Powder Technology 203 (2010) 389–396

Contents lists available at ScienceDirect

Powder Technology

j ourna l homepage: www.e lsev ie r.com/ locate /powtec

Optimized experimental design for natural clinoptilolite zeolite ball milling toproduce nano powders

Amir Charkhi a, Hossein Kazemian b,c,⁎, Mohammad Kazemeini a

a Chemical & Petroleum Eng. Department, Sharif University of Technology, Tehran, Iranb SPAG Zeolite R&D Group. Technology Incubation Centre, Science and Technology Park of Tehran University, Tehran, Iranc Department of Chemical and Biochemical Engineering, Faculty of Engineering, The University of Western Ontario, London, Ontario, Canada N6A 5B9

⁎ Corresponding author. Department of ChemicalFaculty of Engineering, The University of Western OntN6A 5B9. Tel.: +1(519) 661 2111x82209.

E-mail addresses: Hossein.kazemian@uwo.ca, hossei(H. Kazemian).

0032-5910/$ – see front matter © 2010 Elsevier B.V. Aldoi:10.1016/j.powtec.2010.05.034

a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 June 2009Received in revised form 11 December 2009Accepted 27 May 2010Available online 4 June 2010

Keywords:Experimental designPlanetary ball millNatural zeolite nano powderClinoptilolite

Nano powder of natural clinoptilolite zeolite was mechanically prepared by using a planetary ball mill.Statistical experimental design was applied to optimize wet and dry milling of clinoptilolite zeolite. Todetermine appropriate milling conditions with respect to the final product crystallinity, particle size anddistribution, different milling parameters such as dry and wet milling durations, rotational speed, balls topowder ratio and water to powder ratio (for the wet milling) were investigated. Laser beam scatteringtechnique, scanning electron microscopy and X-ray diffraction analyses were carried out to characterizesamples. Results showed that larger than 1 mm particle size of clinoptilolite powder may mechanically bereduced into the size range of less than 100 nm to 30 μm by means of planetary ball milling.

and Biochemical Engineering,ario, London, Ontario, Canada

nkazemian@gmail.com

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Zeolites are valuable inorganic materials having wide variety ofapplications including; molecular sieves, adsorbents, ion-exchan-gers and catalysts. These materials are microporous crystallinehydrated alumino-silicates composed of TO4 tetrahedral (T=Si,Al) with O atoms connecting neighboring tetrahedral in which thesubstitution of silicon by aluminum in framework positions leavesa negative charge behind which in turn may be compensated bysome alkali or alkaline earth cations[1–4].

Nowadays, more than 150 different types of zeolites have beensynthesized while more than 50 types have been discovered in thenature. Amongst natural zeolites; clinoptilolite, with the simplifiedideal formula of (Na, K)6Si30Al6O72 ∙ nH2O; is one of the mostcommonly found mainly in sedimentary rocks of volcanic origin.Clinoptilolite-rich tuff is commercially very favorable due to itshuge and easy mineable resources as well as, its high zeolitecontent. This well known zeolite may be utilized for purificationand separation processes, removal of NH4

+ and heavy cations fromcontaminated water and wastewater, aquaculture, soil fertilizers

and conditioners as well as, for dietary supplement in animalnutrition[4–10].

Recently there has been a considerable growing interest inutilizing nanozeolites due to their advantages over conventionalmicron sizedmaterials. In other words, the reduction of the particlesize of zeolites causes larger external surface areas available forinteraction, shorter diffusion path lengths reducing mass and heattransfer resistances in catalytic and sorption applications, decreas-ing of side reactions, enhancing selectivity as well as, loweringtendencies to coke formation in some catalytic reactions. Up tonow many different methods for synthesis of nanozeolites havebeen reported. All of these are based upon hydrothermal treat-ments which is performed by adjusting effective parameters suchas temperature, process duration and ingredient concentrations inorder to increase number of durable nuclei and reduce crystalgrowth [1]. It is reminded, however, that longer synthesis duration,expensive starting materials (specially organic template), lack ofreproducibility of the synthesis processes and energy consumptionfor separation of nano powders by mean of high speed centrifu-gation are some main issues making the bottom-up chemicalsynthesis of nanozeolites techno-economically unviable processes[11]. As an alternative technique, the zeolite particle size may bereduced mechanically using specially designed ball mills [12–17].In previous researches, production of different synthetic zeolitessuch as Y, X, A, L, ZSM-5 and mordenite was considered utilizingthis method [12–17] where possible changes of milled zeolitecharacteristics when subjected to dry ball milling were well

Table 1Variable parameters and their level in designed experiments in this study.

Level Millingspeed(rpm)

Balls topowderratio in wetmilling (wt.%)

Ball to powderratio in drymilling( No./weight)

WetMillingperiod oftime (h)

Drymillingperiod oftime (min)

Water toPowderratio(vol/weight)

1 450 4.5 0.1 2 10 12 500 5 0.2 3 20 1.23 550 9 4 1.54 600 3

390 A. Charkhi et al. / Powder Technology 203 (2010) 389–396

investigated. However, results have ascertained that high energyball milling, decreased the size of aforementioned zeolites.Furthermore, the XRD patterns revealed that the crystallinity ofthe milled zeolites was also reduced. Thus, collapse of the zeolitecrystal structure renders them useless as molecular sieves,adsorbents or shape selective catalysts. In order to overcome thiscrystallinity problem, wet ball milling of the HY zeolite wasinvestigated. Previous results indicated that by omitting the drymilling step, the crystalline structure of the ground HY zeolite didnot collapse completely, even at long milling durations [13].Furthermore, higher energy efficiencies, lower magnitude ofexcess enthalpies and elimination of dust formation may also bementioned as some other added advantages of grinding in aqueouscompared to dry medium [13].

In order to evaluate the possibility of production of naturalclinoptilolite nano-particles, in this study a combination of wetand dry milling was investigated using a planetary ball mill. Tooptimize grinding conditions, effective parameters were selectedand several sets of experiments were designed based uponstatistical methods. Zeolites were milled at different periods oftime, milling speed, balls to powder and water to powder ratios.Characteristics of final products such as particle size and distribu-tion as well as, crystallinity of samples were examined by Laserparticle size analyzer, scanning electron microscopy (SEM) and X-ray diffraction (XRD) techniques.

2. Experimental section

2.1. Ball milling of clinoptilolite zeolite

Powders of natural zeolite (clinoptilolite-rich tuff) wereobtained from a mine located in north of Iran near the city ofSemnan. Ball milling of clinoptilolite zeolite was performed bymean of a planetary ball mill (PM100; Retsch Corporation). Tooptimize the milling conditions with respect to size reduction andcrystallinity retention, milling parameters such as rotationalspeed, ball to powder and water to powder ratios as well as,grinding time were varied for different experiments. In most of the

Table 2Different conditions of designed experiments in this research.

No. code Dry milling

Zeolite powder amount(g)

BallNo.

Rotational speed(rpm)

Duration(min)

1 322–3311 50 10 550 202 422–4311 50 10 600 203 322–3313 50 10 550 204 422–3313 50 10 600 205 311–2221 100 10 550 106 000–1322 0 0 0 07 212–1233 50 10 500 108 312–1234 50 10 550 109 311–1213 100 10 550 1010 321–2113 100 10 550 20

performed tests a clinoptilolite powder with particle size of largerthan 1 mm was utilized as the starting material in dry milling for atime period of about 10–20 min. The wet milling in water mediawas then carried out at different periods of time between 2 to 4 h.All of the tests were done in a 250 ml stainless steel jar withprotective jacket of zirconium oxide. Zirconium oxide balls of 20and 3 mm were utilized for dry and wet millings; respectively. Thegrinding jars were arranged eccentrically on the sun wheel of theplanetary ball mill. The direction of movement of the sun wheelbeing opposite to that of grinding jars was selected with the ratioof 1:1. A certain amount of zeolite and balls as well as, water in wetmilling were placed in the jar at room temperature and atmo-spheric pressure then sealed and imposed to milling. Due to lack ofappropriate accessories to control the temperature and pressure ofthe jar during grinding , sampling was carried out at the end of thisperiod, at which time the jar was allowed to be cooled down toroom temperature. For characterization step the ground powderswas dried at 30 °C for 24 h.

2.2. Design of experiments

Different sorts of experiments were designed in order tooptimize the appropriate milling conditions for the production ofnatural clinoptilolite nano-powders with higher crystallinity andlower particle size and distributions. Selected variable parametersand their levels are provided in Table 1. For convenience, theshorthand nomenclature of a, b, c, d, e, f, and g were assigned inwhich, (a) the dry milling speed, (b) the dry milling time, (c) theball to powder ratio for dry state, (d) the wet milling speed, (e) thewet milling time, (f) the balls to powders ratio for the wet state andfinally (g) the water to powder ratio; were set. Moreover, differentconditions for each experiment are presented in this table.

2.3. Characterization

The ground clinoptilolite zeolite was characterized with differ-ent instrumental techniques. The morphology of the groundpowders as well as its size studied by means of Scanning ElectronMicroscopy utilizing a LEO 1455vp SEM instrument. XRD patternsto evaluate the crystallinity of the ground powders determined bymean of a STOE STAD-MPDiffractometer inwhich a copper target at40kV and 30 mA (2 hb10_) followed. Furthermore, Particle sizemeasurements of the ground samples performed by laser beamscattering technique through means of a Master sizer 2000apparatus (MALNERN Instruments).

3. Result and discussion

In order to investigate the effect of different ball milling condi-tions on size distribution of the ground clinoptilolite zeolite,

Wet milling

Zeolite amount(g)

Balls weight(g)

Rotational speed(rpm)

Water volume(cm3)

Duration(h)

50 225 550 50 450 225 600 50 450 225 550 75 450 225 550 75 4

100 500 450 100 3100 500 450 120 450 450 450 50 350 450 450 150 3

100 450 450 150 3100 450 500 150 2

Fig. 1. Volume particle size distribution at different milling condition (experiments 1 and 3; see Table 2.).

Fig. 3. Variation of volume weighted mean size in designed experiments.

391A. Charkhi et al. / Powder Technology 203 (2010) 389–396

volume mean size and the size span were studied. The span is ameasure of the size distribution, which may be defined as follow:

Span = d 0:9ð Þ−d 0:1ð Þð Þ=d 0:5ð Þ ð1Þ

where d(0.1), d(0.5) and d(0.9) represent sizes for which 10, 50 and90% by volume of particles in ground samples are smaller than thesesizes, respectively. It is to be noted that smaller span corresponds tonarrower size distribution.

The particle size distribution of ground zeolite powders milledaccording to conditions provided in Table 2 and measured bymeans of laser beam scattering demonstrated in Fig. 1. By utilizingthese results, the span was calculated and shown in Fig. 2. Thevolume weighted mean size and d(0.1) are also shown in Figs. 3

Fig. 2. Variation of span in designed experiments. Fig. 4. Variation of d(0.1) in designed experiments.

392 A. Charkhi et al. / Powder Technology 203 (2010) 389–396

and 4; respectively. The d(0.1) results indicate that throughutilizing planetary ball mill, production of submicron powders ispossible and at different ball milling conditions this value may bevaried between 0.56 and 2 μm for natural clinoptilolite zeolite withinitial size of larger than 1 mm. The volume mean sizes of thesesamples were also changed in the range of 3.3 to 15 μm.

In order to evaluate the influence of each factor on the volumeweighted mean size and span, an arithmetic calculation was

Fig. 5. Influences of milling conditions

carried out. The average effect of each level of variable parameterswas calculated by averaging the results of those experimentswhich were carried out at that level. For example, to calculate theaverage effect of second level of wet speed milling (500 rpm),results of all experiments which were performed at that rotationspeed under wet condition (results of 311-2221 and 321-2113experiments) were averaged. Results of these calculations aredisplayed in Figs. 5 and 6. Data in Fig. 5 indicated that an increase in

on volume weighted mean size.

393A. Charkhi et al. / Powder Technology 203 (2010) 389–396

rotational speed under dry conditions caused initially a decrease involume mean size of ground zeolite while at faster speed of600 rpm, this value increased again. Moreover, increasing of themilling time as well as, ball to powder ratio under dry conditionscaused an enhancement in the volume mean size of groundpowders. On the other hand, increasing of the rotational speedfrom 450 to 550 rpm under wet conditions, decreased the volumemean size; while, at a little bit higher speed of 600 rpm, the volume

Fig. 6. Influences of millin

mean size started to rise again. In addition, the reduction in waterto powder ratio caused lowering of volume mean size. Ultimately,form the data in Fig. 5, it is a foregone conclusion that the volumemean size increased due to the raised ball to powder ratio. It mayfurther be concluded that for production of clinoptilolite powderswith the smallest volume mean size, planetary ball milling shouldbe performed at 1) for dry conditions: i- milling speed of 550 rpm,ii- milling time of 10 min and iii- 5 balls of 20 mm per 50 g of

g conditions on span.

394 A. Charkhi et al. / Powder Technology 203 (2010) 389–396

powders and 2) for wet conditions: i- milling speed of 500 rpm, ii-ball to powder weight ratio of 4.5, iii- water to powder ratio of 1(e.g.; 50 g of water per 50 g of zeolite) and iv- milling time of 3 h.

It is emphasized through Fig. 6 that increased in milling speed,time and balls to powder ratio caused a raise in the span of groundpowders. Furthermore, it is seen that amongst parameters studied,the rotation speed variations had the most influence on this span.On the other hand, increase in speed and time of wet milling led toa wider span. From this figure, it also is understood that increasedball to powder and water to powder ratio under wet milling makesnarrower span.

To study probable changes of morphology for ground powdersafter ball milling as well as to confirm the PSA results, sampleswere subjected to SEM investigation. Some of the SEM images ofground samples are shown in Fig. 7. SEM images indicated thatalmost in all samples, zeolite powder with particles size less than100 nm may be recognized as a separated particle or in the form oflarger agglomerates. Moreover most particles have lost their initialoctahedral shape and converted into spherical, elliptical orirregular shapes. By careful considerations of SEM images, somecrystals with sharp edges and clean surfaces observed; which areabout 200 nm in size. Therefore, it may be concluded that carefulselection of milling conditions may result in production of nanoclinoptilolite zeolite with desirable crystalline structure. Never-theless, mechanical production of zeolitic nano-particles by meanof planetary ball mills may also reduce zeolite crystallinity.

As mentioned in a previous study [14], milling at high speed andprolonged time resulted in collapsed crystalline structure of

Fig. 7. SEM images of original and

ground zeolite. Therefore, apart from size distribution, thecrystallinity study of ground powders is also very importantwhen a choice for appropriate conditions for ball milling ofclinoptilolite samples has to be made. Patterns of XRD for allsamples were obtained and some of them are displayed in Fig. 8.Results showed the increase in time and rotational speed of millingcaused reduction in crystallinity. In the ground clinoptilolitesamples, major peak observed at 2θ equal to 22.49o. With regardsto the intensity ratio of aforementioned peak in original parentclinoptilolite to the ground samples, crystallinity reduction due tothe milling processes was calculated. Results are provided inTable 3.

In order to determine best milling conditions amongst thosementioned in Table 2 above, the following normalized expression;so-called objective function (O.F.), defined to give each experimenta weight for this purpose as a result. The biggest O.F., thus,indicated best milling conditions in which highest crystallinityretained while smallest volume mean size and narrowest sizedistribution also obtained. The weighting factor for each term ofthis expression selected according to its importance for theseauthors. It is given by:

O:F: = 1− Vol weighted MeanMax Vol weighted Mean

� �× 0:35 + 1− Span

Max Span

� �

× 0:15 + 0:5 × Crystalinity:

ð2Þ

ground clinoptilolite zeolite.

Fig. 8. XRD patterns of a) original clinoptilolite zeolite and b) milled clinoptilolitezeolite under condition of experiments 311-2221, c) 422-4311 and d) 321-2113.

395A. Charkhi et al. / Powder Technology 203 (2010) 389–396

As shown in Table 3, best milling conditions belonged toexperiment No. 7 with sample code of 311-2221. It meant thatchoosing following conditions for milling of clinoptilolite gives thebest result with respect to crystallinity and size distribution. Theseincluded; for dry milling i)milling speed of 550 rpm, ii) millingduration of 10 min and iii) balls to powder ratio of 0.1 and forwet milling; i) rotational speed of 500 rpm, ii) milling durationof 3 h, iii) ball to powder ratio of 4.5 and iv) water to powder ratioof 1.

To find the influence of different levels of milling parameter oncrystallinity, a simple arithmetic calculation similar to what wasdone previously for size distribution performed. Results showedthat the increase in milling time and speed under dry and wetconditions caused crystallinity to reduce. In addition, for main-taining the crystallinity, the second level of ball to powder andwater to powder ratios in wet milling had to be selected.

Table 3Milling results in terms of size distribution, crystallinity and defined objective function.

No. Exp. Code d(0.1) (μm) d(0.5) (μm) d(0.9) (μm) Volume w

5 311–2221 0.849 2.462 6.72 3.3096 000–1322 1.601 3.814 8.059 4.417 212–1233 1.199 7.935 20.03 9.5984 422–3313 0.559 4.514 20.496 8.0749 311–1213 1.448 8.093 22.91 10.76310 321–2113 0.746 6.687 20.911 9.0781 322–3311 0.8 5.347 23.37 9.4663 322–3313 0.675 7.944 31.418 12.6228 312–1234 1.904 8.728 35.628 15.0052 422–4311 0.85 4.65 38.265 13.378

a crystallinity is the ratio of major intensity of ground sample to original Clinoptilolite ze

Furthermore, results of experiment 000-1322 in Table 2 designedunder wet conditions indicated that omitting dry milling maycause an enhancement in crystallinity. This issue may be explainedin terms of the following mechanism. Different methods ofgrinding may impose different forces upon particles, such ascompression, shear, attrition, impact or internal forces [18]. In ballmilling, the grinding of samples is carried out by balls which areswept along by the wall and then fall onto the load. Thefragmentation of particles in samples results from the stress,which may be due to compression or shear, imposed upon the bedof particles by the balls. It has been proposed that wet grindingcauses shearing along the cleavage planes whereas dry grindingfractures the crystals [13,19]. It has also been suggested that wetgrinding proceeds with the preferential formation of new surfaceswhile little bulk deformation takes place in the particles [13,20].Thus, crystallinities of zeolite samples obtained by wet ball millingwere considerably larger than those prepared by the dry counter-part. On the other hand, amorphization of crystalline zeolites maybe attributed to breaking the structural Si–O–Si and Si–O–Al bondsand collapsing of the original crystal structure under the action ofintensive mechanical forces [14–17] mainly occurring at the drystage.

4. Conclusion

In this work the possibility of production of nano clinoptilolitezeolite by mean of a planetary ball mill utilizing a combination ofwet and dry conditions was investigated. Statistical experimentaldesign incorporated in order to optimize the effect of keyparameters. Results revealed that by application of appropriatemilling conditions obtained in this study, clinoptilolite powderswith particle sizes less than 100 nm and desirable crystallinitymay be produced. Nevertheless this method led to a wide sizedistribution of ground powders in the range of less than 100 nm to30 μm and crystallinity loss of about 55–100%. Furthermore, basedupon experiments performed under wet milling conditions inthis work, it may be concluded that under certain situationsone may go thru wet milling without any needs for dry millingpretreatments.

Acknowledgments

The Authors acknowledge Dr. Azizollah Kamalzadeh from theInstitute of Scientific-Applied Higher Education of Jihad_e_Agricultureof Iran and Animal Sciences Research Institute in Karaj, Iran, for pro-viding the ball mill equipment. The authors have declared no conflictof interest.

eighted mean size (μm) span Crystallinitya O.F. Normalized O.F.

2.384 0.45 0.604 1.0001.693 0.45 0.591 0.9792.373 0.38 0.419 0.6944.417 0.33 0.392 0.6492.65 0.38 0.387 0.6413.015 0.28 0.370 0.6124.221 0.30 0.351 0.5803.87 0.28 0.271 0.4493.864 0.38 0.265 0.4408.046 0 0.038 0.063

olite.

396 A. Charkhi et al. / Powder Technology 203 (2010) 389–396

References

[1] L. Tosheva, V.P. Valtchev, Chem. Mater. 17 (2005) 2494.[2] R.M. Milton, Molecular Sieves, Soc. of Chem. Ind. London (1968).[3] D.W. Breck, Zeolite Molecular Sieves: Structure, Chemistry and Uses, JohnWiley &

Sons, New York, 1974.[4] J.D. Sherman, Proc. Natl. Acad. Sci. (PNAS), USA 96 (7) (1999) 3471.[5] H. Kazemian, M.H. Mallah, Iran. J. Chem. Chem. Eng. 25 (4) (2006) 91.[6] A. Ansari Mahabadi, M.A. Hajabbasi, H. Kazemian and H. Khademi, Geoderma 137

(3–4) (2007) 388.[7] S.J. Shahtaheri, H. Kazemian, R. Menhaje-Bena, Studies in Surface Science and

Catalysis, 154 B, 2004, p. 1892.[8] H. Faghihian, H. Kazemian, Iranian J Sci. Technol. 26 (A2) (2002) 357.[9] H. Asilian, S.B. Mortazavi, H. Kazemian, S. Phaghiehzaadeh, S.J. Shahtaheri, M.

Salem, Iranian J. Public Health 33 (1) (2004) 45.

[10] F.A. Mumpton, P.H. Fishman, J. Anim. Sci. 45 (1977) 1188.[11] K.T. Matin, D. Bastani, H. Kazemian, Chem Eng. Technol. 32 (7) (2009) 1.[12] J. Xie, S. Kaliaguine, Appl. Catal.: A 148 (1997) 415.[13] K. Akcay, A. Sirkecioglu, M. Tather, O.T. Savasci, A. Erdem-Senatalar, Powder

Technol. 142 (2004) 121.[14] P.A. Zielinski, A. van Neste, D.B. Akolekar, Micropor. Mater. 5 (1995) 123.[15] C. Kosanovic, et al., Zeolites 15 (1995) 247.[16] A.S. Kharitonov, et al., Zeolites 15 (1995) 253.[17] C. Kosanovic, B. Subotic, A. Cizmek, Therm. Acta 276 (1996) 91.[18] C. Frances, N. Le Bolay, K. Belaroui, M.N. Pons, Int. J.Miner. Process. 61 (2001) 41.[19] G. Suraj, C.S.P. Iyer, S. Rugmini, M. Lalithambika, Appl. Clay Sci. 12 (1997) 111.[20] D. Feng, C. Aldrich, Int. J. Miner. Process. 60 (2000) 115.