Geopolymer with Hierarchically Meso-/Macroporous Structures from ReactiveEmulsion Templating

Dinesh Medpelli, Jung-Min Seo, and Dong-Kyun Seo†

Department of Chemistry and Biochemistry, Arizona State University, Tempe, Arizona 85287-1604

We present a simple synthetic route to hierarchically porous

geopolymers using triglyceride oil for a reactive emulsion tem-

plate. In the new synthetic method, highly alkaline geopolymerresin was first mixed with canola oil to form a homogeneous

viscous emulsion which was then cured at 60°C to give a hard

monolithic material. During the process, the oil in the alkalineemulsion undergoes a saponification reaction to be decomposed

to water-soluble soap and glycerol molecules which were then

extracted with hot water to finally yield porous geopolymers.

Nitrogen adsorption studies indicated the presence of mesop-ores, whereas the SEM studies revealed that the mesoporous

geopolymer matrix are dotted with spherical macropores

(10–50 lm) which are due to oil droplet template in the emul-

sion. Various synthetic parameters including the precursorcompositions were examined to control the porosity. BET sur-

face area and BJH pore volume of the materials were up

to 124 m2/g and 0.7 cm

3/g, respectively, and the total pore

volumes up to 2.1 cm3/g from pycnometry.

I. Introduction

OVER the past decades, geopolymers have receivedincreasing attention as an attractive ceramic material

due to the low-energy requirements in their production andtheir promising mechanical properties (compression strength,heat and chemical resistance, etc.).1–4 More recently, newresearch efforts have been geared into utilizing the materialfor nontraditional applications such as evaporative cooling,5

catalysis,6 and drug delivery.7 Success of such emergingapplications of geopolymer materials further requires exploringnew methods for controlling pore structures of the materialsin nanoscale. In this communication, we demonstrate that asimple reactive emulsion templating with biorenewable oilcan produce hierarchically porous geopolymer materials withcoexisting controllable mesopores and spherical macropores,without a need of significantly modifying the conventionalgeopolymer synthetic process.

Geopolymers are typically produced by dissolving solid alu-minosilicate precursors in a highly alkaline solution (typicallywith KOH or NaOH) to form a viscous solution (“geopolymerresin”) and subsequently curing the resin at ambient tempera-tures. Recent studies have shown that geopolymers are inher-ently a nanomaterial exhibiting a dense gel-like structure with5–40 nm-sized amorphous aluminosilicate particles.4,8,9 Theirchemical structure consists of an amorphous, three-dimen-sional network of corner-sharing aluminate and silicate tetra-hedra, with the negative charge due to Al3+ ions in the

tetrahedral sites balanced by the alkali metal ions.1,3,4 Figure 1shows schematic diagrams for the reactive emulsion templatingprocess employed in this work and for the final geopolymerproduct.10 Emulsions are droplets of one fluid (e.g., oil) dis-persed in a second immiscible fluid (e.g., water), which areoften stabilized by a surfactant.11 Mechanically induced drop-let breakup generates (meta)stable emulsions with a distribu-tion of droplet sizes. One novelty of the synthetic design in thiswork is that by employing a vegetable oil (mainly triglyce-rides12), mixing of a geopolymer resin with the oil generatescarboxylate surfactants (soap molecules) in situ through thesaponification reaction of the triglycerides with the highlyalkaline geopolymer resin (hence reactive). The excess oilforms oil droplets which are then embedded in the geopolymerresin. Notably, it has been found in our work that the oil inthe droplets continues to undergo saponification reaction dur-ing the curing of the mixture in our reaction condition, whichturns the originally hydrophobic triglycerides all into soap andglyceride (CH2(OH)–CH(OH)–CH2(OH)). Those moleculesare soluble in water and thus can be extracted by water fromthe cured solid material, resulting in a porous geopolymermaterial (Fig. 1(b), see Section III for details).

II. Experimental Procedure

(1) SynthesisIn the first step of the synthesis, a potassium silicate solutionwas prepared by dissolving an appropriate amount of KOHpellets (Sigma Aldrich, St. Louis, MO) in deionized water in apolypropylene cup in a water bath. A suitable amount offumed silica (Cabot, CA-BO-SIL� EH-5, Bellerica, MA) wasthen added into the KOH solution and the mixture was stir-red with an IKA (Wilmington, DE) mechanical mixer for30 min at 800 rpm to give a clear solution. The geopolymerresins were then prepared by mechanically mixing metakaoli-nite into the potassium silicate solution to form a homoge-nous fluidic liquid. The metakaolinite was produced inadvance by calcining kaolinite (Al2Si2O7�H2O, Alfa-Aesar,Ward Hill, MA) at 750°C for 10 h. Various samples were pre-pared with different water amounts and K/Al ratios but at afixed Si/Al ratio of 2 (Table I). The pH of the resins wasabout 14 for all the compositions. Canola oil (The J.M. Smuc-ker Company, Crisco�, Orrvile, OH), waste vegetable oil(REV biodiesel, Gilbert, AZ) or paraffin oil (Alfa Aesar) werethen added to the resin at a 1:1 oil-to-water volume ratio andmixed for an additional 15 min to give a homogeneous butviscous emulsion. The emulsion was transferred to a polypro-pylene cup and cured in a laboratory oven at 60°C for 24 h.

The cured product was then broken into small pieces(~1 cm 9 1 cm 9 1 cm) and subjected to extraction with hotdeionized water, except for S4 for which hexanes were used.Three series of samples were prepared as shown in Table I toinvestigate the effect of three synthetic parameters on theresulting geopolymer; (1) type of oil (S2, S3, S4, and S5), (2)mole fraction of water (S1, S2, and S7), and (3) amount of

F. Glasser—contributing editor

Manuscript No. 33469. Received July 4, 2013; approved October 23, 2013.†Author to whom correspondence should be addressed. e-mail: dseo@asu.edu

70

J. Am. Ceram. Soc., 97 [1] 70–73 (2014)

DOI: 10.1111/jace.12724

© 2013 The American Ceramic Society

Journal Rapid Communication

potassium hydroxide (S6, S7, and S8). Paraffin oil wasselected (S4) to examine the role of saponification, as paraffinoil is pure hydrocarbons and does not undergo a chemicalreaction with geopolymer resin. To produce a “control” sam-ple (R in Table I), the same synthetic procedure was followedwithout adding any oil.

(2) CharacterizationPowder X-ray diffraction (PXRD) patterns of the finelyground samples were collected using a Siemens D5000 dif-fractometer with CuKa radiation (Berlin, Germany). Car-bon–hydrogen–nitrogen (CHN) elemental analyses wereperformed by employing Perkin-Elmer 2400 Series II CHNS/O Analyzer (Waltham, MA) with a thermal conductivitydetector. Samples for scanning electron microscopy (SEM)were prepared by placing small pieces of the products(approximate cubes of few millimeters in length) on a SEMstub using a copper conducting tape. Samples were then goldcoated for 150 s and were studied using SEM-XL30 Environ-mental FEG (FEI, Hillsboro, OR) microscope operating at10 kV. For transmission electron microscopy (TEM), colloi-dal suspensions of ground samples in ethanol were dried onto copper grids and were studied using JEOL TEM/STEM2010F operating at 200 kV (Tokyo, Japan).

N2 sorption isotherms were obtained with a MicromeriticsASAP 2020 volumetric adsorption analyzer (Norcross, GA)at 77 K. Samples were degassed at room temperature for10 h under vacuum until a residual pressure of ≤10 lmHgwas reached. Specific surface areas were estimated using Bru-nauer–Emmett–Teller (BET) equation, in the relative pres-sure range from 0.06 to 0.2.13 Pore volumes were calculated

from the amount of nitrogen adsorbed at a relative pressure(p/po) of 0.99. Pore size distributions were obtained using theBarrett–Joyner–Halenda (BJH) method assuming a cylindri-cal pore model.14 Total pore volume of the products wasdetermined by pycnometry with deionized water at23°C � 2°C and ambient pressure, whose principle relies onthe permeation of water through the open pore network ofmonolithic solid samples.

III. Results and Discussion

Carbon–hydrogen–nitrogen analysis showed only smallamounts of carbon (0.5 � 0.3 wt%), hydrogen (1.3 � 0.3wt%), and nitrogen (0.005 � 0.002 wt%) in average for allsamples, S1–S8, with the maximum carbon content of 1.2 wt% found for S6. The values compare well with 0.83 wt%C,1.4 wt%H, and 0.005 wt%N for the sample R, which indi-cates that the hot-water extraction removed the organicsproperly. Figure 2 shows SEM and TEM images of the sam-ple S2 as a representative example. The material exhibits dis-crete spherical pores whose diameters range from about 5 to40 lm in Fig. 2(a). A closer look in Fig. 2(b) reveals that thepore wall separating the spherical pores has a finer structurethroughout the matrix. The corresponding TEM micrographsin Figs. 2(c) and (d) show the gel-like nanostructure of thematerial consisting of nanoparticles of about 20 nm that arestrongly fused by necks, which is consistent with previousresults.4,8 The materials were amorphous based on the largelyfeatureless “hump” centered at approximately 27°–30° in 2h,the unique feature of geopolymer, in their PXRD patterns(data not shown).1 The combination of the SEM, TEM, andXRD results surmises that the geopolymer products exhibit a

(a) (b)

Fig. 1. (a) Scheme for the reactive emulsion templating of geopolymer with canola oil and (b) schematic diagram of the resulting hierarchicallyporous geopolymer with a random mesoporous matrix dotted with spherical macropores. The objects in the figures are not scaled.

Table I. Pore Properties of the Porous Products from Various Synthetic Conditions

Sample

Mole

fraction of

H2O (x)

K/Al

ratio Oil used

Surface

area

(m2/g)

Pore

volume†

(cm3/g)

Average

pore

width‡ (nm)

Total

pore

volume§ (cm3/g)

Mesoporosity¶

(%)/total

porosityk (%)

S1 0.63 2 Canola 69 0.44 22 1.5 16/53R 0.68 2 No oil 62 0.16 7 0.69 5.1/22S2 0.68 2 Canola 97 0.53 17 1.7 19/58S3 0.68 2 Canola + paraffin (1/1; v/v) 42 0.40 41 1.7 15/47S4 0.68 2 Paraffin 5 0.03 34 1.1 0.88/31S5 0.68 2 Waste vegetable oil 123 0.37 14 1.6 15/65S6 0.73 1 Canola 55 0.30 17 1.1 13/49S7 0.73 2 Canola 124 0.61 18 1.7 20/53S8 0.73 3 Canola 84 0.70 34 2.1 23/67

†From the pores with width no larger than 150 nm in the BJH desorption pore distribution.‡4(BJH desorption pore volume)/(BET surface area).§Determined by pycnometry.¶From pore volume†.kFrom total pore volume§.

January 2014 Rapid Communications of the American Ceramic Society 71

mesoporous geopolymer matrix made of rather leisurely con-nected amorphous aluminosilicate nanoparticles and thatlarge spherical macropores are scattered over throughout themesoporous matrix [Fig. 1(b)]. It is reminded that the extrac-tion process removed the organic components completely,which strongly suggests that the mesopores in the geopoly-mer matrix are connected and open to allow the solvent andother molecules to flow in and out.

The samples were characterized further by applying BETand BJH analyses to N2 sorption isotherms for the samples,

and the results are summarized in Table I. Figure 3 showsthe isotherms and BJH desorption pore distribution of thesamples with the same precursor composition at K/Al = 2and x = 0.68 but with different types of oil (S2–S5) alongwith the control sample R prepared without oil. All the sam-ples except S4 show a noticeable hysteresis in their isotherms[Fig. 3(a)], indicating the presence of mesopores.15 The sizesof mesopores show a relatively narrow distribution inFig. 3(b) centered in the mesopore region (10–50 nm).Excluding S4, the sample R shows the lowest BJH cumulative

(a) (b)

(c) (d)

Fig. 2. Scanning electron microscopy images in (a) and (b) (scale bar = 50 and 2 lm, respectively) and transmission electron microscopy imagesin (c) and (d) of sample S2.

(a) (b)

Fig. 3. (a) Brunauer–Emmett–Teller isotherms and (b) BJH desorption pore size distribution curves of samples R1, S2, S3, S4, and S5. Allsamples have same composition but they differ in the type of oil (Table I) used in their preparation.

72 Rapid Communications of the American Ceramic Society Vol. 97, No. 1

pore volume (0.16 cm3/g) and the smallest average pore diam-eter (7 nm), whereas the ones prepared with oil containingtriglycerides (S2, S3, and S5) show a significantly higherporosity with the BJH pore volume up to 0.53 cm3/g and theaverage pore width up to 41 nm. The sample S4 preparedwith paraffin oil shows a BJH pore volume even lower thanthe control sample R.

The presence of mesopores (10–50 nm) indicated from thegas sorption studies is consistent with the textural poresamong the fused nanoparticles seen in Figs. 2(c) and (d). Itis noted that the samples maintained their monolithic featureduring the solvent extraction. When water was used forextraction, however, the original pieces broke into smallermonolithic particulates of about 3 mm in diameter, whichmight be due to the large capillary forces exerted by water inthe microcracks that developed during curing. The solidswere robust and did not lose their structural integrity duringsample handling and soaking for water pycnometry,although the quantification of the mechanical strength of thematerials is warranted in the future. Such a structural integ-rity is unusual for geopolymer materials with high totalporosities up to 67% observed in our case (Table I).Theporosity from mesopore structure (given as nanoporosity inTable I) is actually no greater than 23%, and hence thematrix itself possibly maintains its rigidity while the addi-tional spherical macropores increase the total porosity of thematerials.

The negligible nanoporosity (0.88%) for the sample S4 isintriguing because paraffin oil turned out in our experimentsto mix well with the geopolymer resin and could produceporous geopolymer (total pore volume = 1.1 cm3/g). DetailedSEM studies on S4 (data not shown) indicate that the mate-rial indeed exhibits the spherical macropores (20–50 lm) likeothers, but interestingly the pore walls show additional mac-ropores of about 2 lm instead of mesopores. It is suspectedthat the small macropores are open and connected together,as all the paraffin oil could be extracted out according tothe CHN analyses. In any event, the presence of the smallmacropores in S4 instead of the mesopores found in othersamples indicates that the saponification reaction does play asignificant role in pore formation probably by providing thein situ formed surfactant and also water-soluble glycerolbyproduct.

In addition to the oil type, the amount of water and K/Alratio are shown to control porosity of the products as well.With a fixed ratio of K/Al = 2 and canola oil, samples S1, S2,and S7 show an increase in the BJH cumulative pore volume0.44–0.61 cm3/g upon increasing the mole fraction of waterfrom 0.63 to 0.73, whereas their pore widths are more or lessthe same (Table I). Meanwhile, the increase in the K/Al ratioalso significantly increases the pore volume and pore width.With canola oil and x fixed at 0.73, the samples S6–S8, pre-pared with K/Al = 1, 2, and 3, show BJH cumulative pore vol-umes of 0.30, 0.61, and 0.70 cm3/g and the average porewidths of 17, 18, and 34 nm, respectively. The higher amountof KOH in the precursor solution may lead to a more extensivesaponification reaction, which in turn provides a higher poros-ity in the final product. Despite the excess amounts of KOH inthe precursor, it is worthy mentioning that all the productsshowed a neutral pH after the water extraction, indicatingagain that the pore structure is open for permeation of waterin the matrix, hence enabling the removal of the excess alkalinecomponent during the extraction. It is noted that the productswere found to keep their structural integrity and the original

porosity even after prolonged soaking in acidic solutions witha pH value as low as 3.

IV. Summary

We have demonstrated that a simple synthesis of hierarchi-cally porous geopolymers is possible by employing emulsiontemplating with triglyceride oil. The coexisting distinctivemesopores and macropores were characterized using the N2

sorption, SEM, TEM, and pycnometric studies. We have alsoshown that the pore size and/or volume can be controlled bychanging synthetic parameters such as oil type, and waterand alkali contents in precursor solution. Further studies aredue for elucidation of the precise role of those syntheticparameters and potentially others in controlling the porosityand also for quantitative examination of stability of this newclass of porous ceramics under various physical and chemicalstresses.

Acknowledgments

This work was supported by Mattium Corporation through NSF SBIR PhaseII (award no. 1152665). D. M.’s research assistantship was partially supportedby the Center for Bio-Inspired Solar Fuel Production, an Energy FrontierResearch Center funded by the US Department of Energy, Office of Science,Office of Basic Energy Sciences under award no. DE-SC0001016. We gratefullyacknowledge the use of facilities within the LeRoyEyring Center for SolidState Science at Arizona State University.

References

1J. Davidovits, “Geopolymers – Inorganic Polymeric New Materials,”J. Therm. Anal., 37 [8] 1633–56 (1991).

2P. Duxson, A. Fern�andez-Jim�enez, J. L. Provis, G. C. Lukey, A. Palomo,and J. S. J. van Deventer Duxson, “Geopolymer Technology: The CurrentState of the Art,” J. Mater. Sci., 42 [9] 2917–33 (2007).

3J. L. Provis, J. S. J. van Deventer, “Introduction to Geopolymers”; pp.16–26 in Geopolymers – Structure, Processing, Properties and Industrial Appli-cations, Edited by J. L. Provis, G. C. Lukey and J. S. J. van Deventer. Wood-head Publishing, Cambridge, UK, 2009.

4W. M. Kriven, “Inorganic Polysialates or ‘Geopolymers’,” Am. Ceram.Soc. Bull., 89 [4] 31–4 (2010).

5K. Okada, A. Ooyama, T. Isobe, Y. Kameshima, A. Nakajima, and K. J.D. MacKenzie, “Water Retention Properties of Porous Geopolymers for Usein Cooling Applications,” J. Eur. Ceram. Soc., 29 [10] 1917–23 (2009).

6P. Sazama, O. Bortnovsky, J. D�ede�cek, Z. Tvar�u�zkov�a, and Z. Sobal�ık,“Geopolymer Based Catalysts – New Group of Catalytic Materials,” Catal.Today, 164 [1] 16492–9 (2011).

7E. J€amstorp, M. Strømme, and G. Frenning, “Modeling Structure-Func-tion Relationships for Diffusive Drug Transport in Inert Porous GeopolymerMatrices,” J. Pharm. Sci., 100 [10] 4338–48 (2011).

8W. M. Kriven, J. L. Bell, and M. Gordon, “Microstructure and Micro-chemistry of Fully-Reacted Geopolymers and Geopolymer Matrix Compos-ites,” (Advances in Ceramic Matrix Composites IX) Ceram. Trans., 153

227–50 (2003).9J. L. Provis, G. C. Lukey, and J. S. J. van Deventer, “Do Geopolymers

Actually Contain Nanocrystalline Zeolites? A Reexamination of ExistingResults,” Chem. Mater., 17 [12] 3075–85 (2005).

10D.-K. Seo, D. Medpelli, and J.-M. Seo, “Porous Geopolymer Materials”;U.S. Patent No. US20130055924 A1, 2013.

11T. F. Tadros, “Emulsion Science and Technology: a General Introduc-tion”; pp. 1–56 in Emulsion Science and Technology, Edited by T. F. Tadros.Wiley-VCH Verlag, New York, 2009.

12F. D. Gunstone, The Chemistry of Oils and Fats: Sources, Composition,Properties, and Uses. Blackwell Publishing, Oxford, UK, 2004.

13S. Brunauer, P. H. Emmett, and E. Teller, “Adsorption of Gases in Multi-molecular Layers,” J. Am. Chem. Soc., 60 [2] 309–19 (1938).

14E. P. Barrett, L. G. Joyner, and P. P. Halenda, “The Determination ofPore Volume and Area Distributions in Porous Substances. I. Computationsfrom Nitrogen Isotherms,” J. Am. Chem. Soc., 73 [1] 373–80 (1951).

15J. B. Condon, Surface Area and Porosity Determinations by Physisorption:Measurements and Theory. Elsevier Science, Amsterdam, The Netherlands,(2006). h

January 2014 Rapid Communications of the American Ceramic Society 73