively larger particle size of this sample. Zeolites Sigma-1 synthesized by seeded growth (S₁-DDR₁ and S₂-DDR₁) display an additional very broad band at 3400 cm⁻¹ attributed to silanol nests, progressively increasing from S₁-DDR₁ to S₂-DDR₁.

These silanol nests, arising from defects in the zeolite framework, represent clusters of silanols interacting via medium strength hydrogen bonding. These defects in turn can originate from removal of framework atoms or from incomplete condensation. Interestingly, the spectrum of S₂-DDR₁ is characterized by a complex triplet, where, together with the band at 3747 cm⁻¹, two additional vibrations at 3729 and 3706 cm⁻¹ are observed. These have been attributed to silanols in different environments but inside the micropores and terminating inner defects.^23,32 These OHs are slightly more acidic than the band associated with external silanols and are notorious for faster deactivation due to coke retention.²³

ZSM-58 zeolites display very similar features. For S₁-DDR₂ through secondary growth, a broad band is observed with two maxima at 3747 cm⁻¹, 3729 cm⁻¹ and additional overlapping bands pointing to multiple inner defects. The band at 3400 cm⁻¹ is also more prominent in comparison with D-DDR₂.

Some of the spectra are also characterized by the presence of the band at 3668 cm⁻¹ due to the presence of extraframework Al.

The IR spectra obtained from the SSZ-13 samples are rather complex. The Brønsted acidity region is characterized by a triplet with maxima at 3612 cm⁻¹, 3594 cm⁻¹ and 3584 cm⁻¹. Bordiga et al.³³ identified two bands in SSZ-13 with Si/Al = 11.6 at 3616 cm⁻¹ and 3584 cm⁻¹ and ascribed them to protons of the same strength. The band arising at 3616 cm⁻¹ was attributed to the proton located at the 8MR windows while at 3584 cm⁻¹ to the proton isolated from the neighbouring cages. The presence of the third peak in our case implies the presence of additional protons in a different geometric position, most likely within the cages. For all samples, silanols are recognised by a doublet at 3740 (isolated silanols) and 3729 cm⁻¹ and an additionally appearing band centred at 3706 cm⁻¹ for S₁-CHA. Silanol nests are present in all samples as a broad component at around 3520 cm⁻¹, but S₁-CHA is also characterized by a second broad band at around 3490 cm⁻¹. In contrast to parent S₁-CHA, fluoride-modified S₁-CHA-F (Fig. 4D) is characterized by a singlet at 3740 cm⁻¹ representing isolated silanols, while broad components (3490 and 3520 cm⁻¹) related to silanol nests completely disappear. The ratio of the bands in triplet (3612 cm⁻¹, 3594 cm⁻¹ and 3584 cm⁻¹) representing Brønsted acidity seems to remain unchanged, though band absorbance is lower.

Catalytic activity in DMTO. The different zeolites under study were tested in the DMTO reaction. We have recently shown that zeolites with DDR topology have the highest lifetime in MTO reaction at relatively high temperatures (>450 °C).²⁴ In case of SSZ-13, many studies show that, for this topology, milder temperatures (≈ 400 °C) are favourable for longer lifetime.³⁴ For this reason, two different temperatures (400 °C and 450 °C) were selected to allow for a meaningful comparison. Fig. 5 and Table 3 show the obtained lifetimes and related throughput, while conversion as a function of time is presented in the ESI† (Fig. S6). In line with previous papers dedicated to Sigma-1 and ZSM-58,^24,25 zeolites with DDR topology display an up to twice higher throughput at 450 °C in comparison with 400 °C. The fast deactivation of DDR zeolites at 400 °C was attributed to the formation of diamondoid species remaining in the zeolite cages and in contrast to methylbenzenes not reacting further with methanol.²⁴

Fig. 5 Catalytic lifetime of zeolites under study at 400 °C (A) and 450 °C (B) in DMTO reaction.

Table 3 Comparison of catalyst lifetime at different temperatures

Sample	t 80 (h)		Throughput (gDME gcat^−1)		Coke^a (wt%)	Coking rate (mg gcat^−1 h^−1)
	400 °C	450 °C	400 °C	450 °C
a From TGA measurements of deactivated at 450 °C catalyst.
D-DDR1	0.3	0.7	0.4	0.9	3.2	45.7
S1-DDR1	1.1	1.6	1.4	2.0	8.8	55.0
S2-DDR1	0.2	0.4	0.2	0.5	4.3	107.5
D-DDR2	1.3	1.7	1.6	2.1	8.4	49.4
S-DDR2	1.0	1.5	1.2	1.8	9.2	61.3
D1-CHA	2.6	2.6	3.2	3.2	11.8	45.4
D2-CHA	4.5	3.6	5.5	4.4	12	33.3
S1-CHA	3.1	2.0	3.8	2.5	10.5	52.5
S1-CHA-F	3.5	2.9	4.3	3.6	10.2	35.2

---

Interestingly, there is no direct correlation between catalyst lifetime and particle size. Large crystals (10–15 μm) of directly synthesized D-DDR₁ show longer lifetimes than much smaller S₂-DDR₁ at 450 °C (Table 3, 0.9 vs. 0.5 g_DME g_cat⁻¹). Accordingly, the nanosized S-DDR₂ with multiple internal defects displays slightly worse performance than its micrometer hydrothermally synthesized counterpart D-DDR₂. To demonstrate that the observed relation is not topology dependent, the catalytic behaviour of SSZ-13 was also investigated. First of all, in contrast to Sigma-1 and ZSM-58, SSZ-13, except D₁-CHA, show longer lifetimes at 400 °C. However, even at 450 °C there is a noticeable difference in lifetimes for DDR and CHA in favour of CHA. Such a prolonged lifetime of these catalysts is explained on the basis of the bigger dimensions and particular shape of the main CHA cage, which is able to accumulate more aromatics and is believed to stabilize “hydrocarbon pool” species.^35–37 Furthermore, there is a higher probability of blocking the 2D pore network of ZSM-58 and Sigma-1 than the 3D porosity of SSZ-13, leading to a faster deactivation of DDR. According to the literature,³⁸ the diamondoid species can be also formed in CHA cage however below 300 °C leading to the prompt deactivation as in case of DDR. Thus the different optimum temperatures for DDR (450 °C) and CHA (400 °C) can be related to the nature of active species formed at these temperatures and their participation either in aromatic cycle or coke formation. In turn, the formation of different aromatic species is defined by cage architecture, location and strength of the acid site.

The highest DME throughput (5.5 g_DME g_cat⁻¹) is observed for D₂-CHA at 400 °C followed by S₁-CHA (3.8 g_DME g_cat⁻¹). It is obvious that, despite minor differences in size and acidity, noticeable differences are found in DME throughputs. Moreover, D₁-CHA, whose lifetime seems to be temperature independent, outperforms S₁-CHA at 450 °C, clearly showing that the quality of the crystal is more important than its size. Fluoride modification seems to have positive effect on catalyst lifetime as it results in higher DME throughput for S₁-CHA-F in comparison with S₁-CHA, which is especially noticeable at 450 °C (3.6 vs. 2.5 g_DME g_cat⁻¹).

To summarize, an important tendency is found: zeolites prepared by seeded growth deactivate faster than directly synthesized zeolites. As the only substantial difference between these zeolites is the presence of internal silanols and silanol nests, faster deactivation is attributed to the presence of these features.²³

Product selectivities. Product selectivities observed at 400 °C and 450 °C in DMTO are shown in Table 4. Ethylene, propylene and butenes are the main products for both DDR and CHA topologies with some presence of higher hydrocarbons and paraffins. The amount of paraffins and higher hydrocarbons is especially noticeable for CHA catalyst (up to 10%), which has larger dimensions of cage and pore openings. For DDR zeolites, higher hydrocarbons are mostly formed on the external surface as they cannot pass the window opening. Independently on topology, higher temperatures lead to higher selectivities to ethylene and lower selectivities to propylene and butenes as a result of kinetic competition between the aromatic and olefinic cycle.²⁴ The maximum selectivity of ethylene for all catalysts under study (Fig. S7–S9†) coincides with the breakthrough of DME as a result of accumulation methylbenzenes and polycondensed methylated species, responsible not only for ethylene formation but also for deactivation.³⁹ Moreover, under the same conditions, the amount of produced ethylene on CHA catalysts is at least 10% higher than for Sigma-1 and ZSM-58. It is also important to mention that for DDR catalysts, due to very short lifetime and high rate of coking, the carbon balance is not always completely closed. Especially, D-DDR₁, S₂-DDR₁ at 400 °C exhibit high selectivities towards coke formation (>10%).

Table 4 Product selectivities obtained in DMTO reaction over zeolites under study at 400 °C and 450 °C

Sample	Selectivity, %
	C1–C4^a		C2^=^a		C3^=^a		C4^=^a		C5–C6^a
	400 °C	450 °C	400 °C	450 °C	400 °C	450 °C	400 °C	450 °C	400 °C	450 °C
a Integral values calculated during the active period of the catalyst.
D-DDR1	3	2	37	42	39	32	7	7	1	1
S1-DDR1	5	2	25	38	47	43	15	13	2	2
S2-DDR1	1	2	22	32	45	42	10	10	1	1
D-DDR2	4	3	23	39	48	41	15	11	2	2
S-DDR2	2	2	21	32	45	44	13	12	5	5
D1-CHA	7	4	42	50	35	31	12	10	5	5
D2-CHA	4	3	47	55	33	28	11	9	6	5
S1-CHA	6	3	40	51	37	32	13	10	4	4
S1-CHA-F	4	3	41	51	37	31	12	9	6	6

---

Though the method of zeolite preparation (direct vs. seeded) has a noticeable effect on catalyst lifetime, there is no pronounced relation between product distribution and method of preparation.

Discussion

Seeding was shown to be an appropriate way to significantly reduce synthesis time and a tool to control the particle size, which makes it very attractive from large-scale application perspective. Recently Mintova et al.⁴⁰ examined some common issues of zeolite crystal growth, including nucleation via seeding. Among the advantages of seeded growth authors also listed reduced formation of undesired phase and fast nuclei formation. Obviously, the weak point is that seeded growth approach can often lead to aggregation and intergrown crystals.^41,42 This aggregation seems to be correlated with the amount of seeds used. As zeolites in this work were synthesized with very low amount of seeds (0.1% wt), neither aggregation nor intergrowth was observed.

Valtchev et al. fairly mentioned that, although the seeded method is already widely used, the mechanism is barely known.⁴² However, the way this can be achieved is still a matter of discussion. Xiao et al. demonstrated that zeolites crystallize on the surface of seeds i.e. forming core–shell structures.^21,43 The crystal growth proceeds on partially dissolved seeds in alkaline media, but without dissolution to primary building blocks. An attempt to crystallize zeolite BEA only in the presence of primary building blocks resulted in an amorphous phase.²¹ On the other hand, Okubo et al. suggest that seeds are disaggregated into much smaller subunits that, in turn, provide surface for crystal growth, i.e. again through partial dissolution, however, herein building units play the key factor.⁴⁴ Authors showed further that it is possible to synthesize zeolites using seeds of other topology, which have common building units. The mechanisms proposed by Xiao et al.^21,43 and Okubo et al.²⁰ for the SDA-free synthesis of zeolites, in principle, hold well for this work. However, considering the very small amount of seeds used for zeolite synthesis in this work, we tentatively suggest that dissolution of seeds proceeds to much smaller aggregates in comparison to previously mentioned works, giving rise to nuclei which act as a core to build new zeolite crystals. With this little amount of seeds, application of SDA is necessary. The size of the formed nuclei would then depend on the initial size of the seeds and determine the final crystal size. Thus, the 100 nm seeds led to formation of crystals with particle size <500 nm. If the latter are used as seeds these crystals result in the formation of 2.5 μm particle, 2.5 μm seeds will form 5 μm and so on.

Considering that build-up of the crystal starts not from the elementary building unit but already from the pre-formed core, the formation of a framework with T-site vacancies can be anticipated. From this perspective, the secondary growth of building units can be compared – to a certain degree – with random Tetris® blocks which are not always matching the surface and thus leaving some “gaps” in crystal structure. These “gaps” or T-vacancies are saturated with protons from the nearest oxygens giving rise to the hydroxyl nests and/or other silanol defects.⁴⁵ Moreover, during the hydrothermal synthesis, T-vacancies present in seeds are not healed under synthesis conditions and are progressively accumulating in the newly synthesised zeolite crystal. For example, in case of S₂-DDR₁, the core part of the crystal is represented by much bigger fragments of partially dissolved seeds that already possess T-vacancies (as it was also synthesized by seed-assisted method). Thus, the observed increased amount of silanol nests is a consequence of defects accumulation.

The term hydroxyl nest represents the defect where a SiO₄ or AlO₄ tetrahedron is replaced with four hydroxyls. The structural defect can also exist in the form of geminal silanols = Si(OH)₂ or bridged vicinal silanols. It is important to mention that the defects seen via spectroscopic tools (NMR and FTIR mainly) are only the tip of the iceberg, as the majority of them cannot be observed directly. This is the consequence of continuous transformation of structural defects due to dehydroxylation, dehydration or deprotonation happening already at room temperature. The dynamic behaviour of these defects causes drastic changes in characteristic signals obtained by NMR and FTIR.⁴⁵

In general, there are many possible reasons leading to the formation of defects, dealumination being the most common but not the only one.¹⁰ Fernandez et al.⁴⁶ have shown that formation and location of silanol defects can be governed by the SDA during zeolite crystallization. In this case, defects are located in close proximity to the template. In this work, we highlight another origin of defects coming from a mismatching crystal growth.⁴⁵

The catalytic testing of synthesized zeolites in DMTO reaction revealed that zeolites prepared by seeded growth deactivate much faster than directly hydrothermally synthesized materials. In our previous work, we have already extensively discussed influence of defects on prominently deactivating Sigma-1 in the MTO reaction.²⁴ In addition, due to the preferential orientation of lattice planes with an inclination angle of 73° to the disk (Fig. 1D), the diffusion pathway for Sigma-1 is defined by the thickness of the crystal (≈ 100 nm for S₁-DDR₁). This dimension is at least 5 times smaller than for defect-free spherical D-DDR₂. Despite such five-fold shorter diffusion pathway in case of S₁-DDR₁, these catalysts show slightly shorter lifetime in DMTO reaction (Fig. 5), again highlighting the importance of crystal quality. This observation highlights that effect coming from structural defects overweighs crystal size effect. The increased amount of defects facilitates formation of internal coke, significantly affecting the rate of deactivation. This relationship between defects and catalyst deactivation is already a well-established feature explained by the so called “retention effect” of coke species.^23,47 Sazama et al. showed that defective sites in ZSM-5 facilitate hydrogen transfer reactions thus producing higher amount of paraffins and aromatics and leading to twice faster deactivation.⁴⁸ On the other hand, faster crystal growth kinetics in the presence of seeds, might also lead to different distribution of Al in the zeolite framework in comparison to zeolites synthesized by direct hydrothermal growth, thus affecting catalytic performance.⁴⁹ There are many parameters affecting Al distribution like pH, SDA, precursors and synthesis conditions.⁵⁰ According to Dedecek, “the greater heterogeneity of mixtures results finally in higher population of Al pairs and vice versa rather homogeneous up to clear up solutions lead to higher population of single Al atoms.”⁵⁰ Adapted to our case, there is a higher probability of Al-zoning for catalysts prepared by seeding as a consequence of starting from heterogeneous seed-containing synthesis solution. Formation of Al-rich regions in turn can lead to higher coking rates due to higher probability of olefins from the two nearest sites to form aromatics.^47,51 This effect should be further explored.

The comparison of selectivities obtained for defective and non-defective zeolites in this work does not reveal significant difference in product distribution. Logically, for 8MR zeolites, selectivities are rather governed by the cage and window dimensions.^35–37 On the other hand, enhanced hydrogen transfer reactions would speed up coke formation rate. Indeed, the analysis of coking rate (Table 3) expressed as the mass of coke formed during the active period shows that rate of coke formation is almost twice higher for seeded zeolites than for hydrothermally synthesized ones (compare D-DDR₁vs. S₂-DDR₁ and D₂-CHA vs. S₁-CHA).

The results obtained in this work highlight the importance of crystal quality, which turns out to be more important than particle size. Though seeded growth approach allows tuning crystal size on demand and significantly reduces synthesis time, its negative impact on DMTO (MTO, MTH, MTG) catalysis outweighs all listed advantages. Further optimization of zeolite synthetic procedures based on secondary growth approach is required. From this perspective, synthesis of zeolites in fluoride media is rather promising. It has been shown that with the assistance of fluoride, synthesized crystals possess less defects,⁵² while post-synthetic treatment with FeF₃/NH₄HF₂ selectively removes internal silanols from synthesized zeolites improving catalyst lifetime by a factor of ten.⁵³ Valtchev et al. compared synthesis of nano-ZSM-5 by seed assisted method in OH⁻ and F⁻ media, the latter possessing significantly less structural defects.⁵⁴ In this work, the post-synthetic fluoride treatment with NH₄F was chosen as a tool to heal defects in zeolite crystals. IR results clearly demonstrated that such treatment led to selective removal of silanol nests and internal silanols from zeolite crystal. Unlike for ZSM-5, fluoride treatment did not lead to the generation of mesopores in SSZ-13, however led to a noticeable surface area loss. Despite this loss in surface area (≈ 30%) and micropore volume (≈ 20%), S₁-CHA-F showed longer lifetimes in DMTO reaction and lower coking rates. One has to be aware that concentration of NH₄F had to be thoroughly optimized as further increase in F⁻ concentration led to removal of Brønsted acid sites from the zeolite framework (Fig. S5†).

Herein, we have further confirmed the negative impact of defects arising from seeded growth on DMTO performance. It is also important to mention that there are a number of reactions that are catalysed solely by defects. Abate et al. dedicated a paragraph in the recent perspective about the role of defects in such reactions like Beckmann rearrangement, etherification of 5-hydroxymethylfurfural (HMF) to produce, for instance, biodiesel components.¹⁰ Thus, the seeding method to produce defective zeolites can be beneficial for future applications in reactions catalysed by hydroxyl nests.

Conclusions

Zeolites with DDR (Sigma-1 and ZSM-58) and CHA (SSZ-13) topology were synthesized by direct hydrothermal synthesis and seed assisted method. Synthesized zeolites appeared to have similar amount of Brønsted acid sites. However, zeolites prepared by seeded growth contain a significant amount of internal silanols and hydroxyl nests due to the fast crystal growth kinetics. These multiple defects have a negative impact on catalytic performance in the DMTO reaction in terms of lifetime, but did not affect product distribution. Such deactivation behaviour is a consequence of the silanols participating in hydrogen transfer reactions and thus speeding up coke formation rate. These silanols can be selectively removed from by treatment with NH₄F. All zeolites under study exhibited high selectivities to ethylene, propylene and butenes (>90%), selectivities being dependent on reaction temperature (↑T → ↑S₂₌, ↓SC₃₌–C₄₌) and size of cage (↑size → ↑S₂₌, ↓SC₃₌–C₄₌).

Acknowledgements

This research received funding from the Netherlands Organisation for Scientific Research (NWO) in the framework of the TASC Technology Area “Syngas, a Switch to Flexible New Feedstock for the Chemical Industry (TA-Syngas)”.

Notes and references

1. J. Weitkamp, Solid State Ionics, 2000, 131, 175–188.
2. S. B. Wang and Y. L. Peng, Chem. Eng. J., 2010, 156, 11–24.
3. N. Kosinov, J. Gascon, F. Kapteijn and E. J. M. Hensen, J. Membr. Sci., 2016, 499, 65–79.
4. E. E. McLeary, J. C. Jansen and F. Kapteijn, Microporous Mesoporous Mater., 2006, 90, 198–220.
5. J. Gascon, F. Kapteijn, B. Zornoza, V. Sebastian, C. Casado and J. Coronas, Chem. Mater., 2012, 24, 2829–2844.
6. A. Corma, J. Catal., 2003, 216, 298–312.
7. K. Tanabe and W. F. Holderich, Appl. Catal., A, 1999, 181, 399–434.
8. A. Corma, L. T. Nemeth, M. Renz and S. Valencia, Nature, 2001, 412, 423–425.
9. J. D. Lewis, S. Van de Vyver and Y. Roman-Leshkov, Angew. Chem., Int. Ed., 2015, 54, 9835–9838.
10. S. Abate, K. Barbera, G. Centi, P. Lanzafame and S. Perathoner, Catal. Sci. Technol., 2016, 6, 2485–2501.
11. H. Bekkum and H. W. Kouwenhoven, Zeolite Manual for the Organic Chemist, 2012.
12. A. M. Beale, F. Gao, I. Lezcano-Gonzalez, C. H. F. Peden and J. Szanyi, Chem. Soc. Rev., 2015, 44, 7371–7405.
13. J. Gascon, J. R. van Ommen, J. A. Moulijn and F. Kapteijn, Catal. Sci. Technol., 2015, 5, 807–817.
14. C. Gücüyener, J. van den Bergh, A. M. Joaristi, P. Magusin, E. J. M. Hensen, J. Gascon and F. Kapteijn, J. Mater. Chem., 2011, 21, 18386–18397.
15. J. Gascon, W. Blom, A. van Miltenburg, A. Ferreira, R. Berger and F. Kapteijn, Microporous Mesoporous Mater., 2008, 115, 585–593.
16. C. Liu, L. Bai, J. M. Zhang, D. Hu, M. Li, G. F. Zeng, Y. F. Zhang, W. Wei and Y. H. Sun, Microporous Mesoporous Mater., 2016, 225, 312–322.
17. M. Moliner, C. Martinez and A. Corma, Chem. Mater., 2014, 26, 246–258.
18. Q. M. Wu, X. Wang, G. D. Qi, Q. Guo, S. X. Pan, X. J. Meng, J. Xu, F. Deng, F. T. Fan, Z. C. Feng, C. Li, S. Maurer, U. Muller and F. S. Xiao, J. Am. Chem. Soc., 2014, 136, 4019–4025.
19. Q. Sun, N. Wang, G. Guo and J. Yu, Chem. Commun., 2015, 51, 16397–16400.
20. Y. Kamimura, S. Tanahashi, K. Itabashi, A. Sugawara, T. Wakihara, A. Shimojima and T. Okubo, J. Phys. Chem. C, 2011, 115, 744–750.
21. B. Xie, H. Y. Zhang, C. G. Yang, S. Y. Liu, L. M. Ren, L. Zhang, X. J. Meng, B. Yilmaz, U. Muller and F. S. Xiao, Chem. Commun., 2011, 47, 3945–3947.
22. R. Martinez-Franco, Z. Li, J. Martinez-Triguero, M. Moliner and A. Corma, Catal. Sci. Technol., 2016, 6, 2796–2806.
23. K. Barbera, F. Bonino, S. Bordiga, T. V. W. Janssens and P. Beato, J. Catal., 2011, 280, 196–205.
24. I. Yarulina, J. Goetze, C. Gücüyener, L. van Thiel, A. Dikhtiarenko, J. Ruiz-Martinez, B. M. Weckhuysen, J. Gascon and F. Kapteijn, Catal. Sci. Technol., 2016, 6, 2663–2678.
25. Y. Kumita, J. Gascon, E. Stavitski, J. A. Moulijn and F. Kapteijn, Appl. Catal., A, 2011, 391, 234–243.
26. Q. J. Zhu, J. N. Kondo, R. Ohnuma, Y. Kubota, M. Yamaguchi and T. Tatsumi, Microporous Mesoporous Mater., 2008, 112, 153–161.
27. T. Xu, Q. Zhang, H. Song and Y. Wang, J. Catal., 2012, 295, 232–241.
28. I. Yarulina, S. Bailleul, A. Pustovarenko, J. R. Martinez, K. D. Wispelaere, J. Hajek, B. M. Weckhuysen, K. Houben, M. Baldus, V. Van Speybroeck, F. Kapteijn and J. Gascon, ChemCatChem, 2016, 8, 3057–3063.
29. I. Yarulina, F. Kapteijn and J. Gascon, Catal. Sci. Technol., 2016, 6, 5320–5325.
30. C. Baerlocher, L. B. McCusker and D. H. Olson, Atlas of Zeolite Framework Types, Elsevier Science, 2007.
31. S. Bordiga, P. Ugliengo, A. Damin, C. Lamberti, G. Spoto, A. Zecchina, G. Spano, R. Buzzoni, L. Dalloro and F. Rivetti, Top. Catal., 2001, 15, 43–52.
32. F. Thibault-Starzyk, A. Vimont and J. P. Gilson, Catal. Today, 2001, 70, 227–241.
33. S. Bordiga, L. Regli, D. Cocina, C. Lamberti, M. Bjorgen and K. P. Lillerud, J. Phys. Chem. B, 2005, 109, 2779–2784.
34. F. Bleken, M. Bjorgen, L. Palumbo, S. Bordiga, S. Svelle, K. P. Lillerud and U. Olsbye, Top. Catal., 2009, 52, 218–228.
35. Y. Bhawe, M. Moliner-Marin, J. D. Lunn, Y. Liu, A. Malek and M. Davis, ACS Catal., 2012, 2, 2490–2495.
36. J. W. Park, J. Y. Lee, K. S. Kim, S. B. Hong and G. Seo, Appl. Catal., A, 2008, 339, 36–44.
37. J. Z. Li, Y. X. Wei, J. R. Chen, S. T. Xu, P. Tian, X. F. Yang, B. Li, J. B. Wang and Z. M. Liu, ACS Catal., 2015, 5, 661–665.
38. J. R. Chen, J. Z. Li, Y. X. Wei, C. Y. Yuan, B. Li, S. T. Xu, Y. Zhou, J. B. Wang, M. Z. Zhang and Z. M. Liu, Catal. Commun., 2014, 46, 36–40.
39. E. Borodina, F. Meirer, I. Lezcano-Gonzalez, M. Mokhtar, A. M. Asiri, S. A. Al-Thabaiti, S. N. Basahel, J. Ruiz-Martinez and B. M. Weckhuysen, ACS Catal., 2015, 5, 992–1003.
40. J. Grand, H. Awala and S. Mintova, CrystEngComm, 2016, 18, 650–664.
41. N. Ren, J. Bronic, B. Subotic, X. C. Lv, Z. J. Yang and Y. Tang, Microporous Mesoporous Mater., 2011, 139, 197–206.
42. G. Majano, A. Darwiche, S. Mintova and V. Valtchev, Ind. Eng. Chem. Res., 2009, 48, 7084–7091.
43. B. Xie, J. Song, L. Ren, Y. Ji, J. Li and F.-S. Xiao, Chem. Mater., 2008, 20, 4533–4535.
44. K. Itabashi, Y. Kamimura, K. Iyoki, A. Shimojima and T. Okubo, J. Am. Chem. Soc., 2012, 134, 11542–11549.
45. A. A. Sokol, C. R. A. Catlow, J. M. Garces and A. Kuperman, J. Phys. Chem. B, 2002, 106, 6163–6177.
46. E. Dib, J. Grand, S. Mintova and C. Fernandez, Chem. Mater., 2015, 27, 7577–7579.
47. U. Olsbye, S. Svelle, M. Bjorgen, P. Beato, T. V. W. Janssens, F. Joensen, S. Bordiga and K. P. Lillerud, Angew. Chem., Int. Ed., 2012, 51, 5810–5831.
48. P. Sazama, B. Wichterlova, J. Dedecek, Z. Tvaruzkova, Z. Musilova, L. Palumbo, S. Sklenak and O. Gonsiorova, Microporous Mesoporous Mater., 2011, 143, 87–96.
49. N. Danilina, F. Krumeich, S. A. Castelanelli and J. A. van Bokhoven, J. Phys. Chem. C, 2010, 114, 6640–6645.
50. J. Dedecek, Z. Sobalik and B. Wichterlova, Catal. Rev.: Sci. Eng., 2012, 54, 135–223.
51. T. Liang, J. Chen, Z. Qin, J. Li, P. Wang, S. Wang, G. Wang, M. Dong, W. Fan and J. Wang, ACS Catal., 2016, 6, 7311–7325.
52. P. Tian, Y. X. Wei, M. Ye and Z. M. Liu, ACS Catal., 2015, 5, 1922–1938.
53. M. R. Li, Y. P. Zhou, C. Ju and Y. M. Fang, Appl. Catal., A, 2016, 512, 1–8.
54. Z. X. Qin, L. Lakiss, L. Tosheva, J. P. Gilson, A. Vicente, C. Fernandez and V. Valtchev, Adv. Funct. Mater., 2014, 24, 257–264.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6cy02307j

---