Transport enhancement in binderless zeolite X- and A-type molecular sieves revealed by PFG NMR diffusometry

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

http://www.elsevier.com/authorsrights

Author's personal copy

Transport enhancement in binderless zeolite X- and A-type molecularsieves revealed by PFG NMR diffusometry

Dirk Mehlhorn a, Rustem Valiullin a,⇑, Jörg Kärger a, Kristin Schumann b, Alfons Brandt b, Baldur Unger b

a Faculty of Physics and Earth Sciences, University of Leipzig, Linnéstraße 5, 04103 Leipzig, Germanyb Chemiewerk Bad Köstritz, Heinrichshall 2, 07586 Bad Köstritz, Germany

a r t i c l e i n f o

Article history:Received 3 July 2013Received in revised form 16 December 2013Accepted 6 January 2014Available online 13 January 2014

Keywords:A- and X-type zeolitesBinderless zeolite molecular sievesDiffusionWaterTransport enhancement

a b s t r a c t

Pulsed field gradient (PFG) NMR is applied for probing the rate of mass transfer of water molecules in zeo-lite molecular sieves (beads) of type 4A and NaX (NaMSX). Water diffusivities in the binderless speciesare found to notably exceed the diffusivities in the binder-containing beads. Diffusivity enhancementin the binderless species is referred to both the existence of microporous zeolite ‘‘bridges’’ connectingthe individual particles (crystallites) of genuine zeolite structure and the notably larger diameters ofthe transport pores within the binderless beads. Either of these structural features leads to an accelera-tion of long-range diffusion and, hence, to reduced uptake and release times on the individual beads. Bothinfluences act in parallel. In this case the (apparent) activation energy of long-range diffusion can beexpected to be intermediate between the activation energy of intracrystalline diffusion and the isostericheat of adsorption as experimentally observed.

� 2014 Elsevier Inc. All rights reserved.

1. Introduction

To avoid pressure drop in their technical use, such as separa-tion, chemical conversion and thermal energy storage [1–3], nano-porous materials are usually required to be applied as shapedparticles such as beads or pellets. Since nanoporous materials,and notably zeolites as the most attractive host system for thesetypes of application over decades [4–6], are commercially pro-duced as powders, they have to be formed into larger compounds.The conventional procedure of shaping by addition of inorganic(e.g., mineral) binders is well known to be accompanied with adilution of the active species and the influence of the transportpore system (e.g., increasing of the amount of undesired mesop-ores). Recent activities were therefore focused on the productionof binderless monolithic zeolite species [7–11].

In many cases molecular diffusion is among the key processesdeciding about the performance of these materials in their techno-logical application. In addition to their relevance for fundamentalresearch, diffusion studies are thus often of immediate practicalinterest. Being sensitive to molecular displacements over diffusionpath lengths of the order of a few micrometers, the pulsed fieldgradient technique of NMR (PFG NMR diffusometry, see, e.g.,[12–15]) has proved to be particularly suited for in-depth diffusionstudies with such materials [16,17]. As a particular feature,diffusion measurements of this type are able to vary the observed

diffusion path lengths from distances far below up to much largerthan the crystal diameter. In this way, transport resistances in theinterior of the individual crystallite (i.e., in general, the diffusionresistance of the genuine pore network), on the transition betweenthe intracrystalline and extracrystalline spaces (so-called surfacebarriers [18–20]) and during long-range diffusion become accessi-ble by direct experimental observation [15].

The present communication reports about the exploitation ofthese possibilities for exploring the transport characteristics of bin-derless molecular sieves of zeolite type NaMSX (sodium mediumsilicon X [11]) and 4A in comparison with their binder-containingcounterparts [11,21]. As part of the turnaround in energy policyand the thus stimulated search for and development of new mediafor efficient energy storage, water moved into the focus of generalinterest. It has been chosen, therefore, as a probe molecule in thediffusion experiments reported in this communication.

2. Experimental

2.1. Zeolite specimens under study

We have investigated the diffusivity of water in two species ofbinderless molecular sieves of type NaA and NaMSX [11] in com-parison with the water diffusivities in conventional, binder-con-taining molecular sieves produced with the same zeolite type.

The manufacturing of the binderless molecular sieves understudy, referred to as KÖSTROLITH� 4ABF and KÖSTROLITH�

13XBF, is based on mixing of the relevant zeolite powder with

1387-1811/$ - see front matter � 2014 Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.micromeso.2014.01.011

⇑ Corresponding author. Tel.: +49 3419732515.E-mail address: [email protected] (R. Valiullin).

Microporous and Mesoporous Materials 188 (2014) 126–132

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials

journal homepage: www.elsevier .com/locate /micromeso


metakaolin and caustic, followed by granulation. In a following wetchemical reaction non-zeolitic compounds are converted into apurely zeolitic phase. As a consequence, KÖSTROLITH� 4ABF con-sists almost exclusively of zeolite 4A and KÖSTROLITH� 13XBF ofzeolite NaMSX. They are produced as beads forming a networkwith the crystallites of the zeolite powder held together by zeolite‘‘bridges’’ built of the same type of zeolite. In the final step the bin-derless molecular sieves are activated. A detailed description of theprocedure may be found in Refs. [11,22].

To produce conventional, binder-containing molecular sieves,zeolite powder and an inorganic binder are mixed, shaped and acti-vated (binder-containing 4A or NaMSX). Due to the inert binder inbinder-containing molecular sieves the static water adsorptioncapacity is reduced by the amount of that binder (Table 1).

Fig. 1 shows scanning electron micrographs of the samples un-der study, revealing their appearance as crystallites held togetherby bridges which, in the case of the binderless species, are as wellof zeolitic structure as revealed by X-ray diffraction (yielding highcrystallinity) and gas adsortion (showing adsorption capacitynearly identical to that of the pristine zeolites) analyses [11,22].The inserts enframed by broken lines reappear in Fig. 4 as a stan-dard for correlating the measured molecular displacements withthe structural details of the molecular sieves.

2.2. PFG NMR diffusion measurement

The diffusion measurements were performed by means of thepulsed field gradient (PFG) technique of NMR [12–17]. Under theinfluence of field gradient pulses, the intensity M of the NMR signal(the ‘‘spin echo’’) may quite generally be shown to be attenuatedfollowing the relation:

MðdgÞ=Mð0Þ � w cdg; tð Þ ¼Z 1

�1Pðz; tÞ cos cdgzð Þdz ð1Þ

with g, d and t (>>d) denoting, respectively, the amplitude and dura-tion of the gradient pulses and their separation [23,24]. c(=2.67�108 T�1 s�1 for protons) is the gyromagnetic ratio. P(z,t) de-notes the probability that, during time t, an arbitrarily selected mol-ecule within the sample is shifted over a distance z in the directionof the applied field gradient. PFG NMR is thus seen to be sensitive tomolecular displacements (typically of the order of hundreds ofnanometers up to tens of micrometers) during observation timeswhich, depending on the given nuclear magnetic relaxation times,may vary from milliseconds till seconds. For molecular displace-ments in an isotropic medium, the propagator is a Gaussian:

Pðz; tÞ ¼ 1ffiffiffiffiffiffiffiffiffiffiffiffi4pDtp exp � z2

4Dt

� �ð2Þ

In this case, the mean square width (molecular mean squaredisplacement) of the propagator is given by the Einstein formula:

z2ðtÞ� �

¼Z 1

�1z2Pðz; tÞdz ¼ 2Dt ð3Þ

where D stands for the self-diffusivity of the molecules under con-sideration in this medium. Under conditions where Eqs. (2) and (3)

are applicable, the molecules are said to undergo ‘‘normal’’ diffusion[13,15,25].

Inserting Eq. (2) into Eq. (1), the PFG NMR signal attenuation isfound to be given by the relation:

Wðdg; tÞ ¼ exp �c2d2g2Dt� �

¼ exp �c2d2g2hz2ðtÞi=2� �

ð4Þ

where the second equation results by application of Eq. (3).In PFG NMR diffusion studies with beds of nanoporous materi-

als (crystallites), Eq.(4) is strictly applicable in two limiting cases(see [15], section 11.4), namely for mean diffusions path lengthseither much smaller or much larger than the sizes of the individualcrystallites. Correspondingly, the respective diffusivities are re-ferred to as the coefficients of intracrystalline (Dintra) and long-range (Dlr) diffusion. While the intracrystalline diffusivity is exclu-sively determined by the mobility of the guest molecules in thegenuine pore space of the nanoporous crystallites under study,the long-range diffusivity

DIr ¼ pinterDinter ð5Þ

depends on both the diffusivity (Dinter) and the relative population(pinter) of the guest molecules in the intercrystalline space, i.e., thefree space between the zeolites. It is important to note that any con-nection between the individual crystallites, as brought about by thezeolite bridges in the binderless beads considered in this studygives rise to a second contribution to long-range diffusion, originat-ing from mass transfer along these connections. This process ofmass transfer acts in addition to mass transfer through the inter-crystalline space which, by the notation of Eq. (5), has been exclu-sively considered. In this more general case, Eq. (5) has to bereplaced by a relation of the form:

DIr ¼ Dintra;eff þ pinterDinter ð6Þ

where the term Dintra,eff takes account of this part of long-rangemass transfer which occurs over the interconnections betweencrystallites.

The self-diffusivity as recorded by PFG NMR is a physical quan-tity describing molecular translational mobility under equilibrium.It is, therefore, in general different from the (transport) diffusivityDT defined, by Fick’s first law:

J ¼ �DT grad c ð7Þ

as the factor of proportionality between a gradient in molecularconcentration and the molecular flux j generated by this concentra-tion gradient. As a measure of mass transport by diffusion, thetransport diffusivity rather than the self-diffusivity is the actuallytechnology-relevant quantity. It is the key parameter, determiningthe time constant s of molecular uptake or release [15,26]. Forsphere-like particles, e.g., one has:

s ¼ R2=15DT; ð8Þ

with R denoting the mean particle radius.However, except for well-shaped crystallites where, in the last

few years, intracrystalline fluxes have been successfully recordedby the use of micro-imaging techniques [15,27], transport diffusionis not accessible by direct microscopic observation. Given the op-tion of a microscopic measurement of molecular self-diffusion it

Table 1Static adsorption capacities for water and diffusion-relevant porosity data of the zeolite specimens under study.

Equillibrium water adsorption capacity at 25 �C, 55 relative humidity (wt.%) Total macroporosity(%)

Average macropore diameter(lm)

Binder-containing 4A 22.7 31 0.13Binderless 4A 25.6 31 0.61Binder-containing

NaMSX25.6 31 0.21

Binderless NaMSX 31.2 26 0.45

D. Mehlhorn et al. / Microporous and Mesoporous Materials 188 (2014) 126–132 127


is important to recall the close correlation between transport diffu-sion and self-diffusion [15,28,29]: in this context we refer to a sec-ond definition of the self-diffusivity, namely a definition also basedon Fick’s 1st law, Eq. (7), where now, instead of an overall concen-tration gradient and an overall flux, one is considering concentra-tion gradients and fluxes of only the labeled molecules (i.e., of asubgroup of molecules), within an entity of molecules at uniformconcentration, i.e., under equilibrium. One may easily show (see,e.g., [15,25]) that both definitions are completely equivalent.

From the definition of the diffusivities by Fick’s 1st law, how-ever, one may also immediately conclude that the coefficients ofself-diffusion and transport diffusion approach each other if theguest–guest interaction is of only minor influence. This is the casefor, e.g., dominating host–guest interaction or at low guest concen-tration as attained, in general, at high temperatures.

By following the Maxwell–Stefan concept, the interdependenceof self- and transport diffusivity may, more generally, based on theunderstanding that during self-diffusion, i.e., during the process ofcounter diffusion of labeled and unlabeled molecules, the propa-gating molecules have to overcome the drag experienced by ‘‘fric-tion’’ with the host lattice and the counter-diffusing molecules. Byemploying the reciprocal values of the diffusivities as a measure ofthe respective resistances [15,29] one may thus note:

1D¼ 1

DT0þ 1

D�

ii

ð9Þ

DT0 denotes the so-called corrected diffusivity defined by therelation:

DT ¼ DT0d ln pd ln c

; ð10Þ

where d ln p=d ln c, the reciprocal value of the logarithmic deriva-tive of the equilibrium concentration with respect to the gas pres-sure, is referred to as the thermodynamic factor. Being unityunder ideal conditions ðc / pÞ, it is easily seen to represent an ex-tra-driving force for transport diffusion which must be eliminatedif one is concerned with only the drag exerted by the host lattice.D�

ii is referred to as the Maxwell–Stefan self-exchange diffusivity.Its reciprocal value stands for the drag experienced by the guest

molecules on passing each other. Both the thermodynamic factorand the Maxwell–Stefan self-exchange diffusivities are mainly con-trolled by the micropore space and are, hence, comparable for thebinderless and binder-containing molecular sieves considered inthis study. Thus, via Eqs. (9) and (10), increase in the self-diffusivi-ties is immediately seen to as well correspond with an increase inthe transport diffusivity. The behavior observed by consideringmolecular self-diffusion may thus quite generally be considered toalso characterize the situation under transport diffusion as relevantfor the technological application of the nanoporous materials understudy. This analogy, e.g., has been largely exploited for the produc-tion of transport-optimized nanoporous materials [30–32].

In general, and in particular for observation times in the transi-tion regime between intracrystalline and long-range diffusion asconsidered in this study, as well as for distributions in the moleculardiffusivities resulting from sample heterogeneities, the propagatoris much more complex than being expressible by the simple Gauss-ian, Eq. (2). In this case, also the PFG NMR signal attenuation deviatesfrom the simple exponential dependence as given by Eq. (4). Onemay easily show, however, (namely by expanding the cosine func-tion in Eq. (1)), that Eq. (4) may still serve as a reasonable approachfor c2d2g2 � 1=Dt, i.e., in the very first part of the PFG NMR signalattenuation. Now D has to be understood as an effective quantity de-fined by Eq. (3) which, unlike the diffusivity in a homogeneous med-ium, may become a function of the observation time. Throughoutthis study we shall be concerned with this type of ‘‘effective’’ diffu-sivities. In the limiting cases of genuine intracrystalline or long-range diffusion (i.e., for root mean square displacements much lar-ger or much shorter than the mean crystallite radii), these effectivediffusivities do clearly coincide with the genuine coefficients ofintracrystalline and long-range diffusion. For observation timeswithin these time regimes they are independent of the observationtime, as to be required for genuine normal diffusion.

We have performed our measurements with the homebuilt PFGNMR spectrometer FEGRIS 400 [33,34] using the 13-interval pulsesequence [35] for reducing the disturbing influence of internal fieldgradients on our measurements. During the measurement, thezeolite material is contained in closed sample tubes. For sampleactivation, the zeolite material was heated (10 K/h) under continu-ous evacuation up to 400 �C and left, under continued evacuation, at

Fig. 1. SEM images of the interior of the samples under study: binder-containing 4A (a), binderless 4A (b), binder-containing NaMSX (c) and binderless NaMSX (d) zeolitebeads. The (arbitrarily) chosen areas marked with broken lines (7 lm � 7 lm) reappear in Fig. 4 where they may serve as a standard for correlating the recorded moleculardisplacements with the spatial dimensions of the host systems under study.

128 D. Mehlhorn et al. / Microporous and Mesoporous Materials 188 (2014) 126–132


the final temperature for additional 10 h. Subsequently, the mate-rial is loaded with a well-defined amount of the desired guest mol-ecule (in our case water). This is accomplished by chilling the guestmolecules from a calibrated gas volume by means of liquid nitro-gen. For ensuring direct comparability between the different sam-ples, the total amount of guest molecules in the samplesconsidered in this study was chosen to correspond to one half ofmaximum pore filling (data of Table 1).

3. Results and discussion

As a typical example of the primary data determined in ourstudies, Fig. 2 shows the PFG NMR signal attenuations recordedat 80 �C with a binder containing (Fig. 2a) and a binderless(Fig. 2b) zeolite NaMSX for different diffusion times. Signal attenu-ation in the binderless sample is found to be much more pro-nounced than in the binder-containing sample. With Eq. (4), themore pronounced signal attenuation is easily seen to correspondwith higher diffusivities and, correspondingly, larger displace-ments. It is noteworthy that for the binderless sample the slopeof the first part of the attenuation curves is seen to remain essen-tially unaffected with varying observation time. Since this verypart of the attenuation curves determines the effective diffusivi-ties, the effective diffusivities are thus also found to be indepen-

dent of time – indicating that in this case we are recordinggenuine long-range diffusion, i.e., this process which determinesabout the exchange rate between the molecular sieves and theirsurroundings and, hence, about their performance.

In the binder-containing molecular sieve (Fig. 2a), however,decreasing slopes indicate that also the diffusivities decrease withincreasing observation time, revealing the existence of a hierarchyof transport resistances leading to a progressive inhibition ofmolecular propagation.

Fig. 3 summarizes our diffusion data in an Arrhenius plot. Thediffusivities determined for an observation time of 10 ms are indi-cated by large symbols. In addition, selected examples are providedfor illustrating how the diffusivities were found to depend on theobservation times. In most cases the variation in the diffusivitieswith the observation time was found to lead to changes whichwere still within the limits of accuracy of the measurements (beingof the order of the size of the large symbols used in our represen-tation). There were, however, also cases with diffusivities notablydecreasing with increasing observation time (as exemplified, e.g.,by Fig. 2a with the diffusivity data given in Fig. 3 for water in bin-der-containing and binderless NaMSX at 80 �C).

Also indicated in Fig. 3 are the (apparent) activation energiesEapp which result by approaching the temperature dependence ofthe effective diffusivities by a linear Arrhenius plot (brokenstraight lines through the data points for the different samples inFig. 3):

Deff ¼ D0 expð�Eapp=RTÞ ð11Þ

The uncertainty in the activation energies takes account of boththe uncertainty in the diffusivity data and their dependence on theobservation time. It is interesting to note that these activationenergies are intermediate between the activation energies of intra-crystalline diffusion as obtained, e.g., for water in zeolite NaMSX(EDiff-intra � 20 kJ mol�1 [36,37]) and the (isosteric) heat of adsorp-tion (Eiso = 60–70 kJ mol�1 [4,38,39]). Within the closed NMR sam-ple tubes, under the conditions of gas phase adsorption (i.e., forpinter� 1), the relative amount of molecules in the intercrystallinespace is easily seen to follow the Boltzmann relation:

(a)

(b)

Fig. 2. PFG NMR signal attenuation curves observed for water at half saturation fordifferent observation times at 80 �C in a binder-containing (a) and binderless (b)NaMSX type molecular sieve.

Fig. 3. Arrhenius plots of the effective diffusivities of water in the binder-containing (full symbols) and binderless (empty symbols) specimens of 4A(squares)- and NaMSX (circles) type molecular sieves at half saturation. Largesymbols represent the data points determined for an observation time of t = 10 ms.The (in general modest) variation of the diffusivities with varying observation timeis indicated by the bars, with their upper and lower ends indicating the diffusivitiesfor the smallest (t = 3 ms) and largest (t = 30 ms) observation times, respectively.The uncertainty in the effective activation energies (resulting as the best approachof Eq. (9) to the experimental data) takes account of both the uncertainty in themeasurement (represented by the size of the symbols) and the variation of thediffusivities with the observation time (bars).



Pinter � expð�Eiso=RTÞ: ð12Þ

Under the conditions of Knudsen diffusion, commonly impliedfor diffusion in the intercrystalline space [15,39], one has:

Dinter ¼23tR; ð13Þ

where t and R denote the mean thermal velocity and the mean poreradius, respectively. Dinter is thus seen to vary with only the squarerroot of the temperature. Hence, with Eq. (5), the temperaturedependence of Dlr is seen to be essentially given by that of pinter.Thus, Eiso can be considered as a good measure of the activation en-ergy of long-range diffusion within a bed of particles/crystallites iflong-range mass transfer is only possible via periods of mass trans-fer through the free space between the individual crystallites.

From the fact that the measured activation energies are in be-tween the activation energy of genuine intracrystalline diffusionand the heat of adsorption one has to conclude [40,41] that thereare essentially two types of trajectories of the water molecules,namely trajectories consisting of periods interchanging betweenthe intra- and intercrystalline spaces and trajectories exclusivelyextended in the solid phase formed by the individual crystallitesand their interconnection by zeolite bridges.

For assessing the relevance of this conclusion it is important tocorrelate the spatial dimensions of the systems under study withthe diffusion path lengths probed in our PFG NMR measurements.Such a comparison is given by Fig. 4. In this representation, SEMimages of the different samples (left-hand side, regions markedin Fig. 1 by the broken lines) are shown in the same scale as used,on the right hand side, for plotting the root mean square displace-ments of the water molecules as a function of the observation time.The data on the right-hand side of this representation are, via Eq.(3), clearly interrelated with the data of Fig. 3. With the possibilityof their direct correlation with the spatial extensions of the hostsystem, they are thought to serve for a better visualization of thephenomena under study.

We start with recalling that the displacements plotted on theright-hand side of Fig. 4 are average values taken over all watermolecules within the sample. It is interesting to note that, on con-sidering the evolution of the displacements with increasing obser-vation time, one does not become aware of a dramatic deviationfrom a linear dependency. This corresponds with the fact thatthe determined diffusivities were found to vary – if at all – not dra-matically with the observation time. This tendency does, in partic-ular, concern the evolution of the molecular mean displacementsfrom below up to larger than the mean crystallite radii. At a firstglance, this observation may be astonishing: For mean diffusionpath lengths notably smaller than the crystal diameters, only a rel-atively small fraction of molecules ‘‘benefit’’ from the much largermobilities in the intercrystalline space and one might expect a dra-matic change if, with increasing observation time, a notably largerfraction of molecules is able to leave the micropores. Such a nota-ble increase in the overall diffusivities, however, is not observed.This, obviously, is related to confinement effects within both theindividual zeolite particles and the interparticle space.

In both the A- and X-type molecular sieves, water diffusivitiesin the binderless beads (Fig. 4f and h) are seen to notably exceedthose in the binder-containing beads. This difference may be pre-dicted to become even larger by taking into account that – forthe highest temperatures and observation times considered – inboth binderless beads molecular propagation is seen to have at-tained already the stage of normal diffusion while the less than lin-ear increase observed with the binder-containing specimensindicates that, over the experimentally accessible range of observa-tion times, the diffusion path ways are not yet long enough to cov-er all resistances which they have to overcome during the process

of long-range diffusion. This difference was exemplified already bythe PFG NMR attenuation curves shown in Fig. 2, with coincidingslopes for the binderless specimens and anotable decrease withincreasing observation time in the binder-containing ones. By thevery nature of the compacting process, in the binderless beads do-mains of purely microporous zeolitic structure are connected byzeolite bridges which may be expected to contribute to the overalllong-range diffusion in a much more efficient way than the bridgesof non-zeolitic structure connecting the microporous domains inthe binder-containing beads. With the notation of Eq. (6), the ori-gin of this type of transport enhancement is obviously correlatedwith an increase in the ‘‘effective’’ intracrystalline diffusivityDintra,eff which, for binderless beads, may thus be expected to belarger than in the binder-containing specimens.

By inspecting the pore data presented in Table 1, however, alsothe second term in Eq. (6), pinterDinter, may be identified as beingable to potentially contribute to the observed transport enhance-ment in binderless molecular sieves. It is true that, due to theslightly larger macroporosities of the binder-containing NaMSX-type molecular sieves (31.1%, see Table 1) in comparison withthe binderless specimens (26.4%), the relative amount pinter of mol-ecules in the intercrystalline space of the binder-containing speci-mens is expected to exceed the value for the binderless specimenby about 20%. With Eq. (13) and with the data given in Table 1for the pore radii, however, this difference is easily seen to be morethan compensated by the influence of the second factor, Dinter. Thenotably larger mean pore radius in the binderless beads is found togive immediately rise to larger intercrystalline diffusivities and,hence, to an enhancement of the long-range diffusivity by alsothe second term in Eq. (6).

4. Summary and conclusions

As a non-invasive technique for the in situ observation of masstransfer in nanoporous solids, PFG NMR diffusion measurementsare ideally suited for assessing the transport properties of suchmaterials. The present study was dedicated to an in-depth studyof the diffusivity of water in specimens of NaMSX- and 4A-typezeolites aiming, in particular, at an exploration of possible differ-ences between inorganic binder containing and binderless molec-ular sieves beads. For both the zeolite NaMSX- and 4A-typespecimens considered in this study, the binderless shapes werefound to give rise to transport enhancement, which could be quan-tified by an increase in the water diffusivities up to factors of 2–3.

The diffusion path lengths considered in these measurementsdid exceed the size of the individual crystallites. They represent,therefore, the long-range diffusivities Dlr which are responsiblefor the rate of uptake from (or release to) the surrounding atmo-sphere, which (see Eq. (8)) can therefore be expected to be by thisvery factor of 2–3 larger in the binderless species. Since exactlythese rates of molecular uptake and release are often rate deter-mining for the performance of such materials, e.g., in drying orthermal energy storage processes, it is also the output of trans-port-optimized molecular sieves, which contributes to the techno-logical relevance of binderless production routes. Since thediffusivity enhancement in binderless molecular sieves observedin this study can be correlated with special features of the inter-crystalline pore space, notably with an enhancement of the meanmacropore diameters, it should be expected to hold quite generallyincluding, in particular, industrial gases like nitrogen and carbondioxide.

In summary, by avoiding the use of any inert component, thespecial manufacturing process used for the fabrication of binder-less molecular sieve bodies does, obviously, not only ensure theformation of 100% active adsorption matter with performance



enhancement owing to a corresponding increase in adsorptioncapacity. It does also contribute to the observed transport

enhancement. A first microstructural origin of transport enhance-ment may be related to the enhanced efficiency of mass transfer

Fig. 4. SEM images of binder-containing (a) and binderless (b) 4AK and of binder-containing (c) and binderless (d) NaMSX beads and mean diffusion path lengths (root meansquare displacements) of water in these materials (e–h) plotted as a function of the observation time for various temperatures. The length scales of the SEM images and of thediffusion path lengths coincide.



through the zeolite bridges connecting the individual microcrystal-line domains (crystallites) in the binderless specimens, as a conse-quence of the higher flux densities brought about by theirmicroporosity, in comparison with their binder-containing coun-terparts. Second, the very type of the formation of the molecularsieves realized in their binderless fabrication, namely their mutualconnection by a zeolite phase, does, obviously, give rise to a muchmore open (macro)pore space, with a correspondingly reducedtransport resistance, than realized by a binder-mediated shaping.

Acknowledgements

The authors gratefully acknowledge Fraunhofer IKTS Hermsdorffor SEM and Hg porosimetry, Deutsche Forschungsgemeinschaftand Fonds für Chemische Industrie for financial support.

References

[1] N.Y. Chen, T.F. Degnan, C.M. Smith, Molecular Transport and Reaction inZeolites, VCH, New York, 1994.

[2] D.M. Ruthven, S. Farooq, K.S. Knaebel, Pressure Swing Adsorption, VCH, NewYork, 1994.

[3] F. Schüth, K.S.W. Sing, J. Weitkamp, Handbook of Porous Solids, Wiley-VCH,Weinheim, 2002.

[4] D.W. Breck, Zeolite Molecular Sieves, John Wiley & Sons, New York, 1974.[5] D.M. Ruthven, Principles of Adsorption and Adsorption Processes, Wiley, New

York, 1984.[6] S.M. Auerbach, K.A. Carrado, P.K. Dutta (Eds.), Handbook of Zeolite Science and

Technology, Basel, New York, 2003.[7] V. Valtchev, S. Mintova, Microporous Mesoporous Mater. 43 (2001) 41–49.[8] A.G. Dong, Y.J. Wang, Y. Tang, Y.H. Zhang, N. Ren, Z. Gao, Adv. Mater. 14 (2002)

1506–1510.[9] W.C. Li, A.H. Lu, R. Palkovits, W. Schmidt, B. Spliethoff, F. Schüth, J. Am. Chem.

Soc. 127 (2005) 12595–12600.[10] K. Shams, S.J. Mirmohammadi, Microporous Mesoporous Mater. 106 (2007)

268–277.[11] K. Schumann, B. Unger, A. Brandt, F. Scheffler, Microporous Mesoporous Mater.

154 (2012) 119–123.[12] R. Kimmich, NMR Tomography, Diffusometry, Relaxometry, Springer, Berlin,

1997.[13] W.S. Price, NMR Studies of Translational Motion, University Press, Cambridge,

2009.

[14] P.T. Callaghan, Translational Dynamics and Magnetic Resonance, OxfordUniversity Press, Oxford, 2011.

[15] J. Kärger, D.M. Ruthven, D.N. Theodorou, Diffusion in Nanoporous Materials,Wiley-VCH, Weinheim, 2012.

[16] D. Mehlhorn, R. Valiullin, J. Kärger, A. Brandt, B. Unger, Chem. Ing. Tech. 83(2011) 2251–2259.

[17] D. Mehlhorn, R. Valiullin, J. Kärger, K. Cho, R. Ryoo, Materials (open access) 5(2012) 699–720.

[18] J. Kärger, M. Bülow, B.R. Millward, J.M. Thomas, Zeolites 6 (1986) 146–150.[19] F. Hibbe, C. Chmelik, L. Heinke, S. Pramanik, J. Li, D.M. Ruthven, D. Tzoulaki, J.

Kärger, J. Am. Chem. Soc. 133 (2011) 2804–2807.[20] D.S. Sholl, Nat. Chem. 3 (2011) 429–430.[21] K. Schumann, A. Brandt, B.S.F. Unger, in: M. Hartmann, W. Schwieger, S.

Megelski (Eds.), 23. Deutsche Zeolith-Tagung: Book of Abstracts, DECHEMA,Frankfurt, 2011, pp. 47–48.

[22] K. Schumann, A. Brandt, B. Unger, Chem. Ing. Tech. 83 (2011) 2237–2243.[23] J. Kärger, W. Heink, J. Magn. Reson. 51 (1983) 1–7.[24] J. Kärger, R. Valiullin, in: R.K. Harris, R.E. Wasylishen (Eds.), Encyclopedia of

Magnetic Resonance, John Wiley, Chichester, 2011, http://dx.doi.org/10.1002/9780470034590.emrstm0121.pub2.

[25] J. Kärger, Leipzig, Einstein, Diffusion, second ed., Leipziger Universitätsverlag,Leipzig, 2010.

[26] R.M. Barrer, Zeolites and Clay Minerals as Sorbents and Molecular Sieves,Academic Press, London, 1978.

[27] L. Gueudré, T. Binder, C. Chmelik, F. Hibbe, D.M. Ruthven, J. Kärger, Materials(open access) 5 (2012) 721–740.

[28] D.M. Ruthven, in: H.G. Karge, J. Weitkamp (Eds.), Adsorption and Diffusion,Springer, Berlin, Heidelberg, 2008, pp. 1–43.

[29] R. Krishna, J.M. van Baten, Microporous Mesoporous Mater. 109 (2008) 91–108.

[30] J. Kärger, AIChE J. 28 (1982) 417–423.[31] J. Kärger, H. Pfeifer, F. Stallmach, M. Bülow, P. Struve, R. Entner, H. Spindler, R.

Seidel, AIChE J. 36 (1990) 1500–1504.[32] J. Kärger, Ind. Eng. Chem. Res. 41 (2002) 3335–3340.[33] P. Galvosas, F. Stallmach, G. Seiffert, J. Kärger, U. Kaess, G. Majer, J. Magn.

Reson. 151 (2001) 260–268.[34] F. Stallmach, P. Galvosas, Ann. Rep. NMR Spectrosc. 61 (2007) 51–131.[35] R.M. Cotts, M.J.R. Hoch, T. Sun, J.T. Markert, J. Magn. Reson. 83 (1989) 252–266.[36] J. Kärger, Z. Phys. Chem. (Leipzig) 248 (1971) 27–41.[37] H. Pfeifer, J. Kärger, A. Germanus, W. Schirmer, M. Bülow, J. Caro, Ads. Sci.

Technol. 2 (1985) 229–239.[38] N. Avgul, A.V. Kiselev, Y.V. Mirskii, M. Serdobov, Chemistry (USSR) 42 (1968)

768.[39] J. Kärger, P. Volkmer, J. Chem. Soc. Faraday Trans. I 76 (1980) 1562–1568.[40] P. Zeigermann, J. Kärger, R. Valiullin, Microporous Mesoporous Mater. 178

(2013) 84–89.[41] D. Mehlhorn, R. Valiullin, J. Kärger, K. Cho, R. Ryoo, Microporous Mesoporous

Mater. 164 (2012) 273–279.


Transport enhancement in binderless zeolite X- and A-type molecular sieves revealed by PFG NMR diffusometry

Documents