Ultra thin zeolite membranes by the secondary growth of spread by benbenzhou


Ultra thin zeolite membranes by the secondary growth of spread

More Info
									      Ultra-thin zeolite membranes by the secondary growth of

   nanozeolite seeds prepared from 3DOm-imprinted silicalite-1


Pyung-Soo Lee1, Xueyi Zhang1, Jared A. Stoeger1, Abdulla Malek2, Wei Fan1, Sandeep Kumar1,Won
      Cheol Yoo3, Saleh Al Hashimi2, R. Lee Penn3, Andreas Stein3 and Michael Tsapatsis1*

1. Department of Chemical Engineering and Materials Science, University of Minnesota,

Minneapolis, MN 55455, USA

2. Chemical Engineering Program, The Petroleum Institute, Abu Dhabi, United Arab


3. Department of Chemistry, University of Minnesota, Minneapolis, MN 55455, USA

Corresponding author:

Prof. Michael Tsapatsis

Phone : 612-626-0920

Fax: (+1) 612- 626-7246

E-mail: tsapatsis@umn.edu

Homepage: http://www.cems.umn.edu/research/tsapatsis

     Zeolite nanocrystals were prepared from 3-dimensionally ordered mesoporous imprinted

(3DOm-i) silicalite-1 by a fragmentation method involving sonication and pH control.

3DOm-i silicalite-1 with 20, 30, and 40 nm spherical elements and a wide range of crystal

sizes (100 - 200 nm, 500-600 nm, and 1-2 µm) were used. The highest yield (57%) of

isolated spherical elements was obtained for 3DOm-i silicalite-1 with a crystal size of 100-

200 nm and domain size of 40 nm. The zeolite nanocrystals can be used as seeds for epitaxial

growth of silicalite-1. Silicalite-1 nanocrystal suspensions were used to deposit seed layers on

porous α-alumina disks, which were converted to continuous films by secondary growth.

Thin zeolite membranes (300-400 nm) were obtained exhibiting high permeance (ca. 3740

mol/m2/s/Pa at 125 oC) and high separation factor (ca. 120 at 150 oC) for p-/o-xylene.


        Secondary growth from a zeolite seed layer1,2 is an established method for the

fabrication of zeolite films and membranes.3-5 The size, shape, and spatial arrangement of

seed crystals as well as their evolution during secondary growth determine the structure types,

preferred orientation, grain and grain boundary size, strain, and other microstructural

characteristics that consequently determine the performance of zeolite membranes.6-12

        Zeolite seed layers are normally formed on the substrates of interest from zeolite

powders and suspensions by a variety of wet and dry deposition methods.13-17 Dispersable

zeolite nanoparticles with dimensions smaller than 50 nm are desirable for the formation of

densely-populated and thin seed layers that, upon secondary growth conditions, can lead to

well-intergrown, thin zeolite films. Such films are highly sought after as separation
membranes based on the expectation that they may combine high selectivity with previously

unrealized high fluxes. Although zeolite nanoparticles under 50 nm have been reported for

several frameworks (SOD18,19 , FAU20,21, MFI22-26 , MEL27), dispersable and monodispersed

zeolite nanoparticles for coating applications are rather difficult to prepare. Most approaches

rely on control of nucleation and growth, which is rather difficult due to insufficient

understanding of silica speciation and interaction in these complex mixtures. An alternative

approach to generate zeolite nanoparticles is by disassembly or dissolution of zeolite


        3DOm-i silicalite-1 is a type of hierarchical zeolite that is composed of uniform

nanoscale zeolitic spherical elements with their size determined by the 3DOm template

employed during synthesis. The spherical elements have a cubic close-packed (ccp)

arrangement and are interconnected with their respective 12 nearest neighbors, forming single

crystals composed of many oriented, interconnected domains, shown in Figure 1.29 3DOm-i

silicalite-1 has both the micropores of the MFI framework zeolite (with sizes similar to many

industrially important molecules30 ) and ordered mesopores corresponding to the tetrahedral

and octahedral interstitial spaces of a close-packed lattice of spherical elements. Considering

their open structure, their monodisperse and tunable spherical elements under 50 nm, and the

symmetrically positioned connections between the 3DOm-i spherical elements, disassembly

of 3DOm-i particles can be a promising method to produce uniform, isolated MFI


        In this paper, we report the preparation of monodisperse sub-50 nm crystals from the

fragmentation of 3DOm-i silicalite-1 by sonication within a certain pH range. 3DOm-i

silicalite-1 having different crystal size and spherical element size have been used to

determine the yield of zeolite nanocrystals by the fragmentation method. Secondary growth
of the nanocrystals was demonstrated using cryogenic-TEM (cryo-TEM) and implemented

for fabricating well-intergrown, ultra-thin silicalite-1 films. A combination of rapid thermal

processing (RTP) and conventional calcination techniques was used to remove the structure

directing agents (SDAs) from the zeolite films.31 The resulting zeolite membranes were very

thin (less than 400 nm) and exhibited an attractive combination of high flux and separation

factor for xylene isomers.

Synthesis of size-tunable silica nanoparticles, 3DOm carbon, and 3DOm-i silicalite-1

        Silica nanoparticle sols were synthesized by hydrolysis of tetraethyl orthosilicate

(TEOS, 98%, Aldrich) with aqueous solutions of basic amino acid lysine (Sigma-Aldrich) as

described in our previous reports.29,32 Steam-assisted crystallization (SAC) was used for the

confined synthesis of 3DOm-i silicalite-1 in the 3DOm carbon template.29 The composition

of the synthesis sol was 9 TPA2O : 0.15 Na2O : 50 SiO2 : 390 water : 180 ethanol for all

samples. The SAC was carried out at three different temperature and time conditions, 85 o C

for 5 days, at 135 oC for 3 days, or at 180 oC for 2 days.

Fragmentation of 3DOm-i silicalite-1

Dialysis method to study effect of pH Dissolution of 3DOm-i silicalite-1 was performed in a

dialysis tube (Spectra/Por 3, Spectrum Laboratories Inc.), which allowed separation of

silicalite-1 crystals from dissolved silicate species and other ions. A polypropylene beaker

was filled with 1 L deionized water (Millipore Elix, 10MX cm). Tetrapropylammonium

hydroxide solution (TPAOH, 1.0 M in water, Sigma-Aldrich) or L-lysine (Sigma-Aldrich)

was added to the beaker to achieve the desired initial pH value for dissolution. 75.0 mg

3DOm-i silicalite-1 powder was transferred to a centrifuge tube containing 8 mL of the same

solution as in the beaker, to which sonication for approximately 5 min was applied to ensure

the crystals are separated from each other before dissolution. The mixture was then

transferred to a rinsed dialysis tube, which was then sealed and placed into the beaker. The

solution in the beaker was stirred slowly allowing the dialysis tube to rotate. The pH of the

solution in the beaker was monitored regularly. The solution outside the dialysis tube was

changed regularly to a fresh one with the desired initial pH, in order to maintain the pH
within the desired range and remove silicate species produced by dissolution. After the

dissolution was carried out for desired time, the solution outside the dialysis tube was

changed to deionized water, in order to remove the silicate species. After one day, the sol

inside the dialysis tube was collected and dried at 70 oC and 135 oC. 27.0 mg and 4.6 mg

powder was obtained from dissolution pH 9~10 and 11~12, respectively.

Sonication method for higher yield fragmentation of 3DOm-i silicalite-1 Ultrasonication

was used in order to fragment the particles for further characterization and subsequent

deposition onto a porous support for membrane growth. First, 0.1 g of 3DOm-i silicalite-1

was added to 20 mL deionized water in conical tubes. The 3DOm-i silicalite-1 suspensions

were then sonicated (Bransonic Model 5510R-DTH) for 90 min, after which ice was added to

the sonicator water bath to lower the temperature to 15 oC. The water bath temperature was

maintained in a range from 15 oC to 30 oC during sonication. The period of sonication cycles

totaled 18 h for a typical run, but times up to 60 h were investigated. The sols were

unbuffered and the pH ranged from 10 (at the beginning) to 9 (at the end) during the

fragmentation process. After the sonication, centrifugation at 10,000 rpm for 10 min was

performed to separate the clear sol supernatant containing silicalite-1 nanocrystals from the

larger partially fragmented particles. The nanocrystal suspensions were dried at 95 oC for 5 h

and then dried at 180 oC for 1 day. . In order to calculate yields, the final weights of the dried

powders were measured.

Cryo-TEM study of secondary growth of isolated domains

       The silicalite-1 seeds used to study secondary growth were 40 nm spherical elements

isolated from 3DOm-i silicalite-1 crystals using the sonication procedure described above.

After isolation, the seeds were suspended in water at a concentration of 1 mg silicalite-1 /
mL. The suspension was visually stable with no precipitation at room temperature for a

period of at least 3 months. The growth solutions used for this study were prepared as

described by Davis et al.,33 where the compositions are as follows:

                    5 SiO2:9 TPAOH:8100 H2O:20 EtOH (denoted as C1)

                   10 SiO2:9 TPAOH:8100 H2O:40 EtOH (denoted as C2)

                   20 SiO2:9 TPAOH:8100 H2O:80 EtOH (denoted as C3)

       The 1 mg / mL seed suspension was mixed with an appropriate volume of growth

solution so as to yield suspensions that were 0.2 mg silicalite-1 / mL with compositions

consistent with C1, C2, or C3. The suspension was sealed in a centrifuge tube with Teflon

tape to prevent evaporation of the solvent and was heated in a convection oven at 70 oC. After

a certain time, the centrifuge tube was removed and plunged into an ice bath (0 oC). A cryo-

TEM sample was prepared from the suspension immediately afterwards.

       Cryogenic TEM (Cryo-TEM) sample preparation was carried out in an FEI Vitrobot

Mark III vitrification robot. 34 A droplet of the suspension was placed onto a carbon/formvar-

coated copper grid (Ted Pella Inc.) in the climate chamber of the Vitrobot system, where the

temperature was kept at 25 o C and the relative humidity was kept at 100%. The specimen was

transferred under liquid nitrogen to a Gatan 626 DH cryo-transfer specimen holder.

Fabrication of zeolite membrane

        Homemade porous α-alumina disks were used as supports for silicalite-1 membrane

growth.9 We have employed a masking technique as introduced by Hedlund and co-authors.35

A viscous solution of poly vinyl alcohol (PVA, 80% hydrolyzed, Aldrich) was prepared by

dissolving 12.5 g PVA in 10 g of deionized water at 70 o C. First, the solution was spread over

the top surface of the alumina support using a spatula and dried for 12 h at room temperature.

Second, paraffin wax (melting point 70-80 oC, Aldrich) was used to fill the pores of the
alumina disk by vacuum infiltration (20 in. Hg) at 90 oC for 1 h. Excess wax remaining on the

surface of the alumina disk after the 1 h period was removed by a tissue before the wax

solidified. Third, PVA on the alumina surface was removed in tap water (refreshed every 4 h)

at 50 oC for 12 hr, exposing the alumina surface while leaving the pores filled with wax.

After PVA removal, 300 μL of a zeolite nanocrystal suspension (0.5 wt%) was dropped onto

the surface of the alumina disk and dried for 5 h at room temperature. Excess silicalite-1

nanocrystals on the support surface were leveled off by rubbing on weighing paper (VWR

Scientific Products, 4×4 inches). The support was calcined at 550 oC for 8 h in flowing air to

remove the wax. Secondary growth was carried out by placing the seeded alumina disk

vertically in a Teflon liner with C3 composition at 90 °C for 11 h. Rapid thermal processing

(RTP, nominal heating to 700 oC within 1 min, no hold time, and rapid cooling to room

temperature) and subsequent conventional calcination (1 oC/min to 500 o C, held for 2 h) were

carried out to remove the occluded SDA molecules using the set-up reported previously.31 In

contrast to the previously reported RTP method for the calcination of thick zeolite

membranes, no holding time at the maximum temperature was applied.


    SEM images were collected on a Hitachi S-900 and Hitach S-4700 after the samples

were coated by 1 nm of Pt. The samples for TEM studies were prepared by applying a few

droplets of the sol onto a copper grid coated with a holey carbon film (Ted Pella Inc.). The

grid was then allowed to air-dry. Imaging was performed at -177 oC on an FEI Tecnai G2 F30

TEM operating at 300 kV. All TEM images were captured using a CCD camera. Small-angle

X-ray diffraction (SAXD) measurements were taken on a home-built pinhole SAXD line with

a sample-to-detector distance of 100 cm using Cu Kα radiation. N2 adsorption and desorption

isotherms were measured at 77 K on a Quantachrome Autosorb-1 system. The specific
surface area was calculated using the Brunauer–Emmett–Teller (BET) method. Conventional

t-plot methods were used for extracting micropore volume and external surface area from the

nitrogen adsorption data. X-ray diffraction (XRD) patterns were acquired using a

PANalytical X-Pert PRO MPD X-ray diffractometer equipped with a Co source and a Bruker

AXS D5005 diffractometer with a Cu source. A zeta potential analyzer (ZetaPALS,

Brookhaven Instruments Corperation) was used for dynamic light scattering (DLS)

measurements. Particle sizes were calculated using software provided by Brookhaven

Instruments Corporation.

Permeation measurement

        The xylene isomer vapor permeation experiments were conducted in Wicke-

Kallenbach mode with a stainless steel permeation cell fitted inside a Thermolyne 48000

furnace, using an equimolar p-xylene/o-xylene feed as previously reported.9      In order to

analyze the xylene permeate stream, on-line gas chromatography (HP 5890) was used to

determine the molar fluxes based on a methane standard.     For each temperature step in the

permeation experiment the equilibration time was approximately 4 h. Measurements were

taken at various time intervals to ensure equilibration at the desired temperature. The molar

fluxes of the xylene isomer compounds were estimated by the molar fraction at the time of

sampling (based on GC calibration), the helium sweep stream flow-rate, and the membrane

area. For the equimolar feed used here, the separation factor is defined as the ratio of the

molar fluxes.
Results and Discussion


        Figure 1 shows what is meant herein by crystal size and spherical element of 3DOm-i

silicalite-1. The crystal sizes were varied from 100 nm to 2 µm, and the spherical elements

were tuned from 10 nm to 40 nm.

1. Synthesis of 3DOm-i silicalite-1 with tunable spherical element and crystal size

        3DOm-i silicalite-1 can be synthesized inside the ordered spherical cages of a 3DOm

carbon template by steam-assisted crystallization (SAC). By controlling the diameter of the

lysine-silica spheres used to form the 3DOm carbon, the corresponding diameters of the

zeolite domains can be controlled from 10 nm to 50 nm.29 Figure 2 shows SEM and TEM

images of two different 3DOm-i silicalite-1 samples prepared from 40 nm and 20 nm cage

sizes of 3DOm carbon using the same temperature (135 oC), reaction time (3 days), and

chemical composition (9 TPA2O : 0.15 Na2O : 50 SiO2 : 390 water : 180 ethanol). Figure 2

(a) - (d) are for 3DOm-i silicalite-1 with 40 nm spherical elements. The low-magnification

view in Figure 2(a) reveals an average crystal size of approximately 1 μm. The higher

magnification image in Figure 2(b) shows the spherical domains forming the 3DOm-i

silicalite-1. The TEM image of Figure 2(c) reveals an arrangement of spherical elements

well-aligned and close-packed with the appearance of a mesoporous structure. The crystalline

structure of the spherical elements is shown in the high magnification part of Figure 2(d).

Figures 2(e) - (h) are SEM and TEM images of 3DOm-i silicalite-1 with 20 nm domains. As

previously shown in Figure 2(a), Figure 2(e) exhibits average crystal size of 1 μm. In addition

to 3DOm-i silicalite-1, we assume that the few faceted zeolites (e.g., indicated by the arrow
in Figure 2(e)) grow outside the 3DOm carbon template. Counting from SEM images, we

found that 95% by number of the silicalite-1 crystallized inside the 3DOm carbon. The TEM

image in Figure 2(g) confirms an ordered arrangement of 40 nm spherical elements, and a

clear crystalline structure can be seen in the high magnification TEM image of Figure 2(h).

       Figure 3 exhibits 3DOm-i silicalite-1 crystallized inside 3DOm carbon with 30 nm

pore diameters. SAC was conducted at 85 oC for 5 days to crystallize the zeolite particles

seen in Figure 3(a), while those in Figure 3(b) and -(c) were obtained at 135 oC for 3 days and

180 oC for 2 days, respectively. The chemical composition remained the same for the three

cases (9 TPA2O : 0.15 Na2O : 50 SiO2 : 390 water : 180 ethanol). Although the spherical

element sizes of the 3DOm-i silicalite-1 shown in Figure 3 is 30 nm, the crystal sizes of

3DOm-i silicalite-1 varied: 100-200 nm, 500-600 nm, and 1 µm in Figures 3(a), (b), and (c),

respectively. It may be concluded that control of the 3DOm silicalite-1 crystal sizes (while

keeping the spherical element size constant) may be afforded by altering the reaction time

and temperature.

2. Disassembly of the 3DOm-i silicalite-1 to yield isolated spherical elements as single


       Although spherical elements in 3DOm-i silicalite-1 are connected in a close-packed

arrangement, the connections are composed of small amounts of materials so that they may

be broken by dissolution and/or a disturbance such as ultrasonication. If the spherical

elements can be released intact from the host 3DOm-i silicalite-1 then silicalite-1

nanocrystals can be collected. Here, we report that by dissolution of 3DOm-i silicalite-1, in a

certain pH range, isolation of the individual spherical elements from 3DOm-i crystals was

achieved. In a different pH range, silicalite-1 nanocrystals smaller than 50 nm with a

“fractured egg-shell” morphology were formed, further extending the size limit and
morphological variety of silicalite-1 nanoparticles. We also provide a high-yield, additive-

free dissolution procedure for preparing zeolite seeds.

2-1. Effect of dissolution pH

       Depending on the pH at which the dissolution process takes place, silicalite-1

nanocrystals with different morphologies can be obtained from the 3DOm-i silicalite-1. If the

dissolution pH is kept between 9.0 and 10.0, intact domains can be released as shown in

Figure 4(a) and (b) which are representative TEM images of the nanocrystals formed by

dissolving 3DOm silicalite-1 with 40 nm spherical elements (shown in Figure 2(a)-(d)) at this

condition. When the dissolution pH is kept between 11.0 and 12.0, spherical isolated

elements with observable mesopores are formed together with smaller silicalite-1 nanocrystal

fragments having irregular shapes. Figure 4(d) shows the 40 nm etched silicalite-1

nanocrystal with irregular nanocrystals that resulted from such higher pH dissolution. Figure

4(c) is an image of an isolated 40 nm etched nanocrystal, in which the pores can be clearly

observed. Figure S1 shows that similar conclusions can be drawn when using 3DOm-i

silicalite-1 with 25 nm spherical elements.

       Although the etched morphology obtained at the higher pH range does not appear to

be appropriate for the formation of ordered seed layers, the nanoscopic fragments obtained (a

sampling of small fragments in shown in Figure S2) are of fundamental significance because

they demonstrate that zeolite nanoparticles that contain only few unit cells exist and can be

stable with respect to structural transformation, aggregation or electron beam damage

suggesting that formation of zeolite nanoparticles smaller than 10 nm by a suitable

nucleation-growth approach is highly probable.
       The XRD patterns of the silicalite-1 crystals show little change after dissolution

(Figure 5(a)) indicating that the spherical objects shown in Figure 4 appear to have MFI

framework at both dissolution pH ranges. As expected, due to spherical element isolation,

SAXD measurements (Figure 5(b)) show that dissolution under both conditions nearly

eliminated the ccp order. After dissolution, the specific micropore volume of the particles

decreased for both dissolution pH ranges (from 0.129 mL/g to 0.089 mL/g). As seen in the

isotherms of Figure 5(c), dissolution with pH between 9.0 and 10.0 (B) led to wider mesopore

distribution than 3DOm-i crystals (A), which can be attributed to loss of packing order of the

isolated spherical elements. Dissolution with pH between 11.0~12.0 (C) results in greatly

increased total pore volume from the new mesopores, formed within the isolated crystals.

2-2. Dissolution by sonication of unbuffered sols

       The dialysis procedure described above requires pH adjustment and is a low yield

process; therefore, it is not an ideal candidate for scale up. An alternative process that

requires no additives was developed based on the following two observations: (1) Unbuffered

silicalite-1 suspensions exhibit basic pH and (2) silicalite-1 dissolution releases silicic acid

species which reduces the initial basic pH. The process consists of simply sonicating a

3DOm-i silicalite-1 aqueous sol under the sol concentration ranges described in the

experimental section.

       Figure 6 shows a series of SEM images from before and after zeolite fragmentation by

the sonication method using 1-2 µm 3DOm-i silicate-1 crystals with 40 nm spherical

elements. 3DOm-i silicalite-1 suspensions were prepared by mixing the zeolite with

deionized water prior to sonication. The pH of the 3DOm-i silicalite-1 suspension was

10.0±0.3. No other additives were included to the 3DOm-i silicalite-1 suspensions. As shown

in Figure 6(a), spherical elements are connected with other elements by a close-packing
arrangement before sonication. Figure 6(b) is an SEM image of a 3DOm-i silicalite-1 after

sonication for 4 h with the close-packing arrangement of spherical elements exposed. Figure

6(c) is an SEM image of the 3DOm-i silicalite-1 suspension after sonication for 18 h. Region

1 outlined in white is indicative of 3DOm-i silicalite-1 unbroken after sonication. Region 2

indicates faceted zeolite particles grown outside of the 3DOm carbon template. The majority

of the observable area (e.g., Region 3) shows isolated zeolite nanocrystals and no remnant

close-packing structure.

       Further sonication past 18 h did not result in improvements in nanozeolite yield. As

sonication time increased, the pH of 3DOm-i silicalite-1 suspensions decreased. For example,

the pH of the zeolite suspension dropped to 9.3 ±0.2 from 10.0 ±0.3 after sonication for 18 h.

Further decrease of pH results in the observable precipitation of nanocrystals. Furthermore,

no lattice fringes were found from the nanoparticles after sonication for 60 h. Thus,

considering nanozeolite yield, stability, and crystallinity the sonication time for fragmentation

was set to 18 h.

       As shown in Figure 6(c), an aqueous 3DOm-i silicalite-1 suspension consists of

unbroken 3DOm-i silicalite-1, faceted zeolite formed outside of 3DOm carbon, and isolated

silicalite-1 nanocrystals after the sonication. Due to the size difference between the silicalite-

1 nanocrystals and the other two components, they are easily separated by centrifugation. A

SEM image in Figure 6(d) shows the silicalite-1 nanocrystals present in the supernatant

obtained by centrifugation at 10,000 rpm for 10 min.

       Through the control of zeolite crystallite sizes and spherical elements, twelve 3DOm-i

silicalite-1 samples were prepared for fragmentation. Different pore sizes of 3DOm carbon

from 10 nm to 40 nm were used, and crystal sizes of 3DOm-i silicalite-1 were controlled to

100-200, 500-600 nm, and 1-2 µm. Table 1 shows the yield of silicalite-1 nanocrystals with
respect to the starting 3DOm-i crystals by the fragmentation-through-sonication method. The

yield from 10 nm spherical elements was too small to determine. However, 10 nm intact

isolated crystals can be observed by TEM as shown in Figure S3. Silicalite-1 nanocrystal

suspensions can be obtained with appreciable yield as the sizes of the spherical elements

become larger than 20 nm. Two trends were identified from the fragmentation-through-

sonication study. First, a higher yield was obtained from 3DOm-i silicalite-1 having larger

spherical elements. This may be because either they are broken easily by sonication or

isolated bigger nanocrystals are more stable in aqueous solution. Second, the crystal size of

3DOm-i silicalite-1 affects the yield of nanocrystals. Crystal sizes of 100-200 nm exhibited

higher yields than 500-600 nm and 1-2 μm crystal sizes. However, the trend with crystal size

is not monotonic.

       Figure 7 shows characterization data of isolated zeolite nanocrystals from 20 nm

3DOm-i silicalite-1. A low magnification TEM image in Figure 7(a) exhibits silicalite-1

nanocrystals after drying on a TEM grid. The DLS data shown in Figure 7(b) also reveal the

presence of nanocrystals around 20 nm with a small fraction of 50 nm crystals (possibly

small clusters of spherical elements). A high magnification TEM image in Figure 7(c) shows

the crystalline structure of isolated nanozeolite. It is clearly shown that the silicalite-1

nanocrystals maintain a highly crystalline nature after long sonication. To further explore the

crystallinity of the nanozeolite, the sol was dried on a silicon wafer for XRD. Figure 7(d)

shows peaks consistent with the silicalite-1 crystal structure.

        A TEM image in Figure S4(a) shows monodisperse silicalite-1 nanocrystals of 40 nm

dried on a TEM grid. DLS (Figure S4(b)) indicates narrow distribution of 40 nm nanocrystals.

Zeolite nanocrystals of 40 nm exhibit better monodispersity than suspensions containing
particles of 20 nm. A TEM image of 40 nm zeolite nanocrystals shown in Figure S4(c) and

XRD data (Figure S4(d)) clearly show the crystalline structure of silicalite-1 crystals.

3. Secondary growth of silicalite-1 nanoparticles: a cryo-TEM study

       Cryo-TEM images in Figure 8 show the evolution of 40 nm silicalite-1 crystals in C3

solution at 70 oC over a period of 36 h. From the images in Figure 8, it can be seen that in C3

solution, silicalite-1 crystals were among precursor nanoparticles with the size of

approximately 5 nm.33 After 6 h, the seeds showed no obvious growth, as seen in Figure 8(a).

However, the surfaces of the seed particles became rough, which might be attributed to a

combined effect of dissolution by base (TPAOH) in the growth solution and aggregation of

the precursor nanoparticles onto the seeds. After 12 h (Figure S5(a)) and 18 h (Figure 8(b)),

there is clear evidence of growth and as shown by the FFT (fast Fourier transform) of the

image of the grown and original part, the grown portion is crystalline and epitaxially related

to the seed crystal. Growth continued, as seen in Figure S5(b) and Figure 8 (c). Throughout

the growth process, the grown parts appeared crystalline, and the crystallographic

orientations of the seed and the grown parts were identical. From these images, it is apparent

that during growth the particles do not maintain a defined shape.

       In addition to C3 solution, secondary growth was also attempted using C1 and C2 sols.

The SiO2 concentration in C1 is below silica solubility, the SiO 2 concentration in C2 is near

silica solubility, while that of C3 is above silica solubility. Shape change because of

dissolution was observed on the silicalite-1 nanocrystals if they were kept in C1 solution for

36 h. A representative cryo-TEM image of silicalite-1 crystal in C1 is shown in Figure 9(a).

       The same 40 nm silicalite-1 seeds were also kept in C2 solution at 70 oC for 36 h,

whose size increased to approximately 60 nm (Figure 9(b)). Comparing the TEM images of
grown particles in C2 and C3, the particle grown in the latter case always showed lattice

fringes throughout the particles. However, the former showed a thick layer (5~15 nm) outside

the particles without any fringes. Under cryo-TEM condition, defocal series from underfocus

to overfocus conditions were taken to study the crystallinity of this layer. A representative

defocal series (supplementary information S6) shows that lattice fringes do not appear near

the edges. Here, we can conclude that the grown part is not crystalline: the grown particles in

C2 have a core-shell structure where the core (seed) is crystalline but the shell is amorphous.

Based on these observations we selected composition C3 to perform secondary growth for

membrane fabrication.

4. Fabrication of silicalite-1 membrane by secondary growth

        Zeolite nanocrystals made by the fragmentation of 3DOm-1 silicalite-1 were used as

seed crystals to fabricate zeolite membranes following the approach of Hedlund and co-

authors.35 Figure 10(a) is a representative SEM image of a seeded layer of zeolite

nanocrystals deposited on the surface of the alumina disk after the removal of wax by

calcination. The entire support surface was successfully seeded. Since zeolite nanocrystals

were densely-packed on the alumina disk surface, conditions including high temperature

and/or a zeolite synthetic sol with a high silica content were not necessary during secondary

growth. The SEM image of Figure 10(b) shows a cross-sectional view of a zeolite membrane

after secondary growth for 11 h and the removal of SDAs. A well-intergrown zeolite

membrane with thickness in the 300-400 nm range is shown while the silicalite-1 crystal

structure is clearly shown by the XRD data of Figure 10(c). Due to the thin zeolite layer, the

α-alumina peak from the substrate has a much higher intensity compared to the peaks

corresponding to the randomly-oriented silicalite-1 film.
        Figure 11 shows p-/o-xylene isomer permeation data. High permeance was achieved

due to the submicron thickness of the membrane. Permeances in the temperature range of

100-150 oC were higher than 3500 mol/m2 /s/Pa, and the highest flux was ca. 3740

mol/m2/s/Pa at 125 o C. Considering that the limiting flux from the bare alumina disk is ca.

4000 mol/m2/s/Pa, it appears that further improvements of the flux will require different

support materials. High separation factors were obtained at 125 oC (ca. 101) and 150 oC (ca.

121). Such separation factors indicate a high-quality silicalite-1 membrane.6,31,36 Thus, a

high-quality membrane having high flux can be made by the secondary growth of the zeolite

nanocrystal seeds prepared by 3DOm-i silicalite-1 disassembly.


We reported a method of fabricating thin zeolite membranes with nanozeolites obtained from

the fragmentation of 3DOm-i silicalite-1. The spherical element sizes of 3DOm-i silicalite-1

can be controlled by choosing different cage sizes of 3DOm carbon from 10 nm to 40 nm. By

varying temperature and time, the crystal size of 3DOm-i silicalite-1 may also tuned from

100 nm to 1-2 µm. Zeolite nanocrystal suspensions were prepared by the fragmentation of

3DOm-i silicalite-1 through ultrasonication and subsequent centrifugation. The best yield was

obtained from 40 nm domain size from a parent 3DOm-i silicalite-1 with a crystallite size of

100-200 nm after sonication for 18 h. Appreciable yields were obtained for spherical element

sizes down to 20 nm. Although isolated 10 nm crystals were observed, the yields were too

small to be determined. The zeolite nanocrystals were further used as a densely-packed

zeolite layer on a porous α-alumina disk. Relatively short secondary growth time (11 h) and

low temperature (90 oC) were sufficient conditions to form a well-intergrown zeolite layer.

The resulting zeolite membrane was very thin (in the range of 300-400 nm), and high fluxes
(ca. 3740 mol/m2/s/Pa at 125 oC) with a high separation factor (ca. 120 at 150 oC) were

measured for p-/o-xylene.

        The authors gratefully acknowledge financial support from the Petroleum Institute of

Abu Dhabi through the ADMIRE (Abu Dhabi-Minnesota Institute for Research Excellence)

partnership and the National Science Foundation (NSF-NIRT CMMI 0707610). Portions of

this work were carried out at the University of Minnesota, Institute of Technology

Characterization Facility, which receives partial support from NSF through the NNIN


Figure captions

Figure 1. A conceptual schematic of 3DOm-i silicalite-1.

Figure 2. (a) SEM and (b) TEM images of 3DOm silicalite-1 with 20 nm spherical element.

(c) SEM and (d) TEM images of 3DOm-i silicalite-1 with 40 nm spherical element.

Figure 3. SEM images of 3DOm-i silicalite-1 with 30 nm spherical element varying crystal

sizes. (a) 3DOm-i silicalite-1 with 100-200 nm crystal sizes, (b) 500-600 nm, and (c) 1 µm.

Figure 4 TEM images of 40nm spherical elements in different dissolution pHs. (a) and (b)

isolated 40 nm spherical elements formed with dissolution pH 9.0-10.0, (c) and (d) isolated

40 nm spherical elements formed with dissolution, pH: 11.0-12.0.

Figure 5. (a) X-ray diffraction patterns, (b) Small-angle X-ray diffraction patterns, and (c) N 2

adsorption isotherms of 3DOm-i silicalite-1 before and after dissolution. Trace A: as-

synthesized, trace B: after dissolution at pH: 9.0~10.0, and trace C: after dissolution at pH:

11.0~12.0. Dissolution time: 7 days.
Figure 6. SEM images of 3DOm-i silicalite-1 during fragmentation by sonication. 3DOm-i

silicalite-1 with 40 nm (a) before sonication, (b) after 4 h, (c) after 18 h, and (d) isolated

nanozeolite after centrifugation.

Figure 7. Characterization of isolated nanozeolite with a 20 nm spherical element size. (a)

Low magnification TEM, (b) DLS, (c) high magnification TEM, and (d) wide angle X-ray

diffraction after drying on a silicon wafer.

Figure 8. Representative cryo-TEM images of silicalite-1 crystals heated at 70 oC in C3

solution, the particles are among precursor nanoparticles sized approximately 5 nm: (a) after

6 h, (b) after 18 h, and (c) after 36 h.

Figure 9. Cryo-TEM images of 40 nm silicalite-1 seeds treated at 70 oC for 36 hrs: (a) in C1

solution the seeds undergo dissolution; (b) in C2 solution the seeds are surrounded by an

amorphous shell.

Figure 10. Seeded growth of an ultra-thin zeolite membrane. (a) SEM image of the seed

layer on an α-alumina disk support, (b) cross-sectional image of the thin zeolite membrane,

and (c) wide angle X-ray diffraction of zeolite membrane. An asterisk (*) denotes the peak

from the alumina disk.

Figure 11. p-Xylene (filled circles) and o-xylene (open circles) permeances and

corresponding separation factors (triangles) using the ultra-thin zeolite membrane.      The

permeance of p-xylene nearly matches that observed through the blank disc (filled squares).

Table 1. Yields of zeolite nanocrystals from fragmentation of 3DOm-i silicalite-1 by


        (1)      Lovallo, M. C.; Tsapatsis, M. AIChE J. 1996, 42, 3020.
        (2)    Valtchev, V.; Schoeman, B. J.; Hedlund, J.; Mintova, S.; Sterte, J. Zeolites
1996, 17, 408.
        (3)      Snyder, M. A.; Tsapatsis, M. Angew. Chem., Int. Ed. 2007, 46, 7560.
        (4)      Caro, J.; Noack, M. Microporous Mesoporous Mater. 2008, 115, 215.
        (5)      Tsapatsis, M.; Gavalas, G. R. MRS Bulletin 1999, 24, 30.
        (6)      Lai, Z. P.; Bonilla, G.; Diaz, I.; Nery, J. G.; Sujaoti, K.; Amat, M. A.; Kokkoli,
E.; Terasaki, O.; Thompson, R. W.; Tsapatsis, M.; Vlachos, D. G. Science 2003, 300, 456.
        (7)      Choi, J.; Ghosh, S.; Lai, Z. P.; Tsapatsis, M. Angew. Chem., Int. Ed. 2006,
45, 1154.
        (8)      Jeong, H. K.; Krohn, J.; Sujaoti, K.; Tsapatsis, M. J. Am. Chem. Soc. 2002,
124, 12966.
        (9)      Choi, J.; Ghosh, S.; King, L.; Tsapatsis, M. Adsorption 2006, 12, 339.
        (10)     Lee, I.; Buday, J. L.; Jeong, H. K. Microporous Mesoporous Mater. 2009,
122, 288.
        (11)     Liu, Y.; Li, Y. S.; Yang, W. S. J. Am. Chem. Soc. 2010, 132, 1768.
        (12)     Coronas, J. Chem. Eng. J. (Lausanne) 2010, 156, 236.
        (13)     Yoon, K. B. Acc. Chem. Res. 2007, 40, 29.
        (14)     Lee, J. S.; Lee, Y. J.; Tae, E. L.; Park, Y. S.; Yoon, K. B. Science 2003, 301,
        (15)     Lee, J. S.; Kim, J. H.; Lee, Y. J.; Jeong, N. C.; Yoon, K. B. Angew. Chem.,
Int. Ed. 2007, 46, 3087.
        (16)     Matsukata, M.; Ogura, M.; Osaki, T.; Rao, P.; Nomura, M.; Kikuchi, E. Top.
Catal. 1999, 9, 77.
        (17)    Lee, J. A.; Meng, L. L.; Norris, D. J.; Scriven, L. E.; Tsapatsis, M. Langmuir
2006, 22, 5217.
        (18)     Schoeman, B. J.; Sterte, J.; Otterstedt, J. E. Zeolites 1994, 14, 208.
        (19)     Fan, W.; Morozumi, K.; Kimura, R.; Yokoi, T.; Okubo, T. Langmuir 2008, 24,
        (20)     Mintova, S.; Olson, N. H.; Bein, T. Angew. Chem., Int. Ed. 1999, 38, 3201.
        (21)     Holmberg, B. A.; Wang, H. T.; Norbeck, J. M.; Yan, Y. S. Microporous
Mesoporous Mater. 2003, 59, 13.
        (22)   Motuzas, J.; Julbe, A.; Noble, R. D.; Guizard, C.; Beresnevicius, Z. J.; Cot,
D. Microporous Mesoporous Mater. 2005, 80, 73.
        (23)     Valtchev, V. P.; Faust, A. C.; Lezervant, J. Microporous Mesoporous Mater.
2004, 68, 91.
        (24)     Van Grieken, R.; Sotelo, J. L.; Menendez, J. M.; Melero, J. A. Microporous
Mesoporous Mater. 2000, 39, 135.
        (25)     Song, W.; Justice, R. E.; Jones, C. A.; Grassian, V. H.; Larsen, S. C.
Langmuir 2004, 20, 8301.
        (26)     Aguado, J.; Serrano, D. P.; Escola, J. M.; Rodriguez, J. M. Microporous
Mesoporous Mater. 2004, 75, 41.
        (27)     Li, Z. J.; Lew, C. M.; Li, S.; Medina, D. I.; Yan, Y. S. J. Phys. Chem. B 2005,
109, 8652.
        (28)     Stein, A.; Li, F.; Wang, Z. J. Mater. Chem. 2009, 19, 2094.
        (29)     Fan, W.; Snyder, M. A.; Kumar, S.; Lee, P. S.; Yoo, W. C.; McCormick, A.
V.; Penn, R. L.; Stein, A.; Tsapatsis, M. Nat. Mater. 2008, 7, 984.
        (30)     Kokotailo, G. T.; Lawton, S. L.; Olson, D. H.; Meier, W. M. Nature 1978,
272, 437.
        (31)   Choi, J.; Jeong, H. K.; Snyder, M. A.; Stoeger, J. A.; Masel, R. I.; Tsapatsis,
M. Science 2009, 325, 590.
        (32)     Snyder, M. A.; Lee, J. A.; Davis, T. M.; Scriven, L. E.; Tsapatsis, M.
Langmuir 2007, 23, 9924.
        (33)    Davis, T. M.; Drews, T. O.; Ramanan, H.; He, C.; Dong, J. S.;
Schnablegger, H.; Katsoulakis, M. A.; Kokkoli, E.; McCormick, A. V.; Penn, R. L.; Tsapatsis,
M. Nat. Mater. 2006, 5, 400.
        (34)    Iancu, C. V.; Tivol, W. F.; Schooler, J. B.; Dias, D. P.; Henderson, G. P.;
Murphy, G. E.; Wright, E. R.; Li, Z.; Yu, Z. H.; Briegel, A.; Gan, L.; He, Y. N.; Jensen, G. J.
Nat. Protoc. 2006, 1, 2813.
        (35)     Hedlund, J.; Noack, M.; Kolsch, P.; Creaser, D.; Caro, J.; Sterte, J. J.
Membr. Sci. 1999, 159, 263.
        (36)     Xomeritakis, G.; Lai, Z. P.; Tsapatsis, M. Ind. Eng. Chem. Res. 2001, 40,

                               Supporting information

Figure S1. (a) SEM and (b) TEM image of 3DOm-i silicalite-1 particles with 25 nm

spherical elements; (c), (d) isolated 25 nm spherical elements using pH: 9.0~10.0; (e), (f)

isolated spherical elements under dissolution pH: 11.0~12.0. Dissolution time: 21 days.

Figure S2. TEM images of silicalite-1 nanocrystals with at least one dimension smaller than

10 nm obtained by dissolution of 3DOm-i particles with 25 nm domains at pH: 11.0~12.0.

Dissolution time: 21 days. (all scale bars=5 nm)

Figure S3. A TEM image of isolated zeolite nanocrystals of 10 nm.
Figure S4. Characterization of isolated zeolite nanocrystals with a 40 nm spherical element

size. (a) low magnification TEM, (b) Dynamic light scattering (DLS), (c) high magnification

TEM, and (d) wide angle X-ray diffraction after drying on a silicon wafer.

Figure S5. Representative cryo-TEM images of silicalite-1 particles heated at 70 oC in C3

solution, the particles are among precursor nanoparticles sized approximately 5 nm: (a) after

12 h and (b) after 24 h.

Figure S6. A defocal series (cryo-TEM images with defocus -3000 nm to 3000 nm) of a 40
nm silicalite-1 seed treated at 70 oC for 36 h in C2.

To top