Transformation of chemically fine tuned zeolite A precursor into dense lithium aluminosilicates – A comprehensive phase evolution and sintering study

Microporous and Mesoporous Materials 135 (2010) 82–89

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso

Transformation of chemically fine tuned zeolite A precursor into dense lithium aluminosilicates – A comprehensive phase evolution and sintering study Kaliaperumal Selvaraj * Nano and Computational Materials Lab, Catalysis Division, National Chemical Laboratory, Council of Scientific and Industrial Research, Dr. Homi Bhabha Road, Pune 411008, India

a r t i c l e

i n f o

Article history: Received 7 May 2010 Received in revised form 23 June 2010 Accepted 24 June 2010 Available online 1 July 2010 Keywords: Zeolite Phase transformation LAS silicate Powder XRD Phase evolution

a b s t r a c t Beyond their conventional revelation as catalysts, zeolites are being perceived as more challenging materials for modern applications. Thermal recrystallization of zeolite precursors is an efficient method for the preparation of dense aluminosilicate ceramics. However, factors viz., nature of the precursor and the way of firing it etc., influence the product quality. The present report is a systematic and detailed account of a phase transformation of Li modified zeolite A precursor using powder X-ray diffraction (PXRD), thermogravimetry (TG), differential thermal analysis (DTA), scanning electron micrographs (SEM), atomic absorption spectroscopy (AAS), energy dispersive X-ray (EDX) and density measurement techniques. It involves an amorphisation followed by a recrystallization into lithium aluminosilicate (LAS). A comprehensive correlation of PXRD, TG/DTA, SEM, EDX and density data explains the complexity of high temperature phase transition (25–1200 °C). A qualitative phase analysis revealed the mediation of formation of LAS ceramic, b-spodumene by few satellite transitions of SiO2. Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction Many advanced materials fulfilling the property and reliability requirements for emerging applications demand innovative routes of materials syntheses. Ceramic research, in this regard has come up with many alternative precursors and processes. In this context, silicate based ceramic preparation using zeolite precursors is unique in many aspects [1–4]. The meta-stability and the microporosity of this class of highly ordered aluminosilicates were seen as versatile and potential precursors than just as conventional catalysts. Possibilities such as better control over chemical homogeneity, low temperature workability and cost effectiveness due to large availability for bulk production are few key attractions of this route. Though this route was reported [5] as earlier as 1971, it is still in a progressively developing stage [6–17]. As of now, there are various reports found on the preparation of electronic ceramics viz., cordierite (Mg2Al4Si5O18), anorthite (CaAl2Si2O8), mullite (Al6SiO18) and b-spodumene (LiAlSi2O6) using this route [18,19]. Through the conventional routes, the lithium aluminosilicate (LAS), b-spodumene was already known to be a highly successful commercial ceramic, for instance, in electronic industries as substrate materials. Nevertheless, its preparation using zeolite as precursor (zeolite route) is a subsequent development [20,21]. In this * Tel.: +91 20 25 90 22 62; fax: +91 20 25 90 26 33. E-mail addresses: [email protected], [email protected] 1387-1811/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2010.06.014

route, it is possible to fine tune the zeolite precursor for the desired stoichiometry by efficient means such as ion exchanges followed by firing at high temperatures into dense ceramic. It is observed that the precursor stoichiometry and firing approach have strong influences on the property of the final dense phase [21]. Moreover, the chemical homogeneity in the precursor influences the product quality. In the conventional routes, due to the physical mixtures of individual components, chemical homogeneity is possible only at particle level. In contrast, in the zeolite route, as ions are being exchanged it is highly assured at as lower as in atomic level. Contrasting the conventional precursors, the phase transformation course on firing the chemically altered zeolite precursors is distinct due to the enhanced chemical homogeneity. Reports suggest that the phase transformation is considerably complex with various intermediate phases depending upon the nature of the precursor [4,5,21,22]. The phase purity and many other properties such as density of the ceramic essentially decide a candidature in respective applications. A correlated and comprehensive information about the course of the phase evolution and sintering in tandem is though important for a better synthesis control, is however not available in the literature. In this way, the present article reports a systematic and comprehensive study on the phase transformation of zeolite A precursor into lithium aluminosilicate (LAS) ceramic, b-spodumene. Several meta-stable phases formed during the phase transformation between room temperature (RT) and 1200 °C have been

83

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

T2

Temperature, °C

identified and qualitatively analyzed using the standard data files. The powder X-ray diffraction (PXRD) patterns recorded for samples fired at different temperatures revealed the thermodynamics of different intermediate phases. To understand the state of the material on its course of sintering, thermogravimetric (TG) and differential thermal analysis (DTA) data, scanning electron micrographs (SEM), atomic absorption spectroscopy (AAS), energy dispersive X-ray spectra (EDX) and apparent densities for samples heated at various temperatures were obtained and successfully correlated. A comprehensive analysis using these data in conjunction with that of the PXRD provided valuable clue on the onset and the coexistence of different high temperature (HT) phases and their transformations.

Isothermal sintering Natural cooling

T1

Binder burning

t1 t2 2. Experimental

t3

t4

Time, min

Fig. 1. Typical heating program for sintering of LASA3.

The chemical fine tuning of the precursor zeolite was carried out using conventional hydrothermal solution based ion exchange procedure widely used for post-synthesis zeolite modifications as following. About 30 g of as-synthesized form (Na-form) of zeolite A (from Zeolyst Corporation, UK with Si/Al ratio = 1:1) powder was suspended in about 300 ml of 1 M NH4NO3 solution and ion exchange was carried out under constant stirring conditions for about 6 h at 95 °C. This was repeated for three times to get the NH4-form of zeolite A. Later, the filtered powder was dried at 110 °C for 12 h. Similar ion exchange process was repeated on the dry NH4-form of zeolite A with 1 M LiNO3 solution for three times. The complete procedure for multiple ion exchange processes have been explained elsewhere in detail [21,22]. The final dry powder, i.e., Li-form of zeolite A, labeled as LASA3 was used for the further heat treatment and other characterisations. To consolidate the free dry powder, a binder solution of 1 wt.% polyvinyl alcohol (PVA) was prepared as discussed below. About 10 g of PVA [(–CH2CH(OH)–)n of molecular weight 1,00,000 Da] was mixed with 90 ml of boiled distilled water and was kept on flame for 20–30 min with continuous stirring so as to get a clear solution (10% wt./wt.). It was further diluted to 10 times and used as binding agent. The dry LASA3 powder was added with the PVA binder and mixed in an agate mortar followed by drying under IR lamp for about 2 h. The powder was then consolidated into uniform sized pellets with a diameter of 20 mm and a thickness of 5 mm each. Each pellet was made using equally weighing (0.75 g each) powder of LASA3 pressed in an isostatic hydraulic press at a pressure of 5000 psi. These pellets were subjected to heat treatment in an open coiled programmable (Nebourtherm) furnace at air atmosphere. The furnace could hold a set temperature with a sensitivity of ±0.5 °C. The samples were heated at a heating rate of 4 °C/min as per the program shown in Fig. 1. To ensure error free analysis, multiple sampling was ensured by studying a set of at least three pellets for each temperature point simultaneously (T2 as seen in Fig. 1). Several temperature points (T2) were chosen between room temperature (RT) and 1200 °C with an interval of 100 °C, viz., 100, 200, 300, 400. . . 1200 °C. Each set of pellets were held at 525 °C (T1) for about 60 min (t2  t1 as seen in Fig 1) to burn off the binder (PVA). Subsequently they were raised to their intended final temperature point, T2 and soaked for 6 h (t4  t3). The fired pellets were later cooled naturally to RT and crushed into fine powder in an agate mortar. The structural details due to the phase transitions at different temperatures were observed using PXRD patterns recorded in a RIGAKU DMAX III VC instrument equipped with a graphite crystal monochromator and NaI scintillation counter. Nickel filtered Cu Ka radiation (k = 1.542 Å) was used. A powder of standard Si was used as an internal standard for a portion of few samples for lattice parameters calculations and for other samples it was used as external standard. The 2h values

and interplanar ‘d’ spacings were corrected accordingly and refined using least square fitting. A qualitative phase analysis was carried out for the XRD patterns recorded at each temperature point (T2) to identify and estimate various intermediate phases. The reflections were indexed and compared with that of the different crystalline phases as available in the International Centre for Diffraction Data base (ICDD/JCPDS) [23]. Thermogravimetric analysis (TGA) was performed using about 35 mg of LASA3 powder using SETARAM TG DTA 92 instrument in air atmosphere at a heating rate of 5 °C/min. The data was recorded for a range of temperature from 25 °C (RT) to 1000 °C. The derivative data (DTG) and differential thermal analysis (DTA) data were analyzed and plotted. Scanning electron micrographs (SEM) were captured for the samples using LEO/Leica Cambridge, UK stereoscan 440 instrument. Energy dispersive X-ray spectra (EDX) were obtained using BRUKER EDS systems QUANTAX/200 attached with SD detector interfaced with the SEM instrument. EDX spectra were recorded for selected samples to understand the changes in the chemical composition of the materials due to heating. The elemental analysis was done using CHEMITO 201 atomic absorption spectrometer to understand the extent of ion exchanges. The fired pellets after cooling (before crushing into powder for the XRD based phase evolution study), were weighed in a high precision single pan balance with accuracy up to 104 g. The thickness and diameter of each pellet were measured (as an average of five different measurements at different places in the disc) by micrometer screw gauge. These data were used to calculate the apparent density. The comparison between the density before firing (green density) and of the sintered discs (sintered density) was considered as indices of the extent of sintering. The apparent densities of the samples were calculated as follows. Average thickness (t) and the average diameter (d) of the carefully lapped sintered discs of uniform thickness were measured and the volume (V) was calculated considering the sample disc as a cylinder using the equation,

V ¼ pr 2 t

ð1Þ

where r is the radius of the disc (d/2 = r). The weights of the discs (w) were calculated on single pan balances to the precision up to 104 g. The apparent densities of the samples were calculated using the relationship,

Apparent densityðqapp Þ ¼ VolumeðVÞ=MassðwÞ

ð2Þ

The density changes at different temperature points during firing were plotted as function of temperature to understand the extent of sintering.

84

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

3. Results and discussion 3.1. The phase transformation Generally zeolites with low Si/Al ratio such as zeolite A are known to experience loss in crystallinity (partial structure destruction) due to possible dealumination during the post-synthesis treatments such as ion exchanges. However, if thermal and mechanical shocks are avoided by careful treatments, dealumination may be significantly reduced. Fig. 2 shows PXRD pattern of zeolite A sample after three times exchanges with NH4+ ions (labeled as NHA3) and that after three times exchanges with Li+ ions (labeled as LASA3) and those samples heated at different temperatures as LASA followed by the corresponding temperature (e.g., LASA1200 is for LASA heated at 1200 °C). The same nomenclature is followed through out the paper. Elemental analysis by AAS revealed that after three Li exchanges the LASA3 samples has 1.43 wt.% of Li and 0.051 wt.% of Na while the Si/Al ratio has no significant change due to ion exchanges. This indicates the successful exchange of Li into the Na positions of zeolite A. The PXRD patterns show that even after six times of repetitive ion exchanges each for about 6 h with intermittent overnight drying, zeolite A (LASA3) has no significant loss of crystallinity. This assured that the precursor powder is highly crystalline. Fig. 3 shows the multiple plot of PXRD profiles of LASA3 samples fired at different temperatures up to 1200 °C. The patterns labeled as 600 and as 1200 °C in Fig. 3 refer to the significantly intact

crystalline zeolite and the totally transformed dense ceramic phase respectively. The overall course of changes as seen in the PXRD multiple plot may be treated as four major regions (namely A, B, C and D) bordering around 600 °C. Region A (RT to 600 °C) shows that the crystalline zeolite phase being intact until 600 °C and hence this region has least importance in the context of phase transformations. Beyond that, however, the meta-stable zeolite phase sets on an acute loss in crystallinity collapsing down into exclusively amorphous at 700 °C. In the second region B (600– 800 °C), the re-condensation of the amorphous mass into crystalline phases and a simultaneous sintering begins. Region C undergoes complex changes between 800 and 1000 °C (b-quartz to keatite). The PXRD recorded at each step of 100 °C shows that there is more than one crystalline phase strongly overlapping with each other’s reflections. Region D (1000–1200 °C) involves the final densification of the structural transformation from keatite to bspodumene. Fig. 4a shows a narrow 2h range (20–40°) of the PXRD profiles where most of the 100% (I/I0) peaks of the major intermediate phases are seen in the regions B, C and D, (800–1200 °C). Various thermodynamically meta-stable phases in this range of temperatures were identified and compared with the reflections of the respective ICSD files as listed in Table 1. The reflections of individual phases are assigned with labels (Fig. 4a) such as b, M, K, HQ and V for b-spodumene, mullite, keatite, high-quartz, and virgilite respectively along with their respective ICDD references (Table 1).

1000

Intensity, cps

750

LASA3 500

250

NHA3 0 10

20

30

40

50

2-theta, degree Fig. 2. PXRD profiles of samples NHA3 (NH4+ exchanged zeolite A) and LASA3 (Li+ exchanged zeolite A) after three ion exchanges.

Fig. 4a. A PXRD multiprofile (truncated range of 2h = 20–40°) of LASA3 heated at different temperatures in the range from 800 to 1200 °C. Prominent peaks of the phases are assigned as listed in Table 2.

Table 1 Various phases identified during the phase transformation of precursor LASA3 to LAS ceramic b-spodumene. Temperature (°C)

Phases present

Label (as seen in Fig. 4a)

Ref. [23] and ICSD code

800

Mullite Virgilite SS High-quartz b-Spodumene Mullite Virgilite SS High-quartz b-Spodumene Mullite Keatite b-Spodumene Mullite Keatite b-Spodumene b-Spodumene

M V HQ b M V HQ b M K b M K b b

[23a] – 00-079-1276 [23b] – 00-031-0707 [23c] – 01-012-0708 [23d] – 00-035-0797 [23a] – 00-015-0776 [23b] – 00-031-0707 [23c] – 01-012-0708 [23d] – 00-035-0797 [23a] – 00-015-0776 [23e] – 00-012-0708 [23d] – 00-035-0797 [23a] – 00-015-0776 [23e] – 00-012-0708 [23d] – 00-035-0797 [23d] – 00-035-0797

900

1000

1100

Fig. 3. PXRD multiple plots of LASA3 samples heated at different temperatures (provided on each pattern) from 600 to 1200 °C.

1200

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

3.2. The phase evolution 3.2.1. Structural phase analysis A fundamental chemical transformation of the precursor, zeolite A into the LAS ceramic dense phase can be represented as seen in Table 2. The LASA3 precursor composition at room temperature may be generally represented as Li12Al12Si12O48. But for mass transfer through processes such as decomposition or oxidation, in principle, the chemical composition of the precursor should not change during the entire course of firing. However, the observations reveal that it dynamically varies due to new phase formations and changes. It also suggests that the formation of eventually predominating phase, LAS ceramic, b-spodumene is not a single step. The following describes different intermediate phases that form and co-exist in range of temperature from 800 to 1100 °C. The early formation of an LAS intermediate solid solution phase viz., virgilite SS and its transformation into the b-spodumene prominently stretches over the whole range of temperature above 800 °C. The silica polymorphs such as high-quartz and keatite are observed to mediate this conversion. The sample heated at 1200 °C was identified to be LAS ceramic, b-spodumene and PXRD reflections data obtained was matching with the standard data [23]. It was observed to be in tetragonal symmetry and the unit cell parameters for b-spodumene were refined using least square fitting programs. The refined unit cell parameters of the precursor and the final LAS ceramic phase, as shown in the Table 3 were found to have good agreement with that of the literature [23]. However, the co-existence of different thermodynamically less stable satellite phases during the conversion into the LAS phase is interesting to study. The evolution of the intermediate phases may be understood more easily through a stereoscopic view of the case; one view is as an evolution of a single component silicate (SiO2) primary system that exhibits polymorphism. On the other hand, the second view, as that of the silicate system in presence of other single or mixed oxides as aluminum and lithium viz., binary and ternary oxide systems. As a single component system, silica (SiO2) is known to have many polymorphs such as a-(low-) quartz, b-(high-) quartz, tridymite, and cristobalite that are stable at temperatures viz., 573, 870, 1470 and 1713 °C respectively at a pressure of 1 bar. Initially it is observed that certain overlapping reflections were closely matching with that of high pressure phases such as coesite and stishovite. They are known to form at

85

high pressure (>40 kbar) firing that drastically brings down the transformation temperature by several hundred °C. However, a weighted differentiation intensity data during the qualitative phase analysis confirmed their absence as LASA3 samples were essentially fired at ambient pressure. It is observed that the case is not simply of an individual single component (SiO2) system; rather a mixture of primary SiO2 and also a ternary system viz., LiO2–Al2O3–SiO2 (LAS). However, the simultaneous single phase primary transformation of pure SiO2 plays an interactive role in the overall transformation of the LAS system. Literature shows that the differences in the polymorphism of SiO2 with or without the presence of other oxides were not clearly known earlier [24,25]. However, in presence of other cations such as Al and Li, the highquartz phase is observed to preferentially form a separate family of aluminosilicate polymorph namely keatite with a structural formula as MAlSi2O6 [23]; where the positions of M may be accommodating element such as Li. It is important to note that keatite is known to co-exist with either quartz or cristobalite or both [28]. It is interesting to note that the intermediate LAS phases formed during the transformation process have structurally close chemical formula with b-spodumene (LiAlSi2O6). The conversion between such closely resembling iso-structural phases can be perceived as a dynamic variation in the stoichiometric chemical formula, viz., LixAlxSi3xO6 (0 < x < 1). The variable, ‘x’ constantly changes from 0 to 1 the transformation progresses towards final dense LAS phase. As the b-spodumene phase is fully formed, ‘x’ turns to be 1. A close relevance to this may be noted with the Evan’s preliminary report [26] on the phase transformation mechanism however, for a non-zeolite precursor into LAS dense phase. Li reported [27] later that during the conventional formation of LiAlSi2O6 phase (in the ternary system viz., Li2O–Al2O3–4SiO2), the high-quartz phase reconstructively transforms into the keatite phase at elevated temperature. However, the present observation shows that such thermodynamic routes are independent of the precursors. Because, the ternary oxides systems based on physical mixing of constituent oxides are usually expected to have diffusional difficulties of the individual atoms during the phase formations. However, as mentioned earlier, zeolite precursors have high chemical homogeneity to atomic level and expected to be better in coming over such diffusional problems during firing. In the present case, the ease of diffusion could better be explained due to the fluxing nature Li cations at high temperature. The entire course of phase transformation has been illustrated in the Fig. 4b. This provides an over all picture of the formation

Table 2 The overall compositional transformation of the zeolite precursor to LAS phase. Zeolite

Formula

M+

LASA3

Na12Al12Si12O48xH2O

Li+

Composition Precursor (RT)

Ceramic LAS (HT)

Li12Al12Si12O48

LixAlxSi3xO6 (0 < x < 1)

Table 3 Lattice parameters calculated for the zeolite precursor and the dense phase, bspodumene from the PXRD data. Phase

Precursor zeolite

Ceramic LAS

Symmetry (sample)

Lattice parameters Literature (Å)

Observed (Å)

Standard deviation

Cubic (LASA3) Cell volume (Å3) Tetragonal (b-spodumene) Cell volume (Å3)

a = 11.919 – v = 1693.24 a = 7.510 c = 9.208 v = 1552.92

a = 11.9106 – v = 1689.67 a = 7.5061 c = 9.2023 v = 1550.27

±0.0126 – ±6.12 ±0.0101 ± 0.0166 ±5.4

Fig. 4b. An illustration of the evolution of various high temperature phases during the heat treatment of LASA3 from room temperature to 1200 °C. Temperature range has been divided into regions of 200 °C each as discussed in the text.

86

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

and co-existences of different phases in the whole range of temperatures from room temperature (RT) to 1200 °C. Regions A and B are simple with crystalline zeolite and amorphised silica phases respectively. The regions C and D are complex with the formation of several phases such as mullite (M), virgilite (V), high-quartz (HQ), keatite (K) and b-spodumene (b). HQ that shows up after 800 °C has hexagonal symmetry while the final phase, b-spodumene is tetragonal. The transition between these two structures are mediated through satellite phases of SiO2 and b-spodumene namely keatite and virgilite respectively. They go through structural transformations with isomorphous substitutions of various cation positions and it spreads over a temperature range between 800 and 1100 °C. After that b-spodumene emerges as a single largest stable phase. A systematic observation of the entire transformation are comprehensively dealt in the following sections with special focus on the structural, thermal, gravimetric and density modifications in detail.

3.2.2. Thermal phase analysis Unlike in the case of the PXRD profiles (Fig. 4a) that were collected at different temperature points with an interval of 100 °C, the exact onset and fall of a specific phase were revealed by the thermogravimetric analysis (TG) data. The results of the derivatives of TG (DTG) and a differential thermal analysis (DTA) are shown in Fig. 5 (exothermic direction is upwards). The data provided detailed thermodynamic information about the various changes associated with weight losses as discussed below. About 17% weight loss that was observed within the range of room temperature to 300 °C with a rate maximum at 183 °C indicates the loss of occluded water (dehydration) present in the highly porous zeolite A precursor. The dip observed in the DTG data along with a fairly noticeable hump in the DTA data around 325 °C indicates the endothermic removal of NH3 (de-amination) from the un-exchanged NH4-sites of zeolite A, converting them into terminal hydroxyl groups that leads to creation of Brönsted acid sites (H+ ions). From 300 to 700 °C, a slow and gradual weight loss that accounts to about 3.5% is observed during the amorphisation of the LASA3. However, between 500 and 600 °C there is a shallow dip in the DTG along with a spread-out endothermic hump indicating the removal of binder, PVA. The DTA plot shows a significantly sharp endothermic peak (P1) at 800 °C accompanied by an insignificant dip in the DTG plot indicating a meager weight loss (>1%) connected to a sharp absorption of energy with a maximum at 789 °C. This when compared with the PXRD data recorded for the sample heated at 800 °C, indicates that there are two main processes happening in the same temperature range. (i) The dehydroxylation of LASA-OH (Scheme 1). This involves the bridging of

Fig. 5. Thermogravimetry data of LASA3 sample showing the TG/DTG and DTA curves. P1 and P2 are two exothermic peaks due to satellite phase transitions.

LASA3-[O-NH4]-nH20

LASA3-[O-NH4] + n H2O

II. Deamination :

LASA3-[O-NH4]

LASA3-O-H + NH3

III. Dehydroxylation :

LASA3-[OH]

LASA3-O-LASA3 + H2O

I.

Dehydration :

Scheme 1. Different steps of thermal decompositions of lithium modified zeolite A (LASA3) that contains un-exchanged NH4 ions in the extraframe work alkali postions.

dangling Si–O(H) bonds to form fresh Si–O–Si linking bonds which is invariably endothermic and indicated by the small weight loss. (ii) Along with this, the recrystallization of mullite and high-quartz (HQ) phases from the amorphous SiO2 that does not bring in any significant changes in the mass (weight) as seen in the TG/DTA plot. However, in the present case a huge amount of energy for the crystallization is indicated by a sharp and prominent endothermic peak, P1 as seen in the Fig. 5. As listed in the Table 2, virgilite and b-spodumene are though observed to co-exist, mullite and HQ are the dominating phases at this temperature range i.e., between 800 and 900 °C. It is observed that the conversion of high-quartz to keatite is mediated through an intermediate phase, virgilite SS (LAS solid solution). Virgilite phase is known to have a range of varying stoichiometry as the occupancies of its isomorphous sites are due to constantly ‘hopping’ cations as known in literature [27,28]. The second endothermic peak, P2 observed in the DTA curve is not as sharp as P1. However, it is very intense suggesting a prominent phase transformation; P2 is spread over from 855 to 913 °C (more than 50 °C) with large area under the peak suggesting the possibility of more than one process that may occur consecutively or simultaneously. This is confirmed by the changes in the PXRD profile with the evolution of the well-developed b-spodumene phase and the disappearances of other satellite phases such as HQ, virgilite.

3.2.3. Satellite phase mediation The instant appearance of the second and third most prominent peaks of b-spodumene at 2h = 22.7 and 28.19° respectively and the shift of the main peak from 2h = 26.11° to 25.53° unambiguously proves fall of mullite phase and the raising and strong dominance of b-spodumene as a single and strong crystalline phase beyond 1000 °C. It may be interesting to note that the regions A, B and C (up to 1000 °C) may be compared with the known three major adiabatic changes as seen in Scheme 1 [29]. Though the overall transformation is dominantly shadowed by the presence of the 100% (I/I0) peak of mullite at 2h = 26.2°, a closer follow up of the shifts in the peak positions among that of HQ and keatite phases as provided in Table 4 helps to understand the clear transition in the phase. The transformation between HQ and K is known to be so subtle in terms of XRD reflections. Many atoms in the framework undergo meager positional changes and hence, in many cases, it is difficult to be observed in even thermal techniques such as differential thermal analysis (DTA) [27]. Table 4 shows the changes in the peak parameters of the three main peaks of the two phases viz., HQ [reflections hkl = 1 0 0 (d = 4.518 Å), 1 0 1 (d = 3.482 Å) and 1 1 2 (d = 1.887 Å)] and keatite [reflections hkl = 1 1 1 (d = 4.608 Å), 2 0 1 (d = 3.486 Å) and 4 0 0 (d = 1.885 Å)] as the temperature varies in the regions C and D, from 800 to 1200 °C. The superscript marked such as ‘HQ’, ‘K’ and ‘b’ on the peak parameters stand for high-quartz, keatite and for b-spodumene (dense LAS) respectively. The co-existence of these phases and the precise shifts in their characteristic peak parameters at each temperature point (T2) were confirmed using a thorough profile fitting performed on the PXRD data. For e.g., Fig. 6 shows such profile fitted data for a narrow 2h range from 25° to 26.5° which includes four peaks belonging to phases such

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

87

Table 4 Phase transformation of high-quartz to keatite at high temperature (peaks of phase such as high-quartz, keatite and b-spodumene phases are superscripted with the symbols as HQ, K and b, respectively).

as b-spodumene, keatite and mullite (indicated in Fig. 6 as b, K and M respectively). The shifts in the parameter as seen in Table 4, clearly indicate the gradual transformation from HQ to b-spodumene through intermediate phase, keatite. The transformation of HQ to keatite is observed in the temperature range between 800 and 1000 °C (region C) and that of keatite to b-spodumene in the range between 1000 and 1200 °C (region D) as observed in the PXRD profile (Fig. 4a).

3.3. Sintering and density modifications The phase transformation and the sintering occur simultaneously and the sintering may be closely followed by observing the changes in the density of the materials. The phase analysis as understood through PXRD and thermal analysis was found to have close correlation with their densities measured at each temperature point (T2). Fig. 7 shows the plot of the apparent densities calculated for the samples heated at different temperature points. Between the room temperature and 600 °C, there is no significant change in the density. Most of the densification process was observed in the regions B, C and D i.e., between 600 and 1200 °C and is observed to be nonlinear. The differences in the degree of sintering for each temperature regions (B, C and D) of 200 °C each to the total sintering seems to be different. The apparent density (qapp) of the zeolite phase (at 600 °C) was 1.072 g/cc that rose to 1.720 g/cc at 800 °C. It may be noted that at this temperature range (region B), the highly porous (coarse) zeolite phase, after a total structure collapse, recrystallizes into dense HT phases such as high-quartz, mullite and other LAS dense phases (refer Table 2).

Fig. 7. Plot of apparent density measured for the sample fired at different temperatures (from 600 to 1200 °C) as a function of temperature.

This is the region where about 87.8% of the whole densification occurs. This is attributed to the disappearance of micro-pores and cages of zeolite on amorphisation and dense crystalline phases evolve. Region C demonstrates a small change of 5.4% of the sintering due to insignificant growth of the dense phases and there is no prominent phase change. Region D, i.e., between 1000 and 1200 °C contributes to 5.0% of the overall sintering. However, the density difference between 1000 and 1100 °C is corresponding to 4.9% and that between 1100 and 1200 °C there is nearly constant. The change of 4.9% is attributed to the satellite transformation from keatite to LAS ceramic, b-spodumene as mentioned earlier. The plateau observed in the density curve (Fig. 7) after 1100 °C is the evidence for the fact that most of the phase modifications are completed and the 99.9% of sintering has been achieved by 1100 °C. The rise in the density during the final 100 °C range adds a mere 0.1% to the whole densification process (qapp at 1200 °C = 1.81 g/cc).

3.4. Morphological changes

Fig. 6. A portion of the profile fitted PXRD pattern indicates the 100% peak (I/I0) of the three phases b-spodumene (b), mullite (M) and keatite (K) that co-exist at temperature, 1100 °C. Labels of the phases are referred in Tables 2 and 4.

Fig. 8 shows a set of scanning electron micrographs (SEM) of samples sintered at different temperatures viz., RT, 600, 800, 1000, and 1200 °C (Fig. 8a–e respectively). Fig. 8a shows the well defined crystals of zeolite A at room temperature proving that even after multiple exchanges the topology is retained yet. The shape of the cubes reflects the inherent cubic crystal structure of the zeolite A (LTA). The average size of the particle measured is about 1.7 lm. The cubic morphology of the zeolite precursor shows no significant change due to solid consolidation by pressing with binder. However, the negligible loss of morphology observed may be due to mechanical grinding of the pressed pellets into powder for different characterizations such as PXRD, SEM etc., Fig. 8b shows the coalesced amorphous mass observed as reduced finer particles of the collapsing zeolite phase at 600 °C. At higher 800 °C (Fig. 8c)

88

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

Fig. 8. Scanning electron micrographs of unfired LASA3 and that fired at different temperatures viz., (a) RT, (b) 600, (c) 800 (d) 1000 and (e) at 1200 °C. Bottom right shows the EDX spectra for LASA at RT (labeled LASA0030 °C) and for LASA at 1200 °C (labeled LASA1200 °C).

due to the thermal recrystallization of the phases such as highquartz and mullite as observed in PXRD profiles, the formation of bigger particles (2–3 lm) along with the agglomerated mass are observed. Fig. 8d is for the samples heated at 1000 °C and the satellite transformations observed in this range of temperature result in a mixture of agglomerated mass and bigger particles that co-exist as seen in the electron microgram. The consolidation of the powder using PVA binder is believed to have helped the sintering that is reflected in the increment in the average particle size. Though the PXRD reveals the existence of significantly crystalline materials at 1000 °C, in SEM, however it looks highly polycrystalline. Fig. 8e shows the SEM of the sample heated at 1200 °C. It shows the particles with better size distribution though polycrystalline. From the instantly augmented bigger particles and increasing compactness of the particle surface as seen in Fig. 8b and c, it is evident that at temperatures above 600 °C the densification is highly accelerated. This closely corresponds with the data of PXRD and density measurements as discussed earlier. Further, the dense and more uniform particles, as seen in Fig. 8e show that the powder was significantly sintered by 1200 °C to highly dense LAS ceramic, b-spodumene. Right bottom of Fig. 8 shows the energy dispersive X-ray (EDX) spectra of sample at room temperature (before firing) and sample fired at 1200 °C labeled in the figure as LASA0030 °C and LASA1200 °C respectively. The EDX spectrum shows the presence of elements such as Si, Al, O and Na. Due to the detection limitation to lower atomic elements, Li presence is not observed in the EDX data though the presence of trace level of un-exchanged Na is observed. The presence of a insignificant elevation in the base line of

PXRD (LASA1200), the observation of trace level of Na in EDX (LASA1200 °C) and the observation of finer smudged particles in the EDX (Fig. 8e) suggest a possibility of presence of small amounts amorphous substance along with highly crystalline LAS dense phase, b-spodumene.

4. Conclusion The thermal phase transformations of Li modified zeolite A precursor into lithium aluminosilicate (LAS) dense phases in the temperature range between 25 and 1200 °C have been systematically studied in detail. Qualitative phase analysis using powder XRD reveals the involvement of a total structure collapse of the zeolite phase around 700 °C followed by the recrystallization into the LAS phases at elevated temperatures. The formation, co-existence and disappearance of these phases were identified and compared with literature. The correlation of TG, DTA, SEM, AAS, EDX and density measurements along with PXRD provided more details about the satellite transformations at higher temperatures. The observation of various phases suggests the existence of intermediate satellite transitions with in SiO2 as the primary single component system as well as a ternary system viz., Li2O–Al2O3–SiO2 (LAS). Silica forms satellite phases such as high-quartz and keatite. However, keatite is observed to mediate the conversion between the LAS ternary phases, from virgilite to b-spodumene. Thermal characterization along with PXRD reveals the transformation of the zeolite involves different steps viz., dehydration, de-amination

K. Selvaraj / Microporous and Mesoporous Materials 135 (2010) 82–89

and dehydroxylation. Scanning electron micrographs further supplement these observations and further show that the maximum acceleration in sintering is around 800 °C. Density observation throughout the firing process proves that the LAS phase is effectively sintered and most of the densification is completed before 1100 °C. A systematic correlation between different characterization techniques provides better insight into the phase transformation process. Acknowledgement The author is indebted to Mr. Gaikwad, CMC, NCL for helping in getting the SEM and EDX data. References [1] D.W. Breck, Zeolite Molecular Sieves, John Wiley and Sons, NY, 1974, pp. 493– 494. [2] B. Hoghooghi, J. McKittrick, C. Butler, P. Desch, J. Non-Cryst. Solids 170 (3) (1994) 303. [3] R.L. Bedard, E.M. Flanigen, United States Patent, USP 5064790, 12 November, 1991. [4] L.M. Colyer, G.N. Greaves, S.W. Carr, K.K. Foxs, J. Phys. Chem. B 101 (48) (1997) 10105. [5] W.H. Flank, United States Patent, USP 3574647, 13 April, 1971. [6] J. McKittrick, B. Hoghooghi, O.A. Lopez, J. Non-Cryst. Solids 197 (2–3) (1996) 170. [7] M.C. Wang, M.H. Lin, K.M. Hon, Cailiao Kexue 25 (3) (1993) 145. [8] C. Kosanovic, B. Subotic, I. Smit, Thermochim. Acta 317 (1) (1998) 25. [9] G. Dell’Agli, C. Ferone, M.C. Mascolo, M. Pansini, Solid State Ionics 127 (3–4) (2000) 309. [10] C. Weidenthaler, W. Schmidt, Chem. Mater. 12 (12) (2000) 3811. [11] T. Ohgushi, K. Ishimaru, S. Komarneni, J. Am. Ceram. Soc. 84 (2) (2004) 321. [12] R. Dimitrijevic, V. Dondur, P. Vulic, S. Markovic, S. Macura, J. Phys. Chem. Solids 65 (10) (2004) 1623.

89

[13] A.J. Perrotta, D.K. Grubbs, E.S. Martin, N.R. Dando, H.A. McKinstry, C.Y. Huarg, J. Am. Ceram. Soc. 72 (3) (2005) 441. [14] Bahar Hoghooghi, Joanna McKittrick, Eugene Helsel, Olivia A. Lopez, J. Am. Ceram. Soc. 81 (4) (2005) 845. [15] H.J. Lee, Y.M. Kim, O.S. Kweon, I.J. Kim, J. Eur. Ceram. Soc. 27 (2–3) (2007) 561. [16] V. Dondur, R. Dimitrijevic´, A. Kremenovic´, Lj. Damjanovic´, N. Romcˇevic´, S. Macura, J. Phys. Chem. Solids 69 (11) (2008) 2827. [17] B. Liguori, C. Ferone, S. Anaclerio, C. Colella, Solid State Ionics 179 (40) (2008) 2358. [18] M.C. Wang, M.H. Lin, K.M. Hon, Cailiao Kexue 25 (3) (1995) 145. [19] S. Chandrasekhar, P.N. Pramada, Appl. Clay Sci. 27 (3–4) (2004) 187. [20] B. Ruedinger, R.X. Fischer, Eur. J. Mineral. 9 (6) (1997) 1257. [21] K. Selvaraj, V. Ramaswamy, A.V. Ramaswamy, Asian J. Phys. 6 (1997) 132. [22] K. Selvaraj, V. Ramaswamy, A.V. Ramaswamy, Mater. Res. Soc. Proc. 3 (1999) 2033. [23] (a) Inorganic Crystal Structure Database (ICSD), Fachinformationszentrum (FIZ) Karlsruhe, 2007 R.J. Angel, R.K. McMullan, C.T. Prewitt, Am. Mineral. 76 (1991) 332 (ICSD Reference Code 00-79-1276); (b) B. French et al., Am. Mineral. 63 (1978) 461 (ICSD Reference Code 00-310707); (c) Roy, Z. Kristallogr. Kristallgeom. Kristallphys. Kristallchem. 111 (1959) 185 (ICSD Reference Code 01-12-0708); (d) C.-T. Li, D. Peacor, Z. Kristallogr. Kristallgeom. Kristallphys. Kristallchem. 126 (1968) 46 (ICSD Reference Code 00-35-0797); (e) Roy, Z. Kristallogr. Kristallgeom. Kristallphys. Kristallchem. 111 (1959) 185 (ICSD Reference Code 00-12-0708). [24] D.C. Palmer, Rev. Mineral. 29 (1994) 83. [25] L.W. Hobbs, C.E. Jesurum, B. Berger, in: A.B.R. Devine, J.P. Duraud, E. Dooryhee (Eds.), The Topology of Silica Networks, Struct. Imperfections Amorphous Cryst. Silicon Dioxide, John Wiley and Sons Ltd., Chichester, UK, 2000, pp. 3– 47. [26] D.L. Evans, Acta Cryst. A25 (1969) 234. [27] C.T. Li, Acta Cryst. B27 (1971) 1132. [28] B. Martin, Eur. J. Mineral. 7 (6) (1995) 1389. [29] B.I. Shikunov, et al., Izvestiya Akademii Nauk SSSR, Seriya Khimicheskaya, No. 4, 1973, p. 767 (Translated Version – N.D. Zelinskii, Institute of Organic Chemistry).